^i _^i M »PM property of K^t'l Inst, of liealtJl KlAIffi P.O. 4^-^.2j^^'i1^. rV>C' T»^<^'^^ O?^ .aV'- . TJI THE ANATOMY, PHYSIOLOGY, MORPHOLOGY, AND DEVELOPMENT OK THE BLOW-FLY. (CALLII'IIOKA KRVTIIKOCEI'IIALA.) Jl (Stubj) in the tllompiivatibc Jlnatomi) ani) ^Hoiphologn of insects. 11// 1 H PLATES AND / Ll.USTKATlONS EXECUTED DIRECTLY PROM THE DRAWINGS OF THE AUTHOR. B. THOMPSON LOWNE, F.R.C.S., F.L.S., HUNTKRIAN I'ROFESSOK OF COMI'ARATIVE ANATOMY IN THE ROYAL COLLEGE OF SURGEONS ; LECTURER ON PHYSIOLOGY IN THE MIDDLESEX HOSI'ITAL MEDICAL SCHOOL, ETC. \ AUTHOR OF 'THE ANATOMY AND I'HYSIOl.OGV OF THE BLOW-FLV ' (1870) ; LATE PRESIDENT OF THE QUEKETT MICROSCOI'ICAL Cl.UM. VOL. 1. LONDON : rUliLISlIKU FOR THE AUTHOR liV R. ir. PORTLK, iS PRINCES STREET, CAVENDISH SQUAKK, W. 1890-92. 531 (VI-7 V. 1 PREFACE. In 1870 I published a small treatise on the ' Anatomy of the Blow- Fly.' This has now been out of print for nearly ten years. In 1890, when I undertook the present work, a book of about 300 pages was contemplated ; since then, however, it has grown to more than twice that size, and it has been found necessary to divide it into two volumes. The present volume deals with the subject generally — with the anatomy of the larva and the development of the embryo in the egg and of the nymph in the pupa, as well as with the external skeleton and histology of the perfect insect. The second volume will consist of a detailed description of the various internal organs, their development and physiology. The issue of the parts of this volume has been unavoidably delayed. The introduction and the first four chapters appeared in October, i8go, the fifth chapter in April, i8gi, and the remainder in April, 1892. It is hardly to be expected that a work of the present magni- tude can be without errors, but I trust that any which may be found will be unimportant. I have endeavoured to keep matters of fact distinct from the hypotheses and conclusions iv PREFACE. which rest upon them. Many of my views are diametrically opposed to those usually received. In such cases the accepted view has been given as well as my own. Every student will acknowledge that the morphology of Arthropods has made immaterial advances during the past fifty years when compared with that of Vertebrates ; yet an immense number of new observations have been recorded. This is sufficiently proved by the enormously increased litera- ture upon the subject. I have availed myself largely of this accumulated literature, but I have never allowed myself to be guided by mere text- book statements, unless these on investigation have been found consistent with my own observations. Many may possibly depreciate my want of faith in statements which have been repeated until they appear to be almost incontrovertible ; but I am content to await the verdict of posterity on my con- clusions. Some of my views will, of course, be replaced by others, but I have the strongest belief that most of them will withstand adverse criticism, and will ultimately be accepted. If this is so, many of the old statements will disappear from future text-books as completely as the Vertebrate morphology of fifty years ago has vanished from those of to-day. INTRODUCTION. Although the study of the anatomy of insects was formerly cultivated with great assiduity by such men as Swammerdam, Malpighi, Lyonet, Straus Durckheim, and more recently by Leydig, it is a science which has never found a home in England; moreover, the improvements in the compound micro- scope, which have made a new epoch in biological studies, have rather checked than advanced the study of insect anatomy, by diverting investigation into what appeared for a time more fertile fields of research. In the last twenty years, however, many remarkable monographs have appeared in France, Germany, and Russia, and our knowledge has been greatly advanced ; yet, extraordinary as it may appear, there is no really competent treatise on insect anatomy, in English, of more recent date than Newport's article ' Insecta ' in Todd's Cyclopaedia (i8j6). If we wish to make ourselves acquainted with the researches of late years, it is necessary to consult many memoirs in several languages— Graber's work, ' Die Insecten,' to a certain extent, is available to those conversant with German, and Miall and Denny have given us an excellent memoir on the cockroach, which is, however, insufficient as an vi INTRODUCTION. introduction to the study of so highly modified a type as the Blow-fly. I have therefore followed the example of Straus Durckheim, and given a resume of the principles of anatomy, morphology, and histology, applied to insects generally in a series of introductory chapters and sections, in order that the most recent knowledge collected by many observers may be presented to my readers, and enable them to understand more completely the characteristic peculiarities of the very remark- able type I propose to describe, and its relations to other forms of insect life. The term ' Blow-fly ' is applied indiscriminately, not only to several species of a single genus, but to those of other genera and sub-genera. My own researches, which have extended over nearly a quarter of a century, have been chiefly made upon the most abundant form in this country, Calliphora ery- throcephala, and when I speak of the Blow-fly in the following pages, unless otherwise stated, my observations refer to this species only. During the past two years most of my researches have been made in my new laboratory at the Middlesex Hospital Medical School. I am indebted to my Demonstrator, Mr. G. C. Karop, for valuable help, and more especially for the trouble he has taken in correcting my proofs, and to Mr. A. W. Kappel, the sub- librarian of the Linnean Society, for the zeal he has shown in searching out and obtaining for me many of the works I have consulted. CONTENTS. Preface ----... Introduction - - ... . List of Plates ----.. I. Life-History of the Blow-Fly IL An Introduction to the Study of the Anatomy and Morphology of Insects - - . . III. On the General Characters of the Diptera and its Subdivisions, with a Description of the Type-form 'Calliphora Erythrocephala' IV. The Larva of the Hlow-Fly .... Sec. I. External Form and Segmentation „ 2. The Integument - - . . . „ 3. The Head and Mouth Armatuie ■ „ 4. The Respiratory Organs .... „ 5. The Cutaneous Muscle;^ .... „ 6. The AHmentary Canal ... „ 7. The Nervous System .... „ 8. The Sensory Organs and Peripheral Nerve Terminations „ g. The Imaginal Discs .... „ 10. The Ca-lom, Dorsal Vessel and Splanchnic System of Nerves - . . . . . Appendi.x to Chapter IV.— Methods of Study V. The Integumental Skeleton of the Im.\go Sec. I. General Characters of the Exo-skeleton in Insects „ 2. The Head Capsule .... a. General Morphology - - . . b. On the Nomenclature of the Sutures and Sclerites of the Head Capsule - - . . c. The Head Capsule of the Plow- Fly - 25 32 32 35 37 47 50 51 62 71 72 85 93 99 99 106 106 117 119 viu CONTENTS. ClIAl-TER l.AOB Sec. 3. The Exo-skeleton of the Proboscis - - - 127 a. General Morphology ■ - - -127 b. The Sclerites and Morphology of the Proboscis of the Blow-P'ly - - - -134 f. The Proboscis of the Immature Imago - ■ 148 d. Comparative Anatomy of the Proboscis in the Diptcra - • - - - • '49 Si£C. 4. The Thoracic Exo-skeleton - - - - 1 54 a. (Jeneral Morphology - - - - ' 54 b. General Description of the Thoracic Skeleton of the Blow-Fly - - - - - 166 c. The Sclerites of the Thoracic Skeleton - - 171 d. Details of the Exo-skeleton of the Legs - - 190 e. The Wings and Mechanism of Flight - - 198 Sec. 5. The Exo-skeleton of the Abdomen - - - 209 a. General Morphology .... 209 b. The Abdominal Skeleton of the Blow-Fly - - 210 Appkndix TO Chaptkr v.— Methods of Study - - - 211 VI. The Topographical Anatomy ov the Muscles and Vlscera ov the Imago - - - ■ - 215 VII. The Embryology of the Blow-Fly in the Egg - - 230 Sec. I. Early Changes in the Ovum subsequent to Fertilization and ^'ormation of the Blastoderm - - 230 „ 2. Formation of the Membranes and Primitive Band - 241 „ 3. Origin of the Archenteron, Somatopleure and Coelomic Sacs ------- 247 „ 4. The Dorsal Organ of Kowalevski - - - 254 „ 5. The Nymphoid Stage of the Embryo, and its Con- version into the Larva - - - 257 VIII. General Anatomy and Histology of the Blow-Fly - 262 Sec. I. Cells and Nuclei . - - . . 264 „ 2. The Parablastic Tissues . - . - 270 „ 3. Epithelia ------ 277 „ 4. Muscles and Nerves ----- 282 a. The Muscles - - - - - 282 b. The Nerves ------ 288 IX. The Development of the Nymph in the Pupa - - 292 Sec. I. Formation of the Pronymph from the Lai-va - - 296 a. Histolysis of the Larval Muscles - - - 297 b. The Histolysis of the Hypodcrmis, and the For- mation of the Paraderm - - - - 299 c. Relation of the Imaginal Discs to the Paraderm - 303 d. The Contraction of the Paraderm and Evolution of the Discs - - - - - - 305 307 311 CONTENTS. ix e. Histolysis of the Tiacheie of the Larva, and De- '*"^ velopment of the Trache.-e of the Pronymph - 306 / Histolysis of the Fat Bodies and other Larval Tissues ■--... g. Histolysis and other Changes in the Alimentary Canal --.... //. The Position of the Neuroblast and the Develop- ment of the StoniodcXum of the Pronymph - 313 Sec. 2. The Development of the Nymph - - - 315 a. The Position of the Imaginal Discs in the Pro- nynipli 315 /'. Development of the Integument of the Head and Thorax - - . . . .3,3 (. Changes m the Neuroblast and Development of the Peripheral Nerves - - . . e. The Pupa-sheath - - . . . f. Changes in the Alimentary Canal - g. Origin of the Mesoderm - - . . //. Development of the Dorsal and Sterno-dorsal Muscles - - . . . - \\b i. The Tracheal System of the Nymph - - 338 A The Dorsal Vessel and Ccelom - - - 340 Sec. 3. The Development of the Imago from the Nymph - ^44 Appendix to Chapters VI. to I.\.-Methods of Study - - 347 324 330 331 333 BiDLioc;K.\i.Hiios.-l to 3, p. . ; 4 to 10, p. 7 ; H to 15, p. 25 ; 16 to 27 PP- 32, 33 ; 28 to 30, ,,. 62 ; 31 to 35, p. 72 ; 36 to 41, pp. 99, ,00 • 42 to 48, p. 106 ; 49 and 50, p. , ,9 ; 51 to 71, pp. ,27, .28 ; 72 to 81 pp 154 155 ; 82 to 89, p. ,90 ; 90 to 95, p. 198 ; 96 to 115, pp. 230, 231 ; 116 to 121, p. 235 ; 122 to 136, pp. 262, 263 ; 137 to 139, p. 276 ; and 140 to 147 p. 292. ERRATA. rage .7. line 8, for 'hypoblast behind,' read ' hvpoblast behind the Ijlastopore. ' l^age 2,1, last line, for ' ,\vi.' read ' .\iv.' age O5, in footnote, for ' K. L. C read ' C. L. C I age 77, m footnote, /or ' Chironomus ' read ' Corethra.' LIST OF PLATES. I. The Alimentaky Canal of the Larva - - 52 II. The Nervous System of the Larva - - - 64 III. The Nervous System of the Adult Larva - - 70 IV. The Imaginal Discs of the Resting Larva— First Stage 80 V. The Cephalo-Thoracic Segments - - - - 112 VI. The Proboscis 01- THE Imago ■ - -130 VII. The Thorax of the Imago - - - - 168 VIII. Details OF THE Thoracic Skeleton - - - 172 IX. The Legs and Feet - - - - - 192' X. The Wings - - - - - - 200 XI. Male Rlow-Fly - - - ■ 216 XII. Embryology - - ■ - - - 232 XIII. Embryology - - - - 242 XIV. Embryo— Early Stage - - - 250 XV. Embryo— Late Stage - - - - 258 XVI. Histology - - - - - - 266 XVII. Histology - ■ - - 284 XVIII. Histolysis - - - - 302 XIX. Pronvmbh - - - - - 3>6 XX. Pronymbh and Nymph - - - 318 XXI. Nymph ---■-■- 342 CHAPTER I. THE LIFE-HISTORY OF THE BLOW-FLY. The life-history of the blow-fly may be conveniently divided into four stages ; (i) the egg ; (2) the larva ; (3) the pupa, and (4) the imago states. The Egg. — The female insect deposits her eggs in packets on those parts of dead birds or animals which are not covered by hair or feathers, or upon raw or even cooked meat ; and occasionally, it is said, upon the fleshy petals of certain plants (Stapclia, etc.) which have a carrion-like smell. The number of eggs in each packet varies between three or four and a hundred or more. The eggs arc usually fecundated at the moment of deposition, and the larva effects its escape from the egg in from twenty to twenty-four hours. At the end of the first twelve hours the embryo already possesses a rudimentary head with all those parts usually found in insects at a corresponding period of development. Bibliography.— In this and the following bibliographies it has not been my intention to include every book upon the subject, but only those to which 1 am indebted, or which possess historic interest and contain original work. The numerals prefi.xed are given in the text thus [8], when a work is quoted or referred to, and are consecutive. 1. Reaumuu, Df, ' Mdrnoires pour servir i I'Histoire des Insectes,' 4to., torn, iv., Paris, 1738. The fourth and seventh Mdm. are in part devoted to the life- history of the blow-fly, and are exceedingly interesting. 2. Weismann, A., 'Entwicklung der Dipteren.' Leipzig, 1864. .A.lso in Zeitsch. f. wissensch. Zool., Bd. xiii. and xiv. 1863-64. 3. Gl-F.iCHEN, W. F. F., ' Geschichte der gemeinen .Stubenflicge ' (The House-fly). Niirnberg, 1790. Is interesting on account of its antiquity, and gives an account of the life-history of the house-fly. I 2 THE LIFE-HISTORY OF THE BLOIV-FLY. During the remainder of the embryonic stage all the changes which occur are apparently retrogressive. The head is in great part withdrawn into the interior of the embryo, so that when the maggot emerges from the egg, the parts destined to form the head in the perfect insect are found deeply invaginated, and lie far back in the thoracic region, in front of the highly concentrated nervous system. The newly-hatched larva buries itself in the carrion on which the eggs are deposited, and feeds continuously for fourteen days, by which time it has attained its full growth. I believe in warm weather this period is considerably shortened, and that the full growth of the larva is attained in eight to ten days, but the time necessary evidentl}' varies with the temperature and with the condition and nature of the food, and it has been differently estimated by several writers. It is shorter in Musca csesar and Sarcophaga carnaria. The newly-hatched larva measures nearly 2 mm., oryV in., in length when fully extended. It sheds its cuticular integument within two hours after its escape from the egg, and I have actually observed this first moult in larvae hatched in a watch-glass. A change occurs in the form of the mouth armature and the structure of the posterior spiracles at the first moult. As a pair of anterior spiracles are subsequently formed, and further changes occur in the mouth organs and posterior spiracles, it must be inferred that other moults occur, but their number is unknown, as the newly- hatched maggot immediately buries itself in its food. Weis- mann concluded that the larva undergoes at least three. Burmeister [8] erroneously supposed that no moults occur in the larvae of the Muscidae. The full-grown larva when it ceases to feed measures |- of an inch in length when fully extended, and weighs lo to i2 centi- grammes, or about i'5 grains. It leaves the carrion in which it has been nourished and buries itself in the earth. It is not im- mediately transformed into a pupa, but becomes a resting larva. The Resting Larva. — The duration of the resting period varies greatly with the temperature : it may not exceed two days, and may extend to several weeks or even months. THE LIFE-HISTORY OF THE BLOW-FLY. 3 When the larva ceases to feed its crop is greatly distended, but during the resting period it is gradually emptied with the rest of the alimentary canal ; the head is withdrawn within the second annulus, and the whole worm is contracted, and assumes an ovoid form. If disturbed, the head is protruded, and the larva crawls about rapidly, seeking to bury itself again ; after a variable period, however, the contraction of the body becomes permanent, all power of movement is lost owing to the inner layer of the integument and the muscles of the larva having detached themselves from the external cuticle : this is the com- mencement of the pupa state. It has long been known that a number of curious cellular bodies exist connected with the nerves and nerve centres of the larva. These were formerly regarded as ganglia, but Dr. Weismann [2] discovered their true nature. ■ They are the rudiments of the fly ; he named them imaginal discs. It has also been suggested that the largest of these imaginal discs is the invaginated portion of the head of the embryo. Quite recently I was fortunate enough to make longitudinal sections of a newly-hatched larva, which not only demonstrate the fact, but also show the exact nature of the invagination (see Chap. IV., Sec. 3, Fig. 7). During the resting stage two processes are going on simul- taneously, the various larval organs are undergoing rapid dis- integration, and the imaginal discs are unfolding and increas- ing in size and complexity. The Pupa Stage. — In warm weather at the end of two, or at most three days, and in cold weather often after the lapse of several weeks or months, the integument of the contracted larva undergoes a change of colour and texture. At first it turns yellow, then red, and finally black, and with this last change it also becomes hard and brittle ; the insect is now a pupa. The four anterior rings of the pupa-case are separated from those behind by a seam or raphe, and are readily detached as a cap in the fully-formed pupa. It is by this means the imago, or perfect insect, escapes from the pupa-case. If the pupa-case be opened just before it becomes black, I — 2 4 THE L IFE-HIS TOR V OF THE HL O W-FL Y. it will be found to contain nothing apparently but a white cream- like fluid ; but on careful microscopic examination some of the imaginal discs will be detected, and many of the muscles of the larva still remain at its posterior end. The imaginal discs are really very numerous ; fourteen were known to Weismann, and about fifty have been discovered since, besides many scattered groups of cells (histoblasts), from which the internal organs originate ; so that there are in all more than sixty separate discs, which subsequently unite with each other, and form the embryonic fly within the pupa-case. This embryonic fly is known as the nymph. The Nymph corresponds with the chrysalis of a lepidopterous insect, and has its wings, legs, and proboscis folded mummy- fashion over its ventral surface. It is entirely formed from the imaginal discs. All the organs of the larva, without exception, undergo disintegration, and are converted into a creamy fluid — the pseudo-yelk, by which the imaginal discs are nourished. The nymph is also developed in a very peculiar manner. Its head is at first enclosed within its thorax, and its thorax is enclosed within its abdomen. Subsequently, the abdomen is drawn back, exposing the thorax, and the thorax is drawn back, exposing the head. The Imago, or perfect insect, is the fully developed nymph, and escapes from the pupa-case at the end of from twelve to fourteen days in summer; but in winter the pupa stage may last for months, as all development is arrested by a temperature below 45"" Fahr. The preservation of the species in winter depends mainly upon this circumstance. The fly escapes from the pupa-case by pushing off the operculum, or cap. This is effected by the distension of a large bladder-like organ on the insect's forehead— the frontal sac. When it first emerges it is of a pale ash-gray colour, and very soft ; its wings are moist, thick, and crumpled, and the large frontal sac projects from its forehead. The proboscis and many of the internal organs are in a half-developed condition. In two or three hours the newly-escaped insect, which rapidly increases in size by dis- THE LIFE-HISTORY OF THE BLOW-FLY. 5 tending its respiratory sacs with air, assumes a dark colour, and its integument becomes hard and elastic. Its wings are fully developed, and it rises in the air and takes its first flight. By this time the frontal sac is permanently withdrawn into the head, and the external form, characteristic of the mature insect, is attained. There are still, however, traces of im- maturity in the coloration of the integument, and the imperfect hardening of the head and thorax, which is not complete for several days. A few hours appear to suffice for the full sexual development of the male, though three or four weeks are needed for that of the female — a condition which probably prevents the fertiliza- tion of females by males of the same brood. The female may lay several hundred eggs, but these cannot all be deposited at the same time, as the ovaries contain four or five sets in different stages of development. About 180 eggs are matured at one time, three or four times as many remain- ing in a rudimentary condition. The female is fecundated but once ; the eggs are usually fertilized as they are deposited ; but two, or possibly three, fertilized eggs may be retained in the oviduct, so that under exceptional circumstances one or more living larva may be deposited by the female fly. Many species of allied genera normally retain the fertilized eggs in a special ovisac until they hatch, and living larvae are deposited instead of eggs. Saixophaga, Scatophaga, and Tachina are examples of such viviparous flies. The female blow-fly usually exercises discrimination in the deposition of her eggs, and the number laid apparently bears a proportion to the mass of carrion. I have repeatedly seen these insects examining a portion of flesh on all sides, and if eggs in sufficient number are already there, rejecting it as unsuitable for their purpose. The males seldom enter houses e.xcept in cold weather, and are usually found on flowers ; impregnated females frequent carrion and are found in houses ; but those with ripe eggs are seldom seen in 6 THE LIFE-HISTORY OF THE BLOW-FLY. our dwellings until late in autumn, although they abound in butchers' shops, slaughter-houses, and similar places. The question of hibernation is a difficult one to settle. All ' winter flies ' are, I believe, immature, as I have never found one with ripe eggs ; still, I strongly suspect mature females hibernate occasionally. In the winter of 1889 I had a bell-glass in my laboratory full of immature blow-flies, which had recently issued from the pupae. One very cold morning when I went in not a fly was to be seen. I imagined the bell-glass had been lifted, and that my flies had escaped. On closer investigation, however, I discovered them all closely huddled together in a hollow under the base of a cup containing pupte. The cup had been slightly tilted, and every insect had retreated into this narrow space. As soon as the temperature of the room rose to 55° Fahr., the insects emerged from their concealment, and were as lively as ever. TUB FLESH-FLY. I^Sareophaga caniaria. ) CHAPTER II. AN INTRODUCTION TO THE STUDY OF THE ANATOMY AND MORPHOLOGY OF INSECTS. The body in all insects may be regarded as a simple thick- walled tube, the cavity of which forms the alimentary canal. There is no distinct continuous body-cavity, in which the viscera lie, corresponding to the pleuro-peritoneum of a verte- brate, or the continuous coelom of a hollow-bodied worm (nematoid) ; but all the organs are connected together by a delicate sustentacular tissue, formed of branching cells, in the Bibliography.— The general subject of insect anatomy and morphology. 4. SwAMMERDAM, ' Bybel der Natuure.' Utrecht, 1669; Leyden, in Latin and Dutch, Boerhaave's edition, 1738 ; and Leipzig, 175-- 5. Reau.mur, De, ' Mdmoires pourservirh I'Histoire des Insectes.' Paris, 1734-42, 4to. 6. De Geer, ' Mdmoires pour servirh I'Histoire des Insectes.' 410, Stock- holm, 1752-78. Not an anatomical work in any sense, but often quoted ; a treatise on systematic entomology. 7. Fabricius, J. C, 'Philosophia Entomologica.' llamburgi et Kilonii, 1778. A curious little book, which is useful in regard to nomenclature of parts, but of no use as a treatise on structure. His ' Systema Ento- mologiic," Flensburgi et Lipsi.ie, 1775, is a work on systematic ento- mology. 8. BuRMEiSTKR, ' Handbuch der Entomologie.' Berlin, 1832, 8vo. Translated into English by Shuckard, with additions by the author and original notes by the translator. London, 1836, 8vo. 9. Newport, C, Article Mnsecta,' Todd's ' Cyclopa."dia of Anatomy and Phvsiology,' 1836-39. London. Still the best work in English on insect anatomy. 10. Graber, v., ' Die Insecten.' Munich, 1877, small 8vo. The best account of the anatomy and development of insects in a popular form. 8 yJXA TOM \ ' AND .^rORPHOLOG Y OF INSECTS. meshes of which the blood circulates, although there are special cavities, blood, or more properly lymph, sinuses in various parts of the body. The blood of insects is in fact equivalent to the lymph of vertebrates, which it closely re- sembles in its physical and microscopic characters. The somatic nervous system* consists of two parts— a cephalic nerve-centre or brain, which lies in front of the mouth ; and a double ganglionated cord, which is situated on the ventral aspect of the alimentary canal. The dorsal vessel, or heart, is placed immediately beneath the skin of the back, hence the ventral aspect is known as the neural, and the dorsal as the hjemal surface of the animal. The whole body is segmented transversely into a number of annuli or somites, and these are grouped, in the adult form or imago, into three distinct regions, known respectively as the head, thorax, and abdomen. The somites of the head and thorax usually each possess a pair of ventral appendages, jaws or limbs, and one or two of the thoracic somites are generally provided with dorsal appendages— wings. A somite with its appendages is spoken of as a metamere ; the terms ' somite ' and ' metamere ' are synonymous when all appendages are wanting. The external surface of the body is covered and protected by a cuticle, impregnated more or less completely by a horny substance called chitin, which renders it very dense and elastic, and usually forms a complex exoskeleton. The cuticle often consists of many super-imposed laminae, and is formed by an underlying layer of cells known as the hypodermis, which represents the external cellular epithelium of the mollusca and the vertebrata, and which corresponds with the epiblast of the embryo (see Fig. i). The cells of the hypodermis are usually short hexagonal prisms, cemented together by their edges. In the larva the size of these cells increases with its growth, and in the adult larva they are generally very much larger than the correspond- * See Chap. IV., Sec. 7. ANA TO MY AND MORPHOLOGY OF INSECTS. 9 ing cells of the nymph or imago — at least, in the metabola and hemi-metabola. Chitin is a nitrogenous substance more nearly related to mucin than to any other substance found in vertebrates. The large hypodermic cells of many larvae exhibit cup-shaped cavities on their outer surface, which present a great simila- rity to those of the well-known goblet cells of the mucous membranes. Latreille regarded chitin as the result of the degeneration of the external portion of the cell-substance Fig. I. — Sections of the skin of the lilow-fly laiva. — i, a vertical section showing the hypodermis /;, with the super-imposed ciilicular layers ; 2, a similar section, showing the cuticular prisms ; 3, a sensory papilla ; 4, a sub-hypodermic cell. cu. Cuticle ; c, nerve end organ ; /, terminal portion of the end organ ; «, nerve; s, sub-hypodermic tissue ; ir, trachea. [9, p. 882], a view in which I must concur. Chitin is very insoluble in solutions of the caustic alkalies, a property which enables the microscopist to make beautiful preparations of the exo- and endo-skeletons of insects. The cuticle consists of two distinct parts, which correspond with the epiostracum and endostracum of the Crustacea.* I shall, therefore, use these terms to distinguish them. * Huxley, T. H., 'The Crayfish.' London, 18S0, p. 192. lo ANATOMY AND MORPHOLOGY OF INSECTS. The Epiostracum consists of a single external lamina, which differs considerably from the rest of the cuticle in the readiness with which it takes stains. It is apparently quite structureless, or is divided into hexagonal fields. The Endostracum is formed of numerous lamellae, except in the hard plates of the exo-skeleton, which do not usually exhibit a laminated structure. It is often divided by vertical planes into hexagonal prisms, several corresponding to a single hypodermic cell. In the imago, when it first emerges, the endostracum is not usually present except where sclerites are already developed. Sclerites. — This term is applied to the denser cuticular struc- tures which form the exo-skeleton. The hardness and thickness of the several parts of the cuticle are by no means in direct relation to each other. The denser plates are usually thinner than the membranous parts of the cuticle between them. Thus in the imago of the blow-fly the hardest dermal plates vary from 10" to ao'' in thickness, whilst that of the intervening integument in some parts is loo". The flexible portions of the integument are usually transparent and colourless, whilst the sclerites are deeply pigmented and opaque. Syndesmotio Membrane.^ — -The soft flexible cuticle between the sclerites has been termed conjunctiva, but, as the articulations which it forms are known as syndesmoses, I shall prefer the term syndesmotic membrane. Scales, Setae, and Vibrissae are solid or hollow projections of the cuticle. The large scales and setae are hollow, with a giant cell beneath each, the trichogenic cell. It gives off a process which fills the interior of the hair or scale. The large scales and setae are usually articulated with the cuticle of the surface, which forms a kind of socket around the opening or foramen, through which the process of the trichogenic cell enters the seta. The smallest sets are solid, the larger ones are ribbed ; some are deeply channelled on one side. The larger bristles (seta;) of the Diptera are remarkably persistent in number and position throughout very large families. Osten-Sacken regards their arrangement, chcetotaxy, AA^A TOM Y AND MORPHOLOG Y OF INSECTS. 1 1 as important in relation to classification.* The very early development of the setse in the nymph is an indication in favour of his view. The Diptera generally, and the Muscidae more especially, exhibit two very distinct modifications, which Osten-Sacken calls the Diptera chaetophora and the Diptera eremocheta. The former are densely covered with bristles, and use their legs to run, climb, and snatch their prey ; their flight is headlong, and they rarely or never poise over flowers. The olfactory sense appears to predominate. On the other hand, the latter hover and poise on the wing. They are smooth and have no bristles, or very few. Many are brightly coloured. The visual sense appears to predominate, and they only use their legs to alight and when resting. They are helpless in the dark, and are seldom abroad except on sunny days. Dr. A. Forelf says that ' insects organized for an exclusively aerial life depend on their eyes. They generally have little developed antennae, and are absolutely helpless in the dark ; they hardly dare to walk. . . . In other insects the eyes play a very subordinate part. These may be called antennal insects. They can work by night or underground as well as by day.' The blow-fly belongs to the chastophorous division, and is in some degree an antennal insect in Forel's sense, but it has evidently good visual powers. Apodemes arc rod-like involutions of the cuticle, to which muscles are attached. They are frequently very strong, and act as powerful levers. The Thorax in the imago is always provided with three pairs of jointed limbs. The anterior pair are sometimes rudimentary. It consists entirely or in great part of three metameres, known respectively as the prothorax, the mesothorax, and the meta- thorax. Tracheae. — Insects always possess a peculiar respiratory * Osten-Sacken, C. R., 'An Essay on Comparaiive Chatotaxy.' Trans. Entom. Soc, London, 1884. Originally primed in iMilth. der Miinchener Entom. Vereiiis, lid. v., 1881. t A. Forel, ' Beitra835- 17. Schroder, van der Kolk, 'Mdmoire surl'Anatomie et la Physiologic du Gastrus equi,' avec 13 pis. Verhand. d. Kl. Nederl. Inst., D. ii., p. 1-155. 1845. 18. DUFOUK, Li'iON, ' Etudes Anatomiques et Physiologiques sur les Insectes Uipt6res de hi famille des Pupipares.' Ann. Sc. Nat. Zool., ser. iii., torn, iii., 1845. 19. DUFOUR, LliON, ' Recherches Anatomiques et Physiologiques sur les SEGMENT A TION. 33 Segments. — The integument is obviously divided by circular sulci into a series of rings or segments. If the small head be considered the first segment, the second, third, and fourth obvious segments belong to the thorax, and the second bears a pair of pedunculated spiracles. There are no spiracles on the segments which intervene between this and the last, so that this larva is amphipnciistic (see page 28). Each segment has a thickened anterior border, covered by short recurved spines and sensory papillae. The spines apparently prevent a retro- grade movement in burrowing, like the setae of an earth worm. The abdominal segments exhibit a dorsal and a ventral arch, separated by a marked lateral fold, and the ventral arch is divided by a transverse sulcus into an anterior and posterior portion. Although it would appear at first sight perfectly easy to count these segments, authors are by no means agreed as to their number. Dr. Weismann [2] gives twelve, reckoning the Dipt{;res.' Mem. Pres. ?l I'Acad. des Sc. Math, et Phys., torn, xi., pp. 171-360. II pi., 1851. 20. Leuckart, R., ' Fortpflanzung und Entwicklung der Pupiparen.' Abh. Nat. Gesellsch., Halle, Bd. iv., 1S58. 21. SCHEIEER, S. H., ' Vergleichende Anatomie und Physiologie der Ostriden-Larven.' Sitzungbericht der K. Akad., Wien. Bd. xli., pp. 409-496, i860 ; and Bd. xlv., pp. 7-68, 1862. 22. Leuckart, R., 'Die Larvenzustiinde der Musciden.' Archiv. f. Naturgesch. Jahrg. 27, 1861. 23. Marno, Ernst, ' Die Typen der Dipteren Larven.' Verhand. der Zool. Bot. Gesellsch. in Wien. Bd. xix., 1869. 24. Hammond, A., ' The Anatomy of the Larva of the Crane-fly.' Sc. Gossip, vol. xi., pp. 10, 171, 201. 1875. 25. KiJNCKEL D'Herculais, Jules, ' Recherches sur I'Organisation et le Ddvcloppement des Volucclles.' Fol., Paris, 1S75-81. 26. Batelli, Andrea, ' Contribuzione all' Anatomia ed Fisilogia della Larva dell' Eristalis tenax.' Bull, della Soc. Ent. Ital., 1879, pp. 77-120, with 5 plates. 27. VlAi.LANES, H., ' Recherches sur I'Histologie des Insectes et sur les Phdnom^nes histolyliqiie qui accompagnent lo ddvcloppement post- embryonaire de res animaux.' Ann. Sc. Nat. Zool., ser. vi., tom. xvi., pp. 1-348. l8 pis., 1882. 3 34 THE LARVA OF THE BLOW-FLY. head as the first ; Schiner [11], as we have seen, gives thirteen as the general number in the dipterous larva. Weismann apparently overlooked a very obvious segment, between the head and the first thoracic segment. This is usually invagi- nated within the first thoracic segment, so that it cannot be seen except when the larva is forcibly extended ; but New- -e c Fig. 4.— The Adult Larva of the Blow-Fly. a, Lateral view ; b and C, the head of the same in extension and flexion — li, the great hooks ; si, the stomal disc ; mx, the maxilla ; /, the forehead ; a s/>, the anterior fan-like spiracle, with two lenticular bodies in front of it ; d, the posterior extremity of the same : e, the last two segments, showing liie anus (a) flanked by a pair of false feet. The sub- semicircular plate in front of the anus is the ventral plate of the anal {I5lh) segment. port [9] correctly described it. I shall speak of it as Newport's segment. Brauer [14], like Weismann, gives twelve, but believes— and I think correctly— that the last segment is a complex of two. SEGMENTA TION. 35 Although it is convenient to speak of the head as the first segment, it represents the three first post-oral metameres of the embyro. I regard Newport's segment as the last of these cephalic metameres ; the second is only represented by its ventral appendages, the maxillae, which form the greater part of the head of this larva ; and no traces of the first cephalic metamere can be distinguished externally, at least. Adopting these views, the first thoracic segment is the fourth post-oral metamere, and is so numbered in my figures. The three cephalic, three thoracic, and nine abdominal segments, counting the last obvious segment as two, give fifteen post-oral somites, of which the first is apparently suppressed. And the second is only represented by its ventral appendages, the maxillai. Formerly two pre-oral somites were added, making seventeen, a number which has usually been regarded as typical of all insects except the Orthoptera, which generally exhibit eleven abdominal segments. The question as to the morphology of Newport's segment is one of considerable interest, and will be further discussed here- after. The compound nature of the last segment is not, I think, doubtful, but is manifested by the existence of a pair of ventral appendages close to the anus, each consisting of three joints, with a small ventral plate in front of them (Fig. 4, d and e). The concave disc, which contains the posterior spiracles, is fringed with papillae, of which four are large and eight smaller ; five are situated on each side, and two below. 2. THE INTEGUMENT. The Integument is thick and leather-like in the adult larva ; its cuticular portion consists of two distinct parts : the epiostra- cum and the endostracum. The Epiostracum forms a continuous layer over the whole ex- ternal surface, and is inflected into the mouth, anus, and spiracles. It forms the solid recurved spines on the anterior portion of the segments, and the muscles are inserted into it 3—2 36 THE LARVA OF THE BLOW-FLY. by means of fibres of a tendinous character, which run through the whole thickness of the endostracum (Fig. i, /). The Endostracum consists of many laminae, and is more or less distinctly divided into hexagonal columns, several corresponding to each of the cells of the hypoderm. It is perforated by numerous canals, which contain processes from these cells (Fig. I, / and 2) and nerve end-organs, which terminate in the sensory papillae (Fig. i, J, f), immediately beneath the epiostracum. Fig. I (bis). — Sections of the skin of the Blow fly larva.— i, a vertical section, show- ing the manner in which the muscles are inserted into the epiostracum ; 2, a similar section, showing the prisms of the endostracum ; 3, a sensory papilla ; 4, a suh- hypodermic cell, with a terminal trachea in its interior, cti. Cuticle ; c, end org.m ; //, hypoderm cells ; i, sub-hypodermic cells ; /, terminal portion of the end organ ; tr, trachea. Sensory Papillae. — The surface of the integument, especially near the spiracles and on the anterior edges of the rings, is raised into papillae, which consist of both epi- and endostracum. These papillae contain bundles of fusiform cells, which ter- minate in very fine rod-like organs in the epiostracum at the tip of the papilla, these are the terminal organs of cutaneous nerves (Fig. z., 3,1). THE INTEGUMENT. 37 Lenticular Bodies of a highly refractive character are found imbedded here and there in the endostracum, sometimes deeply, more usually immediately under the epiostracum. These bodies vary from -i mm. (o J-y of an inch) to mere granules. They are very hard and brittle and consist of concentric lamellae enclosing an internal cavity. Several are found near the head. They apparently resemble the so-called calcareous bodies of cestoid worms (Fig. 4, h). The Hypoderm consists of large flat hexagonal cells, 70" to 80" in diameter, and 50" or more in thickness. The nuclei of these cells are distinctly vesicular, 40" in diameter (see ' Histology') ; some are cup-shaped next the cuticular layers, and the cavity is filled by the rounded inner extremity of the corresponding cuticular prism. Others are conical and fill cavities in the cuticle below the projecting papillae on its surface. Many send fine processes through the canals of the endostracum (Fig. i, / and 2; h). Basement Membrane. — Several authors have described a thin cuticular basement membrane beneath the hypoderm cells. Such a membrane only exists just before the formation of the pupa. Sub-hypodermal Cells (Fig. i, /, s and 4). — Viallanes [27] first described a fenestrated layer of cells beneath the hypodermis, this is the connective tissue, or so-called peritoneal coat, which is reflected from the surface of the internal organs over the inner surface of the hypoderm ; it supports the tracheal net- work. The individual cells are often stellate, and are con- tinuous with the adenoid reticulum of the blood sinuses. 3. THE HEAD AND MOUTH ARMATURE. The Head. — The only portions of the head which can be recognised externally are the maxilla;, separated on their ventral surface by the mouth, the prestomal and discal sclerites ; and on their dorsal aspect by the epicranium. The Maxillse. — The organs which I have so named form by far the largest portion of the head, and each consists of two parts separated by a slight circular sulcus — a proximal and a 38 THE LARVA OF THE BLOW-FLY. distal joint (Figs. 4 and 5). The distal portion is sub- hemispherical and has two short conical sensory papillae at its extremity, one above the other ; it contains a pair of sensory organs, which terminate in the papillae. These closely resemble the eyes of a leech, except in being devoid of pigment ; they are apparently sensitive to light (see Sensory organs). The proximal portion is sub-cylindrical, and exhibits a very remarkable half-disc on its ventral and outer surface, the stomal disc (mihi), and in front of the stomal disc the orifice of a sac, in which the great hook lies when retracted. oT^^iS®^^ c Fig. 5. — a, the head and mouth of the idult hr\i tf the Ijlow-Fly seen from the ventral surface ; 3, Newport's segment ; 4, the prothor.icic segment ; b, the mouth more highly magnified (i inch objective) ; /, the prestomal sclerite ; ii, the stomal disc ; ni, aperture of the mouth ; //', the labium ; c, transverse sec- tion of the pseudo-trachea; of the stomal disc as seen with ,',t oil immersion lens. The real significance of these organs has been apparently overlooked by Weismann and most subsequent writers, except Macloskie* and perhaps Brauer [14, p. 32], who recognised that the mouth parts of the larva correspond with the proboscis of the imago, without entering into details. This is more remark- able, as the development of the maxillae is easily traced from the unmistakable maxillae of the embryo, from which they scarcely differ, and these were correctly figured and described by Weismann, who, however, entirely neglected the correspond- * See page 44. THE HEAD AND MOUTH ARMATURE. 39 ing organs in the larva, and contented himself by describing the head in the following manner: 'The first and smallest segment exhibits the mouth opening on its ventral surface and two pairs of papilla; on its dorsal surface ' [2, p. 105] ; and although he subsequently describes the mouth organs at con- siderable length, he does not again refer to the maxillaj, except to mention ' two thread-like thickenings in the recently-hatched larva, which spring from the angle of the mouth and run Y\Q. 6.— The head of the Cossus caterpillar after Lyonet. a, antenna ; tml, mandible ; mx, maxilla ; mx", the second maxilla ; /, ligula. obliquely towards the back ' [p. 108] ; and adds, ' the two chitinous ridges of the newly-hatched larva are replaced in the adult by a fan-like group of similar ridges, which diverge from the angle of the mouth and form a half circle at the side of the mouth opening' [p. 109]. These are his only descriptions of the stomal disc. He refers subsequently to the two pairs of papilla; as the antennas and maxillary palpi [p. 120]. A comparison of the maxilla; with those of a Icpidopterous 40 THE LARVA OF THE BLOW-FLY. larva (Fig. 6, mx) will assist the reader in attaining a clear idea of their relations. The Mouth presents a triangular orifice between the proximal joints of the maxillae ; the apex of the triangle is in front and its base is formed by the labium. The Stomal Disc exhibits a nearly semicircular surface, grooved by a number of radiating dichotomously-dividing channels, similar to the pseudo-tracheae of the proboscis of the imago, but without ring-like thickenings (Fig. 5, c). It is of considerable morphological interest, because the principal imaginal disc from which the proboscis of the fly is developed is formed by an invagination of its hypoderm. The channels on its surface open into a cup-shaped cavity on the oral edge of the disc or by a single vessel, which is formed by the union of those of the posterior third of the disc, into the external angle of the mouth. These channels apparently serve to distribute the salivary secretion over the food and also to conduct the food into the mouth. They always contain a con- siderable quantity of grumous fatty material, and their contents are blackened intensely by osmic acid. The Great Hooks are uncinate processes of the cuticular layer of the integument, and when at rest are retracted within special cavities in the maxillae (Fig. 8, j, h), which do not com- municate with the mouth. They are the retractile claws of the maxilla, and resemble the claws on the thoracic feet of some larvae. In the recently hatched larva the great hooks are very small straight chitinous rods, the distal extremities of which are bent at right angles (Figs. 7 and 9), and like the claws on other feet are shed and renewed at each ecdysis. Before the second moult the new and old hooks are occasionally seen side by side, a fact recorded by Weismann. Brauer [14, p. 33] calls these hooks maxillae, but also speaks of them as mandibles, and does not assert their morphological identity with either ; Menzbier* supposed them to lie when re- tracted within the mouth, and regarded them as indurations of * Menzbier, M., ' Uber das Kopfskelet und Mundwerkzeuge der Zwei- flugler.' Bull. Soc. Imp. Nat., Moscow, t. 55, 1880. THE HEAD AND MOUTH ARMATURE. 41 its cuticular lining without morphological significance. In my former work* I drew attention to the marked resemblance of these hooks to certain hook-like sclerites in front of the suctorial disc of the proboscis of the imago ; I did not then know how closely this portion of the proboscis is connected with the great hooks of the larva in its developmental history. These have only a secondary connection by articulation with the internal pharyngeal skeleton (Fig. g, ?) which supports them. They are used as organs of locomotion, and probably assist in the disintegration of the flesh in which the larva burrows. Before entering upon any further description of the head of the larva, I shall explain the relation of its parts from a de- FlG. 7' — A vertical longiludinal median section tlirough the anterior portion of a newly-hatched larva seen with \ inch objective. The external portions of the head are added to the actual section — in, invagination of the forehead ; //, lateral (great) hooks ; iii, median tooth {labruiit) ; /it, stomal disc ; sii, salivary um. As the pharynx is therefore exceedingly coinplex, I shall speak of the whole organ, the skeletal structures and the muscles, as the fulcrum ; that por- tion of the skeleton which belongs to the head-capsule I shall term the cephalo-pharyngeal skeleton, or, for brevity, the ceplialo-pharynx ; that portion developed in the wall of the stomodajum, the pharynx. The term ' fulcrum ' is now in general use ; but it is not absolutely correct, as it is applied to a part of the mouth of the Hymenoptera which has a different morpho- logical value. Between the cephalo-pharynx and the great hooks of the blow-fly larva there is a second sclerite, which is developed in relation with the orifice of the salivary duct and ligula ; it articulates behind with the cephalo-pharynx, and in front with the great hooks. This is the 'connecting piece' of Weismann. I shall term it the hypostomal sclerite, owing to its position in the floor of the mouth. Above the hypostomal sclerite, in each lateral wall of the mouth, there is a short rod of chitin, which abuts upon the fulcrum behind ; I shall term it the parastomal sclerite. The Cephalo-pharyngeal Sclerite may be described as consisting of two vertical plates united by a thin dorsal arch. Each lateral plate exhibits three processes— an anterior inferior, a posterior inferior, and a posterior superior process. The latter I shall term the cornu. In the newly-hatched larva these processes are very slender (Fig. 9, /), but after the second moult they become exceedingly thick and strong. The cornua consist, in the adult, partly of dark and partly of transparent chitin ; they exhibit thick upper edges, which lie in close proximity, so that they enclose a cavity, which contains a blood sinus and the dilator muscles of the pharynx. I shall speak of this cavity as the pharyngeal sinus. The Hypostomal Scleritef is H-shaped in the adult larva ; its posterior processes articulate with the anterior inferior processes of the cephalo-pharynx, and its anterior processes support the great hooks. I have been unable to find any trace of this sclerite before the second moult. * Macloskie, G., in a preliminary note on the head of larval Musca;, held views which are very similar to my own. 'Psyche,' Vol. iv., p. 218, Cam- bridge, Mass., 18S4. t 'Zungcnbein' of Schroder v.d.Kolk. 'Mdmoiresurl'Anatomieet Physiologic du Gastrus Equi.' Verhand. d. kl. Nederl. Inst., Bd. ii., pp. 1-155, 13 pi., 1845. THE HEAD AND MOUTH ARMATURE. 45 The Labral Sclerite of the newly-hatched larva (Figs. 7, m, and 9, .', m) articulates by a pair of processes with the anterior inferior processes of the cephalo-pharynx ; these form the edges of the true mouth. The Great Hooks (Fig. 9, j>, h) present three processes : a head which articulates with the hypostomal sclerite, a coronoid process above, which gives insertion to the retractor muscle, and a discal process below, which is adherent to the centre of the stomal disc. In the adult larva a pair of small sclerites are found lying side by side in front of the crossbar of the hypostomal sclerite ; they are the points of inser- tion for the retractor muscles of the labium. Fig. 9. — I, the cephalo-pharynx and mouth armature of a newly-halched larva ; 2, the mouth armature of the same, seen from its ventral surface ; 3, the same parts from the adult larva seen in profile ; 4, a diagrammatic section through the thickest part of the cephalo-pharynx ; 5, the head capsule of Precilostola, after Brauer ; 6, the internal head sUcleton of the Cossus larva, after Lyonet — in, median ; h, lateral hooks ; //, pseudo-tracliea; ; //, pseudo-labrum and discal sclerites ; x, parastomal y, hypostomal, sclerites ; r/, clypcus ; :' and c", Zopf- griiten or ceplialo-phaiyngcal processes ; t\ lateral plate ; ///(, pharyngeal sinus and muscles of the epipharynx. The pharynx (Figs. 8 and 9, 4) lies between the inferior processes of the cephalo-pharynx. In transverse section its cavity appears nearly cre- scentic. The concavity of the crescent is filled by the pharyngeal sinus ; its lower wall is formed by a plate developed in the cuticle of the alimentary tube, the hypopharynx. The upper surface of the hypopharynx exhibits a series of longitudinal T-shaped ridges, which project into the oesophageal tube. 46 THE LARVA OF THE BLOW-FLY. The dilator muscles of the gullet are inserted into the middle of the lower wall of the pharyngeal sinus, the epipharynx, and by their contraction dilate the alimentary tube, so that the food is sucked into it. When they relax it is depressed on the h3popharynx by its elasticity, like the plunger of a pump ; by this means the food is driven into the crop. As there are no valves to the pharyngeal apparatus, I suspect the plunger itself acts as a valve, its anterior portion rising and descending before the posterior portion, so that the food is swallowed by a Fig. 10. — A vertical and horizonlal diagiammatic section of the larva, to show the relative position of the internal organs and the segments. The segments are numbered on the supposition that the three cephalic post-oral somites exist, as, anterior spiracle ; //;, pharynx ; id, imaginal discs ; 'mJ, wing disc ; m, nerve centre ; jective ; 5, tracheal terminations^, peritraciieal cells (peritoneal coat) ; tr, trachea ; v, vestibule ; s, slits ; h, nuclei of peritracheal cells ; e, mesoblast cells ; pr, fibres of the vestibule^,?-, grating of the slits ; ul, imaginal disc of the anterior spiracle (upper prothoracic disc). Leydig* correctly maintained that these capillaries are de- veloped in the interior of the cells of the connective tissue, but Wistinghausenf has quite recently described them as between * Leydig, Fr., ' Zum Feineren Baudes Arthropoden.' Miiller's Archiv., 1855. t Wistinghausen, ' Ober Tracheenendigungen in den Sericterien der Raupen.' Zeitsch. f. \v. Zool., Bd. 49, p. 565. 48 THE LARVA OF THE BLOW -FLY. the cuticular tunica propria and the cells in the silk-glands of caterpillars, a view from which I must dissent ; from his figures I conclude that his tunica propria is the peritoneal coat, and that the capillaries do not perforate the basement membrane as he supposes. I have never found any instance of such a dis- position ; they are everywhere confined to the mesoblast, and neither perforate the myolemma of the muscle fibres nor pene- trate between the epithelial elements of the hypoblast or epiblast. No organs undergo so many changes during the life of the larva as the arborescent tracheae ; in the newly-hatched insect they are few, but increase rapidly in number as growth pro- gresses, until at length every organ, except the fat bodies, is richly supplied with a network of tracheal capillaries. Before the first moult the main trachea; communicate with the exterior by a single pair of stigmatic slits on the fourteenth segment. After this, at first two, and then three, such slits are found in each posterior spiracle, and a second pair of spiracles appear on the fourth or prothoracic segment, at the second ecdysis. It is usual to speak of the larva; of the Muscidse as being in the first, second, or third stage of development in relation to the number of slits in the posterior spiracles. The Structure of the Tracheae (Fig. ii, 4). — The main trunks and larger vessels exhibit an external coat of thin polygonal cells closely united by their edges. Indeed, in some prepara- tions the boundaries of the cells are not visible, so that the trachea; appear to be covered by a continuous layer of nucleated protoplasm, and were so described by Weismann [2, p. 117]. This appearance is, however, delusive, and the edges of the individual cells are perfectly distinct in properly preserved preparations. These cells do not exhibit indications of division, and increase in size with the growth of the larva, like the cells of the hypodermis. In the largest tracheal trunks of the adult larvae they measure from 60" to 80" (^^ inch) in diameter. In- ternally to the cells there is a thick cuticular intima, with a distinct spiral structure, which gives the vessels their well- known appearance. Sections do not show a spiral thickening of the cuticular membrane. THE RESPIRATORY ORGANS. 49 In the smaller tracheae the intima is apparently structureless, and these end in capillary vessels of from 2" to 3" in diameter in the interior of stellate or fusiform mesoblast cells (Fig. 11,5). The development of the finer tracheae in the interior of cells attached to the external coat of the larger vessels was first observed by Weismann [2, p. 220]. It will be seen that there is a striking similarity between the manner in which the tracheal capillaries of insects and the blood capillaries of vertebrates are developed by intracellular vascularisation. Whether the cuticular intima is continued into the smallest tracheal vessels is doubtful, and the branching corpuscles in which these terminate have been, I believe, frequently de- scribed as plexuses of ganglion cells. They are not readily distinguished when they do not contain air, and the air is rapidly absorbed from them after death by the blood of the insect. They are more easily stained than nerve corpuscles by aniline dyes, and retain the colour longer when washed in water or dilute spirit. Wistinghausen {I.e.) speaks of these stellate cells as a terminal capillary network {Endnctz), and regards all the cells as vessels. The Posterior Spiracles of the adult larva are situated in a pair of nearly round chitinized plates of a dark-yellow colour. Each presents three oblique transverse slits, partially closed by a fine chitinous grating. A vertical section through the stigmatic plate at right angles to the slits is represented in Fig. 11, / and 2. It will be seen that each stigmatic plate is surrounded by a ring of chitin, the peritreme (Fig. 11, pi), which involves nearly the whole thickness of the external cuticle. The inner surface of the cuticular epidermis is covered with a cuticular network of fine fibres, extending across the interior of the slits and forming the grating (Fig. 11, 2). Immediately within the spiracle there is a distinct cavity {v), lined by a similar cuticular network {pr) ; I shall term this cavity the vestibule of the trachea. Its intima resembles the external cuticle, and is quite unlike the proper intima of the tracheal vessels which open into it. The tracheae are usually regarded as derived from a tubular 4 50 THE LARVA OF THE BLOW-FLY. involution of the epiblast. This does not appear to me to be the case ; it is only the vestibule which is so formed, the tracheal trunks arising in the mesoblast as solid cell strings, a fact known to Weismann [2, p. 76]. This mode of development is more consistent with the fact that in some aquatic larvse there are no spiracles, the tracheae being closed. The changes of form in the posterior spiracle which occur, when the larva moults, are not due to a modification of the spiracle, but to the formation of a new stigmatic plate on the outer side of the old one, from the hypoderm cells of the vestibule. The intima of the trachea separates from the external cellular layer, and a considerable space appears between them, which is filled with fluid. A new intima is then formed externally to the old one, the latter is afterwards shed with the stigmatic plate and the other cuticular structures, and the fluid is either absorbed or discharged. The Anterior Spiracles (Fig. n, j) are developed before the second moult, and can be recognised beneath the epidermis of the very young larva as papilla; on the prothoracic segment. In the adult larva the anterior spiracle is fanlike : it presents about thirteen minute orifices. These communicate by short tubes with the tubular cavity of the spiracle, which is apparently entirely filled with minute granules, forming a close yellow spongy mass. I am inclined to regard this spiracle as func- tionally inactive, and it is difficult to understand how it could be otherwise, as it is usually buried in the decomposing fluid on which the animal feeds. Its base is surrounded by the imaginal disc from which the spiracle of the nymph is developed (Fig. 11, 3, id). 5. THE CUTANEOUS MUSCLES. The somatic or skeletal muscles of insects are all cutaneous. In the larva of the blow-fly they closely resemble those of other vermiform larvae, and form a tolerably continuous sheet beneath the hypodermis. Two principal sets are easily distinguished ; transverse fibres which extend from the lateral line (Fig. 4, a) towards the dorsal and ventral surface of the worm, some THE ALIMENTARY CANAL. 51 passing obliquely forward and others backward, and a set of dorsal and ventral longitudinal muscles, the dorsal and ventral recti. These two sets are antagonistic to each other, the transverse fibres diminishing the diameter and increasing the length of the larva, and the recti drawing the annuli together. The individual fibres and bundles of cutaneous muscles would need many pages for the description of their origins and insertions, and the details possess little or no interest. Lyonet has already described the similar muscles of the caterpillar of the goat moth ; I shall, therefore, content myself with a very brief resume. •Generally all the muscles are attached to the integument where more or less obvious sulci are apparent externally, and the muscles of each annulus are repeated in the others ; most of the fibres pass from the edge of one annulus to that of the next, but some pass over two or more without being attached to the integument. The muscles of the ccphalo-pharyngeal skeleton must be regarded as an inflected continuation of the cutaneous sheet ; the fibres and bundles of fibres are chiefly the continuation of the recti : these form special retractors and protractors of the fulcrum, or act upon the labium and the great hooks. The great retractors of the fulcrum arise from the anterior edge of the sixth somite, whilst the protractors arise from the forehead and maxillffi. This fact is a complete refutation of the view held by Hammond,* that the delimitation of the several somites can be determined by the insertions of the somatic muscles. In structure the larval muscles generally differ from those of the imago, and they more nearly resemble the skeletal muscles of vertebrates ; a detailed description will be given in another chapter. 6. THE ALIMENTARY CANAL. a. Comparative Morphology. The Alimentary Canal in insects consists of four parts : the stomodaeum and proctodjeum, developed from the epiblast ; * 'On the Thorax of the Blow-fly." Journ. Linn. Soc. Zool., vol. xv., 1879. 4—2 52 THE LARVA OF THE BLOW-FLY. and the mesenteron and metenteron {niihi), developed from the hypoblast. The terms which have been applied to its several sections are such as indicate incorrect homologies with the alimentary canal in vertebrates. The stomodseum consists of the pharynx, oesophagus, crop, and of a portion of the proventriculus. The Crop. — This is the sucking stomach of Weismann ; it is more properly termed a food-sac. In manducatory insects the term is applied to a uniform enlargement of the oesophagus, which in phytophagous Orthoptera and Coleoptera frequently occupies nearly the whole thorax and a considerable portion of the abdomen. In suctorial insects the crop is a diverticulum of the ventral surface of the oesophagus, usually with a long tubular neck. The neck of the crop often appears as the direct continuation of the gullet, whilst the second part of the oesophagus ascends almost at right angles to the first, and enters the proventriculus. In the imago in the Diptera, Hymenoptera, and Lepidoptera the crop is the well-known honey-bag; it is situated in the abdomen, and when distended with food occupies a considerable portion of that cavity. The Proventriculus of suctorial, corresponds with the gizzard of manducatory insects ; it is usually a thick-walled, almost Description of Plate I. Fig. t.— The Alimentary Canal of the Adult Larv.i, 3inch objective : fv, proventri- culus ; eg, ci^cal glands; r//, chyle stomach ; /, proximal, and i\ distal intes- tine ; m/>, Malpighian vessel ; r, rectum ; ir, tracheae. Fig. 2.— Section of a Portion of the Crop, Jinch objective : ai, cuticle ; ffi, epithelial cells ; />(, peritoneal coat ; g, ganglion cells. Fig. 3. — Proventriculus : mn, median, or stomogastric nerve ; ^', proventricular ganglion ; fg, gastric cxca ; c/:, chyle stomach. Fig. 4.