SSS eee ae —— : er, Paes 2, a : ae ae, ana 2 ~—- H ( Me me Wy af AR TI OT | i ref | =—s> \ ==. i Honea fah Nh? eel aa OPM fd we Wig rtd Sida Midd tic a Gay nN i Weaken ms i NERD a i , + : Maar’ MA WN Wah Talay eeao Mala Fy tas ea aoa) We fh wt pe : NY ie yey aly Nai ei Line NEN Wi SAN enteral aM Saat aee dL v yon Yi ertay i KiNG Lraley AY Bry Aah dor nthe Bh Vs ae i f AMLUDK A APH it evdcoeih mie eraer ay \ BSF a Uy { he lah \ as BAU ROWER A iE) a ¥ ve } nN eh Pr ANOFS sige LOA Deny Ai 7 yh he Ney Ni r ADs tA Le aN y h Dera sth whit is Ww} aN h win: i t! Gusine ait uy cM f N wath eS iy ely ee ete of hin Rp aNy AD ULtiba Mag! D + i , a , j ied Wat? gee) teyiw J g a “te ; Wis ol 168 ZOOLOGICAL BULLETIN EDITED BY C.O. WHITMAN anp W. M. WHEELER THE UNIVERSITY OF CHICAGO VoLumeE II BOSTON, U.S.A. GINN & COMPANY, PUBLISHERS The Atheneum Press 1899 ALL RIGHTS RESERVED CONTENTS “OF VOL IE BRRADAS = ee ee ee In No. 2, Vol. II, of this Auwlletzn the following corrections are to be made: On page 83, line 21. In place of spino-occipital read occipito-spinal. On page 84, line 4. In place of spino-occipital read occipito- spinal. On’ page 85, lines 2,6,24, 26, 27, 32, 35.._In place of spino- occipital read occipito-spinal. The Embryology of the Apterygota III]. Wm. E. Ritter. A Few Facts Concerning the Relationships and Reproduction of Some Bering Sea Tunticates . IV. Epwarp PHEtLrPs ALLIs, JR. The. Homologies of the Occipital and First Spinal Nerves of Amia and Teleosts ll 69-76 77-81 83-97 rE: LU III. IV. CONTENTS (OF VOL. IE No. 1. — July, 1808. ETHAN ALLEN ANDREWS. Filose Activity in Metazoan Eggs . - Cuares B. WILson. Activities of Mesenchyme in Certain Larvae ©; -PaHay. Observations on the Genus of Fossil Fishes called by Professor Cope, Portheus, by Dr. Letdy, Xiphactinus No. 2. — October, 1898. HENRY LESLIE OSBORN. Observations on the Anatomy of a Species of Platyaspis found Parasitic on the Unt- ontidae of Lake Chautauqua AGNES M. CLaypo.e, Pu.D. The Embryology of the Apterygota Wm. E. RITTER. A Few Facts Concerning the Relationships and Reproduction of Some Bering Sea Tunicates . EDWARD PHELPS ALLIS, JR. The Homologtes of the Occipital and First Spinal Nerves of Amia and Teleosts ul PAGES 25-54 55-67 69-76 77-81 1V ie 1B IV. 10g PET: VI. CONTENTS: No. 3. — December, 1808. GEORGE WILLIAM HUNTER, JR. Notes on the Finer Structure of the Nervous System of Cynthia Partita (Verrill) Uric DAHLGREN. The Maxillary and Mandibular Lreathing Valves of Teleost Fishes FLORENCE PEEBLES. The Effect of Temperature on the Regener- ation of Flydra KATHARINE Foot AND ELLA CHURCH STROBELL. Further Notes on the Egg of Allolobophora Foetida No. 4. — February, 1899. M. A. Wittcox, Pu.D. Notes on the Occipital Region of the Trout, Trutta Farto . CoH. TURNER: Notes on the Mushroom Bodies of the Inver- tebrates. A Preliminary Paper on the Comparative Study of the Arthropod and Annelida Brain GusTAV Ersen, Pu.D. Notes on North-American Earthworms of the Genus Diplocardia Maurice A. BIGELow. Notes on the First Cleavage of Lepas . CHARLES A. KoFoIp. On ee Specific Identity of Cotylaspis [n- signis Leidy and Platyaspis Anodontae Osborn . CE. Mc@rune A Peculiar Nuclear Element in the Mate Reproductive Cells of Insects PAGES 99-115 117-124 125-128 129-150 151-154 179-186 187-197 PELE: IV. Ne Vile IIE rT: IV. CONTENTS. No. 5. — June, 1899. C. H. TURNER. A Male Erpetocypris Barbatus Forbes B. F. KINGSBURY. The Reducing Divisions in the Spermato- genesis of Desmognathus Fusca . MicHakEL F. GUYER. Ovarian Structure in an Abnormal Pigeon . CHARLES W. Harairvt. Some Interesting Egg Monstrosities En: Case. A Redescription of Pariotichus Incisivus Cope Howarp AYERS. On the Pithecoid Type of Ear in Man No. 6. — September, 1899. J. Pravrair McMorricu, Contributions on the Morphology of the Actt- nozoa. V. The Mesenterial Filaments in Zoanthus Sociatus (Ellis) . MinniE Marie ENTEMAN. The Unpaired Ectodermal Structures of the Antennata . GARRY DE Norp Hovueu, M.D. Synopsis of the Calliphorinae of the United States FRANK W. PICKEL. The Accessory Bladders of the Testudinata . PAGES 199-202 251—273 275-282 Pi Sry iT i \ \a n he ‘ Volume II. July, 1898. Number I. ZOOMe GVEAT BULLETIN FILOSE ACTIVITY IN METAZOAN EGGS. ETHAN ALLEN ANDREWS. One who has not studied living Foraminifera or Radiolaria can get only an inadequate conception of the remarkable activities exhibited by the delicate protoplasmic extensions that form in such protozoans the thread-like pseudopodia. The figures and descriptions in text-books of zodlogy, in Verworn’s Physiology, in Biitschli’s Prozozoa, or in special monographs naturally fall short of complete expression of the changeableness as well as the extreme delicacy of certain of these processes, though they teach us that the sensitive, contractile, codrdinating powers of protoplasm may here be expressed in filaments of exceeding tenuity and inconstancy of form and position,—in flowing, liquid, apparently homogeneous protoplasm. Such filose phenomena were practically unknown in Metazoa till a recent paper! described their occurrence in the eggs, polar bodies, blastulz, gastrule, and larve of certain echinoderms. © Here the cells put forth protoplasmic threads of excessive delicacy, that may branch and anastomose, elongate or shorten. By means of such filose processes the cells become connected amongst themselves, and, as these connections are living material comparable to the sensitive pseudopodia of many Protozoa, their importance in understanding the coérdination of cells and their subordination to the entire mass, during the animals’ development, was emphasized.} Having been shown the filose threads in living starfish eggs, I have been able to observe the less attenuated ones present in 1 Andrews, G. F., “Some Spinning Activities of Protoplasm in Starfish and Echinus Eggs,” Journ. of Morph. Vol. xii. 18097. 2 ANDREWS. [Vor Ik the eggs of several other animals examined, and thus to extend the known occurrence of filose activities amongst Metazoa so widely that its importance seems strengthened and the proba- bility of its still wider existence increased. The first eggs examined, the frog’s eggs in cleavage and gas- trula stages, yielded when studied alive only the amceboid movement described by Roux; certain sections, however, showed intercellular connections that lead me to expect filose phenom- ena to be present here. A large number of sections of cleavage and larval stages in various frogs and in the salamander Ambly- stoma punctatum were carefully studied. In many sections of the latter animal, prepared and kindly loaned to me by Prof. C. B. Wilson of Westfield, Mass., as well as in certain frog’s eggs, un- doubted intercellular connections exist; but as their filose nature is not demonstrated, they will be but briefly noticed here. In the larva when the medullary folds are closed and the split meso- blast nearly fused on the ventral side, intercellular connections were seen between the large yolk cells, between mesoderm and mesenchyme cells, between the ectoderm cells on opposite sides of the nerve tube, between the ectoderm and mesoderm, and between the entoderm and mesoderm; in fact, cells in all germ layers and in each layer connect with those in another layer and with those in the same. Eliminating the deceitful appear- ances produced by coagulation of liquids between cells, by coagulation of fixative, by vacuolization and shrinkage of the superficial parts of cells, by the throwing off of pellicles, by the edges of drops and vesicles, and by fragments of vitelline mem- brane, as well as by scratches upon slide and cover glass, there still remained the above intercellular connections of undoubted protoplasm. These varied from fine filaments to broad bridges, and were either clear or contained some of the pigment granules of the egg. That they were filose in nature was indicated by their proportions and mode of origin and insertion; yet there was not decisive evidence that they were not produced in other ways either in the normal egg or in the egg when dying. - The next eggs, those of the annelid Serpula, were only exam- ined alive and showed filaments passing out from the surface of the egg toward and to the membrane, both before and after the NOt.) fi8OSE ACTIV IN METAZOAN EGGS. B first cleavage, as elsewhere mentioned.! Later, in the gastrula stage, filaments were seen passing from the ectoderm to the membrane. However, no undoubted case of filaments connect- ing cells was observed in the comparatively few eggs studied. The eggs of a nudibranch mollusk, Jergzpes despectus (2), were examined alive at a later date, and showed similar filose phenomena. In an egg not yet divided and having one polar body formed, numerous fine filose threads were seen pro- jecting from the surface into the wide space between egg and membrane. Most of these filaments were confined to one quarter of the periphery, as seen in optical section; but one isolated, blunt, branched process came up some distance from the others near the polar bodies, which were of unequal lengths, a few longer ones reaching out halfway to the membrane. The longer ones often showed short branches at the tip and swellings along their length, suggesting those on the pseudopodia of filose Protozoa. Moreover, a vibrating particle beyond the finest filaments, and scarce seen with ocs. 6 and 8 and obj. 2 mm., moved out and then back again toward the egg, as if it might have been traveling upon some finer filament not seen. The early stages of several marine lamellibranchs were very briefly examined. In Yoldia limatula an egg before dividing was seen to send out innumerable fine filaments from a thin layer of waving ectosarc. As these filaments were crowded together and radiated directly outward, they looked not unlike the cilia the larva developed, but they were much finer. These filaments did not spring from the entire surface of the egg, but from large areas. At one point a comparatively coarse process, suggesting an icicle in high refractive power and shape, pro- jected amidst the finer filaments. As there was no membrane, all these filose processes projected freely into the sea water. Again, ina much more prolonged study of the eggs of the large nemertean worm Cerebratulus lacteus Verrill, certain filose phenomena were seen before and after the first cleavage. The pear-shaped egg removed from the female had a more or less foto) pronounced prolongation at its pointed pole. From this pro- 1 Andrews, E. A., “Spinning in Serpula Eggs,” American Naturalist. Septem- ber, 1897. 4 ANDREWS. [Vot. II. longation fine filaments were seen to radiate in all directions and to rapidly change, and in other cases comparatively blunt filaments occupied the same place. Before this prolongation was drawn into the main mass, many fine filaments appeared from the opposite blunt end of the egg. As it was being drawn in, fine, short filaments were seen projecting from the surface of the egg round about its base. Under oc. 12, obj. 2 mm., these were much finer than the tail of the large sperms now often present within the egg membrane. When the polar bodies were forming, and for a time after their extrusion, the surface of the egg near these bodies (and sometimes quite gen- erally) sent out very fine filaments, set like cilia close together. Later, during the first cleavage, similar filaments arose from the surface of the egg. They were especially well seen when occurring as stout, stiff-looking, radiating lines arising from the tops of certain remarkable papilla that frequently formed on the sides of the gaping cleavage furrow. As these small papillae armed with tufts of filose processes arose at certain phases in cleavage and then vanished, they suggested some such temporary interconnection of cells as occurs in certain echinoderms;! but the filaments could not be followed from one cell to the other and seemed much too short to furnish any intercellular union by filose activity. In this connection it seems significant that the cleavage is of such a nature as to leave it doubtful, from surface views, whether the blastomeres actually separate entirely as in the echinoderms, or not. In an egg with twenty or more cells fine processes were seen projecting from the profile of a cell favorably placed. In another case, where there were but four cells, a stout filament passed across the space between the inner ends of the cells, near the surface, and made the protoplasm of two opposite cells continuous. On this filamentous bridge there were nodu- lar enlargements that gradually grew smaller as the filament dwindled in diameter and was withdrawn into one of the cells. But as the mode of formation of this connecting filament was not observed, and as the egg subsequently showed abnormal 1 Andrews, G. F., “Some Spinning Activities of Protoplasm in Starfish and Echinus Eggs,” Journ. of Morph. Vol. xii. 1897. ; | ‘ . Novia 2YCOSE ACTIVE TN METAZOAN EGGS. 5 features, this was decided not to be a case of filose activity. In the same egg some shorter, slender pseudopodia projected from one cell, but could not be traced more than a tenth of the distance to the opposite cell toward which they extended over the above-mentioned polar space. The most delicate filose displays were seen near the polar bodies during the first and second cleavages. The egg put forth fine protoplasmic threads that branched and reached up toward the second polar body. In this region a sheet of sub- stance connected the egg with the second polar body, and the filose phenomena in it led to the assumption that it was a flow- ing mass of protoplasm, or that it contained more or less of it. But this, with the remarkable filose activities of the polar bodies, has been described and figured elsewhere! In both a gastero- pod and a lamellibranch, the polar bodies were likewise seen to have filose activities.2 Thus in several great groups of animals the polar bodies may act in a filose way for some time after their extrusion, plainly exhibiting contractile phenomena in their cytoplasm and showing themselves to be still alive and active, so that, whatever may be their import as regards the chromatin they carry with them, they appear as more or less isolated parts of the egg mass, carrying on filose changes of the same nature as those of its other parts. Being convinced that filose phenomena essentially similar to those of certain Protozoa exist also in the great metazoan groups Echinodermata, Annelida, Mollusca, and Nemertina, another attempt was made to find them in Chordata by study- ing some preserved Amphioxus eggs. However, the relation between living and preserved material is so remote in cases of such delicate phenomena as these sought, that little weight could be laid upon the results without a knowledge of the live eggs. Lacking this, it was thought that an acquaintance with the live and the preserved eggs of echinoderms would ‘suffice to enable one to draw tentative conclusions from the preserved 1 Andrews, E. A., “Activities of the Polar Bodies of Cerebratulus,” Archiv. f. Entwicklungsmechanik. Bd. vi. 1898. 2 Andrews, E. A., “Some Activities of Polar Bodies,” Johns Hopkins Univer- sity Circulars. Vol. xvii, No. 132. November, 1897. 6 ANDREWS. [Voroti: Amphioxus eggs. The following three figures illustrate some of the appearances seen in live and in preserved echinoderm eggs, and may aid in justifying the conclusions formed as to the nature of the intercellular connections found in Amphioxus. Fic. 1. Fig. 1 represents a small part of a blastula of the starfish common at Roscoff, France, and is reduced one-half from a camera sketch made by G. F. Andrews in 1894 with oc. 8, obj. 2mm., tube 170 mm. A small opening into the blastula is seen, surrounded by cells that are actively spinning out fine filaments by means of which they are variously connected with one another and with the two polar bodies. These lie, in this case, in the opening or cleavage pore, and are temporarily con- nected to adjacent cells while giving off very remarkable den- dritic pseudopodia, along which protoplasm flows and collects in lumps. The chromosomes are not shown in the polar bodies, Novi.) PIZOSE ACTIVE SN METAZOAN EGGS, i nor in the enlargements on their pseudopodia, into which they were sometimes carried; vacuoles, however, are indicated in the main part of each polar body. The short, apparently blunt filaments projecting from the edge of the cell uppermost in the Fic. 2. figure really represent threads that bent abruptly and extended far inward, and were, in most cases, attached to distant cells of the blastula. This opening into the blastula closes in with change in posi- tion of the adjacent cells, and with changes in the character of the filose activities that lead one to conclude they are instrumental in associating the cells more intimately. Thus in Fig. 2 a later stage shows the cleavage pore reduced to a few small chinks between the cells that have glided over it. The polar bodies were here taken inside, and are represented, lying one over the other, by the black mass to the left. 8 ANDREWS. iVorwil In the lower part of the figure a cell to the left has reached out over one to the right and established a connection with it by means of a broad strand of filaments. An earlier stage of a similar process is shown above, where the same left cell is strongly bound to an upper cell by double strands of filaments. These filaments seem instrumental in drawing the cells together to cover in the cleavage pore. This figure is reduced one-half FiG.23. and otherwise like the first in execution, except that it was drawn with oc. 6. In both these figures the thickness of the finer filaments is exaggerated in drawing, and hence they do not adequately indi- cate the delicacy of these processes. Moreover, as they are constantly changing, and as they contract and draw in when stimulated by certain chemicals or even by mechanical insult to the egg, it is plain that usual methods of preservation will fix but part of these displays, at the most, and that they may be readily broken off in subsequent treatment. However, it was found possible to preserve some of the larger filaments or amalgamations of filaments by special methods, and intercellular ott SS Ss ee a Nom.) - 2iEOSE ACTIVITY IN METAZOAN EGGS. 9 connections and other spinnings have been retained, both in starfish and in sea-urchin eggs for three years. Thus in Fig. 3 a four-cell stage in the common echinus of Roscoff shows filaments passing from cell to cell. These are drawn considerably thicker than the actual threads seen, but otherwise represent them truly. The figure is a surface view camera drawing with oc. 8, obj. 2 mm., tube 160 mm., helped ‘A oR out with ocs. 12 and 18. On the left, above, a series of eleva- tions from one cell seemed to be the remnants of tufts of fila- ments, while the granular matter, imperfectly represented here, partly covering the cleavage pore, appeared to be the same as Hammar’s ectoplasmic layer.! There can be no doubt that these filamentous intercellular connections are the preserved remnants of the active filose threads of the living egg. Some eggs of Amphioxus killed in corrosive acetic and stained in Orth’s new lithium carmine, as well as some killed 1 Andrews, E. A., “ Hammar’s Ectoplasmic Layer,” American Naturalist. December, 1897. 10 ANDREWS. [Vora Tr in Flemming’s fluid and not stained, were kindly placed at my disposal by Professor T. H. Morgan. Both those mounted in balsam and those studied in alcohol showed undoubted filaments connecting the blastomeres in various cleavage stages. From the resemblance of these filaments to those found in preserved echinoderm material there seemed little doubt that this was probably another case of filose intercellular connections, but there must remain some doubt till the live egg is studied. In living material we may expect to find filose displays as remark- able as those in the echinoderms, and, in part at least, more readily observed. In four, eight, and sixteen-cell stages, many eggs showed such marked intercellular connections as the one represented in Fig. 4. These filaments are of clear material that arises from the clear ectosarc of one cell and becomes continuous with that of another cell. Only a few cases of branching were seen, apparently only the main trunks and grosser threads being preserved. Fig. 4 shows an eight-cell stage in Amphioxus, with a definite abruptly curved filament passing from one cell to an opposite one at the bottom of an open cleavage cavity. This is from a camera sketch with oc. 2 and obj. D, not reduced. One cell showed a marked elongation toward another, but no connecting filaments were seen, though they may well have existed there in the live state. Besides the filaments seen connecting cells, as in the above figure, there were other signs of filose activity in these eggs; groups of minute spherules and filaments pro- truded from the ectosarc as if remnants of filose processes. There were also large ectosarcal outflows near cleavage planes, suggesting the amceboid elevations described for the eggs of certain nematodes by Erlanger.! When more highly magnified these intercellular filaments in Amphioxus appear as in Fig. 5, which represents part of another eight-cell stage, drawn with camera, oc. 8 and obj. qs im. Here a filament arose from the ectosarc of one cell, and, after mak- ing a complex bend, difficult to understand, gradually dwindled in diameter till it became continuous with the surface of another 1 Biol. Centrbit. Bd. xvii. 1897 NOs) PRAOSE ACTIVA IN METAZOAN EGGS. II cell on the opposite side of a narrow cleavage space at the center of the rather compact group of eight cells. Such intercellular connections as this would seem to be, in all probability, of the same nature as those seen in live and in preserved echinoderm eggs. And if such be the case, it is especially interesting to find them in Amphioxus, not only as this is in many respects so diagrammatic a representative of the Chordata as to lead one to infer its filose phenomena will be found, in modified form, in various vertebrates, but because the egg of Amphioxus has been so carefully studied by experi- mental methods that the need of an organic intercellular con- nection to explain known facts in embryology is here especially felt. Thus Prof. E. B. Wilson was led to conclude in his study of Amphioxus,! “that the unity of the normal embryo ts not caused by a mere Juxtaposition of the cells... . This unity ts not mechanical, but physiological... . There must be a struc- 1« Amphioxus and the Mosaic Theory of Development,” Journ. of Morph. Vol. viii. 1893. Hey ANDREWS. Verh tural continuity from cell to cell that ts the medium of codrdina- tion, and that ts broken by mechanical displacements of the blastomeres.” The nature of this structural continuity was not surmised, but it now seems evident that it is really, in part at least, brought about by such remarkable, changing, pseudopodial threads as were seen in the echinoderms and preserved, in part, in the Amphioxus eggs I had for examination. That filose phenomena will be found in the blastula and gastrula stages seems most probable, but as yet I have not been able to find remnants of them in the above-preserved material. The figures published by Klaatsch?! in illustration of other problems suggest, at first sight, profuse intercellular connections in these stages of Amphioxus, but it seems more probable that the lines there shown are the results of shrinkage and imperfect preservation, though some filose activity may have furnished an element for certain of the distortions that resulted. As filose phenomena in eggs as far as yet studied are, to say the least, easily overlooked (in the echinoderms, they can be seen only with difficulty, though the main threads from the polar bodies of Cerebratulus and the above connections in Amphioxus are so distinct as to be readily evident with low powers to one searching for them), we need not expect to find them frequently mentioned in the past literature of embryology. Yet some of them must have been seen, even if passed by as of little moment. Thus Professor Conn described and figured? fine lines passing out from the surface of the egg of the gephyr- ean worm TZhalassema mellita Conn to the rather distant membrane. These he regarded as striz and interpreted as indicative of the presence of a jelly-like substance between egg and membrane. Yet an examination of his Fig. 13, Pl. XX, suggests that this egg and its polar bodies will prove to possess filose phenomena.? 1“ Bemerkungen iiber die Gastrula des Amphioxus,” A/orph. Jahrb. Bd. xxv, L8O7.00 Wels Xai: 2 Studies from the Biological Laboratory, Johns Hopkins University. Vol. ii. 1884. 3 In correcting proof I add that the connection made by Professor Flemming (Merkel and Bonnet, Ergebnisse, 1897, p. 279) between certain fine pseudopodia- Now|) 2UZ0SE ACTIMALTY IN METAZOAN EGGS. 12 Conclusions. Filose activities like those of the finer pseudopodial threads of certain Protozoa were seen in the living eggs of Metazoa, in Echinodermata, Annelida, Mollusca, and Nemertina. Study of preserved material makes most probable their existence in Amphioxus and quite probable their existence in Amphibia. Members of other great groups have not as yet been examined from this point of view.! Such filose filaments connect the cells in the eggs and larve of Echinodermata; filaments that are most probably of this nature connect the blastomeres of Amphioxus; filaments prob- ably filose connect the cells in eggs and larvae of Amphibia. Wherever found, filose connecting filaments may be assumed to have the importance ascribed to them on their first discovery in the echinoderms, and to furnish a medium for codrdinating the activities of parts of the embryo. JouHns HOPKINS UNIVERSITY, BALTIMORE, January 29, 1808. like lines seen by him on a polar body of Anodonta in 1873 (Archiv. f. mikr. Anat., Bd. x) and the filose phenomena of Echinodermata as described by G. F. Andrews seems to me probably correct. 1 However, at this later date, March 7, I am able to state that the cleaving eggs of a green Hydra have remarkable ectosarcal displays and some interconnec- tion of cells by filose activities; this in addition to the gross pseudopodia described by Kleinenberg. oo FES J ee ACTIVITIES OF MESENCHYME IN CERTAIN LARVAE. CHARLES B. WILSON. In consideration of the attention which has recently been given to a study of the phenomena of the living cell, the fol- lowing observations may be found of some interest. They were made during the summers of 1896 and 1897 upon mesen- chyme cells in the larvae of certain species of molluscs and nemerteans. Freshly laid eggs of the nudibranch mollusc Tergipes despec- tus (?) were obtained and reared to the veliger stage. Both egg and young embryo are nearly opaque, the gastrula invagi- nation being distinguished only by a slight increase in the density. But after the mantle has been formed and the shell has been secreted, the veliger becomes perfectly transparent. The delicate shell is as clear as the finest glass, and the whole internal structure is now plainly visible. Mesenchyme cells appear very early in development. They arise from large primitive endoderm cells at the posterior edge of the blastopore in the ordinary way for gastropod molluscs, and are set free in the segmentation cavity. At first they are approximately spherical in shape, and float about freely in the liquid which fills the cavity, but they soon begin to elongate, and become spindle-shaped. At this stage in their develop- ment activities were observed which are probably analogous to those described for the polar bodies of certain animals.1 Unfortunately, I did not have with me at the time an objec- tive of sufficient power to bring out the finer protoplasmic movements distinctly. Well-defined amoeboid changes of out- line, however, could be plainly seen. The cell, as it moves about in the liquid which fills the segmentation cavity, puts out blunt pseudopodia-like processes. These change both their 1 Andrews, E. A., “ Some Activities of Polar Bodies,” Johns Hopkins University Circular. November, 1897. 16 WILSON. [LV Oral shape and position. There seems to be a special tendency to their formation whenever the cell comes close to the wall of the cavity, or when two cells approach each other. The processes remain short and blunt, and in no instance were they seen to reach either the wall or a neighboring cell. The putting out of these processes was seen to be accompanied in several instances by corresponding movements of the cell contents very similar to those in an amoeba. Certain disturbances were also noticed in the liquid’ close to the surface of the cell. At the time these were considered analogous to slow ciliary motion, and they were probably caused by “filose”’ action. It was impossible to determine with certainty whether the amoeboid changes ceased and were followed by a period of rest before the cell became per- manently branched; but this would seem probable from analogy. At all events, permanent proc- esses soon appear, the cells be- come fastened in place, and after subsequent development function as muscles. During the development, after the cells become branched, certain activities appear which are of an entirely different nature. A“single cell; or more oltensa pair of these mesenchyme cells, ig as ee ets ee Aa ay be found in close proximity gipes, showing position of the two con- to the internal wall of the mantle Feat eens ak “2 on one (usually the right) side nearly in the icenter (Miss) a); These two cells are in close proximity to each other, and each consists of a body and radiating branches. The body is composed of finely granular protoplasm, and contains a distinct nucleus and several vacuoles. From it branches extend in many different directions. There are usually several fine branches reaching directly across from the body of one cell to No. 1.] MESENCHYME IN CERTAIN LARVAE. i that of the other, which serve to bind the cells firmly together. The other branches are at first of about the same size, but one branch from each cell, extending outward (.c., away from the twin cell), becomes very quickly larger than the rest, sometimes approaching the diameter of the cell itself. These two larger branches are seldom in the same straight line, but usually make a more or less obtuse angle with each other. That they are really branches and not simply a pro- longation of the cell body is apparent from the fact that they are finely fibrous in struc- ture like all the other branches, and not ees eranular like the body. ‘Fic. 2.—The two contractile cells enlarged to show their of the cell. ee ern ee These large branches extend some distance from the cell and then divide dichoto- mously, ending in fine fibrils. The smaller divisions of these branches, together with those of the cell itself, anastomose freely and form a loose reticular network. These details of structure appear to better advantage in the enlarged drawing of the two cells shown in Piece: At this stage of their development they are not attached to anything except to each other, but the network formed by their interlacing branches extends over a considerable area and holds them in position. They appear like ordinary mesenchyme cells, but upon being watched they are seen to possess a peculiar contractile power, which is manifested at intervals. In a few individuals the con- tractions occurred at definite intervals as long as the cell was watched, but more frequently there was a period of rest after a few contractions. Both the time of contracting and the intervals of rest were subject to considerable variation, but the latter was never long enough to enable a camera lucida sketch 18 WILSON. Worle to be made. The drawings in Figs. 1 and 2 were taken from individuals which had been paralyzed with magnesium sulphate. These cells are isolated from everything except the liquid in which they lie, and, consequently, if there be any stimulus previous to contraction, it must be given through the medium of the liquid or it must arise spontaneously in the cell itself. By careful watching, the contraction can be seen to begin in the cell body and travel outward along the branches, though the contractile wave moves so quickly that it practically begins at all points simultaneously. As a result of its action the protoplasm draws together, the cell body becomes more spherical, all the branches, large and small alike, become shorter and thicker, and the whole meshwork of fibrils is drawn in until it occupies much less area than formerly. When it first begins, the contraction is comparatively weak and results simply in a shortening of the branches and fibrils, but as it proceeds it becomes rapidly stronger and stronger. This increase in contraction cannot manifest itself in any further shortening of the branches, for they have already shortened all they are capable of doing. The only way in which the two ends of any branch can now be brought nearer together is by a bending or folding of the fibers upon themselves, and this is what actuallyaroccurs.” “At ‘the iclose on contraction (Fig. 3) the smaller Fe eae em oecis athe cess Chae Ghes- and fibrils have been drawn of contraction. Leitz objective No. 7, in so much that they are twisted into S aaa a corkscrew shape for their entire length. The large branches have also contracted so strongly that their surface becomes wavy or sinuous in outline. The bodies of the cells remain spherical, but become so opaque that neither nucleus nor vacuoles are visible. This twisting or plication of the muscle fibers, whereby their retractive power is increased, is also shown in the retractor muscle of the velum (Fig. 1), and will be noticed later in certain muscles of the nemertean larva. The same thing has a INO. 12)] MESENCHYME IN CERTAIN LARVAE. 19 been observed in “single fibrils in protoplasm, as well as con- tractile pellicles and substance membranes,” and also in Meta- zoan cilia and the muscle bands in rotifers.! It seems to be carried farther here than in the muscles mentioned, for the simple reason that these filaments are unattached and, there- fore, there is nothing to restrain it. After remaining an instant in this extreme contraction, the cells relax, and the return to a normal condition is practically instantaneous. We have here, then, the same power manifested by the single mesenchyme cell, with its branches, that belongs to the more complicated retractor muscle, and that, too, when it is isolated from everything save the liquid in which it floats. This must certainly be a very near approach to a primitive muscular contractility. The contraction lasts one and a half or two seconds, the relaxation occupies but a very small fraction of a second, while the pause or rest varies from two or three to twelve or fifteen seconds. This suggests very forcibly a condition similar to that which obtains in the beating of the heart, with the ~ exception that in these mesenchyme cells the relative duration of contraction and relaxation is reversed, the former being much longer. This same contractile power is also possessed, to a less degree, by the other mesenchyme cells. They may often be seen to contract after they have become branched. The con- tractions are not as rhythmical as those just described, but they-are as automatic. Some of these mesenchyme cells enter the velum during development and become attached to its walls until the whole interior is traversed from wall to wall by their branches, rendering it highly contractile.” In this case, therefore, the same cell which contracts at first automatically may afterward become a part of the muscular network of the velum, where it is under the control of the central nervous system. Similar phenomena were observed in the mesenchyme cells of the pilidium larvae of the Nemertean Cerebratulus lacteus 1 Andrews, G. F., The Living Substance as Such: and as Organism, p. 103. g g p- 103 1897. 2 Lang, Comparative Anatomy, Part II, p. 257. 20 WILSON. [Vor TT: Verrill. These larvae were reared from artificially fertilized eggs, and a full account of their development is in preparation for a subsequent paper. The eggs of this nemertean are opaque during cleavage and gastrulation, but become beautifully transparent on reaching the pilidium stage. The mesenchyme first appears as isolated cells derived from the ectoderm, as observed by Metschnikoff (Zezt. f. wiss. Zool., Bd. xxxix). They move about freely in the gelatinous liquid which fills the space between ectoderm and entoderm. At first they are nearly spherical in outline, but they soon begin to develop processes and become branched, in which condition they are very readily distinguished from the other elements. So long as they remain free floating there is no indication of cell fibers, but simply a nucleus enclosed in cytoplasm. But as they begin to branch they grow larger, and granules appear in the cyto- plasm, while the branches become gradually fibrous in structure. No amitotic division stages, however, were noticed in any of these cells, such as were found by Montgomery in the free- floating mesenchyme cells of the adult worm (Zool. /ahro., Bd. x). The branches hinder the freedom of motion of the cells, and the latter gradually become fixed in position. The fibrils at the ‘extremities Yof the branches “aremthem fastened in place, and from being mere wandering mesenchyme the cell becomes one of the muscles of the pilidium. This transformation was watched several times in the formation of different muscles, and nearly all the intermediate stages were observed. The most important muscle of the pilidium is the one which extends from the apical plate downward to the anterior border of the lappets. The development of this muscle was watched in many different individuals. When the mesen- chyme cells first develop branches, one of them can be seen to become stationary in about the position of the future apical muscle. One of its processes becomes fastened to the apical plate, while another fastens to the wall of the digestive tract, and sometimes a third connects with the aboral wall of the pilidium (Fig. 4). The number of processes is not constant, but the position assumed by the cell is approximately so. No: 1] MESENCHYME IN CERTAIN LARVAE. 