Selyiie 5 \ “a RO ‘ AW <4 Wie Wa ena a aia i Nail eit f ih Vin ted ees Ni] 4 yea Uae eee uy 4 thet ty aie Hits itty ig | ( nh Ny wee af) ears Vt ‘| ends PAAR H LT REA AN IH ER ih Hi stati tihceting Ha NRITHT Aa | ue it beater stata SM HN ROR RT Met ite ArH } nh i i! * fn iBace I iatlan bales yh GA aR aR ER an : fy} wiltrccit vai 4 ah fut Ba Sn TU lye ty BC eg : ayaa Hise DY if kigss ouede of eat Vantvt ae et Mba | aut | Weis raver aril Saye eeeerdels ie pets AW PLES Tat , Daihen PRR yeh cai a 4 Vines at msbelit iH Wnt inthe tae: et Nabe POMPEY ala) wey aba ot SL TH OOM etal nati ) ya ies Nyy! ‘1, ‘ Abeta J ¥ 5 ; ca Soun] bese ta pated Wd petty be Hit taht eh f (eg ; tah ts eae a Th miseries erie is Lae papers nat yh aM abet 2: 4 Nadputy ad ete Bint Ory ittyseaens Silasniy tpt ‘ HY ‘ ett Da i earns tant " CH URN RAS tt ifayitiadtyjal! facia te i vith ie ribs Aad ALM i Sera iM i meeteas ate 4 My Piri Fein ¥) ; ae ; i evince iy Se eats ; 0 ae rt os PLN ans “ four vane yiilacueaea Tt a UE Er / " y 4 He MRA NIT ‘} atimerse me Ee ipeitign “tines i i , x ‘ : t ae tae ; 3 ‘i te aitalisneiyal nt whist atte he se aees Teast anes ty AZAA\AAAAIAAAAAAAA BRAARBAR ¢ ~ AARRRAAAA PARRA Ras phen t eo eeaRamnnennn~ | AAA ARRAS Aa ean NS ARARAIAAA Aaa nA an Eye AP ARAAARAAF AAP a ~~ _A AR RA AIA — ANG Arar y—) Vm! | a= |S NAA, i v—— = ~— fi Nm ne ~PS —— ARAA ie aaa AARARAA ARR AAAR AR RRAR RAR AAR pRA=C EEG RRRRRPREE FNS NSS OS om ~~ EEARARARRRARAR AAA RIEIAIA | a Wa | ARRAL® AA AARARRRERAACARARARARAAR A a nmap anna == RAR AYN AAA V7 ij I reo, i j | | | Soa raasaassnaseeaaanasaen AR AARC ARIA OAY a VAY ~~ |G’ Gin Gn GENE GID ENV ea ae aa aane ap ‘aaaaaaae ann nnnn a MeNNReAAAAAR CRRA RAR AOA AA RAR AAA | <= RRARRABARARARRAAR | A P= ry y ‘an lan| ry | : PRinaln pelea ane aaeasesnnannns Re 2 Ar, ne ee oe i) 5Sn =~ VAAN =A + : ll, i, a, Ss FY BAAR - za PLECEPECR EC E\ANAAAA AC annnnnnnnannes AARAR PARAACAAR AAAS BS BaAnar AAA mn ARAA 3 | Rearocrceneeercenc tn RRR R REAR AARAAA EC moaRAe oo | S\N ee aN AAAA AA A — INN | Aa | Ran AAA ARaAARA ae WAAR RA AAAs a aN Riaz la AAA NAA PNARARRRARARAAAAARAARA AAA ARRARAAAAAS es lo oN SAAS Aa AAAAAAAAAA : asta oanned Anan ANA —JRnSA BARARAAAA AAAS AARAARR Aa RRAaAARAe ay NAAR AAAS a35- mnnenanannanannan NAAR AAA Reena NAAR AAAAARA RRR Rt F chee santas a Tt oa e™\ —S\ \ OY \ eer AA A\ ARARARAAA on CNONGN EN ae \_\ SAAD Ama aAAARAAAAAAS ARARARAA nln \AAAAAAA mila NAA SAAR AR NANA RR AARARARR ARAARAAARARA = A AAAARR AR Raa aA Aa aR RAAARAAARARaa=- og ARARAAAARAA OAR on RAR nnn ieee Aap AAA an bac a a aaARR |p aA al ee a “AAR AAR aa im pa Rae ARANAEAAAAAAAR AAA AS RARR AAR ana i 1 el, | { ~~ s | ete . i. ‘ < : 7, n arg a | my piel LY a i 3 CHIMEROID FISHES AND THEIR DEVELOPMENT BY BASHFORD DEAN PROFESSOR OF VERTEBRATE ZOOLOGY, COLUMBIA UNIVERSITY ‘ Ga. A somion Insti. < ‘Z lgns {Sy eum @N CeLLectTion Lb . Perak $ LS Ta 4 aT ie oe err : ‘ WASHINGTON, D. ©. bv PUBLISHED BY THE CARNEGIE INSTITUTION OF WASHING:ON ~ Se CHIMA:ROID FISHES AND THEIK DEVELOPMENT BY BASHFORD DEAN PROFESSOR OF VERTEBRATE ZOOLOGY, COLUMBIA UNIVERSITY WASHINGTON, D. C. PUBLISHED BY THE CARNEGIE INSTITUTION OF WASHINGTON 1906 < CARNEGIE INSTITUTION OF WASHINGTON PUBLICATION NO. 32 FROM THE PRESS OF THE WILKENS-SHEIRY PRINTING CO, WASHINGTON, D. C. TABLE OF CONTENTS. MIEHOCUCEL OMe a ctorstelctels aioli ol oisieiers everonaiete Shers avers aces eosie Sei elake heels eatele 3 Paleontological evidence as to the position of Chimzeroids........ B Anatomical evidence as to their position... 22 6..42200s0c0s eens 4-9 PAX ONOMYAO le li vam eytOLINS eye aeic ielolotchevete a kar cfolot- = (en's eis eicleles sists, sc 6-7 Collectin sano testeersensitcreraiaeristelcioiter aris evetorteleheraierene: cicker aeiered «vais 7-10 Habits of Chimera colliei; color, size, distribution, movements... 11-18 wextaliditierences. foOdleiayeecie\sieisiets sels sisie cree elevcree eel ie evan OES, 19-21 IBLE COI Operate tevettererereietey hetoreLekoteictesst shazeiaretaieicict=) shel sictetsicteiereieieverste erect 23-25 Modeyohidepositin ce res taeteters teteresaieicisiers/sisiei avalos’ ssstoisveyeisielerousierie a 25-27 Rateof embryonic developmen tepejepe relents isieiec()ie/a1e)s s/everaversseceiere 27 Egouan dutsicapsull carrer -naccispate chet ovalerersrava nie cies overs slcrorelsterare etre sieue ooo 27-40 Generaliiplankotidevelop ment serene ciyacstcte: cielo: cfatieleletecetateus te evaiole oiesche 41 IPLiMmary, CR e me mbPLanesici ctarctelole sietele sisi sisisievovate sual edie eerieters, siete 42-45 DVO MM cpaystomeictatet eiarejercvedere ciedaleietcielslaieieré avsis1s 4]ave sles s/sieis cisoie erste © elerete 45-46 Germuinalaviesicleprmey ster cietete. eeteyste er cistersisiclctorretewarels ove sic tetetretraeree 47 ertilazation ye cca cveicies aie cisis cs eens s/s siapereaie eansieitie Dyas ee tien calls 48-52 CPIM ETILALL O Merete Merartemerereterereley stoke -wchet svereis te: chev ievetererelsvetersvenelsleveje aac ters 52-63 Gastrrlati onleiarciers ovelers svolssel srevsis, ote elerere slays sisal everedel aielever scare cictsle siete 63-92 Hat ya Ctr Dr y.OS cicyere stays isosor loss¥o/s ticle o! sieeie '9:se%s,0. sc ctor ala slay cuneisioienetas 75-101 WAC Serb ry. OS pacpepersne siateleverscteearer sy okeeve siorlewcre eres «i slctekeiay cscvere eysicieners: sts 102-109 Comparison! withvother!Chimzeroidsi icici ociciet siecle oe) ele) s)-ye)sie sts 1s. 6 IOQ-III Solaray allele Chit seta scre.ars oysta) a evskeyays anne ere creverenievere arsrersieeetaiels eevers Cir 114 Oroanopeniyye crs ions oiler crelstrnyoh Welter Acre sicieieicietctersiatiereyc stavenete Savas ve 114-132 Relationshipsiorstossil ChimeroidSecrc,tecre cielo cieieielse clei =! elais)e- 133-155 Summary of evidence bearing upon the position of Chimeroids ...155-156 Bibliograpliyaciscwcsstiele eistele eyoctciens overs cieks clerorisieisi sal giorc:siend wie as cvetsie.s 159-172 Description of plates ier iarisieceteia dleisieraisiciels\ cis eleislste s wisisieiecioler neve 174-194 ERRATA. Page 11, for Regné read Régne. Page 30, second column, after Rhinochimeera pacifica, for 3 read 33. Page 36, fig. 20, for Rhinochimzera read Harriotta (?). A day’s catch of Rat-fish, Chimera colliei, on the beach at Pacific Grove (near Monterey), California. Beside the fish are the float-lines and baskets with trawl. CHIMAROID FISHES AND THEIR DEVELOPMENT. By BASHFORD DEAN, Professor of Vertebrate Zoology, Columbia University. CHIMAEROID FISHES AND THEIR DEVELOPMENT. INTRODUCTION. Chimeroid fishes, a group representing some of the oldest and simplest of backboned animals, are considered in the present memoir with especial regard to their relationship and descent. To this end, attention has now been paid to the plan of their embryonic development, and upon this side evidence has been obtained which, whether of major or minor importance in the study of descent, has at least the interest of newness. For to the embryologist Chimzroids have until recently remained practically unknown, and they are thus the only vertebrate group of their anatomical importance—if ranked as a subclass—to have escaped investigation. On the other hand, from the standpoints of comparative anatomy and paleon- tology these shark-like fishes have received considerable notice, and they have figured in publications of the past half-century as the ‘‘most primitive vertebrates,”’ or, more precisely, as the least modified descendants of the ancestral cranium- and jaw-bearing vertebrate. And in such a role (which I now believe is only partially deserved) they have been given especial importance in problems of descent. The evidence which has been brought forward to demonstrate the primitive nature of Chimeeroids is based in part upon the findings of paleontology ; it is, moreover, as one frankly admits, supported by anatomical facts which are broad in range and which have in many instances been provided by masters in morph- ology. The substance of this evidence is that Chimeroids, although shark-like, are nevertheless widely distinct from the shark, and that they represent a lower plane in piscine evolution. As an aid to subsequent reference, the grounds for this conclusion may now be summarized. PALEONTOLOGICAL. Chimeroids are believed by some to be older than sharks. Their fossils, as Walcott maintains, occur among fragments of ‘‘fish” plates in the Ordovician (Lower Silurian) sandstones. Sharks, on the other hand, do not occur—that is, before the Upper Silurian. Probable it is that Chimeeroids lived unquestionably during the Lower Devonian and, judging from their dental plates, these forms, if Chimeroid, were highly differentiated, even at this early period. Moreover, according to the studies of Jaekel, paleozoic Chimzroids provide the evolutional stages from certain archaic armored ‘‘fishes” to the shagreened sharks. CHIMZROID FISHES AND THEIR DEVELOPMENT. RECENT. Many characteristic structures of living Chimeroids have been referred to as indicating the primitive nature of the group. The following may be cited: Dentition and dermal defenses, by Jaekel (1901), who maintains that the dental plates are primitive or “statodont,” z. ¢., the ancestral condition of the “lyodont,”’ or successional teeth of the later sharks. They have thus, if I understand Jaekel's view correctly, become greatly subdivided, so as to produce the cuspid teeth of sharks. So, too, the larger integumental plates of ancient Chimzroids are believed to have given rise to cuspid scales, and a somewhat similar view was expressed by Pollard (1891). According to Schauinsland (1902), the scales of Callo- rhynchus are of so primitive a nature as to be directly compared to those of the earliest Silurian “sharks.” Finally, Reis (1895) suggests that the curious unpaired tooth of mesozoic Chimzroids finds its homologue only in the ancient Acanthodia. Vertebral column, with delicate ring ‘‘vertebrz,” characteristic of Chimzroids, is, according to Schultze (1817), but the next stage above the notochordal con- dition of the lamprey; to Hasse (1879) it represents a polyspondylous condition ancestral to the diplospondyly of the simplest living sharks; to Gegenbaur (1901) ‘‘less differentiated’’; to Howes (1902) a purely ‘‘chordal type”; to Meyer (1886) ‘‘possibly primitive”; to Rabl (1901) a column which has ‘‘not developed centra.” Cranium and arches.—According to Cope (1870), the autostylism of Chim- zeroids is in itself primitive, in spite of the evidence of its secondary character, which has been assumed on comparative anatomical grounds from the time of Johannes Miller (1838). So, too, Kitchen Parker (1883) inclines, though doubtfully, to its primitive autostyly; and Gadow (1886) appears to have a similar view in stating that dipnoans were descended from a ‘‘simple autostylic form.” The curious labial cartilages are regarded by Howes (1891) and others as homologous with those of hag-fishes. And connected with these the /evator anguli oris, according to Reis (1896), suggests closely the condition in Acanthodian sharks. Allis (1898) also suggests that the jaw muscle (adductor) is of a primitive type (z. e., interbranchial), and in this he follows distinctly the more general conclusions of Vetter (1878), which are, indeed, in the latest time confirmed by K. Fiirbringer. The second branchial arch, it may here be mentioned, has been referred to several times (v. ¢zfra) as retaining archaic features. The labial cartilages, furthermore, are said to be primitive, inasmuch as they represent the most perfect condition of preoral gill-arches known among recent gnathostomes (K. Fiirbringer, 1903, and Schauinsland, 1903); and a presymphyseal cartilaginous element is regarded as a primitive copula between the mandibular and a premandibular arch. In fact, the entire series of copule is archaic (Gegenbaur, 1901). Ribs are absent, a primitive character, according to Goeppert (1895). Fin structures are of peculiar interest. According to Jeffrey Parker (1886), the Chimeroid is the only vertebrate to retain rudiments of a third pair of limbs. Its paired limbs furnish, according to Gegenbaur, M. Firbringer, and Braus, evidence of the origin of the paired limbs from gill-arches. In this connection Howes (1886) EVIDENCE OF PRIMITIVE CHARACTERS. 5 maintains that the paired fins of Chimera are ancestral to those of sharks and dipnoans. Rabl (1901) also refers, but in a different aspect, to the primitive nature of the fins of Chimera. By several writers the unpaired fins are regarded as primitive. The fin spine, as Reis (1896) maintains, shows the granular calcification of the mesozoic Ischyodus. The mixipterygia are ‘‘of less compound construc- tion” (Jungersen, 1898) than those of sharks. Brain, nerves, and sense organs have received considerable attention. Valentin (1842) states that in its brain Chimera is intermediate between cyclostomes and plagiostomes, and his view is shared, more or less distinctly, by Johannes Miller, Mikloucho-Macleay, Gegenbaur, Wilder, and M. Firbringer. To Burckhardt (1893) the Chimeroid brain suggests characters allied on the one hand to the primitive sharks, on the other to the lower ganoids, and according to Studnicka (1895) the forebrain is nearer the primitive form of the selachian brain than even that of Notidanid. Jaekel (1902) holds also that in Chimera, alone among fishes, there appears an epiphyseal opening in the cranial roof. In the matter of cranial nerves Cole (1896) states that ‘‘ Chimzera is unrivalled among vertebrates, first, for the ease with which its nerves may be dissected and, second, for the almost ideal results that may be attained,” as well as for the peculiarity of independent nerve roots, ‘‘archaic and perhaps primitive in type.” Similarly, Flirbringer (1897) comments upon the peculiar conditions of the nerves of the occiput. Collinge (1896) notes also the simplicity of the mucous-canal system, which, he believes, separates widely Chimeeroids and sharks. From the standpoint of the auditory organ Retzius (1884) places Chimeroids in the ancestral line of the modern elasmobranchs. Gegenbaur (1901), finally, notes that the flattened cord is primitive, like that of cyclostomes. Visceral peculiarities have also been given considerable notice. Thus Huxley (1872) refers to the ‘‘almost undeveloped gastric division of the alimentary canal, [and] the relatively small and simple heart.” Gegenbaur (1901) is inclined to regard the few turns of the spiral intestinal valve as the ancestral condition of the gut of Lepidosteus and Ceratodus. Leydig (1851), followed by Mazza and Perugia (1894), suggests that the many small brown glands of the rectum represent the ancestral condition of the digitiform appendix of sharks. Redeke (1899) maintains that in the structure of the kidney Chimeeroids are primitive, since, among other features, they retain a remarkable metamerism and have not the modified Geschlechtsniere of sharks. The foregoing are the principal lines of argument in favor of the primitive position of Chimeroids. Whether they can be maintained in the light of additional evidence, notably on the side of embryology, is a question which will be discussed in the present memoir. To summarize the problem: Are the Chimeroid fishes the least modified descendants of the primitive gnathostome? Or are they, on the contrary, degen- erate, specialized, or widely modified? Are they, in other words, close to ancestral forms which gave rise to sharks, with which they are obviously associated—or are they but modifications of the shark-like form? In spite of the formidable list 6 CHIMAZROID FISHES AND THEIR DEVELOPMENT. of citations as to their phyletic position, every investigator will admit that Chim- weroids have been but little studied—surprisingly little studied, if we consider the morphological problems which they have trenched upon And in this regard we may safely conclude that the obstacle in the way of the investigator has often been a simple one—lack of material for research. For, until recently, good material of Chimzera was relatively rare. Asa deep-water form, it was taken only by special fishermen in special localities, and even then, since it was not a food-fish, it found its way rarely to a market and still more rarely toa laboratory. This, then, has been an obvious reason why embryological material was not early described. It may finally be mentioned that fossil Chimzroids, so important to the general discussion, are rare, and, with very few exceptions, fragmentary. Recent Chimeeroids are included in 4 genera and about 25 species. An idea of their distribution and size may be had by reference to the following table: Taste A.— Avnds, Localities, and Approximate Sizes of Recent Chimeroids. Genus and species. Reference. Locality. Size (+). | | Meters. Callorhynchus callorhynchus........ | Gronovius, 1754, Mus. Ichthyol., I, p. 59, | Australia............. 0 .85 plate iv, figs. 1 and 2. Linn., as spe- | | cies, Syst. Nat., Zoophylae, ro ed., p. | 236. (Followed by Swainson, .Guiche- | _ not, and others.) ( ?=antarcticus) .....| Lacépéde, 1799, Hist. Poiss., I, p. 400, | Australia, S. America.|........ plate x11. (Followed by Swainson, | Guichenot, and others. ) (2=anistralis)) is sjev 575. Fig. 36.—Section of fertilization stage, showing deep entrance pit of a sperm. From the lowermost point arise branching rays. Fig. 37.—Detail of section of specimen shown in fig. 34. | From the path of a sperm astral rays branch in many directions, and at various points (indicated by the dark points) new centers of radiation appear. Fig. 38.—Detail of fertilization stage shown in fig. 34. The present section follows almost exactly the entrance path of a sperm. The latter appears at s, and it isseen that the entrance pit is a delicate tube extending downward in the direction of a sperm. Around the latter appears a well-marked aster, and in this neighborhood, strung along a prominent ray of the aster, are a number of deeply stained “centrosomes.” A similar “centrosome” occurs near the lowermost point of the entrance tube of the sperm. forming a plug-shaped mass twice as deep as wide (fig. 34). This is possibly the homologue of the Panderian nucleus figured in the shark egg; certain it is, how- ever, that the egg of Chimera has not as clearly a marked series of tunics in its yolk arrangement. THE EGG AND ITS MEMBRANES, 47 THE GERMINAL VESICLE. This is eccentric in eggs even as small as 3mm. (cf fig. 30). In the section of a well-grown ovarian egg shown in fig. 33 it lies close to the side of the ese.) “Ihe spireme has here contracted into a minute mass and has given rise to (about) twelve pairs of chromosomes.* These are of remarkably small size, smaller by about one-half than those of a corresponding stage of shark (Pristiurus); and they are also smaller in terms of the germinative vesicle. In Pristiuris (Rickert) the mass of chromosomes at this stage measures 36 « in width and the vesicle 296 ; in Chimzra gl Fig. 39.—Detail of sperm nucleus from section of late fertilization stage. The sperm head is surrounded by a conspicuous aster, in some of whose dendritic rays appear the nodes referred to below. The sperm nucleus itself is undergoing amitotic division. > 475. Fig. 40.—Detail of section of late fertilization stage. Throughout the germinal yolk occur asters which have no apparent reference to nuclear structures. At / many of these asters appear around a large granule of yolk. It will be seen that the rays are formed as lines in the thickened walls of alveoles. Fig. 41.—Detail of section of late fertilization stage, showing asters in germinal yolk. Fig. 42.—Detail of section of late fertilization stage. At 7 the egg-nucleus is shown surrounded by a number of asters. The asters appear to lack centrosomes and centrosphere. Note as before rays formed from rims of alveoles. Fig. 43.—Section similar to the foregoing. A sperm nucleus, however (7), is shown surrounded by asters. the same mass measures 16 and the vesicle 570. In other words, with a germinal vesicle twice the size, the size of the chromatin mass in Chimera is but one-half that of the shark. In the shark the chromatin mass measures about one-ninth the diameter of the vesicle; in Chimera, on the other hand, about one thirty-eighth. This condition indicates again the greater specialization in the egg of Chimera, The chromosomes themselves, it will be remarked (figs. 33 c, 33 D), vary considerably in length ; thus the pair shown at x are apparently longer than those at y and at 2, and a detailed examination has convinced the writer that this difference is a real one, 7, e., not due to the oblique position of the objects. This observation may be mentioned, since it affords an additional suggestion as to the individuality of the chromosomes, recently discussed, e. ¢., by Sutton, Wilson, and Moenkhaus. *Preliminary to first polar division. The number of chromosomes is clearly much smaller than in sharks (36 in Pristiurus and Torpedo). CHIMAZROID FISHES AND THEIR DEVELOPMENT. FERTILIZATION. Fertilization begins, as in sharks, in the uppermost portion of the oviduct and continues throughout the period of the formation of the capsule.* The earliest stage in the writer's material was obtained from a capsule like that shown in plate m1, fig. 13, earlier stages not having been handled successfully. Late stages were secured 454 458° 46 Fig. 44.—Detail of late fertilization stage, showing male nucleus in the process of approaching the egg nucleus. A club-shaped centrosome surrounded by astral rays appears at (right) side of nuclear membrane. >< 475. Fig. 45.—Detail of section similar to the last. A well-marked vacuole (artifact ?) appears at one end of nucleus. Fig. 45A.—Sperm nucleus with aster from section similar to the preceding. Fig 45B.—Sperm nucleus from stage similar to foregoing. The center of aster is to be found in the section below present one. Fig. 46.—Detail of section shown in fig. 34. Sperm nucleus has divided amitotically. This at first suggests a stage of copulating pronuclei. Fig. 46A.—Egg and sperm nuclei in apposition. Rays not conspicuous. Fig. 47.—Early prophase of segmentation nucleus. Two asters are present, one of which (the right) contains two centrosomes. Fig. 47A.—Early segmentation stage. Section passing through segmentation nuclei. No surface furrows are as yet present. >< 190. from capsules about as shown in plate 1, fig. 16. In the present account the stages may conveniently be referred to as early, middle, and late. An early stage is shown in surface view in plate rv, fig. 18, magnified about 15 diameters. This was drawn from a living egg and shows the germinal area somewhat misshapen, due to rupture of the vitelline membrane.+ The germinal area is not sharply outlined ; it is the same color as the remainder of the egg, and * The egg at deposition is undergoing the first stage of segmentation. + This is conspicuous at this stage, glossy and tense. FERTILIZATION STAGES. 49 is only demarked by a slight furrow. Under a dissecting lens a number of minute depressions indicate the points of entrance of sperms. Seven of such points appear in the present instance, and all of them are peripheral; four are close together. In this case sections show that no sperms have entered the middle of the germ. A middle stage in fertilization (plate tv, fig. 19), also examined in the living egg, showed 23 entrance pits. Of these half a dozen are of large diameter and several are minute, a condition which, in comparison with the preceding stage, suggests that the small pits are the early phases of the large ones, and we query, accord- ingly, whether in point of time the entrance of sperms in Chimzera may not prove an extended process (v. zzfra, heading /). In the present specimen it will be seen that the sperms have entered not only the germinal substance but the bottom and even the outer wall of the germinal fosse. Study of sections leads us to conclude: (a) That the tail of the spermatozoon does not enter the egg. In fig. 35 a sperm is shown which has just entered the egg; the middle piece, mf, ends abruptly, and there is no trace of the tail. The entrance pit is not yet sharply formed. (6) That the head of the spermatozoon rotates as it travels inward. Even at the early period above figured, the filamentous character of the sperm head has been lost; it is now spheroidal, surrounded by a light-colored area of the germ. Although hardly within the egg, its axis inclines 45° to the surface, and its middle piece is parallel with the surface, a condition which by analogy with other forms leads us to conclude that it has already begun a process of rotation. Ina later stage in the entrance of the sperm (fig. 38) the lighter-colored portion of the ‘‘ head” points toward the surface of the germ and thus indicates that the rotation has been carried through an angle of 180°. (c) A state of remarkable kinetic activity exists in these stages. In fig. 36 a series of ‘‘astral rays’’ are seen diverging downward from the entrance pit of a spermatozoon (¢f. the observations of Miss Foote in Allolobophora). And from paths traversed by asperm “‘astral rays” arise, sometimes radiating regularly, but usually branching irregularly and forming new groups of radiation. At such points of reradiation darkly staining bodies occasionally appear which remind one of centro- somes. In the present fig. 34 branching astral rays are seen. These, it is found, have arisen around a sperm path. A similar series greatly enlarged is shown in fig. 37, a series of considerable interest, since it shows many ‘‘centrosomes”’ surrounded by bending and irregularly branching rays. The ‘‘centrosomes”’ some- times appear at centers of reradiating rays in sperm asters (figs. 38, 39); at other times they arise without any apparent relation to sperm asters or sperm paths, as around an unusually large yolk granule (fig. 40, the group at the right). As shown in the last figure, more than half a dozen centers of radiation appear around the yolk granule. On the other hand, the two large ‘‘asters’’ shown at the left in the present figure have no apparent relation with the former series, nor are they in the 50 CHIMAEROID FISHES AND THEIR DEVELOPMENT. neighborhood of sperm asters. A similar pair of ‘‘asters’” are shown in fig. 41. On the other hand, the asters shown in fig. 42 are arranged around the male pronucleus, but how they are related to one another can not safely be inferred. In the following section (fig. 43), drawn from the same specimen, a similar radiation occurs around a supplemental sperm head, z. In the four preceding cases it is interesting to observe how perfectly the rays fulfil the alveolar conditions for aster formation as explained by Biitschli. Note in this connection the large size of the alveoli in the immediate neighborhood of the aster. (¢d) The behavior of the germ nuclei in fertilization is similar to that in shark. The sperm which enters the germ in the region nearest to the egg nucleus is the one which accomplishes fertilization; it undergoes the customary form changes while traveling through the germ. In the stage shown in fig. 44 its chromatic material is becoming resolved, and the aster which appears beside it radiates from a centrosome, which is in this case somewhat elongated, situated close to the nuclear membrane. A stage somewhat earlier than the foregoing is shown in fig. 45; this, however, represents a stage in the development of a supplemental sperm head. The foregoing figures are taken largely from late stages in fertilization. A stage from a nearly finished capsule (fig. 46) pictures the union of the germ nuclei, 7. é., corresponding to Riickert’s fifth stage in the fertilization of the ray (Torpedo), as figured in the Kupffer Festschrift (fig. 53 B). On the other hand, fig. 46 a, which at first sight suggests copulating pronuclei, must be construed as picturing a (sperm) merocyte dividing amitotically; for here a third nucleus is found to be pres- ent, above the niveau of the other two. The figure indicates, further, the retention of the aster and an extensive pale-colored area surrounding the nuclei. (ec) The behavior of the supplemental sperm heads is also notably shark-like. In even the middle stage of fertilization they can not readily be distinguished from the early sperm nucleus. Indeed, the nearer they are in a position to the egg nucleus the more difficult they become to distinguish from one another. And conversely those undergo the least conspicuous changes which occur in the margins of the gverm. We have already referred, in fig. 45, to a structure which from its position is apparently the early sperm nucleus. In this phase, at the margin of the nucleus is a vesicular area, at one end of which an aster radiates froma minute centrosome. A somewhat similar appearance occurs in what, from its eccentric position, is undoubtedly a supplemental sperm head (fig. 45 A). Here the vesicular area of the nucleus is less perfectly developed, strands of karyoplasm passing from the nuclear membrane to the large and deeply staining mass of chromatin, a stage, indeed, which may be looked upon as the earlier condition of that of fig. 45. Another sperm head (fig. 45 8) from the same series of sections is intermediate between those of figs. 45 and 45 a. The vacuolated margin is now broken into several discrete areas, and the chromatin is collected into a diffuse mass, irregular in outline.* From this stage the transition is not wide to that of fig. 34, in *The aster lies below the plane of the section. FERTILIZATION STAGES. 51 which is pictured a (sperm) merocyte occurring eccentrically (z) in the germinal area of an egg twin to the preceding. In this the vacuolar area has been practically lost, the aster increased in size, and the nucleus subdivided into a number, probably five or six, of smaller merocytes. Division of this kind has been observed in many instances; and on the other hand no case has been found in which a sperm nucleus divides indirectly. This condition is noteworthy, since it emphasizes on still another line the specialization of the Chimeroid. For in the shark the sperm nuclei may undergo indirect division throughout practically the entire process of cleavage; and when early direct divisions do appear, ¢e. g., in the third cleavage (Riickert in Torpedo), they still show traces of their mitotic ancestry. Indeed, the nearest condition to the presegmentation division of the sperm head in Chimera (fig. 39) occurs in shark only in the period of later segmentation (c/. Riickert, of. cit., pl. m1, fig. 18) In other words, the morphological (or the physiological) result which in the shark is effected only at the end of a series of graduated stages is accomplished by Chimera at a single stroke—a condition worthy of comment, since it affords a palpable case of ‘‘precocious segregation.’ (/) The sperms enter the germ not simultaneously, but during a relatively extended period. The pits formed by the sperms when entering the germ, as already noted, are different in size, and we accordingly infer that, as the sperms themselves do not differ materially in size, nor in all probability in individual activity, the dif- ference in the pits is due to their having been formed at successive periods. This suggestion is borne out by examination of sections. Thus, in fig. 36 an entrance pit is shown, pointing down in the direction of, but not actually connected with, a sperm head lying deep in the germ. And here the pit or funnel has a wide mouth. On the other hand, in fig. 38, a funnel is pictured whose apex is still connected with a sperm head, and its mouth is narrow. The sperm head, in this case, lies in a shallower layer of the germ, and from its structure, also, is clearly a younger stage in development. It follows, therefore, that the former sperm entered the germ at an earlier period than the latter, and that the process of semina- tion is a relatively extended one—relatively, since in sharks all sperms appear to enter simultaneously. The suggestion may, on the other hand, be made that the difference in the behavior of the sperms in the germ might be due rather to their location than to their time of entrance; or, in other words, that the rapidity of their development might be influenced by their proximity to the egg nucleus. This suggestion, however, is not tenable in view of the condition of the fertilization stage (middle stage) shown in plate rv, fig. 19, for here small pits occur side by side with large ones, both in the middle of the germinal area and on the sides. Finally, to contrast Chimzera and shark in stages of fertilization: In Chi- mera the entrance of the sperms is a protracted process; but as soon as the sperms (other of course than the one which fertilizes the egg) enter the germ they divide promptly by amitosis, with the very probable result of producing a greater number 52 CHIMAEROID FISHES AND THEIR DEVELOPMENT. of merocytes in a shorter time.