— a Longitudinal .Section through the Proventriculus: a; oesophagus; /r, trachea ; i d, imaginal disc ; fn>, proventriculus ; cli, chyle stomach. PIQ. 5. — Transverse Section through one of the Gastric Ca*ca, .{-inch objective. Fig. 6. —Transverse Section through the Chyle Stomach ; ( n, cell nests ; ^'C, parietal cells. Fig. 7.— a Transverse Section through the Distal Intestine, J-inch objective : c, amoe- boid cells. Fig. 8.— Transverse Section of a Malpighian Tube, J-inch objective. Fig. 9.— Transverse Section of a Salivary Gland of Resting Larva, showing the formation of the Membrana Propria. THE AI.IMENIARY CANAI. OF THE I.ARVA. THE ALIMENTARY CANAL. 53 spherical organ, formed by the invagination of the posterior extremity of the stomodaeum in the anterior portion of the mesenteron. When it assumes the form of a gizzard, the oesophageal portion is lined by a series of radially symmetrical chitinized folds, which form an efficient organ for the trituration of the food. The form of the proventriculus in the fly is typical of that in all suctorial insects. The Mesenteron consists of the chyle stomach and its append- ages, and of that portion of the intestine in front of the Malpighian tubes. Some authors have named the whole of this part of the alimentary canal the chyle stomach. Although there is no distinct line of demarcation between the anterior and posterior portions of the mesenteron in many insects, the two sections of the tube differ considerably, and in some forms are distinctly divided by a pyloric sphincter into a chyle stomach and an intestine. I shall therefore use the term proximal intestine for the posterior portion of the mesenteron. The Metenteron begins in front of the orifices of the Mal- pighian vessels, as a distinct thin-walled dilatation of the intestine (PI. I., Fig. 1), the lower part of which forms a large diverticulum or caecum in many insects. Beyond this the intestine is much narrowed, and exhibits a thick muscular coat. I shall term it the distal intestine, and shall speak of the dilatation as the sinus. This distal intestine is termed the ileum or colon by different writers. The rectum or proctodasal involution has a very distinct structure, and is always lined by a thick cuticular intima ; it is usually separated from the metenteron by a valve, which is formed by its invagination within the metenteron ; the valve is not unlike the proventriculus in the disposition of its walls. The Glandular Appendages of the alimentary canal are known as the salivary and gastric glands, the Malpighian vessels, and the rectal glands. The Salivary Glands. — This term is applied to all those glands in insects which discharge their secretion into the mouth or pharynx. The largest and most constant have a duct, which 54 THE LARVA OF THE BLOW- FLY. opens at the tip of or under the ligula. In many insects they secrete a true digestive fluid, with proteolytic or amylolytic ferments ; in others they are silk glands (sericteria). I shall term them lingual glands, a term which is equally appropriate whatever their function. The other salivary glands may be distinguished as accessory lingual, labial, oral, or pharyngeal glands, according to the position of the orifices of their ducts. Moseley regards the lingual glands as modified cutaneous glands, homologous with the trachea;. They are certainly developed as involutions of the epiblast at the root of the second pair of maxillse, but differ in their mode of development from the tracheae. The only reason for considering that they are homologous with the latter is the spiral structure of the intima of their duct. The Gastric Glands, or gastric caeca (PL I., Fig. 3, eg) are tubular glands arising from the walls of the chyle stomach. They are frequently very numerous, but as often entirely wanting. The Malpighian Vessels are simple or branched tubules, con- taining large cells, which give them a moniliform appearance. In some insects they are very numerous. In most, if not in all, Diptera there are two, each dividing into two tubes (PI. I., Fig. I, mp). I know no insect in which they are entirely absent. Malpighi supposed these tubes to be hepatic ; Leydig and most modern writers consider them renal. This view I shall show hereafter is not consistent with facts ; and, in spite of the weight of authority, I strongly incline to the theory of Malpighi. The Rectal Glands are situated in the proctodasum. They are frequently wanting, and apparently exhibit great variations in different insects. In the larva of the fly they are entirely absent. In the perfect insect I think they are undoubtedly concerned in the excretion of a substance allied to, but not identical with, uric acid.* * The functions of these and of the other internal organs will be fully dis- cussed in the physiological section of this work. THE ALIMENTARY CANAL. 55 b. The Alimentary Canal of the Blow-fly Larva. General Structure.— The wall of the whole alimentary tube exhibits a peritoneal, a muscular, and an epithelial coat. The Peritoneal Coat (PI. I., Figs. 2,pt, and 6, />) is the most external. It consists of stellate mesoblastic cells, endothelial plates, and tracheal vessels ; it forms part of the wall of a great blood sinus, which surrounds the whole alimentary tract. The Muscular Coat consists of two layers, an external longitudinal and an internal circular set of striated fibres. Those of the pharynx and rectum resemble the somatic muscles ; but the remainder of the muscular coat consists of flat branching bands of less distinctly striated fibres, without any myolemma, which do not exceed 8" to lo" in their broadest transverse diameter. The Epithelial Coat consists of flat polygonal cells in the stomodcEum and proctodseum, and of cubical or columnar cells in the rest of the alimentary tube. In the stomodseum and proctodiEum there is also a thick cuticular intima ; in the mesenteron this intima is either absent or much thinner, but the cells secrete a mucoid layer on their inner surface. In the larvffi of some Nematocera, however, a distinct cuticular lining can be seen in the Hving animal, separating the coarser particles of food from a clear fluid, which intervenes between the cuticle and the epithelium. Hammond [24] describes the same thing in the chyle stomach of the Crane-fly larva. It appears to me that the difference between the mesenteron and the stomodaeum and proctodseum does not consist in the absence of a cuticular layer in the former, but in its non- adherence to the cells. In the anterior and posterior sections of the gut the cuticular intima is only shed at an ecdysis, whilst that of the mesenteron is shed during digestion, and probably only forms a net-like mucoid envelope around the coarser particles of food material. Basement Membrane.— There is a very thin cuticular mem- brane between the epithelial cells and the muscular coat ; and 56 THE LARVA OF THE BLOW-FLY. these cells are firmly cemented together at their base, although they are quite separate towards their apices. The nature and origin of the basement membrane is more apparent in the lingual glands, where it forms the tunica propria. The (Esophagus commences at the posterior inferior part of the fulcrum, and extends backward, as a narrow cylindrical tube, to the posterior border of the ninth segment (Fig. lo). Close to its anterior extremity it gives off a diverticulum on its ventral surface, the neck of the crop, or, more properly, of the food sac ; it then passes between the crura of the supra- oesophageal ganglia or hemispheres, lies above the ventral ganglia, and terminates in the proventriculus. The length and diameter of the oesophagus and the position of the proventriculus varies with the condition of extension or contraction of the anterior segments of the larva. The figure represents a fully-extended larva ; in the contracted state the proventriculus lies close behind the ganglia. The Crop. — The neck of the crop has about the same diameter as the oesophagus. After a short course towards the ventral surface, it curves upwards on the left side towards the back, and dilates into a large sac, which lies over the nerve centres and dorsal vessel. It can be distinctly seen in the living larva through the skin, as it is usually filled with dark gray decom- posing fluid ; when distended, it occupies the dorsal region of three or four segments. The epithelial coat of the oesophagus and crop consists of flat cells, and is separated from the lumen by a thick laminated, cuticular layer, which is deeply plicated, processes from the cells extending into the cuticular folds (PI. I., Fig. 2). When the organ is distended, the folds are obliterated, so that the epithelium has clearly an amoeboid character. When these viscera are contracted, the plicated cuticle may fill the whole interior. The under surface of the crop has a large group of ganglion cells spread over it — the ' ganglion of the crop.' The Proventriculus (PI. I., Figs. 3 and 4) is ovoid, its long axis corresponding with the axis of the body. A longitudinal THE ALIMENTARY CANAL. 57 section through the proventriculus shows that it consists of three layers, the outermost continuous with the wall of the chyle stomach, and the innermost with that of the oesophagus. By traction the oesophagus can be drawn out with the inter- mediate layer, demonstrating that the organ is formed by an intussusception or intrusion of the oesophagus within the mesenteron. Tracheal vessels and mesoderm cells pass into the space between the inner and intermediate layers of the proventriculus; and the oesophagus is surrounded by a tracheal ring, from which these vessels arise, with twelve others, which course over the exterior of the proventriculus and chyle stomach. The epithelium, in the interior of the proventriculus, is divided into two regions by a distinct ring of embryonic cells (PI. I., Fig. 4, id), which is destined to develop in the nymph into the proventriculus of the imago. I shall term it the pro- ventricular ring. The cells of the external wall are cubical ; those of the internal wall exhibit a peculiar feathered appear- ance, and are not unlike the epithelium of the human stomach. Kowalevski* regards the proventricular ring as the rudiment of the stomodaeum of the nymph. This is an error, due to the prevalent view that the whole of the proventriculus is stomodseal. I shall hereafter show that it forms the proven- tricular epithelium only. A large crutch-shaped ganglion (PI. I., Fig. 3, g), the proventricular ganglion, lies in the angle between the dorsal surface of the oesophagus and the proventriculus, at the posterior end of the stomogastric [median) nerve ; ganglionated fibres pass from it into the wall of the proventriculus and chyle stomach. The Chyle Stomach (PI. I., Figs, i, 3 and 6), is 3 mm. long, and broader in front than behind. It has four tubular cascal glands at its anterior end. These glands measure about 2 mm. in length, and resemble the chyle stomach in structure. They do not reappear in the imago. Its epithelial coat consists of * ' Beitrage zur Kenntniss der Nach-Embryologie der Musciden.' Zeitsch. f. w. Zoo)., Bd. xlv., 1887. 58 THE LARVA OF THE RLOW-FLY. large conical cells, of from 20" to 30" in diameter, which are often 80" from base to apex, with large ovoid nuclei. These cells are distinctly striated towards its lumen (PI. I., Fig. 6) ; there is also a layer of sub-epithelial or parietal cells {gc), which are probably glandular, and a large number of very remarkable cell-nests (c «) are scattered between the epithelial and muscular coats. They are formed of small flattened cells, 5" to 6" in diameter, arranged in concentric layers. They are most numerous in the chyle stomach, and are found in the proximal intestine. They are absent in the gastric caeca. These curious bodies are apparently the histoblasts from which the mesenteron of the nymph and imago are developed. The Proximal Intestine is 9 mm. long (fths of an inch). After making a dorsal flexion forwards {the hcemal flexure), it passes backwards (Fig. 10), and forms several coils in the twelfth segment. It becomes much narrowed, and terminates in the sinus. There is a distinct ring of embryonic cells which surrounds the sinus at the orifices of the Malpighian vessel. This Kowalevski regards as the rudiment of the proctodseum of the imago. I consider it the rudiment of the metenteron, which was unknown to Kowalevski. The muscular coat of the chyle stomach and proximal intes- tine has the appearance of a lattice-work, with considerable spaces between the fibres ; that of the sinus is very thin. The Distal Intestine (PI. I., Fig. i, i') is chiefly coiled up with the proximal intestine and the vessels of Malpighi in the twelfth segment. It measures 18 mm., or Jths of an inch, in length. It is very narrow and has a thick muscular wall, which consists chiefly of circular muscle fibres. Its epithelium has a clear mucoid character, and amoeboid corpuscles are seen within and between the cells (PI. I., Fig. 7, c), which frequently exhibit goblet degeneration. The Rectum is not more than i mm. (a'sth of an inch) in length, and descends almost vertically from the dorsal towards the ven- tral aspect of the larva. The epithelium is flat, and is covered internally by a thick longitudinally plicated cuticle. The THE ALIMENTARY CANAL. 59 epithelial cells send processes outwards into the muscular coat. The muscular coat is from 20» to 25" thick, and towards the anus has a strong sphincter, composed of from 20 to 30 circular fibres. Numerous muscle-fibres arise from the integu- ment on either side, and pass inwards and downwards, to be inserted into the peritoneal coat of the rectum by fine tendon- like ends. All the muscle-fibres of the rectum resemble the skeletal muscles of the imago, and each exhibits a close row of central nuclei (see Histology and Histolysis of the Larval Tissues). The limits of the distal intestine and the rectum are not distinctly defined, but the cuticular intima of the latter becomes thinner, and the epithelium exhibits a transitional character towards the distal intestine. In the imago the two are separated by a valve. The Tracheae of the Alimentary Canal. — The great lateral tracheal trunks give off one or more branches on each side in each segment to the alimentary canal. The largest branches are those to the proximal intestine. Two very large trunks join the intestine at the anterior part of the haemal flexure and retain it in its place ; the other large tracheae enter the intes- tinal coil and bind it together. These also send numerous capillary branches to the Malpighian tubules. Intestinal Ganglia. — Viallanes [27, p. 74] describes a distinct plexus of stellate ganglion cells in the muscular coat arranged in four parallel rows, and extending the whole length of the chyle stomach and intestines of the larva of Tipula gigantea. I have been quite unable to find any such structure in the larva of the Blow-fly, but no doubt ganglion cells derived from the proventricular ganglion exist. Visceral Muscular Network. — Weismann [2] has described a network of delicate mu.scle-fibres, which connect the walls of the alimentary canal with the alze of the dorsal vessel ; but I have been unable to detect any such fibres, and, if they really exist, they could scarcely fail to appear in some of my numerous sections. The Lingual (Salivary) Glands are a pair of very large sac-like glands, which consist of a thin cuticular membrane, covered 6o THE LARVA OF THE BLOW-FLY. externally by a reticulum of stellate niesoblast and tracheas, and lined by a very regular pavement epithelium consisting of large hexagonal cells. These glands, in the recent condition, are most beautiful microscopic objects. They lie one on either side of the alimentary canal, and measure nearly a centimetre in length, and '6 mm. in diameter in their widest part. Each has a narrow duct, which joins its fellow beneath the pharynx. The conjoined duct opens through a papilla, which represents the ligula (Figs. 7 and 8). The condition of the secreting cells varies greatly at different periods. In the young larva the cells are small and increase in size with the growth of the insect. In the active feeding adult they measure 100" to iso** in diameter, and form a thin pave- ment 25" thick, enclosing a large cavity, which is filled with clear fluid. In the resting larva the cells rapidly increase in thickness, until at length they reduce the cavity of the gland, so that it appears as a mere stellate fissure in sections. In this condition they are cubical. A chitinous or cuticular intima between the cells and the lumen of the gland has been frequently described in the salivary glands and sericteria of various insects. In the active gland in the fly larva I find no trace of such a membrane, but in the resting larva a thin cuticular intima is often present. The gland-cells exhibit many exceedingly interesting appear- ances. They have a distinctly reticular structure. Their vesicular nuclei can often be separated in the recent condition of the organ. In the resting larva the cells still have numerous granules of secretion towards their inner surface, and are firmly united externally, where they have a distinctly laminated structure, the laminae passing from cell to cell, an indication that the lamellated cuticle is formed from the cells themselves, and not by a secretion poured out on their surface (PI. I., Fig. 9). Towards the gland-duct the cells are considerably reduced in magnitude, and in the duct itself they form a cuticular intima on their inner surface, which exhibits a spiral thickening similar to, but coarser than, that of the tracheal vessels. A ring of embryonic cells intervenes between the cells of the THE A L IMENTA RY CANAL. 61 gland-sac and its duct. This, according to Kowalevski, is destined to develop the corresponding gland in the nymph. The salivary glands of insects have attained a classical interest, owing to the various researches which have been made on the nerve terminals and their relation to the secreting epithelium. In the fly larva their nerves are derived from the proventricular ganglion, which gives off numerous short nerves, ending in the salivary cells. It appears to me that these are processes of the ganglion directly continuous with the proto- plasm of the secreting cells, and that only a few of the cells — those adjacent to the ganglion — receive any. Weismann described a very remarkable structure, which he calls a cell chaplet, connected with the external coat of the two salivary glands. He says : ' It consists of a string of large cells closely united, which hangs like a garland free in the body cavity. Its two ends are connected with the salivary glands, and with a muscular band from the dorsal vessel. It forms an arch, with its convexity directed backwards, in the horizontal plane near the back. This string of cells has no duct.' [2, p. 132.] The Malpighian Vessels are two in number, but each divides into two tubules about two mm. from its junction with the intestine. I cannot give the length of these tubules ac- curately, as they are not easily unravelled, but it is considerably more than a centimetre. They are at once recognised by their dark brown colour, and they become intensely black when treated with osmic acid. The Malpighian tubes have no muscular coat ; but consist of a structureless external cuticle, lined with a single layer of cells, two, or at most three, cells entirely surrounding the tube (PI. I., Fig. 8). Externally they exhibit a moniliform appearance, the large cells projecting as hemispherical protuberances on the surface. The lumen of the gland is small and irregularly flattened. The cells measure 70" to 80", and have large, nearly spherical nuclei. The cell protoplasm is distinctly reticular, and contains 62 THE LARVA OF THE BLOW-FLY. numerous granules of dark brown pigment, of a substance deeply blackened by osmic acid, and of a colourless material which stains with carmine, but is unaffected by osmic acid. Normally there are no crystals, either in the cells or in the secretion, although some preparations exhibit crystals which resemble stearin or tyrosin. These are apparently the result of post-mortem changes. 7. THE NEEVOUS SYSTEM, a. Comparative Morphology. The nervous system in insects consists of two parts, a somatic and a splanchnic system of nerves and ganglia, which have been properly compared with the cerebro-spinal and sympa- thetic systems of the vertebrata, respectively. The former supplies the integument, the organs of special sense, and the skeletal muscles ; and the latter innervates the visceral muscles, and probably has a trophic influence on the secretory epithelia. The somatic nervous system consists of a series of ventral ganglia, united with each other by one or two longitudinal commissures. In the primitive embryonic condition all the ganglia are distinct, and there may be as many as seventeen pairs — four cephalic, three thoracic, and ten abdominal (Brandt [30]). As development progresses, several of the ganglia become fused together, and this fusion is more marked in the higher types of insects. Most larvai have only two centres in the head, a pre-oral and a post-oral centre, with three thoracic and seven or eight abdominal ganglia. In the imago, the two cephalic centres are frequently fused, and the thoracic form but two centres. The greatest concentration occurs in the Muscida;, which have only two centres in the imago, the cephalic and Bibliography :— 28. Newport, G., ' On the nervous system of the larva of Sphinx ligustri and the changes it undergoes in the earliest stages of the pupa.' Phil. Trans., 1832, p. 383. London. 29. Brandt, E., 'Vergleichend. Anatomische Skizze des Nervensystems der Insekten.' Horiu Soc. Ent. Ross., torn, xv,, 1879. 30. Brandt, E., ' Ueber die Metamorphosen dcs Nervensystems der Zweifliigler.' Horai Soc. Ent. Ross., tom. xv., 1879. THE NERVOUS SYSTEM. 63 thoracic, whilst in the larva all are united in a single complex nerve-mass, in which no trace of the primitive separation of the post-oral ganglia remains perceptible externally. The anterior pair of cephalic centres is always pre-oral, and is developed from the pro-cephalic lobes. In the perfect insect it consists of a central pair of hemispherical ganglia, having a more or less convoluted surface united by a commissure. These support two pairs of pedunculated bodies — the corpora fungiform ia — as well as two pairs of sensory ganglia, the optic and olfactory (antennal) lobes. These structures are usually known as the ' brain,' but the term is more conveniently applied to the whole of the cephalic ganglia whenever they are united into a single cephalic centre, and when this is the case the commissures which unite the supra- and infra-cesophageal centres are called crura, and are not usually visible externally. The first post-oral centre, or the infra-oesophageal ganglion, usually consists of three primitive ganglia united ; these belong to the segmented portion of the head, the mandibular, maxillary and labial somites. In the Diptera the first pair are apparently wanting in both the larva and imago. In the Hymenoptera and Coleoptera, three pairs of nerves are given off from these ganglia ; in the Diptera and Muscidje there are only two pairs, whilst the Lepidoptera have only a single pair (Brandt [29]). A small nerve arises, in most insects, from the commissure between the supra- and infra-oesophageal ganglia, on each side. These are usually known as the ' labral ' nerves, but I prefer the term ' pharyngeal ' nerves, as they supply the pharynx as well as the labrum — in the Diptera, at least. The pharyngeal nerves give off a pair of recurrent branches which join the frontal ganglion of the splanchnic system (PI. II., Fig. 2). These are the recurrent nerves; sometimes they have a separate origin distinct from that of the pharyngeal nerves. They are also described as arising from a special ganglion in the commissure — the ganglion of the commissure. Each of the other segmental centres, those of the thorax and abdomen, except the last, usually consists of a single pair of $4 THE LARVA OF THE BLOW-FLY. ganglia. The last is, however, evidently the result of the union of two or more pairs. Each pair of ganglia gives off one, two, or three pairs of nerves, and in the latter case the third pair either arise from the commissures behind the ganglia, or from the posterior part of the centre by a common trunk. This trunk then bifurcates, and its branches unite with the anterior nerves of the succeed- ing ganglion (PI. II., Fig. i, al). They are the alary and lateral nerves of Newport [9], and are of considerable morpho- logical and physiological interest, as their origin from the segmental centres is apparent rather than real, and they consist of fibres, which traverse the dorsal surface of the nerve centres and their commissures, and arise from the brain, so that these probably represent the voluntary motor tracts of a vertebrate. [Newport, 9.] When the primitive ganglia are fused into one or a few complex centres, the determination of their exact number is exceedingly difficult. This is especially the case in the Diptera. Brauer [14] gives a very complete com- parative table, showing the disposition of the primitive ganglia in the larva and imago of various dipterous types, in which he shows that there are always three thoracic and eight abdominal ganglia. I must, however, regard this table as hypothetical to a large extent— although there are obviously ten Plate II.— The Nervous System of the Larva of the Blovit-fly. Fig. I.— The Ventral Chain of the Silk Moth Larva, after Swammerdam : a, antennal nerve ; ?«x, maxillary nerve ; eg, cephalic ganglion ; op, optic nerve ;/, frontal ganglion; x, lateral ganglia?; i, infra-ocsophageal ganglion; al, alar nerves. The ganglia of the thorax and ahdomen are numhered I to 10. Fig. 2.— The Frontal and Lateral Splanchnic Ganglia of the Cockroach. After Hofcr. Fig. 3._A Longitudinal nearly Median Vertical Section of the Neuroblast: a g, antennal ganglion ; s c, stellate cell ; og, retinal disc ; n r, nerve roots. Fig. 4.— a Median Section of the same : a s g, anterior splanchnic ganglion ; com, commissure of hemispheres. Fig. 5.— a Section through the outer part of a hemisphere : r, retinal disc. l.'iG. 6. — A Section through the crus and hemisphere. Fig. 7.— A Longitudinal Lateral Vertical Section through the Neuroblast of the newly-hatched larva. Fig. 8.— The Ventr.il Ganglia of Melophagus, ist Stage, after Leuckart. F,G. 9.— The Ventral Ganglia of Melophagus, 2nd Stage, after Leuckart. Fig. 10. — Dorsal View of the Neuroblast : n e r, ring. Fig. II. — Lateral View of the same : h, hemisphere ; st, optic stalk. Fig. 12. — Nerve Cells from Nerve Root. Fig. 13. — Embryonic Cells of the Cortex of the Neuroblast. Fig. 14.— Embryonic Cells and Trabecula; of the Capsule of the Neuroblast. THE NERVOUS SYSTEM OF THE LARVA. THE NERVOUS SYSTEM. 65 pairs of ganglia in the Tipulidcc, of which the last is a complex of two pairs, and Leuckart figures eleven pairs in the embryo of the Pupiparaj at an early stage of development (PI. II., Fig. 8). At a subsequent stage only nine remain— the penultimate and antepenultimate pair have disappeared (PI. II Fig. 9)- ' '' In the newly-hatched larva of the Blow-fly I have only detected evidence of the existence of six pairs, but ten pairs of nerves arise from the united ganglia. It appears to me that the number is reduced, either by complete fusion of two or more, only in certain cases, and that in others the missing ganglia have undergone complete atrophy. As one centre often gives off at least two, and often three pairs of nerves, I do not regard the numbers of pairs of nerves as evidence of the persistence of a similar number of ganglia. There are apparently never any ganglia developed in the last abdominal metamere— Weismann's twelfth segment, my fourteenth and fifteenth somites. Weismann demonstrates this in 'Chironomus' [2, p. 38], and adduces it as an argument against Zaddach's hypothesis that a pair of primitive ganglia is developed in each somite [2, p. 83]. The Splanchnic Nervous System (PI. II., Fif^s. i and 2) was first detected and figured by Swammerdam [4], and has been carefully studied by J. Miiller,* J. J. Brandt.f Newport [9], and more recently by Bruno Hofer.I It is highly developed in the cockroach ; and exhibits considerable modification, and attains great importance in the larva of the Blow-fly. It consists of two single median and of two pairs of lateral principal ganglia. The median ganglia are the frontal and proventricular ; they are united by a single thick nerve-cord, which lies immediately upon the dorsal surface of the oesophagus. This is the median or stomogastric nerve. The lateral gangha consist of an anterior and a posterior pair. They are connected with the crura of the pre-oral centres, with each other, and with the stomogastric trunk by plexiform nerves, and they supply the labial salivary glands — (B. Hofer). The nerves of the splanchnic system arise from these ganglia and from the stomogastric nerve, and terminate in * Nova .'Vcta, K. L. C. Acad., Bd. xiv., p. T},. t ' Oken's Isis,' 1831, p. 1103. X ' Untersuchungen ii. den Bau der Speicheldrtisen u. d. dazu gehorenden Nervenapparats von Dlatta.' Nova Acta, K. L. C. Acad., Bd. li., 1S87. 5 66 THE LARVA OF THE BLOIV-FLY. ganglionated plexuses in the walls of the viscera, similar to those of Auerbach and Meissner in the vertebrata. b. Structure of the Nerve Centres. The somatic nerve-centres in all insects consist of a central stroma of fibrillated substance, which I regard as analogous to the network of Gerlach, surrounded by a vesicular layer, or 'gray substance.' This is enclosed in a thick capsule of mesoblastic tissue rich in tracheae— the peritoneal capsule- continuous with the sheaths of the nerves. The Central Stroma consists of a fine network of axis cylinders, permeated in many cases by distinct bundles of larger fibres— commissural fibres, and nerve roots. Many of the nerves arise in part from this stroma, and in part from the peripheral nerve-cells. Some portions of the white substance are blackened very easily by osmic acid, and others are scarcely tinged by it. Those parts which become black are infiltrated by an interfibrillar substance of a fatty character. Dietl* named this variety of the stroma ' Mark- substanz,' medullated stroma, or meduUated substance. The medullated substance has been a great difficulty in the investigation of the nervous system in insects. It often appears in sections as a homogeneous or laminated material, or in the form of solid balls, which have been mistaken for giant cells. Dietl correctly described this substance. He says : ' Like ordi- nary stroma, it consists of nerve fibrils, which vary greatly in size from large distinct nerve-fibres, such as are found in nerves, forming commissural bundles, to the finest reticulum of minute axis cylinders, forming a dense stroma ;' but he also says ' it sometimes exists as a homogeneous mass, or in the form of laminse.' Such homogeneous masses are, however, readily resolved into a stroma by treatment with a mixture of ether and chloro- form, which dissolves a fatty interstitial substance. By this method they are seen to differ in no way from the parts in * Dietl, ' Die Organization des Aithropodengehirns.' Zeitsch. f. w. Zool., Bd. xxvii., 1876. THE SOMATIC NERVOUS SYSTEM. 67 which the stroma is more apparent, except in the greater abundance of the infiltrating fatty material. Dietl observes that the medullated stroma onl}' resembles the white matter in vertebrates (the medullary sheath of Schwann) in its reaction with osmic acid. In this I am not inclined to agree with him ; it appears to me that the interstitial substance differs but little from the white substance of Schwann, and chiefly in not being differentiated into a distinct sheath around each fibre. I regard it as an interfibrillar substance, or matrix, in which the axis cylinders, whether bundles of larger fibres or fine stroma are imbedded. I shall distinguish the two forms of stroma as non-medullated and medullated stroma. The Gray or Cortical Substance consists either of large or small ganglion cells. The large ganglion cells are stellate with numerous branch- ing processes, which are continuous with the stroma of the central white substance. Some of the processes are directly continuous with the nerves, but many of the nerve-fibres of the peripheral nerves arise from the stroma. The small ganglion cells are chiefly found in the supra- oesophageal centres, and form a layer many cells thick. They are spherical nuclear corpuscles, surrounded by a very little protoplasmic cell substance, which is prolonged in a fine fibre, and unites these cells in strings or chaplets, ultimately becom- ing continuous with the central stroma. The small round cells resemble those of the nuclear layers of the cerebellum and retina in vertebrates. The external capsule, or peritoneal coat, varies greatly in thickness ; but differs in no way from the mesoblastic tissue covering the other internal organs. c. The Somatic Nervous System and the Neuroblast of the Blow-fly Larva. The central nervous system in the larva of the cycloraphic Diptera generally, differs widely from that of other insects in the close concentration of all the ganglia in a single complex centre, which consists in part of the differentiated nervous 68 THE LARVA OF THE BLOW-FLY. system of the larva, and in part of embryonic structures destined to form the nerve centres in the nymph. Although these parts are intermixed in a complex manner, the cellular elements of each are distinctly recognisable, and those which are active in the larva undergo degeneration, like the rest of the larval tissues, whilst those which are embryonic in type undergo evolution during the formation of the nymph. As the embryonic portion predominates over the active nervous elements, I propose to speak of the whole as the ' neuroblast,' as a more appropriate term than ' central nervous system of the larva,' although the latter forms a portion which is easily recognised, but cannot be very definitely limited, except by a most detailed and elaborate description. The Neuroblast is formed by the fusion of the primitive ganglia of the embryo ; these undergo partial differentiation into nerve-cells and stroma, and remain more or less separated by mesoblastic tissue, so that tracheal vessels are found in the substance of the organ. A large part of the cephalo-thoracic ganglia remains embryonic, whilst a smaller portion of the abdominal, and that only in the more anterior ganglia, retains embryonic characters. I am not aware that any close investigation of this remark- able organ has been made by any previous writer, and with the exception of a description of the manner in which the optic lobes of the im.ago are evolved, by Viallanes, to which I shall have occasion to refer subsequently, I have been unable to find any more detailed description than that which Weismann pub- lished in 1864. He says : ' The central portion of the nervous system of the larva differs remarkably from that of other insects, inasmuch as there is no ventral cord, but a conical nerve mass, the virtual construction of which, from the ganglionated cord, is not betrayed even by a vestige of lateral constrictions. 'The infra-ocsophageal ganglia are intimately fused with those of the thorax and abdomen, whilst the prc-oral ganglia assume the form of " hemi- spheres," are nearly spherical, and are united with the ventral cone by short thick crura, leaving only a narrow space between them for the passage of the oesophagus. The "hemispheres" lie on the dorsal aspect of the ventral cone, so that seen in profile the whole resembles the butt-end of a pistol. ' The central nerve-mass is about one-twentieth of the body length— in a THE SOMA TIC NER VO US S YS TEM. 69 larva l'5 centimetres long, it only measures 74 to 78 mm. All the nerves originate from the ventral cone, and there are twelve pairs, two from the front and ten from the sides.' In this statement Weismann evidently made a slight error, as he subse- quently describes the origin of the optic nerves from the hemispheres ; more- over, a pair of antennal nerves arise from them : both these are, however, remarkable, as they are the nerve-stalks of the antennal and optic discs and although undoubtedly nerves, inasmuch as they contain nerve fibres, they consist, like all the nerve-stalks of the imaginal discs, principally of embryonic tissue. Weismann continues : ' The hemispheres and ventral cone consist of a thin, tough external capsule and its cellular contents ; the cells are lilce the nerve cells of most insects : small and spherical and without visible processes ; they lie close together, and are arranged in no definite order.' In these statements Weismann was only correct in part, the cells he saw were the embryonal elements, and he had not the means at his disposal — the preparation of thin sections — which would have enabled him to arrive at more correct conclusions. He furtlior states that 'the nerve centre exhibits a clear cortex and a dark medulla, and is one of the few examples of a tissue permeated by tracheal vessels. In each hemisphere a tracheal trunk passes, without dividing, ahnost into the centre of the organ, and then gives off a number of fine radial branches. In the ventral cone the greater part of the tracheal net lies on the surface, only in the middle line of the dorsal region a few air- vessels pass into its substance ; these perforate the nerve-mass in a vertical direction and give off branches in a stellate manner to a limited region in its substance.' I shall not give a very detailed account of the neuroblast in this section, as it will be more convenient to do so after describing the central nervous system of the imago, in the section devoted to its development, but shall content myself by describing its most salient features. The Capsule consists of a thick layer of mesoblast formed of a reticulum of stellate cells, and is richly supplied with tracheal vessels. It not only covers the surface of the neuroblast, but dips in between its constituent ganglia (PI. III., cc), more especially in the median line and around the base of the hemi- spheres — an arrangement which accounts for the penetration of the tracheal vessels into the interior of the neuroblast. The Cortical Substance consists in great part of embryonic cells, but stellate and fusiform cells are found in great numbers in the immediate vicinity of the nerve - roots around the oesophagus, and on the dorsal surface of the abdomino-thoracic 70 THE LARVA OF THE BLOW -FLY. ganglia ; indeed, towards the posterior extremity of the organ they appear to replace the embryonic elements entirely. Some of these cells attain very large dimensions, especially in the hemispheres close to the oesophagus (PI. II., Fig. 3, sc); similar cells are also found in the same region in the imago. The central substance, Weismann's medulla, consists of non- medullated stroma ; it is divided into two lateral halves, united by a transverse commissure (PL III., Fig. i), and exhibits several distinct tracts. In the newly-hatched larva the neuroblast is proportionately much longer than in the adult, and exhibits a distinct division of the pro-cephalic portion of the central stroma into three masses (PI. II., Fig. 7), which correspond with the proto-, deu- tero-, and trito- cerebron of Viallanes, and develop into the cerebron and corpora fungiformia, the corpus centrale, and the olfactory (antcnnal) ganglia respectively. In the adult larva the ventral portion of the central substance exhibits a large number of distinct bundles of vertical fibi-es on either side, but contiguous to the middle line. The hemispheres are connected with the post-oral or infra- oesophageal portion of the neuroblast by a pair of thick crura (PI. III., m s), and with each other, above by the supra- oesophageal commissure, and below by the conjoined antcnnal and infra-oesophageal ganglia. The rudiments of the corpora fungiformia (cf), and of the cerebron and optic ganglia, of the imago, can be distinctly recognised (d) ; but the most remarkable structure is a layer of infolded epiblast (yt), from which the retina is subsequently formed in the nymph. This is an imaginal disc, which lies within the capsule of the neuroblast, whilst the eye-disc of Weismann, from which the rest of the eye is formed, lies out- Plate III. — Transverse Sections through the Neuroblasi- of the Adult Larva. Fig. I. — Through the Posterior Part ; and Fig. 2. — Through the Centre of the Hemispheres, a, b, and c f, Parts of the corpus fungiforme ; d, ihe trabecula and its divisions; rr', the ring; rt, retina; c, capsule ; j,/, nuclei of infra- cesophageal ganglia ; 711 s, stroma ; st, optic slalk ; st, sialic of inferior mcso- thoracic disc ; /, median stomogaslric nerve. PLATE III. THE NKRVOUS SYSTEM OK THE ADULT LARVA. THE SENSORY ORGANS. 71 side the capsule of the neuroblast, with which it is connected by the optic stem. The antennal ganglia (PI. III., Fig. 2) exhibit a peculiar grouping of cells in small spherical masses, imbedded in a reti- cular substance. The infra-oesophageal ganglia have a similar structure, but each has a single large group of cells arranged in concentric layers (PI. III., Fig. i, s) with smaller lateral groups (/) scattered through the reticular substance. The description of the splanchnic nervous system will be more conveniently given after I have described the remaining structures of the larva. 8. THE SENSORY ORGANS AND PERIPHERAL NERVE TERMINATIONS. With the exception of the sensory papillae, described on p. 36, the only special sensory structures I have been'able'to Fig 12.-/. A section of the terminal joint of the Maxilla, showing the eye-liUe organs ; -'. A seciion of the eye-like organ (A o'l immersion lens) ; 3. Endings of a nerve in the hypodermis, showing a peripheral ganglion {' d cotes de melon, Viallanes). discover arc the pair of eye-like organs at the extremity of each maxilla. Newport recognised the existence of similar struc- 72 THE LARVA OF THE BLOW-FLY. tures in the larva of CEstrus, and suggested that they are probably eyes. That the larva of the Blow-fly is extremely sensitive to light is certain. These eye-like organs (Fig. 12, / and 2) resemble the simple eyes of the leech, but are devoid of pigment. The two in each maxilla are situated on branches of the same nerve, each is surrounded by a delicate reticular capsule, and is capable of retraction. The epiostracum forms a thin, transparent cap over the surface of the nerve terminals, which are long rod-like cells, with a distinct layer of stellate and fusiform ganglion cells between them and the nerve. The majority of the cutaneous nerves apparently end in the cells of the hypodermis (Fig. 12, 3) in demilunes of granular protoplasm (see Histology), and many of these exhibit a re- markable form of peripheral ganglion, first described by Viallanes, and called by him ' Ganglion a cotes de melon.' The ganglia are small groups of unipolar cells attached to the sides of the nerves. The cells present a striking analogy with those of the root ganglia of the spinal nerves of vertebrates. 9. THE IMAGINAL DISCS, a. General Morphology. The Imaginal discs have been already referred to (pp. j and 21) as the structures from which the nymph is developed. Bibliography :— 31. Lachat et Audouin, ' Anatomic d'une larve apode trouvde dans le Bourdon des Pierres.' Journ. de Physique, de Chimie, etc. Tom. Ixxxviii., p. 228, 1819. 32. DUFOUR, Leon, 'Etudes Anatomiques et Pliysiologiqucs sur une Mouche.' Mdm. Prcs. h I'Acad. des Sc, Math, et Phys. Tom. ix., 1846. 33. Weismann, A., 'Die Metamorphose derCoretlir.i plumicornis.' Zeitsch. f. w. Zool., Bd. xvi., 1866. 34. Ganin, M., 'Tlie Post- Embryonic Development of Insects.' In Russian. Transactions of the 5th Congress of Russian Naturalists. Warsaw (1875), 1876. (S'yezd Russkilch Estestvolspuitatelel Trudui V. Varshavye.) Rdsumd in German, Zeitsch. f. w. Zool., Bd. xxviii., pp. 386-389, 1877 ; and Jahrb. d. Anat. u. Phy., Bd. v., p. 507, 1878. 35. Dewitz, ' Beitrage zur Kenntniss der Postembryonalen Entwicklung der Glicdmassen bei den Insekten.' Zeitsch. f. w. Zool., Bd. xxx., Suppl., 1878. THE IMAGINAL DISC^. TJ, These exist in tlie larva as encapsulated groups of embryonal cells, and vary greatly in form ; for the several discs not only differ from each other, but exhibit considerable variations at different stages of development. Many are capable of easy demonstration in the adult or even in the young larva ; whilst others do not become apparent until the nymph is partially developed. As all the discs preserve an embryonic character, it is probable they are all present in the young larva, or even in the embryo, as distinct groups of cells, but only the larger ones have been actually demonstrated at the earlier periods of de- velopment ; and it is only in the later stages that the smaller discs can be safely recognised. History. — The larger discs in the larva of the Muscidas were first seen by Swammerdam [4], who, from their relation with the nerve-centres, mistook them for ganglia. Lachat and Audouin [31] termed them ' Plaques ' ; Leon Dufour [32], ' Corps ganglionoides ' ; and Leuckart [20] and Scheiber [21] described them as ganglia. Weismann [2] discovered their true nature, and named them ' Iinagtnal Schctbm,' of which the English equivalent, Imaginal discs, is generally accepted. Ganin [34], in a paper in Russian, added much to our know- ledge; but, unfortunatel}', Ganin's work is onl\' accessible to me by short translations of certain portions of it [27, 34]. Morphology. — The view of the nature of the discs which I have adopted (p. 21), although supported by the researches of Dewitz [35], Kunckel d'Herculais [25], and others, is at vari- ance with the views of Weismann [2] and Ganin [25, 27, 34]. Both these distinguished naturalists examined the structure of the thoracic discs with great care, at the earliest stage of de- velopment in which they are recognisable, and concluded that they originate from the nerve-sheaths and the peritoneal tissue of the trachea; to which they are attached. It must be confessed that in the earlier stages of develop- ment, appearances favour the views of Weismann and Ganin ; on the other hand, the hypothesis which derives epiblastic structures from the mesoblastic connective tissues, and severs the development of the Muscidae from all other insects, is one 74 THE LARVA OF THE BLOW-FLY. which possesses such inherent improbability, that even if there were no observations in support of more probable views, one would have been tempted to doubt the conclusions on which it rests, even although the facts observed are substantially as stated by Weismann and Ganin. The most rudimentary discs (Fig. 3, /) do not appear very distinctly until the pupa stage ; they are mere groups of em- bryonic cells, which project on the surface of the hypodermis, and are the rudiments of the abdominal somites. In some sections of larvae I have detected small groups of embryonic cells in the same position ; and, from the characters of the discs, I think it probable that they are derived directly from epiblast cells. Ganin says ' the transformation of the great polygonal cells of the larval hypoderm into small embryonic cells commences at four points in the lateral regions of the segments, and the new formation is preceded by a separation of the hypoderm and muscles of the larva and its cuticle.' He adds, ' these cells arise from the old cells of the larval hypoderm.' These extracts are quoted by Viallanes [27, p. 215], who remarks that, ' The replacement of the larval by the imaginal hypo- dermis is analogous to that of the milk by the permanent teeth in mammals. In the insect four germs exist in each somite, which at a certain period undergo evolution, and de- termine the shedding of the hypodermis of the larva.' Viallanes appears to me, in this and the following statements, to have correctly appreciated the true morphological character of the imaginal discs. He says : ' The facts observed by Weismann in Corethra, and those discovered by Kunckel in Volucella, must lead us to believe that the discs are derived from the hypoderm [or epiblast].* In insects we [frequently] observe that the imaginal discs are united with the hypoderm, either directly, or by a more or less elongated pedicel. The variations which are observed in different species, or in the same species in different parts, * The words in brackets are added by me, and, with their addition, I entirely agree with Viallanes. THE IMAGINAL DISCS. 75 appear to depend on the period at which the discs are first developed. ' In the abdominal region of the fly the discs first appear at the end of the larval period, so that their evolution follows immediately; thus they remain superficial. In the thorax of Corethra — Weismann — and of Volucella — Kiinckel — the discs are formed at the commencement of larval life, and become invaginated, only retaining their union with the integument by a [hollow] pedicel of variable length. In the thorax of the fly they are formed much earlier; they already exist in the embryo in the &gg. It is perhaps on this account that there is no ap- parent connection with the integument ; but further researches on this point are needed. ' It is, perhaps, important to observe that this peculiar method of development is not confined to the Insecta. The researches of Barrois have established the fact .hat in Pilidium, amongst the Nemertids, the phenomena observed are similar, since in these worms the larval exoderm is shed and is replaced by one developed from germs analogous to imaginal discs ; and the analogy is rendered the more striking, as in the different species of Nemertid the same differences are met with as in insects exhibiting a higher or lower degree of metamorphosis ' [27, p. 224]. My own researches show that in the adult larva of the Blow- fly there is precisely the same connection between the epiblast of the disc and the hypoderm as that which has been observed in the ant (Dewitz [35] ), and in Corethra [33] and Volucella [26] (Fig. 15, 2). Moreover, in the maxillary discs and those of the anterior spiracle the sac is still a wide open ampulla, and the inva- gination of the frontal lobes has been already referred to (p. 42). I attribute the failure on the part of Weismann and Ganin to trace the connection as due to the difficulties which attend the investigation in young larvee. In the advanced types of disc, such as the leg discs of the fly (Fig. 15), the invaginated hypoderm exhibits a distinct differentiation into two parts, a disc sac, s, the provisional capsule of Ganin, and the epiblast of the disc. The latter 76 THE LARVA OF THE BLOW- FLY. consists of one or more layers of cells, those on the surface, which is enclosed by the provisional capsule, are distinctly columnar, those beneath, nearer the mesoblastic surface, are small round or fusiform cells. The mesoblastic surface is usually concave, and its cavity is occupied by a layer of stellate mesoblast. The great discs of the head and the inferior pro- and meso- thoracic discs are so intimately connected with the neuroblast by neural stalks, and are apparently so far removed from the hypodermis, that it is not difficult to understand how Weis- mann and Ganin arrived at the conclusion that they originate from the sheaths of the nerves. The rudiments of these discs certainly appear in the embryo, close to the roots of the nerves to which they become adherent, whilst the nerve-centres are in contact with the ventral integument : as the growth of the larva is rapid whilst that of the neuroblast and discs is slow, the latter soon become separated from the hypodermis ; and the pouches which contain the rudiments of the discs are drawn from their points of origin, leaving the neck of the disc-sac as a very narrow tube attached to the growing peripheral portion of the corresponding nerve. The neck of the sac is not readily seen, hence Weismann regarded the bands which he saw in the adult larva connecting the discs with the integument as nerves or of secondary import. The portion of the nerve between the disc and the neuro- blast becomes greatly enlarged towards the end of larval life, owing to the rapid multiplication of the mesoblastic elements of its sheath, which are continuous with the mesoblast of the disc. The epiblast of the disc, as development progresses, usually becomes much folded or invaginated into itself; its cells are then found united by an intercellular substance resembling chitin, and one or more thin layers of cuticle are often shed from its surface. The disc-sac or provisional layer of Ganin consists of thin tessellated cells, or may exhibit transitional characters both at the edge of the disc and in the vicinity of THE IMAGINAL DISCS. 77 the hypodennis. In the later stages it becomes so thin that its original cellular character can no longer be recognised. The Mesoblast of the Disc was first discovered by Ganin ; it consists of a stroma of stellate cells, permeated by fusiform cells, tracheae, and nerve-fibres. It frequently covers the whole outer surface of the provisional sac, but is thickest in the hollow of the disc itself, in which a cavity is usually present, which communicates with the body cavity of the larva, and, like the latter, is filled with blood. The origin of the mesoblast of the disc, like that of the meso- blast of the embryo, is unknown. Kiinckel d'Herculais [25] supposed it to be developed from peritracheal ceils, or from the leucocytes of the larva, but left these alternatives unsupported by facts. Its continuity with the tissue of the nerve and tracheal sheaths is perhaps in favour of the former view ; but it is equally probable, I think, that it is developed by differentiation from the disc itself. This much is certain, it cannot be demonstrated in the earlier stages of disc development. b. The Cephalo-Thoracic Discs. The cephalo-thoracic discs may be studied in the larva ; the abdominal discs are more readily demonstrated during the earlier stages of the development of the nymph. There are nine pairs of discs concerned in the development of the head and thorax. The great cephalic discs, and the maxillary and labial discs, form the head and proboscis, an upper and a lower pro-, meso- and metathoracic disc on each side unite to form the thorax. The Great Cephalic Discs (PI. IV., op d, an d, and pc d, and Fig. 13) are formed, as already indicated, by an invagination of the frontal lobes of the embryo (Fig. 7).* They extend from the posterior extremity of the cephalo-pharynx to the anterior portion of the neuroblast, and are suspended by the cephalo- pharyngeal band and the ring (Fig. 13, cp, and Fig. 14, r). The manner in which they are disposed will be understood by a * Weismann [33] has figured a similar condition in Chironomus ; com- pare his Figs. 17 and 25 with my Fig. 7. 78 THE LARVA OF THE BLOW- FLY. reference to Fig. 14, 2, which represents a transverse section through the region indicated by the line 2 in Fig. 13 : cp is the cephalo-pharyngcal band ; s/), the suspensory membrane, and a, the base of the antennal rudiment. There is a large blood sinus between them, in which the oesophagus lies, into which the dorsal vessel opens through the ring. The suspensory membrane, up, forms its roof, and the KiG. 13.- A semi-diagramimtic reprcstimiiDn of the Head Discs and Ilemispheies seen from above : / 2 -ind j, pi incs of the sections / 2 and 3 in Fig. 14 ; mx, maxillary disc ; //;, phiryngeal , a, antennal ; 0, optic discs ; pf, prefacial rudiment ; cp, cephalo-pharyngeal band ; cr, crop; <://, cerebral hemisphere; ft, retina ; s, optic stalk ; / s, provisional sac. discs its lateral walls. The inner and posterior part of the wall of the disc sac is occupied by the epiblast of the disc ; its outer wall is the provisional membrane. The Ring was first described by Weismann as follows [2, p. 125] : ' The anterior part of the dorsal vessel lies above the THE IMAGINAL DISCS. 79 nerve-centres, and passes between the hemispheres ; immedi- ately in front of these there is a ring of cellular tissue, with a lumen large enough to transmit the vessel. The ring hangs freely in the body cavity, and is fixed by fine tracheal vessels. ' In the anterior part of the second segment ' (my fourth somite) ' a tracheal branch arises from the main trunk, passes inwards and backwards, and ultimately penetrates the hemi- sphere. In its course this trachea is united to its fellow by a transverse vessel which lies in the upper ' (anterior) ' part of the ring; whilst the trachea courses through the side and back part of the ring. The peritoneal coat of these vessels is fused with the tissue of the ring, so that the latter might per- haps be regarded as a development of the peritoneal layer of the trachea. This is, however, not so ; the tracheae have little to do with the formation of the ring, as this is clearly an organ which originates in the embryo ; its form is that of a simple broad finger-ring, the upper segment somewhat notched in the middle line. The diameter of the ring is about '23 mm., measured from before backwards ; the dorsal vessel is attached to it, dilates in front like a funnel, and is finally attached to the pharynx. ' There can be no doubt the ring is a skeletal structure, and in this relation I have not fully described it, for it gives off an anterior and a posterior band.' At this point Weismann's description becomes so complex that I shall only give a resume of his meaning as I understand it. The anterior band con- nects the anterior (upper) part of the ring with the cornua of the cephalo-pharynx — this is my cephalo-phar} ngeal band ; the posterior band is continued over the CESophagus and termin- ates in a transverse enlargement in front of the proventriculus, to which it is firmly attached. This posterior band is un- doubtedly the stomogastric nerve, which is connected through the frontal ganglion with the crura of the hemispheres and the dorsal vessel by numerous nerve-fibres, which lie upon the sur- face of the ring, and which were overlooked by Weismann. I have not the slightest doubt the cephalo-pharyngeal band is the remains of the invagination of the procephalic So THE LARVA OF THE BLOW- FLY. lobes ; and that the ring is an indication of its origin from two lateral halves. The band and ring consist in the adult larva of a solid epithelial tissue, in which the intercellular substance is much developed ; it is exceedingly like cartilage, and closely resembles the endo-thoracic skeleton of the scorpion described by Prof. R. Lankester.* The changes which the ring undergoes in the later stages of the larva are very marked ; in the feeding larva it is nearly vertical, and the nervous elements on its exterior are very distinct ; in the resting larva it grows rapidly and becomes nearly horizontal, so that its upper border becomes anterior and its lower border posterior. The nerve elements are then inconspicuous in relation to the skeletal quasi-cartilaginous substance of the ring itself. The ring and cephalo-pharyngeal band support the great cephalic disc sacs, the provisional membranes of which are continuous with its edges (Fig. 14, s/>). The disc sac is differentiated into the optic, antennal, pre- facial and pharyngeal rudiments (Fig. 13). The Optic Rudiments or Discs are cup-shaped in the feed- ing larva, and lie in front of and above the hemispheres, with which they are connected by the optic stalks. As development progresses, their outer and inferior borders become greatly thickened and folded, so that they are sub-triangular when seen in profile (Plate IV.). Weismann compared the optic disc to a mushroom with an excentric stalk, the optic stalk. The latter is solid at first, but a canal is afterwards formed in its interior, through which the retinal disc ultimately reaches the inner mesoblastic surface of the optic disc. I shall show hereafter that the optic disc is concerned in the development of the dioptric structures of * ' On the skeleto-trophic tissues and coxal glands of Limulus, Scorpio and Mygale.' Quart. Journ. Micros. Sc, vol. xxiv., new series, 1884. Plate IV.— The Neuroblast and the Imaginal Discs co.nnecteu with it IN THE Resting Larva. r. The ring ; d v, dorsal vessel ; a; the oesophagus ; o/> d, optic disc ; an d, antennal disc ; pr d, prefacial riuliincnt and pharyngeal disc ; p/i, cephalo-pharynx ; /r, trachea ; « ,/, prothoracic leg disc ; m/ d, niesotlioracic leg discs ; n, nerve. THE IMAGINAL DISCS. 8i the compound eye (the Dioptron, Mihi), whilst the retinal end organs originate from the retinal disc (Fig. 13, rt). This, which I have not included amongst the head discs, is formed Fig. 14. — Transverse Sections through the head discs (see Fig. 13) : d i\ dorsa vessel ; r, ring; /, provisional sac ; 0, optic ; a, antennal ; /, leg; ica', wing discs ; a-, oesophagus ; s, lingual gland ; pc d, prefacial rudiment ; / fat cell ; ti; trachea ; sp, suspensory membrane ; cp, cephalo-pharyngeal band. on the outer surface of the neuroblast, and is entirely enclosed in its capsule. In its relation with the nerve-centre it presents a striking analogy with the primary optic vesicle of a vertebrate. The Optic Stalk (PI. III., si, PI. IV., and Fig. 13, s) 6 82 THE LARVA OF THE BLOW-FLY. consists of a thick, hollow, neural sheath of stellate mesoblastic cells, enclosing a few nerve fibres. It arises from the upper and outer part of the hemisphere in the young larva, but as develop- ment progresses, owing to the great enlargement of the hemi- sphere, it is gradually pushed downwards and outwards, so that it finally arises from the lower and posterior portion. The nerve-fibres of the optic stalk traverse the mesoblast of the optic disc, and form a small nerve, which passes forward by the side of the pharynx and oesophagus, and probably supplies the simple eyes at the extremity of the maxilla. I cannot be certain of this, however, as I have been imable to trace the entire course of this nerve. Each lateral half of the frontal region is developed from the epiblast of the posterior part of the inner wall of the disc sac, which is connected with the lower border of the optic disc behind, and supports the antennal rudiment in front. The antenna originates from a central papilla, surrounded by two concentric rings. The central papilla becomes the third antennal joint. In front of the antenna a thin band of epi- blast is continued forwards ; it terminates in a bulb-like prefacial enlargement, and sends a process into the cephalo- pharynx, between the cornu and the inferior process, the rudiment of one lateral half of the fulcrum of the imago. Menzbier* and Kiinckel d'Herculais [25] regarded the optic discs and the antennal and prefacial rudiments as distinct, and have deduced theoretical conclusions on the segmentation of the head from this assumption. There is a distinct continuity of the epiblast, and the whole are enclosed in a single disc sac. I am unable, therefore, to admit the validity of such conclusions. The Appendicular Discs of the Head. The maxillary discs (Fig. 13, mx, and Fig. 8, 2) first appear late in larval life as invaginations of the hypoderm of the stomal disc. One appears close to the outside of the attachment of the great hook on each side. The labial discs are represented in the adult larva by a small group of cells on either side of the lingual (salivary) duct near its orifice. * ' Ober das Kopfskelet u. s. der ZvveiflUgler.' Bull. Soc. Imp. Nat., Moscou, torn. Iv., 1880. THE J M AGIN A L DISCS. 