21 Other cells become fastened to the walls of the digestive tract along its anterior border. The branches of these cells anastomose with each other and with the first cell, and from them are developed the strong muscle which enables the larva to retract the apical plate with its tuft of cilia. This muscle becomes attached at first to the wall of the digestive tract, as figured by Verrill (Marine Nemerteans of New England, p. 417). But as soon as the branches begin to anastomose it ; RASS BRN Fic. 4. — Side view of pilidium larva, showing mesenchyme cells in position to form the apical muscle. Zeiss cam. luc. x 575 diams. develops along the line of mesenchyme cells seen in Fig. 4, and becomes fastened to the anterior border of the lappets. In a similar way a transverse muscle is formed just in front of the apical plate. This consists of a single large mesenchyme cell which subse- quently develops long processes reaching from one side of the pilidium to the other. In later development the whole internal surface of the umbrella is covered with a loose meshwork of anastomosed mesenchyme cells, which give the larva so much 22 WILSON. [Vor TP contractile power that it frequently tears the umbrella cells apart by violent contractions when irritated. While floating about freely these mesenchyme cells do not contract, so far as could be observed, but as soon as they begin to form processes they can be seen to contract. Occasionally two cells anastomose with each other before becoming attached to the wall of the pilidium. In such a case they contract irregularly at first, the intervals between contractions being unequal, but later the contractions become rhythmical and very closely resemble those of the opisthobranch gastropods just described. Three or four con- tractions occur in rapid succession and are followed by a comparatively long rest. After the cell branches become fastened to the pilidium wall these rhythmical pulsations cease. The mesenchyme cells now become regular muscles of the larva and contract only when stimulated from the central nervous system. We are witnessing here, then, the passage from an automatic condition, in which the cells contract quite independently from the rest of the larva, into a condition in which every contraction is definitely correlated with that of the other larval muscles. In this nemertean larva the intense contractions resulting in a corkscrew shortening of the branches and fibers occurred subsequent to the fixation of the cells. It was not noticed in any of the free cells even when two of them anastomosed before becoming attached, and all the conditions appeared as favorable as in the veliger larvae. After attachment such a shortening is very noticeable, especially in the apical muscle and the fine radial muscles of the side lappets. In an examination of these larvae, therefore, it is found that: 1. The mesenchyme cells are at first nearly spherical and are free-floating. In this condition they consist simply of a nucleus and cytoplasm; they may put out amoeboid processes, but they do not show any contractile movements. 2. They soon grow larger, become granular, and develop fibrous branches which hinder their free motion, and finally they become fixed in position and function as muscles. INOF 1.) MESENCHYME IN CERTAIN LARVAE. 23 3. In both larvae prior to such fixation cells may be found, singly or in pairs, which pulsate in more or less rhythmical contractions until their branches become fastened to the larval tissue, when the pulsations cease. 4. Both larvae, accordingly, show a well-marked transition from automatic pulsations to muscular contractions dependent upon the central nervous system. STATE NORMAL SCHOOL, WESTFIELD, MASss., April 18, 1808. OLSERVAIONS VON THE GENUS OF FOSSIE PIStomC Aan) BY PROFESSOR “COPE, BORTHBUS, BY DR. _LEIDY, <1 PHAGCTINUS: Os Pee EL ACE Tue earliest reference which we have to any remains of the genus of fishes usually called Portheus is that found in Man- tell’s Geology of Sussex, p. 241, Pl. XLII, 1822. No systematic name is there assigned to this fish. Later, Louis Agassiz, in his Pozssons Fossiles, vol. v, p. 99, referred to Mantell’s descrip- tion, and refigured the materials (of. cz¢., Pl. XXV 4, Figs. 1 a, 16), presenting at the same time additional figures of remains from the same locality (Pl. XXV a, Fig. 3; Pl. XXV 4, Figs. 2, 3). All these he included, with other remains, under the name Hypsodon lewestensis. In 1871, in Proc. Amer. Philos. Soc., vol. xii, p. 175, Pro- fessor Cope established the genus Portheus, founding it on materials collected in the cretaceous deposits of Western Kansas. The type of the genus was called Portheus molossus. Later, in the same volume, p. 330, Cope recognized the affinity of the remains figured by Agassiz, as above cited, to those of Portheus, as well as the fact that other remains had been included by Agassiz under the term Hypsodon which were not congeneric with Portheus. Professor Cope, therefore, restricted Hypsodon to those bones and teeth which differed generically from his own American materials, and included the remainder under Portheus. In this same paper, pp. 333, 335, Cope also referred to Portheus a species which he had described in 1870 (Proc. Amer. Philos. Soc., vol. xi, p. 533) under the name of Saurocephalus thaumas. Both these species and others were fully described in his Cretaceous Vertebrates, published mS 75. At this point it may be noted that in the year 1870 (Proc. Acad. Nat. Sci. Phila., p. 12) Dr. Joseph Leidy described from the Cretaceous of Kansas the spine of a fish which he called 26 | HAY. [Vor. II. Aiphactinus audax, and which, without doubt, belongs to the Saurocephalidae. A more complete description and figures of this fossil spine were given by Dr. Leidy in his Coxtrzbutions to the Extinct Vertebrate Fauna of the Western Territories, p: 200, Pl. XVI, Figs: 0, 10. . This was published in 1373: Professor Cope first recognized the affinities of this spine in a paper in Hayden’s Second Annual Report of the Geological and Geographical Survey of the Territories, 1871, p. 418, where he assigned it to the genus Saurocephalus, in which genus he also arranged the species which he later called Portheus thau- mas. He compares the spine with one obtained from S. prog- nathus, a fish which he later relegated to the genus Ichthyo- dectes, itself a close relative of Portheus. From about this period up to 1874 Professor Cope held the opinion that certain fin remains belonged to Portheus, and probably to the pectoral fin, which it is now pretty certain belong to Protosphyraena. Other spine-like fin rays, whose resemblance to Leidy’s Xiphac- tinus he admitted, he regarded as also belonging to Portheus, and probably to the ventral fins. He claimed, however, that Xiphactinus was distinct from both Portheus and Ichthyo-— dectes ; but he does not specify the points of difference. By the time of the publication of his Cretaceous Vertebrates in 1875, he had become convinced that the fin structures which are now assigned to Protosphyraena did not belong to Portheus ; and to them he gave the name Pelecopterus. He had also learned that the ventral spine-like fin rays of his Portheus did not differ greatly from those of the pectoral fin (p. 204). Of Xiphactinus he says: “ Dr. Leidy applied the name Xiphacti- nus to a genus indicated by a spine in some degree like those regarded above as ventrals of Saurodontidae. Whether it belongs to any of the genera above enumerated, or, if so, which of them, is a question which can only be settled by future investigation ”’ (of. czt., p. 190). Accompanying a considerable collection of specimens of Por- theus collected for me in Western Kansas, in the region of Butte Creek, are many large spines, some nearly complete, others in fragments. Some of these belong to the shoulder girdle which I have figured (Fig. 9), and this, I have no doubt, belongs to No. I.] THE GENUS OF FOSSIL FISHES; 27 Cope’s genus Portheus. No more doubt exists in my mind regarding the generic identity of many of the other spines. Some of these, indeed, were found in a block of soft limestone, and were in close relation to jaws, vertebrae, etc., of P. thau- mas. These spines I have compared with Leidy’s type of Xiphactinus audax, and I find no difference that can be regarded as generic. Both Cope, in his Cretaceous Vertebrates, and Crook, in Palaeontographica, vol. xxxix, p. 119, have described and figured spines of Portheus which differ in no essential respect from Xiphactinus. The genus Ichthyodectes possessed pectoral spines not greatly different in structure from those of Portheus ; but none of them attain the size of those assigned to Xiphactinus and Portheus. Taking all the facts into con- sideration, it seems to me that there can be no reasonable doubt that Xiphactinus is the same as Portheus, and ought to ‘supersede it as a name for this genus of fishes. It is quite probable that X. audax is the same as some one of Professor Cope’s species of Portheus ; but it will require a careful study of well-identified spines of all the species, and a comparison of them with Dr. Leidy’s type specimen to decide the question. For the present, then, we must recognize six American species of Xiphactinus ; vzz., XY. audax (Leidy), X. molossus (Cope), X. thaumas (Cope), X. lestrio (Cope), X. mudget (Cope), and X. lowit (Stewart). In my study of the genus Xiphactinus I have been greatly aided by comparison of its various parts with those of the tar- pon of our southern coast (Zazpon atlanticus). While the tarpon is in many respects quite unlike Xiphactinus, in others it strikingly resembles the latter. Although the two genera undoubtedly belong to different families, these families are closely related, and both belong to the order of Isospondyli. It was in this order that Professor Cope arranged Portheus and its related genera, but he believed that in them he found also characters which indicated relationship with the Siluroids. Such characters I am unable to perceive. Xiphactinus was an Isospondylid, generalized in some respects, but greatly special- ized in others. This specialization shows itself especially in the teeth and paired fins. 28 HAY. [Vou. II. The head of this genus has been described by Cope (Cvez. Vert., pp. 183, 191), by Newton (Quar. Journ. Geol. Soc., vol. XXxill, p. 505), and by Crook (Palacontographica, vol. Xxxix, p. 114). Each of these authors also presents figures of various parts. In the following pages I shall call attention to such features of structure as, in my judgment, are new to science, or which require additional treatment or correction. I regard the identification of the parietals as yet uncertain. Professor Cope was himself in doubt regarding them, and thought that perhaps what he called the supraoccipital might really be the coalesced parietals (Cvet. Vert., p. 183). Further Fic. 1.—Skull of Tarfon atlanticus, seen from right side and partly from below. x 3. on (p. 188) he concluded that the bones which he at first had identified as the epiotics were the parietals. Crook states that the parietals are completely separated by the large supra- occipital. He figures them (Pl. XVIII) as lying laterad of the epiotics, a situation which appears not probable. The small development of the supraoccipital in the tarpon permits the parietals to meet along their whole median borders, while each epiotic (Fig. 1, ef.), by its inner anterior angle, comes into contact with the outer posterior angle of the parietal. Should the supraoccipital now be enlarged we might expect the parietals to be reduced posteriorly and more or less separated. It seems to me that in the four rather complete skulls of Xiphactinus before me, two belonging to the United States National Museum, the others my own, I can recognize the NOA Kei] 7 GENES, OF FOSSIL. FISHES. 29 parietals as wedge-shaped narrow bones which lie between the anterior ends of the pterotics and posterior ends of the frontals on the outside, and the supraoccipital on the median side, I am inclined to believe that the parietals meet along the mid- line in front of the supraoccipital, and really include the elevated surface assigned by Crook to the latter bone, and said by him to be covered with coarse granulations. The posterior pointed end of each bone falls just mesiad of the epiotic. My determination of these bones may be erroneous, but I am wholly unable to find evidences of any suture defining the parietals as located by Dr. Crook. The epiotics have been correctly mapped by Dr. Crook. Professor Cope was in doubt about the opisthotics. At first (Cret. Vert., p. 183) he regarded them as forming the postero- lateral angles of the skull; but, on p. 188, he concludes that these angles are formed by the epiotics, and that the opisthotics are absent. Crook (of. c7¢., p. 115) says that the opisthotics are the largest bones entering into the brain capsule. This I believe to be an error. I am of the opinion that the position and relations of the opisthotics of Xiphactinus are best explained by an examination of the tarpon (Fig. 1, of.). Here what may be regarded as the body of the opisthotic is rather small. Its upper end articulates with the pterotic (fz.0.), while the greater portion of the body lies against the exoccipital (e.0.). It bends forward, sending a small process to the prodtic (f72.). From the lower border of the body of the bone there is sent downward and forward to the basioccipital a broad process which is as large as the remainder of the bone. In passing to the lower portion of the basioccipital, this process forms a bridge across a deep and broad fossa which is excavated in the basioccipital, but the roof of which is formed by the exoccipital. Now, the positions and forms of all the other bones in this region are in Xiphactinus almost identical with those of Tar- pon. There are also the same deep cavities in the side walls of the skulls of the two genera. I believe, therefore, that we are justified in concluding that the opisthotic had somewhat similar form, position, and relationships. Moreover, I am con- vinced that this bone is present in three of the skulls at my 30 HA Ve [VorvIT. command ; although, on account of the distortion to which the skulls have been subjected, the determination is not as satisfac- tory as is desirable. Its lower process appears to have been much slenderer than in Tarpon. In Tarpon the lower process of the post-temporal is attached by a strong ligament to the posterior extremity of the opisthotic ; and, if I am correct in my determination of both these bones in Xiphactinus, they were brought into close connection. The pterotics (squamosals of most authors) were propor- tionally more extensive bones in Xiphactinus than in Tarpon, and formed a more prominent process at the outer and hinder portion of the skull. Each included, I am satisfied, the area marked by Crook as belonging to the parietal. The pterotics furnished the larger part of the articular surface for the head of the hyomandibular. This surface was essentially as it is in Tarpon (Fig. 1, 4m.). As regards the prootics, Professor Cope’s description (C7ez. ‘ert., p. 185) is not far out of the way, though brief. Dr. Crook is less fortunate when he states that the prootics are small. His error arose, if we may judge from his figure of. Ichthyodectes polymicrodus, from his having carved the opisthotic out of the territory of the prootic. The prootics are really the largest of the otic bones. Professor Cope says that with the pterotic and opisthotic this bone bounded a large foramen. This so-called foramen is not really such, but a deep excavation, or fossa, in the side of the skull. In Tarpon this fossa is an inch deep, and about as much in diameter; and it was quite as large in Xiphactinus. In the latter genus the anterior wall appears not to have been completely ossified, so that, in the skeleton, the fossa probably opened widely into the large cavity which lay above the brain, and which will be described further on. Since the cavity just referred to was in life probably filled with the primitive cartilage, the apparent opening from the fossa into it was merely an unossified part of the prootic. In Tarpon the mouth of this fossa is somewhat triangular. Its floor is furnished by the exoccipital and the prodtic, its posterior wall by the exoccipital and the pterotic, its roof by the pterotic, and the anterior wall by the pterotic and the No: i-] HE GHVMES OF FOSSIL FISHES, gyi prootic. The sutures between the adjoining edges of each two of these bones meet in the apex of the fossa. The axis of the fossa is directed inward and upward. Without doubt, the fossa in the side wall of the skull of Xiphactinus was essentially the same as that in Tarpon. In Tarpon there is, as has already been mentioned, an exten- sive fossa on each side of the skull, excavated principally in the basioccipital. This is so deep that only a thin wall of bone separates that on the right side from that on the left. Each fossa is continued forward on the outer surface of the prodtic, becoming narrower and shallower. It is across this fossa that the broad process of the opisthotic is thrown as'a bridge. A somewhat similar fossa existed in Xiphactinus, but on account of the compression suffered by the skulls its features cannot be definitely determined. The prootic of Xiphactinus, like that of Tarpon (Fig. 1, pvo.), provides a portion of the articular surface for the head of the hyomandibular. In Tarpon there are on the external surface of the prootic some four or five foramina. In Xiphactinus I have been able to detect only one of these, that for probably a branch of the facial nerve. It lies just below the anterior end of the articulation of the hyomandibular, and corresponds to that marked 7’ in the figure of Tarpon. In Tarpon this fora- men opens into a canal which runs backward in the prootic and emerges at the hinder border of the mouth of the fossa, above described as being walled in by the prootic, opisthotic, and exoccipital. This canal is then continued backward on the outer surface of the exoccipital beneath the opisthotic. It — or, at least, its hinder portion — serves to. conduct the glosso- pharyngeal nerve. An opening has been found in Xiphactinus in the mouth of the fossa, and doubtless the canal was similarly prolonged both forward and backward. Crook’s statement that the parasphenoid is triangular in section, with the base of the triangle directed upward, is true only when the skull is held in an inverted position.. The error is doubtless due to a slip of the pen: It is also erroneous to say that the finger-shaped processes outstanding from each side of the parasphenoid arise at the union of the parasphenoid and Be FLAY [Vot. IT. basioccipital. They arise about opposite the union of the basi- sphenoid and the parasphenoid. These strong lateral processes are almost wholly absent in Tarpon. In both this genus (Fig. 1, pa.s.) and Xiphactinus there is, on each side, a strong process arising from the parasphenoid to meet the prootic. These processes form the side walls of the muscular canal. This canal was of greater extent perpendicularly in Xiphactinus than in Tarpon. The basisphenoid is a Y-shaped bone, the upper end of which articulates with the prootics, while the lower end rests on the parasphenoid. It is almost twice as long as the corresponding bone in a tarpon of the same size (Fig. I, 0.5.). So far as can be determined from the crushed skulls of Xiphactinus, the form and relationships of the alisphenoids and the orbitosphenoids were very much the same as in Tarpon. In this latter fish both of these pairs of bones are large (Fig. I, als., 0.s.). The alisphenoids meet in the mid-line, below the brain, and thus continue forward the floor of the brain-case. In front of these are the large orbitosphenoids, ankylosed in the mid-line, as in the salmon. There is no distinct presphenoid. In the tarpon the brain-case is roofed over behind by the supraoccipital. In front of this the protic sends upward and inward a plate of bone which meets a similar plate from the opposite prodtic. This roof is continued forward by plates of bone which grow mesiad from both the alisphenoids and the orbitosphenoids. These two pairs of bones also send out great lateral plates, which abut against the postfrontal and the lower surface of the frontal. In the mid-line above the brain, the united orbitosphenoids send upward a more or less interrupted crest of bone. Between the brain-case, as thus roofed over, and the parietals, pterotic, and frontals, there is a great space an inch high and extending from one side of the skull to the other, and in life this is probably filled with the primitive cartilage. The arrangement of this portion of the head may be understood by an examination of Parker’s figures of the salmon (7vamns. Phil. Soc. London, vol. clxiii, pp. 95-145, Pls. I-VIII). In Tarpon there are two foramina in the prootic which open from the outside into the cavity here described. One of these is No. 1.] THE GENUS OF FOSSIL FISHES. 33 found in the lower anterior angle of the great lateral fossa; the other is seen just above the foramen 7’. These foramina are probably closed with membrane in life. They are not found in Xiphactinus. In Xiphactinus the alisphenoids and the orbitosphenoids - appear to have had the same extent and relations, at least as seen from below, and I have no doubt that there was in the skull the same large amount of primitive cartilage that we find in Tarpon to-day. The frontals of Xiphactinus were much broader than they are in Tarpon. In a tarpon whose skull had to one of Xiphactinus the ratio in length of 9.5 to 10.5, the width of the frontals bore the ratio of 1 to 2. Since the breadth of the nasal region of Xiphactinus was little less, we may appreciate Pro- fessor Cope’s characterization of their expression as being bulldog-like. To a broad flat surface of the very stout prefrontal of Xiphactinus was applied the superior articulating surface of the malleolar body of the palatine. In Tarpon the palatine is simi- larly connected with the prefrontal, except that the ethmoid bone sends outward a process which takes part in the articula- tion. Professor Cope states that in the alewife the articulation of the palatine is wholly with the ethmoid. The lower surface of the ethmoid furnishes an articular surface for the anterior condyle of the maxillary. Since this condyle in X. ¢hawmas is much larger than that of X. molossus, we ought to find a corresponding difference in the ethmoids of the two species. There can be no doubt that the orbit of Xiphactinus was surrounded by a ring of orbital bones, just as it is in Tarpon. In a skull of X. molossus before me (No. 1646, U. S. N. M.), the superorbitals are wanting, but the border of the frontals shows distinctly that a row of thick bones has been articulated with it. In Tarpon there are three of these superorbitals. Crook has figured a preorbital in Xiphactinus. In Tarpon the posterior suborbitals are very large, extending backward over the cheek as far as the preopercle. In nearly their whole extent they are membranous, It is certain that 34 TTA [Vior. It: they were quite as extensive in Xiphactinus, and composed of very thin bone. Crook has figured them as extending well back from the orbit, and I find them pressed down on the metaptery- goid and hyomandibular of X. ¢haumas. Dr. Crook states that the ossified sclerotic of Xiphactinus forms a complete ring, meaning, I take it, that it does not con- sist of more than one piece of bone. Having a portion of the sclerotic in my possession which closely resembles those figured by Professor-Cope! (Gre, Vert, Pl. XE, Figs 3; Pl XLII ie: 4), I am inclined to believe that the sclerotic consisted of two separate pieces of bone, and this is the usual condition of the sclerotic of fishes. No one has yet, so far as I know, described the nasals, and I have not succeeded in identifying them. In Tarpon each of these bones is a rugose scale which lies partly on the outer border of the ethmoid. It might easily become detached dur- ing maceration, and this accident may have happened to this bone in the skulls of Xiphactinus that I have examined. The maxillaries and the premaxillaries’ are the most characteristic bones of this genus, and especially on the number and the charac- ter of the teeth borne by them have been founded most of the different species. Fre- quently, however, the premaxillary has been separated from the maxillary. I believe that the species may be identified from the con- dyles of the maxillary. At least, these con- dyles are quite different in the two species which I have been able to examine, X. thaumas and X. molossus, Fig. 2 represents the maxillary of the former species, Fie.73 that of XY. molossus. From these figures it will be seen that in Y. ¢haumas the posterior ieee, “wrecies “condyle (pimernc.) is notched,” behind; winile thaumas, maxillary and f f premaxillary,seen from that of P. molossus is excavated in front. above. x }. : It appears, too, that the condyle is more extended longitudinally in X. thawmas, more transversely in X. molossus. Examining the anterior condyle, the one which No. 1.] RAE GEMESBOF FOSSIL. ALSHES. 35 articulates with the ethmoid, we find that in YX. ¢haumas it is large and elongated, and approaches the posterior condyle within a distance equal to half its own length. In XY. molossus the condyle is much smaller, regularly oval, and far removed from the posterior condyle. It is to be expected that the other species of the genus will exhibit likewise their distinctive characters. I am inclined to believe that these condyles underwent some individual variation, and they have in many cases suffered distortion during fossilization, and this must be taken into account. The left maxilla belonging to the same individual of XY. mo/ossus from which pme{ Fig. 3 was drawn possessed an additional con- fis dylar surface, nearly round and small, just in front of the posterior condyle. It is to be in- ferred that the ethmoids of the various species, and especially the surface of the palatine with See ANT wiichethe postemonr, maxillary condyle artic- smasillay and pre- ulates, will exhibit characters corresponding ene ie eo to those shown by the latter. I call attention to the fact that it is as yet difficult to distin- guish the various species by means of characters furnished by the lower jaws. In the case of Y. molossus there are discrep- ancies between Professor Cope’s description of the number and character of the teeth and one of his figures. The lower jaw of the type specimen is figured on Pl. XX XIX of the Cretaceous Vertebrates and again on Pl. XL, Fig. 1. The statement is made in the text (p. 195) that there are in all 20 teeth; but in the figure last referred to there are 27 teeth represented, and these do not all agree in size either with the statements of the text or with the other figure. The explanation of this discrepancy, evidently, is that the figure on Pl. XL has, so far as many of the teeth are concerned, been erroneously restored. X. thaumas is said (op. ctt., p. 197) to have rather more numer- ous teeth than . mo/ossus, and in the specimen described there ‘are said to be 23. I possess two dentaries which I regard as belonging to X. ¢thauwmas. In these there are 24 Nae 36 HAY. [Vion, Ud teeth, and their sizes and arrangement agree well with Cope’s figure of the dentary of this species presented by him (Crez¢. Vert., Pl. XLIII, Fig. 3). Probably it will be well not to rely too much on the number of the teeth as a specific character. Professor Cope states (Proc. Amer. Philos. Soe, Vol, xa, p75) thatthe teeth of these fishes descend in their alveoli to the depth of an inch. The large teeth really have much. longer roots than thus amdicated- im vthe lower jaw the bases of the large teeth near the symphysis descend nearly to the lower border of the jaw. Fig. 4 p presents a view one-half the natural ee nee ee Se) Of thegsymphyceal, end! sofedae away to show the roots of the teeth. mandible of a species of Xiphactinus, a. seen from the outside. The bone has been broken away so as to expose the roots of the teeth, and portions, too, of these are missing. The teeth in life had a very large pulp, and the cavity containing this pulp had, since burial, been filled up with crystallized calcite. This, in the drawing, is stippled. Where the calcite has fallen out and exposed the inner surface of the dentine the shading has been made by _ perpendicular lines. The broken edges of the dentine itself are shaded by horizontal lines. Cope and Crook have both figured the articula- tion of the lower jaw with ati the quadrate. It appears Fic. 5-— Xiphactinus. Proximal end of lower to me that the figures of ee both are more or less erroneous, or, at least, misleading. Professor Cope (Cvet. Vert., p. 194) states that the articular is distinct, wedge-shaped, short, and supports half the cotylus. He describes the angular as having a prominent angle, like half an ellipse, and extending in a long sword-shaped process ” / ) ee ae No. I.] THE GENUS OF FOSSIL FISHES. 27) on the inside of the ramus to beyond its middle. A lower jaw of Xiphactinus (No. 3782, U. S. N. M.), in almost perfect condition (Fig. 5), enables me to correct some of these state- ments. Cope’s articular is not short, but its continuation for- ward forms the long sword-shaped process that he regards as belonging to the “angular.”’ In short, this articular corre- sponds to the autarticulare of Van Wijhe (Wederland. Archiv. f. Zool., vol. v, pp. 207-320) and originates from the ossification of Meckel’s cartilage. Cope’s angulare is not the true angu- lare, but is Van Wijhe’s dermarticulare, a membrane bone. In Lepisosteus, Amia, and Polypterus these bones remain distinct. Van Wijhe (of. czz., pp. 306, 307) makes the follow- ing statement in speaking of the elements of the lower jaw of the genera mentioned above: ‘Eine Vergleichung mit den Teleostiern zeigt, dass was bei diesen als Articulare angegeben wird durch eine Verschmelzung des Autarticulare mit dem Dermarticulare entstanden ist.’ Here, however, in this Cre- taceous genus of Teleosts, we find these elements still distinct from each other. In the genera of so-called Ganiods referred to above the autarticulare is very short ; but, relying on two good specimens of Xiphactinus and one of Ichthyodectes, I am confident that the proximal end of the autarticulare is continu- ous with the long sword-shaped process described by Cope, and that this process is entirely distinct from the dermarticulare. If the true angulare ever was present in Xiphactinus, it has become consolidated with the dermarticulare. In a specimen of Ichthyodectes there is present a surface to which an angulare seems to have been sutured. Crook represents it as present. Professor Cope’s figure of the lower jaw of Xiphactinus (Cret. Vert., P\. XX XIX) at first sight gives one the impression that the rounded head of the quadrate articulated with a similar rounded head belonging to the lower jaw. The latter, how- ever, is the “ prominent angle, like half of an ellipse,’ and the quadrate was supposed to enter its cotylus mesiad of this angle and well forward. My Fig. 5 shows the jaw seen from within. The cotylus is furnished partly by the autarticulare and partly by the dermarticulare. The head of the quadrate sits in its cotylus on the mesial side of the broad process of 38 HAY. [Von. IL. the dermarticulare. I know of no recent fish which possesses such an arrangement. The tarpon has a very different articu- lation in this region, since it resembles closely the articulation between two vertebrae of a bird. The advantages of such an articulation as that of Xiphactinus are obvious, since this species doubtless preyed on large fishes, and possibly on some of the large aquatic reptiles of its era. Fig. 6 represents another specimen of the jaw of Xiphactinus (No. 1646, U.S. N. M.); At geo is seen the condyle sof the quadrate ; 72777 is the lower end of the epihyal, and ¢.Z. the upper end of the Fic. 6. — X. molossus, showing quadrate, autarticulare, and hyoid bones. x 3. ceratohyal. These bones are closely appressed to the inner surface of the quadrate and of the lower jaw. The entopterygoid, or mesopterygoid (Fig. 7, m.pg.), has a smooth, slightly convex surface sloping inward and upward to form a partial floor for the orbit. Unless its width has been excessively altered by pressure, it was much narrower than the corresponding surface of Tarpon. In the latter the entoptery- gcoid meets the upper anterior angle of the quadrate, these two bones thus excluding the ectopterygoid from contact with the metapterygoid. In Xiphactinus, on the contrary, these two last-mentioned bones have a considerable union (Fig. 7). The bones of the palato-quadrate arch have been described as being devoid of teeth. I have, however, found a consider- No. tz] PAE GENGS, OF FOSSIL FISHES, 39 able patch of small teeth on the entopterygoid (Fig. 7, ¢.), and another smaller patch on the ectopterygoid (Fig. 7, 7.1). In the Tarpon teeth occur on the vomer, parasphenoid, pterygoids, and even on the quadrate. The hyomandibular (Fig. 7, 4m.) is in many respects like that of Tarpon, but, like the other bones of the extinct genus, is of more massive construction. The anterior border of the Fic. 7. — X. thawmas. Hyomandibular and palato-pterygoid bones. x }. bone extends further forward than it does in Tarpon. In the latter the anterior border falls, with a sigmoid curve, in a general downward direction, crossing the posterior angle of the metapterygoid. In Xiphactinus the anterior border of the hyomandibular runs rapidly forward, so as to come into contact with and pass mesiad of the posterior border of the entoptery- goid. The greatest width of the hyomandibular from the articulation of the operculum to the anterior border is nearly equal to the distance from the anterior border to the anterior end of the palatine. In Tarpon the latter distance is about 2.5 times the greatest width of the hyomandibular. However, 40 Veal) [Viorsir. on account of the relatively greater depth of the head in Xiphactinus, the width of its hyomandibular has about the same ratio to its length that we find in Tarpon. In Tarpon the process for articulation of the operculum pro- jects from the hinder border of the bone more than in Xiphac- tinus. In the latter genus the surface for the operculum of X. molossus scarcely passes beyond the border of the bone ; but in X. ¢haumas the surface is at the extremity of a consid- erable process. As a result, perhaps, of its large size, the hyomandibular of Xiphactinus, as well as that of Tarpon, is provided with promi- nent ridges and depressions, and with foramina leading into its interior. Many of these are repeated in the two genera with much faithfulness. In both genera there is found running down near the middle of the outer surface of the bone a high crest, like the spine of the human scapula. This crest has its origin, we may say, in two.low rounded ridges, one beginning at the anterior end of the hyomandibular head, the other at its posterior end, the two ridges converging and meeting opposite the articular surface for the opercular. Here the resultant crest becomes much more elevated, thin, and sharp, and con- tinues to the lower end of the bone. ‘The plane of the crest is directly outward and slightly backward. Both in front and behind the crest is a deep fossa, the posterior one the best defined. The anterior border of the preopercular occupies a part of the posterior fossa. This fossa, in Tarpon, ends above in a deep depression immediately in front of this process for articulation of the opercular; but from the upper border of this depression one or more large canals enter the bone, and, passing upward, emerge by several mouths in another depres- sion on the inside of the bone just below the head of the hyomandibular. It is quite probable that one or more branches of the facial nerve pass downward through these canals. In Xiphactinus the posterior fossa ends above, just as de- scribed, and broad channels are seen passing upward from it in the bone ; while on the outside, just below the anterior end of this hyomandibular articulating surface, there is 2 depression ._- Now] HE GENES OF FOSSIL FISHES, 41 like that found in Tarpon. There can be no doubt, therefore, that the upper end of this hyomandibular is hollowed out simi- larly in the two genera. In a specimen of XY. mo/ossus before me, two bridges of bone are thrown across the upper end of the posterior fossa on the mesial surface of the hyomandibular. On the same surface of the hyomandibular there is a well-marked median crest, in front of which is a broad shallow fossa. It is in the upper end of this fossa that the depression is found that has just been described, and which in Xiphactinus is represented in Fig. 7, d. This depression, it is to be noted, faces the deep fossa which has already been described as occurring in the side wall of the skull. Its significance can only be deter- mined by an examination of a fresh Tarpon. Both depressions probably furnish insertions for muscles. The opercular of this genus is not well known. Cope states that it is thin and broad. Crook figures a portion of the bone, but this reaches downward only about to the middle of the preoperculum. I have a fragment of a bone 50 mm. by 100 mm. which appears to be the opercular of XY. thaumas, and this, too, has every appearance of ending about halfway down the preoperculum. This piece of bone has an articular surface resembling that of Tarpon for connection with the preopercular, and, like Tarpon, there are just below this surface, and on the inner side of this bone, one or two large openings into the interior of the bone. This mention of the opercular may attract attention to it. It appears rather improbable that it is really so short as above described. In each of the three specimens of Xiphactinus before me there is present, attached to the posterior outer angle of the skull, a bone which seems to occupy the position of the post- temporal. If such it is, it was very different from that of Tarpon. It is not much over an inch in length, and less than two inches broad, but very thick. In a specimen of Tarpon the bone is rather thin and much longer. In Xiphactinus, on account of the crushed condition of the skulls, the relations of the bone are hard to make out, but it seems to be connected with the opisthotic and the epiotic. In many fishes the temporal bone is very short and stout. 42 HAG [Vot. II. If I have correctly determined the bone figured here (Fig. 8), the supraclavicular, by its great length, compensated for the shortness of the post-temporal. Its length is about equal to that of the parasphenoid. The shoulder girdle has received very unfortunate treatment. It appears to have been misunderstood by both Professor Cope and Dr. Crook, being by both writers described in an inverted position. Cope gives figures in his Cretaceous Vertebrates as follows: Pl. XL, Fig. 9 (cleithrum?); BE XLII, Bigss2-53 Ek XE, Pigs. 10; 112 Most: of theserdepice the scapula and the parts immediately adjacent. Cope describes the ‘coracoid”’ as a stout flat rod, narrower than the cleithrum (clavicle), and appressed to the inner face of the latter nearly to its distal end (Cve¢. Vert., p. 186). He was unable to state whether or not there was present a precoracoid, but said that the “coracoid”’ occupied the position of the pre- coracoid in some fishes. According to his conception, the scapula was placed high up on the body, although his restoration of the fish Fic. 8.