> bw aed te af Satedel b he Stee baton det ey Ale ser oF har a typed avy 4 stile iftvat ptheta? eae! 1a Hen nd de~ , e, ria on Ja Yaimrebshate) dpaes al w Fd ply? P-n ey rede ATH. aie ieee re a: - : ; Water, > : 7 Mio ihets At as he \y Ade “ , ee * , oo? Mie é teas sia + thal Ba apatites ic ena eer te Pr 7. iid ze i ed P; Psd ty - : id ear tie peed tae cy re f i ; - ol heated oh okads Leama) “oy peteh yipee) bubs Pitt Hee hee pene erie bee } may! og A : os sided wl , ; abies ane insteate ry Nad) , 94 Pe > A » E Magehied deen Hasrasee lpi He ati 3 # F St sf re pS Divscs Sia aanatey Brenig d's) MSs Me Pabst bad ae “ ‘A yy paper af ey Sree “ ae | a a & ’ 3] at te Oe pe per + eerie” ha i} + Ske Me ayaa sir a2. ROELS fe . , rieuee Thee rbot set ype nw Siena gat rahe Taree waginisea Ty leaned Pica ; ai. - +e - J Speengernc ey pipet ees a letep ldver Pe % Le : ; - =A) bap Firat uae ! put Ledettts Dyeitege ay PEER Ter Matec 4 We i ids ated eee y i iy ‘ ouah a Heh Sire tatr te) Ptah we } ‘ “ ey ida : eH forty oe rf in Be) i ah ee ay a : 3 Phytol tal say eiaintidl, eigaTeD aya peemanpeal eats tye Mba . y wee Hat pas’ r 2) Y te, D - Hani tas aah ’ Gis ith ag: sense eS Lisi ener TE Saad al nt rae aia! reash p phat’ ; - ipihi h ities! Napisy pie “ oda ifiee tial 4 “ te use ein inten Hb) Ls < oe hertry sehr ed aRte Poy hi bohide tibbavatry ieee ae 2. od ‘i % oe Hegel az! ee F ; Wargtte ag Behr ph Deyn Sy sea NEI as Srkesibseitaascahet had pate Be the iet SRN RS 3 v4 eteratny as Hes Milage seas ase ata Ae fi atk Raa per epealceateaey | fre rook etter KEIR gc et Ne a ap IL re iat asad HG = 4% yi wigs Tela ity wks eh ye : Ch reba mB may se IY edhe dykes , Li Syareeest rae Me 84, ot td . Q ide rner at si ye aetna 7. ars Mae 44 stead boo ‘ ie) ue | Ug pbichaas St my ¢ Weare aes weal Sarrateral wheteti=4l mt te ewengs Nehrep Sri byt Apes 4¢4/] Je yl pi sperbrey peed wkend weteGe aap Asdae 1 Hos Sad oe we Ertan Hava win “y RS ath oNse oN BINS tec ae ee ae VOC oy ee ee, ier Op Mt werd wired Ate ott ; Digitized by the intemet Archiv In 2010 ttn ‘una from | | _* ; pa Phone one LN ote :.. a A one rn I OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE. VOL. XXX. CAMBRIDGE, MASS, U.S. A. Printed for the Aduseum. 1904-1917. f oe CONTENTS. No. 1— REPORTS ON AN EXPLORATION OFF THE WEST COASTS OF MEXICO, CENTRAL AND SOUTH AMERICA, AND OFF THE GALAPAGOS ISLANDS, IN CHARGE OF ALEXANDER AGASSIZ, BY THE U.S. FISH COMMISSION STEAMER “ALBATROSS,” DURING 1891, LIEUT. COMMANDER Z. L TANNER, U.S.N., COMMANDING. XXX. THE SPONGES. By H. V. Wiitson. 26 plates. July, 1904 . : : : : : : : : : 1 No. 2.— REPORTS ON AN EXPLORATION OFF THE WEST COASTS OF MEXICO, CENTRAL AND SOUTH AMERICA, AND OFF THE GALAPAGOS ISLANDS IN CHARGE OF ALEXANDER AGASSIZ, BY THE U. 8S. FISH COMMISSION STEAMER “ALBATROSS,” DURING 1891, LIEUT. COMMANDER Z. L. TANNER, U.S.N., COMMANDING. XXXII. THE RADIATING ORGANS OF THE DEEP SEA FISHES. BY ROBERT VON LEDENFELD. WITH AN APPENDIX ON THE STRUCTURE OF THE BIRD-LIKE ORGANS OF MALTHOPSIS SPINULOSA GARMAN. By Emanuet Trosan. 11 plates, 1 chart. August, 1905 : ; : ; : : ‘ F : : : a LO5 No. 3.— REPORTS ON AN EXPLORATION OFF THE WEST COASTS OF MEXICO. CENTRAL AND SOUTH AMERICA, AND OFF THE GALAPAGOS ISLANDS, IN CHARGE OF ALEXANDER AGASSIZ, BY THE U.S. FISH COMMISSION STEAMER “ALBATROSS,” DURING 1891, LIEUT. COMMANDER Z. L. TANNER, U. 8. N., COMMANDING. XXXVI. EIN BEITRAG ZUR MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES. Von Emanvet Trogan. 6 plates. October, 1906 . @. ‘ : 5 : : : ‘ é , at WG No. 4.— THE GALAPAGOS TORTOISES. By Samurnt Garman. 42 plates. January, HOEy , ; ; ; : ; ; ‘ j : : : ; 2 RE a dtlemoirs of the Museum of Comparative Zoologqn AT HARVARD COLLEGE. Vout. XXX. No. 1. REPORTS ON AN EXPLORATION OFF THE WEST COASTS OF MEXICO, CENTRAL AND SOUTH AMERICA, AND OFF THE GALAPAGOS ISLANDS, IN CHARGE OF ALEXANDER AGASSIZ, BY THE U. S. FISH COMMIS- SION STEAMER “ALBATROSS,” DURING 1891, LIEUT. COMMANDER Z. L. TANNER, U.S. N., COMMANDING. XXX. THE SPONGES. By H. V. WILSON. WITH TWENTY-SIX PLATES. [Published by Permission of MarsHatt McDonatp and GrorcEe M. Bowers, U. S. Commissioners of Fish and Fisheries. ] CAMBRIDGE, U.S.A.: Printey for the fMuseum. JuLy, 1904. ay le TABLE OF CONTENTS. PaGE PA ERODUCTION (wicca we ue - 3 O15 SYSTEMATIC ACCOUNT OF THE GENERA AND SPECIES. . 16-161 HYALONEMATIDAE HIGVGATLO NIE AM tye foo clk et ws ee LO He ovulsperun |i.) ore. 16 LH, pedunculatum (P1. 3, Figs. 1- 6) 19 H. bianchoratum (Pl. 2, Figs. pica 4 ecraee Gree) eee ; ae ao sp. div. (Pl. 2, Figs. 12- 16) . 25 H. pateriferum (Pl. 1, Figs. 1-13) 28 EUPLECTELLIDAE EUPLECTELLA ae we Er See, oe EH. sp. (Pl. 4, Fig. 4) Sinton x th ne hae etek ol pae ena Je. sp. (Pi. 9, Fig: 9) Skah ee he OU &. delicata (P1. 3, Figs. 7, 8; Pl. 4, LE Seep Td 1 Ae eee ee CAULOPHACUS . . 43 C. schulzei (Pl. 4, Figs. 1, 3, 5-10; IES, Pies. 1-6, 8-10) .. ... 48 Crack Mig. 7). BO RossELLIDAE BAaTHYDORUS. . . d1 B. levis spinosus (P1. 5, eae 11- 43; Per nieeee Oy) ek lk ew ON STAUROCALYPTUS . . ee ge S. sp. (PL. 6, Figs. 4-10) . ae oe FARREIDAE IWAGREA . <>. 57 FE. occa Sti mis (Pl. ¢ 6, Figs. 3, 14 7, Wigs. 1-3, 6). . 57 F. mexicana (P1.7, Figs. 4, 5, 7, 8, Mieke. we ff 60 PPEe Mss -5--< « « ., 62 EURETIDAE EURETE E. erectum . EL. erectum bib cd (inci (Pi 7, Figs. 9,12; Pl. 8, Figs. 1-3, 6) EH. erectum mucronatum ea ‘ Fig. 7) : E. erectum gracile Pl, 8, Figs 4, D8, PEEL, Migs: 1d; 5) E. sp. MELITTIONIDAE APHROCALLISTES A. vastus CoscINOPORIDAE CHONELASMA C. calyx (Pl. 10, Fig. 5) BATHYXIPHUS Bes (El. 10. Fie. 2) . TRETODICTYIDAE HEXACTINELLA . HT. labyrinthica (P1. 10, Figs. 6, 7; PI dt; Figs. 1%) : HI, ventilabrum H. tubulosa ScLEROTHAMNOPSIS S. compressa (P1. 9, Figs. 2, 4, 6-8 10) TSE iss: 13). THENEIDAE ? THENEA : P; eee (Pl 13, Figs. 2-4, £6, (i | nae ae pes T. echinata (P1. 12, Figs. 1-9) T. lamelliformis (Pl. 12, Figs. 10-13; Pl. 13, Fig. 1) T. pyriformis (Pl. - Figs. 5, 8, TG7 RLY: PAGE ANN 4 oo > > 4 TABLE OF CONTENTS. PAGE Pace POECILLASTRA. . . . 101 | PorcrLoscLERIDAE P. tricornis (P1. 18, Figs. 12-14, TYLODESMA.. . . ~ AB Pl. 14, Figs.1-8) . . . . - 101 T. alba I 18, Figs. 51; PL 22, P. cribraria (P1. 14, Figs. 9-12; Pl. Pigs,-2: SY 1- am 136 15, Figs. 1-4; Pl. 16, Figs.1, 3) 105 T. vestibuiaris (P1. 18, Figs. 8, 9; STELLETIDAE Pl. 19, Fig. 1; Pl. 22)5ieaeee PENARES . 111 Pl. 23, Figs. 1-3) . as ae ,. P. foliaformis (Pl. 15, Figs. € 5-11) BM IopHON 2. 143 Barr wcdnrnae I. chelifer ostia-magna el 20, PoLYMASTIA . . 115 Figs. 2, 4, 10, rie fag ls 24, Fig. 1) 143 P.maeandria (Pid 16, Figs. 2, 1-6; I. lamella (Pl. 20, ig 3, 7-9, 12, Pig Wed). s eg ee 13; Pl. 24, Figs. 2-4). . . . 146 TLAPLOSCLERIDAE I. lamella indivisus (Pl. 20, Figs. Proce 118 14-16) 5 3 see ee 149 P. variabilis crassa (PL 41 Figs. I. indentatus (Pl, 19, Fig. 6; PL 6, 9, 12 ele PA Figs. pe 3) LeiGh= 20, Figs. 1 5, 6; Fi, 23, Rig 4) 151 P. similis ese (P1. 17, Figs. AXINELLIDAE G 103-2121 iA) a ee Pian ae 155 OE SUNSETS = a ane 18, Fig. 10; PL P. acapuleensis ie 16, Figs. te 8; 19, Figs. 2,3; Pl. 25, Figs le Pl. 1%, Figs. 1-5, Ava) ee ee ae tice) SE 3) 4). eS, BK. . oS OcEANAPIA. . . 