pe es em ~ b= ey ee mee peal gts meee apd: teen ay ee ae . Sele toes oe 1 ete ewe ene "ge fetetoes preg 4 - ye es wh ee Oh eae} s. * ae is a Ve ee Oe ghee ae er oa 7. Pai Ee aad re ia ieee ae sree ii Oe ete eal * aa eo * henge £4 3 = re + Petri aset 27 se ee ~~ ust oy ae y Pe Pe cat cee Sere: eros! ed Sa eee aes eee openers ek eRe eT aad oS a pcm “ ae ae are a Pte een NS rae ee res - hale me ne on Sie sh) etme 3h wer oY poet “293 Ga ee ne me on a. at neh th et tht tind Pas eee : ss ee os Ci gto Ceca eatin ne eee Spon fy peterh mas eepie Oe ee eee : 4 r abe ALBERT R. MANN -._ LIBRARY New York STATE COLLEGES OF AGRICULTURE AND HOME ECONOMICS ae SS —— A To . } A es ia AT CORNELL UNIVERSITY ‘iT COMPARATIVE ANATOMY OF THE PHANEROGAMS AND FERNS DE BARY Dondon HENRY FROWDE OXFORD UNIVERSITY PRESS WAREHOUSE ~~ AMEN CORNER COMPARATIVE ANATOMY OF THE VEGETATIVE ORGANS PHANEROGAMS AND FERNS BY DR A. DE BARY PROFESSOR IN THE UNIVERSITY OF STRASSBURG TRANSLATED AND ANNOTATED BY F. O. BOWER, M.A,, F.LS. LECTURER IN BOTANY AT THE NORMAL SCHOOL OF SCIENCE, SOUTH KENSINGTON AND D. H. SCOTT, M.A., Pu.D., F.LS. ASSISTANT TO THE PROFESSOR OF BOTANY IN UNIVERSITY COLLEGE, LONDON WITH TWO HUNDRED AND FORTY-ONE WOODCUTS AND AN INDEX @xtord AT THE CLARENDON PRESS 1884 [AW rights reserved | Li- TRANSLATORS’ PREFACE. In producing an English translation of the Comparative Anatomy of the Phanerogams and Ferns, by Professor De Bary, an attempt has been made to Meet two requirements, which have long been felt. In the. first place, those English students who do not read German will now gain access to the most exhaustive work hitherto published on that subject. Though, through unavoidable circumstances, a considerable interval has elapsed between the publication of the original and that of the translation, the book deals so largely with established facts, and in so much less a degree with matters of controversy, that the delay affects its value but little. The Translators have however inserted references to the more important memoirs published since the original was produced. In the second place, by means of this translation it is hoped that suitable English equivalents will have been supplied for numerous technical terms which have not hitherto been translated. Thus, in that part of the book which deals with the arrangement of the vascular bundles (pp. 232-315), the introduction of new English terms has been especially necessary, since this part of the science has not hitherto been treated at length in any English text-book. In conclusion, the Translators wish to record their thanks to Mr. W. T. Thiselton Dyer, Assistant Director of the Royal Gardens, Kew, and to Dr. S. H. Vines, Fellow and Lecturer of Christ’s College, Cambridge, for valuable advice and assistance; also to Mr. W. B. Hemsley, for the good judgment and care with which he has prepared the index. April 22, 1884. TO THE READERS. Tux present volume will complete the Handbook of Physiological Botany, which, since its commencement in 1865, has been edited by the late Prof. Hofmeister. As stated in the Preface to Vol. I, the plan of the book was drawn up in the year 1861, and the sections, which were to be treated according to the state of the science at that time, were distributed into four volumes, and among six contributors. Arrange- ments were made beforehand, so that the Volumes might be expected to appear in quick succession. Two of the contributors retired at the outset, so that, in the year 1866, the fourth volume having first appeared, and then the first section of the second, the programme in the Preface of Vol. I. assigned all the volumes to ‘four contributors, and those still remaining to three. Of these another subsequently retired. Nevertheless, the undertaking was not given up, the preparation of the remaining parts being undertaken by Hofmeister and by the author of this volume. At the beginning of last year Hofmeister was attacked by severe illness, to which he succumbed on the 12th of January of this year. After his death the question of the fate of the Handbook presented itself to the surviving contributors. Among the papers of the deceased were found, it is true, drafts and beginnings of the parts he had undertaken. But they have so much the character of incomplete sketches and fragments, that it was obvious to the undersigned that their publication would neither answer the purpose of the Handbook, nor the intentions of their author. The case standing thus, the remaining parts would have to be undertaken by another’ contributor. Supposing some one to be ready at once, he would have to begin on his own account at the beginning, and the continuation of the Handbook would at best be delayed for years. Nevertheless, if there were a real want, the attempt to continue it would be made. But, in the sixteen years which have passed since the Handbook was planned, the position of our Science has altered. In face of the Literature of to-day, a new work comprising ‘the Mor- phology of the Vascular Cryptogams,’ and ‘the Sexual Reproduction of the Phanerogams,’ can be dispensed with, while a separate treatment of the ‘Algz,’ as at first intended, is hardly possible. On these grounds it has been determined to close the Handbook. As it stands at present, it is arranged as follows :— Vol. I. rst Part. ‘Die Lehre von der Pflanzenzelle.’ By W. Hofmeister. and Part, ‘ Allgemeine Morphologie der Gewichse.’ By the same author. Vol. IJ. ‘Morphologie und Physiologie der Pilze, Flechten und Myxomyceten.’ By A. de Bary. Vol. III. ‘Vergleichende Anatomie der Vegetationsorgane der Gefisspflanzen.’ By the same author, Vol. IV. ‘Experimentalphysiologie der Pflanzen.’ By Julius Sachs. A. DE BARY. J. SACHS, SIRASSBURG AND WURzBURG, June, 1877. PREFACE. Tue preparation of the present volume was begun by the author in the year 1868, afler the other contributors to the Handbook, who had originally undertaken it, had retired. It was fairly far advanced, when in 1867, by reason of other necessary business, it had to be put entirely on one side for almost two years. It also suffered frequent and long interruption at a later time through changes in the official engage- ments of the author. The object of the work, as stated in the programme of the. Handbook, was an epitome of the present knowledge of ‘the Anatomy of the Vegetative Organs of Vascular Plants.’ From the very first, the necessity of numerous confirmatory in- vestigations was apparent, since the descriptions at hand were written at very different times, and by very different authors, and it was only possible to judge of and sift the differences necessarily present in these by actual and personal observation. This led to many researches of my own: new results and new questions appeared. The work soon extended itself beyond the limits originally intended. When one section was successfully finished, and others were in hand, new publications appeared which demanded fresh alteration of what had already been done. Therefore, in order that at least something might result, the necessity finally arose of bringing this work of the Danaids, this supplementary patching and correcting, to a definite conclusion, and of finally closing the work. This was done about three years ago. Since then nothing of importance has been done beyond finishing the revision. That such was the progress of the work may be some explanation and excuse for the frequent unevenness of the performance, Further, the, so to speak, forced hurry of the conclusion necessarily imposed limits. With regard to the contents, the exclusion in the first place of all Paleontology and Pathology, the latter including the phenomena of wounding and healing by Callus, &c., is understood. Also the small sections on the throwing off and fall of the Leaf, &c, were omitted as of minor importance. Further, it was unavoidable that the use of the newest Literature should be limited. Much that has appeared in latter years has, to my sorrow, been consciously and intentionally left unused. On the ground of the above explanations, I particularly beg to be excused for this. From the older Literature I have perhaps cited too much for many, and for others too little. But here also arose the necessity of keeping a definite limit, in order to bring something to completion. On the anatomy of plants such an indescribable amount has been written, that, in a. comprehensive treatise, one or many authors might be cited in reference to every word. To carry this through, even to the extent to which it is done in the section on Epidermis, makes the vill ; PREFACE, description exceed the bounds of convenience and overstep the limit of human power and endurance. I have therefore placed a check upon this also, and, once for all, I will say in answer to possible claims, that every word in this book has had a previous author, printer, and publisher. I hope that I have as a rule referred sufficiently to fundamental works: still it may be emphatically stated that the chief sources and foundations of my work were the writings of Mohl, Nageli, Sanio, Th, Hartig, and in the latest times of van Tieghem, although in some cases I may have omitted to cite them expressly. I presuppose a knowledge of the Text-book of Sachs. When this is quoted, and no further information is given, the fourth edition is always meant. The old Literature is only cited when absolutely necessary, since it lies outside the purpose before us, to write a history of the anatomy of plants. In Sachs’ History of Botany, Zrevcranus’ Physiology, and JZeyen’s Phytotomy and System of Vegetable Physiology, the reader will find what is wanting here. The plan and course of description are more exactly indicated in the Intro- duction. The book deals in the first place with the actual mature structure of the higher plants, and touches upon the history of development only by way of assistance. We do not thereby ignore the fact that the description of the mature condition must necessarily be based upon the history of development, since - that which is termed mature is nothing more than a further advanced part of the whole course of development of the individual. It must therefore always be referred as a matter of course to earlier stages of development, and be coupled with them. But it was the more the object of this work to put that stage of development which is called mature to the fore, since the present overruling preference foi the earlier stages has often brought it about that in the ‘voir venir,’ the things them- selves, which are to be produced, are neglected. I know only too well how far the book falls short of the object indicated in the title. The name ‘Vorarbeiten,’ or ‘Prodromus, of a comparative anatomy would better correspond to the result. That title was only rejected for shortness, sake, and on the consideration that every work should be the predecessor of a better. Most of the figures were drawn by the author on wood, from nature. In case of those copied and borrowed from other books the source is given in each case, I am specially thankful to my respected colleague Sachs for the permission to use the woodcuts of his Text-book, and I should have made still further use of them had not a number of the figures here given been already cut before the earlier editions of the Text-book appeared. I may offer this expression of thanks without presumption not only in the name of the author, but also in that of the reader. Also in the name of both I may add thanks to Dr. von Rostafinsky of Krakau, who has constructed the index of names. A. DE BARY. STRASSBURG, June 15, 1877. TABLE OF CONTENTS. INTRODUCTION Sect. I. 00 ON Au BW rar Il. 12. 13. 14. 15. 16, 17. PART I. THE FORMS OF TISSUE, CHAPTER I. Cellular Tissue. General Introductory Remarks Division I. The Epidermis. . General Definitions 1, COMPOSITION OF THE EPIDERMIS. . Enumeration of the Component Parts . Epidermal Cells . Stomata : . Air- and Water-stomata . Air-stomata . . Water-stomata A . Gaps in the Epidermis . . Hair-structures , ‘ 2. STRUCTURE OF THE EPIDERMAL ELEMENTS, a. Protoplasm and Cell-contents. Epidermal Cells Stomata . Hair-structures &. Structure oe the Walls. Cellulose-Membranes ‘ Intramural and Superficial Deposits Mucilage, Cuticle, and Cuticular Layers Wax PAGE 27 29 30 30 34 45 45 50 54 54 66 67 68 70 73 73 82 x Sect. Sect. 18. 19. 20, ar. 22. 23. 24. 25. 26, 27. . 28. 29. 30. . 31. 32. 33- 34. 35. . 36. 37. 38. 39. 40. 4l. 42. 43. 44. TABLE OF CONTENTS. Restio diffusus Dermal Glands Pulverulent Hairs Digestive Glands. Silicification, Calcification, Couslites: Deposits of Lime Division II. Cork. Origin and Structure of Cork . Division III. Parenchyma. Forms with Thin Walls . Collenchyma, Sclerotic Cells : ‘ ‘ , Endodermis (protective sheath) ’. CHAPTER IIL. Sclerenchyma. General Observations . Short Sclerenchyma, Stone Sdiecnchyas Sclerenchymatous Fibres CHAPTER III. Secretory Reservoirs. Synopsis Sacs containing dort Sacs containing Mucilage . Sacs containing Resin and Gum-resin Sacs containing Tannin CHAPTER IV. Trachee. Synopsis Fibrous ‘iickeniag of tke Walls Bordered Pits ; Transverse Bars . Tracheides . Vessels . Contents of the Traches Thyloses CHAPTER V. Sieve-tubes. Angiosperms ‘ é ‘ ; ; Gymnosperms and Ferns , F é . PAGE 88 88 99 100 102 106 108 115 119 121 126 127 128 135 137 143 145 153 155 156 158 163 164 165 169 172 179 Sect. 49. »» 50. » «BT. » 52. . Internal Hairs TABLE OF CONTENTS. CHAPTER VI. Laticiferous Tubes. . The Latex . The Tubes . . Articulated Tubes . Non-articulated Tubes History. General Observations . CHAPTER VII. Appendix. Intercellular Spaces. General Observations. Development Intercellular Secretory Reservoirs Intercellular Reservoirs containing Air or Water Diaphragms PART Lf. ARRANGEMENT OF THE FORMS Sect. 54. Sect. 55. ” 56. » 57+ ” 58. » 59 FIRST SECTION. ‘PRIMARY ARRANGEMENT. General Observations. Epidermis, Hypoderma CHAPTER VIIL Arrangement of the Traches and Sieve-tubes. I. OUTSIDE THE VASCULAR BUNDLES. Scattered Tracheides . Sheath of Tracheides in Aerial Rosie: Scattered Sieve-tubes . 2. VASCULAR BUNDLES, General Observations . xi PAGE 183 186 189 190 192 200 201 210 217 220 OF TISSUE. A, Arrangement of the Vascular Bundles. a. In the Root. 224 226 227 231 232 xii I. II, Ill, Iv. Vv. VI. VI. TABLE OF CONTENTS, 6. In the Individual Leafy Siem. Sect. 60. General Rules ” ” ” ” ” 61. Dicotyledonous Type. Dicotyledons Gymnosperms ; : : 62. Anomalous Dicotyledons Medullary Bundles 63. Cortical Bundles j 64. Palm-type. Simple Form . 65-67. Modification of the Palm-type 68. Monocotyledonous Seedlings. Aroidez 69. Type of the Commelineze 70-71. Anomalous Monocotyledons 72. Phanerogams with an Axile Bundle 73. Fern-like Plants. General Observations . 74. Equisetum . 75. Osmundacez 76. Isoétes : 77. Psilotum and iiycopouilien : 78. Selaginella . 79. Filices and Hydioplerics: 80-83. Axile Bundle. Tube of Bundles 84. Concentric Rings of Bundles 85-87. Medullary and Cortical Bundles c. Course of the Bundles in the Leaves and Foliar Expansions, 88. Nodes, Stipules . 89. ‘Petiole . go. Lamina. Nervation . 91. Superficial Divarication of the Hutidles 92. Position as seen in Vertical Section 2 Note. Distribution of the Forms of Nervation . ad. Connection between the Bundles of different Orders of Shoots and Branches. 93. General Considerations I. SIMILAR BRANCHES OF LEAFY STEMS, 1. Normal Branches. 94. Type of the Dicotyledons and Dir 95. Other Phanerogams : 96. Fern-like Plants . i1, Adventitious Shoots. 97. II. Roots, B. Structure of the Vascular Bundles. 99. General Observations : ; : ‘ . ‘ S . PAGE 233 235 245 248 248 256 261 264. 267 269 274 277 278 279 279 280 280 282 283 283 289 291 296 298 298 299 305 306 307 307 31l 312 TABLE OF CONTENTS. xiii PAGE 1. BUNDLE-TRUNKS. Sect. 100. Synopsis . . : . : ‘ : ; ‘ . : . 317 » lor. Collateral Bundles. : ‘ : 4 2 . 319 » 102. Bundles of the Leaves of Cycadeze aiid isecies ‘ é g : - 335 » 103. Bicollateral Bundles . ‘ ; ‘ ‘ ‘ ‘ : 3 ‘ . 338 yy) 104-105. Concentric Bundles i 5 : ; : ; ‘ ‘ . 339 », 106, Concentric Bundles of the Ferns ‘ : é ‘ : : + 342 » 107. Radial Bundles . é : ‘ ‘ 5 . : : » 348 » 108. Radial Bundles of Typical Routs j F : : ; ; » 355 » 109. Root-bundles of Abnormal Structure ‘ : ‘ ‘ - ; . 364 »» 110. Imperfect and Rudimentary Bundle-trunks . ‘ ’ ‘ ‘ . 366 2. ENDS AND CONNECTIONS OF VASCULAR BUNDLES. » 111. Endings in the Cortex, and in the Foliar Expansions. 3 . a 4394 » 112. Peculiarities of Leaves of the Conifere . , ‘ 2 é ‘ » 378. » 113. Terminations in Roots and Haustoria . ; é : ‘ ‘ ; (383) » 114. Connections of Bundles. 5 : ‘ ‘ ; a) ae A . 385 C. Development of the Vascular Bundle. » 115. Development of the Individual Bundle. ‘ 388 » 116. Development of the Bundle-system in the Stati, Cecesasinn of ilk Bundles. Their Morphological Position . : , é : : 2 » 117. Development of the Lateral Roots : : ‘ ; on Concluding Remarks é : : . , : ; ; ; » 399 CHAPTER IX. Arrangement of the Primary Parenchyma. Sect. 118. General Considerations . ‘ : z : : : : . . 402 9 I1g9. Pith, Medullary Rays. External Cortex . 2 : ‘ : . . 402 » 120. Petioles, Ribs of the Leaf . a ‘ 3 : : ‘ . 405 », 121. Lamina of the Leaf (Mesophyll, Diachiyaie , : ‘ é : » 406 », 122, Cortex of the Root, Root-cap : » 123. Parenchymatous Sheaths, Endodermis, Starch: layer, —Plerome- death » 414 7 CHAPTER X. Sclerenchyma and Sclerotic Cells. Sect. 124. General Considerations . : : : : ‘ 3 - + 6 4l7 y»» 125. Fibrous Layers and Strands. : : : , : ‘ ; - 417 » 126. Isolated Fibres. bts ck. ake ok GBR gag » 127. Short Sclerenchymatous Elements : : : : 6s . - 425 », 128. Thorns, Prickles, Warts . ; ‘ : : : : é : » 425 » 129. Sclerotic Elements of the Ferns ; : : ‘ ‘ ; : . 426 CHAPTER XI. Secretory Reservoirs. 130. 5 . . . . . . . . . . . . 5 + 431 xiv TABLE OF CONTENTS, CHAPTER XII. Laticiferous Tubes. PAGE Sect. 131. . : : 3 : : " : : : ; : : : - 432 CHAPTER XIIL Arrangement of the Intercellular Spaces. Sect. 132. Spaces containing Air : ‘ F ; ; ‘ - : é . 440 » 133. Intercellular Secretory Reservoirs. ‘ " ; ‘ ; ‘ . 440 SECOND SECTION. SECONDARY CHANGES. CHAPTER XIV. Secondary Growth in Thickness of Normal Dicotyledonous Stems and Roots. I. Cambium. Secondary Thickening. Sect. 134. Origin of the Cambium in the Stem. Intermediate Bundles . ‘ - 454 » 135. Subdivisions of the Secondary Thickening in the Stem . ‘ i » 458 »» 136. Cambium and young Secondary Growth . ‘ ‘ 461 »» 137 General Arrangement of the Secondary Elements as seen in *Reuneveise Section . j A ‘ : “ ‘ ‘ ‘ i . 469 » 138. Their Longitudinal Gouge ‘ , ; , : . 470 », 139. Cambium and Secondary Thickeniag of Raors: : ‘ z : . 473 II. The Wood. 1. DISTRIBUTION AND FORM OF THE ZONES OF SECONDARY GROWTH. * jy AOR es Se, 2. THE TISSUES OF THE SECONDARY Woop. sy EAD. SYNOPSIS: oo awa eG we » 142. Trachee . . : : ; : E : : ‘ 5 ‘ . 478 » 143. Woody Fibres . ‘ s 5 i : i ‘ F Z : . 481 » 144. Cells . et ; : ze - 8 : ‘ ; : : . 483 »» 145. Crystal-sacs. Laticiferous Tubes , . i ‘ : ; 2 . 487 Critical Note. ‘ : : ; : ; ‘ : ‘ ‘ . 487 3. DISTRIBUTION OF THE TISSUES IN THE Woop. 146. General Considerations . é : : 3 ‘ ‘ ‘ ‘ . 488 a. Medullary Rays and Medullary Spots. » 147. Medullary Rays : : : ; ; : : s ; : . 489 » 148. Medullary Spots ‘ : 5 “i : ; : ‘ j » 492 Sect, ” 149. 150. 151. 152, 153. 154. 155. 156, 187. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. 169. 170. 171. 172. 173. 174, 175. 176, 177. 178. 179. TABLE OF CONTENTS, 6 6. Ligneous Bundles. Normal Ligneous Bundles Peculiar Forms . c. Changes of the Tisssues in the Annual Ring. a. Normal Differences of Successive Zones of Thickening. Innermost Ring. Medullary Sheath ‘ Successive Increase in the Size of the Elements Alburnum and Duramen e. Individual and Local Deviations. Differences according to the Thickness of the Annual Ring Rings with Indistinct Limits Individual Variations J. Differences between Non-equivalent Members. Stem, Branches, Roots Fleshy Roots III. The Bast. General Considerations Forms of Tissue of the Bast Parenchyma. Sieve-tubes Laticiferous Tubes Secretory Passages and Sacs Sclerenchyma. Bast-fibres Crystal-sacs Variations according te successive Deiiees Tadividuals: ee CHAPTER XV. Secondary Changes outside the Zone of Thickening. Pith Parts lying autside dia Gambit ZoKe Epidermis . z . . . Cortical Pavelcia iia: Dilatation. Scenndaey Sclerosis Sieve-tubes, Sclerenchyma, Secretory Sacs Phenomena of Disorganisation Periderm. Development and Differentiation of Periderms Superficial Periderm . First Internal Formation of Peritlena Repeated Internal Periderms. Bark , Combination of the different Peridermal onwacons in Weoedy Plants Lenticels XV PAGE 493 498 500 504 505 507 SII 513 514 515 516 519 520 521 525 525 526 529 530 533 535 535 536 541 543 544 547 551 554 558 560 xvi TABLE OF CONTENTS. CHAPTER XVI. Anomalous Thickening in Dicotyledons and Gymnosperms. PAGE Sect. 180. General Observations. Eccentric Stems and Roots : 4 ‘ . 567 »» 181. Synopsis of the Anomalies of Thickening ; - + 367 » 182. Anomalous Distribution of Tissues with Normal ‘Cexihiom, ' Senega- root * 569 » 183. Climbing Bignoniacez 570 » 184. Phytocrene 575 » 185. Malpighiacez, Apocynen, Melepiadler Cdldsirun Tounmeteie 577 » 186. Sieve-tubes in the Wood ; Strychnos, Dicella . 3 ‘ ‘ 5 . 577 » 187. Cambium inside the Ligneous Body: Tecoma radicans . ; 580 » 188, Partial Cambiums and Woody Rings of the Sapindacee . : . 581 » 189. - i in the Calycanthez and Melastoniea: 584 » 190. 53 in the Rhizome of Rheum ; » 585 » 191. Successive renewed Tonics of Thickening . 586 5 192. Chenopodiaceze, Amarantacee, Nyctaginez, Mesembayantbeana, Tetra- gonieze ‘ 5 ‘ ‘ ‘1 : i ‘ ‘ ‘ 590 » 193. Anomalous Dilatation of the old Parenchyma. Intercalary Zones of Thickening in Stems of Lianes . 601 » 194. 5 in Roots, Gunvelvufacen Bis 606 » 195. Cycadeze 608 » 196. Welwitschia 614 CHAPTER XVII. Secondary Thickening of the Stem and Roots of Monocotyledons and Cryptogams. Sect. 197. Stem of Dracaenez, Aloineze, and Beaucarnea . 618 4, 198. Tubers of Dioscoreacez 622 » 199. Roots of Draczenez . 622 » 200. Stem of Isoetes . 623 INDEX 625 ERRATA; P. 41, 1. 7, for Scheidei read Schiedei. P, 41, 1. 8, for aculenta vead aculeata. In description of Fig. 15, for purpurasceus read purpurascens, . a i : \ Jor Sanseviera read Sansevieria. 2 r ie Jor Patschoulivead P. 95, description of Fig. 38 Berea P. 144, 1. 31, for Conocephalous vead Conoce- phalus. P. 185, 1. 7, for Haucornea vead Hancornea. P, 228, 1. 5, for Anselia read Ansellia. P, 230, 1. 32, for Homalonema read Homalo- mena, P. 248, Il. 38 and 41, for Tladiantha read Thla- diantha. P. 253, Il. 32 and 34, for of read from the. P. 477, 1. 9, for Red Fir read Spruce. P. 510, 1. 16, for Drybalanops read Dryoba- lanops. INTRODUCTIOR, Tue body of the plant is composed of parts with definite position, succession, structure, and direction of growth. These we call its members, when we refer only to their share in the structure. Investigation teaches us to recognise members of different, and, in plants of complicated structure, numerous ranks, Roots and leafy shoots; internodes, leaves, segments, and ‘layers of meristem, masses of cells ; finally, the single cell, which may again be separated into members. Each member, of whatever rank, is adapted according to its development to definite physiological work. It becomes the instrument, or organ of this work. As was the case with members, there may also be distinguished organs of different rank— simpler, and gradually more complicated. According as an organ adapts itself to a definite function, it attains properties of form and structure, which are definite, and differ from those of other organs. The description and explanation of the collective phenomena of form and struc- ture constitute the task of morphology. According to the two points of view now put forward, we must distinguish between the morphology of members, and the morpho- logy of organs. The former deals exclusively with the phenomena and laws, according to which the organism is compounded of the members of different rank; the morpho- logy of organs with the properties of structure and form, by which the members become organs, and with the distinction of organs of different rank, according to those properties. Strictly speaking, the morphology of the organ presupposes a knowledge of that of the member, since the origination of a member must precede its evolution into an organ. As a matter of fact, a sharp separation of the two disciplines can hardly be carried out, since both work with the same material, which extends from the realm of the one to that of the other without any definite break. The subject of this book is a part of the morphology of the organs of Plants, which is limited for convenience sake. According to the programme of the Hand- book, of which it is a part, it should treat of ‘ she anatomy of the Vegetative Organs of Vascular plants ;’ it is occupied therefore only with the Phanerogams, and Pterido- phyta, i.e. the Fern-like plants in the widest sense of the word. Further, it pre- supposes such knowledge, gained from other sources, of the outward form of organs of higher rank (e. g. foliage-shoots, roots, &c.) as can be acquired without anatomical investigation, and treats only of their internal structure. Lastly, it is limited to the 1 Sachs, Textbook, second English edition, p. 149. - i 2, INTRODUCTION. vegetative organs. Having regard to the points of view established above, as well as to considerations of the available space, the work has further to presuppose a know- ledge of the morphology of members—otherwise called the general morphology of plants, and general doctrine of the cell—and only to touch on these subjects as far as may be necessary. Since the investigation extends over three great sections of the vegetable king- dom, it will be our task to describe comparatively those phenomena, in which the representatives of these sections correspond, or differ: that is, to offer a comparative anatomy of the vegetative organs. Under the term vegetative organs we include all those organs of the plant which are not organs of reproduction, i.e. which do not, sexually or asexually, serve in the production or direct preparation of germs: the term includes therefore those organs which undertake the entire work of preservation of the physiological individual, and which may take different parts in this work. In the plants in question, which always have members of many grades, those of every sort and rank are to be found developed into vegetative organs: both those belonging to the highest ranks, which are externally apparent, such as roots, leafy shoots, with their internodes and leaves; and of successively lower ranks, such as definite groups of cells, and lastly single cells, or the products of their metamorphosis. ‘But investigation shows that the adaptation to, and participation in vegetative duties, that is, the development into organs of definite function, and corrésponding structure, is far the most commonly and definitely carried out for members of lower ranks, i.e. for - cells, and groups of cells, or the products of their metamorphosis. It is these which in the first instance share among them the vegetative work, and assume a correspondingly characteristic form and special structure. A member of higher rank composed of them is only a vegetative organ of higher rank, inasmuch as it consists of them. The structure characteristic of such an organ is determined by the structure and distribu- | tion of the organs of lower rank, which compose it. The vegetative structure in ques- tion is not universally connected with definite members of higher rank. Equivalent members, it is true, very often develope into equivalent organs: the functions of the leaf, assimilation of carbon, transpiration, &c. are, for instance, usually deputed to leaves: most roots are equivalent to one another in both relations. But, on the other hand, the converse is not uncommon, that non-equivalent members are equivalent organs. In many plants besides the leaves the internodes also, which are connected ‘with them, take part in the functions of the leaf: in others, with ‘leaf-like’ stems, the function and corresponding structure, which the leaves have lost, are transferred ‘to the stems. Trapa natans has part of the petiole developed into a swimming organ: in the floating species of Desmanthus internodes of the stem, and in species of Jussizea certain roots assume this function, and corresponding structure. This being the case, the description of the structure of the vegetative organs must start from the consideration of simpler forms, i.e. the cells. Since the investi- gation of these separately is necessarily followed by that of their connection with others, and of their arrangement into tissues of various grades, the structure of the organs composed of these gradually becomes apparent. Those cells or their derivatives, which have the character of definite vegetative organs, seldom occur, in the plants in question, singly between dissimilar ones; INTRODUCTION, 3 usually similar cells are connected so as to form large groups or masses. An aggre- gation of cells growing in common is termed generally a Zisswe (Tela, contextus; in compound words ieriov’). Each tissue, which is characterised by definite properties, and distinguished from others, is called a “ssue-form or, better, a sort of tissue. For the single cells which belong to a tissue, or for each elemental form derived from such a cell, the term #’ssue-element may be reserved. Tissue-elements which occur singly between dissimilar ones (Idioblasts, after Sachs’ terminology) usually cor- respond in their properties with others which occur in connection with similar elements. They are then to be reckoned as of the same sort of tissue as the latter. In like manner, finally, such tissue-elements as occur only as Idioblasts (for instance many laticiferous tubes) will form together with one another a special sort of tissue. All tissue-elements, which correspond in definite similar properties, are therefore termed collectively a sort of tissue, whether they be Idioblasts, or are connected with like elements. The course of description which is followed in this book may be gathered from what has been said. The first subject is the characterising and distinguishing of the sorts of tissue, which serve as vegetative organs; then the grouping and arrange- ment of these in building up the members or organs of higher rank. In this course of description a difficulty certainly arises: this can only be overcome by the establish- ment of a limit, which is to a certain extent artificial. Those tissues, which act functionally as vegetative organs, are often continuous in the plants in question with those higher members, which are according to their most important adaptation reproductive organs. The member of many Ferns which acts functionally as a Prothallium is principally composed of chlorophyll-containing Parenchyma, similar to that of foliage leaves. This sort of tissue, together with vessels, vascular bundles, &c., takes part in the construction of the parts of the flower of many Phanero- gams, &c. Many peculiarities of vegetative tissue, which appear in these parts, depend much less, it is true, on the properties of the single tissue-elements, than on their arrangement. Since these peculiarities are directly connected with adaptation to the generative process, the study of them should also be in connection with it, and must be excluded from this treatise. In cases of sharply-defined phases of development, the boundary, which must needs be drawn, is evident at first sight. For instance, no one will expect the Fern-Prothallium to be treated of here. But among the Pha- nerogams there often occurs a gradual transition between purely vegetative and reproductive organs. ‘To satisfy in the present case the necessity of a definite limit to the subject in question, all that falls under the definition of flowers, specialised inflorescences, or parts of inflorescences, will be excluded from consideration. * As has been already intimated, the differentiation of the sorts of tissue is a phenomenon which accompanies the development of a part to maturity. Originally the cells of a part differ, it is true, in certain relations, both in form and direction of division; but they correspond in structure, and in the fact that, while they increase slowly in size, they divide repeatedly ; and thus finally produce cells, which develope into tissue-elements. From those phenomena of division, such masses of cells are termed Jerzs/em; and when they form the first foundation of a member frimary 1 Compare Unger, Anatomie, p. 138; Sachs, Textbook, English edition, p. 70, B 2 4 meristem». Th structure the cells of meristem are characterised by having a delicate homogeneous membrane (only in certain exceptional cases thickened and. with flattened pits), and homogeneous, finely granular protoplasm with a nucleus, but with no further recognisable structural elements. By reason of their constant division, they are throughout in uninterrupted connection with one another. In each system of the meristem the divisions pass through a definite number of stages, till they gradually cease. According as this happens, the cells, first formed as members of the meristem, assume those properties, by which the further sorts of tissue are distinguished: great increase in volume takes place, and changes of structure and form; while the latter may result in a partial loosening of the original uninterrupted connection, i.e.-in the appearance of zwéercellular spaces. As compared with the meristem, the tissue elements derived from it attain a great- constancy both of form and structure. They have accordingly been termed permanent tissue, fixed tissue, or mature tissue. If the idea of tissue be understood in the genéral sense stated above, the meristem also is naturally included: we distinguish 5 then on the one hand meristem, or formative tissue, on the other permanent tissue, as the two main categories of tissue. Merely for shortness of expression, however, the term tissue will be used also for permanent tissue in opposition to meristem. In this sense, and according to the previous explanation, the following pages will deal “ the vegetative tissues, which serve as vegetative organs. : + Comparative investigation shows that the cells of the meristem universally cor- respond very closely in the general character of their structure, and the same may be said-of the main phenomena of the vegetative process in the plants we are engaged with. -Answering to this correspondence of origin and functional adaptation, the - forms -of tissue in the whole group on which we are engaged correspond in their main properties, notwithstanding frequent modifications to suit special cases, and the same few tissue-forms everywhere occur. - Tissue-elements of every sort are derived from the cells of the meristem, each has originally the properties of a cell. Accompanying definite development there first appears the fundamental difference, that certain elements refain during their life the: structure and all the characteristic properties of typical cells, others ose the cell-nature. The former are composed of a completely closed cell-membrane and active- proto: plasm, with a nucleus and cell-contents ; they retain the power of independent growth, and remain capable of division; in consequence of this property a meristem may again’ arise from them, and this, in antithesis to the original, is distinguished as secondary meristem (Nageli,1.c.). The latter lose with their development the power of division, and of independent growth; usually they cease entirely to grow, but in many cases a real lasting growth of such elements occurs, resulting from their nourishment by adjoining cells. In their structure, the loss of the cell-nature is indicated either ‘by the complete disappearance of the protoplasmic body, its place being filled by other bodies, usually air or fluids; or by its suffering characteristic changes, which ‘vary according to the individual cases. The latter observation is made with special / INTRODUCTION. 1 Nageli, Beitrige, I. p. 2—Schleiden (Grundz, 3 Aufl. I. p. 283) and Karsten (Veg. Org. d. Palmen) include these masses under the wide term Cambium; Unger (Anat. und Physiol. p. 180), calls them formative cells: Schacht (Pflanzenzelle, p, 165) primary parenchyma. INTRODUCTION. . 5 reference to the sieve-tubes; it remains doubtful whether the contents of these. are protoplasm or not. The cell-walls of the elements in question are wholly or mostly retained. i According to the differences already mentioned, which will be further followed: out in the subsequent special observations, the tissues divide themselves into those which consist permanently of cells (cellulze), and those whose elements are descend- ants, derivatives, or products of change or mefamorphosis of cells. According to their form and other properties these are termed /udes (Tubi, Tubuli), sacs (Utriculi), fibres (Fibre), and are distinguished from cells. Most tissue-elements, of whatever sort, are formed directly and quickly by the metamorphosis of meristematic cells. Exceptions from this rule only occur in certain: cases when cells, after they have acted as such for a long time—even for years—may secondarily pass over to another tissue-form. This takes place in the secondary: development of Sclerenchyma, which will be described in chapters II. and XV. « From this secondary metamorphosis of tissue we must distinguish death, and the: changes eventually connected with it, which appear in certain other cases in thei tissues, such as the dying off of old hairs, cork-cells, cells of the pith of many plants, of the elements of bark, and of the old wood of Dicotyledons, &c. ; appearances which can usually be distinguished with ease from metamorphosis of tissue by the commencement of rotting, weathering, &c. In accordance with the preceding considerations, the distinction of the forms of tissue which act as vegetative organs, and the classification of the study of them, must in the first place be founded upon their structure, that is, on the structure of the single tissue-elements, and the connection of these with one another—whether they be connected with like or with unlike elements. It is obvious that, in organized bodies, certain peculiarities and varieties of structure are connected with certain phenomena of development. And it is not less obvious that varieties of structure are also as a rule correlated with certain varieties of form of single tissue-elements. But experience teaches that between form and structure a cons/ant relation does not exist, or at least not universally ; and that, contrary to the older classifications, which regarded in the first place the form of the elements, this is of only secondary im- portance in the distinction of the tissues. According to the principles now laid down, the following main forms of vegetative tissue may first be distinguished :— I. Cellular trssue, i.e. that which consists of permanent typical cells; with the main subdivisions, epzdermis, cork, parenchyma. Il. Sclerenchyma. Ill. Secretory structures. IV. Vessels. V. Steve-tubes. V1. Milk-tubes. Separate notice of the Intercellular spaces may, with advantage, though perhaps not necessarily, be appended to the study of the tissues. The study merely of the divisions and subdivisions of tissue-form is at all points closely connected with the arrangement of these into combinations of different rank—so as to form vegetative organs of successively higher rank—so far so, that the two modes of study can never be completely separated the one from the. other. ' According to their form the combinations of any rank may be distinguished as Layers, Bundles, Masses (Groups, Nests)—terms, the meaning of which is obvious from 6 INTRODUCTION. the ordinary use of the words: a sharper definition of them is here neither necessary nor possible. When a layer (simple or compound) surrounds a tissue, which differs from it, it is termed relatively to the latter a sheath. ‘Tissues of whatever sort or rank may be mutually continuous for long distances, or throughout thé whole plant. When this is the case, they are said collectively to form a Sys/em, A system of any rank may be composed of systems of lower rank, or may join with others to form a system of higher rank. For instance, the Vascular system is composed in most plants of that of sieve-tubes, and that of vessels; each of the latter is a system of itself ; the two combine as a rule to form the above-named system of higher rank ; while often the system of Sclerenchyma-fibres joins them as a third constituent of the | joint system. According to the point of view from which one starts, one may therefore . distinguish systems in the most various sense, as will appear in the later chapters ; for instance, in the Dicotyledonous stem we may with equal right distinguish a Vascular and an Epidermal system, or a Woody and Cortical system, whereas the latter includes, besides the Epidermal tissue, a part of the Vascular system, and other tissues besides, i Sachs (Textbook, Eng. ed. 1882, p. 79, etc.), in the exposition of the anatomy of the higher plants, starts from the definition of three systems of tissue, which he terms Dermal, : Fascicular, and Fundamental tissue. Under the first term he includes those tissue-forms, : which limit externally such plants as have their cells aggregated in three dimensions of spacé, as a matter of fact, Epidermis and Periderm (cf. § 2, 23 and chap. xv). 3 His Fascicular tissue corresponds in the main to the previously mentioned Vascular system (chap. viii). The name Fundamental tissue includes what remains after the‘ separation of the other two. However much this distinction may be fitted to guide beginners, still, in my opinion, it does not answer its purpose, which is to serve as a basis for a uniform exposition of the various differentiation of plant-tissues. For the names Dermal and Fascicular tissue indicate in Vascular plants systems of tissue, which are positively characterised by definite tissue-forms: but the name Fundamental tissue im- plies the remainder, and this may just as much consist of different positively characterised | tissue-forms, and tissue-systems, which are equivalent to the Dermal and Fascicular systems. But if it is necessary, in the description of the Dermal and Fascicular systems, to make use of a short general term for the tissues over and above these, the terms Fundamental tissue, or Fundamental mass, or Intercalary mass, are very suitable; just as in Nageli’s treatise on the vascular bundles, or fibro-vascular masses, was his distinc- ' tion of these from the rest (‘Proten’), or as was Schwendener’s general term for the parts of the vascular bundle, which, in his exposition of the mechanical adaptation, did, not bear upon his point. And indeed in describing a form or system of whatever rank, some such method must always be used. I think, however, that the distinction of the. forms of tissue must first serve as a foundation for the uniform exposition of the subject‘ which now engages us, and for the choice of terms; then only should follow the investi-" gation, how far these forms of tissue take part in the formation of combinations, and systems of higher rank. . Although on the one hand the course of exposition above brought forward has a definite justification in all cases, and, on grounds which need not be repeated, will be in the present case pursued, and although further the distinction of single forms of tissue must be carried out under all circumstances with regard to the structure of the elements only, still the question arises, on the other hand, if the distinction ‘of, definite systems could not be drawn more naturally on other than the purely histo-! logical grounds above stated ; that is, on such as are derived from the morphology , INTRODUCTION. 7 of differentiation (i.e. the history of development).. Investigation has taught that the original meristem, from its first (embryonic) beginning onwards, maintains a well de- fined differentiation into different—but always meristematic—layers, or segments; and that, in the plants on which we are engaged, in many cases definite sharply marked masses of tissue (I may mention provisionally the axile vascular bundles of many roots) are derived from certain of these layers, others from other layers, just as the different tissue-systems of the animal body are derived from the different germinal layers. The question may therefore arise whether, over and above the distinction of forms of tissue according to definite structure, the exposition of the tissue-systems, and of the construction of the members of the highest rank from them, must not refer to that differentiation of meristem, and take it as the foundation. The determination on this point will depend upon the answer of the further question, whether the origin of each, or of single tissues or tissue-systems, from one and the same portion of the primary meristem, can or cannot be generally proved. For if the latter is the case, if parts of the same tissue-system originate from unlike parts of the meristem, the course of exposition which we are considering must be put aside as impracticable. __ In order to gain clearness on this question, a survey of the original modes « of differentiation of meristems is necessary; and though this lies beyond the strict limits of the subject of this book, still it may with the more reason be here inserted, since in the following chapters reference must constantly be made to that differentiation 1. I. As Hanstein ? has shown, the young embryo of the Angzospermous Phanerogams separates, while still consisting of few cells, allaf which are meristematic, into three layers, or groups of cells, which differ in their arrangement and direction of division; these were termed by their discoverer, Dermatogen, Periblem, and Plerome*’, At the end of the root there is besides these a fourth, the origin of which in the embryo remains to be further investigated. This was called by Janczewski * the Calyptrogen layer: we shall return to this later. The dermatogen layer is separated by a single ‘tangential division of the few cells, which form the original rudiment of the embryo, as a simple peripheral layer of cells. It remains a simple layer of cells, since . all the subsequent meristematic divisions which occur in it take place only by walls at right angles to the surface. It is only at the future apex of the root that other pheno- mena appear, which need not be more fully discussed at present. Further divisions of the cells enclosed by the dermatogen separate an axile longitudinal cylinder, the plerome, from the periblem, which is a zone of tissue lying between the plerome and dermatogen. The pletome consists of cells in which longitudinal divisions pre- ponderate, and which have a corresponding arrangement; the periblem of cells distinguished from the former by more frequent and irregular transverse divisions. 1 [See further Sachs, Ueber die Anordnung d. Zellen in d. jiingsten Pflanzentheilen, Arb. d. Bot. Tnst. in Wiirzburg, Bd. IL; also, Ueber Zellanordnung u. Wachsthum, ibid.—Haberlandt, Scheitelzell- wachsthum b, d. Phanerogamen, Botan. Zeitg. 1882, p. 343. —Nageli, Scheitelwachsthum d. Phanero- gamen, Naturforscher Vers. zu Miinchen, 1877.] 2 Hanstein, Botan. Abhandl. I. On the details of the origination of the embryo, and its great differences in those Di- and Monocotyledons which have been investigated, the reader must be referred to this treatise and to Sachs’ Text-book, English edition, 1882, pp. 585-593. Further, Fleischer in Flora, 1874, p. 369, &c.; Hegelmaier, Bot. Zeitg. 1874, p. 631, &c. s 8 Hanstein, Die Scheitelzellgruppe im Vegetationspunkt der Phanerogamen, Bonn, 1868. * Ann, Sci. Nat. 5 série; tom. XX. 8 INTRODUCTION. This differentiation of the meristem is retained at the punctum vegetationis of the first-stem and of the tap root; further it appears at the puncium vegetationis of all lateral stems and secondary roots. The differentiation certainly varies greatly in distinctness in different cases, and is complicated in roots by the presence of the Calyptrogen-layer. It corresponds exactly to the scheme given above in the apex of the stem of many thin-stemmed water plants, as Elodea, Hippuris, &c.; in these Sanio! first discovered the appearances in question, which were afterwards more fully worked out by Hanstein®. The rounded conical apex of the punctum vegetationis of Hippuris (Fig. 1) is covered bya single layer of dermatogen (¢). Then, passing inwards, follow usually five regularly concentric layers of isodiametric cells, these constitute the periblem which surrounds the plerome-cylinder (f-9). The FIG. 3.—{205) Hippuris vulgaris. Median longitudinal latter has a bluntly conical apex, where it is section through the apex of quite a young leafy shoot, : dee, ee ey ey eee. Oe em ineted Ure Single nell wvaly Pele a P g Se Sats ESSE Bah ieee Aer it widens out. In most other cases the num- ber of periblem-layers is smaller, only 2-3. , The three zones of meristem, though they remain distinct, equally take part in the acropetal growth in length of the end of the stem. Each is continually renewed by divisions in the group of cells (or single cell) which forms their apical portion,. | while as the tissues are removed from the apex the transition from meristem to the definitive tissue takes place. Each is continuous downwards into definite tissues or tissue-systems, which will be mentioned hereafter. That apical group of . cells, or (as in the plerome of Hippuris) single cell, which renews the layer and which hence always introduces the further cell-divisions in it, is called the Zndé#al cell or Lnitial group of the layer. In all Angiospermous plants the dermatogen layer is marked off with similar sharpness from the tissues below it, and is distinguished by its division walls being arranged ‘only at right angles to the surface; with this restriction they run in all directions. But the separation of plerome and periblem does not appear in all cases so sharply marked as in the foregoing instance. Especially where the apex of the stem is broad and flat, it must often be left undecided whether both do not originate from a single common initial group, and are first clearly separated at some distance from the apex, on their gradual transition to definite systems of tissue. The origination -of the normal lateral branches of the end of the stem (i.e. of the leaves and lateral shoots) as emergences of the surface begins beneath the apex by the outgrowth of definite groups of meristem, which were not previously to be distinguished by any special character. These are the initial groups of the emergence. As a matter of fact, both elements of the dermatogen-layer and of the - periblem lying under it take part in the origination, the growth and divisions of both proceeding simultaneously (comp. Fig. 1); but the cells belonging to the dermatogen 1 Bot. Zeitg. 1864, p. 223. * [Kny, Botan. Zeitg. 1878, p. 760.] INTRODUCTION. 9 during all their further growth are divided only perpendicularly to the’ surface, so that the dermatogen-layer is continuous as a single layer also over the branches. The plerome-cylinder of the mother shoot, as far as investigation extends, takes no part in the origination of a branch. In the lateral shoots thus founded the separa- tion of the meristematic mass, covered by dermatogen, into periblem and plerome first appears after some time. Both in this case originate from a common initial group, which is derived from the periblem of the mother shoot. i In the punctum vegetations of the root of the Angiosperms?* the same differen- tiation of the meristem often appears, as in the stem, but sometimes much more definitely. It should be described in the same terms as the latter, so far as the correspondence is exact. To the meristem, from which springs the body of the root, is added in all roots the conical cap, made up of layers of cells, which is known as the Root-cap (calyptra). This covers the meristematic punctum vegelationis, and is increased by it, according as the cells on its outer surface die off. Since this accession originates in certain cases from a special layer of meristem, the latter is, according to Janczewski, to be distinguished as the calypirogen. As has already been said above, we cannot here discuss what genetic relation this bears in the main root of the embryo to the first meristem-cells of the hypocotyledonary stem. We must also take no notice here of the origination of the layers of meristem of lateral roots, our knowledge of which is for the most part due to Janczewski. “” At the active punctum vegetationis of the roots of the Angiosperms, which has already begun to grow in length, four dif- ferent cases of differentiation, which vary ‘according to the species or groups, have been made known to us by Janczewski: ; 4. The meristem at the apex is diffe- rentiated into four sharply marked off layers: plerome-cylinder, pertblem, dermatogen, and ‘pro, 2—4go) Pistia stratiotes. Median longitudinal sec- lastly the calypiragen-layer, which covers the at trou a yomg lateral rot. Rootcap, including latter, and which soon disappears owing to [scrinenis, Between ¢ and pthe periblem, which consist the short duration of its activity. This diffe- the 2Pe* of only single layer. rentiation is found only in two Monocotyledonous water-plants, viz. Hydrocharis and Pistia stratiotes ? (Fig. 2). . ‘ 2, Sharply defined plerome-cylinder, and calyptrogen-layer. Between the two, at the apex of the punclum vegelationts, is an initial group only one layer of cells thick, which splits immediately behind the apex into periblem and dermatogen (i.e. cortex and epidermis). This is the case in most of the Monocotyledons which have been 1 [See further Holle, Botan. Zeitg. 1876, p. 241; 1877, p. 537.-—Eriksson, Botan. Zeitg. 1876, p- 641.—Schwéendener, Scheitelwachsthum d. Phanerogamen-wurzeln, K. Acad. Wiss. Berlin. 1882 ; Botan. Zeitg. 1882, p. 687.—Flahault, Ann. Sci. Nat. sér. 6, Bot. tom. VI. pp. 1-229.} -.* [Kubin, Hanstein’s Abhandl, Bd. III, Heft 4, 1878.] mag 10 INTRODUCTION. investigated ; e.g. species of Allium, and Canna, Hordeum vulgare, Triticum vulgare, Zea Mais (Fig. 3), Stratiotes aloides, Alisma Plantago, Acorus Calamus, (Janczewski), Treub!, as the result of his extended researches, ascribes this differentiation to Jun- cacez, Hemodoracex, Cannacex, Zingiberacee, Typha, Cyperacez, Graminez, Com- melynex, Potamex, Juncaginex, Sagittaria, Limnocharis, Stratiotes. But he differs from Janczewski with regard to Allium, Acorus, and Alisma, since he does not allow the presence of a calyptrogen-layer in the families Liliacez, Asteliez, Xerotidezx, Aspidis- trex, Ophiopogonex, Amaryllidacer, Hypoxidex, Dioscorez, Taccacez, Bromeliacez, ‘Musacee, Orchidex, Palmz, Pandanex, Cyclanthez, Aroidez except Pistia, further in \ é FIG. 3.—Median longitudinal section through the apex of the root of Zea Mazs, from Sachs’ Textbook. a@—qa outer, # inner layers of the root-cap;, s calyptrogen layer; 72 ¢//plerome; g rudiment of a vessel; x, y—r periblem, or the cortex which has developed -from it ; e Epidermis, or dermatogen layer (v=the thickened outer wall of its cells). Above the apex of the plerome cylinder, easily seen between 7 and s, the dermatogen and periblem layers are reduced to two initial layers, which occupy the depressedcentre. According to Janczewskithe initial group should consist ofa single layer. the Iridacez, Pontederiacex, Sparganium, Butomus, and doubtfully in Alisma. He finds rather, covering the sharply defined apex of the plerome, a group of common initial cells two layers thick, from which originate root-cap, dermatogen, and periblem. Hence, the last named families should represent a special type, differing to a certain extent from those first named. : a si 1M. Treub, Le méristéme primitif de la racine dans les monocotyfédones, Leiden, 1876. INTRODUCTION. II 3. (Fig. 4), Plerome-cylinder and periblem sharply defined, the latter overlying the apex of the plerome, and covered by a common initial layer for dermatogen and root-cap. The divisions of the initial layer, which are parallel to the surface of the bluntly conical apex, add on the one hand new cells to the root-cap above the apex, and on the other hand renew the initial layer itself. As the distance increases from the apex of the periblem, which by its growth in length is constantly advancing, the divisions become rarer, and at last cease. The last of: these separates the initial cell into two, one of which is added to the root-cap, the other to the dermatogen as a permanent member of it. We may therefore say with Janczewski that root-cap and dermatogen arise in this case from the calyptrogen-layer. The cells of the dermatogen and root-cap, which owe their origin to the division just described, divide further by walls perpendicular to the surface; from each therefore is produced a section of a layer consisting of several or many cells. In the root-cap each of these sections is so arranged relatively to the similar ones la- terally next it, and to others which have arisen above the apex, as to form a conical hood one layer of cells thick; and the whole root-cap is built up of such hoods fitting one into another. The cells of the sections of the dermatogen undergo extension perpendicularly to the surface, in such a measure that each sec- tion remains for a time consider- ably less extended in that direc- tion than its predecessor, which is farther from the apex. The surface of the dermatogen layer FIG. 4.—(210) Polygonum Fagopyrum. Apex of root in median longitudinal therefore becomes narrower, as tam, fe peleenttum the outer boda of the plremecyiaders «der. the apex is approached, by steps which occur at definite distances from it. Each step is covered by that section of a layer of root-cap which originally corresponded to it, while the edge of the latter abuts on the next lower, that is, the next older step. A not unimportant variety occurs, according to Janczewski, within the type of differentiation in question. In the majority of plants which have been investigated, the periblem consists at its apex of a single initial cell (Fig. 4) or of two such, which lie side by side in one layer; it is below the apex that it first increases to several layers. But in one case, namely, Linum usitatissimum, the apex of the periblem consists of two initial layers. One of these, the inner or lower of the two, behaves as in the first case just described. The other, the outer one, belongs to a layer of cells, which clothes the whole periblem. This, like the dermatogen, divides only perpendicular to the surface, and therefore always remains a single layer. 12 INTRODUCTION. To this type belong the majority of the Dicotyledons. Helianthus annuus, Fago- . pyrum, Raphanus sativus, Myriophyllum, species of Salix, Casuarina stricta, Lmum usitatissimum, and Primulacez? have been carefully investigated. 7 ‘4. The fourth Angiospermous type (Fig. 5) is observed in those of the Cucurbitaceee (Cucurbita) and Papilionaceze (Pisum, Phaseolus, Cicer) which have been investigated. Here a common initial zone extends transversely over the punclum vegetationts " by the divisions in this arise, on the side towards the FIG. 5.—(210) Pisum sativum. Median longitudinal section through the apex of the root, after Janézews! ki, Pp plerome, p—r periblem, ¥ common transverse initial zone, c its lateral continuation, ae root-cap, successive layers, which are added to the conical middle portion of the latter. -On the side of it facing the body of the root arise a massive plerome-cylinder, ; and a periblem many layers of cells thick, having approximately the form of a hollow cylinder open towards the initial layer. From its- margin the transverse initial zoné of meristem curves itself round, so to speak, over the adjoining outer surface of the periblem, and acts here for a further space as initial layer, on the one hand ? Kamienski, Zur vergl. Anatomie d, Primeln, Strassburg, 1875. "2 INTRODUCTION. 13 for the peripheral part of the root-cap covering the sides of the apex of the root, on the other for the dermatogen-layer of the root. The origin of these, i.e. the lateral part of the root-cap, and of the dermatogen, takes place similarly to that of the corresponding parts in the third type. Stak 3 II. In the Gymnosperms the differentiation of the meristem at the punctum vegetationis of the root? is essentially different from the types described for the Angiosperms (Fig. 6). A plerome-cylinder with sharp contour occupies the centre (p-~). The longitudinal rows of cells which compose this converge at the rounded apex towards a small initial group of cells. The plerome is surrounded.by a mantle of periblem consisting of many (e.g. in Thuja occidentalis of 12-14) concentric layers arranged with considerable regularity. Each one of the inner layers (in Thuja, 8-10) of this mantle has its initial group above the apex of the plerome. The division of these cells per- pendicularly to the surface (i.e. ra- dially) brings about the increase of the surface-elements of the layer. At the same time successive divi- sions parallel to the surface, that is, a doubling of the layers, takes place in the apical region (Fig. 6, 2). Since the radial walls of the successive layers fit almost | exactly one on another, the cells of the periblem mantle are ar- ranged above the apex in cor- tespondingly regular rows. As the layers are pushed outwards above the apex by their succes- - sive doubling, division ceases in them, and increase of volume of the cells takes place; those which FIG, 6.—(190) Funiperus Oxycedris, Median longitudinal section through 4 the apex of a lateral root. #—# plerome, surrounded by.about sixteen laye happen to be outermost at the zs F ehien coon uf Hs of periblem, the outermost of which represent the root-cap ; 7 the initial region for-periblem.and plerome, apex become gradually loosened, . and pushed off as a root-cap. Here then it is not possible to distinguish a layer of calyptrogen or of dermatogen; the outermost periblem acts as root-cap covering the ‘meristematic apex. The radially arranged apical prolongation of the periblem is in all cases relatively strongly developed, its height is usually equal to or greater than the whole diameter of the root, rarely (Taxus, Cycas circinalis) it is smaller. According as it is more strongly developed, the arrangement of the cells in rows is more clearly apparent; e.g. Pinus, Ephedra, Zamia integrifolia. a Strasburger, Die Coniferen, &c., p. 340.—Reinke, Morphclog. Abhand). p-1.—Janczewski, Le 14 INTRODUCTION. The differentiation of the meristematic apex of the stem of the Gymnosperms" shows a varying character, which couples it on the one hand with that which appears in the typical Angiosperms, on the other hand with that in the Lycopo- diaceze, while in Araucaria brasiliensis, also in A. Cunninghami, Dammara, and Cunninghamia, the dermatogen, periblem, and plerome remain clearly distinguished, in the extreme apex; in the Abietineze and Cycas these layers run together into a common initial group, which occupies the extreme apex; a separation into the three layers first appears at some distance beneath this in Cycas, and more clearly in the Abietinee. Ephedra is specially interesting, for here, in the same species (E. campylopoda) and apparently fluctuating in the same shoot, the character varies between the two extremes described. At one time there is a dermatogen-layer, sharply defined throughout its whole course, covering the two inner layers which in the extreme apex are more or less clearly separate; in other cases both merge with the dermatogen into a common superficial initial group. A series of similar cases, some corresponding to Araucaria, others approaching the other extreme, were made known by Strasburger’s researches on Taxus, Podocarpus, Saxegothea, Ginkgo, Thuja, Cupressus, Sequoja, and Cryptomeria. ; Here, as in the Angiosperms, the dermatogen and periblem alone take part in the origination of leaves and normal lateral shoots, and in most cases also in the same fashion as there. In the Abietineze, however, there occur also divisions parallel to the surface in the dermatogen of the young leaf. III. As already intimated, the differentiation of the meristematic apex of the Lycopodiacee* corresponds closely with that of the Gymnosperms. The extreme apex of the stem is occupied by a group consisting of 2-4 prismatic cells with their longer axis at right angles to the surface. ‘This is the initial group for the periblem and dermatogen, or rather for a superficial layer which corresponds to the latter. All these initial cells divide-by walls perpendicular to the surface, and the products of this division, as they are removed from the apex by the advance due to the growth of the latter in length, divide again parallel to the surface, forming thus the initial layers of dermatogen and periblem. A plerome-cylinder, which is limited laterally by the periblem, elongates independently through the activity of an initial group or single cell of its own, which occupies the centre of its conically tapered apex, and lies imme- diately beneath the initial group of the outer layers. As Hegelmaier has already as- serted, conditions are to be found, which point to the origin of the initial cells of the plerome from the common initial group at the surface of the apex (by transverse divi- sion). It is thus possible that a fluctuation of the definition of layers occurs here simi- lar to that in the Coniferee. The formation of the leaves starts in the Lycopodiacere from one cell of the outermost (dermatogen) layer, which, after arching outwards, - divides first parallel to the surface, then further. The differentiation of the meristem- atic apex of the root is, according to the investigations of Strasburger and of Bruch- man, similar in the Lycopodiacez to that in the first type of the Angiospermous roots. 1 Strasburger, /.c. p. 323, Taf. 22, 23, 25.—Pfitzer, in Pringsheim’s Jahrb. VIII. p. 56.-—[Dingler, Botan. Zeitg. 1882, p. 795-] ? Cramer, Pflanzenphysiol. Unters. Heft III. p. 10.—Strasburger, Coniferen, p. 336.—Hegel- maier, Bot. Zeitg. 1872, p. 798 ff., 1874, p. 773.—Bruchmann, Ueber Wurzeln von Lycopodium und Asoetes, Jena, 1874.—-Compare also Russow, Vergl. Unters. p, 176. INTRODUCTION. 15 The investigations of Bruchman assign to the root of Isoetes a conformity with the ‘third type of Angiospermous roots. According to the results of this observer, which harmonize with those of Hegelmaier’ with the exception of.one not very im= portant difference, the apex of the stem of the Isoetez is occupied by a small group of initial cells, which are common to the whole meristem. Longitudinal divisions of these form the mother cells of the peripheral layers of meristem, and renew the initial cells; transverse divisions of them supply new elements to the central part of the meristem. A division into the three distinct layers cannot be seen. Very similar to the structure of the apex of the stem of Isoetes is the differentiation of the meristem peculiar to some Selaginellas, and the roots of Marattia. It should therefore be mentioned here, but for the sake of shortness the description of it will be supplied further on’, ; IV. The stem of Isoetes and the above-named Selaginellas and Marattiaceze form the transition between the forms of differentiation of the meristem already described, and that which prevails for the great majority of the Pteridophyta (comp. Fig. 7-9). The characteristic in these cases is this, that the entire meristem of the apex originates from one single common initial cell, which is called, from its position at the apex of stem and root, the apical cel/*. Successive bipartitions divide the apical cell in each case into an apical part, which retains the original position and form, this being compensated again by growth, and remains as the apical cell; and a basal inferior part, which is added to the growing meristem. The latter part is termed the segment-cell*. Further divisions of the segment-cell form the meristem and later tissues. Each portion of meristem, which originates from a single segment- cell, is called a segment. In roots, besides these processes, there is in addition the formation of root-cap, which also originates from the apical cell; this must now be provisionally ignored. The apical cell (Fig. 7-9) has in most of the present cases the form of a three- sided pyramid, with convex base, which is the apical surface (i.e. the outer wall); while the sides are sunk in the meristem. This is the case in all roots of the plants in ques- tion (except those of the Selaginellas, in which the form of the apical cell is doubtful), and in the majority of the apices of stems. In other cases the apical cell has the form of a two-edged wedge, the arched base and the point having otherwise the same ar- ‘rangement relative to the other tissues as in the cases with the three-sided cell: apex of the stem of Salvinia, Azolla, many Selaginellas (S. Martensii®, Kraussiana), and Poly- podiaceze (Pteris aquilina), Polypodium rupestre, Lingua, aureum, punctatum, phyma- todes, Platycerium alcicorne, stolons of Nephrolepis undulata according to Hofmeister’. In the stolons of the last named species, as the apex becomes stronger, the apical cell assumes the three-sided pyramidal form. In Polypodium vulgare it alter- nates between the two (Hofmeister). In the seedling of Selaginella Martensii the apical cell of the main shoot, A Bot. Zeitg. 1874, p. 481. ® (Holle, K. Ges, d, Wiss. zu Gott. 8 Jan. 1876.} “8 Nageli, Zeitschr. f. wiss. Bot. II. p. 121 (1848), III. p. 157. * Pringsheim, Jahrb. f. wiss. Bot. III. p. 491. 5 [M. Treub, Recherches sur les Organes de la Vég. du Selaginella Martensii, Leide, 1877.] ® Hofmeister, Beitr, zur Kenntniss der Gefasskryptogamen II. Abhandl. d, k. Sachs, Gesellsch. d. Wissensch. Bd. V. 16 INTRODUCTION, and of the two branches of the first bifurcation, has, by reason of corresponding divisions, the form of a four-sided pyramid, which however soon passes back to the two-edged form ?. ian Each segment is separated from the apical cell as a tabular cell by a division ‘wall, which is approximately parallel to one of the sides of the apical cell. This wall is called the principal wall of the segment®. Each segment has two principal walls, one (acroscopic) by which it was separated from the apical cell, the other (basiscopic) that which adjoins an older segment. Its omer wall is that part of the free wall of the apical cell which is cut off by the line of junction of the acroscopic principal* wall; its Jaferal walls are the parts cut off by the lines of junction of the same principal wall from the principal walls of the segments, which border it laterally. The principal walls, which cut off the successive segments from the apical cell, are parallel 2% regular successton to the sides or principal walls of the latter. Thus in the case of a two-sided apical cell they oppose the one and the other side of it alter- nately, each fronting the older principal wall; in the case of a three-edged apical cell, they oppose the three sides successively in spiral’ sequence, each facing the third oldest principal wall, and being attached laterally to the two younger principal walls, All the segments therefore of a meristematic apex are arranged (if we ignore sub- sequent displacement) in as many straight rows parallel to the axis as the apical cell has sides. The principal walls of a segment, which has recently been cut off, are inclined to the theoretically straight perpendicular axis of the meristematic apex at an acute angle, which varies according to the form of the apical cell. As growth proceeds, the form of the segment alters, and with it the direction of the principal walls (or rather the planes in which these lie), so that, with reference to a perpendicular axis, they assume a horizontal position. For a thorough discussion of these phenomena, and of the growth of the apical cell itself, cf. Nageli and Leitgeb, Zc. p. 91. The figure 7 A, which is borrowed from these authors, may present the process to the eye. . The segments cut off from the apical cell become gradually divided up into masses of meristem consisting of several or many cells, As the result of the changes of form and position already mentioned as accompanying the joint growth, each ‘of these masses represents part of a more and more horizontal disc, and meets the similarly shaped segments next in age to it at the central line. A transverse section cut at some distance from the extreme apex includes so many united seg- ments as there are straight series of these, i.e. where the apical cell is two-sided, 2; where it is three-sided, 3. The divisions proceed rapidly, and if, as must be done in this case, one disregards lateral outgrowths, they proceed in the same direction and in the same succession in the successive segments. One thus finds the segments of each transverse section in about the same stage of division. In those cases where the successive divisions have been successfully and accurately followed—apex of the stem of Equisetum, Azolla, Selaginella Martensii, partially also in Salvinia, and especially in the roots of Equisetum, Azolla, and many 1 Pfeffer, Entw. d. Keims v. Selaginella.—Hanstein, Bot. Abhandl, Bd. I. ? Cramer, Ueber Equisetum. in Nageli und Cramer, Pflanzenphysiol. Untersuchungen, 3 Heft, p. 21 (1855): INTRODUCTION, 17 Filices and Marsileaceea—we may distinguish in the first stages of the further development of a segment, three sorts of division, differing in their direction and results. They are :— (1) Divisions into flais (Etagentheilungen), that is, splitting of the segment into similar stories one above another by division walls at least approximately parallel to the principal walls. (2) Radial halving, division of a segment into halves lying side by side, but never quite alike, by an approximately (but not exactly) radial wall: in the case of segments arranged in two series, that is, which correspond in the circular transverse section to semicircles, this radial halving divides the section into (unequal) guadranis; in case of segments in three series, into sexéanis; the walls in question are named accordingly. In the first case the division into quadrants is followed either by a second halving by octant-walls (stem of Salvinia, Azolla) or only each larger quadrant is again halved, so that each segment is divided. by two radial walls into three cells (stem of Selaginella Martensii). (3) Dzvision ino straia (Schichtentheilung), that is, division by tangential walls into concentric strata parallel to the surface. These three modes of division, which appear as the first successive stages of division, are followed by further divisions in the three principal directions in each story, and in each stratum. ‘These divisions alternating variously according to the species, finally result in the definitive composition of the segment. Of the above three first stages of division, that marked (3) is seldom the first. Using these figures as above to express them shortly, they usually appear in the succession 1, 2, 3 (apex of the stem of Equisetum, Salvinia), or 2, 3, 1 (root of Ferns), 2, 1, 2, 3 (apex of stem of Azolla); only in the root of Azolla the suc- cessiofi 3, 2, &c. was found by Strasburger. Relatively to the future layers of meristem, ie. to the later developed tissue-layers, the first products of division of the segments are thus, with the exception of the last mentioned case, still common initial cells. _ From the division into strata, marked (3) above, arise layers of meristem, which correspond in their arrangement to the three principal layers of the root of the Angiosperms, i.e. plerome, periblem, and dermatogen. In many cases, though not in all (e.g. in the roots of Ferns and Equisetum), these undergo a similar develop- ment to that of the similar layers of corresponding members of the Angiosperms. They are usually sharply defined, since the walls separating them (like other longitu- dinal walls) in the successive segments fit pretty accurately one on another. As is evident from what has been said, we have to deal, in the phenomenon in question, with more than one, at least two, successive divisions. As is evident from the foregoing account, many differences peculiar to special cases obtain in the very first stages of division. This is the case to a still greater extent in the later stages, which bring about in the segments their definitive compo- sition. To enter with uniform minuteness into the peculiarities of individual cases would lead us much too far. After referring to the special literature, and particu- larly to Strasburger’s description of the many peculiarities of Azolla, we need cite here only a few examples, keeping an eye at the same time upon many relations of ’ form, which have not been touched upon in what has gone before. c 18 INTRODUCTION. In the roots of the Equisetums, Polypodiacex, and Marsileacex (Fig. 7, 8) the division of each segment (4, 4, Fig. 7) begins with the appearance of the sextant wall s. This stands vertically, and, as before stated, is nearly, but not exactly radial; it is attached to the middle of the outer wall of each segment, its inner edge, however, does not extend FIG, 7.—Scheme of the succession of cells in'the ~pex of the root of Equisetum hiemale, after Nageli and Leit- geb. 4 longitudinal section ; B transverse section at the lower end of 4; /.principal walls, s sextant walls, c{‘cambial “ * wall’) the first, e (epidermal wall) the second, » (cortical wall) the third tangential wall ; the successive further tangential divisions between ¢ and ~ are figured.z, 2, 3. # : i . In 4 the figures I—XV1 denote the successive segments; the letters 2, 2, #, , B, the successively older portions of the root cap; 9 epidermis (dermatogen). From Sachs’ Textbook, to the central angle of the latter, but,-curving slightly, meets the central part of the | . lateral wall further from the centre than the angle. The convexities of successive sextant”, walls are as a rule, but not always, homodromous, and turned toward the ascending side of the segmental spiral. The sextants of one transverse section are therefore alternately unequal in form and size, according to the distance of the point of junction of the sextant walls from the angle ofthe segment ; among the cases observed this inequality is greatest » in Equisetum, least inthe Marsileacex. The inequality of the quadrants, octants, etc. of . a transverse section from the above-named plants with two series of segments depends # FIG, 8.—(2g0) A longitudinal section through the apex of the root of Pteris hastata. # transverse section through’ the apical cell of the root and the neighbouring segments of Athyrium filix femina, both after N’ ageli and Leitgeb; .. v apical cell; the other letters and figures as in Fig. 7. From Sachs’ Textbook. os a upon similar conditions. Each sextant is in the second place divided by a tangential wall | (c) into’an inner cell which is usually small, and ‘a larger outer one: the difference in sizé | - between the two is greater the thinner the root is, but, as stated, always so that the outer! cell has the advantage.. The inner cell is the initial cell of the plerome, the outer is in the INTRODUCTION: 1g simplest cases further divided by a tangential wall (e) into two cells, of which the outer is the initial of the dermatogen, the other of the periblem. The dermatogen remains in the simplest cases a simple layer, since its cells suffer only radial divisions alternately hori- zontal and vertical, but it may also be once more divided tangentially. The two other layers are further divided and developed by successive vertical walls in definite order, which need not now be followed; later these are accompanied by transverse divisions into flats (Etagentheilung). In roots which increase much in thickness each outer cell, after the appearance of the wall c, may further divide by a radial vertical wall into two, in which the division by the wall e then takes place. t . . * x FIG. 9.—Apex of the stem of Equisetum. 4 (sso) longitudinal section through a strong shopt of E. telmateja, B view of the apex-of such a sheot from above (Sachs); C,D, £, from E. arvense, after Cramer; Ca diagrammatic’ ground plan of the apical cell and of the youngest segment, ‘D optical longitudinal section of the apex of a stem, # transverse section through / in D, Sin all cases the apical cell. The Roman figures J, H,... (in # read IV for VI) indicate inal cases the segments; the Arabic figures 1, 2, 3... the successive division walls within a single segment; the letters a,d,...in C and D the-successive principal walls. In 4, x,y indicate the highest, youngest rudiment of an annular swelling which will develope into a leaf-sheath; 54 the same older, 4s apical cells of a still older leaf-rudiment, g rows of cells from which the vascular bundles are derived,:z the lowest layers of cells of the segments, 7 the rudiment of the cortex of the internodes; the broad central band between g and g the plerome. From Sachs’ Textbook. ? 2 < a can - ‘In the apex of the stem of-species of Equisetum (Fig. 9) each segment-cell is, according ‘ to Cramer, Reess, and Sachs, divided first of all parallel to the principal walls into two nearly . similar stories (B,C, D), theri follows in each of these the division into alternately dissimilar sextants (F) as in the root; abnormally in many cases (comp, Reess, Pringsh. Jahrb. VI, Pl. X. 8) two sextant walls curved in different directions appear in one segment. The next: division in the sextants is either a tangential perpendicular one into an outer and an inner cell, corresponding to the scheme for the root, or one not exactly radial and perpen- dicular, which is then followed by the tangential division (£). Both cases. happén side C2 20 INTRODUCTION. by side; the arrangement of the inner cells, which may be called initial cells of . plerome, and which suffer thenceforth divisions alternately in all directions, is therefore often irregular. In the outer cells rapid divisions now follow, sometimes parallel to the principal walls, sometimes radial, and tangential-perpendicular. ‘The definite ar- rangement of these is not ascertained. From these, only when they have attained a very advanced stage of development, a superficial layer is marked off, which may be called dermatogen. In the three first stages of division of the segments (which only are well ascertained in this case), the apex of the stem of Salvinia corresponds, according to Pringsheim’s statement, to that-of Equisetum, with such differences as follow from the rows of seg- ments being two in number. 4 For the majority of the apices of stems of the Ferns’ it is doubtful, and requires further investigation, whether and how far the first stages of division of the segments correspond to the scheme derived from the simpler cases above described. At all events, we know from the older statements of Hofmeister (Beitr. II) that the segments undergo directly many and repeated divisions, both in directions parallel to the prin- cipal walls, and radial and tangential. By these the growing meristem is cut up into many layers and rows of cells, which are arranged similarly to the segments, but in which the boundaries of the single segments are not clear. A permanent layer of dermatogen is first distinguishable after numerous tangential divisions; a boundary between plerome and periblem is for the present doubtful. The formation of the leaves begins, in plants which grow with an apical cell, . from an initial cell cut off from a segment; and the leaf itself grows, at least in its . early stages, with an apical cell which forms segments (Tig. 9 A, 45). In the roots of the ferns in question, the formation of the cap starts also from the apical cell, and begins by the cutting off of a cell from the otherwise unaltered’: apical cell near its apical surface, by a wall perpendicular to the axis. This cell has the form of a flat segment of a sphere, and is the first cap-cel/ (Fig. 7, 8, A, 4). This is the initial either of one of the simple layers of cells or sheaths (/, m, , p), froma combination of which the root-cap is built up, or, by undergoing a subsequent transverse division, it is the initial of a pair of such sheaths. Each primary cap-cell is immediately divided by longitudinal walls at right angles to its outer surface, which becomes more and more- convex as the growth at the apex of the root proceeds. It is divided first by a median longitudinal wall into two equal halves, these being again divided into four quadrants by a wall at right angles to the first Each quadrant cell is again divided into two unequal parts by a longitudinal wall, which halves the outer walls, or divides them into unequal parts, and then taking a curved course inwards attaches itself to one of the lateral walls. The further divisions which appear in the eight cells of the sheath thus formed become successively more irregular, and may be followed up in the work of Nageli and Leitgeb. Where the primary cap divides into two, this happens after the first three longitudinal divisions, are completed. et According to Hofmeister, Hanstein, Nageli, and Leitgeb, the rule is that a primary cap-cell is cut off from the apical cell after each cycle of segments, which go ? On the phenomena in Ceratopteris, which differ from those in:the other Feras in the narrower sense, compare Kny, Entwickl. d. Parkeriaceen, Abhandl. d, K. Leo». Acad. Bd. 37 (1875). INTRODUCTION. 21 ‘to form the body of the root. In longitudinal section therefore in the younger transverse zones, each successive segment is laterally overlapped by a new cap (or, as the case may be, by a pair of them). There are, however, exceptions to this rule. Further, each cap-cell abuts on the principal walls of the cycle of segments cut off immediately before it, and this condition remains often for a long time recognisable by this character, that each cap rests with its margin upon the step-like outer walls of two successive segments. This arrangement is obliterated sooner or later by the smoothing down of the steps resulting from growth. “We must now return to the phenomena in many Selaginellas, and in the Marattiaceze, which were before left unexplained. As before stated, a number of species of the first-named genus have on the stem a two-sided apical cell, which forms segments from its two sides. Russow? first drew attention ‘to the fact that in many species—e.g. S. arborescens, Pervillet, Wallichit, Lyallii—there is not a single apical cell, but an apical group of common initial cells. Strasburger? has carefully investigated S, Wadhchit, and found that here, in place of one apical cell, “wo are present, which form segments in conjunction with one another, Each of them has the form of a wedge with narrow rectangular section, and is bounded by five planes ; 2. e. two nearly equal lateral planes, in form of isosceles triangles, the bases of which are the long sides of the rectangle presented by the cell in transverse section; two narrow-rectangular lateral planes, and a fifth also narrow-rectangular, which, is the free apical plane: the lateral planes, like those of simple apical cells, are sunk in the meristem. Both cells are joined by one of their broad triangular lateral faces into a double wedge of corresponding form, and the relative position of this is such that the two triangular lateral faces are perpendicular to the dorsal and ventral faces of the (bilateral) stem, while the joint wall of the pair of apical cells is in a median position, One may therefore shortly name the broad triangular faces lateral, and the narrow rectangular ones the upper and lower. Segments are formed similarly in each of the two apical cells in the following succession. First a principal wall parallel to the lateral faces cuts off a segment almost similar to the apical cell, then two narrow segments of rectangular section are cut off by principal walls parallel to the upper and lower sides. After these follow again lateral segments, &c. Thus four straight rows of segments arise, as from a single four-sided apical cell, the series right and left being wedge-shaped, while the upper and lower segments are rectangular, and are arranged in each case in a double series. The latter form the ventral and dorsal portions of the stem, the former the lateral portions. In the Marattiaceze the apex of the stem is as yet but little investigated, Hofmeister (Beitr. II) ascribes to Marattia ‘cicutefolia a three-sided apical cell. The meristematic apex of the root of these plants, as shown by Harting *, and more carefully described by Russow (/.¢., p. 107), consists of a numerous group of large, polygonal, pyramidal, common initial cells; cap-cells are cut off from these by trans- verse walls near their broader, outer, or apical face: from these the root-cap is formed: near their inner face, cells are cut off which as initial cells form the plerome cylinder. 1 Vergl. Untersuchungen, p. 176. [Cf. Schwendener, iiber Scheitelwachsthum mit mehreren “Scheitelzellen, Botan. Zeitg. 1880, p. 716.] : 2 Botan. Zeitg. 1873, p. 115. 3 De Vriese et Harting, Monogr. des Marattiacées, p. 41, Taf. 4. 22 INTRODUCTION, Further they divide by longitudinal walls, which are similarly directed to their lateral faces but are otherwise apparently irregularly arranged, into daughter cells, of which those nearest the apical point always retain the properties of the common initial cells, the others, as they retreat from the apical point, form the peripheral layers of meri- ‘stem, dividing first by numerous repeated tangential longitudinal walls, which are fol- lowed by others in radial. and transverse directions. In this case also a separation between periblem and dermatogen appears first in an advanced stage of development. Lastly, we may notice shortly the meristematic apex of Ps¢/ofum, which, according to Strasburger, shows according to the quality of the shoot, either a simple apical cell, or an initial group consisting of many members. Reference may be made to the ‘investigations of Nageli and Leitgeb, and of Strasburger?. The foregoing summary shows, first, that the similar differentiation of the meristem at the apex of stems and roots originates in a different way, that is, from different first beginnings, in the different groups of the vegetable kingdom, and in such groups as the Selaginellas and their allies, which are intermediate between the Jarger divisions ; it originates differently even in the single species. Returning to the question, whether in all cases only definite zones of meristem give rise to definite sorts of tissue, the most general answer, according to our present knowledge, is a distinct negative. To be sure this negative does not hold for all single cases. For instance, for the large majority of roots, not only does each of the different layers of meristem correspond to a definite section of a definite system of tissue, but even the separate-parts of each of these sections may often be traced back to its separate initial cells in the apical meristem. It is therefore in this case not only allowable, but preferable, for the sake of clearness, to term the layers of meristem ‘directly the initial layers of the axile vascular strand and its parts, or of the Epidermis, -&c., instead of using the terms selected above. But even in Roots exceptions occur. ‘The Epidermis, for instance, in the Gymnosperms does not originate from a distinct ‘dermatogen layer, so that we should properly speak of a Pseudo-epidermis, if we ‘regard as true Epidermis only the layer of cells derived from a distinct dermatogen ‘layer. In the aerial roots of most Orchids there arises from a distinct dermatogen ‘layer, as will be hereafter shown, a sort of tissue different from Epidermis. The negative, however, of the constant genesis of definite sorts of tissue, or “systems of tissue, from definite zones of primary meristem, holds to a much greater extent for leaf-forming shoots. Here also it is true there are such relations. The ° -system- of vascular bundles of many stems of Phanerogams, for instance, is derived exclusively from the plerome cylinder. The plerome cylinder of the Lycopodiacez is transformed into the axile vascular cylinder ; dermatogen means in the Phanerogams nothing further than young Epidermis, &c. But exactly the Opposite also occurs. The plerome cylinder arising from the inner cells of the segments develops in Azoll/a {and Salvinéa?) into the vascular bundle of the stem. In the stem of: Equisetum * it develops into the—chiefly transient—axile cylinder of Parenchyma, and the system -of vascular bundles develops, according to the data at hand, exclusively from the zone -of periblem. And the whole of the tissues, and tissue-systems of the leaves, which 1 Botan. Zeitg, 1873, p. 118. * [Cf. Haberlandt, Entwickelungsgeschichte des mechanischen Gewebesystems d, PAlanzen,1879.] 5 Compare Sanio, Botan. Zeitg. 1864, p. 224. : INTRODUCTION. : 23 are continuous with the similar and synonymous tissues of the plerome of the stem are formed, according to the data at hand, outside the plerome, being derived, as is the whole leaf, from the periblem and dermatogen, or from the layers of meristem corresponding in position to these. From all this we see then, that definite relations between the original differentiation of the meristem and the formation and arrange- ment of the definitive tissues obviously exist, but that these are not universally the same. If then the course of treatment is to be uniformly arranged, we must for the time being regard especially the dzsirzbution of tissues, while still keeping an eye upon the differentiation of the meristem. In opposition to the foregoing view, another has lately been asserted, since Famintzin! has undertaken to prove that in the Angiosperms definite systems of tissue, namely besides the Epidermis especially the system of vascular bundles, are universally, z. ¢. in all parts of the plant, each derived from the same primary layers of meristem, which are separated even in the embryo, and develop further independently near and between one another like the germinal layers of the animal body. The layers of meristem in question are fundamentally the same above distinguished by us. On the share taken by the dermatogen in the development of tissues there can be no two opinions; the main question therefore is whether the system of vascular bundles arises universally, ze. in the whole plant from the same primary layer of meristem. If we ignore isolated cases of controversy, the plerome or a certain region of it is in stem and root the initial part for the system of vascular bundles, or for the greater part of them. The question there- fore is whether the parts of the system of vascular bundles, which pass from the stem- system into the leaves, and which belong to the latter, also arise from the plerome at the apex of the stem, This could not be otherwise effected than by outgrowths of the plerome pushing between the other layers of the young forming leaf, and growing with these, as was above asserted for the common growth of dermatogen and periblem, Other investigators do not find this; they rather say that the vascular bundles in the leaf, like the other inner parts of it, are derived from the primary periblem, since definite bands of the latter show the corresponding differentiation ; and that they are connected with the system of the stem by reason of the relative position of the said bands of periblem?, Famintzin’s researches certainly afford valuable conclusions on certain processes, but no new result on the main question. When he proves that in foliage leaves, especially in the Papilionacez, the parts of the vascular bundle always arise from quite definite. layers of the meristem, he says nothing new; for as the mature vascular bundles in the leaf have a definite position, this must hold also for their younger stages. He does not. produce proof that these bundle-forming layers arise as branches from the plerome layer in question of the stem, and push themselves between the other tissues of the leaf, and this proof he should have brought in order to establish his view; he com- municates rather observations, which lead to the contrary result. He asserts that the leaf of the Papilionacex mentioned, e.g. species of Trifolium, at a certain age consists of five layers of meristem; the outermost is dermatogen or epidermis, of the four inner only the two innermost are points of origin of the vascular bundles, He further asserts that in an earlier stage within the dermatogen lies only ove layer of meristem cells—which; -according to our preceding statement, must be derived from the periblem of the punctum vegetationis ; and that the four later layers arise from division of the cells of that one. It is clear that thereby the postulated pushing in is excluded, and on this the theory of germinal layers must be founded, ; * Botan. Zeitg. 1875, p. 501.—Beitr. zur. Keimblatt-theorie im Pflanzenreich, Mem, Acad. St, Pétersbourg, 7 série, tom. XXII—Compare also Botan. Zeitg. 1876, p. 540. [Compare further Famintzin, Embryologische Studien, Mem. Acad. St. Pétersbourg, tom. XX VI. No. 10, 1879.] 2 Compare especially Sanio, Botan. Zeitg. 1864, Z.c. 24 INTRODUCTION. With the differences of differentiation of the meristem are always connected those of the mature structure: one may say obviously so, since the causes of the development of the mature structure, which are involved in the properties of the meristem, are different in every case. But while the differences in the differentiation of the meristem correspond in each case exactly to systematic divisions, as distinguished principally on the ground of other phenomena, and especially so in the greater groups—for instance, all Ferns and Equisetums correspond just as closely, and are distinguished from other classes just as much by the differentiation of the apex of their stem and root, as by their reproductive and embryonic processes—the case is often different with the mature structure. The structure of the full-grown stem of Equisetum has no more resemblance to that of a Fern than to that of any Angiospermous plant however distantly related, as regards both external members and internal structure. Similar divergences are found on all sides between the characters of mature plants, and the embryonic or meristematic stages which indicate their relationships. Conversely, there is equally often to be found a convergence of properties of distantly connected species: -and this is clearly expressed externally in the similarity of the most heterogeneous plants which live under like conditions, such as water-plants, the vegetation of steppes, and shores, &c. The reason of these phenomena is easily understood when seen from the point of view of the theory of Descent, and has often enough been stated, The existing form of a species is determined by the inheritance of the properties of its ancestors, and by the changes which these properties undergo through the influence of the environment, ze. the adaptation to the latter. The inherited properties must remain most clearly retained in those stages of ontogenetic development which through all generations are most independent of, that is most protected from external influences, and this is the case with the embryonic and meristematic stages. These recall in each species most completely and clearly the whole series of its reminiscences of descent, or, what is the same, they are more plainly different according to the divisions of the natural system than the later stages. Certainly these are also influenced by heredity, but the results of this may be obliterated and diverted from the original direction by successively accumulated adaptation, As in the external conformation, which must at most be only incidentally touched upon here, so in the internal structure and arrangement of the tissues we may accord- ingly distinguish two series of phenomena, On the one hand those in which we recog- nise the direct effects of the environment (phenomena of direct adaptation), since they appear in plants of the most different affinity, as soon as they are adapted to like con- ditions of life; and since they may change with these conditions of life even in the same individual. It is hardly necessary to cite as proof the different forms of growth, which recur in the most unlike circles of affinity, and the anatomical peculiarities con- nected with these; or the remarkable similarity between species of the same habitat, which are systematically as far apart as possible: of the latter we may quote as the most prominent examples, in the first place, the water-plants, the similarities of which are independent of their systematic position, and will be often referred to in the follow- ing chapters. In the amphibious species the most remarkable varieties of general conformation appear, according as an individual, or even a part of one, lives in or out INTRODUCTION. 25 of the water. Then we may refer to the almost identical form and structure of the roots of the large majority of plants however different systematically, and to the peculiarities, which at once appear in these, where a special adaptation takes place, as, for instance, in the aerial roots of epiphytic orchids, the prop- “Toots of the Pandanacee, Triarteze, &c. On the other hand, there are often to be found phenomena in the structure as well as in the form of the vegetative organs, which may also be derived from adaptations, which have happened in some epoch or other of the phylogenetic development, but which cannot now be certainly referred to their causes; properties, which were acquired at an unknown time, and through unknown causes, are handed down to definite series of successive generations, and at the present time are characters of Species, Genera, Orders, and Classes, these corresponding to those taken from the formation of flowers, embryos, &c. Of the more obvious phenomena of this category, we may mention for example the arrangement of the vascular bundles in the stems and leaves of most of the Monocotyledons and Dicotyledons, the structure of the vascular bundles of the Ferns, of the wood in the Coniferze, and in most Chenopodiacezx, &c. According to the terminology, which calls the properties, by which the divisions of the system are characterised, its characters, we may term the (unexplained) pro- perties of this category (unexplained) anatomical characters. Since the existing anatomical structure of a species is obviously the product of the combination of the two categories of properties, it is to be expected beforehand that, as with external form, in different species, it will be more identical the nearer is their affinity, and the more similar their adaptation. There are instances enough where this is the case. The Conifers, Filices, Chenopodiaceze, Cucurbitaceze may be again cited on the one hand as groups which have from every point of view a similar structure with very similar adaptation; other groups, whose genera and species have ‘very different adaptation, show accordingly very different structural phenomena, for instance, the Ranunculaceze (Ranunculus, Batrachium, Thalictrum, Clematis, &c.), the Primulaceze (Lysimachia, Cyclamen, Hottonia, &c.) To this rule however any fairly extended investigation brings to light numerous exceptions, viz. single species, genera, or groups, which, within their narrower or broader circle of allies, which follow the rule, are characterised by definite peculiarities of structure; these must, it is true, be regarded as inherited consequences of the adaptations of the special ancestors of the plants in question, which however cannot be referred to direct adaptation. Among the numerous cases, which belong to this category, and which will be mentioned in the following chapters, we may cite as ex- amples—the structure of the stem of the Auriculas (Primula auricula, &c.), which differs so remarkably from that of the other Primulas, whose adaptations are however not very different; the wood of Strychnos, Wintera, &c. Examples of this sort show how cautious one must be in adducing and. using anatomical characters for the greater systematic groups; how one must take care not to found such ideas upon the structure of a couple of casually chosen species. The frequent occurrence of such exceptional cases makes the series of phe- nomena, which are to be treated comparatively, highly complicated, and makes useless the attempt, which at once suggests itself, to arrange the single sections, which treat of the forms of tissue and their distribution, rigidly either according to the different 26 INTRODUCTION. adaptational fornis, or according to the ‘systematic divisions: Whether this attempt can ever succeed, will depend upon the results of further investigations, which shall have extendéd over whole families and classes, having regard more comprehensively and completely to @// the questions concerned, than has hitherto, as a rule, been the case. According to the present state of our knowledge there remains for the state- ment of the anatomical peculiarities of the groups, which may be distinguished according to natural relationship or direct adaptation, only this one tolerably con- sistent and practicable course, to start from, the tissues and their arrangement, and to introduce into the general consideration of these the rules and exceptions which obtain for the two kinds of groups above named. z PART 1 THE FORMS OF TISSUE. CHAPTER L. CELLULAR ‘TISSUE, General Introductory Remarks. Sect. 1. The*general properties of cellular tissue are indicated by this name, which has been interpreted above. A knowledge of the structure of the cell is here presupposed. : The sorts of cellular tissue are the Zpidermzs with its individual components, the Parenchyma with its subdivisions, and the Cork. These are distinguished in the first place by their structure, further by the form, arrangement, and mutual connexion of their cells. In earlier periods of vegetable anatomy the form both of the cells, and of the tissue elements, which are not here included under the term, was exclusively or par- ticularly regarded, and according to it were distinguished two main categories of cellular tissue (or of tissues generally); Parenchyma, parenchymatous tissue with cells, that is, elements (parenchymatous cells) of nearly isodiametric form; and Prosen- chyma, Pleurenchyma with particularly long elements, which are connected with one another laterally, and with their obliquely tapered or spindle-pointed ends (Prosen- chymatous cells). Among the former were distinguished a number of subordinate forms according to the special shape, as, Merenchyma, tabular, and stellate paren- chyma, &c., the detailed enumeration of which would now be purposeless +. One may, as is often the case, retain these names to indicate the forms; however it may be better to choose for these forms purely descriptive terms as wanted, and from this point of view to term the two above-named main categories of form on the one hand /sodiametrie cells, on the other elongated or fibrous cells, With reference to the structure of the cells, besides the special properties, according to which the distinctions between them will hereafter be drawn, a difference 1 Compare Meyen’s Phytotomie, and Mohl, Vegetab. Zelle, p. 16. 28 CELLULAR TISSUE, often occurs, which concerns the relative development of mass, on the one hand of the cell walls, on the other of the protoplasmic body and the contents. On the one hand, we have cells with a relatively thin wall, and richly developed protoplasm and contents, characterised by the components of the latter—chlorophyll, starch, sugar, inulin, &c.—as the specific organs of assimilation, and of metabolism, or- chiefly con- taining watery cell sap. On the other hand we find cells whose protoplasm and contents are reduced relatively to the strongly thickened, and often lignified mem- brane, and which accordingly, without giving up the properties of typical cells, or their part in the process of assimilation, obviously participate in the mechanical functions, i.e. the strengthening of the parts to which they belong. The ‘Collen- chymaé of the cortex of herbaceous plants and the sheaths of the vascular bundles of many monocotyledonous roots are examples of the latter condition. One can accordingly distinguish two extreme forms of structure, and call them shortly ¢hen- and ¢hick-walled cells, names which are explained by what has gone before. When with the thickening of the wall there appears a process of lignification—which in itself still needs to be more carefully studied—and a hardening of the wall thus occurs, this process will for the future be indicated by the term Scleroszs. The different grades of wall-thickening are not generally confined to a desis cell-form, or to any one of the sorts of tissue here distinguished on completely dif- ferent grounds; there exist isodiametric and fibrous cells with thin, and with sclerotic’ walls, sclerotic Parenchyma, Epidermis, and Cork cells, &c. But besides this, as may be concluded from what has been already stated, there is no sharp limit between ee the two main forms, even if one ignores the following fact, which should be brought prominently forward, that sclerosis is the commonest phenomenon of secondary metamorphosis which appears in cells. In the large majority of cases, the species and varieties of cellular tissue are dis- tinctly different from one another, and the treatment must start from these cases of marked differentiation and division of labour. But since all are derived from funda- mentally similar meristem, and the properties of the cell remain to all alike, there appear also cases of less complete differentiation aud division of labour, and-tran- sitional forms, to the existence of which attention must be directed from the very first, and which raise permanent difficulties in many single cases in the way of a sharp division of tissues. From the non-equivalent sorts of tissue which originate by metamorphosis of cells, the cellular tissues are, irrespective of their common origin, usually quite clearly distinct, But there are two exceptions to this. Firstly, a sharp limit cannot always be drawn between sclerotic cells and sclerenchyma, which has lost the cell-quality, The secondary sclerenchyma-metamorphosis, which often appears in cells, must lead to transitional forms; and practically it is often impossible to distinguish whether the cell quality remains, or is lost. In many cases therefore the question arises whether a separation of the sclerenchyma from the cellular tissues is to be attempted at all, and to be as far as possible carried out. The frequent occurrence of sharp differentiation answers the question, I think, in the affirmative. Secondly, intermediate cases exist between cells and the secretory reservoirs, in so far as the bodies termed secretions, which fill the latter, as oxalate of calcium, resinous bodies, &c., frequently appear also as constituents of the contents of typical EPIDERMIS. 29 cells, and these, as the quantity of the secretion increases, become like those reservoirs. For judgment upon these intermediate forms, and the possibility of carrying out the separation of the forms of tissue, the same reflections hold good as for the scleren- chyma. The difficulties of distinguishing them in practice are, besides, much smaller in this case than in that of the sclerenchyma. SECTION I. EPIDERMIS. Sect. 2. Epidermis, outer skin, is the name given to the layer of cells which is covered by and produces the caéicie. It constitutes the surface of such plants as are several layers of tissue thick, from the beginning of the differentiation of tissues onwards throughout their life,.or till the beginning of the development of cork, which takes its place. On the stems and leaves of Angiosperms the Epidermis is sharply marked off, even in the young embryo, while still consisting of few cells; in this case, as long as it remains in the meristematic condition, it is termed the Dermatogen layer. This grows, as stated on p. 7, with the stem, leaves, and branches, covering them as a cellular mantle, one Jayer of cells thick. It remains in the large majority of cases throughout its life a simple layer of cells, with exception of hair structures. In relatively few Angiospermous plants divisions of the young epidermal cells appear parallel to the surface, and then in a rather late stage of development; from a simple layer of cells two or several are thus formed. These assume an essentially identical structure, and are then termed many-layered epidermis. Where the differentiation of the apical meristem is other than that characteristic for the stem of the Angiosperms, a permanent outer layer of meristem, derived by successive divisions from initial cells common to it and to other layers, assumes the properties of the epidermis. In plants which grow with an apical cell, definite peri- pheral products of division of the segments serve this purpose ; in Gymnospermous roots the transverse portions of the successive layers of periblem from time to time laid bare by the separation of the root-cap, &c. Comp. above, p.14. In these cases we cannot speak of a many-layered epidermis in the same sense as in the stem and leaf of the Angiosperms, since the genetic relations characteristic of those cases are different ; that term can at most be conventionally used for single cases which in fact scarcely ever occur. In single special cases also in the Angiosperms, the epidermis originates from other elements than the dermatogen. The perforations (and indeed also the laciniae) in the leaves of many Aroideze originate by the early dying off of circum- scribed portions of the young leaf, the original epidermis dying off with them*. Since the edge of the mature perforations is clothed with epidermis, this must be completed from the inner layers of the young leaf, which point moreover remains still to be more 1 Compare Trécul, Ann. Sci. Nat. 4 sér, tom. I. p. 37. 30 CELLULAR TISSUE. closely investigated. Similar relations, which however also require further observation, may hold for the margins of the leaf-segments in the Palms, since these segments originate by the splitting of the continuous young lamina’. In one or few-layered: parts, like the leaf-surfaces of the Hymenophyllacee ? and Hydrillez *, either there: is no differentiation of the epidermis from the parenchyma, or it has been obliterated. One can in this case speak of epidermis only on the ground of the cuticular covering, which is present, or as in the two-layered leaf lamina of the Hydrillez, on the ground of genetic relations. In the many-layered parts also of submerged water-plants the differentiation of epidermis and parenchyma often becomes less plain, as will later be described. = In the overwhelming majority of cases the epidermis is sharply distinguished from the cells which it surrounds. : 1. COMPOSITION OF THE EPIDERMIS. Sect. 3. The following kinds of cells or cell-groups are to be distinguished as parts of the Epidermis :— : (1) Epidermal cells. : : (2) Stomaial, guard-cells, pairs of which enclose a slit-shaped intercellular space, and together with this form the s/oma. (3) Hatr-structures (Trichomes). ; Sxcr. 4. Epidermal cells in the strict sense is the name given to those cells. of the epidermis whose lateral walls are in uninterrupted connection with one. another and with stomatal cells. Exceptions to this occur only in the slightly. differentiated epidermis: of the base of the leaf of the Osmundacez and Isoetes (comp. Sect. 9). The term lateral walls is here used for’ all those which are at right angles to the surface. With reference to the longitudinal axis of growth of the member of highest rank to which they belong, we may therefore speak of superior or inferior lateral walls, and of side or flanking walls; and, in obvious contrast to these, of outer and inner walls. The direction at right angles to the surface, i, e. that of the lateral walls, may be called the Aeighé of the cell ; length and breadth will be used in the same sense as for the whole organ of highest rank, to which they belong. ; . Form of the Epidermal cells (comp. Figs. 10-20, which follow below). a. Epidermis one layer of cells thick, : The general form of the epidermal cells is endlessly various according to the special cases. As a rule the diameters in the two directions parallel to “the surface are equal, or but slightly different, in parts which grow slowly and equally in two or three dimensions, e. g. lamina of many leaves: but the longitudinal diameter is greatly developed in longitudinally extended organs, as most stems, roots, narrow linear leaves, especially of the Monocotyledons, also on the nerves + Compare Mohl, Verm. Schriften, p. 177. “® Mettenius, Ueber d. Hymenophyllaceen, in Abhandl. d. stichs. Gesellsch, d. Wissensch, IX, p- 403. os . Pe, wes ® Caspary, in Pringsheim’s Jahrb. I. p. 49. EPIDERMIS. 31 and ribs’ of slightly elongated leaves. The opposite, that is to say, a great transverse extension, occurs but rarely in the case of the epidermal cells in longi- tudinally extended parts, as for instance in the leaves of Cycas, Encephalartos, Tra- descantia, Crassula, Campelia, Dichorisandra', many Bromeliacez (Pholidophyllum zonatum), and as a peculiarity of stems with clearly marked nodes, as Arceutho- bium, Salicornia. The height is as a rule either much smaller than the larger, or than both of the diameters parallel to the surface, the cells are thus of the form of flat plates overlying the surface; or it is the greatest diameter of all, the cells are then prisms arranged: perpendicularly to the surface; intermediate forms between these two extremes are common enough. The lateral faces are flat, and cut one another with: sharp corners, so that the single cell has the form of an angular plate or prism. In other and not less common cases they are wavy and folded, in which case the hollows and pratuberances of neighbouring cells fit exactly into one another? The extent of the waving may vary, or undulated and flat lateral walls may both occur in equivalent parts of one and the same ‘species, according to the adaptation to different surrounding media. Meyen® noticed this relation (which remains to be further followed in other connections), somewhat indefinitely it is true, for ‘a large number of species of Gentiana,’ in which he found the cells more wavy ‘the damper the region of the atmosphere in which the plant was grown.’ Conversely’ Askenasy* found’ in Ranunculus aquatilis and divaricatus, on’ the submerged form that the epidermal cells of the leaf have flat sides, on the Jand- “form strongly undulated sides.. Also on the amphibious leaves of Marsilea and Sagittaria ° differences occur in the above-mentioned relation. The undulation usually extends equally over the whole height of the lateral wall, but often, e. g.in leaves of grasses and Equisetum %, only over the strip along the outer edge, while the inner part is flat. The outer and inner surfaces of the epi- dermal cells are flat, or to a variable extent convex; the latter either over the whole extent of the cell, or at one (e.g. leaf of Aloe margaritifolia), or two, or several (Equisetum hiemale) relatively small circumscribed spots. Other forms than those possible within the limits laid down are more rare ; e. g. spindle-shaped, elongated, on the leaves of Torreya, Ceratozamia (Kraus, /.¢.) ; the often peculiarly formed swdsidiary-cells surrounding the stomata, together with the hair structures, must be specially mentioned below. One and the same epidermal surface often shows only epidermal cells of nearly similar form—e. g. many smooth stems. Much oftener however considerable differences occur on the same surface. (a) In-relation to the relief of the surface, and (often connected. with this) in the distribution of stomata and hairs; in angular and ribbed stems in relation to the angles, or ribs on the one hand and the faces or furrows on the other; in flat leaves and leaf-like organs in relation to the ribs 1 Kraus, Bau d. Cycadeenfiedern—Pringsheim, Jahrb. Il. p. 318. 2 Treviranus, Verm. Schr. IV. p. 16.—Meyen, Phytotomie, p. 94. 3 Phytotomie, p. 95. * Botan. Zeitg. 1870, No. 13. 5 Hildebrand, ibid. No. I, 6 Mohl, Seis Schriften, p. 262,—Mettenius, He ceanaiaitioen: Pp: 444. 32 CELLULAR TISSUE, or nerves, and the spaces between these ; it is a general rule, that here the epidermal cells over the stalk and ribs are longitudinally extended, and with straight sides, but between the ribs the form and direction of the predominant elongation often alter’; further differences occur in relation to thorns, prickles, teeth, &c. The leaves with stomata arranged together in groups (Begohia, Saxifraga sarmentosa) will be noticed below. (8) Independent of relief, and of distribution of stomata and hairs. To this category belong a number of very different special cases. In the bands of epidermis free from stomata of the leaves and green stems of most Graminez the epidermis consists of longitudinal rows of cells, of which some are elongated; others, alternating pretty regularly with these, are short, that is, broader or at most as broad as they are long. The short ones stand singly or by twos or threes between two long ones; in the two latter cases a further inequality often occurs, in that the upper, or as the case may be the middle cell differs in form and structure from the others *. In the bands of epidermis, without stomata, which cover the peripheral bundles of fibres in the stem and leaves of the Cyperaceze, Douval-Jouve found one to two longitudinal rows of epidermal cells distinguished from the rest by a less prominent outer wall, and instead an inner wall projecting inwards in form. of a strongly thickened cone *. The cystoliths scattered in the epidermis of the Urticaceze and Acanthacex (Sect. 21), the elongated sac-shaped cells rich in tannin scattered or arranged in rows between the isodiametric sinuous elements which Engler* found in the epidermis of Saxifraga cymbalaria and its allies and of Sedum spurium, the large solitary cells in the small-celled epidermis of the leaf of Cymodocea nodosa, and rotunda®, are to be registered as further kindred peculiarities. Then the ‘Interstitial-bands’ on the under side of the lamina, between the nerves of the floating leaves of most if not all species of Marsilea, must be mentioned. They consist of at most three to five rows of epidermal cells, which are distinguished from the ordinary cells with undulating colourless walls by more elongated form, smaller size, deep golden-brown colour of the wall, and homogeneous fluid contents. Many appearances, to be described below with glands and hairs, are directly connected with these. b. An epidermis, more than one layer of cells thick®, appears in the simplest cases, by the division of each original epidermal cell by one or more tangential walls into chambers, which exactly fit one on another. In many cases this happens one may almost say casually to single cells, while the neighbouring cells, which resemble 2 Compare Kraus, /.¢. p. 309. . 7? Compare Bot, Zeitg. 1841, p. 149, pl. I. figs. 10, 17 (Coix), 12 (Sorghum),—Pfitzer, Pringsh, Jahrb. VIL. p.555. Here the descriptions of the older writers and the discovery by Treviranus (Verm. Schr. II) and Meyen (Phytotomie, p. 312, Taf. III. 2, 3) are cited. i. 5 Douval-Jouve, in Mém. de l’acad. de Montpellier, 1872, p. 227. The phenomenon appeared in all the species of the family which were investigated, i.e. of the genera Cladium, Rhynchospora,” ‘ Fuirena, Eriophorum, Schenus, Scirpus, Galilea, Cyperus, Carex, Kyllingia, Hypolytrum, Diplasia. * Botan. Zeitg. 1871, p. 886. 5 Magnus, ibid. 1871, p. 210. : ° Treviranus, Verm, Schriften, IV. p. 11.—Pfitzer, in Pringsh. Jahrb, VIII. p- 16, Taf, VI. EPIDERMIS, 33 them in other points, remain undivided, as for instance in the case represented below in Fig. 29 of Klopstockia, and that cited by Pfitzer of Tradescantia zebrina; or divided and undivided cells (i.e. one or several layered epidermis) stand side by side in about equal quantity, as on the under side of the leaf of Passerina ericoides, and the examples cited by Pfitzer of the leaf of Pittosporum Tobira, undulatum, of the stem of Elegia nuda, Ephedra altissima, monostachya. The upper surface of the leaf of Arbutus Unedo has two layers with their cells fitting one upon another (not taking into account single cells, which remain uniseriate, and undivided), those of Begonia manicata 2-3 (Pfitzer, /.c.), the stem of B. tomentosa 21, that of Pe- peromia blanda2*. In the families to which the three last-named plants belong, the Piperaceze and Begoniacez, and further many species of Ficus (Fig. 18), there is formed on the leaves a much stronger many-layered epidermis, which is divided and developed in a much more complicated way. Pfitzer describes for Begonia sanguinea, ricinifolia, and peltata an epidermis of 4 to 5 layers, while that of B. Drégei and Fischeri on leaf and stem is simple, that in B, Drégei, however, consisting of very large cells. The petiole of B. manicata has a simple epider- mis, with only solitary tangentially-divided cells; the lamina has on the upper surface 2-3 layers with the cells fitting one on another; the inner of these is much higher than the outer; on the under surface (Pfitzer, /.c. Taf. VI. 9) it has two layers, the cells of the inner being more than double as high and broad as those of the outer—this results from the fact that after the tangential division, which separates the two layers, further radial division goes on in the outer, while in the inner only growth of the cells, without division, takes place. In the Piperacex the upper surface of the leaf of all Peperomias in which the point has been investigated® (P. pellucida, magnoliifolia, blanda, pereskiifolia, rubella, galioides, polystachya, incana, arifolia, obtusifolia, argyracea) is provided with an epidermis of more than one layer, while that of the under side is a single layer. In P. arifolia it has two, in others, e.g. P. blanda 2-4, in P. incana 7-8, in P. pereskiifolia 15-16 layers. The high number of layers, and, in those cases where the number is smaller, the con- siderable size of the cells in the inner layers, givés to the epidermis in question a vast thickness, so that it is even in P. incana thicker than the whole remaining mass of the thick fleshy leaf; in P. magnoliifolia and rubella, it exceeds several-fold the rest of the substance of the leaf, and in P. pereskiifolia it exceeds it seven-fold. According to the species the cell-division and growth either proceed simultaneously in all the layers, so that all fit with their cells one upon another ; this is the case. usually in those with two layers, but also in the many-layered P. perestiifolta, where only the outermost layer is, as the result of divisions perpendicular to the surface, smaller-celled and differently arranged from the numerous inner ones (Pfitzer /.c. Taf. VI. i); or (e.g. P, incana) the outer layers become smaller-celled than the inner layers, owing to numerous divisions perpendicular to the surface, and the arrangement of the cells corresponds less in the successive layers. Of the other Piperacez a two-layered epidermis on the upper side of the leaf was found by Treviranus in Chavica maculata, and by Payen in Artanthe colubrina. Miq. As in the Peperomias, so in many species of Ficus, the many-layered epidermis of both surfaces of the leaf is produced by the division of an originally single layer. This stratum . becomes smaller-celled as one proceeds from the innermost to the outermost layer. It has been described for F. bengalensis *, elastica, ulmifolia, pectinata, ferruginea, toe 1 Hildebrand, Unters. iiber d. Stamme d. Begoniaceen, p. 20, Taf. IV. 4. 2 Sanio, Botan. Zeitung, 1864, p. 213. 3 Treviranus, Verm. Schr. IV. p. 11; Physiol. I. p. 449.— Pfitzer, /.c. p. 26. ‘ Treviranus, Verm. Schr. IV. p. 11 (1821). D 34 CELLULAR TISSUE. Carica, laurifolia, Neumanni, nymphzifolia, australis, lutescens, salicifolia’, Its thickness varies according to the species, and is on an average less on the under surface than on the upper. Certain individual cells of the original epidermis remain undivided, and grow to form the sac-shaped cystolith-cells, which project deep into the inner tissue of the leaf (§ 21). " Ficus lutescens and F. ulmifolia have on the upper surface of the leaf an epidermis of two or three layers, on the under surface of only one layer (Schacht. /.c. p. 142, Fig. 10). "A many-layered epidermis has further been described by Nicolai? and Pfitzer (Zc¢.) in the roots of Crinum bracteatum and C. americanum. Lastly, it occurs closely connected with hair structures, on many glandular spots, to be described later, e.g. in Passifiora, on the ends of the leaf-teeth of Drosera, etc. Comp. § 18,20, Sect. 5. Stomata® (comp. Fig. 10-18). Between the cells of the epidermis there lie definitely distributed pairs of cells, whose sides opposed to one another are concave, and between these a slit is left open. The slit extends through the whole height of the epidermis, forming an open communication between the surrounding medium and an intercellular: space inside it, which is called the respzratory cavity (Athemhéhle)*. The apparatus consisting of the pair of cells with the slit is called a pore or stoma® (Spaltiffnung, Porus, stoma), and the cells bordering on the slit stomatal-, pore- or guard-cells. The general form of the mature stoma is in surface view (with medium turgescence) usually nearly elliptical; rarely relatively narrow, usually widely-elliptical (in 162 out of 174 cases observed by Weiss); further it is in some few cases almost circular ®, the special forms being endlessly various according to the species. The irregular three- to four-cornered stomata in Salvinia and Azolla’ form a remarkable exception to the rule: each guard-cell corresponds to one half (in case of the ‘usual elliptical form, a longitudinal half) of the whole form; both are, under medium turgescence, curved in a half-moon or sausage-shape ; they are directly connected by their ends, and by the ends and the convex sides they are joined uninterruptedly with the surrounding epidermal cells. The concave sides are turned towards one another and bound the slit, which is usually elongated in the direction of the division-wall by which the ends of the guard-cells touch one another; in Azolla on the contrary (Strasburger, Zc.) at right angles to this direction. The transverse section of the guard-cell (Figs. 10, 11) is generally round or forms an ellipse ioe Meyen, Phytotomie, p. 311.—Miiller’s Archiv. 1839, p. 264.—Payen, Mém. présent. 4 l’acad. des Sciences, tom. IX.—Schacht, Abhandl. Senckenb. Gesellsch. I.—Unger, Anatomie u. Physiol. p. 190.—Hofmeister, Pflanzenzelle, p. 180.—Weddell, Ann. des Sci. Nat. 4 sér, tom. II. p. 271.— Pfitzer, Zc. p. 25. 3 ; , * Schriften der Physic. Gicon. Gesellsch. z. Kénigsberg, VI. p. 73. * (Cf. further, L. Reinhardt, Einige Mitt. ii.d. Entw. d. Spalt., in Russian, ef. Bot. Jahresber. 1879, p. 30.—Schwendener, Botan. Zeitg. 1882, p. 233.—Tschirsch, Beitr. z. vergl. Anat. d. Spalt. Ap- parats. Verhandl. Bot. Ver. Prov. Brandenburg, Aef Bot. Centralbl. 1881. Bd. VI. Pp. 341. Sachs Vorlesungen, 1882, p. 395-] , * Unger, Exantheme d. PA. p. 43. 5 Spaltiffnung, Sprengel, Anleitg. z. Kenntniss. d. Gewachse; Bau und Natur d. Gewdchse p- 189. Poren, Hedwig, Zerstr. Abhandl. p. 116; Rudolphi, Moldenhawer. Stomata, De Candolle Organograph. végeétale, I. p. 78. Stomatia, Link, Grundlehren, p. 108. The name ‘deena plans (Hautdriisen), later resumed by Link and Meyen, has hardly any further historical interest.—For the history of these parts, so often mentioned since Malpighi and Grew (Anatomy of Plants, pl XLVI) compare Treviranus, Physiol. I. p. 462; Meyen, Phytotomie, p. 97; Pflanzenphysiol. L . 271 : : For details compare A. Weiss, in Pringsheim’s Jahrb. IV. p. 123, &c. aes Compare Strasburger, Pringsheim’s Jahrb. V. Taf. 36: idem, Ueber Azolla, Taf. III. EPIDERMIS. 35 variously inclined to the slit, or bluntly angular; it has usually at the united ends of the cell a different form, and also larger diameter than at the part bordering on the slit, Examples, Persoonia myrtilloides and other Proteacee*, Cycas*, Psilotum, Equisetum, Coniferze, Restiaceze, Grasses, Calycanthus®, Scirpus, Iris, &c. Along the slit, but at some distance from it, run in most cases on each guard-cell two ridge-like protuberances (belonging to the membrane, see Sect. 14), one on the outer, the other on the inner surface, the corresponding ones being continuously connected at the ends of the slit. The ridges are channelled and concave on the side facing the slit, and convex on the other side,.at the free edge they are sharp, and thus appear in the transverse section in the form of sharp teeth. The.outer aperture, the ex/rance (Eingang), and the inner, the ex7f (Ausgang) of the slit are thus bordered by the sharp edges of the ridges ; through the edge of the entrance one enters into the froni cavity (Vorhof), which widens out between the channelled faces; through the edge of the exit into the similarly formed but usually much smaller dack cavity (Hinterhof); the pore-passage (Spaltendurchgang *), which FiG. 11.—Cross section through the leaf of Pinus Pinaster. s guard-cells; # passage of the FIG. ro.—Hyacinthus orientalis, leaf, cross section. e—e stoma; uv the furrow, limited internally by the epidermal cells ; s entrance of the stoma, which has been stoma; c cuticular layers; @ limiting lamellae be- cut through transversely in the middle ; ¢ respiratory tween the epidermal and the hypodermal scle- cavity between the parenchymatous cells, f. (800). From renchymatous cells; g chlorophyll-parenchyma Sachs’ Textbook, (800). From Sachs’ Textbook. widens towards both cavities, leads between the parts of the section of the guard- cells which are broadest, and nearest to one another, from the front to the back cavity. The ridges of exit and of entry are extremely various in form and size, (comp. Sect. 14), they are not uncommonly very small, especially the ridge of exit, and therefore easily overlooked. It is rare for both or for the ridge of exit to be really absent. The latter alone is absent in Elymus arenarius, Bromelia Caratas, Hakea saligna, ceratophylla, Banksia sp.; both in most observed Coniferze° (Fig. 11), Cycadez *, Ephedra, Psilotum, Azolla’. 1 Mohl, Verm. Schr. p. 248. 2 Kraus, /.¢. p. 320. ° * Pfitzer, Z.c. * ¢Figentliche Spaltéffnung,’ Von Mohl, Bot. Zeitg. 1856, p. 697, Taf. XIII. Here the subject is explained.. Many good drawings by Strasburger in his Beitrage z. Entwickelungsgeschichte d. Spaltéffnungen, Pringsh. Jahrb. V. p. 297, Taf. 35-42. 5 Hildebrand, Bot. Zeitg. 1860, Taf. [V.—Strasburger, /oc. c#t. fig. 145- ® Kraus, /.¢.—Strasburger,. fig. 143. T Strasburger, Ueber Azolla, Taf. TIT. D2 36 CELLULAR TISSUE. The size of the mature, full-grown stomata is usually smaller than the average size of the adjoining epidermal cells, often extremely small in comparison with these, e.g. Salvinia; on the same surface, e.g. the leaf-surface, it is in the majority of cases on the whole uniform with slight variations. The absolute size of the space which they occupy in the epidermal surface lies, according to the measure- ments made by A. Weiss! on 1go0 plants, between o‘ooorr™™O (Amarantus cau- datus; length and breadth =o-016™™) and 0-00459™™ 9 (Amaryllis formosissima, length 0-074, breadth o-079™™), in most cases between 0.0002™™ Q and o-ooo8m™m 5, The size of the open slit apparently bears an almost constant relation to that of the whole apparatus, but exact measurements of this have been made only for few cases. The size and form of the slit as well as of the guard-cells vary regularly in the same stoma, according to the turgescence and tension of the membranes of the guard-cells themselves and of the surrounding epidermis; this turgescence and tension depending upon the supply of water, and the effect of light and heat. The curvature of the side of the guard-cells next the slit, and accordingly the opening of the slit, may in each special case increase to a definite maximum, and on the other hand diminish till the slit is completely and firmly closed. With these changes of curvature changes in the general form of the guard-cells are in each case connected. According to H. v. Mohl insolation and supply of water, according to N. Miiller heat and supply of. water, bring about the widening of the slit? The very large stomata of Lilium martagon, candidum and bulbiferum, widen the slit, according to Mohl, to a breadth of ¥/,,,™™ to ¥/,..m™m on the uninjured leaf, at the margins of separated pieces of epidermis to 1/,™™; on the uninjured leaf of Zea mais to 7/,,,™™; on the separated epidermis of Amaryllis formosissima to ¥/,.~™™, The slit remains meanwhile always at least six to seven times longer than broad. Unger* quotes the size of the open slit in Agapanthus umbellatus at 0:000047™™ 1], of Ajuga genevensis at 0.0000137™™Q, The water-pores to be described below (Sect. 8) assume much larger dimensions, as also the stomata on the leaf of the Kaulfussias. The latter are visible to the naked eye as round holes, which are surrounded moreover by a pair of guard-cells apparently incapable of change of curvature. The absolute height of the guard-cells, after what has been already said, requires no description. Compared with the epidermal cells, or the many-layered epidermis of the same surface, the height of the guard-cells is usually insignificant, often very small; at most they are of equal height with them (e.g. Hyacinthus orientalis *), Lilium candidum °, Helleborus niger, Fuchsia * (Fig. 10). The position of stomata relatively to the outer surface of the epidermis is closely connected with these differences. When the height of the guard-cells is equal to that of the epidermal cells the outer surfaces of both lie approximately in the same plane. The same occurs in a series of cases where the height is unequal; here the respiratory 1 Pringsheim’s Jahrb. IV. 2 Compare on the mechanism, which must not here be discussed, and which is not even yet . fully explained, the fundamental work of Mohl, Botan. Zeitg. 1856, p. 697; Sachs, vol. IV of this Handbook, p. 255; N. Miiller, in Pringsh Jahrb. VILL. p. 75. [Schwendener, / ¢.] * Anat. und Physiol. p. 334. * Strasburger, /. c. fig. 14. * Mohl, de. fig. 6. ® Unger, Anat. und Physiol. p. 190. EPIDERMIS, 37 cavity situated beneath the stoma is directly bounded by the lateral walls of the neighbouring epidermal cells, e.g. the leaves of Orchis latifolia (Von Mohl, Z.c.), the very large-celled epidermis of the leaf of the Commelinacez (Strasburger, /. c. Fig. 150), Claytonia perfoliata (1. c. Fig. 120), and many others. More commonly where the height is unequal, the guard-cells lie so that their inner walls fall approximately in the same plane as those of the epidermal cells (comp. Figs. 11, 18, &c.), They form then the bottom of a depression, through which one approaches the stoma from without. This is surrounded by the neighbour- ing epidermal cells, and is often over-arched at its outer margin by outgrowths of these, so that the mouth is considerably reduced. This is the case in the majority of tough-skinned leaves and green stems ; leaf of Polypodium lingua!, Equiseta cryptopora (comp. our Fig. 23, Sanio, Linnwa 29, 385, Taf. III. Milde, Mono- graphia Equisetor.), Coniferze”, Cycadez (Kraus. 2. c.), Monocotyledons, as Aloe®, Agave *, Dasylirion, Hechtia’, Iris °, Allium, Orchidacez, &c., and Dicotyledons, as FIG, 12,—Pholidophyllum zonatum, adult leaf, under surface. 4 superficial view of a piece of Epidermis with a stoma and its subsidiary cells. # median transverse section through a stoma; the guard-cells are pushed outwards by the lateral subsidiary cells, which have been pushed down beneath them (390). Ficus elastica’, australis, Proteacea:*, Nelumbium*, Dianthus Caryophyllus, and many others. Independently of this relation of height the case occurs that the surrounding epidermal cells are so pressed against the stoma that the latter rises a greater or less distance into the air above the outer surface of the epidermis, e.g. leaves of Chrysodium vulgare **, Aneimia Phyllitidis, hirta™, Pholidophyllum zonatum (Figs. 12-16), Nerium Oleander, many Proteacez ”, Helleborus foetidus’, Rhinanthus, species of Primula, many Labiatz, Pyrethrum inodorum, &c. * Rauter, Entw. d. Spaltéffn. von Aneimia, u. Niphobolus. Mittheil. d. natur. Vereins f. Steier- mark. Bd. II. Heft 2 (1870). ? Hildebrand, Bot. Zeit. 1860, Taf. IV. 3 Schacht. Lehrb. Taf. III. p. 24.—-Strasburger, /.¢. figs. 114, 115. * Moldenhawer, Beitr. p. 103.—Oudemans, Comptes rendus, Acad. roy. Amsterdam, vol. XIV (1862). 5 Schacht. 7c. Taf. IV. pp. 9, 12.—Unger, Anat. u. Phys. p. 192. ® Unger, Z.c. p. 191.—Mohl, Verm. Schr. Taf. VIII. 7 Strasburger, /.c. fig. 133. ® Von Mohl, Ueber d, Spaltoff. d. Proteaceen, N. Act, Acad. Leopold. XVI. II, and Verm. Schrift, p. 245, Taf. VII. VIII. ® Schleiden, Grundziige, 3 Aufl. 1. p. 278. 10 Strasburger, 4c. figs. 47, 48. “4 Zc, figs. 50, 57. 2 Von Mohl, Spaltéffn, d. Proteaceen, 7. c. 8 Von Mohl, Zc. figs. 20, 21. 38 CELLULAR TISSUE. From these examples, which might easily be multiplied, the conclusion is drawn that the superficial position of the stomata is the rule for herbaceous less thick-skinned parts, and the depressed position for leathery, succulent, and thick-skinned parts ; — but that this is by no means the case without exception. Further, that corresponding parts of plants of the same family, otherwise of like nature, such as the firm leaves of the Proteaceze and Bromeliaceze, may show the most extreme diversity. As an instructive example the tender-skinned leaves of Salvinia natans may here be cited, the small stomata of which are inserted about half way up the epidermal cells, which are eight to nine times their height’. It is obvious that when the lateral wall of an epidermal cell abuts on a guard-cell, it must present a difference of form and direction, which is in many cases very slight, from those lateral walls.which do not border on stomata. The relations of height of the abutting face follow from what has been said above. The abutting face is in the one series of cases nearly plane, and set perpendicular to the surface, or inclined obliquely to it, in such a way that it converges with the corresponding face on the other side of the stoma, towards the inner face. Both arrangements occur in stomata which are even with the surface, the latter especially in stomata which project outwards. Still cases occur of stomata seated in deep hollows, which abut on their subsidiary cells with a plane perpendicular face*. In other cases the abutting face is concave towards the stoma, and the guard-cells are fitted into the hollow with their convex side, and are therefore more or less completely enclosed by their neighbours. With this is always connected a depression of the stoma (often only slight) below the outer surface: Iris, Amaryllis formosissima °, Graminez, &c. In deeply depressed stomata (cf. the examples given aboye), also in Iris and similar cases, it often happens that the abutting faces are inclined obliquely towards the outer surface, so that they diverge inwards on both sides of the stomata. In this case it comes about that the guard-cells lie mainly on the inner side of the neigh- bouring cells (compare below, Fig. 24, Equisetum). : Irrespective of the faces just described abutting on the stoma, the neighbouring cells are in many cases of fundamentally similar form to those epidermal cells of the same surface which do not abut on stomata, e.g. Lilium, Orchis*, Hyacinthus, Helleborus, Pceonia, Vicia Faba, Sambucus nigra, many Ferns, Salvinia, and many other plants from the most different families®. But in a large number of epidermal layers, especially of foliage leaves, each stoma is on the other hand bounded by one or two or several epidermal cells, differing in form and size from the rest which do not abut on stomata: these not unfrequently resemble the guard-cells themselves. These peculiar neighbouring cells of the stomata are termed its subsidiary cells, or subsidiary cells of the pore®. Their superficial form is generally intermediate between that of the guard-cells and the epidermal cells, or they completely resemble the first. In the latter case the 1 Strasburger, /.c. Taf. XXXVI. figs. 29, 30. * Restio diffusus, fasciculatus, Pfitzer in Pringsheim’s Jahrb. VII, Taf.-XXVII. figs. 1-5. ® Von Mohl, Botan. Zeitg. 1856, Taf. XIII. figs, 2, 4. : * Von Mohl, Botan. Zeitg. 1856. : > Compare Strasburger, /.¢. ° Cellule laterales, V1. Krocker, de pl. epidermide. Pfitzer,.Pringsheim’s Jahrb. VIL. p. 536.— Compare also Botan. Zeitg. 1871, p. 133; Hiilfsporenzellen, Strasburger, 2. c. . s EPIDERMIS, 39 arrangement is such that the whole convex side of each guard-cell is bordered by one subsidiary cell; the stoma thus appears to be surrounded by two pairs of cells, one pair bounding the slit, and one peripheral to these (e.g. Graminez, Proteacez, and the other examples of two lateral subsidiary cells to be cited below); often even by three pairs, since the first pair of subsidiary cells is often surrounded by a second similar pair (Hakea ceratophylla, saligna?, &c.). -If there be a difference of height between the guard-cells and the epidermis, the subsidiary cells often hold an intermediate position also in this respect; where the difference in height is great, they are of equal height to the guard-cells, or a little higher, and with them are fitted either in the outer surface, or at the bottom of the depression. Rarely the subsidiary cells are much higher than the epidermal cells; this is the case in the Scitamineze (Strelitzia ovata, Heliconia farinosa, cf. Bot. Ztg. 1871, Taf. I. and our Fig. 28 B), where they connect the stoma with the Epidermis and Hypoderma. The arrangement of the subsidiary cells may be most intelligibly described in connec- tion with the history of their development, and that of the stoma; this shall therefore here be given. The stoma itself makes its appearance by the bisection of a cell of the epidermis, which may be called its Mother-cell*. The two products of division are the guard-cells. When they separate from one another, as will be described below, a chink appears between them. : The development of the stomata takes place in the epidermis at the close of its meri- stematic (dermatogen-) stage and not quite simultaneously in neighbouring parts, so that one may find the most different stages of development close side by side. The origination of the stomata begins thus: The hitherto almost similar polyhedral cells of the meristematic dermatogen are arranged in longitudinal rows, or irregularly : either all, or the majority or only single ones of these divide into two dissimilar daughter- cells, One of these becomes the Izitial cell of the stoma, the other an Epidermal cell. Where the dermatogen-cells form rows, it is as a‘rule® always the apical, or peripheral part of the cell, which becomes the Initial cell. Exceptions to this are only known among those abnormalities or deformities which will be described below as twin sto- mata. Where the serial arrangement of the dermatogen-cells. is absent, the relative position of the initial cells is also indefinite. The wall which cuts off the initial cell is perpendicular to the Epidermis, or originally only slightly oblique; it either stretches as a plane (transverse-) wall between two lateral walls of the developing dermatogen-cells; or it is curved in surface view to a U-form, and then with its two ends it is attached either to one or two lateral faces of neigh- bouring epidermal cells, or (as a rule in Aneimia) it has the form of a closed ring, which touches no lateral wall. In the last case the initial cell is surrounded laterally bya ring- shaped cell, in the preceding case by a cell of more or less horseshoe shape, In the further growth the three following chief cases occur :— 1. The Initial cell is the direct Mother-cell of the stoma, and the epidermal cells undergo no further division. This is the case in Iris, Hyacinthus, Orchis, Sambucus nigra, Ruta graveolens, Salvinia natans, Selaginella denticuylata, Asplenium furcatum ; Silene inflata, Chrysodium vulgare, the two last have a U-formed wall, the others a plane division-wall‘*: further Aneimia has as a rule an annular wall. 1 Von Mohl, Spaltofin. d. Proteaceen, /.¢,; Strasburger, /.¢. 2 Specialmutterzelle, Strasburger, /.¢, 3 Strasburger, /.c.; Pfitzer, 4c. * Compare Strasburger, Zc. 40 CELLULAR TISSUE, 2. The Initial cell is the direct Mother-cell of the stoma. Soon after it is marked off, along each of its sides a narrow piece of the neighbouring epidermal cells is cut off by walls running nearly parallel to each of the sides: and this takes place— a. Once in each of the contiguous cells: of these there are four, two bordering on the ends, and one on each of the flanks of the stoma: there are therefore four subsidiary cells corresponding in arrangement to these: Tradescantia, species of Commelyna (Fig. 13), Pothos crassinervia (usually), Pholidophyllum (see Fig. 12), Heliconia farinosa ?, Araucaria imbricata*; or 4, 5, and more: Ficus elastica (4-5), Coniferz *, Cycas, etc. ; -also Strelitzia ovata (Fig. 28 A). om . 4. Once in each cell bordering theMianks, so that the stoma is enclosed on either side by a subsidiary cell similar to the guard-cel's, This is the case in Graminex (probably FIG. 13.—Commelyna covlestis ; leaf, development of the stomata and subsidiary cells, surface view 4 very young, B older stage; s in both the initial- and at the same time mother-cell of the stoma; C mature, S guard-cells. From Sachs’ Textbook, * all‘) and the foliage parts of other grass-like plants (Carex, Cyperus, Scirpus, Juncus lamprocarpus, effusus, Luzula maxima), Stanhopea, Aloe soccotrina, nigricans, Musa sapientum®, Claytonia perfoliata, Proteacezx (Protea®), Grevillea robusta®, Lomatia longifolia, etc. c. By repeated division of the subsidiary cells separated as in a and 4, there arises in many cases a double zone or a pair of subsidiary cells on each side. The former is: the case in Dioon’, the latter in Maranta bicolor, Commelyna communis, Pothos _argyrea, Hakea saligna, ceratophylla, and other Proteacee (Strasburger /. c.). 3. The Initial cell is not the Mother-cell of the stoma, but divides further, once or several times in succession, and the result of these divisions is a Mother-cell of the stoma, and one or several subsidiary cells. The chief types of this are :-— 1 Strasburger, 7, c.—Moldenhawer, Beitr. Tab. V. p. 94.—Meyen, Physiol. Tab. V; Phytotomie, Tab, III. pp. 4, §5.—Schleiden, Grundz, 3 Aufl. I. p. 277. i 2 Botan, Zeitg. 1871, Taf. I. 3 Strasburger, /.c.—Hildebrand, Botan. Zeitg. 1860, Taf. II. * Pfitzer, in Pringsheim’s Jahrb. VII. p. 533, &c. * Strasburger, /.¢. ® Mohl, Spaltéffn, d, Proteaceen, /.¢, 7 Kraus, Z¢. p. 335: EPIDERMIS. 41 a, The Initial cell, bounded by a curved, or even U-shaped wall, is again divided by a wall almost similar to the latter into a Mother-cell and a horseshoe-shaped subsidiary cell (Asplenium bulbiferum', Pteris flabellata (Fig. 14), cretica?); or successively by 2-3 curved walls, which alternate in two directions in the surface, and cut one another, into a mother-cell, surrounded by a zone (or in parts a double zone) of half ring- or horse- shoe-shaped subsidiary cells. The longitudinal axis of the subsequent slit is parallel to the chords of the previous curves of division: Cibotium Scheidei (Hildebrand, /. c. Fig. 37-39), Mercurialis perennis, ambigua, Pharbitis hispida, Basella, Pereskia aculenta; or FIG. 14.—Leaf of Pteris flabellata, surface view. 4 very young, ¢ epidermal cells; v subsidiary cell, s (close to 7) mother-cell, the other s initial cell of the stoma. J almost mature, s guard-cells, v and ¢ as in.4. From Sachs’ Textbook. it cuts them at right angles: Thymus serpyllum, Physostegia virginiana, and other Labiatz (Strasburger, /.c.). In the last category but one are also the Equiseta. 4. The Initial cell is divided successively by walls arranged in three directions in the surface into a simple or multiple zone of subsidiary cells, and a mother-cell surrounded FIG. 15.—Surface of leaf of Sedum purpurasceus, 4 young, the initial and subsidiary cells arising by division of the epidermal cells (e); in three of the latter the initial cell is just marked off, in four others these are further divided; the numbers indicate the successive division walls. 2 almost mature, e and numbers asin.4, From Sachs’ Textbook, by it. With few subsidiary cells: Papilionacex, Solanacex, Asperifolie, Crucifere ; with a large number of them: Crassulacex (Fig. 15), Begoniacex *, also Cactacee. 1 Strasburger, /.c. figs. 36-41. 2 Hildebrand, Botan. Zeitg. 1866, Taf. X, fig, 20-23. ; ® Strasburger, /.c.—Compare for details this work so often cited; also the not always precise statements of Karelstschikoff, Zur Entw. der Spaltoffnungen, Bull, Soc. Imp. de Moscou, 1866, 42 CELLULAR TISSUE. c. Initial cell divided by an annular wall into Mother-cell and annular subsidiary cell : Polypodium lingua (Rauter, /.c.). According to what has been said, subsidiary cells of special form originate in all the cases given under 2 and 3: in those under 1 only when the U- or annular-form of the first boundary wall necessitates special peculiarities of form. The mode of formation may be always recognised in the known cases in the mature state, but with varying sharpness, according as the subsequent growth of the cells in surface and height sharpens, _. retains, or obliterates the original distinctions. / ; Oscillations and transitions between the related types are by no means rare. For details comp. Strasburger and Pfitzer /.c. As regards occasional malformations, we must here again return to the twin stomata, i.e. those which appear in contiguous pairs, and refer to Pfitzer’s detailed statement", according to which these may arise by means of many different anomalies of division. : Two normal exceptions must here be somewhat more carefully described, First that of Aneimia, discovered by Link, later for a long time much discussed, and misunder- stood, and finally explained by Rauter, who showed that the same was the case in FIG. 16.—Aneimia hirta; leaf, epidermis. @, 6 mature; @ surface view, 4 section perpendicular to the surface, median through the stoma (375). ¢, @ very young (6co); ¢ surface view with one fully-developed stoma, and five mother-cells as yet undivided; the protoplasm of these has contracted from the delicate membrane through the process of preparation; % unicellular hair; @ perpendicular section through a mother-cell of a stoma with the sur- rounding cells, Polypodium lingua. In these cases the stoma is surrounded by ove annular Epidermal- or subsidiary cell ?. The remarkable point in this phenomenon is nothing more than that the wall of the mother-cell in normal cases has the form of a ring set at right angles to the surface, 1 Pringsheim’s Jahrb. VII. p. 551. 2 Link, Ausgewahlte anatom. Abbildungen, Heft III. Taf. IV. 8.—Ondemans, Bulletin du Congrés de Botanique, &c, 4 Amsterdam, 1865, p. 85.—Hildebrand, Botan. Zeitg. 1866, p. 245.— Strasburger, in Pringsheim’s Jahrb. V.-,¢.; also VII, p. 393, Anm.—Rauter, -, ¢,- 2 EPIDERMIS. 43 * ‘between the outer and inner wall, which touches no lateral wall, and which diminishes conically inwards, In Aneimia (Fig. 16) it, or rather the stoma, is therefore sur- rounded by an annular epidermal cell. Almost the same applies for Polypodium lingua (see above, 3.¢); the annular cell is in its turn as a rule surrounded by a horse- shoe-shaped neighbour, from which it was originally separated by a U-shaped wall. But often (Rauter, Fig. 18) also this wall is not U-shaped, but annular, the stoma is thus sur- rounded by two concentric annular cells. In Aneimia Phyllitidis, and hirta, as in Poly- podium lingua, it happens exceptionally, and in Aneimia villosa (according to Stras- burger) it is the rule that the typically annular walls are U-shaped, and attached to a lateral wall. Those of the mature parts are arranged accordingly. Further it occurs not unfrequently that from one or from both ends of the stoma (and in the former case, according to Strasburger, always the peripheral end) a membrane runs bridge-wise to the nearest lateral wall (Fig. 16, c). In face of the many attempts to explain and interpret this phenomenon it may be remarked, that from the first there is nothing more than the appearance shows at once, that is a membranous band, arranged as described, growing with the other membranes, and requiring an explanation of its appearance no more and no less than any other membrane. The second case, which is to a certain extent peculiar, but which otherwise belongs to the group (3. a), is the formation of the stoma of the Equiseta. It is here stated according to Strasburger (/,c.). The Initial cell, the first appearance of which was not observed, is nearly cubical, the two flanks being parallel to the longitudinal axis of the stem. Near to its own longitudinal axis, thus defined, there appear symmetrically, right and left, two nearly radial longitudinal walls: both are concave on the sides facing one another, and contiguous at their upper and lower ends. The initial cell is thus divided FIG. 17.—Development of the stoma of Hyacinthus orientalis. On the left the mother-cell, just divided ; 4, B successive further stages of development; on the right the formation of the slit is complete (é); the other letters as in Fig. 10, which should be compared (800). From Sachs’ Textbook, into one central biconvex-, and two lateral plano-convex daughter-cells; the two latter lessening wedge-wise inwards, the central one outwards. The central cell is the mother cell of the stoma (it divides later by a longitudinal wall into the two guard-cells), the two lateral ones are the subsidiary cells. The latter assume a form exactly similar to ‘the guard-cells, and over-arch them, so that they cover their whole outer surface, and - only leave a narrow space free above the true entrance of the stoma. Hence the form of the double pair of guard-cells apparently covering one another. In the Equiseta cryptopora of Milde the matter is further complicated by the depression of the stoma with its subsidiary cells (comp. below, Fig. 23). To form the stoma, the mother-cell divides—after, rarely before the completion of the last division, which produces subsidiary cells—into two halves, which are the guard- cells; and the slit appears thus: the division wall between the two splits in its central . part ‘into two lamelle which gradually separate from one another (Fig. 17). This separation .proceeds from the middle towards the ends, and from the entrance and exit 44 CELLULAR TISSUE, towards the passage of the future slit’, The free edges of the ridges of exit and entrance correspond to the inner and outer edge of the original wall of division. ‘The origin of the respiratory cavity by separation of the sub-epidermal cells precedes the formation of the slit. ; The mother-cell of the stoma and the products of its division are of equal height with the other epidermal cells, and lie in the same’ plane as they. The subsequent various unevenness of height and position of epidermal- and subsidiary-cells and of stomata arises through the growth of the cells subsequently to their division. During growth all cells without exception increase in volume. But the passive tension by the internal tissue, which the epidermis of growing, as also of adult parts undergoes, finally brings about, as Pfitzer has shown in the stomata of the Grasses?, a considerable ie, a Mir 1 pyry FIG. 18.—Ficus elastica; leaf, transverse section. e¢—e in each case the thickness of the epidermis; .4 (600) upper side, 4 (390) under side‘of the same very young leaf; in 4, a mature stoma, which remains superficial, and a {tran- situry) hair; in .4 two cystolith-cells, recognisable by their thickened outer wall, epidermal cells still undivided. B (600) upper side, B, (390) under side of a somewhat older leaf, epidermal cells dividing. In 2, x is a younger cystolith-cell, and <; an older one which shows the peg-shaped outgrowth of the wall. C (390) older leaf, under side; division of now three-layered epidermis is ended, stoma depressed, but the final size and form of the parts is not yet attained. upper side of a mature leaf, four-layered epidermis, cystolith-cell (375). diminution of the absolute height and breadth of the part of the stomatal cells which border the slit. (In Zea Mais the breadth soon after the appearance of the slit amounts to 11.4 p, later to 11-6», in the mature state to 5.4.) The same will apply for the other cases above alluded to, in which the part of the guard-cells bordering the slit is narrower and smaller than the connected ends, All the phenomena of development here touched upon are the same, whether the epidermis consists of a single layer, or of several. Only in the latter case (Fig. 18) 1 Von Mohl, Verm. Schriften, pp. 254-257.—Strasburger, /.¢. p. 308.—Pfitzer, Pringsheim’s Jahrb, VIL Zc. ; ? Pringsheim’s Jahrb. VIL, Zc. EPIDERMIS, 45. ‘tangential divisions accompany the extension of the epidermal cells perpendicularly to the surface, either successively from within outwards (Ficus), or the converse (Begonia, _ Peperomia). The extension and division of the epidermal cells above mentioned always begins after the differentiation of the initial cells of the stomata. The stoma itself re- ‘ mains, so to speak, always a single layer, the same is the case with the cells immediately surrounding it (subsidiary cells) in Begonia’, but in Ficus, tangential divisions appear also in the subsidiary cells, which arise according to the type (2, 2); hence a ring of sub- sidiary cells two or three layers deep”. In the growth of the epidermis perpendicular to the surface this difference here occurs, that single stomata fully developed at first, before the tangential division begins, remain at the surface; the majority are matured later, and are deeply depressed below it (comp. Fig. 18). The superficial stomata first developed are surrounded by several partitioned zones of subsidiary cells. For the relative position of the stomata the rule holds, that in elongated parts all the slits run parallel to the longitudinal axis. In parts which do not grow specially in length the slits are arranged apparently without rule in different directions. Exceptions to this rule are the stomata on the stems of Viscum album °, Cassytha, Thesium, Choretrum, Mida, Myoschilus, Anthobolus, Exocarpus, Arceuthobium, Antidaphne, Loranthus, Lepidoceras, Nuytsia*, Colletia®, Santalum album, Sali- cornia®, Casuarina’, Staphylea pinnata, on the under. side of the leaf of Philesia buxifolia. Here the slits run perpendicular to the axis of the whole organ, often (e.g. Salicornia, Arceuthobium, Colletia, Philesia) the epidermal cells are at the same time extended transversely. Secr. 6. According to special differences of form, structure, and arrangement, two varieties of stoma may be distinguished, which may briefly be termed azr-fores (or stomata), and waéer-pores. Both may occur separately, or side by side in one piece of Epidermis. : Sect. 7. The air-pores show the slit itself, during normal vegetation, filled with air; they lead from the surrounding medium directly into the respiratory cavity, which is also filled with air. Its guard-cells are, with exception of the abnormal case of Kaulfussia, we may say always capable of change of curvature, and the slit therefore of variable dilatation. They represent accordingly to a certain extent openings in the epidermis which are capable of closing, through which the air enclosed in the plant communicates with that surrounding it. Their arrangement, presence, and absence are thereby generally defined. Air-pores, and all stomata whatever, are completely absent in roots. Of the other parts of the plant hardly one can be named on which they may not, at least in many cases, be observed *. 1 Pfitzer, Ueber d. Mehrschichtige Epidermis, &c., Pringsheim’s Jahrb. VIII. /.c. 2 Strasburger, /.c. Tab. 41, figs. 135-138, and our fig. 18, both of Ficus elastica. 3 Von Mohl, Botan. Zeitg. 1849, Tab. IX; Chatin, Anatomie comparée des Végétaux, Plantes parasites, Tab. 80, 82. ‘ * Chatin, Zc. Tab. 5, 6, 57, 58: 59 64, 69, 70, 72, 77, 78, 87, 109, 110. 5 Pfitzer, Pringsh. Jahrb. VII. p. 549. 6 Duval-Jouve, Bulletin de la Soc, Bot. de France, XV. (1868) p. 139. 7 Loew, de Casuarinearum caulis foliique evolutione et structura, p. 35. ® Rudolphi (Anat. p. 91) speaks of stomata on the anthers of Le/ium bulbiferum ; Unger (Exanth, p. 127) on those of Capsella bursapastoris, ‘in a pathological state’; on the integument of Canna, Schleiden, Beitr. p. 10; also on the outer margin of the seed in Tudipa, Czech, Botan. Zeitg. 1865, p. 10}.—They exist on Perianths, both with and without chlorophyll, in many 46 CELLULAR TISSUE. The chief place where they occur is the green leaf, surrounded by air, especially the leaves of land plants and floating water plants. Certain land plants destitute of chlorophyll, viz. Monotropa Hypopitys and Neottia Nidus avis’, have no stomata at all. With the exception of the pistil Lathraa squamaria is without stomata, On the contrary, on the leaf of Lathraea clandestina®, as ,also of the Orobanchez ! and Lennoacez ®, they occur in considerable numbers, on that of the Cuscuteze‘ at Teast here and there. On Rhizomes® growing in the ground they are not uncommon, at least in isolated cases; e.g. the young potatoe before formation of the cork-layer’, the tuberous stem of Herminium Monorchis *, the rhizome of Epipogon, &c. In parts which are submerged air-pores are as a rule completely absent, but here also exceptions occur. They are to be found regularly on the submerged primordial leaves and the germinal leaf of the Marsiliaceze ®, on the submerged leaves of the Calitrichinez, Sect. Eucallitriche; Askenasy™ found single ones on the cotyledons of Ranunculus aquatilis normally unfolded under water. The statement of H. Weiss on their occurrence on submerged parts of Najas and Potomageton is not confirmed. ' In water-plants whose leaf can vegetate either submerged or in the air, as Ranunculus aquatilis, the Callitrichineze, Hottonia, Myriophyllum, Marsilia, &c., the occurrence or distribution of air-pores varies according to the above-stated habit. The air-pores occur (perhaps with exception of single cases of their solitary appearance on’submerged parts) only where intercellular spaces, containing plenty of air, are present in the tissue covered by the epidermis. Still stomata are not always present where the latter is the case. Where tissue rich in air alternates with tissue with litle or no air (Sclerenchyma, Collenchyma) there is as a rule in the epidermal tissue covering them a corresponding alternation of spots with and without stomata”. Connected with these are the universal phenomena of absence of stomata on the nerves of leaves ; their occurrence near and between these, their absence on the channels and edges of channelled leaves, petioles, stems, and their presence in the surfaces, or furrows alternating with these; (e.g. leaves of Bromeliacez, Phor- mium, Grasses, stems of the Umbelliferz, Equiseta, &c. : stomata-bearing bands and spots on the young shoots of Hedera, Juglans, Populus", on the sides, and at the plants, in others they are absent. Compare Rudolphi, Anatomie, pp. 85-91; Treviranus, Verm. Schriften, p. 50; H. Krocker, de Plantar. Epidermide (1833), p. 16: A. Weiss, Verhandl. Zool. bot Vereins in Wien, 1857; and especially Hildebrand, Einige Beobachtungen aus der PiancenAna tonite oe ads fe a ? Rudolphi, Anatomie d. Pf. (1807), p. 66. rause, Beitr. z. Anat. d, Vegetationsorg. d. L: i iss.- a = g athreea Squamaria, Diss. Breslau. 1879.] Bow- 3 Duchartre, Sur la Clandestine de l'Europe. Mém. de l'Institut de F rance, 1848, * Unger, Exantheme d. Pfl. p. 49. ; z es zu Solms-Laubach, Die Lennoaceen (Halle, 1871). . ohnfeldt Botan Zeigt. 1881, p. 38. 7 i 8 Prillieux, Ann. sci. ee 5 Ser. Vp. 265, pl. 15. aoa Soa ®° A. Braun, Monatsbr. d. Berlin. Acad. 1870, p- 665. *” Hegelmaier, Monographie der Gattung Callitriche, p. 10. " Botan. Zeitg, 1870, p. 198 “ (Cf. Potonié, Bezichungen zw. d. Spaltéfinungsystem u. d. Stereom. b. d Blatt Sticlen Filicineen, Ref, Bot. Centralbl. 1881. Bd. 8. p. 70 ] , ** Compare Trécul, Comptes rendus, tom. 73, p. 15. EPIDERMIS, 47 base of the petiole of the Ferns (comp. below, Chap. IX). The occurrence of stomata in hollow depressions on the under surface of the leaf of many species of Banksia, and Dryandra * is a special case of the same thing, which derives its peculiar appearance only from the strong. outgrowth of the nerves on the under side of the leaf. On the under surface of the leaf of Nerium oleander there alternate, between the nerves, spots with and without stomata. The latter occur in depressions of the leaf-surface, which are deep and narrow-necked, and covered thickly with hairs ?. On the part or band, which bears. stomata, the air-pores are besides in rare cases limited to circumscribed spots, separated by intermediate areas without stomata: the spots are then also characterised by a special form of the epidermal cells. Thus on the flat underside of the leaf of Saxifraga sarmentosa numerous stomata are collected in circular groups, removed some distance from one another*; on the under side of the leaf of many (but not all) Begonias, e.g. B. manicata, spathulata, Drégei, heracleifolia, two to six or more stomata stand side by side over a great common respiratory cavity *. As a rule there is over large surfaces and bands an almost uniform distribution of air-pores. Their number, both relatively to the number of the epidermal cells and to a definite superficial space, varies within wide limits according to the organ and species, and to some extent according to the condition of the surrounding medium. In the first relation we have, on the one hand, one stoma to almost every epidermal cell, e.g. in leaves of Monocotyledons, as Iris ; on the other hand, as in the stems of many woody plants, Cuscuta, &c., there is one stoma to many hundred epidermal cells®, In the other relation, the maximum numbers found for 1™™qO were 625 (under surface of leaf of Olea Europza‘*), and 716 (under surface of the leaf of Brassica Rapa’). For most foliage leaves the number lies between 40 and 300, rarely higher or lower *. As above stated, on the stems of many woody plants the stomata lie several millemeters, or still further apart, as is conspicuously shown on the formation of lenticels of Sambucus, Acer, &c. (Chap. XV). A like stage of de- velopment being of course assumed, there may be laid down for each part of each species a definite average number, which is, it is true, liable to not inconsiderable individual variations. Karelstschikoff communicates examples of individual variation. On an equal surface (the same field of the microscope, which was not measured) six leaves of Viola tricolor, each taken from a different stock, had on the under surface between 21 and 43, the majority between 30 and 40; on the upper side o to 14, the majority between 9 and 13. 1 Von Mohl, Spaltéffn. d. Proteaceen, Verm. Schriften, p. 245. 2 Amici, Ann. Sci. Nat. XXI. p. 438.—H. Krocker, /.c. p. 13.—Meyen, Physiol. I. p. 291.— Compare Pfitzer, Pringsheim’s Jahrb. VIIL. p. 49. ’ Treviranus, Verm, Schriften, IV. 30. * Viviani, Della struttura degli organ. element, tom. I. fig. 4, p. 151, quoted by Treviranus, Physiol. I. p. 466.—H. Krocker, /.¢. p. 13, fig. 39-—Meyen, Physiol, I. p. 280, Tab. V.~On the development of the groups, compare Pfitzer, Pringsh. Jahrb. VII. p. 451. 5 Compare the figures of Strasburger, Pringsheim’s Jahrb. V; Hildebrand, Botan. Zeitg. 1870, Taf. I. ® Weiss, Unters. iiber die Zahlen- und Gréssenverhiltn. d. Spaltoffnungen. Pringsheim’s Jahrb. IV. p. 124 ff. 7 Unger, Anatom. und Physiol. p. 193. * Compare Weiss, /. ¢. 48 CELLULAR TISSUE, Of the members surrounded by air, stems bearing chlorophyll are rich im stomata if leaves are absent: eg. Equisetum, Salicornia, Casuarina, Colletia, Cactacez, &c. There are eighteen stomataon1™™Q in Cereus speciosus (Krocker). Leafy stems also, whose own foliage-surface is relatively very large, are rich in stomata, e.g. Campanula patula, linifolia, Salvia glutinosa, Polygonum aviculare, Vicia Faba, segetalis, Epilobium palustre, Capsella Bursa Pastoris, Méhringia trinervia, Linum catharticum, Potentilla aurea, and many others (Unger, Exanth. pp. 98-137). Unger ascribes numerous stomata to the green branches of ligneous plants, such as Vaccinium Myrtillus, Rhamnus cathartica, and Frangula. Morren found in Prunus Mahaleb 18, and in Rosa damascena 36 on each 1™™Qq. Similar large numbers are found in related species, in Viburnum opulus, &c. (Stahl, Botan. Ztg. 1873, p. 578). In very many plants, on the other hand, very scattered stomata occur on the stems; e.g. in Prunus domestica seven, in Solanum tuberosum four on each xmm q, or still fewer; only in rare cases there are none at all. From occasional observations on the petiole similar results are obtained as for the stem. Numerous observations of their number and distribution have been made on the parts where they occur in largest numbers, viz. the lamin of green foliage leaves of land and aérial plants. The older observations of Hedwig, Von Humboldt, Sprengel, the copious works of Rudolphi, and other more scattered notices, have been followed more recently by the works of H. Krocker, Unger, A. Weiss, E. Morren, Czech, and Karelst- schikoff?. The very full statistics of Weiss inform us that of 157 species of land plants investigated, the mature foliage leaves have on an average on the space of 1™™q ‘ less than 40 stomata in 12 Species 40—TI00 ” y 42 ” 100—200 53 » 38 yy 200—300, ” » 39 300—400- yy » 12 ” 55° ” »y ot ” more than 600 3 » 3 ” The distribution of the air-pores over the surface of the leaf is in land plants directly connected with that of the air-containing intercellular spaces. It differs therefore ac- cording as the leaf shows a bifacial or centric arrangement of the chlorophyll-con- taining parenchyma, and depends in individual cases upon the number and width of the lacune in this tissue. (Comp. Chap. 1X.) . Herbaceous, flat, horizontal leaves with bifacial arrangement of the Parénchyma have usually stomata on both surfaces. Of 466 such species Karelstschikoff found this to be the case in 450. But of these 37 have on the upper surface only very few, often only solitary ones, lying near the nerves: and the majority are much poorer in stomata on the upper than on the under surface. a Firm, leathery, horizontal, also bifacial leaves, with smooth, shining upper surface, as 1 K. Sprengel, Anleitung z. Kenntn. d. Gewachse, I—Unger, Exantheme der Pflanzen (1833). Anat. und Physiol. d. Pfl. pp. 193, 334.—Compare on the older literature, Meyen, Phytotomie, p. 108; E. Morren, Determination des Stomates de quelques végétaux, Bullet. Acad. Bruxelles, tom. XVI (1864); Czech, Ueber Zahlenverhiltnisse und Vertheilung d. Spaltéffnungen, Botan. Zeitg. 1865, p. 101; A. Weiss, Ueber die Zahlen- und Gréssenverhiltn. d. Spaltoffnungen, Pringsheim’s Jahrb. Bd. IV; Karelstschikoff, Ueber d, Vertheilung der Spaltdffnungen auf d. Blattern, Bulletin Soc. Hist. Nat. de Moscou, 1866. For many details we must here refer to these works, which do not by any means coincide on all points. : EPIDERMIS. 49 Abies pectinata, Nerium, Rhododendron, Ilex, Ficus, Begonias, and many others, have stomata as a rule exclusively on the under surface: the same is the case with many firm herbaceous leaves, as in Glechoma hederacea, Asperula odorata, Trollius europzus, &c. (Karelstschikoff), Betula alba, Pirus communis, Carpinus, &c. (Morren). Rarely the relation is reversed, and with it the inner structure of the leaf also: there are herba- ceous, and even leathery leaves (Pinus sylvestris and its allies, Eryngium maritimum L., &c.) with more stomata on the upper than the under surface; or with the upper surface exclusively bearing stomata, the under surface without them, as Pinus strobus, Thuja spec., Passerina hirsuta’, filiformis, ericoides, and many Graminex with a deeply- grooved. upper side of the leaf, e.g. Aira flexuosa, Calamagrostis epigeios, Stipa pennata, &c., which will be mentioned below. Flat leaves which stand vertically, and most fleshy juicy ones (Crassulacez, many Monocotyledons), bear stomata as a rule, though not without exception, on both sides; either they are equally numerous on both sides, or they preponderate on one side or the other. In this they answer to their centric structure. Leaves, which float on the surface of the water, have stomata exclusively on the upper side, or at least chiefly so, as Callitriche (Hegelmaier /.c.), the floating leaves of Sagit- taria 2, Ranunculus sceleratus*. Further general rules, or laws for their distribution and their relations as to number, cannot for the present be laid down. No general decisive differences, either according to natural affinity and habit, or other conditions of structure of the epidermis, hold good throughout. Further, the proposition that, the more sto- mata there are on a surface, the less their size, and vice versd, is not without exceptions. Of the observed cases of variation and conformity many cannot be referred directly to immediate adaptation. For instance, of the two Lathraas, above cited, which are of the same habit, with similar members and structure, the one has many stomata on stem and leaves, the other none. But on the other hand, the occurrence and distribution of air-pores yields many re- markable examples of the change of structure by direct, often individual adaptation. This is especially the case for the amphibious water plants, and indeed all these, though they belong to the most different families and genera, as Marsilia, Sagittaria, Polygonum, Callitriche, Myriophyllum, Hottonia, Nasturtium, Ranunculus, show the same beha- viour, viz. that where numerous stomata are found on surfaces developed in the air, corresponding surfaces developed under water have fewer stomata or none at all. Marsilia quadrifoliata and other species of the genus * have, according as their habitat is submerged or not, floating leaves, with their upper side only exposed to the air and borne by thin delicate stalks, or leaves, borne on short stout stalks, rising into the air. These aerial leaves have on both surfaces almost equally numerous stomata, sunk slightly beneath the outer surface, between the strongly sinuous epidermal cells. In the floating leaves only the upper surface bears stomata, and on the same area of surface more than _ double as many as the aerial leaf. They lie between less sinuous epidermal cells (comp. above, p. 31), and in M, quadrifoliata, pubescens, diffusa, Ernesti, not depressed; in other species, as M. Drummondii, macra, they are depressed like those of the aerial leaves. A similar difference exists between the aerial and floating leaves of Polygonum amphibium, and Nasturtium amphibium®. The petiole and laciniz of the cut leaves of Ranunculus aquatilis, divaricatus, Myriophyllum, and Hottonia°, which in their normal submerged state are without stomata, form numerous stomata when they develop in the air (Land form). Caruel, Nuovo giornale botan, Italiane, I. p. 194. ? Hildebrand, Botan. Zeitg. 1870. 3 Ascherson, Botan. Zeitg. 1873, pp. 422, 631. * Hildebrand, Botan. Zeitg. 1870, p. t, Taf. 1.—A. Braun, Monatsber. d. Berlin. Acad. 1870, p. 670.—On the inconstant or exceptional behaviour of A/, Zgyptiaca and some others, compare the same. 5 Hildebrand, /.c.; Karelstschikoff, 2. c. § Askenasy, /.c. 5° CELLULAR TISSUE. Sagittaria sagittefolia has stomata on both sides of the aerial leaves, 4-5 times as many below as above, on an equal area of surface’: on the floating leaves they are very rare on the under surface, but numerous above.