st abe ates eo Te ee rie » gly ot GeO Nae emi eS eta s oo chorea Eo fonkm omit me etna mieten de ine oy Md mth! pare: Yr ne Mee ote Pwo or wre! 4 -yoyetoens Y= Oe es ee land se ptt tn einer tm ahve! apr tui ntfs fog insge pa hehe Seal i ri porrehatory ae ih aRie none prerret alae ”" OA ae pe Rah ie gt Mee el aE ee PA OE PRO Re UO Te ee ee el aa “a Nore eee: egy apna me remcsnnty HARVARD UNIVERSITY EEL LIBRARY MUSEUM OF COMPARATIVE ZOOLOGY ait fi eh QUARTERLY JOURNAL OF MICROSCOPICAL SCIENCE: EDITED BY HE. RAY LANKESTER, M.A., LL.D., F.R.S., Linacre Professor of Comparative Anatomy, Fellow of Merton College, and Honorary Fellow of Exeter College, Oxford. WITH THE CO-OPERATION OF ADAM SEDGWICK, M.A., F.RS., Fellow and Lecturer of Trinity College, Cambridge ; AND W. F. BR. WELDON, M.A... F.&:S., Jodrell Professor of Zoology and Comparative Anatomy in University College, London; Fellow of St. John’s College, Cambridge. VOLUME 37.—New Serrtzs. With Mithographic Plates and Engrabings on Wood. LONDON: J. & A. CHURCHILL, 11, NEW BURLINGTON STREET. 1895, EAT el PO, A CHANTAL 7 mons (00%, MOO CUT “yy Se FLY aa0tdaWAg | P ' ery OP LT Tay WAT) hy ae /. ¥ Wetec : ae . ; i wake aoe Paty tat ied: te 4 a MoT | yy j ieil ey by wil Plee inf La Hal i 4 , 7 : 1 - ; = ul e 7 - i { : ty a . ee Ce Ase duce Habit paiue THA Eh SHUNT UTRD k , z a ¥ Baie fr | rn CONTENTS. CONTENTS OF No. 145, N.S., NOVEMBER, 1894. MEMOIRS : On Julinia; a New Genus of Compound Ascidians from the Antarctic Ocean. By W. T. Caiman, reat NaN Dundee. (With Plates 1—3) ; ; Hermaphroditism in Mollusca. By Dr. Pav Petsenzer (Ghent). (With Plates 4—6) A Description of the Cerebral Convolutions of the Chimpanzee known as “ Sally ;” with Notes on the Convolutions of other Chimpanzees and of two Orangs. By W. Biaxtanp Benuaw, D.Sc.Lond., Hon. M.A.Oxon., Aldrichian Demonstrator of Com- parative Anatomy in the University of Oxford; Lecturer on Biology at Bedford College, London. (With Plates 7—11) On the Inadequacy of the Cellular Theory of Development, and on — the Early Development of Nerves, particularly of the Third Nerve and of the Sympathetic in Elasmobranchii. By Apam SEDGWICK, F.R.S. : F On Benhamia cecifera, n.sp., from the Gold Coast. By W. Braxtanp Benuam, D.Se.Lond., Hon. M.A.Oxon., Aldrichian Demonstrator in Comparative Anatomy in the University of Oxford, &c. (With Plate 12) CONTENTS OF No. 146, N.S., DECEMBER, 1894. MEMOIRS : A Re-investigation into the Early Stages of the Development of the Rabbit. By Ricnarp Assueton, M.A. (With Plates 13—17) On the Phenomenon of the Fusion of the Epiblastic Layers in the Rabbit and in the Frog. By Rrcuarp AssHETON, M.A. (With Plate 18) : : : : : b PAGE 19 47 87 103 113 165 iv CONTENTS. On the Causes which lead to the Attachment of the Mammalian Embryo to the Walls of the Uterus. By Richarp AssHETON, M.A. (With Plate 19). The Primitive Streak of the Rabbit ; the Causes which may deter- mine its Shape, and the Part of the Embryo formed by its Activity. By Ricnarp AssHEton, M.A. (With Plates 20—22) On the Growth in Length of the Frog Embryo. By RicHarp AssueTon, M.A. (With Plates 23 and 24) CONTENTS OF No. 147, N.S., MARCH, 1895. MEMOIRS : On the Variation of the Tentaculocysts of Aurelia aurita. By Epwarp T. Browne, B.A., University College, London. (With Plate 25) : : : On the Structure of Vermiculus pilosus. By E.8. Goopricu, F.L.S., Assistant to the Linacre Professor, Oxford. (With Plates 26—28) . 5 : : On the Mouth-parts of the Cypris-stage of Balanus. By THExo. T. Groom, F.Z.8., late Scholar of St. John’s College, Cam- bridge. (With Plate 29) : A Study of Coccidia met with in Mice. By J. Jackson CLARKE, M.B. Lond. (With Plate 30) Observations on Various Sporozoa. By J. Jackson CLARKE, M.B.Lond. (With Plates 31—383) A Revision of the Genera and Species of the Branchiostomide. By J. W. Kirxatpy. (With Plates 34 and 35) Sedgwick’s Theory of the Embryonic Phase of Ontogeny as an Aid to Phylogenetic Theory. By E. W. MacBuripg, B.A., Fellow of St. John’s College, Cambridge; Demonstrator in Animal Morphology to the University of Cambridge PAGE 173 191 223 245 253 269 277 287 303 325 CONTENTS. CONTENTS OF No. 148, N.S., MAY, 1895. MEMOIRS: The Anatomy of Alcyonium digitatum. By Sypnny J. Hickson, M.A., D.Sc., Beyer Professor of Zoology in the Owens College, Manchester; Fellow of Downing College, Cambridge. (With Plates 36—39) E : : ; Note on the Chemical Constitution of the Mesogloea of Aleyonium digitatum. By W. Lanepon Brown, B.A., late Hutchinson Research Student, St. John’s College, Cambridge A Study of Metamerism. By T. H. Morean, Ph.D., Associate Professor of Biology, cs Mawr Nee U.S.A. (With Plates 40—43) . : : 5 On the Colom, Genital Ducts, and oe By Epwin $8. Goopricu, F.L.8., Assistant to the Linacre Professor of Com- parative Anatomy, Oxford. (With Plates 44 and 45) PAGE 343 389 395 477 7 ey e/ j iWtAerrivtee = HORE EA Ge eed vow AO. Oat) el '} | ‘ enti 2 20’! : j ‘ AP Fadtied 4 widedisth. aichio get A FO waite yreew OF. OL) AV y ie Maite PP ose We Pe Ad ee Le ; . my . Tae inet) a ond aot WeayW ; i if nye wae nepal ’ abe ‘ Re fit exalt ae bite a ‘ ° aroine vila Wrad) z ead) &. SUA ed era (woth) hy oth BP | eeinriiie Glee Sith A ae Hagel we Arie it ul A it a > neil Ay eA. a) } | iy df aa ' : rai) ] 1.UNG I | t av PPaAra i. Waites ) i woes ie ei? & d 7? ’ af : ’ ‘ 4 ‘ } . Lio 4 twee UT) ee i) ! We ; tte ANGE 4 R : : . thd ee Vie ee, et Gh aie) it ied 1p > vita) ) ee ‘ l | thi mb be taly wr Al rpiree hil De LP Ate 7 bay FI Pil ae Le) erie oe) one i . + e 7 ‘ a a 2 s 7 : bee _ i io ) « J “ . « 1 ef \ ‘ * ; = 4 - i eae, te 7 =. 7 ‘ On Julinia; a New Genus of Compound Ascidians from the Antarctic Ocean. By WwW. T. Calman, University College, Dundee. With Plates 1—3. Tuer Ascidian described in the present paper was collected in the Antarctic Ocean by Dr. C. M. Donald, of the whaler “ Active,”’ during the Dundee Whaling Expedition of 1892-93, and was placed along with other specimens in Professor D’Arcy W. Thompson’s hands. Its large size and remarkable appearance attracted attention, and closer examination sug- gested its possible identity with the species described by Professor Herdman in his report on the Compound Ascidians of the Challenger Expedition (‘Chall. Rep.,’ vol. xiv, pp 250) as (?)ignotus. The specimens on which this species was then tentatively founded, two of them in the Challenger collection and the third in the British Museum, were in an unsatisfactory state of preservation, so that Professor Herdman was unable to determine the genus, or even with certainty the family, to which the species belonged. Our specimen was pre- served simply in methylated spirit, but on sectioning it was found to be in a better state than might have been anticipated, giving good results with the usual stains and making possible a fairly complete study of its anatomy. In the following description, Maurice’s monograph on Fragaroides (‘ Arch. de Biol., viii, pp. 205—495) has been used for comparison VoL. 37, PART 1.—NEW SER. A 2 W. T. CALMAN. throughout, as being the most complete account we possess of the structure of a Compound Ascidian. Our species, whose relation to Professor Herdman’s (?) ignotus I shall again return to, evidently forms the type of a new genus, which at the suggestion of Professor D’Arcy Thompson (under whose guidance this paper has been pre- pared), I propose to dedicate, under the name of Julinia, to the learned naturalist whose researches on the subneural gland and other anatomical features form classical contributions to our knowledge of the Ascidians. The specimen was found floating on the surface of the sea in the north of Erebus and Terror Gulf, and Dr. Donald tells me that considerable quantities were seen. As there is nothing in its structure to suggest a pelagic mode of life, the colonies may have been torn from their submarine attachment in shallow water. At the same time the remarkable genus Celocormus must be remembered as showing the possibility of a Compound Ascidian leading an unattached life though destitute of any special means of locomotion; moreover in the present case the extreme elongation of the colony renders it improbable that it could have been supported by a narrow base of attachment. External Appearance and General Features of the Colony. The colony is irregularly cylindrical in shape (Pl. 1, fig. 1), measuring 78'5 cm. in length and from 1°5 to 2°5 cm. in dia- meter. One end is torn and ragged, while the other is slightly tapered and smoothly rounded. No attaching fibres such as Herdman describes are preserved. Over large areas, particu- larly along one side, the surface presents a decayed appearance, the zooids being absent or scattered here and there and partially macerated. Over the rest of the surface the zooids are arranged in fairly regular round or oval systems (PI. 1, fig. 2), each comprising from about six to twelve zooids. The colour of the whole is yellowish with a tint as though in life it had been orange, the zooids appearing as lighter spots. The buccal orifices are small and six-lobed. In the centre of ON JULINIA. 3 each system is a large and irregularly-lobed common cloacal opening. In a transverse section through the colony (Pl. 1, fig. 9) the zooids are seen arranged round the circumference. From their proximal ends vascular processes pass inwards, branching aud interlacing in the interior of the colony. In the further structure of the framework of the colony we have to deal with a number of elements more or less obscure, which may be briefly enumerated but which remain here as in other cases imperfectly understood. Just below the ascidiozooids large numbers of peculiar vesicles occur scattered among the vessels. These vesicles are spherical, from ‘3 to °7 mm. in diameter, and consist of a denser layer of the test-substance surrounding a gra- nular mass of variable size and appearance, in which numerous nuclei are occasionally found. It is possible that these masses may represent degenerating zooids like those described by Maurice in Fragaroides, though their great abundance is rather against this view. Many buds in different stages of development occur scattered among these vesicles. The centre or core of the colony is composed of the spongy test-substance, traversed by the vascular processes, which here run more or less in a longitudinal direction. In its minute structure the test resembles that described by Herdman in Colella (1. ¢., p. 78). In section (PI. 3, fig. 32) it presents a reticulate appearance, the cavities occupied by the large vesicular or vacuolated cells known as “ bladder-cells” reducing the matrix to a mere interstitial network. These “ bladder-cells”? are of common, but not universal occurrence among Compound Ascidians. In the immediate neighbourhood of the zooids and of the vascular processes the tissue is denser, forming tracts of homo- geneous matrix-substance which show, besides a few bladder- cells, small test-cells of the more usual type, with scanty protoplasm and irregular outline. Near the surface of the colony numerous groups of yellowish pigment-cells occur. These are irregularly rounded (probably shrivelled by the spirit), highly refracting, and apparently homogeneous. In 4 W. T. CALMAN. the deeper parts of the colony are found scattered here and there a few small crystalline masses (Pl. 1, fig. 11). These bodies dissolved somewhat slowly and without effervescence in hydrochloric acid, but on account of their small number their chemical nature could not be definitely ascertained, and they cannot be certainly assumed to represent the stellate calcareous spicules, e.g. of the Didemnide. Scattered here and there in the test-substance are groups or nests of rather large cells (Pl. 1, fig. 10) containing rounded deeply staining masses of varying size and appearance. The protoplasm is often very granular, and no nucleus is visible. It is possible that these obscure bodies represent the test- phagocytes of Maurice, to which he ascribes the function of absorbing the dead zooids, but it is to be noted that here they do not occur in relation to the masses described above as possibly representing degenerated zooids. General Conformation of the Ascidiozooids. The body (Pl. 1, fig. 3) is divided into two regions, the . thorax and abdomen, united by a narrow neck. The thoracic region contains the branchial sac, and varies in shape accord- ing to the state of contraction. In the most expanded speci- mens it is roughly square-shaped when seen from the side, and broadly oval in transverse section (PI. 2, fig. 14). A number of longitudinal muscle-bands run in the mantle, about four to six on each side. Along the ventral edge of the thorax runs the large undulating endostyle. On the dorsal surface of the animal is the wide atrial orifice, behind which is attached the vesicle described below as probably representing an incubatory pouch. In front of the atrial orifice is the very large atrial languet. The languets belonging to the several zooids of a system meet at their edges to form a composite roof to the common cloacal chamber; seen from the outside this roof is hidden by a continuation of the common test, in the centre of which is the common cloacal opening. As a result of this arrangement, the individual languets are somewhat irregular in shape. There appears to be no such ON JULINIA. 5 continuation of the test-substance on the lower surface of the languets as is seen in Fragaroides. In the abdomen the loop of the gut lies transversely to the dorso-ventral axis, the oval stomach being on the right side. On the dorsal side of the loop is the ovary, containing one or two large orange- coloured ova. In front of this the opaque white vas deferens leads forwards to the cloaca. The heart in its pericardium lies on the ventral side of the abdomen. Branchial Cavity.—Tke buccal siphon is short, ending in front in six simple rounded lobes (Pl. 1, fig. 8). It is lined as usual by a reflected portion of the test, and a strong sphincter muscle is developed in its walls (Pl. 2, fig. 21). The oral tentacles (Pl. 1, fig. 8) are twelve in number, situated on a hexagonal ridge surrounding the base of the buccal siphon. Six of the tentacles, situated at the angles of the hexagon, are larger than the others which alternate with them. Separated from the tentacles by the “ peribranchial zone” is the “ peri- pharyngeal band,” here apparently a simple ridge, not grooved as in Fragaroides. On the ventral side it is continuous with the anterior end of the endostyle, while on the dorsal side it appears to fuse, as in Fragaroides, with the small dorsal tubercle which bears the simple rounded opening of the sub- neural gland. The branchial sac is large and comparatively simple. There are four rows of long and narrow stigmata, each row contain- ing about twenty-four slits on each side (Pl. 2, fig. 12). Between the four rows three shelf-like membranes project into the interior of the sac. These are the “ lames intersériales ” of Maurice, and are continued round the circumference of the sac, being only interrupted by the endostyle. Along the middle of each row of stigmata runs a transverse vessel connected with each of the interstigmatic bars, but not otherwise interrupting the stigmata. A similar arrangement to this occurs in the genus Distaplia. On the dorsal side, but to the left of the middle line, each interserial lamella is produced into a large ciliated languet. Of these the two anterior are placed some- what nearer the middle line than the third. A similar un- 6 W. T. CALMAN. symmetrical position of the dorsal languets to the left of the middle line occurs according to Herdman in Distaplia rosea, and Maurice describes the same arrangement in Fragaroides. The branchial sac is connected with the wall of the atrium by small trabecule placed at the level of the interserial lamelle (Pl. 1, fig. 8). These are few in number, apparently only one on each side for each lamella, though I could not ascertain this with certainty. This condition presents the opposite extreme to that seen in Fragaroides, where an almost continuous membrane connects the branchial sac with the body-wall between each row of stigmata. The wall of the branchial sac, in contrast to that of Fragaroides, is here entirely without muscles, except for the longitudinal bands running in the lips of the endostyle and those at the sides of the dorsal sinus. The minute structure of the interstigmatic bars agrees perfectly with that described by Maurice for Fragaroides. Along each side facing the cleft is the specialised stigmatic epithelium (Pl. 2, figs. 17 and 18), formed by six regular rows of ciliated cells. Each cell forms at its free end a narrow longitudinal ridge along which the cilia stand in a single row. The surfaces of the bar turned towards the branchial and atrial cavities are covered by simple flattened epithelium. In the centre is a blood-sinus bounded by the basement membrane of the epithelium. The endostyle is large aud much undulated. In transverse section (Pl. 2, fig. 20) it shows essentially the same structure as that of Fragaroides. The everted lips are covered by ciliated epithelium. At the bottom of the groove a narrow band of columnar cells bears the enormously long cilia charac- teristic of the Ascidian endostyle. On each side of the groove is the “ glandular ” epithelium, consisting of elongated curved cells tapering towards their free ends, and this side wall of “slandular” epithelium is divided into three parts by two small bands of ciliated epithelium. These two bands are here identical in structure, while in Maurice’s account of Fraga- ON JULINIA. 7 roides the inner or more median of the two appears as a “troisiéme épithélium vibratile de caractéres tout particuliéres.” Just beneath each of the recurved lips of the endostyle a muscle-band runs along its entire length. At the posterior extremity the right lip is continued backwards as a ciliated ridge (“raphé postérieur”’?) which becomes continuous with the esophageal funnel. Digestive Tract.—The funnel-shaped opening of the cesophagus into the branchial chamber is large and convoluted (Pl. 2, fig. 13), and the columnar ciliated cesophageal epi- thelium which covers it is sharply marked off at its edge from the flattened epithelium of the branchial sac. Close to the free surface of the cesophageal epithelium is a layer of deeply staining granules of unknown nature. The csophagus has thick and convoluted walls (Pl. 2, fig. 19), and its hinder end projects for a little distance into the stomach, where it forms a sort of valvular opening (PI. 2, fig. 23). The stomach (Pl. 1, fig. 4) has the form of an oval sac, smooth on the outside but having the lining epithelium thrown into oblique longitudinal folds. This epithelium is very badly preserved in all our specimens, but it can be seen that the cells are columnar and stain deeply. Embedded in the lining epithelium of the stomach, or lying free in its cavity, are certain large oval or pyriform cells (Pl. 2, fig. 25). These have sharp outlines and very granu- lar protoplasm, stained only faintly by hematoxylin and not at all by carmine. The nucleus is large and transparent, and contains a large deeply-staining nucleolus. In one or two cases a small segment, marked off from one end of the cell as shown in fig. 25 a, confirms a belief suggested by the nuclear and other characters that these are parasitic Gregarines. _ A slight constriction separates the stomach from the intestine (Pl. 2, fig. 23). The latter is of nearly uniform diameter throughout its length, there being no division into duodenum, chylific ventricle, and rectum. Its walls are slightly folded, and it is lined throughout by columnar epithe- lium. The terminal part projects slightly into the cloacal 8 W. T. CALMAN. chamber, forming what Maurice calls a “ pavillon anale,” but I have failed to find an anal sphincter, though such in all probability exists. The intestinal gland is well developed. A series of tubules ramifies over the surface of the intestine (Pl. 2, fig. 24); these are lined by somewhat irregular cells, and unite to form a duct which crosses the intestinal loop to open into the stomach at about its middle (Pl. 2, fig. 23). Nervous System.—The nerve-ganglion is oval in form, and gives off numerous nerves. Two large nerves in front and a large median nerve (the cloacal), flanked by two smaller lateral ones behind, are easily seen; but it was found impossible to determine the number or arrangement of the smaller lateral nerves. On the dorsal wall of the dorsal blood-sinus a slight thicken- ing composed of a few small cells (Pl. 2, fig. 22) represents the “cordon viscéral ganglionnaire” of van Beneden and Julin, the remains of the central nervous axis of the larva. In a fortunate series of sections this cord can be traced into con- nection with the cerebral ganglion, which it enters at the root of the cloacal nerve. In minute structure the ganglion corresponds exactly with that of Fragaroides (Pl. 2, fig. 21). Outside is a layer of large nerve-cells, finely granular and with conspicuous nuclei, sending processes inwards—the “grey matter’’ of Julin,— while the interior is composed of the usual fibrillated “ punkt- substanz”’—the “ white matter” of Julin. Subneural Gland.—The subneural gland (Pl. 2, fig. 21) is very feebly developed, consisting merely of a thin layer of cells lying on the posterior face of the nerve-ganglion, separated from it in the middle line by a space representing the back- ward continuation of the duct. The tissue of the gland con- sists of ill-defined rounded cells, many of them containing vacuoles. The duct, on the other hand, is well developed and normal in structure. As in Fragaroides, its anterior wall is con- tinued back over the posterior face of the ganglion for some ON JULINIA. 9 distance, while the ventral wall stops short of this and becomes continuous with the band of glandular tissue. The columnar epithelium lining the duct is reflected at its mouth over the ‘dorsal tubercle,” and in the interior of the duct bears the usual long cilia, one to each cell. It is interesting to note that the extreme degeneracy of the gland itself (carried to a greater degree, I think, than has been described in any other Ascidian) is not associated with any corresponding reduction in the size of the duct. Indeed, even in other cases where the gland is much better developed (e. g. Fragaroides), the duct, both in respect to size and to the richness of its ciliation, seems out of all proportion to any secretory function we can ascribe to the gland. Circulatory System.—The heart in its pericardium lies on the ventral side of the intestinal loop (Pl. 1, fig. 4; Pl. 2, fig. 16). It is a thin-walled, somewhat fusiform tube, but I was unable to trace any vessels connected with it. A space lined by endothelium and lying to the left of the peri- cardium may possibly represent a vestige of the epicardial system. A large blood-sinus runs along the dorsal edge of the branchial sac (Pl. 2, figs. 14 and 22); in its lateral walls run strongly developed longitudinal muscle-bands, while in the middle of its dorsal wall runs the ganglionic nerve-cord above alluded to. The “vascular process” arises near the hinder end of the abdomen, a little to the left side. In the substance of the test these processes branch and probably anastomose, and finally end in club-shaped dilatations (Pl. 1, fig. 6). In section (Pl. 3, fig. 32) each is seen to consist of a simple tube of ectoderm divided into two by a median partition of thin and apparently structureless membrane. No internal layer of connective tissue, as described by Herdman in Colella, could be distinguished, the processes appearing to be exclusively ectodermal, like the ‘‘stolons” of Botryllus (Hjort, ‘ Mitt, Zool. Stat. Neap.,’ x, p. 589), a point which is of interest in connection with the probable origin of buds from them. Numerous blood-corpuscles occur in the cavity of these tubes. 10 Ww. T. CALMAN. Sexual Organs.—The sexual organs lie on the dorsal side of the intestinal loop, the ovary being external or dorsal to the testis. In all the individuals examined several ova were found in various stages of development, the largest being about ‘6 mm. in diameter. They are contained in follicles opening by short canals into the oviduct (Pl. 3, fig. 27). No very young ova were seen, the smallest being about ‘1 mm. in diameter, and nowhere did the ovary present in section the bilateral T-shape described by Maurice in Fragaroides and by van Beneden and Julin in Clavellina. Inside the wall of the follicle in all save the youngest there is a loose layer of cells outside the egg-membrane (PI. 3, fig. 28). This seems to correspond to the inner of the two layers into which the follicular wall splits in Clavellina according to van Beneden and Julin (‘ Arch. de Biol.,’ vi, p. 358), though in the latter case the splitting does not take place until the ovum is nearly ripe. Inside the egg-membrane are the characteristic and problematical “ test-cells,” with large nuclei each containing several nucleoli, and scanty protoplasm. The vitellus becomes more and more coarsely granular as the egg matures, and in the oldest eggs it is broken up into numerous irregular yolk- masses. The oviduct is a thin-walled tube running up alongside the vas deferens (PI. 2, figs. 15 and 16). On the dorsal surface of the animal’s thorax, near its junc- tion with the neck and rather to the left of the mid-dorsal line, there is attached by a narrow neck a spherical thin-walled vesicle of about ‘5 mm. in diameter (PI. 1, fig. 3). The wall of this vesicle is composed of two layers, the outer continuous with the ectoderm of the thorax, while the inner seems to be continuous with the wall of the oviduct. It is difficult to make out the exact relations of the parts, but I believe that the lumen of the oviduct opens into the vesicle. This organ occupies the position of the incubatory pouch possessed by so many Com- pound Ascidians, and its apparent connection with the oviduct naturally suggests that it represents that structure, though I have never found eggs or larvee in it in a single instance. In nearly all the specimens examined hardly a trace of ON JULINIA. al testis could be found, although the vas deferens was usually full of spermatozoa. In one or two cases, however, the testis was developed. It consisted (Pl. 3, fig. 26) of a number of follicles lying in the connective tissue in the intestinal loop, connected by branching tubes with the vas deferens and showing groups of developing spermatozoa in each. In one specimen (badly preserved, unfortunately) the follicles are considerably larger than those figured, and the lower part of the vas deferens is very much distended (to °3 mm. diameter) with spermatozoa. No definite relation between the states of maturity of ovary and testis such as would suggest the occurrence of protandry or protogyny could be demonstrated. The vas deferens is usually about ‘06 mm. in diameter, and is composed of cubical epithelium, probably ciliated. Its hinder part is much convoluted, though not spirally coiled as in the Didemnide; in front it pursues an undulating course to open close beside the anus. The spermatozoa are linear, slightly curved, about ‘001 mm. in length, and bear long flagella. Buds.—Numerous buds in various stages of development are scattered in the substance of the test just below the layer of ascidiozooids. All of them, even the youngest, lie quite free and unattached, usually alongside a vascular process (Pl. 1, figs. 5 and 6). In no case did I succeed in tracing a con- nection with the parent animal, and the origin of the buds must therefore be left undecided—more particularly since Hjort (‘ Mitt. Zool. Stat. Neap.,’ x, pp. 588 and 589) has contested Herdman’s account of the process of ‘stolonial ” budding in the Botryllidz. In view of the obscurity still involving many points in the process of bud development in the Tunicata, it did not seem advisible to attempt a detailed examination of our very minute and imperfectly preserved specimens. One point of interest, however, is the presence of several large cells, apparently ova, in all the buds examined (Pl. 3, figs. 29—31). The appearance of these buds with their contained ova closely resembles those figured by Kowalevsky in Distaplia (‘Arch. f. mikr. Anat.,’ x, 1874). Herdman also 12 W. T. CALMAN. notices the early appearance of ova in the buds in several cases (e.g. Colella), and Hjort (1. c., pp. 604 and 605) has recently called attention to the migration of the egg-cells from the parent animal into developing buds in the case of Botryllus. It is possible that we have here to do with an instance of such migration, but on the other hand, if the buds arise, as seems probable, from the “ vascular processes,” it is very difficult to see how the ova could have reached their destination. The youngest buds (PI. 1, fig. 6, and Pl. 3, fig. 29) consist of a two-layered vesicle, the space between the layers contain- ing loosely scattered cells and one or two ova. Older buds (Pl. 1, fig. 5, and Pl. 3, fig. 31) show the two peribranchial cavities developed at the sides of the branchial sac. In the latter the endostyle can be seen in sections as a pair of ridges near its anterior end. The neural tube lies dorsally to the branchial sac, but its communication with the latter could not be traced. In sections of the abdominal region (Pl. 3, fig. 30) the rudiment of the heart is seen in its pericardium, and on either side of the latter are the two “epicardial tubes,” which, as has been already stated, are in this genus absent or rudimentary in the adult, but which in Fragaroides persist as a great space running the whole length of the body. A long “tail” projecting from the hinder end of the older embryos (Pl. 1, fig. 5) is probably the outgrowing vascular process. Systematic Position.—The unnamed genus referred to by Professor Herdman as (?) ignotus, and already men- tioned above, is so similar in general appearance as well as in habitat to the form here described that the two must be nearly related, if not identical. The poor state of Professor Herdman’s specimens, which accounts for the brevity of his description, may also account for certain discrepancies, as for instance when Professor Herdman describes the tentacles as numerous and equal, and the nerve ganglion as spherical in form. Our species, however, certainly does not belong to the Polyclinidz, to which Herdman refers his (?) ignotus “from the general appearance of the colony and of the ascidiozooids.” The genus being certainly nondescript and ON JULINIA. 13 requiring a name, I give the animal at the same time a new specific one in the face of what uncertainty remains as to its identity with Professor Herdman’s type. As regards the affinities of the new genus, we are confronted by the fact that the diagnoses of recognised families appear to be somewhat artificial, and certainly do not lend themselves very readily to the reception of new forms. Taking, however, the received families as they at present go, Julinia is obviously excluded from the Polyclinide by the absence of a post- abdomen. This negative character, together with the distinct separation of thorax from abdomen, are characters common to the three families of Distomide, Didemnidz, and Diplo- somide. From the last two families our species is excluded by the absence of retractile muscles in the vascular processes, by the large anal languets (absent in Giard’s definition from the Diplosomidz), by the want of the calcareous spicules so characteristic of the Didemnidz, and by the characters of the testis and vas deferens, which latter in the Didemnide is spirally coiled round the single large testis. Narrowing down our genus accordingly to the Distomide, we find that it agrees with that family in the numerous spermatic vesicles of the testis, as well as with individual genera of that family in its incubatory pouch and large atrial languet. The incubatory pouch is common to Colella and Distaplia, while the atrial languet is absent in Colella. The general sum of the characters, then, seems to bring our genus nearest to Dis- taplia, with which latter genus it further agrees in the characters of the branchial sac with its four rows of long stigmata crossed by intermediate transverse vessels. It differs chiefly from the definition of Distaplia in the form of the colony, as well as in the ascidiozooids being embedded in the test instead of forming prominent knobs or lobes. The following is the diagnosis of the new genus: JULINIA, gen. nov. Colony cylindrical, excessively elongated; ascidiozooids com- pletely embedded in the fleshy or gelatinous test; systems 14 W. T. CALMAN, distinct, with well-developed common cloacal cavities ; buccal orifices small, six-lobed ; common cloacal orifice large, irregu- larly lobed; zooids with thorax and abdomen united by a narrow neck; atrial languet very large ; vascular process well developed; branchial sac rather large, with four rows of long stigmata crossed by narrow intermediate bars ; dorsal languets large, three in number, placed to the left of the middle line; stomach with internal longitudinal folds. Species.—Julinia australis, n. sp., with the characters of the genus. Habitat.—Antarctic Ocean. EXPLANATION OF PLATES 1—3, Illustrating Mr. W. T. Calman’s paper “On Julinia, a New Genus of Compound Ascidians from the Antarctic Ocean.” Reference Letters. atr. 1. Atrial languet. 6. Bud. 6.7. Buccal lobes. 4. s. Buccal siphon. b. sph. Buccal sphincter muscle. 6/. c. Bladder cells of test. 47. g. Blood- corpuscles. 47. s. Branchial sac. c. c/.0. Common cloacal orifice. cl. 0. Cloacal orifice. c/. . Cloacal nerve. d./. Dorsal languet. d.. Dorsal ganglionic nerve-cord. d@.s. Dorsal branchial sinus. d. ¢. Dorsal tubercle. ect. Ectoderm. ed. Endostyle. epe. Rudiment of epicardial system (?) in adult. epe. ¢. Epicardial tubes in bud. 7. Outer layer of follicular wall of ovary. jl’. Inner layer of follicular wall of ovary. gi. Subneural gland. ht. Heart. ine. p. Incubatory pouch. ixé¢. Intestine. ind. 7. Interserial lamina of branchial sac. in¢. g/. Intestinal gland. zt. d. Duct of intestinal gland. 7. m. Longitudinal muscle-bands. x. g. Nerve gland. as. Gsophagus, es. f. @sophagus funnel. ov. Ova. ovy. Ovary. ovd. Oviduct. pdr. Peri- branchial cavity. p.c. Pigment corpuscles. per. Pericardium. p. ph. d. Peripbaryngeal band. *.p. ‘‘ Raphé postérieur.” s¢. Stomach. ¢.c. “Test- cells” of ovum. ¢.c’. Test-cells. ¢. Trabecule from branchial sac to mantle. ¢r. 6. Intermediate transverse bars of branchial sac. ¢s. ¢. Tubules of testis. v.d. Vas deferens. v. p. Vascular process. ON JULINIA. 15 PLATE 1, Fie. 1.—General view of colony slightly reduced. Fic. 2.—A single system of ten individuals x 7. In the centre of the system is the irregularly-lobed cloacal orifice; on the outer side of each ascidiozooid the endostyle is seen shining through. Fic. 3.—A_ single ascidiozooid seen from the left side x 20. This figure illustrates the form of the branchial sac, the convoluted endostyle, and the relations of the incubatory pouch and atrial languet. In the abdomen the ovary and part of the vas deferens are seen on the dorsal side of the intestinal loop. Fig. 4.—Ventral view of abdominal region, showing the oblique folds in the stomach wall, and the position of the heart and part of the vas deferens in the loop of the gut. The origin of the vascular process is seen on the lower side of the intestine. Fic. 5.—Bud lying beside vascular process. Fie. 6.—Younger bud near club-shaped end of vascular process. Fic. 7.—View of part of common cloacal chamber seen from within, showing four atrial languets pointing inwards and meeting together. .In one indi- vidual of the system the oral region is also shown. The roof formed to the common cloacal chamber by these converging atrial languets is not to be seen from the outside, it being covered in its turn by a second roof of test- substance. Fie. 8.—Oral region of an ascidiozooid seen from within, showing the six- lobed buccal orifice, the hexagonal ring of twelve tentacles, the peripharyngeal band connected dorsally with the dorsal tubercle and ventrally with the endo- style; the nerve ganglion, the first row of stigmata, and the trabecule con- necting the stigmatic wall with the atrial wall are also seen. Fie. 9.—Transverse section through one half of the colony x 4, showing the position of the ascidiozooids, the distribution of the vascular processes through the matrix of the colony, and the zone of problematic vesicles referred to on p. 3 as possibly representing degenerated zooids. Fic. 10.—Cell elements from the test supposed (p. 4) to be phagocytes. Zeiss, ~zth hom. imm., 2. Fie. 11.—Spicules from test. Zeiss, D. D., 2. 16 W. T. CALMAN. PLATE 2. Fic. 12.—Internal view of dorsal portion of branchial sac, showing the four rows of stigmata, the transverse vascular bars which cross them, the unsymmetrical position of the three dorsal languets, and the interserial lamelle with which these latter are connected. Fic. 13.—The ceesophageal funnel and its connection with the endostyle. Fig. 14.—Semi-diagrammatic transverse section across the thorax at the level of the cloacal opening, showing the proportions of the endostyle and the relations of the dorsal sinus and the degenerate rudiment of the dorsal nerve- cord. Fic. 15.—Transverse section through the constricted neck between thorax and abdomen, showing the relations of csophagus and intestine to vas deferens and oviduct. Fie. 16.—Transverse section through upper portion of abdomen, showing the stomach and intestine in relation to vas deferens and oviduct, and to the heart, pericardium, and (?) epicardial tubes. Fic. 17.—Transverse section of an interstigmatic bar, showing the con- tained blood-sinus, the characters of the ciliated epithelium on its lateral edges, and of the flattened epithelium on its inner and outer sides. Zeiss, 73th hom. imm., 2. Fic. 18.—Surface view of an interstigmatic bar. Fic. 19.—Transverse section of esophagus, showing its folded walls and columnar ciliated epithelium. Fic. 20.—Transverse section of endostyle, showing the epithelial charac- ters of its lips, side walls, and central groove. Zeiss, KH, 2. Fie. 21.—Ventral section of nerve ganglion and associated organs, showing the buccal sphincter, the ganglion with origin of the cloacal nerve, the gland of Julin, and the course of its duct. The long and well-preserved cilia in the duct are all pointing inwards. Zeiss, H, 2. Fic. 22.—Transverse section of the dorsal branchial blood-sinus, showing dorsally the rudimentary dorsal nerve-cord, and laterally, near the attachment of tle sinus to the thoracic walls, its longitudinal muscular bands. LH, 2. Fic. 23.—Longitudinal section of the abdomen, showing the valvular lip of the cesophagus where it enters the stomach, and the entrance into the latter organ of the duct of the intestinal gland. Some tubules of the testis are again seen within the intestinal loop. Fic. 24.—Longitudinal section of part of the intestinal gland. The gland is seen lying against the wall of the intestine, while its duct passes off in the opposite direction towards the stomach. Zeiss, ;',th hom. imm., 2. Fie. 25.—Gregarines from stomach. Zeiss, ;';th hom. imm., 2. ON JULINIA. 7 PLATE 3. Fic. 26.—A portion of the testis, showing its racemose follicles containing spermatozoa. Zeiss, th hom. imm., 2. Fie. 27.—Part of the ovary, showing one ovarian follicle and a portion of another, both in connection with the oviduct. D D, 2. Fic. 28.—Peripheral layer of ovum with its test-cells, surrounded by the two-layered follicular wall. Zeiss, 4,th hom. imm., 2. Fic. 29.—Transverse section of a young bud, showing included ova. Zeiss, D D, 2. Fic. 30.—Transverse section of abdominal region of older bud. In this section we see, besides the cesophagus, the intestine, and a large ovum, the developing heart within its pericardium and the two associated epicardial tubes. D D, 2. Fic. 31.—Optical longitudinal section of a bud of about the same age as that represented in Fig. 30. The branchial sac and the paired rudiments of the atrium are here shown. D D, 2. Fic. 32.—Transverse section of a vascular process, with its surrounding space and adjacent portions of the test-substance. Large bladder-cells, small test-cells, and pigment corpuscles are seen within the matrix of the latter. VOL. 37, PART 1.—NEW SER. B ¥E AVANT? : Ff oy Ay} iraiaii Yaa etre {pl Trip BAO iE Aber i il iL ho mits | } ; 2G Pitt et) ‘ pl] ~ woygit fs 4 hy F wil Se ivi 5 fj ey JCS i> Dl wow? Yi tO) | =e ik é 1) [ Phare ai i j toac thi SAP AT ah ol ae gibb 7st a ea ein off 2h ieee hl ‘ 7 f Uae Deitel ee ie chthobiok fi ep 7 wauily fohulsde ware inl 2s ae . aU) ae Praia ll habs a) rt | BaD ne iy | pili Did Mic pend ew ae hi ives: B pit) iih4 : j (ele } ‘i Pew EOE P eb ) ¥ Bia Pe Ati ’ aS rT Ha i he Pabsialh 04, fon ’ ee Tea ih ae, P = ee | | s rt as ME H/ f 4.84, u Rory 74 ee =e fiaie ray isu nau! } i ; i eee) Al Wael | TT ita E i. i, 3 - , - = ‘2 . - i 2 ie z e q eee oe of : os e ¥ vv = : ' re) si" ' 4 = } @ g ° , v ye ° q 7 oe nih j : Be : oe ae Fi be : HERMAPHRODITISM IN MOLLUSOA. 19 Hermaphroditism in Mollusca. By Dr. Paul Pelseneer (Ghent). With Plates 4—6. ConTENTS. I. HermMapHropitism In Moutusca. 1. Amphineura. 2. Gastropoda. 3. Lamellibranchia. II. Puytoceyetic Evotution or THE HERMAPHRODITE GLAND IN Mo.tusca. 1. Gland with hermaphrodite acini. 2. Gland with separate male and female acini. 3. Gland with separate male and female regions. 4. Separate male and female glands in the same individual. ILI. Puystonocicat Evotution oF HERMAPHRODITISM IN THE MOoLLUscaN INDIVIDUAL. IV. Ortcin or HERMAPHRODITISM IN MoLuusca. 1. Hermaphroditism has arisen from the unisexual state. 2. Hermaphroditism has been grafted on the female organism. V. Conclusions. I. HermMaruropitism 1n Mo.uusca. Hermarxropirism is found to occur in every class of Mol- lusca except the Cephalopoda and Scaphopoda. 1. Among Amphineura, in all the Neomeniide. 2. Among Gastropoda, in four genera of Streptoneura (Valvata, Onchidiopsis, Marsenina, and Entoconcha), and in all the Euthyneura (Opisthobrauchia and Pulmonata). 3. Among Lamellibranchia, in various species of Pecten, 20 PAUL PELSENEER. Ostrea, and Cardium; in Entovalva, the Cycladide, the Poromyide, and all the Anatinacea. 1, Amphineura. (1) Neomeniidx.—In each of the two contiguous herma- phrodite glands of these animals the ova arise in the axial half of the organ, and the spermatozoa in the lateral half (1). (2) Cry ptochiton.—According to Middendorf (2) this genus is hermaphrodite. I have, however, been able to obtain a specimen in alcohol of this form (C. Stelleri, from the North Pacific), and have found it to present a distinctly unisexual condition, as in Chiton. I have also satisfied myself that the sexes are separate in Chitonellus larveformis (= fasciatus) and Cryptoconchus porosus (or monticu- laris). Thus the diccious condition is universal in the Poly- placophora. 2. Gastropoda. (1) Valvata.—The hermaphrodite gland of this genus is formed of acini, ali producing ova and spermatozoa (8). (2) Marsenina and Onchidiopsis (of the family Lamel- lariidze).—While the hermaphroditism of Valvata has been recognised by a fair number of zoologists, that of the two genera just named has only been affirmed by Bergh (4). As the number of monecious Streptoneura is extremely limited, an examination of these two genera was desirable in order that Bergh’s discovery might be confirmed and his state- ment of facts extended. I have only succeeded in obtaining a single specimen of Onchidiopsis greenlandica, and upon it I have made the following observations. The genital gland occupies the posterior part of the visceral mass. Its duct is provided with an accessory mass formed of closely packed, narrow, tubular ceca; it then bifurcates, and the more lateral branch (the right), after first receiving a large bent pouch (fig. 1, rv), immediately thickens before opening to the exterior, and again receives on its right side a large pouch of flattened form. The aperture of this branch is situated in HERMAPHRODITISM IN MOLLUSCA. 21 the mantle cavity, in front of the anus and more to the right ; it is the female orifice (v1). The other branch immediately presents a glandular mam- millated mass with closely packed lobules, and then extends forward under the integument over the neck of the animal as far as a considerable projection of the body-wall on the right side, the penis, which it traverses from end to end (fig. 1, v). The structure of the reproductive apparatus thus differs from that of Valvata in the presence of glands upon the her- maphrodite portion of the duct, and by the absence of a dif- ferentiated “ uterus ”’ in the female portion. The glands of the hermaphrodite duct (fig. 1, x) do not appear to me to be vitellogenous or albuminiparous glands (judging, at least, from the imperfectly preserved specimen which I have examined). I regard them rather as vesicule seminales. As for the two appendages of the female duct, the elongated pouch is a receptaculum seminis (or poche copulatrice), and the flattened pouch (fig. 1, 111) is the mucous or jelly gland. Lastly, the glandular mass of the male duct can be termed the prostate, as in other hermaphrodite Gastropods. If the presence of two genital apertures could not, in itself, demonstrate the hemaphroditism of Onchidiopsis, the struc- ture of the genital gland is sufficient to remove all doubts. This gland is formed of parallel ceca, bifid at their extremities (fig. 3): these terminal divisions are ovogenous, while the proximal portions of the duct are spermatogenous. As a result of this arrangement the gland appears to be composed of a superficial or external female portion, and of a deeper or central male portion. However, these two regions are not demarcated with any regularity, and in the middle portion male and female acini can be seen in sections lying side by side (fig. 2). At all events, the products of the two sexes either do not arise in the same cecum or they do not arise in the same region of a cecum. According to Bergh’s observations, Marsenina appears to 22 PAUL PELSENEER. be constructed upon the same plan,—that is to say, the genital czeca are female in their terminal portion. (3) Entoconcha.—This genus, so degraded by parasitism, has unfortunately not yet been studied by the method of serial sections. As is known, there is a female genital gland in the middle region of the body, and a number of spermatogenic capsules further back towards the posterior orifice of the body. (4) Bulloidea.—The genital gland is here formed of acini, all hermaphrodite (Bulla, Limacina, &c.); but in certain forms (Acton, Pelta, Lobiger) I have found it composed of distinct male and female acini. (5) Aplysioidea.—The same acini produce spermatozoa and ova (Aplysia). I have found, however, that the male and female acini are distinct in certain specialised forms ; the former (male) occupy the central, and the latter the peripheral region (Pneumonoderma, Clinopsis). (6) Pleurobranchoidea.—The genital gland is uniformly composed of hermaphrodite acini in Umbrella. But in Tylodina and all the Pleurobranchide I find that the acini are either male or female, and that the latter open into the former. (7) Nudibranchia.—Meckel was the first to notice in certain Nudibranchs that the genital gland is formed of male and of female acini; but he supposed that these acini of different sexes were without communication with one another, and that in the genital ducts there was a male canal embedded in the female one (5). Nordmann immediately afterwards showed that in Tergipes the ovular acini open into capsules full of spermatozoa; but he took the entire hemaphrodite gland for an ovary merely, and held the sacs full of spermatozoa to be “ poches de fécon- dation” (6). It was R. Leuckart who first determined and correctly inter- preted the constitution of the hermaphrodite gland of Nudi- branchs, especially Eolis (7),—i. e. peripheral acini exclusively ovular, opening into central chambers producing spermatozoa. His interpretation has been confirmed since for a great HERMAPHRODITISM IN MOLLUSOA. 23 number of genera by various authors (Hancock, Bergh, Trinchese, and myself); the same is the case with those Nudibranchs in which the male and female acini are easily distinguished under a lens of low power (e.g. Fiona). It must be noticed, however, that in Eolis (Cory phella) Landsburgii the acini of the hermaphrodite gland produce ova in their distal portion and spermatozoa in their proximal portion (8) ; this arrangement has been recognised as general in all the Elysioidea (Cyerce, Herma, Elysia, Lima- pontia). (8) Pulmonata.—lIn these Gasteropods the genital gland is formed of hermaphrodite acini both in the Stylommatophora (e.g. Helix) and in the Basommatophora (e.g. Auricula). In Siphonaria (Basommatophora), according to Haller (9), each acinus of the hermaphrodite gland is exclusively of one sex, either male or female. In S. Algesirz, which I have studied, I have observed that the conformation of the gland is precisely analogous to that found in Onchidiopsis, the Pleurobranchide, and the Nudibranchia,—that is to say, the peripheral female acini open into the more centrally situated male acini. However, this conformation is not so entirely different from that presented by the other Pulmonata. I know cases, in fact, in which the wall of the genital gland already shows a distinct sexual differentiation upon the two sides of the follicles, and in which the female side exhibits projections which are the rudiments of acini of this sex (Amphibola). It follows, then, from what has been said above, that examples of the various possible modes in which the genital gland is constituted are to be found side by side in all the sub- groups of hermaphrodite Gastropods. 8. Lamellibranchia. (1) Ostrea.—The question of sex in oysters has long been a subject of controversy, and its solution, which presents decided difficulties, is not yet universally recognised. 24 PAUL PELSENEER. The first point capable of immediate demonstration is that there are both hermaphrodites and dicecious oysters. a. Hermaphrodite Oysters. a. Ostrea edulis, Linné, the hermaphroditism of which was first demonstrated by Davaine (10), and subsequently confirmed by Lacaze-Duthiers, Hock, &c. b. Ostrea stentina, Payr. (= plicata, Chemnitz) ; also recognised as hermaphrodite by Lacaze (11). c. Lastly, it has been recently stated that the oyster of the N.W. coast of America is also hermaphrodite ; this, I suppose, is Ostrea lurida. . In the genital glands of hermaphrodite oysters the sperma- tozoa and the ova arise in the same acini, but at different times, so that the products of the two sexes are not often to be seen in the same individual. This perfect alternation is the most striking and distinctive characteristic of the heramaphroditism of oysters ; it explains how it has been believed, and how at first sight it would still be possible to believe, that the oyster is unisexual. The male products are the first to appear. This protandry was discovered by Davaine (13), and confirmed by P. J. van Beneden (14). B. Diecious Oysters. a. Ostrea virginica, Lister, from the E. coast of the United States (15). b. Ostrea angulata, Lam. (= lamellosa, Broc.), the ‘< Portuguese” oyster, from the Atlantic (16). c. Lastly, I have made out the separation of the sexes in a third species, O. cochlear, Poli, from the Mediterranean, which belongs to the same subdivision (Gryphza) as the preceding form. I have not had the opportunity of studying this species alive, but I have had numerous specimens of it of very different sizes, collected at different times of the year. Of all these there was not a single individual which pre- HERMAPHRODITISM IN MOLLUSCA. 25 sented both ova and spermatozoa in the genital gland at the same time. Each specimen had the glands either full of the products of one sex exclusively (fig. 8), or else almost empty (fig. 7). Nevertheless it is impossible to suppose that they repre- sented the successive stages, male and female, of an hermaphro- dite condition: in the first place, because the individuals with male glands were no smaller than those with female glands,— there were female specimens smaller than the males, and males and females of every size; in the second place, because the appearance and conformation of the genital glands differs con- siderably according to the nature of the contents, as in Lamel- libranchs of different sexes. The glands with spermatozoa are formed of ramifications having an appreciably constant dia- meter (fig. 4, 11) ; the glands with ova are more lobulated, and so present a distinct appearance, which shows that O. cochlear is dicecious, like O. virginica and O. angulata. I will add a word here upon the genital apertures of O. cochlear. They are asymmetrical, that of the left being the more anterior (fig. 6, rx), considerably in front of the adductor muscle. The “fente uro-génitale,” discovered by Hoek in O. edulis (17) (where the genital gland opens into the urinary slit), has not in this species the simple appearance which it presents in O. edulis. The genital and renal orifices, although adjacent, are quite distinct ; the genital is more anterior than the renal, less distant from the median plane (fig. 7, v), and directed outwards, whilst the renal aperture is directed towards the axis (fig. 7, x). (2) Cardium oblongum. Lacaze-Duthiers discovered the hermaphroditism of a species of this genus, C. norvegicum, from the Atlantic. I have further observed this condition in a Mediterranean species, C. oblongum, which is grouped with C. norvegicum in a special division or sub-genus (Levicardium). In Cardum oblongum the different acini are each of one sex, either male or female; but the acini of the same sex are 26 PAUL PELSENEER. not confined toa particular region (fig.9) asin Onchidiopsis, for example, although the female acini open into the male acini as in the last-named genus. I have never, however, noticed ova and spermatozoa in the same cul-de-sac, as Lacaze- Duthiers found to be constantly the case in C. norvegicum (18). (3) Entovalva. The structural details of the hermaphrodite genital gland are not known. There exists a single gland, not differentiated into sexual regions, on each side (19). Probably, therefore, it is constituted like that of Ostrea and Cardium. (4) Pecten. The greater number of species of this genus hitherto exa- mined have been found to be hermaphrodite, viz. P. glaber (20), P. Jacobeus (20a), P. maximus (21), P. opercularis (22), P. irradians and magellanicus (28). I can myself vouch for the hermaphroditism of P. glaber and maximus. Moreover I have observed that P. flexuosus, Poli, from the Mediterranean, also has the two sexes united in the same individual (fig. 10). On the other hand, among all the species examined, I have only met with two diccious forms, viz. P.inflexus, Poli, from the Mediterranean, and P. varius, Linné, from the Atlantic, in which latter form this fact was already known (Humbert, 23a). The hermaphroditism of Pecten is recognised with unusual facility. The anterior part of the visceral mass may be readily seen from the outside to be whiter than the posterior part ; the whiter part is the male region. Examination of fresh material, or a section, shows at once that the genital products of different sexes are formed in different regions of the herma- phrodite gland (fig. 10). At the same time, as is already known in the case of certain species, there is only one common genital orifice on each side (opening into the nephridium), and only a single genital duct, which ramifies in the various parts HERMAPHRODITISM IN MOLLUSCA. 27 of the ovotestis (P. glaber, maximus, flexuosus). The male region is thus nearer the genital aperture than the female region. (5) Cycladide. a. Cyclas.—The hermaphroditism of this genus has been known from the time of v. Siebold (24). But the conformation of the genital organs has not been completely described even by subsequent authors (25) ; Stepanoff has merely made known that one portion of the gland is male and the other female. In C. cornea (fig. 12) the genital gland has a superficial position upon each side of the posterior part of the body. It is of scarcely any extent except in length (from liver to nephridium), its height and depth being inconsiderable. It is transversely divided into two portions of different appearance and size, separated by a constriction in which there is merely a ciliated duct uniting the two portions (fig. 11, 11). The most anterior division is more elongated and more appreciably lobulated than the other; it extends as far as the liver. This is the male region of the gland ; the other is exclusively female (he. LL, 11). The hermaphrodite gland is thus seen to be here divided into two portions of different sexes ; and these are no longer in immediate contact (as in Pecten), but already separated from one another, though connected by the intermediation of a duct. Lastly, the posterior (female) portion of the gland is con- tinued backwards by an hermaphrodite duct which terminates at the genital orifice. The latter, which has escaped notice hitherto) is situated outside the visceral commissure, near the most ventral point of the nephridium (where the aperture of this organ is placed), and in front of the posterior retractor muscle of the foot (fig. 12, v1). B. Pisidium (26) and Corbicula (fide v. Jhering in litera) are also hermaphrodite. The conformation of the genital gland appears to be the same in them as in Cy clas (27). 28 PAUL PELSENEER. (6) Anatinacea and Poromyide. Several years ago (28) I made known the hermaphroditism of Tbracia, Lyonsia, Lyonsiella (Anatinacea), and of the allied family Poromyidez (Septibranchia) ; and at the same time I confirmed its existence in Pandora, in which it had already been once affirmed (29). Having thus demonstrated the monecious nature of all the genera of Anatinacea which I had examined, I was led to believe in the hermaphroditism of the entire group, a conclusion which accorded well with Lacaze’s observations upon Aspergillum (80). Since that time I have been enabled to study one other genus of this group, Clavagella, although, in spite of nume- rous efforts, I have not succeeded in obtaining either Anatina or Pholadomya, which appear to be rare. I have observed the same monecious arrangement in Clavagella as in the other Anatinacea previously studied. There can accordingly be no doubt as to the hermaphroditism of this group, since no genus belonging to it has yet been found to be dicecious. Like all the Anatinacea, Clavagella possesses two testes, two ovaries, and four distinct genital apertures. The two testes are symmetrically placed in the pedal projection (fig. 15, vii), i.e. in the ventral part of the body; the two ovaries, which are much more voluminous, compose almost the entire posterior visceral mass as far as the nephridia (figs. 15 and 16). The genital pores are situated behind; those of the testes on the sides of the base of the foot, ventrally or internally to the visceral commissure (fig. 14, vir); those of the ovaries a little further back, dorsally or externally to the same commis- sure (fig. 14, Iv). II. Poytocenretic EvoLution oF THE HERMAPHRODITE Gianp In MOLLUSCA. If any other group of hermaphrodite animals is taken into consideration, one is struck by the general uniformity in the structure of the genital glands. Inthe Mollusca, on the other HERMAPHRODITISM IN MOLLUSOA. 29 hand, as has been shown above, there is a very great diversity in the structure of the hermaphrodite gland. Indeed, it was Leuckart who was the first to recognise that among the Euthyneurous (or hermaphrodite) Gastropods the genital gland is not always formed in the same manner, and that different types of structure are recognisable in it. The facts known to us to-day show that such is the case not only for the Gastropoda Euthyneura, but also for all the remaining groups of hermaphrodite molluscs—the Amphineura, Gas- tropoda Streptoneura, and Lamellibranchia. Four different principal types can be distinguished in the conformation of the hermaphrodite gland of molluscs, but certain forms are transitional between one type and another. 1. The undifferentiated gland, i.e. with acini com- pletely hermaphrodite.—As development shows (vide infra), the most simple condition of an hermaphrodite genital gland is that in which the gonadial wall is still undifferentiated, and produces ova and spermatozoa side by side. This condi- tion is exhibited in— (1) Valvata. (2) A great number of Tectibranchs (e.g. Bulla, Aplysia, Umbrella). (3) Almost the entire group of Pulmonata. (4) Ostrea edulis and stentina. A commencement of specialisation is observable in Neo- meniidz, where each genital gland generally gives rise to male products towards its lateral face, and to ova towards its axial face. Amphibola among Pulmonates (vide supra) also fur- nishes a transition to the following condition. 2. Gland with separate male and female acini (but as yet without separation into different male and female regions).—After the undifferentiated condition comes that in which the gonadial surface shows a clearly marked specialisa- tion into distinct male and female acini,—the acini, however, not forming regions of different sex. 30 PAUL PELSENEER. This arrangement exists in— (1) Various Tectibranchs (e. g. Lobiger, Pelta). (2) Pleurobranchide, Tylodina, and Nudibranchia (except Elysioidea). (3) Siphonaria. (4) Cardium oblongum. It must be noticed here that the female acini open into the male acini, and that the acini of the same sex tend to group themselves together, as may already be observed in various Nudibranchs (where the female acini are the most “‘ eccentric,” i.e. superficial). This arrangement is clearly marked, and furnishes in Onchidiopsis and Pneumonoderma (vide supra) a transition to the next condition. 3. Gland with separate male and female regions (with a common duct).—The type which now presents itself is that in which the acini of the same sex are all united together in such a way as to constitute in the hermaphrodite gland a male and a female part distinct from one another, these two parts having, nevertheless, a single genital aperture and a common duct. This conformation is characteristic of— (1) The hermaphrodite species of the genus Pecten, where the male and female regions are contiguous and form an undivided hermaphrodite gland (fig. 10). (2) The Cycladidz, where it is already observable that the two regions are fairly separated, and only connected by their duct (fig. 12), which forms a transition to the next type. 4. Male and female glands in the same individual entirely distinct from one another, and with special ducts.—The acini of each sex can form a region absolutely separate from that constituted by the acini of the other sex, these two regions each having their special duct, and so form- ing veritable testis and ovary. At the same time the vas deferens and the oviduct may open into a common orifice (Poromyidz), or there may be no common orifice for the two glands. This last condition, which HERMAPHRODITISM IN MOLLUSCA. 31 represents the highest specialisation of the hermaphrodite genital gland, is only found in the Anatinacea among Lamelli- branchs, and in Entoconcha among Gastropods. III. PuoystotoaicaL EvoLuTion oF HERMAPHRODITISM IN THE Mo.uuuscan INDIVIDUAL. The hermaphroditism of molluscs is not, so to say, self- sufficient ; in other words, the eggs of one individual have to be fertilised by the spermatozoa of another. Speaking generally, the two kinds of products are not ripe at the same moment in the same individual: an interval of greater or less duration separates the two periods of maturity of the different sexual elements. : Leuckart was the first to make known that in certain Opis- thobranchiate Gastropods the period of male maturity precedes that of female maturity (31), i.e. the hermaphroditism of these forms is protandric. This protandry ought to be regarded as a general phenome- non in Euthyneurous Gastropods. This is notoriously the case in the Pulmonates (32), and it has been recognised in the various Opisthobranchs which have been studied from this point of view, viz. Lobiger (33), the Thecosomatous ‘ Ptero- pods,” e.g. Clio striata (84), Cymbulia (35), Desmopterus (86) ; Nudibranchs, among which I have observed it in Holis and Elysia; and lastly I may add Clione limacina (Gym- nosomata), in which I have noticed that individuals of a length of 15 millimetres (or less) do not as yet show any ova in their genital gland, but stages in the development of spermatozoa only. Similarly in Entoconcha testicular capsules are only to be observed in individuals in which the eggs are not yet well developed (37). For Neomenia, among Amphineura, protandry is equally probable (38). In Lamellibranchs no general statement can be made with certainty. I have been unable to examine young individuals 32 PAUL PELSENEER. of Anatinacea, or hatched specimens of Cyclascornea smaller than 4 millimetres: at this size both spermatozoa and ova are already to be found, and not infrequently even developing eggs among the gills. But according to the old observations of Davaine (89) and P. J. van Beneden (40), Ostrea edulis is protandric. The alternating activity of the two sexes in the same herma- phrodite Molluscan individual is a perfectly certain fact, and explains various mistakes like that of Saint-Loup, who believed he had found a unisexual organisation in Aplysia associated with external sexual characteristics (41). It must, however, be remarked that this succession of the two sexual conditions may be more or less rapid according to the particular genera under consideration, and that, conse- quently, the alternation is better marked in certain organisms than in others. It seems to me to be more appreciable where the hermaphroditism of the genital gland is most complete (i.e. where the ova and spermatozoa are produced at the same spot, as in Ostrea, Aplysia, &c.) than where there are acini or regions of different sex (Nudibranchs, &c.). As for the general fact that the male precedes the female activity, even in the absence of physiological observations it might have been deduced from the morphological constitution of the genital glands in the adult of a great number of herma- phrodite molluscs. In those forms, in fact, which possess acini or regions of each sex, the male part is always nearest the efferent ducts or the genital aperture: the male products will accordingly be the first to reach them. This arrangement exists in Pecten, Nudibranchs, Pleurobranchide, Onchidiop- sis, Entoconcha, Pneumonoderma, Clionopsis, &c. Cyclas alone constitutes an exception,—a fact for which I can offer no explanation. For the rest, our knowledge of the physiological evolution of individual hermaphroditism in other groups of the animal kingdoms shows that protandry is general in all forms which have been examined from this point of view, e.g. Sponges, Plathelminthes (Trematodes and Cestodes), hermaphrodite HERMAPHRODITISM IN MOLLUSCA. 33 Nematodes (Pelodytes, Rhabdonema nigrovenosa), Myzostomide, parasitic Isopods, Ascidians and Salps, Myxine, Chrysophris, &c. IV. Ontocenetic Evo.ution oF tHE HERMAPHRODITE GLAND IN Mo.uuvusca. According to the theory of the sexuality of the embryonic layers, the female reproductive elements should be of endo- dermic, and the male elements of ectodermic origin. As regards the Mollusca this view has been maintained by Fol (“ Pteropoda”’ Thecosomata, 42). But this theory has not been confirmed for any other mollusc, even for any of the nearest allies of the “ Pteropods,” viz. the Tectibranchs (e. g. Aplysia). I have not had an opportunity of studying the fresh larve of Thecosomatous ‘‘ Pteropods ;” but in sections of various pre- served larve I have been able to make out that the “corps pyriforme” (which, according to Fol, is the testicular part of the future hermaphrodite gland) neither has the structure of a testicle nor contributes at all to the formation of the genital gland. The latter is unique from its origin, as it is in its final condition. The same is the case in the other molluscs hitherto studied. However, according to Trinchese (48), in Bosellia (a near ally of Elysia) the ova are not produced in the same part of the body as the spermatozoa, although this author is unable to explain how the male and female parts of the hermaphrodite acini eventually combine with one another. Now in Elysia, specimens of which I have examined at every age from this point of view, I find that the first ova arise in the same acini in which spermatozoa alone were present in the stage imme- diately preceding. Sothat it is impossible in this case to hold that the hermaphrodite gland results from the fusion of two male and female parts of independent origin. The same fact may be observed equally well in the ‘“‘ Pteropod” Clione (vide supra). VOL. 37, PART 1.—NEW SER. c 34. PAUL PELSENEER. On the other hand, in a form where the male and female acini are sharply separated in the adult state, viz. Pelta (= Runcina), I find that in the young individual ova and spermatozoa arise side by side on the walls of a common pouch. I conclude from this that the undifferentiated condition of the hermaphrodite gland is the most primitive, and the morpho- logical observations recorded above (II) lead also to the same conclusion. The Nudibranchs and Pleurobranchide (with acini either male or female) are more specialised than U mbrella (with hermaphrodite acini) ; Onchidiopsis (with acini either male or female) is more specialised than Valvata (with hermaphrodite acini); Pneumonoderma (with acini either male or female) is more specialised than Aplysia (with hermaphrodite acini), &c.; lastly, those Lamellibranchs with male and female glands altogether separated (Anatinacea) are the most specialised (cf. the closure of the mantle, the reduc- tion of the foot, the complexity of the gill, involving the loss of the external layer of the external branchial lamella). In direct opposition to the theory of the sexuality of the embryonic layers, we find that the hermaphrodite gland has a single origin in the mesoderm (which is itself endodermic), not only in the Opisthobranchs and Pulmonates hitherto studied—where the gland is not yet divided into regions of different sex,—but also in Cyclas (44), where two distinct male and female regions exist (fig. 12, v and xiv). The same mesodermic origin is, moreover, to be observed in the case of both the male and female glands of those forms with separate sexes, such as Chiton, Paludina, and the Cephalopods. These results are also in complete accord with what one finds in other groups of hermaphrodite Invertebrates, e. g. Oligocheta (45), Sagitta, Turbellaria, Plathelminthes, Hirudinea (46). Lastly, the fact that the wall of the genital gland is in con- tinuity with the coelomic epithelium (mesoderm) in Nuculide (Lamellibranchs), Neomeniide (Amphineura), and Cephalo- poda, shows that the genital gland, whether male, female, or HERMAPHRODITISM IN MOLLUSCA. 35 hermaphrodite, is an organ possessing a single mesodermic mode of origin, as in all the other Triploblastica. V. Oricin oF HERMAPHRODITISM IN MoLLUSCA. It has just been shown that, both from the phylogenetic and ontogenetic points of view, the hermaphrodite condition with separate male and female glands is the most specialised, and the undifferentiated hermaphrodite condition (with gland producing spermatozoa and ova at the same spot) the most archaic. Now this last state is that which most nearly approximates to the unisexual condition, since it only differs in the supplementary production of the elements of the other sex,—a phenomenon which is sometimes to be observed as an abnormality in dicecious molluscs (e.g. Anodonta and Ampullaria). The question may, then, be asked in the case of the Mollusca, 1. Whether the hermaphrodite state is not derived from the unisexual; and, in the event of an affirmative reply, 2. Upon which of the sexes the hermaphrodite condition has become established. In the following pages I shall try to show— 1. That hermaphroditism is not a primitive arrangement in the Molluscan phylum (47), and that it has been derived from the unisexual state. 2. That it has become superimposed upon the female condition. : 1. Hermaphroditism has been derived from the unisexual condition.—Let us consider separately the classes in which hermaphroditism is found and those in which the dicecious state alone exists. i. In the classes in which hermaphroditism exists it is abundantly clear that the forms with separate sexes are the most archaic, especially in the conformation of their repro- ductive apparatus (absence of special duct, accessory gland, and penis) : a. Gastropoda Docoglossa and various Rhipidoglossa: the genital gland opens into the kidney. 36 PAUL PELSENEER, B. Lamellibranchia Protobranchia (Nucula, &c.): the genital gland opens at the junction of the ccelom (pericardium) and the kidney. ii. If we consider now the two classes where only the dicecious condition exists, we find that they have preserved many primitive traits, especially in the reproductive apparatus : the genital gland opens into the celom (Cephalopoda) or into the kidney (Scaphopoda). These facts show that in the Mollusca the dicecious con- dition has preceded the moneccious, and by no means (as Rouzand has tried to make out, 48) that “les Gastropodes unisexués se présentent réellement comme les descendants des Gastropodes hermaphrodites, plus ou moins analogues 4 ceux qui vivent encore de nos jours.” The molluscs which are certainly the most archaic are dicecious. Their genital organs present a conformation analo- gous to the primitive disposition observed in the ontogeny of other forms, which can on no account be regarded as the last term of a retrogressive evolution—evolution not being reversible. . But hermaphroditism has been able to establish itself, re- placing the dicecious state in Mollusca of different grades of specialisation— a. In forms where the genital glands open into the ccelom (Neomeniidz) ; B. In forms where these glands open into the kidneys (Pecten) ; c. In forms where these glands have special ducts, ac- cessory glands, copulatory organs, &c. (Onchidiopsis, Euthyneura, &c.). 2. Hermaphroditism has been established upon the female organism.—The hermaphrodite condition, then, is secondary in Mollusca. As for the organisation which has given rise to it, I believe it to be that of the female. Comparative study of the genital ducts and of their develop- ment lends support to this opinion, as we shall at once see. i. In Lamellibranchs, when there is a separate male and HERMAPHRODITISM IN MOLLUSCA. 37 female orifice (Anatinacea), and when the visceral commissure (“connectife cérébro-viscéral”) is sufficiently superficial, the female orifice is found lying outside this visceral commissure like the genital orifice—be it male or female—of all the Lamellibranchs (fig. 13). This orifice is accordingly not a new formation. But the male orifice is within this commissure (fig. 14), and so presents relations to which no counterpart can be found elsewhere ; it is consequently a new formation, and if it was necessary for a male orifice to be produced as a new formation it is evident that the starting-point of the herma- phrodite condition must have been supplied by the female organism, ii. Similarly in hermaphrodite Gastropods the male products are always expelled by a cenogenetic penial orifice, which is wanting in the archaic diccious forms (Rhipidoglossa, —as in the Amphineura and Scaphopods), while the female orifice in the same forms lies in the place occupied by the genital orifice (whether male or female) of archaic Gastropods which possess no penis. And when an approximation of the two apertures, male and female, takes place—the former approximates to the latter in Arion; vice versa in other cases—the disposition is secon- dary, and is accompanied by an increased complexity of the ducts and their appendages. In the most primitive cases the two orifices are always widely separated (Bullide among Opisthobranchs; Auriculide among Pulmonates). In the development of the Pulmonata the penis and vas deferens are new formations (48a) which arise secondarily, the vas deferens only communicating at a late stage with the genital duct properly so called (physiologically hermaphrodite), which leads to the female orifice; that is to say, the genital organs of Pulmonata develop as female organs, which are even- tually modified as as to become hermaphrodite (48b). The opinion formulated above (that in Mollusca herma- phroditism is grafted upon the female organism) is confirmed by the following fact. In hermaphrodite molluscs, whenever 388 PAUL PELSENEER. individuals accidentally present a unisexual genital apparatus, they are always female; it is always the female sex which reappears. a. Clio (Hyalocylix) striata, without penis, cited above (III, note, according to Schiemenz). 6. Cymbuliopsis calceola, without penis (48c). c. Agriolimax levis, without penis, having never been male (adult female; 49). d. Helix aspersa, without vas deferens, penis, and flagel- lum (49a). e. Arion intermedius, without male reproductive organs (49b). VI. ConcuusIons. Granted, then, that in Mollusca hermaphroditism has suc- ceeded a unisexual state, and that it has been superimposed upon the female condition, is this process paralleled in other groups ? 1, The origin of hermaphroditism in a unisexual condition.—I have not been able to make any special investi- gations upon hermaphroditism in the various groups of the animal kingdom. It seems to me, however, that if we examine, by the comparative method, the data which we already possess upon the subject, we must come to the conclusion—opposed to the ordinarily received opinion, which possesses the authority of the names of Huxley, Gegenbaur, Haeckel, Giard, Claus, &c.—that hermaphroditism is secondary, and succeeds a pri- mitively dicecious state. In fact, like myself, who have reached this result for the Mollusca, so Beard and Delage have already formulated this Opinion as concerning two other groups, the Myzostomide (49e) and Cirripedes (50). And Fritz Miller, after a more general consideration of the subject, also argues against the primitive nature of the hermaphrodite state, and shows that in the majority of groups the most archaic forms are unisexual (51). A survey of the various subdivisions of hermaphrodite HERMAPHRODITISM IN MOLLUSCA. 39 animals shows that hermaphroditism is characteristic almost always of specialised forms (52). I will more particularly mention fixation, parasitism, fluviatile or terrestrial life, as specialisations accompanied by hermaphroditism, as the fol- lowing few examples show : Pecten, Ostrea, Aspergillum, mensals Fixed (58) or very Clavagella; various Serpulide sedentary (54), Cirripedes, Myzostomide, Ascidie. Hermaphrodite Beene Entoconcha, Entovalva, Ces- animals. EME 0 oa todes, Trematodes, Hirudinea, cer- tain Isopods, Myxine, Fluviatile or ter- : Valvata, Pulmonata, Oligocheta. restrial In support of the opinion that hermaphroditism is a spe- cialisation of the separation of the sexes, it may be remembered that in a great number of unisexual glands individual variations occur in which the elements of both sexes are produced, one of the two kinds of element being an abnormal product. I will cite merely a few instances: i, Batrachiaus: female frog (55). il, Fishes: female herring (56). iii. Molluses: Ampullaria, Anodonta. iv. Cheetopods: ova in the testicle of Lumbricus (57), - v. Crustaceans: female Apus (58), &c. _ This phenomenon makes it possible to understand that the hermaphrodite state can easily establish itself in certain circumstances where it is useful for the same individual to give rise to the products of both sexes (59). 2. The establishment of hermaphroditism on the female condition.—I cannot at present affirm that her- maphroditisim has everywhere grafted itself upon the female organisation; nevertheless the phenomenon ought to be exhibited elsewhere than among the molluscs. . As a matter of fact, there are several other groups in which the male sex is preserved in the form of individuals, either degraded or not, when there are no longer any females, 40 PAUL PELSENEER. but only normal hermaphrodite individuals. This is the case with— i, Various Myzostomide. ii. Certain parasitic Isopods, e. g. the Cryptoniscide. ili, Various Cirripedes, e.g. Scalpellum, &c. And the study of one of these Cirripedes has led Delage to the same conclusion as that which I have deduced from the study of hermaphrodite molluscs: “Au début, les Sacculines étaient des animaux a sexes distincts, dont les femelles sont devenues hermaphrodites beaucoup plus tard” (60). Lastly, the same seems to be true among osseous fishes in Serranus, according to Ginther (61), viz. that the herma- phrodite individuals are transformed females. This has been confirmed by Brock, who has specially studied the genital organs of fishes. He concludes that “die hermaphrodi- tischen Knochenfische sind weibliche Individuen, in deren Ovarium sich an Stelle einiger Ovarial- lamelle ein Hoden sich entwickelt hat (62). And, on the other hand, the vas deferens of Serranus is not homo- logous with that of other osseous fishes (68) ; it is, accordingly, like that of the Mollusca Anatinacea and Pulmonata (vide supra, v, 2, ii), a new formation grafted on the female condition. In these different cases the establishment of the hermaphro- dite condition has been probably characterised by the follow- ing successive stages :—Production of spermatozoa in a part of the ovaries; reduction of the size of the males and of their number (hyperpolygyny) ; then complete replacement of the female by the hermaphrodite form; and, lastly, the total dis- appearance of the degraded males. Summing up. a. The study of Mollusca, Myzostomide, Crustacea, and Pisces shows that in these groups the separation of the sexes has preceded hermaphroditism ; various cases in other groups tend to show that this is true universally; and the same con- clusion applies to plants. HERMAPHRODITISM IN MOLLUSCA. 41 B. In Mollusca, Crustacea, and Pisces, at least, hermaphro- ditism is grafted upon the female sex. 18. 19. BIBLIOGRAPHY. . Husrecut.— Dondersia festiva, gen. et sp. nov.” (‘ Donders-feest- bundel,’ 1888), pl. ix, fig. 6. Pruvor.— Sur Vorganisation de quelques Néoméniens des cdtes de France,” ‘ Arch. d. Zool. Expér.,’ sér. 2, t. ix, pl. xxx, fig. 52. . Mippenporr.— Beitrage zu einer Malacozoologia Rossica,’’ ‘Mém. Acad. d. Sci. St. Pétersbourg,’ sér. 6, Sci. natur., t. vi. . Bernarp.— Recherches sur Valvata piscinalis,” ‘Bull. Scientif. France et Belgique,’ t. xxii, pl. xx, fig. 2. . Bercu.—‘ Malacologische Untersuchungen,’ Suppl. Heft iv, pp. 261, 273, pl. x, fig. 8; y, fig. 8. . Mecxeit.—‘“ Ueber den Geschlechtsapparat einiger hermaphroditischen Thiere,” ‘ Archiv f. Anat. u. Phys.,’ 1844. . NorpmMann.—“ Monographie des Tergipes Edwardsi,” ‘ Mém. Acad. St. Pétersbourg,’ 1845. . Lrvcxart.—‘ Zoologische Untersuchungen,’ drittes Heft, p. 78, &e. . TrincnEse.— Alolidide e famiglie affini,” ‘ Atti Accad. Lincei,’ ser. 3, vol. xi, pl. Ixx. . Hatter.—‘ Die Anatomie von Siphonaria gigas,” ‘ Arb, Zool. Inst. Wien,’ vol. x, p. 83, pl. ii, fig. 22. . Davatne.— Recherches sur la génération des Huitres,” ‘Mém. Soc. Biol. Paris,’ t. iv, 1853. . Lacaze-Dutuiers.—“ Recherches sur les organes genitaux des Acéphales Lamellibranches,” ‘Ann, d. Sci. nat.,’ sér. 4, t. ii, p. 217. . Ann. Mag. Nat. Hist.,’ ser. 6, vol. xi, p. 262. . DavatnE.—Loc. cit., p. 316, . vAN BeneDEN.—‘ Sur les organes sexuels des Huitres (O. edulis),” ‘Comptes rendus Acad. Sci. Paris,’ t. xl. . Brooxs.—* Development of the American Oyster,” ‘Stud. Biol. Lab. Johns Hopkins Univ.,’ vol. i, part iv, pls. viii, ix. . Boucnon BranpELy.—‘ Comptes rendus Acad. Sci. Paris,’ t. xev. . Horx.— Les organes de la génération de l’huitre,” ‘Tydschr. Ned. Dierk. Ver.,’ Suppl. Deel i, pl. ia. Lacazz-Duruiers.—Loc. cit., p. 222. Vorttzkow.—“ Entovalva mirabilis,” ‘ Zool. Jahrb.’ (‘ Abth, f. Syst. Geogr. u. Biol.’), Bd. v, p. 622. 42 PAUL PELSENEER. 20. Mitne-Epwarps.—“ Observations sur divers Mollusques,” ‘ Ann. d. Sci. nat.,’ ser. 2, t. xviii, pl. x, fig. 1. 20a. Humpert.— Note sur la structure des organes génitaux de quelques espéces du genre Pecten,” ‘ Aun. d. Sci. nat.,’ sér. 3, t. xx, p. 338. 21. Lacaze-Dutst1ers.—Loc. cit., p. 208. 22. Futtarton.—“ On the Development of the Common Scallop,” ‘ Highth Ann. Rep. Fish. Board for Scotland,’ p. 290, pl. viii. 23. Jacxson.— Phylogeny of the Pelecypoda,’’ ‘Mem. Boston Soc. Nat. Hist.,’ vol. iv, pp. 341, 348. 23a. Humpbert.—Loc. cit., p. 338. 24. von Sresotp,—‘ Arch. f. Anat. u. Phys.,’ 1837, p. 383. 25. Lreypic.—“ Ueber Cyclas cornea, Lam.,” ‘ Arch. f. Anat. u. Phys.,’ 1855, p. 59. Sreranorr.— Ueber die Geschlechtsorgane und die Entwickelung von Cyclas,” ‘Arch. f, naturg.,’ 1865, p. 3. 26, Steenstrup.—‘ Underségelser over hermaphroditismens tilvaerelse i naturen,’ 1845, p. 69. 27. Ray Lanxester.— Contributions to the Developmental History of the “Mollusca,” £ Phil. Trans.,’ 1875, p. 1. 28, PrtsenreR.—‘‘ Deux nouveaux Pélécypodes hermaphrodites,” ‘ Comptes rendus Acad. Sci. Paris,’ t. cx. PELSENEER.—“ Sur existence d’un groupe entier de Lamellibranches hermaphrodites,” ‘ Zool. Anzeiger,’ 1891. 29. F. J. H. Lacaze-Dutuiers,—Loc. cit., p. 211. 80. H. pr Lacaze-Duruiers.— Morphologie des Acéphales,” ‘Arch. d. Zool. Expér.,’ sér. 2, t. i, p. 216. 81, Levckart.—‘ Zoologische Untersuchungen,’ Heft iii, p. 76. 82. Since Eisie (“ Beitrage zur Anatomie und Entwickelungsgeschichte der Geschlechtsorgane von Lymneus,” ‘ Zeitschr. f. wiss. Zool.,’ Bd. xix, p. 310), &c. According to Bazsor (‘Ueber den Cyclus der Ge- schlechtsentwickelung der Stylommatophoren,” ‘ Verbandl. Deutsch. Zool. Gesellsch.,’ 1894, p. 57), Agriolimax levis and Limax maximus furnish exceptions and are protogynous, 88. PELSENEER.— Recherches sur divers Opisthobranches,” ‘ Mém. Cour. Acad. Belg.,’ t. lili, p. 21. 84, PeLsenEER.—Ibid., p, 24. 35. Lruckart.—Loc. cit. 86. Cuun.—“ Bericht iiber eine nach den Canarischen Inseln im Winter, 1887-8, ausgefiihrte Reise,” ‘Sitzungsber, k, Akad, d. Wiss. Berlin,’ 1889, p, 543, HERMAPHRODITISM IN MOLLUSCA. A3 37. J. Mitter.—‘ Ueber Synapta digitata und die Erzengung von Schnecken in Holothurien,’ p. 13. 38. Wiren.—‘‘ Studien iiber die Solenogastren,” II, ‘ K. Svensk. Vetensk. Akad. Handl.,’ Bd. xxv, No. 6, p. 47. 89. Davaine.—Loc. cit., p. 316. 40. van BENEDEN.—‘ Comptes Rendus Acad. Sci. Paris,’ t. x]. 41. Sarnt-Loup.—“ Observations anatomiques sur les Aplysiens,” ‘ Comptes rendus Acad. Sci. Paris,’ t. evii. 42. Fou.—‘Sur le Développement de Ptéropodes,” ‘Arch. Zool. expér.,’ sér. 1, t. iv, p. 205. 48. TrincuEsE.—‘ Descrizione del nuovo genere Bosellia,” ‘Mem. Accad. d. Sci. Bologna,’ ser. 5, t. i, p. 776. 44, ZircLer.—“ Die Entwickelung von Cyclas cornea, Lam.,” ‘ Zeitschr. f. wiss. Zool.,’ Bd. xli, p. 562, pl. xxviii, figs. 30, 33. 45. Woopwarp.— Proc. Zool. Soc. London,’ 1893, p. 321. 46. Korscuet and He1pER.—‘ Lehrbuch der vergleichenden Entwickelungs- geschichte der wirbellosen Thiere,’ pp. 118, 119, 129, 223. 47. This opinion, which I expressed several years ago for the Mollusca in general (‘Comptes rendus,’ t. cx, p. 1083), is also shared by Brock for the Pulmonata (‘ Zeitschr. f. wiss. Zool.,’ Bd. xliv, p. 374), and has been since adopted by Lane (‘Lehrbuch der vergleichenden Anatomie,’ p. 816). 48. Rovuzaup.—‘ Recherches sur le développement des organes génitaux de quelques Gastéropodes hermaphrodites,’ Montpellier, 1885, p. 79. 48a. Brocx.— Die Entwickelung des Geschlechtsapparat der Stylommato- phoren Pulmonaten,” ‘ Zeitschr. f. wiss. Zool.,’ Bd. xliv, p. 368. 48b. Brocx.—Loc. cit., pp. 372, 374. 48c. Pecx.—On the Anatomy and Histology of Cymbuliopsis ealeeelg® ‘Stud. Biol. Lab. Johns Hopkins University,’ vol. iv, No. 6, p. 15. 49, von JuERinc.—‘ Jahrb. d. Malakozool. Gesellsch.,’ Bd. xii, p. 207 (1885). Srmrotau.— Zeitschr, f. wiss. Zool.,’ Bd. xlv, pp. 655, 656, 661 (1887). 49a, CoLtince.— Absence of Male Reproductive Organs in Two Herma- phrodite Mollusca,” ‘Journ. of Anat, and Phys.,’ vol. xxvii, p. 237. 49b. Cottince.—Loce. cit., p. 238. 49c, Brarp.—“ On the Life History and Development of the Genus Myzo- stoma,” ‘Mittheil. Zool. Stat. Neapel,’ Bd. v, p. 578 :—< Herma- phroditism, probably all hermaphroditism, had its origin in a unisexual condition.” 50. Detace.—* Evolution de la Sacculine,” ‘ Arch. de Zool. expér.,’ sér. 2, t. ii, p. 704:—* La séparation des sexes chez Sacculina, et probable- 44. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. PAUL PELSENEER. ment chez tous les Crustacés hermaphrodites, est l’état primitif dans le développement ontogénétique et phylogénétique.” F. Miter. ‘ Die Zwitterbildung im Tierreich,” ‘Kosmos,’ Bd. xvii, pp. 821—334. In addition to the various examples given further on, it may be remarked that the only hermaphrodite forms among Echinoderms are the Synaptids, which are the most highly specialised by the reduction of the calcareous plates, and the disappearance of the madreporite and the ambulacral tubes. Similarly in Pisces the hermaphrodite forms (Serranide) belong to the most specialised types of Teleosteans; see Cope, ‘ The Origin of the Fittest,’ 1887, p. 331. Though descended from free ancestors, as their larvee show. Some Spirorbis, Protula, Pileolaria, Laonome. Bourne.—“‘ On Certain Abnormalities in the Common Frog (Rana temporaria),” ‘Quart. Journ. Micr. Sci.,’ vol. xxiv, p. 83, pl. iv. Voer.— Notice sur un Hareng hermaphrodite,” ‘ Arch. de Biol.,’ t. iii, p, 255, pl. x Woonwarp.—‘ Proc. Zool. Soc. London,’ 1893, p. 322. BeErnarp.— Nature,’ vol. xlili, p. 343. Although it is yet a very rare practice to compare the phenomena pre- sented by the two organic “kingdoms,” it may be added here that hermaphroditism, properly so called, is also a characteristic of specialised plants, and that it is wanting, speaking generally, in aquatic plants and the most ancient of terrestrial plants; see especially Errera et Gevaert, “Sur la structure et les modes de fécondation des fleurs,” ‘Bull. Soc. Botan. Belg.,’ vol. xvii, p. 164. Dine, —Loc, cit., p. 704. Ginruer.— An Introduction to the Study of Fishes,” 1880, p. 157 :— “In the European species of Serranus a testicle-like body is attached to the lower part of the ovary, but many specimens are undoubtedly males having normally developed testicles only.” Brocx.—‘ Zeitschr. f. wiss. Zool.,’ Bd. xliv, p. 374. Brocx.— Untersuchungen iiber die Geschlechtsorgane einiger Mure- noiden,” ‘ Mitth. Zool. Stat. Neapel,’ Bd. ii, p. 488. HERMAPHRODITISM IN MOLLUSOA. 45 EXPLANATION OF PLATES 4, 5, & 6, Illustrating Dr. Pelseneer’s paper on ‘‘ Hermaphroditism in Mollusca.” Fic. 1—Hermaphrodite reproductive apparatus of Onchidiopsis green- landica, dorsal view. 1, ovotestis; 11, oviduct; III, mucous gland; Iv, receptaculum seminis ; V, penis; VI, spermiduct ; vil, female aperture ; VIII, prostate; Ix, sperm-oviduct; x, vesicula seminalis. Fic. 2.—Transverse section of the ovotestis of Onchidiopsis. Fic. 3.—Part of the ovotestis of Onchidiopsis, slightly magnified, show- ing male (white) and female (grey) portions. Fie. 4.—Ostrea cochlear, male, left view. 1, rectum; 0, testis; 111, internal plate of left gill; 1v, internal plate of right gill; v, pallial suture ; vi, adductor muscle; vit, auricle ; vit1, ventricle. Fic. 5.—Ostrea cochlear, female, left view. 1, liver, seen by trans- parency through the ovary; 11, auricle; 11, pallial suture; iv, abductor muscle ; v, intestine; vi, ventricle. Fic. 6.—Ostrea cochlear, ventral aspect, mantle and gills removed. I, right genital aperture; 1, right renal aperture; 11, visceral ganglia; IV, outline of the mantle; v, adductor muscle ; v1, visceral commissure ; VII, branchial nerve ; vuiI, left renal aperture; 1x, left genital aperture ; x, palps (the lines of adhesion of the gills to the visceral mass are indicated by dotted lines). Fic. 7.—Transverse section of a female Ostrea cochlear. 1, intestine; 1, right lobe of the mantle; 111, nephridium; 1v, branchial axis; v, genital aperture ; VI, visceral commissure ; VII, external plate of the right gill; v1, internal (adaxial) plate of right gill; rx, visceral commissure (left half); x, renal aperture ; XI, reno-pericardial duct; x11, ovary ; XIII, pericardium; xIVv, ventricle ; xv, left lobe of the mantle ; xvi, rectum; XVII, auricles joined together. Fie. 8.—Transverse section of a male Ostrea cochlear. 1, right lobe of the mantle; 11, branchial axis; 111, testis; Iv, external plate of right gill; v, visceral commissure; vI, internal plate of right gill; v1, visceral com- missure ; VIII, genital aperture ; 1x, nephridium; x, testis ; XI, pericardium ; xu, ventricle (on the pericardial wall of which are modified epithelial cells = pericardial glands); x11, left lobe of the mantle; xiv, rectum; xv, auricle ; XVI, intestine. Fic. 9.—Transverse section of the ovotestis of Cardium oblongum, 46 PAUL PELSENEER. Fie. 10.—Transverse section of the abdomen of Pecten flexuosus, showing male and female halves of the ovotestis. 1, intestine. Fic. 11.—Longitudinal section of the right genital gland of Cyclas cornea (the anterior part is below). 