127 ee ee 158 Aas if e D uy, Hg 8; Pl a A. dendrophora (P1.19, Figs. 4, 5,7; eee PL, 25, Fig. 2). . . 0. sna GELLIUS . . . 130 G. perforatus (PI. 17, Fig. he PL 18, Figet > El. 2k oe G52-EA: 22, Fig. 1). ving xia . . 1301 LITERATURE CITED. . . .162-164 THE SPONGES. INTRODUCTION. Tue collection of sponges with which the following report deals has been found to include forty-seven species and subspecies. Of these, twenty-six, representing thirteen genera, fall in the Hexactinellida, seven, representing three genera, fall in the Tetractinellida, and fourteen, representing nine genera, fall in the Monaxonida. No calcareous or horny sponges and no Lithistids were taken. As was to have been expected, since the expedi- tion was made in unexplored waters, a very large percentage of the forms (thirty-three species and subspecies) prove new to science. ENUMERATION OF THE SPECIES ACCORDING TO THE STATIONS AT WHICH THEY WERE TAKEN. Depth . . Character Serial Number. Latitude. Longitude. A of Species taken. West. Fath- Bott ann ottom, 3354 IN. 7 9 45) 80 50 0} 3822 gn. M. Chonelasma calyx F. Bi. Sch. (sp ?). 3358 N. 6 30 0} 81 44 O} 555 gn. S. Eurete erectum tubuliferum, subsp. nov. Eurete erectum mucronatum, subsp. nov. 3359 N. 6 22 20] 81 52 Oj 465 Rky. Eurete. erectum tubuliferum, subsp. nov. Eurete erectum mucronatum, subsp. nov. Hexactinella ventilabrum Carter (sp ?). 3362 |N. 5 56 0} 85 10 30/1175) gn. M.S. rky. | Thenea fenestrata O. Schm. 3363 N. 5 43 0} 85 50 0} 978| wh. glob. Oz. | Hyalonema pateriferum, sp. nov. 3368 N. 5 32 45} 86 54 30) 66 Rky. Phakellia lamelligera, sp. nov. 3370 N. 5 36 40) 86 56 50) 1384) Rks. and 8. | Staurocalyptus, sp. Eurete erectum F. BE. Sch. (sp?) Hurete, sp. Aphrocallistes vastus F. KE. Sch. (sp ?). 3376 =IN. 3 9 0; 82 8 0/1132] gy. glob. Oz. | Hyalonema pateriferum, sp. nov. 3380 |N. 4 3 0} 81 31 0} 899 Rks. Regadrella, sp. Hexactinella tubulosa F. ¥. Sch. (sp?) urete erectum gracile, Subsp. nov. Lurete, sp. Bathyxiphus, sp. 3381 N. 456 0] 80 52 30/1772 en. M. Hyalonema, sp. 3382 N. 6 21 0} 80 41 0/1793 gn. M, Hyalonema ovuliferum ¥.E. Sch. Caulopha- cus schulzei, sp. nov. Bathydorus levis spi- nosus, Subsp. nov. 3384 |N. 7 31 30} 79 14 0} 458 en. 8. Tophon chelifer ostia-magna, subsp. nov. 3399 N. 1 7 0} 8L 4 0/1740 gn. Oz. Hyalonema, sp. Caulophacus schulzet, sp. nov. Bathydorus levis spinosus, subsp. nov. i 6 THE SPONGES. . Depth t : Latitud Longitude. | in baer ae Species taken, Serial Number. atitude. West. Fath- es oms. 399] It. ey. glob. Oz.| Uhenea fenestrata O. Sch. eee ee a aaa o i Tone sp. Regadrella delicata, sp. nov. Hevactinella ventilabrum Carter (sp ?). Poecillastra tricornis, sp. nov. Penares folia- formis, sp. nov. Oceanapia bacillifera, sp. 1L0V. 3405 S 0 570) 89.38 -0). 63)" B.-s€o.’Sh: Hewxactinella labyrinthica, sp. nov. Poecillastra cribraria, sp. uov. Polymastia maeandria, sp. nov. Petrosia variabilis crassa, subsp. nov. Petrosia similis densissima, subsp. nov. Tylodesma vestibularis, sp. nov. Lophon lamella, sp. nov. Lophon lamella indivisus, subsp. nov. Lophon indentatus, sp. nov. Auletta dendrophora, sp. nov. Tylodesma alba, sp. nov. 3406 S. 016 O! 90 21 30] 551 R. Hevactinella tubulosa F.¥.8ch.(sp?) Hex- actinella ventilabrum Carter (sp?) Scelero- thamnopsis compressa, gen. et Sp. nov. 3413 N. 2 34 0| 92 6 0/1360] glob. Oz. dk. Sp.| Thenea fenestrata O. Schin. 3414 |N.10 14 0} 96 28 0/2232 en. M. Hyalonema pedunculatum, sp. nov. Hyalon- ema, sp. Caulophacus, sp. Thenea lamelliformis, sp. nov. Thenea pyri- fornis, 0. sp. 3415 |N.14 46 0] 98 40 0/1879) br. M. glob. Oz. | Hyalonema bianchoratum, sp. nov. Thenea echinata, sp. nov. 3400 3404 MN be oo S 3425 N.21 19 0j106 24 0) 680} gn. M.and 8. | Farrea occa claviformis, subsp.nov. Farrea, sp. 3430 N.23 16 0/107 31 0) 852 bk. 8. Farrea mexicana, sp. nov. Station recorded as ‘* Acapulco.” Pachychalina acapulcensis, sp. nov. Station recorded as “ Panama.” Gellius perforatus, sp. nov. 17 (Cruise of |N. 0 50 0)137 54 0/2463) It. y. gy. glob. «Albatross ” ooze in 1900.) Hyalonema pateriferum, sp. nov. In addition to the discovery of new forms, some results of general bio- logical interest have accrued from the study of the collection, and are dis- cussed or stated under the respective species concerned. For convenience of reference the more important of these facts may be here classified. Remarkable forms. Hyalonema pedunculatum, p. 15, is noteworthy for the peculiar pedunculate form of body ; Scderothamnopsis compressa, p. 80, for the shrub-like habitus in which it resembles the hitherto unique Sclerothamnus clausii Marsh. ; Regadrella delicata, p. 32, for the character of the sieve-plate region, which may be construed as representing a simpler (although not necessarily a more primitive) phylogenetic condition than the otherwise closely similar Regadrella phoenix O. Schm. Distribution and Habitat. Thenea Senestrata O. Schmn., p. 87, hitherto known only from the Atlantic and Caribbean Sea, is now recorded for the THE SPONGES. 7 Pacific. Some of the Hexactinellids from great depths have been found to live clustering upon one another: Caulophacus, pp. 39, 46, attached to root spicules of Hyalonema ; Bathydorus, p. 47, attached to root spicules of Hya- lonema and to stalks of Caulophacus or siinilar sponge. Morphology in general. Further evidence of a convincing character has been gained that the complex Hurete and Farrea colonies are derived onto- genetically from simple cup-like forms, pp. 60, 61, 66. —— The close simi- larity between the main afferent and efferent canals and their apertures in Poecillastra tricorns is worthy of remark, p. 97.—— The observations of Sollas and of Dendy on the occurrence of a peculiar fenestrated membrane (Sollas’s membrane) in the flagellated chambers of certain sponges have been confirmed for two Tetractinellid genera, Poecillastra and Penares, pp. 99, 103, 109. Pathological phenomena in general. In Euplectella skeletal septa of a reticular character have been found crossing the cavity of the sponge, p. 29. In Hexactinella labyrinthica skeletal masses of a reticular character, such as In Chonelasma similar septa have been observed, p. 70. have been especially described by Weltner, occur, p. 75. In some cases such masses form layers separating one part of the sponge from another. Possibly all these phenomena are pathological and similar, in so far as they may indicate an effort of the sponge to shut off one part (diseased?) of the body from the rest. Morphology of spicules — variations and “pathological” conditions. In Hya- lonema pateriferum, pathological amphidiscs such as have been observed by Marshall and Meyer and F. E. Schulze occur. In these spicules, p. 28, several additional rays of the hexact are developed. Certain details in the structure of the discohexasters of Caulophacus schulzei suggest that a hexaster may arise from a hexact through the development of /aéeral Another case is afforded by branches on the hexact rays, pp. 44, 45. Flexactinella labyrinthica of what seems to be the degeneration of an uncinate into an oxydiact, p. 77. In scopulae present in Sclerothamnopsis com- pressa, the arrangement of the axial canals indicates that the spicule is Abun- dant transitional forms indicate that the protriaenes of Thenea are modi- equivalent to a branched diact, as Schulze has supposed, p. 82. fied dichotriaenes, p. 85. In Thenea pyrifornus a type of metaster is common which may be interpreted as transitional between spirasters and euasters, p. 96. In Penares foliaformis peculiar dichotriaenes are found 8 THE SPONGES. which approach the shape of the Lithistid phyllotriaene, p. 109.