—In the terrestrial forms of the Eucalli- triches the stem and both leaf-surfaces are rich in stomata?, on the submerged forms they are absent on the stem, and occur only solitary on the leaves, on the floating leaves they are numerous on the upper side. A similar relation to that in Sagittaria occurs in the aerial, and the casually or abnormally produced floating leaves of Ranunculus sceleratus, An old statement of De Candolle, according to which leaves of Mentha developed under water have no stomata, is doubtful, and decidedly opposed by Rudolphi*. These facts are in accordance with the constant absence of stomata on certain submerged species, and their presence on closely-related, terrestrial species, e.g. in the genus Tsoetes. How far the finer gradations of distribution are directly caused by the mode of life and condition of vegetation requires careful investigation; in which, besides experi- mental treatment, it is important to compare, not, as has hitherto usually been the case, a large number of casually selected plants, but such as are closely related. By the latter method Pfitzer* has obtained the following result for a large number of indigenous grasses: that for these plants the number and distribution of the air-pores, together with the form of the surface and internal structure of the whole leaf, stand pretty generally in definite relation to the wetness of the locality. On both flat leaf-surfaces are numerous stomata in all marsh- and water-grasses observed (9 species, e. g. Phragmites communis, Alopecurus geniculatus): in numerous meadow- and weed-grasses (34 species, e. g. Alopecurus pratensis, Authoxanthum odoratum, Hordeum murinum, Triticum repens): among the latter Festuca elatior is a remarkable exception, in that stomata appear only on the upper side of the leaf. Almost all grasses inhabiting very dry localities have leaves with well-marked longitudinal folds; the surface is therefore marked with long and narrow furrows, and the stomata are almost exclusively on the sides of the grooves of the upper side of the leaf (12 species, e.g. Aira caryophyllea, flexuosa, Elymus arenarius, Stipa pen- nata). Koeleria cristata and Agrostis vulgaris have, with leaf structure otherwise re- sembling the latter category, numerous air-pores also on the under side of the leaf. The remaining 14 investigated species inhabit bright glades, sunny hills and grass plots; they have leaves flat on both sides, and, like the above meadow-grasses, some of them have stomata on both sides (Avena pratensis, Holcus mollis, Phleum Boehmeri, Poa bulbosa, compressa, nemoralis, Milium effusum), others—perhaps, with exception of Triodia, plants which live only in shady situations—have them only on the upper side (Brachypodium silvaticum, Festuca gigantea, heterophylla, Melica nutans, uniflora, Triodia decumbens, Triticum caninum). Milium is an exception as compared with these. Szcr. 8. Numerous phanerogamic plants, of the most various adaptation, have usually besides the air-pores other stomata different from these, which may be called Water-stomata or -pores*, since, under definite normal conditions, they serve as points of exit for excreted drops of water. These drops in many cases hold in solution large quantities of calcium carbonate, which dries into small scales. These differ accordingly from the air-pores by the slit (and the respiratory cavity below it) being, at least at times, filled with water. They are further characterised, ‘as far as investigations extend, by their guard-cells being immovable, that is, they are incapable of independent intermittent widening. In many cases this is beyond 1 Karelstschikoff, 7. ¢. ? Hegelmaier, /.c. 3 Anat. d. Pfl. p. 69. * Zc, Pringsheim’s Jahrb. VII. 5 [Cf Langer, Botan. Zeitg. 1879, p. §11.—Gardiner, Quart. Journ Micr. Sci, 1881, p.407.] 2 EPIDERMIS, 51 doubt, since here the guard-cells die off at an early stage (e. g. Tropzolum, Colocasia, Aconitum, &c.), or disappear altogether (Hippuris, Callitriche); other cases certainly require confirmation, Finally, there is often besides this a considerable difference of form and size from the air-pores, which sometimes occur on the same epidermal surface with them. The water-pores always lie over the ends of the vascular bundles, the structure of which is described in Chapter VIII; and therefore usually near the margin of the leaf, on the teeth, and, in most known cases, on their upper side: more rarely on other parts of the leaf-surface, singly or in groups, in the latter case often between epidermal cells, which differ from the rest in special form and (smaller) size. Also in closely related species there is, according to the species, in one case a single pore, in another a group of pores. The higher their number at one place, the smailer on the average is their absolute size, and also the difference in size between them and the air-pores connected with them. ‘The absolute size is in extreme cases very considerable, by ‘far exceeding the maxima for the air-pores. According to their shape, one can distinguish two extreme forms of water- pores; on the one hand those with almost semicircular guard-cells, and with a slit always quite small and shor? (Crassula, Ficus, Saxifraga); and on the other hand those with a very large, Jong slit which is always found open, e.g. the huge stomata on the leaves of Aroidew, Papaveracez, and Tropxolum. The largest of the latter are not uncommonly examples of the early death of the guard-cells pre- viously mentioned. The occurrence of water-pores is a very widespread phenomenon, to which the not very lucid statement of Trinchinetti on ‘Glandule periphylle’ réfers'. Recently Met- tenius?, and after him Rosanoff*, Borodin‘, and Magnus®, have paid especial attention to them. The form with relatively Jonger slit is known among land plants in the case of the water-dropping apices of leaves of the Aroidee; in Colocasia antiquorum °, Caladium odorum’, and C. esculentum ®, there are two or three enormously large, wide, open pores. The water-dropping spot on the middle of the under side of the hair-like leaf- apex of Richardia xthiopica has numerous widely open stomata, which are larger and rounder than the air-pores. Further, of Dicotyledons, the following cases, mostly on the authority of Mettenius, are to be mentioned. 7 One relatively very large, wide, open pore is to be found at the apex of the leaf-teeth of the Fuchsias (Fuchsia globosa, &c.), Primula sinensis (rarely 2), (comp. below, Chap. VIII); on the upper side of each tooth (and of the apex of the leaf) in Saxifraga orientalis, cuscuteformis, punctata, Heuchera, Mitella, Soldanella Clusii, Primula auri- cula, marginata, acaulis, species of Aconitum and Delphinium, Eranthis; one or qo in the same position in Sambucus nigra, Valeriana sambucifolia, Doronicum Pardalianches, Ribes triste, Prunus Padus; ‘ree in Cyclamen; a group of 3-6 of them in the same position in Ulmus campestris, Carya amara, Crategus coccinea, Helleborus niger, Geranium macrorhizum; of 6-8 in Crepis sibirica, Helenium autumnale, Verbesina virginica; an about equal, but not quite definite, number of them at the same point in * Linnea, Literaturblatt, pp. 11, 66. 2 Filices horti Lipsiensis, pp. 9, Io. 3 Botan. Zeitg. 1869, p. 883. * Ibidem and 1870, p. 841. 5 Botan. Zeitg. 1871, p. 479. ® Duchartre, Ann. Sci. Nat. 4 Sér. tom. XII. p. 264, pl. 17. 7 E, de la Rue, Botan. Zeitg. 1866, p. 321. ® Mettenius, /.¢, E 2 5% CELLULAR TISSUE, Hieracium sabaudum, Eupatorium verticiJlatum, Platanus occidentalis, Corylus Avellana, Claytonia linoides, Escallonia spec., Aralia racemosa, F erula tingitana. A numerous group of pores is to be found at like points in Tommasinia verticillaris, Archangelica officinalis, Smyrnium perfoliatum, Heracleum flavescens, Eryngium planum, and other Umbelli- fere; Cerastium glabratum, Geum agrimonioides, Aremonia, Potentilla Thuringiaca, and FIG. 19.—Tropzeolum Lobbianum ; upper surface of leaf, 4 (150) epidermis, from the margin, over the end of a strong vascular bundle, with three large water-pores; s air-pore, B, C perpendicular section through water-pores and their immediate vicinity (250). In C some of the cells surrounding the wide respiratory cavity have grown out-into large papillz rising into the cavity. . other species; Alchemilla vulgaris, Ranunculus lanuginosus, and other species; Physos- tegia virginica, Lycopus exaltatus, Hieracium Pilosella, denticulatum (apex of leaf), Rudbeckia speciosa, Senecio vulgaris, and other Composite; Valeriana Phu., Brassica spec., &c, In the examples given, the upper side of the leaf has air-pores also, Galium EPIDERMIS. 53 Mollugo and Rubia tinctorum have practically no air-pores on the upper side, at the apex they have a group of water-pores. In Papaver orientale, somniferum, and other species, 2-3 large pores lie in a small cowl-like depression on the under side of the teeth of the leaf. Tropzolum majus, Lobbianum and other species have over each nerve-ending at the margin of the peltate leaf one very large water-pore, near this 2-3 or 4-5 (Tr. Lob- bianum) additional ones which are rather smaller (Fig. 19). I have not found the pores described by Mettenius and Rosanoff on the callous middle portion of the leaf. Nelumbium speciosum has a group of several pores in the last-named spot. Among submerged water plants we know from Borodin that in Callitriche verna one large open pore lies over the end of the vascular bundle, on the tipper surface of the leaf. In Callitriche autumnalis there lie at the same spot on the young leaf a group of 3-8 open stomata; in the mature leaf the guard-cells of these break down, so that there remains a wide hole in the epidermis, In Callitriche verna also this phenomenon ap- pears in the older leaf: nevertheless I found the guard-cells still intact on leaves several months old. The apices of the leaves of Hippuris behave similarly to those of Callitriche autumnalis (Borodin). On the segments of the young submerged leaves of Ranunculus aquatilis, divaricatus, Hottonia palustris, Askenasy’ found several stomata, which die off with the whole apex before the complete maturity of the leaf. It is doubtful whether these belong to the category in question. Water-pores with a short slit are known in the case of a number of species of Crassula, and Rochea and many species of Saxifraga and Ficus with depressions on their leaves, The leaves of the above Crassulacex? have round “spots or depressions easily seen with the naked eye, either on both surfaces (Crassula portulacea, Lam., _arborescens, cultrata, tetragona, lactea) or only dis- tributed on the upper side (C. cordata, perforata) ; or forming a row just within the margin of the leaf, either on both surfaces (C. lactea, ericoides, Rochea coccinea), or only on the under surface (C. lycopo- dioides, L., C. spathulata), in the latter, one at the base of each notch between two teeth. An ending of a vascular bundle expands beneath the epidermis covering the depression. Scattered between the small, delicate cells of the latter lie, in most species, several (5-8, in C. lactea up to 25) stomata with short slits, which are smaller than the air-pores of the same leaf. In C, perforata and Rochea coccinea (Fig. 20, comp. also below, Chap. VIII) the whole depression consists of ove stoma, exceeding the air- pores in size, and somewhat sunk. The air-pores : i; ‘a FIG. 20.—Rochea coccinea; small piece of are present in most species in large numbers be- epidermis from margin of leaf. S water-pore ; : : irs ith subsidi lis, The scat- tween the large cells of the epidermis of both sur- — {.red'cpots are wartlike outgrowths of the faces of the leaf. In C. cordata they are absent from _uter wall. the upper side, which alone bears depressions. The leaves of the Saxifrages of the division Euaizonia have depressions on the notches of their margin, those of the division Kabschia (Engler) and Porphyrion on their upper side. In these depressions lime is excreted always, or at least while the leaf is young. The base of these, towards which an end of a vascular bundle runs, is constructed simi- larly to the spots in Crassula, delicate and small-celled epidermis with two (S. crustata), or 2-4 (S. Aizoon, longifolia, Rocheliana) large stomata, or one large stoma (S. retusa, oppositifolia, czsia) forming the base of the depression. 1 Botan, Zeitg. 1870, p. 235. ? Magnus, Jc. 54 CELLULAR TISSUE, The depressions on the upper surface of the leaf of some species of Ficus (F. neriifolia, diversifolia, Porteana, Cooperi, eriobotryoides, leucosticta, &c.) have in the main the same structure as in Crassula. The openings which Trécul! describes on the large prickles on the leaf-nerves and petiole of Victoria regia may be here supplementarily mentioned, being doubtful as regards their structure, and requiring further investigation. These prickles enclose a thin vascular bundle, which ends under their apex, and at the apex itself is to be found a depression with one circular opening (ostiole). Finally, while referring to later chapters, it must be remarked that the excretion of water or solutions of lime over the ends of vascular bundles is not always connected with the presence of water-pores. Sect. 9. Gaps in the epidermis other than stomata and their modifications occur normally only in rare exceptional cases. In connection with the water-pores there may here first be mentioned the cracks, which occur regularly at the apex of leaves of the grasses (seedlings of Zea, Secale, Triticum, &c.); from these drops of water are expressed. They arise by irregular tearing of the originally cowl-like apex of the leaf, when this spreads itself out flat as it unfolds.. Gaps of another sort, as found by Milde and King, occur on the middle part of the winged base of the leaf of Osmunda regalis, cinnamomea, Claytoniana, Todea rivularis, and on the ligule of the base of the leaf of Isoetes lacustris. The undulating lateral walls of the epidermal cells leave intercellular spaces between them, which are elliptical or circular in surface- view, and are often as large as the cells themselves. Their distribution is irregular: often many are near one another, even two between two cells, often there are none for a width of several cells. They pass through the entire thickness of the epidermis, and open into the intercellular spaces to be found below them. They are filled either with air or with a colourless jelly of unknown origin. No further examples can be here adduced of gaps in the epidermis, which are not to be classed with stomata; mistakes formerly made with regard to Salvinia and Azolla have been corrected ; the supposed round pores of the Pleurothallidex, again reproduced by Unger®, have been proved to be the insertions of sunken hairs*; and Luerssen has recently shown that the large pores, visible with the naked eye in the leaf of Kaulfussia, . are typical stomata, of huge size and wide cavity, with collapsing guard-cells, and sur- rounded by 2-3 rings of subsidiary cells®. Sect. 10. Such outgrowths above the outer surface of the epidermis as do not belong to the cell wall alone are termed, in the plants with which we are concerned, Hlair-structures (Trichomes, appendages of the Epidermis\. These spring from cells of the epidermis, and are derived from them. We may distinguish as typical forms of hair-structures, Bladders (Papulz), Hairs (Pili, Sete), Scales (Squamz, Lepides, and Palez), and Shaggy hairs (Vill), Warts, and Prickles. ‘These forms are characterised by simple relations of shape, which mostly explain themselves according to their meaning borrowed from the language of ordinary life, and by equally simple differences of structure: Bladders ' Ann. Sci. Nat. 4 Sér. I. 156, p. 13, fig. 10. ? Milde, Monogr. generis Osmunde, p. 86. 5 Anat. und Physiol. p. 194. * Meyen, in Wiegmann’s Archiv, 1837, I. p. 419; Schleiden, ibid. 1838, I; Beitr. p. 8. * De Vriese et Harting, Monogr. des Marattiacées, p. 14, Taf. V. D.—Luerssen, Botan, Zeitg, 1873, No. 40, EPIDERMIS. AE are isodiametric, usually unicellular bodies; Hairs are sac- or thread-like bodies, unicellular, or consisting of a row of cells, simple or branched; Scales are flat membranous structures, always consisting of many cells, arranged in one or several layers; Shaggy hairs are thread-like bodies, consisting of two or many layers or rows of cells; Warts and Prickles are of similar constitution, but are not thread-like, but massive and hard, the warts are blunt, the prickles pointed. Intermediate forms and combinations of these types are common, and may of course be easily named after them. -On each hair-structure may be distinguished the body and the foot. The former is the part which protrudes outwards above the epidermal surface. The foot is the part which lies within this ; it is rarely similar in form to the epidermal cells, especially often it exceeds them in height, as it not uncommonly extends inwards far beyond the inner surface of the epidermis, into the sub-epidermal tissue. The epidermal cells, which surround the foot, may resemble those not bordering on a hair; very often they are quite different from these, and may then be termed subsidiary cells of the hair. Of the various forms of these, that of an annular or rosette-like girdle of subsidiary cells surrounding the foot of the hair recurs especially often (Fig. 21 &). Around the foot of many hairs, or below it, the subepidermal tissue, covered by the epidermis, bulges outwards, so that the foot is borne by an emergence of that . tissue. ‘This may be limited to a slight excrescence, upon which, as its ‘ bulbus,’ the hair is seated, or to a small, stalk-like outgrowth, which in multiseriate shaggy hairs is with difficulty distinguished from the hair itself; but, on the other hand, it may attain considerable dimensions, as in the prickles of Dipsacus’, and species of Solanum, &c., which bear’a hair on their apex, or the fringed scales of Begonia manicata”, The converse condition of the origin of a hair, in a more or less deeply hollowed depression of the surface, is not less common. Small hairs do not always overtop the edge of the depression in which they stand. They fill it completely, or only partially, as those on the leaves of the Pleurothallideze (Pleurothallis, Stelis, Physosiphon, Nephelaphyllum, Octomeria), which (comp. page 54) were wrongly described by Meyen as cavities of the Epidermis. The direction of the body of the hair, as regards the surface which bears it, varies extremely between that at right angles and that parallel to it. The hair-structures of one and the same surface are in the minority of cases all alike, if slight individual differences be disregarded. As examples may be named all known cases of Root-hairs, Leaf of Eleagnez, Bromeliaceze, Leaf and stem of Convolvulus Cneorum, &c. Much more commonly one and the same surface bears hairs of different properties, two to five sorts often occurring close to one another. Comp. Fig. 21. If we disregard the root-hairs, which with very few exceptions (Elodea, Lemna, Ophioglossez) are universally distributed, and reproductive organs, which are not to 1 Schleiden, Grundz, 3 Aufl. I. p. 281. ; 2 Compare Weiss, in Schr. d. zoolog. bot. Vereins. Wien, 1858. 56 CELLULAR TISSUE. be specially noticed here, some few families are distinguished by complete or almost complete absence of hair-structures, as the Equiseta, the Conifere, the Potamez, and Lemnaceze. They occur in the majority of genera and species, though certainly to a very variable extent. Different vegetative adaptation does not determine the presence or absence of hair-structures ; they occur under all states of adaptation, even in submerged species, as Callitriche, Nymphzea, and species of Ranunculus. On the other hand, their. number and development seems certainly to be influenced by the nature of: the environment, since observation shows that in allied species, and in individuals of the same species, the hairiness increases with the sun-light, dryness, and airiness of the spot. But there is no safe foundation for a definite assertion on this point. : As regards the distribution of single forms of hairs through families and genera, the case is similar to that of the forms of foliage leaves. On the one hand there is great uniformity of the majority of species and genera of one family, at least as regards one characteristic form of hair, so that one may speak, for instance, of the bristles of the Borragineze, the short (glandular) capitate hairs and scales of the Labiatz, the stellate hairs of the Cruciferze, the tufted hairs of the Malvacez, the multiseriate shaggy hairs of the Melastomez, the delicate branched hairs accompanying the capitate hairs of the genus Lavendula, the three characteristic forms of hair of most of the Hieracia, &c. On the other hand, in natural families (e.g. Composite, Labiate), and even genera (e.g. Solanum), the most various forms exclude one another ; or one characteristic definite hair-form recurs on corresponding parts in the most remote genera and families, as the stinging hairs on the leaf of Urticacee and Loaseze, the shield-like stellate hairs or scales on those of the Oleacezx, Eleagnex, and species of Solanum, Croton, Bromeliacez, and Ferns; the spindle-shaped, appressed hairs, with central attachment of the Malpighiacez and Cruciferz, &c. The development of hair-structures, both uni- and multicellular, begins, in all certainly investigated cases, from ove epidermal cell, as Lvzt/al cell. This cell protrudes beyond the outer surface of those surrounding it: the part within this surface develops into the foot, the protruded portion into the body of the hair. The growth which ensues is, according to the special case, acropetal, basipetal, or intercalary, as regards the hair itself (Rauter). It is obvious that in forms consisting of more than one cell, divisions accompany growth, and the successive division-walls appear in definite number and position for each case; further, that the definite form and articulation depend upon the successive divisions, and the growth of the cells after the division is complete. In 2- to 4-seriate shaggy hairs, scales, &c., in which the tows of cells are continuous into the foot, and are there represented by two or many cells side by side in the epidermis, e.g. Hieracium aurantiacum and its allies, division of the initial cell perpendicular to the epidermal surface begins almost simultaneously with, or very soon after the protrusion of the body outwards. The development of an emergence bearing a hair begins later than the origination of the hair itself by local growth of the subepidermal meristem, and of the epidermal cells surrounding the initial cell of the hair, : The origination of the hair-structures begins on stem and leaf at a very early age, on the former however, as a rule (but not always), not above the point of insertion of . EPIDERMIS, 57 the youngest leaf*. On the same surface, their formation begins at an earlier stage of development than that of the stomata. ‘The succession of appearance of the hair- structures follows the development of the part of the plant which bears them, but not so thoroughly that the hairs, in their successive appearance, arrange themselves exactly according to the direction of the advancing growth of the leaf which bears them. Not uncommonly new hairs grow out between those already formed.’ Most hairs on the parts named attain their full development with or before the complete unfolding of the bud. The thick covering of hairs, scales, and shaggy hairs in the bud-condition, is generally known. As the bud unfolds, the thickness of the covering decreases, partly as a result of the separation of the persistent hairs on the growing surface ; but partly through the disorganisation of hairs present in the bud during the unfolding, so as to leave behind only rudiments on the unfolded parts, or hardly that?. Even parts, which after unfolding are completely bare, may be hairy in the bud, e. g. the leaves of Ficus elastica®. (Comp. Fig. 18 A, p. 44.) We may accordingly distinguish between evanescent, transitory hairs which are peculiar to the bud, and persistent hairs. Among the latter we may again distinguish, as will be shown below (Sect. 13), between such as persist as Aung hairs, and others which are adherent but dry. In roots the case is different from that described. It is a universal tule, that here the hairs always appear on that part which is just ceasing to unfold, i.e, to extend. The above sentences will give the general points of view for the anatomical con- sideration of the differentiation of the hair-structures. Under this head are ranged structures rich in peculiarities, which have been the object of many works, and therefore have a huge literature to show. In older times more especially the forms, articulation, and functions of the hairs, which do not here concern us, were dealt with‘; in more recent, and the latest times, investigations on the history of development are considered of more importance®. I cite below for the time up to 1867 only the chief works, and + On this fact, which need not be further noticed here, compare Hofmeister, Die Lehre von der Pflanzenzelle, pp. 411, 545, and Rauter, Entwickl. einig. Trichomgebilde, p. 33. [Further, cf. Von. Hohnel, Botan. Zeitg. 1882, p. 145.] : ? Compare Hanstein, Botan. Zeitg. 1867, p. 697 ff. ® Schacht, Abhandl. d. Senckenbergischen Gesellsch. I. * Guettard, Memoires sur les glandes des plantes, &c. Eleven treatises in the Mémoires de l’'Acad. Royale des Sciences, Paris, 1745-1759; altogether 560 quarto pages. Compare A. Weiss, 4,¢.—F. y. P, Schrank, Von den Nebengefassen d. Pflanzen, Halle, 1794, 8vo., with 3 plates.—Rudolphi, Anatomie, p. 117 ff—P. de Candolle, Organographie végétale, I. p. 108.—B. Eble, Die Lehre von den Haaren in der gesammten organ. Natur. Bd. I, Wien, 1831 (only known to me from references). —Meyen, Secretionsorgane d. Pflanzen, Berl. 1837.—Physiologie, Bd. land II (1838-1839).—Bahrdt, De pilis plantarum, Diss. inaug.; Bonn, 1849.—A. Weiss, Die Pflanzenhaare (Abdr. aus Karsten’s Botan, Untersuchungen, Bd. 1); 306 pages, 13 plates, 8vo. * Hanstein, Ueber die Organe der Harz- und Schleimabsonderung in den Laubknospen. Botan. Zeitg. 1868.—J. Rauter, Zur Entwickelungsgeschichte einiger Trichomgebilde; Wien, 1871, with 9 plates (from Denkschr. d. Wiener Acad, Bd. XXXI).—J. Martinet, Organes de sécrétion des végé- taux; Ann. Sci. Nat. 5 série, tom. 14 (1872), pp. 91-232, pl. 8-21.—O. Uhlworm, Beitr. z. Entw. der Trichome, Botan. Zeitg. 1873.—Further, N. Kauffmann, Ueber die Natur der Stacheln; Bullet. Sci. Nat. de Moscou, tom. XXXII. p. 301 (1859); Warming, Sur la difference entre les trichomes et les epiblastémes d’un ordre plus élevé (Abdr. aus Kopenhagen. Videnskab. Meddelelser), Copen- hague, 1873.—C. Delbrouck, Ueber Stacheln und Dornen. Diss., Bonn, 1873.—S. Suckow, Ueber Pflanzenstacheln, etc. Diss., Breslau, 1873. (Further, Reinke, Anatomie d. an Laubblattern vor- kommenden Secretionsorgane, Pringsheim’s Jahrb. X. p. 119.] : 58 CELLULAR TISSUE, refer for the complete enumeration of them to the works quoted, especially those of Weiss and Martinet. . In face of the various facts and opinions it is our first business to determine what one understands by hair-structures or Trichomes. There are two opposed opinions on this point. The advocates of the one apply this name only to outgrowths belonging to and derived from the epidermis—in the sense indicated in Sect. 1—; others apply it to all characteristically formed outgrowths of the plant, to which the conceptions or traditional terms stem, leaf, root cannot be applied, whether these protuberances belong to the epidermis alone, or whether the subepidermal cells, and even the vascular system take part in their formation, e.g. the prickles of the Roses, of species of Smilax and Solanum, and of the Thorn-apple, &c. The foundation of the latter view seems to me to lie less in observable facts, than in the historical fact that outgrowths such as prickles and warts were formerly included among hair-structures, since it was thought that they belonged to the Epidermis}. If we deviate from this view, which is now proved to be incorrect for the majority of cases, the term trichome must also be restricted, and all outgrowths must be excluded from it, which include in themselves more than epidermis. Otherwise a quite unnecessary confusion would be brought into well-founded views and relations, since, if one includes among trichomes all outgrowths of the surface of stem or leaf, one must also include those of the /eaf margin, i.e. all leaf-teeth. If we adhere to the anatomical and developmental facts which are clearly before us, we easily obtain the definition here given of the idea of the hair-structure or Trichome as equivalent to an outgrowth of the epidermis, and the distinction of this from those outgrowths in which more than the epidermis takes part, for which the term Emergences proposed by Sachs (Textbook, 2nd Eng. Ed. p. 161) is suitable. ' The distinction between hairs and peculiarly formed epidermal cells may present difficulties in many single cases, e. g. in the genus Mesembryanthemum, where large cells scattered in the epidermis bulge outwards in M. crystallinum as huge bladders, while in other species they scarcely rise above the surface. But it is just the same whether one calls them hairs or not. The case, described by Uhlworm (J. c., Fig. 28-30), for the warts of Gunnera scabra, which are covered by a piece of epidermis consisting of cells elongated perpendicular to the surface, may be denoted, as above, or one can speak of a group of Jaterally-united prismatic unicellular hairs, or one may (with Uhlworm) term the whole piece of epidermis a multicellular trichome, which in that case forms an exception from the rule of the origin of each trichome from one initial cell. Starting, as is always necessary in defining types, from clearly characterised forms, the above established leading types of hair-structure may easily be separated according to their external development, and they are as a rule easily distinguished, without very exact investigation, by habit and consistency. Their distinction is therefore to be recommended for use in systematic Botany, which has as yet made use of these relations less than they might be employed. Intermediate forms are by no means absent. But these may easily be subordinated, or appended to the types. It is however often indifferent to which of the types a special case is appended, and this is defined as convenience may dictate. One may, for instance, call the flat horizontal appendages of the Eleagnex, or of Polypodium Lingua, stellately branched, multicellular hairs, just as well as stellate scales; or a capitate hair with a large compound head may just as well be termed a long-stalked scale or a dermal wart. Within the main limits, special forms are incredibly various as regards form, special articulation, and direction, &c. The detailed description of them is the subject of the most special systematic study, and their minute classification, though it might have a significance at the times of Guettard and Schrank, can only be idle play at the present time. Here, therefore, we may give only a few details and one or two drawings * Compare ¢.g. Schleiden, Grundz, 3 Aufl. I. p. 271; Unger, Anat, und Physiol. p. 188, EPIDERMIS, 59 ” (Fig. 21) as examples of the leading forms, and for further information refer to the literature above cited, or to any handful of plants. rs. 4 transverse section through a young leaf of Plectranthus fruticosus. c¢ short glandular capitate hairs (150). 2 Cajophora lateritia ; C Hieracium piliferum; leaf, longitudinal with irregular stellate terminal cell (90). dinal section. Explanation in the text (150), 2 Polypodium lingua; under-surface of leaf. Stellate hair, @, in surface view (90); 4 transverse section (150). FIG, 21.—Examples of forms of Hai @ conical multicellular hairs, 4 small capitate hair, transverse section through the carpel; explanation in the text (x50). section. @ thread-like, ¢ short capitate shag-hair; 2 multicellular hair, D Cheiranthus cheiri; under-surface of leaf, longitu 60 CELLULAR TISSUE, I. Hairs, elongated cells or rows of cells, simple or branched. Free ends not much widened, or tapering conically: Filiform and conical hairs; or widened into a head, Capitate hairs (Pili capitati). In the latter case the head is often articulated as a cell- surface, or cell-body; these are transitions to the leading forms II and III, and may be named according to convenience, 1. Filiform and conical hairs. (a) Unicellular and unbranched forms belonging to this group arise, by the arching outwards of the whole or part of the outer wall of one epidermal cell, so as to form a cylindrical or conical protuberance above the neighbouring surface. The whole hair isa single cell, of which a sac-like part of variable size protrudes as the body, the rest is embedded in the epidermis as the foot. To this type belong all root-hairs. They are as a rule partial protuberances of the outer wall of one epidermal cell; when freely developed they are bluntly cylindrical, but by application to the solid particles of soil they assume irregular forms and curvatures}, rarely they are branched (in Brassica Napus observed by Sachs, Textbook, znd Eng. Ed, p. 100), or they may arise in pairs from one epidermal cell. Only in Lycopodium? can special hair-cells be distinguished from the other epidermal cells on the root. From many of the original similar polyhedral cells, a part of the lower end is cut off by an oblique wall as a small cell, which divides further into 2-4 cells: each of these grows out into a hair: the hairs therefore in the mature root are arranged in groups between the elongated epidermal cells. On the foliage-leaf are to be found innumerable further examples. As a remarkable form may be mentioned the conical hairs of many Borraginez, Loasee (Fig. 21, B), Hydrophylleez (Wigandia), Urticex, many Cruciferz, Biscutella, Draba aizoides, Sinapis, Brassica spec.), also of Iatropha urens, and napzifolia. In the stronger forms of this category, whether they sting or not, the base of the conical hair is swollen, and encroaches on the surrounding tissues. It is usually borne on a more or less protuberant emergence, and is surrounded by a rosette of peculiarly formed subsidiary cells, To certain of these hairs (Loasa, Nettles, latropha spec.), which are characterised as a rule by a button-shaped rounding-off of the upper end, and by the nature of their walls and contents (Sect. 13), but by no further anatomical peculiarities of the hair itself or its surroundings, the name stinging hairs (Stimuli) has been given. Compare the figures of Meyen (Secretions-organe), Weiss, Martinet, and Rauter, and the more or less successful figures of the stinging-hairs of the nettle in most text-books. Forms of this nature are especially various in the Loasez (Loasa bryonizfolia, Cajophora lateritia), On the leaf and ¢he carpels of the latter (Meyen /.c. Tab. VIII, B in our Figure 21) are seated two sorts of conical hairs borne by slight emergences, and with their swollen bases surrounded by subsidiary cells, (r) long, smooth, blunt, stinging hairs (a), and (2) shorter ones, having the point oblique to the surface, with a thicker wall and numerous whorls of short points turned upwards (4); further (3) small thin hairs, with a circle of reflexed spicules at the blunt end, and many such laterally, these have a tapering base inserted in the epidermis (c): lastly (4) small 2—- to many-celled capitate hairs (d). The term unicellular branched hair may be applied to those described second in Cajophora (Fig. 21, 4, also c), inasmuch as the spicules or little hooks are short branches. Transitions from the unbranched to the branched form are to be found in the Cruciferz ; in Draba aizoides, D, hispanica, Boiss., side by side with the above-named simple conical hairs, occur others which, though otherwise of like character, are once branched at an acute angle. More richly branched hairs, with many modifications and complications are the prevailing form for the leaf of most Crucifere, though they are not Srehisieely present. The body of these unicellular branched hairs rises from the expanded foot. After a short distance, through which it remains undivided, it splits into 2-4 equal 1 Sachs, Exp. Physiol. p. 186. ? Nageli und Leitgeb, Bau und Wachsthum der Wurzeln, p. 124. EPIDERMIS. 61 diverging branches, which may themselves be repeatedly forked—often, as in Matthiola arborescens, with cymose unequal continuation of the successive forked branches. In the forms that are felt-like to the touch, as Farsetia incana, Matthiola arborescens, Alyssum petreum, Draba spec., the branches rise obliquely from the epidermis up- wards. In others they are parallel to the epidermis, and lying close to it they spread out like a flat star: stellate hairs, e.g. Capsella bursa pastoris, with 2-4 simple rays, Alyssum petrum, with 3-4 rays once or twice dichotomised. If the body of the hair divides close above the outer surface of the epidermis into two conical limbs, both of which are directed in one line parallel to the surface, the form is attained of a spindle lying parallel and close to the epidermis; this at its middle passes over into the foot, which is inserted in the epidermis. Such spindle-hairs, with their longer axis as a rule parallel and close to the part which bears them, are characteristic for Cheiranthus cheiri (Fig. 21, D) and Erysimum canescens; they are also to be found among 3-4 rayed stellate forms in Capsella, Erysimum cheiranthoides, &c. Similar forms occur in other families: unicellular, appressed, very regular stellate hairs, with sharply conical, short, undivided rays, e. g. on the leaves of Deutzia scabra, 3-6 rayed on the upper, usually 9-10 rayed on the under surface. In the Malpighiacez! there is a similar series of forms, though these are less various than those in the Crucifere: simple erect conical hairs, and forked, stellate, and spindle- shaped hairs. The erect branched hairs are simply two-forked, with equal or very unequal branches; many-rayed stellate hairs occur in the genus Thryallis; specially large and remarkable, but otherwise of fundamentally similar form to those of Cheiran- thus, are the unicellular appressed spindle-hairs in this family which are termed by de Candolle (Organogr. p. 103) Malpighiaceous-hairs. Another often-described case of the last-named form are the spindle-hairs (‘climbing hairs’) of Humulus lupulus, with their ends curved like a hook, and borne on an emergence. Further, Weiss (/.c. p. 528) mentions similar appressed spindle-hairs for ‘many species of Galega, Astragalus, Acer, Verbena, and Apocynum.’ (4) Most conical and filiform hairs are multicellular. In the simplest case they are two-celled, so that ove transverse wall separates a foot-cell from one cell of the body; in other cases they consist of more, and even numerous cells (Fig. 21, 4a). As regards the form, the same forms appear again, as in the unicellular hairs. One may even say that the same hair may be uni- or multicellular, i.e. that the formation of transverse walls is of minor importance; thus the long conical hairs on the leaf of Pelargonium zonale are sometimes unicellular, sometimes they have 1 or 2 transverse walls; in the latter case they are somewhat thicker-walled than in the former. Unbranched, multicellular, filiform, and conical hairs are the commonest form of all. Examples: leaf of Cucurbitacee, Solanum tuberosum, and its allies ; most Labiate (Stachys, Salvia, Thymus, Plectranthus, and others, but not all genera); many Composite (Helianthus, Cnicus, &c.); Trades- cantia spec.; the huge yellowish-brown hairs, up to 3 cm. in length, on the base of the leaf of several species of Cibotium, which appear in the shops as Pingawar Djambi, Pulu, &c.? Among the branched forms, in the first place, those described under the unicellular hairs recur as pluricellular. Hairs of the form of a T, that is, stalked spindle-hairs, with pluricellular stalk and unicellular cross-piece in the Anthemidee (Pyrethrum roseum, Tanacetum Meyerianum Sz., Artemisia absinthium, A. camphorata, according to Weiss, /.c.). Stellate hairs with unicellular, often rather irregular star, or even two stars, one above another, on a pluricellular stalk: Hieracium Pilosella and its allies, (Fig. 21, C4, comp, Weiss, Rauter, /. c.) Polypodium lingua has stalked, umbrella-shaped, very regular stellate hairs, in which the foot, the erect stem, the centre, and each ray of the star, are single special cells (Fig. 21, E), In the Hymenophyllex® are found pluricellular forked- and stellate-hairs. As 1 A, de Jussieu, Monographie des Malpighiacées, p. 96, pl. IL 2 Compare Fliickiger, Pharmakognosie des Pflanzenreichs, p. 142. - 3 Mettenius, Die Hymenophyllaceen, p. 65. , 62 CELLULAR TISSUE, examples of the latter may be named the sma// hairs of Verbascum }, the thin-stalked stars of Lavendula Stoechas, &c. Also the short stalked, two- to many-armed hairs of Utricularia? and Aldrovanda 3, in which each arm is a blunt cylindrical cell, belong partly to this category, partly to the tufted hairs to be named below. Hairs not forked, but monopodially branched, are (if we disregard cases like that described in Loasa) always pluricellular. Thus those with scattered, and sometimes repeatedly branched arms in Nicandra physaloides (Meyen, Weiss, /.c.), Lavendula elegans, Rosmarinus officinalis (leaf), on the inner surface of the bud-scales of Platanus (Hanstein, /.c.), those with whorls of branches on the leaves of Lavendula vera, species of Verbascum (e.g. V. phlomoides, the larger hairs). Also those demonstrated by Schleiden‘,’ which cover the leaf of Alternanthera spinosa, belong to this group. Not only is the lower part, which is attached to the foot, composed of 4-5 disk-shaped cells, one above another, but also the upper richly branched part is composed of as many cells as it bears whorls of main branches. The cells are separated from one another by transverse walls folded in deep waves, and each bulges out immediately above the transverse wall, which limits it below, into a whorl! of pointed branches, and here and there, on the rest of the lateral wall, into a single branch. The form often cultivated as Alternanthera amcena shows the same structure in its scattered hairs, but with only weak development of the branchlets. The bodies described by Weiss (/. c., Fig. 76) as branchlets on old hairs of Verbesina gigantea, I was unable to find either in this plant, or in a member of the same genus, and cannot make anything of them. ; 5 Under the name of tufted hairs, already often used, Weiss has judiciously separated a form allied to those under consideration from the forms usually included in the term ‘stellate hairs.’ It arises by the division of an initial cell of a hair by a number of successive walls perpendicular to the epidermal surface, and each of the cells thus produced grows like a simple conical hair, the body of which diverges from the others, while the basal parts remain firmly united. The history of their origin justifies the position of these bodies here, side by side, with the branched, pluricellular hairs, though, as far as the mature state is concerned, one might just as well speak of a tuft of diverging simple hairs. The tufted hairs are either seated in the epidermal layer, or are borne by a thin stalk-like emergence (e.g. felty species of Solanum, as S. marginatum, verbasci- folium, species of Correa), or on the apex of a multiseriate shag-hair (therefore a transitional form); this is the case in many Melastomexe (Tetrazygia eleagnoides®, discolor, angustifolia). Further examples are furnished by very many (all?) Malvacez ® Cistinex ; among the Labiate, Marrubium; species of Croton, e.g. Cr. tomentosus, J. Mill; species of Quercus, Platanus (comp. Weiss, Rauter, /.c.). The single rays of a tuft are usually unicellular, in Marrubium pluricellular. z. Capitate hairs; erect hairs of various forms, whose free end is swollen to form a round or disk-shaped head, the transverse section of which usually exceeds that of the stalk. The head may be part of a cell, or of a unicellular hair (Fig. 21, B, d, glandular hair of Aspidium molle) ; or it may be itself a single cell (Fig. 31-34), or be 2— to multi- cellular, with the cells arranged in the most various ways in one or several layers one above another. Capitate hairs are in the large majority of cases simple. Branched ones are only known where certain branch-endings of ramifying conical hairs bear a head (hairs of the bud of Platanus’). The stalk bearing the head may be reduced to a minimum, to the form of a very small disk—e. g. the glandular hairs of many Labiate (Pogostemon, Plectranthus, Molucella, &c.; Fig. 21, 4, b,c, 38). e 1 Weiss, /.¢, fig. 184. ? Meyen, /.c.; Benjamin, Botan. Zeitg. 1848, p. 58; Schacht, Beitrage, p. 28, ’ Caspary, Botan. Zeitg. 1859, p. 128, Taf. IV. * Grundz, I, 3 Aufl. p. 280. 5 Rudolphi, Anatomie, p, 113. ® Compare Sachs, 2nd Eng. Ed. pp. 43, ror. 7 Hanstein, /.¢. fig. 96. EPIDERMIS. 63 Capitate hairs occur on most leaf-forming plants, especially Dicotyledons and Ferns, as a rule in company with non-glandular hairs, It is true they are absent from many large groups; e.g. (all?) Graminez, Cyperacex, Palms, most Crucifere. To this category belong in the first place the great majority of the universally distributed glandular hairs : in our consideration of these we shall have to betake ourselves to single examples (Sect. 19). Meanwhile we need only remark here, that the glandular hairs are characterised by no special form, but rather by definite properties of the cell walls ; therefore the terms capitate and glandular hair are not equivalent. In the case of many capitate hairs, it is as yet uncertain whether they possess the characteristic properties of glandular hairs, since in the investigation of them no attention was paid to the fundamental point, and since their external development shows no difference from that of glandular hairs. Such cases may therefore remain unnoticed here, and only a few typical examples be cited of non-glandular capitate hairs. The family of the Cheno- podiacex furnishes the longest series of these: they are short hairs with a uni- or pluricellular cylindrical basal portion, which acts as stalk, and bears a relatively large bladder-like apical cell, usually of a round shape, but often irregular. They occur scattered on the leaf of many species of Chenopodium and Atriplex (e.g. Ch. album, Quinoa, Atriplex hortensis*), especially while these parts are young: later the bladder- like terminal cells are easily detached, and then together form a friable ‘meal. In other Chenopodiacez, whose leaves have a permanently white or gray surface, these hairs are so closely packed that their terminal cells (which dry up on mature parts) touch and overlap one another, forming a continuous layer over the epidermis, which does not fall off, e. g. Obione portulacoides, Atriplex rosea, A. nummularia. Hort. Non-glandular capitate hairs occur elsewhere, e.g. on the leaf of the Pelargoniums. The petiole of Pelargonium zonale shows side by side five sorts of hair; two are sharply conical (comp. above, p. 61), the one more delicate, without septa, the other stronger and with one septum; besides these there are three sorts of capitate hairs, (a) glandular with short, usually 2-3 celled stalk, and large unicellular, globular, glandular head?; (4) short-stalked, with inclined, obliquely obovate terminal cell, perhaps also glandular; and (c) elongated hairs bearing on a usually three-celled stalk a large oval or pear-shaped head-cell, not glandular (comp. Weiss, /.c., Fig. 367). Non-glandular capitate hairs with ashort 1-2 celled stalk, and a globular head composed of two cells standing perpendicularly side by side, are very common among the Labiate, together with glands and conical hairs. On the whole, they seem to occur very often as inconspicuous structures. II. Seales. Of the flat outgrowths of epidermis composed of one or few layers of cells, two forms may be distinguished, those which are scutiform, and those which are attached laterally. The former consist of a short stalk or foot, standing perpendicular to the epidermal surface, and a more or less round, umbrella-like disk, attached by its middle to the stalk. This is usually so short that the disk lies almost on the epidermis. It is either wholly a hair-structure, unicellular (e.g. Oleacez) or pluricellular; or is formed, at its insertion, from a small emergence ; or (Shepherdia and other Elzagnez) it is wholly an emergence, i.e. the round scale is seated at its centre directly upon a short emergence. The scale itself consists of radially arranged cells or rows of cells, which arise by corresponding divisions (i.e. arranged, as regards the hair, radially and perpendicularly), The number of these varies greatly, from four (Jasminum) to very many. In scales where the number is large the arrangement is often irregular, especially at the centre, by reason of tangential divisions, which appear in addition to the radial ones. At the periphery the cells usually grow out radially like hairs, so that delicate stellate shapes are produced. It is obvious from what has been said that the more simple forms of this category can hardly be distinguished from stellate hairs, such as those of Polypodium lingua (Fig. 21, 1 Meyen, Secretionsorgane, Taf. II. fig. 1; Weiss, 7c. p. 559, fig. 198. ? Hanstein, Zc. p. 745. 64 CELLULAR TISSUE, E), Platycerium, and from capitate hairs. The families Oleacez and Jasminez * yield an especially complete series of forms, from the 8-celled shield, produced by triple radial division of the initial cell (Syringa), or a 16-celled shield (Fraxinus), to the 3o-32-celled star (Olea Europea). Further examples of the forms of this category are the above- named Elzagnez, single species of Solanum (S. argenteum, Dun., and allied ‘ lepidota’), Croton (Cr. pseudo-china, nitens), Capparis Breynia, Andromeda calyculata, Myrica cerifera2. Further the leaves and stem of Callitriche and Hippuris*, and the long- stalked scales on the leaf of Pinguicula*. Large scutiform scales with pluriseriate, multicellular central part, and radial multicellular margin, cover the leaf of most Bromeliacez, e.g. Hechtia planifolia, stenopetala, Tillandsia usneoides*, Pholidophyllum zonatum, Billbergia clavata, Bromelia bracteata; the young leaves of many Palms, e.g. Klopstockia cerifera °, with scales several layers of cells thick in the middle. As regards their external development, there further belong to this category the circular, shield-shaped, glandular scales of many plants, consisting of few cells (e.g. Thymus, Salvia), or of many arranged in several series (Rhododendron ferrugineum, Humulus lupulus, Ribes nigrum, &c.). The peculiarities of their structure will be treated of later (Sect. 19). Of scales attached at one side the Ferns yield the richest and best known examples, in their so-called chaff-scales, or Palex. Among these occur various intermediate forms -between purely single-layered hairs, such as are many-layered at their insertion, uni- and multiseriate hairs, and shag-hairs. Their relations of size, form, and structure, so often made use of for descriptive purposes, may with a reference to the descriptive literature be here left untouched’. Those large branched scales on the stem of Hemitelia capensis, the similarity of which to leaves of the Hymenophyllums caused them to be described as a species of Hymenophyllum, are not to be included under epidermal structures, since they have vascular bundles, and an epidermis with stomata*. Uhlworm mentions in the case of Alsophila aspera thorn-emergences, which bear on their apex a large scale. In the Phanerogams examples of this category may be sought among those forms which form shag-hairs, inasmuch as these bodies are often developed chiefly in the direc- tion of one transverse diameter, i.e. into many-layered elongated scales. This is the case on the leaf-endings and margins of species of Papaver, in the Melastomez, as species of Lasiandra, Melastoma malabathricum °, &c. To this category belong also the dermal scales, borne on scale-like emergences, of Begonia manicata and its allies. As a special very simple form allied to stellate, tufted, or capitate hairs, may finally be mentioned the scales occurring in the axils of the leaves of Hippuris and Callitriche (Hegelmaier, /.c., Rauter, /.c.). These are borne on a short simple stalk-cell, and appear as a circular fan one layer of cells thick, which is composed of radially arranged elongated cells, or (Pseudo-callitriche) of rows of cells similarly arranged. III. On Shag-hairs (Zotten) (Fig. 21, C, a,c) little need here be added to what * Prillieux, De la structure des poils des Oléacées et des Jasminées, Ann. Sci. Nat. 4 Sér, V. p- I, pl. 2-3. * Rudolphi, /.c. p. 114, where generally are very numerous details, though there is occasionally a confusion with tufted hairs, * Hegelmaier, Monogr. d. Gatt. Callitriche, p. 11; Rauter, Zc. p. 6. * Schacht, Pflanzenzelle, Taf. VII. p. 16.—Lehrbuch, I. p. 280.—Grénland, Ann. Sci. Nat. 4 Sér. III. p. 297, Taf. X. 5 Compare Schacht, Lehrb. I. Taf. IV. pp. to, 11; Pflanzenzelle, Taf. VII. pp. 17, 18. ° How far the scaly or fibrous covering of the unfolding palm leaves consists of hair structures, or of effete drying masses of tissue, requires more complete investigation in special cases. Compare Mohl, Verm. Schr. p. 177, Structura palmarum, § 82. ” On their development, compare Hofmeister, Vergl. Unters. p. 85. * Compare Mettenius, Filices horti Lipsiensis, p. 111, ® Rudolphi, /.c. p. 115. EPIDERMIS, 65 has been already said. The shape of their body repeats that of all single hair-forms, from which it differs only by its articulation—being pluri- or multiseriate, It terminates in a head, or is simply conical, or resembles a tufted hair; the latter for instance in the shaggy hairs of the leaves of Leontodon hastilis and incanus, which fork at their ends into 2-5 stiff conical hairs, in the shaggy hairs of the above-cited'species of Solanum, Croton, and Correa (p. 62), and the Melastomez, where they end in a rich tuft of hairs; to these may be added Osbeckia canescens, and Medinilla farinosa. Its lateral margin is smooth, or toothed and zigzagged by the outgrowth of conically elongated cells; e. g. the conical shaggy hairs of the Hieracia, species of Papaver, and Mimosa; or it even bears tufts of hair (Correa speciosa). In shaggy hairs the foot is very often seated on an emergence. Compare e.g. the capitate glandular shaggy hairs of the leaves of Ribes (Hanstein, Rauter, Martinet, /.c.). The family of the Melastomacez has unusually numerous forms of shaggy hairs with the most various transitions to scales and tufted hairs. IV, In the simplest case Bladders differ from unicellular hairs only in form, and might be called by the same name, were it not too contradictory to the original meaning of the word to call a spherical body a hair. Such unicellular round bladders, with a broad foot penetrating far below the epidermis, or borne by an emergence, are known on the foliage-leaf of Mesembryanthemum crystallinum, Tetragonia expansa and echinata, and Oxalis carnosa’, On the whole leaf-surface of Rochea falcata? and longifolia cylindrical tough bladders rounded at the top arise between the small epidermal cells; they are provided, above their broad foot, with several blunt outgrowths, which almost touch the epidermal surface. They are all of the same height, and are in close juxta- position, so as to form an almost complete covering of the epidermis, In R. coccinea the margin alone is fringed by a single row of such bladders, which are rather elongated to the form of a short thick hair. _ The herbaceous stem, petiole, and under surface of the leaf of many Piperacee— Piper nigrum Hort, Enkea glaucescens, Artanthe elongata—are often but not always covered in the young but almost fully unfolded state by scattered spherical bladders, as large as a pin’s head, which shine like transparent pearls. These prove to be unicellular hair-structures, with a very small foot inserted in the epidermal surface, or projecting further inwards. On older parts they burst, and dry up to inconspicuous black-brown specks. Besides these there occur ‘in the same epidermis very numerous hair-cells, which only differ from the large bladders by their small foot-cell protruding above the outer surface as an inconspicuous papilla: it may therefore be said that many hairs re~ main inconspicuously small, while the minority swell to form the transparent bladders. Just the same appearance for the naked eye, with the same transitory nature, is seen in the round or oval bladders, as large as a millet seed, which Meyen ® discovered in Begonia plantanifolia, vitifolia, Cecropia palmata, peltata, Pourouma guianensis, Urtica macrostachys, in all cases distributed as above in Piper: further in Bauhinia anatomica, especially on the stem when several years old. These he named pearl-glands, Such structures are often observed also on Vitis, Ampelopsis* (A. quinquefolia, Veitschii), Cissus velutina, also on Pleroma macrantha (Melastomacex). These pearl-bladders (those of Pleroma have not been investigated) coincide with those mentioned for Piper in this point, that they are chiefly composed of very large bladder-like cells, which are thin- walled, and contain, besides radially-striated protoplasm, much watery fluid, and a number of brilliant colourless globules of resin or oil. In other points their structure differs. In the Begonias, according to Meyen, they are hair-structures which are allied to capitate shaggy hairs. The body of the pearl consists of about a dozen cells of the above * Meyen, Secretionsorg. Tab. VII. figs. 8-16, 38, 39.—Weiss, /.¢. ; ; ® K, Sprengel, Anleit. z. Kennt. d. Gew. 2 Aufl, I. p. 113, Taf. III.—A. Brongniart, Ann, Sci, Nat. 1 Sér. XXI. p. 453, Taf. 10. 3 Secretionsorgane, p. 45, Taf. VII. , * Hofmeister, Handb. Bd, I. p. 545. 66 CELLULAR TISSUE. character, arranged in two irregular rows, and is borne by a bi-seriate shaggy hair, as stalk. As in Piper there are also found small club-shaped shaggy hairs, from the swelling of which the pearls might have been derived. The pearls of the above-named Ampelidez are, on the other hand, emergences. They consist of several large cells of the character above-stated, and are covered by a protrusion of the epidermis consisting of numerous relatively small hyaline cells. On or near the summit of the body is a stoma, which is widely open, and on old specimens is further ‘extended by rupture at the corners of the slit. Young specimens are seated on the surface as blunt warts, with broad base. In cld specimens the upper part swells so much that the point of insertion appears as a relatively small stalk. The pearls of Urtica macrophylla, and, according to Meyen’s statement, of the other Urticacex, have in the main a like structure, with the difference that they are without the stoma. The pearls of Cecropia and Bauhinia are, according to Meyen, similarly composed; they are also without the stoma, and differ further in their tissue consisting throughout of small very numerous cells. V. Prickles and Warts. It was above stated that the massive outgrowths termed prickles and warts are mostly emergences, in the formation of which epidermis and subepidermal tissue conjointly take part. For the majority of these structures, as the prickles of species of Dipsacus, Rosa, Gunnera, Smilax, Solanum, and Ribes, the Cactacex, &c., this is thoroughly proved by late investigations of the history of their development (Rauter, Kauffmann, Warming, Delbrouck, Uhlworm). Delbrouck and Uhlworm, however, have both shown that exceptions to the predominant law exist, since the prickles of the investigated species of Rubus (R. czsius, idzeus, Hofmeisteri) and those of the petiole of Chamzrops humilis belong to the epidermis, and differ only in form and consistence from shaggy hairs. Further that the warts on the foliage leaf and ‘carpels of Bunias Erucago are at least chiefly derived from the epidermis. If once an anatomical distinction is adopted, it is necessary to separate the above outgrowths of epidermis from emergences of the same or similar form, however closely they may correspond to them—and to-many sorts of hair-formations of most simple structure— .°,. as regards their physiological or teleological significance. . > The often-described oval warts of Dictamnus’, which bear on their apex a short, septate hair, are connected with the above forms. They will be described more in detail below (p. 69). : 2. STRUCTURE OF THE ELEMENTS OF THE EPIDERMIS, (a) Protoplasm and Cell-Contents. Sect. 11. The wall of the Epidermal cells both in one-layered and many- layered epidermis encloses as a rule a delicate protoplasmic sac with distinct nucleus, and within this clear transparent cell-sap, which is either colourless, or tinted with dissolved pigments (Erythrophyll, &c.). It is to this condition (and the colourless ‘membrane) that most epidermal layers owe their great transparency. In the majority of cases chlorophyll and starch are absent from epidermal cells’, This is the case without exception in land plants where the tissue is very thick-walled, and surrounded by air ; often also in thin-walled cells occurring under similar conditions. But in not a few other land. plants more or less numerous chlorophyll grains, eventually with included starch, lie in the peripheral protoplasm. 1 Meyen’s (Secretionsorg.) ‘ Miitzenformige Driisen.’ ‘Compare Hofmeister, Pflanzenzelle, p. 259; Rauter, /.c. Taf. V, VI. * (CE Stohr, Bot, Ztg. 1879, p. §81.] EPIDERMIS, 67 ‘If one surveys the cases in which this occurs, it will be seen that the foliage of plants with delicate leaves, which live in shady places, is especially concerned, such as that of most Ferns, further Impatiens nolitangere, Melampyrum sylvaticum, Galeopsis -tetrahit, Ranunculus Ficaria, Epilobium roseum ; also Listera ovata, and Staphylea -pinnata * may perhaps be added. On the other hand, however, the same phenomenon -occurs also in inhabitants of sunny places, as Mercurialis annua, Lamium purpureum, ‘Caltha palustris, to which examples many others might easily be added. Epidermal cells of parts growing under water are, on the contrary, rich in chlorophyll grains and the bodies included in them, richer even than any other tissue of the species. Thus in the leaves of Ceratophyllum, Aldrovanda, Ranunculus aquatilis, Potamogeton, Hydrillee*, &c. In Elodea canadensis, and its allies, the chlorophyll-containing leaf consists in the main of only two layers, which originate, like scales, from the epidermis ‘of the stem. Brongniart® already showed that in species typically submerged, but which also occur as land plants, such as Ranunculus aquatilis, the submerged ‘epidermis is rich in chlorophyll, while that of the land form is without it, and that an ‘intermediate condition occurs on transition from one habit to the other. But the ‘rule just given is not general for all water plants. Both the amphibious species of Callitriches and C. autumnalis which only occurs submerged, have an epidermis without chlorophyll *. Sect. 12. In contrast with the epidermal cells, the guard-cells of the stomata are always very rich in protoplasm, chlorophyll, and the bodies included in the latter, especially starch grains; in colourless plants only the last-named bodies are present. The subsidiary cells of the stoma resemble the epidermal cells as tegards the properties in question. No peculiar phenomena, i.e. such as do not belong generally to the different cells of the plant, are known for the cell-sap of the epidermal and guard-cells, and the bodies which occur dissolved and suspended in it. ‘It is true they have as yet hardly ever been carefully investigated. This assertion is only confirmed by the casual statements made about oily drops suspended in cell- sap, and masses or drops containing tannin in the Cycadeze (Kraus), about tannin ‘generally in the Crassulaceze, Rosa, Ficus, Camellia, the Saxifragas*, &c.: also about more or less solitary crystals of Calcium oxalate in the leaves of Tradescantia discolor, Begonia manicata, argyrostigma, and Hakea saligna, octohedral crystals in ‘Asplenium Nidus, klinorhombic crystals which completely fill the small cavity in ‘scattered or grouped cells of the leaf of Ilex paraguayensis °. - . Thomas’ describes in the leaves of Pinus Pumilio, Pinaster, and austriaca, ' epidermal cells whose contents are dried up, and replaced by air in consequence ‘of rupture of the membrane. It may, however, be conjectured that this description refers to abnormal! conditions. 1 Sanio, Botan. Zeitg. 1864, p. 196. Compare also Kraus, in Pringsheim’s Jahrb. p. 314. 2 Caspary, Pringsheim’s Jahrb. I. p. 348.—Botan. Zeitg. 1859, p. 125. 3 Ann. Sci. Nat. 1 Sér. tom. XXI. (1830) pl. 17, figs. 3 and 6.—Further, Askenasy, Botan. Zeitg. 1870, Zc. * Hegelmaier, Monogr. p. 9. 5 Compare Sanio, /.c.; Kraus, /.c.; Wigand, Botan. Zeitg. 1862, p. 121; Engler, Botan. Zeitg. 1871, p. 888. , 6 Kraus, Z.c.—Meyen, Physiologie, I. p. 227-—Goldmann, Botan. Zeitg. 1848, p. 557. 7 Pringsheim’s Jahrb. p. 26. . F2 68 CELLULAR TISSUE. Sect. 13. The cells of Hair-structures are, while young, provided like other young cells with a well-developed protoplasmic body, and many while in this condition quickly attain a great size, so that they are specially suitable and easily obtained objects for the study of the protoplasmic body. Mature hairs behave in two different ways as regards their protoplasmic body and their contents. Those of the first category resemble, in short, the epidermal cells, having a permanent protoplasmic body, usually in the form of a delicate sac-like lining to the cell-wall; more rarely the protoplasm persists for a longer time in considerable quantity (stinging hairs of Urtica, Hairs of Cucurbita, &c.).. The cavities in the protoplasm are permanently filled with watery cell-sap (Sap-containing Hazrs). In hairs of the second category the protoplasm and cell-sap dry up when growth stops, and are replaced by air, These persist as air-containing hair-structures, The capitate hairs containing muctlage in Osmunda regalis are hitherto unique, and will be described below (Sect. 19). All root-hairs and a large number of the hair-structures which occur on foliage leaves contain cell-sap. They can be distinguished at once from those of the other category by their transparency. Their protoplasmic body and contents show the ‘same series of various modifications of special character as is the case in the epidermal cells. Most of them, e.g. all root hairs, all (?) stinging hairs, &c., are devoid of chlorophyll. Others have more or less abundant grains of chlorophyll and allied pig- ments. The correspondence with the epidermal cells extends also, as far as is known, to the substances mixed with the cell-sap (comp. Weiss, Z. c. 645). The contents of the often-described stinging hairs have special peculiarities, which are also said to occur in many hair-structures described as glandular hairs of various categories. We know of the erect stinging hairs of the Urticaceze, Loasez, and other plants named above (page 60), which resemble one another so remarkably in structure and form, that the brittle point (Sect. 22) breaks off when touched, and that a fluid issues through the hole thus made, which causes more or less slight inflammation when applied to the human skin, especially if it enters the small wounds caused by touching the hair itself. It is further known of this fluid that it has, like most cell-fluids, an acid reaction, not alkaline as stated formerly’. On the fact that by distilling the nettle plant with sulphuric acid formic acid is obtained, the conjecture has been founded that the latter substance causes the phenomena of stinging. But as a matter of fact nothing is known of the active substance, not even whether it is-to be sought for in the acid fluid, or in the protoplasm *. The apical cells of capitate hairs are often distinguished by very dense proto- plasmic contents, in which resinous substances may be shown to be present. Hanstein (Bot. Ztg. 2. ¢. p. 748) states that in the multicellular capitate hairs of Salvia all the cells may finally be united by the solution of their membranes into one fluid mass (containing Resin or Balsam) which is surrounded by the bladder-like cuticle. In the 1 P. de Candolle, Physiologie, iibers, v. Réper, I. p. 193. ? Von Gorup-Besanez, in Journ. f. pract. Chemie, XLVIII. Pp. Igt. * Compare the neat paper of Duval-Jotive (which however gives no new information), ‘Sur les stimules d’ortie, in the Bulletin Soc. Bot. France, XIV. p. 36. EPIDERMIS. 69 often described club- or egg-shaped warts of Dictamnus, which end in a short hair, there are found cell-contents of a character exceptional for epidermal structures. The cells finally fuse to form intercellular balsam-containing cavities’. These, as Rauter has carefully shown, are multicellular bodies derived from a single epidermal cell, consist- ing of a permanent peripheral layer of epidermal cells with scanty contents, running out into the terminal hair, and an inner multicellular mass. The cells of the latter contain, about the end of their growth, at first chlorophyll; then there appear drops of resin and ethereal oil in increasing quantity; these finally unite to large drops, which fill the cavity formed by the solution of the inner cell-walls (comp. Fig. 22). Beneath the epidermis, but derived in part from it, there arise in Dictamnus similar cavities containing ethereal oil. Martinet (Zc. page 176) describes in Cuphea lanceo- lata long, multiseriate, shaggy hairs con- sisting of elongated cells: in the broad base of each of these is enclosed a central round body consisting of many small isodiametric cells. This resembles the central tissue of the warts of Dictamnus in its position and its contents, which apparently include drops consisting of ethereal oil: but the solution of the cell-walls, which occurs in the latter case, was not observed in the former. Peculiarly formed groups of cells, characterised by their dense contents, which turn brown on drying or treating with alkalies, and which pro- trude little or not at all above the outer sur- face, and only slightly inwards, occur in num- bers in the epidermis of the tubular leaf of Sar- racenia*. They are globular, or flask-shaped, with the neck directed outwards, and consist of about 16 small cells derived apparently from the division of one epidermal cell. Their structure has been well described by Vogl from dried material. Their origin and sig- nification remains still to be investigated. 2 FIG. '22,—Dictarmnus fraxinella; oil-containing dermal Those cells and cell-groups belonging to watts: cut perpendicular to the surface. 4 youngest stage of development; 2 rather older; C (220) median section the epider nis, which have: contents of pecuhar Suche tee tecmes, peer Ramen Hon Sas nature, such as in the above examples, are often described as glands if they are also characterised by special form. This term will be discussed in Sect. 19. Air-containing dry hairs, scales, or shaggy hairs, form dry opaque coverings’ which appear of different colour according to the character of the membranes and 1 Meyen, Secretionsorg. Taf. I. pp. 28, 29.—Unger, Anat, und Physiol. p. 212,—Hofmeister, Handb. I. p. 259.—Rauter, Martinet, 7.c. : 2 A. Vogl, Phytohistolog, Beitrage. Sitzgsber. d. Wiener Acad, Bd. 50 (1864). 7O CELLULAR TISSUE, of the remnants ‘of the cell-contents, and present a lustre which varies according fo: their form and position and: the character of their surface. They preponderate in very hairy plarits. Thus the dense white felt on the foliage of many Labiate (Stachys, Teucrium, Salvia, &c.), Composite (e.g. Gnaphalium), the Verbascums, Banksias, Rubus idzeus, &c.;-the silvery white or brown peltate scales of the above- named Elzagnez, Bromeliacez, Croton, Solanez, Olea spec. ; the rustling ‘ Palea’ of the Ferns; the white crust consisting of dried capitate hairs in the above-quoted (p. 63) species ‘of Atriplex, and Obione, and other Chenopodiacez. (b) Structure of the walls of the Epidermal Elements. Sect. 14. The wall of the epidermal cells is, in very delicate parts, a thin cellulose membrane developed pretty equally all round. In rather more firm parts, in such as are termed herbaceous, and to a greater extent in very tough parts, such as stems and branches of smooth-barked ligneous plants, leathery and fleshy leaves, it is strongly thickened. In rare cases the thickening is almost equal all round, e.g. leaves of Ceratozamia mexicana‘, Pinus sylvestris, and its allies? (Figs. 11, 27; in this case the lumen almost disappears), or is much less on the outer surface than on the lateral and inner ones, as is the rule in the Bromeliacez (Fig. 12, p. 37)* Also in the epidermal cells containing mucilage, which will be described below, the inner wall is -of considerable thickness, often exceeding that of the outer wall. In an epidermis one layer thick and in the outer layer of a many-layered epidermis the outer wall is usually thicker than- the lateral or inner walls. In the above-named tough parts, such as leathery and fleshy leaves, old branches of Viscum, Ilex, Laurus, Menispermum canadense, Palm stems, &c., it is often thickened to such ‘an extent that it occupies the greater part of the whole volume of the cell. The thick outer wall is either sharply marked off from the thin lateral walls or graduates gently into them. The walls of the inner layers of a many-layered epidermis all resemble in the main the lateral and inner walls of the single-layered epidermis as regards. strength and structure, with the exception of isolated peculiar cases which must be mentioned as being extraordinary. ‘ The thickened walls have generally. the well-known structure of cell-membranes, stratification, striation, and pitting, but never fibrous thickening of the walls. The phenomena connected with special peculiarities of substance — cuticularisation, formation of Cystoliths—will be treated of later. There occurs sometimes on the. wavy lateral walls (e.g. under surface of the leaf of Helleborus fcetidus *) at the bottom of a depression a-local thickening in form of an excrescence resembling a fold or doubling of the membrane, which protrudes inward, at right angles to the surface. Pits of the usual form, corresponding on opposite sides, are very common on the lateral and inner walls. As a rule they do not occur on the thick outer ‘walls, e + Kraus, Cycadeenfiedern, 7. c. ? Thomas, /.c. p. 25.—Hildebrand, Botan. Zeitg. 1860, Taf. IV. 3 Von Mohl, Verm. Schriften, Taf. X. 33-—Schacht, Lehrb. I. Taf. IV. fig. 10. * Von Mohl, Verm. Schr. Tab. VIII. fig, 21; Vegetab. Zelle, p.14. Compare also Cohn, Nov. Act. Acad. Leopold. vol. XXII. pars «. : EPIDERMIS, a1 still they are present in a considerable number of exceptional cases. Thus on: the foliage leaves of Coffea, Viburnum Avabaki, Cocculus laurifolius, Cinnamomum: aromaticum, Camellia japonica’, and of Grasses 2, where some of them are arranged perpendicularly to the outer surface; but on the andulatinis corners they are directed obliquely outwards from the lumen of each cell, and facing the neighbouring cell, so that those of two neighbouring cells cross. They occur also in Abies ®, Cycas‘, Lycopodium pinifolium*, and Equisetum hiemale (comp. Fig. 24 B®). The walls of the elongated epidermal cells of the upper side of the leaf of Acropteris australis show a spiral striation, as the result of peculiar pitting (comp. Sect. 30). The free: surface of the outer walls is often quite smooth: but is not uncommonly uneven by reason “of small thickenings protruding outwards: short warts, e. g. in species: of Equisetum, leaves of Sparganium ramosum, Aloe verrucosa, Radula, Crassulaceze (comp. Fig. 20, p. 53), &c.: bands, which are relatively broad and blunt, e.g. leaf of Helleborus niger, foetidus *, Dianthus Caryophyllus, plumarius, or thin and sharp, as in very many leaves and petioles, e.g. Allium Cepa, Eucomis, Rumex Patientia ’, obtusifolius. The bands often run nearly straight and parallel, and are then usually longitudinal relatively to the whole body, rarely (Eucomis) they are transverse; not uncommonly they are wavy and branched (e.g. Helleborus, Pirus communis), and in the majority.of cases they are continuous from one cell to the next. The wall of the sfoma/al ce/ls* is usually, but not always, thinner on the average than that of the adjoining epidermal cells. It is in most, and one may say in regular cases, unequally thickened in such a way that a strongly thickened ridge runs along the entrance and exit of the slit (Fig. 23). These RBCS, protrude on the free sur- face as the above described ridges of en- ge wh trance and exit, which are sharp-edged and a concave towards the slit; rarely both are almost equally strong (Lilium candidum, Ficus elastica); usually the ridge of entrance is much stronger than the other, and in su- perficial stomata of tough leaves it often takes the form of a high and thick wall, e.g. Clivia nobilis, many Proteacez, Pholido- phyllum zonatum (Fig. 12, p.37), Epidendron faa 4 F1G. 23.—Hyacinthus orientalis ; 4ea/, transverse sec- ciliare, Octomeria, Sarcanthus rostratus, &c. tion; ¢—¢ epidermal cells; 5 entrance of the stoma, which is cut in median transverse section ; z the respi- The ridge of exit is often extremely small ratory cavity, between the parenchymatous cells J. (800). From Sachs’ Textbook. (leaf of Pholidophyllum, Dianthus caryophyl- lus, Lomatia longifolia, Sparganium ramosum), or is not present at all (comp. p. 35). The thickened ridges either protrude into the cavity of the cell as flattened swellings, or not at all. The remainder of the wall of the guard-cell, that is the convex side 1 Kraus, /.c. p. 318. 2 Von Mohl, Verm. Schr. Taf. IX. 3 Thomas, Hildebrand, /.c. * Von Mohl, Z.c. Taf. X. 5 Sanio, Linnea, 29, p. 169. § Von Mohl, Verm. Schr. Taf. IX. 6-8. ™ Von Mobl, Z.c. figs. 3-5. ® Compare the papers quoted above, sect. 5, especially Von Mohl, Spaltofin. d. Proteaceen ; idem, Botan, Zeitg. 1856, Z.¢.; and the Inge series of good representations in Strasburger’s work, Pring- sheim’s Jahrb. V. 7% ' CELLULAR TISSUE, which is turned from the slit, the united ends, and the strip of the concave side which borders on the slit is much less thickened. The last-named strip is seen, in those cases where ridges of exit and entrance protrude far into the cavity, asa channel on the inner surface of the wall, which appears in the transverse section of the stoma like a broad pit. When the ridges of exit and entrance are very broad, the middle of the convex side assumes a similar appearance (e.g. Ficus elastica, Fig.18C). T he various modifications of this plan of structure, depending upon the absolute and relative strength of the thickening and its varying protrusion inwards and outwards, require no further description in detail; some of them are evident from the above figures. Further, as regards genuine exceptions, which occur especially in the Coniferee and Cycadez, we may refer to the special literature above-quoted (comp. p. 35). Still the structure of the walls of the stomata of Equisetum, which is anything but clearly explained in the writings of Duval-Jouve and Milde 2 must not be entirely passed over (Fig. 24). The guard-cells themselves are here in no way WY iN ah ee SoS FIG. 24.—Stem of Equisetum hiemale; stoma with the surrounding tissue (390). 4 view from the inner surface, The pair of guard-cells is surrounded by the superposed edye of the pair of subsidiary cells. 2 transverse section of the stem passing in a median direction through a stoma, which lies in a depression of the surface. The narrow entry of the slit is bordered by two flat guard-cells and the subsidiary cells which surround them; each of the latter has a curved pit, turned outwards, Further explanation in the text. The cells of the single-layered epidermis and of the hypodermal schlerenchyma below it have numerous pits. C silica residue of a fragment of epidermis with a stoina, after maceration in Schultze’s mixture and subsequent ignition; seen from outside, The curved figures are the outlines of the prominences of the outer surface. remarkable except in their form, which is flattened obliquely outwards, and in the slight difference in thickness between the ridges of exit and entrance and the other bands of the membrane. But the subsidiary cells, which completely surround them (comp. p- 43), have a thickening on the wall bordering on the guard-cell, in form of bands protruding into the cell-cavity. These diverge in a radiating manner from the slit. Hence the beautiful radiate-marking seen in surface view. The number, breadth, and frequent branching of the radial bands vary according to species. In Milde’s Equiseta phaneropora (at least E. limosum) each band runs over the whole radiate * Milde, Monographia Equisetorum, Nov. Act. Acad. Leopold. tom, XXIV. pars 11,—Duval- Jouve, Histoire nat. des Equisetim de France, Paris, 1865. ? Sanio, Linnea, Bd. 29, p. 389, Taf. I1I.—Strasburger, /.¢, EPIDERMIS, 93 surface, the slit is therefore surrounded by a series of radiate bands. In Milde’s E. cryptopora (at least E. hiemale) the striated wall is traversed about half way between the slit and the convex outer border by a narrow oblique furrow, which is almost parallel to the slit, and this divides two concentric series of radial bands, one of these next the slit, the other on the side opposite to it (Fig. 24). What has been said of the walls of the epidermal cells holds in the main for those of hair-structures. Those walls which separate the cells of multicellular hairs resemble on the whole the lateral and inner walls of the epidermis. Their details of structure are, if possible, more various than their modifications of form. Pro- jections of the outer surface, in form of ridges, warts, or even of those sharp prickles represented in Fig. 21 B, appear in hairs more commonly than in the epi- dermal cells. The stiffness of the hairs depends upon the thickening of the walls, which may proceed till the lumen is obliterated. A hard, rigid hair or shaggy hair, is called a bristle or seta. If it is also conical and sharp one can prick oneself with it, as with the horizontal hard bristles of Malpighia urens, or the rigid hairs of the Borragineze and Cucurbitaceze. In this property, so disagreeable to men, lies the ground for the often-asserted similarity of the hairs of the Malpighiace to the stinging hairs, and of puncturing bristles to prickles. Sect. 15. The cell-walls of the epidermis are cellulose membranes ; a number of other bodies are embedded in, or superposed on this: Cuticular-substance or culin, wax, resin, volatile oils, gum and bassorin, compounds of szlcon, and lime salts, bodies with whose presence remarkable peculiarities of structure are connected. Scr. 16. Of the relatively pure layers of cellulose of epidermal membranes it may be said that in the majority of cases, especially in herbaceous parts, they appear similar to that watery highly refractive modification of cellulose membranes which is characteristically developed in the Collenchyma, to be described later. The detailed investigations necessary for an exact statement on this point have not been made. Allied to these watery cellulose layers are the parts of the membrane of epidermal cells, which have been altered to vegetable mucilage and bassorin; Radlkofer+ has lately drawn attention to the frequent occurrence of these substances in foliage leaves. The thickened inner wall of these epidermal cells consists, espe- cially in their inner layers, in the mature condition, of a vegetable mucilage, which swells in water till its identity is lost, like the mucilage of Linseed, &c. These layers of mucilage are developed especially strongly on the leathery leaves of the Cape Diosmez (Diosma alba, Agathosma spec., leaves of Buku’), where they are found on the upper side of the leaf, which has no stomata, and in the parts of the under surface where stomata are absent. The cells in these parts are of a great height, their outer half has the usual structure of tough cuticularised epidermis. The whole of the inner half, which is often large, is filled with the stratified mass of mucilage, which is limited on the exterior by a level surface. This body swells on addition of water or glycerine to such an extent that it lifts the whole outer parts of * Monogr. d. Gattung Serjania, p- 100 (1875). 2 Compare Flickiger, Schweizerische Wochenschrift f, Pharmacie, Dec. 1873. 74 CELLULAR TISSUE. the epidermis far from the inner tissues of the leaf, and appears itself like’a speciah mucilaginous layer of tissue. . . The same phenomenon, but apparently always developed to a less extent, was found by Radlkofer in the foliage leaves of numerous Dicotyledons ; e.g. Sapindacez, species of Salix, Daphne, Quercus pedunculata, Betula alba, Erica carnea and Tetralix, speciés of Prunus, Genista, Cytisus spec., &c., of Ferns in Botrychium Lunaria. As shown by Radlkofer’s comparative review of the cases investigated, the phenomefon is by no means generally distributed, nor is it generally peculiar to definite. forms of leaf or systematic groups; it is absent, e.g. in Salix alba, amygdalina, Betula fruticosa, Prunus Padus, &c. In the Sapindacez investigated by Radlkofer it is often only single cells, or groups of cells, which show the phenomenon in question. Cutin occurs in form of the cufcle, and in the cusicular layers of the cellulose membrane’. : Cutin is a non-nitrogenous carbon compound, which is completely combustible ; it is dissolved, or destroyed by boiling solution of potash, and by Schultze’s mixture ; it can therefore be completely removed from the epidermis by these reagents. It is only slightly attacked by mineral acids, especially sulphuric acid, which destroy. cellulose. It remains unaltered in ammoniacal sub-oxide of copper after previous: treatment with acids; the same is the case with water, alcohol, and ether. It resists rotting far longer than cellulose. By these reactions. the means are supplied for. isolating the cuticle and the cuticular layers. Iodine preparations, with or without the assistance of sulphuric acid, colour the cuticle and cuticular layers yellow or brown. ‘Aniliné dyes are quickly taken up by them in large quantity, and deep coloration is the result. From a mixture of aniline red and violet, the latter is often (not always) taken up more abundantly *. The cuticle covers the whole outer surface of the epidermis, including the hairs, as a thin, always closely applied hyaline skin, It appears, secreted on the outer surface of the cellulose walls, on the embryo while it consists of only few cells, and covers it henceforward as well as the punctum vegelationis of the stem and all the members which appear on it, always following the growth of these by means of corresponding surface-growth, while its increase in thickness is infinitely less. This continues till the epidermis is eventually thrown off. In rare cases, when the original epidermis is destroyed early, and is then replaced by new elements (the leaves of the Aroidez mentioned on p. 29, and perhaps also of Palms), there appears also over the latter a new cuticle. It is wanting at the punctum vegelationis of the root, but appears behind it, where the differentiation of the epidermis begins. The cuticle is con- tinuous over the surface of the guard-cells, through the stoma, into the respiratory cavity, from the moment when the formation of the slit begins by separation of the two cellulose lamellz. It usually continues over the walls of the respiratory cavity as far as these are formed from the epidermal cells. It becomes gradually thinner as it proceeds inwards, till it ceases where the respiratory cavity is laterally bounded by subepidermal cells. It thus forms at each stoma an open tube which passes + (Cf. Von Hohnel, iiber die Cuticula.—Aef. Bot. Jahresbericht, 1878, I. p. 16.] ? Hanstein, Botan. Zeitg. 1868.. EPIDERMIS. 45 from the slit inwards. In the Cactez, it extends fromthe slit onwards over thé. whole wall of the spacious respiratory cavity, and sends tubular branches with open ends into the intercellular spaces of the neighbouring chlorophyll-containing paren- chyma*. It is wanting as a rule on the inner surface of the epidermis. It is rarely continued from the slits onwards over the whole inner surface of the epidermis, as far'as this borders on intercellular spaces, as a lamella, which is interrupted by the surfaces of insertion of the subepidermal cells (v. Mohl, Z.¢.). This is the case on both the stomata-bearing surfaces of the leaf of species of Armeria, especially A. plantaginea, on the under surface of the leaf of Betula alba, Dianthus caryophyllus,: Euphorbia Caput-Medusz, and the stomata-bearing bands of the leaf of Asphodelus. luteus. In Helleborus niger and viridis the inner cuticle extends from the stomata- bearing under surface, over the upper side of the leaf, which has no stomata. (On the occurrence of cuticle in the deeper-seated intercellular spaces, comp. Chap. VII ; on the peculiar phenomena in Restio diffusus, see Sect. 18.) In the well-established. cases, the cuticle cannot by the means of investigation. at present in use be separated either mechanically, or optically, into separate parts or segments corresponding to the neighbouring cells. By careful maceration with -potash or dilute acids, it may be separated as a continuous skin from large tracts of the underlying cell-membranes. It appears by the action of the above reagents to swell more strongly in the direction of the surface than those membranes. By boiling solution of potash, or Schultze’s mixture, it is transformed into a tough shiny mass, and then completely destroyed without leaving a cellulose residue. It is in. most cases very thin, especially on submerged parts and roots; on aerial parts, not excepting the punctum vegetationis, it is thicker ; only in few cases where it is specially strongly developed (leaf of.Cycas revoluta, Ilex aquifolium), a delicate. stratification can be recognised; as.a rule, there is no sign of this, Its thick- ness is usually equal all over one and the same surface; also on the ridges and warts of the surface so often mentioned the cuticle itself usually runs unthickened over the corresponding outgrowths of the wall (e.g. leaf of Eucomis, Orchis, Helleborus, &c. .Comp. also v. Mohl, Verm. Schr. Taf. IX. Figs. 7,8). Projecting thickenings belonging to the cuticle itself are much more rare; on the hairs of Mono- tropa Hypopitys? there is the most exquisite example of this; the outermost layer of the wall, which is covered with numerous elongated warts, here shows the pro- perties of the cuticle: it is completely dissolved in boiling potash, and leaves the cellulose membrane quite smooth. On some very thick epidermal layers, which form large quantities of wax (Acer striatum, Negundo, Sophora japonica), the cuticle follows the increase of thickness of the membranes only a short time, and then breaks up by irregular splits. : Where the epidermis is delicate the cuticle covers the relatively pure cellulose membrane of the epidermal cells. But where it is thicker, especially when long- lived, the part of the cellulose membrane bordering on the cuticle itself also contains cutin, and consists of layers of cellulose, each of which is permeated by cutin. 1 Von Mohl, Botan. Zeitg. 1845, p. 3.—Unger, Grundziige (1845), p. 25. ? Schacht, Lehrbuch, I. p. 140. 96° CELLULAR TISSUE. According as this is the case, the membrane shows the characteristic reactions of cuticle. Treatment with the reagents, which dissolve cutin, removes it successively: from the persistent cellulose membranes, which retain their original form and struc- ture, though necessarily with considerable loss of substance (comp. Fig. 2g). These layers containing cutin are called cusiculardsed, or cuticular layers". The cuticularisation may extend over all the elements of the epidermis, in- cluding the epidermal cells, the hair-structures, and the cells of the stomata also, On the latter the cuticular layers are it is true often thinner, in correspondence with the smaller size of the cells, but often, especially in the ridges of entry, they are strongly developed, and are continuous with those of the neighbouring cells, e.g. leaves of Clivia nobilis’, Dasylirion*, Epidendron ciliare and other tough-leaved Orchidez, Ficus elastica, &c. The cuticularisation is evenly continuous over the epidermal cells of one surface, and in the above-mentioned cases over. the stomatal cells also, so that one cell resembles another: thus the cuticular layers, in a simple case, appear in a transverse section as an even broad band surrounding the whole epidermis, The cells of the epidermis may be separated from the cuticle, which covers them, both optically and (by help of the above-named .reagents) mechanically. The latter fact is certain, though it is not always easily done. The form, relative thickness, and extension of the cuticle over the cell-walls connected with it is no less various and characteristic in special cases than the other relations of form and structure above described. The following typical forms may accordingly be distinguished :— 1. The cuticular layers form in the great majority of cases a covering on the outer side of the epidermal cells, which is sharply marked off internally from the non-cuticularised membrane. This may be— (a) A layer of almost universally equal thickness, which follows the surface, and does not attain a thickness equal to that of the outer wall; e.g. leaf of Dianthus plumarius, Caryophyllus, Helleborus foetidus, Vanilla, Galanthus nivalis *, &c. Or (4) a thick layer, which follows the outer surface, and projects inwards in the shape of a ridge, of a conical form, in the middle of each lateral wall, and where. several cells join. The projections are usually sharply wedge-shaped towards the © inside, and do not reach as far as the inner wall (Fig. 25). Or they reach as far as the latter, and are continued into the layer (‘ intercellular substance ’) which marks the limit towards the subepidermal layer, and which is also cuticularised; e. g. branches of Jasminum officinale, Ephedra distachya, leaf of Phormium tenax 5, Ilex (Fig. 26), and Pinus (Fig. 27). The non-cuticularised layer (coloured blue with Schultze’s solution), which in all these cases surrounds the cell-cavity, is either relatively thick, and consists of many layers, as in the leaves of Pinus, Ilex (leaf-nerves), many species of Aloe, 1 ‘Von Mohl, Botan. Zeitg. 1847, p. 502. ? Von Mohl, Botan. Zeitg. 1856, Lc. 3 Schacht, Lehrbuch, I. Taf. IV. fig, 9. * Von Mohl, Verm. Schriften, p. 260 ff_—Wigand, Intercellularsubstanz u. Cuticula (1850), fig. 96, &c.—Petunikow, Recherches sur la Cuticula, p. 191, figs. 1, 22 (Bulletin Soc. Imp. de Moscou, 1866). ® Von Mohl, Verm, Schriften, Taf. X. fig. 28, 27. EPIDERMIS, 99 ‘Agave americana, Epidendron ciliare, Dasylirion, Sanseviera zeylanica, and the phylloclades of Ruscus aculeatus ; or, in so far as it borders on the cuticular layers, it is a very thin layer, which in many cases can only be observed with certainty in very good preparations. This is most frequently the case where the epidermis is thick. Examples: leaf of Hakea ceratophylla and other species, Ilex aquifolium (surface of leaf), Hoya carnosa, Taxus baccata (under surface of leaf); one-year-old branches of Viscum album *, Taxus, Rosa canina, Kerria japonica, Ilex aquifolium, Jasminum officinale, Laurus nobilis, Sassafras, Acer striatum °, &c. (c) The whole outer wall is cuticularised, the rest of the wall not. The upper side of one-year-old leaves of Taxus baccata, one-year-old stems of Salix daphnoides, and, according to v. Mohl (Verm. Schr. Tab. IX. 13), epidermis of the stem of Kleinia neriifolia. 2. The cuticular layers and the non-cuticularised part of the wall are not sharply defined one from the other, but rather— (a) either the inmost lamella of each cell-wall is not cuticularised, while the outer layers show the cuticular reaction gradually stronger the further they are from the inmost layer; e.g. stem of Psilotum triquetrum, young stems of Selaginella ineequalifolia, Martensii, &c. ; (4) or the whole wall of the epidermal cells is cuticularised all round: petiole of Arbutus Unedo *, two-year-old branches of Nerium oleander®, leaf of Elymus arenarius °, stem of Klopstockia cerifera’, leaf of Pinus, Abies, Cunninghamia lanceolata, older stems of the above-named Selaginellas. Further, the brown-walled epidermis of the stems and petioles of very many Ferns belongs to this category. 3. The epidermis of the pinne of the leaf of Cycas revoluta may be cited as an exceptional case®. The pitted cellulose walls of the epidermis are covered externally by a thick cuticle, which is stratified, but not separable into cuticular layers. From the cuticle there run narrow limiting bands of cuticular substance between the lateral walls of the cells, to the subepidermal tissue. Where the cuticular layers border on non-cuticularised membrane the limiting surface is either smooth, e.g. in most epidermal layers of branches cited above under No. 1; or it is rendered uneven by numerous small processes, which _ penetrate the cellulose layer like little teeth. Very small processes of this sort are for instance to be found onthe branches of Taxus ®, the leaf of Hoya carnosa ; and larger sharp teeth on the leaves of many species of Aloe (Fig. 25), and the phylloclades of Ruscus aculeatus. Epidendron ciliare has numerous fine teeth both on the surface of the cuticular layers which cover the cellulose wall externally, also on the wedge-shaped ridges, which protrude into the lateral walls, and thirdly on ‘their sharp angular pegs, which protrude further into the lateral angles than the 1 Compare von Mohl, Verm. Schriften, Taf. IX. X. figs. 12, 14, 23, 26.—Vegetab. Zelle, fig. 40. ' —Schacht, Lehrbuch, I. Taf. III. figs. 16, 17, 23-25; IV. 9, &c. * Von Mohl, Botan. Zeitg. 1849, p. 593- 5 Botan. Zeitg. 1871, p. 596. £ Wigand, /.c. p. 78. 5 Petunikow, /.c. pp. 19, 20, fig. 21. 6 Wigand, /.c. p. Ic5. 7 Botan. Zeitg. 1871, p. §77- 8 Von Mohl, Verm. Schr. Z.c.—Schacht, Lehrb, Z.c.—Wigand, /. ¢. fig. 43. 8 Graf zu Solms-Laubach, Botan. Zeitg. 1871, p. 536. 78 CELLULAR TISSUE, ridges. On the epidermis of the branches of Prumnopitys elegans* the processég are very large, in the form of thick plates, which are blunt, often branched, of unequal size, and irregular curvature; so that in their sections separate pieces, cut off from these, lié isolated in the non-cuticularised membrane. On the leaves ‘of many Proteacee (Lomatia longifolia, Hakea ceratophylla, H. Baxteri*) the pro- FIG, 26.—(800) Ilex aquifolium; leaf, 4 transverse section through the midrib of the under side ; @, 4 cuti- cular layers, the inner (4) stains yellow with Schultze’s solution, and is continued into the limiting lamella, which reaches to the sub-epidermal layer. The outer (2) remains uncoloured in Schultze’s solution (only par- tially true cuticle?). ¢, ¢ the non-cuticularised portions of the membranes, From Sachs’ Textbook, compare Pp. 35.—B a few cells of the same epidermis, seen from the outer surface. FIG. 25.—(390) Transverse section through the leaf of Aloe verru- C. de Candolle, in Mémoires de la Soc. de Physique de Généve, XVI. p. 1 (1861). § Verm, Schr, p, 212. PARENCHYMA, 115 SECTION III. PARENCHYMA. Sect. 25. The term Parenchyma is here applied to all internal cellu/ar tissue, 7. e. that which is found within the epidermis or cork-layer. Though as a fact it is in the main identical with Sachs’ ground-tissue (p. 6), in conception it is not so. It has already been stated (p. 5) that in this limitation and classification of tissues the distinction is drawn between those elements which refazz thetr cell-nature, and such as have lost that character. Attention was also drawn to the difficulty in the way of a generally uniform distinction, partly from the incompleteness of our present know- ledge, partly from the undoubted occurrence of real intermediate forms between cellu- lar tissue and many distinct tissues, especially Sclerenchyma. These difficulties did not come prominently forward in Sections 1 and 2. Here, in the internal tissues, they appear frequently, and it is everywhere to be repeatedly pointed out that in the distinctions about to be drawn, definite types must be indicated, which recur univer- sally, but are never sharply distinguished from one another. As regards the distinction of cells from other tissue-elements resulting from the metamorphosis of cells, it may here be again called to mind that the former are distinguished from the latter by the permanent protoplasmic body, in which the nucleus also (always?) remains, or appears temporarily. With these parts, which are directly observable anatomically, the cells retain the faculty of active growth and of division: it is true that this faculty is often enough not manifested, but, in the processes of secondary thickening (comp. Chap. XV), and especially in the phenomena of formation of cork (Sect. 24) ‘brought about by wounding, it may be so generally observed, that it may serve as a very useful character. The chlorophyll-containing parenchyma of a foliage-leaf, for instance, after complete unfolding shows as a rule no further division: the smallest wound immediately induces it. In very thick-walled, sclerotic cells, the direct anatomical determination of protoplasm and a nucleus is difficult, and as a fact is often impossible. Nor is that of the power of division more practicable. In its stead another phenomenon is to be taken into account, namely the periodic appearance and disappearance of starch-grains in many elements, which judging from the nature of their walls may be doubtful. Putting out of account the sieve-tubes (Chap. V), and certain laticiferous tubes (Chap. VI), in which, at all events, peculiar conditions, which need not here be touched upon, are the rule, the formation of starch in all well-known cases is directly connected with an active protoplasmic body. In doubtful cases therefore it is to be regarded as a character which indicates the presence of such a body, so long as it is not proved that it can also occur in spaces without protoplasm, and surrounded by cell-membranes. Abundant starchy con- tents, and especially periodical changes in their amount, must therefore for the present be regarded as a criterion of the cell-quality. In Chap. XIV we shall again return to this subject. Respecting the s/ructure of the cells of the parenchyma, nothing general need be brought forward at present, which would not be included in the doctrine of the structure of the mature vegetable cell, and this we assume to be already known, 12 116 CELLULAR TISSUE. As was stated in Sect. 1, their form is extremely various, and we may here distinguish as the chief types the iso-diametric, or shor¢ forms, and the elongated— fibrous cells, fibrous-parenchyma (‘Prosenchyma’), The further distinction of forms, to which during a certain period much energy was devoted’, has at the present day hardly even an historical interest. However certain definite forms, which are characteristic for definite single cases, must be mentioned. As is the case in cellular tissue generally, the special structure may be taken into consideration, and the distinction may be drawn between ¢hzn-walled and thick- wailed parenchyma, according to the relative development of the membrane on the one hand, and of the protoplasm and cell-contents on the other; but this holds only in extreme cases. In the distinction of subdivisions, the manner of connection of the cells one with another is taken into consideration as one of the characteristic special relations of structure. Thin-walled parenchymatous cells are in most plants the organs of the process of assimilation, and the storehouses of its first products; besides having a relatively thin membrane they are therefore usually distinguished by their contents; w. assimilating chlorophyll, and the most widely-spread direct product of assimilation, starch. According to the preponderance of one or the other of these parts, we may speak shortly of chlorophyll parenchyma, starch parenchyma, in many other cases of o7/- containing parenchyma, &c. The parts of plants which contain chlorophyll, and in which reserve products are laid up, e.g. especially the leaf, cortex of stems, and Rhizomes, are the places where these cells occur in large masses. In contrast to those characterised by the parts of the protoplasm and contents appearing as above described, there are other thin-walled parenchymatous cells, in which, within a protoplasmic sac, which is usually very delicate and slightly de- veloped, all the solid constituents of definite form diminish till they disappear entirely before the cell sap; this sap fills. almost the whole of the cell, and is watery, or contains very thin mucilage. This may accordingly be termed sap-parenchyma, This is wide spread, and as ‘aqueous tissue’ has recently been thoroughly described by Pfitzer? in many thick long-lived foliage-leaves, in which it is situated beneath the epidermis (hypoderma), forming as it were layers strengthening the latter, as in the Pleurothallideze, Bromeliacez, Ilex, Nerium, &c.; or it appears as a middle layer of the leaf, and is surrounded by chlorophyll-parenchyma, as in many succulent plants, e.g. species of Aloe and Mesembryanthemum, and in the leathery leaves of species of Callistemon, Hakea, &c., which will be more fully described in Chap. IX. It occurs in specially large masses in parts without chlorophyll which are rich in inulin or sugar, such as tubers and roots of Composite, Campanulacez, Beta, &c. The cells in question are characterised by their contents, which are almost perfectly transparent and fluid, being sometimes watery, sometimes (species of Aloe) muci- laginous. Their chemical constituents are exactly known only in single cases, as in the above Composite and Beta, they cannot therefore at present be used in distinguishing them generally. * Hayne, in Flora, 1827, II. p. 601.—Meyen, Phytotomie, p. 63.—C. Morren, Bull. Acad. . Bruxelles, tom. V. No. 3.—Compare Mohl, Veget. Zelle, p. 15. * Pringsheim’s Jahrb. VIII. p. 16. PARENCHYMA, 117 The forms of thin-walled parenchymatous cells are in the main nearly iso-dia- metric; but there often occur also elongated-prismatic, spindle-shaped cells, and the like, examples of which, e.g. in the case of the vascular bundles, will be described later; to this category belong also those chlorophyll-containing cells arranged in many leaves perpendicular to the surface, forming the pallsade parenchyma, to be described in Chapter IX. . As above intimated, very great variety of shape is found among the iso-diametric forms. It is only in definite single cases, e.g. in hypodermal sap-parenchyma, that the cells are of such form that all of them are bounded by flat surfaces and sharp edges, and therefore are in uninterrupted connection with one another. As a rule the surface of the parenchymatous cells is more or less rounded, or bears irregular protuberances, or the protuberances themselves are drawn out into long arms: in this case they are mutually connected only by definite parts of their surface, which vary in size accord- ing to the special form. Between them intercellular spacés are left free. Masses of parenchyma in which the latter (which are then usually filled with air) are developed to a great extent are distinguished as /acunar parenchyma, or, comparing it with a bath-sponge, spongy-parenchyma. Compare Chapters VII and IX. The walls of the cells of this category are as a rule cellulose membranes, with ordinary simple pitting. The latter, following the general rule, usually occurs only on the parts of the surface in contact with that of other cells: in cases then where the cells show a decided partial rounding off, and only touch one another with narrowly limited parts of their surface, or only with the ends of protuberances, the pits lie on these spots, and not on the rest of the wall. As regards the surfaces of contact, the same may also occur with dissimilar tissue-elements. When similar cells touch one another by the ends of narrow protuberances, there is often only a single pit on each protuberance ; larger circumscribed surfaces of contact appear as pitted fields on the otherwise smooth wall (Fig. 46). This phe- nomenon, which was known long ago’, and which occurs especially often in round-celled chlorophyll-parenchyma of succulent plants, resembles that of the sieve-plates of the sieve-tubes (Chap. V); but it is incorrect to place it side by side with this?, since the characteristic structure of sieve-plates is wanting in the parenchymatous cells, though the pitted fields also in the parenchyma of the leaf of Cycads, specially of Encephalartos, are distinguished from the rest of the wall by brown coloration in Schultze’s som thccatyleson of Phavestus mul tiflorus, isolated by maceration; 77 solution, and deep red coloration in solution of Anilin 5 the parts of the wall where it bor- ders on intercellular spaces, ¢¢ the Fibrous partial thickenings of the walls are known pitted parts of the surface which bor-* 5 : : der on neighbouring cells; the thin- here and there, e.g. in the form of reticulate or spiral jest points of the pits are shaded dark ie za ns 4B Sachs’ Textbook. fibres, in the watery hypodermal parenchyma of the leaves ©" "7m Seens Bestbve of the Pleurothallideze, and in many roots of orchids ; as reticulate fibres in the 1 See e.g. Schleiden, Grundz, 3 Aufl. I. p. 245. ‘ ae 2 Areschoug, Botan. Zeitg. 1870, p. 305; and Acta Univ. Lund. tom. IV.—Bofézow, in Pring- sheim’s Jahrb. tom. VII. 3 Kraus, Cycadeenfiedern, /.¢. « 118 CELLULAR TISSUE. middle layer of the leaf of Sanseviera guineensis; as longitudinal fibres in the chlorophyll-parenchyma of the leaf of Cycas*. They appear in an exquisite form in the parenchyma of the primary transient cortex of the root of most Coniferze (with the exception of all the Abietineze); this tissue may be best placed in this category?, The cells of the concentric layers of parenchyma, which lie outside the endodermal sheath (§ 27), are in many forms all finely reticulated (Phyllocladus, Podocarpus sp.), or thickened with coarse nets and longitudinal fibres (Cupressus spec., Sequoja sempervirens) ; in Torreya nucifera this thickening is limited to the 2-3 outermost layers, and the innermost layer bordering on the endodermis. In most of the in- vestigated forms, as Taxus, Biota, Thuja, only the latter layer has a fibrous thicken- ing, and, as also in Torreya and Cupressus, each of its radial walls has in its middle one straight, thick, stratified, half-cylindrical, longitudinal fibre, which is continuous over the transverse walls into that of the opposite radial wall, and in all cases fits exactly on to a similar thickening of the neighbouring: cell. In Thuja occidentalis this fibre contains resin according to Reinke. The layer of cells thickened in this manner appears as a closed sheath, with the exception of Frenela rhomboidea, where, according to Strasburger, it shows a break opposite both ends of the row of vessels (Chapter VIII). As a special case, to a certain extent worthy of mention, the tabular-polyhedral chlorophyll-cells peculiar to the leaves of species of Cedrus and Pinus® and many Gramina* may be further cited: these have narrowly infolded bands of wall, and from them broad ridge-like thickenings of the wall protrude inwards. Compare ° above, pp. 35, 78; figs. 11, 27. Luerssen® has recently proved that partial thickenings of the walls protruding’ on the outer surface are a characteristic phenomenon ‘for the parenchyma of many ferns. They occur in the chlorophyll-parenchyma of the leaf of the investigated Marattiaceze, and in the parenchyma of the petiole of the same plants, as well as of numerous investigated Cyatheaceze, Polypodiaceze, and of Todea barbara. They also occur in the stems which have been investigated with reference to this point, e.g. in Ophioglossum vulgatum, species of Polypodium, and Pteris (Luerssen), Aspid. filix mas, Onoclea struthiopteris, Cyathea arborea, Imrayana, Alsophylla microphylla; in Marattiaceze, e.g. M. Kaulfussii, also in the cortex of the root. In most ferns they appear to be wanting in the chlorophyll-parenchyma of the leaf. The pro- trusions of the outer surface occur obviously only on those parts of the wall which border on intercellular spaces, and, as a matter of fact, only on air-containing in- tercellular spaces. In comparison with the thickness of the rest of the cell-wall they are always thin, and when slightly developed they appear in the form of small knots, when better perfected as fine filiform rods, rarely thickened like clubs at their ends ; the longer ones are not uncommonly branched. In relatively few cases they occur * Compare Hofmeister, Pflanzenzelle, p. 168. * Van Tieghem, Ann. Sci. Nat. 5 sér. XIII. p. 187.—Strasburger, Coniferen, p. 346.—Reinke, Morpholog. Abhandl. p. 35. ~* Meyen, Physiologie, I. Taf. VI. 17.--Hartig, Forstl. Culturpfl. Taf, 18.—Thomas, in Pring- sheim’s Jahrb. IV. p. 40.—Compare also Hofmeister, Pflanzenzelle, p- 169. * Karelstschikoff, Bullet. Soc. Imp. de Moscou, 1868, No. 1. * Botan, Zeitg. 1873, p. 641, Taf. VI.—Sitzungsbr. d, naturf. Ges. zu Leipzig, 1873, No. 7. . PARENCHYMA, COLLENCHYMA. 119g singly, e.g. rhizome of Ophioglossum, petiole of Dicksonia antarctica. Usually they are numerous and closely aggregated. The elongated rods springing from the dif- ferent sides of the intercellular space in all directions are irregularly intertwined between one another, so as to form a delicate framework with air in its inter-spaces. The single rods sometimes end free, sometimes they are connected by their branches, or go from one side of the intercellular space to the opposite, and also adhere to the latter. As regards their material, the above outgrowths of the wall are equiva- lent or similar to ‘slightly cuticularised membranes.’ Cellulose colourings cannot be observed in them, but rather they and the outermost layer of membrane which connects them both behave under reagents like the limiting lamella on the surfaces of contact of the contiguous cells; they turn yellow or brown with Schultze’s solution, or with iodine and sulphuric acid, and are destroyed by boiling with solution of potash. It remains for further investigation to determine how far they may ac- cordingly be styled parts of an inner cuticle lining the air-passages. Sect. 26. A-definite special form of /hich-walled parenchyma is distinguished by the name Collenchyma'. It forms thick bands beneath or near to the epidermis, especially in stems, petioles, and nerves of the leaf of herbaceous Dicotyledons (e. g. species of Rheum, Rumex, Beta, and Chenopodium, Aegopodium, herbaceous shoots of Sambucus, Labiatze, Solanacez, Begonias, petioles of Nymphzea, etc. ”), and in the petioles of the Marattias*. In their typical development it is distinguished by the form and structure of the walls of its cells, which are capable of division and contain chlorophyll. The cells are in unbroken connection with one another; only in exceptional cases (stem of Silphium conjunctum and its allies) are the layers traversed longitudinally by intercellular canals. In form the cells are elongated many-sided prisms, with horizontal or obliquely pointed ends: when they are isolated it is usually. plain that they are derived from elon- gated mother-cells, with sharply pointed ends, which are divided by permanently thin transverse walls, or, as the case may be plainly stated, they are chambered‘. The walls are thin at the ends, and along the whole of the middle of the lateral sur- faces of cells, which face similar cells; but along the angles they are provided with a stronger thicken- ing, which protrudes into the cavity of the cell so as ‘ 2 7 FIG. 47.—Epidermis, ¢,and collenchyma. ¢é, of the as to round it off, or it may even project further, petiole ofa Begonia. The epidermal cells are evenly i e é "3 , thickened on the outer wall, where they abut on the while towards the thin middle band of the wall lt collenchyma ; they are thickened like it at the longi- ° . tudinal angles wherever three cells meet ; c/Z chloro- 1S sloped off, or sharply truncated (Fig. 47). In phyll grains; » shin: walled parenchymatous cells ! . ° (550). From Sachs’ Textbook. the stem of the above-named species of Silphium the thickening extends also over the faces opposite the intercellular spaces. The 1 (Cf. Giltay, Botan. Zeitg. 1881, p. 153; also ‘Het Collenchym,’ Utrecht —Ambronn, Pring- sheim’s Jahrb. XII p. 473.—Van Wisselingh, a la conaissance d. Collenchyme, Ref. Bot. Centralblatt, 1882, Bd. XII. p. 120.] 2 Compare Mohl, Vegetab. Zelle, p. 20; Botan, Zeitg. 1844, p. 308.— Unger, Anat. und Physiol. p. 148.—Sachs, Lehrb. p. 24. : 8 Russow, Vergl. Unters. p. 106. * Compare Kraus, Cycadeenfiedern, /.¢. p. 310 (6). 120 CELLULAR TISSUE. thickened parts of the walls have no pits, and are delicately stratified with very fine limiting layers (‘Intercellular-substance’): they swell largely with water, but without becoming gelatinous; when water is removed they contract greatly in all directions (measurements are wanting). In the soaked condition they show, in transmitted light, a characteristic bluish white lustre. With Schultze’s solution they turn light blue’; after slight warming with potash they immediately turn deep blue with solution of iodine in potassium iodide (this is the case in Sambucus, Rumex, Lamium album, Cactacez, Nymphzea). In the same places which are occupied in many plants by cells thus remarkably characterised by the above described properties, there are found in many others layers of cells which differ more or less from these in their form and in the structure of their elements. For instance, the cells of the Collenchyma of the stems of Cacti already mentioned? differ from those described as typical, in their slight elongation, and in the fact that the walls are thickened strongly and uniformly all round, and coarsely pitted. Other single forms approach nearer to the thin-walled or sclerotic forms of parenchyma, without its being possible to carry through any sharp distinction. It is then to a great extent a matter of taste how far one will extend the term Collenchyma, It is now used generally for the form of tissue here described as typical, though it was originally proposed by Link* for the pollen mother-cells with their gelatinous membranes, and was then transferred by Schleiden, at first half in joke, to. the above- mentioned collenchymatous cells of the Cactacez. From the collenchyma must be distinguished the thick-walled forms of paren- chyma, the membranes of which are more or less lignified, and have thus become hard and sclerotic. As the most typical representatives of tissue of this sort may be brought forward the thick-walled cells of the secondary wood of Dicotyledonous trees, which lay up starch periodically, and often prove themselves capable of divi- sion in the case of wounds, or rather healing scars. This will be entered into in more detail in Chapter XIV. In other places than that just mentioned sclerotic cells are to be found widely spread: together with collenchyma and sclerenchyma they form the strengthening apparatus of those parts, and they are connected with both of these tissues by the most various transitional forms. No general specific peculiarities of this tissue can be mentioned in addition to what has been already said; remarkable examples will therefore for the. most. part be mentioned in the chapters which deal with the distribution of tissues. Here we may briefly notice only one series as being specially instructive, and as presenting difficulties in a sharp classification of the tissues, viz., that of the sclerotic cells in the Ferns, In the large majority of these plants there occur in stem, roots, and leaves thick-walled elements, sometimes isolated, but usually in close and often in uninterrupted connection with one another, and combined to form uniseriate or multiseriate layers or bundles; these either lie near the epidermis, or accompany or ensheath the vascular bundles. In the petiole of the Marattiacez they have the properties of collenchyma, as was above stated; also many bundle-sheaths, to be cited later, are directly connected with this tissue as regards their structure. But in the large majority of cases (compare Fig. 48) the walls, which 1 Schacht, Lehrb. p. 195. * Compare Unger, Grundziige, p. 25; Schleiden, Anatomie d, Cacteen, p- 14. 5° Grundlehren d. Krauterkunde, II. p. 199. PARENCHYMA, SCLEROTIC CELLS. ENDODERMIS. 121 are equally thickened all round, or less thickened on one side, are highly ‘ lignified’: in some few cases they are quite or almost colourless (e.g. stem of Lycopodium), usually they are coloured a dark brown. On the chemical property of the characteristic brown substance nothing certain isknown. The sclerotic-tissue elements are generally of elongated prismatic form, either with slightly inclined or sharp-pointed ends, in the latter case they are fibrous cells or fibres. According to the character of their contents, they must, in compliance -with the fundamental ideas above laid down, be for the most .part assigned to the category of cellular tissue, since most of the elements, even those with yery thick walls belonging to the dark brown layers and strands in the ferns, are densely filled with starch grains, which (as was observed in rhizomes of Osmunda regalis) gradually disappear as their age increases. It was not possible by any means to prove that these cells are capable of division. On the other hand, there occur __,,*!% ##—Two sclerotic brown cells from the hypodermal layer of the rhizome of side by side with these sclerotic cells, and often con- Pers aduilina, isolated by chlorate of Potash and nitric acid. 4 more strongly ‘ . re) thickened on one side and with b hed nected with them by quite gradual transitions, elements pits lego) B less thick-walled the aotical thickened till the cell cavity almost disappears, and e420 0f the wall and the pitted wall at showing only the last traces of cell-contents. These, regarded independently, should be accounted as specific sclerenchymatous fibres ; thus, e. g. in the brown sclerenchymatous sheath of the stem of Marsilia salvatrix. Sect. 27. The name Lxdodermis, proposed by Oudemans? for a special case, here denotes generally those peculiar limiting layers to which Caspary” has given the name profective-sheath (Schutzscheide)*. They belong to the category of cellular tissue by reason of the nature of their contents, and their power of independent growth and division, which is so often to be observed, e. g. in roots of Dicotyledons. The endodermis is a sheath consisting in all cases of one single layer of cells. It should also be observed here, that it lies as a rule at the limit between masses of parenchyma and other systems of tissue, especially vascular bundles, and is then to be recognised both by its development and its mature properties, as the layer of the parenchymatous mass bordering on the unlike part. In roots with an axile vascular cylinder the latter is always enclosed by it. The same is the case in stems with an axile vascular cylinder, as Hippuris, Callitriche, Ceratophyllum, Utricularia, Elodea, species of Potamogeton, Corallorrhiza, &c. (compare Chapter VIII), or with a closely compressed axile system of bundles (species of Potamogeton, Hydrocotyle vulgaris, &c.): also in stems of Phanerogams with a strongly developed cylinder containing the vascular bundles, this is marked off from the surrounding mass of parenchyma by a layer of endodermis, e. g. Tagetes patula, and other Composite *, Cobsea scandens, 1 Ueber den Sitz der Epidermis bei den Luftwurzeln der Orchideen, Abhandl. d. Acad. Am- sterdam. Math. phys. Klasse 1X (1861). ? Pringsheim’s Jahrb. I. p. 441; ibid. IV. p. ror. 8 [Cf. also Schwendener, Die Schutzscheide u. ihre Verstarkungen, Abhandl. d. Konigl. Akad. d. Wiss, zu Berlin, 1882.] 4 Van Tieghem, Ann. Sci. Nat. tom. XVI. p, 113. 122 CELLULAR TISSUE. Primulaceze, as Primula sinensis’, Lobelia syphilitica, Rhizomes of Scitaminez, Cyperaceze (e.g. Carex hirta), Acorus gramineus. The same occurs in certain Equiseta. But on the other hand not the whole body of vascular bundles, but each single vascular bundle is in many cases sheathed round by an endodermis. This is the case both in the stem and leaf of almost all ferns and many species of Equisetum, and also in the petioles and leaves (Adoxa moschatellina, Menyanthes trifoliata, species of Primula), and in many stems of Phanerogamic plants, as Nuphar, Brasenia peltata, Hydrocleis Humboldtii, Primula auricula, Menyanthes. Rarely an endodermis occurs in other places than those named: thus in the parenchyma of the stem of many Equiseta, and in many aerial roots, especially of epiphytic orchids, the parenchymatous cortex is marked off both from the vascular bundle and from the tracheal sheath by an endodermis. The relations in the species of Equisetum may here be described according to Pfitzer?, as being specially instructive for the arrangement of the endodermis, which is variable even in closely allied plants. In the parenchymatous ground-mass of the internode there is a ring of vascular bundles equal in number to the angles of the stem (comp. Chap. VIII). In the foliage-stems of E. limosum and E. littorale an endodermal layer surrounds each single bundle. In E. arvense, Telmateja, silvaticum, pratense, palustre (comp. below, Chap. VIII), and scirpoides, this sheath is wanting round the single bundle, but sur- rounds the whole ring externally, curving inwards between two bundles. Besides this outer general sheath, there occurs in E. hiemale, trachyodon, ramosissimum, and varie- gatum a similar inner one, i.e. one bordering the whole inner side of the ring of bundles. In the rhizomes the same phenomena occur on the whole as in the foliage-stem; but in the same species, as more minutely described by Pfitzer, the rhizome and foliage-stem may be similar or dissimilar. Finally, at the points of transition between rhizome and foliage-stems, Pfitzer often found in E. hiemale small strings of parenchyma 1-3 cells thick, as seen in transverse section, which lay between two vascular bundles, and were surrounded by an endodermal layer, the latter either arising as a protrusion from the general sheath, or having no connection with it. The cells of the endodermis (see Figs. 49 and 50) are nearly of the four-sided prismatic form, very often flattened in the direction of the tangent of the part enclosed by them, more or less elongated, with horizontal or oblique ends, and connected uninterruptedly with one another along their radial lateral faces. Their membrane is always delicate when differentiation of tissues begins, and often throughout life it is smooth externally and internally, rarely it is delicately pitted ; but the radial walls are characterised by a fine and usually irregular wavy transverse folding, which is continued over the ends from one radial wall to the other. Further, the undulation extends, according to the special case, either over the whole surface, or only over a band-like longitudinal strip of it. The wall of the cells is further characterised by suberisation, which appears early, i.e. with, the first differentiation of tissue: this always affects the undulated part of the wall, and may also extend, according to the special case, in varying degree over some or all of the other walls. This is the case in the majority of Ferns: a good example of the localisation of the suberisation on the undu- lated bands in the middle of the radial (cellulose) walls is supplied by the root * Von Kamienski, Vergleichende Anatomie der Primeln, /.c. * Ueber d, Schutzscheide der deutschen Equiseten, Pringsheim’s Jahrb, VI, PARENCHYMA, ENDODERMIS. 123 of Botrychium Lunaria. In the root of Ranunculus Ficaria Caspary found most of the cells suberised, more particularly at least on the undulated walls; and, on the other hand, single cells, with no exactly definable arrangement, equally suberised all round. The walls in question may be termed suberised on this ground, viz. that they behave before reagents like the totally suberised membranes of the cork-cells, or like cuticle. (Comp. pp. 75 and 111.) They alone remain behind after the action of concentrated sulphuric acid, even if the acid has destroyed the surrounding cellulose walls. More exact investigations of its chemical relations are entirely wanting. Further, the peculiar refractive properties of corky walls, the dark black contours when seen by transmitted light, belong to the walls in question, Partly in FIG. 49.—Ranunculus fluitans ; transverse section through the vascular bundle of a strong old.adventitious root (225). u endodermis, 7 pericambium, g outer primordial vessels of the diarch uniseriate xylem g—g ; between g—y and # the phloem. FIG. 50.—(375) A piece of the endodermis in tangential longitudinal section. this circumstance, partly in the wavy folds superposed one on another, in not very thin preparations, lies the cause of the often-described phenomenon, that the un- , dulated strips of the radial walls appear in transverse sections as dark points or lines. Another peculiarity, which again recalls the cuticle, is this, that the suberised parts of the walls swell in sulphuric acid and in potash in the direction of their surface. The undulations appear after the action of those reagents to become higher ; but whether this is really the case, or whether they only become plainer for observation, remains to be investigated. . Like many cork-cells, those of the endodermis often remain thin-walled through- out life, e.g. in almost all Ferns, the walls being either totally suberised, or (a point which requires more extended investigation) having a delicate internal cellulose layer. But on the other hand, there appears not infrequently here also a strong thickening superposed internally on the original membrane: this occurs especially in roots of Monocotyledons, the stems of Potamogeton (in many species, as 124 CELLULAR TISSUE, P. crispus, densus, gramineus, no strong thickening occurs), rhizomes of Cyperacezz, e.g. Carex hirta, exceptionally also in roots of Dicotyledons (Primula Auricula). Comp. Fig. 51. In many rhizomes of Monocotyledons, e. g. Carices, several thick- walled sclerotic layers occur in the region occupied in allied plants by the endo- dermis, when fully developed: it remains to be investigated how far these are endodermis. mete The thickening masses are as a rule more or less sclerotic— ore ae C Co ; lignified, or suberised—only in Pr, Ze Se | . rn ‘ 2050S Oyu 2 Auricula do they consist of cartila- ginous, gelatinous cellulose. In relatively few cases they extend in equal thickness round the whole cell (root of Pr. Auricula, of many epiphytic Orchids, stem of Pota- mogeton pusillus); usually they are more strongly developed on the inner side, than on the outer; roots of species of Carex, Cy- perus, Scirpus, Phragmites com- Ses és l munis, Triticum repens, Asparagus, ~SOOWNOS species of Smilax (the so-called 0 ) Ou CF Pp ax ( MS 7s core-sheath (Kernscheide) of Sar- FIG. 51.—Primula auricula (225) ; transverse section through the heptarch saparilla roots), Draczenas, Palms’, vascular bundle of an adventitious root and its surroundings. % pericam- and in the above-named rhizomes of bium ; g the outer primordial vessels of the vascular rays, which alternate Mited saccuchynas w eudokermis; outsile this i rather thicewaiea CYPETAaceze, the stems of Potamoge- cortical parenchyma, with intercellular spaces tetrangular in transverse {on pectinatus, lucens, natans, pre- a longus?: comp. below, Chap. VIIL The thickening masses are stratified and pitted; only in the investigated Dracaenez they are not pitted (Caspary). The undulation is not present on the thickened walls, but it again appears on the original radial walls, if they can be isolated by destruc- tion of the superposed thickening masses, e.g. by sulphuric acid. Usually the thickening and sclerosis extend to all the cells almost equally, but thin-walled cells often occur between the others. They are found quite solitary, e.g. often in the root of Auricula (Fig. 51); or numerous, but alternating not very regularly with the thick-walled cells, in the root of Strelitzia ovata. But in the vascular-bundle- sheath of the aerial roots of epiphytic Orchids 1-2 longitudinal series of cells remain before each vascular group with an unthickened membrane, which turns blue with iodine and sulphuric acid: in their longitudinal course they are here and there inter- rupted by thickened cells *. The elements of the endodermis are, in all exactly investigated cases, cells in the fullest sense of the word, with a protoplasmic body; in Equisetum, according to * Caspary, Zc. p. 108.—Schleiden, Archiv d. Pharmac 1847.—Berg, Atlas d. Pharm. Waaren- kunde, Taf. II, IV.—Mohl, Palm. structura.—Karsten, Vegetationsorgane der Palmen, Taf. III. fig. 2. , ? Caspary, Pringsheim’s Jahrb. I. p. 443. § Leitgeb, Wiener Acad, Denkschr. Bd. 24, p. 207. PARENCHYMA. ENDODERMIS, 125 Pfitzer, they even contain chlorophyll; further, as in all parenchymatous cells, the nature of their contents is very various; many are poor in contents of definite form, or almost empty; very many have abundant starch grains, and even to a remark- able degree in comparison with the surrounding parenchyma. Also in strongly thickened and sclerotic cells there are often abundant starchy contents, as in the roots of Cladium Mariscus, and Carex arenaria according to Caspary, in the stem of Potamogeton natans, &c. In single cases, namely, in roots of Ficaria and Victoria regia, and in stems of Equisetum, Caspary and Pfitzer found the proto- plasmic body of the cells brown, and contracted to a band stretched between the. undulated walls. As was already indicated at the outset, the layer of cells limiting the parenchy- matous cortex from the air-containing sheath, which surrounds it in the aerial roots of the epiphytic orchids and Aroidez, of Chlorophytum and Hoya carnosa, is a special case of endodermis. It corresponds in all important points with the * protective sheaths,’ and is generally distinguished by one peculiarity only, that in each of the longitudinal series, which its cells form, elongated prismatic elements alternate regularly with short roundish or oval cells. Usually all the cells have thin walls, and in that case (according to Leitgeb always) they are undulated on their radial faces; they have, as far as my investigations extend, a complete suberised outer layer, and a delicate cellulose inner layer. But in many species the long cells are strongly thickened and sclerotic, most strongly, and without pits in Oberonia myriantha (Leitgeb, /.¢.). The short cells are always thin-walled. The long cells contain chiefly watery cell sap. The short ones are characterised by relatively abundant, granular protoplasm, and a large nucleus. On the structure of the roots in question, comp. Sect. 56. CHAPTER II. SCLERENCHYMA. Secr. 28. The name Sclerenchyma, introduced by Mettenius?, here indicates those tissue-elements which have not only thickened their walls at the expense of the cell-cavity, but have also lost the cell quality besides. ‘Together with the sclerotic cells of the foregoing paragraphs they form the strengthening apparatus. But while the former, by reason of the nature of their contents, still take an active part in the processes of assimilation and nutrition, the properties which point to this are wanting in the tissues in question; they appear (besides some connection with the transfer of water) to be in the main only strengthening apparatus, or specific mechanical elements, to use the terms of Schwendener. We will not here again return to the practical difficulties in distinguishing this tissue from the sclerotic cells. Comp. p. 115, and Chap. X. The general properties of the sclerenchymatous elements consist in this, that as the thickening and lignification proceed, the protoplasmic body and nucleus dis- appear, and of these and of the products resulting from their activity only remnants together with watery fluid remain behind, partly as not clearly defined granular contents ; often however they take the form-of rather abundant fine - grained starch, which apparently has no further use, as e.g. in the fibrous ring of the outer walls of Aristolochia Sipho, or of crystals of Calcium oxalate, as in many covering tablets of fibrous bands, to be described below, and in the raphide-~ containing fibres of the cortex of the root of Chamcedorea elegans. According to Schwendener * a part of the fluid contents is replaced in the typical sclerenchymatous fibres by air; they always contain some air in the normal condition. The structure _of the walls is in general that of strongly thickened cell-membranes, with their numerous modifications: these will be more readily described in connection with the single forms. According to the form, and the definite relations of structure which vary for the most part with it, we may distinguish two main forms of sclerenchymatous elements, which, however, are not in all cases sharply defined from one another, viz. (1) short sclerenchymatous elements, and (2) elongated elements, or sclerenchymatous fibres. * Abhandl. d. K. Sachs. Ges. d. Wissensch. IX. p. 432. ? Das Mechanische Princip, &c., p. 110. STONE-ELEMENTS, 127 Srcr. 29. The term Short sclerenchymatous elements may be applied to all forms which have not pointed tapering ends; these are sometimes iso-diametric, sometimes moderately elongated. To this group belong— (2) The s/one-elements (‘stone-cells’ of the Pharmacologists), so called after the stony bodies in the flesh and stalk of many pears, which are composed of them, are almost iso-diametric, rarely rod-like elongated derivatives of cells (‘ rod-cells ‘), with stratified, very strongly thickened membrane, lignified to a stony hardness: this wall is perforated frequently by numerous, usually branched pit-canals, of circular appearance in transverse section (Fig. 52). The narrow internal cavity, which usually disappears, is occupied by a watery fluid with a few granules, or often by a reddish brown, apparently formless mass. Stone-elements of this sort are widely spread among the Dicotyledons, especially in sappy, fleshy parts ; in the suc- culent parenchyma they are sometimes isolated, but usually in uninterrupted connection with one another, forming cir. cumscribed groups, or masses, of which the elements bor- dering on the thin-walled tissue may graduate into the latter by the thickening of their walls at this limit being one-sided FIG. §2,—Transverse section of and weaker. In the so-called stout succulent plants, how- a short sclerenchymatous (stone) ever, such as the Crassulaceze, Cactacez, &c., stony formations DA ee dimen Ea are generally wanting. Exquisite examples are supplied by Serna ee the fleshy body of Helosidee, Lophophytum, Langsdorffia!, seaaiiieapccioik ine fleshy tuberous roots, e.g. Pzeonia, Dahlia (Sachs); Rhizomes, e. g. Dentaria pinnata, the pith of Hoya carnosa*, Medinilla spec., and especially the cortex of ligneous Dicotyledons, in which they are mainly derived from secondary sclerosis of parenchy- matous cells, as will be more closely described in Chap. XV. Transitional forms to the sclerenchymatous fibres are’ supplied by the rod- shaped stone-elements of many cortical layers, the short pointed fibres of the Cinchonez, the short and pointed branched stone-elements of the cortex of Firs and Larches, &c. In the Monocotyledons the elements of this category are rare; but we must include under this head the multiseriate dense layers beneath:the epidermis of stems of Palms *, and elements with large cavity, and large pits, which form in the cortex of the root of many Aroids (e.g. Tornelia fragrans) 3-4 layers of cells outside the endodermal sheath of the vascular bundle, and in Raphidophora angustifolia ® also in the inner cortex of the stem a ring of 1—2 layers in thickness. Typical stone-elements are wanting in the Cryptogams. (2) A second form of short sclerenchyma is represented by the peculiar covering plates which Mettenius ® first distinguished in species of Trichomanes under the name ' Hooker, Trans. Linn. Soc. vol. XXII.—Graf Solms-Laubach, in Pringsheim’s Jahrb. VI. p. §30.—Eichler, Balanophorez Brasilienses, Tab. II. 2 Mohl, Ranken- und Schling-pflanzen, p. 89.—Ibid. Poren d. Pflanzenzellgewebes, p. 32. 8 A. Gris, Ann. Sci. Nat. 5 sér. XIV. p. 50. * Mohl, Palmarum structura, pag. VI. Tab. A.C. Verm. Schriften, p. 136. — Botan. Zeitg. 1871, Taf. II. 5 Van Tieghem, Struct. des Aroidées, /.c, ® Hymenophylleen, /.c. p. 418. 128 SCLERENCHYMA. Stegmata, and which,'as shown by later investigations of Rosanoff*, occur not un- frequently among the Monocotyledons also. They always appear on the outer surface of sclerenchymatous or sclerotic bands of fibres (either such as pursue a separate course, or accompany vascular bundles), and are applied to these in longi- tudinal rows, which by the arrangement of their elements lead us to conclude that they arose by transverse division of spindle-shaped cells. The single elements are small, and have the form of flat, or (in the Monocotyledons) plano-convex, usually rectangular plates, with the flat side contiguous with the fibrous band. As regards their structure they are characterised by unequal thickening on different sides, usually also by partial silicification of their walls; they vary extremely in individual cases according to the species or systematic group. In the species of Trichomanes, the wall is strongly thickened on one side, i. e. on the inner face, which is contiguous with the fibrous band. In some few species the thickening is uniform on this surface; it is equally rare to find it so arranged that it occupies the periphery of the inner wall in a ring-like manner. Usually there rises from the middle of the inner wall into the cavity a cushion-like protuberance, hollowed in the middle, or comb-like bands placed symmetrically near the middle. On the varying special forms of these outgrowths, compare /.c. Those outgrowths protruding’ inwards and the region immediately surrounding them are distinguished from the rest of the wall, which shows the cellulose reaction, by their granular appearance and strong silicification, Similar covering plates, perhaps more properly included under the crystal- containing structures, since each contains an aggregation of calcium oxalate, occur, according to a short statement by Mettenius, in certain of the Cyatheacer. The fibrous bands in the stems, leaves, and roots of Orchideze (Pholidota, Stanhopea, &c.*), Palms (Chamerops, Phoenix, Caryota, &c.), of Maranta compressa, Arundinaria spathiflora, have interrupted longitudinal rows of plano-convex stegmata on their exterior. The convex outer wall of these is thin, the inner thickened to a half-spherical rough body, which almost fills the cavity, and consists mostly of compounds of silicon. Often 2-3 such silicified bodies occur in place of one. (Rosanoff found similar silicified bodies also in cells containing chlorophyll and starch on the fibrous bands in the margin of the leaf of Galipea macrophylla, one of the family of Diosmez.) The fibrous bands in the lamina of the leaf of Scita- minez (species of Maranta, Heliconia, Thalia) show small stegmata, the structure of which seems to differ from that just described, and remains to be investigated. SEcT. 30. Sclerenchymatous fibres, of elongated spindle-like shape, with sharp ends, simple or branched, are the form of strengthening tissue which is universal, especially in Phanerogams; they are sometimes in uninterrupted lateral connection, and united, with pectinated * ends, into bundles and sheaths 3 sometimes they are imbedded singly in other tissues. 1 Botan. Zeitg. 1871, p. 749. ? Compare Link, Botan, Zeitg. 1849, p. 750. ® [It is believed that this translation will convey the meaning intended by the use of the word ‘verschrankt,’ the idea being that of an arrangement similar to the fingers of two folded hands, or of two combs (Jectew) with the teeth of the one passing between those of the other. The word pectinatory will be used subsequently in describing the course of the vascular bundles (Chap. VIII, A).] FIBRES. 129 The fibres in question are frequently called also Bast-fidres, or Bast-cells, after a region in which they occur especially often in the Dicotyledons, and, in connec- tion with these terms, Sanio has called those fibres which occur in the secondary xylem, and which belong also in part to this category, bast-fibre-like, or Lbrzform fibres. Comp. Chap. XIV. P, Moldenhawer? calls them fibrous tubes. The name bast, or liber, is at present used for two quite different things. Originally it was used as a topographic anatomical term, for a definite region of the cortex of the Dicotyledonous stem, which is, it is true, as much characterised by definite forms of * tissue occurring in it as by its position (comp. Chap. XV). Among these forms of tissue: - sclerenchymatous fibres are quite generally characteristic; they are present indeed in many cases in very large quantity, and are very conspicuous as compared with the other tissues. On the latter ground, and since the really characteristic structure of this cortical region was not known, they were considered as the essential tissue of the bast-region, and the name bast was transferred from the region to the sort of tissue, but later again used for both without sharp distinction. Since the sort of tissue is by no means limited to this region, the result was that bast was found at other places than in the bast, or that there is bast without bast, in other words that doubt and controversy arose. Now it is in itself indifferent which meaning is attached to the name, and grounds may be brought forward for authorising both the above uses of it, but it certainly cannot be used for two quite different ideas. In the choice to be made accordingly it seems to me decisive that the topographic meaning of the word is the older, and has always been the more usual. Its use will therefore be here limited to the region to be treated of later, and the fibres in question will therefore be called Bast-/ibres, wherever they belong to this region. The form of the sclerenchymatous fibres varies within the above stated limits ac- cording to species and part. Their transverse section is acutely angular, where they are closely united into bundles; it is round in such fibres as lie single and loose in intercellular spaces, as in many leathery leaves, in the foliage of many Aroidez, &c. Those firmly connected into bundles are as a rule sempie, i.e. unbranched, spindle- shaped, usually with continued and gradual decrease of transverse section towards the ends, while the much-elongated forms are usually drawn out at the ends into ex- tremely fine points. This form—subject it is true to many exceptions—is the rule also for the fibres occurring in longitudinally elongated parts, but not closely con- nected into bundles: for instance, for most of the fibres, and even the isolated bast- fibres, which are scattered in the parenchyma of the roots of many palms (Chamzdorea elegans), the petioles and pinnze of Cycadez*, &c. A remarkable peculiarity of form is shown by the very long bast-fibres of many Apocynez, and Asclepiadez (Nerium, Vinca, Asclepias spec.), since they are in their longitudinal course alternately nar- rowly constricted, and then again suddenly distended; the same is the case, in rather irregular form, in the bast-fibres of species of Sida, Urena, and the species of Corchorus which yield Jute *. Even the spindle-shaped fibres, which have just been mentioned as being usually simple, show not uncommonly, when isolated, shortly- and unequally- branched ends, or here and there at other points a branch usually of insignificant size. 1 Beitr. pp. 11-61. . ? Moldenhawer, Beitr. p. 34. 8 S, Wiesner, Microskop. Unters, p. 24 ff, and Idem, Rohstoffe, cap. 11. K 130 SCLERENCHYMA. On the other hand there commonly occur in Phanerogams fibres which are freely and often abundantly dranched, and of a form which varies according to the special place of their occurrence: these usually occur in dissimilar lacunar tissue, with their branches projecting or pushed into its interstices. Inasmuch as these project like many branched hairs into wide, air-containing spaces, as in the Nymphz- ace, Limnanthemum, Aroideze, Rhizophora, the description of them will be more clearly given when we treat of these spaces (Sect. 53), and we need only draw atten- tion here to their connection with the tissues treated of in this chapter. They also _occur more especially in numerous tough, leathery foliage-leaves, though not in the ma- jority of them ; they push their branches into the intercellular spaces of the parenchyma, and appear to serve as strengthening apparatus for that tissue. With reference to the relations of their arrangement, to be treated in Chaps. IX and X, may here be men- tioned the short-branched fibres in the leaf-lamina of Proteacez (Hakea nitida, ceta- tophylla, saligna, &c.1), the long- and finely-branched fibres in the lamina of Olea europea, emarginata, fragrans®, the thick, starlike, short-branched ones of Camellia FIG. 53.—From a transverse section of the leaf of Camellia japonica. P parenchymatons cells, with chlorophyll grains and oil drops; F thin vascular bundle; / branched sclerenchyma fibre. From Sachs’ Textbook, japonica ® (Fig. 53), Statice monopetala, the beautiful stellate, many-armed ones in the lamina and petiole of Fagrzea obovata, and auriculata*. Also the leaf-lamina of the above-named Aroidez, especially the Monsterinez, and the Nymphzacez, may be here again cited. Stellate-branched fibres occur in the foliage-leaf of Sciadopitys, Dammara, Araucaria imbricata ®. Long-branched ones, sometimes of huge size, form at least half of the substance of the leaf in Gnetum Gnemon, and G. Thoa. The relation between breadth and length of the fibres varies greatly both ac- cording to species and in different parts of the same species, and in the self-same part and the self-same bundle it often varies within wide limits. This is to be taken + Meyen, Harlemer Preisschrift, p. 84, Taf. V.—Mohl, Verm. Schr. Taf. VII. fig. 2—Schleiden, Grundz. 3 Aufl. I. p. 277. ? Moldenhawer, Reitrage, p. 61.—Thomas, Zc. Pp. 32. ’ Kraus, Cycadeen-fiedern, 7. c. Pp. 327. * O. Buch, Ueber Sklerenchymzellen. Diss. Breslau, 1872, p. 16. 5 Thomas, /.¢. p. 35.—Mohl, Botan, Zeitg. 1871, p. 8 FIBRES. 131 into account in the statements of average, which require confirmation throughout. The published measurements of simple bast-fibres give for the shortest forms, such as those of Peruvian bark, a relation of about 1:10 to 1: 20; for the longest, found among the Urticaceze, a length exceeding the greatest breadth two or three thou- sand times (to 1: 4000). The branched forms are as a tule relatively short and broad, e.g. Fig. 53, but much-elongated specimens also occur. As examples we may cite the following few measurements found by Mohl? and Wiesner”, and in the Quinine bark by Vogl®, for fibres of bast and bundles; for further details we must refer to Wiesner’s summary, /.c. Where only the length is _ given, the medium of the measurements of breadth given may serve as the breadth. Length. Greatest breadth of fibre. mm. mm, Species of Cinchona, bast . . 0°875— 1°25. . . . 0'031— 0°25, Tilia, bast. . . 1. 2. . . 099 —2°65. . . . average o'or5. Corchorus spec, (Jute), bast. . 08 gi... . » «oro Phormium tenax, leaf. . . . 27 —5'65... ~~. » ~~ O°0T3. Linum usitatissimum, bast . . 20—40 . . . . O15 —O'I7, Cannabis sativa. . . . . . «o&more ... . o115 —o'28, Boehmeria nivea . . . . . upto220 .. . . 0104 —0'08, #sculus Hippocastanum. . . 1°35 —1°8. Bignonia radicans . . . . . 06 = —1°35. Bombax pentandrum . . . . 2°025 — 2°92. Daphne Mezereum ... . up to 3°375. Clematis Vitalba . . . . . 045 —0°85, Bambusa spec. . . . . . . I°8 —3'015. Cocos botryophora. . . . . 0°855— 1°350. Lonicera Caprifolium, bast . . 18:0 — 26'0, Asclepias Cornuti . . .. . up to 26°0, Urtica dioica. . 2. 1. we, up to 77°0. The considerable length of many fibres, together with the occurrence ofthe chambered fibres to be described below, has given rise to the view that a fibre does or may arise by the coalescence of several meristematic cells disposed in a longi- tudinal series*. More exact investigation however can find 4 przor¢ no sound ground for this, and all minute observations have shown that each simple or branched fibre results from the metamorphosis of one cell °. The wall of the sclerenchymatous fibres is thickened to an extent which differs according to each special case, and usually so that the lumen is greatly reduced (centripetal) ; the thickening mass is nearly equally thick all round, or in many cases it projects inwards much more strongly at certain points than at others, e.g. bast- 1 Botan. Zeitg. 1855, p. 876. ? Mikroskop. Untersuchungen im Laborat. d. polyt. Inst. Wien; and Rohstoffe, d. Pflanzen- reichs, cap. II. 3 Die Chinarinden des Wiener Grosshandels, &c. 1867. * Meyen, in Wiegmann’s Archiv, 1838, I. p. 297.—Schacht, in Berl. Acad. Monatsber. 1856, p. 517: Lehrb. II. p. 567.—Hanstein, Milchsaftgef. p. 45. 5 Compare Unger, Wachsthum d. Stammes u. Bildg. d. Bastzellen, Wiener Acad. Denkschr. Bd. XVI; Boehm, Wien. Acad. Sitzungsber. Bd. 53; Sanio, Botan. Zeitg. 1860, p. 210. Further, the statements in Chapter VII upon the intercellular fibres of the Aroidez, and Chapter XIV, K 2 132 SCLERENCHYMA, fibres of Corchorus ‘spec., Abelmoschus tetraphyllus; Sida retusa, &c:1 The thick: ening mass is either continuous, as for instance in most fibres used in manufacture, according to Wiesner’, or in many cases provided with narrow pit-canals, which, especially in the fibres associated so as to form bundles, have almost always the form of narrow, rectilinear, longitudinal, or parallel oblique slits like a left-handed screw 3 FIG. 34.—Pteris aquilina. 4 half of a FIG. 55.—Half of a thick brown-walled sclerenchymatous fibre sclerenchymatous fibre, with fron the stem; B piece of one of these crystals of calcium oxalate im- more highly magnified (550); 2 profile bedded externally in the wall, view of the slit-like pits; C transverse from the stem of Welwitschia section; @ limiting lamella; 4,c inner mirabilis. From Sachs’ Text- layers of the wall. From Sachs’ Text- book. book. (Fig. 54). Still exceptions to this occur in the short fibres, as the simple ones of the Quinine bark, and the branched ones of the leaf of Camellia: here there are narrow canals, not slit-shaped, but round in section, The sclerenchymatous fibres described by Milde‘ as being spirally thickened, which form a many-layered closed -covering on the upper side and on the nerves of the under side of the leaf of Acro- pteris radiata, would be better described as having slit-like pits. ‘Their membrane, which is thickened till the lumen is almost obliterated, has very numerous, regularly * Compare Wiesner, /. o . ? Rohstoffe, p. 305. ® Mohl, Zc, p. 876.—Schwendener, /.c. p. 8, * Filices Europe et Atlantidis, p. 40. . FIBRES, 133 oblique slit-like pits, and these alternate with thick bands of wall of equal height. The epidermal elements overlying this fibrous covering show, as was intimated on p- 71, the same structure. For the more minute structure of the thickening masses, those general rules hold which apply to the structure of the cell-walls', In the fibres united into bundles, and those in the bast, there may often, but not always, be distinguished three dif- ferent concentric systems of layers, or sheaths, the outermost limiting layer, an inner layer, and a middle layer, which is usually much broader and softer. The fibres of the Apocynez and Asclepiadeze are excellent examples of the striation and areola- tion of the wall_—The branched fibres in the leaf of Sciadopitys, Dammara, Araucaria imbricata, Nymphzacez, and especially the colossal spindle-fibres made known by Hooker ?, which lie scattered in all parts of Welwitschia mirabilis, are characterised. by numerous crystals of calcium oxalate which are imbedded in the outer layers of their walls, and which, especially in Welwitschia, attain a considerable size (Fig. 55). The wall of the sclerenchymatous fibres is lignified, to a very variable extent according to the special case: of the bast-fibres used in manufacture, e. g. according to Wiesner, those of Flax, Hemp (light yellow with aniline and sulphuric acid), and of Hibiscus cannabinus turn blue (of different shades) with iodine and sulphuric acid, and with aniline and sulphuric acid not at all or hardly yellow; with the preparation of iodine the fibres of species of Corchorus, Sida retusa, Urena sinuata, é&c. turn yellow or brown, with aniline and sulphuric acid yellow. In the Ferns and Rhizocarpez the fibres of this category have also the above-mentioned (p. 121) characteristic dark- brown colour. In fibres in the bast Sanio* often found the especially thick inner layer of the wall cartilaginous and gelatinous, and that it swelled in water, and turned violet with Schultze’s solution or solution of iodine in potassium iodide (e.g. in Cytisus Laburnum, Morus alba, Ulmus suberosa, Celtis australis, Ficus Syco- morus, Robinia pseudacacia, Gleditschia triacanthos, Quercus pedunculata, Passiflora suberosa); this phenomenon also occurs in various modifications in the fibrous elements of the secondary wood of Dicotyledons, and will be described with the other properties of these elements in Chap. XIV. Conversely it sometimes happens that sclerenchymatous fibres develope from originally collenchymatous cells, in which case the inner layers of the walls become hard and lignified, while the outer retain the original collenchymatous character: e.g. in the bands accompanying the vascular bundles of Eryngium planum, and Astragalus falcatus *. As regards the contents of the fibres we must refer to what was above stated (p. 126) for the sclerenchyma generally. The granular constituents or remnants of the contents, enclosed by many fibres with a larger cavity, e.g. the enlarged parts of those of the Asclepiadez and Apocynex, have repeatedly led to the view that the bast-fibres contain the characteristic Jefe, which exudes on cut surfaces in the Asclepiadeze, Euphorbiaceze, &c., a false idea, which will be discussed in Chap. VI. In the majority of cases, and in all those which have been hitherto noticed, the 1 See Hofmeister, PAanzenzelle, § 27, 28. ? Trans. Linngean Society, vol. XXIV (‘Spicular cells’). 3 Botan. Zeitg. 1863, p. 105.—Ibid. 1860, Taf. VI. 15 and 16. * Schwendener, /.¢. p. 5. 134 SCLERENCHYMA. cavity of the fibres is a continuous hollow, though it is very narrow, and often ceases far from the pointed ends. In the narrow contractions in the Asclepiadez and Apo- cynez it may, it is true, be doubted whether it is not sometimes completely inter- rupted by the thickening of the walls. But on the other hand chambered fibres are often to be found, i.e. such as are cut up into segments or chambers by relatively thin transverse walls continuous with the inner layers of the lateral walls: e.g. in the bast of Aisculus Hippocastanum, in the cortex of roots of Palms, as Chameedorea elegans, Also the chambered fibres in the bast of Vitis, Platanus, Pelargonium roseum, Tamarix gallica’, in the cortex of Aristolochia Sipho, &c., which contain starch for a time, should perhaps be connected with the above, as cases in which the functions of the cell slowly disappear; the same may be said of the fibres produced from col- lenchyma, which are common in the cortex of stems. The chambered elements of the secondary wood of the Dicotyledons, which are also connected with the above, will be spoken of in Chap. XIV. * Compare Sanio, Ueber die im Winter Starke fiihrenden Zellen, &c, (Halle, 1858), p. 12; Botan, Zeitg, 1863, p. 111. CHAPTER IIL SECRETORY RESERVOIRS. Sxct. 31. Bodies of a nature similar to the secretions of the dermal glands (Sect. 19), such as mucilage, and gum, resin, ethereal oils, and mixtures of these designated balsam, milky emulsions of the bodies of both categories which are known in the dry state as ‘Gum-resins,’ are often found laid by in the interior of the tissues; they occur on the one hand in special Sacs, which develope during the differentiation of tissues from definite cells of the meristem: these, retaining their membrane, and growing considerably, are filled completely with the bodies in ques- tion, and thereby lose their original cell-nature ; or they are found in special Zuter- cellular spaces. There occur in many plants other sacs, arranged similarly to the above, which also arise with the first differentiation of tissue from cells of the meristem, and con- tain as their sole or preponderating contents crystals of oxalate of lime. All these places of secretion or reservoirs are closely related to one another. The aggrega- tions of crystals are often associated with large deposits of mucilage in the cavity of a sac, so that one may speak of mucilage-sacs with crystals (e.g. tubers of orchids) or of crystal-sacs with mucilage (e.g. Raphide-bearing sacs), according to the pre- ponderance of one or the other body. As already stated, resin and mucilage often occur mixed together. The form of the sacs merges not uncommonly into that of intercellular spaces filled with the secretory mass, since rows or groups of the former, by absorption of their walls, coalesce to an amorphous intercellular mass. Further, sacs and intercellular spaces with like contents often mutually replace each other, since, in the first place, the same body in different members of the same plant sometimes fills sacs, at other times intercellular spaces, e.g. the red resin of species of Lysimachia and Myrsine ; or secondly, of closely allied plants some have sacs, others intercellular spaces filled with the same secretion, at the same points. Examples of this will be found below, among the Coniferze, Composite, &c. Finally, in families more remote from one another, there occurs only one or the other form of secretion and of the reservoir containing it. The mode of formation of the secretion in the interstitial dermal-glands cor- responds closely with that of the schizogenetic resin-passages, which are to be described below. We must here refer especially to the depressed glands of Psoralea. In consideration of its known properties, calcium oxalate can only be regarded as a body, which is removed from the metastasis of the plant, and is secreted or 136 SECRETORY RESERVOIRS. excreted. Direct observation teaches us the same of the mucilage, resins, and ethereal oils of the dermal-glands. The fact is no less evident that the resins, mucilages, &c., which are laid by in circumscribed reservoirs, e.g. in the resin-sacs of the Laurinez, Piperacez, Zingiberacez, &c., after they begin to be secreted in the meristem, remain, like the calcium oxalate, laid by without further use. In accordance with all these facts we are bound to regard the whole series of the bodies in question, like the secretions of the dermal-glands, as bodies excluded from the constructive metastasis, and to term them, together with these, Secretions. Their occurrence as admixtures of the contents, or as constituents of the membrane of active cells, which may be proved for all bodies of this category, is no argument against this generalisation, since on the one hand calcium oxalate shows plainly that one and the same body may be excreted both in small quantity in an assimilating cell, and in large mass in a special reservoir; on the other hand, in the uncertainty of our present knowledge, a fundamental difference is always possible between what is in the one and in the other case termed, for instance, resin. And finally, this view does not affect that of the application of the secretion to some further uses by the plant, as, for instance, in the well-known case of the hairs on buds. On these grounds we group the whole of the above-described reservoirs together as secrelory-reservoirs. "There may be distinguished reservoirs of crystals, mucilage, resin, &c., according to their exclusive or preponderating contents. Since resin. and ethereal oil occur usually as mixtures, and rarely separate, and since we cannot here enter upon chemical details, which are often uncertain, we shall in the sequel use the words reservoirs of resin, oil, and balsam without claim to exact indication of contents, and usually in connection with the meaning customary for each special’ case. The term gum-resin is used to indicate, but with still smaller claim to accu- racy, the mixture of watery and resinous secretions, which is milky when fresh. According to their structure the reservoirs may be distinguished as Sacs, i.e. struc- tures derived from cells, which retain their walls, and are therefore usually termed cells; and zxfercellular cavities, which according to their form are termed either. passages or gaps. For many of these forms, which vary in structure and contents, the term glands, or zzéernal glands, is in use. It will be difficult to banish it, since it has established itself in the incorrigible terminology of Systematic Botany, although, as the sequel will show, it is not at all wanted. If it is to be retained, it ought accu- rately to be used for all secretory-reservoirs, any other use of it is purely arbitrary and conventional. With those which certainly belong to this category, we must connect many doubtful structures, such as many ‘tannin sacs,’ the ‘ vesicular vessels’ of the species of Leek, and others to be named below: this classification may be corrected when more exact observations have been made. We have already drawn attention to the alternative occurrence of the different forms of secretory reservoirs, in different members of the same plant, or in different genera, or larger circles of affinity. Similar alternative relations occur here and there between reservoirs and latici- ferous tubes (comp. Chap. VI). Even if one discounts the Aroidez and Musacex, the laticiferous tubes of which should perhaps be enumerated in the present chapter, the fact is certain that internal secretory reservoirs are absent from all plants which SACS CONTAINING CRYSTALS. 137 are provided with laticiferous tubes. In the group of Artocarpez, which in common with the majority of its allies is provided with laticiferous tubes, these are absent, according to Trécul, in Conocephalus naucleiflorus, while in their place this plant has mucilage-containing sacs and cavities. Among those Composite, which have been investigated, the Cichoracez are distinguished from the rest by their having lati- ciferous tubes, and by the absence of the oil-ducts present in the others: only in Scolymus are both organs developed. Further, it is often impossible to ignore an alternative relation between the occur- - rence of dermal-glands and internal secretory-reservoirs. In the Cycadez, Conifere, Lauracez, Umbelliferze, Aurantiaceze, and Clusiacez, which have specially large num- bers of the latter structures, dermal-glands are absent or rare. For other families, e.g. the Labiatee, the converse holds. Exceptions, with both sorts of organs side by side, occur not uncommonly it is true, e.g. Dictamnus, and many Compositz with glandular hairs and internal reservoirs. And finally, we must not omit to notice that both organs may be altogether absent, as e.g. in the Gramina, Cyperacez, Palms, many Cruciferee, Ranunculacex, in Taxus alone of the Coniferze, &c. &c. The relations, above brought into prominence, between the different organs which form secretions, should always be kept in view during their consideration, in which the first duty is to separate them according to their structure: we will therefore occupy ourselves first with the sacs containing secretions. .The intercellular reser- voirs will be treated of in Chap. VII, and the intermediate structures will be noticed in a fitting place. 1. Sacs containing Crystals. Sxct. 32. It is known that crystals of calcium oxalate are generally distributed as constituents of the cell contents. In certain sacs they almost exclusively fill the internal space, and these may be distinguished as crystal-bearing sacs. The crystals * consist, as far as is known, entirely of calcium oxalate, which is crystallised either in the quadratic or klinorhombic system—according to Souchay and Lenssen, when quickly deposited it takes the klinorhombic form, with the composition a \ C,0,+2H,0; when the crystallisation is slower it forms as quadratoctahedra of the composition CaO ce \ C,0,+6H,0. The fundamental form of the crystals belonging to the quadratic system is the quadratoctahedron, that of the klinorhombic crystals, which are far commoner in plants, is the hendyohedron: derived forms occur of the most various shape, e.g. klinorhombic columns, klinorhombic plates, twin forms, and blunting of corners. As specially common forms, which can hardly be accurately defined crystallographically, may be named the spear- or needle-shaped crystals, elongated and pointed at both ends, which De Candolle? has termed Raphides. They belong most probably, _ } See Holzner, Flora, 1864, pp. 273 and 556.—Ibid. 1866, p. 413. 2 Organographie végétale, I. p. 126 (fapis =needle). 138 SECRETORY RESERVOIRS. according to Holzner, to the klinorhombic system. Besides these different, singly developed crystals, there often occur. others imperfectly developed, and grown together to angular or stellate groups, which, according to Holzner, may belong as well to one system as to the other. The form and system of crystallisation is indefinite in the case of the quite small crystals, which often occur, and appear more like small granules: on these sharp angles and edges may be recognised with a high power. In the sacs the fully developed klinorhombic forms and the groups almost always occur singly, rarely two together, and fill the greater part of the cell: the Raphides appear always in larger number ; as a rule they are nearly equally Jong and parallel and are closely packed in thé sac in a bundle, so that all the ends in the same direction are in one plane; more rarely they vary in length and direction, as in the cortex of many species of Aloe, e.g. Aloe arborescens, in the parenchyma of Mirabilis, and the very small Raphides in-the numerous crystal-sacs of the Cinnamon- bark of Ceylon. Here thé minute granule-like crystals in an innumerable multitude fill the sac completely, so that in transmitted light it appears to have quite black, densely granular contents: the same occurs in the herbaceous parts of many Solanez}, of Amarantus retroflexus, caudatus, and allies, Sedum ternatum, in the pith and cortex of Sambucus nigra, the cortex of Betula verrucosa, Alnus glutinosa, Staphylea pinnata’, and the bark of the officinal species of Cinchona’, The form of the crystal-bearing sacs is closely related to that of the crystals contained in them, when the latter attain considerable size ; but it cannot at present be definitely stated whether the form of the crystal is dependent upon that of the sac, or the converse. The iso-diametric grouped crystals are contained in sacs re- sembling them in shape, the shorter or longer klinorhombic forms fill sacs of cor- responding shape, which are even of very much elongated prismatic or spindle form, e.g. in the rhizome and leaf of species of Iris*, and in the leaf of Aloe Africana. The sacs containing raphides are elongated in the same direction as the bundle of raphides where the raphides are very large, as in the cortex of Aloe arborescens, in the bulb of Scilla maritima they often attain a great length, in the latter case more than 5mm >, These phenomena appear very striking in the bast-bundles of dicotyledonous plants, the tissue-elements of which are derived from elongated spindle-shaped cam- bial cells. The crystal-bearing sacs arise in this case by transverse division of a cambial cell (Chap. XIV); in those of Guajacum, and Quillaja, which contain a single elongated klinorhombic crystal, few divisions occur: each of the products of this pro- cess (? all of them) becomes one crystal-bearing sac. Also it often happens in plants with small solitary crystals or groups of crystals, that only single products of trans- verse division may develope to crystal-bearing structures. But in very many ligneous plants one cambial cell divides by transverse walls into numerous chambers (20-30), which are hardly or not at all higher than broad, and each of these is filled by a crystal or a group. The general outline of the original cambial cell is meanwhile ? Corda, Beitr. z. Kunde d. Kartoffel, &c., in Hlubeck’s GEcon. Neuigkeiten, 1847, Nos, 58-60. ? Sanio, Monatsbr. d, Berliner Academie, April, 1857. ® Fliickiger, Pharmacognosie, p. 365. * Unger, Anat. und Physiol. p. 123. ® Fliickiger, Pharmacognosie, p. 187. SACS CONTAINING CRYSTALS, 139 retained, while the whole series of chambers may be isolated, remaining still con- nected together like a chambered fibre’. Hartig has called these chambered or septate sacs, crystal-bearing fibres (Krystallfasern).—Similar phenomena occur also in many woods, e.g. Herminiera Elaphroxylon, and on the outer surface of vascular and fibrous bundles. The stegmata of Mettenius on the brown fibrous-bands of Cyatheaceze (comp. p. 128) may perhaps belong more properly to this category. Sacs with very small and numerous crystals, as those of Solanum, Sambucus, &c., usually differ but slightly in form and size from the surrounding cells. As regards the structure of the crystal-bearing sacs, the bundles of Raphides lie at first within a protoplasmic utricle: in all, or at least in all carefully investigated cases, they are enclosed, when mature, by a rather thick layer of homogeneous, transparent mucilage, which is in its turn surrounded by the slightly thickened cellulose wall: the mucilage reacts, in a few investigated cases”, similarly to gum arabic, it swells quickly in water, and disappears (dissolves?), It remains to be in- vestigated how far this mucilage belongs originally to the membrane or to the contents of the cell; according to Frank’s statements respecting the mucilaginous sacs, containing a small bundle of Raphides, in the tubers of Orchis, the latter is probable. The presence of the mucilage is the cause of the quick swelling of the raphide-bearing sacs in water: their membrane bursts, and the Raphides escape with the swelling mucilage, and scatter themselves through the water. In the elongated or spindle-shaped raphide-bearing sacs, which are common, e.g. in the Aroidez, the bursting and escape of the needles usually occurs, as Turpin® has thoroughly de- scribed, at one or at both ends. Hanstein’s* raphide-containing sac-vessels (vesicular vessels, ‘Schlauchgefasse’) are mucilaginous raphide-bearing sacs arranged one above another in long longitudinal rows. These series of sacs occur in large quantity in the parenchyma of many Mono- cotyledons, stem and leaves of Commelinez, Palm stems, e. g. Chamedorea: Hanstein found them of large size in the foliage, stems, leaves, and bulb-scales of many Amaryllide, of the genera Amaryllis, Spreckelia, Crinum, Pancratium, Eucharis, Aloestrmeria, Narcissus, Leucojum, and Galanthus. In these cases they are found 1-2 layers of cells below the epidermis, and especially in the parenchyma of the lower (outer) side of the leaf. In the Liliacez they are less common: they are strongly developed in the leaves of Hyacinthus orientalis, and also in Agapanthus (compare Hanstein, /.c.). In the foliage leaves of Scilla, Ornithogalum, Muscari, there are short series, and isolated sacs, and in the scales of the bulbs of these plants only isolated ones. The stems of Commelinez are best fitted for the investigation of the series of sacs in question. In the growing internodes of these plants, both in the parenchyma of the cortex and of the middle of the stem, there may be observed single longitudinal rows of cells, each of which is loosely filled with a bundle of parallel raphides. The cells are at first 1 Compare Sanio, Monatsbr. d. Berlin. Acad. 1857, p. 261; for further particulars see below, Chapter XIV. . ® Hilgers, in Pringsheim’s Jahrb. VI. p. 286. \. 8 Sur les biforines, Ann. Sci. Nat. 2 sér. tom. VI. p. 5. * Ueber ein System schlauchartiger Gefuisse, etc. Monatsber. Berlin. Acad. 1859, p. 705.— Die Milchsaftgefasse, p. 33- 140 SECRETORY RESERVOIRS. cylindrical. As the internode extends, the length of the cells and of the raphides increases, till in the case of the cells it exceeds their breadth on an average 3-4 times. Hitherto a thin protoplasmic layer, with a nucleus of sharp contour, lines the delicate cellulose wall, As the internode extends further, the cells which remain thin-walled, become 10~20 times longer than they are broad, the protoplasmic parts disappear, while round the bundle of raphides there is seen only transparent mucilage, which shrinks greatly but without turning misty in alcohol, swells quickly in water till it is unrecognisable, turns yellow | with Schultze’s solution, and is not dissolved in potash. Meanwhile the raphides do not increase perceptibly in number or size, they form henceforth a relatively small group in the sac filled with hyaline mucilage. According to Hanstein the members of such a series of sacs coalesce, at least frequently and partially, to continuous long tubes, by the breaking down of the transverse cellulose walls which separate them. But the observa- tions cited in evidence of this are not sufficient to substantiate it. It is true it is often found in longitudinal sections that the raphides are irregularly displaced, and have bored through the delicate transverse walls of the series of sacs; but on the other hand sacs are also found closed at both ends, and dense bundles of raphides in them. And one can often directly see the displacement of the raphides and the perforation of the transverse walls in progress before one’s eyes. The action of water produces this result, the mucilage swells in fundamentally the same way as in solitary raphide-containing sacs. I could not prove to myself the occurrence of spontaneous perforation of the transverse walls, that is of a coalescence of the series to a continuous tube or ‘ Vessel.’ Where I found a perforation already present, it was a gaping rent, such as is seen to be formed when the wall bursts. Further, it is not impossible to suppose that even in the living plant, when too much water is present, walls may burst, and so the same phenomena appear as are seen in sections. According to all these data, which coincide in the main with the statements of Hanstein, the structures in question may be regarded as nothing more than a special kind of raphide-containing sac distinguished by form and arrangement. Rosanoff! was the first to find in the pith of Kerria japonica, Ricinus com- munis, in the sacs which accompany the vascular bundles of the petiole of Aroidex (e.g. Anthurium rubricaule, Selloum, Pothos argyrea, Philodendron Sellowianum), as well as in the parts of the flower of Encephalartos and Nelumbium, that groups of crystals are connected with the membrane; either their apices are in close contact with the lateral wall, or they are suspended by bars of cellulose, which extend from the wall into the cavity, as far as single points of the group; these bars are often branched, and often hollowed like tubes. De La Rue found a similar attachment in the parenchyma of the leaf of Hoya carnosa in the case of small groups of crystals, contained in cells which have protoplasm and even chlorophyll (these, however, in the strict sense do not belong to this category), and also in the leaf and petiole of Aroidez (Pothos crassinervis, &c.). Further, Pfitzer*, following up an older observa- tion of Schacht*, showed that the large solitary klinorhombic crysials contained in the foliage of Citrus, and those in the cortex of Salix aurita, Populus italica, Celtis australis, Fagus sylvatica, Rhamnus Frangula, Acer opulifolium, and Platanus orientalis, are closely surrounded by a cellulose skin, a large part of the surface of which is attached to the cellulose wall of the sac: this skin arises from the protoplasm of the young crystal-bearing cell, which surrounds the crystal: at first it lies free, later it becomes firmly attached to the cell wall. At the point of contact the lateral * Botan. Zeitg. 1865, p. 329.—Ibid. 1867, p. 41.