1, epithelium of the visceral mass; 11, testis; 111, part of the genital duct between testis and ovary; IV, ovary. Fic. 12.—Cyclas cornea, left view, without left gill and pallial lobe. I, anterior retractor of the foot; 11, anterior adductor; 111, palp; Iv, foot; Vv, ovary; VI, genital aperture; vil, visceral ganglion; viiI, posterior ad- ductor ; 1x, posterior retractor of the foot; x, kidney; x1, rectum; xu, auriculo-ventricular slit; x1II, ventricle; xIv, testis ; xv, liver mass. Fre. 18.—Plan of nervous system, and genital and renal apertures in an ordinary Lamellibranch 1, cerebral ganglion; 11, visceral commissure; II, pedal ganglia; Iv, genital aperture; v, renal aperture; vi, visceral ganglia. Fie. 14.—Plan of nervous” system, and genital and renal apertures in Anatinacea. I—III, and v, VI, as in Fig. 13; Iv, female aperture; vm, male. aperture. Fie. 15.—Transverse section of Clavagella passing through the male apertures. I, ovary; II, aorta; 111, mantle; Iv, branchial axis; v, visceral commissure ; VI, spermiduct; vi1, foot; vit, testis; Ix, pad of adhesion of the reflected internal lamella of the gill; x, right male aperture ; x1, branchial axis ; XII, Ovary; XIII, auricle; xIv, intestine; xv, pericardium. Fic. 16.—Transverse section of Clavagella (posterior to Fig. 15), passing through the female apertures. 1, ovary; 1, aorta; 111, intestine; Iv, ovary; v, visceral commissure ; VI, foot ; vi1, pad of ciliated adhesion of the internal reflected lamella of the gill; vii1, female aperture; 1x, branchial axis; x, nephridium ; x1, auricle; x11, mantle; x11, pericardium. CEREBRAL CONVOLUTIONS—‘* SALLY.”’ A” A Description of the Cerebral Convolutions of the Chimpanzee known as “Sally;” with Notes on the Convolutions of other Chim- panzees and of Two Orangs. By W. Blaxland Benham, D.Sc.Lond., Hon. M.A.Oxon., Aldrichian Demonstrator of Comparative Anatomy in the University of Oxford ; Lecturer on Biology at Bedford College, London. With Plates 7—11. ConTENTS. PAGE PAGE 1. Historical : : . 48 9. The Frontal Lobe . ai) 2. Introduction . : Pe!) 10. Summary of Results of 3. Description of ‘“ Sally’s” the Comparison of a series Brain. . : yO of Chimpanzee Brains . 74 4. The Parieto-occipital Fis- 11. List of Chief Papers re- SICH ae ei. as Do lating to the Chimpanzee 5. The Parietal Convolutions. 57 Brain. : ‘ a6 6. The‘‘Affenspalte” in Chim- 12. The Brain of the Orang- panzee and Man . =, bg outang . 3 ° 16 7. The Occipital Lobe . 167 Explanation of Plates . 81 8. The Sylvian Fissure, &c. . 68 PREFACE. Tue brain of the interesting and well-known chimpanzee “ Sally,” which lived for eight years in captivity at the Zoolo- gical Gardens in London, has been recently figured and described by Mr. Beddard (11) in his important memoir on the anatomy of this ape, which he considers to be T. calvus; there are, however, several features about this brain which deserve further notice, 48 W. BLAXLAND BENHAM. After the brain had been described it was purchased by Professor Lankester for the Oxford University Museum; and my attention was particularly directed to it when comparing the convolutions of the anthropoid apes with those of man, for the purpose of exhibiting in the museum a series of anatomical preparations illustrating the relations of man to the apes. One of the most striking features about “ Sally’s ” brain is the absence of a well-defined “ occipital operculum,” the presence of which was supposed by Gratiolet to mark off the chimpanzee brain very strongly from that of the orang, in which it is generally absent. With this disappearance of the operculum is connected certain other modifications in this region of the brain, the most evident of which is the diminu- tion in the extent of the so-called “ Affenspalte” or Simian fissure, and the consequent resemblance of this region of the brain to that of the orang and of man. Gratiolet placed the chimpanzee further from man than the orang, but “ Sally’s ” brain is considerably more human than that of any orang hitherto figured, and it becomes a matter of great interest to compare the gyri and sulci in the brain of this specimen with those of the more ordinary chimpanzee on the one hand and with those of man on the other. In making these comparisons I have the advantage over Beddard in that I am able to make use of the valuable and extensive researches of Dr. D. J. Cunningham, published in 1892, just after Beddard’s com- munication to the Zoological Society ; and in the present paper I shall adopt Dr. Cunningham’s nomenclature, which is in agreement generally with that employed on the Continent. My great indebtedness to Dr. Cunningham’s work is sufficiently evidenced by my constant reference to it. I. Hisroricat. Descriptions, more or less detailed, of the brain of the chimpanzee are fairly numerous. I append a bibliography of the more important papers ; but amongst those described I have only been able to light upon two brains bearing any close resemblance to that of “Sally.” One of these is described and CEREBRAL CONVOLUTIONS—** SALLY.” 49 figured by P. Broca (6).1 In the figure here reproduced (fig. 19) it will be seen that the “operculum” is absent on both sides. The “ Affenspalte,” or Simian fissure (0. 5.), or “external perpendicular fissure ”’ as he called it, is not inter- rupted on the left side, as the second “ pli de passage ” does not rise to the surface. But on the right side the annectant gyrus (2) becomes superficial, and separates the Simian fissure (a. a.) into two parts. The first “ pli de passage” (J.) is present on both sides. It will be seen below that another explanation of these arrangements is possible. It may be well to note that the brain belonged to a young male. The second brain to which reference may be made was described in considerable detail by Dr. Joh. Miiller (9) in 1888, one of whose figures is here reproduced (fig. 25). In it the “ operculum ” is absent on both sides, the occipital lobe and the region of the Simian fissure being more complex than in the ordinary chimpanzee, but not so folded as in Broca’s chimpanzee, or as in “ Sally,” for the second annectant gyrus does not come to the surface on either side. This brain also was taken from a young male, which still retained its milk teeth. Neither of these authors gives any information as to other anatomical characters of the animal to which the brain belonged to enable us to state whether it belonged to the so-called species T. calvus, or T. niger. As long ago as 1866 Sir Wm. Turner (4) described a couple of chimpanzee brains presenting characters which at the time were of peculiar interest, on account of certain views held by Gratiolet above referred to. One of the brains, figured on p. 680 of the paper, had but a feebly developed operculum on the right side, and therewith presented certain other pecu- liarities, but there is no special resemblance to that of “ Sally.” The object of the paper was to show that Gratiolet’s opinions as to the annectant gyri in the chimpanzee were founded on insufficient material. 1 T have to thank Sir Wm. Turner for most kindly calling my attention to this paper. VOL. 37, PART 1.—NEW SER. D 50 W. BLAXLAND BENHAM. 2. INTRODUCTION. In looking over the stock of chimpanzees’ brains in the museum of the Royal College of Surgeons,—for which privi- lege I wish to express my thanks to Prof. Stewart, who most kindly gave me every assistance in the matter,—I came across a specimen in the museum stores which bears considerable resemblance to that of “ Sally,” as well as to that figured by Turner, in that it has lost its occipital operculum, but only on the right side, as it is well developed on the left. This brain belonged to the. late Prof. John Marshall, but has not been described by him so far as I have been able to ascertain. It came into the possession of the Royal College of Surgeons in 1891, but I have been unable to trace its origin. I have thought it worth while to give a figure of the right hemisphere here (fig. 20). On the supposition that the peculiarities of “ Sally’s ” brain are specific, and characteristic of T. calvus, I wrote to Prof. Herdman, of Liverpool, for I heard that he had purchased one of Garner’s chimpanzees which died on its arrival in England ; it was possible that this was T. calvus, and Prof. Herdman most generously acceded to my request to be allowed to examine the brain, and forwarded it to me. I was, however, disappointed, for it does not present characters marking it off for that of ordinary chimpanzees (see fig. 36), and it is as yet uncertain whether the specimen from which it was taken is T. niger, T. calvus, or a third species. In his letter to me Prof. Herdman writes: ‘“‘ Garner declared that this animal was different from calvus, and also from niger. . . . I saw the dead head (after being skinned), and thought it might be calvus.” Thus we have no positive evidence to show whether the peculiarities of ‘ Sally’s”’ brain are characteristic of the species T. calvus; that they are not due to the age or sex of the animal seems certain from the specimens of Broca and Miiller, and it can scarcely be maintained that they are con- nected with ‘ Sally’s” greater intelligence, for both Broca’s CEREBRAL CONVOLUTIONS—‘* SALLY.” 51 and Miiller’s specimens were quite young, and no particular intelligence has been attributed to them. The only sugges- tion remaining is that in the species T. calvus the brain does present peculiarities, and that Broca’s and Miiller’s specimens belonged also to this species. On the other hand, we are still uncertain that there is more than one “species” of chimpanzee. Mr. Beddard, from a careful study of the muscular anatomy of “ Sally’s ”’ limbs, from a comparison of the skull, brain, and other organs, considers that the animal belonged to the species T. calvus, and shows reasons for distinguishing it from the animal described, ana- tomically, by Gratiolet and Alix, and named T. aubryi. Nevertheless, the differences in muscular anatomy, &c., re- corded in the ordinary chimpanzee and in these two other forms, can scarcely be held to establish with absolute certainty the specific distinctness of these three forms: the variations in muscular anatomy in man appear to be every bit as great as in these, and till sufficient specimens of the black-faced bald chimpanzee, known as T. calvus, on the one hand, and of the flesh-coloured, black-red haired form (T. niger), have been as fully and as carefully described as Beddard has done for “Sally,” we shall not be entitled—at least we ought to hesitate—to be positive about the “ specific ”’ differences of these three forms. As we shall see, the convolutions of the chimpanzee’s brain are subject to variations as wide as occur in man, and some of them are of importance. Such features, too, as the shape of the ear are extremely variable, every stage in the folding of the margin being exhibited, from the unfolded condition to a state closely resembling man’s ear. For the present, at least in this paper, the chimpanzee “Sally ” is regarded as being merely a “ variety ” of the single species I’. niger. 3. Description oF “Sauty’s” Brain. Ya The general features, dimensions, and weight of “ Sally’s ” brain have been described by Beddard; 1 have but little to 52 W. BLAXLAND BENHAM. add. He found that the brain “ weighed, after removal of the pia mater and after an immersion of four months in spirit, 83 oz.; it had been allowed to dry for an hour and a half before weighing, it was then damp, but not wet.” He gives the weights of brain of two other chimpanzees (T. niger) as 63 oz. and 63 oz. The brain described by Miiller weighed 64 oz. (213 grammes), whilst that described by Symington (10), though not bearing any particular resemblance to ‘‘ Sally’s” brain, weighed as much as 84 oz. after preservation in spirit ; both these authors tabulate the weights of the brains previously recorded. Mar- shall’s specimen (1861) (2) weighed under similar circumstances as much as9 oz. But the weight of the brain fresh is con- siderably more; thus Symington’s was 13 oz., Marshall’s 15 oz., and Chapman’s (7) 10 oz. 10 grains. In all these cases the animal was young, and Symington suggests from analogy with man, and from comparison of cranial cavities of young and adult chimpanzees, that the weight of the brain in the latter is 15 oz. I have not deemed it necessary to weigh the various brains to which I shall have occasion to refer. Beddard gives the length of ‘ Sally’s” brain as 103 mm., and that of the hemisphere alone as 100 mm. Miller’s measured 92 mm. and Symington’s 95 mm., so that “ Sally’s ” appears to be the largest hitherto carefully measured, as one would expect, since the animal is the oldest chimpanzee whose brain has been described. The general appearance of the brain is more like that of man on asmall scale than is that of an ordinary chimpanzee (see figs. 1, 2,3). The frontal lobes are wider and more rounded anteriorly. But we must not lay too much stress on this point, for there is considerable variation in the shape of the frontal lobes when many chimpanzee brains are examined. There is one polut which Beddard remarks upon, and which differen- tiates “‘Sally’s” brain from any other chimpanzee’s brain described and figured, viz. the very slight “ keel” or ridge on the inferior or orbital surface of the frontal lobes where they CEREBRAL CONVOLUTIONS—‘* SALLY.” 5g touch in the median line. In none of the brains which I have examined is this keel absent ; it is present in the “ Sally ”-like brain in the Museum of the Royal College of Surgeons referred to above; it was present in the brain described by Miiller. A front view of “ Sally’s ” brain thus presents a great contrast with that of any other chimpanzee hitherto examined, and recalls that of an orang or man (see figs. 5—7). Since it is the posterior part of the brain which is so characteristically developed, I will commence with this region. 4. Tue Parieto-occipitaL Fissure. This fissure in “ Sally’s ” brain presents certain features of interest, chiefly in that it is divided into two portions, one of which, the mesial or “internal” parieto-occipital, is confined to the mesial surface; the other extends on to the upper sur- face of the hemisphere, and may be termed the “lateral” parieto- occipital fissure, as the term “external”? has been used in rather a different sense from that which I wish to indicate here. Viewed from the upper surface (figs. 2, 10), the parieto- occipital on this left side appears as a deep, well-marked fissure about 15 mm. long, extending rather obliquely outwards after coming on to the surface, which it cuts at adistance of 75 mm. from the anterior end. This superior or lateral portion of the parieto-occipital is bounded by an “ arcus parieto-occipitalis ”’ (Gratiolet’s ‘seconde plide passage”’), which is, as usual, marked out laterally by the ramus occipitalis of the intra-parietal (p*.) and anteriorly by a small mediad of this branch (n.), the pre- arcal branch. But there is another portion of the parieto- occipital fissure which does not reach the upper surface, and which Beddard overlooked. It is seen in fig. 2 lithographed from a photograph, and also in fig. 10, just in front of the lateral portion. When the mesial surface of the hemisphere is examined (fig. 14), it is seen that the first-named portion of the furrow (lat. p. 0.) is nearly entirely cut off from the remaining vertically disposed part of the furrow (mes. p. o.), for the anterior limiting gyrus passes very deeply down at a., and disappears from view, so that the superficial and the vertical 54 W. BLAXLAND BENHAM. portions of the parieto-occipital are connected by a short but deep sulcus. The vertical portion of the fissure, after thus joining the superior portion, forks dorsally as shown (at 6. and c.) ; below the fork it passes straight downwards on the mesial surface, and is continued on to the tentorial surface of the hemisphere, where it joins the calcarine fissure—the gyrus cunei being quite deep. We thus have the <-shaped arrangement of this system of fissures which Cunningham regards as typical for man, but which he denies to the apes. On the right hemisphere (figs. 2 and 11) the lateral part of the parieto-occipital is entirely separated from the mesial vertical portion, and is completely surrounded by gyri (see fig. 15) ; its mediad end dips for only a very little distance down- wards in the mesial face, and the vertical portion—the main part of the fissure (mes. p. 0.)—comes up to the upper surface of the hemisphere in front of the fissure in question. So entirely cut off is the upper part of the parieto-occipital that I at first took it for a part of the Affenspalte, but its depth and other relations, especially when compared with the arrangement on the left side, show it to be a portion of the parieto-occipital, as Beddard describes it. He, however, overlooked the gyrus which cuts it off from the vertical part of the fissure, which also he did not recognise. The separation of the parieto-occipital fissure into two dis- tinct portions appears to bea rare occurrence. I do not find it mentioned in descriptions of chimpanzees’ brains, and only rarely is it recorded for man. Cunningham does not say very much about it. He figures the “normal” arrangement on p. 43, where the fissure, at a point some distance below its upper end, gives off two branches, one anterior and one posterior, to the main track, which is separated from the calcarine fissure by a super- ficial gyrus cunei; but he represents no gyrus intercuneatus. What has happened in “Sally”? The hinder of the two branches of the parieto-occipital appears to have been cut off— partially on the left, entirely on the right—from the rest of the fissure. Cunningham represents several varieties in the arrange- CEREBRAL CONVOLUTIONS—“ SALLY.”’ 55 ment of the parieto-occipital in man, but in none of them do we find quite these relations. In three instances in which the fissure is duplicated (see his figures on p. 40), the portion which cuts into the upper surface of the hemisphere lies in front of the vertical portion, and not behind it asin “ Sally.” In one case (his fig. 18) this “doubling” is brought about by the gyrus cunei rising to the surface and the “stem” of his <-shaped fissure being prolonged into the cuneus. This was only met with in 4 out of 127 hemispheres examined. In the other case (his fig. 19), which he found only once in the course of his re- search, the gyrus intercuneatus is oblique and becomes super- ficial, the upper part of the parieto-occipital being prolonged downwards in front of the main part of the fissure. Now, in “Sally” it appears to me that the cause of the * doubling” is the same, viz. the rising up of the gyrus intercuneatus (a. in my figures), but the upper part of the original fissure cuts downwards into the cuneus, behind the main part of the fissure, which extends into the precuneus and only just reaches the surface. This part of the parieto-occipital fissure lying on the upper surface and separated from that on the mesial surface may be termed the ‘‘ superior or lateral parieto-occipital.” I wish to avoid the term “ external,” which would perhaps be more appropriate, since the term “ external parieto-occipital”’ has been used in a variety of senses. The only true sense in which the expression should be used, as Cunningham and others have pointed out, is to refer to that part of the vertical incision which appears as the upper surface of the hemisphere, and varies in extent, naturally, with the depth of the incision. But the term has had two other significances attached to it, especially by English writers, namely, to imply (1) true Affenspalte (Bischoff’s “ fissura perpendicularis externa,” in the foetal human brain)—in this sense it is used by Beddard; and (2) the groove between the anterior free edge of the operculum and the parietal lobe. We will now examine the conditions of the parieto-occipital fissure in the normal chimpanzee brain. 56 W. BLAXLAND BENHAM. In a specimen (947 f) in our Museum (see fig. 22) the operculum is normally developed. When this is turned back the Affenspalte is exposed, and in front of it the deep parieto- occipital fissure is seen, extending about three quarters of an inch across the brain (fig. 23). On the mesial face this descends for nearly half an inch (fig. 24), but is cut off from the rest of the parieto-occipital by a conspicuous gyrus (a.). The main portion of the fissure lies in front of it, and only just cuts the edge of the hemisphere (mes. p. 0.). The anterior portion of the fissure (mes. p. 0.) does not branch. Here, when the operculum is removed, we have the same arrangement as in “ Sally.” In another specimen (fig. 26), in which the edge of the operculum is directed obliquely backwards at its mesiad end, we have a further development of the gyrus intercuneatus, so that the very deep “lateral” parieto-occipital (lat. p. 0.) scarcely bends round on to the mesial face at all, and lies almost entirely below the operculum, cutting off the intra- parietal from reaching the Affenspalte. In fact, it looks at first sight as if it were the “ Affenspalte,” and as if the gyrus marked a. were the “ first annectant gyrus.” Such, indeed, it appears to be, for in many cases the “external perpendicular fissure ” of authors is not the same fissure as is here called « Affenspalte,’ although it has been homologised with it. The “internal” or mesial parieto-occipital (mes. p. 0.) enters the calcarine fissure, the gyrus cunei being, as far as I can see, absent (fig. 27). In a brain referred to by Rolleston,' in which on the right side the operculum is less developed than usual (see fig. 30), the parieto- occipital fissure is very deep, and extends for nearly an inch across the surface of the hemisphere (fig.32). Here it is limited by a well-defined ‘‘ annectant ” (arcus). Seen from the mesial surface (fig. 21) there is no trace of a division of the fissure into two portions. On both right and left sides there is but a single, nearly vertical cleft—cut off by the gyrus cunei 1 Rolleston, “On the Affinities of the Brain of the Orang-outang,’ ‘Nat. Hist. Rev.,’ 1861. CEREBRAL CONVOLUTIONS—‘* SALLY.”’ 57 from the calcarine fissure—but presenting no lateral branches, nor gyri coming to the surface. This seems to be the simplest condition, and closely resembles the arrangement in the lower monkeys; some other of our specimens of chimpanzees exhibit this condition ; in others we have seen the gyrus intercuneatus interfering with the continuity, and giving rise to two more or less independent fissures. In a brain at the College of Surgeons (No. 1338, I a*), the left hemisphere exhibits this simple arrangement, whilst the right (1338, I a) possesses the more complicated condition. In the human brain, the separation of the parieto-occipital fissure into two appears from Dr. Cunningham’s researches to be rare, but in the brain (950) exhibited as a typical human brain in the Oxford Museum, and represented here by fig. 16, we have on the left side an arrangement repeating, I believe, that in “ Sally,” namely, the uprising of the gyrus intercuneatus (a.) so as to separate a lateral parieto-occipital (/at. p. o.) from the remainder (m. p.o.). On the right hemi- sphere the parieto-occipital cuts into the upper surface for about an inch, forming the true “ external parieto-occipital,” and this part is continuous with the vertical portion. But in the left hemisphere there lies in front of the “ transverse occipital fissure” or Affenspalte, between it and the vertical part of the parieto-occipital, a transversely placed fissure about an inch in length (dat. p. 0.), bounded anteriorly by a gyrus which is partially overlapped by the “ arcus parieto- occipitalis.” 5. Tue PartetaL ConvoLurions. The fissure of Rolando (“ fissura centralis ”) has the usual wavy character ; it does not enter the Sylvian fissure on either side, nor does it pass on to the mesial face of the hemisphere. The distance of the upper extremity of the fissure from the an- terior end of the hemisphere (between vertical plates) is 45 mm. There appears to be some confusion in Beddard’s paper with regard to the position of this fissure of Rolando. In the plate xxiii, fig. 3, the index line from F, R. is carried to the upper 58 W. BLAXLAND BENHAM. end of the calloso-marginal fissure. This appears to be a slip, yet his measurement of the fronto-Rolandic length given on p. 200 is 54 mm., which is really the distance of the calloso- marginal fissure from the anterior end of the cerebrum. The fissure is thus in front of the middle of the cerebrum and not behind it, as Beddard states. In the normal chimpanzee here figured (fig. 1) the length of the cerebral hemisphere is 95 mm., the fronto-Rolandic length is 55 mm., so that the Rolandic fissure is behind the middle of the hemisphere. The intra-parietal fissure of Turner may be divided, ac- cording to Cunningham, into three constituents: (1) An an- terior post-Rolandic vertical fissure, or “ sulcus postcentralis.” (2) A more or less horizontal furrow, or “ ramus horizontalis ” (“‘interparietal” of Ecker), passing backwards and (3) as a prolongation which bounds the arcus parieto-occipitalis and passes on to the occipital lobe—the ‘‘ ramus occipitalis.” These three constituents may or may not be continuous in man, and to them is added, usually as a separate furrow, a “superior postcentralis,” lying more or less parallel to the Rolandic furrow, and dorsal of the inferior postcentralis. In “ Sally,” as will be seen (figs. 10, 11), this is also the arrangement; the inferior postcentralis (p’.) is continued a short distance upward beyond its junction with the ramus horizontalis (p°.), and above this fissure is a small but well- marked furrow, representing the superior postcentralis (p*.) ; this is continuous on the left side with an oblique furrow (p. s.) lying parallel to and above the ramus horizontalis, whilst on the right side the corresponding furrow is parallel with the long axis of the brain. This secondary sulcus (p. s.) in the superior parietal lobe is represented in Cunningham’s drawing (by the letter c.), and is a very constant furrow in the chimpanzee, though variable in form. The superior postcentralis (p.) is in most of the published figures of chimpanzee brains, as well as in several of those exa- mined by me, a much more definite feature (sce figs. 30, 36, CEREBRAL CONVOLUTIONS—*“ SALLY.”’ 59 20), having a course more nearly parallel with the Rolando than is the case in“ Sally’s” brain. In both of Cunningham’s figures (pp. 203, 204) this furrow continues the direction of the inferior postcentralis (p1.). Cunningham states (p. 230) that the union of superior post- centralis (y?.) with inferior is the usual condition in chim- panzee ; and finds them separated in only two of the brains out of eight examined, and that only on one side. In “Sally,” in Miiller’s chimpanzee, and two others from the Royal College of Surgeons, as well asin Herdman’s, I find the fissure marked by Cunningham as upper part of superior postcentralis, separate from the rest of the system, and in all these instances it loses its parallelism to Rolando, becomes more or less oblique, and enters into connection with the furrow ps. (secondary sulcus of sup. par. lobule). 6. Tur Recion or THE AFFENSPALTE. The most interesting feature in “ Sally’s”’ brain, however, is the condition of the hinder end of the intra-parietal fissure. In the normal chimpanzee brain, and in the lower apes, the hinder part of the parietal lobe is overlapped by a com- paratively thin fold of the occipital lobe, known as the “ occipital operculum.” This lies flat against the parietal lobe, and thus conceals a greater or less amount of the latter. The deep cleft, extending nearly directly backwards and slightly downwards, existing between the operculum and the parietal lobe, is known as the ‘‘ Affenspalte,” or Simian fissure (figs. 27, 29, show its direction well). But some confusion has resulted from the application of this term to the superficial groove between the edge of the operculum and the parietal lobe. This superficial groove has also wrongly been termed the “external parieto-occipital” fissure, since it is apparently continuous with the “internal” fissure of that name. The true ‘‘ Affenspalte’”’ lies behind this latter fissure, and the two are independent. The parieto-occipital fissure is a vertically placed fissure in the median surface of the hemisphere, into which it extends 60 W. BLAXLAND BENHAM. for a considerable distance, and as such is visible also on the upper surface for a greater or less extent, according as the incision is deeper or shallower ; that part of the parieto-occipital (= Huxley’s occipito-parietal) which is seen in the upper surface is the only fissure to which the term “ external parieto-occipital ” is properly applied. I have shown above that this part of the parieto-occipital may become separated from the so-called “internal ”’ parieto- occipital ; but it can always be distinguished from the true Affenspalte, or Simian fissure, or subopercular furrow. In “ Sally’s ” brain (figs. 10, 11) the ramus occipitalis (p*.) sends off a small mediad branch (m.) in front of the parieto- occipital fissure, as is very generally the case in chimpanzees, the existence of which was first mentioned by Rolleston and later described by Sir Wm. Turner (4). This small fissure, there- fore, is of historical interest ; further, it helps to mark out definitely the so-called “pli de passage supérieur externe” which Gratiolet claimed for chimpanzee’s brain. The ramus occipitalis in ‘ Sally”? now curves outwards, forming the outer limit of the “ pli de passage ” or annectant gyrus or “arcus parieto-occipitalis,” and enters the ‘ Affen- spalte.” The arrangement on the two sides of the brain is not iden- tical. The above description refers to the right side. On the left side (fig. 10) the ramus occipitalis appears to bifurcate, a condition represented in some of Cunningham’s figures. The shorter, outer or laterad fork enters the “A ffen- spalte;” the more conspicuous mediad fork (z.) passes back- wards on to the occipital lobe in nearly a straight line. On the right side the ramus occipitalis also appears to bifur- cate, but this is not the case. Owing to the fact that the parieto-occipital fissure has become divided into two portions— the “lateral”? portion lying on the upper surface (lat. p. o.), whilst the “internal” portion is almost confined to the mesial surface and only just reaches the upper surface (mes. p. 0.)— the fissure (v.), which at first sight might be taken for the mediad fork of the ramus occipitalis, is, as we have seen, in CEREBRAL CONVOLUTIONS—“ SALLY.” 61 reality the przparieto-occipital branch of the intra-parietal fissure, which forms the anterior limit of the “ arcus parieto- occipitalis.” This branch (.), which Beddard’s figure repre- sents as joining the parieto-occipital, is separated by a gyrus from this, but as it is overlapped by the superior parietal lobule, its true relations are not seen in a view from above (as, for example, in a photograph). After giving off this small branch the intra-parietal fissure (p*.) curves slightly outwards and enters the Affenspalte, the well-marked obliquely-transverse groove so evidently the remains of the more extensive fissure of the normal chimpanzee. The hinder boundary of the Affenspalte on each side is a low ridge only slightly above the level of the opposite side of the furrow ; this is the remains of the operculum. In describing the terminations of the ramus occipitalis, Cunningham states that the two forks into which it divides are frequently almost at right angles to the main stem and nearly in line with one another, giving rise to a transversely directed fissure, which he identifies with the “ occipitalis transversus ” of Keker. The condition of affairs on “Sally’s” left hemisphere (fig. 10) is, indeed, not very unlike the left side of the figure of the human brain given by Ecker, in which the fissure marked there o., into which the intra-parietal falls, consists of two parts, one laterad of the intra-parietal, which soon forks; the other, mediad, lies behind the parieto-occipital. The question arises, do these two parts correspond to the two fissures in “ Sally’s’’ brain, the laterad (Affenspalte) and the mediad (z.), which I have described as a continuation of the ramus occipitalis? In other words, may the “ occipitalis transversus ’? be composed of two parts originally independent? This is a view held by Eberstaller. There is no doubt but that in the apes the bifurcation of the intra-parietal is an independent thing which may or may not become connected with the Affenspalte. In some cases the outer, in other cases the inner fork enters the Affenspalte, or both forks. Thus in Cunningham’s figure, p. 203, right side, the laterad fork is independent, the mediad joins the Simian fissure 62 W. BLAXLAND BENHAM. but again leaves it, whilst, as we see in “ Sally ”’ on the left side, it is the laterad fork which enters the Affenspalte. Cunning- ham shows that this is the case in the ape, and one of our chimpanzees (fig. 31) somewhat resembles the left side of the brain of Cebus albifrons figured by him in illustration of the point (on p. 223 of his memoir). Amongst our specimens of chimpanzees it is very frequently the case that the intra-parietal does not bifurcate, or if it does the resulting figures are coextensive with Affenspalte. The ramus occipitalis in our specimens usually passes under the operculum and enters the Affenspalte. In one case, however (see figs. 22, 23), the parieto-occipital fissure makes so deep an incision in the hemisphere that it cuts off the intra-parietal from the Affenspalte. We have here a condition similar, as I believe, to that described by Dr. Cunningham for Cebus capucinus, which he uses to support his view as to the homology of the “occipitalis transversus” with the bifurcation of the ramus occipitalis. According to him the fissure regarded as ramus occipitalis lies nearly at right angles to the rest of the intra- parietal fissure, and the parieto-occipital fissure falls into the system at this angle (loc. cit., fig. 45, p. 222). But it appears to me, from observations on C. robustus, that another explanation may be given of the fissure he regards as “ ramus occipitalis,” for in our specimen of this monkey, if the occipital lobe be pressed backwards it is seen that the parieto-occipital fissure is so deep that it extends more than halfway across the hemisphere, and receives, nearly in its middle, the intra-parietal. So that what in C. capucinus is regarded by Cunningham as “ramus occipitalis” is in C. robustus the outer part of the parieto-occipital. The Affenspalte, as Cunningham shows, is quite independent of this transversely placed fissure, as it is too, in our chimpanzee (947 f) just referred to. Now in the human brain do we find any close similarity between the arrangement of the convolutions in this region aud those of “ Sally” ? I have already referred to Kcker’s diagram of a generalised CEREBRAL CONVOLUTIONS—‘* SALLY.” 63 condition of the human fissures, which is closely similar to “ Sally,” the left hemisphere in each case being compared. Again, the left side of the brain figured by Bischoff (and repro- duced in fig. 17) is like the right hemisphere of “Sally” (fig. 11) so far as the “ Affenspalte”’ is concerned. The same oblique direction is to be noted. On the right side of the same brain, too (fig. 18), we have a very similar arrangement, but the intraparietal fissure (p‘.), before entering the “ Affenspalte ” (transverse occipital) sends inwards a short branch (z.) behind the “ arcus parieto-occipitalis.”’ In a human brain in the Oxford Museum, to which I have already referred, we have on the left side (fig. 16) a con- dition of the transverse occipital which seems to support Cunningham’s view. The intra-parietal appears to curve round the ‘“arcus,” and nearly reaches the mesial fissure, where a rather irregular furrow passes backwards from it; this small portion (z.) resembles a part of the “ mediad fork” of the ramus occipitalis of ‘‘Sally’s” left side (on fig. 10). The ‘‘transversus occipitalis ”’ in this human brain would then be, according to Cunningham, the upper (mediad) fork of the intra-parietal. But the general direction of the “ occipitalis transversus ” is so very similar to that on the right side of “Sally’s” brain that it is possible to explain this irregular furrow (z.) by supposing it to be an independent adventitious furrow on the occipital lobe, such as occurs in “ Sally’s ” left hemisphere, which there has gained a connection with one of the forks of the ramus occipitalis. It is still, in fact, a debateable point as to whether there is in the adult human brain a representative or homologue of the “* Affenspalte ” of the ape. There are, indeed, three chief views on the subject: (1) Ecker regards his “sulcus occipitalis transversus ” as such; (2) Eberstaller (to quote from Cunning- ham) draws a distinction between the upper and lower limbs of the sulcus occipitalis transversus of Ecker. In it we can dis-- tinguish a medial and a lateral segment by the point of union with the sagittal portion of the intraparietal furrow. The former (z. in my figure) bounds the arcus parieto-occipitalis 64 W.-BLAXLAND BENHAM. behind, without, in the majority of cases, reaching the mesial border; the latter is the sharply descending end of the arch- like “ fissura interparietalis.” Ecker’s view, originally shared by Cunningham (‘ Journ. Anat. Physiol.,’ 24), isnow combated by him, chiefly, it appears, on embryological grounds, partly on those of comparative anatomy. (8) He would regard the transverse occipital as merely a portion of the “ intra-parietal system ” which has nothing to do with the Affenspalte. It is due to the bifurcation, in fact, of the ramus occipitalis ; the Affenspalte, according to him, is only very rarely represented in man. The “ fissura perpendicularis externa” of Bischoff makes its appearance in the human fetus about the fifth month (according to Ecker and Cunningham) ; it is a “complete” fissure impressing itself upon the ventricle, so that the outer wall is bulged inwards by it. Some time later, apparently during the sixth month, it disappears. Now all these authorities are agreed that this external perpendicular fissure of the human foetus is the homologue of the “ Affenspalte” of the ape. But it is only a transient furrow, though “ complete,” and its place is occupied later on (seventh or eighth month) by the terminal bifurcation of the intra-parietal fissure, which is not a “ com- plete ” fissure, that is, it does not leave its impress upon the wall of the ventricle ; this second or replacing fissure becomes the “ transverse occipital.”’ For these and other reasons Cunningham denies the possi- bility of the suggestion that the latter fissure can be the homo- logue of the Affenspalte. Nevertheless it must be borne in mind that with regard to the “‘completeness”’ of the Affenspalte, Cunningham (p. 69) “cannot tell whether or not it (the elevation of the wall of the ven- tricle) exists in the anthropoid brain. Specimens of these are so valuable that I am unwilling to destroy those I have got, even in a determination of a point of this importance.” He has only observed the bulging inwards of the outer wall of the ventricle at the bottom of the Affenspalte in the Sooty Mangaby and in Cebus. As the Oxford Museum possesses several chimpanzees’ brains, CEREBRAL CONVOLUTIONS—“ SALLY.” 65 Professor Lankester very generously gave me permission to dissect one of them to ascertain what was the fact with regard to this point. I sliced the hemisphere in planes parallel to the median plane till I cut through the posterior cornu of the lateral ventricle. I find that the “ Affenspalte”’ does not causea bulging in the roof or side of the ventricle. In the figure (fig. 9) there appears to be such a bulging, but this is a portion of the ‘ calcar avis’ passing on the roof of the ventricle, and the slight ridge (#)just in front of it lies distinctly in front of the Affenspalte. Miller, who represents dissections of the brain of a chimpanzee, does not figure or describe any such bulging. But even if there were such an elevation of the outer wall of the ventricle in the lower apes, it does not seem to me a consequence that the higher apes would present one. Even if in the ordinary chimpanzee with a well-marked operculum and deep Affenspalte the impress of the latter is exhibited by the outer wall of the ventricle, it does uot necessarily follow that in the condition in which the Affen- spalte is present in “ Sally,” the same impress would be left. As the operculum disappears, and the Affenspalte becomes proportionately shallower, it seems quite possible that the bulging of the ventricular wall should get less and less marked. In the human brain the “ occipitalis transversus,” as we have seen, replaces the “ fissura perpendicularis externa ”’ after the disappearance of all trace of the latter. It does not how- ever follow, because there is no actual continuity of the two grooves, that there is no genetic relation between them. The obliteration of the earlier furrow is probably due to the growth of the nerve tissue bringing about a thickening in the wall of the ventricle, and “ filling up” (if we may use the expression) the furrow; at the same time the bulging on the inner face becomes smoothed out. But the furrow again makes an effort —inheritance is too strong for the vegetative growth,—and succeeds in obtaining its permanent position as the “ occipitalis trausversus,” which is relatively shallower and does not cause any bulging inwards of the wall of the ventricle. We know that ontogeny does not by any means follow VOL. 37, PART 1.—NEW SER. E 66 W. BLAXLAND BENHAM. slavishly the exact steps of phylogeny: the mode of formation of the heart in mammals, for example, differs from that of birds, nevertheless the two hearts in their entirety are surely homo- logous and homogenetic; yet the ontogenetic condition cannot represent a functional phylogenetic stage. In the ontogeny of some Vertebrates the cesophagus becomes closed at a certain period of development, and it again opens out to form the permanent esophagus. We do not deny the homology (homogeny) of these two structures. It appears to me that the facts which have been gathered together with regard to the Affenspalte in the chimpanzee, and the “ occipitalis transversus ” in man,—the fact also that there is considerable variation in position and extent of the fissure in both animals, and that parallel variations occur in both,—are so strongly in support of Ecker’s view, that the mere fact of a slight difference in ontogeny is not sufficient to overturn the homology. Further, what is the course of development of the “ Affen- spalte” in the ape and monkey? Weare, I believe, absolutely ignorant on the matter. In the human brain (fig. 17) it seems at first sight pretty evident that the sulcus occipitalis transversus is formed by the bifurcation of the intra-parietal (p*.) ; the angle formed by the two branches of the fissure lends its weight to this view. But compare with it fig. 18. Here p*. gives off a branch (z.) behind the arcus parieto-occipitalis, which, there is little doubt, is the mesiad fork of the intraparietal, whilst the laterad fork enters the transverse occipitalis. This fissure itself closely resembles the “ Affenspalte” in ‘ Sally” (fig. 11). Thus Cunningham’s interpretation would apply to fig. 17, and Ecker’s to fig. 18. Now, in some chimpanzees, such as that represented by figs. 30 and 32, we have an “‘ Affenspalte””? below the operculum (fig. 32) into which there falls the ramus occipitalis (p‘.), and at this point an angle occurs in the “ Affenspalte.’’ There is, in fact, a very close resemblance to the condition in the human brain (fig. 17), which Cunningham would explain as CEREBRAL CONVOLUTIONS—* SALLY.” 67 bifurcation of the intraparictal. How are we to distinguish between the two? We can scarcely regard the subopercular furrow of the chimpanzees (fig. 32) as anything but the Affen- spalte. Therefore, why should we not apply the same name to an apparently similar furrow in fig. 17? In fig. 31, again, we have an arrangement somewhat like that in the human brain (fig. 18), viz. the ramus occipitalis (p*.) gives off a post-arcal branch, which is distinct from the (greater part of the) Affenspalte. The figs. 17 and 18 repre- sent two sides of a human brain, and figs. 31 and 32 the two sides of a chimpanzee’s. I do not pretend to decide the matter. It would be great presumption on my part, with my slight experience of human cerebral topography, to express any decided opinion on a matter on which such authorities as Eberstaller, Ecker, and Cunningham are at variance. I merely wish to bring forward other facts and arrangements of these fissures which appear to me to have some bearing on the question. 7. Tue Occiritat Lose. The occipital lobe itself is more furrowed in “ Sally’s” brain than in the ordinary chimpanzee, and bears some resem- blance to that of Miuller’s brain. The old terms given by Gratiolet to the gyri in the neigh- bourhood of the parieto-occipital fissure, and to which so much importance was at first attached—the annectant gyri, or “ plis de passage”—are now gradually being discarded. Since Rolleston and Turner showed their presence in chimpanzee, where they are usually concealed by the operculum, the dif- ferential importance of them has gone; and the names, though still sometimes employed, appear to be giving way to more descriptive and systematic terms. The pli de passage supérieur externe and pli occipital supé- rieur are now called by one name, “ gyrus occipitalis primus ” (Ecker), or “ arcus parieto-occipitalis”’ (Eberstaller). This arcus parieto-occipitalis is well seen on the right hemi- sphere, where it curves firstly round the lateral extremity of 68 W. BLAXLAND BENHAM. the “lateral parieto-occipital,’ and is continuous with the gyrus occipitalis primus, which bends round the mediad ex- tremity of the Affenspalte. The gyrus occipitalis secundus (O 2) (deuxiéme pli de pas- sage externe) lies behind the Affenspalte (in fact, is the opercu- lum) and curves round its laterad extremity, becoming con- tinuous with the “ angularis.”’ The third occipital gyrus lies below the oblique fissure (the sulcus occipitalis inferior), and also passes into the third tem- poral gyrus. The calcarine fissure just cuts the edge of the occipital lobe, but is more extensive on the left than on the right hemi- sphere. 8. Tue Sytvian Fissure. The Sylvian fissure in “Sally’s” brain is less vertical than in the brain figured by Miller, though, as Beddard has re- marked, it is more vertical than in common chimpanzees. The “Sylvian angle” in Dr. Cunningham’s sense is 55° in “ Sally ;” the brain received from Professor Herdman has rather a smaller angle, viz. 50°, whilst normally it is, according to Cunningham, 54°5°. With regard to the “anterior limb” of the Sylvian fissure, Beddard has quite rightly figured it as a short, nearly hori- zontal fissure (Pl. 23, fig. 2, F.s.a.). A comparison of a series of chimpanzee’s brains led Cunningham to believe that it is the anterior boundary of the fronto-parietal operculum, and therefore is the homologue of the ‘ramus ascendens” in man. He gives reasons for his conclusion that in Troglodytes the only opercula covering the island of Reil are the fronto-parietal and the temporal. The two anterior ones, frontal, and orbital of man, are absent; and he suggests that part of the orbital lobule in the ape may represent the island of Reil in man. In “Sally” the insula is distinctly limited anteriorly by a portion of the frontal lobe, which is at a much higher plane than the insula, and separated from it by a deep fissure; but it does not sensibly overlap the insula. However, in other CEREBRAL CONVOLUTIONS—* SALLY.”” 69 chimpanzee brains this portion of the frontal lobe below the “anterior limb ” is much more developed, and forms practically an operculum—overlapping, though only to a slight extent, the insula. . Cunningham states that the submerged portion of the insula presents very little trace of an anterior boundary, and that it ascends gradually along an inclined plane until it finally reaches the free surface of the frontal lobe. I do not find this to be the case at all generally. In two brains in the Oxford Museum this overhanging anterior lobe exhibits more or less distinct traces of a subdivision into two lobules. The most distinct case was photographed, and is represented in fig. 8. Here the “anterior limb” of the Sylvian fissure is forked ; one branch is nearly vertical (Sy’.), and forms as usual the boundary of the fronto-parietal operculum; it is the “‘ramus ascendens ” of the anterior limb. Just below it there is another slight but distinct fissure (Sy’’.) separating a small triangular lobe from the rest of the anterior boundary of the insula (fr. op.). This little lobe, I suggest, is the “ frontal operculum” or pars triangularis, and the second fissure is the homologue of the “ anterior horizontal limb” of the Sylvian fissure. Below it the remainder of the lobe will then corre- spond with the orbital operculum, which here distinctly over- laps the insula. A second brain (Oxford Museum, 947 f.) shows somewhat similar conditions, but in a less degree. These may be com- pared with Cunningham’s fig. 1, pl. iv, of the brain of a human foetus of the eighth month. The brain represented in fig. 8 resembles still more closely the figures given by Schafer in ‘ Quain’s Anatomy’ (tenth edition, vol. iii, part 1). He shows that on the right side of this human brain the two branches of the anterior limb of the Sylvian fissure practically join at their commencement, i. e. the ‘pars triangularis”” or frontal operculum is here less developed than on the left side, where it entirely separates the two rami. The greater development on the left side of the pars triangularis may be in relation with the “centre of speech.” 70 W. BLAXLAND BENHAM. It appears to me particularly interesting to find in the chim- panzee a condition so closely resembling that on man’s right side. Cunningham regards the inferior frontal convolution of the ape as representing non-covered insula, and the “ sulcus fronto-orbitalis” as the homologue of the anterior limiting sulcus of the insula, and compares it with foetal human brains at a certain stage. But the position and relations of the fronto-orbital fissure (e. 0.) with regard to the “ anterior limb ” of Sylvius (limit of fronto-parietal operculum) seem to me to be very different. In most of the chimpanzees it passes upwards from the orbital surface of the frontal lobe some distance in front of the anterior end of the Sylvian fissure, and curves upwards and forwards, extending considerably above the so-called “‘ anterior limbs.” It is easy to understand how the folding backwards—formation of orbital operculum—in the fig. 9, pl. iv of Cunningham’s memoirs would give rise to the state of things represented in his fig. 1 of the same plate representing human foetal brain. Here the fissure supposed to correspond with the fronto- orbitalis joins the “‘ anterior ascending limb” of the Sylvian fissure, and then curves backwards. From the facts referred to above, and represented in the photograph (fig. 8), it seems to me that the sulcus fronto-orbitalis may have some other meaning. May it not be part of the system of orbital furrows which has extended upwards ? 9. Frontat Lose. A system of furrows, more or less parallel to, and in front of, the fissure of Rolando, constitutes the ‘ preecentralis.” This precentral fissure arises, according to Cunningham, in three pieces—vertically, a “ preecentralis inferior,” and a “ prze- centralis superior,” with a horizontal ramus, which is usually connected with the upper part of the former. On the right hemisphere of “Sally’s”’ brain (figs. 11, 12) the sulcus precentralis inferior (p. c. 7.) is divided into an upper and a lower portion by a submerged gyrus situated just above CEREBRAL CONVOLUTIONS—“ SALLY.” 71 the junction of the fissure with the s. frontalis secundus (f?.) ; the upper portion of this vertical fissure appears to represent the “ramus horizontalis” (f4.). The pracentralis superior (p. ¢. 8.) is a very well marked, transversely-placed furrow, seen best in the view of upper surface. The lower-end lies behind the upper end of the ramus horizontalis. It is in free communication with the distinct ‘sulcus frontalis primus ” (or superior) (f1.), which runs straight forwards and bifurcates anteriorly. Lying in front of this is an oblique furrow, with its posterior end directed towards the mesial fissure, which it almost reaches: the anterior end bifurcates, and lies in front of, but mediad of, the end of the sulcus frontalis primus. What is this fissure ? It is represented in the left hemisphere by a smaller and unbranched furrow. Is it a portion of the frontalis primus ? or is it a representa- tive of the “sulc. frontalis mesialis ” ? According to Cunningham, the disjointed portions of the sulcus frontalis primus lie in exactly the opposite direction, i.e. the posterior ends of each are laterad (outside) of the ante- rior end of the preceding. From the fact that these disjointed pieces are slightly variable, it will perhaps be safer to leave the matter open, for hitherto the “s. frontalis mesialis”’ is characteristic of the human brain, and has not been met with in the ape. In front of this, again, is a curved furrow which is probably another portion of the s. front. primus. The s. frontalis secundus (or inferior) has been carefully identified by Cunningham, and he has conclusively shown that the sulcus rectus of the monkey’s brain is the homologue of the s. frontalis inferior, and not of the s. frontalis medius as Eberstaller believed. Cunningham therefore agrees with Gratiolet’s views, which are at variance with some more modern views, that the inferior frontal lobe is present in the apes. In “Sally” the sulc. frontalis inferior (secundus) (fig. 12, f?.) 72 W. BLAXLAND BENHAM. is well developed. Posteriorly it is continuous with the pre- centralis inferior, whence it runs nearly directly forwards for some distance and then bends downwards in front of the orbito- frontalis (e. 0.), when it divides into a fore-and-aft nearly hori- zontal fissure (w.), and can be traced forwards to the frontal pole. The anterior branch appears to be the fronto-marginalis of Wernicke. In his description of this fissure in man, Cunningham states that the hinder limb of the fork lies between the two anterior limbs of the Sylvian fissure. Granting that the s. frontalis inferior is identical in man and apes,—and in “ Sally ” it closely resembles the arrangement figured by Cunningham on p. 248 for man,—the hinder branch of the fork ought to have the game relation to the fronto-orbital furrow (e. 0.) as it has in man to the anterior horizontal limb of the Sylvian fissure, if Cunningham’s identification of the fronto-orbitalis is the true one. But here in “Sally” it has not this position; a fact which is adverse to Cunningham’s interpretation of these other furrows. But in other chimpanzees this fissure marked “ w.” is not related to the “s. front. inferior,” but to the ‘‘ medius,” as in fig. 35, where the s. frontalis inferior is short and its anterior end is surrounded by a curved fissure—part of the “ medius ” (f. m.)—which passes downwards and divides near the orbital surface of the brain into a fore-and-aft horizontal furrow (w.), which can be traced round to the frontal pole. This appears to be the s. fronto-marginalis of Wernicke, and this brain resembles in this respect the one figured by Cunningham (fig. 68, p. 290). In other cases the “ medius ” runs in the same direction, but does not divide (fig. 38). The s. frontalis medius in “ Sally” is unconnected with the s. precentralis inferior, aud is quite a distinct fissure on both hemispheres. Turning now to the left side of “Sally’s” brain (figs. 10, 18), we find the sulcus precentralis inferior (p. ¢. 7.) is a well-marked continuous furrow extending from nearly the Sylvian fissure upwards for about half the surface of the hemisphere. There appears to be no ramus horizontalis, unless it is represented by CEREBRAL CONVOLUTIONS —** SALLY.” 73 a small oblique furrow (f.), recalling the “ precentralis medius”’ of some authors. Cunningham in some cases finds in man the ramus hori- zontalis nearly vertical, or composed of.a forked furrow. This he considers is merely an extreme condition to which the term ‘“‘ preecentralis medius’”’? has sometimes been applied. The other frontal furrows are normally developed. It may be worth while to refer to some of the variations in the arrangement of these furrows in our other chimpanzee brains. In most of them the ramus horizontalis is continuous with the preecentralis inferior; it is most characteristically developed in the brain represented in figs. 338, 34. In fig. 38, again, this latter furrow is very well developed, but has no auteriorly directed horizontal branch. Probably the upper part of it, together with the short posterior horizontal (/.) branch, represents the ramus horizontalis. In fig. 28 the ramus horizontalis is continuous with what appears to be a portion of the frontalis primus. It is in both these cases more or less vertical. The frontalis secundus (f%.) is continuous with the precen- tralis inferior, except in fig. 28 and on the brain represented at fig. 30. In the former the frontalis secundus is connected with a downwardly directed vertical furrow, which is very much more distinct in fig. 33. This condition may be compared with that of an eighth month human fcetus figured by Cunningham on p. 250. In the brain represented in figs. 30, 34, on the right side the precentralis inferior (p. ¢. 7.) is provided with a large ramus horizontalis (2.) composed of an anterior sagittal and a posterior constituent directed obliquely upwards. The s. frontalis medius (f. m.) and the “inferior” (f*.) are separate from the precentralis and are nearly simple furrows ; but the “inferior” bifurcates at its extremity. This bifurcation probably represents the s. fronto-marginalis. 8. frontalis primus (/1.) is in three separate pieces, the hindermost being continuous, as usual, with the s. preecentralis superior (p. ¢. s.). On the left side of the same brain (fig. 33) a very peculiar 74, W. BLAXLAND BENHAM. association of furrows occurs, somewhat like that in fig. 35. The ramus horizontalis is connected, as usual, with the s. precentralis inferior. The s. frontalis medius is a curved fissure, concave downwards, lying some little distance in front of the precentralis inferior; from this is given off a branch which runs forward—a modification of the ordinary forking of this fissure. But where is the ‘‘ inferior ’’? Just below the medius and lying parallel to the ramus hori- zontalis is a very short fissure (separated from the preecentralis inferior by a gyrus, which just reaches the surface), crossing the top of a vertical fissure of considerable extent. This latter passes downwards nearly to the anterior limb of Sylvius. This fissure appears to represent the sulcus diagonalis, and the horizontal one at its upper end the s. frontalis inferior. We have in this chimpanzee brain a condition resembling the foetal human brain of eight months figured on p. 250 by Cunningham—the “ inferior,” however, being shorter in the chimpanzee. The brain represented by fig. 38 is of interest in that the s. frontalis secundus is divided into two pieces by a broad gyrus, one portion remaining continuous with the precentralis inferior—a condition resembling that of a seventh month human foetus figured by Cunningham on p. 276, fig. 63; and it is probable that the part of the curved longitudinal fissure in fig. 33, labelled? f2., is a similarly disjointed portion of the “ frontalis secundus.” It is unnecessary to enter into further detail, as an examina- tion of the figures will show the interpretations which I put upon the various fissures. 10. Resunts oF THE CoMPARISON OF A SERIES OF CHIMPANZEE BraIns. 1. The parieto-occipital fissure, originally a simple and single incision on the mesiad side of hemisphere, frequently in the chimpanzee and more rarely in man becomes divided into two fissures by the gyrus intercuneatus becoming superficial. Of these two portions, the superior, or lateral parieto-occipital, CEREBRAL CONVOLUTIONS—“ SALLY.” 75 may come to be more or less entirely on the surface of the hemisphere (figs. 11, 16, and 31), whilst the lower or mesial portion lies on the mesial surface. This “lateral parieto-occi- pital ” is independent of the A ffenspalte, and is not synonymous with what is frequently termed in text-books the “ external parieto-occipital.” 2. The occipital operculum in the chimpanzee presents very great variations in its size; the two hemispheres of a given brain frequently present differences in respect of this feature (cf. figs. 10, 11, 19, 20, 25, 26, 30). In “Sally” this operculum is practically absent, and the subopercular groove or ‘ Affen- spalte ” is fully exposed. 3. This “ Affenspalte,” or Simian fissure of the ape, seems to be homologous with the “sulcus transversus occipitalis” of Ecker, and to be independent of the bifurcation of the intra- parietal fissure. 4, In some chimpanzees there exists a ‘ramus ascendens ” and a “ ramus horizontalis” of the anterior limb of the Sylvian fissure ; these enclose a “ pars triangularis” or frontal opercu- lum overlapping the insula (see fig. 8). Further, I find that the anterior boundary of the insula is more or less distinctly marked, and that there is a distinct and sudden elevation of the frontal lobe at the anterior margin of the insula. In other words, an “orbital operculum” is present in some chimpanzees, though it may be feebly marked in some speci- mens. 5. There is a fissure in “Sally’s”’ brain which may possibly be the homologue of the sulcus frontalis mesialis (figs. 10 and 11, “? f!.”), though I have provisionally taken it to be a dis- jointed portion of the sulcus frontalis primus, since the former sulcus has not hitherto been recognised in any ape brain. 6. Various arrangements of the frontal fissures in chim- panzees resemble those recorded in man, either adult or foetal. 76 W. BLAXLAND BENHAM. 11. List oF THE MORE ImpoRTANT MeEMOIRS DEALING WITH THE BRAIN OF THE CHIMPANZEE. 1. Gratrotet.—‘ Mémoire sur les Plis cerebraux de l’homme et des Pri- mates.’ 2. Marsuatt, J—‘ On the Brain of a Young Chimpanzee,”’ ‘ Nat. Hist. Review’ (new ser.), i, 1861, p. 276. 8. Biscuorr, T. W.— Die Grosshirnwindungen des Menschen,” ‘ Abh. d. k. Bayer. Akad. d. Wiss. Math. Phys. Classe,’ x, 1866. 4, Turner, W.—‘‘ Notes more especially on the Bridging Convolutions in the Brain of the Chimpanzee,” ‘ Proc. Roy. Soc. Edin.,’ v, 1866. 5. Biscuorr.—‘ Ueber das Gehirn eines Chimpanse,” ‘Stzber. d. Wiss. Math. Phys. Classe d. k. Akad. Bayerish.,’ 1871, Miinchen. 6. Broca.— L’Etude sur le cerveau du gorille,” ‘ Rev. d’Anthrop.,’ 1878. 7. Cuapman.— On the Structure of the Chimpanzee,” ‘ Proc. Acad. Nat. Sci. Philadelphia,’ 1879, p. 52. (Gives literature.) 8. Parker.— On the Brain of a Chimpanzee,” ‘ Medical Record,’ 1880. 9. Mijttmr.— Zur Anatomie des Chimpanse Gehirns,” ‘Arch. f. Anthro- pologie,’ xvii, 1888, p. 173. 10. Symineton.—‘‘ On the Viscera of a Female Chimpanzee,” ‘ Proc. Roy. Phys. Soc. Edin.,’ x, 1890, p. 297. 11. Bepparp.— Contributions to the Anatomy of the Anthropoid Apes,” ‘Trans. Zool. Soc.,’ xiii, 1898. 12. Cunnincuam.—“ Contribution to the Surface Anatomy of the Cerebral Hemispheres,” ‘Cunningham Memoirs,’ vii, Roy. Irish Acad., 1892. 12. Tue BRAIN OF THE ORANG-OUTANG. The arrangement of the frontal convolutions in this ape is so varied, and the fissures have in many specimens such pecu- liar relations, that Dr. Cunningham, in the few remarks he makes on the subject (loc. cit., p. 295), confesses his un- certainty as to the accuracy of the identification which he places upon these fissures. As the Oxford Museum possesses two brains of the orang, and as these differ from one another, as well as from that described by Cunningham, it has seemed to me worth while to place on record the plan of these convolutions, and to endeavour in the light of these new variations to arrive at a more accurate determination of these fissures. Even if I am CEREBRAL CONVOLUTIONS—‘* SALLY.” Pari not successful in this—and as so great an authority as Dr. Cunningham has confessed his inability to thoroughly establish his identifications I feel some hesitation in believing that I should be more successful—I shall at any rate have added new facts to the comparatively small amount of knowledge that we possess concerning the orang’s brain. Of the two orangs’ brains one is historical, being that de- | scribed and figured by Rolleston in 1861 (2). It is very well preserved. The other brain, however, is rather soft and slightly injured on the right side. It is, further, rather distorted, the hemi- spheres being pressed over to the left side, so that nearly all the temporo-sphenoidal lobe is thrust under the lower surface, against which it is flattened. It appears to have belonged to a young individual, as it is not nearly so large as those hitherto figured. The cerebral hemispheres measure 84 mm. in length, but it is useless to give other measurements owing to the distortion of the brain. This brain will be referred to as Brain No. 1, that described by Rolleston as No. 2. Orang’s Brain No. 1 (figs. 42—45). The Frontal Lobes.—On the left side (figs. 42 and 44) the s. precentralis inferior (p.¢.7.) is continuous above with an obliquely placed fissure, directed upwards and forwards (A.), which I would identify as the ramus horizontalis, since the poste- rior end of the s. frontalis medius lies below the anterior end of it. About midway along its length, the s. prec. inferior gives off an anteriorly directed furrow, which I would identify as a part of the s. frontalis secundus (f?.). Such an arrangement I have described above in a chimpanzee (fig. 38); and which, moreover, is very similar to that figured by Cunningham for an adult man on p. 262 of his memoir. The s. preceutralis superior (p.c.s.) seems to be in two portions, the lower of which, lying parallel to the ramus hori- zontalis, is continuous with the s. frontalis primus (/'.), which in this hemisphere is represented by only this one fissure. 78 W. BLAXLAND BENHAM. The s. frontalis medius (f.m.) is a simple fissure passing forwards nearly horizontally, its hinder end having precisely the typical position below the ramus horizontalis, from which it is separated by a deep-lying gyrus. The s. frontalis secundus (f?.) is in two parts, the chief of which lies just below the s. frontalis medius and passes forwards to the frontal lobe, giving rise to the “s. fronto-marginalis ” of Wernicke. The other part of the fissure to which I have referred above is in continuity with the s. preecentralis inferior. On the right side of this brain (figs. 43 and 45) the furrows on the frontal region have rather a different arrangement. The sulcus precentralis superior (p.c. s.) is single and of consider- able length, occupying the position of the two furrows of this name on the left side. It lies parallel to the fissure of Rolando, and at about half its length is joined by a longitudinal fissure, which appears to be the main part of the sulcus frontalis primus (f1.). The sulcus precentralis inferior (p.c.i.) is also fairly well developed; upwards it lies in front of the sulcus precentralis superior, and below it nearly reaches the Sylvian fissure, from which it is separated by a deep gyrus. Passing from it ante- riorly are two horizontally directed furrows, the upper of which appears to be the ramus horizontalis (/.), the lower and shorter is prohably a part of the sulcus frontalis secundus (f*.). Where, now, is the sulcus frontalis medius? It ought to lie below the ramus horizontalis. There is a furrow in this position which curves upwards round the anterior end of the ramus horizontalis, part of which possibly represents this sulcus frontalis medius. Unfortunately the brain is so ill pre- served in this region that the exact arrangement of fissures on the lower part of the frontal lobe cannot accurately be made out. Orang No. 2 (figs. 39—41). In Rolleston’s specimen the upper part of the right hemi- sphere has been removed, but is preserved. The following is the arrangement :—On the left side (figs. 39 and 41), lying in CEREBRAL CONVOLUTIONS—“* SALLY.” 79 front of and nearly parallel to the upper part of the fissure of Rolando is a conspicuous fissure which passes for some dis- tance forwards, nearly parallel with the median fissure, for about half the length of the lobe. This appears to repre- sent a portion of the sulcus frontalis primus (f1.), together with the uppermost part of the sulcus precentralis superior (p.¢.8.). At a point a short distance behind its anterior termination a fissure passes downwards and slightly backwards, nearly parallel with the fissure of Rolando, and before it ceases gives rise to another fissure directed nearly horizontally forwards (h.). zak 4 Grains zie! bi aS Mae Ste 4 } reli wv ACP ft pon que ( \ T ; : _ - 4 oer By Biowiens .th* a te Tle ae r ni! 7 ’ ( : ( ; : ‘ ‘ . vii} i a. ane } ' Tar ° ee 7, j Livezie cn . hap ee te esr es | ti eit | 4 vy oer Gee) Le ae x hol wine ania) le Aiud Cus H et (item ? gah iicea Ail? imdpmowen neers 0 SLE ft ib aPTP 2s ¥ rites ‘ ‘ 7) or . 7 : phe Ma ba TTS) aL) gent Met OF a Ph ea wi pais ae j Hix 4) i apn Lhe F: nua mH 3 } $ niry bs { Ai a 5 5 . t A Go nA i edd PP cay A uM ’ Pot ay PE th : ; MA Brojgctel bt See aT pila rigelapll ip Dy (ea Oe Pe SW LL & itlase fs) LS er aa sat pry ee Reh ear forse Det Ga ae att ty alt riers uitas Width eA. A i Wierd Wot bs ay | , bAe tue ia Oks Jie See dil ot jie? >, ee re ar me ae) ea nie dey, ys! ry y Vora ck ew OR See it: beste Te MRT ROI > SV 4) ee GE a UeulAi oy Se y p : Ol eae hina < - poly able haan? MP ae A $2 ee 7, wt alent thts 7 a | EN, i het ee > o risa Loxaten ot eee 7 wept ' Poiiela at) “pdt Wrala dy chris AiG LP ane . F ae » Aronian: xii . ie ; x Ayn erin fit. waite ‘ se gat 8 mee Wi fot ihe. ioe hetieme Aig ail Nites 7 pee le ‘ me na) fo4 ; Pa ye . ' may indo Ob! BRR ie AE ead byt Nate! : } ’ ie mat a equa sail ue nt dag Ss ped {seid} ee ie 2 ws “Willie, ; Weer ON THE CELLULAR THEORY OF DEVELOPMENT. 87 On the Inadequacy of the Cellular Theory of Development, and on the Early Development of Nerves, particularly of the Third Nerve and of the Sympathetic in Elasmobranchii. By Adam Sedgwick, F.R.S. Ir is now more than ten years ago since I first pointed out the inadequacy of the cellular theory of development. That I did so in a very guarded manner need hardly be said; but now, after ten years of mature work, I feel justified in giving a stronger expression to the views which I then formed, and which all my subsequent work has amply confirmed. My words then (in 1883) were as follows :—‘‘In short, if these facts are generally applicable, embryonic development can no longer be looked upon as being essentially the formation by fission of a number of units from a single primitive unit, and the co-ordination and modification of these units into a harmonious whole. But it must rather be regarded as a multiplication of nuclei and a specialisation of tracts and vacuoles in a continuous mass of vacuolated protoplasm.” Again, in 1888, in the preface to my “ Monograph on the Development of the Cape Species of Peripatus,”? I wrote: “It would appear, indeed, that in Peripatus the cells of the adult, in so far as they are distinct and sharply marked off structures, are not, as appears to be generally the case, present in the earliest embryonic stages, but are gradually evolved as development proceeds. In other words, the cell-theory, if it 1 *Studies from the Morphological Laboratory of the University of Cambridge,’ vol. iv, part 1. 88 ADAM SEDGWIOK. implies that the adult cells are derived from embryonic cells which have been directly produced by the division of the ovicell, does not apply to the embryos of Peripatus.” In the days when these words were written it was a general belief among leading histologists and physiologists that the connections which were known to exist in some cases between adult cells had arisen secondarily, and that the primary con- dition brought about by the cleavage of the ovum was a com- plete separation from one another of these units, of which the body was supposed to be composed. There has been, no doubt, a change of opinion since those days, and although many biologists would still maintain that cleavage is complete and results in the formation of separate units which later become connected, there is a constantly increasing number who would consider themselves misrepresented if one imputed to them this belief not long ago universal, and the belief which was supposed to follow from it, that the first stage in the evolution of the Metazoa was a colonial Protozoon. But, as I have said, opinions have changed since those days, and I quote my words, written then, to show that I have long held the view which I am now expressing, and that I was among the first to attack a theory which had even then passed its stage of usefulness, and is now holding men’s minds in an iron bondage. For although opinions have changed on this im- portant subject, and although there are some who think that they have escaped from the domination of this fetish of their predecessors, yet as a matter of fact the cellular theory of development is still rampant, still blinds men’s eyes to the most patent facts, and still obstructs the way of real progress in the knowledge of structure. In order that I may not be met with the statement that such a state of things exists only in my own imagination, that I am putting up a dummy merely to knock it down again, it is necessary that I should give some proof that this hypothesis has still the power which I ascribe to it. What is the cellular theory of development? I am not concerned with what its authors held; what we want to know is, what is the present ON THE CELLULAR THEORY OF DEVELOPMENT. 89 form and extent of it? What is the point of view which it compels its votaries to take ? It is not easy to answer this question ; it is, in fact, as diffi- cult to answer as that other question so often asked of the teacher by his pupil—what is a cell? The source of the dif_i- culty is that we are dealing with a kind of phantom which takes different forms in different men’s eyes. There is a want of precision about the cell-phantom, as there is also about the layer-phantom, which makes it very difficult to lay either of them. Neither of these theories can be stated in so many words in a manner satisfactory to every one. The result is that it is not easy to bring either of them to book. To answer the question—what is the cellular theory of deve- lopment ?—the best plan will be to consider for a moment the ideas which are taught to the student of biology, and which influence him in his future work. We tell him that the cell is the unit of structure, that an organism may consist of a single cell, or of several cells in association with one another: we draw the most fundamental distinction between the two kinds of organism, and we divide the animal kingdom into two great groups to receive them. Asa proof of the importance which we attach to this feature of organisation we assert that a man is nearer, morphologically, to a tapeworm, than a tapeworm is to a parameecium. We tell him that the various structures present in a protozoon are all parts of one cell, whereas in a metazoon the various parts are composed of groups of cells which differ from one another in structure. Finally, when we ask him in the examination to tell us the principal differences between hydra and vorticella, we consider that he is very inadequately prepared if he does not sum them up by saying that hydra has tissues composed of definite cells and is multi- cellular, while vorticella is without definite cellular tissues and is unicellular. Carrying on the idea thus implanted in his mind as to the fundamental importance of the cell, we tell him about the neuro-epithelial cell and the myo-epithelial cell, and we point out their primitive distinctness,—an idea which is still further impressed upon him when he studies the 90 ADAM SEDGWIOK. connection between nerve and striated muscular fibre. Finally, when he comes to study embryology, the importance and dis- tinctness of the cell meets him at every step, from the complete cleavage which he is led to believe is primitive, to the develop- ment of nerves according to the views of His. So much for the student in the schools: now for the investi- gator in the laboratory. He studies the ovum and maintains its absolute isolation in the organism ; or he examines epithelial cells and draws them as isolated structures separated by sharp boundary lines ; or he lahours to prove the continuity between the nerve and muscle, or between the nerve and secreting cell: so much is he dominated by the idea of separate cells that he considers that the burden of proof rests rather with the man who asserts such continuity than with him who denies it. Or, if he be an embryologist, he will talk of, and figure, the proliferation of cells at the primitive streak ; he will describe the nascent ganglion cell sending a process from the developing spinal cord into the anterior root, and he will figure it; he will talk of mesenchyme cells, and figure them for the most part separate from one another. I take it that this is a not unfair account of the training a zoologist receives at the present day, so far as the cell is con- cerned, and of the ideas which dominate him in his later work. He believes that the cell is the unit of structure, and that it forms the basis of organisation in the Metazoa; it is the func- tions of the cell and the relations which it enters into with other cells which forms an important subject of current biolo- gical investigation. Who, then, can deny that the cellular theory of development is still a living power in the school of biology ? That it blinds men’s eyes to the most patent facts, and obstructs the way of real progress in the knowledge of structure, it will now be my endeavour toshow. For this pur- pose I shall deal on this occasion with the origin and structure of two tissues of the Vertebrate embryo—the so-called mesen- chyme and the system of peripheral nerve-trunks. My results are the product of many years’ work, and will, I hope, be published in greater detail and with figures on a future occasion. ON THE CELLULAR THEORY OF DEVELOPMENT. 91 The So-called Mesenchyme Tissue of Elasmobranch Embryos. This tissue is always described as consisting of branched cells lying between the ectoderm and the endoderm. The cells are spoken of as being separate from one another, and from the adjacent ectoderm and endoderm, excepting at points where they are supposed to arise from one of the primary layers. And not only are they described as being separate cells, but they are actually drawn in the author’s figures as separate from each other. This is, perhaps, the best instance that can be given of the bondage in which the cellular theory holds its votaries. For what are the facts? The separate cells have no existence at all! In their place we find, on looking into the matter, a reticulum of a pale non-staining substance holding nuclei at its nodes. It is these nodes, with their nuclei, which are drawn by authors as the separate branched cells of the mesenchyme, and they are constrained by this theory, with which their minds are saturated, not only to see things which do not exist, but actually to figure them. Another erroneous view due to the same cause is the view that this mesenchyme tissue is not continuous with the ectoderm or with the endoderm ; whereas, as a matter of fact, the opposite is the case, for the primary layers are simply parts of this reticulum in which the meshes are closer and the nuclei more numerous and arranged in layers. These are facts of which anyone with an unbiassed mind can convince himself by the simple inspection of a Selachian or an Avian embryo, and they would have been recognised long ago had it not been for the dominating influence of the cellular theory of development. The current views as to the origin of this tissue show just as conspicuously the influence of the same theory. It is said to arise by the budding-off and migration of cells from the walls of the embryonic cclom, from the primitive streak, and from the neural crest ; and the space between the ectoderm and endoderm into which these cells migrate is described as being empty of structural elements. What are the facts? The 92 ADAM SEDGWICK. space between the layers is never empty; it is always traversed by strands of a pale tissue connecting the various layers, and the growth which does take place at the places mentioned is not a formation of cells, but of nuclei which move away from their place of origin and take up their position in this pale and at first sparse reticulum which exists between the layer. As this reticulum, which has always existed, becomes infested with nuclei it increases in bulk, and forms the conspicuous reticulate tissue which is by some authors called mesenchyme. The primitive streak, the walls of the ccelom, and the neural crest, and, as Goronowitsch! has shown, parts of the ectoderm, are growing points where nuclei, not cells, are produced. These facts I described long ago in the development of Peripatus, and it is the recognition of the same processes taking place in the Vertebrata in an even more conspicuous manner that has induced me to again call attention to their importance.” The Origin of Nerve-trunks and the Fate of the Neural Crest. If there is one point more than another on which the cellular theory of development has led anatomists completely astray, it is uponthis one. We may take it that the new views upon the origin of the peripheral nerves began with Balfour’s discovery of the structure which is generally called the nerve crest. Before that discovery nerves were supposed to develop in situ in the mesoderm; after it, there were two principal views as to the origin and growth of nerves: one of these was that cells of the central organ grew outwards as strings to the periphery ; while, according to the other, nerve-fibres are the elongated 1 ¢Morpholog. Jahrb.,’ Bd. xx, 1893. 2 At the same time Peripatus shows certain features more clearly than the Vertebrate; I would refer especially to figs. 24d and 26d on pl. v of my Monograph, in which, while the so-called ectoderm and endoderm are obviously parts of the same layer, or tissue : they are separated by a region in which the vacuoles are larger, the protoplasmic strands less numerous, and nuclei are conspicuous by their scarcity. ON THE CELLULAR THEORY OF DEVELOPMENT. 93 processes of cells either of the central organ or of the ganglia. Both these views are erroneous; and if both were not inspired by the cellular theory of development, they were both promul- gated at a time when that theory was at its zenith. The earlier view, that nerves were developed in situ from the mesoderm, was much nearer the truth. The nerve crest does not, as was first stated by Balfour and afterwards by all authors on the development of nerves, give rise exclusively, or even principally, to nerves and ganglia. It gives rise to nuclei which spread out in, and add to the meso- blastic reticulum, which at all times, i.e. from the very begin- ning, exists between the layers, and to nuclei which become the nuclei of the rudiments of nerve ganglia. The nerves are developments of the reticulum; they are elongated strands of the pale substance composing the reticulum, with some of its nuclei; and their free ends branch out into the fibres of the reticulum, and are added to by the latter falling into the line of the growing nerve. Neither they nor the ganglia appear until the nerve crest is breaking up. The reticulum further gives rise certainly to smooth muscular fibres, connective tissues, and blood-vessels, and probably also to striated muscle. It is also continuous with all the so-called epithelial tissues of the embryo; indeed this latter substance is to be regarded as consisting only of one or more layers of nuclei embedded in the outer part of the reticulum, which is rather denser than elsewhere in correspond- ence with the greater density of the nuclei. Nerves are a gathering up, so to speak, of the strands of the reticulum into bundles, and are formed in that way; or, to put the matter in another way, nerves are a special development of the reti- culum along certain lines. These special developments are generally marked by an increase in the number of nuclei, such increase being particularly great in the neighbourhood of the ganglia. To sum up the matter, the nervous and muscular tissues are, as they were in Peripatus (see my Monograph, p. 131), special developments of the same primitive reticulum, a com- 94, ADAM SEDGWICK. munity of origin which renders their adult relations perfectly intelligible. Further, I have no hesitation in saying that His’ descriptions of the development of nerve-fibres as processes of central or ganglionic nerve-cells, does not apply to Selachians; inasmuch as nerves are laid down long before any trace of nerve-cells can be made out. The neuroblasts of His and of other authors are nuclei lying in a substance which, after death caused by the ordinary reagents, has usually a fibrous structure. This substance is continuous with, and therefore a part of, the reticulum outside. The cell-processes which have been described as growing out from the neuroblasts are merely parts of this reticular substance, the fibres of which become arranged more or less.in the direction of the long axis of the nuclei, and the meshes correspondingly drawn out and narrowed. Many of His’ drawings even show that this is so, and an inspection of the specimens leaves no doubt at all about the matter. Inshort, the development of nerves is not an outgrowth of cell-processes from certain central cells, but is a differentiation of a substance which was already in position; and this differentiation seems to take place from the medullary walls out- wards to the periphery, both in the anterior and posterior roots, and to precede, or to proceed pari passu with, the development of other tissues. The nerve crest is, then, to be regarded as a centre for the growth of nuclei, which spread into the body of the embryo and be- come concerned in the formation of many tissues, nervous tissues amongst the rest. There are many other such centres for the production of nuclei; for instance, I may mention the walls of the ccelom, the caudal swellings, and in the Amniota the primitive streak. All these centres of growth are in so-called epithelial tissues. This is, of course, necessi- tated by the fact that Selachian embryos are at one stage composed entirely—or almost entirely—of these so- called epithelial tissues; as are many embryos, e. g. those of Peripatus and of Amphioxus.! These facts will be dis- 1 The significance of this epithelial structure of the young embryo—this ON THE CELLULAR THEORY OF DEVELOPMENT. 95 puted by many morphologists, but they are easy of proof by the simple inspection of good preparations to minds not warped by the cellular theory as ordinarily taught. In fact, had it not been for the undue persistence of this hypothesis beyond the time of its fruitful life, they would have been recognised long ago, and much needless waste of labour in trying to make the facts of nerve-development conform to the theory would have been saved. The nerve-crest in Selachians (Scyllium, Acanthias, Raia, and Pristiurus) is, as I pointed out some time ago (* Notes on Elasmobranch Development,” ‘ Quart. Journ. of Micr. Sci., vol. xxxiii), from its first appearance, in three pieces.! The first of these pieces reaches from the region of the fore-brain to the hind brain. The posterior limit of it is marked in older embryos by the root of the trigeminal nerve. It gives rise to the reticulum of the front part of the head, and contributes to that of the mandibular arch. The following nerves are formed within its limits :—The trigeminal and its branches, which include the so-called ramus ophthal- micus profundus with the ciliary ganglion and the third nerve (see below). Very possibly other nerves, viz. the fourth, the sixth, and the olfactory, may be also developed from this part of the reticulum, but I have no observations on this point. The manner in which these nerves are laid down may be described as follows:—When the nerve-crest, which in this region of the head very early spreads ventralwards on each side of the brain, is breaking up into the reticulum, certain tracts of it remain unaltered and characterised by a greater density of nuclei. These tracts mark the course of the future nerves and the sites of the future ganglia. They them- collection of the nuclei at the surfaces, as it may be described—I hope to con- sider in another place. Now, I may merely hint that it is probably due to the impress of some well-marked larval phase in earlier stages of evolution (see my article on “ von Baer’s Law, &c.,” in ‘ Quart. Journ. Mier. Sci.,’ vol. xxxvi). 1 Goronowitsch (‘ Morph. Jahrb.,’ Bd. xx) has recently found the same fact for the bird, but he makes no reference to my results on this point. 96 ADAM SEDGWIOK. selves continue to break up, but a kind of core remains which constitutes the foundation of the future nerve and ganglion. The Gasserian ganglion, the ophthalmicus pro- fundus, the mandibular branch of the fifth, and the ciliary ganglion thus gradually emerge from the remains of the nerve- crest—are, so to speak, crystallised out of it. At first they have the form of dense cords of nuclei; but they soon acquire some of the non-staining fibrous substance, which makes its appearance as arule in their central portions, so that fora time sections of these nerves exactly resemble in appearance sections of the nerves of Invertebrata, e. g. Peripatus, Chiton, &c. This description holds for an embryo of 35 mm., beyond which stage I have no observations. The nuclei which have peeled off, leaving the nerve-trunk below, give rise to the muscular and connective tissues of the parts concerned, the reticulum of which is freely continuous with that of the nascent nerve, especially at the free end of the latter. It thus be- comes apparent that these tissues—nervous, muscular, connective, and vascular—are all developed in continuity. While the Gasserian ganglion, the mandibular branch of the fifth, the ophthalmicus profundus, and the ciliary ganglion all crystallise out of the nerve crest; the third nerve does not do so. It arises as a differentiation of the reticulum formed by the breaking up of the nerve crest, and it first makes its appearance as a forward projection of nuclei from the ciliary ganglion. This, by a gradual differentiation of the reticulum, extends itself until it reaches the base of the mid-brain, with which it becomes continuous by means of an increase in the pale fibrous strands which pass between the medullary wall and the reticulum. The third nerve is at first a cord of nuclei and rather dense pale substance. The third nerve,’ there- 1 The continuity of the embryonic tissue which will give rise to the nervous and muscular tissues is well seen in the embryo of Peripatus capensis, and I have already hinted at this fact in my Monograph on the development of that species at pp. 131 and 138, and figured the tissue as nerve musc., pl. x, fig. 5. 2 Tt will be evident, if my observations are correct, that I have found an earlier stage of the third nerve than Dohrn describes in his sixteenth study. In ON THE CELLULAR THEORY OF DEVELOPMENT, 97 fore, presents this interesting and remarkable pecu- liarity in Scyllium and Acanthias; it grows or is differentiated from the ciliary ganglion to the floor of the mid-brain, and not in the opposite direction, as has hitherto been supposed. The proof of this is to be found in the fact that in a Scyllium and Acanthias embryo of 10 to 11 mm. the third nerve can be seen projecting forwards from the ciliary ganglion, and ending in front in the reticulum, short of the floor of the mid-brain. The ciliary or profundus ganglion is at one time—when it is first laid down—in contact with the ectoderm. Later it is shifted inwards, but remains connected for a time with the ectoderm by a cord of cells, which eventually disappears. This point has been seen by van Wijhe. The embryonic medullary wall! is connected with the reticulum by pale fibres similar to those which compose the reticulum, and the nerve-roots, both anterior, posterior, and cranial, are special enlargements of such connecting strands. They are formed at a time when no structures which could be called cells by any but a fanatical devotee of the cellular theory are present, either in the medullary wall or in the ganglionic rudiments, and in a manner which, if closely followed, renders it quite impossible to speak of growths one way or the other, excepting that one can make one assertion—the pale fibrous substance which marks the nerve appears both in the anterior and posterior roots and in the cranial nerve-roots next the central organ, at a time when the white matter (which is composed of this pale fibrous sub- stance) first appears as a thin layer, and in continuity with such white matter. The differentiation outwards pro- ceeds from this point, and the free end of the nerve- rudiment always ends by branching out into the fibres my fuller paper dealing with this subject I hope to examine Dohrn’s results in detail. 