— The occurrence of pathological branching oxeas is recorded for Petrosia, p. 110. Variation. The variability of sponges in regard to points of adult structure is universally recognized. O. Schmidt and Vosmaer, in particular, have laid stress on the phenomenon as bearing upon the problems of systematic-classification. The observations recorded in this report will fall under the following heads : — 1. Variation in body-shape and general anatomy. Attention may be called to the difference in shape exhibited by speci- mens of Hyalonema pateriferum, p. 24, Caulophacus schulzet, p. 59, Thenea fenestrata, p. 84, Petrosia variabilis crassa, p. 114. In Lurete the spiral form of body beset with cup-like outgrowths varies toward a bilateral symmetry, p. 72. In Tylodesma alba massive and lamellate bodies occur, as parts of one continuous specimen, p. 132. In Gellius perforatus a uniform habit of growth may result in very different body shapes, p-. 128. vary, pp. 84-85 (comp. especially Vosmaer, 1882, 1885); also the spicular In Thenea individuals the number and size of the pore areas In Lophon lamella the character of fringe round the osculum, p. 84. the surface varies in different parts of the same individual, owing to the divergence in character of the main efferent canals and their apertures, and to the varying amount of collenchyma round such canals, pp. 142-144. 2. Variation in same individual in the skeletal framework or the skeleton im general. In Chonclasma calyx the tuberculation of the beams of the dictyonal framework varies, p. 71. In Hevactinella ventilabrum there is considerable variation in the way in which the hexacts combine to form the dictyonal framework, p. 79. In Thenea fenestrata there is variation in the develop- In Petrosia ment of the spicular fringes round the pore areas, p. 89. variabilis crassa, p. 116, and in Petrosia similis densissima, p. 118, a skeletal reticulum remains undeveloped in spots. In Pachychalina acapuleensis the skeletal framework in places is fairly regular, although in general irreg- ular, p. 122. In Phakellia lamelligera the skeletal lamellae vary in respect to branching, p. 153. In Petrosia similis densissima over some parts of the surface there are no projecting spicules, while over other parts such spicules are present in considerable number, p. 118.——JIn Gellius per- foratus there is considerable variation in the character of the spicular tufts which project from the surface, p. 128. In Lophon lamella there is THE SPONGES. i) variation in the number of spicules which combine to form the side of a skeletal mesh, p. 146. 3. Uncorrelated variation in the megascleres of an individual, 2. e. variation apparently not correlated with the structural peculiarities of the body- locality. As regards size of the spicules, there seems to be noticeable variation in all sponge species. The shape of one of the macramphidiscs varies considerably in Hyalonema bianchoratum, p. 20. In Caulophacus schulze the principal hexacts are occasionally tuberculated, p. 40, and in the same species the tuberculation of the pentacts varies, p. 43. —— The character of the diact ends varies commonly, e. g. in Bathydorus levis spinosus, p. 49. —— In Farrea mexicana the pentacts vary in respect to tuberculation, p. 96. In Lurete erectum the character of the distal ray of the gastral In Petrosia variabilis crassa the oxea sometimes assumes the shape of a stron- pinules varies extensively, p. 62 (comp. F. E. Schulze, 1899). gyle, or style, p. 115. In Penares foliaformis the triaenes vary con- siderably, p. 109. varies within wide limits, p. 121. In Pachychalina acapulcensis the size of the oxea In Phakellia lamelligera the oxeas and the two kinds of styles all vary considerably in shape, p. 152. 4. Uncorrelated variation in the microscleres of an imdividual. The microscleres as well as the megascleres very commonly vary in size and detailed shape, although in them, as in the megascleres, there is a size and pattern which are characteristic of the individual (species), 2. e. to which the majority of the spicules of an individual conform. Some striking cases of variation are afforded by the micro-oxyhexacts of Hyalonema bianchoratum, p. 19, plesiasters and spirasters of Thenea fen- estrata, p. 86, sigmata. of Tylodesma alba, p. 133. d. Correlated variation in the spicules. In some cases the variation of spicules is obviously not ungoverned by the rest of the body, but is correlated with body-locality. Thus while the pentacts in Hezactinella labyrinthica vary at large in respect to length of the several rays, the pentacts overlying the larger inhalent canals commonly have noticeably short proximal rays. Such a phenomenon would customarily be referred to as “ adaptive.” —— The vari- ation of dichotriaenes toward the protriaene type, round the pore areas of Thenea, pp. 85, 89, 92, 95, is another instance of the same phenom- ~ enon. A complex instance of correlated variation is afforded by the 9 10 THE SPONGES. dermal and gastral pinules of Caulophacus schulzei, which coat the opposite surfaces (pore and oscular) of the body. The two kinds of pinules vary in the same direction in different individuals, and thus the proportionate difference between them is preserved, p. 42. 6. Qualitative variation ? Two sets of individuals living together in the same locality, and which are otherwise indistinguishable, differ conspicuously in respect to a single point. An instance is afforded by Lurete erectum mucronatum, which differs from Eurete erectum tubuliferum in having oxyhexasters instead of onychas- ters, p. 64. Another instance is afforded by Jophon lamella indivisus, which differs from Lophon lanella only in the character of the bipocillus, which is not chelate, p. 145. the two sets of individuals have in each case been separated as subspecies. In order not to confuse the facts with hypothesis, It is idle to dogmatize or to speculate im extenso on the value, from the standpoint of heredity, of the point of difference. Whether this point is inheritable and thus marks off two races, or whether it merely marks off two sets of individuals which started out alike and the offspring of which are alike, and which owe their difference to the action on the individual of the environment, no one can say. The recording of the difference is the duty of the systematist, who, when he has done so, has pointed out an additional case suitable for the experimental study of heredity and environ- mental action. 