— X. molossus. On Pl. LV does not so indicate. Dr. Crook i ee presented figures of the girdle of Xiphac- tinus (of. cz¢t., Pl. XVII), of his Jlchkthyodectes polymicrodus (Pl] XViiand of 7. avzazdes (Pl. XN); - Therejcan. be: no doubt that all these figures represent bones which belong to the side of the body opposite to that to which they are assigned, and that what is regarded as the ventral end is the dorsal. To demonstrate this it is only necessary to compare the figures with the prepared shoulder girdle of a shad. Dr. Crook recognized that Professor Cope’s coracoid was really the pre- coracoid; nevertheless, he has represented it as running ventrally from the coracoid, instead of toward the dorsal end of the ry ZA ht = EE ; Tie GE Lee ay elgg SEALS BLY LILLE SOL Te, Fe, cleithrum. One result of Crook’s error is that the coracoid is brought into a position dorsad of the scapula. The materials employed by both Cope and Crook were defective, that portion of the INOS 1.]] WIE GENES OF BPOSSIL FISHES: 43 cleithrum and coracoid belonging ventrad of the fin articulation being mostly wanting. Fortunately I have on one block both the right and left halves of the shoulder girdle in nearly perfect condition. To one half are also attached some of the remark- able fin rays of this genus. A figure is presented of the right half of the girdle seen from without (Fig. 9). In this figure the cleithrum conceals a part of the coracoid,! but the latter is so broad that a consider- able portion of it is seen. In Tarpon there is along the upper border of the coracoid a long fontanelle between this bone and the cleithrum. If sucha fontanelle was present in Xiphactinus it is con- cealed beneath the clei- thrum. In Tarpon there are two or three foramina : in the coracoid just below Se dhe scala, Theva Fic. 9. — Xiphactinus. Shoulder girdle. x }. life closed by membrane. They are wanting in Xiphactinus. The outer surface of the dorsal limb of the cleithrum of Xiphactinus is broad and convex to the very hinder border. It thus resembles Tarpon, and differs from A/osa sapidissima, in which the hinder portion of this surface is rough and excavated for muscles. In the extinct genus there is an extensive fossa on the inner surface of the upper limb of the cleithrum. The upper half of this fossa lies between an outer and an inner plate of the cleithrum. Further down, the fossa is limited mesially by the precoracoid. There seems to be no such fossa in Tarpon, and that of Alosa is very shallow. In both Tarpon and Alosa the precoracoid is a much less important bone than it is in Xiphactinus. 1 I employ for the elements of the shoulder girdle the terms in common use, except that I use Gegenbaur’s name cleithrum instead of clavicle. For the latter element Dr. Gill has proposed the term proscapula; for coracoid, hypocoracoid ; for scapula, hypercoracoid ; and for precoracoid, mesocoracoid. 44 LIAN [Vor The pectoral fins have been described by Professor Cope and Dr: ‘Crook (Cope, “Cree, “Vert., spp. 186; T1093," 2045" Crook; Palacontographica, vol. Xxxix, p. 119). Neither of these authors compares the fin structure with that of other fishes, although a community of structure is perhaps implied. The large saber- shaped spines, each consisting of an upper and a lower half, are remarkable enough; but when comparison is made with the fins of a shad or of a tarpon the arrangement of all the parts is easily comprehended. The first pectoral ray of Xiphactinus resembles quite closely that of Tarpon. It differed in being, relatively to the size of its owner, somewhat, but not enormously, larger. It differed further in having lost, apparently to the very tip of the ray, the cross-segmentation. In Tarpon this persists in the distal half of the ray. Doubtless, the spine-like rays of Xiphactinus were not so flat as they are now presented to us. It is quite probable that the rays succeeding the first one were, toward their distal extrem- Itles, niot, Only jenosis: segmented, but also longitu- dinally split, as in other fishes. Professor Cope (Cvez. Vert., p. 186) has described the ventral fins and their sup- porting bones. The latter, pelvic actinosts, usually termed the pelvic bones, are called by Professor Cope the femora. Healso figures them (p. 192;Figy of and Pie Figs.) 7,0.7@)0) elipossessma well-preserved specimen of the pelvis and the ventral Big) 20: he wteumasa Pelvicibonesanclibascae finciot A tawwes sane from of fin. x 3. these 1 becomes. evident that the pelvis figured by Professor Cope was very defective. This may be seen by comparing the figures above referred to with my Fig. 10, which represents the pelvic actinosts seen INOS 1: )} DHE GENGS, OF FOSSIL “FISHES. 45 from below. These pelvic bones are, as stated by Professor Cope, massive, and expanded vertically on the outer side to support the facets for the ventral fin. The right and left bones are strongly sutured together. In front of the facets and of the suture the bones become much thinner, but wider. At the same time there is, descending from the outer border of each bone, a crest of moderate height, while from the same portion of the outer border there arises a much higher crest, so that a cross-section of the pelvis in front of the fin articulation would somewhat resemble this figure Ss Just laterad of the inner border of each bone there is found on the upper side a prominent ridge, running from the fin articulation toward the anterior end of the bone. On the lower side, and nearly opposite the upper ridge, is a similar ridge. Elsewhere the bone is very thin. If now the thin por- tions of bone were broken and had crumbled away, there would be left a thick process standing out on each side and two rods, the ridges just described. Such was doubtless the condition of the bones which Professor Cope figured. Just how far forward the pelvic actinosts extended is not known, since those figured by both Professor Cope and myself have been broken. In Tarpon these bones are very long and slender. In Elops, a close ally of Tarpon, they are relatively shorter and also broader behind. Professor Cope has quite correctly de- scribed the facets for articulation of the ventral fin. My Fig. 11 represents the posi- tions and forms of these facets. The bone is that of the right side, and is looked at laterally. The upper facet is for the recep- Fie. 11.—%. thawmas. Side tion! G@ian articular ‘surface on the base of “e™ of smitazen fore’ the upper half of the first ventral ray ; the large lower facet for an articular surface on the lower half of the same ray. The other two facets are for succeeding rays, or possibly for baseosts. There was undoubtedly a disc-like baseost between the upper half of the first ray and the articular surface of the pelvic bone; and there may have been other rudimentary baseosts. 46 HAY. (Vor. II. The planes of the two surfaces which support the first ray approximately coincide and are directed outward, somewhat backward, and slightly downward, a position different from that given by Cope. The first and spine-like ventral ray is constructed like the first pectoral, and may also be compared with that of other Iso- spondyli such as the lake-trout (Cristivomer) and tarpon. Like the first pectoral ray, it seems to have wholly lost its transverse segmentation. These spines, however, show no special physo- clistous characters, as Professor Cope supposed they did. The first ray of my specimen of the ventral fin of 1. ¢*haumas is 38 mm. wide at the base, and was perhaps originally 60 cm. long ; but this was a very large fish, since its upper jaw had a length of 38 cm. Professor Cope states that the first three rays were spines, and that there were probably no additional rays. However, it seems probable that the rays succeeding the first one were much feebler, were segmented, and longitu- dinally split. There were certainly more than three rays, for, in my specimen, I make out six or eight, and there were prob- ably nine. The second ray has only about one-third the diam- eter of the first, and those following become gradually, probably rapidly, reduced. The inner rays must have been very short, since I find finely split and segmented rays at a distance of only 90 mm. from the base of the rays. In Fig. 10 I have represented the bases of the first and second rays. On the anterior border of the first ray the broader upper half of the ray is seen to project some distance beyond the lower half. It will be noticed, also, that on account of the expansion of the pelvic bone in front of the fin articulation, the fin could not be brought in front of a perpendicular to the body at that point. In Cristivomer and Tarpon I find a rudimentary ray in front of, and lying on the base of the first ray. It is short, but has a very long muscular process directed forward and upward. It is more reduced in Tarpon than in Cristivomer. I find no evidence of its presence in Xiphactinus. The vertebral column has been described by Professor Cope (Cret. Vert., pp. 188, 193, 195, 199), and briefly by Dr. Crook (op: ctt:, Pp. 117). “1 have’ noted’ some hitherto undescribed No. I.] tHE GENOSSOF FOSSIL ‘FISHES. 47 peculiarities which are of interest. I have a considerable number of the vertebrae of X. ¢hawmas, including unconnected vertebrae belonging to the anterior portion of the column, and a section about two feet in length from the tail region. I have also numerous vertebrae belonging to some indeterminable species, probably X. molossus. Access is permitted me to vertebrae of probably 1. molossus, belonging to the United States National Museum. These belong mostly to the tail. I will first describe the peculiar structure of the neural arches in the latter region. The drawing pre- sented will assist in the understanding of my de- ; ( seniption (kis. 12). The neural arches here, as else- where in this fish, are con- mected “with the ‘centra by suture, and have usually fallen out before burial, leaving long grooves where their bases were inserted. This was the case with the third vertebra behind the right-hand one shown in Fig. 12. When we come to examine the arches more closely we discover that each lateral half is not a single piece, but consists of two pieces, a basal piece (a.v.a.) and the Fic. 12. —X. thaumas. Two tail vertebrae. x }. arch proper (z.a.). That the proper arch is a distinct piece is shown, not only by the existence of a suture, but likewise by the fact that in the vertebra on the left hand of the figure the arch has fallen out of its place before fossilization. The basal or accessory piece is inserted by a shallow gomphosis into the centrum for nearly the whole length of the latter. It rises high in front and projects so far forward as to come into con- tact with the basal piece of the next vertebrain front. Behind, the basal piece is directed upward and backward in a rather slender process, which abuts against the anterior edge of the basal piece of the next vertebra behind. It is thus seen that these basal pieces provide the anterior and posterior zygapoph- yses. They remind us of the articulating processes of certain other fishes, Mugil, etc. Between the anterior and posterior 48 TAME [Vicrshr processes the basal pieces are excavated to receive the bases of the neural arch, as shown in the figure. The two basal pieces of each vertebra are distinct. Together they seem to forma saddle in which rides the neural arch. I find this same structure of the neural arches in some of the vertebrae belonging to the specimens in the United States National Museum ; but in one section of connected vertebrae an arch like those above described is succeeded in the next vertebra behind by an arch in which every trace of a suture between the arch and the apparent basal piece is lost. This vertebra is shown in Fig. 13. The form of the base of the arch is not greatly different from that of the arch with accessory piece in Fig. 11, and we may even "3%, Tauctor! neural arch. x j. convince ourselves that we’can trace a part of the Fic. 13. — Xiphac- boundary line between the two portions. There is evidently at this point of the vertebral column a sudden change from neural arches furnished with basal accessory pieces to arches without these, or consolidated with them. Further backward the form of the arches becomes modified somewhat, so that they resemble the one shown in Fig. 14. A section 14 inches long and containing 7 vertebrae having arches of this kind is before me. This condition shows us that the neural arches which are provided with basal pieces are confined to the anterior or middle portion of the tail region, while the hinder portion contains no such vertebral structures. We are reminded that in Amia the middle portion of the caudal vertebral column is composed of two rings for each muscular segment, while the anterior and posterior portions have vertebral centra of the ordi- nary kind. It seems as if the tail portion of the vertebral column of the Amioid fishes and of the Fic. 144.-xi. Isospondyli retained primitive conditions longer phachnuss Seu than sthe abdominals portion: ral arch without ‘ ; ; z : accessory piece. It is difficult to determine what explanation is to x t. . . ° be given of the presence of these basal pieces. The so-called zygapophyses of fishes are regarded as being outgrowths from the neural arches, exogenous and not autoge- No, I.] THE GENGS OF FOSSIL FISHES, 49 nous processes. It might be said, possibly, that the basal pieces are the proper arches, while the pieces which are borne on them are the spinous processes. I hold that there are two objections to this view. The first is, that what are sometimes called free spinous processes are always unpaired pieces. The second is, that when the lateral halves of the arches remain distinct from each other and are prolonged into spines, as they are in various fishes, Amia and Salmo, for instance, the spinous portion is never, so far as we know, developed in the embryo as pleces separate from the base of the arcuale. This is true in the case of Amia, which I have investigated. We must, therefore, seek some other explanation. The key to the under- standing of the problem is, it seems to me, to be found in the vertebral column of that primitive fish, Amia. We may call this fish to our assistance since the Isospondyli are believed to have had ancestors not far removed from Amia. In the middle region of the tail of Amia there are for each muscular segment two vertebral rings, the one bearing the arches, upper and lower, the other archless. If a transverse section be taken through the middle of the arch-bearing ring, there will be found an X of cartilage, the upper arms of which are continuous with the cartilage of the neural arch. In like manner the lower arms will be seen to be continuous with the cartilage of the haemal arch. If a section is made similarly through the archless disc, a similar X of cartilage is found ; but the arms project beyond the outer surface of the disc but a short distance. These archless discs are developed in Amia from ossifications arising in the intercalated cartilages, upper and lower, and the arms of the X are the unossified portions of these cartilages. There appears to be no reason why these intercalated cartilages should not sometimes take on a hyper- trophied growth. In the sharks they often become consider- ably larger than the true neural arches themselves. In case these intercalated cartilages should become thus enlarged and arch-like, each might develop a bony investment that would simulate the bony neural half-arch, and thus would rest on the top of its proper epicentum.! 1 For figures illustrating the architecture of the vertebral column of Amia, see the May number of the American Naturalist of the present year. 50 HAY. [Vou. II. Coming now to the anterior region of the vertebral column of Amia, we find that each vertebra is formed through the suppression of certain of the elements which, in the tail region, constitute the vertebral rings or discs, and the union of the remaining elements of each muscular segment into a single mass. The lower intercalated cartilages are suppressed. The upper intercalated cartilages hypertrophy, and their ossifications unite with the bones developed in the bases of the lower arch, thus giving origin to the centrum. The ossification that we might expect to find developing in the base of the cartilaginous neural arch, the epicentrum, is aborted, while the ossification of the enlarged intercalated cartilage, the pleurocentrum, pushes itself into the place of the epicentrum, and thereafter supports the neural arch. Now we have the choice of two suppositions, neither of which, however, may be the true one. We may hold that a distinct bone was developed in the somewhat elongated and projecting intercalated cartilage, and this, of course, rested on the top of the pleurocentrum ; when the latter was pushed for- ward beneath the neural arch to take the place of the aborted epicentrum, this newly developed bone was carried along and was thus brought between the pleurocentrum and the base of the neural arch. Or we may hold that the bone which I have found in Xiphactinus supporting the true neural arch is simply the epicentrum itself, aborted, indeed, in Amia, nevertheless per- sisting in Xiphactinus, but crowded upward out of its original seat on the notochord. Either of the above suppositions presupposes that the upper half of the vertebral centrum takes its origin from the pleuro- centrum. Professor Cope held that the vertebrae of fishes are “intercentra,” that is, have originated in the suppression of all the other elements through the excessive development of the hypocentra. But the very existence, in many genera, of a cartilaginous X in a transverse section of the centrum is proof that its upper portion has been derived from either the bases of the upper arches or the pleurocentra. The deep gashes in the vertebral centra of Xiphactinus,, where the arches have WOT.) RAE GENS OF FOSSIL. FISHES. 51 fallen out, furnish evidences that this cartilaginous X was present. The anterior neural arches of Xiphactinus, probably all of those belonging to the abdominal region, are very different from those of the tail. One of these abdominal neural half- arches, as seen from without, is presented in Fig. rs ; another, seen from the mesial side, is given in Fig. 16. These neural arches are coossified neither with the vertebral centra nor with their fellow bones. The base is hemispherical and planted in a broad excavation in the upper surface of the centrum. The. two excavations of each centrum are close together, and it seems probable that the juxtaposed borders of the right and Fic. 15. Fic. 16. Fic. 15.— X. thaumas. Neural arch near head, seen from without. x 3. Fic. 16.— X. thaumas. Neural arch near head, seen from within. x 3. left elements of each arch are in contact both below and above the neural canal. Behind the base of the neural half-arch is a broad smooth surface (Fig. 16, art. szzf.) looking mesially, and in life coming into contact with a similar surface on the anterior end of the next vertebra behind and looking outward (Fig. 15, art. surf.). These surfaces remind us strongly of the zygapophyses of the higher animals. At the base of each of these half-arches we find a strong rod- like process directed outward and backward. These processes are the epineurals (Fig. 15, ¢..7.). They were confluent with the bases of the arches just as they are in Alosa and Tarpon. It is entirely probable that Xiphactinus and its allies were as bony fishes as our just-mentioned modern genera. The excavations for the insertion of the neural arches are broadest toward the region of the head. Farther backward 52 HAY. [Vou. II. they become longer and narrower. Professor Cope describes all the centra as having on each side two lateral grooves, except the two or three centra near the head, called by him the ‘‘cervicals.”’ However, so far as I can determine, the vertebral centra of the abdominal region have only one lateral groove on each side. Close to the head this groove becomes quite insignificant and is placed close to the pit for the neural arch. The attachment of the ribs deserves notice. They are not joined directly to the centra but through the medium of dis- tinct pieces of bone, the parapophyses. These are very short, and are sunken in circular pits so deeply that they scarcely rise above the surface of the centrum. Each has a concavity for the reception of the head of a rib. Some specimens in my possession have the parapophysial pits empty. In others the parapophyses are present, but without rib heads. In a few the head of the rib yet remains. _ Distinct parapophyses are found in a number of fishes, as Cristivomer, Alosa. It may be here remarked that the vertebral centra of Tarpon are very different from those of Xiphactinus, being very solid, smooth, and wholly devoid of the deep lateral grooves. Most of the neural arches have become coossified with the centra, and appearances indicate that in the young fish there were separate parapophyses, which later coalesced with the centra. The vertebral column has attained a much higher grade of development than that of Xiphactinus. The specimen that I have above referred to XY. ¢thaumas, I believe to be such ; but lest it prove to be something else I shall here attempt a description and a comparison with other species. It certainly is not YY. molossus, since that species has the distal extremity of the maxillary upturned like a saber. Moreover, as I have already illustrated (Fig. 2), the condyles are very different from those of undoubted specimens of UX. molossus. It cannot be X. mudgez, since this species possesses four subequal teeth in the premaxillary, while in my specimen there are present only two teeth. Moreover, the vertical extent of the maxillary behind the posterior condyle is too great. The specimen possibly belongs to XY. lestrio. Cope No: 1. ] HAE GRVOS. OF FOSSIL FISHES: 53 speaks of the two maxillary condyles of that species as being large and separated by a space. This description, though vague, would fit my specimen. But YX. /estrio is stated to have three, and sometimes four, premaxillary teeth. As I have said, I find no evidence of a third tooth. The total length of the upper jaw of my specimen, including the premaxillary, is 380 mm. The height of the maxillary from lower border to top of posterior condyle is 125 mm., almost exactly one- third the length. Applying this proportion to Cope’s figure Oipens esiro M@ret. Vert; Pl XLII, Fig: 1), we find that his drawing of the maxillary would have to end at the right hand within about 8 mm. beyond its present limit, in order to represent the complete bone. It is very evident that a much more considerable piece of that maxillary was wanting. Had this missing portion of that bone had the form and proportions possessed by my specimen, the drawing would have to extend 25 mm. further to the right. This would make the jaw much longer in proportion to its height than my specimen. As a matter of fact, I find no serious discrepancy between Cope’s description of his Portheus thaumas and my specimen. I give description of the upper jaw. Upper jaw heavy and massive ; its height being apparently greater in proportion to its length than in other species, one to three. Premaxillary broadly oval ; its major axis 130 mm., its transverse IIO mm., its greatest thickness 40 mm. Teeth two, the most anterior projecting 55 mm. beyond the bone ; its diameter at base 20 mm. Second tooth 22 mm. long, prob- ably not full grown. Maxillary extending forward against inner surface of premaxillary nearly to the anterior border of latter. Condyles as shown in Fig. 2. Tooth border sinuous, slightly concave just behind premaxillary suture, then convex to beyond large teeth, then again more strongly concave ; finally convex, and rounding into the distal border. Upper border descending rapidly from posterior condyle and concave to point three-fifths of distance to distal extremity, there forming an angle, and again concave until it begins to round into the distal border. In general, the distal third of the maxillary bends downward instead of upward. On dental border there are in front, first, 54 HAY. traces of four or five small and medium teeth, then five large teeth, two of which are yet present and projecting 40 mm., then some 40 teeth from 11 mm. in length to mere points. Extreme height of maxillary 125 mm.; its height 72 mm., at a point on upper border 125 mm. behind the anterior border of posterior condyle. This enters into the total length of the upper jaw 5.3 times, and into height at condyle 1.75 times. I take pleasure in acknowledging my obligations to Prof. F. A. Lucas, in charge of the osteological collections of the United States National Museum, for the generous manner in which he has given me access to such materials as I have needed. Volume IT. October, 1598. Number 2. ZO@roGYCAL BULLETIN OBSERVATIONS ON THE ANATOMY .OF A SPECIES OESEE AGNES EIS sFOUND: PARASITIC ON THE UNIONIDAE OF LAKE CHAUTAUOUA, HENRY LESLIE OSBORN. Tue facts recorded in this article were gathered partly on living material at Chautauqua, New York, in the Biological Laboratory of Chautauqua College of Liberal Arts, and chiefly in the Biological Laboratory of Hamline University, St. Paul, Minn. The fluke was first noticed in July, 1895, in speci- mens of Anodonta which were being used in class work. A few drawings were made at the time, but no attempt at identi- fication could be undertaken then, as the necessary books were not at hand. Later in the year I consulted Bronn’s KA/assex und Ordnungen and decided that it could not possibly be regarded as the usual parasite of the Unionidae Aspidogaster conchicola, and recognized that it bore a remarkably close superficial resemblance to a species designated in that work as the Aspidogaster lenoiri of Poirier (85). Such points as could be determined by a mere surface study of the animal indicated clearly a very decided likeness between Poirier’s spe- cies and mine, but the point which made me hesitate in asso- ciating them at that time was the fact that Poirier’s animal is only recorded from a single region, vzz., Senegal, Africa, and from a single host, TZetrathyra vatllantt, a chelontan. Unfamiliarity with the Trematoda and pressure of other work prevented me from investigating the case at that time, and I did not really take it in hand till the fall of 1897, when I was gathering the facts for a paper, Osborn (’98), on the dis- 56 OSBORN. [Vou. II. coloration of Anodonta shells by a distomid parasite. I planned then to append the facts about Platyaspis to that paper in a short note. But as soon as I looked into the case I found a number of interesting points whose novel character made it necessary to present the evidences in the case in a way impos- sible in a mere note, and hence I withheld the matter and have made it the subject of another article. Poirier referred the animal which he discovered to the genus Aspidogaster, and Braun (92) in Bronn’s Klassen und Ordnungen followed his assignment of the animal to that genus. But Monticelli (92), in his revision of the Aspidobothridae which accompanies the paper on Coty/aspis in Leukart’s Festschrift, showed that Poi- rier’s animal cannot be regarded as an Asfzdogaster, and erected for it the genus Platyaspis, of which it has thus far been the solitary species. So far as I have been able to ascertain, Platy- aspis has never been recognized, excepting from the one locality in which Poirier found and described it. It is, consequently, an interesting and remarkable fact that it should occur in this country and in a very different host, and a fact which should not be presented without a sufficiently detailed account of the evidence to compel belief in the correctness of the observation. I have not been able to decide in my own mind whether the Chautauqua animal is specifically distinct from the African species or not. This is partly because I have not as yet had access to the original descriptions of Poirier, and do not know how absolutely exact his account and the reports of them are. The Chautauqua animal is slightly different from his, but not more so than might be consistent with membership in the same species. In case the Chautauqua animal proves to be distinct, I shall propose for it the name Platyaspis anodontae, and for convenience shall so term it in this article. The facts which are contributed in this paper are derived from studies of preserved material made in the Biological Lab- oratory of Hamline University. At the time of their discovery I made sketches of the living animals, but did not attempt to study them in detail. Last summer I preserved a few by drop- ping them into cold saturated aqueous corrosive sublimate solu- tion. The material has its limitations, but is sufficiently well Now| aVaLOMY OF AISPECIES OF PLATYASPIS. 57 preserved to enable one, by making total preparations and by serial sectionizing, to recognize all the most important anatom- ical features of the animal, and, in addition, to see histological detail enough to supplement the anatomical identification of the organs. But I have not been able to demonstrate on the preserved material the exact relation of the different members of the reproductive system, or to follow out the branchings of the excretory system. This, anda more careful study of the histology, I hope to make during the coming summer with the aid of living material. In the meantime I will report the facts as already determined. I have had only a partial access to the literature of the sub- ject, but gladly acknowledge my especial indebtedness to Monti- celli’s article in Leukart’s Festschrift, in which he gives some account with illustrations of Platyaspis, and to Stafford’s arti- cle on Aspidogaster. These and other articles referred to are indicated at the conclusion of this paper. I am also much indebted to Dr. W. S. Nickerson of the University of Minne- sota, for the privilege of examining his trematode preparations and for friendly advice as to methods of trematode study. HABITs. P. anodontae is habitually, if not exclusively, found, not in the pericardial chamber or cavities of the nephridium, but in the mantle chamber, where it is attached either to the sur- face of the visceral mass, the inner surface of the gill, or to the under surface of the kidney, where the mantle and cloacal chambers communicate anteriorly. While I have not made a sufficiently complete search to be able to assert that it is never located inside of the pericardial chamber or kidney, I feel confi- dent that if it is found there that position is not habitual. This point will receive particular attention in my later work. In its location P. anodontae is thus ectoparasitic, and hence decidedly unlike Aspidogaster, which, according to authors, is habitu- ally found inside the pericardial cavity and nephridium. Thus Stafford (96), p. 8, says: ‘On opening the Anodonta the parasites are often visible in the transparent pericardium. It 58 OSBORN. [Vot. II. was in this organ that the great mass of Aspidogasters was obtained ; and, generally, they were found closely packed into the anterior corners, at the entrance into the kidney and peri- cardial gland. In these latter organs I have found a good num- ber, too, but in no other organ have I succeeded in finding any, although I have taken considerable trouble to find evidence of the migration of the young animals.’’ And a similar impres- sion is given by Huxley (72), Hoyle, Excyclopaedia Britannica, XXIII, p. 540, and Monticelli (92), as well as in most of the current text-books, etc. Its situation is also quite unlike that attributed by Poirier to Platyaspis lenoiri, which is an internal parasite of the turtle. We do not know that Platyaspis does not have two hosts, but the supposition is unlikely in view of the habits of the family Aspidobothridae, and if it should prove to be the fact that it has only one, then these two Platyaspid forms are very differ- ent indeed in their host relations. I am not prepared at present to say much about the host dis- tribution of the parasite, but I can say that in Lake Chautau- qua it is chiefly but not absolutely confined to Anodonta. The following Unionidae have been recognized growing in close proximity: Azodonta plana Lea, Unio luteolus Lam, U. edentula, U. phaseolus Hld., U. gibbosus Brus.; and while Anodonta seems to be the most usual host, the parasite has been noticed rarely in U. luteolus. I have not met with Aspidogaster at Lake Chautauqua, but as it is an endoparasite it may easily be present there and have escaped my notice, since I have not made a point of searching carefully through the pericardium and other organs in which it is reported as likely to occur. I do not, however, imagine that it is very common, for, in case it were, it would surely have attracted attention during the many dissections that have been made by the students. I have not as yet made a strict study of the habits of the parasite. On opening the mantle cavity of the host in air, one of course subjects the parasite to very unnatural conditions; at such time it adheres to the surface of the host by means of its enormous and highly complex ventral sucker, and its anterior Non2i\— ANATOMY OF AUSPECIES OF PLEATVYASPIS. 59 end is moved in an exploring sort of way. Under compression the anterior end is seen to be very mobile, assuming successively an immense number of apparently random shapes with great rapidity and ease. EXTERNAL ANATOMY. The individuals externally resemble (see Figs. 1-3) the form represented in Bronn and Monticelli for P. denozrz so exactly that at the outset I was hardly sure that the animals _M. 10 -Oc. 20 Em ue ae M. G. Dy EMG? qe Fic. 2: Fic. 1.1— Camera lucida dorsal view of P. axodontae as a transparent object, made from a total preparation in Canada balsam, and by reconstruction from serial sections. Fic. 2.— Surface view of the ventral surface of P. axodontae, drawn witha camera lucida. The geni- tal opening was not seen in the animal, but is introduced from the sections (camera lucida). are specifically distinct. There is a division of the body into two parts (see Fig. 3), an anterior and dorsal tubular elongate body, resting upon a broad, flat ventral and posterior portion, 1 The following reference letters are used in all the figures: D., Diaphragm; Em., Embryos; Ex. P., Excretory pore; G. P., Genital pore; Int., Intestine; M., Mouth ; M. G. D., Male genital duct ; Nv., Nerve trunk; Oc., Eye; Ov., Ovary; P., Parenchyma; Sp., Spermary; V., Vagina; Vt., Vitellaria. 60 OSBORN. [Vou. Il. much after the fashion of the relation of a snail’s body to its foot. These divisions are maintained internally to a consider- able extent, as will be seen later, the alimentary and excretory systems being confined to the body, or “neck,” as the dorsal tubular portion is called, while the reproductive system is largely located in the ventral and expanded “foot.” The entire length of the animal from the tip of the foot behind to the tip of the neck in front (in an alcoholic specimen) was, in one case, 1.6 mm. ; Monticelli gives 1.7 mm. for P. lenozri. The foot, or ventral sucker, is 1.3 mm. long and 1 mm. broad, and is thus the most prominent external feature of the animal. There is no differenti- ated oral sucker, the wall of the body at the ante- rior end being thin and Bie a aaeeerrt | ddhieaee and covet (see Fig. 4) unlike the much thickened condition found in Distomids, and generally in the trematodes, where a distinct oral sucker is present. The generative opening (Fig. 1, G. P.) is located in the middle ventral line, near the junction of the anterior region with the foot. The hinder broadened portion of the animal consists of a dorsal portion which shades down impercep- tibly laterally and posteriorly from the tubular anterior portion and fades out posteriorly to form the broad, flat dorsal sur- face of the sucker. This latter extends into an extremely thin rim all around the edge of the foot, and includes a flap which extends in front of the junction of the sucker with the anterior region of the body. On the hinder dorsal surface of the body, near the extreme posterior end, is located the opening of the excretory system (Figs 1, Ex: Shs): The ventral sucker itself is subdivided, the plan of its sub- divisions being entirely unlike that of Aspidogaster, and as dis- tinctly similar to that of Platyaspis. The surface of the sucker (see Fig. 2) is regularly subdivided by transverse and longi- tudinal folds into compartments, of which there is a distinct peripheral series of 20 compartments and a median series of g, making 29 compartments in all. The precise position of the No. 2.] ANATOMY OF A SPECIES OF PLATYASPIS. 61 ridges which border these compartments is shown in the cam- era lucida drawing of the ventral surface. This number of compartments is different from that given by authors for P. lenoiri, where there are 25 compartments, 18 in the periph- eral series, and only 7 in the median row, in place of 20 in the peripheral and 9 in the median row, as here. I have, however, not determined how constant the plan of this sub- division of the ventral sucker is, but I have found the number given in my figure in about twenty cases. INTERNAL ANATOMY. The animal is covered with the usual trematode body wall, consisting of a thin and delicate cuticle, beneath which, in sec- tions, the ends of muscle fibers are recognizable. Internally Fic. 4. Fic. 5. Fic. 4. — Transverse section, passing through the eyes, level Oc. in Fig. 1, No. 12 of series (camera lucida). Fic. 5. — Transverse section, No. 31 of series. The left of the section is the right side of the animal (camera lucida). the inter-spaces between the organs are filled in with the usual parenchymatous tissue. This is subdivided into that of the body and that of the foot by a transverse diaphragm, as seen in Fig. 6. The alimentary and excretory systems lie wholly dorsal to this structure, as well as the terminal ducts of the reproductive system, while the gonads themselves, and the vitellaria, are wholly ventral to the diaphragm. The alimentary system begins with a widely dilatable funnel- shaped “ pre-pharynx,” surrounded by the extremely mobile and 62 OSBORN. [Vot. II. thin-walled homologue of the oral sucker. This narrows rapidly posteriorly and leads into a pharynx, oval in outline, and com- posed of muscular tissue and cuticularized on its outer surface. Posterior to this chamber there is, as in Aspidogaster (Stafford, Fig. 1) and Stichocotyle, Nickerson (94), a very short oesopha- gus, whose wall is cuticularized and not glandular, followed immediately by a single, rather large tube, the intestine. This runs down the body to near the posterior end (section No. 88 of Fig. 1), where it ends on the same level as the opening of the Fic. 6. — Transverse section of P. azodontae, serial No. 54, passing through the ovary. The right side of the figure is the left of the animal (camera lucida). excretory system. This tube has a uniform diameter (see Int., Figs. 5, 6), and is lined with tall glandular cells, which are, many of them, peculiarly vacuolated at the outer end (z.