—Compare also De la Rue, ibid. 1869, p. 537» 3 Flora, 1872, p. 98, Taf. III. ; ° Abhandl, Senkenberg. Gesellsch. z, Frankfurt a M. I. p.1g0, Taf. VII. fig. 21. f. . if SACS CONTAINING CRYSTALS, I4I wall of the sac is often strongly thickened, especially in Citrus, where the crystal appears inserted in the very thick lateral wall, or in a conical excrescence of it. In the septate sacs of the wood of Herminiera elaphroxylon’ there lies in each of the almost cubical segments one klinorhombic plate, with one side fitted into the strongly thickened inner wall of the sac, while the rest almost fills the cavity of the segment. I could not in this case find a membrane surrounding the crystal. The space not occupied by the crystal in all these sacs is in the mature state apparently filled with water. Many crystals of whatever form, with the exception of Raphides, appear to lie free within the membrane of the sac, being either closely surrounded by it, but without attachment, or suspended in an apparently watery fluid. Thus, e.g. the large crystals of Iris, and the crystalline granules. For all these cases, however, it remains to be more definitely determined whether a gelatinous coat or an attachment to the wall is present or not. The general occurrence of such a condition is attested by Payen in his statements on the occurrence of silicious coats round grouped crystals, and of membranous sheathing layers of a granular substance which turns brown with iodine. Compare Hofmeister, Pflanzenzelle, p. 393. Crystal-containing sacs occur in all parts, and in all tissues of plants; they appear most abundantly, and often in very great quantity [the stem of Cereus senilis contains in the dry substance more than eighty-five per cent. of calcium oxalate (Schleiden)], in the parenchyma of sappy foliage, and in leathery leaves, bordering closely on the vascular bundles, and arranged in rows which follow these, in the bast and pith of - dicotyledonous plants, often also in the secondary xylem-parenchyma (Pterocar-us santalinus, Heematoxylon, &c.), and in the medullary rays of the wood (e.g. Camellia japonica, Vitis; compare Chapter XIV): where large air-containing intercellular spaces are present they are often particularly numerous at the limits of these, and project into them: e.g. Aroidez, Pistia, Myriophyllum. They occur in most families, and usually in all genera and species of a family: in those in which regular crystal-bearing sacs are rare, or absent, the calcium oxalate is often deposited in the form of small crystals in the contents of parenchymatous cells, or, as in the Cupressinese, Taxinese, Ephedra, and Welwitschia®, in the cell- membranes. The more generally this rule applies the more worthy of attention is a series of exceptions. In the Equiseta no oxalate of Jime is observed anatomically. The same is the case in most Ferns, Graminez, and the Potamez (with the exception in the Phanerogams of the parts of the flower). Still exceptions occur in many of the above families: such as crystals in the Epidermal cells of Asplenium Nidus, in the covering plates of the Cyatheaceze (comp. p. 128), numerous clusters of crystals in the parenchyma of the stem of Panicum turgidum. On the other hand, crystal-containing sacs, or crystals, are not found at all in certain species or genera of families, the members of which, as a tule, contain them in large quantity. In Nicandra physaloides and Petunia nyctaginiflora I found no crystal-containing sacs, while the rest of the Solanacez, which have been in- 1 Hallier (Botan. Zeitg. 1864, Taf, III) gives the outline of these cells correctly, but with a wrong description. ; - 2 Compare Graf zu Solms-Laubach, Botan. Zeitg. 1871. 142 SECRETORY RESERVOIRS, vestigated, have them in abundance. According to Gulliver, crystal-containing sacs are wanting in Tulipa silvestris, Fritillaria Meleagris, Lilium Martagon, candidum, aurantium, whilst most other Liliacee have them in plenty: Sparganium has many Raphides, the species of Typha have no crystals. Among the Lemnacez?, no Wolffia has crystals; the Lemnz and Spirodelz have numerous raphide-bearing sacs, the latter having also many clusters of crystals. The form of the crystal-bearing sacs, and of the crystals within them, is cha- racteristic for many divisions, families, and species”; still general and absolute rules cannot be Jaid down. In most families of Monocotyledons Raphides occur exclu- sively, or they preponderate largely, and often occur in vast quantity, e. g. Liliacez, Orchidaceze, Bromeliacez, &c. But in species of Allium there are no Raphides, and, as far as is known, crystal-containing sacs are entirely absent. Instead of thése there lies in the middle of each cell of the subepidermal parenchyma, on the outer side of the young scales of the bulb, one prismatic crystal, or several grown together ® (this is specially well seen in A. sativum). In others, e.g. the Araceze, sacs containing Raphides and clustered crystals occur side by side, often in the same section. In the Iridaceze only large columnar solitary crystals are to be found. While the Musacee have only sacs containing Raphides, there occur in the Marantacez and Zingiberacee only other forms of crystals. In the Dicotyledons there are most frequently found clustered crystals, or klinorhombic solitary crystals, or both forms together, often also with granular crystals, while Raphides are entirely absent. In certain cases, however, in these plants also the latter occur exclusively or in preponderating quantity. Finally, a few further examples may be added to the above. For further details the reader must be referred to the authors cited, and for the phenomena in the bast of Dicotyledons to Chap. XIV. Clustered Crystals occur exclusively or greatly preponderate in the foliage of Cheno- podiacex, Caryophyllacez, Gactacex, Lythracex (Gulliver), and very many other families: and, according to Sanio*, in the bast of Juglans regia, Rhus typhinum, Viburnum Oxy- coccos, V. Lantana, Prunus Padus, Punica granatum, Ptelea trifoliata, Ribes nigrum, Lonicera tatarica. Solitary klinorhombic crystals in the foliage of Citrus: in the bast of species of Acer, Poma- ce, of Quillaja Saponaria, Robinia pseudacacia, Virgilia lutea, Melaleuca styphelioides, Ulmus campestris, Guajacum, Berberis vulgaris®, &c. Also in Abies pectinata. Sanio found solitary klinorhombic crystals and clustered crystals together in the bast of Quercus pedunculata, Celtis australis; A’sculus Hippocastanum, Hamamelis virginiana, Morus alba, Salix cinerea, Fagus sylvatica, in species of Populus, Gleditschia triacanthos, Carpinus Betulus, Ostrya virginica, Corylus Avellana, Tilia parvifolia, Spirea opulifolia. Solitary klinorbombic crystals together with clustered crystals, and sacs containing granules, are found in the bast of Betula verrucosa, and Alnus glutinosa (Sanio). Sacs containing granules alone, in Sambucus nigra. Raphides are absent in the examples of Dicotyledons hitherto enumerated. They are numerous and preponderate in the leaves of species of Galium and allied genera, in the + Hegelmaier, Lemnaces, p. 33. * For details of the cortex of woody dicotyledons cf. Sanio, /.c.; there are very numerous details respecting the leaves in Gulliver, Annals and Magazine of Natural History, vol. XI, XII, XIII, XIV, XV, XVI. . ° Hanstein, Milchsaftgefisse, p- 36. * Monatsber. d. Berliner Academie, April, 1857. 5 Compare Sanio, 1 ¢ SACS CONTAINING MUCILAGE, 143 foliage of Vitis, Cissus, and Ampelopsis: in Vitis they are found in the wood also, and, together with klinorhombic crystals, in the bast, in the cortex of Cinnamonum Zeylani- cum, and Olea Europea: in the foliage (leaves and stems) of species of Impatiens, Me- sembryanthemum, and Phytolacca, in Nyctaginex, and CEnotheree. It will be seen from what is above stated that here also differences occur in equivalent parts of plants closely allied to one another and of similar habit, and that one cannot designate any form of crystal-containing sacs as a general peculiarity of a family, or asa phenomenon of adaptation, Among the often-quoted Solanacex the majority (e.g. Solanum tuberosum, Dulcamara, species of Nicotiana, Scopolia atropoides, Jochroma Warczewiczii) have very numerous sacs with granules throughout the parenchyma of the stem (in the leaves often clustered crystals). Jochroma coccineum has granule- containing sacs in large quantity in the parenchyma of the pith, but in the cortex only solitary prismatic crystals: in Atropa Belladonna the granule-containing sacs are entirely wanting in the foliage: it has been above noted that crystal-containing sacs are com- pletely absent in Petunia. All these statements hold for true crystal-containing sacs, and it must always be borne in mind that, besides these, smaller solitary crystals of all forms may occur in the contents or in the membranes of other tissue-elements. The crystal-containing sacs appear while the tissues are still young, usually when the tissues begin to differentiate’; in the leaf of Citrus, according to Pfitzer, their formation begins when it is about 3 ctm. in length: they are developed in greater quantity only in the almost fully unfolded, but still tender leaf, when its cells attain their last definitive extension and thickening of the membranes, They retain unaltered through life that size and structure which they have attained when the differentiation of tissues is complete. In ‘connection with the crystal-containing sacs must be mentioned the occurrence of cells containing Cystoliths, which are found in the Acanthacez, and many Urti- cacez (Pilea), in the epidermis of which they have been described above (p. 105), also scattered in the parenchyma of the cortex, and even of the pith. As regards their structure, all that has above been said of the same structures in the epidermis holds good. 2. Sacs containing mucilage. Sct. 33. When vegetable mucilage and gummy bodies, occurring within the tissues, do not belong to the cell contents of assimilating parenchymatous cells (as is the case with the plentiful mucilage in roots of Borraginacez, e.g. Symphytum, Cynoglossum, &c., or that of mucilaginous sappy parenchyma, e.g. in species of Aloe, comp. p. 116), they fill completely or almost entirely the cavity of special mucilage-containing sacs*. Such sacs occur in the parenchyma of the Malvacez, Tiliacee, Sterculiaceze, in the cortex of the officinal Lauraceze, the Ulmez, the Cac- taceze*, and of the tubers of Orchis, also in the cortex of the firs (Abies pectinata and its allies). They are in all cases distinguished from the cells of the surrounding parenchyma by their greater size, and are distributed between these either singly, or 1 Compare Hilgers, /.c. ? Trécul, l'Institut, 1862, p. 314.—A. B. Frank, Ueber die Anatom. Bedeutung, &c., vegetab. Schleime; Pringsheim’s Jahrb. V. p.161, Taf. XV, XVI.—Idem, Zur Kenntniss d. Pflanzenschleime, Journ. f. pract. Chemie, Bd. 95. 5 Schleiden, Anatomie d, Cacteen, p. 8, where, by the way, the structure is not rightly re- presented. 144 SECRETORY RESERVOIRS. in groups or rows, usually without any clearly recognisable order. In the tubers of Orchis they appear fairly regularly as wide sacs completely filling the meshes of a network, which is composed of plates of starchy parenchymatous cells, one or more layers thick, which face in all directions*, When lying in water they appear as intercellular cavities filled with swollen mucilage, and were described as such by the older anatomists *. If the swelling of the mucilage be prevented, e.g. by treatment with alcohol, the space enclosed by an outer cellulose membrane appears either entirely filled with the firm mass of mucilage, or partially, so as to leave an unimportant central cavity. The mass of mucilage shows in the majority of cases—Malvacee, Cactaceze *, Laura- cex,—the structure of a very thick, abundantly and delicately stratified cell- wall; it often has even pits (Malvacez), and is, as regards its origin and morphological signifi- cance, nothing more than a cell-wall which has thickened strongly at the expense of the internal cavity. According to Trécul’s statements this point may, it is true, be doubted, and new investigations desired. In other cases, and as types of these the tubers of Orchis may be named, the mass of mucilage has no such stratification: it developes from a drop of mucilage, which appears first like a vacuole within the protoplasm, and surrounds a bundle of Raphides lying near to the nucleus: this drop as it grows completely displaces the protoplasm and nucleus, while the bundle of Raphides remains imbedded in the mature mass of mucilage. The sacs in the cortex of the silver-fir appear to correspond to these in structure, their development remains to be more accurately investigated. In later stages of life the sacs often appear swollen up in the living plant to structureless masses (e. g. in Althzea rosea): these then fill up cavities in the pa- renchyma of various form and size, according as they have been derived from the swelling of one or several sacs. These latter structures may best be connected with those mucilage-containing cavities, originally derived from a group of swollen mucilaginous cells, which are described in the parenchyma of the Lime (cortex, leaves, bud-scales*): also the ‘gum’-beating cells, and cavities formed by the swelling of these, which were described by Trécul5 in the parenchyma of the branches of Conocephalous naucleiflorus. This author himself doubts whether the gum- cavities of the species of Quiina described by him ® belong to this category, or to the products of secondary disorganisation. . Subject to the same doubt, the small masses of mucilage scattered through the parenchyma of the stem and leaf of Welwitschia’ may be mentioned here. The sacs with which we are now dealing may be distinguished from those gummy and mucilaginous products: of disorganisation which may be derived in a secondary manner from the most various tissues, by their originating directly from the meristem ; they often appear as its first recognisable product of differentiation, 1 Frank, /.c.— Berg, Atlas z. pharm, Waarenkunde, Taf. 23. * Meyen, Secretionsorgane, pp. 22. 5 Wigand, in Pringsheim’s Jahrb. III. p. 149, Taf. VII. 6. * Frank, Beitr. z, Pflanzenphysiologie, p. 113. 5 Comptes Rendus, tom. LXVI. p. 575 (1868). * Ibid, tom. LXIII. p. 717 (1866). 7 Compare Hooker, Welwitschia, pp. 11, 19. SACS CONTAINING RESINS AND GUM-RESINS. 145 differing from the cells of the surrounding parenchyma in their more rapid growth, and in the absence of any formation of even transitory chlorophyll or starch. It is plain that the cases of subsequent swelling of sacs are to a certain extent connected with those of secondary disorganisation. On the other hand, it cannot be denied that these structures are closely related to mucilaginous epidermal cells (p. 73), and sacs containing raphides (p. 139), and to sclerenchymatous elements, e. g. those of the bark of Punica. 3. Sacs contatning resins and gum-resins. Sect. 34. Sacs, which from the moment of differentiation of tissues are per- manently filled with the above-named bodies, the resin being usually accompanied by ethereal oil, occur as characteristic components of numerous families, or of single genera and species ; in the latter case they usually represent at certain places the ‘intercellular reservoirs, which occur in other parts of the same plant (e. g. Tagetes, Lysimachia), comp. Sect. 50. Taking the extreme cases into account, we may distinguish two forms of these sacs, shor¢ and Jong. ‘The former are of almost iso-diametric and usually roundish foyn, and have thin, smooth, homogeneous membranes which, in the cases hitherto ‘investigated (Laurus, Camphora, Acorus Calamus, Zingiberacez, Canella), give when mature a yellow instead of a blue coloration with iodine and sulphuri¢ acid, and are not destroyed by strong action of the acid’. Protoplasm is apparently absent in the mature sac, which is completely filled by one homogeneous variously-coloured “drop .of resin, or by an aggregation of several of these. Sacs of this category lie solitary, or in small groups in the parenchyma (primary or secondary), with the cells .of which they are strongly contrasted by their highly refractive contents, and are often distinguished by their more considerable size, in the Zingiberacea, Acorus, Piperacez, Lauracez’, Magnoliaceee (Magnolia, Drimys, Liriodendron*,) Canellacee, ‘in the cortex of Croton Eleuteria, and its allies (Cascarilla bark), Galipea officinalis (cortex Angusturze*), and Aristolochiacee. In the majority of the above groups ‘and genera the sacs in question are the only reservoirs of the characteristic secre- tions. But Galipea has also, according to the statements of Engler *, intercellular reservoirs in the primary parenchyma, In the root of Acorus ® Calamus and gramineus, the inner of the two superficial ayers is composed of regular prismatic resin-sacs. Van Tieghem ascribes a similar structure to the roots of Xanthochymus pictorius, and Rheedia lateriflora, in which the intercellular reservoirs present in the stem and leaf are absent, and are replaced ‘by these sacs (comp. below, Sect. 50). It remains to be investigated whether the ‘hypodermal layer of tissue containing drops of oil and resin, which is described in the roots of Valeriana, belongs to this category”. 1 (Compare Zacharias, Bot. Zeitg. 1879, p. 617.] ? Unger, Anatomie und Physiol. p. 210. 3 Treviranus, Beitriage, figs, 34, 35. * Compare the figures in Berg’s Atlas z, Pharm. Waarenkunde, which refer to the above-named barks and other drugs produced from the families above cited. 5 Studien iiber d. Rutaceen, &c., Halle, 1874. ‘ * Van Tieghem, Struct. des Aroidées, Ann. Sci. Nat. 5 sér. tom. VI. p. 175. * Compare Meyen, Secretionsorgane, p. 63, Taf. VI. fig. 22. L 146 SECRETORY RESERVOIRS. In the primary parenchyma of the roots of many Lysimachias and Myrsine also, and in the secondary parenchyma.of many Composite, sacs are found as sub- stitutes for the intercellular reservoirs which are present in other parts of the same plants; their mode of occurrence will be more accurately stated in the subsequent section which deals with these structures. The development of the short resiniferous sacs, and especially the history of development of their contents, is still uncertain, and requires thorough in- vestigation. I have termed the other category Jong sacs because they either permeate the tissues as long tubes, which are simple, i.e. arise from one greatly elongated cell, retaining its original wall, and are arranged singly or in longitudinal series, or they form long series, which follow a similar course, though the single members of these are but little elongated. These two special forms may graduate into one another in a single plant (e.g. in the Convolvulaceze) according to the extension of the members to which they belong. Most of the long sacs here grouped together have been only partially investi- gated, or very unequally in their different relations, it is therefore possible that to a certain extent quite heterogeneous structures stand provisionally side by side ;.gthe features common to them all are sufficiently indicated by the present treatment of them. The contents of the sacs in question, at least when fully developed, usually consist of a milky mass of resinous bodies (in the widest sense) and watery solutions or mucilage. Their distribution varies in special cases; the rows of sacs in species of Allium and the sacs of the Cinchonez permeate the parenchyma alone. But most of these structures accompany the vascular bundles or lie in the secondary bast, being arranged more or less similarly to the laticiferous tubes of various families. In many plants, e. g. certain Aroideze and Musacez (comp. Sect. 47), they occupy exactly the position which is held by the laticiferous tubes in other nearly allied forms, these being absent in the above plants. All this points to a near relationship, both morphological and physiological, with the laticiferous tubes, or at least with certain organs ascribed to this category; many of the sacs in question are hence frequently described as laticiferous tubes. It has at all events been assumed for many of them that they arise, as is the case with articulated laticiferous tubes, by coalescence of longitudinal rows of cells, a view which is generally incorrect. When. however the sacs of this category are arranged in a linear series, e. g. in Convolvulacez, Acer, Allium, it appears that such coalescence may occur here and there for a short distance by perforation of transverse walls, or even of thin pits on the lateral walls; it is however always difficult to distinguish (and I have not been quite certain in any case) whether the perforations obsérved are spontaneous, or were formed in the process of preparation. The organs to be placed in this category are the following: the series of sacs filled with latex in species of Alizum, and perhaps also of Aloe; those to be described below (Sect. 47) in the Arodez, and Musace@ ; the series of sacs described as ‘laticiferous vessels’ in the Convolvulacee : and closely allied to these in structure and arrangement are those of the Sapofacee (Sideroxylon, Bumelia, Isonandra), those which run along the vascular bundles of many Cynaree, and finally those of Sambucus, Cinchona, Ladenbergia, and Acer. It must remain undecided whether the resiniferous SACS CONTAINING RESINS AND GUM-RESINS, 147 elements in the parenchyma of the secondary wood of Coniferze belong to this series. (Comp. Chap. XIV.) As doubtful cases may here be mentioned the sacs filled with pigment, in Sanguinarta, Glaucitum, and Macleya, which will be again noticed in Sect. 47; the sacs or cells containing gum-resin which accompany the vascular bundles of species of Aloe; lastly, the various ‘cells’ filled with peculiar pigments (usually watery solutions), such as those in the roots of Rheum’, Rubia, &c. The following particulars may be quoted concerning those organs of this series which have been more accurately investigated. 1. Hanstein” discovered in those species of Allium in which he sought them (A. Cepa, fistulosum, ascalonicum, &c.), large wide sacs, which he grouped with the series of raphide-containing sacs of other Monocotyledons, as sac-like vessels (vesicular vessels). As regards their form and arrangement they closely resemble these latter sacs, e. g. those of the Amaryllidacee, though they differ from them in structure, and especially in the character of their contents. In the scales of the bulb of species of Leek they appear as numerous opaque lines, which are visible to the naked eye, and run longi- tudinally like nerves. They lie near the outer surface of the scale, between the second and third layers of parenchyma, The single sacs are of circular transverse section, and wider than the cells of the neighbouring parenchyma which abut closely upon them; they are much longer than broad; are often somewhat swollen below their flattened ends, and are arranged in longitudinal series one above another (Fig. 56). At the base of the scale they are often shorter than above, and not unfrequently bear sac-like branches, which connect neighbouring series with one another as transverse or oblique anastomoses: here also rows of sacs occur in close longitudinal ~-,. aggregation. ~ The sacs are filled with a granular cloudy fluid, which appears to the naked eye on the surface of sliced onions as a pale milk, in the sac itself it is cloudy, but still transparent. The nature of the constituents of these contents has not yet been thoroughly investigated: I have not been able —_ FiG.s6.—AlliumCepa; longitudinal section through to substantiate the obvious conjecture that they Sere Ce ee BE cae especially contain the oil of garlic. Raphides or eae a oe wa another and are divided by the pitted transverse wall g—g; they are othér crystals are entirely absent. A large, represented as cut in half longitudinally ; sg their con- = tents coagulated by potash; the longitudinal wall slightly elongated nucleus is still to be found in behind this abuts on another lower sac, and shows ace tak are nee too od, The walk of he Se eee _ sacs are colourless and delicate, so that in sec- tions they are squeezed in laterally by the turgescent parenchymatous cells: where they touch the latter they are smooth or have solitary small round pits: but over the whole surface where two sacs are in contact with one another the wall is, like a coarse sieve, covered with crowded, round—not perforated—pits, while between these lie rather thick bands of membrane. This is the case both with the transverse walls, 1 Unger, Anat. p. 206. 2 Z.¢., compare p. 139 148 SECRETORY RESERVOIRS. and with the less common lateral surfaces of contact of sacs which are contiguous longitudinally. The rows of sacs are continued into the foliage leaves: they have here a similar position and structure to those in the scales of the bulb, but the sacs are much more elongated, and the fluid contents less cloudy. Similar sacs to those of Allium have hitherto been found only in species of Triteleia (Hanstein). 2. The ‘Sap cavities’ in the leaves of officinal and other species of Aloe are of doubtful relationship to the structures under discussion. They accompany the longitudinal vascu- lar bundles in the form of a band of prismatic sacs, which presents in transverse section a semicircular multiseriate appearance: these sacs have flat ends and are arranged one above another in longitudinal rows. The length of one sac varies, according to Trécul’s measurements, e.g. in A. vulgaris, from o'40™™ to 1°30™™, while the width is consider- able, being, in the species above named, as great as o'ro — 0°13 ™™. They are thin-walled, and are filled, according to the species, locality, and time of year, with ‘sap’ of varying intensity of colour, or even with colourless ‘sap’ (Aloe arborescens, plicatilis), which is homogeneous, or it contains suspended in it spherical drops, which vary in number, size, and special structure. It has been asserted that by disorganisation of single sacs, cavities are formed in the band, which contain the same ‘sap.’ It is not improbable that the gum-resin, which is the ‘ Aloe’ of the shops, is derived from these sacs, but even this is not certain. The band is marked off from the surrounding chlorophyll-parenchyma by a layer of apparently prismatic, rather flattened small cells or sacs, which often also con- tain coloured sap, and in these Fltickiger saw, in A. soccotrina, after slow evaporation of the ‘clear, tenacious, beautiful yellow contents,’ distinct yellow plates (of Aloin?) crys- tallise out. For further details about these organs, which still require exact investigation, see Unger, Anat. u. Physiol. p. 206; Fliickiger, Pharmacognosie, p. 106, and the detailed account of Trécul, Compt. rendus, 1 Mai, 1871; Ann. Sc. Nat. 5 série, t. XIV. p. 80, Numerous species of Haworthia and Aloe ciliaris have, according to Trécul, no secretory sacs, , 3. In the stems of Sambucus (S. nigra, S. Ebulus) there occur in the cortex outside the vascular bundles, and especially in the periphery of the pith, longitudinal lines, which turn dark brown on drying, and in this condition have even been regarded as Fungi!. Accord- ing to Dippel’s exposition?, which I find to be confirmed in all essential points, these lines consist of elongated, spindle-shaped sacs of very considerable length and breadth, which are tapered at both ends. The transverse section of these sacs is round, and the breadth varies between 0°025™™ and o'164™™ (Dippel). It is stated by Dippel that the length of the mature sac usually exceeds 18 — 20™™:; the only one isolated by him without injury was 14™™ in length. They seem to me to reach a considerably higher figure, and even to equal the whole length of an internode, that is to attain a length of 20° and more: but it is difficult to decide this for certain, owing to the difficulty of isolating them intact. At all events these lines consisting of sacs, which turn brown on drying, traverse the whole length of the internodes, and even pass through the nodes from one internode into the next. ‘The membrane of the sacs is rather thin and colourless: in older internodes it is thickened and stratified, and has round or oval, non-perforated pits, The contents are, when young, a cloudy, finely granular, rather tenacious mass, which fills the whole cavity. In older stages this mass is often attached, partially or entirely, to the walls, and the central cavity is then filled with an apparently watery fluid: in old parts it assumes an homogeneous, firmly gelatinous character, and ' Compare Oudemans, Over eene bijzondere soort von buizen in den Vlierstam (Sambucus nigra), tot hiertoe voor een fungus (Rhizomorpha parallela Roberge) gehouden. Verslag. k. Acad. von Wetenschappen, Natuurkunde, 2 Reihe, tom, VI. (1872). * Die milchsaftfiihrenden Zellen der Holunderarten. Verhandl. d. Nat. Vereins f. Rheinland u. Westphalen, Jahrg. 22, pp. 1-9, Taf. I (1866). SACS CONTAINING RESINS AND GUM-RESINS. 149 a red-brown colour. The nature of the material composing this mass is not clear (comp. Oudemans, /.c.). According to the reaction with salts of Iron it contains much tannin: it swells in water, alcohol, ether, glycerine, alkalies, acetic acid: it diminishes in volume in mineral acids and salts of metals: the originally colourless mass is coloured by most acids (also by iodine and sulphuric acid), alkalies, and metallic salts (with exception of compounds of iron) a reddish brown, by Schultze’s solution it is coloured blue. Carmine and aniline colouring matters are taken up by it in very large quantity. Each of these peculiar structures originates, as Dippel has shown, from one simple cell, which grows to a great length. The observation of their development in the youngest internodes leads unmistakeably to this view. In the highest internodes of Sambucus nigra tangential longitudinal sections, which include the peripheral zone of pith uninjured, show cells with the above characters of the contents scattered in the parenchyma: these are easily brought into prominence by their deep aniline-staining ; they are scattered through the parenchyma: the longest are almost of equal length to the internode, the shortest are hardly twice as long as broad. In more elongated inter- nodes, up to 5™™ long, the first sacs are already considerably lengthened, while new cells, some placed alongside, some above the first, assume the same characters. Such conditions at first sight allow of the assumption that the sacs arise from rows of coalescing cells, but this is confirmed by no direct observation. In older and quite mature sacs the mass of contents is easily divided, especially after the action of potash}, into cylin- drical pieces sharply limited by transparent bands: these resemble cylindrical cells arranged in longitudinal series: but from evidence derived from direct observation of young stages of development they can only be regarded as products of the action of the reagent. Drawings like Fig. 9 of Oudemans, /.c., represent doubtless something else than the development of the sacs. 4. The sacs containing gum-resin or latex which, according to Karsten, occur in all species of the genera Cinchona and Ladenbergia, appear to be closely allied to those of Sambucus”. They are found, like these, partly in the periphery of the pith, partly in the young outer cortex, close to the bast layer. In many species (C. heterophylla, obtusifolia, &c.) they remain small, and are difficult to recognise even in the cortex when two years old. But in other barks, such as those derived from Cinch. scrobiculata, ovata, umbellulifera, &c., they attain a width of roop to over 300p, in C. lancifolia (?), according to Vogl, even of 700 p, and a length of at least several millimetres. As far as I could see they have conical closed ends: Karsten’s statement that they arise from the coalescence of longitudinal rows of cells certainly requires further investigation. The sacs have a rather thick wall, which shows cellulose-colouring after treatment with potash. Their contents, which include much tannin, are described as milky in the fresh state ; in the dried bark they are so shrivelled that the sacs usually appear empty. 5. A great number of Cynares°, and many Vernoniacea, have in the stem, petiole, and stronger ribs of the leaf on the outer side of the vascular bundles, or of the fibrous band which limits them, a group of sacs filled with a fluid made milky by numerous resinous (?) drops: this exudes on cut surfaces in the form of small white milky drops, and is thus visible to the naked eye. In old sacs the contents coalesce to a very glutinous string. In many species, e.g. Lappa, Cirsium lanceolatum, the sacs are placed not only at the outer, but also at the inner margin of the vascular bundles. The sacs themselves have a spindle-like form; they are closed at both ends, and attain in the mature plant a con- “1 Compare Hanstein, Zc. p. 21. ? Karsten, Die medic. China-Rinden Neu-Granadas, Ges. Beitr. p. 382.—Berg, China-Rinden d. Pharm. Sammlg. zu Berlin.—Idem, Atl. d. Pharm, Waarenkunde.—Vogl, China-Rinden d. Wiener Grosshandels.—Fliickiger, Pharmacognosie, p. 566. 8 Trécul, Des vaisseaux propres .. . des Cynarées laiteuses . .. L’Institut, 1862, p. 266.—Vogl, Ueber Milchsaftgefasse in der Klette; Botan. Zeitg., 1866, p. 193. 150 SECRETORY RESERVOIRS. siderable length, exceeding 3—4™™; they have a moderately thick membrane which shows no important peculiarity. The above sacs may occur in certain species of a genus, and be absent in others. Trécul found them in Cirsium arvense, oleraceum, lanceolatum, anglicum, palustre, przaltum: Carduus nutans, crispus, tenuiflorus, Onopordon acanthium: Carlina vulgaris, longifolia, salicifolia: Jurinea alata, Notobasis syriaca, Tyrimnus leucographus, Galactites tomen- tosa, Durizi, Silybum marianum, Echenais nutans, Arctium lanuginosum; Lappa com- munis :—Vernonia eminens, noveboracensjs, przalta:—but they are absent, according to the same author, in Vernonia flexuosa Sims., and in the Cynarez of the genera Cynara, Rhaponticum, Acroptilon, Serratula, Carduncellus, Centaurea. 6. The secretory sacs of the species of Acer’, which are usually called laticiferous vessels from their milky contents, are of cylindrical prismatic form (on the average about 1™™ long and 50-60 broad in A. platanoides), and are arranged in rows longitu- dinally one upon another. Their colourless cellulose walls are as a rule completely closed, the terminal surfaces, which fit one on another, are horizontal or oblique, the lateral surfaces often have short sac-like protrusions, and with these pit-like thinner-walled pro- trusions they press sometimes between the limiting surfaces of neighbouring parenchy- matous cells, sometimes on the lateral walls of other similar sacs. In surface view these thinner-walled protrusions appear as broad, round or transversely elliptical, clearly marked pits, which are smooth, not latticed. I never saw any perforation of the ends, The open lateral communications described by Hanstein, between neighbouring sacs, by means of perforated lateral protrusions, I was also unable to find in sections which had not been macerated: but they were often found just as represented by Hanstein (/.c., Fig. 6) in macerated preparations, even if these (from A. platanoides) had been prepared only by boiling in water. The contents of two sacs, coagulated by boiling into masses, then hung together by a short bridge, which loosely filled a corresponding canal. How far these conditions exist in the living plant, or have arisen as products of maceration, ie. by rupture of a closed lateral protrusion due to boiling, I must leave undecided. The sacs are solitary or in groups of 2-4, surrounded by parenchyma: they lie on the limit between the phloem of the vascular bundle and the bundle of sclerenchymatous fibres, which surrounds this externally, at that point in fact where in other plants the first-formed sieve-tubes stand (comp. Chap. VIII): they are also found in the primary bark of branches, and in the pétiole and ribs of the leaf. They do not extend further into the parenchyma of the leaf, nor are new ones formed in the secondary bast. Among the species investigated, they are largest and most numerous in A. platanoides. They are developed very early in the internodes, and seem to have special significance during their early stages; still, according to Hartig, they remain filled with sap in the branches of A. platanoides for about ten years. In A. saccharinum and monspessulanum their sap appears, according to Hartig, not to be milky. 7. The peculiar resins of the Convolvulaces? occur in sacs, sometimes as nearly homogeneous masses, but more frequently forming with watery solutions milky mix- tures: these sacs have usually been called ‘laticiferous vessels’ from the latter peculiarity of their, contents. The saes have been observed in all investigated herbaceous species: they are found, according to the species, in stem, roots, ribs of leaves, or only in certain of these parts: they occur especially in the parenchymatous cortex, and in the bast of the stem and roots: they are ranged one above another in rows, which run longitudinally through the members, and are isolated, or numbers are grouped together: the latter is the case, ' Hartig, Naturgesch. d, forstl. Culturpflanzen, p. 545.—Botan. Zeitg. 1862, p. 98.—Hanstein, 2.¢. * Trécul, Des Laticiferes des Convolvulacées; Comptes Rendus, tom. LX. (1865), p. 825.— A. Vogl, Ueber Convolvulus arvensis; Schriften d. Wiener Zool. Bot. Gesellsch, 1863, p. 258.— Idem, Zur Kenntn, d. Milchsaftorgane d. Pfl.; P'ringsheim’s Jahrb. bd. V. p. 31. SACS CONTAINING RESINS AND GUM-RESINS, 151 especially in the tuberous roots of Ipomoea Purga, where they form numerous annular zones}. Each single sac of a series in slightly elongated members, e.g. in the tuberous roots cited, is short, being not longer, or even shorter than broad: in elongated internodes they attain a considerable length, and an extended cylindrical form with flat or slightly curved ends. The contents of the sacs are a mass of resin, mixed to a variable extent with watery fluid, and presenting therefore an appearance which varies in different cases (comp. Trécul, /.c.); in many investigated cases it contains tannin. The walls are thin, homo- geneous, and apparently soft, and show, as far as investigated, no cellulose coloration. With iodine and sulphuric acid they turn yellow: long treatment with the acid does not destroy them. On the fresh surface of section through the sac-bearing parts the contents of the sacs exude as ‘latex,’ the more abundantly the longer and more numerous the sacs are. In fresh plants it often appears that the pressure of the contiguous turgescent parenchyma, which presses the milky fluid out from the cut sacs, can burst also the transverse walls, which are not touched in cutting the section, and press out the contents of more deeply- seated members of a series of sacs at the surface of section. I was unable to prove to myself (after investigations on stems and rhizomes of Convolvulus arvensis, Calystegia sepium, dahurica, Pharbitis hispida) either that there is a perforation or solution of the septa within the living plant, and a formation in this manner of long sacs by the coales- cence of shorter ones, or that there is a genetic connection, as stated by Vogl, between long sacs and sieve-tubes. 8. The reservoirs of the milky secretion in the Sapotacez resemble those of the Con- volvulacez in many points. Since these are but little known, a report by Herr K. Wil- helm upon an investigation of them conducted by him may be here inserted. He investigated especially Bumelia tenax W., and Sideroxylon mastichodendron Jacq., with which, as far as can be concluded from comparison of dried material, Isonandra Gutta coincides in the main. The latex of these plants occurs in completely closed sacs, which are always surrounded by parenchymatous elements, and differ from these fusdamentaily only in their contents, This is literally true for the ixzer cortex, the reservoirs of latex here found have exactly the form and size of the neighbouring parenchymatous cells. In the outer cortex and in the pith the laticiferous elements are usually distinguished from the rest by their con- siderable length and breadth, as well as by their arrangement in uniseriate strings, which run longitudinally through the axis in question, and may be followed to within a short distance of the pusctum vegetationis, The outer cortex and pith are thus traversed by single rows of laticiferous sacs, which are arranged, at least in the youngest parts of the stem, radially and tangentially perpendicular; new elements are constantly added to these from the apical meristem. As the rows pass downwards in the stem, their originally parallel arrangement is disturbed by the increase of the intermediate parenchyma: they suffer tensions and fractures,—their several parts however remain nevertheless in con- nection, and at the same time their character as series of distinct sacs is retained. No single case was observed which necessitated or even supported the assumption that, in the living plant, a coalescence of neighbouring members of tubes had taken place as a typical occurrence. Also in the inner cortex, in the phloem of the vascular ring, it was never possible to prove with certainty a coalescence of parallel laticiferous sacs, or of such as touched at their ends, so as to form extensive reservoirs. In tangential sections the primary ar- rangement shows no regularity: they usually lie scattered and solitary, but sometimes several occur near to, or one above another, between large parenchymatous cells of similar form. In radial sections they appear sometimes to form long longitudinal + Compare Berg, Atlas, Tab, XXIII. 152 SECRETORY RESERVOIRS. strings, But careful investigation and comparison of corresponding transverse sections shows that they never appear in the same radial planes (or only in rare cases, and then only few of them), but in the large majority of cases in diferent planes. As the laticiferous sacs of the bast-ring draw nearer to the outer cortex, they lose the milky nature of their contents: they appear constantly more watery, while the sacs them- selves become gradually more and more compressed, and finally unrecognisable. The above relations of distribution and arrangement of the laticiferous reservoirs hold also for the petiole. In the lamina laticiferous sacs appear as elements or concomitants of the nerves, or are found here and there solitary in the parenchyma, in which case they are always characterised there by considerable size. The membrane of the single laticiferous sacs, whether in the outer or inner cortex, the pith, or the leaf, appears in the large majority of cases of equal thickness throughout. It is as a rule very thin, and equal to that of the neighbouring cells of the parenchyma, or even thinner. However I recognised in many sacs of the inner cortex partial thickenings of the walls: these appear distributed at many points: they appeared swollen, and a slight protuberance of the outer surface of the cell usually corresponded to such points.—The membranes of the laticiferous sacs are colourless; they give a b/ue reaction with Schultze’s . solution, but usually the colour is fainter and less pure than that of the surrounding parenchyma. The contents of the laticiferous sacs have sometimes the character of an emulsion, and accordingly appear white by reflected light, showing under the microscope as a finely . granular dark mass—this is always the case in the reservoirs of the inner cortex; sometimes they form more or less refractive plugs resembling homogeneous masses of resin: these are usually colourless, or light yellow, and fill the cavity of the sac com- pletely. These plugs can easily be isolated from sections under water, when their be- haviour under solvents can be investigated. They occur chiefly in the outer cortex. Carbon disulphide, chloroform, and benzol dissolve the mass almost completely: ether ..Jeaves a considerable granular residue. Alcoholic solution of iodine colours it golden yellow. Addition of alcohol removes the resinous appearance, and makes the mass itself dark and finely granular. If freshly isolated pieces of latex be exposed to concentrated . sulphuric acid, they are dissolved gradually with a yellow colour: dilute sulphuric acid first produces swelling, while homogeneous drops escape from the substance, the outline of which soon becomes indistinct ; their substance then gradually dissolves in the surround- - ing fluid, and colours it yellow. Meanwhile the original contour of the string of latex is retained: it appears that one substance is extracted while the other remains undissolved. The latter was in many cases found still undissolved after long immersion (two days) in sulphuric acid. Potash produced no apparent change. In many laticiferous sacs of the outer cortex dusky-looking contents may be found, con- sisting of numerous drops of the most variable size: they dissolve immediately in water. One is tempted to assume that the before-mentioned resinous contents develop gradually from this latex, which is easily soluble—they occur however in the highest regions of the stem immediately below the punctum vegetationis—nevertheless this phenomenon de- serves attention, viz. that the resinous plugs in the outer cortex assume, after treatment with alcohol, an appearance which coincides remarkably with that of the contents of the _inner cortex. The latter may be completely dissolved by warming with dilute potash: when treated less strongly with this reagent there are sometimes formed from it numerous small crystals or isolated large ones, which disappear quickly on adding acetic acid. The laticiferous reservoirs of the pith were not accurately investigated as regarded their contents, which coincided exactly in optical properties with that of the sacs in the outer cortex}, 4 * [On Resin-sacs in Hypericum, compare v. Héhnel, Bot. Ztg. 1882, 149.] ‘ TANNIN-SACS, 153 4. Tannin-Sacs. Sect. 35. The secretion of Zannzn in large quantity in the sacs of Sambucus suggests the idea of placing other sacs or cells with large quantities of tannin in the category of secretory sacs. It is true the presence of this body in large quantity is not decisive, since it occurs also in other places, as in the epidermal cells, and in many plants, especially ligneous ones, particularly in the assimilating, starch-forming parenchyma, and since, as far as we know at present, it is at least undecided whether it appears as a secondary product of the constructive metastasis, as is the case with calcium oxalate, or as an integrating transitional member of it. Further, with the exception of the tannin, too little is known of the structure, and especially of the character of the contents of the organs, which are possibly to be distinguished as tannin-sacs, for us to be able to decide whether, and when, they are to be regarded as secretory sacs, or only as parenchymatous cells rich in tannin. But there are a number of organs which, as far as may be concluded from information at hand, have apparently lost the properties of cells, and are the points of secretion of mixed substances requiring further investigation, amongst which tannin takes permanently the most prominent place under the reagents at present in use: these organs correspond further in many cases, in their early appearance and position with regard to the vascular bundles, to the secretory sacs of Sambucus, the Cynarez, Aceracez, &c., also to many intercellular, secretory reservoirs, and may therefore be substitutes for these. Awaiting more exact investigation, and excluding such as contain starch as well as tannin, we may here introduce these organs as Zannzn-sacs. They occur as elongated sacs, especially near to the vascular bundles, in the parenchyma of the stem and petiole of many Ferns (Marsilia, Polypodiacez, Cyatheaceze, Marattiacez', &c.). Among the families of Monocotyledons, the Araceze and Musacez should be mentioned as having those rows of sacs, to be described in Chap. VI, which accom- pany the vascular bundles. Also the laticiferous tubes of these plants, consisting of coalesced sacs, would be better placed here than with the rest of the laticiferous tubes in Chap. VI. Of the Dicotyledons certain Leguminose may without doubt be mentioned here. In Phaseolus multiflorus Sachs? found in the phloem of the primary vascular bundles of the stem and leaves (but not continued into the root) longitudinal rows of prismatic tannin-sacs arranged singly or in small groups. They form in transverse section a broken series of curves. A similar arrangement appears in a similar place in the transverse section of the branches of Robinia pseudacacia*. The sacs are here 6-8 times as long as broad, cylindrical, with rounded ends, and only attached to one another by the flattened middle of the terminal surfaces. A group of rather wider series of sacs, with longer members filled with tannin, lies in these trees in 1 Von Mohl, Baumfarne, Verm. Schriften, p. 113.—Martius, Icones pl. Crypt. Brasil. Taf. XXXI and XXXIII. Compare also Karsten, Vegetationsorgane d. Palmen, /.c. p. 205.—Trécul, Comptes Rendus, Mai, 1871, and Ann. Sci. Nat. 5 sér. tom. XII. p. 373-—Russow, Vergl. Untersuchungen, 2 Unters. iiber d. Keimung d. Schminkbohne, Wien (Acad.), 1859. 5 Hartig, Forstliche Culturpfl. p. 546. 154 ‘ SECRETORY RESERVOIRS. the pith, just opposite each of the vascular bundles: besides these, short sacs are found scattered in the pith. Many, but not all Leguminosz are rich in tannin, which is distributed in the tissues in various ways, but in a very constant manner in single species, genera, &c., often without doubt in tissues that do not correspond. The same holds for the Rosiflorze. It remains to be investigated whether also, in many.of the cases of the occurrence of tannin enumerated by Trécul, we have to deal with secretory sacs}, 1 Compare Trécul, Du Tannin dans les Légumineuses; Comptes Rendus, tom. LX. p. 225.— Du Tannin dans les Rosacées, Ibid. p. 1035.—See also Sanio, Bem. iiber den Gerbstoff u. s. Verbrei- tung, &c.; Botan. Zeitg. 1863, p. 17.--Wigand, Ibid. 1862, p,121. CHAPTER IV. TRACHES. Szct. 36. Under the above name all those tissue-elements may be grouped which have the following characters: (1) their walls become thickened as they differentiate from the meristem, and lignified to a variable extent, while the thickening is arranged in a fibrous manner, or with bordered pits, or rarely with transverse bars; (2) almost simultaneously the whole protoplasmic body and organised contents of the cells, which are being transformed, disappear altogether, their place being taken by air, or by clear watery fluid. The larger and more elongated tubes falling under this definition were distinguished even by the old anatomists?, as Trachea, Vessels, Tubes (Vasa, Trachez, Fistule); more recent investigations ? have led to the recognition of two subdivisions, which, according to Sanio, miay be distinguished as (1) Zracheides, and (2) Vessels (Vasa) or Trachea, in the narrower meaning of the word. I shall use the name Trachez in this book only as a collective term for both, and specially also in those cases where it is not certainly decided whether a tube belongs to the one or to the other subdivision. As will be thoroughly discussed in later chapters, the chief points of occurrence of the Tracheze are the vascular bundles and woody bodies. It may however be at once pointed out here by way of explanation that the above localities are by no means the only ones in which Trachez are found. Tracheides are found solitary in the parenchyma (Sect. 55) in many plants, and form the root-sheath characteristic of the aerial roots of Epiphytic orchids *. Vessels and Tracheides correspond in the general points of structure : both alike have very various special forms; transitional forms are to be found between them, which will be described more especially in the secondary wood (Chap. XIV). The distinction between them depends entirely upon the mode of connection of the 1 Malpighi, Grew, Anat. Plant.—Compare Treviranus, Physiol. I. p. 82; Link, Philosoph. Bot. p. go, &c. 2 Sanio, Botan. Zeitg. 1863, p. 113.—Caspary, Monatsbr. d. Perliner Acad. July, 1862,— Caspary terms the organs here called Tracheides, as far as they were the subject of his investigations, ‘ Leitzellen,’ : 2 It need hardly be said that in extending the term Trachez to all tissue-elements which correspond in the above peculiarities of structure, without regard to their place of occurrence, it includes also the well-known air- or water-containing, usually fibrous-thickened elements of the leaves of Sphagnum and Leucobryacez, they being in fact for the most part Tracheides. Compare on these elements of the mosses, Von Mohl, Verm. Schr. p. 294, and Schimper, Monograph. d. Torf- moose. 156 TRACHEE. elements one with another, and upon certain phenomena of structure of the walls, which will be stated in the description of the latter. The walls are, as already indicated, always interruptedly thickened, the thickening masses following the well-known rules for cell-membranes’, and being either pitted, or forming fibrous bands, or both. The form of thickening is either uniform over the whole wall of an element, and even of many contiguous elements, or it varies at different points of one wall-surface, or on different sides of one tube, according to the nature of the neighbouring tissue: these differences are found to be especially frequent in the secondary wood (Chap. XIV). Vessels in which these varieties occur have been called mixed (Vasa mixta)?. That part of the wall of the Trachez which is slightly or not at all thickened is always a very delicate, almost immeasurably thin film. According to the form of the thickening mass there are distinguished— (1) Trachee with fibrous thickening bands, under which head are ranged— (a) Trachee with spiral fibrous thickening (spiral vessels). (8) Zrachee with annular fibrous thickening (annular vessels). (c) Trachee with reticulate fibrous thickening (reticulate vessels). (2) Pitted or dotted Trachea. (3) Zrachee with transverse bars (Tr. trabeculatee). The Trachez with spiral fibrous thickening (included in I. a) were termed true tracheze (Trachées katexochen) by Mirbel and P. de Candolle (Organogr. I), owing to a false conception of the structure of these and other forms, while the annular and reticulate vessels were called false Tracheze (fausses trachées) or striped vessels, Vaisseaux rayés, the latter being moreover confused with pitted vessels. The above forms, especially those with fibrous thickening, often merge into one another, so as to form ‘Vasa mixta.’ The pitted vessels show in many cases, as will be more thoroughly detailed later, protuberances of the inner surface in the form of fibres usually having a spiral course, more rarely in the form of transverse bars, which traverse the cavity, and give the character to the form (3). Sect. 37. In the walls with fibrous thickening the strengthening bands extend inwards from the unthickened membrane, usually as relatively narrow flattened ribands, appearing in section of elliptical or rounded-rectangular, or almost quadratic form: in depth (i.e. perpendicular to the surface of the wall) they are less, or not more strongly developed than in breadth (comp. Figs. 56*, 57). They are frequently very flat, broad plates, in the latter case often broken by short, small slits or de- pressions of the inner surface, e.g. in the spiral or annular tubes of Commelina tuberosa*; rarely they are deeper than they are broad, e.g. the closely-wound fibres of the later developed spiral vessels in the stem of Artanthe elongata and other woody Piperacez, and especially the annular and spiral bands, like sharp fluting, which protrude far into the cavity in the Trachez of the stem of many Cactez ¢, and 1 Hofmeister, Pflanzenzelle, § 25.—Sanio, /.c. * Compare P. Moldenhawer, Beitr. p. 185; Von Mohl, Verm. Schr. pp. 278, 279. * Von Mohl, Ueber den Bau der Ringgefasse, Verm. Schriften, p. 285. * Schleiden, Mém. prés. Acad. St. Petersbourg, sér. VI. tom. IV.—Compare Grundziige I. (3 Aufl.) p. 259—Trécul, Ann, Sci. Nat. 4 sér. tom. IL. pl. 19. TRACHEZ, 157 of the leaves of many Mesembryanthema, e.g. M. stramineum. A less common form of the fibres, corresponding to the bordered pits, is that of which the section has the outline of a short-armed recumbent t~, while the fibre is attached to the thin wall by the free end of the single (here horizontal) arm. This is the case in the closely- wound spiral tubes, which show transitional forms to the reticulate, and, in many woody stems, are first fully formed when the extension of the internode is ended, as is the case in Artanthe elongata, Nerium, Convolvulus Cneorum. The single arm is ; =p: qe) RIS atet Seer ee Eanes ANSE Meee ets seesett GY Pasbieet DAES Baas eae 4 FIG. 56#,—Piece of an annular vessel from the stem of Zea Mais. % the thin wall, on which the limits of the adjoining cells are visible, ~ annular fibres, 7 one of these cut through, 7 the strata of the same (550). From Sachs’ Textbook. FIG. 57.—Saururus cernuus. Piece of a radial longitudinal section through a vascular bundle of the herbaceous stem; ~ in- most, narrow, distorted annular-vessel. To the left of this successively (z) spiral vessel, with loosely wound single fibre, which in two places runs back into itself, so as to form a ring; the thin wall between the spirals of the fibre has given way; (2) spiral vessel with very narrow flattened curve, cut longitudinally in half, with exception of the upper margin; (3) scalariform reticulate vessel ; /selerenchyma (or bast) fibres. The curves of the spiral fibres rise in the drawing in the opposite direction to the course they really pursue (375). in these cases almost always smaller than the two others; in the above-named Artanthe it is very inconspicuous compared with the other strongly-protruding parts. The Zrachee with spiral fibrous thickening show considerable variety in the number of the fibres, and the steepness and direction of their coils. Their number is often only 1-2 in the narrow tubes, which are first formed when the differentiation of tissues begins, in others 4 or more, and it rises in many cases, e.g. the petiole of Musa, to 16-20. The steepness of the coils is greatest in those tubes which are developed earliest, before the extension of the part to which they belong has ceased: since in these the coils are separated from one another by the | 158 TRACHEE. extension which the tube itself undergoes’. By this process a spontaneous separa- tion (tearing off?) of the fibre from the elongating wall may occur*. If the tube developes later, during or after the close of the extension of the part, the coils are less steep: when several fibres are present they are then arranged at the minimum distance from one another. The coils rise in most cases (when seen from without) from right to left, that is like the thread of a left-handed screw, or, according to the terminology adopted in Botany, the spiral is right-handed. The opposite direction occurs in Pinus sylvestris (Mohl): in the wood of Vitis vinifera, Berberis vulgaris, Artemisia Abro- tanum, Bignonia capreolata, the inmost first-formed tubes are right-handed, the outer later-formed ones left-handed. Where the spiral fibre is interrupted, both the opposite directions of inclination may occur at different heights in one vessel, e. g. the stem of Cucurbita *. Not unfrequently, and especially in the closely-wound forms, the spiral fibres are branched, or their coils are connected in a bridge-like manner by oblique or transverse fibrous bands. It is a no less common phenomenon that one fibre at the end of a tube, or at other points, should run back into itself, thus forming a ring. This phenomenon characterises the series of numerous transitional forms between the spiral tubes and the annular and reticulate tubes, while it gives rise in the latter to a number of special forms of the net. It should be added, in connection with the annular tubes, that the distance of the rings from one another follows the same rules as the steepness of the inclination of the spiral fibres. Besides those above mentioned, there is among the reticulated tubes a varied series of special conformations of the net. Reticulated tubes, the meshes of which are elongated transversely, and arranged on one surface of the wall in a row one above another, being thus comparable to the rungs of a ladder, are called ladder-like or scalartform vessels, and have frequently been con- founded with the pitted vessels, which show a like surface of wall. (Comp. Fig. 56*.) Individual peculiarities of the tracheides in the sheath of the roots of Orchidaceze will be brought forward again in Sect. 56. ' Sxct. 38. It is well known from the general doctrine of cells, that it is only the relative size of the unequally thickened parts of the membrane which gives rise to.a general distinction between reticularly thickened and pzHfed membranes, and that therefore there is no sharp limit between these two forms. The wall of the puted trachecdes shows sometimes szmple pits, i.e. not bordered, sometimes bordered pits*. The term pit is applied to a gap in the internal thickening of. the wall, this gap being closed externally by a piece of membrane which is only slightly or not at all thickened. It is in fact a canal varying in length according to the extent of the 1 Von Mohl, Veget. Zelle, p. 26. 2 Compare Sachs, Textbook, and Engl. Ed. p.go. With this must not ibe confused that ‘unrolling’ of spiral fibres, which occurs when a part is torn, and the often-cited ‘ power of unrolling’ of spiral vessels, The latter phenomenon comes about simply, ‘by the rupture of the delicate unthickened membrane when the part is torn, while the tough fibre, to which the delicate and easily overlooked tatters of the ruptured wall are attached, is drawn out. 3 Von Mohl, Verm. Schriften, pp. 287, 321.—Sanio, /.c. p. 124. * (Cf. Russow, Zur Kenntniss des Holzes, insonderheit des Coniferenholzes. Orig, communica- tion to Bot. Centralblatt, Nos. 1-s, 1883, in which paper the other principal writings on this subject are referred to.] a BORDERED PITS, ‘ 159 thickening, which traverses the wall transversely. If the canal be equally wide throughout or narrowing outwards, we have a non-bordered pit. On the other hand, the term bordered pit is applied to those in which the canal widens suddenly towards the outside, i.e. towards the non-thickened part of the membrane, so that it is here broader than at the part of the canal bordering on the internal cavity. In the surface view of the wall the boundary of the unthickened portion of the membrane — IAMATEIS FIG. 58.—Pinus sylvestris; radial longitudinal section through FiG. s9.—Ephedra helvetica. Wood (230) ; the wood of a branch; a—e ends of tracheides with bordered a@ member of a vessel, 4 tracheides seen from Pits (¢’, 2” in surface view; c & piece of a young wall of a the radial side, isolated by maceration with tracheide, with still immature bordered pits; further develop- Schultze’s mixture ;_/the oblique ends of the ment. of these, successive narrowing of the canal a—c; d@-and e member of the vessel in surface view, with mature condition; s¢ large pits on the limiting surface be- two rows of large open bordered pits; at x,* tween tracheides and cells of medullary rays (550). From Sachs!’ two closed bordered pits. Of the tracheides, Textbook. is only drawn in outline, the surface of the other is putin. (The direction of the split- like pits is reversed ; they rise in reality from left to right.) may be seen surrounding the limit of the section of the canal like a border or halo (Fig. 58). The widened outer part of the pit, the limit of surface of which is the halo, is called the cavity of the pit; in the canal itself may be seen the outer aperture, which leads into the cavity of the pit, and the zzzer which borders on 160 : -TRACHEE, the lumen (comp. Figs. 59-60). The cavity of the pit is in most cases originally and often permanently of the form of a plano-convex Jens (‘ half-lens-shaped ’), since the outer surface of the thickening of the membrane, which borders it on one side, is concave, while the unthickened portion of the membrane, which limits it on the other, is flat. The canal is, according to the extent of the thickening of the membrane, either extremely short, so that a sharp-edged opening leads from the lumen of the tube into the cavity of the pit, e.g. in the thin-walled tracheides of the spring wood of Pinus; or, when the thickening is greater, it is elongated, and widens outwards suddenly into the cavity of the pit, e.g. autumn wood of Pinus, pitted vessels of Nerium, Fraxinus, wood-eléments of Convolvulus Cneorum, Pteris aquilina (Figs. 61, 64), &c. The above general description of the structure of the bordered pit is said to hold for those uncommon bordered pits, not belonging to our present subject, which occur in certain ce//s}, and for those on the limiting surfaces between Trachexw and other elements; and.it is clear that between these and the non-bordered pits only the above-described difference of form exists, which corresponds exactly to that between the flattened and the / shaped fibrous thickenings. .On the surfaces abutting on elements of another order, the bordered pits of the Trachee either correspond to non-bordered pits on the walls of these, or they are opposite to an unpitted wall. But where Trachez with bordered pits are contiguous with one another, the bordered pits correspond to one another in such a way, that on each limiting surface all the cavities of the pits of the one fit exactly over those of the other. The plano-convex cavities are thus applied to one another in pairs, so as to form the ‘lens-shaped pit-cavities ’ (comp. Figs. 58-62), and each of these is divided by a thin flat lamella of membrane (the limiting lamella) into. two halves. This is the case in the first instance in all investigated cases. Also in mature Trachee this condition always remains permanent, FG Sor Tranevere secon brow te seconirywerd scan easily be proved in old wood of the section having passed through the oblique wall separat- Pinus, Ephedra (Fig. 60 5) 2 As a rule ing two members, and in fact through the middle of an open bordered pit (pore), and to the left of this, through the however the originally plane limiting lamella margin of a pit. Besides this vessels and tracheides are transversely cut at 2 and 4, through the middle of bordered grows in surface in such a way, that it be- pits of the lateral walls, which have the limiting lamella thickened in a knob-like manner on both sides; at ¢ the comes larger than the median plane of the thickening is on one side. : : lens-shaped double cavity, and therefore bulges in a convex manner to one side, and comes into close contact on this side with one of the concave walls of the cavity of the pit (Fig. 60 c); meanwhile it remains a very delicate film, but is-always, in the cases investigated, thicker in the middle than at its margin. In Pinus sylvestris (and its allies), as first shown by Sanio, the thicker part has the form of a relatively broad plate with a sharply marked * Compare the figures of the endosperm of Phytelephas (?). Schleiden, Grundziige, 3 Aufl. p. 232. ? Compare Hofmeister, Pflanzenzelle, p.175. , BORDERED PITS. 161 margin; in Ephedra, which is the most striking instance I know of, it is shaped like a flat biconvex lens ; in other relatively small, or at least narrow pits (Cassyta pani- culata, vessels of Nerium, Pteris aquilina, &c.), it appears as a hardly perceptible swelling. ‘The thicker portion always lies, like a lid, upon the outer aperture of one of the pit-canals. The corresponding bordered pits of neighbouring Tracheze are accordingly closed by the limiting lamella, which occasionally remains plane, but as a rule is applied to one wall of the pit-cavity. On account of its delicacy, and the small size of the whole pit, the limiting lamella, in the form in which it usually occurs, has not hitherto been clearly recognised. In contradiction to Hartig+ alone, and following the statements of Schacht and Dippel?, the pit-cavity was regarded as being in the mature state in open communication on both sides with the adjoining cavities of the tubes, while the few cases in which the limiting lamella was observed were considered as exceptions. Sanio® has recently clearly proved that in Pinus sylvestris the case is as stated above. I find his statements confirmed in all cases subjected to exact investigation, both in the tracheides of the wood of that tree, and in those of Abies pectinata, excelsa, Juniperus communis; also in the tracheides, and lateral walls of the Trachez of Ephedra, and Welwitschia: further in the lateral walls of the ‘scalari- form vessels’ of Ferns (Pteris aquilina); the tracheides of the secondary wood of Draceena, Cordyline paniculata; the Trachez of the wood of Convolvulus Cneorum, Statice monopetala, the large-pitted vessels of the wood of Cassyta (C. paniculata, R. Br.), Nerium Oleander, &c. Extremely good preparations, which are not always easy to obtain, always show the case as described: it should then be characterised at least as the regular condition, and that which is distributed over the most various divisions of the vegetable kingdom. Further investigations must show whether ex- ceptions occur. While the above fundamental conditions of structure remain constant, the special form of the bordered pit is very variable (compare the Figures 58-62, and what follows in Chaps. VIII and XIV); firstly, according to the length of the canal, which depends upon the extent of thickening of the walls; this has already been mentioned above; secondly, according to the special form of the pit, and of the canal with its outer and inner aperture, as seen most distinctly in surface view of the wall, and according to the relative size of the diameter of these parts in each pit. All these parts have the forms generally characteristic of pits, which (in surface view of the membrane) vary in individual cases between a circle and a narrow slit. In the same pit all the parts are alike in form, or very similar, as is the case in the circular pits of the tracheides of the wood of Pinus, and the slit-like ones of most scalariform vessels, thus giving rise to the appearance of the bordered pit in surface view as two or three concentric outlines, which differ only in size (e.g. Fig. 58, 61 B). On the other hand, the form of the parts may differ in the same pit, either so that they all differ from one another, or one from the rest, and this may occur in all possible . combinations. ‘Thus there is a narrow elliptical inner aperture, and a circular outer aperture of the canal, which diminishes greatly towards the outside, with irregularly ? Compare especially Botan. Zeitg. 1863, p. 293. ; 2 Schacht, De maculis (pits), &c. Programm. Bonn, 1860.—Dippel, Botan. Zeitg. 1860, p. 329. ® Pringsheim’s Jahrb. Bd. IX. [See also Sachs, in Arbeiten des Bot. Inst. in Wiirzburg, IT. p. 294.] M 162 TRACHEZ, circular outline of the relatively very large cavity in species of Cassyta: a narrow slit- like inner aperture, and a very small circular outer aperture of the canal, with a broadly elliptical cavity in Eleagnus acuminata: a long and narrow slit-like inner aperture, and a short slit-like outer aperture, with a circular cavity in Aleurites triloba—all these cases occur on the large-pitted vessels of the wood’. As regards the relative diameter of the different parts of a pit it is obvious, from what has been premised, that it is always larger for the pit-cavity than for the outer aperture of the canal. The latter is either as large as the inner aperture, or it is smaller than this, and the canal thus becomes narrower outwards to \ 9? ool 000 0000 0000 00 qe “oN 00000000%0 Ogg DU) FIG. 6r.—Pteris aquilina Rhizome; 4 (142) end, about 4 of a short member of a vessel ; the oblique ladder-like end surfaces“ and a part of the lateral wall in surface view; Ba piece of A at x, magnified 375 times; C (375) thin longitudinal section through part of a lateral wall, where two vessels touch one another; D (375) a similar section through the oblique wall /and its margin adjoining the lateral wall. At/the pits are open. a varying extent, and with a form corresponding to the above description : the inner aperture, when of a slit-like form, and differing from the cavity, is always narrower, but often longer than the greatest diameter of the latter. Slit-like bordered pits placed close to one another may thus coalesce internally, in numbers from 2-6, into a common slit, as Mohl found (/.¢. Figs. 6, 10, 15) in Aleurites and Eleeagnus, and 1 See Von Mohl, Ueber den Bau der getiipfelten Gefassse, Linnzea, 1842; Verm. Schriften, p. 272, Taf. XII. BORDERED PITS, ETC, 163 Sanio (2. ¢. 125) in the wood of Tectona grandis, Fraxinus, Tamarix, &c. This must originate in the thickening of the membrane lasting longer at the inner side than at the outer, and altering its original direction at a later period. On the vessels of the wood of Mahonia aquifolium Sanio found round bordered pits, arranged in left- handed oblique series, with the inner apertures serially coalescent into long slits, while between these the thickening of the walls protruded inwards in form of spiral bands. . The arrangement of the bordered pits on a wall-surface differs in no way from the known rules for the arrangement of pits generally. They are arranged on a surface in perpendicular, horizontal, or, especially when of slit-like form, in oblique spiral series: in the latter case the spirals are almost always left-handed: the number of these series varies according to the special cases, and on equal areas it varies on the whole inversely with the size of the pits. We may cite as examples of extreme cases, on the one hand, the usually loose series of large round-bordered pits on each radial face of the tracheides in the wood of Pinus, and the several loose series of large pits on the wide vessels of the wood of Cassyta (Mohl, 2c. Fig. 1); on the other hand the walls covered with close and small pits, which are found surrounding the large vessels in the vascular bundles of the stem of the Cucumber, the tubers of Dahlia?, many Dicotyledonous woods, as Quercus, Nerium, &c., &c.; in these cases the margins of the pit-cavities are separated from one another by quite narrow bands or ridges of the wall. A special case of frequent occurrence may here be mentioned, viz. the trans- verse, slit-like bordered pits, which are characteristic of almost all Ferns (Fig. 61), and which also appear in many Dicotyledonous woods, as Cheilanthus arboreus, Vitis, &c., these being arranged like the rungs of a ladder, in one or few longitudinal perpendicular rows on a single wall-surface. The wall-surfaces on which they occur may be called scalariform or ladder-like surfaces, whilé these Trachee, together with the similar reticulate tubes above mentioned’, have been termed scalariform or ladder-like vessels, Vasa scalariformia, also scalariform tubes. It is useful to dis- tinguish them from the reticulate vessels with non-bordered, transverse pits, either as bordered scalariform surfaces or Traches, or to reserve specially for them the name of ladder-like or scalariform surfaces. Sect. 39. As one of the above-described forms of wall-thickening there are found, in some few cases to be cited below, ingrowths of the thickened portions of the membrane, which protrude in a conical or bar-like form into the cell-cavity, or are stretched transversely across it: and those Trachez in which these are largely deve- loped may be distinguished by the name, introduced above as No. 3 (p. 156), viz. Trachee with transverse bars. The bars are very largely developed in the narrow primary tracheides occupying the corners of the vascular bundles of the stems of the stronger species of Lycopodium, and in the margin of the vascular bundles of the leaves of Juniperus* (comp. Chap. VIII). They are here somewhat flattened cylindrical fibres, branching irregularly on all sides, and with the branches connected on the one hand one with another to form a net spread through the cavity, and 1 Compare Sachs, Textbook, 2nd English Ed., p. 26. ? Von Mohl, 2. ¢. ® Compare Link, Elem. Phil. Botan. Ed. 1, p. 95; Von Mohl, Veget. Zelle, p. 27; Unger, Anatomie und Physiologie, p. 172. * Compare Von Mohl, Botan. Zeitg. 1871, p. 12. M 2 I 64 TRACHEZ., attached on the other to the, thickened lateral wall of the tracheide. In the tracheides of the leaves of Juniperus (Fig. 62), their points of attachment and origin are especially the thick swollen margins of the bordered pits, in the Lycopodia the spiral or reticulate bands with which the lateral wall is thickened. At the corners of the vascular bundles of the leaves of Biota orientalis the swollen margin of the bordered pits is often elongated into blunt cones, which protrude into the. cavity, but here end blind, without branching or coalescing with one another. or with the opposite wall. As a rare and anomalous phenomenon Sanio* found single simple bars, stretched transversely from one wall-surface to the opposite one, in single tracheides of the wood of Hippophae rhamnoides, and Pinus sylvestris: in the latter they are stretched between the tangential walls, and where they occur at all they traverse in the same direction the entire length of long radial series of tracheides, as far as the cambial zone. (Comp. Chap. XIV.) * FIG. 62.—Juniperus communis; leaf, FIG, 63.—The same (22s) ; vascular bundle, g xylem, ¢ single sclerenchymatous fibre transverse section (600), 2 parenchy- at the outer limit of the phloem; ¢ margin consisting of tracheides with bordered pits matous cell; next to it tracheides of the and transverse bars, The parenchymatous cells near and between the latter are corner of the vascular bundle, with bor- shaded with dots, dered pits, and reticulately-branched transverse bars. The parts lying below the surface of section focussed are shaded. . Further details on the structure of the walls of the Tracheze will be described in later chapters, especially the VIIIth and XIVth. Sect. 40. The wall of the Trachez, whatever be its structure, is in the one series of cases a completely closed membrane (Zrachetdes) ; or it is broken through at the limiting surfaces between elements placed serially one above another, and originally completely closed, the series thus coalescing to a continuous tube, which is called a vessel. ‘The tracheides therefore differ from the vessels only in the absence of the holes which occur in certain of the walls, and thus connect the cell-cavities. Tran- sitions between them occur in the secondary wood of ‘Dicotyledonous plants (comp. ‘ Chap. XIV), e.g. Leguminose, inasmuch as with otherwise similar characters the holes are absent in one case and present in another. Holes are also to be found in * Botan. Zeitg. 1863, p, 117.—Pringsheim’s Jahrb, Bd, IX. p. 59. TRACHEIDES. VESSELS. 165 the elements of the root-sheath of many Orchids (Sect. 56), but these had better be termed generally tracheides, since the connection in rows which is characteristic of vessels is absent. : The Zracherdes are in some few definite cases—-ends of vascular bundles, trans- fusion tissue, the root-sheath of Orchids—short, even iso-diametric sacs: as a rule they are of the form of elongated, spindle-shaped, fibrous cells, pointed at the ends, and with round or polygonal transverse section. They usually remain microscopically small, their length, which is a large multiple of their breadth, reaches 0-16m™™ to about 1-00™™m; this is the case in the wood of most Dicotyledons': or it rises to 4mm, as in the later annual rings of Pinus?: in many cases however they attain great dimensions: the large spindle-shaped spiral and annular tubes in the stem and petiole of Musa and Canna® attain a width of 0-08 to o-10mm, and are always more than 1°™ in length; the spiral tubes of Nelumbium speciosum have, according to Caspary, a length of over r2°m, and width of o-567™™, The great majority of Trachez belong to the category of tracheides: for instance, the tracheal elements of all peripheral ends and expansions of vascular bundles, of the secondary wood of the Coniferze, Cycadeze, most elements of the secondary wood of woody Dicoty- ledons, almost all Trachez of the Ferns, in the widest sense—vessels are only known to occur in Pteris aquilina, and in the root of Athyrium filix femina ‘—the Tracheze of the vascular bundles in stem and leaf of the Cycadez and Coniferz°, of many, though far from all Monocotyledons, and numerous Dicotyledons*. Many even of the most striking elements with fibrous thickening, usually described as vessels, belong to this series. To the already-cited examples of Canna, Musa, and Nelum- bium we may, according to Caspary’s work quoted above, and referring to this for further details, add the following examples: the ‘ vessels’ in the vascular bundles of Stratiotes aloides (stem), Caladium nympheifolium, Pistia Stratiotes, Acropera Loddi- gesii, Aerides odorata, Alisma Plantago, Sagittaria sagitteefolia, Hydrocleis Humboldtii, Musa spec. (vessels in this case in the root), Brasenia peltata, Nuphar luteum, pumilum, Nymphzea alba, gigantea, Victoria regia, Monotropa Hypopitys. A general view of the occurrence of tracheides and of true Trachez will only be possible when the necessary arduous investigations have been extended over a larger number of cases than has hitherto been the case. Sect. 41. A vessel arises from a series of originally-separate cells, placed one above another, by the perforation, at the close of the process of thickening, of the division walls between the members of the series, the latter being then termed the members of the vessel. The rows of cells, above indicated (pp. 9 and 11, in Figs. 2 and 4) by the letter v, which extend to the apex of the plerome, and similar ones marked g in Fig. 3, p. 10, are rudiments of vessels. The members are also always easily distinguishable in a mature vessel, and are 1 Sanio, Botan. Zeitg. 1863, p. 114. % Sanio, in Pringsheim’s Jahrb. VIII. p. 4o1, &c. 3 Compare Unger, Anat. und Physiol. p. 171, and p. 218, Fig. 92 4. * Russow, Vergl. Untersuchungen, p. 103. 5 Mettenius, Beitr. zur Anat. d. Cycadeen, p.258. [See also V. Héhnel, Ueber das Vorkommen von Gefassartig-zusammenhingenden Tracheiden-strangen in Coniferen-hélzern, Bot. Ztg. 1879, p. 329-] ® Caspary, Monatsbr. d, Berl. Acad., Juli, 1862. . 166 TRACHEE. separable from one another, their limits being marked by their margins, which are always permanent, often also by other portions of the perforated dividing wall: these portions have the structure of a thickened double cell-membrane, and consist of two thickening plates and a simple limiting lamella between them. Schultze’s mixture, or hot solution of potash, destroys the limiting lamella, and thus separates the mem- bers from one another. ' The form of a member of a vessel is as a rule cylindrical or prismatic, the breadth being throughout almost uniform, or diminishing quite gradually towards one end: more rarely each member widens in the middle to a barrel-shape. The length of a member is usually greater than the diameter: it is very much greater in the vessels with a loose spiral, or in annular vessels, which develop before the extension of a part is complete, and thus grow greatly in length as the part elongates. The members of such vessels as arise after the extension of a portion of a stem or root is complete are composed of short members, these being barely longer or even shorter than they are broad, e. g. the wide-pitted and reticulate vessels of old stems of Cucurbita, Cobza, Vitis, &c. (comp. Chap. XIV). Successive members of a vessel are as a rule of almost similar form through long tracts, though they often decrease gradually in width. The general form of the vessel may be concluded from these data: such as are composed of short barrel-shaped members were dis: tinguished by the old authors as rosary-shaped: Vasa moniliformia. The walls, by which the members of the vessel are in contact with one another, are either horizontal, in which case those of the successive members fit exactly on one another, and together form the septum (comp. e. g. Fig. 3,), or they are more or less oblique, and the inclined faces of successive members here also fit exactly throughout so as to form the oblique septum (Figs. 59-61): or the ends are oblique and pointed, and only a part of the opposed faces of successive members is united so as to form a septum, near and above which the pointed end forms a blind and often irregularly-formed continuation. The perforation of the septum is always brought about thus: on the delicate primary membrane one or several flat large pits are formed by the typical pro- cess of thickening; the unthickened parts of the membrane are then at once dissolved and disappear, while the thickened bands of membrane, connected directly with the thickenings of the lateral walls, remain persistent. In almost all cases on horizontal septa, and not uncommonly on oblique ones, there appears one single pit, or one single round or elliptical opening, which then always occupies the greater part of the surface of the septum, and often, especially in thin-walled vessels, the whole surface with exception of a very narrow peripheral band. On the other hand, strongly-inclined septa, and very rarely horizontal ones (Avicen- nia), retain in most cases several or many openings included within the thickened margin, and separated from one another by thickened bands. These are in some few cases round, e.g. in the Trachez of Ephedra! (Fig. 59), usually they have the form of slits of varying breadth, and arranged parallel in rows, whence the expres- sion ladder-like perforated septa (Fig. 61). The slits are usually almost at right angles to the longitudinal axis of the vessel, and the series of them are similarly ? Von Mohl, Ueber den Bau d. grossen getiipfelten Gefasse von Ephedra; Verm. Schr. p. 268. VESSELS, 167 arranged; thus the narrow, closely-grouped transverse slits in the ladder-like oblique walls of the pitted vessels in the wood of the Betulacez, Ericacez, of Corylus, Carpinus, Pteris aquilina; the round openings of the oblique walls of Ephedra arranged in 1-2-3 rows, &c. Rarely the slits are parallel to the longitudinal axis of the vessel: e.g. vessels of Hieracium vulgatum, Onopordon Acanthium, in which irregular reticulate openings also occur. In an Avicennia Sanio found the hori- zontal septum surrounded by a sharply-marked thickened margin, and the whole remaining surface covered with many irregular, round or slit-shaped, bordered openings’. There is no constant relation between the form of thickening of the lateral wall and the form of perforation. Nevertheless most vessels with fibrous- thickening have round openings, and very many vessels with bordered pits have ladder-like perforations. In pitted vessels, however, simple openings are also frequent, and Sanio found ladder-like perforations in the spiral vessels of species of Casuarina, Olea europza, and Vitis. In thin-walled vessels—such as most of those with fibrous thickening, and thin-walled pitted vessels, e. g. in the wood of Betulacee and of Tilia—the margin of the opening of the septum is smooth and thin, corresponding to the margin of very flat, not bor- dered ‘pits. In thick-walled vessels it is thicker, and has the structure of a pair of corresponding bordered pits opened by disappearance of the limiting lamella, and with but small difference of width between the pit- cavity and the wide orifice of the pit: it consists therefore of two acutely-diverging lamelle. In many vessels this structure as of a bor- dered pit is extremely striking, e. g. in the large solitary openings of the pitted vessels in the wood of Nerium, Fraxinus, Convolvulus Cneorum (Fig. 64), Pirus torminalis ®, in the serially- arranged round openings of the ves- sels of Ephedra (Fig. 60, g), and the small slits of the scalariform yes- Zi, Stc:Cowolvalis Cuetrum, moat, te ee rat ae sels of Pteris~ aquilina (Fig. 61). It oes At ¢the very delicate closing membrane of the bordered pit also occurs very plainly in the thick and closely-wound spiral vessels in the stem of Nerium. In other cases, even when the margin of the opening is very thick, it is often only slightly indicated by a small indentation running over the limiting lamella of the margin: e.g. the pitted vessels of Cucurbita, Juglans, Acer monspessulanum (Dippel, /.c.). The history of develop- ment shows, in the cases of the latter category, that the opening arises by the 1 Sanio, Botan. Zeitg. 1863, p. 121.—Von Mohl, Ze. 2 Compare Dippel, Botan. Zeitg. 1860, p. 329. 168 . TRACHEE, disappearance of the limiting lamella in the face of a young pair of wide corre- sponding bordered pits. Since the structure of the bordered edges of the apertures corresponds to that of closed bordered pits, the practical decision whether in a given case an open or a closed pit is present is necessarily extremely difficult, if the parts of the septum in question be small, and resemble closely the bordered pits of the lateral walls in form and size; and the difficully is the greater, since in such cases intermediate forms between the open and closed bordered pits—which are never exactly alike— occur at the limits between the septum and the lateral walls. The apertures at the middle of the septum in the pitted vessels of Ephedra (e.g. Fig. 59, 7) are ° larger than the closed bordered pits of the lateral walls. But at the margin of the septum there occur not unfrequently round bordered pits (Fig. 59, x), which, though resembling the apertures in’ form and size, are closed like the pits of the lateral wall. In the large scalariform vessels in the rhizome of Pteris aquilina (comp. Fig. 61) the transverse, bordered, slit-like pits of the lateral walls are always closed. The strongly-inclined septa or terminal faces of the obliquely pointed members of the vessels show a quite similar ladder-like series of slit-shaped pits, like those of the lateral walls, but with the difference that here the silit-like pits are wider, and the thickened intervening bands which separate them are thinner than on the lateral walls. In the middle of the septum the slits are open, as is plainly shown in sections through vascular bundles which have been injected with glue and then dried, if the sections be then treated with water and the glue dis- solved. The bands between the slits then separate from one another (Fig. 61, D, /). Towards the ends of the intervening wall the pits are, on the other hand, of the same width as the open ones, but closed by a limiting lamella: on the corners of the lateral walls they become gradually narrower, and like those of the lateral walls. The vessels are not unfrequently branched, two or more series of members being attached laterally to one member. When such branches run parallel or converge, they may attach themselves again laterally to a single series of members; a vessel may thus be in its longitudinal course alternately single and double?. As regards the adsolute size of the vessels, there is nothing to oppose the view that their ength may equal that of the whole plant, or at least may be very great. At all events, on following the vascular bundles through long distances, member is found attached to member, while blind-ends are rare, except in the ends of the peripheral expansions of the plant. The width of the vessels is extremely unequal, and changes variously according to the point of their occurrence in a given plant, and according to the single species or genus. It may be said generally that the diameter does not on the average exceed that of narrow fibrous cells in those vessels which appear first in stems and roots, before the extension is complete (spiral and annular), and in those which traverse the nervation of the leaf. Those formed in stems and roots at the end of the process of extension, or subsequently, may in many cases attain much greater width, while this does not prevent others of the smallest calibre occurring with or near the former. For examples of this see Chap. XIV. Vessels of greatest width occur in the central part of the vascular bundles of many * Von Mohl, Palm. Structura; Verm, Schriften, p. 142. TRACHEZ, I 69 Palm stems* (compare Chap, VIII), where they attain a diameter of o-28omm (Mauritia armata) to 0-562™™m (Calamus Draco); in the wood of many climbing and twining plants, e.g. Cucurbita, Cobwa, Phytocrene?, Ampelidez, in which also the width may rise to 0-3 — o-5mm, &c.3 The vessels of greatest width are always pitted vessels with short members. After what has above been said on longitudinal course and branching, it need hardly be noticed further that in one and the same vessel the width (and with it the form of thickening of the walls) may often change in successive parts of its course, i, e. in its successive members; for instance, Mohl states the diameter of the above vessels of the Palm-stems at the lower ends of the bundles as o-orrmm, As regards the material composing the walls of Trachex and Tracheides, it is certain that they are, when first formed, cellulose membranes, and that they consist, when mature, of more or less lignified cellulose. The lignification is present to a very varying extent according to the special case; in hard, firm parts more than in soft, sappy parts; the Trachez of delicate foliage-leaves, or of sappy stems, &c., often show an almost pure cellulose reaction, A very remarkable phenomenon, to which Burgerstein has recently again drawn attention, is the surprisingly early appearance of lignification in many vessels, It is beyond the scope of this work to enter minutely into the process of lignification: it cannot at present be exactly stated how far it shows peculiarities in the several organs in question. Reference may therefore only be made here to works upon the subject: the summary of the older results in Hofmeister, Pflanzenzelle, Sect. 30, Kabsch, Pringsheim’s Jahrb. III, and the newest investigation of A, Burgerstein, Sitzungsber. d. Wiener. Acad. Bd. 70, July, 1874. Sect. 42. All Trachez are alike in the peculiarity that when they are fully formed the protoplasmic body disappears entirely, without leaving any vestiges behind, as is the case in dried-up cells. The membrane alone remains of the components of the cell. The space surrounded by it is filled in the mature tube with very dilute watery fluids, which may here be called shortly water, or with air, or with both together. The large majority of Trachez are entirely or for the most part filled with air at the time of full development. The extremely thin layer of fluid on the inner surface, which is always difficult to observe, is often beyond anatomical demonstration: even in cases of excessive supply of water in bleeding parts air bubbles occur in the fluid contained in them*. It is only in lateral extensions of the vascular bundles of certain plants (Transfusion tissue, Chap. VIII) and in the endings of bundles that they are exclusively filled with water. The same holds for the rudi- mentary Trachez of many water plants. A remarkable exception to this occurs very generally in plants which contain latex, or resinous, or tannin-containing secretions, whether the latter be stored in the sacs treated of in Sections 33 and 34, or in intercellular reservoirs (Chap. VII). A greater or less number of vesseds are often filled in these plants for a greater or less distance with latex, or with some such characteristic secretion. No fixedrule is to be found as to the position of these vessels relatively to the other normal air-containing vessels, or to the secretory reservoirs, How the secretion gets into the vessels is not 1 Compare Von Mohl, Bau des Palmenstammes ; Verm. Schr. p. 142. ? Mettenius, Beitr. zur Botanik, p. 50. 5 (Compare Westermaier u. Ambronn, Lebensweise u. Structur d. Schling- u. Kletter-pflanzen, Flora, 1880,] * Compare Hofmeister, Flora, 182° pede 170 TRACHEE. explained in plants without laticiferous tubes, though plausible conjectures may be made on the subject. The same holds as a matter of fact also for plants with latici- ferous tubes, but there are controversies on this point, which we shall return to in Chap. VI. The frequent filling up of the cavity in the Trachez, e. g. in old layers of wood of Coniferee and many Dicotyledons, with resin or resin-like masses is undoubtedly a phenomenon of incipient degradation and disorganisation, It will be further treated of in Chap. XIV. In old or damaged, large, tubular Trachez the internal cavity is not unfrequently partially or completely filled with parenchymatous cells (Fillzellen), which in the wood of the chestnut drew the attention of Malpighi?: they have since been frequently described, and have been termed by the anonymous writer in the Botanische Zeitung? Zhyloses (Thyllen). They may arise where a Trachea borders on parenchymatous cells, and in fact from those cells themselves, which grow into it. A small part of the membrane of a parenchymatous cell adjoining an unthickened point on the wall of a Trachea (as a rule a pit) grows to an excrescence protruding into the cavity of the latter: it contains protoplasm, usually with a well-marked nucleus, and expands from a blunt -and short cylindrical form to a round, often voluminous bladder, and -finally cuts itself off as a special cell from the rest of the cavity of the cell which produced it by means of a division-wall, formed at its point of entrance into the Trachea, Thus there always arise at first solitary bladder-like cells protruding from the wall into the cavity of the tube. The process may be arrested at this point: but often the- phenomenon is extended quickly over numerous points of a portion of a tube, so that the latter gradually becomes entirely coated internally with the cells, and these, as they extend, gradually fill it up completely. This often happens to such an extent that the tube is entirely filled by thyloses flattened into polyhedral forms by reciprocal pressure. Further, a multiplication of them by division has been observed in many cases’. The parenchymatous cells bordering on a tube take part unequally without recognisable rule in the formation of thyloses: some throw out thyloses at one point, others at several, others again not at all. The formation of fresh thyloses may continue for a long time in a portion of a vessel: in vessels several years old (e.g. in an eight years’ old layer of wood of Vitis, Reess, 2c.) the first beginnings of new thyloses often occur in close proximity to others apparently several years old. The cellulose wall of the thyloses, which is at first delicate, is later thickened in woody plants, and often has corresponding pits on the surfaces of contact with other thyloses. In the same plants starch may be stored up in their contents, as is the case in normal parenchymatous cells. The formation of thyloses has been observed in Monocotyledons (Arundo 1 Anat. Plant. p. 9, Tab. VI. fig. 23. ? Vol. for 1845, p. 225. ©vAAis=sac, bag, reservoir. In this treatise the older literature on the subject is referred to. For more recent information see Reess, Zur Kritik der Bohm’schen Ansicht iiber die Thyllen, Botan. Zeitg. 1868, p.1; Unger, Ueber d. Ausfullung alternder u. verletzter Spiralge- fasse durch Zellgewebe, Sitzungsber. d. Wiener Acad. Bd. 56 (1867). ® Trécul, Sur lorigine des bourgeons adventifs; Ann. Sci, Nat. 3 sér. tom. VIII (Maclura),—A. Gris, Ann, Sci, Nat. 5 sér. tom. XIV, p. 38 (Cissus). LEE os THYLOSES, 171 Donax, Canna, Hedychium, Strelitzia, Musa, Palms), and in the wood of very many Dicotyledons, both in one-year-old stems (Canna, Cucurbita, Bryonia, Cucumis, Solanum tuberosum), Euphorbia helioscopia, &c., and especially in long-lived stems of Dicotyledonous woody plants, where they are very widely distributed, and easily observed phenomena, e. g. in Vitis, Quercus, Sambucus, Platanus, Robinia, &c. But in the roots of Dicotyledonous trees, which have been examined for them (Quercus, Fraxinus, Fagus, Betula, &c.), they do not occur, or extremely rarely': in the roots of herbaceous plants, however, they occur in large quantity: Pharbitis hispida, young strong roots of Cucurbita, Urtica, Rubia, &c. The tubes in which thyloses appear are in most cases typical, wide-pitted vessels : but in Canna (and also in Musa and its allies) they are also the above- mentioned (p. 165) wide, fibrously-thickened non-perforated, tracheides. In the pitted vessels of many Dicotyledonous woody plants the formation of thyloses is a regular phenomenon, which appears in the normal uninjured plant, though not extending to all pitted vessels. In Robinia pseudacacia it is stated that the pitted vessels of the wood (and, according to Gris, all the vessels) begin to be filled with thyloses in the autumn of the year in the spring of which they were formed, and that these are at times filled with starch. Other woods behave in the same way as regards the time of first formation of thyloses, but no definite rule has been recognised for their occurrence or absence; e.g. in Vitis, Quercus robur, Platanus, according to Reess. Injuries by which the vessels are opened are, as far as investigated, without influence on the formation of thyloses in woody plants. In the large tracheides of the stem of Canna, however, they occur, according to Unger, if these have been injured, e.g. cut into, and then exposed to air or water. These facts may afford starting-points for the inquiry into the still unknown causes of a formation of thyloses, which cannot be further noticed here. 1 Von Mohl, Botan. Zeitg. 1859, p. 294, and earlier. CHAPTER V. SIEVE-TUBES. Scr. 43: The Sieve-tubes, Tubi cribrosi, were first clearly distinguished by Th. Hartig’, in the year 1837, as essential constituents of the bast and of the vascular bundles of Phanerogams, and were in some cases designated by the above name, while in others they were termed sieve-fibres. After lying unrecognised for many years, Hartig’s observations were confirmed and extended, especially by Mohl, Nageli, and Hanstein ?. The chief points of occurrence of the organs in question are those above mentioned ; they are rarely found elsewhere. They are present in both Phanerogams and Ferns. They have been most thoroughly investigated in the Angzosperms. They may therefore be treated of first as they occur in the latter plants, and afterwards the peculiarities to be found in the other divisions may be-added. The articulation of the sieve-tubes in the plants in question is throughout similar to that of the vessels treated in the preceding chapter. They arise from longitudinal rows of elongated, cylindrical, or prismatic cells, and these remain always clearly distinguishable and separable as their members. On the faces with which the members are mutually contiguous they come into open communication through the sceve-plates or sieve-fields, which are circumscribed portions of the wall, by means of numerous very small, perforated pits, the Zores of the sieve. The form of the members of the tubes is that above stated. Their ends are limited by ove flat, or slightly concave wall (concave on the under side); and this is either almost horizontal, or at.most slightly oblique, and in that case as a rule slightly broader than the middle of the member; or it is very strongly inclined, and cuts the lateral-wall on one side at a very acute angle, so that each end of a member is bevelled on one side like a chisel. The inclination of the terminal surfaces is in the latter case—though not invariably and exactly—towards the radial plane. 1 Vergl. Untersuchungen iiber die Organisation des Stammes d. einheim. Waldbaiime; in Jahresber. iib. d. Fortschritte d. Forstwissenschaft, &c. p. 125.—Compare further Hartig, Vollst. Naturgesch. d. forstl. Culturpfl. Berlin, 1851; Botan. Zeitg. 1853, p. 571.—Ibid. 1854, p. 51. 2 Von Mohl, Einige Andeutungen iiber d. Bau d. Bastes, Botan. Zeitg. 1855, p. 865.—Nageli, Ueber d. Siebréhren, Sitzsber. d. Miinchener Acad. Feb. 1861.—Hanstein, Die Milchsaftgefasse u. verw. Organe, &c. Berl. 1864.—Mohl calls the members of sieve-tubes ‘/atticed cells’ (Gitterzellen). P. Moldenhawer had already distinguished them in part as ‘vasa propria, but confused them with other elements under this name. [See further Wilhelm, Beitrage z. Kenntniss d. Siebrohren-apparates Dicotyler Pflanzen, Leipzig, 1880.—Janczewski, Sur les tubes cribreux, Mém. Soc. Cherbourg. 1881. —Idem, Et. Comp. sur les tubes cribreux, Ann. Sci. Nat. 6 sér. tom. XIV. 1882.—Russow, Sur la structure et le développement des tubes cribreux, Ann. Sci. Nat. 6 sér. tom. XIV. 1882.] SIEVE-TUBES. 173 The first of these two chief forms, which may be termed that with transverse or flat ends, is by far the most general, and is almost exclusively present in the ‘primary’ vascular bundles (Chap. VIII); the second, or sharp-ended form, pre- ponderates equally in the secondary bast of Dicotyledons. Exceptions are however found to this rule, e.g. the beautifully-bevelled members of tubes in the vascular bundles of stems of Calamus, and the roots of Aroidez (e. g. Philodendron Imbe) ; and on the other hand the flat-ended tubes of the secondary bast of Fagus sylvatica, Quillaja saponaria, Ficus elastica, Maclura, &c. The size of the members of the sieve-tubes varies, especially in different species, no less than that of the vessels. The same rules hold for the length of their members as for those of the vessels: but the maxima of diameter of the latter are not attained by the widest sieve-tubes. The widest sieve-tubes attain a diameter on the average not more than o.02™™ to 0.08™™; Cucurbita; species of Bignonia, Phytocrene, Calamus, &c. On the other hand, extremely narrow and insignificant ones are to be found, especially in many, but not all, succulent plants, and in such as have latex (e.g. Asclepiadaceze, Crassulacese, &c. Comp. Chap. VIII). A few measurements of large members of sieve-tubes may be given below, but with the remark that the measurements of length in long members are only approximately made, or from single specimens, owing to the extreme difficulty of neatly isolating such delicate organs. Length. Diameter. Internodes of mm. mm, Cucurbita Pepo . . . «. ~ 0-370—0-450 . . . « 0-045—0-050 Lagenaria vulgaris . . . . O125—0-200 . . . . 0:025—0-040 Calamus Rotang. . . . over 2™m + ee 4 © 01030—0-050 Potamogeton natans. . . . 0275 «ee « t0 0.025 Bignonia spec. (Mohl) . . . to 1-35 see 2 0450 Vitis vinifera, bast . . . . about 0-6 Root of Philodendron Imbe . to more than 2™™, As regards the longitudinal and lateral connections of the members, and the branchings which in certain cases are thus produced, the same rules apply in the main as in the case of the vessels. The walls of the sieve-tubes are always delicate, not lignified, colourless cellulose membranes. The sieve-plates characteristic of them only occur on those surfaces where the members abut on similar elements. The sieve-plate is a sharply-limited part of the wall, like a large shallow pit, which is originally, and often throughout life, less thickened over its whole surface than the wall surrounding it. It is thickly covered over its whole extent with round or polygonal secondary pits, which are separated from one another by narrow bands of membrane: it thus resembles a fine, sieve, net, or lattice (Figs. 65-73). ‘The sieve-plates of members of tubes which are contiguous fit with all their secondary pits exactly on one another, and in them the intervening wall disappears when the differentiation of tissues begins, so that holes— the steve-pores—appear, through which an open communication is established between the neighbouring members. The original width of the sieve-pores differs according to the special case. The widest occur in the Cucurbitaceze, where the largest (Cucurbita, Lagenaria) attain a size of 5 » and more: most of them are much narrower ; in the above-named Cucur- bitaceze only 2 wide; also for the bast of Bignonia spec. Mohl states it at 2p, . 174 SIEVE-TUBES. which would be too high an average for most plants: in many Angiosperms with small sieve-tubes they are certainly narrower, often being on the limit of clear recognition. Further in the same plant, and even in plates lying close together, the width of the pores is very unequal: in the large tubes of Cucurbita and Lagenaria, where exact measurement is possible, the diameter of the pores of neighbouring, and otherwise equally developed plates may differ by three times (comp. Fig. 65). In one and the same plate the difference in size of the pores is usually small, if at all recognisable » ‘according to Niageli they are, at least in Cucurbita, wider on the average at the middle than at the margin of the plate. Very considerable differences in size and form on the same plate are rare. (Comp. Hanstein, /.c., Taf. III. Fig. 4, Cucurbita.) Fig. 65. Fig. 66. Fig. 67. FIG. 6s—67.—Lagenaria vulgaris, mature internode of stem (37s). Fig. 65 and 66, transverse sections through one and the same bundle of sieve-tubes (or vascular bundle) ; » wide-meshed sieve-plate, occupying the whole horizontal end of a member, exposed, in surface view; 2, @ similar one with narrow pores, Z another, injured at one margin in cutting the section, callous, the pores still slightly open; the original cellulose sieve recognisable through the callus; 0 sieve-plate covered by the contents coagulated with alcohol; @ delicate parenchyma. Fig, 67, lateral view of two members of a sieve-tube attached end to end, the plate callous, so as to close the pores completely, the orginal sieve is recognisable in the middle between the two masses of . callus, According to the developmental data to hand, which however are not extensive on this question, the above-described simple structure makes its appearance on all sieve-plates on their first development. Many retain it long, or even throughout life ; others alter, by assuming the condition termed by Hanstein callous, The change consists in the thickening of the bands of membrane in all directions. They swell perpendicular to the surface to three or more times the original thickness, and become convex on their inner side: in the direction of the surface they expand so that the original pores are contracted to narrow cylindrical canals, which widen out like funnels only between the convexities of the inner surface. The single bands of “membrane of one plate often take a different share in the callous-thickening: this increases or decreases gradually on one plate from the middle towards the edge; in this matter both sides may be alike, or the reverse: the general form of the callous plate may thus be biconvex, biconcave, or plano-convex, &c. Often the inequalities of thickening are irregularly distributed over one face. The callous thickening may lastly extend in the direction of the surface so as to close the canals completely. Sieve- plates may often be found covered with a thick mass of callus, which is not perforated, ? [On the Callus compare Russow, Botan. Zeitg. 1881, p. 723.] . SIEVE-TUBES, 175 and in which the canals are only indicated by transverse striae, and by funnel-shaped depressions of the surface: in others even these indications are not noticeable (Fig. 67, 76). : The callous plate always consists of three lamellz, one central, and two applied to this laterally, one on each side: each of these belongs to one of the members of the tube. The middle lamella is the original cellulose sieve. The lamelle of callus are in the fresh condition homogeneous, colourless, apparently soft, and having by transmitted light the peculiar bluish lustre of gelatinous membranes: they are coloured yellow by solution of iodine in potassium iodide, and by Schultze’s solution a deep brownish-yellow: in sulphuric acid they swell till their outline is completely lost. A similar swelling results from the action of alkalies, especially solution of potash, and of Schultze’s mixture. By these reagents the callus mass may be completely removed from the persistent cellu- lose sieve. At the margin of the sieve-plate, next the adjoining membrane, the callus mass stops rather abruptly. From all these phenomena it is concluded that the callus mass is formed by apposition upon the original cellu- lose sieve. The conditions of its appearance and its physiological significance require further investigation: ac- cording to some few experiments on Cucurbita and Lage- naria, the callous thickening seems in these cases to appear and increase with the age of the sieve-tubes, and in the first-formed (peripheral) tubes of a vascular bundle it seems to advance very quickly till the sieve is entirely closed. In many plants sieve-tubes may be found side by side without callus, and with callus in the most different stages, e. g. Lagenaria : in others, e.g. in the bast of Quillaja, only callous sieves are known, but they are always open. In the bast of many ligneous plants —Vitis, Tilia—I find all the sieve-plates completely closed by callus in the winter time; in the height of summer they are open and not callous. (Comp. Figs. 69, 74; and 76.) The sieve-plates are always placed on the terminal faces of the cylindrical tubes. If these faces are horizontal, A or only slightly inclined, each has throughout the properties th of one sieve-plate, which may be termed a simple transverse —_ry¢.6e.—cucurbita Pepo; mature plate: this is so in all the above-named cases of the original Sftwo sicretabes, with large callus vascular bundles and primary bast of Angiosperms (comp. fontents Pe aerertg ae Figs. 65-67), in the slightly inclined terminal faces of Ca- ee rug icc fas bak . : : member into the other. lamus, in the secondary bast of Fagus and Quillaja. ‘On strongly-inclined terminal faces the sieve-plates are arranged like the scalari- form openings of vessels in series one above another, and are, like these, separated from one another by narrow bands of membrane : they usually form a single series, rarely they form here and there several irregular rows. Examples of this are supplied 176 . SIEVE-TUBES, by the above-mentioned secondary bast of ligneous ‘Dicotyledons, e. g. Phytocrene, Bignonia, Tilia, Juglans, Vitis (Figs. 69-70), Betula, Populus, Pirus communis’, &c., the very oblique terminal faces of Calamus (Fig. 71), Philodendron Imbe, &c. In the secondary bast of Vitis the occasional horizontal ends of members have also a ladder-like structure. Sieve-plates are distributed in different ways, according to special cases, on the lateral faces of members of tubes, where these adjoin other similar members. On most forms with simple transverse plates, as Cucurbita, they are not un- commonly absent on the sides, or they occur irregularly, and are then usually C i ; He, a FIGS. 69, 70.—Vitis vinifera, Bast of a branch several years old, 1°™ in thickness, in Summer (beginning of July). Fig. 69 (145) Tangential section. s, s sieve-tubes, the inclined terminal surfaces, and a horizontal scalariform one, are cut through longi- @tudinally, with the exception of one at the upper edge, which is seen obliquely in superficial view. 72,2 medullary rays, Fig. 70 Radial section, two scalariform terminal surfaces of sieve-tubes in superficial view, separated from one another by narrow paren- chymatous cells (375). FIG. 71.—Calamus Rotang (Spanish Reed), End of a member of a sieve-tube isolated by maceration (375). relatively small. Where the lateral wall adjoins elements of another category iso- lated, usually flat, pits are found: in these forms especially the lateral wall is very soft and extensible ; after maceration in potash its inner layer may be drawn out to some * Compare Von Mohl, /.c.; Dippel, Mikroskop, p. 251, &c.—Von Mohl’s fig. 11 of Pirus repre- sents the partial surface-view of three oblique terminal faces. : SIEVE-TUBES OF ANGIOSPERMS. 19% length. Among these forms the tubes of the secondary bast of Ficus elastica and Fagus sylvatica appear to be exceptions, since in them the lateral faces, turned towards the periphery and middle of the stem, are covered thickly with sieve-plates, which are only separated from one another by narrow fibre-like bands’. These plates on the lateral-walls are extremely delicate, and it cannot be determined whether they really have pores, or are only portions of the wall having the latticed appearance of sieve-plates. In the tubes with ladder-like terminal faces the series of plates is continued, usually quite gradually, from these to the neighbouring lateral surfaces, and especially on to the radial ones : in the bast of the Dicotyledons the arrangement is always such that the plates on the lateral surfaces are smaller and wider apart than on those which are terminal. The contents of the fresh intact sieve-tube as it lies in water appear as trans- FIGS. 72, 73.—Two large sieve-tubes of Lagenaria vulgaris in longitudinal section, where two members join, after action of alcohol and iodine solution. In Fig. 72 4° is the non-callous widely-perforated horizontal plate, as seen in longitudinal section; ~ the contracted sac-like contents, with the dense aggregation of slime. Pro- cesses from this traverse the pores on the left side; on the right side (7) they have been torn out of these in cutting the section, In fig. 73 the transverse plate (g) is oblique; thus the half of it present in the prepara- tion is seen in section, and obliquely in surface-view. The coagulated slimy contents have been quite sepa- rated from it in cutting the section, and the plate, which rested on it, has been so turned that its whole sur- face, which before abutted on the intact sieve-plate, is turned towards the observer. The processes, which before fitted into the pores, appear upon it as rings (375). parent as water. More exact investigation shows that the wall of each member of the tube is invested by a continuous thin layer of almost homogeneous slimy substance resembling protoplasm. This layer surrounds a central watery fluid, to which must be ascribed the alkaline reaction “ characteristic of the contents of bundles of sieve-tubes, at all events in Cucurbita. At one end, or more rarely at both ends, of the member it encloses an apparently dense lustrous aggregation of slimy substance, which lies upon the sieve-plate, either as a thin lamella, or as a plug of relatively considerable 1 Von Mohl, /.c.; Dippel, Mikroskop, p. 255- 2 Compare Sachs, Botan. Zeitg. 1862, p. 257. N 178 - SIEVE-TUBES. height. Usually this aggregation of slime is found at one end only of the member, and in that cage, in Cucurbita according to Nageli, in # of the instances (taking the whole plant into consideration) it is at the upper end, that is, on the under surface of the sieve-plate. In very many cases numerous very small grains of starch are im- bedded in the slime, and especially in the terminal aggregations of it. - Briosi found these in the stems and leaf-stalks of 129 out of 146 species investigated. At the sieve-plates the slimy contents are continuous through the pores from one member of the tube into the adjoining one. It may be seen, especially in callous plates, and when coloured yellow with iodine, filling all the pores and demonstrating, like a natural injection, the open communication through them (Figs. 68, 74). Where however the size of the parts makes an exact investigation possible, it may be seen that it does not pass as a homogeneous mass equally from one member into the i FIGS. 74, 75.—Vitis vinifera, Bast ; from the same branch as Fig. 69, prepared on the same day (600). FIG. 74.—Tangential section through the ladder-like limiting surface of tvo members of a sieve-tube 4 and 2, In 4 the dense plug of slime, contracted by alcohol, sending blunt processes through all the sieve-pores into 8; @ a grain of starch, FIG. 75.—Radial section, after action of absolute alcohol, destruction of the starch by brief action of strong solution of potash {the latter has caused a slight swelling of the membrane), and subsequent washing out of the potash, and treatment with iodine. Part of a ladder-like wall in surface view; beneath it the shrunken slimy contents of an adjoining member, which sends capitate processes—upwards as the preparation lies—through the pores into the other member. FIG. 76.—Bast from a branch several years old, and 1’5 ©™ thick, of the same plant in winter, Callous closed wall between two members of a sieve-tube, tangential section (400). other, but that the peripheral layer of the one member sends processes into the pores, which they fill, and end blind at the limit of the adjoining member: the processes either end simply at the surface of the sieve-plate, or are more or less swollen, and rise above it into the cavity of the adjoining member, while at the point of transit through the plate they fit into corresponding holes in the peripheral layer of the member, which they enter (Figs. 72-75). As far as is at present known the processes always extend on one sieve-plate to one side only, thus from the member a to 4, and not also conversely: further, they extend from the surface on which there is the larger collection of slime to the other. They are in their turn also filled with slimy contents. According to Briosi’s statement, that the starch-grains often stick in the r ' Briosi, Ueber allgemeines Vorkommen von Starke in den Siebréhren, Botan. Zeitg, 1872, p- 305.—Compare also Sachs, Exp, Physiol. p. 383, &c. STEVE-TUBES OF GYMNOSPERMS. 17g sieve-pores, they must often also include starch-grains: This requires more exact proof. The starch-grains are doubtless densely crowded on the sieve-plate, and especially so on the pores. But they cannot so easily and generally enter and pass through the pores, since they are often larger than these: for instance, in Vitis, at the period of most active vegetation, they are on the average twice as broad as the pores. (Fig. 74, a.) The structure described above is found in fresh intact sieve-tubes. But it appears much more plainly after the action of reagents. On treatment with alcohol the peripheral layer, resembling protoplasm, immediately coagulates along the sides of the members, separates from the membrane, and contracts to a relatively thin, folded, but still closed sac, which occupies the middle of the member (Figs. 68, 72, 74). On the face which is in contact with the sieve-plate, and which is attached by the processes in the pores, the sac retains its original width, or at least that of the perforated part of the wall: it thus widens more or less rapidly in a conical manner opposite these faces: the processes which enter the pores alter their form and position but little or not at all. Iodine preparations produce the same changes in form, and colour the whole peri- pheral layer and the terminal aggregations of slime deep yellow to yellowish brown ; the starch-grains violet*: this coloration appears much more quickly in the parts in question than in the callus-masses, so that by this means these two parts may easily be distinguished from one another: this renders the understanding of the structure more easy, especially in slightly-thickened and widely-porous sieve-plates, inasmuch as in this case the figures of the sieve-plates on the one hand, and on the other of the plates of slime (with their processes), which cover the sieve-plates, are necessarily similar in the surface-view of the plate, and are often difficult to distinguish at first sight. According to the above behaviour with alcohol, and preparations of iodine, and other known chemical reactions *, the slimy contents of the sieve-tubes, i.e. both the lateral peripheral layer and the terminal aggregations, consist in the main of an albuminoid substance similar to protoplasm. It is doubtful whether it should really be termed protoplasm, less because of the slight differences between the iodine reaction of the slime and of the protoplasm of the surrounding tissue in Cucurbita °, than because protoplasm is a body which is characterised not only by its material composition, but also by a definite organisation or structure, which expresses itself in protoplasmic movements, differentiation of nuclei, &c., and since phenomena such as the above have not been observed in the contents of sieve-tubes. Sect. 44. In the Gymmnosperms and Fern-like plants* tubes are found, in similar places to the sieve-tubes of the Angiosperms, which, from their great similarity to these, are doubtless rightly included under the same term, but differ in certain points from them, and especially in the character of their contents. The sieve-tubes of those Gymmnosperms which have been investigated—e. g. Larix, Abies pectinata, Juniperus, Sequoja gigantea, Salisburia, Ephedra, Gnetum, Encepha- lartos—are similar in form and average size of the members to those commonly found in the bast of ligneous Dicotyledons, which have the ends of the members Het Compare Briosi, 7. ¢. 2 Compare Sachs, Flora, 1863, p. 38. 3 Nageli, /.c. p. 16. 4 (See Janczewski, /.c.; also Russow, /.¢., and Strasburger, Zellhdute, p. 57.) N 2 180 SIEVE-TUBES, bevelled like a chisel. They may, like these, attain a considerable width, e.g. 0.030mm in the secondary bast of old roots of Abies pectinata. The oblique terminal faces are directed, both in the stem and in the roots, towards the radial planes (medullary rays). Sieve-plates are distributed uniformly in one or two Jongitudinal rows over the terminal faces, and the whole remainder of the radial lateral face. They form roundish spots, separated by high intervening portions, or are rarely elongated transversely, and separated by narrow ring-like bands: these spots are coarsely latticed, while in the cavities of the coarse lattice the very delicate sieve-structure is seen (Figs. 77, 78). Considering their close similarity to like parts of Dicotyledonous plants, there is no reason to doubt that the channel through the narrow sieve-pores is open. But this has not been directly proved, and the proof has hitherto been impossible, since the tubes in the plants in question are filled almost exclusively with watery fluid. The masses of starch-containing slime, giving the reactions of pro- toplasm, which in the Dicotyledons send their processes through the sieve-pores, have not yet been discovered in the plants in question : on the walls of the tubes there are attached internally some few very small granules, which turn yellow with iodine. The nature of the materials composing the fluid contents re- quires further investigation. Further, I was unable to find a formation of callus, with the exception of a doubtful case in the root of Abies pectinata. Among the Ferns a number of plants have relatively large and wide vascular elements, and among these such as are, from their position (comp. Chap. VIII), and their structure, to be enumerated among sieve-tubes. This is the case in many Polypodiacez, e. g. Pteris aquilina (Fig. 79), Marsiliaceze (Mar- silia Drummondi and its allies), Cyatheacez, Osmundacee, Ophioglosseze, according to Dippel in the Equiseta, and at least Endofa menber esiee- the larger Lycopodia?. i eee ae aa In the Equiseta and the Ophioglossez they consist, ac- Get Go cording to Dippel and Russow, of tabular prismatic members, very. fne ‘pores indicated Which stand one upon another in longitudinal rows with hori- tse PyPoins Zontal, callous, sieve-like, transverse walls. The lateral walls have no sieve-pits. In the other cases cited, the members of the tubes are fitted one on another with pointed ends (in Marsilia also with horizontal ones), and have sieve-plates both on the latter, and also on the whole of those lateral surfaces which are contiguous with similar elements. These are usually elongated transversely, forming, according to the width of the surface of wall, one or several rows: in these rows they are either crowded closely, and separated only by narrow bands of wall (Fig. 79, B), or they are at a considerable, and then usually a variable distance from one another. The sieve- * Dippel, Bericht d. 39. Versamml. deutscher Naturforscher zu Giessen, 1864, p. 146, Taf. IV. —Idem, D, Mikroskop, pp. 195, 203.—Russow, Vergl. Unters, pp. 5, 101, 118, 129, 142. SIEVE-TUBES OF CRYPTOGAMS, 181 plates are not callous. Their pores, as far as they can be recognised, are very narrow and round; and in Marsilia, according to’ Russow, very numerous on one plate; in the cases investigated by me (Pteris aquilina, Cyathea, Alsophila spec., Osmunda) they are less numerous, and relatively far distant from one another. , The wall of the tubes is thin at ‘the sieve-plates ; the rest of it is strongly thickened, stratified, and soft, and apparently swells in water. a quantity of watery fluid, and a thin peripheral layer, coloured yellow by iodine, which contains throughout, and especially at the ends of the members, and on the lateral sieve-plates, numerous very small granules which adhere closely to the wall. In dried-up tubes the ends are also found filled with a homo- geneous brown mass. These granules are not starch: they These tubes contain turn a deep yellow with preparations of iodine: maceration in Fi! 78 —Encephalartos pungens. Bast of an old dilute solution of potash destroys them only partially even after stem. Part of the radial many days. Their dense aggregation and their tenacious hold wall of a sieve-tube (375). on the sieve-plates usually prevents a clear decision on the permeability of the pores: but I believe that I have clearly seen in thin longitudinal sections in Pteris aquilina that the granules of con- tiguous sieve-tubes are connected by thin filamentous processes which tra- verse the transverse pores (Fig. 79, ¢). The tubes are not inferior in width to the medium and thicker tubes of the Gymnosperms. The length of the. single members is considerable, in the cases investigated (Pteris aquilina, Cy- atheaceze) it is 1-3™™. In the Mar- silias they attain, according to Russow, the length of one whole internode, that is, of several centimetres, a statement which may have its origin in the ease with which the ends of members may be missed in tubes prepared free by maceration, In the larger indigenous Lycopo- dia (L. clavatum, annotinum) there. occur in the vascular bundles of the stem organs which, in their position and width, have great similarity to the members of sieve-tubes of the above Ferns. They are prismatic and elon- gated, so that their pointed ends can seldom be seen in sections. Their contents correspond also to those of Gy, : i e a eS fs 3 for Ge ° a i & 7 Hi FIG. 79.—Pteris aquilina, Rhizome. 4 end of a member of asieves tube, isolated by maceration (142); & part of a thin longitudinal sec- tion. This has approximately halved two sieve-tubes, 51 and s2, which are so drawn, that the plane of section faces the observer, and the uninjured side lies behind. The latter is represented as lighter, while that which lies in the plane of section is darker. s1 abuts to the right on parenchyrnatous cells which have been cut through, tothe left on 52; at the back again, its broad surface covered with sieve-plates, is contiguous with a sieve-tube, while to the left and at the back it abuts with a smooth wall on parenchymatous cells. sg borders with its whole smooth-walled posterior side on parenchyma (the nucleus is indicated in two of the cells), right and left on sieve-tubes ; c,c sections of walls bearing sieve-pits (375). 182 SIEVE-TUBES. sieve-tubes. Neither Hegelmaier* nor I could find the clearly-latticed sieve-plates, almost like those of Pteris aquilina, which Dippel described on their lateral walls, On the other hand I saw on the whole lateral wall numerous small pits, solitary or in groups, to which were attached those peripheral granules, like those of Pteris aquilina, which turn yellow with iodine: but from these it could not be determined whether there are open sieve-pores or not (comp. also Chap. VIII). In the smaller Lycopodia, the Selaginelle, and in very many Ferns with small vascular bundles composed of narrow elements (comp. e.g. below, Fig. 160, Polypodium vulgare), the position in which the sieve-tubes occur in the above instances is occupied by elements of similar form, and general character of contents and walls, but without distinct sieve-plates or sieve-pores. Whether the latter are really absent, and whether these elements are only the morphological equivalents of sieve-tubes, remains to be further investigated both in these cases and also in Lycopodium. I am the less inclined, in the doubtful cases, to deny the presence of sieve-tubes, and prefer the more to treat the question as an open one, because these very organs counsel one to be circumspect, for twenty-two years ago no botanist, with the exception of Hartig, had any idea of the characteristic structure of the most conspicuous of them. ? Botan. Zeitg. 1872, p. 778. CHAPTER VI. LATICIFEROUS TUBES. Sect. 45. Certain plants, belonging to families or genera to be named below, known as plants which produce milk on injury, contain, in tubes of definite structure and of definite mode of development, a milky opaque fluid, which bears the name of Jatex ; after this the tubes themselves may be called Laticiferous tubes 1. The tubes traverse the parts continuously for long distances, adjoining more _especially the turgescent, usually parenchymatous elements. They are themselves completely filled with the milky fluid, their walls, though often strongly thickened, are always soft, and easily compressed. Thus if a tube be injured at any point, the pressure exercised by the adjoining turgescent tissues forces the milk out through the opening. The wall of the milk-tubes is always a soft, apparently watery cellulose mem- brane, which readily shows the characteristic blue coloration with preparations of iodine. Details of its structure will be given below. Within the wall neither protoplasm nor nuclei are to be seen®. It is true many forms of coagulated, finely granular latex, e.g. that of the Cichoracez, resemble coagulated protoplasm, or there remains here and there, in partially emptied tubes after action of alcohol, solution of iodine, &c., a coat which looks like a coagulated protoplasmic lining to the wall. Further investigations will therefore perhaps be able to prove the presence of a protoplasmic body. Still, as it is difficult to obtain sharp an- atomical evidence of such a body, and as our present knowledge both of the physio- logy and chemistry of latex is scanty, we may regard the contents as being fluids. As the name implies, all latex consists primarily of a watery, in itself transparent fluid, in which numerous undissolved small bodies are suspended as an emulsion. In most cases both parts, the fluid and the bodies, are colourless, and the milk white: more rarely the milk is orange-red (Chelidonium), or sulphur-yellow (species of Argemone) ; but it is not possible to define exactly in these cases what share each of the parts takes in the coloration. The clear watery fluid contains, as is shown by analyses of those sorts of latex which are used technically, very various bodies in solution ; others, as indicated by the phenomena of coagulation, in a highly swollen state. In these two forms there gene- rally occur in latex, varieties of gum, sugar, small quantities of albumen, often Pectic 1 Milchsaftgefasse, Vasa lactifera, lactea, or Lebenssaftgefisse, Vasa laticis of authors. 2 [On this point compare the papers cited below by Treub, Scott, Bower, and Schmidt.] ~*~. 184 LATICIFEROUS TUBES. bodies (said to occur, e.g. in species of Lactuca), further tannin in many plants, espe- cially the Aroidez, Musa, also in the Cichoracez, and Euphorbia Lathyris: peculiar constituents soluble in water are found in many sorts of latex used officinally in the dry state, as e.g. morphin combined with meconic acid in opium; and lastly, the greater part of the ash which appears in analyses. As regards the form in which the latter occur in the living plant it should be mentioned that salts of malic acid, especially malate of lime, occur in very large quantity in the latex, at least of many Euphorbias. In the officinal Euphorbia (E. resinifera, Berg) the latter salt is found in large quantity: in the latex of one-year-old plants of E. Lathyris it occurs in autumn in so large quantity that when a drop of latex escapes into the air it imme- diately precipitates innumerable crystals’. , As soon as latex comes in contact with the air, and still more quickly on treat- ment with water, alcohol, ether, or acids, coagula appear in the hitherto apparently homogeneous, clear fluid itself, and independently of the aggregation of the insoluble bodies, described by Mohl (Bot. Ztg. 1843, No. 33). The coagula collect together, and separate with the insoluble bodies from the clear fluid. These phenomena of coagulation, which appear under the action of so various agencies, point especially to a complicated composition of the fluid, and deserve further investigation. : The suspended bodies are present in the fluid in varying quantity, and the cloudiness of it is also variable, according to the age of the part, and according to the species: e.g. Morus, Nerium, and Stapelia show slight cloudiness; most species of Ficus and Asclepias have dense white milk. Excepting the starch-grains of the Euphorbias, which will be described below, the bodies have the form of round granules. They are in most cases—e.g. Euphorbia, and all plants with reticulate tubes—immeasurably small, and, when in free drops, they show active Brownian movement. The latex of the Artocarpez and Morez shows larger granules. They have in Ficus Carica an average diameter of 3u (1.4 to 5.6), and show concentric stratification, as found by Caruel?, the larger having three layers of almost equal thickness surrounding a small nucleus, the smaller only two layers. The outermost layer is sharply distinguished by different refrangibility from the inner ones. The granules of the latex of Ficus elastica, Broussonetia papyrifera, Maclura aurantiaca, have in the main the same structure; also, though less distinctly, the very variously sized ones of Morus nigra: all these granules are soft and sticky, and readily adhere and collect together after leaving the plant. The slightly milky latex, which escapes from young petioles of Nerium Oleander, contains pale, apparently homogeneous, spherical bodies of unequal size, two or more of which are often adherent to one another: the larger of these exceed those of the .Fig. in size. Much larger spheres are described in the case of Musa. On the chemical nature of the granules the existing chemical analyses tell us that —leaving out of account relatively very small quantities of substances characteristic of special cases, such as the alkaloids of Opium, which are insoluble in water—they may be generally designated on the one hand as resins, on the other as consisting of 1 The chemical definition of the crystals I owe to the kindness of Professor Fliickiger. 2 Sur les granules particuliers du suc laiteux du figuier, Bull. de la Soc. Bot, de France, XII (1865), p- 273- , LATEX, 185 Caoutchouc. There are also found relatively small quantities of fat, and wax-like bodies: of the latter a large quantity is described only in Galactodendron (Solly’s Ga- lactin). Resins are abundant, e.g. in the Euphorbias, and in Opium. Caoutchouc, on the other hand, is stated to exist in the latex of very numerous species, belonging to the most various families of Dicotyledons. It forms sometimes the large majority of the constituents which are insoluble in water, as in the Euphorbiacee (species of Hevea), Artocarpeze (species of Ficus, Castillea), Apocynacez (species of Haucornea, Urceola, Landolphia, Vahea), which yield the Caoutchouc of commerce, to which, according to existing statements, might be added the Asclepiadacez (Calotropis gigantea) and Lobelia Cautschuk?. In other cases, according to certain unreliable statements, it occurs in small quantities in many sorts of latex, e.g. in that of Lactuca virosa and Papaver somniferum. It is still uncertain whether the constituent described as Caoutchouc or ‘india-rubber’ is universally the same chemically distinct body. Further, it is uncertain whether the body or bodies included under this name are the sole constituent of the granules of latex, or whether these each consist of a mixture of different substances. The above-mentioned differentiation of the granules of Ficus, which certainly consist chiefly of Caoutchouc, into layers of different refrangibility, makes the latter view more probable in this case. Besides the bodies described, there are found in the latex of the Euphorbias numerous starch-grains*. In the herbaceous (Tithymalus) species they are usually of the form of cylindrical or spindle-shaped rods, which in E. Lathyris grow to 55h in length, and top in thickness, in E. Cyparissias to 40n in length, and 6p in thickness ; more rarely they have a roundish form, or (especially in E. Myrsinites) rather swollen ends. In the shrubby and succulent species of hot latitudes they are shaped like a flattened rod, and appear from the narrow side linear-spindle-shaped ; seen from the broad side they show a massive broad central part, and much widened, roundish spatula-shaped, often lobed ends*. Also in others but by no means all of the Euphorbiaceze rod-shaped starch-grains occur in the latex: e.g. spindle-shaped ones in Exceecaria sebifera, Miill., staff-shaped in Hura crepitans*. How far the blue coloration, which Hartig ° saw appear with iodine in glycerine in the latex of Cheli- donium, and which Trécul ® saw with iodine in that of Nerium, Cerbera Manghas, &c., after boiling with potash, arises from very small starch-granules, remains to be further investigated. The above observations, compiled from data at hand, will show sufficiently well how little is known for certain of the anatomy of latex, which has been entirely neglected since Mohl’s work of the year 1843, and how much may be expected from renewed investigation. The same holds with regard to the chemical conditions. A number of investigations have been made in this direction on sorts of latex, such as Opium, Euphorbium, &c., which are used technically or medicinally, without giving any 1 Compare on the plants which yield Caoutchouc, Collins, Report on the Caoutchouc of Com- merce, London, 1872; Wiesner, Rohstoffe des Pflanzenreichs, p. 153. 2 Rafn (Pflanzenphysiol. p. 88) first noted them; they were first recognised as starch by T. Hartig, 1835; Erdmann and Schweigger-Seidel, Journ. f. pract. Chemie, Bd. V. p. 4. % Compare Meyen, Physiol. /.c.; Nageli, Starkekorner, p. 428; Weiss und Wiesner, Botan, Zeitg. 1861, p. 41, 1862, p. 125. * Vogl, in Pringsheim’s Jahrb, V. 5 Botan. Zeitg. 1862, p. 100. ® Comptes Rendus, tom, LXI (1865), p. 156., 186 LATICIFEROUS TUBES. ground for deciding upon the possible chemical changes which result from drying in the air. Reference must be made to the technical, and especially the pharma- cological literature, for information on the above-mentioned investigations’. Here only some few results may be given of analyses of fresh specimens of latex, or of such as were kept so as to prevent drying up, by way of giving a rough idea. Faraday ? investigated the latex of a rubber-tree of the family Euphorbiacee— Hevea elastica, Siphonia elastica, Pers.’ probably H. guyanensis—which had been sent to England in closed bottles. The fluid contained in 1000 parts— Water with an organic acid : F ds : . ; - §63 Caoutchouc ; 7 3 : A : : . , . 317 Albumen . 7 5 . . : : 5 . . 19 Bitter substance (with much nitrogen) and some wax : i 413 Bodies insoluble in alcohol, soluble in water. : . ‘ 29-1 The preserved latex of Galactodendron utile contains, according to Heintz*, in 100 parts— Water. é ‘ ‘ 6 é - 573 Albumen . 3 5 5 é ‘ - 0-4 Wax (C3,H;,O3) - 5‘ ‘ ‘ i 5-8 Resin (C3, H 5,0.) . , : a - 3r4 Gum and Sugar 5 , ‘ ‘ : 47 Ash . 3 z . : , : 3 O-4 ‘Weiss and Wiesner * investigated the fresh latex of some indigenous Euphorbias. In the slightly acid latex of E. Cyparissias they found in 100 parts— ‘Water ‘ Z r : . + 72913 Resin _ : 7 3 5 ‘ . 15°72 Gum ‘ 7 ; F ‘ : . 3°64 Sugar and extractives ‘ 7 , rs 4°13 Albumen , . : 7 7 ‘ . O-14 Ash , - ‘ 7 3 3 a é 0-98 For comparison with these may be given the composition, as found by Fliickiger', of Euphorbium, i.e. the fixed residue of the latex of E. resinifera— Amorphous resin (C9 Hy O,) « ‘ ‘ 38 Euphorbon (Cy; H 4,0.) . . 3 ‘ 22 Mucilage ® ; ° 7 : ; A 18 Malates, especially of calcium and sodium? 12 Other constituents of ash ‘ - , 10 100 Sect. 46. The tubes themselves, in which the latex is contained, are all alike in certain points of structure and arrangement, but may be divided, according to form and development, into two categories, articulated and non-articulaied laticiferous tubes. Each of these categories is peculiar to definite families, the ar/culated tubes being 1 Compare Wiesner, Rohstoffe des Pflanzenreichs ; Flickiger, Pharmacognosie; Fliickiger and Hanbury, Pharmacographia; Rochleder, Phytochemie, &c.; also Meyen, Physiol. IT. 7. ¢. * Compare Berzelius, Jahresbericht for 1827 (German by Wohler), p. 246. 3 Poggendorff’s Ann. 65 (1845), p. 240. * Botan. Zeitg. /.¢. 5 According to Fliickiger and Hanbury, Pharmacographia, p. 504. * Probably inclusive of starch or its products of decomposition, ™ Compare the former statement for E. Lathyris. se LATICIFEROUS TUBES. 187 found in the Crchoracee, Campanulacee, Lobeliacee (and, according to Trécul, in Gun- delia Tournefortii, one of the Cynaracee), the Papayacee, many Papaveracee (Papaver Roemeria, Argemone, Chelidonium, but not Glaucium, Macleya, Sanguinaria), many Arodee, and Musacea. The non-ariiculated tubes are found in the Luphorbiacee, Urticacee in the wider sense (including the Artocarpee, and Morez), Apocynacee, and Asclepiadacee. The properties common to all are, firstly, that they traverse the whole length of the mature parts of the plants as continuous, and, with rare exceptions (Musa, Che- lidonium), frequently-branched tubes; and not only do they traverse each single member of the plant on its own account, but they send branches from these into all the like and unlike lateral appendages. Secondly, all laticiferous tubes have, as was above noted, soft, apparently very ‘watery, smooth or flatly pitted cellulose walls, which often show the characteristic iodine reaction of collenchymatous walls (cf. p. 120). These are in many cases very delicate, and without recognisable minute structure: thus, e.g. almost universally in the Aroidez, in Vinca, Asclepias curassavica, and in other cases in the thin branches of higher order. The membrane of the stronger stems and branches of most tubes is thickened, and appears as though swollen in transverse sections, having delicate stratification, while striation is also seen, especially in the strongly-thickened tubes of woody stems (species of Euphorbia, Nerium). The thickening increases with age. Even in the very thick membranes no sculpture of the surface can be recognised : on the other hand delicate transversely-oval pits are often seen, e.g. in the tubes of Plumiera alba, in those of the base of the stem of Campanula Medium (Trécul), and in old stems of Lobelia syphilitica. In the base of the ‘stem of species _of Argemone are found crowded together band- and knot-shaped thickenings which protrude inwards. The wall of the tubes of Plumiera alba may be split, according to Trécul, into spiral bands of rop—15p in breadth. Further, pits occur less commonly on the lateral walls than appears at first sight, especially in articulated laticiferous tubes, since the lateral wall of these, especially when old, has often very numerous and short protrusions, which give, in surface-view, the figure of pits with delicate contour. I could never find support for those state- ments, according to which the pits of the lateral wall have the structure of sieve- plates. As will be more completely described in Chapter XII, but must here be only briefly stated, the laticiferous tubes permeate the whole body of the plant, in most cases as a continuous system, sending branches from the stem into all lateral members. These branches often push their way between the elements of all regions, and of all tissues of other sorts than themselves. As regards the main branches or stems of the tubes however it is generally the case that they follow the vascular bundles, i.e. the wood and bast, as concomitants of, or sometimes even substitutes for the sieve-tubes. In this course they often approach very closely in point of distance to Tracheze, especially vessels, and on this circumstance, together with the above-mentioned (p. 169) occurrence of apparently coagulated latex within the vessels of laticiferous plants, depend the controversies on the anatomical relations between these two series of organs. The fact is that the trachee of the furthest ends of vascular bundles in the laminz of leaves are often accompanied by branches of laticiferous 188 LATICIFEROUS TUBES. tubes, and are in immediate contact with these’: further, that in the xylem of the stem of the Papayacez the laticiferous tubes are directly and firmly attached to the large vessels, sometimes throughout their length, sometimes by single ends of their branches?: again, that a similar relation exists between the laticiferous tubes which accompany the vascular bundles of many Aroideze and the trachez which belong to them®. Finally, it is an indubitable fact that in sections through plants having latici- ferous tubes, numerous vessels are often to be found filled with coagulated masses, which appear very similar to the coagulated latex of the plant, and which also have its characteristic colour, e.g. in Chelidonium reddish yellow*. This latter phenomenon often occurs very conspicuously in roots, and under conditions which do not allow of the idea of a flow of the latex from a cut surface. Trécul concludes ‘from this series of facts, observed by him in many cases, that in all plants with ‘laticiferous tubes at least single branches of the tubes come into direct contact with tracheze, and open communication is set up by perforation of single portions of the wall at the points of contact®. He even states that he has directly observed the points of perforation, e.g. in Lobelia laxiflora. Other observers, among whom I must place myself, according to my investigations up to the present time, have been unable to find such contact and communication of milk-tubes with the trachez, with the exception of the above-mentioned cases of the Aroideze and Papayacez: on the other hand, they have seen that where branches of milk-tubes run from the cortex to the pith, they take a course by the medullary rays through the woody or vascular ring. ‘Trécul’s statements accordingly require further proof: in the first place, that on the’ general contiguity of milk-tubes and vessels, and secondly, that regarding the open anastomosis of milk-tubes into vessels, which should be confirmed in the case in which it was observed by him, and, at all events, in the Papayacez and ‘Aroidez. According to the present data these anastomoses occur very seldom, and it is extremely difficult to prove them with certainty by direct observation, and to distinguish them from non-perforated pits. If they do really occur, still it is not proved that they are proper to normal. tissue, and not really pathological phenomena, i.e. ruptures in the thin places of contact of the tubes, which arise in the same way, ‘through the pressure of the turgescent parenchyma, as the discharge of latex on surfaces of section. The apparent presence of coagulated latex in vessels is evidence at first sight of the existence of open pores: the normal existence of these is however made the more. doubtful by the fact that, as far as investigated, the occurrence appears to be quite irregular and inconstant, and that milky or resinous coagula are found in vessels even in such plants as have no laticiferous tubes, but closed secretory cavities, without any open connection with the vessels. If an anastomosis, or even a mere contiguity of milk-tubes with the trachez, ‘does not occur in most cases, and I consider this most probable, there is then no ex- planation of the occurrence of what appears to be coagulated latex in the latter: for such an explanation it is however necessary, in the first place, to answer the question, 1 Compare, c.g. Hanstein, Milchsaftgefasse, Taf. IX. fig. 13 (Lactuca virosa). 2 Trécul, Ann. Sci. Nat. 4 ser. VIL. p. 289, pl. 12 (1857); Comptes Rendus, tom. 45, p. 402. 5 Compare Hanstein, /.c.; Van Tieghem, Structure des Aroidées, 2.¢. Taf. II. fig. 1, pp. 6-8. * [Compare von Héhnel, Milchsaft in Tracheen, &c. ef. Bot. Jahresbericht, 1878, I. p. 30.] > Compare especially Comptes Rendus, tom. LX (1867), p. 78. LATICIFEROUS VESSELS. 189 whether those coagula are really latex which came as such from the tubes, and not products of coagulation of fluids which had diffused through the walls of the vessels. Sect. 47. The distinctions of the two categories of laticiferous tubes depend upon certain phenomena of their development and form. The articulated series, as types of which those of the Cichoracez, Papaveracez, and Papayacez may serve, arise from series of elongated cells of the meristem (or cambium), which coalesce, by perforation of their septa, to form continuous tubes. In the simplest case, which occurs in Musa and Chelidonium (Figs. 80, 81), the tubes remain simple, or are branched, and con- nected in a net-work only inasmuch as one series of their original members may FIG. 80.—Chelidonium majus; tangential section through the FIG. 81.—Chelidonium majus; stem, cortex, radial secondary cortex of an old root; #—72 and 6—é portions of milk- section, part of a milk-tube with a perforated sep- tubes between the cells of the parenchyma, At a—a@ mm passes tum at s (225). éeneath the parenchymatous cells (225). continue its course from any given point as two diverging series, and conversely, The septa between the original members are here perforated only in the middle by one or few holes; their margin is persistent ; occasional large openings in the lateral wall also occur in rare cases, where two tubes are directly contiguous. In the majority of really typical cases the septa between the members of each series soon disappear completely, so that in the mature tube no trace of them re- mains. Occasionally in such cases single septa may persist through life. The tube puts out lateral protrusions, usually at numerous points, which force Igo LATICIFEROUS TUBES. themselves between the neighbouring unlike tissue-elements, and grow out to cylindrical branches, which sometimes remain short, not longer than broad, and sometimes attain a considerable length. Some of these protrusions end blind, others join with similar ones from neighbouring tubes, or with the trunks of these, and open communication is formed by disappearance of the wall at the point of contact. Where two tubes run longitudinally side by side, they are further directly connected by numerous large perforations of the wall of contact. Thus there arises a net of communicating tubes which is usually very complicated, with meshes of most various form and size, and with blind branches of various length and direction, imbedded in ! 3 ws puagae|| aay skal Ie Fic. 82.— Tangential section from the FIG, 83.—Scorzonera hispanica; ~ slightiy magnified tan- cortex of Lactuca virosa, with three reti- gential longitudinal section through the bast of the root. In culately joined milk-tubes (225). the, parenchyma the reticulately connected milk-tubes; 5 piece of a miik-tube and its surroundings, more highly mag- nified. From Sachs’ Textbook. surrounding —usually parenchymatous—unlike tissue (comp. Figs. 82, $3). This net- work, as above stated, extends throughout the whole plant. Non-reticulate articulated tubes, such as those of Chelidonium, aré branched, at least at the points of insertion of lateral ramifications, and send out branches from these points into the latter. Sect. 48. The non-articulated laticiferous tubes do not show a net-like LATICIFEROUS CELLS. 1gol anastomosis in any well-constituted case ; all their branches, which are often very numerous, end blind (Fig. 84). Anastomoses may occur between their branches in the nodes of many plants (leafy Euphorbias), but this is quite uncertain. Each tube arises, not from a series of coalescing cells, but from one single cell of the meristem, which grows so as to form a long branched sac, and forces its branches be- ( tween the other tissue-elements. The statements concerning their first develop- ment differ greatly. According to Schmal- hausen’s investigations on species of Euphorbia and Asclepiadacez, to be more fully stated below, some few cells of the meristem in the cotyledonary node of the embryo, outside the plerome, are the starting-points of the milk-tubes. These begin, even in the young embryo, to elongate into cylindrical sacs, and to penetrate with their growing ends be- tween the neighbouring cells into the cotyledons and towards the end of the root: they have at an early stage oc- casional branches in the cotyledonary node. Ad ‘/udes in the primary cortex, the leaves, and the pith of the mature plant are dranches of these few sacs, which are present in the young embryo. From the embryonic stage onwards their ends extend to close (6-8 cells) beneath the e rimar’ ing points, and grow on- P 1 ary grow 5 po 8, 5 FIG. 84.—4 part of a trunk of a milk-tube, with its stronger wards with these, sending branches, which branches prepared free and spread out, hardly larger than ? natural size, from the stem of Euphorbia splendens. 4@/ ends have the same properties, into the lateral of treut and branches are broken off. B from the stem of buds, leaves, and roots, as soon as these Mae DR ean Cantante eectesoseuenee make their appearance. Lastly, they at branch further and elongate in the meristem and the young parts so as to form the final system of tubes. The whole plant, e.g. a shrub of Euphorbia the height of a man, has thus only few, much-branched milk-tubes, the ends of the branches of which extend on the one hand into all growing points, and grow with these to an unlimited extent, on the other hand they are distributed in the mature tissues in the manner described, and end blind. As a matter of fact pieces of tubes an inch long, with hundreds of branches, may be teased out of macerated portions of the stem without finding a single anastomosis, or any other blind ending than those of small lateral branches (comp. Fig. 84 A). : According to observed facts the poss¢b/ity cannot be denied, that in later stages of development of a plant, especially in the nodes, single cells of the meristem develope into new milk-tubes, and may coalesce as branches with those which originate from the embryo. However, the occurrence of this phenomenon, if indeed it occurs at all, 192 LATICIFEROUS TUBES. is very restricted, and not clearly proved hitherto by any observation. I was unable, even in the milk-tubes, which are present in large quantities throughout the secondary bast of Morus, Ficus, Maclura, and Nerium, to prove that they arise successively afresh from the cambium (Chap. XIV), and are not branches of the original tubes which thrust themselves into the layers of secondary bast. A positive proof that the latter is the case has, it is true, not been as yet obtained. The peculiarities of form and structure of milk-tubes in special cases of their occurrence will, in order to avoid unnecessary repetition, be described together with the treatment of their arrangement in Chap. XII. The peculiar juice which flows from milky plants, and the peculiar tubes which contain it within the plant, were known to the fathers of vegetable anatomy, but with- out their having sharply distinguished them from the reservoirs of resin and other se- cretions, described in Sect. 34, and from the intercellular passages which so frequently have contents of similar appearance. The terms then in use, viz. Succi proprii, and reservoirs of such peculiar juices, referred rather to these latter forms of tissue and cavities, or to their contents’. After many more or less successful attempts to distin- guish them accurately and separate them (the history of which may be found in Meyen and Treviranus), C. H. Schultz-Schultzenstein from 1823 onwards? drew the attention of his contemporaries to the tubes with which we are engaged, and earned for himself in his later works, especially two large publications in the year 1841°, the merit at least of isolating by maceration the network of tubes of the Cichoracez, Campanulacee, and the tubes of the Euphorbias, &c., and drawing them for the most part correctly in their main points. It is true his works were rendered unpalatable by his terrible views on circulation, or, as he calls it, Cyclosis of the milk, or sap of life (Latex) in the ‘Vasa laticifera,’ and that which was really good in his observations was hidden by the mass of preposterous statements and representations into which he let himself be led by the idea that networks of vessels of sap of life must be universally present. Misunderstood sieve-tubes (the structure of these was quite unknown before 1837), fungal hyphze (comp. Cyclosis, Taf. IV), and many objects which cannot be defined ac- curately from the figures, were confounded with real milk-tubes Almost simultane- ously with these larger publications of Schultz, and in connection with them, Meyen* gave, on the whole, very good representations of the structure of a number of latex- tubes, but without giving any close attention to the most elaborate networks of tubes such as occur in the Cichoracez and their allies. The knowledge of the structure and distribution of the tubes was further advanced by Hanstein 5, Dippel ®, and other authors, to be named later, who have worked at their development; Hartig” having first sharply distinguished the articulated from the non-articulated forms, and Unger ® * Compare Treviranus, Physiol. I. p. 137, &c.; Meyen, N. Syst. d. PAanzenphysiol. IT. p. 371, &c. ? Ueber den Kreislauf des Saftes im Schéllkraut. 8 Die Cyclose des Lebenssaftes in den Pflanzen, Nov. Acta Acad. Leopoldino-Carolin, vol. 18, Supplem. II. p. 336 S. 33 Taf.—Mémoire pour servir de réponse aux questions de l’Acad. des Sci. pour l’Année 1833; Mém. prés. 4 l’Acad. des Sci. tom. VII. p. 104 S. 23 Taf. * Die Secretionsorgane der Pflanzen (1837), Physiol. II. pp. 376-386. * Die Milchsaftgefasse, &c. Berlin, 1864. * Entstehung der Milchsaftgefasse, Verhandl. d. Bataafsch. Genootschap. &c. te Rotterdam, tom. XII. p. 3 (1865). * Botan. Zeitg. 1862, p. 99. : * Anatomie und Physiologie, p. 157. LATICIFEROUS TUBES, DEVELOPMENT, 193 ‘having given a short and clear, though not quite correct review of the leading types. A great service in advancing the knowledge of these organs was further rendered by Trécul, who, in a series of papers published since 1862, described his very numerous observations, and thereby gave the impulse to the new works on the development of these organs. It is true Trécul* inclines again to the old idea of circulation, and places the milk-tubes nearer other reservoirs of ‘sucs propres’ than is allowable from an anatomical point of view. Finally, Vogl? has supplied a number of valuable contributions and confirmatory observations. The history of the origin of the milk-tubes, so indispensable to a clear under- standing of their structure, remained long in the dark®. Unger’s view *, according to which (from observations on Ficus benghalensis) they arise from rows of cylindrical cells by coalescence, found no response, and opinions remained undefined, till in 1846 the often-mentioned anonymous writer in the Botanische Zeitung expressed the opinion, as the result of an extended series of investigations, that each milk-tube is originally an intercellular passage without a wall of its own, which is subsequently provided with a membrane peculiar to it, through the agency of the adjoining cells. The contradiction by Schacht® of this view, which was at first not unfavourably received, and the subsequent answer by Trécul, called forth the series of newer works, by which the anonymous writer was refuted, and a clearer knowledge of the matter was gained, at least in many particulars. Unger formulated his view afresh in 1855 °,.in these words. The milk-tubes are ‘shorter or longer, cylindrical, irregular or branched cells, filled with opaque milky or dark coloured juice, which fuse with one another in longitudinal rows or at their points of branching.’ He lays 1 From the series of Trécul’s papers which treated of ‘sucs propres,’ those on milk-tubes may here be named: Des vaisseaux propres en général et de ceux des Cynarées laiteuses en particulier, L'Institut, 1862, p. 266.—De la présence du latex dans les vaisseaux spiraux.... et de la circulation dans les plantes, Comptes Rendus, tom. 45, p. 402 (1857).—Des laticiféres dans les Papavéracées, Ibid. tom. 60, p. 522 (1863).—Sur les laticiféres des Euphorbes, &c., Ibid. tom. 60, p. 1349.—Lati- ciféres et liber des Apocynées et des Asclépiadées, &c., Ibid. tom. 61, p. 1349; L'Institut, 1862. p. 215. —Des laticiféres dans les Chicoracées, Ibid. tom. 61, p. 785 (1865).—Des laticiféres dans les Cam- panulacées, Ibid. p. 929.—Des vaisseaux propres dans les Aroidées, Ibid. tom. 61, p. 1163 (1865), et tom. 62, p. 29 (1866).—Matiére amylacée.... dans les vaisseaux du latex de plusieurs A pocynées, Ibid. tom. 61, p. 156 (1865).—Rapport des laticiféres avec le systéme fibro-vasculaire, Ibid. tom. 51, p. 871 (1860).—Rapports des vaisseaux du latex avec le systéme fibro-vasculaire; ouvertures entre les laticiféres et les fibres ligneuses ou les vaisseaux ; Ibid, tom. 60, p. 78 (1865).—Des vaisseaux propres et du tannin dans les Musacées, Ibid. tom. 66, p. 462 (1868).—Most of these works we:e reprinted in the Annales des Sciences Naturelles; all in Baillon’s Adansonia, tom. VII-IX. 2 Ueber die Intercellularsubstanz und die Milchsaftgefasse in der Wurzel des gemeinen Lowen- zahns, Sitzungsbr. d. Wiener Acad. Bd. 48.—Beitriige z. Kenntniss der Milchsaftorgane d. Pfl., Pringsheim’s Jahrb. V. 3 [On this subject see further Schmalhausen, Beitr. z. Kenntniss d. Milchsaftbehalter d. Pflanzen, Meém. de l’Acad. Imp. de St. Pétersbourg, XXIV. 1877.—Faivre, Compt. rend. hebd. t. LXXXVIIL— D. H. Scott, Development of articulated laticiferous vessels, Quart. Journ. Micr. Sci, 1882,—E. Schmidt, Botan, Zeitg. 1882, p. 435.—Treub, Comptes rend. 1 Sept. 1879, Archives N éerlandaises, t. XV]. - a des Wiener Museums f. Naturgesch. bd. II (1840), p. 11, where it is attempted to adduce a proof, weak enough even for that time.—Endlicher u. Unger, Grundziige (1843), p. 40. 5 Botan. Zeitg. 1851, p. 513. 6 Anatomie und Physiologie, p. 157. oO 194 LATICIFEROUS TUBES. special stress: upon the coalescence of originally separate cells, since he places all milk-tubes among his ‘ cell-fusions, and only cites as cases where the cells do not coalesce to tubes, those of Chelidonium, in which plant he overlooked the perforation of the septa, and those of Sanguinaria, which must be entirely excluded from the category of milk-tubes. All investigators acceded in the main to Unger’s view. Dippel and Hanstein made thorough investigations, which proved it clearly for many cases (articulated tubes). Schacht had already published at an earlier date * an excel-, ‘lent history of the development of the tubes of Papaya from coalescing cells of the meristem, and greatly shaken thereby his own view repeatedly proclaimed since the above-cited work of 1851, according to which the milk-tubes are no special form of tissue at all, but only ‘ Bast cells,’ i.e. sclerenchymatous bast fibres, filled with latex’, All the later authors mentioned, who expressed their views on the subject, extended the above theory of coalescence to all milk-tubes, both articulated and non- articulated. The first objections to this are to be found indicated by Hartig, but have recently been more clearly brought forward by David*®. As may be seen in the above statements, the two sorts of tubes should be treated separately. For the articulated tubes it can no longer be doubted, after the excellent descrip- tions of their development by Schacht, and particularly those by Dippel, that they arise, as above described, by coalescence of cells. Without the development having been traced step by step, this follows in the case of the tubes connected into a net from the fact that in an earlier stage there are only simple cells of the meristem in the position afterwards occupied by the network. ‘The fact may be proved most: clearly in the secondary cortex of the Cichoracez. In the tubes of Chelidonium, which are not connected into a network, the limits of the original cells remain partially preserved through life. The xon-articulated tubes offer much greater difficulties. Most authors since 1846 have simply extended to them the same history of development as was observed for the articulated tubes; only Dippel and David attempted to get to the bottom of the question by direct observation. Dippel followed the tubes, in Ficus Carica and Euphorbia splendens, into the youngest meristem of the growing point, and close to the latter he found here and there septa in the tubes: he found such septa also in occasional preparations of old tubes of Euphorbia Cyparissias, Asclepias curassavica, Nerium Oleander, and Vinca minor, and concludes from this that they arise by coalescence. David arrived at a quite different result for the families named in the title of his dissertation, of which he investigated the following species specially: Euphorbia splendens, Caput-medusz, Lathyris; Ficus elastica, Carica; Nerium Oleander, and Hoya carnosa. According to him each non-articulated milk-tube is one cell, ‘a milk- cell,’ originating at an early stage by the elongation of one single cell of the meristem, which branches, and thrusts itself between the elements of the surrounding tissue. All the branches of each of these cells end blind and closed: the cell may grow to a great ' Monatsbr. d. Berliner Academie, 1856, /.¢. ? [F. O. Bower, on Gnetum, Q. J. M. S. 1882.] * Ueber die Milchzellen der Euphorbiaceen, Moreen, Apocyneen, und Asclepiadeen; Dissert. Breslau, 1872. LATICIFEROUS CELLS, DEVELOPMENT, 195 length, in E, splendens, e.g. more than r2™M, in E. Lathyris, the length of an inter- node plus the leaf belonging to it, that is about 2ocm. As the plant grows, new milk-cells are formed in the growing point of the stem: the tubes of the leaves are only branches of those which traverse the stem. It is obvious that this view scarcely differs except in name from Schacht’s theory of bast cells. David’s chief proof of his statement was obtained by him by macerating sections of the apical meristem with potash, and then teasing out from them the young milk-cells, which are at first short and spindle-shaped, but gradually become elongated and branched. In non-macerated sections also he was able to find the required early stages. The examination even of the mature specimens shows that David's view is impossible, for, as above shown, the tubes, e.g. of the Euphorbias, can be followed for any distance, and numerous blind peripheral ends of branches may be found in leaves, cortex, and growing points, but never a completely closed tube which is of less length than the whole plant. Were the tubes originally formed in succession as single completely closed cells of the growing point, they must have coalesced with one another to form the mature tubes. As a matter of fact those solitary spindle-shaped initial cells of the milk-cells do not exist. The tubes run continuously into the furthest meristem of the growing point, their ends can be followed to within 6-8 cells of the extreme apex: their course is variously curved both in radial and tangential direction between the de- veloping parenchymatous cells; longitudinal sections must thus cut off portions of them, which are roundish, or spindle-shaped, or cylindrical, and look exactly like cells of such form, especially if the section be very thin and transparent, or if the delicate, very readily swelling membrane be swollen by maceration in potash. David’s young milk-cells are such sections of tubes: they were represented by Dippel, but were tightly explained. Dippel’s view does not rest, like that of David, on mistakes of observation, which might easily be avoided: it is, from analogy with articulated tubes, extremely probable : still 1 was unable, though I investigated the matter repeatedly, with every prejudice in its favour, to find the view confirmed by observation. If one examines the ends of stems in species of Euphorbia, Stapelia, or Ficus, which are elongating, but before secondary thickening begins, the last ends of the tubes and their branches always extend, as has been repeatedly stated, into the furthest meristem of the growing point, and of its youngest leaves and branches, and I was never able to observe in their cavity traces of septa or of any coalescence of originally separate cells. Where septa appeared to be present in the young ends of tubes, and this was not rarely the case, continued and careful observation of the preparation always led to the conclusion that these were not within the tube itself, but belonged to cells above or below it. The state of things above indicated is best seen in radial longitudinal sections, which have become quite transparent by maceration for one or several days in very dilute solution of potash, without very considerable swelling of the cell-membranes, and which are thick enough to allow of following the tubes for a considerable distance. Accordingly I can only explain the septa, which Dippel represents as present in the young ends of milk tubes, as walls outside the tubes, or perhaps also as the contiguous walls of two tubes cut obliquely. The septa described by him as occurring here and there, I also have certainly been able to find not uncommonly at the nodes (but only there), in 02 196 LATICIFEROUS TUBES, Euphorbia Lathyris: they are thick transverse cellulose plates in the main tubes, They perhaps show, as above stated, that new branches and continuations of the tubes coming from below arise from cells of the meristem, which coalesce with them, Perhaps they are subsequently formed structures in the originally continuous sac. Observations on the growing plant, from the stage of germination onwards, having constantly proved the existence of tubes which extend continuously into the extreme meristem, there was reason for supposing that these arise in the embryo in small number only, and that, once formed, they grow on with the plant in such a way that the whole system of tubes of the stock, exclusive of the layers of secondary cortex, is derived from their elongation and branching. In order to test this suppo- sition, Herr J. Schmalhausen undertook in the laboratory at Strassburg an investiga- tion of the development of tissues in the ae of species of Euphorbia (E. Lathyris, Myrsinites, Lagascze). I append the summary of his results up to the present time word for word, as it was given to me, with the remark that what is said on the coalescence of branches has, as above indicated, always appeared doubtful to me. ‘The first initials of the laticiferous tubes appear at a very early stage of develop- ment of the embryo of Euphorbia. At that moment, when the cotyledons begin to make themselves prominent, there are single cells, lying approximately in the same transverse section of the embryo, which first distinguish themselves from those surrounding them by a peculiar refractive property of the cell-walls, which gives them the appearance of being swollen. At this time the plerome cylinder is clearly marked off at the apex of the root from the three-layered mantle of periblem and dermatogen by a boundary line, which appears sharply in optical section; while in the upper cotyledonary portion of the embryo no arrangement in layers is visible. Where the limit between the plerome cylinder and cortex of the end of the root ceases at the upper end of the latter, that is in the part belonging to the cotyledonary octants, the cells in question may be seen so placed that the line separating the plerome cylinder from the cortex leads up to the base of these cells. These original cells of the laticiferous tubes extend greatly at first in different directions, so that they can now be easily recognised by their great size. As the embryo continues to grow the cells increase in length, and, thrusting their upper and lower ends between the cells of the sur- rounding tissue, they put out processes upwards into the cotyledons, and downwards into the end of the root. Besides this, lateral processes are formed, which compose a felt in the node of the embryo, and surround its growing-point. Thus the sacs of the embryo are formed, not by coalescence of cells, but by apical growth of the processes of the original cells, which force themselves between the cells of the embryo: where two processes meet with their ends, the wall separating them is often absorbed, and a coalescence of the sacs takes place, as is the case also in the nodes and between the branches of the main tubes in young leaves (or cotyledons) : this occasionally occurs also in the apex of the root, ‘The best proof that the sacs have an independent apical growth, and do not arise by coalescence of cells, is obtained at the apex of the root, where they have a direct course. The processes of the original cells, which grow into the apex of the root, are arranged approximately in two concentric layers: one series of them belongs to the plerome cylinder—they permeate a layer of cells of the apex of the root, which is subsequently found to be within the layer of endodermis; the others LATICIFEROUS CELLS, DEVELOPMENT, 197 are found in the and or 3rd layer of cells beneath the surface. Both series have an almost straight course, and not unfrequently they may be followed through their whole length from the node of the embryo to the apex of the root. Before reaching this, they terminate with diminished diameter, measuring hardly half that of the sur- rounding cells. Behind the tapering end the lateral walls are bulged out, with teeth, which fit between the surrounding cells: further back the diameter of the sac increases, the teeth are smoothed down, and their walls are only slightly sinuous: at the top, where the diameter of the sac is not less than that of the surrounding cells, or even greater, its walls are smooth. The appearance of such a sac gives the impression as though it were only with difficulty that it could find room between the cells to push in its apex, and that it endeavours, by extension, to fill up all possible cavities. The surrounding cells may at this time have firmer walls, while those of the tube are soft—therefore the young milk-tubes cling to the surrounding cells, and push out teeth between them. Later the walls of the tubes may become firmer, the in- equalities smooth themselves out, and as the other tissues gain more space, the walls become quite even. Growing in this manner the tubes permeate the ripening embryo to the extreme apex of the root, and reach the seat of its future growth, viz. the limit between the apex of the root and root-cap. ‘In the germinating seed other remarkable phenomena become apparent. While the apices of the laticiferous tubes grew up to this point between the cells of a slowly dividing, and slowly growing tissue, they are now surrounded by a quickly growing tissue, in the focus of growth of the apex of the root. Accordingly they enlarge here to the diameter of the surrounding cells, and even exceed them sometimes, and terminate with bluntly rounded ends, which often appear swollen, at the limit between the root-apex and the root-cap. The ends of the tubes are easily recognised by means of their dense contents ; their blunt ends—like plastic masses—are easily found, and clearly marked. But there are never seen even traces of septa in course of solution, which would certainly be found here, if the tubes were formed by the disappearance of the walls separating the cells from one another. From the examination of longitudinal sections through the apices of roots of seedlings, it must certainly be concluded that the laticiferous tubes of the root of Euphorbia have an independent apical growth, and grow on continuously at the apex with the other tissues of the root. ‘ At the apical point of the stem the behaviour of the milk-tubes is much more difficult to observe, since they here take a very irregular, crooked course, and there- fore cannot be followed for long distances. Terminations of the tubes are some- times to be found above the youngest leaves, but their connection with tubes lower down could never be proved’: in the nodes a felt of tubes is always formed, from which branches run up towards the apical cone. There is no evidence against the tubes having here a growth fundamentally similar to that in the apex of the root, and none to support the idea that new milk-cells are successively formed in the growing point, which would subsequently grow out to tubes. “It appears as though all the tubes of the Euphorbia plant owe their origin to a process of branching of the original cells formed in the embryo. 1 I was able to prove this repeatedly in E. splendens and trigona.—De Bary. 198 LATICIFEROUS TUBES, ‘In Asclepiadeze and Apocyneze (apparently also in Ficus) the same seems to be the case also in the apex of the root of the germinating seed. The tubes are much narrower and difficult to follow. In the apex of the root they are distributed throughout the cortex. My observations on them are in other respects still very incomplete.’ The origin of the above-mentioned milk-tubes in the secondary bast, which is formed from the cambium, in Ficus, Morus, Broussonetia, Maclura, and Nerium, is not explained by the above observations ; and I am unable, either from published accounts, or from my own observations, to propound any well-founded view whether they arise as branches of those in the primary cortex, and make their way into the secondary bast, or whether they are successively formed anew from the cambium, and in that case are in no direct connection with the primary tubes. Maclura aurantiaca may be especially recommended for the further pursuit of this question. However great the differences in development, above noticed, between the non- articulated and the articulated tubes, still the frequently abundant branching of single members of the latter, and perhaps the coalescence of branches of the former to form anastomoses in the nodes, constitute a transition between the extremes of both categories; and even without these, other common points of structure and dis- tribution, and of functions as indicated by the latter, would certainly lead to uniting both as one sort of tissue. Brief allusion has above been made, in accordance with the views of Dippel and Hanstein, to their near relations to the sieve-tubes: this subject will be returned to in Chap. XII. These relations are mainly physiological, and topographically-anatomical. The assertions as to close histological relations between milk- and sieve-tubes are, I believe, wrong, or at least exaggerated; for instance Vogl’s assertion that both the resin-passages of the Convolvulacez, which do not belong to this category, and the true milk-tubes, e.g. in the Campanulacez, develope from sieve-tubes, or at Jeast may do so. Such a development never occurs. Again, it is especially stated by Dippel, both that the septa in articulated tubes—e.g. in Chelidonium and Papaver—are perforated like sieve-plates; and also that the lateral walls (e.g. Papaver, Cicho- raceze, Carica) are often provided with sieve-plates, and that thus there are in these ‘cases intermediate forms between sieve-tubes and articulated milk-tubes. I could never find this phenomenon, but rather only smooth, delicately outlined pits on the lateral walls, or wide, sometimes grouped perforations. Schmalhausen’s investigations on this point also gave the same result. Further, Dippel’s drawings of these structures show little similarity to true sieve-plates of dicotyledonous plants. The septa often resemble them, it is true, inasmuch as they are perforated by more than one pore ; but the pores are coarse, wide, and irregular, and have none of the other characteristics of structure of sieve-plates (comp. Figs. 80, 81). On the contrary, the sharp difference of structure between the sieve- and milk-tubes is always particularly clear, where one would a priori the most expect to find intermediate forms, i.e. where both are closely side by side, as is the case in the secondary bast. ... The view, so long held, as to the near morphological relations between milk- tubes and sclerenchymatous fibres—bast-cells ’—has been above refuted. -It was excusable that Mirbel at the beginning of this century confused the two organs, and came in 1835 to the conclusion that all the sclerenchymatous fibres of the secondary LATICIFEROUS TUBES, 199 bast are ‘latexiféres*.’ But Schacht’s joy at having got rid of the laticiferous vessels and recognising them as branched ‘bast-cells,’ which he has repeatedly expressed since 1851, was obviously based upon his failure to recognise the true, delicate milk-tubes in Hoya and herbaceous Euphorbias, a failure which had less excuse at his time; further upon the opinion that the latex of these plants was contained in the thick- walled sclerenchymatous fibres, which are often branched, and upon cognate and further confusion of the latter with the well-distinguished thick-walled milk-tubes of Nerium and succulent Euphorbias. Finally, it is undeniable that in the family of the Papaveracez, in the Aroidex, and Musacez, there is a near relation between the milk-tubes and peculiar sacs con- taining colouring matters and tannin. In the first the milk-tubes are absent in the rhizome of Sanguinaria, in Glaucium, Macleya, and the sacs containing colouring matter appear in their place: the other above-named genera are without these, and have laticiferous tubes. All those Aroideze and Musacez which have been examined have tannin-sacs, variously distributed; in certain forms there are also found milk- tubes extremely rich in tannin, in place of which in many Aroidez there are only rows ef tannin-sacs. (On this point compare also Chap. XII.) All these anatomical relations remain inexplicable, till we know better than at present the physiological significance of their different contents. Here we cannot do more than draw attention to them. The anatomical facts above mentioned suggest that under the name Milk- tubes there are at present united two sorts of structures, which do not correspond in function, namely, in the first place those of. the Aroideze and Musacez, containing more especially tannin; on the other those of other milky plants, which have little or no tannin, and which are closely related to the sieve-tubes. 1 Mirbel, Exposition de ma théorie, &c,, Paris, 1809, p. 247, &c.—Idem, Ann. Sci. Nat. 2 sér, tom. III. p. 143. CHAPTER VIL APPENDIX. INTERCELLULAR SPACES. Sct. 49. Between the elements of mature tissue there are often cavities, which are grouped under the term intercellular spaces. These arise in two ways in the original masses of cells, which, at least when in the state of meristem, are always in uninterrupted connection. Firstly, by separation of permanent tissue-elements, as the result of their unequal surface-growth in different directions, the original common walls splitting, while the common limiting layer which was originally present is—perhaps always—dissolved. Secondly, by disorganisation, dissolving, or in many cases rupture of certain “ransitory cells, or groups of cells, which are surrounded by permanent tissues. We may call the first mode of origin schizogenelic, the second lysigenetic, and, if a special term must be adopted for the mechanical rupture, it may be called rhexigenelic. According to the stage of development at which the formation of intercellular spaces takes place, one may, with Frank?, distinguish as profogenetic those which are formed on the first differentiation of tissues, and as Ays/erogenetic those which appear subsequently in old mature tissues. According to their contents intercellular spaces fall into two categories. The first are filled with bodies or mixtures of the same sort as the secretory sacs treated of in Chap. III, or the secretory cavities of the epidermis; they are nearly related to these anatomically and physiologically, and spaces of these two categories not un- frequently appear even to act as substitutes one for the other. They may be designated secretory intercellular spaces, or zwéercellular secretory reservoirs. In treating of them reference has often to be made to the other, not intercellular, secretory organs. The others contain from the very first only air, or in rare cases water. They form together a special ventilating apparatus for the tissues. The stomata of the epidermis (p. 34) are a part of this apparatus; they are a special case of schizogenetic and protogenetic spaces, which usually contain air, but also water in the special cases mentioned. The same intercellular spaces rarely take part in both functions. Thus in the parenchyma of Lysimachia Ephemerum, where the fixed, red, resinous secretion, to be more accurately described below, partly covers the wall of the cells bordering on the ? Beitr. zur Pflanzenphysiologie, p. ror. INTERCELLULAR SECRETORY RESERVOIRS. 201 air space, in some places as a thin layer, in others as thick masses filling the cavity, while again in other places it is entirely absent. The intercellular spaces will be here more exactly treated, following the two main categories as defined by the character of their contents. INTERCELLULAR SECRETORY RESERVOIRS. Sect. 50. Hysterogenetic reservoirs of this category arise in old masses of tissue of long-lived plants from subsequent metamorphosis. Therefore, in order to avoid repetition, their treatment may be passed over for the present, and be resumed in Chaps. XIV and XV. We shall then speak here of protogenetic forms only*. These may be distinguished by their contents, and be named as those which contain vesiz and e¢hereal oil, or mixtures of both, or Balsam, further as containing mixtures of gum, oY mucilage, or gum-resin. According to form on the one hand there may be distinguished elongated, tubular canals or passages, permeating the tissues for long distances, and having a rounded or angular transverse section ; and short, circumscribed, round or rather long, completely closed hollows or cavities, the latter being also indicated by the word of many meanings, viz. g/ands (comp. p. 92), and being further distinguished as internal glands, to avoid confusion with the outer glands belonging to the epidermis. Between the general quality of the secretion and the form of the reservoir there is no generally constant relation; there are passages with balsam or mucilage, and cavities with the same contents, &c. On the other hand, in both cases there is as a rule a very constant, similar character of the reservoirs, according to the families, genera, or species in which they occur, so that they give to these very constant anatomical characters. The Coniferze are a striking exception owing to the variety of form of their resin-reservoirs in the different genera. Unimportant exceptions occur here and there in other families: among the Composite, e.g. Tagetes patula has short closed sacs in the leaves, instead of the passages which are found in the other parts of this plant, and in the leaves of allied species. The same holds for the leaf of Mammea americana, in contradistinction to the other parts of this tree, and the leaves of other Clusiaceze. Some further cases of this sort will be mentioned in the subse- quent special descriptions, p. 206. As already indicated, the reservoirs in question occur only in certain classes, families, and genera, and especially in those which have no other places for the pro- duction and storing of those secretions. A comparison of the statements in Chaps. I and III will make this clear, and will also call to mind the absence of any special secretory organs in the vegetative body of many plants. We may also shortly mention the mutual substitution and relation of the reservoirs in question with the milk-tubes in different genera and species, e.g. in the Composite and Aroidez : this has been above indicated, and wiil be again treated in Chap. XII. In rare cases a mutual substitution is found to take place between sacs and cavities with the same secretion, in different parts of the same plant. This also occurs 2 Compare Frank, /.¢c.—N. Miiller, in Pringsheim’s Jahrb. V. p. 387.—Van Tieghem, Ann. Sci. Nat. § sér. tom. XVI.—[Also Szyszytowicz, Sekret-behilter d. fliichtigen Oele. Ref. Bot. Centralbl. 1881, Bd. 8. p. 259.] 202 INTERCELLULAR SPACES, most prominently in Myrsine africana, and many species of Lysimachia, where the characteristic red resinous secretion occurs in sacs in the root, and in roundish intercellular cavities in the other parts of the plant. The reservoirs in question may be arranged, according to their form and the nature of the secretion, in the following systematic groups. 1. Mucilage- and gum-passages in the Marathacee, many Lycopodia, the Cycads, species of Canna, and Opuntia, and some Araliacee. Mucilage-containing cavities of lysigenetic origin occur in single cases, mentioned above among sacs (p. 144). 2. Resin, ethereal oil, emulsions of gum-resin of different quality, according to the special case, and often little known as regards chemical relations, occur— (2) in passages in the Conifere, Alismacee, and Arodea, the ‘ubifloral Composite, Umbellifera, Araliacee, Pittosporee, many species of Mamillaria, Clusiacee, Anacar- diacee, the genera Azlantus, and Brucea of the family Simarubee. (4) In short cavities in the group Ruiacea, in the sense of Bentham and Hooker (exclusive of Simarubez and Zygophyllez), in species of Hypericum, many species of Oxalis, Myrtacee, Myoporee, species of Lystmachta, Ardisia, and Myrsine: perhaps also in Gossypium. The wail of all secretory spaces, which is made up of that of the neighbouring cells, is, with the exception of the above-mentioned special case of Lysimachia Ephemerum, always completely closed, the lateral walls of the cells, 1.e. those perpendicular to the surface of the space, being joined together without interspaces. When, as is the case in the petiole of Marattiaceze and Cycadez, the original cells, which limit a passage, are separated laterally from one another by continued growth of the surrounding tissue in a peripheral direction, the enclosure is completed by the next outer layer of ‘parenchyma. The number of cells limiting the transverse section of a cavity varies according to the particular case. The small resin-passages, which run transversely between the larger longitudinal ones in the secondary cortex of Cussonia, are, according to N. Miiller, at first at least, slit-like spaces between the partly separated walls of two rows of cells, and are thus limited in transverse sections by two cells. Most of these spaces appear in transverse section to be limited by 3, 4, or many more cells, the number of which may increase as the space widens by divisions placed radially with regard to the space. Fig. 85. The cells surrounding the secretory reservoir have generally the properties of parenchymatous cells. In form they either do not differ, or only slightly, from the cells of the neighbouring parenchyma, e. g. in the young roots of the Composite, in the leaves of Ardisia crenulata: or, as in most cases, they are sharply distinguished, so that one might speak of a peculiar lining of the wall, or epzthed/um of the inter- cellular space. According as the peculiar character extends to one or several layers of cells surrounding the space, there may be distinguished one-, two-, or several-layered epithelium. The enclosure of the several-layered epithelium of the resin-passages in the leaves of Pinus Strobus, sylvestris, Laricio, &c., and of the roots of Philodendron, in a completely. closed sheath of sclerenchymatous fibres, which are wanting in the homologous passages of other closely allied plants, is worthy of notice. The cells of the epithelium are generally prismatic in elongated passages ; their greatest diameter usually lies in the direction of the length of the passage, only in the leaves of Cycads (Kraus, /.¢.) do they lie transversely. Their transverse diameter is usually much INTERCELLULAR SECRETORY RESERVOIRS. 203 smaller than that of the neighbouring parenchyma, so that they appear very different from it in transverse sections: in rare cases they are distinguished from it by their greater width (roots of Composite, branches of many species of Rhus according to Trécul). Their inner surface is often slightly convex towards the passage; in the mucilage- passages of the Marattiaceze it is even elongated and conical ; in those of the leaves of Lycopodium the cells bulge in a club-like manner. Where an epithelium can be distinguished in isodiametric cavities (Lysimachia punctata, and its ally, Myrsine) the cells are flattened towards the surface of the cavities. The wall of the epithelial cells is delicate, in resin- and balsam-passages often coloured brown or yellow. In the mucilage-passages of old leaves of Cycas revoluta alone it is stated by Trécul? that. they are strongly thickened on the side next the passage. FIG, 85,—Hedera Helix ; transverse sections of the young stem (800); g resin-passages. These are in 4, B, C young, and have appeared between four and five rows of cells, in the secondary cortex w 4, at the limit of the zone of thickening c; 4 wood; in Dand £ older, larger passages; 4 bast; ~~ parenchyma of the outer cortex. From Sachs’ Textbook, There is a great lack of investigations on the protoplasm and contents of the cells of the epithelium, and of the surroundings of the secretory reservoirs generally. It seems to be certain for all cases that they contain smail masses of the secretion, which then filter in some way in mass through the membrane into the reservoir. The cells surrounding the young oil passages in the roots of Composite have clear contents, in which, in Helianthus annuus’, large quantities of tannin are shown to be present by reagents; this is also the case with the oil in the passages. In Tagetes patula *, at the transition from the root to the hypocotyledonary stem, a clear violet ’ L'Institut, 1862, p. 315. + Sachs, Botan. Zeitg. 1859, p. 183. 4 Van Tieghem, 4c. p. 113. 204, INTERCELLULAR SPACES. cell-sap makes its appearance in the cells bordering on the passages: further up and throughout the stem orange-yellow granules, which turn blue with iodine, are also found fixed upon the wall of the cells next the passage. In the limiting cells of young passages of the secondary cortex of Pittosporum Tobira, Miiller mentions numerous starch grains with a covering of yellow oil. The epithelial cells of the reservoirs in the leaf of Ginkgo, in the stems of many Composite, e. g. Solidago levi- gata, contain chlorophyll grains: those which surround the round resin-reservoirs in the leaf of Ardisia crenulata are, with exception of the necessary peculiarities of form, in no way different from those of the rest of the chlorophyll parenchyma, to which they belong. In many plants which secrete resin (Coniferze, Anacardiaceze, Umbelliferee, Ara- liaceze, Composite) Miiller has by staining with Alkanna found drops of resin not only in the cells bordering on the reservoirs, but distributed widely in the surrounding tissues. Without wishing in the least to combat the correctness of these observations, I regard further careful investigation of the contents of these cells as the more desirable, since very thin sections, which Miiller states that he used. almost exclusively, are not the most suitable preparations for the study of protoplasm and cell contents. The contents of the intercellular secretory reservoirs form in most cases a homo- geneous fluid mass, or an emulsion-like mixture without characteristic peculiarities of structure. Their chemical properties are indicated approximately, and for the present sufficiently, by the names above used. The red secretion in the cavities of species of Lysimachia and Myrsinez, which, from its solubility in alcohol, may, till further in- vestigated, be classed with the resins, is an exception to this rule, since it appears in a fixed form, and with apparently crystalline structure, as will be described below. The same seems to be the case, judging from incomplete investigation, in many species of Oxalis. The large mucilage-passages, 4™™ wide, of the Opuntias are characterised by containing numerous and large crystals of calcium oxalate embedded in mucilage. As regards the material characters of the milky contents of the passages of Mamillaria angularis, and its allies, which will be more accurately described below, I have come to no certain conclusion. The mode of development of the intercellular secretory reservoirs is for the Jysz- genetic forms generally as follows: that, in a group of delicate cells, which arise by definite meristematic divisions, and which correspond in form and arrangement to the . future reservoir, the secretion appears at the expense of the original protoplasmic body, and that subsequently the walls of the cells are dissolved, and the separate secretory masses coalesce. Of reservoirs of this category, containing gum and mucilage, only the passages of the periphery of the petiole of the Marattiaces have as yet been carefully investigated, and in their case it is shown, that the elements of the simple row of cells corresponding to the subsequent cavity of the passage are filled with the secretion before they are broken up. Nothing is known of the form in which they first appear. In the investigated reservoirs of ethereal oil and resin the secretion first appears as small drops in the protoplasm of the cells about to be broken down. These increase quickly in size and number, and coalesce after the disappearance of the walls, to larger masses. Where the original cellular body consists of several INTERCELLULAR SECRETORY RESERVOIRS. 205 “ layers, the process of breaking down and of coalescence progresses centrifugally (comp. Fig. 86, and above, p. 69, Fig. 22). This mode of development holds for the gum-passages in the periphery of the petiole of the Marattiacew, for the mucilage-passages of the Opuntias, which require further investigation, and perhaps also of the Mamillarias ; again for all secretory cavities investigated, with the exception of those of the Lysimachias, Myrsinez, and species of Oxalis. Doubtful cases will be named below. The schizogenehc spaces (comp. Fig. 85, p. 203) appear sometimes between cells, which resemble those surrounding cells which do not border on secretory reservoirs, both in arrangement and origin: sometimes they are produced by peculiar divisions of special initial meristem-cells. The first case is found to occur in the above-mentioned spaces of Lysimachia Ephemerum, and in the slit-like transverse passages of Cussonia also noted above. Further, the large longitudinal passages in the secondary cortex of the same plant and of other ligneous plants arise between the common .corners of junction of four rows of cells, which are produced from the cambium in the same way (Chap. XIV) as its other products. The same holds, with many special modifications it is true, for the formation of the resin-passages in the secondary wood of the Abietinez’. The prismatic longitudinal passages at the inner limit of the primary cortex in the roots of Composite are, excepting in the appear- ance of the characteristic contents, formed in the simplest and most common case, in the same way as those air-containing cavities found in the external layers of the cortical parenchyma, viz. at the corners of contact of four rows of cells. On the other hand, the primary passages in the pericambium of the roots of the Umbelliferee, which will be described later (Chap. XIII), are the result of special meri- stematic divisions, The passages in the primary cortex of the Abietine, and in the leaves of Cycas and Alisma, may each be traced back, according to Frank and N. Miiller, to a row of initial cells, which divide longitudinally by successive crossed walls: the four daughter cells then separate at the angle of contact to form the passage. As the plant grows there occurs, as has been shortly noted above, a widening of the passage, and growth of the cells which enclose it, in a direction tangential to its surface: this is accompanied by increase in number of these cells by radial division, e.g. from the four original cells to 6 or 8 in Pinus and Alisma (Frank); in cases of long- continued growth in thickness of the part, and correspondingly great widening of the passages, the number may reach much higher figures, e.g. cortex of Coniferee, Rhus, Pittosporum, &c. On the other hand the cells surrounding the space may also divide in a tangential direction, and the originally single limiting or epithelial layer may be doubled, or divided into several layers, e.g. in the passages of species of Philodendron, the cortex of Pittosporum, Hedera (Fig. 85), and the leaves of Pinus. This mode of origin of several layers of epithelium from the originally simple limiting layer is not proved for all cases where they occur, and another origin is in many cases quite possible for the outer layers. The origin of the secretion contained in the schizogenetic spaces, looking at it from the purely histological side, and neglecting the chemical questions, is in my opinion not clear, and requires further investigation. It is obvious that it, or at least 1 Sanio, in Pringsheim’s Jahrb, IX. p. 99. 206 INTERCELLULAR SPACES, the material for its formation, must be derived from the cells, which closely or immediately surround the space. Where it is possible, as e.g. in the case of the above-mentioned resinous secretion around the reservoirs, to prove its presence in the cell-contents, it may be assumed that it passes as such from the cells into the reser- voirs, of course not by filtration in great masses, but, as Miiller assumes, by diffusion through the membranes in successive small quantities. This view is supported by the fact that, according to Miiller in the Coniferze, and Sachs and van Tieghem in the roots of Composite, the intercellular spaces are first present without the characteristic secretion, and that this first appears in them at a later stage. On the other hand there occur, as above stated, cases where the secretion is also of a resinous nature, but is not proved to occur as such in the cell-contents in the vicinity of the reservoir. Further, Sanio has recently expressly stated that the resin- passages of Pinus are filled with the secretion even from the time of their first origin}. So far as my investigations extend, the young stage of the passages of the root of Compositze, when the secretion is absent, is at best very transitory: the secretion appears very early, and may be easily overlooked in the narrow young passages: it may be really absent, especially in transverse sections, since it may have flowed out. Since, in the dermal glands of the epidermis, secretions, which are in every re- spect similar to those under consideration, are often to be first observed anatomically as constituents of the cell-wall, and further since the intra-mural glands (p. 96), when regarded purely histologically, are merely a special case of schizogenetic secretory cavities in the epidermis, the question arises, whether in all cases the secretions of schizogenetic reservoirs are not to be regarded as constituents of the cell-wall. The actual observations supporting this view are of at least equal weight to those supporting the other, while none of the latter exclude the correctness of the former view. All secretory passages, with the exception of the few named on p. 205, are of schizogenetic origin ; of these however the mucilage-containing passages in Canna must be more exactly investigated; also besides those cavities already mentioned in Lysimachia, Myrsinez, and Oxalis, those which take the place of passages in many short leaves of Conifers deserve further attention. Further descriptions in detail of the secretory passages must so often have reference to their arrangement, that, to avoid repetition, they must be given subsequently in Chap. XIII. Here those which are found in species of Mamillaria will alone be described. In the literature I find it only briefly mentioned that these plants have latex (De Candolle), which is contained in passages (Unger). Investigation shows, firstly, that the Mamillarias are entirely withcut those mucilage-sacs (p. 143) which occur in the allied genera. I also find no trace of intercellular secretory reservoirs in small species, as M. glochidiata and the like; and in an unnamed, very robust species. On the contrary, M. angularis, Hystrix, and Zuccariniana have a complex system of branched passages. These are limited, as seen in transverse section, by one 4-5 celled layer, or by two, or even three concentric layers of delicate cells, flattened tangentially to the passage: the width of the passages is about equal to that of one large cell of the parenchyma. They contain a thick, uniformly finely- granular, colourless juice, which emerges on section in large white drops, and hardens ‘Le. p. 101, INTERCELLULAR SECRETORY RESERVOIRS. 207 quickly in the air without change of colour. This must be a peculiar mixture: water, alcohol, ether, benzine, and alkalies do not alter it on the whole, though each reagent may dissolve a small quantity of it. When burnt it leaves behind a very small residue of ash. I have not obtained a clear idea of the mode of origin of the passages of the Mamillarias. The crystal-containing passages of the Opuntias, which may attain a width of 3™™, are apparently of lysigenetic origin; in the mass of mucilage the cells, from the dis- organisation of which they arise, are still partially recognisable. Some details concerning the secretory cavities remain to be added, and statements may also be looked for regarding their occurrence and arrangement. (a) The Myrtaces, judging from observations on numerous species of Eucalyptus, Melaleuca, Callistemon, Eugenia, and Myrtus, are generally provided with oi/-cavities. In horizontal leaves these are particularly numerous on the upper surface, though not confined to it, and their epithelial layer is contiguous with the epidermis, in which those cells which touch the epithelium differ from the rest in form and size. In Myrtus communis, for instance, two semicircular epidermal cells are contiguous with the wall of the cavity at the upper side of the leaf: these cells are distinguished from the rest by their lateral walls not being undulated, and by their being only half as high. In the outer cortex of the branches, according to investigations on species of Eucalyptus, they are separated by some few layers of parenchyma from the epidermis. They have an approximately spherical form: the larger ones may be seen with the naked eye as transparent points, others are smaller, c.g. in the leaves of Eugenia australis. The cavity, which is filled with mixtures of oil and resin, is limited by a continuous epithelial layer composed of tabular cells. According to Frank? the cavities in the leaf of Myrtus communis are of schizogenetic origin. A round, thin-walled cell, lying beneath the epidermis, divides successively into 8 octants; these separate at their central point of contact, so as to form directly an intercellular space filled with oil, and this gradually assumes the form of the spherical cavity; the original eight epithelial cells meanwhile stretching tangentially, becoming flattened, and occasionally dividing. This description is contradicted by Martinet’s short statement, according to which the oil-cavities of the Myrtacez arise like those of Citrus in a lysigenetic manner, a view which, though I have not observed their development myself, I must consider to be correct, front the agreement of the mature cavities with those of the Rutacez. (4) The presence of oi/-cavities is a general and characteristic phenomenon in the members of the group Rutacee in the sense of Bentham and Hooker, i.e. the families or divisions of the Rutacez, Diosmex, Boroniex, Zanthoxylez, Flindersiee, Toddaliee (Skimmia), Aurantiacez, Amyridee?. The Simarubex and Zygophyllee are excluded from this group, and have no oil-cavities. The distribution of the organs in question, and their relation to parenchyma and epidermis, is, as far as investigated, the same as that in the Myrtacex. In the stem of Dictamnus and Correa alba they lie directly under the epidermis, in the leaves of species of Agathosma and Diosma especially or exclusively on the under side of the leaf—relations which are found also in the Myrtacex. They also correspond to these in average form and size. Their origin is in all-cases lysigenetic. Even Frank’s drawing for Ptelea trifoliata does not contradict this, though according to his description of the process of development their origin is schizogenetic. Rauter ® gives a very exact history of development of the oil-cavities on the upper surface of the leaf of Dictamnus (Fig. 86). A cavity (A) originates from two cells, one an epidermal cell, the other a cell of the hypodermal parenchyma. The first divides successively into four cells, placed crosswise ’ Beitr. p. 125. 2 Compare Engler, /.c. on Amyris; also Van Tieghem, /.c. p. 173- 3 Trichomgebilde, &c. /.c. p. 21. 208 INTERCELLULAR SPACES. in the epidermal layer; each of these is again divided into an inner cell next the parenchyma, and a superficial cell (d, c). The superficial cells increase further so as to form the portion of the epidermis which covers the cavity (B, c,d). The inner ones take a direct part in the formation of the cavities. The chief part of the cavity originates it is true from the products of division of the primary parenchymatous cell (4, p, p), which divides by alternate horizontal and vertical divisions into numerous daughter cells; these, together with the similarly formed products of division of the inner epidermal cell, form a compact round body of numerous small cells. In the protoplasm of all cells of this body, which are at first very granular, there appear, after the division and growth have ended, more and more numerous drops of ethereal oil ; then the delicate membranes are dissolved, and the oil drops coalesce to large drops (o in C), The process begins in the middle of the body, and proceeds cen- trifugally to its surface. The oil-containing cavity thus formed is limited, with the exception of the epidermis, by cells of the surrounding parenchyma, which are more or less flattened towards the surface of the cavity; these cells, being in uninterrupted con- tact with one another, completely enclose the latter. The cavity in the hairy dermal warts of the Dittany arises, as described p. 69, in the same way. The oil-cavities of Ruta have in the main the same development. That of Citrus differs at most in small secondary peculiarities, which need not be described in detail here. Martinet, who described them, has traced their origin back to a condition in which the transverse section shows three small cells in the epidermal layer, with abundant protoplasm, and beneath these three inner cells. The arrangement of these cells is such, in these youngest cells and in rather later stages (/.c., Fig. 234 and 235), that it is probable that their mode of origin is the same as that described by Rauter for Dictamnus. It can hardly be doubted, after investiga- tion of mature or half-formed stages, that those of the other members of the group of Rutacex have FIG. 86.—Dictamnus Fraxinella; oil-reservoirs fundamentally the same origin. of the upper side of the leaf; transverse sec- tion; C (200) mature; 4 and 2 successive young stages of development (320). Further description in text. After Rauter from Sachs’ Textbook. The cavity is always completely and smoothly iso- lated by the close connection of the cells of the sur- rounding tissue, so that in good sections it may be taken for a single large cell lying between these. As is especially evident in the chloro- phyll-containing parenchyma, the cells of the limiting layer do not differ fundamentally in structure from those of the mass of tissue, in which the cavity lies, Obvious remnants of partially dissolved delicate cell-membranes are not unfrequently to be found; these form a more or less irregular covering to the wall. It is possible that in many cases . the mass of delicate cells, instead of being dissolved, persists in whole or in part. Many figures of Engler appear to point to this. But here, even with tolerably good pre- parations, it is possible to be deceived, since the sections often do not cut the cavity in a median plane, but lay bare portions of their wall, which then appear in surface-view as dense, multicellular bodies. (c) The spots, recognisable with the naked eye as transparent points, in the lamina of Hypericum perforatum, and its allies, are oil-cavities of flattened spherical form, which’ occupy almost the whole space between the portions of epidermis of both leaf-surfaces which cover them, and are separated from the lower epidermis by at most one layer of parenchyma, Their structure, or that of the tissue surrounding them, is fundamentally INTERCELLULAR SECRETORY RESERVOIRS. 209 that described for the group of the Rutacee. Their origin, judging from observations described, can hardly be other than that stated by Martinet, viz. lysigenetic, though Frank describes it as schizogenetic. Cavities of the same sort occur in the superficial parenchyma of the cortex of the stem. In those of the stem of Hypericum balearicum Unger! found papillose and hair-like processes rising from the wall into the cavities. Many species of Hypericum, e.g. H. calycinum, canariense, hircinum, &c., have no transparent spots recognisable with the naked eye: it is uncertain whether the oil- cavities are absent in these cases, or, more probably, are only smaller than in the punctate species, or are in some way hidden, (d) The bodies scattered in the parenchyma of species of Hypericum, which contain violet colouring matter, and are described by the above authors as glands, may here be men- tioned, and recommended for further research: also the similar bodies in Gossypium. In the leaves of some species of Hypericum, these consist of spherical, loose aggregates of round cells; the colouring matter apparently lies between them. In the leaves of Gossypium they are round, undoubtedly lysigenetic cavities, which are filled with the violet colouring matter, soluble with difficulty in alcohol. (e) Among the Myoporee the species of Myoporum have numerous round oil-cavities of unequal size in the leaves and the outer cortex of the branches. The cavities are superficial, and separated only by one or two layers of cells from the epidermis, which is arched convexly outwards, e. g. M. parvifolium: in M. tuberculatum, on the contrary, according to Unger ', they occur in the middle of the chlorophyll-parenchyma of the leaf. ‘They are surrounded by 1-3 layers of flattened cells. As far as investigated, their origin appears to be lysigenetic. (f) In the parenchyma of species of Lysimachia, of Myrsine africana, and Ardisia crenulata a resinous body is found in intercellular spaces ; it is soluble with difficulty in alcohol, easily soluble in ether, of a bright brownish red colour, and forms fixed, often almost brittle masses: its chemical properties require closer investigation. Its mode of distribution differs in special cases, and of those plants investigated it is simplest in Lysimachia Ephemerum. In the root of this plant it appears in most of the usual prismatic, air-containing, intercellular spaces as a finely granular covering on the wall of the adjoining cells: here and there it is interrupted, while in thickness it varies from that of an insignificant layer to such a bulk as to fill up the passage. In the very lacunar parenchyma of the pith and cortex of the stem, it has fundamentally the same distribution, but is less regular, owing to the irregular form of the cavities: in neighbouring cavities it is contained in very unequal quantities, in many not at all. When secreted in quantity, it forms on the cell-wall a convex covering, striated perpendicular to its surface. Finally, it is found in the leaves, in the cavities of the chlorophyll-parenchyma, as thick, irregu- larly shaped masses, with indistinct radial striation: the cavities are surrounded some- times by ordinary cells, sometimes by a more or less distinct layer of flattened cells. The secretion is strongly localised in the leaves and the cortex of Lysimachia punctata ? and of Myrsine, in the leaf of Ardisia, and of most Lysimachias. Here round reservoirs, which appear as points to the naked eye, are found in the parenchyma: they are surrounded by about eight flat, closely connected, chlorophyll-containing cells, and filled with the dense red secretory body, which is in Myrsine radially striated in a remark- able way. The reservoirs arise schizogenetically, and contain the secretion as soon as they are visible. In the cortex of branches of the above Ardisia, the reservoirs are rarely round, usually elongated, spindle-shaped, appearing as little strokes more than 1™™ jn length. In the roots of Lysimachia vulgaris and punctata, and of Myrsine, the secretion does not lie in intercellular spaces, but in single sacs or cells, not distinguished in size or form from the surrounding cells of the parenchyma: in each of these there is a body of the 1 Anatomie und Physiologie, p. 213. 2 P, Moldenhawer, Beitr. p. 162.—Meyen, Secretionsorgane, p. 61. P 210 INTERCELLULAR SPACES. same structure as that in the reservoirs of the leaf, which does not completely fill the cavity, being surrounded, while young, by colourless granular contents (protoplasm ?), In the roots of Ardisia I did not find the red secretion. (g) Many, though not by any means all the species of Oxalis from the Cape and America have, on the under surface of the leaf, and running towards the margin, rather prominent reddish bands, which are mentioned in descriptions as glands or wales. In the species—not exactly defined—which I have investigated, these bands are reservoirs quite similar in the colour, consistency, and radiate structure of the secretory mass, and also in the structure of the surrounding tissue to those of Lysimachia punctata and Ardisia. They lie in the chlorophyll-parenchyma, and are separated by but one layer of its cells from the distended epidermis of the under surface of the leaf. Thorough investigations were not made '. INTERCELLULAR SPACES CONTAINING AIR AND WATER. Sect. gr. Air- and water-containing intercellular spaces occur, on the one hand, in many vascular bundles, and these will be treated of in Chap. VIII; on the other hand, they are a characteristic component of large masses of thin-walled assimilating Parenchyma. (ntercellular spaces are absent only when the parenchyma forms definite sheaths. s The cavities in question extend between all cells, so that each one of the latter borders on one or several. They together form, as will be again mentioned below, a continuous system throughout the plant, which opens into the stomata, where these are present. The spaces sometimes contain water in the vicinity of the water-pores, elsewhere they normally contain azr, i.e. a mixture of gases similar to atmospheric air, in which the proportion of oxygen and carbonic acid varies with the activity of the processes of assimilation and respiration *. The whole volume of the air-spaces varies greatly in special cases, and is often very large in proportion to the volume of the part of the plant which is not filled with air. Approximate measurements, which Unger*® made on leaves and petioles of 41 species of plants, gave as minimum 77 parts by volume of air to 1000 parts of the leaf in Camphora officinalis, and as maximum 713 to 1oco in Pistia texensis, The air of the vessels and that diffused in the cell-sap, which would be also pumped out, is not taken into account in these statements: but it appears that this is on the whole of minor importance, and that on the other hand the figures obtained would be sometimes lower, sometimes much higher, if the single masses of parenchyma com- posing the leaf were investigated separately. It was long ago known that the whole volume of the air-spaces in réltion to ‘that of the whole plant is largest in plants of all classes and families which grow in water, or in moist positions, and on the other hand in those which inhabit dry places, as many Composite, Umbellifere, Labiate, Grasses, &c. with hollow stems or petioles. According to their gradual, and not distinctly limited differences of relative r [See fithien Russow, iiber sekretfiihrende Intercellular-gange der Acanthaceen, &e., or pat. 1880.—Bot. Centralbl. 1881, Bd. 5. p. 365.] * Compare Sachs, Experimentalphysiologie, p. 262. 5 Beitr. z. Physiol. d. Pflanzen. I. oibegsbte d.Wiener Acad. Bd: XIU. p. x6 SPACES CONTAINING AIR AND WATER. 211 width, the air-spaces may be divided into Znéerstitial spaces, such as are of smaller volume than the adjoining elements; cavities, Jacune, of almost equal or slightly larger volume; passages, chambers, and hollows of much greater relative volume. The tissue permeated by air-spaces may accordingly be termed lacunar, chambered, &c., and that with only narrow interstitial spaces may be called (relatively) dense. The air-spaces of dense and lacunar parenchyma always arise schizogenetically at the first commencement of differentiation of the tissues. In preparations seen by transmitted light they may be seen close behind the growing-point, by reason of the air contained in them, as black bands between the cells. Comp. e.g. in Fig. 3, p. 10, the regions marked m and r. ; The interstitial spaces of dense tissue run as a rule between the rounded corners of the cells, which are closely connected over the greater part of their walls, as narrow angular canals, the number of their sides being equal to the number of cells which border on them. Thus, e.g. the numerous three-sided prismatic interstitial air-spaces in regular polyhedral cells arranged in alternating rows, the four-cornered ones between the cells arranged in radial and concentric rows which are not alternate, in the inner primary cortex of many roots (Fig. 51, p. 124). More rarely they form narrow slits, standing serially one above another, and separated one from another by connected portions of the walls, these slits lying between the limiting surfaces of two contiguous cells, as in the dense tissue of the leaf of Myrtacez, species of Scirpus, the parenchymatous lamellze of the cortex of Pilularia and Marsilia: this condition is allied to that of the many- armed lacunar tissue. Of lacunar parenchyma there may be dis- tinguished two chief forms, which however cannot be sharply separated. The first (Fig. 87) arises through the unequal growth in surface of all cells as they pass from the meristematic state, in such a way that at certain points they put out processes which may grow to long arms, at other points these are not formed. The ends of the processes of adjoining cells remain connected with one pe i Sa esac gn eenveree seston another; between the other parts of the sur- [Re, 3 of the circumference of the stem, that of the leaf-trace 1. The number p of the bundles entering one leaf is, as in other investigated species, an uneven number: 5, 7, to 21 and more, Of the p bundles of one leaf the two marginal ones, /./, Fig. 100, converge as they enter the node (1), and unite there at once with the median bundle (m) of the next higher leaf, which descends perpendicularly between them. This united bundle then descends further perpendicularly through the next internode, and forks in the next node (2), close above the median bundle, which passes out at that point, and each of its two shanks unites with the bundle, which descends side by side with it. The course of all other bundles corresponds to the scheme. The number x of the bundles of one leaf-trace in an internode is thus in Foeniculum = p—1, while in cases which follow the scheme * Carried out in 1873,-in the Botanical Institute at Strassburg, by Herr von Kamienski. + COURSE OF THE BUNDLES IN THE STEM. 243 exactly it is= p. The transverse section of the internode of Hydrocotyle shows, e.g. six alternating unequal bundles, if three enter each leaf: that of Foeniculum shows, when the number from each trace is equal, twelve (Fig. 101) if seven enter each leaf, and sixteen if nine bundles enter each leaf. As the result of inequality in number of successive traces in the growing plant, there occur also in Fceniculum deviations from the above special scheme. Those modifications of the course of the bundles, which certainly occur in the flowering shoots of the Umbelli- fere in question, and in many with deviating conformation, have not been investigated. 15. Leaves opposite, the pairs decussating more or less exactly. The bundles of one pair pass perpendicularly down through two internodes, and then curve, in the second lower node, sometimes converging symmetrically, sometimes both turning in the same direction, and then descend further, and unite with those belonging to lower leaves. This condition is plainly seen in the youngest stages only, later a second shank is usually formed at the point of curvature, so that the bundle becomes branched, and sits astride of that directly below it (Fig. 102). Further in many allied cases, the lower ending of the bundles becomes quite indistinct, by their coalescing laterally by means of intermediate bundles which appear very early (Chap. XIV). To this category belong, according to Ndageli and Rohrbach (/.¢.), Fraxinus excelsior, Vinca minor, Apocynum hypericifolium, species of Phlox, Veronica incisa, Calluna vul- garis, Hypericum quadrangulum, Androsemum, Euonymus europxus, Species of Alsine, Spergula, Cerastium, Dianthus and Silene, also Galium and Rubia. Figs. 102 and.103 may illustrate the arrangement for the special case of Cerastium. Fig. 102 is the scheme for the course of the bundles of a shoot with the cylindrical surface reduced to one plane; ab, cd, ef, gh, are the bundles of foliage leaves, the letters standing at their point of exit from the ring. Below the node marked de these bundles of the trace are alone present. Above dc there are also others, viz. p, 0, 2, the bundles of the terminal flower-stalk, and 4i, /m, of which pairs one enters each of the branches in the axils of the leaves g and & (comp. Sect. 94). All these bundles have a place in the ring as shown in Fig. 103, which represents a transverse section through the internode above ef. 16, Leaves in whorls: traces consisting of one bundle, which runs down more than two internodes. Trevirania longifolia, Russelia juncea. 17. Leaves opposite: traces of three or four bundles, which unite at the second lower node with those of the next lower pair: not pectinated. Antirrhinum majus, Ruellia maculata, Bignonia serratifolia, Tecoma radicans, 18. Leaves opposite and decussate. Traces of 2 bun- dles, not pectinated. -Anagallis arvensis, Stachys angusti- He te SM a a) Fig, 102. Fig. 103. FIGS. roz and 103.—Cerastium frigi- dum, after Nageli. Fig 102, Scheme of ‘bundle-arrangement ; explanation in the text.—Fig. 103 (20). Transverse section through a shoot in the internode above e,f of Fig. 102. The letters indicate the same bundles in both figures; e, / are already branched in Fig. 103, in the sheath-like connate bases of their pair of leaves. : E folia, Satureja variegata Hort. (Nageli, /. c.), and many other Labiate, Nepeta Cataria, Melissa officinalis, &c. Two bundles, which are united. in the petiole to form one, separate from one another at once in the stem of the Labiate (Figs. 104 and 105), R2 244 PRIMARY ARRANGEMENT OF TISSUES, : and pass down the corners, between which the leaf is placed, through two inter- nodes. At the second lower node they unite with those of the next lower trace, after having traversed one internode close side by side with these. The transverse section below the apex of the stem thus always shows eight bundles, in groups of two beneath the corners; those of each pair are of unequal size, the stronger belonging to the leaf-trace of the nearer pair of leaves, the weaker to that of the next pair. The bundles of one corner soon unite, as vessels appear between them. z E, The transverse section then shows four bundles, which later unite to form a closed ring (Chap. XIV), b> 19. Leaves opposite, traces of 3 bundles: the lateral I bundles pectinating with those of the next pair. Clematis Vitalba, Viticella; Atragene, Urtica Dodartii, Lonicera spec., Acer pseudoplatanus, Philadelphus coro- narius, Tagetes lucida, T. signata Bartl., Humulus Lupu- lus, Centranthus ruber, Aesculus macrostachya, Euphorbia _® Oy Ne kK ad i Lathyris. ) ( h The foliage shoots of the above plants, though ( corresponding in the above points, differ in the unequal length of the course of the traces. The median bundles ; sometimes insert themselves at the next lower node, sometimes at the next but one, sometimes still lower. The lateral bundles also pass through one, two, or | 2 ZB several sections of the stem. Giving the reference to F ( ‘ J” Nageli’s work, we will here describe only the very f simple examples of Clematis and Atragene (Figs. 106, 107). The pairs of leaves decussate at right angles. The six corners of the internodes, of which two opposite ones corresponding to the median points of the leaves are rather more prominent, alternate regularly, The width of the three-bundled leaf-trace is about 115°. The median bundles (ad, gk, gn, xt) pass through one internode, divide at the next node into two shanks, and insert themselves with these on the lateral bundles of the pair of leaves at that point. At first there is always but one shank present, and the two median bundles of the same pair have (according to two observations) a symmetrically converging curvature. The formation of the second shank often appears at a ieee in, coon late stage in Cl. Viticella, or is entirely absent. after Ndgeli, Fig. 104. Scheme of the vases The two lateral bundles of the leaf (dc, a bi, lar system in the end of the shoot, the cylin- lm, &c.) also pass through one internode, they curve drical surface being exposed in one plane. ab, de, fe, gh, tk, the traces of successive in vergi . pairs of leaves, the letters being plared at the a converging manner at the next node, and insert nodes from the Sabet ‘ats only one bundle themselves on those same lateral bundles, with which gs). Transverse section through a young im the shanks of the median bundle unite. In Cl. Viticella ee raga eet spreetis fhe leaf-trace in this condition is usually complete: in cated by the same letters as in the above. Cl. Vitalba a second shank is formed at the point of curvature of the lateral bundles also: this curves to the opposite side, and coalesces with a median bundle of the node. The trans- verse oc of the young internode shows six bundles of the leaf-trace (Fig. 107, p- 246). The axillary branches have also six bundles in their lowest internode, which unite. to two on entering the stem. These two unite at once right and left with the median bundle of the leaf which bears them, Fig. 104, Fig. 105. COURSE OF THE BUNDLES IN THE STEM. 24.5 20. Leaves opposite: traces consisting of three bundles, the lateral bundles of the same pair united from the first. Mercurialis annua and M. perennis. 21. Leaves opposite: traces consisting of five bundles, the two lateral bundles of the same pair united from the first. B. GYMNOSPERMS!. As has been above mentioned, the vas- cular system in the stems of the Conifere does not differ from that of the Dicoty- ledons: they will therefore be mentioned here only as special instances of the type of the Dicotyledons. The seedling of most of them has two opposite cotyledons which grow green on germination, and rise above the ground; rarely (Ginkgo, Araucaria, section Colum- bea) they remain in the ground. More than two occur exceptionally in many genera, and constantly in Taxodium (4 to 9) and in the Abietinez, in the sense of Strasburger, i.e. Linnzus’ genus Pinus. The number of the cotyledons differs here according to the species, and varies in the same species within wide limits: e.g. in Abies pectinata between 4 and 7, in Pinus sylvestris between 3 and 8, in Pinus Pinea between 8 and 14. Discounting single exceptions to be named below, one bundle enters the short hypocotyledonary section from each cotyledon; where there are two cotyledons both bundles run vertically downwards, and soon undergo coalescence to form the root-bundle: where the num- ber is higher, two or three bundles coalesce immediately after entering the hypocoty- ledonary section to form one, so that the number of the bundles of the trace in the latter is smaller than that of the coty- ledons. Statements by Lestiboudois ((.c., pp. 25 and 26) lead to the assumption that in Cupressus pyramidalis and Abies bal- samea the bundle, which passes from the cotyledon into the stem as a simple bundle, splits at the node into two shanks, and that the opposite shanks of two neighbouring bundles unite into one, which (alternating with each pair of coty- ledons) descends perpendicularly. The cotyledons of Araucaria brasiliensis? have Sambucus nigra. FIG, 106 (40).—Clematis Viticella, after Nageli. End of a branch made transparent by removing the surface and treatment with potash, showing the course of the leaf-traces. The outgoing ends of the bundles are somewhat distorted by slight pressure; the two highest pair of leaves a @ and y § have as yet no developed bundles. each 8 vascular bundles, and these each unite in the cotyledonary node to two, so 1 Nageli, Z.c—Lestiboudois, /.c.~A. B, Frank, Botan. Zeitg’ 1864, p. 150.—Geyler, Gefiss- biindelverlauf in d. Laubblattregion d. Coniferen, Pringsheim’s Jahrb. VI.—Strasburger,”: Die Coni- feren und Gnetaceen. 2 Strasburger, /.¢. p. 369. ’ 246 PRIMARY ARRANGEMENT OF TISSUES. that from the two cotyledons four bundles of the trace pass down into the hypocoty- ledonary section. In all investigated cases the bundles of the trace of the first epicotyledonary leaves insert themselves on the cotyledonary bundles at or close below the cotyledonary node. With the single exception of Ginkgo, the trace of foliage and scale leaves passing FIG. 107 (20).—Clematis Viticella. Transverse section through a young internode; further explanation in the text. FIG. 108.—Juniperus nana, after Geyler, 4 scheme of the longitu- dinal course of the bundles, the cylindrical surface being reduced to one plane; the whorls of three members are slightly displaced spi- rally; 4= bundles of the buds. B (16) transverse section through a young shoot; 1, 2, 3, the bundles, which pass out in the order of the figures to a whorl; 2 the bundles which pass into the axillary buds. through the stem is in the Conifere a single bundle; and this is still the case even where the leaves have several bundles, and where these arise, as in Dammara and the broad-leaved Araucarias, by splitting of the single bundle of the trace while still in the node. In the investigated species of Juniperus, Frenela, Cupressus, Callitris, Libocedrus, Thuja gigantea Nutt, Chamecyparis eri- - coides Hort., the leaves are arranged in two- or many-membered alternating whorls, Their traces consisting of a single bundle descend undivided through one internode, and fork about the middle of the second internode into two shanks, which insert themselves right and left upon the bundles of the trace of this internode (Fig. 108). Thuja occidentalis, Th. plicata, and Biota orientalis have’ on the other hand, it is true, the same whorls of two alternating leaves as their nearest allies above named; but the opposite traces of each pair of leaves pass perpendicularly downwards through two internodes without dividing; they then curve over the leaf-trace, which emerges at the znd lower node, both turn- ing in the same direction (more rarely converging symmetrically), and affix themselves laterally on the bundles which emerge at the and, 3rd, or rarely the 4th lower node (Fig. 109). In the numerous Conifere with spirally arranged leaves, Chamecyparis glauca Hort., Widdringtonia juniperina, the in- vestigated species of Taxodium, Glyptostrobus, Cryptomeria, Sequoia, Cunninghamia, Pinus in the sense of Linnzus, Podo- carpus, Saxegothea, Taxus, and Araucaria, the one-bundled leaf- traces have a course corresponding in its main points to the scheme described for Iberis (Fig. 110), Each bundle descends independently through a definite number of internodes, and then curving towards a definite lower bundle, affixes itself laterally on it, and coalesces with it further down. The number of the definite lower bundle, on which it is affixed, varies according to the individual case, but is constant for each single case, and follows the series 2, 3, 5, 8, 13, 21,.... The direction in which the apposition on the coalescing bundle takes place, is also con- stant for each individual case, and’ is defined by the number of the coalescing bundle, so that the apposition on the 3rd, 8th, or 21st lower bundle is in a descending direction, that on the 5th, 13th, or 34th in an ascending direction with reference to the leaf-spiral (Geyler, /.c.). The same rule above described is followed as regards the course of the bundles of the trace in the leaves of Cephalotaxus Fortunei, Torreya grandis, and, as far as at present known, of Dammara australis: these leaves appear at first in whorls, but are subsequently displaced spirally. Ginkgo also belongs to this category, since the two bundles of the trace, after running separately through 1-3 internodes, unite to one, which curves in a kathodic direction above the 5th lower trace, and inserts itself in about the 8th lower internode in an anodic direction on the 5th lower trace, with which it pursues a united course in the 9-11 internode. Among the Gnetacen, Ephedra vulgaris has in each of the two cotyledons two vascular COURSE OF THE BUNDLES IN THE STEM, 247 bundles, these enter the hypocotyledonary stem, which thus contains four bundles of the cotyledonary trace, The epicotyledonary section contains eight bundles, four opposite each cotyledon. In the cotyledonary node these eight bundles unite in pairs, each of which passes down between two cotyledonary bundles, and splits within the hypocotyledonary section into two shanks, which insert themselves on the next cotyledonary bundles. Below this point of union the four cotyledonary bundles unite to form the bundle of the root. Each of the other leaves of Ephedra, which, as is well known, are scale-like, and are arranged in accurately alternating whorls of two, also contains two bundles. In E. vulgaris the two-bundled leaf-trace enters the ring of bundles at its own node (1) ; takes a perpen- A a7 718 33 20 25 22 49 ee \ 1 20| 1s \e Ww 9 | |a \79 o ‘ea <0 4 4 a | a FIG. 110.—Pinus sylvestris, after Geyler. Scheme of the vascular system in the young shoot, the cylindrical surface being reduced to a single plane; leaves arranged 8/21, in a ‘right-handed spiral. The figures indicate the successive bun- dles of the leaf-trace, which are represented as broad bands. Each pair of converging bundles (represented as thin lines),. near the emerging bundles o—9, goes to an axillary shoot. The traces unite in descending order, each with the eighth lower one. FIG. 109.—Thuja plicata, after Geyler. 4 scheme of the bundle system, the cylindrical surface being reduced to one plane, 2 (x6) trans- verse section through a young shoot. 1, 1 the bundles, which pass directly into a pair of leaves; outside each, below the surface, is a resin pas- sage, A. dicular and parallel course downwards through two internodes, and inserts itself in the third node laterally on the trace which emerges at node 2, each bundle joining with that laterally next it. In the node there appears at an early stage a transverse girdle of tracheides, which unites the bundles. Strasburger states for Ephedra campylopoda, that between the two bundles of the trace. of each leaf there runs a ‘ complementary bundle’ which arises from the girdle of tracheides: this passes through one internode, from the node of the pair of leaves, to the next transverse girdle. In Ephedra altissima, according to the same author, the two bundles of the trace of a leaf take a separate course only in their own internode, and in the next are united to a single bundle. E. vulgaris has therefore eight bundles of the trace in the internode, of these two opposite pairs belong to the same pair of leaves : E, campylopoda has ten, E. altissima only six. The species of Guetum have on their foliage shoots decussating pairs of leaves, separated from one another by elongated internodes. Each leaf contains four or five bundles, accord- ing to the species'. According to some few investigations on Gn. Thoa the leaf-traces follow the above-described scheme for the Umbellifere. The ten-bundled trace of each pair of leaves passes down through two internodes, and unites in the second node with the 1 Strasburger, /.¢. p, 115. 248 PRIMARY ARRANGEMENT OF TISSUES. trace of the next lower pair. The bundles of the trace of successive pairs all pectinate with one another. Besides this, other coalescences must occur at or below the node, since in the internodes investigated the transverse section showed only eighteen bundles instead of twenty as assumed according to the scheme, and as actually occurs in Gn. Gnemon. On Welwitschia comp. Chap. XVI. : To avoid repetition, the Cycadee will also be described in the chapter cited. II. Anomatous DicoTyLEpDons. Secr. 62. A not irconsiderable number of Dicotyledons; some Cycadez, and Welwitschia differ in their bundle-system from that which characterises their allies, in that the primary bundles are not arranged in a simple ring. Either they contain a ring of bundles arranged according to the usual type, while additional bundles are found either within these, that is in the pith, or outside them, that is in the outer cortex; or the bundles are arranged in several, often not sharply distinguished circles, or so arranged that they appear in transverse section irregularly scattered between other tissues, with the exception of the most peripheral bundles, which may be distinguished as a ring well marked off from the outer cortex. These more or less remarkable exceptions to the main type either occur in quite isolated species in typically formed genera and families (e.g. Umbelliferz), or in numerous species of otherwise typically formed genera (e.g. Begonia), or they are characteristic of certain genera, or small families (e.g. Nymphzeaceze, Calycanthacez, Podophyllum, Diphylleja), more rarely even of large families, as Piperaceze and Melastomacez. But even in the latter, exceptions occur to that grouping of the bundles which holds for the majority of their allies. The above phenomena are based either upon an oblique radial course of the bundles of the leaf-trace, or upon the appearance of cauline bundles besides the bundles of the trace which are arranged in the typical ring. Discounting the Nyctaginez, which will be described below (Chap. XVI), many Amarantaceze, &c. with medul- lary bundles, the following cases belong to this category. a. Medullary bundles. 1. All bundles belong to the leaf-trace : some on entering the stem are arranged in the typical ring, and have a radially perpendicular course in it: others, passing further inwards, are therefore medullary, and are either.scattered through the pith or arranged in rings. To this class belong most Cucurbitacee, species of Amarantus and Euxolus, Phytolacca dioica, the Piperacee, doubtless also the herbaceous Ber- beridee, Podophyllum, Diphylleyja, Leontice ; further species of Papaver Thalicirum, and Aciea. The bundles of the climbing Cucurbitacez’ (Cucumis, Cucurbita, Bryonia, Tladiantha, Cyclanthera pedata) are arranged in two rings (Ecbalium elaterium, which has no tendrils, has only one circle of bundles): those of the outer ring are opposite the corners of the stem, and are of equal number with these, e.g. five in Cucumis sativus, Cucurbita, Tladi- antha dubia, Cyclanthera pedata, seven in Bryonia dioica: those of the inner ring alternate ' Bernhardi, Beobacht. tiber PAanzengefasse, p. 20.—Sanio, Botan. Zeitg. 1864, p. 227.— Nageli, 7c. p. 77. COURSE OF THE BUNDLES IN THE STEM.. 249 with those of the outer, so that their outer portion falls between the latter, but their number is not always the same as that of the outer series, one being left out (e.g. 4 in the specimens at hand of Tladiantha). The bundles of: both rings are, as far as investigated, bundles of the trace, which pass downwards on the average through two internodes: it remains for further investigations to ascertain their course exactly, a matter which is made very difficult by the early ap- pearance of irregular transverse anastomosing bundles in the nodes. Of those Amarantacez, which have been investigated, some, namely species of Celosia, Gomphrena, Alternanthera, Freelichia, and Achyranthes, have the primary ring of bundles, and pith cylinder (the latter without bundles), which are typical of the Dicoty- ledons. In Amarantus? caudatus, A. retroflexus, and Euxolus emarginatus A. Br. the numerous (e. g. 11) bundles, which are arranged in the base of the leaf in a curved series concave upwards, separate from one another in the node, taking however a steep down- ward course: some form a ring with certain bundles, which descend from above, others penetrate deeper into the pith. Thus there are formed within the ring of bundles several irregular medullary rings, in which the bundles belonging to individual leaves remain grouped together. The middle of the pith has no bundles. The median bundles of each leaf-trace appear to penetrate deepest into the pith. Lower down the bundles of one trace unite, after traversing several internodes separately. An accurate investi- gation of their course remains still to be made. Slender specimens of Euxolus lividus Mog. show similar, but simpler conditions. Phytolacca dioica has (according to Nageli, /.c. p. 118) three bundles of the trace for each leaf. The two lateral ones descend in a radially perpendicular direction in the stem between the pith and outer cortex; they split first into two, then into several shanks, and these together form the ring of bundles. The median bundle enters the pith, but hardly deeper than 2 the radius of the pith, it there descends through 8-12 internodes, and then again unites with the ring. It describes a curve convex inwards, the strongest curvature of which is in its upper part: it reaches nearest to the middle of the pith in the 3rd and 4th internodes, The transverse section through a mature internode thus shows 8-12 bundles, which are free within the ring. The course of the bundles, which in the aerial stem of Podophyllum, Diphylleja, and Leontice? are distributed in the transverse section throughout the pith, almost as in the Monocotyledons, and of those which, in Papaver orientale (often also in P. somniferum), in Actza racemosa, Cimicifuga foetida, and species of Thalictrum, form an irregular 2- or 3-seriate zone round the pith, requires further investigation: it can hardly be doubted that they belong to the leaf-trace. Further investigation must show whether the medullary bundles, which occur in Statice or the Plumbaginex*®, belong to this series: I found no such bundles in the species of Statice, Armeria, and Plumbago which I examined. Since P. Moldenhawer and E. Meyer it has been known that in the internodes of all Piperacee which have been investigated, with the exception of Verhuellia, there are medullary bundles, usually arranged in a circle, within a ring of bundles, which in the woody species (Piperex) subsequently undergo secondary thickening. Rarely more than one inner circle is present, e.g. in Peperomia variegata sometimes 2, in P, incana and obtusifolia 3-4, in Piper geniculatum z, in Artanthe cordifolia 4*. The number of the 1 Link, Grundlehren der Anatom. und Physiol. der Pflanzen, pp. 144, 148.—Unger, Dicotyle- donenstamm, p. 108. 2 C. H. Schultz, Vaisseaux du latex, &c., Mém., prés. Acad. d. Sciences, VII (1841).—Sanio, 2, €. DP. 230. ; 3 Russow, Vergl. Unters. p. 153.—Schwendener, Mech. Princip. p. 143. 4 P. Moldenhawer, Beitr. p. 5.—E. Meyer, De Houttaynia et Saurureis, p. 39.—Unger, Bau, &c. des Dicotyledonenstammes, p. 68, &c,—Karsten, Veget. Org. d. Palmen, Ac. p. 148.—C. de Candolle, Mémoire sur Ja famille des Pipéracées, Mém. Soc. phys. de Genéve, Taf. XVIII. p. 2.— 250 PRIMARY ARRANGEMENT OF TISSUES. bundles, both of the inner and outer circle, varies, at least in many species, in the suc- - cessive internodes of the same shoot. ‘They run perpendicularly down the individual internode. In most species transverse anastomosis in the node makes it difficult to follow their further course. Most recent authors have arrived at the result that the bundles are partly common, partly, and especially the inner ones, cauline: the in- florescence certainly contains cauline bundles. Karsten alone, in the year 1847, ex- pressed another view for the Piperacez, according to which all the bundles are bundles of the leaf-trace: this view is confirmed and generalised by the more thorough investiga- tion's of Weiss, The latter can only be epitomised here, reference being given to the original work, since the latter only appeared after this book had begun to pass through the press. Peperomia galioides shows the course most clearly. The leaves are arranged in whorls of five, each has a single bundle. The bundles enter the node in the outer circle, and pass down in it through one internode: they then curve inwards, and, forming the inner circle, pass down the second internode: in the node below the latter they insert themselves on the bundles of the next higher whorl, which here curve into the pith. The transverse section through the internode shows two concentric series of five bundles each. P. brachyphylla has decussating whorls of two leaves each. Each leaf contains three bundles, one median and two lateral. ‘The median bundles run in the peripheral circle through two internodes, then turn inwards, and, after a further course in the pith through one internode, insert themselves with their tapering ends on a medullary bundle. All lateral bundles of the trace run through one internode in the peripheral circle, then curving in the next lower internode into the pith, they run further in the pith through one internode, and insert themselves, also with tapering ends, on the medullary bundles of the third internode.’ In each node accordingly six bundles pass inwards; in the pith of the internode however there are only four: certain bundles must therefore unite on enter- ing the pith. Weiss found the course similar, but more complicated and less regular, in P, rubella, and in P. variegata and incana, which have alternating leaves: of these the former has a leaf-trace of twelve bundles, the latter of seven. Weiss, in common with Karsten, found in the woody Piperacex (Piper, Artanthe, species of Chavica) that the bundles of the many-bundled leaf-trace, which embraces the stem, descend, where there are one or two medullary circles, through at least one internode in the peripheral circle; then curving into the pith, they traverse a second internode in the medullary circle, and finally insert themselves on medullary bundles of a lower internode. On passing from the outer to the inner circle two or three bundles may unite, and the number of the medullary bundles may thus vary with little regularity in the successive internodes. Where there are more than two circles (Artanthe cordifolia) the medullary bundles traverse at least two internodes. It is instructive that the course of the vascular bundles in the Piperacez closely re- sembles that of the Commelinez (§ 69). 2. All bundles belong to the leaf-trace. After entering the stem they pass over into a network of bundles, which branches irregularly on all sides. To this series belong the Nymphzacez, Gunneracez, Primula Auricula, and its nearest allies, perhaps also many Balanophorez, In the first three groups a definite number of bundles are seen to enter the stem from each leaf, and immediately after entering they pass over into a network of bundles, which are irregularly connected both in the direction of the surface of the stem, and also in radial planes by oblique and Sanio, Botan. Zeitg. 1864, p. 193.—F. Schmitz, das Fibrovasalsystem d. Piperaceen, Diss, Essen, 1871.—J. Weiss, Wachsthumsverh., &c. d. Piperaceen, Flora, 1876. COURSE OF THE BUNDLES IN. THE STEM. 250 transverse anastomoses, the network being still further complicated by the insertion of the bundles of roots and buds. Transverse and longitudinal sections through the stem show ‘irregularly scattered’ bundles, cut through in different directions; the former thus.remind one superficially of transverse sections of stems of Mokecotyte. dons, of which however only those of the Aroideze with irregular reticulate con- nection of bundles can be more closely compared with them. The structure of species of Gunnera has been more exactly investigated by Reinke !, In G, Chilensis Lam, (G. scabra R. B.) one bundle of the trace enters the short hypo- cotyledonary stem from each cotyledon: the two unite, after a perpendicular course, to form the axile root-bundle. The cotyledons are immediately followed by a pair of almost opposite primordial leaves, which decussate with them: then come the further leaves in spiral arrangement, all of them being separated from one another by extremely . short internodes. From each of the primordial leaves three bundles, which unite in the . cotyledonary node, enter the hypocotyledonary stem: here the two united traces alter- nate with those ofthe cotyledons, and unite with them lower down to form the axile bundle. Exactly at their point of entry into the centre of the stem they are mutually united by a horizontal bundle, parallel to the surface of the stem, and by one or a few, which traverse the middle of the stem obliquely. Branches from the latter pass to the cotyledonary bundles. From each of the leaves, which next follow the two primordial leaves, three, bundles enter the stem, and from the successively higher ones a larger number (not exactly stated). Each successive leaf-trace resembles the first, inasmuch as immediately on entering the stem it is connected with the network of bundles by con- necting bundles in all directions: only the number of bundles of every sort and direction increases in proportion as the axis of the seedling 2 mm. thick increases to the swollen stem of 50mm. thickness. The appearance of the bundles of all categories takes place almost simultaneously. In conformation and structure the following species coincide with G. Chilensis: viz. G. petaloidea, bracheata, insignis, commutata, peltata, manicata. G. perpensa L., to which is allied G. macrophylla, has in its stem of 1°™ in thickness, and with rather longer internodes, for the most part bundles with a longitudinal course: most of these are united into a hollow cylindrical net, within the parenchymatous cortex, and have transverse and oblique connecting bundles, which traverse the pith. In the short internodes of the leafy stem of G, magellanica, which is 2-3 mm. thick, 3-4 vascular bundles run longitudinally: these are connected, diréctly with one another, by convergence, to form pointed elongated meshes, and also at the modes they are united by transverse anastomoses, at the points of entry of the three-bundled leaf-trace. The elongated internodes of the stolons of this plant have usually only one concave band- shaped axile bundle: sometimes this splits for a short distance into two. Finally the thin stems of G. monoica and prorepens show in their internodes usually two bundles, which are here and there united to a single one, into which the (one-bundled ?) leaf-traces run. In the elongated internodes of the stolons they have one axile bundle. In Primula Auricula® the one-bundled leaf-traces of the cotyledons and of the first leaves unite, running almost horizontally into the middle of the stem, to form one axile bundle which traverses it. The bundles which enter from the subsequent leaves run for a distance—not defined by any constant number of internodes—side by side, and then unite with one another, or with the axile bundle. As the plant grows stronger the number of bundles entering one leaf increases to twenty, these pass obliquely down the stem, and are here united by irregular and oblique branches and anastomoses, which run * Morphologische Abhandlungen, Leipzig, 1873, p. 47, Taf. 4-7. 2 Vaupell, Ueber d. peripherische Wachsthum d. dicotyled. Rhizome, Liste, 1855.—Von Kamienski, Zur Vergl. Anatomie d. Primeln, Diss. Strassburg, 1875. 252 PRIMARY ARRANGEMENT OF TISSUES. radially and tangentially. In transverse section there appears a ring of 15-20, widely separated, stronger bundles, which correspond to the central bundles of the base of the leaf: surrounding the ring numerous smaller bundles are irregularly scattered; these are derived from the lateral ones of the base of the leaf: in the space within the ring are seen the transverse sections of the connecting branches, which here also run in ll directions, Pr. Palinuri, calycina, and marginata resemble Pr. Auricula. Other species of Primula, as Pr. sinensis, spectabilis, and elatior, have a typical Dicotyledonous ring of bundles, which are very soon united laterally by intermediate bundles, On their special course, and the peculiarities of many species, especially Pr. farinosa, compare Kamienski’s work. In the Nymphwzacex the system of vascular bundles of the stem (Rhizome) is usually a network of anastomosing bundles difficult to disentangle, from which those for the leaves, roots, and peduncles branch off at certain places, and which in stronger stems, e.g. of Nuphar luteum, traverse the whole internal part of the stem, which lies within a sharply limited cortex, even to its very centre. It may be seen from Unger’s figure’ how chaotic this structure is. Nageli (/.c. 121) tried to explain the matter by the investiga- tion of weaker rhizomes of Nymphza alba. I reproduce his description as follows, Inter- nodes shortened. Leaves arranged spirally. The transverse section shows between pith and cortex a circle of separate bundles, which is divided usually into three, rarely into four groups, which can be recognised with the naked eye. The three groups are of unequal width: they vary continually throughout the length of the stem, and are related to the arrangement of the leaves. The bundles of the circle are often connected with one another, so that, when seen in surface view, they appear as a network. -The middle of the pith is traversed by a central bundle, which throws off now and then a branch to the net-work. From the base of the leaf five bundles enter the stem: three of these lie rather higher and form the true leaf-trace. Their lateral bundles separate widely from one another, and weave themselves in with the net-like circle at two almost diametrically opposite points, so that the trace is about 180° wide. The median bundle also loses itself usually at once in the net. But sometimes it passes inwards through the pith, after forming some anastomoses with other bundles, and unites with the central bundle. In one stem it was the 8th and 13th, in another the rst, 6th, 11th, 18th, and 32nd leaves, the median bundles of which passed to the centre, while those of all the other leaves remained in the outer network. In the first example the 8th and 13th, in the second the rst, 6th, 11th, and 32nd leaves were on the upper side of the procumbent stem, the 18th on its lower side. ; An independent growth of the central bundle at its apex was not observed: Nageli therefore considers it as a sympodium of median bundles. I found in weak rhizomes of Nuphar pumilum that the transverse section resembled _ that described in the case of Nymphza: an irregular ring of 8-12 bundles, and a central, often very eccentric, and frequently branched bundle being seen: the latter is in rare cases entirely absent from the section. The bundles of the ring form a net with elongated meshes, the main meshes being limited by the bundles of the leaf-trace, between which smaller bundles, usually pushed back rather further into the pith, form an irregular net- work. The leaf-trace consists of three bundles, and is about 120° wide, the median bundle forks in the node into two shanks diverging at an obtuse angle: each of these in its descending course is united with the lateral bundle of its own side. I never saw the median bundle curve to the middle line of the stem, but rather I here saw only a bundle, which ran irregularly from side to side, and here and there gave off a lateral branch and anastomosed with the peripheral net. An independent ending of this bundle beneath the growing-point was not to be found. Besides this I have examined but few preparations, and wished in the above remarks only to point to N. pumilum as an object well suited for the elucidation of the structure of the stem in the Nympheacez. ? Anatomie und Physiologie, p. 235. COURSE OF THE BUNDLES IN THE STEM. 253 3. Bundles of the trace and cauline bundles are both present. The bundles of the trace are arranged in a ring, the cauline bundles are in the pith. To this series belong species of Begonia, Orobanchex, species of Mamillaria, Melastomacez, some Um- belliferze, and Araliaceze: further in the main also Nelumbium, In the Begonias medullary bundles are common. Hildebrand? found them in 28 ‘species out of 128: for instance in B. Evansiana, laciniata, Rex, xanthina, etc. Accord- ing to Hildebrand’s description of stems in the mature state (following the course from below upwards) these seem to be for the most part cauline. In the first internodes of the seedling they are wanting, they branch from those of the ring in higher internodes. They run parallel and perpendicular in the internode, in the nodes they anastomose with one another and with those of the ring: as a rule none pass through the node without connection with others. In B. Hiigelii, Muricata, and luxurians Hildebrand saw 1-3 medullary bundles enter, without previous anastomosis with others, directly into the middle of the petiole, a phenomenon which occurs less frequently in others: e.g. B. laciniata. From the plexus in the node medullary bundles pass on into the next higher internode, and other bundles branch off to pass into the ring. In some stems single bundles in successive nodes pass successively into the pith, and again into the ring, and finally into a leaf, ‘but this whole course is rendered very indistinct by anastomoses.’ The many points, which still remain doubtful, arise partly from the difficulty of following the course of the bundles in the Begonias with accuracy. It is made the more probable that the medullary bundles are for the most part cauline by the fact that, according to Sanio’s investigation of B. Evansiana (/. c. 224), they appear later than those of the ring 2, Aralia racemosa® and A. japonica have within the typical Dicotyledonous ring a second, consisting of small, widely separated bundles. In other species, e.g. A. papyrifera, the second ring is absent. Besides this ring there are, according to Sanio in A. racemosa, other single bundles scattered in the pith. All these bundles, which run perpendicularly in the internode, anastomose in the nodes without passing out to the leaves: their de- velopment proceeds much later than that of the ring. As regards the distribution of vessels and sieve-tubes the outer medullary bundles have a position the reverse of that of the bundles of the ring, the inner ones are irregularly arranged. Comp, § 101. In the stem of some few Umbellifere *, Silaus pratensis Bess., Peucedanum Oreose- linum Mch., Opoponax Chironium K., Ferula communis, and an undetermined form of Taurus, medullary bundles have been observed within the ring: as many as 13 in Silaus, 20 in Opoponax, 82 in the plant of Taurus (Reichardt), at least 100 in the flowering stems of Ferula communis. They are distributed in the transverse section over the whole pith: their number varies in successive internodes, e. g. in Silaus, Ex. I. 13, 11, 10, 9, 7,3; Ex. II. 10, 8, 7, 7, 6, 13 Ex. III. 9, 8, 5, 3, 1: Peucedanum Oreoselinum, Ex. I. 22, 20, 18, 17, 14,73 Ex. II. 20, 18, 17, 17, 12, 6; Ex. III, 15, 13, 10, 7, 3. According to the consistent accounts of Jochmann and Reichardt the medullary ‘bundles do not pass out into the petiole; they are cauline. They take a parallel and perpendicular course through the internode, dividing here and there, and united for short distances: at the node they anastomose by means of connecting bundles with one another, and with those of the ring: from the anastomosing bundles in the node those medullary bundles start which traverse the next internode. Those of the lowest inter- node above the root affix themselves on those of the ring of the same internode, or 1 Anatomische Untersuchungen tiber die Stamme d. Begoniaceen, Berlin, 1859. 2 [Cf. Westermaier, Ueber das markstandige Biindelsystem der Begoniaceen, Regensburg, 1879.] 3 Sanio, Zc, p. 226. * De Candolle, Cxpmnogmaahle, I, p. 184, Taf. I1I.—Jochmann, de Umbelliferarum sstructura, 1854.—H. W. Reichardt, Ueber das centrale Gefassbiindelsystem einiger Umbelliferen, Wiener Acad. Sitzungsbr. Bd, XXI (1856), s. 133. 254 PRIMARY ARRANGEMENT OF TISSUES. rather spring from them: the same is the case with those of the lowest internode of a branch, so that the latter are not in direct continuity with those of the stem (Reichardt), The occurrence of medullary bundles is a purely specific property. Of eight investigated species of Peucedanum they appear only in P. Oreoselinum: they are absent in Silaus tenuifolius. They are not present in the one-year-old seedling of S. pratensis. Some Mamillarias+ have, within the typical ring of bundles of the leaf-trace, a second in the peripheral parts of the pith, this being composed of numerous small cauline bundies,. In M. angularis, and an undetermined similar species, there are about thirty of these. They ascend the stem parallel to those of the leaf-trace, are strongly undulated in a radial, and especially in a tangential direction, and anastomose at acute angles. They are absent in young seedlings and in young shoots, and appear later at some height above the base of the shoot, springing from the inner side of the bundles of the trace. I was unable to find anastomoses with the leaf-trace, or the ring of secondary wood, excepting at their point of origin. In other Mamillarias, as M. pusilla, glochidiata, &c., I looked in vain for the medullary bundles, even in the mature shoot. Of other Cactacex, forms of Echinocactus and thick ones of Cereus (e.g. C. candicans?) have a medullary system of bundles, which on account of its peculiar relation to the lateral shoots will be described in Sect. 94. The small Orobanchex have in their stems only the typical Dicotyledonous ring of vascular bundles, Strong stems of the more robust forms, as O. elatior Sutt., rubens Wallr., caryophyllacea Sm., Rapum Thuill., and Cistanche lutea?, haye, inside the ring, and scattered in the pith, small bundles, the Up number of which is variable, and is large in strong specimens, i They are cauline, and in young specimens take an undulating / longitudinal course through the stem, anastomosing here and there, and ending blind beneath the apex of the stem. Comp. Fig. 111. In fully developed flowering stems they gradually cease below the inflorescence, curving outwards, and uniting with the 3 ~ bundles of the ring. It is only in Cistanche lutea, where they occur in considerable quantity, that they run freely in large num- bers to the extreme apex of the inflorescence. In Epiphegus FIG. rr.—Orobanche Rapun americanus and Conopholis*® there are found in the base of the {natural Size). ‘Bud of a flower. main stem three concentric rings of bundles: in the latter genus ing Shoot, median longitudinal a section, 4—) ring of bundles, the bundles composing them are contiguous with one another in consisting of leaf-traces; within is a : : n this are the cauline bundles. at radial series. It is uncertain whether these belong to this category, the base, where it I ‘ . : Off, is a network of tke ati, oF to that of radially diverging bundles of the trace, or are pro- which anastomose one with an- i other, and with the bundles of ducts of a secondary cambium, . : the leaf-trace. In the flowering stems of species of Balanophora, according to Géppert’s* description, the same or a closely similar arrange- ment occurs to that in O. Rapum: numerous branched bundles are scattered within 2 ring composed of bundles of the leaf-traces. In the Helosidee the ring of bundles is atone present in the elongated rhizomes, in the tubers and Inflorescences there are branched, scattered bundles >, Two cataphyllary leaves and one foliage leaf succeed one another regularly on the 1 Von Mohl, Verm. Schriften, p. 115. ® Graf zu Solms-Laubach, de Lathroee generis positione systematica, Diss. Berlin, 1865, pp. 8, 14, and Pringsheim’s Jahrb. VI. p. 522. 5 Chatin, Anat, Comp. Taf. XVIII. * Ueber den Bau der Balanophoren, N. Act. Carol, Leop. vol. XVIII. Suppl. 1.—Compare also po eres Trans. Linn, Soc, London, XXII; Graf zu Solms, in Pringsheim’s Jahrb. 1. Ps 529. 5 Compare Eichler, in Flora Brasiliensis, Fasc, XLVII, COURSE OF THE BUNDLES IN THE STEM. 255 rhizome of Nelumbium speciosum’, The internode between the latter and the next following scale leaf is elongated (as much as 4 feet), the others are short. The elongated internode (Fig. 112) has six blunt angles, so that when horizontally placed one surface is turned upwards (0), another downwards (uz), and one edge to each side right and left. It is traversed by 6 large air-canals (/), corresponding to the angles, one small axile one, and two small ones, corresponding to the upper surface, and in regular cases (slight deviations occur) by about 252 vascular bundles with a perpendicular course, the arrangement of which in the transverse section is according to Wigand as follows. Firstly, an izner system of twelve bundles, disposed in two concentric alternating circles (1 and 2) of six each, at equal distances from one another in each circle, in the inmost circle (1) one bundle is opposite the upper, and one opposite the lower surface; secondly, a middle system, extending to the outer limit of the air-canals: it consists of four concentric rings (circles 3-6); in each of the circles 4, 5, 6, one bundle alternates with two air-canals: in circle 3, in addition to the otherwise similar arrangement, there are on each side between each two aw FIG, 112,—Scheme of the transverse section through the internode of a rhizome of Nelumbium speciosum, after Wigand. o upper, # lower side; Zair canals, The figures 1—11 indicate the successive circles in which the bundles are arranged. The bundles of the circles 3 and 5 have an orientation the reverse of that of the others, which is indicated by direction of the dashes affixed to the round spots. lateral lower air-passages three bundles, and on each side between the two lateral upper ones two bundles, that is six more bundles are present than in the other circles: the bundles of four circles, with exception of the last-named six, form regular radial rows. Thirdly, the peripheral system, between the outer side of the air-canals and the surface of the stem, consists of 4-5 concentric circles of bundles (7-11), the inmost (7) having 18 bundles in pairs alternating with the radial rows of the middle series and with the air-passages ; the next (8) consisting of 45 bundles, of which two are between the two of one pair, and 1 According to Wigand, Nelumbium speciosum, Botan. Zeitg. 1871, p. 816, &c. See also Trécul, Ann, Sci. Nat. 5 sér. I. p. 162, &c. 256 PRIMARY ARRANGEMENT OF TISSUES, three between each two pairs of the above; the bundles of the next (9) alternate with those of (8), those of (10) and (r1) alternating irregularly with those next within them. Of these bundles all those of the peripheral series are, according to Wigand, ‘true leaf-bundles, since they traverse only one internode and then run into the leaf-organs,’ The course of the rest is difficult to describe with certainty, because of the complicated branchings of lateral shoots and roots in the node, and it requires further investigation. The bundles of the inmost circle of the middle series (3) seem to be cauline, since ‘ it has not hitherto been possible to prove that they contribute to the lateral organs,’ and they ‘apparently always traverse one internode only, and lose themselves in the node, their place being taken by fresh ones in the following internode.’ The same holds perhaps for the other members of the middle series, which are radially arranged. Of the inmost series the four lateral bundles of the inner side (1) give off branches to the roots, the upper and under ones of the same circle give off branches ‘ directly or in- - directly’ to the leaves. All six bundles of this circle ‘are however - distinguished by the fact that they alone of all the bundles of the stem traverse all internodes and nodes up to the growing point,’ while on the contrary the six bundles of the same series, which alternate with them (2), ‘each belong to one elongated internode only, and to the roots (which arise at each node), and then end in the node, just above the root-region.’ The medullary bundles of the Melastomacee will be described below, p. 259, in con- nection with the rest of the bundle system of these plants. b. Cortical bundles. Sect. 63. A relatively small number of Dicotyledons is characterised by having a bundle system in the internodes arranged typically in a ring, and outside this other bundles, which traverse the cortex. These cortical bundles are sometimes bundles of the leaf-trace, which run for a certain distance outside the ring, and later curve into it: as in the cases of Lathyrus Aphaca, and Pseudaphaca above described, the Casuarinas, and many Begonias: also the cortical bundles of the Cycadez, to be described in ‘detail in Chap. XVI, belong to this category, perhaps also Nepenthes. Sometimes they are certain bundles, belonging to many-bundled leaf-traces, which never enter the ring, but form with the similar ones belonging successively to upper and lower leaves an independent cortical bundle-system only connected with the ring by anastomoses at the nodes: this is the case in the Calycanthes, many Melastomacez, also in Arceuthobium Oxycedri. In many succulent plants with reduced leaves, as Salicornia, Cactez, they are branches of the bundles of the leaf-trace, which are branched and distributed like the bundle-expansions of the leaf lamina, and will therefore be treated of where these expansions are described. Finally, in the winged Rhipsalidacez, according to Véchting, the peculiar case occurs that the bundles of the leaf-trace are mainly cortical, while a ring of bundles, ‘surrounding a pith, and quite similar to the typical Dicotyledonous ring of bundles of the leaf-trace, consists at least for the most part of cauline bundles. The young foliage shoots of the Casuarinas' (Fig. 113) are covered with whorls of small leaves, united into a long sheath: the average number of leaves of one whorl varies according to the species (4-20), The leaves of successive whorls, and the ridges, which run down the backs of the leaves, alternate in successive internodes. One vascular bundle enters each leaf. From the point of insertion of the sheath it passes * Compare Low, De Casuarinearum caulis foliique evolutione et structura, Berlin, 1865. COURSE OF THE BUNDLES IN THE STEM. 257 into the periphery of the stem, and there runs, parallel to it, in the cortex as far as the next node: it then curves inwards, and ranging itself with those of the same whorl of leaves around a narrow cylinder of pith, it descends per- pendicularly through a second internode. At the lower limit of this, i.e. at the 2nd node from the point of exit into the leaf, it is inserted (according to Léw with a short fork) on the bundles which here pass out into the cortex. The transverse section of each internode thus shows 2 con- céntric circles of an equal number of alternating bundles: the one peripheral, composed of the bundles of the trace of its own whorl of leaves, the other axile (forming the woody ring at a later period), consisting of the bundles of the trace of the next higher leaf. In Begonia angularis Raddi, Hildebrand found one bundle in the outer cortex of each of the six angles of the stem. All the six in one internode form together the trace of the next higher leaf, which includes 3 of the circumference of the stem. They pass perpendicularly downwards in the angles @ to the next lower node, and here curve into the ring of es uae pantar gene bundles. Their further course in the latter has not been scheme of the course of the vascu- investigated. Many internodes have less than six angles, and [at bundles in the median longitu. dinal section of a young branch. correspondingly fewer cortical bundles: others may be en- —*~4 successive whorls of leaves; bundles of whorl, 2 ending in node tirely without either. The-same is the case with Begonia a, 3in4, ging, &c. tomentosa, with this difference, that the number of the cortical bundles is ‘indefinite’ and often very large, while some of them often run through two internodes in the cortex. Arceuthobium Oxycedri* has decussating pairs of leaves, each leaf has three bundles of the trace, one median and two lateral. The latter converge, and enter the stem: here they unite and descend opposite those of the other leaf of the pair, and separated from them only by a narrow band of pith, to the next node, where they insert themselves upon the bundles which there pass out. The thin median bundle of each leaf pursues an in- dividual course through the cortex, and is also inserted at the next lower node. The transverse section through an internode thus shows two decussating pairs of bundles, one stronger axile pair, and one weaker and peripheral. In the Calycanthacex* three bundles pass out into each of the opposite and decussate leaves, one stronger median, and two weaker lateral ones, ‘The median ones are arranged in a ring in the stem. Each runs down through two internodes, and then affixes itself at the node on the median bundles which there pass out. The lateral bundles ‘(rather later developed) pass perpendicularly down the stem in the cortex outside the ring of bundles: at the next node they insert themselves on the cortical bundles which there pass out. Thus in each internode the transverse section shows the ring of bundles, and outside it four cortical bundles. In the node each cortical bundle is connected with the ring by a short radial transverse bundle, and by another with the median bundle which next passes out, and again, by a stronger horizontal girdle-like bundle, with the next cortical bundle belonging to the same side of the stem. In the seedling two bundles enter each cotyledon: they descend with the median bundles of the first three leaves through the hypocotyledonary axis, the transverse section of which thus shows six bundles. The cortical bundles of the first two leaves only extend down to the cotyledonary node. 1 /.¢., compare p. 253. 2 Graf zu Solms-Laubach, in Pringsheim’s Jahrb. Bd. VI. p. 523. 3 Mirbel, Ann. Sci. Nat. XIV (1828),—Gaudichaud, Archives de Botanique, II. p. 493 (1833). —Woronin, Botan, Zeitg. 1860, p. 177. s 258 PRIMARY ARRANGEMENT OF TISSUES. ° In the internode of Nepenthes* there is found an inner typical bundle-ring, surrounding the pith, which later undergoes secondary thickening, and externally other bundles are 4 Ws MY ary SZ a} a 1 ie a 1, any, ay 7 ' t 1 to \ 4a says by FIG. 114.—Centradenia rosea. End FIG. 115.—Osbeckia canescens (magnified about 25). Longitudinal of a shoot, halved longitudinally; the half of the end of a shoot, prepared like Fig. 114, and seen from without. three lower pairs of leaves and the epi- Six decussating pairs’ of leaves numbered in order ; in 5 and 6 as yet no dermis removed, made transparent by bundles, in 4 the median bundle is just visible. Further description in potash, and seen from wzthout. The the text, . ‘ pairs of leaves and the bundles bélonging to them numbered in order; 2 the median ones, Z the lateral bundles. One leaf of the fourth pair (4) covers the ‘ growing-point. In the node of pair 3 the transverse girdle is as yet imperfectly developed. Further description in the text. Magnified about 25. to be seen in the very broad outer cortex: these are partly bundles of the leaf-trace, which passing in a radially-oblique direction through the cortex, gradually enter the 1 C, H. Schultz, Vaisseaux du Latex, Zc. compare p. 249.—Korthals, Verhandelingen, /.¢. compare p. 226. ~ COURSE OF THE BUNDLES IN THE STEM. 259 ‘inner ring ; partly small bundles, the origin of which remains to be decided; these pass close béneath the epidermis, and are connected with one another by oblique branches. A thorough investigation is in progress?. : In the Melastomacex the course of the cortical bundles is partly the same as the above, and partly similar to that in Calycanthus. It may here be described in connection with the other peculiarities of arrangement of the bundles in this family (comp. p. 256)”. The stem has four angles, and bears decussate pairs of opposite leaves: those of one pair are either equal, or, as in many Centradenias, of unequal size. Each pair faces two opposed surfaces of the stem: these may be called the surfaces belonging to that pair, and the others the intervening surfaces. In the simplest case investigated the trace of the single leaf which enters the node consists of three bundles, one median and two lateral: in many species it consists of more than three, through multiplication of the lateral bundles on each side. The bundles which enter the stem pass, in many species, directly into the bundle-ring without forming cortical bundles: Sonerila margaritacea, Medinilla farinosa, Sieboldii, magnifica, Cyanophyllum magnificum,. Clidemia parviflora, Miconia purpurascens, Lasiandra Hoibrenkii®. In the other investigated species the median bundle always enters the bundle-ring, usually without, rarely after previously giving off cortical bundles; the lateral ones either run, as in Calycanthus, down the angles of the stem as cortical bundles, or they enter the ring, after having given off cortical bundles as branches. The cortical bundles are always connected at the node both with one another and with those which pass to the ring, by a transverse girdle of horizontal branches: they run from this to the next lower transverse girdle, and insert themselves on the latter. In the ring the bundles of the trace always pass down through several internodes, they pectinate in various ways with those of lower pairs of leaves, remaining simple, or splitting into shanks. For numerous differences of arrangement according to the number of bundles and the species, compare Véchting, /. c. The cortical bundles are always derived, as above stated, from the bundles of the trace. In the simplest case (Centradenia rosea, Fig. 114) they are the lateral bundles of the three-bundled trace. From the base of the leaf one stronger median bundle m,—m,) and two weaker lateral ones enter the node. All the median bundles are arranged in the ring: they pursue an individual course, directly, or with slight curvature at the nodes, through three internodes (two according to Véchting), and then unite laterally with such as come from lower nodes. ‘The lateral bundles give off one branch on each side at the node, which runs transversely through the outer cortex, the one to the out-going median bundle of the same leaf, the other to a similar one, which comes from the nearest lateral bundle of the opposite leaf. These branches together form the transverse girdle of the node, which girdle is further connected with the out-going portions of the bundles by small branches. From the point of departure of the branches of the girdle each lateral bundle runs perpendicularly through the cortex of the corner of the stem, and inserts itself on the transverse girdle of the next lower node. As an example of the other case, i.e. that cortical bundles and transverse girdles are branches of the bundles entering the ring, Osbeckia canescens (Figs. 115, 116) may be described, Astrong median bundle m, - 2 enters from each leaf, into the corresponding side of the internode, and here descends perpendicularly, in the bundle-ring: in the next lower node it splits into two shanks (v, and ), which enclose between them the united lateral bundle, which there enters the stem: these shanks may be further followed down three internodes. A strong lateral bundle passes from each leaf almost horizontally through the cortex into the middle of each of the intervening sides: it here unites with 1 E, Zacharias, Ueber die Anatomie des Stammes der Gattung Nepenthes, Strassburg, 1877. ? Véchting, Bau, &c. d. Melastomaceen, in Hanstein’s Bot., Abhandl. III. Compare also Criiger, Botan. Zeitg, 1850, p. 178.—Sanio, Ibid. 1865, p. 179.—Hildebrand, Begoniaceenstamme, p. I. 3 The names quoted are partly derived from Véchting’s work, and partly garden names, for the correctness of which I will not answer, $2 260 PRIMARY ARRANGEMENT OF TISSUES. a similar one from the opposite leaf, and the united bundle (/,~/,) then curves into the ring. Its course in the ring may be gathered from Fig. 115. The lower endings of the bundles in the ring were not investigated. .Opposite its point of exit into the leaf the ‘median bundle gives off a branch on either side, which takes a curved and almost horizontal course through the cortex to the nearest angle, and here unites with the out-going lateral bundle. The transverse girdle accordingly arises from the last-named branches of the median bundles, and the horizontal portions of the lateral ones. From the latter sections of the girdle there ‘arise near each angle one or two cortical bundles (r), which run perpendicularly to the next lower node, and here insert themselves on the transverse girdle. If two cortical bundles are present, they stand radially before one another, and one often inserts itself on the other before reaching the lower girdle. To the above description it may be added that the two lateral bundles which have been mentioned in each leaf-trace are each formed by the coalescence of two lateral bundles (/ and p, Fig. 115) of the base of the petiole. For further peculiarities in other species re- ference may be made to Vochting’s work: it may be added that the examples here selected have been retained, since the corresponding woodcuts were finished two years before that work appeared. Species with wide wing-like projecting angles often have several cortical bundles placed radially one before another in each of these angles (e.g. Hete- rocentron subtriplinervium, Lasiandra macrantha 3-4, Centradenia grandifolia 5-7), these anastomose now and then with one another; they originate as branches either of median or lateral bundles of the first order, or from lateral bundles of a higher order, if such are present. Besides the cortical bundles there are also cauline medullary bundles in most of the Melasto- macez (compare Fig. 116). These are usually to be found also in those species in which the cortical bundles are absent. Only Sonerila margaritacea FIG. 116.—Osbeckia canescens. Thick transverse section through a node, which is rather further developed than 3 and less than 2 in Fig. 115; it is transparent, and seen from below. The transverse sections of the bundles of the leaf-trace which are nearest the beholder and in the surface of section are represented as dar&. m1 the median bundles of the trace passing out at the node; / the lateral bundles of the same; ~ the cortical bundles which ‘pass down from the node; o those which pass upwards from it; #2 the median bundles of the next higher node, which branched above the node; with them there alternate, opposite each corner, single bundles, with regard to which it is not quite certain whether they are the lateral ones of the next higher pair of leaves, or shanks of the median bundles of the second higher pair. The lighter spots in the pith indicate the transverse sections of the cauline bundles. In this node they were not yet plain, and have been ‘drawn in the figure from another preparation (40). is without either: it alone of the species investigated has a quite typical Dicotyledonous bundle-system. In the simplest case a single bundle is found in the centre of the pith, e.g. Medinilla farinosa, Sieboldii, or this may even be sometimes present, sometimes absent, as in Eriocnema marmorata and Centra- denia rosea. In the transverse section of the in- ternode of other species there are several bundles which lie in the middle of the pith, e.g. Melastoma igneum, Lasiandra Maximiliani 1-3, Medinilla- mag- nifica 2-4, Melastoma cymosum 8-10; finally, others have many of them scattered over the whole trans- verse section, e.g. Heterocentron subtriplinervium 18, Miconia chrysoneura and Cyanophyllum magnificum 30, 40, and more. The medullary bundles run perpendicularly through the internodes. In the nodes they are connected by oblique or transyerse branches of varying number one with another, and with the bundles of the ring. From the network or felt thus formed proceed those of the next higher internode. In the large majority of cases they arise much later than the bundles of the leaf-trace which are in the same transverse section, and they do not pass out to the leaves. The medullary bundles are usually relatively small, and characterised by peculiarities of structure to be described below (§ 105). COURSE OF THE BUNDLES IN THE STEM, 261 Some of the Rbipsalidex! have round stems, others angular and winged; both forms have in transverse section a circular or elliptical ring of bundles, which, especially in the winged species, is surrounded by a very broad succulent cortex, In the round forms, such as R. Saglionis; salicornioides, the one-bundled leaf-traces pass slightly. obliquely downwards through the cortex into the ring, which is originally formed from them alone, but later secondary intercalary_ bundles also appear (Chap. XIV). In the winged forms the leaves are seated only on the angles. The bundles of the trace enter these, and run, in the main tangentially-perpendicular and radially-oblique, down through the cortex; they enter the ring about the level of the next lower leaf, and then descend further in a perpendicular direction, They thus form the portions of the ring corresponding to the angles. But the portions of the ring between these, which correspond in the elliptical ring of two angled forms (e.g. Lepismium radicans, Rhipsalis carnosa), to the broad sides of the ellipse, as seen in transverse section, and which comprise the greater part of the ring, are here composed of cauline longitudinal bundles, connected here and there by oblique anastomoses: on these the common bundles insert themselves in the region indicated. These cauline bundles correspond to the secondary intercalary bundles, which complete the woody ring of typical Dicoty- ledons, and which will be described in Chapter XIV; but they are distinguished from these by their appearing on the first primary differentiation of tissues. Finally, in all Rhipsalidee branches come off from the bundles of the trace during’their course through the cortex: these with their further branches form a cortical network of bundles (which is further strengthened by the bundles which come from the axillary buds). The special form and development of this varies according to the species, in the winged species it is exclusively or chiefly expanded in the wings, in a radial direction. Compare Véchting, /.¢. It is not known how far other Cactacex, which are winged and have a cortical net- work of bundles, correspond to the winged Rhipsalidee in the course of the panes of the leaf-trace and intercalary bundles, II. Tue Type or THE Pans. Sect. 64. The stem of most Monocotyledons does not show the bundles in the transverse section of an internode arranged in a simple ring, but within a peripheral zone of cor/ex, which has no bundles, there is a circular surface, in which either several concentric irregular and interlocking series of bundles are arranged round a central portion without bundles (pith), as e.g. in many stems of Grasses, ' which later become hollow; or the bundles lie scattered over the whole surface. Thus instead of the ring of bundles of the Dicotyledons there is here a cylinder, which contains the bundles; the zone surrounding the cylinder, which was called the cortex, corresponds to the cortex of the Dicotyledons: the terms pith and medullary rays may be used in a comparative sense for the bands of other tissue (as a matter of fact parenchymatous) which lie between the bundles in the cylinder. The arrangement of the bundles in transverse section in the Palm type depends upon the radially-oblique course of the leaf-traces. This must first be demonstrated for that form which may be called the simple Palm type: with this are further connected a number of more or less divergent phenomena. ae 1 Vochting, Morpholog. und Anat. d. Rhipsalideen, Pringsheim’s Jahrb. IX. p. 326. 262 PRIMARY ARRANGEMENT OF TISSUES. a. Simple type of the Palms. ’ Since Mohl's Anatomy of Palms! the following chief dharasreretion have been known for this type. All bundles of the cylinder (with doubtful, and extremely unimportant exceptions 'to be mentioned below) are bundles of the leaf-trace. The base of the leaf encloses the whole circumference of the stem, or at least the greater part of it. The leaf- trace always consists of a number of bundles, and usually of many, in strong shoots of two hundred bundles: its width is 3 of the circumference of the stem, or more, or not much less, The bundles curve from the base of the leaf into the cylinder, and pass downwards in the latter; some at its surface, running almost radially perpendi- cular; others are radial and oblique, penetrating first in a curve convex upwards and inwards, towards the longitudinal axis of the cylinder, then turning outwards and gradually passing towards the surface of the cylinder; as they approach it, they assume an almost perpendicular position. All bundles descend through many internodes, and finally unite in the outer portions of the cylinder with such as emerge lower down, inserting themselves on these sometimes in a tangential, sometimes in a radial or oblique direction. Up to this point of insertion at their lower ends the bundles pursue an individual course. The coalescence of the lower ends of descending bundles with such as emerge lower down occurs with such frequency that the whole number of the bundles in equally strong successive inter- nodes remains about the same. Where the successive internodes and leaves grow stronger the number of the bundles increases, and wece versa. The number of the internodes which a bundle traverses cannot be exactly defined. Further, the bundles of a leaf-trace, which curve towards the middle of the cylinder, do not all penetrate to the same depth; on an average the median bundle of a series penetrates the deepest, the others less deeply the further they are from the median one, the marginal ones pass almost perpendicularly down the surface of the cylinder: where there are several series, those of the inmost penetrate as a rule deeper than those of the outer, which are equally distant from the median bundle, The necessary consequences of the course described are, firstly, that the bundles in the transverse section of an internode are more crowded the nearer they are to the surface of the cylinder, a phenomenon which is especially striking where the bundles are distributed over the whole surface of section of the cylinder. Secondly, that the successive traces pectinate, and cross one another with their curved bundles. Mohl’s celebrated scheme, which is here reproduced -in Fig. 117, shows this condi- tion in radial longitudinal section, but starting from the incorrect assumption that all bundles of one trace are almost equally curved and are tangential and perpendicular, that is that they lie at the surface of a cone which has a funnel-shaped opening at the top. If it is assumed that the leaves alternate with a divergence of } and embrace the stem, and that the bundles are. tangentially perpendicular, their course in the stem would be represented with greater exactitude by the scheme of a radial 1 De Palmarum Structura; Monachii, 1831; Verm. Schr. p. 129; translated in the Ray Society's Reports and Papers on Botany, 1849.—Nageli, Beitr. I. Zc. COURSE OF THE BUNDLES IN THE STEM. 263 section through the median planes of the leaves, Fig. 118. But the assumption of a tangentially-perpendicular course only holds good for bundles which are also radially perpendicular. As Meneghini’ first asserted, and Mohl also conceded (Verm. Schriften, p. 160), and as Naegeli proved more accurately, each radially curved bundle runs also tangentially oblique, with a spiral curvature, which is stronger the stronger the radial curvature. Nageli found e.g. that the median bundle of a leaf of Chame- dorea elatior Mart. made 14 revolutions in passing through six internodes: in the sixth it had not quite accomplished half the distance from the centre of the stem to the inner surface of the cortex, on its way outwards. In stems with very short internodes and closely packed bundles the spiral curvature is at once visible in the transverse section, and very plainly in the bundles of the stem of Xanthorrhcea, é& 4g 4m 4 3 3 z |_| a 2 2 2m 2 ¢ 7 4 bin ._ FIG. 117.—Mohl’s scheme of the vascular system of F1G, 118.—Scheme of the vascular system in the Palm- Monocotyledons. Successive leaves, or rather suc- type, assuming that the leaves alternate in two rows, and cessive nodes, are numbered in order. embrace the stem. Successive leaves numbered in order. zm median bundle, which pass almost horizontally to the middle of the stem: the peculiar appearance so often noted? in transverse sections of this plant depends upon the above peculiarities, Finally, many variations of direction from that hitherto described as uniform may occur in the course of a bundle, such as curvatures outwards and inwards, &c., which never appear to be constant. The above description holds both for the preponderating number of cases, in which those bundles which penetrate deepest reach the middle of the stem, and also for those where, as in the stems of grasses which become hollow, a broad central part (pith) remains free from bundles. Where the internodes are short, as e.g. in the well-known preparations by maceration of the stems of Dracwna Draco, the course can easily be recognised in the main features. Where the internodes are ' Ricerche sulla struttura del caule nelle piante monocotiledoni, Padova, 1836. ? De Candolle, Organographie, I. Tab. VII. VIII.—Schleiden, Grundziige, 3. Aufl. II. p. 160, 264 PRIMARY ARRANGEMENT OF TISSUES, : very long and thin, as e.g. in the Grasses, or egg- or spindle-shaped, as in the so-called pseudo-bulbs of epiphytic Orchids, it appears otherwise at first sight. The internode appears to be traversed by bundles, which are parallel, or convergent towards both ends: of these some may be seen to enter the leaf at the node, while many run into the next internode. One may however be easily convinced. that here also their course is such as that above described. In young stems, which are still short, no variation from it is to be traced. In the subsequent elongation of. the internodes of the stem of Grasses to 20-50 times their original height or even more, that portion of all the bundles which turns outwards while descending is so elongated, that at first sight it does not appear to. deviate from the perpendicular in any one of the six or more internodes which it traverses: the general view of its course is rendered especially difficult by a complex felt of transverse bundles in each node* (comp. Sect. 95). In the pseudo-bulbs. besides this elongation there is also a transverse swelling of the internode in the middle, and a consequent curvature of the bundles. According to the statements of Unger? and Millardet® there is found in plants of this category, Grasses, Palms, Dracena, Yucca, Narcissus, Galanthus, Leucojum, Pandanus, in addition to the bundles of the leaf-trace a further system of cauline bundles, which run upwards outside the cylinder, converging towards the growing-point; accord- ing to the rule which holds for cauline bundles in the Phanerogams, these are formed later than the bundles of the leaf-trace. Leaving out of account the Commelinacez, which will be described below, and the secondary formation of wood in Dracena, Yucca, and their allies, I could not convince myself of the presence of such cauline bundles, which might it is true be easily mistaken for perpendicular bundles of the trace. As has been repeatedly stated, the cylinder traversed by the bundles is in most of these cases sharply marked off from the inner surface of the cortex. The cortex itself is of variable thickness, in Rhizomes it is usually of great thickness, in aerial stems it is often relatively very thin. b. Modifications of the type of the Palms. Sect. 65. According to the scheme laid down in the previous paragraphs all the bundles pursue an individual course up to their point of final insertion at the periphery of the cylinder, this being sharply limited from the cortex, which in the internodes is without any bundles. 2 In many cases this scheme is subject to the following modifications: (1) the bundles receive oblique or transverse connecting branches, or anastomoses, during their course: (2) before they reach the periphery of the cylinder in their curved course they unite with others belonging to lower leaves (Sect. 66): and (3) cortical bundles appear outside the cylinder (Sect. 67). Each of these phenomena may occur separately or combined with the others, Anastomoses of the bundles of the leaf-trace one with another—besides those * Von Mohl, Palmarum Structura, Tab. Q.—Schleiden, Grundziige, 3. Aufl. II. p. 158. 2 Le p. 54. ® Mem. de la Soc. des Sciences Nat. de Cherbourg, tom, XI. p. 4. COURSE OF THE BUNDLES IN THE STEM. 265 which arise through the’ insertion of bundles which go to the branches and roots— occur in the greatest abundance, so as completely to mask the typical course of the bundles, in the tuberous stems ‘of certain Aroidez to be described below. In other stems with short or slightly elongated internodes they appear as an unimportant phenomenon which occurs occasionally. They are however numerous and characteristic in the greatly elongated internodes of the flower-stems and foliage shoots of many Cyperacez, Scirpus palustris, lacustris, and their allies, of Papyrus, species of Cyperus, and of Pontederia cordata?. These ‘halms’ have this peculiarity, that their longitudinal bundles are connected in a reticulate manner one with another by small horizontal or oblique branches, like those of the foliage- leaves of Monocotyledons (Sect. 91). The transverse branches run in the dia- phragms (but not by any means in all), which separate the air-cavities one from another. According as the longitudinal bundles are scattered over the whole transverse section of the halm, or (as in Sc. palustris and its nearest allies) only form a ring inside the chlorophyll-parenchyma of the cortex, there are found also transverse branches throughout the whole thickness of the halm, and in the most various directions, or only in the zone occupied by the ring. Srct. 66. The phenomenon of insertion of bundles of the leaf-trace during their curved course through the middle of the cylinder, and before they reach its periphery, on others which emerge lower down, and of their further descent in union with the latter, is also of frequent occurrence in the Aroidez to be described below. Further it occurs in Pandanes?, Bromeliaceze (Ananassa, Tillandsia acaulis, Hort:), and apparently also, according to Karsten*, in many Palms, especially Martineza aculeata. Whether this union of their course follows definite rules for certain bundles of a leaf-trace remains for more exact investigation to decide. i Szct. 67. The cortex is free from vascular bundles in many of the Monocotyle- dons of this category, if those bundles be disregarded which pass to the leaves at the nodes, and those which enter branches and roots: such bundles must appear in all or almost all transverse sections where the internodes are very short. On the other hand a special cortical system of bundles may be distinguished from the cylinder in certain cases. This consists in the simplest case of bundles of the leaf-trace, which after their entry into the stem descend first in the cortex through one or several internodes, and then enter into the cylinder. This is the case in stems of certain Aroidee, in many Rhizomes, as Carex hirta (but not in C. disticha), where all bundles run through one internode in the cortex: Scirpus lacustris, Typha, Sparga- nium, &c. In other cases however it consists of bundles, which do not take part, or at least not directly, in the construction of the cylinder: Palms, Scitaminez, and many Bromeliacez. : In the cortex of those Palms which have been investigated, smail bundles, which are arranged in irregular concentric rings, are found in the cortex, outside the dense periphery of the cylinder. P. Moldenhawer has compared the region in which they lie with the bast of Dicotyledonous trees (Chap. XLV), Mohl has named it the 1 Duval-Jouve, Diaphragmes vasculiféres, /.c.; compare p. 216. 2? Van Tieghem, Ann. Sci. Nat. 5 sér. tom. VI. p. 195. 8 Karsten, Veget. Org. d. Palmen, p. 98. 266 PRIMARY ARRANGEMENT OF TISSUES. fibrous layer. It is weakly developed in tubular stems like Calamus (Geonoma, Bactris, Hyospathe, Desmoncus, Calamus), and in the cylindrical stems of Mohl (Mauritia, Oenocarpus, Kunthia, Astrocaryon sp., &c.): more strongly in Rhapis flabelliformis, Phoenix, Jubea spectabilis ?): and most fully developed in Mohl’s Cocos-like stems: Cocos, Leopoldinia, Syagrus, Elais, Corypha sp. Mohl regards the bundles of the fibrous layer as being, at least in part, bundles of the leaf-trace, the lower ends of which pass out from the cylinder into the cortex, and pass down to the base of the stem as fine threads, either undivided, or split up into many thin branches (Cocos). More recent investigations? have shown that the bundles of the fibrous layer are not the ends of bundles which pass through the cylinder and emerge from it lower down, but that, as Mohl® also allowed to be the case with some of them, they pass from the base of the leaf directly into the fibrous layer. Here they run almost perpendicularly, and frequently show splittings and branchings, or fusions ; the latter especially in a tangential direction, both in the internodes, and also in the insertions of the leaves, so that those bundles which enter from a leaf are in direct continuity with those which descend from higher leaves. Most of the bundles in question in the fibrous layer consist of a number of sclerenchymatous fibres: in a strict sense therefore they do not belong to this category. But in others according to Mohl’s figures, and according to investigations on species of Chamzedorea and Rhapis, there are solitary small sieve-tubes, and in single cases also one or a few small trachee. And while some continue as purely sclerenchymatous bundles into the petiole, others, after their entry into the petiole, assume the structure of fully organised vascular bundles‘. Further, according to a note by Schacht, there is a connection with or transition to the vascular bundles of lateral roots, though the fibrous bundles do not arise ‘as branches’ of these. The cortical system of bundles of the Palms is accordingly a direct continuation both of the system of vascular bundles, and of the system of purely sclerenchymatous bundles in the leaves, and connects them one with another. Another cortical bundle-system occurs in the stem of Ananassa and Tillandsia acaulis Hort. The-parenchyma of the thick cortex is here traversed by numerous bundles of the trace, which pass obliquely downwards from the leaves into the sharply: limited cylinder. Other thin but also compiete vascular bundles pass from the base of the leaf into the cortex: here they pass down, sometimes close beneath the surface, sometimes deeper, but always far removed from the cylinder, through several inter- nodes, and then insert themselves on one of the main bundles, and with it enter the cylinder. The general direction of their course is approximately perpendicular, or arched according to the dome-like shape of the end of the stem: they are besides curved in a sinuous manner to a very variable extent. It is doubtful whether the anastomoses of the bundles of Ananassa, described by Unger, are these cortical / 1 Wossidlo, Quzedam additamenta ad Palmarum anatomiam, Diss, Vratisl. 1860, and.Nova Acta Leop. Carol. vol. XXVIII. : 2 Schacht, Lehrbuch, I. p. 327.—Niageli, 7.c. p. 132.—Wossidlo, Zc. 3 Palmarum Structura, p. xviii; Verm. Schriften, pp. 155, 184. * Mohl, Wossidlo, /.c. 5 Dicotyledonenstamm, p. 50, figs. 23, 24.—Anatomie und Physiologie, p. 232. COURSE OF THE BUNDLES IN THE STEM. 2.67 bundles, or the above-mentioned isolated communications which occur within the cylinder. In most Scitaminez (Musacez, Zingiberaceze, Cannaceze’), as far as descriptions go, the vascular system in the cylinder is of the Palm type; but outside the cylinder there is a system of complete cortical vascular bundles. According to Wittmack’s description of Musa Ensete they are ‘completely limited to the cortical layer, and have a very sinuous, aimost zigzag course, especially in the lower part. Owing to their crowded arrangement and their frequent crossing, it was very rarely possible to follow their traces far. But in favourable cases it appeared that they approach rather near to the epidermis, and then proceed upwards parallel to the surface: but when- ever they thus approach the base of a leaf, they curve inwards, and form anastomoses with the chief vascular bundles (bundles of the trace), till finally they themselves enter one of these leaf-bases, together with the large bundles which pass into it from the interior of the stem. They here usually turn to the outer or inner wall of the leaf-sheath, rarely they may be seen penetrating into the central regions, which are traversed for the most part by the main bundles.’ Wittmack found the same arrangement in all the nine species of Musa investigated by him, in Strelitzia reginz (weak), in the rhizome of Curcuma Zedoaria, in the flower-stalk of Phrynium viola- ceum, and Calathea grandiflora: Meneghini found it previously in species of Ravenala, Hedychium, and Canna. Sect. 68. As far as hitherto investigated, the single vascular bundle of the cotyledon in the seeding of the Monocotyledons of the above types runs directly into the axis of the main root, e.g. Allium Cepa?: or the cotyledon contains several bundles, and these unite in the cotyledonary node, and then similarly pass into the bundle of the root: e.g. Palms*. The bundles of the leaves which follow after the cotyledons show the typical course, with such modifications as result from the small number of bundles and the shortness of the internodes to be traversed. They unite in the cotyledonary node with those of the cotyledon and of the root. . The elongated internode between the insertion of the scutellum and the first sheathing leaf in the seedling of Zea Mais shows an abnormal arrangement. It contains a ringlike mass of vascular tissue, which surrounds a wide pith: this ring is continued at the point of insertion of the scutellum into the bundle of the first root. The annular mass originates by the coalescence of the lower ends of the numerous bunales of the first foliage leaf- traces with the trace of the first sheathing-leaf. The latter contains as a rule two bundles situated right and left and in front of the median line. In the node they both bend in- wards, and somewhat to the rear, and anastomose on reaching the node by means of a curved connection: they then spread out and pass down in the ring. A branch goes also from the curved connecting bundle perpendicularly downwards. The traces of the foliage- leaves consist of many bundles, and are as wide as the whole circumference of the stem: the lower ends of the lowest pass beneath the node of the sheathing-leaf between and alongside of those of the latter, and with them they form the ring. According to investi- gations hitherto made—which require further completion—it appears that the individual 1 Meneghini, 7. c.—Wittmack, Musa Ensete, Halle (Linnzea), 1867. 2 Sachs, Botan. Zeitg, 1863, Taf. III. -§ Mohl, Palm. Struct. p. xliv. Tab. P.—Sachs, Botan. Zeitg. 1862, Taf. IX.—Compare also the data on this point in van Tieghem, Symmetrie de Structure, &c., Ann. Sci. Nat. § sér. tom, XIII. 268 PRIMARY ARRANGEMENT OF TISSUES, bundles of the trace cannot still be distinguished in the ring. More rarely there are in the sheathing leaf in addition to the two lateral bundles two other smaller bundles placed symmetrically in the posterior half, near to the median line of the sheathing-leaf. There was found in one case in the transverse section of the internode, in the centre of the anterior side and outside the ring, a small isolated bundle, the origin and course of which remained doubtful. The terms anterior and posterior are in all cases to be understood here, so that the side next the scutellum is the posterior. In conclusion, a short connected description may here be given of the peculiarities of the bundle-system in the Aroidez and Pandanacez, so often alluded to above *. A number of forms do not differ from the Palm type, except in the fact that in many of them the bundles run for a long distance in the cortex before entering the cylinder, Asecond category is distinguished from the first by the bundles being united on their descending curved course within the cylinder, and at a considerable distance from its surface. In transverse sections therefore there are found within the peripheral layer ‘compound’ bundles, i.e. such as are cut through at the points of coalescence or separa- tion, Finally, in a third group the bundles are not only united on their entry into the middle of the cylinder, but are connected by anastomoses in all directions, and in such a way in specially good cases that, as in the Nympheacez (p. 252), there is close beneath the growing-point a complex network, which branches in all directions, and takes up the leaf-traces as they enter, so that the typical bundle system can only be recognised by slight indications. ; To the first category belongs in the first place the rhizome of Acorus gramineus and A. Calamus: the majority of the bundles descend obliquely, as in the above-mentioned Cyperacez, through several internodes in the thick cortex, and this, as is especially well seen in A. gramineus, is traversed by bundles arranged in several rings in transverse sections. Further a number of epiphytic forms with elongated internodes belong here: all the Monsterinez investigated by v. Tieghem (species of Heteropsis, Monstera, Raphidophora, and Scindapsus), with bundles of the trace, which sometimes enter the leaf at once at the node, but for the most part traverse the cortex through two internodes before passing into the leaf: further the investigated species of Anthurium and Pothos, in which also cortical bundles are present, which vary in number and distribution according to the species, with the exception of A. Miquelianum, which belongs to the simple Palm type. The second category is connected with the above forms by the investigated species of the genus Philodendron. As in the former, some of the bundles traverse the cortex through (two) internodes before they pass out into the leaf. In Ph. micans all the bundles take an individual course through the internodes, and are only united at the nodes, the joint bundles then pass downwards, and end at the periphery of the cylinder. In other species (Ph. Rudgeanum, hastatum, tripartitum) the points of coalescence are within the cylinder in the internodes also, so that the transverse section often shows ‘compound’ bundles side by side with simple ones. According to v. Tieghem this arrangement is found, subject to many modifications according to the elongation of the internodes, the presence or absence of cortical bundles, &c., and specific peculiarities, in all investigated Aroidex with moderately elongated internodes and unisexual flowers (species of Homa- lonema, Aglaonema, Dieffenbachia, Syngonium, and others to be named below), and of those with bisexual flowers in Calla palustris, Lasia ferox, and Spathiphyllum. To this series belong also the stems of the Pandanez (Pandanus javanicus, pygmzus). The thick, cylindrical (Alocasia), or almost tuberous stems, with short internodes, of other unisexual Aroidex, Alocasia odorata, Colocasia antiquorum, Caladium esculentum, Dracunculus, Arum, Richardia aethiopica, &c. belong to the third category, since their bundles are not 1 See van Tieghem, /.c, | COURSE OF THE BUNDLES IN THE STEM. 269 only frequently united within the stem, but are connected in a reticulate manner by anastomoses. Many forms, such as Syngonium, are transitional between the second group and the third. In small branches of rhizomes of Richardia xthiopica also the typical curved course may be easily recognised, especially in the ends of the branches; in Alocasia, Dracunculus, Caladium esculentum the system of vascular bundles forms ‘a complicated net, in which it is impossible, by the most careful dissection of a macerated stem, to trace one bundle with certainty even for a‘short distance.’ Still even here it may constantly be recognised, especially at the apex of the stem, how the bundles pass from the base of the leaf with a curved course to the middle of the stem, and from thence downwards and outwards, ta FIG, 119. FIG. 120. FIGS. 119 and 120.—Tradescantia albiflora. Fig. rr9, Outline and vascular bundle-system of the end of a stem cleared by potash, and divested of part of the cortex and cylinder by longitudinal sections: The surface of section is turned towards the observer. Those bundles which in this position run in a higher plane are drawn darkly, those in lower planes more lightly. The successive leaves are indicated by the figures 15; leaf 6 is just formed at the growing-point ; opposite the median bundle of r is an axillary bud. The bundles of leaves 1, 2, and 3 are visible; /the separate, ¢the united portions. From 3 are to be seen four marked /; two of these (right of the median line) run into the leaf; two others, median and anterior, are cut off at their point of exit. From leaf 2 are three marked /; the middie one cut off. The lines s show the course of the cauline bundles. Magnified about 25. 2 : « FIG. 120 (40).—Transverse section through a young internode. In the middle the four united bundles; externally eight separate ones, the three principal ones marked J; then the circle of ten cauline bundles. IV. Type oF THE CoMMELINACES. Sect. 69. The bundles in the stem of those Commelinacez which have been investigated, and of many Potamogetons, have a course which differs from that of most Monocotyledons, and resembles rather that of the Piperacez (p. 249) and Mirabilis (Chap, XVI). This is seen particularly clearly in the plant widely known 270 PRIMARV ARRANGEMENT OF TISSUES, ~ in gardens as Tradescantia albiflora, and will be described first in that example (Figs. 119 and 120). From the sheath-like base of the alternate distichous leaves, which embraces the stem, as a rule eight bundles curve into the node (1), and thence descend perpen- dicularly to the next node (2). In the internode they are at about equal distances laterally from one another, and at varying distances from the middle of the stem, but at Jeast 4 of the radius from it. Close above node (2) they curve towards one another and the middle of the stem, and unite in the node itself in pairs to form four bundles. These four bundles are stronger than the original eight: they are arranged crosswise near to the middle of the stem, and run perpendicularly downwards to the next node (3), where each inserts itself on the point of junction of two which pass out at node (2). Each internode accordingly shows in transverse section (Fig. -120) in the first place twelve bundles, four inner ones, arranged crosswise, and around them an irregular circle of eight weaker ones (4). Besides these twelve bundles of the leaf-trace there are also usually 11-12 bundles (of subsequent development) which are arranged in a circle outside the eight external bundles of the trace. This circle, together with the rather small-celled parenchyma between the bundles, marks off the cor/ex from the cylinder which contains the vascular bundles. Some of these bundles lie also further inwards, between the eight outer bundles of the trace. These 11-12 bundles do not pass out into the leaves, but run up into the youngest internode, passing almost per- pendicularly through the internodes, curving slightly inwards at the nodes, and passing near to the outgoing bundles of the trace. Irregular short transverse bundles connect them in old nodes with one another, and with the bundles of the trace. Deviations from the above numbers are often found, e.g. in the lower internodes of lateral shoots, where there are often in all only 18-19 bundles visible in the transverse section—e.g. 3+ 6 bundles of the trace, and 10 cauline bundles, I found fundamentally the same arrangement both of the bundles of the trace and of the cauline bundles in all the other Commelinacee investigated: Commelina agraria Kth., C. procurrens Schl., Tradescantia zebrina, virginiana, Spironema fragrans, Dichorisandra thyrsiflora, D. oxypetala, Maravelia zeylanica. But in all these the number of bundles of each category is higher than in Trad. albiflora, especially in the last-named six species with a thick stem, and leaves with many bundles. The arrangement of the separate and united bundles is accordingly more complicated and requires further investigation. Potamogeton natans (Fig. 121) has alternating leaves in two rows: these are often displaced from this arrangement (by torsion of the stem?): the leaf-trace consists of three bundles, the width of the leaf-trace is about 180°. The three bundles of the latter curve towards the middle of the stem, and pass separately down one internode, the stronger median bundle being nearer the middle line than the two lateral ones. In the next node they all three coalesce to a single bundle, which then passes down to the second node, and here inserts itself at the point of coalescence of the next lower trace (rarely one of the lateral bundles continues a separate course up to this point of insertion, Fig. 121, x). In the internode there appears accordingly in the bluntly rectangular transverse section (Fig. 122) of the ‘cylinder’ which con- tains the bundles, one large bundle at each end of the smaller diameter ; these are Opposite one another: one of these (1) is the united trace of the second higher leaf, COURSE OF THE BUNDLES IN THE STEM. 271 the other rather smaller one (2) is the median bundle of the next higher leaf. At each end of the longer diameter is a small bundle (2, 2): these are the lateral bundles of the last-named leaf. On both sides of each lateral bundle, that is opposite each corner of the rectangular transverse section, there is further a small cauline bundle (s, s). The median bundles appear first, then the lateral ones, and the cauline bundles much later. At an early stage irregular anastomoses appear in the nodes between all of them, the small cortical bundles of sieve-tubes and fibres also taking part in these (p. 232). Among the other Potamogetons which have been investigated, P. perfoliatus has fundamentally the same bundle-system. Also the rhizome of P. pectinatus FIG. 121.—Potamogeton natans (40). End of the stem, made transparent; the outer layers of tissue removed in the lower part by longitudinal sections; successive leaves and their belongings numbered in order, #¢ median bundle; / lateral bundles of the leaf indicated by the figures; ¢the united portions; x a bundle of leaf 4 which runs an exceptional distance separately. The median planes of the leaves 1—3 lie alternately right and left in the plane of the paper (opposite the middle of 2 is the axillary bud belonging to it). Above 3 is a rotation, so that the median plane of leaf 4 is at the front, while that of 5 is at the back. The median bundle of 5, 57, therefore runs deep down ‘into the preparation. The lateral bundles of 5 and the cauline bundles are (as yet) not visible ; =the stipular sheath © of the base of the leaf. appears, according to an incomplete investigation, to resemble it. In other species the structure resembles that described, but is simpler, and reduced in proportion to the average decrease in size of the leaves. ‘ P. lucens and P. gramineus have a leaf-trace consisting of a single bundle, which does not divide into three bundles till its exit at the node into the leaf. Each passes 272 PRIMARY ARRANGEMENT OF TISSUES. down, near to the middle of the stem, and close to that of the next lower leaf, through one internode, and then unites with the latter in the node (comp. Fig. 124). Ina transverse section of the internode there are accordingly two bundles of the leaf- trace, which are close to the centre in the diameter between the median lines of the two rows of leaves. A small vertical cauline bundle appears at a later stage than the leaf-traces and near to them, and this lies in the radial longitudinal plane at right angles to the plane of the median lines of the leaves: in the nodes transverse and oblique anastomoses appear at an early stage, as in P. natans. P. densus shows fundamentally the same structure, with the striking difference that each bundle of the leaf curves almost at right angles into the middle of the stem, and inserts itself in the next lower node directly on the bundle which there passes FIG. 122 (145).—Potamogeton natans. Axile body of the internode, containing the vascular bundles; transverse section. 7 endodermis thickened on one side (with starch) ; outside this the lacunar cortical parenchyma (with much starch); Z air cavities. Explanation of the figures in the text. The groups of delicate tissue of the numbered areas are the phloem portions; the wide meshes in them are the sieve-tubes of the bundles; the areas in which the figures stand are their vascular portions (for the most part converted into cavities). Between the bundles is starchy parenchyma, and sclerenchymatous fibres, with a narrow lumen, which appears as a dark point. out, so that only one axile sympodium of the leaf-trace is present besides the two cauline bundles. In the upright stems of P. pectinatus (Fig. 123), in P. pusillus and Zanichellia palustris this axile sympodium is alone present, without the two cauline bundles, to which fact we shall return later. P. crispus shows a somewhat different arrangement, which will be described below. ‘ It is not improbable that this type of vascular bundle-system is allied to that of Hydrocharis, Stratiotes, and their allies, but further investigations are wanted on this point. : The transverse section of the stolons of Hydrocharis Morsus Ranz ' shows four bundles * Rohrbach, Beitr. zur Kenntniss einiger Hydrocharideen, Abhandl. d, Naturf. Ges. z, Halle, Bd. XII, p. 75. COURSE OF THE BUNDLES IN THE STEM, 273 arranged crosswise, two larger and two smaller, the similar ones being opposite one another. A few layers of cells below the epidermis a circle of eight to ten small bundles is found, which run perpendicularly and separately through the internode. In the short thick stem of Stratiotes aloides' the bundles—all of which descend from the leaves—unite ‘after numerous anastomoses to one central bundle and eight ° Fic. 123.—Potamogeton pectinatus; end of the shoot (4o}. Thick median section, cleared with potash, parallel in the lower part to the median planes of the two rows of leaves ; higher up, from leaf 5 onwards, this plane Eee rotated through almost 90°, Successive leaves numbered successively 16; v the opposite longing to the corr dingly numbered leaf; s asguamzda intravaginalis of leaf 3. In the middle is plainly seen the vascular portion of the sympodium of vascular bundles; the upper- most bundle, which is clearly seen, goes to leaf 6 (which lies at the back of the section); that which goes to 5 is cut transversely, since 5 was turned towards the observer, owing to the above-noted rotation. In the axils of each of the leaves (up to 5) an axillary bud is plainly seen. That belonging to leaf 2 lies deep in the section ; its parts are indicated by 4, and the vascular bundle which enters it by £4. That belonging to leaf 3 is still small; but the development of the first vascular bundle which enters it is beginning in the angle where the bundle 3+ (which runs into leaf 3) unites in the node with that which comes from 4, In the axil of leaf x is a strongly-developed axillary shoot ; 21 and 22 its scale leaves, /1 and 22 its two first foliage leaves; the vascular bundle which enters /2 is just beginning to be developed. In the axil of 72 a secondary axillary shoot is appearing; its first vascular bundle is beginning até or nine peripheral ones. In the thin stolons these bundles run perpendicularly through the internodes,’ 1 Rohrbach, Z.¢. p. 94. T 274 - PRIMARY. ARRANGEMENT OF TISSUES. Vv. Anxomatous MonocoTyLepons. Sect: 70. Under this term may be grouped some examples of vascular bundle- systems, which differ fundamentally from that of the very. great majority of Mono- cotyledons. Some of these are found in certain water-plants, the rest in certain Dioscoreze, the bundle-system of which approaches very closely to that of the Dico- tyledons. Potamogeton crispus, while it approaches very closely in other anatomical properties to other members of the genus, is distinguished from them by the course of the bundles in the stem. Comp. Figs. 124, 125. g 13 2 , Uz 3 U, FIG. r24.—Potamogeton crispus (40). End of a shoot. Longitudinal section parallel to the median planes of the two rows of leaves, cleared by potash. The successive leaves numbered 1, 2—10; Wie UD eevee the sheaths of the corre- sponding leaves; the sheaths of the upper leaves were obscured in the pro- cess of preparation, and are partly omitted in the drawing. The median vascular bundles of leaves 9 and 10 are just beginning to develope; the seven highest leaves are still without bundles. FIG. 125.—Potamogeton crispus (40). End of a shoot. Thick ‘median longitudinal section perpendicular to the median plane of both rows of leaves, cleared hy potash; the sheaths were obscured by the process of preparation, and are omitted in the drawing. The successive leaves are nuinbered in order. The series of unevenly numbered leaves are uppermost, nearest the observer; the even numbers are below. The same is the case with the median bundles, which go to the respective series and unite in the axile bundle 7, The median bundle is plainly seen as far as the sixth leaf from the apex; the Jateral bundles united to form the two bundles 2, Z, beginning at the eleventh leaf from the apex {s) in the node; the development of 4 and 5 is not yet completed downwards through the internode, In the internodes the air-cavities are first formed from below upwards, and from the outside inwards, In each of the sheathing leaves, which alternate in two rows, the bundles pass out at the node. The median bundles run down through the internodes in the manner described for P. lucens and gramineus (Fig. 124). The lateral ones (Fig. 125) pass on each side almost perpendicularly from a- bundle which traverses the stem perpendicularly, and corresponds exactly in position and relatively late appear- ance to the cauline bundles of the other species, so that the arrangement of the COURSE OF THE BUNDLES IN THE STEM. 275 bundles in a transverse section of the internode is the same asin them. The lateral bundles of the stem of P. crispus are however not cauline. The development of their tracheze begins at the nodes, and proceeds from each of these towards the middle of the next upper and lower internodes (Fig. 125, 4, 5). The course of the three bundles in the stem of Zostera marina’ is just the same as the above: one axile bundle is built up sympodially from the median bundles of the leaves, while two lateral ones, which lie in a plane cutting the median-plane of the bi-seriate leaves at right angles, give off lateral bundles to the leaves; these however require further investigation. The arrangement of the bundles in the transverse section of the internode of Zostera differs from that in Potamogeton, since in the latter the lateral bundles are close to the central bundle, while in the former they run at some distance from it, near to the surface of the stem. In Cymodocea zquorea Koen,’ the median bundle of the seven bundles of the trace runs obliquely down to the middle of the stem, it there passes perpendicularly to the next node, where it joins the median bundle, which there passes out. The transverse section of each internode thus shows a central bundle. Near to the periphery of the stem 20-25 small bundles (in weak stems fewer) arranged in two concentric circles pass perpendicularly through each internode. Each of these, ac- cording to Bornet, divides at the node into two, of which one ascends into the next higher internode, the other either curves out into the leaf, or unites either with a neighbouring peripheral bundle, or with the axile bundle. Besides this a complex net- work of anastomosing bundles is formed in the node between the different bundles. The peripheral ones appear to be cauline, but this requires to be further investigated. As far as may be judged from transverse sections Cymodocea isoetifolia exactly re- sembles other species. After what has been said it need not be stated in detail how Hydrocharis and Stratiotes might, as far as our present knowledge goes, belong equally well to this as to the préceding section: this will have to be decided by further investigation. Scr. 71. In the foliage-shoots of Tamus and Dioscorea Batatas the vascular bundles are arranged according to the Dicotyledonous type, that is in a ring sur- rounding the pith. It is true that here the bundles pass an unequal distance into the pith, but this also occurs in typical Dicotyledons. Nageli (/.c. p. 123) gives the following description of Dioscorea Batatas. The leaves are sometimes spirally arranged, sometimes in decussating pairs. The leaf- trace consists of three bundles. When the arrangement is decussate (Figs. 126, 127) its width is about 120°. If their course in a tangential direction be first con- sidered (Fig. 126), the six bundles of one pair of leaves pass nearly straight down two internodes, the lateral bundles (dc, ef; 42, lm; op, rs; uv, yz) pectinating at the first internode with the lateral ones which there enter. Above the bundles of the second lower node the two lateral bundles of one trace converge towards one another, and insert themselves on lateral bundles of the next lower pair; but the median bundle (a, d; 4, #; 4 x) divides into two shanks which unite with the same 1 Compare Magnus, Botan. Zeitg. 1872, p. 216. 2 Bornet, Recherches sur le Phucagrostis major, Ann. Sci. Nat. 5 sér. tom. I. Compare especially p. 39, pl. 6. fig. 1, and pl. 11, fig. 1. T2 276 PRIMARY ARRANGEMENT OF TISSUES, lateral bundles~ The leaf-trace is here much contracted, and usually forms by ‘coalescence one single mass. The six bundles of one pair of leaves thus go through only two internodes before they coalesce with lower ones, and the transverse section through an internode shows twelve bundles (Fig. 127), of which six pass out at the next node and six at the node above it. These twelve bundles would be arranged in a circle if they had a radially perpendicular course. But this is not the case: they penetrate further into the pith as they pass down lower. Their radially oblique course is however restricted almost entirely to the nodes: the same bundle thus appears at unequal distances from the centre in the two internodes. The lateral bundles, which in their first internode are nearer the centre than the median ones, differ still more from them in this respect in the next lower internode. The transverse section through an uct v Y £# | Sy ae ~s v oy sy % ely? VUU ALI z poy Fig. 126. Fig. 127. eer FIGS. 126, 127.—Dioscorea Batatas, with decussating pairs of leaves; after Nigeli. Fig. 126. Scheme of the vascular bundle-system in the end of a shoot, the cylindrical surface being reduced to a single plane. Fig. 127. Transverse section through an internode, the same bundles being indicated by the same letters as at the bottom of Fig. 27; further explanation in the text. internode therefore shows four bundles which project further inwards and form a rectangle,—these are the lateral bundles of the next higher pair of leaves,—and eight outer bundles. On the longitudinal course where the leaves in Dioscorea are arranged spirally, and on Tamus, compare Nageli, /.c. The course of the bundles in Monocotyledonous stems has recently been the subject of extended researches by Falkenberg?. The chief result of that work which * P. Falkenberg, Vergl. Untersuchungen iiber den Bau der Vegetationsorgane der Monocotyle- ‘donen, Stuttg. 1876. : COURSE OF THE BUNDLES IN THE STEM, 277 concerns us here was briefly stated in 18741: it is the discovery of a new form of bundle system in the /oldage- or flowering-stems of Lilium, Tulipa, Fritillaria, Cepha- lanthera, Epipactis, and Hedychium. The course of the bundles is such that the bundles of the leaf-trace penetrate downwards, and for different distances inwards to- wards the middle of the stem, and then affix themselves on corresponding bundles of lower leaves, wethout having previously curved outwards. For the rest we may here refer to Falkenberg’s comprehensive work, and print off the above paragraphs with- out alteration as they were written down about four years ago, since they coincide in the main with his work. VI. Puanerocams witH an Axite BunDLeE. Sect. 72. A number of plants with reduced foliage and roots living in water or marshes, and some in damp humus, belonging partly to the Monocotyledons, partly to the Dicotyledons, show the bundle-system of the stem united into one bundle, which is surrounded by a thick cortex, and traverses the middle of the stem longi- tudinally : from it bundles pass at the nodes into the leaves. With this simplicity of course there.is usually connected a considerable simplification in the structure of the bundles, which always shows peculiarities: we shall return to them in Sects. 105 and 110, As regards the coarser structure, and especially the relations to the bundles of the leaf-trace, which it may be said require more exact study in many cases, there may be distinguished two chief forms. Firstly, axile bundles, formed or developed sympodially from weak bundles of the trace, which approach closely to one another, and coalesce longitudinally: thus they do not differ in their first origin from the typical bundle-systems of the Phanerogams: secondly, such as are cauline, and grow acropetally with the end of the stem; the bundles which run to the leaves are given off from the former at the nodes, or apply themselves to them as branches : thirdly, those cases in which the axile bundle is built up of longitudinally coalescent bundles of the leaf-trace together with cauline bundles are connected with the two former cases as intermediate forms. To the first category belong the following Dicotyledons: Bulliarda aquatica ac- cording to Caspary’s account®, Hottonia, Elatine Hydropiper, hexandra, also E. Alsinastrum 8, and probably Trapa natans: of Monocotyledons, Potamogeton pectina- tus and pusillus, to which may be added Zanichellia* and Althenia®; also Ruppia® and its allies. To the second category the following Dicotyledons belong: Aldrovandia’, } Botan. Zeitg. 1874, p. 732. 2 Schriften d. Physical. dconom. Gesell. zu Kénigsberg, Bd. I. 1860. 3 (Comp. Friedrich Miiller, Struktur einiger Arten von Elatine, Flora, 1877, p. 481.] * Schleiden, Beitr. p. 215.—Caspary, Pringsheim’s Jahrb. pp. 383, 440. 5 Prillieux, Ann. Sci. Nat. § sér. tom. II. ° Compare Irmisch, Ueber einige Arten d. Familie d. Potameen (Abhandl. d. Naturwiss. Vereins f. Sachsen und Thiiringen, 1858), p. 44. 7 Caspary, Botan. Zeitg. 1859, p. 126, Taf. V. Ibid. 1862, p. 193. 278 PRIMARY ARRANGEMENT OF TISSUES. Hippuris*, Callitriche*, Myriophyllum *, Ceratophyllum‘, probably Utricularia, and the non-aquatic genus of Piperaceze Verhuellia®: of Monocotyledons, the Hydrillee, Elodea canadensis, and Hydrilla verticillata®; Najas’, and the rhizomes of the root- less Orchids which grow in humus, Epipogon Gmelini, and Corallorrhiza innata*. In Corallorrhiza and the stolons of Epipogon a branch passes from the axile bundle into each of the biseriate scale-leaves: in the coral-like rhizome of Epipogon, which has short internodes, the branches which pass to the leaves are absent, according to Reinke. According to the structure and development of the bundle the following plants may be placed in the third intermediate category ; perhaps Myriophyllum, Hippuris, and Elatine Alsinastrum: further the larger Potamogetons belong to this series: their original arrangement of cauline bundles and bundles of the trace has been described above on p. 272. In the series of described: species of this genus, and of allied forms such as Zannichellia and Althenia, with which Elodea, Najas, &c. are also connected, there are to be found all stages of simplification of composition (and of structure) of the axile bundle: leaf-traces consisting of one or more bundles, running side by side with cauline bundles through the internodes, are found in the stronger, more leafy forms, at one end of the series, while at the other a single cauline bundle is present, which unites at the node with the bundles from the leaves. VII. Fern-iixe Prants. Sect. 73. In the young seedlings of all the investigated forms of this series the bundle-system of the stem is a sympodium of leaf-traces consisting of one bundle (which with the exception of Equisetum developes in an acropetal direction). The first bundle, which usually ends blind in the foot of the embryo, curves after a very short course through the stem into the first leaf; from the point of curvature the development of a bundle, which runs out into the second leaf, begins. In the case of the subsequent leaves the same conditions prevail. In Isoetes, Equisetum, and Osmundacew, this construction of the bundle-system out of distinct leaf-traces is permanent even in the mature stem. The same holds perhaps for many Ferns with a simple axile bundle. In the Lycopodiums, and Selaginellas, the axile bundle which traverses the stem, or the two or more bundles 1 Nageli, Beitr. /.c. p. 56.—Sanio, Botan. Zeitg. 1865, p. 191. ; a 2. c_—Hegelmaier, Monogr. d. Gattung Callitriche. Idem in Martius, Flora Brasil. asc. 67. * Vochting, Zur Histologie u. Entwicklungsgeschichte v. Myriophyllum, Acta Acad. Leopold. XXXVI (1872). * Schleiden, Beitr. p. 216.—Unger, Anatom. und Physiol. 198.—Sanio, Botan. Zeitg. 1865, p. 192. _ Schmitz, Flora, 1872. * Caspary, in Pringsheim’s Jahrb. I. Idem, Verhandl. d. Naturforscher u. Aerzte z. Kénigs- berg, 1860. 7 Compare Magnus, Beitr. z. Kenntniss d. Gattg. Najas, p. 48. * Irmisch, Beitr. z. Morphologie und Biologie d. Orchideen.—Schacht, Pflanzenzelle, p. 268. Idem, Lehrbuch, IJ. p, 21.—Reinke, in Flora, 1873. COURSE OF THE BUNDLES IN THE STEM. 279 ‘of many Sélaginéllas, may, as regards development, be considered as a cauline ‘bundle, the corners of which are composed of the sympodially united leaf-traces of a single bundle. ‘On the. other hand, the Lycopodiaceous plant, Psilotum triquetrum, -has only a cauline bundle without leaf-traces. Also in the case of Marsilia and Pilularia -a similar view may be held, in common with Nageli, according to the development of ‘the bundles. In the majority of the Ferns there is an obvious connection between ‘the form and arrangement of the leaves, and of the bundles which enter them; in a number of cases, especially in those forms, to be described later, with a reticulate stem-system, and one bundle for each leaf, the stem-system may be recognised as being composed of the constituent leaf-traces+. But in very many cases such a separation cannot be carried out according to our present knowledge without arbi- trary treatment, but rather a bundle-system of varying form and complication may be distinguished in the stem, from which bundles for the leaves are given off at certain points. The following description must accommodate itself to the facts: in each case those stem-systems, which may arbitrarily be recognised as being composed of leaf-traces, will be associated with those with’ which they correspond most closely in their real structure. There may accordingly He distinguished on the one hand the types of Equi- setum, Osmunda, Isoetes, on the other the various series of types of the Ferns, which are connected by numerous intermediate forms: under the latter the Lycopodiums and Selaginellas may be ranged as peculiar instances, and are here co-ordinated merely for synoptical reasons, Sect. 74. Equisetum’?. The weak bundles of the stem are arranged in a ring separating the pith and cortex. From the median line of each tooth of the leaf- sheath one bundle enters the stem, it here passes perpendicularly down one inter- node, and then divides, at the next lower node, into two short shanks, each of which affixes itself on the nearest lateral bundle which here passes out. Where the number of teeth of successive sheaths is the same the bundles of successive inter- nodes alternate as they do. Sect. 75. Osmundacese*. Comp. Figs. 128-130. The mature rhizome of Osmunda regalis has leaf-insertions arranged with a divergence of 3%, and short ‘internodes. Its centre is occupied by an irregular bluntly five-cornered prism, with a thickness of about 6™™ in strong specimens: this consists of a vascular-bundle- cylinder (Ring), a narrow sheath of delicate-celled parenchyma surrounding the latter, and a parenchymatous pith surrounded by the ring of bundles, and with brown sclerénchymatous cells scattered through it. The prism is enclosed by a cortex 2-5mm in thickness, which is dark-brown and sclerotic, but contains much starch: this is traversed by vascular bundles, also surrounded by a thin sheath of delicate parenchyma, on their oblique upward course from the ring into the leaves (Fig. 128). 1 See Holle, Botan. Zeitg. 1875, p. 265, &c. . ? Nageli,:Zeitsch. f. wiss. Bot. 3. p. 143; Beitr. /.c. p. 57—Cramer, in Nageli und Cramer, Pflanzenphys. Unters. Heft 3, p. 21 —Hofmeister, Vergl. Unters. p. 93.—Duval-Jouve, Hist. Nat. des Equisetum de France, 1864. * Géppert, Flora, 1848, Taf. IV. A.—Unger, Denkschr. d. Wiener Academie, Math. Na Eieg: Classe, Bd. VI (1853).—Milde, Monogr. Osmunde, p. 32. 280 PRIMARY ARRANGEMENT OF TISSUES. One bundle enters each leaf: the arrangement of the bundles in the cylinder is quite q aD 4 op aye 10-2 Fig. 129. Fig. 128. FIGS. 128 and 129.—Osmunda regalis. Fig. 128. Transverse section through a strong stem, seen from above, i. e. from the ‘apex of the stem; magnified about twice. 7 lowest bundle of leaf-trace, from which a root-bundle passes through the cortex. Fig. 129.—Sketch of the bundle-ring in the former figure, more strongly magnified, 1. Bundle of the lowest trace cut through just at its point of entry into the ring, and with one of the two root-bundles, which are here attached to it. The figures 1—13 indicate the bundles of the trace of the thirteen successive leaves, which are visible in the transverse section; No. ro is abnormally united with 2 Compare Fig. 130. similar to that of the Coniferze with alter- nating leaves with a single bundle (Fig. 130). From one leaf # one bundle enters the cylinder, and runs almost perpendi- cularly downwards, as a rule through 13 in- ternodes, and then, when close to the leaf n-13 perpendicularly below it, it curves towards the ascending side of the bundle belonging to the leaf 7~8, and unites with it. In the cases investigated the insertion and coalescence took place occasionally even after a shorter course, e.g. in the case of the bundle ro in Fig. 130, eight IS ATT ele SS ms +S — ads SX « monde ees WAT ANI 1 T FIG, 130.—Osmunda regalis. Scheme of the vascular system in the stem, the cylindrical surface being reduced to a single plane; arrangement of the leaves 5/13. The bundles of the leaves num- bered in genetic series at their point of exit; each has two roots inserted close to the point of exit, and indicated by short transverse strokes. On 2 and 10 is to be seen the anomaly observed in the specimen upon which the scheme is founded, viz. that 10 inserts itself on 2 close to the point of exit of the latter, instead of con- tinuing its course to 2—5. internodes below the point of exit. The bundles are strongest at their point of exit from the cylinder, and are of horseshoe shape. In the petiole they retain this form, or are at least half-moon-shaped. In the cylinder of the stem, as they pass downwards, they decrease in thick- ness at first gradually, and then quickly, and assume a wedge-shaped transverse section. Here they are separated from one another by narrow bands of parenchyma (medul- : lary rays). The structure of the whole transverse section (Fig. 128) may be deduced from the above description.— The bundles, which pass into the first leaves of the seedling, unite to an axile bundle without any pith: this gradually extends into the ring of bundles surrounding the pith. In Todea africana and T. hymenophylloides are seen phenomena exactly similar to those in Osmunda, which need not be described in detail here. Sect. 76. While thus the. bundle-system of these plants, and of the Equiseta, may well be ranged under the type of Dicotyledons, and is specially allied to that of the Conifers (Juniperus, Widdringtonia*), the species of Isoetes have in their extremely shortened tuberous stem an axile bundle without pith, as is the rule in submerged water-plants: this bundle is built up sympodially by the coalescence of the inner ends of the one-bundled leaf-traces. Phylloglossum may also belong to this category ?. Sect. 77. Psilotum and Lycopodium. The leafy stem of Psilotum triquetrum® has one vascular bundle with 2-8 corners which project more or less from the surface. It is cauline throughout, the small leaves have no vascular * Compare above, p. 246, and Geyler, /,c., especially Taf. IV. ? Compare Mettenius, Botan, Zeitg. 1867, p. 98. 3 Nageli, Beitr. 7, ¢. p. 52. COURSE OF THE BUNDLES IN THE STEM. : 281 “bundles, Still, according to Nageli there is a relation between the corners of the bundle and the insertions of the leaves. ‘At some distance (about 3-8™™) per- pendicularly below each leaf one corner of the bundle projects very strongly, and gradually loses itself below, but rather more quickly above. The corners of the bundle gre therefore the more numerous in a portion of a stem, as the vertical rows are more numerous of the otherwise irregularly arranged leaves, which can only with difficulty be referred to a cyclic arrangement.’ The leaves of the Lycopodia! are arranged, according to the species and indi- vidual, in alternating whorls of two or more members, or spirally with a divergence of $> Ya» rs» &c. Each contains one thin vascular bundle. The stem is traversed by one strong, almost cylindrical axile bundle, in which the symmetrically distributed bands of tracheides, to be described in Sect. 107, form external protrusions, which, like the bundles passing into the leaves, and the above-mentioned corners in Psilotum, consist of narrow spiral tracheides, and may like them be briefly termed corners. The bundles of the leaves insert themselves (when followed from the base of the leaf) after a curved downward course through the cortex, on the corners of the axile bundle. At the beginning of the differentiation of tissues, there is at first a bundle of spiral tracheides at the corner, which forms a direct continuation of that which passes into the leaf: it passes down through some internodes, and then. inserts itself on the point of curvature of one which passes out lower down. It is only later that the more internal masses of larger tracheides are developed. From these facts, and according to the phenomena of development, the axile cylinder may be characterised as a cauline bundle, on the corners of which the sympodially united bundles of the leaf-trace are directly inserted. The same facts, however, admit equally well of our speaking of a polyarch (Sect. 107) axile bundle, which gives off branches from its corners into the leaves. The origination of the bundle which passes into the leaf follows very soon after the protrusion of the young leaf itself. The development of each bundle of spiral tracheides begins where the bundle inserts itself upon the point of curvature of a lower one, and proceeds towards the apex of the leaf in question: then from the point of curvature it proceeds again in the same direction to a higher leaf. This happens very rapidly, at r least in the stem itself, so that only in favourable cases (in L. alpinum) could Hegelmaier find a bundle of spiral tracheides, of which the portion passing through the cortex to the leaf was not already equally developed with the lower part, which passes down the corner; and Cramer ascribed to L. Selago a simultaneous development of the whole bundle from its lower point of insertion to the apex of the leaf—while Hegelmaier found a basipetal direction of development in the leaf itself. . The corners of the axile bundle, as also the rows of leaves, differ greatly in number in different species, individuals, and shoots of different rank of one individual, and correspond to one another as a rule neither in number nor arrangement in one and the same shoot, while apparently the correspondence is less close the greater the number of both. It is true Hegelmaier found a correspondence of both in 75 per cent. of the last branchings of L. alpinum which are covered by four rows of leaves in decussating pairs, and in about 60 per cent. of the branches of L. complanatum. But in most cases the number of 1 Nageli, Zeitschr. f. wiss. Bot. Heft 3 and 4, p. 132,—Cramer, in Nageli und Cramer, Beitr, Heft 3.—Hegelmaier, Botan. Zeitg. 1872, p. 789, &c,—Sachs, Textbook, 2nd Eng. Ed. p. 468. 282 PRIMARY ARRANGEMENT OF TISSUES, corners is smaller than that of the rows of leaves: in L. Selago, e.g. with whorls of five members (that is ten rows), there are 4-6, in L. inundatum with 3 arrangement of the leaves 4 or 5, in the above-mentioned branches of L. alpinum 3, &c. From Hegelmaier’s statement that in the main vegetative axes of L. clavatum and L. annotinum, with a diver- gence of 2, 2, 3, there were found 10-17 corners, it would appear that there is a higher number of corners than of rows of leaves. * Where the rows of leaves correspond exactly to the corners, all the bundles of one row of leaves insert themselves on the same corner. In other cases one corner may take up bundles from one row only, but must also often take them from more than one row, Usually it only takes up the bundles of two neighbouring rows, but sometimes also single bundles of more distant rows’. The bundles insert their inner end irregularly, some- times on the right, sometimes on the left, and sometimes on the inner side of the next lower bundle, Sect. 78. Selaginella. A number of species, forming no doubt the majority, such as S. Martensii, S. helvetica, pubescens, rupestris, &c., have in each shoot one axile, ribband- or plate-like vascular bundle, the faces of which in relation to the ground are directed upwards and downwards, while the margins are lateral to the right and left: in some, as S. pubescens, the bundle is provided at the middle line of its under surface, and near to each lateral margin with a sharp band-like process. The leaves have each one small bundle, and these behave in their course and insertion one on another, and also on the cauline por- tion, similarly to the leaf-bundles of Lycopodium. -They are inserted on the bundle at its margins: in the species with two double rows of leaves, one facing towards each side, the bundles of the two corresponding rows (that is those from one row of upper and one of lower leaves) insert themselves on each lateral margin: in S. rupestris with leaves in many rows, the bundles of several rows pass to each margin. S. Kraussiana, Galeottii?, and most other articulate ® have, in- stead of one axile bundle, two which run near the middle line, each follow- ing one double row of leaves: each takes up the leaf-bundles on its own FIG. 13t,—Selaginella inzequalifolia ; transverse section of the . * stem (150). From Sachs’ Textbook. side, V1Z. those from one upper and one lower row of leaves, at its outer margin, with the same arrangement, as in the first-mentioned series of cases. * Compare Cramer, /.c. p. 14, Taf. 30, 31. ” Nageli, Beitr. 7c. p. 53.—Hofmeister, Vergl. Unters. * A. Braun, Monatsbr. d. Berliner Academie, 27 Apr. 1865. COURSE OF THE BUNDLES IN THE STEM. 283 ~ Other species of the genus show a different arrangement of the bundles in the stem, such as ‘2 median, 3 median, 3 forming a triangle, or numerous scattered bundles’ S. ineequalifolia shows three median ones (Fig. 131). S. Lyallii has in its strong main shoots, which emerge above the ground, ten or twelve bundles distributed in the trans- verse section in three parallel equidistant rows in an almost quadratic surface; where the number is ten they are so arranged that three roundish ones form two opposite sides of the quadrangle, while one transversely extended one occupies each of the two other sides, and two other roundish ones lie in the middle of the quadrangle and form, with the two transversely extended ones, a row of four parallel to the rows of three. Other transverse sections of the same shoot show in place of one of the transversely extended bundles two contiguous round ones, which are doubtless products of its division. The course of the bundles and the insertion of the bundles of the leaves have not as yet been investigated in those shoots which have other than one axile bundle or two lateral ones. In 8. spinulosa, which has homomorphous leaves in many rows, there is a single axile bundle of roundish transverse section (and with a structure differing from that of other species): the leaf-bundles insert themselves on it on all sides. Finices anp HyproprEeriDE&”. Sect. 79. It has been already stated that there is always in the seedling of these plants one axile bundle composed of the single, acropetally developed bundles of the leaves. In many cases each lateral shoot begins with one such bundle. In a number of forms this structure is persistent in the mature stem. In the large majority the bundle extends itself, and forms a tube, which surrounds a paren- chymatous cylinder of pith, and is enclosed in a parenchymatous cortex. At the insertion of each leaf the tube has a gap, the folvar gap, from the margin of which the bundles start, which go into the leaf; ,at other points it is closed, or reticulately perforated. Of this semple bundle-tube—or ring of bundles as it appears in transverse section—several special forms may be distinguished. In relatively few cases there are in addition to the simple tube accessory, me- dullary, and cortical bundles, or there appear several concentric tubes or rings. a. Axile bundle and simple bundle-tube. Sect. 80. One axzle bundle, from which a branch goes to each leaf, traverses, as in submerged Phanerogams, the floating stem of Salvinia and Azolla. It occurs also in the rhizomes of Pilularia minuta, exceptionally also in weak plants of P. globulifera*, in the investigated stems of species of Hymenophyllum *, Gleichenia, 1 Compare A. Braun, /.c. 2 Mohl, Structura caudicis filicum arborearum, &c. in Martius, Ic. Plantar. crypt. Brasil. Tab. 29-36.