CLE P2L5 S>54 iSl^S'rli: Cornell University Library QE 779.P25S5 SHurian stromatoporoids of America^ 3 1924 004 578 898 CORNELL UNIVERSITY LIBRARY ^^..... ,, f^S^-. UNIVERSITY OF TDK^NTO STUDIES GEOLOGICAL SERIES No. 6: SILURIAN STROMATOPOROIDS OF AMERICA. By W. A. PARKS THE UNIVERSITY LIBRARY, 1909: PUBLISHED BY THE LIBRARIAN UNIVERSITY OF TORONTO STUDIES COMMITTEE OF MANAGEMENT Chairman: Robert Alexander Falconer, M.A., LittD., LL.D., D.D., President of the University Professor W. J. Alexander, Ph.D. Professor W. H. Ellis, M.A., M.B. Professor A. Kirschmann, Ph.D. Professor J. J. Mackenzie, B.A. Professor R. Ramsay Wright, M.A., B.Sc, LL.D. Professor George M. Wrong, M.A. General Editor; H. H. Langton, M.A., Librarian of the University SILURIAN 8TR0MAT0P0R0ID8 OF AMERICA (EXCLUSIVE OF NIAGARA AND GUELPH) BY WILLIAM ARTHUR PARKS, Ph.D. ASSOCIATE PROFESSOR OP GEOIOGY UNIVERSITY OF TORONTO 56- -f / r«? ? ^/ e PREFATORY NOTE The present paper forms the third of a series of articles on the Stromatoporoids of America. The first and the second con- tributions are known respectively as No. 4 and No. 5 of the Geological Series of University of Toronto Studies. The above papers deal with the forms from the Guelph and Niagara forma- tions. The present article is divided into three parts. Part I contains a description of the Helderbergian forms. Part II treats of the examples from the Silurian area around Hudson Bay, which must be regarded as of about the same horizon as the Niagara. Part III deals with certain Niagara forms which have come to the notice of the author since the appearance of Niagara Stromatoporoids. It is hoped, therefore, that the three pamphlets now issued will together include all the forms known from the Silurian of America. Whatever merit this work may possess is, in large part, due to the kindness of the gentlemen who have afforded me an op- portunity of studying the material in their collections. To the following palaeontologists, and to the institutions which they represent, I therefore wish to express my most sincere thanks :— Dr. J. F. Whiteaves and Dr. H. M. Ami, Geological Survey of Canada. Dr. J. M. Clarke, State Geologist, State of New York. Dr. Charles Schuchert, Yale University. Dr. Ray S. Bassler, United States National Museum, Wash- ington. Dr. F. A. Bather, South Kensington Museum, London. Dr. E. O. Hovey, American Museum of Natural History, New York. Mr. T. E. Savage, Illinois State Geological Survey. In addition to the literature already referred to in the intro- duction to Niagara Stromatoporoids, mention must be made of the excellent description of Helderbergian Stromatoporoids, by [243] 4 Prefatory Note Dr. George Herbert Girty, in the second volume of the 48th Re- port of the Regents of the New York State Museum, 1894. The only other original description of a Helderbergian form is con- tained in Stromatoporidae of the Upper Silurian, by Dr. J. W. Spencer. (Bull. Mus. University of Missouri, 1882.) The literature bearing on the forms from Hudson Bay has already been listed, vv^ith the exception of a short preliminary article by myself in the 22nd volume of the Ottawa Naturalist, 1908. [244J SILURIAN STROMATOPOROIDS OF AMERICA I. Helderbergian Stromatoporoids The Helderbergian rocks of the State of New York furnish a most interesting series of " Milleporoid " Stromatoporoids, which are closely related to one another and to certain forms be- longing to the coralline limestone of the Niagara. In many of these specimens the ultimate fibre is excellently preserved, and yet the gradation from one form to another is so imperceptible that some doubt may be expressed as to their specific differentia- tion. While it has been thought advisable to retain all the forms previously described, it must be admitted that new species could be easily created which would stand on as firm a founda- tion as some of those herein included. The variation of struc- ture in different parts of the coenosteum of the same specimen is so great that care must be exercised in creating new species lest the multiplication of forms results in making none of the species determinable. The most important feature revealed by the study of these forms is with regard to the minute structure of the fibre. Stu- dents of the Stromatoporoids will recall the work of Baron von Rosen, Ueber die Natur der Stromatoporen, in which the affinities of the group to the sponges is strongly maintained. Plate I, figure 2, 3f that work exhibits a vertical section of Stroinatopora typica, in which the ultimate fibres of the skeletal matter are arranged in horizontal and vertical rows so that a rectangular network results. Von Rosen interprets these ele- ments as sponge spicules. Dr. Nicholson, in his description of 'Actinostroma astroites* refers to this plate, but he interprets the appearance as being due "to a sort of decomposition, in conse- *British Stromatoporoids, p. 145. [245 J 6 Parks : Silurian Stromatoporoids quence of which the thick and reticulated skeleton-fibre becomes broken up into innumerable minute, dark-coloured, vertical and horizontal lines." Precisely the appearance represented in Von Rosen's figure is shown by examples of Stromatopora constellata from the coralline limestone of New York; it is also to be ob- served in Syringostroma ristigouchense from Cap Bon Ami, and it is presented in greater or less perfection by all the examples from the Helderbergian of New York. It is impossible to be- lieve, in the face of this e\'idence, that such a widespread struc- ture is the result of mineralization. In the species Actiiiostroiiia astroites, Rosen sp., we have, according to Nicholson, an example of the genus in which the radial pillars occur to the number of twelve or fifteen in the space of one mm. Nicholson comments on the similarity of this structure to that of the minute fibre of Stromatopora typica, but, as above stated, he believes that the resemblance is only acci- dental and due to the alteration of the latter species. In this connection it is interesting to record the discovery of a similar species growing beneath a colony of Stromatopora constellata. The radial pillars are a> numerous as 24 to the mm. With their connecting arms, tliey build up a rectangular network of exactly the same size of mesh as the o\"erlying Stromatopora constellata; in fact, the cocui'steum is exactly the same as that of the latter species, but -K'ithout the astrorhisal and vertical canals. We. have therefore two facts established beyond question: — first, the existence of a rectangular arrangement of fibre in a large number of Stromatoporoids ; second, the occurrence of precisely similar fibre without the \ertical and horizontal canals. If this latter form is the skeleton of a colony of zooids, each individual must have been less than the 24th of a mm. in diameter. Such ex- treme tenuity is hardly to be imagined in the case of any Coelen- terate. The genus Stromatopora, at least as far as the present examples are concerned, is e\-idently nothing more than an ex- tremely fine Actiuostroma pierced by numerous vertical tubules — the so-called zooidal tubes. The above considerations and many other facts which have recently come under my observation tend strongly away from Ni..h;ilson's conclv.sii.ns as to the Hydrozoan affinities of the [246] Parks : Silurian Stromatoporoids 7 Stromatoporoids. If the position taken above with regard to the relation of Actinostroma to Stromatopora can be more fully established by further evidence, then the tvv^o divisions of " Mille- poroid " and " Hydractinoid " types must be abandoned as the main basis of classification. Nicholson's argument in support of the Hydrozoan character of Stromatoporoids rests largely on the interpretation of the horizontal partitions in the vertical tubes as " tubulae." I am convinced that, in some cases at least, these partitions result from the continuation of horizontal strands of skeletal fibre across the pore. I do not, however, consider that enough evidence has yet been obtained to justify such a sweeping change in the method of classification as must result from a complete proof of the principles suggested above. ORDER— STROMATOPOROIDEA, Nicholson and Miirie Section A. (" Hydractinoid " Group) Fa»«:/y— ACTINOSTROMIDAE, Nich. Genus — Clathrgdictyon^ Nich. and Murie. Clathrodictyon jewettIj Girty — Plate XVI, Figs, i and 2. Clathrodictyon jewetti, Girty, Rep. Mus. State of New York, 48th, vol. ii, p. 298, pi. vi, figs. S and 6, 1895. Dr. Girty's description is as follows : — " Coenosteum known only as a small fragment. Latilaminate, possibly massive. Sur- face characters not known. Laminae very conspicuous in vertical section. Radial pillars persistent through one interlaminar space, usually terminating in the laminar partition. Tangential sec- tions show the cut ends of the radial pillars, which are apparently not connected by arm-like processes, but by a continuous wall. Monticules are present associated with astrorhizae and axial tubes, but the character of the astrorhizae has not been deter- mined. The skeleton does not show the ' pillar and arm ' structure characteristic of Actinostroma, nor the reticulated skeleton common in Syringostroma. " The species resembles C. striatellum, Nicholson, in many particulars, but the latter is without astrorhizae, monticules, etc., [247] 8 Parks : Silurian Stromatoporoids which form a striking- feature in the Actinostroinidae. It differs from Nicholson's species also in the more normal pillars and regular mesh. All the sections studied have been cut from a small fragment which alone represents the species. In none of them do the radial pillars extend through a number of inter- laminar spaces, and although an apparent prolongation through two or even three such spaces may occasionally be observed, the occurrence is too rare to influence the generic determination of the specimen." To this description there is little to add. I have prepared two sections which are here produced (Plate XVI, figs, i and 2). The astrorhizal systems are very inconspicuous, and there is no direct evidence as to the ultimate structure of the skeletal matter. It has an indistinct granular appearance, but is very much min- eralized. There is no sign of the double-based pillars charac- teristic of C. stnatcllum. and the fibre is coarser than in that species. The spacing of the laminae is very like that of the form described by myself as C. rectum from the Niagara of Louisville, Ky., but the pillars are not so strictly vertical and are less regular in their development. Locality — Lower Pentamerus limestone, Cedarville, N.Y. The type specimen is in the Peabody Museum, Yale University. Section B. (" Milleporoid " Group) Fa/Mi/y— STROMATOPORIDAE, Nich. Genus — Syringostroma, Nicholson. In view of Dr. Girty's objections to the generic relations of certain species, it seems necessary to define the genus Syringo- stroiua somewhat more closely than has hitherto been done. There can be no doubt that Nicholson's general conception of the genus was that of a " Milleporoid " type, in which the reticulated tissue is aggregated into distinct vertical pillars, connected at intervals by laminae. The genus Actinostroma, on the other hand, consists of individual vertical fibres, or dense pillars, con- nected by laminae, which arise as whorls of arms developed from the pillars. The essential difl^erence is, therefore, that the pillars of Syriugostroma are reticulate and those of Actinostroma are [248] Parks: Silurian Stromatoporoids 9 not. As already pointed out, the skeleton-fibre of certain species of Stromatopora is precisely the same as that of a fine-grained Actinostroma. It may be said, therefore, that Syringostroma is a sort of compound Actinostroma — a Stromatopora in which the vertical canals are so arranged that they leave round pillars of tissue between them, thus simulating the structure of Actino- stroma. It is obvious, therefore, that transitions will naturally occur between Syringostroma and Stromatopora. In fact, the whole series of forms described by Dr. Girty are intermediate in character. They all show the pillar-like structure in vertical section, but in tangential section it is less pronounced. Could this appearance not be revealed at all, it would be necessary to remove the whole series from the genus. Thin sections, however, show more or less distinctly the severed ends of these pillars, but in one species at least so indistinctly that it seems better to remove the form to the genus Stromatopora. Bearing in mind that the pillars and laminae are merely parts of the same general tissue, we have two variables — the thickness of the laminae and their distance apart. If the laminae are thin, far apart, and connected by round pillars, we have a typical Syringostroma. On the other hand, if they are thick and close together, so as to be separated by the ramifications of the astrorhizal canals only, we have a typical Stromatopora. The following species are described by Dr. Girty, who ascribes them all to Syringostroma — 5". ccntrotum, S. consimile, S. microporum, S. barrettl, and 5". foveolatnm. In the order named they show less and less of the typical Syringostroma structure and approach more and more to the Stromatopora type. I propose to remove the last-named species to the latter genus, but it must be admitted that the dividing line is difficult to draw. Of Dr. Girty's species I find three easily distinguishable types — 5". centrotum, S. foveolatuin, and a type