¢., next the sur- face of the membrane), as noted by Nickerson in Stichocotyle. The excretory system was only recognized in sections, and, as to its terminal portions, its minor divisions, and their rela- tions to the inter-spaces of the parenchyma, must be studied upon living specimens. The excretory pore is clearly visible in sections and in surface views of total preparations. It is located in section No. 83, and it is a single opening, and not double, as reported for Aspidogaster by Stafford. The sections on which this conclusion rests are not shown in this article, but will be given in connection with my later paper. Two enlarged terminal collecting excretory vesicles are seen meeting beneath the surface pore in horizontal sections. They can be traced forwards on either side, but ultimately they are lost in the par- enchyma. These points are indicated in Fig. 1. Name awa TOM OF AY SPEGIES OF . PEATVASPIS. 63 The nervous system was only imperfectly seen. I did not recognize it at all in the living animals, and it barely shows in the total preparations, but in transverse sections it is clear that there is a large band of fibers crossing the pharynx dorsally —far forward (in section No. 10), and extending down ventrally around the pharynx so as to more than half encircle it. A lateral nerve can be traced posteriorly in a few sections. It shows at Nv. im Pig. 4: Just posterior to this nerve band there are located two sym- metrically disposed organs. They are shown at Oc. in Figs. 1 and 4, where their location is indicated. They are, apparently, invariable structures, very noticeable in living animals, and I have found them in every individual that. I have examined without any exception. They are spherical bodies located in the parenchyma, deeply below the surface and near the ante- rior boundary of the pharynx ; they are posterior and close to the cerebral nerve mass. They are spherical and apparently hollow. The surface is pigmented; the pigment, in the form of minute grains, is clearly visible under the immersion lens ; these grains are, apparently, scattered on all parts of the sur- face of the sphere, but they are much more closely deposited on the inner and upper side. I have not thus far recognized any lens. I have from the first considered them eyes ; their invariable presence, their position in the neighborhood of the cerebral nerve mass, and the presence of pigment demanding this identification. If, however, we accept them as eyes, we must recognize that P. anodontae differs in possessing them from adult trematodes generally. It is well known that eyes are present in early stages of the trematodes, but they are not hitherto recorded of adults, so far as I am able to learn. The accounts of P. lJenoiri do not mention this point, and the illustrations do not shed any light on the question ; so far as can be ascertained from them, these organs are wanting in the African form. There is no room for doubt as to the Chautau- qua animals being adult; the condition of the reproductive sys- tem at large and the presence of eggs and embryos settle that. It is, perhaps, hardly worth while to speculate on the matter now, but I cannot help noting the possible correlation between 64 OSBORN. Versi: the comparatively free life of P. axodontae and the possession of eyes, in contrast with the absence of eyes in the strictly endoparasitic P. /exozrz and its allies, the other genera of this family. The location of the chief organs of the reproductive system agrees closely with that indicated for P. /enxozrz in the figure in Bronn (Pl XX; Fig. 1); ‘and it is also very similar ‘tosthe arrangements found in Aspidogaster. I have not been able to trace all of the windings of the ducts by the section method ; their intricacy has made it impossible to do so; but I feel rea- sonably sure of the identification of the portions which I have introduced into the partially diagrammatic Fig. 1. The sper- mary is single. It is recognized by the presence of small sper- matic cells, but no spermatozoa were recognized in any of the sections. The organ is oval, large, and located about on a level with the hind end of the intestine, and ventrally to it. The nuclear material indicated, possibly, some activity in the tissue, but no mitotic figures were visible. I have thought of two sup- positions by which to account for their absence, vzz., the organs may not have been in a state of activity at the time; and, second, the methods of preservation may not have been adequate. I found in staining that the presence of the cuticle interfered with the action of reagents, and it is quite possible that the germinal cells, if active, got into a state of rest before the rea- gent used in fixation had had time to take effect. The almost invariable presence of the embryos in the vagina seems to indi- cate that the animal is mature and that, consequently, these organs are or have been active. There is a single ovary. It lies on the right side of the body (see Figs. 1 and 6), near the middle, and ventrally to the intestine, and below the diaphragm. The vitellaria are also conspicuous, lying scattered through the ventral portion of the flattened body, near its margin. I have not as yet succeeded in tracing the ducts which connect the different portions of the female reproductive organs. The terminal portions of the reproductive system have been identified with reasonable cer- tainty. The generative opening is visible in the mid-ventral line of section No. 23. This places it in front of the foot, in No.2) AWATOMY OF A SPECIES OF PLATYASPIS. 65 the position indicated in Fig. 2, the same position as that assigned to it in P. denozrt, in Aspidogaster, and in Sticho- cotyle. Two distinct passages lead posteriorly from this com- mon opening, the male and the female ducts. These have not been followed back so as to enable me to base their identifi- cation upon a connection, respectively, with the spermary and ‘ ovary. However, I feel tolerably sure that the one on the right side is the male passage and that on the left the female, as indicated in Fig. 1. The latter contains a small number of oval chitin-enclosed capsules, usually about six, which I am inclined to regard as embryos. They are conspicuous in total preparations, and in sections the chitinous capsule is seen sur- rounding a mass of protoplasmic nucleated cells. These objects are, apparently, identical with similar structures located in the passage leading to the uterus in Poirier’s figure (Pl. XX, Fig. 1). According to that figure the passage is one which leads directly from the ovary, and receives a duct from the yolk gland and vitellaria in its course. The objects in the duct are very differ- ent indeed from the embryos of most flukes, including the innumerable small embryos of the closely allied Aspidogaster ; but their situation and their chitinous covering are so identical with those of the fluke embryos at large that there can be no doubt that these are embryos, but extremely interesting from their unusual size. It is obvious that in Platyaspis we have to do, not with an immense number of small embryos, as in the flukes generally, but with a few large ones. If we accept the view that these objects are embryos, we are then able to identify the passage containing them as the vagina, an identification which locates that organ as it is located in P. lenoiri and Aspidogaster, but not in Stichocotyle, where it is on the right side (Nickerson, ’94, p. 477). I might add that it is some additional evidence in favor of this identifica- tion that the wall of the organ agrees histologically with that of the homologous organ of Aspidogaster. The other of the two passages opening at the genital pore is thus indicated to be the cirrus organ, the terminal portion of the spermiduct. In favor of this view, in addition to the points mentioned in connection with the identification of the other as 66 OSBORN. [Vou. II. the oviduct, is its histological structure, which closely resem- bles that indicated by Stafford for Aspidogaster. SYSTEMATIC POSITION. The question of the systematic position of the Chautauqua Platyaspis does not at present admit of a final answer. There can be no doubt of its generic position. Its anatomy agrees so completely with P. Zexozri in all essential particulars, and is so completely unlike that of the other genera of the Aspidoboth- ridae in all generic points, that it can, I think, be finally stated that it is a species of Platyaspis. The only divergences thus far recognized from Poirier’s spe- cies P. /enotri are in the number of the compartments of the ventral sucker and in the presence of eyes.: As for the first of these, it would be necessary to study the case of the Ameri- can species more fully to determine whether the number of compartments is a constant feature ; so far as is at present known it is constant. And it would be necessary to study the African species as well, to determine whether the account of Poirier is to be regarded as absolutely and exactly true and invariable. If such should prove to be the case, it would fur- nish good grounds for regarding the American form as specific- ally distinct. As for the point about the presence of the eyes in one case and their absence in the other, it is possible that the organs are not functional eyes, but only rudiments, which are more distinct in the American form than in the African. They may be present in the African form, but less distinct, and so may have escaped notice. At all events, it is at pres- ent impossible to decide that the animals are specifically dis- tinct. Still, since they are so widely apart in home and habit, at least so far as our present knowledge of them goes, it appears, on the whole, best to recognize them by distinct names. '72 85 '92 92 ‘94 98 SZ a LOM YVNOl ASSPECTES OF (PEATYVASELIS. 67 LITERATURE. HuxLey. Anatomy of Invertebrated Animals. Hoye. Article Trematoda. Zxcyclofpaedia Britannica. POIRIER. ‘Trematoda, nouv. ou peuconnu. Bz. de la Soc. Philom. Paris, 7 sér., x, pp. 20-42, pls. 1-4. (Quoted from Broun. Original not seen by me.) Braun. Article Trematoda. Bvonn’s Klassen u. Ordnungen. Vol. iv, p. 897. MOonrTIcELLI. Article Cotylaspis. Festschrift 70. Geburtstag R. Leukart’s. NICKERSON. On Stichocotyle nephropis Cunningham, a Parasite of the American Lobster. Zool. Jahrb. 8, pp. 477-480, Taf. 29-31. OsBoRN. Observations on the Parasitism of Anodonta plana Lea. Zool. Bulletin. Vol. i, p. 301. BIOLOGICAL LABORATORY OF HAMLINE UNIVERSITY, March 22, 18908. ae. THE EMBRYOLOGY OF THE APTERYGOTA. AGNES M. CLAYPOLE, Pu.D. Up to within the last few years comparatively little work has been done on the embryology of the lowly insect forms included in Brauer’s group Apterygota, and until we reach the recent date of 1892 no studies of serial sections have been reported. Hence there are among earlier works many misinterpretations of super- ficial features. Much of the early work was done on members of the subdivision Collembola, including the more lowly aptery- gote insects. The first contribution came from the system- atist Nicolet. Unfortunately, but meager data of this article have been obtained. It can merely be stated that some time previous to 1869 Nicolet published studies on Podura aquatica, Desoria cinerea, Cyphodetrus agilis, Sminthurus ornatus, and Orchesella sp(?). He established three facts: that holoblastic cleavage exists among the Collembola; that their eggs are spherical ; that an amnion and serosa are wanting. In 1871 Packard! gave the results of studies on the form Tsotoma walkert, not a thysanuran, but a collembolan. Only the stages after germ-band formation are mentioned. This 1s said to arise as a complete girdle and to show early 6—7 seg- ments. These were identified as antennae 1, mandibles 1, maxillae 1, with possibly a second, thoracic legs 3. Later rudi- ments of a spring appear on the fifth abdominal segment; an unpaired median labrum is also developed. At no time are any tracheae present, and the larvae on leaving the egg resemble the lower Collembola more than the adult Isotoma. A cuticle like that of crustaceans appears during development. In 1875 Oulganine? published studies on Achorites tubercu- latus Nic., Anurophorus fimetarius, and two species of Degeeria. 1 Packard, Jr., A. S., “ Embryological Studies on Diplax, Perithemis, and the Thysanurous Genus [sotoma,” Peabody Acad. Sez NO. Die) LO 7K. 2 Oulganine, W., “ Sur le Développement des Podurelles ” (Extrait du Russe par M. de Korotneff), Arch. de Zool. Exper. Tome iv. 1875. 70 CLA VPOLE. [Vor. II. The eggs all undergo equal holoblastic cleavage, resulting in a uniform single-layered blastoderm soon rendered many layered by rapid growth. The surface becomes crenated and ridged and forms a cuticle, also crenated. The blastoderm becomes smooth, and outlines of the embryo appear. The second em- bryonic layer is said to arise from a definitely placed area lying between the head and tail of the belt-like germ band. A “dorsal organ’ is probably present, though not described as such. Nine pairs of appendages appear: antennae I, mandi- bles 1, maxillae 2, thoracic legs 3, abdominal 2. One pair of those on the abdomen forms the collophore, and the other the spring. Poduridae are found to resemble the lower arthropods - in the following respects : (1) holoblastic cleavage ; (2) absence of amnion; (3) possession of blastodermic cuticles; (4) the formation of the intestine from the middle germ layer. In 1882 Lemoine * added studies on the Collembola Azwro- phorus laricis and Sminthurus plumbeus, two species differing widely from each other in habits and form. The first is small, springless, inactive, and colonial, while the second is large, with a well-developed spring, active, and solitary. The eggs of Anurophorus found in April and May were clear and easy to observe, while those of Sminthurus found in November and December were slow in development and difficult to study. Superficial cleavage, accompanied by secondary yolk cleavage, was true of the former, and the latter showed very unequal holoblastic cleavage with a blastoderm early formed of two layers. ‘Dorsal organs” appear in both, persisting up to the time of hatching. The entoderm is said to arise from two in- pittings, one in front and one behind the “ dorsal organ” and at several other places on the periphery. The germ band, at first forming a belt surrounding the whole circumference of the egg, shows 12-13 segments: 1 cephalic, 3 mandibular, 3 tho- racic, and 5-6 abdominal. Two membranes appear, one of which is clearly connected with the “dorsal organ.” A _ collo- phore develops in both forms, but the spring is rudimentary in Anurophorus, a condition also true of the tracheal system. 5 Lemoine, P., “‘ Recherches sur le Développement des Podurelles,” 45s. Franc. p. advance. des Sct. La Rochelle. 1882. Ne 2, Lae EMBRYOLOGY OF THE APTERYGOTA. 71 Grassi* gives the first mention of thysanuran development. He describes three features in the developmental processes of Japyx. The cleavage is distinctly superficial; an amnion exists, and also a ‘dorsal organ.” Ryder ® published studies on Anurida maritima Guen., giving the following results. After the formation of the germ band two membranes invest the embryo, the inner one being cre- nated. The embryonic area forms a nearly complete belt sur- rounding the egg, and seven pairs of appendages can be made out: antennae 1, mandibles 1, maxillae 1, thoracic legs 3, and the collophore on the first abdominal segment. A rudimentary spring is reported as still visible on the fourth abdominal seg- ment of the hatched larva, but no trace remains in the adult. Wheeler® gives the first account of section views of an apterygote insect. He shows the existence of an intercalary segment with appendages in the head of Anurida maritima, placed between the mandibular and antennal segments. The “dorsal organ”’ is shown in section and is definitely homolo- gized with the “‘indusium” of Xiphidium. Later studies all include internal structure as seen in sections. In 1896 Heymons’ published a short account of his work on Lepisma saccharina, the highest of the Thysanura. The eggs of this species are oval and about 1 mm. in their longest diame- ter; cleavage is distinctly superficial, and an extremely small germ band early appears. This is found to sink immediately into the yolk, still, however, retaining its connection with the serosa, the extraembryonic part of the blastoderm, by a thin membrane, the amnion. As the embryo sinks in, the amniotic cavity becomes large and distinct, but always retains connection with the outside by the open amnion pore. The amnion is hence never constricted from the serosa. By later growth of the germ band the pore is opened and the amnion is retracted 4 Grassi, B., “I Progenitori degli Insetti e dei Miriapodi l’Japyx e la Campo- dea,” Atti accad. Givenia Sci. Nat. in Catania. (3), vol. xix. 1885. 5 Ryder, J. A., “The Embryology of Anurida Maritima Guen.,” Amer. Nat. Vol. xx. 1888. 6 Wheeler, W. M., “ A Contribution to Insect Embryology,” Journ. of Morph. Vol. viii, No. 1. 1893. 7 Heymons, R., “Ein Beitrag zur Entwicklungsgeschichte der Insekten Apter- ygota,” Sitz. Berichte Acad. Wiss. Berlin. 1896. 72 CLAYPOLE. [Vor. II. ) to form a “dorsal organ”’ similar to the structure of that name in the pterygote insect. The author proves two points: (1) that Thysanuran cleavage is superficial, differing from the Col- lembolan type, and (2) that embryonic membranes are formed homologous with those of the Pterygota. Hence Lepisma is an intermediate form transitional between the Collembola and higher insects. This paper is followed in 1897 by a longer and more com- plete study of the same form at the hands of the same worker, Heymons.’ In this he shows that some of the cleavage nuclei migrate from the center to form the blastoderm, while others remain in the yolk as yolk-cells. The gastrula has the form of a circular depression instead of the typical groove, and as soon as a two-layered condition is attained the germ band sinks into the yolk. While buried in the yolk the germ-band segments and paired appendages appear. First antennae, post-oral in position, next distinct intercalary appendages, mandibles, and two pairs of maxillae with a median unpaired labrum. The maxillae early split in two longitudinally, and the maxillary palps remain clearly homologous with the thoracic legs. Paired abdominal appendages appear on each segment except the 11th, getting progressively smaller from the first pair to the Ioth. | After several weeks the larvae hatch and are chiefly distin- guished from the adults by their white color and the absence of the styli and cerci. The reproductive cells appear at an early stage in the hind end of the embryo and are clearly of ectodermic origin; after much migrating they enter the primitive somites and form follicles segmentally arranged in the female. The mesenteron was described as arising from yolk-cells that migrate from the yolk and multiply to form a continuous layer enclosing the yolk and is hence entodermal in origin. In conclusion it is clear that the Thysanura show strongly marked pterygote peculiarities, and the conditions described suggest the author’s opinion that the formation of embryonic envelopes is due to increase of yolk material in the egg. 8 Heymons, R., “ Entwicklungsgeschichtliche Untersuchungen an Lepisma sac- charina L.,” Zeztschr. f. wiss. Zool. Bd. \xii. 1897. NOs2:|' THE EMBRYOLOGY OF THE APTERYGOTA, we At almost the same time Uzel® published a series of articles on the two forms Campodea and Lepisma. His work on the latter practically confirms that of Heymons and may be omitted from this review. He determines that the eggs of Campodea are spherical, about 0.4 mm. in diameter, undergo superficial cleavage, resulting in a blastoderm spread uniformly over the whole surface. No secondary yolk cleavage exists. The germ band arises by migration of cells from all parts of the blasto- derm to form a belt encircling nearly the whole egg. A « dor- sal organ’”’ appears between the head and the tail, but neither an amnion nor a serosa is developed. Paired appendages are early distinguished, also a median unpaired labrum. Of the paired appendages the following are found: one pair of antennae, a pair of distinct intercalary appendages, one pair of mandibles, two pairs of maxillae, three pairs of thoracic feet, and nine pairs of abdominal structures. An interesting point is noted in the permanent retention of the intercalary appendages as lateral folds round the adult mouth. Later the 1—7th abdomi- nal appendages split longitudinally, the outer part forming the permanent styli and the inner the abdominal sacs. _ There is to appear in the new /ournal of Morphology” a paper giving the results of my studies on Anurida maritima. Merely a summary of the most important points will be included in this consideration. It was found that the ovary was ex- tremely simple in form, like that of a myriapod. Each ovum is associated with nutritive cells and the germinal vesicle early disappears.. The egg is spherical, about .27 mm. in diameter, with at first slightly unequal holoblastic cleavage; this is even- tually lost after a large-celled morula stage has been formed, and the blastoderm rises by migration, as in eggs with super- ficial cleavage. It assumes a two-layered condition at once, the entoderm remaining dormant in the yolk. A ‘dorsal organ’’ is formed between the two ends of the belt-like germ band, the latter early showing the usual pairs of appendages 9Uzel, H., “ Vorlaufige Mittheilungen iiber die Entwicklung der Thysanura. Beitrage zur Entwicklungsgeschichte von Campodea staphilinus,” Zool. Anz. INTSeUZC UZ Ose 3G 177% 10 Claypole, A. M., “The Odgenesis and Embryology of Anurida Maritima,” Journ. of Morph. Vol. xiv. 1808. 74 CRAVE OL: [Vou di together with a pair on the intercalary segment which takes part in the formation of the adult mouth. These are homolo- gized with the second pair of Crustacean antennae. Yolk is found enclosed with reproductive cells, causing their very rapid development. Anurida agrees with the rest of the Collembola in showing characters allying it strongly with the lower arthropods. Summing up the present state of knowledge regarding Apterygote embryology, it is found that at least fourteen species of Collembola and three of Thysanura have been studied with more or less care. This work confirms by its results the opin- ion that the Apterygota possess truly primitive characters and also show transitions to the higher and lower Arthropoda. Cleavage among the Collembola, as far as determined, shows many types: equal holoblastic, unequal holoblastic, holoblastic becoming superficial, and truly superficial; while on the other hand the Thysanura show only the superficial type, whether the eggs are spherical or oval. It is unfortunate that in most cases the size of the eggs is not given, and in many instances the method of cleavage is unknown. Still a comparison of the three available forms is instructive. ANURIDA. Spherical, .27 mm. in diameter. Cleavage holoblastic becoming superficial. CAMPODEA. Spherical,.4 mm. Superficial cleavage. LepISMA. Oval, 1.mm. Superficial cleavage. The apparent discrepancy between the two sizes given for the eggs of Anurida by Ryder and myself is readily explained. by the fact that the measurements were taken at different stages. There is a marked increase in size during develop- ment. No early stages are described for Isotoma, whose size is the smallest yet recorded (.15 mm.), but some of Packard’s so-called gerni-band figures suggest strongly that they are pos- sibly stages showing the first cleavage plane appearing. Enough is given in this short series to indicate a regular increase in the size of the egg and a pfarz passu loss of holoblastic cleavage in passing up the scale of apterygote insects. The only trace of the total cleavage remaining in Lepisma is shown in the yolk ee No.2.] THE EMBRVOLOGY OF THE APTERYGOTA. S i ie cleavage; it is markedly significant that nothing of this kind occurs in Anurida after the holoblastic condition is lost, though the yolk is fused into a solid mass and nuclei are scattered through it ; such secondary cleavage is reported in Anurophorus, where egg cleavage is superficial. It is equally clear that the amnion and serosa are absent in the Collembola, the embryonic membranes formed having the nature of “ Blastodermhauten.” In Anurophorus, Achorutes, Degeeria, Sminthurus, and Anurida these membranes show some amount of crenation and hence have powers of expansion. In all cases they are found in connection with the so-called “dorsal organ,” which is a structure clearly homologous throughout the Collembola, being similarly placed and similar in development. A similar structure, similarly placed, is also found in Campodea and Japyx. Heymons considers the “dorsal organ,’ caused by the invaginating cellula envelopes, homologous with these. But the distinct and early appearance of this organ and the simultaneous presence of the amnion and serosa in Japyx and Campodea are clear evidences against such an homology. This is still further confirmed by reference to the structure described by Wheeler as the “indusium”’; this is without doubt, as he states, the homologue of the apterygote “dorsal organ,” and is certainly distinct from the structure that rises later during the elimination of embryonic envelopes. It is possible that a structure similar to the earlier stages of the indusium may exist in Lepisma, but no such specialized later developments would be expected as those found in the Orthoptera. It is also interesting to see the clearness with which certain facts are indicated as to the appearance and fate of certain appendages. The collophore is without doubt a fused pair of abdominal feet ; the spring has a similar origin on the fourth or fifth abdominal segment, and according to Uzel the styli and ventral bladders rise directly from abdominal appendages. There is almost unanimous evidence that an intercalary seg- ment exists in the apterygote head placed between the antennae and the mandibles, disappearing in some cases but remaining to form permanent mouth-parts in others. It is reasonable to 76 CEAV POLE: consider them the homologues of the second pair of Crustacean antennae. In every way the apterygote insect appears to be truly primi- tive ; no evidences of wings appear, and many points in shape of eggs, cleavage, embryonic membranes, and appendages show resemblances to the lower Arthropoda. One question has been purposely left untouched in this brief review : that of gastrula- tion. It is only from studies based on sections that safe con- clusions can be drawn, and the difficulties introduced by the method of germ-layer formation described for Anurida render further facts necessary concerning these processes in other forms before general principles can be safely deduced. WELLESLEY COLLEGE, WELLESLEY, MASS. ‘x AY Rise: wep WebACT Ss CONCERNING THE RELATION— Sairo aND REPRODUCTION OF SOME BERING, SEA TUNICATES. WM. E. RITTER. WHILE President Jordan was engaged, as commissioner in charge of the fur-seal investigations for 1896, in studying the natural history of the seals of the Pribilof Islands, he collected a considerable number of tunicates. These he kindly intrusted to me for study. They proved to be so interesting that during his second summer’s work (1897) in the same capacity he encouraged the enthusiastic young zodlogists, R. E. Snodgrass, A. W. Greeley, and Trevor Kincaid, who accompanied him, to give particular attention to collecting these animals. The result was a large, well-preserved collec- tion, the study of which contributes substantially, in several directions, to our knowledge of the group. These contri- butions will appear in detail as a part of the final report of the scientific investigations made by the commission, to be published later by the United States Government. Some of the facts brought to light are, however, of sufficient con- sequence to make worth while their publication in advance of the report itself. I consequently present them here. As indicated by the title of the note, they relate to the affinities of the Bering Sea tunicate fauna and to the reproduction of some of the species studied. The collection contains eleven species, ten of which are new to science. These are distributed among seven genera in the following way: Boltenia, Styela, Aplidiopsis, and Synoicum, each one species ; Dendrodoa and Polyclinum, each two species ; and Amaroucium, three species. So far as I am able to deter- mine, no tunicates have before now been described from this portion of the world, the northern species hitherto known hav- ing come from the North Atlantic and Arctic oceans, mostly 78 RITTER. [Vor. Il. from the vicinity of the Scandinavian peninsula. The addition of these species to the others already known from far northern seas increases quite to a certainty the probability that there is a distinct Arctic tunicate fauna. The clearest indication of this is afforded by the presence in the collection of the species of the genera Dendrodoa and Synoicum. The single species of the first-mentioned genus hitherto known was described by MacLeay in 1824 from Winter Island (north of British America). Herdman has expressed doubt as to whether or not MacLeay’s genus is really distinct from Sztyela. From the two species now at hand I have convinced myself that the genus is thoroughly valid —much more so than many others that receive general recognition. This, then, appears to be one characteristically Arctic genus. The other genus above mentioned, Syxozcum, seems to be quite as character- istically Arctic. The first species belonging to it was made known by Phipps (1774), and more fully described by Savigny in 1816, and came from Spitzbergen. Since then another species from Lofoten Islands, north coast of Norway, has been: described by ‘Sars. This, then, seems to be another genus characteristically northern. Of the other species the one belonging to the genus Ap/z- diopsis has its nearest ally in A. sarszz Huitfeldt-Kaas, from Lofoten; and two of the three species of Amaroucium appear to be more closely related to A. mutable Sars, from Hamer- fest, Norway, than to any other member of this large genus. The one representative of the genus Lo/tenza I identify as LB. elegans Herdman, from the north Atlantic; so that six of the eleven species may be said to be characteristically far northern, three of them very pronouncedly so, they belonging to genera that are exclusively of this character. The genera Polyclinum and Amaroucium are both cosmo- politan in their distribution; they are almost sure to be represented in any considerable collection of compound ascidians from any part of the world, so that it is only by comparing among themselves the different species in each genus that anything significant as to distribution can be learned. No. 2.] SOME BERING SEA TUNICATES. 79 The facts which I here present relating to reproduction pertain to Synxozcum' alone. They are, in outline, as follows: On examination the colonies are found to contain zooids in various stages of degeneration, as well as those in a normal condition. Some of these degenerating individuals are with- out the thorax; others, again, are lacking both thorax and intestinal loop, the post-abdomen alone being present, this lat- ter, however, retaining quite its normal form and structure. In still other zooids the post-abdomen, which alone remains, is reduced from its original club shape to a spherical form. The post-abdomen, as with all the polyclinidae, lodges the heart, the epicardiac tubes, the sexual organs, and a variable quantity of mesenchymatous tissue, the cells of which contain a characteristic granular material which apparently is food yolk. This last-mentioned substance constitutes, in this species, by far the major portion of the bulk of the post- abdomen at the time when the latter becomes free from the rest of the zooid. The ova at this time appear to be all contained in the compact band-shaped ovary, and are in many stages of growth. They are all, excepting the very largest, almost entirely free of yolk; they possess neither recognizable follicular epithe- lium nor ‘test”’ cells, and ¢hey are distinctly amoeboid in form. Careful examination of the ova discovers that many of them, particularly the larger ones, contain within the substance of the cytoplasm other cells in various stages of disintegration. They are ingesting other cells; they are clearly amoeboid in habit as well as in form. Beside the amoeboid ova contained in the ovary there occur, in some of the post-abdomens that have become more nearly spherical in form, ova in which the amoeboid character is wholly wanting, they being quite spherical in form and regu- lar in outline. In these ova, which are also considerably larger than the largest amoeboid ovarian ova, the cytoplasm is no 1 Synoicum is a compound ascidian in which the colony is composed of a number of lobes arising from a common basal mass. Each of these lobes con- sists of a groundwork, or matrix, of firm, homogeneous testicular substance, in which are imbedded a small number of zooids. 80 MTT Pa: [Vou. II. longer homogeneous and clear, but is filled with granular sub- stance. In some of these last-described ova the nucleus still maintains the large, clear, spherical, vesicular character which it presents throughout the amoeboid period. In others, how- ever, it is indistinguishable. This last condition probably indi- cates the period of maturation. In addition to these several stages of development of the Ovarian ova, numerous stages, from the two-celled stage on- ward, in the development of the embryos have been found. Finally there occur numerous packages of tadpoles, each package containing from ten to sixteen or more individuals, situated in cavities of the semi-cartilaginous test of the colo- nies. These cavities are almost perfectly spherical, are remote from the surface of the colonies, and are entirely closed. They contain nothing but the closely packed tadpoles; and after these have been picked out the firm, smooth walls of the cavities remind one of bullet molds. The tadpoles themselves are enveloped by an unusually thick layer of what in all probability corresponds to the test formed at an early time in the embryonal life of all ascidians. But it contains an unusual number of cells, and in addition bodies of various kinds, which I can account for in no other way than by supposing them to be remnants of the parental zooids which produced the ova. In fact, there is little room for doubt about the nature of some of them.. Thus, in one instance in particular, a small cluster of them resembled the large yolk containing mesen- chyme bodies of the adult zooids so strikingly that I should not have thought of questioning their nature but for the remarkable position in which they occurred. Besides these bodies, pieces of fibers are found which are almost certainly remnants of the muscle fibers of the paternal mantle. In some instances the tadpoles are in an advanced stage of metamorphosis while still contained in the cavities. Mature spermatozoa, as well as others in various stages of development, are abundant in most, if not in all, of the post- abdomens. ‘ ee 3 j No. 2.] SOME BERING SEA TUNICATES. 81 Unfortunately, the collection does not contain sufficient speci- mens of this species to enable me to answer several questions of fact that arise from a consideration of the observations pre- sented. However, the facts that we have scarcely admit of misinterpretation. When the post-abdomen first becomes free from the parent zooid, the ovary contained in it has a large number of ovarian ova in various stages of growth. That some of these mature, become fertilized, and develop into tadpoles is proved by direct observation. When the full tadpole stage is reached, only a very limited number of individuals —ten to sixteen — is present in each cavity, and the cavities contain nothing else than the tadpoles. The ovarian ova are distinctly amoeboid in form and certainly contain ingested cells. The conclusion seems inevitable that by far the larger portion of the ova of each ovary are consumed as food by the few of the same ovary that develop into embryos; furthermore, that the granular material (food yolk) of the parental mesenchyme cells is also made use of as food by the growing embryo, and that probably other tissues of the parent zooid are used, to some extent at least, in the same way. The absence of follicular epithelium and “test” cells from the ovarian ova is undoubtedly correlated with the amoeboid nature of the ova; but it is quite possible that their absence is more apparent than real. Some of the cells ingested by the ova may represent either follicular or “test” cells, or both. The observations also seem to indicate that the test, or ‘“cellu- lose mantle,”’ of the late embryos and tadpoles engulf various portions of the parental zooids, and this suggests that the embryos are in some way nourished by this means. Such a process, however, would be quite remarkable, and further observations on the point are greatly to be desired. UNIVERSITY OF CALIFORNIA, February 10, 1898. an fit etOwmOLOGInS OF THE OCCIPITAL AND FIRST SPINAL NERVES OF AMIA AND PEEROSLS: EDWARD PHELPS ALLIS, Jr. In a recent and extensive work Fiirbringer (No. 2) treats of those nerves of vertebrates that lie between the vagus, or vago accessorius, and the first free spinal nerve. This last nerve, although not definitely so defined by him, is seen, by inference, to be the first nerve posterior to the last one that issues from the cranio-spinal canal either through a foramen in the cranium or by an aperture that lies anterior to a dorsal vertebral arch segmentally related to the cranium. The nerves that lie between this first free spinal one, so defined, and the vagus are all included under the general term spino-occipital, and are subdivided into two groups. The nerves that are assigned to one of these two groups are said to have belonged, with their associated skeletal elements, since an early phyloge- netic period, to the occipital region of the skull, and they are accordingly called the occipital nerves. Those belonging to the other group are said to have acquired their relations to the cranium by a more recent assimilation of their associated skeletal elements, and to be as yet but incompletely emanci- pated from the spinal nerves. As they thus represent an intermediate stage between the nerves of the first group and the free spinal ones they are called the spino-occipital nerves. These three names for the nerves here under consideration will be adhered to in the present article, although I think the adoption of them in the present state of our knowledge of the subject a needless complication, and even a possible source of error or inconvenience. From Fiirbringer’s several special descriptions of these two groups of nerves, and his several general statements regarding them and their associated skeletal elements, it is seen that he considers as occipital nerves all those that issue from the 84 ALLIS. [Vou. II. cranium in that part of it that lies between the posterior limit of the protometameric cranium of Sagemehl’s descriptions (No. 6, p. 526) and the posterior limit of the paleocranium ; and that he considers as spino-occipital nerves all those that issue in that part of the auximetameric skull of Sagemehl’s descriptions that lies posterior to its protometameric portion. These definitions of these nerves seem at first sight to be morphologically concise and definite. A little consideration will, however, show that two suppositions can be made regard- ing the segmental position of the nerves thus incorporated in the skull. They can either lie, morphologically, between the dorsal arches of two adjacent assimilated vertebrae, and so become enclosed between those vertebrae as they fuse with each other and with the skull; or they can lie, morphologi- cally, posterior to the dorsal arches of the assimilated ver- tebrae, become first incorporated in those vertebrae and then with them in the skull. Under the first supposition the last nerve incorporated in the skull would lie anterior to the dorsal arch of the last incorporated vertebra, and anterior to the intermuscular septum related to that vertebra. Under the second supposition it would lie posterior to the same arch and septum. Each nerve of the series would accordingly belong, under the first supposition, to a trunk muscle-segment one anterior to the one it would belong to under the second supposition. From Fiirbringer’s descriptions it is not evident, in each particular case, to which of these two categories the nerves under consideration belong. That, in ganoids and tele- osts at least, he considers them as belonging definitely to the first category is evident from his statement regarding Polyodon. Of this fish he says (No. 2, p. 449): ‘Fir mich bildete das Verhalten an dem untersuchten Exemplare von Polyodon, wo die erste dorsale Wurzel (a) eine durch ein partielles Ligament markirte Stelle des Schadels passirt, das entscheidende Krite- rium. In dieser Stelle erblickte ich die noch nicht vollkom- men verwischte Grenze zwischen dem selachierartigen (proto- metameren) Cranium und der Wirbelsaule, und in der dorsalen Wurzel diejenige des urspriinglichen ersten Spinalnerven, der nun zum ersten occipito-spinalen Nerven (a) geworden ist.” No. 2.] AMIA AND TELEOSTS. 