83 Kunckel d'Herculais [25] figured and described three pairs of appendicular discs in the resting larva of Volucelia — a pair of mandibular, a pair of maxillary, and a pair of labial discs. The two latter correspond with those of the Blow-fly larva, but I have been unable to find any traces of mandibular discs. Weismann knew nothing of the appendicular head discs. The Thoracic Discs are arranged in two groups. Four neural discs attached to the second and third pairs of nerves lie beneath the neuroblast and the great cephalic discs. These are the inferior, pro- and mesothoracic discs. Four pairs are closely related to the great tracheal trunks, the upper pro-, meso- and metathoracic, and the inferior metathoracic discs (Fig. 16). Fig. 15.— Leg Disc< : /, the piothor.ncic leg disc from tbe adult larva; 2, longi- tudinal section of.lhe mesothoracic leg disc of the same ; ^, transverse section of the same ; a, neural stalls ; l>, neck of sac ; c, nerve after traversing mesoblast of the disc ; m, mesoblast ; .f, sac of disc ; /;, hypoderm ; tr, trachew. The inferior thoracic discs may be called for brevity leg discs. The prothoracic pair are enclosed in a single sac, and are connected with the hypodermis by two distinct necks (Fig- 15. I b), and with the neuroblast by a pair of nerves. The mesothoracic leg discs are not united ; they lie below and a little behind the prothoracic discs. The concentric structure which the leg discs exhibit in optical section is due to the arrangement of the epiblast (Fig. 15). The central papilla is the rudiment of the last tarsal joint, and 6—2 84 THE LARVA OF THE BLOW-FLY. the ridges which surround it become the other tarsal joints, except the outermost, which is the rudiment of the femoro- tibial part of the leg. The most external portion of the epiblast becomes the sternal region of the thorax. The leg discs, in an early stage of development, exhibit only the central papilla, and the number of concentric rings increases as development progresses. The superior prothoracic disc (Fig. ii, j) is formed by an involution of the hypoblast at the base of the anterior spiracle ; it appears first towards the end of larval life, and, according to Weismann, it is the only thoracic disc not present in the Fig. i6.— Group of tracheal discs from an adult larva : tc, wing disc ; li, superior, and /, inferior metathoracic discs ; tt\ main trachea ; tr , modified tracheal vessel, seen with an inch-objective. young larva. At the end of the resting stage, or early in the pupa stage, the mesoblast of this disc surrounds the main tracheal trunk, and unites it with the mesoblast of the superior mesothoracic disc. The superior mesothoracic, or wing, disc (Figs. 14, J, and 16) lies above and in front of the great cephalic disc. It is the largest of all the imaginal discs. Its epiblast is corrugated, so that it presents numerous concentric lines when seen in optical section. The wing projects from the surface of the disc as a conical papilla. THE CCELOM AND DORSAL VESSEL. 85 The superior metathoracic disc is placed behind and below the superior mesothoracic disc. It is exceedingly like the latter, but much smaller. It is attached by bands of mesoblast to the wing disc, and to the metathoracic leg disc. One or more of the trachese which are in relation with the wing disc are usually entirely surrounded by minute embryonic cells, which are probably the rudiments from which the wing tracheas of the nymph are developed (Fig. 16, tr). Similarly modified tracheae are also found in relation with the other discs in the resting stage of the larva. 10. THE C(ELOM, DORSAL VESSEL, AND SPLANCHNIC SYSTEM OF NERVES. The Coelom. — By the term coelom, I include all the tissue interspaces between the hypodermis and the viscera. It con- tains a reticulum of cells, which divides it into larger and smaller blood sinuses. The cells may be classed in the following groups: endothelioid, stellate, connective, adenoid, and fat cells ; but numerous transitional forms occur, and the whole are probably modifications of the primary stellate mesoblast. Under the term endothelioid, I include the cells of the sub- hypodermic layer, those of the peritoneal coats of the tracheae and viscera, and true endothelial plates which bound the larger blood sinuses. They pass by numerous transitional forms into the stellate connective cells, in which the tracheal capillaries are developed. Under the term adenoid, I include certain strings of cuboid or spheroidal cells, which often attain a large size, and are frequently multi-nucleated ; sometimes as many as four or five nuclei are present in each cell. They form Weismann's cell- chaplet (see page 61), the pericardial septum, and the fringes of the dorsal vessel. The fat cells are also united in strings, and form a large reticular sheet, the fat body or omentum. This appears to the naked eye as a glistening, opaque, white sheet of tissue, which floats out of the body cavity when the integument is slit up 86 THE LARVA OF THE BLOW-FLY. under water. It consists of a median dorsal and two lateral lobes. Sections show that the omentum is much folded and convoluted, that it occupies the greater part of the body-cavity, and has large blood sinuses between its folds. The cells of the fat body (Fig. ig,/ 6) measure about o'i5 mm. in diameter ; they consist of a reticular protoplasm, in the meshes of which granules and globules of fat are imbedded. These are so numerous that they conceal the nucleus. Sections show that the cells are bounded by a thin cuticular membrane. These cells have large vesicular nuclei, which undergo very remarkable changes during the development of the nymph. Fig. 17. — Blood corpuscles {leucocytes'] of the adult larva : /, living corpuscles, showing the amosboid condition, in d the nucleus is also amroboid ; .?, the same, treated with magenta, showin;,' the various appearandes produced by the action of the re.igent ; 3, a living cell in several stages of direct division, all drawn with ,'.t oil immersion lens. (For details see ' Histology of Tissues.') The fat body, although adherent in places to the larger tracheal vessels, has no tracheal capillaries developed on its surface. When removed from the larva and exposed to the air, it rapidly assumes an inky hue, probably the result of o.xidisation. The contents of the cells are not easily acted upon by osmic acid, unless the outer wall of the cell is ruptured. THE CCELOM AND DORSAL VESSEL. 87 The cells of the fat bodies increase rapidly in size with the growth of the larva, and, except in very young larva;, do not appear to increase in number. The fat body is undoubtedly the principal store of nutrient material for the development of the imago. The Blood of the larva consists of an opalescent spontaneously coagulable fluid, and has a large number of amoeboid corpuscles 5" to 6" in diameter (Fig. 17). It permeates all the tissue spaces of the ccelom, but there are several distinct blood sinuses through which the direction of the blood stream appears to be constant. The largest are the great ventral and the pericardial sinuses. The Great Ventral Sinus commences between the imaginal discs of the head, and extends forwards to the pharynx and maxillae as the cephalo-pharyngeal sinus, and backwards sur- rounding the neuroblast and the alimentary canal as far as the posterior border of the 12th somite, where it is lost in spongy tissue, through which the blood ascends to the pericardial sinus. The Dorsal Vessel (Figs. 18 and 19). There is no organ in the larva the study of which presents such difficulties as the dorsal vessel. It is a muscular tube which extends from the posterior transverse tracheal trunk to the ring, to both of which it is attached. As Weismann pointed out [2, p. 121], it consists of three parts, which I shall distinguish as the ventricle, the intermediary portion, and the aorta. The Ventricle (Fig. 10, dv) is ovoid, flattened from above downwards and constricted at two points, so that it consists of three chambers. It lies in the dorsal or pericardial sinus, and is separated from the intestines by a partial septum — the pericardial septum. The Pericardial Sinus lies immediately beneath the dorsal integument of the nth, 12th, 13th, and 14th somites. The ventricle only partially fills the cavity, which contains tufts of fine trachete and a quantity of lymphoid tissue, indistinguishable from the lymphoid tissue of vertebrates. This is especially abundant on either side of the ventricle. 88 THE LARVA OF THE BLOW-FLY. Viallanes [27, p. 66] describes the dorsal vessel of a young dipterous larva, which he refers to the genus Ctenophora, in which he observed tufts of fine tracheas in the pericardium, and the connection of the dorsal vessel with the transverse anasto- motic tracheal trunk. He adds : Fig. i8. — The dorsal vessel : /, a semi-diagrammatic representation of the two posterior sections, showing the large cells of the pericardial septum, the alse rnuscularcs and the cell fringes ; 2, the anterior eNtremity of the aorta and the ""g ; 3^ ^ transverse section of the intermediary portion ; 4, a longitudinal sec- tion of the same, with a pair of valves, seen from aljovc, sliowing the internal cells, cell fringes and aLx musculares ; 5, a lateral view of a valve pouch — [V o" immersion lens. ' It appears, therefore, that in this larva the dorsal vessel is an arterial heart, since the blood which enters it has passed THE CCELOM AND DORSAL VESSEL. 89 through the floating branches of a rich tracheal arborescence. Thus the respiratory function appears to be localized in the last segment, and there are few tracheae in other regions of the body.' The paucity of tracheae in the young larva of the Blow-fly was observed by Weismann, but in the adult larva they are Fig. 19. — .Sections through the pericardial sinus, with the dorsal vessel in situ : /, near the anterior, and 3, near the posterior extremity of the ventricle ; / s, pericardial sinus ; //', cells of fat body ; a //;, ala; niusculares ; i s, ventral valve ; m, dors.al recti muscles ; ;', intestine ; / c, cells of the pericardial septum ; tr, trachea. abundant, a fact which does not, however, render the presence of numerous tracheae in the pericardial sinus less interesting. The Pericardial Septum (Figs. 18 and 19) forms the floor of the pericardial sinus. It consists of a double row of large ovoid cells. Weismann [2] states that there are thirteen on each side of go THE LARVA OF THE BLOW-FLY. the middle line, but I have been unable to determine their number, which appears to me variable. These cells are oval in transverse section and measure "i mm. in long diameter ; they stain deeply and have large oval nuclei. They are connected with the muscular alse of the ventricle, and I have frequently traced a direct continuity of their cell substance with a muscle fibre. The Muscular Alae of the Ventricle consist of three groups of diverging muscle fibres on each side (Fig. i8, /), which are inserted in part into the cells of the septum and in part into the middle line of the ventral surface of the ventricle. The finest of these fibres are not more than 2^ to 3" in diameter, and are very distinctly striated. They are surrounded by a myolemma, which also encloses the large cells of the septum. The muscular alas appear to arise from the peritoneal coat of the great lateral tracheal trunks, and they divide and subdivide in their course towards their insertion. Valves. — The cavity of the ventricle communicates with the pericardial sinus by a series of slit-like openings, guarded by valves. It is exceedingly difficult to determine the exact number of these openings, as they are most readily studied in transverse sections. There are certainly two lateral openings to each chamber, and one or more ventral slits. There are also several openings (four ?) at the posterior extremity of the heart. The ventral openings (Fig. ig, 2) are between the cells of the septum, and the alar muscles pass into the wall of the ventricle immediately behind them. The valves which guard the openings are thin membranous flaps, which project into the ventricle, and are not nucleated projections of the wall of the cavity. I have been unable to distinguish valves between the chambers of the ventricle. The second, or intermediary, part of the dorsal vessel (Fig. 18) varies considerably in diameter and in the form of its cross section at different points. It is circular at its origin, becomes pentagonal in the middle of its course, and circular again near its termination. It measures from 'is mm. to THE CCELOM AND DORSAL VESSEL. 91 ■25 mm. in diameter, and is smallest at its ends. It dips downwards from its origin, and its termination is near the central axis of the body. It is not enclosed in a pericardial cavity, but the pericardial septum is continued forwards as a fringe on either side of the vessel, consisting of a double or triple row of cuboid cells. The cell fringes give insertion to small alar muscles at intervals which arise from the lateral tracheal trunks. They terminate in front in Weismann's cell chaplet. When the dorsal vessel is removed from the body the intermediary portion contracts and throws the cell fringes into very regular plications. The cells which form the fringes and cell chaplet are 25" to 50» in diameter. The marginal cells arc cuboidal, and those nearer the dorsal vessel are spheroidal, and usually contain from two to five nuclei. Similar cell chaplets exist in the imago, and are undoubtedly young fat bodies ; and in many insects fat bodies are found attached to the dorsal vessel in the place of cell fringes. I am inclined to regard these structures, therefore, as the young fat cells of the nymph (see Development of the Nymph). The intermediary portion of the dorsal vessel contains several double valves opening forwards. These are pouches of the lateral walls. The muscle fibrillae have a somewhat spiral arrangement at the valves, which gives the appearance of a St. Andrew's cross when both sides of the tube are seen superimposed (see Fig. 18, ^ and 5). The third part of the dorsal vessel, or aorta (Fig. 18, 2), measures about i mm. in length. It lies upon the neuroblast and terminates in the ring, from which a number of line muscle fibrilljE are prolonged forwards, and are attached to the meso- blast of the cephalic discs and to the posterior extremity of the cephalo-pharynx. These fibrillar lie in the cephalo-pharyngeal blood sinus, and were described by Weismann. Structure and Morphology. — Viallanes [27, p. 58] gave a resume of what was certainly known of the dorsal vessel in 1882, and I do not find that anything has been added to our knowledge since. He said : ' It consists of a tube with two lateral rows 92 THE LARVA OF THE BLOW- FLY. of nuclei in its walls.' He dismissed the researches of Biitschli, Dohrn and Jaworowski in the following words : * The study of certain embryos leads to the conjecture, which has never been certainly established, that it is formed from a double row of cells, and that each nucleus represents a cell. Thus, it may be said to consist of a series of hollow segments, each formed of two cells, joined in the middle line.' Viallanes claims to have demonstrated the junctions of the cells by staining the intercellular substance with nitrate of silver, and to have observed muscular fibrilhe, which pass from segment to segment imbedded in the substance of the cells ; and he concludes that it is morphologically a capillary blood vessel with muscle fibrillse imbedded in its cellular walls. Weismann [2] regarded it as a hollow muscle fibre. I am inclined to consider it a hollow fibre of peculiar construction, and think that its true nature is as well represented by Weis- mann's as by Viallanes' hypothesis. Its walls are certainly cellular, and consist of muscle fibrillae, but whether the cells should be regarded as a bed in which the fibrillse lie or a lining intima having an endothelial character is a point not easily determined. It is certain that a very slight modification of one of the skeletal muscle fibres with central nuclei would render it practically identical with the dorsal vessel if the valves are left out of consideration. As these are, however, a most important element in its construction, it appears to me that Wcismann's view is only an approximation to the truth. The nuclei of the dorsal vessel are arranged with great regularity. They are about "i mm. apart, and each measures about 15'' to 20" in diameter. The nuclei are surrounded by more or less granular protoplasm. The muscular layer is ex- ternal to this cell substance, and consi.sts of fibrillae 2" to 3" in diameter. They are chiefly longitudinal in direction. They are most distinctly striated, and the transverse striae surround the whole tube. The Splanchnic Nervous System consists of a series of visceral ganglia. Those of the pharyngeal sinus, of the crop and THE SPLANCHNIC NERVES. 93 proventriculus are the largest. They are all united with a central ganglion, which I shall term the median ganglion. The median ganglion is situated on the dorsal surface of the oesophagus, immediately behind the commissure of the hemispheres. It is a pyramidal enlargement of the median splanchnic nerve. It receives several nerve cords from the crura of the hemispheres at its base, and gives off branches from its apex which are attached to the posterior margin and external surface of the ring ; they surround the dorsal vessel and supply it, and terminate in the ganglion of the crop. The portion of the median nerve in front of the central ganglion divides into several branches, which form a plexus around the oesophagus. These end in the ganglion of the pharyngeal sinus. The Ganglion of the Pharyngeal Sinus consists of a number of large stellate nerve cells, the processes of which terminate in the intrinsic muscles of the pharynx. Many of these cells measure 20" to 30" in diameter ; they are not closely packed together, but are scattered in the posterior part of the sinus and lie chieily close to the pharyngeal epithelium. Beside these ganglion cells there is a considerable quantity of lymphoid tissue and a group of cells similar to those which form the pericardial septum in the sinus (Fig. 8, /). The posterior portion of the median nerve remains undivided and terminates in the ganglion of the proventriculus (page 57). This ganglion not only supplies the proventriculus and chyle stomach, but gives off several ganglionated nerves on each side to the salivary glands (page 61). Further details of structure and the manner in which the ganglion cells are related to the parts they supply will be given in the histological section of this work. APPENDIX TO CHAPTER IV. METHODS OF STUDY. The great improvements of modern histological research are the perfect fixing of the tissues, the use of a nuclear in the 94 THE LARVA OF THE BLOW-FLY. place of a diffuse stain, dehydration with alcohol and mounting of soft tissues in Canada balsam in the place of fluids, glycerine, and glycerine jelly, which all have a most destructive action so far as the finer details of structure are concerned. The use of serial sections stained after they are cut has almost superseded the old methods of dissociation with needles, although this method is not without advantages, when the tissues are properly fixed. I believe I have tried all the principal methods recommended at various times, but have finally adopted the following almost exclusively. 1. Fixing. — The ' fixing' of the tissue elements is of primary importance, and the majority of the preparations I have seen of parts of insects are rendered worthless by defective methods. All the ordinary fixing fluids fail in the Blow-fly larva, owing to the great impermeability of the integument. I have found the following methods good : 1. The integument may be slit up longitudinally under Flemming's solution,* and the parts removed with fine-pointed glass rods. Preparations of the several tissues should be examined within half an hour or less of their immersion. 2. The living larva may be injected by means of a hypo- dermic syringe, with a mixture of peroxide of osmium, i % solution, and absolute alcohol, i part to 20. This is the best preparation for the demonstration of Newport's segment and the stomal discs. Larvae so prepared may be dissected under water or Flemming's chromo-acetic solution ;-f or the parts removed may be placed in absolute alcohol, and afterwards mounted. 3. The larvae may be fixed by heat : a few should be placed in a test-tube with water, heated to the boiling point, and allowed to cool. The opacity of the larva indicates the coagu- lation of its blood. Larvas heated in absolute alcohol, Mayer's method,! almost always burst. Heating in water is the best method of preparing larvae for cutting sections, as all the * Chromo-aceto-osmic mixture ; an aqueous solution of chromic acid o'25, osmium peroxide 01, and glacial acetic acid o'l %. t The above without the osmium peroxide, t Mitth. Zool. Stat., Neapel, ii., p. 7, 1S81. METHODS OF STUDY. 95 tissues are perfectly imbedded in the coagulated blood. Sections cut by a mechanical microtome, from heat-coagulated larvae imbedded in paraffin, were prepared in the following manner. 2. Imbedding. — Heat-coagulated larvie are cut longitudinally or transversely and placed in absolute alcohol for two or three days, transferred to chloroform, and left until they sink. I adopt the following modification of Giesbrecht's method :* The chloroform is gradually saturated with fragments of hard paraffin. When it will dissolve no more it is placed in an oven and gradually heated, adding paraffin a small piece at a time, until the temperature of 130° Fahr. is reached ; the chloroform evaporates slowly, and sufficient paraffin must have been added in the process to prevent rapid evaporation. The concluding stages are performed in an open vessel. In twenty-four hours or longer the last traces of chloroform have been removed. The imbedding mass is allowed to cool, and the specimen is cut out and prepared for the microtome. The specimen must be placed in position before the paraffin is allowed to cool. It is important to perform the whole operation at as low a temperature as will melt the paraffin. Any undue rise of temperature will render the integument horny and destroy the specimen. All my serial sections were made with the Cambridge rocking microtome. 3. Staining and Mounting. — Staining is most satisfactorily effected after cutting the sections. I adopt the following pro- cess, which leaves nothing to be desired : I. The section ribbon should be attached to the glass slip with equal parts of white of egg and glycerine, Mayer's formula. t I effect this as follows : I spread a thin film of gljxerine albumen on the slip by a camel's hair pencil, which has been previously wetted with distilled water. If not wetted the glycerine albumen becomes * Zool. Anzeig., p. 483, vol. iv., 1881. f Joiirn. of Roy. Mic. Soc, new series, iv., p. 317, 1884. 96 THE LARVA OF THE BLOW-FLY. frothy and the air bubbles remain. Adjust the paraffin ribbon on the slip and bring it everywhere into contact by means of a clean wet camel's hair pencil. Place the slips in an oven at 100° Fahr., and heat to the melting point of the paraffin. Allow the slips to remain at this temperature at least two hours. Prepare the following baths : 1. Spirit of turpentine. 2. Methylated spirit. 3. Equal parts of methylated spirit and distilled water. 4. Distilled water, 100 c.c, 10 % sol. hydrochloric acid, I c.c. 5. Ehrlich's logwood.* 1. The slips with the specimens attached are taken from the oven and immersed in bath i whilst still hot. They should remain from one to twenty-fours hours. Prolonged immersion does no harm, but a trace of paraffin remaining prevents staining. If the bath is not fresh it is as well to place them for a few minutes in a second turpentine bath. 2. The slips are removed from the turpentine and placed face downwards in 2. The turpentine sinks to the bottom of the spirit ; a second bath of spirit may be used with advantage. They should remain at least ten minutes. Transfer to bath 3 for a few minutes, and to bath 4 for from five minutes to a quarter of an hour. They should be moved about in bath 4 until the acidulated water lies smoothly on the slip. Pour a * I prepare Ehrlich's Logwood (Ha;matoxylin) stain as follows : Hasmatoxylin Crystals (the best), 2 grms. Absolute Alcohol, 100 c.c. Dissolve and add — Distilled Water and Glycerine, 100 c.c. of each. Alum, as much as the mixture will dissolve. Finally add 5 c.c. of glacial acetic acid. This stain should be kept at least a year in a corked bottle. If kept in a stoppered bottle, the vessel should hold at least 600 c.c, and the stopper should be removed occasionally to change the nir. I believe the time necessary to render the stain as good as possible may be much abbreviated by shaking the vessel frequently, and I know its improvement is due to the oxidisation of the alcohol. When recently prepared it is useless. For the original formula sea Zeitsch. f wissensch. Mikr., iii., p. 150,* 1886, and Journ. R. Micros. Soc. Lond., 2nd ser., vi., p. 1090, 1886. METHOnS ai' STUDY. 97 few drops of 5 over the slide, and cover with a bell glass for an hour or two. 3. Wash well in bath 4, and judge by the colour of the specimens, which should be orange-yellow, not brown, or they are stained too deeply. If too deeply stained leave them in the acid bath until they assume the desired tint. Practice is needed in judging of this. 4. Place the slides face upward in a large vessel, holding half a gallon, of any hard water. The colour of the prepara- tions will gradually change to a brilliant blue — this may require an hour or more — I use ordinary London water. It is advantageous to place the specimens in a shallow dish, through which a gentle stream of water is flowing for half an hour, as a trace of acid in the albumen film causes the colour to fade. I use a porcelain photographer's bath, and allow the water from the tap to flow in a thin stream into one corner and out of another. Specimens so prepared exhibit perfectly definite nuclear staining. If thought desirable, a second diffuse stain as neutral carmine may now be poured over the slide and washed off when the pre- paration is sufficiently stained. I do not think, however, that such double staining serves any useful purpose, as the logwood usually gives sufficient colour to differentiate all the tissues — whilst the double stain often obscures details. The lime salts in the water used is a powerful mordant of Haemato.xylin. I formerly rendered the water bath feebly alkaline, a glass rod dipped in liq. ammonije used to stir the bath is sufficient. Neither soda nor potash should be used, and the advantage of using ammonia is doubtful. 5. Transfer to a bath of 50 % alcohol, and then to a bath of methylated spirit. Wash the slide with a little absolute alcohol — I c.c. is sufficient — allow it to drain for a moment, and drop clove oil over it with a pipette, moving it as a photographer does a plate he is coating with a film or with varnish. Drain by setting the slides on end, under a glass shade, on a sheet of filter-paper. If there is any cloud it may be removed 7 98 THE LARVA OF THE BLOW-FLY. by placing the slides for a few moments in an oven at ioo° Fahr. Clouds are produced by water or alcohol. A water cloud cannot be removed, and depends on the presence of water in the methylated spirit. Alcohol clouds arise from the use of too much clove oil. Clove oil will not dissolve alcohol, but displaces it. Setting the slide on end usually dissipates any alcohol cloud. The specimens must on no account be allowed to dry at any stage of the process ; if they do they are ruined. The greatest danger is after the washing with absolute alcohol. The methylated spirit used must be the best, and entirely free from gum. After draining off the clove oil I mount in Canada balsam dissolved in xylol. Staining in bulk before cutting may be effected with borax carmine, but is often unsatisfactory. Picro-carmine may also be used. I find that immersion of the part of the larva to be stained in solution 4 for some hours facilitates the penetration of the stain. In no case can an entire larva be satisfactorily stained, as the integument is practically impervious. Whatever process is adopted, no good can result from improperly-fixed tissues. CHAPTER V. THE INTEGUMENTAL SKELETON OF THE IMAGO- 1. GENEEAL CHARACTERS OF THE EXO-SKELETON IN INSECTS. Willis, as long ago as 1692, recognised certain analogies between the cuticular skeleton of an insect and the osseous skeleton of a vertebrate. Geoffroy St. Hilaire was the chief exponent of the ideas which Willis enunciated, whilst Audouin and his followers, amongst whom the great majority of the present school must be included, persistently regard all such views as misleading. No doubt, as far as details are con- cerned, the followers of Audouin are right ; nevertheless, it must be admitted that analogies exist, in the modern sense of the word, whilst the most recent literature on the subject, the writings of Patten and Gaskell, show that even in the wider sense in which the word analogy was used by St. Hilaire, there is more truth in his theory than has been admitted by the disciples of Audouin. Bibliography.— The following works, with those quoted on page 7, deal more especially with the external skeleton. 36. Savigny, J. C, ' Mdmoires sur les Aniniaux sans Vertdbrds,' premier part, 8vo, Paris, 1816. This work is the foundation of the present nomenclature applied to the parts of the mouth in insects and arthropods generally. 37. Latreille, p. a., ' Observations Nouvelles sur TOrganisation ex- t^rieure et gdn^rale des Animaux Articul(fs, et h. Pieds Articul^s, et application de ces connoissances h la nomenclature des principales parties des memes Animaux.' M^m. du Museum d'Hist. Nat., torn. viii., p. 169, 4to, Paris, 1822. 38. AuDOurN, v., 'Recherchcs Anatomiques sur le thorax des .A.nimaux 7 loo THE INTEGUMENTAL SKELETON OF THE IMAGO. The skeletal structures of insects, both internal and external, undoubtedly consist of indurated cuticle, both the epiostracum and endostracum taking part in their formation. These hardened or chitinized cuticular structures form plates, rings, or complex solid or hollow parts, resembling small ossicles, and are denominated sclerites. The internal sclerites which form an cndo- skeleton are either indurated inflections of the epidermis, or indurations of the cuticular layer of the alimentary canal, or of the tracheal vessels, yet the former are apparently replaced in some arthropods, by a tissue closely resembling cartilage, as, for example, the ento-thorax in the scorpions ; and this tissue is regarded by Lankester* as of mesoblastic origin— a fact which, taken with many others, indicates a similarity which is not by any means a mere superficial resemblance between the carti- laginous skeleton of a vertebrate and the indurated epidermal skeleton of an insect. Further, in both sub-kingdoms it is indubitable that the skeletal structures afford the most important morphological characters, and preserve, more or less perfectly, the original segmental character of the embryo in the adult organism. In both sub-kingdoms the skeleton is modified in accordance with * ' On the skeleto-trophic tissues and coxal glands of Limulus, Scorpio and Mygale.' Quart. Journ. Micros. Sc, vol. xxiv., new series, 1S84. Articulds et celui des Insectes Hexapodes en particuliire.' Ann. Sc. Nat. Zool., torn, i., 1824. 39. AUDOUIN, v., art. 'Insectes,' Dictionnaire Classique d'Histoire Na- turelle, torn, viii., 8vo, Paris, 1825. This article gives a very complete rdsumd of the views of Audouin, which are the basis of the modern nomenclature of the thorax. 40. Straus Durckheim, 'L'Anatomie Comparde des Animaux Arti- culds. Auquelle on a joint I'Anatomie descriptive du Hanneton,' 410, Paris, 1828. One of the most masterly memoirs ever published. The nomenclature of the parts is, however, entirely enipiricr\l, and founded on supposed analogies with the parts of vertebrates. 41. MiALL AND Denny, 'The Structure and Life-History of the Cock- roach,' 8vo, London, 1886. This litile work is an excellent introduction to the more extended study of the Insecta, and should be read by every student. GENERAL CHARACTERS OF THE EXO-SKELETON. loi the needs of locomotion. Such considerations led Geoffroy St. Hilaire and his followers to apply the names of various parts of the vertebrate skeleton to those of the insect skeleton which, as they conceived, serve similar functions. Although, perhaps, without exception, such terms are not jus- tified, and in a strictly morphological nomenclature must be rejected, their rejection would be exceedingly inconvenient, and would lead to the adoption of many new and unfamiliar terms. Definitions.— In all insects the sclerites of the head, thorax and abdomen, except those of the pre-oral region, form a series of annuli or rings, and correspond more or less closely with the surface of the primitive somites of the embryo; hence their morphological import. As this portion of the skeleton cannot be properly called axial, I shall term it somatic, to distinguish it from the skeleton of the appendages of the metameres, which maybe properly termed appendicular. The sclerites of the somatic skeleton are chiefly plates, and the limits of these are marked by seams or sutures. The appendages of an insect are also segmented, hence the term Arthropoda, and the several segments of an appendage are termed joints ; this term is therefore used in two distinct senses : it may either mean a segment of an appendage, or the articulation between two sclerites. Ambiguity does not neces- sarily arise from this usage, although it would be better to adopt a more consistent nomenclature. The segments of an appendage are so generally called joints in works on entomo- logy, that it is difficult to avoid this usage. I shall there- fore always speak of the union of two or more sclerites as an articulation, except when the word joint is qualified, as in the expressions ball-and-socket, or hinge-joint, when it may be used without ambiguity in its more usual sense. The different forms of articulation maybe classed under two groups, Sutures and Arthroses. Sutures are seams between the sclerites, and are chiefly seen in the somatic skeleton. Arthroses are movable articulations, such as occur in the appendicular parts of the skeleton. A suture may be either a symphysis or a syndesmosis. 7—2 102 THE INTEGUMENTAL SKELETON OF THE IMAGO. Symphysis (Fig. 20, /) signifies the union of two sclerites by an inflected chitinous ridge between them. Syndesmosis (Fig. 20, 2) is the union of two sclerites by a soft, flexible portion of the cuticle. Syndesmoses are frequently imbricate ; that is, one sclerite overlaps the other, protecting the soft integument between them from injury. Arthrosis. — An arthrosis is a kind of locked syndesmosis: two joints of an appendage are united by syndesmosis, but pro- cesses on the exterior of one fit into hollows on the exterior of the other. These may form a hinge, or peg joint, or, in some Fig. 20. — The principal forms of Articulation (Diagtanimatic) : /, symphysis; 3, syndesmosis ; 3, ball-and-socUet-joint, from a section of the tarsus of the fly ; 4, ginglymus or hinge-joint ; J, coxa, with a hemispherical articular surface, show- ing its relation to the line of attachment of the syndesmotic membrane. cases, a ball-and-socket (Fig. 20, j, 5). A hinge joint, ginglymus (Fig. 20, 4), admits of flexion and extension in one plane only ; a peg joint, of rotation only; and a ball-and-socket, amphi- arthrosis, admits of more or less free movement in any direction. There are other forms of articulation, but these do not need special terms for their description ; they are generally a modi- fication of one or other of the above. Sutures may be arranged in three classes, in accordance with their morphological value ; median, primary, and secon- dary sutures. Median Sutures are either dorsal or ventral, and are indicative of the bilateral symmetry of the developmental process. They represent the furrow between the two halves of the primitive GENERAL CHARACTERS OF THE EXO-SKELETON. 103 band of the embryo in the ventral region and the raph6 in the dorsal region, formed by the union of the two lateral halves of the somatopleure on the dorsal surface of the yelk. Primary Sutures are the remains of the inflections of the epiblast between the primary somites of the embryo. Secondary Sutures are either symphyses or syndesmoses, separating the sclerites of a metamere. Morphological Considerations. — The median dorsal sutures are the first to disappear in the more highly-modified types. In the higher forms of the Insecta the median sutures are rarely present, but their position is indicated by complex inflections of the integument in the ventral region. Such are the entothoracic and endocranial processes which support the ganglia of the ventral chain. The consolidation of the somites, which is characteristic of the most highly-modified types, leads to the obliteration, or partial obliteration, of the primary sutures, whilst the secondary sutures become larger and more numerous, and frequently form important syndesmoses. These are especially concerned in the mechanism of flight and the respiratory movements of the thorax. Secondary sutures also often form strongly-marked internal ridges and processes, which afford attachment to muscles and add greatly to the solidity of the external skeleton ; thus the secondary sutures appear to take the place of the primary sutures, with advanced complexity. The more highly - developed a part, the more numerous and important are its secondary sutures, and the greater the tendency to the consolidation of several somites, with the consequent obliteration of the primary sutures between them. These laws are exemplified in the structure of the head and thorax, and the confusion which exists in the nomenclature of the sclerites of the cephalic and thoracic skeleton depends mainly on the fact that they have hitherto remained unrecog- nised. Although the secondary sutures are very similar in corre- sponding somites of widely-divergent types of insect, their disposition and number have possibly only an indirect mor- phological significance, and result from similar adaptive I04 THE INTEGUMENTAL SKELETON OF THE IMAGO. modifications. In the simpler or more generalised forms each somite is protected by a simple annular sclerite, or by a dorsal and a ventral semi-annulus. The consolidation of the cuticular layers in the higher types commences in the inflected ridges corresponding with the primary and secondary sutures, or in relation with the insertion of the more powerful muscles; hence numerous secondary sclerites occur. These are frequently distinct in the adult nymph or young imago, but are often con- cealed in the adult imago by large deposits of chitin between them and the hypodermis. A very important fact, to which I would draw attention, is the close resemblance which exists between an e.xternal skeleton developed directly from the epiblast of the embryo and an external skeleton formed entirely from the epiblast of imaginal discs ; the secondary sutures in both are as closely related as the primary ones. Whether they first appear in the embryo or in the nymph, they represent the same functional adaptation of the skeleton to the needs of the organism. The same ridges and syndesmoses appear on the wing-bearing segments, whilst no traces of them appear on the wingless prothorax or abdominal somites. If, as I have already stated,* it is probable that the larva and nymph states in the Metabola are interpolated stages of develop- ment, we might expect to find a greater divergence of structure from that of the primitive type in the nymph and young imago, with a tendency to a return towards the primitive type in the more mature insect ; and we actually see a more complex condition of the skeleton in the young imago than in the adult insect. In the vertebrate skeleton each bone, as a rule, is formed from several ossific centres, which subsequently unite with each other. In the young insects there are often several sclerites, which are afterwards united to form a single plate, but the process of union is different : the component sclerites are not usually fused by the gradual growth of each and their subsequent union, but by a continuous deposit of chitinous laminae on the internal surface of the whole, so that * Page 19. GENERAL CHARACTERS OF THE EXO-SKELETON. 105 with advancing maturity there is a tendency to return to a simpler form of skeleton. A classification of the various sclerites indicative of their morphological significance is not possible with our present knowledge. It is not probable, for example, that any one of the smaller plates which occur in the prosternal region of a young imago in the Diptera represents the sternal plate of a more simple somite, or the presternum of an archaic insect, nor is there any reason to regard the consolidated presternum of the adult imago, formed by the union of these several sclerites, as homologous with the more simple prosternum of a more generalised type. Generally, the sclerites which appear at the earliest period of development are either superficial plates, which are sub- sequently united by a continuous deposit of chitin on their internal surface, or chitinized portions of inflected folds of integument. In the former case, I shall term them exo-scleritcs, and in the latter cndo-sderitcs. The continuous deposit of chitin by which these are united with each other I shall term the scleral matrix ; and the sclerites formed of two or all these elements compound sclerites; those which consist of only one element simple ; and sclerites which are formed entirely from the scleral matrix I shall term matrix sclerites. Lastly, certain sclerites occur only in the more generalised forms of insects, such, for example, are the bows of chitin which partially surround the limbs of the Cockroach (Periplaneta) ; these are formed in folds of the syndesmotic integument between the dorsal plate and the limb. I think it probable that these belong to the limb rather than to the thorax, and that they represent two limb joints between the coxa and the trunk, corresponding to the two basal joints of the limb in some Crustacea. If this be so, they are clearly indicative of an under- lying archaic type. They are not present in the more highly modified types of insect, and such sclerites may be termed evanescent, as they tend to vanish in the higher forms. Apodemes.— This term was applied by Audouin to endo^ sclerites which form lever-like rods on which the muscles act. io6 THE INTEGUMENTAL SKELETON OF THE IMAGO. Sometimes there is a distinct movable articulation between the apodeme and the plate on which it acts. Such apodemes were distinguished by Audouin as epidemes ; both terms are useful. Endopophyses, Entosterna, and Diaphragmata are terms applied to the internal inflections of the skeleton. They are plates of chitin, which form partial septa dividing the body cavity, either giving attachment to the muscles, strength to the skeleton, or support to some of the internal organs. The diaphragmata are transverse septa ; the entosterna are consolidated median sutures in the ventral region. Endopophysis is a general term for any internal inflection which has not received a distinctive appellation. 2. THE HEAD CAPSULE. (a) General Morphology. The head capsule of an insect may be said to be a subovoid capsule with an anterior opening, the mouth ; and a posterior opening, the occipital foramen, with the edges of which the skin of the neck is continuous. It is usually described as con- sisting of an epicranium, which forms the convex upper surface and sides of the head capsule, of a plate on its under surface, Bibliography : — 42. Lkmjk;, K., ' Vom Bau des thierischen Koipeis,' Tiibingen, 8vo heft i 1864. 43. Balfour, 'Comparative Embryology,' Lond., 1881. 44. RoLLESTON's ' Forms of Animal Life,' 2nd edit. Revised and enlarged by W. Hatchett Jaclcson. 8vo, Oxford, 1888. Beside the works on insect Anatomy and Development already quoted, the reader who desires to attain further information on the most recent morphological speculations should consult : 45. Gaskei.l, W. H., ' On the Origin of the Central Nervous System of Vertebrates,' ' Brain,' vol. xii., pt. i, 1879. 46. Bkllonci, 'Sur la Structure et les Supports des Lobes Olfactifs dans les Arthropods Supdrieurs et les Vertcbrds.' Archiv. Ital. de Biologic, torn, iii., p. 191. 47. Gaskell, W. H., 'On the Origin of Vertebrates from a Crustacean- like Ancestor.' Quart. Jour. Mic. Sc, vol. xxxi., pt. 3, 1890. 48. I'ATIEN, W., ' On the Origin of Vertebrates from Arachnids.' Quart. Journ. Mic. Sc, vol. xxxi., pt. 3, 1890. THE HEAD CAPSULE. 107 called the gula, and of a rim surrounding the dorsal and lateral parts of the mouth, termed the epistome. The upper lip or labrum is attached to the anterior edge of the epistome, and the lower lip or labium is articulated with the anterior edge of the gula. The great compound eyes occupy a larger or smaller area on each side of the epicranium. Three simple eyes are situated near the median line on its dorsal aspect, and a pair of antennae are articulated with it in front of, or above the great eyes. Sometimes the antennae are separated from the epistome by a considerable space, which is properly termed the face. Historical ResumS.— Hitherto the head capsules of different orders have not been compared with sufficient care to establish a uniform and satisfactory nomenclature, and entomologists have contented themselves with regional terms, such as cheeks {gcnce), forehead (frons), and vertex, or have borrowed morphological terms, such as clypeus, rostrum, epistome, labrum, and labium, and applied them without due regard to the homologies of the parts designated. For example, the term clypeus, originally applied by Fabricius to the part now called the labrum, has been used indiscriminately for every part of the dorsal or anterior surface of the head by different writers. Although numerous attempts have been made to establish a uniform morphological nomenclature, so far as I can judge, the only serious one ever made to differentiate and name the sclerites of the head-capsule is the work of Robineau-Desvoidy [49], and his observations were confined to the head capsule of the ' Myodaires' (Muscida). The work of comparative morphologists has been, hitherto, entirely founded on the hypothesis that the dorsal surface of the head consists of the sternal region of several pre-oral somites, and that the antennae, great eyes, ocelli, and even the labrum, like the mandibles, maxillae and labium, are modi- fied ventral appendages homologous with the thoracic limbs. Such views rest upon no secure foundation, and have done much to retard the advance of knowledge ; they originated from the statements of Savigny, and it is only recently that any io8 THE INTEGUMENTAL SKELETON OF THE IMAGO. departure from his assumptions has found a place in recognised text-books on morphology. Balfour [43, vol. i., p. 408] , however, says ' the antennas can hardly be considered to have the same morphological value as the succeeding appendages ; they are, rather, equivalent to the paired processes of the pre-oral lobes of the Chsetopoda.' And Hatchett Jackson [44, p. 497] states that the head in insects exhibits no trace of segmentation, except the appendages of the three first post-oral metameres. Fig. 21.— The I'ostevior surface of the Head Capsule of an immature Dragon-fly (Libvlliila i/t'/iiYssa] : <•, epicranium ; /c, paracephalon ; J, jugum ; mif, mx\ and /«.v", the lateral piers of the mandibular and two maxillary segments ; r, galea of maxilla ; /*, laliium ; lb' lateral lobe of labium (galea). The mandibles and lacina; of the maxilla are not represented. So far as I know, no one has as yet attempted to show the limits of the somites which bear the mandibles and maxillaj ; even those who apparently recognise the non-segmental character of the anterior part of the head capsule arc silent on the position and relations of the three post-oral ccphahc somites ; yet these segments are sufficiently distinct to be easily recognised in the head capsule of immature specimens of Libel- lula depressa and many other insects (Fig. 21). In order to attain a true conception of the nature of the THE HEAD CAPSULE. 109 head capsule, our investigation must not be limited to a mere inspection of the sclerites of which it is composed, or to a study of the appendages which arise from it. Just as the parts of the skull of a vertebrate have a definite relation to the enclosed nerve centres, so the parts of the head capsule have similar relations to the nerve centres of the insect. It was formerly supposed, and the idea originated in the mind of Dohrn,* that the neural surface of a vertebrate and an insect correspond ; in other words, that the ventral surface of . an insect represents the dorsal surface of a vertebrate. Such an hypothesis is, however, clearly untenable in the light of modern discoveries, and may be said to be practically defunct. With regard to the nature of the pre-oral nerve centres, they were, and still are, regarded by many, as similar to the post- oral neuromeres, or paired ventral ganglia — a view which I regard as equally untenable, and reject with Dohrn's hypothesis and the theorj' that the head capsule consists of several pre-oral metameres. The older writers on insect anatomy recognised strong relationships between the arthropod and the vertebrate brain. Even as lately as 1864, Leydig [42, p. 185] pointed out that in both classes of animals the brain consists of special paired ganglia, arranged in three groups— the fore-brain, the mid-brain, and the hind-brain. To the fore-brain he attributed the function of volition, to the mid-brain visual, and to the hind-brain co-ordinating functions. Faivre stated that in Dytiscus these co-ordinating functions are located in the infra- oesophageal ganglia. Leydig included the infra-oesophageal ganglia, and regarded them as the representatives of the hind-brain. In my Presidential address, delivered at the Annual General Meeting of the Quekett Microscopical Club, February 22, 1889, I made the following statements : ' The theory of segmentation was formerly applied to the vertebrate skull, and originated in the brains of Oken and Goethe. Professor Huxley, in his lectures on the vertebrate skull, published in London in 1864, disposed for ever, I believe, * ' Ursprung der Wirbelthiere,' Leipzig, 1875. no THE INTEGUMENTAL SKELETON OF THE IMAGO. of these views, and showed clearly that the skull is developed from parts which do not undergo segmentation. My own researches in the embryology of insects have sufficiently shown me that the brain and head capsule of insects are also developed from structures — the procephalic lobes of Huxley — which undergo no segmentation. I am, indeed, convinced that there are no prestomal segments in insects. 'Viewed in relation to development, the brain in insects con- sists of a central ventricle and two hemispheres, which are themselves hollow. The central ventricle contains a transverse and longitudinal commissure, the corpus centrale, and is con- nected by its posterior wall with the median eyes or ocelli. There is thus a close correspondence between the brains of vertebrates and those of insects. The so-called antennal lobes correspond to the olfactory bulbs, the central ventricle to the third ventricle, and the ocelli to the pineal gland, or pineal eye, where the latter is developed. The hemispheres are cerebral lobes, and the pedunculated bodies are merely isolated convo- lutions of the surface.' At that time I thought that such views would be regarded as a return to ideas long laid aside, and I confess I was unable to explain the position of the alimentary canal, which led me to suspect that, however striking the analogy between the two, it must be analogy only. Gaskell's Views. — It was not until I had read Dr. Gaskell's paper on the origin of the central nervous system in verte- brates [45] that a more complete light was thrown upon the whole subject. Still further confirmation is afforded by the papers on the origin of vertebrates from a Limuloid ancestor by Gaskell and Patten [47 and 48] quoted at the commencement of this section. Both these authors bring a large amount of evidence to show that the brain in the lowest forms of vertebrates corre- sponds very closely with the cephalic nerve centres of the arthropods. The most startling consequence of these views is Gaskell's theory that the alimentary canal of the vertebrate is a new THE HEAD CAPSULE. in structure, formed on the ventral surface of the embryo, and that the original track of the alimentary canal is represented by the ventricles of the brain and the central canal of the spinal cord, which are the persistent remains of a structure corresponding with the alimentary canal in the Arthropoda. Although I feel that very great difficulties stand in the way of the complete acceptance of this theory, numerous facts tend strongly towards its verification; and my own observations cer- tainly accord very well with those of the two authors named. These facts were observed by me long before I heard of the views of Patten and Gaskell, and only needed a theory, such as they have put forth, to bring them into accord, and to show that there is a far closer relation between the vertebrate and arthropod types than has been suspected — at least, of late. I therefore avail myself of their views as a working hypothesis. Nerve Centres.^ The unexpected discovery of the formation of the primitive archenteron.as a dorsal invagination in the embryo of the larva of the fiy (p. i6), is in some sense a confirmation of the theory that the central canal of the primitive nervous system of vertebrates is the morphological representative of the alimentary tract ; it has, however, yet to be shown that the primitive neuromeres or ganglionic centres are related to this invagination in insects. Hitherto I have not been able to trace any such relationship, but I regard it as far from un- likely that the cells from which the nerve centres arise are primarily derived from the outer surface of the archenteron. In embryos not more advanced than those represented in the figure, I have been unable to detect any trace of nerve centres, but it is possible that they are derived from the invaginated part of the blastoderm, although they first appear on the inner surface of the ventral epiblast, at a later stage. The dark cellular mass which surrounds the archenteron, and extends into the head, is certainly suggestive of some such relation- ship. Leydig held that the infra-oesophageal ganglia are the repre- sentatives of the cerebellum and medulla, and it is rather remarkable that Gaskell has not apparently seen that such a 112 THE INTEGUMENTAL SKELETON OE THE IMAGO. view cannot be maintained if the alimentary tract represents the central canal of the nervous system. The great corpora fungiformia more probably represent the cerebellum. The post-oral metameral ganglia, from which all the true nerves arise, are infra-oesophageal, and probably, with the thoracic ganglia, correspond to the medulla oblongata. The loss of co- ordination due to the destruction of these ganglia is no argu- ment, as it is impossible to destroy them without destroying the peduncles of the supra-cesophageal ganglia. The manner in which the fibres of each crus diverge and form three bundles has not, so far as I know, been observed by previous writers ; these three bundles correspond to the three sections of the brain : the anterior bundle goes to the fore-brain, the middle bundle to the mid-brain, and the great posterior bundle forms the stem of the internal and external calix of the corpus fungiforme. This divergent arrangement of the nervous cords which form the crura is clearly indicative of an arrangement of the centres in front of the mouth, not in a linear series, but as a divergent series all equally related to the segmental neuromeres. Tn the brain of the more generalised Insecta — as, for example, that of the Cockroach {Pcriplanda) — the trabeculas (Balkcn) appear to me to represent the thalamencephalon, the cauliculi (Vordcrkorn) the corona radiata, and the peduncles {Hinterast) the superior peduncles of the cerebellum. The oesophagus lies below the ventricle of the cerebroid Description ok Plate V. A series of diagrams illustrating the theorctic.il structure of the skeleton in insects. Fig. I. — Side view. Fig. 2. — Median section. Fig. 3.. — Ventral view, with the integument removed on the right side to show the nerve centres : a, antenna ; a ir, anterior cephalocele ; c, compound eye ; c/>c, epicephalon ; /', /-, /', thoracic legs ; «/ cc, posterior cephalocele ; /n <", paracephalon ; s*, s^, s^, 4th, 5th, and 6th somites ; .t/, stomoda.'um ; i. to VI., neuromeres of the ventral chain ; /, antennal ganglion ; s, upper part of the crus of the cerebron ; j, corpus fungiforme ; 4, median commissure surrounded by the central ventricle ; J, cerebroid ganglia ; 6, optic ganglia ; 7, commissure of the optic ganglia. THE CEPHALO-rilOKAnc SEGMF.NTS. THE HEAD CAPSULE. 113 ganglia, and below the median commissure of the optic lobes and corpora fungiformia. The nerves to the organs of the mouth and face arise from the infra-cesophageal ganglia, whilst those which supply the auditory organ are derived from the sixth or seventh neuro- meres, or segmental ganglia, which in the insect are. the posterior thoracic ganglia. Biramous Appendages. — If, as Dr. Gaskell suggests, the limbs of the arthropod correspond with the visceral arches of the vertebrate, and the branchial clefts are the spaces between them, we should have seven such clefts between the auditory capsule and the mouth, representing the cephalo-thoracic limbs, corresponding with seven pairs of cephalic metameral nerves as the primitive number in vertebrates. That no vertebrate is known in which these arches exist does not, however, militate against the argument, as the persistence of the cranial nerves renders it certain that a number of somites have disappeared in the process of evolution. The double character of the embryonic appendages in the Crustacea, and in the maxillae of insects, as well as in the thoracic limbs of the rudimentary fly nymph, is certainly very suggestive of the double character of the pterj'gomaxillary arch, or even of the hyomandibular in vertebrates. It is true Balfour [43, vol. i., p. 542) states that in insects double or biramous appendages never exist ; but this statement is certainly erroneous, though, except in the nymph of the Diptera, the thoracic legs have not hitherto been seen in insects in the biramous stage. I cannot doubt that the biramous thoracic feet of the fly nymph, described by Weismann, and seen by myself, are indicative of a primitive biramous con- dition. Development and Morphology. — The head of an insect consists undoubtedly of a pre-oral and a post-oral portion, and the latter is formed from three segments (PI. V., Figs, i, 2, and 3), whilst the former, although divided into three distinct regions, is only segmented in a very different sense ; it does not exhibit a series of metameres. 114 THE INTEGUMENTAL SKELETON OF THE IMAGO. The pre-oral part is either developed from a cap of blasto- derni, the head-cap, or from a single pair of imaginal discs, which represent the invaginated head-cap and procephalic lobes. The structure of the head is also the same, whether it is developed from the blastoderm directly, or from the invagi- nated imaginal discs. The procephalic lobes (PI. V., pa c) are thickened lateral portions of the head-cap, and are connected with the segmented post-oral primitive band by a pair of crura which surround the stomodjeal pit. The procephalic lobes bear the great eyes (c) and the antennae (a), which are undoubtedly olfactory organs. The head capsule will be seen therefore, to consist of a segmented portion behind, which I term the ' metacephalon,' and laterally of two plates in front of the metacephalon, bear- ing the great eyes and the antennae; these I term 'paracephala' (PI. v., pa c). The paracephala are united in front, and form the epistomum and the labrum (pi). The median region behind the epistomum, or metalabrum which I regard as a preferable term, exhibits three distinct parts ; two of these are frequently bladder-like swellings, the anterior and posterior cephalocele (PI. V., a cc and p cc). The facial region is developed from the former and the forehead from the latter, the simple eyes are situated in the forehead, or on the posterior cephalo- cele. Behind the forehead there are two plates, which extend forwards from the metacephalon ; these form the epicephalon. That portion of the procephalic lobe which lies in front of the crus unites with its fellow, and curves downwards and backwards over the mouth, forming the prefacial region (mes I and pi). So that there is a close resemblance between the crura of the prefrontal lobes and the trabecula; cranii of the vertebrate. The anterior and posterior cephalocele correspond with the thin portion of the blastoderm which intervenes between the procephalic lobes. The posterior cephalocele is the fore-head (Vorderkopf) of the German embryologists. In the Dragon- flies it persists as a bladder-like swelling, on which the simple THE HEAD CAPSULE. 