* It thus appears that the early divided merocytes are equivalent morphologically, and probably, therefore, physiologically, to late merocytes in sharks. Furthermore, at the time of fertilization the number of sperms entering the egg of Chimzra appears to be greater than in sharks. The average number reckoned by Rickert in Pristiurus is about 16; in Chimera it is at least 24, judging from the number of entrance funnels in late stages of fertilization. Again, in Chimera the sperms form and long retain definite paths in the germ which are unknown in other forms; so also are their entrance funnels more conspicuous. The sperm nuclei, furthermore, as well as the segmentation nuclei, are the more special in terms of the entire egg, since they are smaller than in sharks (in diameter about one-half). Also, as evidence of specialization—for these structures have clearly a special physiological value—witness the great number of asters and their ready mode of appearance, ¢. g., around vagrant yolk granules of extraordinary size. In point of histological differentiation of the germinal cytoplasm, finally, we observe in Chimera conditions unparalleled in the shark. We recall here the differ- entiation of typical Schaumflasma, the light areas surrounding the nuclei, and the extensive development of astral rays. SEGMENTATION. Drawings of the living germ, plate tv, figs. 20-29, give a general idea of the process of segmentation. And in surface view this resembles distinctly the usual conditions in shark. There is the same type of germinal area in which cleavage lines appear, anda marginal zone which apparently circumscribes the area of cleavage. In the first of these figures the germinal area appears convex, although some- what flattened above, and is separated from the surrounding germinal yolk by a narrow fosse. The furrow which appears to traverse it is sharpest and deepest in the middle of the germ and fades away at the margins. The surrounding zone of germinal yolk, however, extends widely over the surface of the egg and lacks a sharply marked outer boundary line. Its inner boundary, 7. ¢., at the fosse, shows a number of small eminences’ These, as sections also show, correspond to the eminences formed in the shark blastoderm by peripheral (sperm) merocytes; they are more numerous at the corresponding stage and are more regularly disposed around the germ. In further detail: Sections of the present specimen demonstrate that in spite of the single apparent furrow the present stage represents not the first, but the third cleavage, two cleavage furrows having been retarded (? suppressed), for there are found to be present six segmentation nuclei. This condition, it may be remarked, occurs in certain specialized sharks (Torpedo). The first division of the segmentation nuclei is accompanied by no trace of a surface furrow. Such a stage is shown in fig. 47 a. The nuclei are here somewhat widely separated from one another and are in the resting stage; the only indication *Recent examinations of the sections of the fertilization stage which yielded fig. 46 a (Chimera Emb. 12, in my cabinet) show that no less than 88 (sperm) merocytes are present. Thus in Chimera as many merocytes are present in a presegmentation stage as in Torpedo (Riickert) in a stage of fourth cleavage. SEGMENTATION STAGES. 53 of cleavage is in the arrangement of the germinal cytoplasm around the nuclei. This is expressed in such a way that the germinal yolk rises between them like a wall. The second stage in which cleavage is seen at the surface is shown in plate rv, fig. 21, from the egg companion to the one shown in plate tv, fig. 20, but incubated longer (about forty minutes). Here a second furrow is noticeable. The resulting ‘‘ blastomeres” are unequal in size, one of them being as large as two of Fig. 48.—Section passing between the point | and | of the segmentation stage shown in PI. IV, fig. 20. In the present section the line which appeared to indicate first cleavage lies below the point 7; below this a vacuole is present whose lateral extent gives one the impression of the width of the furrow noted in surface view. >< 35. Fig. 49.—Section through a segmentation stage corresponding to PI. IV, fig. 22. It will be observed that some of the cleavage lines do not open to the surface, as at c. On the other hand, one of the spaces between the blastomeres opens into a fissure-like vacuole, a. Fig. 50.—Similar section of early segmentation stage, in which, as at 7, a nucleus appears without any neighboring cleavage furrow. A line of thicker germinal yolk appears in its place. Fig. 51.—Section of early cleavage stage in which, when viewed from the surface, deep, fissure-like vacuoles appear as cleavage lines. the others. As in the companion stage the furrows fade away at the margin of the germinal area, and this is again surrounded by a somewhat regular ring of merocyte eminences. It may be mentioned that these characters are materially modified, 7. e., as far as surface view is concerned, when the egg is hardened, e. g., in acetic sublimate. And in sections it is found that the circumgerminal fosse and merocyte eminences disappear and what was interpreted as surface furrows in the living egg appear as long and wide vacuoles. Thus in fig. 48, a section transverse to the 4 CHIMAROID FISHES AND THEIR DEVELOPMENT. first cleavage line of the stage of plate rv, fig. 20, the fosse will be seen to have dis- appeared, and the line which indicated cleavage, so conspicuous in surface view, now appears under the point / as a vacuole triangular in section, its apex touching the surface of a germ.* Vacuoles, we note, are abundant in early stages; several are present in the section of the germ just referred to, and from their arrange- ment they suggest the division of the germ into blastomeres. Three stages of early cleavage are shown in surface view in plate tv, figs. 22, 23, and 24, the first as an opaque object, drawn from the living egg, the second and third as translucent objects, in the last cases the germ having been removed, placed in a watch glass, and examined by transmitted light during the process of fixation. In these three preparations there is considerable irregularity in the surface charac- ters; in the first the margins of the blastomeres are rounded, in the others angular; outwardly they appear to represent fourth and fifth cleavages; in section, how- ever, single ‘‘ blastomeres ’"’ are sometimes found to contain several segmentation nuclei. It was observed that the resting and dividing nuclei were sometimes found in the same section, and it follows accordingly that in Chimera the synchrony of cleavage is early lost. Four later stages of segmentation appear in plate ty, figs. 25, 26, 27, and 28, all drawn under conditions of transmitted light, the living specimens having been removed and examined in watch glasses. In the first of these the germ is well marked off from its circumgerminal zone; in the rest some of the marginal cleav- age lines were traced half-way across the circumgerminal zone, and in a few instances these lines could be followed quite across it. In these stages continued subdivision of the ‘‘ blastomeres” has taken place, those in the central position becoming divided oftener than those near the periphery. As in earlier stages, some of the cleavage lines are probably not expressed at the surface, and are due only to vacuoles; the latter are lineal in surface view, sometimes wide, sometimes narrow, occasionally almost attaining the surface, at other times lying fairly deep in the germ. (Cf figs. 49, 50,51.) Sometimes, as in fig. 49, they are actually continuous with cleavage furrows, as at a, and considering the relation which they often bear to nuclei (¢. 2., 2zfra, under the heading ‘‘gastrulation”), we conclude that in some cases—even, indeed, in many cases—they are homologous to cleavage spaces, z. e., that they are cleavage spaces which fail to become expressed at the surface of the germ. This conception appears to be applicable even when the vacuoles appear in the peripheral region of the germ in fertilization stages. Thus in fig. 34 the masses of germinal yolk separated by the vacuoles (under the points marked with an asterisk [*]) usually bear sperm nuclei which, as we know by analogy, will cause ‘‘segmentation.” - Accordingly, even in this position vacuoles may be compared to intercellular spaces, at least from the standpoint of developmental mechanics. In fig. 52 a section of a segmentation stage corresponding to plate tv, fig. 26, shows that cleavage has by this time extended deep into the germinal area. Hori- zontal divisions have occurred, irregularly however, for in some places the blastoderm On *There is thus a possibility of there having been an open furrow in the living egg. CLEAVAGE STAGES. on on varies in thickness from one to five cells. Noteworthy is the irregularity of the yolk wall out of which blastomeres are segmenting, as at x, and into which deep inter- cellular spaces are continued, becoming confluent below, as at v, v, v, with vacu- oles like those described in earlier stages. It may be remarked that in this stage the vacuoles pass deeply into the yolk. Another stage of late segmentation (or blastula), corresponding in general with plate rv, fig. 27, is shown in the series of sections, figs. 53-56. In the section, fig. 53, which passes near the center of the germ, the conditions differ little from the preceding stage. We observe that continued divisions have taken place and that there is still the same outcropping of blastomeres from the yolk wall, as at a and x, following mitoses. An advancing character in this stage is the general flattening of the germinal wall, as at the point y, a preliminary step toward the formation of the floor of the segmentation cavity, and possibly indicating fore and aft differentiation of the germ. It may be remarked that this is the first stage in whicha conspicuous zone of merocytes was seen. These are numerous under the central blastomeres, most numerous under the peripheral blastomeres, and then rapidly decrease in Fig. 52.—Section of late cleavage stage (corresponding to Pl. IV, fig. 26). x, Mass of germinal yolk from which a blastomere is being budded out; z’, vacuoles which are continuous with intercellular spaces. number peripherad. The three sections, figs. 54-56, illustrate suchaseries. The first of them, fig. 54, indicates the relation of the above-mentioned vacuoles to intercellular spaces, as at the points marked with an asterisk (*); the second and third, figs. 55, 56, are instructive as showing the extension of a nest of cells, x (it is the same group in both sections—it appears, however, at the left in the lower section, since this has been turned over on the slide), beyond the margin of the circular mass of blastomeres—instructive, since it suggests that the outlying region of the germ (circumgerminal zone) is still little different from the germ itself in its cell-forming nature. It is also to be observed that deep fissures representing intercellular spaces (fig. 56, 2), extend peripherad through the germinal yolk, corresponding to the marginal furrows described in the eggs of ganoids and dipnoi, of Heterodontus, even of amphibia. There is here accord- ingly aregion in which, side by side, occur small blastomeres, large yolk masses (bearing nuclei), and undivided yolk; there is no gradual transition from the yolk to the large blastomeres and from these in turn to the small ones, which, as we have seen, correspond in size with blastomeres of the center of the germinal area. We observe, furthermore, that the small blastomeres arise in any 56 CHIMAEROID FISHES AND THEIR DEVELOPMEN neighboring position, as in figure 55, at a, 6, or ¢, budding out directly from the yolk. In such cases the important question remains whether the nuclei which pass into the blastomeres are derived from segmentation nuclei or from Figs. 53-56.—Sections of late segmentation or early blastula stage (corresponding to Pl. IV, fig. 27). The first section traverses the mid-region of the segmented area, the others progress marginalward. x, Blastomeres which have recently budded out of the yolk ; yy, wall of sub-germinal yolk; 2, cleavage line passing deep into the sub-germinal yolk ; *, vacuoles which come to the surface of the germ and form intercellular spaces. > 35. sperm-nuclei. In the latter event, judging by analogy, they would show amitoses only; in the former they should multiply by mitosis. Examined in this light it is found that the nuclei which here pass into the blastomeres show amitoses, and THE DIVISION OF THE YOLK. 57 they might thus be regarded as of sperm-head origin. On the other hand, it might still be claimed that the nuclei of such blastomeres were derived from the segmentation nuclei, fora more careful examination shows (1) that amitoses exist in the blastomeres in the central region of the germ, and (2) that no mitoses are found in the zone of the merocytes, where we may reasonably expect that some nuclei are present which are derived from segmentation nuclei. The problem is, nevertheless, a difficult one, and hardly to be answered in the present outline of Chimeroid development. We point out, however, that two criteria which have been given a prominent place in the discussions of shark development can not be employed in the present instance, viz. (1) the number of the chromosomes which would naturally give a clue as to the origin of the nuclei can not be estimated in the merocytes, since they are here undergoing only amitotic division; (2) the size of the present merocytes can not prove an important element for comparison, since they range from minute to large, and in shape from spherical to greatly elongated and irregular. An important phase of the cleavage in Chimera has naturally been introduced by the foregoing discussion, 7. ¢., as to the segmentation of the egg in its extra- germinal region. We have seen that as segmentation progresses nuclei (whether segmentation or sperm-merocyte) spread peripherad. Their presence can be determined in sections ; and in surface view, in the later stages, e. 2, plate tv, fig. 28, cleavage lines can be seen passing outward in the region of the circum- germinal zone. ‘That these lines are actual furrows is shown in such a section as that of fig. 56 (at the right). Such marginal furrows, however, are usually minute in size, and are often, in surface view at least, difficult to follow, a difficulty which may be due either to the blending of these delicate lines with the color of the circumgerminal zone, or to the partial or total confluence of the adjacent rims of the cleavage furrows, ¢. ¢., as in the marginal blastomeres of Cryptobranchus (according to Ishikawa). The distinctness of these lines, however, increases after the circumgerminal zone is passed, and they later give rise to what we must regard as the most remarkable feature of the segmenting egg. To follow this process: In fig. 57 a late stage of segmentation is shown; the germ is at g; nearby are nests of blastomeres (c/. fig. 56, 2), two of which are of such size as to appear in the figure, as at g’; from the germ radiate furrows, two of which, a and a’, have become conspicuous at the periphery of the egg, where they may have merged with similar furrows, or indeed, on the other hand, there is a possibility that the long furrow passing between the points marked with an asterisk (*) may be the deflected continuation of the lines a and a’. In figs. 58 and 59 furrows are seen arising from or near the circumgerminal zone, and examination shows they deepen as they proceed peripherad. In the egg shown in fig. 59 the outgoing line subdivides and marks out superficially a narrow segment of yolk. In the same egg, but in the anterior region, we note another marginal line, 6. Examined from below this stage is of considerable interest 58 CHIMAROID FISHES AND THEIR DEVELOPMENT. (fix. 60), for it shows that while the cleavage lines have failed to express themselves on the dorsal side of the egg, they yet appear conspicuously on the vegetal side. Thus the line in fig. 59 at a passes sharply inward (fig. 60), subdividing into a series of 59 60 b & up d Figs. 57-61.—Eggs showing progressive cleavage of the yolk mass. In a blastula (57) a conspicuous fissure is noted between the points * and *. cleavage lines which in turn merge with other cleavage lines (d—g) passing downward and inward from the equatorial zone of the egg, In some cases well-marked yolk masses are outlined, as at the point marked with an asterisk (*), suggesting large yolk-filled blasto- meres on the ventral wall of an amphibian blastula. It may be remarked that the lines here described are not mere surface mark- ings, for during the process of hardening an egg, e. g., in acetic- sublimate, one may separate the yolk masses by aid of dissecting needles, and in this process it becomes clear that the lines are in reality fissures dipping deeply into the substance of the egg. Indeed, in the former specimen it was found that the mass marked with an asterisk (*) could be removed ex é/oc from the remaining mass of yolk. It is evident, accordingly, that in this stage the egg is being divided up on its ventral side into a number of large yolk masses; that these masses stand in relation to the entire egg very much as do, é. g., in the frog’s egg, the blastomeres of the lower pole to this entire holoblastic eee; further, that the fissures which accomplish this result, like cleavage lines on the vegetal side of the holoblastic egg, are interconnected with a series (a—g, in fig. 60) of cleavage lines which pass downward and inward from different points in the equatorial region of the egg. Between the stages shown in figs. 59 and 60, and those in figs. 61 and 27, which are older by about nine days, observations are lacking. It is nevertheless clear, by comparison of these stages, that the yolk masses shown in fig. 60 have separated from one another widely as the fissures deepened, and that, as the masses became more distinct, their condition of surface tension—in view always of the syrupy consistency of the egg—caused them to round out their contours to the degree shown in fig. 61. THE DIVISION OF THE YOLK. 59 In dorsal view the latter stage shows few large yolk masses, and these are distinct from one another, although closely pressed together. The yolk masses, it may be mentioned, were removed separately (in the living condition) without causing their rupture, and it was then seen that the fragmentation of the yolk mass had progressed further than was at first evident, for, lying below and on either side, against the ventral wall of the capsule, were many small masses of yolk (¢/. fig. 61, at the right, and fig. 27, somewhat behind the embryo), their contours rounded out for the most part; but a thick, creamy or syrupy fluid in which they lay made it further evident that in some cases the yolk masses had broken down. This fluid, we remark, was observed in specimens of this stage only when the larger masses were separated; but if sea-water was injected (by pipette) between the larger masses as they lay in the open capsule, it would dissolve the underlying creamy yolk and the entire contents of the capsule would become hidden from sight in the resulting milky fluid. One might conveniently digress at this point to follow the fate of the yolk masses above mentioned. The blastoderm appropriates only a small portion (which has been estimated as about one-tenth of the volume) of the entire egg. This separate yolk mass is shown slightly shaded in figs. 27 and 61, and the blasto- derm, with its attached embryo, has as yet inclosed only a small portion of it. A similar stage is figured in plate vin, fig. 47, and a somewhat later one in plate vim, fig. 48. In the latter the blastoderm is seen to have almost inclosed the yolk. It completely incloses the yolk and forms a diminutive yolk sac in the embryo shown in plate vit, fig. 49, and a similar condition occurred in the embryo of plate 1x, fig. 50. These features are dwelt upon in order to show that the behavior of the blastoderm in appropriating but a portion of the yolk is a normal phenomenon. And I note that the condition shown in fig. 61 has been observed on three occasions by myself, and that similar conditions were recorded by Dr. Wilbur.* In supplementary evidence upon this point we may again refer to the embryo of plate vim, fig. 49, for in this the yolk sac, although of miniature size, is evidently normal, since it exhibits a well-developed vitelline circulation. ¢ The yolk-masses other than that appropriated by the blastoderm undergo con tinued subdivision. This is in progress in fig. 61, where the large yolk mass shown in the lower part of the figure is being divided into three smaller ones. We have already referred to the pasty fluid present among the lowermost yolk masses in this stage. In the egg capsule from which the embryo of plate vim, fig. 49, was taken, no extra embryonic yolk masses were found, but the egg capsule contained a fluid so cream-like as to conceal completely the embryo and lead me to infer that the egg was addled, almost causing me to throw away this valuable stage. We can only conclude, therefore, that the creamy fluid was due to the continued breaking down ‘ *In his early letters Dr. Wilbur referred to these conditions doubtfully; he was then represented normal appearances."’ +Theyolk sac measured about half an inch in length and a quarter of an inch in breadth; accordingly at this stage of development it represents but about one-tenth the volume of the egg of an Elasmobranch of similar size ‘not sure whether they (e. g., Spinax niger). 6o CHIMAROID FISHES AND THEIR DEVELOPMENT. of the yolk masses of the earlier stage.* That this fluid was nutritive to the embryo was also evident, since the external gills were dilated at various points with brilliantly colored blood knots, and in these, as I later found, numerous erythrocytes were undergoing division. And this condition in the gill filaments is the more clearly correlated with the presence of the milky fluid, since in similar eve capsules (sharks and rays), where this milky fluid mass is lacking, blood knots on the external gills are also absent. One infers, moreover, that the milky fluid, which from its included yolk is highly nutritive, may also be passed as food into the mouth of the embryo and assimilated in the gut. But to this I will refer at a later point. The entire process of the fragmentation of the egg of Chimera, it will be seen, is worthy of especial comment. Unlike the eggs of other vertebrates, and unlike, indeed, those of invertebrates, unless we include a somewhat generic resemblance in certain mollusks (¢. ¢., Neritina, Blochmann, 1887) and in certain digenetic trematodes, the present egg follows in its development two distinct paths, 7. ¢., a small portion of the egg develops in the direction of producing the embryo with its complete though diminutive yolk sac; the remaining portion, about nine-tenths of the bulk of the egg, proceeds to undergo a process of repeated fragmentation to the end that it may be appropriated by the embryo secondarily. To account on phyletic grounds for this extraordinary and ‘‘unnatural”’ plan of development, one must, I believe, start with the premise that the fragmentation of the egg is a process comparable with total cleavage. This premise we may accept on the following evidence: (1) The fragmentation, like cleavage, is progressive. (2) Although the cleavage lines have never been followed conclusively from the rim of the blastoderm into the deep fissures which initiate the fragmentation, they have at least been observed in late stages of segmentation to pass out over the circumgerminal zone in the direction of the peripheral fissures (c/. in this regard the evidence of Heterodontus).t (3) The yolk masses give evidence of being nucleated. There is in the first place evidence that the nuclei travel peripherad. In the stages of plate tv, figs. 25-27, nuclei are found to have occupied the circumgerminal zone, 7. ¢., they have traveled outward a distance equal to about three-quarters of the diameter of the blastoderm. In an early gastrula, furthermore (plate v, fig. 31), and in section, fig. 63, they have proceeded outward a distance equal to twice the diameter of the blastoderm. Now, on the evidence of progressive centrifugal movement of the *The reader may reasonably query at this point how it happens that the creamy nutritive material is not washed out through the openings of the capsule during the respiration of the young. This result has, I take it, been avoided in the course of the evolution of this process in two ways: (1) By retarding the appearance and growth of the capsular openings until the nutritive material is partly consumed ; (u) by the great density of the creamy fluid, for if the nutritive fluid be heavy (and experiments with the living eggs have convinced me that this is the fact), a moderate current of sea-water could be passed over it without causing it to be washed away. fAnnot. Zool. Jap., tgo1, vol. rv, pt. 1, pp. I-7- THE REDUCED SIZE OF THE YOLK SAC. 61 nuclei, and as this nucleated area (in diameter) is measurably greater than that of the yolk mass which the blastoderm comes to inclose, it follows that nuclei are present in some of the outlying yolk masses. Of this, however, we must none the less admit that no direct proof is at hand, since no sections of these outlying yolk masses were made. In this connection I observe that if the embryo-bearing yolk mass be examined even under a low power (plate vit, fig. 48a) one obtains a fairly con- vincing picture of its holoblastic character. (4) The foregoing evidence is none the less strong if, conversely, we consider that on no other morphological ground, save that of cleavage, using the word ina broad sense, can this progressive and normal fragmentation be explained. Accepting, then, the premise that these divergent paths in the development of the egg of Chimera took their origin in a holoblastic egg, the present con- ditions may well have been developed on somewhat the following lines: In the primitive Chimeroid the egg resembled that of Cestraciont; it was probably, however, not as large as that of the recent Heterodontus, but its cleavage fissures were deeper and more numerous. The embryo at that stage had the usual external gills of the selachian. The next stage would be attained when the eill filaments, passing beyond the stage of the well-known trophonemata, came to appropriate the white of the egg which was contained in the deep cleavage fissures, a process which in time caused or accompanied (a) the deepening of the fissures, and in further time (4) a rupture at the bottom of the fissures. Through such a process yolk material came to escape and mingle with the albuminous contents of the deep fissures. Such a process, we may now assume, was naturally followed by adaptative changes in the trophonemata, which in the end accelerated the growth and differentiation of the embryo. In short, at this evolutional stage the embryo was receiving through a (morphologically) indirect channel an amount of nutri- ment which rivaled that derived from the vitelline circulation. ‘The result was what one would have anticipated, z. e., the down growth of the vascular blastoderm was retarded, while the fissuring of the yolk-mass became deepened and _ the trophonemata further modified. The line of evolution thus carried on in the egg will be seen to involve the fate of the yolk sac, viz., in determining how great an amount of the yolk could be diverted from it. In the present species (C. col/ie7) about nine-tenths of the egg has been diverted, while in the Callorhynchids, where the yolk sac is known to be larger, possibly not more than half. In the foregoing process it is suggested that the first steps in the disinte- gration of the yolk mass were found in cleavage phenomena. It should, however, be admitted that the cleavage may not have been equivalent to that of the usual holoblastic type. The nuclei which spread peripherad may have been sperm-nuclei; and in this event the peripheral furrows are special phenomena, unconnected, possibly, in phylogeny with the cleavage lines in the holoblastic egg. Certainly in favor of such an interpretation is the fate of the disintegrating yolk masses, since such a fate is paralleled somewhat by the sperm-nuclei in the shark egg. It 62 CHIMAROID FISHES AND THEIR DEVELOPMENT. is opposed, on the other hand, by the conditions in the egg of Cestracion, where the peripheral furrows, similar in general regards, are known to be continuous with those of true cleavage. The question, therefore, can not be answered finally until evidence is forthcoming to distinguish the kinds of nuclei present in the extra- embryonic yolk masses. Meanwhile, judging at least from the behavior of the nuclei in the circumgerminal ring, I think it is not at all improbable (c/ Gastru- lation) that in these masses both sperm and segmentation nuclei are present. Returning again to the development proceeding at the animal pole of the egg: We recall that in the sections figs. 53-56 there was shown a stage of late seg- mentation, or an early blastula, such, for example, as pictured in plate tv, fig. 27. In a slightly later stage (plate tv, fig. 28) an increased number of blastomeres are present, and there is still an indefinite condition in the periphery of the germ, blastomeres being continued irregularly over the ring-like circumgerminal zone. On the other hand, in plate rv, fig. 29, a stage is figured earlier than the preceding, Fig. 62.—Section of blastula. sc, Segmentation cavity. but showing a well-marked line of demarcation between the blastomeres and the cir- cumgerminal zone. It seems evident, accordingly, from this and similar instances, that considerable variation occurs as to the time at which the marginal relations of the germ are established. Thus in the stage first referred to (figs. 53-56) the circumgerminal zone was traversed by radial fissures and invaded by nests of cells; in asimilar stage (plate tv, fig. 29, sectioned in fig. 62) the same region is solid and yolk-filled, forming a compact border to the germ. In contrasting these two stages one observes that, while they can differ little in point of age, judging from the number of blastomeres in the cross section of the middle of the germ, they yet have marked differences in their relation to the yolk; the former has around it and under it ‘‘fine yolk” (Riickert); the latter has its fine yolk contracted into a thick mass lying immediately below the germ, a condition which may be the immediate cause of the failure of marginal blastomeres to express themselves in a peripheral direction. We observe that in fig. 62 the fine yolk is pervaded with vacuoles which, from their shape and relations, are evidently equiv- alent to inter-blastomeral spaces, a conclusion which is supported both by the nucleated character of the masses of fine yolk thus outlined and by the continua- tion of the inter-blastomeral spaces with the distal ends of the vacuoles. The fine yolk, in short, is already coming to be formed into blastomeres, and it is interesting to note that a blastomere, which is found on the boundary line between the fine GASTRULATION. 