of which 5". consimile is the most important represen- tative. The practical distinction of S. consimile, S. barretti, and .S'. microporum I find to be a matter of great difficulty and of doubtful value. [249] 10 Parks: Silurian Stromatoporoids SvRiNGOSTROMA RiSTiGoucHENSE, Spciicer, sp. — Plate XVI, Figs. 3, 4, and 5. CoENOSTROMA RISTIGOUCHENSE, Spencer, Bull. Mus. Univ. State of Missouri, p. 49, pi. vi, figs. 12 and 12a, 1884. Syrincostrgma RISTIGOUCHENSE, Nicholson, Mon. Brit. Strom., p. 97, pi. xi, figs. II and 12, 1886. Syrincostrgma ristigguchense, Nicholson, Ann. and Mag. Nat. Hist., April, p. 324, pi. viii, figs. 6-8, 1891. Actingstroma RISTIGOUCHENSE, Girty, Rep. State. Mus., N. Y., 48th, vol. ii, p. 298. Syrincostrgma ristigguchense, Whiteavcs, Can. Rec. Science, p. 138, i8g6. Dr. Nicholson's description is as follows : " The coenosteum in this species is massive, more or less definitely latilaminar in growth, and readily splitting into thick strata concentric with the surface. The laminae are gently curved, and there are either no astrorhizal eminences (mamelons) or but very inconspicuous ones. Astrorhizae are, as a rule, \'ery well developed, being of large size and much branched. They are arranged in vertically superimposed systems, and have their centres about one cm. apart. " Vertical sections [PI. XVI, figs. 3 and 4] show the skele- ton to be composed of thick, close-set, parallel, radial pillars, which are minutely porous in structure. The radial pillars are separated by narrow zooidal tubes, about five or six pillars occupying a space of 2 mm., measured transversely, while about seven ' concentric laminae ' occupy the same space, measured vertically. Tangential sections [PL XVI, fig. 5J show the large, rounded ends of the transversely divided radial pillars, placed close together, and united in a stellate manner by whorls of deli- cate radiating connecting processes or ' arms. The rounded or sinuous pores included within the hexactinellid network thus formed represent the zooidal tubes as seen in section. In the fact that the coenosteum consists of well developed radial pillars, united at corresponding levels by whorls of connecting processes, Syriiigostrouia ristigfliicliciisc resembles an Actinostroma, while in the minutely porous structure of the skeleton-fibre it entirely resembles a typical Stromatopora." The fibre of this species is not so well preserved as in many examples of Syriugosfronia and Stromatopora from the State of New York, nevertheless its reticulate character is sufficiently shown in the better parts of many sections. The vertical fibres are quite well marked, and the horizontal reticulation is apparent [250] Parks: Silurian Stromatoporoids ii in transverse section. The horizontal fibres, however, do not maintain a direction at right angles to the vertical with the same degree of persistency exhibited by some of the New York species. In some places, however, as viewed in vertical section, sufficient proof is obtainable that the square network is the basis of struc- ture, although a more irregular network is produced by the flex- ures of the horizontal fibres. The figure shown (PI. XVI, fig. 4) is considerably restored, and it represents the appearance above described in a more striking manner than has really been seen over a section of such extent. It will be observed that the gross horizontal laminae are in some cases quite thick, and not formed of a single whorl of fibres, as must be the case in an example of the genus Actinostroma. The continuity of the radial pillars is also a function of the thickness of the section ; in very thick slices they are strongly marked and continuous for considerable dis- tances, but in very thin sections this appearance is entirely lost, in consequence of the fact that the astrorhizal canals invade the sides of the pillars so as to give them the appearance of being entirely interrupted at intervals. Dr. Nicholson considers the round or angular pores which appear where the tangential section follows a lamina to be the severed ends of zooidal tubes, and he considers the vertical interspaces observed in vertical section as being the longitudinal sections of the same structures. Now, where the tangential section follows an interlaminar space the severed pillars are surrounded by a completely open annulus, which is always in free communication with the astrorhizal canals. What, then, has become of the zooidal tubes? Either they are non-existent or their walls, on two sides at least, were composed of soft tissue. In either event, it is manifestly wrong to interpret the vertical interlaminar spaces as zooidal tubes. Doubtless pores pierced the laminae between the pillars and passed into the interlaminar spaces, but it is not in accord with the facts to state that these pores represent the habitation cavities of zooids. Locality. — Helderbergian, Dalhousie, Cap Bon Ami, New Brunswick. [251J 12 Parks: Silurian Strom atoporoids Syringostroma centrotum, Girfy. — Plate XVI, Figs. 6, 7, 8, and 9, and Plate XVIII, Figs. 6 and 11. Syringostroma centrotum, Girty, Report, State Iiluseum, New York, 48th, vol. ii, p. 293, pi. vii), figs. I and 2, i894- Dr. Girty's description is as follows : " Coenosteum massive, spheroidal, often attaining a large size. Intermittent concentric growth results in the formation of latilaminae, which are usually conspicuous. The surface is thickly covered with rounded emin- ences or mamelons, and astrorhizae are numerous but minute. The presence or absence of an epitheca has not been ascertained. " No specimens in this collection are entire, but all evidence points to an originally spheroidal form for the coenosteum. One specimen appears to have had a diameter of about 27 cm. when entire. The concentric character of the structure is usually quite striking. Some specimens are readily separated into thin sheets or latilaminae. . . . Tangential sections do not remain parallel to the surface in this species, except over small areas, for the curvature is not regular, but flexuous, and the latilaminae are more or less foliaceous and imbricating. Thus extended sections cut the monticules at all angles, and appear like panels of curly maple. " In specimens which break along the latilaminae the sur- face is seen to be vermiculate and porous, thickly covered with prominent conical elevations. This characteristic vermiculate structure is shown also in tangential sections, and is not due to weathering and preservation. Astrorhizae are numerous, but small and inconspicuous, as if they were merely the usual porous structure intensilied. They are distributed over the surface, and are often to be found on the sides of the monticules. The monti- cules, as shown by radial section, are usually superimposed throughout one latilamina, but in two successive latilaminae this may or may not be the case. They are often pierced by straight central canals, directed radially. These canals often extend through one whole set of monticules. They have no proper walls, and, therefore, cannot be referred to ' Caunopora ' tubes or tubiculous annelids. When broken transversely, the mamelons are seen to be distinctly porous. The pores, or canals, are often arranged in concentric series, coincident with the cut edges of [252] Parks: Silurian Stromatoporoids 13 the intersected laminae, and evidently represent sections of the astrorhizal canals. " Weathered fractures and properly oriented sections show that the radial pillars are strong, parallel, and continuous through several layers of the coenosteum. They are united at more or less regular intervals by concentric partitions, which have the porous structure above described. These concentric laminae appear to be composed of inosculating fibres, forming a reticulate skeleton, and not of lateral arms given off in a whorl around each pillar, as in the genus Actinostroina. Vertical sections through a monticule show that the radial pillars are not parallel as elsewhere, but are inclined at a slight angle away from the imaginary axis of the monticule." In vertical section (PI. XVI, figs. 6 and 7) the features so well described by Dr. Girty are clearly shown. Both the hori- zontal and vertical elements are flexuous and fairly continuous across the slide ; of the former about seven occupy the space of I mm., while the latter are much more widely spaced, averaging about three to the mm. This wide spacing of the laminae and their extreme thinness (usually only one strand of fibre) consti- tute the chief means of distinguishing this form from its allies, as seen in vertical section. Tangential sections (PI. XVI, figs. 8 and 9) are very char- acteristic. The close-set astrorhizal systems on raised mamelons give a series of concentric rings and radial canals where severed, strongly suggestive of " curly maple," as stated by the author of the species. The cut ends of the radial pillars are more apparent than in the other species referred to Syringostroina by Girty. As the pillars, like the rest of the coenosteum, are composed of a reticulation of fibres, the interspaces between which are only a little less than the diameter of the pillars themselves, it cannot be expected that the cross sections of the pillars should present a strictly round appearance. The skeletal fibre is not as well preserved in this species as in some of the others. In vertical section (PI. XVIII, fig. 11) the horizontal components are not well defined; the vertical fibres are more distinct and heavier, frequently forming the borders of the pillars. The general square mesh of the reticula- tion is indistinctly shown. Tangential sections of the fibre '(PI. [253] 14 Parks: Silurian Stromatoporoids XVIII, fig. 6) also show the greater strength of the vertical elements, the cut ends appearing as distinct dots. The horizontal connecting arms are not so clearly shown, and were doubtless of a more delicate character. The highly magnified tangential sec- tion, presenting the appearance described above, is quite charac- teristic of the species. Locality. — Helderbergian, Cedarville, N.Y. ; Herkimer Count)r, N.Y. The illustrations are prepared from sections No. 1756 and No. 1757, New York State Museum, Albany, N.Y. Syringostroma consimile, Girty. — Plate XVI, Figs. 10, 11, and 12, and Plate XVIII, Fig. 14- Syringostroma consimile, Girty, Rep. State Mus., N. Y., 48th, vol. ii, p. 297, pi. vii, figs. 3 and 4, 1895. With regard to the general shape of the coenosteum, Dr. Girty states: "Coenosteum massive, large and sub-spherical. Latilaminate structure shown by sections or on weathered sur- faces. The laminae are disposed in pointed, wave-like folds, which are not superimposed. Skeleton-fibre, finely porous ; tissue, reticulate; astrorhizae, few and \'ery large, sometimes provided with a tubular axis." The whole coenosteum is composed of exceedingly delicate vertical and horizontal fibres, about twenty-five of which appear in the distance of i mm. This constitutes the reticulate skeleton- fibre of the organism. Passing through this tissue are vertical interspaces, about eight of which appear in the space of i mm. Astrorhizal systems are present, the horizontal canals of which are confluent with these interspaces : but whether or not the ramifications of the astrorhizal canals communicate with all the interspaces, it is impossible to state. Nevertheless, I am inclined to believe that such is the case. The vertical spaces are interrupted at certain levels by hori- zontal laminae, which may consist of several meshes of the ulti- mate network or may be reduced to a single fibre. The thicker laminae are fairly horizontal, and may occur to the number of seven in a mm. The finer ones are more distant, and are apt to turn up into the pointed folds referred to by Dr. Girty. This feature is presented by parts only of the coenosteum, and there- fore cannot be considered as of diagnostic value. [254] Parks: Silurian Stromatoporoids 15 The astrorhizae appear to be about 8 mm. apart ; they do not usually possess an axial tube, and their horizontal canals are considerably wider than those of 5". constcUata. Vertical sections (PI. XVI, figs. 10 and 12) show a reason- able amount of constancy as far as the vertical interspaces and pillars are concerned, from eight to ten of each appearing in the space of a mm. These interspaces are enlarged where they are confluent with the astrorhizal canals. Each " pillar " is fairly continuous, and in some cases is seen to show a dark line on each side, which is produced by the vertical fibre of the reticulation. Much less constancy is exhibited by the gross horizontal elements — the laminae. In the lower part of the figure they are seen to be thick and continuous ; in the upper part they are thin and bent upwards into pointed loops. In certain parts of the plate, where the preservation is particularly good, the rectangular meshwork of the skeletal fibre can be distinctly discerned. Tangential sections (PI. XVI, fig. 2) show the wide hori- zontal canals of the astrorhizae and the vermiform skeletal mat- ter. The distinction between laminae and the interlaminar spaces is not well marked, owing to the thickness and contiguity of the laminae. In very thin sections the cut ends of pillars in the inter- laminar spaces are more apparent; certain of the vermiform patches resolve themselves into rows of partially coalesced dots. Even on the low scale of magnification of this figure the fibrous, reticulate character of the ultimate fibre can be distinctly