—Verm. Schriften, p. 108.—Hofmeister, Beitr. zur Kenntniss d. Gefasskryptogamen, II.— Abhandl. d. K. Sachs Gesellsch. d. Wissenschaften, V. p. 602.—Stenzel, Ueber d. Bau u. d. Wachs- thum d. Farne, Nova Acta Acad. Leopold. Bd. 28.—Mettenius, Ueber den Bau von Angiopteris, Abhandl. d. K. Sachs Gesellsch. d. Wissensch. IX. p. 500.—Trécul, Ann. Sci. Nat. 5 ser. tom. X. p. 344, and tom. XII. p. 218. 3 Russow, Vergl. Unters. p. 13. * Mettenius, Hymenophyllez, /.c. (compare above, p. 126). 2,84 PRIMARY ARRANGEMENT OF TISSUES. Lygodium?, and also of Schizza, and the leafless stolons of Nephrolepis. The bundle has usually a circular transverse section, in Salvinia rotundifolia it is horse- shoe-shaped. Sect. 81. In many Ferns the original axile bundle widens out as the stem grows stronger into a tube, which is for the most part closed all round, and has only at each node, below the insertion of the leaf, a relatively small slit or /olzar gap, through which the medullary parenchyma is connected with the cortex, and from the margin of which one or several bundles pass into the leaf. To this series belong for the most part forms with a thin creeping rhizome, and leaves alternating in two rows: the investigated species of Marsilia, normal specimens of Pilularia globulifera? with a very small foliar gap, from the lower margin of which a foliar bundle arises. Most species of Dennstaedtia (D. tenera, scandens, davallioides, punctilobula) have a tube, which is closed as far as the foliar gap; the bundle which enters the leaf arises from the whole margin of the gap as a continuous concave plate, which is only occasion- ally split up at its base for a certain distance into several bundles lying side by side. ‘The same structure is found in all species of Microlepia and Hypolepis, in the species of Phegopteris, which are allied to the latter genus, and in the species of Pteris of the section P. vespertilio, aurita;’ further in Polypodium Wallichii, and conjugatum, to the bundle-tube of which attention was first drawn by R. Brown, and in which a bundle passes into the leaf from each side of the narrow slit-like foliar gap*. Of the Hymenophyllacez, Loxsoma has a closed tubular bundle 4, the foliar gaps of which have not been described. Of the Schizeeaceze® the species of Schizzea may perhaps be mentioned here: but for reasons which will be given below (Sect. 106) Russow has correctly placed them in our previous category: otherwise they have been as yet but little investigated. Among the Ophioglossacez the above described structure occurs in the rhizome of Botrychium Lunaria®. Hofmeister’? found in Ophioglossum vulgatum the network of bundles of the rhizome, which belongs to the next category, sometimes coalescent for a certain distance to form a closed tube. Sect. 82. Most Ferns with an ascending or upright rhizome or stem, with leaves in many rows, and but slightly elongated internodes, are distinguished fundamentally from the type just described in this point only, viz. that the foliar gaps are relatively large, and the bands of the bundle-tube, which separate them, are relatively narrow. The tube has accordingly the form of a Wet, the meshes of which are the foliar gaps. From the margins of the meshes branch the foliar bundles, which there run obliquely upwards through the cortex to the insertion of the leaf. ‘The bundles of the meshes of the stem are, according to the species, relatively narrow, of round or elliptical transverse section, or, as in stems of the Cyatheaceze, broad, or band-shaped plates, with their margins often curved outwards: the bundles which pass into the leaf show the same varieties of form: their number for each leaf is constant within narrow limits in the mature plant of each species, but varies in different species 1 Russow, Zc. ? Russow, 2c. * Mettenius, Angiopteris, p. 544. * Mettenius, Hymenophyllaceze, p. 418. : ° [Comp. Prantl, Morphologie d. Gefasskryptogamen, Heft 2, Leipzig, 1881.] Russow, /.c. p, 117, &c. 7 Beitr. III. p. 664. COURSE OF THE BUNDLES IN THE STEM. 285 from one to very high figures. Where several bundles pass out, they often anastomose with one another in a reticulate manner immediately after leaving the foliar gap: this is especially the case in the Cyatheacez. From the various combinations of these different relations result the most various individual forms of the net, and of the grouping of the bundles in the transverse section of the stem: a cylindrical pith is always surrounded by them. To this type belong numerous Polypodiacez, a number of the Cyatheacez, of the Schizzeaceze, Aneimia, of Ophioglossacez, Ophioglossum (O. vulgatum, O. pe- dunculosum). Peculiarities may be subsequently described in a few examples. The seedling of Aspidium Filix mas begins with leaves arranged with 3 divergence: their solitary bundles are united sympodially in the stem to one axile bundle. Above the 5th-6th leaf the stem increases greatly in thickness, the 4 arrangement passes over to 2, and from the point of outward curvature of the bundle of the highest leaf of the + arrangement the formation of the reticulate bundle-tube begins. Each leaf receives one bundle from the lower angle of the rhombic mesh or foliar gap, upon which its base is seated: or, in other words, two bundles run into each leaf, arising from the point of exit of those which pass into the two next lateral older leaves: these converge acutely towards their own point of exit, and are there united into a single bundle. By the repe- tition of this formation the network of rhombic meshes is built up. Where the arrange- ment is 2 there pass up to leaf 9 one bundle from 6 and one from 7, to leaf 7 one from 4 and one from 5, &c. Inthe second year the plant becomes much stronger, the leaf arrangement passes over to 4, which divergence is retained in the mature plant, or passes over into 28,7. Each leaf now receives several bundles from the margin of its foliar gap, at first five, in mature and strong stocks seven: one arising from the lower angle, and six from the sides of the mesh; of the latter two weaker ones on each side, belonging to the lower half of the mesh, and one stronger one belonging to the upper half. The structure of the meshes is the same with 3, arrangement as with 2; at the lower angle of each, where the median bundle passes into the leaf, two bundles which descend from the centre of the two next lateral older leaves come into contact—that from the one side following the parastichies composed of every third leaf, that from the other the parastichies composed of every fifth leaf (comp. Fig. 132). The transverse section of the mature stem thus cuts eight vascular bundles (where the arrangement is g4, 10-12 bundles), which form a circle round a wide pith: outside this in the cortex are seen the bundles which pass obliquely into the base of the leaf, arranged in different number and order according to the position of the section. The vascular bundles of the stem are weak compared to the mass of the parenchyma, in transverse section they are roundish or flattened externally and internally (Fig. 133). According to the numerous investigations of Hofmeister, Stenzel, and Mettenius, fundamentally the same structure—even the narrow, reticulate bundles, which are weak in arborescent stems—is found in Onoclea, Struthiopteris, in all investigated species of Blechnum (incl. Lomaria), Woodwardia, Asplenium, Phegopteris, species of Aspidium with a stem having leaves in more than two rows, in Ophioglossum, and Aneimia. The individual peculiarities depend partly upon the form of the meshes corresponding to the elongation of the internodes—thus very elongated meshes are found in the runner-like branches of the rhizome of Struthiopteris, Aspidium cristatum, in the creeping stems of A. Thelypteris, quite short and broad ones in Aspl. Filix feemina—partly upon the number and arrangement of the leaf-bundles which arise from the margin of a gap. Most 1 A. Braun, Schuppen d. Tannenzapfen, Nova Acta Leop. vol. XV. p. 278. % Hofmeister, Beitrige II.—Stenzel, 7.c. 286 PRIMARY ARRANGEMENT OF TISSUES. of the investigated species of Aspidium have according to Stenzel’s account three or five bundles for each leaf, one median, one from the lower angle, the others arising in pairs from the sides of the mesh: in Aspidium Thelypteris the median one is absent, according to Stenzel (Tab. V. 18), one bundle passing into the Jeaf from each side at the middle of the very elongated mesh. Blechnum Spicant (Stenzel, Tab. II. 5) has two lateral ones, one arising on each side, close to the lower angle of the mesh, BI. brasiliense however has seven, one median and three pairs of lateral ones, arising from the lower half of the mesh. In Asplenium Filix foemina, the mature rhizome of Struthiopteris?, in Aneimia, and Ophio- glossum the same arrangement is found, which appears only in the young plant of Filix mas, viz. each leaf receives only one median bundle from the lower angle of its gap. It is remarkable, according to Stenzel’s account (/.c. Tab. II. 3), that in the scale leaves of the elongated runners of Struthiopteris the median bundle is absent, and in its place a bundle runs on each side from the middle of the long mesh, Among the Cyatheacez, Dicksonia (Balantium) antarctica, Karsteniana, Cibotium Schiedei, glaucescens, Plagiogyria biserrata, Alsophilia pruinata, blechnoides?, have—in contradistinction to their nearest allies to be described below—the same structure as is now under discussion. The appearance of the transverse section of most of these plants, which differs so strikingly from that of the Polypodiacez, depends partly upon the form FIG. 132.—Aspidium Filix mas; natural size. F slightly magnified ; D end of the stem, FIG. 133-—Aspidium Filix mas; . the leaves of which are cut off, excepting the highest ones; 4 transverse section of a transverse section through a strong petiole; w roots; Za similar end of a stem, the network of bundles exposed by paring stem, with 8/21 leaf-arrangement; off the cortex (g); # mesh of the net, with insertions of the foliar bundles; C base of the natural size. petiole, with a lateral bud %, longitudinal section; w root. From Sachs’ Textbook. of the vascular bundles in the stem itself, these being broad plates, usually curved out- wards at the margin, with narrow foliar gaps; partly upon the massive dark brown bands of sclerenchyma surrounding these bundles; partly upon the large number of thin foliar bundles, or the presence of one or a few broad channel-shaped ones; finally upon the very oblique ascent of the foliar bundles through the cortex, and the frequent anasto- moses in this part of their course between those belonging to one leaf (comp. Fig. 141, p. 293). If the internodes are short the transverse section shows a circle surrounding the pith, composed of bundles elongated in the direction of the periphery, or curved out- wards in a horseshoe-like manner, and with brown sheaths: between these are medullary rays of unequal breadth, according as the section has passed through foliar gaps at vary- ing height; outside the ring of bundles are those bundles which are passing to the leaves; where the internodes are elongated and the foliar gaps are of relatively small size, the transverse section may show one closed annular bundle, which is only broken here and 1 Hofmeister, /.¢. ? Mettenius, Angiopteris, p. 524. COURSE OF THE BUNDLES IN THE STEM, 287 there (by a foliar gap), and opposite these points are foliar bundles in the cortex, such as Karsten? illustrated in the case of Alsophila pruinata. Sect. 83. If the structure just described for stems with leaves in several rows be imagined to be transferred to horizontally growing stems with leaves alternating in two rows, there are thus obtained foliar gaps.in two alternating rows right and left, Bocas * eS FIG, 134.—Davallia dissec- ta; rhizome; slightly mag- nified. 4 vascular bundle- system, the cylindrical sur- face reduced to a single plane; o upper bundle, 2 lower bundle, & point of in- sertion of a leaf, at < point of origin of a lateral shoot ; # transverse section, After Mettenius. FiG. 135.—Aspidium coriaceum ; rhizome; slightly magnified ; after Mettenius. 4 bundle-system, the cylindrical surface reduced to a single plane; meaning of the let- ters as in Fig. 134. SB transverse section, FIG. 136.—Polypodium fraxinifolium ; rhizome ; slightly magnified ; after Mettenius. 4 bundle- system, the cylindrical surface reduced to a single plane ; o upper bundle, J insertion ofa leaf, X points of origin of the lateral shoots. #8 trans- verse section. limited by one bundle with a median upper course, and by one median one below, and by alternating transverse bundles between the two. This arrangement is found in many forms as described, or with unimportant modifications. Comp. Figs. 134, 135. é * Vegetationsorg. d. Palmen, Taf. IX. fig. 1. ? Mettenius, Angiopteris, p.544. Special deviations and irregularities described by Trécul, 7. c, 288 PRIMARY ARRANGEMENT OF TISSUES. Their creeping rhizome shows a circle of bundles in transverse section. Of these, one passing along the middle of the upper side (the upper bundle, o Figs. 134, 135) and a second passing similarly along the under side (the lower bundle, u/) are distinguished by their band-like form and greater size from the other weak ones, which are opposite to the two rows of leaves. Both the stronger bundles are connected regularly, at distances corresponding to those separating the leaves, by transverse bundles curved convexly upwards, or bent at an angle, so as to form a net, the meshes of which are the foliar gaps. From the margins of these arise (in addition to the bundles for the lateral shoots x) the foliar bundles (4) which converge opposite the point of insertion of the leaf, which is usually relatively small, and run almost radially perpendicular in the stem up to that point: these bundles may anastomose one with another, and with the upper and lower bundles by solitary thin transverse connections. ‘The transverse sections of the foliar bundles are the small bundles seen in the transverse section of the stem; they there form together with the upper and lower bundles either a circle, as above stated, or in flattened stems an elliptical figure, which is often compressed in such a way that the upper and lower bundles have a central position, while the foliar bundles are outside. The arrangement described occurs in a simple form, with specific modifications as regards the number of the foliar bundles, the form of the gaps, strength of the bundles, &c., in Asplenium obtusifolium, A. resectum, Acrostichum brevipes, A. Lingua, A. simplex, A. Melanopus, Polypodium altescandens, P. tenellum, Nephro- lepis ramosa, Aspidium albopunctatum, A. coriaceum (Fig. 135). In the Davallias there arises the further complication, that the branches springing from the margin of the foliar gap do not run directly or with unimportant anastomoses to the leaf, but first form a network of fine bundles, which stretches over the foliar gap, and sends a certain number of branches into the leaf. According to the number of these foliar bundles (which varies according to the species) the network is simpler (D. parvula, pedata, heterophylla), or more complicated, and with more numerous meshes (D. bullata, dissecta (Fig. 134), elegans, pyxidata, canariensis, &c.). A more considerable deviation from the structure described appears in other creeping stems of Ferns with leaves in two rows: here, not only is the foliar gap covered in by a network of bundles, but also instead of the lower bundle two or more reticulately anastomosing bundles are present, and the lower bundle is as it were split up into a network of bundles (Fig. 136). Where the number and arrange- ment of the bundles are very simple, as e.g. in Polypodium aurisetum, piloselloides, cayennense, or where the upper and lower bundles and their transverse connections bordering on the foliar gaps are strongly distinguished from the rest by their size, as in Platycerium alcicorne, the structure may be referred easily to the scheme with ‘upper and lower bundles. But often the upper and lower bundles and all the anas- tomoses are of such uniform strength, and the meshes of different order so irregular, that the foliar gaps can only be distinguished at both sides of the upper bundle, where the bundles pass out into the leaves. In place of the tube regularly per- forated by foliar gaps there is in extreme cases as it were a complex irregular net- work, the relations of which to the more simple type can only be recognised as indicated by the regularly alternating ‘foliar meshes’ 4. As extreme examples may be named, Polypodium vulgare, sporadocarpum, aureum ; numerous species of COURSE OF THE BUNDLES IN THE STEM. 289 Polypodium and Acrostichum axillare.show numerous intermediate forms between these and the simple scheme with upper and lower bundles. For further details cf. Mettenius, Angiopteris, p. 552, &c., Taf. VII-X. &. Several concentric rings of bundles. Sxct. 84. A number of Fern-stems with leaves in many rows—species of Pteris, and Saccoloma, Marattiaceze, Ceratopteris—show in the transverse section of the stem several concentric rings of bundles, similar to one another in form and thick- ness. As far as is known, these cases are connected with those forms above described, in which the bundles running from the ring to a leaf pass gradually, and for a long distance obliquely upwards through the cortex, and are connected by anastomoses both one with another, and also with those belonging to neighbouring leaves. The middle of the stem is traversed by one axile bundle, or in most cases by a relatively narrow tube of bundles, surrounding a narrow pith. From these inmost bundles there arise at regular distances, and in close relation to the arrangement of the leaves, flattened or narrow bundles, which branch out at once into broad reticulate layers: these do not pass out directly into a leaf close to their point of origin, but run upwards and towards the surface of the stem through a number of internodes, and finally pass out into leaves, or divide into branches, which pass out in succession. Each of these layers of bundles has the form of a portion of the surface of a cone, which widens upwards: each is surrounded by a layer of similar form (and a zone of parenchymatous cortex separating it from the latter), and arises at its lower end from the inmost bundles. At the points of insertion of the leaves there are anastomoses between the successive zones, i.e. between those which are passing out, and the next inner ones, which run further. The rings which appear in the transverse section are the transverse sections of the conical zones: their number in any given transverse section depends upon their special course, particularly upon the inclination of the conical surfaces, which is closely connected with the elongation of the internodes. The siniplest case occurs in Pteris elata var. Karsteniana, and P. podophylla Sw.'; also in P. Orizabe, and P. gigantea®, In the two first-named species there is, aceording to Mettenius, a second narrow bundle-tube within an outer one, the former being split sometimes on one, sometimes on two sides. Portions of the latter curve directly out- wards into the leaves; portions of the first, turning outwards, enter the gaps formed by the passing out of a portion of the outer ring, and coalesce with the outer tube, which rises laterally from the base of the leaf. In Saccoloma inaquale there is a similar arrangement (Mettenius). Saccoloma adian- toides*® shows in transverse section (Fig. 137) at least three Closed or split rings repre- senting so many conical zones, of which the outermost alone gives off broad and flattened concave portions into the leavgs (closely arranged with g& divergence) ; the middle one, curving outwards, enters the gaps thus formed; finally thes inner one fills up in the same way the gaps made in the middle one by this ‘outward curvature, 1 Mettenius, Angiopteris, p. 535, Taf. VI-.12-16. 2 Karsten, Vegetationsorg. d. Palmen, /.c. p.193. 3 Mettenius, /.c. p. 531, Taf. VI—Karsten, 7c. p. 194 (Dicksonia Lindeni). _ U 299 PRIMARY ARRANGEMENT OF TISSUES. The inmost of these zones differs in individual cases: in some, investigated by Mette~" nius, it is composed of two flat small bundles, which appear curved in transverse section ; in others, on which Karsten’s statements appear to be based, there is a solid cylindrical bundle. Karsten says of this that it ends freely in the pith below: this does not coincide with the data given above, chiefly after Mettenius, and requires further examination. Of the Marattiacee with several rings of bundles as seen in transverse section, the cylindrical stem of the species of Danza with on an average three rings in ‘transverse section, may according to the notes of Mettenius’ belong to this category. Of the three rings the outer con- sists of numerous filamentous bundles, the two others of broad and flat ones. These stems how- ever require more exact investigation. The same may be said of the inverted conical-shaped thick tuberous stem of Angiopteris evecta®, of which Mettenius has thoroughly’ investigated and de- scribed one large specimen only. Transverse sections of the stem show 5-6 irregular zones or rings which ‘pass over one into another. From an irregular network of’ bundles surrounding the F1G, 137.—Saccoloma adiantoides; transverse sec- pith bundles arise in a succession: correspond, ton through the stems after Mettenius; natural ee. ing to the arrangement of the. leaves, and pass separated from the outer ring; 4 and ¢ bundles of two obliquely upwards and outwards; each of these successively higher leaves, appearing as protrusions of a z ess the outermost ring. The bundles which enter the leaves soon becomes wider, and splits into a rather ir- sre finely ondslated, The bundles ofall ngs a= SU” yegular network rising obliquely outwards and up- chyma. In my specimens the inmost ring of bundles is wards, and having the form of a portion of a conical sometimes narrow, sometimes it is replaced by a single ‘ 5 round bundle. surface, the bundles of which are reticulately con- nected with those of neighbouring equivalent por- tions. A number of branches, derived from that zone which is at the time the outer- most, enter the base of each leaf (into the back and sides of its basal part), and, to take the place of those which have passed out, a corresponding portion of the next inner zone passes upwards into the outward zone from below the axils of the leaf in question, and. of the two next lower lateral leaves. Portions of the third zone enter the gaps thus formed in the second, and are reticulately connected with the next outer zone. Further anastomoses between the branches of the successive zones are formed at the very point of insertion of the leaf, and two bundles belonging to the inner side of the. base of the leaf are given off from the second zone, which enters the gaps of the outer zone. In the stems investigated the lower zones had small, almost cylindrical bundles, which formed wide-meshed irregular nets; the upper zones (near to the stationary end of the dead stem) had broad flattened bundles with narrow reticular meshes ; the zones of intermediate position were also. intermediate as regards the reticular form ; the bundles which enter the leaves were of similar form to those of the zones from which they arose. Accordingly transverse sections showed at different heights either several concentric annular zones, often irregularly connected by oblique bands (the portions cut through in their course into outer zones), consisting of small roundish bundles separated by abundant parenchyma—corresponding to the usual arrangement in the tuberous stems of Marattiacee; or on the other hand rings, of which the outer at least consist of broad flattened pieces, separated by sonte few bands of parenchyma. For further peculiarities, comp. Mettenius, /.c. 1 J.c. p. §24.—Compare also Brongniart, Archives du Muséum d’ Hist. Nat. tom. I. p. 439, Tab. XXXII, and Karsten, Vegetationsorg. d. Helnet, Taf. IX, fig. ro. 5 ? Brongniart, Ac.” COURSE OF THE BUNDLES IN THE STEM. agt The investigation of a young stem of Angiopteris showed me a completely typical bundle- tube, interrupted by wide foliar gaps; two strong foliar bundles arise below at the lateral margins of the gap, and ascend obliquely through the cortex, within which they divide into the branches which pass out into the leaf. The concentric zones of thin bundles in the stem of Ceratopteris thalictroides may also belong to this category, but require further investigation. Comp. Mettenius, /. c. Pp. 530. State c. Accessory medullary and cortical bundles in addition to the simple tube of bundles. Scr. 85. Among the Cyatheacee, according to Mettenius, the above-mentioned forms (p. 286) have only the typical tube, perforated by foliar gaps, and consisting of flattened vascular bundles, the margins of which next the gaps are curved out- wards. Other species, including the majority in the genera Cyathea and Alsophila, have, in addition to the vascular bundles thus disposed, small bundles, which originate from the foliar gaps and traverse pith and cortex, there forming a delicate open network. The relatively thin bundles from the margin of the foliar gap, which pass into the petiole, arrange themselves so that in a transverse section of its insertion, or in the leaf-scar, they are arranged in a curve, convex downwards, and simple, or with the ends turned inwards above: it consists of few bundles in small leaves, e. g. in young specimens of Hemitelia capensis 13-14 bundles, of Alsophila radens KIf. 4, Cyathea arborea Sm. 131; on the other hand, in strong specimens or species they form two curves with the ends curved inwards, and consisting of many bundles, one of these curves being convex downwards, corresponding to the lower edge of the foliar gap and spring- ing from it, the other convex upwards, and corresponding to the upper margin. The in- curved ends of both curves are directed down- wards and towards the middle of the leaf-scar F1G. 138.—Cyathea Imrayana; natural size, Two 5 old leaf-scars from a dead stem; @ lower, 4 higher on so that their bundles form on each side two thestem; a with four, 4 with two medullary bundles nearly parallel series running to the middle of one the scar*. Compare Fig. 138. Further, in the space surrounded by the single curve or by the upper one, a relatively small number of bundles pass out into the petiole—e.g. as described by Mettenius in Hemitelia capensis, Alsophila radens, and Cyathea arborea two each, in a species of Cyathea 7, in Alsophila Haenkei 4, in Cyathea Imrayana 2 or 4, in C. ebenina 2. These do not arise from the margin of the foliar gap, but are connected both inside and outside it with the margin itself, as well as one with another, by numerous strong anastomoses (usually sheathed with sclerenchyma): they then run through the foliar gap downwards into the pith. (Figs. 139, 140, m.) Immediately after their entry they pass with a steep curve inwards and down- wards, and divide into branches diverging acutely downwards; these sometimes 1 Mettenius, Angiopteris, /.c. Taf. V. : ; ae 2 Mohl, in Martius, Icones, Z.c., Verm. Schriften, p. 110,—Numerous valuable details in Trécul, ie. XIL. p. 270. U 2 292 PRIMARY ARRANGEMENT OF’ TISSUES, pass on further in the middle, sometimes at the periphery of the pith, and some insert themselves at an acute angle on similar branches from lower leaves, others end blind. .At the foliar gap the bundles, which are themselves about as thick as a bristle, are surrounded by brown sclerenchyma, or supported by it on one side, and these sheaths of sclerenchyma, which are closed or open on one side, accompany the bundles for a long distance downwards; sometimes they also anastomose with F1G. 139.—Cyathea Imrayana; natural size. From a dead stem, the soft parts of which had rotted away; - the stem is halved longitudinally, and seen from within, The figure ‘shows a piece of the tube of vascu- lar bundles with one foliar gap, 2—?, and a leaf-scar exactly halved longi- tadinally. From this the bundles of sclerenchyma which accompany the “vascular bundles may be seen to branch off, and from there those which accompany the medullary bundles arise as branches and pass down into the pith with many branchings and anastomoses; some of them’have blind pointed endings ; the one which extends furthest downwards has anastomosed with _ one coming from alowerleaf. Ofthe bundles which end at the leaf-scar, me originates from the pith, the rest from the mafgin of the foliar gap. ~ FIG. 140.—Cyathea Imrayana; axile longitudinal section through the same stem as Figs. 141 and 142; natural size. The section is about « 3mm. thick, and for the most part transparent; the S/ack bands of ‘ sclerenchyma and #a/e vascular bundles here represented in one plane do not all Jie exactly in this plane. but are‘all near it. Certain parts of the chief sclerenchymatous-sheath s—s’ which show through, and at the bottom two portions of the surface of the stem seen obliquely, are shaded off as they pass from the surface of section. The letters a, 5, s’, have the same meaning as in Fig. 141; 7 cortical bundles, 0 leaf- scars, 62 vascular bundles passing out into leaves, zw insertions of roats, m a foliar buridlé running info"the pith; above x blind ending of a dullary bundle ( ined under the mi ) = similar sheaths, which descend from leaves side by side with or below them, some- times they diminish downwards and end blindly in a point, while the vascular bundles continue their downward course alone beyond these endings (comp. Fig. 14°): _ COURSE OF THE BUNDLES IN THE STEM. 293 According as they follow the first or the second course, the sclerenchymatous sheaths themselves form either a tough net traversing the whole pith, as e.g. in the case of a stem which is before me under the name of Cyathea ebenina, or they pass from each leaf inwards and downwards as a sheaf of bundles, with blind and pointed ends, which show frequent anastomoses one with another, but only fewer or quite solitary anastomoses with the sheaves of bundles belonging to other leaves: e.g. Cyathea arborea, Hemitelia capensis (Mettenius), C. Imrayana, and most other species; in Alsophila miicrophylla and villosa the vascular bundles in the pith are only accompanied by isolated spindle-shaped bundles of sclerenchyma, which are not connected into sheaths till the foliar gap is reached (Mettenius). In most of the dried stems, when subjected to investi- gation, the delicate unsheathed parts of the vascular bundles cannot usually be seen, the tough sclerenchymatous bands alone being clearly preserved. Since, as above stated, the course of these bands is a copy of that of the vascular bundles, the description given for the latter will suffice also for them. Many but not all species have, besides the medullary bundles, acces- sory cortical bundles also. In C. Im- rayana (Fig. 142) these arise from FIG. 141.—Cyathea Imrayana: transverse section through the living stem ; natural size; seen from above. At J, c,d, foliar gaps. All guzte ddack bands and points are transverse sections of sclerenchyma; all the paler ones are vascular bundles, In and near the foliar gaps, especially 2 and 4, are root-bundles going to the periphery ;_/channels at the base of the leaf; @ vascular bundles of the main tube; s outer, s’ inner plates of its sclerenchymatous sheath. Inwards | from s’ the pith with its bundles; outwards from s the cortex with its bundles. bundles which ‘pass into the leaves, and close above their point of departure from the foliar gap, and in fact from most, but not from all those which branch off from the lateral and lower margin of the gap. They descend with a steep curve from their point of origin into the parenchyma of the cortex; some of them, after pursuing an individual course for a short distance, unite each with another coming from the same foliar gap; most of them however pass almost straight downwards in the middle of the cortex, and in the neighbourhood of the adjoining lower foliar gaps they either affix themselves at an acute angle on bundles which there arise, or they end blindly. The cortex is accordingly tra- versed by a network of bundles with elongated meshes, which are sometimes quite closed, sometimes open on one side. In the stems in question the cortical bundles, of the thickness of a bristle, usually have no sclerenchymatous sheath; some few, especially those which arise from the upper part of the lateral margin of the gap, are however often accompanied for: about 1°™ from their point of origin by such a sheath, in the form of a channel, which is open outwards; comp. Fig. 142. In a dry strong stem, bearing the name C. Imrayana (not the same used for the accom- panying figures), a structure, similar to that just described can still be clearly seen. 294 PRIMARY ARRANGEMENT OF TISSUES, But in this case there is this further peculiarity, that from the upper part of each lateral margin of a foliar gap two or several bundles arise, which unite after a short course into one, and that these bundles from their point of origin onwards are sur- rounded by a thick sheath of sclerenchyma. Together with the latter they form on each side of the foliar gap a cone several millimetres thick at the point of origin ; they either taper to a point towards the adjoining lower gaps, and end blindly in the parenchyma, or coalesce with the margin of the adjoining lower gaps, and with a cone which there arises. The sclerenchymatous sheath of the cone shows slits here and there, through which unsheathed branch bundles emerge and turn down- wards. Cones, fundamentally similar to those described, the points of which end blindly in the cortex, and indi- cate with certainty the presence of a system of cortical bundles similar to that in C. Imrayana, were first found by Mettenius‘ in a dry stem of Also- phila Haenkei. In other species a system of cortical bundles is un-. known, partly because of the difficulty. of finding it in dry stems subjected. to investigation; in many species however (for instance Cyathea arborea and Alsophila microphylla may be named with certainty) it is altogether. absent. Secr. 86. Most species of Denn- F steditia have, as above indicated, a FIG. 142.—Cyathea Imrayana; piece of a living stem with simple tube of bundles which is closed four bases of petioles, the outer layers of cortex being peeled bundles which avse from tiem and pass into the leaves, wien With the exception of the narrow foliar which descend through the cortes, are exposed the later S@PS- Within this, and near the upper Patt pein treet Mier ner ecm, side in the horizontal stem, there are ane parts are held together in their natural position, Natural in the pith of D. rubiginosa one, in D. cornuta several small bundles with circular transverse section, in D. cornuta these form a tube alternately closed and again split into 2-3 bundles. At the base of a shoot the medullary bundles arise from the inner surface of the tube, at the foliar gap they approach the latter, and divide into a few branches, of which some anastomose with the margin of the gap, others enter the leaf with those which start from the gap, the third series (or single bundle) ascending further in the shoot as medullary bundles ®. The somewhat more complicated arrangement in Chrysodium vulgare, on which compare Mettenius, 4c, may be placed in connection with the above. ' Angiopteris, p. 528, Taf. V. A good figure of a transverse section of a Fern stem with cortical bundles is given by C. H. Schultz, Mém. présent. de l’Acad. des Sciences, tom. VII (1841), pl. 22. ? Mettenius, Angiopteris, p. 540. COURSE OF THE BUNDLES IN THE STEM. 295 Sect. 87. While in the last-named cases accessory medullary bundles occur, and in many Cyatheacez accessory medullary and cortical bundles in addition to a typical tube of bundles, there is found in Pér7s aguilina and Polybotrya Meyeriana a bundle-tube constructed as in the type with a strong upper bundle, and this is strengthened by a much divided cortical system of bundles. In the seedling of Pteris aquilina?, till the development of the seventh or ninth leaf, there is one axile bundle, which traverses the stem starting from the point of union of the first leaf with the first root: in transverse section it is deeply grooved, and half-moon-shaped: bundles pass from it into the leaves; after the formation of the seventh to ninth leaf ‘the stem forks.’ Both branches of the fork increase rapidly and considerably in thickness, and the course of the vascular bundles in them is altered. The lateral +: opening of the axile bundle is widened, then the | upper half is separated from the lower ; there are now two bundles, an upper and a lower one, which split now and again into thinner branches, while these are soon again united. When the branches of the fork have attained the length of about 6m, and a thickness of about 4mm, weaker branches come off from both the bundles: these run near the surface (in the cortex), and here form a peri- pheral network with long and narrow meshes, in which the upper central bundle is distinguished Sa nasser Gee a 3 from the rest by its greater width. This structure is retained by the mature rhizome (Fig. 143): the number of the peripheral bundles rises to twelve in the transverse section. Two thick brown plates of sclerenchymatous fibres lie between the inner and outer systems of bundles, and are only separated from one another at the two sides of the stem by FIG. 143.—Pteris aquilina; stem (rhizome) ; slightly enlarged. 4 transverse section, 7 brown sclerenchymatous layer of outer cortex, ? colourless soft parenchyma, 7 vascular _ bundles of the inner zone, ag the broad upper bundle of the outer zone, #7 the plates of brown sclerenchyma separating the two zones. B the upper main bundle of the outer zone of the stem (s¢) and its branches (s¢’, s¢’”), with the branch bundles which pass into a leaf (4) prepared free; —z outline of the stem. a narrow slit filled with parenchyma: they are often joined at one side, often even all round so as to: form a closed tube. Branch- bundles from both nets of bundles pass into the leaves and branches: roots arise from the outer ones only. At those points of exit, as also at the base of the petiole, the two nets anastomose by single transverse bundles. Throughout the whole of the rest of their course they are without connection with one another in many weak specimens ; in strong rhizomes, according to Stenzel, thin connecting bundles pass from the margins of the inner bundles to the lateral outer bundles, while the upper and lower bundles of both systems are connected by single short branches, which pass through holes in the bands of sclerenchyma. ' “In the main axis of Polybotrya Meyeriana? an inner network is found sur- rounding a narrow pith, and composed of 3-7 strong bundles arranged in a circle when seen in transverse section; it corresponds in composition to that above 1 Hofmeister, /.c. p. 620.—Mettenius, Angiopteris, p. 561.—Stenzel, /,c, 2 Mettenius, /.c. p. 559, Taf. VIL. 296 PRIMARY ARRANGEMENT OF TISSUES. described for the Polypodiacez with a reticulately divided lower bundle, having long narrow, rather irregular meshes: foliar meshes and points of origin of lateral shcots appear only in regular alternation, on both sides of a clearly distinguishable upper bundle. Outside this net is a peripheral one, showing in transverse section fifteen to about fifty bundles, which rarely form a single circle, but are usually arranged at the upper side of the stem in 2-3 irregular series, at the lower side of the stem ina curve. The bundles of the peripheral net are thin and connected in elongated meshes, in the arrangement of which there was as little regularity to be recognised as in that of the obliquely ascending connecting-bundles, by which the outer net is joined with the inner at many points. Three branches arising from a single mesh of the inner net and g—12 peripheral ones enter each lateral shoot; into the base of the leaf there pass two or several inner, and. 9-24 outer ones. All roots arise from the outer net, c. Bundle system tn the leaves and foliar expansions. Sect. 88. The bundles, which leave the bundle-cylinder of the stem and pass out into the leaves, run as a rule towards the margin and apex of the leaf. The bundles may run from their point of departure from the cylinder of the stem onwards, just as they did in the latter, as for instance in a leaf of a Conifer, where one bundle passes out from the cylinder and runs without division up to the apex of the leaf; or divisions. of the bundles and coalescences of separate ones may occur at any point, so that the number to be seen in the transverse section of the leaf is not the same as at the point of exit itself. Examples of this, when it takes place in the node itself, have been given above, Sect. 61; in the case of their further course the pheno- menon is universally known, and will be treated in detail in the following paragraphs. In most cases the bundles at the odes pass, without division, or after splitting into branches which run side by side, through the cortex into the leaf. But in certain individual cases there appear special branches at the node itself, which lie in the parenchyma of the cortex; these are here connected into a net or a transverse girdle; often branch bundles pass downwards from the node into the cortex of the adjoining internode. Branches peculiar to the node (which do not pass out from it) are often found where several bundles emerge, forming oblique or curved connections between these. This is the case both in many traces with numerous bundles, belonging to solitary (alternating) leaves, e.g. Lathyrus Aphaca (comp. p. 240), where the median bundle has a curved transverse connection with the lateral ones; Viola elatior, Platanus, which will be mentioned below: also in whorls of two or more with leaves having one or more bundles. In the case of leaves with one bundle Hanstein? found a transverse girdle connecting the bundles at the node in numerous Rubiacez with whorls of two or more members (species of Asperula, Rubia, Galium, Hamelia chrysantha, Houstonia coccinea, Bouvardia mollis); on the other hand, other Ru-. biaceze (Coprosma ligustrina, Exostemma floribundum according to Hanstein) show no transverse girdle. In whorled leaves with several bundles transverse girdles have ? Ueber giirtelformige Gefassstrangverbindungen, Abhandl. d. Berliner Academie, 1857, p. 77. BUNDLE-SYSTEM IN THE LEAVES. 297 been above described (pp. 257-259) in Calycanthus and Melastomacez. Hanstein found them in Sambucus, in species of Valeriana, Centranthus, Valerianella; Scabiosa; Knautia, Succisa, Dipsaeus ; Dahlia, Bidens cernua, and tripartita ; Guizotia oleifera ; and Nageli in Humulus. In most plants with opposite leaves, e.g. the Labiate, Asclepiadeze, Caryophyllaceze, Caprifoliaceze excepting Sambucus, many Composite and others enumerated by Hanstein, the transverse girdles do not occur. Branch bundles passing down through the cortex may here be mentioned, andl have already been described above for many Cyatheacee. Also in the strict sense the medullary bundles of these latter Ferns belong to this category, as being ‘ appen- dices’ which run downwards from the bases of the leaves. Among the Monocoty- ledons no examples belonging strictly to this series are known, still the bundles running through the cortex, as described above for Bromeliacez and Palms, correspond in their course to some extent to those under discussion. The same may be said of the above-described cortical bundles of Melastomaceze and Calycanthus. Among the Dicotyledons, however, branches passing downwards from the node through the cortex occur elsewhere ; in the first place in the foliaceous corners of the so-called winged stems, e.g. in species of Lathyrus (L. silvestris, latifolius, Nissolia, &c.), secondly, and in the most striking manner, in many (but by no means all) succulent plants: Salicornia, species of Mesembryanthemum, Cactee. The course and ramification of the bundles in these cases and in the winged corners closely resemble that to be described in the leafy expansions, the cortex here assumes completely the anatomical (and physiological) properties of foliar expansions. In the species of Salicornia! the short scaly leaves are arranged in decussate pairs. One bundle, which splits up immediately at its point of exit into three branches, passes from the node into each leaf: of these a single median one runs to the apex of the leaf-scale, and a lateral one on either side passes perpendicularly downwards into the cortex. ‘These branches, of which there are six for.each pair of leaves, give off throughout their whole course numerous ramifications, which anastomose frequently in a reticulate manner. From the apices of the leaves downwards the cortex of the whole internode, which attains a length of 2%, is traversed by a tubular network of bundles, which is closed (not interrupted as stated by Duval), and which ceases immediately above the next lower node, without any connection with that which there arises. In many species of Mesembryanthemum, e. g. M. imbricatum, M. crystallinum, &c., but by no means in all species of the genus, thin branch bundles, divided into reticulately anastomosing branches, pass downwards from the node into the cortex; here again they do not reach the next lower node. In Cactex, Epiphyllum truncatum, species of Cereus, Mamillaria, the above-mentioned (p. 261) Rhipsalidacee*, &c., reticu- lately connected branches come off from the main bundles, which run to the apex of the rudimentary leaves, and pass through the cortex, forming a continuously anastomosing network between the neighbouring main bundles. The bundles, which enter s#pules, and other appendages of the base of the leaf which often have a glandular surface, usually arise as branches from those which enter the main leaf (e. g. Prunus, Passiflora, Tropseolum, Medicago, Liriodendron *, 1 Duval-Jouve, Z.c. (see above, p. 226). * Vochting, Zc. 8 Nageli, Beitr. Zc. 298 PRIMARY ARRANGEMENT OF TISSUES. Coprosma ligustrina, Exostemma floribundum ‘, Quercus’); from the anastomoses at the node; or from the transverse curves, e. g. in those Rubiacez which have them, as is seen in an especially striking way in the large leafy stipules of the Stellate, and in Sambucus Ebulus (Fig. 44), &c. In less common cases special lateral bundles of the composite leaf-trace, which traverses the stem, pass into the stipules*. In Viola elatior one lateral bundle of the three forming the leaf-trace passes into each stipule, and in addition branches from curved transverse anastomoses, which appear in the node between the bundles of one trace. In Platanus occidentalis the < sheathing stipule has 7-9 bundles at its ; base; of these the two stronger lateral ones atise from the outermost lateral bundles of the leaf-trace (which has 7-9 bundles), the rest unite to two, sometimes to one or i three bundles, which separately enter the | bundle-ring of the stem. In Humulus i Lupulus’, of the three bundles of each t leaf-trace which enter the stem, the median j UTM one passes into the petiole, and each lateral Hie) dep-—aGbecs Shahi) alter Hagaiein. Sohne wk OME: THO’ stipule as its middle nerve. the bundle system in two successive internodes ; the cylindrical surface reduced to a single plane. Teaves in decussate pairs ; Each lateral one 1s connected at the node each leaf receives one median bundle. %, and on either side two lateral ones, s’, s//; of these the inner stronger one, s’, by a transverse bundle on the one hand descends undivided into the internode; the outer one, s/’, divides in the node into two shanks; one inner one which with its median bundle, on the other with pursues a solitary course downwards, and an outer «which coalesces at once with the sitnilar one of the opposite leaf (z). the lateral one of another (opposite) leaf ; By this union in the node the transverse girdle is formed, from which the bundles 1 pags into the stipules, The further course from the transverse girdle thus formed the ofthe leatrace ofeach pur, whit consist of twelve Dundes; Jateral nerves of the stipules are given off . in regular succession, so that the girdle itself is compounded in a sympodial manner of outwardly curving bundles. Sect. 89. The bundles run as a rule straight through the fe/iole towards the lamina. When they are numerous they are here also branched and connected by anastomoses. The bundles, when more than one, are arranged on the transverse section either in a curve open upwards, or in a ring, or distributed over the whole surface of- transverse section. Large leaves, e.g. of Leguminosae, Umbelliferz, Palme, Aroides, Cycadex, Ferns, &c., yield various special examples of these relations, which have been often made use of in the Ferns for systematic purposes *. . Sect. go. In the lamina of leaves, whatever their form, in the peripheral leaf- members of any stem, and in the leaf-like branches (Phyllodes, Phyllocladia, &c.), x s’n|™ s * Hanstein, /.¢. 2 A. B, Frank, Botan. Zeitg. 1864, p. 378. 3 Nageli, /.c. pp. 75, 92,114. : * Compare e.g. Grew, Anatomy, Tab. 49.—Presl, Gefassbiindelvertheilung im Stipes der Farne, Abhandl. d. k. Bohm. Ges. d. Wissensch. 5. Folge, Bd. V.—Reichardt, in Denkschr. d. Wiener Academie, Bd. XVII. - Duval-Jouve, Et. sur le pétiole des Fougéres, Hagenau, 1856.—Trécul, Ann. Sci. Nat. 5 sér. X and XII, and the descriptive literature of Ferns.—Reichardt, Ueber d. centr. Gefassbiindelsystem einiger Umbelliferen, Wiener Acad. Sitzungsberichte, 1856.—A. B, Frank, Botan. Zeitg. 1864, p. 380, [De Candolle, Anatomie comparée des Feuilles, Geneva, 1879.] BUNDLE-SYSTEM IN THE LEAVES. 299 parts which will all be included under the term “eaf-expansions, the. bundles dis- tribute themselves along the surface, sometimes ending free, sometimes anastomosing with one another in a reticulate manner. The bundles, especially those in the flattened expansions, lie as a rule in the protrusions or ridges of the surface, which are known as nerves, ribs, or veins. The course of these, the wervation or rzbding, and that of the vascular bundles often exactly coincide, both phenomena are therefore usually indicated by the same name. There is no further objection to this convenient use of the term; but it must be pointed out that two different phenomena are thus dealt with, one belonging to the external formation of the parts, and referring to the relief of the surface, and another, which refers to the inner structure ; and that though the two phenomena are always closely correlated, they do not coincide always, or in all points. The phenomena of surface-relief as seen in coarse ribbing may be here assumed as known, the reader being referred to the literature on this subject, especially to that of Pteridography and Palzontology*. In treating of the arrange- ment of the vascular bundles we must take into consideration (1) the divarication of the bundles, that is their course in the direction of the surfaces of the leaf-expansions (Sect. 91); (2) their position—as seen in a transverse section—within the other tissues (Sect. 92). Sect. 91. As regards the divarication of the bundles it must be premised that for the middle portion (Rachis, Petiolus communis) of compound and deeply-divided leaves, and for the main ribs of many, especially of large leaves, the same rules hold good as have been above stated for the course of the bundles in the petiole. The bundles are continued from a petiole, which contains several bundles, into the main nerve, and here are arranged in a channel open upwards, or in one or several circles, and are connected by anastomoses one with another: in their course towards the periphery some of them pass out into the branches of the nerves, others give off branches into them, while they decrease in number and size proportionately. The stronger lateral nerves of a lamina may also contain several bundles, e.g. Quercus pedunculata. A.B. Frank (/.c.) has given an exact description of the course of the bundles of the leaf in this plant, and notes on the same in other plants. The bundles which pass into the foliar expansion, either as branches from the above-named parts, or directly from the nodes, either remain unbranched, or give off branches, cften up to high orders, their strength diminishing as a rule with each higher order, but in a degree which varies greatly in each individual case. The bundles and branches of whatever order either end free in the foliar expansion, or unite and anastomose with others. Free ends sometimes lie at the periphery of the foliar expansion, in flat leaves especially at the margin and point, occasionally also at the surfaces: sometimes they 1 As chief works and sources of information the following may here be cited: L. von Buch, Ueber die Blattnerven, and Die Gesetze ihrer Vertheilung, Monatsbr. d. Berliner Academie, 1852, p. 42.—C. von Ettingshausen, Die Blattskelete der Dicotyledonen, Wien, 1861, fol. ; and the following articles of the same author from the Sitzungsberichten (S.) and Denkschriften (D.) der Wiener Academie: Apetale (D. XV); Papilionaceze (S. XII) ; Bombaceze (D. XIV); Celastrineze (D. XIII) ; Euphor- biaceze (S. XII); Loranthaceze (D. XXXII); Graminez (S, LIL. 1).—For the Ferns compare Mettenius, Filices Horti bot. Lipsiensis. 300 PRIMARY ARRANGEMENT OF TISSUES. lie internally : zw/ernal endings. The bundles which end at the margin and point may be termed, in accordance with the terminology of the coarser nervation, apically directed (acrodrome) and marginally directed (craspedodrome). The points of end- ing of strong marginally directed bundles are often the ends of teeth and laciniz. Anastomoses may appear between branches of any order, between equivalent and non-equivalent ones, and at any point of the foliar expansion. They give the bundle system the form .of a net (reticulate veins), which varies greatly in individual instances. An allied special form, which is especially common in flat leaves, is found where anastomosing bundles describe curves close within the margin and following its.outline: these are curved (camptodrome) bundles, according to the terminology of nervation. : The extremely various individual cases, which arise by various combinations of the above phenomena, group themselves under two main types, namely expansions with bundles having a separate course, which end free, without anastomoses; and such as have anastomosing bundles. 1. Bundles with separate course and free ends are found in the rudimentary and submerged leaves of many Angiosperms of the most various orders, in the foliar expansions of all Gymnosperms, with the exception of Gnetum and Stangeria, and in the leaves of many Ferns. One unbranched bundle, or the ramifications of branched ones, traverse the foliar expansion, and end free either internally or usually at the margin. Rudimentary scale-leaves of Angiosperms often have their vascular bundles thus arranged, when they are present at all ; the same may be said of the cotyledons of Mo- nocotyledonous plants with one median bundle, or with two running near to the middle line, or with more than two. The cotyledons of Dicotyledonous plants have been but little investigated with regard to the relations in question, many have certainly a reti- culum of bundles, even when they are ‘single-nerved.’ These most simple forms of leaves have been but little attended to in relation to the structural conditions under consideration. Of larger foliar expansions many submerged leaves of Dicotyledons (Batrachium, Myriophyllum*) with one bundle in each segment of a leaf belong to this series; also Pseudocallitriche with one median bundle*; Elatine Alsinastrum with one median bundle, which usually gives off some marginally directed branches into the narrow submerged leaves, &c. ; similarly among Monocotyledons, e. g. the rudimentary simple median bundles of the Hydrilleze. In the foliage of land-plants there is one simple apically directed bundle in each of the scale-like rudimentary leaves of the Casuarinas and of Arceuthobium, as also in those of Equisetum and Ephedra, which resemble them in habit. Among Gymnosperms the foliar expansions of all Conifers* belong to this category: the leaves of the Cupressinex, of Taxus, Phyllocladus, &c., with one median bundle; those of the Abietineze which usually have two very close together, with median parallel course, rarely (Abies Pindrow) with one simple bundle; the double leaves of Sciadopitys with two which have a parallel course near the median ' Askenasy, Botan. Zeitg. 1870, Pp. 196.—Véchting, Myriophyllum, /.c, * Hegelmaier, Monogr. d. Gattg. Callitriche, p. 31. * Geyler, /.¢. (see above, p. 245).—Thomas, in Pringsheim’s Jahrb. IV, p. 43.—Strasburger, /.¢, ‘BUNDLE-SYSTEM IN THE LEAVES. gor Iine. The leaves of the broad-leaved Araucarias; and: of species of Dammara and. Nageia, have more than two bundles, which run unbranched from the base towards the apex, and end some of them at the apex, others below it. In Ginkgo the two bundles which pass from the petiole into the lamina, branch repeatedly into marginally directed forks. On Phyllocladus compare Strasburger, Z. c. The pinne of Cycas contain one median bundle, those of most Cycadez numerous unbranched bundles, which run parallel or slightly curved from the base to the apex. In Stangeria they are traversed by one middle nerve, in which 6-8 bundles run side by side; these give off branches laterally, which are arranged in a pinnate manner, and sometimes curve towards one another and anastomose close to the margin’. The leaves of Gnetum have, as far as is known, a typical reticulate bundle-. system ; in the leaf of Welwitschia there is a peculiar arrangement, which will be described below. Among the Pteridophyta, besides the Equiseta already referred to, may here be cited the awl-shaped leaves of Pilularia, Iséetes, Lycopodium, and Selaginella; also the leaves and portions of leaves with fan-like, dichotomous, branched bundles, representing the Cyclopteris-nervation (e.g. Adiantum, Marsilia), and those with ‘bundles branched once or repeatedly in a pinnate manner, all being disconnected and marginally or apically directed, which compose the nervation of Cznopteris, . Ctenopteris, Pecopteris, Tzeniopteris, Sphenopteris, Eupteris, and N europteris *, 2. The foliar expansions with anastomosing bundles may be divided according to the arrangement of the latter into two subordinate types, which may be termed the striated and reticulate. (a) In the striated type numerous bundles run separately and parallel along the leaf-expansion, the median ones running straight to the apex, the rest diverging the more from this straight course the nearer they are to the margin, and the more the boundary lines of the latter depart from parallelism. Most of these bundles curve towards one another close to the margin, and unite so that each one affixes its acro- scopically-curved end on the basiscopic side of the one next it-in the direction of the median line. Free ends are rare. Throughout their course the bundles are con- nected in a ladder-like manner by thin transverse branches. The former bundles may accordingly be called shortly longitudinal bundles, in contradistinction to the transverse branches. This arrangement is found, as far as is known, almost exclusively in the Monocotyledons and in the leaves of the majority of their families, also in the phylloclades of Ruscus and Myrsiphyllum, in the latter cases with transitions to the reticulate form. Some few families of Monocotyledons, such as the typical Aroidez, Dioscoreze, Taccacez, and many Smilaceze, are exceptions. Of plants which are not Monocotyledons, the leaves of Welwitschia and of many narrow- leaved species of Eryngium, as E. pandanifolium, E. junceum, &c., belong to this category, or are at least allied. 2 According to the course of the longitudinal bundles we may here again distinguish two subsidiary forms, which are it is true connected by intermediate 1 Kraus, in Pringsheim’s Jahrb. IV. 7. ¢. * Compare Mettenius, Filices Horti Lipsiensis, p. 2, &c. 302, PRIMARY ARRANGEMENT OF TISSUES, examples (e.g. in the Draceenas). In the one, which may be called the longijudinally' striated, all the bundles run, in the manner indicated, separately from the base to the apex of the leaf or lamina. In the other, with pinnate sirtation, numerous bundles enter the midrib of a flat leaf, and pass through it towards the apex. They then pass one after another from the midrib into one or other half of the leaf, giving off numerous branches into it; only one or few of them extend to the apex of the leaf itself. All the bundles and branches which pass into the halves of the leaf are arranged in a pinnate manner, and have an acroscopically curved direction. The pinnate arrangement is characteristic of the group of Scitaminez, of the broad- leaved Draczenas, Curculigo, many species of Hamanthus (e.g. H. coccineus), Eucharis amazonica, &c. The longitudinally striated arrangement is characteristic of the majority of ordinary, linear, tapered leaves of Monocotyledons, also for the fan-shaped leaves of Palms, and the Foliola of the pinnate-leaves of the same. family. The longitudinal bundles which traverse the Monocotyledonous leaves of this category are often of almost equal strength, inasmuch as they enter the leaf from the stem as so many bundles of the trace, or from the node, as almost equivalent branches of one bundle of the trace (e.g. species of Potamogeton). On the other hand, it not unfrequently happens that they are not of uniform origin, some arising as branches from the others, but all pursuing fundamentally the same course, This has already been noticed above for the pinnately striated forms; the same occurs also in the longitudinally striated. The longitudinal bundles of one leaf are not uncommonly of almost equal strength; in other cases of very unequal strength. In Palm leaves Mohl distinguishes bundles of three different strengths. Frequently one median bundle, which exceeds the rest in strength, is found in longitudinally striated leaves; in many leaves of Orchids with five and more projecting ribs (e.g. Stanhopea, Acropera, Maxillaria squalens) there is one bundle in each rib, which is distinguished by its size from the rest, which are not prominent. In the pinnately striated leaves of Heliconia farinosa, the ends of the bundles which pass out from the midrib are much stronger than their branches which pursue a similar course, a difference which cannot be recognised in similar leaves of allied plants, e.g. Phrynium setosum. From the example from the Orchidacex it cannot be con- cluded that in striated leaves generally the strength of projecting nerves must cor- respond to that of the enclosed bundles. In the keel-like projecting midrib of species of Carex, and in Pandanus pygmeus, there is a bundle, which far exceeds the rest in strength; in the thick midrib of Zea Mais and other large leaves of grasses there are several with the same arrangement as in the flat halves of the leaf, and resembling those which traverse the latter, with the exception of the rather stronger median bundle. The transverse branches which connect the longitudinal bundles like the rungs of a ladder are often almost equal to them in strength,—e. g. Rhapis flabelliformis, Vanda furva,—but usually much weaker, being even reduced to a single vascular tube, or row of tracheides, for instance those in the pinne of species of Chamzdorea, the leaves of Curculigo, Zea, &c., which are even hard to find. Their number on a-given surface varies according to the species: on the average the distance between two may be about 1mm, it is often greater, rarely they are much more closely arranged—in Phrynium setosum on the average 1o--12 inadistance of 1™™, They run either almost exactly transverse to the longitudinal bundles, so that the whole system of bundles consists of rectangular meshes; or they have a more or less oblique direction. Further, they pass either from one bundle to the next, or in very many cases they pass the next bundle, only touching it externally, and run to the second or third lateral bundle. It is often seen, especially in leaves with alternating stronger and weaker bundles, that they connect those of equal strength, running past the intervening ones of unequal strength. The transverse bundles BUNDLE-SYSTEM IN THE LEAVES, ~ 303 above described (p. 265) in the halms of many Monocotyledons have the same course as in the leaves, , It is only rarely that in leaves of Monocotyledons single transverse branches end blindly in the surrounding tissue. It is more commonly the case for a longitudinal bundle to arise as a branch from one of them, : The huge leaf of Welwitschia is traversed longitudinally by very numerous strong parallel bundles, these being con- ’ nected in a ladder-like manner by transverse branches. The transverse branches either run at right angles from the longitudinal bundles, or obliquely, and sometimes directly and simply from one longitudinal bundle to the other, some- ‘ times converging and anastomosing one with another in the narrow intervening spaces. Here and there a transverse branch ends freely in the parenchyma, without reaching the next longitudinal bundle; and from each of the transverse connections there often starts one short branch, which also ends blind in the parenchyma, and is always directed to- wards the base of the leaf (Fig. 145). The bundle-system is accordingly similar in most points to that of the longitu- dinally striated Monocotyledons, but the numerous internal free ends correspond to the usual condition in the reticulate Dicotyledons.—In the leaves of the above-named species of Eryngium only transverse branches between the parallel longi- FIG, 145.—Welwitschia mirabilis ; a i aig piece of the network of vascular, tudinal bundles are found; in other similar narrow-leaved bundles in the leaf, prepared free} species, as E. aquaticum, there are also free endings and reti- 2#nified about 4. & the edge of the piece nearest the base of the culate anastomoses, leaf. _ (4) In the reticulate type (Fig. 146) the bundles which enter the leaf undergo. branching of higher or lower order, and the branches are distributed over the whole surface, run in different directions, and are sometimes connected into polygonal or curved meshes, or some- times end free, internally or peripherally. Meshes of higher order are enclosed in those of lower orders. The marginal sides of all marginal meshes form to- gether in flat leaves a sympodial bundle, which follows the margin: this is more or less near to the actual margin, and is not uncommonly situated in the extreme margin itself (e.g. Quercus pedunculata, Banksia, Lauracez, Cocculus laurifolius, and many other cori- aceous leaves). The bundles which end free internally arise as branches from the sides of the meshes, and terminate in the area enclosed by these, often after repeated short branching. FIG, 146.—Psoralea bituminosa (40). The As far as is known all foliar expansions of the Sit branching: of the bundles in a piece of Dicotyledons, with the few exceptions above quoted, belong to this type. The leaves of many land-plants, such as those of many species of Trifolium, which, from the appearance of their coarse nervation, seem to belong to the type with bundles pursuing a separate course, are still no exception; nor are the small linear “one-nerved” foliage-leaves of Dicotyledonous plants, such as species of Erica, Passerina, Fabiana imbricata. Among Monocotyledons the 304 PRIMARY. ARRANGEMENT OF TISSUES. Dioscorez belong to this type, and also many Smilacex, especially Smilax, Taccacez, Lapageria, Philesia, &c.; among the Gymnosperms, Gnetum ; finally, the reticulate Fern-leaves, which represent in Pteridography the types of nervation of Gonio- phlebium, Phlebodium, Doodya, Marginaria, &c. The typical Aroidez, the broad- leaved Potamogetons, and Hydrocharis belong also to this category, but appear as intermediate forms between the type under consideration and the striated type. It is well known that the bundles of this type enter the expansion singly or several together in a midrib, which runs towards the apex, and give off from it branches of the first order, arranged in a pinnate manner, into the two halves of the leaf (Folia penninervia) ; or several separate main bundles diverge from the insertion of the leaf, and are also in their turn pinnately branched (Folia palmatinervia, peltinervia, tripli- nervia, &c.), The branches of higher orders are also sometimes arranged in a pinnate manner, sometimes they show (true or false?) dichotomy. The number of orders of branching usually ranges in the Phanerogams from five to eight. In the Ferns the branching is simpler in all respects than in the Phanerogams of this series. It has already been noted above that branches of each order may end free or anastomose peripherally or internally. As regards the occurrence of these modes of ending the following cases occur :— 1. In rare instances reticulate connection between ai? branches, free ends occurring at most only in the apex of the leaf. In many succulent plants—species of Sempervivum, Mesembryanthemum—all free endings are absent, or at least have not hitherto been proved to exist: but an exact investigation of this point is still to be desired. Byt by no means all succulent plants belong to this type: Salicornia, e.g. has numerous internal endings in its cortical reticulum of bundles (p. 297), and the shrubby species of Crassula have numerous peripheral ones. Among Monocoty- ledons those Aroids which have been investigated (species of Anthurium, Pothos, and Monstera, Calla, Richardia) belong to this series, and further Hydrocharis and Pota- mogeton.- All'these have a free end at the apex of the leaf. In the two last-named genera the main bundles have a course similar to that of the longitudinal bundles of the striated type, the transverse branches which connect them are repeatedly branched, and the branches are connected into a net with angular meshes. In the Aroidex also the course of the main bundles réminds us of that of the striated Monocotyle- donous leaves. Between them branches ‘of many orders form a complex angular network. Free internal ends are absent, or present only rarely here and there. 2. Free internal endings within the meshes, and no free peripheral endings. Close to the margin of the flat leaf runs the sympodial marginal bundle, which limits all the marginal meshes externally, and gives off no branches in a peripheral direction. The leaves of species of Ficus and Banksia, of Cocculus laurifolius, Buxus, Quercus pedunculata’, and Psoralea (Fig. 146), may be named as examples of this pheno- menon ; apparently very many leaves, especially tough long-lived ones, with an entire margin, resemble these. Still I must cite no further examples, since the existing works on the coarser nervation do not permit of a certain decision whether short and thin free-ending bundles pass peripherally from the sympodial marginal bundle or not. z Frank, ‘Botan. Zeitg. 1864: p. 380. BUNDLE-SYSTEM IN THE LEAVES, 305 3. Both internal and peripheral ends occur in all (?) Ferns of this category, though in many(e.g. Ophioglossum vulgatum, pedunculosum, Platycerium) the peripheral ends are but few; further in numerous Dicotyledons, and species of Smilax and Dioscorea. In the flat leaves of the Dicotyledons the terminating bundles sometimes run along strong (marginally directed) nerves, sometimes they are small short branches which terminate at the margin, and especially in the marginal teeth. It is a phenomenon of especially frequent occurrence that two or more branches, one from each side, unite near to the margin with one stronger, marginally-directed bundle; they then run together outwards to the margin, so that in other words the free ends pass off from the sympodial marginal bundle. This is the case in the leaves of Primula sinensis, Papaver, Brassica, Fuchsia, Calendula, Cucurbita, Mercurialis, and Camellia japonica. The leaves of the Cupuliferze’, Betulacez, Myrica, Planera, Ulmus, species of Tri- folium, Tropzolum, &c. are further examples belonging to the present category, which however have not been exactly investigated as regards the last-mentioned condition. Small branches go from the sympodial marginal branches towards the margin, in the investigated species of Smilax and Dioscorea. Sct. 92. As regards the arrangement of the bundles as seen in a transverse section of the foliar expansion, they run, of course with the exception of their peripheral ends, within the other tissues, not superficially, those in the ribs being enclosed by the collenchyma and sclerenchyma, or by the aqueous tissue (p. 116), which forms the mass of the projecting rib, or rarely by the chlorophyll-parenchyma, which extends in greater or less bulk into the rib. For those which do not lie in prominent ribs, that is, for the smaller branches of most ramifying bundles, and for all the bundles of many fleshy Monocotyledonous leaves, the ru/e holds, that they lie close within, or below the inner limit of the chlorophyll-containing palisade-cells or rows of cells, which are perpendicular to the surface of the leaf, but are not embedded in that tissue. Thus in the bifacial leaves (comp. Chap. IX) they lie in the spongy paren- chyma, where this borders on the palisade layer; in leaves whose tissues are arranged on the centric type, at the periphery of the (colourless) middle layer; in interme- diate forms, such as Dianthus Caryophyllus and Crassulacez, at the point where the rows of cells which run inwards perpendicularly from the whole leaf-surface meet at the middle of the leaf. In the first and last cases the bundle-system is accordingly extended, in reference to a horizontal leaf, in a horizontal plane, in the second case on the surface of a much-flattened hollow body. I know of no exceptions to this rule in bifacial leaves. In concentrically constructed expansions with a relatively thin middle layer, consisting only of a few layers of cells, all the bundles are often enclosed within the latter (e. g. leaves of Statice monopetala, Phyllodes of Acacia marginata), or the stronger bundles are within the middle layer, and only the thinner branches at its outer limit (Hakea ceratophylla, Acacia longi- folia, Huegelii). In the leaf of Agave americana the thick middle layer is traversed by several series of bundles, which, with the exception of the central ones, stop short towards the margin; they run parallel to the surface of the leaf, and are connected by anastomoses: besides these there is an external series which extends round the whole leaf, at the limit between the chlorophyll layer and the middle layer. In each of the 1 Von Ettingshausen, Blattsk. d. Dicotylen, Taf. I, II, &c. x 306 PRIMARY ARRANGEMENT OF TISSUES. longitudinal lamella of the thick middle layer of the leaves of Typha and Sparga- nium, which are separated by large air-cavities, there lie 1-3 longitudinal bundles: numerous smaller ones are placed at the outer limit, abutting on a bundle of hard hypodermal sclerenchyma. In the thick leaves of the Mesembryanthema (M. lingue- forme and its allies, M. barbatum, imbricatum, stramineum, &c.) the main branch- bundles run longitudinally through the centre of the middle layer, and send out branches on all sides in an obliquely apical direction, which extend with their reti- culately connected ultimate ramifications to its outer limit. Also in the thick leaves of the Crassulaceze, and even of the Semperviva, divergences of the bundle- branches and meshes are found towards the surfaces: this is most strikingly seen in the branch-bundles which end at the surfaces in the thick-leaved species of Crassula to be described below (Sect. 111). On the arrangement of the bundles peculiar to the fructiferous leaves of Platy- cerium, which, from its form, belongs to this category, and upon the peculiarities of vascular bundles in the sporangium-bearing leaves of Ferns generally, which will not be further treated in this work, the Pteridographic literature should be referred to, especially Mettenius, Filices horti Lipsiensis. It is well known that within all types there is found the greatest variety in the direc- tion of the bundles, and of the corresponding ribs of different orders, in their divergence, number, and relative strength. It is the province of the special description of plants to enter into the details of these conditions of nervation. The examples of amphibious plants, of leaf-like phyllode branches, of plants with rudimentary leaves and cortical networks of bundles, show that the bundle-system may be immediately dependent upon different adaptations. Sometimes it is altered accord- ing to different adaptations of morphologically equivalent members: the submerged leaves of the water Ranunculi and of Elatine Alsinastrum belong to the first type with a separate course of the bundles, while the aerial leaves of the same species belong to the reticulate type: sometimes a similar bundle-system appears in morphologically different members subject to similar adaptation: phylloclades of Myrsiphyllum, and Ruscus, as compared with foliage leaves of allied plants. On the other hand, the different main and subsidiary forms of the vascular system cannot in most cases be referred directly to adaptative causes. Within a narrower or wider circle of relationship the same type of nervation occurs, whether the adaptation be similar or different, and the converse is also the case. Further the nervation is to a great extent independent of the form of the leaf. After what has been said above it is superfluous to adduce examples of this. Among the large divisions of the vegetable kingdom the Dicotyledons show the greatest uniformity of ground-plan of nervation, since their aerial foliage leaves, with the one ex- ception of the narrow-leaved Eryngia, all belong to the reticulate type; and individually they show the greatest variety, since in this type variations and combinations of dif- ferent points of detail are possible, and really exist. Among the Monocotyledons the great majority of forms belong to the striated type; which shows generally an extraordinary uniformity in the main phenomena. Only the few above enumerated families and genera appear as remarkable exceptions, since some of them correspond exactly to the reticulate type of Dicotyledons, while others approach it. Among the Gymnosperms Gnetum alone (in accordance with other points in its mor- phology which approach nearest to the Dicotyledons) has a truly Dicotyledonous nerva- * tion: the pinnz of Stangeria have only single marginal anastomeses: all other forms have bundles with separate course, CONNECTION OF THE BUNDLE-SYSTEMS, 307 The variety of the vascular system in the large foliar expansions of the Ferns is worthy of observation: here even in a narrow circle of relationship, e.g. in the genera Polypo- dium and Aspidium (in the sense of Mettenius, Fil. hort. Lips.), some species have separate bundles of the most simple arrangement, others reticulate veins; here also there are found sometimes in one and the same species, sometimes especially in different species, all intermediate forms between the most different types. The number of the bundles on a given area is always small in Ferns as compared with Angiospermous Phanerogams, but the plan of their distribution, e.g. in Ophioglossum vulgatum and Platycerium, is often the same as in the reticulated leaves of Dicotyledons, d. Connection of the bundle-systems of shoots and branches of different order. Sect. 93. The bundle-system of the lateral, similar or dissimilar branches of one relatively leading axis is continuous with that of the latter, and inserts itself upon it. The form in which this is brought about depends in the main upon the morpho- logical quality of the leading and lateral axes, the morphological point of origin of the latter, and the course of the bundles within the axes under consideration. Specific peculiarities are found besides in many cases. According to these relations the following summary may be subdivided thus :— I, SIMILAR BRANCHES OF LEAFY STEMS, 1. Norma Brancues'. a. Dicotyledons and Gymnosperms with a ring of bundles. Sect. 94. The normal branches of the Dicotyledons and Gymnosperms treated in Sect. 61-63 are in the large majority of cases axillary; we shall therefore speak here of these exclusively. There are but few relevant investigations on the rarely occurring extra-axillary branches, some few of which are referred to below. The primary bundle-system of the axillary lateral shoots, when it consists of leaf-traces, shows four main forms of insertion on that of the leading shoot. In most cases it unites itself at the point of insertion of the branch into two ora few bundles: these insert themselves, at the node of the leaf which bears the shoot, on those bundles of the trace of the leading shoot which border on the gap of the bundle-ring (gap of the leaf which bears the shoot) formed by the exit of the median bundles of the trace. In a second, apparently less numerous series of cases, the two or few bundles of the base of the branch enter the ring of bundles of the leading shoot, at the node of the leaf which bears the shoot; they pursue an individual course down to a lower node, and here insert themselves like bundles of the leaf-trace. In these two cases the arrangement of the bundles of insertion is always such that there is direct continuity between the pith of the leading and lateral shoots. In a third series of cases the bundle-system of the lateral shoot inserts itself 1 In the sense of Sachs, Lehrb. p.174, 2nd Eng. ed. p. 171. X 2 308 PRIMARY ARRANGEMENT OF TISSUES. externally, at the node of the leaf which bears it, with numerous bundles, on the many-bundled ring of the leading shoot, so that the pith-cylinders of the shoots of the two orders are only connected by narrow medullary rays. The fourth category is represented by many Cactaceze of very peculiar character, to be described below. A number of instances of the first and second category have been carefully investigated by Nageli and others who followed him in investigating the course of the primary leaf-traces'. In those belonging to the first series, and these are the most numerous, the bundle-system of the axillary shoot is united at the point of in- sertion into two bundles, which may be called the bundles of insertion. These affix themselves to the bundles of the trace at or immediately below the node of the leaf, which bears the shoot, either— (2) Upon the bundles descending from above, which form the lateral limit of the gap of the leaf which bears the shoot, the one being inserted on the right, the other on the left—Iberis amara, Lupinus (axillary shoots of the cotyledons, Fig. 94, p. 238), Passiflora Vespertilio (axillary tendrils), Antirrhinum majus, Urtica Dodartii, also Pisonia ; Juniperus (Fig. 108, p. 246), and the short shoots (the bundles of needles) of Pinus (Fig. 110, p. 247); or (4) on the bundle or bundles of the trace of the leaf itself, which bears the shoot—Anagallis arvensis (axillary peduncles), Clematis (p. 245). In Satureja variegata both the cases designated (a) and (4) are found to occur. In Galium and Rubia first the two bundles of the trace of the first pair of leaves, and then those of the second pair of leaves of the axillary shoot insert themselves, usually at the node, on the bundle which passes into the leaf bearing the shoot; the same is usually the case in Russelia juncea, and sometimes in Spergula arvensis. In the second series of cases the two bundles of the axillary shoot pass, at the node of the leaf which bears it, into the bundle-ring of the leading shoot, and pursue an individual course down through one (e.g. Aristolochia, Fig. 96, p. 239, Lathyrus Aphaca, Figs. 98, 99, p. 240), two (e.g. Cerastium frigidum, Figs. 102, 103, p. 243), and even three internodes (e. g. axillary peduncles of Viola elatior), and then insert themselves on bundles of the leaf-trace. For further examples and details, see Nageli, Z.c., and above, Sect. 61, p. 235. In the species of Galium, Rubia, Spergula,- and Russelia above enumerated there is, according to Nageli, either direct insertion at the node of the leaf which bears the shoot, or the bundles pursue an individual course down through one internode or more. In Vitis vinifera usually three bundles pass from the axillary shoot, and also from the extra-axillary tendril into the main shoot, and pursue a separate course through one internode. The two or three bundles of insertion of the axillary shoot arise either by the coalescence of the several bundles of the leaf-trace of the lowest internode to two at the point of insertion, e.g. Clematis; or these are the single bundles of the trace of the two lowest leaves, e. g. Galium. The connection of the axillary bundle-system with that of the leading shoot is however not limited to the bundles of insertion above described. According to Frank’s investigations? on Taxus, Quercus, Bidens, and Solidago, there appear’ con- * Compare above, § 61. ? Botan. Zeitg. 1864, pp. 154 and 382. CONNECTION OF THE BUNDLE-SYSTEMS. 309 necting bundles, one in Quercus, in Taxus about three, which run down from the upper margin of the gap of the leaf-bearing shoot to the bundles of insertion. By means of these and further ‘completing bundles’ the point of connection of the pith of the main and axillary shoots is soon enclosed bya ring. The completing bundles doubtless belong to the secondary formations of intercalary bundles (Chap. XIV); the same is probable for the above connecting bundles, but their relation to the leaf- traces of the axillary shoot requires more exact investigation. As regards the insertion of the bundles of axillary shoots, two or more of which are seated one above another, Frank’s statements may here be reproduced word for word (i.e. p. 388), but the subject is recommended for further research. ‘In Rubus two buds are seated in the axil of the leaf closely one above another, their vascular bundle-systems are united in the lower parts one with another, and are connected with the vascular system of the stem as though they belonged to one single axillary bud. After both lateral series (i.e, the branches of the two original bundles of insertion, De Bary) have united at their anterior ends (i.e. next the leaf which bears the shoot) each separates at the middle, and the anterior halves close up to form a circular system for the lower bud. The remaining posterior halves soon unite at their anterior ends and form the vascular bundle-system of the upper bud. The posterior parts of both circles of vascular bundles are here also closed by descending bundles, which accordingly arise in the case of the lower bud from the vascular bundles of the upper.—In the axillary buds of Lonicera Xylosteum, of which often as many as four are seated one above another, but which are usually separated some distance one from another, and the uppermost of which can hardly be distinguished externally from an adventitious bud, the lower parts of the vascular bundle-systems are also inserted between the members of the vascular bundle-ring of the stem, but in this case each system is independently connected with the stem, since the gap of the vascular bundle-ring of the mother shoot closes above each bud, and only opens again immediately below the insertion of the next, at which point the bundles of the bud arise on both sides from the margins of the open vascular bundle-ring.’ The third, less common case of insertion of the axillary bundles outside the closed bundle-ring of the leading shoot occurs in the Umbbelliferee, though not in all of them. The bundles of the leaf-trace of the lowest internode of the axillary shoot unite in this case at the node of the leaf which bears the shoot, to form one cortical bundle, which immediately divides into two arms; these pass off right and left, and embrace the bundle-ring of the leading shoot transversely like a girdle. From this girdle branches arise in pairs side by side and pass downwards. Each of these pairs bestrides one bundle of the trace of the leaf; which bears it, from above and outside, and inserts itself on its two sides, at the point where it curves outwards from the bundle-rings of the leading shoot. ‘The bundle-system of the axillary shoot is thus attached outside the ring of the stem by the pairs which bestride the same number of emerging bundles of the leaf which bears it. It thus embraces either the whole circumference of the node, bestriding all the emerging bundles of the stem-embracing trace of the leaf which bears it—Foeniculum, species of Heracleum, Cherophyllum, Myrrhis, and Archangelica; or a part of it bestriding only some few bundles -of the leaf which bears it—Ethusa Cynapium. A continuity of pith between the shoots of both orders is thus only possible by the narrow medullary rays. The phenomenon stands thus in the mature plant’. The exact study of the 1 If I am not mistaken, it has been long ago described by C, F. Schimper as ‘Astkorb;’ I have not succeeded in my attempts to find the reference. 310 PRIMARY ARRANGEMENT OF TISSUES, history of development remains still to be made. The arrangement can still be re- cognised after considerable secondary thickening of the stem; the stem-embracing insertions of the branches which are thick, but attached by thin bases, weave in a basket-like manner round the node of the leading axis: it is especially developed in the perennial subterranean shoots of Myrrhis, species of Cheerophyllum, &c. Many Umbelliferee have, as is obvious from the continuity of the pith-cylinder on both sides, another form of axillary insertion, which remains to be more exactly investigated: thus Silaus pratensis with its medullary bundles noticed on p. 253. The same form of axillary insertion as in the above-named Umbellifere is found in Aralia japonica; it remains to be investigated whether the same is the case in other Araliacea2, and in other families in which the leaf-insertion, and perhaps also the bundle-system, resembles that of the Umbellifere, e.g. Ranunculaceze with alter- nating leaves. The above-mentioned fourth, and very special case of bud-insertion occurs in species of Echinocactus, and some of Cereus with thick shoots (C. candicans?). Its development requires investigation. In the mature state the leaf-traces having one bundle are found united to form sympodial bundles, which are separate and per- pendicular, and of equal number to the angles of the stem (Echinocactus), or are connected in a reticulate manner; between them are broad medullary rays. The leaf-bundles run slightly obliquely, almost horizontally upwards towards the lower margin of the spine-cushion, that is towards the point of insertion of the rudimentary leaves. Just above each foliar-bundle and in a direction almost parallel to it, the thick cortex of the stem is traversed by some few vascular bundles, which are near one another, and have their xylem-portions turned towards ‘one another; these belong to the axillary bud formed above the rudimentary leaf, and attain a consider- able strength as soon as the bud developes into a shoot. These bundles of the bud now pass through the medullary rays, between the sympodia of the leaf-trace in the stem, into the pith, and there branch freely in all directions, their branches being united one with another to form an elaborate plexus traversing the whole pith. This system of bundles of the bud is only directly connected with the sympodia of the leaf-trace by single short connecting bundles at the points of passage through the medullary rays. In the Opuntias, Cereus speciosissimus, &c., and also in the Rhipsalidaceze’ this phenomenon is wanting; the bundles of the bud, as far as investigated, are inserted, in the manner usual for Dicotyledons, partly on the cortical bundles, partly on those of the bundle-ring: medullary bundles are altogether wanting. In the Mamillarias, which have medullary bundles (p. 254), no connection between these and the young lateral shoots has as yet been discovered. The above-named plants have accordingly a system of medullary bundles, which differs fundamentally in its significance from the others described above on P. 253. Where these latter, and where cortical bundles occur in Dicotyledons, the axillary insertion occurs, as far as is known, in one (usually the first) of the typical forms, with the addition of direct connection between the medullary or cortical bundles of the leading shoot and of the lateral shoot. * Compare Véchting, 4c. (p. 261). CONNECTION OF THE BUNDLE-SYSTEMS. 3II 5. Monocotyledons and Phanerogams with axile bundle. Szct. 95. Among the Monocotyledons belonging to the Palm-type (Sect.65—6 7)the numerous bundles of the lowest internode of the normal axillary shoot enter, in the Palms’, Draczenas, Liliaceze, Aroidese, Orchidaceze, &c., at the node into the bundle- cylinder of the main shoot, and pass obliquely downwards and inwards with the bundles of the leaf which bears the shoot, inserting themselves successively on peri- pheral bundles of the latter, without reaching the middle of the cylinder. In many cases, as in the rhizomes of Acorus, the axillary bundles do not penetrate further than to the surface of the cylinder of the main shoot, but spread themselves out, with abundant branching, for a great distance downwards, over the nearer longitudinal half of the main shoot, and interweaving, and here and there uniting with the bundles at the surface of the cylinder, they form a dense plexus of bundles, which is sharply limited on the side next the cortex. In those forms which have been investigated, Zea, Saccharum, Coix, Arundo Donax, &c., numerous bundles pass from the lowest internode of the axillary shoot transversely into the node, and here branch very freely, their branches spreading over the whole transverse section of the node, but only slightly in a vertical direction, and thrusting themselves between the bundles of the main axis, which run perpendicularly through the node, and between one another, and here and there inserting themselves on the bundles of the main axis. The whole series of axillary bundles forms at the node a complex and confused felt, expanded and attached in the manner indicated, and having the form of a transverse disc, which in the above-named large species reaches a height of several millimetres”; its origin is not obvious in the mature state, but it is clearly seen in young stages of development that it is formed by the insertion of bundles, starting from the axillary shoot. In the Commelinez with thin stems, as Tradescantid* albiflora, Commelina agraria, several internal bundles are found in the basal internode of the young axillary shoot (comp. Sect. 69)—e. g. three or four in the Tradescantia named—which enter the node of the main shoot, and here, turning downwards above the outgoing median bundle’ of the leaf which bears the shoot, insert themselves at the point of union of the inner bundles. In somewhat older axillary shoots there are further peripheral, doubtless in part cauline bundles, which insert themselves on the cauline bundles of the main shoot. In thick-stemmed Commelinez, such as Maravelia zey- lanica and species of Dichorisandra, the internal bundles of the lowest internode of the axillary shoot unite to a single thick bundle, which passes almost exactly horizontally into the node of the leaf which bears it, and inserts itself in the middle of this, with some few branches, on the internal bundles which descend there. In Tradescantia virginiana several bundles pass from the axillary shoot into the node of the leaf which bears it, and there divide into branches, which are interwoven as in the nodes of the Grasses between the internal bundles of the main shoot, and insert themselves on them, 1 Mohl, Palm. Struct. p. 31.—Compare also Falkenberg, /.c., and the note above on p. 276, 2 Compare von Mohl, /.¢. Tab. 9; Schleiden, Grundz. 3 Aufl. IL. p. 158. 312 PRIMARY ARRANGEMENT OF TISSUES. In the node of Potamogeton natans the bundle-system of the young axillary shoot coalesces so as to form a single bundle, which inserts itself on the median bundle of the leaf which bears it, at the point where it curves outwards. The other investigated Potamogetons—P. lucens, gramineus, pectinatus, and pusillus—show a quite similar insertion, with the difference that the bundle which comes from the axillary shoot is not inserted on the bundle passing into the leaf which bears it, but on the axile sympodial bundle which passes downwards at the node. Comp. Fig. 123, p. 273. As far as is known, the naturally very simple relations of insertion in those Phanerogams which have an axile bundle resemble those in the last-named Mono- cotyledons. c. Fern-like plants. Secr. 96. Among the Filicineze there is sometimes forked, sometimes Mono- podial branching; in many species, as in Aspidium Filix mas, Athyrium Filix foemina, both forms of branching are found side by side. Monopodial branching occurs undoubtedly in the Salvintacee, Marsiliacee, and in many Lidices. Very many Filices show a shoot-system, apparently composed of a main axis and lateral shoots, in respect of whichit is a matter of controversy whether it is a Monopodium, or a Sympodium composed of unequally developed successive forks, It will here be treated as a Monopodium, in accordance with the conclusion of Mettenius. But I remark distinctly that I only accept this conclusion in order to simplify the description Aere to be given, and that I leave the controversy in question quite undecided. The fact that the insertion of the bundles of the actual lateral shoots appears in many cases in point to support the conclusion of Mettenius cannot by any means decide the question, since unequally strong growth of originally equivalent shoots may also have as its result an originally unequal arrangement of their vascular bundle-system. The normal lateral shoots of the plants of this series, with the exception of many Hymenophyllums and Davallias, which in this point also are the subject of controversy, are not axillary, nor have they even any other constant relation to the insertions of the leaves. They arise either from the stem, and sometimes close to the point of insertion of the leaf, laterally or at the back of the latter, sometimes at a great distance from it, between two leaves; or they arise on the back or sides of the base of the petiole itself, often, as in the ordinary branchings of Aspidium Filix mas, which may be accordingly considered as belonging here, at a considerable distance from the point of insertion, in the above-named example about 2-3 centimetres from it, Comp. Fig. 132, C, p. 286. In the undoubtedly Monopodial branching of many Filices with more than two rows of leaves, the vascular bundle-system of the lateral shoot is united as a rule towards the point of origin to one thin and not hollow bundle, which is inserted on one bundle of the main axis. This is the case in the lateral shoots of Aspidium cristatum, spinulosum, Blechnum Spicant, Athyrium Filix foemina, Polypodium al- pestre, Alsophila aculeata, &c.', which appear at or below the back of the base of the * Hofmeister, Beitr. 2.¢.—Stenzel, /.¢.; compare p. 283. CONNECTION OF THE BUNDLE-SYSTEMS. 313 petiole. The single thread-like bundle, which enters the narrow base of the branch, is inserted, in the first-named plants, at the lower margin of a foliar gap of the stem; in the specimen of Alsophila investigated several bundles are thus inserted, each of which passes into one of the numerous shoots which arise round the base of one leaf. The bundle of the lateral shoot of Struthiopteris has the same point of at- tachment as in Aspidium cristatum; it is not thread-like, but has the shape of a plate with a channel-like external groove, which gradually widens as it passes from the point of attachment, and closes up to form a tube opening obliquely outwards and downwards. Through the tube and channel the pith of the lateral shoot has con- tinuous connection with the parenchyma of the main shoot, while where the bundles are filiform at their point of origin this continuity does not exist. The lateral shoots seated on the petiole in Aspidium Filix mas have as a rule the same thread-like insertion of the bundles on one bundle of the petiole. More rarely the bundle-system of the lateral shoot arises as three bundles from so many bundles of the petiole, or directly as a tube filled with pith from the margin of a gap in a widened band-like bundle of the leaf. Finally, in a species named as Diplazium giganteum, Stenzel found this latter form of insertion in the branchings arising from the stem below the leaves: each branch has its own small gap in the network of bundles of the stem, from the margin of which arises the tube-like system of the branch. In the Ferns and Rhizocarps with elongated stems bearing two rows of leaves, and with lateral shoots arising from the stem, not from the petiole, when there is a simple axile bundle present, there is naturally an insertion of the bundle which passes to the lateral shoot on that which traverses the main shoot. In the Marsilia- ceee with a tubular bundle, that which enters the branch passes directly off in a tubular form from the margin of a corresponding gap in the tubular bundle of the main axis, through which gap the pith of the two axes is continuous. In the Ferns with clearly marked upper and lower bundles (see above, p. 287), in most of the described cases?, the bundle-system of the lateral shoot is united, as in those ferns with leaves in many rows, into one bundle, which arises from the next lower transverse bundle separating the foliar gaps: it usually has the form of a channel open towards the pith. This is the case in Aspidium albopunctatum, coria- ceum (comp. Fig. 135, p. 287), Acrostichum Lingua, brevipes, most Davallias, and other above-named species, on the details of which Mettenius’ description must be referred to. According to Trécul’s investigation®, however, it appears that in A. coriaceum the bundle passing to the branch is inserted both on the transverse and on the lower bundle, at the angle between the two. Among the Davallias some species (D. stenocarpa, divaricata) differ from the rest, inasmuch as a closed transverse bundle limiting the foliar gaps is absent, but in its place, and in the direction which it has in other species, two bundles run, the one arising from the upper, the other from the lower bundle; both converge obliquely towards the apex, and enter the lateral shoot as upper and lower bundles without coming into contact. The arrange- ment in D. cherophylla described by Mettenius is more irregular still, but should be placed in this category. 2 Mettenius, Angiopteris, p. 546. 2 Ann, Sci, Nat. 5 sér. tom, XII, p. 242, 314 PRIMARY ARRANGEMENT OF TISSUES. Of the Ferns with divided upper and lower bundles, which are connected by intermediate forms with those above described, and with those in which there is only a complex, often irregular network of bundles instead of two bundles (see p. 288), this only may be stated generally, that in’ many simply constructed forms, e.g. Polypodium squamulosum, the bundle-system of the lateral shoots arises as a simple bundle from a definite mesh in that of the main axis. In the large majority of these cases several thin bundles, which enter the lateral shoot, arise from the margin of definite meshes (comp. Fig. 136, p. 287). Their number varies according to the species, from two to eight, as far as present data extend, and is always smaller than that which enters a leaf of the same species. Where medullary bundles are present, and there is continuity between the pith of the main and lateral shoot, branches separate from the medullary bundles of the former and enter the latter, e.g. Polybo- trya Meyeriana; where the bundles of the lateral shoots insert themselves as a simple not hollow bundle, such a branching does not occur. When the lateral shoots arise from the base of the leaf, modes of insertion of bundles are found which are similar and subject to similar variations to those of the bundles arising from the stem: this is shown by the above-mentioned example of Aspidium Filix mas. There are but few thorough investigations on this point. No exact investigations have been made either on the insertion of the bundles at the rarely occurring points of branching of the Osmundacez, or on the mode of origin of these branchings. In the Equiseta the bundle-system of each branch is united into a single bundle, and inserts itself externally at the node of the main shoot on the angle of branching of one of the bundles descending from the next higher leaf-sheath (comp. p. 279)!. Where the branching appears and remains as a forking of the main axis, the whole bundle-system also divides into two parts, each of which enters into one branch of the fork: both systems are fundamentally similar to one another and to that of the main axis. Among Ferns the rhizomes of Pteris aquilina? are—if we disregard for the present the above-mentioned controversies—an exquisite example of this. Comp. Fig. 143, p. 295. Also in Athyrium Filix foemina, according to Hofmeister, the phe- nomenon is frequent ; in Aspid. Filix mas it is rare; these cases are regarded, it is true, by Mettenius as an apparent forking, derived from early stronger development of monopodially formed lateral shoots. , In the Lycopodia and Selaginellz, whose branchings are always forked—though not always quite equal—the insertion of the vascular bundles which enter the branches is generally indicated by what has been said above. In the Selaginellz .with two lateral bundles in each shoot, as in 8. Kraussiana, Martensii, &c. (comp. p. 282), of the four bundles destined for each branching either three diverge acutely from one bundle of the main shoot, and the fourth is the continuation of the other which traverses the main shoot; or each of the latter divides into two branches which are very close to one another at their entry into the fork. In S. Martensii the former + See Stenzel, 7, ¢, Taf, 1V. Fig. 13. ? Compare Hofineister, Stenzel, 72, cc. CONNECTION OF THE BUNDLE-SYSTEMS. 315 case has been observed’, in S. Kraussiana? both cases are found; in the latter there is also a more or less complete transverse anastomosis at the point of branching. The passage into the lateral shoots has not been exactly investigated in those Sela- ginellze which have several bundles in their shoots, 2. ADVENTITIOUS SHOOTS. Sect. 97. It is a general rule for adventitious shoots that their bundle-system is always inserted on those vascular bundles, or points of the wood or bast body of the main axis, which are nearest to their point of origin. Since such shoots may arise at the most heterogeneous points, and at the most various periods of the life of the plant, the individual cases show great variety. If the normal course of the bundles is known, their relative arrangement is completely determined by what has been said above. II. ROOTS. Sxct. 98. In the forked roots of the Isoeteze, Selaginella, and Lycopodia, the vascular bundle forks as in dichotomous stems. Roots are found as lateral branches on members of their own kind, as well as on stems, rarely on leaves®; some appear in definite morphological positions, e. g. at definite points of the leaf-insertion ; others are without arrangement: the former may be called normal, the latter adventitious lateral roots. The invariably endogenous formation of lateral roots takes place in or close to vascular bundles or masses of wood or bast. Their vascular bundle is inserted directly and without branching on the nearest one of the main axis, or it divides into branches, which connect themselves with several bundles of the axis. The former simple insertion occurs obviously in members with a simple axile bundle—thus in almost all roots; also in stems constructed on the Dicotyledonous type, and in the Ferns. Splitting of the bundle of the root into several shanks, which insert themselves on several bundles, is a common phenomenon in the stem of Monocotyledons. It does not occur however in all species; e.g. in the rhizome of Carex hirta each root- bundle inserts itself simply on one peripheral bundle of the stem-cylinder. The branches or shanks, into which the root-bundles about to be inserted are divided, separate at the periphery of the bundle-cylinder, they then insert themselves, in a first series of cases on the bundles which are present at that point, without penetrating more deeply into the cylinder of the stem: this is the case in the investigated Orchidee, many Commelinex, Aroidez, Richardia ethiopica, Philo- dendron spec., with few short shanks, diverging chiefly upwards and downwards ; Acorus with more abundant branching ; Calla palustris* with a ring of roots, whose bundle-insertions together form a transverse girdle at each node of the rhizome. 1 Compare Nageli und Leitgeb, Entstehung, &c. d. Wurzeln, Taf. XVIII. Fig. 11. ? Hofmeister, Vergl. Unter. Taf. XXIII. p. 4. 3 [Compare Mangin, Origine et Insertion des racines advéntives .. . chez les Monocotylédones, Ann. Sci. Nat. sér. 6, tom. 14, 1882.] 4 Van Tieghem, Struct. des Aroidées, /.¢, 316 PRIMARY ARRANGEMENT OF TISSUES. In another series of cases the bundle of the root divides at the outer limit of the stem-cylinder into numerous branches, which, diverging in all directions, enter between the bundles of the stem, and penetrate with a sinuous course to the middle of it, and then insert themselves, some of them further out, others further within, on bundles of the stem. This is the case in the Palms+, where the penetrating bundles do not reach the middle of the stem, in the nodes of the Grasses, and thick-stemmed Commelinez. Also in the thick primary lateral roots of Pandanus the bundle-system of the corresponding lateral axis inserts itself on that of the main axis in the manner just described, i.e. numerous radiately diverging and undulating branches pass between and up to the longitudinal bundles, which will be described below (Sect. 108): also in the Palms, according to Mohl, the bundle of the lateral roots splits in a similar way into thin branches, and penetrates between the elements of the hollow cylindrical bundle of that which is relatively the main root. The phenomena in the above Commelinez should be rather more exactly described. Several lateral roots arise at the nodes, especially on the side opposite to the axillary shoot, and somewhat higher than its point of insertion, and the point of union of the bundles of the leaf-trace. The bundles of the latter penetrate horizontally into the stem as far as the cylindrical surface occupied by the ring of cauline bundles (see p. 269). Here they divide into horizontally diverging shanks, and these together form a slight transverse girdle passing round the whole periphery. In the thin stems of Tradescantia albiflora and Commelina agraria this girdle is without centripetal branches, In Tr. zebrina, virginiana, and Maravelia zeylanica numerous branches pass down from it in a centripetal direction, they are distributed transversely through the whole node, with sinuous curva- tures, and anastomose with the descending bundles of the stem, and with those which enter from the axillary shoot, thus forming a felt, which is less dense and deep than that which is characteristic of the nodes of Grasses, but is similar to it, B. STRUCTURE OF THE VASCULAR BUNDLE. Sect. 99. The vascular bundles are strands which consist of trachee and sieve-tubes as their essential parts. Both are accompanied by parenchymatous, and often by sclerenchymatous elements. The structure of the bundle is determined by the juxtaposition of all these component parts. A bundle undergoes in its course more or less considerable changes. Cross- sections of the same bundle, taken at a distance from one another, may show the greatest differences in the number and distribution of the individual elements; Fig. 147, for example, represents the part of a bundle of the leaf-trace of Acorus Ca- lamus passing through the leaf, while Fig. 148 shows its lower end as existing in the stem ; in the intermediate tract the one structure passes gradually over into the other. Such differences come out most conspicuously on comparing the peripheral ends of the bundles, where they are spread out in the leaves or periphery of the stem, with the other parts of them. It is therefore expedient in considering them to distinguish s * Von Mohl, Palm, Struct. p. xix. Tab. J, A, and Verm, Schr. pp. 157 and 172, STRUCTURE OF THE VASCULAR BUNDLE. 317 between dundle-ends and bundle-trunks, although a sharp limit between the two cannot be shown in any case. To the category of dundle-trunks belong chiefly the bundles which pass through the stem, roots, leaf-stalks, and thick nerves of the leaf. The description of their structure must start from these organs. Most bundle-trunks, however different in details, possess the composition which has above been indicated in a general way. In comparatively few cases their structure is simplified by essential organs disappearing or remaining rudimentary. A distinction must therefore be drawn between cucomplete and complefe bundle-trunks. Here we shall first speak of the latter. I. Bunpie-TRUNKS. Sct. 100. The essential tissue-elements of the complete bundle are trachez and sieve-tubes." The two are always so arranged that one longitudinal portion of FIG. 147.—Acorus Calamus; cross-section through FIG, 148.—Cross-section through the the periphery of the flower-stalk (145). ¢ epidermis; 3 concentrically arranged lower end of a small vascular bundles, with a sclerenchymatous sheath bundle of the leaf-trace in the stem (145). on the outside. In the middle is a large vascular bundle; The delicate and small-meshed phloem w its phloem; g outer large trachez of the xylem; 2 occupies the middle, and is surrounded . interceliular passage at the inner side of the latter. The by aring of scalariform netted tracheides ; cross-section through the leaf shows the same structure. outside this is parenchyma. the bundle includes all the trachez, while another, or in rarer cases more than one other, includes all the sieve-tubes. Thus in every bundle we have to distinguish between that part which contains the trachez (tracheides or vessels) and that which contains the sieve-tubes, or expressed more shortly, between the xylem and the phloem. In both parts the characteristic tissue-elements are as a rule not present alone, but are placed between rows or layers of cells (cf. p. 5) in such a way that all or most trachez or sieve-tubes are in contact with the latter at least at one point. It is true that in the case of very small bundles this intercalation is not uncommonly absent in the xylem; but then the few trachez of which this consists, border, for the most part at least, on the cells which encircle the bundle. More rarely masses of 318 PRIMARY ARRANGEMENT OF TISSUES. trachez, consisting of several or many rows as seen in cross-section, without inter- calated cells, occur in the xylem of thick bundles, as in the Marattiaceee, Osmun- daceze, and Ophioglossez?, and in the bundles of the leaf of Yucca, the stem of Fritillaria?, &c., to be described below. The external surface of a vascular bundle is marked off from the surrounding non-equivalent tissue in various ways, and not uncommonly in such a manner that the surrounding parenchymatous elements pass over directly and quite gradually into those which lie within the bundle itself, between the tracheze and the sieve-tubes, A smooth, sharp, external boundary is then not present at all, although the bundle itself, in so far as it consists of its essential elements, always stands out sharply. More frequent no doubt is the complete or partial limitation of the bundle by means of a distinct sheath, strand-sheath, or bundle-sheath *, in the sense of the word fixed in general at p.6. This occurs, firstly, in the form of the endodermis (p. 121), or other specialised parenchymatous layers (Chap. IX); secondly, in the form of strands or layers of sclerotic fibrous cells, or especially sclerenchymatous fibres, which border the bundle on one side or encircle it all round. All these sheaths are as a rule limited on the inside towards the bundle with the same sharpness as on the outside towards the non-equivalent surrounding mass; they can therefore just as well be assigned to the bundle itself as to its surroundings, its boundary being fixed either at the external or internal surface of the sheath. The latter is the usual and expedient practice in the case of parenchymatous, and particularly of endodermal sheaths, especially on the ground of developmental phenomena in almost all roots; yet it must not be left out of consideration that in the case of the endodermis of most bundles in Fern-stems the history of development rather leads to the opposite result. Cf Sect. 106. As regards the sclerenchymatous strands and sheaths which accompany the bundle longitudinally, it has long been customary to assign them to the vascular bundles, or, as was done by Nageli, to regard them at least as essential concomitants of the latter, and to call the united strand formed by them and the vascular bundle the fibro-vascular strand. Both from former sections, and further from the general comparative consideration of the distribution of sclerenchyma (Chap. X), it follows that these fibrous sheaths and strands do not strictly speaking belong to the vascular bundles, but to a special system of tissue, which may or may not have a common course with the latter over certain tracts. In spite of all these theoretical considerations, the anatomical treatment of the vascular bundles cannot by any means leave sheaths of whatever kind unregarded where they occur. The elements of the vascular bundle are, as far as investigations reach, almost everywhere and always in uninterrupted connexion, both among themselves and with those of the surrounding sheaths. The only not uncommon exceptions are that the xylem, especially in collateral bundles, shows intercellular spaces containing air at its inner edge, and that spaces containing secretions lie in the outer regions of the bundle. Cf. Chap. XIII. ? Compare Russow, Vergl. Unters. p. 117. ? Von Mohl, Palm. Struct. Tab. G. Fig. 11. * C. H. Schultz, Die Cyclose, /.¢. (p. 192) p. 246.—Sachs, Textbook (2nd Eng. ed.), p. 124.— Compare also Russow, Vergl. Untersuch. SFRUCTURE OF COLLATERAL BUNDLES, 319 According to the arrangement of the xylem and phloem, three main forms of bundles are to be distinguished, which are designated the collateral’, the concentric, and the radial, One and the same bundle may at different points of its course pass over from one to the other of these forms (cf. Fig. 147, 148). * 1. Collateral Vascular Bundles. Sect. 101. Collateral bundles are with rare exceptions characteristic of the stem and foliage-leaves of the Phanerogams, also of the stem of Equiseta, Ophioglossez, Osmunda, and Todea (?)%, Among parts belonging to the category of roots they occur only in the tuberously developed roots of Dioscoree (D. Batatas), Ophrydeze, and perhaps those species of Sedum which are related to S. Telephium, Cf. p. 233. In the most numerous and the typical cases they consist of a xylem and phloem portion, each of which borders longitudinally on the other with one surface, and with the remainder of its periphery on the non-equivalent surrounding tissue. A special subordinate form, to be called the double collateral or dicollateral, is distinguished from the usual one by the fact that two phloem groups lie on opposite sides of one xylem group. These will be described last, and at present only the simple collateral bundles will be discussed. The orientation of collateral bundles is in the usual cases, which we may call normal, always such that the xylem is turned towards the middle, and the phloem towards the periphery of the whole organ to which they belong. Accordingly we can use the terms inside and outside as regards the bundle in a general sense, calling the edge turned away from the phloem the inner edge, and using corresponding words for the remaining sides. In the bundle-ring of the typically constructed Dicotyledons and Gymnospernis (p. 235) allthe xylem portions lie, in consequence of the orientation mentioned, in an annular zone which immediately surrounds the pith, while all the phloem portions occupy a zone concentric with the former and external to it. The former, together with what is added later by secondary new-formation, is traditionally known as the woody ring or woody mass, the second as the éas¢, das/- ring, bast-zone or tnner corlex, and the two parts of the bundle are accordingly called the woody-part, and the dast-part or cortical-part—xylem and phloem*, the nomen- clature originally adopted for the Dicotyledons being extended to the structurally similar parts of all vascular bundles, without regard to arrangement and orientation. The same orientation is also the rule for the stems of Monocotyledons, and for all leaves or portions of leaves in which the bundles are placed in a ring around a central part from which they are absent. Where on the other hand the bundles in a leaf, or portion of a leaf, have an arrangement other than the annular one just mentioned, their phloem is turned towards the morphologically lower leaf-surface, and their xylem towards the upper, thus preserving the same orientation as in the stem if referred to the latter, the leaf being supposed to be in the erect position. As regards Dicotyledons with the typical single ring of bundles, no exceptions to these rules are known, unless perhaps in species of Strychnos (cf. Chap. XVI). 1 Russow, Vergl. Unters. * (Compare Haberlandt on collateral bundles in the leaves of Ferns, Bot. Ztg. 1881, p. 467; idem, 1882, p. 217.] 8 Nageli, Beitr. I, 320 PRIMARY ARRANGEMENT OF TISSUES, In stems, leaf-stalks, and leaf-ribs with several concentric rings of bundles, or bundles scattered in cross-section, exceptions certainly occur, though rarely; also very rarely in the lamina of flat leaves. In the positions first mentioned precisely the opposite orientation to that in the normal case is shown by certain bundles in the stem of Nelumbium (Fig. 112, p. 255), namely by those of circles 3 and 5 in the intermediate series; further by the medullary bundles in the stem of the Araliz mentioned on p. 253, by the four cortical bundles in the internode of Calycanthus, by those of the middle one of the three concentric bundle-circles in the leaf-stalk of the Lime+, and several others. In stems and leaf-stalks many bundles which are scattered as seen in cross-section have an 77regular orientation, i.e. with the two parts facing neither directly outwards nor directly inwards. Those especially which anastomose or branch often show torsions which divert them from the normal orientation, near the points of branching or of union. Examples of this are found in many leaf-stalks, e. g. Aralia japonica, Aroidez ; in the pith of Silaus and other Umbelliferze mentioned at p. 253; and in the interior of the stem of Aroidez and Pandanez (cf. p. 268). In the leaves of Typha and Agave Americana (p. 305) the bundles running through the middle layer, which is destitute of chlorophyll, all have their vascular part turned towards the upper side; in those on each side which border on the chlorophyll-parenchyma the vascular part faces the ,middle of the leaf. Of the longitudinal bundles lying in one plane in the leaf-lamina of Draczena reflexa, the median one has norma! orientation, while all the rest have their xylem turned towards this, and their phloem towards the edge of the leaf. The general form of the cross-section of collateral bundles is as a rule round or oblong; in the latter case the greater diameter passes as a rule through the middle of the outer and inner edges. In the stems and leaves of several Mono- cotyledons this unequal extension on different radii of the cross-section amounts to a marked lateral flattening of the bundle; e.g. leaves of Scitamines, Asphodelus luteus, Hemerocallis, Hyacinthus, Pandanus, leafy-stem of Canna, &c. Other shapes are rare; such as the horseshoe-shaped cross-section of the bundles in the stem of Osmunda (Fig. 128, p. 280), the annular section in the stem of Botrychium (p. 284), and in the petioles and ribs of several Dicotyledonous leaves, as those of Eriobotrya japonica, Veronica speciosa’, the pulvinus of the leaf-stalk of Mimosa pudica, &c. In these latter cases the inner edge of the ring is always the inner edge of the bundle, not only as regards orientation, but also as regards structure. Ignoring these annular bundles, and ignoring the places where several bundles meet, which we shall describe below, both the whole collateral bundle and each of its two parts form an approximately, though never exactly monosymmetrical body, with its plane of symmetry passing through the middle of its outer and inner edges. The arrangement of the elements in the bundle is also in harmony with this approximate monosymmetry, as will be shown below. The number of elements in each part, and the resulting thickness of the bundle, is extremely different according to the individual cases: many sappy herbaceous ? Compare Frank, Botan. Zeitg. 1864, p. 381. ? Areschoug, Om bladets inre byggnad. Lund's Univ. Arskrift, tom. IV. STRUCTURE OF COLLATERAL BUNDLES. 321 plants, especially Monocotyledons, and water and bog-plants, have only a group of a few (3-6) tracheze, and a phloem portion limited to about 20 or fewer elements; in the stems and leaf-stalks of many Aroidez the number of the trache falls in many bundles to 2 and 1‘; on the other hand, the thicker bundles of Monocotyledonous plants, and above all the thick bundles of the Ferns above mentioned, and of the leaves of Dicotyledonous land-plants, present very high figures. It is an obvious a prior’ conjecture that a definite ratio exists between the size of the bundles, especially of their vascular parts, and their number, and that both are definitely related to the extent of the transpiring and assimilating leaf-surface, as well as to the vigour of the root system and the arrangement of the roots. A number of facts, which have partly been stated here, partly in Sects. 61-71, and are still to be considered in Chap. XIV, point to such relations. The comparison of nearly-related species inhabiting the water and the land respectively, demonstrates among the former a considerable diminution in the development of the bundles (cf. e. g. Figs. 153 and 154), which may extend to the entire disappearance of the xylem. A sufficient basis, however, for the attainment of general results, available for more than individual phenomena, is still wanting, so that we can here only point out these manifest relations without entering into them more minutely. The structure of the two parts of the bundle, in so far as it is brought about by the juxtaposition of the different sorts of tissue, and by their peculiarities, as described in former chapters, will in the case of the Osmundacez and Marattiacee be ‘mentioned further on, under the head of concentric bundles in Ferns (Sect. 106). In the other collateral bundles, especially those of the Phanerogams when they consist of more than one or two elements,— (a) The xylem is built up of tracheze and (parenchymatous) cells. At its inner edge lie a few narrow, spirally or annularly thickened trachez, which are the first products of the differentiation of the tissues, and are thus to be called primitive elements (Protoxylem of Russow). For the reasons given at_p. 157 it is especially these primitive elements in which, in the mature plants, the spiral threads are ‘steeply wound or quite distorted, and the rings widely and often irregularly separated from one another. The primitive elements themselves are not uncommonly com- pressed by. the expansion of their neighbours, and here and there manifestly destroyed. In the Coniferz, Equiseta, and Ophioglosseze, the primitive elements are tracheides; the same may be the case in the plants or parts of plants mentioned above at p. 165 as wholly destitute of vessels. In the other cases they. are called -vessels, and in most instances no doubt rightly so, although just as in the case of the primitive traches, few very accurate investigations of this really not very essential distinction have been undertaken. Outside the primitive elements wider ‘tracheze follow, which are tracheides or vessels according to the individual cases mentioned in Chap. IV, and especially in Sect. 40%. Their development takes place successively, advancing from the inner edge of the bundle outwards, and as a rule at a time when the elongation of the entire part to which they belong is nearly at an end. The thickenings on their walls therefore have a successively denser arrange- ment: dense spiral and annular trachez, then reticulated and pitted trachezx follow 1 Compare van Tieghem, Struct. des Aroidées, /.¢. 2 Compare details in Caspary, Zc, ’ 322 PRIMARY ARRANGEMENT OF TISSUES, one another in succession from within outwards, with gradual transitions, or with the omission of one or the other intermediate form. As regards the occurrence of particular forms of thickening, it may at any rate be given as a tule that among the Monocotyledons the series ends with the development of dense fibrous thickenings, annular and spiral fibres, reticular fibres, and surfaces with non-bordered pits or scalari- form markings (pp. 158, 163). Bordered pits or scalariform slits here only occur in the collateral bundles of stems which are long-lived and relatively slow-growing, as those of many Palms, and Arundo Donax, and many rhizomes. The same often applies to the stem and leaves of such Dicotyledons as form no secondary wood, though here exceptions occur even in relatively short-lived parts, as for example in the leafy stems of Thalictrum flavum and aquilegifolium. Trachez with bordered pits appear in the stem of most Dicotyledons and Gymnosperms which form secondary wood, at the point where the latter joins on to the primary bundles (cf. Chap, XIV), and occur further in the outer part of the bundle-trunks which traverse the leaf-stalks and leaves. The elements of this part are in many cases ‘, especially among Coniferze, very similar to those of the secondary wood in the stem generally, both in form and structure; yet this minute agreement is by no means of universal occurrence. As the structure of the walls changes from within outwards, the width of the tracheze increases in comparison with that of the primitive elements, and this change may take place according to the individual cases either gradually or suddenly, either constantly or so that narrower tracheze again succeed the wider. In very small bundles, containing only one or two trachez, the latter form a group which, though varying in individual cases, does not require a further description. On the other hand, larger bundle-trunks, which consist of numerous elements, show remarkable differences both in their arrangement and in their gradually changing width.. The collateral bundle-trunks of most Dicotyledons and Gymnosperms show the trachee ranged in rows running from within outwards, which touch each other laterally, or are separated by rows of non-equivalent elements. In every row the tracheze become wider towards the outside, and in thick bundles soon attain an average size, which remains uniform in the whole region (Figs. 157, 158, 183). In the large bundles of many Dicotyledonous leaves the width of the elements first increases successively, and then diminishes again to an average size, which further towards the outside remains constant, e.g. leaves of Camellia, Ilex, Rosmarinus, species of Eucalyptus, &c. The difference between the narrowest and widest trache is in all these cases moderate, especially in comparison with many Mono- cotyledons; the largest is 2-3 times as large as the smallest, or even less. The absolute size of the tracheze is also moderate; in the leaves they are generally very narrow, falling far short of the average width of the tracheides and fibres of the secondary wood. Cf. the dimensions given in Sect. 40 and Chap. XIV. In the thicker bundles of the Monocotyledons other conditions are the rule. In most cases the trachez here form two .main rows as seen in cross-section, which diverge like the limbs of a V. At the point where the two limbs cut one another, or inside it, lie the primitive elements. The end of each limb is usually occupied by a * Compare Frank, Botan. Zeitg. 1864, pp. 167, 393: J STRUCTURE OF COLLATERAL BUNDLES, 323 trachea, exceeding the primitive elements many times in width, with dense spiral, or narrow reticular thickenings on its wall; and this either forms the end of a continuous or interrupted row of successively wider trachez, or it follows suddenly on much narrower ones. In the middle between the two limbs there are either no vessels (even the whole Phloem may be included here, e.g. the leafy stem of Asparagus i and Tamus communis), or the middle is occupied and more or less filled up by a group of narrow, densely reticulated or pitted vessels, as for example in the Grasses (Fig. £51); and this group may spread out even beyond the external edge of the large tracheze at the ends of the limbs (Fig.147). In the laterally flattened Mono- cotyledonous bundles mentioned above, the trachez lie in an interrupted single, or in places multiple row, running from within outwards. In this it is usual for one or a few narrow primitive elements to be followed on the outside by one or a few trachez of considerable width, e.g. by a very large spiral tracheide in the leaf- stalk of Musa? and Canna, &c.; further outside there are either no more trachee, e.g. in the leaf of Pandanus*, or some relatively very narrow ones, e.g. Musa, Canna, Heliconia, &c.: the broad bundles in the stem of many Palms, especially . Calamus‘, though not laterally flattened, also show the same character; in Calamus some narrow spiral vessels are followed by a single pitted vessel of enormous width (cf. p. 169), and there are no others further outside. The phenomena mentioned give rise to the characteristic habit of most Mono- -cotyledonous bundles, which is especially evident in cross-section. They occur also among those members of the class which, like the Dioscorez, behave differently to the others with respect to the arrangement of the bundles (cf. p. 276). On the one hand, however, they are not confined to Monocotyledons, for the bundles in species of Ranunculus, and especially of Thalictrum, belong to or are closely connected with the form first described for Monocotyledons, while those of Nelumbium stand in the same relation to the second form. On the other hand, their distribution is by no means universal even in those families of Monocotyledons in which they prevail. The cross-section through the bundle-trunks in the leafy stem of Fritillaria im- perialis and in the leaf of Phormium tenax ® shows, for example, a triangular group. (widening towards the outside) of moderately, wide trachez, which increase but little in size towards the outside, and show the regular character and succession in the structure of their walls. The thicker bundle-trunks in the leaf of Yucca filamentosa (from which the thinner ones only differ in the number and size of their elements) show a thick xylem, broadly triangular in cross-section, in which the primitive ele- ments are succeeded by a likewise broadly triangular group of spiral vessels with a closely-wound fibre, which are of moderate and tolerably uniform width. With this group is immediately connected on the outside a zone, consisting of about four cross- rows of narrower reticulated and pitted vessels, united at all points, which on the other side borders on the phloem. The bundles of the Equiseta, to be further described below, agree essentially ‘ Von Mohl, Z.c. Tab. G. - 2 Von Mohl, .c. Tab. G, fig. 3. 8 Von Meyen, Phytotomie, Taf. VIII. : * Von Mohl, 7.c. Tab. D, F. 5 Yon Mohl, Z.c. Tab. G. ¥ 2 324 PRIMARY ARRANGEMENT OF TISSUES. in ‘the arrangement of their trachese with the two-limbed bundles of Mono- cotyledons. The arrangement of the parenchymatous cells of the xylem results for the most part from that stated in the case of the trachez. In bundles with many rows of tracheze they form similar rows interpolated between the latter, with the form of narrow long medullary rays, round which, in a longitudinal direction, the lines of trachee run with slight undulations, alternately receding from one another and coming into contact at the ends of the rows of cells. In bundles of the Monocotyledonous type, with less regular seriate arrangement, they form single longitudinal rows, or groups of various form between the trachese. ‘The cells themselves are elongated in different degrees, with horizontal or oblique ends, their walls delicate or considerably thickened and lignified ; in the latter case their distinction from tracheides is often difficult. (5) The phloem of collateral bundles consists of sieve-tubes, and thin-walled, elongated, prismatic cells, for which Nageli’s term Cambiform-cells is to be reserved. With reference to the more special structure three cases, not all equally well known, . are to be distinguished. : 1. In the more accurately investigated Monocotyledons (e. g. the Grasses, Fig. 10, with which the Equiseta appear also to agree), and no less in very many Dicotyledons, as Ranunculaceze, Umbelliferae (Feeniculum, &c.), Vitis, Aristolochia, also Cucurbita, &c., the cross-section of the phloem shows the two kinds of meshes, which have been known since Moldenhawer, and Moh!’s Anatomy of Palms ; some wider and polygonal, which are the cross-sections of the sieve-tubes, others narrower, square or rectangular, or frequently narrow and obliquely four-sided, representing the cross-sections of the cambiform-cells, The. latter stand isolated among the sieve- tubes, distributed with varying regularity, so that as a rule each sieve-tube border’ on another with one part of its lateral walls, and with another part on a cambiform-cell. -: Traced longitudinally the cambiform-cells form. rows between the sieve-tubes, and parallel to them. Regarded individually they are as a rule shorter than, seldom as long’ as the joints of the sieve-tubes. Both from their arrangement in cross-section, and on tracing them in the longitudinal direction, it often has the appearance as if the cambiform-cells arose with the elements of the sieve-tubes from one mother-cell, the latter dividing longitudinally into a daughter-cell which becomes the sieve-tube element, and another which either becomes a cambiform-cell without further division, or is divided by cross-walls into several of them’. On. this point, however, more accurate investigations must be undertaken* The cambiform-cells have delicate non-lignified cell-walls and finely granular protoplasm, with a nucleus elongated in the longitudinal direction. On the structure of the sieve-tubes nothing need here be added to what was said in Chap. V. 2. In several, perhaps in numerous Dicotyledons (e.g. in the leaf-stalk of Olea Europea (Fig. 156), in the stem of species of Lobelia, Crassulacese, Cacteze*; and in several, especially succulent, Euphorbiz, as E. Caput Medusz), -the cross-section of the phloem shows, among wider thin-walled elements, numerous or sparsely scattered groups of much narrower cells, each group often appearing from its size 1 Compare Vochting, Melastomeen, p. 16. ? [Compare Wilhelm, 7.¢, (see p. 172).] * Compare Véchting, Rhipsalideen, Zc. Tab. 52. STRUCTURE OF COLLATERAL BUNDLES. 325 and arrangement as though it had proceeded from the longitudinal division of one of the wide elements. So far as investigations extend, the narrow elements are here sieve-tubes, which may be accompanied by narrow cambiform-cells; the wide ones are cells, whether they be called cambiform or simply parenchyma. More minute and extended investigations of these elements are wanting; they are usually curtly disposed of under the name cambiform or soft bast. In the thick bundles, often mentioned above, of Dicotyledonous leaf-stalks and ribs, the paren- chymatous rows, resembling medullary rays, are continued directly outwards from the xylem into the phloem, and in the latter pass between bands of tissue, appearing irregularly narrow-meshed in cross-section, which no doubt contain the sieve-tubes. 3. In the primary bundles of the Coniferze, especially of the leaves, where there is no disturbance from the secondary growth which quickly comes on in the stem, and also in the large bundles of the leaf of Welwitschia, the cross-section of the phloem shows regular rows of similar elements with soft membranes, which are in a very high degree capable of swelling. They are either present alone (Figs. 63, 157), or a single row consisting of relatively wide parenchymatous cells runs here and there between them, e.g. in the leaf of Dammara alba. The general form of the regularly serial elements first mentioned is elongated prismatic. How far they are sieve-tubes or cambiform-cells is undecided. The development of the elements of the phloem of collateral bundles begins at the external edge and proceeds towards the xylem, and thus in the opposite order to that in the latter, i.e. centripetally in the phloem, if the centrifugal direction is maintained in the xylem. In the middle of the bundle, on the border between the two parts, active meristematic division may still be going on when the elements of the external and internal edge are completely differentiated. ‘The outermost, first-developed elements of the phloem (Russow’s Protophloem) are often distinguished from those which follow by their smaller width, and thicker, apparently gelatinous walls ; as regards their nature, however, they are in the cases now in question, which admit of more accurate investigation, partly sieve-tubes, partly cambiform-cells. In the thicker bundles they not uncommonly become com- pressed in the radial direction, owing to the expansion of the surrounding tissue, while their walls apparently swell up to the obliteration of their lamina—a pheno- menon which ensues to a much greater extent in the old sieve-tubes and cambiform- cells of the secondary bast. (Cf. Chap. XV.) The boundary between. phloem and xylem is in most cases on the whole sharply marked, owing to the contrast between the delicate and non-lignified membranes on the one hand, and the characteristically thickened and lignified membranes on the other. The elements which form the boundary on the side of the phloem, no doubt always have the characteristics of cambiform-cells; it is not known that a sieve-tube ever borders directly on a trachea. In the thick bundles of the great majority of Dicotyledons the cells on this border-surface long remain capable of division, and when the differentiation of tissues has once begun at the periphery, their divisions take place chiefly in the tangential direction, parallel to the outer edge. Owing to the arrangement of the cells in radial or tangential rows thus determined in the zone indicated, the boundary is pretty sharply defined, even in cases where the actual divisions soon come to an end. .In the stems of Dicotyledons and Gymnosperms 326 PRIMARY ARRANGEMENT OF TISSUES. with secondary growth in thickness, this border-zone maintains its capacity for division, and becomes a portion, or a starting-point, of the cambial ring (Chap. XIV). In other bundles the cessation of the divisions takes place early in each transverse portion, simultaneously with the general differentiation of the tissues ; and according to the degree in which this happens, the radially arranged boundary layer becomes less clear. It is therefore also, especially in Monocotyledons, the less possible everywhere to fix a sharp boundary between phloem and xylem, the less numerous and crowded the lignified elements in the latter are. The word Cambiform used above was first adopted by Nageli (Beitr. I. p. 4). For the meristematic strand giving rise to a vascular bundle, and consisting of thin-walled longi- tudinally extended cells in which longitudinal division goes on for some time, Nageli uses the traditional and certainly ambiguous name Cambium, for which Sachs has re- cently substituted Procambium. The tissue of the phloem which has arisen from this cambium and passed over into the permanent condition is called by N&geli in its totality the Cambiform, i.e. cambium-like tissue, as its elements are so similar to that of the cambium, in their elongated form and thinness of wall, that former observers have actually identified them with it. Our present knowledge imposes the necessity of severing the sieve-tubes from Nageli’s Cambiform, as a distinct kind of tissue. The name there- fore, once being in existence, remains reserved for their characteristic companions, and may the better be used for them as it is for the most part literally applicable to them, even when the original meaning, indicated in Chap. XIV, and differing from that above- mentioned, is restored to the word cambium, The structure of the phloem of collateral vascular bundles of the main stems has been described in the preceding paragraphs; the description was based primarily on the numerous investigations now existing of such objects as afford a clear and certain insight, owing to the size of the elements in question. It was already mentioned under 2 and 3, that in certain cases differences occur from the type described under 1, and that doubts exist as to the structure. Besides these definitely characterised cases, there are many others, especially concerning the smaller vascular bundles, in which we know nothing more of the structure of the phloem than that it consists of thin-walled, narrow, . and elongated elements, the delicacy and smallness of which, as well as the tendency of their walls to swell, which interferes with their preparation, is deterrent to accurate , research. Where these difficulties have already been successfully overcome, the struc- ture described has always manifested itself. An uninterrupted series of transitions leads from the cases with relatively large, easily intelligible elements, to those where they are most delicate and difficult. No grounds whatever exist which would compel us to assume an essentially different structure. I therefore think that I am right in stating that the structure described exists in all vascular bundie-trunks here in question, and the more so as it is really not long since the largest sieve-tubes were first clearly recognised, as remarked at p. 182, and I do not doubt that further investigations, which are in any case a necessity, will justify what has been said. In the xylem of many collateral bundles an intercellular passage occurs, which follows the course of the whole bundle, sometimes next to or within the otherwise persistent xylem, sometimes so that, though the latter is originally formed, the tracheze become destroyed and degraded as the parts expand. In numerous Monocotyledons, the Equiseta, and some Dicotyledonous water- plants, at the inner edge of the bundle, where the primitive trachee are placed, a passage is formed by the peripheral extension of the surrounding cells, i.e. schizogene- tically (p. 200), while the external part of the xylem attains perfect development and is persistent. This severing of the original continuity of tissue goes on within. the STRUCTURE OF COLLATERAL BUNDLES. 327 wall of the primitive tracheze ; the latter are attached to the wall of the passage, and if its expansion is considerable they may be laterally removed from one another, and as the separation usually takes place before the elongation of the parts is complete, they become simultaneously distorted in the longitudinal direction, and: reduced to thickening fibres adhering to the wall of the passage. This process often attacks annular tracheze, the rings of which, in cases of great elongation, then become shifted to a long distance from one another. The width attained by the passage is various, sometimes equal to that of a moderate vessel, sometimes to the cross-section of the whole persistent part. The cells actively engaged in its formation undergo, on considerable expansion, divisions which are radial with reference to the passage, and remain as a rule thin-walled. These passages contain air, with the exception of some submerged plants to be mentioned below. All collateral bundles of the stem of Equiseta show a relatively wide passage at the inner side of the xylem, The same phenomenon occurs very widely in the leafy stems (halms) and leaves of numerous Monocotyledons, but not in their rhizomes: thus in the stems of Hydrocharis, Butomus, Sagittaria, Alisma, Juncacex, Kyris, Cyperacee, Acorus Calamus, Leucojum, and Commelinez (Tradescantia albiflora, zebrina, Lyonii!), As the names show, most but not all of these plants inhabit water or bogs. Further, the pheno- menon does not occur in all the bundles of the same part; e.g. the smaller bundles in the leaf and scape of Acorus Calamus have no passage, while the larger bundles have a very wide one; and further, nearly related plants, agreeing in their mode of life (e.g. Grasses, Cyperacex, Commelinez), often show a different character with reference to the width of the passage, and its presence or absence, Of Dicotyledons only some water- plants belong to this series, namely, the water Ranunculi and Nelumbium, besides those to be mentioned below. Both have a passage on the inner side of the larger, but not of the smaller vascular bundles of the stem. In a number of water-plants the process described extends to the complete destruction of the whole xylem. The latter originates at an early age in the form of a few annular tracheides, or of a large number arranged in a bundle, all of which become both separated from one another laterally, and torn longitudinally, as the part elongates. The inconspicuous separate rings or fragments of rings remain adhering to the wall of the passage, the phloem, which is usually strongly developed, is alone persistent. As far as the investigations extend, these passages contain water. The leaf-trace bundles of the internodes of Potamogeton natans and its allies belong to this series ;—in their cauline bundles and in the nodes the tracheze are persistent—also the bundles of the leaf-stalk and flower-stalk of species of Nympheza and Nuphar, and of Brasenia peltata*. In many bundles of the plant last- named a portion of the vessels are persistent; they thus belong to the former category ; in their rhizomes no passages occur in the vascular bundles.—On allied phenomena in the non-collateral bundles of other water-plants, comp. Sect. 110. A formation of passages in some degree different from that described occurs in the flower-stalks, leaf-stalks, and leaves of Aroideze, especially of those with unisexual flowers, as Colocasia, Caladium, Richardia*. In the xylem only a few tracheides— 1 Compare Frank, Beitr. 2c. p. 138. * Compare Caspary, Berlin. Monatsberichte, 1862, /.¢.—Trécul, Ann. Sci. Nat. 4 sér, tom, 1, p. 151. 5 Duchartre, Recherches sur la Colocase, Ann. Sci. Nat. 4 sér, XII,—Unger, Beitr. .. Physiol. 328 PRIMARY ARRANGEMENT OF TISSUES, 2-4 in each cross-section—are formed, between delicate narrow cells; they are elongated, and ranged in longitudinal rows, one above the other, like vessels. One of these rows becomes expanded, according to van Tieghem’s description apparently passively, to form a passage surrounded by narrow cells, the thickenings on its wall disappearing in places. The very oblique, fibrously thickened (unperforated ?) end- surfaces, with which the articulations are in contact one with another, are persistent ; in cross-sections, therefore, the passage often appears divided by a septum into two unequal portions. The other tracheides, lying partly inside, partly outside the dilated one, remain narrow and delicate, with annular or spiral fibrous thickening. In the scape and leaf-stalk the passages increase in width with their distance from the periphery, the outermost bundles have only a narrow row of tracheides in place of them. They accompany the bundles into the lamina of the leaf, extending into the thick strand into which the bundles are united at its apex: here they lie close side by side in great numbers, forming the often-described water canals of the leaves of Aroideze. In the leaves also they are accompanied by rows of tracheides which are not dilated, with which transverse anastomosing branchlets are everywhere connected, the petiole not excepted. In the thicker bundles of the leaf of Sparganium ramosum a wide passage is produced, according to Frank’, in the same manner as among the Aroidee. The contents of the passages consist of air and watery liquid, while among the Aroidez latex containing tannin is also present in places. Comp. p. 188. The sheaths of collateral bundles consist either of simple parenchyma, or, rarely, of the form described as endodermis; or, lastly, and indeed in the majority of col- lateral bundle-trunks, of strands of sclerenchymatous or collenchymatous fibres, which accompany the bundles, whether it be as a tube encircling the whole bundle, or as a strand which partly surrounds the circumference of the bundle. In the latter case it rarely borders exclusively or principally on the xylem, but usually on the phloem, or only on its outer edge. These sheaths and accompanying strands are‘to be separated from the bundle itself, and considered in the following chapters. Here we have only to mention that those which consist of sclerenchymatous fibres may not uncommonly border immediately on the elements of the bundle, and even insert their own elements between the latter, so that the limitation, especially as seen in cross-sections, ceases to be clear, and the arrangement of the specific parts of the vascular bundle is often influenced in a peculiar manner. A gradual transition from the sclerenchymatous elements of the sheath to the cells of the xylem very often takes place in cases where the latter are provided with strongly thickened and lignified membranes, e.g. in stout bundles of Monocotyledons. (Fig. 150.) The sclerenchymatous sheath is generally very sharply limited on the side of the phloem; the latter lies as a uniformly thin-walled mass of tissue between the sheath and the xylem. It happens in rare cases that here also the sclerenchyma of the sheath penetrates deeply into the phloem, and is continued as far as the thick-walled d. Pf, Wiener Acad. Sitzgsber. Bd. XXVIII. p. 111.—De la Rue, Botan. Zeitg. 1866, p. 816.—Van Tieghem, Struct. des Aroidées, 2. c. } Beitr. p. 137. ho ae, STRUCTURE OF COLLATERAL BUNDLES. 329 éells of the xylem. In the stem of Rhapis flabelliformis most bundles have a phloem which is somewhat crescent-shaped in cross-section; it is surrounded by a thick sheath forming a bundle of fibres, which projects strongly towards the outside; round the small xylem, consisting of a few vessels surrounded by thick-walled cells, the sheath is feebly or not at all developed. In some of the inner bundles a ridge-like projection about three layers thick passes from the sheath to the thick-walled cells of the xylem through the middle of the phloem, severing the latter into two symmetrical halves. Essentially the same phenomenon appears much more conspicuously in the stem of Calamus. The xylem, as already stated at p. 323, always shows some narrow spiral vessels, and outside these a very large pitted vessel; around and between them lie thick-walled elongated cells. In the smaller peripheral bundles the single phloem, semicircular in cross-section, lies on the outside of the pitted vessel, and is surrounded by a strong sheath of fibres, which is continued round the xylem. In most bundles the phloem is divided by a broad continuation of the sheath, extending up to the pitted vessel, into two portions lying right and left of the latter. Each of these portions consists of some large sieve-tubes (comp. p. 176) standing in a row parallel to the circumference of the vessel, together with the accompanying cambiform cells. This row also may once more be interrupted by sclerenchyma, so that individual sieve- tubes with their accompanying cells stand isolated in the sclerenchyma of the sheath*. In the mature leaf of species of Pandanus, the phloem appears at the first glance to be wholly absent even in thick bundles, and in its place a strong narrow strand of sclerenchymatous fibres seems to border immediately on the xylem’. More accurate investigation shows isolated sieve-tubes between the sclerenchymatous fibres. In younger bundles they are easier to find—perhaps also more numerous; it seems as if here, as in other bun- dles, sieve-tubes present at an early period were afterwards crushed by the neighbouring cells and rendered un- recognisable, but this has still to be investigated. As the parts belonging to and bor- dering on the vascular bundle must in fact be considered in connection with one another, it seems more ex- pedient here, in conclusion, to collect together some figures referring to collateral bundles, explaining them with reference to the paragraphs above, than to insert them among the preceding descriptions of the in- dividual parts. Fig. 149. Cross-section through a mature internode of Equisetum FIG. x49. 1 (Compare Kny, Verhandl. d. Bot. Ver. Prov, Brandenburg. Bd. XXIII, 1881, pp. 94-109; also his Bot. Wandtafeln, V.] ? Compare Meyen, /.¢. (p. 323).—Van Tieghem, Ann. Sci. Nat. 5 sér. VI. p. 197. 330 PRIMARY ARRANGEMENT OF TISSUES, palustre (145). « Endodermis, 7 axial air-canal, at x remains of the membranes of shrivelled pith-cells. In the middle a vascular bundle surrounded by parenchyma, with~ out a distinct sheath. At the inner edge of the bundle lies a wide intercellular passage, in which the letters r, ¢,s are written. ¢ an annular fragment, adhering to the wall, of the membrane of a primitive tracheide for the most part destroyed. persistent annular tracheides. g groups of the last developed, likewise permanent, annular and reticulated tracheides, distinguished from the surrounding tissue by the shading of their walls. s the phloem; in this the wider lumina belong to the sieve-tubes (cp. p, 180), the narrower ones, some of which are granularly dotted, to the cambiform-cells, The double-contoured bands on the outer edge of the phloem, inside the layer of cells following u, indicate the collapsed primitive elements of the phloem (Protophloem). Figs, 150 and 151 represent two cross-sections through a bundle of the leaf-trace of Zea Mais, at different parts of its course. Fig. 150 (550), from Sachs’s Text-book, is from FIG. 150, the stem. g-g,s,r,/, the xylem, w the phloem. In the latter v, w indicate the sieve- . tubes, between which the narrower cambiform cells stand in regular distribution. At the outer edge of the phloem are its narrower, thick-walled primitive elements. On the inner edge of the xylem in an intercellular passage, /, bounded on the outside by the ring, r, of a primitive annular vessel, partly destroyed by the longitudinal extension. ss spiral vessel. g, g large vessels with (unbordered) pits, or narrowly reticulated. Between the phloem and s, g and g a transverse group of narrow pitted vessels, A sheath composed of sclerotic lignified elements goes all round the bundle; 4, » thin-walled parenchyma outside the sheath. a@ outer edge, i inner edge of the whole bundle. In the lamina of STRUCTURE OF COLLATERAL BUNDLES. 331 the leaf and the upper part of the leaf-sheath the bundles are similar to that in Fig. 150 though on the average smaller. Fig. 151 (145), on the other hand, is taken from the place where the bundle* passes through the basal portion of the leaf-sheath of a young plant. The bundle . itself, with its group of vessels at g, is in all parts smaller than that in the preceding figure, but otherwise similar to it, as is clear without explanation, even in the finer points of structure not represented in the figure. Here however there is a single-layered sheath, consisting of deli- cate parenchymatous cells, square in cross-section, around the whole bundle. Outside this, and separated by it from the phloem, is the thick strand of collen- chyma, sc. Both the latter and the bun- dle are surrounded by the parenchyma- tous layer sz, which is rich in starch (the starch-sheath, Sect. 122). e-e Epidermis of the outer surface, Fig. 152. Cross-section through a vascular bundle of the internode of a creeping stem (runner) of Ranunculus repens (225). xx annular and spiral ves- sels at the inner edge. ¢ pitted vessels in the external region of the xylem; be- tween and around the latter delicate- walled elongated parenchymatous cells. s phloem; the larger meshes are sieve- tubes, the smaller ones, partly dotted, are cambiform cells. At the inner border of the phloem are delicate cambi- form cells in rows. Ex- ternally to its constantly thin-walled peripheral ele- ments, the bundle is sur- rounded by a thick scle- renchymatous sheath, which is only interrupted on the outer border of the xylem. Outside this is large-celled, thin-walled parenchyma. The longitudinal section through the bundle, if not quite accurately median, would be similar to that of Saururus, represented in Fig. 57, p. 157. Fig. 153 (225) represents the cross-section of one of the larger bundles in the internode of Ranunculus 332 PRIMARY ARRANGEMENT OF TISSUES, fluitans, The structure is similar to that of the preceding species, but simpler. gx tracliea with annular and loosely spiral fibres, bordering on an intercellular passage in which the letter is written. g wider trachez with densely spiral and reticulated wall. + five relatively large sieve-tubes between narrow cambiform cells. The whole bundle is surrounded by narrow elongated prismatic cells, showing non-lignified cellulose walls, and these pass over gradually on the outside into large-celled parenchyma with abundant starch, As regards the surrounding layer marked wu it remained doubtful whether it has an indi- cation of endodermal structure on those walls which are radial with reference to the bundle. : Figs. 154 and 155, from the fully elongated hypocotyledonary stem of Ricinus.communis (from Sachs’ Textbook), represent that form of bundle, with radially disposed xylem, which predominates among Dicotyledons, together with the initial stages of the secondary growth in thickness, which here follows immediately on the formation of the primary bundles. In the cross-section (Fig.154) g,¢,¢ are the rows of vessels alternating with rows of thick-walled cells, beginning at the inner edge of the bundle with the primitive elements, which are distinguished by the thick shaded walls. ¢ narrower, g wider pitted vessels, y phloem consisting of sieve-tubes, cambiform and delicate parenchyma, with three bundles of sclerenchymatous fibres on its outer border, at 4, On the boundary between xylem and phloem, the formation of the zone of cambjum and secondary growth (c-c) has begun by means of tangential divisions, and is continued, starting from the sides of the vascular bundle, over the parenchymatous zone cd, cé: cf. Chap. XIV. parenchyma of the FIG, 153. pith, r of the outer cortex. Between the layers containing the letters 4 and r is the parenchymatous sheath containing starch-grains (Starch-ring, Chap. 1X). Fig. 155 is the radial longitudinal section through a bundle of the same structure as Fig. 154. The letters r, 4, c, a indicate the same things as in the former figure. # bands of parenchyma from the phloem. s innermost and narrowest spiral vessel. s’a wider one. / scalari- form reticulated vessel. ¢ mature vessel with bordered pits; at g the perforated trans- verse wall between two elements. 7 a pitted vessel still immature, the borders of the pits not yet developed. 4, 4’ thick-walled cells of the xylem ; on the wall of ¢ and 7’ the boundaries of cells which have been cut through are visible. 4”, 4” narrow tracheides(?). 333 et % Cal (OOdalnAr' ASAT ee — (OU Os FIG 154. STRUCTURE OF COLLATERAL BUNDLES. FIG. 155. 334 PRIMARY ARRANGEMENT OF TISSUES, Fig. 156. Cross-section through the vascular bundle in the midrib of the leaf of Olea Europza (375). s-s the phloem, consisting of wide (parenchymatous ?) cells, and scattered groups of very narrow elements (sieve-tubes ?); comp. p. 325. / sclerenchymatous fibres, forming a girdle round the outer edge of the phloem, and occurring scattered on the inside of the xylem. The very dense xylem borders on the phloem internally; the primitive elements at its inner edge do not appear clearly; its larger external portion consists of radial rows of thick-walled pitted trachex, which alternate with bands of parenchyma, The latter are indicated by the granular dotting of the lumen. # paren- chyma. FIG. 3156. Fig. 157. Cross-section through a vascular bundle in the leaf of Welwitschia mirabilis (145). Anuninterrupted zone of very thick-walled and elongated sclerenchymatous fibres surrounds at / the outer edge of the phloem, and at /’ the inner edge of the xylem. Inside the zone f follows the phloem, which is crescent-shaped in cross-section, consisting of narrow, radially arranged, elongated elements; their structure could not be detected with certainty, nor are they drawn quite accurately in the figure, because the great swelling of the membranes makes it impossible to spread out the cross-section of the phloem in one plane, in such a preparation as that figured. The inner portion of the xylem enclosed by the fibrous sheath /’ consists of tolerably wide, elongated prismatic cells, connected without intercellular spaces, with thick almost gelatinously soft. mem- branes. Between them are inserted numerous, very narrow, compressed and distorted, spiral and annular tracheides, with thick, distorted, fibrous thickenings; sp; they doubt- less represent the primitive elements of the bundle. Further outside follow the persistent trachex, placed in tolerably regular rows, here and there alternating with delicate cells, and in general increasing in width in each row from within outwards; first annular and spiral trachex, with dense and very broad thickening layers, then reticulated and (at g) large pitted vessels, with bordered pits and round perforations in the cross-walls. 7, ¢ are the rows of pitted and reticulated trachez in cross-section ; they surround the bundle STRUCTURE OF COLLATERAL BUNDLES. 335 and are to be described below in Sect. 112. They also occur on the right-hand side of the figure, and may be known by their notched outline. p parenchyma of the meso- phyll, with an indication of the small crystals of Calcium oxalate imbedded in the walls. Sect. 102. The bundles of the leaves of Cycadez and Isoetes, with which perhaps those of Phylloglossum agree, are different from the typical, simply collateral bundles in the arrangement and development of their parts. FIG, 157. The peculiarity of these bundles, expressed generally, consists in the fact that the essential elements of the xylem are reversed in position and order of develop- ment, as compared with those in the typical cases. In the plants first named there are the additional peculiarities that tracheides appear at a later stage on the border of the phloem, and that the bundles are in many cases united in pairs. The bundles of the leaf-trace of the Cycadeze begin in the stem, according to Met- tenius?, as simple collateral bundles, and run out in the same form into the sheathing bases of the leaves, but before entering the petiole of the foliage-leaves, or the apex of 1 Abhandl, d. K, Sachs, Gesellsch. d. Wissensch. VII. p. 573. 336 PRIMARY ARRANGEMENT OF TISSUES, the scale-leaves, they assume a changed structure, which they maintain throughout ‘their whole course in the leaf. In the round cross-section of the bundle (Figs. 158,159), a small group of narrow spiral tracheides (sf) (the primitive elements of the xylem) occupies about the centre. From this an uninterrupted group of large prismatic pitted tracheides (the inner portion of the xylem) extends towards the inside; this group occupies the entire inner side of the bundle, and has in cross-section the form ofa FIG. 158.—Cycas revoluta. Petiole of a small foliage-leaf belonging to a young plant; vascular bundle, cross-section (223). Explanation in the text. The spiral tracheide sf is connected with the inner pitted tracheides ¢ by means of a group of annular and reticulated tracheides. Inc here and there fragments of the large crystals of Calcium oxalate. sector of a circle with the centre at the primitive tracheides. The remaining part of the bundle, lying outside that described, is principally formed of thin-walled elements, ranged in radial rows: next the outer edge are several concentric rows of sieve-tubes (s), separated by delicate parenchyma; the outermost row which bounds the. bundle, being in the mature condition compressed in the manner often described, and thicker- walled than the rest, bordering the outer edge as a narrow shining band (£); on the boundary of the xylem, so far as present investigations extend, only prismatic cells ‘occur, without sieve-tubes. To these parts is finally added. an external portion of the xylem, developed in centrifugal order, outside the primitive tracheides ; this forms a small group of pitted tracheides (2) ranged in irregular radial rows, which are separated from one another, from the primitive tracheides, and from the inner portion of the xylem, by thin-walled elements. As the bundles become thinner in STRUCTURE OF COLLATERAL BUNDLES, 337 their course through the leaf, the thickness of the individual portions diminishes, especially the external portion of the xylem. In many species, at any rate, the bundles in the cross-section of the leaf-stalk are generally arranged in the figure of an inverted 9, with the limbs directed towards the upper side. In the constricted part of the figure either they are separate, the constriction being open; or they approach one another in pairs, the inner vascular portions of each pair being turned towards one another, and the scalariform vessels of each in uninterrupted connection over a broad surface. In Zamia longifolia one such coupled bundle is present near the middle of the cross-section; in Dion there are about six of them forming a row vertical to the surface of the leaf’. Other species of Zamia show, according to Mettenius, a less regular grouping of the bundles and of their connections. In Cycas revoluta the bundle is encircled by a thick-walled sclerotic sheath, fn Sp a pate Li 0 [pi Uy] LB '=q q UR Hi ) OOO 0 OO 6) Ke FIG. 159.—Cycas revoluta (225). Longitudinal section through a similar bundle ot the same leaf-stalk, in the direction y—7, Fig. 158. The letters have the same meaning as in that figure, 2 reticulated tracheide with narrow slits ; the fibres of the narrowest (first) spiral tracheide are distorted. sharply limited both inside and out, which consists of short, wide, angular elements (cc), with pitted or narrowly reticulated walls, and cavities often filled up by large crystals of Calcium oxalate. In the other species investigated, there is no sheath which can be sharply distinguished from the surrounding parenchyma; at most some of the sclerenchymatous fibres, which are scattered through the whole tissue, stand around the periphery of the bundle. , : According to Russow?, the feeble leaf-bundles of Isoetes closely approach those of the Cycadez as regards structure. They are collateral and their orientation is normal. Their xylem consists of narrow prismatic parenchymatous cells, with some narrow, spiral, and reticulated tracheides standing between them. The primitive elements appear, according to Russow, on the boundary of the phloem, while the others, which are nearer the inner edge of the bundle, attain their development later. On the inside of the phloem thin-walled prismatic elements can be distinguished, but 1 Mettenius, /7.c. p. 573, Taf. 1, fig. 10. 2 Lc. pp. 140, 155. 338 PRIMARY ARRANGEMENT OF TISSUES. no evident sieve-tubes ; on its outer border are thick-walled cells, which latter in the terrestrial species assume the characteristics of tough fibrous cells. On the border of the phloem and xylem there lies in most species in the middle of the bundle a single intercellular canal, while in I. Engelmanni there are usually three of them; their origin is not clear’. The radial walls of the layer of cells bordering these canals have in I. Engelmanni (Russow) and I. Durieui the characteristics of the radial walls of the endodermis. No such structure is present at the periphery of the bundle. In connection with Isoetes, Phylloglossum may be mentioned, as the short description by Mettenius? shows at least this one point of agreement, that its vascular bundles only contain a few delicate tracheides, with annular thickenings, or spiral fibres which can be unwound. The phloem is at any rate very inconspicuous; according to Mettenius it is even in many cases wholly absent. Sect. 103. The form of collateral bundles above designated as the doudle or bicollateral is distinguished from the simple collateral form by having two groups of phloem, one being situated as in the latter on the outside of the xylem, and a second on its inner side. In all other respects they agree with the simple form’. As the type of this form of bundle are to be mentioned in the first instance all leaf-trace bundles of the Cucurbitaceze, and in fact of all species investigated‘. Both groups of phloem have the typical structure described (p. 324, 1), and are especially remarkable for the size of their sieve-tubes (chap. V). They are frequently connected by means of a narrow band, fringing the lateral edge of the bundle, and containing some sieve-tubes, so that in these cases the bundle, strictly speaking, belongs to the concentric type. The xylem is constructed altogether-on the collateral type; on the inside are narrow annular and spiral vessels ; towards the outside are reticulated vessels becoming gradually wider, and finally very large pitted vessels with short articulations. The latter are surrounded by broad layers of cells, some of which are elongated, with thick pitted walls, while others are short elements with undulated surfaces fitting into one another, and round-meshed reticular thickenings on their walls. It remains to be investigated whether, or how far these elements should be called tracheides. A bicollateral structure is presented by the leaf-trace bundles of many Dicoty- ledons, which belong to the ring: Melastomacee ®, Cichoriacez, Solanacee, Ascle- Piadeze, and Apocynez ; Strychnos, and Daphne. In many of these the inner phloem is so widely separated from the rest of the bundle, that it may be regarded as a distinct strand of sieve-tubes ; in other cases distinct strands of sieve-tubes occur side by side with the inner groups er phloem of bicollateral bundles, e.g. Cichoriacez, Sola- num tuberosum, and dulcamara; comp. p. 231. Of the Myrtaceze already mentioned at P- 231, Eucalyptus globulus decidedly belongs to this series. All the investigated species of Eucalyptus, Metrosideros, Callistemon, Melaleuca, and Myrtus, have on the inner side of the primary bundles a group of tissue consisting of delicate narrow elements, and, according to what has been found in the case of Eucalyptus globulus, it is very probable that these groups have the same nature as those in the latter * Compare A. Braun, Isoetes-Arten d, Insel Sardinien, Monatsbericht, d, Berlin. Acad. 1863. * Botan, Zeitg. 1867, p. 99. 8 (Petersen, Ueber das Auftreten bicollateraler Gefassbiindel in verschiedenen Pflanzenfamilien. Engler, Bot. Jahrb. 1882, p. 359-] * Compare Dippel, Mikroskop, p. 225; Bryonia, 5 Vochting, Jc. STRUCTURE OF CONCENTRIC BUNDLES, 339 plant. No accurate investigations of them however lie before us, and the example of Welwitschia (p. 335) shows that one must be cautious in coming to a decision on apparently bicollateral bundles. On the behaviour of Trapa, which here remains to be mentioned, comp. sect. 105. 2. Concentric Bundles, Secr. 104. In concéntric bundles one of the two parts occupies the middle, and is encircled by the other. Of the two cases here possible, the one, namely that in which the phloem occu- pies the middle and is surrounded by the xylem, occurs in the lower ends of the leaf- trace bundles of many, but not of all rhizomes of Monocotyledons, where they lie at the periphery of the bundle-cylinder in the stem, e. g. Iris germanica, Cyperus aureus, Papyrus', Carex arenaria* (but not, for example, C. disticha and C. hirta), Acorus Calamus and A. gramineus*®. This form arises no doubt from the collateral bundle, as in its course the xylem gradually surrounds the phloem more and more on both sides, until the latter is completely enclosed ; where it does occur however this form must be distinguished from the typically collateral. The structure and the surround- ing tissues present no generally valid differences from collateral bundles. The phloem, which is round as seen in cross-section, is as a rule surrounded by a single, or rarely by a multiple ring of reticulated or pitted vessels, with parenchymatous cells interspersed between them. Comp. fig. 148, p. 317. Scr. 10g. The other possible case, that the xylem occupies the middle and is surrounded on its whole surface by the phloem, occurs in individual Dicotyledons with anomalous distribution of the bundles, also in isolated cases among the Cycadeze, and is characteristic of the entire group of Ferns, with a few exceptions, partly mentioned above. Among Dicotyledons the medullary and cortical bundles of the Melastomacez* may first be mentioned. In these the centre is occupied by a few narrow vessels, which are scattered among delicate prismatic cells, the vascular group being sur- rounded by a delicate ring, consisting of sieve-tubes and cambiform cells, In feeble bundles only one single, narrow, spiral vessel often occurs, and even this may be absent, so that we then have the sieve-tube bundles mentioned at p. 231. According to Reinke’s description, all the bundles of the stem of the species of Gunnera, especially G. scabra, also belong to this series ; so also do those in the stem of Auriculas (cf. p. 251). In the leaf of the plants last-named the bundles are collateral, and are arranged in a row as usual among Dicotyledons. The collateral structure also holds good for the smaller bundles of the stem, even when almost circular in cross-section; on the one side is a small group of narrow, primitive, spiral trachez, with larger reticulated vessels external to them; on the other side is the small phloem, the whole being surrounded by delicate cells, bounded externally by the endodermis. On the other hand, the larger bundles of the stem of Pr. auricula show a concentric arrangement; the narrow primitive elements are in the middle, 1 Link, Icones Anatomicz, Tab. V. figs.1, 9; IX. fig. 6. 2 Treviranus, Physiol. I. p. 195, Taf. III. 8 Van Tieghem, /.¢, * Sanio, Botan, Zeitg. 1865, p. 179.—Vochting, Melastomeen, /. ¢, . Z2 340 PRIMARY ARRANGEMENT OF TISSUES, surrounded successively by the wider vessels, the phloem, and the endodermis. It is manifest that this structure may come about owing to the frequent junctions of the smaller collateral bundles. The above-mentioned isolated occurrence of concentric arrangement in the Cycadez was found by me in some of the small bundles in the petiole of Dion; they had a round xylem, surrounded by a phloem with its elements in radial rows. Lastly, the axial bundle (described at p. 277) in the internodes of several Dicoty- ledonous water-plants must be placed here: namely, Hippuris, Trapa (?), Callitriche, Bulliarda, Elatine?, Hottonia, and Myriophyllum. It consists in general of a central xylem, completely surrounded by a phloem, both parts usually having abundant delicate parenchyma between the essential elements. In the cases of Hippuris, Trapa, Hottonia, and Elatine Alsinastrum, the persistent vessels are arranged in an interrupted ring, which surrounds a relatively thick cylinder of parenchyma (‘pith’), The phloem is bounded on the outside by an endodermis. In the leaves of these plants, with the exception perhaps of Callitriche, the bundles are collateral, with normal orientation. The axial bundle of Verhuellia (p. 278) also seems to belong to this series, though Schmitz could only detect that it consists ‘of a strand of prosenchymatous cells, in the middle of which runs a single spiral-vessel.’ All the cases here cited, with the exception of the three elucidated by Sanio, Véchting, and Hegelmaier, require still more accurate investigation. In Bulliarda aquatica, according to Caspary’s description, the middle of the stem is occupied by a thin cylindrical strand consisting chiefly of elongated cells, in which, about midway between periphery and centre, lie two indistinctly separated groups of annular and spiral vessels, which run out to the leaves. In Elatine Alsinastrum the axial cylindrical strand consists permanently, as regards its main bulk, of much elongated cells; a few cell- layers inside its periphery, one vessel for each leaf of the whorl next above it first appears, and in the node this vessel bends out into the leaf at right angles; or, to express the fact differently, the vascular elements running into the leaf here abut on the cauline vessels. The vessel itself appears to be the continuation of the one which has passed out at the next lower node. Later on other isolated wider vessels are formed side by side with the original ones; in cross-section all are arranged in an irregular ring. Sieve-tubes lie in the zone outside the vessels. Axial vessels are not present at the beginning; after those of the leaf-trace however 1-2 vessels appear and are permanent, Hottonia appears (judging from very incomplete investigation) to behave similarly, apart from the obvious differences due to difference of Phyllotaxis, and with the further distinction that axial vessels do not occur. The often investigated axial bundle in the stem of Hippuris* shows in its early stages, as first accurately described by Sanio, annular and spiral trachez at its centre, which are scattered among thin-walled prismatic cells, and are cauline, with acropetal growth. At a later stage vessels are formed at the periphery of the cylinder, and from these the bundles branch off, which pass transversely through the cortex into the leaves. They are connected with one another in the node, and in the cross-section of the internode they represent an irregular, many-layered and often-interrupted ring, in which the vessels increase in width in the centrifugal direction, according to their order of origin. Outside the vascular ring lies a many-layered ring of prismatic cells, and between the * (F. Miiller, Struktur einiger Arten von Elatine ; Flora, 1877.] * Von Mohl, Verm. Schr.; Palm, Structura, Tab. g, fig. 2.—-Nageli, Beitr. /.c. p. 56.—Sanio, Botan. Zeitg. 1865, p. 191. : STRUCTURE OF CONCENTRIC BUNDLES, 341 latter are small bundles of sieve-tube, each consisting of one, or rarely two sieve-tubes! surrounded by a layer of cambiform-cells.—Both vessels and sieve-tubes run vertically and separately in the internode. In the node the equivalent elements anastomose with each other, and the two kinds of elements unite to form the vascular bundle, and enter the leaf. The peripheral elements of the axial strand which we have described are persistent. The cauline axial tracheal elements begin to disappear on the first appear- ance of the peripheral vessels, and become so crushed by the surrounding prismatic cells that a mature bundle encloses a dense parenchymatous ‘pith’ within the peripheral vascular ring. _ One bundle runs out into each of the leaves, which form multifoliate whorls, and usually each has a separate course from the rest ; yet it not uncommonly happens that a common trunk arises from the vascular cylinder of the stem, and then, towards the periphery of the stem, divides into two or even three leaf-strands. In the stem of Callitriche? the thin axial bundle, consisting chiefly of delicate elongated prismatic cells, contains at its apex an axial annular or spiral vessel, which grows acro- petally, and projects far above the last node which contains vessels; close by this a second (and a third) soon appear. The first two primordial vessels have a position in the internode corresponding to the two opposite leaves of the adjoining node; in the node a small bundle branches off from them for each leaf. As the internodes become elongated 2-12 wider annular or reticulated vessels appear, by the side of and somewhat external to the primordial elements; they are arranged in two irregular groups, and are persistent, while the primordial vessels soon disappear in the internode, and are replaced by an axial intercellular passage, to the wall of which their remnants adhere. This pas- sage may subsequently be filled up again by the luxuriant growth of the neighbouring cells, In the node the axial passage between the vascular elements, which are here densely crowded, is absent. The vascular group is surrounded by a small zone of phloem, consisting of a few rows of narrow elements, and bounded on the outside by the éndodermis, In Trapa the wide central portion of the axial strand consists, in the fully elongated internode, of loose, large-celled parenchyma, traversed by numerous longitudinal air- passages ; a relatively narrow, peripheral, annular zone consists of thin-walled prismatic cells. Among the latter are large vessels arranged in a circle at wide intervals. Their primitive elements appear to run out into the leaves, but to be distorted and indistinct, each being in many cases replaced by an air-passage when the internode has attained its definitive elongation. The large persistent annular vessels are apparently of Jater origin. Outside and inside the vascular circle Sanio® found a circle of scattered sieve-tube- bundles, each of these consisting of one sieve-tube with horizontal cross-walls, surrounded by a layer of cambiform tissue. Myriophyllum spicatum* has in the young internode, when elongation begins, in the middle of the axial cylinder a central group consisting first of one and then of 2-4 spiral vessels, which are in close contact with each other. This group is cauline and grows acropetally, and in the node branches grow out from it centrifugally into the leaves, which are ranged in alternating, usually quadrifoliate whorls. In the leaf the vessels are united with a small phloem to form a collateral bundle. From the node the four phloem- bundles of the whorl—which require further histological investigation—run down, as a 1 The sieve-tube nature of these elements is disputed by Russow, who however includes them in his protophloem. 2 Nageli, 7.c—Hegelmaier, Monogr. d. Gattg. Callitriche—Idem in Martius, Flora Brasiliensis, fase. 67. 3 Botan. Zeitg. 1865, p. 193. * Véchting, Zur Histologie und Entwickelungsgesch. v. Myriophyllum, Nova Acta Leop. XXXVI. 1872. : 342 : PRIMARY ARRANGEMENT OF TISSUES. ' Jeaf-trace, radially and tangentially vertical, in the periphery of the axial cylinder. Each passes through two internodes, and at the third node, above one of the leaf-bundles which pass out here, splits into two short, strongly-diverging limbs, each of which attaches itself to the nearest of the bundles coming down from the second node,—The main mass of. the axial cylinder consists permanently of thin-walled prismatic cells. When elongation is complete, the axial bundle of spiral vessels disappears, while round its circumference — larger thick, and usually reticulated, vessels appear, which are scattered and arranged in irregular rings. Also the number of the supposed sieve-tubes at the periphery increases with age, so that the original arrangement may become indistinct. Sxcr. 106, The vascular bundles in the stem and leaves of the Ferns* belonging to the divisions Polypodiaceae, Cyatheacez, Hymenophyllacee, Gleicheniaces, Schizseacee%, and Marattiaceze, to which the Selaginelle are to be added, are of various size and form; in cross-section they may be circular, elliptically trapezoidal, band-shaped or plate-like ; the outline of the broad ones being even, wavy, folded in a furrow-like manner, or with the edges bent in; others are annular or tubular (e.g. stem of Marsiliaceze, Microlepia, Dennstedtia, &c., see p. 284), or forming peculiar symmetrical figures resembling an X, V, U, ©, &c.: the bundles of the leaves may be similar to those of the stem to which they belong, or very different from them. Comp. Figs. 128-141. Their structure is as uniform as it is distinct from that of most other forms of bundle. Comp. Figs. 160, 16. The middle is occupied by the xylem, the form of which is either identical with that of the entire bundle, or similar to it, or in various degrees different ; the former for example is the case in the annular or band-shaped bundles, and also in the approximately cylindrical ones of the stem; while the latter condition occurs especially in leaf-stalks, in such a manner that the symmetrical figures- mentioned are peculiar to or especially marked in ‘the xylem, the outline of the whole being simpler. The former may even be severed into two symmetrical groups in one bundle, as for example in the leaf-stalk of Aspidium molle, Polypodium phymatodes. The xylem consists, as regards its main mass, of wide, long, prismatic to spindle- shaped, scalariform tracheides with bordered pits (comp. p. 165), and only in rare cases of scalariform vessels, with septa perforated in a scalariform manner (Pteris aquilina, p. 162). Between, or more rarely on the outside of these, lie, at definite points, some narrow spiral and narrow scalariform tracheides, which are the primitive elements at the origin of the xylem; from these the development of the wide tracheides starts, and advances centrifugally with reference to each point of departure, though, it may be, centripetally with reference to the whole bundle. The position and number of these primitive groups are different in the individual cases. In bundles presenting an angular or unilaterally elongated cross-section, one such group lies at or near each corner, or at each end of the greater diameter of the section, as in the flat bundles in the stem of most Selaginellas (comp. Fig. 131, p. 282), * Von Mohl, Structiira filic. arborearum, /.c.—Link, Icones selects, Heft. III. und 1V.—Met- tenius, Angiopteris, /.c—Karsten, Vegetationsorgane der Palmen, Zc. pp. 117, 130, &c.—Dippel, Verhandl, der Naturforscher-Versammlung zu Giessen (compare p. 180), and Mikroskop, p. 198 ff— Trécul, Sur la position des Trachées dans les Fougéres, &c.; Ann. Sci. Nat 5 sér. tom. X. p. 344, tom. XII. p, 219 ff—Russow, Vergl. Untersuchungen. —With reference to the form of the bundles, compare also Presl and Reichardt in the works cited at p. 298. * [See Prantl, Unters. z, Morph, d. Gefiisskrypt, Heft. IT. Leipzig, 1881.] STRUCTURE OF CONCENTRIC BUNDLES, 343 where they lie in the corners themselves, and are continued into the leaf-bundles, which join on here (comp. p. 282); and in the band-shaped or symmetrically many- rayed bundles of the leaf-stalk of Ferns'. In bundles which are elliptical in cross- section, e.g. in the rhizome of Pteris aquilina, their position corresponds approxi- mately to the foci of the ellipse. Besides these primitive groups occupying the ends and corners, others may be present in the same xylem, e.g. in the band-shaped, symmetrically curved bundles from leaf-stalks, represented by Russow; a median group occurs in Gleichenia vulcanica, Aneimia Phyllitidis, Marsilia Drummondii; in Asplenium Filix-femina there is a median group and an intermediate one in each side between this and those at the margin; in Balantium Culcita there are two inter- mediate groups, and so on. In roundish or circular and in annular xylems several primitive groups are scattered about the cross-section ; five for example in the annular FIG. 160.—Polypodium vulgare ; rhizome; cross-section through a weak vascular bundle (225). _s phloem region, sieve-tube structure not clear; sf narrow spiral tracheides of the xylem, the wider elements which constitute the Majority are scalariform tracheides ; « endodermis apparently derived by tangential division from the same layer of mother-cells. with the parenchymatous layer adjoining it on the inside. Outside # is parenchyma. The pitting on its cell-walls is essentially the same everywhere ;. in the figure it has only been indicated at certain points. Those of its cells which border on w are thicker-walled on the inside than elsewhere, bundles of the stem of Marsilia Drummondii ; six, according to Russow’s description, in the vascular cylinder of Trichomanes radicans, three near the middle in the round axial bundle in the small stem of Selaginella spirulosa, In very small bundles there is often only a single primitive group present, which is more or less eccentrically placed, e.g. small bundles in the rhizome of Pteris aquilina, Angiopteris (Mettenius, Zc. 517). In the large flat bundles in the stem of Cyatheacez the primitive groups have only recently been discovered by Trécul. Here they occur in the form of narrow reticulated tracheides at the edges bordering the leaf-gap, enclosed among the scalariform tracheides, or in a narrow notch between them; and from here they proceed, or send branches into the bundles of the leaf-stalk. In consequence of early 2 CE. Russow, ‘4c, Taf. X; especially abundant details, in Trécul, Zc. 344 PRIMARY ARRANGEMENT OF TISSUES, compression and distortion they are only to be detected in the mature stem with great difficulty, and often there are only traces of them. The xylem is either composed of tracheides. only, without any non-equivalent elements interposed between them, or there are groups and rows of parenchymatous cells containing small starch grains, intermixed with them?. The two conditions are distributed according to species and perhaps genera, not according to the form of the bundle. The first, for example, occurs in Marsilia and Pilularia, where the xylem is an uninterrupted, one to three-layered ring of tracheides ; also in the axial bundles of the Selaginellas, and in many flat, round, and angular bundles of Polypodiacez, e.g. in the stems of Polypodium vulgare (Fig. 160), P. Lingua, Davallia pyxidata ; the petioles of Asplenium auritum, Scolopendrium vulgare, and many others (cf. Russow, /.c.). In the bundles of the Marattiaceze also, no parenchymatous cells, or extremely few of them, occur among the tracheides. The other case occurs, for example, in the relatively thick cylindrical xylem of the rhizomes of Trichomanes radi- ate As VOX) cans, Gleichenia, and Lygodium, in GN PACE O% Is the annular bundle in the rhizomes a BC A 6o2) ay of Microlepia and Dennstezedtia, in the <1 RS 89¢ STN round or flat bundles of the stems of Pteris aquilina (Fig. 161), Poly- podium fraxinifolium, Platycerium alcicorne, Alsophila microphylla, Cy- athea Imrayana, and arborea ; in the bundles of the leaf-stalks of Tricho- manes, Aspidium Filix mas, molle, Lygodium and many others (comp. Russow, /.¢.). The axial bundle in the stem of the Schizzas is also placed in this category by Russow, : and no doubt rightly, for although a many-rowed uninterrupted ring of tracheides appears to surround a thick axial pith-cylinder of paren- chyma, yet it is not separated from the latter in the manner to be de- FIG, 161.—Pteris aquilina. A quarter of the cross-section through scribed below, which is characteristic a large vascular bundle of the stem; cf. Fig. 143, p. 295. S spiral tracheide, g—g wide scalariform vegsels (cf. p. 162), sf sieye-tubes, ¢ the of all other bundles of Ferns belonging protophloem of Russow, sg endodermis, # the parenchyma surrounding it, containing starch-grains; K thickened portions of the wall of the here but the two are in immediate ’ vessels between the rows of scalariform pits. Between 4 and sg, and ithe poten, expel rung S.2'¢ uumggtet pererchymatour contact, so that the strand in ques- tion must naturally be regarded as a xylem with a coherent axial cylinder of parenchyma. In the leaf-stalk of Tricho- manes pinnatum and elegans?, and in that of species of Aneimia, Gleichenia and Schizzea, very thick-walled, lignified, sclerenchymatous fibrous cells are added to the trachez ; there is a thick bundle of them in each corner of the V which the xylem * Russow's ‘companion cells’ (Geleitzellen), ? Mettenius, Die Hymenophyllaceen, p. 421. STRUCTURE OF CONCENTRIC BUNDLES, 345 forms in Trichomanes, and of the approximate T which it forms in Schizeea pectinata ; in Gleichenia dichotoma and polypodioides they lie isolated and often separated from the tracheides by parenchymatous cells along the edges of the V-shaped xylem. The xylem is surrounded in all cases, in annular bundles both within and without, by a many-layered complexus of tissue, which is to be regarded as the phloem, Fig. 161. One or a few layers of parenchymatous cells, containing starch, and similar to those of the xylem, border immediately on the latter: Outside the parenchymatous layer comes an annular zone, which contains the sieve-tubes, though in the smaller bundles, as already mentioned at p. 181, the latter are certainly not always to be clearly dis- tinguished. When distinctly developed they form a usually single, but in some places double annular ring, and are in contact with each other by means of those walls which are radial with reference to the centre of the bundle. This zone is then followed all round on the outside by a likewise annular zone of elongated fibre-like elements with narrow lumina, characterised by thick, brilliant, and soft walls. These are called by Dippel bast-fibres, and by Russow protophloem, because they appear as the primitive elements of the phloem, and here also it remains doubtful whether they are to be reckoned among the sieve-tubes, or regarded as special organs. They are partly in immediate contact with the indubitable sieve-tubes, and even frequently in- serted in the same circle, while in other cases they are separated from them by small parenchymatous cells. Finally, a one or few-layered sheath of parenchymatous cells containing starch, which are often tolerably wide, and always differ from those outside the bundle in their form and their (smaller) size, completely surrounds the zone of sieve-tubes and fibres, and, apart from a few exceptional cases to be mentioned below, this is in its turn enclosed by a single-layered endodermis, which limits the bundle sharply on the outside. This consists of prismatic, usually inconspicuous cells greatly flattened from without inwards, with a moderately thick, usually brownish membrane, which soon becomes cuticularised, and which in the radial walls is easily torn across, so that in sections the whole endodermal sheath is often split, and difficult to recog- nise. In especially favourable cases (e.g. species of Polypodium) every cell of the endodermis stands exactly in front of a cell of the parenchymatous layer adjoining it on the inside, so that the common origin of the two from one layer of mother-cells is recognised at once. Even where the latter is not the case, the origin of the two is the same, at least among the true Filices and Marsiliacez'. Among the plants belonging to this series the Marattiaceee and Selaginelle are destitute of an endodermis?. The bundles of the former appear simply inserted in the parenchyma, and this applies both to the petiole and to the stem. So at least I found it in young stems of Angiopteris, and I can only suppose that the figure cited by Russow from De Vriese and Harting’s Monogr. des Marattiacées (Taf. VII. Fig. 3, 4), according to which the case would be different in the stem of Angiopteris, represents the section of a root passing through the stem, for in the root the endodermis is always present 8. In the Selaginella the phloem is surrounded by a dense layer of small-celled parenchyma. 1 Russow, Z.¢. p. 195. 2 (Compare Treub, Recherches sur les Organes de la Vég. du Selaginella Martensii, Leide, 1877.] ® Compare Sachs, Textbook, 2nd Eng. ed. p. 420, supplementary remarks. i 4 ! t 2 3 £ 346 PRIMARY ARRANGEMENT OF TISSUES. In petioles, when the xylem is concave or notched, strands of cells 3-4 rows thick; which are distinguished from the rest of the parenchyma by their very wide lumen, are to be found in its depressions and furrows, sometimes just in front of the primitive tracheides ; ‘in longitudinal sections they are conspicuous from the fact that their walls are irregularly bent in and out, and are connected with those of the neighbour- ing cells in such a manner that large cavities or intercellular spaces arise; in old bundles it is usually found that the walls have become brown.’ Russow calls them cavity-parenchyma (Liickenparenchym). The tracheides bordering on them usually have very irregularly developed spiral bands. Examples: species of Asplenium, Cyathea microlepis (Dippel), /¢., Aspl. Filix femina, Cyatheaceze}, e.g. Cyathea medullaris. The wide cells in Osmunda regalis to be mentioned below may also belong here. From its general distribution among the Fern group in the widest sense, one may term the structure of the bundle just described the Fern-type. At the same time different degrees of deviation from the type occur within this group. Those de- scribed in the Marattiaceze and Selaginelle are trifling. The Lycopodiaceze, of which . we shall treat in the next section, are closely connected with the Selaginellz in the structure of the bundle, as well as in other points. In this respect the Equiseta are most widely different from the Fern-type ; their strictly collateral bundles, which most closely resemble those of Monocotyledons, were described above (p. 329). Besides these, the Ophioglossez, and in part at least the Osmundacez, have colla- teral bundles. The two parts have normal orientation in the round or flat bundles, the xylem is similar to that of Ferns, with some narrow spiral tracheides (primitive elements) at its inner edge; its’main bulk consists of large prismatic tracheides, which in Ophioglossum (pedunculosum and vulgatum) show narrowly scalariform reticulate thickening without pit-borders, while in Botrychium they have very thick reticulate fibres, with elliptical bordered pits in the meshes of the reticulations. Parenchyma is present in the xylem of the annular bundle in the stem of Botrychium rutzfolium, in the form of radial bands resembling medullary rays; I could not find any in the examples of B. Lunaria, which I investigated. The phloem appears very similar to that of the typical form; the wide elements presumed to be sieve- tubes still require more exact investigation (p. 180). The bundles of the petiole, and the small bundles of the stem of Ophioglossum, which in cross-section are arranged in a circle, are not bounded externally by any kind of distinct sheath. The bundle of the stem of Botrychium Lunaria and rutzfolium, which is annular in cross-section, is encircled on the outside by an endodermis, the cells of which do not differ from those of the surrounding parenchyma, except in the exquisite undulating longitudinal bands in the middle of their radial side-walls. In Osmunda (compare p. 280) the bundles of the stem are collateral. The xylem, where it enters the circle of bundles, is horseshoe-like in cross-section, and during its descending course becomes narrowed to a wedge-like form; internally it borders directly on the parenchyma of the pith; it has the same structure as in the typical Ferns, and has hardly any parenchyma inserted among the scalariform tracheides. The groups of xylem are separated from one another in the whole longitudinal 1 Russow, /.¢,. STRUCTURE OF CONCENTRIC BUNDLES, 347 course of the bundles, by medullary rays 6-10 cell-layers in breadth. Round this ring of separate groups of xylem runs a common annular phloem- region, which is similarly constructed to that of the typical Fern-bundle: outside each group of xylem are first some layers of small-celled parenchyma, then an almost uninterrupted zone of large sieve-tubes running round the whole stem ; this zone is usually one layer thick outside the xylem-groups, while in front of the medullary rays it is many-layered, and projects into them like a wedge. The layer of sieve-tubes is next immediately bounded on the outside by a layer of transversely elongated, partly thick-walled elements, which in their turn are separated from the brown sclerotic tissue of the stem by a many-layered zone of parenchyma. Outside the transversely elongated zone runs an endodermis, which in the mature condition can be recognised by the brittleness of its radial walls. In the petiole of Osmunda the runnel-shaped xylem is surrounded by a similarly-formed zone resembling the phloem of typical Fern-bundles, which in the mature state is bounded on the outside by a very indistinct endodermis; this zone however, as also stated by Dippel, only contains sieve-tubes in its broader convex half. On the con- cave side it is parenchymatous, and distinguished in cross-section by 10-12 small groups of conspicuously wide cells, which still need investigation’, In the stem of Todea Africana and T. hymenophylloides the structure of the vascular bundle is like that described for Osmunda, only the form of the xylem is in some degree dif- ferent in consequence of the fusions of laterally adjoining bundles. In the lowest part of the leaf-bundle, which has the same shape as in Osmunda, sieve-tubes are, in T. Africana at least, present on the concave side as well. In the leaf-stalk of T. Africana I found the endodermis scarcely recognisable, while in T. hymenophyl- loides it is very clear. The axial strand, which the collateral bundles in the stem of Isoetes unite to form, consists of a roundly angular mass of short and irregularly spindle-shaped, reticulated and spiral tracheides, and of thin-walled parenchymatous cells irregularly distributed between them, these elements together forming the xylem. The latter is completely surrounded by a transparent mantle of shortly-prismatic or tabular cells, with contents clear like water, and a strongly refractive membrane, which is provided with broad and very delicate pits, but no clear sieve-pores. Russow is no doubt right in considering this mantle as a peculiarly imperfect phloem of the axial strand, especially as the equivalent parts of the leaf-bundles pass over into it directly. With reference to its phenomena of growth it will have to be further spoken of in Chap. XVIII. This may probably be the most fitting place to mention the axial bundle, which tra- verses--longitudinally the leafless stolons of Nephrolepis tuberosa, N. acuminata, and N.exaltata* In the structure, form, and centripetal development of its xylem this agrees completely with the 5-6 rayed radial bundles in the roots of Ferns to be described below. Here also, as in the latter, phloem-groups alternate with these rays, and appear to con- ’ tain relatively wide sieve-tubes, but I am doubtful whether the narrow primitive elements of the phloem do not also completely surround the rays of the xylem. At any rate the ~ : whole inner part of the bundle is surrounded by about two layers of very narrow elements, and the latter usually by two layers of wider parenchymatous cells, on which 1 Compare Dippel, Russow, é.¢. 2 Trécul, 2. ¢.—Russow, Zc. p, 100. 348 PRIMARY ARRANGEMENT OF TISSUES. the endodermis borders externally. The xylem consists in the middle of wide scalari- form tracheides, and interstitial bands of Parenchyma. According to all the data, which however require to be more exactly established, the bundles described may represent an intermediate form between the concentric and radial Fern-bundles. According to Russow the likewise axial bundle of the stolons of N. pectinata and rufescens has not the structure described. 3. Radial Bundles. Sect. 107. The radial bundles are closely connected with the concentric by means of those in the stem of Lycopodium, and by the diarch forms, which occur in many roots. In the typical cases they are distinguished by the fact that in the radial bundle the xylem forms several bands running out radially from the centre, between which lie the same number of groups or bands of phloem alternating with them, In - all radial bundles the development of the characteristic elements, both of the xylem and phloem bands, begins at the periphery, and proceeds thence with varying celerity towards the middle. The primitive elements, which in the xylem are here also distinguished by. their narrowness, form the peripheral ends of the rays. As these thus form the points of departure of the development of xylem, it is usual to speak of the number not of the rays, but of the starting-points—of di- to polyarch bundles’. Radial bundles occur in the stems of Lycopodiacez, and in the filiform stolons of Nephrolepis; and in all roots, with a few exceptions mentioned at p. 319. The axial strand which traverses the middle of the stem of Lycopodiacez agrees, with the exception of its radial structure, with the bundles above described in the stem of the Selaginellz (with the exception of S. spinulosa), which, in the struc- ture and development of their xylem, correspond to the diarch or oligarch radial forms. : In the stem of Pszlofum? this strand is cauline, not receiving or giving off leaf- strands. In the branches which appear above ground the cross-section of the whole is almost circular, bounded on the outside by an endodermis, which only differs from the surrounding parenchyma in its undulating radial walls. The xylem is triarch to pentarch and octarch; its not always equidistant rays are separated from the endo- dermis by one or a few layers of relatively narrow, elongated, prismatic parenchymatous cells, and consist at their peripheral ends of a group of narrower reticulated tracheides (I did not find spiral tracheides), and towards the centre of a few rows of scalariform tracheides ; these rows do not reach to the middle of the bundle, but abut ona strand of elongated prismatic, pointed sclerenchymatous fibres, which traverse the middle. The rest of the substance of the bundle consists of thin-walled prismatic parenchyma, in which, especially at its periphery, are scattered some few-celled groups of somewhat narrower and thicker-walled sieve-tubes. This designation is at least justified by the appearance of the smooth lateral walls, agreeing with that in the Ferns, by the granular contents adhering obstinately to the walls, and by the absence of nuclei, which is very conspicuous on comparing these elements with the surrounding cells ; on the thin oblique terminal-surfaces of the articulations I.believe that I have directly 1 Nageli, Beitr. 7c. p. 10. * Nageli, 2. c.—Russow, /.¢. p. 131. STRUCTURE OF RADIAL BUNDLES, 349 seen delicate sieve-pores. In the subterranean shoots of the rhizome (Nageli and Leitgeb’s rhizoides) the bundle is very weak and rudimentary in its development; I find only a flat or three-cornered xylem, consisting of a few, frequently only 3-6, re- ticulated and scalariform tracheides, separated here and there by thin-walled elements; . the peripheral tracheides are but little narrower than the internal ones; the xylem is completely surrounded by 2-4 layers of delicate spindle-shaped cells. I could see nothing of any sieve-tubes. The vascular bundle of Tmesipteris appears from Russow’s statement to have a similar structure to that of Psilotum. In the cylindrical axial strand of the stems of Lycopodium (comp. p. 281) the xylem consists of a number of plates or bands, the peripheral corners of which are each formed by a group of narrow tracheides (comp. p. 163), the above-described points of attachment of the leaf-trace bundles, while the larger inner part consists of wider scalariform tracheides. (Comp. Fig. 162.) The number and arrangement of the plates and their relations to the rows of leaves vary with the species, and with the vigour of the individual shoots. Of the latter relations we have already spoken above. As regards the other con- Coro <2 ditions which come under con- f ea % 7 sideration’, among the native species investigated L. inundatum has 3-3 plates, united in the middle to form EO Re ie a body which has a stellate cross- ei section, thus constituting a tri- to pentarch radial xylem. The latter is however even here not unfre- quently in so far irregular, that one _,,,1ic,t4¢- tacmpaium Chaneayparine, Crosesecon of 2 shot, or the other plate separates from nine genes conexito.e right the cross-section of a bundle running into a the rest, so as to have an isolated course for some distance, and then again to unite with the others. Four radial plates united in the middle are present in the ultimate ramifications of the heterophyllous species, as L. complanatum, and as a rule L. alpinum, though here deviations occur in 20-30 per cent. of the cases. In the stouter axes of the last-named species, as well as in L. clavatum, annotinum, and Selago, the number of the vascular plates is higher, being in proportion to the thickness of the shoots ;—in stout main stems of L. com- planatum and alpinum, for example, it amounts to 11 and 13, in those of L. anno- tinum and clavatum to 17, but diminishes again in the weaker ramifications to 4 and 3. In these cases the plates are only partially, or scarcely at all radially convergent; most of them rather form separate bands in the decidedly bilateral prostrate main-shoots of all species possessing them (Fig. 162), these bands being slightly curved, with their convex surface always directed towards the lower side of the stem, and their corners lying chiefly right and left ; they are further united with each other in a great variety of ways, sometimes radially, sometimes so as to form loops. Their union and 1 Hegelmaier, /.c. p. 790. 350 PRIMARY ARRANGEMENT OF TISSUES, " separation vary in successive sections of their longitudinal course. Feeble branches of higher order once more show a more radial arrangement and mode of union, In the non-bilateral stems of L. Selago radial union of all the 4-6 plates is, according to Hegelmaier, the more frequent case, while irregular winding and grouping are more rare. For further details compare the treatises cited at p. 281. The intermediate spaces between the vascular plates, which are usually smaller than the latter, are occupied by the one or more masses of phloem of the bundle, each constituting a correspondingly-shaped group of elongated prismatic parenchymatous cells with oblique ends, and apparently oily contents, among which lies a usually simple intcrrupted row of wider sieve-tubes, represented by the wider, somewhat more strongly contoured meshes of Fig. 162 (comp. p. 181). The walls of all the elements of the phloem are soft, swell strongly in water, and become blue with solution of iodine in potassium iodide. Between the peripheral angles of the vascular plates, and alternating with them, lies in each phloem-portion a small group of thick-walled, narrow, elongated fibrous elements,—the primitive elements of the phloem. Round all the corners runs a zone of prismatic’ parenchyma, usually two cell-layers thick, of the same or similar cell-form and structure to that of the phloem, but in most species (L. clavatum, annotinum) distinguished by inter- cellular spaces, and loose, easily separable connection of the cells. A sheath, con- sisting on the average of two layers of tangentially elongated celis, possessing thin walls, cuticularised according to Russow, and not undulating, surrounds the whole vascular bundle, and unites it with the inner cortex, which according to the species is parenchymatous or sclerenchymatous. - The stout roots of Lycopodium clavatum', Alpinum, and species of similar growth, have essentially the same structure as the stems. In the two species men- tioned the xylem is hexarch to dekarch, very often heptarch, and then so arranged in the simplest most regular case as to form three separate plates, one being diametral, while two stand symmetrically in front of the two surfaces of the first; these two are concave, with U-shaped cross-section, and with the concavity turned towards the periphery. Every plate diminishes in breadth in the centripetal direction, and often consists in the middle of only a single scalariform tracheide. Irregularities and inter- ruptions of the plates occur similar to those in the stem. In the heptarch or octarch examples of L. clavatum investigated, I almost always found one of the concave plates larger, and of narrow horseshoe-like cross-section, the other smaller and much flatter, with a separate, in cross-section elliptical or wedge-shaped, vascular strand (in itself monarch), lying in front of its slightly concave outer surface. Other arrange- ments however occur, and these are sometimes most irregular and involved. The structure of the surrounding tissue and of the spaces between the vascular plates is the same as in the stem. In the branches of these roots the number and arrangement of the plates become reduced and simplified as the thickness diminishes ; their last ramifications—and in L, Selago and inundatum all roots of every order of ramifica- tion—have only a vascular group surrounded by a phloem, which is perhaps only parenchymatous (?). In the branches of the root of the stouter species first-named, the * Nageli und Leitgeb, Entstehung, &c. der Wurzeln, p. 117, &c—-Van Tieghem, Ann. Sci. Nat, 5 ser. tom. XITL STRUCTURE OF RADIAL BUNDLES, 351 former consists of a few small vessels lying on one side of the cylindrical bundle. In the roots of L, Selago and inundatum a strongly-curved, diarch vascular plate, sickle- shaped in cross-section, lies, according to Russow’s description, inside the cylindrical phloem, the sieve-tubes being situated between its limbs. = Sect. 108. In the great majority of roots the axial bundle which traverses them is of very regular radial structure, which in its principal characters is uniform in all cases *. The approximately cylindrical bundle is surrounded by an endodermis, which is either permanently undulated, or is only so at first, becoming sclerotic in the mature condition. According to its origin the endodermis is not to be assigned to the bundle, but forms the innermost (limiting) layer of the surrounding cortex. The xylem is according to the particular case diarch or polyarch, and its starting-points, corresponding to what are afterwards its peripheral corners, all lie at equal distances from one another: in diarch bundles at diametrically opposite points of the circular cross-section; in all other cases removed from each other by the fraction of the periphery determined by their number (3, 4, &c.). From the starting-points vascular plates develope ina radial direction, and in centripetal order of development ; and these either meet in the centre or do not reach it, but remain separated by a parenchyma- tous or sclerenchymatous mass, which permanently occupies the middle of the bundle. The same number of phloem-groups alternate with the vascular plates, to which they thus correspond in number and arrangement. The xylem and phloem-rays are separated from each other by delicate paren- chymatous cells, and in fact two layers of the latter may as a rule be distinguished between each xylem and the next phloem group; more rarely only a single layer is present, or there are more than two. On the outside an uninterrupted zone of parenchyma, which usually forms a single layer, more rarely two layers, and rarely several, constitutes the limit of the whole bundle towards the endodermis. In the case of the Ferns, Nageli and Leitgeb have called this limiting layer the Perdcamdium, a name which"it may here bear generally, even where, as in Equisetum, its origin is different from that in the cases for which it was first introduced. In Monocotyledons however cases are not rare, where the outermost vessels border directly on the endo- dermis, and the Pericambium is thus interrupted at every xylem-plate, and only surrounds the phloem-rays. Van Tieghem calls the whole of the cells, which are interposed between the groups of xylem and phloem, and thus unite them into a dense cylinder; conjunctive tissue (tissu conjonctif). The latter forms, as has been said, the usually two-layered bands between the xylem-plates and phloem-groups, and is continued inwards between the former in cases where they do not meet. Externally it borders on the pericam- bium. The latter is called by Van Tieghem in the case of the Phanerogams the thizogenic layer, from the function which it performs in the origination of lateral roots. 1 Nageli, Beitrige, /.c. p. 23.—P. van Tieghem, Recherches sur la symmeétrie de structure dans les Plantes vasculaires. I. La racine. Ann. Sci. Nat. 5 sér. tom. XIII.—Nageli und Leitgeb, Ent- stehung d. Wurzeln, Miinchen, 1867.—Nicolai, Zc. (compare p. 231).—See also Link, Icones anatomicze.—Schacht, Lehrbuch, p. 167, etc. 352 PRIMARY ARRANGEMENT OF TISSUES. For the structure of the individual parts few general rules are to be given, except those which hold good generally for vascular bundles and their sheaths. The xylem-plates consist of one or more rows, which, according to the particular case, are uninterrupted in the radial direction, i.e. one trachea follows on another; or they are interrupted by the interposition of non-equivalent (parenchymatous or sclerenchymatous) elements. For the special nature of the trachez, i.e. whether they are vessels in the strict sense or tracheides, the rules and difficulties stated at p. 164 apply. The first-formed vessels or tracheides, which occupy the corners, are always narrow, the later ones, following in a centripetal direction, become suddenly or successively wider. The latter are always pitted or reticulated vessels (or tracheides); the narrow peripheral ones are as a rule also reticulated or annular vessels, with dense and fine thickening fibres, the prevalent direction of which is transverse. For short distances however the fibre has not uncommonly in these cases also a simply spiral course. Closely wound spiral fibres, which can be unrolled for a long distance, occur more rarely, e.g. in the roots of Tornelia fragrans, Cucurbitaceze, Anthriscus Cerefolium (Van Tieghem), Phaseolus (Dodel), Cycadeze (Mettenius), and Coniferz. The structure of the phloem-rays, where they are well developed, is essentially the same as in the typical collateral or concentric bundles. In feeble roots of Mono- cotyledons they are not uncommonly reduced to one sieve-tube with narrow-celled surrounding tissue (e. g. Triglochin maritimum, Aponogeton, Hydrocleis Humboldtii, Potamogeton lucens, comp. Van Tieghem, Zc. Taf. VI), this being of typical structure, only small. It is therefore to be supposed that the typical structure belongs to them generally, though they still require more exact investigation, especially in the small- celled bundles of Dicotyledons. I should also wish to extend the last remark to the roots of Conifers, in the primary bundles of which, according to Janczewski’s more recent statement’, sieve-tubes are said to be wholly wanting. The number, and with it also the arrangement and relative breadth of the xylem and phloem-rays, the extent and distribution of the tissue occurring around and between them, lastly the special structure of the particular forms of tissue, and thus the entire structure of the root-bundle, vary, sometimes in different roots of the same species, sometimes according to the species and the larger systematic divisions. In the former relation the general rule holds good that as the thickness of the roots diminishes, not only does the number of the tissue-elements in the bundle diminish, but also the number of its radial plates, if in the thicker specimens this exceeds two. Further slight individual differences, which cannot be referred to difference in thick- ness, occur among members of the same species.‘ In the other relation, besides the obvious identity or similarity of structure of closely related forms with similar adap- tation, the great conformity of structural plan in all divisions of vascular plants is to be emphasised. For none of them can a special structure be stated as everywhere characteristic of the group. Van Tieghem’s first plate shows the almost identical cross-sections of young roots of Cyathea medullaris, Allium Cepa (main root of the seedling), Taxus, and Beta. Smaller differences between subdivisions of the larger’ classes are often more sharply expressed. The existing investigations give rise to the following rules :— 1 Ann. Sci. Nat. 5 sér. tom. XX. p. 34. STRUCTURE OF RADIAL BUNDLES. 353 1. In almost all Dzcofyledons where the point has been investigated, the original bundle of the root is oligarch, usually with 2, 3, or 4 rays, more rarely with 6 or 8, while higher numbers occur exceptionally. In the mazn-roots the xylem-plate is usually diarch-diametral, triarch, or tetrarch; higher numbers occur rarely, whether in single individuals (as 5-7 in specimens of Vicia Faba, and perhaps even 12", instead of 4), or as the rule for ‘certain species, as most Amentaceze (Quercus sp. 6-8, Alnus 5-6, Castanea 6-12, Fagus 8, Carpinus 4), Alsculus (6), Coffea (8), &c. None of these ‘humbers are constant absolutely and without exception even for the particular species. Whether a definite number can be characteristic of one of the larger genera or of a natural family (apart from occasional individual variations) is not to be decided from the existing data. At any rate this is the case in several families of which a dozen or half-a-dozen representatives have been investigated. Diarch xylem-plates occur, for example;-in the main root of all investigated Cruciferze FIG, 163.—Ranunculus fluitans. Cross-section through the vascular bundle of a strong old adventitious root (225). ze endoderinis, 2 pericambium, gy external primordial vessels of the diarch uniseriate xylem g—g; between g—g and ? is the phloem. (Brassica, Raphanus), Fumaria, Caryophyllacez, Vitis, Urtica, Umbelliferse (Anthris- cus Cerefolium, Foeniculum, Petroselinum sativum, Carum Carvi, Coriandrum, Daucus, Pastinaca sativa v. Tieghem), Chenopodiaceze (Beta, Atriplex, Spinacia), Mirabilis, Centranthus, and Valeriana, and in Tagetes erecta among Compositae; tetrarch xylems as a rule occur in the investigated Cucurbitaceze (Cucumis, Cucur- bita, Lagenaria, Luffa), Euphorbiaceze (Euphorbia, Ricinus, Mercurialis sp.), Tropz- olum majus, Convolvulus tricolor; generally the numbers 2 and 4 appear to be predominant. But on the other hand considerable differences between the forms investigated occur in the case of the higher numbers of the Cupuliferee above 1 Compare van Tieghem, /.c. p. 223. In the case cited it was doubtful whether the main root or a strongly developed lateral root was.in question. Aa 354 PRIMARY ARRANGEMENT OF TISSUES, mentioned ; and among the Papilionacez, of which more numerous representatives have been investigated than of other families, a considerable variety of the conditions in question is to be recorded; as a rule the xylem-plates are diarch in Lupinus varius and Trigonella, triarch in Pisum sativum, Lathyrus sativus, Orobus vernus, Vicia sativa, Ervilia villosa, Ervum Lens, Hedysarum coronarium, Onobrychis sativa, and Medicago sativa; tetrarch in Phaseolus, Dolichos lignosus, and Cicer arietinum : lastly, higher numbers than feur occur, as mentioned above, in Vicia Faba. In the dranches of the root the numbers remain as a rule the same, or diminish if they were greater than two. In suds¢diary roots springing from the stem they often increase, in correspondence with the thickness of the roots ; amounting, for example, to 7,9, 11 in Cucurbita maxima, 5, 6, 8 in Lagenaria and Luffa (van Tieghem), 4-5 in Phaseolus, 5-8 in Valeriana; the adventitious roots on the rhizome of Nymphza alba have 6-10 rays; in Nuphar luteum there are as many as 24; in an aerial root of Clusia flava van Tieghem found 13 rays, and so on. The converse case, however, also occurs; there is a diarch xylem-plate in all the adventitious and lateral roots of Tropzolum majus, the main root being tetrarch. The orientation of the parts in the cases investigated is such that, in the case of” diarch and tetrarch structure of the mazz root, the surface or one of the two inter- secting surfaces of the xylem-plates always coincides with the median plane of the two cotyledons, which diverge at an angle of 180°. In the triarch main roots of Pisum, and the other triarch Papilionaceae mentioned, the planes of two xylem-plates fall according to van Tieghem in the median planes of the two cotyledons, which only diverge at an angle of 120°. For higher numbers exact statements are wanting. In all Phanerogams the plane of the diarch xylem-plates of Jaseral roots always lies in the median plane of the main axis from which they arise, and the same applies, so far as investigated, to one of the planes in tetrarch xylems. The original structure of the individual bands of tissue shows—within the general plan of structure of root-bundles—few peculiarities characteristic of Dicoty- ledons. As regards the xylem-plates the usually very gradual increase of the width of the vessels in the centripetal direction is worthy of remark. Only as an exception, . in the polyarch subsidiary roots on the rhizome of Primula Auricula and Nym- phaeaceze does the case usual in Monocotyledons occur, namely, that the short row of vessels, which does not reach to the centre, consists of a few narrow peripheral vessels, and then of one or several which are very wide (Fig. 164). In most cases belonging to this series the one- or few-rowed plates constitute radial bands, narrow in cross-section, separated by relatively very broad interstices. These bands either meet in the middle, or they are separated, or connected together, by means of a parenchymatous axial strand. In stout polyarch subsidiary roots, and in the upper part of stout main-roots, where they pass over into the hypocotyledonary stem, this axial parenchymatous mass, the ‘pith’ of the root, is of considerable thickness. Rarely, among Dicotyledons, the axial parenchyma connecting the xylem-rays is represented by a strand of sclerenchymatous fibres, e.g. in the subsidiary roots of Stachys sylvatica, Mentha aquatica, and Hedera Helix (v. Tieghem). A peculiarity, which so far as 1 am aware only occurs among Dicotyledons, is the presence of a bundle of sclerenchymatous fibres, roughly crescent-shaped as seen in cross-section, on the outside of the phloem-groups of triarch and.tetrarch roots of STRUCTURE OF RADIAL BUNDLES. 355 Papilionaceze (Pisum, Phaseolus). The fibrous bundle lies inside the pericambium. In other respects the phloem still requires more exact histological investigation in these cases. In all roots of Dicotyledons investigated a pericambium, consisting of one, or in many cases of several layers, completely surrounds the xylem-pldtes. Those peculi- arities of its structure which are related to the development of lateral roots, the resin- canals which sometimes occur in it, and other points connected with it, will have to be discussed below (Sects. 117 and 133). This original structure of the roots of Dicotyledons is, however, permanent in but few cases; in most cases, and in many species immediately after its origination, it is altered by the secondary growth in thickness, starting from the inside of the phloem-rays, of which we shall treat in Chap. XIV; comp. Fig. 165. Hence result essential and actual differences from other roots, especially those of Ferns and Mono- cotyledons, in which, with the ex- ception of many roots of Draczena ', these secondary changes are want- ing. It must, however, be expressly stated that the changes due to secondary growth in thickness occur in by no means all roots of Dicoty- ‘ FIG. 164.—Primula Auricula (225). Cross-section through the heptarch ledons, and thus do not establish vascular bundle of an adventitious root and its surrounding tissue. # peri- “ anand ‘ cambium; g the external primordial vessels of the xylem rays, which any generally valid distinction be- alternate with the same number of phloem groups s, and are separated from the latter by thin-walled parenchyma; « endodermis, outside which tween these and the others. Apart is tolerably thick-walled cortical parenchyma, with intercellular spaces quadrangular in cross-section. J from those cases where, as in the subsidiary roots of Stachys silvatica, Mentha aquatica, Lysimachia nummularia, Myrio- phyllum, and Hippuris, the secondary growth in thickness is infinitesimally small, and as such even doubtful, because in other cases also the innermost vessels uniting the plates are developed very late, this secondary growth is completely absent in a number of subsidiary roots. For instance’if/:tHdse of Gunnera®, the Nymphzeaceze, Ficaria ranunculoides, and Primula Auricula, to which cases it may be anticipated that more extended investigation will add-others. The fact that a rudimentary secondary growth in thickness occurs at the points of insertion of the roots in question (in Ficaria and Nuphar) has no effect on thé condition of their much greater portion. The fact that in roots of Dicotyledons sclerosis of the endodermis but rarely occurs no doubt stands in the closest relation with the occurrence of secondary growth. Such sclerosis, however, occurs for example in the adventitious roots on the rhizome of Priniula Auricula and.Ranunculus repens ; comp. Figs. 164 and 165. 1 Compare Caspary, Pringsheim’s Jahrb. I. p. 446.—Falkenberg, 7. c. p. 197. ? Reinke, Morpholog. Abhandl. p. 58. 3 Van Tieghem, /.¢, p. 266, &c. Aaz2 356 PRIMARY ARRANGEMENT OF TISSUES, 2.. The axial bundle of the root in the Gymnosperms? is in general similarly con- structed to the ordinary one of Dicotyledons. Its original structure is always altered very early by secondary growth from the cambium ; the sclerenchymatous fibres in the periphery of the phloem region of Dion described by Reinke may have owed their origin to this. Qver the angles of the xylem-plates the pericambium is single-layered in Taxus, Thuja, and Biota; many-layered (from 3 and 4 to 7 cell-layers thick) in species of Podocarpus, Pinus, and in the Cycadeze investigated. ! The xylem-plates consist at their outer corners of tracheides, with the fibrous thickenings generally characteristic of this region; in their internal, later developed portion they consist of pitted tracheides, such as are characteristic of the wood of Gymnosperms. Among the Coniferz the Cupressinee and Taxinez have diametral and diarch xylem- plates in roots of all degrees, or more rarely triarch ones. In the Abietinex higher FIG. 165 (145).—Ranunculus repens. Cross-section through the vascular bundle .of an old adventitious root. w endodermis, 7 pericambial layer, g external primordial vessels of the tetrarch xylem, ~ large axial pitted vessel. In the pitted vessel x surface-view of a roundly perforated cross-wall. A narrow zone of secondary wood Has been deposited on the primary xylem plates extending from gto 7; the cells between this and the phloem groups are tangentially divided ; cf, Chapter XIV. numbers and with them greater individual variations are the rule, though here no con- stant relation exists in the main root between these variations and the number of the cotyledons, which, as is well known, is likewise variable and always more than two. In Abies excelsa, for example, van Tieghem found in 13 seedlings a triarch root-bundle, the cotyledons numbering 7, 8, 9, or 10; in a specimen with 6 cotyledons the bundle was diarch, in one with 8, tetrarch. The numerous investigations of the observer mentioned established similar relations for the species of the genus Pinus in the narrowest sense (P. Pinea, halepensis, sylvestris, &c.). The number-of the xylem-plates here amounts 1 See van Tieghem, /. c.—Strasburger, Coniferen und Gnetaceen, pp. 340, 360, &c.—Mettenius, Beitr, 2, Anatomie d, Cycadeen, p. 595, &c.—Reinke, Morpholog. Abhandl. I. . STRUCTURE OF RADIAL BUNDLES, 357 to 3-6, rarely 7. They are distinguished from those of the closely-related Abietinez by their form, which may be compared to that of a Y. Each of them is, literally speaking, diarch; and begins externally with two rows of about five narrow tracheides, touching - the pericambium at two separate points; they converge towards the inside and abut on each other. From their point of junction a radial row of tracheides, 1-2 layers thick, extends in the centripetal direction, without reaching the centre of the root. In the angle of the Y lies a resin-canal surrounded by delicate cells. The roots of Ephedra which have been investigated have diametrally diarch xylem-plates. Among the Cycadex the xylem in the subsidiary and branch-roots, which have been investigated in numerous species, is usually diametrally diarch, the two original plates meeting in the middle, or being separated by parenchyma. The same holds good for the investigated main roots of Cycas revoluta and Zamia furfuracea. More rarely, in the thick subsidiary roots of usually diarch species, the bundles are three-rayed. Ina hybrid Ceratozamia van Tieghem found three or four xylem-plates, and in a specimen of Zamia muricata Mettenius found six in the main root. In the subsidiary roots of Cycas revoluta, when the centripetal development of the plate is already advanced, some scattered narrow reticulated vessels appear at the sides of its peripheral corners, as if secondarily; whether these constitute the first beginnings of the secondary growth re- mains to be decided. 3. Among the Monocotyledons there are first of all many thin main roots of the seedling, which in the structure of their axial vascular bundle are indistinguishable from those of Dicotyledons and Gymnosperms. In the case of Allium Cepa, with a diametrally diarch, and sometimes triarch xylem-plate, this has already been mentioned above; A. Porrum and Lilium Martagon are characterised by a similar structure of the main-root, Tulipa Gesneriana shows the deviation that its pericambium consists of two layers instead of one. Bulbine annua has three xylem-plates, which do not meet; Iris Monnieri has four. The weaker roots of all degrees are essentially similar to those just described. Stouter main-roots, such even as those of species of Asphodelus, Canna, and Asparagus officinalis, and then those of the Palms (Phoenix, Seaforthia elegans), and above all the subsidiary roots springing from the stem (which in this class, as is well known, usually far exceed the main-roots in thickness), though in the great majority of cases they maintain the typical plan of structure, yet become polyarch as their bundles increase in bulk, and also show a more varied differentiation, due to differ- ences of many kinds in the structure of the tissue-elements. Comp. Figs. 166, 167, 168. « First of all, as regards the number, arrangement, and form of the groups of tissue of these typical roots of Monocotyledons, the number of the xylem and phloem rays rises from 5-10 up to 20, 50, and more. The thick roots of Iris, Asparagus, Smilax (Sarsaparilla), Palms’, &c., are examples of a high degree of polyarchy. The phloem-bands are always small, consisting of relatively few elements, their cross- section being roundish or radially elongated. The xylem-bands, consisting of one or a few rows of elements, usually begin at the periphery with a short uninterrupted radial band of narrow trachez, which become gradually wider towards the inside. These are suddenly followed in the centripetal direction by one or a few, very wide, reticulated or pitted vessels. The latter are usually separated from the peripheral 1 Von Mohl, Palm, Structura, Diplothemium maritimum, Tab. I. 358 PRIMARY ARRANGEMENT OF TISSUBS. part of the row by one or more layers of interstitial cells. In the thicker polyarch bundles these large vessels are often confined to certain rows, while in the others, which alternate irregularly with the former, they are absent; or the case frequently FIG, 166.—Acorus Calamus. Cross-section through a vascular bundle and the neighbouring part of the cortex of an adventitious root. s endodermis; 2, g narrow primitive vessels; g larger internal vessels, not yet completely developed; #4 phloem-groups. From Sachs’ Textbook. FIG. 167.—Very thin cross-section through the vascular bundle of an older adventitious root of the same plant (145). s endodermis, g primitive vessels, w phloem groups. The axial mass of cells, which is still thin-walled in . Fig. 166, is here sclerotic, and the internal vessels are completely developed. occurs that two neighbouring rows converge at an acute angle towards a large vessel, forming in cross-section the figure of a V, in the angle of which the large vessel lies. Xylem-plates do, however, occur among Monocotyledons also, in which the elements STRUCTURE OF RADIAL BUNDLES, 359 become wider quite gradually in centripetal order, e.g. in many Orchids, as Stan- hopea sp., Epidendron ciliare, &c. Here also the number of the vessels of a plate following one another in the radial direction is small, on the average 4-6, not un- frequently still fewer. In the Carices investigated the row usually consists of a single narrow peripheral vessel, or two lying side by side in the tangential direction, and. one wide internal pitted vessel, the latter being separated from the former by at least three layers of parenchymatous cells. A second narrow pitted vessel may lie between the two. The peripheral vessels often occur without the corresponding wide one, so that a row can no longer be spoken of. In smaller roots or bundles, as in the main roots mentioned above, the weaker roots of all degrees among the Grasses (Secale, Triticum), and in weak adventitious roots of Tradescantia virginica, the xylem-rows either meet in the middle of the bundle, or converge towards one or two wide vessels passing through the centre of the bundle, which originate very early, but attain their development very late. The xylem-rows sometimes come into direct contact with these, or are sometimes separated from them by a few interstitial cells. In the thicker typical roots of Monocotyledons the radial xylem-plates do not nearly reach the centre. The latter is occupied by a thick cylinder of parenchyma or sclerenchyma, at the circumference of which the system of xylem-plates often forms a relatively narrow ring. In the great majority of roots of Monee) t ledors the xylem-ring is surrounded on the outside by an uninterrupted pericdtisiutn, which is one layer thick over the xylem-plates, and the outside of which borders on the endodermis. It rarely con- sists of two layers over the xylem-plates, as in the main roots of Tulipa Gesneriana mentioned above, and in roots of Sarsaparilla. All the roots of Graminez investi- gated form a remarkable exception to this rule (Oryza’, Secale, Triticum, Zea, Coix, Sorghum, Hordeum, and Paspalum spec.?), as in these the pericambium is, as a rule, interrupted by the rows of vessels which thus border directly on the endodermis. Even here, however, a small pericambial cell often lies between the endodermis and the outermost vessels, e.g. in Maize. The same occurs among the Cyperacee in species of Carex. In certain cases the narrow pitted vessel borders closely on the endodermis, e.g. in C. foenea, folliculata, divulsa, and hirta; or both this and the usual arrangement, in which a pericambial cell is present outside the vessel, may occur in different parts of one and the same cross-section. According to van Tieghem, other species of Carex, such as C. brizoides, only show the latter typical arrange- ment; the same is the case in species of Cyperus, as C. longus and C. alternifolius. In these typical roots the structure of the single tissue-elements shows a variety of individual differences, both as regards the vessels, and no doubt the sieve-tubes also, though the latter in most cases still require more accurate investigation. Into these differences we cannot here enter at all minutely. The mass of cells, forming simultaneously longitudinal and concentric rows between and inside the xylem-plates, shows sometimes a typically parenchymatous, sometimes a typically sclerenchymatous structure, or a form intermediate between the two. And indeed these peculiarities either extend uniformly to the whole interstitial mass of tissue in question, or are 1 Nageli und Leitgeb, Zc. . 2 Van Tieghem,./.c. (See further, Klinge, Vergl. hist. Unters. d. Gramineen u. Cyperaceen- wurzeln, Mém. de Acad. Imp. St. Pétersb. VII. Sér. Tom. XXVI. No. 12, 1879.) ' 360 PRIMARY ARRANGEMENT OF TISSUES. different in certain zones and groups. Of the combinations which are here possible the following usually occur: (1) the entire interstitial mass of cells, including the cylindrical axial portion, remains thin-walled and parenchymatous, e. g. adventitious roots of Tradescantia virginiana, Curcuma longa, and Clivia nobilis (Acorus Calamus is intermediate, i.e. its parenchyma has very strong walls). (2) The whole mass of cells mentioned becomes sclerenchymatous, e. g. Carex divulsa, Cyperus alternifolius, and no doubt most adventitious roots of Cyperacez and Grasses; also Curculigo: recur- vata.. (3) The tracts of cells between the xylem- and phloem-plates are sclerenchy- matous, forming together with them a dense firm ring around an axial strand of parenchyma, with intercellular spaces containing air: e.g. roots of Smilax (Sarsa- parilla) with very extensive parenchyma containing abundance of starch, most aerial roots of Orchidacee (e.g. Epidendron ciliare, Oncidium sphegiferum), and many Palm roots (cf. Mohl, /.¢.), in which scattered sclerenchymatous fibres may occur again inside the axial thin-walled parenchyma, e.g. Chamzdorea elegans. The pericambium remains in most cases thin-walled and parenchymatous, even where it borders on sclerenchyma, though it may itself eventually be involved in the sclerosis, either wholly or in part; the former, for example, is the case in Sarsaparilla roots, the latter among Orchids, e.g. Epiden- dron ciliare, where opposite each xylem-plate two rows of its cells usu- ally remain very delicate, while the others, like those adjoining them on the inside, become greatly thickened. The usually one-sided sclerosis of the endodermis, which is very frequent, though by no means universally dis- tributed among long-lived Mono- cotyledonous roots, has been dis- ; cussed in Sect. 27. FiG, 168.—Philodendron Imbe Hort. Halens. Cross-section Deviations from the type of through a thick subsidiary root, slightly magnified. The axial vascular bundle, and on the right the entire cortex, are shown; 2 structure of Monocotyledonous roots external limit of the xylem-rows. The obliquely shaded radial curvoundiog an intercelularpastage containing tc." hitherto considered occur in dif- ferent degrees as regards the ar- rangement of the forms and regions of tissue, though their structural conditions remain the same. In thick root-bundles, some or all of whose xylem-plates converge in pairs in the form of a V, the groups of phloem lying inside a V are often smaller than those between two V’s. The latter are frequently large, radially-placed plates, while the former are roundish groups. This occurs in an exquisite form in the aerial roots of an Aroid cultivated in Halle under the name of Philodendron Imbe (Fig. 168), and also in Palms; cf. Mohl’s figure of Diplothemium maritimum already cited. In Chameedorea elegans this inequality goes further. In the corner of the V lies a small roundish group of phloem; between every two V’s a similar group lies towards STRUCTURE OF RADIAL BUNDLES. 361 the outside, while further inside a second occurs, which is elliptical in cross- section, and is separated from the outer group by interstitial sclerenchyma, in which vessels frequently lie. Thus both an inner and an outer row of phloem-groups are here present in the otherwise typical bundle. It has been mentioned above that the xylem-plates not uncommonly converge towards one or two axial vessels, though it may be without coming into immediate contact with them. Such axial vessels frequently occur isolated, in the middle of a thick parenchymatous or fibrous cylinder, and are separated from the inner edges of the radial plates by many layers of cells: This kind of structure is found here and there as an individual peculiarity of many roots, as for example in the Sarsaparilla of Veracruz’; in Carex folliculata I found, on the same stock, roots of the structure usual in Carex, with a thick dense sclerotic axial-cylinder, and others in which the middle of the latter is traversed by about 5 moderately large, prismatic, pitted vessels, which are in contact with one another. These trifling forms of deviation constitute the transition to those more con- spicuous cases, in which numerous vessels, as well as groups of sieve-tubes, occur scattered in the whole of the cylinder inside the radial ring, a phenomenon which is characteristic of the thick adventitious roots of many epiphytic Aroidex, of those Musacez which have been investigated, of the Draczenez, Pandanez, (Pandanus, Freycinetia, Cyclanthus), and of the Palms Iriartea exorrhiza and I. preemorsa. All the investigated roots of terrestrial Aroidez, as well as those of many epiphytic species, present the usual typical structure; but in the thick aerial roots of other forms, scattered wide vessels, and very large sieve-tubes, isolated or occurring in pairs, and accompanied by cambiform tissue, are distributed throughout the wide and constantly sclerenchymatous cylinder inside the radial ring. The two kinds of elements do not lie in the radial rows; Tornelia fragrans, Heteropsis ovata, Monstera surinamensis, Adansonii, Raphidophora angustifolia, Scindapsus pictus, Philodendron micans, and Anthurium digitatum ?, are examples of this structure. The same occurs in species of Strelitzia, and no doubt in other Musacez °, Essentially the same arrangement is present in roots of Dracenz and Pandanea, with the sole difference that the axial tissue in which vessels and sieve-tubes are distributed is not homogeneous, but around the vessels and small phloem-groups consists of scleren- chymatous fibres, while between them it is formed of parenchyma, in which, in the case of Pandanus, lie wide intercellular passages containing air and scattered bundles of fibres. The ring has likewise sclerotic interstitial tissue between the radial xylem- and phloem- groups, the number of which even in moderately thick (1.5 cm.) roots of Pandanus amounts to nearly 200 of each. The cross-section of such roots therefore presents first the typical, relatively narrow ring, surrounded by a many-layered -pericambium and an endodermis, and then, inside this, a wide space filled by parenchyma, in which numerous thick strands run longitudinally. Each of these strands consists of a many-layered mass of sclerenchymatous fibres, in which are enclosed one or more isolated wide vessels or groups of them, and one or more small phloem-groups separated from the vessels, while more rarely only one, or neither of the two forms of tubes are present. The position of the two in the strand varies irregularly. The distribution of the strands in the parenchyma appears in equivalent roots to be somewhat different according to the species. Among 1 See Berg, Atlas d. pharmac. Waarenkunde, Taf. III. g. ® Van Tieghem, /.¢. p. 149. » Compare Wittmack, Musa Ensete, Halle (Linnza), 1867, p. 62. 362 PRIMARY ARRANGEMENT OF TISSUES, the Pandanez, for example, I find them, as seen in cross-sections, isolated and irregularly distributed, in the thickest roots of Freycinetia nitida of the Berlin Gardens; in Pandanus pygmeus (graminifolius of gardens) they are arranged in transverse rows (parallel to a diameter), which are separated from one another by broader bands of parenchyma. In P. odoratissimus, two or more strands separated by narrow bands of parenchyma are placed together in groups, and the groups scattered between broader masses of paren- chyma. As the thickness of the roots diminishes the conditions of structure described become simplified. A branch-root of Pandanus pygmzus, 1-2 mm. in thickness, has, for example, inside the radial ring about 2-3 large vessels, and the same number of phloem- groups, enclosed in homogeneous fibrous sclerenchyma, which is directly continued into the ring. Branch-roots of Dracena reflexa about 1 mm. in thickness have a thoroughly typical structure, the radial ring surrounds a thin-walled axial cylinder of parenchyma. It is only in thicker roots that an irregularly placed strand of sclerenchyma containing vessels appears, such strands becoming very numerous as the roots increase in thickness. The roots of Iriartea, finally, which are an inch in thickness, are distinguished from those last described, first by the fact that their bulky vascular mass is not cylindrical, but deeply furrowed, having in cross-section the form of a star with about ten blunt and usually bifid rays; further by the fact that the radial ring also is divided up into scleren- chymatous bundles, enclosing the vessels and phloem-groups, and radial bands of paren- chyma, which are sometimes narrow, 1-2 layers in thickness, sometimes many-layered, and which separate the bundles from one another. The middle of the star also consists mainly of thin-walled parenchyma, often with lacunz, which is directly continued into the radial bands of the ring, and in which bundles of sclerenchyma, each containing one or more vessels and phloem-groups, lie scattered. Inside each sclerenchymatous bundle the vesse]s are surrounded by 1-2 layers of parenchymatous cells, those of them which belong to the ring standing in direct connection with the many-layered pericambium, An endodermis, which is thickened here and there, appears according to Mohl’s figure to surround the star. Finally, in the entire parenchyma, both of the star and of the cortex which surrounds it, numerous small bundles of sclerenchymatous fibres lie, each enclosing in its centre 1-2 thin-walled elongated elements (perhaps sieve-tubes?). The xylem- plates in the ring appear short and irregular in cross-section, their radial arrangement and alternation with the phloem-plates is according to Mohl’s figure often indistinct, though in general to be recognised. The development of the elements, both in Iriartea (Karsten) and in the roots of Pandanus, begins at the periphery of the ring, and in general proceeds centripetally. According to all these phenomena, the series of large roots just described are immediately connected with the type of Monocotyledons as special cases, in which the anatomical differentiation becomes more varied, with the more considerable size. The system of bundles, which traverses the tuberous roots mentioned at Pp. 233, is entirely different in structure from the bundles last mentioned. In Dioscorea and Sedum all the bundles are typically collateral. The same holds good for the Ophrydez, with the limitation that the vessels are only very sparingly developed. Each bundle is surrounded by a separate endodermis, 4. In the Z7ices in the widest sense, the Marsiliaceze, and the Equiseta, with a few exceptions to be mentioned below, the axial cylindrical bundle of the roots does not deviate in its differentiation from the types hitherto regarded*. Its xylem is in the great majority of cases, with the exception of the Marattiacez, diametrally diarch, beginning externally on each side with some narrow, fibrously-thickened tracheides lying side by side, which are succeeded in the centripetal direction by one or a few rows of wider, often large scalariform tracheides, of the structure usual in Ferns ; (true vessels only occur in Athyrium Filix femina *). Cf. Fig. 169. In Botrychium, the * Compare Nageli und Leitgeb, van Tieghem, Russow, 7//. cc. * Compare p. 165. STRUCTURE OF RADIAL BUNDLES, 363 tracheides, which form several rows, are of a different structure, similar to that described in the case of the stem and leaf at p. 346, and are all of approximately equal and relatively small width. Triarch and tetrarch bundles sometimes occur in thick roots of species, which are usually diarch; triarch-bundles have been observed in Pilularia, Equisetum, Botrychium, Blechnum brasiliense, and Cyathea medullaris, tetrarch in Equisetum, the Blechnum above mentioned, and Cyathea. In the species of Trichomanes? investigated triarch to octarch bundles usually occur, diarch bundles being rare, while, on the other hand, the latter are characteristic of the roots of Hymeno- phyllum. On the monarch bundles of some species of Trichomanes, see below. FIG, 169.—Adiantum Moritzianuin (225). Old root, cross-section. 4—/: hairs of the epidermis cut through, 1 endodermis, #c pericambium, 7 primitive tracheides of the diarch xylem alternating with two phloem groups. The xylem-plates are in most cases united in the middle, in the thinner bundles often by means of a very large vessel (e. g. Equisetum), or of a row consisting of two large vessels, crossing the diametral pair of plates at right angles (Fig. 169). In other respects various subordinate differences of form occur, e.g. a regularly elliptical cross-section of the diametrally united plate in Osmunda, Todea, &c., &c. The arrangement of the phloem-groups corresponds to the general plan of ? Mettenius, Hymenophyllaceen, /. c. p, 420.—Russow, Z.c. p. 95. 364 PRIMARY ARRANGEMENT OF TISSUES. root-structure; their histological peculiarities are essentially similar to those of the stem of the same plant, and like the latter still require more exact investigation. The pericambium appears, as a rule, as a single layer all round, but it also occurs with two layers; this is the case only outside the phloem-groups in Aspidium Thelypteris, and all round in Polypodium ireoides*; in Osmunda and Todea there are several layers all round. In the Equiseta, in contrast to the other forms belonging to this series, all the cells of the pericambium stand precisely in front of those of the endodermal sheath, and together with these have arisen from the division of the innermost cortical layer. In the other vascular Cryptogams the latter forms the endodermis only, while the pericambium originates by tangential division of the plerome-cylinder surrounded by the cortex. In the endodermis, apart from various subordinate differences of form, the cells. lying in front of the corners of the xylem-plates are the initial cells of the lateral roots, and are often distinguished from the others by their more considerable size. The structure of the endodermis is otherwise essentially the same as in the bundles of the stems of the same plants. In the roots of Cryptogams, the orientation of the diametrally diarch xylem- plates is always such that their surface cuts the median plane of the next higher order of ramification at right angles. Those arising on the stem appear, according to species, either to have the like orientation with reference to its median plane, or to have their surface coincident with the median plane of the stem. The axial root-bundles of the Marattiacez ” are distinguished from those of the other Ferns, with which their structure otherwise agrees, by their tetrarch or polyarch xylem*. The number and length of the radial plates increases, in the same species, with the thickness of the roots; the former may amount to 18-20. In the thicker roots we frequently find them converging in pairs, and as seen in cross-section united to form the figure V. In the roots occurring above the ground the xylem-plates do not reach to the middle of the bundle; in the thin, 4-5 rayed branches underground, they meet, according to Russow, in the middle. The very thin root-bundle of Azolla*, which differs in its development from that of the Ferns, has, according to Strasburger, a usually triarch xylem, consisting only of spiral tracheides. Besides this there are only some inconspicuous elements lying inside the pericambium, and constituting a doubtful phloem. Sct. rog. A structure departing from the general radial type of root occurs in the rhizophores of Selaginellz *, in the true roots of the same plants, in the thinner roots of Lycopodia, and in the roots of Isoetes and Ophioglossum: with the excep- tion of the rhizophore of Selaginella Kraussiana, the peculiarity of this structure consists in the fact that the usually monarch xylem either occupies one side of the bundle, the phloem lying on the other—the arrangement being thus collateral—or that the former is at least strongly approximated to one edge of the phloem which surrounds it. Most roots or rhizophores belonging to this series are dichotomously 1 Van Tieghem, 2. c.—Compare also Nageli und Leitgeb, p. 83. * Meyen, Haarlemer Preisschrift (1836), Tab. VII. ° (Cf. Holle, Kénigl. Ges, d. Wiss. zu Gétt. Jan. 8, 187 76.] * Strasburger, Ueber Azolla, p. 48. * [M. Treub, Recherches sur les organes de la Vég. du Selaginella Martensii. Leyden. 1877.] STRUCTURE OF RADIAL BUNDLES, 365 branched, and show a definite orientation of the parts of the bundle in the successive bifurcations. The structure in question might therefore be regarded as characteristic of dichotomous roots, if it were not that those of Ophioglossum are always entirely unbranched’, for there is no basis of fact for van Tieghem’s supposition, according to which this unbranched root would be the favoured branch of a root which has already undergone bifurcation while still inside the cortex of the stem which pro-. duces it, its other branch not coming to development. Among the dichotomous rhizophores? and roots of the Selaginelle, in the first place, the rhizophores of S. Kraussiana are distinguished by cylindrical vascular bundles,. in which the middle of the central and centrifugally developing xylem is occupied by the narrow primitive tracheides, while the periphery is formed of wider scalariform tra- cheides. The phloem completely surrounds the xylem as a many-layered small-celled zone; histologically it still requires more exact investigation*. To form the bundles of the first pair of roots, the bundle of the rhizophore is uniformly severed into two halves, in which a group of narrow primitive tracheides occupies one edge of the xylem, while the development of the elements proceeds from this point towards the other broader side. The xylem is thus monarch, similar to the usual collateral bundles, from which those in question are distinguished by the fact that the phloem completely surrounds the whole xylem. The structure last described belongs to all the investigated roots of Selaginella, and to the rhizophores of S. Martensii. These bundles also divide at the dichotomies in such a manner that the plane of division passes through the primitive group and the edge of the xylem lying diametrically opposite to it. In primary axes arising from the stem, the orientation of the bundles with a unilateral group of primitive tracheides is such that that group faces the base of the stem. In the dichotomous branches it always lies on the inner side, turned towards the other branch of the pair. At every bifurcation therefore each bundle proceeding from the division of the main-bundle undergoes a torsion of 90°, and indeed this takes place gradually inside the main axis, in such a way that the two bundles run on side by side from the point where they separate, which is above the point of bifurcation of the root, as far down as the latter. Only in the first dichotomous branches of the rhizophore of S. Kraussiana does the same orientation come about without torsion. : The feeble bundles in the seeks of Teaches show in the general structure of their uni- laterally monarch xylem, and in its orientation in the dichotomous branches, the same behaviour as those of Selaginella. As regards the elementary composition of this part they are distinguished by the fact that it consists only of a few rows of annular and reticulated tracheides, without scalariform vessels. So far as investigations extend, the phloem is feebly developed, and confined to the side remote from the primitive tra- cheides ; as seen in cross-section it has the form of a narrow crescent-shaped band, wh ile its histological structure is indistinct. The position of the bundle in the root is from the first slightly eccentric, and in the dichotomous branches always approaches the other branch of the pair. The eccentricity, which is caused by a mainly one-sided extension of the cortex and of its large air-cavities, increases with the thickness of the roots. According to Mettenius’ short statement ®, the structure of the xylem and the eccentric position of the bundle in the roots of PAy/loglossum are similar to those in Isoetes, The bundle approaches that side of the always unbranched root which is basiscopic with reference to the stem. 1 Compare Holle, Botan. Zeitg. 1875. Holle’s observations are not in agreement with van Tieg- hem’s statement that the roots of Botrychium, with typically radial bundles, are dichotomous (p. 315). ' 2 Nageli und Leitgeb, /.¢. p. 124. 3 (Cf. Treub, 2. ¢.] * Hofmeister, Beitrage z. Kenntniss d. Gefasskryptogamen, I.—Nageli und Leitgeb, 2. ¢, p, 131. 5 Botan. Zeitg. 1867, p. 99. 366 PRIMARY ARRANGEMENT OF TISSUES. The thin roots of Lycopodium, already described at p. 350, are immediately related to the above. . Lastly, the roots of Ophioglossum are to be mentioned here 1, In the circular cross- section of the axial bundle it may be seen that the half, which with reference to the parent axis is basiscopic (lower), is formed of tracheides, which are united together without inter- cellular cavities, and are similar to those of the stem (p. 346). The upper edge of the bundle is formed by a half-ring, usually two layers in thickness, of relatively large, wide sieve-tubes. Between this phloem and the xylem lie some layers,—numbering on the average three,—of delicate prismatic, narrower elements, destitute of starch, the sieve-tube nature of which is doubtful; as a rule a layer of delicate cells separates the xylem from the endodermis, while the sieve-tubes border immediately on the latter. According to van Tieghem the last statement often holds good also for the two middle tracheides of the lower edge. The development of the tracheides begins at one corner of the segment of the circle, and proceeds from this point round the convex edge, and from this again towards the phloem. The endodermis, like that of the stem, only differs from the other cortical parenchyma in the undulating bands on its radial walls. 4. Imperfect and rudimentary Bundle-trunks. Sect. 110. The vascular bundles described above occur in land plants possessing - foliage which is rich in chlorophyll, and also in the stems and leaves of parasites which contain no chlorophyll, or only traces of it, as in the Orobanches, Cuscutas, Lennoaceze, &c. As indicated at p. 321, they are ce/eris paribus, as a rule, all the more developed the greater the development of the leaf-surface. Conversely, the development of the vascular bundle-system diminishes in every respect with that of the leaf-surface exposed to the air; and this is the case firstly with regard to its differentiation into individual bundles and their branches, as is clearly shown by its reduction to an axial strand in the stems of many submerged plants (cf. pp. 277, 321), and by its simplification in the submerged leaves of am- phibious plants (p. 306); and secondly as regards the anatomical differentiation of the individual bundle. In the latter, while the plan of structure remains the same, a diminution of the essential tissue-forms may be recognised, as, for example, the com- parison of the bundles in the stem and root of Ranunculus repens (Figs. 152, 165) with those of R. fluitans (Figs. 153, 163) teaches ; and in fact the diminution is chiefly in the xylem, while the phloem remains the same, or is less reduced. Further, as the characteristic elements, and especially the trachez, progressively diminish, deviations from their usual typical arrangement also occur. Again, there may be complete dis- appearance of the tracheal elements, and finally of the sieve-tubes also, so that the entire bundle is replaced by a strand of uniform elongated cells. Lastly, we find the absence even of any rudimentary indication of a vascular bundle, as in the tiny swimming frond of the Wolffias, which is a large-celled mass of parenchyma, covered by an epidermis, which has stomata on the surface in contact with the air. These cases of imperfectly developed bundles are to be contrasted with the complete ones hitherto considered. They are divided into two main categories, namely, those which originate as complete bundles, and then by disappearance of the xylem become more or less incomplete—being thus bundles with a transitory * Van Tieghem, Russow, /.¢. * Hegelmaier, Lemnaceen, p. 31. RR IMPERFECT AND RUDIMENTARY BUNDLES. 367 xylem,—and secondly, those which remain imperfect from the beginning. Both forms are connected step by step with the complete bundles by various intermediate forms, which have already often been mentioned above. This especially holds good of those bundles which decome imperfect by disap- pearance of the Trachez. In many herbaceous plants with collateral bundles an intercellular passage appears in place of the primitive vessels when the development of the tissues is complete, as was described at p. 327. In a series of other plants, which are submerged or partly submerged aquatics, all the vessels in most of the bundles disappear at once throughout a long part of their course, after they have originated as annular or spiral vessels. In place of the . xylem an intercellular canal (filled with water) occurs in the mature bundle, and on its walls the remnants of this thickened membrane may remain preserved. On the other hand, the phloem of the bundles is persistent, and, in many of the cases in question, very well developed. These phenomena present many variations according to the particular cases, and are especially conspicuous in the stems of the Potamo- getons, and of the submerged plants connected with them, which have an axial bundle, or a very simple bundle-system?. Even in those among these forms in which, as in P. natans, distinct bundles of the leaf-trace and common bundles can be dis- tinguished, the want of vessels on the one hand, and on the other hand the crowded position of the bundles, often gives the whole bundle-system a structure which at the first glance is difficult to decipher, and into this we here have to enter somewhat more minutely. The course of the leaf-traces, and of the four cauline bundles in the stem of Pota- mogeton natans and perfoliatus, was described above at p. 272. All the bundles are at their origin collateral, with normal orientation. In the node all their parts are persistent, and soon become irregularly united by anastomoses. In the whole internode, on the other hand, the entire xylem disappears from the bundles of the leaf-trace with the beginning of the more intense elonga- tion, and is replaced by an approximately cylindrical narrow intercellular passage, bordered by narrow elongated cells”. In P. perfoliatus the same holds good also of the four cauline bundles; in P. natans, on the other hand, the few (1-3) reticulated and annular tracheze of the latter are usually persistent. The phloem-portions of all the bundles are very well developed and persistent. All the bundles are further closely ap- proximated to one another, being only separated by a few layers of parenchymatous cells containing abundant starch, and traversed by small groups of sclerenchymatous fibres. The bundles are grouped to form an axial strand, rectangular as seen in cross- section, which is marked off from the lacunose cortical parenchyma by an endodermis, which becomes sclerotic subsequently (Fig. 170). Inside this, one bundle faces each longer side of the rectangle ; opposite one of these two sides is a larger bundle, the sympodial one, descending from the second leaf above: facing the other side is a somewhat smaller bundle, the median one of the next higher leaf, belonging to the internode. One of the lateral bundles of this leaf faces each of the shorter sides ; the four cauline bundles face the four angles. A group of sieve-tubes lies at the 1 Compare p. 277, and the literature there cited. ? A. B. Frank, Beitr, z, Pflanzenphysiol. p. 135. 368 PRIMARY ARRANGEMENT OF TISSUES. outside of the wide intercellular passage of the sympodial bundle, while two some- what smaller groups stand symmetrically right and left of the centre of its inner side. In the remaining bundles the intercellular passage or vascular group is bordered on the outside by an arched group of sieve-tubes. In the other Potamogetons investigated (lucens, gramineus, densus, crispus, pec- tinatus, and pusillus), in Zanichellia, Althenia, Cymodocea, and Zostera, the trachez in the node are persistent, while those in the internode are all transitory. To every bundle an intercellular passage corresponds, which is surrounded on the outside by phloem. Where severa] bundles traverse the internode, they approach each other closely in a manner similar to that described for P. natans ; in the case of the two leaf-trace bundles in the internodes of P. lucens and gramineus this goes so far, that SOD Bint Tay TAY aL LETRA RI >< gS eee \> FA KO rr? CTs FIG. 170 (145).—Potamogeton natans. Axial mass of the internode, contaming the vascular bundles ; cross-section. 2 unilaterally thickened endodermis, containing starch; outside the latter lacunose cortical. parenchyma, with abundant starch; Z air-cavities. Explanation of the numerals at p. 272, The delicate groups of tissue of the numbered circles are the phloem; the wide meshes in the latter are the sieve-tubes of the bundles; the circles in which the numerals stand are the xylem portions, usually converted into cavities. Between the bundles is parenchyma containing starch, and sclerenchymatous fibres with a narrow lumen appearing as a darker. point, their intercellular passages, which are turned towards each other, are only separated by one layer of cells, or in most cases are united to form a single passage. The single axial sympodial bundle, which (without cauline bundles) traverses the internode in the upright stem of P. pectinatus and pusillus (Fig. 171), has, in the manner of a concentric bundle, a central intercellular passage replacing’ the xylem- group, and this is completely surrounded by a relatively bulky phloem, containing large sieve-tubes, and externally limited by the endodermal sheath, which eventually becomes sclerotic. Re The bundles in the stems of Zanichellia and Althenia behave quite similarly to the forms last mentioned, only with the difference that the phloem is very delicate and slightly developed, and consists of elongated cells with a few indistinct sieve-tubes. Elodea and Hydrilla, of which we shall speak later on, are also connected with these cases, ‘ IMPERFECT AND RUDIMENTARY BUNDLES, 369 The vascular groups of Cymodocea xquorea! and Zostera behave, both in the nodes and in the internodes, like those of the Potamogetons. The intercellular canal derived from the xylem lies in the small peripheral bundles on the inner side; outside this is a radially elongated phloem-group containing two or three large sieve-tubes ; in the thicker axial bundle the intercellular space occupies the middle, and in the case of Cymodocea has at its periphery four sieve-groups placed cross-wise in the transverse section; in Zostera it is completely surrounded by a broad phloem, as in Potamo- geton pectinatus. : The small stems of the Hydrillex and of Aldrovanda vesiculosa consti- tute transitional forms between the bundles which decome incomplete and those which remain rudimentary. Elodea canadensis and Hydrilla verticillata have an axial bundle of es- sentially similar structure to that of Zani- chellia. The one or two axial tracheze present in the young rudiment of the stem, which send off a branch into each leaf, are incomplete from their first origin ; their walls are only thickened with seg- ments of rings, and disappear every- where—even in the node—on the com- mencement of active extension?. Ac- cording to Caspary, Aldrovanda shows an axial bundle of 8-9 annular trachee, which, together with their branches going to the leaves, are persistent in the nodes, but disappear in the inter- nodes during extension, and are here re- placed by a passage surrounded by ' delicate-walled elongated elements, which have not been: more minutely investigated. Of bundles which remazn rudimentary, those in the small stem FIG. 171.—Potamogeton pectinatus (80). Cross-section through an internode of the upright stem. g intercellular passage replacing the xylem which has disappeared; z unilaterally thickened endo- _ dermis; between ~ and g the phloem, with wide sieve-tubes. ’ Between x and the epidermis ¢ is the lacunose cortex; 4 bundles of sclerenchymatous fibres; @ a similar bundle, with a small group of sieve-tubes in the middle. of Ceratophyllum and Najas are im- mediately related to those just described. According to Sanio the former are bundles which are at all times destitute of vessels, and consist of a mantle of narrow sieve-tubes and elongated cells, between which lie several small intercellular pas- sages, each produced by the absorption of a row of cells*; and inside this mantle are some layers of parenchyma, which surround an axial passage derived from the absorption of a multiseriate strand of cells. In the case of Najas the mature structure is similar to that of Elodea; sieve-tubes have not been described, and are doubtful; the axial canal arises from the absorption of a row of meristematic cells, 1 Compare Bornet, Ac. * Caspary, Sanio, /.c. (p. 278). * Compare Frank, Beitr. p. 143, Bb 370 PRIMARY ARRANGEMENT OF TISSUES. A structure deviating from that described occurs in the bundles of the small submerged stems of the floating Utricularias, the Lemnacez, the Podostemacee, Vallisneria spiralis, and in the rhizomes of Epipogon and Corallorhiza, which inhabit humus, all of these likewise remaining rudimentary. In Utricularia vulgaris? an axial, approximately cylindrical bundle is present, and gives off a branch into each of the so-called leaves. Its smaller half, which with reference to the horizontally floating stem is the upper, consists of elongated prismatic cells, usually with flat ends and thick collenchymatous walls; the lower half consists chiefly of thinner-walled elements, namely, of wide sieve-tubes and numerous narrower prismatic cells. Near the boundary between the upper and lower half, but within the latter, next the centre, lies a single row of very long, wide tracheides, with their pointed ends applied to one another, showing alternately annular and spiral thicken- ing. In the small and younger specimens no other tracheides occur. In very thick stems, on the other hand, I found a second single or double row of annular tracheides, lying at the side of those mentioned, near the periphery of the bundle. These are similar to the former in structure, but only about half as wide. Their development seems to take place very late. All the tracheides are persistent, no intercellular spaces whatever are present in the bundle. The bundles in the frond of Lemna? consist in the native species of a thin row of annular tracheides, surrounded by one or a few layers of elongated cells. Spirodela polyrrhiza has several parallel rows of tracheides instead of one ; in Lemna valdiviana, on the other hand, the development of tracheides is suppressed. The absence of any vascular bundles in the Wolffias has already been mentioned above. The bundles of the Podostemacez, which are still much in want of accurate investigation, consist, according to Trécul’s short statement °, of a bundle of fibres and some small annular vessels, the latter being often absent in old stems and replaced: by a cavity* Vallisneria also requires more exact investigation °. The axial bundle in the rhizome of Corallorhiza contains in the middle two multiseriate strands of reticulated tracheides with narrow transverse meshes, and from these the simple bundles branch off for the distichous leaf-rudiments. It further consists of elongated, usually thin-walled, elements, which still require more ° exact investigation. In the rhizome of Epipogon the bundle consists, so far as investigations extend, only’ of uniform, moderately elongated cells, with oblique ends and thin walls. In the reots of those aquatic plants, which in the stem possess incomplete or small bundles containing passages, the bundles, which here, it is true, are always feebly de- veloped, may be complete, and provided with a persistent xylem constructed according = to the radial type, e. g. Potamogeton lucens*®. As a rule, however, they here also either become incomplete by disappearance of the vessels, interceliular passages appearing in their place, or they remain rudimentary. ' Compare van Tieghem, Ann. Sci. Nat. 5 sér. tom. X. p. 54. ? Hegelmaier, Lemnaceen, p. 48. 8 Archives du Muséum d’Hist. Nat. tom. VI. p. 4 * [Compare Warming on Podostemacez, Mém. Acad, Roy. Copenhagen, 1881, &c.; also Cario, Tristicha hypnoides, Bot. Ztg. 1881, p. 25.] ° (Compare Franz Miiller, die Entwickelung von Vallisneria spiralis, Hanstein’s Abhandl. Bd. III. Heft. 4, 1878.) * Van Tieghem, Ann, Sci. Nat. 5 sér. tom. XIII. p. 164, pl. VI. ENDS AND CONNECTIONS OF THE BUNDLES. 37t The former process goes on as a rule more slowly in the roots, the vessels persist longer, than in the stems of the same plants, The structure of the root-bundle (peri- cambium, alternating xylem and phloem rays) here preserves its typical character, although the number both of the individual rays, and of the elements composing each of them, is low, the former being reduced to four or to two, the latter often to one. The mature root has therefore the typical bundle-structure, with the exception that in the place of the 2-4 xylem rays, each of which is represented by a vessel, and in place of the large central vessel, which in many cases unites the rays, an intercellular passage is present ; e.g. Aponogeton, Alisma, Hydrocleis (van Tieghem). In the root of Elodea canadensis, as in the stem of the same plant, the disappearance of the 4-5 peripheral vessels, and of the larger central vessel, which is separated from them by an annular layer of parenchyma, takes place immediately after their first, incomplete formation. - Among those root-bundles which. on the other hand, remain rudimentary, those of Najas belong to this category. They consist of two layers of elongated delicate cells, and these surround an axial passage, which arises by the absorption of a row of meristematic cells. In the root of Vallisneria, according to van Tieghem, there is only an annular layer of elongated cells enclosing an axial passage, and surrounded by the endodermis, and this constitutes the rudiment of the bundle. The delicate bundle in the root of species of Lemna shows in cross-section essentially the same structure ; the middle is according to Hegelmaier occupied by a row of cells (not by a passage). Spirodela polyrrhiza shows the same character, but with the difference that the row of cells in the middle is developed into a persistent row of narrow annular tracheides, II. Enps anp ConNnecTIONS oF THE VaScuLAR BUNDLES. Sect. rrr. The ends of the vascular bundles, as was shown above at Sect. gt, lie in the foliar expansions, and in the cor/ex of many plants, either as zu/ernal ends ceasing in the parenchyma, or forming anastomoses, or as peripheral ends at the edge or surface of the leaves; in the stems of Cyatheaceze, described at p. 291, they also end in the interior of the pith. With the ultimate degrees of ramification the thickness of the bundles usually diminishes, both the number and the size of their elements being reduced; at the extreme free ends they often, but not always, show a terminal dilatation. Xylem and phloem do not behave alike in this respect. Clearly characterised sieve-tubes are, it is true, often still present in the thicker bundles of the foliage leaf, e.g. in those of the leaf-nerves ; in the last orders of ramification they are no longer to be found, the latter consist either of trachez alone, or of these and of delicate elongated cells accompanying them, the sieve-tube nature of which is no longer recognisable. Where and how the sieve-tubes cease and end has not hitherto been clearly made out in any case, and this point deserves more accurate investigation. The trachese always form the direct continuation of the xylem of the thicker bundles. The ultimate zz/ernal ends and anastomosing branches of the bundles consist only of one or a few rows of short tracheides with fine spiral thickening, or reticulate thickening with narrow transverse meshes; their wall is often quite smooth in parts, appearing as if immature, e.g. in the leaf of species of Chamedorea, and Zea Mais (Fig. 175). Whether vascular perforations occur in these rows is at least doubtful, and not easy to decide. 372 PRIMARY ARRANGEMENT OF TISSUES. The dilatation of the terminal branches above-mentioned comes about either by dilatation of the individual tracheides, or by increase in the number of rows. The end surfaces of the tracheides, bordering on the parenchymia, are usually cut off sharply, either transversely or obliquely. Characteristic differences corresponding to the several main forms of distribution of the bundles, or to the larger systematic divisions, are not to be observed, except that in general the thickness of the single terminal branches diminishes where the branches are numerous. Thus in the Ferns investi- gated, the relatively few final ramifications are comparatively thick, consisting of several rows, while in the foliage of Dicotyledons, where the reticulations are abundant and internal ends are present, the bundles are at last, as it were, broken up into single, or in places double tracheal rows, which end free with short branches (Figs. 172, 173). The transverse branchlets of most Monocotyledons consist of FIG. 172.—Psoralea bituminosa (40). Ultimate FIG. 173 (225).—Ultimate ramifications of vascular —' ramifications of the bundles in a piece of a bundles from the leaf-lamina of Psoralea bituminosa 5 leaflet; at is the edge of the latter. branched rows of tracheides, the ends marked x being torn off, the others terminating free. The whole branched row is surrounded by large cells containing chlorophyll; outside these are the circular transverse sections of some cells of the dense palisade parenchyma of the leaf, © one, or of quite a few rows of tracheal elements (Figs. 174, 175). The leaf of Welwitschia is to be mentioned as an exception to the rule, for its very numerous transverse branchlets (comp. Fig. 145, p. 303) have, so far as can be determined, the structure of complete vascular bundles provided with a multiseriate xylem and phloem, and it is only the short, thick, free-ending branches, springing partly from the angles of the cross-branchlets and partly direct from the longitudinal bundles; which consist exclusively of tracheides, the latter being inserted between the elements of the surrounding parenchyma. As follows from what has been stated, the ultimate vascular branches often border directly on the parenchyma which in the lamihz of foliage-leaves contains ENDS AND CONNECTIONS OF THE BUNDLES, 373 Chlorophyll. Its elements, in so far as they stand in immediate connection with the trachez, approach on the one hand the latter, on the other hand the typical paren- chymatous cells in their form. This is the rule for the leaves of Monocotyledons and Dicotyledons; the exceptional case in which even the ultimate transverse branchlets are completely enclosed by stout sclerenchymatous sheaths occurs rarely in thick leaves of Mono- cotyledons, e.g. Rhapis, Vanda furva, In the leaves of Ferns the branches of the bundles are, so far as investigated, always ensheathed by one or a few layers of elongated cells destitute of chlorophyll?, of which the outermost often has the: structure of an endodermis up to the immediate neighbourhood of the free ends. At the free ends themselves the rows of tracheides pass over into the chlorophyll-parenchyma, through the intervention of some elon- gated smooth-walled cells. In the. case of many ends of bundles, which according to their local position must be called fer7- pheral, no essential differences from the internal ends are to be mentioned. On the other hand, the structure shows peculiarities in those numerous cases, where the bundles run to parts of the epi- dermis described in Chap. I, which are distinguished by water-pores and water-filtration, by excretion of lime, or by glandular structure and secretion. Of the cases belonging to this series, in the first instance, the ends of the bundles in the furrows of Fern-leaves excreting water and lime (p. 106) are closely similar to the internal terminations in these FIG. 174.—Zea Mais, Cross-section through a feeble (lower) leaf- sheath, about 2cm, higher than Fig, 151, p. 331. ¢ epidermis of the outer surface, bordering internally on a hypodermal bundle of scleren- chymatous fibres, s¢, On the inside of the latter one of the smaller lon- gitudinal vascular bundles abuts; + annular vessel, x air-cavity, ¢—¢ pitted vessels; outside ¢,7,¢ are first narrow reticulated and pitted vessels (with a darker outline), then the phloem vv; ¢ transverse branchlet springing from the tracheze of the vascular bundle and consisting only of a few rows of tracheides. FIG, 175.—Cross-section through the lamina of the leaf of a young plant of Zea Mais. o epidermis of the upper, w of the lower surface; # hypodermal strands of sclerenchyma ; g1 and gz two small longitudinal vascular bundles in cross-section; 41 with three, ¢3 with two narrow vessels; both with a small phloem, consisting in g2 of only three elements; gs transverse connecting branch between gi and gz, consisting of a row of tracheides with partial fibrous thickening, and, like the two longitudinal bundles, enclosed directly in parenchyma containing chlorophyll leaves*, They show a knob-like swelling, in consequence of a sudden increase in the number and size of the tracheides, the latter being very short, with narrow 1 Cf. Mettenius, Fil. hort. Lips. p. 9. ? Mettenius, /. ¢. 374 PRIMARY ARRANGEMENT OF TISSUES. reticulations and pits, or with spiral fibres ; one or two layers of delicate cells ensheath the whole end of the bundle, and separate it from the thin-walled epidermis of the furrow. The ends of the bundles of the leaf-teeth of Drosera, and of the inner surface of the foliar pitchers of Nepenthes, are likewise placed close under specialised portions of the epidermis, and in point of structure stand next to those just described. The leaf of species of Drosera (in particular of D. rotundifolia) has at its edge and on its entire upper surface numerous filiform teeth with broadened ends’. Those of re aul iT ERRELEE TOCA UTM TOGTTC! — ] 1 FIG. 176 (145).—Drosera rotundifolia. End of a tooth from the upper side of the leaf; axial longitudinal section. s—s the bell-shaped layer directly sur- rounding the end of the bundle; its lowest, yery long cells, s, standing with their narrow outer wall inthe epidermal surface, belong, as shown by Warming, and as their appearance indicates, to the primary epidermis; all the others, accurring above the apex, proceed from the subepidermal meristem. The same holds goad of the cells of the second layer occyrring above the apex: in so fae as they border on the apical cells of the innermost layer they proceed from the sister-cells of the latter. Those, on the other hand, which belong to the lower edge of the knob, as wellas the entire gutermost layer, are derived from the primary epidermis, the surface are, apart from differences of length, similar to one another; they are filiform processes, somewhat conically tapered, but swollen at the end to form an approximately ovate head. They consist of some layers of elongated cells, in the middle of which are one or rarely two narrow spiral vessels or rows of tracheides (I will not state definitely which) branching off from the net of bundles of the leaf-lamina, and having a straight course ; the whole is covered by a single epi- dermal layer, the cells of which are likewise elongated. In the middle of the knob-shaped end the spiral vessel enters a group of closely-connected, short, reticulated and spiral tracheides, which has, as a whole, an ovate form, and constitutes the main bulk of the terminal portion. The layer of cells surrounding the ves:els terminates below the middle of the group of tracheides, in the form represented in Fig.176. At the point of transition to the head the epidermis first becomes short-celled, and then suddenly passes over into the three-layered covering of the sur- face of the knob, which, as Warming has shown, is derived partly from the primary epidermis, partly from the layer of meristem lying below the latter. The innermost layer of this covering forms a bell-shaped single stratum, consisting mostly of elon- gated cells, which is in immediate contact above and at the sides with the group of tracheides, while below, at the edge of the bell, it ends in the outer surface of the epidermis. The membranes of its cells are smooth and firm, similar in their reactions to those of an endodermis, the walls vertical to the surface being undulated. From the edge of the bell onwards the surface of the knob is covered by two layers of cells, which are thin-walled, and are distinguished in the fresh condition by their dense, intensely red contents. The inner of these two layers does not reach quite to the edge of the bell, and consists of small, isodiametric, polyhedral cells, which are very delicate, and are in uninterrupted connection one with another. It is everywhere covered by the outermost layer, which is continued immediately from the edge of the bell over the whole surface of the knob, and consists of polygonally prismatic cells in uninterrupted connection. The diameter of these cells ver- tical to the surface increases successively towards the apex of the knob; here it is about twice as long as the diameters coinciding with the surface, while at the base it is about equal to them. The delicate outer walls of these layers, which are covered later on with the sticky secretion (p. ror), show a very delicately undulated outline at the edges.—The teeth of the edge of the leaf are expanded at their ends to the form of a spatula or long * Compare Meyen, Secretionsorgane, p. 51.—Trécul, Ann. Sci. Nat. 4 sér. tom, III.—Nitschke, Botan. Zeitg. 1861, Nos. 22, 23, &c.—Martinet, Ann. Sci. Nat. 5 sér. tom, XIV.— Warming, /.¢.5 compare above, p. 57.—Darwin, Insectivorous lants. ENDS AND CONNECTIONS OF THE BUNDLES, 375 spoon, and have on the upper, somewhat concave surface of the latter the same three superficial layers as the middle teeth, while the edge and under-side consist of an ordinary one-layered epidermis. Under the three-layered portion of the surface lies a group of tracheides, corresponding to it in position, which are of the same structure as those of the middle teeth, and are connected with the network of bundles of the lamina by means usually of 2-3 spiral vessels. On the inner surface of the pitcher of Nepenthes branches 1-2 vessels thick, coming obliquely from the network of bundles, end directly beneath the basal part of the digestive glands considered at p. 101, but by no means beneath all of them. Most of these organs, on the other hand, have no bundle-ends extending to them like those of other plants mentioned above, Most bundles, which terminate beneath water-pores or glandular parts of the epidermis *, consist of rows of tracheides, perhaps also of vessels, which run parallel towards the point of termination, diverging directly beneath the latter at greater or less angles, and then ending blind. The structure of the terminal elements is the same as in the internal ends; in species of Crassula reticulated tracheides with unusually large meshes occur. In all cases investigated, except Crassula, rows of delicate, smooth-walled cells, elongated in the same direction as the trachez, lie among the latter without any definite arrangement. These become more numerous ne Fig. 177. Fig. 178. FIGS. 177, 178.—Primula sinensis. Fig. 177 (40). Outline of a tooth of the leaf with its branches of the bundles. The thick- est of the latter ends below the water-pore #.—Fig. 178 (145). ‘Longitudinal section vertical to the surface of the leaf, through the middle of a similar tooth. o upper, # lower surface of the leaf; 4 water-pore, below which is the air-cavity, and then the epithema, the cells of which are drawn somewhat too large. It contains a little chlorophyll throughout, and the bundle of tracheides, g, ends in contact with it. Under the epidermis on both sides is parenchyma containing abundant chlorophyll. as the tracheze between them terminate or diverge, and pass over gradually into a group of small delicate cells, which covers the ends of the vessels; these are in their turn immediately covered by the epidermis, and with reference to the former relation may be called the Cover, Zpzthema”, of the end of the bundle. Either a single 1 Compare p. 50, § 8, and p. go. Most of the literature there cited refers also to the subject treated of here. [See also Gardiner, on the development of the water glands in Saxifraga crustata, Quart. Jour. Micr. Sci. July, 1881.] 2 éni@nua, the cover. 376 PRIMARY ARRANGEMENT OF TISSUES. bundle ends in one group of epithema, or two or more converge, SO as to end in a common group. The former, for example, is the case in the large leaf-teeth of Fuchsia, Primula sinensis (Figs. 177, 178), and Cucurbita, where the bundle-end is a thick, short strand, derived from the union of several convergent bundles inside the edge of the leaf; also beneath the furrows of the surfaces of the leaf of species of Crassula, in the glandular spots of the surfaces of the leaf of species of Malpighia, &c. The latter is the case in many teeth and crenations of the leaf, especially the broader ones, e.g. Brassica!, Papaver, Tropzolum (Fig. 179), and many others. The arrange- ment of the parts is here such that one bundle coming from the middle of the leaf,. and one or more marginal ones on each side, converge towards the epithema, and end at its circumference. Both cases, namely common and special epithemata, often occur side by side, e.g. in leaves of Crassula, in the marginal furrows of the leaf of Saxifraga Aizoon, elatior, and their allies (comp. Unger, /.c.) FIG. 179.—Tropzolum majus (40). Course of the ends of the vascular bundles in a piece of the edge of the leaf at its median indentation, At e epithema, above which lie the water-pores figured at p. 52, fig. 19. In the furrows or spots (mentioned in Sect. 8) of the surface of the leaf of certain species of Ficus, which have incorrectly been supposed to be without vascular bundle- terminations, in the glandular ends of the petiolar appendages of Passiflore, Mal- pighiaceze (e.g. Stigmaphyllum), and Amygdalez, and in the glandular prominences or depressions of Acacias, divergent rows of trachez with short articulations run into an epithema lying below the epidermis. These rows may belong, according to the particular case, to one or more bundles, or in several cases may equally well be considered as one bundle or as several. In the furrows of the leaves of Ficus, according to investigations on F. neriifolia and diversifolia, a disc-shaped group of epithema lies below the epidermis, either over a knot of the network of vascular bundles, or over a single bundle ; in the former case the vascular bundles meeting in the knot, break up, as it were, towards the side of the furrow, into numerous short rows of tracheides directed towards the latter; in the second case a bunch of tracheides branches off from the bundle and enters the epithema. ? Compare Unger, /.c. (above, p. 327), Taf. 2, fig. 17. ENDS AND CONNECTIONS OF THE BUNDLES. 377 The glandular spots on the under side of the leaf of Prunus Laurocerasus also lie over a knot or a narrow mesh of the net of bundles, and from this some vessels or tracheides, which are not numerous, branch off into the epithema lying below the glandular epidermis. In the petiole of Passifora cerulea and its allies a bundle ending below the epithema enters the cylindrical appendages or teéth, which end with a concave glandular surface. The conditions are similar in the appendages of the petiole of Amygdalee. Several bundles running to the glandular terminal surface enter the broad round appendages of the petiole in Stigmaphyllum., . In the Acacias the glandular spots of the appendages of the petiole behave very dif- ferently according to the species (cf. p. 98). A number of small bundles, here and there reticulately connected, enter the elongated, wart-shaped projection of the base of the petiole of A. lophantha, running towards the free surface, and here ending in the epi- thema. Below the flat wart-like prominence on the upper edge of the base of the phyllodes of A. marginata, and A. calamifolia, numerous isolated short trachee branch off from the neighbouring strands of the net of bundles, and without being united into distinct bundles turn towards the epithema and there end. The same is to be seen here and there at the base of glandular pocket-like depressions of A. latifolia and its allies, but the vascular ramifications are here sparse and very short, while, on the other hand, the thick xylem groups of the bundle-net often border directly on the epithema. As regards the characteristics of the group of epithema, in many cases it scarcely deserves a special name, as it is nothing but a small-celled parenchyma, which, on the one hand, passes over immediately into the other larger-celled parenchyma of the organ, and, on the other hand, into the interstitial cells of the end of the bundle. So, for example, in the glandular appendages of the petiole of the Passifloree, and in most ends and teeth of leaves. Here the epithema is distinguished from the lacunar chlorophyll-parenchyma by the smaller size of its cells, and by their containing little or no chlorophyll, the places occupied by the epithema thus differ- ing in their pale colour from the green surface of the leaf. The epithema passes over quite gradually on all sides into the large-celled chlorophyll-parenchyma ; the water- pores of the epidermis lead immediately into the intercellular passages of the latter. In the leaf-teeth of Papaver orientale even all transitional forms are present between the small parenchymatous cells ofthe epithema and the tracheides of the bundle-ends. These epithemata have a very different form and extent, according to the shape and size of the bundle-ends and teeth of the leaf; in the narrow ends and teeth of the leaves of Fuchsia, Callitriche*, and Primula sinensis, for example, they are quite small bodies, showing only a few cells in section, and lying immediately below the large stomatal cavity, which belongs to the water-pore, situated at the end. In the broad teeth of the leaf of Papaver and Brassica, and in the crenations of Tropzolum, the group of epithema, which receives several ends of bundles, is a many-layered small-celled parenchymatous mass, approaching 1™™ in breadth. On the other hand, however, many ends of bundles run out into epithemata, which are sharply distinguished and limited. The furrows of the leaves of species of Ficus, Crassula, and Saxifraga, the glandular petiolar appendages of the Acacias, and many other cases, are examples of this. At the parts of the leaves indicated at p. 51, of species of Crassula, and of Rochea coccinea (Figs. 180-182), a thick bundle runs 1 Borodin, /.c.; compare p. 51. 378 PRIMARY ARRANGEMENT OF TISSUES. vertically to the surface under each epidermal furrow, ending some distance within the latter in a conical or hollow-conical expansion. Upon or in this expansion rests an epithema, of oval or elongated form according to the species, and extending to the epidermis, which contains water-pores. Its cells are on the average about one fifth the size of those of the surrounding chlorophyll-parenchyma, they are roundish or slightly elongated in the same direction as the vascular elements, with watery, colour- less contents. They are almost uninterruptedly united with each other, and even the cavities below the water-stomata are small. In the furrows of the leaf of the Saxifrages mentioned above, the end of the vascular bundle is expanded into a large epithema of an approximately conical form, with its base resting on the epidermis of the furrow. Its structure is very similar to that of Crassula, its cells are elongated in the same — direction as the tracheze, and the whole body, like the vascular bundles themselves, is ensheathed by a layer of cells which are very rich in tannin. The epithemata in FIG, 180.—Crassula arborescens. Longitudinal section, vertical to the leaf-surface, through the apex of a leaf. Magnified 30—40. e—e/ epidermis; g vascular bundle, divided into two branches, which terminate below a small- celled epithema in broadly conical ends consisting of short tracheides; the branch at on the upper, that at p’ on the lower surface of the Jeaf. the furrows of Ficus have an approximately discoid general form, are round-celled, aud in other respects are also similar in structure to those of Crassula. The same also holds good in general of those epithemata which lie below the glandular portions of the epidermis. How far the nature of the contents of their cells shows remarkable peculiarities, still remains to be more exactly investigated. Sect. 112. In the leaves of the Conzfer@, as stated above, the finer ramifications of the bundles are absent; the leaves are traversed either by a number of bundles of approximately equal thickness, or in most cases by a single median one; in most Abietinee by a median pair of bundles running close side by side, only. separated from one another by one or two layers of elongated cells (e.g. Abies excelsa, pectinata, Pinsapo; Cedrus Libani; Pinus Pinaster, Laricio), or by a thick strand of sclerenchymatous fibres. The bundles are collateral, and their orientation is ENDS AND CONNECTIONS OF THE BUNDLES, 379 normal. ‘Towards the end they are tapered, and the xylem and phloem diminish in such a manner, that here also the ultimate termination consists only of one or a few rows of short tracheides. They are distinguished by the fact that within the lamina of the leaf, as if to replace the finer branches, the edge of the xylem is expanded throughout its entire length into a border, consisting of rows of short tracheides inserted in the parenchyma of the leaf. This fringe of tracheides, which was first accurately described by Frank in the case of Taxus’, and afterwards more generally demonstrated by Mohl?, is absent, or at least extremely feeble, in Larix europzea alone of forms known to me: it arises, in the case of the Abietinez, exclusively from the outer, remote edges of the pair of bundles, in the other cases investigated from both sides of each of them. » SY cory ry i Ss J Fig, 181. Fig. 182, FIGS. 181, 182.—Rochea coccinea (200). Fig. 181. Fragment of the epidermis from the edge of the leaf. S water stoma; s air stoma, with subsidiary cells. The scattered dots are wart-shaped projections of the outer walls.—Fig. 182. Section through the edge of the leaf vertical to the surface. e—e epidermis, S water stoma, subsidiary cell, 4 somewhat thick vascular bundle in cross-section. The meshes with thicker and double outlines are the sections of bundles of trachez which run to neighbouring bundles; the more delicate outlines are those of the accompanying elements. A short strand goes off from 4 and runs towards 5; the tracheides of which it consists diverge and embrace the delicate-celled group of epithema lying between 4,2, and S. All round is large-celled chlorophyll-parenchyma. It is attached to the xylem by means of one or two longitudinal rows of tracheides, more or less frequently interrupted by parenchymatous cells, and from here projects on each side into the surrounding parenchyma; in most of the species it has the form of a plate, which is either plane or a little curved, and approximately follows the direction of the leaf-surfaces (Taxus, Cephalotaxus, Torreya, Taxodium semper- virens, Cunninghamia (Fig. 183), Juniperus (Fig. 184), Thuja, Thujopsis, Gingko), or is curved from each side round the body of the vascular bundle, and is separated from the latter and from the plate on the other side only by a few rows of parenchymatous 1 Botan. Zeitg, 1864, p. 167, Taf. IV. 2 Ibid. 1871, p. 10. Mohl calls the tracheide-border Transfusion-tissue. (See also Zimmer- mann, tiber das Transfusions-Gewebe, Flora, 1880, p. 2.] 380 PRIMARY ARRANGEMENT OF TISSUES. cells. The border is in fact curved round the xylem in Sciadopitys, Araucaria brasiliensis, Cryptomeria, and Dammara ; round the phloem in Abies pectinata and Pinsapo. In Abies excelsa and the Pines (P. silvestris, Laricio) it is split as it were into two plates on each side at its place of attachment, which, in a manner still to be more accurately described, are bent, the one round the xylem, the other round the phloem, so that the pair of bundles is completely surrounded by the border of tracheides. Sane sa The plates of tracheides are in many cases, especially in Podocarpus Meyeriana Endl., of nearly equal thickness throughout in every cross-section; in the other 7 et Ul FIG. 183.—Cunninghamia sinensis. Cross section through the leaf (220). # lower, o upper surface; # resin passage; xs sclerenchymatous fibres ofthe hypoderma, s those scattered in the parenchyma; ¢ xylem of the median bundle; ¢tracheide border of the latter. Below, towards the resin-passage, is the thin-walled phloem; the white band at its boundary towards the parenchyma surrounding the resin passage is the compressed primordial tissue of the phloem; g transversely elongated parenchymatous cell of the middle of the leaf. cases mentioned, except Abies excelsa and the Pines, they are thicker on their outer edge, i.e. that remote from the vascular bundle, than on the inner attached edge, in consequence of increase in the width and number of layers of their elements ; this often takes place to such an extent that the cross-section becomes wedge-shaped, e.g. Taxus and Podocarpus Thunbergii. The tracheides of the border are in general arranged in fairly regular rows both in the direction of the length and of the breadth of the leaf; these rows often show interruptions, which are filled up by parenchymatous cells, but all are at some points in immediate connection with one another. At the inner edge, which is attached to the xylem of the bundle, their form is similar to that of the tracheides of the latter, ENDS AND CONNECTIONS OF THE BUNDLES. 381 namely elongated; but they are on the average shorter and wider, and have terminal surfaces, which are but little oblique, and may even be horizontal. As their distance from the inner edge increases, their length rapidly diminishes, while their width increases, so that in the outer part of the border they are not longer, and are often even shorter than they are broad, being similar in form and size to the neighbouring parenchymatous cells. These conditions appear in a special form in Podocarpus Meyeriana, Thuja gigantea, and in the Pines and Abies excelsa. In the two first-named the border is very broad, it projects deeply into the middle of each half of the leaf in the form of a flat wing. Its tracheides, with the exception of the innermost, are, with reference to the diameter of the leaf, much broader than long, thus having their greatest diameter directed towards the edge of the leaf; they form rows running towards the latter, FIG. 184.—Juniperus communis (225). Median vascular bundle of the leaf: g xylem; c isolated sclerenchymatous fibres at the outer border of the phloem; ¢ border, consisting of tracheides with bordered pits and cross beams, The parenchymatous cells occurring near or between the latter are dotted in a granular manner. which are often interrupted, but are in contact with one another, and might be termed a narrow-meshed net of uniseriate vascular bundle terminations. In the connate sheathing-base of the flat pair of leaves of Thuja gigantea, the border of tracheides of each leaf is expanded into a low wing, which runs to meet that of the opposite leaf, and unites with it to form a transverse girdle. In the Abietinez last-mentioned?, the pair of vascular bundles lies in an approxi- mately cylindrical central portion of the leaf, which is free from chlorophyll, and is se- parated from the surrounding chlorophyll 5, j— Mohl, Verm. Schriften, p. 265, Taf. X. * Compare Kraus, Cycadeenfiedern, /.c.—Von Mohl, Botan. Zeitg. 1871, p. 7.—Thomas, in Pringsheim’s Jahrb. /. c. SCLERENCHYMA AND SCLEROTIC CELLS. 419 Fig. 27, p. 78). They form a thick uninterrupted layer on the upper side of the leaf of Jacquinia ruscifolia, an often interrupted layer on the lower side of this leaf, and: on both sides of that of Theophrasta ornata and species of Olea, This form of the strengthening apparatus is absent, however, in most leathery leaves, and in fact this is the case even in such species as are closely related and similar to those mentioned, e.g. most leathery leaves of Orchids, Pholidophyllum zonatum, Zamia integrifolia, Taxus, Cephalotaxus spec., Tsuga Canadensis, Abies amabilis, &c. A continuous two- or many-layered fibrous investment further appears in the aerial roots of Philodendron Imbe, Rudgeanum, and other species’. In the Cyperacez, e. g. species of Carex, the outer layer of the cortex of the root is often sclerotic in a high degree, and throughout many strata of cells. The second form of distribution of sclerenchymatous masses is that in which they lie at a greater distance from the epidermis, in the inner regions, being united so as to form either a continuous annular layer or isolated strands. The former arrangement occurs in a number of stems, in such a form that the fibrous ring lies in the external cortex, bordering within and without on parenchyma- tous layers: shoots of Berberis vulgaris; Caryophyllee, as Dianthus plumarius, Gypsophila altissima, Silene Italica; Cucurbitacee; and climbing Aristolochiz, as A, Sipho*. Usually, however, the fibrous layer lies on the outer border of the bundle-ring or cylinder, in such a manner that it includes the vascular bundles, or in the latter case the outermost of them, or that they rest against it. In this case, and in Berberis*, and doubtless also in the other plants mentioned with it, the ring belongs, according to its origin, to the outer layer of the plerome, it marks, more or less ‘sharply, the outer doundary of the plerome. This phenomenon is of most frequent occurrence in Monocotyledonous stems. It occurs in the halm of many Grasses and of many Cyperacez and Juncaceze, and in fact sometimes in combina- tion with the occurrence of hypodermal fibrous ridges, which, as processes of the ting, unite the latter with the epidermis, e.g. Piptatherum, Molinia, Bromus spec. *; or penetrate from the outside close up to the ring without reaching it (Alopecurus pratensis, Panicum turgidum, Juncus paniculatus); or, lastly, show sometimes one, sometimes the other condition (Cladium Mariscus). Other plants belonging to the _ families mentioned show the sclerenchymatous ring connected only with isolated ridges of the epidermis, or without this connection, and only with projecting ribs, which correspond to the insertion of vascular bundles, on its outside; e.g. Rhynchospora alba, Juncus bufonius, Pennisetum longistylum. To the latter cases is related the smooth and sharply limited sclerotic ring, which, with numerous individual modifica- tions, forms the outer boundary of the cylinder, and includes or supports the peripheral vascular bundles in most Monocotyledonous foliage-stems: Restiaceze, Eriocaulonez partly, Commelinez, Melanthacez, Liliaceze, Smilaceze, Tamus, Iridez, Orchidez, Alismacez, Typhaceze, &c.; and in most rhizomes, also belonging to the families mentioned above. The same phenomenon of a sclerenchymatous ring directly supporting or * Van Tieghem, Struct. des Aroidées, /.¢. * Compare Treviranus, Physiol. I. p. 209.—Caspary, Pringsheim’s Jahrb. I. p. 444. —Sanio, Botan. Zeitg. 1864, p. 222.—Von Mohl, Palm. Struct, Tab. H.—Mettenius, /. ¢. * Compare Schmitz, /.¢. (p. 393). * Compare Schwendener, /.c. Ee 2 420 PRIMARY ARRANGEMENT OF TISSUES, including the bundles, recurs in numerous Dicotyledonous stems: Caryophyllez, e.g. Silene catholica, Saurureze, Podophyllum, species of Thalictrum (also in the petiole), Papaver, Plantago, Trientalis, Hypocherris radicata, &c.* The converse case, that a continuous layer of sclerenchyma supports the whole inner side of the ring of vascular bundles, is rare in Dicotyledonous stems. This is the case in woody Piperaceze—Artanthe, Chavica spec. 2, and is especially characteristic in the shoots of Bougainvillea spectabilis. In tough firm organs, longitudinal fibrous strands, which are not distributed in the hypoderma, and which in the whole or a part of their course stand in no direct relation to the vascular bundles, are more frequent than continuous annular layers, Examples of this are afforded by the cylindrical or prismatic strands in the parenchyma of the leaf and petiole of Marantacez, Palms, Draceen, and Pandanus, which, as regards their position, are connected with the hypodermal strands by various transitional forms ; e.g. the strands in the interior of the leaf of Welwitschia (Fig. 187), and in the cortical parenchyma of Ephedra, the strands in the internodes of many Potamogetons, mentioned at pp. 232 and 271, and represented in Fig. 171, p. 369; the little bundles occurring in the parenchyma of the stem-cylinder of many Palms (As- trocaryum, Cocos, Leopoldinia, Lepidocaryum spec. *) between the vascular bundles and those which traverse the cortex of most Palm-stems; the numerous strands. in the rhizome of Acorus‘, in the cortex of the root of Phoenix, Cocos spec.*, and of the Pandanez, and in the axial cylinder of thick roots of the same (comp. p. 361) and of the Iriarteze. In many of the cases mentioned, the fibrous strands have an isolated course, without connection with the fibres accompanying the vascular bundles; e.g. in the roots mentioned, in Ephedra and Welwitschia; also in the leaves of Draczena, as far as my experience goes. The leaf of D. reflexa °, as seen with the naked eye, is traversed longitudinally by more than thirty nerves, of which about eighteen are stronger and darker, anastomosing here and there by means of fine transverse branchlets; these are the vascular bundles lying in the middle lamella of the leaf (comp. p. 320), and paler nerves alternate with them. The latter are simple small bundles of fibres. They do not lie in the middle lamella of the leaf, but below the two surfaces, separated from the epidermis by one or two layers of chlorophyll-parenchyma. In the interval between two vascular bundles they usually lie 3-5 together (next the edge 1-0), so that the total number amounts to nearly fifty, which cannot all be clearly distinguished with the naked eye. The smallest are only 5-7 fibres thick, the strongest contain about three times that number. The bundles taper off gradually and terminate below the apex of the leaf, and immediately above its base, and do not anastomose. Nor did I see them enter the cortex of the stem. In the case of other fibrous bundles, the anatomical relation to the sheaths of the vascular bundles still needs investigation. As regards those of * Compare Schwendener, /.c., p. 143, and the references in the notes, pp. 248-2590. ? Sanio, Botan. Zeitg. 1864, p. 214. * Von Mohl, Palm. Structura; Verm. Schriften, pp. 155,170. * Van Tieghem, Struct. des Aroidées, /.c. 5 Von Mohl, Palm. Struct. p. xx. * The determination of the species not quite certain. SCLERENCHYMA AND SCLEROTIC CELLS, 421 the Potamogetons it has already been stated (p. 232) that they anastomose in the node with one another and with the vascular bundles. Those in the cortex of Palms, as already described at p. 266, after running through many internodes, are sometimes prolonged into purely fibrous bundles, which pass out into the leaves, and sometimes pass over into the fibrous investment of vascular bundles, which likewise make their exit into leaves. Connected with this is the phenomenon likewise mentioned above, of the occurrence, in the plants last named, of intermediate forms between purely fibrous strands and those which contain small complete vascular buridles or single sieve-tubes. It may here be the best place to recall to mind the fibrous strands which sur- round a secretory passage in the leaves of Pinus and the roots of Philodendron. Comp. p. 202, Fig. 185, p. 381; Fig. 168, p. 360. The phenomena hitherto described show that the arrangement of fibrous strands and fibrous sheaths is in a high degree independent of the course of the vascular bundles; but that, on the other hand, there are also close relations between the two. A further phenomenon, corresponding to the attachment of the vascular bundles in stems to the rings or sheaths described, or their inclusion in the latter, occurs widely, especially in the Monocotyledons mentioned above, inasmuch as the vascular bundles are attached to the inner edge of the hypodermal fibrous ridges which project in- wards, or are included in the latter. In flat leaves with fibrous ridges passing verti- cally through them from one side to the other, the latter often have a bundle inserted in the middle, or several one above another near the middle. Side by side with these, other ridges or strands often occur, at least among Monocotyledons, which are quite similar to the former in structure, but contain no vascular bundles, whether they are so far related to one of the latter that they stand opposite to it, or whether even this relation is absent. Tn so far as the vascular bundles are attached to or included in those of the sclerenchyma, the latter stand to the former in the relation of sheaths. The same relation, as already stated above (Sect. 99 et sq.), is also of general occurrence in the case of those vascular bundles which are not attached to continuous sclerenchyma- tous sheaths or to hypodermal strands; the sclerenchymatous fibres follow their course as bundle-sheaths (p. 318), which may serve both as a local protective ap- paratus for the single bundle, and as a strengthening apparatus for the entire organ. Together with the vascular bundles they form jibrovascular bundles (p. 318). Sclerotic endodermis may take part in this function, as was stated in former paragraphs. Between sheaths which simply follow the bundle, and attachment to sclerenchymatous bundles the position of which is otherwise determined, the most various intermediate forms occur, as shown in detail by Schwendener. Apart from these relations of position, the fibrous sheath of the bundle is either closed all round, or parily interrupted, or only pariial, i. e. limited to a relatively small portion of the circumference. The first-mentioned relation of complete sheathing obviously occurs in these of the cases above-mentioned, where the bundles are com- pletely inserted in a closed ring of sclerenchyma, It further exists in other forms of insertion or attachment, and in the case of individual fibro-vascular bundles, especially, but not exclusively, in Monocotyledons. In isolated collateral fibro- vascular bundles the fibrous sheath is then rarely of approximately equal thickness 422 PRIMARY ARRANGEMENT OF TISSUES. all round, e. g. in the rhizome of Carices. In the majority of cases it is thicker on the outside, where it embraces the phloem, than on the inner border of the bundle; the converse relation occurs more rarely, e.g. in the rhizome of Scirpus palustris, in the periphery of the stem of Saccharum officinarum, Bambusa spec., and other species described by Schwendener. And further its thickness usually diminishes at the lateral edges of the bundle, so that here, especially next the limiting surface of xylem and phloem, it is often only 1-2 layers of fibres in thickness, while at the outer or inner edge its thickness amounts to many layers. This is the case, for example, in most bundles of Acorus and Zea (comp. Fig. 147, p. 3175 Fig. 150, p. 330), and many other Monocotyledons; in the stem of Saururus and its allies, in the leathery leaves of species of Melaleuca, Eucalyptus, Eugenia, Callistemon, &c. To the latter condition is related the most frequent form of local zuser- ruption of the fibrous sheath, which consists in the presence of a gap of greater or less extent, filled up by comparatively thin-walled parenchyma, next the lateral edges of the limiting surface of phloem and xylem. Such gaps, or ‘avenues’ as Schwendener calls them, from the surrounding parenchyma to the vascular bundle, are a widely distributed phenomenon in the region indicated, among the bundles of tough parts of Monocotyledons, e.g. in the stem of Canna, the leaves of Typha, Musa, Yucca, and Phormium spec. The bundle of Ranunculus repens represented in Fig. 152, p. 331, is a good example of this, as is also that of Welwitschia, Fig. 157, P- 335; the same condition obtains in the bundles of many tough Dicotyledonous leaves, e.g. species of Hakea and Lomatia. Lateral avenues also appear to occur occasionally in the leaves of the Myrtaceze mentioned above. In a species of Bambusa investigated by Schwendener, an avenue exists on the inner side of the internal bundles of the stem, in addition to the two lateral ones. The fibrous strand, which is of immense thickness on the inner. edge, is here divided by a transverse lamella of parenchyma into a narrow, thicker-walled section bordering directly on the xylem, and a broader, thinner-walled peripheral section. The former is usually interrupted by two short bands of parenchyma, which lead from the transverse lamella to the xylem. The partial fibrous sheath of collateral vascular bundles usually occurs in such a form that the phloem is supported in its whole extent, or only at its outer edge, by a more or less strongly developed fibrous mass, often only by a small group, or even by single scattered fibres. This is the prevailing rule in the leaves and stems of Dicotyledons (Figs. 154, 156, pp. 333 and 334), and is besides not uncommon in Monocotyledons, e. g. in the leaf of species of Crocus, Agave, and Draczna, in the sheath of the leaf of Zea (Fig. 151, p. 331), and in the stem and petiole of Aroidex, as Arum and Colocasia. Comp. also the small bundles of Acorus, Fig. 147, p. 317 The converse condition, that the partial fibrous sheath embraces the inner edge of the xylem, is more rare: it occurs in the smaller bundles in the periphery of the halm of Papyrus, the halm of Cyperus vegetus and other Cyperaceze (comp. Schwen- dener, /.¢.). The fibrous strands which appear in company with the vascular bundles often occur, as shown by the examples of Grasses and Cyperaceze adduced above, side by side with those otherwise arranged, but in very many cases they are present alone. The latter holds good for most of the Dicotyledons which are not expressly SCLERENCHYMA AND SCLEROTIC CELLS. 423 mentioned above as examples of the contrary. Among the tough stems of Mono- cotyledons, the Bambusez investigated show the same behaviour. In many stems of Monocotyledons the sclerenchymatous strands occur chiefly, though not ex- clusively, in company with the vascular bundles; this is the case in the thick stems of Zea, Saccharum, &c., Palms, and Pandanez; also in the stems and petioles of Aroideze, as Colocasia, Arum, and many others. It is the rule in Monocotyledonous stems, and petioles resembling them in structure, that both the relative and absolute thickness of the strands accompanying the vascular bundles, as well as the thickness of the walls of their elements, increase as they approach nearer to the periphery of the bundle-cylinder. In the Aroidexz mentioned, the bundles, with the exception of the outermost circles, are without a fibrous covering. In most Palm stems, the periphery of the bundle-cylinder, which is surrounded by a narrow cortex, is formed of massive fibrous bundles, separated by narrow bands of parenchyma; on the inner side of each of them a small vascular bundle is attached or inserted; this region therefore consists chiefly of firm masses of sclerenchyma, while the bundles in the interior of the stem, as follows from their general course (p. 262), stand farther apart, and have in every respect a weaker fibrous investment. Finally, it scarcely needs to be especially stated, that in the case of vascular bundles which are accompanied by fibrous strands and which gradually become longitudinally united a union of the fibrous strands also takes place. If the latter happens earlier than the union of the vascular bundles themselves, the latter, as seen in cross-section, appear inserted, two or more together, in one fibrous strand, as is conspicuously evident in the periphery of Palm-stems, and especially in the stems of the Pandanez. Lastly, we must here once more call attention to the fact that the fibrous strands are often derived from collenchymatous elements. Those strands, or those sections of them, which belong to parts characterised by long-continued capacity for growth and curvatures due to growth, show collenchymatous properties as long as this capacity is maintained, or they show an intermediate character between sclerenchyma and collenchyma; e.g. the base of the sheaths of leaves in the Grasses (Fig. 151), and the above-mentioned stems of Aroideze. Sect. 126. In certain relatively rare cases, isolated sclerenchymatous fibres, or fibres united through part of their course to form small bundles, occur in the parenchyma, external to, and usually side by side with, the strands and layers described. To this category belong, in the first place, those branched elements projecting into the air-spaces, which under the name of internal hairs have already been minutely «described above (Sect. 53, p. 221) in the case of the Nymphzacez, Limnanthemum, Rhizophoree, and many Aroidez. Elements more or less similar to those mentioned occur elsewhere fixed in dense parenchyma. As isolated cases of this sort may shortly be mentioned the unbranched fibres in the cortex of the root of Chamedorea elegans, already dealt with at p. 129; also the branched fibres in the pith of Carissa arduina‘, and the often branched fibres, 6-14™™ in length, which Trécul? found in the cortex of Euphorbia rhipsaloides, and in the pith and cortex of E. xylophylloides. 1 A. Gris, 2.¢., compare p. 403. * Comptes Rendus, LX. p.-1349. 424 PRIMARY ARRANGEMENT OF TISSUES, Isolated fibres, sometimes ramified, sometimes unbranched, occur as a widely’: distributed and characteristic phenomenon in the parenchyma of the. cortex and leaves of many Gymnosperms, and to this mode of occurrence, as well as to that. described in the case of the Nymphzaces and Aroidex, their appearance in a number of tough Dicotyledonous leaves is related. Many of the phenomena belonging to this category, and the literature referring to them, have already been mentioned in Sect. 30 (p. 130), to which reference must therefore be made. Among the Gymnosperms, many Cycadee (Dion, Ceratozamia, Encephalartos, &c.), and several Coniferz (e.g. Cunninghamia, Fig. 183, p. 380), show longitudinal unbranched fibres, occurring isolated or in small groups, in the parenchyma of the petiole and leaf, The same applies to the cortex of Ephedra. Stellately-branched fibres lie scattered in the chlorophyll-parenchyma in Sciadopitys, Dammara, and Araucaria imbricata. The latter recall the fibres, differing from those of the hypoderma, by which the entire paren- chyma of Welwitschia, even including that of the floral organs, is abundantly permeated, the resemblance consisting especially in the presence of great numbers of crystals of calcium oxalate deposited on their surface: they are thick and short spindle-shaped’ elements, with short protruding branches here and there at their pointed ends, and with’ a very thick, much stratified, lignified wall (comp. Fig. 55, p. 132, and Fig. 187, p. 408), In the stem these fibres are directed without order towards different sides. In the leaves, where they are on the average somewhat narrower than those of the stem, they lie asa rule, not without exception, in the middle lamella, parallel to the surface of the leaf, with their longitudinal axis sometimes directed longitudinally, sometimes transversely or obliquely with reference to that of the leaf; they usually stand about at right angles to the surface of the leaf, on both sides of the middle lamella, in the chlorophyll tissue tra- versed by fibrous bundles; they reach the middle lamella with one end, and the inner. surface of the epidermis with the other, and often have a hook-like bend at the latter, or are even wedged in between the inner parts of the epidermal cells. In the parenchyma of the third genus of Gnetacex, Gnetum, at least in the species investigated, sclerenchyma- tous fibrous cells are no less abundant than in Ephedra and Welwitschia; in Gnetum Gnemon they are present in the entire parenchyma of the external cortex, here running longitudinally, and branching rarely or not at all, also in the pith of the nodes, and in the leaves near their surfaces, especially the upper, to which they are nearly parallel. in Gnetum Thoa they occur in the same way in the outer cortex, but especially in the pith of.the nodes and in the leaves; in the regions last mentioned’ they are abundantly and variously ramified, in the leaf their size is very unequal, sometimes very considerable, Comp. p. 130. The leathery leaves of all these Gymnosperms are thus strengthened by a complicated’ sclerenchymatous frame-work. Among Dicotyledons the leaves of Camellia and Fagrza: are characterised by numerous fibres scattered in the parenchyma, which are abundantly and irregularly branched (Fig. 53, p. 130); this also applies to the leaf of species of Olea. In Olea europza the fibres are very irregular in their ramification and direction, sending out branches on all sides as far as the under surface of the epidermis; in O. fragrans ‘they extend, usually without branching, right across the whole leaf at right angles to the surface, and branch in a more or less pedate manner at the upper and lower epider-’ mis, so that they act as columns, connecting the two epidermal layers together.’ (Thomas.) In the Proteacez mentioned at p. 130, rod-shaped, more or less ramified sclerenchymatous fibres stand between the palisade-like chlorophyll-cells, at right angles to the inner surface of the epidermis. They are as high as, or somewhat higher than the parenchymatous layer on either side which contains the chlorophyll, and with their usually branched ends they adhere on one side to the epidermis, while on the other they are attached to or wedged into the middle layer of the leaf, which in the thick-leaved species is destitute of SCLERENCHYMA AND SCLEROTIC CELLS. 425 chlorophyll. Still stouter, shortly-branched fibrous cells stand in the palisade-parenchyma on either side of the leaves of Statice purpurea, approximately at right angles to the. surface, but without reaching the epidermis. Sct. 127. Short sclerenchymatous elements occasionally appear in the primary structure of the Phanerogams, united, like the fibres, to form hypodermal strands or sheaths, or as portions of them. This is the case with the hypodermal layers in Palm-stems (Cocos, Elais, Astrocaryum vulgare, Mauritia armata, Klopstockia, Chameedorea Karwinskiana) mentioned at p. 127, which are interrupted under every stoma by thin-walled parenchyma, and also with the annular layers also mentioned at p. 127, in the stems and roots of Aroideze, and with the stegmata. Some special phenomena to be placed in this category are described by Pfitzer* in the case of the foliage-stem of Restiaceze. Restio diffusus has single-layered double longitudinal rows of rod-shaped, radially elongated sclerenchymatous elements, which alternate with longitudinal bands of chlorophyll-parenchyma. Willdenowia spec. and Leucoplocus show the same structure, with the difference that the hypodermal layers are broader, consisting of 3-4 rows. In Elegia nuda and species of Restio (R. teclorum, paniculatus, incurvatus, &c.), Thamnochortus, Will- denowia, Hypolzna, Ceratocaryum, and Leucoplocus, the very large air-cavities under the superficially situated stomata are bordered by a ring of sclerenchymatous elements, elongated and placed at right angles to the epidermis, and converging in a curve towards the inside; they are in uninterrupted lateral connection with one another, with the exception of slit-like interstices, by means, of which communication between the stomatal cavity and the intercellular spaces of the neighbouring parenchyma is effected. In the external cortex of Dicotyledonous woody plants, short sclerenchymatous elements often appear in conjunction with fibres, as annular sheaths, the formation of which, on account of their connection with the processes of secondary growth, will be dealt with in Chap. XV. Their occurrence in sappy masses of parenchyma and in the pith of Dicotyledonous plants has been already spoken of on p.127. The small groups or nests which lie scattered in the nodes of many Potamogetons (P. crispus, densus, gramineus, perfoliatus, &c.), near the vascular bundles where they anasto- mose and pass out into the leaves, furnish an isolated case of the occurrence of these elements. Sect. 128. Among sclerenchymatous masses, which strictly speaking belong to the category of hypodermal tissues under consideration, those remain to be specially mentioned to which massive hard emergencies owe their strength: e.g. tough prominent warts, as those of Aloe verrucosa, thorny teeth of leaves, as in Ilex Aquifolium, Agave, and Aloe, and spines and thorns of the different morphological categories. The epidermis itself no doubt always takes part in the sclerosis. The sclerenchymatous elements are in several cases short, e.g. warts of Aloe verrucosa, thorns of roses; usually they are elongated. The sclerenchyma, together with the sclerotic epidermis, either forms the entire mass, or it surrounds other internal tissues. E. g. the sclerenchymatous cylinder of the stem of Berberis vulgaris * (p. 419) sends a t 1 Pringsheim’s Jahrb. VII. p. 561. 2 Mettenius, Hymenophyllacez, 7. ¢. p. 439. 426 PRIMARY ARRANGEMENT OF TISSUES, massive branch into the thorn-leaf, which in the broad base of the thorn forms as a thick plate the larger lower half; the narrower upper half is thin-walled parenchyma, in which the vascular bundles lie. As the thorn tapers off the hypodermal scleren- chyma increases in relative extent, at the cost of the other tissue, in such a manner that in the cylindrical upper part there are only feeble vascular bundles in the middle, enclosed in scanty parenchyma, and completely surrounded by the scleren- chymatous mass, while the apex is formed of the latter and of the epidermis exclusively. The ends of the petiolar, stipular, and branch thorns of Astragalus aristatus, Halimodendron, Robinia, Maclura, Crategus, and many thorny teeth of leaves, show a similar structure. For further details compare the works cited at p. 57, by Delbrouck, Uhlworm,-Suckow, &c.; v. Mohl, Palm. Struct. p. 7; Lestiboudois, Comptes rendus, Tom. 61, pp. 1034, 1093. Sect. 129. The sclerotic elements of the Ferns and Hydropteridez ’ are, as ex- plained in Sects. 26 and 28, sometimes fibrous cells containing starch, sometimes specific sclerenchymatous elements; the division of labour between the two forms is not however strictly carried out, and a sharp severance of the two is not possible. Their arrangement corresponds on the whole to the rules stated for the Phanerogams in preceding. paragraphs; even the peculiarities to which attention is to be called can generally be brought under these rules as special cases. : Sclerotic Aypodermal masses of tissue are absent in the stems or rhizomes of many Ferns, the hypodermal zone being only distinguished from the internal parenchyma, into which it gradually passes over, by the closer connection, smaller width, and somewhat thicker walls of its cells; e.g. Polypodium vulgare, pustulatum, Davallia elegans, Acrostichum vexillare, Angiopteris; Marsiliacee with a many-layered dense hypodermal zone of parenchyma, which passes over internally into the inner zone, which Fic, 188—Osnunda regalis; is traversed by a circle of wide air-passages. Many stems, eee Rom above, te fromthe and especially the thicker ones, have on the other hand a Tea on Suet Ieee — distinet hypodermal sclerotic zone, consisting of several or aerial eben ware many layers of uninterruptedly united elements, which in ee the true Ferns always have brown membranes; in most cases this zone does not border directly on the epidermis, but is separated from it by some layers of thin-walled parenchyma. This is the case, for example, in the Cyatheacez, Polypodium Lingua, Platycerium, Davallia pyxidata, &c. The scleren- chymatous ring, which lies in the cortex of the thin stem of Hymenophyllee, may just as well be mentioned here as among the instances of sclerenchyma accom- panying the vascular bundles, to be brought forward later. The same applies to the dark-brown mass of sclerotic elements containing starch, which in Osmunda regalis and Todea hymenophylloides forms the principal part of the stem, and which is every- where sharply marked off from the relatively small colourless tracts of parenchyma. containing the vascular bundles. Comp. pp. 346 and 279, and Fig. 188. ° Sclerotic annular layers bordering directly on the epidermis occur more rarely, e.g. in the ? Compare the literature cited above, §§ 73-87, on the structure of the Fern-stem, SCLERENCHYMA AND SCLEROTIC CELLS. 427 rhizome of Pteris aquilina and Polybotrya Meyeriana. The sclerosis which occurs at an early period in the cortex of the root of many Ferns was mentioned above - at p. 413. In the petioles and ribs of the leaves in Ferns, it is a general rule that a more or less strongly developed sclerotic hypodermal layer, often interrupted by the bands and islets covered with an epidermis without stomata mentioned at p. 405, lies directly beneath the epidermis, which itself not unfrequently takes part in the sclerosis. The collenchymatous zone in the petiole of the Marattiacee is more deeply seated, and separated from the epidermis by several layers of es walled parenchyma (comp. p. 120). According to Mettenius, strands usually branch off from the hypodermal sclerenchyma in the petioles and ribs, which accompany the finer ramifications of the vascular bundles into the lamina of the leaf. Among strictly hypodermal sclerenchymatous masses in the lamina of the leaf, the continuous layer of Acropteris australis has already been described at p. 132. Such masses occur in the form of nerve-like bands in many leaves of Marsilice and Ferns. In the lamina of the aérial leaves of Marsilia coromandeliana, trichopoda, muscoides, and distorta!, narrow colourless bands of sclerenchyma run between and in a similar direction to the nerves; some of these are small strands or even isolated fibres, adjacent to . the epidermis of the lower surface of the leaf; others are thicker, extending through the entire thickness of the leaf, from one epidermal surface to the other. Hypodermal strands distributed like nerves are described by Mettenius as occurring in the segments of the lamina of the leaf of Todea Hymenophylloides, Polypodium solidum, Pteris pinnata, Davallia elata and elegans. In other Ferns the edge of the leaf or of its segments is entirely or partially rimmed by a hypodermal (many-layered) strand of sclerenchyma, which is prolonged continuously into the sclerenchyma of. the petiole. This is the case in Polypodium Lingua, sporadocarpum, Brownianum, Asple- nium lucidum, Polybotrya cervina, Meyeriana, Aspidium falcatum, Adiantum denti- culatum, &c.? The strands described by Mettenius as Nervi spurii, in the leaves of many species of Trichomanes, may also be mentioned here, although, as running through a usually single-layered lamina, they do not strictly belong to this category. They consist of one or a few rows of elongated elements, which are usually accom- panied by stegmata (p. 128). Around the vascular bundles-in the stem, roots, petioles, and the stouter ribs of the leaf, sclerenchyma and sclerotic cellular tissue is in many cases entirely absent, the bundle or its endodermis is surrounded by thin-walled parenchyma, which differs scarcely if at all from that lying further away from it, and is never sharply marked off. This is no doubt the case in most roots; in the stem and petiole of the Marattize, in the rhizomes of Aspidium Filix mas, Onoclea Struthiopteris, Polypodium vulgare, Davallia pyxidata, &c. Pteris aquilina may also be mentioned here. In several roots on the other hand, and in most stems and petioles of Ferns, distinct sclerotic sheaths and strands occur in company with the vascular bundles. This is the case, for example, in the roots of many Polypodia, as P. ireoides, vulgare, &c., 1 A. Braun, Monatsber. der Berlin. Acad. 1870, p. 693. 2 Mettenius, Hymenophyllacez, p. 438.. 428 "PRIMARY ARRANGEMENT OF TISSUES. ae Blechnum occidentale, and Scolopendrium vulgare’, in such a form that one or more layers bordering directly on the endodermis acquire brown walls, strongly thickened chiefly on the inner side, so that these layers form either a sheath going uniformly round the whole axial bundle, or one which is interrupted, or at least thinner, over the corners of the xylem-plates. ‘The sclerotic sheath in stems and petioles has, in the first series of forms, the same position-relative to the endodermis as in the roots. And indeed the sclerosis very often at first affects only the inner walls bordering on the endodermis, and the lateral walls of the layer of cells which is in direct contact with it; the outer walls of this layer, like those of the surrounding parenchyma, are not sclerotic, e. g. stem and petiole of Polypodium Lingua, and pustulatum, and the stem of Davallia elegans. The converse condition in the thickening of the walls occurs rarely: petiole of Blechnum brasiliense; or the sclerotic thickening of the wall exists all round, though it may be weakest on the outside: Polypodium Phyllitidis”» In other cases the endodermis is surrounded by an uninterrupted sclerotic sheath, consist- ing of one or more layers (rhizome of Polybotrya Meyeriana and Hymeno- phyllez), or by an interrupted sheath, i.e. by one or several many-layered strands of sclerenchyma adjacent to it. This is the case, for example, in the thizome of Platycerium alcicorne, and in very many petioles. In the very frequent case where the bundles in the latter have projecting corners and de- pressed incurvations of: their surface, a definite relation between the latter and the strands of sclerenchyma exists; e. g. they lie on the concave side of the runnel-shaped bundles in the petiole of FIG, 189 —Cyathea Imrayana; cross-section through the hving stem, natural size; seen from above. Atd,c,d, foliar gaps. All the Balantium Culcita and Cyathea medul- qgutte black bands and points are cross-sections of sclerenchyma; ; . % the paler ones those of vascular bundles. Within or adjacent to the laris, in the corners of the figure x foliar gaps, especially @ and 4, are root-bundles on their way to the i. i periphery ; JS furrow of the base of the leaf; @ vascular bundle of which the bundle shows as seen in fee nsec cm eames” cross-section, in the péliole: of Scslo~ : pendrium vulgare, &c. (Comp. Rus- sow, /.c. and the special pteridological descriptions.) In a second series of forms, the sclerenchyma accompanying the bundle is not adjacent to the endodermis, but is separated from it bya zone of delicate parenchyma, usually consisting of many layers of cells. This is the case in the stem of Todea africana, but more especially in that of the Cyatheacez. In most of the latter, e.g. Cy- athea arborea, Imrayana (Fig. 189), Alsophila microphylla, and many others, the band- 1 Van Tieghem, /.c., p. 66, Taf. 5. ? Russow, /.¢, p. 81. SCLERENCHYMA AND SCLEROTIC CELLS, 429 shaped main bundles of the stem are immediately surrounded by a many-layered zone of delicate parenchyma ; this is completely enclosed by a stout sheath, likewise consisting of many layers, and reaching a thickness of above 1™™, which consists of acutely spindle-shaped, closely connected fibrous cells. From this sheath strands of various dimensions branch off, which accompany the medullary and cortical bundles (with the exception of the thinnest unsheathed branches), and those which pass out into the leaf; they seldom enclose these completely as they do the main-bundles ; usually they are open and runnel-shaped, and placed on the inner, medullary side of the bundles, from which, however, they are separated by a broad layer of parenchyma (comp. Sect. 85). In many Cyatheacez, as Alsophila pruinata, blechnoides, and species of Cibotium, a// the bundles of the stem are accompanied only by open strands or plates of sclerenchyma, which lie opposite their inner side. Connected with these are the two thick brown plates of sclerenchyma, often fused to form a tube with only a narrow opening on one side, which run longitudinally through the rhizome of Pteris aquilina in the middle of the parenchyma, between the outer and inner tube of vascular bundles (Sect. 87). To this category belongs also the axial strand of sclerenchymatous fibres, running inside the annular vascular bundle, and often, it is true, bordering directly on the endodermis, in the rhizome of species of Pilularia and Marsilia, In the larger species of Marsilia, as M.“Drum- mondii, and salvatrix, the outer side of the axial annular bundle is also sur- rounded by a tough brown sheath of sclerenchymatous fibres, which passes over internally into thinner-walled layers of cells containing abundant starch. In M. quadrifolia this sheath is replaced by a many-layered annular zone of cells bordering on the endodermis, which are rich in starch, and have brown but thin walls. Isolated small bundles of sclerenchyma with brown membranes, or even isolated fibres, are observed here and there in the parenchyma of Ferns, e.g. the very hard small bundles in the pith of many Cyatheaceze, where, it is true, they are often con- nected with those accompanying the vascular bundles as their ramifications, though they may often have an independent course, e.g. Alsophila microphylla. On the other hand, they are entirely absent in several species, e.g. in Alsophila pruinata, blechnoides, species of Dicksonia and Cibotium, according to Mettenius. Small strands, consisting of only a few fibres, ending blindly in the parenchyma above and below, or isolated fibres, traverse the parenchyma longitudinally in the rhizome of Pteris aquilina, Polypodium Lingua, Osmunda regalis, &c. In the winged edge of the base of the petiole of Osmunda and Todea similar brown fibrous tracts and fibres are arranged so as to form pinnate bands‘. As regards the distribution of sclerotic elements in the small stem of the Lycopodia and Selaginellg, similar general rules and also similar variations prevail, as in the case of the thinner stems of Ferns. In the stouter Lycopodia, as L. clavatum, alpinum and Chamecyparissus (Fig. 162, p. 349), a bulky, many-layered ring of fibres surrounds the axial vascular bundle. A narrow annular zone of slightly sclerotic elements lies in the middle of the thin-walled cortical parenchyma of the foliage-stem of Psilotum. In the rhizomes of this plant and in the stems of Lycopodium Selago ! Compare Milde, Monograph. Generis Osmunds ; Vindob. 1868. 430 PRIMARY ARRANGEMENT OF TISSUES. sclerotic layers are absent. In the stems of Selaginella the sclerosis is sometimes limited to the epidermis (S. spinulosa); in most species it further affects a hypo- dermal zone, which gradually passes over internally into thin-walled parenchyma; in S, rupestris the entire tissue of the stem is in the highest degree sclerotic, with the exception of the zone of lacunar parenchyma, which in this, as in all other species investigated, directly or indirectly surrounds the vascular bundles. (Comp. Fig. 131, p. 282.) The roots of the Selaginella and Lycopodia present essentially similar phenomena to those of the stem, with reference to the conditions in question. CHAPTER XI. ARRANGEMENT OF THE SECRETORY RESERVOIRS. Sect. 130. The primary arrangement of the secretory reservoirs presents little of interest, and it has often been impossible to avoid the mention of it in preceding sections. The present Chapter, which is necessary simply for the sake of consistency, must therefore be a short one, and confine itself chiefly to references to former paragraphs. The sacs containing crystals lie, in the plants enumerated in Sect. 31, to which they are peculiar, in the primary parenchyma, sometimes ‘scattered’ between its cells, sometimes more regularly distributed, or at definite places in definite grouping. On their ‘distribution in the pith of Dicotyledonous woody plants, see p. 403; on their accumulation on the wall of the air-passages of Water-plants, Aroidez, &c., comp. p. 2193; on the series of raphides accompanied by mucilage in the Mono- cotyledons, see p. 139. It may here be mentioned more definitely than was done in Séct. 31, that the vascular bundles are occasionally, but by no means universally accompanied by series of elements containing crystals; e.g. the petiole of Cycas revoluta (p. 336); and the medullary bundles of Melastomacez, as Heterocentron and Centradenia spec., with longitudinal rows of sacs containing crystalline aggrega- tions on their outer side, &c. The sacs containing mucilage lie in the primary parenchyma of the plants men- tioned at p. 143, and in fact principally in the foliage and cortex, usually scattered without any generally determinate order; their more regular distribution in the tubers of Orchis has already been mentioned at p. 144. The same scattered position in the primary parenchyma of the foliage, pith, and especially of the cortex, prevails in the case of the short sacs containing resin and gum-resin, of the families mentioned at p. 145. The Jong sacs of this category, and the tannin-sacs, have already been discussed at p.146, &c. Comp. further Sect. 48, especially p. 199, and Chap. XII. CHAPTER XII. COURSE OF THE LATICIFEROUS TUBES. Sxcr. 131. The laticiferous tubes’, in most plants which are characterised by their occurrence, traverse the entire body of the plant as a continuous system. Exceptions to this rule seem, however, to occur; in the roots of Asclepias. curassavica, and Cornuti, and of Periploca, I could not find them, but will not assert their absence with complete certainty; in the roots of Ficus elastica I only find them in the secondary bast. As regards their relation to the other tissues, we may call them, as already indi- cated above, companions, or in some places even representatives, of the sieve-tubes. The latter relation is especially manifest in the secondary bast of many plants, to which we shall have to return, below, Sect. 163. In the primary groups of tissue the Jaticiferous tubes are distributed— (2) In the roots within the phloem of the vascular bundle. Only in the case of the Euphorbiz investigated do others occur in addition, which arise as branches from those of the cotyledonary node, and lie close under the epidermis, separated from the latter only by a few layers of cells (comp. p. 196)”. (2) In stems, petioles, and ribs of the leaves, the main courses or main trunks of the tubes lie chiefly in the tissue surrounding the phloem-portions of the vascular bundles, following the longitudinal course of the latter, and, as seen in cross-section, scattered without strict regularity among the surrounding parenchyma. If the phloem is sheathed by a strand of sclerenchyma they lie outside the latter. In addition to these tubes, other smaller onés occur in certain cases, e.g. Cichoriaceze and Papaver, which run in the phloem itself. In milky plants provided with phloem-portions towards the pith, or with separate medullary bundles of sieve-tubes, these also are accompanied by laticiferous tubes (comp. p. 338). In the foliar expansions, the tubes on the one hand still follow the higher orders of ramification of the bundles; on the other hand, in the majority of cases, they send out branches which leave the paths of the vascular bundles, force their way in all directions between the cells of the parenchyma, and end blindly, sometimes in the interior of the latter, sometimes at the inner surface of the epidermis. In the case of Siphocampylus manettizeflorus, Trécul even states that the ends of the branches pass between the cells of the epidermis, as far as its outer surface. In many milky Dicotyledons, branches of the tubes also traverse the cortex of the stem, partly in the internal parenchyma, partly hypodermally. In the succulent Euphorbiz possessing + See Chapter VI, and the literature there cited. ? [Compare Scott, /.c., p. 143. (See p. 193).] COURSE OF THE LATICIFEROUS TUBES, 433 rudimentary or very fugacious leaves, and in the Asclepiadez, they branch off at all points from the main trunks, and are distributed through the massive cortical paren- chyma i in all directions as far as the epider mis, having an oblique and crooked course. In non-succulent stems with well-developed, persistent foliage leaves, longitudinal tubes occurring in the hypodermal cortical layer, and branched off from the main trunks in the nodes, are, judging from Trécul’s statements and the phenomena in Euphorbiz to be communicated below, at any rate much more frequent than has been represented by most previous observers. In many cases also, branches, which are usually thick tubes, pass from the main trunks into the pith; in the case of medullary bundles of sieve-tubes, they branch off from the main-stems which accompany the latter (Hoya, Asclepias spec.) In plants without medullary sieve-tubes, as Ficus and Euphorbie, they arise as branches from the main trunks, principally, but as shown by the succulent Euphorbiz not exclusively, in the nodes. According to the particular case, they are either inserted in the parenchyma throughout the entire thickness of the pith, e.g. Ficus, or they are confined to its periphery, e.g. Euphorbiz. _ According to the particular families, and even the species, the general rules stated above are subject to manifold variations. The most important details will be given below, reference being made to Chap. VI, and the special works there cited. Some data referring to the secondary wood and secondary bast, belonging to Chaps. XIV and XV, may here be anticipated. I. ARTICULATED LATICIFEROUS TUBES. 1. Cichoriacess. In the stem of those species investigated, which have ordinary col- lateral bundles within the boundary of the plerome, the tubes in the first instance lie around ‘ the phloem of each of these bundles, ‘Their longitudinal main trunks here form, as seen in cross section, a single curved row, often interrupted by parenchymatous cells, at the boundary towards the cortical parenchyma; their numerous transverse anastomoses pass along the outer surface of the approximately semi-cylindrical phloem, which is usually . destitute of any sclerenchymatous support. ‘These peripheral tubes are the largest. In the stems of Chondrilla, Taraxacum, Apargia, and Cichorium, no other laticiferous tubes are present, The bundles in the lower part of the stem of Sonchus tenerrimus, Picri- dium tingitanum, and, in a slight-degree, of Lactuca virosa, have sieve-tubes on the inner side of the xylem also, and are then accompanied here also by laticiferous tubes, which are connected with those outside by branches passing right round the vascular bundle .