63 and coarse yolk is composed half of fine and half of coarse yolk. It is quite prob- able, therefore, even from this single observation (cf also zzfra), that the region of the coarse yolk is not as inert as one is at first inclined to believe, an induction which suggests at once that the fewer and larger fissure-like vacuoles in this region are equivalent to the vacuoles of the fine yolk, or in other words, to intercellular spaces. A final point of contrast between the foregoing stages: In the former the blastomeres are relatively compact; in the latter there is a general inter-blastomeral space which marks an early state of the definite cleavage cavity. It is probable, as noted for the former stage, that the anterior end of the germ can now be distinguished. GASTRULATION. The stage shown in surface view in plate v, fig. 30, and in sagittal section in fig. 63, is probably the most valuable of the author’s early Chimeroid embryos. For it may be accepted as providing a key to the problem of gastrulation not only in this form but in sharks as well. Its discovery is none the less a fortunate one, since it is a stage which has every appearance of being brief, and therefore easily overlooked. In diameter it differs little from the blastula above described (fig. 62), but its depth is notably greater. Comparing these two stages, we conclude that the deep subgerminal region of the earlier stage (fig. 62), which was traversed by vacuoles, has been replaced by the deep-lying mass of cells of fig. 63. We observe that this thickening of the cellular mass has not yet been accompanied by an extension over the surrounding region; the mass is at present compact, sub- spherical, lying in a smooth depression of the germinal wall. At one end of the cellular mass the segmentation cavity, below the letters sc, represents all that remains of the intercellular spaces of earlier stages. Near the opposite end is a small archenteric cavity, a, communicating with the surface through the opening 6p. The archenteron is regular in outline, its marginal cells forming a somewhat epithelial lining (fig. 63 8). It has probably arisen by an invagination in pre- existing cells, since the cells lining its outer half are slightly pigmented and closely resemble those of the surface of the blastoderm. Especially noteworthy is this— that behind the archenteron, z. ¢., between it and the germinal wall, are several rows of cells. We have, therefore, evidence that in Chimera a gastrula is formed whose blastopore is located not af the rim of the early blastoderm but near it. Jt is thus a condition in which the merging of the cells of the blastoderm with the surrounding yolk does not yet take place in that zone of the blastoderm in which the archenteron ts forming. Ne have here, accordingly, a condition which throws light upon the origin of the gastrula of sharks, confirming in a striking way the interpretations of C. K. Hoffman (1896, Morph. JB., p. 210). 64 CHIMAEROID FISHES AND THEIR DEVELOPMENT. Continuing our examination of the present specimen, it will be observed that the growth of the cell mass is taking place at both anterior and posterior margins. Cells are still being contributed to the cellular wall behind the archenteron, judging at least from their relations to the yolk—such a blastomere, for example, as that near 2 having become detached from the germinal wall. And at the extreme anterior region of the blastoderm many cells are being budded out of the germinal wall. Thus, as shown in the detail (fig. 63 a), it will be seen that in an overhanging portion of the germinal wall, as below and between the points marked with aster- isks (*), a row of half a dozen cells are clearly outlined in the wall—a condition which indicates strongly a similar origin for the adjacent cells. At lower points of the Fig. 63.—Sagittal section of earliest gastrula. «, Archenteric cavity. 6, Blastopore. sc, Segmentation cavity. I-V indicate position of nuclei in yolk region. Figs. 63 A and B.—Details of foregoing section at anteriormost and posteriormost margins respectively, germinal wall, finally, cellular additions to the blastoderm are being made. Note- worthy in the present section are the vacuoles which pass deeply into the yolk and suggest, as we have already noted, modified or suppressed lines of cleavage; espe- cially well marked are those occurring in the fine yolk on either side of the blasto- derm, since they form a series of vertical fissures and mark off masses of fine yolk containing nuclei. The vacuoles also occur throughout the neighboring coarse yolk, and in connection with their appearance there we note the presence of merocytes which have traveled, as at Iv, m1, 1 or 1, far out over the yolk. We note, lastly, the way in which the fine yolk passes down in rifts into the coarse yolk, for this suggests again the modified holoblastic condition of the egg. The next stages in gastrulation deal with the extension of the blastoderm over the yolk. Thus in fig. 64 is given a sagittal section of a stage in which the diameter GASTRULATION. 65 of the blastoderm has doubled and during this growth it has lost the compact char- acter of the earlier stage. We recognize, however, in the cellular mass (at the left in the figure, Av) the group of cells which formed the ventral lip of the blastopore, and from a detail of this region, fig. 64‘, we conclude that the blastopore, 6f, has 64" Fig. 64.—Sagittal section of gastrula slightly older than the preceding. >< 35. a, Archenteron; 4, position of former blastopore ; 472, Cells of posterior lip of blastopore; sc, Segmentation cavity. Fig. 64'.—Deetail of preceding section showing the region of the blastopore. Fig. 6411.—Lateral section from the series from which fig. 64 was drawn. become closed, owing probably to stress arising from the rapid extension backward of the entire blastoderm; and we note in this connection the greatly compressed character of the cells. Parenthetically, we may also call attention in another 66 CHIMAEROID FISHES AND THEIR DEVELOPMENT. section (fig. 64 A) to the great number of amitoses occurring throughout the germinal wall, and on the other hand, the absence of mitoses in this region. Referring again to fig. 64, we interpret the cavity a as the archenteron of the earlier stage which has deepened and reached the germinal wall, still preserving its smooth posterior boundary, but dilated anteriorly and fading into a mass of detached cells. And we identify the cavity sc as the segmentation cavity now enlarged and with irregular offshoots. Another section of the present specimen shows, near the side, fig. 64", the marginal extent of the dilated archenteron and the obliteration in this region of the segmentation cavity. Its major interest, however, is in contributing data concerning the relation of the blastoderm to the yolk. Especially at the anterior end we observe that the cellular elements, of great size, have recently become detached from the yolk. The details in the study of this specimen deserve especial comment, for they indicate an intimate functional relationship between mitosis and amitosis.* To illustrate these conditions a number of details are given in figs. 64 A-N, all drawn from the foregoing specimen, but from selected sections. We may first refer to the character of the merocytes. In a detail of the anterior end of this specimen, fig. 64 A, over fifty merocytes are present, most occurring in the fine yolk, but some ina superficial rift of coarse yolk which spreads inward toward the blastoderm. We observe: numerous amitoses; the masses of fine yolk whose distinct outlines suggest polynuclear blastomeres; at one point (near 1) a nucleus surrounded with fine yolk, altogether suggesting a single blastomere; blastomeres formed out of the yolk (2 and 3); a large clear blastomere (4) which appears to have budded out of the germinal wall; and (5) a small clear blastomere, which has undoubtedly been derived from the adjacent yolk. At the opposite end of the blastoderm (fig. 64.8), and within it, is a large cell containing many nuclei, some of which are in amitotic division, and similar appearances are observed further along in the same section, fig. 64 1 and J. In the first of these, 1, a large blastomere has broken up into three smaller cells, in the largest of which the nucleus has subdivided amitotically into at least half a dozen smaller ones; in the second, J, a blastomere has divided and in each resulting *The merocytes here considered are regarded as products of the segmentation nuclei. The difficulty, however, in distinguishing finally between the merocytes derived from the segmentation nuclei or from the sperm-heads has already been commented on (p. 57). Figs. 64 A-N. Details of sections of preceding stage. (See page 67.) A. Detail of germinal wall at extreme anterior end of blastoderm. 1-5, cells which are arising, or have recently arisen from the germinal wall. There can be little question from the yolk-filled character of some of these that they have recently arisen from the germinal wall (i. e., they can not be cells which are being passed into the germinal wall, as His suggests). Such a cell as that indicated at 5, although destitute of yolk material, is so far from the remaining cells of the blastoderm that it could only have been budded off from the germinal wall. BL. Detail of the posterior rim of blastoderm showing the origin of blastomeres from the yolk wall. Observe that some of the cells are filled with coarse yolk ; others, 3, have relatively little. The cell, 2, just separated from the germinal wall, contains a number of (amitotic) nuclei. C. Detail of wall of germinal yolk. 1, 2, 3, Nuclei arising amitotically, passing in the direction of the floor of the subgerminal cavity. 4, Blastomere arising from the germinal wall. 5, Blastomere undergoing amitosis. 6, Blastomere arising from the germinal wall, and showing aster. D. Origin of blastomeres from the germinal wall. 4 and 5, Blastomeres recently separated. 1, 2, and 3, Nuclei about to be passed into blastomeres. FE. Yolk-filled cell arising from the germinal wall, and exhibiting typical mitosis. Adjacent is a blastomere whose nucleus is dividing amitotically. F. Blastomeres newly arisen from the germinal wall. G. Vesicular nuclei in region near surface of germinal wall. HI. Vesicular nucleus, undergoing amitotic division, with adjacent vacuolar spaces. I, J, Ky and L. Cells of blastoderms in some of which amitosis is taking place. Mand N. Cells of blastoderm dividing by atypical multiple mitosis. MITOSIS AND AMITOSIS. 67 Figs. 64 A-N. Details of sections of preceding stage. (See bottom of opposite page.) 68 CHIMEROID FISHES AND THEIR DEVELOPMENT. blastomere the nucleus has undergone, or is undergoing, amitosis. Now in these instances there can be no question that the amitotic divisions are taking place within the blastoderm itself, in a region where, by analogy, mitosis alone should occur, and where later, indeed, in the same form, mitoses alone are found. And we are thus constrained to conclude either that amitosis and mitosis are processes not as immutably different in fate as is generally assumed, or that at a later period the amitotic blastomeres undergo disintegration within the blastoderm. But to continue: In fig. 64 c, from a neighboring section, we observe as before amitosis occurring within the germinal wall, and this type of nuclear multiplication appears active to an extraordinary degree, as the detail, 1, indicates. Moreover, with this activity, there is evidence from the greatly elongated character of some of the nuclei, 2, 2, and from the evident trails which occur, e. g., at 3, that these nuclei are passing rapidly in the direction of the surface of the yolk. We note also that cellular increments, e. g., in such a cell as 4, are arising from the germinal wall, and, as in the former specimen, amitotic division is present, 5, in the blasto- derm proper. Adjacent to this, and in as close relation with the germinal wall, there is also evidence of mitotic division, 6. We have seen that in this section the cell 4 is arising out of the germinal wall; if any doubt exists as to possibility of cells to arise from the germinal wall at this late stage, we may refer to the detail shown from a neighboring section in fig. 64p. Here is present a row of cells arising in this manner: in the wall itself occur the nuclei 2 and 3, of which the latter is passing into a lobe-shaped process budding outward from the germinal wall. From their position we may safely conclude that 4 and 5 have arisen in a similar way. We observe, finally, that the nucleus in cell 5 is undergoing changes in the direction of amitotic division. Another interesting detail is given in fig. 64 ©. We have here two cells which appear to have arisen side by side from the germinal wall; the cytoplasm of one is clearer, more differentiated apparently than its neighbor, which contains fine yolk, yet the nucleus of the cell lacking in yolk is undergoing amitotic division, while that of its neighbor is dividing mitotically. In other sections in this series we note the following details: Fig. 64 Fr, a cell half budded from the germinal wall, also a pair of cells evidently in stage of telophase, of which the lower appears to have just budded out from the germinal wall; fig. 646, two reticular nuclei in the germinal wall, products of amitotic division (cé fig. 64c), in one of which are two large chromatin masses; fig. 64 4, nucleus undergoing a complicated series of amitotic divisions; this occurs near the surface of the germinal wall, and we note the presence of vacuoles, three in number, lying immediately above the main masses of the dividing nuclei; fig. 64 1, within the outline of a single large blastomere occur- ring in the blastoderm proper, three cells appear, and two of these appear to have been derived from the largest, in which we observe as many as half a dozen nuclei; fig. 645, a cell in a late stage of division which shows three nuclei already GASTRULATION. 69 separate in one of its daughter cells, and the nucleus in the other about to undergo amitotic division; fig. 64 kK, a nucleus similar to the last occurs in a cell high up in the blastoderm, and near it a cell which has undergone amitosis; fig. 64 m and n, two cells which are undergoing an extraordinary type of division; they contain many asters, conspicuous centrosomes, but no chromosomes; in N the cell is sub- dividing into three daughter cells. The significance of these phenomena is commented upon on a later page, in the discussion of megaspheres and yolk nuclei in their relation to germ layers. To resume the question of gastrulation: The third stage in the writer’s material is represented in surface view in plate v, fig. 31. It differs from the earlier stage shown in this way, plate v, fig. 30, in the following regards: (1) The circumgerminal ring, which had gradually been extending and carrying its nuclei peripherad, has faded out over the surface of the yolk, its proximal zone now alone noteworthy. (2) There is a conspicuous antero- posterior differentiation. The region of the blastopore is indicated by a short transverse shadow, marking the cavity of the archenteron, and the segmentation cavity is denoted by a broad transverse area. the ends of which as they approach the rim of the blastoderm bend backward, giving a somewhat crescentic shape. Three sections of this blastoderm are figured, the first, fig. 65, is sagittal, the second, fig. 654, passes between the points 4—A, shown in the surface view, the third, fig. 658, between the points 6-4. Comparing the sagittal section fig. 65, with that of the earlier stage, fig. 64, we observe increased growth at the posterior rim of the blastoderm; the germinal wall instead of shelving forward, now shelves backward, especially near the surface of the egg, still having below a sharp shoulder®™ against which lies the remains of the posterior lip of the blastopore, Avz, of earlier stages, —this region, in short, is being overgrown by the blastoderm as it progresses hindward. The archenteron thus remains, as at a, separated only imperfectly from the segmentation cavity, sc, which is now of great size. This condition, indeed, is well shown in the more lateral sections, figs. 65 A and B, and they indicate as well the narrow limits of the archenteron; the sides of which, it will here be seen, are practically confluent with the sides of the segmentation cavity. As in the preceding stage, noteworthy relations exist between the blastoderm and the yolk. We observe, for example, that in the more lateral section a tongue of coarse yolk passes inward close to the surface of the germinal wall, and we obtain evidence that the row of neighboring cells has been formed by actual outbudding. In these cells amitosis occurs, as in the previous stage. These cells, it may be remarked, do not long remain in their subjacent position, for, identified by the coarse yolk they contain, they can be traced into the blastoderm and are found widely scattered among other cells. * Vis-d-vts is a second shoulder which corresponds to the anterior germinal wall of the stage shown in fig. 63. In both regions, then, the blastoderm has overgrown the surface of the egg. 70 CHIMAROID FISHES AND THEIR DEVELOPMENT. The fourth stage in gastrulation appears in surface view in plate v, fig. 32, and is but two days older than the preceding, plate v, fig. 31. It shows the following advances: (1) The embryo proper makes its appearance in a small depression, and is conspicuous on account of the dark adjacent shadows; (2) the center of the blastoderm rises as a circular plateau, leaving at its base a well-marked flattened rim; surrounding this the circumgerminal ring has largely faded away, its nuclei yeenses SEEN aa ee ae : Re Fig. 65.—Sagittal section of gastrula slightly older than the preceding. a, Archenteron. se, Segmentation cavity. pm, Posterior mass of cells (in region of ventral lip of blastopore). Fig. 65 A.—Section parallel to preceding, but situated further at the side. Fig. 65 B.—Section parallel to preceding, but more nearly marginal. Fig. 65 C.—Detail showing coarse yolk. now having passed far out over and into the yolk;* (3) anteriorly the segmentation cavity is becoming restricted to a small area, appearing in surface view as a light- colored tract near the rim of the blastoderm. Two sections of this blastoderm are figured, one sagittal (fig. 66), the other (fig. 66) passing between the points 4A-A shown in surface view. Contrasting the sagittal section (fig. 66) with that of the * The figure represents the circumgerminal zone as too wide and conspicuous, an inaccuracy which was noticed too late for correction. GASTRULATION. 71 earlier stage (fig. 65), we notice that (a) almost the entire flattened rim of the blastoderm has been added; that (4) in the posterior portion of this rim the ectoderm is already differentiating the medullary plate of the embryo, w; that (c) the major erowth has taken place backward—in witness of this, contrast the distance between the anterior end of the archenteron and the posterior rim of the blastopore in these two stages; that (7) in this connection the main cell-mass extends itself dorsalward and becomes the plateau-like region of the blastoderm; that (¢) the germinal wall rising abruptly beside the archenteron in the earlier stage becomes excavated in Fig. 66.—Sagittal section of gastrula in which the embryo is appearing. a, Archenteron. bp, Region of blastopore. 7, Region of outermost margin of ventral lip of blastopore. m, Thickening in medullary plate of embryo. pm, Posterior mass of cells (in position of ventral lip of blastopore) . s¢, Segmentation cavity. Fig. 66 A.—Section (lateral) parallel to the preceding. the later, a portion of its material, at least, being represented by the spongy mass of cells which now forms the floor of the archenteron, @; that finally (/) there is evidence that the posterior rim of the blastoderm is rolling inward, the surface of the blastoderm growing more rapidly in this region than the lower layer with which it is connected. Detailed examination of the sections, however, leads us to the belief that the process of inrolling extends only as far as the point 7 (fig. 66); just above this the inrolled rim of the blastoderm merges with the cells arising from the germinal wall, and from this point inward openings occur between the cells and communicate with the archenteron. The section (fig. 66 A) already referred ea CHIMZROID FISHES AND THEIR DEVELOPMENT. to as passing slightly to one side of the sagittal plane shows favorably the thickening of the ectoderm at the side of the medullary plate and its inbending. We here observe also the reduced size of the segmentation cavity, the thickening of the cell mass roofing the archenteron, and the thinning out of the mass of cells, pm, forming its floor. We may at this point consider conveniently the general bearing of the process of early gastrulation in the Chimzroid. We have seen that: (a) In an early stage an archenteron was present (fig. 63), whose ventral wall was composed of cells and whose axis was at right angles to the surface of the blastoderm. (4) Ina second stage, the area of the blastoderm had increased, and the blasto- pore was closed (fig. 64); its position (fig. 64 A), however, accurately located, but more posterior than in the first stage; also the archenteron has greatly increased in size. (c) At a third stage (fig. 65), the location of the blastopore can not be accurately determined, although it is certainly near the hindmost point of the blastoderm; the archenteron is less definite, and its long axis, which remains parallel to the neighboring germinal wall, becomes tilted backward, as indicated by the arrow in the figure: and the cells, A, which correspond to the ventral (posterior) wall of the archenteron, now occupy a position further under and further forward than in earlier stages, in consequence of the hindward extension of the blastoderm. (7d) Finally (fig. 66), this hindward extension is so expressed that the position of the early blastopore shifts under the rim of the blastoderm and comes to appear at the point $f; concomitantly the archenteron increases in size, its axis lying nearly parallel to the surface and its ventral wall developing extensively both in thickness and in (anterior) extension. From these conditions it follows that in the later gastrulation of Chimera we are dealing with a reopening of the blastopore of an earlier stage. Accordingly, in contrast with gastrulation in sharks, Chimera preserves the primitive blastopore within the blastoderm itself. This stage, however, is an evanescent one. In connection probably with a change in nutritive values, whereby the yolk is passed to the archenteron from a source more and more postero-ventral there is a constant tendency for the cells of the archenteron to be drawn, both in ontogeny and in phylogeny, closer to the source of nutriment. For this reason the cells of the archenteron multiply more rapidly from below than from above (7. e., the region where primitively they were invaginated from the ectoderm) with a result that the blastopore becomes of less and less importance in early stages. It is suggested, also, that during this growth there is a constant convection of the cells of the blastoderm, in the process of which elements formed in the region of the posterior wall of the archenteron pass downward and forward. Pari passu, the posterior rim of the blastoderm, including the region of the blasto- pore, extends first backward, then downward and inward; it thus comes finally to lie under the rim (7. ¢., the later rim) of the blastoderm. GASTRULA OF SHARK AND CHIMAZERA. 73 We have emphasized these conditions of growth in Chimera, since they serve, I conclude, to explain the gastrulation of the shark, a process so puzzling that Samassa (1895) has even gone so far as to deny its presence, sensu stricto, in this group. According to the present interpretation the primitive shark had, like Chi- mera, a blastopore which opened xzear but not at the rim of the blastoderm; in this position it next became a rudimentary organ, since, apparently, the conditions governing the increase of cells in the archenteron suffered a change—inasmuch as they came to receive their nutriment directly from the neighboring germinal wall instead of indirectly, 7. ¢., through a process of continued invagination at the blastopore. Accordingly, in the development of modern sharks the blastopore A C Se ‘ Fig. 67.—Diagrams comparing gastrula of Chimera and Selachian, 4 and 4, Earlier and later stages in gastrula of Chimera colliei. Cand 2, Earlier and later gastrula of shark (mainly after Ruckert). a, Archenteron. @/, Dorsal lip of blastopore. s¢, Segmentation cavity. 07, Ventral lip of blastopore. fails to appear within the blastodermal disc, since here it has long been functionless. But obviously the blastopore would again become important in the economy of eastrulation, if nutritive material were brought into its neighborhood by any process in the growth of the blastoderm or in the encroachment of the germinal wall. Thus we may infer that it would again become a functional organ when its position was transferred to the rim of the blastoderm. In this position it still occurs exceptionally, as C. K. Hoffman has shown in Acanthias, * or it may indeed reopen deeper under the rim of the blastoderm, as the majority of investigators maintain. *In a letter, which I am permitted to quote (July, 1903), from Professor Hoffman, the comparison is accepted as follows: ‘‘In Chimzra the blastopore is located eas and in Acanthias a/ the rim of the early blastoderm. For the rest the archenteron and the open blastopore of Acanthias agree entirely with those of Chimera. Acanthias forms the bridge (in this regard) between Chimera and other sharks and furnishes us the key to the problem of gastrulation of the other sharks.’’ 74 CHIMAEROID FISHES AND THEIR DEVELOPMENT. To make the comparison of the gastrulz of Chimera and Shark more concrete we have figured two stages side by side (fig. 67, a and c, B and p). We need only add to the foregoing text the remark that the archenteron and segmentation cavity are more distinct in Chimera, and that the differentiation of the embryo takes place in a more restricted area. We append also (fig. 68, a, B, C) a scheme expressing our interpretation of the mode of origin of the meroblastic gastrula in this form. In a is pictured a sagittal section of an early gastrula of a holoblastic type, and between the points marked with asterisks is indicated the narrow zone below which the amount of yolk is supposed to have notably increased. In 8, the second stage in this evolution, is a condition not unlike the late gastrula in Chimera: The yolk mass still segments, and the ventral lip, 7/, passes inward and forward as the dorsal lip rolls backward and inward. In c, finally, is attained the condition in sharks: Archenteron and segmentation cavity merge; segmentation is lost in the Fig. 68.—Diagrams suggesting origin of meroblastic character of egg of shark. Al. Sagittal section of early gastrula of holoblastic egg (e. g., Petromyzon). 2. Section showing conditions similar to those in Chimera colliei (cf. fig. 66). C, Section of gastrula of shark. «sch, Archenteron; 7/7, ventral lip of blastopore ; sc, segmentation cavity. yolk mass, and the latter comes to pass its nutriment into the blastoderm indirectly, 2. é., as nourishment for the growth and multiplication of the cells already formed, instead of directly, 7. ¢., in the form of new yolk-filled blastomeres, and from this process there results a smooth germinal wall. This interpretation agrees in general with that lately restated by Ziegler (Lehrbuch Entwicklungsgeschichte, 1901, pp. 352-353); it differs in the interpretation of the fate of the ventral lip of the blastopore. According to the older view the ventral lip remains more or less passive, in the present interpretation it has undergone a marked change; the cells which primitively formed the ventral lip of the blastopore are to be sought in the region zv/, on the floor of the archenteron. The de facto ventral lip of the blastopore (7. e., in all stages but the earliest) is accordingly a secondary structure, which arises from the new conditions attending the overgrowth of the blastoderm. LATER GASTRULA. 75 LATER GASTRULA. Surface views of three later gastrulae are pictured in plate v, figs. 33-35, a series in which the body of the embryo becomes distinctly differentiated. In the first it occurs as a lip-like thickening, the blastoderm itself having become some- what larger in diameter and flatter than in the previous stage. In the present specimen, which was examined after my interest was aroused in the matter of the peripherad migration of the yolk-nuclei, these structures could be seen* spread out widely over the neighboring surface of the yolk. The second stage, plate v, fig. 34, resembles outwardly a shark embryo at Balfour’s stage B; the light area in the anterior and median portion of the blastoderm, which marks the cleavage cavity, 1s, however, larger than in any selachian hitherto described. In the third stage, plate v, fie. 35, the embryo arises as a knob-like eminence, its tail end projecting some- what over the edge of the blastoderm; anteriorly the surface of the blastoderm becomes thin and transparent, and it here assumes a peculiar vesicular character. DETAILS OF THE LaTER GasTRULA OF PLATE V, FIG. 35. This stage, although scarcely later than Balfour's stage 8 in shark nomencla- ture, is remarkable for the concentration of its elements. Thus, if we compare it in point of size with a similar stage in Pristiurus, measuring it always in terms of its blastoderm, it is of much smaller size. At this stage the length of an embryo of Torpedo measures about one-third the diameter of its blastoderm, that of Pristiurus about one-eighth, and that of Chimera not more than one-twelfth. Moreover, its component parts are already more highly differentiated. A number of details of this stage are given in plate vi, fig. 39, and figs. 39 A-E. In the first of these (fig. 39) the embryo with its adjacent blastoderm is viewed as an opaque object; it appears next in similar position (A) but as a transparent object, showing ectoderm, entoderm, and archenteron. Behind the embryo the surface of the yolk shows a series of lines representing either surface fissures or vacuoles, related, as we have concluded, to lines of cleavage. In the following figures the embryo is viewed from an antero-dorsal direction (8), postero-dorsal (c), postero- median (pb), and postero-ventral (cE). The mesoblast is well indicated in plate v1, fig. 39 B, also the extent of the thickening of the ectoblast forming the posterior margin of the embryonic body. In connection with these figures we may refer to the series, fig. 69 a—m, drawn from sections of this embryo cut parallel to the neighboring rim of the blastoderm (7. ¢., transverse, although slightly oblique to the axis of the embryo), and point out the following features: (1) The size and definiteness of the gut, an important factor in establishing the contour of the embryonic body; the gut acquires the cavity, g¢ (which communicates with the yolk region only for a short space near the rim of the blastoderm, ¢, and accumu- lates around its anterior end the bulk of the mesoblast, wes). (2) The fusion of ecto- and entoblast occurring not merely a/the tail end of the embryonic body but *The circumgerminal zone is, however, shown too distinctly in the present figure; its color should resemble rather that in plate v, fig. 34. 76 CHIMA‘ROID FISHES AND THEIR DEVELOPMENT. far forward, almost to the end of the embryonic gut, the band of fusion extending in a narrow zone as denoted at x, fig. 69H. (3) The concentration of the yolk- entoblast under the embryonic body; this becomes conspicuous quite in front of the embryonic gut, then merges with the gut, then separates from it, and, as the lumen of the gut opens out ventrally, it proceeds backward in a layer finally rounding Fig. 69 A-M.