seen. High magnification shows that the skeleton is composed of vertical and horizontal fibres, arranged precisely as in the genus Actinostroma. This ultimate structure is common to all the Hel- derbergian Stromatoporidae, and is seen in species from other formations as well. With regard to the size of the mesh, the present species is intermediate between Stromatopora foveolata, in which there are 30 strands to the mm., and Stromatopora con- stellata, in which from 22 to 24 occur in the same space. The fibre is beautifully preserved in the type specimen ; a very fortu- nate section is shown in Plate XVIII, fig. 14. Here the cut ends of the vertical fibres are excellently revealed, and also the fine horizontal connecting elements. Only occasionally is this minute structure shown to such perfection. Usually the angularity of the mesh and the identity of the pillars are lost by mineralization.. [255] i6 Parks: Silurian Stromatoporoids Stromatopora constellata, on the one hand, and Syringo- stroma centrotuin, on the other, present two extremes, by compari- son with which these Helderbergian forms are best determined. The present species differs from S. constellata in finer arrange- ment of the vertical elements, and in the fact that these vertical ele- ments are strictly iiitcrlaininar. Despite the general resemblance, it must be remembered that this distinction is of generic import- ance. Tangential sections may be readily distinguished by the wider astrorhizal canals of the present species. From Syringo- stroma centrotum it may be distinguished by the lack of the '■■ curly maple " effect in tangential section, and by the more closely set and thicker laminae. The upturning of the laminae into pointed folds does not seem to be a constant feature. The cross sections of the pillars are less distinct than in 5*. centrotum, because the laminae are closer together and because there is a tendency for the pillars to unite into vermiform rows. The species is much further removed from a typical Syringostroma than is ^. centrotum. Locality. — Helderbergian, State of New York. The type is from the outlet of Skaneateles Lake, and is in the Peabody Museum, Yale. Syringostroma barretti, Girty — Plate XVII, Figs, i and 2, and Plate XVIII, Fig. 5. Syringostroma barketti, Girty, Rep. State Mus. New York, 48th, vol. ii, p. 296, pi. vii, figs. 5 and 6, 1895. The type specimen of this species I have been unable to obtain, but two examples from the New York State Museum appear to belong to the species, and on them the following de- scription is based. Before attempting this description it will be necessary to give Girty's account in full, as follows : — " Coenos- teum large, hemispherical, spreading. Latilaminae distinct, more or less labyrinthine towards the centre, on the periphery flowing in broad folds. They end abruptly on the under side and are attached directly without an epitheca. Laminae parallel and gently flexuous. Astrorhizae not numerous, but large and con- spicuous. The nucleus of an astrorhizal system is sometimes represented by an axial tube, and the laminae at that point are [256] Parks: Silurian Strom atoporoids 17 often elevated into a low monticule. Skeletal tissue finely- fibrous, but a little coarser than in 5". ccntrotum. " This species is characterized by the infundibuliform, con- centric growth and the flat base without an epitheca ( ?). With- out the aid of thin sections, the outer surface of the type specimen appears dense, fine-grained, and structureless, except for latila- minae, which separate in unusually thin sheets. Sections near the surface are without monticules, astrorhizae and axial tubes, exhibiting only the uniform porous skeleton and fibrous struc- ture. The same surface characters are presented by the basal portions and suggest an epithecate condition, but it has not the polished surface and concentric wrinkles characteristic of the epitheca in Favosites. Lower Pentamerus, Indian Ladder, New York." Vertical sections of the examples in my possession which come nearest to the description given above (Plate XVII, Fig. 2), show that the laminae are bent into folds, but this feature is not altogether absent in 6". consimile, and is therefore not diag- nostic. However, a more irregular manner of growth is shown by the greater difficulty in cutting sections which are either strictly horizontal or vertical over any considerable area. The vertical pillars are somewhat heavier than 5". consimile and are farther apart, only four or five occurring in a mm., whereas seven or eight appear in 5. consimile. The horizontal laminae are of about the same average thickness as in the other species, but they are somewhat more irregular. In the middle of the photograph the rectangular reticulation of the skeletal matter can be distinctly seen. Tangential sections (Plate XVII, fig. i) show that some of the astrorhizae have an axial canal and that the horizontal channels are much ramified. Dark bands representing the severed laminae are more conspicuous than in 5". consimile, but this appearance is dependent on the thickness of the section. The severed ends of the pillars can be seen in certain parts of the section as well as their tendency to unite into vermiform rows. It is doubtless this tendency that makes these structures so much more apparent in vertical than in tangential section. The skeletal fibre (Plate XVIII, fig. 5) presents, in tangen- [257] i8 Parks: Silurian Stromatoporoids tial section, a mesh of about the same size as that of 5". consimile. The fibre itself appears to be somewhat heavier, and distinct dots representing the ends of the vertical elements are not to be seen. These features as well as the rounded shape of the meshes are probably due to a less perfect state of preservation. Dr. Girty himself expresses doubt as to the advisability of separating this species from S. consimile. I regret that I have not seen the type of S. harretti, for I have doubt as to the iden- tity of my specimens; they do not, in all particulars, agree with Girty's description or figures. The pillars are more continuous, but sections which are not quite vertical are almost exactly like the figure given by Girty. I cannot agree with that author that the two species may be distinguished by the character of the laminae which in i'. consimile " are angular and independent in their flexures," and in J?, barrctti " gently cur\-ed, proximate and parallel." The statement is true for the latter species, but I have not observed the angular inflections as a constant feature in the former. The most practical way of distinguishing the forms is by the larger size and wider spacing of the pillars. The curved laminae are also a means of identification as far as the type speci- mens are concerned, but I have never found that the exact man- ner of growth of the coenosteum as a whole is of determinative value. Like S. consimile, this form differs from Stromatopora con- stellata in the thinner and more widely separated laminae. The tangential sections are remarkably alike, but 5". constellata is built on a much coarser scale, both with regard to the minute fibre and in the gross reticulation. The tangential section of 5". harretti is much finer than one would expect from an examina- tion of the \'ertical section. 5". harretti, in its vertical section, approaches closer to 5". centrotuni than does 6". consimile; how- ever, it may be distinguished from the former species by the lack of the " curly maple " effect and by the much stronger Stroma- topora-like appearance. Locality. — Lower Pentamerus, Lidian Ladder, N.Y. The present figures are prepared from specimens from Clarksville, N.Y. (Section No. 736, New York State ]\Iuseum.) [258] Parks : Silurian Stromatoporoids 19 Syringostroma microporum, Girty — Plate XVII, Figs. 3 and 4. Syringostroma microporum, Girty, Rep. State Mus. New York, 48th, vol. ii, p. 296, pi. vi, fig. 7, 1895. Dr. Girty's description is as follows : — " Coenosteum mas- sive, latilaminate, shape unknown, probably spherical or sub- spherical. Surfaces parted along the latilaminae are covered with numerous monticules, which are, howeAer, of small size. They are usually pierced by axial tubes and provided with well developed astrorhizae, seldom visible on the lamellar partings. In vertical section the skeleton is seen to be composed of per- sistent radial pillars and the customary concentric laminae. Tangential sections show the laminae to have a finely porous structure, the skeletal fibre being of the characteristic reticulated tissue common to other Helderberg Syringostromas. Tangential sections through the monticules show only the cut ends of the astrorhizae. When the section lies in the lateral slope of a mon- ticule, the latter is seen to be provided with an extensively spread- ing astrorhizal system." This species approaches so close to S. consimile that it would be better to consider it as a variety only. The preserva- tion of the type specimen is not nearly so good as in the case of the latter species, and I am inclined to think that this is respon- sible for the difference shown by Girty's figures of the two spe- cies. In all these fine-grained forms the thickness of the section is so important that wonderfully different results may be obtained. However, the type specimens show some differences and the description of the present form is best expressed in terms of those differences. 5". microporiim has very small monticules placed three or four mm. apart ; S. consiniilc has astrorhizal sys- tems the same distance apart, but monticules are absent. I have long been of the opinion that the presence or absence of small monticules is not a feature of specific value. The spacing of the pillars is the same in both species and the laminae are like- wise separated by a similar interval. The upturning of the laminae into points in 5". consimile is by no means universal and cannot therefore be used to distinguish it from the present spe- cies. Extended vertical sections of 5. microporum, in spite of the presence of the monticules, do not shovv' continuous axial [259] 20 Parks: Silurian Stromatoporoids tubes in the astrorhizal centres to as great an extent as similar sections of 6^. consimilc. The upturning of the laminae on approaching the axial tubes is also much more pronounced in 5. consimilc. This is the most evident difference between the two species. The skeletal fibre is less well preserved in the type of S. microponiin. but shows the same rectangular network as the other type. I believe the two to be pi'actically identical, but I should not be justified in destroying an author's species with- out having had an opportunity to examine more material. Locality.. — Lower Pentamerus limestone, Cedarville, N.Y. The type specimen is in the Peabody Museum, Yale University. (7r;;«j— STROMATOPORA, Goldfiiss Stromatopora foveolata, Girty, sp. — Plate XVII, Figs. 5, 6 and 7, Plate XVIII, Figs. 4 and 10. Syringostroma foveolatum, Girty, Rep. State Museum, New York, 48th, vol. ii, p. 29s, pi. vi, figs. 8 and 9, 1895. Dr. Girty's description is as follows : — " Coenosteum mas- sive and of large size. Outer surface and point of attachment not known. Division into latilaminae usually apparent. In one specimen the latilaminae are folded into regular hemispherical elevations, having a diameter of about 50 mm. Radial pillars continuous and usually large. Viewed in vertical section, the laminae are thick and the interlaminar spaces narrow. Tangen- tial sections show the laminae to have a comparatively dense structure, but the organism may still be referred to the Mille- poroid type. The pores, which in 5. ccntrotuui appeared as transparent vermiculate patches, in this species are represented only by minute spots, sometimes connected by transparent thread- like bands. The skeleton fibres are large, with fine reticulations, giving the tissue as a whole a dense consistency. The concentric laminae, likewise, are usually thick and hea\'y, the vacuoles of small size, and the intercolumnar spaces minute. Astrorhizae are frequent, and well developed monticules appear to be pre- sent, chiefly associated with astrorhizae. They are indicated in the section by a darker shading, but their ele\'ation is so slight tliat tangential sections often show the astrorhizae equally well on all sides. The monticules sometimes have tubular axes. [260] Parks: Silurian Stromatoporoids 21 When weathered, the coenosteum often develops httle pits or conical depressions about four mm. apart, which are rather characteristic of the species." To this excellent description there is nothing to be added as far as the gross anatomy is concerned, but it must be pointed out that the skeletal fibre is essentially the same as in Syringostroiiia consimile, consisting of delicate ver- tical fibres connected by horizontal ones, giving rise in vertical section to a rectangular mesh and in horizontal section to an irregular network. The fibres are somewhat heavier than in 6". consiniUc, and, especially in horizontal section, build a finer network. Vertical sections (Plate XVII, figs. 5 and 6) show very pro- nounced differences from the species of Syringostroma already described. In the latter the pillars occupy the space between more or less continuous laminae; in this species the pillars lie between open bands which represent the planes over which the systems of astrorhizal canals are thrown. The bands of pillars are therefore laminar, while in the other species they are inter- laminar