115 or median eyes— ocelli— are situated. When it is closed by plates of chitin these are usually, a triangular sclcrite some- times obviously divided into three, which bears the ocelli, the ' epifrontal sclerite ' {Mihi), and two flat plates below or in front of the epifrontal, which I term ' mesofrontals.' The median eyes, or ocelli, correspond very closely with the so-called pineal eyes of the lower vertebrates, and just as the pineal eyes are evanescent, so the median eyes of insects are frequently absent. They are very rarely present in the Coleoptera ; but I have one specimen of Cicindela maritima with a pair of ocelli, the only specimen I have seen in which they are present. In the Muscidffi the frontal sac consists of a great part of the posterior cephalocele withdrawn into the interior of the head, between the meso-frontals and the antennal ridge, a ridge developed by a process from each prefrontal lobe be- tween the two cephaloceles. The anterior cephalocele persists in the Dragon-flies and in the Homopterous Hemiptera as a very large sub-hemispherical protuberance, but it is more generally closed by a pair of plates, which subsequently meet in the middle line, forming the facial plate. In front of the facial plate the profronta] lobes meet in the middle line, and form the base of the upper lip, or labrum ; this portion of the prefrontal lobes curves back over the mouth. I term the whole of that part of the head which is formed by the union of the prefrontal lobes in front of the anterior cephalocele the 'prefacial region.' It is usually distinctly divided into three parts— the meta-, meso-, and pre- labrum. The former is termed the episteme; the mesolabrum forms a rostrum in many insects; the prelabrum is commonly known as the labrum. In the Diptera, at least, the pro-, meso-, and metalabrum are generally each protected by a pair of lateral sclerites, which are at first distinct, but become fused in the mature insect. It will be seen, therefore, that not only are there three primary divisions of the cephalic nerve centres, excluding the cephalic neuromeres, but three corresponding enlargements of the primitive brain case or head capsule. Plate V. shows the 8 ii6 THE INTEGUMF.NTAL SKELETON OF THE IMAGO. relations of these parts and of the segmental region of the head. The figures are diagrammatic, and only intended to give the results of a long series of observations, made partly on embryos and partly on various adult and larval insects. The head capsule is represented in its embryonic condition, and the nerve centres in the form they assume in the adult of the most highly differentiated Insecta. I regard the anterior cephalocele {a cc) as the vesicle of the olfactory lobes (/), the posterior cephalocele (p cc) as the vesicle of the cerebral hemispheres (-7) and their median ventricle. The epicephalic region (ep c) is probably more complex, and corresponds with the optic commissure, the corpora fungiformia {■!), and the dorsal arches of the first three ventral neuromeres (i., ii., m.), whilst the great optic lobes (6') apparently belong to the paracephala (pa c). The latter frequently overlap the segmental parts of the head capsule, especially in the Orthoptera, thus bearing some resem- blance to the opercula in fishes. In the Fly-nymph the median parts of the head capsule lie in a deep cleft between the two paracephala, and in close proximity to the ganglia with which they correspond, so that the head appears to be open in the middle line ; but sections show that this appearance is due to the deep infolding of the inner edges of the paracephala. I shall hereafter show that the paired lateral invaginations of the epiblast, from which it is stated the pre-oral centres are developed in some arthropods, are sensory vesicles — a view which is borne out by the development of the compound eyes, thus establishing another relation between the arthropod and the vertebrate. The nature and relations, as well as the developmental history, of the nerve centres and sense organs will be fully dis- cussed in the sections devoted to the nervous system. THE HEAD CAPSULE. b. On the Nomenclature of the Sutures and Sclerites of the Head Capsule. Before attempting a detailed description of the head capsule of the Blow-lly, I shall define the terms which may be \ery fifenerally applied to the sutures and sclerites which are found in the head capsule of an insect. In the Earwig, Foi-ficula (Fig. 22), or the Cockroach, the median epicranial suture will be readily recognised (Fig. 22, V x). Its anterior extremity bifurcates and marks the anterior limit of the paired epicephalic sclerites. The diverging branches of the epicranial suture I term ' epifrontal sutures ' (Fig. 22, x, y). On either side of the epicranial suture a faint suture (h y t) separates the para- cephalon from the median region of the head capsule. The antenna is articulated either directly or through the medium of a small sclerite — the torulus (/) — with the anterior border of the paracephalon {p). Immediately above the antenna in Forficula, and in many other insects, there is a dark spot ; it marks the junction of the internal skeleton of the head with the anterior border of the paracephalon. The occipital foramen is surrounded by a scleritic ring — the metacephalon, which consists of the dorsal and lateral parts of the three first post-oral somites. The sternal portion of these somites forms the gula, or greater part of the lower surface of the head, in the Coleoptera. In the Earwig and many Orthoptera, however, the gula is ap- parently internal. Although the axis of the head is more or less vertical, the gula remains nearly horizontal, and is concealed by the mandibles and maxilla; ; under these circumstances it is little developed in the middle line, but assumes the form of two lateral bars, which support the mandibles and join the para- cephala in front. These lateral bars are, in part at least, developed from the crura of the procephalic lobes (Fie. 22, 2, g). The base of the labium, (Fig. 22, 4, mx-) forms with the cardines of the maxillai (Fig. 22, 4, Jnx^) the greater part 8—2 u8 THE INTEGUMENTAL SKELETON OF THE IMAGO. of the under surface of the head — this I term the 'pars basilaris.' The lateral parts of the metacranial annuli some- (SEses^ F,G 22 -The Integumental Skeleton of thellead a{^x.V.^xss\g(Fo,fiadaaurHuhr,ay. fi.teral aspect; -^ interior of one lateral half; j, anterior, an, parafrontal ; v, vertex and epifrontal plate. trapezoidal in the female, and narrow and triangular in the male. In the male the whole head is smaller, and the eyes are more obliquely placed. Seen from in front, the head capsule of the mature imago exhibits the following parts : the frons, the face, and the epistome in the middle line ; with the anterior margins of the paracephala and the great, or compound, eyes on either side. THE HEAD CAPSULE. 121 The Frons exhibits three sclerites, a triangular epifrontal, and two subquadrate frontals (Fig. 23, j, /). The three simple or median eyes are situated near the angles of the epifrontal. On either side of the frontal plates there is a narrow sclerite, the parafrontal ; it is formed in the anterior edge of the paracephalon. The Face assumes the form of a semi-elliptical shield ; it is bounded above by the antennal ridge, and below by the epi- stome ; the large third joints of the antertnse almost conceal it. Two sclerites are developed in this region— my mesofacials, the fovese of Desvoidy ; these are distinct in the young imago, but become subsequently fused into a single plate (Fig. 24, /, m/). The Antennal Ridgfe is a distinct sclerite in the young imago, continuous with the facial edge of the paracephalon ; it presents a median process, which descends between the antennfe, with a deep emargination, on either side. This ridge corresponds with the toruli and the antennal rostrum; it is distinct in the young imago, but becomes fused with the mesofacials in the adult (Fig. 24, /, r). The lunula is seen immediately above the antennal ridge. The anterior edges of the paracephala overlap the face slightly, and form a setigerous chitinized fold on either side of the face ; these are the facialia of Desvoidy. The Epistome (Fig. 23, j, c) is but little developed in the Blow-flies; it consists of a narrow, somewhat salient, ridge, join- ing the facialia below the face. In the Sjrphida; it forms a large beak-like shield, in which several distinct paired sclerites are developed. The epistome in the Cicadae is very similar to that of the Syrphida;. The Paracephala exhibit three distinct sclerites in front, the parafrontals, the facialia, and the genai. The latter appear to be the anterior and lower part of the great lateral plates of the head, extending forward beneath the great eyes. Fig. 23, j, exhibits a lateral view of the head of an immature imago ; the folds of the paracephala, from which the parafrontals, the antennal ridge, the facialia and the epistome are subsequently developed, are very distinctly seen. 122 THE INTEGUMENTAL SKELETON OF THE IMAGO. The Antennae hang vertically in front of the mesofacial plates; they are usually said to consist of throe joints and a bristle! They are really six-jointed, as Desvoidy pointed out, with a greatly enlarged third joint, the three terminal joints being reduced to the form of a three-jointed bristle. The large third joint is cylindroid, and slightly prismatic. It contains a complex sensory organ, which is undoubtedly olfactory (see ' Sensory Organs '). The Posterior Surface of the head capsule consists of the lo 24 _/ 1 he I'ace of the Jilow-Hy : ep, epislome ; / facial edge of paracephalon ; /«, unula ; mf mesofacial plate ; r, rostrum ; (r, torulus. 3, the Metacephalon of the same : h, basal scleritc orgula ; bs, basal suture ; cc, cotyloi.l cavity ■ ,ir dorsal ring ; ep 0, epioccipital sclerites ; j^ andy-', first an,l secon, the palpus ; pi, prelabrum ; /, ligula ; pa, paraphysis ; d s, discal sclerite ; //; s, thyroid sclerite ; ep f, epifurca ; s s, sesamoid sclerite ; st, stomal or hypoglossal sclerite ; s v, salivary valve ; f, fulcrum ; //•' to tr*, trachea; ; /«' to /«", muscles ; s g, saliv.iry gland of the oral disc. Part of the rostrum only is represented, i j to ep f is the haustellum, consisting of the theca, prelabrum and disc. THB PROBOSCIS OF THE IMAGO. THE EXO-SKELETON OE THE PROBOSCIS. ,31 which the palpi are always developed.' He regards the palpi of the Diptera as labial, and not as maxillary, a conclusion which appears to me to be unwarranted. They are, without the slightest doubt, maxillary palpi. If the proboscis of the Diptera represents the first, and not the second, pair of maxilla, the four-jointed sheath of the Hemiptera must unquestionably have the same morphological value. I have repeatedly examined the mouth organs of the larger Cicada, and the conclusion at which I have arrived is, that the sheath exhibits no indications of its morphological identity with either the labium or the maxilla which would enable one to judge whether it represents the one or the other ; but there is no evidence that both pairs of maxillae are developed, or that either pair of setas should be identified as representatives of the first pair. Just as nearly everyone has accepted the statement that the greater part of the proboscis of the Diptera represents the labium, so they have accepted the statement that the sheath of the lancets in the Hemiptera is a modified labium ; yet this view which Savigny initiated depends mainly upon the supposed identification of the second pair of setse with the maxillas. Against this identification, I would observe, there is no evidence that a seta can be homologous with a maxilla, and the manner in which these setae are connected with the head shows at once that they are not the maxillae. Latreille identified them with the terminal lobes of the maxillae, a much more correct view of their homology.* Before entering upon a detailed examination of the proboscis with a view to elucidate its real nature from a morphological point of view, it is essential to acquire a definite idea of the nature and structure of the labium and maxillaj in insects. The Labium.— Erichson [52] discusses the mouth parts of insects ; in especial relation to the structure of the labium, or lower lip, he says : ' The third pair of jaws, in the Insecta, form a considerable * Cuvier, ' R6gne Animal,' Nouvclle edit., 8vo, Paris, 1S29, torn, v., p. igo. 9 ■ 32 THE INTEGUMKNTAL SKELETON OF THE IMAGO. portion of the under-lip, which is composed of these united with the mentum, or chin, and the tongue {ligiila). The third pair of jaws have the chin behind and below them, and the tongue above and before them (between their terminal lobes on their oral surface) and always united with them through its coalescence with the remarkable labial palpi.' The nature of this tongue has led to much discussion. Savigny called it the ' hypopharynx,' and erroneously supposed it to be a mere process of the floor of the pharyngeal cavity. It is really always a tongue-like process of the floor of the mouth, which either overhangs or surrounds the extremity of the duct of the great salivary (lingual) glands, and it arises from the base of the labium. It may be fleshy as in the Cock- roach, reduced to the form of a hollow seta as in the Diptera, or many-jointed as in Bees. It is frequently fringed by tactile or gustatory bristles and papillas. When Savigny named it the ' hypopharynx ' in the Flies, he entirely overlooked its real nature. In the more generalised Insecta, although the appendicular portion of the labium, the modified second pair of maxillae, con- sists of parts which can be more or less readily recognised as corresponding with the parts of the first pair, they are never so large or complex, and there is a distinct tendency for them to become obsolete in all the more highly modified forms. In the Lepidoptera nothing remains but their palpi, and in the Bees they are reduced to the form of a pair of scales, the paraglossa:, and a pair of long-jointed palps. The palpi are apparently the last parts which remain in the most highly modified types, and even these are generally regarded as obsolete in the Diptera and Hemiptera— in which, according to received views, the second pair of maxillae (labium) play such an important part in the structure of the suctorial mouth. The MaxiUffl (or first pair of maxillae). — Brulle very carefully figured and compared the maxillai of numerous insects. I give a copy of some of his figures of the maxillae and labium of a few well-marked types (Fig. 26). A typical maxilla consists of THE EXO-SKELETON OF THE I'KOBOSCIS. .33 a basal joint, the cardo, c, which supports the stipes, s, or second joint. Three lobes are articulated with the stipes— the first, or external lobe, is usually reduced to a small scale which sup- ports the palpus ; it is the palpiger, /. The second lobe, or upper lobe of Kirby, frequently forms a large hood in which the third, or internal, lobe lies when at rest. This second lobe is the galea, g. It almost always consists of, at least, two joints, and is usually soft and fleshy. In a few Coleoptera, as the 1' IG. 26. — Details of the Maxilla and Labium, after Brulle : /, maxilla of Locusta viiidissima ; 2, of Blaps ; j, of Pepsis sp. ; 4., of Xylocopa violacea ; s> of "n Australian species of .'Eschna : ,?•, galea; /, lacina : g' , sous-galea ; s, stipes; />', palpiger ; c, cardo ; 6, labium of Copris Isidis : /', lacina or inner blade of second maxilla ; lig, ligula ; ^, outer lobe (galea) ; /, palpus. Tiger-Beetle {Cicindela), the galea is palpiform, and is usually termed the second maxillary palpus [Newport, 9, p. 8go]. The third lobe, or lacina, I, is frequently a cutting blade, and some- times has a small claw, the uncus, articulated with it near its extremity. In some phytophagous insects the lacina is con- verted into an obtuse lobe covered with setae. The improbability that so complex a structure can become a mere seta is, to my mind, very great ; but when it is 9—2 134 THE I NTEGU MENTAL SKELETON OF THE LMAGO. entirely removed from the part of the head to which the maxilla is invariably attached in manducatory insects, it is impossible to admit its homology with that organ. A seta might represent one of the setse of the lacina.or of the galea, but not the maxilla itself. The following synonymy will be useful to those who desire to study the maxilte of insects generally : Cardo, Kirby and Spence. Style, Audouin. Sous-Maxillaire, UruU^. Stipes, Kirby and Spence, Burmeister. Maxilla, Newman. Maxillaire, BruUd. Lacina, Macleay. Stipes, Erichson. Mando, Burmeister. Inter- niaxillaire, Straus Durckheim and Brulld. Palpiger, Palpigcre, Straus Durckheim, Burmeister and Audouin. Galea, Fabricius. Upper lobe, Kirby and Spence. Uncus, Kirby and Spence. Premaxillaire, Brulld. b. The Sclerites and Morphology of the Proboscis of the Blow-fly. The proboscis of the Blow-fly consists obviously of three parts : a basal portion, the rostrum ; an intermediate portion, the haustellum ; and a terminal portion, the oral sucker. In the use of the term ' rostrum ' I have followed Fabricius ; ' haustellum ' is also one of his terms, but I have used it in a different sense to that in which he employed it. Kraepelin terms this portion of the organ, with the oral sucker, the true proboscis, but no definite terms have hitherto been applied to the several parts. The Rostrum or basal portion of the proboscis is a mem- branous cone, attached by its base to the epistome, the genai, and the pars basilaris of the head-capsule. It supports the haustellum at its apex. The rostrum is somewhat flattened on its anterior surface, and presents a horse-shoe-shaped sclerite, which forms a hinge on the epistomum, being connected with its oral margin by syndesmosis. I term this sclerite the clypeus. In using the word 'clypeus' in this sense I have followed the nomenclature usually adopted in the Diptera (see page 43). The clypeus articulates with the labrum through the intervention of a tract of syndesmotic integument, and as Fabricius applied the term to the whole labrum, this use of it is not inappropriate. THE EXO-SKELETON OF THE PROBOSCIS. '35 The Clypeus (Figs. 2 j, j c, and 25, 6) forms the external dorsal portion of the fulcrum, or cephalo-pharyngeal skeleton. This supports the haustellum at its distal extremity. The Palpigerous Scales and Palpi. — On each side of the rostrum, near its apex, there is a convex scale (PI. VI., p s), which sup- ports a single-jointed clavate palpus — the maxillary palpus — at its inner proximal angle. These scales are united in many of the Diptera by a curved plate of chitin, which embraces the posterior surface of the rostrum. Sesamoid Sclerites. — I apply this term to two small nodulated fusiform rods of chitin close to the apex of the rostrum on its Fig. 27. — A portion of the Proboscis of the Blow-fly, showing the relations of the prelabrum with the theca ligula and pharynx : ap, apodeme of the labrum ; pli, the pharynx ; /y, hyoid sclerite ; ///;, prepharyngeal tube ; >■ >; part of the rostrum ; ///, the prelabrum ; lig, the ligula ; // /; h, the haustellum ; ,c,(, sesamoid sclerite ; //' s, thyroid sclerite ; vi »i, muscles ; Ir, trachea ; IJ, lingual, and sd, salivary ducts ; sv, salivary valve. ventral aspect. They give insertion to the long retractor muscles of the rostrum (Fig. 27, ss). The Haustellum, or middle joint of the proboscis, consists of two valves, a large ventral or posterior valve, which I shall term the thcca, and a narrow styliform dorsal or anterior valve, the prelabrum. The prelabrum lies in a groove on the anterior face of the theca, the edges of which overlap it. The theca ends distally in the great two-lobed oral disc, or sucker. The Oral Sucker is a fleshy oval disc, deeply cleft at its anterior 136 THE JNTEGUMENTAL SKELETON OF THE IMAGO. margin. The edges of the cleft are continuous with the margins of the groove in the theca, and are united as far as the edge of the disc by a remarkable bead and channel joint — the thick edge of one lobe of the disc fits into a corresponding cylindrical channel in the other (Fig. 31). The distal surface of the disc is channeled by the well-known 'false tracheae.' In the centre there is a deep longitudinal fissure, which extends into the tubular mouth situated between the KiG. 28. — /, Tlie fulciuin : a, proximal, and /', cHstal cornua ; c, median raphe ; (/, chitinous plalc, uniting the clypeus with the cpistome ; 3, side view of the same : J>>i, pharyngeal tube ; //, lateral plate ; j, transverse section through the rostrum : ///, pharynx ; ////, mesolabrum ; a/>, apodeme of the |irclabrum ; iiix, maxilla ; ///, muscles ; Ir, trachea; ; ./, a vertical section of the mouth : //, hyoid ; //, prepharyngeal tube ; r, inner plate of the prelabrum ; /, ligula ; sii, salivary (lingual) duct ; si, stomal or hypoglossal plate ; il s, discal sclerite ; «, nodulus ; c/i, cochleariform sclerite ; s s, sesamoid sclerite ; 5, the furca. labrum and the theca. The proximal surface of the sucker is convex and covered by setae ; those near its margin are ver\- long and form a fringe. Morphology of the Rostrum. — Transverse sections of the rostrum show that it consists of two parts (Fig. 28, j), a median anterior or dorsal tube {ml), which is continuous with the cavity of the labrum — this represents a mesolabrum, and THE EXO-SKELETON OF THE PROBOSCIS. 137 a latero-ventral tube (inx), which has the same relation to the mesolabrum that the theca has to the prelabrum. In the nymph, at an early stage of development, the separation of the pre- and mesolabrum is not apparent, and the whole lies in a groove, so that the mouth at this period extends back to the epistome ; it then very closely resembles the labrum and sheath of the Hemiptera. Subsequently the fulcrum is developed between the meso- labrum and the grooved part of the rostrum, and the meso- labrum becomes inseparably united with the ventral portion of the rostrum by the fusion of the outer and inner plates of the lateral walls of the fulcrum (Fig. 28, '.i Ip). I regard the latero-ventral portion of the rostrum as the basal portion of a pair of united maxilla; ; it bears the palpigerous scales and the maxillary palpi. It is probable, however, that part of the thin integument on its posterior or ventral aspect is common to the maxillae and the rudimentary labium, so that it may be regarded as an extension of the ventral integument between the insertion of the labium and the maxillae. In the more generalised Insecta (Fig. 22, 4) it is easy to see that the mere obliteration of the deep grooves between the bases of the labium {mx-) and the maxillae {mx^') would result in the formation of a rostrum similar to that of the Blow-fly. Indeed, the fusion of the bases of the maxillas and labium is far more complete even in the most generalised forms than has hitherto been supposed. The Fulcrum, or cephalo-pharyngeal skeleton (Fig. 28, / and 2), has been compared by Kraepelin [70] to a Spanish stirrup-iron with a double foot-plate. It may be described as consisting of a flattened tube, the pharyngeal tube, formed by two pharyngeal plates, a dorsal or anterior, epipharynx, and a ventral or posterior plate, hypopharynx. The epipharyngeal plate exhibits a strong median raphe, which projects into the pharynx. A pair of horns, cornua (a, b), project at either end of the pharyngeal tube, and give insertion to several muscles. On either side of the pharyngeal tube there is a somewhat triangular lateral plate (//>), formed by the adjacent walls of the 138 THE INTEGUMENTAI. SKELETON OF THE IMAGO. mesolabrum and the base of the maxilla. This plate consists of two laminae towards the pharynx, but the two are inseparably united throughout the greater part of its extent. The anterior or dermal edges of the lateral plates of the fulcrum are con- tinuous with the clypeus, so that the mesolabrum is bounded by the clypeus, the lateral plates of the fulcrum, and the epi- pharyngeal plate ; it contains the dilator muscle of the pharynx, which, by withdrawing the epipharyngeal from the hypo- pharyngeal plate, draws fluid through the tubular mouth into the pharynx, so that the fulcrum is a powerful instrument of suction. It also forms a movable support for the whole haustellum. Morphology of the Fulcrum.— Gerstfeldt [55] is, I believe, the only author who has anticipated me in the statement that the maxilla; enter into the composition of the fulcrum, but he merely observes, 'The anterior lancet (labrum) shows distinctly, by the presence of a median raphe, that it is formed of two halves, which must be the blades of the maxilla {Kicfcrladcn). They rest upon a piece extending backwards (the fulcrum), which appears to be the united stipites, from which two slender, nail- shaped parts diverge downwards and backwards. These Straus termed glosso-pharyngeal apophyses in Melolontha, they are analogous with the cardines ' [55, p. 25] . I so far agree with Gerstfeldt as to regard the outer portion of the lateral plate of the fulcrum as a portion of the stipes of the maxilla. His other statement, that the nail-shaped processes represent the cardines, is undoubtedly incorrect, and the fusion of the blades of the maxillae with the labrum is by no means proved. Although I formerly held this view myself, I now think it far more probable that the lacinae are undeveloped in all those flies in which only two median lancets exist. The lacina; of the maxillas are present, however, when paired lancets enter into the composition of the mouth armature, as in the Tabanidae. Menzbier [60] states that the fulcrum is developed in the chitinous lining of the stomodaeum, and regards it as an internal organ. This view is inconsistent with its developmental history THE EXO-SKELETON OF THE PROBOSCIS. 139 and its connection with the clypeus. It has, however, been very generally adopted. Retraction and Extension of the Proboscis. — The proboscis is an erectile organ. It is llaccid and folded on itself when not in use, but is capable of being rendered rigid by the injection of air into the extensive tracheal sacs which lie in its cavities. When at rest the only rigid parts are the fulcrum, the prelabrum and the theca. The two lateral halves of the flaccid stomal disc are folded together, so that the stomal surface is concealed. The disc is also flexed on the ventral surface of the theca. The haus- tellum is folded forward on the rostrum, so that the back of the labrum hes against the clypeus, and the fulcrum turns upon the epistome, so that its dorsal surface looks downwards and back- wards. In this position it lies entirely within the head capsule, and the maxillary palpi rest one on either side of the haustellum, which is also more or less completely withdrawn into the head capsule. The thin integument of the ventral surface of the rostrum is invaginated into itself, and lies in folds above and behind the retracted proboscis. When the proboscis is in use, it is projected from the head capsule, and rendered stiff by the injection of air into its trachea. Its varied movements are brought about by the action of muscles; the flaccid proboscis is also folded and withdrawn into the head capsule by means of retractor muscles. The nature and precise mechanism of the move- ments of the proboscis will be further discussed in another section. The Prelabrum is the rostellum of Fabricius, the labrum-epi- pharynx of Mcnzbier and most recent authors. It is a hollow prolongation of the mesolabrum, and lies in the groove of the theca. It exhibits two pairs of sclerites, one pair on its dorsal convex surface and one pair on its oral concave surface ; the latter form the roof of the long tubular mouth. These sclerites are all fused in the fully-developed imago, but the two dorsal are readily separated from the two oral sclerites in the young imago, and also in the skeletons of mature insects prepared with caustic soda or potash. 140 THE INTEGUMENTAL SKELETON OF THE IMAGO. The separation of the sclerites forming the upper and lower surface of the labrum gave rise to the view that one pair repre- sent the labrum and the other the epipharynx. As the labrum does not consist of a single plate in any insect, but is always a hollow process, this is not a tenable view (Kraepelin [70]). The external (dorsal) pair of sclerites are fused together at their proximal extremity, where they articulate with a pair of Fig. 29.—/, A section throiigh the mouth in the plane a little below that marked by the line pi, Plate VI. ; w, cavity of the prelabrum containing muscles ; «, edges of the groove in the theca ; The trophi of a Flea removed from the head and seen from behind : a a, hinge of the maxillae ; *, support of the ligula ; //, palpi ; p s, palpigerous scale ; /, lacina ; lig, ligula ; psl, pseudolaliium. 2, A portion of the same, seen from its inner side : j^, process of the hinge supporting the pseudolabium ; c, articulation with the peristome ; /, process of the hinge which strengthens the palpigerous scale ; il, articular head supporting the lacina ; f, articular extremity of the lacina ; /;, the base of the pseudolabium cut through ; the other letters as which are exceedingly confused and conflicting, I must refer my reader to Taschenberg and Gerstfeldt [55] . That both pairs of setas in the Diptera are parts of the 154 THE INTEGUMENTAL SKELETON OF THE IMAGO. maxillae is sufficiently obvious in Tabanus, where they are supported by a massive hollow cardo, or basal joint, and also in the curious sand-fly (Simula rcptans), one of the Biblionidse, the female of which is a ferocious blood-sucker. In these insects the palpiger bears a broad trenchant blade, and the lacina is strong and serrated as in the Fleas. Synonymy of the mouth organs of the Diptera ; the following are the most important : Rostrum, Fab. Kopfkegel, Kraepelin. Russelstiel, Graber. Labnun, Clypeus, Fabricius. Properly prelabrum, Mihi. Operculum, Mihi [62]. Labrum of authors generally. Labrum-epipharynx, Menzbier. Fulcrum, Menzbier and Dininiock. Pharynx, Meinert. Palpiger, Mihi. Mandible. Superior lancet. Lacina, Mihi. Inferior lancet. Maxilla. Palpus, Maxillary palpus, Kirby and Spence, Gervais. Labial palpi in some Diptera. Savigny. Ligula, Hypopharynx, Savigny. Pseudolabium, Mihi. Labium, Savigny, and generally. Oral Disc, Mihi. Labial palpi, Burmeister, Erichson and Kraepelin. The other parts mentioned in the above description have not hitherto received names except where the synonymy has been already indicated. 4. THE THORACIC EXO-SKELETON. a. Oeneral Morphology. The Thoracic Exo-skeleton exhibits great uniformity, even in the most dissimilar insects. It always consists of three highly modified somites, which Audouin [38] named the pro-, meso-. Bibliography : — 72. JURINE, L., 'Observations sur les ailes des Ilymenopteres.' Mdm. Acad. Sc, Turin, xxiv., 1820. 4to. 73. Chabrier, J., ' Essai sur le vol des Insectes.' Mdm. du Museum, torn. vi. to viii., 1820-22. Paris, 4to. THE THORA CIC EXO-SKELETON. 1 5 5 and metathorax, to which the first abdominal somite, Latreille's segment m6diaire, is added in many of the Hymenoptera. The nomenclature of the several sclerites which form the walls of these somites is still unsatisfactory and confused. This has arisen partly from the difficulty of defining the limits of the somites, but chiefly from an attempt to make the three somites conform to an ideal type which originated in the mind of Audouin [38 and 39] . Audouin examined the meso- and meta- thorax in a great number of insects, and found the same parts represented, more or less distinctl}', in all ; and there is no doubt that the meso- and metathorax, when each has a pair of wings, exhibit a close resemblance to each other. The attempt, however, to identify the sclerites of one with the other is not always successful, and when this process is extended to the prothorax it leads to nothing but confusion. Audouin took the most highly-modified wing-bearing seg- ments as his type, and expected to find the same structures in the less highly - modified wingless prothorax. Instead of deriving the more specialised from the more generalised, he set up an ideal highly specialised type and attempted to make the less specialised somites of the body conform to his ideal. 74. Latreille, p. A., ' De quelques Appendices Particuliers du thorax de divers Insectes.' M^m. du Museum, torn, vii., 1821. 75. Macleav, 'An Explanation of the Comparative Anatomy of the Thorax of Winged Insects, with a review of the present state of the Nomenclature of its parts.' Zool. Journ., vol. v., p. 145, 1830. 76. SCHIODTE, J. C, 'Bidrag til Kundskab om Insekternes Thorax, nied fortriiiligt Hensyn til Latreille's Theorie om Segment m^diaire og til Forekomsten og Fordelingen af Spiracula Thoracica.' Overs. Danske Vidensk. Sclskabs Forhandlinger, 1856, p. 135. An English rdsumd in Ann. Nat. Hist., ser. iii., vol. xv., p. 483. 77. LoEW, Herman, ' Monograph on N. American Diptera,' vol. i., 1886. 78. Hammond, A., 'On the Thorax of the Blow-fly.' Journ. Linn. Soc. Zool., vol. XV., 1S79. 79. (}oscH, C. C. A., 'On Latreille's Theory of " le Segment Mddiaire "' (in English). Schiiidte Nat. Tidsskr., Rieke iii., Bd. xiii., 1881-83. 80. Brauer, F., ' Ucber das Segment Mddiaire Latreilles.' Sitzungb. Akad., Wien.. Bd. 85, 1882. 81. OsTEN - Sacken, C. R., 'An Essay on Comparative Chictotaxy.' Trans. Entoin. Soc, Lond., 1884. Originally published in Mitth. der Miinchener Ent. Ver., vol. v., 1881. 156 THE INTEGUMENTAL SKELETON OE THE IMAGO. Audouin's own words, when translated, are : ' If we wish to study the anatomy of an insect's thorax, we ought, after dividing it into three segments, to seek for a sternum on the inferior surface of each ; for an episternum, a parapteron, and an epimeron on the flank. We should also search for an entothorax, a peritreme, and a trochantin. I say that we should seek for them, not that we should find all these in each insect ; very generally their union is so complete and intimate that they cannot be demonstrated, but as in other cases these pieces are present, it is more rational to conclude that the same material is utilised in all than to suppose new creations are perpetually occurring ' [39, p. 126]. Audouin's ideal seg- ment of the exo-skeleton has been generally adopted as typical ; and each segment of the thorax is said to consist of a sternum, two lateral plates, the episternum in front and the epimeron behind united by an oblique internal ridge, and of four dorsal plates, one in front of the other, named respectively the prescutum, the scutum, the scutellum, and the post-scutellum. The united episternum and epimeron are also termed the pleuron, and the four dorsal plates form the tergum— hence a segment is described as consisting of a dorsal arch, the tergum, a ventral arch, the sternum, and of a pair of pleura between the dorsal and ventral arches. The term ' pleuron ' has been unfortunately applied in quite a different sense in the Crustacea— in these it means a lateral prolongation of the dorsal arch, which forms the gill cover in the decapods. By the rule of priority the term should certainly be used as Audouin used it ; I shall therefore call the united episternum and epimeron the pleuron, and suggest that the pleuron of the Crustacea should be called the epipleuron. In the prothorax a sternum and a pleuron are still recog- nisable, but it is doubtful, I think, how far these correspond with the same parts in the meso- and metathorax. There is usually a dorsal arch formed of one or more sclerites in all three thoracic somites, but sometimes this is reduced to a mere rim, and the homologies of the several dorsal plates of the meso- and metathorax are all exceedingly doubtful. THE THORACIC EXO-SKELETON. 157 Audouin's ideal segment in its simplest form represents the type of a wing -bearing somite, but if we seek the more primitive condition, it must be amongst terrestrial forms in which the skeletal ring in each somite consists of a dorsal and ventral arch, each developed from two lateral halves. Macleay [75] supposed that each thoracic segment consists of four united sub-segmental annuli, but, so far as I can see, this view is entirely unsupported by facts, and the whole evidence of development is adverse to it. Another, and perhaps more tenable, view is held by Patten [48] . He regards each segment as the result of the fusion of two primary somites or metameres. This hypothesis origi- nated from the fact that the ventral lateral appendages of the somites are frequently bifurcate, or consist of an exo- and an endopodite, in the more generalised Arthropoda, a character which persists in the maxillae of insects, and, as I have already observed, and shall show hereafter, also in the thoracic legs of the fly at an early period of development. In support of his hypothesis. Patten [48] makes the following statements : 1. ' In Scolopendra each neuromere, or pair of segmental ventral ganglia, has four pairs of nerves, two probably motor and two sensory ; and in ail arthropods, the neuromeres of which have been carefully studied, each exhibits two transverse commissures. 2. ' In Acilius each segment has two pairs of tracheal open- ings, spiracles, one pair easily seen near its anterior margin, and one pair very rudimentary and difficult to recognise near its posterior edge. 3. ' The double character of the segments of Julus is evidenced according to Heathcote by the duplication in each segment of the cardiac ostia, arteries, neuromeres, trachea;, and legs ; whilst in Scorpio the neuromeres present a distinctly double character.' Whilst admitting the apparent validity of Patten's arguments, I would observe, however, that his hypothesis is not supported by developmental evidence, so far as the Insecta are concerned, 158 THE INTEGUMENTAL SKELETON OE THE IMAGO. and the three thoracic segments are always represented by three metameres, both in the larva and in the embryo. Although a ventral groove appears in the somites of the abdomen in the larva of the Muscidae (see Fig. 4, a and d), nevertheless, each thoracic segment in the imago is developed from two pairs of imaginal discs only. I am also unable to see that the structure of the thorax of the imago is more easy of interpretation by the assumption that each segment consists of two united rings. Yet I cannot regard the supposition as unsupported, although I am unable to accept it as proven. The Ventral Appendages of the thorax, or legs, are divided into a series of joints or segments. The basal joint is known as the coxa, or hip ; the second is the femur, or thigh* ; the third the tibia, or shank ; and the terminal joints, usually five in number, form the tarsus, or foot. The bifurcate form of the ventral appendages already al- luded to appears to be a common phenomenon in all the more generalised forms of arthropods. At first sight, although apparent enough in the maxilla;, this condition seems to be entirely absent in the thoracic legs of insects. It is usual to consider that one of the divisions of the limb, that corre- sponding with the expedite in the Crustacea, is suppressed. During the evolution of the thoracic limbs, in the imago of the Blow-fly, from the lower thoracic imaginal discs, they are, however, bifurcate in what I shall term the third stage of development. At this stage the coxa, femur, and tibia are repre- sented by an elongated sac (Fig. 34, ~, and -/, ex), which is united at its open extremity with both the thoracic wall and the five- jointed tarsus. A longitudinal septum and two constrictions afterwards appear and separate the femur and tibia. It would seem, therefore, that the femur and tibia are developed from the outer limb of a bifurcated appendage, and the coxa forms a basal joint common to both parts of the limb. This remarkable phenomenon was first observed by Weis- mann [2, p. 167], but it appears to have attracted little or no attention since. * 1 regard the trochanter as part of the femur. THE THORACJC EXO-SKELETON. 159 If we compare the thoracic leg of an insect with that of a crayfish, we recognise in the five joints of the insect's tarsus the representatives of the five joints of the crustacean Hmb known as the basipodite, ischiopodite, meropodite, carpopodite, and propodite, whilst the claw-like dactylopodite is represented in insects by two tarsal claws. In the crustacean the basal coxopodite corresponds with the coxa of the insect, whilst the exopodite of the generalised Crustacea corresponds with the fenioro-tibial portion of the limb in the insect. The bladder-like femoro-tibial rudiment closely resembles Fig. 34. — Kive stages in the development of the leg in the nymph, showing the manner in which the femoio-tibal joints are formed from the exopodite. /, The rudiment of a limb in the second stage of development. ^, The same in the third stage. 7, 4, and ^, Three successive stages of the same : e, coxopodite ; ex, exopodite ; eit, endopodite ; s, sternum ; c ', coxa ; /, femur ; /, tibia ; / to 5, tarsal joints. the exopodite of some Crustacea, and the rudimentary limb of the fly-nymph takes us back to the primary bifurcate condition still retained in the thoracic limbs of the generalised Crustacea. This much is certain, the five tarsal joints in the Blow-fly are all differentiated distinctly before any trace of the segmen- tation of the femoro-tibial portion of the limb is apparent, and this is developed from a process which closely resembles the exopodite of the crustacean limb. The Dorsal Appendages of the thorax, or wings, are highly i6o THE INTEGUMENTAL SKELETON OF THE IMAGO. characteristic of tiie imago in the Insecta, but are confined to the meso- and metathoracic somites. The wings are sac-like prolongations of the syndesmotic integument between the dorsum and the pleuron ; each wing, therefore, exhibits two layers of thin integument one above the other. These are closely united in the mature adult, but in the immature imago or fully - developed nymph they are separated by a layer of spongy cellular tissue, of a reticular character, the spaces of which are blood sinuses continuous with those of the thoracic cavity. In the nymph this tissue is permeated by a brush of dichotomous sub-parallel tracheal capillaries. The upper and lower layers of integument are termed the upper and lower wing membranes. Diagrammatic repre- sentations of the wing-bearing somites are given in most works on elementary comparative anatomy. I shall not, therefore, reproduce them here. The Homology of the Wings. — The wings are developed from the edges, epipleura, of the dorsal plates of the meso- and meta- thorax. The edge of the dorsal plate of the prothorax closely resembles the epipleuron of the Crustacea, but never attains the characters of a wing. In the Cockroach larva both the meso- and metathorax have precisely similar edges to the dorsal plates ; within these the wings are developed. The rudi- mentary elytra of the female Cockroach are manifestly a modification of the edge of the tergum of the mesothorax. In many aquatic larvae leaf-like appendages occur on both the thoracic and abdominal dorsal plates, in the position of wings. These leaf-like appendages contain tufts of tracheae, and have the function of gills. Graber says : In the young larvae of the Termites, which live in damp places, similar tracheal gills occur on the thoracic and abdominal dorsal plates, and the development of the wings is effected by the enlargement and modification of those which belong to the meso- and metathorax [10, Bd. i, p. 190]. The tracheal gills of insects are, moreover, very similar to THE THORACIC EXO-SKELETON. l6l the dorsal gills of the annelids, so far as their position is concerned. It appears probable, therefore, that the wings are a modification of the respiratory organs of some ancestral form. Although our knowledge of the more generalised Insecta goes far to render it certain that the earliest forms of insect life were entirely terrestrial, and that the aquatic habits, like the vermiform condition of the larvae, are acquired, or inter- calated modifications, the conditions observed in the Termites, and the existence of temporary tracheal brushes in the rudi- ments of the wings of the fly-nymph, show that the terrestrial habits of the earliest and most generalised insects cannot be used as arguments against the view first enunciated by Gegen- baur,* that the wings of insects are modified gills which have entirely lost their respiratory, and have assumed a new function, that of aerial locomotion. Development of the Wings. — The wings first appear as papillae of the epiblast, which soon exhibits a cavity filled with stellate mesoblast and blood. This sac-like wing becomes broad and flattened, so that it presents an upper and a lower wall, and a tuft of nearly parallel and very numerous tracheal vessels is developed in the wing cavity. The walls of the wing sac become plicated in fan-like folds radiating from its attachment to the thorax. The angles of the folds become thickened, and form the primary nervures. Achitinous epidermis is deposited on the surface of the wing sac. In many nymphs this attains considerable thickness, and appears externally when the larval integument, under which the wing is formed, is shed. The first cuticular layer does not persist in the imago, but is shed after a second cuticular layer — the permanent cuticle of the wing which becomes the wing membrane — is deposited. The nervures are developed on the convexities of the primary folds, and in the corresponding grooves of the opposite surface of the wing sac. The tracheal vessels of the wing remain only in the young * ' Grundziige der Vergl. Anat.' i62 THE I NTEGU MENTAL SKELETON OF THE LMAGO. nymph ; they are afterwards absorbed. When the imago emerges from the pupa, or when the cuticle which covers the nymph is shed, the mesoblast and hypoderm of the wing are both present, so that the wing is still a sac communicating with the thoracic cavity and permeated by a rich plexus of blood sinuses. It is small and thick. After the emergence of the imago the wing rapidly expands in length and breadth, at the same time decreasing in thickness. The upper and lower surfaces are drawn together by the contraction and atrophy of the mesoblast, and subsequently of the hypoderm, except at its junction with the thorax and in the course of the largest nervures ; where, the wing cavity persists, and its mesoblast develops small tracheae, blood sinuses and tendinous cords connected with the smaller wing muscles. The greater number of the nervures become solid chitinous rods, and appear as convex ridges on both surfaces of the wing. Between these the wing membranes come into contact, and all the epiblastic and mesoblastic cells disappear. Structure of the Wings. — The wing membranes are supported and extended by branching hollow or solid nervures, which form a reticulated framework. These are chitinized thicken- ings of one or both wing membranes. The number of nervures, their manner of branching, and the extent of the reticulation between their branches, vary greatly, but there is a general agreement in the arrangement of the principal nervures at the attachment of the wing with the thorax in all insects. In the Dragon-flies the four wings are alike, and these insects exhibit a very generalised type of wing. In other insects the wings deviate more or less from this type, without, however, altogether departing from it. The wings of the Dragon-flies exhibit an anterior and a posterior margin, a rounded apex, and a truncated base which corresponds to its thoracic attachment. The anterior part of the base is folded fanwise, so that it exhibits three ridges and two furrows above, and two ridges below. This portion of the wing is strengthened by five strong nervures, one corresponding to each ridge. THE THORACIC EXO-SKELETON. 163 The posterior part of the base remains membranous, and forms an axillary fold which checks the forward movement of the wing. I shall term the dorsal nervures costal, sub- costal, and patagial; and those of the ventral ridges hypo- costal and median nervures. The costal nervure forms the anterior thick margin of the wing. The wing membrane at its attachment to the thoracic wall dips downwards and backwards between the costal and hypocostal, or second nervure ; it then ascends nearly perpendicularly to the sub- costal. It descends backwards to the median, and reascends to the patagial nervure. The three superior nervures terminate on the dorsal aspect of the thorax in three tubercles, which articulate with the edge of the dorsal plate. The anterior tubercle supports the marginal nervure. The median tubercle is a complex sclerite, very narrow above where it bears the subcostal nervure, but broad below where it supports the hypocostal ; the posterior tubercle terminates in the median and patagial nervures. The wing joint, therefore, consists of three parts ; each capable of independent movement, and acted upon by special muscles. I shall term them respectively the pro-, meso-, and metapterygium. These are present in the wings of every insect I have examined. The pro- and mesopterygium are generally more largely developed in the anterior wing, whilst the metapterygium is more developed in the posterior wing. Movements of the Wings. — The wings move in the horizontal and vertical planes ; the forward movement is extension, the backward flexion. The upward and downward movements are elevation and depression. The wings are further capable of rotation on their own long axis ; this action is similar to the feathering of an oar, and is of the utmost importance in the mechanism of flight. I shall term it rotation. The Wing Joint admits, therefore, of flexion, extension, elevation, (Icpix'ssion, and rotation, although the e.xtent II 1 64 THE INTEGUMENTAL SKELETON OF THE IMAGO. of each varies greatly in different insects. In the generalised Neuroptera and even in jEschna, the only movements which are at all extensive are those of elevation and depression ; rotation is limited, and flexion and extension are still more limited — hence the wing joint is comparatively simple. Rotation is mainly effected by the simultaneous elevation of the pro- and depression of the metapterygium, or vice versa. In the Hymenoptera and Diptera rotation and flexion are provided for by the segmentation of the pterygia, which consist of complex sclerites uniting the rigid nervures with the thoracic wall. These sclerites form a wing root which has some resemblance to the carpus of a vertebrate. The sclerites articulate with each other by complex surfaces, not only in series, formed by the segmentation of the roots of the main nervures ; but laterally, those of the propterygium articulating with those of the mesopterygium, and those of the meso- pterygium with those of the metapterygium. In the more highly modified Insecta there is a constant tendency towards the reduction of the wings to a single pair, either by a locking mechanism which unites the posterior border of the anterior with the anterior border of the posterior wing, or by their function as an organ of flight being in abey- ance in one or other pair. A few insects in all orders are apterous. In the Diptera the anterior wings only are organs of flight ; the posterior pair are greatly reduced in magnitude, and form complex sensory organs, which are known as halteres, or balancers. Owing to the almost equal development of both the anterior and posterior wing roots and systems of nervures in the Diptera, the wing often appears to consist of both an anterior and a posterior wing united, its anterior half closely resembling the anterior, and its posterior half the posterior, wing of the Hymenoptera. There can be, however, no question as to their being the homologues of the anterior wings only, nor as to the homology of the posterior wings with the halteres. The only authority who ever questioned thir> THE THORACIC EXO-SKELETON. 165 homology was Latreille, who regarded the halteres as abdo- minal appendages. I shall show hereafter that Latreille was clearly misled by the attempt to prove that the first abdominal segment enters into the composition of the thorax in the Diptera. The posterior margin of the wing is usually prolonged, in all the Diptera with a heavy abdomen, in the form of two semicircular scales, the anterior of which is the squamula, or lesser wing scale, and the posterior the squama, or great wing scale. Those Diptera which possess wing scales are termed calypterate. Modifications of the Thorax. — Graber says, in speaking of the thorax in insects generally: 'It is especially an organ of locomotion ; in the more terrestrial forms the three thoracic segments are mobile on each other ; in aerial insects, on the other hand, they are consolidated and form a rigid case, in which the segment bearing the largest pair of wings is most developed. The sternal region is chiefly developed in relation to the legs, and when the three pairs are equal in size, the three sterna are also more or less equally developed, whilst the pleura and terga are largest in the wing-bearing segments. Whenever the wings are reduced, there is a corresponding reduction of the dorsal arch ' [10, vol. i, p. 85]. In illustration of the above statements, I would observe that in the Coleoptera, in which the mesothoracic wings are reduced to wing covers, or elytra, which serve as mere protective sheaths, the dorsal arch of the mesothorax is much reduced. In the Diptera, on the other hand, it is the metathorax which has a rudimentary dorsum, and the whole segment is greatly reduced. The coleopterous metathorax resembles the dipterous mesothorax, and may be compared with it, almost every sclerite of the one occurring in the other. In all aerial insects the prothorax is also greatly reduced in size ; thus, in the Lepidoptera and Hymenoptera the dorsal arch forms a mere collar (collare). It is also separated from the sternal portion by a syndesmosis, a fact which led Kirby and Spence to deny the prothoracic origin of the collare. i66 THE INTEGUMEXTAL SKELETON OF THE IMAGO. In the Diptera the dorsal arch of the prothorax forms a narrow rim, which is not visible without dissection. It would, perhaps, be hazardous to give a more detailed description of the thorax applicable to insects generally, and any attempt to describe the various modifications of the exo-skeleton would be foreign to my purpose. I shall, however, in the special description of the thorax of the Blow- fly, give such comparative details as are necessary to the correct interpretation of the homologies of its several parts. b. General Description of the Thoracic Skeleton of the Blow-fly. The nomenclature I employ is a modification of that pro- posed by Osten-Sacken [81] for the sutures, and, as far as possible, that of Audouin [38 and 39] for the sclerites of the meso- and metathorax. The Thoracic Skeleton forms a subovoid capsule, with a cervical opening in front, an abdominal opening behind, and three pairs of ventral foramina with which the basal joints of the legs articulate. There are two large spiracles on each side, the anterior and posterior. It presents for examination a dorsal, a lateral, a ventral, an anterior, and a posterior aspect. The Dorsal Aspect (Fig. 35) exhibits two transverse sulci, prescutal and postscutal. The anterior or prescutal sulcus is faint and separates the prescutum (j*) from the scutum (-=■) ; the posterior or postscutal sulcus is deep and well marked, and divides the scutum from the scutellum (/). The Prescutum is very convex from before backwards ; its anterior part, which cannot be seen from above, descends vertically to the border of the cervical opening. The Scutum (.') exhibits a lateral projection, the scutal spine {Mihi) {4), behind the anterior root of the wing (5). In front of the spine is a depression in which the anterior wing root lies, the pre-alar fossa ; behind the spine there is a smaller fossa, the post-alar fossa; and externally to this a deep triangular fissure, the supra-tympanic fissure. * The numbers and letters refer to all the figures. THE THORACIC EXO-SKELETON. 167 The Scutellum (/) is a pouch-like projection overhanging the posterior part of the thorax and the base of the abdomen (Plates VII. and VIII.)- It is connected with the scutum by a pair of divergent ridges, which arise from its outer anterior angles; these are the scutellar bridges of Loew [77]. De- scending from the root of the scutellar bridge, a process of the scutellum forms the upper edge of the posterior alar- apophysis (Plate VII., b /). It bounds the outer edge of Fig. 35.— Dorsal view of llic Thorax of the Hlow-fly : /, Scutellum ; 2, scutum ; 3, prcscutum ; 4, scutal sjiine ; J, proplerygium ; 6, costal, 7, subcostal and hypocostal nervures forming the remigiuni ; S, patagium ; 9, lobulus of the wing ; 10, scjuamula ; //, squama ; 13, scutellar bridge ; 13, posterior alar apophysis. the supra-tympanic fissure, and the great wing scale (//) is attached to it. The Lateral Aspect is represented in Plate VII., Fig. i. Near its anterior margin the large anterior spiracle (/p) will be readily recognised by its orange-yellow colour. Start- ing from the upper margin of the spiracle, and extending back to the wing root h, there is a deep suture, the dorso- pleural suture. It is the anterior part of an extensive syndesmosis {e, f, g, h), the alar syndesmosis. Descend- ing vertically from the dorso-pleural suture between the plates numbered i8 and 28 is the mesopleural syndesmosis. «68 THE JNTKGUMENTAL SKELETON OF THE IMAGO. Bounding the mesosternum (jo) above and behind is a curved suture, the sterno-pleural ; and extending upwards from the sternopleural suture to the posterior spiracle is the hypo- pleural suture. The following plates will be readily distinguished : A large square plate {i8) continuous with the mesosternum in front of the sterno-pleural suture. This I term the lateral plate ; its nature has been much discussed. I believe it is a portion of the great mesosternum, which in the Bees (Bombus) and in many Orthoptera always extends upwards to the dorsum. Brauer incorrectly regards it as the episternum. It is certainly not the episternum of Audouin. In some Lepidoptera and other insects the sterno-pleural suture reaches the spiracle, when the lateral plate becomes a separate sclerite. The plastron of the mesosternum ( jo) is the large plate below the sterno-pleural suture. Behind it are the two sclerites of the coxa of the intermediate leg {p, q) and the lateral part of the metasternum (^j). Above the latter, the posterior spiracle is situated ; and between the hypopleural, sterno- Description of Plate VII. Fin. l.TheKxterior, and Fk;. 2, The Interior of the Thoracic Skeleton : /, scutellum ; 3, scutum ; j, prescutum ; 4, posterior dorso-pleuval diarthrosis ; j, great alar apojjhysis ; 6, scutal pouch ; 7, ridge on alar apophysis ; S, small tympanic plate ; 9, posterior parascutal plate ; /<>, apodeme of parascuta ; //, anterior parascutum ; / », uncinate process ; /j, head of alar apophysis ; 14 and ij, anterior alar fossa ; /6, great ampulla ; //, parapteron ; jS, lateral plate ; /9, anterior spiracle ; -'o, paratreme ; ^/, tympanic ridge ; 22, lateral plate of post scutellum ; 2j, mirror ; 24, tympanic plate ; sjj, costa ; 26, epimeron ; 27, ridge between epimeron and episternum ; ^S ; episternum ; -^ly, great entopleuron ; jo, mesoplastron ; j/, hypotreme ; J2, epitrochlea ; jj, prodorsum ; j^, neck ; jj, process of epitrochlea ; j6, root of the haltere ; ?7, ridge above the posterior spiracle ; jS, scaphoid process of the mesophragma ; ji^, the mesophragma ; 40, spiracle ; ^/, epimeron, and 43, episternum of the metathorax ; ^j, mctaplastron ; 44, entosternum of metathorax ; ^j, furca of mesosternum ; ^6, posterior coxa ; ^7, intermediate coxa; ,/S, 41), and 50, articular heads on the meso- and meta- sterna for articulation with the coxa; ; 5/, vertical plate of metasternum ; 52, mesoplastron ; _jj, vertical plate of mesosternum ; a, part of the shield of the postscutcUum ; /', c, posterior alar apophysis ; d /, supra-tympanic fissure ; g, great alar apophysis ; //, post-alar fossa ; /•, hypopterygium ; /, saccuius ; m, great tympanic membrane ; //, tympanic, bulla ; 0, halter ; /, posterior, and , hypopterygium ; ha, hamulus ; epc, epicosta ; f, lower extremity of the fissure in the ascending process. Fig. 8.— The Internal .Surface of the Pleuron : n, extremity of the tympanic ridge ; sac, sacculus ; /', spines forming ridges on the inner surface of the thoracic wall. Fii;. 9.— The parapteron and its relations, seen from the Interior of the Thorax: a, anterior process ; .«/, part of the mesopleural syndesmosis ; b, long process ; pe, pre-epaulet ; lip, hypopterygium ; ha, hamulus. The wing and sacculus have been removed. Fig. 10.- External .Surface of the Metapleuron. Fig. II.— The Internal Surface of the same: ap, apodeme of the halter ; «', part of the first abdominal segment ; r, ridge between episternum and epimeron ; .v, rail on which the hook of the posterior coxa slides. The other references to the figures in this plate are the same as in Plate Vll. PLATE VIII. DETAILS OF Tllli THORACIC SKELETON. I THE THORACIC EXO-SKELETON. 173 spiracle and the paratreme ; below and in front it is continuous with the plastron, for a short distance ; its other relations are sufficiently indicated in the figures. Near the anterior extreinitj' of its superior border there is a small tubercle, which articulates with a corresponding depression in the paratreme ; this articu- lation is the anterior dorso-pleural diarthrosis, it forms the anterior extremity of the dorso-pleural syndesmosis. Many different views have been held as to the homology of the lateral plate — several of these have been already alluded to. Hammond attempts to identify it with the parapteron of Audouin, but a more careful study of Audouin's writings would have convinced him of his error. In many insects the corresponding plate is, apparently at least, quite removed from the plastron of the mesosternum by the extension of the pleuron forwards ; I believe, however, a narrow neck will be found on careful examination uniting this plate with the plastron. The Manubrium (I'late VIII., Figs, i and 2, ,>-) probably represents the body of the prosternum. It is a narrow quadrilateral plate, somewhat wider in front than behind ; it articulates with the plastron by the manu- brial suture. Its inner surface exhibits three sub-parallel ridges — one in the median line, and one on each of its edges ; these terminate behind in the manubrial suture. The lateral and anterior edges of the manubrium are continuous with the fle.vible syndesmotic integument of the prosternal area. There is a small pouch-like projection on the external surface of the manubrium, close to the manubrial suture, and a corresponding hollow on its internal surface. The manubrial suture forms an inflection on the inner surface of the ventral thoracic wall, from the extremities of which a pair of strong fusiform processes extend to the lower borders of the anterior spiracles. These pro- cesses may be termed the 'hypotremata.' The Hypotreme (Plates VII., Fig. 2, and VIII. Figs. 2 and 4, j/) extends obliquely across the anterior coxo-sternal foramen, and terminates in three ridges— two ascend and surround the spiracle, and the third curves down- wards and forms the external edge of the epitrochlea. Tiie hypotremata give great strength to the anterior part of the thoracic wall, acting as a kind of stretcher, keeping the external wall of the thorax from being drawn towards the middle line by the action of the muscles which move the head and the anterior legs. As they extend from the suture between the pro- and mesosternum, they may be regarded as the internal continuation of this suture. The spiracle is situated between the pro- and mesothorax, and the suture which forms a ridge in front and behind it is also continuous with the hypotreme. 174 THE INTEOUMENTAL SKELETON OF THE IMAGO. The epitrochlear sclerite lies in front of the descending process of the hypotreme, and is prothoracic, whilst the lateral plate and the plastron lie behind it, and are both mesothoracic. The hypotreme is a cylindrical rod, between the nianubrial suture and the lower margin of the spiracle, and has probably become detached from the posterior margin of the coxo- sternal foramen by the atrophy of the inflected hypodermis, from which it is developed. 1 therefore regard the hypotremata the manubrial suture and the spiracles as representing the primary suture between the pro- and meso- thorax. The Metasteraum is the largest sclerite of the metathorax. It is shaped somewhat like the sphenoid bone of the human skull. It consists of a median longitudinal inflection of the integu- ment supporting a metafurca, of a narrow transverse inflection, and of two lateral plates— the metapleura of Osten-Sackcn. These must not be confounded with Audouin's metapleura ; I term them ' metaplastra ' (Plate VIII., Fig. 4, 4.3). The metaplastra are the only parts of the metasternum which are easily seen externally, but when the intermediate and posterior coxae are carefully separated, a narrow ridge can be distinguished between them : this is the lower edge of the transverse inflected plate ; it unites the metaplastra with each other, and is continuous with their anterior and posterior margins. Each metaplastron is separated from the mesostcrnum in front by the transverse ventral suture and the intermediate coxa ; behind it is in relation with the metapleuron and the first abdominal ring, and externally it surrounds the posterior spiracle. Seen from its internal surface, the metasternum is cruci- form. The transverse vertical plate is the inflection between the intermediate and posterior coxae ; it is joined behind, in the middle line, by a median plate, which is triangular ; the upper edge of the latter supports the metafurca, its lower edge articu- lates with the posterior coxae. At its junction with the transverse vertical plate it supports a capitellum {4^) with which the four coxae articulate. Its posterior angle bears two divergent capitate processes {4S), which articulate with the posterior coxae. The Metafurca (//) is a long hollow trough from which the THE THORACIC KXO-SKELETON. 