—Transverse sections through early embryo and neighboring blastoderm of stage corresponding to that of plate V, fig. 35. The series passes from in front backward. e, Yolk region intruding between caudal folds; 7¢, gut cavity; 7”, mesoblast; Y, yolk lying in cavity of gut; 2, fold near posterior end of embryo where ectoderm and entoderm merge. outward on either side. (4) The presence in the cavity of the embryonic gut of small masses of the disintegrating segments of the egg (fig. 69 G, 2c), which serve probably as food, interesting in connection with the fate of the yolk in Chimera (c/. in stages of Plate vit). Contrasting the foregoing conditions with those in an elasmobranch in stage B (e. g., as shown by the Zieglers, Archiv f. mikr. Anatomie, PRECOCIOUS DEVELOPMENT IN EARLY EMBRYO. TL Bd. xxxix, Taf. 11), we note that the Chime- roid although smaller is much less flattened out; that the gut which is flattened against the yolk in the elasmobranch is in Chimeera defi- nitely formed and provided with a distinct lumen; that the lateral contour of the embryo’s body in Chimeera is already developed, the ectoderm in the hinder region fusing with the entoderm; that the yolk entoblast thickens in the median axial line, a feature lacking in the shark, but important doubtless in the early assimilation of the yolk. From the foregoing details one is led to conclude that in Chimera ‘* gation” has been developed to a noteworthy degree. In spite of the small size of the embryo, both relatively and actually, it has already made strides in the direction of attaining its definite form, outstrip- ping in these regards the elasmo- branch; thus it has already developed eut outline, definitely arranged the mesoblast, separated practically the sides of the embryo from the blasto- derm, and has specially concentrated the yolk entoblast in the axial region. Accordingly, in these regards, Chi- mera stands separate from the elas- mobranchs; transitional, however, is Callorhynchus, judging from figures recently given by Schauinsland (e. 2, in his plate x1). An idea of the complicated nature of the blastoderm at this stage (plate v, fig. 35), both in itself and in its relation to the yolk, may be had by examination of fig. 70. This repre- sents part of a section which passes through the blastoderm transversely, somewhat in front of the embryo. precocious segre- Fig. 70.—Detail of section of preceding embryo. The section is transverse and passes near middle of blastoderm. It shows particularly the early differentiation of the vacuolar area. a-b, Peripheral zone of blastod erm ; b-c, central region of blastoderm; 1-5, centers of proliferation of ecto- derm into mesoderm ; 6-7, Lower ends of these proliferations in their relation to entoderm ; 8, amitosis in spongy trabeculae ; 9, grouping of mesoderm cells to form vessels; 10-11, centers, large and small, of germinal-yolk in which or near which nuclear elements are dividing amitotically . CHIMAEROID FISHES AND THEIR DEVELOPMENT. NI ioe) We notice, first of all, that the peripheral zone of the blastoderm (between the points a—é) is less complicated than its central portion (between d-c). The periph- eral zone is, however, more highly differentiated than in a similar region in an elasmobranch (cf Riickert's memoir in Kupffer’s Festschrift, plate vu, fig. 75); witness the definite character of the ectoblast and yolk entoblast, and the gigantic size of many of the mesoblast cells. But it is in the central region of the blasto- derm where the conditions are most extraordinary; we observe, that at many points, 1-5, masses of cells extend downward from the ectoderm, proliferating in ridges, sometimes giving rise to root-like processes. These terminate below either freely, or they may actually fuse with the entoblast; at various points, 6, they he close to the entoblast; at 7 is shown a point where they become continuous with the entoblast (the continuity to be traced in the serial sections). They thus form the spongy meshwork which we have already noted in the surface view of this stage, a condition of complication, which, as far as I am aware, is unknown in the extra-embryonal blastoderm of so early a stage in any other vertebrate. We note in connection with the spongy character of the blastoderm the presence of many large cells (unshaded in the figure), some of which, like many in the neighboring spongy trabecule, are undergoing numerous divisions (amitotic) as at 8.* To understand the meaning of this spongy blastoderm one should first consider it in its prospective value. Later specimens show that in this region appear blood-vessels, and in the present early preparation—and even indeed in earlier ones, we are evidently dealing with the beginnings of vascular structures. In fact in the trabecula themselves we find at various points (9) the cells already grouped together so as to form cavities, and in the latter large granular cells are undergoing subdivision, in the direction evidently of blood-building. In this character again, it will be remarked there is given an important instance of the precocious mode of development of Chimera. In other words, in this form at a period which outwardly suggests stage B of the shark the vascular development in the extra- embryonal blastoderm is (approximately) equivalent to the shark’s in stage rE. *We have here again evidence against the commonly accepted view (of Flemming, Ziegler, and von Rath) as to the significance of amitosis. Admitting that these cells come to form blood and blood-vessels, it must also be granted, as the following evidence shows, that the blastoderm becomes part of the young fish, and therefore the behavior of its cellular components is not to be compared with that of the vitellophagous periblast nuclei in the teleost. Of course it will be seen, on the other hand, that the adherent of the Flemmingian view might object that although the blastoderm itself was a permanent structure of the embryo it might none the less contain provisional cellular elements (nutritive). He will admit, however, that this rarified view as to the fate of component elements of the blastoderm receives little support from the examination of related elasmobranchian structures. The present evidence, it seems to me, favors the view that amitosis is but a symptom of early and rapid cell- multiplication. Such a need for rapid division often occurs in evanescent structures, and hence it may happen that this type of division has been given less consideration than it is justly entitled to, from the standpoints both of cell physiology and cell philosophy. In this matter I need merely mention, in view of the scope of the present paper, that there is rapidly accumulating a mass of evidence against the decadent character of amitosis. In the nature of such evidence are the observations of Conklin (Am. Nat., Oct., 1903) on the egg follicle cells of Gryllus; Kellogg's results on similar structures in Hydrophilus (Science, Mar. 4, 1904); also H. L. Osborn's observations on Fasciolaria (Science, Feb. 5, 1904) in which amitosis occurs in stages of gastrulation; Boeke's statements that in teleosts mitotic may arise from amitotic nuclei (Petrus Camper, vol. u, Afl. 2, pp. 161, 1902); finally, Child's ‘‘Amitosis in Moniezia” (Anat. Anz., vol. xxv, 1904). EXTRA-EMBRYONIC BLASTODERM. 79 The complexity of the foregoing conditions (fig. 70) applies as well to the yolk region as to the blastoderm itself. Without enter- ing into undue detail we may note the following: Extreme vacuo- lization of the subgerminal region (the vacuoles at the right in the figure are indicated by dotted lines); they usually occur in or in association with the lighter areas of the germinal yolk. If we regard the vacuoles in the earlier stage as retaining the character of intercellular spaces, they have by this time undergone, in part at least, change of function, serving now as nutriment purveyors to the yolk ento- blast. In this connection we find that at various points, 10, the coarse yolk 1s traversed by fine yolk in rifts, whose ests that of the vacuoles of es. In this fine yolk, more- shape sugg earlier stages. over, many nuclei are present, and, judg- ing from numerous amitoses, dividing rapidly. In addition to these rifts of fine yolk, we note that there occur at many places throughout the coarse yolk small areas of fine yolk, 11; these have in nearly every case nuclei in or near them, and we have thus ground for regarding the yolk region of the egg not as a syncytium pur, but rather as a mass of yolk-filled cells whose boundaries have broken down, but whose individuality as cells has not yet been wholly lost. A second section of the extra-embry- onic blastoderm of this stage is shown in fig. 70 A, a detail of a section passing | through the blastoderm considerably in front of the preceding section. Here is indicated even better than before the presence of giant cells which have arisen from the yolk, migrated outward, and are undergoing division in the region immediately below the ectoderm. At one point (@) a yolk cell of gigantic size is shown (unshaded); at other points (6, 6, 6) similar yolk cells are undergoing division by amitosis. Figs. 70 A and B. —Sections of stage of early embryo figured in fig. 69. A. Portion of extra-embryonic region in section corresponding to fig. 69 B. Xe NSS) B. Embryonic and extra-embryonic regions in section similar to fig. 69 C. 80 CHIMAROID FISHES AND THEIR DEVELOPMENT. Into this region extends a delicate layer of mesoblast (vz); and here and there groupings of cells of this layer, as at v, suggest the formation of blood vessels. Under the gigantic cell, one notes that the cells of the mesoblast layer are of remarkable size. Another section, given in fig. 70 B, pictures details of a section similar to fig. 69G. This illustrates particularly a subgerminal zone containing large yolk nuclei; of these some are situated close to the surface of the subgerminal wall, and one (7) has passed into the entoderm. This obviously. cannot be confused with the adja- cent entoderm cells, if only on account of its greater size. In this section a special area of formative yolk is shown underlying the periphery of the blastoderm. Under the embryo itself the formative yolk attains the surface notably at the sides of the embryonic body, and it is from this region that the cells appear to be passed into the embryo. Less activity is probably present in the ventral median line, on account of the quantity of coarse yolk which is here present. Later GAsTRULA. EMBRYO WITH OPEN MEDULLARY FOLDs. This stage, figured in surface view, plate v, fig. 36, and in detail, plate v1, fig. 40, may be compared with Balfour's stage p in elasmobranch. In spite of the conspicu- ous growth of the embryo, the blastoderm, it may be noted, remains remarkably small in size. In this stage the blastoderm of Chimera shows a well-marked central area, which on closer examination is found to be made up of spongy mesh- work; there is also a somewhat thickened rim, and a marginal zone, the latter shown in sections to be formed of peristomial mesoblast. Beyond the limits of the blastoderm the surface of the yolk showed faintly diverging lines which suggested cleavage planes. (C/. plate v1, fig. 39 A.) The embryo itself, when viewed as a transparent object, plate v1, fig. 40, shows shark-like medullary folds, more delicate, however, and narrower in proportions. The tail folds are less conspicuous ; the mesoblast concentrating in this region shows on each side a dark area, the rela- tions of which are referred to later. DETAILS OF STAGE D. Transverse sections of the embryo and the neighboring blastoderm in this stage are pictured in fig. 71, A-1. Thus beginning with a section through the tail folds, we see in B, ectoderm and entoderm continuous in the chordal region. In this section the mesoderm merges with the entoderm not at the sides of the chordal region, but near the margin of the blastoderm, thus suggesting the theoretical condi- tion in the origin of the mesoblast advocated by Graham Kerr. In section c (at the left side, the plane being slightly oblique) the side of the blastoderm is coming into functional connection with the yolk; the notochord is here being folded off from the entoderm; the latter is now a thick, flattened layer, its outer half lying apposed to the yolk wall. In p the section shows the beginning of the neural folds; below them is a well formed layer of mesoblast, also the dorsal wall of the gut; the gut lumen appears at 2 at its side the dorsal wall of the gut shows a wide contact with DETAILS OF EARLY EMBRYO. 81 ial Hin Ha Mi Figs. 71 A-E and continued F-I on page 82.—Transverse sections of late gastrula shown in Plate V, fig. 36. The sections pass forward ; the first of the series, A, traverses the tail folds; the last, I, the head region of the early embryo. a, Tongue of mesoblast cells representing the urogenital anlage; 0, megaspheres in process of passing through the yolk-entoblast; 0c, body cavity; ¢, points in extra-embryonic region where the ectoderm cells are being proliferated into the blastoderm; (/; gut cavity; ty, megasphere appearing in peristomial mesoblast. 5 CHIMAROID FISHES AND THEIR DEVELOPMENT. the yolk, and here at various points yolk nuclei are clustered, having evidently an important physiological relation with the overlying blastoderm; we note at aa tongue of mesoblast cells which projects medianward; this occurs but in a few sections, and evidently corresponds to the dark area noted in surface view; it resembles, however, so closely the ‘‘lame intermédiaire’ (Swaen and Brachet) Figs. 71 F-I. (For description and lettering see page 81, A-E.) in the teleost, that, if for no other reason, we are led to suggest that it represents the precocious beginnings of the excretory system. In £ the notochord has sepa- rated from the yolk, the gut lumen becomes narrowed, and lateralward the first trace of a body cavity (dc) appears. We observe that the margin of the gut passes directly into yolk-entoderm, the distinctness of its lower boundary having DETAILS OF EARLY EMBRYO. 83 faded away, and, part fassu, the yolk nuclei have greatly increased in number. — In the region where the yolk-entoderm approaches the lumen of the gut it thickens and sinks downward, leaving as the floor of the gut cavity a wedge-shaped mass of ger- minal yolk. At the outer rim of the yolk-entoderm we observe that it becomes con- tinuous with the mesoblast; in other words, recalling sections p and G, the peristomial mesoblast of Chimzera which now arises is zo¢ con¢Znuous with the gastral mesoblast. We have thus a reason for inquiring whether gastral and peristomial mesoblast Figs. 71 J-N.—Details in sections of foregoing embryo (figs. 71 A-I). J. Region of peristomial mesoblast. €, ectoderm; €!, cells recently derived from €; €7¢, entoderm; 7, peristomial mesoblast. XK. Detail of subgerminal yolk region showing cellular arrangement of merocyte elements. Z. Lying in the subgerminal yolk is a megasphere, which, on the evidence of the overlying vacuoles, is in the process of rising towards the yolk entoderm. M. Similar megasphere passing into the yolk entoderm. V. Megaspheres similar to preceding, but representing a somewhat later stage of passage into the blastoderm. are as intimately related as we have generally assumed.* A condition of the peristomial mesoblast is figured in detail in j, and it proves of considerable interest, since the region of mesoblast proliferation is of wide extent. Not only are cells budded out from the marginal mass 7, but we observe also that cells are added to the mesoblast from the neighboring ectoderm; thus at ¢’ is a cell which has been derived from the ectoderm ¢, where, by the way, a syncytium is now present; and *Cf. the current view as to the secondary confluence of blastopore and yolk ‘‘blastopore,'’ as summarized in Ziegler's Handbuch der Embryologie, pp. 352 and 353. 84 CHIMZEROID FISHES AND THEIR DEVELOPMENT. at the point ¢ a mitosis is taking place preliminary to budding off another mesoblast cell. Weconclude that the cell e’ has been derived from the layer e, and not from the cell mass #, when we consider (1) that its granular contents agree in character with the layer e¢ rather than with the mass 7 ; (2) that a continuous boundary line sepa- rates the mass w from e’; and finally (3) that the cell e’ is connected with the layer by a protoplasmic process, above which a nucleus in mitosis is present. Less evident, from this section at least, is the question whether cells are added to the Figs. 71 O-II.—Details in sections of foregoing embryo (continued from page 83). O. Detail showing transition between yolk region and the cells of the blastoderm. m. Yolk nucleus lying against the wall of the vacuole v’. v. Zone of large vacuoles. m’. Yolk nucleus now lying within a vacuole, and transformed into Y. Zone of small vacuoles and fine germinal yolk. a yolk-surrounded blastomere. ye. Yolk entoderm. P. Detail of blastoderm, showing at y the division by mitosis of a megasphere lying in the yolk entoderm. Q. Detail of section near the marginal region of the blastoderm, showing single megasphere, y, lying free in the space between ectoderm and yolk entoderm. yn. Yolk nuclei undergoing division by atypical mitosis and by amitosis. R-ZT. Details showing various phases of division in yolk nuclei. peristomial mesoblast from the entoderm more proximal in position. At some points one is inclined to admit that such a cell as shown in J, ev¢, is being budded off into the tongue of mesoblast. (Cf the condition shown in the section n. ) In the section Fr, the notochord is again continuous with the entoderm; the gut region rises, and its lumen is now walled with cells save in its median-ventral line. Here a thin wedge of yolk intrudes. Especially noteworthy is the relation of the yolk to the yolk-entoderm in this region. The latter has again a more MEROCYTES AND BLASTODERM. 85 distinct ventral line of boundary, broken only at points, as at 6 and 4, where cells from the yolk are entering. There can be no question in this regard since the entering cells are distinguishable as large in size, circular in outline, and granular in content. (Cf sections 1, mM.) Another noteworthy feature in this section is that some of the ectoderm cells as at ¢ and ¢, give off amceboid processes and, I am led to believe, later become detached, contributing to the growth of the mesoblast. A detail of this condition is shown in section J. We may finally note that the body cavity, 4c, reaches its maximum size in this region of the embryo. In G the floor of the gut becomes cellular ; the notochord is again separate from the gut wall; and as before merocytes contribute directly to the growth of the yolk entoderm. Inu the last-mentioned character is seen even to better advantage, for not only are the large yolk-cells passed into the lateral yolk entoderm, but they appear also high up in the central gut wall, as at g, and in the region of the peristo- mial mesoblast, as at m2. In 1, finally, a section is shown passing through the region of the head tip, which now projects forward above the blastoderm. On either side of the gut the mesoblast is distinct, differing in this regard from the condition shown in an elasmobranch (cf. Ziegler’s figure 19, 1, Arch. f. mikr. Anat., Bd. xxxrx, Taf. rv). In the neighboring blastoderm, as in the shark, the mesoblast is limited to a small tongue of peristomial cells. Before concluding an account of this stage two of its features still deserve comment. (1) The fissuring of the yolk region. The fissures are usually vertical, as indicated in all the foregoing sections, and may, as we have already seen, be regarded as homologous with cleavage spaces. (2) The mode by which merocytes become cells of the embryo. This heading, however, deserves to be treated ina more formal way. THE TRANSFORMATION OF MEROCYTES INTO CELLS OF THE BLASTODERM. In this connection a number of details of sections of stage p have been figured, figs. 71 K-11, and in examining the series we find evidence, first of all, that merocytes move froma lower into a higher zone of the yolk. Thus, in fig. 71 0, the merocytes are elongated in the direction of the yolk-entoderm.* Also in the three sections L, M, and N we observe a great yolk cell (megasphere)f first deep in *That this is connected with a migration of these elements in the direction of the surface of the cell mass is known by analogy — witness the behavior of slime cells in the skin of amphibians and fishes (e. ¢., Homea). {+The megaspheres can have little to do with primitive ova, since they occur widely scattered throughout the blastoderm. Thus in fig. 71H one is arising at the extreme rim of the blastoderm, in F several are seen midway between the embryo and the rim of the blastoderm, in 1 one occurs near the middle of the floor of the gut; others appear in mesoderm and others still in ectoderm. In these several regions they are seen to undergo division, losing more and more of their appearance as megaspheres (v. fig.71 Pp). It can not be believed, therefore, that these elements are to be regarded as primitive ova, destined to carry the segregated germ plasm into the embryonic genital folds, for this would involve a conception of primitive ova traveling about extravagantly, from the gut wall to the rim of the blas- toderm, a conception the more improbable when we consider that the urogenital region, to which primitive eggs naturally belong, is already indicated by this stage, as at a, fig. 71 G. On the other hand, it follows, I believe, that the evidence provided by Chimera strengthens materially the position of Riickert that the megaspheres in elasmo- branchs are to be regarded not as primitive ova but as highly specialized bearers of nutriment, capable of carrying into the midst of embryonic tissues centers of new formative energy. These as single large cells could be passed through the intervening tissue more effectively than could the many small cells to which they give rise, for the resistance of an embryonic tissue to the penetration of cells is obviously proportioned to the surface-contact of the invading cells. 86 CHIMA:ROID FISHES AND THEIR DEVELOPMENT. the germinal wall, next having just emerged from the germinal wall, and finally having almost passed through the niveau of the yolk-entoderm. We have even data indicating how the upward migration of such a megasphere takes place. One sees in La line of vacuoles appearing between the megasphere and the yolk- entoderm, and it follows clearly that the vacuoles, by a process of coalescence, provide a less resisting space into which the megasphere can rise. Regarding, in the next place, the fate of the megaspheres, I think that there can be no doubt that they serve to bear nutriment to the tissue which they enter. In some cases, as in fig. 71 P, at y, they undergo mitotic division (after having divided only by amitosis in the yolk), and their descendants can not be distinguished from the neighboring cells. In other cases, m and Nn, they become closely surrounded by cells, entoderm in the present case, which form around them a syncytium, and appear to serve as nutriment distributors; witness for example the grouping of the cells around the large megasphere in m, and the radiating arrangement of the cells adjacent to the cluster; even the ectoderm is budding off a cell at the point nearest the megasphere. In a word, I think we can fairly conclude that in Chimera, even in this late stage, cells are constantly being added to the blastoderm from the germinal wall. This condition maintains in the case of the megaspheres, as we have just noted, and it holds equally good for other types of cellular additions to the blastoderm. We thus observe in 0 (a detail of section G) that between the yolk-entoderm (1c), and the wall of fine yolk (4) is a vacuolar zone,* in which merocyte elements are being ferried over to become cells of the blastoderm; thus at 2’ is a vacuole into which the merocyte (7) is about to pass. It is to be noted, however, that cells may also appear in the finer yolk, and thence by the mediation of an enveloping vacuole be passed upward into the vacuolar zone, thence to the blastoderm (c/. in fig. 71 0, at 7’). That throughout these stages there is a general transformation of the yolk from coarser elements into finer elements there can be no question. Deep in the yolk appear nuclei surrounded by spherical masses of finer yolk, in turn surrounded by masses of coarser yolk, in turn more or less irregularly by a system of vacuoles (= intercellular spaces) fig. 71 kK. There is, to be sure, a greater or less amount of coalescence of these yolk elements, and in the zone close to the entoderm we observe that the nuclei with their surrounding fine yolk have come to merge into a single layer (= the zone of merocytes of the subgerminal wall). It is from the elements of this layer in turn that some cellular additions to the blastoderm are made. The nuclear changes which occur during the process of their “levitation” are worthy of especial comment, for while the cells of the yolk-entoderm now divide by mitosis (as in 0), the nuclei of the region below the vacuolar zone divide amitotically, *Similar conditions have been observed in the early stages of teleosts (cf, among others, Hoffmann, Zeit. wiss. Zool., vol. xLv1 (1888), pl. xxxv, a paper, by the way, which is too little referred to in recent work on teleostean embryology). MEROCYTES AND BLASTODERM. 87 and under varied and striking forms—albeit in a series more or less gradational (7. e., showing more decided mitotic character) as one passes from a lower to a higher zone in the yolk substance. To illustrate various types of division: In pr, in a sphere of fine yolk is a nucleus about to divide amitotically*; in s a similar nucleus has undergone such a division, in this case four nuclei resulting. In a somewhat similar case, T, noteworthy growth in two of the resultant nuclei has occurred; they have, in fact, passed out of the sphere of finer into the coarser yolk. In u three similar and large nuclei result. In v, which represents a later stage of the condition shown in T or u, and is drawn similarly from deep in the yolk region of a section (¢. g., as seen at several points in kK), continued amitosis occurs; here one of the larger nuclei, especially, 1s seen to be budding off a small nucleus, and it has already‘apparently budded off several. In w, a similar detail indicates the great rapidity with which nuclei may arise; a large nucleus at one point has given off a small one, while at a neighboring point almost simultaneously (judging from the close position of the small nucleus) it is budding out a long process which is about to be separated not into a single new nucleus but into two. In x seven nuclei have arisen from a single center (? sphere substance) in the fine yolk, and of these one has undergone rearrangement in its chromatin material. Of this a dense mass occupies the center of the nucleus and is connected with the nuclear wall by a series of radiating linin strands. In y a somewhat similar nucleus is shown in detail; at one side it is apposed to the finer yolk (= ? sphere substance) and here the mass of chromatin approaches, indeed almost touches the nuclear membrane (for nutritive reasons?). In another nucleus, z, the chromatin mass shows a doubled arrangement, preliminary, as it appears, to a stage in division shown in Aa, FF, and possibly in gs. In turn the doubled nucleus in cc is obviously a further stage than AA, but it shows also around it a series of (five) smaller nuclei which, from their radiating arrangement around the dividing nucleus in the center of the fine yolk, are possibly the descendants of a similar type of nuclear division. In pp a nucleus shows a less distinct doubling of its chromatic elements than Aa—cc. And in EE a distinct threefold division occurs. GG represents a stage in division carried further than cc, the neighboring nucleus having probably arisen from a similar division. In uu are two neighboring nuclei, the products, we conclude, of a division like that of GG and cc: but, curiously enough, they are undergoing division in different ways. The upper, near which appears an attraction sphere and centrosome, has arrayed its chromatin in two masses nearly equal in size, each suggesting a confused series of chromosomes; the lower is simply passing out a portion of its chromatic substance into the fine yolk. In u, the last of the series given, two nuclei appear; they are evidently products of such a division as GG, and each in turn is about to undergo division. The lower one is noteworthy, since the division of the chromatin material is practically completed in the middle of the nucleus. It may be said in general that the nuclear processes which here approximate mitosis (cc or HH) are observed in the region immediately subjacent to the yolk entoderm. *A similar condition in the embryonic germ cells of Loligo appears to be due to rapid growth, and is not followed by fragmentation (Miss Sturges, Science, 1899, Feb. 3, pp. 183-184). 88 CHIM4ROID FISHES AND THEIR DEVELOPMENT. In summary: The evidence which is thus provided strengthens the conclusion that in the gastrulation of Chimera amitosis is not to be interpreted in accordance with the current view, 7.¢., as a process of decadent cell division. It is conditioned, rather, by rapid growth and multiplication of nuclei, since its products may resume mitosis when the usual rate of cellular division is attained. Moreover, the products of amitotic division in the blastoderm of Chimera, are too many and too widely scat- tered to warrant the belief that their cellular descendants can play no part in producing permanent organs LaTeR GASTRULA. EMBRYO WITH PaRTLY CLOSED MEDULLARY FOLps. This stage is figured in surface view, plate v, fig. 37, and enlarged, viewed as a transparent object, in plate v1, fig. 41. It corresponds approximately with Bal- four’s stage F in the shark. Comparing the blastoderm of this with the preceding stage, we find that it has increased but little in size. The spongy region, however, which occupies its central portion appears more prominently, and we observe a noteworthy thickening in the region of mesoblast (gastral) extending outward on either side of the embryo. The details of the embryo are well seen in a /o¢o preparation. The medullary folds arch over and meet in the median line, fusing in the posterior third of the embryo’s length. In front of this, after a slight interruption, the folds meet again, then diverge to a degree suggesting the corresponding stage of shark. The tail folds are conspicuous at this stage, and we observe that the gut has arched upward, a transverse line showing where a neurenteric canal is to open below. On either side in this region the mesoblast is thickened, fading away laterally. Here are forming the extensive caudal veins. Other vascular details are shown in the antero-median vessel (apparently vitelline vein) which appears immediately in front of the head and spreads out widely over the blood-producing region. We note also transverse larger vessels, the vitello-intestinal, extending outward on either side to about an equal distance. Gastral mesoblast is conspicuous in this stage; in this may be traced about a dozen somites, the anterior ones extending far forward. DETAILS OF FOREGOING STAGE, CORRESPONDING TO BALFOUR'S STAGE F. Sections are shown in fig. 72 A-E passing through the blastoderm in a plane transverse to the axis of the embryo. In the first, which passes through the tail region of the embryo, we observe that the mesoblast bands (wes) are continuous with the entoderm not in the region adjacent to the notochord but marginally (¢/. the view of Graham Kerr as to this place of origin in the vertebrate gut pouches); near by the entoderm (ev?) thickens conspicuously, then thins again as it passes into the notochord. Only at the open notch between the tail folds does the lumen of the nerve tube pass over into the wide space (¢/. fig. 71 A) which is coming to form the cavity of the gut. It will be seen that it is especially the thickening of the entoderm and the constricted origin of the mesoderm which in the transparent preparation (plate v1, fig. 41) causes the appearance of a dark band in the region of the tail folds of the embryo. In fig. 72 B similar conditions in gastral mesoblast and DETAILS OF EARLY EMBRYO. Figs. 72 A-E.—Transverse sections and details of the blastoderm shown in fig. 72. The sections pass anteriorward from the region of the caudal folds, shown in section A, as far as the “neck” region of the embryo, section E. d. Beginnings of segmental duct. mes. Mesoderm. ect. Ectoderm at the point where this becomes continuous with the yu. Yolk lying free in the gut cavity. mesoderm in the tail folds of the embryo. x, Urogenital anlage. ent. Entoderm. 