1 Tnasmuch as the nerve-crest is derived from the medullary wall and gives rise to mesodermal structures, the medullary wall itself gives rise, in part, to mesoderm. VOL. 37, PART 1.—NEW SER. a 98 ADAM SEDGWICK. of the reticulum. The only exception to this rule is the third nerve of Scyllium and Acanthias (and probably others), which is undoubtedly differentiated from the ciliary ganglion to the floor of the mid-brain; but this is, perhaps, more an apparent exception than a real one, because the ciliary ganglion belongs to the fifth nerve and the order of fibrous differentiation is normal, viz. from the root of the fifth nerve, through the ciliary ganglion, to the floor of the mid- brain. I commend this observation on the development of the third nerve to the physiologist, with a view to a renewed investigation of its functions. It is rendered the more inter- esting by the fact that in Lepidosiren it is commonly stated that the area of the third nerve is supplied by the ophthalmic branch of the fifth, the third nerve being absent.1 I have already, in my ‘ Notes on Elasmobranch Develop- ment,’ stated my reasons for believing that the views put forward by Hensen as to the origin of nerves were nearer the truth than those of any other zoologist. I have, in this paper, shown not only that the network required does exist, but also how it arises, and how it gives rise to the rudiments of the peripheral nerve-fibres. Minot, in his ‘ Human Embryology,’ p. 624, says that Hensen’s theory of the origin of nerves “cannot be adopted because the outgrowths of the nerve-fibres have been observed; moreover, Altmann has pointed out that the fibres seen in the embryonic mesoderm are really pro- cesses of the mesoderm cells, and, as shown in the excellent fig. 2 of his plate, are quite distinct both from the ecto- derm and endoderm.” (Theitalics are mine.) This passage is, according to my work, full of errors; for I maintain, as the result of long and careful observation, extending over many years, that the outgrowth of nerve-fibres from cells in the ganglia and medullary wall not only has not, but cannot 1 This statement rests on Hyrtl’s work. It must, however, be remembered that his specimen was confessedly rotten in its nervous tissues, and by the fact that v. Wijhe (‘ Nied. Arch. f. Zoologie,’ Bd. v, 1882) has found the third nerve in Ceratodus. Parker does not deal with the brain and nerves in his memoir on Protopterus. ON THE CELLULAR THEORY OF DEVELOPMENT, 99 be observed ; that the fibres in the embryonic mesoderm are not processes of mesoderm cells (as they are always figured), which have no existence, but are parts of the reticulum which has always existed from before cleavage onwards, connecting together the various parts of the developing ovum, and that this reticulum is not separate from ectoderm and endoderm, but freely continuous with both, they being but parts of it. The almost universal practice of drawing this reticulum as composed of separate branched cells is a most remarkable instance of the manner in which a theory can blind men’s eyes to the most obvious facts. Before concluding this general account of my work, I may mention one or two other points of general interest which I have noticed. Firstly, I may mention that in Scyllium there are a number of anterior roots next the head, varying in number from three to five, according to the age of the embryo, without posterior roots. They no doubt give rise, as has beeu suggested by others, to the so-called anterior roots of the vagus. Secondly, Balfour was quite correct in the account he gave of the origin of the sympathetic ganglia in Elasmobranchs.! The ganglia arise as swellings on the posterior roots of the spinal nerves, and soon become removed from the latter, so as to form isolated masses connected with the spinal nerves by a cord. These masses eventually become united longitudinally into a chain. I may add to Balfour’s account this fact, viz. that no sympathetic ganglia are found within the area of extension of the vagus ganglion. Or, if I am not correct in applying the term “ vagus ganglion” to the posterior part of the vagus—the part which lies dorsal to the gill-slits and gives off the branchial nerves, it would be better to say that sym- pathetic ganglia are not found in the region of the branchial slits, but begin immediately behind these structures. Thus, in an embryo of 22 mm. the vagus ganglion and branchial 1 T have not examined mammals on this point, but I think Paterson’s memoir (‘ Phil. Trans.,’? 181) does not carry conviction. On the contrary, there is, 1 think, in it internal evidence which inclines me to the view that he has not got to the bottom of the matter. 100 ADAM SEDGWICK. region of the fore-gut ends at the level of the fifth anterior root, and the first posterior root and the first sympathetic ganglion occur at the level of the sixth anterior root. In older embryos, in which the branchial region extends much further back and overlaps a number of fully-formed spinal nerves, the original sympathetic ganglia which were formed in connection with the ganglia of these spinal nerves thus over- lapped are found to have disappeared. The first sympa- thetic ganglion appears always to be just behind the branchial region, as in the adult, and sympa- thetic ganglia are formed in Scyllium in connec- tion with nerves which are without a posterior root. Gaskell reproaches v. Wijhe with not knowing the true meaning of a sympathetic ganglion, and one is tempted to ask, does Gaskell himself know much more about it, or throw any light upon the question? He says (‘Journal of Physiology, vol. x, p. 162) that a sympathetic ganglion is the ganglion of the anterior root of a spinal nerve which has travelled to a variable distance from the central nervous system. As Dohrn (seventeenth study, ‘Naples Mit.,’ Bd. x) very properly in- sists, this view is at variance with the known developmental history of the ganglion—which I am able to confirm so far as its nuclei are concerned, and with the reservations necessitated by the views set forth in this paper—and I am now able to state that it is at variance with the fact that sympathetic ganglia are entirely absent from those spinal nerves in which the posterior root fails to reach its full development. In fact, one may say of these ganglia that they are always absent when the posterior roots are not developed. With regard to the fate of the neural crest described in this paper, I should mention that I strongly hinted that it gave rise to nuclei which entered the reticulum in my ‘ Notes on Elas- mobranch Development,’ p. 581, published in 1892; and that Goronowitsch arrived independently at the conclusion that it broke up into mesenchyme in the bird, and published his results at some length in 1893 (‘ Morph. Jahrbuch,’ Bd. xx) ; but Goronowitsch failed to recognise the reticulum, and he was ON THE CELLULAR THEORY OF DEVELOPMENT. 101 unable to appreciate the full significance of the facts he de- scribed in their bearing on the question of the origin of nerves. Platt approximated to the truth with regard to the third nerve in her account of it as growing from the ciliary ganglion to the brain, but retained the error of her predecessors in regarding it as a cellular object, and not as a differentiation of the reti- culum. Minot has a characteristic comment on Platt’s statement. He says (‘Human Embryology,’ p. 639), “This view rests probably on erroneous interpretation of observation, for it cannot be admitted that a motor nerve is formed by ganglionic fibres”! (The italics are mine, as is also the note of admiration.) = ) Sera) Wa) Cb till ae el ae eo ee a rise 2. Ge SRG ankle As or 6 Vor Wee BRR Bynes (ie Re |) *, jixhaw his eerie PP. el bea BR Te HAE od sa rip Jy a ey - € ' a a ; ‘ 4 j et i J i t ives a > ’ 0 ‘ a i € v i + = * _ Fy A - > > a a ® ON BENHAMIA C@OCIFERA, N. SP. 108 On Benhamia cecifera, n. sp., from the Gold Coast. By W. Blaxland Benham, D.S¢c.Lond., Hon. M.A.Oxon., Aldrichian Demonstrator in Comparative Anatomy in the University of Oxford ; Lecturer on Biology in Bedford College, London. With Plate 12. Some time this year I received a large specimen of Ben- hamia from Professor Jeffrey Bell for examination and identi- fication. My best thanks are due to the authorities of the British Museum for their permission to make the examination. The worm, which turns out to be a new species, to which I give the name Benhamia cecifera, was collected by Captain Torry at Axim, in the Fantee Country, on the West Coast of Africa. We already know a number of species from this side of Africa, as well as from the East Coast and inland, and we know pretty certainly that this continent 1s the home of the genus. This new species is of considerable size, measuring 510 mm. (20 inches), with a diameter of 17 mm. at the clitellum, and gradually diminishing posteriorly. Its average diameter is about 12 mm., and it tapers only very slightly to within a few segments of the hinder end, where it rounds off. There are 310 segments. The colour of the worm is dirty brown (in spirit), and the hinder end is not very sensibly lighter than the rest of the body. In each segment there is a circle of small dark brown pig- 104 W. BLAXLAND BENHAM. ment spots at the level of the chete, giving the impression at a casual glance that the chetz are in a circle round the body. A similar style of pigmentation has been noted by Michaelsen in the case of B. affinis and others. The dorsal pores are large and conspicuous, commencing after the 4th segment ; as usual these are absent (or obliterated) in the clitellar region. When the animal was handled a con- siderable quantity of a brown fluid issued from the pores. The chxtz, which appear as black dots, have the usual arrangement, and present no pecularity of shape or ornamen- tation. They are invisible, if indeed they are present, in the 2nd segment. As I had only one specimen, which had to be returned to the British Museum, I could not, of course, examine the body-wall microscopically. The four couples are equi- distant from one another, being about 2 mm. apart. Some of the anterior segments, except the first three, are biannu- lated, the first annulus being nearly twice as large as the second, and bearing the chete. The prostomium is not dovetailed into the peristomium ; the latter is not more than half the size of the 2nd segment. The clitellum covers eleven segments, viz. x111 to XXIII in- clusive ; the intersegmental groove is still distinct detween the 13th and 14th segments, but elsewhere, at any rate dorsally and laterally, the limits of the segments composing the cli- tellum are not recognisable, except by bands of darker pig- ment. As in all species of this genus, there is a ventral area in the middle of the clitellar region which in this instance presents specific characters. In Segments x11, xiv, and xv the clitellum is ‘‘ complete,” but in the remaining segments it ceases ventro-laterally, 1. e. just below the outer series of chetz, and forms a distinct margin (fig. 1, m.) limiting a more or less rec- tangular depression, the “ ventral field.”’ This is wider ante- riorly than posteriorly, being about 12 mm. across on Segment xvit and about 8 mm.on Segment xx11. Anteriorly the ventral field is bounded by the hinder margin of Segment xv. Within this “ventral field” lie the four prostate pores, a pair on each of the Segments xvii and xix, The two pores of ON BENHAMIA CM@CIFERA, N. SP. 105 each side are connected by a longitudinal groove, as in other species of the genus. But what marks this species as distinct from all others hitherto examined is the presence and arrange- ment of numerous small pits, no doubt having some function in relation to copulation. The figures give a better idea of the arrangement of the pits than any detailed description. It will be seen that for the most part they form transverse rows on Segments xv, XVI, XVII, XIX, Xx, xx1,xx11. There is a single median pit on Segment xxii, and on the 18th segment three short longitudinal rows occur, one median and a pair of lateral rows. “ Copulatory organs ” (“ pubertats Tuberkeln”’) have been noted by Michaelsen in various species, especially B. affinis and B.inermis. In the latter they are paired, and occur on segments in front of the clitellum, but they appear from his figures and descriptions rather as tubercles than as pits, and have altogether a different arrangement. These pits in the present worm mostly have a well-defined margin, not raised above the general level of the surface. In others there appears to be a papilla on one side of the pit, pro- jecting to a greater or less extent into the cavity of the pit, as shown in fig. 2, B. C. But in no case do the papille project to the exterior, nor are they visible except on careful inspection. Two pairs of similar pits occur in relation to the sperma- thecal pores, viz. on Segments vir and vi (fig. 1, p’. p’.). The anterior pair lies on the hinder margin of the segment, the pos- terior pair on the anterior margin of the second annulus—this one is more laterally placed than the former. In the opened worm there were no sacs or other structures projecting into the cavity of the worm’s body ; the pits appear to be limited to the body wall. The spermathecal pores have the usual position on the boundaries of Segments vii-vi11 and vii1-1x, in a line with the inner series of chetz. The oviducal pore is small, but it is visible just in front of the ventral (inner) chetz of Segment xiv. On Segment x11 a pair of circular whitish areas, each with a central depression, 106 W. BLAXLAND BENHAM. is situated just behind and slightly laterad of the inner chete. These probably have some copulatory significance. They are, however, different from the “‘ pits” above mentioned. Internally in this segment is a low but rather extensive muscular promi- nence occupying the whole length of the segment, which, I presume, is the wall of some structure opening by the above- mentioned pore. Turning now to the internal anatomy, there are the following points to be noted. As in other species of the genus, certain of the anterior septa are very much thickened, viz. the septa 9/10, 10/11, 11/12, 12/13, and 13/14; the next two septa are also thick, but less than the foregoing five. In front of the first of these thickened septa, the viscera are wrapped together by a delicate membrane, due no doubt to the thin septa of these segments being pushed backwards by the large gizzards ; owing to the tenuity of the septa here, and to the fact that they are ruptured in merely turning the organs aside, it is difficult in a dissection to trace them out. The first septum in the body is distinct, and separates Segments 1v and v; con- sequently septa 5/6, 6/7, 7/8, and 8/9 are represented e the enveloping membrane just referred to. The Alimentary Tract.—Amongst the internal organs the most noticeable feature of novelty is presented by the intestine. The two gizzards, characteristic of the genus, are of large size, and appear to belong to the Segments v and v1. The cesophagus is narrow, and provided with the usual three pairs of “ calciferous glands” lying in Segments xv, xvi, and xvi1. Whether carbonate of lime is present, as has been deter- mined for other species, I am not in a position to affirm, as I did not think it worth while to explore the glands. As is frequently the case, the first gland on either side is the smallest. In the genus Benhamia the wall of these glands presents a variety of patterns, owing to its folding ; these have not been figured carefully, but have been described by Michaelsen and by Horst as horizontal folds in some cases, vertical folds in others. But in the present species the foldings are very claborate, so that the surface of the gland, as well as ON BENHAMIA C@CIFERA, N. SP. 107 its shape, suggest a much convoluted cerebral hemisphere (see fig. 4). Each gland is, on the whole, reniform, and connected to the narrow cesophagus by a short, narrow, but distinct duct. The glands are supplied by a large vessel from the dorsal trunk, and are further closely connected to the septa. In the 20th segment the gut enlarges suddenly and forms a wide, thin-walled, dark brown coloured “ sacculated intestine,” constricted of course by the septa. In Segment xxix, however, the intestine takes on the peculiar character (unknown in its details in any other earthworm) which is referred to in the specific name “ cecifera.” In each segment, commencing at the 29th, the intestine gives rise to a finger- shaped diverticulum on each side, which is directed upwards and arches over the dorsal blood trunk (fig. 3, c.). These ceeca extend as far as the 52nd segment, the last half- dozen gradually diminishing in size, till the intestine resumes the ordinary sacculated condition. Each such cecum arises by a comparatively broad base from the upper surface of the side of the main gut, and gradually narrows as it curves upwards, to end in a blunt rounded apex. The apices of the pair of czeca of any segment overlap each other above the dorsal vessel. These ceca are thin-walled, and quite unlike the well- known ceca which lie in the 26th segment of the ordinary Pericheta. The generative organs conform to the usual Benhamia type. There are two pairs of sperm-sacs, in segments x1 and x11, each being much divided and having a lobulated appearance. The two pairs of prostates are very long, narrow, yellowish tubes of the usual acanthodriloid pattern; the muscular duct is very delicate at its origin from the gland, enlarging in its slightly winding course, to pierce the body wall: they lie in segments xvii and x1x (fig. 5). I could detect no penial chet; these special chete have been noted as lacking in other species of Benhamia. It is a point on which I laid some stress in my note on “ The Genera 108 W. BLAXLAND BENHAM. Trigaster and Benhamia,”’! where I pointed out the various differences between the two genera. Some of these differences have been bridged over by various species of Benhamia, though in my opinion the genus Trigaster still stands as a valid genus. Though I could find no penial chetz on careful examination of the partially dissected individual, I should not like to affirm that they do not exist. There are several strong bands of muscle in the Segments xvii to xxi (fig. 5), arising close to the nerve-cord and passing outwards to be attached to the body-wall near the dorsal midline; they have a similar ar- rangement to those figured by me for Moniligaster (indicus) robusta, A. G. B.,? and it is possible that special cheetz may lie below these muscles—but I did not wish to do too much injury to the specimen. The spermathece, which lie in Segments vir and viii, differ in size, as has already been noted in several species. The anterior pair are smaller than the posterior pair, and whereas the latter is provided with a small ovoid diverticulum, those of the 7th segment are without one (fig. 6). Each spermatheca consists of a thin-walled, somewhat ovoid sac, with a longish and comparatively stout, muscular, glistening duct, which widens as it approaches the body-wall. I found no structures in these segments corresponding with the pits noticed in the account of the external anatomy. The vascular system calls for a few remarks, but I have not attempted to make a complete study of this system ; never- theless, from what I have observed and from what previous observers have noted, I believe the vascular system of Ben- hamia and Acanthodrilus would repay a careful study on the lines of and in relation to Dr. Bourne’s thorough account of the vascular system in Megascolex® and Moniligaster.* 1 «Ann. Mag. Nat. Hist.,’ 1890, p. 414. ? Quart. Journ. Mier. Sci.,’ vol. xxxiv, pl. 32, fig. 3. $ «On Megascolex ceruleus, &.,” ‘Quart. Journ. Mier. Sci.,’ vol. Xxxil, p. 49. 4 “Moniligaster grandis,” ‘Quart. Journ. Mier. Sci.,’ vol. xxxvi. ON BENHAMIA C@OCIFERA, N. SP. 109 In Benhamia cecifera the dorsal vessel gives origin to four pairs of very large hearts, lying in Segments 1x, x, XI, XII. These may be, and I suspect are, connected also with a supra- intestinal trunk, as Horst has described in certain species of this genus; but I could not, without injury to the worm, determine the point. In Segments vir and vir two other pairs of hearts arise ; these are smaller, and differ in their appearance from the posterior pair, being more irregular in their swellings. The large posterior hearts remain greatly distended till close to the ventral trunk, to which they are joined by a short narrow vessel (fig. 8). The heart in Segment vii has a less diameter and a longer, narrow vessel connecting it with the ventral trunk. The heart in Segment vir is much shorter. About midway between the dorsal and ventral trunks it suddenly narrows ; this narrow part opens into a small spherical dilata- tion, whence arise two delicate vessels ; one goes forwards to the side of the second gizzard, the other continues downwards to enter the ventral vessel. This anterior heart recalls most curiously the condition of the heart in Segment rx of Megas- colex figured by A. G. Bourne (loc. cit. pl. 8, fig. 4). The dorsal vessel terminates here in Segment vir after giving rise to the hearts (figs. 7, 8), but its place appears to be taken by a pair of “latero-longitudinal ” vessels (I use this term topographically, and with no implication as to their homology with those so named by Bourne in Moniligaster). These (J. 7. v.) arise, as far as I could determine, as branches of the dorsal trunk just in front of the third pair of hearts, i.e. in Segment 1x, and each runs forwards to the pharynx, giving off vessels to the gizzard and other structures. The blood-vessels on the body-wall present a feature which I have not seen noticed in any previous account. In each seg- ment there is a small vessel having a circular direction along the inner face of the body-wall. I do not know its relation to the main trunks, but, as it is readily recognisable with the naked eye: on each side of its course, it gives rise to a fairly regular series of small globular dilatations ; these were more 110 W. BLAXLAND BENHAM. readily noticeable along the dorsal wall, but could he traced all round (fig. 9). Viewed under the microscope it is seen that each dilatation is connected by a short narrow vessel to the circular vessel; and from the dilatation a vessel passes away, which soon dips into the muscular coat of the body wall and then subdivides. I have not traced them further ; but the regular arrangement and large size of these dilatations, which recall those previously described and figured in the case of Lumbricus, Trigaster, &c., is deserving of mention. The nephridial system consists of a series of “ micro- nephridia,”! arranged in a row around the inner surface of the body-wall, near the hinder septum of most (probably all) of the segments. This nephridial system is specially abundant in Segments x11, xiv, and xv, and most markedly towards the ventral surface of these segments. SUMMARY. Benhamia cecifera is, then, characterised (1) by the number and arrangement of the peculiar copulatory pits on the ventral field; (2) by the possession of a number of peculiar finger-shaped ceca arising from the intestine; (3) by the form of the foldings of the wall of the calciferous glands, and possibly (4) by the termination of the dorsal vessel in Segment vil; (5) by theextent of the clitellum ; (6) by the position of the first dorsal pore between Segments rv and v. AFFINITIES. In its size (length 510 mm.) and number (310) of segments it might be grouped with certain other large species of the genus found on the West Coast of Africa, viz. B. Schlegelii, B. Buttikoferi, B. rosea, and B.inermis. But it differs from each of these in the above characters. Firstly, it is marked off by the forward position of the first dorsal pore ; this structure has been found of specific value in our European 1 This term, employed by Vejdovsky, is preferable to my term of “ plecto- nephridia,” for we have no evidence that in every case there is a true network any more than we have a real network in all “ capillary networks.” ON BENHAMIA C@OCIFERA, N. SP. Tit earthworms, Allolobophora and Lumbricus, and is probably of similar value here. In extent of the clitellum (Segments x11t—xxt11 inclusive) it recalls B. rosea, Mich., in which it ceases on the 22nd segment. But in this species the paired spermathecal and oviducal pores are connected by transverse grooves; further, penial chete are present, and the shape of the ‘ ventral field ” is quite different. Again, no mention is made of copulatory pits or tubercles. To B. inermis, Mich., our present species presents a greater likeness, in that there are no penial chzetz but there are copulatory tubercles (‘‘ pubertats-tuberkeln ”’) ; these are, how- ever, arranged on quite a different plan. There appear to be about twelve or more pairs of them; the first pair on Segment vu, then follow seven pairs situated at the intersegmental grooves of as many segments, all in front of the clitellum ; there are four more pairs within the clitellar region. Each pit leads into a sac projecting into the body cavity. This associa- tion of the presence of copulatory papillz and pits with the absence of penial cheete is worthy of note. Another large species, B. itoliensis, is known only from its anterior end. This species was found at Itoli, on the Victoria Nyanza, and though in general it appears to resemble B. cecifera, yet the possession of penial chetz and absence of copulatory pits (as well as other features) serve to distinguish the two from one another. 112 W. BLAXLAND BENHAM. EXPLANATION OF PLATE 12, Illustrating W. Blaxland Benham’s paper, “On Benhamia ceecifera, n. sp., from the Gold Coast. Fic. 1.—Ventral view of genital segments of Benhamiacecifera. (Nat. size.) The segments are numbered. All the genital pores are shown. spth. Spermathecal pores. pt. p?. Copulatory pits in their neighbourhood. cop. Copulatory (?) organs of a different nature on x1v. m. Latero-ventral margin of clitellum bounding the “‘ ventral field.” The arrangement of the copulatory pits, represented as small dots, is shown. Fie. 2.—Three copulatory pits from the “ ventral field.’ In B. and Cia papilla (p.) is more or less exposed in the pit; in 4. no papilla is to be seen. Fic. 38.—A portion of the intestine exhibiting the characteristic finger- shaped ceca (c.), arching upwards over the dorsal blood-vessel (D. v.) ; in two cases the blood-vessels (v.) to this ceca are shown. Fic. 4.—The three pairs of calciferous glands (g/.), with their vessels from the dorsal trunk (d. v.). Fig. 5.—The prostate duct (d.), with copulatory muscles (m.), which are seen passing into the next segment. pr. A portion of prostate. .c. Ventral nerve-cord. spt. Septum. Fic. 6.—The two spermathece of the right side. d. Muscular duct. div. Diverticulum. Fie. 7.—The three most anterior pairs of hearts, lying in Segments vir, vill, and 1x. The dorsal vessel (D. v.) is seen ‘o terminate in Segment vit. spt. Septum 9/10. Fic. 8.—The same three hearts seen partially from the side, showing the relations ventrally. 7. 7. . Latero-longitudinal vessel arising in Segment 1x. g. Its branch to gizzard. g’/. Branch from first heart to the gizzard. s. Swelling from which this.arises. V.v. Ventral vessel. Fic. 9.—A small portion of the body-wall of a segment, to show the dilata- tions on the circular vessel (c.), as well as the position of the micronephridia (z.). S'., S*. The anterior and posterior septa. EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 1138 A Re-investigation into the Early Stages of the Development of the Rabbit. By Richard Assheton, M.A. With Plates 13—17. INTRODUCTION. Ir is now nearly fifteen years since Ed. van Beneden' published his account of the development of the early stages of the rabbit ; and although I do not see that one can as yet describe every detail of even the earliest embryology of the rabbit strictly epigenetically, still it seems probable that van Beneden took far too little heed of the extrinsic causes which may direct the course of development. The effect upon the development of the presence or absence of such structures as the albuminous layer or zona radiata has been almost entirely ignored, apparently because they are matter outside the ovum. So, again, the size and structure of the uterus have hardly received their proper share of attention. The present paper and my other “ On the Causes which lead to the Attachment of the Mammalian Embryo to the Walls of the Uterus,” tend, I hope, to show how many of the details of the earliest stages may be ascribed to the direct maternal influences. That is to say, the inherited force is an energy 1¢Ta maturation de l’ceuf, &c., Bruxelles, 1875; ‘La formation des Feuillets chez le lapin,’” ‘ Arch. de Biologie,’ vol. i, 1880. VOL. 87, PART 2,—NEW SER. H 114 RICHARD ASSHETON. which would of itself produce not a specific embryo, but an amorphous monster unless directed by the influence of (in the rabbit) inanimate coats and the walls of the uterus. I have come to the following conclusions upon certain dis- puted points. (i) Van Beneden’s description of the segmentation I consider to be inaccurate. (ii) I find no trace of van Beneden’s blastopore. (iii) I find no trace of a “ gastrulation.” (iv) There is no evidence in support of Robinson’s! specula- tions concerning the existence of a hypoblastic wall to the blastocyst surrounded subsequently by the epiblast. (v) Rauber’s layer fuses with the inner layer of epiblast as described by Balfour? and Heape.® (vi) This fusion has but slight morphological significance, since its existence and disappearance are caused mechanically by ontogenetic conditions. (vii) The growth round of the hypoblast is apparent only, being due to the presence of a zone of specially active epiblast surrounding the embryonic disc, which zone is to be considered to be the equivalent of the trager in other rodents. I have endeavoured to show how important is the albuminous layer, and how I believe that it is possible to account for many details of change up to the end of the sixth day by strictly epigenetic processes; and since these processes during this time are almost all directed by environment as between the embryo and the maternal influences rather than between cell and cell of the embryo itself, it follows that the palin- genetic features of the development must be reduced to a minimum. This paper is based upon the examination of upwards of 300 embryos between the 24th and 168th hours. The embryos have been examined fresh, and after treatment with various reagents —Perenyi, osmicacid, picric acid, Flemming, Hermann, chromic 1 Quart. Journ. Mier. Sci.,’ vol. xxxiii. 2 «Comparative Embryology,’ vol. ii. 3 «Quart. Journ. Mier. Sci.,’ vols. xxiii and xxvi. EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 115 acid, &c., and sections have been examined of all stages from the 30th hour onwards. A large portion of the work for the present paper, and the three or four other papers I hope to publish shortly, was carried on while I held the post of Demonstrator of Zoology in the Owens College. To the kind consideration of the late Professor Arthur Milnes Marshall Iam indebted for many opportunities for work which I should otherwise have found difficult to obtain. The remainder of the work has been done in the Morpho- logical Laboratories of Cambridge through the permission of Mr. Sedgwick, to whom I wish to express my gratitude for the said permission and for his kindness in reading my papers and offering many valuable criticisms. CHAPTER I. SEGMENTATION OF THE Ovum. Since the accuracy of van Beneden’s account of the early stages of the development of the rabbit depends to a great extent upon the way in which the earliest cleavage planes succeed one another, I think it advisable to give in some detail the results of my own researches on this point. In fact the very first segmentation division, according to van Beneden, gives rise to an important question, which is as follows: is there any essential difference between the two first segments, and do the descendants of one segment give rise to the epiblast cells, and the descendants of the other to the hypoblast cells ? To this van Beneden replied that there is always a difference in size between the two first segments, and that the smaller of the two is the more granular, and that from that one are derived all the cells of the inner mass, and that the larger clearer segment gives rise to cells which gradually surround the descendants of the small segment and form the wall of the future “ blastodermic vesicle.” From the descendants of the small segment van Beneden 116 RICHARD ASSHETON. derived the hypoblast and subsequently the mesoblast ; from the descendants of the larger, the epiblast. As far as the conversion of the whole of the inner mass into hypoblast and mesoblast is concerned, van Beneden has changed - his opinion, admitting that Kolliker’s (and others—Balfour, Heape) careful research in this matter is a truer account than his own. Again, Heape, on the mole, has shown that a different inter- pretation may be placed upon the fact as seen in the rabbit, although he finds that there is a difference in the size of the two first segments. The results of my researches show that there is, if not always at any rate very frequently, a difference in size between - the first two segments ; sometimes the difference is very marked, but more usually it is to be found only by careful measurement. In one rabbit I examined the following results were obtained. Rabbit 48, was killed 244 hours after coitus. In the thin- walled part of the Fallopian tube of the right side I found four embryos at about 30 mm., 40 mm., 42 mm., 50 mm., from the funnel mouth of the Fallopian tube respectively. No. 1, the highest up, had not as yet divided. The ovum was spherical with a distinct nucleus, the ovum not occupying the whole space within the zona radiata. No. 2, the second on the right side, was divided into two segments, but this I did not measure. No. 3, the third and lowest on the right side, was also in two segments. I could detect no differences at all in colour or texture, but on measuring the longest and shortest diameters of each the following results were obtained. One of the segments measured as follows: longest diameter ‘1 mm., shortest diameter (076 mm. The other segment: longest diameter ‘096 mm., and shortest ‘068 mm. The specimens were examined and measured in the fresh condition in a drop of aqueous humour of the rabbit. The measurements were taken with a micrometer eye-piece, No. 8, and Zeiss D objective, which gave a magnification of 54, mm. for each division in the eye-piece. EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 117 In giving the measurements I have reduced these to actual measurements, giving them in decimal fractions of one milli- métre. No. 4 was also in two segments, but was not measured ; each segment was composed of an inner denser and outer clearer part. In the Fallopian tube of the left side I found six ova, which were scattered along an area between two points 35 mm. and 65 mm. from the upper end of the Fallopian tube. The whole length of the Fallopian tube was about 100 mm. Three of the ova were in the wide thin-walled upper part, and the other three were just within the lower thicker-walled portion of the Fallopian tube. No. 1, that is to say the one nearest to the funnel end, was unsegmented. It appeared, how- ever, to be on the point of dividing (fig. 1). It did not occupy the whole of the cavity within the zona radiata. It was spherical, and two polar bodies lay between the ovum and the zona, the one larger than the other. The nucleus, however, though not very distinct, undoubtedly seemed to be double or dumb-bell shaped. The diameter of the ovum was ‘11 mm. No. 2 (v. fig. 3) was in two segments. The difference in size between these two was very marked. It was quite obvious without measurement. This specimen showed the greatest difference I have hitherto met with. There was no perceptible difference in density or in colour, either while fresh or after treatment with nitrate of silver followed by picro-carmine and aniline blue-black. The measurements of the two segments of this specimen were as follows : Of the larger the longest diameter was ‘101 mm., the shortest diameter ‘(088 mm. Of the smaller the longest dia- meter was ‘(08 mm., and shortest was ‘064 mm. The nuclei were clear and round. No. 3 was in two segments, but far more equal in size. Nuclei round and clear and distinct. Measurements were— The larger segment . : aod) ome. < 078: mm; The smaller segment. . : SUG sos 6 “OSA 5, 118 RICHARD ASSHETON. I could detect no difference except as regards size. No. 4. I could make nothing of this one; it may have been an unfertilised ovum or pathological. No. 5 was in two segments. Here again the only difference I could detect was in size, as follows: Larger segment . : : . ‘086 mm. Xx ‘072 mm. Smaller segment : : =; “0840-— 5. << 2066 2 No. 6, the lowest down, that is to say the nearest the uterus on the left side, was still unsegmented, being spherical. Both polar bodies were visible between the ovum and the zona pellu- cida. The ova from another rabbit, No. 51, were examined fresh in aqueous humour of the rabbit. In the left Fallopian tube I found only one ovum, situated almost midway between the two ends. This specimen was not yet segmented, but the ovum had retracted from the zona pellucida, which seems to be a sign that it was a fertilised ovum. In the right oviduct I found four ova. No. 1, the one nearest the funnel opening of the Fallopian tube, was in two segments. I could determine no difference in character, but on measurement a slight difference in size was made evident. Larger segment : ; . 0924 mm. x ‘070 mm. Smaller segment, ; : s», “0980:. 55) xX Db4nts After being carefully examined and measured while fresh, this specimen was transferred to a 2 per cent. solution of osmic acid for thirty seconds, and subsequently stained with a mixture of aniline blue-black and picro-carmine, and mounted in glycerine jelly—a very pretty preparation. At no period could I detect any difference except in size between the two segments. In this specimen the two polar bodies were placed far apart from each other, a peculiarity I have not noticed in any other (v. fig. 4). No. 2 was also in two segments, and was examined, measured, EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 119 and drawn in aqueous humour. The result of measurement was— Larger segment . : : . ‘099 mm. x ‘068 mm. Smaller segment . F : 5 UO SK BOLO No.3 was also in two segments, and was examined, measured, and drawn in aqueous humour. Result of measurement : Larger segment . : : . ‘LOO mm. x ‘079 mm. Smaller segment. F ; i, OS Geb ne CORE OS No. 4, the nearest specimen to the uterus, was as yet unseg- mented, possibly an immature or unfertilised one, as the ovum completely filled the space within the zona pellucida. I could not find any trace of polar bodies. All three specimens from this rabbit, which were in two seg- ments, I examined most carefully, to find, if possible, evidence of the marked difference in appearance described by van Beneden. I treated all with 2 per cent. osmic acid and used weak staining solutions of picro-carmine and aniline blue- black, of picro-carmine, of Beale’s carmine and _ aniline blue-black respectively, but to no purpose. They were then mounted in glycerine jelly, and now, after the lapse of a year, I still fail to find any difference between the two segments. From another rabbit I obtained two specimens in the two- segment stage, but was unfortunately only able to make a very cursory examination. One of these I thought I distinguished as having a larger darker segment and a smaller clearer one, but I could only examine it while lying on the folds of the Fallopian tube, the unevenness of which might easily have caused one segment to appear darker. So I cannot advance even this instance in support of there being a marked difference in appearance between the two segments. Another rabbit, with ova aged 254 hours, gave me four, of which two were in the two-segment stage, and two showed four segments. Here again there was no perceptible difference. I did not measure these while fresh. One I have since measured, 120 RICHARD ASSHETON. and find that the longitudinal and transverse axes of the one measure exactly the same as those of the other. A tabular view of the measurements of these two first cleavage segments may be of interest, showing how the size of one specimen may vary with that of another, and the variation in size of the two segments of one and the same specimen. Larger Segment. Smaller Segment. Longest Shortest Longest Shortest Diameter. Diameter. Diameter. Diameter. Rabbit 48— Specimen No. 1 100 mm. X ‘076 mm.| *096 mm. x ‘068 mm. ‘ se TOM a x O88 y; “080° ,, «x02 |, Me. ey Ati 100 ;, x 078 ,, | 098 ,, x ‘074 ,, ah heel ‘086 5. x O72 . | "Usa. © sone Rabbit 51— Specimen No. 1 0924.4 XScO7Ou. 092025. Sex 0G aR ak 099 ,, x 068 ,, | 090 ,, x -060 ,, ee 100 5, © "079 4, | “098 4, x 07s Each segment of each specimen when examined immediately after the death of the animal showed a denser, more granular inner portion, and a clearer, almost hyaline outer layer, the nucleus being situated in the denser inner portion (v. fig. 2). This difference is more evident in the early stages of segmen- tation, up to the time that there are twelve to sixteen seg- ments, than later. Stage with Four Segments. The second plane of cleavage seems to be at right angles to that of the first. It appears that the two segments divide about, if not exactly, at the same time. ‘This occurs about twenty-five or twenty-six hours after coition. The time between the segmentation of one into two and between two into four appears to be nearly two hours. Since there may be some difference in size between the two primary segments, it follows that there is also very frequently a difference in size between - EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 121 the cells of the four-segment stage, and, as one would expect, it may appear that two segments are small and two larger, as in fig. 6. Development between the 26th and 72nd Hours. Stages of Segmentation between Four-Segment Stage and Commencement of Cavity of the Blas- todermic Vesicle. The further cleavage of the four segments does not occur in each segment at the same moment. A stage of five segments or even seven may often be found. In a rabbit killed 27+ hours after coitus (Rabbit 36) I found in the right Fallopian tube two ova in the usual locality for this age, which is about 40 mm. above the uterus; one of them was in five segments, the other in seven. The five- segment one is shown in fig. 7. In this specimen it will be seen that two (Z2.) spheres are almost exactly the same size, and that these are considerably larger than two (S*.) of the remaining three, and slightly larger than the third (S*.). This specimen I surmise may have been one in which the two primary segments were unequal, and that each of these divided into the approximately equal spheres, and that at the moment of examination one of the daughter cells of the primary smaller (?) one had divided again into two very nearly equal spheres (S*.). It should be noted that in van Beneden’s account of the process the larger primary segment gives rise to the more rapidly dividing daughter cells forming his so-called epiblast ; while in this case it seems to be the descendant spheres of the smaller primary segment which appear to be the more ready to undergo division. Fig. 14 is a camera lucida drawing of the other embryo of the same Fallopian tube, seen as a transparent object. In this specimen one segment (L?.) is larger than the others. The four marked S*. were approximately equal, and slightly larger than the two (L*.). May this be interpreted as follows ? The ovum which gave rise to this embryo divided into two 122 RICHARD ASSHETON. slightly unequal spheres. The smaller of these two divided into two, each of which has divided into two, the result being four approximately equal spheres (S°., S®°., S°., S*.). The larger primary spheres divided into two, one of which is still undivided (L?.), the other having divided into two (marked Z?.). Here again there seems to be a tendency for the descendant spheres of the smaller primary to undergo division first. In the Fallopian tube of the left side of the same rabbit I found six embryos, of which some were in seven segments, others ineight. I had not time, unfortunately, to measure all these or draw them carefully. The embryos of this rabbit seem to be unusually far advanced for their age. The stage with eight segments is a very common one to find between the 29th and 44th hours. This may be accounted for by there being a rather long resting stage after the production of the eight segments. One rabbit (Rabbit No. 19) presented a very curious condition of its Fallopian tube, a condition, I believe, that has been noticed before, but I cannot remember by whom. The rabbit was a very large, healthy English doe. Both Fallopian tubes were almost filled with ova. Onexamining the Fallopian tubes with a lens before opening I imagined I had come across a marvellous find of embryos—in all about fourteen “ ova” were to be seen shining through the wall of the left Fallopian tube. Of these only three turned out to be embryos, and were mingled with at least twelve apparently disintegrating un- fertilised ova. The right Fallopian tube likewise contained a number of disintegrating ova, as well as four perfectly normal embryos. Some of these unfertile ova were more thickly coated with albumen than is usual for normal eggs, possibly due to their having been a long time in the Fallopian tube. It is a matter of curiosity why these unfertilised or pathological ova had not passed down the Fallopian tube, but had allowed the fertile ova to pass them, as one at least had succeeded in passing the whole twelve bad ova. One can hardly believe that all these ova had left the ovary at the same time and together with EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 123 the ova which became fertilised and developed into normal embryos. The corpora lutea are often very difficult to make out at this early stage, but I am pretty certain that there were cer- tainly not more than five corpora lutea in one ovary and six in the other. Fig. 9 is drawn from one of the fertile ova of this rabbit. I have reproduced it here to show the appearance of the denser inner and clearer outer layer of protoplasm of each sphere, de- scribed above as occurring in the earlier stages. Possibly this appearance is constant in later stages, but as the cells get smaller the difference becomes less easy to distin- guish. I have avoided the use of the terms macrosphere and micro- sphere, because a morphological significance is sometimes given to these terms which I think is not advisable, at any rate in the case of the rabbit. At the same time it is necessary, in discussing van Beneden’s description and deduction, to pay especial attention to the question of the fate of the segments derived from the two first spheres. It is hardly necessary to point out the very great difficulty in determining the fate of the descendant segments of the two primary segments when size is the only character on which we can rely, and when even in size there may be no difference. Again, if the ovum may sometimes divide into two segments of different size, may not the two primary segments also segment unevenly? I can see no way of determining the question except by watching the division of one specimen, and this is, as far as I have tried, impossible. Even if it were accomplished, it must be under conditions which can hardly be called normal, and therefore not a very safe ground on which to base either facts or theory. Certainly, in cases where the primary division is so unequal as in fig. 8, the subsequently formed segments could, on the whole, probably be classed as larger ones derived from the 124 RICHARD ASSHETON. large primary segment (fig. 3, Z.), and smaller ones derived from the small primary segment (v. fig. 3, S.). If, then, we rely upon the only character at present possible, that is to say, upon the size of the segment, we are bound to conclude that not only does the process of segmentation proceed with great irregularity, but that also there is no evidence of the descend- ants of the larger primary segment ever forming a cap and subsequently enveloping the descendants of the smaller primary segment. I believe, rather, that descendant segments of the larger primary segment become intermingled quite irregularly with the descendants of the smaller primary segment, for this un- doubtedly affords a better explanation of the appearance of embryos like that shown in fig. 12 or fig. 15. Fig. 11 is an outline drawing of a specimen taken from a rabbit which was killed at the completion of the 39th hour. In this animal seven of the eight embryos found were in the eight-segment stage. It is extremely difficult to mea- sure the segments with sufficient accuracy to be of any service, and although I measured them I shall not give the results. In this case, as shown by the fig. 11, there was very little dif- ference in size, and none, so far as I could judge, in texture. There was one curious feature which is worth mentioning, but I do not attach any importance to it. The larger polar body was visible between the embryo and the zona radiata. The smaller of the two was inside, or rather mingled with the seg- ments of the embryo. It is quite possible that this frequently occurs, for the polar bodies seem very often to disappear en- tirely. It would be of interest to determine whether, when this is the case, the polar bodies, ever under the altered con- ditions, acquire a renewed activity and give rise to segments whose descendants become part of the embryo. I have as yet no evidence as to which sphere commences the next series of cell division. At about the 47th hour the embryo has the typical morula form, and is made up of a number of segments (sixteen to twenty), which very frequently present great diversity in size. EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 125 Nor is there any regularity, as far as I have observed, as to the location of the large and small spheres. Figs. 12, 15, and 17 show this very well. Fig. 12 was drawn with camera while the specimen was still fresh in a drop of aqueous humour of the rabbit. After the drawing had been made the specimen was placed in Perenyi’s fluid and subsequently freed from the zona pellu- cida with fine needles. The segments were then separated one from another. In all there were seventeen segments, and of very different sizes. Fig. 18 shows three of the segments thus separated. Except in size I could detect no difference. Another specimen from the same Fallopian tube was placed in 4 per cent. solution of silver nitrate for two minutes, and after having been washed in water and exposed to the sun for a few hours was embedded in paraffin and cut. Fig. 20 is from a section through about the centre of this specimen. The nuclei of the individual segments are not at all distinct, excepting where they have been cut almost through their centres, as no other stain except silver nitrate has been used. (Of all methods of fixing the early segmenting stages, I believe none answer so well, as far as concerns the preservation of the correct shape of the spheres, as a weak solution of silver nitrate allowed to act for not more than two minutes. Next to silver nitrate I believe osmic acid, 2 per cent., is best.) Fig. 15, which is a specimen from a rabbit killed sixty-six hours and a half after coition, exhibits in a very remarkable way the great difference in size which may sometimes occur between the several segments. From the forty-fifth to the seventieth hour the segmentation proceeds slowly, and, I am inclined to think, sometimes very irregularly, as shown by the last-mentioned specimen (fig. 15). In sections of these stages I do not notice anything particu- larly remarkable, except that I have completely failed to find any constant character whereby the inner cells can be distin- guished from the outer. Fig. 21 is of a section through the 126 RICHARD ASSHETON. centre of a rabbit embryo aged seventy-six hours and a half. This was preserved in Perenyi’s fluid, stained with borax carmine, cut and mounted in series. The one I have drawn is the fourth of eight which pass through the actual embryo itself. The embryo at this stage is composed of a number of cells or segments, as far as I can see all similar in character, though varying a good deal in size. The cells in the centre are no doubt pressed closely together in the living state, the several clefts, with the possible exception of one, being artificial. This just-mentioned excep- tion, the more regular and continuous slit marked C. BL., isin all probability the first commencement of the slit which ulti- mately enlarges into the cavity of the blastodermic vesicle. In only one specimen have I found certain cells to take the stain better than others. I have drawn two figures (18 and 19) from the series of sections in which this occurs, in order to show how such an appearance as that described and figured by van Beneden (pl. iv of his paper) may arise, and at the same time to show how the interpretation put on it by him cannot be held to be sound. In this specimen the embryo has contracted very much, and is lying quite free from the zona pellucida and albuminous layer. In fig. 18 the majority of the cells at the surface are seen to be slightly darker than the mass inside, with the exception of two at the point 2. This is undoubtedly hke van Beneden’s figure of an optical section (pl. iv, fig. 1). But if we look at another section, fig. 19, we find here that again the cells of the surface layer have mostly stained darker, with the exception of one at 2’. Hence we are obliged to believe that in this specimen there existed at least two of van Beneden’s blastopores, for the light-coloured cells which show at the surface in fig. 18 are at almost the opposite pole to that at which the lighter-coloured cells of fig. 19 show at the surface. In sections of other specimens of about this age (seventy- two hours) fixed with osmic acid 2 per cent., I have failed also EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 127 to find differences of any value between the inner and outer cells. The same may be said of sections of specimens treated with silver nitrate. Fig. 16 is a drawing through the centre of a specimen from a rabbit seventy-two hours after coition, fixed with nitrate of silver + per cent., exposed to the light and stained with picro- carmine. Summary up to the Seventy-second Hour. The original description of the segmentation stages by Bischoff I believe to be in the main correct. I cannot find any evidence to support van Beneden’s view of the origin of the inner mass of cells from a smaller, more darkly staining primary segment only ; or for the origin of the outer layer of cells from a larger, more lightly staining primary segment only ; or, again, for the growth of the descendants of one of the two primary segments round the descendants of the other primary segment. It must be borne in mind that the very fact of van Beneden’s description of the later stages of development (the origin of the mesoblast and hypoblast from the inner mass, and that of the epiblast entirely from the outer layer) having been shown to be wrong by several authors (Kolliker,' Heape, Balfour, &c.), made it almost certain that van Beneden’s description of the segmentation stages was incorrect, or, at any rate, that the interpretation he put upon the supposed facts could no longer be held to be sound. My account is briefly as follows : 1. The ovum about the twenty-fourth hour after coition divides into two segments, one of which is usually larger than the other, there being much variation in this respect. 2. Each of these segments again divides about the twenty- sixth hour after coition, each dividing very nearly at the same time. These four segments now resulting may vary in size. 3. The third series of divisions takes place about the twenty- 1 A, Kolliker, ‘ Festschrift zur Feier des 300 Jahrigen bestehens der Julius- Maximilians- Universitat zu Wiirzburg,’ 1882. 128 RICHARD ASSHETON. eighth hour after coition. There is now less unanimity of action in point of time of the division of the several cells, so that it is fairly common to find embryos with five or seven segments. Here, again, four cells may be distinctly smaller than the remaining four, or all may be almost exactly equal in size. There is no difference to be detected in the character of the contents of the spheres. 4. The segments continue to divide with less and less regu- larity, so that the descendants of one primary segment become mingled with those of the other, there being much difference in size between the several segments. 5. The result of this continued activity is the formation of a solid “morula” of cells of varying size but similar character, and it is impossible to apply the term epiblast or the term hypoblast to any part of the embryo as yet. The ovum when it leaves the ovary is enclosed within a sheath or protective investment, the zona radiata. As the ovum proceeds down the oviduct this protective investment is further strengthened by the deposition on its outer surface of a thick coat of some albuminous substance which is secreted by certain cells of the epithelial lining of the oviduct, and which forms a tough strong membrane which has an important influence on the future mode of development. CHAPTER II. Tur Formation or THE BLAsToDERMIC VESICLE. The Fourth Day (73rd to 96th Hours). The most noticeable feature of this day’s events is the commencement of a cavity within the morula, which cavity enlarges enormously. What the object of this cavity is we can pretty safely infer, as also we can pretty safely conclude that the causes which bring it about originated in the organism in connection with the diminishing size of the ovum of the distant ancestral animal; but how this cavity is actually produced in the EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 129 embryo at the present day is a question of very great dif- ficulty. It can hardly be looked upon as a cavity comparable to the segmentation cavity of Amphioxus or Amphibia, or the Lam- prey or Ganoids, for it has a fate different from that in any of those animals. The fate of the segmentation cavity of the above-mentioned animals is to disappear and take no part in the formation of the cavities of the adult. The cavity of the blastodermic vesicle whose formation we are about to discuss never disappears, never, at any rate, in the rabbit, as part of it remains as the cavity of the alimen- tary canal of the adult. Part of it undoubtedly is comparable to the archenteron; possibly all of it is, for it becomes the gut-cavity of the adult. The first cause which produces the cleft that subsequently enlarges into the cavity of the blastodermic vesicle may be a more active growth of the outer layer of cells. Undoubtedly there is after this time a more active growth of the cells of the outer layer. They increase much more rapidly than the cells of the inner mass. . It is not easy to explain why the energy which up to a certain time causes a solid morula should do so no longer. Why should not the morula steadily increase in size, but be stilla morula? Although I believe that the subsequent vast increase in size of the blastodermic vesicle cavity is due to the diffusion inwards of fluid derived from the uterus, still this can hardly be the cause of the first starting of a cleft as in fig. 21. This seems to be best explained by the assumption of an in- crease in rate of growth of the outer cells over the inner cells of the morula. That such an increase does exist the following table provides evidence. In a median section through an embryo: Hour. Number of cells cut through in outer layer. Inner layer. 76th . : : : Loe, : . ; 27 82nd . : . : 24. : : : 24 Ser OMe sweets: be i Bi; : - teh 100th . : ° . 4A 4 : : 28 VOL. 37, PART 2,——NEW SER. I 130 RICHARD ASSHETON. This, though rather a rough method of investigation, shows sufficiently well that there is a proportionally greater rapidity of increase of the cells on the outside over that of the inner cells. Growth of the embryo must surely be dependent upon nourishment from without, when the bulk of the mass increases as it does from the 70th hour. It is the cells upon the outside of the morula which are in the most favourable position for the acquisition of such nourishment from the fluids of the uterus or Fallopian tube. No doubt during the early stages of seg- mentation the energy of division is derived from the nutriment —yolk, &e.—contained within the ovum itself. As this be- comes used up the embryo will become more dependent upon external sources. This will mean that the externally placed segments will become more favourably placed for growth than the internally placed segments. May not this gradual ex- haustion of intrinsic nutriment be a determining factor in the cessation of the increase of the embryo as a morula and causa- tion of the first commencement of a cavity ? This may give rise to the first cleft, but in itself it can hardly account for the large increase in size of the blasto- dermic vesicle. The cells are so very delicate it is hardly con- ceivable that they could cause the great expansion of the very tough albuminous wall. It seems far more likely that the force which causes the expansion is due to an osmotic current being more rapid inwards than outwards, either simple or more probably assisted by the vital activity of certain cells of the embryo, as is supposed in the case of diffusion in intestine, and suggested by Heape in connection with the mole. Or the diffusion process may be a simple physical process, but the nature of the liquid after entering the cavity may be so changed by the activity of the cells as to render its diffusibility less when once inside than before its entrance from the uterus. However this may be, the most noticeable fact of the development of the rabbit embryo during the fourth day is the commencement and enlargement of the cavity of the blasto- dermic vesicle. EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 1831 Up to the moment of the beginning of the cavity there does not appear to be any pressure upon the walls of the embryo by zona radiata and albumen layer ; in fact, the embryo may often be found to be slightly retracted from the zona radiata in the fresh state. The diameter of the cavity within the zona radiata is, in the unsegmented ovum, about 0°11 mm., and up to the time that the cleft appears it has not greatly enlarged, measuring only about 0°12 mm. On the appearance of the cleft the cells of the outer layer become pressed hard against the zona radiata and flattened, but no great increase in diameter of the outer border of the zona radiata is as yet recognisable, although the thickness of the zona radiata is diminished, apparently being of a com- pressible nature, and therefore becomes compressed between the pressure from within the blastodermic vesicle and the resistance afforded by the tough albumen layer from without. At this time, being in the very firm lower part of the Fallopian tube, the resistance afforded by the albumen layer is very likely aided by the walls of the Fallopian tube itself. In this condition the embryo usually passes into the uterus, although sometimes specimens may be found in the uterus (but very rarely) in which no cleft has as yet appeared. The embryos, I think, pass rather suddenly through the last 4 to 6 mm. of the Fallopian tube at some time between the 75th and 80th hours after coition. No increase takes place in the thickness of the albuminous layer after entering the uterus, but by the stretching of it caused by the expansion of the blastodermic vesicle it rapidly thins. The zona radiata thins so much as to be hardly per- ceptible by the end of the fourth day. Van Beneden’s figures, 5, 8, 9, 10, 11, on pl. iv of his paper, taken from optical sections, represent very well the appearances presented by the embryos during these changes. As van Beneden’s figures represent optical sections, it is, I think, advisable to give a series of figures drawn from real sec- tions, as none have hitherto been published. 182 RICHARD ASSHETON. Figs. 20 to 34 are all camera drawings, and each is magnified 465 times. Figs. 20 to 25 show sections through the whole embryo, the subsequent figures through the embryonic disc, or a part of it only. Fig. 20, and also fig. 16, are sections of silver nitrate preparations. In both cases there is no difference between the cells at the surface and those towards the centre as regards their colouring or nuclei. Those inside are certainly more com- pressed than those on the surface. So also it frequently happens that a large segment may be so placed that although the greater part of it may be said to belong to the “inner mass,’’ yet a small part of it may appear on the surface, as 2. in fig. 20. If such a specimen is examined in optical section, this individual cell might very likely give rise to such an appear- ance as van Beneden describes, and lead to the idea of an inner mass partially surrounded by an outer layer. The difference in colour which van Beneden describes is totally absent in real sections, and I can find no greater opacity of the inner mass in optical section than what can be equally well explained by the greater thickness through which the light must of necessity pass in viewing a sphere in optical sections. I am not able to offer any explanation of the condition figured by van Beneden in his second figure ; I can only say that I have have not been able to find it. Fig. 21 is a section through an unusually large specimen. This specimen was preserved in Perenyi. This specimen I believe shows the earliest stage in the formation of the cavity of the blastodermic vesicle (C. BL.). There is as yet no evidence of an internal hydrostatic pressure, but the outer cells form a compact layer, and seem at one point to be as it were lifted away from the inner cells, leaving a slight cleft (C. BL.). In fig. 22 this cleft has increased considerably. It must be noticed that it does not extend through more than about 240°. EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 13838 Through the remaining 120° the inner mass is still as much part of the wall of the vesicle as the outer layer of cells. In fig. 23 the cavity shows a further increase in size. The outer layer of cells (O. ZL.) now show signs of having become stretched, due, I believe, to the rapidly increasing hydrostatic pressure within the blastodermic vesicle. The walls of the vesicle are now, while living, firmly applied to the zona radiata and albumen layer. The space shown between the vesicle and the zone in the figure is due to reagents. The same remark applies to all the figures from the 22nd onwards. I have not been able to distinguish sharp lines of division between the outer layer cells in section after the commencement of the blastodermic vesicle cavity. In surface view there are certain lines of divisions, which, as van Beneden has shown, are very clearly brought out by silver nitrate. In this fig. 23, which I believe to be a median section, there is a fairly clearly defined line marking the inner mass from the outer layer, which is less perceptible in fig. 22, and which becomes very distinct in later stages, such as in fig. 28. It is now possible to distinguish clearly an inner mass as separate from the outer layer. This separation seems to be caused simply by the tension being more acute in the outermost of the segments of the mass J. M. in fig. 22, causing these seg- ments to be more stretched than the remainder, because they are more directly united with the already separatedand stretched cells O. L. Lines of division between the segments of the inner mass can usually be found, but varying greatly in definition. I cannot say to what extent these several segments may be really divided. The inner mass is certainly connected in some way with the outer layer, but whether by direct protoplasmic union I am quite unable to say. Fig. 24 presents no new characters excepting a tendency to flattening of the inner mass. Im figs. 25, 26, 27, the inner mass is flattened out still more, so that it presents a lenticular form in section instead of a circular outline as it did in figs. 22 and 23. How does this change in form come about ? 134 RICHARD ASSHETON. Van Beneden (p. 29) describes the process thus :—“‘ La masse cellulaire endodermique s’aplatit 4 la face profonde de Vectoderme et la surface de contact entre les deux feuillets s’étend peu a peu.” This is a quite correct account of the appearances but in that the description tends to suggest an inherent inclination on the part of the inner mass cells to flatten themselves; it is, I think, right to look for some other explanation. The inner mass is attached in some way to the outer layer, and accordingly it seems only probable that this mass should be drawn out with the stretching of that part of the outer layer to which they are attached, during the expansion of the whole blastodermic vesicle by the increasing hydrostatic pres- sure within. Figs. 22—27 are all sections of preserved specimens, and so mostly crumpled to a certain extent. If, however, the angle subtended by the arc of the wall of the blastodermic vesicle to which the inner mass is attached is measured in optical sections of fresh specimens of ages corresponding to my figs. 22 to 27,it will be found that the angle is nearly constant. This angle is about 80° (see my diagram, fig. 41). Up to this time there is no indication of a separation into what one may call epiblast and hypoblast. There is a separation into outer layer and inner mass, but this I believe to be of no palingenetic importance whatever. The separation I have tried to show is due to the individual circumstances of ontogeny entirely. So also, and as I shall endeavour to show in the next chapter, I believe the ultimate separation of epiblast and hypoblast is also devoid of all palin- genetic influences such as any form of invagination, or epibole, which is usually spoken of as “ gastrulation.” The several segments of the inner mass change their shape in response to the change of their individual environment. In figs. 20, 21 they are compressed and hence polygonal. As the pressure is removed during the growth of the cavity C. BL., those towards the cavity become rounded (figs. 22—27), which character tends to become universal in figs. 28, 29, with excep- HARLY STAGES OF DEVELOPMENT OF THE RABBIT. 185 tion of some which subsequently come under the influence of tension, to which I shall refer again. CHAPTER: if. Toe APPARENT EXTENSION OF THE HyPpoBLast. The Fifth Day (97th to 120th hours). In the description of the shape and growth of the blasto- dermic vesicle, and the discussion which follows, I call the part of the blastodermic vesicle to which the inner mass is attached the upper pole, the part immediately facing this and furthest removed from it the lower pole. The zone midway between the poles I call the equator. The part whereat the central portion of the inner mass is attached to the outer layer, that is to say that part where there are three layers of cells, I call the embryonic disc, but I do not thereby mean to say that only that area and none other takes part in the formation of the adult. The events of the fourth day are concerned almost entirely with the enlargement of the embryo. During the fifth day this enlargement continues, but other important developments now take place. It is during this day that we can first detect a separation into the two well-known embryonic layers, epiblast and hypoblast, and as to the mode of origin of these layers I fear I cannot quite agree with the description and explanations of any former writers on the subject. As I shall point out in a future paper, I can see no necessity for the occurrence of any folding in or growing in of cells from without, usually known by the term ‘“‘ vastrulation,” and most certainly I can find no evidence of it whatever. As far as I know, only one investigator has stated that he has found actual evidence for this in the case of the rabbit. This is Keibel, in his work “ Zur Entwickelungs- geschichte der Chorda bei Saiigern,” ‘ Arch. fiir Anat. und Physiol.’ 1889. In this Keibel only found his “ blastopore ” 136 RICHARD ASSHETON. in one specimen, in which the albumen layer and zona radiata had been removed (thereby rendering it very likely that a tear may have occurred), and he was quite unable to find it in other specimens. If such an aperture really existed, it would cer- tainly be perceptible in surface views of specimens stained with silver nitrate. Van Beneden makes no mention of such an aperture. I have hunted carefully but can find none, either in surface views of specimens treated with silver nitrate and other reagents, or in series of sections; in fact, I should be greatly surprised to find such an aperture at the time given by Keibel for its appearance. It seems to me quite useless to look for it, since it is sup- posed to represent a turning inwards, such as can be seen in Amphibian ova, for the Mammalian embryo at this age is utterly unlike the Amphibian embryo at the corresponding age, and therefore the conditions which lead to the apparent turning in as seen in Amphibia are hardly likely to be existent in the case of the Mammalian embryo. It may be well to point out that this is not the view of so great an authority as Dr. Oscar Hertwig, who says, _ p. 90 of his text-book, when speaking of this spot, “ Von dieser Stelle aus, nehme ich an, hat sich schon auf einen noch friiheren Stadium das untere Keimblatt durch Um- schlag eines kleinen Bezirks der einblatterigen Keimblase entwickelt.” The changes that occur during the fifth day affect both inner mass and outer layer. The outer layer consists throughout the fifth day of flattened lenticular cells, whose boundaries are easily recognisable in surface views, and I have nothing to add to van Beneden’s careful and accurate description of them. They are of very various shapes and sizes, the lines which mark their boundaries may be at any angle to each other. The boundaries may number six, five, or four, the most usual form being the bexa- gonal. As the vesicle enlarges the cells become rather smaller, and become much less regular in outline, so that by the seventh EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 137 day the cells of the outer layer may be almost circular, or may have any number of sides curved or straight. Changes in the inner mass.—These are very important during the fifth day. Up to now the cells of the inner mass have been very uni- form in character. During the process of flattening of the mass (a stretching out as above suggested) the individual cells have likewise become flattened, and each is somewhat lenticular in shape (figs. 27 and 28). The inner mass presents an approximately circular outline when viewed from above or below, as in fig. 38, but here and there single cells (HY. J.) seem to jut out somewhat beyond the others. So we may say at this moment that the embryonic disc is several layers of cell thick at the centre, but thins out toward the periphery (figs. 26, 27). The whole embryo is now spherical. Early on the fifth day, somewhere, as a rule, between the 96th and 100th hours, the cells which were noticed before as jutting slightly from the sides of the “ inner mass ” may now be seen to be quite separated from it, as in the diagram fig. 39, HY. L., standing out clear and round. These are not easily seen in the perfectly fresh specimen, but are brought out beautifully by certain reagents, as, for instance, Flemming’s strong solu- tion, or a filtered mixture of one part Perenyi with one part picro-carmine ; indeed, almost any reagent that stains slightly the cells and not the albuminous layer. At the same time the cells of the centre of the inner mass have become more flattened, and can now be seen to form a patch of cells, nowhere more than two cells thick. That is to say, the embryonic area is composed of, in all, three layers of cells, the outer a definite membrane continuous with the general wall of the vesicle, and two other looser layers of cells slightly flattened but much thicker and rounder than the cells of the outer layer. Beyond the periphery of this mass, a number of cells very much rounder may be seen scattered irregularly over the inner surface of the thin outer layer, extending over an arc of 138 RICHARD ASSHETON. about 60° from the upper pole in all directions (fig. 28, HY. @,): From this moment the apparent changes for some hours seem to concern the innermost layer of cells of the embryonic disc, and the straggling cells. The hypoblast, as a perfectly definite layer, is formed by the time that the blastodermic vesicle measures ‘5 mm. in diameter, that is, by about the 102nd hour after coition. It is not, how- ever, as yet by any means a continuous membrane, it is a net- work or a fenestrated membrane. For this reason in section it appears to be represented by isolated cells lying beneath the embryonic disc (v. fig. 29, HY.). Certain cells can be detected in earlier stages which from their being more lenticular, and more inwardly placed, can probably be described as hypoblast cells (v. fig.28, HY.). This stage is about the ninety-eighth to one hundredth hour, when the diameter of the blastocyst measures about ‘36 mm. Also all the apparently—and in many cases I believe actually— isolated cells (HY. I. in figs. 28, 29, 87—40) can from the moment of their apparent wandering be termed hypoblast. By the separation of the cells round the periphery of the inner mass, and certain others of the more inwardly placed of the inner mass into what we may call hypoblast, there are left cells in between the hypoblast and outer layer, which do not show the same tendency to be flattened or drawn out as the others. As is well known, this intermediate cell layer forms part of the epiblast, and so may be termed the inner layer of epiblast. To this I shall refer later. I wish now to refer to the straggling cells (HY. J.) of the inner mass, and consider why they apparently wander round the inside of the blastodermic vesicle. That they arise from the inner mass I have no doubt. I have explained above the flattening of the inner mass as due to being drawn out by the general expansion of the walls of the vesicle, and to this may also be due the original isola- tion of certain cells round the periphery of the inner mass as seen in figs. 28, 37, and 38. EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 139 If the cells of the inner mass do not multiply quickly enough, and if they are not connected together firmly, and if they adhere by some means or other to the outer layer, then, as the outer layer expands, there must be a tendency for the general separation of the cells of this inner mass, and this will be most apparent at the edges. That is to say, the centre, originally thicker than the edges, will become thinner, and at the periphery the edges will be drawn out to an irregular or regular fringe according to less or more regular local growth and expansion of the outer epiblastic layer. If the spread of these is due to this cause, and if the growth is equal over the whole surface of the sphere, then, in a section taken through the centre of the sphere and the centre of the embryonic disc, the are along which these isolated cells are found should subtend an angle equal to that subtended by the compact embryonic disc up till now; that is to say, an angle of about 80°. But on measurement at a rather later stage, 102 hours after coition, represented in fig. 89, the angle is found to be very considerably wider than 80°, in fact somewhere between 110° and 130° (the limit of these cells in question being very irregular), and by the eighth day an angle of 200°. Hence it would seem that this does not account for the apparent growth round. If, however, we have any reason to suppose that the cells of the outer layer multiply more rapidly and thus allow the expansion of that part of the sphere to take place quicker around a zone bounded roughly by the edges of the compact embryonic area on the one hand and some other parallel line about the original equator on the other hand, we could still consider that the suggested cause is the correct one. I think we have, for these reasons : The blastodermic vesicle of 96 hours is a sphere. The blastodermic vesicle of 120 hours is no longer a sphere, but is a body of such a nature that a horizontal plane taken through its longest diameter will not pass through the equator, but will be nearer the upper pole than the lower. 140 RICHARD ASSHETON. The circumference of this plane and of all planes taken parallel to it are, I think, as yet circles. There is nothing so far to indicate which will be anterior or posteriorend of the embryo. The blastodermic vesicle of about the 140th hour is in shape of the same character, but more markedly so. . In the blastodermic vesicle of the 175th hour sections taken parallel to the equator are now no longer circles, but are ovoidal. The fact that there are now a long and a short axis may be entirely due to the pressure upon the vesicle exerted by the walls of the uterus; but I want to point out more especially that the horizontal sections are not true ellipses, but have one end larger than the other. The large end corre- sponds approximately, but by no means always exactly, with the future posterior end of the embryo. A median longitudinal vertical section of this stage is very instructive, for the asymmetry is very marked. One end, the future posterior end, is much more bulky than the anterior end. The embryonic area is always placed nearer to the anterior than to the posterior end. Fig. 42 shows four stages, namely, at the 100th hour, the 125th, 140th, and 175th. A. is the more anterior, P. the more posterior end. The thick black line in each case is the outer layer of cells or epiblast; the thin black line represents those cells of the inner mass which become separated to form the hypoblast. It is quite clear that a very great change in shape as well as size has occurred between the 96th and 168th hours. How has this been produced ? I have argued that there is a hydrostatic pressure within the blastodermic vesicle causing it to expand. The pressure is sufficient to cause the stretching and flattening out of the cells of the wall of the vesicle, and also to cause the stretching and expansion of the very tough albumen layer. On the other hand, we know that the pressure is not so great as to rupture either the vesicle wall or the albumen layer; in other words, the vesicle wall and albumen layer are sufficiently strong to resist the hydrostatic pressure for, at any rate, some con- siderable time. The albumen layer becomes continually more EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 141] and more stretched, and quite late it does, as a matter of fact, rupture, but at a time (the ninth day) that does not con- cern the present question. No addition is ever made to the thickness of the albumen layer after the embryo leaves the Fallopian tube; it is an inanimate structure. The cellular wall of the blastodermic vesicle is, on the contrary, living, and capable of adding to itself at any part of its area. Like the albumen layer, it becomes greatly stretched and becomes very thin (v. figs. 21—29), and unless it received additional matter (i.e. multiplication of the cellular units) it would rapidly thin out altogether. After about the 100th hour the cellular wall ceases to get any thinner. Up to this moment we must suppose that the rate of increase of hydrostatic pressure has been in excess of the rate of addition of material to the cellular wall of the vesicle and so has stretched it, but from now there is no appreciable thinning out of this cellular wall (v. figs. 27, 28, 29, and 34). This means, I believe, that the increase of cellular tissue just balances the increase of hydrostatic pressure, and so the vesicle grows in size, the thickness of the cellular wall remaining unaltered. Now there is no reason, as far as I can see, to doubt the albumen layer being equally tough on all sides of the embryo. The albumen layer is not secreted by the embryo, but is applied by the Fallopian tube. Although preserved specimens and sections are not well adapted for this purpose, still my figures show that there is no regularity at all in these preserved specimens, such as a thicker part of the albumen layer being present over any one part of the embryo in stages up to the 96th hour (v. figs. 16, 20, 21, 22, 23, and 24), and in the fresh specimens the outer and inner limits of the albumen layer present in optical section true concentric circles. Therefore I do not think we can attribute any subsequent difference in thickness of the albumen layer to a difference in texture acquired by that albumen layer during its deposition. At a later stage we do find a difference in thickness occurring as a constant character. The difference I find is as follows: 142 RICHARD ASSHETON. The most marked difference concerns the part of the albu- men layer adjoining the embryonic disc. This is in the later stages very much thicker than elsewhere. Figs. 30 and 31 are portions of the upper and lower poles of an embryo taken from a rabbit of the one hundred and forty-fourth hour. The albumen layer is at the upper pole about twice as thick as it is at the lower pole. This is a perfectly constant feature, and becomes more and more marked the later the stage. Hydrostatic pressure exercises its influence equally in all directions. If it encounters less resistance in one direction than in another it will cause greater effects in that one direc- tion. I have argued above that at any one moment (between one hundredth and one hundred and sixty-eighth hours) the hydrostatic pressure within the vesicle on the one hand, and the living cellular wall of the vesicle together with the non- living albumen layer on the other hand, are in a state of equilibrium. The blastodermic vesicle is always taut, but does not rupture. But the next moment the hydrostatic pressure has increased on the one hand, and the albumen layer has become thinner and the cellular wall has increased its material —and still the equilibrium is maintained. It seems to me to be clear that the degree of rapidity of increase of the cellular wall must be a factor in the resistance afforded to the hydro- static pressure by the two walls of the vesicle. If this is so, then it follows that at any area where the increase of cellular tissue is greatest, there the hydrostatic pressure will exert its greatest influence, and there the albumen layer will, as a consequence, be thinnest. This is, of course, dependent on the close attachment of the cellular layer to the albumen layer. I am bound to confess I do not find it easy to prove that there is no sliding of the albumen layer over the cellular layer; but, on the other hand, I see no evidence to suggest that there is such a sliding. Accordingly, I take it that the great thickness of the albu- men layer adjoining the embryonic disc over that adjoining the lower pole shows that there has been less stretching and EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 148 therefore less growth at the embryonic pole than at the lower pole. It may be said that this is due to the fact of the cellular wall at this point being practically several layers thick. But it must be remembered that the cells of the inner cell-mass are very loosely arranged (figs. 28, 30, 35) at this time, and certainly do not give me the impression of being under great tension, as are the outer cells. As far as the tension is concerned, I believe the outer layer has to bear it very nearly all, even in the embryonic disc. Sections show how very thin it is here. Granted that they are only very equally stretched, it follows, I think, since the albumen shows not the same amount of stretching as elsewhere, that the multiplication of cells must have been going on more slowly in the embryonic region. Now from measurements of the albumen layer it will be seen that the thinnest place is not usually at the lower pole, but somewhere between the equator and the upper pole. This, then, marks out as an area a zone of more rapid multiplication of cells, that area which lies just outside of the embryonic disc. Now we know that upon the eighth and many subsequent days there is a very great activity evinced by the cells of a zone immediately surrounding the embryonic disc; it is the zone called by Duval the ectoplacental area. If it is an area of very great activity upon the tenth day, of great activity upon the ninth and eighth days, when does it begin to be an area of activity ? Why not upon the seventh or sixth, or even on the fifth day ? I believe that it does begin as early as the fifth day, and that it is to the presence of this area of more active cell division round the embryonic disc that the shape of the blastodermic vesicle is as I have described it to be (v. fig. 42), and that it is to this that the apparent growth round of the hypoblast cells is due. The ectoplacental area is, as is well known, much more strongly developed all round the posterior end of the embryo 144, RICHARD ASSHETON. than anteriorly. Although present to a slight extent at first anteriorly, it never develops greatly. This means that the zone of special activity is at a later stage more intense poste- riorly than anteriorly. So apparently it is at an earlier stage, and causes the greater bulkiness of the posterior end of the blastodermic vesicle of the seventh day, and causes the appa- rent throwing forward of the embryonic disc (v. “175 hr.,” fig. 42). Table of Measurements of Thickness of Albumen Layer at Different Parts of the Vesicle in Mil- limetres. Five Days. Six Days Four Hours. Embryonic area , ; ‘0060 : : 0058 Placental zone, 4. . : 0024 ; : "0022 Placental zone, P. . ; 0030 ; : "0024 Lower pole ; : : ‘0036 : 5 "0034 To sum up the above section: (1) There is no growth round of the inner mass cells over the surface of the outer layer cells in the sense of a migration. It is only an apparent growth round produced by the more rapid growth of a zone of the wall of the vesicle immediately surrounding the embryonic disc, in which zone the marginal cells of the inner mass lie. (2) The presence of such a zone accounts in a great measure for the shape assumed by the blastodermic vesicle during the fifth, sixth, and seventh days. (3) A zone of such a character undoubtedly exists during the eighth, ninth, and following days, giving rise to the ecto- placenta. (4) Such a zone of activity accounts for the varying thick- ness of certain parts of the albumen layer. To return for a time to the question of the growth round of the inner mass cells. It must be remembered that the cells never completely surround the cavity of the blastodermic vesicle. The lower pole is always, in the rabbit, one-layered only as long as it exists. EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 145 There is only one other alternative that will account for the apparent growth round of the inner mass on the inner wall of the blastodermic vesicle, as far as I can see, and that is actual active migration of the cells in question. Of course it is difficult to bring evidence to show that they have not migrated, since it is not possible, I fear, to follow the process in any one and the same specimen. At the same time I cannot find any evidence to prove that they have actually migrated. If a cell does migrate, like an Ameceba for instance, one would expect to find evidence of protoplasmic protuberances or pseudopodia. Of this I can find no trace. The majority of the cells seem at first (figs. 38, 39) to be quite isolated from each other, and to be approximately spherical, whether examined in the perfectly fresh condition or after treat- ment with various reagents. They are, it is true, slightly flattened on the side by which they adhere to the vesicle wall (fig. 28, HY. J.). Certain of the cells here and there are connected by threads of protoplasm, but this, I think, is not a sign of pseudopodic activity, but merely indicates the final stage in division between the two cells. I have no doubt that these cells divide rapidly after a time, though I do not think much activity of division takes place during the first few hours after the apparent migration begins. If one of these inner rounded cells is undergoing the process of division, then, as the wall on which it rests expands, the two dividing halves of the inner cell will be pulled apart, and a strand of protoplasm connecting the two cells may remain for some time. Of course it is just possible, I suppose, that these rounded inner cells might migrate by means of a rolling motion con- sequent on ‘*‘ streams” of protoplasm within them, as do some protozoa. But is such a phenomenon known anywhere in the metazoan body? I cannot think we are justified in assuming this without evidence, for when examined in the fresh con- dition no such protoplasmic activity can I notice. When the cells are so isolated as I believe them to be during VOL. 37, PART 2,—NEW SER. K 146 RICHARD ASSHETON. the 98th to 108th hours or thereabouts in the rabbit, I cannot conceive that the migration can be accomplished otherwise than by an actual migration, for which there is no evidence, or else by a process such as I have attempted to describe above. This I believe to exist. If the inner layer was not composed of isolated cells, but was a compact membrane, then it might creep round by means of its own interstitial growth, although I do not think it would in that case be a thin smooth membrane. Such is the account of the growth round of the hypoblast in the mole. Heape makes no mention of any isolated cells at the edges; he says of the hypoblast, ‘‘it extends laterally by virtue of the multiplication of its cells, which at the same time become much flattened.” It seems to me to be much more likely that the layer would be flattened if they were drawn out aloug with the epiblast cells ; but I have no evidence at present whether the same cause can be attributed to the spread of the hypoblast in the mole as I suggest for the rabbit. In fig. 42 the dotted lines radiating from the centre of the smallest blastocyst indicate the amount of growth of the several segments up to the ove hundred and seventy-fifth hour. Until the one hundredth hour I imagine the growth of all parts of the wall of the blastocyst to be equal. From that moment there is a greater activity in an area around the embryonic disc, which causes the inner layer to be apparently carried further round the anterior of the blastocyst. This zone of activity by the one hundred and fortieth hour has become more marked still, and now shows itself to be more intense posteriorly than anteriorly, which latter character is plainer still at the one hundred and seventy-fifth hour. Eventually the activity culminates, upon the addition of resistance afforded by the walls of the uterus, in the production of the ectoplacental area, or part of it. To this I have referred in another paper, and also to the hypothesis that this zone of activity, in the absence of a tough albuminous coat, results in certain other rodents in the pro- duction at once of a heaping up of cells—the trager. In the embryonic disc region there is very much less activity EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 147 exhibited. The outer layer in that region is very much attenuated, and shows very little sign of activity. The inner layer consists all this time of rounded or ellipsoidal cells, and, like the outer layer of epiblast of that region, shows very little if any sign of activity. So little activity of division does there seem to be in the inner layer of epiblast, that there is a distinct tendency for the several cells to become slightly separated as in fig. 30, which gives rise to the very irregular and speckled appearance of the embryonic disc of the one hundred and twentieth hour. Very probably this is caused by the slight stretching of this region. It is more noticeable at the edges than towards the centre. Whether there is any palingenetic meaning in this double- layered condition of the epiblast I have discussed in another paper. For the present, I think the right view to take of the condition is that derived from the study of the actual way in which the separation has originated, and to regard it as a con- sequence of ontogenetic circumstances only. The Outer Layer of Epiblast.