175 median thoraco-abdominal muscles arise. The posterior ex- tremity of the metafurca is pointed, and dips down between the posterior coxa; ; the anterior extremity rests upon the mesofurca. The Dorsal Valve. — If an incision be made through the anterior part of the scutum horizontally backwards until it meets the anterior extremities of the dorso-pleural syndesmosis, the syndesmotic membrane is easily divided, after cutting off the wings close to their roots, as far back as the anterior external angles of the scutellum. Here the dorsum articulates with the post-scutellum by a diarthrosis, the posterior dorso- pleural joint. With a little pains the articulation can be broken through and the remainder of the syndesmotic membrane be- tween the lower surface of the scutellum and the post-scutellum divided. In this way the thorax can be separated into a ventro- pleural and a dorsal valve by the natural articulation, which permits these parts to move upon each other by the action of the powerful sterno-dorsal and longitudinal thoracic muscles. The dissection should be made in an artificial exuvium (see Appendix to chapter). The Under Surface of the Dorsal Valve. — The dorsal valve con- sists of part of the prcscutum, of the scutum, ) is an inflection of the posterior angle of the scutum, and is separated from the dorso- pleural costa by a deep fissure. It terminates in front in an articular head, which supports the anterior wing-root. Its inner edge is nearly parallel with the inner edge of the dorso-pleural costa. The scutellum, seen from its under surface, has the form of a conical pouch, the anterior margin of its inferior wall being deeply crescentic, with a thickened rim, the scutellar rim. The anterior extremities of the rim each give off two divergent processes, at the common root of which there is a circular 1/6 Tin-: INTEGUMENTAL SKKI-ETON OF THE IMAGO. cup, this cup articulates with a hemispherical process of the post-scutellum, formin=: the strong posterior dorso-pleural diar- throsis {J^). The inferior process of the rim of the scutelluin (h d) supports the posterior wing-root ; this is the alar process of the scuteilum. The anterior and superior process (7) extends forwards on the inner edge of the great alar apophysis and forms its articular head (/?). The crescentic rim of the scuteilum is united with the post- scutellum by soft syndesmotic integument. The Supra-tympanic Fissure (/). — The external surface of the great alar apophysis looks downwards, and forms the anterior part of a triangular inflection of the integument above the great wing-scale. This inflection opens and closes and permits of considerable variation in the capacity of the thoracic cavity ; when closed its superior or anterior half rests upon its posterior or inferior half, in which a complex tympanic mechanism lies, and a minute spiracle opens. These will be described in detail in another chapter. The whole arrangement of the supra-tympanic fissure is similar to the infolded gusset of a pair of bellows— the apex of the fold is just in front of the posterior dorso-pleural diarthrosis. The dorso-pleural costa is the inner edge of the floor of the alar fossse. It is prolonged forwards as a sutural ridge, which bifurcates and surrounds the paratreme. The presutural ridge is the inner edge of the presutural sulcus. It joins the two dorso-pleural costse and supports a cup- shaped cavity in each, which articulates with the head of the parapteron {17). A pair of small plates, which form the inner wall of the posterior alar fossa, articulate with the scutal portion of the dorso-pleural costa. These I term the 'anterior and posterior parascutum.' The Anterior Parascutum (//) is a distinct oblong plate. Its anterior external angle is prolonged in a curved hook {12) — the uncinate process — which embraces the anterior head of the dens, a small sclerite on which the propterygium turns when the wing is elevated or depressed. The Posterior Parascutum (p) is inseparably fused with the THE THORACIC EXO-SKELETON. 177 edge of the scutum. A small apodeme projects between the parascuta and unites them. It gives insertion to a slender muscle with a long tendon, the action of which is to depress the alar edge of the parascuta and assist in the elevation of the wing. In removing the dorsum from the thorax, it will have been observed that the prescutum was divided by a horizontal incision just above the anterior spiracle, so that only its posterior part enters into the formation of the dorsum. The prescutum is very convex in front, so that its anterior part forms the anterior wall of the thoracic cavity (Plate VIII., Figs, i and 2). The inner surface of this portion of the thorax is best examined by cutting off the anterior part of a skeleton, or artificial exuvium, just behind the anterior spiracle. The prescutum will be seen in such a preparation to be bounded in front by the prophragma and the two paratremes. The Prophragma {pp) is a membranous inflection of the edge of the prescutum. The attached margin of the prophragma corre- sponds with the junction of the anterior edge of the meso- thorax and the posterior margin of the rudimentary prothorax. Its extremities are continuous with the superior and inferior edges of the paratremes, so that the lateral parts of the pro- phragma split into two lamina;, like the e.xtremities of the hypotremes. The central portion of the prophragma has a free convex posterior border, and exhibits a median projection, strengthened by a distinct bifurcate sclerite. Its upper and lower surfaces are horizontal. The Prodorsal Arch {pd) is the inflected edge of the cervical opening of the thorax. It is divided in the median line above into two lateral halves by a distinct suture. Its outer extremities articulate with the anterior margins of the epitrochlcar sclerites, and with the inner and lower part of the paratremes. Each lateral half is divided into two by a distinct oblique suture, which extends from the attached edge of the pro- phragma downwards, outwards and backwards, and termi- nates in a strong apodeme, which projects into the thoracic cavity, the prothoracic apodeme. 178 THE INTEGUMENTAL SKELETON OF THE IMAGO. The Paratreme {20) is seen best from the interior of the thorax. It is a distinct and irregularly quadrate sclerite, bounded above and internally by the prescutum and the pro- phragma, below by the extremity of the prodorsal arch, and externally by the spiracle, above which it just reaches the lateral plate and the dorso-pleural syndesmosis. Although the limits of the paratreme are not very readily seen externally, they are marked by a very distinct sutural ridge projecting into the thoracic cavity. The Epitrochlear Sclerite is a quadrangular vertical plate, articulating with the paratreme and prodorsum above, and with the condyle by its inner margin. Its external surface is convex, and projects in a tooth-like process, overhanging the anterior coxa in front. The external edge of the epitrochlea is formed by the descending process of the hypotreme. The Sclerites of the Prosternal Area are nine in number : the manubri\nn, two pairs of clavicukc, anterior and posterior, and two pairs of pectoral sclerites, the greater and lesser pectorals. In the immature imago these sclerites are all distinct chitinous thickenings of the syndesmotic integument ; but in the mature insect the clavicuhc are fused with the manubrium and the epitrochlear sclerites. The Claviculae. — The anterior clavicula is a thin rod of chitin, extending from the anterior edge of the manubrium to the epitrochlea. The posterior clavicula is somewhat shorter and broader, tapering at its extremities. The Great Pectoral is a triangular sclerite, which flanks the manubrium, with its short edge adjacent to the clavicula. The Lesser Pectoral sclerite is a minute convex caudate fold between the great pectoral and the anterior coxa. It is broadest behind, and is covered with minute bristles. I'he clavicuhu unite with each other and with the pectorals in the adult insect, and so complete the foramina with which the anterior cox;v articulate. The integument of the prosternal area is continuous with that of the neck. The Neck is very narrow, sub-cylindrical, and covered by syndesmotic in- tegument, in which seven sclerites are apparent. These are a median ventral sclerite — the sella, to which allusion has already been made; a pair of acces- sory pieces in front of the sella— the cornicula; ; and two pairs of lateral sclerites : a large posterior pair — the condyles ; and a small anterior pair— the epicondyles. There is no dorsal sclerite in the neck of the fly, although two dorsal sclerites exist in the neck of the cockroach [41]. The Sella (Plate VIII., Fig. 3) is a saddle -shaped sclerite, with two remarkable transparent lobes in front, which form the lateral and upper walls THE THORACIC EXO-SKELETON. ,79 of a small median pit in the cervical integument. These lobes are covered with fine sensory bristles, and receive a large number of nerve filaments, which end in special sensory cells connected with the bristles. This sensory organ resembles similar, but smaller, organs situated on the flexor sides of the limb articulations, and, like these, it is probably concerned in indicatmg the movements of the parts adjacent to it, by giving rise to sensory impulses originating from movements of the head and fore-limbs, by which the walls of the pit in which it lies are alternately stretched and relaxed. The Comiculae are two small curved sclerites in front of the sensory lobes of the sella, which form the anterior wall of the pit. The Condyles are subconical sclerosed pouches on the under and outer part of the cervical integument. They exhibit three surfaces : a posterior surface— which rests against the anterior surface of the thorax— an inferior and an external surface. The inferior surface exhibits a curved ridge fringed with hairs, with its convexity inwards. This ridge resembles a rudimentary appendage ; indeed, the whole condyle, by a little modification and greater development, might become a chela similar to the post-oral chela; of Arachnids. The condyle is present in every insect I have examined. The Epicondyle (Fig. 24, >-•) is a rod - shaped sclerite, which easily separates from the condyle in the young imago, but which appears to be inseparably united to it in the adult insect. The epicondyle articulates with the cotyloid cavity of the occipital ring. A small sclerite, the first jugular sclerite of Kiinckel d'Heiculais lies in front of the epicondyle (Fig. 24, y'). Morphology of the Prothoracic Region.— The discovery of the prophragma has not hitherto been recorded. It undoubtedly marks the posterior limit of the dorsal portion of the pro- thorax, and confirms the views of Brauer [80] and Hammond [78] as to these limits. I regard the following sclerites as undoubtedly prothoracic — the prodorsum, paratremes, and epitrochleas, the sella, manubrium, claviculas and pectorals. The cervical sclerites also probably belong to the prothorax. I have spent much fruitless labour in the endeavour to find some correspond- ence between these parts and those of the meso- and meta- thorax. My epitrochlea is certainly the trochantin of Audouin and the rotula of Straus Durckheim. Both the episternum and the epimeron, if they existed, should be behind the epitrochlea, but they are not recognisable. The paratreme appears to me to correspond with the operculum of the Hymenoptera in its double relation with the prodorsum and the spiracle. Ham- mond assigned it to the prothorax. I cannot accept Brauer's 12 i8o THE INTEGUMENTAL SKELETON OF THE IMAGO. view that the diaphragma represents the inflected post- scutelUim, either in the prothorax or the mesothorax, but think it is indubitable that the prodorsal arch and the paratremes represent the collar of the Hymenoptera. With regard to the morphological significance of the cervical sclerites, amongst which I have included the sella, they must either be regarded as a portion of the complex prothorax, or as the remains of one or more segments which are no longer distinct in either the embryo or the nymph. I prefer to regard them as prothoracic. Complex prothoracic sterna are fre- quently seen even in the Orthoptera, for example, in the greatly elongated prothorax of Mantis ; and, as I have already observed, the structure of the meso- and metathorax is no guide to that of the prothorax. The prothorax appears to mc to exhibit strong indications in favour of the view held by Patten and others that the thoracic segments have resulted from the fusion of two or more primitive metameres. But any attempt to determine their limits, or even their number, in the absence of direct developmental evidence, could be nothing but guesswork, although it is evident that many metameres have disappeared in the process of evolu- tion. The Nomenclature of the Spiracles. — Whether the anterior spiracle should be regarded as prothoracic or mcsothoracic, or the posterior as metathoracic or abdominal, has given rise to much discussion, which I regard as futile, although the views in which it originated are of some interest. Oken's idea of the origin of wings from respiratory organs was adopted by Latreille and De Blainville, and Blanchard ap- pears to have considered them as everted tracheal sacs, the thoracic attachments of which are homologous with the spiracles, and this led him to assert that ' there is never any spiracle on either the meso- or metathorax of a winged insect,' an opinion which has been frequently repeated. Palmen showed that the tracheal gills of Ephemera arise independently of the spiracles, and are in no way mere modifications of them, and Blanchard's statements rest solely on the fact that THE THORACIC EXO-SKELETON. i8i the larvae of the Lepidoptera have no spiracles on the segments corresponding with the meso- or metathorax. Weismann [2] assigns the anterior spiracle to the prothorax on developmental grounds. Hammond [78] adopts the same view. Brauer [80], on the other hand, says it is meso- thoracic. Gosch [79] pointed out the verbal character of the discus- sion, and his statements may be summarised as follows : It is the rule to describe inter-segmental abdominal spiracles as belonging to the segment immediately behind them, and this rule is apparently justified by the fact that when the abdominal spiracles are segmental they are always near the anterior border of the sclerite on which they occur. In the case of inter-segmental thoracic spiracles the rule has been reversed, and they have been ascribed to the segment in front of them. Hence the anterior spiracle has been named pro- thoracic. It has been asserted that the anterior spiracle in the Coleoptera always remains attached to the prothorax when the latter is separated from the mesothorax. In my ex- perience it as often remains on the mesothorax. I do not, however, think that such evidence is admissible. Weismann's statements are far more important ; the anterior spiracle of the nymph is certainly prothoracic, but it is by no means certain that the anterior spiracle of the imago, which is developed behind that of the nymph, is also prothoracic, as Weismann believed. It will probably be argued by some that the spiracles are segmental appendages. I regard them as fissures developed in relation with the internal tracheae, and hold that such fissures may be either segmental or inter-segmental. I have already shown that the anterior spiracle in the Blow-fly lies between the pro- and mesothorax. The posterior spiracle is between the metasternum and the tympanic bulla, and the bulla is undoubtedly mesothoracic. It is therefore situated between the meso- and metathorax. The Pleural Kegion is bounded above by the wing roots, in 12 — 2 1 82 THE INTEGUMENTAL SKELETON OF THE IMAGO. front by the posterior edge of the lateral plate ; below by the sternopleural suture; and behind by the hypopleural suture, the tympanic bulla and the attachment of the great winglet or squama. It is a complex region, which is far larger in the Syrphidse and Volucella than in the heavy-flying Muscidae. It contains the following sclerites, which form part of the thoracic wall : The episternum (Plate VII., 28), the epimeron {26), the costa {23), the epicosta (Plate VIII., Fig. 7, cpc), and the parapteron (Plate VII., / 7). The Pleuron (Plate VIII., Fig. 8) is sub-triangular, divided into two by a vertical suture, which projects as a distinct ridge on its inner surface. This ridge is united below with the great entopleuron. The part in front of the suture is the episternum ; that behind it is the epimeron below, and the costa above. The Episternum is an irregular quadrilateral plate, terminating above in two curved processes. These processes are sub parallel ; the anterior sup- ports the pre-epaulet of the wing. The posterior process has a hemispherical swelling behind, and below it, the great ampulla (Mihi). The process curves over the great ampulla, and ends in a strong hook, the hamula, which articu- lates with one of the sclerites of the propterygium. The inner surface of the great ampulla gives origin to a powerful muscle, which acts on the wing root — the ampullar muscle (Mihi). The episternum is bounded in front by the mesopleural syndesmosis, below by the sternopleural suture, behind by the epimeron, and above by the great ampulla. The Epimeron is much thinner than the episternum ; it is irregularly cor- date, with the emargination above. Seen from the interior of the thorax, it exhibits a series of radiating ridges, which commence at its margin and form the sutures between the adjacent plates of the thoracic wall ; the strongest extends to the external angle of the post-scutellum. The epimeron articulates— below with the meso- and metasternum, from the latter of which it is separated by the hypopleural suture ; behind with the tympanic bulla, and above with the costa. The Costa is a thin shell-like plate. The outer surface is convex, and looks downwards and outwards ; it is covered with fine bristles. Its upper edge is continuous with the syndesmosis of the wing, and adjacent to a small sickle-shaped sclerite— the epicosta. The Epicosta presents a strong, chitinized, hollow, curved protuberance in front, which forms a kind of handle to the sickle. I term this projection the 'lesser ampulla'; it gives origin to the lesser ampullar muscle. Kunckel d'Herculais [25, p. 98] incorrectly names the epicosta, the ' parapteron,' a term already applied by Audouin to a very different part. In Volucella and the Syrphidx- the epicosta is prolonged behind as a long cylindrical flexible process, covered by fine sctx. This tail-like process was first described by Chabrier [73, torn, viii., p 398], who fancifully com- THE THORACIC EXO-SKELETON. 183 pared it with the long wing feathers of the birds of paradise. Its use is unknown, but it is probably a sensory appendage. The Paxapteron (Audouin), seen from the exterior of the thorax (Plate VII., 77], is a strong heart-shaped nodule, which lies in the mesopleural syn- desmosis, just below the dorsoplcural suture. It is the head of a powerful lever-like apodeme, which gives insertion to the anterior wing muscles. The Post-scutellar Region, or Dolium (Plate VIII., Fig. 5), forms the greater part of the posterior aspect of the thoracic wall. It is bounded above by a syndesmosis, which unites it with the scutellum, and by the posterior alar apophysis, and the insertion of the wing scales ; and below by the upper margins of the posterior spiracles and of the thoraco- abdominal opening and mesophragma. The sclerites which form this region are those of the post- scutellum and the tympanic bullae. They are united by symphyses, which project internally as strong ridges, and form a kind of sub-hemispherical cap, the dolium, which is supported by the lateral plates of the metasternum, with which it is firmly connected by strong internal ridges. It is also articulated with the scutellum by the posterior thoracic diarthroses and by syndesmotic integument. The Post-scutellum (Plates VII. and VIII., 22) consists of a median shield and two large lateral plates. The median shield is a very massive sclerite, sub-quadrate in outline, with a concavo-convex centre, convex on its outer surface. Its superior angles are prolonged to form the con- vex heads of the posterior thoracic diarthroses, and articulate with corresponding concave articular surfaces at the roots of the alar apophyses of the scutellum. Its inferior angles pro- ject backwards and upwards into the abdominal cavity and support the first abdominal ring. The upper edge of this plate is united with the scutellum by syndesmosis, and is strengthened by a thick ridge. Its lateral edges articulate with the lateral plates of the post-scutellum. Its inferior edge has a rim on its posterior surface, the post- dorsal arch, and its edge articulates with the mesophragma. The OS cornutum of Jurine appears to be the upper edge of the post-scutellum. i84 THE INTEGUMENTAL SKELETON OE THE IMAGO. The lateral plate of the post-scutellum (Plate VIII., Fig. 5, /') is sub-triangular, with the apex of the triangle in relation with the inferior angle of the median shield of the post-scutellum, to which its posterior edge is firmly united ; its anterior edge articulates by symphysis with the tympanic bulla ; and its superior edge, united with the posterior alar apophysis of the scutellum, forms the inferior margin of the supra-tympanic fissure. This edge is much thickened by a strong internal inflection, which projects into the thoracic cavity ; it is con- tinued forwards and downwards into the costal margin of the epimeron (Plate VII., Fig. 2, 22). The posterior thoracic diarthrosis is greatly strengthened by the inflected superior and posterior edges of this plate. The Tympanic Bulla (h), seen from the side, appears sub- hemispherical, but from behind sub-triangular (Plate VIII., Fig. 5, tp)i It is the segment of a cone with a convex base. The apex of this cone is in relation with the insertion of the halter. Its upper edge forms a symphysis with the post- scutellum, and its lower edge the upper margin of the posterior spiracle. The lower edge also articulates with the two pro- cesses of the metasternum, which form the anterior and posterior margins of the spiracle. The convex base of the tympanic bulla is deeply notched above and in front, so that a semicircular opening is left between it and a rod of chitin, which extends from the meso- pleuron to the upper margin of the tympanic bulla. This is a portion of a strong ridge, the tympanic ridge, between the posterior thoracic diarthrosis and the costal margin of the epimeron (Plate VII., Fig. 2, 21). The notch, or foramen, in the bulla is closed by a tense mem- branous integument {ni), the membrana tympani major (Mihi), which looks upwards and forwards beneath the great wing scale. The tympanic bulla, seen from above and in front, resembles a kettledrum, the membrana tympani major being the drum skin. The lowest point of the posterior wing root rests upon the drum membrane. The inner edge of the membrane is attached to the alar apophysis of the scutellum. THE THORACIC EXO-SKELETON. 185 The Mesophragma ( ,"•(;) is a partial concavo-convex septum, which projects downwards and forwards into the thorax from the lower edge of the shield of the post-scutellum. It is convex behind, and attached above to the post-scutal shield. Its free inferior edge is deeply notched in the middle line, and separated from the lateral walls of the thorax by a narrow fissure on each side, which transmits several small muscles. The mesophragma has a small, almost square, process, directed forwards and outwards near the upper edge of the posterior spiracle; it arises at the junction of the free and attached margin of the mesophragma. Internal Ridges.— The internal surface of the postero-lateral region of the thorax may be readily examined by making a section in the vertical plane, through the junction of the dorsocentral and lateral parts of the dorsum. This region exhibits two sets of ridges, or inflected sutures— one set radiating from the posterior thoracic diarthrosis.and the other from the inferior angles of the post-scutellar shield. One ridge is common to the two sets ; it is the suture between the shield and the lateral plate of the post-scutellum. Two other ridges radiate from the posterior thoracic diarthrosis. The upper one is the great alar apophysis, the lower the 'tympanic ridge.' The tympanic ridge crosses the tympanic notch, and forms the straight margin of a transparent semicircular plate— the mirror. The convex margin of the mirror is attached to the thoracic wall, dividing the tympanic bulla into two parts (33). Its plane is nearly at right angles to the membrana tympani major, from which it is separated by a narrow space above its free margin. It resembles the mirror of the Cicadiv;. The tympanic ridge extends as far forward as the costal edge of th« epimeron. The ridges from the lower centre are the common one already mentioned, that between the post-scutellum and the tympanic bulla, and the upper margin of the posterior spiracle. Tlie one between the tympanic bulla and the post-scutellum ends in the tympanic ridge, which it joins at a right angle immediately behind the insertion of the great tympanic membrane. The Morphology of the Post-scutellar Kegion.— The morphology of this region has given rise to very great differences of opinion. Brauer first correctly ascribed it to the mesothorax. Ham- mond follows Brauer, but incorrectly includes the metasternum in the mesothora.x. Adopting Brauer's view, the posterior spiracle lies between 1 86 THE INTEGUMENTAI. SKELETON OF THE IMAGO. the meso- and metathorax, and the mesophragma is situated behind the spiracle and immediately in front of the meta- pleuron. Owing to the great development of the dorsal region of the mesothorax, the dorsal arch of the metathorax is pushed downwards and backwards, and lies on the abdominal surface and lower edge of the post-scutellum, with the halteres between it and the metapleura. The post-scutellum of the metathorax of the Cockchafer, the tergum of Straus Durckheim, is very similar to the post- scutellum of the mesothorax of the fly ; and like the latter, is divided into a central shield, two lateral plates and two tympanic bullae, although the latter are feebly developed. The great mobility of the scutellum on the post-scutellar region, and the compact union of the latter with the meta- sternum, have led some to regard it as a portion of the meta- thorax; but the relations of the parts amongst themselves, and those of the mesophragma and the halteres show conclu- sively that the post-scutellar region is mesothoracic. The Metathorax consists of the metasternum, which has been already described, the metapleura and the post-dorsum. The Metapleuron {4r and 42) bears the same relation to the halter that the niesopleuron does to the winp, but, like the appendage which it sup- ports, is reduced to very small dimensions. It is in the form of a nearly vertical isosceles triangle, with its ape.\ upwards. Its internal surface (Plate VIII., Fig. 11) e.xhibits three well-marked ridges. The central ridge is nearly vertical ; its upper extremity bifurcates and forms a pair of rounded horns, which surround the lower half of the base of the halter. The anterior and posterior ridges are the sutures, in front between the metapleuron and the metasternum, and behind between the posterior edge of the metapleuron and the abdomen. The metapleuron is therefore divided, like the pleuron, into two parts — the episternum and the epimeron of the metathorax. The Episternum of the metathorax (^42) articulates in front with the metasternum, and below with the posterior co.xa ; above with the post- dorsum and halter, and behind with the epimeron. The Epimeron of the metathorax {41) articulates below with the posterior coxa ; behind it is united by syndesmosis with the dorsal arch of the first abdominal segment ; above it articulate*; by its cornu with the halter and post-dorsum ; in front it is united by symphysis with the episternum. The Post-dorsum is reduced to the form of a thickened fold, which unites THE THORACIC EXO-SKELETON. 187 the cornua of the nietapleura ; it has on the lower edge of the post-scutum, and is connected by syndesmosis with the dorsal arch of the first abdominal segment. The Thorax as a whole. — The numerous complex sclerites of the thoracic wall are united into three groups ; those of each possess little or no movement on each other, and are united by symphysis. The whole dorsum forms one portion of the thoracic wall, the sterna and pleura a second, and the post-dorsum, or post- scutellum, with the tympanic bullae, a third. The sterno-pleural portion of the thorax somewhat resembles an old Spanish ship-of-war. I term it the ' carina.' The post- dorsal portion is firmly attached to the carina and forms a kind of poop, the ' dolium.' The dorsum is attached to the carina by the elastic prescutum in front, and rests upon the four diarthroses. Between the anterior and posterior diar- throses the dorsum and the pleurae are united by syndesmoses, and a loose syndesmosis connects the scutellum with the dolium behind the posterior diarthroses. The tympanic fissures are oblique extensions of the dorso-pleural syndesmoses, and open and close like the gussets of a bellows. This permits of an increase or diminution of the convexity of the dorsum. The mesopleural syndesmosis also allows of some variation in the convexity of the carina. The great muscles which affect the magnitude of the thoracic cavity are the dorsales and sterno-dorsalcs (Plate XI.). The contraction of the latter diminishes the convexity of both the dorsum and the carina, closes the tympanic fissure, and increases the breadth of the thorax ; whilst that of the former increases the convexity of the dorsum and carina, closes the mesopleural suture, opens the tympanic fissure, and shortens the longitudinal, but increases the vertical and trans- verse diameters of the thorax. Increased convexity of the dorsum and carina renders the dorso-pleural syndesmoses tense, whilst diminished convexity relaxes them. The effect of these movements will be con- sidered hereafter. 1 88 THE INTEGUMENTAL SKELETON OE THE IMAGO. The elastic recoil of the thoracic wall acts in antagonism to both sets of muscles, which only counteract each other to a certain degree, both sets increasing the transverse diameter of the cavity. The Median Segment of Latreille. — Before concluding this somewhat lengthy section, I would say a few words on a subject which has been a fertile source of error in relation to the morphology of the sclerites of the thorax in the Diptera. Latreille discovered that the ventral portion of the first abdominal segment enters into the composition of the thorax in the Hymenoptera. This is especially well seen in those with a sessile abdomen, and anyone who will examine the thorax of a Cimbex or Trichosoma can easily verify the fact. Latreille calls this segment the ' segment mediaire.' The median segment of Latreille bears a spiracle, but there is a second spiracle in front of this in Cimbex, between the metasternum and the dorsum, corresponding with the posterior spiracle of the Diptera. This spiracle was unknown to Latreille ; it was discovered by Professor Schiodte in 1856 [76] . Latreille erroneously supposed the posterior thoracic spiracle of the fly to be the first abdominal one, and thought that the first abdominal segment enters into the composition of the dipterous thorax. He was, however, more consistent than his followers, for he denied the existence of any homology between the halteres and the second pair of wings. He clearly saw that his theory fell to the ground unless the halteres are regarded as abdominal organs. At the present day no homology has been more clearly established than that of the halteres and the posterior wings, yet many who smile at Latreille's idea, that they are abdominal appendages, have readily adopted his view of the existence of a median segment in the dipterous thorax. I have shown, I think clearly, that so far from this being the case, the meta- thoracic tergum is abdominal in these insects, the metapleura being on the abdominal side of the mesophragma. THE THORACIC EXO-SKELETON. 189 A very complete history of the opinions held on Latreille's ' segment mediaire ' is given by Gosch [79J . The remaining sclerites of the thoracic skeleton are those of the legs, wings, and tympanic apparatus, to each of which I shall devote a separate section. Nomenclature and synonymy of the thoracic skeleton : A. Prothorax, Audouin. Collare, Chabrier, etc. Manitruncus, Kirby and Spence. Corselet, Straus Durckheim. a. Presternum, Burmeister. Prosternal area, Mihi. Contains the following sclerites : Sella, Mihi. Cephalo-thorax, Mihi, oliin. Pectorals, Mihi. Claviculae, Mihi. Manubrium, Mihi. b. Neck-sclerites, jugulares, are : Condyles, Mihi. 3'""= jugulaire, Kiinckel d'Herculais. Epicondyles, Mihi. 2""= jugulaire, Kiinckel d'Herculais. c. Prodorsum, or prodorsal arch, Mihi. d. Epitrochlea, Mihi. e. Paratreme, Mihi. Scapula, Burmeister; oliin. Humerus. f. Internal parts : Prophragma, Mihi. Post-scutellum, Audouin. Hypotremata, Mihi. Belong also to mesothorax. B. Mesothorax, Audouin. Consists of : a. Sternum. Meso-sternuni, Audouin. Plastron, Chabrier. Entothorax, Audouin. Furca and median plate, Mihi. Lateral plates, Mihi. Great Entopleuron, Mihi. b. Dorsal region : Prescutum, Audouin. Scutum, ibid., which bears the Great Alar Apophysis, Mihi. Processus Styloideus, Chabrier. Scutellum and Post-scutellum, Audouin. Supra-tympanic fissure, Mihi. Mesopbragma, lirauer. Costa, Chabrier. Lateral Plate of Post-scutellum, Mihi. c. Lateral region : Pleuron, Audouin. Parapteron, Audouin. Clavicle, Kiinckel d'Herculais. Episternum and Epimeron, Audouin. Costa, .Straus Durckheim. Tympanic Bulla, Mihi. Bolium, Mihi. The tympanic bull;K and post-scutellum. I90 THE INTEGUMENTAF. SKELETON OF THE IMAGO. C. Metathorax, Audouin. a. Metasternum. Transverse and Vertical Plate, Mihi. Plastron, Chabrier. Metafurca, lirauer. b. Dorsal Arch, Mihi. c. Metapleuron, Audouin. Epistemum and Epimeron, Audouin. d. Details of the Exo-skeleton of the Legs (J'latc IX.). (Ear a gcficral description of the ventral appendages of the thorax \legs\ set p. 158.) The three pairs of legs differ chiefly in the form of the co.xse ; the remaining joints are very similar in all. The Coxae of the anterior legs are tubular and prismatic (Fig. I, ex) ; those of the intermediate pair, scaphoid, or boat- shaped (Figs. 2 and 3) ; and of the posterior, pyramidal (Fig. 4). Each coxa is consolidated and protected by three sclerites ; an anterior, a posterior, and an internal plate. In all the anterior plate is the largest and strongest. It exhibits an internal longitudinal ridge, which terminates Bibliography :— 82. Power, Henry, 'Experimental Philosophy,' in three books, con- taining new experiments, microscopical, mercurial, magnetical, 4to. London, 1644. 83. HOOKE, • Micrographia.' London, 1667. 84. LEEUWiiNHOEK, A., ' Anatomia rerum cum animatarum turn inanima- tarum ope Microscopiorum.' Lugd. Bat., 1687. 85. Leeuwenhoek, 'Select Works, containing his Microscopical Dis- coveries ;' translated by Samuel Hoole, plates, 4to. London, 1798- 1807. 86. Dereham, The Rev. W., ' Physico-Theology,' second edition, 1714. An ingenious teleological disquisition, containing a note on the fly's foot, p. 374, and many curious notes on insects. 87. INMAN, Thos., ' On the Feet of Insects.' Proceedings of the Liverpool Literary and Philosophical Soc, No. vi., p. 220. Liverpool, 1849. 88. West, Tuifen, 'The Foot of the Fly; its Structure and Action elucidated by Comparison with the Feet of Other Insects,' Part I., with 3 plates. Trans. Linn. Soc, vol. xxiii. (1859), t86i. 89. LowNE, B. T., 'On the so-called Suckers of Dytiscus, and the Pulvilli of Insects.' Monthly Micros. Journ., vol. v., 1871. THE THORACIC KXO-SKELETON. 191 distally in a peg — the anterior malleolus ; this malleolus («0 articulates with a depression in the trochanter. The posterior sclerite always exhibits a thin fenestra (/), surrounded by a thick rim ; and the latter supports a second peg — the posterior malleolus, which also articulates with the trochanter. When the limb is fully flexed on the coxa, the trochanter lies in the fenestra. The internal coxal plate is smaller than the others, and completes the inner aspect of the coxal joint. The Anterior Coxa is united by a loose syndesmosis with the sclerites of the prosternal area, and with the margin of the anterior sterno- coxal foramen. Its anterior plate forms a diarthrosis with the projecting process of the epitrochlear sclerite. The movements of the anterior sterno-co.xal articulation are very free : the anterior limbs are not only used as legs, but almost as arms, serving to clean the face and proboscis, and in climbing. These actions depend chiefly on the great mobility of the sterno-coxal articulation, as the prothorax — the manu- truncus of Kirby — is immovable on the mesothorax. In the predacious Diptera the anterior legs are used in seizing their prey, as in Empis and Dolichopus. The Intermediate Coxa is lodged in an elliptical depression between the meso- and metasternum (Plate VII., Fig. i, p, q), so that its movements are greatly restricted. The anterior sclerite articulates at its outer extremity by a peg (Fig. 2, a) with a deep socket in the mesosternum, on which the coxa rotates about its long axis ; this joint permits of a rowing movement of the femur in the horizontal plane in running, and forms what I call a roller-joint. This kind of articulation is very highly developed in some of the geodephagous Coleop- tcra, as, for example, in Passalus. The wide-open proximal margin of the co.xa is connected with the edges of the sternal foramen by a loose syndesmosis, which forms a kind of conjunctiva, in which the coxa moves. The inner plate articulates with the capitellum at the posterior ■extremity of the mcsothoracic entosternum. 192 THi: IXTEGUMENTAL SKELETON OE THE IMAGO . The Posterior Coxa is articulated with the metasternum, the epimeron of the metathorax, and with the ventral plate of the first abdominal segment. Its movements are more extensive than those of the intermediate coxa, as it is capable of abduc- tion, adduction and rotation. The latter movement is limited by a hook which overhangs a ridge at the inferior margin of the metapleuron, on which it travels as on a rail ; this hook springs from the anterior sclerite. The inner sclerite articulates with the capitellum on the entosternum of the metathorax. The freedom of movement in the posterior coxo-sternal articulation permits the posterior legs to be used in cleaning the wings and abdomen. The Femur is a tubular joint; its proximal extremity is par- tially separated from the rest by deep lateral vertical inflections and forms the trochanter. The constricted part of the femur is strengthened by a ridge-like fold, which is received in a furrow in the trochanter (Fig. 6). This arrangement permits of more or less movement in a horizontal plane, passing through the axis of the femur. The trochanter is slipper-shaped (Fig. 5). Its proximal DliSCRII'TlON OK PLATK IX. Fig. I. — The Left Anterior Leg seen from behind. Fig. 2. — The Right Intermediate Coxa seen from behind. Fig. 3. — The Left Intermediate Coxa seen from in front. Fig. 4. — The Left Posterior Coxa seen from behind : ex, anterior, and cx^, posterior, and ex", internal coxal plate ; f, the fenestra ; fe, femur ; h, hamulus of posterior coxa ; m, anterior, and ;«', posterior malleolus ; tr, trochanter ; ss, sensory plate ; .', tibia ; /', tarsus ; u, unguis ; p, pulvillus. Fig. 5. — The Left Anterior Trochanter seen from behind and above. Fig. 6. — The Proximal Extremity of the Left Anterior l"enuir, with the Trochanter removed. Fig. 7. — The Femoro-Tibial Articulation seen from its inner aspect : g, groove in the femur ; ex, exterior apodeme ; Jl, insertion of the flexor muscle. Fig. 8. — The Pulvilli, Claws and Planta of the Anterior Tarsus ; ventral aspect. Fig. 9. — Lateral view of the same. Fig. 10. — Upper portion of the Last Tarsal Joint seen from its under surface. Fig. II. — A portion of the Pulvillus of Carabus granulatus, showing the trumpet- shaped seta', after Tufi'en West. Fig. 12. — Tarsal Seta; of Exoletus ha.'morrhoidalis, after Tuffen West. Fig. 13. — Tarsal Setas: a, of Mylabris Cichoria;, after TutTen West; b, of Calli- phora erythrocephala, seen with j'j oil immersion lens. THE LEGS AND FEET. THE THORACIC EXO-SKELETON. 193 margin receives the malleolar pegs laterally, and is prolonged internally as an apodeme, which lies within the inner coxal sclerite, and gives attachment to the extensor muscles. The Coxo - Trochanteric Articulation is very complex, and is capable of flexion and extension, also of rotation on the long axis of the femur, and of rotation at right angles to the long axis of the femur on the long axis of the coxa. The anterior malleolar peg is horizontal, and the posterior vertical. The elevation of the anterior edge of the trochanter throws the tibia forwards, and draws the posterior malleolar peg from its socket ; this permits of the depression of the distal end of the femur on the anterior malleolar peg. The elevation of the posterior edge of the trochanter rotates the femur backwards on its long axis, and carries the insect forwards over the tarsus ; this movement withdraws the anterior malleolar peg from its socket, and replaces the posterior ; the femur is then swung forward on the vertical posterior malleolar peg. In extreme flexion of the femur on the coxa both pegs are in their sockets, and the coxo-trochanteric articulation is locked. Thus in running, when the foot and tibia are thrown back by rotation of the femur, the latter swings forward in the horizontal plane on the posterior malleolar peg ; but when the foot is brought to the ground in front of the femoral plane, the distal end of the femur is depressed — that is, the femur is extended on the anterior malleolar peg; a rotation of the femur then urges the insect forwards, and unlocks the anterior and locks the posterior coxo-trochanteric diarthrosis. The movements of the shank of the femur on the trochanter are apparently very slight, and depend entirely on the elasticity of the integument uniting these parts. The Tibia is sub-cylindrical, slightly curved on its long axis, and thickest at its proximal extremity. It forms a hinge with the femur, the distal extremity of which is hollowed out below for its reception in extreme flexion. The proximal extremity of the tibia ends in two curved ridges (Fig. 7), which articulate with two grooves in the interior of the lateral portion of the distal extremity of the 194 THE INTEGUMENTAL SKELETON OF THE IMAGO. femur (g). The flexor muscle is inserted into the syndes- mosis {^fi) between the femur and tibia on the ventral surface of the limb, and the extensor {ex) into the femur dorsally to the articular apodemes. The movements of the femoro-tibial articulation are strictly limited to flexion and extension. The Tarsus, or foot, consists of five joints, each contracted' at its proximal extremity, and exhibiting a capitellum, or head, which articulates with a concave socket on the dorsal distal margin of the joint above it. The tarsal articulations are true ball-and-socket joints, which permit of considerable lateral movement, as well as of flexion, extension, and limited rotation. The ventral margins of both the proximal and distal extremities of these joints are united with the adjacent joints by loose syndesmoses. (Fig. 20, o, is a diagrammatic representation of this form of joint.) The plantar surfaces of the tarsal joints support combs of stiff bristles, which are used in cleaning the surface of the integument and the setas with which it is covered. The comb is most conspicuous on the anterior tarsi, and is rudimentary or absent on those of the intermediate legs. The terminal tarsal joint supports the pads, pulvilli {p), and the claws, ungues {u), as well as a plate — the planta or empo- dium — at its distal extremity. The Claws, or Ungues, are strong, curved, hollow sclerites. Each has a sub-hemispherical head, which articulates with a notch in a thick cordiform swelling on the under surface of the dorsal aspect of the distal margin of the last tarsal joint. The claws of the posterior tarsi are more slender and less curved than those of the four anterior feet. The upper edges of the claws are united by syndesmosis with the distal extremity of the last tarsal joint. Some writers regard the plantae, claws, and pulvilli as a sixth tarsal joint ; the rudimentary tarsus of the nymph, how- ever, has only five joints (Fig. 34), so that the claws and pulvilli must be regarded as paired appendages of the last joint. The planta is a sclerite in the extremity of the limb, and not a distinct limb segment. THE THORACIC EXO-SKELETON. 195 The Planta (PI. IX., Figs. 8 and 9) is a subquadrate plate, triangular in tiie posterior tarsus, which articulates with a deep emargination in the plantar edge of the last tarsal joint ; it bears a long median setose spine, which projects between the pulvilli. Its inner (upper) surface gives insertion to the strong apodeme of the flexor tarsi muscle, and its lateral edges articu- late with the stalks of the pulvilli. These edges exhibit a series of parallel plications or ridges. The Pulvillus is a membranous, somewhat pyriform flattened sac, the narrow neck of which is strengthened by a chitinous ring articulating by syndesmosis with the edge of the planta, .?nd with the distal margin of the last tarsal joint. This ring supports a fan of chitinized ridges, which radiate over the dorsal surface of the pad (PI. IX., Fig. 10). The pulvilli are the cushions, by means of which flies and some other insects climb windows or walkover the lower surface of glass, or other smooth and polished bodies. With an oil immersion (iV) the pad is seen to be covered on its under surface with papillae, arranged in very regular rows, each papilla having a minute orifice at its extremity. Towards the edges of the pad these papillae become longer, and give place to long, hollow setae, with trumpet-shaped orifices, which form a dense fringe projecting beyond the edges of the pad. Henry Power [82] is the first author who mentions these pads, and his description of the feet of the fly is so curious that I shall give it in extenso. He says : ' She (the fly) hath six legs, but goes only upon four ; the two foremost she makes use of instead of hands, with which you may often see her wipe her mouth and nose, and take up anything to eat. The other four legs are cloven, and armed with little clea's, or tallons (like a Catamount), by which she layes hold on the rugosities and asperities of all bodies she walks over, even to the supportance of herself, though with her back downwards, and perpendicularly invers'd to the Horizon. To which purpose also the wisdom of Nature hath endued her with another singular Artifice, and that is a fuzzy 13 196 THE INTEGUMENTAL SKELETON OF THE IMAGO. kinde of substance like little sponges, with which she hath lined the soles of her feet, which substance is always repleated with a whitish, viscous liquor, which she can at pleasure squeeze out, and so sodder and be-glew herself to the plain she walks on, which otherways her gravity would hinder' (p. 5). This statement of Power's was subsequently controverted by Leeuwenhoek [84], who supposed that the minute hairs on the pads have a hold even on the polished surface of glass, and act as tenterhooks ; whilst Dereham [86] suggested that the pads act as suckers, and that the insect is supported by atmospheric pressure. This view has been the popular one ever since, but Blackwall* and Inman [87] reverted to Power's original sugges- tion, which is certainly the correct one. The minute size of the papilla; and setae on the under surface of the pads formerly made their investigation exceed- ingly difficult, but the great similarity of these to the so-called suckers of Dytiscus, and the pulvilli of the Harpalidse, has long been recognised. Tuffen West [88] compared the pads of the Diptera with the pulvilli of the Coleoptera, and gives a large number of beautiful and elaborate figures of the pads of various insects, which are most accurate in all their details. West supposed that the separate seta; act as suckers. In 1871 I showed at the Royal Microscopical Society that the great water-beetle remains suspended by its so-called suckers in a very perfect air-pump vacuum [89]. Flies also walk perfectly well over the exhausted dome of an air-pump. Gilbert White, in his ' Natural History of Selborne,' records some interesting observations which he made on flies in autumn, when they are feeble. He states that they often struggle to remove their feet from glass as if they were firmly glued to its surface. In my former work on the Blow-fly [62], I made the following statements, which still represent my views on this subject : ' There is no essential difference in the pads of flies and the pulvilli of beetles, moths, and other insects ; a similar fluid is * ' Remarks on the Pulvilli of Insects.' Trans. Linn. Soc, Lond., vol. xvi., p. 767. 1883. THE THORACIC EXO-SKELETON. 197 secreted in all. The only difference is that the pads of flies are membranous and transparent, instead of hard and opaque. ' The feet of the smaller house-fly {Musca corvina) are the best to show the manner in which the viscid fluid exudes from the extremities of the trumpet-shaped hairs, as they are very large in this species, and a glistening bead of fluid can be seen plainly at the extremity of each hair by placing the living insect under the microscope. The footprints left upon glass by flies consist of rows of dots corresponding to these hairs ; this is best seen in those of the lesser house-fly from their greater size. ' The whole appears precisely analogous to the manner in which caterpillars and spiders suspend themselves by silken threads. In both cases the fluid is exuded from minute pores, and bears the weight of the insect, the only difference being in the nature and quantity of the fluid exuded. Much discussion has arisen as to the manner in which flies liberate their feet, and it has even been objected that they would become so firmly adherent after a time that the insect would be glued to the spot. Nothing can be more simple than the arrangement by which the foot is liberated, and in the healthy insect the secretion probably never becomes solid as long as it remains in contact with the foot. It is sufficiently glutinous, even in the fluid, or rather semi-fluid, state it assumes as it exudes, to sustain the weight of the insect, when the strain is put equally upon all the hairs, of which there are about 1,200 on each pad; but when the pad is removed obliquely, so that each row is detached separately, the resistance amounts practically to nothing. ' The direction and length of the hairs upon the pad are so adapted to the oblique direction in which the strain is put upon them when the tarsus is straight, that the insect has a perfectly secure hold ; this is immediately released as soon as the tarsus is curved, which is effected by the long slender tendon of the flexor tarsi. In the small house-fly the pads themselves are capable of being curved, for the tarsal tendon branches, and is inserted into the distal extremity of each pad.' 13—2 198 THE INTECUMENTAL SKELETON OF THE IMAGO. The adhesive fluid which covers the plantar surface of the pads is secreted by a pair of simple saccular glands. These occupy the greater part of the cavity of the four last tarsal joints. The secretion exudes through the hollow papillae and hairs, often so rapidly when a fly is captured and held between the finger and thumb by the thorax, that it forms small glisten- ing drops on the pads, and may be collected on a slip of glass, when it immediately becomes solid. The tarsal secretion solidifies equally well under water, in which the coagulum is quite insoluble. e. The Wings and Mechanism of Flight. {Plate X.) It is well known that the wings of insects generally present three well-marked areas or regions— the costal, median, and posterior or anal areas. These areas are easily distinguished in the wings of the Diptera ; the costal area is bounded behind by the great subcostal nervure, and in front by the margin of the wing. The median area is also called the disc or discal area, a term which I prefer to median. The disc is bounded behind by the median nervure. The posterior area I shall term the ' patagium '; it is folded when the wing is flexed, and is supported by the Bibliography :— 90. BORELLUS, J. A., ' De motu animalium,' 410, 2 vols. Lugd. Bat., 1 7 10. 91. I'ETTlGREW, J. Bell, 'On the Mechanical Appliances by which Flight is attained in the Animal Kingdom.' Linn. Soc. Trins., Lond. (1867), 1868, vol. XXVI. 92. Marev, E. J., 'Animal iVIechanism : a Treatise on Terrestrial and Aiirial Locomotion,' 1873 ; 2nd edit., 1874. Internat. Sci. Ser., vol. xi., Paris and London. 93. Lendenfeld, R. von, ' Der Flug der Libellen ; ein Beitrag zur Anatomie und Physiologic der Flugorgane der Insecten.' 7 pl.ites and figures. Sitzungb. der K. Akad. Math. Naturwissen. CI., Wien. Bd. Ixxxiii., pp. 289-376, 1881. 94. AnOLPH, E., ' Die Dipterenfliigel, ihr Schema und ihre Ableitung.' Nov. Act. C.L.C. Acad., Bd. xlvii., 1885. This paper is entirely systematic. 95. Amans, p. C, ' Comparaisons des Organes du Vol dans le .Serie Animal.' Ann. Sc. Nat. Zool., ser. vi., torn, xix., 1885. This paper is of great length, but is inaccurate and does not advance our knowledge of the subject. THE THORACIC EXO-SKELETON. I99 mctapterygium, from which all its nervures arise. Behind the patafTium in the Blow-flies a small segment is separated from the rest of the wing by a notch (I term this the ' lobulus ')> ^.nd behind the lobulus are the squamula and the squama. The nomenclature of the nervures which I have adopted will be sufficiently understood by the figure (PI. X.). The Wing-Roots. — The wing of the Blow-fly is apparently supported by only two roots ; an anterior, from which the nervures of the marginal and discal areas arise, and a posterior, which supports the nervures of the patagium. The anterior wing-root consists, however, of the united pro- and mesopterygium. It presents five carpoid sclerites ; these support the marginal and the common root of the subcostal and discal nervures, which I term the ' remigium.' Its move- ments are quite independent of those effected by the meta- pterygium. The carpoid sclerites are arranged in a proximal and a distal series. I term the proximal sclerites the epaulet and the dens ; they articulate with the dorsum and with the sclerites of the distal row. The latter are the sub-epaulet, the coracoid, and the unguiculus (Mihi) ; they articulate with the plcuron, with the proximal sclerites, and with the marginal nervure and the remigium. The Dens (PI X., Figs. 7-9) consists of a body and three processes. The body is irregular in form, narrowed in front, forming the neck, which supports an oval convex scale, seen between the epaulet and the coracoid. The neck terminates in a rounded head concealed by the scale, this articulates with a socket formed by the uncinate process of the anterior parascutal plate. Behind, the dentate process projects downwards and backwards, and lies upon the posterior surface of the unguiculus. Internally, there are two processes, an anterior, which projects inwards and forwards and receives the insertion of the accessory elevator muscle of the wing. The second or posterior process is far larger ; its distal extremity articulates by a cup-shaped cavity with the head of the great alar apophysis ; a line joining the head of the dens with this cup-shaped cavity may be termed the axis of the dens — it is the axis on which the wing is raised and depressed. The direction of this axis undergoes rotation about the head of the dens with increased convexity or flattening of the dorsal valve. Increasing convexity of the dorsum produces the descent of the wing, pushes the alar apophysis forwards beneath the head of the dens, and renders its axis nearly vertical, so that the 20O THE INTEGUMENTAL SKELETON OF THE IMAGO. anterior margin of the wing descends in a semicircular arc, with its convexity behind and its upper limb vertical. During the ascent of the wing the alar apophysis is drawn back, and the axis of the dens is directed backwards and slightly upwards, hence the anterior margin of the wing ascends in a curve, the upper half of which is convex in front. When the wing is at rest and flexed, the axis of the dens is nearly horizontal, and the dentate process depresses the patagium and assists in folding the wing. The Dens occupies the posterior superior part of the propterygium, and only a small portion of it can be seen externally, the dentate process with the neck and the scale. It is apparently represented in the Hymenoptera by two distinct sclerites, the omoplate and signioidea of Chabrier [72, vol. viii., p. 73]. The Epaulet (PI. X., c) is a large scale fringed by black bristles, which articulates with the dens behind and with the parascutal plate and the coracoid by syndesmosis ; it covers the sub-epaulet, but does not appear to take any pai t in the formation of the wing-joint, only protecting it in front and above. It does not correspond with the tegula of the Hymenoptera, with which it has been confounded, as this is in front of, and not behind, the anterior spiracle. Description of Plate X. Fig. I.— The left wing of the Blow-fly. Fig. 2. — The roots of the same more highly magnified. Fig. 3. — The base of the marginal nervure and of the Remigium of the left wing. Fig. 4. — The base of the left Remigium seen from in front and below. Fig. 5. — The dorsal aspect of the base of the right Remigium. Fig. 6. — The Epaulets of the right winjj, upper and anterior surface. Fig. 7. — The left Dens seen from above and in front. Fig. 8. — The left Alar Apophysis and Dens seen from the under surface. Fig. 9.- — The left Dens seen from the interior of the thorax. Fig. 10. — The left Unguiculus seen from behind. Fig. II. — The right Unguiculus seen from in front. Fig. 12. — The left Parascuta, Dens and Alar Apophysis. The following references arc the same in all the figures in wh ch they occur : a, a', apodemes of the pre-epaulet ; aa, alar apophysis ; ap, anterior apodeme, and ap' , posterior apodeme of the remigium ; cr, coracoid ; d, dens ; dp, dentate process of the dens ; e, epaulet ; e', sub-epaulet ; cd, epaulet or .scale of the dens ; /;, hypopterygium ; ha, hamulus ; /', conoid ; /, lobulus ; w, marginal nervure ; mp, metapterygium ; pap, posterior alar apophysis (alar apophysis of the scu- tellum) ; //, posterior, and ps, anterior parascuta ; r, remigium ; s, squama ; s', hy|)opterygial sclerite ; sqa, squaniula ; st, stirrup of the unguiculus ; tb, part of the tympanic bulla ; tm, tympanic membrane ; ti, unguiculus ; «', uncinate process of the anterior parascutum ; A, the deltoid ; T, the tau. Nervures of the wing: m, marginal; /, mediastinal; z, sub -costal; J, radial; 4, ulnar; S, median ; 6, .sub - median ; 7, anal ; , Fig. 39) is produced by the contraction of the primitive band, and it communicates with the yelk in front of the ccelomic sacs and behind the procephalic lobes. As the primitive band contracts, the head and tail-folds of the amnion separate, so that the cavity, p, is only divided from the inter- amnial space by its roof of parablast. It becomes the pericardial cavity. Segmentation of the Somatopleure and Ccelomic Sacs.— Even before the epiblast of the somatopleure covers the dorsal Fig. 41.— a dorsal view of an embryo before the somatopleural plates meet on the dorsum of the yelk, showing their segmentation. surface of the yelk, the annular segmentation of the larva becomes apparent by the formation of deep inflections between the segments. These inflections at first only involve the epiblast, but soon extend to the mesoblast. The ccelomic sacs form two cellular plates, with a third incomplete plate between them (PI. XIV., Fig. 2). The external and internal plates become adherent to the involutions of the epiblast between the somites, so that each ccelomic sac is divided into a series of secondary cavities, corresponding in number to the annuli of the epiblast. I have not hitherto seen this stage of development in any of my sections, but Graber gives figures of embryos in which each somite exhibits a distinct cavity between the outer and inner ccelomic plates. ORIGIN OF THE ARCHENTERON. 253 The coelomic plates ultimately form the somatic muscles, and I am inclined to think that the inner plate (the Darm- driisenblatt of the Germans) enters into the formation of the muscles of the alimentary canal. Cholodkowsky*> has very carefully studied the development of the ctclomic cavity in ' lilatta Gcrmanica,' and arrives at the following conclusions, which are quite consonant with the view I have given above of the changes which the coelomic sacs undergo in the ISlow-fly embryo : 1. ' In "Blatta Gcrmanica" the body cavity is developed from eighteen pairs of hollow somites, which originate from the segmentation of the epiblast and coelomic sacs.' 2. 'The cavities of the somites each divide, as in Peripatus, mto three parts, one of which is probably homologous with the segmental funnel.' 3. ' The somatic cavities ultimately become broken up (aufgegeben), and open into the periccelomic space, which corresponds with the segmentation cavity, and is permeated by the parablast.' 4. ' The ccclom, or body cavity, in its definitive stage, has therefore a three- fold origin ; it consists of the primitive cavities of the somites {i.e., the cavities" of the segmented coelomic sacs), of the primitive segmentation cavity, and of a true schizocoele, which is subsequently formed.' I am unable to trace any appearance of a true schizocoele in the Fly embryo, but think it probable that only a small portion of the inner coelomic plate enters into the formation of the alimentary canal. If this is so, a schizoccele may be assumed to e.xist, although the invasion of its cavity by spongy parablast masks its true character. 5. 'The heart is developed from the primitive segmentation cavity.' I think this statement should be : the segmentation cavity forms the pericardium, not the heart. 