7. Variation towards other species or subspecies. A class of spicules in one subspecies may vary in considerable number towards a condition characteristic of a subspecies inhabiting a different locality. An instance is afforded by Lurete erectum gracile, in which the tuberculation of the gastral pentacts and hexacts is sometimes very similar to that found in the other subspecies of Lurete erectum, p. 66. Or a form of spicule characteristic of one species may occur infrequently in a related species. For instance, in Cawlophacus schulzei the pinuli ocea- sionally have the shape characteristic of C. latus and C. elegans, p. 42. A striking case is afforded by Farrea occa claviformis, in which a few gastral clavulae were found closely similar to the peculiar clavulae of Farrea con- volvulus, p. 5d. §. Constancy of character in spicules. It often happens that while in a single individual the size of a particular spicule may vary within wide limits, the character remains fairly constant, THE SPONGES. 1] e.g. pinuli of Hyalonema bianchoratum, p. 20. The character of a spicule even in minute details may apparently become fixed for the species. Thus in a specimen of Hyalonema ovuliferum the rays of the micro-oxyhexacts have the same sudden terminal curving exhibited by the corresponding spicules of Schulze’s type specimen, although the two sponges were taken 49° of latitude apart, pp. 13, 16. A form of spicule which in some sponges varies greatly in size, in other species varies but little. Thus in Gedus perforatus the sigmata vary only slightly, p. 129, whereas in Tylodesma alba they vary greatly, p. 133. I do not undertake a comparative consideration of the geographical dis- tribution of the forms making up the collection. Such a consideration would demand a knowledge of the actual systematic value to be attached to many species recorded in the literature of sponges. And such knowledge is not to be had at present. In modern sponge literature, e.g. in the two great monographs of Schulze and Sollas (Schulze, 1887; Sollas, 1888), the species conceived are, as it seems to me, what H. M. Bernard contends for in his interesting recent discussions (Proc. Cambridge Phil. Soc. Vol. XI. Pt. 1V.; Verhdlg. V. Intern. Zool.-Congress) of the species-question as affecting the method of recording certain data, viz. homogeneous morpho- logical groups. The sponge species are often very homogeneous, because represented by single specimens. That such groups answer always to natural species, as we understand the word when we speak of the human race, Passer domesticus, Littorina litorea, or other organisms which we know in great number, is not only open to doubt, but is excessively improbable. It is, I suppose, from this latter point of view (the envisaging clearly the a priori probability that sponges in general exhibit those individual and local differences which all species known intimately exhibit) that O. Schmidt was led to record in literature the existence of such species as Farrea facunda. Perhaps Farrea facunda is a “natural species,” but the data at hand make such a statement only a subjective assumption. Or when the distinguished systematist Topsent expresses the opinion (1902, p- 12) that five species of Poecillastra recorded by Sollas probably represent the variations of two or three species, one is justified in saying “ perhaps, but the known specimens differ in certain definite respects.” Such sub- jective interpretations of differences perhaps always affect the manner in which we record the occurrence of certain morphological peculiarities in association with geographical and bathymetrical site. But whereas once |e THE SPONGES. they were rampant, to-day they are reduced to a minimum, with the result, as I have said, that the species of modern sponge literature are strik- ingly homogeneous groups, which need not be thought of as always corre- sponding to natural races. That this method of precise analysis is the only method capable of yielding trustworthy data, seems to me incontestable. That it may result in temporarily recording more species than exist in nature, will only trouble those who incline to the view that the one excuse for systematic zodlogy is to provide them with a handy collection of names for the animal kingdom. The data which are thus accumulating as to the occurrence of this or that peculiarity of structure in a certain locality are growing rapidly through the labors of systematists. Scarcely begun is the accumulation of the almost equally important data (comp. Poléjaeff, Report on the “ Challenger Keratosa,” p. 85), as to what peculiarities of structure are due to a difference in the physiological state of individuals belonging to the same race. Such knowledge, to be acquired through continuous observa- tion of living individual sponges under normal and under modified condi- tions (experimental method) may be expected to bring about the union of many recorded species. Another most important class of data can only be revealed through the physiological study of the race, viz. through the breeding of sponges. And with the increase in the number of marine laboratories at which observations may be carried on continuously through- out the year, the inauguration of such studies may be anticipated. —— The modern statistical method of considering the differences between indi- viduals of such groups as are procurable in large numbers is a refinement of what is commonly understood as systematic work, and a promising field for those acquainted with the structure of sponges. Such studies, by revealing the kinds and the extent of structural modifications which oceur among individuals not separable into morphologically definable groups, may be expected to provide invaluable special cases for experimental study. It is through the combination of these several classes of data that we must hope to learn the limits of the natural groups of sponges as they exist to-day. When such trustworthy definitions of natural groups are at hand, the facts of the geographical distribution of the species will doubtless become intelligible. THE SPONGES. 13 In the classification of the Hexactinellida the changes introduced by Schulze and Ijima in the system of the former, as laid down in the “ Chal- lenger” Report on the Hexactinellida, have been in general adopted, where they concerned the types treated of in this Report. For the Tetractinellida I make use of Sollas’s system. Topsent in a recent discussion (1902) pro- poses certain changes in Sollas’s treatment of the streptastrose forms, but the changes proposed especially concern the definition of the genera and the two subdivisions, and do not materially alter the classification of the group. Lendenfeld’s very extensive changes in the classification of the As/rophora (1894, 1903) do not seem to me an improvement on the system of Sollas. In dealing with the Monaxonida I employ Topsent’s group Hadromerina (1898) and also follow this author (1894 @ ) in the division of the Hal- chondrina into families. In regard to spicule ternunology the usage except in minor particulars and with respect to a few terms has been practically uniform since the “Challenger” Reports. The useful list of Schulze and Lendenfeld (1889) includes terms employed in the “ Challenger” Reports and others as well. > Some of the latter offer no advantage over the “ Challenger’’ terms, and have not been generally adopted, e.g. amphiox for oxea, amphistrongyl for strongyle, amplhityl for tylote, chelotrop for calthrops. Vosmaer (1902) in a recent paper full of interest discusses some of the forms with regard to which there is not a uniform usage. Prominent among these is the streptaster. Sollas (1888) included under this head a long series of forms, which he divided into plesiasters, metasters, spirasters, ampluasters, and sant- dasters. Of these the first four, and especially the first three, imtergrade freely. Vosmaer thinks it impossible to carry out in practice the distinction between plesiasters, metasters, and spirasters, and would designate them all spimspirae, including under this term some at any rate of Sollas’s amphi- asters. Schulze and Lendenfeld (1889) use spraster, amphiaster, and sani- daster in the sense of Sollas, and do not use plesiaster and metaster, but employ the term streptaster for spicules which in Sollas’s terminology would fall under these two heads. Lendenfeld (1903, p. 12) continues to use the terms ampluaster and spiraster, but does not employ streptaster, metaster, plesiaster, nor sandaster. The sanidasters (Sollas) are included under microrhabds (Lendenfeld), e.g. in Tribrachion schnidtti Weltn. The spirasters and metusters of Sollas are together included in spirasters (Lendenf.), e.g. in Pachastrellu (Poecillastra) schulzei (Soll.) The plesiasters (Sollas) are passed 14 THE SPONGES. over to the ozyasters of Lendenfeld, e.g. in Ancorina (Thenea) fenestrata (O. Schm.). Thus Vosmaer and Lendenfeld do not agree, and they both differ from Sollas. In this matter Topsent (1902) adheres to the terminology of Sollas, and I likewise employ it. That the types singled out from the streptaster series by Sollas exist is of course indisputable. That they also intergrade, cannot be questioned. And this latter fact makes it necessary, whatever technical terms be employed, to describe the spicules of each species. Nevertheless Sollas’s subdivision of the streptasters and his technical terms greatly facilitate reference to the spicules, and also make for accuracy of description. By combining the terms the transitions between the types may in a measure be indicated, e.g. in Poecillastra cribraria the microscleres of the dermal membrane (Plate 14, Fig. 12 a) are typical syrasters, while those of the parenchyma are plesiasters (Plate 14, Fig. 126) or plesiaster- metasters (Plate 14, Fig. 12¢, 12d) and more rarely typical metasters. The lists of generic synonymy that are given include references to memoirs in which the genus as a whole is defined or in some way dis- cussed, but are by no means complete guides to the species of the several genera. In stating the size of tapering spicule rays, the thickness given is the greatest thickness, unless mention is made to the contrary. In the case of some macerated skeletons of Hexactinellids only a direct comparison with types or with determinable specimens could give any warranty for an identification. And even then a doubt, expressed by a query, remains as to the species, although the direct comparison enables one to say that forms agreeing in dictyonal framework with certain described species occur in such a region. After having made provisional identifications of the forms included in the collection, I found that before the work could be completed, it would be necessary to examine certain types deposited in European museums. The trustees of the University very kindly granted me leave of absence for the year 1902-03, for which I offer to them and to President Venable my THE SPONGES. 15 sincere thanks. To the trustees and to President Gilman of the Carnegie Institution I wish also to express my hearty thanks for a generous grant which enabled me to carry out my plans. My year was spent chiefly in Berlin, in the laboratory of Geheimrath F. E. Schulze, although visits were made to the museums in London, Paris, and Leyden. To Geheimrath Schulze I am under lasting obligation, not only for the permission to occupy a working place in his Institut, but for the generosity with which he allowed me to make use of his library, photographic atelier, and collections, in particular his magnificent collection of microscopical prepa- rations of the Hexactinellida, and finally for the helpful suggestions and friendly aid with which he responded to all of my calls for assistance. To another friend in Berlin, Professor Wilhelm Weltner, Custos in the Museum fiir Naturkunde, I am likewise under deep obligations for assist- ance in the use of the admirable collections of the Museum, and for aid of many kinds. To Geheimrath K. Mobius, Direktor of the Museum fiir Naturkunde, I offer my respectful thanks for the use of the photographic atelier, and for the kindly permission to make free use of the library and collections in the Museum. My respectful thanks are also due to Geheimrath E. von Martens of the Museum fiir Naturkunde for facilities allowed me during the course of my work in the Museum. To the following gentlemen also I beg leave to express my thanks: to Professor E. Ray Lankester, Director of the British Museum (Natural History), for permission to examine types, and to Professor T. Jeffrey Bell and especially to the curator of sponges, Mr. R. Kirkpatrick, of the same Museum, for courtesies shown me during my visit; to Professor Edmond Perrier, Director of the Muséum d’Histoire Naturelle, for permission to examine the collections, and to Professor E. L. Bouvier of the Muséum for courtesies shown me during my visit; to Professor E. A. Jentink, Director of the Rijks Museum in Leyden, for permission to examine types, and to Dr. R. Horst of the same Museum for courtesies shown me during my visit. Finally I desire to thank Mr. Agassiz not only for the opportunity of studying the valuable collection upon which I now report, but for the patience with which he has waited for the report. UNIVERSITY OF North CaARoLina, CuHapeL Hitt, N. C. Jan. 22, 1904. SYSTEMATIC ACCOUNT OF THE GENERA AND SPECIES. HEXACTINELLIDA O. Schmidt. AMPHIDISCOPHORA F. E. Schulze. HYALONEMATIDAE Gray. Hyalonema Gray. 1832. Hyalonema Gray, 1832, p. 79. 1887. Hyalonema Gray, Schulze, 1887, p. 189. 1893. a mE + 1893 a, p. 28. 1894. a i = 1894, p. 18. Hyalonema ovuliferum F. E. Schulze. 1899. Hyalonema ovuliferum ¥. Wi. Schulze, 1899, p. 13, Taf. ii. Figs. 9-12. Station 8352. One fragmentary specimen, comprising the lower end of what must have been a larger sponge than Schulze’s type. The fragment is a solid, elongated, and bilaterally compressed mass, through the middle of which the root spicules pass as a compact bundle. The root spicules are broken off 10 mm. below the rounded lower end of the sponge, and do not quite project from the upper end. Over the upper end of the fragment the dermalia and hypodermalia and the peripheral layer of parenchymal hexacts are absent, and this end doubtless represents the place at which the upper part of the body was broken away. The sponge is compact, although soft and easily torn. The surface, which is much injured, shows the apertures of numerous small canals not exceeding 1 mm. in diameter. The piece is 50 mm. long, with transverse diameters of 25 and 15 mm. respectively. . At the extreme lower end the sponge tissue round the emerging root tuft is not differentiated to form a hard and dense mass (basal collar-pad), THE SPONGES. 