—This is by far the most active of the embryonic layers of the fifth day. It is in an active condition of growth during the whole of the day, and thereby allows of the expansion of the vesicle. The character of the cells seems to be just as it was during the latter part of the fifth day. The Inner Layer of Epiblast.—This layer seems to be a region of rest throughout the whole of the sixth day. There is very little sign of cell multiplication. The cells are more or less circular in outline when viewed from above, and oval when seen laterally. They are rather scattered, and thus give rise to the speckled appearance noted above. They seem to be clearly separated from the outer epiblast above them and from the hypoblast below them, and since they are either quite separated from each other, or connected only by fine strands of protoplasm at certain minute spots, they are simply pulled apart by the expansion of the blasto- dermic vesicle, and are not individually stretched and flattened, 148 RICHARD ASSHETON. as are the cells of the outer epiblast and hypoblast, where they are more intimately connected one with another. In the latter case, that of the hypoblast, cells with a tendency to the features characteristic of both layers of epiblast are to be found. CHAPTER IV. CHANGES THAT OCCUR DURING THE SixtH Day (120TH To 1447H Hours). If we may call any period of the development of an animal unimportant, it is to the period between the 120th and 144th hours of the development of the rabbit that this epithet might be applied. During this day the blastodermic vesicle increases very greatly in size, and assumes very markedly the shape described in the last chapter as being characteristic of the later stages of the development of the vesicle, prior to its attachment to the walls of the uterus. The blastodermic vesicle is no longer a sphere. It will be found by measurement to have longer and shorter equatorial axes, and a polar axis which is of less length than the shortest equatorial axis. The following measurements are taken from specimens of the earlier part of the sixth day : L t Short 3 equatorial equatorial Polar axis. pt re axis. - No.1. 5 days 5 hours . .| l'1 mm. | 1:05 mm. | 1:00 mm. | 0°55 mm. B02. Bois Beek heel BB gy |) 088) Re 0 -vee (1) Hypoblast of the Embryonic Disc.—This is now a continuous layer, sufficiently so as to show lines of demarcation between the cells when treated with silver nitrate. This con- tinuous membrane extends a short distance beyond the peri- phery of the inner epiblast layer. The cells composing this membrane are completely flattened. (2) Hypoblast beyond the Embryonic Disc.—The EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 149 scattered hypoblast cells have now become much more numer- ous, and are scattered more evenly over the portion of the wall upon which they are to be found. Many of them, possibly all of them, are now undoubtedly connected by more or less fine protoplasmic threads. These scattered cells, although such conspicuous objects during the fifth day, are now extremely difficult to make out, and can very easily escape notice. They are more numerous now nearer to the embryonic disc, and merge gradually into the continuous layer just described. They are now very much more flattened. Towards the line of their outer limit they present more the characters of the former day, being fewer and rather rounded and more isolated. Consideration of the Extent to which the Shape of the Cells of the several Layers may be attri- buted to Mechanical Causes. At this time we find cells of two very different types, with cells showing all intermediate stages between the two. The first type is the rounded, almost completely isolated cell, such as those of the inner layer of epiblast, or near the outer limit of the hypoblast ; the second type is the flattened or stretched cell of the outer layer of epiblast or embryonic hypo- blast, continuous with its neighbours around all its edges; and thirdly, forms of cells intermediate between these two types. How far can we hold this difference of form to be due to the environment of the individual cell apart from its own inherited tendencies ? Whether there is any cell in the embryo at this time quite separated from all others I am not certain. Of course they are all contiguous to one or more other cells, but possibly there is an actual protoplasmic union between one cell and another. This is certainly the case very frequently with the cells of the hypoblast, which at first sight seem to be quite separate. In sections, and in surface views, there is no distinct connec- tion to be made out between the rounded cells of the inner 150 RIOHARD ASSHETON. epiblast layer. But when the embryonic disc is broken up in such a way as to scatter and tear apart the various cells, then there undoubtedly appear to be strands passing from some of these rounded cells to others of the same layer. At the same time I cannot say positively whether these strands are really connections between the cells in question, or whether they are fragments and shreds derived from the tearing apart of the hypoblast or outer epiblast layer, between which they are placed. But however this may be, the cells of the inner epi- blast layer are, at the time I am speaking of, either isolated or else connected only at certain spots of small area. These are of the rounded type. At the outer limit of the hypoblast there are also cells, some of which, I believe, may be quite isolated; others are connected to each other by a few strands of protoplasm. These approach very closely to the rounded type of cell. This is the type of cell which I believe to be the most natural, by which I mean the least influenced by its environment. This is the type of cell which first comes into existence in the segmentation of the ovum, when, within the protecting investments, the cells, at first uninfluenced by pressure or tension from without, or from each other, assume their natural or spherical contour. As the seg- mentation proceeds, the inner segments become the more com- pressed, and assume polygonal forms. After the establishment of the cavity of the blastodermic vesicle, the outer cells, by the pressure from within increasing more rapidly than their rate of multiplication, are drawn out into thin plate-like cells. These outer layer cells are from the first connected with each other by their edges, and form a continuous membrane, a condition without which in all probability the formation and enlargement of the blastodermic vesicle could not be produced. As long as this tension within is maintained at arate greater than the rate of multiplication of the cells, the cells retain their flattened condition. What of the inner mass cells? Upon the removal of the pressure of the outer layer, the more outwardly placed of the EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 151 inner mass cells re-assume, at any rate on their free surfaces, the rounded contour which is natural to them. As the blastodermic vesicle expands, the inner mass, which is adherent to the wall of the vesicle either by actual proto- plasmic connections or otherwise, is drawn out into a lenticular shape. I have tried to show that there is a zone of the wall of the vesicle which, by greater activity of multiplication of cells, admits of more rapid expansion of that part. Upon this zone rests the edge of the lenticular inner mass. The expansion of the zone is in direction radially from the embryonic poles. Hence the outermost cells of the inner mass, i.e. the cells at the edge of the lenticular mass, will tend to be separated more rapidly than the innermost. This will tend to isolate these cells from others of the inner mass. Let us suppose that all the cells of the hypoblast layer are dividing at a uniform rate. I think it is reasonable to suppose that the existence of strands connecting cells of this kind together have their origin in past cell divisions. Accordingly on this supposition the con- necting strands will be more numerous and the nuclei nearer together, and the meshes of smaller area in the embryonic disc hypoblast than in the hypoblast outside that area. The hypo- blast cells of the embryonic area will differ from those of the extra-embryonic area in this way : (i) The embryonic cells will have more and shorter, and so presumably stronger, strands connecting them with their neigh- bours than will the extra-embryonic hypoblast cells. (ii) The embryonic hypoblast cells will have connecting strands upon all sides, whereas the outermost extra-embryonic cells will have them upon one side only. Is it possible for the flattening of the embryonic hypoblast cells to be due to their becoming stretched by the tension produced by the extra-embryonic hypoblast cells (with which they are in direct connection by means of the strands just men- tioned) being removed in all directions by the rapidly expanding zone of the outer epiblast? If so, it is possible to account for all the shapes of the cells composing the embryo at this age. 152 RICHARD ASSHETON. As I have stated on a previous page, the hypoblast of the embryonic area is a network at first. Also, I believe that at first many of the outermost cells of the extra-embryonic area are really isolated. These will be under less tension than those near the embryonic pole, as they will, if they are connected at all with other hypoblast cells, have connecting strands upon one edge only. Hence these preserve for a longer period their rounded contour. The ultimate conversion of the isolated cells into a network (or a series of networks) and of the networks into thin continuous membranes, and of thin continuous membranes into columnar membranes, would seem, therefore, to be but the result of increase of rate of multiplication over rate of expansion. The inner layer of epiblast cannot be said to have come into existence as a layer until after the formation of the hypoblast. Until that moment it formed, together with the future hypo- blast, the inner mass. It was impossible, except in as far as could he premised from their position, to say from their characters which would be inner epiblast cells and which hypo- blast cells (v. fig. 28). What I believe takes place is this. Those cells of the lenticular inner mass which, being at its edges, are removed by the expansion of the wall of the vesicle, and those which are in direct connection with the cells so removed, become by virtue of their position the future hypoblast ; the remainder become the inner layer of epiblast. That is to say, those cells of the inner mass which are not influenced by the expansion of the vesicle, as above described, and are accordingly upon that part of the wall, though not actually as yet part of it, which is least affected by the expanding forces, become the inner epiblastic layer. Of all the cells, therefore, in the embryo at this time, these (the inner epiblast layer) are least affected by external causes. These cells are the more free to assume their natural shape, which I believe to be spherical, and are only slightly flattened between the two layers, outer epiblast and hypoblast. EARLY STAGES OF DEVELOPMENT OF THE RABBIT, 153 CHAPTER V. CHANGES THAT OCCUR DURING THE SevenTH Day (145TH To 168tH Hovrs). The Fate of the Outer Layer of Epiblast (or Rauber’s Layer) in the Embryonic Disc. The embryos have now grown to such a size as to cause them to respond more effectually to the impulses set up by contractile movements of the muscular walls of the uterus, and therefore we find them much further advanced along the uterine tube, and more scattered. They have not, how- ever, taken up a permanent position as yet, for although this may occur in some cases during the last few hours of the seventh day, more usually it does not take place until the early part of the eighth day. It will be best to describe the course of events in the three layers separately. Outer Layer of Epiblast.—Very little need be said of the greater part of this layer, no change except such as has been de- scribed as occurring during the fifth and sixth days takes place. But special attention must be given to that part of the area which lies over the patch of inner layer of epiblast, i.e. embryonic disc. Inner Layer of Epiblast.— During the early part of the seventh day the cells of this layer are just as described in the preceding chapter. They extend now over an area of about °6 mm. The general outline of the mass is still circular. Each cell is distinct and rounded, with very large nucleus ; and with nearly all stains that I have used they stain only lightly. Early on the seventh day these cells show signs of greatly increased activity. They multiply, become pressed together, and now form a very compact layer at the same time as certain changes occur in the outer layer of epiblast. The course taken by these two layers during the next few hours, and its significance, have been very differently described 154 RICHARD ASSHETON. by different authors; indeed, very opposite views have been held during many years. It is an interesting question from the extreme difficulty of the investigation and from the morpho- logical problems connected with its solution. As regards the actual facts, there have been three quite distinct accounts. Most observers have noticed three layers at this stage: (1) an outer thin layer; (2) a middle thick layer; (3) an inner thin layer. These accounts are very briefly as follows: (1) Van Beneden maintained that the inuer layer is not epiblast at all, but is mesoblast, and that the outer layer becomes thickened over the embryonic disc area, and gives rise by itself to the epiblast of the embryo. (2) Rauber, Lieberkuhn, and later Kolliker and others, hold that the outer layer is quite transitory, and that during the seventh day it splits up, degenerates, and disappears entirely, taking no part in the epiblast formation of the embryo. The epiblast, they hold, is wholly derived from the inner layer of ovoid cells, which van Beneden took to be mesoblast. (3) Balfour and Heape contend that both layers persist as the epiblast of the embryo, but the two layers fuse together by the growth downwards of the outer layer cells in amongst the inner layer cells. There cau be no doubt now that van Beneden was wrong. The middle layer certainly forms part if not all of the perma- nent epiblast, and this is acknowledged by van Beneden himself (‘ Arch. Biologie,’ vol. v). It requires some careful examination to determine whether Rauber, Lieberkuhn, and Kolliker, on the one hand, or Bal- four and Heape, on the other, were right, or whether all were wrong. The surface views given by van Beneden (pls. v and vi) and Kolliker (pl. ii, figs. 13, 15, 16, and 19) are, I think, quite correct. The interpretation van Beneden put upon his figure he no doubt now admits to be wrong. The interpretation put upon them by Kolliker I believe to be correct, namely, that the large areas are cells belonging to the outer epiblast layer, and the smaller ones are cells belonging to the inner epiblast layer. I EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 155 have drawn a figure (fig. 36, Pl. 17) from a specimen aged six days five hours. This is very similar to Kolliker’s aged six days. It isa silver nitrate specimen. This was drawn when the extreme upper surface was in focus. On focussing down it is possible to find small cell outlines under the large areas, but finer than those of the cells (HP. F.). On focussing still further down, the outlines of the hypoblast layer are visible as very fine, very wavy lines, forming as a rule more regular areas. The small cells (EP. I.) are present throughout the embryonic disc, but in some places are much more marked than others; that is to say, in some places the silver nitrate produces the characteristic black marks between the cells more strongly than in others. For this there must be a reason, and this seems to be that in some parts the small cells are at the surface (#P. J.); at others (HP. O.) they are covered by some other body, this other body being certain cells of the outer layer of epiblast (Rauber cells). This is an intermediate stage. LHarlier the outer layer is quite continuous over the whole of the embryonic disc, as shown in Kolliker’s figs. 11 and 12, pl. 11 (see also my sections of these stages, figs. 27, 28, and 30). In these the outlines of the outer layer are distinct as large polygonal areas. In the earliersstages the outlines of the inner layer of epiblast are not sharply defined in silver nitrate specimens, apparently because they are but loosely arranged. But in Kolliker’s figs. 11 and 12, which represent specimens at the latter end of the sixth or beginning of the seventh day, the small cells, i. e. the inner layer of epiblast which now forms a more compact layer, show outlines which are visible below the outer layer cells, although faint. I entirely agree with Kolliker that this outer layer (Rauber’s layer) now becomes broken and, as it were, torn up into isolated cells or little patches of cells, and I think that the cause of this may be as follows. I have above given reasons for supposing that the cells of the embryonic disc region are comparatively inactive during the fifth day and first part of the sixth day. The outer layer of cells (Rauber’s layer), I 156 RICHARD ASSHETON. pointed out, has been stretched to a very high degree of tenuity. From the middle of the sixth day the embryonic disc area becomes an area of increased activity. We have at this moment, therefore, an outer extremely attenuated membrane under high tension due to the hydro- static pressure within. Each “ cell’’ of this membrane repre- sents an individual minute centre of protoplasmic activity. These are so drawn out that a few will extend over a con- siderable area. Closely pressed against these is a layer of rounded or but slightly compressed cells, the inner epiblastic layer. Hach cell of this layer also represents an individual minute centre of protoplasmic activity, and it will be seen that in any given area of the embryonic disc for one of these little centres of activity of the outer layer (Rauber’s) there are three, four, or five of the little centres of activity in the inner epiblastic layer. The hypoblast, although it will tend to press the inner layer of epiblast cells firmly against the Rauber layer, need not necessarily be affected by what I am going to describe, because at no time does the hypoblast, after its initial separa- tion from the epiblast, appear to be at all firmly attached to the epiblast. Note the readiness in which it stands away from the epiblast in sections (v. all my figures, and those of Kolliker). To return to the two epiblast layers. It is clear now that if the two epiblast layers of the embryonic disc acquire an in- creased activity, then if for any given area the inner epiblast contains, say, three times more cells than the outer epiblast, then the inner epiblast will increase its bulk, roughly speaking, three times more rapidly than the outer epiblast. This might produce several effects. It might produce a heaping up of cells; it might produce an arching inwards of the inner layer of epiblast ; it might produce an extension of the inner epiblast by sliding over the outer epiblast (or rather be- tween the outer epiblast and hypoblast) ; or it might cause the EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 157 further stretching of the outer epiblast, or it might cause its rupture. I believe it causes the rupture of the Rauber layer ; whether the several fractures between the cells are sharp, or whether, as is possible, the cells are pulled apart so as to produce large meshes through which the inner layer cells come to the sur- face, I cannot say. If the latter case, although the Rauber layer as a membrane would be obliterated, still the continuity of protoplasm would not be broken ; but I think, from considera- tion of all the appearances, the fractures are fairly clean. We now come to the question of the fate of the Rauber layer cells. Kolliker says they disintegrate and disappear; Balfour and Heape say they pass into the inner layer of epiblast. I am fairly confident that the latter view is correct. If my description and hypothesis are correct, the Rauber cells, as soon as they become broken up, are no longer under such a high degree of tension. If they have any vitality in them they will, as they grow, be able to assume other shapes than ‘‘squamous.” They will become intermingled with the cells of the inner epiblast layer. Their nuclei will always form a knob, and thereby perform the part of the thin edge of a wedge to the whole cell, and so on recovering what we must allow was their more normal shape, lost only under pressure of the fluid within, will tend to become intermingled with the cells of the lower layer. No doubt some will get settled earlier than others, and often upon the eighth day it is possible still to detect a Rauber cell not yet accommodated. Many stages of this process, I believe, are shown in my figs. 32, 33, and 84. These are sections from the same specimen, which was taken from a rabbit at the 153rd hour, preserved in Flemming’s strong osmic-aceto-chromic fluid, and stained with a weak borax car- mine. In fig. 34 to the left is the thin layer of outer epiblast (formerly continuous with Rauber’s layer alone) (EP. O.), lying between the albumen layer (ALB.) and hypoblast (HY.). To the right is the fused or fusing inner epiblast and “ Rauber cells” or outer layer of epiblast. The greater part is stained only slightly, and contains numerous large rounded nuclei 158 RICHARD ASSHETON. (EP. I.). The outline of cells is hardly visible. Here and there, as though filling up interstices, are cells which stain much more darkly (HP. O.), and in the character of their nuclei appear more like the cells of the outer epiblast beyond the border of the embryonic disc. Some of these are only little wedge-shaped bodies on the surface (fig. 32, EP. O.), others pass right through; others seem to send flaps over the surface of the inner layer cells (fig. 833, HP. OR.). The last- named cell (fig. 33, HP. OR.) is very curious. It will be noticed there are several rather like it. These almost look as though they were being enclosed quite accidentally by the inner layer of cells. It seems to me quite possible for this to be very often the case. Take such a spot as that to which the line EP. O. runs in fig. 80; here, of course, the bend in the outer layer is in all probability artificial. But if a gap existed, as frequently happens, between the cells of the inner epiblast, then if at a neighbouring spot the outer layer became ruptured, the tension would be removed, and the Rauber cell might very well become enfolded quite passively, as I believe is taking place in fig. 33, EP. O.R. There is no reason why a cell thus enfolded should die, in all probability it would grow and multiply like any other cell of the region. (It may be observed that of all the cells in the whole blasto- dermic vesicle at this time, none are so badly placed for nutri- ment [provided we allow that nutriment is all the time being received from the fluids of the uterus] as the cells of the outer epiblast layer over the embryonic disc. For they here lie be- tween the thickest part of the albumen layer and the thick inner epiblast layer, which may account for the period of in- activity of this layer at this moment.) I cannot leave this question without a short discussion of the morphological bearings of the events connected with the fusion of these two layers. First of all, I wish to call attention to Heape’s description of the mole’s development. The mole is very like the rabbit in its developmental history. Just after the separation of the hypoblast layer, Heape EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 159 describes the appearance of a cavity between the outer layer of epiblast and the inner layer of epiblast. It is hardly a cavity, for it is partially filled by very “stellate” cells. With the formation of this the inner layer of epiblast becomes arched inwards, a process termed “ temporary inversion of layer ”’ by Heape. Subsequently this arch flattens out again, and it and the outer layer cells and stellate cells between them all fuse together to form the permanent epiblast. I have not studied the mole, but from Heape’s description this seems to be an almost exact parallel to the process which occurs in the rabbit, with this exception: whereas in the rabbit the increase in activity of the inner layer of epiblast gives rise to a rupture of the outer layer of epiblast, in the mole one of the alternatives suggested above is taken, and the inner epiblast bulges inwards, leaving a loose space threaded across with “stellate ” cells between it and the zona radiata. Subsequently on the expansion of the whole blastodermic vesicle the plate flattens out again and the “stellate” cells and other outer layer cells become intermingled with those of the plate and form the permanent epiblast. Heape regards this as a kind of inversion, and the stellate cells as oWirager.’” I cannot agree with Heape in considering the stellate cells he mentions as being equivalent to trager cells, and certainly I do not think that the “ Rauber cells ” of the rabbit are in any way connected with “ trager””—but of this I shall say more in a later paper. As regards the meaning of the fusion of the two layers, I do not see that it need necessarily have any morphological signi- ficance at all. It may be merely an accident of development. At the same time I cannot entirely neglect certain occurrences in another group of Vertebrates, and have discussed them in another paper. By the union of the two layers the embryonic dise acquires a very much more distinct outline, which is now practically circular; its outline is considerably more regular than before the junction just described has taken place. 160 RICHARD ASSHETON. Hypoblast.—(1) Hypoblast of embryonic area. This seems to have become in very slight measure changed. It is now undoubtedly a continuous membrane in the region of the embryonic area. This condition seems to extend a distance from the embryonic area equal to about the diameter of the embryonic area, beyond which it becomes a network and passes insensibly into— (2) The straggling cell portion of hypoblast. This part of the hypoblastic layer retains its irregularly scattered condition of the sixth day, but certain features may be remarked upon which were unnoticed before. The cells are more thickly scattered about; they are more irregular, having entirely lost their rounded form, and are more flattened. Many in all parts may be seen to be connected to- gether by fine filamentous strands, not only in the close proximity of the embryonic hypoblast, but also near its periphery. Again, the outer limit of this zone is much more marked, and is, in fact, now rendered very plain indeed. It forms a well- marked edge, very irregular it is true, but an almost if not quite continuous edge. Along this edge the cells are slightly crowded, and rather elongated in the equatorial plane of the vesicle. What I mean may be made out from fig. 40, and an idea of the general history of events connected with the deve- lopment of this part of the hypoblastic layer may be derived from the four figures 37—40. The extent of area covered by the two parts of the hypoblast is now rather more than half the whole area of the inside of the wall of the blastodermic vesicle. Another point of interest may be noticed. There is a strong tendency for this line of limit to be thrown into small folds or waves, as shown in fig. 40, Pl. 17. These two characters may be observed thoughout the area under discussion. If an embryo of this age is cut in two, and the cut edge examined under a high power, these characters are seen very clearly. That is to say, the cells forming this limiting line are themselves rather more rounded or ‘ hog- EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 161 backed,” and the joining strands are curved and even arch away from the epiblast, and in some cases undoubtedly the cells themselves seem to stand away from the surface. It is, however, quite possible that the sinuosity of the line and arching away may be the result of reagents, as I have not examined this edge in a fresh specimen with success, EXPLANATION OF PLATES 138—17, Illustrating Mr. Richard Assheton’s paper, “ A Re-investigation into the Early Stages of the Development of the Rabbit.” List oF REFERENCE LETTERS. A. Anterior end of blastodermic vesicle. 4ZB. Albuminous layer acquired in Fallopian tube. C. BL. Cavity of blastodermic vesicle. HM, D. Embryonic disc. #P.J. Inner layer of epiblast. 2#P.O. Outer layer of epiblast. EP. OR. Cell of outer layer of epiblast becoming “accidentally ” included in the inner layer of epiblast. HY. Hypoblast of embryonic disc. AY, J. Hypoblast of region beyond embryonic disc. J. /. Inner mass of cells of blastodermic vesicle. JZ. Larger of the first two segments. JZ?. Supposed second generation of layer of first two segments. JZ*. Supposed third genera- tion of layer of first two segments. O.Z. Outer layer of cells of blastodermic vesicle. P. Posterior end of blastodermic vesicle. 2. B. Polar body acci- dentally enclosed. PZ, 8. Sinuous protoplasmic junction between two hypo- blast cells. S. Smaller of the first two segments. S?. Supposed second generation of smaller of first two segments. §°%. Supposed third generation of smaller of first two segments. 2., 2’. Deceptive appearances suggesting a van Beneden blastopore. Z. Zona radiata. All the figures excepting Figs. 37, 38, and 41 have been drawn with the help of a camera. All those figures of which the magnification is 465 times were drawn with Powell and Lealand’s 4 apochromatic oil immersion ; the others with Zeiss or Reichert’s lenses. VOL. 37, PART 2,—NEW SER. L 162 RICHARD ASSHETON. PLATE 13. Fic. 1.—Fertilised ovum. Rabbit 243 hours. x 165. Fie. 2.—Ovum in two segments, from same rabbit as above. 243 hours. x 165. Fic. 3.—Ovum in two segments, showing great difference in size, from same rabbit as above. 243 hours. xX 165. Fic. 4.—Ovum in two segments, from rabbit 24 hours. Polar bodies sepa- rated. x 165. Fie. 5.—Ovum in two segments, from rabbit 253 hours; drawn after mount- ing in Canada balsam. xX 165. Fic. 6.—Embryo in four segments, from same as preceding. 254 hours. x 165. Fic. 7.—Embryo in five segments. 273 hours. xX 165. Fie. 8.—Embryo in eight segments. x 165. Fic. 9.—Embryo in eight segments. x 165. Fic. 10.—Embryo in seven segments. xX 165. Fig. 11.—Embryo in eight segments, showing internally-placed polar body. 89 hours. x 165. Fic. 12.—Embryo of 47th hour, showing contrast in size of the several segments (seventeen segments). X 165. Fic. 13.—Isolated segments of same specimen (47th hour). x 165. Fic. 14.—Embryo (27% hours) in seven segments. PLATE 14. Fie, 15.—Embryo (663 hours), showing great difference in size of seg- ments, Fic. 16.—Section of a specimen preserved in silver nitrate } per cent.; stained picro-carmine, and cut in paraffin, 72 hours. x 465. Fic. 17.—Embryo, showing contrast in size of segments. Fic. 1§.—Section of embryo (72 hours) preserved in Perenyi; stain, borax carmine. Fie. 19.—Another section of the same embryo. Both showing what might be a van Beneden blastopore, but on opposite sides! Fie. 20.—Section through rabbit embryo of the 47th hour. Cut in paraffin. Preserved 4 per cent. silver nitrate 2 minims, water and sunlight + hour; no other stain. xX 465, EARLY STAGES OF DEVELOPMENT OF THE RABBIT. 163 PLATE 15. Fic. 21.—Section through rabbit embryo of the 77th hour. Removed from Fallopian tube. Preserved in Perenyi; stained borax carmine; cut in paraffin. This seems to be an unusually large specimen. x 465. Fie. 22.—Section through rabbit embryo of the 80th hour. Removed from uterus. Preserved in Perenyi, and stained in borax carmine; cut in paraffin. x 465. Fie. 23.—Section through rabbit embryo of the 83rd hour. Removed from uterus. Preserved in Perenyi, and stained in borax carmine; cut in paraffin. xX 465. Fic. 24.—Section through rabbit embryo of the 80th hour. Removed from the same uterus as Fig.22. Preserved in Perenyi, and stained picro-carmine ; cut in paraffin. x 465. Fie. 25.—Section of embryo of rabbit of the 80th hour. Taken from uterus (same as Figs. 22 and 24). Preserved in Perenyi, and stained in borax carmine, and cut in paraffin. x 465. PLATE 16. Fre. 26.—Section of the embryonic disc of rabbit of 96th hour. Taken from the uterus and preserved in Perenyi, and stained in aniline blue black ; cut in ‘paraffin. x 465. Fie. 27.—Section of embryonic disc of rabbit of 96th hour. Taken from the uterus. Preserved in silver nitrate 4 per cent. 25 minims; water and sunlight 2 hours ; alcohol; borax carmine; cut in paraffin. x 465. Fie. 28.—Section of the embryonic dise of a rabbit of the 100th hour. Taken from the uterus, and preserved in Perenyi. Stained in borax carmine, and cut in paraffin. x 465. , Fie, 29.—Section through the embryonic disc of a rabbit embryo of the 103rd hour. Preserved in Perenyi; stained borax carmine. X 465. Fic. 30.—Section through a portion of the embryonic dise of a rabbit embryo of about the 140th hour. Preserved in Perenyi; stained picro- carmine. The hypoblast is now nearly a perfect membrane in the embryonic area. xX 465. Fic. 31.—A portion of the lower pole of the same section, x 465. Fies. 32, 33, and 34.—Portions of sections of the embryonic disc of a rabbit of the 153rd hour. Preserved Flemming strong formula; stained in borax carmine. The fusion of layers is taking place. x 465. 164 RICHARD ASSHETON, Fic. 35.—Portion of a section of the embryonic dise of a rabbit of the 125th hour. The hypoblast here is seen as consisting of cells with very con- spicuous central portion, but these are connected by fine and apparently dis- continuous strands—really a network. x 465. PLATE 17. Fic. 36.—Surface view of a portion of the embryonic area of a rabbit of about the 150th hour. Prepared with nitrate of silver. This corresponds to the sections Figs. 32—34. The shading is diagrammatic, and represents only my interpretation of the areas as defined by the silver nitrate lines, which are drawn by camera. Fic. 37.—A diagram of the embryonic area of the rabbit of the 96th hour. Fie. 38.—A diagram of the embryonic area of the rabbit of the 100th hour. Fic. 39.—A surface view of a portion of the wall of the blastodermic vesicle of the rabbit of the 103rd hour. Seen from within. Preserved Flemming; stained borax carmine. Xx 350. Fic. 40.—A portion of the wall of the vesicle, showing the termination of the inner layer at the 144th—150th hours. x 350. Fic. 41.—A diagram to show the constancy of the angle subtended by the inner mass during the earlier stages of the growth of the blastodermic vesicle. Fie. 42.—Camera drawings of the blastodermic vesicles of a rabbit of the 100th, 125th, 140th, and 175th hours, as seen from the side, arranged so as to show the supposed variation in rate of expansion of the different zones under the influence of—(1) increase of hydrostatic pressure within; (2) resist- ance (mechanical) of cellular wall of vesicle and albuminous layer, which are supposed to be nearly constant factors ; and (3) the vital activity of the cellular wall of the vesicle, which by hypothesis is supposed to vary along certain zones and areas. FUSION OF EPIBLASTIC LAYERS IN RABBIT AND FROG. 165 On the Phenomenon of the Fusion of the Epiblastic Layers in the Rabbit and in the Frog. By Richard Assheton, M.A. With Plate 18. In my paper upon the early stages of the development of the rabbit I have given evidence in support of the views held by Balfour and Heape concerning the fate of the outer layer of epiblast over the embryonic disc of the rabbit embryo of the seventh day. The two layers of epiblast gradually fuse together, and cells from each take part in the formation of the permanent epiblast. In the description I have given of the process, and in the attempt I have made to explain how the fusion is brought about, I have regarded the phenomenon as being entirely accidental and of no morphological importance. It is, however, only right to point out that there is a fusion of two epiblastic layers in certain Amphibians which seems to have a deeper meaning, and to which in some way the condi- tion in the rabbit may be comparable. As far as I know, there is no account published of this phenomenon, and I am not aware that anyone else has noticed it. Therefore, it seems to me, a short account of the facts as they appear in young embryos of Rana temporaria may be of some interest. For the purpose of following the fusion of the two epi- 166 RICHARD ASSHETON. blast layers in the frog, it is best to examine the sections un- stained. In Rana temporaria the epiblast is from a very early period divided into two layers—an outer called the epidermic layer, and an inner called the nervous layer. If a section is taken, say, transversely though the neural plate of a tadpole about the time of the folding up of the neural folds, a section is obtained of which fig. 1 is a drawing. The two layers are seen to be very sharply and distinctly divided, the outer or epidermic layer of epiblast (Z. EP.) is a single cell in thickness. The cells are much more deeply pigmented than are the cells of the inner or nervous layer of epiblast (H. NE.) which forms a much thicker layer. The cells of this layer are very closely packed in the region of the neural plate through which this section is taken; so much so as to render it impossible at this stage, at any rate, to distin- guish the boundary of the cells—if, indeed, there are any distinct boundaries. The nuclei are large, but are only seen with difficulty without staining. Some indication of the boundaries of the cells is to be seen in the slight increase of pigment along certain lines. In fig. 2, which is taken at a slightly later stage, but before the neural folds have completely closed, the cells of the epidermic layer may be seen to have become much elongated, their inner borders being no longer truncated, but mostly pointed, and seem to be growing into the mass of nervous epiblast. In fig. 3, which is from a section of a tadpole of about 3 mm. to 34 mm., in which the neural tube is now completely closed and separated off from the skin, the appearance of the darkly pigmented cells is very remarkable and instructive. I can see no reason to doubt that the darkly pigmented cells of the epidermic epiblast, seen to be elongated in fig. 3, have by this time elongated and passed right through the mass of nervous cells (or nuclei) and spread out into fine filaments on the further side of the nervous layer. I cannot say for certain whether these fine filaments anastomose or not, but, however FUSION OF EPIBLASTIC LAYERS IN RABBIT AND FROG. 167 that may be, they seem to form that which may, as a whole, be regarded as a reticulum, which is comparable to the myelo- spongium of His. The outlines between the cells of the nervous layer are quite imperceptible in unstained specimens of this stage. The nuclei of the nervous layer are only seen with great difficulty. The boundaries of the epidermic layer cells can in most cases be readily perceived, owing to the deeper pigmentation of the cells of this layer. The processes (PR.) are sharply and clearly defined, and are heavily loaded with pigment. The bodies of the epidermic epiblast cells remain as yet lining the interior of the neural canal, although there is cer- tainly a tendency for the nuclei in some of them to move more inwards. Also there seems to be a tendency for the epidermic cells to become pressed apart by the nervous cells. This is more marked at this stage in the spinal cord than in the brain, as may be seen in fig. 4. I think we may take these appearances as a conclusive proof of the occurrence of an intimate fusion of the two epiblastic layers in the frog. Observe the condition of the auditory vesicle in fig. 3. Here the walls of the vesicle are composed entirely of the light- coloured elements. There is no trace of the dark, deeply- pigmented strands such as are to be seen in the central nervous system. This is a further piece of evidence that the dark strands in fig. 3 are derived from the epidermic layer. For, as is well known, in the frog the nervous layer alone gives rise to the auditory vesicle—so it is a most significant fact that the dark strands should be entirely absent in this case. After this stage (3—5 mm.) it is extremely difficult to trace the fate of the epidermic and nervous cells respectively. Up to now, the fact of the greater pigmentation of the former has rendered the inquiry easy. The origin and meaning of the pigmentation is obscure, but its presence in the frog seems to be due to two causes, separate at any rate chronologically. Firstly. Pigment is present in the unfertilised ovum as a 168 RICHARD ASSHETON. superficial layer covering the upper pole. Hence for a long time we find the superficial layer of cells after segmentation to be more deeply pigmented than the more internally situated segments. Secondly. Pigment seems to be in some way connected with the actual protoplasmic activity, as it appears internally wher- ever division of cells takes place. So also I believe the intensely black appearance of the pro- cesses of the epidermic epiblast which I have been describing is in some way connected with their intense activity just evinced by their growth inwards—which seems to be very rapid, and rather sudden. When at a later period certain groups of nerve cells show a similar intense activity, there is a similar deposition of pigment in and around their processes, as, for instance, in the develop- ment of the ganglion habenula as shown in fig. 5. This pig- ment in each case becomes greatly lessened after the period of intense activity of growth has passed by. The Question of Early Separation into Neuroblastic aud Spongioblastic Elements. Although the evidence to be drawn from the figures accom- panying this paper is far from being conclusive, yet I think it tends very strongly towards the inference that the epidermic layer of epiblast in the frog gives rise to the spongioblastic elements, and the nervous layer to the neuroblasts. His has shown how the spongioblastic network precedes the development of neuroblasts and nerve fibres and forms an irregular network with angular processes by no means unlike the dark strands in my fig. 3. These processes are not at all like the processes of nerve cells, which are always more tapering and less knotty. Further, at this stage there cannot be found any trace of definite nerve-fibres. These do not appear until later, till the tadpole has attained a length of about 6—64 mm. I can only interpret these figures as showing the con- version of the epidermic layer of epiblast into a supporting FUSION OF EPIBLASTIC LAYERS IN RABBIT AND FROG. 169 framework in between the cells of the nervous layer. It is quite possible that we ought to regard the nervous layer of epiblast as comparable to the germinal cells of His rather than to the fully-developed neuroblasts. If this is correct, we then must conclude that in the frog there is a very early separation of the neuroblastic from the spongioblastic elements. Comparison between Rabbit and Frog. Is it then possible that the condition in the rabbit has its parallel here in the frog? It is true that the epiblast is double only over a certain area of the embryo in the rabbit, whereas in the frog it is double throughout. In the frog the nervous layer soon becomes much thickened along the future dorsal surface of the embryo, and over the rest of the embryo the nervous layer becomes reduced to a layer of one cell only in thickness, like the epidermic layer. Now, although it is extremely difficult to trace the history exactly, I am almost sure that the area over which the inner layer of epiblast cells in the rabbit is found, corresponds to that area in the frog over which the nervous epiblast remains thick, or becomes thicker—i. e. to the neural plate. In the frog the whole of the neural plate does not become folded up to form the neural tube, but the outer lateral portions of the anterior part remain outside of the tube, giving rise to the ganglia of the anterior cranial nerves. I have endeavoured elsewhere to bring evidence to show that the epiblastic wall of the anterior part of the body of the rabbit embryo includes more than the double-layered part of the embryo, i.e. more than the so-called “ embryonic disc.” Whether the ‘‘ embryonic disc”’ goes to form the neural tube and ganglia of anterior cranial nerves as in the frog, or whether it forms only the neural tube, I have no evidence to offer. The embryonic disc is precisely in the same position relative to the primitive streak as is the neural plate to the blastopore of the frog. I am well aware that the epiblast is not at first double in all 170 RICHARD ASSHETON. mammals. For instance in the opossum, according to Selenka, it isa single layer. But it also is not always double in the Amphibia. In Triton it is at one time only a single layer of one cell in thickness. Why there should be this early differentia- tion into spongioblastic and neuroblastic elements in one and not in another so comparatively closely allied animals it is not easy to guess. Possibly it may be that, since the spongio- blastic elements of the nervous system are the first to show activity of growth in the nervous system (His, and above), then in those animals in which, owing to various individual causes, the epiblast is many-layered, the outermost layer of “cells” or centres of activity being, by reason of its external position more favorable to processes of respiration and so in a con- dition more favorable to active growth, it will be this layer of epiblast that will take on itself the earliest phase in the further development of the nervous system. Although I think the rabbit’s condition can be quite well explained without reference to the above, on the other hand there may be some deeper meaning in it, for which reason I have thought it best to make notice of the condition in the Anura. To make this parallelism complete and certain it is necessary to show that in the rabbit the cells derived from the outer layer give rise to the spongioblastic tissue, and those derived from the large inner layer cells to the neuroblastic tissue. - This I cannot do. After a while the cells which have origi- nated in each of the layers become equally active, but I have as yet been unable to trace their fate respectively. FUSION OF EPIBLASTIO LAYERS IN RABBIT AND FROG. 171 DESCRIPTION OF PLATE 18, Illustrating Mr. Richard Assheton’s paper, “On the Phe- nomenon of the Fusion of the Epiblastic Layers in the Rabbit and in the Frog.” List oF REFERENCE LETTERS. AU. Auditory vesicle. 2. NH. Nervous epiblast. #. HP. Epidermic epiblast. M#. Neuroblast. PJ. Stalk of pineal gland. PR. Processes of epidermic layer cell. 7. #. Band of nerve-fibres passing from ganglion habenula to the ventral side of the brain. Fig. 1.—Transverse section through a portion of the neural plate of a frog embryo (Rana temporaria) during the process of folding up of the neural plate. Unstained. x 165. Fie. 2.—Transverse section through corresponding region, but at a slightly later period, of a frog’s embryo. Unstained. x 165. Fig. 3.—Transverse section through the corresponding region after the separation of the neural tube from the skin. Tadpole, 33 mm. Unstained. x 165. Fic. 4.—Transverse section through the spinal cord of the same specimen as Fig. 3. Unstained. x 165. Fic. 5.—Transverse section through ganglion habenula of a frog tadpole of 13 mm., showing deeply pigmented neuroblastic processes. Aniline blue- black. x 350. il 2S ‘er had io WYRE 9.0% co. oerae vo wahey " palin - li eo S - ie 2 » a : an i 7 en Ba, cn mali . 7 — 7 : . Ss are ve 7 ‘1! Bit ik “* ere Morro and ae = i. Oe E 4 eg 'F ale i @ el seer re ey h ti Lose yi +1 | i yo ae ; = awa Ati sala eu widen! ee P| Aid 7 ' Hwisad slit 73 Liat} ‘Sait 1 tee hie alt UT Bre idaast —s i ; Tr) see) ter Gant : * Or iatin lf a a | snes tiie NAO ive? yonOtn a > joa bheviayd SL? Josie bettie Tr see iviGina Gwe iy ed, Moule, ao ee tide hy hate A Ua qa shoe 7 rite) Gi) 10 Ohie [earee Ag ey 7 path So Tal Ler ull ii add ¥ dguvndl idtinde dow Teele Aislin oy uitinlyl iv Gobel) MOA T4108) 4 ehhh 2 [= dle t 3 Toile a de ind Bian mk) ir eri thy oma glk ot win wn Devedh: ae 08) Ko della wall 2 creyee eli) Sug) OP EG : Add unith ey Posty witmoueriay wih wis Stet) Cuil, a oie A eh ey; ; AL Wasa’ i agra at Anil es mi) merry! Lifiss faipud- ad! woe 7 : «palate mnimeaes Yi Sor bax ha ie eget: Gulove 4 Konan 5 ; fis © ; eet se Se ee i aa : me nd Spar a a Miyareaiiy rtiane® Hane” Pati nailin teat g : 5 - odiig tt Hiaahia(s eee qe xd bit Bit entice “dl es Boer in . _ | peu ag %~ ' _— - * 4 - rt neal > ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 173 On the Causes which lead to the Attachment of the Mammalian Embryo to the Walls of the Uterus. By Richard Assheton, M.A. With Plate 19, I. Tue First ATTACHMENT OF THE RaAspsit EMBRYO TO THE WALLS OF THE UTERUS. By the end of the eighth day, if not actually attached to the walls of the uterus, the embryos have become definitely located, and their presence is made evident from the exterior of the uterus by a bulging in the wall opposite the mesometrium, Frequently by this age they are actually attached to the uterus, and cannot easily be extracted without damage. It must also be noted that the embryos by this time no longer lie anyhow in the uterus, but when definitely located the position of the embryo to the uterus is such that the embryonic area is always towards the mesometric wall of the uterus. This is a very important fact. It is probably necessary for the development of the rabbit that it should be thus situated. It would be very awkward if the embryo became fixed in any other position. The shape of the uterus at this stage, and the shape of the blastodermic vesicle at this stage, are so beauti- fully adapted one to the other as to render any other position almost impossible. The blastodermic vesicle by this time is a slightly elongated body whose lower pole is semicircular in transverse section, while the upper pole is much flattened. If 174 RICHARD ASSHETON. a piece of uterus is blown out with water, then hardened, and a transverse section cut, the cavity will be seen to be bounded by a semicircular wall on the abmesometrial side, and, by reason of the two largely developed “ placental” lobes, a very flattened wall on the mesometrial side. Fig. 1, Pl. 19, is a figure of the section of the uterus during the early days of pregnancy, or in such part of the uterus in which no embryo is present during rather a later stage in the earlier part of pregnancy. JM. is the mesometrium; C. is the cavity of the uterus. The general external outline of the uterus is circular; so is the inner muscular system, but the latter is eccentrically placed to the outline of the uterus. Within these muscular coats, which form the most resisting part of the uterus, comes the soft connective tissue and epi- thelial lining of the uterus. The loose connective-tissue layer seems to be the most yielding, and, as a consequence of the blastodermic vesicle within the cavity of the uterus, the folds or lobes, as they appear in transverse section (PL. L., PP. L., OP. L.), diminish in size, and as the body within increases in size, one by one they begin to disappear. The obplacental lobes have quite disappeared by the middle of the eighth day, and the periplacental lobes can no longer be described as lobes or folds after the beginning of the ninth day. The placental folds being the largest, and also upon the less expansible side of the uterus, become much flattened but never entirely dis- appear. About the time that the blastodermic vesicle becomes de- finitely located, its shape is in transverse section as fig. 4. The lining of the uterus is seen in fig. 1 to be thrown into folds. When the blastodermic vesicle (fig. 4) is inserted into the cavity, the folds become pressed out to a greater or less extent : thus the folds called obplacental are entirely obliterated (OP. L., fig. 2), the periplacental are nearly obliterated, but the placental lobes being very much larger, and also being supported by a larger mass of muscular tissue, are hardly com- pressed at all. ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 175 The cavity of the uterus is thus bounded on the mesometrial side by an almost straight line, while the opposite wall is circular. A body having the shape shown in fig. 4, in passing down the tube by virtue of the continual slow contractions of its walls, could hardly fail to become so fitted as to lie with its flatter surface against the flatter wall of the uterus. Since the flatness of one surface of the blastodermic vesicle is caused by the embryonic dise and changes connected there- with being upon that surface, it follows that in the rabbit the embryo always comes to lie up against the mesometrial side of the uterus, which is, under the circumstances, by far the most favorable position for its future development. For in this position it will be less exposed to the tension which must necessarily arise upon the opposite side in the subsequent expansion of the uterus by the accumulation of fluid within the blastodermic vesicle. The blastodermic vesicle does not by any means become attached with the centre of its flat surface always exactly adjoining the median cleft between the two placental lobes. It may sometimes be a little to one side or the other, but it is never very far out. As the blastodermic vesicle expands, it eventually fills up the whole cavity of that section of the uterus in which it happens to be. With further expansion it necessarily exerts a pressure upon the walls of the uterus, and this pressure is made apparent without by a swelling, or protuberance, occurring upon the side of the uterus away from the mesometrium. The first visible sign of the commencement of the formation of the placenta—or, at any rate, of those structures which eventually cause the placenta to come into existence, appears very shortly after the appearance of this swelling. I allude to the papille and protuberances of epiblast. I said just now the first visible sign because I believe that the outgrowths are due really to a continuation of those causes which have given to the blastodermic vesicle its present form and shape, and described in a former paper. The immediate cause of the origin of the papille seems to be 176 RICHARD ASSHETON. the pressure which now is applied to the vesicle from without by reason of the resistance offered by the elasticity of the walls of the uterus to the internal hydrostatic pressure. First of all, let us consider what is the nature of these papilla-like growths. Sometimes during the eighth day, it may certainly be as early as the seventh day four hours, when the uterus is swollen very considerably by the pressure of the contained blastodermic vesicle, here and there it may be noticed in transverse sec- tion that over all the lower surface of the vesicle certain of the epiblast cells are no longer so much flattened, but the nuclei appear rounded instead of oval in section, and the proto- plasmic part of the cell much more distinct and granular— altogether more comfortable-looking (v. a., fig. 5). A little further along, at 6., may be seen a couple of such nuclei ina mass of granular protoplasm. At c. a group of three or four, or more, of such nuclei in a mass of granular protoplasm. Outside this may be seen the torn remnants of the much attenuated albumen layer. The portion of the walls of the vesicle here figured was part of the lower pole of the blasto- dermic vesicle of an embryo of seven days four hours. Al- though I did not succeed in taking it out of its place in the uterus unhurt, it was, nevertheless, not yet attached to the walls of the uterus at any point. It may be noted that here there is no trace of an inner layer or hypoblast. In fig. 6, another piece of the wall of a vesicle of about the same age cut in situ in the uterus, this piece shows a por- tion of the side of the blastodermic vesicle, or, at any rate, a portion not so far removed from the embryonic pole as that drawn in fig. 5. Here in fig. 6 the same features are to be seen in the epiblast as described for fig.5. There is here, how- ever, a well-developed hypoblastic layer. Fig. 8 is a section of the upper part of the wall of the vesicle, of that part which closely adjoins the embryonic disc. This specimen is from an embryo from the same rabbit as that from which fig. 5 was drawn, and almost exactly the same size. Here it will be noticed that not only are the cells ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 177 here and there large and granular with round nuclei, but the whole of the epiblastic layer is now composed of thickened cells, almost columnar with rounded nuclei, and for a given length there are many more nuclei than there were at that region during the sixth and seventh days. In fig. 9a piece of the same region a few hours later is shown, and here the epiblast is seen to be so thickened as to be actually several cells thick,—in fact, a proliferation of cells is taking place out- wards. Figs. 10 and 11 are still later stages of the same area. I believe that to understand the first outgrowth of these papille, both the small ones scattered over the lower pole, and the area nearer to the embryonic disc, the great change that the embryo has now undergone in the physical and mechanical conditions must be taken into careful consideration. During the fifth, sixth, and seventh days I described in my previous paper the growth of the walls of the vesicle as being due to the hydrostatic pressure within it, together with the multiplication of the cells of the walls of the vesicle, support being rendered to the delicate wall of the vesicle by the albumen layer. The cells of the wall (the epiblast) are very much flattened because of the tension produced by the hydrostatic pressure. The hydrostatic pressure is sufficient to keep the cells always taut, and any increase in size of a cell owing to growth, or any aggregation of cells by multiplication, is prevented by their being flattened out by the internal pressure as soon as formed. Thus all the cells are uniform in thickness, and extra activity of one part shows itself by extra expansion of that are of the vesicle. What happens, however, when the walls of the uterus, by reason of the great size now attained by the blastodermic vesicle, affords supports to the walls which was hitherto want- ing? It must decrease the ratio of hydrostatic pressure to the rate of growth of the cell wall. If the amount that the cells are stretched is constant when the hydrostatic pressure is to the rate of growth of the cell VoL. 37, PART 2.—NEW SER. M 178 RICHARD ASSHETON. walls as, say, 10 is to 10, then when the hydrostatic pressure (P) is to the rate of growth (R) of the cell walls as say 8: 10, it follows that the cells will not be so much stretched as when Rad: sas 0. The actual pressure within the vesicle no doubt does not diminish, more probably it increases, but the relation of P: R is altered by the fact that additional resistance is now added without by the walls of the uterus. At any rate, it upsets the ratio formerly existing. The result, I believe, is that now the rate of growth of the cells of the wall is as compared with the rate of increase in size of the blastodermic vesicle greater than it was before, so that when a cell ‘‘ grows” and “ divides ” it no longer becomes at once stretched out, but forms a rounded granular cell, or group of cells, as in figs. 5 and 6. The cells which are for the time inactive will remain flattened, for elasticity is not an attribute of protoplasm. This is the case at the lower pole of the blastodermic vesicle and at the lower sides. In the region near the embryonic disc (where it has all along been assumed that there is a more active growth) it is found that almost every cell has evinced signs of activity, for here the whole area has become thickened, and the cells far more closely packed than they were, and almost columnar instead of being flattened (fig. 8). It is, I believe, usual to describe the first attachment as occurring between the ectoplacenta of the embryo and the placental lobe of the uterus. This I am convinced is not an accurate statement. The first actual attachment is between the lower parts of the blastodermic vesicle and the periplacental and obplacental folds. The exact course of the procedure I am doubtful about, but I believe it to be a combination of at least two main causes, but it may involve more; or possibly it is wholly due to only one of the two I am about to mention. The first attachment is effected by means of the “ papille,’’ or thickened spots of epiblast of the lower pole of the embryo already described (figs. 5 and 6, a., b.,¢.). In fig. 7 one of ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 179 these thickened spots is shown to be on the point of effecting an attachment. It may be noticed to be wedge-shaped in section; it is a little blunt cone. Practically each papilla assumes this shape, and is being pressed against the epithelium lining the uterus. In this figure (and almost any number might be drawn showing the same characters) the papilla certainly looks as though it were piercing the epithelium by reason of the pressure from within the vesicle. Of course the actual hydrostatic pressure will be the same at a, as it is at e., but nevertheless a greater pressure will be exerted on the uterine wall at the apices of any knobs on the wall of the vesicle than at the area between them if the uterine wall is in a state of tension, which undoubtedly it is at this time. If we consider that in all probability the uterine epi- thelium is a softer material than the muscular and con- nective tissue outside it, it is all the more probable that the softer (if it is softer) uterine epithelium will give way between the two. I think it quite possible that this may be the only neces- sary cause, and by this means the “ papilla”? reaches the capillary system of blood-vessels in the uterine connective tissue, a point of the utmost importance to the welfare of the embryo. On the other hand, although I have no doubt that the additional pressure which exists at the points of these knobs is of much importance in that it causes very close contact between the uterine epithelium and parts of the wall of the blastodermic vesicle, yet it is more than possible that the breaking down of the uterine epithelium at those points is not due to mechanical pressure alone, but to a chemical or a physiological process, such as absorption of the uterine cells by the vital activity of the cells of the knobs. But the possi- bility should not, I think, be lost sight of that the first breaking through of the uterine epithelium at these points may be entirely due to mechanical pressure alone. A point to which special attention must here be drawn is 180 RICHARD ASSHETON. that the moment after the uterine wall comes to take the place of the albumen layer as a support, then the greater part of the expansion of the combined vesicles must be that part furthest removed from the embryo, because it is here that the uterus wall is infinitely less resistent, and with it will go that part of the vesicle to which it is attached. Onthe Importance of the Albumen Layer and Zona radiata. Those who have followed my account of the development of the blastodermic vesicle of rabbit between the eightieth hour and the time of attachment of the vesicle to the walls of the uterus at about the 170th hour, must have seen how important a part the albumen layer plays in producing the form actually assumed by the vesicle. I pointed out how that the thin cellular wall of the blastodermic vesiele would be itself quite able to withstand the hydrostatic pressure within ; and, further, that it is not until the vesicle has attained such a size as to stretch the walls of the uterus as to cause them to afford the support necessary to prevent the bursting of the vesicle, that the albu- men layer was lost. By this time the albumen layer has become exceedingly thin, and the disappearance of it is brought about by its rupture. Traces of it may be found curled up and crumpled several days later. Once burst, its function, as far as I can judge, is at an end. Now it has almost certainly at least one other very important influence upon the development of the rabbit. By its presence until the eighth day it absolutely prevents the cellular tissue of the blastodermic vesicle from coming into close contact with the cellular tissue of the uterus. The embryo up to the eighth day is as free probably from the protoplasmic influence of the mother, as is the egg of a bird after it has been covered with.a thick calcareous shell. There is also, it will be remembered, another coat to the ovum, the zona radiata, which is present from the very first, and has a similar effect to that which the albumen layer has, but being less thick its effect is more evanes- cent. The two coats may be considered as one. ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 181 As regards the rabbit. Ihave shown in considerable detail how many events are, or at least may be, explained by external influences (such as albumen layer, hydrostatic pressure, pressure of uterus, rupture of albumen layer) acting in conjunction with a simple steady force or energy, the primary centre of cell mul- tiplication. In doing so, it will be remembered that in connection with the actual forms assumed, and phases passed through, in the rabbit, the presence of the albumen layer, the size of the cavity of the uterus, and even the shape of the walls of the uterus, had important consequences ascribed to them. If my line of argument has been correct, the effects produced by the albumen layer and the size and shape of the uterus must be very different or absent altogether in forms in which these conditions are different or absent. As regards the size of the ovum itself, there is very little difference amongst mammals. Let us examine the case of a rat,—a rodent, and so not very distantly removed genetically from a rabbit. And yet how different is the form assumed in the earliest stages of develop- ment! Do the conditions differ from those in the rabbit, and if so, how? They differ in these respects : (i) There is no zona radiata or albumen layer. (ii) The diameter of the uterus is very much smaller, and the lumen proportionately smaller still. (iii) The walls of the uterus are of a more uniform thickness. In the rabbit the zona radiata is so quickly covered by the albumen layer that it cannot be said to have in itself much effect as a support or protective coat. In the rat there is no zona radiata or albumen layer. The ovum develops freely in the cavity of the uterus unprotected by any coat. It would seem to pass very rapidly down the Fallopian tube. Robinson found one early stage of segmenta- tion; from this he figures the mesial section, which shows parts only of twelve segments, so we may conclude that his specimen was a fairly early stage of segmentation—comparable, perhaps, to that of the rabbit of the forty-eighth hour, 182 RICHARD ASSHETON. The segmented ovum of the rabbit lies within a thick, tough spherical coat, and is spherical. The ovum of the rat is not thus placed within a spherical mould, and can therefore take other forms, which from the conditions would seem to be necessarily disc-like or oval. The blastodermic cavity (“ Dottersackhéhle” of Salenka, “ vitelline cavity ” of Robinson) would seem to be produced by a hydrostatic pressure from within, as in the rabbit, but owing to there being no resistance of albumen layer the walls of the cavity of the vesicle offer less resistance. Thus there is less tension, and this has the effect of producing a shape of no symmetry except such as is given to it by the walls of the uterus in which it lies. In the rabbit an early tendency to rapid growth of the area immediately surrounding the embryonic disc gives rise to an expansion of that part of the vesicle; so, also, the same tendency to arapid growth occurs in the rat round the corresponding area, but with very different results. A heap of cells accu- mulates, which gives rise to the irregular mass growing among the cavities of the uterus, and usually called Trager. After a while the blastodermic vesicle becomes fixed in the walls of the uterus, and the condition then is similar to that in the rabbit, i.e. the necessary resistance is supplied, and a more or less spherical swelling appears upon the uterus, in which the embryo develops in safety. As regards the actual fixing of the rat embryo, this is no doubt primarily due to the pressure exerted between the irregularities of the walls of the uterus, the “triger” (and other external layer cells of theembryo), this pressure being brought about by— (i) Multiplication of cells of trager and general growth of embryo. (ii) Hydrostatic pressure within the blastodermic vesicle. Thus we see that what corresponds to the trager in the rabbit becomes obvious in the rat at a very early period; whereas in the rabbit its presence (essentially as an area of increased energy) is marked by the continued expansion of the vesicle, as described in a former paper, ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 183 Then, again, let us look at the conditions of the mole. Here is an animal, an insectivore, which is considerably re- moved genetically from a rabbit. Yet here we find that the development is extremely like that of the rabbit. It is true that there is a slight approach to the inversion condition, but only slight; and I think that what appears to be an inversion for a short time is very possibly in no way connected with the inversion such as that of the house mouse, field mouse, rat, or guinea-pig. The separation figured by Heape (‘ Quart. Journ. Micr. Sci.,’ 1883, Pl. 29, figs. 24 and 25) seems to be due to quite other causes, namely, temporary and rather sudden acceleration of growth of the embryonic disc itself, and not of the area just beyond it. For the cells which become separated from the epiblast, and which Heape compares with the trager, afterwards again unite and fuse with the epiblast and form together the epiblast of the embryonic disc; whereas in all other forms any cells which have once become separated as “trager”’ never participate in the formation of the permanent epiblast. I should say these cells correspond rather to the outer layer of epiblast in the embryonic disc in the rabbit, and that in the mole the representatives of the trager cells would be those which at this moment (Heape, figs. 24 and 25) form part of the wall of the vesicle beyond the region of the embryonic disc, as in the corresponding region of the rabbit. Until the time of this temporary bending in of the embryonic area, the development of the mole is exactly comparable to that of the rabbit. The ovum segments and forms a spherical morula. A blastodermic cavity and vesicle are formed almost exactly comparable to the rabbit, and the hypoblast is formed in the same way; the blastodermic vesicle is a spherical bladder- like body ; and may not the reason of this be that the ovum develops up to a certain point under quite similar conditions, that is to say, surrounded by a thick protective covering of zona pellucida and ‘‘ mucous layer ” (v. Heape) ? This mucous layer develops rather differently, being applied according to Heape in the uterus, and not in the Fallopian tube. 184 RICHARD ASSHETON. The coat, however, is very much less firm, and offers apparently less resistance than that of the rabbit, for, later, after the vesicle has filled the cavity of the uterus the walls very soon become moulded into the shape of the cavity of the uterus (v. Heape, Pl. 28, figs. 8 and 9). Of course it is necessary to take into consideration the fact that the actual lumen of the uterus in the mole is very much less than in the rabbit, and that therefore the developing vesicle will be influenced by the resistance offered by the muscular coat of the uterus at a time when the vesicle is much smaller than in the case of the rabbit. Now to consider another case where we have pretty complete records, i.e. the guinea-pig. During the early segmentation stage the ovum of the guinea-pig is surrounded by a zona radiata. We find that in the development of the early stages the embryo assumes just the same form as in the mole or rabbit; but after the embryo has reached the uterus the zona is ruptured, and from this moment the embryo is naked and goes through phases which are certainly more like those of the rat than those of the rabbit or mole. It can hardly be necessary to point out that whereas at first the conditions were similar to those of the rabbit or mole, from this moment they resemble more closely those of the rat. In this question I have only gone into any details in the case of the rabbit. But still, although I can only judge of the other forms from either rather superficial personal examination or from the writings of others in which this point has not been given any special prominence, I think I may be allowed to make a few general remarks upon such other forms in which marked divergence of shape and structure occurs in the method of formation of the blastodermic vesicle and germinal layers. Amongst the Carnivora, from observations on the dog by Bischoff and Coste, and on the cat by Fleischmann, it seems clear that the condition most nearly resembles the rabbit. The ova are in all these cases invested with a firm coat, ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 185 either zona pellucida or albumen-layer, and more probably both. These are present until as late a period as in the rabbit. There is no inversion in either case. The shape is somewhat different in both cat and dog. The vesicles are spherical until about the time that the resistance of the uterine walls is encountered, after which time they assume a much more oval shape. Whether this is due chiefly to the effect of the walls of the uterus, or whether due in part toa modification in intensity or of the extent of the trager area, it is impossible to say without careful investigation. Amongst the Ungulates we have descriptions of the deer by Bischoff, the sheep by Bonnet and Coste, the pig by Keibel. Although the literature on the point I am now discussing is very scauty for this group of animals, we know two facts. Firstly, that the vesicle, at first spherical, grows to an enormous length—as much as 180 mm. in the pig (Keibel), or the whole length of the uterus in cases where there is only one embryo. Also we know there is no inversion. They are surrounded by a zona radiata or some other investment, but it would seem to be very thin. Bischoff says there is no albumen-layer in deer. As regards inversion, the Ungulates, as far as at present known, behave like the Carnivora and rabbit. But after a certain point the vesicle lengthens enormously. It seems probable that this lengthening is due partly to the large cavity in which the vesicles lie. All the Ungulates of which we have any record are large. The lumen of the uterus in the pig is enormous as compared with the lumen in the uterus of a rabbit. Also there seems to be evidence to show that the zona radiata or albumen layer is thinner, and offers therefore less resistance, and accordingly has less effect in resisting the pressure of the walls of the uterus, which, it must be remem- bered, actually lie in contact with each other. We must also suppose that although the pressure is sufficient to prevent the vesicle from retaining an approximately spherical form, it is not as yet sufficient to cause a fixation of the em- 186 RICHARD ASSHETON. bryo to the walls of the uterus. What actually brings about the fixation in the embryo in Ungulates I do not know. It would seem possible, from the frequency of a cotyledonary placentation, that a number of spots of pressure arise at considerable intervals amongst the folds of the uterus, and so bring about local areas where the rate of growth of the cellular wall is in excess of the rate of expansion of the elongated vesicle. In connection with the elongation of the blastodermic vesicle, I may draw attention to a fact in the development of the rabbit. It follows from what I have said, that if the albumen-layer in the rabbit was less thick, or absent, its development would be very different. So it happens that after the rupture of the albumen-layer there is a tendency for the vesicle to elongate during the eighth and ninth days. The normal shape of the vesicle upon the ninth day is shown in fig. 3. It is important for the development of the space (C. BL.) that after the rupture of the albumen layer the passages (UT.) should be closed; otherwise, unless the thin wall of the blasto- dermic vesicle is very much stronger at that point than _- hitherto, it could not withstand the pressure from within, which must now be considerable, and would rupture, and the liquid would escape and the swelling (C. BL.) would collapse. In the rabbit, no doubt, the mucous membrane of the approxi- mated walls of the uterus at this point (UT7.) is thrown into folds and pressed together by becoming a little bent inwards, and by this means tends to block the passage, but also, no doubt, the horns of the vesicle (H.) also help to obliterate and plug up the cavity at these points. Not unfrequently it occurs that either one of the horns (or both) extends through this narrow passage (UT.) and is pro- longed (the albumen layer having ruptured by now) for a short distance along the cavity of the uterus. Sometimes a horn seems to be prolonged for a distance of as much as 34 mm., but the greater part of this is non-cellular (Z.), the origin of which I have not traced, ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 187 SUMMARY. Section l. (i) The blastodermic vesicle of the rabbit becomes first attached to the walls of the uterus by its lower pole. (ii) This attachment of the lower pole is regarded as the re- sult of a mechanical pressure of certain spots or knobs of thickened epiblast of the blastodermic vesicle upon the epithe- lium of the uterus (the pressure being the hydrostatic pressure within the vesicle), whereby the uterine epithelium is pierced and the knobs of epiblast become embedded in the connective tissue below. (iii) These knobs of epiblast, as also the thickening along the trager region, are regarded as being the direct result of a destruction of the equilibrium between the rate of increase of the hydrostatic pressure within the blastodermic vesicle and the rate of growth of the cellular wall of the vesicle, under which conditions the epiblast had hitherto remained practically at a constant measure of thickness. The destruction of the equilibrium is brought about by the additional pressure put upon the expanding blastodermic vesicle by the resistance of the uterine walls. This is a con- tinuation of the same series of forces which in my former paper were supposed to account for the peculiar shape of the blastodermic vesicle and apparent growth round of the hypo- blast. (iv) In the trager region the attachment is effected in a rather different manner. Here the activity of the epiblast is greater than at the lower pole, so that here the thickening is of a more general and more rapid character, which results in a somewhat extensive and irregular area instead of isolated knobs of thickened epiblast. Also the soft tissue imme- ‘diately underlying the epithelium of the uterus is here very extensive. Thus the conditions are not such as to cause a perforation of 188 RICHARD ASSHETON. the epithelium. Instead of this, the epiblast of the embryo becomes very thick and amounts to almost a loose proliferation of cells, which cells become eventually pushed into the irregularities and glands of the placental lobes. Section 2. (i) The primary cause of “inversion”? is the fact that the embryonic area is at one time a region of less activity, and is surrounded by a zone of greater activity in connection with the future formation of placenta and a space within which the embryo can develop unaffected by pressure external to itself. (ii) The occurrence of “inversion” is determined by the production of a heaping-up of cells at an early stage by this zone of greater activity and so forcing the embryonic area inwards. (iii) Inversion is prevented by causes which impede the heaping-up of tissues around the embryonic area. (iv) Foremost among these preventing causes is the presence of an investing coat, either zona radiata, or albumen layer, or mucous coat, &c., which— (1) By hindering a close connection between the cells of the blastodermic vesicle and the uterine walls ; (2) By affording a strong support to the walls of the blasto- dermic vesicle— allow the blastodermic vesicle to assume, under the expanding influence of the hydrostatic pressure within, shapes due more to its own inherent tendencies and less to the effect upon it of the pressure of the uterine walls. (v) The thicker and more lasting these coats are, the more marked are the intrinsic characters of the blastodermic vesicle, and the longer deferred is the impression of characters due to the physical effects of the uterine walls. For instance, both the rabbit and the dog have investing coats. In the rabbit it is far thicker, so in the rabbit it is found that the vesicle assumes a shape which can be best accounted for by intrinsic causes, while in the dog, where it is much thinner, the uterus seems to be the more potent ATTACHMENT OF MAMMALIAN EMBRYO TO UTERUS. 189 factor in moulding the shape of the vesicle. In the rabbit the vesicle attains a far greater size than in the dog before becoming attached to the uterus, though this may, no doubt, also be partly owing to the far greater resistance offered by the uterus of the dog. (vi) When the lumen of the uterus is large, and where the investing coat is present, but delicate, the blastodermic vesicle may grow to a great length, and possibly the development of villi and consequent placental attachment may be brought about by local developments of regions of pressure. (vii) Even in the rabbit, after the rupture of the albumen layer the blastodermic vesicle may become extended along the cavity of the uterus to a length of over 20 mm. EXPLANATION OF PLATE 19, Illustrating Mr. Richard Assheton’s paper “On the Causes which lead to the Attachment of the Mammalian Em- bryo to the Walls of the Uterus.” List oF REFERENCE LETTERS. a. Epiblastic papilla. 4ZB. Albumen layer. 4. Epiblastic papilla. Cavity of uterus. c¢. Epiblastic papilla. C. BZ. Cavity of the blastodermic vesicle. e. Wall of the blastodermic vesicle. ZC. Ectoplacental cells. ZZ. External longitudinal muscle layer. HIZB. Embryonicarea. HP. Epiblast. H. Horn of the blastodermic vesicle. HY. Hypoblast. JC. Internal circular layer of muscle-fibres. IM. Mesometrium. MHS. Mesoblast. P. LZ. Placental fold of the uterine mucous membrane. OP. LZ. Obplacental fold of uterus. PP. L. Periplacental fold of uterus. U7. Cavity of the uterus. U7. EP, Epithelium of the uterus. Z. Non-cellular continuation of horn of vesicle attached to the remains of albumen coat. Fic. 1.—Transverse section of the uterus of a rabbit before any distension has been caused by the presence of an embryo. Fic. 2,—Transverse section of the uterus of a rabbit on the seventh day after fertilisation. The distension is chiefly at the expense of the weaker 190 RICHARD ASSHE'TON, wall, i.e. the obplacental folds. At this moment the location of the embryo can only with difficulty be recognised from without. Fie. 3.—A diagram illustrating a longitudinal section of the blastodermic vesicle and uterus of a rabbit about nine days after fertilisation. The epi- blastic papilla and ectoplacental area are omitted. The cavity of the uterus is seen to be obliterated by means of the horns of the vesicle (H.). Fic. 4.—A transverse section of the blastodermic vesicle about the time that it becomes finally located. Fic. 5.—Section of the lower pole of the blastodermic vesicle of a rabbit. Seven days four hours. x 465. Fic. 6.—Section of the side of the blastodermic vesicle of about the same age. x 465. Fie. 7.—Section of the lower pole of the blastodermic vesicle of a rabbit (after attachment) and of the uterine epithelium. One of the wedge-shaped papille is seen to be piercing the epithelium. x 465. Fic. 8.—Section of the ectoplacental area of the blastodermic vesicle of about the same age as that of which Fig. 5 is a section. x 465. Fic. 9.—Section of ectoplacental area later than Fig. 8. x 465. Fic. 10.—Section of ectoplacental area still later. x 465. Fig. 11.—Section of ectoplacental area about the time of its attachment to the placental lobes. x 465. THE PRIMITIVE STREAK OF THE RABBIT. 191 The Primitive Streak of the Rabbit; the Causes which may determine its Shape, and the Part of the Embryo formed by its Activity. By Richard Assheton, M.A. With Plates 20—29. Tue following pages are offered as a contribution towards the elucidation of the mode of formation and the function of the primitive streak in the rabbit. Although it is necessary to make incidental remarks upon the formation of the mesoblast and notochord, I have not given here a full description of the results of my own observations upon those points, but I hope to be able to do so, and to dis- cuss the results of former authors upon this and other animals, inalatercommunication. In the present paper I have confined myremarks almost entirely to the rabbit, and to my ownsugges- tions towards an explanation of certain phenomena connected with the structure we call the primitive streak in that animal. Of recent years the theory of concrescence has been again brought into great prominence in attempts to account for the growth in length of the Vertebrate embryo. The present paper will, I hope, tend to show that there is no trace of such an occurrence in the rabbit, and that the growth in length of the embryo can quite as well—and to my mind infinitely more easily—be accounted for by a process of addi- tion of new cellular units between the pre-existing embryo and an area of rapid cell-production. I have attempted to show which part of the embryo is to be regarded as the “ pre-existing embryo,” and which as being due to the activity of the area of rapid cell-production—the primitive streak. 192 RICHARD ASSHETON. I have also tried to explain how the changes in shape of the primitive streak are brought about. I begin with a short account of this area of activity (which I shall speak of as the secondary area or centre of cell- production—as compared with the process known as the segmentation of the ovum, which I call the primary centre of cell-production) which adds certain details of interest to the accounts already given of the rabbit primitive streak by other authors, notably by Hensen, Kolliker, and Rabl. The Earliest Signs of the Formation of the Secon- dary Area of Proliferation. Until the middle of the seventh day the embryonic disc is circular in outline. After the fusion of the inner and outer layers of epiblast, the outline of the embryonic disc becomes very much sharper. This state continues until about the middle of the seventh day, when, although the anterior periphery of the embryonic disc retains this character, the are of the hindermost quadrant becomes very much less distinct, and is no longer a semicircle. On measuring the embryonic disc very soon after this, it is found to have long and short axes, the long one usually parallel to the long axis of the vesicle, the short to the short axis of the vesicle. The lengthening has been in the antero-posterior direction. The anterior end presents much the same shape and features as before, except that its periphery describes an are of a larger circle than before. It is still sharply defined. The posterior end is, if anything, more ragged, and is undoubtedly at this moment thinner than the rest of the embryonic area. The measurements are as follows: Blastodermic vesicle, long axis. F : 4°] mm. My » short axis . : ; Bu 4; uf » polar axis . : ; 3°86 2s Embryonic area, long axis . ‘ : M357 'g bs ‘ transverse axis . . 126 | This ragged border I believe to be the effect of the setting up of an area of increased activity of cell-production either THE PRIMITIVE STREAK OF THE RABBIT. 