6. ' The fat bodies are derived from yelk cells, which at a certain stage are amoeboid, and wander into the body cavity.' This is merely another form of the statement that the fat bodies are of parablastic origin. 7. With regard to the origin of the coelomic sacs in Blatta, the memoir quoted gives little information, but at the moment of going to press I received a much-extended memoir of Cholodkowsky's on the development of Blatta Germanica.t In this paper he states that the mesoblast (his inner germinal layer) is derived from an invagination which takes place in the ventral median line, the primitive trace (Primitivrinne). I think that the drawings given by Cholodkowsky of the primitive trace, which are accurate representations of sections, do not bear the interpretation he puts upon them. In this I am in unison with Tichomirow, who denies that the primitive trace has anything to do with the formation of the mesoblast. * Cholodkowsky, N., 'Zur Embryologie von Blatta Germanica,' Zool. Anzcig., Bd. xiii., p. 137, 1890. . t Cholodkowsky, ' Die Embryonalentwicklung von Phyllodromia (Blatta) Germanica," Mem. de I'Acad. Imp. des Sc. dc St. Petersbourg, torn, xxxviii. No. 5, 1891. 354 THE EMBRYOLOGY OF THE BLOW-FLY IN THE EGG. In this connection I would mention that Metschinikoff* figures coelomic sacs in Simulia exactly similar to those I claim to have discovered in the Fly embryo, although they are unlettered in his figures, and are not referred to in his text. As Metschinikoflf only saw them in optical section, he apparently mistook them for the remains of the yelk. Similar appearances are also seen in the embryo of Chironomus. 4. THE DORSAL ORGAN OF KOWALEVSKI. Kowalevski [97] first described a structure in Hydrophilus which he termed the dorsal organ. He regarded it as a thickening of the outer provisional membrane, the serosa, and states that it afterwards forms a tube and sinks into the yelk. Grabert has more recently investigated the subject, and accords with Kowalevski; but both Graber and Kowalevski regard the whole of the blastoderm beyond the primitive band as serosa. In Lina, Graber derives the dorsal organ from the amnion, and Cholodkowsky states that it is formed from the serosa in Blatta.J It appears to me that these differences arise partly from a confusion of nomenclature, and partly from the very extra- ordinary changes of position which the amnion is believed to undergo by Graber. The dorsal plate always originates from a portion of the blastocele which is in contact with the yelk. Cholodkowsky says that its development in Blatta Germanica is well seen, and that the dorsal margin of the posterior amnial fold ends in a thick plate of contracted serosa, consisting of tolerably high cylindrical cells, the dorsal organ. This plate becomes shortened, and forms an oval fossa, which is in- vaginated in the food-yelk. From the bottom of the invagina- tion large, irregular, often amoeboid, cells arise and wander into the yelk. Cholodkowsky's figures convince me that his dorsal organ is my mesenteron. The scattering of the cells in the yelk, which * Metschinikoff, E., ' Embryologische Studien an Insecten,' Zeitsch. f. w. Zool., Bd. xvi., i866. t Graber, V., ' Vergleichende Studien iiber die Keimhiillen und die Ruckenbildung der Insecten,' Denkschrift. K. Akad. Math. Naturwiss. CI Wien, Hd. Iv., i888. % Mem. de I'Acad. St. Petersbourg, torn, xxxviii. THE DORSAL ORGAN OF KOWALEVSKI 255 he believes occurs, is, I think, the result of imperfect fixation ; and I have no faith in the method employed by Cholodkowsky (chromo-nitric acid). Graber gives beautiful figures of the dorsal organ in Hydro- philus, Melolontha, Pyrrhocoris, and Musca, although in the latter he mistook it for a ventral invagination. In all these insects it becomes an epithelial tube. Graber, after describing the formation of the dorsal tube by the invagination of the dorsal plate (his dorsal ptychoblast) and its sinking into the yelk, says : ' The cells give off amoe- boid processes, which probably feed upon the yelk.' At a later stage, according to Graber, it separates into its con- stituent cells, which become wander-cells in the yelk. In these statements Graber and Cholodkowsky agree so far as Hydrophilus and Blatta are concerned ; but Graber has more recently described the same organ in the Blow-fly embryo as the rudiment of the hind gut. The identity of Graber's hind gut in the fly embryo with Kowalevski's dorsal organ and with my archenteron is certain. The only questions which remain uncertain are : Are Graber and myself right in maintaining that it is part of the alimen- tary canal in the Muscidae ? And is Graber right in regarding it as the hind gut, or am I right in regarding it as a typical gastrulation forming the archenteron ? I hold that the dorsal plate, when it first appears in the fly embryo, is neither covered by the amnion nor serosa, and that these membranes extend over it subsequently. Compare Graber's diagram {I.e., p. 254; Fig. 20, p. 18), which shows the membranous fold of the amnion and serosa overlapping it in Hydrophilus. Of course, if the whole blastoderm outside the primitive band is called serosa, and not only its folds (Falten- blatter), then the dorsal plate is developed from the serosa. Graber, in Lina, derives it from the amnion, which he thinks first splits in the ventral median line, then travels through the yelk to the dorsum, where its torn edges reunite. This theory appears to me so improbable that I cannot entertain it. The idea that the membranes, amnion and serosa, undergo 256 THE EMBRYOLOGY OE THE BLOW-FLY IN THE EGG. rupture during development is very old. Weismann conceived that in most Insects the development of the primitive band is accompanied by a rupture of the blastodermic vesicle. He divided Insects into two groups, one in which he believed such splitting occurs, which he named regmagene Insecien ; others in which he believed no rupture to occur he styled aregma- genc. According to the same authority, Chironomus, Simuha, Pulex, Donacia, and Phryganea are 'regmagene,' whilst Musca and Melophagus are ' aregmagene.' E. Metschinikoff * in 1886, said on this point : ' Although it must be admitted that the primitive band has not the same rela- tion to the remainder of the blastoderm in different insects, it must not be concluded that it is differently developed since no tearing of the blastoderm occurs. The variation is due to the complete or incomplete character of the amnion.' In Weismann's ' aregmagener Insecten ' no amnion is de- veloped, and, as I have already stated, an amnion is developed in the fly embryo. If I am correct in this, unless the Pupi- parae are without an amnion, Weismann's second class, re- garded as insects in which the amnion is not formed, does not exist. According to my observations, the development of the fly embryo is very similar to that of Chironomus, as given by Weismann, and the apparent rupture of the serosa described by him in the latter is merely the result of the shortening of the primitive band and the separation of the head and tail folds. I confess that the exact relation of the dorsal plate to the amnion has greatly puzzled me, although, if Graber's figures are correct (PI. XIII.), it is clearly first formed before the serosa exists, and, if my interpretation of them is right, when the amnion is a thick-walled epithelial tube ; and is only after- wards covered by the extension of the amnion and serosa over the dorsal surface of the embryo. It is important to remark that both the dorsal involution of ° L.c, p. 254. THE NYMPHOID STAGE OF THE EMBRYO. 257 the dorsal plate and the proctodaeum exist at the same time in Hydrophilus (Graber) and in Blatta (Cholodkowsky), as this is strong confirmatory evidence in favour of my statement that the proctodaium and dorsal involution are distinct in the fly embryo. Moreover, it disposes of Graber's contention that the dorsal invagination — my archenteron — is a hind gut, and it establishes the relation of the Malpighian tubules with the mid-gut, and not with the proctodaeum (see Development of Malpighian Tubules). 5. THE NYMPHOID STAGE OF THE EMBRYO AND ITS CON- VERSION INTO THE LARVA. The Nymphoid Stage of the Embryo. — About twelve hours after mpregnation the embryo has assumed what I term the nym- FlG. 42.- -An embryo about twelve hours old in the Weismann.) nymphoid stage. (After phoid stage; the procephalic lobes meet in the middle line above, and the rudiments of the cephalic post-oral appendages attain their maximum development. The head-capsule in front of the mouth and between and below the procephalic lobes pro- jects as a vesicular swelling — the fore-head. The remainder of the embryo is segmented as in the larva. In this condition it has a closer resemblance to the nymph than at any inter- mediate period ; the thoracic appendages are, however, entirely wanting, and the rudimentary cephalic appendages exhibit a verj' generalized type. Formation of the Larva. — After the twelfth hour the retro- gression of the developmental process becomes apparent, and 258 THE EMBRYOLOGY OF THE BLOW-FLY IN THE EGG. at the same time sulci appear in the epiblast, separating the larval segments. The fore-head and procephalic lobes rapidly disappear, and the rudimentary cephalic appendages undergo the following modifications. The anterior pair disappear. According to Weismann, they approach each other and form a median outgrowth in the anterior wall of the stomodaeum. On Fk;. 43. — The anterior extremity of the Hlow-fly eml)ryo in five successive stages, showing the transformation of the mouth organs : /, a side view of the head of the embryo before the segmentation of the coelomic plates, after Weismann ; 2, a later stage of the same ; $ and 4, still later stages ; j, a ventral view of the head in the same stage as./ ; (5, a ventral view of the head in the mature embryo ready to escape from the egg. the same authority the median tooth of the larval mouth armature represents their remains, and is therefore composed of the united mandibles. I am by no means sure that such is its origin, and, as already stated, regard it as the labrum. The second pair of appendages, the maxillae, become parallel with each other, and form the maxillae of the larva. The third pair Descrii'tion of Pi.ate XV. Two sections through the head of an embryo during the involution of the head discs (procephalic lobes). Fk; I. — A section through the fore-head and maxillse : //;, the fore-head, showing the rudimentary brain and brain vesicles, v ; the single layer of cells enclosing the vesicles, probably becomes the retinal disc ; the cleft, inin, is the orifice of the stomodxum ; iiix, the maxilla ; r', the maxillary disc ; s g, salivary duct. The cellular mass, i\ is probably part of the rudimentary pharynx. Flc. 2. — A sectiim through the paracephala (procephalic lobes) of the same embryo : d V, dorsal vessel ; tiiiis, rudimentary muscles ; tg, ganglion of the ventral chain ; St, stomoda-'um ; »', antennal, and i-, optic imaginal discs ; s g, salivary duct. EMUUYO, LATE STAGE. THE NYMPHOID STAGE OF THE EMBRYO. 259 unite behind the mouth and rapidly undergo a great reduction in size ; they form the labium. There is no doubt in my mind with regard to the fate of the procephalic lobes and fore-head, although the meagre facts which have been hitherto recorded are by no means in perfect accord with my view. Weismann is the only author who claims to have observed the phenomena, and his observations are somewhat vague, and in some points, I think, incorrect. The changes which take place during the involution of the procephalic lobes will be more readily understood by a refer- ence to Figure 44 than by the most lengthy description. They are probably brought about by the rapid growth of the anterior part of the stomodaeum ; and by the increased size of the somites developed from the somatopleure, which causes the prothoracic segment to advance towards the anterior extremity of the embryo. I have sections which show that the procephalic lobes form a double involution on each side, and, from the condition of the great cephalic discs in the newly-hatched larva, I think it is manifest that their antennal and optic rudiments are formed by the invagination of the procephalic lobes. Although I have not been able to demonstrate the invagination of the fore-head at this period, it undoubtedly rapidly disappears, and the invagination of this part is so clearly seen in the newly-hatched larva, that I can see no reason to doubt that it forms the cavity above the labrum already described (p. 41). This communicates below with the sacs which enclose the antennal and optic rudi- ments or discs. Weismann says: ' When the fore-head is definitely separated from the procephalic lobes, it appears as an obliquely truncated process, having the form of a four-sided prism. Its ventral surface is convex, but neither this nor the upper flat surface exhibits any median fissure. After the anterior maxillae become parallel, the fore-head is bent towards the abdominal surface, and rapidly assumes a position in which its anterior truncated surface becomes ventral ; its ventral surface covers the mouth and lies upon the primitive band, which is only separated from 26o THE EMBRYOLOGY OF THE BLOW-FLY IN THE EGG. Pn;_ AA Three diagrams representing the manner in which the head discs and nervous system of the embryo arc related — / and 2, before the retraction, and ?, after the retraction, of the fore-head to form the cephalo-pharyngeal sac : c, the crop ; g, cephaHc ganglion ; /', antennal, and P, optic discs ; iiix, 6rst pair of maxillx ; /■/<, cephalo-pharynx ; r, the ring ; s, salivary duct ; st, stomodx'um ; z\ hollow ventricle of the brain ; v g, ventral ganglia. The invaginated epiblast in 3 is represented by a dark single line in / and 2. The dotted lines in _' and 3 trace the course of the tracheal vessels of the discs, which arise close to the position of the future anterior spiracle. / and 2 represent embryos in the stage from which the sections represented in I'late XV. were taken ; j exhibits the relations of the parts in the young larva. THE NYMPHOID STAGE OF THE EMBRYO. 261 it by a transverse line. The mandibles at this period lie in the mouth-cleft, and sink more and more deeply into it. It is remarkable that the fore-head takes no part in the formation of the larval head, but is entirely invaginated ' [2, p. 64]. On this point Weismann adds : ' After the invagination of the fore-head within the mouth, both pairs of maxillae grow forwards ; especially the anterior, which become considerably enlarged, project beyond the fore-head, and draw the pro- cephalic lobes together, whilst the latter become smaller and smaller, and unite with the maxilla; [2, p. 67] . With the exception of the last statement, Weismann's account appears to me to be quite correct. Weismann believed that the dorsal sensory organs on the maxillae of the larva represent antennae, and hence he supposed that the procephalic lobes unite with the rudimentary maxilla. Van Rees [147] confirms by his views and observations an opinion which I expressed as early as 1872 [142], that the pro- cephalic lobes are invaginated to form the cephalic discs, although Graber is led to a different conclusion, as he asserts that these discs in the fly embryo are at first composed of a single layer of cells [114]. Certainly, his drawings lead to such a conclusion, but in the face of the evidence afforded by my sections and the observations of Van Rees, I cannot regard Graber's sections as conclusive. It is possible that certain sections, through the thick edges of the discs, would present these appearances, but I think it probable that further investi- gations will show that the disc cavity which exists at the earliest stage of development is never entirely obliterated. Graber's sections are apparently made from embryos in a more advanced condition than mine, and it may be that at this stage the disc cavity is relatively small, and that the edges of the disc form considerable solid outgrowths, over which the cavity subsequently extends; but on this point further observations are needed. In newly-hatched larvae the disc cavities are certainly distinct. CHAPTER VIII. THE GENERAL ANATOMY OR HISTOLOGY OF THE BLOW-FLY. It will perhaps be useful to the general reader, as well as to the student, who may not be conversant with the peculiarities of the tissues in insects, to devote a chapter to their considera- tion, before entering upon a description of the developmental Bibliography.— As recent histological literature is very extensive, the reader who wishes for an extended Bibliography is referred to Quain's 'Anatomy,' loth edit., 1891, vol. i., part 2, 'General Anatomy or Histology.' The following works are more or less especially devoted to the histology of Insects, or are quoted in the text : 122. Leydig, F., 'Zum feineren Bau der Arthropoden.' Miill Archiv 1855. 123. Leydig, F., 'Lehrbuch der Histologic des Menschcn und derThiere.' Frankfort-a-M., 1857. 124. Weismann, a., ' Ueber die Zwei Typen contractilen Gewebes und ihre Vertheilung in die grossen Gruppen des Thierreichs, sowie iiber die histologische Bedeutung ihrer Formelemente.' Zeitsch. f. ration. Medicin, Bd. xv., 1862. 125. Landois, H., ' Beobachtungen iiber das Blut der Insecten.' Zeitsch. f. w. Zool., Bd. xiv., 1864. 126. BuTSCHLl, 0.,'ZurEntwickelungsgeschichteder Biene.' Zeitsch. f. w. Zool., Bd. XX., 1870. 127. Balbiani, E. G., ' Sur la structure du Noyau des cellules salivaires chez les larves de Chironomus ' Zool. Anzeiger, 1880, p. 662. 127. Viallanes (see page 25). Ann. Sc. Nat., ser. vi., T. xiv. 128. Levdig, F., Zelle und (^.ewebe. Neue Beitriige zur Histologie des Thierkiirpers, 8vo, Bonn, 1885. 129. Carnov, J. B., 'La Cytodi<5r6se chez les Arthropodes.' La Cellule, T. iii., Louvain, 1886. 130. Frenzel, J., ' Einiges iiber den Mitteldarm der Insecten sowie iiber Epithelregeneration.' Archiv f. mikros. Anat., I!d. xxvi., 1886. 131. Shaffer, C., 'Beitriige zur Histologic der Insecten.' Spengel. Zool. Jahrbuch, Bd. iii., 1887. 132. Faussek. v., ' Beitriige zur Histologie des Darmkanals der Insecten.' Zeitsch. f. w. Zool., Bd. xlix., 1887. CELLS AND NUCLEL 263 changes which occur in the pupa, and before giving a detailed account of the various organs of the imago. Upon making a careful examination by the aid of the micro- scope, it will be found that the number of distinct forms of texture or tissue is comparatively small, and, although many differences may be observed in different parts, a study of the transitional forms shows that all the tissues may be classed under one or other of the following heads : 1. Epithelial tissues. 2. Amoeboid cells. 3. Connective tissues. 4. Muscular tissues. 5. Nervous tissues. When any tissue is examined with the microscope, it may be resolved into a number of parts, which are repeated over and over again. These parts are termed tissue-elements, and many tissues retain certain structural elements which are termed cells, and which closely resemble the primary cell from which the whole body is developed. Others, as the muscles and nerves, although developed from cells, soon become so modified that their cellular character is not apparent at first sight, but a careful study shows indications of their cellular origin. In addition to the cellular constituents, many of the tissues exhibit a more or less homogeneous intercellular substance, 133. Rath, Otto Von, ' Hautsinnes-organe der Insecten.' Zeitsch. f. w. Zool., Bd. xlvi., 1888. 134. BuTSCHLl UND ScHEWl AKOFF, ' Ueber der feineren Bau der quergstiiften Muskeln der Anhropoden.' Biol. Centralblatt, Bd. xi., P- 33. 1891. 135. SCH.AI'ER, E. A., ' On the Minute Structure of the Muscle-Columns, or Sarcostyles, which form the Wing-Muscles of Insects.' Proc. Roy. See. Lond., vol. xlix., p. 280, 1891. 136. SCHAFER, E. A., ' General Anatomy or Histology.' Quain's ' Anatomy,' edit, x., vol. i., pt. 2., published separately. London and New York, 1891. 18 264 GENERAL ANATOMY OR HISTOLOGY OF BLOWFLY. termed the matrix. This matrix may be laminated, and the cuticle of insects has been regarded by some as substantially an intercellular matrix. Other matrices exhibit a distinct fibrillation, as in fibro-cartilage, but I am in doubt if such are ever found in insects, although fibrillation of the cuticular layer has been described in Crustacea and Arachnids. The various tissues may be arranged according to their origin. Thus the epithelial, nervous and muscular tissues are developed from the cells of the blastoderm, whilst the amoe- boid cells and the connective tissues originate from the parablast. The principal modifications exhibited by cells and nuclei may be conveniently considered first, and a description of the several tissues classified under the subjoined heads will follow : 1. The parablastic tissues. 2. The epithelia ; and 3. The muscles and nerves. 1. CELLS AND NUCLEI. Cells. — The parts called cells are corpuscles, which are very similar in insects and in all other animals ; indeed, they cannot be said to differ in any essential character from those of which the human body is built. In the fly the largest cells are li^ of an inch in diameter, or twice as large as the largest in the human body, whilst the smallest do not exceed 17707777 of an inch, and are smaller than any of those found in the tissues of man. Some cells have a distinct outer wall, or cell-wall ; all possess a main substance termed cell-substance, or protoplasm ; and all have a central body imbedded in the protoplasm, termed the nucleus. The Cell Substance, or protoplasm of a cell, consists of a fine reticulum, or network, radiating from the nucleus, termed spongioplcmn, and of an apparently structureless substance, which occupies the meshes of the spongioplasm, known as liyaloplasm or ciichylcina. CELLS AND NUCLEL 265 An amcebiform or amoeboid cell consists only of cell-sub- stance enclosing a nucleus ; it possesses the property of spon- taneous movement by means of pseudopodia, like an amoeba, and is capable of ingesting solid and fluid materials, which appear as globules or granules in its interior. Many mature cells have a large quantity of granular material, and others have oil, as, for example, the fat cells, stored in their protoplasm — Cell contents. Such stored material is usually formed by the cell, and is sometimes present in so large a quantity that the cell-substance forms a mere wall around the contents. It is from this form of cell, which was first discovered, that the appellation has been extended to all cells. It has therefore acquired a new and technical meaning. Non-amoeboid cells are usually called ' fixed,' and are frequently surrounded by a distinct film, or cell-wall, either secreted by or formed from the protoplasm of the cell. The cells in Insects are united either by the fusion of their proto- plasm, when the spongioplasm of one cell may be traced into those adjacent to it, or by an intercellular cement material, and rarely by a more copious matrix. Mucoid Degeneration. — ^The part of the protoplasm of a cell which is exposed, either by reason of the cell forming a portion of the surface of the body, or of the lining of the alimentary canal, is sometimes observed to become clear and transparent, and to stain very feebly or not at all. This part of the cell consists of mucin, or of that modification of mucin termed chitin. It is the result of a change in the character of the cell substance itself, and, as this material exhibits none of the vital properties characteristic of proto- plasm, its formation may be regarded as a kind of degenera- tion. The great cells of the fat bodies of Insects are surrounded by a thin pellicle of chitin, which forms the cell-wall. "Whether the mucin or chitin formed by cells is to be re- garded as an excretion or as a modification of the cell substance has long been a matter of speculation. From the changes which occur in the cells of Insects, I have no hesitation in 18—2 266 GENERAL ANATOMY OR HISTOLOGY OF BLOiV-FLY. regarding it as a direct modification of the cell substance, and not as an excretion formed in the cell and poured out on its surface. I shall hereafter recur to this point. The Nucleus. — The appearances presented by nuclei are much more diverse than those exhibited by cells. The nucleus exists in two conditions — the resting stage, which is charac- teristic of a cell which is not undergoing division, and the active nucleus, in the dividing cell. The Resting Nucleus in insects is generally enclosed in a thin capsule, and consists of a clear plasma, nucleoplasm, in which a reticulum, or long coil of fibre, is seen; this fibre is termed the nuclear fibre or nuclear skein. It consists of granules, chromatin granules — so called from their affinity for such dyes as logwood or carmine, nuclear stains, when properly applied (see Appendix to Chap. IV.). The chromatin granules are united by a substance which does not stain — achromatin. Dkscuhtion of Plate XVI. Cells and Nuclei, exhibiting Iheir principal modifications. Fig. I. — A large epithelial cell from the hypoderm of the larva, with striated cell substance, a vacuolated nucleus, and a nuclear nucleolus. Fig. 2.- — Two hypodermal cells from the larva in the resting stage. Fig. 3. — Multinucleate pericardial cells from the young nymph, possibly young fat cells. Fig. 4. — An amceboid cell from the blood of the young imago, loaded with droplets of oil. Fig. 5. — An amoeboid cell from the blood of the young imago. Fig. 6. — An amteboid corpuscle from the same, enclosing a vacuolated nucleus. Fig. 7. — A similar cell with a single droplet of oil in its interior. Fig. 8. — The cell-capsule of an exhausted fat cell from an adult egg-laying female, showing stellate daughter cells. Fio. 9. — A young fat cell from the imago. Flc. 10. — The nucleus of a fat cell from the nymph filled with leucocytes. Fk;. II. — An epithelial cell from the hypodermis of a larva, showing the intra- cellular reticular spongioplasm ; the nucleus enclosing a nuclear nucleolus. Fig. 12.— The nucleus of a salivary cell from the larva, showing the nuclear thread. Fig. 12a. — A portion of the nuclear (chromatin) thread, showing the chromatin beads. Fig. 13. — Goblet cells from the proximal intestine of the imago. Fig. 14.— Secretory epithelial cells, from the proximal intestine of the imago. Fig. 15. — Rodded s.ilivary cells from the sericterial glands of the imago. Fig. 16. — Epithelial cell from the distal intestine of the imago, with a stri.aled border. Fig. 17. — Large cell from the rectal papilla. PLATE XVI. HISIOLOGV. CELLS AND NUCLEL 267 In the resting nucleus the nuclear fibre is sometimes replaced by a reticulum, or even by scattered masses and granules of chromatin. Growth of the Cells of the Larva. — During the larval stage the epithelial cells of the hypodermis and alimentary canal, as well as those of the salivary glands and fat bodies, do not undergo multiplication, cell-division ceasing with the hatching of the egg or soon after. The rapid growth of the larva results from the increase in the size of the cells, so that when full-grown these attain a magnitude which is quite exceptional. Those of the lingual (salivary) glands frequently measure 150" to 200" (t^t inch), and have nuclei from 50" to 60" in diameter. Nuclei of the Salivary Cells. — The nuclei of these gigantic cells are very favourable objects for the study of the resting nucleus. If a larva is opened and immersed for half an hour in Flem- ming's mixture (see p. 94), and then washed for twenty-four hours in 75 per cent, alcohol, the salivary glands can be dissected out, and after staining with Erhlich's hsmatoxylon, the individual cells can be mounted in balsam in the usual way. Such a preparation exhibits the structure of nuclei admirably. Or the cells may be separated after a few minutes' immersion in Flemming's mixture and examined in it; under this treat- ment the nucleus in each cell exhibits a very delicate investing membrane. In some cases I have seen such a nucleus divested accidentally of all ceil protoplasm as a distinct vesicular spheroid, and I have even observed the rupture of the investing membrane, and the escape of the contents — a clear, apparently fluid substance, in which the nuclear threads lie in a more or less dense tangle. The nuclear thread consists of large bead-like chromatin granules, varying from 2" to 4" in diameter, flattened where they are in contact with each other, sometimes exhibiting the appearance of a series of cups and balls. The stained granules are invested by an unstained substance, which unites them with each other. Balbiani [126] described a similar con- dition in the nuclei of the salivary glands of Chironomus, and Viallanes [27, p. i6g] states that Henneguy announced to him 268 GENERAL AXAIOMY OR HISTOLOGY OF BLOW-FLY. that he had observed the same thing in Musca. Eimer* has recorded the appearance of the non-stained thread uniting the chromatin granules. Carnoy [129] describes the nuclear thread of insects as com- posed of numerous minute chromatin granules, united by a clear material, and states that they sometimes surround a central cavity, or, in exceptional cases, form a spiral fibre. I have never seen these appearances, and this is the more remarkable as Carnoy describes them as occurring in the hypodermis of the Blow-fly larva. My failure to do so is probably due to a different method of preparation. When fresh nuclei are examined in Flemming's mixture, or even in normal saline solution.f I have repeatedly seen the nuclear threads divided into transparent discs, which remind me of rouleaux of human blood corpuscles, except that they are smaller and colourless. They are highly refractive. It is a fact worthy of note that, although the nuclear skein is so well developed in the large nuclei of the salivary cells, the fat bodies, and the epidermis, these cells never undergo division, and are destined to be completely destroyed by the action of phagocytes in the first days of the pupa stage (see Chap. IX.). Nucleoli. — It frequently happens that one or two larger, highly-refractive masses are seen suspended in the nucleoplasm ; these are termed nucleoli. Some regard them as mere isolated masses of chromatin ; others, as distinct in nature and chemical composition. I am inclined to think that the history of nucleoli has yet to be written. Some are undoubtedly merely the remains of a degenerated or degenerating nuclear thread. Such masses are sometimes comparatively large and irregular in form ; they are seen in numbers in the degenerated tissues of the pronymph. Others, which are distinguished by Carnoy as nuclear nucleoli, resemble a minute nucleus, and have granules of chromatin in * Eimer, ' Ueber den Bau des Zellenkerns,' Archiv f. mik. Anat., Bd. xiv. t X o"65 solution of common salt. CELLS AND NUCLEI. 269 their interior. Some nucleoli exhibit distinct amoeboid move- ments ; such appear to me to be young nuclear nucleoli. These nucleoli are possibly the nuclei of immigrant phago- cytes, which have found their way into the large nuclei of degenerating cells : compare the nuclei of leucocytes (Fig. 17), which exhibit all the varieties of form ascribed by Carnoy to nuclear nucleoli. In some cases at least this explanation is undoubtedly correct. The Paranucleus (Nebcnkern). — It sometimes happens that a more or less shrivelled, irregular mass of material is found out- side a nucleus. This mass stains readily, and has been termed the paranucleus; such paranuclei are occasionally seen in epithelial cells. The nature of the paranucleus is unknown, but, in some cases at least, I believe I have traced its origin to a degenerated nucleus. After division one nucleus remains, and the other undergoes degeneration. I shall hereafter recur to this subject. The Active Nucleus. — Before and during the division of the nucleus, a phenomenon which always precedes the division of a cell, the nuclear fibre is seen to undergo remarkable changes in its arrangement. These changes are spoken of as karyo- mitosi.s, or karyokinesis. The student will find an account of them in any modern text- book on histology. When these changes occur, the division of the nucleus is said to be indirect. The indirect process of nuclear division is of very wide, if not universal, occurrence in both animals and plants. The old view was that a nucleus undergoes simple fission, a change which is termed the direct mode of nuclear division. Many observers have entirely denied its occurrence since the discovery of indirect division. At one time I thought that the large nuclei of many insect tissues would afford favourable objects for the study of indirect cell-division, but, although the nuclear thread is well developed in them, they unfortunately do not divide. And I have been unable to observe indirect nuclear division in insects satisfac- torily except in the early stages of spermatogenesis. Although 270 GENERAL ANATOMY OR HISTOLOGY OF BLOW-FLY. nuclear spindles have been described in the ova, and I have no doubt of their occurrence, 1 have not observed them in any of my preparations. The cells of the embryo and of the imaginal discs are so small that they are not favourable objects for the observation of karyomitosis. Large multinuclear cells occur, however, in both the larva and imago in the form of cell-chaplets (see pp. 6i and 85), in which the nuclei are undoubtedly undergoing multipli- cation, and in these I have been quite unable to discover nuclear figures. Nor, so far as I know, has anyone ever claimed to have observed the phenomena of karyomitosis in these cells. The direct division of the nucleus has been observed by me in the blood corpuscles, leucocytes, of the larva, but under circum- stances unfavourable for the observance of karyomitosis if it occurs (see page 271) ; so that I regard the occurrence of direct division as an open question, but still hold that it is probable. 2. THE PARABLASTIC TISSUES. The tissues or tissue elements traced to the parablast or mesenchyme (see page 235) are the blood corpuscles, the con- nective reticulum which forms the bed in which the tracheal capillaries lie, the fat bodies, the oinocytes, and the multi- nucleated cell-chaplets. Phagocytes are probably merely modi- fied blood cells, but whether all phagocytes are derived from these, or indeed solely from the parablastic tissues, although probable, cannot be positively asserted. The Blood Corpuscles are amoeboid cells, measuring from 6 to 12" in diameter. They are far more abundant in the resting larva and pronymph than in the imago. In the larva they exhibit resting nuclei, with a single, highly refractive nucleolus and little chromatin. In this stage I have observed the direct division of the cells. The process occupies about half an hour. The nucleus is first withdrawn to one end of the cell (Fig. 17 bis, 3a), and a constriction then occurs in the equator of the cell. This THE PARABLASTJC TISSUES. 271 subsequently disappears and the nucleus returns to its central position. Such changes recur many times, the equatorial con- striction of the cell being more marked with each repetition of the phenomenon. At length the nucleus is seen to be elongated and it ultimately divides, and the two new nuclei rapidly separate ; the constriction then becomes deeper and deeper, and in a few minutes the cell is divided completely by it. The process is precisely similar to that which has been described in the agamic division of the Infusoria. Fig. \T bi$. — Blood corpuscles (/<.•»(■»(-_)'/«) of the. -idult larva: /, living corpuscles, show- ing the ania'boid condition, in d the nucleus is also amoeboid ; _', the same, treated with magenta, showing the various appearances produced by the action of the reagent ; j, a living cell in several stages of direct division (all drawn with ,'„ oil immersion lens). The blood corpuscles of the resting larva, nymph and imago, when fixed by steam and properly stained, sometimes exhibit distinct but irregular nuclear figures — asters, diasters, etc. — and several nuclei are frequently seen in one cell ; hence it might be inferred that karyomitosis occurs in the division of the nucleus, as above described, although the conditions of the observation are such that this could not be seen in the living corpuscle. The method I adopt for the examination of the nuclear figures 272 GENERAL ANATOMY OR HISTOLOGY OE BLOWFLY. of blood corpuscles is Schiifer's.* The blood is spread in a thin layer on the cover-glass, and fixed by holding it in the steam of boiling water, or passing the cover rapidly through the flame of a spirit-lamp. In this way the corpuscles may be fixed in the amoeboid condition. The preparation is then stained by putting a drop of Ehrlich's hsematoxylon on it for a few seconds. It is then washed rapidly with distilled water, and placed in hard water until the stain becomes blue, after which it is mounted in balsam in the usual way. Some of these preparations are very beautiful, and exhibit both the intracellular reticulum, spongioplasm, and the intra- nuclear network, chromatin fibres. The appearances are pre- cisely similar to those seen by Schiiferf in the white corpuscles of the newt. As in the newt, they frequently exhibit pyriform nuclei drawn out into long tails. When the fresh blood of the larva is treated with a solution of magenta, the corpuscles swell up and discharge a granular mass, which leaves a perfectly transparent stroma around the nucleus (Fig. 17, 2, g h i). H. Landoist and others have described crystals in the blood of Insects. Whenever I have observed crystals, I have always been able to trace them to some reagent used in the preparation. Staining with lithia carmine invariably gives rise to crystals of lithia salts. The same author states cor- rectly that the blood plasma contains a globulin which is precipitated by COo, or acetic acid and other reagents. He also remarks that magenta produces the appearance known as ' Robert's macula.' Viallanes devotes a chapter to the con- sideration of the state of the blood of the larva immediately before its metamorphosis [27, pp. 131-136], which he sum- marises in the following manner : ' The cells of the blood of the larva are analogous to the leucocytes of Vertebrates, and are typical embryonic cells.' * Schiifer, E. A., ' On the Structure of Amoeboid Protoplasm,' Proc. Roy. Soc.,vol. xiix., 1891. t Ibid. X Landois, H., ' Krystallisation in Insectenblut,' Zeitsch. f. w. Zool., Ba. xiv. THE PARADLASriC TISSUES , rjl Phagocytes. — It has long been known that amoeboid blood cells arc capable of enclosing molecular material, bacteria, etc., in their protoplasm, in point of fact, of feeding like amoebae. During the histolysis of the larval tissues, the number of white blood cells rapidly increases, and many are seen with several nuclei. They are found surrounding, penetrating, or imbedded in the larval tissue in process of degeneration. These cells ultimate!}' become loaded with globules of oil, and even with solid fragments of muscle. In this condition they were termed by Weismann [2] granule cells {Kornchcnkugdn). He further distinguished large and small granule cells. The large ones Viallanes terms ' corps rosea.' They appear to be merely overgrown and loaded leucocytes (PL XVI., Fig. 4, and PI. XVIII., Fig. 6). These arc very numerous in the blood of the nymph, and some still remain in that of the young imago. The large granule cells are generally seen to be surrounded by young leucocytes, which appear to be feeding upon them, and some may be actually observed pene- trating them by means of pseudopodia. It is clear to me that there is a continual transference of granules from cell to cell, those which are overloaded being attacked by younger and more vigorous leucocytes, which become granule cells, and are in turn themselves attacked by a succeeding generation of phagocytes, until the whole of the effete tissues of the larva are assimilated (see ' Histolysis of the Larval Tissues,' Chap. IX.). The Connective Reticulum resembles the adenoid or retiform tissue of Vertebrates, and consists of a network of stellate branching cells and endothelial plates. It permeates the whole body cavity, and forms the subhypodermic cellular layer and the so-called peritoneal coat of the traches and viscera. The smaller tracheal capillaries are excavated in the stellate cells of this reticulum.* * The intracellular origin of the tracheal capillaries h.is been observed in the larva: of Lepidoptera and Ichneumonidioby Hermann Meyer, Zeitsch.f.w. Zoo!., Bd. i. In Dipterous larva; by Weismann [2]. In Lampyris by Wielowiejski [138]. In the fat bodies of Luciola italica by Emery [139]. Under the 274 GENERAL ANATOMY OR HISTOLOGY OF BLOW- FLY. The best part for the demonstration of this tissue is the thin wall of the great abdominal air sacs, which is covered by a network of flat stellate cells. Tufts of branching cells, also belonging to the reticulum, are found on the terminal branches of the tracheal vessels, and sections exhibit it in the blood sinuses and between the muscles. The evidence on which it is believed that this adenoid reti- culum is developed from the parablast, although not direct in Insects, is on the whole very conclusive. It is precisely similar to the parablastic tissue of the Echinodermata and Coelente- rata described by Hertwig, Kjrotneff, and others. In parts of the body, more especially in the pericardium, it is converted into a true cytogenic tissue, from which amoeboid leucocytes are developed. It is very abundant in the coelom of the nymph, and is apparently formed by the metamorphosis of the leuco- cytes, which abound at this period, and the corresponding tissue in the larva originates from the deep cells which lie beneath the blastoderm, and not from the cells of the coelomic pouches. Lastly, Schaffer [131] has traced the development of leucocytes to the peritoneal coat of the tracheae, and to groups of cells closely related to, and probably merely a part of, the fat bodies. The Fat Bodies of the larva have already been described (p. 85). Those of the imago are subcutaneous groups and strings of multinucleate cells. Three kinds of cells have been described in the fat bodies of many Insects — Fat cells. Intercalated cells {EingcsprcngtcZcUen, Oinocytes), and Cytogenic cells (Blut-heerde). The fat cells of the larva become cytogenic in the pupa, but whether from the immigration of leucocytes which multiply in their interior, or by actual proliferation of their nuclei, is a matter which cannot be definitely settled (see Chap. IX.). In the Blow-fly larva the oinocytes appear to be distinct from the skin of Corethra larvie by Wielowiejski, Zool. Anzeig. 6 Jahr 9. A typical example of the formaiion of intracellular canals is seen in the development of the capillaries of Vertebrates, and the tracheal capillaries of Insects are developed in a precisely similar manner. THE PARABLASTIC TISSUES. 275 fat bodies, nor are the true cytogenic cells connected with them, as they are according to Schaffer [131] in the larvae of Lepidoptera. The multinuclear cell chaplets, which occur in both the larva and imago, are closely related to the fat bodies. Those of the imago at least appear to be young fat bodies, and the oinocytes of the imago are probably fat cells, the fat granules of which have been absorbed for the nutrition of the great food-yelks (eggs) in the mature female. Recently the origin of the fat bodies in Insects has attracted considerable attention, owing to its import in relation to the parablast theory. C. Schaffer has, I think, satisfactorily traced the development of the larval fat bodies in the Blow-fly to the peritoneal coat of the tracheal vessels, but both he and Graber also trace them in various insects to the hypodermic layer of the integument. I think, however, the figures given by these authors indicate that they had not to do with young fat bodies, but with tangental sections of the hypodermis. Both authors regard the cellular layer of the traches as also of hypo- dermic origin, and as due to invagination, a view which I regard as untenable (see Tracheal System). The Fat Cells of the Imago differ from those of the larva (p. 86) in being frequently multinucleate. Butschli [126, p. 558, PI. xxvii.. Fig. 43], Claus,* and BoIIes Leet concur in deriv- ing them from multinucleate cell-chaplets similar to that described by Weismann in the larva ; and my own researches lead me to the same conclusion, except that the great cell- chaplet of the larva certainly undergoes histolysis in the pupa, although other and similar chains of multinucleate cells appear in the young imago, and ultimately become f^xt bodies. The fat cells of the imago are loaded with fat granules, and appear very similar to those of the larva, except that they are smaller. These cells usually exhibit a single vesicular nucleus. In the immature state they are mixed with smaller cells, * 'Zeitsch. f. w. Zool.,' Bd. xxv., p. 266, PI. XIV. t 'Kecueil Zool. Suisse.,' torn, ii., p. 391. 276 GENERAL ANATOMY OR HISTOLOGY OE BLOW-ELY. some with two or four nuclei, in which fat granules are abun- dant. These are the proliferating cells from which the fat bodies are developed. In the mature Insect, cell-capsules (mother cells), containing hardly any fat granules, or entirely exhausted, are seen amongst the fat cells, and contain a reticulum of stellate nucleated cells, closely resembling the segmentation cells of the food-yelk (Fl. XVI., Fig. 8). Intercalated Cells. — The fat bodies of many Insects have peculiar pigmented cells scattered amongst the fat cells. Graber describes these as containing yellow or green granules and rod-like crystals, and he regards them as imperfectly developed fat cells. I have never found such cells in either the larva or imago of the Blow fly, unless the cell-capsules of exhausted fat cells (mother cells), which contain the stellate cells described above, are their representative. If this be so, Graber's ' eingesprengte Zellen ' are more probably exhausted than imperfectly developed fat cells. I have not had an oppor- tunity of examining the insects in which intercalated cells arc described by Graber and others. Oinocytes. — This term is applied by Wielowiejski to certain chains of cells which occur — a pair in each segment — imme- diately beneath the integument in the larvae of some Diptera, which he regards as representing the intercalated cells of the fat bodies. This interpretation of their nature is, I believe, incorrect. Nor is the term ' oinocyte ' a good one, as it has also been used for the intercalated cells of the fat bodies. Wielowiejski says : ' The oinocytes in the Brachycerae re- semble those of Chironomus, and are situated in the same position, at the sides of the abdominal segments.' He The reader may consult the following papers on the origin and nature of Oinocytes (Wielowiejski), and Intercalated cells, Schaltzellen (Leydig) : 137. Graber, V., ' Uebcr den propulsatorischen Apparat. der Insecten.' Archiv f. mikr. Anat., Bd. xi. 138. WiEU)wn.jsKi, H. RrnER von, ' Studien iibcr Lampyriden.' Zeitsch. t. w. Zool., Bd. xxxvii., i88i. 139. Emery, ' Ueber die Leuchtorgane von Lampyt-is italica.' Zeitsch. i. w. Zool., Bd. xl., 1884. EPITHELIA. 277 describes these as five extraordinarily large cells, four of which lie close together, commonly forming a rectangular cord, whilst the fifth is separated from the rest by a cell's diameter, and always contains two nuclei; the others have only one in each cell. This description may be applied to the same organs in the larva of Musca, except that all the cells usually have twa nuclei or more, and I am by no means clear that the number is constant. They undergo complete histolysis in the young nymph. I am greatly tempted to regard these remarkable structures as the remains of segmental organs, similar to those of the Chaetopods and Peripatus. Although their developmental history is unknown, Graber gives a figure [114, Fig. 128] of ' the lateral body cavity ' of Stenobothrus variabilis, in which the cavity, lb, W, is apparently a segmental organ, and the cells, o', are, I suspect, Wiclowiejski's oinocytes and represent the remains of a segmental tube. 3. EPITHELIA. Epithelial Cells differ from those of the parablast in being developed directly from the outer or inner layers of the blasto- derm, and in not exhibiting an amoeboid stage. The individual cells are generally united by a scanty intercellular substance, termed cement substance. It is usual to distinguish the several varieties by their form, as tabular, or flattened, spheroidal and columnar cells. An epithelium may consist of a single layer of cells, when it is termed simple, or of several layers, when it is said to be stratified. Many epithelial ceils possess cilia, but such are never found in the Arthropoda. The epithelial cells of Insects may be classed as chitino- genic, secreting, and absorbent cells, on the one hand, and sensory cells on the other. The cells derived from the epiblast are chiefly chitinogenic ;. those from the hypoblast arc secreting or absorbent. They 278 GENERAL ANA TOM Y OR HISTOLOG V OF BLO W-FL Y. may be further subdivided into mucigenic and serous cells. The sensory epithelial cells are all epiblastic in origin and usually chitinogenic. The epithelial cells of the embryo are distinguished by their small size and their arrangement in several layers. As the cells of these layers interlock, they cannot be termed stratified in the strict sense of the term ; they may be termed tran- sitional, according to the usual nomenclature, but it scarcely indicates their nature. I prefer to term such epithelia embryonic. In the Blow-fly, at least, this variety is only seen in the embryo, in the imaginal discs, and young nymph. The epithelia of the larva are remarkable for the rapid growth of the individual cells, which, as already mentioned, attain extra- ordinary dimensions in the adult larva. The epithelial cells of the imago are very variable in size, but are usually far smaller than those of the larva. Embryonic Epithelia. — The cells are very small, rarely exceed- ing 5" to 5" in diameter, and are multiplied with great rapidity, probably by indirect cell division, although, from their small size, it is difficult to demonstrate nuclear figures. The appearance of the nuclei is, however, indicative of karyo- kinetic changes. The cells are cubical, columnar, or fusiform, exhibiting one or more thin cuticular laminse on their free sur- faces. They are glued together by a firm intercellular cement material. Chitinogenic Cells (hypodermic cells).— The cells which lie beneath the cuticular epidermis of the skin, and those which form the subcuticular layer of the stomodaeum and procto- djeum, are the principal chitinogenic cells. The cells beneath the cuticular skeleton are properly termed the hypoderm. They are either flattened or columnar. Those of the larva have already been described (p. 37). In the imago the hypoderm cells are far smaller, and are either columnar or tesselated. The latter are usually found under the transparent syndes- moses, where no muscles are inserted. The cells into which the muscles are inserted are columnar. Those at the bases of the setae are spheroidal, and large in proportion to the size of E PIT HE LI A. 279 the setae ; some at the base of the larj^est are gigantic. The cells which support the setas are termed trichogenic. Many of the epithelial cells beneath the cuticular skeleton disappear entirely in the adult imago, others persist, and some of these are pigmented with orange-brown or red pigment granules. The cells beneath the cuticular lining of the stomodseum and rectum in the imago are tesselated, except those of the rectal papilla;, which are large, cylindrical, and probably glandular (see Rectal Papillae). Mucigenic or Muciparous Cells are found in great numbers in the metenteron. They resemble the goblet cells of Mammals (PI. XVI., Fig. 13). The absorbent cells of the alimentary canal are probably all mucigenic. They exhibit a more or less marked basilar border, which in many Insects is so distinctly fibrillated that it re- sembles a mass of cilia. This condition is well seen in some sections of the intestine of the immature imago of the Blow- fly, in which the basilar border splits into tufts, so closely resembling tufts of cilia that they might easily be mistaken for them, except that, when the living cells are examined, there is no movement. In the adult insect the basilar border is much thinner (see Alimentary Canal). Serous Gland Cells are seen both on the free surface of the alimentary canal and in the salivary tubules in the imago. They are distinctly striated like the cells of the pancreas, or rodded like those of the kidney in Vertebrates. The cells of the Malpighian tubules contain various granular substances, oil and pigments, like those of the vertebrate liver (see Malpighian Tubes). The cell-substance also frequently exhibits intracellular channels. There are also small convoluted subcutaneous wax glands (see Tracheal System), the cells of which are very small, but loaded with fatty matter like those of sebaceous glands. Cell-fibrillation. — The cells, which intervene between the muscles and the cuticular integument, of the long columnar variety, are often palisade-like and undergo longitudinal fibril- 19 28o GENERAL ANATOMY OR HISTOLOGY OF BLOW-FLY. lation, forming a kind of tendon. This tibrillation commences next the muscle, and eventually nothing is left of the original cell but a stellate core of protoplasm enclosing the nucleus, surrounded by tendinous fibres. A similar transformation of the cells of an epithelium into fibrous tissue occurs in the proventriculus, but the fibres are much branched, and form a close meshwork comparable with that of the elastic cartilage of Vertebrates, e.xcept in the extreme fineness of the fibres, which more closely resemble those of white fibro-cartilage (see Proventriculus). Development of the Integument of the Imago. — The epiblast of the inuigiual discs is differentiated into two layers, a superficial one of small and a deep one of large cells. The larger epiblast cells from which the setae are developed remain as the hypoderm, but the great majority of the cells of the superficial layer become chitinized throughout their entire substance, and ulti- mately so intimately united that all trace of their original limits is lost. They are transformed into cuticle, and form the super- ficial part of the exo-skeleton. As early as 1870 I pointed out that the epiostracoid layer, which I then termed the protoderm, is cellular in character. From more recent investigations, I have now no doubt I was correct, for the examination of a series of nymphs at the stage of development in which the seta; are being formed cannot fail to convince the observer that the cuticular external layer is formed from the epiblastic cells, not by the secretion of cuticular lamella;, but by the actual conversion of the cells into a skeletal epidermis, under which a second layer of cells is developed, the true hypo- derm is of the imago, chiefly by the extension of the trichogenic cells of the epiblast beneath the smaller cells of the outer layer. The Cuticular Structures. — The cuticular epidermis of the imago does not differ materially from that of the larva (see p. 9). Where the sclerites are dense, however, the laminated structure is less apparent or entirely absent. Such parts are also opaque and deeply coloured. Wherever the cuticular integument remains transparent, it is distinctly laminated, and EP I THE LI A 28 1 may even exhibit lines vertical to its surfaces, corresponding with the divisions between the subjacent cells. Whether the cuticular layers of the skin, the dermal skeleton, is to be considered, as Leydig held, as an indurated exudation from the cells, or as a modification of the cell-substance, is a question which is no longer open to doubt. The direct con- version of the epiblastic cells of the nymph into cuticular tissue is perfectly obvious (see Development of the Nymph), and I suspect that the laminated endostracoid layers are also formed in the same way, as well as the internal cuticular membranes. This view is further supported by the gradual thinning ot the hypodermal cells as the cuticular la3'ers are developed, and by the eventual disappearance of the bypoderm beneath the harder sclerites of the adult imago ; by the manifest fusion of the bases of the cells and the delamination of the basement cuticle of the sericteria of the larva ; by the sculpturing of the surface of the epidermis, and by the presence of minute and often complex setse on its surface. The formation of the branching setae on the surface of the membranous parts of the proboscis is quite inexplicable on the hypothesis that the cuticular layers are formed by a fluid or semi-fluid excretion from the cells. The laminated structure of the cuticle is, it is true, sometimes unbroken by any indication of areas corresponding with those of the sub- jacent cells, but this is sufficiently explained by the manifest fusion of their bases, which precedes the process of chitini- zation. Leydig regarded the epidermis as a fibrillated matrix, but, as far as I know, there are no indications of fibrillation in the dermal cuticle of Insects. In the Crayfish, however, the hypodermis is said to be fibrillated, but this fact is no argument against its cellular origin, as a similar fibrillation occurs in the wall of the proventriculus of the imago of the Blow-fly, and is undoubtedly the result of the fibrillation of the extremities of the epithelial cells of which it is composed. In the larv;e of some Diptera the cuticle is divided distinctly 19—2 282 GENERAL ANATOMY OR HISTOLOGY OF BLOW-FLY. into fields corresponding with the cells from which it is developed, and this occurs when the latter are united by an intercellular cement. The existence of this cement-substance can be demonstrated by staining with silver nitrate. Pore Canals frequently exist in the cuticle. The largest pass into the hollow setae and transmit a process from a trichogenic cell ; smaller, frequently branching, canals pass into papillae, or are lost in the thicker parts of the cuticle. All contain un- differentiated processes of the cells, and may be compared with the dentine tubes of a tooth or the canaliculi of bone. The cuticular lining of the stomodaeum and proctodaeum are thin and laminated ; that of the rectum has papilla-like teeth and ridges on its surface in the vicinity of the recto-metenteric valve, and hair-like processes over the rectal papilla". Cuti- cular membranes also line the tracheae and invest the indi- vidual fat cells. The basement membranes, on which some epithelia rest, must also be regarded as cuticular lamelhe. 4. THE MUSCLES AND NERVES, a. The Muscles. The Somatic Muscles of Insects exhibit three varieties, which Weisniauu [124J distinguished as larval muscle, leg muscle, and wing muscle. Although in their ultimate structure these three forms of tissue are precisely similar, the manner in which the ultimate elements are arranged to form fibres differs, so that except when examined with adequate magnifying power the general appearance of each is very dissimilar from that of the others. In the larva each muscle consists of one or more large fibres. Each fibre has a distinct muscle - sheath, sarco- lemma, and nuclei are seen at intervals between the sarco- lemma and the muscle-substance. The muscle-substance exhibits distinct transverse striation, and, as Weismann pointed out, the fibres are precisely similar to those of Verte- THE MUSCLES AND NERVES. 283 brates. Except in being colourless, they closely resemble the pale muscle fibres of Mammals. The ordinary or leg muscles in the imago exhibit no proper sarcolemma, and have their nuclei arranged in the centre of the fibres. Their transverse striae are more marked than in the muscles of the larva, and the fibres frequently exhibit the appearance of being divided into ' muscle cases' by transverse septa — ' Krause's membranes.' Each nmsr.le consists of one or many fibres, either inserted in a bundle into a common apodeme, or on one or both sides of the apodeme in a penniform or bipenniform manner. More rarely the fibres are inserted into the integument through the medium of fibrillated hypodermal cells, which closely resemble tendons (PI. XVII., Fig. 5, b). The wing muscles, sternodorsales and dorsalcs, are softer than the ordinary muscles. Each consists of several giant fibres without any sarcolemma, but surrounded by tracheal vessels and epithelioid cells. Each fibre consists of numerous fasciculi of fibrilh-e, separated from each other by numerous nuclei. The fibres are easily broken up into their constituent fibrillae, and only exhibit the faintest indications of transverse striation, Leydig [123] described the wing muscles of Dytiscus as yellow, possessing but little solidity, and as very easily separated into their constituent fibrilia;. In the fresh condition the wing- muscles of the Blow-fly are so soft that they appear almost semi-fluid. They have a bluish-gray colour. The Visceral Muscles of the Blow-fly are distinctly striated ; those of the alimentary canal consist of fusiform (often branched) flattened cells, with a single large ovoid nucleus embedded in each cell. These cells are united into fenestrated laminae by an intercellular cement substance, and, except that they exhibit very distinct transver.se stride, they bear a close resemblance to non-striated muscle fibre. The muscle fibres of the dorsal vessel differ entirely from those of the alimentary canal. The dorsal vessel consists of a muscular tube formed of very fine transversely striated fibrilla, which branch and form a longitudinal network around the 284 GENERAL ANATOMY OR HISTOLOGY OF BLOW-FLY. whole tube. Nuclei are seen at very regular intervals, one on either side of the dorsal vessel. According to Viallanes [27], it consists in Eristalis of a series of segments, each containing a pair of nuclei, one on either side. The segments are united by fine lines of cement material, which is stained by nitrate of silver. I have been unable to obtain definite results with silver nitrate either in the larva or imago of the Blow-fiy, but I think it probable that the dorsal vessel consists of Description of Plate XVII. The Muscular and Nervous Tissues of the Fly. FiC. I.— Sarcostyles : a, from the ordinary muscles ; b, the same when stretched ; c, appearances due to the overlapping of the sarcostyles when stretched ; d, a portion of a muscle fibril, exhihiting longitudinal stria? and beads due to the overlapping of the sarcostyles ; c, a fibril showing the so-called muscle-caset ; /, sarcostyle from one of the great thoracic (wing) muscles ; g, another sarcostyle from the same, showing a granular appearance, probably due to post-mortem change (all seen with a ,'j oil immersion lens). Fig. 2.— a transverse section of a portion of one of the fasciculi of the great dorsal muscle : « h, nerve fibres ending in the muscle. Fig. 3.— a longitudinal section of a fasciculus of the same. Fig. 4. — Two stages in the development of the great dorsal muscle : a, an early stage, showing two rows of muscle cells, turn, and a group of parablast cells, />; *, a later stage, in which the fibrillation of the muscle has commenced (ihe muscle cells are fused) ; it, muscle nuclei ; k, nuclei of the muscle-sheath derived from the parablast dV oil immersion). Fig. 5. — Tendon-like terminations of the muscles : 5a, connective reticulum between the tergum and the young dorsalis muscle, from a nymph on the fourth day of the pupa stage (,V oil immersion); j/', tendon-like insertion of an ordinary muscle from an immature imago. Fig. 6.