85 It is thus evident, in so far at least as the ganoids and tele- osts are concerned, that the most anterior spino-occipital nerve of Firbringer’s nomenclature must lie, morphologically, between the protometameric part of the cranium and the first posterior assimilated, or partly assimilated, vertebra. The most posterior spino-occipital nerve, where there are more than one, should then lie, necessarily, between the last and the next to the last assimilated vertebrae, and the first free spinal nerve between the last assimilated vertebra and the first free one. These necessary relations of the last-named nerves to the skull and vertebrae, thus definitely indicated by inference, seem not to have been carefully borne in mind by Fiirbringer in his general definitions and conclusions, although it is sufficiently evident that they are of primary importance in any attempt at comparison. In Amia calva, Firbringer found, as I had found independ- ently of him (No. 1), four nerves between the vagus and the first free spinal nerve. The most anterior of these four nerves is said by him to belong to the occipital nerves of his nomen- clature, the other three to the spino-occipital ones. The occip- ital nerve is designated by the letter z, the other three by the letters a, 6,andc. The nerve next posterior to the nerve c is said to be the first free spinal one, and is designated by the number 4. In other fishes, in which there may be other occipital or spino-occipital nerves, not found in Amia, the additional occipital ones are said to always lie anterior to the one occipital nerve of Amia, and the additional spino-occipital ones always posterior to the three spino-occipital ones of that fish. In teleosts, Fiirbringer finds but two spino-occipital nerves, and he considers them as the homologues of the nerves J and c of Amia. On page 465 of his memoir he says, that the occipital nerves are wholly wanting in all teleosts, and that the existence of the first spino-occipital one has not yet been estab- lished in any teleost known to him. On page 543 he further says, that in teleosts, not only all the occipital nerves but also the first spino-occipital nerve is “vollstandig riickgebildet.”’ The nerve next following the nerve c is said to always be, as in Amia, a free spinal one, and it is accordingly designated, as in that fish, by the number 4. 86 ALLIS. [Vor. II. In my work on Amia, already referred to, I fully described all the occipital and first free spinal nerves in that fish, giving at the same time their relations to the anterior muscle-segments of the trunk, and the relations of these segments and their myosepta to the bones of the skull, to the anterior vertebrae, and to the bones of the shoulder girdle. Similar descriptions of these nerves, and of the segments and bones they are related to, form part of a work I have now nearly finished on Scomber scomber. The dissections of this fish have been made under my direction, in my laboratory here at Menton, by Dr. J. Dewitz, and can be briefly summarized as follows : The sixth intermuscular septum is the first one that extends from the mid-dorsal to the mid-ventral line of the body. The ventral parts of the fifth and fourth septa, as seen on the inner surface of the body wall, run downward, from the vertebral column, on to the dorsal edge of a large accessory shoulder- girdle bone, and there end. On the outer surface of the body the fifth septum runs downward and forward to the hind edge of the clavicle, at about the middle of its length, and there ends. The fourth and more anterior septa run downward to and end at the dorsal edge of the same bone. The sixth muscle- segment, the one that lies immediately in front of the sixth septum, is thus the first one that extends ventrally the full length of the clavicle, and the fifth septum is the one that marks the apparent septal position of the ventral end of the clavicle. The fifth septum of Scomber is thus, in its relation to the clav- icle, the apparent homologue of the same septum in Amia. Centrally the fifth septum is attached to the second free vertebra of the fish, the fourth septum being attached to the first vertebra. Articulating with each of these two vertebrae there is, on each :sidé,.assingle-rib, which lies in. theyinter muscular septum attached to the vertebra, at the line where that septum is intersected by the horizontal muscle-septum. On the third and next following vertebrae there are, in addition to these horizontal ribs, ventral ones, which lie along the inner surface of the trunk muscles, in the mesial edges of the septa of the vertebra to which they are related. In one specimen a short rudimentary ventral rib was found on the second vertebra also. No; 2.] AMIA AND TELEOSTS. 87 The second and third intermuscular septa have their central attachments on the occipital part of the skull, the large occipito-superclavicular ligament lying in the third septum, with its outer end in the horizontal line of the outer ends of the horizontal ribs. The line of attachment of the first septum traverses the hind end of the posterior process of the intercalar, and the pedicle and three other processes of the suprascapular are en- veloped in, or lie in definite relations to, different parts of it. The anterior muscle-segments, on each side of the head, extend forward on the dorsal surface of the skull in two deep grooves, the lateral one of which corresponds closely in position to the temporal groove of Amia. This groove lies, however, in Scomber, on the dorsal surface of the parietal and frontal bones instead of, as in Amia, between those bones and the chondrocranium. The anterior margin of the muscle-segments in Scomber extends forward slightly beyond the posterior por- tion of the supraorbital lateral canal, covering externally that canal, while in Amia it only reaches, approximately, the hind edge of the frontal bone. The temporal extensions of the trunk muscles, which are certainly secondary adaptations, thus extend considerably farther forward in Scomber than they do in Atmia. In Amia the first intermuscular septum has the same general relations to the intercalar and to the pedicle of the supra- scapular that the first septum in Scomber has. The fourth and fifth septa have their central attachments to the two occipital arches, and each usually contains one of the two occipito-supraclavicular ligaments of the fish. In one larval fish these two ligaments were found in the third and fourth septa. In Scomber the dorsal and ventral roots of the nerves of the fifth and sixth trunk-segments both traverse foramina that perforate, respectively, the first and second free vertebrae of the fish, the foramina in each vertebra lying posterior to the intermuscular septum that has its attachment to the vertebra. Both nerves have dorsal, ventral, and horizontal branches, and from each nerve a communicating branch is sent dorsally, but 88 ALETS: [Vor Ii morphologically forward to the dorsal branch of the next anterior nerve. The ventral branch of the nerve of the fifth segment sends a large branch forward to join a nerve formed by the fusion of the nerves of the three next anterior segments. From the large nervous trunk, so formed, a branch is sent downward and forward to the sternohyoideus muscle, the remainder of the trunk, as the nervus pterygialis, continuing downward and backward to the pectoral fin. After giving off this anterior branch, the main nerve continues downward and enters the ventral fin, no other branch being sent from it to the pectoral fin. The ventral branch of the nerve of the sixth segment sends an important branch forward to join the nerve of the fifth seg- ment, the branch joining the latter nerve distal to the point where the anterior branch of that nerve is sent forward to join the three next anterior nerves. As this branch thus forms part of a plexus which is evidently the so-called brachial plexus of the fish, it is highly probable that the nerve of the sixth segment takes part in the innervation of the pectoral fin. After giving off this branch the main nerve continues down- ward and enters the ventral fin. The nerve of the seventh segment has no perceptible con- nection with the anterior nerves. It thus, in all probability, takes no part in the formation of the brachial plexus, and consequently no part in the innervation of the pectoral fin. Anterior to the nerve of the fifth muscle-segment, between it and the vagus, there are in Scomber but three nerves. The two posterior ones are represented by both dorsal and ventral roots ; the anterior one by a ventral root only. All of these roots traverse foramina in the occipitale laterale, the foramina lying close together and varying in number from two to five. The one or two foramina of the posterior nerve were always found separate and distinct from those of the two anterior ones, and they lay posterior to, and close to, the third intermuscular septum. Whether the foramina of the two anterior nerves also lay posterior to this septum, or were traversed by it, or lay anterior to it, was not noted. The five roots issue close No. 2.] AMIATAND "LELEOS TS. 89 together, and the common ganglionic mass formed on them lay always posterior to the septum. From this ganglion three dorsal, three ventral, and two hori- zontal branches arise, but as the anterior one of the two latter branches soon separates into two nearly equal parts, there are thus three horizontal branches, in all, associated with the ganglion. The three dorsal and three horizontal branches are distributed to the fourth, third, and second muscle- segments, in a manner similar to that of the corresponding beaches in the posterior segments. The three, ventral branches unite to form a single nerve which, after being joined by a branch of the nerve of the fifth segment, is dis- tributed, as above stated, to the sternohyoideus muscle and to the muscles of the pectoral fin. As the branches of these three segmental nerves all arise from a single ganglion, there were naturally no anterior communicating branches associated with them. There was no indication whatever of a separate nerve related to the first muscle-segment, and no branches of the nerve of the second segment could be traced forward into it. In Amia, the four spino-occipital nerves belong to the second, third, fourth, and fifth muscle-segments, there being in Amia, as in Scomber, no separate nerve related to the first segment. The nerves of the second and third segments issue from the cranium through foramina in the occipitale laterale, the next two issuing through apertures in the membranes that fill the spaces between the cranium and the occipital arches. The first two nerves are represented by ventral roots only, the other two by both dorsal and ventral roots. All four of the nerves take part, as do the nerves of the corresponding segments in Scomber, in the innervation of the sternohyoideus muscle, and a part of the fourth nerve joins the nerve of the sixth muscle-segment to form the nervus pterygialis. Posterior to the nerve of the sixth segment several other nerves enter, independently, the pectoral fin. We thus see that the first six muscle-segments of the trunk of Scomber closely agree, in their relations to the dermal bones of the cranium and shoulder girdle, with the corresponding gO ALLES: [Vor Gk segments of Amia; and that the nerves related to these seg- ments in the two fishes, that is, the first five postvagal nerves, agree even more closely with each other in their general periph- eral distribution. The relations of these several nerves and segments to the skull and vertebrae are, on the contrary, totally different in the two fishes; for the fourth and fifth intermus- cular septa have their respective attachments, in Scomber, to the first and second free vertebrae, while in Amia they have their attachments to the two occipital arches. This marked difference in the two fishes would find an evident and simple explanation in the assumption that the first two free vertebrae of Scomber were, in Amia, partly incorporated in the occipital part of the skull. But this assumption is directly opposed to Fiirbringer’s general conclusions, according to which it must be assumed that the first two free vertebrae in the two fishes are strictly homologous. Under the first assumption there would be, in the two fishes, a marked accord in the nerves and muscle-segments of the region. Under the second assumption there are marked differences to be explained and accounted for. In Scomber, for instance, the fourth muscle-segment lies between the hind end of the skull and the first free vertebra, and it is innervated by the posterior one of the three nerves that issue through the foramina in the occipitale laterale. The next, or fifth, muscle-segment lies between the first and second vertebrae, and is innervated by the first free spinal nerve, the roots of that nerve traversing foramina that lie in the first vertebra close to, but posterior to, the intermuscular septum that has its attachment to that vertebra. In Amia the first free spinal nerve innervates the muscle-segment that lies between the hind end of the skull and the first free vertebra. The homologue, in Scomber, of the first free spinal nerve of Amia is, accordingly, in so far as the morphological relations of the nerves to the skull and vertebrae are concerned, the last spino-occipital nerve, and not the first free spinal one. The insufficiency of Fiirbringer’s definitions is thus at once evident, for an examination of the skull alone in the two fishes would not in any way indicate that the last spino-occipital nerve was not, in each, similarly related to the last occipital vertebra. No. 2.] AMIA AND TELEOSTS. gI But even if this difference in the morphological relations of the nerves to the vertebrae were evident in the skull alone, Scomber would still present a marked exception to Fiirbringer’s general formula ; for, if the most anterior spino-occipital nerve of this fish is considered as nerve 0 of his nomenclature, the most posterior one would necessarily be nerve d, and not nerve 4; and such a nerve is not given, or its existence intimated, in any of the teleosts considered by him. If, on the contrary, the most posterior nerve is to be considered as nerve c, the most anterior one would be nerve a; anerve said by him to be absolutely wanting in all teleosts. The successive incorporation of vertebrae in the occipital part of the skull is attributed by Fiirbringer, primarily (unmit- telbar), to the reduction and disappearance of the myomeres that give to the vertebrae in question their movements relative to each other and to the skull (No. 2, p. 548). This same reduction and subsequent disappearance of the anterior muscle- segments is also said to precede and be the primary cause of the reduction and disappearance of the nerves related to them (No. 2, p. 543). Why, then, is there, in the adult of both Scomber and Amia, an anterior muscle-segment, relatively well developed, without any indication whatever of a separate spinal-like nerve related to it? And why is it that in Amia the last so-called occipital vertebra is incorporated in the skull after the fish has passed the age represented by a 50 mm. specimen, and yet, between the age represented by a 12 mm. larva and the adult fish, there is no related reduction in the number of myotomes? As there are, both in the adult and in larva, four muscle-segments ante- rior to the one that, in the adult, lies between the last assimi- lated vertebra and the next anterior one, some reduction in this number might have naturally been expected. In Acipenser ruthenus, according to Sewertzoff (No. 7, p. 232), there are always, in the adult, two or three spino-occipital nerves anterior to the one that innervates the most anterior myotome. The temporal extensions of the trunk muscles certainly rep- resent to some extent, in Scomber and in Amia, independent invasions of the cranial region, for in Amia these muscies lie g2 ATES. [Vot. II. internal to the parietal bone, while in Scomber they lie exter- nal both to that bone and to the frontal. This seems to indi- cate that Amia and Scomber represent separate lines of descent from some fish in which the trunk muscles had not as yet in- vaded the temporal part of the skull to the extent they have in these two fishes. In Scomber, the muscles extend farther forward than they do in Amia. If, then, there are in Scomber two less anterior myomeres than there are in Amia, and the anterior seg- ments in both fishes are in process of reduction, what is the explanation of this independent and apparently aggressive activ- - ity in the muscles ? Furthermore, aside from the fact that the last spino-occipital nerve perforates the occipitale laterale, I find no indication what- ever, in the skull of Scomber, of the incorporation in it of either of the two occipital vertebrae of Amia; and the simple fact that this nerve is incorporated in the occipital part of the skull is not necessarily any indication whatever, in any fish, of its being a spino-occipital rather than a post-occipital one. My work thus leads me to conclude, not only that the spino- occipital and first free spinal nerves in Scomber and Amia are homologous structures, but also that the first two free vertebrae of Scomber are represented in Amia by the two incompletely incorporated occipital vertebrae. In this my conclusions are directly opposed to those arrived at by both Sagemehl and Fiir- bringer in their comparisons of Amia with other teleosts. Sewertzoff (No. 7, p. 240), in his examination of the skull of Amia, simply confirms Sagemehl’s earlier observations ; that is, he finds three spino-occipital nerves instead of four. In Lepidosteus osseus, he says (No. 7, p. 238) that Balfour and Parker’s investigations show that the myotomes in embryos of that fish extend forward to the ear capsule, exactly as his own investigations show that they do in embryos of Acipenser ruthe- nus. In the adult Lepidosteus, he finds on each side of the head, in addition to the two foramina said to have been previ- ously described by Gegenbaur, a third and more posterior one which “durch eine enge Ritze, wie durch einen Riss mit dem hinteren Rande des Bogens verbunden ist”’ (No. 7, p. 239). The anterior of these three foramina perforates the occipitale NO,-2.] AMIA AND TELEOSTS. 93 laterale, the other two the ‘‘angewachsenen ” occipital arch of the fish. The posterior foramen is said to resemble exactly the foramina found in the dorsal arches of the free vertebrae, and Sewertzoff hence concludes that it unquestionably gives passage to a spinal-like nerve. This nerve is said by him to “belong”’ to the so-called occipital arch of the fish and to indi- cate, with the next anterior nerve, that that arch is formed by the fusion of two dorsal vertebral arches instead of representing but one such arch, as Gegenbaur asserts. The drawing which accompanies Sewertzoff’s descriptions seems to me to show, beyond question, that the nerve here under consideration, and the following spinal ones, each innervate the muscle-segment that lies immediately posterior to the arch the nerve in question perforates. The last nerve that perforates the skull is, accord- ingly, a post-occipital and not a spino-occipital one, exactly as in Scomber ; and as it seems, both from Sewertzoff’s figure and descriptions, to have been but recently, and still incom- pletely, incorporated in the skull, this may account for its appar- ent absence in the specimen described by Gegenbaur. This nerve in Lepidosteus is considered by Sewertzoff as the homo- logue of the last spino-occipital nerve in Amia, and the two ver- tebral arches said to be represented in the single occipital arch of Lepidosteus are accordingly considered as the homologues of the two partly assimilated occipital arches of Amia. If the posterior spino-occipital nerve of Lepidosteus is, as it seems to be, a post-occipital nerve, this comparison is evidently not correct. With Acipenser, so fully described by Sewertzoff, I am un- able to make any comparison, the embryos of Amia that I have as yet investigated not having been sufficiently young to show whether or not a certain number of the anterior postotic somites disappear in this fish without giving origin to permanent muscle- segments. It seems to me, moreover, that there is some con- fusion in Sewertzoff’s descriptions. On pages 224-8 of his memoir he says, that in stage B of Acipenser the first myo- tome posterior to the ear capsule still exists, but is relatively much reduced in size. The first dorsal root in the specimen representing this stage lay opposite the fifth myotome on the 94 ALLIS. (VoL. II. right side of the head, but opposite the sixth myotome on the left side. The first ventral root lay opposite the fourth myo- tome on both sides of the head. In stage 42 the first myotome is said to still exist, and the nerves to have the same relations to the myotomes as in stage B. In stage C the reduction of the anterior myotomes is said to have advanced no further than in the preceding stages. The first two myotomes in this stage are then said to have no related nerves, the third and fourth to have ventral roots related to them, and the fifth to have a complete spinal nerve. Later he says of this same stage, ‘‘ ver- schwunden sind das vordere Myotom (J7/1) des Stadiums £2 und die vordere ventrale Wurzel (sf.d1)’’ and “jetzt ist das vor- dere Paar der dorsalen Wurzeln, (welches gegeniiber den Myo- tomen des 6ten Paares, sf.d2, Fig. 1, dag), von beiden Seiten gleich entwickelt.”’ That he has here in some way changed the numbering of the myotome seems evident, but it is, never- theless, not at all certain, for although he says in one place that the number of myotomes has changed, he says in another that it has not changed. Whether this uncertainty in the num- bering is perpetuated or not in the descriptions of later stages is difficult to judge. So far as the adult is concerned, the defi- nite statement on page 232, that there are no myotomes ante- rior to the post-occipital one, shows a condition totally different from that found in Amia. The post-occipital myotome is said to be innervated, as it is in Amia, by the first free spinal nerve. In the Characinidae, Sagemehl (No. 5, p. 58) found but one spino-occipital nerve, and he considered it as the homologue of the middle one of the three spino-occipital nerves found by him in Amia; that is, as the homologue of nerve 6 of Fiir- bringer’s nomenclature. The first nerve posterior to this nerve is said to lie posterior to the stapes, and to innervate the muscle-segment that lies between the first and second vertebrae. The stapes is said to represent the dorsal arch of the first ver- tebrae, and the claustrum to be the homologue of the posterior occipital arch of Amia. As the nerve 4 in Amia lies anterior to the anterior occipital arch of that fish, there are thus, ac- cording to Sagemehl, two nerves missing in the Characinidae, one of which would be the homologue of the nerve ¢ of Amia, No. 2.] AMIA AND TELEOSTS. 95 and the other the homologue of the post-occipital or first free spinal nerve. The latter nerve, although wanting in the Chara- cinidae, is said to be found in Silurus glanis, and to lie in that fish between the claustrum and stapes. There are thus, according to Sagemehl’s descriptions, ex- actly the same number of spinal, or spinal-like, nerves indi- cated in the occipital part of the skull of the Characinidae as are found in Scomber, and they have exactly the same relations to the vertebral components of the skull. The same is true, according to his descriptions (No. 5, pp. 527, 543), of many other teleosts, among which may be mentioned Esox, Umbra, Perca, the Gadidae, Cyprinodontidae, and Cyprinidae. In the Cyprinidae the nerve c is said to be wanting, as it is in the Characinidae. In the other fishes named, excepting Esox, it is said to be found. Whether it is or is not found in Esox is not stated. Fiirbringer, however, gives it in this fish (No. 2, p. 466). In the Characinidae, Fiirbringer gives nerve 4, differing in this from Sagemehl. He agrees with the latter author as to the absence in these fishes of nerve c. My work would incline me to think that the nerve considered by both these authors as nerve 6 was in reality nerve c, and that nerve J had been missed by both of them in dissection. In Carassius, Sewertzoff says (No. 8, p. 423) there are three dorsal vertebral arches in the occipital part of the skull. In Amia I found (No. 1, p. 727) that the same number of arches were indicated in the region occupied by the cartilaginous occipitale laterale, and that this number of vertebral arches corresponded to the number of muscle-segments. The muscle- segments in Scomber indicate a similar number of vertebral arches in the occipital part of the skull of that fish, Scomber thus agreeing in this with Carassius. In Salmo salar, Harrison (No. 3) gives two persistent occipi- tal muscle-segments, and says that a third and more anterior segment, found in embryos twenty-four days old, disappears entirely after that age. The first persistent segment is said to have no spino-occipital nerve related to it. The second segment is said to be related to the hypoglossus, which nerve in young stages is found ‘von demselben Bau als die iibrigen”’ spinal 96 ALEELS: [Vo. Il. nerves, but in older ones is usually represented by a ventral root only. The third segment lies between the hind end of the skull and the first free vertebra, and the nerve related to it is said to be the first spinal nerve. This nerve, however, leaves the vertebral canal with the hypoglossus, ‘durch eine einzige Oeffnung zwischen dem Occipitale und dem ersten Wirbel.” What this opening may be in or through is not indicated, but the fact that the hypoglossus traverses it warrants the supposi- tion that, in the adult, it must be enclosed in the hind end of the skull. The post-occipital nerve of Salmo thus probably agrees, in this respect, with the corresponding nerve in Scomber. Young larvae of Salmo also agree with Scomber in the number of occipital muscle-segments, but there is, in Salmo, one less spino-occipital nerve than in Scomber. In Necturus, Platt (No. 4) says that the first postotic somite aborts and disappears without giving rise to muscle fibers, and that this is true also for all other vertebrates above the Selachii of which she knows. If it be assumed that Amia agrees in this with Necturus, the nerve of the fifth muscle-segment is seen to be, in both these animals, the anterior nerve of the brachial plexus. The dorsal arch next posterior to this nerve is, in Amia, the posterior occipital arch. In Necturus it is the arch of the third free vertebra. The skull of Amia would thus, under this assumption, contain three vertebrae found free in Necturus. If, on the contrary, it be assumed that the first postotic somite of Amia does not disappear, but gives origin to muscle fibers, the two occipital arches of Amia would correspond to the dorsal arches of the first two free vertebrae of Necturus, as they do to the dorsal arches of the same vertebrae in Scomber. The occipital arch of Necturus would then represent the entire cartilaginous occipitale laterale of Amia, if that structure represents but a single vertebral element, or the posterior one of the three vertebral elements that enter into it, if there are three. Which of these two suppositions, if either, is the correct one can only be known after the investigation of larval stages of Amia earlier than any I have as yet examined. . PALAIS CARNOLES, MENTON, April 30, 1898. Nor2:] AMIA AND TELEOSTS. 97 bo BIBLIOGRAPHY. ALLIS, EDWARD PHELPS, JR. The Cranial Muscles and Cranial and First Spinal Nerves in Amia Calva. /ourn. of Morph. Vol. xii, No. 3. March, 1897. FURBRINGER, MAx. Ueber die Spino-Occipitalen Nerven der Selachier und Holocephalen und ihre vergleichende Morphologie. Festschr. 70. Gebrtstg. C. Gegenbaur. Ba. iii. HARRISON, R. G. Die Entwicklung der unpaaren und _paarigen Flossen der Teleostier. Arch. f. mikr. Anat. Bad. Ixiv, Heft 3. 1895. PLATT, Jutta. The Development of the Cartilaginous Skull and of the Branchial and Hypoglossal Musculature in Necturus. JZorph. Jahrb. Bd. xxv, Heft 3. Dec. 3, 1897. SAGEMEHL, MAx. Beitrage zur vergleichenden Anatomie der Fische. III. Das Cranium der Characiniden, nebst allgemeinen Bemerkungen uber die mit einem Weber’schen Apparat versehenen Physostomen- familien. Morph. Jahrd. Bd. x, Heft 1: 1884. SAGEMEHL, MAx. Beitrage zur vergleichenden Anatomie der Fische. IV. Das Cranium der Cyprinoiden. Morph. Jahrb. Bd. xvii, Hlett)4.. Oct. 23, ror. SEWERTZOFF, A. Die Entwicklung der Occipitalregion der niederen Vertebraten im Zusammenhang mit der Frage tiber die Metamerie des Kopfes. Budi. de la Soc. Imp. des Nat. de Moscou. No. 2. 1895. SEWERTZOFF, A. Beitrag zur Entwicklungsgeschichte des Wirbeltier- schadels. Anat. Anz. Bad. xiii, No. 16. May 22, 1897. Volume LI. December, 1598. Number 3. ZOOLOGICAL BULLETIN. NOTES ON THE FINER STRUCTURE OF THE NERVOUS SYSTEM OF CYNTHIA PARTITA (VERRILIC) GEORGE WILLIAM HUNTER, Jr. In the fall of 1897, while working upon the morphology and finer structure of the nervous system of Cynthia partita (Ver- rill), after noting the papers of Von Lenhosseck ('95), Dehler (95), McClure ('96), and Miss Lewis (96), I was led to look for the centrosome and sphere in the cells of the central ner- vous system. I was directly prompted to this investigation by an examination of the plates of Van Beneden and Julin’s ('84) early paper on the central nervous system of the Ascidians. In Pl. I, Figs. 2 and 3, these authors represent ganglion cells with excentric invaginated nuclei. Careful study showed the same thing to be true for the ganglion cells of Cyxthza. I was, however, not immediately successful in staining the centro- some, although later material killed in more favorable reagents showed that a structure homologous with the centrosome and sphere of authors exists in the tunicate ganglion cell. The incomplete notes on fibrillar structure of the nerve trunks are given in view of the recent papers of Apathy ('97) and Bethe (98). It is hoped in a later paper to give a more complete account of the cell structures and their relation to the neuron. METHODS. Several fixing fluids were employed. They were found to be of extremely varying utility as preservatives of the finer structure of the central nervous system. The fluids of Flem- {00 HUNTER. PVOEssnI. ming, Hermann, Von Rath, and aqueous or alcoholic solutions of corrosive sublimate gave uniformly favorable results. Speci- mens were left in Flemming or Hermann from one to two hours, and in corrosive from one-half an hour to six hours, according to the size of the specimen. The shorter periods gave better results. The method of Von Rath was somewhat modified. Specimens were left in his_ picro-acetic-platini- cosmic mixture from one to four hours, washed six hours in methyl alcohol, twenty-four hours in pyroligneous acid, and several days in weak alcohol, before leaving permanently in 95%. Such specimens, passed through xylol or oil of bergamot, imbedded in paraffin, and cut from two to three mzcra thick, gave the most satisfactory results, especially when stained in Heidenhain’s iron-haematoxylin. Of the other killing fluids used I found Lang’s fluid, Gilson’s mixture, and Perenyi gave the most satisfactory results. Chromic, chrom-acetic, chrom- nitric, and corrosive-acetic mixtures shrink the cell-body badly, giving it a vacuolated and fibrillar appearance. Formalin (ex- cept in very weak solution), picro-formalin, and picric mixtures were of even less value, destroying the cell elements greatly. As stains, Heidenhain’s iron-haematoxylin, with safranin and Biondi-Ehrlich as controls, were employed for general work. The methylen blue-eosin mixture of McClure, and cyanin and erythrosin were used to demonstrate the chromophilous sub- stance in the nerve cell. To demonstrate the structure of the cell prolongations and nerves, thin sections were stained from two to three days in iron-haematoxylin and the stain only partly drawn out. This method gave very favorable results. STRUCTURE OF THE NERVE CELL. The cells of the so-called brain differ greatly in size, the largest being situated most peripherally, the smallest most internally. The largest ganglion cells measure 12 mzcra X 16 micra, and have a nucleus measuring 4 micra X 9 micra. Those of medium size, composing the greatest number of cells in the ganglion, average 7 micra X 14 micra, with nuclei measuring 3 micra X 6 micra. The smallest cells are 3 to 4 mzcra across, No.3.] WERVOUS SYSTEM OF CYNTHIA PARTITA. IO! and 5 to 6 micra in length, with nuclei measuring 3 X 4 mzcra. It can be seen that the nucleus is proportionally largest in the smallest cells, therein frequently taking up a large part of the cell-body. The nuclei of the smallest cells are much richer in chromatic matter than are the nuclei of the larger cells, and may easily be confounded with the so-called neuroglia nuclei. Under the 1-12th oil immersion (Zeiss) the cell appears to have a granulo-fibrillar structure. The granular masses, which stain with haematoxylin and basic analins, are irregular in shape and size, and look in places as if they were made up of smaller granules. They are usually found concentrated in certain regions of the cell, z.¢., the extreme periphery, around the nucleus, and sometimes near the center of the cell surrounding the centrosome and sphere. A regular concentric grouping of these granules was scarcely ever found. In general the larger granules are found near the periphery of the cell. They are frequently found forming a reticulum or arranged in rows. It seems probable that these coarse granules are homologous with the chromophilous substance of the vertebrate nerve cell, as well as with that substance in invertebrates (McClure, Pflige, Lugaro, and others). This is shown by double staining with methylen blue and eosin or erythrosin. Such methods show the cell to be made up of two differently staining elements — a varying number of irregular masses which stain with methy- len blue, and a ground substance finely granulo-fibrillar or homogeneous, which takes the red stain. This ground sub- stance seems to be made up of two portions: a semi-fluid (hyloplasm) and a granulo-fibrillar part. In general the blue- staining substance may be said to be restricted to the more peripheral parts of the cell. The masses vary much in size, small granules as well as large masses being seen; the former existing nearer the center of the cell than the latter. In cells containing an excentric invaginated nucleus the area opposite the inpushing of the nuclear membrane is seen to be made up of very fine granules which stain red. In such cells the chro- mophilous masses were found disposed in the peripheral portion of the cell, around the nucleus. 102 LOM T FR. [VoL. II. Besides the granules first described, small groups of refrac- tive bodies — probably pigment granules —are found in some of the cells. Vacuolated spaces are frequent, but are so much increased in size by poor preservation that I am inclined to believe them artifacts. Such spaces may be filled with the hyloplasm of Nansen, Montgomery, and others. A fibrillar or reticular structure for the nerve cell could not be absolutely proved, although the fibers from the cell process can be followed for some distance into the cell-body. The fre- quent arrangement of the granular portion of the cell into a sort of network suggests a reticular framework of fibers as indicated by Cajal and Van Gehuchten in the vertebrate nerve cell, or Pfliige in invertebrates. The nerve cell is surrounded by a thin membrane and in the large ganglion is surrounded by a capsule of fine fibers (neuroglia of authors, or connective-tissue sheath). This capsule can best be seen in cells that are somewhat shrunken. The nucleus is irregular in contour and appears circular, ovoid, kidney-shaped, or, in ex- treme cases, cup-shaped. Rarely the invagination has appeared to cut the nucleus into two distinct parts. Never has the nucleus been found to occupy a central position in the cell; it is always excentric, and frequently situated in an outpocketing of the cell. In Sylar Seleere: ete ye Nat unipolar cells it is usually found some and sphere; nucleus at axis-cylinder at the opposite end from the cell sudateall Von Rah Testgenaioain. process; but in many cases it is forced by the action of the cen- trosome close to the axis-cylinder end of the cell (see Fig. 1). The nuclear membrane is very prominent and stains deeply with haematoxylin and basic analins. The nuclear process of Schultze (’79), Rhode ('96) was found, but it seemed to be an artifact. Binucleated cells were rarely seen. In large cells the chromatin exists in small particles collected against the nuclear membrane and scattered through the nucleus. No. 3.] WEAKVOUS SYSTEM OF CYNTHIA PARTITA. 