193 upon, or just within the posterior border of the previously cir- cular embryonic disc (see PS., Pl. 20, fig. 2). Whether this activity concerns in the first instance both the cells derived from the outer layer of epiblast and those derived from the inner layer of epiblast equally, I cannot form at present a definite opinion. Very soon the centre of this ragged area becomes thickened, and the embryonal area has now the appearance shown in fig. 3. The original outline of the embryonic disc is fairly distinct, but of course slightly larger than in the earlier stage (fig. 1). This thickening, which is due to a greater accumulation of cells at this point, would seem to be the result of a greater activity of cell multiplication at this spot. As to the cause of this sudden increase of activity I have no suggestion to make. It is, of course, quite possible that the activity was present from the first, and that it is only owing to changed conditions that it now becomes evident as a heaping- up of cells. To give a strictly epigenetic explanation some such cause ought to be adduced, but for the present I am not able to suggest any, and so must ascribe it to a palingenetic cause. It is not possible to point out the exact boundaries of this area of increased activity, but probably if that portion of the epiblast which is distinctly thickened is taken to be the area of increased activity, I think the error will not be great. Up to now, and for some considerable time to come, there is no change noticeable in the condition of the hypoblast. The darkened area PS. in fig. 3 may be taken as the area of increased activity. The posterior border of this area is semicircular, the anterior border is conical. The darkest part of this area is about the position of the posterior border of the original embryonic disc. Fig. 4 is a slightly later stage. The anterior portion of the embryonal area still shows a circular darker part, which I take to correspond approximately to the original circular embryonic dise of the sixth day. The secondary area of activity is of the same nature as in the preceding figure, but its anterior conical part is slightly more pronounced. On either side of it there are VOL, 37, PART 2.—NEW SER. N 194 RICHARD ASSHETON. two parts which are very much less opaque, but still more opaque than the extra-embryonic part of the blastodermic vesicle. The specimen of which fig. 4 is a drawing was cut into a series of transverse sections, that of which fig. 4.4 is a drawing, into a series of sagittal sections. Fig. 17, Pl. 20, is a very nearly median sagittal section of the specimen represented by fig. 4 a. In this drawing A. is the anterior end, and P. the posterior. The epiblast (ZP.) of the embryonic disc is thickened throughout. The under or inner surface of the epiblast is even along the anterior half, while along the posterior it presents a very irregular, and towards the extreme posterior end, broken edge. The hypoblast (HY.) is separate all along, and forms a con- tinuous layer. Itis specially thickened at the two points where it underlies the extreme anterior and posterior ends respectively. There are now between the hypoblast and the epiblast, towards the posterior end, certain scattered cells, some apparently en- tirely, others only partially separated from the epiblast. These are the first mesoblast cells which make their appearance. The figs. 13—16 on Pl. 21 are from transverse sections of _ the specimen shown in fig. 4. They are drawn on a larger scale, and, taken together with section 17, demonstrate pretty accurately the structure of the specimens figs. 4 and 4a. Fig. 13 is taken through the anterior part of the specimen along the line 13 in fig. 4. The central portion only is given in fig. 13 (see fig. 13, A.). This section carries out the evidence derived from the longitudinal section (fig. 17), and shows that the under or inner surface of the epiblast is quite smooth ante- riorly. There is no median thickness of the hypoblast, and no sign of mesoblast cells. In fact the anterior region of the embryonal area has not greatly altered from the condition of the circular embryonic disc upon the day previous. The posterior end, however, is very different. Fig. 16 passes through the speci- men (fig. 4) along the line 16—that is to say, through the very densest portion of the area. Here we see the epiblast enormously thickened, and a certain THE PRIMITIVE STREAK OF THE RABBIT. 195 number of cells seem to be lying separate or forming a loose network between the epiblast and hypoblast. These are meso- blast cells. As the sections pass more posteriorly this primitive streak knob becomes smaller and ends rather abruptly, as seen in the sagittal section, fig. 17. But forwards the thickening of the epiblast is continued much further, getting less very slowly, as figs. 14 and 15 show, which are taken along the lines 14 and 15 in fig. 4, This narrow band of thickened epi- blast is seen as the irregular lower surface of the epiblast in fig. 17. It extends very nearly halfway, along the length of the embryonal area. I think it is pretty clear that this is part of the same area of proliferation as that through which fig. 16 passes, as will appear more certainly later on. Hence the most anterior point to which this reaches is a point of very great importance. This is marked A. PS. in figs. 3 and 4. The proliferation of cells is so rapid at the posterior end as to cause an eminence in transverse section (fig. 16; evident also in fig. 17). The three sections, figs. 14, 15, 16, differ only in amount of proliferating area and accumulation of cells proliferated. This I take to mean a difference only in intensity. Excepting in intensity there is no difference sug- gested by the structure of the secondary area of activity in the three sections. The secondary area of proliferation is now fully established, and has changed in shape from an ill-defined spot in fig. 2 to a more and more elongated area in figs. 3 and 4. In fig. 5 it has still further lengthened. Figs. 21, 22, 23 are sections taken across this specimen at corresponding spots to figs. 14, 15, 16 respectively, and must be compared one with the other. It may be noticed that there is here ex- tremely little difference in the extension transversely of the area of proliferation. The cells, which appear to be entirely separated, extend further from the middle line on each side, but the area of fusion itself (between epiblast and mesoblast) differs very little. This is especially noticeable in the most 196 RICHARD ASSHETON. posterior of the sections, namely, figs. 16 and 23. The actual area of proliferation is about the same. In the anterior sections, if there is a difference, it is that in the more advanced specimen (fig. 5) the sections show a smaller extent of proliferating area than in the younger specimen. In specimen fig. 5 there is now a very slight groove along the surface of this area, so slight as to be scarcely recognisable in sections fig. 22 (P. GR.). There is another point of interest. Up to this moment the extreme anterior end of the secondary area of proliferation has almost shaded away imperceptibly into the epiblast of the future neural plate, at any rate it has always been the most slender part of the structure. Now it has become a very well-marked spot—it forms almost a knob—and is henceforth recognisable as what other authors have called “ Hensen’s node.” Fig. 21 passes through the centre of this. Fig. 6 is a later stage still. In this the secondary area of proliferation has attained about its maximum length, and at this moment shows the greatest degree of distinction into different parts: (i) hindermost, a broad area of proliferating surface, its posterior border forming an arc of a circle, while anteriorly it tapers off into (ii) the middle region, a long narrow strip of proliferating area deeply grooved along its length ; more deeply grooved anteriorly and less posteriorly (fig. 26). This again passes abruptly into (iii) the most anterior part of the primitive streak or Hensen’s node, a very much thickened and compressed region of short extent (vide section, fig. 25). This stage is the height of development of the primitive streak. The secondary area of proliferation is at this stage a perfectly typical primitive streak, with well-developed primitive groove. In the preceding figure (fig. 5) the groove is very shallow and wide, and easily recognisable in the whole specimen. In fig. 6 the groove is very deep and narrow, and is only with difficulty seen in surface views as its walls are nearly approxi- mated, and it appears when seen by transmitted light as a THE PRIMITIVE STREAK OF THE RABBIT. 197 narrow dark line instead of a wide lighter line when seen under similar conditions at the earlier stage. In all the stages described hitherto there has been no fusion of the hypoblast with the proliferating epiblast. Im stage fig. 5 at the extreme anterior end of this area, where there is a special thickening of the primitive streak, the hypoblast is in very close contact (see figs. 20 and 21). In this stage (fig. 6) the fusion is complete, and anteriorly the hypoblast in the middle line and the epiblast and mesoblast are continuous in the mass spoken of as Hensen’s node. Figs. 24, 25, 26, and 27 are sections taken along the lines 24, 25, 26, and 27 in fig. 6. A section taken more anterior still than fig. 24 corresponds to figs. 13 and 19. The last three sections (figs. 25, 26, and 27) should be com- pared with figs. 21, 22, and 23, and figs. 14,15, and 16. Here again it will be noticed that there is no great increase or diminution of extent in transverse section of the area of active proliferation. The most posterior parts differ hardly at all except in the spreading of the separated cells (meso- blast). It must be noticed that there is now a region hitherto not to be found. The section drawn (fig. 24) passes through this region, which is just anterior to the primitive streak. The section through the middle region (fig. 26; compared with figs. 22 and 15) differs in one particular only (besides ex- tension of mesoblast). Instead of the area of proliferation appearing at the surface on a level with the rest of the outer layer, it lies at the bottom of a deep groove. The sides of the groove are not “ fused” with mesoblast ; it is only the floor of the groove that is participating in the addition to the meso- blast. Fig. 25, the section through the anterior region of this area of proliferation, shows an intensified form of the state in the preceding stage (fig. 21). Here the groove suddenly ceases, and even in most instances gives rise to an eminence. This eminence so overhangs the groove posteriorly as to give rise to 198 RICHARD ASSHETON. a lumen in a section through its posterior half, as shown at P. GR. in fig. 25. The origin of this groove—the primitive groove—I shall discuss in a later paragraph. The presence of this groove necessitates that the thickened epiblast of the neural plate anteriorly, and the thickened epi- blast of the floor of the groove posteriorly, should be at dif- ferent levels. In this it is possible that any intrinsic growth of the neural plate may tend to produce an overlapping, and so give rise to what has been termed a notochordal canal. The secondary area of proliferation having now attained its maximum length and most complicated form, it henceforth tends to return to its original shape, and becomes less and less like a typical primitive streak. In fig. 7 no very great change has taken place as regards the area in question, with the exception that the primitive groove is not nearly so deep or narrow. It resembles more closely in transverse section the stage of fig. 5. Also the layer of meso- blast cells on either side of the primitive groove are thicker now than in fig. 6. Fig. 8 shows the area still at its greatest length, and figures of sections taken through the anterior end and the middle are given on PL. figs. 80 and 31, and should be compared with figures of the corresponding sections from the earlier stages. The most marked difference is the entire absence of the groove. Compare fig. 31 with fig. 26. The apparent canal into Hensen’s node has gone, and the outline of the embryonal area adjoining this region has changed shape. In fig. 4 the outline is arched away from the primitive streak. In fig. 5 it is less arched, in figs. 6 and 7 it is nearly straight, in fig. 8 it is becoming again curved. From this moment the length of the secondary area of proliferation becomes less. In fig. 10, in which five protovertebre are visible, this area has diminished to about one third of its former length. Figs. 32, 33, and 34 are three sections through such a speci- THE PRIMITIVE STREAK OF THE RABBIT. 199 men. The most posterior, fig. 34, is in no way different from fig. 27, except that the proliferating area may be slightly broader. A similar remark may be offered with regard to the most anterior section, fig. 32, namely, that except the area of pro- liferation may beslightly broader, it very closely resemblesfig. 25. In both figs. 32 and 34 the mesoblastic plates on each side of the fused area are thicker than in figs. 25 and 27. The section that passes through the middle of the length of the primitive streak presents the most noteworthy appearance. In this, fig. 83, the breadth of the area of fusion is very much greater than in any of the specimens previously described. There is no groove; in fact, there is, in the place of a groove, actually aneminence (P. S.). The cells proliferated are much more numerous and more crowded within a given space as compared with fig. 31, and still more so as compared with fig. 26. The total length of this secondary area of proliferation is now only about one third of what it was when at its maximum length as in fig. 6. Thus it has apparently become shorter and thicker. In an older stage, when there are as many as twelve proto- vertebrz formed, the foundations of the most bulky portion of the body have been laid down. The secondary area of pro- liferation is now less conspicuous. Its length is not much more than one fifth, if so much, of its greatest attained length. In stages much later than this it is not possible to observe it in surface views, as such parts as remain are then placed at the extreme tip of the tail, after the complete development of which it disappears. To sum up—the secondary area of activity arises as a small spot excentrically placed to the primary centre of activity. It increases in magnitude, then becomes elongated, and very much reduced in breadth towards the centre of its length, and is deeply grooved. Rather suddenly, after attaining its greatest length, the groove disappears entirely; the area becomes much shortened, and thicker at the spot where it had beenso thin. Now, instead of a groove, there 200 RICHARD ASSHETON. is a ridge along the median line of the area; it still furthershortens and diminishes in size, but does not finally disappear until the last segment has been formed (or until sufficient material has been added to allow of the formation of the requisite number of segments). The above description draws attention to the most note- worthy features of the primitive streak itself, from the moment of its first appearance. At no time during its existence, or before its existence, is there any trace of any form of concrescence such as Duval has described for the Avian primitive streak. Is it possible to account for the changes as seen in the rabbit by a consideration of the conditions under which the blastoderm of the rabbit is placed ? The following pages contain what seems to me a likely ex- planation. The Process of Elongation of the Primitive Streak. In the exercise of its functions the secondary area of activity is essentially centrifugal in its results. While it is, so to speak, spread out flat upon a plane (as during the stages represented in figs. 83—10), the most noticeable feature of its action is the production in every direction of a sheet of mesoblast, of almost circular outline. When after ten days or so it becomes a freely- growing knob, it produces a mass of cells, the outline of which still is circular, the area itself being now also circular in outline. If the nature of the area, as shown by its action and by the form assumed when perfectly free, is to be radially symmetrical, why does it assume a linear form? Why does it become dis- torted? Is it an ontogenetic distortion, or is it connected directly with phylogenetic causes ? I believe the actual elongation, or rather all the changes it undergoes, including the grooving, is due entirely to the ontogenetic influences, and has no recapitulatory meaning. In one sense it may be said to be due to the phylogeny ; but only in that, unless the mammalia had been descended from animals which had large yolked eggs, the ontogenetic development of a THE PRIMITIVE STREAK OF THE RABBIT. 201 rabbit would not, in all probability, have been upon the lines on which it is—provided a rabbit could ever have been evolved under other circumstances. What I mean is: a bird ora reptile has a large egg because the embryo obtains its nourishment from yolk, the presence of which causes the egg to be so large. A mammal does not require the yolk as a nourishment, and therefore the eggis small. A mammal is enabled to develop within the body of its mother because, we presume, originally the embryo was protected within the membranes of a large egg within the oviduct of the mother, and it was only by the substitution, no doubt very gradual, of (i) placental nourish- ment instead of the yolk, and (11) fluids exerting considerable pressure on the uterine walls instead of yolk, albumen, and shell, that it was possible for a mammal to dispense with a large-yolked egg and to be developed under the conditions in which we now find it. Thus it is to the phylogeny that the conditions under which development takes place to-day are due. Now these conditions were imposed upon the development at a compara- tively very recent period in the evolution of the rabbit. So when we find that certain forms assumed by certain centres of activity, which centres of activity themselves undoubtedly date back to infinitely earlier epochs than the date of these superimposed conditions, are due entirely to the con- ditions of this day, we must be very careful in drawing any morphological conclusions from those forms assumed. If, for instance, I can show that the changes in form of the primitive streak from a nearly circular spot to a linear expression, and its groove which appears during part of its existence, are due to these present conditions alone, then to draw any conclusions such as frequently are drawn—as, for example, the theory that the line-like primitive streak and its groove represent the lips and aperture of an elongated ancestral gastrular mouth, is impossible. If the blastodermic vesicle, at the time that the secondary area of activity is established, became an object freely suspended in some fluid, and was under no influence of internal hydro- 202 RICHARD ASSHETON, static pressure, it seems probable that the immediate effect of this rapid proliferation of cells at one spot would be to produce at first a mass of cells or a knob, as in reptiles, and then a wrinkling of the thin wall of the vesicle, and eventually a pro- jection or outgrowth. This is, indeed, what takes place a little later, when, by reason of the attachment to the uterus of the walls of the blastodermic vesicle immediately surrounding the embryo, the circumstances are such as to fulfil these condi- tions of equilibrium, The result at this time is that a definite projection and rolling over is effected—the tail fold. But as yet the blastodermic vesicle is unattached and lies freely within the cavity of the uterus, but is continually ex- panding in size as described before, owing to the combined effect of increasing hydrostatic pressure and multiplication of the cells of its walls. I believe that the conversion of the almost circular spot in fig. 2, or pyriform area in fig. 3, to the linear expression of figs. 6 and 7, and back again to the pyriform area of fig. 10, is inti- mately connected with two facts: (i) Growth of the embryo taking place from different centres. (11) Expansion of the whole area, due in this case (rabbit) to the hydrostatic pressure within the blastodermic vesicle. Ina former paper, “‘ A Re-investigation into the Early Stages of the Development of the Rabbit,” I ascribed the apparent growth round the hypoblast as being due to its being carried round by the outer wall, due to the combined effect of the greater area of special activity around the embryonic disc, aud increasing hydrostatic pressure within. May not a similar explanation be given to account for the apparent growth outwards of the primitive streak mesoblast from the primitive streak? In this case the area of special activity is much more concentrated and vigorous than in the former, and is therefore itself evident as an area of rapid growth. All the cells towards its periphery will be as liable to be removed by the expansion of the walls as the outer layer cells, as with them the inner mass cells towards the periphery are, according to my hypothesis, apparently removed in the former case. THE PRIMITIVE STREAK OF THE RABBIT. 203 Having drawn attention to this possibility, I must leave the consideration of the mesoblast for the present. What are the conditions under which the blastodermic vesicle exists during the lengthening of the primitive streak and the formation of the deep primitive groove ? In the previous paper mentioned above, I have accounted for various changes in shape of the blastodermic vesicle, as well as the apparent growth of one layer over another, by pointing out how these changes could be produced by increasing hydrostatic pressure within, together with more rapid cell-division in one part, and less rapid cell-division in another part. So, I believe, the same principle holds good with regard to the lengthening of the secondary area of proliferation into a streak, and to the production of the median groove. The blastodermic vesicle is still expanding rapidly by reason of the increasing hydrostatic pressure. The walls of the vesicle are thin except at one point, the embryonic disc, where the wall is thick and compact. Somewhat suddenly, a spot at the posterior edge of this embryonic disc becomes extremely active—the secondary area of proliferation. This gives rise, as we have seen, to a mass of cells produced very rapidly at one spot (fig. 3), forming a second compact disc. We have now two very distinct compact masses in the outer layer or epiblast, one formed by the presence of the inner layer of epiblast and its fusion with the outer layer, the other by the very rapid proliferation at one spot caused by the appear- ance of the secondary growing point or primitive streak. While this is proceeding, the whole blastodermic vesicle is still expanding from the hydrostatic pressure within. Now what effect, if any, will this hydrostatic pressure have in regard to these two compact masses? In the first place there will be a tendency for them to part. There will be a tendency to part from another cause, namely, the counteracting effect of their re- spective growths ; but this would tend to produce a space (i. e. mass of cells) between the primary growing point and the ante- rior border of the secondary growing point (i.e. anterior end 204 RICHARD ASSHETON. of primitive streak). This takes place, no doubt, and isa very conspicuous feature at a later stage, as described later, but it does not account for a lengthening of, and within, the secondary area of proliferation itself. The lengthening of the, at first, circular patch into a streak, i.e. the change from fig. 11 to fig. 12, I believe to be due almost entirely to the expansion of the blastodermic vesicle by the hydrostatic pressure within. The expansion of the vesicle is quite sufficient to allow of this. It is not easy to make very satisfactory measurements in support of or against this view. If we take a specimen about the stage shown in fig. 3, we find the secondary area to measure not less than 38 mm. antero-posteriorly, and the diameter of the whole vesicle about 4°5 mm. to 5 mm. If we take another specimen in which the secondary area has attained its maximum elongation, we find the area to measure about 1:08, which is rather less than three times the length it was in stage fig. 3. The diameter of such a blastodermic vesicle is about 10 mm. to 12 mm., which is from rather under two and a half to rather over two and a half times the length it was in stage fig. 3. I may, perhaps, make my meaning clearer by reference to figs. 11, 12, Pl. 20. In fig. 11 the point M is the centre of the primary area of cell-production, i.e. the centre of the embryonic disc. The point N is the centre of the secondary area of cell-production. The grey represents the circular embryonic disc; the white represents the proliferating area of the primitive streak. Really this diagram represents no actual stage. Fig. 3 is very near it. If my account of the course of events is correct, there never could be a stage which would be exactly repre- sented by this diagram, although it theoretically represents the condition. Now the tendency of each of these two areas considered separately would be to expand equally in all directions, from the centre Min the one case, and Nin the other. But these two areas are not separate, they even overlap, and therefore THE PRIMITIVE STREAK OF THE RABBIT. 205 they must have some disturbing effect upon each other; and the line along which that disturbance will be most marked will be along a line drawn from the centre of the one to the centre of the other—from M to N. I have argued before that while the increasing hydrostatic pressure within the blastodermic vesicle causes its expansion, yet a more rapid expansion of one part of the wall than another may be brought about by the more rapid cellular growth of that part of the wall. In other words, areas of weakness are produced upon which the hydrostatic pressure takes more effect than elsewhere. So at this stage must there not be a line of weakness between the two points M and N in diagram 11, an area along, at any rate, that portion in which the influences of both centres are at work, in which there will be less resistance offered to the ex- pansive force of the hydrostatic pressure than elsewhere? It seems to me that this is extremely likely, and that a figure such as that represented in diagram fig. 12 must be the result. The anterior half or portion of the secondary area of pro- liferation will be drawn out as indicated in the diagram by the separation of the black dots; the posterior portion, being unin- fluenced by the primary area, will not lose its radial symmetry. Similarly the embryonic disc will be influenced, and its hinder portion drawn out as indicated by the black crosses on the two diagrams, its anterior border not losing its radial sym- metry. It must be remembered, however, that the secondary area of proliferation is in its function of cell-production more vigorous now than the primary area, and becomes still more vigorous for the next two days. Also the visible line of the primitive streak does not even at this comparatively early stage represent the whole of the effects of the secondary area of proliferation. No doubt the epiblast immediately surrounding it owes its origin to the energy of the area in question, but to what extent it is not easy to determine. At any rate it is probably not less than that part beneath which primitive streak mesoblast lies. So that a more accurate figure of the conditions of the two areas of activity respectively 206 RICHARD ASSHETON. would probably be such as I have drawn in fig. 43, on Plate 22, to be explained presently. I conclude, therefore, that the lengthening of the primitive streak is due to the expansion in an antero-posterior direction of the most anterior part (half?) of the secondary area of cell-proliferation due to the conditions under which the embryo is developing. May not the primitive groove be also due to ontogenetic conditions ? It will be noticed that it does not exist during the early stages of the formation of the primitive streak (vide figs. 1— 4a, and the sections of these). It will be noticed also that it is at its greatest development at the time the primitive streak reaches its greatest length, and from that moment it becomes shallower and very rapidly dis- appears altogether (vide fig. 8 and sections). It is deepest and most distinct along that part of the streak which is thinnest, and gradually shallows and disappears towards the thicker posterior end of the streak. The sections through the stages in figs. 4, 5, and 6 show that the groove commences at the same time as the spreading of the mesoblast. The mesoblast is a reticulum connected with the area of pro- liferation, and is seen to spread out from both sides of the drawn-out proliferating area. I have explained above how this spreading out of the meso- blast may be due to the general expansion of the walls of the vesicle of this region by the increasing hydrostatic pressure. If the proliferating area produces cells more quickly than they can be removed laterally by the expanding walls of the vesicle, a heap of cells will be the result at this spot, or if the prolife- rating area is linear a ridge will be formed. If the cells pro- duced exactly balance those removed, the surface will remain flat ; there will be no ridge. But what will happenif the cells produced along the prolife- rating area are removed more rapidly than they are produced ? If the deficiency of production is very great, then probably the cells—few in number—will be torn apart and removed as THE PRIMITIVE STREAK OF THE RABBIT. 207 isolated cells. But if the deficiency is only slight, surely the result will be that the meshes of the reticulum of mesoblast will be drawn out into larger meshes, and the connecting filaments become finer. It seems to me that this must mean that a tension between cell and cell must be produced, and not only between cell and cell of the reticulum, but between the reticulum on each side of the primitive streak and the actual proliferating area itself. Is it not possible for this tension to produce such a groove as that shown in figs. 22 and 26? Posteriorly there is no groove. Here the supply of cells is greater than the removal, and a heap is formed ; but along the narrowest part of the proliferating area the conditions favour the formation of a groove as I have suggested, for not only is the proliferating area very attenuated, but lateral dragging of the mesoblast is in the same direction along a considerable length. It follows that upon the above explanation of the lengthening of the primitive streak and formation of the groove any increase in activity of the proliferating area might lead to the oblitera- tion of the groove, and even perhaps a cessation of the increase in length of the streak. Also any modification which either prevented the increase of hydrostatic pressure within the vessels, or which prevented the portion of the wall containing the embryonal area from being affected by the increase of hydrostatic pressure, would bring about the obliteration of the groove and cause the cessation of lengthening of the streak. In the rabbit it seems probable that the latter event occurs. During the stages represented by figs. 1 to 5, that is during the formation of the greater length of the primitive streak, the blastodermic vesicle lies freely in the uterus. About the age represented by fig. 7 the papille are appearing by which the lower pole becomes attached to the obplacental portion of the uterus. The ectoplacental area is thickening, but is still quite smooth and quite free to slide over the surface of the placental lobes; but between that time and the time represented by fig. 8, the ectoplacenta has become irregular, the albumen 208 RICHARD ASSHETON. layer has ruptured, and the ectoplacental region is firmly attached to the placental lobes. From this moment the upper pole of the blastodermic vesicle can only be expanded by the hydrostatic pressure as much as the placental lobes and the thick mesometrial wall of the uterus will allow it. As is well known, this portion of the uterus expands now very little and slowly; by far the greater part of the swelling of the blastodermic vesicle concerns only the ab-embryonic pole of the vesicle and the obplacental lobe of the uterus. In this way a very large part, if not all, the tension is rather suddenly removed from the actual embryonal area and immediately surrounding tissues, and this is exactly coincident with the attainment of the maximum length of the primitive streak and with the rather sudden obliteration of the primitive groove. No doubt to all appearances the primitive streak of a bird is at the time of its greatest development extremely like the primitive streak of a mammal of the type of development such as we find in the rabbit. If we accept Duval’s description for the bird, and also the explanation I have offered above for the rabbit, as being both correct, we are bound to conclude that it is really only a coin- cidence that the secondary area of cell-production in each case should assume a linear form. Such a coincidence is unlikely, though of course not impossible. An Attempt to determine which Portion of the Embryo is derived from the Cells proliferated from the Primitive Streak. Those who have followed my description of figures will have noticed that, according to the account, there is one part.of the embryo laid down as the immediate result of the segmentation of the ovum, and that the most conspicuous part of this is a circular patch—the embryonic disc. Subsequently a renewed activity of cell-production takes place upon the posterior border of this embryonic disc, giving rise to new tissue, and THE PRIMITIVE STREAK OF THE RABBIT. 209 continues active till the formation of the extreme end of the tail is completed. Is it possible to determine approximately which parts of the embryo are formed by the two centres of growth respectively? Evidence may, I think, be brought to show that the primitive streak in the rabbit is the growing point of the whole of that part of the animal which is situate posterior to the head, while from the primary centre of activity is formed the part of the embryo anterior to the first protovertebra. In other words, the secondary centre of growth is responsible for the metamerically segmented part of the animal. During the later stages of development the line of demarca- tion between the two—which perhaps is never absolutely definable—becomes less and less distinct. For instance, the heart, which is formed distinctly in the region of the primary centre of growth, becomes located more posteriorly; while parts of the nervous system, owing to the more rapid growth of the neural tube, become moved forwards, so as to bring into the primary region portions whose origin has been due to the secondary area of activity. If we regard the secondary area of proliferation or primitive streak simply in its functional capacity, neglecting for the moment all preconceived ideas of its morphological or recapi- tulatory meaning, and consider it only as we find it in the rabbit embryo, it seems to me that we are bound to describe it as the visible expression of an area of intense protoplasmic activity, which is continuously budding off cells, the most con- spicuous of which are those from its lower surface, the primi- tive streak mesoblast cells. These cells, according to my hypothesis, are rapidly removed from this area, being carried away by the expanding wall of the blastodermic vesicle, to which they are closely approximated. This part of the wall is enabled to respond to the expanding influence of the hydro- static pressure within, by itself receiving rapid additions of cells from the proliferating area. On this hypothesis we may take the outline of the primitive streak mesoblast as indicating also the outline of epiblast VOL. 37, PART 2,——NEW SER. ) 210 RICHARD ASSHETON. derived from the primitive streak, at any rate for the stages up to about the one drawn for fig. 6. At this time mesoblast cells are separated from the hypoblast, both at points where there is already existing primitive streak mesoblast, and at points where up to now there has been no mesoblast, e. g. under the part of the embryonal area in front of the primitive streak. After this stage it is extremely difficult to determine in many places what the origin of the mesoblast cells has been. As, however, this involves the question of the formation of the whole mesoblast, I must leave the further discussion for another paper. I must now refer again to the figures of the embryonal area (figs. 3 to6). I have given above my explanation of the change in shape, and of the process of elongation of the proliferating area. Has the whole difference in length between the embryonal areas (figs. 5 and 6) been due to the elongation of the primi- tive streak itself? In the following paragraphs the measurements given are taken from measuremeuts made upon photographs of the specimens, drawings of which are given in figs. 3, 4, 5, 6, 8, and 10. The measurements, reduced to their natural magnitude, are given in millimetres. It is evident that growth has taken place somewhere, for the embryonal area in fig. 6 measured 1°72 mm., while fig. 5 measured but 1°38 mm. It is due hardly at all to lengthening of the primitive streak area, as the primitive streak area of fig. 6 measured 1:02 mm., and that of fig. 5 measured 1 mm. Of course one must not suppose that the primary area of growth has ceased to produce effects. Its effects hitherto have been to produce a circular patch. If we suppose it still to exercise a like influence the diameter of the head plate will have increased anteriorly the same as it has from side to side. Fig. 5 measures transversely ‘94mm. Fig. 6 measures trans- versely 1:1 mm. Therefore we may say that at least (16 mm, of the length may have been due to growth of the anterior THE PRIMITIVE STREAK OF THE RABBIT. 211 portion. So by adding to 1:38 mm. (length of fig. 5) the sums of ‘02 mm. due to lengthening of the primitive streak and ‘16 mm. the amount due to growth of anterior end, we have ‘16 mm. still to account for to produce the length shown in fig. 6. Where has this extra length been acquired? In transverse section it is evident that there is now a region which was not to be found in the preceding stage. A section through this region is given in fig. 24. The most noteworthy feature is the notochordal thickening. A few sections further forward this thickening is quite absent. It is surely legitimate to ascribe the appearance of this new region as being the result of cells budded off from the front end of the primitive streak. This becomes much more evident in the later stage (fig. 8). In this the total length of the embryonal area is about 3°11] mm. Fig. 6 measured 1°72 mm. There has therefore been an increase of 1:39 mm. The proliferating area measured the same as in fig. 6, namely 1:2 mm. There is at the anterior part of the embryonal area a very well marked circular area. The neural groove along this area is narrow, while the neural groove along the embryonal area between the above- mentioned circular area and the anterior end of the primitive streak is broad and shallow. The junction between the two corresponds almost exactly with the circumference of the circular area at the point Z. Does this anterior circular patch represent either exactly or approximately the area of influence of the primary centre of activity ? I wish to point out the possibility of all that part between the anterior end of the primitive streak and the posterior border of the circular anterior area (namely that part of the embryonal area which is marked by the shallow, broad, neural groove) having been formed entirely by the activity emanating from the primitive streak area, or, to be more exact, the anterior and antero-lateral parts of that area of proliferation ; the parts of the embryonal area in front of this region having been formed chiefly as the result of the activity Ai? RICHARD ASSHETON. of the primary centre of growth (the most anterior portion entirely from the primary centre, the most posterior possibly from both). The circular area measures transversely 1°2 mm. Therefore we may suppose that at least ‘1 mm. of the increase in total length of the whole embryonal area may be due to the activity _of the primary centre. This leaves ‘97 mm. to be accounted for. If we measure 1°2 mm. (the total length assumed to be due to the primary area of activity) from the anterior end we come to a spot Z. The space between this spot Z. and the anterior end of the primitive streak, 4. PS., measures ‘97 mm. This seems to me to be evidence that this space, Z. to A. PS., is the area which has been added between the stages fig. 6 and fig. 8. The question remains, where have the cells composing this portion of embryonal area originated ? They must have arisen either (i) from an area of rapid proliferation in front, or (ii) from a general area of proliferation in situ, or (il) froma general area of proliferation behind. Absolute proof is, as far as I can see, impossible, but I will take each alternative separately, and give my reason for believing that the latter only can be the true solution. (i)—(a) From the very first moment of the development the effect of the primary centre of activity is in the ontogeny of the rabbit to produce an embryo of a radially symmetrical figure, whereas the area in question between Z. and A. P. S. shows no sign of a radial symmetry. (6) By the time that the area in question is developed, and while it is being developed, there is no sign of one spot from which all pre-existing structures could obtain material. For these reasons I think growth does not take place in front. (ii)—(a) Growth of the area in question is, comparatively speaking, rapid, but there is no sign of the rapid multiplication of undifferentiated cells, every cell within this area is distinctly either epiblast, hypoblast, or mesoblast. (6) This area continues to grow rapidly in length, and the “ mesoblastic somites,’ which are a characteristic feature of THE PRIMITIVE STREAK OF THE RABBIT. ots this area, are continually being added on posteriorly, showing that the differentiation is at any rate taking place from before backwards. (c) If this is a centre of growth there should, under the cir- cumstances, be some sign of a radial disposition of tissue, which there is not. For these reasons I think we cannot conclude that there is any positive proof of a rapid addition of new cell material, although an arrangement of the mesoblast cells into meso- blastic somites undoubtedly takes place in situ. (iii) Since, therefore, we cannot find probable proof of growth of new segments either anteriorly or in the area in question itself, and since we do find an area immediately behind the area in question and continuous with it in which there is an undoubted rapid proliferation of cells (which are all similar and connected, and only become epiblast, hypoblast, or mesoblast according to which region they may adjoin), I think it may be considered that there is very good evidence indeed that the growth of the embryo of the area in question has taken place by the addition of indifferent cells from this area —the primitive streak. Although such measurements as the above are useful in support of my contention, they are not satisfactory, because every specimen of exactly the same stage is by no means always exactly the same size. A very useful landmark is supplied by the band of tissue, or rather an area which gives rise at first to the mesoblast form- ing the pericardium, and rather later the endothelial lining of the heart. This area is clearly marked as a circular band in fe LOC. In this stage the neural tube, owing to its more rapid growth, has become folded forwards so as to hide the anterior part of the pericardial band PC. In fig. 28, Pl. 20, which is a median sagittal section of this specimen, it is seen at the point PC. In fig. 9, an earlier stage, before the neural tube has become thrust forward, this band is seen to follow the contour of the 214 RICHARD ASSHETON. neural tube (NP.). So also, but less distinctly, in the earlier stage (fig. 8). In the stages earlier than this it is not perceptible in surface views, but may be easily recognised at its first appearance in stages like figs. 4 and 4a. At this time—its first appearance—it shows itself as a slight tendency to a more rapid growth of the hypoblast immediately underlying the anterior and lateral edges of the embryonal area. Figs. 17 and 18 show the anterior edge cut at PC. Sections 13 and 14, if continued to the edges, would show the same slight thickening of the hypoblast. Figs. 35—38, Pl. 21, show the subsequent history of this thickening. Fig. 35 is a transverse section through the lateral edge of the anterior part of the embryonal area of a stage between those represented in figs. 4 and 5. It is in front of the primitive streak area. There is as yet in thisregion no mesoblast. The hypoblast (HY.) shows, however, signs of increased activity, and is thicker than before. This is specially the case at the edge of the thickened epiblast. Fig. 36 is a slightly older specimen, still rather younger than fig. 7. Here the hypoblast (HY.) is thickened over the whole area, extending under the thickened epiblast, but more especially so at the edge of the epiblast. A certain number of cells are to be seen lying between the epiblast and hypoblast, marked PC. in my drawing. These I believe to have been budded off in situ from the hypoblast (HY.). Cells are budded off from the hypoblast all over this area of the anterior part of the embryonal disc, but more thickly at this region round the edge of the thickened epiblast than elsewhere. Fig. 37 is a section from the same region of a later stage, a stage with two mesoblastic somites (vide fig. 9). There seems to be here still a slight proliferation of cells from the hypoblast of this region, but not so great as before. The cells formerly budded off, which we can call mesoblast, have become arranged so as to leave a slight cavity between THE PRIMITIVE STREAK OF THE RABBIT. Al |e) them (PC.). This cavity gives rise to the pericardial cavity, the inner thick wall to the muscular wall of the heart, and peritoneal lining of the cavity around the heart; the outer thinner wall to the peritoneal lining of the rest of the peri- cardial cavity. The cells, which seem now to be budding off from the hypoblast of this spot (marked END. H.), give rise, I believe, to the first of the cells which, in fig. 38, are seen to be forming into a tube, which is the endothelial lining of the heart (END. H.). Fig. 38 is from an older stage, an embryo with seven meso- blastic somites, a little older than fig. 10. Although it is not possible to give a decided opinion, I am inclined to think that at this stage cells are still being budded off from the hypoblast to form the endothelial lining of the heart. At the same time I must mention that at an age intermediate between figs. 37 and 38 I have found a stage in which a mass of cells lies between the PC. cells and the hypoblast cells, showing no trace of a budding off either from the hypoblast or the mesoblast mass marked PC. In fig. 88 the endothelial cells may be seen to be attached to both the pericardial cells and the hypoblast cells, though the latter give one the impression of being concerned in the pro- duction of the cells in question rather than the former. This band, originating as a thickening of the hypoblast at a time when the embryonal area is hardly at all affected by the primitive streak activity, seems to be of great use in preserving the outlines of the embryo due to the primary centre of growth. Now no part of the pericardial thickening seems to be affected in the growth of the embryo caused by the secondary centre of activity, the primitive streak, excepting that it is absent posteriorly. Where it is present, that is, anteriorly and laterally, it forms almost an accurate “‘ segment ” of a circle, the circumferential boundary of which extends through about 270°, or three quarters of a whole circle. From this I argue that all the parts of the embryo which are formed within the outer peripheral boundary of the pericardial 216 RICHARD ASSHETON. band may be ascribed to the activity of the primary centre of growth (segmentation of the ovum) ; all posterior to this due in the main to the secondary centre of growth (primitive streak). Referring again to fig. 8, it will be seen that this again indicates the point Z.as the line of demarcation between the two areas. This spot is marked by a very distinct difference in the character of the neural groove, which is seen in the later stages also. It is at this spot that the first protovertebre are formed (vide fig. 9). As development proceeds, the fore-brain, whose outline is in figs. 8 and 9 concentric with that of the pericardial band, becomes thrust forward (vide fig. 10), so as when viewed from above to be no longer concentric. So also there is a similar thrusting forward of the whole of the dorsal portion of the embryo, including the protovertebre, and the pericardial band no longer serves as a landmark of the same character. The Shortening of the Primitive Streak. Although the primitive streak becomes very much shortened, Tam not at all sure that its area is diminished during the process. The hinder end of the streak is about the same width through- out its existence up to such a stage as fig. 10, and somewhat later, but the elongated anterior portion is extremely narrow. After the shortening the anterior part is almost as wide as the posterior part. So I doubt whether there is any diminution of area during the contraction. After the cessation of tension has been effected by the close attachment of the surrounding walls of the blastodermic vesicle to the uterus, the result of growth must be soon to cause pres- sure of the nature of a thrust in the tissues which are most actively growing. This, we know well, brings about a very considerable thrust in the direction of the longitudinal axis of the embryo, causing the head and tail folds. There is also the tendency that has all along existed for the secondary area of proliferation to be radially symmetrical. This, together with the pressure which is known to exist in the direc- tion of the longitudinal axis, seems to be quite sufficient cause to THE PRIMITIVE STREAK OF THE RABBIT. 217 bring about the compression of the primitive steak, without assuming a conversion of it in situ, as it were, into the embryo. On Pl. 22 will be found six diagrams which illustrate my con- ception of the lines of growth between the stages of figs. 1 and 8 of Pl. 20. Fig. 40 represents the embryonic pole of the blastodermic vesicle before there is any visible sign of the appearance of the secondary area of cell-production. The grey central area is the embryonic disc. The vesicle is expanding approximately equally in all directions ; this is indicated by the concentric circles round the embryonic disc. The dotted line is an imaginary line drawn outside the embryonic area. All parts of the vesicle outside this line are only slightly affected by the subsequent origin of the secondary area of cell-production, as seen by the next diagrams. In fig. 41 the secondary centre of activity has become esta- blished, its centre being about the spot marked with a cross in the posterior region of the embryonic disc (grey). The effect is as yet slight, leading to a little more rapid expansion of that part of the embryonic disc indicated by the ellipticity of the lines of growth. This represents the stage of fig. 2, Pl. 20. In fig. 42 the activity of the secondary area of cell produc- tion has become more intense. It is very concentrated, and is now marked by a heaping-up of cells. The outline of this actual thickening must not be taken as being the boundary of the part of the wall of the vesicle due to the secondary centre of activity, for the outermost cells produced thereby will no doubt be stretched and flattened. Fig. 43 represents a stage intermediate between figs. 5 and 6 on Pl. 20. The primitive streak is at its greatest development. The anterior part of the embryonal area, which owes its existence to the primary centre of activity alone, is shown to have its posterior borders distorted as explained in the earlier part of this paper. It is practically the same condition as that illustrated by fig. 42, but more pronounced. The outline of the part of the wall of the vesicle due to the secondary centre of activity is de- 218 RICHARD ASSHETON. rived from the outline of the primitive streak mesoblast, which upon my hypothesis would approximately mark this area. Immediately in front of the anterior end of the secondary area of activity (i.e. in front of the front end of the primitive streak) there will be a lesser tendency to expansion, and accordingly a heaping-up of cells proliferated from the front end of the primitive streak, forming the thickening known as *‘Kopffortsatz.” From this moment the increased intensity of action of the secondary area of cell-production over the primary is shown by the fact that the outline of the former increases its diameter four times by the stage illustrated by fig. 8, whereas the latter’s increase is scarcely perceptible. Fig. 44 illustrates the conditions of fig. 8. The two centres of growth have become further removed from each other by the interposition of new cell material proliferated chiefly by the secondary area. In other words, growth in length of the embryo has now very markedly taken place. This growth in length is represented in the diagram by the area within the curves of which the one marked 1 is the outer- most. The region of the embryonal area due to the primary activity has recovered from its temporary distortion, and has regained its radial symmetry to a great extent. This can be detected in the series of drawings figs. 1—8. The secondary area of activity also, upon the diminution of the tension to which it had been subjected, tends also to assumea more radial form (fig. 45, with which fig. 10 may be compared), Ultimately the counteracting effects of the two centres cause each to become tilted over, so that instead of lying in the same plane they lie in different planes parallel to each other, and at right angles to the original common plane. In fact they assume their natural positions, On this explanation it is clear that most, if not all, of the ectoplacental region is really derived from the primitive streak, which may perhaps account for its much greater activity than that part of the blastodermic vesicle which arises directly from the primary centre of activity. THE PRIMITIVE STREAK OF THE RABBIT. 219 EXPLANATION OF PLATES 20—22, Illustrating Mr. Richard Assheton’s paper on “ The Primitive Streak of the Rabbit; the Causes which may determine its Shape, and the part of the Embryo formed by its activity.” List oF REFERENCE LETTERS. A, Anterior end. AM. Amniotic cavity. 4. PS. Anterior end of primi- tive streak. #ND.H. Endothelial lining of heart. 4#P. Hpiblast. AN. Hensen’s node. HY. Hypoblast. M4. Centre of primary area of cell-pro- duction. MHS. Primitive streak mesoblast. MHS. HY. Hypoblastic meso- blast. 2. Centre of secondary area of cell-production. CH. Notochord. NG. Neural groove. WP. Neural plate. iP. Posterior end. PC. Peri- cardial band. P. GR. Primitive groove. P. PS. Posterior end of primitive streak. PS. Primitive streak. Z. Point which marks approximately the boundary between the results of the primary centre of growth and that of the secondary centre of growth. PLATE 20. Fie. 1.—Surface view of the embryonic dise of a rabbit embryo of the 150th hour. x 18. Fic. 2.—Surface view of the embryonic disc of a rabbit embryo, at the earliest moment at which the primitive streak activity is perceptible as a thickening. Age about 158 hours. x 18. Fic. 3.—Surface view of the embryonal area of a rabbit embryo, in which the primitive streak is very evident. Age about 168 hours. x 18. Fic. 4.—Surface view of the embryonal area of a rabbit embryo, in which the process of lengthening of the primitive streak is well advanced. Age about 168 hours. x 18. Fic. 44,.—Surface view of the embryonal area of a rabbit embryo, slightly more advanced than the preceding. Age about 168 hours. x 18. Fic. 5.—Surface view of the embryonal area of a rabbit embryo, in which the primitive streak has become greatly elongated and is grooved slightly. Age about 172 hours. x 18. Fic. 6.—Surface view of the embryonal area of a rabbit embryo, in which the primitive streak has attained its maximum length, and is most deeply grooved. This figure has unfortunately been drawn slightly larger than the photograph from which it was copied, on which the measurements in the text were made. The error is about 75. Age about 180 hours. x 18. 220 RICHARD ASSHETON. Fic. 7.—Surface view of the embryonal area of a rabbit embryo, in which the primitive streak has become much shallower. Growth in length of the embryo has now taken place. Age about 188 hours. x 18. Fic. 8.—Surface view of the embryonal area of the embryo of a rabbit. The primitive groove has entirely disappeared. The ecto-placental region is now firmly attached to the placental lobes of the uterus. Age about 192 hours. xX 18. Fig. 9.—Surface view of embryonal area of the embryo of a rabbit. Age about 196 hours. x 18. Fie. 10.—Surface view of the embryonal area of the embryo of a rabbit. The primitive streak has now become compressed to its former dimensions ; the mid-dorsal portions of the embryo have become thrust forward, and the head-fold has commenced to be formed. Age about 200 hours. x 18. Fie. 11.—Diagram to illustrate the relative positions of the two areas of cell-production. Fic. 12.—Diagram to illustrate the effect produced upon the two areas of cell-production by the increasing hydrostatic pressure within the blastodermic vesicle. Fie. 17.—A sagittal section through the specimen of which Fig. 44 is a drawing. It is taken along the line 17. x 100. Fig. 18.—A median sagittal section through a specimen intermediate be- tween Figs. 5 and 6. xX 72. Fie. 28.—A median sagittal section through Fig. 10. x 54. Ftc. 39.—A horizontal section through a portion of the sheet of mesoblast surrounding the primitive streak. The portion drawn was a lateral portion. x 350. PLATE 21. Fies. 13, 14, 15, 16.—Transverse sections through specimen Fig. 4, along lines 183—16. x 175. Fics. 19, 20, 21, 22, 23.—Transverse sections through specimen Fig. 5, along the lines 19—23. x 175. Fies. 24, 25, 26, 27.—Transverse sections through the specimen Fig. 6, along the lines 24—27. x 175. Fries. 29, 30, 31.—Transverse sections through a specimen similar to that drawn in Fig. 8, along lines corresponding to those marked 29, 30, 31, in Fig. 8. x 175. Fics. 32, 33, 34.—Transverse sections through the anterior end, the middle, and posterior end of the primitive streak of a specimen slightly older than that of which Fig. 10 is a drawing. xX 175. THE PRIMITIVE STREAK OF THE RABBIT. 221 Fic. 35.—A transverse section through the edge of the embryonal area of a specimen rather older than Fig. 4, on a level corresponding to the line 13. x 175. Fie. 36.—A transverse section through the corresponding region of a specimen like Fig. 7. x 175. Fie. 37.—A transverse section through the corresponding region of a specimen like Fig. 9. x 175. Fic. 38.—A transverse section through the corresponding region of a specimen rather older than Fig. 10. x 175. PLATE 22. Fie. 40.—Diagram showing lines of growth, i.e. expansion of blastodermic vesicle of the embryonal area and surrounding parts of a rabbit embryo, corresponding to Fig. 1, Plate 20. Fic. 41.—A similar diagram of a slightly later stage, corresponding to Fig. 2 at the time of the first appearance of the secondary centre of growth, which is placed eccentrically to the centre of the primary area of growth. Fies. 42, 43, 44.—Similar diagrams illustrating the increase in importance of the secondary centre of growth. These correspond to Figs. 3 or 4, between 5 and 6, and to 8. Fie. 45.—A portion of a similar diagram of the posterior end only of the embryonal area of a specimen, corresponding to Fig. 10, or a little later. ‘ay ys . WS Rs A ee eh ‘ ty oese ahi Rev rg Bite Dif tis = Ae CN ee w > Be - @ a SE me Shae eat a0 | ~ o os = “9 rey tohgel Pays si a 7 : 7 ah ie sd iuher Jah pall wat art ey ’ a a} Opt ee PRS “hh pS] Leip Pek yy gh! 5 Pelle wae i a aq heal ith ti Richly {nas «i was : Aa vel 5 Welw Lay yon 0 Te ee “dé Ging « ur > wae oe og ee ee ae 7 ro Tee A er? 2S g” i, : i dF whh \\ie0 jen opin nt ee fe ; a ie ~ l | a 1% saa SSA TAL T were 10 cagil ange Ob onion AR ¥ WiiZee wwe © batarael yay, od Ce nis 0? Yt 1 ine Lal Dias fd ty comstadal te a ‘ wot Lat catty: we ti pel eae ae asl ‘- - a ey is Podine 7 Ah 7 Ae gah ws / “ 7 rhe ie i A bal Ps | cl pilee gov} 4 ‘> Wy iuila @ Te pe hiegw ri ' ae ” GROWTH IN LENGTH OF THE FROG EMBRYO. aoe On the Growth in Length of the Frog Embryo, By Richard Assheton, M.A. With Plates 23 and 24. In some previous papers on the development of the rabbit I have attempted to show that there are two main centres of growth, each in itself tending to produce a radially symmetrical form; but since these two centres of growth are situated eccentrically to each other the resulting embryo is cylindrical, and subsequently bilaterally symmetrical. I endeavoured to show that no concrescence occurred in the rabbit, and that no theory of concrescence was necessary to account for the facts. I wish now to indicate the manner in which two centres of growth can also bring about the corresponding results in the frog embryo, without any concrescence of the dorsal lips of the blastopore. In a paper in this Journal Dr. Robinson and I discussed the question of the formation of the archenteron in the frog, and came to the same conclusion as Moquin-Tandon (in Anura) and Houssay (in Axolotl), that the archenteric cavity was due to a splitting amongst the cells in situ, and not to an invagi- nation or overgrowth of surface cells. Recently Jordan, and Umé Tsuda and Morgan have made very interesting communications upon the subject. The former author, after an able summing up of the evidence on both sides, concludes (p.331), ‘The evidence thus far adduced 224, RICHARD ASSHETON. for invagination is, to say the least, inconclusive ;” and on the other hand (p. 332), “ There is not a shred of evidence to show that the large cells at first surrounding the mouth of the blas- topore are not subsequently pushed in by the ingrowth of ectoblast cells. No positive evidence whatever exists to prove either the impossibility of invagination or the likelihood of no invagination. I find it difficult to gather the reasons that have influenced Houssay and Robinson and Assheton to adopt the view that invagination does not occur.” Jordan then describes (pp. 333-4) “ocular evidence that the small cells around the lips of the blastopore are actually infolded.” Morgan, after a description of interesting experiments after the method of Roux upon the living egg, says, “‘ The statement of Robinson and Assheton that no portion of the archenteron in the anura is formed by invagination is certainly incorrect, as I hope to show in a later paper.” Of course this latter statement must depend upon the exact meaning to be attached to the word archenteron. My own conception of the term archenteron is that cavity which in the embryo is supposed to represent the digestive cavity of a hypothetical ancestral “ gastrula,”’ no matter how this cavity was brought about. If, however, by archenteron is meant any subsequent pro- longation of this cavity, such as would represent a post-gastrula condition ancestrally, then certainly such a statement was inaccurate. When Dr. Robinson and I made the statement referred to, we regarded as archenteron part of the cavity which I now con- sider to represent a post-gastrula condition. In other words, I agree with Morgan and others to a certain extent as regards the growth over of the dorsal lip of the blastopore, and consider only the most anterior part of the gut cavity of the frog’s embryo at the time of the closure of the blastopore as being formed by a splitting, and as representing the true archenteron. Accordingly, in my opinion, the sentence (referred to above) by itself accurately describes the facts, but in the context in GROWTH IN LENGTH OF THE FROG EMBRYO. 925) which it stood I admit that I now think it inaccurate. My reasons for so thinking I will now proceed to give. Again, Morgan and Umé Tsuda say that we “ apparently at the outset have orientated the embryo wrongly, for they state the segmentation cavity has a roof which ultimately becomes the anterior wall of the gastrula; for the anus which marks the posterior end of the embryo appears at the opposite side of the ovum,—that is, on the floor of the segmentation cavity.” I cannot understand their objection to this paragraph. Figs. 7 to 11 on Pl. 24 are all placed with what we conceive to be the dorsal surface (D.) directed towards the top of the plate. As regards the ocular evidence of an invagination spoken of by Jordan, it is a pity that more details are not given of the observations. Is it possible to trace a cell, or a spot on the surface some distance from the lip of the blastopore, to gradually approach and fold over the edge and so disappear, or do only cells actually on the edge seem to be affected by the process ? What is the cause of the invagination? I can quite well imagine that individual cells at the edge may, by multiplication of their neighbours or themselves, be pushed over the edge, as also might cells on the inner edge appear to be pushed out- wards, if we could see that edge. The splitting theory still seems to me to be the more probable for the commencement of the archenteric cavity and its extension forwards. But, as I shall point out a little further on, there is un- doubtedly an apparent overgrowth, and I think certainly an actual overgrowth of the lower pole cells by the dorsal lip of the blastopore, together with the lateral and ventral lips, as they are formed at a later period. This process, however, should not, I think, be compared with the process of gastrulation so called, or formation of primitive archenteron, but should be considered to be intimately connected with the growth in length of the embryo. In other words, to follow the same line of argument that I have used in the description of the rabbit embryo, the VOL. 37, PART 2.—NEW SER. Pp 226 RICHARD ASSHETON. formation of the primitive archenteron is by a process of split- ting, and is the direct effect of the primary centre of growth; whilst the continuation of the cavity produced by an overgrowth is the direct effect of the secondary ceutre of growth, producing the elongation of the animal. The splitting process in the frog corresponds in results to the invagination process of Amphioxus, while the overgrowth of certain parts of the white pole of the ovum of the frog by the dorsal, and subsequently lateral and ventral lips of the blasto- pore, together with the continuation of this process in the for- mation of the tail, corresponds to the elongation of the gastrula in Amphioxus, by means of what Hatschek called the polar cells. I shall now attempt to explain what I believe to be the actual method in which the splitting is brought about. The frog’s egg segments, as has been described by many observers, more rapidly at one pole than the other. ‘This is, I think, universally supposed to be due to the greater accu- mulation of yolk granules at the “ lower” pole, which thereby hinder the segmentation activity at that pole. If we admit that “ yolk ”’ determines the inequality of the process known as segmentation, we must admit it also in the case of each cell. If it is true of the segmented ovum, it is equally true of the unsegmented ovum. ‘To say that yolk being more plentiful in one part of a cell than in another hinders the activity of the protoplasm, is the same as saying that a cell divides into two parts, which in magnitude are in inverse ratio to the purity of the protoplasm contained. In other words, the result of asimple process of cell division, such as we see in the segmenting ovum, is two cells equally balanced as regards protoplasmic energy. Fig. 15 on Pl. 24 is a diagram of a vertical section of the unsegmented ovum of the frog. The circles 1 to 7 represent diagrammatically what I imagine to be the distribution of yolk, as determined from a considera- tion of the segmented ovum. The space No. 1 is that region in which segmentation is most retarded, and so presumably the region in which yolk is GROWTH IN LENGTH OF THE FROG EMBRYO. 227 most abundant. The space No.2 contains less yolk to a given area than No. 1, No. 3 less than No. 2, and so on. For the sake of simplicity we may regard the outer space only. This may be supposed to contain protoplasm of a uniform degree of purity. Accordingly division of this space will be such as to produce two spaces whose areas are equal. This is about the spot marked by the line (a), and will repre- sent the third furrow of segmentation, that is the first hori- zontal furrow. Similarly, the next horizontal furrows will be about the spots 6 b, the next at cc cc, the next atdddddddd, and so on; always resulting in a balance of protoplasmic energy on each side of the furrow. In this way the frog’s egg becomes segmented more and more rapidly in the upper hemisphere than in the lower. For a considerable time there is an almost complete absence of horizontal furrows in the lower hemispheres. This point is very well seen in Umé Tsuda’s figures Iv, v, of Plate 24, ‘ Quart. Journ. Micr. Sci.,’ vol. xxxv, part 3. Another effect is that as segmentation proceeds there is a continual increasing disparity in size between the cells of the black pole and those of the white. Whereas at first the superficial area of the cells of the extreme upper pole bears to the superficial area of the cells of the extreme lower pole the ratio of 1 to 2, at the time of the commencement of the blasto- pore it bears the ratio of 1 to 5. In this way there is a gradual apparent creeping of small (black) cells over the surface of the egg—though in reality it is conversion of large cells into smaller in situ, as, I believe, is now generally accepted. My diagram fig. 15 gives the idea of no segments in the white or lower hemisphere of the ovum. This is because it deals only with horizontal furrows. The segmentation energy may be said to produce its effects along the area of least resistance. Is it not possible that the commencement of the archenteron may be a continuation of this same process ? 928 RICHARD ASSHETON. The effect up to now has been to produce a fairly sharp line of demarcation between small and large cells upon the surface at the point x in diagram, fig. 15. On the supposition that this diagram represents fairly accu- rately the distribution of yolk, it is clear that as this line advances it encounters greater and greater resistance. May not a time come when it will find the path of least resistance to be inwards and backwards, as in diagram 16? Diagrams 15 and 16 are inaccurate for later stages of seg- mentation, because they do not show a segmentation cavity. Fig. 12 is a more accurate representation of a completely segmented egg. A. is the black upper pole (the anterior wall of the future embryo); P. is the white lower pole (posterior end of the future embryo) ; sg. the segmentation cavity. The letter A. points to the smallest cells of this stage, y. p. to the largest. There is a gradual merging of the one into the other, not along the surface, for here the line is much sharper, but along the cells to which y. and a. are directed. My idea is that the continuation of the segmentation process is the conversion of, first, the cells y., then the cells z. into smaller ones, and in this way a layer of small cells will be produced lying up against the mass of much larger cells y. p. This layer I have indicated in the fig. 12 by the dotted line. If a section of an embryo of the stage in which the blasto- pore is nearly complete is examined, it will be seen that there is such a layer of cells along the floor of the segmentation cavity. Fig. 21 is an outline camera drawing. Such details as are shown were not drawn by camera. The smallest cells are those forming the lip of the blastopore. The point a. represents the at present most anterior limit of the archenteron. More anteriorly, however, following the lines a a., a a., there is what I take to be a differentiation of the yolk-cells, that is a splitting up into smaller cells, which cells, upon the splitting hypothesis, will eventually form the GROWTH IN LENGTH OF THE FROG EMBRYO. 229 roof of the archenteron, and come to lie up against the epiblast now forming the roof of the segmentation cavity, as shown diagrammatically in figs. 12 and 13. The differentiation as seen at the point aa. must, on this hypothesis, be considered to be the effect of the direct continua- tion of the process of differentiation on the surface of the ovum (whereby the epiblast is separated), that is a direct continua- tion of the process of segmentation. This line bounded above by small cells, below by large cells, constitutes a line of separa- tion or a split, which is, I believe, the first commencement of the archenteron, and is a result of the primary centre of activity, comparable to the events of the first five days in the development of the rabbit, or to the formation of the gastrula in Amphioxus. But although corresponding in effects, the only really homologous feature is the presence of a primary centre of activity, or process of segmentation of the egg; the actual directive agencies being in each case ccenogenetic and entirely different. The conversion of the narrow slit into a spacious cavity is to be considered to be due, at any rate in part, to the effect of the secondary centre of activity, to which I shall refer again. I must now refer to the experiments made by Roux and Schultze and Morgan and Umé Tsuda upon the developing egg by following natural or artificial spots, which experiments I have myself repeated during this spring. It is impossible to repeat these experiments without becom- ing convinced that there is a change of relative position between certain spots on the ovum,—for instance, the dorsal lip of the blastopore, and the most inferior spot upon the white pole of the ovum. These two spots, as seen from without, un- doubtedly approach one another before the complete formation of the circular blastopore. But the question to what extent this approximation is carried, and whether by a concrescence of the lateral lips of the blastopore, or by a rolling under of the white pole, or by a growth over the upper lip without concres- cence, is answered differently by the several observers, My own suggestions are as follows, 230 RICHARD ASSHETON. The Secondary Area of Cell Production. I think that every one will agree that after the closure of the blastopore, the embryo grows in length by the proliferation of cells at the spot which formerly formed part of the lips of the blastopore. There is very little doubt that rapid growth at this spot takes place before the final closure of the blastopore ; the question is, when does this growth begin ? Again, it must be remembered that growth of the embryo as a whole, derived from the rapid multiplication of the cells in this area, is growth in length. It is the secondary area of growth comparable to the secondary area of growth or primitive streak of the rabbit. If there is such a growth backwards of the blastoporic lips before their closure, there will then be a portion of the future gut cavity of the embryo that will have been formed, not by a splitting nor by an invagination, but by a growth backwards of the blastoporic lips. Amphioxus, after the completion of the process of invagina- tion, begins to grow in length. According to Hatschek’s account this was largely due to the activity of two pole cells. Recently, however, Wilson has stated very clearly that these pole cells are “a myth.” They never exist at any time, but the posterior region of the larva of later stages ‘is rapidly growing, and numerous mitoses may be observed in all the cells in the region of the mesenteric canal.” Now although the process of invagination produces the double-layered condition of the embryo of Amphioxus, and at the same time the cavity of the archenteron, yet it is only the anterior part of the archenteron that is formed in this way. There is a posterior point of the archenteron which is formed, not by invagination, but by growth of the blastoporic lips. This must be so, whether we accept Hatschek’s or Wilson’s description of the secondary growing point. The exact line of demarcation between the two parts I have no means of showing. It is not easy to say at what moment GROWTH IN LENGTH OF THE FROG EMBRYO. 231 the secondary growing point becomes a functionally active area in Amphioxus. It is possible for it to become established as soon as a blastoporic lip is formed, and not before, because a characteristic feature of this secondary area of proliferation is that it should produce cellular units to all existing cellular layers. I have in a previous paper tried to locate this line of demar- cation in the rabbit. The moment of origin of the secondary area of proliferation in the rabbit is fairly well marked. Of Amphioxus I cannot speak. Can we find it in the frog? The frog is by no means so simple as the rabbit, but is more amenable to experiment than is Amphioxus. The frog differs in one respect from Amphioxus, which is of importance in reference to the question now under discussion. In Amphioxus the blastoporic lip is formed at the same moment apparently at all points of its circumference. In the frog it is formed at one point first, namely, at the future dorsal region, and many hours elapse before the lip is formed ventrally. Hence it is possible for the secondary area of proliferation to become established much sooner dorsally than ventrally. In my account of the frog given above, the production of the split forming the primitive archenteron, which I believe to represent the process of gastrulation of Amphioxus, although no invagination occurs, is to be considered, like the invagina- tion process in Amphioxus, as the result of the primary area of proliferation, and of itself would tend to the production of a radial symmetry. The very moment this split begins, a portion of the blastoporic lip is thereby formed. If my supposition is right, that the secondary area of pro- liferation may be established as soon as there is a mass of cellular tissue in connection with all the primary layers, it is clear that possibly the secondary area of proliferation may start immediately upon the formation of the dorsal lip of the blastopore, and not delay until the whole blastoporic rim is completed. Accordingly, on this view, there will be an extension of gut 232 RICHARD ASSHETON. cavity anteriorly by means of a splitting, the result of the primary area of activity, and posteriorly by means of a growth backwards of the dorsal lip of the blastopore, the result of the secondary area of activity, comparable to the corresponding parts in the rabbit, formed previous to the eighth day, and upon and subsequently to the eighth day respectively. In the rabbit and in Amphioxus the lining of the archenteron of the primary area is completed before the secondary area of pro- liferation has become established, but in the frog afterwards ; and so the linings of both parts of the gut cavity are formed together. Many actual experiments and observations have been made upon the eggs of the frog with the object of demonstrating the mode of formation of the blastopore, and the relative position of the blastopore when it has a completed margin to the originally black and white poles of the unsegmented ovum. Such attempts have been made with varying results by Roux, Schultze, Hertwig, Morgan, and Umé Tsuda, and although the experiments described are in many cases contradictory, yet there seems to be no doubt that the dorsal lip of the blastopore does overgrow a portion of the whiter side of the embryo prior to the completion of blastoporic lip ventrally. I have myself made similar experiments in repetition of Roux, and I am quite convinced that this overgrowth does occur to a certain extent, but I am equally sure that it is incorrect to assert that the neural plate is formed entirely upon the lower (white) pole of the ovum. The dorsal lip overgrows the white segments, at any rate apparently, but so do the lateral and ventral lips as they are formed. It is only because the dorsal lip is formed first that this part seems to overgrow the white pole to so large an extent. The overgrowth is a part of the same process which pro- duces the lengthening of the embryo. If the whole blastoporic lip could in the frog be formed at once the embryo would, I suspect, change rapidly from a sphere to an oval, as does the embryo of Amphioxus (v. Hatschek, figs. 80 and 34). GROWTH IN LENGTH OF THE FROG EMBRYO. 233 In the frog the dorsal lip cannot of itself grow outwards and so produce an oval embryo until the rest of the blastoporic lips are formed. Unless it remains inactive it must follow the contour of the ovum. That it does not remain inactive I have convinced myself, and therefore I agree with the above- mentioned authors that a portion of the white area passes out of view of the observer by becoming hidden by the advancing dorsal lip of the blastopore. Experiments in marking Parts of the Ovum. General Remarks.—I find— (i) That it is impossible to fix the egg in any one position so as to prevent with certainty the rotation of the ovum within the vitelline membrane without injuring or distorting the ovum. (ii) That, accordingly, any fragment which exudes from the ovum through the aperture made in the vitelline membrane when pricking the ovum in order to mark one spot is useless as a landmark. (iii) A scar upon the ovum itself, fixed to the ovum and within the vitelline membrane, is the only mark which can be relied upon for drawing conclusions as to the relative rate of growth, and a change of position at different points upon the surface of the ovum. (iv) A severe injury by pricking naturally produces much abnormality of development; whereas a very slight injury, although admirable for a short observation, is apt to recover and so get lost and obliterated after many hours. Some of my own experiments I will now briefly describe. Outline figures are given upon Pl. 23. Figs. 1a—1ld show the results of an experiment. Here the puncture was very small, and made midway between the two horns of the developing blastoporic lips (fig. la). Three hours and a half later the blastoporic lips were completely marked. By this time (fig. 14) the mark was distinctly closer to the dorsal lips of the blastopore. Three hours later the mark had approached the dorsal lip still nearer, but the ventral 234, RICHARD ASSHETON. lip of the blastopore had gained a little upon the mark. After a lapse of eight more hours the blastopore was very much smaller, and the mark was found partly covered by the dorsal lip. In this (fig. 1d) the ventral lip had gained much more upon the mark than had the dorsal lip. Figs. 2a—2d are figures of a specimen which had a natural mark upon the white pole of the ovum. The drawings were made at 5.30, 8.30, 10.40 p.m., and 8.30 a.m. I thought the scar was part of the embryo, but upon the blastoporic rim reaching it the scar became partly scraped off on to the rim. Both this specimen and the last show that the apparent over- growth of the dorsal lip of the blastopore is much more marked at first than afterwards. This is well illustrated by figs. 4a,45. In this specimen a mark was made after the complete formation of the blasto- pore (fig. 4a) near the centre of the unenclosed yolk. Fig. 4b is the same specimen seventeen hours afterwards. I do not know to which lip the mark was approximated. Figs. 83a—3g show a similar apparent overgrowth of the dorsal lip, and also that this overgrowth is greater during the earlier period of blastopore formation. Figs. 54, 5c, and 5d are from embryos which have completed the closure of the neural plate. All these were, at the moment of the first signs of the blastoporic lip, pricked near to the margin between the black and white at the point most distant from the commencing blastopore, and equidistant with the latter from the equator of the ovum. Fig. 5d shows the scar a little to the left of the spot where the blastopore has closed. Fig. 55 shows the scar upon the side of the embryo about its middle, both dorso-ventrally and antero-posteriorly. Fig. 5c shows the scar upon the ventral edge of an unclosed blastopore. In this specimen the injury was very severe, a large mass (ewo.) exuded, and, as always follows in such a case, an abnormal embryo was formed. Fig. 5a shows the spot near which they were all pricked. Ten embryos were pricked in the centre of the lower pole of GROWTH IN LENGTH OF THE FROG EMBRYO. 235 the ovum in the blastula stage before any trace of the dorsal blastoporic lip could be detected. Of these when preserved, at which time the normal ones were from 43 mm. to 5 mm. in length, seven showed no trace of the injury externally, and seemed to be quite normal. One showed no injury, but was rather abnormal in shape. Two failed to develop beyond the blastula stage. Another specimen was pricked, as shown in fig. 6a, on both sides of the blastopore, on one side upon the lip, on the other a slight distance away from the line where the lip was appa- rently about to develop. Fig. 64 was drawn ten and a half hours afterwards. After the blastopore had closed I was unable to detect the injuries. There is no doubt that one must be very cautious indeed in drawing conclusions from injuries made upon eggs. This is especially so with injuries made upon the more active part of the embryo, i.e. the more deeply pigmented cells. A very in- considerable injury is sufficient to produce an abnormality. Three slightest punctures possible upon the rim of the blasto- pore equidistant from each other are sufficient to prevent the closure of the blastopore, while one only, if at all severe, will have the same effect. This clearly must be the case, as the closure of the blasto- pore is an effect of increase of bulk of certain parts of the walls of the embryo, and if this increase in bulk is, through pricking these walls, prevented or delayed by the letting out of matter, and thereby obviating the necessity for the blasto- poric lip to advance, the blastopore does not close; and so also injuries to the white pole or yolk plug by allowing the escape of material from that area, and thereby diminishing the bulk of the part of the embryo that can be covered by the advancing lips of the blastopore, hastens the closing of the blastopore. My own experiments are in part confirmatory, but mostly contradictory to those of Roux. They are upon the whole confirmatory of those performed by Morgan and Umé Tsuda. Roux asserts that the dorsal lip of the blastopore passes over 236 RICHARD ASSHETON. the lower pole of the ovum through at least 170°. The ventral lip according to him does not advance at all. Morgan and Umé Tsuda consider that the ventral lips and lateral lips advance, but not to so great an extent as the dorsal lips. They also notice the “first overgrowth of the dorsal lip of the blastopore is more rapid than the later growth ; that is, the approach to the points of injury is faster at first.” I quite agree with the latter authors that “itseems..... most probable that the blastopore does not start at the equator of the egg, but some distance below that circle.” Now my experiments do not give evidence of an overgrowth by the dorsal lip of more than 60° or 70° from the moment of the first commencement of the dorsal lip, and to the closure of the blastopore. More probably, I think, the apparent overgrowth is even less. According to Roux the overgrowth is at least 170° to 180°. If Roux is right in both his suppositions, namely, that the dorsal lip moves over the white pole, and to an extent of 180°, I cannot understand how the last remaining portion of the blastopore to remain open should show so white a piece of yolk plug. This piece of yolk plug is as white as any part of the surface of the ovum of the frog ever is. There is a considerable amount of variation in the pigmentation of the unsegmented ovum. It is extremely rare in England to find eggs in which there is deficiency of pigment over an area sub- tended by so great an angle as an angle of 120°. An area where there is almost an absence of pigment is much more restricted. Very frequently the less pigmented area extends over a much smaller are. On Roux’s supposition, the part which remains longest uncovered ought to be grey, if not quite black. It is, as far as I have observed it, an almost invariable rule to find the yolk plug at its latest stage intensely white. I have only once seen embryos which in the gastrula stage showed a darkened blastopore, and these were from eggs which in the unsegmented stage were so intensely pigmented GROWTH IN LENGTH OF THE FROG EMBRYO. 937 that the lower pole was only slightly lighter in colour than the upper, and this for an area not greater in extent than that sub- tended by an angle of 50°. Yet in these the blastopore, though very dark, was quite as light as the lightest part of the unseg- mented ovum. If the centre of the blastopore at the moment it is in the stage represented in fig. 20 is not either the lower pole, or some spot extremely close to the lower pole, of the unsegmented ovum, the intense whiteness of this spot must have been pro- duced by a disappearance of pigment previously existing. Is there any evidence of this? I cannot think of any. On the contrary, there is evidence of increasing pigmentation, as, for instance, in the epiblast-cells as they form upon the sur- face, in the cells of the splitting archenteron, and even in the white cells themselves. If the final position of the blastopore is, as Roux supposes, at the equator, and at a spot removed 170° from the spot of the first commencement of the blastopore, surely the uncovered part of the surface would be dark, if not black, and certainly not intensely white. At the moment the definite outline of the ventral lip of the blastopore is formed the anus of Rusconi thus fashioned is not of uniform tint. The dorsal part of the area is much lighter in colour than the ventral part, fig. 19. When, how- ever, it has diminished to the condition of fig. 20, it is of uniform tint and intensely white. This is accounted for by the more rapid closure of the ventral lip from this moment, as my experiments and those of Morgan and Umé Tsuda demonstrate, as also may be seen by examination of sections as described in a former paper (Robinson and Assheton). Except for a slight advance of the dorsal lip of the blasto- pore my experiments do not support Roux’s. According to Roux, injuries made at the point 2 in fig. 5a ought to have appeared upon the dorsal side in the medullary folds a little way anterior to the blastopore. Instead of which one was in the median line ventral to the _ blastopore, one was laterally placed on a level with the blastopore, and one laterally placed but far forwards. So, again, if the lips of the 238 RICHARD ASSHETON. blastopore concresce as Roux assumes, then marks made upon the lips should show upon the dorsal surface somewhere along the neural folds. In no case did I ever find this to occur. As the result of such injuries, I either found the scar upon the lateral lip of the blastopore when completed, or else as in figs. 6 a, 6 6, further removed from the blastopore but in the same relative positions. I never found a defect in the neural plate except when the dorsal, or near the dorsal lips of the blastopore, was injured. This spot is unfortunately at the same time the most interesting to injure, and the most delicate, and most liable to produce abnormalities which prove very little. There is certainly no need to assume a concrescence, as the facts can be equally well accounted for by other means, which to my mind agree far better with the development of other Vertebrates than does the concrescence theory. In other words, I believe that as in the rabbit, so in the frog, there is evidence to show that the embryo is derived from two definite centres of growth, the first, and phylogenetically the oldest, being a protoplasmic activity which gives rise to the anterior end of the embryo (= gastrula stage) ; the second, which gives rise to the growth in length of the embryo: which centres of growth occupy the same relative positions in loca- tion and in sequence of time, and probably to each are due the same parts of the embryo. In the rabbit the area is a spot which assumes for a time a linear form, but is unaffected by its change of shape in the functions it has to perform. So in the frog, although at first crescentic, then circular, then linear, and ultimately a knob, its function is precisely the same as in the rabbit, and is unaffected by the change in form. From the moment of its first appearance it performs its one function—that of adding on new cellular units to the previously existing embryo. One difference of effect is that owing to the manner of its coming into existence, one portion arising before the other, that portion—the dorsal—becomes functional before the GROWTH IN LENGTH OF THE FROG EMBRYO. 