90 CHIMZZROID FISHES AND THEIR DEVELOPMENT. entoderm prevail; the thickened ectoderm at ect marks a point at which this layer is making cellular additions to the mesoblast; it represents the marginal point where the tail fold and the margin of the blastoderm meet. At other points also, the mesoblast is receiving increments; in addition to the gastral mesoblast we note cells arising from the wall of the yolk-entoderm midway between the cavity of the gut and the periphery of the blastoderm, and we see further that an invasion of Figs. 72 F-K.—Details of the region of the yolk-entoderm of fig. 72. In /* the region is indicated in detail which lies immediately below and at the side of the arching wall of the gut. (Cf. fig. 72 E.) a. Large vacuolar nucleus which appears on the point of undergoing reconstitution m. Megasphere. into a cell of the yolk-entoderm. $2. Subgerminal zone. b. Nucleus similar to foregoing, but in a less advanced condition. v. Vacuolar zone. e, ce’, d. Cells which have recently been differentiated out of the germinal wall. ye. Yolk entoderm. G, H, and | illustrate particularly the zone of reconstruction of yolk-entoblast cells from yolk nuclei. In J a telophase occurs, repre- senting a rare condition in the subgerminal zone. In K, similarly, a telophase occurs in a megasphere. The latter has, however, passed through the zone of vacuoles and lies in the yolk-entoderm. In this neighborhood, however, as we note at the left, a syncytial condition may be present. cells from the periphery of the blastoderm has occurred, in the form of a crease- shaped invagination. In c the dorsal wall is sharply distinguished from the sides of the gut. On the floor of the latter appear small masses of yolk, 7, which can only serve, as already noted, as ingested nutriment. The mesoblast in this region shows considerable differentiation; myotomes are sharply marked off; the gono- DETAILS OF EARLY EMBRYO. OI nephrotomal zone is of notable size; at @ and in the adjacent cell-mass (at the left) are the beginnings of the pronephric tubules; and below at x appears the thickening of the mesentoderm whence arises the posterior portion of the pronephric duct. In a section, p, passing through a more anterior region of the embryo, the urogenital structures are practically undifferentiated; the mesoblast extending continuously from the notochord to the periphery of the blastoderm. In this region the mesoblast probably receives little or no increment from the yolk-ento- derm, judging from the latter’s smooth surface, save only at or near the margin of the blastoderm. Below the yolk-entoderm in this region the subgerminal zone of nuclei is more conspicuous and definite than in the early stage, fig. 71 E, and this zone, indeed, appears with even greater prominence in the more anterior section, fig. 72 E (to be contrasted with fig. 71 F or G). It will here also be seen that divi- sion of the mesoblast into splanchno- and somatopleure is occurring, and that the lateral wall of the gut is more definitely established. A detail, shown in F, indicates the more special relation of the subgerminal zone to the marginal cells of the gut cavity. The subgerminal zone is here reduced to a narrow tongue (cf. also £), which inserts itself under the thickened mass of cells at the base of the gut wall, in the direction of the lumen of the gut. In the present detail the base of the gut wall is shown at gw, the yolk-entoderm at ye, the vacuolar layer at v, and the subgerminal zone at sgz. We note first of all the narrow- ness of the vacuolar layer, through the intervention of which we have seen (fig. 71 0) yolk nuclei become cells of the embryo, a condition indicating the specialization of this region. In this zone (v), furthermore, we see large nuclei which are evidently in transition between yolk and embryo, and at m a megasphere which has just passed through it, the vacuoles becoming reconstituted below. Most significant in the region of the rim of the gut wall is the concentration of the elements of the subgerminal zone, coarse yolk, fine yolk, lacuna, vacuoles and yolk nuclei of different kinds, the continuation (to the left) of the vacuolar layer, and the compounding of its vacuoles—characters which are obviously to be interpreted as more special and complicated than in the earlier stage. A few additional details may be cited. In G, where nuclei are passing through the vacuolar zone and becoming cells, we observe that at ¢c a nucleus which has been taken into a large vacuole (a process forming now a reconstituted cell), is still dividing amitotically, and that at ¢’ a similar division has recently occurred, indi- cating in both cases, as we have before remarked, that the difference between ami- totic and mitotic division is one of degree rather than of kind. In n, a detail from a section close to fig. 72 E,a point is figured where merocytes and newly constituted yolk-entoderm cells occur in such confusion that it is difficult to say where the layer of merocytes terminates and where the cells of the embryo begin. And the same is true of the detail shown int. In the last figure, on the other hand, merocytes are still multiplying, even at a point close to the yolk-entoderm. In J, a detail of the vacuolar region, cells are arising from merocytes; at 4 a merocyte, less vesicular than a, adjoins a vacuole into which it will probably pass, judging from transitional conditions (cf. the neighboring c). And even in the vacuolar layer such newly 92 CHIMAROID FISHES AND THEIR DEVELOPMENT. constituted cells may divide, and by mitosis, although this is not of the usual type (cf.atd). In this connection, finally, in k a detail is given showing that megaspheres as they pass into the yolk-entoderm present more or less evident mitosis, witness the conditions # and m (cf. also fig. 72 F at wz). Parenthetically, just below the megaspheres here mentioned are vacuoles into which merocytes are passing. EARLY EMBRYOS FROM THE COMPLETE CLOSURE OF MEDULLARY FOLDS TO OPENING OF GILL-CLEFTS. An early embryo attached to its blastoderm is shown in plate v, fig. 38. This may be contrasted with the stage of closing medullary folds shown in same plate, fig. 37. In the blastoderm we observe that the spongy central area has increased notably in size and that it has even extended to the anterior rim of the blastoderm. We note also that asymmetry has made its appearance, the embryo now lying some- what on its right side. The present blastoderm has increased more rapidly at its left, and here a lobe-like eminence is produced hindward over the yolk. The entire size of the blastoderm is scarcely larger than in the preceding figure. The embryo is shown in detail, plate v1, as an opaque, fig. 41°, and then as a transpar- ent object, fig. 41°. In general this stage corresponds with Balfour's stage G in shark; it differs, however, in the definiteness of its structures, for the anterior region has already become quite highly differentiated in spite of the fact that the tail region is still flattened out against the yolk and hardly protrudes beyond the rim of the blastoderm. About 22 segments are present in this stage. The head rises above the blastoderm and the divisions of the brain and the optic vesicles are formed, and it is an evidence of the high specialization in development that the embryo of this large-eyed form should possess large optic vesicles at this early period, 7. ¢., before the tail end of the body is established,—a fact of considerable interest from the standpoint of embryonic adaptation. In this stage two gill-slits are appearing, ¢’, g™. The region of the pronephros is marked out at fz, the heart at h, the anterior cardinal vessels at ¢, and the vitello-intestinal at 0. In the tail region the neurenteric canal is distinctly seen at z. DETAILS OF THE PRESENT EmBryo (STAGE G),. A series of selected transverse sections of this embryo may now be passed in review to indicate the more prominent advances, figs. 73 A-uu. The anterior sections A-p pass through the ectoderm inclosing the tip of the head and show a conspicuous median infolding (recessus olfactorius impar) which in surface view eives the appearance of separating a ‘‘forebrain”’ from a “‘right optic vesicle,” the sections having been cut in the plane indicated by the dotted line in plate v1, fig. 41°. The next section (£) touches the distinct end of the central nervous system, the wall of which is more extensively traversed in Fand c. Inu, 1, and J, the lumen of the forebrain is traversed. In xk and L, representing many sections, the cavities of the optic vesicles appear, and we observe here closely apposed to the ventro- median wall of the brain a mass of cells which in later sections is seen to constitute the anterior end of both notochord and gut. In sections m and N this cell mass forms a conspicuous ventral keel, in n the lumen of the gut first appearing. In o and p DETAILS OF EARLY EMBRYO. 93 Figs. 73 A-S.—Transverse sections of embryo shown in plate V, fig. 38, and in plate VI, figs. 41 and 41 A. These begin at the head end of the embryo, section A, and extend through 47 sections to the tip of the tail, section UU (see pages 94 and 95.) 9g, Gut cavity; ”, notochord; $, somite. o4 CHIMA:ROID FISHES AND THEIR DEVELOPMENT. Figs. 73 T-EE—Continued. be. Body cavity. nc. Neural crest. g*, g™. Evagination of gut wall to form the sn. Subnotochordal rod. first and second gill openings. y. Yolk lying free in cavity of gut. h. Heart. DETAILS OF EARLY EMBRYO. 95 Figs. 73. FF-UU.—Continued. w. Wedge-shaped mass of yolk which comes to pass into the ventral wall of the gut cavity. 06 CHIMAROID FISHES AND THEIR DEVELOPMENT. we distinguish in this ventral cell mass a lower lumen-bearing area and above a thickened mass, on either side of which, attached but not fused, lies a solid mass of mesoblast. In Q we distinguish notochord and gut (#, g); on either side of the notochord the mesoblastic somite (s) bears a cavity. In r the mesoblastic sacs are well separated from both notochord and gut, and the notochord itself, greatly reduced in size, shows a compressed and almost longitudinally subdivided appear- ance. Ins, the body of the embryo is becoming flattened on its side; the lumen of the gut is deep and narrow; closely apposed to its sides are the mesoblastic masses whose lumen now becomes greatly reduced; on the dorsal median wall of the gut appear the beginnings of a subnotochordal rod. From this stage onward the lumen of the central nervous system becomes notably reduced. In T the section passes through the embryo in the plane where the neck region flattens out over the yolk. Here we note the distinct subnotochordal rod (sz) and the flattening mesoblast which now forms a delicate band almost surrounding the gut. In the surface view of this region, on the other hand, only the thickened proximal ends of the mesoblast masses can be distinguished. In vu, where the neck is flattened out, the heart appears at /#; and in the upper region of the gut we note the thickening of the wall of the gill-slit, the cavity of which is seen in the preceding section at g*. In v, as indeed in some of the earlier sections, a thickened neural crest appears at zc. In w the body cavity (4c) is becoming conspicuous. In x the somato- and splanchno- pleure spread out widely peripherally; in the gut we notice in the thickening of the lateral walls an out-bending for the second gill-slit (c/ in z, g"') and in the cavity of the gut in this and in many sections following we find masses of yolk. These masses, sometimes small, as in sections z, AA, BB, EE, sometimes large, as in Y, CC, bp, are unquestionably budded out (as in EE and HH) of the ventro-median wall of the gut. On account of their abundance and range in size we can not conclude that they are artifacts, but, on the other hand, if we regard them as normal structures, it is natural to assume that they serve as food material, and are assimi- lated by the gut in the usual way. This conclusion, simple as it seems, is none the less difficult, since it attributes to Chimera an embryological process which appears to be unknown in the vertebrata and only remotely paralleled among invertebrates. If, accordingly, we accept the present evidence, it follows that Chimera is to be regarded as the terminal member of an evolutional series, at one end of which were forms whose yolk-laden cells contributed directly to the growth of the young; next came those whose yolk-filled cells contributed indirectly to the growth of the young through various processes, typically through the intervention of merocytes; and finally, in Chimera, the mode of nutrition by merocytes is supplemented by a still more oblique process, z. é., one which passes fragmented yolk material from the zone of merocytes directly into the lumen of the gut. Continuing the sections: InGc, and in many sections following, a wedge-shaped mass of yolk material (zw) is converging toward the ventro-median line of the gut (v. also p. 76); in LI, it becomes subdivided, and in mm appears a small recess which may also contain this nutriment (? anlage of liver). In jy and in following YOLK AND YOLK-ENTODERM IN EARLY EMBRYO. 97 sections the pronephric duct appears, at first only on the left side, as an ectodermal keel, beginning about the plane of the 8th somite. Thence, passing backward, it merges with the somatopleure at about the plane of the 12th somite, after mM. In this section the subnotochordal rod appears for the last time. In 00 the notochord dips into the dorsal wall of the gut; and in pp it forms an evagination of its wall. QQ and rr are sections through the neurenteric canal, and ss to uu through the tail end. Two further details of this stage are shown in figures 74 and 75. The former of a section close to that of fig. 73 LL, the latter from a section close to fig. 73 G, representing only a detail of the extra-embryonic blastoderm lying under the region of the head. Fig. 74 has been given to illustrate the ingress of yolk material through the ventral wall of the gut, for here is seen the wedge of yolk protruding through the thickened mass of yolk-entoderm cells, but under conditions which bespeak the complicated nature of the process. For the rest, there is here not a mere rupture which admits the yolk into the cavity of the gut, but an attendant Fig. 74.—Detail of section of early embryo shown in fig. 73 LL. Yy Yolk plug pressing into cavity of gut; ¥’, ¥'’, y''’, layers of yolk of different consistencies. series of changes of which the “‘rupture”’ itself is, with fair probability, the terminal member. Thus the wedge-shaped mass of yolk (7) is composed of fine yolk; it next passes through a transitional zone (7) into the coarse yolk (y”). And on either side of the wedge lies a layer of very coarse yolk (7), which obviously comes into close physiological rapport with the neighboring layers, for this thickens as it approaches the yolk-wedge, and here it is filled with nuclei of extraordinary size. Indeed on one side (left) we note that this layer of coarse yolk is separated from the yolk-entoderm by a layer-like offshoot of the fine yolk (y”) from near the point of the wedge. We observe also the relation which the bordering yolk-entoderm bears to the point of the yolk-wedge, for this layer is here many times thicker than in neighboring regions. The yolk-wedge, in short, which passes into the cavity of the gut stands in specialized relation (1) to the usual mass of yolk, 7. e., spreading out fan-shaped below, thus securing a large surface of contact; (2) to the lateral areas of coarse yolk; (3) to the lateral masses of yolk-entoblast, and (4) finally, as we 98 CHIMAZROID FISHES AND THEIR DEVELOPMENT. have already seen, to the walls of the gut, since it passes to them yolk masses, large and small, and perhaps also dissolved yolk material. In evidence of the nutritive value of this material witness numerous mitoses in the adjacent (inmost) cells of the entoderm—one of which appears in the present section. In fig. 75 a detail is given of the process by which yolk-cells are passed into the tissues of the embryo. In this portion of the extra-embryonic blasto- derm the mesoderm occurs only as detached (mesenchymatous) cells (wz); the ectoderm forms a_ single-celled layer, and the entoderm a closely formed cellular mass (ye). Between the entoderm and the yolk is the usual zone of vacuoles (v). At meg a large volk-filled cell (cf pp. 83 e¢ seg.) pro- trudes from the yolk into the entoderm, the cells of the latter affording little bar to its progress upward. In this connection we note that the huge cell (weg) lies now within a vacuole in whose wall yolk-nuclei appear; indeed at one point a yolk- nucleus has actually entered the vacuole. In the same figure at meg’ is a large cell (cut not quite through the middle) which has evidently had a similar origin to meg; for from its size it can not be confused with a neighboring cell of any germ layer. It contains coarse yolk, and on account of its irregular outline, judging from earlier instances, it has probably undergone division by amitosis. Fig. 75.—Detail of extra-embryonic region of embryo of fig. 73. e. Ectoderm; 1, mesoblast; meg, eg!, gigantic yolk-cells ; v, vacuole; ye, yolk-entoderm. ADDITIONAL EMBRYOS OF THIS PERIOD. A second embryo of this period, 7. ¢., prior to the breaking through of gills and mouth, is shown on plate vn, figs. 42, 42* and 42”, and on plate vim, fig. 42°. The present specimen is badly bent in its trunk region, but in other regards it may be readily compared with the earlier stage, plate vi, fig. 41. The chief advances include: (1) the modeling of the trunk, in whose hindmost region only appears the former flattened condition; (2) the appearance of auditory sacs (az); (3) the model- ing of optic vesicles (04); the protrusion of the forebrain region into a frontal knob (£). The general shape of the head, as shown in dorsal view, already suggests the adult condition, in spite of the small size of the embryo. This now measures only 2.5 mm., not allowing for the bent trunk region. The tail at this stage protrudes beyond the rim of the blastoderm, its tip budding out like a knob beyond the flat- tened caudal eminence. About 25 somites are present. A third embryo, plate vu, figs. 43 and 43°, shows over sixty somites, and gives us a picture of the young Chimera at about the end of the first month of incubation, In this stage over sixty somites are present, and the tail bud has EARLY EMBRYOS. 99 grown out conspicuously. The broad flattened trunk terminal of the preceding embryo is here represented, and at a the anal region, a point anterior to which the number of somites corresponds in a general way to that in the earlier stage. Noteworthy advances include: (1) A more definite modeling of the regions of head and trunk. The latter has now lifted up above the surrounding blastoderm, and the head (including the chin region) has separated from the yolk-wall. (2) The gill-slits are now conspicuous, although, as sections show, they have not yet broken through; we note that the spiracular slit s, evidently the equivalent of g’ in the former stage, is of considerable size; behind it occur three prominent depressions and the trace of fourth and fifth. (3) The appearance of pronephros and pronephric duct; the pronephros itself is situated at the plane of the ninth, tenth, eleventh and twelfth somites, as can better be seen in the transparent preparation in the same embryo, plate vu, fig. 43°. (4) The knob-like terminal eminence of the head region has greatly increased in size. A fourth embryo of this period is shown as a transparent preparation in plate vil, fig. 44. It contains a greater number of somites than the preceding, over 80 as opposed to over 60, but in many regards it appears to be less advanced in devel- opment. Thus we note that its head region appears somewhat less mature than in the former embryo; the chin is less definitely established and so also the gill-slits are shallower and the optic and auditory vesicles and the pronephros less definite. The tail, moreover, is less pointed, even bulbous where the terminal growth is taking place. In this stage we note the presence of a conspicuous postanal gut. The details of the vascular supply of the gill region are well shown; the spiracular artery is conspicuous, and, further hindward, we observe the duct of the pronephros ( Azd) and the postanal gut Aag. A series of characteristic sections of this stage is given in figures 76 A-N. These show a general correspondence to the conditions of the young shark. In fig. c the premandibular head cavity (fv) is shown; in p the mandibular (#7). In this section also we observe that the mouth has not yet broken through. In later sections, as in E, F, H, I, and k, we note that the gill-slits have not been com- pleted; fusions of the gut wall with the ectoderm have, however, occurred. We note in section M, passing through the pronephric tubules, that the relation of these structures corresponds closely to that in the young shark. A subnotochordal rod, conspicuous in the earlier stage, is here represented only in a rudimentary condition, as in N; the gut has separated from the notochord and the main vascular trunks now appear in the region formerly occupied by the subnotochordal rod. The present stage corresponds closely with that of the shark in which the mesoblast bounds a continuous myo-, nephro-, and splanchno-ceele. In Chimera, however, continuity in these regions is less clearly marked, a feature which evinces greater develop- mental specialization, 7. e., in masking an archaic condition and preparing the way for the prompter growth of structures useful to the young fish. 100 CHIMAROID FISHES AND THEIR DEVELOPMENT. THE RELATION OF THE BLASTODERM AND YOLK AT THIS STAGE, As already noted (p. 58), the egg of Chimera has by this time undergone a process of fragmentation. The bulk of the egg subdivides in the direction of pro- ducing for the embryo nutriment to be appropriated vza gills and gut; a single mass Figs. 76 A-N.—Transverse sections of the embryo shown in plate VII, fig. 44. m, Mandibular head cavity ; p7”, premandibular head cavity. YOLK-SAC OF EARLY EMBRYO. fot only, representing about one-tenth the bulk of the unsegmented egg, is reserved for the yolk-sac of the young fish. In the stage last described (7. ¢., of plate vu, fig. 44), in spite of the advanced characters of the embryo, the blastoderm has not increased vastly in size beyond that shown in plate v, fig. 38. It has, however, as we see in plate vin, fig. 47, constricted marginally, becoming cup-shaped, as it continues to envelop the small yolk mass. How far it has succeeded in inclosing the yolk is perhaps better seen in the details of the last figure, shown in figs. 47* and 47”. The relation of yolk and blastoderm is pictured in detail in fig. 77, a section passing through the blastoderm parallel to the long axis of the embryo. At the points #4 and wé’, the rim of the blastoderm comes in contact with the yolk ; above #é the blastoderm is thickened and spongy; for, as a sign that the body of the embryo lay adjacent, this region is richly vascular. Noteworthy here is a deep sub- marginal sinus (vs) whose posterior wall (c) is cellular. We have in this condition a physiological parallel with the submarginal space in ganoids, and more directly even with Kupffer’s vesicle in teleosts. On the ventral side of the blastodermic cap (on the left in the figure) the vascular sponginess is largely lost; and the blastoderm is thin, save only at its rim (w6’). And here in place of a deep submarginal sinus, a number of distinct blood- producing vesicles appear (ms’) scattered distally in a narrow zone of finely divided yolk (/y). From another standpoint, finally, the present section is noteworthy. For it shows that the entire yolk-sac is divided into masses which are largely separated from one another by a system of fissure-like vacuoles. Closer inspection shows nuclei scattered irregularly through these masses of yolk, and, everything considered, I think we can therefore justly conclude that the yolk-sac at this stage, in spite of its relatively large size, is a totally segmented structure comparable with the yolk-sac of Amia or Ichthyophis. In the present case, it is true, the yolk masses (blastomeres) show a condition of greater or less attachment to their neighbors, and each mass will usually contain more than a single nucleus. But even in this event, the comparison will, I believe, hold. In some cases the shape of the yolk masses is distinctly blastomere-like, as between the vacuoles (v) in the present section. Viewed from this standpoint, accordingly, Chimzera has retained a primitive embryological character, holoblastic cleavage; but we can hardly fail to observe that this character has lost much of its primitiveness inasmuch as the blastomeres are polynuclear and the intercellular spaces obviously adapted as reservoirs of nutriment. Fig. 77.— Section of extra-embryonic region and of upper part of yolk-sac of stage of plate VIII, fig. 47. ce. Cellular area; mb, 7b’, margin of blastoderm; ™8, marginal sinus; //, fluid yolk; ¥, vacuoles. CHIM#ROID FISHES AND THEIR DEVELOPMENT. LATE EMBRYOS. EMBRYOS FROM THE APPEARANCE OF GILL-OPENINGS TO THE TIME OF HATCHING. Four specimens illustrating this period are described on the following pages. The first of these, shown zz ¢ofo (plate vu, fig. 45), illustrates a stage in which all five gill-slits are clearly shown, but of these only the first has completely broken through, that lying immediately below the auditory vesicle. In front of this the spiracular cleft is faintly outlined. The entire head region is modeled clearly, and the anterior end of the embryo has separated from the blastoderm as far back as the region immediately behind the heart. The tail has greatly elongated and has entirely lost the bulbous terminal which we noted in the earlier embryo. The region immediately adjacent to the embryo is divided up into a spongy mass by many blood-vessels; we observe also that the blastoderm has almost com- pletely inclosed the attached yolk mass, a small yolk plug only being visible at the hinder end of the yolk-sac. This condition is shown in plate vii, fig. 48. Here, through the rim of the blastoderm one can faintly see the extent of the submarginal sinus which was noted in the preceding stage. From it now extend many vessels, as indicated in the figure. The region of the yolk plug is figured in plate vm, fig. 48°, as viewed under a dissecting microscope. It shows an interesting condition in connection with the holoblastic behavior of the yolk; for a number of irregular masses are visible, outlined, it appears, by vacuoles, and suggest yolk-filled blasto- meres. It will be observed, however, that the contours of the yolk masses are less definite as they approach the irregular rim of the blastoderm. (C/ fig. 77.) Sections of this stage are shown in the adjacent figures. In the first (fig. 78) the mouth (¢. ¢., its hinder portion) and auditory vesicles are traversed; the mouth has not yet broken through nor has the neighboring gill-slit, the hyomandibular. We note that the auditory vesicle is now a thick-walled sac opening broadly at the surface; that a subnotochordal rod is present; that the brain wall in this region (hind- brain) is remarkably thick and asymmetrical, and that the fifth ventricle is corre- spondingly reduced in diameter. A section through the mid-trunk (fig. 79) indicates that in this region the trunk is spread out more widely than in the corresponding or, in fact, in any stage in the shark. The splanchnoccele (sfc) is of great size, and its walls, both splanchnic and somatic, contain large spaces. The myoccele is virtually obliterated, although its margining cells have not fused across its earlier opening into the gononephroceele. The last region is not clearly demarked; at Az a pronephric tubule appears in the position usual in elasmobranch. At d/ the early condition of the dorsal fin corresponds closely with that of a shark embryo. A second embryo (plate vu, fig. 46) slightly older than the preceding, was one of the specimens received from Dr. Wilbur. It had with it only a small fragment of the blastoderm, and at the time of preservation the embryo appears to have turned in a position nearly transverse to its usual one. At this stage the tail protruded widely over the rim of the blastoderm, and it follows, therefore, that, probably as an individual variation, the blastoderm has not as completely inclosed DETAILS OF LATER EMBRYO. 103 A more detailed examination of this embryo the yolk as in the former specimen. shows that two gill-slits have broken through. The mouth, moreover, is more nearly completed, the mandible appearing and the visceral region having a more advanced The pronephros is conspicuous. The pectoral fin is present as a longitud- inal dermal ridge. The tail, judging from its twisted condition, is evidently capable of active movements. This, however, in its detailed structure, as shown in a trans- parent preparation (plate vin, fig. 46°) is still distinctly immature; its tip retains a neurenteric canal (zc), and a postanal gut (fag). In the latter the irregularity The present embryo measured about 20 mm. contour. at the point x is probably artifact. in length. eo) Fy, g ANN? eo S5e P avAY 1) a 1% Hc ERE . if Y co me UEHALG Best, ot aco: NA eon a : , Gs Ae Ly Fig. 78.—Transverse section passing through the posterior head region of embryo of plate VII, fig. 42. At the right the section traverses an auditory vesicle and the hyomandibular evagination. The latter fuses with the ectoderm, which here invaginates, but no opening has as yet been formed. Fig. 79.—Transverse section through the middle of the trunk region of the preceding embryo. df. Ectodermal anlage of dorsal fin. pn. Pronephros. spc. Body cavity. hb. Hyomandibular evagination. sn. Subnotochordal rod. A third embryo of this stage is pictured in plate vim, fig. 49° to °. about 35 mm. in length, and was observed living. It was this embryo whose It measured capsule was taken accidentally on a trawl line during one of the writer's visits at As already noted, it was found developing in a creamy fluid. Pacific Grove. When placed in a watch-glass, its general position and color were as here repre- sented. It lay for a while on its side, its diminutive yolk-sac extending outward from the body and the delicate tail region showing constant undulatory movements. Most conspicuous were the bright-colored vessels on the yolk-sac, which outlined a The visceral cavity showed red through vitelline circulation obviously shark-like. the delicate wall, and in the gill region there were prominent bead-like dilatations, brilliant in color. One notes the bright red spot under the eye, which was later 104 CHIMAROID FISHES AND THEIR DEVELOPMENT. found by sections to represent the spiracle.