17 although the principalia here exhibit the usual modification. The free portion of the root tuft is 3 mm. in diameter, and includes about 12 spicules, which are in the neighborhood of 400 » thick, none exceeding 450 w in thickness. In the smooth principal oxyhexacts the rays are 250 to 850 pw long, with a basal thickness of 12 to 40 u. In the peripheral region of the paren- chyma these spicules are abundant, and for the most part regularly dis- posed, so as to produce roughly cubical meshes. In the deeper part of the parenchyma the hexacts are rare, and are arranged without regularity. ~The smooth oxydiacts measure 750 to 1700 » in length, with a thick- ness of 8 to 20 w near the median enlargement. They are comparatively scarce in the dermal membrane and peripheral parenchyma, but very abundant in the deeper parenchyma, where they run in all directions, often arranged more or less distinctly in bundles. In the smooth, hypodermal oxypentacts there is no trace of the distal ray. The tangential rays vary from a length of 200 yw, with a basal thick- ness of 16 pw, to a length of 1000 yw, with a basal thickness of 80 ». The proximal ray is commonly half again, or twice as long as the tangential rays. The spicules are very abundant, and almost alone are concerned in forming the hypodermal meshwork, hypodermal diacts being rare. The smooth micro-oxyhexacts are very abundant throughout the paren- chyma. The rays are 24 to 30 yw long, slender and straight nearly to the end, where they are rather suddenly and distinctly curved. The dermal pinules are very long and slender. The tangential rays are nearly cylindrical and then rather suddenly pointed, beset with scat- tered sharp microtubercles. On the distal ray the teeth are sharp and short, not exceeding 10 w in length (measured along upper border of tooth), becoming gradually reduced in size toward the upper and lower ends of the ray. The upper end of the distal ray in the spicules scattered over the general surface is broken off, but the ray becomes very slender above, and the shape indicates that it terminates in a long point. In the common sizes of this spicule the tangential ray measures 60 to 90 p long by 12 p» thick; the distal ray, 750 to 900 p» long, thickness of the lower smooth part of ray, 16 w. At the extreme lower end of the specimen the pinules have distal rays only about 4 as long as elsewhere. In Schulze’s type specimen the dermal pinules had a distal ray 300 to 400 » long. As regards this point, therefore, my specimen differs from the type, but the 3 18 THE SPONGES. difference cannot be regarded as of importance, since the character of the pinule is the same in both sponges. The larger variety of macramphidisce has a total length of 200 to 320 yu. The length of the bell-shaped umbel is slightly less than $ the total length of the spicule, and 1,5, times the width of the umbel. The shaft bears at its middle a circle of small protuberances, and usually other protuberances are scattered irregularly along it. In some cases there are three circles of protuberances, one in the middle, and one toward each end of the shaft. The shaft is very slender, excluding tuberosities about 5 w thick. The spicule is abundant in the dermal membrane, and is present also in the parenchyma, particularly in the peripheral region. The ellipsoidal form of macramphidisc has a total length of 60 to 80 p. The greatest width of the umbel is about equal to its length, and the umbels nearly meet at the equator of the spicule. The arrangement of the protuberances on the shaft varies. More commonly they are aggre- gated into one group at the middle. Frequently, however, there may be two such groups, on opposite sides of the middle point of the shaft. In some spicules the protuberances extend nearly over the entire length of the shaft. The spicule is abundant in the dermal membrane, and also present in the parenchyma. Small amphidiscs, which collectively may be referred to as micram- phidises, are abundant in the parenchyma. How abundant they are in the dermal membrane is impossible to determine, since the surface is injured. The smallest of these spicules are micramphidiscs of the common type, with umbels about 4 the total length of the spicule, or somewhat less. The total length of the spicule is 16 to 24 ». The umbels are as wide as deep, and with numerous (about 16) teeth. The shaft is gen- erally smooth and enlarged in the middle; in some of the larger spicules spinose in the middle. Other small amphidiscs in which the umbels closely approach or reach the equator of the spicule, and in which the shaft is spinose, form a series leading up from a length of 30 p to the ellip- soidal form of macramphidise 60-80 » long. As an example may be given a spicule measuring 30 w long by 18 » wide, in which the teeth of the opposite umbels meet at the equator. —- Still another type of small amphi- disc 30 to 50 » long, resembling in general the elongated form of macram- phidisc, is abundant in the parenchyma. The shaft bears sharp tuberosities, wud the umbels are bell-shaped, somewhat deeper than wide, and some- —_—— tM THE SPONGES. 19 thing more than } the total length. Very rarely a spicule is found intermediate in size and character between these and the smallest indu- bitable specimens of the elongated macramphidisc, which measure about 110 » in length, and themselves are not common. In the small amphi- dises, up to lengths of 50 yw, it is only possible to count the umbel rays with accuracy, in apical or approximately apical view. In such view it may often be seen that the number of rays is more than 8, being in the neighborhood of 14. At the extreme lower end of the specimen acanthophorae are found in considerable abundance. They include diactines, tauactines, and staurac- tines, in which the ends are spinose, and commonly rounded and enlarged, although sometimes pointed. In the tauactines and stauractines the rays are subequal or very unequal, 12 to 16 yw thick, and commonly less — often much less — than 200 p in length. I have examined preparations of Schulze’s type, and find that the only tangible point of difference between the two sponges is the difference in the length of the distal rays of the dermal pinules. On the other hand, the dermal pinules, the two forms of macramphidisc, and the micro- oxyhexacts, have in the two sponges the same character, even as to many minute details of structure, such as the sudden curving exhibited by the rays of the micro-oxyhexact. The type specimen was taken to the west of Prince of Wales Island (55° 20’ N.. 136° 20’ W.), at a depth of 2869 