239 ventral, and so the dorsal part of the embryo is developed more quickly than the ventral. I believe the true way of regarding this area of secondary proliferation, both in the rabbit and in the frog, is as a single area, whether circular, annular, or linear, whose sole function is the addition of cellular units to the posterior end of the previously existing embryo. Its form is the result of secondary or ontogenetic causes. The exact line of demarcation is not easy of determination. Very careful marking of the dorsal lip might give it as far as the nervous system is concerned, but organs, no doubt, change their relative position somewhat as they develop. The brain is certainly thrust somewhat forwards. I brought forward evidence to show that in the rabbit this point was about the level of the first mesoblastic somite. It is, at any rate, possible that metameric segmentation may be directly due to this process of elongation. It seems always to be closely connected with it. If so it may be due to this, that the anterior mesoblastic somite of the frog is the smallest, and each for succeeding five or six becomes longer dorso-ventrally than its preceding neighbour. For upon my suppositions of the non-concrescence of the blastoporic lips, and of the unity of the nature of this area of secondary proliferation, then, since the first part of this proliferating area to be formed is that part adjoining the dorsal surface, those parts in the mid-dorsal line, e. g. neural plate, will be the first, and at first the only part of the embryo to receive additions from the proliferating area. As the lateral lips of the blastopore are formed, more and more of the lateral plates of mesoblast will receive additions, so that in this way it is possible that the gradual increase in size of the first six mesoblastic somites in the frog may be connected with the gradual development of the area of proliferation. Every one is agreed that there is a certain part of the neural plate formed on an area anterior to the first commencement of the blastopore lip. The point in discussion is to what extent does this pre-blastoporic formation exist ? 240 RICHARD ASSHETON. Morgan and Umé Tsuda conclude that all except ‘“ the thickness of the medullary folds”? round the dorsal lip of the blastopore is formed by the growth of the lip. Roux shows the same in his figure. Pfluger, however, thinks it possible that a considerable length of the anterior part of the nervous system is formed in this black hemisphere, and with Pfliiger I quite agree on this point. I find that the neural plate in normal embryos at the time it becomes visible on the surface extends through fully 170°, if not more, while the distance through which the dorsal lip of the blastopore travels I cannot make out to be more than 70° at the most; that is, from a spot a little below the equator to the lower pole, or perhaps a little beyond it. Figs. 7 to 14 represent diagrammatically the views put forward in this paper. Fig. 7 is the fully segmented frog’s egg, the white pole placed to the right of the paper; the black pole or roof of the segmentation is placed to the left, as repre- senting the future anterior end of the embryo. All the others, 8—14, are arranged similarly, Fig. 8 represents the stage at which the dorsal lip of the blastopore has become established. Up till now there has been but one general centre of growth. From this moment the secondary centre of growth is in existence, and we have now the commencement of the con- version of an embryo radially symmetrical into an embryo bi- laterally symmetrical. As yet only the dorsal part of this secondary area of proliferation is in existence, and accordingly the dorsal part of the embryo is developed more rapidly than the ventral, as the annexed figure 9 shows. In this figure the ventral part of the secondary area has just been completed, and now the whole of the secondary area of proliferation, the homo- logue of the whole of the primitive streak of the rabbit, is complete, and new material is added to ventral and lateral and dorsal parts of the embryo, as diagram fig. 10 illustrates. The shape now rapidly changes, and the radial symmetry is lost and the bilateral symmetry acquired, fig. 11. The neural plate is indicated in fig. 11 by the continuous GROWTH IN LENGTH OF THE FROG EMBRYO. 241 line ; the dotted line represents only approximately the sup- posed division between the parts of the embryo derived from the primary and secondary areas of proliferation respectively. The subsequent fate of the secondary area of proliferation (or primitive streak) I have, with Dr. Robinson, described and discussed in a former paper. The three figures 8, 9, and 10 represent my views of the extent to which the white pole becomes overgrown by the dorsal lip of the blastopore. The neural plate is indicated as in fig. 11 by the continuous line. The extreme anterior end of this part of the epiblast has been obtained by subtracting the amount due to overgrowth from the total amount observed when definitely established. It does not necessarily follow that the distance through which the edge of the dorsal lip of the blastopore advances represents the total growth in length due to that part of the secondary area of proliferation. In order to advance, the lip of the blastopore has to exert pressure upon the “ yolk plug,” causing it to be forced inwards. It thus follows that an equal force must be exerted in the other direction. What effect this has will depend upon the strength of the resistance offered by the anterior wall of the embryo. From the fact that the archenteron is a slit in the stage represented by figs. 8, 21, and 17, and is a spacious cavity in the stage represented by figs. 9, 13, and 19, it seems likely that the growth of the blastoporic lip is rendered evident, not only by the amount of white yolk plug covered, but also by the arching up of the dorsal roof of the archenteron. But, on the other hand, the arching up may be in great part, if not eutirely, * due to its own interstitial growth ; though I do not think this is likely, for it seems to me to require the thrusting energy of the blastoporic lip to account for the obliteration of the seg- mentation cavity. If Schultze’s idea of the apparent overgrowth of the white yolk plug being due to a rolling inwards of the white pole were correct, ought not the ventral end of the segmentation cavity to become obliterated before the dorsal? But it is the dorsal part that first disappears. VOL. 87, PART 2,.—NEW SER. Q 242 RICHARD ASSHETON. DESCRIPTION OF PLATES 23 & 24, sueate Mr. Richard Assheton’s paper ‘On the Growth in Length of the Frog Embryo.” CompLete List oF REFERENCE LETTERS. A. Anterior end. aa. Small cells in floor of segmentation cavity. 47. d. Dorsal lip of blastopore. 4/. v. Ventral lip of blastopore. D. Dorsal surface. exo. Exovate. P. Posteriorend. sg. Segmentation cavity. y. Dorsal lip of blastopore. z. Floor of segmentation cavity. z. Yolk-cells, which will be overgrown by dorsal lip of blastopore. y. p. Yolk-plug. x. Spot at which the blastopore commences. PLATE 23. Fic. 1.—a—d. Ovum pricked in centre of white between the developing lips of the blastopore. Fie. 2.—a—d. Course taken by a natural mark on the white pole. Fic. 3.—a—g. A mark was made near the centre of the white pole. Fie. 4.—a, 6. Ovum was marked in centre of blastopore when first out- lined. Fic. 5.—a—d. Three embryos marked at the cross in 5a. In 4 the mark was on the side; in ¢, ventral to blastopore; in d, in which it was very indistinct, at the side of the blastopore. Fic. 6.—a, 4. This embryo was marked at the sides of blastopore. PLATE 24. Fic. 7.—Frog embryo before appearance of blastopore. The white (or posterior) pole is placed to the right of the observer. Fic. 8.—Frog embryo at the time of the commencemeut of the blasto- pore, which is shown as a dark crescentic groove. The dorsal cap represents that part of the epiblast which will form the anterior part of the neural plate. Fic. 9.—Frog embryo at the moment of the completion of the ventral lip of the blastopore. The dorsal lip has grown over the white pole through an are of 50° to 60°. The dotted line indicates that part of the embryo supposed to be derived from the secondary area of proliferation. The arrows indicate from which portion of the rim the respective parts have been formed. Fic. 10.—Frog embryo at the time of first appearance of neural plate, visible only in sections. The blastopore is much reduced. The ventral lips GROWTH IN LENGTH OF THE FROG EMBRYO. 243 have closed more rapidly than the dorsal. Dotted line and arrows indicate same features as in Fig. 9. Fie. 11.—Frog embryo when the neural plate is a conspicuous object ex- ternally, and is deeply grooved. The embryo has become very distinctly elongated, owing to the growth due to the secondary area of proliferation. This was drawn with a camera. The dotted line and arrows represent my own interpretation of the facts as in the preceding figures. Fic. 12.—A section of the same stage as Fig. 8, semi-diagrammatic. The dotted line indicates the location of the split amongst the cells, whereby the archenteron is supposed to have originated and become prolonged forwards. Fic. 13.—A diagram of a stage intermediate between Figs.9 and 10. The portion of the gut cavity indicated by the dotted line represents that part which I suppose to be formed by the splitting amidst the yolk-cells. The future continuation of this slit is represented by a prolongation ventral- wards of the dotted line. The posterior part of the roof of the gut cavity, marked with small dots, is that part formed by the active growth of the dorsal lip of the blastopore, which is shown by the diagonal shading. Fie. 14.—A diagram of a later stage, such as Fig. 11. In this the seg- mentation cavity has become entirely obliterated. The secondary area of proliferation is completed and active all round the blastopore. Fie. 15.—A diagram to show the sequence of the horizontal furrows during segmentation of the frog’s egg. Fig. 16.—A diagram to show in which direction the process of segmentation will incur least resistance, on the supposition that the yolk is distributed as indicated by the intensity of the shading. Fics. 17—20.—Figures of the frog’s egg during the formation of the blastopore, to show which part of the surface of the ovum forms the yolk-plug. Lach figure is arranged in the same position. Fic. 21.—A section of a frog’s egg of a stage intermediate between Figs. 8 and 9. The ovum was drawn with camera. 4 ’ CDW i pila’ | peer © 20-3) =o Libihe © re tu Laat! ‘ i : ‘? > . wep ji Ley hia) Bis yb Sihl LF ln § MMi Whit. Melo’ fae yt . Oks OTHER DOAN Mae Yo HE a BrwoRe avihei xaeuw baw aae a dwelt od) pale aiden saa lesa f ped iz aa ajalltast one divide «egndite ou ae lituadd di oa ty Serene eee: Ott Eitomityil (ie ens rae wl ergyetare +] Letting these BP time ome lt a clini: Vee cain unt Woe vk xf vivig | ots is a pre Waa ba Fr owt faven. Ay’ tee 4. OSes wae NAT yeas gmt yay, tae | orks et t slp weanaly = ior iim DL aae A awe an boot gwar helt Wind Wiel 7 el pier Sin nwo |e le all ae ek eee ON aw iit sia ine Uil M) Tis Ti Me MOLW pea « Wer, Ae late 0 in die ack Gh ae ey Aleta: AAR eae een cee yi) au, afi } in WA a oe arelarla ? ite: Th we Vie wt Ai eOL oO Pai? ‘OPE iat! vue cone al ae ijtel yi ailicaeiaue @ ie Poverty ab Ail etl? Wy Tact tlh ow yy a, 10" Toe" * YA Jaga Sale wh +,/ale 1» Lt ea Aitl fo A eoie ovninin vl ey Tpadyy ae at rab ie ' Sede ghee ; get tal ative &ih ar. i +. or ) VIEL) POM |r ALLE te wiry ily as “ ah} an Hum rs i Te in irae “et a Ws ris eraluie O46 orl) chal otha Tit ie ator: (eu: BES ert * » = -_ ‘Ss ~ *) ares 4 er Olah fie. HObsiyoite® a1) phos Ce Pn ee aint ue A Ai Paw te B47 b q' baa) 7 iy i aye * ae, PEN Viog¢ Loe : ] qa ¥ cul ot yb? a ike Bie Ter, Ge Via! ‘en a . byt fi r. bl—.0S* VE Spy tt es knee Mewar! crib Wo ie Stang od feel eno » ity ‘f Tie L * i= ier hee wy han iiiabh otha aa. MERA Jewel sale 4 ie ih ney eet ferrrah “ s Abeta OF w& W had a ated wiht J uk _ ‘ ; Ln | i my ; : i ; af Bs JAA —E pte, aa) ss - - ie 7 . * . 7 7 4 (4 at a a. - t 169 ORT OG Ar GO 09 VARIATION OF TENTACULOCYSTS OF AURELIA AURITA. 245 On the Variation of the Tentaculocysts of Aurelia aurita, By Edward T. Browne, B.A., University College, London. With Plate 25. Ir was a suggestion from Professor Weldon that led me to examine a large number of specimens of the ephyre and adult stage of Aurelia aurita for the purpose of finding out the variation in the number of tentaculocysts, and if a variation occurred among the ephyrz to see how far it affected the adults. All the specimens were collected and preserved at Plymouth by the officials of the Marine Biological Association, and I sincerely thank the director, Mr. Edward J. Bles, for the loan of so many specimens. The ephyrz are divided into two sets; the first collected during the spring of 1893, the second specially obtained for me during the spring of 1894. The ephyra of Aurelia normally has eight arms, each bear- ing a tentaculocyst, four perradial bundles of gastric filaments, and four mouth lappets. The first table gives the numerical variation of the tentacu- locysts of 8359 specimens collected in 1893, VOL. 37, PART 3,—NEW SER, R 246 EDWARD T. BROWNE. TaBLe I, The Numerical Variation of the Tentaculocysts of 359 Ephyre collected in 18938. benteanloestta qpeeenaia: Percentage. Six 5 : < : 4 : - cy eet Seven : ; : 8 : . 2) 33 Hight (normal) F . 278 : : « Ts Nine = . a 29 A : 5 Gel Ten). ‘ : eos ; : 2 Eleven : ; 19 3 ‘ « 1333 Twelve ; , a ae ‘ : a oe Thirteen ; ; ; 3 : , = 10:6 It will be seen from this table that no less than 81 speci- mens (22°6 per cent.) are abnormal in possessing more or less than eight tentaculocysts, and that the range of variation extends from six to thirteen tentaculocysts. There are only 12 specimens (3°3 per cent.) with less than eight tentaculo- cysts, and the remaining 69 specimens (19 per cent.) are above the normal number. TasueE II. The Numerical Variation of the Tentaculocysts of 1156 Ephyre collected in 1894. Number of Number of tentaculocysts. specimens. Percentage. Five . ‘ . ‘ 1 : : _ — IK vs : : ; 6 , J 5 OS Seven ; , PINS 5 5 2 ow Hight (normal) © . . 883 , : Bye (so! Nine. : F s sb : e os eM Ten . : : 5 alll . : A Eleven ‘ 5 100 : : eae (21 b Twelve : : ara 4 : 5 .) ae Thirteen 5 : F 3 : : » PAQS Fourteen : s ; i : é oo The second table shows in detail the variation of the tentaculocysts of 1156 specimens collected in 1894. On VARIATION OF TENTACULOCYSTS OF AURELIA AURITA. 247 comparing it with the first table it will be seen that the percentage of abnormal specimens is nearly the same. In the first set 22°6 per cent., and in the second set 20°9 per cent. of the ephyrz are abnormal. The decrease is mainly due to a falling off in the number of specimens with twelve tentaculocysts amounting to 23 per cent. By taking a larger number of specimens the range of varia- tion has extended from five to fourteen tentaculocysts, but only one specimen of each of the two extremes has been found. The ephyra with five tentaculocysts (fig. 1) has four perra- dial arms, equal in size; but the fifth is interradial and about half the size of the other arms. The variation in the number of tentaculocysts does not affect the other organs of the body, which may vary inde- pendently of one another. Two specimens have only three bundles of gastric filaments instead of the normal four; both have four mouth lappets; but one (fig. 2) of them has six and the other seven tenta- culocysts. Six specimens have six bundles of gastric filaments. and six mouth lappets; three possess eleven tentaculocysts (fig. 8) and the others have twelve. A few curious abnormal growths of the arms were also observed. One specimen (fig. 4) has a perfect double arm with two tentaculocysts, like two arms united together. Another specimen (fig. 5) shows a bifurcation of an arm, each branch terminating with a tentaculocyst. Perhaps the most interesting monstrosity is that which occurs in a specimen (figs. 6 and 7) with a large outgrowth on the aboral side of the umbrella. The outgrowth has two arms, and one of them bears a tentaculocyst. There are also seven other arms, with tentaculocysts, in the normal position and a vacant place for two more. Adult Aurelia. The adult specimens of Aurelia were also collected at Plymouth during the summer of 1894, and belong to the 248 EDWARD T. BROWNE. same generation as the ephyre taken in the spring of that year. The umbrella of these specimens varied from 14 to 24 inches in diameter. The tentaculocysts of 383 specimens were examined, and the number possessed by each specimen is recorded in Table ITT. Tasre ITI. The Numerical Variations of the Tentaculocysts of 383 Adult Aurelia collected in 1894. Number of Number of tentaculocysts. specimens. Percentage. SILK) a. ‘ : 2 : . am “GE Seven : : 5 ale d ; sa AEE Hight (normal) : ~ 296 : : dS Nine. : 3 oy EGS 5 4 to 8656 Ten . 3 16 4°] Eleven 10 ‘ : 7 oO Twelve . : ‘ 7 ; ‘ sy 8 Thirteen 0 = Fourteen 0 oe Fifteen il 2 There are 87 specimens (22°8 per cent.) with a variation in the number of tentaculocysts, 20 having less than the normal number and 67 showing an excess. On comparing the abnormal number of tentaculocysts of the adults with those of the ephyra stage, it will be seen from the percentages that there is only a slight difference. The ephyre have 22°6 per cent. abnormal in 1893, and 20:9 per cent. in 1894; the adults show 22°8 per cent. Itis clear from these figures that the abnormal ephyre do not appear to suffer from their abvormality, but are able to reach in safety the adult stage. The figures also show a slight increase of ab- normal forms in the adult stage. This may be due to an insufficient number of adult specimens; the small number is due to their scarcity at Plymouth. On comparing the 359 ephyre taken in 18938 and the 383 adult specimens taken in 1894, it will be seen that the per- VARIATION OF TENTACULOCYSTS OF AURELIA AURITA. 249 centages of abnormality are very close for nearly the same number of specimens. It is probable that ifa thousand adults could have been obtained at Plymouth the percentage of ab- normal forms might have been closer than 2 per cent. of the ephyre taken in 1894, The adult specimens were taken at random out of large jars; and it is interesting to note how close the percentage of abnormality of each complete hundred comes to the mean abnormality. The first hundred showed 23 per cent., the second 22 per cent., and the third hundred 24 per cent. of abnormal forms. An examination of the spe- cimens does not show that any particular position on the margin of the umbrella is favoured either by an increase or decrease of the tentaculocysts. Eighteen specimens possess seven tentaculocysts, and in eleven of these the missing tentaculocyst is a perradial one, and in seven it is adradial. The presence of an extra tenta- culocyst may either affect the symmetry of a single quadrant or one half of the umbrella, and in a few cases by being very close to another not even upset the symmetry. Ten specimens with nine tentaculocysts show that the extra tentaculocyst is in one quadrant of the umbrella, and thirteen specimens have one half of the umbrella containing five tentaculocysts about equal distances apart, and the other half possessing the normal four. Five specimens have eight tentaculocysts occupying their normal positions, and an extra one only separated from a normal one by a few marginal tentacles. When the tentaculocysts exceed nine in a specimen their position is by no means constant, and a different arrange- ment occurs in almost every specimen. In some the tentacu- locysts are about equal distances apart, and in others one half of the umbrella contains the greater number. A few specimens have three tentaculocysts very close to- gether, and usually separated by a few marginal tentacles. One specimen of an adult Aurelia has fifteen tentaculocysts with the normal number of genital pouches and arms. This exceeds the maximum number reached among the ephyre. None of the adults have, however, either thirteen or fourteen 250 EDWARD T. BROWNE. tentaculocysts. Their absence is probably due to the exami- nation of an insufficient number of specimens. A variation in the number of tentaculocysts does not inter- fere with the other organs of the body. There appears to be a correlated variation between the number of genital pouches and buccal arms, and eight specimens show it. One specimen has three genital pouches, three buccal arms, and nine tentaculocysts. Three specimens have three genital pouches, three buccal arms, and each one has traces of a fourth genital pouch and a fourth arm. Two have eight tentaculocysts, and the other has ten tentaculocysts. One specimen has five genital pouches, five buccal arms, and eight tentaculocysts. Three specimens have six genital pouches, six buccal arms; two have eleven tentaculocysts, and one has twelve tentaculocysts. I have not given drawings of these specimens, as somewhat similar forms have already been figured by Ehrenberg (1) and Romanes (2). Mr. Bateson, in his recent book, ‘ Mate- rials for the Study of Variation,’ gives abstracts and figures from the papers of Ehrenberg and Romanes. He also gives a table which shows that 26 specimens (1°49 per cent.) out of 1763 adult Aurelia, washed ashore on the Northumberland coast in 1892, have more or less than four genital pouches. The Plymouth specimens show 2°08 per cent. with an abnormal number of genital pouches. It may be difficult to notice in a few years what effect the numerical variability of the tentaculocysts has upon the adult Aurelia. The variation shows a tendency for an increase of the tentaculocysts, since the specimens displaying an increase are about three times as numerous as those possessing a diminished number. Whether this will change the present characteristic features of the species or not can only be found out by examining the ephyre and adults at long intervals of time, and comparing the results with previous records. VARIATION OF TENTACULOOCYSTS OF AURELIA AURITA. 251 References. 1. Enrensere, C. G., 1834.—‘Abh. K. Ak. Wiss.,’ Berlin, pp. 199—202, plates. 2. Romanss, G., 1876.—! Journ. Linn. Soc. Zool.,’ vol. xii, p. 528 vol. xiii, p- 190, plates xv, xvi. 8. Bateson, W., 1894.—‘ Materials for the Study of Variation,’ pp. 426— 429, fig. 128. DESCRIPTION OF PLATE 25, Illustrating Mr. Edward T. Browne’s paper ‘‘ On the Varia- tion of the Tentaculocysts of Aurelia aurita.” Fie. 1.—Ephyra with five arms. Aboral view. xX 25. Fig. 2.—Ephyra with six arms, three bundles of gastric filaments, and four mouth lappets. Aboral view. x 35. Fic. 3.—Ephyra with eleven arms, six bundles of gastric filaments, and six mouth lappets. Oral view. x 25. Fic. 4.—Ephyra with a perfect double arm, seven tentaculocysts, four bundles of gastric filaments, and four mouth lappets. Aboral view. x 40. Fic. 5.—A portion of the umbrella of an ephyra, showing a bifurcation of anarm. X 40. Fic. 6.—Ephyra with an outgrowth of two arms from the aboral side of the umbrella. Aboral view. x 35. Fic. 7.—A lateral view of Fig. 6. xX 25. . 3 PS oh ie hs eGR REY - : A = sel oe Pro st Os : = jy A iasey To “nt ‘D a) Ya) a= a hy — — ud By oe 4d onmey S07? pf) >i ay son dort Cee ee Walivne otha A do mesmo) et ietal s . Py deli lee ge EA io ee lo urieiegh A 6 Shy “iv ¢ tAjsalb nav Ga ul| gee wey DALe i ae 3 = . , are ie “iG A, ne ed Baa. Ci 90a oO |: yess prea ato hs) ora), was fifient we Wy Woe eye A We Neen ifad s ge) eu, ie vyign & - Lh » ay? Mild @dlive GOTO ea bs a i’ / <> ~~ / - eh ’ ‘ i 7 i , ee , _ —_ 7 . as > - a 7 ; od = = er é 5 = ie Jal Pr) ¥ 7 2 a > ‘ _ be oO 7 ¢ Ly : 7 i . * ° , n i . ene ann: San Ant si ia Or WaT aioe 1 Pulpoha al yyy Wut iit - COQ ETS, Th anvrataae 7 —— tr ‘shite i bd ; aa ; me, 1 ot ah i) ‘ — —_ «¢ ' , | -. rs —_ a f ' — * é aot eit ‘ L_— ad ! ' > 9 Par MOUTH-PARTS OF THE CYPRIS-STAGE OF BALANUS. 269 On the Mouth-parts of the Cypris-stage of Balanus. By Theo. T. Groom, F.Z.S., Late Scholar of St. John’s College, Cambridge. With Plate 29. Tuer buccal mass of all ordinary Cirripedes consists, as is well known, of three pairs of jaws, together with the labrum and palps. The form and disposition of these mouth-organs in the various genera of the Thoracica has been well investi- gated by Darwin, and the differences observed between the different forms in respect to these parts have been utilised for purposes of classification. The morphological significance of the jaws has, however, never been satisfactorily ascertained, and the current views are all practically based upon supposi- tions, the actual development having been traced in no case. Darwin (Nos. 1 and 2) regarded the frontal filaments seen in the larva as the first pair of appendages, and taking the eyes to mark the first segment of the head, and believing that the prehensile antennules arose within the fronto-lateral horns, and that the three pairs of jaws arose in front of the Nauplius appendages, regarded the mandibles of the adult as the appen- dages of the fourth segment of the typical Crustacean, and the two pairs of maxillz as the fifth and sixth respectively. The eyes, however, are now no longer regarded as modified appendages ; the frontal filaments, too, have been more cor- rectly regarded as sense-organs, and the fronto-lateral horns have been proved by Claus and others to be glandular processes of the carapace. 270 THEO. T. GROOM. Krohn, Willemoes-Suhm, and Lang (Nos. 3, 7, and 9) have proved that the prehensile antennules of the Cypris-stage in the Thoracica arise from the first pair of Nauplius appendages (antennules), and Delage (No. 10) has shown the same for Sacculina. With reference to the fate of the remaining Nauplius appendages, Metschnikoff, Willemoes-Suhm, and Lang (Nos. 8, 7, and 9) believed that both pairs were lost, and the first- mentioned supposed that the mandibles and two pairs of maxillz were all new structures formed inside a fourth pair of appendages seen behind the third pair of Nauplius appendages. Claus, on the other hand, supposed (No. 8) from the analogy of other Crustacea that the mandibles of the Cypris-stage arose from the third pair of Nauplius appendages, the latter not being lost like the second pair, but greatly reduced in size. This observer did not, however, succeed in proving his point, and I shall attempt in the following remarks to show that this view is the correct one. The Cypris-stage, as is well known, possesses a well-defined buccal mass. This, on account of its concealment within the carapace, its small size, and delicate nature, as well as the close packing of the component parts, is difficult to make out satisfactorily. Darwin found in the Cypris-stage of Lepas australis “all the masticatory organs of a Cirripede in an immature condition.” Pagenstecher (No. 4) describes the mouth-organs of the Cypris-stage of Lepas pectinata after fixation as consisting of imperfect lobes and papille without bristles or teeth. Willemoes-Suhm (No. 7) did not succeed in satisfactorily separating the mouth-parts in Lepas fascicularis, but describes the buccal mass as consisting of “three parts all very rudimentary.” Claus (No. 6) also describes three pairs of gnathites in an undetermined Cypris-stage. These had the form of simple outgrowths; the mandible was largest and connected with the labrum by a finger-shaped palp referable equally to the upper MOUTH-PARTS OF THE CYPRIS-STAGE OF BALANUS. 271 lip or to the mandible; the two hinder pairs Claus describes as giving the impression of a single pair of appendages. Of the two species of Balanus the Cypris-stage of which I have carefully examined, B. perforatus, on account of its minute size, is much less favourable for purposes of investiga- tion than B. balanoides. Repeated dissections of the pupa of the latter form have enabled me to form a tolerably good idea of the constitution of the buccal mass. Viewed from behind (fig. 3), a pair of appendages (mz?.), clearly identical in position and relations with the lower lip, or second pair of maxille of the adult, is recognisable. These are attached on each side behind along an oblique straight line, and are somewhat broader at the base than at the apex. Like the second maxille of the adult, they are closely applied to one another in the middle line. Seen in side-view (fig. 8) they are somewhat lanceolate in shape, the apex generally projecting somewhat forwards ; the hinder margin is convex, while the anterior is somewhat sinuous. Like the remaining mouth-parts, and like those of Lepas, as described by Pagen- stecher, they are simple lobes devoid of teeth or bristles. At the junction of these appendages with the first pair of maxillz may be seen a conical chitinous process (figs. 1, 3, 4, 6). The first maxille (mz!.) are situated externally to the second, and project somewhat behind the latter (fig. 1, cf. also fig. 11). They are similarly convex externally, but are simply concave on the inner side ; they lie in planes inclined outwards at a greater angle than the second maxille (cf. fig. 11). Like the remaining mouth-parts they are broadest externally, narrowing towards the centre. They do not attain quite the vertical height of the second maxille (fig. 1), The mandibles are large, and occupy the outer part of the buccal mass (figs. 1—3), being found externally and anteriorly to the first pair of maxille (figs. 1—3; cf. also fig. 11). The main portion of the mandible passes distally into an inwardly- curved process with a somewhat truncated end, giving the appen- dage much the form of the adult mandible or first maxilla, though, as already remarked, no teeth or hairs are present, Paresh THEO. T. GROOM. Anteriorly the basal portion of the mandible gives off a some- what inwardly-curved palp-like process (figs. 1—5, palp.). This has all the appearance of being an integral portion of the mandible, and is filled by a mass of developing muscle con- tinuous with that of the rest of the appendage, and running, together with the muscle of the two pairs of maxille, to the anterior part of the thorax (fig. 1). The anterior margin of the basal portion of the mandible (figs. 1 and 5) is apparently very short, though this is difficult to ascertain definitely, since in teased preparations the buccal mass commonly tears away from the head immediately in front of the palp; it is, thus, easy to see why Claus regarded the palp as belonging equally to the mandible or labrum. The evidence, too, afforded by the muscle-supply must be taken with caution, as Nussbaum (No. 11) describes a muscle supplying the mandible in the adult Conchoderma as having its origin in the palp. The evidence, however, afforded by the immature Cypris-stage to be described immediately appears to clearly show that the palp belongs to the mandible. The three pairs of gnathites, according to Claus, are situated beneath a labrum; the labrum, however, as a definite promi- nence, is exceedingly small at this stage in Balanus, and projects as a very minute lobe between the distal ends of the palps. The site of the future labrum is, nevertheless, recog- nisable as a broad area in front of and between the mandibles and their palps. The jaws are directed ventrally, so that this area is to be described as situated anteriorly to, rather than below the jaws. Some stages of an undetermined species of Balanus ob- tained from Messrs. Sinel and Hornell in Jersey in the spring of the present year (1894) were specially valuable as throwing light on the question of the fate of the Nauplius appendages, and on the homologies of the jaws of the adult. The antennules of the Nauplius, as already pointed out, have long been known to give rise to the prehensile antennules of the Cypris-stage, and these, according to Darwin, can often be recognised in the adult. The fate of the antenne of the Nauplius has not been MOUTH-PARTS OF THE OCYPRIS-STAGE OF BALANUS. 273 traced. Fig. 12 shows the position of the base of the antenna and its socket in the last Nauplius-stage of the Balanus in question. The socket is seen to lie at the side of the labrum, just behind that of the antennule ; the innermost part of the base of the mandible projects as a pivot, which fits into a slight indentation in the side of the labrum. The origin of the mandibles is well behind the mouth and attachment of the labrum. As the Nauplius prepares for the moult which trans- forms it into the Cypris-stage the body contracts, and many of the structures belonging to this latter stage can be seen beneath the cuticle. If the larva be now carefully dissected out of the cuticle, or be examined just after the moult, it shows peculiar features in which it approaches the Nauplius-condition, and which are not found in the perfect Cypris-stage. Figs. 9 and 10 show the larva in this stage. The edges of the carapace have not yet met in the mid-ventral line, and allow good views of the ventral surface of the animal. The prehensile antennules can be seen at the anterior end of the labrum, which, in most cases, is still recognisable, often as a well-developed structure. The sides of the labrum are indented, as in the Nauplius-stage, by the remnants of the antennz, which are now undergoing histolysis and absorption, but are easily recognisable. Behind these can be seen the mouth-parts of the Cypris-stage already well-formed. These consist of the mandibles with their palps, and the first and second pairs of maxille. In the more advanced stages the labrum and antenne are small (fig. 11) or have practically disappeared, so that the larva has attained the con- dition (so far as these organs are concerned) seen in the perfect Cypris-stage. A series of stages were obtained showing the gradual reduction of these parts. A highly important feature is that the mandibles, clearly forming with their palps a single appendage, can be detected on the site of the mandible of the Nauplius, behind the antennz and labrum. The labrum shows in some cases more or iess clear indications of the median and lateral lobes found in the Nauplius at the distal extremity. It is clear, then, that the palps can neither be regarded as equivalent to the antenne, nor to the lateral lobes 274 THEO. T. GROOM. of the labrum of the Nauplius, but belong to the mandibles. Judging by analogy the palp will represent the ramus of this appendage, the mandible proper being the gnathobase. Of the two pairs of maxill, the first pair are indicated after the first moult undergone by the Nauplius by a row of bristles which has been termed the extra-maxillary are (No. 12). Ata later stage this becomes a small foliaceous appendage (fig. 13) provided with a number of bristles, and already correctly regarded by Claus as an early condition of the “ outer maxilla’? (No. 8). Inside this the first maxilla of the Cypris- stage clearly arises (fig. 13). The second pair of maxille appear later,simultaneously with the six pairsof cirri; they arise, as described by Claus, just in front of the first pair of the latter, and are clearly serially homologous with the thoracic appen- dages. Further details as to the origin of the two pairs of maxille will be given on a future occasion. It may be regarded as tolerably certain from what has been said above that : (1) The antennz of the Nauplius become definitely lost with the moult resulting in the production of the Cypris- stage. (2) The biramous mandibles of the Nauplius become reduced at the same time to the small mandibles, the ramus being probably preserved in the form of the small palp. (3) The first pair of maxille arise behind the mandibles, and at a later date, as a small pair of foliaceous appendages. (4) The second pair of maxille arise still later just in front of the first pair of thoracic legs (cirri). The mandibles, first maxille and second maxille are accord- ingly developed consecutively in the order named. They cannot be regarded as parts of a single or fourth pair of appendages as Metschnikoff maintains, neither can the two pairs of maxille be regarded as outer and inner parts of a single pair of appendages as Claus suggests. The mandibles of the Nauplius are not lost as many authors have maintained, but give rise, as Claus supposes, to the mandibles of the Cypris-stage. The jaws represent, then, the third, fourth, MOUTH-PARTS OF THE OYPRIS-STAGE OF BALANUS. 275 and fifth pairs of appendages of the typical Crustacean series. This conclusion is quite in harmony with that of Nussbaum, as deduced from his beautiful anatomical investigations of the adult. Cirripede, “ Die Anordnung der Musculatur spricht dafiir, dass jeder der drei Kiefer aus einem Beinpaare hervor- gegangen ist” (No. 11). In the fate of the mandibles of the Nauplius and in the constitution of the head of the adult the Cirripedia thus con- form to what is recognised as typical among other Crustacea, BIBLIOGRAPHY. 1. C. Darwin.—‘A Monograph of the Sub-class Cirripedia,’ vol. i, “ Lepadide,” 1851. ; 2. C. Darwin.—‘A Monograph of the Sub-class Cirripedia,’ vol. ii, ‘** Balanide, Verrucide, &.,” 1854. 3. A. Kréun.—“ Beobachtungen iiber die Entwickelung der Cirripedien,” ‘Arch. f. Naturgesch.,’ vol. xxv, 1859. 4. A. PacEnstEcHER.—“ Beitrage zur Anatomie und Entwickelungsges- chichte von Lepas pectinata,” ‘ Zeitschrift f. wiss. Zool.,’ vol. xiii, 1868. 5. E. Metscunrxorr.—“ Ueber die. Entwickelung von Balanus bala- noides,” ‘ Sitzungsb. der Versamml. Deutsch. Naturf. zu Hannover,’ 1865. . C. Craus.—‘ Die Cypris-ahnliche Larva der Cirripedien und ihre Ver- wandlung in das festsitzende Thier.,’ Marburg and Leipzig, 1869. fon) 7. R. von WitLemors-Sunm.— On the Development of Lepas fascicu- laris and the Archizoéa of Cirripedia,” ‘Phil. Traus. Roy. Soc.,’ vol. elxvi, 1876. 8. C. Cuaus.— Untersuchungen zur Erforschung der genealogischen Grundlage des Crustaceen-systems,’ Wien, 1876. 9. Arnotp Lanc.—‘‘ Ueber die Metamorphose der Nauplius-Larven von Balanus,” ‘Mittheilungen der Aargauischen Naturforschenden Gesell- schaft,’ vol. i, 1863—1877. 10. Yves Dezace.—* Evolution de la Sacculine,” ‘ Archives de Zoologie expérimentale,’ vol. ii, 1884. 11. M. Nusspaum.—‘Anatomische Studien au Californischen Cirripedien, Bonn, 1890. 12. T. T. Groom.—“ On the Early Development of Cirripedia,”’ ‘Phil, Trans. Roy. Soc.,’ vol. clxxxv, 1894. 276 THEO. T. GROOM. EXPLANATION OF PLATE 29, Illustrating Mr. Theo. T. Groom’s paper, ‘On the Mouth- parts of the Cypris-stage of Balanus.,” Fie. 1.—Side view of buccal mass of Cypris-stage of Balanus balanoides. Mandible. Mandible with (palp) its palp. mz. and mz. First and second maxilla, showing chitinous process between their bases. duce. m. musi. Muscles supplying buccal mass. Thoracic muscles. Muscles filling up the first segment of the thorax. Fic. 2.—View of same from in front. Letters as in Fig. 1. Also, adductor. Adductor muscle of carapace. Fic. 3.—View of same from behind. Fie. 4.—Similar view, the form of the appendages being shown by trans- parency. Fie. 5.—Mandible and its palp. Fic. 6.—View of first maxilla from outside. Fic. 7.—View of same from side. Fig. 8.—View of second maxille from side. Fie. 9.—Side view of an immature Cypris-stage of an undetermined species of Balanus from Jersey. adductor. Adductor muscle of carapace. anf}. Antennule. azé?. Antenna. caudal app. Caudal appendage. compd. eye. Compound eye. /r/. aperture. Aperture of fronto-lateral gland. ma. First maxille. ma. Second maxille. palp. Palp of mandible. Fic. 10.—Ventral view of similar larva. Lettering as in Fig. 9. Also, ant', muscles. Muscles to antennule. cp. edge. Edge of carapace. Fig. 11.—Ventral view of mouth-organs of a larva in a somewhat more advanced stage. The labrum is now greatly reduced, and the sites of the antenne barely recognisable. Fie. 12.—View of labrum and bases of antennules, and antenne with their sockets (ant'. sock., ant. sock.) of the last Nauplius-stage of a species of Balanus from Jersey. mandible sock. Socket of mandible. Fie. 13.—View of the developing maxille (mz'. and mz*.) of the Cypris- stage, as seen ina similar Nauplius. The small setigerous first maxille are external, while the two pairs of pupal maxille are underneath the cuticle; the first pair of these has the tip slightly withdrawn from the external maxillz, inside which they were formed, STUDY OF COCCIDIA MET WITH IN MICE. 277 A Study of Coccidia met with in Mice. By J. Jackson Clarke, M.B.Lond. With Plate 30. Nearty all that has been published on Coccidia of mice is contained in the writings of Kimer! and Th. Smith.? Ludwig Pfeiffer,® in reference to what these authors have said on this subject, concludes: ‘“‘ Unsern frithern Auseinandersetzung nach, handelt es sich hier um das Schwarmercystenstadium einer Coccidie: das dauercystenstadium ist unbekannt.” In order to supply data which may fill the gap thus indicated, I venture to submit the following observations. A white mouse, kept in a place previously occupied by rabbits, was found dead. Dissection revealed the presence of large numbers of Coccidia in every part of the alimentary tract beyond the cardiac opening of the stomach. There was no other abnormal feature. Sections of the intestine showed an altered condition of the epithelium, but although decom- position had not begun when the tissue was fixed I could not, apart from the encapsuled parasites collected in the lumen of the gut, have been certain that epithelial infection by Sporozoa was present, so soon after death do some of the intracellular protozoa lose their characters. The contents of the intestine were spread on a clean slide, 1 Kimer, ‘ Ueber die Hi- oder Kugelformigen sogenannten Psorospermien, &e.,’ Wiirzburg, 1870. 2 Th. Smith, Washington, ‘Journal of Comp. Med. and Surg.,’ 1889. 3 L. Pfeiffer, ‘ Protozoen als Krankheitserriger,’ 1891, p. 57. VOL. 37, PART 3,—NEW SER. T 278 J. JACKSON CLARKE. placed on damp blotting-paper in a Petri’s dish,! and kept at room temperature (May—June). The ripe free parasites were uniformly much smaller (aver- aging 18 x 13 mw) than the Coccidium oviforme of the rabbit (averaging 32 x 22 mu). They were also of a more rounded shape. When first examined the capsules of the free parasites were completely filled by granular protoplasm, having the usual dense, round, central body ; they were examined on successive days for a fortnight, and I may briefly indicate the results by saying that with the modification in size previously alluded to they underwent the same changes as does C. ovi- forme of the rabbit when placed under similar circumstances. Thus on the sixth day most of the parasites had subdivided (fig. 2) into four granular spheres, of which some possessed a clear oval corpuscle placed as far as possible from the point of meeting of the segments (fig. 3). On the eighth day the sub- divisions (sporogonia) had assumed an oval form, and some pre- sented a capsule and a differentiation into spores and granular residual matter (nucléus de reliquat). Many of the spores showed the dumb-bell form described by Leuckart in C. ovi- forme of the rabbit (see fig. 5), and in one or two cases I was able to see the division of the C-shaped body into two comma- shaped spores, as described by Balbiani. This arrangement of lasting spores is, as in the case of the rabbit’s Coccidium, often departed from, and appearances such as are depicted in fig. 6 occur frequently. In order to have material for histological study, and to test the time required for the manifestation of the disease, a young healthy mouse, whose faeces were found to be free from Coc- cidia, was fed with bread-sop containing some of the material taken from the Petri dish after twenty-two days’ exposure to air.” 1 This consists of two shallow glass dishes, one of which is larger than the other, and when inverted forms a cover which excludes dust from the smaller one. 2 T have given the days on which the various appearances were first noted, and I do not wish to convey an impression that on any given day all the STUDY OF COCCIDIA MET WITH IN MICE. 279 No change was noticed in the animal until six days after the parasites were administered. On this day its coat was noticed to be rough, the abdomen was distended, and the thighs some- what drawn up towards the belly, so that the animal’s gait was stiff. The stools were softer than normal. The next day blood! was noticed about the animal’s anus, and this sanguineous discharge contained coccidia of the same dimen- sions as those described above. On the seventh day the animal was distinctly ill and suffering, so it was killed. After death the stomach was found to be distended, the contents consisting chiefly of parasites, many of which showed the formation of sickle-spores* as shown in fig. 8. The large and small intes- tine contained practically little beside ripe Coccidia. The stomach, intestines, and portions of the liver and kidneys were fixed in saturated solution of corrosive sublimate and hardened in the usual way. After embedding in paraffin, sections were cut and stained in Ehrlich’s acid hematoxylin and eosin (Gribler’s wasserléslich). Histological Examination. In the stomach the glands of the cardiac end contained free Coccidia, many of which were encapsuled and filled with small swarm-spores as shown in fig. 8; but the most striking feature consisted in masses of minute bodies exactly resembling the swarm-spores within such capsules, but stained and lying free on the surface of the mucous membrane, and distending the ducts coccidia presented the same appearance. On the contrary, up to nineteen days after the parasites were placed in the moist chamber, nearly all the phases could be observed at the same time; but on the date last named all the parasites were subdivided into sporocysts, each of which comprised two spores and a nucleus de réliquat, as the residual body has been termed by Aimé Schneider. 1 This phenomenon is of interest in connection with an observation recorded by Hess (‘Schweitz Archiv fiir Thier.,’ Ziirich, 1892, vol. xxxiv, p. 125), to the effect that the “Rothe Ruhr” of cattle is associated with the presence of a coccidium. 2 Compare R. Pfeiffer, ‘ Beitrage zur Protozoon-Forschung,’ Berlin, 1892. 280 J. JACKSON CLARKE. of the glands. In the latter situation the accumulated spores, devoid of any containing capsule, had an appearance which recalled the well-known Sarcosporidia—a resemblance which will be referred to again in a subsequent part of this paper. That these small free bodies were swarm-spores derived from the coccidia was shown conclusively, I think, by a close exa- mination of the epithelial cells against which they lay. Many of the cross-sections of the pyloric glands had the appearance shown in fig. 9, where some of the lining cells contain one or several minute bodies, many of the same size and appearance as the small free bodies lying in the lumen of the gland. From the minute intracellular parasites which lay in the cell-protoplasm a gradation of forms could be traced up to the encapsuled swarm-sporing parasites; but in the stomach the full-sized intracellular parasites were very few in number compared with those about to be described in the intestine. This was probably due to the majority of the parasites having subdivided into swarm-spores without the formation of a cap- sule, a process which occurs in C. oviforme in the acute disease, and which will be described immediately in the small intestine in the mouse. Small intestine.—Sections taken from various parts all presented similar features. The epithelial cells of the glands of Lieberkiihn were almost every one infected, and in the majority of them the parasites were spherical with numerous “corps albuminoides de réserve,’ and a small central body which stained with eosin, and in the lumen of the gut lay the encapsuled free parasites. Sections of the small intestine showed that there the epithelial cells were as much infested as in the stomach, but the parasites presented a somewhat different aspect, as is shown in fig. 10, which represents a cross-section of one of the crypts. The epithelial cells show signs of having pro- liferated, being in places three deep instead of forming a single layer, and the sickle-shaped swarm-spores are wanting. Most of the intracellular parasites are in the shape of granular spheres, the granules staining with eosin; but some, such as STUDY OF COCCIDIA MET WITH IN MICE. 281 the one marked a, have larger granules! staining deeply with hematoxylin. One parasite (6) has surrounding it a thick capsule, and is in the same phase as that shown in fig. 1. The granules and the central body took the eosin and not the hematoxylin of the stain. Other parasites were larger than this, but still devoid of a capsule (c), and represented, I be- lieve, an intermediate phase between the common form (/) and the phase represented by the bodies d and e, which measured respectively 34 x 19 and 35 x 22. The marginal part is represented as seen in optical section, and is marked by a close- set series of short rods which stained a deep blue with hema- toxylin. The lighter dots in the bodies represent the surface rods seen out of focus, and the series of darker dots in their inner parts are represented as seen in focus, though they are placed on the surface of the parasites. I have met with para- sites in this phase in the liver and intestine of rabbits suffering from acute coccidial disease, and as I have nowhere seen a full account of this particular phase, I may introduce two figures (11 and 12) taken from sections of a rabbit’s liver. These short rods I believe to be of the nature of chromosomes, and the parasites in this phase to be almost wholly nuclear in con- stitution. I once believed them to be wholly nuclear, until lately, at the suggestion of Professor Marcus Hartog (Cork), I looked for a layer of protoplasm, which I have found to be present, but so thin is this layer that it is difficult to make out. Some of the modes of subdivision of this phase of the parasite are illustrated in fig. 10, d, and figs. 11 and 12. The large intestine contained innumerable coccidia both within the epithelial cells and free, but I could find none in the phase of sickle-formation as in the stomach, nor were there any with peripheral rods as in the small intestine. The liver and kidneys showed no abnormal features. I have not yet been able to procure Eimer’s original paper, but it is freely abstracted by Leuckart.* I have been able to 1 These granules, like those staining with eosin, are in some cases “corps albuminoides,” i. e. stored food material. 2 Leuckart, ‘ Parasites of Man,’ Hoyle’s translation, 1886, pp. 197 and 219. 282 J. JACKSON CLARKE. conclude that the parasites of Eimer (Eimeria, A. Schneider) are the same as those described in this paper. The chief grounds for this conclusion consist in the round form Eimer refers to in many of his parasites, and to their dimeusions, 18H xX 18 and 26 x 16, approaching more nearly to the parasites described above than to C. oviforme in the rabbit, where the intestinal parasites average, I have found, 32 u x 22 p. In order to test the transmutability of C. oviforme into Eimeria, I fed two young white mice whose excrement I had found to be free from coccidia with food mixed with the dung of a rabbit containing spore-ripe C. oviforme. The result of these experiments was negative. A third experiment gave a positive result. A very old white mouse, whose feces had been found to contain no coccidia, was given food mixed with some material which had been taken from the cecum of a rabbit, and kept for two months in a moist chamber. There were many spore-ripe coccidia in this material. Five days after the administration of the parasites the mouse died, and its alimentary tract was found to be filled with coccidia like those in the other mice referred to above. I may now conclude these observations by indicating the conclusions to which they point. Conclusions. 1. That the Sporozoa described above as they were found in white mice belong to the Coccidiidea, and, like C. oviforme, to the Disporez. 2. That they are probably identical with the Sporozoa de- scribed by Eimer in the intestines of mice, and named Eimeria by Schneider. 3. That Eimeria is probably only a variety of C. ovi- forme, and may be but a modification of C. oviforme (although two out of three experiments made with a view of solving the point gave negative results) determined by the STUDY OF COCCIDIA MET WITH IN MICE. 283 smaller dimensions of the epithelial cells of the intestine of the mouse as compared with those of the rabbit. 4. That the appearance of large numbers of swarm-spores in the gastric glands of the mouse is very similar to that pre- sented by the Sarcosporidia, and suggests that the latter is but one phase of a Sporozoon which may have in other phases a form resembling that of the Coccidia. A similar conclusion to this has been arrived at by L. Pfeiffer as the result of com- paring Klossia with the Sarcosporidia. October 3rd, 1894. EXPLANATION OF PLATE 30, Illustrating Mr. J. Jackson Clarke’s paper, “‘ A Study of Coccidia met with in Mice.” Fic. 1.—Ripe Coccidium, as seen with 4, 0. i. immediately after removal from the body of the mouse. Fig. 2.—Coccidium after six days in the moist chamber. Fic. 3.—The same after the same length of time. Fic. 4.—The same after eight days. Fie. 5.—The same, 10th day. Fic. 6.—The same, 12th day. Fie. 7.—Section of a duct of a cardiac gland (mouse’s stomach). Fie. §—Encapsuled parasite subdivided into sickle-shaped swarm-spores, examined fresh from the mouse’s stomach. Fic. 9.—Cross-section of duct of a pyloric gland. Fic. 10.—Cross-section of a gland of the small intestine of a mouse, show- ing two large naked intracellular parasites with close-set peripheral bars of chromatin. The darker points on the surface of the parasite appear to mark the commencement of subdivision. Fic. 11.—Naked and free coccidium from a section of a rabbit’s liver, showing the mode of subdivision of a naked parasite with peripheral chromatin bars. Fie. 12.—The same, showing another phase of subdivision. fi Neit 4 a tlan Al dametees MEA id | 2yi e FAt eee) on peti TAL an Paden a ay ie iAMiy nh bn ) Loe : > eonaye, ohhh gue i ? cee dickies i it x oF op 8Vty nee @ ’ 4 vs . ike Ch Qe yor il a 4S etry we \' : sifu aaa pde:! 44 Sr tr Le ve) ' re , ae Faé { ‘ f j | i Vid } Aare h “+ Py t ¢ " i | P : i i / j j a 1 FP y + fi z f . i ! aL apy ‘ # * - ih soa! POW Ban tty f ps i” i Japs h Moe Oe! ee 4 Se lle wo) yee r Shaye vi a nee AJL 7 ees abana ve Seba Tee itn Ae ae ivi ut) a ae Wr=4 ce ott aa — oe any iieae © a baie es nek my ; = ms ~~ Va 4 OBSERVATIONS ON VARIOUS SPOROZOA. 285 Observations on Various Sporozoa, By J. Jackson Clarke, M.B.Lond. With Plates 31, 32, and 33. Certain problems propounded in the last few years demand for their solution a closer study of the intimate structure of the sporozoa than has hitherto been found necessary. The varia- tions of nuclear form presented by these organisms call for especially close examination. For some of the higher members of the group this work has been well begun by Wolters,! who examined Clepsidrina blattarum, Monocystis magna and agilis in Lumbricus agricola, and Klossia in the kidney of snails. The description of the nuclear changes in the common gregarines of the earthworm is particularly complete, and since I have been able to confirm many of Wolters’s observations it may be of interest if they are briefly reviewed here. He found in Lumbricus agricola both Monocystis magna and agilis constantly present, and almost to the exclusion of other species. In both, at every stage, a distinct nucleus was present. In the former it was relatively large and oval, in the latter of a rounded form. Under compression the nuclear membrane ruptured, but the contents did not escape, and when the compression was removed the nucleus tended to return to its primitive form. Thus the nucleoplasm appeared to be of a solid structure. In M. agilis this was also observed. The nuclear membrane was found to 1 Max Wolters, ‘Die Conjugation und Sporenbildung bei Gregarinen,” ‘ Archiv fiir mikroskop. Anat.,’ p. 99, 1891. 286 J. JACKSON CLARKE. be strong and sharply defined, and in the young parasite there was a Single large nucleolus. Later the outline of the nucleus had become irregular and the nucleoli more numerous. In M. agilis, and in one instance in M. magna, Wolters encountered what he has named the flame-nucleus, a condition in which the nuclear membrane has disappeared, and the nuclear substance is prolonged at various points into the protoplasm of the body of the protozoon. As soon as the nucleolus has broken up into subdivisions the parasites are ripe for conjugation. It was found that after the syzygium was formed the nucleus of each parasite moved to the periphery, became elongated, and soon exhibited a typical nuclear spindle with the chromatin now massed together at the middle of the spindle. The chromo- somes were very small. The division of each nucleus took place, and half of each nucleus was extruded as a polar body. Meanwhile the surfaces by which the parasites adhered to each other became altered in such a way that instead of the sharply marked line of division previously seen throughout a complete series of sections of a syzygium, there was at one part of the applied surfaces a communication through which the two parasites fused together. In each parasite the polar body was extruded on the surface opposite this communication, towards which, after the polar bodies had been formed, the two nuclei moved, and having reached the spot at the same time they fused together. After this a nuclear spindle was to be seen in each half of the syzygium, and could be distinguished, by its position close to the area of communication, from the spindles which were concerned with the formation of the polar bodies. Thus it appeared that the conjoined nuclei had undergone division, and that the daughter nuclei were again subdividing. The resulting two nuclei moved towards the periphery in each half of the syzygium, and there formed two spindles. This Wolters found to be the case in complete series of sections. These spindles were smaller than those of the polar bodies. By repeated subdivision these peripheral spindles and resulting nuclei increased in numbers and became surrounded by proto- plasm constituting the sporogonia, which arranged themselves OBSERVATIONS ON VARIOUS SPOROZOA. 287 peripherally, and also in spaces extending from the periphery into the remains of the body substance of the parasite. The single nucleus of the sporogonia subdivided into eight, and meanwhile the sporogonia became surrounded by capsules constituting sporocysts (pseudo-navicelle), the substance of which split up into eight crescentic spores, each of which contained a nucleus. Such are the chief results obtained by Wolters with regard to the Monocystides, and the thorough- ness of the work and the beautiful drawings by Nussbaum, at whose instigation the work was undertaken, go far to establish the conclusions arrived at. I will now detail some of the features I have obtained by examining the seminal vesicles of Lumbricus agricola, taken in the month of May. I have not had the opportunity of making such complete serial sections as Wolters, and in so far my criticism must be incomplete; still, the observations recorded below may be found of some interest. In this place it is advisable to describe the methods employed. Wolters found, and I have had the same experience, that Flemming’s fluid did not give good results with gregarines. Wolters’s results are chiefly based on the examination of sections of material fixed in saturated solution of picric acid. I have employed a method more commonly adopted, and which I have found most satisfactory, not only for gregarines, but for Coccidium oviforme and for animal tissues in general. Small portions of the tissue are placed for twenty-four hours in Foa’s reagent, i.e. a mixture of equal parts of a saturated solution of corrosive sublimate in normal saline solution and a 5 per cent. solution of bichromate of potassium or Miiller’s fluid. Then the material is transferred for twenty-four hours to running water, and afterwards placed on successive days in 30, 60, and 90 per cent. alcohol. After that they are placed in absolute alcohol, and after saturation with chloroform are embedded in paraffin, care being taken that the bath does not reach a temperature higher than 50°C. ‘The sections were cut with a Minot’s microtome, and fixed on the slide with albumen and glycerine. After the usual process they were 288 J. JACKSON CLARKE. stained with Ehrlich’s acid hematoxylin diluted with distilled water, and when they had assumed a brownish pink colour were transferred to a bath of tepid tap water and left for at least two hours. Then for two or three minutes they were stained with a solution! of Gribler’s water-soluble eosin, dehydrated, cleared by xylol, and mounted in the usual way. With regard to Wolters’s description of polar bodies, I can only say that structures which are explicable only as of this nature are of frequent occurrence in M. agilis. They resemble flattened nuclei, and are placed beyond the surface of the parasite and lie between it and the capsule. A nucleus is usually to be seen within the parasite close to such bodies, and frequently remains of a spindle can be made out passing from the nucleus peripherally towards the polar body, as with Wolters I am inclined to regard it. I have not been able to give sufficient time to the investiga- tion to place me in a position to criticise Wolters’s description of the fusion of the nuclei after extrusion of the polar bodies. I have been able to confirm Wolters’s view of the origin of sporogonia in the main, but some modifications are, I think, required of the process as described by Wolters, who would appear to say that every nuclear division after the fusion and re-separation of the original nuclei proceeds by regular mitosis. Against this, such appearances as I have sketched in Pl. 31, fig. 1, may be objected. The drawing represents a syzygium of M. agilis. It was surrounded by a connective-tissue capsule. About the middle of each half of the syzygium is a mass (a) which I think can only be regarded as nuclear. These masses are composed of fine granules, most of which are coloured purple? by hematoxylin. Amongst these are coarser granules, which give with the same reagent a deep blue reaction characteristic of chromatin in both animal and vegetable cells. These chro- 1 This was obtained by dropping a few drops of a strong alcoholic solution into a watch-glass filled with distilled water. ? This is Ehrlich’s metachromatic reaction. ‘‘ Metachromatisch d. h. in Einer dem angewoéhnten Farbtone abweichenden Niiance farben,” Ehrlich, ‘Gesammelte Mittheilungen,’ 1891, p. 2. OBSERVATIONS ON VARIOUS SPOROZOA. 289 matic granules are arranged in lines (0) which radiate out into the substance of the parasite, in the body of which numerous similar granules can be seen, and they are often joined together by achromatic filaments. Besides these granules numerous typical spindles (c) and nuclei are to be seen, placed for the most part at the periphery of each part of the syzygium. I regard the granules as a phase of mitosis, and the purple granules of the nuclear bodies as chromatin in a modification previous to that in which karyokinetic activity begins. Wolters seems at one time to have held a similar opinion to mine in reference to the radiating strings of granules, and to have relin- quished it, I think, on insufficient grounds. ‘‘Um diese Spindeln sah ich bei priiparaten welche durch Flemmingsche ldsung abgetodte waren, viele stark farbende Kornchen in der Substanz vertheilt, ebenso hier und da, auch weit ab von den Spindeln, in den Syzygiten. Ich war geneigt dieselben als chromatische Substanz anzusprechen. Spiatere Untersuchun- gen an Hoden, die ich mit Umgehung dieser Lisung abtédtete und hartete, zeigten nichts davon, sodass ich von meiner ansicht zuriick gekommen bin, ohne eine befriedigende Erk- larung dieser K6rnchen geben zu kénnen.” In a second drawing, Pl. 31, fig. 2, of part of the periphery of another couple of M. agilis I have represented some of these granules (a) on a larger scale. In this case some of the deep blue granules were surrounded by material which was coloured by the eosin of the stain. On comparing these with the sporo- gonia (¢c) lying at the periphery a close similarity of struc- ture was observed, the body of the sporogonium staining in the same tone with eosin as the material investing the chro- matic granules. The nuclei of the sporogonia all stained of the same deep blue as the granules, and they presented a great variety of form. In some mitotic processes seemed to have begun, and the same holds good for the sporogonia of Mon. magna, as is shown at (d) in the lower of the two in fig.3. Iam able to confirm Wolters’s description of the forma- tion of sporocysts and spores. Whilst speaking of sporocysts I may mention that I have encountered in the seminal vesicles 290 . J. JACKSON CLARKE. of Lumbricus agricola prismatic pseudo-navicelle exactly like those described by Bosanquet! in the body-cavity. The observations made by Wolters on Clepsidrina blat- tarum are, unfortunately, scanty. Much remains to be done with regard to the earlier phases, both of the poly- and the mono-cystides. More importance attaches to Wolters’s description of certain phases of Klossia helicina. This parasite is of great interest from the position it holds be- tween the Gregarines and the Coccidia. From the fact that conjugation has not been observed, I, like L. Pfeiffer,” should be inclined to place it with the latter group. First described by Kloss* of Frankfort, it has since been described by A. Schneider and by L. Pfeiffer (loc. cit.). To the latter author I am indebted for many beautiful preparations of the parasite in Helix hortensis and Succinea Pfeifferi. Pfeiffer has arrived at some interesting conclusions based on the study of this parasite. Of these the more important are—lst, the phenomenon of multiple * infection, as many as fifteen parasites being found within a single epithelial cell; 2nd, that when multiple infection occurs only one of the parasites reached maturity; and 3rd, that the size attained by the parasite is determined by the size of the epithelial cells of the kidney of the species of snail infested by the parasite, though in all cases the sporogonia have the same dimensions, the number of sporogonia varying with the size of the parasite from which they are derived. The general features of Klossia I have never seen better than in some common grey slugs which I examined in July, 1892. The slugs were found in a hollow in the rocks below the falls of the river Shin, in Sutherland. With them were 1 W. C. Bosanquet, ‘Quart. Journ. Micr. Sci.,’ 1894, No. 148, p. 421, fig. 19. * L. Pfeiffer, ‘Protozoen als Krankheitserriger,’ 1891, p. 72. 3 Hermann Kloss, ‘ Senkenbergische Abhandlungen,’ vol. i, 1855-6. 4 LL. Pfeiffer compares this with what occurs in the Sarcosporidia, the micro- sporidia in Coccidium salamandre, and in the Coccidia of the kidneys of he goose and the dog. OBSERVATIONS ON VARIOUS SPOROZOA. 291 numerous examples of Helix hortensis. The kidneys of the snails and the slugs were all alike infested by the sporozoa, but only in the slugs did I find swarm-sporing side by side with the ordinary mode of reproduction by sporocysts. One of these parasites, as seen in a teased preparation of a slug’s kidney, is shown in fig. 4. The nucleus (@) was large and oval, and showed a single large nucleolus (4). In fig. 5 is repre- sented a cell (a) as seen in a section, and containing a parasite subdivided into four sporocysts (c) (others were present out of focus), each of which contains six crescentic spores (d). In some of these a single nucleus could be made out. The spores were very large, averaging 12, in length. All the sections of the kidneys of several slugs showed a marked infection, and in most of the sections swarm-sporing was well seen. Fig. 6 shows an example of this. A much hypertrophied cell (a) contains a large sporing parasite, and six smaller parasites (e). The sickles (4) are very large, 20uin length. Some of them are undergoing farther subdivision.! The capsule of the para- site has ruptured, and at the point of rupture are some free sickles, which also are undergoing farther subdivision (c). The appearance of swarm-sporing in a fresh teasing is shown in fig. 7, where some detached sickles (4) are present. The colour reactions in these sections were not good, probably owing to the slugs having in the first instance been placed in Scotch whiskey, so that they will not serve as a basis for comparison with Wolters’s descriptious ; but since I have been able to find in Coccidium oviforme all this author encountered in Klossia, and also many additional features, I will pass to the consideration of this more familiar parasite. For examination I chose a highly infected liver in which the lesions were still in process of evolution. Fig. 8 shows an average appearance of a portion of the epithelial lining of a cyst as big as a small pea. All the larger parasites show signs of nuclear activity. The chromatin, in nearly every instance, gave a typical deep blue reaction to acid hematoxylin. The most abundant form 1 The large sickles formed in swarm-sporing would thus appear to have the equivalents of sporogonia, not of spores. 292 J. JACKSON CLARKE. in which Coccidium oviforme usually presents itself in sections is the spherical granular body represented in fig. 9 to the left. The granules (4) stain slightly with eosin, but remain transparent. They have, no doubt, the same signification as the Gregarina corpuscles! in the higher Sporozoa, i.e. they serve as stored food. The ‘‘ dauerform” (L. Pfeiffer) of the parasite is represented in the same drawing to the right. In it the granules have disappeared, and the nuclear body (a), which is round in the spherical-granular phase (a’), is oval in the encapsuled parasite. In neither case, however, does the nucleus give the reaction of chromatin, but is stained by the eosin. One other phase of the parasite may be mentioned in passing. This is a small, dense, spherical body, devoid alike of granules and of nuclear body, and staining throughout without eosin. Sometimes a parasite possessing a thick oval capsule is found to have broken up into sickle-shaped swarm- spores, but the rule is that when the parasite multiplies whilst still within the body of the host, it has either no capsule at all or only the delicate so-called primordial capsule. It is to the changes which lead to subdivision of the parasite within the host that I wish to direct attention. R. Pfeiffer? first de- scribed swarm-sporing, but chiefly in fresh specimens, and without any detailed account of nuclear processes. LL. Pfeiffer (loc. cit., p. 45) described distinct hematoxylin-stained nuclei in the process of swarm-sporing, and says, ‘‘ Kingehendere Untersuchungen sind hier noch sehr nothig, dar der mehr oder weniger akute Krankheitsverlauf von Massgebenden einfluss ist auf die Vermehrungsweise des zugehdrigen Para- siten.” Among the parasites in actively extending lesions of the rabbit’s liver some present a distinct ‘‘ geflammte Kern,” like that shown in fig. 10 (ce). Such nuclei take only the eosin of the stain, or at most a slight tinge of purple at their edge. Most 1 The chemical nature of these bodies is not yet determined. They dissolve n alkalies and mineral acids, and are not fat, nor do they contain lime. See Max Wolters, loc. cit. 2 R, Pfeiffer, ‘ Beitrage zur Protozoenforschung,’ Berlin, 1891. OBSERVATIONS ON VARIOUS SPOROZOA. 293 of the nuclei which are preparing for subdivision show distinct radiating processes, which stain deep blue with acid hematoxy- lin. Sucha nucleus is shown in fig. 11 (a). The food-granules in such parasites have become diminished in numbers, and in some cases stain more deeply with eosin. The next stage is shown in fig. 12, where the nucleus has subdivided into two parts with the formation of a typical spindle. Wolters was unable to observe spindle-formation in Klossia, but from the identity of some of the phases of nuclear structure in C. oviforme, and in Klossia as described by him, it is probable that in both as also in Gregarines, spindle-formation takes place. In this way two or three (fig. 1, a and a’) nuclei are formed. Fig. 13 shows the subdivision of a separated portion of the nucleus. . bay, rey ot) BD ~ PDAS Wh Ve) | WET ee a8 ips - iag his ig ye note LA ho .: mi ne) ilfps | a fs praia (} A iy ay) picks age iii braiit a Pi) is ark ooh t OF B.S F it ¥ . | = me —".t oP. Caintidlaeaionn Ty joy, eo pone aaah 4 art) fark) 3 Jee Ses) Se ie ad AO eRe PS eC ee war tate. pi eee yi Leica Wesel it vend Pipi marin : ietieioh wh biie awe os@pdls celal fe Diesels Lb ty Vidensi! ult i La iu) 4466 ol ee Cee Ten ae Ope thep a silks wild * pi 7 - - “st era re yh) Oe Sgr 5) ie Ae eA} Ss ISS ol BTEE ‘ oe wane Ty yiasl > limape LOPS vy fork SR A Biche — sini / eiaidulanun oi ulmanise Tne sae ae Hhans( Gurr Gpeowtt ¥ Tig hat, B) els OE ral nel. Te) Lee rd a a Serpe OE eat dts. Maa wii 4 Ment Chobe sim jal peg cca | age Lah L <,i4ea als ‘ 19) Ai we Ai)" Y dd nih ype Leagan jem tt adios ong a lb ‘i Stay Sole ee Jules ts Sa SCRE Hote tie fone amen oN — ; a sieG > p Svitireh Grins ih Pulte eA Need Sou : : eu’ F aia CUP Nev WED Pow sl Ud te nie ot Re ideal veoh . WVHA lipin Chie pid alike Foliar) nis see OA ter Be | Nive auido or outages hae inhi a bt SUR iN ta eetay wt. sib 4 Riss rT lished’, aolaouh” » 2) Vth opt Sh 7} CR a a * qiWonk, “Nk iis | ite Selatan “a ees — Hy ; eo! er —. bg 7 pig _ an +e 1 , he pes > yon a A STUDY OF METAMERISM. 395 A Study of Metamerism. By T. H. Morgan, Ph.D., Associate Professor of Biology, Bryn Mawr College, U.S.A. With Plates 40—43. TABLE OF CoNTENTS. PAGE I. Typical Forms of Modification in Annelids : - : . 395 II. Variation in the Position of the Reproductive Organs. . 403 III. Abnormalities in Front of or including the 15th Metamere . 407 IV. Study of Embryos . : ‘ : : ; 5 . 418 V. Abnormalities at the Posterior End : : : : : . 421 VI. Modification of Internal Structures . : : : : ~ 425 VII. Study of Polychetous Annelids and Leeches . : : . 428 VIII. Summary. : : - : : . 431 IX. Modifications in Ancens of Arairopods : : : : . 435 X. Abnormal Metamerism of Locust . ; : : : . 437 XI. Study of Colour-bands of Echinoderms . : : : . 439 XII. Regeneration in Earthworms . ; : : : ‘ . 445 XIII. General Conclusions . ‘ ; F : ; : : . 460 I. Typrcat Forms or MoptrFicaTion. Tuat there were occasionally to be found in the Annelids irregularities in the serial repetition of the rings seems to have been known to several of the earlier writers on descriptive and systematic zoology. The fact did not attract more than a passing attention, and such irregularities were relegated to that waste-heap of abnormalities from which subsequent in- vestigation has often drawn valuable material. Simultaneously in 1892 two articles appeared dealing with 396 T. H. MORGAN, these abnormal conditions, one by C. J. Cori (18), and the other by the present writer (88). Both writers pointed out the general interest attached to these modifications, and their importance for an interpretation of the general problem of metamerism. Subsequently a third paper appeared (10), re- cording the presence of similar abnormalities in many groups of Annelids, but without making any attempt to solve the problem itself. My own paper was only a preliminary notice, and until the present time I have not had an opportunity of fully describing the material that I had at that time already accumulated, studied, and drawn. The present paper attempts to give a full consideration of the facts only touched on before, and to extend over a wider field the conclusions reached. The modification of the rings, segments, or metameres of the Annelid fall into two general classes,—not, however, sharply separated from one another. The first of these I shall speak of as the compound metamere, in contradistinction to the normal or simple metamere. In the earlier: paper the shorter term “ split metamere” was used, although it was there pointed out that the term was misleading. The other type of modification may be spoken of as the spiral metamere or spiral modification, or briefly as the spiral. The simplest case of the compound metamere is represented by the diagram on Pl. 40, fig. I, a,3,c. Fig. I, a, is a dorsal view of a compound segment. The segment is double on the right side, normal on the left. Fig. I, B, is the same, turned over so as to be seen from below, showing a similar doubling below on the same (right) side of the body. Fig. I, c, repre- sents the compound segment as opened along the “split.” It is conceived to be transparent, and the dotted line to represent the lower surface outlines. The general appearance is that a metamere has been split into two on one side of the body, but has retained on the other side its normal structure. The question at once arises, are we dealing here with a case of division of a metamere on one side of the body, or is it a case A STUDY OF METAMERISM. 397 of three half-metameres united together? It is one of the aims of the paper to answer this question. To the other class of modifications belong the spirals, one of the simplest of whichis shown in fig. VII, a, B, c. The spiral is produced by a combination of two half-compound metameres, i.e. one metamere is “split”? below, but the same one is normal above. The next metamere is split above and on the same side of the body, but not below. The posterior arm of the anterior half-compound metamere is continuous at the side with the anterior arm of the posterior (upper) half-com- pound metamere. The same fact may be stated in a different and perhaps a clearer way. A half-metamere on the right side of the body unites in the mid-line below with a metamere in front of it, and above with a metamere behind it, fig. VII, c. We have formed as a result of this arrangement one more half-metamere on the right side of the body than on the left side, as is best shown by the construction in fig. VII, c. We may proceed to study more in detail the many kinds of modification that group themselves around these two types. We shall find that nearly all of the geometrical combinations conceivable are to be found in the worms themselves. Category 1, fig. I, a, B, c.—This form has already been described as the type. It is equally common on the right and left sides of the body. Category 11, fig. II, a, B, c (above, below, and reconstructed, as seen from above).—In this form one (above) of the lines of the split turns to one side (posteriorly) to fuse with the septal line. This leaves, as shown in the reconstruction, a free end forming a modification of the compound metamere. This modification may appear either on the right or left side of the body, and the free end may occur above or below, anteriorly or posteriorly. Category 111, fig. III, a, B, c.—By a simple shifting of the lines in the last form we pass to the third category. Here both of the lines of the extra half-segment turn towards the same septal division, and we have formed an intercalated half- metamere, 398 T, H. MORGAN. The variations of this category are formed by the lines turning both anteriorly or both posteriorly, and on the right or left side of the body. In reality the half-segment is simply intercalated between two perfect segments ; and when we speak of the lines turning either forward or backward we only imply that the intercalated piece has encroached more on the one or the other segment. Category rv, fig. IV, a, B, c.—Here three half-segments of one side unite with one half-segment of the opposite side, so that the right side appears to be “ split”? by two incomplete divisions into three half-segments. This may be called a double compound metamere. Variations arise owing to the depth to which the incomplete lines (splits) pass towards the middle line. A combination of this form (really only a ha/f- double compound) with others is shown in Pl. 40, fig. XIV. This same modification may be carried a step further, and in one case I found four half-metameres of one side united to one of the opposite. See Pl. 42, fig. 47. Category v, fig. V, a, B, c.—Here on the upper side appear two eompound segments—one on the right side, the other on the left. On the lower surface, B, two perfect metameres are found. By tracing the division lines between these seg- ments from the lower to the upper side we see the result is brought about by the failure of these two septal lines to meet above ; each turns somewhat to one side. The reconstruction shows that a middle connective is left to unite the two seg- ments above. This is brought about by the fusion above of the middle of the right posterior segment with the left anterior. The variations are few. The incomplete meeting may take place either above or below, and the lines turn anteriorly or posteriorly. Category vi, fig. VI, a, B, c.—The arrangement found here inay be produced from fig. v by extending the free ends of the segment lines till the one meets the segment in front, and the other the one behind. A reconstruction of such a figure gives the spiral shown in fig. VI,c. This spiral form terminates A STUDY OF METAMERISM. 399 both anteriorly and posteriorly in free ends. The two meta- meres involved take two complete turns around the body. Variations of this (like those of category v) readily suggest themselves. Category vit, fig. VII, a, 8, c.—This spiral form has been already described as a type of the spiral. Category vitt, fig. VIII, a, 8, c.—On the upper side are two half-compound segments opening in opposite directions (a). On the lower side three normal segments are present (8). The reconstruction (c) shows that the posterior arm of the first half-compound segment takes a complete turn around the body, and then unites with the second half-compound segment as its anterior arm. The spiral involves three segments. The middle of the three, instead of forming its proper union above, unites above on the right side with the segment anterior to it, and on the left side and above with the segment posterior to it. The variations of this form consist in having the union above or below, on the right and on the left, or on the left and on the right side. Category 1x, fig. IX, a, 8B, c.—This form is merely a longer spiral than that of category vir. The spiral turns one and a half times around the body, beginning and ending as in vi. Three segments on one side of the body correspond to four on the opposite. This process may theoretically be continued through an indefinite number of segments, but generally a few turns suffice to bring the spiral to a close. The most extreme case that I have found involved twelve and a half turns to the spiral (Pl. 41, fig. 26). Category x, fig. X, a, B, c.—This spiral is an extension of that of category vi11. It involves a greater number of meta- meres, and in consequence the spiral is longer. Two turns are taken around the body. In both categories 1x and x all the variations found possible for vi1t and 1x are here applicable. 400 T. H. MORGAN. Fig. XI, a, B, shows a further extension of vit, with three turns. There are several combinations of the preceding categories that are possible, and some of these I have found. It is not particularly important to discuss these possibilities, as they all reduce back to forms already described. As an example, however, one such is given in fig. XII, a, B, c; the reconstruc- tion, c, sufficiently explains the conditions. It will be noted that five half-segments of one side correspond to three of the opposite. Whenever a double half-compound metamere is introduced a more complicated form of spiral results (see fig. XIV, and Pl. 42, fig. 47, at x1). This causes a double spiral, i.e. two spirals, to take a parallel course, as shown in fig. XIV. Oneor both of the spirals may end in a half-compound metamere— both so end in fig. XIII. This reconstruction will, I think, sufficiently explain the conditions, so that a further description would be superfluous. The double spiral may be formed in other ways. Two suc- cessive compound metameres may be introduced in such a way that a new spiral is started along with the one already present. Such cases are shown in fig. XIII, and in fig. 47, Pl. 42, at x’, Several variations of these combinations suggest them- selves, and several have been found amongst the worms, but the discussion of these variations would not add materially to the former cases, and may be omitted. Even a triple spiral was found in one case, as shown in fig. 47 at x*. It was of short duration, owing to the introduction of compound meta- meres in such a way that the spirals were quickly absorbed. Lastly, the series of compound metameres, double and triple spirals shown in fig. 47, Pl. 42, were all drawn from the same worm. We find 134 half-metameres on the right side and 118 half-metameres on the left. In all there were fifty-two perfect rings, and the numbers in the plate between consecutive groups of compound metameres, spirals, &c., in- dicate the number of perfect rings in that locality. An exa- mination of this reconstruction will show how far it is possible A STUDY OF METAMERISM. 401 for the variations—abnormalities—to be present, and the worm still be able to reach sexual maturity. In another worm con- taining many abnormalities there were 131 half-metameres on the right side of the body and 189 on the left side. In all there were seventy-nine normal rings. The number of half- rings in these worms is noticeably very large as compared with the normal number. It should also be noticed that the abnormalities are not confined to any one part of the body, - but scattered throughout the whole length of the worm. We may now study the relative proportion in which the different modifications, described above, are found. Ina lot of 318 worms 218 were found to be normal externally with respect to the metameres, and 100 abnormal. This is in the proportion of 1 to 2°18, or in general one worm of every three was abnormal. In the same worm there often occurred more than a single abnormality. Thus in the above 100— With. . 1 abnormality : ‘ : 65 - : . 2abnormalities . : ? 16 er 3 eo as : 2 : 10 » more than 3 os ‘ = : 9 100 An examination of the material shows that there is practi- cally no difference in the number of compound metameres on the right as compared with the left side of the body. The examination shows, however, that a much larger number of “‘ splits” are present on the dorsal surface of the body than on the ventral surface. In a compound metamere we have a split above for one below, but in such spirals as shown in figs. vri1, x, &c., we have two above and none below, and such spirals have their “splits” on the dorsal surface in the greater number of cases. I regard this as important, as it gives a clue to the solution of the problem. An examination of the material shows that there are thirty- two cases of spirals beginning and ending above (category vi1r), and only one case of a spiral beginning and ending below. 402 T, H. MORGAN. There were ten cases of failure of lines to meet above as in categories v and vi. No cases below. There were twenty-five cases of spirals beginning above or below, and ending on the same side of body—below or above. Amongst these twenty-five cases there were fifteen that began above and ended below, and ten that began below and ended above. If we compare the data of the thirty-two cases with the twenty-five cases we find that we have ninety-nine cases where the split was on the dorsal surface, and twenty-five cases where the split appeared on the ventral surface.’ That is to say, there were nearly four times as many cases of false unions above as below. If we arrange the different forms of abnormalities found under the categories given in the preceding section, we find the relative proportion of the abnormalities in the 100 abnor- mal worms to be as follows. In this count no difference is made as to whether an abnormality occurred once or twice on the same worm. It is based on the results of all abnormalities found present. Category I, II . : ‘ : ; 42 - inf = 5 : : : 1 3” Wo VI. ° . . . 10 Vito: : : : . 20 } ” 96 { se Tie Fis ; : 5 : 6 % VILE. 5 ; : : 16 | uf dee # < 5 5 ‘ s} oe =f xI, &e. : ‘ 4 ; 8 Mixed forms too irregular for classification . a 1 These numbers were derived from the data given above as follows: 32 (spirals beginning and ending above) x2 = 64 US Gees “3 above and ending below). —— fi |) TO > ss below and ending above) . wg 10 (cases where lines failed to meet above) . = 10 99 15 (spirals beginning above and ending below) . = 15 OCs pS below and ending above) . = 10 25 A STUDY OF METAMERISM. 403 II. Vartations 1n THE Position oF THE REPRODUCTIVE Organs. With the hope of throwing additional light on the modifica- tions in the arrangement of the metameres, I was led to examine carefully those cases where the false arrangement occurred in the segments anterior to the fifteenth. Here there are definite landmarks, viz. on the outside are the openings of the vasa deferentia, oviducts, and seminal re- ceptacle, and on the inside are the organs belonging respectively to these openings. The openings of the vasa deferentia on the fifteenth segment are the only ones of these openings that are readily and with certainty to be made out, so that I have confined my study largely to the modifications in the position of these openings. When we come to study the internal arrangement of the organs, then the position of ovaries and seminal receptacles will also be considered. In utilising the fifteenth segment as a point of departure, it was first necessary to determine whether this was in reality a fixed point. In a lot of 799 worms, twenty were found with the vasa deferentia opening abnormally. In twelve cases of the twenty the openings were on the same segment, but this was not their normal segment (fifteenth). In six cases the openings (right and left) were not on the same segment. In two cases the openings were doubled on one side or the other. To summarise: A. Openings on the same segment, but not on the 15th . 12 B. Right and left openings, not on the same segment <6 c. Openings doubled on one side . : ; ae A. The following table records the openings of the vasa deferentia in the first category. No. of Worms. E vasa deferentia opened on 10th metamere. 1 EP) 3) llth » 2 ” » 12th » 4 2 2”? 14th 2 4 » » 16th ,, 12 404. T. H. MORGAN. In those cases in which the openings of the vasa deferentia occur on a segment anterior to the 15th metamere, we may be dealing with a case of incomplete regeneration of the ante- rior metameres. In fact, in the first case given there was evidence for believing that the anterior end had just regene- rated, and even the second case gave slight evidence of the same thing. That all of the cases can be explained in this way is, I think, highly improbable. We find, as the table shows, the greatest variation on each side of the 15th meta- mere, pointing in favour of individual egg variation rather than regeneration. ‘This view is also strongly supported by an examination of the internal conditions. Again, those cases recorded above in c and p show conclusively that the openings of the vasa deferentia may shift, and in these cases regenera- tion is largely out of the question. In order, however, to satisfy myself completely, i.e. experi- mentally, on this point, I tried by (artificially) cutting off the anterior metameres of a large number of worms to determine what proportion of these regenerated the full number of anterior metameres. The results are given in a later section. B. There were six cases out of the twenty in which the openings of the vasa deferentia were in different metameres of the same worm. These are figured in Pl. 42, figs. 48—58. In the first case, fig. 48, the 15th metamere has on its right side (left of figure seen from below) an opening of the vas deferens. The left side of the 15th has no opening, but on the 16th metamere of that side occurs the opening of the left vas deferens. One such case is recorded. In fig. 49 another variation is found. Here the 15th meta- mere has on the left side of the body its proper opening, but there is no opening on the right side of that segment. On the 16th metamere, however, and on the right side, the right vas deferens opens. Four such cases are recorded. In fig. 50 we find that the right side of the 15th metamere has an opening, but not the left side. On the 14th metamere we find the opening of the left vas deferens. One such case is recorded. A STUDY OF METAMERISM. 405 There can be no doubt but that an examination of a larger number of worms would show other combinations, but these are sufficient to show that variations in the position of the vasa deferentia do occur. c. Two worms in this lot showed a doubling of the openings of the vasa deferentia on one side. In fig. 51 we find on the 15th metamere a pair of openings, right and left; but in addition to these we find on the left side of the 16th segment another opening of a vas deferens. In fig. 52 we find the 15th metamere with its pair of openings, the right somewhat enlarged. On the left side of the 14th metamere another opening is present, as shown in the figure. Dissections were made of nearly all of the worms recorded in categories a, B,c. There is always the possibility of mis- takes when only surface views are examined, but fortunately in all cases recorded the evidence from internal structure has supported the conclusion from the external appearance. More- over additional data of considerable interest has, I think, been gained. We may examine first the results of dissections of worms in which the paired openings of the vasa deferentia are on some other segment than the 15th. No. 1. Openings of the vasa deferentia on the 10th meta- mere, fig. 53. The ovaries were present two metameres in front of the vasa defereutia openings, i.e. the ovaries were in the 8th metamere. All of the other parts of the reproductive system were present and normally developed. The brain appeared in the Ist metamere. No. 2. Openings of the vasa deferentia on the 11th meta- mere. Only one ovary was found in the 9th metamere (left side) ; the other probably lost in the process of dissection. No. 3. Openings of the vasa deferentia on the 12th meta- mere. ‘The ovaries were found in the 10th metamere, fig. 54. The first segment seemed incompletely regenerated. No. 4. Openings of the vasa deferentia on the 14th metamere, fig. 55. The ovaries in the 12th metamere. 