(Trécul). Finally, where, as in the investigated species of Lactuca, Sonchus, Scor- zonera, Tragopogon, and Hieracium, small medullary bundles of sieve-tubes are present, the latter each contain some laticiferous tubes, which run parallel to the sieve-tubes, and anastomose with one another between them, without however having an open com- munication with the sieve-tubes. According to Trécul, anastomoses take place along the entire internode between the nets of tubes accompanying the different vascular bundles. In all plants of the family in question, they are especially numerous in the nodes, both between the peripheral tubes accompanying the vascular bundles, and be- tween these and the medullary tubes, the latter occurring together with the anastomoses - described at p. 231, between the groups of sieve-tubes belonging to the two regions. In the nodes the nets of laticiferous tubes of the stem are further continued into those of the petioles and of the axillary branches. -In the petioles and ribs of the leaves, the nets of tubes accompany the vascular bundles with the same arrangement as in the stem, and finally their branches, which end Ff 434 PRIMARY ARRANGEMENT OF TISSUES. blindly, either terminate together with the last vessels in the parenchyma of the leaf, or similar terminal branches are sent out as far as the epidermis. For the course of the tubes in the peduncle and receptacle, essentially the same rules hold good as for the stem: some branches of them accompany the small vascular bundles which traverse the pistil, corolla, and stamens (Hanstein). In the roots the tubes lie in the phloem-groups of the original vascular bundles, and thus within the pericambial zone—at least in Tragopogon and Scorzonera hispanica. They never enter the xylem groups unless it be in the ultimate ramifications of the bundles, where they are in close contact with the tracheides, and in the nodal anasto- moses, which pass through the medullary rays, between the medullary and the peripheral ~ bundles, where they may come to lie close to the vessels, 2. The nets of laticiferous tubes of the Campanulacem and Lobeliacem are in general quite similar in form to those of the Cichoriacex. In their distribution a differ- ence is in so far observable, that the chief seat of their occurrence is the internal phloem- region, which lies towards the xylem-portions of the bundles. In the periphery of the phloem and the parenchyma of the external cortex they are in many cases entirely absent, or very isolated (Lobelia inflata, syphilitica, urens, Adenophora Lamarckii, Phy- teuma Halleri, spicata, Campanula sibirica, medium, ranunculoides, grandis, lamiifolia) ; more rarely they are abundant (Tupa salicifolia, Isotoma, Centropogon, Piddingtonia spec.), and especially in Tupa Feuillei, Ghiesebrechtii, and Musschia aurea, where they penetrate up to the epidermis. In the case of Siphocampylus manettizflorus, Trécul states that individual ends of branches penetrate as far as the surface of the epidermis, and even project there as small papill. In those Campanulacez which have sieve-tubes on the inner side of the woody ring, or in the medullary bundles, the phenomena above mentioned are complicated by the occur- rence of laticiferous tubes accompanying the latter, as in the case of the Cichoriacez with similar structure. They are absent in the xylem, in the secondary ring of wood, and in the parenchyma of the pith in all Campanulacez investigated, and in many Lobeliacez. In other plants of the latter family on the other hand, e.g. Centropogon surinamensis, Tupa salicifolia, Ghiesebrechtii, Feuillei, Siphocampylus manettizeflorus, microstoma, and Lobelia laxiflora, Trécul found them scattered at the periphery, and more or less deep in the interior of the pith, and found that these medullary tubes are in communication with the cortical ones by means of branches traversing the woody ring. 3. The laticiferous tubes of the Papayacea, investigated in Papaya vulgaris, Vascon- cellea monoica, cauliflora, and microcarpa, form in the stem of these plants an abundantly ramified and anastomosing net-work of tubes, extending both through the highly paren- chymatous primary and secondary wood, and also through the medullary rays and the bast-region. The main trunks have an approximately vertical course and form interrupted rows approximately concentric with the circumference of the stem; the separate portions of these rows everywhere alternate variously in wood and bast, with parenchyma, vessels, and sieve-tubes. The neighbouring tubes are connected laterally with each other by means of exceedingly numerous wide anastomoses, Similar anastomoses occur in the most various particular forms, between the different groups and rows, both in radial and tangential direction, the radial connections being often effected by long transverse branches with an approximately horizontal course. Blindly ending branches and branch- lets further occur with varying frequency. The tubes are usually separated from the vessels by at least one layer of parenchyma, some however are in contact with them, and as discussed above (Chap. VI) there is perhaps. open communication between the two. Into the pith of the internodes, which soon disappears, the tubes do not enter; they form, however, a complex network, anastomosing on all sides in the medullary disk, which is persistent in the nodes. In the primary cortical parenchyma, Dippel alone found isolated tubes, accompanied as a rule by parenchymatous cells containing crystals near the outside of the bundle of bast-fibres, and connected with the internal tubes by very elongated horizontal transverse branches. COURSE OF THE LATICIFEROUS TUBES, 435 In the leaf-stalks and ribs of the lamina they follow the vascular bundles, often accom- panying and touching sieve-tubes and vessels, In the parenchyma of the leaf they end with numerous anastomosing branches. After the commencement of secondary thickening the root has a similar structure, and a similar distribution of the laticiferous tubes to that in the stem. 4. Among the milky Papaveracesw two types of laticiferous tubes are to be dis- tinguished. The one is represented by the investigated species of Papaver, Roemeria, and Argemone, and shows tubes which arise from elongated elements, but only rarely allow traces of the original cross-walls to be recognised in the mature condition; these are connected into a net by more or less numerous anastomoses, In the stem and petioles they lie in tangential curved rows in the phloem of the vascular bundles, in each one of which they anastomose in the transverse direction, though there are no anastomoses between those of the different bundles of an internode, In the (secondarily thickened) root, in the cortical parenchyma, and especially the bast layer, and also in the parenchyma of leaves, pericarps, &c., they end in an abundantly ramified net. The other type is represented by Chelidonium, and is characterised by the facts that the cross-walls of the elements merely have one or more large perforations in the middle, while on the other hand their edge remains preserved, and that reticulate connections ‘do not occur. On account of the partial persistence of the cross-walls, the tubes appear at the first glance like rows of cells, the articulation of which comes out all the more sharply, as they are usually somewhat constricted at the cross-walls (cf. Figs. 80, 81, p. 189). Sometimes when two tubes are in direct lateral contact, perforations appear to occur in the lateral wall also. In the older roots the tubes are often branched, owing to the fact that a series of elements is continued from one point into two diverging series, which meet at an acute angle; and the individual elements are short, on the average 2—4 times as long as broad, being of about the same length as the neighbouring parenchymatous cells, and elements of the sieve-tubes. In the parts of the plant above ground, on the other hand, the elements are very elongated, so that their ends more rarely come into view in preparations. In the (older) roots, the tubes in the bast-layer are so distributed in groups, forming concentric, irregularly interrupted rows, that every group usually lies in the neighbourhood of a small group of sieve-tubes, surrounded by masses of parenchyma containing starch. In the stems and petioles narrow laticiferous tubes lie scattered in the vascular bundles within the phloem, and at the periphery of the xylem; they further occur externally to the vascular bundles, on the outside of the fibrous bundles bordering on the phloem, and also.isolated in the peripheral (cortical) parenchyma, In the lamina of the leaf and the parts of the flower the system of tubes ends in the reticulate form described above for other cases. In other Papaveracez, especially Macleya cordata and species of Glaucium (I investigated Gl. luteum), no doubt also in Eschscholtzia (which, however, requires further investigation with reference to the statement of the anonymous writer in the Botanische Zeitung, 1846), and in the Fumariacez, no laticiferous vessels whatever are known. A red sap, which is on the whole clear, and mixes both with water and alcohol without turbidity, appears conspicuously on cut surfaces of the rhizome of Sanguinaria: it is contained in large, thin-walled, roundish or shortly cylindrical cells or sacs, which are abundantly dis- tributed through the whole parenchyma, sometimes isolated between (starch-containing) parenchymatous elements, sometimes, and especially in the cortex, forming continuous longitudinal rows (cf. Hanstein, /.c., Taf. 1). In the stem and petiole, which I have not investigated, these sacs are elongated cylindrical or prismatic. Neither their walls nor their contents show the properties characteristic of laticiferous tubes; they were therefore mentioned above at p. 147. The same holds good for the sacs filled with a clear galdich-vellew sap, which are scattered through the parenchyma in the root of Glaucium luteum, and which, according to the form of the contiguous elements, are more or less longitudinally-extended.. In the stem and foliage of species of Glaucium (cf. Trécul, /.c.) they are absent. In the Ff2 436 PRIMARY ARRANGEMENT OF TISSUES, rhizome and stem of Macleya cordata, similar sacs, sometimes very much elongated, are scattered in large numbers about the periphery of the ring of vascular bundles, and in the medullary rays. As the parts grow old the reddish-yellow colour of the sap dis- appears. The elongated sacs bordering on the fibrous bundles of the bast then acquire thickened walls, like sclerenchymatous fibres. ' sg, According to their structure, the latex- and tannin-tubes of many Aroidew! belong to this part of our subject. Their main trunks lie in the periphery of the phloem of the vascular bundles, usually two or more together, in collateral bundles situated as a rule symmetrically on. each side, in the ‘compound’ bundles less regularly distributed. .In their most perfectly developed form, in Caladium and its allies, Alocasia, Xanthosoma, Syngonium, &c., they constitute thin-walled sacs, about equal to the sieve-tubes in . width, following the longitudinal course of the bundles; at the limiting surfaces of the surrounding parenchymatous cells they send out numerous, pointed or blunt, blind - protrusions, which penetrate between the latter. Other protrusions are extended to form. longer tubular branches, penetrating between the surrounding tissue-elements, and “ these also sometimes end blindly, often showing somewhat enlarged ends, while some- times they meet similar branches of neighbouring trunks, and enter into open communi- cation with them, The network of tubes thus formed is extended throughout the parenchyma, not only between the vascular bundles, but also not uncommonly in the parenchymatous cortex, as far as the under side of the epidermis. . . . The branches of the network of tubes also run towards the trachez, attach their ends to the latter, and, as stated above, frequently enter into direct communication with them. ' The contents of these tubes consist of a finely granular fluid, which, according to Trécul, is milky in species of Syngonium and Xanthosoma,.in other cases only a little turbid, and very rich in tannin, so that after the action of iron salts or potassium bichro- mate the net-work of tubes comes out darkly stained. -Other Aroidee (Richardia africana, Arum vulgare, Dracunculus, Aglaonema simplex, Dieffenbachia Seguine, and species of Philodendron) have no Jaticiferous.tubes, but contain .. longitudinal rows of elongated cylindrical or prismatic sacs in the periphery of the phloem, with the same arrangement as the trunks of the net-work of tubes described above. . They contain the same turbid contents as the tubes described (in Dieffenbachia Seguine according to Trécul they are without tannin), but they are separated by cross-walls :. and are without lateral anastomoses, only sending out short blind protrusions between the limiting surfaces of the neighbouring parenchymatous cells, A third category—Heteropsis, Lasia, Scindapsus, Monstera, Anthurium, Acorus, &c.— is wholly destitute both of the tubes described, and of sacs. That these nét-works of tubes arise from the union of originally separate, branched cells, is shown both by the history of their development, and by comparison with the sac-tubes of the second category. .. 6. In the species of the genus Musa, the vascular bundles in the stem, the petiole, the : midrib, and the lamina of the leaf (especially also in the fruits) are accompanied by wide laticiferous tubes which are arranged symmetrically, 2-6 around each bundle, and in fact around both phloem and xylem; usually, however, they are not in direct contact with the bundle, but are separated from it by 1-2 layers of parenchymatous cells, The tubes in the stem and petiole are unbranched, and each consists of a row of cylindrical sacs . standing vertically one above another, which. are about four times as long as broad, and are united to form a cantinuous tube by means of a wide round opening in every cross-wall. Round every cross-wall the tube is somewhat constricted, as was well represented even by P. Moldenhawer. The tubes contain large, homogeneous, strongly refractive spheres * Karsten, Monatsber, d. Berliner Acad. 1857; Gesammelte Beitriige, p. 253.—Trécul, Van Tieghem, Hanstein, 7c, ARRANGEMENT OF THE LATICIFEROUS TUBES. 437 (of resin ?), suspended in a fluid, which is almost always in the highest degree rich in tannin; in M. zebrina alone it is frequently destitute of tannin, according to Trécul. After the action of alcohol or potassium bichromate, a very sharply defined coating. of the wall appears, resembling a contracted primordial utricle, and this likewise shows the tannin reaction (investigated in M. Cavendishii). Besides that in the laticiferous tubes, tannin also occurs as a principal constituent of the contents in single scattered, short, parenchymatous cells, and in isolated cambifurm cells in stem and petiole. The other Musacez investigated have no laticiferous tubes, in spite of the fact that in . the rest of their structure they agree entirely with Musa. In place of them spaces usually filled with tannin are found in cross-sections of Urania speciosa and Strelitzia, in the neighbourhood of the thicker vascular bundles; these appear to be the laticiferous . tubes; the longitudinal section, however, shows that they are only parenchymatous cells or sacs with the contents indicated, which do not even form uninterrupted longitudinal _ rows one with another. Those which succeed one another at different heights rather belong, sometimes to one and the same, sometimes to various other rows of parenchyma. In the rest of the parenchyma, and in the phloem portions of the vascular bundles, scattered tannin-sacs occur, as in the equivalent parts of Musa, Heliconia speciosa and H. Bihai, according to Trécul, never show cells or sacs filled with tannin, with the exception of single scattered ones in the phloem of the vascular bundles, The same applies to H. pulverulenta Lindl. Finally in Ravenala madagascariensis, &c., Trécul could _ not find tannin at all, only the wall of individual cells of the leaf-sheath showed an indi- cation of blue colouring after the action of iron sulphate for twenty days. II. NON-ARTICULATED LATICIFEROUS TUBES. 7. Euphorbiaces. Of this family a number of species of Euphorbia have been in- vestigated with accuracy. In the shrubby, more or less succulent forms of hot countries, as E. splendens, E. Caput Medusz, canariensis, rhipsaloides, &c., the stem shows a relatively thin ring of vascular bundles, or of wood and bast, which encloses a bulky - succulent pith, and is surrounded by a layer of cortical parenchyma, which is also thick. The stouter, thick-walled, main trunks of the laticiferous tubes run close to the outside of . the ring of bundles; they are scattered in the cortical parenchyma, either singly or in small groups. Their course is in general longitudinal, though not rectilinear, but undulating both in the radial and tangential direction. They give off numerous branches which are further ramified through several degrees (cf. Fig. 84, p. 191). Those of the first degree - do not differ from the main trunks, their direction also is the same. The branches of - higher degrees become successively narrower and thinner-walled, those of the last degree * have blunt blind ends, The direction of the higher degrees of ramification is very various; sometimes their course, like that of the primary tubes, is longitudinal; some- times they penetrate in a curved and undulating course between the cells of the cortical parenchyma, sometimes individual branches pass through the latter towards the surface, and reach the inner side of the epidermis, where they end blindly, either at once, or ‘ after running for some distance below it. Other less numerous branches pass through the medullary rays into the pith, in the peripheral region of which they divide into numerous thick branches, each of which usually has an isolated longitudinal course. We therefore find laticiferous tubes scattered in the parenchyma, on the medullary side of the wood also. Reticulate anastomoses do not occur. Where an H-shaped connection between two tubes is found, this depends merely upon the form and direction of blindly ending branches. In those species which have developed leaves (E. splendens) the laticiferous tubes enter them, at first following the vascular bundles, and then sending out numerous branches from these, through the parenchyma, which run in the most various directions, and at last end blindly, 438 PRIMARY ARRANGEMENT OF TISSUES. The tubes are uninterruptedly continuous throughout the entire plant. No trace of an articulation ever occurs (with exception of the formation of isolated cross-walls, mentioned at p.196). It has never been found possible to dissect out from a piece of a stem, a tube of which even one of the main branches or trunks was closed blindly at both ends. Trécul isolated a portion of a tube from the stem of E. globosa, the united ramifications of which were together 93'5 ™™ in length, and had 120 points of branching; yet 7 main branches and many smaller ones were torn off. I have isolated many main branches from the stem of E. splendens up to 50-7o™™ in length, without ever finding a blind end to them (i.e. apart from the shorter lateral branches). In the growing-point of the stem, branches, and roots, and in young rudimentary leaves, the laticiferous tubes extend close up to the extreme apex, even before the development of the first elements of the vascular bundle, and are always to be clearly recognised as branches which start from the tubes in older parts, and penetrate into the meristem in course of formation, Even in the seedling (investigated in the case of embryos of E, resinifera Berg.) the laticiferous tubes behave as described above, with the sole difference that their ramifi- cations are as yet less abundant than in the older plant. It follows from these facts, that all the tubes of the plants in question are branches of a few primary ones, which, originating in the embryo, grow on and on with the stock, and put out their branches into the newly formed-meristems and tissues. Cf. Chap. VI. The embryo and seedling of the native, herbaceous Euphorbiz (Tithymalus, Klotzsch and Garcke) show the same points of origin of the tubes as the species discussed above, while the mature plant shows the same con- tinuity of all of them as branches of a few primary trunks, which first: appear in the cotyledonary node (cf. p. 196). In corre- spondence with the entire growth and struc- ture, the course of all their branches is different from that in the succulent shrubby forms, especially in the stem. In the internodes of E. Lathyris (cf. Fig. 190) the thicker main branches of the tubes lie in the parenchyma of the cortex, outside the fibrous strands which support the phloem of the vascular bundles, occurring isolated and in small numbers between the parenchymatous cells. They i here have a fairly straight course through the & 24 internode, and are little, if at all, branched. FIG. 190.—Euphorbia Lathyris; radial longitudinal section In the node, on the other hand, abundant through half the node and the neighbouring portions of the inter- nodes of a mature stem, together with the base of a leaf; slightly magnified ; # pith, 2 portion of a bundle of the leaf- trace, A secondary wood, Z bast layer of the stem. The dark lines marked + are the laticiferous tubes; the short black lines and points in the node are fragments of the web of laticiferous tubes occurring there. From this two laticiferous tubes belonging to the bast, and two hypodermal ones are seen passing upwards, while four pass towards the pith; » hypodermal laticiferous tubes, ramifications, with a variously curved and in- tertwined course, occur (Fig. 190), sending out from here further branches, some of which ascend, with the arrangement de- scribed, into the cortex of the next higher internodes, some enter the leaves and axillary buds, at first following the vascular bundles, while others finally, in these regions, pass between the vascular bundles into the pith of the stem, and then descend scattered singly between the parenchymatous cells of the per- manently succulent peripheral portion. The inner part of the pith, which soon dries up, contains no laticiferous tubes. Finally, numerous thin branches proceed from the node}; these pass along the internodes near the epidermis, and together form a hypodermal system of tubes. They ascend vertically from the node into the internode next above, and usually run between the first and second hypodermal layer of parenchyma, more rarely directly below the epidermis, with a straight upward course and ramifying here and ARRANGEMENT OF THE LATICIFEROUS TUBES. 439 there at an acute angle. At the next point of insertion of a leaf, which they reach on their way, many of these branches, preserving the same position relatively to the surface, bend out into the petiole, pass through this to the lamina, and here divaricate near the epidermis of the lower surface, The great majority of the branches, which are dis- tributed and terminate in this region, certainly belong to the hypodermal system in question. The latter stands in no other connection with the main trunks belonging to the bundles, than that which depends on its points of origin in the nodes; it is separated from them in the internodes by many layers of lacunar cortical parenchyma. Its presence may always be recognised, even on the most superficial observation, by the fact that from almost every slight prick into an internode, though far from penetrating the cortical parenchyma, drops of latex exude. The great majority, at any rate, of the Tithymalus- Euphorbiz have essentially the same arrangement of the laticiferous tubes; e.g. E. Cyparissias, sylvatica, Characias, Peplus, Lagascz, also E. Myrsinites, which is charac- terised by especially large and numerous tubes; many differences occur, it is true, according to the particular species, which chiefly affect the greater or less frequency of hypodermal tubes, the presence or absence of medullary ones, &c., and may here be passed over. No accurate investigations of the laticiferous tubes of other Euphorbiacex exist. Hanstein says: ‘Where the development of the latex itself is inconsiderable, as in Ricinus, Mercurialis, and other genera, there we find its vessels also less abundantly distributed, and less conspicuous. They possess but scanty ramifications and anasto- moses, but on the other hand they have more strongly thickened walls.’ Vogl mentions the Jaticiferous tubes in the outer and inner cortex of Hippomane Mancinella, and calls attention to the great similarity of those of Hura -crepitans with those of the succulent Euphorbiz. 8. The laticiferous tubes of the Urticacesw, Apocynes, and Asclepiades, agree, so far as investigations extend, with those of the Euphorbiz in all essential points, as regards both their form, structure, and ramification, and their development and distribution. This agreement becomes especially conspicuous if the succulent leafless Asclepiadex, of the genera Ceropegia and Stapelia, be compared with the Euphorbiz of similar habit. On the average the laticiferous tubes of the families in question are narrower and thinner-walled than those of the Euphorbiz, but very thick ones occur, e.g. in species of Nerium and Ficus. With respect to the abundance ‘in which they occur in the cortex and pith of the stem, the divarication of their branches in the parenchyma of the leaf, &c., the same differences between individual species of each family prevail, as within the genus Euphorbia. While, e.g. in thick-leaved species of Ficus, they extend their branches abundantly through the parenchyma of the leaf, up to the epidermis, in the leaf of Humulus, according to Hanstein, they are confined to the vascular bundles, and do not extend into their ultimate ramifications. Especially abundant subepidermal rami- fications are described by Trécul in the leaves of the Asclepiadex, Echites peltata, and Arauja sericophora. Among Asclepiadex and Apocynez, a larger number of forms have been investigated, especially by Trécul, e.g. Hoya carnosa, species of Asclepias (A. Cornuti, curassavica, &c.), Physostemma, Centrostemma, Cryptostegia, Stapelia, Ceropegia, Echites, Arauja, Nerium, Vinca, Apocynum, Plumiera, Tabernemontana, and many others. Among the Urticacez, more accurate investigation has been chiefly confined to species of Ficus (F. carica, elastica, repens). Minute comparative investigations on the course and development of the tubes are however still to be desired, not only for the orders last mentioned, but also for the Asclepiadez and Apocynez (cf. p. 198). CHAPTER XIII. PRIMARY ARRANGEMENT OF INTERCELLULAR SPACES * Sect. 132. The air-containing, and sometimes water-containing intercellular spaces have been described in Sect. 51, and in the paragraphs treating of the stomata, the parenchyma, and the structure of the vascular bundle, and their arrangement was ‘described at the same time. Here it is only necessary again to bring forward the fact already mentioned on p. 210, that all the air-spaces in question together form a connected system of communicating tubes throughout the whole plant. Where there are stomata this system opens first into the ‘respiratory cavities’ below them, and then through the slits themselves it communicates with the outside, and thus in the terrestrial and swimming plants, which we have more especially under con- sideration, it also communicates directly with the surrounding atmospheric air*, Secr. 133. Of the dutercellular secretory reservoirs the short cavities have also already been dealt with as regards their distribution, on p. 207. It therefore remains still to treat of the arrangement and course of the secretory passages and canals, and at the same time to take into consideration many pheno- mena of their structure, which were passed over before (comp. p. 206), with constant reference of course to Sect. go. The secretory passages traverse the members of the plant longitudinally at first as prismatic tubes, which usually acquire a round or elliptical transverse section: rarely they appear as more or less elongated sacs with both ends closed blindly, as in the exceptional cases of Tagetes and Mammea, quoted on p. 201, and in many Coniferee. In the great majority of cases they are in open communication throughout the plant, forming a system of tubes, which branches and anastomoses—especially, but not exclusively, at the nodes—and may send out blind, or also anastomosing branches into the foliar expansions. Their arrangement in the members varies greatly according to the groups and ‘even the species, nevertheless it is regular and constant within each of these circles of relationship. Besides those which appear constantly in one species or group of different rank, there are in many cases accessory passages, which may occur in varying number according to the individual or species, or may even be absent, e.g. leaves of Pinus, medullary passages of the -Terebinthaceze, Coniferz, &c. + (Compare v. Héhnel, Verhiiltniss der Intercellularraume zu den Gefassen, Bot. Ztg. 1879, p- 5413 also, Ueber Harz-raume im Kork-gewebe, Bot. Ztg. 1882, p. 161, and iiber gefassfiihrende Holzer mit Harzgingen. Ibid.] ? Compare Sachs, Experimentalphysivlogie, p. 254. INTERCELLULAR SECRETORY RESERVOIRS, 441 Any other form of tissue, and any region may contain secretory passages, even the primary xylem of the bundles. Still in this relation also the phenomena are con- stant, according to species and groups, while in individual cases all possible com- binations occur, as is obvious from the following synopsis of the most important known examples, and from the works therein cited, which are to be compared for further details. For the sake of clearness, and in order to avoid repetitions, the secondary changes belonging to the subjects of Chaps. XIV and XV will often be mentioned in the following pages. The mucilage-canals of the Marattiacee' traverse the parenchyma of the pith and cortex of the stem in great numbers, and branch and anastomose frequently. They are con- tinuous from the cortex into the roots, in which they pursue a directly longitudinal course towards the apex, and here end in the meristem; and into the foliar organs, having a similar straight course in the petiole and rachis, with few branches and anastomoses. Their endings in the foliar organs have not been exactly investigated; thorough descrip- tions of their course in the segments of the lamina are also wanting. The leaf of Lycopodium inundatum (also of L. alopecuroides)? is traversed on the pos- terior side by a mucilage-canal, which runs from apex to base: here it enters the cortex of the stem for a short distance, and there ends blindly. The mature canal is limited, as in Marattia, by closely-connected cells of the adjoining parenchyma, but irregular club- shaped cells are seated upon the latter, which project like hairs into the cavity of the canal, In the young leaf these cells form a strand 4~5 rows thick, where the future canal will be, and they have the form of angular meristematic cells; as the leaf unfolds they separate from one another, while the surrounding tissue extends to a corresponding extent, and they elongate to a club-shape, while the mucilage appears between them. They thus constitute the epithelium of the passage, which is still more dissociated than that in Marattia. A similar small passage is found in the marginal expansions on the dorsally winged ridges of the leaves of the spike in L. annotinum. In the stem of the Cycadee mucilage-passages are also distributed through the paren- chyma, in specially large numbers in the cortex: they have branchings and anastomoses. But branches from those of the stem always pass into the leaves, and there end. They traverse the petiole and rachis of the pinnate leaves longitudinally, in varying number according to species and individual—in one small leaf of a seedling of Zamia longifolia I found, e.g. in the petiole only two, in the larger leaves of stronger plants they are numerous, and their distribution in the parenchyma is generally irregular. They enter ‘the pinnz only in forms of Dion, Encephalartos, and Stangeria®, in the first genus running sometimes above the vascular bundles, in Encephalartos between them, and indeed alter- nating regularly with the parallel vascular bundles in the same plane with them and at equal distances; in the pinna of Stangeria they lie above and below the vascular bundles of the rib, without passing out laterally into the lamina. In the specimen examined one central bundle, and one lateral one near it on either side ran into the rib. One mucilage- passage lies between the middle bundle and.the upper epidermis, and one on each side between the lower epidermis and the gap which separates the middle one from the lateral ‘bundles. Among the Conifere all investigated species, with the single exception of Taxus, have resin-passages or resin-reservoirs, which vary in distribution and number according to the species. Starting from the leaves‘, in those examples which have one median vascular bundle or pair of bundles—as in the investigated species of the Cupressinez, Sequoiex, 1 Harting et de Vriese, Monogr. des Maratt—Frank, /.¢. 2 Hegelmaier, Botan. Zeitg. 1872, p- 844. ; % Kraus, Cycadeenfiedern, /.¢. p. 328. * Thomas, Coniferenblatter, /.¢, 442, PRIMARY ARRANGEMENT OF TISSUES. Taxinee, the genera Saxegothea, Dacrydium, Podocarpus (excepting the section Nageia), and Tsuga with exception of T. Douglasii Carr—there is one constant resin-passage between the bundle and the epidermis of the lower surface of the leaf, either close to the latter, often as a keel or ridge projecting outwards as in species of Juniperus, Thuja, and Biota, or deeply embedded near to the bundle, as in Cunninghamia (Fig. 191). Besides these there are in many species (e. g. Cryptomeria) accessory passages corresponding in their arrangement to those which are constant in the Abietinez. One of these lies (with the exception as above of Tsuga) at each lateral margin of the leaf close to the upper surface; either this alone is present, e.g. always in Larix and Cedrus, or three are also accessory ones in the hypoderma, the number and arrangement of which vary FIG. 191.—Cunninghamia sinensis; transverse section through the leaf (220). 2 Jower, o upper surface; % resin passage, xs hypodermal fibres, s sclerenchymatous fibres scattered in the parenchyma, g xylem of the median bundle, ¢ its border of tracheides, Below, and towards the resin Passage, is the thin-walled phloem; the white band at its margin towards the Parenchyma surrounding the resin Passage is the compressed primordial tissue of the phloem. g transversely-elongated parenchymatous cells of the middle of the leaf. according to both species and individual: e. g.in the needles of Pinus sylvestris to such an extent that 1-22 of them have been observed. In the leaves of Sciadopitys 4-10 passages lie under the epidermis, which are dis- tributed in varying symmetry over the margin and lower sides in relation to the simple or double leaves. The leaves of Araucaria, Dammara, and Ginkgo, with several vascular bundles, .are traversed by at least as many passages as bundles ; they alternate with the latter in almost the same plane. The passages in most cases pass continuously through the elongated leaves from the base upwards, and end blindly at a distance from the apex, which varies sometimes with the individual, sometimes with the species, Thus the median Passage in species of Podocarpus, stops far below the middle of the Jeaf. In the lamina of Ginkgo, in place of the INTERCELLULAR SECRETORY RESERVOIRS. 443 uninterrupted canals, there: are between the vascular bundles short cylindrical sacs 1™m or more in length, which are closed blindly at both ends. In the scale-like leaves of many Cupressinex, as Thuja, &c., the passages are of course short, and relatively broad, and may more properly be termed gaps or cavities. The passages and cavities of the leaves are continuous from the insertion of the latter into the primary cortex, and there pass perpendicularly downwards. In the transverse section they form a ring lying in the cortical parenchyma, and are generally grouped according to the arrangement of the leaves. At any rate in a large number of forms they end blindly above the insertion of the lower leaves, without open communication with other passages. Thus in the investigated Cupressinee with whorled leaves, as Thuja, Biota, Juniperus. In J. communis e.g. a large passage enters the stem from each leaf, and there runs downwards, in one of the three angles, to a:point close above the plane of insertion of the next lower whorl, and there ends. On the other hand, in Pinus sylvestris, Abies excelsa, and, according to Mohl’s state- ments ?, in the Abietinee generally, the passages which come from the leaf, after descend- ing through numerous internodes, open into others belonging to lower leaves; the point of confluence corresponds to a widening of the passage on which the higher one is inserted. The passages of the primary cortex are therefore connected into a system of communi- cating canals. The composition of this, as well as its distribution in individual forms, remains to be more exactly investigated. The great majority of the Conifer have no others in the primary tissues of the stem besides the cortical passages above mentioned. This is the case in all investigated Taxinex except Ginkgo; most Cupressinex, Podocarpus, Cedrus, Abies, Tsuga, Pseudo- larix.—Araucaria Cookii and Brasiliensis, and Widdringtonia cupressoides, have besides, according to van Tieghem, a passage in the phloem of the primary vascular bundles, which stops short before the exit of the bundle into the leaf. In the species of Pinus s. str., Larix, Picea, Pseudotsuga there is also a passage which is not continued into the leaf, but it does not lie in the phloem, but in tle xylem of the primary bundles. Finally, Ginkgo biloba has large passages in the pith in addition to the cortical ones. In transverse section there are one or two present, and so arranged that they correspond to the insertions of the next higher leaves. Nevertheless they end blindly both down- wards and towards the petiole, though the canal situated in the latter above the vascular bundles lies in the ideal prolongation of the medullary canal situated opposite the leaf in question. : In the Root the passages of the primary cortex are wanting in all investigated Conifere, and in most cases also those of the vascular bundles. In the latter, however, they are found in certain species or groups; thus, according to van Tieghem, in Araucaria Cookii and Brasiliensis there are five in each phloem portion of the diarch bundle, in Widdring- tonia cupressoides one in the same position. The Cedars and Firs (Cedrus Deodara, Abies pectinata, balsaminea, Brunoniana) and Pseudolarix Kaimpferi have a canal in the middle of the radical bundle. In the Pines (Pinus s. str.) and Larix one passage lies between the two shanks of each vascular plate, described on p. 357. Alismaceex and Butomes. For Alisma Plantago Meyen, and especially Unger 2, have given exact descriptions of the canals with milky contents, while Frank has cleared up the history of their development. According to Unger’s description the passages are absent from the roots, but are distributed throughout the rest of the plant. In the rhizome they traverse the parenchyma, forming a network branching in all directions, and with a course independent of the vascular bundles. Those which enter the petiole and peduncle branch off from this network, and then take a longitudinal course, being 1 Botan. Zeitg. 1859, p. 333- ? Meyen, Phytotomie, Taf. XIV.—Unger, Das System der Milchsaftgange in Alisma Plantago, Denksch, d. Wiener Acad, Bd. XIII. 1857.—Van Ticghem, /. c. 444 PRIMARY ARRANGEMENT OF TISSUES. connected on the way by occasional transverse anastomoses. Those of the peduncle only occur in the hypodermal parenchyma. In the lacunar tissue of the petiole lie numerous small vascular bundles at the periphery, and five arranged in a curve in the middle. Ex- ternally to each peripheral bundle, and between it and the hypodermal layer of cells, lies one passage; one wider one, with its epithelium abutting directly on the epidermis, alternates with each pair of peripheral bundles. Around the inner bundles there is in the layer of parenchyma surrounding them, one passage opposite the points of insertion of each of the plates of parenchyma separating the air cavities. In the lamina of the de. veloped foliage-leaves the passages appear on both sides immediately beneath the epidermis: their main trunks accompany the chief vascular bundles of the leaf: their very abundant branches together form a completely closed net, the meshes of which do not coincide with those of the network of bundles. The linear primordial leaves of the young plant have only three passages, which accompany the three vascular bundles, and only come together at the apex of the leaf. Those which enter bracts and sepals are only connected one with another at the base of these organs by anastomoses, and then run parallel towards the apex, and end blindly short of the latter. In Sagittaria sagittifolia the passages are arranged in the cortex of the stolons in two circles, the one peripheral, the other internal, near to the cylinder of vas- cular bundles; in the petiole, which has a structure like that of Alisma, pas- sages are found between the epi- dermis and those vascular bundles which do not abut directly upon it; there are others, which are arranged between the bundles, and singly at the points of union of the plates of paren- chyma which separate the air cavities: (van Tieghem). Similar conditions to those in the Alismacez described are found in Hy- drocleis Humboldtii, for the details of which reference must be made to Schleiden? and van Tieghem. Here also the passages are wanting in the roots. Among the Aroidee, according to FIG. 192.—Philodendron Imbe Hort. Halens; transverse section of a the observations of Trécul? and van “ihetihe wise corer'g'on margin cttaeaion mse ane Tieghem %, there are passages contain- obliquely-shaded radial bands, w, are the phloem groups ; 2 periderm; ing resin and ethereal oil in the genera é fibrous bundle surrounding a laticiferous intercellular passage. & a Philodendron, Homalonema, Schisma- toglottis, and gum passiges in many species of Aglaonema. The other genera of the family in question, as far as investigated on this point, have no passages: this applies on the one hand to all those which have true milk-tubes, and on the other to those whose vascular bundles are accompanied neither by milk-tubes nor by rows of tannin-contain- ing sacs, comp. p. 436. The resin passages of numerous investigated species of Philo- dendron traverse all the members of the plant longitudinally as narrow canals, apparently (but nothing is stated explicitly on this point) in such a way that they are all connected one with another at the nodes and other points of insertion. In the lateral roots, the stem, and petiole, they are scattered in the parenchyma, forming in the cortex of the root 3, 4-5, or even 8 (Ph. Melinoni) more or less regular concentric rows (Fig. 192) i * Grundziige, 3 Aufl, I. p. 267. * Comptes Rendus, tom. LXII. p. 29 (1866). Structure des Aroidées, /.c. INTERCELLULAR SECRETORY RESERVOIRS, 445 and occurring in stem and petiole either in the peripheral zones of parenchyma only, and even within the hypodermal collenchyma, or also (Philodendron hastatum, tripar- titum, micans) internally, between the vascular bundles. In the leaf-lamina they run in the parenchyma between the tertiary branches of the bundles and parallel to them; either about the middle-plane of the leaf (e.g. Ph. micans, lacerum, crinipes, Imbe, &c.) ; or near to the lower surface of the leaf, and only separated from its epidermis by, 1-2 layers of cells (e.g. Ph. eximium, Rdeeanuk, Sellowianum, pinnatifidum, cannefolium, &c.). The passages of Schismatoglottis, Homalonema rubescens, and H. Porteanum resemble those of Philodendron in fundamental points of form and course, but with the limitation that in the stem of H. rubescens, instead of elongated canals, there are found elliptical cavities from 0-25™™ to 0.50™™ in length and 0-20™™ to 0-38™™ in width. It is remark- able that, according to Trécul’s account, the canals and cavities are entirely absent | in H. Wendlandii. The parenchyma of the stem of Aglaonema marantzfolium is traversed throughout its whole length by gum- (or mucilage-) passages about 0-24™™ wide, which, however, are not continued into the leaves nor into the peduncle. In A. simplex these are absent. Van Tieghem found similar gum-passages in the petiole and midrib of the leaves of Monstera surinamensis, in the cortex of the stem, and in the petiole of Rhaphidophora pinnata, and in the lower part of the petiole of Anthurium crassinervium, while they are absent in M. Adansonii, Rh. angustifolia, and Anth. violaceum. The above-mentioned resin-passages have, as was above indicated, a typical structure; in the roots of Philodendron their epithelium, which is composed of 2-3 layers, is sheathed by 2~3 closely connected layers of narrow elongated sclerenchymatous fibres (comp. p. 202). The resin-cavities in the stem of Homalonema are enclosed by several layers of thin-walled cells arranged in radial rows (apparently derived by division from ' so many primary cells); the inmost project uniformly in a convex manner into the cavities. The gum-passages above mentioned are surrounded by a layer of small cells, - often projecting, as in the Marattiacez, into the passage: these differ but slightly from - those of the adjoining parenchyma. In the rhizomes of the investigated species of Canna, and also in the lower portions of the flowering stems, there are numerous passages: these are filled by a clear transparent mucilage, which oozes out in glistening drops when they are cut through. The passages are absent from the cortex, and are very numerous within the periphery of the vascular cylinder: in the middle they are less frequent. ‘They traverse the rhizome longitu- dinally, their endings have not been observed ; anastomoses and points of branching were found here and there. Their walls are composed of small cells with abundant proto- plasm, which often project as irregular papille into the passage. There has been as yet no thorough investigation of their development. Among the Composite’ all investigated forms of the section Tuifore have a system of oil passages characterised by complexity of composition and uniformity of arrangement. There are no investigations. at hand concerning the Labiatiflore. In the Ligulifloral Cichoriacee they are absent, with the exception of isolated cases to be mentioned ultimately. * “In the roots of the Corymbifere and Cynaree the passages lie in the innermost portions of the primary cortex, and when typically arranged they form a simple curved series opposite each phloem group of the axile vascular bundle, thus alternating with two xylem plates of the latter. According to the usual plan of structure of the roots, the cells’ of . the inner layers of parenchyma are in these plants also arranged in regular radial and concentric rows; between the angles of junction of any set of four there is a 4- or 3- angled intercellular passage; the inmost layer has the properties of the endodermis. In the - simplest case the angular intercellular passages, lying at the point indicated between the - endodermis and the next outer layer of parenchyma, assume the properties (i.e. the * Van Tieghem, Canaux Sécréteurs, /.¢. 446 PRIMARY ARRANGEMENT OF TISSUES. contents) of oil-containing passages. There they together form the curved series, and are separated from one another laterally by a single cell only. According to van Tieghem the oil passages, especially in Tagetes patula, arise at the point indicated, between con- centric layers of cells, which are derived by tangential division of the originally simple endodermal layer, the latter remaining simple opposite the plates of xylem, In the seedling of Helianthus annuus both cases may be found side by side in the same main root. Iff addition to these normal passages there are not uncommonly other peripheral ones, which arise in the same way between the. layer surrounding the endodermis and the next outer layer. All the passages are at first narrow, the two lateral and outermost ones of each normal curved series are triangular, the rest quadrangular. When a peri- pheral series is present, its members may coalesce with those of the inner series by splitting of the cell-wall separating them. The number of the passages of any normal curved series or group shows individual variations even in one and the same root, but ‘comparison of a number of cases shows that there are different average numbers which are characteristic for groups, genera, and species. These numbers are highest in those Cynaree which have been investigated—the following quotations refer in each case to a single group opposite one mass of phloem;— there are ro and more in the diarch main roots of Carduus pycnocephalus, Silybum marianum, Xeranthemum cylindraceum, the diarch or triarch roots of Centaurea atro- purpurea, Echinops exaltatus ; 15-20 in the diarch main root of Cirsium arvense; 12-15 in the tetrarch subsidiary roots of Serratula centauroides. Calendula officinalis has 8-10; Venidium calendulaceum 3-5. E In the Senecionez the number decreases; e.g. Helianthus annuus, in the tetrarch main root, 5-87; Gnaphalium in the diarch root, 5-8; Tagetes patula, in the diarch main root, 5-7; Arnica Chamissonis, Tanacetum vulgare, tetrarch, 4~6; Cotula matri- carioides triarch, 2; Achillea millefolium the same, 1-3; Senecio vulgaris, tetrarch, 2, sometimes united into one; Pyrethrum Parthenium, triarch, 1, rarely 3, &c. Among the Asterez van Tieghem found 6-8 in a triarch root of Inula montana, but only one in Bellis perennis, Erigeron glabellus, Aster, Conyza, and species of Solidago. In the latter cases, especially in Solidago limoniifolia, the canal may be greatly widened, by the separation of the cells which originally limited it externally, so that it extends as far as the next outer layer, or even continues, by further separation of cells, into several layers which lie further out. Among the Eupatoriex a triarch root of Tussilago Farfara showed 5—7 passages; a similar one of Ageratum conyzoides had 2-3; Petasites niveus and Eupatorium aroma- ticum have 1 each, which is extended as in Solidago. As the primary cortex is stretched by the secondary formation of wood and bast, the primary passages of the root, at least in the investigated Senecionez, remain in their place, while they increase to a variable extent in width, and the cells surrounding them in number by division. Compare the details on Tagetes patula in van Tieghem /, c. and the drawing of Radix Artemisiz in Berg. Atlas, Taf, XV. In the stem of the Composite in question the oil passages are only absent in relatively few exceptional cases: Echinops exaltatus, Gnaphalium citrinum, according to van Tieghem. In by far the most numerous cases they are continuous from the root through the stem with its branches and leaves, but subject to branchings, or increase in numbers, which will be described below. They are seated in the primary tissue of the stem always in close contact with the outer side of the plerome sheath, which in the Compo- site (p. 415) may be followed from the hypocotyledonary portion through the whole stem, covering the outside of the ring of vascular bundles. In the hypocotyledonary stem the passages have the same structure and arrangement as in the root; -higher up, and especially from the cotyledonary node onwards, they change their arrangement, accord- ing to that of the vascular bundles, in a manner to be mentioned immediately ; they are * Compare Sachs, Botan. Zeitg. 1859, Taf. VIII. Fig. 7. INTERCELLULAR SECRETORY RESERVOIRS, 447. r + separated from the plerome-sheath by a special layer, often consisting of numerous small epithelial cells. In addition to those which traverse the primary cortex, there are in certain species, but not all, other passages situated at the periphery of the pith, but these are only to be seen above the cotyledonary node. As regards the special distribution in the transverse section of the stem van Tieghem quotes the following special cases, (1) Only cortical passages are present in contact with the plerome-sheath. (a) Only one passage in the middle of the outer margin of each main leaf-trace bundle; Senecio vulgaris, Kleinia ficoides, Cineraria maritima, Flaveria contrajerva, Bellis perennis, Petasites niveus, Baccharis halimifolia, &c. (4) The same, but in addition as many passages opposite the outer margin of each united leaf-trace (faisceau réparateur) as single bundles of the trace coalesce above to form the united trace: Aster. (c) On each side close to the phloem of each main bundle of the trace is one passage : Tagetes patula, Arnica Chamissonis, Tanacetum vulgare, Cotula matricarioides, Ana- cyclus Pyrethrum, Pyrethrum Parthenium, Santolina Chamecyparissus, Achillea Mille- folium, Zinnia elegans, Inula montana, Cirsium arvense, &c, (d) An unequal number, e.g. 3-5 passages opposite the outer margin of each main bundle: Centaurea atropurpurea, (e) A group of passages opposite each lateral margin of the phloem of each main bundle: Silybum marianum. : (2) Cortical and medullary passages are present, the latter opposite the xylem of the bundles. (a) Medullary passages only opposite single bundles, e.g. two: Ageratum conyzoides. (4) Opposite each -bundle of. the leaf-trace is externally a cortical, and internally a medullary passage: Solidago limonifolia. (c) Opposite each bundle of the leaf-trace one medullary and several cortical passages occur: Serratula centauroides, Dahlia variabilis. (d) Near each bundle is one group of medullary and one of cortical passages: Carduus . pycnocephalus, Spilanthes fusca, (e) Opposite each bundle is a curved series of medullary and another of cortical pas- sages: Helianthus tuberosus. — The petioles and leaves have no oil passages when they are absent in the stems bearing them, while they are rarely wanting when the latter do contain passages: Xeranthemum cylindraceum, Cirsium arvense, radical leaves of Lappa grandiflora. In most cases oil passages are present in the leaves, and especially as direct continuations or branches of cauline passages ; often also others occur which may be called accessory. The former may, however, be called fascicular, since they accompany the bundles, either directly, like the cortical ones, or in close proximity to the parenchymatous or endodermal sheath, which each individual bundfé takes with it from the stem into the leaf. In single plants, as Tussilago Farfara and Cineraria maritima, they are intercalated in the sheath itself. Their number and arrangement about each bundle, as seen in transverse section, differ similarly according to the species, but still more variously than in the stem, as may be seen from the examples quoted by van Tieghem, /.c., pp. 118 and 133, &c. They either accompany the branches of the vascular bundles through the lamina (on this point closer investigation is required), or they are limited to the midrib or rachis, as in the leaf of Tagetes patula, where they do not enter the lateral segments of the leaf. Their form, average size, and limitation are, at least at first, the same as in the root. In addition to these fascicular passages there are in single cases, in those species which form passages in the secondary bast of the stem, others situated in the last-formed parts of the phloem of the petiolar bundle: e. g. Helianthus annuus. Van Tieghem found accessory passages in the leaf of Solidago limoniifolia: beneath the _ epidermis of the lower surface, and separated from it by 1-2 layers of collenchymatous cells, there is a series of 3-5 narrow canals on either side of the mid-rib. Again, in 448 PRIMARY ARRANGEMENT OF TISSUES. Tagetes patula: neither the cotyledons nor the lateral portions of the lamina receive fascicular passages in this plant. On the other hand there is on either side along tie margin, in the parenchyma of the lower side ‘of the leaf, an unbroken series of oil-con- taining (schizogenetic) sacs, closed blindly on both sides (comp. p. 201). ; ‘Finally, in the Cichoriacez the oil passages are wanting in all parts in most species investigated, and in the stems and leaves of all species. But in the root of Scolymus grandiflorus van Tieghem found them in groups of five, with exactly the same position, origin, and structure as the primary ones of the roots of the Senecionee. And in the diarch main root of Cichorium Intybus and Lampsana communis there are also rudiments of passages, there being the characteristic cell-divisions, but no opening of the passage at those points of the endodermis where they arise in other Composite. Scolymus is therefore the one known member of the Cichoriacex which really has oil passages in addition to the laticiferous tubes. The case seems to be different with the relations between the occurrence of oil passages and that of the sacs with milky contents, which accompany the vascular bundles (comp. p. 149). At least van Tieghem found both organs side by side in the upper part of the stem and leaves in Cirsium arvense and Lappa; it is true that in the leaves the passages soon stop, and the sacs become the more numerous. The same seems to be the case in other species, e.g. according to Trécul’s and van Tieghem’s statements ‘in Cynara Scolymus ; but on this point further investigations are wanted. All investigated Umbellifere 1 have without exception a very complex system of longi- tudinal usually anastomosing sap-passages, the contents of which are ethereal oils with resin or milky mixtures of these bodies with mucilage and solutions of gum. In the primary tissues of the root the passages lie exclusively at the periphery of the vascular strand, directly within the endodermis. Their formation starts from a portion of the one-layered ring of pericambium situated in each transverse section opposite the angle of each vascular plate. The number of the cells of this portion is always even; e.g. 6, 10, 12; the radial wall, which separates the two central ones (it may be called the middle wall), is opposite the outermost vessel of the plate, in the extension of its median plane. On both sides of it lies an equal number (e.g. 3, 5...) of passage-forming cells, These are at first rectangular in transverse section, and somewhat radially elongated. Each then divides into two cells by a wall, inserted at the middle of its outer wall, in- clined at about 45° to the radial wall facing the elongation of the vascular plate, and meeting this further out than its middle ; of these two cells one is large and irregularly 5-angled in transverse section, and one small and triangular. . The triangular cells lie at the outer limit of the pericambial layer, the 5-angled cells extend through its whole ’ thickness; the middle radial wall of the portion of the ring is met by two, all the rest by one of the inclined walls. At the angles between each triangular cell and ‘its 5-angular .. sister-cell an oil passage arises by splitting of the walls; this passage is situated at the middle radial wall, and is quadrangular, being limited externall¥ by the two originally triangular cells which remain small, internally by two 5-angled cells; at all other radial walls, however, a triangular passage is formed, limited externally by one triangular, and internally by two 5-angular cells. Opposite each vascular plate both of two- and of many-rayed bundles there arises a curved series of passages, the number of which must always be uneven: one central one and on either side of it the same number of lateral ones, The absolute number vdries in species and individual between five and thirteen . Opposite each vascular plate. The middle quadrangular one of the passages of each group is the largest, and the width of the rest diminishes as they are further removed from it. . In addition to these passages, which correspond to the vascular plates, one srhaller one appears rather later in the middle of each sieve-group. It is pentagonal in transverse * Jochmann, De Umbelliferarum Structura, Berlin, 185 4.—Trécul, Comptes Rendus, tom. LXIIL Ppp: 154, 201 (1866).—N. J. C. Miilier, in Pringsheim’s Jahrb. V. Zc.—Van Tieghem, Ann. Sci. Nat. 5 sér. XVI. Compare above, § 50. INTERCELLULAR SECRETORY RESERVOIRS. 449 section, and is bounded externally by two cells of the pericambial ring, internally by three cells of the phloem group. The whole number of the passages in one root-bundle may accordingly be very large, e.g. 2x 11 next the xylem, plus 2 in the phloem, in the diarch main root of Pastinaca; 4x5 next the xylem, plus 4 in the phloem, in tetrarch adventitious roots of Génanthe pimpinelloides, &c. Adventitious roots like those of the last-named plant may retain this primary structure for a long time, or even throughout life. Usually, however, and in all investigated main roots, it is altered at an early stage. The cambium which appears in the bundle (Sect. 139) produces externally a massive, and for the most part parenchymatous secondary cortex ; the primary cortex, together with the endodermis, is simultaneously thrown off with an abundant formation of periderm (sect. 176) starting from the pericambial cells lying outside the passages, and strong increase and division of the pericambial cells lying within the passages, The passages thus come to lie near to the inner side of the periderm, at first with their original arrangement, later, as the result of the increasing growth in thickness of the cortex, more and more dislocated. They then represent the canals described by Trécul, which lie under the peridermal covering of the roots of Umbellifere. The hypocotyledonary stem of the seedling retains approximately the structure of the bundles and arrangement of the passages of the root till close below the cotyledons. One passage branches out from the hypodotyledonary stem to each of the three vascular bundles which enter there, and has in the cotyledon a position close to and opposite the phloem of the bundle. Though there are no definite statements on this point, it cannot be doubted that a connection exists in the cotyledonary node between the passages hitherto described and those found higher up in the stem. The arrangement of these differs in most cases from those hitherto treated of. In the outer cortex of the internodes there is as a rule one passage opposite each vascular bundle, or each of the stronger ones (Fig. 193). Since one angle of the stem or one hypo- dermal collenchymatous bundle corresponds usually to each of the latter, especially the stronger ones, the passages are thus, as Trécul states, also opposite the strands of collen- chyma. Instead of one only, 2, 3, or even 4 passages may be opposite to one broad strand of collenchyma. According to the species these passages are either near to the periphery, close to or even in the strand of collenchyma; or they have a more internal position in the parenchyma between the latter and the vascular bundle. In the creeping stem of Hydrocotyle vulgaris there is a passage, at the inner limit of the non-collenchy- matous cortex, closely opposite each vascular bundle, with its epithelium directly adjoin- ing the inner face of the endodermis (p.121); the same arrangement appears in the branches of Bupleurum fruticosum, In addition to the passages mentioned, which may be called fascicular, there are in many species more or less numerous ones situated in other positions in the outer cortex; thus e. g. according to Trécul, in all regions of it, from the epidermis to the limit of the ring of bundles in Smyrnium Olusatrum, A2gopodium Podagraria, and Sison Amomum. Trécul distinguishes, according to-the conditions of arrangement indicated, ten types, the number of which may be increased if necessary. Almost all Umbellifere have oil passages in the pith of the internodes of the stem; but Trécul cites Bupleurum Gerardi and ranunculoides as exceptions to this. Also in Hy- drocotyle vulgaris and Xanthosia rotundifolia J find no medullary passages. The flower- ing branches of Bupleurum fruticosum show, according to the same author, in the upper internodes numerous medullary passages alternating with the inner margins of the vas- cular bundles; in the lower internodes the number of them diminishes successively, but they seem, according to the description, which I cannot fully understand, to be present originally also at the base of the branch, and to be crushed by secondary extension of the surrounding cells of the pith. In the seedling of Fceniculum officinale the medullary passages are often entirely absent in the first internodes (comp. Fig. 193); in higher in- ternodes they appear quite isolated at first, but in large numbers in the stronger plant. Gs 450 PRIMARY ARRANGEMENT OF TISSUES. Where the pith is permanent, e. g. in the stems of species of Ferula and the rhizome of Imperatoria Ostruthium ', the passages may. be scattered through the whole pith. In the numerous species with internodes which become hollow they are limited to the persistent periphery of the pith (Anthriscus vulgaris, Myrrhus, Carum Carvi, Heracleum sp.) If they are originally formed in the middle, and this is a point which is not decided, they disappear with the cells of the pith surrounding them. In some.cases, however, the passages persist in the middle of stems which become hollow, either surrounded by some layers of pith-cells, and standing freely and singly in the hollow (Smyrnium Olusatrum), or em- bedded in permanent lamellz of pith, which extend from the periphery into the hollow (Heracleum Sphondylium). ; ; ; The passages described run through the internodes as a rule with a straight, longi- tudinal course, and few branches or anastomoses, There are, however, numerous branches at and near to the nodes, by which all anastomose one with another, and are extended into the leaves and axillary shoots. Blind endings have not been ob- ‘served. Also the sacs described in the old rhizome of Imperatoria are only huge dilatations of the passages. The passages have a similar distribution in the petiole to that in- the stem. Anastomoses, even of reticulate form, occur at the points of insertion of the segments of divided or compound leaves. The branches finally enter the lamina. Here, according to Trécul’s observations on Angelica silvestris, Opoponax, Impe- ratoria, Smyrnium, Ferula tingitana, Lageecia, &c., and also in Eryngium, they accompany the vascular bundles, in such a way that they traverse the nerve both on the upper and on the under side, in the latter position they are on the average larger and more numerous ; there is FIG. 193,— Transverse section through an one in each of the smaller nerves, in the larger there are nternode of a young plant of Foeniculum offi- cinale (40); pith surrounded by the partly. num- bered vascular ring. Between the bundles the cambium zone connecting them is indicated; the small circle outside the stronger bundles is the transverse section of an oil passage; in each of the blunt angles of the stem the transverse section of a fibrous bundle is. indicated in form of a seginent of a circle. Compare p. 241. often several; further they are connected and in open communication by means of their ultimate branches so as to form a network similar to that of the bundles. The sap-passages of the Araliacez contain resin in most cases investigated; according to Trécul they contain gum in Aralia chinensis, spinosa, Panax Lessonii, P. crassifolium, &c. According to Trécul’s investigations of numerous species of the genera Hedera, Paratropia, Cussonia, and the plants already named, they are as generally distributed in this family as in the Umbellifer, and the general plan of their arrange- ment and course in root, stem, and leaves, as well as “heir various modifications according to single species, correspond so closely to those in the Umbbelliferz, that they need not be thoroughly entered into here, but reference may be made for many details to Trécul? and N., Miiller (/.c.). In the roots of Hedera Helix and Aralia Sieboldtii van Tieghem found the number and arrangement of the primary groups near the vascular plates to be not always so very regular as in the Umbellifere, and those in the phloem to be either in contact with the pericambium, or completely enclosed in the phloem. The numerous anastomoses between the radially undulated passages of the primary (and also secondary) cortex of the branch of Paratropia macrophylla observed by Trécul may here also be mentioned, and the statement of the same author, that in the lamina of Panax Lessonii and crassifolium the passages seem to occur in the mid-rib only, and apparently do not follow the lateral branches of the vascular bundles. * Compare Berg, Atlas z. Pharm. Waarenkunde, Taf. 22.—Wigand, Pharmacognosie. ? Des Vaisseaux propres dans les Araliacées ; Comptes Rendus, tom. LXI. p. 1163 (1863). INTERCELLULAR SECRETORY RESERVOIRS. 451 The family of the C/usiacez1 has especially numerous passages containing gum-resin, and in fact all members of it with exception of the genus Quiina, which is separated as a special group. According to Trécul’s and van Tieghem’s investigations there are in the stem and roots of the true Clusiacez three chief forms of distribution of the passages. In the genus Clusia they lie only in the primary parenchyma, in the stem they occur also in the pith, but are absent from the vascular bundles and the secondary cortex. A second category has passages in the phloem of the vascular bundles in addition to the above-mentioned places; there being one in each primary phloem-group of the root- bundle, in each primary bundle of the stem, and further in the secondary bast; e. g. Mammea americana: the same distribution, with the exception of those in the primary bundles of the stem, is seen in Calophyllum Calaba. Thirdly: the passages are wanting in the cortical parenchyma of the root, but are present in that of the stem, and in the primary phloem-groups and the secondary bast both of the root and of the stem: Rheedia lateriflora, Xanthochymus pictorius. They are present in the pith of the stem in Rheedia, but not in Xanthochymus, - The passages run longitudinally and anastomose one with another through the whole plant: more rarely in the internodes, but always, and in various individual forms described by Trécul, in the nodes. From the latter, branches of the passages go into the petiole, anid through this further into the lamina. In these members they are found in most species only in the parenchyma and the watery hypoderma. Only in Mammea americana a number of the vascular bundles entering the petiole take the passage which traverses their phloem with them out of the stem: in the case of the median bundle it goes nearly to the apex of the leaf. The number of those which enter the petiole is usually high, and varies according to species. Trécul states, e.g. that there are 30 in Rheedia lateri- flora, about 40 in Xanthochymus pictorius, 14-20 in Calophyllum Calaba, and more than 200 in Clusia rosea. Their distribution in the parenchyma varies according to the species. Finally, in the lamina also they run without relation to the vascular bundles, even cross- ing these in certain cases, and branched here and there, but without visible anastomoses. According to the arrangement in the massive leaf-substance there may be distinguished a system of internal passages situated with vascular bundles in the inner chlorophyll paren- chyma, and a hypodermal system, of which the passages are always narrower. For details on these relations see Trécul, /.c. Of the forms investigated Mammea americana ‘alone has no passage in the lamina except that which traverses the mid-rib, but has instead a round resin-containing cavity embedded in the parenchyma in each mesh of the network of vascular bundles. 4 Pittosporee ?. The root of Pittosporum Tobira shows originally opposite each vascular plate a group of oil and resin passages of similar origin and arrangement to those in the Umbellifer. The number of the passages in each group is certainly smaller, and their arrangement often less regular than in this order: there is a central quadrangular one, and usually on either side of it a smaller, triangular one. The passages in the phloem-strands are absent in Pittosporum. By the same process of secondary formation as in the Umbelliferz the passages are subsequently pushed outwards, under the periderm, while they widen considerably, and the cells limiting them increase in number. In the primary tissue of the stem there is only one passage in the outer part of the phloem of each vascular bundle, and these passages of the stem are continuous with those of the root at the limit of the two members. The bundles which enter the leaf are each accompanied by one passage, which retains the same position as in the stem, and divides in the lamina into branches, which also follow the branches of different rank of the vascular bundles. 3 1 Meyen, Physiol. II. p. 384.—Anonymous author in Botan. Zeitg. 1846.—Treécul, in Comptes Rendus, LXIII. pp. 537 and 613 (1866).—Van Tieghem, /.c. 2 Miiller, 7.c—Van Tieghem, /.c. Gg2 452 PRIMARP ARRANGEMENT OF TISSUES. The same conditions of arrangement as in Pittosporum are found in the primary vascular bundles of the branches and leaves of Sollya heterophylla, and Citriobatus multiflorus. Bursera spinosa, however, never has. sap-passages, according to van Tieghem. Cactee. The passages containing latex in many Mamillarie (comp. pp. 202, 206) traverse the whole stem, and are scattered through the parenchyma. They are rare in the inner pith, but numerous in the zone of parenchyma situated between the ring of wood and the inner circle of cauline bundles (comp. p. 254), in the whole cortex, and the mamillz. They are branched in all directions, and all branches communicate openly one with another; those which enter the mamille run within them near to the axile strands of xylem, and give off numerous branches, which are repeatedly branched and run through the chlorophyll-parenchyma straight towards the surface, many of them as far as the simple hypodermal layer of collenchyma. In other Cactez these passages are not found. Those found by Schleiden! in Opuntia peruviana differ fundamentally from them. I have investigated them in O. robusta. They lie here close to the outer limit of the phloem (not, as Schleiden states, in it) of the bundles of the trace, which are connected into a net, and follow them in their lon- gitudinal course. They are apparently of lysigenetic origin, being cavities in the parenchyma, which attain a width of 3™™, and are filled with swollen, sometimes still recognisable cells, and numerous conglomerated crystals of calcium oxalate embedded in the mucilage. The investigated Attica etiate molle, Spondias cytherea, Pistacia vera, Lentiscus, Rhus aromatica, suaveolens, Cotinus, Coriaria, virens, Toxicodendron, typhina, glauca, elegans, semialata’, and villosa—are, as regards the disposition of their passages containing mixtures of gum-resins, distinguished by the fact that they are situated in the stem and leaves in the phloem of the primary vascular bundles. Further, in addition to this, there are others in the secondary bast of the stem, from which, in Rhus viminalis, blindly ending branches penetrate here and there horizontally into the medullary rays of the xylem: finally, in many species (Rhus toxicodendron, typhina, glauca, elegans, viminalis, semialata, and Spondias cytherea) there are medullary passages. In the root a relatively large passage lies in the middle of each phloem portion of the primary, usually 2 or 3 rayed vascular bundle. In the secondary layer of bast new ones are’ successively added. o The phloem portion of the primary bundles of the stem is limited from the parenchy- matous outer cortex by a strong bundle of sclerenchymatous fibres of half-moon shaped transverse section, and the fibrous bundles are almost in contact with one another at their margins, thus forming together a ring surrounding the outer cortex. Outside this there is no resin passage, but there is a thick one immediately within it in the phloem of each bundle. In the secondary cortex, which appears later internally, new ones are formed successively in the strands of bast. The medullary passages vary in number according to the species; Trécul gives as the number, e.g. in a transverse section of a branch of R. semialata 58, typhina 25, viminalis 5-12; the larger number are situated especially at the periphery of the pith, the minority scattered irregularly. According to successive transverse sections it appears that a part at least of the medullary passages ends blind in the pith. The cortical passages, as far as they belong to the secondary bast, are connected also in the internodes by more or less numerous tangential anastomoses. In the nodes the cortical passages anastomose both with one another, and also with the medullary passages by means of branches, which follow the vascular bundles out into the leaf; from the plexus of anastomoses the passages are continued down into the next internode and into the leaf. ' Anatomie der Cacteen (Mém. prés. Acad, S. Pétersbg. tom. IV), p. 358, Taf. VIL. 4. ? Trécul, Des Vaisseaux propres dans les Térébinthinées. Comptes Rendus, tom, LXV. p. 17) 1867.—Van Tieghem, 7. ¢. INTERCELLULAR SECRETORY RESERVOIRS, 453 The branches observed by Trécul in Rhus viminalis, which pass into the wood, pass off nearly at right angles from the cortical bundles, and into the medullary rays, without reaching the medullary passages: The vascular bundles passing into the petiole, which are arranged in curves in its transverse section, and branch in their further course, take with them one passage each from the stem: these passages have the same position as those in the primary bundles. The same is the case with the stronger branches of the bundles, while the passage is often absent in the weaker ones. In addition to these passages there are in the petiole of Rhus semialata medullary passages also, which lie 1-3 together on the inside of the strongest bundles; in the petiole of Spondias cytherea one is placed opposite the inner margin of the median bundle. A similar arrangement, the details of which may be read in Trécul’s work, is found in the mid-ribs of leaves and leaflets; these contain several vascular bundles, which turn their phloem, and also the resin-passages sometimes towards the upper, sometimes towards the lower surface of the leaf. All lateral ribs contain but one passage turned towards the lower surface of the leaf, and even this is absent from the . last branches of the bundles. In Rh. semialata and glauca Trécul saw the passages of the leaf-lamina anastomosing in a reticulate manner like the vascular bundles, which they accompany. According to van Tieghem’s investigations of Bursera gummifera, and the transverse sections of branches of species of Balsamodendron and Protium figured by Marchand’, there is found in these balsam-bearing trees of the order Burseracex a structure of the cortex similar throughout to that described in the Anacardiacex, and the same distribu- tion of the gum-resin passages in the roots, the stems and their branches, and the petioles. In the genera Ailantus and Brucea*, now placed among the Simarubee, there are longitudinally-running sap-passages, as in many species of Rhus, at the periphery of the pith of the stem ; in Ailantus glandulosa as many as 60; in other regions of the stem they are wanting. They appear, according to accounts at hand, to traverse the successive internodes, and to give off branches at the nodes into the leaves; this is not, however, dis- tinctly stated. At all events the passages are again found in the petioles and the mid-rib of the leaflets, and in the pith-like parenchyma surrounded by vascular bundles arranged in curves or rings, or lying between these. They are not present in the lateral ribs which are given off from the mid-ribs of the leaflets. 1 L, Marchand, Recherches pour servir 4 histoire des Burseracées; in Baillon, Adansonia, tom, VIL. p. 258, pl. VIII, et tom. VIII, pp. 17, 74, pl. IL, TL. 2 Trécul, Vaisseaux propres des Térébinthacées, /.c. SECOND SECTION. SECONDARY CHANGES. CHAPTER XIV. SECONDARY GROWTH IN THICKNESS OF NORMAL DICOTYLEDONOUS STEMS AND ROOTS. I. Campium. GENERAL ARRANGEMENT OF THE SECONDARY THICKENING. Srcr. 134. In the Dicotyledons possessing an axial bundle, in the anomalous forms enumerated on p. 250, sub. 2, in the Berberideze and Ranunculacee men- tioned on p. 249, and in the Peperomize (p. 249), the primary arrangement of the vascular bundles and of the tissues surrounding them in the s/em undergoes no change after extension is complete. The same statement applies to a relatively small number of forms, the stem of which possesses a normal ring of bundles, consisting of collateral leaf-trace bundles of normal orientation, as in Saurures, and species of Ranunculus. In the very great majority of Dicotyledonous stems, on the other hand, the com- pletion ofthe primary groups of tissue is followed’ by the formation of new additional elements, in consequence of which secondary changes occur in the pre-existing, primary tissues (Sect. 54, p. 224). : These changes proceed from the ring of bundles, and that is the case both in the typical instances where the latter alone is present, and also in other cases where, besides this, medullary and cortical bundles occur. The changes in question consist chiefly in the fact, that from a meristematic zone, called the Camdium or Cambial ring, which passes through the ring of bundles, new elements are added to the latter in the direction of the transverse diameter of the stem, which thus receives a (secondary) growth in thickness, through the addition of new elements. In short-lived stems this growth may soon cease; in long-lived ones, and especially in ‘woody’ plants, it endures throughout life. As these changes go on, the number and arrangement of the primary leaf-trace bundles, and of the primary medullary rays which separate them, either remain as they were originally, or new bundles, the termediate Bundles, separated by medullary rays, appear between the original ones, their appearance sometimes immediately succeeding the primary differentiation of tissues, sometimes occurring later. Apart from the numerous modifications of detail which thus become possible, the origin and position of the Cambium, and the arrangement of the elements pro- SECONDARY THICKENING. NORMAL DICOTYEEDONS. 455: duced from it, which constitute the secondary thickening, are in the main outlines the same in the very great majority of stems of Dicotyledons and Gymnosperms. The phenomena belonging to this category, to be treated of in the present Chapter, may therefore be termed those of the normal growth in thickness, and the stems in question be termed the normal*stems, while those showing a different behaviour may be contrasted with them, as azomalous forms (Chap. XVI). 1. The origin of the Cambium takes place in cases without simultaneous: or previous formation of intermediate bundles, in the following manner. The delicate cells arranged in radial rows, lying on the internal border of the phloem of : he leaf- trace bundles (p. 325), remain meristematic ; divisions by means of tangential walls FIG. 194.—Cross-section through the fully elongated hypocotyledonary stem of Ricinus communis. ¢ gy leaf- trace bundle; ¢c,cé cambial zone; ¢é interfascicular segments of the latter, arising by tangential division of the parenchyma of the medullary ray. For further explanation compare p. 332, From Sachs’ TextbBok, go on in them in the succession to be described below, and the vascular bundle thus grows in the radial direction. Sooner or later the tangential divisions are continued from the lateral edges of the phloem-groups, through a band in each medullary ray connecting these groups with one another, so as to give rise to a meristematic zone in this band also. (Comp. Fig. 194.) The meristematic annular zone, which is thus formed from the bundles, and from the segments belonging to the medullary rays, is the cambium. If the leaf-trace bundles in any cross-section are of unequal thickness, the interfascicular completion of the cambium begins at the edges of the thickest, 456 SECONDARY CHANGES, and then extends, in succession, to those which are less thick. In addition to the example represented in Fig. 194, the process described occurs in the Menispermez, Casuarina, and Begoniz investigated, in Berberis and Mahonia’*,in Aristolochia Sipho and its allies, Atragene, and the woody Piperacez. Cucurbita is also to be men- tioned here, in so far as the two concentric rings of bundtes in its stem (p. 248) behave, as regards the relation.in question, as a single ring curving alternately inwards and outwards. In Cucurbita, as well as in Aristolochia Sipho and Atragene, the behaviour described is most clear, because it goes on especially slowly. In the former plant particularly the growth in thickness by means of meristematic new formation first goes on in the vascular bundles alone, the cells of the parenchymatous medullary rays following it for a long time only by means of radial extension; only at a relatively late period does tangential division begin in the latter, in the order described. 2. No doubt in the majority of cases, the first appearance of the cambium is immediately preceded by the formation of caudene intermediate bundles, and this happens as follows :— (a2) In the broad medullary ray between two leaf-trace bundles, distinct collateral vascular bundles appear, which are again separated from the leaf-traces by distinct medullary rays. The formation of the cambium then starts from the two kinds of bundles, in the same way as in (1). This process takes place in a very evident manner in the internodes of Clematis Vitalba. When the completion of the six leaf- trace bundles (p. 244) has begun, an intermediate bundle appears between each two of them in the usual manner (p. 389). The twelve bundles of the ring which are then present are separated from one another by the same number of radial bands, each about ten cells broad. The cells of the latter now assume parenchymatous properties, while the twelve vascular bundles increase and develope their elements in the radial direction ; finally, the completion of the cambial ring takes place by means of tangential division of an interfascicular layer of cells, starting from the edge of the six leaf-trace bundles. Each intermediate bundle runs longitudinally through the entire internode, and is only attached by its ends to the trace-bundles, in the nodes. A further case, belonging to this category, and characterised by. the very early appearance of the intermediate bundles, is represented by the Rhipsalides with winged corners to the stem, spoken of at p. 261. How far the same process takes place elsewhere in internodes is not certainly known at present; the ‘complementary bundles’ of Ephedra campylopoda, noticed above at p. 247, may belong to the same category. It at any rate occurs very generally in the hodes, even among forms belonging to category 1, in order to form the oblique or reticulate connections of the trace bundles, which here appear in all cases at an early period. (6) Between each two leaf-trace bundles collateral intermediate small bundles of normal orientation appear, consisting only of one or a few radial rows of elements, and separated both from the leaf-trace bundles and from one another by equally narrow, non-equivalent radial bands (parenchyma). Their longitudinal * Sanio, Botan, Zeitg. 1863, p. 373. SECONDARY THICKENING, NORMAL DICOTYLEDONS. 457 course is undulated in such a manner, that within the internode they show alternately lateral union and separation, at short vertical intervals, both with one another and with the leaf-trace bundles, and thus form 4 net with vertically elongated, narrow meshes, which are filled up by the bands of parenchyma. The whole process begins at the lateral margins of the leaf-trace bundles, and is continued from the latter through the interfascicular segments of the original ring. It may therefore be said, that the leaf-trace bundles coalesce, by means of successive increase in breadth at their lateral margins, so as to form a closed ring, which is only traversed by the narrow radial bands of parenchyma, and in which the primary leaf-trace bundles only remain characterised by the fact that they project more deeply into the pith (and by the special structure of this projecting portion). In the typical cases, the whole body thus formed may actually be termed one collateral vascular bundle, with an annular cross-section. ‘The origin and orientation of the cambial zone is, in the cases in question, essentially the same as in those brought forward above. ‘This coalescence of the leaf-trace rudiments to form a ring appears most clearly in those internodes which contain only a few trace-bundles, with a simple course. In the internode of Euonymus latifolius! the rudiments of the traces of the adjoining pair of leaves, each consisting of a single bundle, appear first at two diametrically opposite points, between the pith and external cortex; then the traces of the next higher pair appear in the middle of the spaces between the first. From the lateral margins of these four bundles, according to their order of origin, rapidly succeeding divisions extend through the interfascicular bands, so as to form the small-celled rudiment of the closed ring of bundles; and finally the formation of the definitive tissue takes place in the latter, beginning with new divisions and consecutive differentiation at the points where the formation of the ring started, and advancing towards completion in the same direction as the latter. The whole ring, including the original leaf-trace rudiments, consists finally, especially in the xylem, of alter- nating radial bands of bundles, and of non-equivalent elements, which show the longitudinal course already stated. Fraxinus* behaves quite similarly, and most Rubiacez, Asclepiadez, Apocynez, &c., having a very regular radially arranged ring of wood, are also connected with this type. Comp. Sects. 61, 63. The formation of the closed ring is less evident in the case of internodes possessing numerous leaf-trace bundles, which from the first are separated by very narrow interfascicular bands, e. g. Acer, Sambucus *, &c.; the result, however, is es- sentially the same. How far the closing of the ring proceeds exclusively from the coalescent margins of the leaf-trace bundles, or also from small intermediate bundles arising like those of Clematis, remains to be investigated for each particular case. (c) The coalescence of the vascular bundles to form a continuous ring may go still further, so that no alternating dissimilar radial bands appear between the original bundles, but the whole of the tissue forming these zones assumes the struc- ture of a vascular bundle, if this éxpression be allowed for the sake of brevity; i.e. it consists of the elements of vascular bundles, with similar structure and arrangement to those of the later developed portion, and of the products of secondary growth in + Sanio, Botan. Zeitg. 1863, p. 360. 2 Compare Nageli, Beitr. 1. p. 95 3 Thid. Zc 458 SECONDARY CHANGES. the bundles of the trace. The descriptions which follow below will explain this in greater detail. The formation and orientation of the cambial zone are once more the same as in other cases. Hartig! and Sanio state that this condition exists. in the case of Ephedra monostachya, Cheiranthus Cheiri, and Miihlenbeckia complexa, Hieracium, Pyrethrum, Galium, Plantaginez, and other plants to be mentioned imme- diately. I find it in Cobzea, Crassulaceze ? (Sedum spec., Sempervivum arboreum, Echeveria pubescens), Caryophyllee? (Dianthus plumarius, Silene italica), Rumex lunaria, Campanula Vidalii, Lobelia syphilitica, Xanthosia rotundifolia, and Centradenia grandifolia. Many other Melastomacee, and, according to Chatin’s * description, the Rhinanthacez also, appear to behave in the satne way, but here it is not admissible to draw any certain conclusion from one or the other species, as to the behaviour even of its nearest allies ; thus, for example, Rumex alismifolius, in contradistinction to the R. lunaria just mentioned, possesses the structure given under (4). Comtp. also below, Sect. 147. Sxct. 13g. According to the traditional terminology, the part of the ring lying on the inside of the cambial zone, and including in itself the whole of the xylem groups, is called, in stems of the Dicotyledonous type, the wood or ligneous body (xylem of Nageli), while everything that lies outside the cambial zone is called the cortex. The latter is divided into the Jas? zone, bast or liber* (phloem) which, limited internally by the Cambium, includes and is characterised by all the phloem- groups of the ring, and the ex/ernal cortex *, Duhamel’s Enveloppe cellulaire, lying outside this. In the ligneous body the elements of the vascular bundles form s/rands with the arrangement described, zood-strands; the bast has similar das/-strands, cor- responding in their arrangement to the wood-strands; or, if one will, the two may be termed xylem and phloem s/rands. The bands of non-equivalent tissue-—consisting in the great majority of cases of parenchyma—lying between the strands, and having a radial course, as seen in cross-section, are called medullary rays. Each of the latter consists of a portion belonging to the ligneous body, the medullary-ray of the wood (‘Markstrahl katexochen’ of Nageli), and of a portion lying in the bast-zone (cortical medullary-ray, cortical ray of Nageli). Those medullary rays which are formed on the first origin of the woody ring pass through from the pith to the external cortex. They have accordingly been termed /arge medullary rays, in con- tradistinction to those which arise later, and do not reach the pith, and are thus in this respect smailer rays. With reference to their origin at the first commencement of the woody ring, the former have also received the name of the original, primary rays. The genetic relations which are indicated by the latter term are not, however, the same in all cases for the anatomically similar large medullary rays, as follows from what has been stated above. In the case described under (1) they are identical with the original rays, and thus the expression primary medullary rays is appro- 1 Botan. Zeitg. 1859, p. 94. * Regnault, Ann. Sci. Nat. 4 sér. tom. XIV. p. 87.—Hartig, Z.c. ® Anat. Comparée, p. 221. * «Liber, seu interior corticis amictus, ligno contiguus, fibris reticulatis . . . . compositus.’ Malpighi, Anat. Plant. cap. I. ® Compare above, p. 236. SECONDARY THICKENING. NORMAL DICOTYLEDONS., 459 priate, according to the strict sense of the words. In the condition described under (2) (a) and (4), on the other hand, the large medullary rays have originated secondarily from the primary ones, primary rays in the sense of the first case having no more existence after the completion of the ring of wood. The plants mentioned under (2) (c) have, according to what has been said above, neither large nor primary medullary rays; those which appear later in the wood of Ephedra are all small ones not reaching the pith. The cambial zone, finally, is divided into portions belonging to the strands and to the medullary rays, into fascicular and zterfascicular portions, according as it borders on a medullary ray or on a strand of wood or bast. For Sanio’s terminology, which differs from the above, and the employment of which would seem to be attended with almost insuperable difficulties, comp. Bot, Ztg. 1863, P- 372. In the cambial zone growth in the direction of the radii of the cross-section of the stem goes on, with pauses in winter; growth is followed by corresponding cell- divisions; of the products of division those bordering on the wood and bast are in each case added to these as definitive tissue ; a zone lying between the two, however, remains meristematic, and from this the process of new formation is repeated. The masses of tissue which are added by this process to the wood and bast are the secondary wood, and the secondary bast. In the normal Dicotyledonous type, the differentiation of the two in the entire secondary growth remains essentially the same as, or at any rate quite similar to, that of the ring of bundles immediately after the first formation of the cambial zone and the intermediate bundles. On the side of the wood, new elements, equivalent to the first, are constantly added to the existing medullary rays, in their original direction, in such a manner that in absolute dimensions and number of cells they either main- tain everywhere the same height and breadth, or, as they extend in the radial direction, they increase gradually, and usually relatively little, in breadth; the latter is the case especially in broad many-layered medullary rays, as in the stem of Quercus, Casuarina, Clematis, Atragene, &c. The entire ligneous body accordingly remains divided into the same number of mazu strands or main sections as that of the strands existing between the large medullary rays, on the completion of the ring; and these strands become successively broader towards the outside, being wedge-shaped as seen in cross-section. On the one hand, they consist as before of elements equivalent to their original ones, and the differences successively appearing in them are to be described below; on the other hand, radially-arranged plates of non- equivalent tissue occur within the strands, which are essentially similar to the large medullary rays in structure and direction: there are the small, short, secondary medullary rays, which sever the main section or strand into partial sections. In each successive zone of secondary growth new small medullary rays appear, each of which however, when once started, grows on in the radial direction, like the first medullary rays. Every main section of the wood is therefore divided up by medul- lary rays, which become successively more numerous and successively extend less deeply towards the pith. On the cortical side completely similar conditions obtain to those in the wood; 460 SECONDARY CHANGES. the bast-zone continues to be divided into main sections by the large medullary rays, which at the boundary of the cambium are continuously increased by equivalent elements, and here for the time being show the same breadth as in the wood; each of these main sections is divided into partial sections by secondary, successively smaller medullary rays, which are a prolongation of those of the wood. The equiva- lent rays and sections of the wood and bast fit one on another at the cambial boundary; the younger rays, which in the wood penetrate less deeply towards the pith, penetrate to a lesser distance outwards in the bast; the wedge-like widening of the portions of the strands is necessarily in the opposite direction in the bast to that in the wood, as seen in cross-section. From this distribution, slight deviations, still to be included under the normal type, occur here and there. As such are to be mentioned— (a) The discontinuous medullary rays of Hartig’. In Fagus and in exotic woods certain medullary rays in the wood do not extend to the border of the cam- bium, but end externally within the ligneous strand. It remains, however, to be investigated, whether this phenomenon, which appears on examining cross-sections through medullary rays of no great height, always depends on an actual termination of the ray, and a formation of bundle-elements at its exterior limit, or whether it may not perhaps be due to a vertical upward or downward curvature of the ray, in con- sequence of which its radial prolongation comes to lie in a different surface of cross- section from that through which the knife has passed, and in which the inner portion lies. (6) The medullary spot, characteristic of many woods, which will be described below. (c) The appearance of secondary intermediate bundles arising from the cambium, inside the older medullary rays. This phenomenon occurs in the internodes of Atra- gene alpina in a form which quite agrees with the normal processes, especially in the allied species of Clematis. In the specimens investigated the internodes in their first year all showed the six leaf-trace bundles only, distributed as in Clematis (p. 244), separated by six large medullary rays, and, like the latter, furnished with secondary growth by means of a zone of cambium extending all round the stem. In some of the older internodes, at least two years old, this structure is permanently maintained, while the secondary growth continues; they thus correspond to the type given above under 1. In others, however, intermediate bundles appear, in or after the second year, and in fact, in the most regular case, one appears on each side of each median bundle of the trace, so that the total number of the bundles now amounts to ten. In many cases one, two, or three of the four intermediate bundles are absent, and the total number of bundles is accordingly nine, eight, or seven. The longitudinal course of the intermediate bundles is that given in the case of Clematis; in one case only two of them occurred between two trace-bundles, in a short internode; they anastomosed with each other irregularly in their undulating course. The structure and subsequent growth in thickness of the intermediate bundles are similar to those of the six trace- bundles. Clematis Vitalba frequently shows a similar phenomenon, so that in the twelve. 1 Botan. Zeitg. 1859, p. 94. SECONDARY THICKENING, NORMAL DICOTYLEDONS, 461 large medullary rays, or in some of them, small intermediate bundles subsequently appear, which have an undulating course‘. Here, however, the process appears to be rare; I could not find it in my material, even in stems an inch thick. In the broad medullary rays of the wood of Casuarinz the formation of secondary inter- mediate bundles seems to occur constantly, and therefore, strictly speaking, separates the stems in question from the normal Dicotyledonous type*. The inner, oldest portion of these bundles extends vertically, without interruption, through the whole internode. After some years intermediate bundles appear further outside, which are at first very small, but constantly become thicker as secondary growth goes on, and these frequently anastomose with one another and with the neighbouring main bundles, in an irregularly undulating course. In the older wood, even in branches one inch thick, the space corresponding to the original medullary ray is divided up by an irregular net, with pointed meshes, of small bundles. Menispermum canadense pre- sents the same phenomenon in a less conspicuous degree. To what extent similar bast-bundles, corresponding to the secondary bundles of wood may occur, still remains to be investigated. Sect. 136. When once formed, the cambial ring constantly increases in thick- ness and circumference, while successive reciprocal differentiation of wood and bast goes on. The growth of the transverse diameter of its individual cells takes place in a much smaller degree than the general growth, though in many cases they show an increase within definite limits. On the other hand, a continual increase of the number of cells goes on, by division of the existing ones. The general growth, the divisions, and the differentiation of the products of the division in the direction of wood and bast, proceed during the periods of active vegetation; they cease during the periods of rest in winter. The course of the successive divisions and differentia- tions is obviously only to be determined by investigation during the period of vegetation, especially at its commencement. In order to test and confirm the result thus obtained, a comparison of the resting stage in winter is of service. In order to explain the general course of the divisions, we may, in the first instance, consider the transverse section alone, and assume that the cambial layer, as well as the immediately contiguous youngest layers of wood and bast, consist of entirely similar cells, ranged in radial rows; this assumption applies with tolerable exactness in the case of the bundles of the wood. The following rule was first established by Sanio * in the case of Pinus sylvestris. (See Fig. 195.) All cell-division proceeds from a single, annular, meristematic layer of cells, one row thick as seen in cross-section, which may be called the Zuztal layer. In this every (initial) cell divides by a tangential longitudinal wall into two daughter- cells, one of which once more becomes an initial, while the other becomes a /zssue mother-cell ; the latter change may happen either to the inner of the two daughter- cells, which is added to the wood, or to the outer, which is added to the bast. In the quite regular case of Pinus sylvestris, each tissue mother-cell then divides once by a tangential wall, and the two products of its division become directly converted 1 Compare Sanio, Botan. Zeitg. 1863, p. 127. ? Goéppert, Linnea, XV. (1841), p. 747, Taf. IV. Fig. 7.—Sanio, 2. ¢. ® Pringsheim’s Jahrb, Bd, IX. 462 SECONDARY CHANGES. into elements of the tissue. But even here, in the case of Pinus sylvestris, deviations from the latter rule may be demonstrated. Of the first products of division of the tissue mother-cell, one divides once more before passing over into definitive elements of the tissue : in the case of the wood this is always the outer one; in the case of the bast it is usually the inner, more rarely the outer; or each of them divides once more. Thus in the former case three, in the latter case four elements of the tissue are derived from one tissue mother-cell. The investigation of good transverse sections through the active and the resting . cambial zone of the most various Dicotyledonous and Coniferous woods confirms Sanio’s main result, ob- tained in the case of Pinus sylvestris, as regards the J bundles of the wood (comp. Figs. 196, 197). In every radial row there is one initial cell, dividing tangentially, i r from the division of which a new initial cell and a C tissue mother-cell proceed in each case. After one or two further divisions of the latter, the definitive elements of the tissue are formed. From what is known of Pinus sylvestris, very various differences of detail as regards the divisions by which the latter elements are produced are to be expected as soon as more extended minute investigations of this diffi- cult subject have been undertaken. In most of the cases investigated the longitudinal divisions in the tissue mother-cells take place ex- clusively in the tangential direction, the elements of q the tissue are therefore always originally arranged in iC radial rows, and deviations from this rule are the result of subsequent displacement. Exceptions, however, occur c in the bast of many plants, on the origin of the sieve- ( ubes; here the mother-cell of a member of the latter is divided by one or more excentric walls, directed neither radially nor tangentially, into a member of a sieve-tube, and cambiform cells. Comp. p. 324. This no doubt applies to all the numerous cases of irregular grouping of the sieve-tubes; it is, however, undecided FIG. 19%—Pinus sylvestris; cambiat HOw far displacements occur here, in consequence of zone; crosé-section through a radial row ; after Sanio(6s¢), Aside tewardsthewooa, SUbSequent longitudinal growth of the elements, which z (conjectural) cambial initial cell. On the : : . s side of 7 towards Hare twin-cells of the | Might give rise to the same result in the arrangement wood; on the side of ztowards the bast are twin-cells of the bast; the cell bordering of the latter. on 7 towards the bast is the still undivided ge . . ‘ 8 a4t tissue mother-cell for the bast, if the inter- The variations mentioned in the differentiation pretation of z stated above is correct; Sete . : otherwise the former is the initial, and | and division of the tissue mother-cells are sufficient to an as yet undivided tissue-mother-cell for the wood. show that the initial cells of a cambial ring do not always fit exactly one on another with their radial surfaces, and that even the equivalent products of their divisions in the entire series of radial rows form with one another annular zones which are not smooth, but are often interrupted. Comp. Fig. 196, and Sanio, /.¢., Taf. 5-8. In addition \ Wo jl? oy, ee —oe— H SECONDARY THICKENING, NORMAL DICOTYLEDONS. 463 to this, even casual observation shows that ithe secondary growth on the side of the wood is almost always far more abundant than on the side of the bast. Both the successively developed elements of the wood and bast, and the cells of the initial layer, in certain cases, increase for a time in size, while growth in thickness proceeds, as will be described in greater detail below; from a definite period onwards, however, a constant average size is assumed by all the elements which arise subse- quently, while in other cases the average size remains approximately equal from the beginning. The number of elements in the tangential zones, and hence the number of radial rows, must therefore be constantly increased as growth in thickness proceeds, we C_) . SUAS ie O > O60D 4) A LOY A XO (Ke : Cc 3 Bee UT, a O {SOA SITY ee : AGN FIG. 196. FIG. 197. FIG. 196.—Sambucus nigra; young internode; cross-section (220). P,P limit of the parenchyma of the external cortex. Between P—P and »—+r primary zone of bast (phloem); ¢ cambial zone; g, g (pitted) vessels in process of formation; ~—% parenchyma of the pith. The cross-sections with a double outline, at and above s, are the spiral vessels of a leaf-trace bundle. FIG. 197.-—The radial row x—x from the cambial layer of Fig. 196 (600); 7 seems to be the cambial mitial cell, just divided: & side towards the wood; ~ side towards the cortex. and this takes place by means of radial divisions of the initial cells into two equi- valent daughter-cells, which then perform the functions of initial cells in the manner described. Assuming that the woody cylinder within the cambium undergoes no further enlargement, and that all the secondary elements successively formed are of equal size, it is shown, by a simple theoretical consideration}, that for each radial 1 Nageli, Dickenwachsthum des Stengels, &c., bei den Sapindaceen, p. 15. \ 464 SECONDARY CHANGES. row to divide once into two, it is necessary that as many new elements should be formed in the radial direction as are already present on the radius of the cambial ring. ‘An equal increase in the radial and tangential directions, so that two outer cells would correspond to every inner cell, could only take place if the diameter of the cell were equal to the radius’ (i.e. only in the innermost layer of cells of a stem assumed to be destitute of pith). ‘When the radius of the cambial ring has the length of 50, or 100, or 1000 wood-cells the radial rows must be pro- longed by 50, or 100, or 1000 cells in order to’ be doubled once.” The former of the above assumptions holds good exactly for the boundary of the cambial zone towards the wood; the second only applies to particular cases. In other cases, in which a successive increase of size of the secondary elements goes on, the relation is still less favourable to the multiplication of the radial rows. In fact, according to these considerations, the radial divisions in the initial layer must take place rarely in comparison to the tangential ones, and indeed the former are found to appear here and there in individual cells, in the course of successive secondary growth, without any demonstrable order of succession. In the medullary rays the course of the secondary growth and of the divisions is in general essentially similar, yet at least in the most frequent case, to be described below, of radially elongated, parenchymatous elements of the medullary ray, it may be simpler as regards the divisions, the cells following the growth of the wood- strands for a longer time by radial extension, and the divisions happening more rarely than in the strands; they produce on thé one side a new initial-cell, and on the other a new tissue-cell directly, and without any further previous divisions. In addition to the divisions by vertical longitudinal walls, which alone have hitherto been regarded, /ransverse divisions occur in the formation of the secondary parenchyma, and no doubt oblique ones also in the origination of small medullary rays. These can only be discussed below, after describing the conditions of form of the cambial cells. If we wish to designate by the name cambium a zone strictly distinguished from wood and bast, it consists, in accordance with what has already been sdid, of two, or rather of three, different layers of cells, namely (1) the single zzfial layer, and (2) the “sue mother-cells, including (a) those of the wood-side, and (4) those of the bast-side. The contingent modifications in the case of the medullary rays need not be repeated here. On the two layers of mother-cells border the products of their division,.which already belong to the wood or to the bast. As the definitive forma- tion of these requires some time, and they must at their first origin be similar to their mother-cells, a sharp distinction between them and the cambial zone is usually very difficult in practice, even in the condition of winter’s rest; in descriptions they are therefore usually comprehended under the term cambium. We may distinguish them from the true cambium as young wood or young bast; in cases, however, where this distinction is not practicable, or is a matter of indifference, while on the other hand 4 distinction is required between the mature wood and bast and the collection of zones just described, the latter may be included under the general name of she zone of young secondary growth, or the young secondary growth, the term_young secondary growth being understood as opposed to the developed secondary growth, consisting of wood and bast. SECONDARY THICKENING, NORMAL DICOTYLEDONS. 465 . The séructure of the cambium, and of the young wood and bast, has been given above, as regards the arrangement of the cells when seen in cross-section. In each portion of the cambial zone, the form of the ceils is identical with or similar to the general form of the elements of that section of the mature wood and bast which borders on it in the radial direction. (Figs. 198, 199.) Parenchymatous medullary rays are bordered by portions of the cambium, in which the shape of the cells is almost identical with that in the rays themselves, except that the relative radial diameter is on the whole smaller; while the bundles of wood, and the rare medullary rays composed of fibrous cells, are bordered by cambial cells having the form of elongated fibrous elements. Regarded more minutely, the form of these elongated cambial cells is very uniform in all the known cases. It was first accurately recognised by A. Braun’, though only in- dicated by him, and has lately been fully described by Velten*, It is that of a rectangular prism, of which the radial transverse diameter is smaller than the tangential (on the average about half as long), while the ends of it, owing to an inclination of the radial lateral walls to the radial plane, form a sharp edge, lying radially: and almost horizontally. The inclination of the lateral walls is usually confined to one side, and then directed alter- nately to the right and left, more rarely (e. g. Caragana arborescens, Cytisus Laburnum) the two radial surfaces are inclined to one another like a roof. The steepness of the inclination varies, partly according to the individual case, partly, in its average degree, according to species; and is on the whole the greater, the more the cells are elongated; the relatively short cambial cells of Tilia parvifolia, for example, have their end-surfaces inclined at about 45°, the relatively very long ones of Hamamelis ee Re virginiana are quite gradually tapered and pointed, as three-year-old branch, during the winter's rest (March); tangential sec- seen in the tangential view. Very slightly inclined or tion (ts). acadthe zone of secondary growth and cambium, containing a horizontal terminal surfaces ‘only occur in isolated cases, — fe"Wlary yay above.and bordering on especially above, or at the side of medullary rays. (Fig. ¥°" 199.) From the conditions of form described, it follows that the elongated cambial cells, which, as meristem, are in uninterrupted contact, also form, as a rule, un- interrupted longitudinal rows, and only form alternating horizontal rows in the case of uniform inclination of the two radial surfaces. The absolute average size of the cambial cells varies according to the species, as will be stated below (Sect. 153). In certain cases it remains on the average the same during the whole growth in thickness, or increases continuously for a series of years, until a size is attained which remains approximately constant. And in fact this * Monatsber. d. Berliner Acad. 7 Aug. 1854, p. 50 d. Sep. Abdr. Anmerkg. * Botan. Zeitg. 1875, p. 811.—Compare also Sanio, Botan. Zeitg. 1863, p. 108.—N, Miiller, Bot. Unters. Heft IV... Hh 466 — SECONDARY CHANGES. increase extends to all the diameters, so that the general form remains, if not exactly, yet approximately, the same. ae 4 The séructure of the cambial cells is given in its most important points, by the Fig. 199. Fig. 2o0r. FIGS. 199—201,—Fraxinus excelsior. Internode of the stem, two years old, during the winter’s rest, beginning of March (375). Fig. 199. Tangential longitudinal section through the layer of the zone of young secondary growth, bordering directly on the mature wood; ~ medullary rays.—Fig 200. Radial longitudinal section. D, 4 mature wood ; ¢, ¢ limiting layer between this and the young secondary growth ; the latter has rows of round pits (shown with too dark an outline) on the radial Jateral walls, and passes over on the right into the inner zones of the bast; + medullary ray —Fig. 20x. Transverse section. cg shaded ; c—c limit between this and the y 4h mature wood of the previous year, oung secondary growth succeeding it on the left; 4 bi ast ; ~ medullary rays. statement of their meristematic properties, They are furnished with densely granular pro- toplasm, and with a well-defined nucleus, which is spindle-shaped, and in the elongated SECONDARY THICKENING. NORMAL DICOTYLEDONS, 467 cells is elongated in the same direction as they are; in-the medullary rays of many woody plants (Vitis, Begonia) they contain chlorophyll, and in winter small starch-grains (Vitis, Aristolochia Sipho, &c.). Their cellulose walls are thin and delicate at the time of active growth, yet even here the difference between the radial and tangential surfaces, to be mentioned immediately, frequently and perhaps always appears, or is at least indicated. At the commencement of the winter's rest, the tangential walls of the elongated cells remain smooth and relatively thin; the radial walls, on the other hand, become considerably thickened, the highly refractive thickening mass being interrupted by a single longitudinal row of roundish pits. In the medullary rays a similar thickening takes place, and on their limiting surfaces towards the elongated cells there is a formation of pits corresponding to that on the latter (Figs. 198, 199- zor). On the recommencement of the period of growth, the thickening mass appears to be again dissolved, at least in part. These peculiarities of structure are shared by the cells of the young wood and young bast, as well as by the cambium. These cells may also enter with the latter into the condition of winter's rest. If, during the latter period, the zone of young secondary growth between the mature wood and bast be investigated, concentrically and radially “arranged layers of cells are found, with the radial walls thickened as described; on the one hand, they are sharply limited towards the youngest mature wood, which is distinguished by its thick lignified walls; on the other hand, towards the bast, though here the limitation is less sharp, at all the points where the mature elements have slightly thickened and non-lignified walls. The number of the concentric cambium- like layers varies, frequently even in immediately contiguous radial rows; a fact which finds its explanation in the want of uniformity in the course of the cell-divisions in the latter, as mentioned above. In the simplest case only the single initial layer lies between the mature elements of the wood and bast: in its perfect form I have ob- served this only in the case of Juniperus communis (comp. below, Fig. 207). Usually the cross-section shows 2-4, or even more concentric layers of apparently similar, tangentially flattened cells, and only very accurate investigation teaches that these are non-equivalent, one being always the initial layer, while the others are partly tissue- mother-cells, partly young wood or young bast. The latter fact often appears especially clearly on the side of the wood, when growth begins anew after the winter’s rest, for then certain of the cells are found undergoing extension to form members of vessels, directly and without further divisions (e. g. Vitis vinifera). Thus, even during the winter’s rest, the layers of the true cambium are not distinguished either among themselves, or from the young bast and young wood by any characteristic structure ; on the contrary, the entire zone of secondary growth, whether consisting of all its possible parts, or of the cambium, or the initial layer alone, may enter upon the winter’s rest, and then assumes everywhere the same structure. ' Having finished the description of the zone of secondary growth, we still have to return to the transverse and oblique cell-divisions taking place in it, which were left unexplained above. In the medullary rays these do not occur, or, if they do, are irrelevant. In the elongated elements, on the other hand, they appear as a constant and essential phenomenon. (2) In the tissue-mother-cells they appear universally where short parenchyma- tous cells which do not belong to the medullary rays, and septate fibrous cells, are Hha . 468 SECONDARY CHANGES. formed within the secondary wood and bast; wood-parenchyma, bast-parenchyma, &c. (Fig. 202.) The transverse divisions take place once or oftener, and accordingly the height of the products of division is unequal from the first; from the different inser- tion of the transverse walls on the lateral walls, and the relative thickness of the two in the mature tissue, it may be conjectured that the transverse divisions take place in the one case in the early condition of the tissue-mother-cells (wood and bast paren- chyma), in the other case in cells already belonging to the young wood or young bast (septate fibrous cells); on this point, however, more exact investigations have not as yet been made. Comp. below, Sect. 144. Transverse divisions of the tissue mother-cells, or of the young wood-cells, are found in some instances before the formation of vessels with short members, The members of wide vessels with horizontal limiting surfaces may be shorter than the cambial cells from which they arise; the difference in length may, however, be due to the fact that, as they become wider, their height faces into the horizontal position’. In other cases how- ever, e.g. in that of Vitis observed by Cohn’, and in that of Acacia longifolia observed by Sanio *, the shortening of the members of the vessel is so considerable that it does not find a sufficient explanation in this displacement, but (4) In the initial layer those transverse and oblique di- visions of the elongated cells have to take place, by means of which new small medullary rays, consisting of short Spe in ee are parenchymatous cells, originate within the bundles of the gential longitudinal section through = wood. This follows almost with certainty from the fact the innermost layer of bast of the is diminished by displacement of the oblique terminal sur-" rather necessitates the supposition of a transverse division. same branch as Fig. 198; magnifed that the medullary ray, from its first appearance onwards, as in the latter. s members of sieve- tubes; ¢ a sieve-plate lying deeper = extends through the initial layer, towards wood and bast. than the surface of section; 77 small medullary ray,two cells highs The = The only other possible supposition would be that its remaining elements are parenchy- matous cells of the bast, ihe origin _ first origination is due to divisions which extend, in the Tiree || Same direction, through cambium, young wood, and young bast. The first formation of a small medullary ray from elongated cambial cells is hard to observe, and is at present not very clearly known*. From the position of very small medullary rays, only one cell in breadth, and one or a few cells in height, as seen in tangential sections through the cambium and the zone of secondary growth, it may be stated that they originate either by single or repeated transverse division of an end of an elongated cambial cell, or by the cutting off of a portion of the radial lateral wall of the latter, by means of a (mussel-shaped) wall of division, with its concave side turned towards the 1 Sanio, Botan. Zeitg. 1863, p. 122. ? Bericht iiber d. Verhandl. d. Schles. Gesellsch., Bot. Section, 1857, p. 44. * Pringsheim’s Jahrb. IX. p. 56. * See N. Miiller, Bot. Untersuchungen, IV. p. 181,—Velten, Botan. Zeitg. 1875, p. 842. - ‘ SECONDARY THICKENING. NORMAL DICOTYLEDONS. 469 tadial wall in question. (Comp. Fig. 202 m.) In the first cell thus formed, further transverse divisions may then ensue. In the frequent case where a medullary ray, originating as described, increases in height in the successive zones of secondary growth, both as regards its absolute size and the number of its cells, the various possible modes in which new cells may be added may easily be perceived, but the actual process has not been established with certainty. Further, it is not clear whether a small medullary ray may not be formed by repeated transverse division of an entire elongated cambial cell, or even of several one above another. Sct. 137. As follows from their mode of development, the secondary elements of the wood and bast are always at first arranged in radial rows, if we except many groups of sieve-tubes. A growth in thickness of the masses of tissue lying inside the actual zone of secondary growth, such as would result in a displacement of the radial rows, takes place in the case of the wood exclusively during the development of the innermost layers, which are affected by it in the way of displacement or tangential extension ; at a later period nothing of the kind occurs; in the case of the bast the conditions are no doubt different, owing to the continual widening of the zone of secondary growth, but the displacements following from this, which are to be described below, affect only the old external zones to any considerable degree. The secondary elements must therefore maintain their original radial arrange- ment: (1) When the form and length which they had in the cambial stage undergoes little or no change on their definitive development. (2) When, although they become larger than the cambial cells, they maintain a form similar to the latter, and when, in particular, they have terminal surfaces in- clined only towards the radial plane, where they abut on and penetrate between one another. The first case occurs almost without exception in the medullary rays of normal structure (for the exceptions in the case of Atragene, Casuarina, &c. see below), in most parenchymatous masses of the wood, and in short tracheides, e.g. those of Cytisus Laburnum, &c. It is true that the height and breadth here frequently in- crease somewhat, even in the medullary rays‘, on the transition from the cambium to the definitive condition, but only in a slight degree, which does not alter the general grouping. The second case applies to the elongated elements (fibrous cells, woody fibres, tracheides) of the wood of many, and the bast of most plants which form wood. Even, however, in casés belonging to this category, the longitudinal growth of the parts passing out of the cambial condition is trifling, as will be shown in the following paragraphs. The conditions in question are present, almost without exception, in all parts of the wood and bast of the Conifer; in the Dicotyledons they are tolerably general in the soft bast, with the exception, however, of those cases which are characterised by irregular groups of sieve-tubes. The sclerenchymatous fibres of the bast maintain their radial arrangement, for example, in Carpinus, Corylus, Ostrya, Liriodendron, and Magnolia acuminata and 1 Compare Hofmeister, Pflanzenzelle, p. 164. 470 SECONDARY CHANGES. tripetalat. The above-mentioned elongated elements of the Dicotyledonous wood preserve their radial arrangement, for example”, in Cunonia capensis, Viburnum Opulus, Staphylea, Hamamelis, Nerium, many Asclepiadez, Rhus typhinum, Jatropha: Manihot, Laurus nobilis, Camphora, Aesculus, Verbena maritima, Broussonetia, Catalpa, Paulownia, Hydrangea hortensis, Justicia carnea, Fuchsia, Melastomacez’, &c., which are chiefly, though not exclusively, plants with leaves in whorls; not but what with similar phyllotaxis a different arrangement of the elements may occur, as will be shown by facts to be mentioned immediately. The original radial arrangement is, on the other hand, disturbed or obliterated : (1) In groups of elongated elements, which show a great elongation on transition from the cambial condition to that of mature tissue, in the course of which they insert their tapering ends, which are the principal seat of growth, between each other, and acquire terminal surfaces which are inclined not only towards the radial rows, but also in other directions, or even show curvature and torsion of their ends. (2) When certain of the originally similar elements undergo considerable growth in the transverse directions, on attaining their definitive development. The first case perhaps occurs during the formation of many irregular groups of sieve-tubes, but this has not been minutely investigated, and is doubtful. It certainly takes place, however, in the case of those fibrous elements, which in the mature condition do not show any regular serial arrangement, and which often grow to many times their original cambial length: as in the groups of sclerenchymatous fibres in the bast of many Dicotyledons, e.g. Tilia (Fig. 211), in the fibrous cells, fibres, and elongated tracheides in the wood of Leguminosz (Cytisus Laburnum, Caragana, &c.; most beautifully, on account of the contrast with the other elements of the wood, which maintain the cambial form and length, in Herminiera Ela- phroxylon, comp. Sect. 150), Ulmus suberosa, Morus alba, Celtis australis, Tamarix gallica, Ilex aquifolium, Cornus sanguinea, Pyrus, &c. The second case occurs universally in the development of wide vessels. Ori- ginally similar to the other elements, the members of the vessel often become expanded to many times their initial size; the neighbouring elements thus become not only displaced, but often transversely deformed in the direction of the surface of the vessel, compressed, or even completely crushed, so that mere rudiments remain‘. According to the degree of expansion, and the number of the wide vessels in any portion of the transverse section, the general arrangement of the elements is influenced by them. It need scarcely be mentioned that all the phenomena described may occur in various degrees, so that. cases intermediate between the extremes may exist. Sect. 138. The collective zones of secondary growth, cambium, young and mature wood, bast, &c., which have hitherto been regarded with immediate reference to a single transverse section of stem and root, if 4aced longitudinally, are continued, as uninterrupted layers, both through the successive portions of the same axis, and ? Hartig, Forstl. Culturpfl. p. 256.—Sanio, Botan. Zeitg. 1863, p. 107. * Compare Sanio, Zc. pp. 107, 115. 3 Vochting, Zc. * For a minute description, see Velten, Botan. Zeitg. 1875, p. 809, &c, SECONDARY THICKENING, NORMAL DICOTYLEDONS. 471 from the main ‘axis into the lateral shoots; and in fact each of the distinct layers of any transverse portion is continuous with the equivalent and simultaneously formed layer of the successive succeeding portions; the cambial layer of a shoot formed in one year, or its yearly production of wood being continuous with the like layers in the next year’s shoot, &c. The longitudinal course of the stngle elements, especially of fhane which are elongated, as. appearing in the direction of the ‘long grain’ of the wood and bast, presents a series of remarkable phenomena, which, in a certain though not strict sense, are independent of those hitherto discussed.’ In treating of them we will here leave ‘out of consideration the /orszon of the entire masses of wood and bast in twining stems, in so far as this is in immediate relation to the torsion of the whole. twining part, as the description of the latter phenomenon forms no part of the present work’. In stems with a straight and vertical growth, the elements in question usually have their longitudinal axis in an oblique position, deviating from the vertical, and this is the case both in bast and wood, the latter having been the subject of more exact investigations, which are here to be principally regarded *. The deviation from the vertical is usually less conspicuous in the direction of the radial plane; although it must take place in this direction in the case of the above-mentioned bundles with elements penetrating irregularly between each other, e.g. in the wood of Fraxinus and Cytisus Laburnum. It is clearly seen, even on observation with the naked eye, in the case of the Guaiacum wood, in which the fibres of successive concentric layers have their ends passing between each other ob- liquely in the radial direction (though to a less degree than in the tangential direction). In many woods the oblique position in the tangential’ plane appears more clearly. Asa rule its direction is'the same for all the elements. of each concentric layer, and on observation of the surface is indicated by an oblique.‘ grain’ or striation, running round the whole stem. The angle at which the strize cut the vertical varies, partly according to the species, partly in individual cases. According’ to Braun it reaches its maximum—as much as 45°—in Punica Granatum; then follow Sorbus, _Aucuparia (up to 40°), Syringa vulgaris (up to 30°), A’sculus Hippocastanum (10° 20°); smaller values are more frequent, e.g. usually 4°-5°, rarely as much as 10° in Pinus sylvestris, 3°-4° in Populus pyramidalis, Betula alba, &c. ‘In many cases,’ says Braun, ‘ especially in Pinus, I have convinced myself that specimens with shorter internodes usually show greater degrees of torsion, than those with longer ones.’ The inclination is also said to alter with the age of the tree, becoming greater in the later secondary layers in Punica, and smaller in Pinus. The direction of the inclination has been found to be invariably the same in the case of many trees; right-handed (in the sense of Mechanics) in A’sculus Hippo- castanum, left-handed in Populus pyramidalis. Other trees show one direction as the rule, the other as the exception, e.g. Pyrus communis, Carpinus, chiefly right- 1 On this subject reference may be made to H. de Vries, in Arbeiten des Botan. Instituts zu Wiirzburg, Heft III, and the earlier literature on the subject which is there cited. 2 See A. Braun, Ueber den schiefen Verlauf der Holzfaser, Monatsber. d. Berliner Acad. 7 August, 1854; Botan, Zeitg. 1869, p. 747; 1870, p. 158. - 472 SECONDARY CHANGES. handed Salix alba chiefly left-handed. Further, it either remains the same in the successive secondary layers of the same stem, or, in many kinds of trees, as Pines and Firs, the direction changes, becoming reversed after a number of similarly inclined layers. Among 167 species of Dicotyledonous woody plants and Conifers, to which Braun’s investigations extend, the oblique grain is present in 111}; in the rest, e.g, Pinus Cembra, Populus monilifera, Ulmus campestris and effusa, Fraxinus excelsior, Clematis Vitalba, it has not been observed. The arrangement of the fibres in the Guaiacum wood is different from their uni- form obliquity round the entire stem in the trees hitherto mentioned. Here the grain curves backwards and forwards with short undulations, in each layer of wood, often cutting the vertical at 45°, assuming different directions in successive narrow layers, _ not in the broad (annual?) rings. This arrangement, showing a different direction in every small subdivision of the wood, and the interlacing of the elements, in addition to the radially oblique position and interweaving above mentioned, are the causes of the impossibility of splitting the Guaiacum wood in the radial, and the difficulty of doing so in the tangential direction. The facts mentioned, and especially the reversal of direction, in successive layers of wood, are sufficient to show, what all accurate investigation confirms, that the oblique grain is a purely anatomical phenomenon, independent of the external con- formation of the plant. It is also only perceptible externally in the case of injuries which lead to the splitting of the tree in the direction of the grain, such as frost cracks, splitting of the cortex in the direction of the grain of the bast, e.g. in Tilia, Syringa, Juniperus, and Thuja, or in the case of an excessive local swelling of the layers of wood, starting from branches or roots, as in many trees (Punica, Car- pinus, Populus pyramidalis); this leads to the formation of ridges, which run round the stem obliquely, in the direction of the grain. A plausible anatomical explanation of the oblique position of the elongated elements of the wood is afforded in a general way by their conditions of length. As will be shown below, the elongated elements in a number of woody plants increase successively in length for a series of years. As the total length of any portion of the stern remains unchanged during the secondary growth in thickness, and as, further, no enlargement of individual cells at the cost of others which become obliterated occurs, either in the cambium or its products (with the exception of the relatively inconsiderable phenomena connected with the expansion of vessels mentioned above), but on the contrary, all the cells of any layer parallel to the periphery grow, and become larger, or at any rate not smaller, it follows that the progressive increase in length of the elongated elements must result in their position becoming oblique; in the one case this affects the cambial cells themselves, in the other the fibrous elements in process of differentiation. It may at once be added that in the case of stems with the fibres in a tangentially vertical position, as, for example; Fraxinus, the length of those belonging to successive layers must remain the same, or any difference in length must be equalised by radial obliquity only, a point which has still to be investigated. These considerations render the phenomenon intelligible it its main outlines, but by no means explain all the details. It is open to question whether the above-mentioned differences in- length are sufficient by themselves to SECONDARY THICKENING. NORMAL DICOTYLEDONS. 473: explain the extent of the angle of inclination; the similar direction of the inclination in the individual ‘layers, its reversal, and more especially its alleged diminution in the later layers in Pinus, &c., all remain to be explained. The basis of such an explanation is to be sought in a more complete determination of the size and form of the organs in question, than has yet been made. The undulated course of the woody fibres, which appears on cicatrised wounds, &c., and gives the character to knotted and gnarled wood, may here be excluded from minute consideration, as it is a pathological phenomenon ?. Sect. 139. Plants with typical Dicotyledonous structure of the stem, as well as the majority of those Dicotyledons and Gymnosperms which are anomalous in this respect, form, with very rare exceptions, a cambial ring in ‘¢he roof? at an early period, and this, when once present, shows a completely similar growth and new production to that in the stem, although in each particular case definite special differences exist between stem and root, which are to be discussed below.. The first origin and orientation of the cambial ring must, on the other hand, be different from that in the stem, in consequence of the different original general structure in the root. It originates in the axial vascular bundle. The process begins with growth in thickness and tangential division of that layer of cells which borders on the inner surface of the phloem-groups, and in fact it proceeds from the middle of each of these groups towards its two lateral margins, and thus also towards the outer corners of the xylem-plates. The products of the tangential division are cambium and young secondary thickening. In the usual case of abundant growth in thickness, the tangential divisions eventually reach the pericambial cells lying outside the xylem-plates, and are continued over the latter, thus uniting the originally separate portions of the cambium to form a closed ting. As follows from the original arrangement and form of the xylem- and phloem-groups in the root-bundle, the general transverse section of the cambial ring, at its origin and before its completion, is a figure following the outline of the xylem body; in the diarch bundles it is a narrow ellipse, in those with more than two rays it is a polygon, with as many blunt corners and concave sides as xylem-plates are present. As, however, the cambio- genetic production of tissue on the side of the wood takes place unequally all round, and is the more abundant the nearer it is to the original starting-points of the formation of cambium, the phloem-groups are rapidly shifted towards the outside, and the concavities of the ring become flattened, so that its transverse section soon assumes a permanent, approximately circular form. It is only in the comparatively rare cases of slight growth i in thickness that the concavities between the original xylem-plates are permanent, and that the union of the original portions of the ring round their angles fails to take place; the formation of cambium is in some few cases entirely wanting, comp. p. 355. In other cases the tangential division in the neighbourhood of the xylem-plates is conspicuously less than that opposite the original phloem-groups; the cells in the former position * Compare Schacht, Lehrbuch, p. 67—Goppert, Nachtrage zu d. Schrift iiber Inschriften, &c., u, tiber Maserbildung, Bresl. 1870.—Idem, iiber die Folgen ausserer Verletzungen der Baume, Bresl,, 1873.—Ratzeburg, Waldverderbniss, 1,—Nordlinger, Forstbotanik, I. p. 4 ; ? See Van Tieghem, Ann. Sci. Nat. 5 sér. tom. XIII. p. 185, pl. 3, 4, 8 474 SECONDARY CHANGES. follow the growth in thickness chiefly by means of radial extension, so that, as seen in transverse section, the small-celled ring appears interrupted at the xylem-plates by rows of larger cells, e. g. Cucurbita, Urtica (Figs. 203, 204). The production of tissue by the cambial ring or its segments, when once originated, maintains the same course as that described in the case of the stem. Any special differences in the succession of the divisions are, at the present time at least, unknown. The mass of tissue given off internally is to be termed the wood, the peripheral mass, the bast. Both are divided into wood and bast bundles, which correspond to one another in the same way as in the stem, and may be compre- hended under the name s/rands, and into radial bands composed of non-equivalent elements, which alternate with the strands, and in the root no doubt always consist of parenchyma; these are the medudlary rays. The original large medullary rays are obviously excluded in the case of the root. As regards the arrangement of those rays which are present, and the dis- My {? MH < int Mais a HAN tes Hi hed he Nes L iy ee i fe VA Ro iN 7 ORI CL =a Fig. 204. FIG. 203.—Urtica dioica; transverse section of a small subsidiary root from the rhizome (80). g original diarch xylem- plate ; crossing it are two secondary bundles of wood, separated by broad bands of parenchyma (medullary rays) ; outside each, at s, is the bundle of bast belonging to it. Secondary cortex is present all round, with numerous scattered sclerenchymatous fibres, indicated as dark spots; the whole bounded externally by periderm. c layer of cambium or young secondary growth. The primary cortex has been cast off. FIG. 204.—Cucurbita Pepo; transverse section of the main root of a young plant (40). g xylem body of the axial bundle, its four rays united in the middle by a large pitted vessel; four secondary bundles of wood alternate with them; s the original and secondary phloem-bundles. The primary cortex of the root has been replaced by periderm and cast off. ss position of the strands which is determined by them, two main*types are to be distinguished, though these are not sharply contrasted in all cases. (x) A main medullary ray (usually very broad) appears opposite the angle of each original xylem-plate, the same number of main strands alternating with the main medullary rays, e.g. Centranthus, adventitious roots of Tropzolum, Urtica dioica (Fig. 203) among diarch roots; the main root of Cucurbita (Fig. 204), Phaseolus, Convolvulus tricolor, and many others among the tetrarch forms; adventi- tious roots of Cereus grandiflorus, Clusia, Cucurbita, and Artanthe among the poly- arch root-bundles. Comp. van Tieghem, /.c. (2) The whole periphery of the primary bundle acquires fascicular elements, SECONDARY THICKENING. NORMAL DICOTYLEDONS, 475 between which only smal! medullary rays exist, with an arrangement not corresponding to that of the primary xylem-plates. The entire secondary growth thus forms a cylindrical strand, without main medullary rays, e.g. Taraxacum, Scorzonera his- panica, Rubia, Thuja, Taxus, Cupressus, &c. On the connection of the bundles of secondary wood with the primary plates, see Sect. 152. For the further growth of the bundles and medullary rays when once formed, the successive subdivision of the former by new medullary rays, and the structure of the cambial layer, the same general rules hold good as in the stem. The divisions in the cambial zone also appear to follow in general the same rules as in the latter, though this remains to be investigated. The frequent appear- ance of secondary intermediate strands in the broad medullary rays, especially of fleshy roots, is worthy of notice. As a rule, the beginning of secondary growth takes place in the root imme- diately after the differentiation of the primary tissues; in the cases mentioned of feeble development on the other hand, as well as in the adventitious roots of Clusia, Cereus, arid Piperaceze, it begins relatively late, so that each portion of the root at first remains for some time in the primary condition, Il. Tue Woop. 1. Distribution and form of the zones of secondary thickening. Sect. 140. In the native Dicotyledons and Conifers the wood acquires in every period of vegetation an increment of growth, the development of which begins in spring with the unfolding of the buds, and, with the exception of the roots of Dicotyledonous trees, reaches its end in autumn, starting afresh after a period of rest in winter; in the roots of native Dicotyledonous trees, on the other hand, it continues to make slow progress throughout the winter, and only reaches its end on the beginning of a new period of growth, whereupon it immediately begins again’, The product of each period of secondary growth, corresponding in our climate to an annual period, is, as a rule, distinguished from that of the earlier and later periods, by definite differences of structure in the limiting layers, which are to be described below. It is therefore called an annual zone, annual layer, or annual ring, and its limiting layers just mentioned are called spring-wood and autumn-wood. The consideration of the zones of secondary thickening may conveniently start from those which are severed into distinct annual layers, especially as this is by far the most frequent case. The form of the annual rings has been investigated in the case of trees and shrubs, It is well known, and does not require any detailed statement here, that their average breadth shows great variations in the same individual, according to age, and to the action of more or less favourable conditions of vegetation’, and that under 2 Von Mohl, Botan. Zeitg. 1862, p. 313. _ 2 Compare the works of Nordlinger and R. Hartig, to be cited below.—Further, H. de Vries, Biaticss des Druckes auf d. Bau, &c. des Holzes, p. 96; Flora, 1872, p. 241, 1875. 476 SECONDARY CHANGES. similar, or equally favourable conditions, the average breadth varies considerably according to the species. Compare, for example, the broad rings of Paulownia and Ailanthus, with those of Citrus and Cornus; Pinus silvestris and Abies pectinata with Taxus, &c. In the stem of the young tree the breadth of the annual rings increases, under otherwise similar conditions, for a number of years, then remaining for a series of years at an average maximum, but decreasing again in advanced age. In the yearly shoots formed on the thickened stem the average maximal breadth of ring is attained in the first year, or in the first few years’. It is shown even by superficial observation, which may be easily confirmed by more minute research, that the yearly secondary growth in the lateral branches and -roots of a tree is less than that in the stem. The breadth of each ring is, in the regularly developed s/em, uniform all round, though even in this case it may be unequal in consequence of unequal acceleration of the growth on different sides, the ring thus becoming undulating or eccentric, even to such an extent as to be wholly absent on the deficient side. The rings of one and the same cross-section of the stem often show the most various differences in all these respects, forming, as it were, records of the history of its growth and nutrition ; in certain woody plants, to be mentioned below, such inequalities of growth are typical. Unilaterally unequal development of the rings,and consequent eccentric thickening, are the rule for the /ateral branches of the stem and of the roofs; and in fact in the lateral branches of most Dicotyledonous woody plants, the upper szde is the favoured one, e.g. Acer pseudoplatanus, Alnus, Carpinus, Cornus, Corylus, Crategus, Cytisus Laburnum, Euonymus, Gleditschia triacanthos, Fagus, Tilia, Prunus spec., Robinia, &c.?; on the other hand, in the Coniferous woods, and according to Nérdlinger in Castanea, the under side is favoured. In smadl stems also, which for a series of years have grown upright, and increased uniformly in thickness all round, but have then been permanently brought into the inclined position by the pressure of snow, Nordlinger found that the rings formed from the time when the stems became oblique, were eccentric, and that in the Pines, Firs, and Larches the under side, in Oaks and Beeches the upper side is favoured. In the lateral roots of trees at the places where they arise from the stem, the upper side which is continued into the latter is the one favoured; at a greater distance from the stem the under side usually has the advantage, according to Mohl’s opinion’, but this point has not been decided for certain, Centrically developed roots of trees are, however, not actually rare. In the case of our forest trees a series of investigations have been instituted on the average amount of the annual secondary growth of the stem in its successive transverse sections from base to apex, which of course always determines the general form of the stem*. In some cases the successive surfaces of the transverse section * Nordlinger, Der Holzring, p. 14. ? Compare Nordlinger, Holzring, p. 20.—Hofmeister, Allgem. Morphologie, p. 604. 3 Botan. Zeitg. 1862, p. 274. : * Von Mohl, Botan. Zeitg. 1869, p. 1.—Nordlinger, Der Holzring, Stuttg. 1872.—R. Hartig, in Dankelmann’s Zeitschr. f. Forst- u. Jagdwesen, Bd. III; and Botan. Zeitg. 1870, p. 505.—We may refer to these works for the older, very defective literature, and for many details not strictly be- longing to our present subject. i SECONDARY THICKENING. NORMAL DICOTYLEDONS. 477. of the individual layers has been determined, giving the ‘growth in mass;’ while in others the (radial) dameter, the breadth of ring has been measured, on the successive dimensions of which the form of the stem depends. The growths in mass and breadth of ring in a layer do not necessarily correspond one with another, because the former may be greater in its lower portion, where its periphery is larger, though the. breadth of ring is smaller, than in its upper portion where the ring is broader. The following rules may hold good as giving the consistent result of the published investigations, which extend to the Oak, Beech, Alder, Silver Fir (Abies pectinata, D. C.), Scotch Fir (Pinus sylvestris L.), Red Fir (Abies excelsa Poir.), Larch, Weymouth Pine (Pinus Strobus L.), &c. (1) In the shaft, i.e. in the unbranched stem between the basal ‘stock’ and the crown, the annual growth in mass increases in the case of trees which stand free, from above downwards; according to Nérdlinger the average diameter of the ring always increases simultaneously, while, according to R. Hartig; this may increase or decrease, or remain the same, which agrees with the results of the radial measure- ments made by Mohl on three trees grown in the open, which showed increase in the upward direction. In closely planted trees, the average breadth of the rings increases in the upward direction, according to Nordlinger, Mobl, and R. Hartig, while according to R. Hartig’s statement, which has been-reasonably disputed by Nérdlinger, the growth in surface or mass remains throughout approximately the same. Trees, the crown of which has become s/unfed in consequence of their being closely planted, show a diminution of the secondary growth in this direction, namely, from above downwards, which may extend even to its complete absence in the lower portion. (2) In the crown the growth increases in the downward direction, both in the stem and branches. (3) In the basal s/ock a considerable increase in the secondary growth and the average breadth of ring takes place in the older stems, i. e. in the outer layers; this starts from the upper side of the insertions of the roots, and may extend upwards to a different height (o-3-3 metres and more), according to the particular case. At the points of insertion of vigorous lateral roots, the secondary growth is locally increased in such a manner that the well-known projections of the stock, separated by furrows, and in the case of tropical trees attaining huge dimensions, may arise in the course of a few years. The ‘forms of growth’ mentioned under (1) are subject to change in the same individual, when successively planted free, and in contact with others. For each particular species of our forest trees one or the other form of growth is the tule, and this depends on whether they usually grow in close contiguity throughout life, either wild or in forest culture (e.g. Beech, Silver Fir, Red Fir), or whether in their later years they become free (e.g. Scotch Fir, Larch, Oak, Alder)}. The dependence of the general form of the stem, whether it be more conical or cylindrical, on the conditions mentioned is self-evident. In the same way it is clear, that in the stems of exotic plants which deviate from the cylindrico-conical form, as the spindle- or barrel-shaped stems of Bombaceze’, the progression of the > Compare R. Hartig, Botan. Zeitg. 1870, p 513. * See e.g. the figure of Bombax Munguba in Martius, Fl. Brasil. Tab. physiogn, X, 478 SECONDARY CHANGES. annual sécondary growth from below upwards must be different from that in our trees, in so far as the form is dependent on the thickness of the layers of wood, and not on that of the cortical layers or masses of pith. How far the one or the other explanation is the right one has not been made out in all cases in these plants, The Mamillaric, for example, cited by Mohl as examples of barrel-shaped stems, owe their form, not to the successive increase or decrease in thickness of the layers of wood, but to that of the cortical masses of parenchyma. 2. The forms of Tissue of the secondary wood. ‘Sect. 141. The forms of tvssue of which the secondary wood is composed’ belong chiefly to the categories of ce//s (comp. pp. 5, 115, 121), ¢rachea, and scleren- chymatous elements, especially sclerenchymatous fibres. They sometimes have their characteristic anatomical peculiarities, and the division of labour which these in- dicate, rigorously developed and carried out; sometimes, however, the division of labour is arranged in such a manner that while an element has the essential pecu- liarities and functions of one of these forms of tissue, and must therefore be classified with it,.it further shares in the characteristics of another form. In the tough strong woods of trees and shrubs, which have been the chief subject of investigation, the latter holds good of the elements of a// forms, in so far that they are—in various degrees—thick-walled and sclerotic ; a phenomenon which is not characteristic of the secondary wood generally, but only of that of the ‘ woods’ so called in ordinary phraseology. In the very hard secondary wood of Convolvulus Cneorum, for example, all the elements both of the bundles and of the medullary rays are in the highest degree sclerotic; in the soft, fleshy, chiefly parenchymatous wood of the stem of Carica, Cheirostemon, many succulent roots, &c., only certain individual elements have that character. In the stem of Clematis Vitalba, the parenchyma of the medullary rays takes part in the sclerosis, in that of Atragene it does not, and so on. In many woods one or the other form of tissue may be absent, and its functions be undertaken by others, as will be shown by the examples to be mentioned below. Sect. 142. The ¢vache@ of the secondary wood appear partly in the form of vessels, partly as tracheides*. Of the forms of vessels, as distinguished by the structure of their walls, the reticulately thickened are present exclusively or principally in succulent soft woods, as in the stem of the Papayacez, and in many fleshy roots (Sect. 159). Reticulated vessels with large meshes are further characteristic of the wood of the Crassulacez’, even of the species with hard wood. Reticulated vessels occur, together with pitted ones, in the Caryophyllez, and may often be found in herbaceous Dicotyledons, which have been comparatively little investigated. The wood of the Mamillariz and of species of Echinocactus and Melocactus contains only spiral and annular trachez, and in fact both vessels and tracheides: some have feebler thickening fibres, the 1 Sanio, Ueber die im Winter Starke fiihrenden Zellen des Holzkérpers., Halle (Linnzea), 1858. —Id Botan. Zeitg. 1863, p. 85, &c. The latter also contains detailed citations of the older literature. 2 Compare Chapter IV. * Compare Regnault, Ann. Sci. Nat. 4 sér. tom. XIV. p. 87. SECONDARY THICKENING. NORMAL DICOTYLEDONS. 479 transverse section of which is almost isodiametric, while those of others are ridge- shaped and deeply projecting as-described at p. 156. The former are chiefly vessels, the latter usually tracheides ; but the separation of the two forms is hard to carry out with certainty on account of the difficulty of establishing the presence or absence of the vascular perforations. The species of Opuntia and Cereus have reticulated vessels, which may be accompanied by trachez with the ridge-shaped thickenings above mentioned. The cases described, however, constitute exceptions as compared with the great majority especially of arborescent or shrubby Dicotyledons. In the latter the vessels of the secondary wood are pitted vessels, the wall, apart from the pits, being either smooth, or with a fine spiral fibre on the inner side (comp. Fig. 205). The pits of the vessels are bordered and correspond one with another, at least on those surfaces which border on vessels or tracheides (comp. Sect. 38). On the surfaces bordering on non-equivalent elements, various conditions occur, which will be stated below, chiefly according to Sanio. Where pitted vessels are adjacent to sclerenchymatous fibres with unbordered pits (comp. Sect. 143), the pitting may be wholly absent (Olea europza, Fuchsia globosa according to Sanio) ; in most cases pits are present, though they are always less numerous than those on the surfaces contiguous with trachex, and they also differ from the latter in form. On the surfaces adjoining fibres with unbordered pits, the pits on the wall of the vessel are bordered, but smaller than those on the boundaries of vessels, in Hedera Helix, Euonymus latifolius, europeus, and Syringa vulgaris; they are unbordered in Sambucus nigra, racemosa, Acer platanoides, Salix acutifolia, hippophefolia, Populus pyramidalis, Asculus Hippocastanum, Rhamnus Frangula, Aucuba japonica, Pittosporum Tobira. The surfaces adjacent to parenchyma and fibrous cells have sometimes bordered, sometimes unbordered pits, and sometimes both kinds, The first, for example, is the case in Quercus pedunculata, Diospyros virginiana, Juglans regia, Porlieria hygrometrica, Spartium scoparium, Caragana arborescens, Sophora japonica, Acacia Sophora, Morus alba, Daphne Mezereum, Ribes rubrum, Syringa vulgaris, Casuarina equisetifolia, Hibiscus Rosa sinensis, Peonia Mutan, Ficus Sycomorus, Olea europa, Nerium Oleander, Tamarix gallica, Punica Granatum, Justicia carnea; these pits are unbordered in Hedera Helix, Sambucus racemosa, nigra, A’sculus Hippocastanum, Rhamnus Frangula, Syringa Josikxa, Solanum Dulcamara, Populus pyramidalis, Salix hippopheefolia, acutifolia, Vitis vinifera, Magnolia tripetala, acuminata, Hydrangea hortensis;—both kinds occur in Bombax Ceiba, Ficus rubiginosa, Jatropha Manihot, Fuchsia globosa and Eugenia australis. As a rule the pits bordering on cells are relatively large ; they are rarely small and very numerous (Hydrangea hortensis); where two kinds occur the bordered often differ ‘from the unbordered also in their superficial outline. The pits of the vessel always correspond with the unbordered ones of the non-equivalent elements, and in fact those of the latter are always of the same breadth as the border of the pits on the vessel. As an exception we must mention the wall destitute of pits with which the vessels border on the fibrous cells in Punica Granatum. Where the lateral walls of the members of a vessel have a spirally thickened inner layer in addition to the pits, the spirals are always present on the surfaces adjacent to trachez, with one exception to be mentioned at the end; in certain cases they also exist on the surfaces adjacent to all the other, non-equivalent elements (Tilia parviflora, Pitto- sporum Tobira, Prunus domestica, Laurocerasus). In other cases the spirals are absent on the surfaces which border on parenchyma, while they are present on the others (Amygdalus communis and other Amygdalezx); or they are absent where vessels border on one another and on parenchyma, and are only present on the surfaces which are adjacent to fibres. 480 SECONDARY CHANGES. The different forms of perforation of the transverse wall vary according to the particular cases (species). The vessels are relatively thin-walled in most true woods, even in very hard woods, and are often remarkably delicate (Camellia japonica). More rarely the thickness of their walls is equal to that of the thickest-walled of the elements which accompany them, e.g. Fraxinus excelsior, Ornus, Nerium Oleander, Piperacez, Convolvulus Cneorum, The Zracheides (Fig. 20g) either constitute the only tracheal elements of the wood (Coniferze, Wintereze), or they occur together with other tracheal organs, especially vessels. They are characterised as tracheides by the properties NG indicated in Chap.1V. The structure of their wall is AY in the case of the Coniferze and Winterez that of pitted vessels with bordered pits; in most Taxineze there are also spirals on the inner side of the wall. In the second case the same holds good, with the addition that they then resemble the members of the narrower vessels of the same wood, either in all points, except the vascular perforation, or at least in possessing similar bordered pits to those of the vessels belonging to the same wood. As regards the spirally or an- nularly thickened inner layer, they usually agree with the vessels accompanying them, but not always: in Pyrus communis, Sorbus Aucuparia, and Staphylea pinnata the vessels have spirals, the tracheides not, in Philadelphus coronarius the converse is the case. In particular tracheides of many plants, we find, as an exception, isolated thickenings of the wall, projecting inwards in the form of blunt cylindrical pegs, or of bars running transversely from one side to the other. MONI NON FIG. 205.—Cytisus Laburnum; tangential section through the same autumn wood as Fig. 198, p. 465. 5 intermediate cells, »—2 medullary ray; in the second cell from the top is a peg-shaped thickening of the wall. The medullary ray is bounded on the left by an unperforated tracheide, on the right by a narrow vessel; at gis the perfcration of the transverse wall of the latter. The lower longitudinal wall of the tracheide to the left was preserved in the preparation; the upper one, all but a very small piece, was removed in cutting the section; in the other tracheides and vessels the longitudinal wall facing up- wards is drawn, the spiral fibres being shown in reversed direction, and the really.bordered pits with only a single outline. Both forms were observed by Sanio in Hippophe rhamnoides, the latter in Pinus silvestris', and casually by me also in Drimys Winteri. The transverse bars lie in the radial direction, at least in Pinus and Wintera, and are continued as a single bar through many elements of a radial row. Peculiar transverse ‘lines, which Sanio found on macerated tracheides of Casuarina (/.¢. p. 117), still require explanation. Where tracheides occur together with vessels or sclerenchymatous fibres, or both, two extreme cases may be distinguished as regards their form, namely, on the one hand, those which on the whole resemble the members of the smaller vessels in length and width, and abut on one another with a relatively slight inclination of their terminal surfaces; on the other hand, more elongated, ‘ fibre-like’ forms, with long acuminate ends, sometimes even forked (Hippophe, Casuarina torulosa, Staphylea pinnata), which penetrate between one another and between the non-equivalent elements: 1 Compare above, p. 164. SECONDARP THJCKENING. NORMAL DICOTYLEDONS. 481 . The average length, expressed in hundredths of a millimetre, is, for example, accord- ing to Sanio’s determination, in Tracheides most Fihriform similar to vessels. Tracheides, Fagus silvatica. . . 39 2. 6. 1 ee ew ee 75 Cunonia capensis . . 69 . . . . 1... 97 Casuarina torulosa . 45... 1. ues 104 3 equisetifolia 48 . . . 2... 1. 78 Hamamelis virginica. 70 . bow a e286 Sheperdia canadensis 19 . 2 + 45 The first category includes those tracheides which also approach the vessels most nearly in the structure of the wall, while those belonging to the second are in this respect also Jess similar to the vessels, and approach the fibres in every point. If we survey the entire series of the woods investigated, there is on ‘both sides, as well as between the two principal cases just distinguished, a continuous transition between the extremes '- The average thickness of the walls and the nature of the thickening layers * generally vary in accordance with the other differences and similarities, if the excep- tional cases of very thick-walled vessels already mentioned be left out of account. The peculiar tubes, containing air, which form the floating apparatus of many Leguminous woods (Herminiera, &c.) must here be mentioned by way of supplement. In order to avoid repetitions, the description of them follows below, in Sect. 150. Sect. 143. The sclerenchymatous fibres of the wood, shortly termed woody fibres (Fig. 206), are generally distinguished from those elongated tracheides which are more or less similar to them in form, by the structure of their wall. The latter is always destitute of the spirally-thickened innermost layer—although the whole wall may be striated and capable of splitting in a spiral direction,—and its pits, which are invariably slit-shaped, and lie in a left-handed oblique direction, are always present in relatively small numbers, often very sparingly, and differ in details from those of the accompanying vessels. While in many cases they are bordered here also (Quercus, Daphne, Liriodendron, Fraxinus, &c.), they are usually not bordered (Sambucus, Hedera, Clematis Vitalba, Syringa vulgaris, Ligustrum vulgare, Euony- mus latifolius, Celastrus scandens), or so small that the presence of the border is difficult to determine. Both kinds of pits are mentioned by Sanio as occurring in Jatropha Manihot, those with a border being the more numerous. In this plant Sanio found two kinds of pits on the surfaces of junction between the fibres and the medullary rays, namely, small slit-shaped ones on the radial lateral surfaces of the cells of the medullary ray, but large round ones on their horizontal corners, towards which a small pit runs from each adjoining cell of the ray. -In the other cases investigated, the pits of the fibres are of approximately equal size, on whatever form of tissue they may border. The walls of the fibres are thickened, in a manner which accords generally with the fundamental rules holding good for all cell-walls; their usually comparatively thick middle layer® is as a rule homogeneous, at least without any conspicuously ' Compare Sanio, /.¢. pp. 117, 118. 2 See Hofmeister, PAlanzenzelle. p. 196. : 5 Hofmeister, Z.c. 1i 482 SECONDARY CHANGES, apparent finer stratification arid striation’; its thickness, however, is very unequal in different species and individuals. The walls are as a rule lignified. A not un- common exception, however, occurs, inasmuch as one of the layers is of a manifestly soft, cartilaginous, gelatinous consistency, and is then excluded from the process of lignification, becoming violet immediately on treatment with preparations of iodine. This gelatinous layer (comp. p. 133) is as a rule the innermost one; it surrounds the lumen immediately, either as a narrow border (Jatropha Manihot, Morus alba), or usually as a ‘thick, ap- parently swollen mass, filling the greater part of the lumen. In rare cases a layer enclosed between the lignified ones shows the characteristics in question; these-often extend to the whole of the wall lying inside- the - outermost limiting layer’(‘ primary membrane’). Lastly, the gelatinous layer may often: be distinguished by its refraction, even when it is stained like a lignified membrane, by preparations of iodine. - The occurrence of the gelatinous layer is strik- ingly irregular.’ Sanio found it especially among Leguminose (Cytisus Laburnum, Sarothamnus, So- phora japonica, Caragana’ arborescens, Gleditschia triacanthos), where it is of quite usual occurrence ; S N also in Ulmus suberosa, Celtis australis, Hakea VN ia, Ailz LINN) suaveolens, Morus alba, Broussonetia, Ailantus, SINE Fuchsia globosa, Eugenia australis; Castanea, Dios- S N E ‘pyros virginiana, Corylus Avellana, Ostrya vir- NINE ginica, Populus pyramidalis, Betula alba, Alnus glu- N tinosa, Enckea media, Eucalyptus cordata, Calycan- VA\ thus floridus, Amygdalus communis, Prunus Lauro- S ‘cerasus, Jatropha Manihot, and Ficus Sycomorus, and aA he supposes that it occurs much more generally, It Z N “ is, however, by no means generally characteristic of EAN all the fibres of these woods, as it may be present or BAIN absent in different parts even of the same annual. BR ring; while its occurrence is often rare, and even so =| S isolated (Betula, Alnus) that one may repeatedly investigate a wood without finding it. Nor is its presence or absence connected with any definite special form of structure in other respects, or with the average thickness of the wall, It cannot there- fore be regarded as a characteristic peculiarity of the fibres, the less so, as in particular cases (Hamamelis, Fagus silvatica, Casuarina) it also occurs in elements which, according to their other properties, belong to the class of tracheides which resemble vessels. What was stated above in the case of the tracheides most.closely resembling fibres, applies. YL UM i : = Fig, 206. . Fig. 207 FIG. 206.—Cytisus Laburnum ; three-year-old branch during the winter’s rest-(March). Tangentiak section (145). a@5c¢d@ the zone of secondary growth and cambium bordering on the autumn wood, 4 of the previous year, and containing a medullary ray . above, “FIG. 207.—Cytisus Laburnum, Outlines of a selected short woady fibre, from the youngest annual ring of the same branch as that from which Fig. 206 istaken (x45). 0 s - generally to the fibres themselves as regards form and size. Sanio (2c. 106) adduces many examples of the occasional bifurcation of their acute ends. On the average their length exceeds that of the neighbouring tracheides the more, the more the latter * Compare Sanio, /.c. p. 105, ® For details see Sanio, Z.¢. p. 103. SECONDARY THICKENING. NORMAL DICOTYLEDONS. 483 resemble vessels, comp. Figs. 206 and 207; in extreme cases it may be approxi- mately equal to that of the tracheides (Syringa), or even somewhat less (Ribes). . The following mean values, as determined by Sanio in hundredths of a millimetre, may -serve to demonstrate the relations of length :— Tracheides. Fibres. Sophora japonica . . . . . 16. 1 1 we 95 Sarothamnus scoparius . .. 417... . . 56 Ulex europeus . . . . .. 16... . . 103 Celtis australis . 2. . . . . 26... 4, . 87 Cordia pallida. . . . 1... 270.0. 1 dE Rhamnus cathartica 2... 28 . 2. wee 52 , Esculus Hippocastanum . . . 26... . . 43 Tilia parviflora... 1. . . 3r 6. 1 we we 46 Salix acutifolia . . 2. 1. 1 330. 2 wwe 53 Rhustyphina . . . 1. 1 32 4.2.4.0. 04 «35 ; Rhamnus Frangula. . « . . 24 «© « «es 44 Quercus pedunculata. . . . 49... « . 80 Prunus Laurocerasus . . . . 56... . . 126 Populus pyramidalis . . . . 39 2. «1 ee 45 Hakea suaveolens . . . .. 26... .°. 81 : Eucalyptus cordata. . . . . 34 4... + « 60 Periploca greca. . 2... . 28. . «ws 36 Daphne Mezereum. . . . . 15... . + 21 Spirea chamedryfolia. . . . 33>... =. + 38 Syringa vulgaris. . . . . 2 50... 2. 81 -Ribesrubrum . ... . 6. 49... es 47 In the contents of the woody’ fibres shrivelled remains of protoplasm and of formed constituents of the contents can only be detected in rare cases where the wall is very thick and the lumen very narrow, as in the tough fibres of the wood of Viscum, and perhaps also of Leguminosz, Quercus, &c. Further attention must, however, be directed to this point, which is difficult to make quite clear, on account of the scantiness of the ‘remaining contents and the thickness of the wall. Even in the cases just mentioned air is certainly present, in addition to the remnants of the contents. In most woody fibres, however, the lumen contains nothing but air and water. It is manifest that they agree in this point with the tracheides, nor does it admit of doubt that in so far as this is the case they take part in the functions of the Jatter, and thus that we here have a case of the above-mentioned phenomena of incomplete division of labour, The sharp severance of the two organs cannot therefore be carried out without violence and uncertainty, especially as the characters assigned to them,.and in particular the bordering of the pits, are on the one hand variable in different cases, and‘on the other are difficult to determine in practice, in the case’ of very small pits. It will therefore constantly be necessary to speak of tracheides resembling fibres, and of fibres resembling tracheides.. On the other hand, however, many cases of sharp differentiation exist, as in the Leguminose mentioned, Quereus, and many others; these cases render the distinction necessary, and by taking these as the starting-point, it can be carried out even in the less clear cases. | 'Szcr, 144. The cells of the secondary wood may be divided according to their form, into fibrous cells, and short parenchymatous cells, : : 1. The fibrous cells resemble the woody fibres more or ‘less closely in Poy riz 484 "SECONDARY CHANGES. “ Like the latter they proceed from the longitudinal division of an elongated spindle- shaped tissue-mother-cell of the cambial zone, without any transverse divisions. In the thick-walled forms, a subdivision of the lumen into compartments may subse- quently occur, by means of thin transverse walls, as is the case in the chambered sclerenchymatous fibres (p. 134): these are sepiate fibrous cells. The elements in question, as follows from what has been stated, are products of the cambium, in which the cellular qualities persist permanently, or disappear slowly. In their further characteristics they are closely related to the other elements of the wood, namely, on the one hand, to the woody fibres, and on the other to the short- celled parenchyma. Of the two subordinate forms which result from these relations, - the former may be termed fibrous cells in the strict sense, while the latter may bear the name of znfermediate cells! (Ersatzfaserzellen) given by Sanio. a. The former agree in their form, and in the structure of their walls, with the woody fibres, and thus certainly take part in the functions of the latter, into which they may gradually pass over completely. They are distinguished from them by the nature of their contents. The latter almost always contain starch (comp. p. 115): in Spirza salicifolia Sanio found traces of chlorophyll; this appears more abundantly in the septate fibrous cells of the one-year-old wood of Vitis vinifera and Centradenia grandifolia. Tannin is contained in the fibrous cells’of Vitis, and traces of it in that of Syringa vulgaris, while in the other woods investigated it is absent in these elements, even where it occurs in other cells. In the wood of Punica Granatum, all the elements of which, with the exception of the vessels, are filled with starch*, the grains contained in the fibrous cells are- on the average considerably larger than those of the other cells. Besides the plants already mentioned fibrous cells containing starch occur in the wood of Berberis vulgaris, Mahonia Aquifolium, Begonia muricata, angularis, Sambucus nigra, racemosa, Cheiranthus Cheiri, Salix cinerea (root), Ligustrum vulgare, Syringa vulgaris, Clematis Vitalba, species of Acer, Vitis vinifera, Celastrus scandens, Euonymus . europzus, latifolius, Hedera Helix, Acacia floribunda, Robinia pseudacacia, Ficus elastica, rubiginosa, Sycomorus, Rhus Toxicodendron, Tamarix gallica ; fibrous cells with slightly granular contents occur in Ephedra. Septate fibrous cells occur in Coleus Macraei, Hydrangea hortensis, Fuchsia globosa, Aucuba japonica, Celastrus scandens, Euonymus latifolius, europzus, Spirza salicifolia, chamedryfolia, Ceratonia, Hedera Helix, Pittosporum Tobira, Eugenia australis, Rubus idzus, Justicia carnea, Ficus Sycomorus, rubiginosa, elastica, Bignonia capreolata, Tecton3 grandis, Rhus Cotinus and Toxicodendron, besides the plants already mentioned; either the non-septate predominate (e.g. Spirea salicifolia), or the septate, e.g. Vitis, Hedera, and Punica. Starch has always been found in the septate cells, although in small quantities, except in Punica and Ceratonia, In the case of Justicia earnea only, Sanio states that he found the cells containing air, ‘no doubt abnormally.’ _b, Sanio’s zntermediate cells (Ersatzfaserzellen) (Fig. 205, p. 480), agree with the short-celled parenchyma of the ligneous bundles which will be next described, in all : 1 [In justification of the use of the term ‘intermediate cells’ in place of a more strict translation of the term ‘Ersatzfaserzellen’ introduced by Sanio, it may be pointed out that, though the inter- ‘mediate cells do replace the short-celled parenchyma in some few cases, that is not their constant character. It is thought that the use of the term ‘intermediate cells’ will avoid this difficulty, while it brings the real character of the cells more prominently forward, viz. that they are intermediate in form between fibrous and short parenchymatous cells; compare Sach’s Textbook, 2nd Engl, Ed. p. 950] ? A, Braun, /.¢.; compare p. 471. SECONDARY THICKENING. NORMAL DICOTYLEDONS. 485 properties, with the exception of their shape, and the form of their pits, which in many species is not rounded but slit-like. Sanio’s term owes its origin to the fact that they not only frequently accompany the short-celled parenchyma of the bundle, but in many woods exclusively represent or replace it, as in Viscum, Porlieria, Cara~- gana arborescens, andi Spirzea salicifolia. 