in position. The spaces between the pillars, while they open into the astrorhizal passages, are not coincident with them ; they are true pores passing through the laminae, and the " pil- lars " are not true pillars at all, but vertical sections of the tissue of the laminae lying between the open tubes. This constitutes the difference between Syringostroma and Stromatopora, and it is for this reason that the species is removed to the latter genus. While the gross horizontal elements vary in thickness, an average of four occur in the space of one mm. They are separated by much thinner open bands, representing the hori- zontal pores. A lamina of average thickness consists of four or five bands of ultimate fibres. These fibres are more or less con- tinuous across the pores, forming the so-called " zooidal tabulae " ■ — an interpretation that I am not inclined to accept. Tangential sections (Plate XVII, fig 7) show the true Stromatopora type of structure — wide-spreading astrorhizal canals and the dense severed laminae pierced by pores. The microscopic character of the fibre is beautifully shown in this species; the horizontal and vertical arrangement is dis- tinct in spite of the extreme fineness of the reticulation. Measured [261] 22 Parks : Silurian Stromatoporoids either horizontally or vertically as many as 30 strands to the mm. can be seen. As the fibre itself is somewhat heavy a dense appearance is presented under lowf magnification. Under higher powers, vertical sections (Plate XVIII, fig. 10) show both hori- zontal and vertical fibres and the square network. Tangential sections (Plate XYIII, fig. 4) present an irregular network which is much finer than that shown by any other figure on the plate. This species is even more typically a Stromatopora than is S. constellata, which we have adopted as a standard for comparison. The pores piercing the laminae are more regular and more dis- tinctly pores. About eight of them occur in one mm., and they are separated by interspaces much wider than themselves. The gross vertical elements, if they ever existed, are so interrupted by the astrorhizal canals as to be indistinguishable. In this species the astrorhizae are about 3 mm. apart ; in 6". constellata they may be distant from one another as much as 7 mm. From Syringostroma centrotiim the distinction is easily made by means of the thick, closely set laminae and the dense, uniform appearance of the tangential section, which shows no indication of severed pillars or concentric rings. Locality. — Lower Pentamerus, Cedarville, N.Y. Dr. Girty's types are in the Peabody Museum, Yale. The figures here pro- duced are prepared from sections No. 1733 and No. 1736, New York State Museum, Albany, N.Y. No. 2270, Am. Mus. Nat. Hist., is probably the same species, but the general structure is a little coarser. Cherry Valley, N.Y. Stromatopora corallifera, sp. nov. — Plate XVII, Figs. 8 and 9, Plate XVIII, Fig. 7. Stromatopora corallifera, Hall and Whitfield, MSS. (Am. Mus. Nat. Hist. Labels.) The coenosteum in this species has a habit of growth be- tween cylindrical and hemispherical, i.e., the laminae arch strongly upward, and, as the growth is more rapid in the centre, a cylindrical form results. The type specimen is about 10 cm. high by 8 wide. The cylinders are sometimes conjugate and invested by laminae common to the whole mass. The most important feature in the minute structure of the [262] Parks: Silurian Stromatoporoids 23 skeleton is the very slight differentiation of the laminae as dis- tinct structures. Nevertheless the concentric manner of growth is clearly shown. The fibrous, reticulate structure of the ulti- mate skeleton is pronounced. This fibre is pierced by numei-ous vertical pores of which about six occur in one mm. Between these pores the skeleton-fibre appears as vertical columns when viewed in section, the column being of greater diameter than the pore itself. The horizontal fibres of the general tissues are continued across the pores to the number of 15 or more in a vertical distance of one mm. (zooidal tabulae?). In the true Stromatoporas the definition of the gross laminae depends on the development of layers of horizontal canals; in this species these canals seem to be extremely variable, sometimes no farther apart than the horizontal fibres of the general tissue, sometimes at much greater intervals. The whole effect is to obliterate the concentric laminae unless the delicate lines occurring to the number of 15 in one mm. are to be interpreted as such. The astrorhizae are very large with exceedingly long and tortuous horizontal canals. Vertical sections (Plate XVII, fig. 8) suggest the genus Sy lingo stroma in the long pillar-like portions between the verti- cal pores; the defined laminae of that genus are, however, want- ing. The continuity of the vertical elements is extremely variable, in some places they are cut off by horizontal canals at minute in- tervals and in others they continue across many strands of the ultimate fibre. That they are not pillars, but interspaces bie- tween pores, is seen by the fact that the two margins of an indi- vidual are not parallel, but each conforms to the curving of the pore which borders it. The frequent lateral coalescence of the " pillars " also proves that they are not individual elements. Gross laminae which result from the arrangement of the astro- rhizal systems at regular intervals are very ill defined. The general concentric manner of growth is, however, clearly seen in the horizontal elements of the skeletal fibre. As in 5". foveolata, these are partially continuous across the pores, especially if the section is thick. If these fine horizontal lines are interpreted as zooidal tabulae, then there is no trace of horizontal elements in some parts of the coenosteum. [263] -'24 Parks: Silurian Stromatoporoids The cut ends of the large astrorhizal canals can occasionally be seen, but the horizontal canals are more commonly repre- sented by depressed openings between the fine laminae. Tangential sections (Plate XVII, fig. 9) show the ex- tremely well developed astrorhizal systems, with much ramified horizontal canals, the somewhat irregular pores piercing the laminae, and the general Stromatopora-like aspect of the tissue. The ultimate fibre has the same arrangement as in the other species; it is a little coarser than in 6". foi'colata and a little finer than in Syringostroma consiinilc (Plate XVIII, fig. 7). The above description is founded on sections which represent the average appearance of the tissue. A great deal of variation is, ho\ve\er, presented by different portions of the same coenos- teum — possibly as great a difference as exists between forms herein ascribed to different species. From i'. constellata the present species differs in its finer ultimate fibre, the more ex- tended astrorhizae and the much less distinct gross horizontal laminae. From 5". fovcolata it may be distinguished by the thin- ness and variability of the laminae. The astrorhizal systems are not superimposed as they are in 5". fovcolata, and there is no up- turning of the laminae towards axial astrorhizal tubes as in that species. Locality. — Helderbergian, Port Jervis, N.Y. Type specimen No. -^V-a, Am. Mus. Nat. Hist., N.Y. Probably the same is No. 22 7_o of the same museum. Perry County, Penn. [264] Parks : Silurian Stromatoporoids 25 II. Silurian Stromatoporoids from Hudson Bay. ORDER— STROMATOPOROIDEA, Nuh. and Murie Section A (" Hyclractinoid Group") Fa^MV-v— ACTINOSTROMIDAE, Nich. Genus — AcTiNOSTROMA, Nich. and Murie. Actinostroma tenuifilatuMj Parks. For description, see Niagara. Stromatoporoids, page 10. It is with some hesitation that specimens from Southampton Island are ascribed to this species. The coenosteum is large and hemispherical, readily exfoliating, and presenting low, ill defined and irregularly distributed monticules on the parted surfaces. The skeletal fibre is exceedingly delicate and much altered by crystallization. About six concentric laminae appear in i mm., measured vertically. These elements are thin, but are much more sharply defined than the radial pillars. With regard to the nature of these latter structures considerable doubt must be expressed. They are very badly defined in all the examples, and appear as continuous lines only exceptionally. Without the comparison of a number of sections one would pro- nounce the organism a Clathrodictyon, near to C. vesiculosnm minutum. The better preserved portions, however, when cut very accurately in a vertical direction, .show the continuity of at least some of the pillars. Locality. — Southampton Island, Hudson Bay, Canada. Actinostroma tenuifilatum var. inflectum, Parks. Plate XIX, Figs, i and 2. Actinostroma inflectum, Parks, Ottawa Nat. vol. xxii, p. 27, 1898. " While fragments only are available, the inference is obvi- ous that the coenosteum is of a hemispherical shape, and that it reaches considerable dimensions. "Vertical sections (Plate XIX, fig. i) show it to be com- posed of delicate horizontal elements, the spacing of which is extremely variable, as many as ten or as few as three laminae ' [265] 26 Parks : Silurian Stromatoporoids occurring in the space of i mm. Tlie concentric layers are con- nected by continuous radial pillars, which occur to the number of six or seven in a mm. Instead of maintaining a horizontal direction, the laminae are bent sharply upwards at intervals of about I mm. As each overlying lamina follows the same course, and as the identity of the lamina is lost at the apex of the fold, the coenosteum appears to be traversed by vertical columns made up of loose vesicular tissue. These columns do not show the compact structure of those of Nicholson's Stylodictyon, but the general appearance in vertical section is very suggestive of that genus. A similar arrangement is not uncommon in different Stromatoporoids, and it is very questionable whether it is a feature of generic or even specific value. These inflected portions doubtless enclose astrorhizal systems, but horizontal canals are not perceptible. Owing to the upturning of the laminae it is difficult to prepare sections which follow the course of the pillars over any considerable extent. In consequence, one may easily mistake this species for a Clathrodictyon. " Tangential sections (Plate XIX, fig. 2) do not reveal any astrorhizal canals, nor is the preservation sufficiently good to show the whorls of connecting arms typical of the genus. Noth- ing is presented by such sections beyond the cut ends of the pillars and the obliquely severed upturned edges of the laminae. Typical examples are easily distinguished from A. tonnfilatum, but intermediate forms connect the two species, so that one is tempted to regard the examples under discussion as representing a variety only of the latter species." To the above description there is nothing to be added. A careful study of the significance of inflected laminae has con- vinced me that no specific value can be attached to that peculiar- ity. I have, therefore, reduced the species to the rank of a "\'ariety. Locality. — Pagwachuan River, Station 641, and near mouth; Little Current River, 17 miles from mouth. Type col- lected by Mr. AV. J. AA'ilson, and stored in the Museum of the Geological Survey, Canada. ACTINOSTROMA TENUIFILATUM Z'O". CYLINDRICUM, Z'ar. IIOZ'. The mode of growth of the coenosteum is distincth- cylin- [266] Parks : Silurian Stromatoporoids 27 drical, each cylinder being about 15 mm. in diameter, and attain- ing a length of 8 or 10 cm. The cylinders may remain separate or they may coalesce into a conjugate mass surrounded by lami- nae of the common coenosteum. This cylindrical manner of growth is by no means confined to the present species, but is seen in many others, and will be referred to again in the course of this article. The laminae are wound around a central axis, and they likewise cover the end of the axis as a series of superimposed, deeply convex caps. Each cylinder therefore grows upwards as well as by increase in diameter. The minute structure of the fibre is identical with that of Actinostroma tenuiftlatiun. The laminae are more closely set in the initial layers of each cylinder and increase in width of spacing towards the periphery. On the average, six occur in a distance of I mm. The radial pillars are continuous and a little more closely set than the laminae. The only difference from the type of the species is seen in the cylindrical manner of growth. Locality. — Silurian, Southampton Island, Hudson Bay. Type specimen in Museum of Geological Survey, Canada. Actinostroma franklinense, sp. nov. — Plate XIX, Figs. 3 and 4. Clathrodictyon franklinense, Ami, MSS. Geol. Sur. Canada, Cruise of the Neptune, p. 329, 1906. cf. Actinostroma intertextum, Nich., Men. Brit. Strom., p. 138, pi. xiii, figs, 8- 1 1, 1892. The shape of the coenosteum cannot be stated, as the species is known from a small fragment only. However, it was probably of an explanate character, growing over the substratum and reaching a thickness of at least 15 mm. The most striking feature in the minute structure of the organism is the sharply defined continuous radial pillars which occur to the number of six in a distance of i mm. The horizontal laminae are somewhat more closely set, but they are much less continuous and defined structures. Astrorhizae have not been observed. Vertical sections (Plate XIX. fig. 3) show the sharply defined continuous radial pillars and the discontinuous, more