* Further details of the gill region are given in fig. 49°. In the various figures given of this embryo we note a number of advancing structures: (1) The eyes are now well formed, protrude widely from the head, and are provided with a conspicuous lens. (2) The region of the snout shows distinct modeling. Olfactory pits are present and are separate from the rim of the mouth. The snout region, it will be seen by reviewing the preceding figures, notably plate vu, figs. 44-46, does not cor- respond to the greatly dilated eminence which forms the cap-like knob surmounting the head. This appears rather in the region of the forebrain, and the writer does not, therefore, agree in the conclusions of Schauinsland (who, however, it will be borne in mind, examined Callorhynchus, not Chimera) as to the fate of this singular organ. It has, we suggest, the function of providing for the growth of the contour of the antero-dorsal head regiont rather than for the framework of the snout, as Schauinsland suggests. (3) The mouth has broken through, and its margins are thickened. It shows distinct movements, although at irregular intervals, in the living young. Between the rim of the upper jaw and the eye appears the spiracle, and in a remarkably anterior position contrasted with that of an elasmobranch. (4) The five gill-arches (plate vii, fig. 49") show well-developed lamellz on their anterior margins, and from these are produced the external gills. The latter extend outward on either side to a distance equal to about the diameter of the head between the eyes. The presence of dilated spaces, blood-filled, in the external gills has already been recorded. It is worthy of note, perhaps, that when the present specimen was preserved masses of yolk (plate vim, fig. 49”) were found adhering to the gill-filaments, a fact which may have some significance, since the blood- dilated spaces appeared at points adjacent to the attached yolk masses. In this stage, it may be added, the fifth gill-sht has not as yet broken through. (5) The fins are well established. The lobe of the anterior dorsal fin, however, shows as yet no trace of a spine. The paired fins are distinct lateral folds, much as in the young shark; in fact, the pectorals are even precociously large. It may be added that the metameral elements of the fins were conspicuous in the living embryo, since blood-vessels were present and appeared in a series of brilliant spots. The ventral fins are drawn together immediately behind the anus, and no trace appears of a clasping organ or of a third pair of limbs.{ The general arrangement of the fins is best seen in plate vim, fig. 49. (6) The yolk-sac, in spite of its small size, was perfect. Its structure is delicate, for at first its contour was smooth, but after the embryo had been kept living for several hours in sea-water, it was noticed that the surface of the sac *This is not in the position in which Solger (Morph. JB., 1876, pp. 219-221) expected it to appear, 7. e., behind the articulation of the mandible. +Possibly as a larval organ to protect the head when in contact with the wall of the egg-capsule. + Cf. T. J. Parker, Nature, vol. xxxIx, p. 625. With regard to the non-appearance of mixipterygia, which certainly occur early in Chimeroid ontogeny (c/. zfra, plate 1x, fig. 50f, also text), it is possible, of course, that the present embryo was a female. DETAILS OF LATER EMBRYO. 105 loosened at several points, giving the wavy contour noted in the figure of the entire embryo. The arrangement of the vessels is clearly shown, and one traces the posterior umbilical veins and the anterior vitelline arteries. At first sight the yolk-sac seemed to be attached anteriorly throughout the length of the heart region. Later examination, however, showed that a single stalk, albeit a very short one, connected the sac with the trunk in a fashion very much as in the young shark. (C/. plate vii, fig. 49°). (7) The lateral line sys- tem of organs is already established. At either side of the eye sensory pits are present and the backward growth of the lateral line could be traced as far as the anal region. DETAILS OF THE FOREGOING EMBRYO. In the sections, figs. 80-83, are shown details of the foregoing embryo. In the first of these (fig. 80), in the eye structures one observes the proportionally enormous size of the lens. Particularly noticeable, also, are the elaborately branching vessels lying be- tween the eye and the mid- brain (v, 7), a symptom again Fig. 80.—Transverse section through the eye region of the embryo shown in of the embryo’s precocious plate VIII, fig. 49. growth. In the following Vv, Vy Branches of anterior internal carotid. section (fig. 81) the spiracle (s) is seen to be continuous. Sections through an external gill-filament show, even more conspicuously than in shark, the presence of both vein and artery (a, v). In the same section we observe one of the blood-filled dilatations (d@), which have already been commented upon. This appears at or near the end of the gill-filament. In fig. 82, a section through the pelvic region just anterior to the anus, we observe on the right side the opening of a segmental duct at sd. Beside it, at mé, appears a mesonephric tubule. On the opposite side of the body a corresponding tubule, wz/, opens directly into the body cavity. One observes in the same section a dilated caudal vein at cv, and above it the caudal artery. In the section, fig. 83, we observe that the unpaired fins are already well established and that they are made up largely of mesoblast. The caudal vein and artery appear as before, and the section traverses numerous muscle plates. 106 CHIMZROID FISHES AND THEIR DEVELOPMENT. The latest embryo in the writer’s material, one of the specimens secured by Professor Wilbur, measured 51 mm. in length. Its age was said to be six months. (Plate 1x, figs. 50 and 50° to ®.) It is decidedly like the adult Chimzra, as can be seen from the figures; it has well-established snout (in which sensory grooves and == pits appear), paired and unpaired fins, and clasping organs, show- ing that the present specimen was a male. On the other hand, two prominent embryonic characters still appear, viz, the yolk-sac (which in the present specimen is preserved only in part) and the external gills, a tuft of which is seen protruding from below the opercular folds. The external cills are shown in plate ix, hey 50°, Grandi Their degree of differ- entiation is indicated in fig. 50%, in which we note that in each fila- ment one of the compo- nent vessels is less con- torted than its neigh- Fig. 81.—Transverse section passing through the otic vesicles of preceding embryo. At the sides external gills are shown. bor, the filament thus 4, artery; C, dilated blood knot in external gill; 8, spiracle; V, vein. presenting a crinkly appearance when viewed under a low power. Occasionally a terminal dilatation is seen. It will be noted that some of the filaments attain great length, although in general they are fewer in this than in the earlier stage, a process of reduction having set in at certain points.* In lateral view this embryo shows fragments of yolk attached to its side and to its paired fins, a condition probably artifact, although deserving mention, since in the younger stage yolk masses were observed attached to the gills. Before making the present sketch, a portion of the opercular fold and the neighboring external filaments were removed. The sensory canals are well indicated; that of the lateral line has now passed down the side of the body and has entered the tail region. The mandible is well established. In plate 1x, fig. 50°, we observe the extent to which the opercular folds overlap the tuft of external filaments; we here observe also that the frontal clasping organ *Cf. also Schauinsland (of. c7¢., Taf. xvi). DETAILS OF LATER EMBRYO. 107 is long, narrow, and relatively of great size, suggesting its origin from an anterior fin spine, and interesting in connection with paleontological data (cf. figs. 132-137). In fig. 50%, an idea is had of the extent of the overgrowth of the opercular fold on the ventral side of the head, and here is shown also that the external gill-filaments arise only from the anterior wall of the gill-slit, and that the external filaments increase in length as they pass toward the middle of each flap. A detail of the ventral fin is shown in fig. 50°. Here the mixipterygium is but a further differen- tiation of the base of the ventral fin (c% plate vin, fig. 49°), and the anterior clasping organ (acl) evidently represents the fin’s anterior segmental elements (radialia) (cf. also fig. 112). The mouth region in this stage is noteworthy, since it shows that not only are the anterior and posterior dental plates (ad@/ and pdp) present, but also a series of other eminences which are best interpreted as rudimentary dental plates. Similar structures are now described in detail in the work of Schauinsland on Cal- lorhynchus (v. zzf7va). The present figure also indicates the early stages in the curious lip cartilages of the Chimeroid. They arise at the sides of the mouth and suggest at this 5, 97 Transverse section through the region of the ventral fins of stage the corresponding structures in preceding embryo. shark. In view of the recent work Hee a oe eee ee of Schauinsland and of the younger Fig. 83.—Transverse section through the tail region of the preceding Firbringer (Morph. JB., 1903, vol. ube XXXI, pp. 360-445), we recognize with interest the unpaired element at the mandib- ular symphysis which is held to represent the homologue of the basihyal of the hyoid arch. (C% fig. 111.) In commenting further upon this stage we note that in the eye the iris is well established, and that in the umbilical sac the yolk material is arranged in conspicuously concentric lamellz (plate 1x, fig. 50°). The Skull.—The skull at this stage may be compared instructively with that of a late embryo of Callorhynchus figured by Schauinsland in Taf. xvu, figs. 124, 125, 126, of. cet. The present figs. 84 a—p were, like the figures mentioned, prepared from wax-plate models. The embryo referred to by Schauinsland is more advanced than the present one, although the difference in age does not appear to be conspic- uous. On the other hand, the figures of a younger Callorhynchus shown in Schauinsland’s Taf. xvi, figs. 130 and 131, can not be compared satisfactorily with the present specimen of Chimera, for its skull was evidently far less mature, a large part of the model having been based upon outlines of procartilage. A study of the foregoing figures indicates that the skull of Chimera is, at a corre- sponding growth period, the more highly modified; the orbits are larger, the snout 108 CHIMAROID FISHES AND THEIR DEVELOPMENT. region is wider and more compressed, the palato-quadrate is reduced and _trans- ferred to a more anterior position, nor is it as distinct an element as Schauinsland figures it in the kindred genus. As further evidence of the more modified character of the skull of Chimera, we observe that the preorbital ridges are curiously flattened, forming together a transverse brow-plate in the young skull; and that the Figs. 84 A-D.—Reconstruction of skull of Chimzra embryo shown in plate IX, fig. 50. The model is shown in lateral, three-quarters dorsal, and caudal aspects. a-b, Anterior and posterior points at which the palato-quadrate element has fused with the cranium ; @¢, roof of auditory capsule; ¢/, ceratohyal; 0, fora- men through which the ophthalmic nerve passes out of the cranium; /08, foramen through which passes the superficial branch of the ophthalmic nerve; gh, basihyal; Ach, hypochordal portion of the basis cranii; 7”, hyomandibular; p, pharyngobranchial; pdf, palato-quadrate fissure; p70, preorbital process; 7/40, postorbital process; 7, median rostral cartilage; 8, spiracular cleft later retained as the foramen through which the hyomandibular branch of the seventh nerve passes to the under side of the skull; V-+ VJJ, foramen for fifth and seventh nerves. postorbital ridges are reduced in size. We note also the greater width of the cranium in Chimera and the lesser development of cartilage in the region between the orbits. In short, we can justly conclude that at corresponding stages the skull of Callorhynchus more closely resembles that of a young shark than does the skull of a young Chimera. The proportions in the case of Callorhynchus are distinctly DETAILS OF LATER EMBRYO. 109 shark-like, so also are its early rostral cartilages. In Chimera, on the other hand, the developmental processes, evidently abbreviated, produce larger orbits, larger auditory organs (from this is due the broadening of the skull noted above), coales- cence of foramina, and altogether a more mature modeling of the head.* In this form, moreover, we find in the hyoid arch more perfectly developed ceratohyal and greatly reduced pharyngobranchial elements. It follows, I conclude, in view of these and other evident specializations, + that one can not reverse the order of comparison and regard Chimera as resembling the more closely the ancestral type from which in turn Callorhynchus and sharks developed. COMPARISONS WITH OTHER CHIMAEROIDS. Before concluding the account of the later embryonic stages of Chimeroid, which we have hitherto based upon C. collte?, reference should be made to the conditions known in other genera and species. In Callorhynchus.—Schauinsland has already given many observations upon the young of Callorhynchus. It appears from his figures that there are little outward differences in the development of stages corresponding to those of plate vil, fig. 45, of Chzmera collet, and that of Callorhynchus in Schauinsland’s Taf. xu, fig. 105. Also there are but minor differences between the present plate vir, fig. 42, and Schauinsland’s Taf. xin, fig. 98. We may thus compare also the present plate vu, fig. 45, with Taf. xiv, fig. 107, also plate vu, fig. 43°, with Taf. xiv, fig. 106. In a later stage, contrasting Chimera in plate vi, fig. 49, with Schauinsland’s Taf. xv, figs. 116 and 117, we can not fail to note the more shark- like conditions in the Australian species, and this is even more evident if we contrast the still later stage of C. collie? given in the present plate 1x, fig. 50, with Schau- insland’s Taf. xv, fig. 121. Observe in this connection the less tapering tail of Callorhynchus, a more distinct second dorsal fin and the early appearance of the row of dorsal scales which suggest closely the conditions shown in Scyllium by Paul Meyer. In referring to the latest embryonic stages in Callorhynchus, outline drawings may be given of specimens preserved in the department of ichthyology in the British Museum, figs. 86-88, and in the Copenhagen Museum, fig. 89.f And these may in turn be compared with the outline of the young Callorhynchus, fig. 85, figured by Parker and Haswell in their Text-book of Zoology. An examination of these figures shows that the absorption of the yolk-sac takes place, as one would expect, while the embryo is still inclosed within the capsule. In fig. 85 the yolk- sac is of irregular shape, rather large, and the embryo still retains its external gills. In figure 86 the sac, still large, is somewhat bilobed, a condition which becomes *Cf. also in this connection the more advanced condition of the mixipterygia in Chimera (plate 1x, fig. 5of, and Schauinsland's Taf. xv1, fig. 120). +In morphological regards cf. the reduction of dermal defenses, great size of head, reduction of caudal region, differ- entiation of dorsal fin, specialization of clasping organs, modification of brain. $For the privilege of examining these valuable specimens the writer is indebted to Mr. Boulenger and to Dr. Winge. hie CHIMAROID FISHES AND THEIR DEVELOPMENT. Figs. 85-89.—Latest stages in the development of Callorhynchus within the capsule. Nearly actual size. 85. Callorhynchus “ antarcticus.”” Detail of figure given by Parker and Haswell. 86. Callorhynchus sp. Stage in which the yolk-sac is reduced in size. A, Outline of dorsal fin when unfolded. B, Detail of frontal clasping organ with surrounding row of dermal denticles. 87 and 88. Callorhynchus sp. Late stages in absorption of yolk-sac. The foregoing three figures are after specimens in the Bntish Museum. (Cf. p. 34.) 89. Callorhynchus “ antarcticus,” showing very late stage in the absorption of the yolk-sac. After sketch of specimen from New Brighton, New Zealand, preserved in the Zoological Museum at Copenhagen, “TARVA:” OF CHIMERA. ltl intensified in the later stage, fig. $7. In fig. 88 the sac is still irregularly bilobed and in figure 89, where it has been almost completely taken into the embryo, the anterior lobe is still present. Observe in connection with these figures that the shape of the sac is obviously correlated with the shape of the embryo inclosed within the capsule.* A further consideration of these figures leads us to conclude that in the latest stages of development the embryo of Callorhynchus rests on its side, and in this position the dorsal fin is observed to lie neatly tucked against the side of the body, the dermal web of the fin being folded under the depressed spine. So also the paired and unpaired fins are closely apposed to the sides of the body, the continuous dorsal and anal fins folding closely around the side. The dorsal fin folds over the trunk towards the left side of the embryo. An outline of its margin, slightly raised, is shown in fig. 86 a. In this stage the appearance of the frontal clasping spine is indicated in fig. 868. This corresponds obviously to the con- dition which is figured in a younger stage by Schauinsland in his Taf. xv1, fig. 122. In Chimera.—The only late embryonic stage known to the writer is the one preserved in the Jardin des Plantes and figured by Professor Vaillant in his ‘*Travailleur” report (1882), a specimen which the writer had the opportunity of examining through the courtesy of its describer. This specimen (fig. 90 4, B), probably of C. affints, was dredged in the Bay of Biscay, together with fragments of its egg-capsule. A small yolk-sac is adherent; this is of spherical form, and appears to have been delicately connected with the body of the embryo. It is possible, of course, that the present spherical form of the yolk-sac may have been the result of the specimen having been freed from the capsule, for under this condition the yolk-sac would probably have assumed its present shape. It may be noted that the surface of the sac was deeply creased with blood-vessels, somewhat as indicated in fig. 90 a. Noteworthy in this specimen is the great length of the hinder trunk (and tail) which, it will be seen, is proportionately longer than in Chimera colliet, and much longer than in Callorhynchus, and it is also to be mentioned that the long urostyle shows that the continuous dorsal fin could not have extended function- ally into this posterior region. Clasping organs are developed, and, as shown in fig. 90 B, they attain even now a considerable size, about one-third their adult (proportional) length. This condition is noteworthy as indicating again a precocious type of development, sexual characters having been differentiated, although the embryo is small in size and provided with a considerable yolk-sac. IMMATURE YOUNG. Four stages of ‘‘larve” of Chimera collied are shown in plates x and x1, to illustrate especially changes in outward form, proportions, and coloration. The specimens figured in plate X were secured by the A/éatross during its work on the Pacific coast, and were kindly placed at the writer’s disposal by the United States National Museum. The youngest specimen figured (fig. 51 and figs. 51° and °) was evidently lately hatched. It still shows the scar marking the point of intrusion *The irregular outline of the yolk-sac would, by analogy, probably be filled out if the living embryo were removed from the constricting capsule. I12 CHIMAEROID FISHES AND THEIR DEVELOPMENT. of the yolk-sac (plate x, fig. 51°, ys); its form at this stage is probably modeled somewhat differently from that of the latest stage of the encapsuled embryo; thus the membranes of the unpaired fin in the tail region are probably less marked than in the earlier stage (comparing the embryo of plate 1x, fig. 50) in which this mem- brane serves as an organ for carrying out the water used in the respiration of the gO Fig. 90.—Late stage of Chimera affinis (?). The yolk-sac is largely resorbed. The present is the type specimen of Professor Vaillant, and is preserved in the ichthyological museum of the Jardin des Plantes, bearing the number 42392. In spite of its small size (its total length is only about || cm.) it shows a well-developed mixipterygium (B). A detail showing the vascular supply of the yolk-sac is indicated at A. Fig. 91.—Detail of early Chimera monstrosa, showing larval coloration. After sketch of specimen in the museum of Tromsoe. Colors are indicated, @, ashen, ”, white. Fig. 92.—Detail of Chimera monstrosa indicating final larval coloration. After sketch of specimen in the museum in Copenhagen. encapsuled young. The advancing characters of the earlier young may best be followed by contrasting figs. 51, 52, and 53. The changes thus observed are: In proportions.—The head length of the embryo, measured for example anterior to the base of the dorsal fin, decreases as we ascend the scale; in the earlier stage it measures about 20 per cent. of the entire length, in the latest about 16 per cent. ; the eye alters little in size, but the region of the head lying below the eye increases notably; the shape of the pectoral fin changes progressively; almost as wide as high in the first figure, it becomes nearly twice as high as wide in the latest stage. So, “LARVA”? OF CHIMARA. 113 too, the unpaired fins change proportions notably; in the stage shown in plate x, fig.52, their width is much greater proportionately than at other stages. Ln shape.—The shape of the trunk undergoes noteworthy changes. In the stage shown in plate x, fig. 53, itis much longer proportionately than in the earlier and later stages. We note also that the dorsal fin (or rather that portion of it posterior to the first dorsal) changes from continuous to lobate and then again later to a lower and less lobate form, during progressive development. Ln color.—Pigmentation appears progressively. In the youngest stage the pig- mented areas are dorsal. In the stage of plate x, fig. 53, pigmentation is more marked on the sides of the body than at any other stage. The sharpness in the coloration of the distal margin of the dorsal fins is most conspicuous in the stage of plate x, fig. 52; also a distinct larval coloration is noticed in the pectoral fin, a well- marked color being present along the anterior margin of this fin and in the anterior portion of its dermal web. Observe also the distinct patch of pigment at the base of the dermal web in plate x, fig. 53. Noteworthy, further, is the progressive increase in the number of pigmentless blotches; few in fig. 51%, they become numerous in fig. 53°, and small and most numerous in the stage of plate x, fig. 54. Similar changes in coloration affect the region of the eye. A late stage in the development of Chimera collici may finally be referred to in plate xt. At this age the young fish has attained nearly mature size (7. ¢., about three-quarters of that of the adult), although it is still distinctly ‘‘larval.’’ Its coloration is darker (cf. fig. 1), making the small pigmentless spots more con- spicuous. The margins of the fins, on the other hand, are pigmented, and with these we may contrast the fin margins in the adult, figs. 1 and 2, especially in the latter figure, where we observe that the anterior rim of the paired fins, notably the ventral, are pigmentless. We observe also distinct changes in proportions from the earlier stages; the length of the fish anterior to the anal region is now scarcely more than one-half the total length; in the earlier stage figured it is less than one- third. In the present specimen, a young male, the ventral fins partly uncover the mixipterygia; the ventrals are small in size, surprisingly so when we consider the length of the entire fish. At this growth period the young of this species occur in schools and sometimes appear in shallow water. * In other species similar changes in colors and proportions are probably present in ‘‘larval’’ young. In one species, Chimera monstrosa, they are present in even a more marked condition. In a young specimen preserved in the museum of Tromsoe, to which the writer's attention was kindly called by Dr. V. Storm, the coloration was brilliantly marked. Although not larger than the specimen shown in plate x, fig. 53, it had developed dorsals sharply marked with black, pectorals with an ashen blotch and with a white anterior rim, a pattern which has been reproduced from a sketch in the present fig. 91. It is evident, moreover, that in C. monstrosa this stage is of brief duration; for in a second and equally well- *The present specimen was taken, together with 22 others, in a water depth of less than ro feet, near Port Wash- ington, Puget Sound, June, 1896, in a single haul of a herring seine. In this locality Chimera is rarely taken in shallow seines. The specimens measured from 30 to 40 cm. 114 CHIMHROID FISHES AND THEIR DEVELOPMENT. preserved specimen of this species—one which was examined in the Copenhagen collection—the colors had notably changed. The pigmented margins of caudal and postdorsal fins had become reduced to a dusky band, and the marking of the pectoral was limited to a mere fuscous blotch at the fin tip (fig. 92). The length of this specimen was but about two inches greater than the former one. From the foregoing notes we may justly conclude that Chimera undergoes a series of ‘‘larval” changes. That these are adaptive remains still to be proven, a verdict which, it may be remarked, applies equally well to many if not all the “larval” changes of teleosts, but the fact that such changes do occur in the hatched young is noteworthy in its bearing on the specialized nature of Chimeroid development. It is also, I believe, significant that the ‘‘larval”’ coloration of the young of Chimera monstrosa occurs at an earlier relative period than in C. colltez (2. e., that the distinctness of coloration, which in C. col/iec—a smaller species by the way—is shown in a specimen twelve inches in length, is attained in #onstrosa by the time the young measures but about seven inches), for this denotes that the structures of sostrosa are the more highly differentiated and that this species is of later origin. In another direction it contributes testimony as to the abbrevia- tion of developmental processes. ORGANOGENY. INTEGUMENT AND DENTITION. In the major problem of the position of Chimzeroids the evidence of scales and dentition claims an important place. For the question has been raised repeatedly whether the dentition of these fishes is fundamentally different from that of sharks, and whether the characteristic tritoral plates may not have retained primitive gnath- ostomal characters (Jaekel). And it has similarly been queried (Pollard) whether the present integumental defenses of Chimzroids may not prove the rudiments of a complete body armoring. We may accordingly review at this point the evidence in the matter of integument and teeth afforded by a study of the recent forms, both in adult and in embryonic condition. It has long been known that recent Chimeeroids retain shagreen-like structures. These occur in greater or less number (a) on either side of the median dorsal line; (4) in connection with sensory canals, especially in the suborbital region; and (c) in the male as organs of retention 7x copzo. (a) Shagreen-like scales on either side of the median line are most numerous in Callorhynchus, where they form rows, each including about a dozen scales, in three definite tracts, 7. ¢., in the head, between the first and second dorsals, and between the second dorsal and the caudal fin (figs. 93 A and B). In Harriotta they are smaller and less numerous. In Khinochimera they are tumid and uncalcified, occurring along the fleshy anterior margin of the caudal fin, obsolescent elsewhere. In Chimera they are rudimentary or absent. These scales occur, therefore, in a regressive series, at one end of which stands Callorhynchus, at the other Chimera; and it is significant, I believe, that a condition closely similar to Callorhynchus occurs in sharks, e. g., Pristiurus and Scyllium, as figured by Paul Meyer, who, SHARK-LIKE DERMAL DENTICLES. 10 es Figs. 93 A-F.—Dermal denticles of Callorhynchus. A. Dorsal aspect of young Callorhynchus “‘antarcticus” (Chili), measuring 16 cm. in length. The dorsal denticles are conspicuous ; their disposition and number is indicated. B. Dorsal aspect of well-grown Callorhynchus “‘antarcticus’” (Chili), “measuring 50 cm. in length. The dorsal denticles are reduced. In a specimen (Australian) measuring 92 cm. they do not appear. C, Isolated denticle from the back of a late ‘‘embryo."" < 42. After Schauinsland. D. Isolated denticle of a late ‘‘ embryo" (shown from side). > 42. After Schauinsland. E. Row of four denticles from the back (in front of second dorsal fin) of a late “‘embryo."" After Schauinsland. F. Enlarged denticle from similar situation in “* adult"’ specimen. % 21. After Schauinsland. a6 CHIMAROID FISHES AND THEIR DEVELOPMENT. however, does not refer to these structures in connection with Chimera (MT. Zool. Stat. Neapel, vi, p. 221 e¢ seg.). In further detail: in the dorsal scales of Callo- rhynchus, as Duméril and others have shown, the individual scales are furcate at their base, and the free points of the base project forward and embrace the pre- ceding member of the series, thus rendering the row of scales stronger and more compact (cf esp. Duméril, Garman, and Schauinsland). (Figs. 93 c-E.) It has further been shown by Schauinsland that these scales present notable shark-like features in their development; they first arise, like shagreen denticles, as an out- erowth of the derma; they then differentiate odontoblasts, by which in a centrifugal direction dentine is laid down; and at the end of the process a pulp cavity remains and a basal plate perforated by small nutrient canals. In Schauinsland’s words we further note that ‘‘in the latest embryonal stages the denticles, and especially their tips, acquire a greater and glassy transparency (vitrodentine), by which they become more and more differentiated from the substance of the (basal) plate. In short, developmentally speaking, the dermal denticles of Callorhynchus represent the most primitive scales which occur among living selachians. Through the presence of a basal plate perforated by dentine tubules, they suggest the scales of the oldest palaeozoic selachians.’’* Schauinsland illustrates his foregoing remarks with two excellent figures, one showing in section an early stage (of. c¢., Taf. x1x, fig. 139) in the development of the dermal cusp, the other a late stage in which the cusp presents a thick cortical layer of vasodentine (zé7d., fig. 140), projecting its tip beyond the epidermis. On the basis of the foregoing observations, therefore, we may conclude that, as far as these body scales are concerned, Callorhynchus is distinctly shark-like; there is not the slightest embryological evidence that this Chimzroid had ever ganoid- like scales. We might even, I think, go farther than Schauinsland, and point out resemblance with more typical selachian conditions; for this author, while main- taining that ‘‘the epidermis takes no part in the formation of the denticle,’’ and admitting that he ‘‘was unable to demonstrate the presence of enamel,’’ shows nevertheless in his earlier figure that the cells of the epidermis are arranged over the dermal papilla in a wayt that is more than suggestive of an enamel organ— an emphatically shark-like character; and we may further conclude that the base of the denticle perforated with tubules is not merely characteristic of denticles of Silurian forms but of later sharks as well (c/ Rose, ve trabeculo-dentine in Anat. Anz., 1897, p. 36). In connection with the presence of scales arranged near the dorsal line, it has already been commented on (Schauinsland) that these structures are relatively more prominent in the late embryo than in the adult, although no explanation of this phenomenon has yet been advanced. I may accord- ingly hazard the opinion that they have been retained in this position owing to their importance as larval organs—possibly for the purpose of enabling the well- *Cf. Rohon, J. O., Uber fossile Fische vom oberen Jenissei, Mem. Acad. St. Petersburg, 1889, and Die ober- silurischen Fische von Oesel, JZem. Acad. St. Petersburg, 1893. He refers to denticles of Thelodus-like forms which the recent researches of Traquair have associated with fishes which are in some regards shark-like. +Cf. e. g., Jentsch, B., Beitr. z. Entwick. u. Struktur d. Selachierzihne. Leip. 1897, fig. 6. SHARK-LIKE DERMAL, DENTICLES. 17 developed young to maintain its position in the egg-capsule, possibly also for the purpose of protecting the delicate dorsal fin, 7. e., by keeping it from rubbing against the walls and the roof of the capsule, during the movements of the young fish. According to this view the dorsal scales of the young Callorhynchus after the time of hatching are to be looked upon merely as rudimentary organs.