metres on a muddy bottom. Hyalonema pedunculatum, sp. nov. Plate 3, Figs. 1-6. Diagnosis. Body pipe-shaped, produced below into a peduncle bent upon the body. Gastral surface deeply concave. Canals very small, and consistency dense. Dermal and gastral pinules with bushy distal ray, 320-440 » long, ending above inacone. Micro- oxyhexacts with straight or slightly curved, minutely denticulate rays, 50-60 ,« long. Macramphidises of one kind, 120-180 » long, with wide umbels, which nearly reach the equator of the spicule. Station 8414, one specimen. The lower part of the sponge (Plate 3, Fig. 5) forms a peduncle-like process, strongly bent upon the morphological vertical axis of the body. The peduncle was broken across at a short distance from the body, the actual lower end of the sponge not being present. The upper or gastral surface exhibits a deep, narrow concavity, shown in sectional view in the 20) THE SPONGES. ~ figure, which represents the sponge after a part of the lateral surface has been sliced off. A ridge such as usually marks the passage of the upper or gastral into the lateral or dermal surface is distinguishable only on one side of the body, and does not contain peculiar marginal diacts. The gastral surface to one side of the deep cavity exhibits a sharp protuberance, which superficially suggests the central conus found in some species of Hyalonema, At another point (in the left of the figure) the periphery of the body, where gastral and dermal surfaces meet, is produced into an ear-like lobe overarching a lateral concavity. The expanded part of the body has a greatest horizontal diameter of 24 mm., and a similar depth. The peduncle, which is flattened, is 10 mm. thick in one transverse axis. and 5 mm. thick in the other. At one side of the stalk-like process, just beneath the surface, traces of the bundle of root spicules remain. Three of the spicules are still in place. These are broken across at their lower end, but may be followed upward in the sponge as far as the union of the peduncle with the expanded part of the body. In addition, seven cavities remain, from which root spicules have been pulled out. These cavities are mostly 350 to 400 » in diameter, while the actually remaining spicules range in diameter from 75 to 200 p. The entire body is very dense and firm, the canals being 0.5 mm. and less in diameter. These small canals open in some abundance on both dermal and gastral surfaces. The dermal and gastral membranes are not reticulate, but exhibit where uninjured a thick furze of projecting pinuli, which is discernible to the eye. The parenchymal macroscleres are smooth oxyhexacts and smooth oxydiacts. In the oxyhexacts the rays are subequal and tapering. In the commoner sizes the ray measures 420 x 24 » to 1 mm. x 70 wp. The spicule is only fairly abundant. The oxydiacts are exceedingly abun- dant, scattered in all directions in the upper body, arranged predominantly lengthwise in the stalk. There is commonly an enlargement, often very slight, at or near the middle, and the rays taper evenly. The spicule measures 1-2 mm. in length, 12-20 » in thickness. The hypodermalia and hypogastralia are abundant and alike. They are smooth oxypentacts with no remnant of the distal ray, the other rays tapering evenly. The size varies considerably. In a common size the tangential ray measures 370 » x 36 yw, the proximal ray 670 » x 40 p, but much smaller spicules are abundant. THE SPONGES. 21 The micro-oxyhexacts are extremely abundant throughout the par- enchyma. The rays are straight (Plate 3, Fig. 3) or slightly curved (Plate 3, Fig. 1), taper gradually to a fine point, and are beset with very minute sharp prickles. The ray is 50-60 yw long, and 4 yp thick at the base. Both the straight and curved varieties are common, although the form with straight rays predominates. The dermal and gastral pinuli are alike. They are large, strong pen- tact pinuli (Plate 3, Fig. 4), with no trace of the proximal ray. The distal ray is 320-440 » long, with a greatest diameter of 30-60 yp, anda basal diameter of 10-16 p. The stouter forms, in which the distal ray has a distinctly fusiform outline, as in the figure, are the typical spicules, although slenderer forms in which the distal ray is nearly cylindrical occur. The ray ends above in a terminal cone, and the upwardly projecting scales are narrow and sharp. These degenerate in the lower third of the ray to prickles, below which the ray is smooth. The scales ,in the thickest part of the ray have a length, measured along their upper border, of about 16 w. The tangential rays are 40-50 w long and 8-12 p thick; about cylindrical and then suddenly curving to the point; with a few scattered, sharp, microtubercles. On the surface of the peduncular part of the body, only a few pinuli remain, and over much of the dermal surface they have been lost. On the gastral surface and on the uninjured parts of the dermal surface, they are thickly crowded. In the walls of the larger canals a few canalar pinuli are to be found. They are of the same general type as the dermal and gastral pinuli, but with a shorter and relatively more slender distal ray. Macramphidises of the type shown in Fig. 2, Plate 3, are very abundant in the dermal and gastral membranes. The shaft is smooth, and the umbels closely approach the equator of the spicule. The umbels are wide, evenly rounded, not truncated apically, and include 8 tongue-shaped rays, which are rounded at the free end.” A typical spicule has the follow- ing measurements: total length of spicule, 160 ; width of umbel, 100 p; depth of umbel, 70 »; greatest width of umbel ray, 24 p. The total length of the spicule commonly varies from 120 to 180 yp, but larger spicules are exceptionally found which reach a length of 200 to 250 pw. The umbel rays are sometimes not strictly rounded at the free end, but round-pointed, and occasionally even sharp-pointed. Mesamphidiscs having the character shown in Fig. 6, Plate 3, are 22 THE SPONGES. abundant in the walls of some of the canals, and are scattered in the parenchyma. The shaft bears small, scattered, irregular tubercles, often a circle of tubercles, in the middle. The umbel has a deep bell shape, and is more than } the total length of the spicule. The umbel rays are 8 in number, long, narrow, and pointed. A typical spicule has the following measurements: total length, 604; length of umbel, 244; width of umbel, 20. The total length of the spicule varies from 40 to 65m. Micramphidiscs of the common Hyalonema type, 25 to 30 p long, are fairly common in the parenchyma. Transitional forms between these and the mesamphidiscs are easily found. The skeletal resemblances between this species and /Hyalonema bian- choratum, sp. nov., are striking, involving as they do the characteristic spicules, macramphidises, dermal pinules, and oxyhexacts (comp. Plate 2, Figs. 2, 5, 10, and 11). Nevertheless the form of body is very different in the two species, and in H. bianchoratum there are two types of macramphidise. Moreover, the macramphidise of H. pedunculatum never assumes the shape with truncated