irregular laminae. [267] 28 Parks : Silurian Stromatoporoids Tangential sections (PI. XIX, fig. 4) exhibit the cut ends of the pronounced pillars as round dots. Where the section fol- lows a lamina the dots are connected by irregular lines. The whorls of arms, characteristic of the genus, cannot be observed. This species can be distinguished by the very sharp defini- tion of the radial pillars and the irregular character of the laminae. The form is remarkably like Actinostroina iiifcrtextum, Nich., from the ^\'enlock of England. In fact, it is possibly identical. The only differences are the slightly finer structure and the lack of distinct whorls of connecting arms. It is readily admitted that this latter feature may be due to mineralization, in which case the species are identical. Locality. — Silurian, Beechy Island, Lancaster Sound, Can- ada. Type specimen in Museum of Geological Survey, Canada. Name proposed by Dr. H. M. Ami in the Cruise of the Neptune. I have listed the species as new because Dr. Ami has, as far as I am aware, issued no description or figures. Genus — Clathrodictyon. Nich. and lilurie. A large number of forms belonging to this genus occur in the rocks of Silurian age around Hudson Bay and Davis Strait. It is possible, but highly inadvisable, to create numerous new species, for, although minute differences from known species may be detected, it is not difficult to account for those differences as mere local variations. Clathrodictyon striatellum, d'Orb. For synonomy and description, see Gnclph Stromatoporoids, page 14, and Niagara Stromatoporoids, page 2^. Several specimens from Southampton Island are probably referable to this species. The horizontal laminae are well defined, but the vertical elements are very indistinct. So much are they destroyed that it is impossible to make statements as to their minute character. The spacing of the laminae is somewhat finer than in typical examples of 5". striatcllitm and coarser than in C. vcsicidosmn ininutum. Clathrodictyon vesiculosum, Nich. and Mnrie. For synonomy, etc., see Niagara Stromatoporoids, page 14. [268] Parks: Silurian Stromatoporoids 29 This species is common, and presents some peculiarities of growth, as the coenosteum is quite variable in form — sometimes hemispherical, sometimes explanate, and occasionally forming cylinders or ramose branches, a habit common to different species in this area. The fineness of the fibre is also variable, as many as 15 fibres to a mm. occurring in some examples. To the locali- ties mentioned in Niagara Stromatoporoids must be added Southampton Island, Hudson Bay. British Museum specimen No. P2270 from Cape Riley, N. Devon, is an extremely fine- grained example with unusually large astrorhizal canals. To this species may also be referred certain specimens, in a poor state of preservation, from Griffith Island, Grinnell Land, and Bessels Bay, Greenland. Clathrodictyon vesiculosum minutum. Parks. For description, see Niagara Stromatoporoids, page 19. Locality. — Southampton Island, Hudson Bay. Clathrodictyon cystosum^ Parks. For description, see Niagara Stromatoporoids, page 21. A form closely allied to or identical with that type of this species in which the laminae are but little inflected occurs on Southampton Island. The coenosteum forms large hemispherical masses encrusting Favositcs. British Museum specimen No. P6993 from Bessels Bay, Greenland, probably belongs to this species. Clathrodictyon cystosum lineatum. Parks. — Plate XX, Fig. 9. For description, see Niagara Stromatoporoids, page 14. This variety is represented by large masses on Beechy Island, Lancaster Sound. Some slight differences from the type are observable but a comparison of the figures will show that these are not of specific or even varietal value. Several specimens in the Geological Survey of Canada collection and British Museum specimen No. 56702. Clathrodictyon cystosum z'ar. cylindricum^ var. nov. Tlie coenosteum in the type specimen is about 8 cm. in width [269] 30 Parks: Silurian Stromatoporoids and 6 cm. high. The basal portion is lost. The whole mass con- sists of conjugate upright cylinders, fifteen of which occur in the type. These structures have an' average diameter of 15 mm., and they project at the top of the coenosteum as prominent rounded points. The surface is smooth, or minutely pitted. The minute arrangement of the skeletal fibre is so like that of Clathrodictyon cysiosum that it is impossible to remove the form from that species. The cylindrical manner of growth is, however, characteristic, and on that alone the new variety is founded. The arrangement of the coenosteum into conjugate cylinders is a feature common to several species and to different genera. It indicates a manner of growth only, and cannot be considered of specific value. An insensible transition can be observed between typical Clathrodictyons exhibiting this peculiar manner of growth and the genus Cyclodictyon, in which the general fibre is traversed by radiating secondary pillars. Locality. — Silurian. Southampton Island, Hudson Bay, Can- ada. Type specimen in Museum of Geological Survey, Canada. Clathrodictyon drummondense, Parks. — Plate XX, Figs. 7 and 8. For description, see Niagara Stromatoporoids, page 26. Two specimens from Soutliampton Island represent this species. They have a globose shape, and they are about 50 mm. in diameter. The surface is characteristically rough. The two vertical sections shown in the plate indicate the variation to be seen in a single coenosteum. Figure 7 is from a section cut in a vertical direction through a well preserved part of the skeleton. Figure 8 exhibits the variable aspect of sections, not quite ^-erti- cal, passing through altered parts of the coenosteum. ACTIN0DICTY0N_, geil. IIOV.* The coenosteum is composed of horizontal laminae after the type of such forms as Clathrodictyon z'csictdosnin. i. e., forms in which the vertical pillars are formed by inflections of the laminae. *For this generic name I am indebted to Professor W. S. Milner, of the University of Toronto. {aKTlg-ivoc: and ^iktvov.) [270] Parks: Silurian Stromatoforoids 31 The general fibre does not, therefore, differ from that of Clathro- dictyon. It is, however, pierced by large, spongy pillars at more or less regular intervals. These pillars are quite analogous to those of Lahechia, but, instead of traversing a tissue of vesicular cells, they pass through the tissue of what is otherwise an ordin- ary Clathrodictyon. In all the species known the laminae are disposed in a concentric manner around an axis, giving rise to a cylindrical form of coenosteum. These cylinders, if single, may produce a ramose form, and, if conjugate, an irregular mass, in which the individual cylinders are apparent only as botryoidal elevations on the surface. The manner of growth has never appealed to me as being of generic or even of specific importance. The new genus, therefore, must rest on the peculiarity of the secondary pillars. It differs from Stylodictyon, in that the pillars are not formed by the upward inflection of the laminae and that the tissue is not reticulate. It is interesting to note that as the pillars become more numerous the ordinary elements of the tissue decrease in importance, and eventually become so attenuated that they might be interpreted as zooidal tabulae. I am convinced that the structures resembling tabulae in the species about to be described are to be interpreted as the residuum of horizontal laminae. The genus presents certain strong resemblances to the family Idiostromidae, and particularly to Idiostroma itself. This resemblance is seen in the cylindrical manner of growth and in the presence of a large open tube in the axis of some of the cylinders. The specimens are too small to justify an opinion as to the character of these tubes ; they may be analogous to the tubes of Idiostroma or they may be enclosed corals. These structures, however, are not always present. They do not open to the surface, and they do not send out branches into the general tissue of the coenosteum. The resemblance to Idiostroma is best seen in the species to be described as Actino- dictyon loivi. Here the uniform development of large pillars and the attenuation of the horizontal elements makes the similarity very marked. The fibre, however, is not porous, but dense, and there is no evidence of secondary large canals radiating from the axial tube to the surface. [271] 32 Parks : Silurian Stromatoporoids AcTiNODicTYON CANADENSE, sp. fiov. — Plate XX, Figs. I and 2. The coenosteum is known from a small fragment only, and although some observations might be made as to its form, it seems advisable to await the discovery of more material. The laminae are arranged in a concentric manner around a centre, which in the t3'pe specimen is occupied by a coral or axial tube. These elements are extremely variable and much crumpled, their upward inflections giving rise to vermiculate ridges and points which are continued upwards as the A-ertical pillars of the general reticulation. The spacing of the laminae is also quite variable, as many as six or as few as three appearing in a vertical distance of one mm. Passing through this general fibre are stouter continuous pillars disposed at intervals which vary from I up to 3 or 4 inm. These pillars are of greater diameter than the general fibre and appear to be of less dense character, as they disappear with greater rapidity as a section is reduced in thick- ness. Vertical sections (Plate XX, fig. i) show the much crumpled character of the laminae with their upwardly inflected points bearing the pillars, which may or may not reach the lamina next above. The large secondary pillars are conspicuous. A close examination of extended sections seems to indicate that these large pillars first appear as continuations through several interlaminar spaces of the ordinary vertical elements. When, however, the pillar has acquired strength, it passes with a con- siderable degree of independence through the general tissue. Tangential sections (Plate XX, fig. 2) vary much with the thickness of the section ; the one figured is fairly thick and con- sequently all the elements are heavier than in a thinner prepara- tion. The cut ends of the pillars (small) are occasionally seen as distinct dots ; more usually, howe\'er, the dots are replaced by crescentic bands representing the severed edges of the upturned laminae. In this respect the section is very suggesti^'e of Clatlirodicfyoii fastigiatiim and its allies. Where the section follows a lamina a very characteristic feature is seen in the large round pores which pass through the lamina. At first I was in- clined to regard these pores as analogous to those exhibited by StromatoporcUa, and to regard the large pillars as tubes. By [272] Parks: Silurian Stromatoporoids 33 comparison with the other species of the genus and by the close examination of several sections, I am convinced that such an interpretation is impossible. The cut ends of the larger pillars are not conspicuous, but can occasionally be made out. This lack of definition is doubtless owing to the less dense character of these structures. This species is the most Clathrodictyon-like of the four to be described; it represents the lowest stage in the development of the genus. Locality. — Silurian, Southampton Island, Canada. Type specimen in Museum of Geol. Sun, Can. AcTiNODiCTYON Lowi, sp. nov. — Plate XX, Figs. 3 and 4. This species also is known from a small fragment only. The coenosteum presents the same concentric arrangement of its constituent parts as the previous species. Like that species also the elements are disposed around a coralline branch or large axial canal. The horizontal elements of the coenosteum are very delicate and closely set, as many as fifteen occurring in the space of one mm. They are inflected upwards, but the continuation of the flexures into distinct pillars cannot be made out. The horizontal course of these elements is so interrupted by the much larger secondary pillars that their continuity can scarcely be discerned. The large pillars occur to the number of four in the distance of one mm.; they are irregular in their course, unite with one another and are not always round in cross section. Vertical sections (Plate XX, fig. 3) are difiicult to prepare in a direction following any considerable number of pillars. The figure is drawn from a vertical longitudinal section and in consequence the radiating character of the pillars is not shown. The delicate horizontal elements are much disrupted in places, but are to be observed in parts of the section where no large pillars occur. The irregularity of these latter elements, their coalescence and their greater frequency distinguish this species from A. canadense. Tangential sections (Plate XX, fig. 4) show the cut ends of the pillars, their union either by approximation or by the aid [273] 34 Parks: Silurian Stromatoporoids of the laminae and the tendency to irregular form. As in the case of A. canadense, the round pores passing through the laminae are a conspicuous feature. The strong resemblance between the cross section of A. lowi and the similar section of A. canadense is very striking. The vertical sections, however, are so different that no confusion can possibly result. If I had the present species alone, it is very probable that greater hesitation would be shown in the interpre- tation of the delicate horizontal elements. Their resemblance