103 These chromatin granules are held in place, as is the nucleolus, by a finely fibrous achromatic network. In large cells one nucleolus is always present, rarely two. If two are present, one is larger than the other. The nucleolus is frequently observed to be vacuolated. It is often found suspended in the achromatic network of the nucleus, but just as frequently is it found against the nuclear wall. In deeply invaginated nuclei the nucleolus is found against the nu- clear wall at the bottom of the invagination, as if the wall had been pushed in until it had reached the nucleus. In the smaller cells of old specimens as well as the cells of young animals quite a different state exists. The chromatin gran- ules are more evident, being larger, staining deeply, and apparently more numerous than F'. 2.—Cross-section of periphery of brain of young ‘ Cynthia, showing absence of connective-tissue sheath. in the large cells. The pc., large peripheral ganglion cells; .7., neuroglia nuclei; 7., neuroglia and nerve fibers. Von Rath. Iron- haematoxylin. 5 x oc. 6 (Zeiss). nucleus, as has been noted, is much larger comparatively than in larger cells. The nucleolus is small. There appears at first sight little difference between these nuclei and those of the so-called neuroglia cells, but a closer investigation shows the latter to be more oval and elongated and to have a less prominent nucleolus than the nerve cell (see Fig. 2). In ganglion cells of young specimens killed a few days after metamorphosis the nucleus is very rich in chromatin, and presents much the same aspect as is shown by the smaller ganglion cells of the adult specimens. The nucleus is pro- portionately very large, occupying the greater part of the cell- body. In the larger peripherally placed cells of the young 104 HUNTER. [Vou. II. specimens the nuclei do not occupy proportionately as much space as in the smaller cells, but still much more space than in cells of corresponding size in older specimens. The nucleus is rarely indented, usually ovoid or nearly spherical, is rich in chromatin, and contains a small nucleolus. In general the con- dition is more nearly that of the adult ganglion cell (Fig. 2). THE CELL Process AND NERVE TRUNKS. The cell process is undoubtedly fibrillar (Schultze, Flemming, Pfliige). A decided entrance cone was frequently observed. In other cases the fibrils appeared to enter the cell-body, and spread out in the cortical part of the cell. This was especially noticeable in material killed in Flemming. In rare cases where only one large fiber or bundle of fibrils seemed to enter the cell, it would be traced for some little distance. This may be sim- ilar to the intra- cellularaxis cylin- der of Binet (’94). The structure of the cell proc- ess in the cen- tral system was extremely diffi- cult to make out, but a satisfactory picture could be obtained from young specimens. In such animals the connective-tissue sheath surrounding the central ganglion is not developed, and ganglion cells are frequently found projecting into the loose connective tissue surrounding the ganglion. Indeed they are often near nerve trunks entirely free from the cell mass of the ganglion. Such a cell is shown in Fig. 3. The nerve trunk of which it is a part would be shown in the next section. The fibrillar Fic. 3. — Section parallel and exterior to nerve trunk, showing outlying ganglion cell. Young specimen. g.c., ganglion cell; e., epithelium of peribranchial cavity; 2. leucocyte; n.f., nerve fiber. Von Rath. I[ron- haematoxylin. 5 x oc. 6 (Zeiss). No. 3.] WERVOUS SYSTEM OF CYNTHIA PARTITA. 105 structure of the cell process is here plainly seen, as well as the characteristic wavy course of the fibrils. The fibrils do not form an entrance cone, but seem to spread out in the cell-body, especially toward the periphery. In the axis cylinder the fibrils appear to hold an irregular course, and do not run absolutely parallel. These fibrils are separated from each other by a homogeneous substance which does not stain with haematoxylin, and but slightly with eosin or erythrosin. This is the peri- fibrillar substance, probably the hyloplasm of authors. The structure of the sheath is very difficult to make out; it appears to be almost homogeneous or very finely fibrillar, as described by Apathy and Bethe. No myelin substance could be proved. In the nerve trunks the individual processes can rarely be differentiated in longitudinal section, and then only in very small, loosely con- structed nerves, such as are found in the dorsal lamina. But in cross- section the structure is much easier to make out.) yli- sections): ‘be soaked for twelve hours Fic. 4. — Cross-section of nerve trunk in dorsal lamina, show- in the iron-alum solu- ing fibrilsand sheaths. Adult specimen. 2./, nerve fibrils ; tion, and forty # eight m.c., muscle fibrils in cross-section. Von Rath. Haema- toxylin. Camera drawing. 75 x oc. 6 (Zeiss). hours or more in the haematoxylin solution, then the stain is only slightly drawn out, so that the section looks black, a very good idea of the nerve trunk can be obtained. Fig. 4 shows such a specimen. The nerve trunk is usually free from any connective-tissue envelop and is found lying free in the connective-tissue space (mesenchyme). The sheath of the cell process stains a dull blue or blue black, while the fibrils take a deep black. These fibrils may be united into a little group inside the sheath (when a sheath is present) or scattered indifferently through the perifibrillar substance. Often they are found concentrically arranged just inside the sheath. Sometimes only one or two large fibrils are found. It seems probable that one of these stages merges into 106 HUNTER. [Vot. II. another, as Bethe holds, and not that it is an exhibition of ana- tomica! difference between motor and sensory roots, as Apathy seems to believe. In many cases the sheath does not stain, and the fibrils appear to be loose in a non-staining perifibrillar substance. In such conditions they are usually grouped into small bundles of from two or three to a dozen fibrils. In thick sections the characteristic wavy course of the fibrils, as de- scribed by Apathy, Bethe, and others, can be seen. No heavy connective-tissue sheath is found surrounding the smaller nerve trunks. There is, however, a thin connective-tissue sheath about some of the smaller nerves, which, in the main nerve trunks, becomes quite noticeable, and forms a decided capsular sheath around the ganglion. In the central nervous system the conditions are more diffi- cult to make out. The structure seen could be best explained by the elementary network of Apathy or the anastomosis of Bethe. The sheath appears to be lost, and the interior of the ganglion (neuropile of authors) seems to be made up of fibers of different sizes, crossing and recrossing each other. These fibers are frequently seen to branch or divide dichotomously, but no clear cases of anastomoses have been made out. This is difficult because of the widely different courses taken by fibrils in the ganglion. Intermingled with the nerve fibrils and almost indistinguishable from them are the so-called neuroglia fibers. Neuroglia nuclei are scattered through the ganglion as well as through the nerve trunks. Whether the fibrils just described are homologous with the primitive fibrils of Apathy and Bethe, the author is not prepared to say without further research. Such, however, seems to be the case. One interesting fact with regard to a comparison of my re- sults with those of Nansen, who worked on the nerve tube of Ascidians (see Nansen, Pls. XXI, XXII), might be given here. It was observed that the nerve trunks, as well as the central ganglion fiber mass, when treated with chromic mixtures such as Nansen used, gave results much like those exhibited in his plates. The shrinkage caused by the chromic acid gave the nervous tissue the appearance of a number of tubes. If, however, these so-called tubes were followed to the ganglion Nos3.] WERVOUS SYSTEM OF CYNTHIA PARTITA. 107 cells it was found that not the hollow portion of the tube but its wall seemed to make up the axis cylinder. In specimens killed in Von Rath, Hermann, Flemming, or sublimate, fluids which gave much less distortion and shrinkage, the clear area between the so-called walls of the nerve tubes is seen to be filled with fine fibrils. These fibrils in chromic material are evidently shrunken and lie close to the wall of the “tube.” Indeed some specimens show the fibrils lying against the wall of the “tube.” CENTROSOME AND SPHERE IN THE TUNICATE GANGLION CELL. In nerve cells containing excentric, flattened, or invaginated nuclei, as well as in many cells not showing this nuclear dis- turbance, were found the structures which I have taken homol- ogous with the centrosome and sphere of Von Lenhosseck, Fic. 5. — Centrosome and sphere in ganglion cells (Cyzthia). 1, Von Rath; 2, 3, 5, 6, 8, Flemming; 4, Perenyi; 7, chrom-oxalic. Iron-haematoxylin. Camera drawings. qs x oc. 6 (Zeiss). Dehler, Lewis, McClure, and others. In most general terms the structure can be spoken of as consisting of three parts. Beginning from the outside and going inward we have first : an outer coarsely granular zone —the granular zone of McClure and Lewis. The area of this zone varied greatly (see 108 HUNTER. [Vor. II. Fig. 5). In some specimens it was as much as three-fourths of the cell diameter; in others it was much smaller and less pronounced. It is made up of the coarse granules of the periph- eral part of the cell-body. Next is found a clearly staining area, homogeneous or finely granular, which always contains one, often several, black deep-staining granules, the centrosome or central bodies of authors. This clear area corresponds to the sphere of Von Lenhosseck or the disc of McClure. Such an area may be of considerable size and contain visible radiations which extend out into the surrounding cytoplasm (Miss Lewis), or may be reduced so as to be almost or completely wanting (see Fig. 5). The central bodies are of variable number. One large granule is frequently found ; perhaps two is the most constant number. This last statement seems especially true for young cells. The above-described type of centrosome is often met with, but there are many modifications. In some of the cells of a ganglion may be found a centrosome with well-developed astral rays, presenting the appearance found in leucocytes. In other cells of the same ganglion (see Fig. 5) may be found a centro- some with the typical archoplasmic condensation around it. In still other cells the centrosome may have little or no conden- sation of cytoplasm about it, and may exist as a deep-staining granule in the cell. Again, such a centrosome as last men- tioned may be made up of several granules which seem to be more or less solidly welded together. All these forms or states of centrosomic activity may be present in one or the same section (see Fig. 5). This figure shows that the centrosome structure is not a fixed one, such as Von Lenhosseck pictures, but extremely variable in form, more so than Miss Lewis fig- ures. It seems evident from this that the centrosome, as a dynamic center, is of varying importance in different cells. This is further shown by the differing amount of condensation in other cells, as well as the manner in which the nucleus is indented. In these cells, all described being from adult speci- mens of varying age, we see represented various-sized cells. There seems to be no restriction as to size, although the centrosome is much easier to prove in the larger cells. No.3.) WARVOUS SYSFEM.OF CYNTHIA PARTITA. 109 It would be difficult to give the proportion of cells found which contained centrosomes, as in many specimens after stain- ing no such structure can be proved. The nucleus, although excentric, appears ovoid or circular, and no concentric arrange- ment of the cytoplasm can be observed. In such cells, how- ever, the centrosome may exist as a granule, although no such state has been proved. Several very young animals from two to three mm. long were killed shortly after metamorphosis. In such specimens the ganglion cells, although nearly as large as in the adult, contained nuclei much larger in proportion than those contained in the adult cells. The nuclei were much richer in chromatic material than in the adult. The most striking feature noticed Fic. 6. — Cross-section of brain (Cyzthza) shortly after metamorphosis, showing centrosomes in ganglion cells. 4.c., large peripheral cells; c.¢., connective-tissue cells. Von Rath. Iron-haematoxylin. Camera drawing. 5 x oc. 6 (Zeiss). was the fact that a very large proportion of the cells was found to contain centrosomes, although in most cases the sphere and radiations were lacking. It cannot be positively stated that all ganglion cells at this stage contain centrosomes, but certainly a very large proportion do, as can be seen by a glance at Fig. 6. The centrosome in these cells is usually double, z.e., two cen- tral bodies are found. There seems to be no common axial relation between the direction of the two bodies and the long IIo HUNTER. [Vot. IT. axis of the cell. In general a very small clear area may be said to surround the central bodies, but it is small compared with the same area in cells of older specimens. A very slight condensation is frequently found, but it is also slight as com- pared with older cells. Rarely, if ever, are astral rays found. In some few cases a decided granular condensation of the archo- plasmic type is found. But in the majority of cases the cen- trosome in the young cell differs from the same structure in the old cell, by existing without protoplasmic rays extending from the central body, frequently without any condensation of cytoplasm about it, and often exists as one or a pair of deeply staining granules, situated in the central part of the cell-body. More than all, it differs in the amount of mechanical influence exerted on the cell structures. In the cells of the young Cyz- thia, where the nucleus is proportionately so much larger than in older cells, we would expect a most decided invagination and excentricity. But such is rarely the case. Exceptionally do we find a nucleus with a decided invagination, and flattened nuclei are rare. The nucleus is, however, always excentric. It is round or ovoid in shape, rarely flattened or pushed into an outpocketing of the cell-body, as is observed in older specimens. These facts can only be explained on the supposi- tion that the centrosome does not exert any decided mechani- cal influence on the cell protoplasm, as is seen by the absence of a disc, sphere, or radiations. Indeed in many small cells the centrosome seems pushed to one side by the larger nucleus. The centrosome does not seem to have any fixed position in the cell-body. It was most frequently found between the nucleus and the cell process near the center of the cell. It was also frequently found to lie between the nucleus and that part of the cell most distant from the cell process. It might even lie laterally between the nucleus and the cell membrane. Such positions appeared to be normal. The question of the function of the centrosome is of extreme interest, although with our present data it is very far from being solved. Von Lenhosseck has little to say with regard to its probable function, and, with Dehler, seems to consider it a centrosome once actively functioning in division but left over ol No. 3.] WERVOUS SYSTEM OF CYNTHIA PARTITA. {11 in the resting cell. Miss Lewis believes the structure homolo- gous with the centrosome and sphere in dividing cells. McClure thinks the central bodies and disc found in He/ixr are morphologically equivalent to the centrosomes and sphere com- monly found in other cells. The papers of Buhler and Schaffer I have not seen. It is evident that, at least in certain stages of its existence, the centrosome has a mechanical influence in the cell proto- plasm. As we have seen in young specimens of Cynthia, a condensation of cytoplasm about the central bodies, with the accompanying indentation of the nucleus, is lacking. But in such cells as contained the centrosome, with its sphere and radiations or condensation, a marked mechanical force seems to be excited. This was shown by the excentric position of the nucleus, the flattening or invagination of the nuclear mem- brane on the side toward the sphere, the condensation and concentric arrangement of the cytoplasm about the central body, and the frequently found radiations extending toward the periphery. However, no instances of mitotic division were found. Binucleated nerve cells were seen, and cases where nuclei were so flattened and distorted by the invagination as to be nearly divided into two parts, but in no case anything like mitosis was found. Recent investigation seems. to point to the fact that nerve cells, although they may remain for a long time in a so-called embryonic state, z.¢., as neuroblasts functionally inactive, never divide as adult cells. No cases of mitotic division of nerve cells have been yet placed on record, so far as known to the author. It would seem, then, that another explanation must be found for the presence of the centrosome in the ganglion cell. More recently is advanced the theory that the centrosome may be left over in the cell from its embryonic state to be called forth into activity by seasonal conditions. Such a view was hinted at by Von Lenhosseck in his reference to Meves’s paper 1In Miss Lewis’s second paper, “ Studies on the Central and Peripheral Ner- vous Systems of Two Polychaete Annelids,’ Proc. Amer. Acad. Arts and Sci- ences, vol. xxxiii, No. 14, 1898 (which came too late to be inserted in the biblio- graphical list), she pictures ganglion cells containing two spheres; but she con- cludes that the ganglion cells, after an early embryonic period, never divide. rE? HUNTER. [Vor on the centrosome in the tendon of Achilles of the frog. Meves thinks the centrosome a permanent cell organ which in old cells may not be functionally active. Von Lenhosseck points out that Meves’s observations were made on winter frogs, and thinks that perhaps with the renewal of life activities the cell might divide again. The author has not yet concluded any experiments in this direction, as his material was limited to a killing period of three months, June to September. Such experiments in Cynthia would be difficult, because probably no actual hibernation period takes place, although the life activities may be reduced in winter. One interesting fact might be noted, however; if the central ganglia of several specimens, killed in the same fixing fluids and exposed to same conditions of technique, are sectioned, stained in iron-haematoxylin, and examined, it is found that some specimens show nearly all the cells of the ganglia to con- tain centrosomes and spheres, with accompanying indentation of the nuclei, while other specimens show few if any cells in this condition, and the centrosome, if present, not surrounded with a sphere or radiations. In other words, at a given time of the year (summer), certain ganglion cells in some animals are observed to contain centrosomes, while corresponding cells in other animals seem to lack this structure. It is important to note that the age of the specimens cannot be taken into consideration, and this may be an important factor. It seems to the writer that the centrosome in the ganglion cells must have a meaning other than cell division. Might it not serve the same function as it appears to have served in certain cells possessed of protoplasmic movement, such as leu- cocytes, giant cells of bone marrow, embryonic blood corpuscles, pigment cells, etc. ? In leucocytes it has been shown by Flem- ming (91) and Von Rath (’95) that the centrosome is apparently not engaged in mitotic division, as a sphere and central body are found existing in cells in which the nuclei are divided, seem- ingly by amtosis. In the liver cells of an isopod (Porced/zo) the attractive sphere is not concerned in the division of the cell. Other like cases have been observed. In such cells as the pig- ment cell the centrosome appears to be a dynamic center, caus- ing contraction or expansion (chromatophore of cephalopods). ot No. 3-] WERVOUS SYSTEM OF CYNTHIA PARTITA. I13 It is well known that, in early life at least, the ganglion cell is migratory. Such a cell is shown in Fig. 3. It has wandered out from the ganglion (shown in next section), and is probably on its way to the periphery. This cell is observed to contain two centrosomes. It is worthy of note that in the mammalia the only ganglion cells in which centrosomes are found (so far as known) are those of the spinal and sympathetic ganglia. The cells of the spinal ganglia have probably migrated from the central system; the cells of the sympathetic are proved to have migrated from such a source. In the spinal cord cells border- ing the central canal migrate out into the cord. These cells are the neuroblasts. The above facts seem to prove that the ganglion cell in certain stages of its existence has the power of locomotion. Might not the centrosome preside over the locomotor power of the cell— in the ganglion cell as well as in the leucocytes ? The theories of Englemann and Cajal, in regard to the move- ment and growth of the ultimate ends of the cell process, are interesting from this standpoint. According to these authors, the cell processes are capable of growth and may branch, form- ing more and more complex endings. These endings may at one time be in connection with one set of cells, at another time _ with another set, thus giving many new pathways between different cell groups at different times. Here again is the idea of movement of parts of the cell. Could the centrosome influ- ence such movement ? Would such movement, if it existed, be homologous or analogous to the movement in pigment cells? Such questions cannot yet be answered. This suggestion in regard to the possible function of the centrosome in the ganglion cell must not be taken for fact or theory. It is only suggestion. Much more work must be done and many more facts gathered before such a view could be taken for a theory. But it is hoped that at some future time the problem may be successfully attacked and solved. 114 HUNTER. [Vou. Il. SUMMARY. The principal points treated in this paper are as follows : I. The demonstration of the fibrillar nature of the nerve process as opposed to the “nerve tube”’ of Nansen. Positive proof of the elementary network of Apathy and Bethe is lack- ing. Such a view could, however, best explain the structure of the neuropile of the ganglion (brain) of Cynthia. II. The presence in the nerve cell of ganglion bodies of different size, which color strongly with methylen blue and which are of different chemical structure from the groundwork of the cell. These bodies are undoubtedly homologous with the chromophilous substance in many invertebrates (Pfliige, McClure) as well as in vertebrates. The ground substance of the cell appears granulo-fibrillar. Frequently fibrils may be proved in the cell, especially near the process, and in the periphery of the cell. A cone of entrance was frequently found. III. The existence of the centrosome and sphere in the ganglion cell. This structure was found in adult as well as in young specimens killed a few days after metamorphosis. In the young cell the structure more frequently existed without radiations, and with little or no cytoplasmic condensation about the central body or bodies. The centrosome was proved in a greater proportion of cells in young specimens. In older speci- mens the centrosome and sphere, although not limited to cells of certain size, was proved in fewer cells proportionately. When found it exhibited all possible variations from the central body with little or no cytoplasmic condensation to a decided sphere with cytoplasmic radiations extending almost to the periphery of the cell. In the latter case the nucleus was deeply invaginated and pushed far to one side of the cell, while in cells with little or no radiation and small sphere the nucleus was often ovoid, or only slightly flattened. It is hoped in a later paper to give a more complete account of the fibrillar structure of the nerve trunks, and to throw some light, if possible, on the function of the centrosome in the ganglion cell. HuLL ZOOLOGICAL LABORATORY, May 15, 1898. | , | No. 3.]| WERVOUS SYSTEM OF CYNTHIA PARTITA. I15 BIBLIOGRAPHY. APATHY, S. Das leitende Element des Nervensystems und seine topo- graphischen Beziehungen zu den Zellen. AZ7¢t. aus d. zool. Stat. =u Neapel, Bd. xii. 1897. BETHE, A. Das Central-Nervensystem von Carcinus mznas. ATCA mtkr. Anat. Bd.li. 1898. Binet, A. Contribution a l'étude du Systéme nerveux sous-intestinal des Insects. Journ. d’Anat. et Phys. 1894. CajJAL, R. La fine Structure des Centres Nerveux. Croonian Lect. Proc. Roy. Soc. 1894. DEHLER, A. Der feinere Bau der sympathischen Ganglionzelle. Arch. f- mtkr. Anat. Bd. xlvi. 1895. FLEMMING, W. Ueber Theilung und Kernformen bei Leukocyten und iiber deren Attractionspharen. Arch. f. mikr. Anat. Bd. xxxviii. I8gl. Lewis, M. Centrosome and Sphere in Certain of the Nerve-cells of an Invertebrate. Anat. Anz. Bd. xii. 1896. McCuiure, C. F. W. On the Presence of Centrosomes and Attraction Spheres in the Ganglion Cells of Helix fomata, with Remarks upon the Structure of the Cell Body. Princeton Coll. Bull. Vol. viii. 1896. McC.ure, C. F.W. The finer Structure of the Nerve Cells of Inverte- brates. Zool. Jahrb. 1897. MEvEs, F. Ueber die Zellen des Sesambeins in der Archillessehne des Frosches (Rana temporaria) und iiber ihre Centralkérper. Arch. f. mtkr. Anat. Bd.xlv. 1895. Montcome_ry, F. H. J. Studies on the Elements of the Central Nervous System of the Hetronermertini. Journ. of Morph. Vol. xiii. 1897. NANSEN, F. The Structure and Combination of the Histological Elements of the Central Nervous System. Bergens Museums Aarsberetning. 1886. PFLUGE, M. Zur Kenntniss des feineren Baues der Nervenzellen bei Wirbellosen. Zezz. f. wiss. Zool. Bd. xl. 1895. VAN BENEDEN AND JULIN. Le Systéme nerveux central des Ascidies adultes et ses rapports avec celui des larves urodeles. Arch. d. Biol. 1884. VAN GEHUCHTEN. L’Anatomie fine de la cellule nerveuse. Za Cellule. 1897. Von LENHOsSECK, M. Centrosome und Sphere in den Ganglionzellen des Frosches. Arch. f. mikr..Anat. Bd. xlvi. 1895. Von Ratu, O. Ueber den feineren Bau der Driisenzellen des Kopfes von Anilocra Meditterranea Leach in Specellen und die Amitosenfrage im Allgemeinen. Zezt. f. wiss. Zool. Bd. xl. 1895. SCHULTZE, H. Die fibrillare Structur der Nervenelemente bei Wirbel- losen. Schultze’s Arch. 1879. THE MAXILLARY AND MANDIBULAR BREATH- ING VAEVES OF TELEOST FISHES. ULRIC DAHLGREN, InsTRUCTOR IN HisToLoGy, PRINCETON UNIVERSITY. WHILE watching the living fishes in the aquaria of the United States Fish Commission at Woods Holl, Mass., the writer noticed that the jaws were scarcely moved in breathing, the mouth being kept open all the time, except when used for biting or for yawning, or other acts than breathing. Further, when the fish was facing the observer and when the light was favorable a pair of large and well-developed membranous valves were seen inside the mouth, opening and shutting with a perfect automatic motion as the fish breathed. One of these valves, which were both situated just inside of the teeth, depended from the roof of the oral cavity, while the other arose to meet it from the floor of the oral cavity just in front of the tongue. They were crescentic in shape, widest directly in front, and tapering down laterally to a point just behind the angle of the mouth. Their lines of attachment to the surfaces of the oral cavity were concentric with the teeth. In texture they were semi-transparent and extremely flexible and strong. A few minutes’ observation was sufficient to demonstrate that these structures were valves of great importance in breathing; and an examination was made of the literature on the subject. No mention of such valves appears in the standard works on ichthyology and comparative anatomy, with the exception of Owen,! who says: “The folds of membrane behind the upper and lower jaws, of which ‘internal lips’ the swordfish and dory afford good examples, seem intended to prevent the reflux 1 Owen, Anatomy of Vertebrates, vol. i, p. 413. London. 1866. 118 DAHLGREN. [Vot. II. of the respiratory stream rather than the escape of food from the mouth.” Gunther! makes no mention of the organs in question, and states that “the water used by fishes for respiration is received by the mouth and by an action similar to that of swallowing is driven to the gills and expelled by the gill openings.” I have found several differences between the acts of swallowing and of breathing in the teleost fishes. Wiedersheim 2 states that “fishes breathe by taking in water through the mouth and, by the contraction of the latter, forcing it out again through the gill slits.” The use of the word “through” in the above quotation leads me to infer that the mouth opening is meant, and not the oral cavity. A. B. Macallum? has mentioned these structures in his article on the “ Anatomy of Amiurus,” where he says: “ Behind the pads of teeth and running concentrically with them are folds, one above and one below, arising from a relaxation of the lining membrane; that behind the maxillae is largest, but both may be absent. In one specimen of Amzurus nigricans the fold reached downward and backward into the cavity of the mouth fully one-half inch.” No mention of the function of these folds of membrane is made. These valves have been observed in operation by the writer in over fifty species of fresh-water and marine fishes, and no teleost has been found which does not possess them. Since no accurate description of them and of their function has become part of our recent manuals or text-books, and since, on the other hand, for want of such knowledge it has been impossible to clearly describe the act of breathing in fishes, I take this opportunity of calling attention to these valves and of demon- strating their value as organs of breathing. These valves will first be described as they appear in the common sunfish, Lzpomotis gibbosus (Linn.). (See Fig. 1.) 1 Gunther, Introduction to Study of Fishes, p. 136. Edinburgh. 188o. 2 Wiedersheim (translated by W. N. Parker), Comparative Anatomy of Verte- brates, p. 278. London. 1896. 3 Proc. of Canadian Inst., N.S. vol. ii, No. 3, p. 387. Toronto. ‘ 4 # ; F : IN'O3./] VALVES OF TELEOST FISHES IIg In this fish they are highly developed and efficient. The upper valve is a sheet of membrane hung from the roof of the oral cavity and covered by a continuation of its mucous mem- brane. Its line of attachment is slightly bow-shaped, and lies directly behind the teeth and parallel with them. The valve is broad with a straight lower edge, in the middle of which is a thickened tooth-like projection. This projection Fic. 1. — Maxillary and mandibular breathing valves of sunfish, Exfomotis gibbosus (Linn.). Seen from in front (A), and from behind (4). w+x., maxillary breathing valve ; #7., mandibular breathing valve. is the lower end of a vertical median thickening of the valve. The lower valve is two-thirds as broad as the upper, and in other respects is similar to it. The median thickening is perhaps not so marked, and the valve tapers more at each lateral extremity. Sections show that each valve is a membrane of elastic connective tissue covered with a mucosa. The mucosa possesses a well-developed layer of smooth muscle, while a layer of the same kind of muscle extends from anterior surface to posterior surface at the base or line of attachment of the valve. In death the valves are found lying close to the surface of the 120 DAHLGREN. [Veni oral cavity with their free edges pointing backward, and if the specimen has been hardened they are more easily seen. (See Bigs2>) In fresh dead specimens they are very hard to detect because of their flexible texture and tapering edges, which allow them ee Ventral See ——— SX SSS Fic. 2. — Head of flounder, Paralichthys dentatus (Linn.), seen from left (upper) side. The shaded area represents a vertical median section of the mouth and oral cavity to show the position of the breathing valves in an alcohol hardened specimen. to adhere so closely to the mucous membrane of the oral cavity. Both valves are placed with their edges pointing downward and backward at an angle of less than 45° to the axis of the body. This angle is increased to about 80° when the valves are struck by the regurgitating stream of water at the beginning of expiration. N@:53a] VALVES VOM, TELEOST FLSEES. I2I A number of young black bass, Wicropterus salmoides (Lac.), were carefully kept in aquaria with running water, and when they had become perfectly tame, in some weeks’ time, observa- tions were made on the rate and manner of breathing, with particular reference to the valves under consideration. Each breath requires two acts: one of inspiration and one of expiration. During inspiration the oral cavity is enlarged by moving the opercular frames outward, thus requiring an incoming stream of water to fill the extra space produced. This stream enters the oral cavity at the mouth, which at the beginning of inspiration is open about one-fourth of its normal maximum extent. During inspiration the mouth is opened about ten per cent more, this motion being due only to the connection of the jaws with the opercular frames. The mandibular and maxillary breathing valves are flattened back against the top and bottom of the oral cavity by the entrance of this stream of water. Meanwhile water would enter at the gill openings, which are widening, was it not for the branchiostegal membranes which, although they are attached to the opercular frames, move independently of and contrary ‘to them, closing the entrance automatically by the demon Of-themwater-that tries to enter. (See Fig. 3.) The opposite act of expiration now takes place, the opercular frames moving inward to reduce the space in the oral cavity. The water tries to leave at the mouth, but catching on the edges of the breathing valves and then striking their posterior surfaces it forces them up into such a position that their edges meet and all further progress is stopped. The water then leaves at the gill openings. During expiration the mouth is slightly shut, both this and its opening during inspiration being unavoidably due to the attachment of the jaws to the opercular frames, and not to an effort to retain or let out the water. One fish of this lot was six and one-half inches long. When at rest and free from recent excitement, the number of breaths each minute was forty, with the temperature of the water at 10% ° Centigrade. This rate was very constant, and the half-yawns which the 22 DAHLGREN. [VoEsa: fish occasionally gave did not disturb the rate because they also occurred at regular intervals. The fish was taken out in a scoop-net and held gently in a wet cloth while both valves were Yj N NN \ ae ZL ee, \ . we SNH \ . oF : 5 . avity << Mouth Ocsophagwus Oral C Yy ~~ Fic. 3. — Diagrammatic representation of the pump-like structure of the teleost oral cavity. The anterior portion of each figure is represented in longitudinal vertical section, the posterior portion in longitudinal horizontal section. A, inspiration; 4, expiration. Arrows represent water pressures ; double arrows represent motions of opercular frames. a, branchiostegal valves; 4, maxillary and mandibular breathing valves. cut in their median line from edge to line of attachment, thus practically destroying their usefulness as valves. When returned to the water the fish darted about, but soon settled down and ‘ a 1 NOS] VALVES OF TELEOST FISHES. 123 in twenty minutes had apparently recovered from the effects. The rate was now fifty-nine per minute, and the manner of breathing had entirely changed. The enlargement of the oral cavity was fully a third greater than before (on a rough esti- mate), and at or before the beginning of expiration the mouth was tightly closed with an effort in order, apparently, to prevent the regurgitation of water. In four days this fish was breathing normally again and the valves were apparently repaired, the scar being visible on the edge as a notch, with a white line running from this notch to the line of attachment of the valve. SUMMARY, In the light of the above observations and experiments the act of breathing in the teleost fishes may be described as follows: The respiratory stream enters the oral cavity by the mouth and leaves by the two gill openings, coming in contact with the respiratory surfaces of the gills just before it passes out. It is urged on its course by the pump-like construction and action of the oral cavity and its two sets of valves, an anterior set, which are those under consideration, and a posterior set, the branchiostegal valves. In inspiration the stream enters at the mouth, in response to a dilation of the oral cavity produced by the outward lateral movement of the opercular frames. At the same time water is prevented from entering at the gill openings by the branchiostegal valves which, although they are attached to the opercular frames, move independently of and contrary to them; so that, while this outward movement of the frames extends the gill openings, the branchiostegal valves close them automatically by the action of the water which tries toOlenter. In expiration the water is forced out of the gill openings by a corresponding contraction of the oral cavity. At the same time the water is prevented from regurgitating through the mouth, not by the contraction of the latter, but by the automatic operation of the maxillary and mandibular breathing valves 124 DAHLGREN. which move as accurately and efficiently as any of the heart’s valves. Caught on their posterior edges by the first movement of regurgitation, they snap together and completely prevent any water from leaving the oral cavity by the mouth which, meanwhile, is left partly open, almost as much open as during inspiration. That these valves are of value as breathing organs is evident upon casual observation; that they are of much importance is shown by the compensatory action brought about by injury; that they are not of immediate vital importance is proved by the fish’s ability to get along without their services until they are repaired. Pater OF TEMPERATURE ON THE REGENERATION OF HYDRA. FLORENCE PEEBLES. Ir has been shown that temperature has a marked effect on the regeneration of Planaria torva. Lillie and Knowlton (2) have proved by experiment that the optimum temperature at which regeneration is completed is 29°.7 C.; the minimum, about, 34> iC; During the last month I have made a series of experiments on Hydra viridis and Hydra grisea, in order to test the effect of temperature on the regeneration of the hypostome and tentacles. A transverse cut was made through the reproduc- tive zone of the polyp just posterior to the ring of tentacles. The body thus deprived of hypostome and tentacles was sub- jected to a gradual rise of temperature. The dishes in which the Hydras were placed after the operation were partially sub- merged in a water bath in which the temperature varied from 26° to 32°C. Readings were taken during the day at intervals of four to six hours; the variation was never greater than five degrees. Control experiments were made at room temperature, which ranged between 18° and 24° C. Observations were made at intervals of twelve to twenty-four hours, and regen- eration was considered complete when the new hypostome and tentacles had attained their normal size. Normal Aydras were also placed in the water bath, in order to determine what degree of warmth the uninjured polyps could endure without apparent disturbance. In one series of experi- ments the temperature was raised to 38° C., and as a result not only the injured, but the normal, polyps died. At the end of several hours they had completely disintegrated. That the rise of temperature up to 32° C. produces a marked decrease in the time required for regenerating the lost parts is seen in the following tables. In Table 1 the range of tem- perature and the percentage regenerated at a given time are 126 PEE B OBS (Vor. II. recorded. Forty-six individuals were used in a series of six experiments. TABLE 1 — LivpRA ViRIDIS: Tots Per Cent REGENERATED. 