— Transverse sections of ordinary muscle fibres near their insertion into the integument, showing a concentric arrangement of the muscle substance {-^ oil immersion). Fig. 7.— Karyokinetic figures and cell-plates in an ordinary muscle fibre, from a nymph on the twelfth day of the pupa state. The isotropous substance is seen to form broad bands between the cells {^^ oil immersion). Fig. 8.— a. Transverse .sections of ordinary muscle fibres of the adult imago, showing nerve-ternnnals ; b, a more highly magnified representation of the nerve-terminal (iV oil immersion). Fig. 9.— a transverse section of the labial nerve of the proboscis, with a tracheal vessel attached to its sheath dV oil immersion). Fig. 10.— a similar nerve seen in the longitudinal section, showing the ganglion cells. Fig. II. — A ganglion cell from the thoracic ganglion. Fig. 12.— Two small ganglion cells from the optic ganglion. Fig. 13.— a sensory seta showing the trichogenic and ganglion cells at its base. A process from the ganglion cell apparently ends in the nucleus of the trichogenic cell. PLATE XVII. B5 I' ISS'I d Ir d C C f\ -<^»N 51. HISTOLOGY. THE MUSCLES AND NERVES. 285 such united muscle segments, an idea first suggested by Jaworowski.* The Ultimate Structure of all these forms of muscle exhibits a perfect uniformity of type. It is not my intention to give even a resume of the very various accounts of the minute structure of muscle which are still under discussion. For details on this subject the reader may refer to the tenth edition of Quain's ' Elements of Anatomy,' in which he will find a good biblio- graphy of the recent literature on the subject. The only memoir which agrees with my own observations is by Biitschli and Schewiakoff [134], and their results so completely accord with my own, that I shall give a resiuiu' of them. According to these authors, each fibre, muscle cell, or bundle; of primitive fibrillse, consists of two varieties of protoplasm — a contractile fibrillar substance and ordinary protoplasm, inter- mediate substance. The contractile fibrils, or primitive fibrillce, are fine fibres with their long axes parallel to the long axis of the fibre. In transverse sections they appear prismatic, highly refractive dots, surrounded by the less refractive protoplasm, the intermediate substance or sarcoglia of Kiihne. The trans- verse sections of the fibrilla; give rise to the appearance known as ' fields of Cohnheim.' The intermediate substance, sarcoglia or sarcoplasm, is distinctly reticular, but the reticulation is very irregular. The muscle nuclei are imbedded in this sub- stance. The saixolemma may be either a distinct membrane, as in the larval muscles, or merely a more or less condensed layer of reticular sarcoplasmic fibres. Each contractile fibril is formed of two kinds of material, one anisotropous, the other isotropous. The anisotropous sub- stance consists of prismatic or strap-shaped segments ; these are united in a linear series by the isotropous substance, which they regard as a cement material. The authors quoted state that the primitive fibril, contractile substance, is also reticular. Their figures are, however, more readily understood than their description, and it appears that the prisms of anisotropous substance are formed by still * Siizungbericht, K., Acad. Wien, Bd. Ixxx., 1879. 286 GENERA r. ANATOMY OR HISTOLOGY OF BLOW-FLY. smaller elements united together. The anisotropous substance is readily stained, the isotropous remains unstained, by carmine and aniline stains. I shall use the nomenclature adopted by Schiifer [136], and term the primitive iibrillae sarcostyles. Each sarcostyle (PI. XVII., Fig. I) consists of a series of sarcomeres, anisotro- pous substance, cemented together by a cement material, isotropous substance. A fibre consists of a number of sarco- styles, imbedded in a more or less abundant intermediate substance, in which the muscle nuclei lie. This I regard as the undifferentiated reticular protoplasm of the muscle. With regard to the alleged structure of the sarcomeres, I have made many preparations from the wing muscles, in which they are most easily isolated, and it appears to me that each consists of a highly refractive flat or prismatic rod, in whicii I have sought in vain for evidence of more minute structure. It may be, however, that the means at my disposal are inadequate for any further resolution of structure, and as in every other point my observations entirely agree with those of Biitschli and Schewiakoff, I only wish to record the fact that I have been unable to resolve the sarcomeres, as they have done, into more minute constituent rods. My rod- like elements, saixomeres, measure "ooi mm. in diameter, and from '002 to 'ooG mm. (2" to 6") in length, whilst the ulti- mate elements described by the authors named are only '0006 to "oooSmm. in length. I3y teasing the wing muscles in Flemming's mixture, it is quite easy to isolate the sarcostyles, and it is only a little more difficult to attain the same results from the muscles of the larva. The ordinary muscles of the imago cannot be so separated, but the thinnest sections indicate that they only differ from the other forms of muscle in the small quantity of interstitial substance, and the greater density of the reticulum, •which plays the part of a sarcolemma. When the muscle fibrillae are stretched, the sarcomeres become narrow in their equator, and they then appear strap- shaped (PI. XVII., Fig \,b). THE MUSCLES AND NERVES. 287 Most of the appearances which have been described in striated muscle fibres are, I believe, diffraction phenomena arising from the super-position of many layers of sarcomeres ; others arise from the coagulation of the sarcoplasm, or inter- mediate substance. To the first I attribute the bright dots and dark transverse lines, and to the latter the cracks which are seen in transverse sections, and the reticulated appearances described by C. F. Marshall in specimens prepared with gold chloride.* The development of the ordinary Skeletal Muscles, Meg muscles,' of the imago is readily observed in the nymph and immature imago, and throws much light on the nature of the muscle fibrilic'E. In the youngest state each fibre appears as a row of cells placed end to end. In some cases the cells are fused together into a multinucleated protoplasmic cord ; there is at this period no trace of transverse striation. At a later stage the nuclei exhibit karyokinetic figures, and divide in a plane transverse to the fibre. A bright line, a cell plate, appears between the two demiasters into which the nucleus separates (PI. XVII., Fig. 7). The fibre increases in breadth during this process, but the distance between the cell plates diminishes with each division. Fibrilla; next appear at the periphery of the fibre. The cell plates are Krause's mem- branes, and form the isotropous cement material of the fibrilla;, the cell substance between the plates being differ- entiated into sarcomeres ; a portion remains undifferentiated in the centre of the fibre around the nuclei. The superficial portion of the cell forms a kind of sarco- lemma, which is more firmly attached to the cell plates than to the intervening material. The bulging of the peripheral portion of the fibre from the imbibition of fluid, formerly relied on as demonstrating that the membranes of Krause are true septa, is due to this fact. The membranes of Krause, mj' cell plates, are not per- manent ; they disappear as plates, and are only represented by the isotropous material of the sarcostyles in the fully formed muscle. * Quart. Joiirn. of Miciosc. Science, vol. xxviii., i8!>8. 288 GENERAL ANATOMY OR HISTOLOGY OF BLOW-FLY. Many of the muscle fibres exhibit two or three concentric layers of fibrilla;, especially at the ends of the fibres (PL XVII., Fig. 6), where these are separated by tendinous tissue derived from the hypodermis, into which they are inserted. The larval muscles are also developed from rows of cells, and are at first non-striated. I have not succeeded in seeing the manner in which the fibre is subsequently converted into bundles of sarcostyles. The wing muscles are developed from rows of muscle cells, but these are at first imbedded in a mass of parablast (PI. XVII., Fig 4, fl). During their development the nuclei of the muscle fibre disappear ; but those of the parablast cells remain between the fasciculi. The origin and develop- ment of the wing muscles will be further considered in the next chapter. b. The Nerves.* The Structure of the Peripheral Nerves. — My observations on the structure of the larger nerve trunks, from the centres to their primary divisions, agree with the description given by Waldeyer. Each nerve trunk is surrounded by a nucleated sheath (PI. XVII., Fig. lo) continuous with the capsule en- closing the ganglion from which it arises. The sheath is sub- divided by branching longitudinal septa. The spaces enclosed by these septa contain fasciculi of exceedingly fine nerve fibrils, which in transverse sections appear as dark specks surrounded by a granular fiuid. The whole closely resembles a gray or sympathetic nerve from a Vertebrate, consisting of several fasciculi of fibrillae. It may be described as a number of fine axis cylinders surrounded by plasma and enclosed in a connec- tive-tissue sheath. Viallanes [27] , I believe, first noticed that the nerves of Insects contain nuclei at intervals, chiefly at the angles of * For a general description of the nerve centres, see p. 66. A more detailed account will be given in the chapters devoted to the nervous and sensory organs. THE MUSCLES AND NERVES. 289 bifurcation, distinct from and larger than the nuclei of the sheath. These he named ' axis cylinder nuclei.' I have fre- quently been able to demonstrate the fact that these are really the nuclei of fusiform bipolar ganglion cells. On the proximal side of each cell the process is an axis cylinder, but on the distal side it is a tubular nerve, which has some resemblance to a' medullated nerve fibre (PI. XVII., Figs. 9 and 10). Such large fibres are seen in transverse sections of the branches of nerves surrounded by fine axis cylinders. Although the larger tubular fibres are blackened by osmic acid, they have far less fatty matter in their composition than the medullated nerves of Vertebrates. It appears to me that the proximal fibres should be regarded as nerve-root fibres, belonging properly to the central ganglia, whilst the larger distal fibres are true nerve fibres. At each bifurcation of the nerve one or more ganglion cells occur, so that the number of larger fibres increases. I regard the cells which appear at the angles of bifurcation as trophic nerve elements, similar to those of the ganglia on the posterior spinal nerve- roots of Vertebrates. Whether all the nerve fibres, or only the sensory ones, pass through these ganglion cells is a matter on which I am still doubtful ; but those which do, terminate in peripheral ganglion cells, the branching processes of which again assume the form of simple axis cylinders. I have not observed any such terminal ganglion corpuscles on the motor nerve fibrils which end in the muscle fibres. Hence I think it probable that all such large tubular nerve fibres are sensory in function. Although I have compared the large nerve tubules with the medullated fibres of Vertebrates, I have never been able to find any indications of ' nodes of Ranvier,' or the rodded structure characteristic of the latter. The nerve tubules described by Weismann and others correspond with the large nerve fibres. Motor End Organs.— I have sought in vain for well-marked motorial end plates both in the larva and the imago of the Blow-fly. The nearest approach to such organs are small, more 290 GENERAL ANA TOM Y OR HISTOLOG V OF BLO IV-FL Y. or less triangular specks of protoplasm, which is readily stained by carmine, immediately beneath the sarcolemma of the muscle fibre at the point at which the nerve enters the latter (PI. XVII., Figs. 2 and 8). I have found it exceedingly diiBcult in my sections to demonstrate the nerve terminals in muscle at all, but in favourable sections have found a fine plexus of fibrils between the muscle fibres, from which branches pass into the muscle substance. In the imago the individual nerve fibrils are less than i" in diameter, but in the larva they are twice as large. In some sections it appears that the terminal nerve fibrils are connected with the transparent isotropous substance, but on this point I am by no means convinced, although the observation is in accord with the statements of Engelmann and Foetlinger. The well-marked motorial end plates, described and figured by Viallanes from the larva of Stratiomys and Tipula, appear to me to be the connective cells which exist on the terminal branches of tracheal capillaries, and I strongly suspect that the so-called motor end plates described in various insects are in reality the terminal cells of tracheal vessels, which are readily mistaken for nerves. Sensory Narve Terminals. — The cutaneous nerves either end in ganglion cells or in special sensory organs. Their terminal branches are much larger than those of the motor nerves, and are distinctly tubular. Occasionally a distinct axis cylinder, which stains deeply with logwood or carmine, appears in transverse sections as a central point in the nerve cylinder. The terminal ganglion cells on the sensory nerves frequently give off a process which enters a trichogenic cell, and either passes to its nucleus (PI. XVII., Fig. 13), or ends in a granular crescent on one side of the cell. These crescents are precisely similar to the menisci or tactile discs described by Ranvier in the pig's snout. The terminal ganglion cells also give off pro- cesses which penetrate the hypodermis and end in pore canals in the cuticle. Such nerve-endings are readily demonstrated in the larva. Some of the cutaneous nerves of the larva, instead of ending in a single large ganglion corpuscle, have a group of small gan- THE MUSCLES AND NERVES. 291 glion cells near their termination (Fig. 12 his, j), and end in demilunes of granular protoplasm on the inner surface of the hypodermis. Beside the end organs of the special senses, which will be described hereafter, many of the cutaneous nerves termi- nate in remarkable fusiform bi-polar cells. One or several such cells are enclosed within a capsular prolongation of the nervc- sheath. The central pole of each cell is continuous with a tubular nerve fibre ; the peripheral pole is prolonged as a highly refractive cylindrical process, which lies in the axis of one of Fifi. 12 its. — /. A section of the terminal joint of the Maxilla, showing the eye lilie organs ; 2. A section of the eye like organ (^\ oil immersion lens) ; j. Endings of a nerve in the hypodermis, showing a peripheral ganglion ('1} cotes dc mdon,' Viallanes)— all from the lilow-fly larva. the transparent moderate-sized setae so abundant on certain parts of the insect. Such sensory seta; are numerous on the prosternum and the lips of the proboscis. They are pro- bably organs of touch. Many of these encapsulated nerve-end organs contain peripheral cells, which are smaller than the central fusiform cells, and are not apparently connected with the nerve. They are perhaps concerned in the nutrition of the end organs, and resemble the outer cells of the taste buds of Vertebrates. CHAPTER IX. ON THE DEVELOPMENT OF THE NYMPH IN THE PUPA. Dr. Weismann's great work [2] on the after-development of the Muscidae was the first memoir on the subject which can be regarded in the light of a scientific exposition. When we remember the primitive modes of histological research which were then in use, and the unsettled state of the opinions entertained as to the manner in which the tissues of the animal body are developed, we must regard this memoir of Weismann's as one of the most wonderful records, if not the most won- derful record, of brilliant discovery which has appeared in Bibliography. 140. Heroi.d, M., ' Eiitwickelungsgeschichte der Schmeuerlinge,' plates, 2 voN., 4t()., Cassel, i8i v 141. Weismann, a., 'Die Kntstehung des vollendeten Insects in der Larve und Puppc' Abh. Senckenb. Natur. Gesellsch., Bd. iv., 1863. 142. LowNE, B. T., ' Notes on the Development of the Nervous System of the Annulosa,' Monthly Microsc. Journal, January, 1871, pp. 260, 26r. 143. Metschnikoff, E., ' Untersuchungen iiber die intracellulare Ver- dauung bei wirbellosen Thieien.' Arbeiten aus dem Zool. Inst, zu Wien, Bd. v., 1883. 144. Bakfurth, D., 'Die Kiickbildung des Froschlarvenshvvanzes.' Archiv f. Mik. Anat., Bd. xxix., 1887. 145. KOWALEVSKI, A., 'Beitriige zur Nachenibryonalen Entvvickelung der Musciden.' Zeitsch. f. w. Zool., Bd. xlv., 1887. 146. .See aUo Zool. Anzeig., Bd. viii., pp. 98, 123, and 153, 1885. 147. Rkes, J. VAN., ' Beitriige zur Kenntniss der inneien Metamorphose von Musca Vomitoria.' Sprengel's Zool. Jalirbuch, lid. iii., Abth. f. Anat. und Phy., 1889. Perhaps the most important paper on the subject since those of Weismann [2] and Ganin [34J. THE DEVELOPMENT OF THE NYMPH. 293 the whole range of zoological science in the present half of the nineteenth century. The only observations of importance made previously to the appearance of Weismann's exposition occur in De Reaumur's seventh memoir in the fourth volume of his ' History of Insects '; and as these appear to me to possess a very high interest in relation to the most recent views propounded by Van Rees, I shall give an abridged translation of Reaumur's statements. He says : ' I have spoken elsewhere of those worms which are nourished in the intestines of the horse, and which only leave them when they are about to be transformed into pupte. These insects remain within the pupa -shell much longer than the Flesh flies, and are much longer in becoming flies. I opened the shell of some eight days after the transformation, and was able to withdraw the whole insect from the shell without injury. In these I could observe neither wings nor legs, nor any of the parts proper to a nymph; they presented the appearance of elongated ovoids. . . . Probably all those flies which form the pupa-shell from the larval skin undergo this metamorphosis previously to becoming nymphs. I term it the spheroidal or ellipsoidal metamorphosis.' He continues : ' With much address and patience, one may convince himself that this occurs in the Flesh flies,' and ' the worms therefore pass through a metamorphosis which is addi- tional to that which caterpillars and the larvae of most of the four-winged flies undergo.' If we open the pupa-case of one of the Flesh flies five or six days after the transformation of the worm, we find a well-formed white nymph, provided with all the parts of a fly; the legs and wings, although enclosed in sheaths, are very distinct. The sheaths are thin, and conceal nothing. The proboscis of the fly is seen lying upon the corselet and the lips, the aguillon and its sheath are distinct ; the head is large and well developed, and the compound eyes are very recognisable. But how has our insect quitted its second form, the spheroid stage, to take this third form, that of the nymph ? 294 i'HE DEVELOPMENT OF THE NYMPH. ' There is nothing more easy than to get a number of fly pup«, and, after boiHng them, to remove the pupa-case. In this way we can watch the progress of events. At the end of two or three days the very short legs project at the anterior extremity, and on the next day the wings appear ; in another day the extremity of the proboscis, and afterwards the whole organ becomes apparent. The head first shows itself in those nymphs in which the legs have almost reached the posterior extremity of the abdomen. ' I have recognised that it is not that the parts grow day by day, as the appearances would lead us to believe, but that they exist preformed, and the mechanism of their evolution is very simple. ' I have already spoken of a cavity seen at the an- terior extremity of the ovoid, which contains the exuviated hooks and darts (mouth armature) of the larva. I have observed a similar cavity at this stage in all insects which pass through the ovoid condition, and there is a little horn bearing a stigma at two opposite points of its margin. I conclude that these stigmata belong to the corselet of the fly, and the parts which appear day by day are really en- sheathed in the cavity at the anterior extremity of the ovoid nymph. ' In order to prove that this is really the case, I press upon an insect in the ovoid stage in which the extremities of the feet only are seen, and succeed at length in suddenly producing a nymph. I achieve by pressure a result which requires several days for its accomplishment in the natural state of things. ' The most essential parts of the nymph and the imago, the head, wings and legs, are therefore lodged in the body-cavity of the worm before its first transformation, each enclosed in its envelope; for when they appear each is so enclosed. It is as if all the parts were invaginated, like the fingers of a glove withdrawn into the hand.' No doubt Reaumur believed, from the analogy of what he knew to be the case in the Lepidnptera, that the limbs— wings, THE DEVELOPMENT OF THE NYMPJf. 295 head, etc. — of the Hy are in a far more advanced condition within the larva than they really are; but the main point remains, he recognised the fact that the parts of the nymph are at first invaginated, and that their appearance on the surface is due to evagination. And, further, he recognised a distinct stage in the evolution of the nymph, when all these parts are invaginated, and when the whole organism appears to be little more than a simple sac, containing a fluid or semi- fluid material ; I shall refer to this stage as the pronymf>h. The observations of Reaumur quoted above agree very closely with those of Van Rees, and with the following state- ment which I made in 1872 : ' As yet a complete series of investigations are wanting, but I have traced the steps of development sufficiently to allow me to state that the great procephalic lobes which exist in the half-developed embryo become folded inwards, and lie one on either side of the alimentary canal during the whole period of larval life. These involuted procephalic lobes — and they are nothing else— form a portion of the imaginal discs of Weismann, whilst the eyes, antenna', and mouth organs are ultimately developed from cellular outgrowths at the bases of the same structures, just as they are in the Crustacea ' [142]. The above is, I believe, the earliest notice of the invagination theory of the origin of the imaginal discs, which I think more recent researches have placed beyond a doubt. The development of the imago within the pupa-case will be considered under the following heads : 1. The formation of the pron\'mph from the larva. 2. The development of the nymph from the pronymph; and 3. The development of the imago from the nymph. The first two form the subject of the present chapter, to which a resume of the third has been added. The details of the development of the various organs of the imago will, however, be more conveniently elucidated in the second volume of this work. 296 THE DEVKLOP.UKN'r OF THE NYMPH. 1, THE FORMATION OF THE PKONYMPH FROM THE LARVA. From the commencement of the pupa stage to the end of the second or middle of the third day.* The Paraderm. — I shall show hereafter that the whole integu- ment of the nymph is developed from the epiblast of the imaginal discs, but after the larval tissues have undergone histolysis (see p. 22), and are converted into a cream-like pseudo-yelk, the imaginal discs, which are as yet concealed in their provisional capsules, are united with each other by a cellular membrane, which encloses the pseudo-yelk. Van Rees regards this as the larval hypodermis ; but the larval hypo- dermis has long before undergone complete histolysis, and the cellular covering of the pseudo-yelk is a new formation of parablastic origin. It is a temporary structure, destined to be replaced by the epiblast of the discs. I have therefore termed it the paraderm. The Pronymph. — The parablastic sac enclosing the pscudo- yelk is Reaumur's ovoid stage, my pronymph. When the paraderm has been replaced by the ectoderm of the discs, the nymph is fully formed, and takes the place of the pro- nymph. Moreover, a cuticular layer is shed at this period, constituting the pupa-sheath of Weismann, which corresponds with the hard covering of the nymph in the Lepidoptera and other insects with obtcctate nymphs (see p. 20). The changes which occur before the formation of the pupa- sheath correspond nearly with those which take place in the caterpillar before the ecdysis of the last larval skin, whilst those which occur after its formation correspond with the changes which take place in the chrysalis, or nymph stage of the Lepidoptera. If these views are correct, Reaumur's idea that the Diptera * The dates from the commencement of the pupa stage given in this work are much longer than those given by Van Kees, but correspond pretty closely with those of Wcismann's memoir. They are only approximate, as so much depends on temperature. FORMATION OF THE PRONYMPH FROM THE LARVA. 297 exhibit an extra stage not separated as distinct in other Insects is substantially justified. In the Lepidoptera the phenomena of histolysis are partly carried out in the larva and partly in the nymph, whilst in the Diptera they are completed in a much shorter time, and a new stage of development becomes manifest. In the Lepidoptera the imaginal discs are united before the larval skin is shed, whilst in the Diptera their union is delayed until the process of histolysis is almost complete. The Histolysis of the larval tissues proceeds from before backwards, and from without inwards. It commences in the anterior segments during the resting stage of the larva, and is not complete in the posterior segments until after the forma- tion of the nymph. The cells of the hypodermis and the muscles are first attacked. The separation of the hypodermis from the larval cuticle is the first change which occurs after the pupa state is assumed, and, like the histolysis of all the tissues, proceeds from before backwards. It is easy to remove the anterior segments of the pupa-case on the first day of the pupa, but it is not possible to remove the posterior seg- ments of the case before the end of the second, without injury to the pronymph, as the larval muscles and hypoderni remain attached to the cuticle longer than those of the anterior segments. a. Histolysis of the Larval Muscles. Kowalevski was the first who actually demonstrated the manner in which the histolysis of the larval tissues is effected, but he was led to undertake the investigation by the writings of Metschnikoff [143] on intracellular digestion. All the observations made before Metschnikoff's great discovery of the part played by the white blood corpuscles, phagocytes, may be passed over in silence, as they have been completely super- seded and shown to be erroneous by the investigations of Kowalevski, first published in 1884. All that was actually known before that date may be summed up in the following words. A number of large granule cells, called ' Kcirnchen kugeln ' by Weismann, and ' corps rosea ' by Viallanes, make 20 — 2 298 THE DEVELOPMEM' OF THE NYMPH. their appearance whilst the muscles and other tissues of the larva undergo disintegration. The change is preceded and accompanied by a great increase in the number of blood corpuscles, and the whole contents of the pupa, except the rudiments from which the imago is developed, assume the form of a white cream-like fluid, or pseudo-yelk, the cellular elements of v/hich, granule cells, are similar to the cellular elements of the great food-yelk of Birds and Reptiles. As my own observations confirm those of Kowalevski in almost every detail, I shall give the results of his investigations. He says : ' If we investigate the changes which are going on in the young pupa by sections, it will be seen that the muscles and other tissues are completely surrounded by the blood plasma, and that, more especially in the anterior part of the body, vast numbers of blood corpuscles, leucocytes, adhere to the muscles. Pupa; one or two hours old already exhibit indica- tions of the penetration of the sarcolemma by these corpuscles. Generally some of the muscle fibres exhibit one or two leuco- cytes in their interior close to the sarcolemma, whilst others still remain without any. A few hours later many corpuscles are seen within the muscle fibres. Minute cracks then appear radiating from these cells, and processes of the cells or the cells themselves may be seen lying in these cracks. Soon afterwards minute fragments of the muscle substance are seen entirely imbedded in these leucocytes, which are thus converted into the well-known granule cells. There is never any difBculty in distinguishing the nuclei of the leucocytes from those of the muscle, as they are smaller and more spherical. New cells are now seen continually passing into the cracks, and the latter extend so that the muscle becomes more and more broken up. When the whole muscle is permeated by the leucocytes, these assume a spherical form and separate from each other. The sarcolemma has by this time disappeared ; it is probably so perforated by the passage of leucocytes that it allows the fibre to fall to pieces. The muscle nuclei, which Viallanes believed to proliferate, are removed in the same way as the rest of the FORMATION OF THE PliO NYMPH FROM THE LARVA. 299 muscle ; they are surrounded and enclosed by immigrant blood corpuscles. Frequently, however, the nuclei remain after the complete histolysis of the fibre, and are isolated by the falling apart of the granule cells. In this case they lie free in the blood, but are ultimately seized upon and disinte- grated by amoeboid corpuscles. These nuclei, when enclosed within the leucocytes, lose their vesicular form and become converted into spheroid or ovoid masses of material, which stain deeply with carmine. The blood corpuscles attack the muscles with such energy that on the second day (third day) scarcely any remain which are not converted into granule cells. These, although loaded with angular and spheroidal pieces of muscle, do not cease to move by minute pseudopodia. Moreover, they continue to feed and attack the cells of the fat bodies.' There are two points in the foregoing description to which I would add a few words from my own observations. Kowalevski appears to think that all the leucocytes (phago- cytes) enter the muscle and fat bodies from the blood ; at least, he says nothing of their rapid multiplication within these tissues. Many of the leucocytes within the muscles and fat bodies contain from four to six or eight nuclei, and are evi- dently undergoing rapid proliferation. The second point is in relation to the muscle nuclei. It is true these are frequently seen surrounded by a clear proto- plasmic area in the substance of the muscle, or free in the blood ; but I cannot convince myself that the areole is a phagocyte. It appears to me that the leucocytes attach them- selves to the exterior of the nucleus and perforate its capsule by sending a pseudopodium into its interior, after which the nuclear fluid disappears, and the chromatin falls into a mass which exhibits no definite structure. The remains of the nucleus then pass into one of the adjacent leucocytes and disappear. b. The Histolysis of the Hypodermis, and the Formation of the Paraderm. Reaumur was acquainted with the fact that the transforma- 300 THE DEVELOPMENT OF THE NYMPH. tion of the resting larva into a pupa (see p. 3) is accompanied by a separation of the hypoderm from the overlying cuticular layers. The changes which occur in the hypoderm have been variously described, and they are by no means easy to follow. Weismann says, speaking of the third-day pupa : ' The thorax of the nymph is already formed by the union of the imaginal discs, but it is not, as one would expect, enclosed within the larval hypodermis. It lies immediately beneath the hornj' pupa-shell. The hypodermis and muscles of the thoracic segments have undergone degeneration and have changed into a fine granular mass, which mixes with the blood in the interior of the developing nymph' [2, p. 165]. Weismann, who be- lieved that the hypodermis of the abdomen of the larva changes directly into that of the imago, gives no nearer details upon the subject. In spite of this direct observation of Weismann's, which is perfectly correct, Graber [10, Bd. ii., Figs. 163 and 178] gives a well-known schematic representation, as the result of the observations of Ganin and Viallanes more especially, and represents the Fly-nymph as consisting of a simple abdominal cellular integument, the hypodermis of the larva, with a double thoracic integument, the larval hypoderm, enclosing the newly- developed thorax of the nymph. I may at once observe that this scheme is entirely erroneous. The most recent memoir on the Metamorphosis of the Blow-fly is by Van Rees [147], and his observations agree in the main with my own, which only differ in this, that what Van Rees regards as the larval hypoderm I regard as a new formation, developed after the histol3'sis of the larval hypoderm, my paraderm. Sections of the resting larva, and of the pupa in its earliest stage before a trace of colour appears in the cuticular shell, exhibit unmistakable histolysis of the hypoderm, the cells of which are invaded by leucocytes at an earlier period than even the larval muscles. All the phenomena described as occurring in the muscles likewise occur in the cells of the larval hypo- derm. The degenerating cells are swollen, their protoplasm FORMATION OF THF PRO.WMPH FROM THE LARVA. 301 becomes spongy and scarcely stains at all with ha;matoxylin, borax- or picrocarmine, and the vacuoles contain blood cor- puscles and multinucleated phagocytes. The nuclei of the epithelial cells are large and contain a large quantity of clear substance in which a small reticulum of chromatin is imbedded. This subsequently gives place to a mass of readily stained material on one side of the nucleus. The nuclei appear to be perforated by phagocytes. During these changes the hypoderm becomes widely separated from the larval cuticle ; and a very thin cuticular layer is developed on both its outer and inner surfaces. Sections made from pupae an hour or two older show that the whole pronymph is covered by a layer of cells, which differ from those of the larval hypoderm. These new cells, my paraderm, have far smaller nuclei, which are apparently solid. The cells arc no longer spongy and vacuolated, and their protoplasm, unlike that of the degenerating hypodermic cells, stains intensely with haematoxylin and carmine. Like the latter, they lie between the two cuticular laminae already men- tioned. It is exceedingly difficult to trace the origin of the paraderm. It either originates from the leucocytes developed within the cells of the hypodermis, or the hypodermic cells undergo a complete rejuvenescence. This later hypothesis appears to me most improbable, and I regard it as almost certain that the paraderm originates from leucocytes, and is a true parablastic tissue, similar to that which Korotneff has described as investing the yelk in the egg of Gryllotalpa before the appearance of the epiblast. Indeed, I think some of my sections (PI. XVIII., Fig. 4) show that the paraderm is formed outside the degenerating hypoderm, and consists at first of small but rapidly growing amoeboid cells. This parablastic layer is continuous with the pedicles of the imaginal discs, as Van Rees described it. I only differ from him in no longer regarding it as the hypoderm of the larva. It is gradually absorbed after the epiblast of the disc becomes, as it were, engrafted upon it ; for during the subsequent growth of the disc the large parablastic cells disappear beneath it, so that 303 THE DEVELOPMENT OF THE NYMPH. the small epihlastic cells of the disc itself take its place and form the body-wall of the nymph. Although neither Kowalevski [145] nor Van Rees [147] recog- nised the differences between the larval hypodermis and the para- derm which they regard as the larval hypoderm in the pupa stage, Viallanes [27] says : ' The hypodermic cells have become thicker than they were in the larva ; their contours are effaced, so that it is not possible to limit the extent of adjacent cells. The protoplasm is not only more abundant, but it has acquired a property not exhibited in the larva— it stains readily with carmine and hasmatoxylin. The nuclei are, moreover, pro- foundly altered.' Though there are inaccuracies in his de- scription, I quote it to show that the changed appearance of the cells has been observed. These new cells are certainly not Description ok Platk XVIII. The Histolysis and Regeneration of the Alimentary Canal, and the SlriicHin; of the Imaginal Discs. Fig. I.— a transverse section through the chyle stomach of a pupa four days old ; rf «, degener.uing larval epithelium; e e, embryonic epithelium developed from the scattered histoblasts of Kowalevski ; // elongated fusiform cells ; //, para- blastic layer consisting chiefly of amoeboid cells ; p>i, phagocytes feeding on the larval epithelium. Fk;. 2.— a transverse section of the ruilimentary mesenteron of the nymph, from a pupa five days old : /, the paraderm by which the larval hypodermis is completely replaced. Fig. 6. — Large granule cells from a pupa five days old. Fig. 7.— The edge of one of the wing discs from a pupa four days old, showing its relation to the cells of the paraderm and the amoeboid celh:, phagocytes, by which the latter is removed. Fic;. S.~ The paraderm from one of the intersegmental abdominal folds : /, superficial cells; /', palisade-liUe cells; , epiblast ; m, mesoblast ; />, parablnstic reticulum ; v, provisional capsule. PLATE XVIII. HISTOLYSIS. FORMATION OF THE PRONYMl'H FROM I HE LARVA. 303 thicker than those of the larval hypoderm, nor are the nuclei larger, as Vialkmes states, but smaller. Although transverse sections of young pupa; show that a considerable space exists between the pupa-shell and the larval hypoderm — which is filled by a serous fluid, and which remains for some hours after the paraderm is complete — the latter subsequently comes into close relation with the pupa-case, except at its anterior pole. This arises from the shortening of the pronymph, and is the result of the still further invagination of its anterior within its posterior segments. The anterior part of the abdominal wall, or, rather, of that part of the paraderm which has the abdominal discs attached to it, now forms a double fold, enclosing the parts derived from the thoracic segments. I term this the ' anterior abdominal fold.' The anterior abdominal fold is deepest on the ventral aspect of the pronymph, and is probably produced by the last con- tractions of the longitudinal ventral muscles of the larva. Subsequentl}' the whole paraderm undergoes contraction with important results ; for I regard this contraction as one of the essential factors in the evolution of the discs. Van Rees ascribes all the contractions which give rise to the definitive form of the body of the nymph to the larval muscles ; yet these contractions occur at a time when, if any larval muscle remains, it is entirely detached from the integument of the pronymph. Kowalevski actually observed slow but rhythmic contractions of the body of the pronymph after the whole of the larval muscles have been disintegrated by histolysis. I have no doubt whatever of the contractile power of the paraderm. When the paraderm is completely developed, the thoracic portion of the pronymph is entirely covered by the anterior abdominal fold, whilst the cephalic portion is similarly in- vaginated within the thoracic. c. The Relation of the Imaginal Discs to the Paraderm. I have already drawn attention to the true morphological character of the imaginal discs (p. 75). At the time I wrote 304 THE DEVELOPMENT OF THE NYMPH. that section of my work I was unacquainted with the masterly memoir of Van Rees, and I would here remark that my con- clusions agree so closely with his that, without this disclaimer, it might be supposed I had derived them from him without acknowledgment. I shall perhaps do best to give a translation of the thesis which Van Rees supports in his memoir. After giving the history of the views of his predecessors, he says [147, p. 22] : ' It appeared to me that there is only one possibility which leads to a complete solution of the problem before us, that the imaginAl discs are not only ectoderm, but are invaginations of the ectoderm itself. I can only understand the manner in which the Muscidae are developed by supposing that the ancient progenitors of the Flies had imaginal discs, which, like those of the Tipulidae, lay in immediate relation with the larval hypoderm ; and that in later generations these were continually drawn more and more deeply into the maggot until they assumed their present positions. ' The fact that each imaginal disc in Corethra is supplied by a nerve and a tracheal vessel affords us a clue to the manner in which the relation of the discs to the hypoderm has been maintained. Although at length the imaginal rudiment may appear as a mere appendage of the nerve or tracheal vessel, it has nevertheless a neck or hollow pedicle closely applied to the nerve or trachea. This is readily seen when the neck is short, but in the extreme case we have probably to deal only with a difference of degree, and not of kind, so that we may conclude that the direct relation between the disc and the hypoderm is always maintained, and that the insertion of the pedicle into the hypoderm indicates the point where the disc must have lain in the ancestral form. The object of my researches has been to demonstrate this postulated connection between the disc and the hypoderm.' Van Rees brings no embryological evidence to support his thesis, and merely remarks that he has traced the pedicles of the wing-discs into the hypoderm in the half-grown larva, in FORMATION OF THE PRONYMPH FROM THE LARVA. 305 which the pedicles of the leg-discs are more distinct than in the full-grown maggot. As the discs of the wings and halteres are the only ones in which I have not succeeded in tracing any connection with the hypoderm, I am exceedingly gratified to find Van Rees places the connection on the same basis as that of the other discs by direct observation. With regard to the function of the pedicles in relation to the evolution of the discs during the formation of the nymph, Van Rees has shown by transverse sections that the provisional cavity opens upon the surface by the shortening and ultimate opening out of the pedicle. His paper is illustrated not only by drawings of the discs at the period of their evolution, but by diagrams. d. The Contraction of the Paraderm and Evolution of the Discs. In the pronymph stage the imaginal discs all enlarge rapidly; their provisional cavities also become distended with fluid and extend into the pedicles. The latter become shortened, until at length the disc sacs lie in immediate relation with the paraderm. A ring of small cells now appears at the junction of the disc sac with the paraderm in the position previously occupied by the pedicle, and the adjacent large flat paraderm cells begin to contract. The area of each cell diminishes and its thickness increases, so that the nuclei of adjacent cells are drawn together. At the same time an orifice appears in the centre of the small cells, which opens into the disc sac, and this rapidly enlarges, so that these sacs are converted into open pouches. The cells of the provisional capsules undergo the same con- traction and thickening as the paraderm cells, a change which brings the disc to the surface. When this is complete, the provisional capsule becomes part of the paraderm. The effect of the gradual but continuous contraction of the paraderm has been already alluded to. This contraction occurs chiefly, but not exclusively, in that part which corresponds with the abdomen of the nymph, where the evolution of the disc 3o6 THE DEVELOPMENT OF THE NYMPH. epiblast takes place slowly. The paraderm of the thoracic region is removed during the rapid increase of the thoracic discs, which grow over its surface and subsequently unite with each other except on the dorsum ; where the paraderm remains, however, the only covering of the pseudo-yelk long after the complete evolution of the head. When the head is thrust forwards from the interior of the thorax, the cephalic discs are still in a very rudimentary con- dition, and are not united with each other except by the paraderm. Van Rees ascribes the evolution of the head and thorax to the contraction of the muscles of the larva before their final degeneration, but it is certain that these are in an advanced state of histolysis, and have lost all connection with the integument before it occurs. The evolution of the head and thorax is not, therefore, the result of muscular contraction, but of the organic changes which take place in the cells of the para- derm. These are slow and continuous, and the contraction of its surface, accompanied by increase in its thickness, continues until it is finally replaced by the ectoderm of the discs. e. Histolysis of the Tracheae of the Larva and Development of the Tracheae of the Pronymph. The tracheal vessels of the pronymph are developed from the anterior superior thoracic disc, which surrounds the spiracular trunk of the larva, and from the vessels already described (p. 85), which exist in the larva in relation with the imaginal discs, and exhibit an outer coat formed of small embryonic cells. The greater part of the larval trachese undergo active histo- lytic changes; the external cellular coat is entirely stripped from them by the action of phagocytes, and the naked intima collapses. The intima of the great longitudinal trunks is seen for three or four days lying in the pseudo-yelk. As Weismann observed, it is severed from its connection with the persistent portion of the tracheal system very near the anterior spiracle. It con- tracts to about half its original length, lies entirely in the FORMATION OF THE PRO NYMPH FROM THE LARVA. 307 posterior half of the pronyinph, and is ultimately withdrawn. It is found still attached to the cuticle of the larva around the posterior spiracles, on the inner surface of the pupa-case after the escape of the imago. Weismann supposed that the remains of the longitudinal trunks of the larva are withdrawn at the time the imago escapes from the pupa; but as these exuviae are entirely disconnected with the pupa-sheath and lie outside it, it is evident that they are expelled from the body of the pro- nymph before the pupa-sheath, the shed epidermis of the nymph, is formed. Indeed, I have found them in the pupa- shell freed from the body of the nymph on the third day of the pupa stage. The shedding of the tracheal exuviae appears to be dependent on the contraction of the abdominal paraderm, and to occur at the same time as the expulsion of the cuticular intima of the rectum of the larva from the pronymph. f. The Histolysis of the Fat Bodies and other Larval Tissues. The Fat Bodies during the formation of the pronymph separate into their component cells, so that the contents of the body cavity has the appearance of a granular fluid even before the histolysis of the muscles is complete. The fat cells, however, undergo histolysis very slowly, so that, as Weismann and Ganin observed, many persist even in the imago when it emerges from the pupa. The histolysis of the fat bodies commences, like that of the muscles, at the anterior extremity of the larva, nnd proceeds from without inwards, so that in the pronj mph their cells are separated from each other, except where they are in immediate relation with the remains of the alimentary canal of the larva. Kowalevski claims to have observed the immigration of the leucocytes into the fat cells in the following manner. He says : ' The breaking up of the fat bodies can be observed in the living nymph. I have several times withdrawn one from the pupa-case on the third or fourth day, and kept it alive in white of egg for more than twenty-four hours. The head vesicle is fairly transparent, and it is possible to watch the 3o8 THE DEVELOPMENT OF THE NYMPH. disintegration of the contained cells. I saw small granule cells attach themselves to a fat cell and crawl over it ; others subsequently made their appearance, so that two hours from the beginning of the observation the whole fat cell was covered by these leucocytes; indeed, in this state it resembled a segmented ovum in the morula stage. The appearance per- sisted for a long time, until at length a nest of granule cells entirely replaced the fat cell, and these finally separated and scattered themselves.' I have not repeated this observation of Kowalevski's, and I am unable to understand how he obtained a satisfactory view of the process in ihe entire nymph. But I have taken living fat cells from the pupa, spread them over a thin cover, and fixed them with steam, subsequently staining and mounting them. Such preparations sometimes exhibit leucocytes in the act of entering a fat cell with remarkable clearness. The changes which occur in the fat cells after the penetration of their external membranes by leucocytes are exceedingly interesting. At first a group of leucocytes is observed sur- rounding the nucleus ; these multiply with great rapidity, and extend in ever increasing numbers outwards towards the cell- capsule. As the leucocytes increase in number, the fat granules gradually disappear and give place to vast numbers of cells, which differ in no perceptible manner from the blood corpuscles of the larva. The cells nearest the nucleus of the fat cells are far larger than those near the periphery, and the former are usually multinucleate. The latter probably originate by the division of the former. During the process the envelope of the fat cell disappears, and eventually the peripheral leucocytes become scattered in the blood. For a long time after this the central larger cells cohere around the nucleus, but ultimately they separate and leave the nucleus, which, after a longer or shorter time, appears to be attacked like the muscle nuclei by phagocytes. Isolated nuclei are frequently seen free in the blood of the nymph, and these are usually full of cells which are set free as leucocytes PI. XVI., Fig. lo). FOr). The contraction of this layer is apparently the cause of the shortening of the gut, and probably not only causes the epithelium to accumulate in a mass, but also in- vaginates the muscular coat of that portion which has no imaginal cells between it and the epithelium, or, in other words, draws the degenerating and contracted intestine into the interior of the new layer of imaginal cells. The similarity of the fusiform cells to the muscular elements of the intestinal wall is certainly very striking. The further development of the new mesenteron is one of the most extraordinary phenomena which occur in the pupa state. As it does not take place until the fourth or fifth day, it will be described in the next section. h. The Position of the Neuroblast and the Development of the Stomo- daeum of the Pronymph. The description given by Kowalevski of the manner in which the great crop of the larva disappears differs entirely from that of Van Rees. The Russian naturalist believed that it degenerates in situ. The statements made above agree more closely with those of Van Rees. Kowalevski's error originated by his not having observed the remarkable change which takes place in the position of the neuroblast at this period. The axis of the central nervous system, which in the larva is parallel with that of the body, turns through a right angle on the first day of the pupa stage, so that it is transverse to the long axis of the pupa, and transverse sections are no longer across, but parallel to the axis of the nervous system, the dorsal surface of which looks backwards, and the posterior extremity of which is ventral. At the same time there is a great increase in the size of the infra-cesophageal ganglia, which now lie in front of, and not below, the supra-oesophageal nerve centres (Pi. XIX., 5, 0- The crop, when it is withdrawn into the oesophagus, forms a swelling which lies in front of and above the infra-cesophageal ganglia, which Kowalevski mistakes for the hemispheres, or 21 — 2 314 MHE DEVELOPMENT OE THE NYMPH.) supra- oesophageal nerve centres. This change of position in the nerve centres makes it appear as if the crop still remains dorsal to the hemispheres in transverse sections until the real nature of the nerve ganglia upon which it lies is recognised ; for the infra-CBSophageal nerve centres, without careful examina- tion in their new position and greatly increased development, are readily mistaken for the hemispheres. Van Rees, without remarking the changed position of the neuroblast, observed the withdrawal of the crop towards and afterwards into the oesophagus, whilst Kowalevski drew his conclusions entirely from an erroneous interpretation of the appearances presented by transverse sections. Kowalevski derives the new stomodaeum from the proven- tricular ring, \'an Rees from a group of cells which appear at the point of invagination of the crop. My own view is that the remains of the larval stomodaeum become invested by a layer of parablast ; and that the new stomodaeum is developed mainlj- from the embryonic cells which form part of the head discs, from the lingual discs which are seen on either side of the labium at the oritice of the salivary duct, and probably from the imaginal rudiments in the larval oesophagus, described by Van Rees, which I have not been fortunate enough to identify. These may possibly form the new crop, at first a short diverticulum of the oesophagus corresponding with the origin of that of the larva, which, it will be remembered, is a ventral diverticulum of the oesophagus, and not a dorsal one, as is generally believed. I have already alluded to the discontinuity of the stomodajum and proventriculus of the pronymph. My sections of pupae of the second day show no trace of the oesophagus immediately in front of the proventricular ring, and still further forward one comes suddenly upon the blind end of the degenerating oesophagus. This is surrounded b\' numerous parablast cells. I strongly suspect that the blind posterior end of the oeso- phagus is connected with the proventriculus by a solid cord of parablast during the whole metamorphosis, although at this FORMATION OF THE PRONYMPH FROM THE LARVA. 315 early stage I have been unable to demonstrate the connection. In a later stage the solid parablastic cord uniting the blind end of the new stomodaeuni with the rudimentary proventriculus of the nymph is sufficiently apparent to leave no doubt in my mind upon the subject. The development of the new stomodseum is by no means easy to trace, owing to the rapid changes which occur during the evolution of the head. According to Kowalevski, it is at first a wide, short tube, which only subsequently becomes narrow and elongated. Van Rees failed to observe its first stage, but it is evident that before the evolution of the head it must be a very much shorter tube than when the head appeais in front of the thorax. 2. THE DEVELOPMENT OF THE NYMPH. From the. end of the second day to the fifth day of the pupa sta^c. a. The Position of the Imaginal Discs in the Pronymph. The disposition of the imaginal discs in the pronymph is seen in PL XIX., which represents a vertical section in the median plane, from the rectum (;-) to the neuroblast (s, t). In front of the neuroblast the section is taken through the discs, about I mm. to the right side of the median plane, except in the case of the wall of the oesophagus and the rudimentary mouth (/, e), which are shown in the median plane. Hence this figure is so far diagrammatic ; it is a construction made from a series of sections, and not a drawing of any one section. It will be seen by reference to the plate that the paraderm, represented by a double outline enclosing a series of short black lines, is invaginated above, and forms the roof of a cavity (/, h). This I term the ' cephalic involution.' Below and on either side of the cephalic involution there is a crescentic invagination of the paraderm, communicating laterally with the sacs of the 3i6 THE DEIELOPMENJ- OE THE NYMPH. wing discs {w), and below with those of the leg discs (/ h I). This I term the ' thoracic involution.' The thoracic discs are still enclosed in the disc sacs, but these open into the cavity of the thoracic involution : those of the wings (w) above and laterally ; those of the upper metathoracic discs, which arc not represented in the figure, behind and below the orifices of the wing sacs ; and those of the leg discs in order from before backwards on the ventral surface of the invaginated thoracic wall. The anterior stigmatic (upper prothoracic) discs are at this period conical invaginations around the main trunks of the anterior stigmatic tracheae of the larva (c). The cephalic involution exhibits a roof which consists entirely of paraderm, and a floor composed chiefly of the cephalic discs. These are arranged in the following manner : In front of the stomodaeum are the maxillary and lingual discs ((.'), with the orifice of the sericterial ducts between them. Above arid behind the stomodaeum is a tongue-like process (/), from which the pro-, meso- and metalabrum are developed. Behind and laterally are the antennal rudiments {^), and the optic discs of Weismann (/?) ; the latter rest upon the hemispheres of the neuroblast {s). It will be observed that all the imaginal discs are placed in Descrii'tion ok Plate XIX. A sagittal section of the pronympli on the second day of the pupa, conslructecl from a series of longitudinal and transverse sections. The section represents the median plane except in front of the neuroblast, .i, t, where it is taken through the principal di>cs and gives the relation of the parts both in the median jjlanc and about I mm. to the right of the median plane. The stomodiieum is in the median plane, and the disc sacs in the plane of the anterior spiracles. a, Cephalic involution of the larval integument ; h, anterior larval spiracles ; c, shed intima of the anterior spiracular trachea of the larva ; il, mouth-armature of the larva ; e, labial disc ; /, labrum ; .4', antenna ; h, optic disc ; /, anterior leg disc ; L, intermediate leg disc ; /, posterior leg disc ; m, sericterial gland of larva ; h, cavity of the mesenleron of the pronymph, containing the intestine and Malpighian vessels of the larva ; 0, dorsal vessel ; p, parablast surrounding the mesenteron, and connecting it with the rectum and stomoda-um ; r, rectum ; s, hemisphere ; /, ventral neuroblast ; 11, crop ; ~m, wing disc. The dark continuous outline within the chyle stomach of the larva represents the new epithelium of the mesenteron (compare PI. XVI II., Fig. 2, e). THE DEVELOPMENT OF THE NYMPH. 317 the same relative positions as the parts subsequently developed from them ; it is only the paraderm which is reversed by the invagination, its thoracic extremity beinp^ in front of its cephalic portion. As the rudimentary nerves and tracheae of the nymph are all attached to the discs, no disturbance of their relations occurs during the process of evagination, and the reversal of the parablast, which is attended by its contraction, takes place without any disturbance of the internal organs. The Evagination of the thorax, and afterwards that of the head, is undoubtedly effected by the contraction of the ab- dominal paraderm, which forces the pseudo-yelk forwards into the thoracic and cephalic cavities. As has been already observed, Reaumur came to the conclusion that the process is a mechanical one, and Weismann made the following remarks on the evolution of the head [2, p. 173] : ' On the fourth day of the pupa stage the head is pushed forwards out of the hollow in the thorax in which the discs are developed ; it then unites with the front of the thoracic integu- ment. What the nature of the force is which produces the forward movement of the head may be subject to discussion, but I can assert with confidence that it is not the result of growth, but is purely mechanical. I once found the thorax well developed in a fourth-day pupa, but no head was visible ; the latter appeared after the preparation had lain for som.e hours under the pressure of the cover-glass, although the nymph was of course dead.' This is merely a repetition of Reaumur's experiment under somewhat different conditions. As Weismann, however, makes no reference to Reaumur, I think it evident that he was not at that time conversant with his memoirs. Reaumur, Weismann, and more recently Van Rees, have ascribed the evolution of the head and thorax to the con- traction of the larval muscles before their final degeneration. I think there is little doubt that the last muscular contraction gives rise to the shortening of the pronymph, by which it is drawn back in the pupa-case. This retraction releases the pharyngeal skeleton of the larva, which is subsequently found 3i8 THE DEVELOPMENT OF THE NYMPH. depressed upon the ventral surface of the pupa-case in a dry and brittle condition. The formation of the anterior abdominal fold is also probably due to muscular contractions, but the reduction in the size of the abdominal region, which gives rise to the evolution of the head and thorax, is certainly not muscular. It occurs after the muscles are all completely detached from the body wall, and when most of them are in a far advanced stage of histolysis. These changes are clearly due to contraction of the paraderm, the cells of which increase in thickness as they diminish in the extent of their surface. Moreover, the character of the contraction differs entirely from the result of a muscular act. The reduction of the magni- tude of the abdomen occurs very slowly, and gives it the form characteristic of that of the nymph. The anterior abdominal fold is obliterated and the thorax exposed, whilst the latter is subsequently distended, and the head is evolved by the flow of pseudo-yelk from the abdomen forwards. b. The Development of the Integument of the Head and Thorax. As has been already stated, the integument of the thorax is developed from the epiblast of six pairs of imaginal discs. The manner in which these overgrow and replace the paraderm has been described by Van Rees [147] , and, except that he terms the paraderm ' the larval hypoderm,' my own observations confirm his statement that the edge of each disc overlaps the paraderm and extends by the development of new cells, whilst Description ok Pi.aie XX. The exterior of the pronymph and nymph, and the anterior spiracles of the nymph. KiG. I.— The thorax of the nymph at the end of the third day of the pupa, seen from its ventral surface (after Weismann). KiG. 2. — The same seen from its dorsal surface. KiG. 3.— The anterior spiracular apparatus of the nymph on the ninth or tenth day of the pupa stage : s c, stigmatic cornu ; s, intersegmental spiracle ; / r, tracheal vessel. Fig. 4.— The pronymph at the end of the third day, showing the position of the abdominal imaginal discs. Fig. 5. — The nymph on the sixth day of the pupa stage. PLATE XX- PRO.NYMni AND NYMPH. THE DEVELOPMENT OF THE NYMPH. 