to the zooidal tabulae so often referred to cannot be denied, but I hold this interpretation to be quite impossible in view of the resemblance between the two species. The entirely analogous structures in A. canadense are undoubtedly the horizontal com- ponents typical of the Clathrodictyons. As the supposed Hydro- zoan affinities of the Stromatoporoids rest on the interpretation of these structures as tabulae, it is obvious that some doubt is cast on that interpretation by the facts stated above. Location. — Silurian, Southampton Island, Canada. Type specimen in Museum of Geol. Sur., Can. Species named after Dr. A. P. Low, Deputy Minister of Mines, Canada, by whom the specimen was collected. AcTiNODiCTYON NEPTUNi, sp. nov. — Plate XX, Figs. 5 and 6. The coenosteum consists of the cylindrical masses charac- teristic of the genus, some of the cylinders being as much as 30 mm. in diameter. In the type specimen several of these struc- tures are grouped together without any apparent regularity. The structure of the coenosteum is very like that of A. can- adense, but the centres of the cylinders are not occupied by corals or axial tubes. The laminae are of the same type as those of A. canadense; they are, however, a little more closely set and are not so distinctly inflected upwards into points. The large pillars are arranged in rows along the length of the cylinders, being about 8 mm. apart both horizontally and vertically. Vertical sections (Plate XX, fig. 5) show the radiating pillars traversing the general reticulation of Clathrodictyon- like fibre. Tangential sections (Plate XX, fig. 6) exhibit the cut ends [274] Parks : Silurian Stromatoporoids 35 of the minor pillars less distinctly and the ends of the large pillars more distinctly than in A. canadense. The round openings so characteristic of tangential sections of the genus are also to be seen. It will be observed that the differences between the two species are only of slight importance — possibly not of specific value at all. It is quite possible that A. canadense and A. nep- funi may be proved to be specifically identical by the discovery of intermediate forms. Locality. — Silurian, Southampton Island, Canada. Type .specimen in Museum of Geol. Sur., Can. Named after the ship Neptune, in which the specimens were brought from the north by Dr. Low. AcTiNODiCTYON KEELEi sp. nov. — Plate XIX, Figs. 5 and 6. The coenosteum in this species is most distinctly cylindrical and is represented by two fragments, one from a branch 45 mm. in diameter, and the other from a branch of 20 mm. diameter. The surface presents a pitted aspect and exhibits astrorhizae about 15 mm. apart. The concentric laminae are much more distinct than in the other species of the genus. An average of six appear in one mm. and they are connected by fairly distinct vertical pillars. The formation of these pillars by upward flexures of the laminae is apparent, but to a less degree than in the other species. The large pillars are evidently formed by the occasional continuation of the minor pillars through several laminae. Vertical sections (Plate XIX, fig. 6) show the fibre to be coarse, but this is perhaps due to mineralization, as certain parts of sections have a much finer appearance. The continuous pillars are irregularly disposed and are evidently composed of super- imposed minor pillars. The cut ends of the large canals of the astrorhizal canals are conspicuous. Tangential sections (Plate XIX, fig. 5) show the cut pil- lars, the vermiform severed laminae, and the round pores char- acteristic of the genus. A large astrorhizal canal is seen crossing the section. The figure is considerably restored, as the state of preservation of the specimen is far from good. [275] 36 Parks: Silurian Stromatoporoids This form is very closely related to an ordinary Clathro- dictyon; it differs only in the concentric arrangement of the laminae and in the superposition of certain pillars. I am con- vinced, however, that this feature is not merely accidental, as the continuity is much too extended to be so explained. It dif- fers from Actinostroma in that these long pillars are only occa- sionally developed and in the very different aspect of the trans- verse section. Although the generic features of Actinodictyon are less distinctly shown than in the other examples here de- scribed, I am inclined to regard the form as closely allied to those from Hudson Bay. Location. — Gravel River, Mackenzie District, Canada. Type .specimen in Museum of Geol. Sur., Canada. Named after Mr. Joseph Keele, of the Geological Survey of Canada, by whom the specimens were collected. Section B. (" Milleporoid " Group) Fa;;n7y— STROMATOPORIDAE, Nick. Gcnns — Stromatopora, Goldfuss Stromatopora amii, sp. nov. — Plate XIX, Figs. 7 and 8. The coenosteum of this species forms large hemispheric masses 15 cm. in diameter by 10 cm. in height. The point of attachment is not shown, but it was probably of much less diameter than the adult coenosteum. Instead of being formed of the ordinary concentric laminae, the skeleton consists of a series of coalesced tapering cylinders. Each cylinder is, how- ever, composed of the ordinary elements. The laminae are strongly arched upwards and are continuous o\'er the tops of the cylinders, so that these structures grow upwards as well as out- wards. This cylindrical manner of growth is common to many Stromatoporoids and is nothing more than a strong expression of the same tendency that leads to the formation of " mamelons." The upper ends of the cylinders appear on the surface as distinct rounded eminences situated from 10 to 15 mm. apart. The whole exterior surface is composed of a fine irregular network of vermiform ridges. Distinct astrorhizae appear to be present, but are not well defined. [276] Parks : Silurian Stromatoporoids 37 The most conspicuous elements in the fibre are the vertical pillars, which, although very irregular, appear to the number of three in the space of one mm. The horizontal laminae are spaced a little wider than the pillars and are even more irregular in their development. These vertical elements are not to be coh- sidered as pillars in the strict sense of the word ; they are seldom round, but appear to be the downward continuation into the coenosteum of the vermiculate ridges observed on the surface. The laminae are separated by the large canals of the astrorhizae and are pierced by well defined, round pores. Very conspicuous are the arched " tabulae " crossing the intercolumnar spaces and also appearing in the astrorhizal canals. There is no trace of the characteristic reticulated fibre of the genus Stromatopora. The coalescence of vertical and horizontal elements is, however, perfect, and the absence of the typical fibre may be the result of mineralization. Vertical sections (Plate XIX, fig. 8) show the rather pro- nounced pillars only in part of the plate, probably because at this point the vermiculate ridges are cut at right angles to their direction. A much more common appearance is that presented by the rest of the section. The thick, irregular and interrupted laminae are even less conspicuous than the pillars. The arcuate " tabulae " are distinct ; they may be observed crossing the ver- tical interspaces and also in the horizontal canals. In some sections these structures present many peculiarities, curving into circles, coalescing with each other, and even simulating the vesicular arrangement characteristic of the fundamental fibre of Stromatocerium. While not prepared to state that these struc- tures are not true zooidal tabulae, I am convinced that some other explanation would accord better with all the facts observed. Tangential sections (Plate XIX, fig. 7) show scarcely any trace of the cut ends of pillars as such. The general appearance is that of a true Stromatopora. Very conspicuous are the large round openings which appear where the section follows the course of a lamina. The peculiar " astrorhizal tabulae " are very characteristic. This form is most closely allied to Stromatopora citrteri, Nich. It differs, however, in the cylindrical manner of growth, [277] 38 Parks : Silurian Stromatoporoids the presence of astrorhizae and the character of the skeletal fibre. There is also a strong resemblance to Idiostroma roemeri which is shown in the vermiculate character of the surface, the cylin- drical manner of growth and the general structure of the skeletal fibre. The presence of axial canals in the cylinders is extremely doubtful; they are certainly absent for the most part, but some evidence of their occasional occurrence is not wanting. The minute fibre appears to be dense; if this is really the case the form can be ascribed to neither Stromatopora nor Idiostroma, but it might be considered as an extreme example of the genus 'Actinodictyon. Such a determination is, however, carrying the interpretation of the " tabulae " as residual laminae beyond the point of absolute proof. This species is particularly interesting in that it shows a possible connection between Clathrodictyon and Actinodictyon on the one hand, and Stromatopora and Idiostroma on the other. Locality. — Silurian, Southampton Island, Canada. Type specimen in the Museum of the Geol. Sur., Canada. Named after Dr. H. M. Ami, of the Geol. Sur., Canada. Stromatopora carter:, Nicholson. — Plate XIX, Figs. 9 and 10. Stromatopora carteri, A'ich., Mon. Brit. Strom., p. 174, pi. i., figs. 6 and 7, pi. xxiii, figs. I and 3, 1891. Stromatopora carteri, A'icli., Ann, and Mag. Nat. Hist., ser. 6, vol. vii, p. 314, pi. ix, figs. 5 and 6, 1891. Stromatopora carteri, Parks, Ottawa Naturalist, vol. xxii, p. 28, 1908. Nicholson's description is as follows : — " The coenosteum in this species is of considerable size, massive, irregular in shape, and composed of gently undulating or curved latilaminae, which vary from 2 to 4 or 5 mm. in thickness in their central portion. The under surface and mode of attachment are not known, but the upper surface is without distinct eminences or ' mamelons,' and shows simply an irregular vermiculate tuberculation. Astro- rhizae are not developed in any recognized form. " As regards internal structure, the skeletal fibre is about one-sixth mm. in diameter, and is coarsely porous. Vertical sec- tions show that each latilamina is composed of very distinctly developed radial pillars, which are separated from one another bv equally distinct zooidal tubes, and which reallv run con- [278] Parks : Silurian Stromatoporoids 39 tinuously from the bottom to the top of each latilamina ; though they appear to be more or less broken up, if the plane of section is slightly oblique. About seven radial pillars, with the inter- vening zooidal tubes, occupy a space of two mm. measured trans- versely. The zooidal tubes are furnished with a moderate num- ber of well developed transverse partitions or 'tabulae.' The radial pillars are connected at varying intervals by irregular horizontal or oblique processes, but these do not give rise to dis- tinct ' concentric laminae,' and the skeleton thus forms a loose and open reticulation, in which the vertical elements are far more conspicuous than the horizontal. As a result of this, tangential sections show the cut ends of the radial pillars, either as separate structures, or, more usually, as united by the irregular horizontal processes already spoken of in such a way as to give rise to ver- miculate and sinuous rows, which inosculate with one another and form a lax network. A characteristic feature is the peculiarly loose and open nature of the reticulated coenosteal tissue. Another characteristic feature is the total or almost total absence of the branched and radiating astrorhizal canals, which are so conspicuous in most species of Stromatopora. In most examples of the present form no trace whatever of these structures can be found, and in none is there more than the merest indications of their existence. The surface, therefore, is simply smooth or gently undulated, and is entirely without ' mamelons.' " S. carteri, as far as known, is entirely confined to the Wenlock limestone of Britain ; and my specimens have been prin- cipally obtained from the single locality of Ironbridge, in Shrop- shire, where the species is not altogether uncommon, though vastly more rare than is 5. typica, Rosen. The species has not hitherto been certainly recognized in the Silurian rocks of Got- land and Esthonia. Mr. Whiteaves has submitted to me for examination a fragment of a species of Stromatopora, obtained from a loose boulder on the banks of the Hayes River, Hudson Bay Territory, which very closely approaches in its characters to 5". carteri, though my material is not sufficient to justify me in asserting that it is absolutely identical wih the latter." To this full description there is nothing to be added; a [279] 40 Parks : Silurian Stromatoporoids small and ill preserved specimen from Hudson Bay approaches very closely to Nicholson's description and figures. The coarsely fibrous character of the tissue cannot be discerned, as the speci- men is too much altered; in all other respects, however, it con- forms very well except that the transverse section (Plate XIX, fig'. lo) seems somewhat finer than the similar section figured by Nicholson. The resemblance of the vertical sections (Plate XIX, fig. 9) to Nicholson's (Mon. Brit. Strom., pi. xxxii, fig. 2) is perfect except for the lack of preservation in the minute fibre. Locality. — Silurian, Station 641, W. J. Wilson, Pagwa- chuan