* And it may be pointed out, in this connection, that when these enlarged dorsal scales are developed in shark embryos they appear only tx those forms tn which development takes place tn ege-capsules.t (6) Small dermal plates have long been known to occur in Chimzroids in connection with the sensory-canal system. Pollard makes a special reference to those situated in the suborbital canals, and Schauinsland gives the following notes upon them (of. c7t., p. 13): In the immediate neighborhood of the mucous canals—I have investigated those only situated on the head—there also occur dermal calcifications. I find there (in transverse section) in the floor of the canal (in the neighborhood of the skull) a large plate, and in addition at its sides and bounding it four to six conical caps of dentine. The development of these is like that of the denticles, save that the plate contains no pulp cavity, while the lateral small hard structures present such a cavity, if indeed only in a narrow form, and filled with few cells, whereby they come to resemble a small denticle. These calcifications are also probably only the rudiments of former dermal denticles which came to sink down at the same time that the epidermis was invag- inated to form the mucous canals; in this process they lost their primitive form and underwent degeneration. In adult, and especially in a number of fossil Holocephali the slime canals are surrounded by a great number of closely compressed rings formed of calcified and bony material ; these had their origin through a process of pressing together the single dentine-like bony caps noted in the embryo. In the matter, then, of the character of these plates in living forms, we may again conclude that they are equally derived from solitary dermal denticles, shark-like in type. There is no evidence, on the side of embryology at least, that these plates result from a breaking down of larger structures. It is only necessary to note further that these structures in Callorhynchus are most marked in their likeness to the selachian condition, and that they are least marked in the case of Chimera. { (c) In all recent Chimzroids numerous denticles are present in the male, 7. ¢., on the frontal clasping organ, on the mixipterygium, and on the anterior pelvic clasping organ. These denticles have a transparent, almost glassy character. In the frontal clasping organ of Callorhynchus, they occur not only at the tip of the organ itself, but also proximalward and at the front and sides of the depression into which this clasping organ fits; but in the other genera, the denticles are limited only to the tip of this organ. It follows, accordingly, that in Callorhynchus appears again a more shark-like character, 7. ¢., a greater number of denticles spread over a larger extent both of the clasping organ itself, and of the sheath into which the *In a specimen of Callorhynchus ‘‘ antarcticus’’ (Australia), measuring 92 cm. in length, the dorsal denticles have disappeared. +The tubercles in the encapsuled Scyllium (de Philippi, Paul Meyer) may well have a similar function. By Paul Meyer they are described (of. czt., p. 224) as rudimentary organs, viz., the remains of the ancestral annelidan parapods! tAs to the condition of these dermal elements in fossil Chimaroids, v. figs. 138 and 139; by evidence thus obtained the conclusion becomes definite, 7. e., that the shagreen of recent forms has been greatly reduced from a_ condition altogether shark-like. ae CHIMAROID FISHES AND THEIR DEVELOPMENT. clasping organ is usually depressed. In this connection we call to mind the great size of the clasping organ in the young Callorhynchus, suggesting its origin from an anterior fin spine (cf. figs. 132-137); its small size in Chimera on the other hand indicates the later derivation of this genus. This induction is also supported by a study of the clasping organs connected with the ventral fins in the antero-pelvic clasping organ of Callorhynchus. We observe that this structure is furnished with many dermal denticles—4o or there- abouts in the case of Callorhynchus antarcticus, according to Duméril, whereas in the various species of Chimera and in Rhinochimzra the number is reduced, varying usually from about six to three. In the mixipterygium shagreen denticles occur plentifully. In the case of one arm of this trifid organ in Chimera collie? the denticles extend proximally as far as the base of the organ. In the other two arms the shagreen is limited to tracts near the tips. An abundant supply of these denticles is, however, present, repre- senting, in fact, tracts of shagreen. In Chimera monstrosa, on the other hand, the amount of the shagreen is less, a condition which furnishes another reason for regarding this species as the more modified. In RhAznochimera pacifica, as the writer has already noted (Jour. Sci. Coll. Tokyo, vol. x1x, p. 10), the shagreen at the tip of the mixipterygium is greatly reduced. In Callorhynchus, on the other hand, it is as abundant as in the case of Chimera colliet. DENTAL PLATES. These have always been the stumbling-block in comparing Chimeroid with sharks, for by only superficial comparison have the tritoral areas in the dental plates of Chimzroids been regarded as equivalent to the teeth or clusters of teeth in the shark. Nor has paleontology as yet been able to elucidate the problem, even to the degree in which it has thrown light upon the origin of the dental plates in the lung-fishes. In fact, as we shall later note, the study of the dentition of fossil Chimzroids leads us at the present time tono decisive results. The develop- ment of the dental plates might therefore be looked to to furnish evidence as to the nature of these structures. For it is well known that through embryology a flood of light has been thrown upon the mode of origin of the dentition of lung fishes. Accordingly, we conclude that one of the most important sections of Schauinsland’s memoir on Callorhynchus is devoted to the question of the mode of origin of the dental plates. Schauinsland’s account, indeed, is of such value in the present connection that I have been led to quote it in freely translated form (of. cz¢., pp. 13-16): In even their earliest stages the dental plates are laid down as distinct elements, 7. e., four above and two below, and there is at no time a definite indication that these are composed of simpler elements which have fused together. The upper anterior plates are certainly simple ; the remaining pairs, however, show along their hinder (caudal) border a somewhat trifid arrangement. In this region, too, the plates with their three ridges pass into a fold of the skin, and here their growth takes place. (No trace appears even in earlier stages of the median (unpaired) mandibular tooth which has been described in fossil Chimeeroids.) If we regard the three ridges as rows of teeth which have become fused together, they would have obviously a certain similarity to the dental plates of dipnoi or even of teleosts (e. g., Anarrhichas); and we THE ORIGIN OF THE DENTAL PLATES. 119 might accordingly regard the anterior plates as premaxillary or vomerine, although in the latter regard, 7. ¢., ve premaxillary and vomerine elements, we query whether we can justly introduce this comparison in the holocephali. On the other hand, if the comparison be a legitimate one, we might even go farther and regard the more median ridge of the large plates of the mouth-roof as equivalent to the fused vomerine teeth, and look upon the remainder of these plates as having arisen from fusion of the elements in a double row of palatine teeth. Of course, however, such an interpretation would be purely hypothetical. One is inclined to look upon the anlage of a dental plate as the product of a single and enormously enlarged dental papilla, circumscribed by a dermal fold, the induplicature of which is deepest at the posterior margin of the papilla. The first deposition of hard material begins at the outer surface of the papilla, and takes the form of a thin cap of dentine, soon, however, the tooth- substance appears below at the points where the plate is to come in contact with the cartilage of the head. And almost at the same time trabeculee and lamellae appear between, 7. ¢., in the substance of the plate, and produce a meshwork of spongy tooth-substance (pulp-dentine). The mode of origin of the plate resembles closely that of bone when derived from connective tissue (e. g., in Sphenodon). The mesenchyme cells in the papilla are collected together closely at certain points and become transformed into odontoblasts, and from these, peripherally, the dentine takes its origin. It may be remarked that the dentine is sometimes laid down in an irregular way, with branching processes, its canals ramifying, unlike the parallel canals of true dentine. Occasionally trabecules of the dental mass, especially in older individuals, show a somewhat lamellar structure, and those which are first differentiated, that is, those lying inner- most, are distinguishable from the later lamellae by their capacity to become stained. As already noted, the entire dental plate is finally formed of a meshwork of dentine-like material, whose trabecules thicken with age, so that finally the plate attains a high degree of hardness. The spaces between the meshwork represent collectively a large, greatly branched pulp cavity, whose cells in part have retained their former reticular arrangement, in part have become odonto- blasts, as far at least as they become opposed to the trabecules. In the various ramifications of the pulp cavity blood-vessels are often present. Enamel is not deposited ; nevertheless the epidermis cells must have a certain influence on the character of the dentine, since the dentine becomes glassy in character when in contact with the epidermis, but remains unchanged when- ever the epidermis is lacking. The dental plates are fastened to the head cartilage by means of a firm layer of connective tissue, which indeed here and there may enter the substance of the plate, and for still stronger attachment claw-like outgrowths arise from the base of the plate, especially from its anterior and lateral portions. Finally, I must refer to the presence of remarkable structures in the dental plates, which occur only within the ridges above referred to. These take the form of a chalky mass, which appears in cleared preparations and can be traced throughout the entire length of a dental ridge ; it is partly inclosed within the meshes of the trabecules of the dentine, and by these partly again broken up into rounded masses and processes. In transverse section this chalky mass presents the appearance of a section of a many-rooted tooth, while in longitudinal section its substance appears continuous, although greatly fenestrated. A more detailed examination shows that we are here dealing with an especial variety of dentine; that is, differentiated from odontoblast-like embryonic cells, whose processes grow deeply down and develep canals which from their parallel arrangement recall strikingly those of typical dentine. In any event, the material in question can more accurately be designated as dentine than can the remaining spongy substance of the dental plate. From the latter it is also distinguished in remaining colorless after treatment with the usual stains for bone, and especially in retaining permanently, even in the grown Callorhynchus, its soft and uncalcified condition. It may be noted that this soft dentine is not present in the youngest embryonic stages; it appears shortly after the caudal ends of the plate are established and extends gradually from a hindward into a more anterior position. It has nothing to do with the origin of the hard structures of the plate, since it appears after these have been laid down. It usually appears somewhat deeper than the outer surface of the plate ; later it often comes to lie in close contact with it, and even extends thence inward, not infrequently coming to be associated with the remaining meshwork of the dentine. What the significance of this structure is remains in any event doubtful, and only with reserve do I express the opinion that these soft masses of dentine represent the rudiments of former rows of single or iH CHIMHROID FISHES AND THEIR DEVELOPMENT. already fused teeth, which had primitively passed from behind and taken up a position on the dental ridges. In the case of these teeth (a similar process occurs in the ontogeny of Cerato- dus) spongy dentine, or bone-like masses, were differentiated in the course of phylogenetic development, and these became finally of greater value for purposes of nutrition than the separate teeth ; and they accordingly fused together, overgrew the teeth, and in the end completely enveloped them. And since the teeth had no longer their primitive function, they came to lose their limy structure and degenerated, remaining in the condition in which we see them to-day. While their arrangement in three rows possibly indicates an alliance with the higher forms, their mode of successional growth suggests the origin of the rows of teeth of selachians. The results of the foregoing observations of Schauinsland, it will be seen, are disappointing to those who on a frior? grounds anticipated that the dental plates of Chimzroids would in the ontogeny of recent species be found to be formed of the coalesced bases of separate tooth elements, which, in their turn, would of course be homologous with those of sharks. One may, nevertheless, I believe, take a somewhat more hopeful view of this problem, in view of the evidence above provided. In the first place, however, in order that there may be a better understanding of the terms of the problem, it will be found expedient to review briefly the characters of dentition known among the more prominent types of recent Chimezroids, for there is room for the belief that Callorhynchus, in spite of its many archaic features, may prove to have modified the conditions of its dental plates, or at least parts of them (the ‘‘tritors’’), more completely than some of the other forms. To this end we may compare the dental characters of Harriotta with those of RKhinochimeera, as representing extreme types in Chimeeroid dentition. In fig. 94 4 are shown in Harriotta the dental plates and the roof of the mouth; in fig. 948 the dental plates, tongue region and floor of the mouth, and, in figs. 94c and 94p, corresponding regions are shown in Rhinochimera. Contrasting these forms, we notice that in Harriotta the dental plates are studded with peg-like eminences, some of which, both in the upper and in the lower ‘‘jaws,’’ form together tumid tracts or ridges. These peg-like eminences, ‘‘tritors,’’ are found to pass deep into the substance of the dental plate; thus, where the plate is flattened and more or less transparent, as at the anterior margin, the peg-like structures are seen to pass backward, forming long and narrow cores. ‘These are evidently of hard, bony texture, for they often stand out from the plate-like ridges when the intervening basal portion of the plate is worn away. We also observe that the adjacent mucous membrane of the roof, sides, and floor of the mouth is studded with distinct papilla. These, it will be seen, correspond to the ‘‘tritors,’’ in size, prominence, and closeness in arrangement, and may, I believe, from the evi- dence of similar structures in the mouth region of various fishes, be looked upon as homologous with tooth-forming papille.* It will thus be observed, as in figs. 94 4, 948, that they occur within the stomadeal region; they are absent in the dorsal wall of the pharynx; they are present, however, on the floor of the mouth, and are *In a recently published paper on the oral and pharyngeal denticles of elasmobranchs (Proc. Zool. Soc., 1905, I, pp- 41-49), Imms gives reasons for homologizing similar structures in sharks with teeth. He did not, however, find the papillae present in the specimen of Chimera monstrosa which he examined. COMPARISON OF DENTAL PLATES. I21 continued along the floor and sides of the pharynx. In Rhinochimera, on the other hand, the dental plates have become thin and have developed hard cutting edges, giving the mouth an almost beak-like appearance. In the plates tritoral areas are reduced to thread-like elements, so delicate that they become difficult to Fig. 94.—Dental plates, and roof and floor of mouth of: A, B, Harriotta raleighana. C, D, Rhinochimeera pacifica. distinguish even in the hard anterior pair of ‘‘vomerine’ plates; and in con- nection with the obsolescence of the tritoral areas, it is now interesting to observe a ereat reduction in the number and size of the papillae of the mouth. Thus on the roof of the mouth there occur no papilla throughout the wide tract immediately behind the palatine plates. 22 CHIM4EROID FISHES AND THEIR DEVELOPMENT. XN Cy IA Figs. 95 to 103.—Dental plates of Chimeeroids. Fig. 95, Callorhynchus “‘callorhynchus” ; 96, Harriotta raleighana; 97, Chimera phantasma ; 98, C. mediter- ranea; 99, C. monstrosa; 100, C. mitsukurii; 101, C. affinis; 102, C. collie; 103, Rhinochimera pacifica. COMPARISON OF DENTAL PLATES. 123 Comparing now a series of the dental plates of Chimzroids (figs. 95 to 103), we may first place side by side those of Callorhynchus and Harriotta (figs. 95, 96). It then becomes clear, I think, that the ridges in the dental plates of the former genus correspond to the clustered tubercles in Harriotta, a comparison which is well borne out by the embryological studies of Schauinsland, for it will be recalled that the separate ridges of Callorhynchus were shown to consist of a mass of chalky centers in which the lamella of dentine were parallel to one another, although their substance, as was noted, remains uncalcified (cf fig. 105). A similar state of affairs, it may be remarked, occurs in the posterior part of the large tumid ridges in Harriotta, for these ridges and their tritors can be readily sectioned. On the other hand, the anterior eminences of the same tumid ridges are found to be much harder than the neighboring bony plate, and may with less question, therefore, be regarded as representing true teeth. Indeed, it is, after all, a matter of minor importance that these tritoral elements have never hardened in the case of Callorhynchus; for when we consider the thickness and hardness of the surrounding bony plate, we are led to conclude that this may well have usurped the function of the separate denticles, and that these therefore remain undeveloped. The same rudimentary condition is probably true of the minute tritoral points which one finds along the anterior margin of the vomerine plates in Callorhynchus. Continuing the comparison, one can with fair definiteness understand the relations between the dental plates of such forms as Harriotta and Chimera phantasma. For, in the latter, the wide tritors at the base of the palatine and mandibular plates (fig. 97) are evidently homologous with the clustered tubercles in Harriotta. In C. phantasma, however, the crushing surfaces of the plate are smoother and less extended. In C. mediterranea (fig. 98) the dental plates have become more oblique (slanting) in their manner of attachment, the posterior flange of the plates intruding deeply below the mucous fold in the roof of the mouth. In C. monstrosa (fig. 99) the tritoral areas of the palatine plates are less numerous, while in the mandibular plates they are more abundant, but show less clearly the peculiar banded structure of the foregoing specimen. In C. mectsukurii (fig. 100) the conditions are not widely different from those in the species from the Mediterranean. A peculiar arching appears in the palatine plates, and the ridges on the posterior face of the mandibular plates, although smaller, are more con- spicuous. In C. affinis (fig. 101) the proximal tritoral areas were not observed, and altogether the grinding margin of the palatine and mandibular plates was narrower. In C. collie? (fig. 102), while the tritoral ridges on the posterior faces of the pala- tine and mandibular plates are (usually) conspicuous, the grinding edges of these plates are exceedingly narrow. And in Rhinochimera pactfica (fig. 103), finally, we attain a condition, as we have already noted, in which the tritoral areas are reduced to obsolescence, the entire distal margin of the plate functioning as a cutting edge. From what has already been said regarding the dental plates in C. colliec (p.19), I think we may safely conclude that a wide range of variation occurs in the dental plates of Chimeroids. Thus the tritoral structures may vary in number, size, I24 CHIMAROID FISHES AND THEIR DEVELOPMENT. and arrangement; in fact, one might even go so far as to maintain that from a large series of dental plates of one species of Chimera one might obtain variants which, separately considered, would be placed with other species. Moreover, from the function of these crushing plates, it is not unnatural that marked differences should appear in specimens of different ages and from different localities (¢. 2., from those individuals which have lived upon different food material). In short, we incline to the belief that changes in the dental plates of Chimzeroids do not predicate as wide divergences in lines of descent as one would naturally expect. From the standpoint of adaptation, furthermore, admitting the extreme value of physiological adaptation in dental plates within the limits of the present group, we obtain a suggestion why phylogenetic changes are not recapitulated favorably in their devel- opment. Ina form, for example, like Callo- thynchus, in which the basal (trabecular) por- tion of the plates has become greatly devel- oped in the adult, we naturally expect that there will be less oppor- tunity—shall we say : 5 : Fig. 104.—Callorhynchus callorhynchus. Dental plates and neighboring mouth parts of late time?—for the tritors embryo (about 110 mm. long). After Schauinsland. to recur in develop- Fig. 105.—Callorhynchus. Detail of middle ridge of mandibular dental plate of specimen ment ina separate and slightly younger (about 95 mm. in length) than the preceding. The dental ridge is seen see. : eh as a transparent object. After Schauinsland. finished form. they Fig. 106.—Callorhynchus. Dental plates of “larva’’ measuring about 16 cm. After spec- do appear, they appear imen in museum of Columbia University. regu larl y on ] y in ‘‘family”’ or in ‘‘generic” form, soon to be remodeled or erased. ‘Thus we find in Callorhynchus, according to the figures of Schauinsland, that these tritors do occur in later embryonic stages (fig. 105), although this author does not refer distinctly to the relation of dermal cusps to tritors in Chimezroid plates. Following briefly the problem of the dentition of Chimezroids, we may again refer to the presence of numerous papillz in the mouth region of these forms. For, by analogies in other fishes, these structures may well represent rudiments of discrete denticles. It is, therefore, of particular interest that in the case of Callorhynchus, where the dental plates are heaviest and largest, we find a corresponding increase in the size of the papillae. For it may be suggested that papilla which have become calcified either singly or in groups, have retained their dentitional (and ancient) trend in evolution, while those which remain soft have survived because they have undergone a change of function. The similarity in dental and non-dental structures is shown strikingly in the roof of the mouth of Callorhynchus (fig. 104), after Schauinsland. That shown in the roof of the mouth of Chimera (plate rx, fig. 50°), although not DENTAL PLATES OF LARVAL CHIMAROIDS. 125 as conspicuous, is none the less suggestive when we compare it with the strictly tritoral conditions shown in Harriotta, fig. 94 4. DENTAL PLATES OF LARVAL CHIMAROIDS. Furthermore, if one compares the dental plates in Chimzroids of different stages of growth, one is impressed with the evidence of larval adaptations. The plates of a Chimeroid recently hatched (C. collec) are surprisingly large in size, but instead of spreading out in 107 the form of crushing plates, they protrude marginally, forming Py relatively high edges and 108 function evidently in cutting. Moreover, the substance of these juvenile plates is glassy (cf Schauinsland, ve vitrodentine) rather than horn-like or chalky, and their margins are sharp and brittle. It is clear, therefore, that the plates grow during earlier stages, notably at their outer or secant margins, and it is a probable conclusion that this condition of growth is corre- lated with the special feeding requirements of the young. In later stages the plates broaden and thicken, the secant edges become less and less conspicu- : Figs. 107-109.—Haniotta raleighana. Dental plates (somewhat diagrammatic) gradually the tritoral *°* pie . ous, and st adua J a of three individuals measuring respectively 10, 49, and 64cm. At A the areas appear. The Jatter,, at lateral aspect of the yomerine and palatine plates is given. least in the species examined, are developed first vaguely, in extended tracts or ridges, and in these there later arise discrete eminences. This is the condition indicated above in Callorhynchus (cf also with fig. 95 the juvenile plates shown in fig. 106); it is even more marked in Chimera collie, and it is to be observed in such a form as Harriotta. Of the last form we may introduce sketches of three stages of the dental plates.* In the first (fig. 107) the plates are frail, although well formed, and with secant prosilient edges; they have already ridges outlined and their clouded color (especially in the palatines) is probably due to the presence of vitrodentine. In the second stage (fig. 108) a number of distinct tritoral emi- nences appear. And in the final stage (fig. 109), the largest specimen of Harriotta recorded, the tritors are well differentiated. *For the privilege of examining this unique material the writer is indebted to the United States National Museum. He wishes especially to express his thanks to its Assistant Secretary, Mr. Richard Rathbun, and to its assistant curator of Fishes, Mr. Barton A. Bean. 126 CHIMROID FISHES AND THEIR DEVELOPMENT. From the foregoing characters in ‘‘larval” dental plates, and they are certainly in the general line of Garman’s observations,* we conclude that among the many specializations in the young Chimeroid may be included a larval dentition, 2. @., preceding the appearance of tritors. It may also be remarked that the tritors themselves, when they come to appear in the different forms of Chimeeroids, occur in point of time in interesting sequence. In Callorhynchus they appear in the embryo (95 mm.), while it is still encapsuled, but they fail to develop into typical structures; in their place there appear calcified ridges representing collections of tritors. In Harriotta tritors become functional at a period shortly after hatching, and from this time onward increase both in size and number. In Chimera they occur at a later period, develop slowly, and even in the adult are relatively few, and the plates themselves early develop secant margins. In Rhinochimera, finally, they appear only in the adult, and even then in rudimentary form. In the Chim- eroid series, there is thus, I think, such evidence of progression, even in recent forms, that we can hardly assume with Garman that from a condition like that in Rhinochimera arose the dental plates of the other genera. On the contrary, in the case of Rhinochimzera we are dealing evidently with a terminal form, one in which the tritors fail to develop perfectly even in the adult. + CONCLUSIONS CONCERNING THE DENTAL, PLATES OF RECENT CHIMAROIDS. A comparison of a series of the dental plates of recent Chimezeroids, as we have seen, strengthens the view that these structures are compound, 2. ¢., formed of separate denticle-like elements, homologous with the dental plates of certain sharks, e. g., Cestracionts. The tritors, according to this view, represent dental eminences, simple or compound. But more doubtful is the homologue of the dental plate itself. It may represent either the fused bases of teeth like the Cestraciont, or a structure entirely saz generis, 7. e., fused by a hardening of the connective tissue accumulated around the bases of the true dental plates. According to the observa- tions of Schauinsland the embryological facts support more or less distinctly the origin of the tritoral ridges from many tooth-like eminences dentinal in structure. On the other hand, the same evidence tends to regard the substance of the dental plate itself as independent of the tritors. An examination of the larval dentition of Chimeroids throws, I think, a side-light on the foregoing discrepancy, for it is found *Garman, however, interprets these characters (Proc. New Eng. Zool. Club, rgor, vol. 1, pp. 75-76) not as larval- isms, but as primitive; thus, according to him “the teeth of Rhinochimera are of a much less differentiated form than those of any other of the recent genera of the group; that is, their later stages are more like the earlier, and presum- ably more like the teeth of primitive Chimeroids ; they approach those of the extinct myriacanths and the very early conditions of the teeth of other living Chimzroids, Chimera, Callorhynchus, and Harriotta. In advanced stages. the teeth of Harriotta differ from those of Rhinochimera in possessing several series of tritors which in superficial aspect resemble, in shapes and arrangement, certain crowns of placodont teeth. On the teeth of Rhinochimera there are no tritors; the teeth of the very young of the other living genera are similar; this no doubt isa mutual resemblance to those of a common ancestor, an index to derivation. * * *'' To this interpretation, on the other hand, there are two somewhat critical objections: (1) that in Rhinochimera, as this author has later observed, there are present tritoral points, small, it is true, but tritors none the less; and (2) that his conception of the dental plates of fossil Chimezroids (e. g., Myriacanth) is not valid, for whatever be the puzzles of the dental plates of fossil Chimzroids they have always tritoral areas. {They may be expected to appear in a more perfect condition in very old individuals, somewhat as they develop in the late rather than in the young larve of Chimera. SKELETON. 