319 the underlying paraderm is removed by phagocytes, which exist in great numbers near the edge of the growing disc. During this extension, however, the disc also increases in size by the growth and multiplication of all its cells, whilst the paraderm undergoes a corresponding contraction in its whole extent. In this way the edges of the discs are ultimately united with each other, first in the ventral region, and later, on the dorsal aspect of the nymph. The Upper Prothoracic Discs (PI. XIX., c and Fig. 11) are seen as pouch-like involutions of the anterior edge of the cephalic involution, surrounding the shed intima of the pro- thoracic spiracular tracheal trunk of the larva. These discs subsequently undergo evolution, i.e., their inner surface, that next the shed intima of the tracheae, becomes external, and they thus become the horn-like stigmata of the pronymph (stigmatic cornua). The rudimentary prothoracic tergum (dorsum of the prothorax), which is far larger in the pronymph than it is subsequently, is also developed by the extension of the edges of the anterior stigmatic discs (PI. XX., Fig- 4)- The Stigmatic Cornua of the pronymph (PI. XX., Fig. 3, s c) have a simple trumpet-like orifice, rapidly become highly chitinized, and acquire a yellow colour. They are subsequently shed, and are replaced by intersegmental spiracles developed between the pro- and mcsothorax. These are the anterior spiracles of the nymph (PI. XX., Fig. j, s). The Intersegmental Anterior Spiracles of the nymph resemble the anterior spiracles of the larva in having digitate extremities, but differ in possessing only four or five digitations. Weis- mann confounded them with the stigmatic cornua, and it is only recently that I discovered that the two are distinct and exist simultaneously. The coexistence of two spiracles, one on, and the other behind, the prothorax, is of extreme interest, setting at rest questions as to the nature of the imaginal spiracles and the morphology of the thorax which have already been discussed, and bearing out the views I have expressed on the subject. 320 THE DEVELOPMENT OF THE NYMPH. The Leg and Wing Discs are first exposed by the retraction of their provisional membranes. Their thoracic edp^es unite with each other in the median line of the ventral surface (PI. XX., Fig. i), subsequently from before backwards, and last of all on the dorsum. The appendages are developed as diverticula of the body cavity, which rapidly grow in length and diameter, and assume positions similar to those of the lepidopterous nymph (PI. XX., Fig. 5). The dorsal surface of the mesothorax (PI. XX., Fig. 2) is completed by the union of the right and left wing discs. The upper prothoracic discs do not meet until the mesothorax is much enlarged, and the metathoracic discs remain separated in the median line above, or, if thej' meet at all, only do so by a narrow isthmus. The mesothorax, when it is first closed dorsally, is a narrow ring no wider than the mesothoracic segment of the larva, but it increases rapidly from before backwards, and even at first exhibits a depression marking the future position of the post-scutal sulcus (p i65). The Head first appears in front of the thorax as a thin, bladder-like projection from its interior. As soon as it is entirely exposed, it is seen to exhibit five convexities, one lateral protuberance on each side corresponding to the optic disc and three median vesicles. The frontal vesicle is most prominent ; it has the facial (antennal) vesicle on its ventral surface, and a narrow rim representing the occipital surface of the head, the posterior vesicle, above and behind it. The thin- walled head consists in great part of the paraderm of the cephalic invagination with the cephalic discs, which form a comparatively small part of its surface ; for some hours after it appears in front of the thorax a great part of it consists of large flattened cells, very distinct from the embryonic epi- thelium of the disc epiblast. The Neck. — The constricted neck is developed very slowly by the ingrowth of the parablastic layer, which forms a septum between the head and thorax. Sections through this region, and also those between the segments of the abdomen, exhibit THE DEVELOPMENT UF THE NVMPH. a remarkable palisade-like structure (PI. XVIII., Fig. 8), which Van Rees has mistaken for the developing discs. The large cells of the paraderm give off processes and form new cells beneath them, which enclose a network of blood sinuses. These reticular septa are well developed between the abdominal segments when the discs are quite small, and they are not replaced by permanent structures until the development of the nymph is far advanced. The Proboscis is developed from two median processes (PI. XIX., c and/). The upper and larger of these, /, repre- sents a layer of imaginal cells, which, in the larva, lie within the pharyngeal sinus (p. 44), and which is withdrawn from the sinus by the shedding of the cephalo-pharyngeal sclerite. It V ic, 45 - \ iieurobl.Tst, d y, '^Iii^hiiil; ihe proboscis, ssfl of ilif young nymph. I mil cIl , irclmilLion, inil iIoimI becomes the pro-, meso- and metalabrum. The lower and smaller process, c, is formed by the coalescence of the four appendicular discs of the head (p. 82), and it is from this that the rostrum, pseudolabium and ligula are developed. The maxillary discs (p. 82, Figs. 8, 2, I ; 13, mx, and PI. XV., Fig. I, i"), which are first situated on the inner surface of the stomal disc of the larva, increase rapidly in size during the first few hours of the pupa stage, when they lie one on either side of the cephalo-pharynx. After the latter is shed, the discs unite with each other and with two small groups of imaginal cells, one on either side of the orifice of the united sericterial ducts, to form the lower median process, e (PI. XIX.). 323 THE DEVELOPMENT OE THE NYMPH. The small groups of imaginal cells at the orifice of the seric- terial ducts represent the labium of the larva in which they are developed, and they form the ligula of the imago (p 147) ; hence, as has been already stated, I regard the pseudolabium, which is developed from paired maxillary discs, corresponding with the maxillae of the embryo, as representing the galeae of the maxillae, and not the true labium, as is usually held. On the fourth day of the pupa the mouth organs closely resemble those of an hemipterous insect (Fig. 45). The upper and lower processes lie upon the pectus and form a labrum and haustellum, between the basal portions of which the fulcrum is subsequently developed. The oral lobes do not appear at this stage, but are subsequently developed from the sides and apex of the haustellum, and by the seventh or eighth day the organ has assumed its definitive form (PI. XXI.). Weismann's account [2] of the manner in which the proboscis is developed is exceedingly indefinite. He says : ' Its several parts originate in a manner which is totally unlike that in which the mouth organs of the larva are formed ;' and, after describing the origin of the latter from paired rudiments (lateral appendages), continues: 'The parts of the proboscis first appear in the pupa as structures similar to those which are ultimately developed from them ; thus the underlip does not originate from paired rudiments, like the under-lip of the larva, but as a hollow, grooved process.' This is undoubtedly my lower process c, the origin of which I have traced to the coalescence of the paired maxillary rudiments. Weismann, moreover, admitted that he was unable to observe the earlier stages of the development of the proboscis. Menzbier [50], on purely theoretical grounds, concluded that the head is developed from six pairs of discs, and he supposed that three pairs unite to form the proboscis ; Kunckel d'Hercu- lais, as has been already mentioned, described the maxillary discs in Volucella, and these are the only facts which have been previously recorded as to the manner in which the pro- boscis is developed in the Diptera. The Antennae are developed as hollow processes of the great THE DEVELOPMENT OF THE NYMPH. 323 head discs. Weisrnann says : ' The formation of the antennse resembles that of the legs. First an oval furrow appears, which has a considerable extent and encloses an ovoid projection. The latter resembles the nucleus of the leg disc. Very soon three concentric furrows appear within it, so that the ovoid body is divided into three segments, the three rudimentary antennal joints." And he adds : ' Even on the second day of the pupa the antenna has scarcely any resemblance to the developed organ.' In Plate IV. the appearance of the antenna is repre- sented in an advanced stage of the resting larva, on which Weismann relied for his conclusions. In a still younger larva, however, from which Fig. 13 was constructed, the antennae are well seen in sections, and bear an unmistakable resemblance to the developed organs. The folds which surround the antenna at a more advanced period, and mask its true charac- ter, are developed from the part of the disc which becomes the face and forehead, and not, as Weismann supposed, from the antennal rudiment. These folds possibly form the two basal joints of the antenna, which are well seen in some of my sections on the second day of the pupa, when they are almost as large as the third joint. In comparing the development of the antenna; with the legs, Weismann was probably biased by the belief that these organs are homologous with ventral appendages. The difference in the development of the two structures is very marked. The antennae are first mere ridges of the head disc, enclosing the rudiments of the antennal nerves, and never become biramous at any stage of their evolution. Nor are they ever, enclosed in separate provisional capsules distinct from those enclosing the optic discs. Changes in the Disposition of the Eye Discs. — The eye discs in the resting larva (Fig. 13) lie in diverticula of the cephalic in- volution. Their corneal (outer) surfaces look outwards and forwards, and are covered by a provisional membrane. In the next stage the provisional sacs have become part of the cephalic involution, and the eye discs are thickened and plicated (PI. IV.) 324 THE DEVELOPMENT OF THE NYMPH. and form part of the wall of a great pyriforin sac extending from the brain to the pharynx. In the pupa on the third day they have become sacs, the corneal surface looks towards the interior of the sac, and is concave, whilst the outer convex surface is covered by an ex- pansion of the optic stalk which unites them with the brain (see Fig. 46). Van Rees finds great fault with Viallanes' figure of the eye disc, which represents it with the outer surface covered by a provisional membrane, and states that the surface which looks towards the exterior is covered by mesoblast. He has figured and described the discs in their third stage as they appear in the pupa, the concave corneal surface bounding the lumen of a sac. During the evolution of the head, the eye discs again become convex on their corneal surfaces. Van Rees gives a perfectly correct description of the con- dition of the parts from which the head is developed, as they are found in the last larval and early pupa stages, but, had he examined them in a series of sections from larvae a few hours younger, he would have found that Viallanes' figures are as correct as his own. He says : ' The forehead and eye discs of the adult larva extend from the brain, with which the latter are connected by nerve stalks, to the pharynx. The eye discs extend over the front of thq brain and resemble mushrooms, whilst the antennal discs, which are continuous with the edges of the eye discs, at first form a tube and later a cone, with its apex attached to the pharynx.' This description scarcely differs from Weismann's. c. Changes in the Neuroblast, and Development of Peripheral Nerves. It is not my purpose at present to enter into the details of the development of the nervous system, which will be more con- veniently given in a future chapter, but to present to the reader a review of the great changes which occur during the evolution of the neuroblast. THE DEVELOPMENT OF THE NYMPH. 325 In the resting larva and pronymph the neuroblast increases rapidly in size, and, as already stated, its position varies at different stages of the development of the cephalo-thoracic discs. Fig. 46. — a semicliagrainiiialic scciion of one of the hemispheres, the optic cup and optic disc from a pronymph on the third day of the pupa stage : a c, anterior commissure ; a g, antenn.il Ranglion ; d, optic disc ; a, n;4ophagus ; s, optic slalk : r t, retinal ilisc ; /• .v, retinal stalk, connecting the ciip-llUe ilisc with the optic ganglion, which is seen surrounding its inner extremity ; tr, trahecula. During the evolution of the head discs on the second or third day of the pupa state, the hemispheres, which up to this period remain approximately spherical, become pyriform, and a groove 326 THE DEVELOPMENT OF THE NYMPH. makes its appearance, dividing each into a proximal and a distal portion. The external or distal portion is larger, and subsequently develops the great optic ganglion and the retina of the compound eye ; the smaller or proximal inner part be- comes the supra-oesophageal nerve centre. I have already referred to a remarkable epithelial disc, the retinal imaginal disc (p. 70) as it occurs in the larva (Fig. 13, rl), which has not been seen by any previous observer. A com- parison of this disc with the cerebral vesicle of the embryo (Fig. 44) renders it probable that the disc cavity is the remains of the vesicle or part of the vesicle itself. Moreover, the disc cavity of the retinal disc is prolonged as a central cavity through the optic stalk (see Fig. 13 and PI. IV.). In the pupa on the third day this disc has greatly increased in size, and its rela- tions are represented in Fig. 46. In this stage it forms an optic cup, the inner surface of which ultimately becomes applied to the inner surface of the optic disc (op d), and forms the retina of the compound eye. This disc and its stalk, the optic tract, form the greater part of the external portion of the de- veloping hemisphere. Sections through the optic cup and tract give very various and perplexing appearances, and it is only by a fortunate series that I was able to interpret its true structure. It will be at once apparent that tangental sections exhibit a closed sac, and, as the epithelial disc rt is folded and corru- gated, some are not readily understood. Again, it is impossible to get a single section which shows the canal in the optic stalk and the retinal disc, as the optic tract is curved. It is not so difficult, however, to demonstrate the further evolution of the retina from the retinal disc. As my own views on the structure and morphology of the compound eye are greatly at variance with those which are generally received, I must defer the further consideration of this subject until I give the full details of my investigations. In the meantime I would refer my readers to my published papers on the subject.* * Lowne, B. T., ' On the Compound Vision and the Morphology of the Eye in Insects,' Trans. Linn. Soc, and series, Zool., vol. ii., 1884. ' On the Struc- ture of the Retina of the Blow-fly,' Journ. Linn. Soc. Zool., vol. xx., 1889. THE DEVELOPMENT OF THE NYMPH. 327 The corpora fungiformia and antennal ganglia are differen- tiated from the inner or smaller portion of the hemisphere by the third day of the pupa, but the ventral cord between the infra- cesophageal and thoracic ganglia is not developed until the separation of the head and thorax is well advanced. Previously to the forward movement of the head these remain closely united with each other as in the larva. Weismann [2] observed the pyriform condition of the hemi- spheres in the pupa, and described the manner in which they are divided mto two parts by a groove ; and Viallanes [27] has given many particulars on the changes which the optic ganglia undergo, and some of his observations and figures lead me to believe that, if he had not been misled by received views, which, as I have shown elsewhere, have no substantial basis, he would probably have discovered the manner in which the retina is developed in Insects. Of the optic stalk (Stiel) which unites the optic disc to the hemisphere, Weismann says : ' It still appears on the fifth day of the pupa as a nervous cord, but on the twelfth day it can be no longer seen.'* He concludes, however, that it has by this time spread out into an invisible layer over the whole surface of the ganglion. That he should have arrived at such a conclusion is scarcely consonant with the general careful character of his work. If, as he states, and as is certainly the case, the optic stalk disappears entirely between the fifth and twelfth day, the opinion that the radial striae (which, he remarks, ap- pear later between the optic disc and the optic ganglion) are the nerve fibres which existed in the eye stalk is not based upon any evidence, and appears to me contrary to the observed facts. My observations lead me to the conclusion that the whole of the nerve fibres of the optic stalk disappear, and that nothing remains but a thin layer of connective tissue, uniting the capsule of the optic ganglion with the edges of the optic disc, within which the retina grows outwards until its outer surface comes into contact with the inner surface of the parts developed from the optic disc — my dioptron. * These times refer to Sarcophaga, which is a weclv longer in (he pupa than the Blow-fly. 22 328 THE DEVELOPMENT OF THE NYMPH. Development of the Peripheral Nerves.— -Weismaiin held it as probable that towards the end of the pupa stage, ' when the differentiation of the limbs into skin, muscles and nerves occurs, the newly-formed nerves come into relation with the limbs through the medium of the nervous portion of the imaginal pedicle ' Van Rees states that in every stage of the pupa which he examined the nerves were demonstrable, extending from the neuroblast to the tissue of the discs. This is my own view of the subject, although in the early stages these so-called nerves consist in the main of formative tissue, cells and fibres, but contain no true nerve elements. It is very difficult to ascertain in what manner the fibres are developed, but it appears pro- bable that they first appear at the central end of these conduct- ing cords, and grow towards the limbs, etc. ; but on this point I am guided entirely by analogy, and have at present no abso- lute evidence to offer. I hope, however, to have something more to say on the subject when I describe the nervous system, as I am at present engaged in some new researches on the subject. As, however, it is one presenting extreme difficulties, it is very probable that I may have to leave the development of the peripheral nerves an unsolved problem. d. The Development of the Integument of the Abdomen. Ganin [34] was the first to recognise that the abdominal integument of the imago is developed, like that of the head and thorax, from imaginal discs. He says : ' Weismann quite correctly states that the hypodermis of the eight abdominal segments of the larva is not destroyed, but he was wrong in supposing that the hypodermis of the larva is directly converted by the division of its cells into that of the imago,' a change which Weismann further states occurs at the end of the second or beginning of the third day. Ganin, it is said, discovered that the imaginal rudiments of the abdomen are already present from the first days of larval life ; yet Viallanes says, ' According to Ganin, the hypoderm of the larva becomes transformed into THE DEVELOPMENT OF THE NYMPH. 329 imaginal cells.' I am at a loss to reconcile this statement with the general tenor of Ganin's work. Viallanes supposes that the larval hypoderm is destroyed before the discs enclose the pseudo-yelk. In this I agree with \\\m, but only in a certain sense. Viallanes states that the pseudo-yelk is only covered by a fine cuticle before the discs grow over it. In this he was certainly wrong. Kowalevski, on the other hand, found that the discs grow over a cellular layer, which is only removed from the central area of the disc as its growth advances, and that they always overlap this layer by their edges. This view is also supported by Van Rees, but both authors regard it as the hypodermis of the larva. It is my paraderm. The Imaginal Discs of the Abdomen (PI. XVIII., Fig. 5, d) are two in number on each side of each segment, except the last, which only has a single pair of discs visible externally. The ventral pair lie one on each side of the anus, and are in- vaginated like the thoracic discs. This fact is noted by Kowalevski. The position of the discs is well seen in PI. XX., Fig. 4, which represents a very beautiful preparation made in the following manner. The pronymph was removed from the pupa-case, after heat-coagulation had been effected by boiling in water for a few seconds ; it was immersed for three hours in a o'l per cent, solution of hydrochloric acid, transferred to a mixture of glycerine and water, gradually strengthened until it contained 50 per cent, of glycerine, and afterwards stained in borax carmine to which an equal volume of glycerine had been added. By this treatment the imaginal discs are the only part of the integument which receives the stain, and their limits are beauti- fully seen. My specimen also exhibited a third and a larger more faintly stained region on each side of each segment, midway between the dorsal and ventral disc. These areas correspond with the large subcutaneous cells, which Wielowiejski termed oiuocytes (see p. 276). Van Kecs believes that each abdominal segment has a third 22 — 2 330 THE DEVELOPMENT OF THE NYMPH. dorsal disc on each side much smaller than the others, but I have found no trace of it. It cannot be doubtful that the two pairs in each segment correspond with the dorsal and ventral discs of the thoracic segments, from which they only differ in having no appendages and in their smaller size. The union of the abdominal discs occurs from before back- wards, and in the ventral median line, long before they unite in the dorsal region, so that even on the sixth day of the pupa the dorsal vessel can be seen pulsating beneath the transparent parablast, which still covers a considerable portion of the dorsal surface of the abdomen. e. The Pupa Sheath. On the fifth day the nymph may be regarded as completely formed. The wings, legs, segmentation of the body, and for- mation of the head, thorax and abdomen as distinct cavities, foreshadow the form of the imago, and resemble those of a lepidopterous nymph. But these parts exhibit, as it were, a rude outline of the perfect insect ; the joints are as yet only indicated by furrows, none of the setae are developed, and the embryonic epiblast of the imaginal discs has only partially replaced the paraderm in the dorsal region. The discs from which the abdomen is developed have not as yet coalesced with each other. Amongst the small epithelial cells of the discs other and larger cells have made their appearance ; these are the tricho- genic cells from which the setae are afterwards developed. At this stage the whole surface of the disc epithelium is seen to be covered by a thm chitinous cuticular layer, which during the next day or two separates from the underlying epithelium, and forms a loose sheath, which is not cast off until the fly emerges from the pupa-case. The formation of the nymph is, therefore, accompanied by two virtual ecdyses, that of the larval cuticle and that of the pupa-sheath, after which the cuticle of the imago is developed. Thus, as Weismann observed, ' we find three cuticular skins, THE DEVELOPMENT OF THE NYMPH. 331 one over the other, in the ripe pupa — the pupa-case, the pupa- sheath, and the epidermis of the imafjo.' The pupa-sheath is not, however, separated at once, but remains attached to the nymph at the inflections between the segments, and more especially between the head and thorax and the thorax and abdomen ; it becomes greatly thickened in these folds before its final separation (see PI. XXI.). Weismann first described the manner in which the pupa- sheath is formed. Reaumur was aware of its existence, but he thought that the rudimentary appendages are enclosed in it on their first appearance. Weismann correctly stated that when they first appear on the si'.rface they are not covered by any sheath. f. Changes in the Alimentary Canal. The changes which the alimentary canal undergoes in the first two days of the nymph stage (part of the third, fourth, and fifth days of the pupa) are of so extraordinary a character that I am greatly surprised to find that they had never been observed by anyone previously to my having undertaken their investigation. Kowalevski correctly described the formation of a new epithelial layer in the chyle stomach by the union of the imaginal rudiments discovered by Ganin, but his investigations cease with the beginning of the third day of the pupa state. He gives a description of a section in which a new epithelial layer surrounds the so-called corpus luteum [145, Fig. 21], which he regards as the shed epithelium of the larval chyle stomach ; but he fails to show what becomes of the remainder of the larval intestine, or in what manner the great saccular mesenteron of the nymph is formed. There are 34 mm. of hind-gut, metenteron, in the larva, and 30 of these do not reappear in the imago, in which the meten- teron only measures about 4 mm. Yet no one seems to have considered the impossibility of the total disappearance, not only of 30 mm. of intestine, but also of the great Malpighian tubes of the larva, without leaving a trace of their ever having 332 THE DEVELOPMENl' OF THE NYMPH. existed. By opening the pupa on the third day, or even after the legs of the nymph reach tlie middle or posterior third of the abdomen, it is easy to remove the v^fhole alimentary canal, and it will be seen that the hindgut is scarcely changed in length, and is almost exactly as it was in the resting larva, whilst the Malpighian tubules, increased in thickness and paler in colour, still exist, although their cells are much vacuolated and eroded. At this period it will be found that the intestinal coil formed by the metenteron and the Malpighian tubes lie in a cavity, bounded dorsally by the remains of the fat bodies of the larva and ventrally by the chyle stomach, which is strongly curved, its ventral surface convex and its dorsal surface concave (PI. XIX.). In sections (PI. XVIII., Fig. 2) it will be seen that the chyle stomach is covered by a layer of parablast cells on its ventral surface, and that these cells connect it with the remaining lobe of the fat body and surround the whole intestinal coil, between the origin of the Malpighian tubules and the posterior three or four millimetres of the gut. The cavity bounded by these parablast cells also contains a quantity of coagulated fluid. On the fourth day these changes are complete, and the chyle stomach is found to be split open on its dorsal aspect, so that its cavity is continuous with the provisional cavity in which the hind-gut and Malpighian tubes lie (compare PI. VIII., Fig. 2, and PI. XIX.). The cylindrical tube of new imaginal epithelium becomes thinner and thinner on the dorsal aspect of the chyle stomach, until at length it forms a crescent in section. The ends of this crescent grow upwards over the wall of the pro- visional cavity, and enclose the whole hind-gut and the remains of the Malpighian tubes, which form the so-called corpus luteum. The posterior part of the metenteron at the same time becomes converted into an impervious cord of parablast, which connects the temporary mesenteron with the posterior extremity of the nymph. This cord is eventually contracted ; thus it acts as a kind of gubernaculum, by which the new proctodseal rectum, formed by an invagination of the posterior segment of 'IHE DEVELOPMENr OF THE NYMPH. 333 the nymph, is brought into relation with the posterior extremity of a new metenteron formed by the elongation of the mesen- teron of the nymph. The nature and origin of the clear, coagulable fluid which surrounds the 'corpus luteum.'or degenerating larval intestine, is uncertain. It resembles blood, but contains no granule cells. The remains of the larval intestine are seen to be penetrated by and surrounded by leucocytes and multinuclear phagocytes, by which they are ultimately disintegrated. Crystalline sub- stances resembling leucin and tyrosin are also frequently present at a later period. The fluid is absorbed during the later stages of development. The chyle sto.nach, pro.ximal intestine, Malpighian tubes, and metenteron of the imago are all developed from the archenteron of the nymph. (See Alimentary Canal of Imago.) g. Origin of the Mesoderm. Mesenchsrme and Mesoderm. — Unfortunately, whenever a new term is introduced into science, it is years before it attains any definite meaning. A crowd of investigators seize upon it and apply it to every possible appearance which exhibits even a remote resemblance to that for which it was originally coined. Hence the most inextricable tangle of misconceptions originates. I shall not attempt to criticise the various meanings to which the term ' mesenchyme ' has been applied, but to define the exact meaning I ascribe to it in these pages. By mesenchyme, or parablast, I mean those groups of cells which have been at one time in their history wandering amoeboid corpuscles, but which may form a continuous rete or network, or an epithelioid layer, and which are either converted into connective tissue, return to their amoeboid form, or remain as amceboid cells. Whether these cells are budded off from the blastoderm, or whether they are the off- spring of yelk cells derived directly as amoeboids from the mother organism, may remain an open question ; but that they exist in the food yelk is beyond doubt, and unless they become 334 THE DEVELOPMENT OF THE NYMPH. connective tissues or endothelial cells, they arc as a rule temporary structures, and form no part of the adult organism. By mesoderm I mean cellular layers derived directly from the epiblast or hypoblast, or both, the cellular elements of which have never been wander-cells— amoeboid corpuscles— and from which, as a rule, the muscular tissues are developed. Whether there are muscle fibres which are developed from wander-cells is, I think, doubtful ; but it is possible that there are, as Hertwig maintains. Structure of the Imaginal Discs.— During the evolution of the imaginal discs, the details of their structure become far more apparent than in the earlier stages of their development, and since the issue of the first part of this work in i8go I have seen good reasons for modifying what I then wrote concerning the nature of the mesoblast of the discs (p. 77). What I then regarded as mesoblast I now hold to be parablastic tissue. In Plate XVIII., F"igs. 5, g, and 10, I have given details which were less distinctly seen in the younger discs which I then described. In the preparations represented it is perfectly easy to distinguish two layers of cells in the disc itself, a super- ficial layer of columnar epiblast and a deeper layer of small round cells, which I now regard as mesoblast, between the stellate parablast (which I formerly termed mesoblast) and the epiblast. From a re-examination of the discs at an earlier stage, I have come to the conclusion that a thin layer of small mesoblastic cells is present even in the youngest discs, whilst the stellate parablast is certainly developed from wander-cells during their evolution in the pronymph. Other groups of cellular elements are also developed during the evolution of the discs. Groups of outlying ovoid cells appear in the parablast around the tracheal vessels of the nymph. These have been described by Van Rees as the mesenchyme (mesoblast?) of the disc. I believe they are also parablastic cells, and that they are subsequently absorbed. Similar groups and strings of ovoid cells also appear in some sections between the stellate parablast and the small- celled THE DEVELOPMENT OF THE NYMPH. 335 mesoblast of the disc, and extend into the parablast. I have as yet been unable to determine the fate of these cells, and am doubtful whether they are of parablastic or mesoblastic origin. I am tempted to believe, from their arrangement in strings, that they are developed from the mesoblast, and are rudimentary muscle cells ; but as I am uncertain as to their nature I shall term them intermediate cells, from their position between the mesoblast of the disc and the parablastic plug which fills the provisional cavity. The stellate parablast extends over the outer surface of the provisional capsule, and when the latter is distended by the de- velopment of the disc, ultimately entirely replaces its epithelial elements. The various views which have been held as to the origin of the mesoblast of the discs are sufficient evidence of the diffi- culties which beset the investigation. I believe, however, that the recognition of the true character of the parablast, and its extreme importance in the developmental process, will do much to clear up many difficulties, and affords the clue to a recon- ciliation of many conflicting views. Viallanes everywhere figures and describes my parablast as the mesoderm of the discs, and naturally arrives at the follow- ing conclusion : ' M. Ganin thinks that the mesoderm in each disc is derived from the exoderm, but a certain number of new facts lead me to think this is not general, but that in many cases the mesoderm of the discs is jormcd from embryonic cells (blood corpuscles, etc.) scattered in the f;eneral body cavity.' Kowalevski, on the other hand, describes and figures my intermediate cells as mesoderm, and thinks they are developed from wander-cells. Van Rees entirely discards the term ' meso- derm ' and uses the word ' mesenchyme.' He says : ' As matters stand, I must declare myself in favour of Ganin's hypothesis, and hold that this tissue arises from the cells of the epithelium of the discs. It is on this account that I use the term " mesen- chyme " in the place of the old term " mesoderm," which is not only less definite, but incorrect, if we accept the views enunciated by the brothers Hertwig concerning its origin in their coelom 336 THE DEVELOPMENT OE THE NYMPH. theory.' What this means is beyond my comprehension, but it is clear to me that Van Rees' mesenchyme, in the regions from which the muscles are developed, consists of different elements, both parablastic and mesoderm cells. As to the origin of these cells, his opinions are, in my judgment, incap- able of being maintained, as he has not been able to distinguish the mesenchyme (parablast) from the mesoblast. Neither do I regard his account of the development of the wing muscles as consonant with facts which have been repeatedly observed by myself and others. Whether Ganin really recognised the mesoderm as the inner disc cells, as his statement as to their origin leads me to think, is a question I cannot settle, as I am only acquainted with his work from the translations of Hoyer and others, as already stated. h. Development of the Dorsal and Sterno-Dorsal Kuscles. Van Rees is of opinion that the dorsales and sterno-dorsales of the imago (' Brust Muskeln ') are developed directly from larval muscles. The development of the muscular system of the imago has hitherto presented great difficulties, and the most diverse opinions have been held as to its origin. Weis- mann thought that the muscles are formed from granule cells ; Kijnckel d'Herculais derives them from the discs ; Ganin from the disc mesoderm ; Viallanes retroverted to a view similar to Weismann's, and regarded them as developed from cells originating in the fat bodies ; and Kowalevski returns to Ganin's view. In the face of the confusion which has arisen from not defining the exact meaning of mesoderm, it is very difficult to understand how far Ganin's and Kowalevski's views correspond. Put it is clear that great differences of opinion exist on this point. Van Rees, on the evidence of several sections of the pupa in the first day, which show three pairs of dorsal muscles of the larva in the second thoracic segment, in a less degenerated condition than those around them, concludes that these become the wing muscles (dorsales). Accepting all Van Rees' facts, it THE DEVELOPMENT OF THE NYMPH. 337 appears that these muscles become surrounded by a vast num- ber of cells, parablast or mesenchyme, and that this mass of cells persists and forms the nidus in which the wing muscles appear. There is nothing to show that one particle of the larval muscle remains. I am far from denying the facts alleged by Van Rees, but there is no question that the wing muscles are developed from mesoderm cells which grow from the meso- thoracic disc, and that these grow into a mass of parablast. That this cellular parablast consists of leucocytes, which ac- cumulate around certain dorsal muscles of the larva, is possible ; but the larval muscle fibres described by Van Rees as the last to degenerate in the thorax form no part of the new muscles. The dorsales and sterno-dorsales are at first minute and slender strings of cells imbedded in an abundant parablast. They first appear on the third day of the pupa state as six, not three, bands of cell tissue, chiefly parablastic in origin, of which nothing remains in the end except the nuclei, with, per- haps, some remnant of their protoplasm giving off long stellate processes, which form a connective reticulum for the support of the new muscle fibres. The development of the dorsales and sterno-dorsales in Chironomus can be distinctly watched in the living animal, and it is clear that, although developed between the fibres of the larval muscles, the latter take no part in their formation. I must hold, therefore, that the develop- ment of the muscles of the imago is not from those of the larva, although it scemslikelythat the vast numbersof leucocytes which surround these muscles may form coherent parablastic tissue into which the mesoderm cells of the discs grow. It may e\en happen, in the case of the wing muscles, that fragments of the larval muscles sometimes remain longer than in that of other muscles, although I have not observed the fact ; but the new muscles neither originate from the nuclei of the old ones, nor in their substance by the transformation of existing muscle tissue. Neither do they arise from the leucocytes which sur- round the larval muscles. They are developed in every case from cells which grow from the discs themselves, and the para- blast which surrounds these cells forms at most a connective 338 THE DEVELOPMENT OF THE NYMPH. reticulum in which the muscles lie — a connective reticulum which, like the connective tissues generally, is, therefore, of parablastic origin. i. The Tracheal System of the Nymph. The investigation of the changes which occur in the tracheal system in the nymph and pronymph stages is most difficult. The vessels are not easily traced in sections, in which of course no air remains after successful imbedding, and it is not until the nymph stage is far advanced that the tracheal vessels are easily demonstrated. The earlier stages of their development have been described by Weismann, and the methods he adopted — those of ordinary dissection — although far from satisfactory, throw more light on the process than the newer methods of section-cutting. I am inclined to agree in most points with Weismann, but confess that I should feel more satisfied if my sections showed the conditions, which he described, clearly. Such rough dissections as can be made bear out Weismann's statements, and the difficulty of tracing the tracheae in sections in the early stages of the pupa is so great that, until some further methods of research can be invented, I fear little pro- gress is likely to be made. In pupae a few days older, however, the evidence of sections is unequivocal. In the earliest sections of pupae I have only found the large tracheal trunks in various stages of degeneration and re-forma- tion, with an occasional group of small twigs cut through transversely, which are recognisable by their bright, highly- refractive intima. Weismann says : ' The tracheal system of the pupa is very peculiar, and differs not only from that of the larva, but from every known variation of the respiratory system in insects. Only a small portion of the tracheae of the pupa is formed in relation with those of the larva ; the greater part is developed independently. ' Two great respiratory trunks are common to the larva and THE DEVELOPMENT OF THE NYMPH. 339 pupa, but there is this difference, whilst those of the larva extend the whole length of the body and end in a stigma at each pole, those of the pupa are short and possess only anterior openings, the stigmatic cornua. From these the main trunks extend backwards for a short distance, and then suddenly break up into small twigs, which without further sub- division form a tuft comparable with a horse-tail (' Pferde- schvvanz ') and float freely in the pseudo-yelk' [2, p. 169]. It is these vessels which were first observed by Viallanes in the rudimentary wing, to which allusion has already been made (p. 160). They do not last long, but soon disappear, or only a small number remain, as I have sought for them in vain on the fifth or sixth day of the pupa. Weismann adds : ' The main trunks in the nymph give off side branches, and are united in front by a transverse vessel.' These are clearly the vessels to the several discs, and are developed from the trachea; of the larva, which are distin- guished by the small embryonic cells of their external coat (see p. 85 and PI. IV.). The transverse commissure does not unite the main trunks, as Weismann thought, but the branches to the head discs, which become the anterior extremities of the main trunks of the nymph. The separated intima of the larval trachea; is shed by being withdrawn through the stigmatic cornua during their evertion, and remains attached to the pupa-case. Almost directly after this the intersegmental spiracle of the nymph is seen to be connected with a diverticulum of the wall of the main trunk close behind the stigmatic cornu. During the separation of the pupa-sheath from the cellular ectoderm of the nymph, the intima of all these trachea; separates from the cellular wall, a distinct space filled with fluid intervening, and soon after this a new intima appears outside the old one. This is the intima of the trachea; of the imago. The intima of the tracheae of the nymph, the stigmatic cornua, and the intersegmental spiracles of the nymph separate with the pupa-sheath, but the shed intima of the tracheje is not withdrawn from the trachea; of the imago until the latter 340 THE DEVELOPMENT OF THE NYMPH. draws itself out of the pupa-sheath as it escapes from the pupa- case. The Pro-imago. — It is evident that the separation of the pupa- shonth and of the intima of the tracheae is a virtual ecdysis. From this period the nymph becomes gradually transformed into the imago. I regard the nymph as completely formed as soon as this ecdysis has occurred, but as all the subsequent events are gradual developmental processes, it is impossible to say that either the nymph stage ends or the imaginal stage commences at any time, unless the actual ecdysis of the pupa- sheath and larval skin be regarded as the commencement of the imago state. To avoid circumlocution, it will be exceed- ingly convenient to term the nymph, from the shedding of the pupa-sheath to the escape from the pupa, the pro-imago, and to regard all those parts which exist at the time of the separa- tion of the pupa-sheath as parts of the nymph, and all those subsequently developed as parts of the pro-imago. For example, I shall speak of the tracheal system as it exists before this period of virtual ecdysis as the tracheal system of the nymph, and the new tracheal system which is subsequently formed as the tracheal system of the pro-imago. k. The Dorsal Vessel and Coelom. The dorsal vessel of the imago is apparently developed directly from that of the larva. At every stage of development I have always found the dorsal vessel intact. Herold [135] observed that in the Lepidoptera the pulsations of the dorsal vessel continue throughout the whole period of the pupa sleep. Newport confirms this in Sphynx ligustri, as he asserts that the vessel pulsates throughout the whole pupa stage, although its beats almost cease during hibernation, a period, however, in which development is also at a standstill. Kiinckel d'Herculais saw the heart beating in the pupa of Eristalis until the eighth or ninth day, after which he states that it stops or beats feebly for a day or two, but that after- wards it pulsates regularly. THE DEVELOPMENT OF THE NYMPH. 341 Kowalevski [145] states that in the Blow-fly pupa it is seen pulsating on the third day as it does in the larva, but that later its pulsations are irregular, and he observed that the anterior and middle portions of the dorsal vessel remain almost unchanged during the whole pupa period. Weismann, on the other hand, concluded that in the Muscidfe the dorsal vessel degenerates and is rebuilt in the pupa. He says: 'Although direct observations on the cessation of its pulsations have not been made, it may be concluded from its changes of histological structure that after a certain time con- tractions are no longer possible.' ' In the first days of the pupa state the dorsal vessel remains unaltered. The histolysis of the great tracheae of the larva, to which its alar muscles are attached, removes its point d'appui. Its position is still, however, in the middle line of the back, although it is probably much folded by the shortening of the body of the larva. The ring (see p. 78), also degenerates on the fourth or fifth day, and removes its anterior attachment. The isolation of the vessel at this period is exceedingly diffi- cult, as it has become very fragile, and is evidently in the first stage of histolysis. As an organ, it is not broken up, but is redeveloped by a similar process to that which has been observed in the intestine and Malpighian vessels.' The main point to which I would draw attention in Weis- mann's statement I have printed in italics. The rest is not entirely accurate. The ring is not lost as a point of attach- ment, although it is profoundly altered. As an epithelial struc- ture it no longer exists, but a vast number of cells appear around it before it disappears, and these cells form a dense network which remains apparent in the pro-imago, and is a conspicuous object in sections on the tenth day of the pupa, or even later. This network of cells lies in the nymph above the hemi- spheres, and subsequently behind them, and appears to travel back over their surface as the head is developed. The muscles which form the alas of the pericardium are, it is true, removed, and a pericardium can no longer be said to 342 THE DEVELOPMENT OE THE NYMPH. exist in the nymph ; yet I see no reason to believe that the dorsal vessel ceases to pulsate, and in the numerous sections which I have examined I can see no evidence of any change in structure. It remains the same as in the larva until the last two or three days of the pupa state, and does not exhibit any traces of degeneration. Kowalevski says that when treated with osmic acid the boundaries of its constituent cells (muscle cells) become more distinct, and that its transverse striations are less marked on the third day of the pupa than in the larva. As, however, the demonstration of the muscle cells of the dorsal vessel by osmic acid is very uncertain at all stages, and the muscular striations vary in distinctness in different prepara- tions, I am not inclined to regard the above statement as important, and Kowalevski leans to the opinion that the dorsal vessel does not undergo histolysis. The last-named author erroneously supposed that the dorsal vessel is more deeply seated in the larva than in the pupa ; in the posterior segments of the larva, as already stated, it is placed immediately beneath the integument. This portion is lengthened during the pupa stage, whilst the aortic section, which is deeply seated, becomes shortened. Although I believe that the dorsal vessel persists during the whole period of the pupa sleep and performs its function, its form is so changed during the last few hours of this period that the dorsal vessel of the imago cannot be regarded as identical with that of the larva ; and I think it probable that all its muscle cells are re-formed from embryonic cells, and gradually replace the muscle cells of the larval heart, since on the tenth or eleventh day of the pupa the dorsal vessel is lined by a double row of musculogenic cells, and from this period its lumen and the thickness of its walls rapidly increase. The Coelom, Cell Chaplets and Pericardial Cells. — The coelom in DeSCRII'TION 01' I'LAl'K XXI. A median sagittal section of a male nymph from a pupa seven days old, showing the relations of the nerve centres, dorsal vessel, stonioda.'iim, archenteron, and recto- cloacal pouch. The coelom is almost entirely occupied by cellular elements derived from the fat bodies of the larva. THE DEVELOPMENT OF THE NYMPH. 343 the nymph is a continuous cavity between the body wall and the newly-formed alimentary canal, which contains the pseudo- yelk. It soon, however, becomes permeated by a parablastic network which binds together the alimentary sac and the em- bryonic rudiments from which the internal organs of the imago are developed. I believe this network is formed from the cells which cover the extremities of the young tracheal vessels. The cell chaplets, which form the fringes of the dorsal vessel, and some at least of the great pericardial cells of the larva, remain and retain the same relative position as in the larva. The pericardial fringes grow rapidly by the multiplication of the cells of which they are formed. Their function and their ultimate destiny are, however, unknown. Kowalevski [146] discovered that if the larvas are fed with flesh impregnated with carmine, methyl blue, or silver salts, the peri- cardial fringes, the great pericardial cells, and the cell chaplet of Weismann become intensely coloured, and that the cells of the pericardial fringes remain coloured, not only during the whole pupa stage, but for some days after the imago emerges from the pupa. These cells multiply so rapidly, according to the above-named author, that they form a network which clothes the whole of the posterior part of the mesophragma. He also observes that the cell fringes of Weismann disappear soon after the histolysis of the salivary glands is complete, and that only seven pairs of pericardial cells remain in the imago. The six posterior pairs degenerate. Flies fed with syrup coloured with cochineal, according to Kowalevski, soon exhibit coloration of the pericardial cell fringes. I have frequently found pericardial cells in the nymph on the third day scattered amongst the constituents of the pseudo- yelk of the posterior part of the abdomen. These undergo histolysis like the fat bodies, and are probably the cells from the posterior part of the pericardium. The peristent pericardial cells are probably the elements from which the alar muscles of the pericardium of the imago are developed, as I find them embedded in the substance of their fibres. 23 344 THE DEVELOPMENT OF THE NYMPH. 3. THE DEVELOPMENT OF THE IMAGO FROM THE NYMPH. From the fifth day of the pupa state to the escape of the imago. The development of the nymph, as already stated, may be regarded as complete when the pupa sheath is separated from the subjacent cellular layer. From this period until the imago emerges from the pupa, the process of development may be conveniently divided, as Weismann [2] suggested, into two stages. The first commences on the fifth day, and terminates Fig. 47. — A section through the .iljdominal integument of a nymph from a pupa seven clays old: e f, outer celKilar layer of small epiblast cells ; /(, inner cellular layer of hypoderm and trichogenic cells. The cuticular layers are developed l)etween the outer and inner cells. about the end of the seventh. During this period the external form of the nymph undergoes rapid evolution, and by the end of the seventh day differs but little from that of the young imago when it is ready to escape from the pupa-shell. The integumental setae are developed from the large trichogenic cells, and the small cells of the epiblast between them increase in number so rapidly that the whole integument becomes minutely corrugated, the great seta; arising from the ridges. The hollows of the rugae are occupied by minute setie which DEVELOPMENT OF THE IMAGO FROM THE NYMPH. 345 closely resemble cilia. These form the down-like hairs which cover the body of the imago (Fig. 47). Subsequently the cellular integument is seen to consist of two layers of cells : a superficial layer of very small ones, which ultimately become converted into a chitinous membrane ; and a deep layer of large cells, the trichogenic cells beneath the great setas, and of medium-sized cells derived from them, which form the hypodermis and the deeper layers of the cuticle de- veloped after the escape of the imago from the pupa (see p. 280). The corrugation of the integument, which is especially marked in the abdominal region, permits of its expansion after the escape of the imago from the pupa-case. By the end of the seventh day the articulations between the limb segments begin to assume definite characters, and the pads and claws of the tarsi are formed from their terminal lobes. The larger wing nervures are present as distinct ridges, and the oral Icbes of the proboscis are so completely developed that the similarity of the mouth to that of the Hemiptera, so marked at an earlier period, is no longer obvious. The second period, from the end of the seventh day to the escape of the imago from the pupa, is marked by scarcely any changes of external form, but is chiefly remarkable for the rapid development of the internal organs, which make very slow progress during the first period. Throughout this second period the development of the contents of the head- capsule exhibits a marked advance over that of the thoracic and abdominal organs, and the contents of the thorax are further advanced than those of the abdomen. Whilst the thoracic ganglia remain comparatively rudi- mentary, the brain exhibits great complexity, and scarcely differs from that of the adult. The least developed organs of the head, on the tenth day of the pupa, are the compound eyes, in which pigment first appears on the ninth or tenth day ; but the simple eyes are already fully developed, and are ap- parently more perfect than in the adult imago. These organs in the pupa are more like the simple eyes of spiders than those of the adult insect. 23—2 346 THE DEVELOPMENT OF THE NYMPH. The stomodaeum and proctodaeum are also far more ad- vanced than the mesenteron ; this on the tenth or eleventh day is a mere thin-walled sac, with a coiled csecal prolongation of its posterior extremity, which becomes the metenteron. At the junction of the latter with the mesenteron the Malpighian vessels are seen as short caeca ; they are chiefly developed during the last two days of the pupa stage, but the rectal papillas are almost as well developed on the tenth day as in the perfect insect. The thoracic muscles and ganglia are very imperfect at this period, and Weismann observed that if the nymph is removed from the pupa-case even on the eleventh day, it exhibits no trace of muscular movement. The great size of the blood sinuses and the late redevelopment of the dorsal vessel have been already referred to. The tracheal system of the imago remains very inconspicuous until after its escape from the pupa, so that Weismann states that, ' of all the organs, the trachea; are the last developed ' [2, p. 235]. In this he was wrong, as the principal tracheal trunks of the fly are all present on the ninth or tenth day, although the smaller vessels are apparently developed later. The sexual glands form a marked exception to all the other internal organs, as they exist in a rudimentary condition in the young larva, and probably in the embryo, and undergo progressive development, which is complete in the male a few hours after its escape from the pupa, and in the female, only some weeks later. APPENDIX TO CHAPTERS VI. TO IX. METHODS OF STUDY. No great difficulties have been experienced by me in the preparation of sections of either embryos or nymphs, except in the early stages of the pupa, when the pupa-shell cannot be removed. The difficulty in this case arises partly from the imper- meability of the pupa-shell, partly from the fact that paraffin will not adhere to it, and partly from the extreme friability of the sections— perhaps due to the difficulty of fixing the tissues owing to the impermeability of the pupa. In preparing embryos, it is advantageous to remove the chorion, or egg-shell, after heat coagulation, but I have pre- pared very good sections occasionally when this has not been done, staining them after cutting. I have found it impossible to stain them in mass unless the shell has been removed. Young pupa; should be divided with a razor after heat coagula- tion, and the extreme ends of the pupa case cut off. In pupae after the third day the whole shell must be removed. I have found collodionising the sections advantageous in young pupae. This is easily effected by painting the cut surface of the paraffin block with thin collodion, which dries almost instantaneously. As the study of the development in the egg and pupa necessi- tates the preparation of a large number of sections, the process of staining these after they are cut, although it gives the best results, is far too laborious ; it is necessary, therefore, to have a good method of staining en masse. 348 APPENDIX TO CHAPTERS VI. TO IX. The best method is Lang's picro-carmine and eosin.* The eosin in some way acts as a carrier for the carmine, and is afterwards washed out with 70 per cent., and then with abso- lute alcohol. It is impossible to over-stain, and the nymph should be left from four days to a week in the stain. My friend, Brigade-Surgeon Scriven, has been indefatigable in preparing serial sections of nymphs, which he has very kindly placed at my disposal, and he adopts the above method most frequently. His specimens are very beautiful and definite, and I take the present opportunity of thanking him for his valuable assistance. Viallanes used collodion as a material for imbedding the nymph, and I have obtained very fair results from its use. I am not, however, prepared to recommend it for serial sections, and I have not found it to possess any advantages over paraffin. When celloidin or collodion are used, either for imbedding, or for collodionising paraffin sections, oil of cloves should not be used as a clearing agent. Equal parts of xylol and solid carbolic acid may be employed instead, but the slides must be well washed with xylol to remove all traces of carbolic acid before they are finally mounted with balsam. The details of these processes and much valuable information on technique will be found in the second edition of the ' Micro- tomist's Vade-Mecum,'t which is a far better work than the first edition, from which I formerly quoted. The outer covering of the pupa or embryo is easily refiioved with a little practice. Perhaps the best method is to cement the eggs or pupje to a slide with gum or shellac in creasote ; the outer covering can then be cut with a sharp needle, and the embryo or nymph removed. This must be done under dilute alcohol — 50 per cent, is sufficient — with a dissecting micro- scope. * Equal parts of a solution of picro-carmine, Wcigert's or Ranvier's, and a 2 per cent, aqueous solution of eosin. t 'The Microtomist's Vade-Mecum,' by A. Hollos Lee, second edition, London, 1890. APPENDIX TO CHAPTERS VI. TO IX. 349 I then make a drawing of the external form of the cmbr3o or nymph, which is subsequently a great aid in the interpreta- tion of sections, and transfer the preparation to 75 per cent., and after an hour or two to absolute alcohol. In a week or ten days the preparation should be transferred to the staining solution. I think it is advantageous to place specimens intended to be stained with picro-carmine in a solu- tion of picric acid in 50 per cent, alcohol for a few days, before transferring them to alcohol. In preparing sections of the imago, it is necessary to cut the insect into two or more parts, or to remove a portion of the integument, and then to fix the tissues either in absolute alcohol with ten to twenty drops of a solution of osmic acid to half an ounce of alcohol, or to place them in a solution of picric acid in 75 per cent, alcohol. Fixation by heat is inadmissible, as the trachea; swell and displace and vacuolate all the tissues. It is most important to bear in mind that specimens hardened in picric acid must never be wetted with water subsequently. I have not found that such dilute osmic solution as I have recommended prevents subsequent staining with hematoxylin. For logwood-staining after cutting, I consider Miiller's fluid far the best fixative for imagos. A week or ten days is suffi- cient before transferring to alcohol. Such specimens must be washed in large quantities of dilute alcohol, 50 per cent, to remove the chromates. Whenever it is possible, young imagos should be used, as the chitin of the adult is most difficult to cut. I have, however, succeeded in cutting egg-bearing females without serious frac- ture of the integument. To do this it is necessary to be very careful that the temperature of the paraffin used for imbedding does not rise even for a few minutes much above its melting- point, and any such rise of temperature is always most destruc- tive. Neither must the insects be kept in absolute alcohol more than a few days previously. Such insects should, I think, always be fixed with Miiller's fluid. I have found eau-de-Javelle and eau-de-Labarraque, so much extolled, exceedingly destruc- tive to the internal organs, so that they are quite inadmissible. 350 APPENDIX TO CHAPTERS 17. TO IX. The rapid fixation of the tissues of the imago, so essential to good results, can only be insured by first washing the whole insect in alcohol to remove the waxy secretion from the integu- ment, unless absolute alcohol is used as a fixative. An exhaust- ing syringe is useful to assist the permeation of the insect by the fixative fluid, but, if alcohol be used for washing, is not essential ; indeed, of late years I have not used it. Benzoline has recently been recommended as a substitute for chloroform before imbedding in paraffin. I cannot, however, recommend it, as I am sure it is more destructive to the tissues than chloroform. It has also been recommended instead of turpentine for removing the paraffin from sections. It certainly does so more rapidly, but I consider its use dangerous, and, although cheaper than turpentine, I do not think it should be employed. Xylol is equally dangerous, as its vapour is most inflammable. When used, it ought to be kept in a small bottle, and employed with great caution. A serious accident may readily occur from any carelessness. v> «««r, ^' '•■ ^