48 hrs. 72 hrs. 26-27° 100% 100% Zino 97% 100% Base 88.9% 88.9% (1 dead) The rate of regeneration in Table 1 may be compared with the rate for Hydra viridis at room temperature, given in Table 2, where the results from forty-five individuals in eight experiments are given. TABLE 2.— HYDRA VIRIDIS. Per Cent REGENERATED. 48 hrs. 72 hrs. TEMPERATURE. 18-24° 37.8% 100% It is readily seen that the number completely regenerated at room temperature in forty-eight hours is much smaller than under higher temperature. Ina recent paper (3) I noted that the rate of regeneration of Hydra viridis is much more rapid than that of Hydra grisea. In connection with this it is interesting to find that when the temperature is raised there is a larger reduction in the time required for regeneration in Hydra grisea than in Hydra viridis. Table 3 is the record of twenty-eight individuals in five experiments. . TABLE 3.— Hypra GRISEA. Per CENT REGENERATED. 48 hrs. 72 hrs. TEMPERATURE. " a No. 3] THE REGENERATION OF HYDRA. 27 At room temperature, 7.¢., 18—24° Cj the regeneration is much slower. In five experiments, in which nineteen polyps were injured, there was no regeneration completed at forty- eight hours; and at the end of seventy-two hours a very small number were complete, as Table 4 shows. TABLE 4.— HyprRa GRISEA. Per CenT REGENERATED. 48 hrs. | 72 hrs. | 96 hrs. 18-24° 0% 26.3% | 94.7% TEMPERATURE. In order to show the great difference in the rate of regen- eration of the two species and the effect of the higher tem- perature, a record, in which several sets of experiments are combined, is given in Table 5. TABLE 5.— Comparison oF H, GrIsEA AND H. VIRIDIS. TEMPERATURE. PER CENT REGENERATED. fT, grisea. 48 hrs. 72 hrs. 96 hrs. 18—24° 0% | 26.3% | 94-7% 26-32° 75% 100% 100% Ff. viridis. 18—24° 37.8% 100% 26-30° = 08.5% 100%, Owing to lack of material, I was unable to try the effect of lower temperatures on //ydra grisca. The results obtained from a series of five experiments on A/ydra viridis show that when the polyps are subjected to cold, regeneration is greatly retarded. Table 6 is the record of thirty-eight individuals at a low temperature where the thermometer was kept at 12° C. TABLE 6.— HyprRA VIRIDIS. Per CENT REGENERATED. 96 hrs. 120 hrs. 130 hrs. 144 hrs. 168 hrs. 13.1% 23.7% 34.2% 71% 100% 128 PEEBLES. The change in appearance of the injured region was so slight from day to day that observations were less frequent than in the experiments where the temperature was higher. It will be seen from Table 6 that by a lowering of the temperature a delay of twenty-four to ninety-six hours results in the process of regeneration. On the other hand, an increase of tempera- ture brings about an increase in rapidity of the rate of regen- eration of twenty-four to forty-eight hours for Hydra grisea, and also a slight increase for Hydra viridis. While making these experiments with temperature, I tried the effect of colored lights upon the regeneration of Hydra. Four colors were tested — red, blue, green, and yellow. These colors were obtained by making solutions of congo red, copper sulphate, anilin green, and potassium bichromate, respectively. These were tested with the spectroscope and found nearly monochromatic. A number of experiments were made and also control experi- ments in darkness and in diffuse daylight, but the process of regeneration was in no way influenced by any of the colors. The experiments noted here were made in the Physiological Laboratory of Bryn Mawr College, and were directed by Dr. J. W. Warren, to whom I wish to express my thanks for sug- gestions and assistance. BryYyN Mawr, PA., June, 1808. REFERENCES. 1. DAVENPORT, C. B. Experimental Morphology. 1897. LILLIE AND KNOWLTON. On the Effect of Temperature on the Development of Animals. Zodlogical Bulletin. Vol. i, No. 4. 1897. 3. PEEBLES, F. Experimental Studies on Hydra. Arch. f. Entw. d. Organismen. 5. Band, 4. Heft. 1897. 4. Wixson, E. B. The Heliotropism of Hydra. Amer. at. Vol. xxv. 1891. N HORDE NOES ON THE. EGG OF ALLOLO-— BOPHORA FOETIDA. KATHARINE FOOT AND ELLA CHURCH STROBELL. PREFACE, In the autumn of 1894, while I was studying living eggs from the ovaries of A/lolobophora foetida with a Zeiss 2 mm. immer. lens., Dr. Whitman suggested that I kill the eggs under this high magnification in order to observe the effect of the fixatives. I made many attempts, but was unable to over- come the technical difficulties sufficiently to make the method of any value. The following spring I experimented with eggs from the cocoons; but found it impossible, without injury to the egg, to hold it steadily in the field while applying the fixative. This summer, by the aid of the Bausch and Lomb compressor, Miss Strobell and I have been able to get more satisfactory results, for it has been possible with this compressor to hold the eggs firmly and yet so gently that they continue to develop normally, forming the polar bodies, etc. Ziegler’s (10) classic work on the living Nematode eggs led me in the spring of 1896 to commence the study of the living (cocoon) eggs of Allolobophora foetida, and at that time I began a comparative study of the living and fixed cytoplasm of these eggs. It was my aim to be able to place side by side illustrations of the living cytoplasm with illustrations of the same stages killed by different fixatives, hoping by that method to support or correct my earlier interpretations.! Since the spring of 1897 Miss Strobell has been associated with me in this work ; and, as our results have been attained 1 T quote the following from a paper sent to press December, 1897, and which will appear in Journ. of Morph., vol. xiv, No. 3, 1898. “As I am at work on a paper which will give the results of a comparative study of the living and fixed cytoplasm in these eggs, I shall omit here any description of the living normal cytoplasm.” 130 FOOT AND STROBEEE, [VoL. II. by our combined efforts, we unite in bearing the responsibility of their publication, — parts of the following paper being written by each of us. Owing to the difficulties of obtaining large numbers of these eggs at definite stages of development, it will require both time and patience to make a comparative study of much value. In the present paper we shall give a brief account of such of our results as we can support with photomicrographs; and I shall also give certain results obtained from a further study of the deutoplasm of these eggs, some data bearing on the formation of the vesicles between the first and second maturation spindles, and a few notes on shrinkage. With orange-methyl green I had differentiated the following structures in the cytoplasm (6), the network and archoplasm (polar-ring substance) selected the orange, while the nucleoli, sperm-granules, centrosomes, and the large and small granules in the spindle, attraction spheres, and cytoplasm selected the methyl green. I did not suspect that many of these granules might be deutoplasmic, for at that time I was fully convinced that xylol or xylol balsam would, in all cases, dissolve out the deutoplasmic granules in a few hours. Further investigation showed me that the time required to dissolve the deutoplasmic granules is very inconstant. In some cases it will require days, again it will require as many weeks, and in a few cases they form an insoluble compound with the stain, and cannot be dissolved at all — this last being true even when the sections have been subjected. to exactly the same technique, in the one case the deutopiasmic granules staining weakly and readily dissolving out, and in the other staining deeply and remaining insoluble. These facts led me to suspect their identity with many of the above-mentioned methyl-green gran- ules, and I am greatly indebted to Miss Strobell for assisting me in my experiments to determine this point. KATHARINE Foot. To test the surmise that some of the granules differentiated with methyl green were identical with the deutoplasmic gran- ules, a number of eggs were submitted to the following technical tests. The eggs were killed in chromo-acetic, as this fixative : , > fy ; eS os et No. 3.] ZTHE EGG OF ALLOLOBOPHORA FOETIDA. 1S shows a definite structure of the cytoplasm and an approxi- mately definite distribution of the granules ; osmic acid was then used, not as a fixative, but as a stain which might select the deutoplasmic granules alone. After a few minutes in 1% osmic, the eggs were hardened, imbedded, sectioned, and mounted in glycerine without further staining. An examination of the sections showed the granules intensely blackened, and as they were apparently the only constit- uents in the cell that responded to the osmic, we have designated them osmophile granules (Photo. ian. Vext-hew f° ~The. nu- cleoli in the pronuclei, and when present throughout the cyto- plasm, were not blackened by the osmic, and the centrosomes, which were occasionally visible in the unstained sections, were Fic. 1.— Section of odcyte, 24 pena nomine aULOWMish-yeMone, With the ere 2 een oe cece granules. Osmophile granules in cytoplasm diaphragm open, the intensely and sphere drawn with camera lucida. Rays black osmophile granules are cp renee most sharply differentiated from the centrosomes and the green- ish refractive nucleoli. Photographs were taken of many of these sections, and numerous camera sketches were made, and a comparison of these with the sketches previously made of the granules that had been differentiated with methyl green (6) showed beyond question the identity of many of the latter with the osmophile granules. The above-mentioned unstained sec- tions were then soaked in warm xylol twenty-four hours and left mounted in xylol balsam until the osmophile granules had so nearly disappeared (at first the xylol merely removes the blackening caused by the osmic) that their presence could be detected only by comparing the sections with photographs which indicated exactly where to look for each granule. The sections were then stained with orange-methyl green, and the faded osmophile granules selected the green and became again distinctly visible. Other sections were stained with acid fuchsin before being mounted in glycerine, and these showed 132 LOOT AND STROBELL. [Von.cil, a striking contrast of red nucleoli and centrosomes, with black osmophile granules. One unfertilized odcyte, first order, contained eleven nucleoli (?) distributed in and near the spindle and throughout the cyto- plasm. The acid fuchsin was subsequently removed from these structures with 70% alcohol, and they could no longer be seen without the aid of sketches to indicate their position, while the black osmophile granules remained as sharply differentiated as before. In unstained preparations (such as described above) we have found these granules at all stages of the development of the egg, from the smallest odcytes (or oogonia) to the seg- menting ova, and in the former, one is occasionally found so exactly in the center of the yolk nucleus that in a stained preparation it would be unhesitatingly pronounced a centro- some. Tiny osmophile granules are often seen within the attraction spheres, spindle, and cones, though they are far more numerous throughout the rest of the cytoplasm. In some cases one, two, or more osmophile granules have been seen apparently in the exact center of a sphere (Text-fig. I). Throughout the rest of the cytoplasm, both their distribution and form vary greatly, even at exactly the same stage of the development of the egg. As to distribution, they are sometimes quite evenly distributed throughout the cytoplasm (Text-fig. I), and again large areas appear to be entirely free from them. As to size, they are sometimes tiny microsomes (Photo. 8 and Text-fig. I), and again many of them are as large and homogeneous as nucleoli (Photo. 12), while in many cases they appear as masses — aggregations of individual granules (Photo. 13). Whether any one of these conditions is distinctive of the normal egg we are unable to determine at present. There are equally marked variations in their response to stains ; after a short immersion, iron haematoxylin removes the blackening caused by osmic fixatives —and, as a rule, it does not stain the granules. After prolonged immersion (three days) the osmophile granules show only a faint indication of the stain, and in these cases, when mounted in xylol balsam, they entirely dissolve out of the sections (cf Photos. 13 and 14). In excep- tional cases, however, they have formed an insoluble compound No. 3.) 7HEZ £GG OF ALLOLOBOPHOKA FOETIDA. 133 with the stain and cannot be dissolved out. Photo. 9 is a section of an ovarian egg, in which these granules stained intensely with iron haematoxylin, all the sections in this slide giving the same reaction. The next slide of sections of the same ovary was treated similarly, and the osmophile granules did not stain. In a few cases we have completed the entire process of killing, sectioning, staining, and mounting, within ten hours; and in these cases the granules have responded sharply to the stain, but we have not repeated this method often enough to test its value. Formation of Vesicles between First and Second Maturation Spindles. — In earlier papers (4—6) one of us has stated that at the telophase of both the first and second maturation spindles, the chromosomes assume the form of small vesicles, corre- sponding in number to the number of the chromosomes. Mead (8) has seen in Chaetopterus similar vesicles at the telophase of the second maturation spindle. He describes the formation of these as follows : ‘“‘ When the chromosomes have reached a position near the poles of the spindle, each of them swells up to form a vesicle, in which, at first, two distinct rows of granules may be seen. Later, each chromosome exactly resembles a miniature nucleus.’’ Whether the chromosomes of the second maturation spindle of Al/olobophora foetida form their vesicles in this way, we are unable to state, as we have not yet secured preparations of the second maturation spindle showing ‘a 7) O the metamorphosis of the chromo- p C es somes to vesicles. Of the first Oo maturation spindle, however, we > Cc vi have preparations showing these ae : Fic. II. — Some forms shown by the chro- Eiansivional stares, sandy they’ SUS= ~~ i 0.3 ukia: dle telophase of. He fet Bese aimceiod Or tonmatiomadiitering ™stvraton spindle, showing the (proba. ble method of formation of the eleven front hat ofthe second maturation — vesicles! which are present a little later, spindle of Chaetopterus. Text-fig. oe ae pope Cas Ii represents a number of forms assumed by the daughter-chromosomes of the tetrads of the first spindle. At the telophase of this spindle (Photo. 11) we find these forms both in the polar body and egg. They 134 FOOT AND STROBELL. [Vou. II. suggest that the vesicles of this stage are formed by the two parts of each chromosome uniting to form a ring. Living Cytoplasm.— The cytoplasm of the living egg of Allolobophora foetida presents a dissimilar structure at different stages of its development, and it has been our aim to demon- strate this difference and to determine, if possible, what structures in the sec- tions can be identified with those seen in the living egg. In an egg at the pro- nuclear and first cleavage stages we have a definite cytoplasmic feature, which is not so pronounced in the less mature egg. This fea- ture is conspicuously shown in Photo. 1 in the form of a beeen globules of vanyens SUaoe Fic. III. — Living unfertilized egg, from perfectly which appear to be an ap- white cocoon. Only the larger osmophile granules represented. Chromosomes distinctly visible. Cam- era, 2 mm. immer., oc. 2. proximately transparent non-miscible substance suggesting drops of sap or oil. We have designated them as sap globules rather than oil globules, as they do not blacken with osmic. In the living oocyte, first order, we have not been able to demonstrate this structure (Text-fig. III), but in odcyte second order we find very tiny sap globules (Text-fig. IV) which gradually increase in size as the egg matures (Photos. 4—5, I); (the globules develop- ing whether the egg is fertilized or not). Fic. IV. — Peripheral sap globules after for- mation of rst polar tively large — the early stages of a pathological body. Traced from ae : photographic nega. CONdition being apparently expressed by a too tive of living egg. In abnormal eggs the sap globules are rela- ae rapid development of the.cytoplasm. The nor- mal enlargement of the globules, as expressed in later stages, cannot be due to individual growth, for the increase in the size of the egg is not at all commensurate with Nos3:) 7HE-£GG OF ALTLOLOBOPHORA FOETIDA. 135 the enlargement of the sap globules. It would seem, rather, that this sap, which in some form must also be present in oocyte first order, is molded into the more definite shapes of the later stages by a rearrangement of . the other constituents of the cytoplasm. If an egg is gently pressed until a tiny break is made in the outer membrane, the larger globules become somewhat constricted in form when flowing through this aperture as represented in Text- fig. V. After escaping from the pres- Fic. V.—Sap globules pressed out : ; of living egg. (Pronuclear stage.) SUme.Or the -aperrure, they regain their “wreehand ketch. ‘To. the night spherical shape. ee ae When an egg at the pronuclear stage _ large globules above is the result has been kept in water too long, the glob- "78 °F fv smaller ones, ules fuse and swell, many of them increasing in size two or three times their diameter, giving the egg a vacuolated appearance. This rapid and definite response to abnormal conditions sug- gested to us their value as a guide to determine the relatively harmful effect of killing fluids, for any fixative producing a marked disturbance of the sap globules or the surrounding substance would probably cause the globules to fuse or break up at once. Our method has been as follows (under a Zeiss 2mm.) : First, to focus on the periphery of an egg showing A A Bike BGs V0: Fic. VII. Fic. VI.—A. Sap globules of living egg. 2B. The same globules after 15 minutes in 1 per cent. osmic. Zeiss 2 mm. immer., oc. 2 (camera). Fic. VII.— 4A. Sap globules of living egg. 2B. The same area after 15 minutes in chromo- acetic. Zeiss 2 mm. immer., oc. 2 (camera). pronounced sap globules, sketching about half a dozen of these with the aid of the camera (Text-fig. VI, A) ; second, one of us then applies the fixative, while the other closely watches the effect produced on the sap globules ; third, after 15 to 136 FOOT AND STROBELL. [Vot. II. 30 minutes, the same globules are again sketched to illus- trate the effect. produced by the fixative (Text-fig. VI, B).. A comparison of Text-figures VI and VII will show the relative injury to the form of the globules produced by 1% osmic acid and chromo-acetic. Photo. 3 further illustrates the effect of chromo-acetic on the sap globules and a comparison of this Photo. with Photos. 2 and 5 will show that osmic acid and corrosive sublimate are far less injurious. In Photo. 2 many of the sap globules in the center, and the line of globules on the right, were sketched from the living egg (camera lucida), and a comparison of the sketch with the photograph shows scarcely perceptible change in structure. Such changes as occur later in corrosive sublimate and osmic preparations are probably pro- duced by the alcohols and imbedding. In order to test this we have fixed, stained, hardened, and cleared eggs under a Zeiss 2 mm. immer., oculars 2 and 4. We have repeated many times each step in the technique, selecting, as in the case when the fixative alone was tested, a definite number of sap globules upon which to center our attention. As, how- ever, the egg becomes more opaque during the process of hard- ening, this method does not promise to be as satisfactory as a complete comparison of sections with freshly fixed material of the same stage. The above-mentioned method of applying the fixative while focusing on the periphery of the egg (under such gentle pressure that the egg continues to develop normally) has been supported by applying the fixative to a thinner layer of cytoplasm, obtained by gently pressing an egg until it is flattened to the outer membrane. The form of the sap globules remains unchanged under this pressure, and their reaction to the fixative appears to be the same as when the egg is not sub- jected to pressure. Whether any part of the sap globules becomes coagulated by the fixatives and takes part in forming the network seen in some sections we are unable to determine. If the globules are present in any form, they do not stain, for the vacuoles seen in the sections (Photos. 6, 16-18) un- doubtedly answer to these structures. How far the breaking or fusing of the sap globules may be responsible for definite features of certain fixatives we are not prepared to answer until No.3.) 7HE EGG OF ALLOLOBOPHORA FOETIDA. 137 we can support our conclusions with a larger number of photo- graphs of sections. In the living egg there appear to be at least four constit- uents of the cytoplasm. 1. The above-mentioned clear, approximately transparent sap globules of varying sizes (Photos. I, 2, 4, 5). 2. Dense, opaque, deutoplasmic granules, varying in size from tiny points, scarcely visible under a magnification of one thousand diameters, to those plainly seen with the low powers. In form, size, and distribution they appear to answer to the above-mentioned osmophile granules (page 130) (Photos. 4, he7-Ovi2;stoeand sbext-tic, 1) “These cranules dance about with great activity in eggs that have been kept too long in artificial media —this abnormal activity being probably due to pathological disturbances in the rest of the cytoplasm. 3. Nucleolar-like bodies, strongly refractive in the living egg — that do not blacken with osmic. In the sections they react to the stains selected by the nucleoli of the pronuclei. The sperm-granules (4 and 6) react to the same stains. 4. Lighter areas which are relatively free from the sap glob- ules and osmophile granules — these areas being represented by the polar rings, spindle, attraction sphere, and the inter- spaces of the sap globules. This substance does not blacken with osmic and appears distinctly granular in fixed eggs. It stains intensely, and in the sections appears more opaque than the rest of the cytoplasm —even in those cases where the osmophile granules are not dissolved or faded out. Thus the lighter areas of the living egg are the darker areas of the sec- tions. (Cf Text-fig. III and Photo. 17.) A study of the living ege suggests no fundamental difference between that part of those lighter areas which contributes to forming the polar rings, spindle, and sphere (archoplasm (§5)), and the part occu- pying the interspaces of the globules. We have been able to demonstrate a difference by differential staining of the sections (5), thus far succeeding only with the double stains. We have, however, additional data on this point arguing strongly for the specific nature of archoplasm. Chromosomes in the Living Egg. — The rarity of the cases in 138 FOOT AND STROBELL. [Vou. II. which we have been able to see chromosomes within the spindle led us at first to think that the living eggs which showed this exceptional feature must be abnormal, for, as a rule, the chro- mosomes are not visible until the egg dies naturally, or is killed by a fixative. It was not until we were able to watch an egg develop normally after the chromosomes were seen, that we were convinced these exceptional cases were due to other causes. This does not appear to be wholly dependent upon the position of the spindle, for often in cases where the spin- dle is in the most favorable position, and clearly indicated, the chromosomes cannot be seen, even with the highest powers. Text-fig. III represents a living oocyte, first order, taken from a freshly deposited cocoon. The chromosomes at the metaphase were so distinct that two of them were readily traced with the camera. We watched this egg until it con- stricted off its first polar body, and then we killed it in chromo- acetic, stained and hardened it, and in every respect it appeared to be anormal odcyte, second order. We would accentuate the fact that the chromosomes in the living egg showed exactly the same form as those seen in sections, as this possibly indicates that the chromosomes are less sensitive than the cytoplasm to the action of the fixatives. The fact that these centers of activity are more staple than the cytoplasm, one of us suggested in a former paper (7), where it was stated that the pronuclei continued to develop long after the cytoplasm was unquestion- ably abnormal. The egg represented in Text-fig. III was below the average in size, thus transmitting relatively more light ; this fact probably accounting, in part, for the relatively distinct outlines of the structures within the egg. The broad, clear rays, which could be traced from the attraction sphere almost to the periphery, do not appear to correspond to the rays so clearly outlined in chromo-acetic sections (Photo. 15), but rather to those of osmic acid sections (Photo. 17). This photograph is technically poor, owing to the fact that it is taken from a section 10m thick. It is introduced only because it is the best example we have (in sections) of a structure that can be compared to the rays of an attraction sphere in the living egg. We are not yet pre- No. 3.] ZHE EGG OF ALLOLOBOPHORA FOETIDA. 139 pared to discuss the finer details of these structures, for we feel we must wait until we can control the shrinkage of the eggs killed in osmic acid, before placing much confidence in the morphological details seen in sections. This shrinkage occurs in the alcohols. When formalin is substituted as a hardener, the shrinkage is much reduced, but the use of formalin prohibits sharp staining. The spinning phenomena, which has been seen by Mrs. Andrews (1) during the formation of the polar bodies in other eggs, we have not yet been able to detect, but in view of the exceptional cases in which we have seen the chromosomes and other details in the living egg of Allolobophora foetida, we are not prepared to say that the above- mentioned spinning phenomenon does not occur. Shrinkage. — A comparison of the size of sections of eggs at a given stage with the size of the average living egg at the same stage shows that, at some point or points in the tech- nique, a large amount of shrinkage has occurred, in some cases amounting to one-half the diameter of the living eggs. With a view to determine when the shrinkage occurs, we have first measured the living egg and then each step in the subsequent technique. In this manner we have tested twenty-eight fixa- tives, the compound fixatives and their component parts, each in varying strengths and varying the time the egg was immersed in the fixative from five minutes to twenty-four hours. An attempt to formulate the data gathered from these experiments has shown that the action of a given fixative upon eggs, even at the same stage of development, is extremely inconstant. But as a general rule, subject to many exceptions, it may be said: First, certain fixatives shrink the living egg, and in these cases relatively little shrinkage is produced by the subsequent treat- ment with the alcohols, e.g., strong chromic acid and, in most cases, corrosive acetic (strong). Second, certain fixatives do not shrink the living egg, and in these cases they shrink more or less during the subsequent treatment with alcohols, e.g., weak osmic acid, .1% to 1%, and corrosive sublimate. Third, certain fixatives swell the living egg, the subsequent treatment with alcohols producing a slight shrinkage —the final result being a mounted egg with almost the same diameter as the 140 FOOT AND. STROBELL. [Vorz- Es living, ¢.g., chromo-acetic,! strong osmic acid, platinum chlo- ride. Fourth, the amount of shrinkage caused by the fixative is dependent upon the stage of development reached by the egg, the unfertilized egg being much more sensitive to the fixative. The hundreds of eggs that we have measured have served merely to impress us with the fact of the inconstant effect of the fixatives and subsequent technique —this inconstancy speaking for the individuality of each egg. As _ shrinkage must be an important factor in determining the final distribu- tion of the cytoplasmic elements, we hope to be able to collect enough data on this point to be of service. PHOTOMICROGRAPHY., Preface. — In the autumn of 1893 and the winter of 1894, my friend Dr. Charles G. Fuller, of Chicago, successfully pho- tographed a full series of my sections of the egg of Adlolo- bophora foetida, illustrating successive steps in the maturation and fertilization of the egg.? The work was done with the Zeiss horizontal photomicro- graphic camera, Zeiss microscope with apochromatic condenser, Zeiss projection oculars 2 and 4, and Zeiss apochromatic lenses 16-2 mm. immer., 140 aperture. Artificial light was used. I am glad of this opportunity to express my indebtedness to Dr. Fuller. The good quality of his work will speak for itself when the photographs are published. These photographg were shown at Woods Holl in the sum- mer and early autumn of 1894, and, as far as 1 am aware, they were the first photomicrographs of sections showing the proc- esses of maturation and fertilization of the egg.— KATHARINE Foor. Our work has been done with a Bausch and Lomb vertt- cal photomicrographic camera, Zeiss microscope, apochromatic 1 In 1896 (5) I regarded chromo-acetic as the most reliable fixative, giving as one reason for this, that eggs measured before killing, and after mounting, gave almost the same diameter. At that time I had not measured the eggs at each step in the ¢echnigue, and the measurements were not extended to sections. — K. Foor. 2 As these photographs have no especial bearing on the details discussed in this preliminary, I shall reserve their publication for a future paper. Nows.| 2HE Z£GG OF ALLOLOBOPHORA FOLTIDA. I4I lenses 16-2 mm. immer., 140 aperture, Zeiss compensating and projection oculars. We have not used a magnification beyond one thousand, finding this will reproduce details that can be clearly seen only with a 2 mm. immer. and ocular 8. With a 2 mm. immer. projection ocular 4, diaphragm of ocu- lar at o, and the longest draw, the magnification attained is nearly one thousand [about 960]. When the diaphragm of projection ocular 4 is adjusted to 10, the magnification is much less ; ¢.g., a draw giving a magnification of 520, with the ocular adjusted at 10, will give a magnification of 670, with adjustment at 0. We tested our magnification by measuring the object with a Zeiss micrometer eye-piece, then taking the measure of the photograph in microns, and dividing the latter by the former. On account of the difference in magnification dependent upon the adjustment of the projection ocular, we found this the only accurate method. Light —clear daylight ; sun shining, but not on mirror. Time—as near noon as convenient. Exposure —as a rule, fifteen to thirty seconds for sections.! It gives us pleasure to express our indebtedness to Prof. Henry Crew, of Northwestern University, for recommending to us the following methods of developing and printing, and for instruction in their use. Plates —Seed 27. Developer — Metol. Printing paper — Kirkland’s Lithium. We wish also to express our obligation to Mr. J. G. Hubbard for our first lessons in developing and printing. The experiments of a year with photomicrography have con- vinced us of its utility as a practical aid in cytological investi- gation, and we hope in this paper to argue for its more general adoption. The impossibility of photographing fine cytological details, which can be readily illustrated by a drawing, has been urged by Flemming (3) and others as the principal argument against the use of the camera in cytological work, Wilson’s atlas (9) and Erlanger’s photographs (2) serving to support these objections. Those who have attempted to photograph cytological details 1 For photography by daylight it is necessary to have a time shutter in the camera. 142 FOOT AND STROBELL. (Vor. Ti. realize the following technical difficulties. At a magnification of one thousand, which is often necessary to bring out these details, not enough light can be transmitted through our thinnest sections to admit of focusing delicate structures on the ground glass of the camera; ¢.g., even with the aid of the best focusing glass the small centrosome shown in Photo. 15 cannot possibly be seen on the ground glass. It is barely pos- sible to see this detail through the microscope with a 2 mm. immer. and ocular 8, and we were astonished to find it so dis- tinctly reproduced in the photograph. The ring chromosomes in Photo. 11 further illustrate this point. It is impossible, however, to focus such details on the ground glass, and it has been our aim to devise some method of over- coming this difficulty by discarding the ground glass as a factor in focusing. Selecting a structure that could be clearly focused on the ground glass at a magnification of one thousand (a sharply stained nucleolus, for example), we first focused through the microscope, making a note of the exact position of the pointer on the face of the micrometer screw. We then slipped the camera down, focusing the nucleolus on the ground glass. The position of the pointer on the micrometer screw then indi- cated exactly the difference between the two foci. In order to facilitate the accurate measurement of this variation in focus we marked off into twenty parts each of the twenty-five divisions that are designated on the face of the micrometer screw. The difference in focus proved to be about ,3, of one of the twenty-five divisions; ¢.g., with the pointer at the 5 mark for the focus through the microscope, to get the focus on the ground glass, turn the screw until the pointer indicates 4 and }f. We have tested the accuracy of this method by photograph- ing at 960 diameters such a detail as the centrosome in Photo. 15. We took five photographs in as many minutes, keeping all the factors unchanged except the focus. One photograph was taken at what we calculated to be the correct focus; 2.€., 3, above the focus through the microscope. Two were taken above this point and two below. Trying this several No. 3.) (7HE EGG OF ALLOLOBOPHORA FOETIDA. 143 times, we found that the variation of ; almost unfailingly caught the desired focus. As two or three photographs can be taken in as many minutes, we generally take three of each preparation, one at the tested focus, one ,!5 above, and one 1, below this point, in case any difference in temperature, thickness of the cedar oil, or any other unforeseen factor should affect the focus. The few minutes required to develop these extra negatives is time well spent, for occasionally the focus above or below the tested one proves the best. This variation in focus can be just as readily settled for any magnification, and it does away entirely with the hopeless effort of attempting to focus fine details on the ground glass. We have tested the method with different magnifications and different lenses, and find it works admirably in all combina- tions, but it seems unnecessary to give figures for the different tests, as the variation is undoubtedly a point that must be settled for each microscope. A method of work that can aid the biologist in seizing accu- rately and rapidly any points of interest his material offers, and enables him to retain them in a form convenient for comparative study, must certainly be of great value in the laboratory. Photomicrography appears to us to fill just such a need. A dozen photographs of a variety of features can be taken in the time required to reproduce any one of them by a careful draw- ing. The printed photographs can be kept in a form service- able for frequent reference, and the impression first made by the preparation not allowed to fade. Possibly a photograph is less intelligible than a simplified sketch to any one unfamiliar with the preparation, but cannot this be said of the preparation itself ? We have collected over two hundred sketches and as many photographs, illustrating features in our sections we wished to preserve for comparative study. Of the relative values of these two methods there can be no question, in every case the photographs proving to be the more valuable aid in recalling the preparations. We are not pleading to replace the sketch with the photograph, but we would argue for the use of both, letting the photograph speak for the preparation and the 144 FOOT AND STROBELE. sketch for the investigator’s interpretation. One of the criti- cisms of Erlanger’s photograph, most commonly heard, empha- sizes the worthlessness of his preparations. Is not this the strongest possible argument in favor of photomicrography ? Woops Hott, October io, 1898. PAPERS) REEE RRED 10: 1. ANDREWS, GWENDOLEN FOULKE. Some Spinning Activities of Pro- toplasm in Starfish and Echinus Eggs. Journ. of Morph. Vol. xii, No. 2. 