406 T. H. MORGAN. Seminal receptacles in the 8th and 9th metameres. In another worm similarly modified the ovaries were also present in the 12th metamere. No. 5. Openings of the vasa deferentia on the 16th meta- mere. The ovaries were present in the 14th metamere. We may next examine those cases where the openings of the vasa deferentia are, right and left, on consecutive meta- meres. No. 6. Three worms were dissected in which the right vas deferens opened on the 15th metamere, and the left on the 16th, figs. 56 and 57 (projections from above). In two of the three worms the ovaries showed a similar variation, appearing on the 13th (right) and 14th (left) metameres,—that is, each two segments in front of its corresponding vas deferens. In the third worm both ovaries appeared in the same metamere, viz. the 13th. No. 7. Two cases recorded with the left vas deferens opening on the 15th metamere, and the right on the 16th. The ovaries varied correspondingly, and were found in the 13th (left) and the 14th (right) metameres (see fig. 58). The seminal receptacles showed a similar displacement, and were found in the 9th and 10th (left) and 10th and 11th (right) metameres. No. 8. The worm drawn in fig. 52 in which the left vas deferens is doubled (14th and 15th metameres) was dissected, and a corresponding doubling of the ovaries of the same side was found. No. 9. In another case of doubling the vasa deferentia of the right side opened on the 16th and 17th metameres; that of the left side was single and appeared on the 16th metamere (fig. 59). The ovaries varied also, appearing in the 13th and 14th metameres of the right side, and in the 14th of the left side. The variation here is uot the same in the ovaries and vasa deferentia on the right side, for if the ovaries had varied to correspond they would have appeared in the 14th and 15th metameres of the right side. No. 10. The paired openings of the vasa deferentia are found on the 15th metamere. In addition there is another A STUDY OF METAMERISM. 4.07 opening on the left side of the 16th. The ovaries are found in the 13th metamere (fig. 60, a). No. 11. The paired openings of the vasa deferentia are found on the 15th metamere. On the left side of the 14th there is an additional opening. One pair of ovaries are found in the 13th (fig. 60, B). No. 12. One case of doubling is recorded for L. terres- tris (the preceding cases were all for L. foetidus) (fig. 74, a, 74, 8B). The opening of the right vas deferens is doubled, and opens both on the 15th and 16th metameres. The left only appears on the 15th. Dissection showed both ovaries present in the same segment and not doubled. The dissection shows also the course taken by the vasa deferentia (fig. 74,8); and we find a single vas deferens on the right side that supplies both openings of that side. I regret that on account of the difficulty in dissecting out the vasa deferentia in L. foetida I did not get further evi- dence as to the method of doubling of the external openings in those worms. III. ABNORMALITIES IN FRONT OF OR INCLUDING THE 15TH SoMITE. In a lot of 799 worms (L. feetidus), twenty, as recorded in the preceding section, showed abnormalities in the position of the openings of the vasa deferentia. In this same lot, after the removal of these twenty (leaving 779), there were thirty- two worms with abnormalities in front of or including the 15th metamere. This is in the proportion of 1 : 24. In 1893 I again collected material in order to obtain a larger number of abnormalities in the anterior ends of the worms. From 957 worms eighteen were found showing modifi- cations of the anterior metameres. This is in the proportion of 1 : 53. The material from these two sources, with a few additions from other lots of worms, has been brought together and figured in Pls. 41 and 42. VOL. 37, PART 4,—NEW SER. EE 408 T. H. MORGAN. The modifications may be classified under the following categories : . Part of a metamere compressed. . Failure of external septal lines to meet. . Spirals in front of the 15th metamere. . Spiral involving the 15th metamere. . Results of the introduction of compound metameres. A. Plate 41, figs. 1—3. In fig. 1, 4 B, the 4th metamere is not fully developed on the right side of the body. The ring is narrower, and has less pigment on the right side. The vasa deferentia open on the 16th metamere, as shown in B. Fig. 2, a B, shows another worm with the 2nd metamere similarly modified. The vasa deferentia open on the 15th metamere. Fig. 3, a B, shows the 5th metamere reduced on the left side. The vasa deferentia open on the 16th metamere. Conclusion.—It is not a little surprising to find in two of the three cases where an anterior metamere is partially unde- veloped, that the openings of the vasa deferentia have shifted one segment posteriorly. The data are too small to lead to any definite conclusion. B. Plate 41, figs. 4—11. In fig. 4 we see that the line of division between the 9th and 10th metameres failed to meet exactly on the upper surface. The vasa deferentia opened normally, if we count the 9th and 10th as true metameres. In fig. 5 the line between the 5th and 6th metameres failed to meet above. Vasa deferentia normal. In fig. 6 the line between the 4th and 5th metameres failed to meet. Vasa deferentia normal. In fig. 7 the line between the 3rd and 4th metameres failed to meet. Vasa deferentia normal. In fig. 8 the line between the 13th and 14th metameres failed to meet. In fig. 9 the line between the 12th and 13th, and in fig. 10 HoOOwW Pe A STUDY OF METAMERISM. 4.09 the line between the 11th and 12th. In these last three cases all failures were above, and in all the vasa deferentia were normal. In fig. 11 the line between the Ist and 2nd failed to meet below; in fig. 12 between 18th and 14th below. In fig. 13 the line between the 5th and 6th metameres failed to meet at the sides. Both in figs. 11 and 13 the vasa deferentia were normal, but in fig. 12 there were indications of an additional opening on the left side of the 16th metamere. Conclusion.—In nine out of the ten cases the position of the openings of the vasa deferentia were not affected by the failure to meet of metameres anterior to the openings, nor was a shifting to be expected on a priori grounds, since the modi- fication is clearly only a lack of perfect union between the right and left half-metameres (fig. 13 perhaps excepted). Fig. 12 is, then, a case of doubling of the openings of one side, like those described in the last section, and can have nothing to do with the lack of perfect union of the more anterior segments. It is noticeable that in seven out of these ten cases the imperfect union is on the dorsal surface. In only two cases is it found below, and in one at the side. We will return toa consideration of the point at another time. C. Figs. 14—19, 20—26. Two cases are to be considered under this head: 1st, those in which the spiral introduces no new half-metameres on either side; 2ndly, those cases where a new half-metamere is introduced in front of the 15th metamere, so that one half of 15 becomes united to 16 (15—16), and the other half that remains is united to half of 14 (15—14). lst case.—In fig. 14 a spiral is found involving the 7th, 8th, and 9th metameres. It belongs to category viit (Pl. 40, fig. vi11). No new half-metameres are introduced, and the vasa deferentia open on their normal (15th) metamere. In fig. 15 a similar spiral involves the 12th, 13th, and 14th metameres. Vasa deferentia as before, normal. 410 T, H. MORGAN. In fig. 16 a similar spiral involves the llth, 12th, and 13th metameres. Vasa deferentia normal. In fig. 17 a spiral is present that involves the 12th, 13th, and 14th metameres. The lower end of the figure (posterior end of the spiral) turns to one side and ends in a point. The openings of the vasa deferentia were not apparent on the surface, but dissection showed the ovaries were in the 138th metamere, that is within the spiral. Presumably, then, the vasa deferentia were also in their proper metamere. In fig. 18 a longer spiral involves the 6th, 7th, 8th, and 9th metameres. The spiral belongs to category x (Pl. 40, fig. x). This form does not introduce any extra half-meta- meres, and we find the vasa deferentia normal in position. In fig. 19 a similar (reversed) spiral is found. It includes the llth, 12th, 13th, and 14th metameres. The vasa defe- rentia open normally. Conclusion.—In these six cases where a spiral is intro- duced in front of the 15th metamere the position of the openings of the vasa deferentia is not affected. No better demonstration is needed to show that the spiral is due simply to a rearrangement of the segments involving their union across the middle line. It will also be noticed that in all of these cases the false union is made on the dorsal side of the worm. 2nd case.—In fig. 20, a, B, c, are drawn the upper and lower surfaces of a worm, and (in c)'a reconstruction of the spiral. The 2nd, 3rd, and 4th metameres are involved, and an extra half-metamere (3rd) is introduced upon the left side of the body. The modification belongs to category vit (Pl. 40, fig. vir). We find corresponding to this intercalation that the openings of the vasa deferentia are affected, aud appear on different but consecutive segments. Each, however, opens on its own 15th half-metamere. In other words, the 15th half-metameres still develop their proper reproductive openings, and in consequence of an interpolated half-metamere in the spiral anteriorly the two openings do not fall into the same metamere, A STUDY OF METAMERISM. 411 In fig. 21, a, B, c, a somewhat similar spiral involves the 9th, 10th, and 11th metameres, and also introduces an extra half-metamere on the left side. The external openings of the vasa deferentia were not found, but dissection showed that the ovaries were present in consecutive segments, i.e. each in the 13th half-metamere of its side. The explanation is the same as in the last case. In fig. 22, a, B, c, we find a spiral belonging to category 1x (PI. 40, fig. rx). An extra half-segment is introduced. Nevertheless the openings of the vasa deferentia seemed both present on the same ring (metamere). This metamere is made up of the 15th half-metamere of one side and the 16th half- metamere of the other. In fig. 23, a, B, c, D, are shown two modifications in the anterior end. The first of these is brought about by an incomplete lateral union (c) of the line between the 5th and Gth metameres. More posteriorly a spiral is intro- duced that involves the 13th and 14th metameres. An extra half-metamere, as shown in D, is introduced on the left side. The external openings of the vasa deferentia were not seen. Dissection showed that only one ovary was developed, and that on the right side (or at least only one was found). This was in its normal half-metamere. The seminal receptacles were in their normal segments (9th and 10th), which showed that the more anterior modification (failure of lines to meet) did not affect subsequent metameres. In fig. 24, a, B, c, a short spiral involving the 5th and 6th metameres is present, and introduces an extra half-metamere on the left side. External openings of vasa deferentia were not found. Dissection showed that the ovaries were present in the same segment (13th of one side, 14th of other side). I could not fully decide whether this case was due to a true in- tercalation of a half-segment or to an overgrowth of the middle line by the right half 6th metamere, due to an incomplete dorsal growth of the left half 5th metamere. The latter inter- pretation would be more in accordance with the presence of the ovaries in the 18th—14th metamere, which would then be only the 13th metamere. Moreover, as this is the only case 412 T, H. MORGAN. recorded of such a modification, I think the latter interpreta- tion is probably correct. In fig. 25, a,B,c, arather complicated spiral involves the 10th, 11th, 12th, and 13th metameres,as showninc. An additional half-segment is introduced on the left side. Nevertheless both of the openings of the vasa deferentia appear on the same seg- ment (15th—14th). Conclusions.—The number of worms recorded is too small to give any conclusive data. In two cases (figs. 20, 21), where an extra half-segment is introduced, each opening of the vas deferens is on its normal half-segment, and therefore on con- secutive rings. In two other cases (figs. 22, 25) the openings of the vasa deferentia occur on the same ring (14th—15th in one case, 15th—16th in the other). In a third case only one ovary was found (fig. 23), and in a fourth the nature of the spiral was doubtful. D. Figs. 26—82. Two kinds of modification belong to this category: Ist, those where the spiral passes through the 15th metamere ; and 2nd, those cases where the 15th metamere begins or ends a spiral. lst Case.—In fig. 26 is shown a dorsal view of a worm with a spiral involving the 11th to the 24th metameres. The spiral winds around the body thirteen times. The number of half- segments is not increased by the spiral. Looking at the lower surface of the worm the vasa deferentia are found on the 15th metamere, and are in no way affected by the fact that the metamere is part of a long spiral, and not a closed ring. In fig. 27, A, B, C, a spiral involves the 10th to the 16th meta- meres. Looked at from below, B, the vasa deferentia are seen on the 15th metamere. The reconstruction in c shows that the 15th metamere is a part of the spiral, and still carries the openings of the vasa deferentia. In fig. 28 a spiral is found involving the 13th to the 24th A STUDY OF METAMERISM. 413 metameres. The openings of the vasa deferentia were not seen on the outside, but dissection showed the ovaries to be present in the 13th metamere. Conclusion.—In these three cases where the spiral in- volves the 15th metamere the reproductive organs appear in their normal halfmetameres. Im all of these cases the false union of metameres that produces the spiral is above. The lower surfaces of the half-metameres join one another as under normal condition,—that is, the 14th below and on the right side joins the 14th, &c. On the dorsal side the reverse is true,—that is, the 14th joins the 15th half-metamere, &c. As the reproductive organs are in the lower part of the segment, we would expect to find them still in their normal segment (and not in any way modified), and this we do find in all cases recorded. 2nd Case.—In fig. 29, a,B,c, a spiral is drawn involving the 14th, 15th, and 16th metameres. The spiral begins in the 14th half compound metamere (14—14). The openings of the vasa deferentia are present in their normal half-metameres. In fig. 30, a,B,c, a spiral involves the 15th, 16th, and 17th metameres. Here, as in the last case, the openings of the vasa deferentia are not affected. In fig. 31, a,B, a similar spiral involves the 15th, 16th, 17th, and 18th metameres. Here a half compound metamere (15—+5), beginning the spiral, carries the openings of the vasa deferentia on their normal 15th metamere. In fig. 32, A,B, a spiral involves the 15th, 16th, and 17th metameres. The openings of the vasa deferentia occur on their normal metamere, which forms the anterior limb of the half compound metamere (15—+5) that begins the spiral. Conclusion.—In none of these last cases does the presence of a half compound metamere involving the 15th metamere affect the openings of the vasa deferentia. General Conclusion from Category D.—Nearly every case points unmistakably to the conclusion that, however falsely the segments may unite in front of the 15th metamere, or so that the 15th metamere is involved in a spiral or half 414 T, H. MORGAN. compound metamere, still the openings of the vasa deferentia appear on their 15th half-metamere. It is also very noticeable that in by far the larger majority of cases the false unions of the metameres are on the dorsal surface. EK. Figs. 33—46. The results recorded in the preceding section are, as has been said, fully in accord with the statement that, however varied the union of the half-metameres across the median line may be, still the openings of the vasa deferentia occur in most cases on their normal (15th) half-metamere. In the present category (£), where we should hope to find this same relation to hold, the matter stands otherwise. In only one case, viz. in that first given, does the anticipated result follow. Many of the cases are unintelligible, or nearly so. I used in my earlier paper the first figure referred to above to illustrate the explanation offered as to the value of the compound metamere (there called the split metamere). I still believe that it does this, but I was not then aware how much in the minority, as far as numbers go, this modification really is. I fear that by picking out this one case as an illustration in my short preliminary communication I exaggerated the value of the modifications as furnishing evidence of my view. I was careful to state only that this was “one of the most in- structive cases,. .. . and gives us a clue by which to interpret the split metamere.” This statement I still think is true, but by far the weightier evidence for my conclusion is now fur- nished by the cases recorded in the preceding sections rather than in this one. Pl. 41, fig. 33, 4,8, shows the anterior end of a worm having the 10th metamere compound (10—+1%) ; there is, therefore, one more metamere on one side of the body (the left) than on the other. Correspondingly the openings of the vasa de- ferentia are not on the same metamere, but on consecutive | ones. Each opening is on the 15th half-metamere of its side. Each half-metamere has developed its normal structures, but A STUDY OF METAMERISM. 415 owing to the shifting of the metameres, due to the introduc- tion of an extra half-metamere, the halves of the 15th meta- mere do not unite with one another. In fig. 34 two modifications of the metameres are introduced in front of the 15th metamere. Anteriorly the line between the 4th and 5th metameres fails to meet above. The 7th metamere is compound, so that an extra half-metamere is introduced on one side (the left); nevertheless the two vasa deferentia open on the same segment (16—15). In fig. 35, a,B,c, we find again two separate modifications. A small half of the 3rd metamere forms above a union with half of the same segment on the right side, and below it is wedged in between the 2nd and 4th metameres. Again, further back a half-segment is introduced on the left side between the 8th and 10th metameres (or, if the more anterior extra half be counted, between the 9th and 10th). The openings of the vasa deferentia were not made out, but the ovaries occurred in consecutive segments, i. e. on the 13th half-metamere of each side if the more anterior half-metamere (3rd) is not counted. If this was counted, then the ovaries would lie in the 14th half-metamere of the left side, and in the 13th half-metamere of the right side. In fig. 36, a,B,c, a spiral! is introduced that involves the 11th, 12th, and 13th metameres, as shown in c. An extra half- metamere (12th) is introduced on the right side. The open- ings of the vasa deferentia were not seen, but dissection showed that the ovaries are in consecutive segments. In the left side the ovary occurred in the 14th half-metamere, and on the right side in the 16th half-metamere. In fig. 37, a, B, the 12th metamere is compound, and we find that the vasa deferentia open on correspondingly shifted meta- meres. But, besides this, the openings of the right side of the body have doubled, so that the 15th—16th metamere has two openings, right and left. In fig. 38, a,B,c (lateral), a half-metamere is introduced between metameres 3 and 4. The openings of the vasa ! This case belongs to a preceding category, D. 416 T. H. MORGAN. deferentia are on the same segment. (It seems to me very doubtful whether the surface line of the 4th metamere is due to a half-metamere introduced.) In fig. 89, a,B,c, a compound metamere occurs on the 15th segment (15—15). The opening of the right vasa deferentia alone was visible on the exterior. The ovaries were found in the 13th metamere. In fig. 40, a,B, we find a spiral involving the 8th, 9th, and 10th metameres, introducing an extra half-metamere on the right side. The vasa deferentia open, nevertheless, on the same metamere (15th—16th). Another modification starts at the 15th metamere, and involves a few more posterior segments. In fig. 41, a,B,c, we find four compound metameres (see c), followed by a spiral that involves segments immediately in front of the region of the opening of the vasa deferentia. Fifteen half-metameres of the right side correspond to twenty of the left side. The openings of the vasa deferentia appear on the same ring, and this ring is made up of the 16th half- metamere of one side and the 2lst half-metamere of the other. In fig. 42, 4,B,c, we find a compound metamere (3—3), fol- lowed by a complicated spiral (c) that involves the 14th to the 21st metameres. The openings of the vasa deferentia were not seen on the outside. Unfortunately I have no records of the dissection of this worm. In fig. 43, a, B, c, we find two spirals and a compound meta- mere in the anterior region. The openings of the vasa de- ferentia occur in the last spiral. The opening on the left side is in the 18th half-metamere, and on the right side in the 16th metamere. In fig. 44, a, B, a compound metamere (6—#) is first found, followed by a spiral belonging to category vii, Pl. 40, fig. vit, and this followed by another compound metamere (18—15). Three extra half-metameres are introduced on the right side. The openings of the vasa deferentia are 1 Belongs to another category, i.e. D. A STUDY OF METAMERISM. 417 doubled on both sides. On the left side the openings are found on the 15th and 16th half-metameres. On the right side of the body the openings appear on the 20th and 2lst half-metameres. In fig. 45, a,B,c, we find the 6th metamere somewhat reduced on the left side. We have an intercalated half-segment (15th) on the right side, followed by a spiral involving the 15th to the 19th metameres. The latter introduces an extra half-metamere on the right side. The openings of the vasa deferentia were not found. Dissection showed the ovaries in the 18th half- metamere of the left side, and in the 21st half-metamere of the right side (in the half-metamere that follows the 18th on the other side). In fig. 46, a,B,c, an extremely modified anterior end of a worm is drawn. As the reconstruction shows, many extra metameres are introduced on the left side of the body. In the exterior of the 24th and 25th left half-metameres swollen areas seem to point to openings of the vasa deferentia, but none appear on the right side. In the dissection of this worm no ovaries could be found, and no data throwing any light on its “ make-up” were obtained. Conclusion.—It is very difficult to reach any conclusion in the face of so much conflicting evidence. It would seem as though an intercalated segment might or might not affect the position of the reproductive organs. If to get out of the dilemma we assume that some of these cases are not due to the intercalation of true half-metameres, but to unilateral division of metameres, then other difficulties are encountered as great or greater than those that appear on our first assump- tion. It was only after the preceding pages were written and the hopelessness of any explanation realised that a partial solution—or what seems at first to be such—suggested itself tome. Such cases as those drawn in figs. 41 and 46 might be due to a regeneration in the adult of anterior meta- meres. We see that this might offer an easy escape from thedilemma. It seemed also to explain why the segments in some of these worms that carry the openings of the vasa deferentia 418 T, H. MORGAN. and the ovaries are so far from the anterior end. Here more segments might have regenerated than were lost. Fortunately we are dealing with an hypothesis that could be tested by experiment. I at once set to work to determine whether or not the same number of segments reappeared when the anterior ends were cut off. I also wished to find out whether a greater number of modifications than in the average worms were thus produced. A later section gives the result of these experiences. IV. Stupy or Emsryos. Capsules containing embryos of L. foetidus were collected, and the embryos killed immediately on their emergence from the cocoon. Out of 170 embryos there were twenty-five that showed abnormalities in the arrangement of the rings. This is in the proportion of 1 to 5:2. In other words, there are fewer cases of abnormalities in young worms than amongst adults. In the latter the proportion was one to two. This apparent contradiction finds its explanation in the fact that the adults are often found regenerating lost metameres, and proportionately the number of abnormalities in these newly formed parts is greater than in the embryos. This will be taken up more fully in another section. The number of metameres in the young worms that have just left the capsules depends to some extent, as the following table shows, upon the size of the young worms. ‘The size seems to be in general connected with the number of embryos that develop in the capsule. Amongst the worms in the same capsule, however, there are variations in size, and we can only make some such general statement as this, that where many embryos are present in the same capsule they are each smaller when they leave the cocoon than when fewer embryos are present. The following table indicates, in a rough way, the rela- tion between the size of the embryos and the number of metameres : A STUDY OF METAMERISM. Size of Worm. 419 No. 1 111 metameres Very large. ene 102 ef Large. 3 100 : ms awed: 85. a Small. shat : ‘ 85 3 : F Bs Sets ; : 78 . : ‘ a ey 67 : Very small. Numbers 6 and 7 came from a capsule that contained 12 worms. The following table gives the record of twelve young worms (including those of the preceding table), showing the number of metameres present : 85 100 102 85 107 106 11 : : 78 9 ‘ : ; ; 67 Seal : 87 Pa . 85 , 12 97. 12 ) 1110 92°5 aonronrrFr wnr I attempted to determine whether the young worms that had abnormalities came in proportionally larger numbers from individual capsules. Capsules were also examined to see whether the number of embryos in a capsule was a factor in the production of abnormalities. This examination ought also to show whether the eggs from certain worms had a tendency to produce abnormally joined metameres; if, for instance, in all the capsules that contained a large number of embryos many were abnormal, it would follow, with a great deal of pro- bability, that the crowding (or smaller amount of food, &c.) brought about (directly or indirectly) the result. If, on the other hand, certain capsules, regardless of the number of 420 T. H. MORGAN. worms present, showed a large percentage of abnormalities, then the cause probably arose in the egg. The following tables give the data that I have collected. The number of recorded cases is smaller than I could have wished, but the evidence is all in one direction and fairly con- vincing. The capsules were isolated and the worms collected as soon as they crawled out. Number of embryos in a capsule. Number of abnormal worms. No. 1 6 ‘ : ‘ 1 the 3 0 aye 6 2 si 6 0 San gs 12 il 5 (6 5 1 ar Oe 3 1 eae ss 4 0 ar Ae 12 1 ” 10 5 ] 55 ake 12 1 ” 12 8 if Conclusion.—In none of the capsules is there a prepon- derance of abnormalities, nor does there seem to be any rela- tion between the number of worms in a capsule and the presence or absence of wrongly united metameres. We must conclude that neither of the possible causes sug- gested above is active in producing the abnormal embryos. We get no clue from these data as to any outside forces producing the result ; and although we have not by any means exhausted all the possibilities, still it seems not improbable that in each particular case local internal conditions determine the result. All of the commoner forms of abnormalities recorded for the adult worms are also found in the embryo. This applies to abnormalities in the anterior part of the body as well as in other parts. There are more abnormalities in the tail end than in the head end. A few typical abnormalities found in some of these embryos (young worms just out of the capsule) are shown in PI. 42, A STUDY OF METAMERISM. 421 figs. 75—78. In fig. 75, a,B, isdrawn a compound metamere ; in fig. 76, 4, B, a spiral of category vir; in fig. 77 a spiral of category vill; and in fig. 78 a longer spiral of category x. A number of adult worms was examined, and the following table shows the number of metameres inthem. Mature worms were chosen as far as possible. No: 101 ae 102 me 92 ea: 85 (perhaps regenerating). nea oF EP) 6 97 Pe iL 105 at KOL - ; ‘ : 95 Pree |S ie : é : 101 lO 88 (probably regenerating). ro es 9 Or 85 513 eM alin? 84. mela : 95 15 ) 1416 94°4 average. It will be seen, if we may judge by the figures given, that few, if any, new segments are added to the worm after it leaves the capsule. Another count of 25 worms gave the following figures :— 99, 89, 85, 100, 90, 96, 95, 95, 97, 101, 101; 105, 101, 106, 88, 84, 88, 95, 100, 92, 100, 91, 105, 92, 96. These numbers give an average of 99°6. If we combine this with the average of the 15 cases given above (94°4) we get 97 for the average number of segments of 40 worms. V. ABNORMALITIES AT THE PostreRIon EnpD. In one lot of 525 worms 40 were found showing newly formed posterior ends due to regeneration. In these 40 worms we find— 422 T. H. MORGAN. A. Regenerated portion just beginning to show metameres. 4 B. Regenerated portion normal a. Number of new metameres : : 6 Bs hiss - : : 6 c. Regenerated portion having one abnormality . 12 a. Number of new metameres ; : 19 b. a3 55 6 (p: Be e 30 ee, if 16 é. ss As 25 iy 0 2» 3 UE Bs $3 : q 4 h. 0 = 5 ; ll a 39 33 4 je ~ oS : : (several) k. 5 5 : : 15 l. as 33 : : 6 p. Regenerated portion with two abnormalities. 7 OF (Number of uew metameres not recorded.) E. Regenerated portion with three abnormalities . . 4 rv. Regenerated end with more than three abnormalities . 12 In p, £, and F the number of new segments was not counted, but measurements were taken of the regenerated portions for comparison with c, and were as follows: c. D. E. F. 1 18 mm. 26 mm. 27 mm. 25 mm. 2 IBY Ay ilies 20). ¥s5 gore 3 10 ” 21 ” 9 oP) 25 3 4 (ae 16 8 ss 22 =, 5 Bes SS 4 ie" oo. 6 oh GE Sie es 18 mm. Hl ee 7 avs 6) 98 ,, 19; 5 8 ae 164 mm. i es 9 Aes 145; 10 2 ”? 13 ” ll 2 5 12, 12 2 + As 12) 80 , 13\) 294 14) 65 mm, average length. 182 mm, A STUDY OF METAMERISM. 423 These measurements show conclusively that the greater number of new abnormalities occur not more frequently at the beginning of the newly formed portion than throughout all the later period of growth. Summary.—In the 40 cases mentioned above 4 were not well enough developed to furnish any data. In the36 that remain there were only two that did not show any abnormality, and it is noticeable that each of these had only formed at the time a few new segments (six each). We must conclude from the data that about 18 worms out of every 19 would show abnor- malities in the regenerated posterior end. In these same worms the number of abnormalities in the region of the body anterior to the regenerated end was recorded. This gives us data to show whether the irregularities are due to inherited peculiarities of the tissue or to the conditions acting during regeneration. In the 12 cases recorded in c 9 were normal in front of the regenerated portion, and 3 showed ab- normalities,—that is, as 1to 3. In the6 cases recorded in p, 5 were normal in front and | abnormal (1 to 5). In the 4 cases in £ there were no abnormalities in front. In the 12 cases recorded in r there were 2 cases unrecorded, and the remaining 10 had no abnormalities in front. To sum up, the evidence points unmistakably to the con- clusion that the abnormalities found so frequently in regene- rated portions are due to the conditions acting during regene- tion (from within or from without), and in no way connected with an hereditary. tendency to be more abnormal in one case thaninanother. That is to say, the tissues of a worm that has developed normally from the egg are just as apt to develop irregularly in regenerating as are the tissues that have de- veloped irregularities during embryonic growth. Heredity seems to have nothing to do with causing the abnormalities. In fact, one might say that the tissues inherit a strong tendency to regenerate normal metameres, but the means at command are so imperfect that abnormal results are frequent. Will the large proportion of abnormalities present in regene- rated worms, taken in connection with the number of worms found VoL. 37, PART 4,—NEW SER. FF 424 T. H. MORGAN. regenerating naturally, account for the difference between adults and embryos? We have seen the number of worms found showing regenerated (and regenerating) posterior ends to be 1 in every 18 (1 to 12); therefore in a lot of 225 we should expect to find 19 worms having new posterior ends. In our lot of 225 worms we found 100 abnormal worms. If we deduct from this 100 the number of those that should have regenerated tails (19), we should have 81 cases remaining. The proportion would then be 81 to 225, or 1 to 2°8 (approxi- mately 1 to 8). We found, however, that the proportion of abnormal to normal embryos in A. foetida was only 1 to 5. The difference that is still found is probably due to several causes, and these are not in the least of a hypothetical nature. Firstly, after a time a new regenerated end cannot be distin- guished from the rest of the body. Secondly, a small per- centage of regeneration must also take place in the anterior end. Thirdly, the difficulty of seeing the abnormalities in the embryos is much greater than in the adult, and defects may have occasionally escaped even a careful examination. Fourthly, the data is drawn from too small a number of cases to make an exact agreement very probable, even if it did exist. In the light of these conditions I think the closeness of the result is as near as could be expected. A number of worms were examined in which the posterior end of the body was regenerating, to see if any could be found in which the tip of the “ tail” showed, in process of formation, modifications of the typical arrangement. Several were found, a few of which are drawn in PI. 43, figs. 79—81. The first of these (fig. 79, 4,B,c) shows a compound metamere in process of formation from the terminal piece or telson. On the side that forms the “split” the telson is proportionately longer than on the other. In fig. 80, a, B, c, the end of the body is very much modified, and a spiral is in process of formation. The reconstruction (c) shows sufficiently the conditions present. In fig. 81, a, B, c, another modification is present. Above and on one side of the telson the outline of a metamere is seen, while A STUDY OF METAMERISM., 425 below this line only extends a third across. Whether this would ultimately form a complete ring, or whether we have here the beginning of a spiral starting with a half-compound metamere below, is uncertain. Several vertical longitudinal sections were made through these “ tails,” but no additional information was gained from the sections. The irregularities in the positions of the rudi- mentary body cavities was very apparent. VI. Mopirications oF INTERNAL STRUCTURES. Before any conclusion can be reached as to the value of the irregularities seen on the surface, we must examine the condi- tion of the internal organs, particularly the arrangement of the septa. We can formulate this general statement, that the ar- rangement of the septa generally conforms to the curves of the lines seen on the surface. This means that the phenomena are deep-seated, and that all the struc- tures of the body are involved in the new arrangement, and not merely the external surface lines. Exceptional cases are not uncommon, in which we find two principal departures from the rule. First, the arrangement of the septa does not always agree with the surface lines. Some- times the arrangement of the septa is more perfect, and some- times simply different. Secondly, the arrangement of the somatic attachment of the septa is sometimes different from the splanchnic attachment. Usually in this case the somatic portions of the septa follow the lines found on the surface, while the splanchnic attachment differs in its arrangement, and is usually more irregular in its form. If a worm having a compound metamere be opened we find the septa arranged as shown in fig. 61, 4,B. One septum (s) reaches only to the mid-dorsal line. If we remove the diges- tive tract we find the same septum ending freely below, just before reaching the middle line(s’). This half-septum is found to correspond to the “split” in the compound metamere, and lies between the two half-metameres of that side. 426 T. H. MORGAN. The septum is attached on the somatic side to the body-wall along the “split.” Centrally it is attached to one half of the splanchnic wall of the digestive tract. We should find on sectioning such a worm that the half-septum had a double wall, and was like a true half of a septum in every respect. We should see also, as a result of this arrangement, that the body cavity of the single half-metamere is continuous in the mid- ventrai and mid-dorsal lines with two half-metameres on the opposite side. The number of the nephridia corresponds to the number of the half-metameres, as shown in fig. 62, so that there is one more nephridium on one side than on the other. The nervous system is often modified, as shown in fig. 62. Here each of the half-metameres is seen to receive its full number of nerves from the ventral cord. This is not always the case, for, as shown in fig. 61, B, the double side gets only the supply normal for a single half-metamere. In figs. 63 and 64, B, we see other irregularities in the nerve-supply. This may be connected with the degree to which the half- septum approaches to the mid-ventral line. When it falls far short the nerve-supply is less than when it reaches the mid- line. The nephridia, however, that lie laterally in the body are invariably doubled. In fig. 64, a, the surface line between the half-metameres of the compound metameres is not very extensive. The figure represents the body-wall flattened out (the worm had been previously opened). Fig. 64, B, shows that the septum s is attached to the somatic wall over a correspondingly small area. The condition of the septa in the spiral modification is very interesting. One of the simplest cases is shown in fig. 65, A,B. The first figure shows a short spiral beginning and ending above (category vit1). On opening the worm from below (fig. 65, B) the septa are seen to follow the same arrangement, both on the body-walls and on the intestine. There results a single spirally winding septum beginning with the half-septum of the compound metamere anteriorly, and ending with another compound metamere poste- A STUDY OF METAMERISM. 427 riorly. In other words, there is a continuous body cavity (cclom) lying between the coils of the septum, and this cavity is continuous from the anterior to the posterior end of the spiral. Fig. 66 shows a similar spiral, the only difference between this and the last being that in the former (fig. 65) the anterior end of the spiral abuts against the septum in front, while here it ends freely. Fig. 67, a,B,c, shows a similar but longer spiral. Opened dorsally the septa are found as in B, and it will be seen that they follow the course of the outer spiral. In the middle line the septa over the intestine show a tendency to irregularity (even more so than the figure shows). In c the septa are drawn as seen from below. They run obliquely over the lower wall of the digestive tract. In both B and c the figures were drawn from preparations made by dissecting free the digestive tract and its attached septa from the body-walls. These examples will suffice for the regular forms, where surface spirals and septal spirals agree; but this, as stated above, is not always the case. In fig. 68, 4B, we find on the surface a spiral shown in a; that belongs to category x (Pl. 40, fig. x). Opening the worm we find that the septa are also spirally arranged, but the spiral is shorter than the surface spiral by one turn, although both begin and end on the upper surface. Fig. 69, a B, shows a parallel case, where the internal spiral is shorter than the external. In this case the septal spiral is shown only in its somatic attachment. In fig. 70, a,B,c, we find a surface spiral like the last. We find the septa arranged dorsally as shown in B, and on the ventral as shown in c. It will be seen that the external and internal (septal) spirals belong to different forms (cate- gories). ‘The external belongs to category x1 (Pl. 40, fig. xr), while the septal belongs to category 1x (extended), Pl. 40, fig ix; Fig. 71 shows a form in which the external line between two metameres failed to meet in the mid-dorsal line. In dissection 428 T, H, MORGAN. the arrangement of the septa was found to be normal ; that is to say, the septum has met normally across the mid-dorsal line, but the external lines have failed to do so. In fig. 72, a B, a surface spiral is drawn, and in B the arrange- ment of the septa beneath is shown. The septa shows an abnormal arrangement, but this does not correspond to the surface lines: it is more of an approach to the normal. Fig. 73 is from a compound metamere but in which the septa form a simple spiral making a little more than one turn of the body. It is not always easy to picture to one’s self the relation existing between septal and surface lines when the two do not correspond. In the simpler cases it is easily understood. When two spirals are formed of different lengths, it is due to the fact that the septa sometimes unite with one another beneath the surface before the surface spiral is brought to an end. That is, if both start together one may continue for a longer time than the other. We must conclude, then, that while the two usually vary together, yet they may vary inde- pendently. 1 have dissected a far greater number of worms than recorded above, and from that number have selected those given as the most interesting examples to illustrate the main points that concern us at present. Many other rela- tions have suggested themselves, but I have only wished to go into the subject as far as concerns the matter in hand. VII. Stupy or Potycu#tous ANNELIDS AND LEECHEsS. A large number of species of polychztous Annelids have been examined, and many modifications in the arrangement of the metameres have been found in them. From this number I have chosen a very few for description, since a large number of cases have been given in the papers by Cori and Buchanan. The modifications that are figured here are all taken from specimens of Amphinome. In fig. 82, a B, we find above B a compound metamere with the split on the left side. In the ventral side of the worm, a, we find the left anterior half of the compound metamere A STUDY OF METAMERISM. 429 wedged in ventrally between two segments, while the other left half completes the ring. We have here evidently a modi- fication of the compound metamere similar to category 11, Pl. 40, fig. 11. In Pl. 48, fig. 88, a, B, we see a metamere imperfectly deve- loped on the right side. The dorsal parapodium and gill are absent on this side, but the ventral parapodium is present, as seen in side view in B. We have here probably a case of incomplete development of a metamere of one side. Another somewhat similar modification is shown in fig. 84, AB. The first of these shows the ventral side, and the middle segment of the three drawn is seen to be imperfectly deve- loped. The ventral parapodium of the right side (left of figure) is present, but the metamere does not reach as far dorsally as the line of dorsal parapodia (see B). On the left side of the body we find the metamere completely undeveloped. We seem to have here a half-segment of the right side that has not fully developed even on that side, but has overgrown the mid-ventral line on the left side. Fig. 85 shows a spiral beginning and ending on the dorsal surface. Seven segments are involved, giving five turns to the spiral. The large number of abnormal forms found in Amphinome is in part due no doubt to the frequent regeneration of por- tions of the body that takes place. The segments are broad, and for this reason it is surprising to find abnormal com- bination of the metameres so frequent. In other poly- chetous Annelids, where the metameres are very narrow, the false unions and imperfections of the metameres are very numerous. Cori has figured several modifications, and I have found similar ones that are exceedingly difficult to explain as simply due to modifications of half-metameres. Such cases are particularly common in polychetous Annelids. The majority of these very abnormal forms are, I think, due to regeneration rather than to egg-variation. Since the method of regeneration is more irregular than the growth of the 430 T. H. MORGAN. embryo from the egg, it becomes much more difficult to explain on the metamere assumption these variations due to regeneration. A number of leeches were examined to see whether the annuli that make up the broad metameres ever showed false unions like those found in the earthworm’s metameres. Such modifications are not rare. In fig. 86, a B, the line between the second and third annuli of the ninth metamere failed to extend above over to the left side of the body. Again, in this same metamere we find that the line between the 5th and 6th annuli turns forward to join the line between the 4th and 5th, so that the last annulus of this metamere forms a compound annulus with the first annulus of the next metamere, producing a short spiral form. A more complicated spiral is shown in fig. 87, 4,8. The 16th and 17th metameres are involved. The spiral com- mencing in the 16th metamere runs over into the next meta- mere, as the reconstruction B shows. The annuli are also imperfectly joined at the sides, as shown in B. This last result is due to a failure of the annuli to unite perfectly with one another. A third spiral-form, involving a double compound annulus, is shown in fig. 88, 4, B. A half-compound annulus above and on the right side forms a spiral with a half-double com- pound annulus above and on the left side. The first half- compound ring is the 3rd. We see that the annuli of leeches show many of the same sorts of false unions that are present in the metameres of the earthworm, but in the leech this represents only a superficial alteration. The modifications in the leech would correspond to those modifications occasionally found in the earthworm, where the septa are normal but the surface markings abnormal in their arrangement. It will be noticed that in the figures given no new annuli are introduced into the metameres, and only the method of union of the annuli varies, This result may only be due tothe A STUDY OF METAMERISM. 431 relatively few leeches examined, although the number was sufficient to show that the modification involving additional half-annuli, if it occurs, must be rare. VIII. Summary. In the papers of Cori (18) and myself (32) the same ex- planation was offered as to the origin of the simpler cases of abnormal unions. Cori said, “ Nun kann es sich ereigen, dass in der einen K6rperhalfte wahrend der Entwickelungsperiode ein Ursegment mehr gebildet wird, dem auf der Gegenseite kein Ursegment entspricht. Auf diese Weise wird die Bildung jener Falle von Schaltsegmenten verstiindlich, wie sie im vorherge- henden anatomisch beschrieben worden. Betreffend das Ver- haltnis eines Schaltsegmentes zu den iibrigen Metameren des Korpers sind zwei Méglichkeiten vorhanden. Es kann sich ein Schaltsegment vollstandig ausbilden, und kann von den anderen Segmenten abgegrenzt bleiben oder es geht eine Verbindung mit dem vor oder nach folgenden Metamer ein.’ My own statement was very similar:—‘‘* We know that in the embryo the metameres are laid down right and left of the middle line of the body in blocks of mesodermal tissue; that under normal conditions these hollow blocks come to lie exactly opposite (right and left of) one another, so that the opposite pairs unite across the median dorsal and ventral lines. If we conceive that the blocks are slightly displaced on one side, or that two consecutive blocks of one side are smaller than two of the opposite side, we may have, as a necessary mechanical result of the relative position of the block, a split metamere.” ‘To account for the spiral arrangement I said, and I am still of the opinion that this is the true explanation, “Tf we imagine one of the mesodermic blocks of one side to be larger either above or below (but not both above and below), so that above (let us say) it opens into two body-cavities of the opposite side, while below it opens into but one, then we have produced the conditions necessary to start the spiral. Each of the consecutive blocks on the same side as the supposed larger block will open below into its proper opposite block, but 432 T, H. MORGAN. above (on account of the first displacement) into the one lying in reality behind it. . . . The spirals when once started do not run on continuously, but end after passing around the body several times. The ending likewise finds its explanation in the inequality of the blocks of opposite sides.” I did not in this preliminary paper consider a point with respect to the ending of these spirals that may be spoken of here. If a spiral once started was dependent in order to end, on another accidental displacement appearing, so that the spiral is, as it were, satisfied, the chances are that most spirals would reach a prodigious length before terminating. As a matter of fact the spirals are generally short, the shorter the commoner. This shows, I think, that the conditions, after a spiral has been started, are of such a nature that the chances are the spiral will soon end itself. It is easy to offer a formal explanation of why this is so. The shifting of the blocks by the double union established at the anterior end! will tend to displace the block behind, so that two of one side will come to correspond to one of the other. In the preceding section of the paper there are in reality two topics that have been discussed in close connection with each other. The question of the shifting of the reproductive openings was considered in connection with the origin of the compound metameres and spirals. The reason for doing so was based on the hope that the one relation might help in the interpretation of the other. Moreover, in respect to the abnormal position of the open- ings of the reproductive organs, we have considered first those cases where, in worms with normal metameres, the openings have shifted or doubled ; and secondly, those cases where there were both abnormal metameres and abnormal openings of the reproductive organs in the same worm. The former of these subjects has been already noticed by previous authors. Beddard (5) in 1886 published some most interesting observations on variations in the reproductive organs of Perionyx excavatus. 1 It is assumed that the spiral starts at the anterior end. But it is also possible that the shorter spiral arrangement involves all the segments, A STUDY OF METAMERISM. 433 In a lot of 489 individuals there were found seventeen that showed such variations. Benham (6) described in 189] a variation in the openings of the reproductive organs of Lumbricus herculeus. The vasa deferentia appeared on the 14th and 15th metameres.! With respect to the abnormal unions of metameres, and the relation borne to the position of the reproductive organs, there is little to be added to the conclusions reached in the sections dealing with abnormalities in front of or included in the 15th metamere. ‘T'wo cases were there sharply separated,—those cases where there is a half-metamere more on one side than on the other (C, second case, E), and those cases where the spiral does not introduce an additional half-metamere on one side (B, C, first case, D). The results from the latter cases are in full harmony with the interpretation of the spiral stated above. In those cases where an additional half-metamere is introduced there are certain examples that seem to verify the same statement ; but there are other cases, and these form the majority, where the results will not bear this interpretation. We have, then, to assume, in some of the cases recorded, either that the half- metameres are not equivalent to the normal, or else that the reproductive organs do not give unfailing evidence of the change that has taken place. Anyone who has studied the facts will, I think, agree with me that it is far more probable that the reproductive organs have appeared on a wrong (?) half-metamere than that the half-blocks are not equivalent. This leads us to another side of the problem, as yet not dealt with. It is natural to think of every half-block of one side as having a half-block that belongs to it on the other side; in other words, to look upon one of these blocks as predestined for the other, and we marvel to find them separated. I think there are no real grounds for such an 1 Bateson’s memoir, ‘ Materials for the Study of Variation,’ that has just reached me, refers to numerous cases of abnormalities in the reproductive organs of earthworms recorded in a paper by Michaelsen, ‘Jahrb, Hamb, wiss. Anat.,’ vill, 1891. I have not had access to this paper. 434, T. H. MORGAN. expectation. If the facts recorded in the preceding pages indicate anything clearly, it is that the union of the blocks across the middle line is the natural result of their size and of their position. Under normal conditions the blocks are so placed that they unite pair for pair. If the order is dis- turbed they unite differently. This is brought out most strikingly in that case where there were 134 half-blocks on one side and 118 half-blocks on the other. Here throughout the body there are very few half-segments that are united with the one opposite bearing the same numerical relation from before backwards. In the light of these statements I think we ought not to look to any one half-metamere as predestined to carry under all conditions the openings of this or that organ. Rather should we expect that those segments, which happen to get so placed in the body of the worm that they correspond to the normal position for a particular organ, will go ahead and develop that organ. We might say that the position in which an organ will develop is determined by a definite region of the body irrespective of how that region has gotten into the necessary position, provided this happened when the cells were still undifferentiated. The question is too large to discuss here, and the facts too meagre. There is a liability to err if the statements made above in regard to the methods of union of the metameres be applied to all cases. All that I contend for is that the majority of cases will bear this interpretation. Sooner or later one is sure to come across individuals where it is impossible to see the appli- cation of the theory. Particularly have I found this true in the Polycheta. I have met such cases quite often in examin- ing the regenerated anterior ends of worms where the head had been artificially cut off very obliquely. Sometimes pieces are left, where the new head joins the old body, that cannot be interpreted as half-metameres. Whether all of these unusual cases are the result of regeneration I am not prepared to say. Many of them seem to be; others, so far as I know, may have come in from the egg, A STUDY OF METAMERISM. 435 The results that have, on the whole, been the most puzzling to me, and which I cannot pretend to entirely explain, are some of those cases where the internal and the external spirals do not agree. It seems quite certain that in the far larger number of cases the false unions are at bottom, the result of imperfect joining of the mesodermic blocks, and that the ectodermic grooves mould themselves on the internal arrangement. But in cases of disagreement it seems clear that the outer grooves have to a certain extent an independent individuality, and may behave differently from the spirals in the mesoderm. When we find similar spirals in antenne, &c., it is not, after all, so surprising that the outer body-wall of the worm should show this independent variation. IX. Mopirications 1n ANTENN#H OF ARTHROPODS. The antennz of Arthropods are made up of a series of seg- ments or rings. In some Arthropods the segments are long and relatively few in number. In others the segments are narrow rings and very numerous. The latter kind are more likely to show variations in the arrangement. The antennz and antennules of the lobster (Homarus vulgaris) I found showed many variations. Perhaps the lobster’s antenne and autennules may owe many of these false arrangements to the fact that if they have been lost new ones will regenerate, but that all the abnormalities come from this source is extremely improbable. Ihave found in the antenne of the lobster nearly all the sorts of variation in the arrangement of the rings that are to be found in the earthworm. A few of these variations are shown in Pl. 43, figs. 90 to 95. Fig. 90, a,B, shows a compound ring, Fig. 91, A,B, shows a double compound ring, where three half- rings of one side are united to one of the other. Both ofthese modifications are common, particularly the former, at the broad base of the antenne. In nearly all cases the ‘‘splits” run in from the sides. Rarely a modification similar to that shown in fig. 92, a,B, is 436 T. H. MORGAN. found. Here the “split” is found over the broad surface of one side, and not over the other. ‘This condition is never found in the falsely joined metameres of the Annelids, for even in those cases where a metamere appears below, and not above, it is also found to extend over and on one side (lateral) of the body. Fig. 95, a, B,c, shows a spiral formed by a union of three compound rings above and one below, as shown in the re- construction. Fig. 93, 4,B,c, shows a longer and more complicated spiral. In the middle of this spiral we find a “double spiral” developed. Fig. 94, a,B,c, alsoshows a complicated spiral. These spirals in the antennz are similar to the spirals found in the earth- worm, and need no further description. The antennules of the lobster also show the same modifica- tions. Both the antennz and the antennules are flattened from above downwards, and the splits extend as a rule on the lateral sides of the appendages. ‘The basal joints of nearly all antennz show imperfect rings. The greater number of abnor- malities occur in the proximal portion of the antenne, and are found in less and less frequency as far as the middle region. They are rarely found in the distal ends of the antenne, and it is important to notice that distally the rings become much longer relatively to their breadth. Another interesting condi- tion is found. By far the greater number of “ splits” occur on the convex side of the antenne and autennules, particularly in the latter. There results a larger number of half-rings on the convex side. In one case there were ten more half-rings on the convex side of one antennule than on the opposite side. When we see that it is normal for the antennules to bend outward, i.e. to turn their convex side towards the middle line, it seems fair to draw the conclusion that the tendency of the antennule to turn to one side is the direct or indirect cause of the greater number of rings on the convex side. There is no reason for believing that the greater number of rings causes the turning, since the antenna is bent whether more rings are present on one side or not. Processes that go A STUDY OF METAMERISM. 437 on in the cells of the antenne before each moult will deter- mine where the new lines of division between the segments come in, and in some way the bent condition of the antenna seems to effect to a large extent the formation of the new lines. I have also examined the development of the antenne in the lobster. At the third (?) moult the terminal division of the endopodite of the antenna contains within it a rod of cells, and these show alternate constrictions to form the lines between consecutive rings. ‘These lines of constriction are exceedingly near to one another, since the segment containing the new series is considerably shorter than the total number of segments that ultimately develop from it at the next moult. It is easy to see that under these cramped conditions the arrangement of the division lines might easily be disturbed by local causes. Au examination of the antenne of certain insects where regeneration of the antenne of the adult is not probable (and that in the larva hardly possible) shows that in those forms where there are a large number of rings, and each ring very short, variations exist in the arrangement, while in those forms with long rings such variations are absent (or not found commonly). In the locust I have not seen any misformed rings in the antenne (in as many as fifty individuals examined). In the cockroach (Blatta), where the rings are narrow, compound seg- ments and short spirals are frequently found both in the larvee and the adults. X. AspNnorMAL MetTAMERISM OF LocUwstT. It seemed to me highly improbable that such specialised metameric forms as the Crustacea and Insecta, in which the metameres are so definitely limited in function, number, and position, should show any variations comparable to those in the earthworm. However, I had brought to me a locust (‘grasshopper ”’) in which there was a half-segment more on one side of the abdomen than on the other. In fig. 89, a, is drawn a dorsal view of this locust with the 438 T. H. MORGAN. wingsremoved. Corresponding to the 5th metamere of the abdo- men we find a compound metamere, so that one half-metamere of the left side corresponds to two half-metameres of the right side. A figure of the abdomen from below is drawn in fig. 89, B, and the abdomen seen from the right side is seen in fig. 89, c. On account of the interpolated half-metamere the abdomen is bent towards the left. The interpolated segment reaches exactly to the median line above and below. The abdomen ends normally, having all of the accessory structures of the male perfect in all details. Fig. 89, c, shows that each of the half-metameres on the right side of the compound metamere bears stigmata, so that there is one more of these on the right than on the left side of the body, i.e. eight on the right (one more than normal) and seven on the left. If we assume each of the right halves of the compound metamere equivalent to a true half-metamere of the series, then each of the half-metameres following the compound metamere is joined to a half-metamere that is not its normal half. As a consequence, one new additional half-metamere will have to be formed at the end of the series on the right side to make a perfect ending to the abdomen. In so highly modified a form as the locust this view seems very improbable. An alternative view would be to suppose that a half-metamere has been intercalated between the 4th and 5th, or between the 5th and 6th, and that the intercalated half has united across the middle line to the 5th in common with the normal half of the 5th. We might think of this intercalation as due to a half of one metamere dividing into two, or we may suppose that when the ventral mesoblast broke up into a series of blocks it formed on one side one more block than on the other (one more than the normal). If the last view be true—and it seems more probable than any of the other suggestions—we see at once the futility of trying to explain the conditions on an assumption of predestined right and left half-metameres. When we recall that the whole ventral plate of the insect becomes segmented at the same A STUDY OF METAMERISM. 439 time, and when we recall the complicated modifications of the terminal segments, it seems highly probable that posterior to the compound metamere no shifting of the segments of one side can have taken place, and that no one segment has divided, but that the plate as a whole has broken up into a greater number of half-metameres on one side than on the other. XI. Srupy or THE CoLouR-BANDS OF ECHINODERMS. In the summer of 1891, while in Jamaica at the Marine Laboratory of the Johns Hopkins University, I began a study of the colour-bands on the arms of the Ophiurians in the hope of getting results that might help towards a solution of the problem of metamerism. The work was laid aside for two years, during which time the colour was so far removed from the alcoholic specimens as to render renewed study unprofit- able. My earlier results, although not carried very far, point to certain conclusions that are not without interest in the present connection. In several species of Ophiurians (“ brittle- stars’) the arms are banded at regular intervals by pigment of a different colour from that of the rest of the arm. Each band of pigment is confined to a single segment—metamere— of the arm. In fig. 96 three such pigmented segments are shown. The colour of the band in this case is a rich orange- red, while the rest of the arm is an olive-green. Along the mid-ventral line there runs a longitudinal narrow line of orange-red pigment not shown in the figure. We find that a definite number of uncoloured segments alternate with a coloured segment. Three uncoloured seg- ments lie between the coloured segments, i.e. every fourth segment is coloured. Occasionally, however, the regularity of the arrangement is disturbed. Fig. 97 shows a common variation. Instead of finding a fourth coloured segment between the upper and lower coloured segments of the figure, we find two consecutive segments coloured, each over half its extent. The middle line of the arm marks the extent to which each is pigmented, the first half-band lying to the right and the more distal to the left. VOL. 37, PART 4.—NEW SER. GG 4.4.0 T. H. MORGAN. We see that the right and left sides of the arm vary independ- ently, so that each forms its half-coloured band. In the present case something has disturbed the regularity of the process, so that the colour has appeared too soon on the right side, as the figure shows. In fig. 98 we have asomewhat similar case. Here the half- bands are separated by an entire uncoloured segment. In other cases that I have seen the half-bands may be even farther separated by uncoloured segments. From a study of these and similar variations we find that the following relations exist. When a segment is only half coloured on one side it is generally, though not invariably, followed, sooner or later, by a segment coloured only on the opposite side. We might say that the colour-ring had split, and its two halves had appeared on different segments, so that when we find a half-band on the right we would expect to find its other half on the left, and vice versa. When half-bands appear, one of the halves seems always to come in too soon proximally. The other half then appears on its normal segment or beyond it. My material has been too limited to warrant any attempt at further explanation of the phenomenon, and no doubt a more extensive study would show exceptions to these statements that would render an explanation still more difficult. The three following tabulations of the colour-bands show the main variations that comein. The x indicates a coloured ring. The figures between these indicate the number of uncoloured segments. The half-coloured bands are indicated by x L. } and x } R., depending whether the colour is on the left or right side of the arm as looked at from below. 441 MUETAMERISM. AVSiLUDYS On First REcoRD. on >.< Gel >.< oe &6 Cel >< xX XO Ko KO Kod Xoo XO K OD KX OD X Gel >< Ge Se Ge) >< Ge) > Se Gel 5 Gey OS Gelb 2S Gr) Xe on on ine) 4 Ge) > Gel > GP) Gel >4 ar) >.4 Ge) >. Xoo Xo KN Fa XR daaiacs KX oO Koo X xX 4 x mel ox Gel 9 Gel < ine) 3K xcs Xo KX oO X co xX co ee pa HAM ANG XK IW@dam XK oo ees Hx x x pS aa XN xX oo XN XK x Broken. oo Ko K OD K OD K OD KX OD KX OD KO XK OD c2 X 0 K 20 X od X xX KX oO Ko XK OO Xoo KX oo X od K oO co KX oD K oO K OD Broken. Broken. H. MORGAN. T. 4.4.2 Seconp ReEcorp. pe a Pe xX om KO a KX WAIN GR IN INR KH KS XK OD wi IN Hi GR OK OD XK OO ere eae ak ix 45% - x * pa pa X om KX od K a ANG NIN Ge ier XK OD XK CO i re eS x x x a od Koo KX OD XK ay AG AN ied GV ig rn KX OD X OD XK CD a Sil ee x x x 443 STUDY OF METAMERISM. A Tarrp REcoRpD. S ec KO WM KH KOO K OD Keo KOO KH Kod Karrinanninee | KA Ka Xa Xa 4 ie ox x x = e 8 ee MR XD Xe0Xo2 Xe9 X29 Kemet ccs ets emar X | XA XH XK KAKA KA Xo Xe lee eel oe ee x x x X™ XOX KD KD KN | Ka XA KA xX Xm Xoo XK Xo KD KR | Xa XA XA XG pa pa A esl acl XK OD KK ay iG &K Alo XO KAKA KS XA XA KA XA XR Xa XRXGKR 4 Xen See OC x Xx Xx x x 444, T, H. MORGAN. A few of the more obvious relations that one finds in these tables may be pointed out. In some cases exceptions to the statements made above are found. At ain the third record, last (fifth) column, we see that a half-colour band is present on one side, but not on the other. And again this is seen at a in the third record, fourth column. In the third record, first column, we find in one region many irregularities present, although taken altogether the same number of half-rings are present on the right and left sides. In the third record each arm was found to have a new distal end that had regenerated. In this part every other segment was coloured. Whether this was a permanent or only a tem- porary coloration I cannot say. It would be interesting to find out whether in the same in- dividual similar variations showed a tendency to appear on the different arms. Even the limited data of the three tables give a slight amount of evidence in favour of such a view. The regular arrangement of the colour-bands on every fourth segment finds an interesting parallel in certain Annelids, where the coloured rings bear a more or less definite relation to the metameres of the body. Andrews (1) makes the following statement in regard to the arrangement of coloured bands in the polychetous Annelid Procerea tardigrada, which belongs to the family Syl- lidee:—‘‘ The female has a dark dorsal transverse band upon Somites 3, 6, 8, 9, 18, 17, 21, 25, 27, 29, 32, 35, 38, 42, 46, 49, 51, 58, 56, 57, 70, 71, 74, 77. . . . The non-sexual form has pigmented bands like those of the female, but arranged according to a definite law or general rule, to which the bands in the female conform also; bearing in mind that the female is formed as a cut-off part of the non-sexual stage, separating almost always just posterior to the thirteenth somite, and hence having thirteen less somites than that stage. In 110 individuals carefully studied, only three had the bud formed just posterior to the fourteenth somite; seventy-nine had an evident bud just posterior to thirteenth somite. “ Having tabulated the arrangement of the coloured bands in A STUDY OF METAMERISM. 44.5 these 110 individuals, there results the general rule that the bands occur upon the third and fourth somites, then upon every other or alternate one up to and including the twelfth, then (in the region of the bud) upon every fourth one up to and including the twenty-fifth, then upon every fifth one up to and including the forty-first, after which the exceptions become so numerous that no rule is evident. The examina- tion of so many cases shows a definite tendency to limitation in the bands to certain somites in the anterior region, and a greater and greater irregularity in the posterior region.” After illustrating some cases of failure in the normal arrangement of the bands, Andrews adds, “ These facts seem sufficient to indicate that we have in this Syllid a marked tendency to the acquirement of a regular metameric marking, which, however, does not coincide with the metamerisation of the somites, but tends to follow a special law best expressed in the oldest part of the body in which certain alternating coloured and not coloured somites are distinguishable—a series of groups or combinations of somites thus following one another.” It is not without importance to find in the typically meta- meric Annelids regular serial markings following definite laws in each portion of the body. Groups of metameres seem here to act as a unit. This case is certainly paralleled by the colour-bands on the arms of the brittle-stars. The serial repetition of the appendages of the Crustacea furnish examples of somewhat similar regional variation. It is not improbable that if we find an explanation for one set of phenomena we will be able to explain them all. XII. Regeneration 1n Eartoworms. In the winter of 1887-8 I made a small number of experi- ments to determine the extent of regeneration in the earth- worm. Again, in the spring of 1892, another series of experiments were started, but an accident spoiled the results. In the past winter of 1893-41 made a more elaborate and sys- tematic attempt to work out the same problem. I am much 4.4.6 T, H. MORGAN. indebted to Miss Elizabeth Nichols, Fellow in Biology, Bryn Mawr College, who began this study of regeneration with me. Many of the early experiments were largely carried out by her ; and later, when the work devolved on me, I profited much by the results of the previous work. There were several main problems that I wished to work out. First, the extent to which the earthworm could regene- rate; secondly, the number of new segments that would reap- pear in the anterior end after the removal of a definite number ; thirdly, the presence or absence of abnormalities in the re- generated anterior segments. Certain rough results were at first obtained, which showed that when many segments were cut off only a few segments replaced them. That is to say, there was no apparent connec- tion between the number of segments that were cut off and the number that regenerated. Four and rarely five new segments came back. I then set to work to determine what result would follow when only a few segments were cut off, for obviously if four came back when one, two, or three were cut off, the result would appear as though the reproductive organs had all shifted posteriorly. The tables below that are first given show the results of these latter experiments. One of the most conspicuous results was the great decrease in size that the worms suffer during the period of regeneration. When only a few segments were cut off the regeneration was soon accomplished, and no great decrease in the size of the body of the worm was obvious; but where many segments were cut off, and regeneration only took place after several months, or not at all, the body dwindled until it got to be less than a half of its original size. I have not made a histological study of these worms to determine what organs have suffered most during the period. The worms used were L.(orAllolobophora) fetidus, which live in manure heaps. They were kept in ordinary flower-pots filled with the manure in which the worms were found living. The pots stood in about an inch of water, and each was covered A STUDY OF METAMERISM. 44,7 with a glass plate. The pots stood in a large glass case in one of the rooms of the laboratory. The temperature of the room varied from about 70° F. during the daytime to about 50° F. at night. Under the same conditions the normal worms kept perfectly healthy. TaBLe I. Two segments cut off, 3/16, 94. Killed, 4/28, ’94. 2 segments regenerated : ' . vas def. 15 2 + cp : : : eee 2 22 2? ° ° ° 5) 15 2 a 5 : : : a ko Two segments cut off, 3/4, 94. Killed, 4/28, ’94. 2. segments regenerated ‘ : . vas def. 15_ Two segments cut off, 3/16, ’94. Killed, 5/5, 794. 2 segments regenerated. Sem. recept. 9, 10, 11 (normal) Tase II. Three segments cut off, 3/4, 94. Killed, 4/28, 794. 3 segments regenerated : ° . vas def. 15 3 3 33 (and piece of 4) 3 os » : : “ ay LB 3 x 3 : ‘ ‘ fee lb eee a : : ae ar » (and small piece of 4) 2 5 - (with indications of athird) ,, 14 2 23 29 33 be) Three segments cut off, 3/16, 794. Killed, 4/28, 794. 3 segments regenerated : ; . vas def. 15 es om 3s A ‘ : : aa Lo 2 a .s (sem. recept. 8, 9, 10) 3 36 ‘5 (and little piece of 4) a 25 Three segments cut off, 3/16, ’94. Killed, 5/5, 94. 2 segments regenerated (sem. recept. 8, 9, 10) 2 a % Os Pr ) vas def. 14. 448 T. H. MORGAN. Taste ILI. Four segments cut off, 3/16, 94. Killed, 5/5, °94. 4 segments regenerated : ; . vas def. 15 3 ” 2 . . . ” 14 3 4 5 5 . : ceeds 3 ” ” Ge . . 3 14 Be 5 %9 : ; ; eg 3° ” ” . : : $5 14 3 ” FP) a 9 TA 3 ” ” . . . ” 14 2 ” ” (+ 1/3 of 3) ZA 5 i137 3 ” ” (sem. recept. 8, 9, 10) ye ee = : A 3 sy ale Four segments cut off, 3/4, 94. Killed, 4/28, 94. 3 segments regenerated ‘ . vas def. 14 4, “ = (+ 4 of 5th) (8rd segment imperfect) so be 4, mt a (a small piece of 4 had been left) . ae ls TasLe IV. Five segments cut off, 3/16, 794. Killed, 5/5, 794. 4 segments regenerated 2 : . vas def. 14 3 ae os (+ 4 of 4th) ‘ 3 ke Five segments cut off, 3/16, 94. Killed, 4/28, 794. 4 segments regenerated. 3 33 ” 4 33 ” Five segments cut off, 3/4, 94. Killed, 4/28, ’94. 4 segments regenerated (+ 4 of 5th) . vas def. 15 eae a5 (+ 4 of 3rd) : a as A STUDY OF METAMERISM. 449 TABLE V. Attempts made to cut off 1 and 2 segments, 3/4 and 3/16, ’94. Killed, 5/2, 794. 1 segment regenerated - ; . vas def. 14 1 ”» 2” C . ey) 15 1 Pe - (indications of a divi- sion into 2) 2 ee 2 segments ms ope eel iy Conclusions from Tables I—V.—In all cases recorded (6 cases) where two segments were cut off two regenerated. When three segments were cut off (14 cases) three grew back in nine worms and two grew back in five worms. In those worms (14) where four segments were cut off, four segments regenerated in five worms, and three segments grew back in eight worms, and two segments and a third! in one worm. ; When five segments were cut off in no case did five grow back. In some of these worms a little more than five must have been cut off, so that a small piece of the sixth was taken off (see Nos. 2 and 6 and 7 of Table IV) and regenerated. In four worms four segments grew back, in two worms three segments grew back, and in one worm only two segments grew back. Taking these four tables together, we see that up to four segments lost, and including this, there is a tendency for the worm to reproduce the number lost. This was actually done in all the cases where two segments were lost, in nine cases out of fourteen where three were lost, and in five cases out of fourteen where four were lost. More than four segments the worm does not seem to be able to regenerate, as a rule. The data given in Table V are not as accurate as in the pre- ceding tables, because there was some confusion in the numbers of the pots containing these worms. So far as the figures go, they show that in three cases two segments must have been cut off, that one individual then regenerated two, and two 1 The amputation was oblique, and cut off a part of the fifth segment. 450 T. H. MORGAN. individuals regenerated only one segment each. In one case one segment must have been cut off and one regenerated. The results of these tables have a direct bearing on the con- ditions found in adult worms, and recorded in a previous section. Those worms in which both openings of the vasa deferentia were found on a segment anterior to the fifteenth may have been, in most cases, the result of the loss of ante- rior segments, and a subsequent incomplete regeneration. We cannot affirm that all cases were the result of the process, because, in the light of other facts, it is not improbable that such variations may have come from the embryo. These records also show us that regeneration of the anterior end will not account for any of those cases where the vasa deferentia open on a metamere posterior to the fifteenth, for in no cases were more segments generated than amputated. What the result would have been in the cases recorded below where a worm regenerated many segments, after amputation far posterior to the fifteenth segment, I cannot tell. The worms were not kept for a long enough time to determine whether or not reproductive organs would ever have appeared. In the preceding and following tables it will be noticed that in many cases the position of the openings of the vasa deferentia is recorded, and when these were not found the segments con- taining the seminal receptacles (9—10—11 normally) are re- corded. These landmarks were located after regeneration had taken place. This served as a check for those cases where the number of segments amputated had been previously recorded, and in the other cases gave fairly accurate evidence as to the number of segments that had been cut off. In the next two tables,—the results of the first series of ex- periments,—a large number of recorded segments were cut off to find the limit of the power to regenerate anterior segments. A STUDY OF METAMERISM. 451 TasLE VI. Segments cut off, Oct. 12, 93. No. cut off. No.of worms. Noy. 14— Jan. 1— (1) 4 segments regenerating. (1) 4 32 » S 10 (4) (1) 4 Bb + ame as Nov. 14 (1) Imperfect. Jan. 80— May — (1) aa (1) 6 segs. regen. (1) ni38 ai 12 (5) ¥ GQ) see (1) 5 segs. regen. (1) ves (1) 335,» (1)? 5 segments regenerating. Jan. 30— Apr. 14— May 18— 14 (5) (1) 5 seg. regenerating. bea aie (1) Imperfect. (1) (1) (1) Dead. (1) Jan. 30— Apr. 14— May 18— 16 (4) (2) Alive. (1) Alive. (0) Alive. Jan. 30— April 14— 19 (2) (2) Both regenerating. (1) Had regenerated 4 or 5 segments, (1)? Jan, 30— 19—24. (5) (1) Regenerating imperfectly. Jan. 30— (0) 24. (5) (3) Not regenerating. Dec. 2. Jan. 3— Jan. 30— Apr. 13— 26 (5) (3) (2) (2) (0) 27 (3) (2) (2) (1) (0) 452 T. H. MORGAN. TaBLE VII. Segments cut off, Jan. 12, 94. No. cut off. No.of worms. Apr. 14— May 18— 19 (1) (0) 21 (1) (0) 22 (3) (2). (1) Regenerated (2) Very imperfectly imperfectly. regenerated. 23 (1) (1) » (0) 24 (1) (0) 25 (1) (1) (1) 26 (1) (1) (0) 27 (1) (0) The results from Tables VI and VII show that one cannot say definitely, ‘here the power of regeneration ends.” The figures show that some worms regenerate where others fail to do so. It may be that the possibilities are different for different worms, or that at the time of operation certain worms were in better condition than others, or the external conditions (bac- teria, &c.) may have been different in different cases. The tables show that posterior to the twelfth segment the power of regeneration rapidly decreases. Worms that have lost more segments than this number may live for some time, and heal up the wound, or even regenerate imperfectly. But sooner or later the majority of these die. Occasionally re- markable exceptions are found. In Table VI the fifth record shows that a worm that had lost nineteen segments regenerated four or five new ones, and in the next tables more remarkable cases still will be recorded. It is a tedious operation cutting off a definite number of segments from a living worm. In the three following tables the number of segments cut off was not counted at the time, but could be calculated with approximate certainty after re- generation by the position of the vasa deferentia or segments containing the seminal receptacles, or, when the amputation was behind these, something like an approximation could be obtained by utilising the anterior end of the clitellum, or even the posterior end of the body. A STUDY OF METAMERISM. 458 Taste VIII. Anterior segments cut off, 1/22, 94. Killed, 4/18, 794. No. of segments regenerated. Vasa def. Calculated No. cut off. 3 13 bee 3 (4) 13 B (3) 3 (4) 14 4 (4) 3 12 6 3 12 6 3 (irregular) 14 4 3 14 4 4 12 te 4 12 r 4 (irregularly united to body) 138 6 4 12 7 4 (3) 13 6 (3) 4 (2 + 3) 14 eee 5 15 5 Anterior segments cut off, 1/22, 94. Killed, 5/5, ’94. No. of segments regenerated. Vasa def. | Calculated No. cut off. 3 13 5 - 3 +4+4+9) 15 3 (+) 3 (with compound 1—4) 14(orl4—15) 4 3 (3 + ¢ irregular) 13 5 (+) - 3 (irregular) 12 6 v 4 12 re 4 12 a 4 14 55 4 12 ty 4 (2) 14 5 (4) 4 (2) 12 7 (3) 5 (4) 14 6 (2) ~ 454, T. H. MORGAN. Tanne IX. Anterior segments cut off, 3/4, 794. Killed, 4/28, 794. No. of segments regenerated. Vasa def. Calculated No. cut off. i 14 2 1 14 ae 1,(4) (vas def. double +4 — 14 lor 2 (?) 3 (3) 15 3 (2) 3 3 (3) 4 4 4 4 vas def. 13 6 4 14 5 The Tables VIII and IX add a large number of data to those of the preceding tables. It will be noticed that there is one definite case of five new segments coming in for five cut off. There are also a number of irregular methods of union of new and old parts, and several cases of imperfectly formed new rings. The next table covers much the same ground as the preceding, but is interesting because the worms had been kept for a much longer time than those in the preceding table, and because the last record shows a case where from thirty to forty segments were in all probability cut off, and yet three and a half segments had regenerated. A STUDY OF METAMERISM. 455 TaBLEe X, Anterior segments cut off, 10/18, 793. Killed, 2/11, 794. No. of segments regenerated. Sem. recept. Vas. def. Calculated (Normal, 9 —10—11) No. cut off. 2 (2 + ¢ irregular) 5—6—7Z 11 6 (+) 3 7—8—9 13 5 3 4—5—6 -— 8 3 (+) (worm very small) -- 10 8 (2) 4, 5—6—7 — § 4 S$—9I—10 — 5 eI — 12 7 5. (irregularly joined) —6—7 (?) os 9 (?) 5 ~~: 11 9 5 (small piece) 5 (very small piece) 4. (F) = 13 6 G) 6 (+ 3) —7—8 = ‘10 (3) 3 (4) small worm, with only 65 old segments and old anus. 30—40 In addition two other worms, where union of new head and body was too irregular to count even approximately the segments. TaBLeE XI. Anterior segments cut off, 11/15, 93. Killed, 4/14, 794. No. of segments regenerated. Sem. recept. Calculated No. cut off. 3 9—10—11 3, 5 6—7--8 8 4 — § Not regenerated 10 in front of clitellum 15 3 ere 34, 19 a (mouth present) 9 a ne 16 Scarcely regenerated (small piece) 3 or 4 very irregular 9 m 3 16 3 imperfect segments 8 re ss Ly 4 (1st partially divided) vi = _ 8 4 A * above) 15 AS - 10 5% (1st imperfect) * 15 _ ss 10 15 or more regenerating 63 segments in front of anus 385—40 This table is interesting because it shows a large number of WoL. 37, PART 4.—NEW SER, HH 456 T. H. MORGAN. worms regenerating imperfectly a few segments in front of the clitellum. Moreover the last case is particularly instruc- tive, because here again 30—40 segments were probably cut off. The number of worms that were operated on at the be- ginning of this experiment was unfortunately not recorded ; hence we do not know what the per cent. of mortality has been. TaBLE XII. Anterior segments cut off, 11/18, ’93. Killed, 1/21, 94. No. of segments regenerated. Sem. recept. Vas. def. Calc. No. cut off. 2 —3—4 8 9 3 — 10 8 3 (united irregularly) 8—9—10 14 4 3(+ $44) Sie - oe) 4 (+ 4) 6—7 —8 12 7 (+) 4 (+ 4+ 4d irregularly joined) — 11 8 (+) 2 3—4—5 9 8 3 —4—5 — 9 3 (2) 45—6—7 = ‘Bice a 7—8—9. 13 6° 4 (3) —6—7 — 8 (3) 3 13 in front of clitellum 12 3 (3) ll » ” 14 3 12 %3 23 13 3 (or four imperfect) just cE ” 23 a 17 23 2» 8 4 14 33 33 ll 5 16 33 > 9 5 14 ” ” ll 2 (or three very imperfect) 4 5 + 21 2 (or three very imperfect) 14 y » il Very imperfectly healed 7 ” ” 18 Imperfectly healed 12 es » 13 Healed—evidence of few rings 9 s a 16 Beginning to regenerate 10 99 » _ 1 oy) 0 “15 ” ae ~ 15 Healed, not regenerated 15 ie a 10 Healed—regenerating (?) 5 ” » 20 8 je more segments. Rings faint. Small piece } 56 old segments in seals of anus. — A STUDY OF METAMERISM. 457 In several instances in the preceding tables records are found of worms that, in addition to the formation of new segments, completed parts of segments accidentally cut off in the operation. The following records are of worms in which purposely very oblique amputation had been made. Taste XIII. Anterior segments obliquely amputated, 3/4, 794. Killed, 4/28, 794. §. segments completed (9—10—11)sem.recept. vas.def.— 9 aS », (dorsal) — esi Be ll ” ” (lateral) = J 15 2 % 9—10—11) ,, Ee 4 = », (lat. vent.) +] newseg. (9—10—1]) 3 — 4 ” ” » +3newsegments — eA 6 me » (lateral) +3 53 —_ — 7Q@ 5 », (lateral) = i ae 7 s », (lateral) +2 ‘5, (8—9—10) 8 at The last table shows very clearly that the power to com- plete segments that have in part been removed is much greater in the earthworm than the power to regenerate whole segments. In the table we find as many as twelve segments completed, and in a comparatively short time. Such a number of segments is never or rarely regenerated when the anterior segments are cut squarely off. / The latter records of this same table show that when ante- rior segments are entirely cut off, in addition to those cut obliquely, that we have both a regeneration of those lost (within a limit) and a rebuilding of those injured. Moreover, although the results are too few to speak with entire confidence, it would seem that the presence of segments completing them- selves does not interfere with the formation of the full comple- ment of new regenerated segments. These facts, it seems to me, throw an interesting light on the problem of regeneration. They are too few to warrant at present any speculation. It is my intention to make a fuller and more accurate study of these phenomena. The obliquity of the cut ‘was lateral in most cases, in a few 458 T. H. MORGAN. the slice was taken off the dorsal surface, and in others off the ventral. In a few cases no record was made. Whether or not the injured rings complete themselves more readily at the side than dorsally or ventrally cannot be determined from the table, for the position of the parts cut off in the first instance was not recorded. In connection with the preceding experiments another was carried on. ‘The anterior ends that were cut off were in many cases kept alive to see what power of regeneration remained in them. It became evident very soon that the length of the life of the pieces was in a general way proportionate to their linear length. Those with a few segments died in the course of a week ; those with more lived longer, &c. No cases of survival of as few segments as fifteen were ever found. The two fol- lowing tables show to what extent longer pieces remained alive, but with one exception (twenty-four anterior segments) they all died after a time. It is surprising to find this to be the case, for such pieces contain the mouth (so that the piece may feed) and all of the important (?) organs of the body. TaBLe XIV. Record of anterior segments. Date—Oct. 12. Nov. 14. Dec. 2. Jan. 4. ‘No. of anterior No. of segments. worms. 12 (5) (1) (0) a (5) (4) (2) (0) 16 (4) (0) 19 (2) (2) (1) (0) 19—24 (5) ~- (2) (0) 24 (5) (5) (4) (3) May 5 (0) 26 (5) (5) (3) (1) Apr. 26(1) Jan. 12. Apr. 14. May 5. May 18. 19 (1) (0) 22 (3) (3) (3) (?) (0) 23 (1) (1) (0) 24 (1) Regenerated a half-inch. 26 (1) 0) ( 27 (1) (1) (0) A STUDY OF METAMERISM. 459 The worms that had the anterior segments amputated, the history of which is recorded in Tables I—XIII, were examined from time to time during the period of regeneration. One striking result was often apparent. When the amputation of the anterior segments had taken place obliquely, the new seg- ments grew out approximately at right angles to the cut surface. It was not uncommon to find a new regenerating anterior end with its axis inclined at as much as (or even more than) 45° to the long axis of the body. I have no records as to whether the same thing happens when the oblique surface is above or below. Those referred to were from lateral oblique sections. This result is comparable to that of Barfurth on the regene- ration of the tadpole’s tail. Here the new part appeared at right angles to the cut surface, and subsequently swung round into line. To get the same result from the tail of the tadpole and the head of an earthworm suggests that there is some fundamental law of growth underlying both phenomena that should be more extensively investigated. It was my intention to examine the power of regeneration in young worms for comparison with the adult, but only a few experiments have been made, and there has not been sufficient time to allow the completion of this side of the work. Miss Adelene M. Fielde (14) has published a few fragmen- tary notes on the power of regeneration in L. terrestris. Pieces containing twenty to thirty segments from the posterior end of the worm lived forty days, but did not regenerate at either end. In these pieces new half-segments had been in- serted, the authoress affirms, hecause she could not find any such modifications (compound metameres) in other worms! In nine worms five anterior segments were amputated. These “wholly regenerated.” In ten worms five anterior and twenty to thirty posterior. These were found regenerating. 