River, Hudson Bay, Canada. Boulder on Hayes River, Hudson Bay, Canada. Stromatopora wilsoni. Parks. — Plate XIX, Figs. 11 and 12. Stromatopora wilsoni, Parks, Ottawa Naturalist, vol. xxii, p. 28, 1908. As this species is known from a single fragment only, it is impossible to do more than infer the shape of the coenosteum. Judging from the curvature of the laminae it probably was of hemispherical shape and in all probability of considerable size. The coenosteum is made up of distinct latilaminae, about 8 mm. in thickness. Instead of being simply concentric the horizontal elements are folded upwards into low rounded monticules, about 15 mm. apart, so that the surface presents a botryoidal appear- ance. This curvature of the laminae added to the very poor state of preservation makes it difficult to arrive at a clear under- standing of the structure of the tissue. Distinct vertical pillars occur to the number of five or six in the space of one mm. It has not been observed that these pillars are continuous through the whole thickness of a latilamina. The vertical interspaces be- tween the pillars are less in diameter than the pillars them- selves. The horizontal laminae are likewise thick, and, though less pronounced than the pillars, they may be traced continuously for considerable distances. Astrorhizal systems are present and there is some indication that they are superimposed, giving rise to axial canals. The skeletal fibre, though now destroyed, was probably of an original fibrous character. The vertical section (Plate XIX, fig. 12) is so much restored that too much confi- dence must not be placed on it as a means of identifving this [280] Parks : Silurian Stromatoporoids 41 species. The vertical elements are seen to be more pronounced than the horizontal ; the change in direction of the laminae is exhibited and the conspicuous cut ends of the astrorhizal canals appear as round spaces. Tangential sections (Plate XIX, fig. 11) are more satis- factory. The severed laminae appear as dense tissue pierced by the round openings representing the vertical pores. The branches of the astrorhizal systems appear as irregular openings ; they are discontinuous because, owing to the curvature of the laminae, it is impossible to prepare a section which lies in any one interlaminar space. While convinced that this species is distinct from any other, I must admit that the description is very inadequate; it is only the desire for completeness that induces me to describe the form at all. Better material will doubtless be obtained and result in modifications of the above description. That the form is distinct I believe for the following- reasons : The fibre is much finer than that of Stromatopora cartcri and 5". amii on the one hand, and coarser than that of .S". constellata on the other. The astro- rhizal systems are more distinct than in 5". carteri, and the hori- zontal laminae are better developed than in that species. While I do not consider the botryoidal manner of growth as of specific value, I have never observed it in specimens of 6". constellata, and I have never prepared tangential sections of that species in which distinct, connected astrorhizal canals did not appear. The species differs from S. constellata also in the more irregular course of the pillars and the occasional very large openings of the astrorhizal canals as seen in vertical section. However, with a single, very poorly preserved specimen as type, the species must be regarded as resting on an insecure foundation. Locality. — Silurian, Pagwachuan River (near mouth) Hudson Bay, Canada. Type specimen in Museum of Geol. Sur., Canada. Stromatopora constellata. Hall For synonomy and description, see Niagara Stromato- poroids, page 44. This form {S. hndsonica, Dawson,) is common in the Silurian area west of Hudson Bay, but there is no example in the large collection from Southampton and Beechy Islands. [281] 42 Parks: Silurian Stromatoporoids III. Additional Niagara Stromatoporoids AcTiNosTROMA TENUissiMUM, sp. iiov. — Plate XVIII, Figs. 2, 3, 9 and 12. With regard to the general shape of the coenosteum there is little to be said, as the only specimen is intercalated between two different species of Stromatopora: it appears, however, to have had an incrusting habit with a tendency to form a hemi- spherical mass. The organism is made up of extremely fine vertical fibres, about 24 of which appear in the space of one mm., and delicate connecting arms, arising in whorls after the manner of a typical Actinostroma. The horizontal " laminae " formed by these arms are not nearly as regular and persistent as the vertical elements. The interspaces between the fibres are much larger than the fibres themselves, so that the general appearance of sec- tions is that of a delicate open network. The coenosteum is interrupted at intervals of about 2 mm. by horizontal (concen- tric) partings, giving to the skeleton a pronounced latilaminar structure. Certain large open canals or interspaces traverse the general tissue, but these are so far apart and so irregular that I am of the opinion that they are quite foreign to the organism. A^ertical sections which follow the course of the pillars (Plate XVm, fig. 3) are very hard ti> prepare owing to the extreme tenuity of these elements. However, the examination of an extended section leaves no doubt as to the practical con- tinuity of the fibres through a latilamina. The horizontal laminae are less continuous but may be distinctly seen in certain parts of a section. Tangential sections (Plate XVIII, fig. 2) show the cut ends of the pillars with connecting arms of about the same diameter, so that the general appearance is that of a " hexactinellid " network. To exhibit the gross anatomy photographs rather than drawings have been inserted in order to eliminate any possible [282] Parks : Silurian Stromatoporoids 43 " personal equation," as this species must be considered of the utmost importance in discussing the affinities of the Stromato- poroids. The drawings of vertical and tangential sections on a scale of magnification of 30 were prepared, as were the other figures on the same scale, on the photographic print, from which the silver was afterwards removed. Plate XVIII, fig. 12, shows that the fibre differs from that of an ordinary Actinostroma only in its extreme tenuity. The tangential section (Plate XVIII, fig. 9), except for a little sharper definition of the radial pillars, is precisely like that of Stromatopora constellata. In its general anatomy this species is certainly very closely related to Actinostroma astroites, Rosen, It is, however, much finer, as in that species only from 12 to 15 fibres appear in the space of one mm. The resemblance of the two species is also seen in the lack of continuity in the horizontal laminae and in the pronounced latilaminar subdivisions. The presence of astro- rhizae has not been established, but, lacking a surface, and con- sidering the extreme delicacy of the fibre, it is impossible to say that they do not exist. Nicholson has commented on the resemblance of Actino- stroma astroites to Stromatopora typica, but he has dismissed the comparison on the ground that the resemblance is due merely to the alteration of the latter species. Whatever justification there may be for his view, I am convinced that in Actinostroma tcnuissimum and in Stromatopora constellata we have two organisms of exactly the same ultimate skeletal fibre, differing only in the fact that the latter species is pierced by a series of vertical and a series of horizontal canals. Further, it is very sig- nificant that the two species should be found growing together, conforming to the same curvature, and separated by no greater an interval than that which divides the latilaminae of the former species. It is further significant that the Helderbergian species of Stromatopora and of Syringostroma herein described present the same ultimate structure. The conclusion seems obvious that the vertical and horizontal canals are not essential to the life of the organisms as a class. As long as these canals are interpreted as pores, their presence or absence may not necessarily be vital, but if the vertical passages are considered as the habitations of [283] 44 Parks : Silurian Stromatoporoids zooids, those zooids are an essential part of the organism, and it is difficult to explain how life could be maintained without them. The species must be considered as an Actinostroiiia, despite its extreme delicacy and its intimate relation to Stromatopora. Should the conclusions given above be more fully established, a complete revision of the classification of Stromatoporoids will be required. Locality. — Niagara limestone, Scoharie County, New York. The type and only specimen is in the Museum of the State of New York, Albany, N.Y. Genus — Aulocerium, gen. nov.* Nicholson has made the genus Lahechia include those forms originally described as Stromatoceriiim, so that species of which the coenosteum consists essentially of a series of vesicles pierced liy pillars must be ascribed to Lahechia. The genus about to be described is precisely like Stromatocerimn, or Labecliia, in the fact that the skeleton is composed of rows of vesicles, with the convexities directed outwards. It differs, however, in that there are no pillars whatever. Instead of possessing these radial struc- tures the coenosteum is pierced by distinct tabulate tubes of large size. These tubes resemble the " Caunopora tubes " presented by many species, but \\hether they are essential to the organism or whether they may be absent in some examples of the same species it is impossible to say, as only one fragment is available. In this fragment the tubes are an integral part of the coenosteum. The one species here included cannot be placed under Lahechia, as there are no radial pillars. It cannot be called Stromato- cerinni. as Nicholson has abandoned that genus, and it does not seem advisable to revive it. AuLocERiUM sAVAGEi, sp. uov. — Plate XVIII, Figs. 13 and 15. The coenosteum is known from a small fragment about 30 mm. sq., with a thicknes'^ of .20 mm. The general form was probably laminar, and the organism probably attained consider- * avlb^, tube ; uripiou, honeycomb.— For this generic name I am indebted to Professor W. S. Milner, of the University of Toronto. [284] Parks : Silurian Stromatoporoids 45 able dimensions. The skeletal matter consists of rows of vesi- cles, with the convexities outwards, about six vesicles appearing in the space of i mm., measured vertically. The maximum extent of a vesicle horizontally is something over ^ mm. The cavities appear to have been hollow, and their walls are very delicate. At a fairly uniform interval of 2 mm. large tubes of about J mm. diameter pierce vertically through the vesicular tissue. The walls of the tubes are thin— of about the same calibre as the fibre of the vesicles. The tubes are crossed by four or five tabulae in the distance of i mm. These tabulae are strongly curved downwards, and are likewise of about the same weight as the walls of the vesicles. I am of the opinion that these tubes are not foreign to the organism, for the following reasons. They arise suddenly, of full diameter, at various levels in the coenosteum. They are of the same diameter as a vesicle. They are evenly spaced. Their substance is of the same micro- scopic appearance as the rest of the coenosteum. They are inti- mately united, even blended, with the vesicular tissue. Vertical sections (Plate XVIII, fig. 13) exhibit the general features described above. Tangential sections (Plate XVIII, fig. 15) show the severed vesicular walls and the cross-sections of the tubes. The inner circle to be observed in some of the tubes represents the cut edges of the depressed tabulae. The vesicles are not equal in size, and bands of smaller cells appear at certain intervals. The severed edges of the vesicles of such a zone appear in the upper right-hand part of the figure. It is possible that these small meshes mean that the tissue of the vesicular walls is not continuous, but reticulate. No trace of discontinuity is, however, shown in vertical sections. This species, the only representative of the genus, is very interesting, in that it may throw some light on the origin and significance of the pillars in Labechia. Here we have open, tabu- late tubes taking the place of, and doubtless performing the same function as, the pillars of Labechia. If these tubes represent the habitation cavities of zooids, it is a fair assumption that the pillars of Labechia are of the same nature. Locality. — Niagara limestone, within five feet of the base, [285] 46 Parks: Silurian Stromatoporoids Wilmington, 111. Part of the type is No. 1267N, University of Toronto Museum ; the other part is in the possession of Mr. T. E. Savage, Geological Survey, Illinois, Urbana, 111. The species is named for Mr. Savage, who collected the material and through whose kindness the opportunity of description was afforded. Strom ATOPORA constellata. Hall. — Plate XVII, Figs. 10 and 1 1 ; Plate XVIII, Fig. 8. For synonomy and desci-iption, see Niagara Stromatopo- roids, page 44. The examination of a number of excellently preserved ex- amples of this species and the comparison of them with the similar species from the Helderbergian has led me to materially alter the opinion previously expressed as to the interpretation of the structures observed. Without the information now at hand there was no reason