1897. 2. v. ERLANGER, R. Beitrage zur Kenntniss der Structur des Proto- plasmas, der karyokinetischen Spindel und des Centrosoms. Arch. f. mikr. Anat. Bd. xlix. 1897. 3. FLEMMING, W. Zelle. LZyrgebnisse der Anat. u. Entwicklungsge- schichte, Merkel u. Bonnet. Bd. vi. 1897. 4. Foor, KATHARINE. Preliminary Note on the Maturation and Ferti- lization of the Egg of Allolobophora foetida. Journ. of Morph. Vol. ix, No: 3. 1894. Be Yolk-nucleus and Polar Rings. Journ. of Morph. Vol. xii, Nour. 2896: 6. The Origin of the Cleavage Centrosome. /ourn. of Morph. Vole xii Non. 0607: We The Cocoons and Eggs of Allolobophora foetida. /ourn. of Morph. Vol. xiv, No. 3. 1898. 8. Mrap, A. D. Some Observations on Maturation and Fecundation in Chaetopterus pergamentaceus Cuvier. /Journ. of Morph. Vol. x, Nomr:; 1895. 9g. Wirson, E. B. An Atlas of the Fertilization and Karyokinesis of the Ovum. Macmillan & Co. New York, 1895. Io. ZIEGLER, HEINRICH ERNST. Untersuchungen tiber die ersten Ent- wicklungsvorgange der Nematoden. Zez¢. f. wiss. Zool. Bd. 1x, lett Se rOO5e 146 FOOT AND STROBELL. EXPLANATION OF PLATE A. THE reproductions for the following plates were made by Edward Bierstadt of New York. His work was so admirably done that neither strength nor definition has been sacrificed by the process of reproduction. In order to economize space, only a small portion of the negatives of Nos. I-5, 10, 11, 13-15, has been reproduced. The eggs of Nos. 2-5, 10, were fixed under a Zeiss 2 mm. immer., one of us applying the fixative while the other watched its effects. All the photographs were taken with Zeiss 2 mm. immer. projection ocular 4. PuHoto. 1. Living egg (stage, telophase of first cleavage spindle). Periphery of egg, showing part of a polar ring, surrounded by sap globules. Egg slightly col- ored with weak methyl green. The negative of this egg being a little sharper, it was chosen in place of one of an unstained egg of the same stage. In these cases the light was transmitted through a part of the egg fully too uw thick; thus a very sharp negative could scarcely be expected. X 500. Medium, distilled water. Puoro. 2. Periphery of a segmenting egg, after 20 minutes in corrosive sub- limate. Delafield haematoxylin. Before killing the egg, many of the sap globules were sketched with the camera lucida, including the line of five on the right. The fixative produced no perceptible change in their size or contour. (Cf No. 5.) x 500. Medium, distilled water. PHOTO. 3. Periphery of odcyte, second order, after chromo-acetic (15 mm.). Stained with alum cochineal, hardened in alcohol and cleared in xylol. The small sap globules have fused, forming irregular patches. X 500. Medium, xylol. PuorTo. 4. Periphery of egg just after formation of second polar body. Slightly flattened. Killed in .1 % osmic acid (15 mm.). Gentian violet. X 500. Medium, distilled water. Cf. size of sap globules with those of egg of a little later stage (No. 5). PHOTO. 5. Periphery of egg at pronuclear stage. Pronuclei one-half maximum growth. Egg pressed until the cytoplasm reached the outer membrane. This was done undera 2 mm. immer., and the sap globules appeared neither broken nor fused by the gentle pressure. A comparison of the photograph with a living egg at the same stage shows them to be the normal size. 1% osmic acid, 1 hour. Unstained. X soo. Medium, distilled water. The sharp black specks are osmo- phile granules; those out of focus appear as faint rings. PHoTo. 6. Section (4 “) through cytoplasm and one polar-ring of egg at telophase of first cleavage spindle. Fixative, chromo-acetic. Hardened in 40 %, formaldehyde, 26 hours. Stain, iron hematoxylin. X 340. Medium, xylol balsam. PLATE. A; PHOTOGRAPHS BY FOOT AND STROBELL. ARTOTYPE, E. BIERSTADT. e) 145 HOOT AND SLTROBELL, EXPLANATION OF PLATE B. Puoto. 7. A small piece of the cytoplasm of an odcyte, second order. Un- stained. The egg was killed in 1% osmic, and after one hour crushed on the slide. This was to demonstrate the presence of the osmophile granules to com- pare with those of No.9. X goo. Medium, distilled water. Puoro. 8. Ditto. Changing focus in order to reproduce the three tiny gran- ules, two on the left and one on the right of the preparation. In No. 7 the two on the left (out of focus) appear as tiny faint rings. PHoTO. 9. Section (3 «) through the cytoplasm of an ovarian egg at free end of ovary. Granular aggregations of archoplasm and osmophile granules, such as are found in the o6gonia or tiniest odcytes, and in eggs of later stages, both in the ovary and cocoon. Cf. Nos. 4, 5,7, 8, 10, 12,13. Fixative, chromo-acetic followed by osmic. Hardened in alcohol, 24 hours. Iron haematoxylin. X 860. Medium, xylol balsam. PuHoTo. to. Nucleus in a macromere of a segmenting egg, showing two nucleoli and chromatic thread. Fixed under Zeiss 2 mm. immer., 1 % osmic acid. Dela- field haematoxylin. Exposure 1% minutes, as the light was transmitted through the entire egg. X goo. Medium, distilled water. PHoTo. 11. Section (3 ~) of telophase of first maturation spindle. Cf Text- fig. II. Fixative, chromo-acetic. Hardened in 5% formaldehyde, 48 hours. Iron haematoxylin. X 860. Medium, xylol balsam. Puoro. 12. Section (3 “) of odcyte, second order. Killed in chromo-acetic, washed in water, followed by 1% osmic acid for 30 minutes, to differentiate the osmophile granules. Hardened in alcohol. Unstained. X 500. Medium, glycer- ine. Cf Text-fig. I. Attraction sphere and the refractive sperm indicated, although the egg is unstained. eS PEATE, 2B: rd y ner Saar lane ett hy thes PHOTOGRAPHS BY FOOT AND STROBELL. ARTOTYPE, E, BIERSTADT. 2 se ior m, ay 150 FOOT AND STROBELL. EXPLANATION ©OF BEATE C: PuHoro. 13. Section (4 «) through cytoplasm of odcyte, second order. About one-third of the negative reproduced. Fixative, Hermann. Hardened in 10% formaldehyde, 21 hours. Unstained. %X 870. Medium, glycerine. The photo- graph was taken to demonstrate the two aggregations of osmophile granules, and the smaller ones throughout the cytoplasm. C/ No. 14, which is the same section, after staining in iron haematoxylin 24 hours, and dissolving out the osmo- phile granules with xylol and xylol balsam. A careful comparison of the two photographs will show that they are focused on the same plane, the stained prepa- ration, however, showing no trace of the black granules demonstrated in the unstained section. No. 14 shows that the stain has differentiated certain granules that are not seen at all, or are very faintly indicated in the unstained preparation The section was finally stained deeply with methyl green, with the view of deter mining whether any granules could be differentiated within the vacant places formerly occupied by the black osmophile granules. None could be seen. PHOTO. 14. See No. 13. PHOTO. 15. Section (4 ~) through upper pole of first maturation spindle. A centrosome in the sphere, and equally sharply stained granules beyond the sphere. Fixative, chromo-acetic. Hardened in alcohol. Iron haematoxylin. X g6o. Medium, xylol balsam. PHoro. 16. Section (3 4)! of egg at pronuclear stage, showing a part of one of the pronuclei which have attained about one-half their maximum growth. (After other fixatives, the pronuclei have a very different appearance.) Aggregations of archoplasm (polar-ring substance) at periphery; possibly some of the same sub- stance around the pronucleus. Fixative, .2% osmic. Hardened in 5 % formalde- hyde, 22 hours. Iron haematoxylin, 28 hours. X 540. Medium, xylol balsam. Puoto. 17. A very thick section (10 u) of an oocyte, first order. Lower pole of first maturation spindle, and an indication of the chromosomes which are approach- ing the lower pole (spindle at anaphase). This section was photographed merely because it showed thick rays from the attraction sphere, which strongly suggest those seen in the living egg (represented in Text-fig. III, at a little earlier stage). The sap globules in this egg are larger than is the rule in normal eggs at this stage. No. 18 is the following section photographed to show the row of sap globules completely surrounding the attraction sphere, and to compare this attrac- tion sphere with that of Photograph 15, in which case the egg was killed in chromo-acetic. Fixative, 1% osmic acid. Hardened in alcohol. Iron haema- toxylin. X 500. Medium, xylol balsam. PHOROs-16. See JNo: 17; 1 We have tested the accuracy of our microtome by measuring the diameter of the eggs, both before and after sectioning, and counting the number of sections; e.g., this section is one of a series of thirty, and the largest of these measures 94m in diameter, PHOTOGRAPHS BY FOOT AND STROBELL, 7 c PEATE Cy ARTOTYPE, E. BIERSTADT. Volume LI. February, 1899. Number 4. LOOROCTOA BU isk | Me NOMES=ON- THE OCCIPITAL VREGION ‘OF: TEie TROUT, TRULTTA, BARTO: M.A. WILLCOX, PH.D: PROFESSOR OF ZOOLOGY IN WELLESLEY COLLEGE. Axsourt a year and a half ago I undertook an invest.gation for the purpose of determining the number of segments in the hinder part of the head in the Teleostei. Only when the work was nearly complete did it come to my attention that Harrison, in his paper entitled «‘ Die Entwicklung der unpaaren und paari- gen Flossen der Teleostier,’’ had covered much of my ground. Under these circumstances it seemed unwise to continue the investigation. But as my results embody a few new points, I give herewith a brief summary of them. The material was obtained from the Ziirich fish breeding station and consisted of eggs of the salmon, Sa/mo salar, and the trout, Z7u¢ta fario. When they came into my hands, the salmon eggs had already been developing twenty-nine days in water of 10° Centigrade, and required twenty-eight days in water of about 7°-8°! to hatch. The age of the trout eggs was unknown, but I estimated them to be about four days older ; they required twenty-two days in water of about 7°—8° to hatch. The material was preserved in an aqueous solution of corro- sive sublimate, to which was added 20% of glacial acetic acid ; it was then imbedded in paraffin, sectioned, and stained with haemalum. The greater part of the work was done upon the trout, the salmon being used only for comparison. 1The temperature was not determined at the time, so that this is merely an approximate estimate based upon the observations of others. E52 WILLCOX. [Vot. II. I take as the basis of my description the youngest of the trout embryos. The following data will indicate their approxi- mate age: The length was a little less than 1 cm.; the spinal ganglia were formed, but the anterior ones were still connected by a longitudinal commissure with the vagus; the operculum had covered the first branchial arch. The characteristic struc- tures of a segment at this stage are: (a) a pair of myotomes ; (6) a spinal nerve with a ventral root and a dorsal ganglion which is not as yet directly connected with the spinal cord by a dorsal root ; (c) a portion of condensed mesoblast forming the anlage of the axial skeleton. These structures I will take up successively in order to compare the condition in the most anterior segments with that which obtains farther back. The myotomes are well differentiated and in general similar ; they extend forward to the foramen for the vagus, the first one lying close behind the condensed mesoderm which invests the ear. Striation of the fibers is already present but is not uni- form, being generally more pronounced in the deeper parts. The time at which striation appears seems to be variable; in a second specimen nine days older than this it is barely indicated. The anterior three myotomes lie laterad of the parachor- dals and accordingly represent post-otic cephalic segments; they resemble the posterior ones but lie more laterad, as if pushed out by the enlargement which forms the hind brain. In front of the lateral portion of the most anterior one on the right side is a small triangular mass of tissue in which a few unstriated longitudinally disposed muscular fibers are to be seen, and which undoubtedly represents another nearly atrophied myo- tome, making the number of post-otic cephalic segments four. 1 propose at the earliest opportunity to examine younger embryos in the hope of finding this first myotome better developed. In salmon embryos I have found no trace of it, though I have not made a search exhaustive enough to warrant me in assert- ing that it is absent. In trout embryos one day older than the one just described, it has disappeared ; the succeeding (second) myotome pair also eventually atrophies, though I am unable to say just when. Seven days after hatching it shows no trace of degeneration ; in trout of fourteen days it has entirely dis- NO:4:| 2HE OCCIFITAL REGION OF THE TROCT. 153 appeared. It undoubtedly corresponds to the temporary myo- tome pair found by Harrison in Salmo. The anlage of the axial skeleton consists at this time of con- densed mesoderm aggregated on either side of the chorda, especially along the dorso-lateral and ventro-lateral lines, extend- ing up the neural canal nearly to the top of the spinal ganglia and broadening anteriorly into the parachordals. This anterior broadening begins opposite the fifth myotome. The mesoderm shows no trace of segmentation, except that it is marked at intervals by ridge-like vertical thickenings corresponding to the myosepta. Such ridges are present also on the parachordals opposite the myosepta between segments 2 and 3, 3 and 4, 4 and 5. Cartilage is present in an embryo nine days older; it has the form of paired rods (neural arches) flanking the spinal cord. Each rod lies with its ventral end in a myoseptum, but crosses the myotome obliquely, so that its dorsal end lies in or near the next anterior myoseptum. The foremost rods cross the fifth myotome pair; they are considerably smaller than the others, and are closely connected by condensed mesoderm with the parachordals. They are obviously the anlage of the neural arch (occipitalbogen), which in the Salmonidae fuses with the skull. In specimens of twenty-one days after hatch- ing, no fusion has yet taken place. The parachordals in my youngest embryos are largely chondrified; the cartilage exists as a continuous mass which shows no suggestion of segmenta- tion. According to Stohr, that portion which lies behind the otic capsules (“hintere Parachordalplatten’’) chondrifies on either side from a single center. We come now to the nerves. The two temporary segments are altogether without them, but the first permanent segment has a rudimentary one; the second permanent segment has a typical spinal nerve which differs in no respect from the suc- ceeding ones, and in my oldest embryos (twenty-one days after hatching) shows no sign of degeneration. Harrison states that in salmon somewhat older the dorsal root is atrophied. The rudimentary nerve of the first permanent segment is present in my youngest embryos. It is better developed on the right side, and here has the typical structure of a spinal nerve, differing 154 WILLCOX. from the succeeding ones only in being much smaller. On the left the ventral root is wanting. On each side the dorsal ganglion is connected by a longitudinal commissure anteriorly with the ganglion of the vagus, posteriorly with the spinal ganglion of the next nerve. This rudimentary nerve has dis- appeared entirely, or, at most, has left only a trace in embryos nine days older ; in those which have been hatched fourteen days, the nerve of the second permanent segment innervates also the first segment. This nerve and the one belonging to the next succeeding segment leave the neural canal by the same foramen, namely, the one between the first neural arch and the parachordal. They correspond with those considered by Harrison to be the hypoglossal and the first dorsal. I have not traced their distribution, and am therefore unable to express an opinion on this point. SUMMARY. The cephalic region of the trout consists then of at least four segments. These are represented by more or less perfectly de- Diagram representing the z x x changes which take place Veloped myotomes of which the first two inthe cephalic region of —palrsathophy in the course of development the trout between the fifth = and tenth weeks of devel: “The skeletal anlage shows no trace of seg- 9 ditory or- . 5 % eee oe ee Lmentatione.wkbhe third segment has a rudi- gan; m‘*-m‘, myotomes; ni-n}, spinal nerves; nat- mentary spinal nerve which early atrophies ; na®, neural arches; p, para- 4 i chordal; % position of the ‘fourth has a typical one which three vagus. Shaded or dotted Weeks after hatching shows no sign of de- structures are permanent; those which are left light generation. The condition cannot be better nee: aid represented than by a diagram similar to . that already employed by Sewertzoff. This investigation was carried on in the laboratory of the University of Ziirich, and I gladly take this opportunity of thanking Professor Lang for his kindly aid and interest, as well as for the generous way in which material was placed at my disposal. NOTES ON THE MUSHROOM BODIES OF THE INVERTEBRATES. A PRELIMINARY PAPER ON THE COMPARATIVE STUDY OF THE ARTHROPOD AND ANNELID BRAIN. Cc. H. TURNER. Musnroom bodies, fungiform bodies, pedunculated bodies, are synonyms that have been applied to certain peculiar struc- tures found in the insect brain. These bodies were first dis- covered by Dujardin.!. Although rediscovered by Leydig ? and Rabl-Rueckhard,? yet aided by osmic acid, the microtome and staining fluids, Dietl* was the first to give a complete descrip- tion of the whole organ. Thanks to the researches of more recent investigators, it is now well known that the mushroom bodies occur in all classes of insects, and that they reach their highest development in the //ymenoptera. Dietl,* Berger,® and Viallanes® have found in the Decapod brain structures which they consider homologues of the mushroom bodies of the Hexa- pod brain. Kenyon,’ as the following quotation shows, thinks all of these men are mistaken. ‘Special swellings found on the brains of certain of the Crustacea have been compared to them (the mushroom bodies), but it is seriously doubted, I think, whether such swellings or cellular heaps are properly to 1 Dujardin, “ Mémoire sur le Systeme nerveux des Insects,” Azz. Scz. Vat. Ser. 3, tome xiv, pp. 195 e¢ seg. Pl. IV. 1850. 2 Leydig, Vom Bau des thierischen Korpers. pp. 232 e¢ seg. 1864. 3 Rabl-Rueckhard, “Studien iber Insektengehirne,” Reichert und Du Bots- Raymond's Archiv. f. Anat. pp. 488, 489. 1875. ; 4 Dietl, “Die Organisation des Arthropodengehirns,” Ze7t. f. wiss. Zool. Bd. XXViii, pp. 488-517. 1876. 5 Berger, “ Untersuchungen iiber den Bau des Gehirns und der Retina der Arthropoden,” Ard. Zool. Inst. 2 Wien. Bd. i, Heft 2. 1878. 6 Viallanes, H., ‘Etude Histologique et Organologique sur les Centres Ner- veux et les Organs des Sens des Animaux Articules,” Azz. Scz. Nat. Zool. et Pal. Tome xiv, pp. 405-455, Pls. X, XI. 1893. 7 Kenyon, F. C., “The Brain of the Bee,” Journ. of Comp. Neurology. Vol. Vi, pp. 133-210, Pls. XIV-XXII. 1896. 156 TURNER. [Vot. II. be homologized directly with them. In neither Retzius’ figure of the brain of Astacus fluviatalis, nor in Bethe’s figures of the brain of Carcinus moenus, can I find cells having the relations and the appearance of those I find in the bee. I have noticed Fic. 1. —Section through the mushroom bodies of Cecrofza larva. nothing resembling the structures in Isopods or Amphipods, nor have I found indications of them in the brains of Pauropus, Polyxenus, Scolopendrella, Lithobius, nor even in several forms of Thysanura that I have examined. If cells homologous with those filling the cup-like calyx of the mushroom bodies of the bee are at all present in these forms, they are so undifferen- tiated as to be indistinguishable from the general mass of cells about them.” More recently Hamaker ! has homologized certain groups of cells found in the brain of WVerezs with the Hexapod mushroom Fic. 2.— Section through the brain of Caméarus. bodies. He bases this conclusion upon the following facts: (1) the cells of that type are confined to the brain; (2) they 1 Hamaker, J. J.,.“*The Nervous System of Mereis virens Sars,” Bull. Mus. of Comp. Zool. at Harvard College. Vol. xxxii, No. 6. 1898. No. 4.] BODIES OF THE INVERTEBRATES. 157 are intimately connected with the neuropil; (3) they have small nuclei, and very little cytoplasm ; (4) they are arranged in rows radiating from the neuropil. While now at work upon a comparative study of the Arthro- pod and Annelid brain, several preparations have been exam- ined which throw light upon the distribution of the mushroom bodies. Since it will be some time before these studies can be completed, certain discoveries bearing directly upon the distri- bution of the mushroom bodies are described in this prelim- inary paper. The author does not consider this the place to record the bibliography, nor to discuss the technique, nor to acknowledge Fic. 3. — Section through the brain of Verevs. his indebtedness to those who have in any way aided him in these studies. All such information will be given in the final paper. The mushroom bodies are composed of two factors, cells and fiber-tracts. The cells are minute bodies having small nuclei and almost no cytoplasm. In this respect they resemble Deiter’s corpuscles of the vertebrate brain. Compact masses of these cells crown each stalk. In these nidi the cells are arranged in rows which radiate from the top of each stalk. In each half of the brain the principal fibers of the mushroom bodies are collected in a stalk which lies, more or less erect, in a plane which cuts the longitudinal axis of the brain nearly at right angles. The top of each stalk may be either unbranched or bifurcated or ramosely branched. The lower portion of the stalk usually gives rise to two branches, one passing outwards (laterad) and the other inwards (mesad). 158 TURNER. [Vor. II. The preparations at my disposal make possible the demon- stration of these bodies not only in the Hexapods (Fig. 1), but also in the Decapods (Fig. 2), and in the X7phosura, and in cer- Fic. 4. — Section of Polynoe. tain Polychaeta (Figs. 3-5), etc. In the Polychaeta the stalk seems to be unbranched (Fig. 5), although in Verezs (Fig. 3) and Polynoe (Fig. 4) there is a slight indication of a bifurcation ; in the Hexapods (Fig. 1) and the Decapods (Fig. 2) it is bifur- Fic. 5.— Section of Leszdonotus. cated, while in Zzmulus (Fig. 6) it is ramose. Indeed, in Limulus it is so much branched that it simulates, in structure, the vertebrate cerebellum. At present it is not possible to state whether these mush- No. 4.] BODIES OF THE INVERTEBRATES. I J 59 room bodies occur in all Polychaeta or not ; but it is possible to assert that they exist in Vereis, Lepidonotus, and Polynoe. Nor is it possible to aver their existence in all Crustacea; but it is certain that they occur in Camébarus and Limulus. Both of these questions will be considered at length in the final paper to which is relegated a discussion of several fiber tracts con- nected with the mushroom bodies. It is thought that sufficient facts and figures have been given to demonstrate that the mushroom bodies occur in the Hexapods, Decapods, Azphosura, and certain Polychaeta. In each case these 2 6? eoseseee ~ °aeo < = Fic. 6. — Sagittal section through the brain of Lzmeulus. bodies lie in the front portion of the supra-oesophageal gan- glion. Is it not logical to conclude that the front portions of these brains are homologous? This statement runs counter to the generally accepted view. Goodrich,’ who has recently in- vestigated this question, does not think that the homologue of the Annelid prostomium, with its archicerebrum, occurs in either the insects or the Crustacea. But, since the mushroom bodies constitute the major portion of the protocerebrum of Hexapods and Decapods, and since the mushroom bodies occur also in the archicerebrum of certain Polychaeta, it follows that at least the major portion of the protocerebrum of the insects and the 1 Goodrich, E. S., “On the Relation of the Arthropod Head to the Annelid Prostomium,” Quart. Journ. Micr. Sci. Vol. xl, pt. ii, new series, pp. 247-268, 12 figs. 1897. 160 TURNER. Crustacea is the homologue of the archicerebrum (supra-oesopha- geal ganglion) of the polychaete annelids. CLARK UNIVERSITY, SOUTH ATLANTA, GA. September 14, 1808. REFERENCE, LETTERS. C.M., cells of the mushroom bodies. S.A7., stalk of the mushroom bodies. All figures were drawn with a camera. Figures 2 and 6 are drawn to the same scale. Figures 3, 4, and 5 are drawn to the same scale, but are enlarged more than 2 and 6. Figure rt is enlarged more than any of the others. NOTES ON NORTH-AMERICAN EARTHWORMS OFS tHE GENUS” DIPLOCARDE GUSTAV EISEN, PH.D. WirtH the discovery of a species of Dzplocardia possessing the spermiducal pores in xx, it becomes advisable to include in this genus my genus, A/eodrilus. This genus was established some years ago for a species from Baja, California, Aleodrilus Keyesz, and based on the position of the spermiducal pores in xxi instead of in xix, as in Dzplocardia communis, the only species known at that time. Later finds of new species of this genus show that Dzplocardia verrucosa Ude stands inter- mediate between the two first-mentioned species, possessing the prostate pores in xx. The location of the spermiducal pores in Olzgochaeta is gen- erally considered of generic importance, and it is very rare that we find variations in this respect in the same genus. But this character ought to be accompanied by others in order to serve as a genus characteristic, provided the distance between the respective somites is not very great. If we except the absence of penial setae in A/eodrilus, there are no other characters which would help to sustain the genus, since we have a com- plete series regarding the location of the male pores running through three successive somites. As few species of Dzplo- cardia are known, and no confusion will ensue, a fusion of the two genera Aleodrilus and Diplocardia will help to simplify the already extensive nomenclature of the terrestrial O/zgo- chaeta. Through the kindness of Prof. Frank Smith, I have had opportunity to examine all the various Dzplocardia species at his disposal, and am able to add a few observations on minor points. I have also received several new species from North Carolina, which I describe here preliminarily, reserving a more detailed description for an illustrated paper now in press. As many more species of Dzplocardia are likely to be found in the United States, a review of the species, so far known, is of 162 EISEN. [Vou. II. considerable interest, especially as the original descriptions are scattered, and not readily compared. I wish to call especial attention to the position of the spermathecal pores, which in some species are post-septal, in others pre-septal, while in at least one species some of the pores are pre-septal, while others are post-septal. The existence of sexual spermathecal setae is of the greatest interest. The sculptures of these setae vary in different species, and in a detailed description should be carefully noted. Another interesting feature in the anatomy of at least one species, and probably in several, is the posterior ‘“ glandular crop’ of the intestine, found in xiv and xv. _ It is of a totally different structure from the gizzard, and resembles greatly the glandular crop which I have once described in Pontodrilus Michaelsent, and which in this species is also situated poste- riorly. With our extended knowledge of new species, it will be necessary to modify the definition of the genus as given by Ude, the last one to define the genus. Several of the char- acters considered by him generic are now seen to be only specific. In the following I have endeavored to mark the thickening of some septa in a way that it could be readily recognized. The number of bars above the Roman numeral indicates the comparative thickness of the septa. Thus, one marked with three bars is about three times thicker than the one marked with one bar, etc. DIPLOCARDIA GARMAN. Definition. — Setae, eight, in four couples, lateral and ventral. Pental setae, present or absent. Spermathecal setae, present or absent. Pyvostomtum divides somite i more or less. C7zte/- lum saddle or ring like, generally xiii—xviiil. Ovzducal pores xiv. Spermathecal pores, two or three pairs, either post- septal or pre-septal. Spermiducal pores on xix, XX, OY XXI, according to species. Prostate pores on somites next anterior and posterior to the spermiducal pores. The pores on each side connected by a groove. A genttal zone generally present, i No. 4.] NORTH-AMERICAN EARTHWORMS. 163 with or without papillae. /ztestine with two gizzards, gener- ally inv, vii Ocesophagus either with or without folds contain- ing calcic concretions, but never with calciferous diverticula, as in Benhamia. Sometimes a glandular crop in xiv and xv. Sperm sacs, one pair pre-septal in ix, one pair post-septal in xii. Two pairs testes in x, xii Two pairs sperm funnels in x, xi. Pyvostates, two pairs, opening anteriorly and poste- riorly to the sperm ducts. Spervmathecae, two or three pairs, each one with a diverticulum near the center. Dorsal vessel, double or single. Wephridia, meganephridia, generally without coelomic mantle. Acanthodrilidae. As far as known, confined to the United States and to northern Mexico. The genus Diplocardia differs thus from Acanthodrilus in having two successive gizzards, Acanthodrilus having only one. From Benhamia, which genus possesses two successive giz- zards, three pairs of calciferous diverticula, and numerous micro- nephridia, Dzplocardia is distinguished by its meganephridia, of which there are two in each somite, and by the absence of calciferous diverticula of the tubular intestine. From both Benhamia and Acanthodrilus, as well as from all other genera of the family, Dzplocardia is characterized by the position of its male or spermiducal pores. Key to Species of Diplocardia. I. Spermiducal pores in somite xxi, no penial setae. (A/eodrilus.) SP. t. D. Keyesi (Eisen). II. Spermiducal pores in somite xx. aS 2e D. verrucosa Ude. Ill. Spermiducal pores in xix. A. Spermathecae, two pairs. Sp. 3. Both pairs of spermathecal pores pre-septal, or posterior to the setae ; sexual spermathecal setae present in viii and ix. D. Ezsent (Michaelsen). Sp. 4. The pair of spermathecal pores in viii are post-septal; the pair in ix are pre-septal. Sexual spermathecal setae in viii and ix. D. Michaelseni n. sp. Sp. 5. Both pairs of spermathecal pores are post-septal, sexual sperma- thecal setae in vili—x. D. Udet un. sp. Sp. 6. Both pairs of spermathecal pores are post-septal, no sexual spermathecal setae. D. riparia Smith. 164 EISEN. [Vor. II. 4. Spermathecae, three pairs. Sp. 7. Penial setae straight, about one-half longer than ordinary setae. D. communis Garman. Sp. 8. Penial setae sigmoid, several times longer than ordinary setae. a. Penial setae not ornamented. D. singularis Ude. B. Penial setae ornamented. WD. stmgularis, subsp. n. caroliniana. DipLocaRDIA KEYES! (EISEN). Definition. — Color, flesh, marbled violet, no pigment. Szze, 7omm.by § mm. Somites,150. Prostomium divides somite 1 about one-half. Dorsal pores, the most anterior one in 1 vill/Ix. Spermiducal pores in xxi. Spermathecal pores, two pairs, in vill and ix, in front of setae ab. Prostate pores in xx, xxii. Oviducal pores in front of setae a. Setae all ventral ; a—6 slightly larger than c-d; a—a larger than 6-c. No sculpture. Penzal setae none. Spermathecal setae not differentiated. C/ztellum ring-like anteriorly, posteriorly saddle-shaped. Gezztal zone not distinct, two parallel grooves in % xx—-% xxii; groove almost straight, with a knob at each apex ; concavity turned ventrally. Seféa, thickened are : vi/vii, vii/vill, vill/ix, 1x/x, x/Xx1. Oesophagus without calcic concretions. Gzszards v, vi. Sac- culated intestine xv. Dorsal vessel single, not covered with chloragogen cells. Hearts in x, xi, xii, with large pulsating divisions ; no chloragogen cells. Wephridia, meganephridia, no coelomic mantle. Testes x, xi. Sperm funnels x, x1. Sperm ducts, which join at xii/xili in a common muscular sheath ; fuse in xx/xxi. Sperm sacs, one pair pre-septal in 1x, one pair post- septal in xii. Sperm masses in x, xi. Ovitducts in xiv. Prostates confined to one somite each, small, tubular, thicker at apex. Spermathecae, two pairs in viii, ix; distal end knob-like; the duct is very. slender and long, with a very minute wart-like and ear- shaped diverticle, about the middle of the duct. Habitat.— Ensenada de Todos Santos, Baja California, Mexico. DIPLOCARDIA VERRUCOSA UDE. Definition. — Color, pink. Size, 65 to 75 mm. by 2% to 3 mm. Somttes 100 to 125, body round, of even thickness. No. 4.] - WORTH-AMERICAN EARTHWORMS. 165 Prostomium divides somite 1 by one-half. Dorsal pores, most anterior one vill/ix (or x/xi). Spermzducal pores on xx. Sper- mathecal pores on anterior % of somites ix, x, somewhat dorsal to setae ad. Prostate pores on xix, xxi. Oviducal pores interior to setae a, no glandular ridge. Se¢ae sigmoid, very faintly orna- mented. Distance dd more than ¥% the periphery; c-d some- what larger than a—d ; a—a three times, and 6-6 two and a half times larger than a—0, no setae ad in xx. Penial setae curved, not ornamented. Spermathecal setae not differentiated. C7ztel- lum saddle-shaped, xiii—xvill. Genttal zone, a rectangular field from posterior % XVili-% xxii, extending laterally to center between d-c. Two deep grooves from % xix—% xxi, the con- vexity of which is outwards, except in the center of xx, where it is turned towards median line; one median papilla on xxii; one pair papillae on xix in line with setae 6; one pair papillae on xix and xxi, interior to grooves ; one pair papillae exterior to grooves on each of xix, xxi, xxii (two pairs papillae on each of xix, xxi, and three papillae on xxii). Sepfa, thickened are: ———— x Vi/Vii, Vii/Vill, Vill/ix, 1x/x, X/Xi, Xi/X11. Ocesophagus, no calciferous folds or thickenings. Gzzzards in v, vi. Sacculated intestine commences in xvi. Dorsal vessel single. Flearts, three pairs in x, xi, xll. Vephridia, meganephridia, com- mence in ii, pores intersegmental in front of setae d@. © Testes, x, xl. Sperm funnels, x, xi. Sperm ducts open in central part of groove in xx. Sperm sacs, one pair pre-septal in ix, one pair post-septal in xii. Ovzducts open in front of and interior to setae a. Prostates very thin, even, bent in four folds, confined to one somite each. Sfermathecae, two pairs in viil, ix, retort- like, with a small, short-stalked, ear-like diverticulum below the center. No specialized spermathecal setae. Habitat. — Omaha, Nebraska. (See note on page 172.) DIPLOCARDIA EISENI (MICHAELSEN). Definition. — Color, dorsally gray or pigmented, clitellum violet gray. Szze, 150 mm. by 2mm. Somites, 165; viii—xiil, smoother and wider than the others. Pyrostomztum divides somite 166 EISEN. pVer ls i about one-half, with the lateral margins strongly converging. Dorsal pores, most anterior one on xi, first distinct one on xili. Spermiducal pores on xix in line with setae a. Spermathecal pores Vill, 1x, posterior to setae a, in line with ad. Prostate pores Xvill, Xx, in line with setae 6. Oviducal pores near median line, surrounded by a zone. Se¢ae, sigmoid, with numerous fine bars ; a—a about ;/5, dd, 3 the whole periphery ; d—c is shorter than a—a ; a—0, shorter than c-d; a—b, % as long as d-c; a-b, slightly shorter than c-d. Setae 0 in xix is present, a is absent or present. Penzal setae rudimentary or very small, in the body wall of xviii and xx. Sfermathecal setae differentiated and ornamented in villi and ix. C7@ztel/um ring-shaped in xill—xvii, saddle-shaped in xvili. Gezztal zone, a quadrangular glandular ventral zone in xvill—xx, in the corners of which lie the prostate pores. The two grooves are curved ventrally. No depressed area and no papillae. Seta, thickened are : Vi/vil, Vil/Vill, Vili/ix, 1x/x, x/X1, xi/xii. Gizzards v, vi. Sacculated intestine commences in xviii, a dor- sal typhlosole. Dorsal vessel alternatingly double and single in vi-xv. ffearts, four pairs in x—xill. Mephridia, megane- phridia, commence in ili. Testes in x, xi. Sperm funnels x, xl. Sperm ducts join, but do not fuse until at the male pore in xx. Sperm sacs, one pair pre-septal in 1x, one pair post-septal in xll. Ovzducts large. Prostates, two pairs in viii, ix. A large sac-like part and a thinner, irregularly bent, muscular duct; a small, stalk-like diverticle with a knob-like apex. Flabitat. — Florida. DIPLOCARDIA RIPARIA SMITH. Definition. — Color, brown anteriorly and dorsally, clitellum dull coppery colored. Szze, 220-250 mm. Somttes, 136-157. Prostomtum divides somite i by one-half. Dorsal pores, most anterior one on anterior margin of xi, near x/xi. Spermiducal pores, xix. Spermathecal pores, two pairs in viii, ix, anterior to setae ab. Prostate pores, xviii, xx. Oviducal pores, xiv. Setae as in DD. communis, no ventral setae ad in xix. Distance —s No. 4.] NORTH-AMERICAN EARTHWORMS. 167 a—a = b-c; a-b very little larger than c-d. Pental setae, xviii, xx. Spermathecal setae not differentiated (?). Clztellum saddle- shaped in xili-xvilil. Cevztal zone, no rectangular ventral zone; a ventral depression in xvii—xxi, deepest in xviii and xx. A pair of crescent-shaped grooves curved ventrally, from center of xvill-xx. Two papillae very close to median line, between xxi/xx. One median papilla xvi/xvii, one pair papillae xvii/xviii, one pair papillae xx/xxi, one pair papillae xvii/xvili. Gzzzards v, vi. Sacculated intestine commences xviii. Dorsal vessel single. Mephridia, meganephridia, a small pair in il. Teszes in x, xi. Sperm funnels x, xi. Sperm sacs, one pair pre-septal in ix, one pair post-septal in xil. Pvostates xviii, xx. Sperma- thecae, two pairs in vili, ix, with a large ear-like diverticulum, which is very prominent and exteriorily slightly racemose. Anterior and posterior spermathecae are of the same size. Habitat. — Havana, banks of Illinois River, Illinois, U. S. A. DIPLocARDIA MICHAELSENI 7%. SP. Definition. — Color, flesh. Szze, 45 mm. by 2 mm., hardly tapering posteriorly. Somztes, 63. Prostomium divides somite icompletely. Dorsal pores, most anterior iv/v. Spermiducal pores xix. Spermathecal pores, one pair pre-septal in ix, one pair post-septal and almost central or median in vill. Prostate pores xviii, xx. Oviducal pores xiv, in front of and anterior to setae a, close together. Setae all ventral; a—-a= 3 ab; a-a about one-third larger than d-c; d6-c= about 2 a—b. Penial setae present at spermiducal pore. Spermathecal setae present in viii, ix; setae a and 6 being differentiated and sculptured. Clitellum ring-like, dorsally xili-% xvili; ventrally xiv—xvii. Genital zone, a deep central, oval pit in xvili-xx, surrounded by an elevated ridge. A pear-shaped ventral and median papilla in xxi and % xxii, and a similar papilla in % xxii and xxiii. Grooves between prostate pores are straight.