2. The short parenchymatous cells of the wood present some differences accord- ing as they belong to the ligneous bundles or to the medullary rays. Thus the parenchyma of the bundles, usually called ‘wood parenchyma, is to be distinguished from the parenchyma of the rays. The parenchyma: of the bundles has ‘been principally investigated in the case of hard woods, and it is to these that most of the following statements apply. They may, however, with some obvious modifications, be also extended to the soft, fleshy, suc- culent, or at least loose woods, which consist chiefly of parenchyma, as in the stems of Papayaceze, Bombax, and Chirostemon, and in many roots. The cells of the typical parenchyma of the bundles arise by predominant trans- verse divisions of the elongated tissue mother-cells in the cambial zone. Accordingly they are arranged in simple, or in some places multiple, longitudinal rows, each of which has a somewhat spindle-shaped form, resembling that of the mother-cell. This grouping comes out most conspicuously when they lie isolated in non-equivalent tissue, less so where they are united to form larger masses. The length of the spindle-shaped groups is usually less than that of the fibrous cells, more rarely, as in Vitis, it is on the average equal to it. The form of the individual cells is that of a more or less elongated prism, with horizontal or oblique terminal surfaces ; it is obvious that those which form the ends of a group must further show a conical tapering. Those which border on wide vessels are often flattened in the direction of the circumference of the vessel, and elongated transversely, in consequence of the expansion of the members of the vessel at the expense of their surroundings (p. 470). In many cases the cells which surround a group of contiguous vessels on oppo- site sides are connected by means of flatly tubular outgrowths of their lateral walls, which penetrate between two vessels, and fit on to one another at their ends. The ' outgrowths are frequently branched, frequently they have blunt ends without fitting on to others. This phenomenon is explained by Sanio', and no doubt correctly, by the unequal growth and partial displacement of single rows of parenchymatous cells originally lying between the rudiments of the vessels. It occurs in Casuarina, Mela- leuca imbricata, Platanus occidentalis, Celtis australis, Ficus Sycomorus, Cordia pallida, and especially in Tectona grandis and Avicennia spec.; and it has also been found by Sanio in the intermediate cells of Porlieria. The wall of the parenchymatous cells of the bundles is in the harder woods always provided with roundish or elliptical pits which are not bordered; they are never slit- shaped, or arranged in regular oblique rows; on the sides which are in contact with vessels they are usually larger than on the others, though exceptions to this rule occur (Betula alba). The pitting goes all round, even over the transverse walls, and the latter are of equal thickness with the lateral, or the thinner lateral walls—a point of 1 Zc. p. 94, where the further details are to be compared. 486 SECONDARY CHANGES, t distinction from the septate fibrous cells. ‘The walls, though lignified in the mature condition, are uniformly distinguished from those of the tracheze and fibres of the. same ‘wood by their lesser thickness. Exceptions to this rule are very rare: Magnolia acu- minata and tripetala, Liriodendron tulipifera, Gymnocladus canadensis, and Amorpha fruticosa, in which the radial walls of those woody parenchymatous cells which lie in the autumn wood are not inconsiderably thickened. Spiral or’ annular fibres are always absent. The parenchyma of the bundles in soft fleshy woods is in general only distinguished from that above -described, by its usually larger cells, and less thickened walls. The-nature of the cell-contents is in-general characterised by the term paren- chyma. In most hard woods the starch-grains, which are stored up periodically, during the winter’s rest, form the chief contents; chlorophyll and tannin occur here and there ; -the former, for example, in the wood of Cobzea scandens. The parenchyma of the medullary rays consists, in the great majority of secondaty woods, of cells-which have essentially the same properties as the parenchyma of the bundles in the same plant, without being exactly similar in every point. In woods -which are not fleshy and succulent the walls of the, cells of the medullary rays are as a rule lignified, like those of the woody parenchyma. Exceptions to this rule occur in many twining and climbing plants, in which the cells of the medullary rays remain unlignified, delicate, and capable of yielding to pressure and tension, e.g. Menisper- mum canadense, ‘Aristolochiz, Atragene alpina. ‘The lignified medullary rays of Clematis Vitalba, which agrees so closely in every respect with Atragene, show how- ever that this phenomenon is by no means generally characteristic of plants with the habit mentioned. The form of the cells of the medullary rays is usually that of a rectangular prism, often with rounded corners, and roughly comparable to a brick; in thin medullary rays, filling up a narrow mesh between the ligneous bundles, the: cells which occupy the angles of the mesh have a corresponding wedge-like form. Usually the cells are chiefly elongated in one direction, and are either procumbent with their greatest diameter directed horizontally and radially; or wprdghé, with. their greatest diameter . vertical. The former is by far the most frequent case. Cells standing vertically occur, for example, in Asclepiadez (Periploca, Hoya, Asclepias curassavica), Nerium, Drimys Winteri, and Medinilla farinosa, In the medullary rays of Camellia japonica pro- cumbent and vertical cells occur in groups. Medullary rays-with procumbent cells ate always easy to distinguish from parenchyma of the bundles, even where they traverse the latter, because the longitudinal diameters of the two kinds of cells cross one another ; in the case of the upright cells this distinction is often less simple on account of the similar direction of the longitudinal diameters. — But few minute investigations on the structure of the cells of the medullary rays exist, and many details are still to be discovered. From what is already known, it- may however be asserted that the cells of a medullary ray are as a rule similar to one . another, apart from irrelevant differences, some of which follow directly from what has _ been said. But few exceptions to this rule are known. In the medullary rays of Aristo- Tochia Sipho, Sanjo? found smaller cells, containing finely granular starch, arranged | in Det pe loys : SECONDARY THICKENING. NORMAL DICOTYLEDONS, 487 irregular reticulate rows, between larger, empty, dried-up cells, a condition which recalls the pith of the Roses, &c. In the medullary rays of Atragene alpina annular zones of two kinds alternate from within outwards, the one consisting of a few rows of relatively narrow, closely connected cells, the other of somewhat larger cells connected to form an irregular, coarsely lacunar tissue, but otherwise similar in structure to the narrow cells, This structure owes its origin to the fact, that with each thickening of the ligneous bundles the medullary ray receives an increment of growth, which remains on the whole smaller in the radial direction than that of the ligneous bundles as regards the number and size of its cells. Certain cells or groups of cells either follow the general growth, or are dragged apart in a purely mechanical manner in consequence of it, so as to form the lacunar zones. In general, though not exactly, each pair of dissimilar zones corresponds to an annual ring. In the wood also the cells of the parenchyma are no doubt always accompanied. at least by narrow intercellular interstices containing air. In particular cases, to: be mentioned below, they surround wide passages containing secretions, : Scr. 145. Forms of tissue other than those discussed in the preceding para- ‘graphs are in most woods either absent, or, at any rate, of small importance. The most widely distributed are the sacs containing crystals, which, wherever they occur, accompany the parenchyma of the bundles or of the medullary rays; e.g. Legumi- nose, as Pterocarpus santalinus, Hematoxylon *, , Herminiera (pp. 139 and 141), Vitis, &c. s Laticiferous tubes are abundantly developed in the wood of the Papayaceze, which is chiefly parenchymatous. Their reticulately connected branches are here distributed between the elements of the parenchyma, and are in contact with: the vessels, In other plants containing latex, only those branches which run from the -cortex into the pith pass through the secondary wood. No doubt they always exist earlier than the latter, from the primary differentiation of tissues onwards, and sub- ‘sequently become enclosed, and also of course stretched, by the secondary growth. Where one of these connecting branches borders on woody fibres, the latter are not uncommonly bent inwards so as to follow its course, as is indicated in Fig. go, p. 438, ‘by the oblique shading. Comp. also Chaps, VI and XII. The classification of the eleménts of the secondary wood given in Sects. 141-145 is based on that which Sanio, in agreement with T. Hartig?, has given in his fundamental works cited above, p. 478. It differs, however, from the latter in some points, Apart from the medullary rays, Sanio classifies the elements of the ligneous bundles as follows :— I. Parenchymatous System. 1, Woody parenchyma. z, Fibres representative of the woody parenchyma. II. Libriform System. 3. Simple undivided bast-like wood- ests or wood-fibres : Libry ‘orm tissue. . } 4. Septate Libriform tissue. ” III. Tracheal System. 5. Tracheides. 6. Vessels. His System II includes both our woody fibres and our fibrous cells, the two are placed together on account of their form and the structure of their wale while no | primary 1 Fliickiger and Hanbury, Pharmocographia, pp. 176, 188, Compare especially Botan. Zeitg. 1859, p. 92. 488 SECONDARY CHANGES. importance is attached to their contents. The intermediate fibres are separated from the fibrous cells, also on account of the structure of their walls, and placed, with the parenchyma of the bundles, in Category I. The rest of the classification resembles our own. Inso far as nothing further is aimed at than an intelligible arrangement of the forms of tissue of the secondary wood, Sanio’s classification is without doubt as intelli- gible as our own, and perhaps more so. Both also suffer from the same defect, namely, that the categories distinguished cannot always be sharply separated, and in particular that intermediate forms occur between fibres and tracheides, &c., as has often been stated above. Both, however, afford guidance in each particular case, in accordance with the system adopted. There would therefore be no reason for undertaking alterations in Sanio’s arrangement, if it were not an essential object to refer the forms of tissue in the secondary wood to their proper place among those distinguished in the plant generally, including those outside the secondary wood. It can admit of no doubt that the elements of the secondary wood are not organs sui generis, but belong to the forms of tissue which have been characterised in this book as trachez, sclerenchymatous fibres, and cells; the last being distinguished from the others by permanently containing protoplasm, or in doubtful cases by their periodically varying store of starch (comp. pp. 5 and 115). The known phenomena of the secondary wood present, as I believe, no argument against the general classification of the forms of tissue carried out in this book, for the occurrence of intermediate phenomena cannot avail as an argument against the distinction of typical forms, In the face of these facts it was, however, necessary to depart in some points from Sanio’s classification, which was based on other considerations. I willingly grant that the sharp separation of the cells from the other elements is often inconvenient for the practical description or identification of woods, as it is not always an easy matter to establish the cellular quality. In most cases indeed the presence of starch is a certain character, both in the fresh and in the dry wood. In its absence, however, the distinction is in the latter often impossible or only possible with great diffi- culty. In the wood of Cobza, for example, which has been preserved dry, it can scarcely be determined with certainty whether the numerous reticulated elements with large pits are short tracheides or parenchymatous cells, for in the structure of their walls they resemble reticulated trachez, which might occur, and all the formed contents have become indistinguishable. In the fresh plant, on the other hand, the presence of proto- plasm and with it the cellular quality may be recognised, at least up to the third year, by means of the large chlorophyll grains. Experiences of this kind are, as we have said, inconvenient, though, on the other hand, they are certainly instructive, as they point to the necessity of investigating woods in the fresh living condition, more than is usually done. They would, however, only present a serious objection to the classification arrived at, if the problem of the anatomy of wood were to be sought in the construction of a convenient ‘key’ for description and identification. 3- Distribution of the tissues in the wood. Sxct. 146. The distribution of tissues in the wood, and the consequent structure of the latter, is uniform in the successive annual rings, with the exception of certain differences to be specially treated of below; the single annual ring may therefore be considered first. In the few cases without annual rings the description applies to the entire woody ring. As has already been often mentioned, alternate radial bands of unlike structure are nearly always apparent at the first glance: (1) the medullary rays, and (2) the ligneous bundles. This fact determines the main subdivisions of the exposition. Where this alternation of unlike radial bands does not exist, as in the cases men- SECONDARY THICKENING. NORMAL DICOTYLEDONS, 489 ‘tioned on'p. 458, and in many fleshy roots to be mentioned below, the entire mass of secondary wood is to be regarded as a single cylindrical ligneous bundle. a. Medullary rays and medullary spots. Sect. 147. In the woods which have been principally investigated, every zone possesses a large number of medullary rays of various rank. Since new secondary ones always arise as the ligneous mass grows in thickness, their number increases with that of the successive layers. And further, as in the successive layers their size either remains approximately uniform, or at least increases in a much lower degree than their number, it may well be assumed that the proportion between the space occupied by them, and that occupied by the ligneous strands, remains approximately the same in all the successive layers. This assumption agrees with the observation that where the increase in number of the medullary rays is very slight, an especially striking dilatation of the original ones occurs ; e.g. Atragene. No detailed researches on this relation exist. The number of the medullary rays in the surface of cross-section is apparently about in inverse proportion to their size, i.e. to their breadth, and no doubt their height also, Nérdlinger ‘ undertook a great number of enumerations, which, although according to his own judgment they are not exactly reliable for woods with numerous medullary rays, on account of the serious difficulties, and are also taken from casually selected annual rings, yet give definite proportional numbers for the very different conditions obtaining in the particular species. He gives, for example, in a breadth of 5 millim., for Aristolochia Sipho 9, Clematis Vitalba 10, Cytisus Laburnum 19, Robinia pseudacacia 20, Acer pseudoplatanus 33, Abies pectinata 37, Abies excelsa 44, Acer platanoides 47, Acer saccharinum 53, Quercus pedunculata 64, Alnus glutinosa 78, Aisculus rubicunda 84, Euonymus europzus, Punica-Granatum 10, Rhododendron maximum 140 (the highest: figure determined). The average breadth of a medullary ray varies according to the species, by Nérdlinger’s measure- ments (/.¢.), from 1™™ (Quercus Cerris) to o’o1g ™™ (Asculus, Buxus, Castanea, Euonymus europzus, Hamamelis, Juniperus communis, virginiana, Kcelreuteria, ‘Ligustrum vulgare, &c.). It amounts to about 0-025 mm, according to the same author, e. g. in Abies pectinata, Pinus, Larix, Taxus haceats, Syringa vulgaris, &c., to about 0-05 ™™ in Acer pseudoplatanus, dasycarpum, Juglans, Robinia psstid: acacia, Sambucus nigra, &c., to about o-1™™ in Ailantus, Alnus incana, Cytisus Laburnum, Gleditschia, Platanus acerifolia, &c. Whether the measurements were made on sections of sufficient thinness for the determination of absolutely exact numbers may remain undecided. The Aezghi of the medullary rays is not less variable in different species than their breadth, but has received much less attention in the published investigations, In woods without intermediate bundles, and in Clematis with one intermediate bundle to each primary one, the height of the primary rays is equal to that of the internodes, and thus amounts to 100 or 200mm, In the smallest secondary rays of the Abietineze, which are only 1~2 cells high, it scarcely exceeds 0-025 mm, ; * Querschnitte von Holzarten, Band 2, p. 5. Ago SECONDARY CHANGES. - , These relative magnitudes may be better determined according to. the number of cells, or layers of cells, which compose the medullary ray in breadth and height, than according to absolute measurements. Those under 0-025 ™™ in breadth are no doubt all of them only one cell broad (at most two or a few in the middle), and are thus ‘ uniseriate’ as seen in tangential or transverse section, as, for example, in almost all Coniferze, and. generally in the. narrowest of the above examples; while in like manner the broader ones are always pluri- or multiseriate. Similar rules obviously ‘hold good for the relative heights, and here also the number of cells varies, according to the particular case, from very high figures, down to one or two. Certain woods possess medullary rays of /wo different sizes, with which differences of the minute structure are also usually connected; e.g. those of the Abietinez, to be described below, which differ in the presence or absence of a resin-canal; the broad high multiseriate rays, and numerous low uniseriate ones between them in Quercus and Fagus; small secondary rays, ‘which are triseriate in the middle, between the much larger primary ones, in Casuarina’, &c. The medullary rays are sharply defined, and exactly fill the meshes between the ligneous bundles, which run in a curved course around them. An exception to this rule is described by Schacht” in the case of the wood of the root in Araucaria brasiliensis, in which the uniseriate rays, consisting of irregularly undulated cells, are ‘united by rows of similar cells, which run between the tracheides of the ligneous ‘bundles vertically, from one ray to another, lying above or below it. A further exception is formed by the medullary spots to be described below. In the great majority of cases the medullary rays consist only of parenchyma. In many woods they collectively constitute the main mass of the parenchymatous tissue which is everywhere distributed between the other elements: in many cases (Winterez) this is represented by them alone, in others (Coniferze) at least to much the greatest extent, The succulent parenchyma, which forms the greater part of the wood of the stem in Carica and Vasconcella, is principally formed by the ee, broad, and high medullary rays. Exceptions to this purely parenchymatous structure rarely occur: As such are to be mentioned in the first instance the medullary rays of many Abichinee,. all investigated species of Pinus in the narrower sense, Cedrus, Larix, Tsuga canadensis, ‘Abies excelsa, and balsamea—and of Sciadopitys, which consist of /wo finds of elements, namely, parenchymatous cells, and /rachezdes of similar form to the latter, distinguished by Hartig ® as ‘ fibres,’ Among the Abietinex, the Pines (Pinus), Firs (Picea excelsa), Larix and Pseudotsuga, have two kinds of medullary rays: larger ones, which contain, in their many-layered central portion, a resin-canal which runs horizontally into the bast, and is not in communi* . gation with other canals of the wood and bast ‘, and smaller single-layered rays, usually only a few (1-12) cells or elements in height, and destitute of a resin- -canal, The other * trees mentioned have medullary rays of only one kind, and of the structure last mentioned; * Compare Géppert, Linnea, Bd, XV. p. 747+ —Low, Diss. de Casuarinearum . «.. Structura, Berl. 1865. ? Botan, Zeitg. 1862, p. 412, Taf. XIII. 15. * Forstl. Culturpfl. p.13, Taf. V. Compare also his Jahresber. (1837), p. 145. * Hartig, Naturgesch, d. forstl, Culturpfl. p. 95, Taf. 5—Von Mohl, Botan, Zeitg. 1859, p. 334+ SECONDARY THICKENING. NORMAL DICOTYLEDONS. 491 they seldom exceed the height and breadth stated;—in Cedrus they are as much as 50 cells in height, and often more than one cell bread in the middle. In the species first mentioned above, the medullary ray consists firstly of somewhat elongated, pris- matic, procumbent cells, which, on their surfaces of contact with one another and with , the tracheides of the ligneous bundles, have, according to the species, one or more large, unbordered pits; in the latter case they are really not pitted, in so far as the thickening and the pit, strictly speaking, only belongs to the tracheide+, , Secondly, tracheides occur in addition to these cells of the medullary ray, which they re- semble in form and position. Their walls, where they border on equivalent elements, and on tracheides of the bundles, have bordered pits of smaller size than those of the latter; in many species of Pinus (e, g. P. silvestris, and Laricio), and’ Sciadopitys, they farther have irregular thickening ridges, projecting inwards like teeth, on their upper and lower sides; towards the cells of the medullary ray they only have extremely scattered small pits, which, so far as I could see in P. silvestris, are -unbordered, Each of the radial rows of which the medullary ray consists, is, as far as investigation extends, composed _ exclusively of one of these two kinds of elements, and in fact in a ray more than two ‘elements in height, the upper and lower wedge-shaped edgés always consist of one to three series of tracheides. In the middle of the medullary ray there then lie either rows of cells only, or rows of tracheides alternating with the latter. E. g. they ‘occur. in the following order, succeeding one another from above downwards (or conversély), the Roman numerals indicating the rows of tracheides, the others the rows of cells, and the letters (a), (6), &c. the particular medullary rays investigated, | Pinus silvestris, wood of the stem: (a) II, 4, I, 1, II. (4) I, 2,1, 3, L. () 1, 3, IV, 3, IL. (2) 1,2,1. () Wy4, I, &e. nie europea, stem: (a) 1, 1, 11, 6,1. (4) I, 1,1V, 9,1. (c)' I, 14, I, &e. : . Small medullary rays, only two Plements in height, are in P. silvestris often composed . of tracheides only?. : The. second exception to the usual, purely parenchymatous structure, occurs in several plants which form but little wood, and consists in the fact that the medullary ray is not formed of parenchyma, but of elongated, sclerotic fibrous cells. It has been primarily observed in the perennial stems of the suffrutescent Begoniz *, e. g: B. angularis, muricata, Hiigelii. The very large and broad medullary rays of the secondary. wood here consist of upright, very much elongated cells, which abut on one another with oblique, sometimes acute, sometimes blunt, terminal surfaces, very like the cambial cells of the hgneous bundles, and acquire lignified, sclerotic walls of considerable thickness, with small pits. The cells have scanty contents, and fre- quently even contain starch. The broad medullary rays form collectively a tough ting into which the relatively. narrow ligneous bundles are fitted. ‘A similar structure occurs in many herbaceous stems of Umbelliferz, as in Cherophyllum, Myrrhis, Seseli, Daucus, and Eryngium +, though here more minute in- vestigation is needed, with reference to some doubtful points suggested by Jochmann ; it perhaps’ occurs frequently i in herbaceous Dicotyledons. Of the cases adduced by .. +See Fig. 58, p. 159. Further, Higncnics Pflanzenzelle, p. 175.—Sanio, in Paphins Jahrb. VIII. -* @°For further details see Kraus, Bau d. Nadelhélzer, Wiirzburger Natuaviss: Zeitschr. Ba. v— Goppert, Monogr. d. Fossilen Coniferen, Harlem, 1850. For figures see especially Goppert, 2.c.; also Schacht, Baum, 1 Aufl. p. 202; Lehrbuch, I. p. 233. _§ Hildebrand, Begoniaceen-stimme, p. 24... =. so : ‘Jochmann, Umbelliferarum Structura, p. 10. 492 SECONDARY CHANGES, Schwendener (Mechan. Princip, p. 148, sub. 3), Tropzeolum, Impatiens, Centranthus, and Cachrys, Zerhaps belong to this category, the others not. In any case all the exceptions of the category last-mentioned constitute transi- tional forms to those mentioned at p. 458, where there is no sharp lateral limitation of distinct ligneous bundles by radial bands of non-equivalent structure, and hence medullary rays cannot be distinguished at all, whether it be that only the large medullary rays are absent and small secondary ones appear later, as in Ephedra and Cobzea, or that rays of every rank are absent throughout, as in Crassulacez, Centra- denia, Rumex Lunaria, and Campanula Vidalii. Sect. 148. In many woods, e.g. constantly in species of Alnus and Sorbus, accumulations of parenchymatous cells occur, constituting as it were local hyper- trophies of the medullary rays; these were first described by T. Hartig as celular passages, subsequently by Nordlinger as medullary spots, and by Rossmissler as repetitions of the pith. According to the investigations of these authors and of Kraus, these structures appear in cross-section in the form of elongated spots, usually at the outer side, but not unfrequently in the middle of an annual ring, with their greatest diameter following the periphery of the ring, while they often form considerable annular segments extending through go° and more. In the vertical direction they extend like passages for distances of several feet, sometimes ending blindly, sometimes branched here and there, not unfrequently crossing one another in their irregular course. They are often conspicuous to the naked eye by their brown colouring, e.g. in the trees mentioned; in other cases they are colourless, e.g. Populus monilifera, and tremula. They consist of irregularly polyhedral, ‘irregularly arranged cells, with thick pitted walls, contents including starch, tannin, &c.; the latter are generally brown in the dry wood, and chiefly contribute to the colouring of the spots. This may also partly depend on very thin-walled, compressed (partly disorganised ?) cells at the circumference of the spot, as described by Kraus in the case of Sorbus torminalis. They owe the names given them to the similarity of their thick-walled cells to those of the pith, especially of its periphery. The medullary rays coming from the middle of the stem enter the inner side of the spots, their cells becoming. broader as they approach the latter, and assuming more and more the characteristics of those of the spots; thus the medullary rays pass over from within outwards into the passages ; they further coalesce with them laterally. On the outside new medullary rays start from the passages, and as regards their direction these are either independent of those coming from within, or lie in the same straight line with them. Local swellings of the medullary rays, due to increased breadth and number of their cells, which latter may assume irregular forms, are immediately connected in structure with the smaller spots or passages of this kind; and this also applies to the longitudinal union of neighbouring rays, by means of small groups of parenchymatous, more or less irregular cells, which abut on them. Such structures occur in Abies alba, balsamea, Pichta, in Cunninghamia, Cupressus sempervirens, and frequently also in Abies pecti- nata. In the Coniferze, and in Liquidambar, hysterogenetic and lysigenetic resin- ‘ Hartig, Forstl. Culturpflanzen.—Nérdlinger, Querschnitte, Bd. I.—Kraus, Nadelhdlzer, /.¢, p. 162. SECONDARY THICKENING. NORMAL DICOTYLEDONS. 493 canals’ often arise in them?; in the wood of Prunus avium, they are; according to Wigand ?, a principal starting-point of the disorganisation which produces the cherry- gum. The dilatations of the medullary rays and the medullary spots occur, it is true, relatively seldom; they are, however, characteristic of many woods, both Dicotyledonous and Coniferous, Ac- cording to Kraus and Nérdlinger they have been frequently observed in Betula alba, dahurica, populifolia, Crategus oxyacantha, monogyna, pyracantha, cor- data, Cydonia vulgaris, Pyrus prunifolia, Amygdalus communis, Cotoneaster mi- crophylla, Prunus spinosa, Salix aurita, Caprea, bicolor, Rhus Cotinus, Liihea grandifolia, Pterocarya caucasica, Vac- cinjum Myrtillus, Vitex incisa, Calluna vulgaris, Erythroxylon grandifolium, Guazuma ulmifolia, and Liquidambar styraciflua, besides the plants mentioned above ; while they occur rarely in Alnus viridis, Catalpa, Magnolia acuminata, and Salix triandra, I leave unmentioned the cases marked (?) by Nd6rdlinger. Among the Coniferous woods Kraus found them in Abies balsamea, Pindrow, Pichta, Picea orientalis, and Juniperus excelsa; in Abies pectinata, Cedrus Deo- dara and Larix, he only found swellings of the medullary rays. In Dicotyle- donous woods the passages occur chiefly in the lower part of the stem, and are thence continued into the roots; they FIG. 208,—Pinus silvestris ; radial longitudinal section through also extend, however, into the apices and the wood of a branch. a@—¢ ends of tracheides, with bordered pits (¢’, ’’) in superficial view; cb portion of the young wall of a branches. though here they are less tracheide with bordered pits still immature ; the further develop- 2 e ment of the latter and the contraction of the canal in the succes- numerous and constant. In the Coni- sion a—c; d, e mature condition; s, ¢ large pits on the boundary . cheid id cells of th dull: ) Fi ferous woods their course has not been Between tracheides and cells of the medullary ray (550). From minutely investigated; according to Dippel’s statements (/.c.), which concern this particular point but little, those in the white Fir which contain resin canals may be traced longitudinally for considerable’ distances, b. Zhe ligneous bundles. Sect. 149. The structure of the ligneous bundle within an annual ring differs according to the presence or absence of the particular forms of tissue, and according to the distribution of those which are present, and the form and structure of each 1 Kraus, /.c.—Dippel, Botan. Zeitg. 1863, p. 253. ? Pringsheim’s Jahrb. III. p. 118. 494: a SECONDARY CHANGES. element individually. In addition to these there are differences, to be discussed below; which are presented by the structure of the same form of tissue In the spring and autumn wood, and by the general grouping of all the elements of the ligneous body. 1. Insome few woods the bundles are composed of only oné form of tissue, to the exclusion of all the rest, and they then consist of tracheides, which, apart. from the general differences between spring and autumn wood, have everywhere the same structure. In these cases the parenchymatous system is represented only by the medullary rays, which are inserted everywhere in great numbers between the bundles, Among Dicotyledonous woods the Winieree belong to this category; namely, Drimys Winteri and its allies, Tasmannia’ aromatica?, and Trochodendron aralioides’, the systematic position of which is doubtful: among Condfere, Taxus baccata_ belongs here, according to Sanio, but in the case of this tree Hartig and Kraus state that scanty bundle-parenchyma is present. : In the other Coniferze the “-acheides, of which- the. main mass of -the wood is 7 FIG. 209.—Junipérus communis; stem; cross-section through the autumn wood, bast, and cambium during 7 the winter's rest (end of September). %—/ outermost series of the autuntn wood; 4, 4 series of bast-fibres. At X there is only one caubial cell between % and 6; m—m medullary rays. homogeneously composed, are accompanied by bundle-parenchyma, which sometimes gccurs in single vertical rows, scattered among the tracheides, while sometimes, as in many Abietineze, it forms the coating of the resin-canals, . In the Conifer (Figs. 208 and 209) and the other plants just mentioned, the tra- cheides are arranged in radial rows; they are quadrangular, as seen in cross-section, when those which belong to neighbouring radial rows ‘stand opposite to one another, hexagonal or pentagonal when the radial rows are alternate; their ends are elongated and sharp, owing to inclination of the radial surfaces (comp. p. 469). The radial surfaces have large . corresponding bordered pits, which in the Winterex, the Araucariz, Dammare, and.the wood of the root in other Coniferz, form two or more longitudinal. rows, while in the 1 Géppert, Linnza, Bd. XVI. p. 134.+-Kraus, /.¢. ? Eichler, in Flora, 1864, p. 451. ag 8 SECONDARY THICKENING. NORMAL DICOTYLEDONS. 495. stem-wood of the rest of the Coniferz they form only a single longitudinal row, apart from individual exceptions, which are frequent, for example, in Larix, In the Coniferz pits only occur on the tangential surfaces in the autumn wood. Taxus, Cephalotaxus, and Torreya further show spiral or annular thickenings on the i inner surface of the wall of the tracheide*. Comp. Chap. IV. aS The resin-canals, surrounded by parenchyma as epithelium, occur in the ligneous bundles of the same Abietine which possess horizontal canals in the medullary rays (p. 490). They run longitudinally, and, as seen in transverse section, lie scattered in a ring in the external region of every annular layer. Their number varies according to species and individual; Von Mohl? counted, for example, on an equal transverse surface embracing several somal layers, in Pinus nigricans 190, P. silvestris 124, Larix europxa 128, Picea excelsa 78. 2. The ligneous bundle of all Dicotyledonous woods, except those mentioned above, and of the Gnetaceze (Ephedra, Gnetum) always contains vessels, and at least one of the forms of cells distinguished; of the: Jatter bundle-parenchyma and inter- mediate fibres are usually present together. These are further accompanied, accord- ing to the species, by some or all of the other tissues characterised in Sects. 142-144. Of the combinations thus possible, the following have been demonstrated, according, to Sanio’s statements, in the trees and shrubs investigated. 1. Vessels, Tracheides, Bundte-parenchyma, Intermediate fibres. (2) Only with bundle-parenchyma: Ilex Aquifolium, Staphylea pinnata, Rosa. canina, Crategus monogyna, Pyrus communis, Spirza opulifolia, Camellia, &c. (6) Only intermediate fibres: Porlieria. (c) Both forms of cell: Jasminum revolutum, Kerria, Potentilla fruticosa, Casua- rina equisetifolia and torulosa, Aristolochia Sipho, and many others. 2. Vessels, Tracheides, Fibrous cells, Bundle-parenchyma, Intermediate fibres. (a) Only bundle-parenchyma: fibrous cells non-septate: e.g. Sambucus nigra, racemosa, Acer platanoides, pseudoplatanus, campestre. (6) Bundle-parenchyma and intermediate fibres, fibrous cells non- septate = Ephedra monostachya, Berberis vulgaris, Mahonia*. (c) Bundle-parenchyma, fibrous cells septate and non-septate: Punica, Euonymus latifolius, europzus, Celastrus scandens, Vitis vinifera, Fuchsia globosa, Centradenia grandifolia, Hedera Helix, &c. ' (d) All four forms of cell: Miihlenbeckia complexa, Ficus (?). web oe 3. Vessels, Tracheides, Woody fibres, Bundle-parenchyma, Intermediate fibres, This is the prevalent, one may almost say the typical, combination. (a) Only bundle-parenchyma: Sparmannia africana, Galyeanthes Rhamnus ca. thartica, Ribes rubrum, Quercus, Castanea, Carpinus spec., Amygdalez, Melaleuca, Callistemon spec., &c. (4) Only intermediate fibres: Caragana arborescens. (c) Both forms of cell, Most Dicotyledonous woods no doubt belong to this series, e.g. Salix, Populus spec., Liriodendron, Magnolia acuminata, Alnus glutinosa, Betula alba, Juglans regia, Nerium, Tilia, Hakea suaveolens, Ailantus, Robinia, Gleditschia SBEC es Ulex europeus, &c. 1.Compare Hartig, Kraus, Géppert, /. ¢.—Von Mohl, Schacht, Botan. Zeitg. 1862. ? Botan. Zeitg. 1859, p. 340. , 8% Compare Sanio, in Pringsheim’s Jahrb. IX. p. 55. 496 SECONDARY CHANGES. > 4. Vessels, Woody fibres, Parenchyma, Intermediate fibres. (a) Both kinds of cells : Fraxinus excelsior, Ornus, Citrus medica, Platahus, &c, (6) Only intermediate fibres: Viscum album. (c) Only bundle-parenchyma: Avicennia. 5. Vessels, Fibrous cells, Parenchyma. Cheiranthus Cheiri, Begonia. Here no doubt also belong many of the Crassulacez and Caryophyllez which still need more minute analysis, 6. Vessels, Fibrous cells, Parenchyma, Woody fibres (?). Coleus Macrzi, Eugenia australis, Hydrangea hortensis. 4. Vessels, Tracheides, Woody fibres, Fibrous cells (septate), Parenchyma, Intermediate Sires. Ceratonia siliqua, Bignonia capreolata, doubtful however-as to the woody fibres according to the existing data. Although no strict law holds good, without exception, for the distribution of these tissues in the woods containing vessels, still certain general rules may be given. The vesse/s occur in all the layers of the annual ring, but are usually more fre- quent in the internal than in the external portion. Only Bombax Ceiba shows the opposite condition, according to Sanio. It is not uncommon, however, for their frequency to show no difference from within outwards (Acacia Sophora, floribunda, Enckea media, Artemisia Abrotanum); or only a slight one, the pores, or groups of pores which they present in cross-section, being scattered uniformly through the wood, e.g. Laurus nobilis, Adsculus, Acer, Populus. The vessels rarely form the main, fundamental mass of the wood (Avicennia), As a rule they lie in small groups in the non-equivalent fundamental mass, and these groups are either isolated or arranged in more or less interrupted radial bands or concentric zones (Hedera Helix). They are either everywhere of about the same width, or more usually their width is greater in the inner part of the annual ring, and diminishes gradually or suddenly towards the outside. With this difference in size, a difference in structure is in many cases also united, in so far as the narrow vessels have spiral fibres, while the wide ones have none (Morus alba, Broussonetia papyrifera, Gymnocladus, Virgilia lutea, Celtis aus- tralis, Ulmus suberosa, Catalpa, and Robinia pseudacacia); or the structure is the same in all (Quercus pedunculata, Castanea vesca, Fraxinus, Amorpha fruticosa, Sophora japonica, and Periploca). In the above-mentioned ‘ parenchymatous’ woods, as Bombax, Carica, &c., and also in the roots still to be discussed below, the bundle-parenchyma forms the main mass in which the vessels and other elements are inserted in groups. In the solid ‘woody’ woods its arrangement has a regular relation to that of the vessels. It usually accompanies the latter, either in such a manner that it surrounds each vessel or group of vessels singly—paratracheal parenchyma according to Sanio, e.g. Enckea media; or it forms tangential bands, alternating with similar ones, con- sisting chiefly of tracheides or fibres, in or at the side of which the vessels stand: Sanio’s metatracheal parenchyma. The latter, for example, is the case in the spring wood of Tectona grandis, in the autumn wood of Fraxinus, in the autumn and spring wood of Amorpha fruticosa, Sophora japonica, Robinia pseudacacia, Gleditschia triacanthos, Gymnocladus, Virgilia, Caragana arborescens, Paulownia, Morus, Brous- sonetia, Ailantus, Tamarix gallica, &c. In Casuarina equisetifolia, torulosa, Hakea SECONDARY THICKENING, NORMAL DICOTYLEDONS, 497 suaveolens, and other Proteaces, species of Ficus, Cordia pallida and many others, every annual ring has several concentric bands of metatracheal parenchyma. According to Sanio, parenchyma always occurs scattered between the tracheides in Dicotyledonous woods, with the exception of Casuarina, where it only occurs in the metatracheal arrangement, and of Rosmarinus officinalis. It is absent between the typical woody fibres, according to Sanio, with the exception of Edwardsia grandiflora, Ulex europzeus, Celtis australis, Olea europzea, and further of Hibiscus Rosa sinensis, where it actually forms tangential bands between the fibres. In Tamarix gallica it occurs even between fibrous cells which contain starch. In some few cases of the wood of roots, which consists chiefly of masses of parenchyma, the latter are the seat of passages containing secretions: root of Inula Helenium' and Opoponax Chironium ; Trécul’s statement respecting Oenanthe crocata perhaps also refers to this subject ?. For the zxfermediate fibres, the same rules hold good as for the bundle-paren- chyma, because they occur either accompanying or representing the latter. Although the woody fidres may occur in all the layers of the annual ring, they are present in especially large numbers in its central portion, in the case of hard woods. They here usually-form the fundamental mass, in which the other elements, particularly vessels and parenchyma, are imbedded; in many woods, e.g. Robinia and Gleditschia, they occur only in the central portion of the ring, and are absent in the spring and autumn wood. In woods which are mainly parenchymatous (as Bombax and Cheirostemon), and in that of Avicennia, which chiefly consists of vessels, the fibres are only imbedded in small groups, or singly, between the elements of the funda- mental mass. For the fidrous cells, whether septate or non-septate (Berberis, Clematis, Vitis, Tamarix, Punica, &c.), the same rule of distribution holds good as for the woody fibres. As shown by the above statements as to their occurrence, the two elements, which are similar in form and in the structure of their walls, may mutually represent each other, indeed it is doubtful whether the case mentioned under 7, of the simul- taneous presence of both, ever occurs. The ¢racheides may likewise form by themselves the fundamental mass of the wood, representing, as it were, the two tissues last mentioned; this is the case in the combination mentioned above under 1, e.g. Pomacez, Camellia, &c. They then always belong to the ‘fibriform’ category, resembling woody fibres in shape, and in the character of their walls. Where, on the other hand, they occur in conjunction with fibres and fibrous cells, they are present chiefly in the neighbourhood of the vessels : and in fact they occur next them in small numbers, and isolated, when the latter are scattered singly or in small groups in the annual ring, and are all alike (e. g. Punica, Fuchsia globosa, Ceratonia, and Nerium). When, on the other hand, two kinds of vessels, distinguished by their size, and usually also by special structure, are present, then the tracheides accompany the small ones, rarely the large ones also (Quercus peduncu- lata, Castanea vesca, and Periploca). According to the mode of occurrence of the small vessels, they are then either distributed with the latter in the whole annual ring (Ulex 1 See Berg, Atlas d. pharm. Waarenk. Taf. X. 2 Trécul, /.¢.; compare p. 448. Kk 498 SECONDARY CHANGES. europzus, and Rosmarinus), or limited to its outer portion (Morus alba, Broussonetia, Catalpa, Paulownia, Sophora japonica, Gymnocladus canadensis, Robinia pseudacacia, Corylus, Carpinus, and Ostrya)—Secondly, the tracheides are in many cases most abundant in the outer part of the annual ring, or confined to it, even apart from their connection with particular vessels. In the annual ring of Ribes nigrum, Syringa vulgaris, Ligustrum vulgare, Euonymus europzus and latifolius, they successively increase in frequency towards the outer border of the annual ring, until they form the fundamental mass in which isolated vessels and fibres, or fibrous cells are imbedded ; while in the inner portion of the ring the fundamental mass consists of fibres or fibrous cells, and the tracheides occur isolated, side by side with the vessels. Or the tracheides occur only at the outermost, autumnal limit of the annual ring, while the latter otherwise only contains vessels of one kind, as in Tilia, Salix hippophefolia, acutifolia, Populus tremula, pyramidalis, Rhamnus Frangula, Juglans regia, cinerea, Pterocarya, Diospyros virginiana, Betula alba, Alnus glutinosa, Laurus nobilis, Cam- phora ; Acer pseudoplatanus, platanoides, campestre ; Sambucus nigra, and racemosa. Sxcr. 150. Certain phenomena which depart in some degree from the usual rules, but belong nevertheless to the normal Dicotyledonous type, are here to be mentioned as special cases by way of supplement; they chiefly concern the wood of species or families which are distinguished by special adaptations, and by peculiarities of form correlated with these. Most of these phenomena still require more minute investigation, for which the following paragraphs will only contain indications. We have, first, to return to the woods enumerated on p. 458 and p. 492, which are destitute of medullary rays. In those cases where only the primary large medul- lary rays are absent, while small ones soon appear, as Ephedra, Cobzea, and no doubt Xanthosia also, the above general rules for the structure and distribution of the organs hold good for the main mass of the wood. As far as the absence of the medullary rays extends in the cases mentioned, and in cases of their entire sup- pression throughout the whole secondary wood, the main and fundamental mass, with isolated exceptions to be mentioned below, consists of thick-walled fibrous cells or fibres—the latter being gradually developed from the former: these are usually elongated, though short in Echeveria pubescens, and have sharply pointed ends and a regular radial arrangement. Leaving out of consideration the xylem portions of the original leaf-trace bundles, which project into the pith, and belong to the medul- lary sheath to be described below (Sect. 152), the vessels—perhaps also series of tracheides of otherwise similar structure—and groups of bundle-parenchyma are inserted in the fundamental mass in the more strongly developed wood. The grouping of the vessels and of the parenchyma, and the quantity and quality of the latter, vary according to the particular case. In the aérial stem of the Crassulacee’ investigated, which has a somewhat strongly developed ligneous body, the parenchyma consists of elongated thin-walled cells, which remain unlignified, and accompany the vessels in longitudinal rows. The vessels and parenchyma are either quite isolated and scattered in the mass of fibres, at most two or three occurring together—Sedum maximum; or some of them form larger groups, consisting of as many as twelve elements—S. populifolium, and Echeveria * * Compare Brongniart, Arch, du Muséum d’Hist. Nat. tom. I.—Regnault, 7. ¢. SECONDARY THICKENING, NORMAL DICOTYLEDONS. 499 pubescens; or they form with the parenchyma extensive groups, constituting trans- versely elongated, occasionally irregular, interrupted transverse bands— Semper- vivum arboreum. Lastly, in the small creeping stems of Sedum reflexum, they form, together with the delicate cells, thick, multiseriate, continuous annular zones, which alternate with similar fibrous zones. Other species however, even those with thick stems, as Crassula lactea, Sedum ternatum, and Echeveria pubescens, form mere traces of secondary wood, which are scarcely worth mentioning. In the Caryophyliee investigated! (Dianthus, Gypsophila, Silene spec., and Arenaria graminifolia), especially in their rhizomes, thin-walled, long-celled parenchyma, often forming large irregular islets or annular segments, is inserted between fibrous masses of similar form, but in the foliage stems it may be absent, e. g. Gypsophila altissima. The vessels are numerous, and occur among both forms of tissue, often forming interrupted radial rows; in the fibrous groups they are often unaccompanied by parenchyma, In the foliage shoot of Dianthus plumarius the main mass of the very slight secondary thickening consists of vessels. In the stem of Rumex Lunaria, the pitted vessels, accompanied by rows of bundle- parenchyma or: by intermediate fibres, lie in very regular but interrupted radial rows in the fundamental mass, and the latter consists of fibrous cells densely filled with large starch-grains. With the exception of the starch, the same holds good for Centradenia grandifolia (comp. p. 484). In Campanula Vidalii the vessels are very isolated, and scattered in the radial series of fibrous cells, while I was unable to find any parenchyma or intermediate fibres accompanying them. The RAinanthacee require further investigation. The secondary wood of the thicker, fleshy stems of Cactee* must again be men- tioned here. The peculiar structure of the ligneous bundles, vessels, and tracheides, in the Mamillarie, Echinocactus, &c., has already been repeatedly noticed above (comp..p. 478), and it must here be added, that in the genera mentioned the whole secondary ligneous bundle consists of the spirally and annularly thickened trachez, usually with a deeply projecting fibre, between which very thin-walled bundle- parenchyma is imbedded, in single longitudinal rows. The medullary rays of all ranks resemble the latter as regards the delicacy of their cells. In other Cactez, especially species of Cereus and Opuntia, the secondary ligneous bundle consists of tough woody fibres, and the scattered reticulated vessels which are usually accom- panied by parenchyma (see p. 479). The parenchyma, including that of the medul- lary rays, may be very thick-walled and lignified, e.g. in Cereus speciosissimus. In the Opuntiz short tracheides are also present, with the above-mentioned . thickenings on the wall, which project inwards in the form of plates, and are usually annular. These elements are sometimes distributed singly in the interior of the bundle (O. tunicata, O. robusta), sometimes at its edges (O. cylindrica, ramulifera, andicola). Lastly, the wood which forms the floating apparatus of the stems of certain Leguminose of the genera Aischynomene and Herminiera, which vegetate on the surface of the water, may deserve mention. Their structure appears to be very 1 Regnault, Z.c. p. 118, pl. VI. 2 Compare Brongniart, /.c.—Schleiden, Anat. d. Cacteen, /.¢. (p. 156). Kka 500 SECONDARY CHANGES, uniform in the various species of these floating woods, as far as the existing de- scriptions extend’, Hallier’s figures (/. c.) and Figs. 51-53 in Schleiden’s Grundziige (3 Ed.), I. p. 261, which at any rate represent a similar object, may serve to illustrate its coarser characteristics. The following short description refers especially to the wood of the Ambatsch of the White Nile (Herminiera Elaphroxylon=Aidemone mirabilis Kotschy ”) :— The extremely light wood has no distinct annual rings. It consists, in its main mass, of elements which must be called tracheides (until undried material has been investigated), because in their condition as observed, they contain only air, without a trace of proto- plasm or remnants of cell-contents. They stand alternately in radial rows, and have the form of hexagonal upright prisms, about three times as high as broad, with terminal sur- faces inclined to the radial plane at an angle of about 45°, either unilaterally, or bilaterally, like a roof, Their thin colourless membrane is very beautifully thickened on the entire terminal surface, by a fine delicate network of fibres, while on the radial, and in a lesser degree on the tangential lateral surfaces, it is provided with small groups of simple minute pits. / The mass consisting of these tracheides is traversed (1) by very numerous parenchy- matous medullary rays, containing starch, which are 1-10, on the average about six cells high and one cell in breadth, and further by some larger medullary rays, several cells broad in the middle; the cells of the medullary rays are elongated and procumbent. (2) By narrow bands arranged in irregular and often interrupted concentric annular zones, which are themselves crossed by the medullary rays; these bands chiefly consist of long-pointed fibres, with their ends overlapping one another, in a radially and tangen- tially oblique direction. In these bands, or more correctly on their inner (medullary) side, lie large pitted vessels, usually isolated, rarely forming short radial rows of few elements, and in both cases at long distances from one another laterally, being separated by several subdivisions of the wood, as détermined by the medullary rays, Every vessel, or every group of vessels, is partly surrounded by a single layer of intermediate fibres containing starch, or.by parenchymatous cells, with a moderately thickened pitted wall, 2-4 of which stand one above another; and very thin-walled, narrow parenchymatous cells, likewise containing starch, are continued in a single layer over the inner surface of each fibrous band. Between these the above-mentioned (p. 141) chambered crystal-sacs are here inserted. Tracheides, intermediate cells, members of the vessels, and also the medullary rays of medium altitude, are everywhere of approximately equal height, and lie with their ends in the same horizontal planes, thus forming regular horizontal layers. Their form and arrangement (with the obvious exception of the members of vessels) are similar to those usual in cambial cells, so that it may safely be assumed that they are derived from a cambial zone consisting of cells similar to them in height and form. The fibres, on the other hand, are (as estimated) at least double as long as the other elements mentioned, and must thus have undergone the corresponding elongation (and displacement) on their differentiation from the cambial zone, c. Changes of the individual forms of tissue in the annual ring. Scr. 151. In those of the cases just brought forward, in which the distribution of the forms of tissue is different in the successive layers of an annual increment of growth, and where the stricture of the autumn-wood of one year is thus different from that of the neighbouring spring-wood of the next year, a limiting line must appear between * Hallier, Botan. Zeitg. 1859, p. 152; 1864, p. 93. * Compare Schweinfurth, Beitr. z. Flora Athiopensis, p. 9. SECONDARY THICKENING. NORMAL DICOTYLEDONS. 501 the two. In addition to this cause of the demarcation of the annual rings, which con- sists in the distribution of non-equivalent tissues, and is not always present, there are two other causes, consisting in differences of form and differences of structure of equivalent tissue-elements, appearing in the successive, and especially in the extreme zones of an annual increment of growth. The first of these phenomena consists universally, in a shortening of the radial diameter, and thus in a fangential flattening of the elements, at the outer limit of the autumn-wood, and this is due to increasing cortical pressure’. Changes in the average length may, as has been stated, be connected with this.» The second cause, not present in all cases, consists in an increase in the thickness of the wall, and some- times in further changes of its structure. These changes, especially the shortening of the radial diameter, affect both the elements of the medullary rays and those of the ligneous bundles. They come on either suddenly or gradually, and in fact this latter difference depends, partly on species, partly on differences in the vigour of development of the annual rings in one and the same wood, These conditions are shown most simply and clearly in the woods of the Conzfere’, the bulk of which consists only of tracheides and medullary rays. In the procumbent elements of the medullary rays, the shortening of the radial diameter at the autumnal limit is not very conspicuous, though it does occur. The tracheides however are rela- tively wide at the spring limit of each annual layer, usually quadrangular as seen in cross- section, though sometimes pentagonal or hexagonal, and their radial diameter is equal to the tangential, or even somewhat greater; at the autumnal limit, on the other hand, they are always much flattened, i.e. the radial diameter is shortened. According to N. Miiller*, a diminution of their length in the autumn-wood is connected with this, in the case of the Fir. To this is further added an increase, accompanying the flattening, not only of the relative, but of the absolute thickness of the wall; and the appearance of pits on the tangential surfaces of the wall, while in the spring elements, with a wide lumen, the latter are limited to the radial surfaces. By way of example, these conditions may be illustrated by the mean magnitudes found by Mohl‘ in the case of a well-grown specimen of Pinus sylvestris, thirty years old; they are expressed in Paris lines :— Spring-wood, Autunn-wood. Radial diameter. . . . . O0204 . +» 00056 Stem Tangential . . . . 4. «. OOr42 . . . « « O-0%42 Thickness of wall . . . . oo0rg . . . 4. 00031 Radial diameter. . . . . 010232 . . . . « 00094 Root . : Tangential . 2... . . oo16r . . . . . Oor6r Thickness of wall . . . . O0018 . « . « 2 00035 The sudden or gradual appearance of these differences, and the relative thickness of the zones with narrow and wide lumen, here depend in general on the thickness of the annular zones, as will be shown below. 1H. de Vries, 2.¢. (p. 475). 2 Géppert, Monogr. d. foss. Coniferen, /.c.—Von Mohl, Botan. Zeitg. 1862, p. 225, &c.— Kraus, /.c, 5 Botan, Untersuchungen, IV. z, p. 190. * Lie p. 237. 502 SECONDARY CHANGES, Deviations from the typical structure occur, in so far as groups of unusually thick- walled tracheides may appear in the different regions of the annular ring. These, as seen in cross-section, form band-shaped, annular segments, of a brownish yellow colour, similar to that of the autumn wood; in Pinus sylvestris they always occur in the innermost annual rings, and often also in the outer ones. Comp. Sanio, Pringsh, Jahrb. IX. ror. In the Dicotyledonous woods the demarcation of the annual ring is further influ- enced by the distribution of non-equivalent forms of tissue, and of the changes which each of these shows in the successive zones of the annual ring. The shortening of the radial diameter at the autumnal limit, sometimes occurring suddenly and sometimes gradually, and either amounting to a decided flattening or only occurring in a slight degree, according to the special case, is here also general : leaving out of consideration the vessels, which will be specially mentioned below, it is approximately uniform in the forms of tissue which occur in contiguity, more rarely it takes place to a very unequal extent: as in the marked and sudden flattening of the tracheides of the autumn-wood, while the radial shortening of the fibrous cells which accompany them is gradual and relatively much slighter, in Clematis Vitalba, and Mahonia Aquifolium. With reference to the increase of the absolute thickness of the wall at the autumnal limit, a distinction must be drawn between the particular forms of tissue, for the latter differ from one another in average thickness in all the parts of any’ wood. In the case of equivalent forms of tissue both conditions occur, according to the kind of wood, namely, either a marked increase of the thickness of wall at the autumnal limit or its approximate constancy throughout. The latter, for example, is the case— In the bundle-parenchyma of Gleditschia triacanthos, Ailantus glandulosa, Sophora japonica, and Caragana arborescens ; In the fibrous cells of Berberis, and Mahonia ; In the vessel-like tracheides of Betula, Alnus, Populus, Salix spec., Magnolia acuminata, Sambucus nigra, &c.; In the thick-walled fibriform tracheides of Cornus sanguinea, Syringa vulgaris, and Buxus sempervirens. Increase of the thickening of the wall at the autumnal limit takes place for example in— The bundle-parenchyma of Gymnocladus, Morus alba, Broussonetia, Paulownia, and Amorpha fruticosa ; In the woody fibres of Laurus Camphora, Jatropha Manihot, and Carpinus Betulus; In the vessel-like tracheides of Caragana arborescens, Carpinus Betulus, and Ostrya virginica ; In the fibriform ones of Syringa Josikea, Philadelphus coronarius, Kerria japonica, &c. Contrary to the rule holding good for other woods, Sanio observed a decrease of the thickness of the wall at the autumnal limit, in the tracheides of Staphylea pin- nata, both as compared with the inner ones of the same annual ring, and with those of the spring-wood of the next following ring. As regards the degree of increase in the thickness of the walls at the autumnal SECONDARY THICKENING. NORMAL DICOTYLEDONS. 503 limit, no accurate comparisons of the different forms of tissue have been carried out, but one may assume from the appearances, that it takes place in an approximately uniform ratio in all. Since then, as was shown above, the different forms of tissue of the same wood usually have a very unequal average thickness of wall, the total relative bulk of the walls at the autumnal limit must depend in the first instance on the dis- tribution of the non-equivalent elements in the annual ring, and secondly on the increase of thickness, which has just been discussed, in each individual form of tissue. Thus, where, for example, the autumnal limit consists chiefly or exclusively of paren- chyma, as in Morus alba, Broussonetia, Fraxinus, Robinia pseudacacia, Caragana arborescens, Amorpha fruticosa, Virgilia, Gleditschia, Catalpa, Paulownia, Ailantus, &c., or of vessel-like, thin-walled tracheides, as in Betula alba, Alnus glutinosa, Populus, Salix spec., and Cytisus Laburnum, while in the inner part of the ring thick-walled fibres prevail, then the autumnal limit is, on the average, thinner-walled than the next inner-portion of the ring. The manifold possible combinations and modifications in this respect follow from what has been already stated. : “The Vessels in most Dicotyledonous woods behave similarly to the other elements, ///// i 80 far as they decrease on the average in width from the spring to the autumn limit of the annual ring, though without, as a rule, undergoing any conspicuous tan- gential flattening. With this a decrease in the number of the vessels from within outwards is often connected. Both differences may appear very gradually, as in the annual ring of the Salicineze, Pomacez, Fagus, Buxus, Cornus sanguinea, &c., and may even be scarcely noticeable (Enckea media, Ulex europzeus). In other woods, as Quercus pedunculata, Fraxinus, and Castanea, they appear suddenly, in such a manner that the first spring-zone of the ring, which is rendered highly porous by numerous very wide vessels, is succeeded on the outside by less numerous and much narrower vessels, occurring between the other elements. According to the degree in which this peculi- arity of the spring-wood, which may be shortly characterised by the name ‘ porosity,’ clearly appears, and coincides with other differences of structure between the annual limits and the middle portion of the ring, the boundary comes out more or less sharply even as seen with the naked eye. The trees last-mentioned form examples of especially sharp demarcation of the annular rings, and this is also the case with Fraxinus, Robinia pseudacacia and Gleditschia triacanthos, as the boundary is marked both by large vessels, and by the other conditions of structure mentioned above. If, on the other hand, the vessels, while of approximately equal width, are uniformly distributed singly or in groups, as in Enckea media, Ulex europzeus, and no doubt Olea europza also, the limit of the annual rings may be perceptible with difficulty, or scarcely at all, not only when slightly magnified, but even under microscopic investigation. These various degrees in the demarcation of the annual rings, at once suggest the conjecture that cases of their entire suppression may occur. This unquestionably takes place as an individual peculiarity. Many plants, e.g. the Araucariz, appear to have a special tendency to this, and such individual phenomena will be discussed below. Plants, in which the demarcation of the annual rings is constantly absent as a specific peculiarity, are certainly rare, and the statements respecting such cases have been so often disputed, that I am unwilling to adduce even those plants in which I was myself unable to find limits between annual zones,—namely the woody Piperacez, Opuntia, 504, SECONDARY CHANGES. Mamillariee and Cacteze, and Cobw#a scandens,—as certain examples of plants which are constantly destitute of annual rings’. Whether in tropical trees, as well as in our temperate zones, the annual ring as anatomically distinguished always represents the yearly growth, or whether there are trees with half-yearly rings, i. e. such as form two annual rings in the year, corresponding to the two periods of vegetation falling in one annual period, as has been stated to be the case in Adansonia digitata, appears to me to be a question which ought to be tested, but one that does not belong to the present anatomical exposition. Attention may here, perhaps superfluously, be recalled to the fact that in woods with alternating concentric bands of non-equivalent tissue, as Ficus, Casuarina, &c., markings are present, which resemble annual rings on superficial observation, but from which the true annual rings are distinguished by the characteristic structure of the autumn wood, d. Normal differences of successive zones of growth and annual rings. Scr. 152. The first innermost annual ring of every wood necessarily shows certain essential differences in its structure from all the later ones. In the stem, instead of the characteristic spring-wood, it has at its internal limit the groups of spiral, annular, and reticulated vessels (or tracheides) corresponding to the primary vascular bundles, but absent from the intermediate bundles; these usually project more or less into the pith, and, together with the neighbouring tissue-elements, belonging partly to the ligneous ring, partly to the pith, the structural peculiarities of which have been often mentioned in former sections, they have long been distinguished collectively as the medullary sheath or medullary crown, Corona. In the root, the internal limit of the first annual ring has a different structure, corresponding to the conditions described on p. 473. The secondary wood here closely surrounds the original axial xylem-plates. It is quite a prevailing rule that these latter remain separated from the secondary vessels and tracheides by at least one layer of connective cells*. No case is known in which the secondary elements border immediately on the outermost narrow, spiral, and annular vessels of the primary plates. On the other hand, it often occurs that the inner pitted vessels (or tracheides) of the original plates stand in immediate connection with equivalent elements of the secondary wood; this has been observed in Taraxacum and Ranunculus repens (Fig. 165, p. 356). The cells which usually border on the xylem-plates consist of the inner layer of those connective cells which were originally present in this position, while from their outer layer the cambial ring bordering on the phloem-group has been derived (pp. 351, 473). In those cases where the pitted vessels are contiguous, it still remains to be investigated whether the latter are derived directly from the above-mentioned connective cells, or from a cambial zone which, in the latter case, must originate in the innermost layer of connective cells. All the tissue-elements of the region in question usually become very thick-walled, and have relatively narrow cavities, especially in * Compare Link, Philos. Bot. p. 136.—Meyen, Physiol. I. p: 361.—Treviranus, Physiol. I. 235. —Unger, Botan. Zeitg. 1847, p. 267.—Schacht, Lehrb. II. p, 62.—Sanio, Botan. Zeitg. 1863, p. 392- ® Compare Botan. Zeitg. 1844, p. 367. 3 See Van Tieghem, /. «. SECONDARY THICKENING. NORMAL DICOTYLEDONS. 505 true woody plants. The accurate distinction of the individual elements one from another therefore presents great technical difficulties; even in good transverse sections the recognition of the original xylem-plates between their thick-walled next neighbours is often very difficult, besides which the narrow vessels at their angles often seem to become indistinct owing to pressure from the surrounding tissues. In addition to these peculiarities of the internal limit of the wood in stem and root, further peculiarities exist in the inner secondary ligneous mass itself. In rare cases, especially in the stem of Mahonia Aquifolium, Berberis vulgaris, Pelargonium roseum, and Solanum Dulcamara, Sanio found septate fibrous cells in the first annual ring, while they are absent in the succeeding ones. And further, in the first and next-following annual rings of the stem and its branches, in many though not all Dicotyledonous woods, although the elements characteristic of the species are all present, yet their characteristic arrangement does not appear clearly till later, as it is merely indicated in the former region. ‘The groups of vessels, and the parenchymatous zones of Hedera Helix, Quercus pedunculata, Juglans, Casuarina, &c.’, are examples of this. The characteristic structure of a wood is therefore not always to be recognised with clearness from its inner rings. Similar phenomena, which have not at present been accurately investigated, may occur in the case of roots. Sect. 153. The most remarkable phenomenon of this category is the change in the average size of equivalent elementary organs, as regards both their width and length, which accompanies the growth in thickness. In most woods this change takes place in such a manner that the average size increases during a series of years, and then attains a definite value which remains constant in the succeeding years. The change in question takes place in a different manner in the stem, branches, and roots of the same tree, and in their different transverse zones; in some cases it depends on’a corresponding successive increase in size of the cambial cells, while in others it is independent of this. The most complete measurements referring to this point were carried out on Pinus silvestris by Sanio*, who determined the mean length of the tracheides, and their mean tangential breadth in the autumn wood. He thus sums up his results :— 1. In the stem and branches the tracheides everywhere increase from within outwards, throughout a number of annual rings, until they have attained a definite size, which then remains constant for the following annual rings. 2. The constant final size changes in the stem in such a manner that it con- stantly increases from below upwards, reaches its maximum at a definite height, and then again diminishes towards the summit. 3. The final size of the tracheides in the branches is less than in the stem, but is dependent on the latter, inasmuch as those branches which arise from the stem at a level where the tracheides are larger, themselves have larger tracheides than those which arise at a level where the constant size is less. 4. In the gnarled branches of the summit, the constant size in the outer annual ring also at first increases towards the apex, and then falls again; but here irregu- larities occur, which may be absent in regularly grown branches. 1 Sanio, Botan, Zeitg. 1863, p. 397. 2 Pringsheim’s Jahrb. VIII. p. 401, &c, 506 SECONDARY CHANGES. g. In the root the width of the elements first increases, then falls again, and next rises to a constant amount. An increase in length also takes place, but could not be exactly determined. The absolute size of the elements is not the same at the same place in different individual trees, but these differences do not infringe the general rule. In order to illustrate the absolute dimensions, some of Sanio’s determinations (cf. Zc), taken from a main axis 110 years old, may here be cited. ML, = mean length, MBr. = mean breadth of the tracheides, the latter determined in the autumn wood. A. Disk 21 years old, from the summit. B. 4, 35. 99-9) above the thick branches of the crown. C. 5, 72. 5)» from the shaft, 36’ above the ground. D. 5) 105. 5) 45 Close above the ground. The dimensions are expressed in millimetres. A, B. Annual ring. Annual ring. 1 MBr.: 0-016; ML.: 0-78 1 MBr.: 0-016; ML.: 0-80 T4949 ” » 74 I5 » 2 » 2-60 18 5 ”. (2.20 17» ” x «274 20 yy n” » -29T 18 yy ” » 282 aI yy 0.026 yy 282 IQ 45) 3 9) «2°82 20 4 ” » «—-282 22» ” yy «282 35) 35 0-028 y» 2°78 C. D. Annual ring. Annual ring. 1 MBr.: 0-017; ML.: 0-95 1 MBr.: 0o-of1; ML.: — 17 5 ” » = 2-74 20 5 ” » «187 19 » ” » 3°13 29 » ” » 248 31 ” » 369 39° ” 9 260 37 9 ” » 3°87 3I0 oy ” sy 265 38 yy ” » «3°90 46 ,, ” y 265 39 oo» ” » 4:00 60 ” 265 49 » ” 9 4104 80 5, ” » 26) ° 43» ” » 4°09 105 » 0028 » = 265 45» ” » 4°20 46 55 ” = «4°20 72 34 0032 ” 4°21. In the Dicotyledonous woods investigated1, the conditions in question vary ac- cording to the species; some show no increase in the size of the elements in the successive rings: e.g. Mahonia Aquifolium; or only an inconsiderable one, as in Berberis vulgaris, where only the vessels in the spring wood become wider. In the other cases investigated (Caragana arborescens, Sophora japonica, Saro- thamnus scoparius, Acacia longifolia, Carpinus Betulus, Quercus pedunculata, Cornus sanguinea, Rhamnus cathartica, and Ficus elastica), an increase in length takes place from within outwards, which affects the individual forms of tissue unequally. The * Sanio, Botan. Zeitg, 1863, /.c.—Pringsheim’s Jahrb, IX. p. 52, &c. SECONDARY THICKENING. NORMAL DICOTYLEDONS. 507 fibres always increase in length. The cells of the bundle-parenchyma show no change. The members of vessels and the tracheides behave differently according to the kind of wood. By way of example we may reproduce Sanio’s statements regarding the mean Tength of the elements in a stem of Quercus pedunculata 130 years old, with very narrow annual rings. The dimensions are expressed in millimetres. Members of Woody fibres. Tracheides. large vessels. ust Annualring . . 1. 1. O42 2 2 6 6 039 6 6 6 ee end 5 ee ee O60 ~ 6 & we O84 « 2s we (OCR5 4th 9 eR Ge OR ow eee OBR as ok we RE Three outermost rings . . 1-22 . . . » O72 . «. « . 0°36, An increase in the width, and in the thickness of the walls of the elements, accompanying the increase of the latter in length, is likewise evident in many Dicotyledonous woods, especially in the case of the large vessels of the spring-wood. Quercus pedunculata is a typical example of this. Sanio found the mean radial diameter of these vessels in the third annual ring =0-08™™; its definitive dimension, which is not attained before the sixth year, amounts to 0-31—0-33™™. As regards the change in the definitive constant length of the elements at different levels of the stem, only one investigation, on @ szmgle Birch-stem, exists, which may be referred to in Sanio’s treatise. The increase of the ligneous elements in length and width is, in some of the cases observed, the immediate consequence of a corresponding successive enlarge- ment of the cambial cells; while in others this is not so. The former is the case in the Coniferze, where both the successive enlargement of the cambial cells, and the relatively trifling increase of size, which the ligneous elements undergo after their origination, are distinct. In the Pine, for example, the tangential diameter of the winter cambial cells is more than twice as great (o-026™™) in a stem a hundred years old as in a shoot one year old (o-or2™™); in a one-year-old apical shoot of the same species, on the other hand, the length of the cambial cells is o-87™™, that of the tracheides in the autumn-wood 105mm, The other category includes, for example, Rhamnus Frangula, Cytisus Laburnum, Caragana arborescens, and probably most Leguminose. As far as the investigations extend the cambial cells here undergo little or no increase of size, as their distance from the pith becomes greater. The increase in size is thus due to the ligneous elements themselves, which have already been developed from the cambium; the woody fibres, which are especially affected by it, become in Cytisus Laburnum, for example, six times as long as the cells of the cambium. Srct. 154. In many woody plants there further arise those differences of physical properties between successive zones of growth or annual rings, on which is founded the technical distinction between sap-zwood (Splint, Alburnum, Aubier) and duramen or rzpe-wood (Kern, Herz, bois parfait). The sap-wood is wood which has been completely developed from the cambium, and which possesses the anatomical characteristics above described, and the physiological properties which depend on them. It has a light whitish or yellowish wood-colour. In many trees, as for example Acer pseudoplatanus, platanoides, and Buxus, the alburnum condition 508 SECONDARY CHANGES, undergoes, according to Nérdlinger, no change, at least -as regards the outward appearance and coarser physical characteristics of the wood. These are termed alburnum trees by Nérdlinger'. In most woods changes of the chemical? and physical character, and in a less degree of the structure also, take place sooner or later with the increasing age of the successive zones. These changes represent the beginning of that process of degradation which ends with decomposition, and as they proceed the wood ceases to be available for its original physiological work’, with the exception of the purely mechanical function in the case of those woods which become harder. Externally a darker colour appears, which varies in the different kinds of wood, and may become deep black, as in the Ebony woods, dark- green as in Guaiacum wood, or red and violet as in the dye-woods of the Czesalpiniz, Pterocarpus, Hzematoxylon, &c. With these changes is very often connected an increase of the specific weight and hardness, and a diminution of the contained water*; in a word the technical value of the wood is created or enhanced by these changes. In this case especially the terms ‘ ripe-wood’ and ‘ heart-wood’ are used ;— the two names being applied rather arbitrarily to the particular cases, or used for successive stages. The formation of hard and lasting duramen is however only a special case of the incipient process of degradation. The latter may also lead quickly to the opposite result. According to Noérdlinger', if I understand him rightly, ‘many soft woods, e.g. Canadian poplar, and several kinds of Willow,’ have a brown heart-wood, which is not distinguished from the alburnum either by higher specific gravity, or by hardness and durability, but, on the other hand, shows a tendency to rapid decom- position, accompanied by the growth of mould. The complete disorganization of the wood, which in many cases takes place in old living trees, is an allied phenomenon which will be discussed below. Apart from these changes, anatomical investigation does not demonstrate any alteration in the original structure and the original thickness of wall of the cells and tubes, on the formation of the heart-wood, but merely changes in the material properties of the walls, and in the contents. The membranes are infiltrated by heterogeneous organic bodies, and the latter often appear in the interior of the elements, which they fill up more or less completely, as well as in any cracks and crevices that may be present. The colours of heart-woods are those of the in- filtrated substances. Qualitatively, the latter are usually combustible organic compounds, showing extreme variety in detail, if all the cases be taken into consideration. While referring to the technical literature °, it is here sufficient to call to mind the dyeing substances and chromogenetic bodies of the dye woods, and the infiltrations of resin in the wood of many Conifer, Guaiacum, &c. As regards these substances, it is by no * Technische Eigensch. d. Hélzer, p. 28, &&c. > Compare the summary in Hofmeister, Pflanzenzelle, p. 247, and the technical literature. * Rossmissler, Tharander Jahrb, IV. p. 186 (according to Nordlinger), * On these subjects, which do not concern us further, compare Nordlinger, Z.c.,and Wiesner, Rohstoffe, cap. 13. 5 Le. p. 36. ° Compare the summary in Wiesner, /. ¢. « SECONDARY THICKENING. NORMAL DICOTYLEDONS. 509 means to be asserted that in the woods in question they alone saturate the heart- wood, and may not perhaps be mixed with other bodies which have not been minutely investigated, or that differences may not prevail between the mixtures which occur in the membranes and those in the cavities. On these questions no sufficiently accurate investigations exist. In a number of Dicotyledonous woods, which do not serve as dye-woods (Ailantus, Prunus domestica, spinosa, Amygdalus communis, Zanthoxylon fraxineum, Rhamnus cathartica, Sorbus Aucuparia, Gle- ditschia, Periploca, &c.), bodies were found by Sanio? both in the interior of the vessels and in the membranes, which on their first appearance in the interior of the vessels were colourless, but afterwards came to differ from one another in their yellow or red colour, while they agreed together in their high resistance to all solvent agents. Caustic potash does not change them; Schulze’s mixture at boiling- point first produces discolouration, and then solution. It is thus quite possible that the appearance of a definite body, or of a series of substances, which are nearly related, and vary in the different kinds of wood, may be generally characteristic of the process of metacrasis which produces the duramen, and that the appearance of resins, of special colouring matters, and so on, may only be a phenomenon peculiar to particular cases, and accompanying the former changes. In Caragana arborescens Sanio found that only air is at first contained in the vessels of the yellow heart-wood; at a somewhat later stage he found a yellow body with the properties above-stated; in the discoloured central portion, which is limited by a red ring, and even in the red ring itself this body had again disappeared. This phenomenon can afford no argument against the view that the appearance of the infiltrating substances in question is characteristic of the formation of dura- men, but can only indicate that in certain cases they may undergo displacement, or further processes of rapid decomposition. Inside the vessels and cell-lumina the infiltrating bodies of whatever kind appear at first in the form of deposits on the wall, often in two layers; here and there, where they are accumulated in large quantities, they show hemispherical prominences, projecting inwards, or sometimes form biconcave plates extending transversely through the lumen; the masses are homogeneous, rarely granular (Castanea vesca), frequently with flaws and cracks. All these phenomena point to the fact that they first appear in the fluid form, and afterwards harden. Very extensive accumulations fill the lumina completely, as for example in the Guaiacum and Ebony woods. In the vessels and crevices of the Campeche wood, greenish crystals (Hematoxylin ?) sometimes occur *. It is remote from the purpose of this book to discuss the question, so difficult in the defective state of chemical knowledge on these subjects, as to the origin of the infiltrating substances, and more especially to investigate how far they are derived from the transformation of other pre-existing substances, at the same places in which they occur, to what extent they have been conveyed to these places from elsewhere, and in the latter cases what their origin is. Attention may 1 Botan. Zeitg. 1863, p. 126.—Compare also Hartig, ibid. 1859, p. 100. ? Fiiickiger and Hanbury, Pharmacographia, p, 188. 510 SECONDARY CHANGES, ‘here only be briefly called to the fact, that the anatomical conditions already adduced point to the occurrence of somewhat complicated processes, and that the expressions just used are not intended to prejudge any developmental theory whatever. As a last stage of the changes described, and one which leads to a result opposite to the formation of heart-wood in the technical sense, a displacement of the normal membranes (extending to their complete disappearance) by masses of resin and balsam is stated to occur locally in the woods which produce these substances; in other woods, which do not form resin, a transformation into dis- organised masses of mucilage and gum takes place. The wood of Pinus Strobus and Abies pectinata, which is infiltrated with balsam, dissolves away, according to Wigand ', into a resinous mass, and as this goes on, the walls of the tracheides and cells diminish in thickness, becoming lost eventually in the structureless mass of resin. The canals, an inch in width, filled with balsam, which, according to Karsten ?, traverse the wood of species of Copaifera, and similar phenomena reported in the case of the stem of Drybalanops aromatica® can also, according to the data before us, scarcely arise otherwise than by partial disorganisation and solution of the ligneous elements. The formation of cherry-gum by disorganisation of the wood of the Amygdalezx, as described by Wigand, which no doubt belongs to a great extent to the province of pathology, starts sometimes from vessels, sometimes from groups of wood- parenchyma, or medullary spots. Comp. Wigand, 2. ¢. In addition to the organic infiltrated substances, considerable accumulations of Silicic acid occur, as found by Criiger* in the old wood of plants which are charac- terised by extensive silicification of almost all their parts, namely of the Chrysobalanee, Hirtella silicea, Petraea volubilis, P. arborea, and of Tectona grandis, They occur in the Iumina of the cells and the vessels (the latter exclusively in Teak wood) as amorphous masses, filling up the space more or less completely. According to the published statements they are not present in the alburnum. The data before us do not allow of any decided judgment as to further differences between alburnum and duramen, as regards the incombustible constituents which they contain. In the contents of the ce//s of the wood, both in the medullary rays and in the ligneous bundles, and also of the thyloses where they occur, a further essential change takes place on the formation of the heart-wood, in addition to those described. This consists in the permanent disappearance of those constituents which characterise the living cell, especially of the universally distributed starch, and in their replacement by air or infiltrations®. The disappearance of the starch from the cells coincides fairly exactly, in the cases investigated, with the appearance of the other characters of the heart-wood. * Pringsheim’s Jahrb. III. p.165.—On the other hand compare also Dippel, Botan. Zeitg. 1863, p. 256. ? Botan, Zeitg. 1857, p. 316. * Compare Fliickiger and Hanbury, Pharmacographia, p. 202. * Botan, Zeitg. 1857, p. 297. ® Sanio, Starkefiihrende Zellen, p. 19 (1858).—Id. Botan. Zeitg, 1860, p. 202.—A. Gris, Comptes Rendus, 1866, tom. 70, p. 603. SECONDARY THICKENING. NORMAL DICOTYLEDONS. 51r As has often been indicated in the preceding pages, the transformation or degradation of the alburnum into duramen takes place in different species of trees at a different average age, and in some gradually, in others suddenly. In the stem of Fraxinus excelsior when forty years old, and in that of the Birch when thirty- five years old, abundance of starch was found by Gris in a/7 the annual layers; the former species is, according to Nérdlinger, characterised by a very broad alburnum, while Betula alba is an ‘alburnum tree.’ In a stem of Fagus sylvatica ninety-five years old, investigated by Gris, there was abundant starch in the fifteen outermost annual rings, then a gradual decrease in its amount from these up to the thirty-fifth ring, while further inside it was wholly absent. According to Nérdlinger, Quercus pedunculata has 8-13 annual layers of alburnum; in an oak-stem fifty-eight years old, Gris found that the cells ceased to contain starch somewhat suddenly at the sixteenth ring, and in one ninety-eight years old, at the twenty-first ring. Robinia Pseudacacia has 3-5 rings of alburnum, which are sharply marked off from the dark heart-wood, both by their containing starch, and by their light colour: Castanea vesca behaves in a similar way (Gris and Nérdlinger). According to statements based on those coarser differences which are of technical importance, the diversity in this respect is very great. It is equally evident from the facts already mentioned, that in the same species of tree, and even in the same individual, stem, or branch, numerous individual variations occur, within definite specific limits, according to the age and vigour of development, especially of wood-formation, and at different levels of the stem. This is the case whether the relative thickness of alburnum and duramen be deter- mined according to the number of annual rings, or according to absolute measure. Even on different sides of the same cross-section the number of annual rings showing the properties of alburnum is often different *. What has been hitherto said applies to wood enclosed in the uninjured stem. It is well known that changes not unfrequently occur in the wood in consequence of injuries, and that the extent of these changes, which are similar to those characterising the formation of heart-wood, shows a definite relation to the position and extent of the wound. The processes of decomposition to which they owe their origin may often be different to those which accompany the formation of duramen?, On the other hand, the phenomena occurring in both cases are often so similar—as for example the conversion into resin in the Conifer, and the formation of black, hard wood in the Ebenacee—that we may suppose the same process, which usually comes on more slowly, to be accelerated or excited by injury. e. ILndtoidual and local deviations. Sect. 155. Within the limits of the typical characteristics, which have been explained in the preceding pages, the structure of the annual ring in the same species shows definite variations, according to differences, however produced, in the vigour of its development; the single feebly developed ring, or even portion of a 1 Compare especially the detailed statements respecting the Oak in Duhamel, Physique des Arbres, I. p. 46, &c. ? Compare e.g. Noérdlinger, Zc. p. 37. 512 SECONDARY CHANGES, ring, of a tree which has otherwise grown vigorously, shows essentially the same character as the whole of the rings of a feebly thickened tree. This character is different in Coniferous woods, and in the Dicotyledonous woods investigated, In the former, according to Mohl, the relative thickness of the spring-wood with wide lumina, and the autumn-wood with narrow lumina, and the more or less sudden transition from one to the other, varies quite universally, according to the thickness of the rings, the variation in the stem being, as a general rule, in the opposite direction to that in the root, subject to specific or perhaps individual modifications’. In the former, the outer zone of the ring, with narrow lumina and thick walls, forms a larger portion of the whole, and is the less sharply marked off from the interior the thinner the ring is. In the root it is the more developed the thicker the ring; in the feeble annual rings which are prevalent in roots, and which in the White Fir (Abies pectinata), for example, are on the average only about o-2mm broad, it often consists of only 3-1 layers, and is sharply limited towards the internal zone, which has wide cavities. In the Dicotyledonous woods investigated (Fraxinus, Fagus, Quercus peduncu- lata, Morus, Broussonetia, Rhus, Sophora, Gymnocladus, etc.)*, the middle portion of the ring becomes reduced, as the thickness of the whole diminishes, in such a degree that in extreme cases the ring consists exclusively of spring-wood and the autumnal limiting layer. This condition appears most sharply in Morus, Rhus, and the Leguminosz mentioned, where the woody fibres are confined to the middle region in well-developed rings, but wholly disappear in feeble ones. In this respect the wood of stem and root shows a general agreement, and as the annual rings, in roots which have once attained a thickness of two or three inches, are, as a rule, extremely thin (o-25™™ and less in thickness), it follows at once that there is a considerable difference in the general structure of the wood of stem and root, which is increased by further differences in the structure and distribution of the elements, which will be described below. ‘ As in Coniferous woods the outer portion of the annual ring contains relatively the largest mass of lignified membranes, and consequently possesses the greatest density, strength, and hardness, while in Dicotyledonous woods, owing to the structure and distribution of their elements, this is the case with the middle parts of the ring, the lesser density and strength of the wood of the root, as compared with that of the stem, is a necessary consequence of the facts above stated. The stem-wood, however, as technically made use of, when taken from well-grown trees, for which the excep- tions to be mentioned below do not hold good, diminishes in the case of the Conifere in density and strength as the thickness of the annual rings increases, while in Dicotyledonous woods the reverse is the case. These rules, however, undergo a considerable modification in the case of stems, owing to local changes in the relative breadth of the autumnal wood at different levels on the stem, as Sanio” found in some well-grown main axes of Pinus sylvestris. In these cases the relative breadth of the autumnal wood increases in 1 Compare von Mohl, Z.c. p. 238. ? Compare von Mohl, Botan. Zeitg. 1862, /.¢.—Sanio, ihid. 1863, p. 397. 8 Pringsheim’s Jahrb, IX. p. 115. SECONDARY THICKENING, NORMAL DICOTYLEDONS. 513 every annual ring, independently of its total breadth, from apex to base, so that, for example, the proportion between autumn- and spring-wood in the same three outermost annual rings amounts to 1: 10-6 at a height of 27™, and to 1: 2-5 ata height of 1™, The strength and technical usefulness of the wood at different levels on the stem is thus very unequal. According to the statements of technical authorities, who value the wood of the upper trunk of Dicotyledons and Conifers less than that of the lower, it may be conjectured that similar conditions to those found by Sanio in the Pine are of general occurrence in the stems of trees; yet these differences in technical usefulness might also have other causes, e.g. differences in the formation of heart-wood, or the causes might be various in the various cases. More extended investigations are still required before any general rules or laws can be laid down. Sect. 156. The indistinctness of the limits between the annual rings, which in many woods is typical (comp. p. 503), may also occur in those with a typically sharp demarcation, as an individual phenomenon, especially where the rings are slightly developed ; in cases of eccentric ring-formation this often takes place in such a way that two distinct rings on the stronger side run together into a single one on the weaker. Such an obliteration of the boundaries of rings has often been observed, in the case of Dicotyledonous and Coniferous trees, both in the wood of the stem, and more especially in that of roots’, Both in the cases of unilateral coalescence of otherwise distinct rings, and in those where the number of the existing complete rings is less than the known age of the wood in years, a distinction must be drawn between the partial or total suppression of the grow/sh in thickness during a period of vegetation, and the suppression of the demarcation of rings where growth takes place. Both cases may occur, the former, for example, has been demonstrated in stunted trees*; when a case comes under investigation, it can, as a rule, scarcely be deter- mined subsequently which of the two phenomena has taken place, As already indicated, the Araucarias appear, according to the existing data and controversies‘, to have a special inclination to the individual differences in question during the occurrence of growth in thickness, Schacht denies the demarcation of the annual rings in A. brasiliensis, even in the face of Goppert’s reply. Kraus, on the other hand, describes two pieces of the stem of the same species, one of which had concentric zones, but not a trace of an annual boundary, while the other had eleven annual rings, characterised by sharply defined autumnal and spring-wood. In a portion of the stem of a well-grown specimen of A. excelsa (cultivated in the open ground) I found sixteen rings, which were so sharply marked when seen with the naked eye that one is astonished to hear of its being necessary to seek them with high powers in thin sections. This is explained by the fact that every ring consists principally of thick-walled, tolerably uniform tracheides, and only contains a narrow zone of more thin-walled elements (spring-wood) at the boundary of the next inner ring. These are but little wider than the thick-walled ones, a sharp boundary between the two being as little observable as a . -distinct flattening of the latter. For these reasons the limit between the rings is by no 1 Compare Noérdlinger, Techn. Eigensch. d. Hélzer, p. 130. ? Compare T. Hartig, Forstl. Culturpfl. p. 86.—Von Mohl, Botan. Zeitg. 1862, p. 228, &c.— Kraus, Bau d. Nadelhdlzer, /.¢. p. 146.—Nordlinger, Der Holzring, p. 21. 3 Compare R. Hartig, Botan. Zeitg. 1870, p. 527. * On these compare Schacht, Goppert, Botan. Zeitg. 1862, and Kraus, /.c, ul 514 SECONDARY CHANGES. means conspicuous under microscopical investigation, while the middle of the thin-walled zone appears to the naked eye as a sharp line of demarcation. Possibly the same conditions may exist in the stems of A. brasiliensis, which have been the subject of controversy. A branch of the same specimen of A. excelsa, about 2°™ thick, with narrow rings, shows the same characteristics in some parts, while in others there is a decided flattening of the tracheides at the autumnal limit. The question whether the formation of two successive rings in one period of vegeta- tion occurs as an individual deviation from the rule, even in our woody plants which typically form a single ring, is essentially foreign to the present anatomical survey. It must be remarked, however, that this phenomenon is stated to occur as a rare exception, and in fact as a consequence of the interruption of the summer growth by external causes (frost, drought, attacks of insects, hailstorms 1, &c.), As regards the anatomical characteristics of these anomalous double annual rings, no other statements have been published, than that the boundary between them is usually indistinct: only in the case of shoots of Sambucus nigra, interrupted by a hailstorm in very vigorous vegetation in 1846, does Unger mention ‘two distinct woody rings,’ Sect. 157. The existing investigations show that the structure of most woods remains essentially constant, within the limits defined by the preceding statements, But exceptions even to this rule occur, the most remarkable of which are afforded by the Ash (Fraxinus excelsior?). In a well-grown tree of this species, the annual ring, which is about 2~3™™ in breadth, shows on the inside a zone of spring-wood, consisting of slightly thickened fibres, between which wide vessels surrounded by bundle-parenchyma are inserted; then follows externally the thick middle layer, consisting of thicker-walled fibres, with scattered smaller vessels, likewise surrounded by bundle-parenchyma ; finally, at the outside, there is the autumnal limiting layer, consisting of several rows of bundle-parenchyma, with small, very thick-walled vessels, In very thin annual rings, the reduction of the middle layer takes place as described above. In the case of very luxuriant young trees, grown on damp soil, with annual rings over 12™™ thick, V. Mohl found the fibres less thick-walled, and the vessels, especially the large ones, narrower than in moderately thick rings. Sanio found a specimen which differed conspicuously from the usual form, in having concentric zones of parenchyma, containing narrow vessels in the middle layer ; in one piece only there was an annual ring similar to the usual type. Another stem, which was stunted, and at an age of fourteen years was only 15™™ thick, showed a strikingly feeble development, and in the thinnest rings actual suppression, of the charac- teristic spring layer, in contradistinction to the rule holding good for narrow annual rings, while the vessels were everywhere remarkably narrow. The mean width of the largest of the latter in one ring was o-o7™m, that of the large vessels in V. Mohl's broad-ringed stems was 0-17™™, while in normally grown trees it is about o-26™m,— Sparmannia africana, as more minutely described by Sanio, Lc. P- 399, shows a conspicuously different structure, even in successive transverse portions of the same stem, or on different sides of the same ring, broad bands of irregular large-celled parenchyma being sometimes present, sometimes absent. ‘ Unger, Botan, Zeitg. 1847, p, 265.—Nérdlinger, Holzring, p. 10, *? Von Mohl, /.c. p, 269.—Sanio, Botan, Zeitg. 1863, p- 398. SECONDARY THICKENING. NORMAL DICOTYLEDONS. 515 f. Differences of the secondary wood in non-equivalent members of the same plant. Sect. 158. The ligneous body of the stem and its branches has, in the plants in question, the same structure, within the limits of deviation defined in the preceding pages. In trees, however, differences in dimensions exist, owing to the fact that not only the thickness of the annual rings, but also the size of the tissue-elements, is less in the branches than in the stem. This at least is the case according to the more accurate investigations before us, which were carried out on Coniferous trees, “and are mentioned at p. 506?. A far less general agreement prevails between the special structure of the ligneous body in the stem and its branches on the one hand, and in the roots of the same plant on the other. Here, on the contrary, there are two different extreme cases; plants with the wood of the root completely similar to that of the stem, and others with the opposite character; between the two extremes there are of course many intermediate cases. The first of the two cases occurs in Gymnospermous and Dicotyledonous trees and shrubs. Although the wood of their roots is never quite similar to that of the stem in form, structure, and distribution of the tissue-elements, yet the differences only affect relative dimensions and subordinate variations of structure”. Differences of the former class consist, firstly, in a considerable reduction, in the case of the root, of the average thickness of the entire annual ring, which, though subject to great varia- tions, may sink to minimal dimensions, in the White Fir, for example, to o-117™™, although, on the other hand, it sometimes reaches 2-3™™; in Dicotyledonous woods it may even be smaller than the average diameter of the vessels present in the annual ring, in which case the ring must have an undulating outline, widened at the vessels. Secondly, there are differences in the width and thickness of wall of equivalent portions of tissue. These affect details in the structure of the walls, and, in the case of Dicotyledonous woods, the distribution of the non-equivalent forms of tissue in the annual rings. These relations once more appear most clearly and simply in the Coniferous woods. According to V.Mohl’s measurements, the tracheides of the spring-wood in the root of the White Fir are distinguished from those of the stem by this fact, that their radial diameter is on the average 4, and their tangential diameter 4 greater, while their length is also greater; those of the autumn-wood by their greater radial diameter and wider lumen. To this is often added a diminished absolute thickness of wall in the tracheides of the root, and, Zerhaps independently of this, a greater softness of the wood; further, the fact that the development of the thicker-walled autumn wood-cells diminishes with the thickness of the annual rings, and in very thin ones is almost suppressed. Similar relations, to be compared in V. Mohl and Schacht, Z.c., reappear in other Abietineze. The differences between the wood of the root and that of the branches, which Schacht compared with it, are as regards the width of the tracheides even greater than in the case of the stem-wood, for reasons stated above. In the wood of the root in Coniferze, the pits on the radial sides of the tracheides are often arranged in two longitudinal rows, whether the side 1 Compare also von Mohl, Botan. Zeitg. 1862, p. 461.—Schacht, ibid. p. 409, &o. * Von Mohl, Botan, Zeitg. 1862, pp. 225, 269. Ll2 516 . SECONDARY CHANGES. be in contact with one neighbouring tracheide, or with two (as is the case where the elements form alternate rows), while in the stem the presence of a single row of pits is the rule. (Comp. p. 494): : The wood of the root in Dicotyledonous woods is generally also distinguished from that of the stem by its greater ‘porosity’ and softness. On the one hand, this character is closely related to the changes in the structure of the annual rings which accompany their decrease in thickness, as described above at p. 512. As in the woods investi- gated the middle strong part of the rings is reduced in the slightly developed ones ta such an extent that it may even be entirely absent, and these rings thus consist principally of the wide and relatively thin-walled vessels of the spring-boundary, the distinction indicated must necessarily result., In themselves the large vessels of the wood of the root are inferior in average width to those of the stem-wood, in the case of the Ash and Oak. In the Beech, on the other hand, and in a less degree in the Birch and Aspen also, the average width of the internal vessels, even in relatively thick annual rings, is greater than in the stem. This increase in the relative extent of the total area of the cavities of the vessels is, according to V. Mohl, the only anatomical cause of the greater porosity of the wood of the root in the Beech and Aspen. Further, in other cases a more or less considerable increase in the width of tracheides and cells is found (amounting to 3 in Berberis), and a corresponding decrease in the average thickness of their walls. Besides Berberis, this is the case in Fraxinus, Betula, and Quercus. Sect. 189. The second of the two extreme cases distinguished within the general plan of structure, namely, an extremely different anatomical composition of the wood of the root, as compared with that of the stem and its branches, is of very general occurrence among herbaceous Dicotyledons, especially perennials and biennials, the roots of which store up reserve substances such as starch, inulin, &c., and apparently large quantities of water. The distinction is no doubt most strikingly marked in the fleshy tap-roots of cultivated plants, Brassica Rapa and Napus, Daucus, Raphanus, &c.; but these are only special cases of a phenomenon which is of very general occurrence. The most general anatomical character of these roots consists in the reduction of the specific woody elements as compared with the parenchyma, This is brought about in different ways :— (1) By feeble development of the entire ligneous body lying inside the cambium as compared with— a. The persistent parenchymatous primary outer cortex, or b. The relatively very thick, also chiefly parenchymatous, secondary bast. (2) By the development of a relatively small quantity of specific woody elements, i.e, vessels and fibres, in the ligneous body, which as a whole is strongly developed.» As examples of (1) a.-the annual subsidiary roots, approaching 2™™ in thickness, of some ‘ Asclepiadez are to be mentioned. In the shrubby Asclepias curassavica the wood of these roots is more than 1™™ thick, cylindrical, and similar in structure to the wood of the stem. In the thick roots of the rhizome of A. Cornuti and Vincetoxicum officinale _ the greatest diameter of the original diarch xylem-plate is less than 0-3™™ ; the breadth of the secondary mass of wood attached to it is only half as much; all the rest, except the slight zone of bast, is primary cortical parenchyma. The subsidiary roots of the Piperacez further belong to this category, as also do those mentioned at p. 355, which show no thickening of their vascular bundles, or mere indications of it, SECONDARY THICKENING. NORMAL DICOTYLEDONS. 517 The far more frequent case mentioned under (1) 4., to which wé shall have to return in considering the changes of bast and cortex, occurs, for example, in the root of Taraxacum, Rubia, and Umbellifere. A root of Taraxacum which is now before me, 4™m in thickness, has, for example, a ligneous body only about 0-5™™ jn diameter. Lastly, case (2), which especially belongs to the present subject, occurs in its most typical form in Brassica and Raphanus. The main mass of the Radish and the Turnip is formed of the chiefly parenchymatous wood; the bast and external cortex are not more than 1-2™™ thick. Between the typical cases. mentioned under (1) 4, and (2), a number of intermediate forms occur, with but slight difference between the mass of the wood and that of the bast and external cortex, e.g. roots of the Umbelliferee, Scorzonera hispanica, Rheum Rhaponticum, &c. But in these cases, as the relative thickness of the wood increases, the proportion of parenchyma to the specific woody elements increases in favour of the former, if it is allowable to enunciate a general rule for cases which show such great variety in detail. In the roots in question, the tracheze of the wood, in all known cases, are exclusively vessels, with a wall which is reticulately thickened (and then often with ‘scalariform transverse meshes), or has bordered’ pits, the latter structure being not uncommonly found on the surfaces which abut on other vessels, the former on those adjoining non-equivalent tissue. Their average width is considerable; wider and much natrower ones frequently occur together. They are always immediately accompanied by rows of elongated prismatic cells with pointed or horizontal ends, which may be called fibrous cells in the sense defined above, and the nature of the contents of which still requires more accurate comparative investigation; they are further accompanied by short-celled parenchyma, which in form and position corre- sponds to the bundle-parenchyma, Tracheides do not seem to occur, yet this point also still needs further investigation. Between the ligneous bundles, which are thus constructed according to the general rule, medullary rays of different orders are inserted as the wood becomes more strongly developed, in a manner which likewise corresponds to the general rule. So far as is known these are always parenchy- matous, and their elements are generally to be distinguished from those of the bundle-parenchyma by their form and position, and the special nature of their contents, according to the rules described above; but in the particular cases the differences may be either very clear or not distinct. Leaving out of consideration the cases marked (1) a. where the ligneous body is quite small, the formation of massive parenchyma is apportioned between the ligneous bundle and the medullary rays in two principal forms, between which of course there are again many intermediate cases. (x) Natwow ligneous bundles are separated or divided up by broad parenchy- matous medullary rays. They consist principally of vessels and fibrous cells, the latter being usually narrow, thick-walled, and lignified; the medullary rays are thick masses of parenchyma, their cell-walls for the most part thin and not lignified.. The above-mentioned cases, with strongly developed main medullary rays, belong to this category, as Urtica, Cucurbita, Symphytum officinale, &c. Comp. p. 474, Figs. 203, 204. (2) In most really fleshy roots the main mass of the parenchyma in the ligneous body belongs to the ligneous bundle itself. In its most internal portion, bordering 518 SECONDARY CHANGES, on the primary xylem-plates, the latter consists of vessels lying somewhat near together, and only separated by narrow one- or few-layered bands of usually non- lignified parenchymatous or fibrous cells. If, as is usually the case here, the main medullary rays are absent, one can scarcely speak of medullary rays at all, they are only indicated by single radial bands of parenchyma. In cases with a relatively small wood (Taraxacum, Rubia, &c. (1) 4.) this condition is permanent. Where, on the other hand, the wood is largely developed (2), as in Rheum, Scorzonera his- panica, Pastinaca, and the swollen roots of Brassica and Raphanus’, the formation of parenchyma in the ligneous bundle increases with the progressive growth in thickness. The bundle is principally composed of parenchymatous cells with non- lignified walls, decided longitudinal extension, and radial arrangement; and in this massive thin-walled parenchyma of the bundle lie the vessels, which form closely united groups, or are farely quite isolated, and are accompanied by narrow, usually non-lignified fibrous cells. As seen in transverse section they form interrupted radial rows, increasing in number in the centrifugal direction, and concentric zones which are also interrupted, while in their longitudinal course they form a net with pointed meshes. The medullary rays are inserted between the parenchymatous masses of wood. Their cells are in many cases distinguished from those of the ligneous bundle by their form, which is usually radially procumbent, and by differences in their contents. This is the case in Rheum, whete the procumbent cells of the numerous rays, which are only 1~3 cells in breadth and penerally only 6-10 cells in height, ‘are distinguished by containing an abundant yellow colouring matter (Chrysophanic acid) from the upright parenchymatous cells of the bundle, which chiefly contain starch; also in the cultivated root of the Parsnip, &c., where the uniseriate to triseriate cells of the medullary rays, densely filled with small starch-grains, contrast sharply with the narrower, elongated cells of the bundle, in which the ‘starch is less abundant. On the other hand, many of the cases referred to above, in which there is ho sharp boundary between the parenchyma of the rays and bundles, belong to this category. As séen in cross-section, the rays may indeed be distinguished in their middle part by the greater radial elongation of their cells, by their general course, &c.; but they pass over quite gradually into the adjacent parenchyma of the bundle. So, for example, in Scorzonera hispanica, Raphanus, Brassica, and the fleshy swollen roots of Daucus, The preceding short statements and examples are only intended to call attention to the most remarkable structural phenomena in the wood of fleshy roots. Reference may be made to the descriptions of officinal roots in the pharmacological literature (Wigand, Fliickiger, and Berg), and especially to the representations ir? Berg’s Atlas, for illustrations of special characteristics, which are extremely variable in different species, and of the no less varied intermediate forms between those ligneous masses which consist chiefly of parenchyma, and those which are, in various degrees, more woody, i.e. which agree more in structure with the wood of the stem. The variations in the structure of the wood, which may occur within the same species in different individuals, sometimes no doubt as an effect of external conditions, and in the same * Compare Nageli, Beitr. I. p. 25. SECONDARY THICKENING. NORMAL DICOTYLEDONS. 519 individual in roots of different order and thickness, are considerably greater in these cases than any similar phenomena which have been observed in the wood of the stem. The most conspicuous examples of this are once more afforded by plants which, in their wild form, have thin roots, but in many cultivated varieties are pro- vided with fleshy swollen roots, as species of Brassica, Raphanus, Daucus, &c. In the main root of the wild Daucus Carota, the ratio of the thickness of the wood to that of the surrounding cortex (bast-layer), expressed according to the radii of the transverse section, is 8:3. The somewhat firm wood consists in the bundles partly of narrow fibrous cells, which are at least 8—10 times as long as broad, and pointed at both ends, and are provided with a moderately thickened membrane with small pits, and partly of rather wide vessels, arranged in radial bands, the walls of which show almost exclusively transverse bordered pits. Between the bundles there are numerous medullary rays, consisting of one or more layers of large, approximately isodiametric, parenchymatous cells, Further details, which might be mentioned, especially as regards the innermost portion of the wood, may here be passed over as non-essential. In the cultivated yellow carrot, the proportion between the radius of the cross-section of the wood and that of the surrounding, chiefly parenchymatous, cortex (bast-layer), is approximately as 1:7. The vessels, at least in the great majority of cases, are reticulated, with transverse meshes, fibrous cells are wholly absent, and are replaced by wide, thin-walled, parenchymatous cells, which abut on one another with horizontal surfaces, and are on the average twice as long as they are broad. Although indications of medullary rays may be recognised, they are not sharply marked off from the parenchyma of the bundles. It seems to me remote from the purpose of this book to give a synopsis of all the woods investigated, which might serve as a key to their identification. Some assistance towards the latter object may be obtained from the preceding pages. Reference may further be made to the literature cited, especially that of Pharmacology; e.g. to Wiesner’s Rohstoffe des Pflanzenreichs, Hartig’s Forstl. Culturpflanzen, and his treatise, Zur vergl. Anat. der Holzpflanzen, Bot. Ztg. 1859, p. 93, and more especially to Sanio, Ueber die Zusammensetzung des Holzkérper’s, &c., Bot. Ztg. 1863, p. 401. Joseph Méller’s copious ‘ Beitrage zur vergleichenden Anatomie des Holzes,’ Vienna, 1876, could not be made use of for the present work. Ill. Tue Bastr?. Srct. 160. The cambial ring of Dicotyledons and Gymnosperms with normal growth produces on ‘its outer side the secondary layers of bast; these are added to the original bast-zone of the stem, which is represented by the phloem-portions of the vascular bundles. A similar process goes on in connection with the primary phloem- groups of the root, in the manner described above. The secondary zones are directly continuous with the original ones, and form, together with the latter, the entire zone or mass of bast. Its external limit is formed by that of the primary phloem-groups and of those portions of the medullary rays which lie between them. By means especially 1 [Compare Moller, Anatomie der Baumrinden, Berlin, 1882.] 520 SECONDARY CHANGES, of the former, it is sharply marked off from the non-equivalent tissues of the external cortex, more especially in those cases most frequent in stems, where the original groups of phloem -are supported or enclosed on the outside by sclerenchyma. Nageli* has called this external limiting zone of the bast-layer the cortical sheath, a term corre- sponding to that of medullary sheath, used for the internal boundary of the wood, - The original structure of this limiting zone is evident from the descriptions given in preceding paragraphs; the structure of the bast as a whole depends partly upon the former, partly upon that: of the secondary increment of growth to be described here. A further modification, however, results from the fact that the structure of the bast must undergo a constant change, so long as the volume of the body enclosed by it is increased by means of the activity of the cambium; for each zone thus under- goes, after its first formation, a constantly increasing extension in the direction of its surface, by which it must be in some way or other affected. So long as growth in thickness goes on, a constant transition must of necessity take place between the original conditions and those which have been modified by the peripheral extension, ‘and in fact the successive stages of change must always come under observation in every mass of bast which is regarded as a connected whole. This fact must always be borne in mind. For purposes of description, however, it is necessary to separate the initial structure from the changes due to peripheral extension. The former will be immediately considered here, the latter in the next chapter, The differentiation of the bast (Sect. 135) is similar to that of the wood, in so far as it consists of principal and partial strands of various degrees, which are separated from one another, or divided up by the large and small medullary rays (shortly termed dasfsrays). The equivalent rays and strands of wood and bast correspond, and fit on to one another in the cambial zone. The original form and ‘size of the medullary rays, and the consequent course of the strands, are the same as in the adjoining wood. Sect. 161. Among the forms of fssue, sieve-tubes and parenchyma are charac- teristic, without exception, of the secondary bast of Dicotyledons and Gymnosperms ‘with normal growth. They are, at least very frequently, accompanied by sacs con- ‘taining crystals (comp. p. 141); further, by sclerenchymatous elements, and especially ‘by elongated fibrous cells, the das¢-fidres ; not uncommonly also by short sclerenchyma, ‘slone-sclerenchyma (stone-cells) ; while, finally, laticiferous tubes and secretory reservoirs characterise the bast of certain species or families. As has already been partly shown by the descriptions of the forms of tissue given in former chapters, especially Chap. V., all these elements in the bast, with the obvious exception of the scleren- ‘chyma, possess delicate, non-lignified, soft walls. Nageli has accordingly introduced the collective term soft dast for all those portions of the bast which are not Sclerenchymatous. In most cases the elements of the soft bast are originally narrow, and continue to resemble the cells of the cambium, from which they are derived; they are often difficult to distinguish from the cambium, and from one ‘another, especially as seen in transverse section. For this reason, but to a much greater extent owing to the fact that the softness of the tissues makes it somewhat ‘difficult to obtain good preparations, the structure of the bast, and its distinction from ' Dickenwachsthum, &c., d. Sapindaceen, p. 13. SECONDARY THICKENING, NORMAL DICOTYLEDONS, 521 the cambium, remained for a long time extremely obscure, and the accurate repre- sentations given by Th. Hartig as early as 1837 failed to be understood, until Mohl, in 18557, brought them into deserved honour. For the same reasons later investi- gations often leave much to be desired, and the special anatomy of the bast is but insufficiently treated of by most authors. Sect. 162, The main fundamental mass of the medullary rays always consists of parenchyma. The form and arrangement of its cells are identical with, or very ‘similar to those of the adjacent wood. It likewise occurs in the strands as a constant constituent, and, like the parenchyma of the wood, it is usually derived from single or repeated transverse divisions of the tissue-mother-cells, in those portions of the cambium which correspond to the strands, and its original arrangement agrees Fig. 210. Fig. 211, FIG. 210,—Cytisus Laburnum, tangential longitudinal section through the innermost layer of bast of the same branch as Fig. 198, under the same magnifying power as the Jatter ; s members of sieve-tubes, ¢a sieve-plate lying deeper than the surface of section, # small medullary ray, two cells in height. The remaining elements are cells of the bast-parenchyma, the origin of which from the transverse division of cambial cells becomes clear on comparison with Fig, 198. FIG. 211.—Juniperus communis, small stem. Transverse section through the autumnal wood, bast, and cambium, during the winter's rest (end of September) ; #—A external rows of the autumnal wood, 4, 4 series of bast-fibres, At « there is only one cambial cell between # and 4; #—mz medullary rays. with this mode of development (Fig. 210, comp. also Fig. 198, p. 465); more rarely it arises without transverse division, from longitudinal division only of the ‘tissue-mother-cells, and then corresponds to the intermediate cells of the wood. The Sveve-‘ubes (Chap. V.) are constant, specific constituents of the strands in ‘the normal soft bast of Dicotyledons. They are always accompanied by parenchyma, and in most woody plants are in general so arranged that the sieve-tubes form single, biseriate, or pluriseriate, tangential rows, which may be interrupted by parenchyma, and alternate with tangential rows of the same. The original radial 1 Compare note on p. 172. “ 522 SECONDARY CHANGES. arrangement of the secondary elements often continues to be maintained here, or is at least recognisable; hence, in every radial row derived from the cambium, one or more sieve-tubes always alternate more or less regularly with parenchyma. This arrangement appears with quite diagrammatic regularity in the bast of the Cupressinee and many Taxinee*, The transverse section of the bast (Fig. 211) here shows regular rows, both in the radial and tangential direction. Every fourth tangential row consists of fibres; of the three which lie between two fibrous rows, the middle one is parenchymatous, while the outer and inner each form an inter- rupted layer of sieve-tubes. The parenchymatous cells are approximately similar in width to the sieve-tubes (Juniperus communis), or are wider (e.g. Thuja occidentalis). In the stem of species of Pinus (P. Strobus, nigricans, silvestris), and also of Abies pectinata, irregular tangential rows of wide parenchymatous cells alternate with a FIG, 212,—Tilia argentea. Transverse section through the inner bast, cambium, and autumnal boundary of the wood ofa branch seven years old (cut in November) (220). The wood is drawn without the details of structure of the inembranes; 4 external limit of the autumnal wood, which is sharply limited by the dark outlines of its tangentially flattened elements ; c—c cambium and zone of young secondary growth ; 27—22 small medullary rays; in the one to the right are three fibres (/); 4 crystal-containing sacs,-with crystals, some of which were broken by the razor. In the bast-strand, between the two rays, three fibrous bands (7) are drawn ; alternating with them is the soft bast, consisting of sieve-tubes s, of gr ly dotted ( i ) cells, ining abund: starch and protoplasm, and of other, somewhat wider elements, bordering on the fibres, distinguished by clear watery contents and pitted walls, multiseriate zones of radially arranged sieve-tubes*% In the-old root, but not in the stem of the White Fir, I often find, between two radial rows of sieve-tubes, radial bands of parenchyma, which are uniseriate and resembie the medullary rays, but do not lie in the same straight line with the medullary rays of the wood. * Hartig, Forstl. Culturpf. p. 95, Taf. 9,10.—Von Mohl, /.c. p, 891,—Graf zu Solms-Laubach, Botan. Zeitg. 1871. * Von Mohl, /. ¢—Hartig, /.¢. Pp: 13, 35, Taf. 5. - SECONDARY THICKENING, NORMAL DICOTYLEDONS., 523 In the bast of Dicotyledonous woody plants, the arrangement of the two kinds of tissue, so far as can be decided from the existing data, is less regular than in the Coniferous woods first mentioned, owing to the fact that the tangential series of the one form of tissue are sometimes single, sometimes double or multiple, and are interrupted by interpolated elements of the other form; while the average width of the adjoining tissue-elements of the two kinds not uncommonly shows considerable differences, which are usually in favour of the sieve-tubes, e. g. Tilia and Vitis ; more rarely in favour of the parenchymatous cells. The narrower parenchymatous cells which accompany the sieve-tubes here show the same arrangement and the same characteristics as the camdiform cells de- scribed at p. 324, in the case of the primary vascular bundles; they should therefore be designated by the same name. They appear in transverse section as narrow, three or four-cornered meshes contiguous with the tubes, sometimes on one side of them, sometimes on more than one, but never (?) on all sides, each side, however, never having more than one (Fig. 212). Traced longitudinally (Fig. 213) they form, in most cases at any rate, series, each member of which is many times shorter than the adjoining member of the sieve-tube, and is derived from the trans- verse division of the tissue mother-cell. In Tilia I rarely found them equal in length to the members of sieve-tubes. The fre- -quency of the narrow cambiform cells seems to be very unequal in the different wy particular cases; for example, I find few Ee HO ao of them in Pyrus ~and Spirea aulmifolia, FIG. 213.—Vitis vinifera ; bast of a branch several years old, rem. in thick in (beginning of July). T: if while they are numerous in Tilia. "More ~section (145). s,s sieve-tubes, the inclined scalariform terminal ‘5 ~surfacespand one which is horizontal, being cut through longi- accurate details are to be expected from _ tudinally,-with the #xception of one.at the upper edge, which is e 5 is seen obliquely in superficial view. #, 2% medullary rays;onthe further investi gations. boundary between these and the strand of sieve-tubes are sacs “ , at 3 containing crystals, In spite of the irregularities described, the radial and tangential seriation of the elements continues to be maintained in many cases in its principal outlines, and each tangential row contains for the most part either sieve-tubes or parenchyma, as is especially shown by longitudinal sections which follow its course. The most various plants, e.g. besides those mentioned, Populus, Salix, Punica, Ficus, Sambucus, Fagus (Mohl), Aisculus and Ribes, show the greatest agreement in /A7s respect, although great differences occur between them in the average size of the elements and the special form of 524 . SECONDARY CHANGES. the sieve-tubes (comp. p. 173). It may here not be superfluous to add, that, so far as my experience extends, the same statement also applies essentially to the common’ officinal barks, and that it is only the dried condition in which they usually come under observation which has hitherto, to some extent, impeded the clear recognition of the actual conditions. : The elements of the soft bast in woody plants of the families Apocynez (Nerium Oleander), Asclepiadeze (Asclepias curassavica), Convolvulaceze (Convolvulus Che- orum), and Campanulaceze (Campanula Vidalii), are less regularly arranged than in the cases hitherto described, and this is no doubt also the case in the Cichoriacea, and in the plants mentioned at p. 324 (2), which are characterised by very small sieve- tubes in the primary bundles. Between relatively wide parenchymatous cells, which in some degree maintain the serial arrangement, groups of narrow elements are found in these instances, which, as seen in cross-section, show very various, triangular, or polygonal forms, and have apparently arisen from repeated longitudinal divisions taking place in every direction in the original mother-cells of the tissue. The narrow elements which are thus grouped are the sieve-tubes and cambiform cells; in the Cichoriaceze, Campanula, and Lobelia, the laticiferous tubes also occur ‘in connection with the groups of narrow elements. The arrangement of the elements of the bundle in the soft-bast, as just described, prevails both in the stems and roots of woody plants, so far as. the existing investi- gations extend. There have been few minute investigations on the stems of herbaceous plants, but, according to the existing observations, they are not essentially different from woody plants with reference to the conditions herein question. , Those roots which consist chiefly of parenchyma, with a very thick secondary bast, as described above at p. 516, show in the latter a distribution of the sieve-tubes which’ is similar to that of the vessels in the massive parenchymatous wood of the root. A strand of bast corresponds to every ligneous strand, and the former contains relatively small groups of narrow sieve-tubes, accompanied by narrow elongated cells, and enclosed in a mass of large-celled parenchyma, * : The more special distribution, relative amount, and form of the tissues under con- sideration is very various, according to the particular case, and often even in closely related plants. In most cases the bundles, as seen in cross-section, form relatively narrow radial bands, lying in the same straight line with the ligneous strands, between rays of wide-celled parenchyma, which essentially, though not always quite exactly, form the continuation of those of the wood. The strands consist of narrow, elongated cells, which may even be pointed like a spindle (the latter, e.g. in the wild form of Daucus Carota), and of sieve-tubes lying between them, which are always scantily developed, and are likewise narrow, The individual radial bands, like the ligneous strands, are either continuous, or are more or less interrupted by interpolated large-celled parenchyma. Examples of this arrangement are afforded by many roots of Umbellifere, Scorzonera hispanica, Cichorium, Argemone, &c. In the roots of Rhubarb (Rheum undulatum and Rhaponticum) essentially the same distribution reappears, but is modified by the forms and relative numbers of the histological elements under consideration. The uniseriate, procumbent, parenchymatous rays of the wood are continued without interruption through the zone of bast. The strands separated by them consist, as regards much the greater portion of their mass, of large, upright, parenchymatous cells, filled with starch, and between these lie the SECONDARY THICKENING, NORMAL DICOTYLEDONS. 525. narrow sieve-tubes, which are very isolated, and therefore easily overlooked, and are accompanied by narrow, elongated cells. Finally, an arrangement deviating further from the usual rule is represented by the bast of the roots of Taraxacum. Narrow, concentric, annular zones containing the tubes, here alternate regularly with similarly arranged, broad zones of large-celled parenchyma, which are on the average about sixteen layers of cells in thickness. The _zones containing the tubes consist of narrow cells, numerous milk-tubes, and scanty sieve-tubes; the large-celled zones consist of thin-walled cells, which are placed in very. regular radial and vertical rows, corresponding to the original cambial form and arrangement. At numerous points the radial series of these elements pass through the zones containing the tubes, so as to interrupt them, without however consisting of special ray-parenchyma distinct in form from that of the annular zone. The roots of Chelidonium and Papaver are intermediate in structure between those of Taraxacum, and of the first category. ; Sect. 163. In plants which have laticiferous tubes (p. 186), the non-articulated ones may be absent from the secondary bast; at least I did not find them in ‘it in Vinca, Asclepias curassavica, and the Euphorbias; in most cases they are present, and as regards the articulated tubes especially, this is always the case so far as investigation extends; they are then characteristic companions or representatives of the sieve- tubes. The large non-articulated tubes in Ficus, Maclura, and Morus follow singly the lines of sieve-tubes. The articulated, usually reticulate tubes form groups. together with the sieve-tubes, which anastomose, both within each individual. strand of bast, and with those of neighbouring strands, by means of connecting branches. In the bast of the Papayaceze the net of milk-tubes has a relatively scanty development, at least in comparison to its complexity in the wood. In the other plants belonging to this category the milk-tubes are always relatively very. numerous, as is especially conspicuous in the shrubby stems (Sonchus pinnatus and Campanula Vidalii), and in the roots of Cichoriacez, Campanulacee, and Pa- paveracez: as their number increases the sieve-tubes become proportionately reduced. In the strands of bast of the roots of Cichoriaceze (Lactuca virosa, Taraxacum) only scanty, narrow sieve-tubes are present, as has already been mentioned above; in the secondary bast of the root of Platycodon grandiflorus, I did not find the latter at all, though I do not wish to assert that they are entirely absent. This mutual representation appears in the most striking manner in the bast of the roots of Papaveraceze: Papaver Rhoeas and Argemone mexicana have only very isolated sieve-tubes side by side with the highly developed net of laticiferous tubes ; in Chelidonium majus the former are more numerous, although the milk-tubes pre- dominate; Glaucium luteum has no milk-tubes, but, on the other hand, has large groups of sieve-tubes. Sect. 164. The occurrence of protogenetic secretory passages in the soft-bast has already been to some extent noticed in anticipation in Chapter XIII. They only. occur in those plants and parts of plants which also possess them in the primary tissues, and by no means in all even of those. Their position appears to be always within the bundles, not in the medullary rays. Among the Dicotyledonous families in question their occurrence in the secondary bast of stem and root in the Terebinthacez, Burseracez, and Clusiaceze has already been mentioned above. So far as is known they are present in the 526 SECONDARY CHANGES. same regions in all Umbelliferas, and in the Araliaceze ; where the bast is strongly developed they form, as seen in transverse section, interrupted radial and concentric series, differing in arrangement according to the species’. In the secondary bast of the stem and branches of Pittosporum Tobira, the passages appear relativély late; Van Tieghem found four concentric rows in a branch 10™™ thick. They were not found in the secondary bast of the root of this plant. Among the Composite which contain passages, some also have them in the secondary bast, e.g. Helianthus and Centaurea atropurpurea; in Inula Helenium the bast of the root contains wide passages, which are closed blindly, so far as is known, at both ends, and coated by a delicate epithelium: they resemble those in the secondary wood. Other Composite have no passages in the region mentioned, but, on the other hand, have scattered sacs, filled with secretions, in the parenchyma ©’: of the rays, e.g. Echinops and Tagetes patula (comp. p. 203). The Coniferze, which have such great numbers of protogenetic resin-passages in their other tissues, do not form them in the secondary bast, except in a few cases. The exceptions include, firstly, the blind ends of the horizontal passages in the medullary rays of the Abietineze mentioned on p. 490, which extend into the zone of bast. According to Van Tieghem, longitudinal passages occur in the secondary bast _ of Araucaria Cookii and brasiliensis, and of Widdringtonia cupressoides, which were mentioned on p. 443. The spaces filled with resin which occur in other Conifere are subsequent, hysterogenetic products of disorganisation, which will be discussed in Sect. 173. Secretory sacs, other than those containing crystals, appear in the soft bast in the plants mentioned in Sects. 33-35, and are sometimes scattered without any perceptible regularity, while they sometimes have a definite arrangement, also stated in the paragraphs mentioned. Sect. 165. The sclerenchymatous fibres of the bast, bast-fidres, or according to the older terminology ‘bast-cells’ in the strictest sense, have the form and structure generally described in Chap. II. With reference to the latter, it may further be mentioned that the lamella which forms their boundary, whether towards elements of the soft-bast, or towards one another, is in these cases especially often an unlignified membrane of cellulose, which surrounds the more or less lignified, thick membrane of the fibre as a distinct sheath? Comp. Figs. 211, 212. The bast-fibres are entirely absent in the bast of many plants; both in its secondary portion, and at the outer boundary of the primary phloem. This is the case in the stems ard branches of Ribes, Viburnum Lantana 8, Pittosporum Tobira, undulatifolium, Citriobatus multiflorus, Porlieria, Centradenia grandifolia, and Ber- beris vulgaris, and in the roots of many herbaceous Dicotyledons. Thus they do not universally form an essential constituent of the bast. In the cases where they are present, which certainly form the majority, they occur— * Compare the figures of roots of Umbelliferse in Wigand, Pharmacognosie, and Berg, Atlas, Taf. 8, 9, 14, 22. * See Graf z. Solms-Lambach, Bot, Zeitg. 1871, p. 516, &c. * Hanstein, Baumrinde, p, 17. SECONDARY THICKENING, NORMAL DICOTYLEDONS, 527 (1) Only at the outer limit of the primary bundles, surrounding their phloem- portions (comp. p. 422), and not in the products of secondary growth. This is the case in the stem and branches of Fagus, Betula, Alnus, Platanus, Viscum, Menispermum, Viburnum Opulus, Convolvulus Cneorum, Nerium, Cornus, Punica, Camellia japonica, Drimys Winteri, Ephedra distachya, Abietineze*, &c. (2) Both at the outer limit above-mentioned, and also in the interior of the secondary bast. This latter condition is no doubt the most frequent, especially among woody plants. With reference to the relative amount and distribution of the fibres, it presents very various modifications. In the medullary rays fibres are rarely present, e.g. isolated ones in Tilia, comp. Fig. 212, The principal forms of their distribution in the bundles are the following :— , (a) Concentric layers or rows of fibres alternate regularly with similar layers of soft bast. The layers of both kinds in neighbouring bundles fit approximately, though not always quite exactly, one on another, so that they form annular zones, extending round the whole stem, and interrupted by the medullary rays. This phenomenon occurs with especial regularity, as already mentioned at p. 522, in the Cupressineee and many Taxinez, where every fourth secondary tangential row of cells becomes a uniseriate zone of fibres, which separates two triseriate zones of soft bast from one another. Comp. Fig. 211. Among the Dicotyledons no such strict regularity exists. The fibrous layers always consist, on the average, of two or more tangential rows, and the number of these rows changes in the same individual, both in the successive annular zones and within each individual portion of the bundle; as follows from these facts, the thick- ness of the zones of soft bast is also unequal. The conditions in question are also extremely various, according to the species. In the case of many species, however, a regular alternation of concentric zones of fibres and soft bast, of definite average breadth, takes place within the limits of deviation indicated, the original radial and tangential seriation of the fibres of each portion of the bundle being sometimes maintained, as in Vitis, Spirzea ulmifolia, Pterocarya caucasica, and species of Acer, though in most cases their position as seen in transverse section becomes irregular, owing to the displacements due to longitudinal extension (p. 470): Tilia, Cheirostemon, Sparmannia, Malvaceze, Medinilla, species of Salix, Ladenbergia globosa®, Vas- concella monoica, Guaiacum, and Clematis Vitalba ; comp. Figs. 212 and 214. (4) Concentric zones of fibres, alternating with soft bast, may still be dis- tinguished generally, and in places are even regularly arranged; on the whole, however, they are irregular, as they are both interrupted in each bundle by elements of the soft bast, and are also unequal in number and dissimilar in arrangement in neighbouring bundles. This condition, with numerous variations according to species and individuals, and in the average number and breadth of the successive zones, is characteristic of the bast of very many woody Dicotyledonous plants, 1 Compare Hartig. Forstl, Culturpfl. pp. 13, 212, 326, &c,—Hanstein, /,c. p. 21.—Schacht, Der Baum, p. 381.—Von Mohl, Z.¢. p. 891. ? Berg, Atlas, Taf. 29. 528 SECONDARY CHANGES. e.g. Quercus, Corylus, Carpinus, Pyrus, Juglans regia, Sambucus nigra’, Daphne Mezereum, Rhamnus Frangula, Simaruba officinalis’, Ulmus °, Glycine sinensis; Quillaja, Olea europza, and Populus pyramidalis. In these cases the fibres of each group are seldom radially arranged; their arrangement as seen in cross-section is usually irregular. . ’ (c) The bast of numerous other Dicotyledons contains fibres scattered through- out the soft bast, singly or in small groups, as seen in cross-section; traced FIG, 214.—Sparmannia Africana; branch, transverse section (80). Below A—/ is wood; above 2—A is first the cambial zone, then higher up and towards the outside is the layer of bast, the outer limit of which lies at 4; m larger medullary rays; the single radial rows marked < X are the smaller ones. Alternating with the medullary rays are narrow strands of bast, consisting of alternate groups of fibres and of soft bast with narrow cavities. On the external boundary of the bast are sacs or cells with stellate crystals. ¢ remains of the epidennis; ? periderm; s remains of a sac containing mucilage, after the mucilage has been washed out. : longitudinally they are also isolated, or form narrow strands, anastomosing with others at an acute angle; sometimes they are distributed over the tranverse section in large numbers, as in the external zone of bast of Ladenbergia magnifolia, and in the bast of most Cinchonas‘, also in Ficus elastica, Morus, and Celtis’; 1 Von Mohl, Z.¢. p. 879. ? Berg, Atlas, Taf. 28, &c. 5 Hartig, 2.¢. p. 466. * Compare Berg, Atlas, Taf. 29-35. 5 Hartig, Forstl. Culturpfl. p. 450. SECONDARY THICKENING, NORMAL DICOTYLEDONS. 529 sometimes they occur in relatively very small quantities, as in the inner zone of bast of the Ladenbergia last-mentioned, and in the officinal barks of Cinnamomum, Croton Eluteria’, Larix europzea *, and Mahonia Aquifolium. As the fibres are included in the general serial arrangement of the bast, it follows that the more closely they stand, the more distinct in this case also are the inter- rupted, radial, and concentric zones which they form, as e.g. in Cinchona macrocalyx, figured by Berg, Z.c., Taf. 35, and very beautifully in Laurus Sassafras. In general, numerous intermediate cases occur between the forms of distribution enumerated here and under 4, as was to be expected beforehand. In this respect the cortices of Cinchonacez present an instructive series of gradations. Ulmus also deserves to be again mentioned here. The appearance of shor/ sclerenchymatous elements (stone-elements) in the bast will be considered in the next chapter, in order to avoid repetitions. Sect. 166. Sacs containing crystals are often a characteristic, sometimes even a predominant, constituent of the secondary bast ; their occurrence however is as far from being universal as is that of the fibres. They lie both in the medullary rays and in the bundles, in the latter usually forming longitudinal rows with short articulations; each of these rows is derived from a single mother-cell, and may often be isolated as a connected series; they were termed septate sacs at p. 139 (comp. Fig. 213). Guaiacum and Quillaia have already been mentioned above as forming at least partial exceptions to this mode of arrangement; it may be left an open question how far similar, i. e. short, isolated sacs, may further occur elsewhere in the bundles, or not ; no minute investigations on this point have been published, and the statements as to the distribution of crystals, which are chiefly based on trans- verse sections, do not admit of any decision on the question. The form of the crystals is either klinorrhombic, or that of clusters, raphides, or granules; one or more definite forms are characteristic of each particular case (comp. p. 142, and Sanio, /. c.). With reference to the presence or absence of the crystal-containing sacs, and their distribution in the former case, the following phenomena have been observed, which however, at least so far as my own observations are concerned, still require re-investigation. I indicate those forms of the crystals, which have been chiefly observed in transverse sections, in parentheses (K =klinorrhombic, C = clusters, R = raphides). 1, Crystals are absent from the secondary bast: Drimys Winteri, Fraxinus, Syringa, Jasminum fruticans, Mahonia Aquifolium (?), Laurus Sassafras, Cinnamomum aromaticum (Cinnamon Cassia), Clematis Vitalba, Atragene, Aristolochia Sipho (?), Camellia japonica, Sorbus Aria, and, according to Hartig (Forstl. Culturpfl.), also Cornus, As regards the sacs containing the crystals, the Cupressinex, Taxinex, and other Conifere, and Ephedra, are also to be included in this series, as in these the Calcium oxalate is not deposited in the interior of sacs or cells, but is intercalated in the membranes, In most of the plants just mentioned I find that crystals are also absent from the primary cortex. 2. Crystals are contained in the secondary bast (and then usually or always also in the primary bast, and in the external cortex). In these cases they occur— 1 Berg, /.c. 36, 37. 2 Hartig, Forstl. Culturpil. p. 13. 53° SECONDARY CHANGES. (a) Both in the medullary rays and in the bundles: Nerium Oleander (K), Simaruba officinalis (K), Canella (C), Platanus, Cinnamomum zeylanicum and its allies (small raphides, chiefly in the medullary rays), Juglans regia (C, Sanio), Acer platanoides (R), Sparmannia Africana (C), Carpinus Betulus, and Corylus Avellana (K, C, Sanio). () In the bundles, exclusively, or to much the greater extent : species of Salix (C, K), Pyrus communis (K), Punica (C), Ribes (C), Guaiacum (K), Galipea officinalis (R, K), Maclura aurantiaca (K), Ulmus (K), Quillaia (K), sculus (K), Rhamnus Frangula (C), Quercus pedunculata (K, C), Betula verrucosa, Alnus glutinosa (K, C, and granules, Sanio), and Porlieria hygrometrica (K). , (c) Exclusively, or to much the greater extent, in the medullary rays, and when the latter aré’ of considerable breadth, most abundantly at their boundary towards the bundless: Vitis (K, R), Tilia (K, C), Cheirostemon (C), Olea europza (very small R), 4. Ficus elastica (K), Croton Eluteria (C), Pistacia Lentiscus (C), Prunus Padus (C, K), P. avium (C), Kerria japonica (G), Berberis vulgaris (scanty K according to Sanio), Lonicera tatarica (Sanio), and Sambucus nigra (granules, Sanio). The crystal-containing sacs, especially those which are septate and contain klinorrhombic crystals, appear in many cases in company with the bast-fibres, as pointed out by Schacht*; e.g. species of Acer, Pomaceze, Ulmus, Quercus, Salix, &c.? Clusters often oceur exclusively and in large quantities where fibres are absent, e.g. Punica and Ribes ; though the entire absence of the tissue last- mentioned may also occur in connection with the entire absence of crystals, e.g. Drimys Winteri. See ES No constant relation however exists between the presence or absence of the two forms of tissue mentioned, or between any definite form of tissue and of the crystals, as is evident from the facts stated. Thus klinorrhombic crystals are especially often present in abundance where fibres are absent in the secondary bast, e. g. Porlieria and Nerium; while, on the other hand, the klinorrhombic companions of the fibrous bundles are absent in Juglans regia and many other cases, And further, where the fibrous bundles are accompanied by crystals, the occurrence of the same or another form of crystal in the soft bast is by no means excluded, In the soft bast the rows of sacs containing crystals have in most cases an irregular, scattered position in the transverse section, On the other hand, they are often arranged in concentric zones, which alternate regularly with zones destitute of crystals. This is the case in Punica Granatum, where the whole bast appears regularly striated.in transverse section, owing to the fact that uniseriate zones, con- sisting almost exclusively of crystal-sacs, alternate with zones each consisting of a few rows of cells, which are destitute of crystals; these zones are interrupted by numerous uniseriate medullary rays, which are also destitute of crystals (comp. Fig. 215, and the beautiful figure in Berg, Atlas, Taf. 40, which however is not quite correct in the details), In species of Ribes also, the bundles, which are separated by broad medullary rays without crystals, consist of multiseriate zones of soft bast, like- wise destitute of crystals, which alternate regularly with usually uniseriate uninter- rupted zones, containing clustered crystals %, Sect. 167. As regards the changes of structure corresponding to the successive * Der Baum, 1 Anf. pp. 228 and 238. * Compare Sanio, /.¢. * Compare Hanstein, Baumrinde, Figs. 15-17. SECONDARY THICKENING. NORMAL DICOTYLEDONS, 531 zones of secondary growth, to the non-equivalent members, and to individual dif- ferences, there is far less to be said in the case of the bast than in that of the wood, partly, no doubt, owing to the really greater simplicity of the phenomena, partly because it is better to pass over a number of the known changes and to discuss them in the next chapter, and lastly, in no small degree, because, for reasons already stated above, there is still an absence of any very minute or extended investigations on the subject. In woods, the ligneous elements and cambial cells of which increase succes- sively for a time in size (p. 505), the same is generally the case in the bast, as would be expected beforehand, on account of the enlargement of the cambial cells. This is evident on observation in the case of numerous Coni- fers and Dicotyledonous woods, e.g. Tilia, Fagus, and -\exmlioty Nerium, also Vitis and Cobea. It is obvious that Aes OIKSC) : only those internal zones must here be taken into con- SU SsiseShe sideration, which have not as yet undergone any sub- =2(— LOUAS) = sequent dilatation. No very accurate investigations have been published on the degree and the persistency of the general increase in size, nor on the possibly un- equal participation of the particular forms of tissue in =} @ 5 (\ es these changes. 3 Ome JE Between stem and branches differences in the size _-) S@ eee of the elements appear to exist similar to those in the wood, but these also have not yet been more closely investigated. In the roots of trees and shrubs, so far as the investigations extend, the special structure of the bast is very similar to that in the stem of the same plant, e.g. Vitis, Sambucus, Tilia, and Punica. Wherein the differences, which no doubt occur, con- M @ m “oT VOOM [= ar FIG. 215.—Punica Granatum (220) ; trans- verse section through the inner part of the bast of a branch six years old. c side towards the cambium; 2, # two medullary rays; between the latter is a strand of bast; s, sieve-tubes. It is uncertain which of the other cavities belong to elements of the latter kind. Between the empty cavities of the elements of the soft bast are transverse zones of crystal-containing sacs; the clusters con- sist, cannot be stated at present. It has been repeatedly tained in them are only indicated by shading. mentioned above (pp. 516 and 524) that a different relation exists between the foliage stems of herbaceous plants, and the roots which belong to them, especially when the latter are fleshy. The thickness of the secondary zone of bast added in a definite space of time is very variable, both in its relation to the simultaneous growth of the wood, and according to absolute measure, whether in different species and individuals, or in non-equivalent members of the same plant. In both relations the extreme cases are represented on the one hand by the fleshy roots, consisting chiefly of bast (p. 516), and on the other by the woody stems of trees and shrubs. Especially in trees with a persistent cortex, which is not thrown off by the formation of bark (Sect. 177), as, for example, the Fir and Beech, the difference of thickness between wood and bast is, as is well known, very considerable, and the absolute thickness of the latter small. In the common Beech with smooth bark (Fagus silvatica) the entire bast-zone in a stem 100 years old is, according to Hartig ', scarcely more than 1™™ thick. Woody plants which periodically throw off their Forstl. Culturpfl. p. 212. MM 2 532 SECONDARY CHANGES, cortex by the formation of bark show a more vigorous production of bast, corre- sponding to the activity of this process. This phenomenon appears as an instructive individual variation in those stems of Fagus silvatica occasionally occurring which are called Stone-beeches, and are conspicuous from their thick, furrowed bark. The new production of bast no doubt takes place most abundantly in those stems, which, like the vine, annually renew their entire bast-zone, and throw off that of the previous year. Le It is remote from the purpose of the present description to enter into the causal relations of these phenomena. Generally valid anatomical characters of the boundary between successive zones, corresponding to the annual rings of the wood, may perhaps still be discovered in the case of the bast, but cannot be determined from the existing data. Even in those cases where regularly alternating concentric zones of non-equivalent tissue appear in the bast, especially fibrous zones alternating with soft bast, their number varies according to the year, the age, and the individual, and a determination of the annual boundaries is therefore usually uncertain. As an example, Hartig’s? statement may be cited, that in the case of Willows and Poplars with a smooth cortex, the number of zones of bast-fibres is smaller than the age in years, amounting only to 3-4 for every 10-15 years, while, during the development of thick bark-forming cortices, 2-4 fibrous zones arise annually. The species of Acer? form in the first years either one, or (at the base of the annual layer) two, successive fibrous zones, but even from the sixth year onwards the relation changes in such a manner that often only 20-25 fibrous zones correspond to roo years. The numbers are more regular in Tilia, where, according to Hartig*, four fibrous zones appear at the base and one at the apex of the shoot in its first year, to which two or three are added in the second year, and in each succeeding year, on the average, two; this is also the case in Pyrus communis, which, according to Mohl*, forms one fibrous zone annnally. In those woody plants also, which renew their cortex every year, and in which the limit of the annual increment of growth is sharply defined by the layer of periderm formed at its outer side (Sect. 177), similar differences to those just mentioned appear. Lonicera Caprifolium and its allies annually form one zone of fibres and one of soft bast; Clematis Vitalba usually two of each®; Vitis vinifera generally forms two Bbratg zones alternating with soft bast, at the close of the first period of vegetation, while in later years 3-5 are generally formed annually * 1 Lc. Dp. 444. 2 Le. pe 547 3 Zc. p. 560. * Botan. Zeitg. 1855, p. 880. 5 Hanstein, Baumrinde, pp. 72, 77. ° Hanstein, 2.¢, p. 61.—Von Mobhl, Zc, p. 879. CHAPTER XV. SECONDARY CHANGES OUTSIDE THE ZONE OF THICKENING. Sect. 168. In the normal Dicotyledons and Gymnosperms the mass of wood and bast developed from the cambium is bounded on the one hand by the pith, and the pre-existing ligneous zones adjacent to it, on the other by the external cortex and the pre-existing bast. It is clear, a przorz, that this surrounding tissue may, and to some extent must, undergo changes, in consequence of the cambiogenetic secondary growth. As stated above, no change in the pzth is necessarily involved in the actual course of the processes of secondary growth described, or as a consequence of them. As a matter of fact, however, it has been asserted, especially by Duhamel}, that in trees and shrubs the medullary cylinder diminishes in thickness, and may at last wholly disappear as the secondary growth of the wood proceeds. As regards the majority of cases, this view has been given up, as depending upon imperfect observation, and in fact the only demonstrable anatomical change in the pith during the phenomena of secondary growth is, that it sooner or later, rapidly or slowly, dies off and dries up. The possibility of a change in the pith caused directly by the growth in thickness is not indeed excluded a@ prior?. For if wood and bast increase in thickness and circumference, and the external cortex does not yield to this enlargement of the circumference in a corresponding degree, am increasing pressure will be exercised on the pith, which may lead to anatomical changes in the latter. In what cases and in what form such changes may possibly take place, are questions which have not been investigated, and to the solution of which there is scarcely any safe clue; the possibilities will not be discussed here. That such cases occur is however shown by the change of form in the pith of the internodes of Aristolochia Sipho, which accompanies the growth in thickness. The young internode, up to the age of one year, has an approximately circular transverse section. The pith, which is of the same form, or is broadly elliptical, as seen in transverse section, consists, like the medullary rays, of permanently delicate and soft- walled cells, and is surrounded by a circle of 11-13 leaf-trace bundles*, without intermediate strands. The external cortex enclosing the latter contains a strong 1 Physique des Arbres, I. p. 37; detailed discussion in De Candolle, Organographie, I, p. 168, * Compare Nageli, Beitrige, p. 82, Taf. VIII. 534 SECONDARY CHANGES. ting of sclerenchyma (p. 419), and is covered by the very tough epidermis, which is for a time persistent. The vascular bundles take an unequal share in the cambio- genetic growth in thickness, which makes a vigorous start in the next year. The median bundles of the two next higher leaves, occupying two diametrically opposite segments of the circle, and their next neighbours, grow less strongly in thickness than those situated in the two other segments; the increase is greatest in the three bundles which occupy the middle of each of the latter segments; it chiefly affects the xylem. During this unequal growth in thickness no perceptible change at first occurs in the form of the cross-section of the whole internode, and even at a later period the change is but trifling. On the other hand, those ligneous bundles, which grow more strongly than the others, press with their inner edges against the pith, the cells of which become compressed in the direction of the corresponding radii of the transverse section, and the general form of the pith is changed in such a manner that its transverse diameter, lying in the direction of the greatest growth in thickness, constantly becomes smaller, while the diameter at right angles to the latter remains unchanged. In an internode five or six years old the pith is merely a narrow band, the shorter diameter of which scarcely amounts to 5 of the (original) longer one. The cause of these phenomena manifestly lies in the fact that the cortex undergoes too small an extension for the volume of the growing wood, and by its resistance presses the latter against the pith, and compresses it. Allied species of Aristolochia behave in a completely similar way. The disorganisation of the pith in the shrubby species of Astragalus, which yield gum-tragacanth, may further be mentioned here, though not actually standing in direct causal connection with the anatomical processes accompanying growth in thickness‘. It consists in the conversion of the cellulose-membranes into a mucilage which is capable of swelling up greatly; the change extends from the pith to the medullary rays, and chiefly affects the cells of the middle of the latter, and of the pith, while those which border on the ligneous strands are changed in a less degree, or not at all. In the living plant the mucilage is already present in a highly swollen condition, and is kept in high tension by the pressure of the surrounding resistent tissue *. On injury to these tissues it flows out, and when a plant is left to itself it may spontaneously burst the surrounding tissue, and exude from the cracks in the form of the strings which on drying form the tragacanth of commerce. In many species, e.g. A. rhodosemius, the mucilaginous disorganisation begins early, even immediately below the apex of the stem®; in others it appears to come on at a later period. As regards the gradual dying off of the pith in old woody stems, the disappear- ance of the cell-contents, and especially of the store of starch, similar rules hold good to those for the formation of heart-wood. Comp. pp. 403 and sro. Leaving out of consideration the displacements obviously occurring in the cases of Aristolochia and Astragalus, no anatomical changes in the wood, directly caused by the growth in thickness, are known. On the influence of the cortical pressure on the formation of autumn-wood, which is indirectly connected with the above, comp. Von Mohl, Botan. Zeitg. 1857, p. 33. * Flitckiger and Hanbury, Pharmacographia, p. 153- ° Graf zu Solms-Laubach, Botan, Zeitg. 1874, p. 69. SECONDARY CHANGES OUTSIDE THE ZONE OF THICKENING. 535 the remarks and citations upon p. 500; on the processes of degradation in wood when ageing, the relation of which to these phenomena is, to say the least, very doubtful, see Sect. 154. The primary medullary bundles of the forms here in question—Piperacez, Begonize, Aralize and Umbelliferee, Mamillariz, and Melasto- mace (comp. Sect. 62)—undergo, so far as is known, no secondary anatomical changes. Sect. 169. All the parts lying outside the active cambial zone, namely, the entire primary cortex and the secondary cortex for the time being, necessarily undergo progressive changes with the progressive growth in thickness, These consist in— (z) Growth of existing tissue-elements, new formation of equivalent ones from them, and subsequent metamorphosis (p. 5): Sects. 170, 171. (4) Compression, displacement, and déstruction of existing tissues: Sects. 172, 173. (¢) New formation of non-equivalent forms of tissue out of existing ones: Periderm, Sects. 174-179. The process indicated by (a) only affects the cellular tissues: Epidermis and Parenchyma. Sect. 170. In the majority of the cases of vigorous growth in thickness, the epidermis is destroyed at an early period, cork or bark being formed; further details will be given below. Stems, however, are not wanting, whether with weak or with very vigorous growth in thickness, in which the epidermis follows the latter by its own growth during a considerable period. This is the case in many herbaceous plants, and in woody plants with a smooth, green, cortical surface, as long as the latter is present. The condition of the surface mentioned is in fact dependent on the persistency of the epidermis, the cells of which, being filled with sap, allow the colour of the sub-epidermal chlorophyll to show through. As examples of stems and branches with vigorous secondary growth of the wood, which, for some years at least, retain the epidermis, may be mentioned: Viscum album‘, species of Ilex, the ever- green Jasmines, Menispermum canadense, Aristolochia Sipho and allies, Sophora japonica, Negundo, and many others. -In Acer striatum the epidermis is still for the most part present in a living condition, and following the growth, even on stems a foot thick, and forty years or more old. The long-lived epidermis of the woody plants mentioned is provided from the first with thick strongly cuticularised outer walls, which sometimes contain and excrete a large amount of wax (comp. pp. 76, 82). Its original structure under- goes relatively unimportant changes during growth. These consist in an increased thickening of the cuticularised external walls, the surface of which covered by the cuticle usually remains smooth; but in Acer striatum as well as in Negundo and So- phora japonica it becomes cracked as the thickening proceeds, the cracks penetrating from outside into the external cuticular layers, which do not follow the. growth, and successively breaking them up into crumbling fragments. In Acer striatum the cracks coincide with the dilated bands of the external cortex, to be described below ; and a new excretion of wax rods takes place in each case on the surfaces newly. laid bare by the cracks ; it is on this that the white striation of the cortical surface ! Von Mohl, Botan. Zeitg. 1859, p. 593. - 536 SECONDARY CHANGES. depends’. In addition to these changes, the growing epidermal cells, as they become larger in the direction of the circumference, i.e. broader, divide successively by walls which stand at right angles to their transverse diameter, and to the surface, and abut on the inner surface of the original wall. This successive multiplication of the epidermal cells takes place in such a degree, that the average breadth and general form of the individual cells remain approximately the same, or only undergo inconsiderable changes. The epidermal cells of the stem of Acer striatum when the latter is 200™™ thick, are, for example, scarcely twice as broad as those of a shoot one year old, and 5™™ in diameter. Sect. 171. Parenchyma forms the main mass of the primary external cortex, the medullary rays of various degrees in the bast-layer, and the parenchymatous groups in the bundles of the latter. Until a zone of cortex is thrown off by the formation of bark, which may take place sooner or later, but in many instances never occurs at all, and which will be discussed below, the parenchyma follows the cambiogenetic secondary growth by corresponding growth of its own, in all the parts in which it exists. The parenchymatous mantle of the external cortex increases successively in width, while the medullary rays of the bast, and the parenchymatous elements of the bundles, increase in breadth in the centrifugal direction (Fig. 214, p. 528). The several portions of the bast do not always participate in the same degree in this phenomenon, which may shortly be termed Dilatation of the parenchyma. If attention be directed to cases of extreme difference, it is found that in the one case dilatation of the entire parenchyma of the bast takes place in an approximately uniform proportion, as each annular zone becomes shifted outwards. In all the radial bands, and thus most clearly in the medullary rays of every degree, the parenchymatous cells increase uniformly, and quite gradually in breadth, in the centrifugal direction. The intermediate non- equivalent tissues, which do not grow with them, especially sieve-tubes and bast-fibres, thus become uniformly removed one from another, and the more so the further they are from the cambium; as in Salix fragilis and allies, Punica, Rhamnus Fran- gula*, Spirzea ulmifolia, Pyrus communis, and #sculus. In the other extreme case the dilatation is unequal in the various radial bands of the transverse section; it amounts to little or nothing in the bundles, and is most active, either in all the parenchymatous rays, or in some of them. Between the lateral limits of these dilated rays the arrangement, and lateral distance from one another of all elements of the tissues, remains approximately the same. This behaviour occurs, firstly, in a number of stems, which are constructed on the type described at p. 458, and the large medullary rays of which are broad and multiseriate; e.g. stems of Meni- spermum, Aristolochia, and Piperacez ; here the dilatation is brought about, at least to the greatest extent, by the large medullary rays, and, indeed, the latter all take an approximately uniform share in the process, in their entire height. The strands of bast therefore remain similar in form and arrangement to the phloem portions of the original vascular bundles, from the further development of which they have arisen, They are not, however, wholly without share in the dilatation, as a slight * For details compare Botan. Zeitg. 1871, p. 605, &c. ? Compare Berg, Atlas, Taf. 39, 40, ‘SECONDARY CHANGES OUTSIDE THE ZONE OF THICKENING. 537 widening of their parenchymatous elements, and through them of the whole bundle, takes place here also. This category includes Tilia, and other woody plants provided with a similarly grouped bast. In young shoots of Tilia the zone of wood and bast is traversed by numerous large parenchymatous rays extending to the pith, most of which are uniseriate, but many, e.g. in transverse sections of a branch of Tilia parviflora now before me, the seventh, ninth, or fourth, and so on, are bi- or triseriate; the latter is at any rate the case at the cambial limit, and sometimes as far as the pith, while in other cases they become uniseriate before reaching the latter. The dilatation begins in the biseriate and triseriate rays; whatever strands and rays lie between them, take, in the first instance, no part in the process. As growth in thickness proceeds, a constantly increasing number of the original uniseriate rays take part in the dilatation. Later on successive small secondary rays are also in- volved in it. When a ray takes part in the dilatation it is, as a rule, bi- or pluriseriate at the cambial boundary. The result of these phenomena is in the first instance the severance of the bast-layer into the often described’ groups of bundles with a wedge-like form widened towards the cambium, as seen in cross-section, and rays, alternating with the former, widened in the opposite direction, and further, the successively occurring subdivision of the first groups of bundles into more numerous narrower ones, separated by rays. The number of groups in a transverse section, rises, for example, in branches of T. argentea now before me, from 45, in an internode 6™™ thick, to 138 in one 28™™ thick. In order to complete this description, which has been given in the first instance with reference to the transverse section, it must be added that the medullary rays are of considerable height—the larger ones measuring more than a hundred cells—and their ends situated in quite different transverse sections all round the stem; and that the dilatation begins in every ray about at the middle of its height, and thence advances upwards and downwards. Quite similar phenomena to those in the bast of the Lime reappear in many other plants, and in members of different value, e.g. in the stem and branches of Hibiscus syriacus, Pterocarya, and Galipea officinalis, and in the highly parenchyma- tous roots of Umbelliferze (Archangelica, Levisticum, &c.), Glycyrrhiza, and many others *, Intermediate cases between the extremes, represented on the one hand by Salix fragilis, Spirzea ulmifolia, and Punica, and on the other by Tilia and Menispermum, would naturally be expected to occur from what has been stated. In these every transverse section shows radial bands of the bast, which are more or less strongly dilated, in various gradations. Examples are afforded by Sparmannia africana, represented in Fig. 214, p. 528, with numerous strongly dilated rays, and more slightly dilated ones in the divisions limited by them; further, in very various gradations, by the Quinine barks, Croton Eluteria, Simaruba officinalis, Cinnamomum zeyla- nicum, &c.° In the structure of the dilated masses of parenchyma the increase in the 1 Compare e.g. Schacht, Der Baum, p. 198; Lehrb. IJ. p. 50.—Hanstein, Baumrinde, Taf. I, * Compare Berg, Atlas, Taf. 37. 6, 8, 9. 5 Berg, 2c. Tab. 29-38. 538 SECONDARY CHANGES. tangential diameter, or breadth, of all cells taking part in the dilatation, is most conspicuous, and follows obviously from what has been stated. In proportion, however, to the increase in breadth, successive radial bipartitions take place, by means of which the original breadth of the cells is approximately restored, and the number of the cells in each tangential row increased in a corresponding degree. These phenomena also take place in the endodermis of the stems, mentioned at pp. 121 and 415, and of the roots, so long as it is not thrown off and thus excluded from the growth. An increase in the average breadth of the individual cell no doubt takes place, judging from estimates. It appears to rise rapidly to an approximately constant value, and then to maintain this during the succeeding divisions, so that cells of the same layer in a stem a foot thick are no broader than in one as thick as one’s finger, though they are of course more numerous in a corresponding degree, The final, constant, average dimensions are relatively little in excess of those existing originally at the beginning of growth in thickness; they may be estimated to amount to scarcely more than two or three times the latter. Accurate measure- ments are still to be undertaken. Cell-division in directions other than the radial, resulting in a multiplication of the concentric layers of parenchyma, is rare, at least in the cortex of those woody plants which have been principally investigated, and apart from the formation of periderms: those special] cases in which it does occur will be discussed below (Sect. 172); it remains to be investigated whether or not such division also takes place in many fleshy roots. The structure of the cell-walls and of the contents, the periodic variation in the amount of starch in the latter, &c., remain the same in the principal mass of the parenchyma during the dilatational changes, in certain cases throughout life, in others for a time. Sooner or later, however, changes may occur, and in fact (2) dilatational changes of the collenchymatous hypodermal layers (p. 404), and (4) processes of secondary sclerosis. The collenchymatous hypodermal layers of the cortex of stem and branches in woody plants always follow the dilatation uniformly for a time in their whole circumference, while they maintain their original characteristic structure ; this process often continues uniformly, so long as they are not thrown off by the formation of bark; the question whether, in many cases, their walls decrease somewhat in thickness, as the extension proceeds, remains to be more accurately investigated. In some cases, which might no doubt be multiplied by further observations, namely, in Tilia, Acer striatum, and Aisculus, other conditions prevail. At certain points the collenchymatous cells show a considerably greater growth in breadth than in the intermediate tracts, and all their walls, both those originally existing and the portions added by subsequent growth, decrease considerably in thickness. They permanently assume the appearance of thin cellulose-walls, and are thus sharply distinguished from the neighbouring thick, brilliant, collenchymatous walls. With this is con- nected, at least in the case of Acer striatum, a diminution in the amount of chlorophyll, which is apparent to the naked eye. The process begins in a few small portions of each transverse section, spreads laterally from these points, and involves new regions lying between the first. The thick-walled collenchyma in this way becomes subdivided into constantly smaller portions, lying isolated among the altered tissue, and at last disappears, as these portions also become involved in the process SECONDARY CHANGES OUTSIDE THE ZONE OF THICKENING. 539 of transformation. Until this occurs the external cortex appears traversed by the altered bands, which I have already termed elsewhere ‘dilatational bands’. Their occurrence indicates an unequal participation of alternate bands of the cortex in the dilatation, and this will find its explanation in dissimilar mechanical conditions, which are still to be more minutely studied. In Tilia the first dilatational bands of the external cortex correspond exactly to the most dilated medullary rays ; in Acer striatum they often coincide with the intermediate spaces between the outer- most bundles of bast-fibres, which are filled with parenchyma, but they do not occupy this position exactly or constantly. In the peripheral layers corresponding to the original collenchyma the con- nection of the cells remains close and approximately unbroken, if the local in- terruptions due to the lenticels, to be described below, be left out of consideration. In the inner, lacunar, portion of the external cortex the cavities originally existing grow in the direction of the dilatation. Further, and often extensive, interruptions of continuity may be produced both in the region indicated, and also, though less frequently, in the more deeply situated layers of the bast, as the effect of the tension, distortion, and pressure which the tissues undergo during growth in thickness. In proportion to the originally lacunar structure of the external cortex, it therefore becomes traversed to an increasing extent, as dilatation proceeds, by broad, crevice- shaped cavities, and is often split up into irregular concentric lamellz, e.g. species of Prunus and Pyrus, Asculus, &c. Coase Seo my Oe ae is fay al ye lar A 9 m p 4 ees Ee yh