to doubt Nicholson's conclusion that the vertical pores represent the habitation cavities of zooids. I am now inclined to believe that this opinion is erroneous. Growing beneath a colony of 6". constcllata I have found an exceedingly fine-grained Stromatoporoid having the typical structure of an Actinostroma. The vertical elements in this species are about r}j mm. apart; the horizontal laminae are less pronounced, but, with the vertical threads, build up a reticular network, which differs from Actinostroma onlv in its extreme tenuity. The ultimate fibre of the overlying 5. constellata has exactly the same structure. Further, this same reticular fibre is exhibited by Syringostroma ristigoiichcnsc. S. ccntrotum, S. consimilc, Stromatopora foveolata, and other Helderbergian species. It seems an obvious conclusion that all these species represent similar fine reticulations traversed h\ systems of ver- tical and horizontal pores. The existence of the fine tissue with- out pores, especially vertical pores, proves that zooids are not essential to the life of the organism. This is at least good nega- tive evidence that the vertical pores of 5". constcllata are not zooidal in character. In the various Helderbergian species named above the horizontal laminae may consist of a single layer of fibres or of several. The laminae are separated by the ramifi- cations of the astrnrhizal canals, which leave certain vertical [286] Parks: Silurian Stromatoporoids 47 supports of the original tissue between the laminae. If these supports are superimposed and circular, we have the genus Syringostroma; if they are not circular, whether superimposed or not, we have a true Stromatopora. S. constellata belongs exactly here. It consists of a fine Actinostroma-like reticulation, with alternate layers of non-areolated tissue (the laminae) and of open fenestrated and canal-traversed tissue (the interlaminar spaces). There is absolutely no difference, in character or origin, between the tissue of the laminae and that of the pillars. The interlaminar passages are placed in communication with one another vertically, to a greater or less extent, by finer pores pass- ing through the laminae (the so-called zooidal pores). In the first place, if these pores are zooidal in character, why do they always communicate with the astrorhizal canals? This com- munication is seen to perfection in vertical sections of Stromato- pora foveolata (Plate XVII, figs. 5 and 6). In most figures of Stromatopora the connection between vertical and horizontal tubes is complete — so complete that it is impossible to state where the one ends and the other begins. In my own figures of 5. constellata (Niagara Strom., Plate XIII, figs. 7 and 8), if we interpret the vertical pores as zooidal, what has become of the astrorhizal canals? It is absolutely necessary to restrict the imaginary zooid to a single lamina. This leads to an absurdity : Some laminae are only of the thickness of a single fibre, there- fore the zooid could have a vertical extent of about -H mm. In S. constellata the laminae are comparatively thick, and the pores are seen to be crossed by " tabulae." This circumstance is used by Nicholson as the strongest argument as to the zooidal character of the tubules. These so-called tabulae are merely ■fibres of the general reticulation passing across the pore. If, with Nicholson, we consider the pores to be continuous through several laminae, each " tabula " is generally coincident with a lamina. Sometimes this is not the case, and a " tabula " crosses in an interlaminar space. The explanation is the same. In Plate XVII, fig. 10, occasional fibres are seen to cross the open spaces between the gross vertical elements in the parts of the section which are evidently interlaminar. Such a fibre, seen in vertical section, could be easily mistaken for a " tabula." [287] 48 Parks: Silurian Stromatoporoids It would be premature to state that no zooidal tubes occur in any Stromatoporoids, but I am convinced that there is no direct evidence of their existence in any species of the genus Syringostroma, and that the vertical pores of 5. constellata and its allies from the Helderbergian are not to be interpreted as such. Both tangential and vertical sections (Plate XVII, figs. lo and ii) show that 5'. constellata is closely allied to Syringos- troiua. The columns in the interlaminar spaces are sometimes quite round, but the laminae are much too thick and separated by too slight an interval to allow of its reception in that genus. A tangential section, magnified thirty times, is introduced in order to make a comparison with the other species possible. (Plate XVIII, fig. 8.) Locality. — Niagaran. The distribution of this species has been given in Niagara Stromatoporoids. The figures here pro- duced were prepared from sections No. 851 and X'o. 853, New York State Museum, Albany, N.Y. Stromatopora clarkei, sp. uoz: — Plate XVII, Fig. 12; Plate XVIII, Fig. 1. This species is known from one small specimen only, and this is so intergrown with other species that no information as to the general shape of the coenosteum is available. The horizontal and vertical elements are poorly defined, so that the structure of a typical Sfromatopora is presented. Division of the coenosteum into latilaminae is not strongly marked, but is apparent. The ultimate fibre is porous, but a rec- tangular arrangement of the constituent threads has not been observed. The tissue is pierced by large vertical passages and is traversed horizontally by anastomosing astrorhizal canals. Vertical sections (Plate XVII, fig. 12) show from four to five vertical pseudo-pillars in the space of one mm. These pil- lars are not strictly parallel and the vertical passages between them may, in some cases, be quite flexuous. Some of these pas- sages are continuous across a latilamina — a distance of from i^ to 2 mm. Owing to the small size of the specimen, it was impossible to prepare many sections, so that a more regular [288] Parks: Silurian Stromatoporoids 49 appearance than that presented by the figure might be obtained from better oriented sections. Horizontal elements are very slightly shown, but appear as thin lines crossing the vertical passages. These structures may be interpreted as tabulae, but I have great doubt as to the cor- rectness of this view, and for the following reasons. The fine lines are continuous across several vertical passages, at least in many instances. Similar lines are seen to run horizontally through the horizontal canals. The lines are seen to partly maintain their identity in crossing the pillars. Tangential sections (Plate XVIII, fig. i) present a very characteristic appearance. The horizontal canals of the astro- rhizal system lose little in diameter as they depart from the centra, and as they have about the same calibre as the vertical passages and are confluent with them a uniform open reticula- tion results. The even character of the section is further in- creased by the slight development of horizontal laminae; so that no conspicuous darker patches, pierced by pores, are observable, as is the case in most examples of the genus. The species is easily distinguished from S. cotistellata by its coarser character, its less pronounced laminae, and the charac- teristic arrangement of its astrorhizal canals. Locality. — Niagara limestone, Scoharie County, New York. The type and only specimen is in the Museum of the State of New York, Albany, N.Y. [289] 50 Parks: Silurian Stromatoporoids EXPLANATION OF PLATES XVI TO XX (All figures ndt otherwise speciHed are enlarged ten times) 6 — Syringostroma centrotum. Plate XVI Fig. l—Clathrodictyon jewetti. Vertical section. Cedarville, N.Y Fig. 2— " " Tangential section. Cedarville, N.Y. Figs. 3 and 4 — Syringostroma ristigouchense. Vertical section. Cap Bon Ami. " " Tangential section. Cap Bon Ami. Vertical section. Cedarville, N.Y. From N. Y. State Museum, slide no. 1757. Vertical section. Cedarville, N.Y. From N. Y. State Museum, slide no. 1757. Tangential section. Cedarville, N.Y. From N. Y. State Museum, slide no. 1756. Tangential section. Cedarville, N.Y. From N. Y. State Museum, slide no. 1756. Vertical section. Outlet of Skaneateles lake, N.Y. From type in Yale Museum. Tangential section. Outlet of Skaneateles lake, N.Y. From type in Yale Museum. Vertical section. Scoharie county, N.Y. From U.S.N. M., no. 36845. Fig. 5— Fig. 6—: Fig, 7— Fig. 8— Fig. 9— Fig. 10 — Syringostroma consimile. Fig. II- Fig. 12— Fig. Fig. Fig. Fig. 4- Plate XVII I — Syringostroma barretti. Tangential section. Clarksville, N.Y. From N. Y. State Museum, slide no. 735. 2 — " " Vertical section. Clarksville, N.Y. From N. Y, State Museum, slide no. 736. 3 — Syringostroma niicroponim. Tangential section. Cedarville, N.Y. From type slide in Yale Museum. Vertical section. Cedarville, N.Y. From type slide in Yale Museum. The right of the figure is restored. Vertical section. Cedarville, N.Y. From N. Y. State Museum, slide no. J736. Vertical section. Cedarville, N.Y. ^ N. Y. State Museum, slide no. Tangential section. Cedarville, N.Y. N. Y. State Museum, slide no. Vertical section. Port Jervis, N.Y. Am. Mus. Nat. Hist., no. 2270 — 3. Tangential section. Port Jervis, N.Y. From Am. Mus. Nat. Hist., no. 2270 — 3. Fig. 10 — Stromatopora constcUata. Tangential section. Scoharie county, N.Y. From N. Y. S. Museum, slide no. 8si. Fig. Fig. Fig. Fig. 5 — Stromatopora fovcolata. Fig. 7- 8 — Stromatopora corallifcra. From 1736. From 1733- From [290] Parks: Silurian Stromatoporoids 51 Fig. II. — Stromatopora constellata. — Vertical section. Scoharie county, N.Y. From N. Y. S. Museum, slide no. 852. Fig. 12 — Stromatopora clarkei. Vertical section. Scoharie county, N.Y. From N. Y. State Museum specimen. Plate XVIII Fig. I — Stromatopora clarkei. Tangential section. Scoharie county, N.Y. From N. Y. State Museum specimen. Fig. 2 — Actinostroma tenuissimum. Tangential section. Scoharie county, N.Y. From N. Y. State Museum specimen. Fig. 3 — " " Vertical section. Scoharie county, N.Y. From N. Y. State Museum speci- men. Fig. 4 — Stromatopora foveolata. Tangential section x 30. As above. Fig. S—Syringostroma barretti. Tangential section x 30. As above. Fig. 6 — Syringostroma centrotum. Tangential section x 30. As above. Fig. 7 — Stromatopora corallifera. Tangential section x 30. As above. Fig. 8 — Stromatopora constellata. Tangential section x 30. As above. Fig. 9 — Actinostroma tenuissimum. Tangential section x 30. As above. Fig. 10 — Stromatopora foveolata. Vertical section x 30. As above. Fig. II — Syringostroma centrotum. Vertical section x 30. As above. Fig. 12 — Actinostroma tenuissimum. Vertical section x 30. As above. Fig. 13 — Aulocerium savagei. Vertical section. Wilmington, 111. From Geol. Sur. 111. specimen. Fig. 14 — Syringostroma consimile. Tangential section x 30. As above. Fig. IS — Aulocerium savagei. Tangential section. Wilmington, 111. Geol. Sur. 111. specimen. Plate XIX Fig. I — Actinostroma tenuifilatum intlectum. Vertical section. Pagwa- chuan river, Canada. Geol. Sur. Can. speci- men. Fig. 2 — " " " Tangential section. Pagwa- chuan river, Canada. Geol. Sur. Can. speci- men. Fig. 3 — Actinostroma franklinense. Vertical section. Beechy island, Lan- caster sound, Canada. Geol. Sur. Can. specimen. Fig. 4 — " " Tangential section. Beechy island, Lan- caster sound, Canada. Geol. Sur. Can. specimen. Fig. 5 — Actinodictyon keelei. Tangential section. Gravel river, Mackenzie river, Canada. Geol. Sur. Can. specimen. Fig. 6 — " " Vertical (longitudinal) section. Gravel river, Mackenzie river, Canada. Geol. Sur. Can. specimen. Fig. 7 — Stromatopora amii. Tangential section. Southampton island, Hud- son bay. Geol. Sur. Can., slide no. 2234. Fig. 8 — " " Vertical section. Southampton island, Hudson bay. Geol. Sur. Can., slide no. 2235. [291] 52 Parks: Silurian Stromatoporoids Fig. 5 — Stromatopora carteri. Vertical section. Pagwachuan river, Hud- son bay, Canada. Geol. Sur. Can. specimen. Fig. 10 — " " Tangential section. Pagwachuan river, Hud- son bay, Canada. Geol. Sur. Can. specimen. Fig. II — Stromatopora wilsoni. Tangential section. Pagwachuan river, Hud- son bay, Canada. Geol. Sur. Can. specimen. Fig. 12 — " " Vertical section. Pagwachuan river, Hud- son bay, Canada. Geol. Sur. Can. specimen. Fig. Fig. Fig. Fig. Fig. Fig. Fig, Fig, Fig, Plate XX I — Actinodictyon canadcnse. Vertical section. Southampton island, Hudson bay. Geol. Sur. Can. specimen. 2 — " " Tangential section. Southampton island, Hudson bay. Geol. Sur. Can. specimen. 3 — Actinodictyon lowi. Vertical section. Southampton island, Hud- son bay. Geol. Sur. Can. specimen. 4 — " " Tangential section. Southampton island, Hud- so.n bay. Geol. Sur. Can., slide no. 2271. 5 — Actinodictyon neptuni. Vertical section. Southampton island, Hud- son bay. Geol. Sur. Can. specimen. 6 — " " Tangential section. Southampton island, Hudson bay. Geol. Sur. Can. specimen. 7 — Clathrodictyon drummondense. Vertical section. Southampton island, Hudson bay. Geol. Sur. Can. specimen. 8 — " " Vertical section. Southampton island, Hudson bay. Geol. Sur. Can. specimen. 9 — Clathrodictyon cystosum lineatum. Beechy island, Lancaster sound. Geol. Sur. Can. specimen. [292] UNIVERSITY OF TORONTO STUDIES GEOLOGICAL SERIES NO. 6 Plate 16 9 PARKS, PHOTO MIC. AND O E 1_ COCKAYNE, BOSTO^ UNIVERSITY OF TORONTO STUDIES GEOLOGICAL SERIES NO. 6 Plate 17 A PARKS, PHOTO MIC. AND DEL COCKAYNE, BOSTOr- UNIVERSITY OF TORONTO STUDIES GEOLOGICAL SERIES NO. 6 Plate 1>&1'.