277 that the dental plates of the adult are attained only after a process of metamor- phosis, during which the marginally high, delicate, glassy, and secant plates of the young are worn down and give rise to the adult dental plates, broad and thick, studded with tritors. Obviously, therefore, if we accept the view that a larval dentition is present, it is clear that the substance of the dental plate can better be regarded as a ‘‘precocious segregation” of the basal elements of teeth, 7. e., along the outer marginal rim of the plates, than as a new and independent accession to the materials of development. All will admit, however, that the requisite proof of this conclusion can be presented only by paleontology. Ona later page the evidence in this regard is summarized. SKELETON. The vertebrate column of Chimeroids represents, according to Hasse (1879) a polyspondyly, which he regards as typifying the ancestral condition in sharks. The column of Callorhynchus was examined from the standpoint of embryology by Schauinsland, whose conclusions I summarize as follows: That the early growth of the chordal sheath resembles that of many sharks, inasmuch as its substance is invaded gradually, and only at few points, by mesen- chyme cells. That cartilage appears quite late in development. That in each segment (metamere) appear both neural and interneural plates, as well as corre- sponding (7. e., double) hemal arches, especially throughout a greater portion of the tail region. That these cartilaginous arches do not grow around with their bases the secondary chordal sheath; this is only overgrown by a stout sheath of connective tissue; the latter together with the arches on the one hand and the secondary chordal sheath on the other forms the secondary vertebrae, but the secondary chordal sheath is not divided into separate (primary) vertebra—the segmentation of the column being indicated only through these parate arches. Schauinsland, in brief, has been able to find no vertebral centra, in the sense in which they occur in other fishes ; and my own studies upon Chimera have been no more successful in this important quest. No centra are found in either early or late ‘‘larval’’ stages. Nor do they occur, as I suspected they might, after the fashion of gerontic structures, in very large individuals. At the most, in the latter case, there was a fusion of neural and hemal arches occurring in the region near the occiput, but nothing which could be interpreted as definite centra. There is still, none the less, the possibility that some form of centra were represented in the ancestral Chimeeroid, and that they were gradually lost in ontogeny; indeed, as we shall later note in the Jurassic Squaloraja and Myriacanthus, centra appear to have been present in the anterior region of the column (figs. 138 and 140 Cc), where in all recent Chimeroids, indeed, the most perfect neural and hemal supports appear. The development of the skull has already been illustrated in several stages of Callorhynchus by Schauinsland, and in a single late stage of Chimera by the present writer. The results of their observations are briefly these: The chimzroid cranium, instead of developing as a uniform trough-like brain-case (shark), appears, even in early condition, in a wonderfully complete form ; it incloses the hindbrain, 128 CHIMEROID FISHES AND THEIR DEVELOPMENT. the forebrain, and the nasal region; and it early develops conspicuous ridges which evidently support and protect the eyes. And it is the latter organs, it may. safely be said, which have played the most important part in modifying the growth of the cranium. For the orbital region is of enormous size, occupying no less than 50 per cent of the entire length of the cranium;* and, correlated with this, between and above the huge optic capsules, the growth of cartilaginous structures is retarded. It follows, accordingly, that while the posterior and anterior parts of the chondro- cranium are well developed, its mid- or orbital region is largely unformed, and this is, I take it, the reason, the principal reason, that holocephaly has been developed, to weld strongly together the anterior and postertor parts of the cranium where primi- tively the orbital walls came to be suppressed as the eyes increased in size. Certain it is that the wide palato-quadrate elements extend like firm beams between the anterior and posterior moieties of the skull, and afford at the same time a support for the great optic capsules. And in this result appears a suggestion why the palato- quadrates appear so early and are so large in size; in fact, in no stage examined has it yet been found that these palatine elements are altogether separate from the cranium. In the earlier stages described (Callorhynchus) they are separate only for about half their length, and from the details of that stage it is even doubtful whether ereater separateness ever occurs in the development of this element, earlier stages showing probably a prochondrial continuum—very much as one sees it in the prochondrium of the paired fins of sharks. The skull of the Chimeroid, in a word, is specialized even in early ontogeny; witness, among other regards, the enormous size of the posterior clinoid process, the huge fosse for the infundibulum, the exag- gerated preorbital processes, the median frontal crest, and the interorbital vacuity. It is true, on the other hand, that certain skeletal structures in the chimeroid head retain a primitive character—possibly because they have been spared func- tional changes by the very fact that the palato-quadrate element has fused with the cranium. As primitive features we may here mention: (1) The _ perfect condition of the copule of the branchial arches. (2) The presence of a pha- ryngeal element in the hyoid arch which resembles the pharyngobranchials of the hinder arches. (3) The relatively large and discrete labial cartilages, as probable premandibular arches, and finally (4) the presence of a symphyseal cartilage as (Schauinsland, K. Fiirbringer) the probable serial homologue of a basihyal. These characters are expressed, slightly schematized, in fig. 111, and may be compared with the corresponding structures in sharks (fig. 110). In these figures serially homologous parts are indicated by shaded or unshaded areas. It should be mentioned, in passing, that even the branchial region of Chimeeroid, in spite of the foregoing primitive characters, is not without convincing evidence of precocious specialization—witness the early appearance of the supporting extra- branchials of the hyoid arch, which are prophetic of the opercular flap of the adult. The problem of rostral cartilages receives no evident solution in Chimeeroid development. The anterior azygous process of selachians, which rises from the *TIn the skull of the shark (e¢. ¢., Scyllium) at a corresponding stage the orbit occupies about 30 per cent of the entire length of the cranium. ROSTRUM, BRANCHIAL-ARCHES, FINS. 129 nasal septum (usually. its base) is probably represented in the element which Schauinsland has figured as sf in his plate xvum, figs. 124, 126. However, in the Chimeroid the rostral supports (vy! and 7*) later developed into long and separately jointed elements. Quite doubtful, on the other hand, are the homo- logues of the paired dorsal elements in the selachian rostrum, those figured, e. ., by Kitchen Parker in Trans. Zool. Soc., vol. x, plate xxxviIt, fig. 1, as dér; they are possibly the homologues of Schauinsland’s elements s in the figures quoted. Equally doubtful is the more dorsal azygous element (Schauinsland’s 7, v. the present fig. 111), which folds forward and becomes a main support of the produced snout in Callorhynchus; it certainly finds no homologue in sharks, and in view of the history of the frontal clasping organ in Chimeeroids (wv. figs. 132~1 37) I am inclined to interpret it as an element, @. ¢., a fin support, transposed from a hinder position, * a view which is the less difficult to accept when one considers the metamorphosis to which the head roof has been subjected by the precocious growth of the eyes. Figs. 110 and 111.—Skull and branchial arches of Shark and Chimeroid compared. B\-B5, Branchial arches; BH, basihyal; bt”, basis trabecularum (Kitchen Parker); C; copula; C'B, ceratobranchial; 73, Epibranchial; HTB, hypobranchial; 1’, “ anteriormost lip cartilage” (Kitchen Parker); M, mandible; PB, pharyngobranchial- The history of the fins and their supports, finally, gives additional evidence as to the modified nature of later Chimzroid development. We may comment, for example, upon the appearance of lobate dorsal fins, the anterior with its spine, at an early period, and the prominence of the paired fins, the pectoral, for example, having at one time a greater proportional size than in the adult. We observe also the precocious appearance of the mixipterygia and the antero-pelvic appendages (note especially plate rx, fig. 50‘; also fig. 90, and Schauinsland’s Taf. xv1, figs. 120 and 125), a well-marked character which in such early embryos can hardly be regarded as primitive. Nor is the plan of development of the paired fins to be looked upon as yielding any evidence in favor of Gegenbaur’s archipterygium theory. Thus, the pectoral, for example, appears not asa lobate organ, contracted, shortly to bud out radial structures, but as a lappet of a lateral fold which shows in the early stages distinct metameral elements (cf. especially plate vit, fig. 49, and Schauinsland’s Taf. xxtv, fig. 174).* The paired fins, in short, develop like those of *This translocation of anterior fin-rays is by no means uncommon, associated, too, with change of function, e. ¢., Lophius, Antennarius, etc. Even the sucking disc of Remora might here be cited. 130 CHIMAEROID FISHES AND THEIR DEVELOPMENT. young sharks, save that, as in the case of many other chimeroid structures, the rate of growth is accelerated; the lateral-fold beginnings extend over fewer body segments and are higher (proximo-distally), leading us to conclude that in this mode of early fin growth the Chimezroid exhibits the same relation to the shark that the tel- eost bears to the ganoid. Especially convincing evidence as to the modified nature of the chimeroid fin is produced by the development of the ventral ‘‘claspers’”; for these, the antero-ventral hooks and the mixipterygia, are to be regarded as highly modified radials. The antero-ventral clasper, it is clear, has not yet been evolved in the sharks, unless the greatly enlarged anterior lappet of the ventral fin be regarded as its equivalent; but there is good foundation for the belief that in Chimeeroids between the antero-ventral organ and the compressed lappet of the € Cf Fig. 112.—Ventral fin and appendages in Chimera colliei. A, Fin of young specimen (31 cm. in length) ; ventral aspect showing mixipterygia and antero-ventral clasper, the latter still connected by dermal crease with the anterior rim of fin; ¢, mixipterygium with lips unfolded; B, skeleton of foregoing fin, showing the arrangement of the supports (radials) of the branches of the mixipterygium; C, skeleton of fin, adult; D, skeleton of ventral fin of Cestracion (Heterodontus japonicus), adult, for comparison with foregoing. pelvic fin there formerly existed a number of radialia; witness, for example, the rudiments of the segmentation of the basal plate from which the antero-ventral organ arises (fig. 112, nerve and vessel openings in B and c),f or better still, the radials which persist in the anterior reach of the fin of the Jurassic Chimeeroid, Squaloraja (fig. 138, av). The mixipterygium also bears testimony to having been closely connected with the radials of the base of the fin; thus in one stage in development, c/. fig. 112 B, the base of the mixipterygium bears rudiments of radialia, and the trifid tip is in itself a relic of a clustering of distal radials. These observations are clearly in line with Jungersen’s, who, while admitting that the ‘‘appendix-skeleton of the Holocephales is of less compound construction than that of Plagiostomes,” calls attention to the ‘‘ wide separation of the whole organ (7. e.,) *In the adult Chimzroid the basal articular element of the pectoral fin is usually termed (as in Cestracion) mesopterygium, and it is regarded (Gegenbaur, rgor) as including also the propterygium; Schauinsland, however, has shown (of. cit., Taf. xxiv, fig. 174) that the bibasal character of the fin is due to the obsolescence of the metapterygium. The articular basal is, therefore, the propterygium. With this result the present writer is in accord. +This conclusion was originally suggested by Gegenbaur (1901) on the evidence of adult anatomy. VISCERA. D0 the mixipterygium) from the fin proper; the highly specialized form of the primary skeletal parts—against the simpler form in the Plagiostomes (as the simple rod- like shape of the terminal joint * * * ),’’ the presence of ‘‘particular copulatory organs,’’ and infers finally that ‘‘the Holocephales by no means occupy a primitive position among the Selachians.’’ (Danish Ingolf Exped., 1, pp. 20-21). VISCERA. In the development of its viscera, also, Chimera indicates a high degree of specialization. This, for example, may be noted in the following structures: Mesenteries.—No continuous mesentery is observed even in later embryonic stages of development. Thus, in the embryo shown in plate vn, fig. 45, the mesentery is clearly reduced to the string-like supports for vessels and ducts which characterize the adult. In the same stage only a rudiment of a ventral mesentery is present. Gut.—In no stage is the gut of the same proportional length as in the shark. In the latter (Pristiurus) the length of the digestive tract (measured from mouth to anus) decreases in length between stages k to o from 55 per cent to about 50 per cent of the total length of the embryo; in Chimera in similar stages from less than 30 to about 15 per cent. In other words, the gut of Chimzera develops in a much more restricted body region; and from early stages it appears as a short tube of wide caliber. The stomach dilatation, we may thus conclude, fails to become expressed, and the intestinal valve, instead of undergoing the further spiral devel- opment of sharks, makes but a few turns (about four) and then increases rapidly in the width of the infolded band. Gills—The gills exhibit greater changes in their ‘ ‘metameral”’ series than sharks. Thus the hyobranchial cleft, even in as early a stage as kK, is notably the largest and by stage o the opercular fold has attained almost its adult proportion. On the other hand, the fifth gill-furrow, although clearly indicated, e. g., stages K, L, M, fails to become a functional gill-slit. And the spiracle, even in a favorable stage, is littke more than a tubular rudiment; it never develops respiratory filaments and is lost by stage N. Another evidence of precocious development is shown in the mode of growth of the external gills. These filaments are from the beginning (about stage kK) of large caliber (c/ Schauinsland’s Taf. xtv, fig. 110), 2. ¢, they at once assume nearly their functional size. Accordingly they do not arise in a uniformly developed vertical series, but on account of their extraordinary diameter bud out one after another as the gill-bar increases in size. ‘Their later specialization in developing blood-producing dilatations has already been noted (pp. 60, 106). Kidney.—The restricted length of the visceral cavity is accompanied by modi- fications of the excretory system. Of the pronephros I am unable at present to give a detailed account, and will note only that it is smaller and more difficult to trace thanin the shark. The mesonephric tubules, on the other hand, are long and coiled irregularly; they appear early and are clustered in a deep stroma along the dorsal wall of the body cavity. Their early condition, therefore, does not, 132 CHIM4ROID FISHES AND THEIR DEVELOPMENT. embryologically at least, indicate a primitive segmented condition (Redecke), and I am led to suggest that the ‘‘segmentation”’ of the kidney of the adult arose secondarily. The absence of the Geschlechtsniere in Chimzroids, moreover, I also interpret as a secondary reduction, an accompaniment of the enormous develop- ment of kidney in a short body cavity, a process which caused an enlargement of functional nephric tubules, an obliteration of rudimentary ones, and more direct and special means of carrying out-gonadial products. In favor of the last inter- pretation are the great size and elaborate regional differentiation of Wolffian and Millerian ducts. NERVOUS SYSTEM. The following features in the development of the system may be mentioned as indicating that the Chimeroids have been subject to wider changes than kindred sharks. Reduction of Cord.—The cord in the region of tail and hinder trunk, repre- senting about 60 per cent of its entire length, is greatly modified. Contrast in this region the diameter of the cord, its histological differentiation, the size, number, and character of the roots of the spinal nerves. The flattening of the cord in the hinder trunk and tail region is, therefore, hardly to be compared to the condition in Cyclostomes. Flexure of Brain.—I\n Callorhynchus (cf Schauinsland’s Taf. xxir) the brain shows extraordinary flexures; in the region of the midbrain its axis changes direc- tion by almost 180”. Size of Infundibulum.—In early stages the infundibulum attains great size; and concomitantly the dorsal wall of the diencephalon is compressed between the forebrain and the optic lobes. Separation of Hemispheres.—In this regard the early condition is more marked than in any other fish-like vertebrate. Observe also the separation of the entire forebrain from the midbrain. This in Chimera begins in early stages and in the adult attains remarkable proportions. IIL. FOSSIE CHIMAEROIDS: THEIR SIGNIFICANCE IN THE STUDY OF RECENT FORMS. The evidence of paleontology in the problem of Chimzeroid descent is import- ant, although one must frankly admit that it is still lacking in essential details, for not only are fossil Chimeroids rare, but they occur with but few exceptions in fragmentary form. An outline of the distribution in time of the genera of Chimeroids is shown at the bottom of page 134. In this has been omitted reference to the supposed Silurian Chimeeroid Dectyorhabdus priscus Walcott, for reasons which are stated below. Among the genera given, it will be seen that three, doubtfully Chimeeroid, are Devonian, representing together about 16 species; one, probably a Chimeroid, is Permian, and four are exclusively Jurassic. From this time onward the greatest number of genera flourished in the Cretaceous, representing at least 50 species, and one of these genera, Ischyodus, extends from the Jurassic into the early Miocene. Another, a Cretaceous genus, Callorhynchus, is, as we have seen, represented by half a dozen species at the present time. With this plan of distribution in mind, we may summarize our knowledge of fossil Chimeeroids with reference especially to their advancing characters. THE QUESTION OF A SILURIAN CHIMAROID. Paleozoic Chimeroids claim evidently our closest attention, and we should consider first of all the question of the ‘‘ fossil Chimzeroids”’ described by Walcott in 1885. At Caiion City, Colorado, in the Ordovician (Upper Silurian), the United States Geological Survey obtained a number of narrow, ribbon-shaped fossils which were described by Walcott as Dictyorhabdus priscus, and were regarded provision- ally, on account of their general shape and transverse striation, as vertebral columns of a chimeroid fish.* In spite of the relative abundance of these fossils, however, no Chimeera-like dental plates, spines, or kindred structures were found, a condition the more remarkable since in the matrix there occur innumerable frag- mentary ‘‘fish’’ remains. It is therefore doubtful whether so delicate a structure as the vertebral column of a Chimeroid would be preserved if no traces were present of associated spines, heavier cartilages, and dental plates. The chimeroid nature of the fossils, moreover, becomes more doubtful still if they are closely scrutinized. *Walcott, it should be stated, refers doubtfully to their chimzroid nature. The ‘‘correlation is based entirely upon the resemblance between the fossil form and the calcified sheath of Chimcera monstrosa. This resemblance is too striking to be passed over, although there are certain differences that render it of less value in classification than at first." 133 134 CHIM#ROID FISHES AND THEIR DEVELOPMENT. In the type specimens (e. g., one shown in fig. 113) we note that the calcified ‘“‘rings,’’ which were compared to the ‘‘centra” of Chimera, are not rings at all, in the sense that they occur in, for example, Squaloraja, but suggest rather a series of more or less irregular lines of growth. These, indeed, are not transverse to the long axis of the fossil, but at one side pass obliquely into root-like processes, com- pared by Walcott to “‘lateral rib-sockets or supports,’’ structures which, it must be admitted, are altogether unknown in chimeroid anatomy. The foregoing evidence, accordingly, seemed inadequate for associating this fossil with Holocephali, and an examination of the types in Washington did not yield me any more convincing basis of comparison. Nevertheless the very suspicion of a Silurian Chimeroid was of sufficient interest to warrant an attempt to secure more perfectly preserved material. Accordingly, in 1896, I took the opportunity of visiting the type locality, and may now add the following details: The horizon, the age of which is now generally admitted to be Ordovician, was readily located, and Dictyorhabdus was found to be fairly abundant. ad Prane Ve STAGES OF FERTILIZATION, SEGMENTATION, AND BLASTULA. (Preparations magnified about 15 diameters. All drawings from fresh material. Figs. 22 to 28 from camera drawings of embryos which had been removed from the egg and viewed as transparent objects.) Fig. 18.—Late stage of fertilization. The oblong shape of the germinal area is due to artifact. The preparation illustrates the number and size of the entrance pits of spermatozoa and the extent of the marginal groove. Fig. 19.—Later stage of fertilization. This indicates the extent of the marginal groove and the difference in size of the entrance pits of the spermatozoa. Fig. 20.—Stage showing in surface view a single furrow. As already noted, however, this stage is not one of first segmentation, since it contains several segmentation nuclei. Surrounding the germinal area is a narrow groove margined outwardly by eminences containing sperm nuclel, Fig. 21.—Stage similar to foregoing, but showing at the surface four “‘ blastomeres.” Fig. 22.—Stage of early segmentation. Here the marginal areas containing sperm nuclei are far less conspicuous. Fig. 23.—Stage similar to the preceding. Fig. 24.—Stage of segmentation. Fig. 25.—Stage of late segmentation. Blastomeres in resting stage. Fig. 26.—Stage of late segmentation. Fig. 27.—Stage of late segmentation. The darker color of the central blastomeres indicates a greater depth in this region of the germ. Fig. 28.—Blastula. In this stage inter-blastomeral lines were traced over the light-colored circum- germinal ring. Fig. 29.—Blastula. Viewed as an opaque object, and showing a sharply marked boundary between the blastoderm and the circumgerminal ring. 180 y CHIMAERA . : PLATE IV: Bashiomt Neat, del ; : hh Ansty Wernera Wintel, Frankturt®™ — PLATE: Ve BLASTULA, GASTRULA, AND KARLY EMBRYOS. (Preparations magnified about 15 diameters. In Figs. 30-34 the circumgerminal zone has been inaccurately litho- graphed ; it should appear less conspicuous, its outer margin merging insensibly into the surrounding yolk.) Fig. 30.—Late blastula, showing especially the extent of the circumgerminal ring and its irregular margin. Fig. 31.—Early gastrula. The transverse shadow at the lower end of the germinal area represents the beginnings of the archenteric cavity. Fig. 32.—Early gastrula, showing the extent of the archenteric space. Fig. 33.—Gastrula, showing the appearance of the head region of the embryo. In this preparation merocytes could be distinguished in the outer part of the circumgerminal ring. Fig. 34.—Gastrula, showing the early embryo and the extent of the segmentation cavity. Fig. 35.—Gastrula, slightly older, showing the early vascularization of the blastoderm. Fig. 36.—Gastrula, showing early embryo at a stage corresponding with Balfour’s stage c in the shark. Fig. 37.—Blastoderm, showing embryo at a stage corresponding with Balfour’s stage Fr in the shark. Fig. 38.—Blastoderm and embryo at a stage corresponding with Balfour’s stage G in the shark, 182 SSF es = . = ‘ 7 u PLATE V er, Frankfurt v Lith Ansiv Wernera Wi CHIMAERA Dean dei. Prare- VI. DETAILS OF KARLY EMBRYOS. a. Archenteron. n. Neurenteric opening. ec. Ectoderm. op. Optic vesicle. ent. Entoderm. pn. Pronephric region. gi. gli, First and second gill-clefts. v. t. Vitello-intestinal vein. h. Heart. Fig. 39.—Detail of embryo shown in plate 1, fig. 35, viewed as an opaque object. Fig. 39 a-e.—Same embryo viewed in various positions as transparent object. Fig. 40.—Embryo shown in plate v1, fig. 36, viewed as a transparent object. Fig. 41.—Embryo shown in plate v1, fig. 37, viewed as a transparent object. lig. 41 a.—Embryo shown in plate vi, fig. 38, viewed as an opaque object. Fig. 41 6.—Embryo shown in plate v1, fig. 38, viewed as a transparent object. 184 Vi. PLAT ERA CHIMA elt 40 oO” 4 SC « 41” | Vite S 3 2 ae g 1 % c iy og A Puate VII. LATER EMBRYOS. (Preparations magnified about 25 diameters.) a. Anus. ot. Otic vesicle. at, Atrium. ot. o. | External opening of otic vesicle. bd. Epiphysis. p. Pineal outgrowth. C. Conus arteriosus. p. a. g. Postanal gut. cre: Caudal eminence. pn. Pronephros. Coe: Caudal vein. pn. ad. Pronephric duct. 2, a". Gill slits. Die Pectoral fin. k. Cephalic knob. & Spiracle. op. Optic vesicle. Sage Sinus venosus. Vv. Ventricle. Fig. 42.—Embryo, age about 25 days, corresponding to Balfour’s stage G (+) of shark. This embryo bent during the process of fixation. It shows especially well the knob-like out- growth, & in the region of the forebrain. Fig. 42 @ and J,—Anterior region of preceding embryo. Shown in nearly lateral and in dorsal aspect. Fig. 43.—Embryo, age about 29 days, corresponding approximately to Balfour’s stage 1 in shark, Fig. 43 6.—Anterior region of specimen similar to preceding. Fig. 44.—Embryo, age about 31 days, corresponding approximately to Balfour’s stage J in shark. It shows a bulbous caudal thickening. Fig. 45.—Embryo, age about 40 days, corresponding approximately to Balfour’s stage kK in shark, The circular area under the letters g.’ ¢.’” was found to be artifact. Fig. 46.—Embryo, age about 45 days, somewhat more advanced than Balfour’s stage L in shark. At the time of fixation the embryo probably twisted, so that its axis came to lie nearly parallel to the neighboring margin of the blastoderm. (Length of embryo 20 mm.) 186 CHIMAERA PLATE VU Dean del Lith Anstv Wernera Winter, Frankfurt?M Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Prate VIII. EmpBryoOs OF CHIM#RA COLLIEI. ch. Notochord. sp. Spiracle. . n. Neural tube. x. Irregularity in line of postanal gut. Possibly artifact. n.c. Neurenteric canal. y- Yolk. p. a. g. Postanal gut. y. s. Stalk of yolk-sac. 42 c.—Ventral aspect of the head region of specimen shown on plate vu, fig. 42. This indi- cates especially the extent of the stomadeal invagination. 46 a.—Detail of tip of tail of specimen shown in plate vi, fig. 46. It illustrates especially the character of the caudal knob and the extent of the postanal gut. 47.—Embryo and blastoderm shown attached to irregular mass of yolk. The embryo is of the stage shown in plate vil, fig. 44. It will be seen that a deep crease marks the line of sepa- ration of blastoderm and yolk, ». 47 a, 6.—Figures showing the foregoing specimen in natural size. These give an idea of the extent of the yolk mass around which the blastoderm is growing. 47 c.—Margin of blastoderm in the region which, in fig. 47, is concealed by the tail. It is here slightly nicked, and a line of fusion can be traced in the direction of axis of embryo. 48.—Embryo of stage shown in plate vu, fig. 45. The blastoderm has by this stage almost entirely inclosed the yolk mass noted in fig. 47. A small portion of the yolk is, however, seen in lower part of figure. The figure also shows, although indistinctly, a line of fusion passing from the embryo in the direction of the rim of the blastoderm. 48 a.—Preceding specimen shown in posterior aspect. This indicates the extent to which the rim of the blastoderm has inclosed the yolk. The irregularity in its margin is due probably to artifact. In the yolk itself masses can be distinguished, even under a low power, which suggest separate blastomeres. The exposed surface of the yolk is somewhat irregular, suggesting that a portion of the yolk material has recently become detached. The blastomeres themselves are loosely associated, so that some of them could be removed with dissecting needles. Their peripheries are not quite as distinct as the present figure indicates. 49.—Late embryo. Age unknown (probably five or six months), corresponding approximately to Balfour’s stage N in shark. Although this specimen was examined living, and was apparently uninjured, its body cavity was filled with blood cells. Observe also the enlarged blood-knots in the external gills and the position of the spiracle denoted in this figure by the small red spot immediately above the rim of the upper jaw. (Embryo’s length 35 mm.) 49 a.—Dorsal aspect of preceding specimen. This pictures more clearly the blood-knots of the external gills. 49 4.—Ventral aspect of preceding specimen. This shows especially the masses of yolk, y, attached to the external gills; also the point of attachment of the stalk of the yolk sac, y. s. 49 c.—Detail of facial region of preceding specimen, indicating the extent to which the gill arches protrude at the side of the head. The gill filaments are cut away, but from their bases one observes that they occur only on the anterior rim of each gill slit. 49 d.—Lateral aspect of preceding specimen. This pictures again the gill region from which the external filaments have been removed. The spiracle, sg., is seen immediately under the eye. 188 , cs Fig. Puate IX. LATE EMBRYO OF CHIM2RA COLLIEI. a.ci. — Antero-pelvic clasper. mix. Mixipterygium. a. a. ~. Anterior dental plate. p. a. p. Palatine dental plate. fu0. Frontal organ. 50.—Late embryo, age about six months, corresponding approximately to Balfour’s stage P in shark. Lateral view. The attachment of small masses of yolk to the side of the embryo is probably artifact. The opercular fold has here been partly cut away, so as to expose the gills. % about 3. g. 50 a.—Ventral aspect. External gills removed from the left side. . 50 6.—Dorsal aspect. ig. 50 ¢c.—Anterior aspect. External gills removed from the left side. Observe particularly the large size of the frontal clasping organ. . 50 @—Ventral region, showing extent to which the opercular fold has overgrown the gill lamellae. A detail is given as to the origin of the external filaments. ig. 50 e.—Region of the mouth. This shows especially the appearance of the sensory canals and the early condition of the dental eminences and of the labial cartilages. . 50 £—Ventral fin, showing the early condition of the mixipterygium and of the antero-ventral clasping organ. . 50 g.—External gill filaments, giving detail of vein and artery. Igo CHIMAERA w r PLate X. ““LaRVa’’ OF CHIMARA COLLIEI. These specimens were dredged off the Californian coast by the U. S. Fish Commission steamer A/batross, in water of about 300 fathoms. Figures are of nearly natural size. Fig. 51.—Newly hatched young. Length about ro cm. This shows especially the great width of the pectoral fin, the relatively large eye, and the lack of lateral coloration, Fig. 51 @.—Dorsal aspect of foregoing specimen. Observe particularly the large size of the open- ings of the auditory organ, aw. Fig. 51 6.—The ventral aspect of foregoing specimen. At y.5. is shown the scar, marking the point of entrance of the yolk-sac, Fig. 52.—Young of about 12.5 cm. This specimen shows a marked differentiation of the dorsal fin, also noteworthy changes in coloration, Fig. 52 a.—Dorsal aspect of foregoing specimen. Fig. 53.—Young, 18.5 cm. in length. This exhibits an extreme degree of pigmentation, Fig. 53 a.—Dorsal aspect of foregoing specimen. Fig. 53 4.—Ventral aspect of foregoing specimen. 192 CHIMAERA PLATE X Prats XI IMMATURE SPECIMEN OF CHIM4RA COLLIEI. This was drawn from a freshly taken specimen and is intended to represent the fish in its natural colors ; it does not, however, give an adequate idea of the brilliantly metallic shades of the living fish, or of the translucency of the snout region. At this stage the fins are deeply pig- mented. Natural size. 194 CHIMA Bashtord De CHIMAERA PLATE XI. tik vie ‘ { ie: 7 |" ' so ‘ cs = rk se | te c ‘ ‘ . Ay . sind oe : a 3 { ; a i — - s «3 2 = i a ‘ —= , : e¢° = i ARAA AS re ~NARAAAAREA ~~ ammsannan nnn nr BAA RAR AAA RAARAR ry rN SAA lan ~~ AAA AAAAAAA VARA AAA — Ap AA ARRAY ~ => — ss QAO ~~ V ~ = RB RRARARARA = eeeeccccrecreeeecteee : RARAAGRARAAnaaaaee ARAAA AAA AARBARAAARAR A AAA ennannaan=*= AAA AAR saaannne a AA Rama mn z A RAARRAAARARS AAAaD = as ana a~aRR v= sociccanasasaeeneeOlts AR ARRRERA A AAA, AAA I~ i W/ = manne ARABA sananaa PARAR ssa AA ARRAA ARR aR AA SIAN AP A anna AARAA AAA, Aaa aan aaana AA aaanan BARRA ARAB RR RRs ~a mannan RAPA AARAARARA- VEAAREY a ml BS ANRAS © AAARA A . Pinenpanimannamannnen an ARAAARAA ~~) AAA aRam~ala ala Manna naaaaaa RRARRRE \AARREARRA Ae. am ~ AAA Valle aiayaiae ARA Y AARAAREEER AANA AAA AA RRARRRSAR, aap a a ARAN AAA RAN NANA AAA AE a ce ce 6 I RR co RPA RE RRR RARER AAA AVF AR ARE =Car | \ : (ee ananne | ea | | RRRARRE RAZ ARRRAAAAABA ) Cow obo AAsmAsagaasaaaseo" eee nanan Pasaaamane AAARAAs | ARARAm a . RAANPACAAARS man: eco Aaa l=) } annaaa nace : ) lanes! | | RAAB AAA SNS AANA ee PP clots pan nnn aasanananananan anannaannneeen aa lena ARARARAAAAA anna mean: NAARAAsaacseanaes ZF MN i)