4.60 T. H. MORGAN. XIII. Generat Concivusions. The solution of the problem of metamerism has often been attempted by morphologists with varying success. It has become more and more evident that the problem is a difficult one, and I think the results have shown that the final solution can only come with a better knowledge of the fundamental relations of the parts of the body to one another, with an acknowledgment of the imperfection of the phylogenetic method, and with a better insight into ontogenetic laws. Darwin said over thirty years ago, in the ‘ Origin of Species,’ “We need not here consider how the bodies of some animals first became divided into a series of segments, or how they became divided into right and left sides with corresponding organs, for such questions are almost beyond investigation.” The attempt of morphologists to solve even the simpler of these phenomena, viz. metameric repetition, shows how true Darwin’s words remain even to-day. It may, therefore, seem doubtful whether anything will be gained by a new analysis of the problem of metamerism, or by a critical consideration of the numerous theories already advanced. It would certainly be unwise to add any new speculation to that already afloat. In the following pages no new theory is offered, and I have attempted no more than a consideration of those methods which have proved sterile or erroneous as contrasted with the methods that have been fruitful and suggestive. During the last fifteen years the methods of the phylo- genists have been applied to the solution of metameric repe- tition. Comparative anatomy has been the point of departure ; but speculation has leaped far beyond its legitimate boundaries. The results have shown that the method of phylogenetic inter- pretation is subjective rather than objective, and the conclu- sions have given, therefore, at most a probable course of evolution, and often only a conceivable process of transition. It is only fair to say that in many cases the speculation has been advanced tentatively, as suggestion rather than conclu- A STUDY OF METAMERISM. 461 sion, and the admirable work on which the speculation has been based has been in no degree vitiated by the attempts of the authors to push their conclusions beyond the limits deducible from their immediate results. The Ceelenterates have formed the basis of Sedgwick’s theory (34) of the origin of metamerism. The radial actinian has been worked over into a bilateral metameric form, and all the details of structure of the higher forms (celom, nephridia, gill-slits, trachea, &c.) have been evolved from the hypothetical actinian ancestor. Wilson (37) has supported the same view, “but only as a suggestion for further investigation of the facts.” The Turbellaria have been the starting-point of Lang (25) and Meyer (29); but each author has described an entirely different course of transition from the flat-worm to the Annelid. The Nemertian claims have been pressed forward by Hubrecht (22) and Balfour (2), and hinted at by others. The Echinoderms have proved refractory, yet Wagner! has made out a possible phylogeny from these to metameric forms, and Haeckel (15) reversing the process, made a star-fish out of five fused Annelids. The Enteropneusta, declared unsegmented by Bateson (3) and segmented (32) by the present writer, have been believed, nevertheless, by both authors to throw light upon the problem of metamerism. With this divergence of opinion it is not surprising to find as great a divergence of method. In one case a radial form has given the starting-point, in the others a bilateral form. But the ways in which bilateral forms have been transformed into the metameric form have been very different. The budding theory has played perhaps the most conspicuous part. Nearly all of the older writers looked upon segmented forms as animal colonies—Quatrefages, Cuvier, Owen, Duges, Geoffroy- St. Hilaire, Lacaze-Duthier, Herbert Spencer, and Perrier. The same idea is often found in later speculation as well. In 1 Quoted on the authority of Meyer (29). I have not been able to find the original, 4.62 T. H. MORGAN. the budding theory a short bilateral form is supposed to have elongated by a repetition of itself to form a long chain of united individuals. Other authors—Hubrecht, Lang, Meyer —have supposed metameric repetition to have appeared in a long animal which split up secondarily. Hubrecht (22, 23) has supposed that a long animal, such as a Nemertian, is con- tinually in danger of injury from without, and the animal has met this danger by acquiring a remarkable power of regenera- tion! hose animals that had the main organs of the body repeated were better able to regenerate any pieces that were broken off; for such pieces would contain, in all probability, all of the essential organs of the body. Lang’s (23) description of the origin of metameric repetition from the flat-worm, Gunda segmen tata, is too well known to need rehearsal. In this elongated form the repetition of the digestive diver- ticula have been the centres around which the metameres have been built up. Meyer (29) has contended that the repetition of the meta- meres is an expression of the alternate bendings of the sides of the body of a free-swimming elongated worm. A pair of elon- gated gonad-pouches have been broken up into a series of meta- meric compartments, owing to the swimming movements of the body, and around these as centres the metameres have evolved. The attempts to find the solution of metamerism within the metameric groups have been equally unsuccessful. The simpler Annelids, such as Polygordius and Protodrilus, have been interpreted as.archaic forms by Hatschek. This explana- tion has been rejected by Kleinenberg, Eisig, and Meyer. The lower Vertebrates have been equally difficult to interpret. Lankester (28) found that the tail of Appendicularia showed muscle-plates with corresponding ganglia in the nerve-cord. The Ascidian larva has no similar structures. Brooks (8) has argued that Appendicularia is a very old archaic form; while Willey (89) has interpreted the tadpole larva of the Ascidians as a secondary larval form derived from a fixed ancestor. By inference, therefore, Appendicularia is a sexually mature larva, A STUDY OF METAMERISM. 4.63 and its segmented tail either an antefact or a secondary acquirement. Bateson (8) has attempted to show that Balanoglossus is related to the Chordata, and is unsegmented. I (82) have defended the same supposed relationship, and described Balanoglossus as segmented. Spengel (86) believes Balano- glossus not to be related at all to the Chordata. Amphioxus is a typically segmented form, but has given no clue as to the origin of its metamerism. By many morphologists (Semper, Dohrn, Van Wijhe, Minot, &e.) the group Chordata is supposed to be derived directly from segmented Annelids. But the Annelid-vertebral con- nection has as many opponents as defenders, and is one of the most illustrious results of the phylogenetic method. It must be admitted, of course, that morphologists are dealing with complex and difficult problems in attempting to unravel the past connections between the larger phyla of the animal kingdom, and that their speculations have been in many cases ingenious working hypotheses; but the results show, I think, that the method is easily carried too far, and that, after many trials, it has not led us to any definite con- clusion. Ontogeny has been also fruitful in speculation. The cry has been that Ontogeny tended to repeat Phylogeny, and larval forms without end have been set up as archaic remains. The rise, culmination, and decline (?) of the Gastreea theory illustrates in a concrete case the history of the ontogenetic method ; and the Nauplius theory teaches a lesson of caution that ought not to be forgotten. The ccelom theory, built up on the splendid results of Agassiz, Metschnikoff, Kowalewski, Lankester, Hatschek, and Hertwig, while a most suggestive working hypothesis, has led to no settled conclusion; for the well-ascertained fact that in many groups (Sagitta, Amphioxus, Brachiopods, Echinoderms, &c.) gut-pouches give rise to the celom has not led us to any decision as to whether we have here an ontogenetic performance or a phylogenetic repe- 4.64, T. H. MORGAN. tition. In the Annelids, for instance, where there is a typical ceelom developed, there are no traces of gut-pouches, and this has to be interpreted as a secondary loss. Hatschek has suggestively remarked (15), ‘‘ The two mesodermal teloblasts may correspond to the celom-sacs from which they were derived by a reduction in the number of their cells. In fact, we find this method of formation only in those cases where the number of cells of the embryo is very small.” Whatever be the conclusion reached as to the formation of the ccelom, we shall still be far from a decision as to how and when the ccelom repeated itself in the metameric forms. The trochosphere of the Annelids, Mollusca, and Mollus- coidea (?) has been utilised by Hatschek to support the whole family tree of the higher Metazoa (the Vertebrates perhaps excepted). Kleinenberg has interpreted the trochosphere as a recapitula- tion of ahydro-medusa. E. B. Wilson and others have rejected the trochosphere ancestry, and believe the larva to be cceno- genic. Whitman wrote in 1887, “In spite of volumes devoted to the discussion of the subject, the larva of Poly- gordius still remains a morphological puzzle.” - The trochosphere theory got a strong support from Semper’s discovery of Trochosphera equatorialis.’ Even Kor- schelt and Heider, who, as a rule, are most circumspect in accepting embryological speculation, wrote in 1890, ‘‘ Hochst wahrscheinlich liegt in der Trochophora der Anneliden die ontogenetische Recapitulation einer Stammform vor, welche den Anneliden Mollusken und Molluscoiden gemeinsam war und von der aus sich diese Thierstiimme als selbststiindige Gruppen abzweigten.”” Recognising the relationship existing between the trocho- sphere larvee of Mollusca and Annelida, embryologists have not been satisfied to postulate only the archaic nature of the larva, ‘ Semper’s Rotifer was known long before Hatschek’s theory, and sug- gested to me the name “trochosphere” for the larval form (‘ Devel. of Lynneus,’ 1875), which some years latcr Hatschek adopted from me with the change of the word to “ trochophore.’—KH. Ray LanKusTER, A STUDY OF METAMERISM. 465 but have gone further, and postulated it as the ancestral adult. They have been led to believe that such a minute few-celled larva has evolved into the complicated segmented Annelid, and into the equally complicated but unsegmented Mollusc. The resemblances between the adult Mollusc and Annelid they have been content to call “ parallel developments.” For all the evidence we have at present we might satisfy the facts just as well by assuming that a large many-celled unseg- mented form stood as the bottom of these two groups, having a trochosphere as its larval form. The Rotifers and related forms would then be interpreted as arrested forms. The one conclusion would be, I think, as justifiable and as easily maintained as the other, and both impossible to demonstrate. In the face of so much conflicting embryological indecision, it seems to me we have in reality arrived with certainty no nearer to the solution of metamerism. There have been only a few attempts to explain the meta- meric repetition as the result of mechanical action. Hu- brecht’s (22) explanation for the Nemertian is scarcely a me- chanical explanation, since it presupposes a repetition of parts and an ability to regenerate in the worm itself. Kennel’s (24) ingenious attempt to interpret division of animals as the result of external stimuli is scarcely a mechanical explanation. His (19) has attempted to explain the metamerism of the Vertebrate as an embryological phenomenon—as the result of series of breaks occurring in the two lateral mesodermic sheets of the embryo. Meyer’s (29) view, referred to above, is distinctly a mechanical hypothesis. The repetition that he assumes to have come into a long Nemerto-Turbellarian an- cestor was the result of the movements of the body in swim- ming. Many objections are easily raised against Meyer’s romantic speculation, As Hatschek has pointed out, those Annelids that are adapted for swimming are heteronomously segmented, while Meyer’s hypothesis seems to demand first a homonomously segmented form as the result of its own activity. In the second place Meyer’s sketch starts off on Lamarckian 4.66 T. H. MORGAN. principles. Although many naturalists still admit the prin- ciple of use and disuse as a factor of organic evolution, perhaps an equally large number reject the explanation. Hence, until we get definite proof of the truth or falsity of any such theory, it ought not to be used as the starting-point on which to build up other or new theories. Caldwell (11) has offered a brief and interesting attempt to explain metameric repetition. So brief, indeed, is the presenta- tion, and so obscurely worded, that I am not certain that I have entirely understood the meaning of the author. It isa mechanical theory par excellence. The theory assumes that as early as the blastula! stage the endoderm, as well as the mesoderm, is represented in cells or groups of cells at the surface of the sphere. Now the endoderm may before inva- gination get separated into two parts, owing to an early elongation of the blastula, so that when gastrulation sets in it may take place at two regions of the surface. One of the regions may contain much more endoderm than another, giving an oral or an anal gastrulation asa result. Similarly the mesodermal “ Anlage” may be pulled apart and turn in with the endoderm at one region or the other, or may even have been left along the line where the two endodermal masses separated. In many cases the whole of the mesoderm may be turned into the gastrula cavity with the endoderm, and subsequently set itself free from the endoderm by one, two, or many gut- pouches. When many gut-pouches arise, they mark the be- ginning of metameric repetition. The reason for many pouches appearing in some forms is to be explained as due to an early elongation of the invaginated endoderm forming the archen- teron, so that the mesodermal “ Anlagen” get pulled out and broken apart. If, as Caldwell supposed, there is pre-formation for the mesoderm, there must be for all the other organs of the body, and this the author admitted. 1 The author says planula, which makes his explanation obscure ; unless he means technically a blastula. A STUDY OF METAMERISM. 467 It is difficult to see how by simple elongation of a larva to meet a supposed larval advantage, in the first place the organs in the ectoderm should get separated into exactly the same number as the number of gut-pouches, and in the second place that all of these should correspond so as to give organs repeated symmetrically in both ectoderm and mesoderm. I cannot, I admit, resist the conviction that metameric repeti- tion is far too difficult and fundamental a problem to be ex- plained as the result of mechanically pulling out the supposed Anlagen of all the organs of the body. The phenomena of metameric repetition and apical growth are closely associated together. Whether the latter stands in any causal relation to the first cannot be definitely asserted, or if it could we should have no means of determining whether the relation is an ontogenetic or a phylogenetic one; whether the apical elongation was established in the embryo or in the adult (if, indeed, we can draw any line of any value between the embryonic and adult appearance of any organs). Nevertheless it is important to emphasise the fact of the connection, for we find both in Vertebrates and Annelids the two phenomena closely bound up together. Further, we have the interesting case of the star-fishes and brittle-stars, where at five radial points there is apical growth, and the five arms are segmented. In the higher plants also there is a repetition of similar parts (phytomeres) and apical growth. Whether the repetition of the calcareous skeleton and tube- feet of the arm of a star-fish is a repetition comparable to the repetition in the Vertebrate and Annelid will depend largely upon definition of terms. As to the fact of a symmetrical repetition and the presence of apical growth there can be no question, and that is the main point. There are no grounds for assuming that the repetition of the parts of the star-fish arm was ever connected with any attempt of the animal, in the past, to reproduce itself at five equidistant points, nor would such a suggestion be believed by anybody for a moment. The method of growth in these arms, where there is a terminal piece carrying the eyes and a subterminal growing region, 468 T. H. MORGAN. is so similar to the method of elongation of the Annelid body that even the most casual observer must be impressed by the comparison. The greatest drawback to any attempt to refer the two cases to a common method of growth (not phylo- genetic, of course) would probably be met hy the statement that the repetition of a metamere is something entirely different from the repetition of the vertebral ridge in the star- fish’s arm. If we can succeed in breaking down this conven- tional and artificial definition of the value of metameric repetition as compared with other repetitions of the body we shall have made a step forward Iam confident. This question will be more fully dealt with in the next section. I have asked the opinion of eminent botanists on several occasions as to the meaning of the repetition of the parts of the higher plants, particularly the Phanerogams. So far as I can learn, the botanical phylogenists have not had a much better time of it than the zoological phylogenists. The former have had the immense advantage of much paleontological ma- terial, but, as I understand, even with this the great gaps come just where the evidence is most wanted. I have not, however, found any botanist who believed that in the past a single stem and leaf or two leaves (phytomeres) represented the ancestral plant which grew long by repeating itself, as the embryo plant does at present. Haeckel, in his ‘ Generelle Morphologie, 1866, used the term “ promorphology ” to include the fundamental relations of the parts of an animal to one another, in the same sense that the relations of the axes of a crystal are the expression of its form. The aim of promorphology, Haeckel said, is to deter- mine the ideal fundamental form by a process of abstraction, and to discover the natural laws according to which organic matter develops its outer form. The relation of the parts, i.e. the form, results with absolute necessity from the architectural union of the constituent parts, in the same sense that an inor- ganic crystalline form results from the union of crystalline material, and from its relation to its environment. Whether morphology will be ultimately driven to an inter- A STUDY OF METAMERISM. 469 pretation of the symmetry of organic form as the expression of physical laws of protoplasm, rather than due to slow adapta- tion of an amorphous or irregular substance to its surround- ings, is too large a question to attempt to discuss. Whatever the explanation may be, there are certain well-established facts in this connection that have, it seems to me, an important bearing on metameric repetition. The relation existing between bilateral and radial symmetry is one of the most suggestive fields of promorphology. We get from this source more suggestion as to a possible interpre- tation of the facts of metameric repetition than from any other source. It would lead too far to attempt anything like a full discussion, but I may cite two cases that will serve as simple illustrations of a large field of inquiry. The radial symmetry of a sea-urchin is a very perfect type of five-rayed structure. The test, made up of a mosaic work of calcareous plates, is a marvellous piece of detailed fitting. Yet there are several cases on record of individuals that have a sixth ray (antimere) introduced. Each of the six antimeres may be a perfect copy of the others, as well as of the normal. The same condition is not uncommon in the star-fish and other Echinoderms, but owing to the lack of a mosaic calcareous skeleton the result is not so impressive. Again, in other individuals one of the five antimeres may be entirely or in part omitted, and yet the surface shows a perfect symmetry. The point here to be emphasised is that a whole section of the body (antimere) may be introduced or omitted, involving the introduction or loss of all the organs belonging to such a division. More remarkable still are the triangular tapeworms de- scribed by Leuckart and others. A strongly marked bilateral animal repeats occasionally one of its halves, so that we may paradoxically speak of the worm as composed of three halves. Instead of a somewhat flattened bilateral animal, there results ! Bateson points out two cases which are to be distinguished. There may be a division into two of one antimere, or there may bea redistribution of the whole material into six parts. The text refers to the latter. 4.70 T. H. MORGAN. a radial (triradiate) form, having three sets of longitudinal organs where normally there are only two. Two other facts in connection with these variations ought to be emphasised. The scolices of the tapeworms arise on the large bladderworm by invagination of the surface wall. The relation of the inva- gination to the surface of the sphere is a radial rather than a bilateral relation, and it is interesting to find this occasionally expressed in the triangular scolices. In the second place, Leuckart records finding on the same bladder (Cysticercus) both radial and bilateral scolices. This seems to show that the radial type need not have come in from egg variation, but is the result of conditions acting at the time of the formation of the scolices. Now both of these cases, the sea-urchin and the triangular tapeworms, show that a complete section of the body may be repeated and intercalated symmetrically amongst the other parts. The new portion has appeared at once and fully equipped, duplicating the structures of the body that lie in a similar axial position. These and many other similar cases show us, I think, very positively that the variations appearing in a radial animal must have come simultaneously and all together into the anti- meres. Moreover I think no one will doubt that the relation existing between the repeated organs in a radiate animal is at bottom the same relation existing between the right and left sides of the body of a bilateral animal. Mivart (31) and Brooks (9) have emphasised the further fact that the relation between the right and left sides of the body is the same relation that exists between the serially repeated parts of a metameric animal. If this line of argument be admitted, it puts the problem of metamerism into a large category of well-established facts. That the final explanations of these facts is closely bound up with the solution of some of the most fundamental problems of biology is self-evident. To hope, therefore, to solve the problem of metamerism in A STUDY OF METAMERISM. A471 the simple ways already tried by phylogenists and embryo- logists is, it seems to me, a vain hope. But a more vigorous study of the facts of metameric repetition from the standpoint reached above might lead us in the right direction. A study of a larger number of facts than we possess at present will at least tell us whether or not metamerism is to be re- ferred to this category. If this should prove true, we shall know where to search for the key to the problem. Even if there is no immediate hope of reaching a definite conclusion, yet we shall have gained much if we can find in what direction the solution lies. BIBLIOGRAPHY. 1, AnpREws, E. A.—* Report upon the Annelida Polycheta of Beaufort, North Carolina,” ‘ Proceed. U. S. Nat. Museum,’ vol. xiv, 1891. 2. Batrour.—‘ Comparative Embryology,’ 1879. 8. Batrson.—“ The Ancestry of the Chordata,” ‘Quart. Journ. Micr. Sci.,’? xxvi. 4. Batrson.—‘ Materials for the Study of Variation,’ 1894. 5. Brpparp.—‘ Proceed. Zool. Soc.,’ 1886. 6. Bennam, W. B.—‘ Annals and Magazine of Nat. Hist.,’ ser. 6, vil, 1891. 7. Bere, R. S.—‘ Untersuchungen iiber den Bau und die Entw. d. Ge- schlechtsorgane der Regenwirmer,”’ ‘ Zeit. f. wiss. Zool.,’ xliv, 1886. 8. Brooxs.—“ Salpa in its Relation to the Evolution of Life,’ ‘Stud. Biol. Lab. Johns Hopkins Univ.,’ May, 1893. 9. Brooxs.—‘ Lucifer,” ‘Phil. Trans.,’ 1882 (reprinted); ‘ Morpho- logical Monographs,’ 1884. 10. Bucuanay, F.— Peculiarities in the Segmentation of certain Poly- chetes,” ‘Quart. Journ. Mier. Sci.,’ xxxiv, 1893. 11. CatpwreLi.— Blastopore, Mesoderm, and Metameric Segmentation,” ‘Quart. Journ. Micr. Sci.,’? xxv, 1885. 12. Ciavs.— Zur morphologischen und phylogenetischen Beurtheilung des Bandwurmskérpers,” ‘ Arb. aus d. Zool. Inst. Wien,’ viii, 1889. 18. Cori, C. J.—‘* Ueber Anomalien der Segmentirung bei Anneliden,” ‘ Zeit. f. wissen. Zoologie,’ liv, 3, 1892. 14. Frecpp, ADELENE M.—* Observations on Tenacity of Life and Regene- ration of Excised Parts in Lumbricus terrestris,” ‘ Proceed. Acad. Nat. Science,’ 1885. VOL. 37, PART 4,—NEW SER. II 4,72 T. H. MORGAN. 15. 16. Li: 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34, 35. 36. 37. Harcxet.—‘ Generelle Morphologie,’ 1886. Harscuex.— Studien tiber Entwicklungsgeschichte der Anneliden,” ‘Arb. a. d. Zool. Inst. Wien,’ Bd. i, 1878. Hatscnex.— Ueber Entwicklungsgeschichte von Echiurus,” ‘ Arb. a. d. Zool. Inst. Wien,’ iii, 1880. HatscueK.—‘ Lehrbuch d. Zoologie,’ 1888. His, W.—‘ Uisere Korperform,’ 1874. Haacxe.—* Zur Morphologie der Seeigelschale,” ‘ Zool. Anzeiger,’ viii, 1885. Hertwie, O. and R.—‘ Die Celomtheorie,’ 1881. Husrecut.—* On the Ancestral Form of the Chordata,” ‘ Quart. Journ. Micr. Sci.,’ xxiii, 1883. Husrecut.—“ The Relation of the Nemertea to the Vertebrata,” ‘Quart. Journ. Micr. Sci.,’ xxvii, 1887. KenneL, J. v.—‘ Ueber Theilung und Knospung d. Thiere,’ Dorpat, 1887. Lane, A.—‘ Der Bau von Gunda segmentata,” ‘Mitt. a. d. Zool. Station Neapel,’ iii, 1881. Lane.—‘ Ueber d. Hinfluss d. festsitzenden Lebensweise auf. die Thiere,’ Jena, 1888. LANKESTER.— On the use of the term ‘ Homology’ in Modern Zoology,” ‘Ann. and Mag. of Nat. Hist.,’ 1870. LankEsTER.—“The Vertebration of the Tail of Appendicularie,” ‘Quart. Journ. Micr. Sci.,’ xxii, 1882. Meyer, Ep.—“ Die Abstammung der Anneliden,” ‘ Biolog. Central- blatt,’ x, 1890. MIcHAELSEN.—“ Oligochaeten des Naturhistorischen Museums im Ham- burg,” iv, ‘ Jahrb. d. Hamburg wiss. Anstalt,’ viii, 1891. Mivart.—‘ On the Genesis of Species,’ 1871. Morean.—“ The Development of Balanoglossus,” ‘ Journ. Morph.,’ ix, 1894. Morean, T. H.—‘“ Spiral Modification of Metamerism, *Journal of Morphology,’ vii, 2, 1892. Sepewick.—“On the Origin of Metameric Segmentation,” ‘ Quart. Journ. Mier. Sci.,’ xxiv, 1884. Spencer, H.—‘ Principles of Biology,’ vol. ii, appendix B, 1867. SprencEL.— Die Enteropneusen,” ‘Fauna and Flora,’ Neapel, 1893. Wi.son.— The Mesenterial Filaments of the Alcyonaria,” ‘ Mittheil, a. d. Zool. Station Neapel,’ v, 1884, A STUDY OF METAMERISM. 473 88. Witson, E. B.—‘‘ Embryology of the Earthworm,” ‘Journ. Morph.,’ iii, 1889. 39. Wittey, A.— Studies on the Protochordata:” I, ‘ Quart. Journ. Mier. Sci.,’ xxxiv, 1893. 40. Woopwaxrp, M. F.—* Description of an Abnormal Earthworm possess- ing Seven Pairs of Ovaries,” ‘ Proceed. Zool. Soc.,’ 1892. DESCRIPTION OF PLATES 40—43, Illustrating Mr. T. H. Morgan’s paper, “‘ A Study of Metamerism.” PLATE 40. Type forms of modification of metameres of Annelids. See pages 395 to 403. Fie. I, a,B, c.—Compound metamere (split metamere), as seen from dorsal (A), ventral (B) side, aud reconstructed in c, as seen from above. Fic. II, a B c.—Modification of last. Fic. III, a 8 c.—Inserted half-metamere. Fic. IV, A B c.—Double compound metamere. Fie. V, a B c.—Failure of lines to meet dorsally. Fic. VI, a, B, c.—Spiral of two metameres. Fie. VII, a, B, c.—Spiral, with one more half-metamere on one side than on other. In all five half-metameres. Fie. VIII, a, B, c.—Spiral, with same number of half-metameres on each side. In all six half-metameres. Fie. IX, a, B, c.—Spiral of type VIL, but longer. Fic. X, A, B, c.—Spiral of type VIII, but longer. Fic. XI, a, B.—Spiral of type VIII, but still longer. Fic. XII, a, 8, c.—Spiral formed by combination of half-compound meta- meres. Fie. XIII.—Spiral, in part a double-spiral, formed by introduction of half-compound metameres. Fic. XIV.—Spiral, in part a double-spiral, formed by the introduction of a half-double-compound metamere. 474, T. H. MORGAN. PLATE 41. All figures are from the anterior end of L. (Allolobophora) fetidus. Fic. Fie. Fic. Fie. above. Fie. Fic. Fic. Fic. Fic. Fic. Fic. Fic. Fic. Fic. Fic. Fic. Fie. Fig. Fic. Fic. Fig. Fic. Fic. Fic. Fic. Fic. meres. Fie. Fia. Fic. Fic. Fie. See text pages 407 to 418. 1, 4 B.—Fourth metamere, incompletely developed on right side. 2, A B.—Second metamere, ditto. 3, A B.—Fifth ditto, left side. 4.—Failure of septal line between 9th and 10th metameres to meet 5.—Ditto between 5th and 6th ditto. 6.—Ditto 4th and 5th ditto. 7.—Ditto 3rd and 4th ditto. 8.—Ditto 13th and 14th ditto. 9.—Ditto 12th and 13th ditto. 10.—Ditto 11th and 12th ditto. 11, 4 B.—Ditto 1st and 2nd below. 12.—Ditto 13th and 14th below. 13.—Ditto 5th and 6th at side. 14.—Spiral of category vi11, involving 7th, 8th, and 9th metameres. 15.—Ditto, involving 12th, 13th, and 14th. 16.—Ditto, 11th, 12th, and 13th. 17.—Ditto, 12th, 13th, and 14th. 18.—Ditto category x, involving 6th, 7th, 8th, and 9th. 19.—Ditto, 11th, 12th, 13th, and 14th. 20, A, B, c.—Ditto category vil, involving 2nd, 3rd, and 4th. 91, a, B, c.—Ditto, 9th, 10th, and 11th. 22, A, B, c.—Ditto category 1x, involving 6th, 7th, 8th, and 9th ditto. 23, A B, C D.—T wo modifications—categories 11 and v. 24, 4 B c.—Short spiral; origin doubtful. 25, AB C.—Spiral virr and compound metamere 1 combined. 26.—Spiral of category vii, very long, involving 11th—24th meta- 27, A B c.—Ditto, long, involving 10th—16th. 28.—Ditto, 13th—24th. 29, a B c.—Spiral category vu, involving 14th, 15th, and 16th. 30, A B c.—Ditto category vitl, involving 15th, 16th, and 17th. 31, a B.—Ditto, 15th, 16th, 17th, and 18th. A STUDY OF METAMERISM. A75 Fic. 32, a B.—Ditto, 15th, 16th, and 17th. Fic. 33, A B.— With compound metamere 10—22. Fic. 34, a B.—Ditto 7—2. Also category v. Fig. 35, a B c.—Ditto 9—.8,. Also category II. Fic. 36, A B c.—Spiral vil, involving 11th, 12th, and 13th metameres. Fig. 37, A B c.—Compound metamere 12—32. Fic. 38, A B c.—Category 1. Fic. 39, A B c.—Compound metamere 15—35. Fic. 40, A B.—Spiral vir, involving 8th, 9th, and 10th. Also another modification. Fie. 41, A B c.—Four compound metameres I and spiral vit. Fic. 42, a B c.—Compound metamere 3—3, followed by complicated spiral. Fic. 43, a B c.—Two spirals vit, and compound metamere II. Fic. 44, 4 B.—Compound metamere 6—. Spiral viz. Compound meta- mere 13—13. Fie. 45, AB c.—Intercalated half-metamere 15. tu. Spiral involving 15th—19th metameres. Fic. 46, a B c.—A much modified anterior end. See construction c. PLATE 42. All figures of L. foetidus except Fig. 74 (L. terrestris). Fie. 47.,—All of the abnormalities drawn from one very abnormal worm. The numbers inserted between the spirals, compound metameres, &c., give the number of normal rings that have been omitted. Fics. 48—50.—Vasa deferentia opening on different metameres. The figures as seen from below. Fies. 51, 52.—Vasa deferentia doubled on one side. Seen from below. Fires. 53—55.—Both vasa deferentia on 10th, 12th, and 14th metameres respectively. Projections from above. Fies. 56—58.—Vasa deferentia on alternate segments. Diagram as seen from above, so that right and left of Figs. 53—60 are reversed as compared with Figs. 48—52 (below). Fries. 59, 60.—Vas deferens doubled on one side. Projections from above. Fics. 61—64.—Dissections of compound metameres to show arrangement of septa, &c. Fies. 65—70.—Dissection of spirals to show condition of septa. Fig. 71.—Failure of surface line to meet above. Septa were normal. 476 T. H. MORGAN. Fic. 72, A B.—Disagreement between surface line (a) and septa (B). Fie. 73.—Dissection of a compound metamere. Septa do not agree with surface lines. Fie. 74, 4 B.—Vas deferens doubled on right side. L. terrestris. Fies. 75—78.—Abnormal arrangement of metameres in young worms at time of emergence from coccoon. Fig. 75, category 1; Fig. 76, category vir; Fig. 77, category vii; Fig. 78, category x (Pl. 40). PLATE 43. Fics. 79—81.—Regenerating posterior ends of the body of L. fetidus, See text for description. Figs. 82—85.—Abnormal arrangement of the metameres in Amphinome. Fies. 86 —88.—Abnormal arrangements of the rings of leeches (Macrob- dilla decora). Fic. 89, A 8 c.—Compound metamere 5—5 in abdomen of locust. Fics. 90—95.—Abnormal arrangements of the rings of the lobster, Homarus americanus. Fic. 96.—Nine segments of the arm of a brittle-star (Opliiuridea) from Jamaica, as seen from below. Every fourth segment is a pigmented colour-ring. Fic. 97.—Nine segments of another individual, showing the middle pig- mented ring broken, Fic. 98.—To show a further separation of colour-rings. ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 477 On the Celom, Genital Ducts, and Nephridia. By Edwin 8S. Goodrich, F.L.S., Assistant to the Linacre Professor of Comparative Anatomy, Oxford. With Plates 44 and 45. Tue chief object of this paper is to call attention to a theory of the homology of the coelom which has been gradually gaining ground abroad, but has not, I venture to think, received in this country the notice which it deserves. The theory I refer to is, that the cavity which we know as the coelom in the higher Coelomata is represented by that of the genital follicles in the lower types of that grade. Ever since Hatschek wrote the often-quoted words: “ Die secundire Leibeshohle verhalt sich wie die Héhle der Geschlechtsdriise der niedrigeren Formen,” and pointed out that true meso- blastic metamerism is really due to the repetition of the gonads (48), so many favourable new facts have been brought to light, that from a suggestion the statement has become a well- established theory. In a most interesting and suggestive paper on the ancestry of the Annelids, Dr. E. Meyer has set forth the theory in some detail (81). After showing that the ccelomic cavities in the Polychetes are quite comparable in development and function with the genital follicles of the Planarians, he further maintains that the theory throws a flood of light on various otherwise obscure questions, such as the bilateral character of the coelom, its invariable connection with the genital cells, the absence of a truly unpaired prosto- 478 EDWIN S. GOODRICH. mial celom, &c. He then treats of the nephridia and genital ducts, and it is with this part of the question that I wish more especially to deal in this paper. Meyer holds that the ne- phridium of the Platyhelminths is represented by the so-called head-kidney in the first segment, and by the tube of the nephridium in the trunk segments of the Annelids, while the genital duct of the Platyhelminths is represented by the wide funnel of the trunk nephridium which develops, independently of the tube from each genital or ccelomic follicle, and becomes grafted on to it afterwards. Unfortunately, the author restricts his remarks almost entirely to the Polychztes and those forms which appear directly to lead up to them. Although the theory has been at all events partially adopted by other writers (Lang, 71; Korshelt and Heider, 67), no one, as far as I am aware, has pushed it to its logical conclusion, and applied it to all the groups of Celomata. This is what I shall attempt to do in this paper. First of all, however, there is one thing to be noticed with regard to Meyer’s general statement about the nephridial funnel, namely, that since the publication of his own researches on the Polycheta, and of those of Vejdovsky and others on the Oligocheta, there can be no doubt that the nephridial funnel in the latter forms part of the true nephri- dium (is, in fact, derived from the end-cell), and is not a grafted genital duct, as is the case in some at least of the Polycheta. An unprejudiced review of the well-established and recently ascertained facts concerning the development of the excretory organs and genital ducts of the Colomata must, I think, inevitably lead us to the conclusion that we have been confus- ing two organs of totally different origin under the one name nephridium—the one organ the true nephridium, the other the morphological representative of the genital duct, which may be called the peritoneal funnel, to avoid confusion. Further, that while on the one hand in certain groups such as the Planaria, Nemertina, Hirudinea, Cheetopoda, Rotifera, Entoprocta, besides the genital ducts or peritoneal funnels, we find true nephridia in the adult; on the other hand, in such ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 479 groups as the Mollusca, Arthropoda, Ectoprocta, Echinoderma, and Vertebrata, there are in the adult no certain traces of true nephridia. In these latter groups, as we shall see, the peri- toneal funnels (primitive genital ducts) take on the excretory functions of the nephridia which they supersede. In the following brief review of the various classes of Ceelo- mata, I shall endeavour to show that the two kinds of organs can always be distinguished ; that the first, the nephridium, is primitively excretory in function, is developed centripetally as it were, and quite independently of the ccelom (indeed, is probably derived from the epiblast), possesses a lumen which is developed as the hollowing out of the nephridial cells, and is generally of an intracellular character, is closed within, and may secondarily acquire an internal opening either into a blood space or into the ccelom (true nephridial funnel as opposed to the peritoneal funnel) ; and that the second kind of organ, the peritoneal funnel, is primitively the outlet for the genital pro- ducts, is invariably developed centrifugally as an outgrowth from the coelomic epithelium or wall of the genital follicle, is therefore of undoubtedly mesoblastic origin, and possesses a lumen arising as an extension of the ccelom itself. In the series of diagrams illustrating this paper, based on the most recent and accurate researches, it has been my con- stant endeavour to interpret the author’s results correctly, and not to distort the facts in favour of the theory here advocated. PLANARIANS. The nephridia of the Planarians, as is well known, are formed of a main duct, which branches out into fine tubules ending blindly internally in flame-cells (fig. 1); they do not develop beyond this “ pronephridial ’’ condition—protonephri- dium of Hatschek (55). The arrangement of this single pair of nephridia is extremely variable; the two organs may join and open by a median external pore near the mouth or behind, or they may open by a number of pores at the sides. Gunda segmentata, a most interesting form described by Professor Lang (69), possesses longitudinal main trunks into which 480 EDWIN S. GOODRICH. open the fine branches ending in flame cells, and from which pass segmental ducts to the exterior, corresponding to the segmentally arranged gonads. It is by the breaking up of such a system into separate organs that Lang would derive the nephridia of the higher Ceelomata. Unfortunately, we know little about the origin of the ne- phridia in this group. Lang has described, in Discocelis, paired ingrowths of the epiblast, which he believes give rise to the nephridia (70). This observation strongly supports his theory as to the phylogenetic derivation of the nephridia from epidermal glands ; and, indeed, it seems pretty certain that an incipient excretory organ to be eflicient must have been derived from, or at all events situated close to, the surface layer in order to get rid of its excretory products. It is in the Planarians, a group undoubtedly primitive! in some respects, that we should expect to discover the celom in its first stages of development, and, in fact, we do seem to be able to trace it from its first appearance. In some Accela (Graff, 42,44), and other simple forms, the gonads consist merely of the genital cells lying freely in the parenchyma. In others, these cells become surrounded by an epithelium formed by the adjacent cells; the epithelial sacs, one on either side of the body, may then become hollow, while the wall grows out to form two tubes, the genital ducts (peritoneal funnels). Another important stage is presented by these organs in Gunda segmentata (Lang,69). Here the genital follicles are repeated segmentally, the first pair being ovarian, the rest testicular sacs. If these 1 One of the most useful lessons of modern research has been to teach us with what great care the word “primitive” should be applied to any group of existing animals. A few years ago naturalists readily derived one group of living animals directly from another, apparently more primitive; but their genealogical trees are now becoming reduced to bushes, in which the branches spring from a common base. Nevertheless, it is true that certain groups may retain, either in their general organisation or in some particular details, characteristics of the ancestors from which they have diverged. The Plana- rians, with their complicated nephridial and genital apparatus, their deeply- sunk nervous system, yet generally archaic plan of structure, are a striking case in point. ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 481 follicles were larger, Gunda segmentata could be calledatruly segmented animal.! The inner ends of the genital ducts are formed as outgrowths from the genital follicles : ‘‘ Der Oviduct ist bei Gunda segmentata, wie bei Planariatorva anfangs ein solider Zellenstrang. Zweifellosensteht er durch Wucherung aus dem soliden ovarium selbst, ahnlich wie die Samenleiter Aus- wiichse der Hoden sind” (observations confirmed in his later work, 70).2 The oviduct becomes hollow and ciliated, and grows backwards to the genital pore; the vasa efferentia fuse to form the main sperm-duct. Tosum up,then. In the Planarians the excretory organs area pair of pronephridia, probably derived from the epiblast; the gonads arise from a mass of cells in the mesoblast, which may become hollowed out into a genital follicle (coelomic sac) from the wall of which arise the genital cells. The follicle grows out to form the genital duct (peritoneal funnel), which joins an epiblastic invagination at the genital pore (fig. 1).° 1 £. Meyer believes (81) that the ancestor of the Annelids possessed a pair of long genital follicles, and that metamerism was brought about by their being broken up at intervals, chiefly to facilitate its serpentine motion; each portion would then have acquired its own duct to the exterior. It seems to me more probable that the metameric arrangement of the genital follicles is more directly due to that “tendency” to repetition by a sort of budding, which is seen in the case of the gonads, the penes (Anonymus), and even the pharynx (Phagocata; Woodworth, 112) amongst the Planarians, and again, perhaps amongst the Mollusca (for a full discussion see Bateson, 3). 2 Jijima’s account of the development of the sperm-ducts differs somewhat from that of Lang, but he derives both from the mesoderm (60). 3 The spaces contained in the connective tissue or parenchyma have been sometimes compared with the ceelom; these spaces seem rather to represent the vascular system of the higher Coclomata. I need not treat here of the homology of the vascular system, which is probably of quite separate origin from the celom. Professor Ray Lankester’s view, that the blood-system is simply a liquefaction, as it were, of the mesoblast, seems to me to agree per- fectly with the facts. Moreover, the theory held by many authors that it is directly derived from the blastoccel appears to be quite untenable. As Pro- fessor Lankester has pointed out to me, if this were the case we should expect to find the blood-spaces best developed amongst the Diploblastica ; now it is just in the (adult) Colenterates that it is entirely absent. Indeed we may say that the blood-space, or hemoccel, does not appear in phylogeny 482 EDWIN S. GOODRICH. About the Cestodes and Trematodes it need only be said that they are built (so far as concerns the question discussed in this paper) upon essentially the same plan as the Planarians;? while as to the Nematodes, of the development of which we know too little, it seems probable that here also the genital follicles represent the coelom, while the body cavity is a blood- space corresponding in its relations to the parenchyma of the Planarians. RoviFrEeRaA. The Rotifers, recently described in great detail by Platte (87, 88) and Zelinka (115), agree in the general structure of the nephridia, genital follicles, and genital ducts, so closely with the Platyhelminths that they may be dismissed with a very few words. The nephridia are a pair of branching tubes ending internally in flame cells, and opening behind into the cloaca.’ The colom is represented by a pair of genital follicles, one of which only is generally developed; the wall of the follicle is produced backwards to form the genital duct, or peritoneal funnel opening into the cloaca. The development of the nephridia has not yet been thoroughly worked out. Zelinka has traced them to a group of cells of doubtful origin. He adds: ‘das Exkretionssystem konnte ich in so fern mit Sicherheit auf das Ektoderm zurickfihren, als es bestimmt nicht auf das Entoderm bezogen werden kann ”’ (115), until the mesoblast has been formed, and, generally speaking, that the greater the development of mesoblast the more definite is the vascular system. I should rather consider the connection of the blood-spaces with the blastoccel as arising for purely mechanical reasons, so to speak, and in no way of phylo- genetic significance. The temporary continuity of the blastoccel with the blood-space during the ontogeny of certain of the higher forms would seem to be connected with that method of development by the folding of germ-layers or sheets of tissue, which no one would look upon as primitive. During this process, the cavities which will form the vascular system later on, are ine- vitably continuous with the space left between the germ-layers. 1 Fraipont maintains that the flame end-cells of the nephridia of the Cestodes and Trematodes communicate internally by means of a small lateral aperture 39). : 3 As in the case of some Platyhelminths, the nephridia have been stated to open internally (Eckstein, 28). ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 488 ENTOPROCTA. This group also, from our present point of view, differs little in its organisation from the Planarians. A pair of nephridia—short tubes, generally with an intracellu- lar lumen and ending in a flame-cell or a group of cells with cilia,—open by amedian pore in front of the genital pore (fig. 4) (Hatschek, 47; Harmer, 46a; Ehlers, 29; Davenport, 26). Possibly in some forms they open internally (Joliet, 61),andin Urnatella they may be connected with a system of branching tubules ending in flame-cells (Davonport, 26). The origin of the nephridia is doubtful; Hatschek traced them in the embryo to cells which he believed to be mesoblastic (47). Embedded in parenchymatous tissue are a pair of genital follicles (fig. 4). From each a typical peritoneal funnel leads to a median pore (Ehlers, 29). Mo.uvsca. We now have to deal with a group of animals in which, as I shall endeavour to show, the embryo is provided with a pair of true nephridia; later, when the two ceelomic sacs belonging to the unsegmented trunk have acquired a considerable size, the peritoneal funnels formed from their walls take on the excretory function, whilst the nephridia degenerate !, True nephridia have been described in all the groups of Mollusca except the Isopleura and the Cephalopoda, and have recently been the subject of a special study from Dr. R. von Erlanger (34, 35). They consist of short tubules formed, as a rule, of one or of a small number of cells, pierced by a canal which communicates with the exterior by a pore behind the velum (fig. 20). The inner end of the nephridium is provided with a flame-like bunch of cilia, or with a flagellum. An internal opening into the blood-space or head-cavity has been observed in some forms, such as Lymneeus, Helix (de Meuron, 79; Jourdain, 62; Fol, 38; Sarasin, 92; Wolfson, 111; v. Erlanger, 35, &c.), Teredo (Hatschek, 50), and perhaps in 1 This view evidently obviates the difficulty some have felt as to the presence of two pairs of so-called nephridia in an animal composed of one segment. 484 EDWIN S. GOODRICH. Cyclas (Ziegler, 116). In other cases, such as Paludina, and Bythinia (v. Erlanger, 32, 33), the nephridium appears not to open internally, but to remain in the pronephridial stage. The first stages in the development of the nephridium of the Mollusca have, unfortunately, not yet been satisfactorily worked out. Rabl (88a) derives the nephridium in Planorbis from a large cell which also gives rise to the mesoblast. Wolfson traced that of Lymnzus, which, he says, is derived from a large in-wandering epiblastic velar cell on either side (111). Hatschek (50), v. Erlanger (32, 33), and others trace it to cells which they consider to be of mesodermal character, but the exact origin of which is not clear.! In some cases a late epidermal invagination is said to take place at the nephri- diopore which forms the peripheral end of the duct (v. Erlanger, 32, 34). We must now examine the development of the excretory organs of the adult Mollusca, which appear to be nothing but peritoneal funnels. An accurate description of the develop- ment of Paludina has recently been given by R. von Erlanger, and although the ontogeny is much modified owing to the asymmetry of the adult, I shall begin with it as it is the only detailed account we have. The genital follicles or ceelomic sacs (pericardium) arise as a cavity on either side, a hollowing out of the mesoblast (fig. 20). These cavities enlarge and fuse below the gut. On either side a thickening takes place on the ventral wall of the coelom, which here grows out in the form of a typical peritoneal funnel (fig. 21). The right peritoneal funnel then enlarges and fuses with an epidermal invagination, which forms the outer end of the excretory organ. ‘‘ Was die Niere anbelangt,” says von Erlanger, “so bin ich der Ansicht, dass der secernirende Abschnitt derselben aus dem Mesoderm 1 IT think we may safely say that there is nothing which precludes the pos- sibility of the nephridia being primitively derived from the epiblast, as they appear to be in the Platyhelminths, Rotifers, and Cheetopods; although the fore- casts may sink in at a very early stage and thus become included in that “ An- lagecomplex” which we call mesoblast. In Helix, de Meuron derives them from the epiblast, but he is not positive about the internal extremity (79). ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 485 stammt und dass diejenigen Beobachter, welche ihr aus dem Ektoderm enstehen lassen, entweder nur den ausfiihrenden Theil der Niere beriicksichtigt haben oder, was noch haufiger geschicht, die beiden Abschnitte nicht in ihren Zusammenhaug erkannten: Der ausfithrende Theil wird namlich von Allen, mit Ausnahme von Rabl, aus einem Theil der Mantelhdhle abgeleitet ” (82). On the left side, to which the genital function is restricted, the gonad develops from the wall of the celom ; then, together with the rudimentary left peritoneal funnel, it becomes constricted off from the main division of the colom (the pericardium), forming a small genital sac. From the wall of this sac the genital duct grows out, and joins an epidermal invagination, like the peritoneal funnel of the right side. Doubtless the genital duct is really the left peritoneal funnel, as v. Erlanger suggests : ‘Ich konnte.... feststellen, dass die Anlage der Genitaldriise in der urspring- liche linken Halfte des Perikards ensteht, und zwar ungefahr da, wo sich die rudimentire linke Niere zuriickgebildet hat. Eben so ensteht auch die Anlage des Ausfiihrganges an der Stelle, wo der rudimentire Ausfiihrgang der linken Niere sich befand, und scheint einfach aus diesem hervorzugehen.”” These observations have been fully confirmed in the case of Bythinia (33), and the conclusions, as we shall see shortly, are supported by the comparative anatomy of the Mollusca in general (fig. 22). Ziegler has described the development of Cyclas cornea, which in some respects is less modified, the adult being symmetrical (116). Here also the ccelom (pericardium) arises as a right and left follicle. The two cavities enlarge, surround and fuse below the gut. On either side the organ of Bojanus develops as a peritoneal funnel, which meets an epidermal invagination (fig. 22). In the Mollusca, then, we find at an early stage a pair of true nephridia (Urniere, head-kidneys), possibly of epiblastic origin. The ccelom develops as two cavities in the mesoblast, genital follicles, from the walls of which grow out two peritoneal funnels, the organs of excretion and carriers of the 486 EDWIN S. GOODRIOH. genital cells of the adult. It can hardly be doubted that primitively the renal organs (organs of Bojanus, nephridia of authors) functioned as genital ducts. Such is, indeed, still the case in Chetoderma, the Neomenians, and the Zygobranchia. Also in the most primitive Lamellibranchs, such as Nucula and Solenomya, the peritoneal funnels still retain their original function. As Pelseneer has shown (86), gradually a separate genital duct has been split off, which in Anodon, Cardium, &c., opens independently. Intermediate stages are found in such forms as Pecten, and Spondylus, where the genital cells are shed into the kidney itself; and in Arca, Ostrea, &c., where the kidney and genital duct open into a common cloaca. Likewise in the Chitons a separation has taken place of the genital region of the ccelom from the renal ; the gonad then acquires special ducts, which may not be homologous with peritoneal funnels. In the Cephalopoda, although the coelom remains continuous, special apertures serve for the escape of the genital cells ; whether these should be con- sidered as a second pair of peritoneal funnels is also doubtful.! (The aberrant form Rhodope veranii would appear to be amongst the Mollusca, if it be really of that class, the only one which retains true nephridia in the adult. It is provided with a pair of branching tubes, opening by a common pore, and ending internally in flame-cells. The coelomic or genital fol- licles, ovarian and testicular, are small and numerous. Their ducts join to a common hermaphrodite duct, which again di- vides into two openings by male and female pores [48, 12].) DInopui.vs. For our purpose it will be convenient to treat this genus separately, and not with the Archiannelida which will be in- cluded with the Polycheta. It is of great interest, since, in some species at least, the nephridia are metamerically repeated whilst the colom is represented by a single pair of genital follicles. 1 Such a case may be compared to that of certain Nemertines, where nu- merous genital follicles and as many genital ducts are present in one segment, and to the similar arrangement in the Vertebrates. ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 487 In Dinophilus apatris, Korschelt has described flame-cells which he believed to be connected with a longitudinal canal opening posteriorly (66). E. Meyer, however, has figured in D. gyrociliatus five pairs of typical closed nephridia, or pronephridia, ending in flame-cells internally, and disposed metamerically according to the outer signs of segmentation (80). Harmer (46) describes five pairs of very similar nephridia in the female D. teniatus. In the male there are four pairs of nephridia, perhaps with internal openings.! The testes are a large right and left sac, which fuse across the median line ; as Harmer himself says, it is ‘‘ possible that in the connective-tissue lacune of the body of Dinophilus we have the representative of the so-called ‘primary body-cavity,’ whilst in the fully developed male the ‘secondary body-cavity’ is repre- sented by the cavity of the testis, with which the funnels of the vesiculze seminales are connected.”’ It might be added that these funnels, which form the inner openings of the sperm-ducts, have all the appearance of true peritoneal funnels, comparable to the genital ducts of the Platyhelminths, and the other groups already spoken of. The paired ovarian cavities appear to be provided with only very degenerate ducts, reduced to mere pores in D. vorticoides and D. apatris (compare the Archiannelida and the female ducts of certain Oligochzta such as the Enchytreeids). The structure of Dinophilus might be explained in one of two ways. Hither it has acquired a number of nephridia, whilst retaining the primitive single pair of genital follicles ; or it is a degenerate form which has lost its metamerically repeated genital follicles, whilst retaining a number of separate nephridia. Our present knowledge does not enable us to con- clude for certain which of these explanations is the correct one.” Histriodrilus Benedeni (Histriobdella homari) may 1 The fifth pair is possibly, according to Harmer, represented by the distal portion of the genital ducts leading to the penis (compare with certain Poly- cheeta where the nephridium fuses with the peritoneal funnel). 2 In a paper which has just appeared Schimkewitsch describes a ladder- like nervous system, and traces of segmentation in the developing mesoblast (‘ Zeit. f. w. Zool.,’ Bd. lix, 1895). VOL. 37, PART 4,—NEW SEk. KK 488 EDWIN 8S. GOODRICH. be closely related to Dinophilus. It possesses five pairs of pronephridia (four in the female), and a distinct coelomic or genital cavity also opening by one pair only of peritoneal funnels to the exterior (Foettinger, 37). Here it seems to be pretty certain that there was originally a segmented cclom ; there is still a ventral chain of ganglia. NEMERTINA. The nephridia of the Nemertines consist essentially of a longitudinal canal on either side of the anterior region of the ali- mentary canal, opening to the exterior by one or by several pores situated laterally at more or less regular intervals (von Kennel, 63; Oudemans, 85; Hubrecht, 59; Burger, 15,17). Internally they have been stated to open into the blood-vascular system ; the latest researches, however, do not support this view (15), and Birger has shown that in several species the nephridial canals give off fine branches ending in bunches of flame-cells (17) (figs. 2, 3, and 24; in these diagrams the nephridia are represented in the same region as the gonads). Although the development of the nephridia has not been followed out in detail, Hubrecht (58) and Birger (18) have traced their origin from direct invaginations of the epiblast. In this group of elongated worms the genital follicles are numerous, and generally arranged in pairs, alternating with the intestinal ceca. Each follicle communicates with the ex- terior by a duct or peritoneal funnel, formed as an outgrowth from its wall at a comparatively late period (figs. 2, 3, and 24). It is hardly necessary to emphasise the striking similarity between this metameric arrangement of follicles with their cor- responding ducts and the almost identical metameric coelomic follicles and genital ducts of the Chzetopods.! OLIGOCH ATA. Thanks to the numerous researches of Profs. Hatschek, 1 R. 8. Bergh, in 1885, in a paper which I have not seen, pointed out the resemblance between the genital follicles of the Nemertines and the eelomic cavities of the Chetopods; but he considered the ducts of the follicles to be homologous with the nephridia of the latter group (6). ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 489 Bergh, Wilson, Vejdovsky, and others, we have now a very detailed history of the development of the nephridia in the Oligochetes. The observations of Hatschek, Wilson, and Bergh do not coin- cide in many particulars; but all these discrepancies have been so admirably reconciled by Vejdovsky in his careful work on Rhynchel]mis and other forms (101), that we can, I think, be now quite confident that we have a satisfactory account of the embryology of these organs ; more especially since these results have been in many respects amply confirmed by the researches of Whitman, Bergh, and Biirger on the Hirudinea. The development of the nephridium in Rhynchelmis has been most carefully described by Vejdovsky (101). In the first stage he figures it consists of a large cell (trichterzelle or funnel-cell) within or on the posterior surface of the septum. This “ funnel-cell”’ divides and gives off a string of small cells behind, from which is developed the canal of the nephridium. A large vacuole appears between the two cells formed from the funnel-cell itself, in which a flame-like flagellum is developed. Vacuoles now arise in the posterior string of cells, fuse together, and form the lumen of the canal which communicates with the end-chamber containing the “ flame” (figs. 6 and 25). Finally this chamber opens into the ccelom, and the posterior loop joins the skin; a communication is established with the exterior by means of a secondary invagination of the epidermis—the end- vesicle. Quite similar is the development of the nephridium in Stylaster and Tubifex (Vejdovsky, 100). Bergh (9 and 10) traced back the nephridia in Criodrilus and Lumbricus to a large cell, the funnel-cell, lying close to the epiblast and between each successive pair of solid mesoblastic somites. When these become hollowed out, the funnel-cell buds off posteriorly a chain of cells, the future canal of the nephridium ; vacuoles appear in these cells, as in Rhynchelmis, to form the lumen. Meanwhile the funnel-cell itself, which has retained its large size, pushes through the mesoblast to reach the ceelomic cavity in the segment in front; here it divides, acquires cilia, and becomes the funnel of the adult nephridium, 490 EDWIN &. GOODRICH. The posterior canal grows to the surface, where it opens through the epidermis ; in some cases there is here an invagination of the epidermis to form the end-vesicle (Vejdovsky). As to the origin of the funnel-cell—the forecast of the whole true nephridium: it arises from the primitive cell row, or nephric cord, formed by the repeated division of one of the teloblasts on either side. In the earlier stages this teloblast and the nephric cord to which it gives rise lie on the surface of the embryo; thus the funnel-cells are epiblastic in origin. From the nephric row one cell enlarges and enters into con- nection with each successive segment, as described above (fig. 5). To judge from the figures of Bergh, Wilson (108 and 109), and Vejdovsky, in some forms, such as Dendrobena and Lumbricus, the funnel-cells give off the chain of posterior cells, whilst separating from the nephric row, thus remaining for some time in connection withit. In other cases, such as Criodrilus, the funnel-cells appear to separate first.! In the embryo of most Oligochetes the nephridia of the first segment are developed precociously to perform the excretory functions at an early stage (fig. 25). Vejdovsky (100 and 101) has described these organs in Rhynchelmis, Chetogaster, AZolosoma, Nais, Allolobophora, Lumbricus, Dendrobzena, &c., and Bergh described those of Criodrilus (9). They consist of fine canals with an intracellular lumen, or some- times of wider tubes; they'are often ciliated, and occasionally end internally in a flame-cell; they appear to be always blind within. Externally they open either by a median dorsal pore or by two lateral pores on the first segment. In fact they closely resemble the closed pronephridia of the Platyhelminths, Entoprocta, and other groups we have already examined, or the pronephridial stage of the trunk nephridia. The origin of the cells which form the (larval) nephridia of the first segment has not been traced; but since they arise (in some cases at least) before the division of the promesoblast cell, 1 Bergh denies the derivation of the funnel-cell from the nephric row. Wilson rightly traced the development of the main body or canal of the nephri- dium from the nephric row, but failed to discover the origin of the funnel- cell itself. ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 491 Vejdovsky considers it probable that they are derived from the epiblast, a conclusion which agrees with the known develop- ment of the posterior nephridia. The mesoblast in the Oligochetes is formed, as in all Annelids, as two germ bands, which become broken up into separate somites. The hollowing out of these gives rise to the coelomic follicles, which increase in size, surround the gut (a stage resembling what we find in the Nemertines), and fuse below it (figs. 6 and 25). The transverse septa, between adjacent follicles, become pierced, allowing a communication from one to the other (Kowalevsky, 68; Hatschek, 48; Wilson, 109; Vejdovsky, 101). From the wall of certain of these follicles the gonads are developed, whilst others remain sterile. The number and position of the fertile follicles varies consider- ably according to the family of the worm in question, and even amongst different individuals of the same species (Woodward observed an earthworm with seven pairs of ovaries; 118, 114). The genital ducts (peritoneal funnels) develop as a thicken- ing of the celomic epithelium in the fertile segments, which grows outwards towards the epidermis, with which it fuses (figs.6,7,and 25). Vejdovsky, who has followed thedevelopment of these organs in several forms, such as Stylaria, Cheetogaster, the Enchytreids, and Tubificids, says: “Die Anlage des Samenleiters wiederholt sich nach dem oben Dargestellten in iibereinstimmender weise bei allen bisher beoachtete Familien. Die zuerst zum Vorschein kommende Anlage des Samentrichters besteht aus einer Zellvermehrung des Peritoneums an den Dissepimenten der betreffenden Segemente” (100). His observations have been confirmed by Bergh (8) and Lehmann (74) in the Lumbricids.t In the case of the male ducts, these organs may be further complicated by the fusion of two con- 1 Beddard (4) tried to show that the genital ducts were derived from the nephridia in Acanthodrilus. The few facts he brings forward from the very scanty material at his disposal do not, I think, prove his case. ‘The theory of Claparéde that the genital ducts of the Oligochetes were the modified nephridia of the genital segments was founded on an erroneous notion, since thoroughly disproved by Vejdovsky’s observations. 492 EDWIN 8S. GOODRICH. secutive peritoneal ‘funnels, and by the invagination of the epidermis at the genital pore to form an atrium and penis. We may now sum up the main characteristics of the nephridia and genital ducts in this group, which has been treated at length owing to its great importance (figs. 5, 6, 7, and 25). The nephridia of the Oligochetes are probably of epiblastic origin. They develop from large cells (‘‘funnel-cells” ), arranged metamerically outside and between each pair of somites. They pass through a more or less disguised prone- phridial stage (comparable to that permanently retained in flatworms, &c.); in the first (most forms), and sometimes in the trunk segments (Chetogaster) they never develop beyond that stage. In the other segments the nephridia grow towards, and open into, the ccelom by means of a funnel formed from the original ‘‘funnel-cell.”! The genital ducts, on the other ai are peritoneal funnels of undoubted mesoblastic origin, which grow outwards from the metameric genital follicles to open to the exterior. They thus have no connection with the nephridia, and differ from them entirely in their development. HIRUDINEA. So closely do the ceelom, genital ducts, and nephridia of the Leeches agree in their development with those of the Oligo- cheetes, that their history need only be rapidly sketched. Birger (16 and 19) has carefully traced, in several forms, the origin of the whole nephridium proper (funnel and canal) from a large “ funnel-cell,” which comes to lie in the hinder wall of each ceelomie follicle. Just as in the previous group of worms, this cell buds off a row of cells behind which constitute the canal; the “funnel-cell”’ then divides up into a ring of small cells, which form the funnel of the adult nephridium (figs. 8and 9). This organ remains closed in some forms, such as Hirudo, but opens into the ccelom in others, such as Nephelis. 1 The branched so-called plectonephric condition of the nephridia in certain earthworms has recently been shown to arise by the secondary subdivision of originally paired nephridia (Vejdovsky, 102; A. G. Bourne, 18). ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 493 Meanwhile the lumen of the canal in the posterior chain of cells becomes hollowed out. The peripheral end of the canal fuses with an invagination of the epidermis, the vesicle, by means of which it opens to the exterior. The first origin of the “ funnel-cells,’’ from which the nephridia are formed, has not been traced in detail; it seems quite probable that they are derived from the nephric rows described by Whitman (107) in Clepsine. (They are possibly the large cells mistaken by Whitman [106] for the forecasts of the testes, as suggested by Bergh.) As in the Oligocheta and Polycheta, so in the Hirudinea the nephridia of the anterior segments develop precociously in the larva. Bergh (7) has traced their origin as outgrowths from the epiblastic cell-rows. In Aulastoma there are four pairs, which never develop beyond the pronephridial stage, i.e. do not open internally. Bergh was also unable to find external openings (compare the nephridia of Capitella, 30 and 81). The ccelom develops in a normal manner as a hollowing out of the paired metameric blocks of mesoblast. Most of these ceelomic or genital follicles are fertile: an anterior pair develop the ovaries on the peritoneal wall; several posterior pairs develop the testes. That part of the ccelom which surrounds the gonads generally becomes partially separated off as a peri- gonadial ceelom (Bourne, 12a; Birger, 16). That the genital ducts are peritoneal funnels is shown by their development, although it seems to be somewhat modified. The oviducts arise from the ccelomic epithelium surrounding the ovary, and fuse with the two ends of a forked, but median invagination of the epidermis (fig. 9) (Biirger, 19). The vasa efferentia are similarly formed from the testes (fig. 28); they grow outwards and forwards, fusing with those in front. The most anterior join the ends of a median forked invagination of the epidermis (Nussbaum, 84; Biirger, 19). The complete genital ducts thus closely resemble those of some Planarians (Gunda), and differ essentially from those of the earthworm only in the number of peritoneal fnnnels which contribute to their formation (compare also the Vertebrates). 494, EDWIN S. GOODRICH. ARCHIANNELIDA AND PoLycH#Ta. The first origin of the nephridia in these worms is not so well known as in the case of the Oligocheta. It is, however, to be remarked that in the only case where the forecast of the nephridium appears to have been traced from the beginning, it has been found to arise from the epiblast (the head-kidney in Nereis; Wilson, 110).!| This would agree with what we have seen occurs in most, if not all, of the groups we have already examined. E. Meyer (80) has given us a most excellent description of the development of the nephridia in Polymnia and Psygmo- branchus. They arise on either side from large cells, situated close to the epiblast between each pair of mesoblastic somites, with which they have no connection at this early stage. These large cells, which seem to me obviously homologous with the “ funnel-cells” of the Oligocheta and Hirudinea, divide, form- ing a short chain of cells within which an intracellular lumen becomes hollowed out (figs. 10, 11, and 26). The mesoblastic somites become hollowed out to form the genital or celomic follicles, from the posterior wall of which the ccelomic epithe- lium becomes pushed out, forming a typical ciliated peritoneal funnel, which fuses with the internal blind end of the nephri- dium (figs. 11 and 26). This specialised portion of the coelomic epithelium forms the wide-mouthed funnel of the adult nephri- dium (fig. 12). The lumen of the nephridial duct becomes intercellular by the multiplication of the cells which constitute its wall, and breaks through at its point of junction with the funnel on the one hand (opens, in fact, here into the ccelom), and establishes a communication with the exterior through the epidermis on the other. Such is the history of the wide- mouthed segmental organs of compound origin of the trunk. The nephridia of the first segment (or sometimes of several 1 Hatschek (48, 51, 54), Salensky (91), and von Drasche (27) all consider the cells from which the nephridia are developed to be of mesoblastic origin, but the evidence on this point is not convincing. Possibly, however, as sug- gested for the Mollusca, they have secondarily come to be derived from the mesoblast, ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 495 anterior segments) develop in the same way as the posterior, but never pass beyond the pronephridial stage (head-kidneys) ; they end blindly internally with a typical flame-cell (fig. 26). Meyer figures as many as five pairs of such pronephridia in Nereis cultrifera (80)'. Although the head-kidneys of the Mollusca undoubtedly occasionally open internally, v. Drasche and Hatschek seem to have been mistaken in describing an internal opening in these organs in the Cheetopods (see Frai- pont, 40; Meyer, 80). Fraipont appears to attribute a very similar history to the wide-mouthed trunk ‘“‘nephridia”’ of Polygordius as Meyer has described for the trunk “nephridia” of the Tubicolous Poly- chetes, though his statements are less precise: “ Le méso- blaste est representé de plus au niveau des muscles obliques par une masse de cellules assez confuse. C’est dans ce groupe de cellules situées au dessus des muscles obliques contre les champs musculaires longitudinaux que ce différencient les entonnoirs des organes ségmentaires et plus tard encore les organes sexuels, C’est un simple épaississement du péritoine ” (40). We see, then, that in the Polycheta nephridia are developed from large cells, which may be compared to the “ funnel cells,” giving rise to the nephridia in the Oligocheta. Whilst, however, in the latter the pronephridium acquires an opening into the ccelomic follicle independently of the peritoneal funnel, which acts as a genital duct ; in the former, the Polycheta, the pronephridium may acquire an opening into the ccelomic follicle in the region where the peritoneal funnel is formed, fuse with it, and become an organ of double function—excretory and genital. In many cases division of labour leads to the restriction of the genital function to one set of ccelomic follicles and their funnels, and of the excretory function to another set 1 There can now, I think, be no doubt that the head-kidneys are simply the precociously developed nephridia of the first segment; they do not open into the ccelom for the very good reason that at this stage there is, as a rule, no ccclom for them to open into. They preserve the same relations as the Platy- helminth nephridia (Hatschek, 48, 51, 54; Meyer, 80). 496 EDWIN 8S. GOODRICH. of celomic follicles and their funnels. This may lead to a corresponding differentiation of structure ; in the first set the peritoneal funnel becomes the most important part, in the second the nephridial portion (such specialisation has been well described in many forms by Eisig, Meyer [80], Trautzsch [98], Cunningham [25], Marion and Bobratzky [78], &c.). The fusion between the nephridium and peritoneal funnel does not occur in all Polychetes ; fortunately, we appear still to have all the intermediate stages between this condition and that of the Oligochetes. Meyer (80) has shown that in Nereis the nephridia of the first five segments have the typical pronephridial structure with a flame end-cell, and that in the posterior segments this end-cell (judging from his figures) opens into the ceelom (true nephrostome ; compare Rhynchelmis).' I have also described in the Lycoridea (41) a ciliated region of the ccelomic epithelium which I believed to be the peritoneal: funnel. It is, however, to Hisig that we owe the description of what appear to be intermediate stages. In Dasybranchus and Tremomastus we have conditions in which the peritoneal funnels (Genitalschlauche) are separate from the nephridia and open independently (fig. 14), and in which the two organs are con- nected but still open separately (fig. 13). Perhaps in certain segments of these forms and of Capitella we have the more usual Polychete arrangement, in which the peritoneal funnel no longer acquires an independent opening (81). ARTHROPODA. It will be best to begin our review of this group with a brief recapitulation of the development of Peripatus, which has been so excellently described by Mr. Sedgwick (94), and Dr. von Kennel (64). Soon after the metameric somites have been hollowed out to form the celomic follicles, the upper half of each coelomic cavity becomes nipped off from the lower half. From the wall of each of these lower ccelomic sacs a peritoneal 1 Such would appear to be the condition in Protodrilus, where the nephri- dial funnels figured by Hatschek (52) are small, and provided with a flagellum. ON THE C(@LOM, GENITAL DUCTS, AND NEPHRIDIA. 497 funnel is formed as an outgrowth, which fuses with the epidermis (figs. 15 and 16). V. Kennel maintains that in Peripatus Edwardsii the mesoblastic funnel is met by an epiblastic invagination, and the question arises as to whether this invagination represents a true nephridium, in which case the segmental organ of Peripatus would be a compound organ similar to that of Psygmobranchus (see above), or whether it is merely a secondary invagination of the epidermis, such as occurs more or less pronounced in almost every case where a tube opens on to its surface (vesicle of the nephridia in the Hirudinea, peripheral end of the genital ducts of the Oligo- cheta, &c.). I am inclined to take the latter view, and con- sider the segmental organs of Peripatus as purely peritoneal funnels which have assumed the excretory functions. Whilst these organs have developed in this way, the dorsal or genital halves of the somites in the posterior segments have become fused, forming two genital tubes communicating posteriorly with the undivided ccelomic follicles of the last segment. The peritoneal funnels of this segment retain their primitive func- tion, and develop into the genital ducts (fig. 17). The peri- pheral ends and median portion of the ducts are probably derived from the epidermis. The history of the genital and excretory organs of the other groups of Arthropoda can easily be brought into agreement with the development of these parts in Peripatus. However, whereas in the latter all the celomic follicles give off peritoneal funnels, in the Crustacea, Arachnida, Myriapoda, and Hexa- poda the peritoneal funnels are only fully developed in a very few segments. The shell glands and green glands of the Crustacea have been shown to bear the relations of peritoneal funnels (Grobben, 45; Weldon, 104; Marchall, 77; Allen, 1) and develop from the ceelomic follicles. ‘ Bei Daphnia,” says Lebedinsky (78), “entwickelt sich . . . die Schalendriise als die Ausstiilpung der Somatopleura, welche sich zur Max.’ richtet und hier sich mit dem Ectoderm vereinigt.” The same author describes the development of the coxal gland of Phalangium as a typical peritoneal funnel derived from a ccelomic follicle (73), 498 EDWIN 8. GOODRICH. which description agrees with that of Laurie (72) of the origin of the coxal gland in Scorpio, and of Kingsley (65) in Limulus. The genital cells are derived from the walls of the ccelomic follicles! of many segments (Heymons, 57; Wheeler, 105). The dorsal portion of these follicles generally fuse to form con- tinuous tubes, or a median genital sac (as in the Crustacea). The ducts are the peritoneal funnels of one segment. The particular segment selected, so to speak, for this purpose varies much in position in different groups, and also according to sex. Wheeler, in his admirable account of the development of the Orthoptera (105), shows that the ccelomic follicles of all the abdominal segments at an early stage begin to develop peri- toneal funnels, but that those of one segment only reach the exterior and form the genital ducts (fig. 30). As far as we can see, therefore, there are no certain traces of true nephridia in the Arthropoda. The segmental organs, the green glands, the shell glands, and the genital ducts are all developed as peritoneal funnels.” SIPUNCULUS. The development of the excretory organ of Sipunculus nudus has been described by Hatschek (53). The nephridium arises from a large cell (? “ funnel-cell” ) which comes to hein the wall of the ccelomic follicle. This cell divides, forming a chain of cells in which a lumen is developed. The outer end joins and opens on to the epidermis; the inner end grows towards and opens into the celom. The ciliated funnel is formed from the cclomic epithelium. From this it would appear that the excretory organ of Sipunculus is of a double origin, formed by the junction of the nephridium with the peritoneal funnel, as in most Polychetes. It functions as the carrier of both genital and excretory products. 1 Sedgwick holds that they are, in Peripatus capensis, derived from the hypoblast. If this be the case, it must be considered as due to some secondary modification. 2 The possibility of the segmental trachez of the Arthropods being derived from the true nephridia should not be lost sight of. ‘The trachee arise com- paratively late, as a rule, as invaginations of the epidermis, and it seems not ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 499 PHORONIS. In the larva of Phoronis, Caldwell describes a pair of nephri- dia, slender ciliated canals blind internally (21). Their origin is still doubtful. The ccolom is developed as two pairs of follicles, of which the larger “ posterior” pair is alone fertile. The excretory organs of the adult consist in Phoronis psammophila and Ph. Kowalevsk1ii of a pair of peritoneal funnels leading to the exterior from the ‘ posterior” ccelomic follicles (Cori, 23). According to Caldwell (22), they are developed in connection with the nephridia of the larva, and would appear to be of a compound nature like those of Polychetes. In Phoronis australis both the anterior and posterior pairs of celomic follicles are provided with their peritoneal funnels, which open by a common duct (Benham, 5). These funnels in Phoronis serve both as renal and genital ducts. Ectorrocra. No true nephridia appear to have been found in these Polyzoa. On the other hand there are two peritoneal funnels, an excellent account of which has been given by Cori in Cristatella (24). They differ from those of Phoronis only in that they open by a common median pore. BRacHIOPODA. Here the ceelomic follicles, formed as paired archenteric pouches, are provided with a pair of wide-mouthed peritoneal funnels opening to the exterior. They function as the carriers both of excretory and of genital products, and possibly the distal end of the organ represents the true nephridium (also in the Ectoprocta), which has fused with the peritoneal funnel (Morse, 83; Blochmann, 11). impossible that they may be formed not from the nephridia, but from those late epidermal invaginations which, as we have seen, so generally occur in connection with the external opening of the peritoneal funnels. In the Ar- thropods above mentioned these funnels disappear in those segments which possess trachez. 500 EDWIN S. GOODRICH. SAGITTA. Archenteric diverticula give rise to the three pairs of coelomic follicles present in the adult Sagitta. The genital cells are precociously developed, and come to lie in the two posterior pairs of follicles (Hertwig, 56). Whether the genital ducts— which in the male segment, at all events, open into the ceelom by ciliated funnels—are peritoneal funnels, or are partly formed from true nephridia, cannot be decided with our present in- complete knowledge of their development. EcuHINODERMA.! Only a very brief reference can be made to this highly modi- fied group. In the larva we find a right and left coelomic follicle, the enterocceels (derived from unpaired or paired archenteric diverticula), which may give rise to a second pair of cceelomic follicles by constriction. The anterior follicles then develop peritoneal ciliated funnels (fig. 18), which open to the exterior, fusing with paired epiblastic invaginations. As arule only the left peritoneal funnel becomes developed (Field, 36; Bury, 20). The genital cells are developed from the wall of the posterior coelomic follicles (MacBride, 75, 76); but how far the genital ducts can be likened to peritoneal funnels is quite uncertain. There appears to be no trace of true nephridia. VERTEBRATA.! As is well known, the coelom in the Vertebrates arises by the hollowing out of a series of metameric blocks of mesoblast. In the lower forms (Balanoglossus and Amphioxus) several of the anterior celomic follicles are formed directly as pouches from the wall of the archenteron. From these follicles, pro- duced by either method, peritoneal funnels are developed 1 Tf the treatment of these last two groups (Echinoderma and Vertebrata) seems to be somewhat too brief and dogmatic, it is that space will not allow me to discuss the subject in extenso. Moreover, we are here treading on such uncertain ground that I do not feel competent to treat of the structure of these animals in full detail, and offer these remarks merely as a suggestion. ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 501 which communicate with the exterior. The gonads develop from the wall of the follicles; in the lower forms a large number are fertile, but in the higher forms the genital products become restricted to a very limited region. No true nephridia have been discovered in this group ; for, until the development of the interesting segmental tubules described by Weiss (103) and Boveri (14) in Amphioxus is known, it is not possible to decide for certain on their nature. Hemichorda or Enteropneusta.—lIn the anterior or pro- boscis region is a coelomic cavity which appears to represent a fused pair of follicles (fig. 23), as is evidenced by the fact that it communicates with the exterior by paired peritoneal funnels (ciliated proboscis pores) in Balanoglossus Kupfferi and B, canadensis, and occasionally in Ptychodera minuta and B. Kowalevskii (Spengel, 96 and 97), and by its development as a bilobed sac in B. Kowalevskii (Bateson, 2).!_ The collar region contains a second pair of ccelomic follicles, also provided each with a peritoneal funnel or collar pore (Bateson, 2; Morgan, 82; Spengel, 97). Behind these is a third pair of follicles which do not develop funnels, but become of large size and encroach on neighbouring segments, sending back blind poste- rior prolongations. The posterior region contains a series of fertile ccelomic follicles, the gonads (fig. 23), each provided with a peritoneal funnel leading to the exterior.” Although the metamerism of this region is not very definitely pronounced, possibly having become obscured through degeneration, the genital follicles bear a remarkable resemblance, both in their structural relations and in their arrangement, to those of the Nemertines, as has already been noticed by Schimkewitsch (93 aud 98a). As in the Nemertines, so also in the Enteropneusta, several genital follicles may occur in one segment. 1 Compare the development of the corresponding anterior pair of cavities in Amphioxus (Hatschek, 49). ? If this explanation be correct, the gill-slits of Balanoglossus are meta- meric and intersegmental as in other Vertebrates. If not correct, we have an almost incredible state of things in which the genital cells are shed into a cavity which is not the ccelom. 502 EDWIN S. GOODRICH. Craniata.—Owing mainly to the recent researches of van Wijhe (99), Riickert (89 and 90), Semon (95), and others, it is now generally concluded that the Craniates originally pos- sessed a metameric series of pronephric tubules, or peritoneal funnels, opening independently to the exterior. In existing forms the funnels, which grow out from the ccelomic follicles, sometimes reach the epiblast as solid rods (fig. 19) ; they then fuse with each other to a common duct which opens into the cloaca (fig. 31), constituting what is known as the pronephros and its pronephric duct. In his excellent paper on the de- velopment of the renal system of Ichthyophis, speaking with regard to E. Meyer’s theory, Semon says: ‘‘ Die segmentalen Ausfiihrgange der segmentalen Genitalfollikel oder Ursegmente iibernehmen neben ihrer urspriinglichen auch noch exkre- torische Funktion, sie werden zu Vornierkanalchen ”’ (95). To conclude our general survey it may be said: that the ceelom can be traced from its smallest beginning as a cavity or cavities in which are developed the gonad-cells: it grows gradually in size and importance until it becomes the body cavity in which the viscera rest ; that the genital ducts (with a few possible exceptions due to secondary modifications) are homologous throughout the Celomata; that the nephridia, which have often been confused with these ducts, can always, when they occur, be distinguished from them ; and finally, that the coelom may secondarily acquire a renal function, in conse- quence of which the peritoneal funnels supersede the nephridia proper as excretory ducts. List or WORKS REFERRED TO. 1, Auten, EK. J.—‘‘Nephridia and Body-cavity of some Decapod Crustacea,” ‘Quart. Journ. Mier. Sci.,? N. 8., vol. xxxiv, 1892-93. 2. Bateson, W.—‘‘ The Early Stages in the Development of Balano- glossus Kowalevski,” ‘Quart. Journ. Mier. Sci.,’ N. S., vol. xxiv, 1884. ‘The Later Stages,” &c., id., vol. xxv, suppl. 1885, and vol. xxvi, 1886. _ 8. Bareson, W.—“ The Ancestry of the Chordata,” ‘Quart. Journ, Mier. Sci.,’ vol. xxvi, 1886. ON THE C@LOM, GENITAL DUCTS, AND NEPHRIDIA. 5038 . Bepparp, F. E.—“ On Certain Points in the Development of Acantho- drilus multiporus,” ‘Quart. Journ. Mier. Sci.,’ N.S., vol. xxxiii, 1892. . Benuam, W. B. B.—‘‘ The Anatomy of Phoronis australis,” ‘Quart. Journ. Micr. Sci.,’ N. S., vol. xxx, 1889-90. . Bereu, R. §.— Die Exkretionsorgane der Wirmer Kosmos,” Bad. ii, 1885. . Bereu, R. $.—‘* Die metamorphose von Aulastoma gulo,” ‘ Arb. Zool. Inst. Wirzburg,’ Bd. vii, 1885. . Bereu, R. 8.—“ Untersuchung tiber den Bau und die Entwicklung der Geschlechtsorgane der Regenwiirmer,” ‘ Zeit. f. Wiss. Zool.,’ Bd. xliv, 1886. . Brereu, R. 8.—* Zur Bildungsgeschichte der Exkretionsorgane bei Crio- drilus,” ‘ Arb. Zool. Inst. Wirzburg,’ Bd. viii, 1888. . Bereu, R. $.— Neue Beitrage zur Embryologie der Anneliden,” ‘ Zeit. f. wiss. Zool.,’? Bd. 1, 1890. . Brocumany, F.—‘ Untersuchungen iiber den Bau der Brachiopoden,’ Jena, 1892. . Boumie, L.—“ Zur feineren Anatomie von Rhodope Veranii (Kdlliker),” ‘Zeit. f. wiss. Zool.,’ Bd. lvi, 1893. 12a. Bournz, A. G.—“ Contributions to the Anatomy of the Hirudinea,” 13. , 14. 15. 16. Li: 18. 19. 20. 21. ‘Quart. Journ. Mier. Sci.,’ vol. xxiv, 1884. Bourne, A. G.—‘ On Certain Points in the Development and Anatomy of some Karthworms,” ‘Quart. Journ. Mier. Sci.,’ N.S., vol. xxxvi, 1894. Boveri, Th.—“ Die Nierenkanalchen des Amphioxus,” ‘Zool. Jahrb.,’ Bd. v, 1892. Bircer, O.— Untersuchungen iiber die Anatomie und Histologie der Nemertinen,” ‘ Zeit. f. wiss. Zool.,’ Bd. 1, 1890. Bireer, O.—“ Beitrage zur Entwicklungsgeschichte der Hirudineen,” ‘Zool. Jahrb.,’ Bd. iv, 1891. Bircer, O.— Die Enden des exkretorischen Apparates bei den Ne- mertinen,” ‘ Zeit. f. wiss. Zool.,’ Bd. liii, 1892. Bircer, O.—“ Studien z. e. Revision der Entwicklungsgeschichte der Nemertinen,” ‘ Ber. Nat. Gess. Freiburg,’ Bd. viii, 1894. ; Burezr, O.—“ Neue Beitrage zur Mntwicklungsgeschichte der Hiru- dineen,”’ ‘ Zeit. f. wiss. Zool.,’ Bd. lviii, 1894. Bory, H.— Studies in the Embryology of the Echinoderms,” ‘ Quart. Journ. Mier. Sci.,’ N.8., vol. xxix, 1888-89. CaLpwELL, W. H.— Preliminary Note on the Structure, Development, and Affinities of Phoronis,”’ ‘ Proc. Roy. Soc.,’ vol. xxxiv, 1883. vou. 37, PART 4,—NEW SER, i 504 EDWIN §. GOODRICH. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. CaLpweLt, W. H.— Blastopore, Mesoderm, and Metameric Segmenta- tion,” ‘Quart. Journ. Micr. Sci.,’ N. 8., vol. xxv, 1885. Cort, C. J.—‘‘ Untersuchungen tiber die Anatomie und Histologie der Gattung Phoronis,” ‘ Zeit. f. wiss. Zool.,’ Bd. li, 1891. Cort, C. J.—“ Die nephridien der Cristatella,” ‘ Zeit. f. wiss. Zool.,’ Bd. lv, 1893. Cunnincuam, J. T.—‘Some Points in the Anatomy of Polycheta,” ‘Quart. Journ. Mier. Sci.,’ N. 8., vol. xxviii, 1887. Davenport, C. B.—“On Urnatella gracilis,” ‘Bull. Mus. Comp. Anat., Harvard Coll.,’ vol. xxiv, 1893. DrascuE, R. von.—“ Beitrage zur Entwicklung der Polychaten,” Wien, 1884-5. Ecxstrin, K.—‘ Die Rotatorien der Umgegend von Giessen,” ‘ Zeit. f. wiss. Zool.,’ Bd. xxxix, 1883. Enters, E.—