poles and flattened sides which is common in HH. branchoratum (Plate 2, Fig. 1). The resemblance is thus only partial, although close in the parts concerned. Hyalonema bianchoratum, sp. nov- Plate 2, Figs. 1-11. - Diagnosis. Body cup-shaped with deep gastral cavity. Numerous efferent canals open independently on the gastral surface ; their apertures covered in by the gastral membrane. Micro-oxyhexacts with minutely denticulate rays about 50 » x 4»; rays commonly slightly curved or straight. Dermal and gastral pinuli are alike, and pentacts; distal ray 250-530 » long, comparatively stout, with long appressed upper spines and outwardly projecting lower spines, with a terminal cone. Macramphidiscs of two types. In one type the umbel is ; to 3 total length, with acutely pointed rays; umbel width greater than its depth; total length about 450 ». In the other type umbels closely approach equator of spicule; rays rounded at the end; umbel wider than deep; total length, 150 p-250 p. Station 5415, one specimen. Sponge body (Plate 2, Fig. 6) is cup-shaped and, possibly owing to packing, is laterally compressed. The entire cup is 50 mm. deep, and has a greatest width of 75mm. The cavity of the cup is 30 mm. deep, and the wall about 12 mm. thick. The wall thins away toward the free edge, but a marginal fringe separating the inner or gastral surface from the outer THE SPONGES. Za or dermal surface, is not present, possibly owing to the bad preservation of the specimen. The root spicules have been pulled out. The wall of the cup 1s excavated by numerous canals 4 mm. and less in diameter, and the consistency of the sponge is soft and flabby. The dermal membrane is badly injured, the gastral membrane less so. The latter is not separated from the sponge tissue, but simply passes over the apertures of numerous efferent canals. The parenchymalia principalia are chiefly oxydiacts, which vary greatly in size and considerably in details of shape. There are many slender, often slightly curved, nearly cylindrical forms, with slight enlargement at or near the middle showing an axial cross. Common sizes are 1.5 to 2 mm. long by 20 thick. The ends are sometimes rounded and slightly enlarged. The spicule really tapers slightly from the middle toward the ends, and when the length is short (0.5 to 1 mm.) the outline becomes noticeably fusiform. There are some similar but much larger forms, connected by intermediate stages with the above. These may reach a size of 6 mm. x 34y. There is no median enlargement. There are other stouter diacts of a distinctly fusiform shape, ranging in size from 600 » x 20 p to 1350 w» x 60. These exhibit an enlargement with axial cross. The enlargement may be faint or conspicuous, and may or may not extend quite round the spicule. This diact is not common in the interior, but is the predominating form at the dermal and gastral surfaces. Other principalia are smooth oxyhexacts, which are scattered through the parenchyma in some number. ‘The rays, which are not always of the same length, commonly range from 200 to 850 » in length. A consider- able number of large forms are present, having a ray length up to 3.5 mm. The hypodermal and hypogastral pentacts are alike, and are strong, smooth oxypentacts, with no remnant of the distal ray. The tangential rays commonly vary in length from 350 to 850 ». The proximal ray may be somewhat shorter than, or two or three times as long as, the tangentials. Micro-oxyhexacts are abundant throughout the parenchyma. The spic- ules are strong spicules with rays 40-60 w long and 3-4 yp thick at the base; commonest size of ray about 50 4x4 y. Very minute, sharp den- ticulations are scattered along the ray, which tapers to a fine point. Spicules the rays of which are slightly curved, as in Plate 2, Fig. 10, predominate, but abundant straight-rayed spicules (Plate 2, Fig. 11) occur, and not infrequently spicules are met with in which the rays are 24 THE SPONGES. conspicuously curved as in Plate 2, Fig. 9. When the rays are curved, the curvature is of the usual character, viz. opposite rays of a diameter are bent in opposite directions. The dermal and gastral pinuli are alike, and are pentacts. The distal ray (Plate 2, Figs. 2, 3) is comparatively stout, ending above in a ter- minal cone. In the upper half of the ray the spines are long, narrow, and appressed. Below they are shorter and project outward in hook-like shape, degenerating farther down into a few prickles, The lowest part of the ray is smooth. The tangential rays bear a few scattered, sharp microtubercles. The range of size is considerable. Large spicules are abundant (Fig. 2, Plate 2) in which the distal ray is 530 uw long, with a greatest thickness of 64 » and a basal thickness of 24 wu ;. tangential rays, 44 4 x 16; spines on distal ray reaching a length of 40 p, measured along their upper border. Smaller spicules (Fig. 3, Plate 2) are abundant down to a size in which the distal ray is 250 w long, with greatest thick- ness of 32 uw, and a basal thickness of 10 »; tangential rays, 30 w x 10 p. Still smaller sizes (Fig. 4, Plate 2) occur, although not commonly, in which the distal ray may be only 175 y long. Some of the shortest spicules have tangential rays as long as are met with in any of the pinuli, the length reaching 60 p». Although the size of the pinuli varies within such wide limits, the character remains fairly constant. The character -is expressed especially in the terminal cone, the long and narrow appressed upper spines, and the lower hook-like spines. The macramphidises are of two types. The first type of macramphidisc, Fig. 8, Plate 2, is only moderately abundant. It occurs, in the present condition of the specimen, both at the surface and in the interior. The shaft is slightly expanded, not always symmetrically, at the middle, and is smooth or bears one or two scattered small, sharp protuberances. The umbel is rather evenly rounded, although somewhat truncated at the apex, where there is a depression. It is considerably wider than deep, and its depth is from } to 2 the total length of the spicule. There are 8 rays which are acutely pointed. The variation in size is not great, and a typical spicule has the following measurements: total length, 455; depth of umbel, 122 1; greatest width of umbel, 188 1; greatest width of umbel ray, 36 »; thickness of shaft, at the middle of spicule, 30 p. The second type of macramphidise is more abundant. It occurs espe- cially at the surface, or in the peripheral parenchyma. In this spicule THE SPONGES. 25 (Figs. 1 and 5, Plate 2) the shaft is smooth and only slightly thickened at the middle. The rays, 8 in number, are wide, tongue-shaped, rounded at the free end, and closely approach the equator of the spicule, rays of opposite umbels alternating. There is a depression at each pole of the spicule. The precise shape of the umbel varies. In many spicules (Fig. 5, Plate 2) it has an evenly rounded outline, while in others (Fig. 1, Plate 2) it is truncated at the pole and flattened on the sides. The two varieties shade into each other. A typical spicule with evenly rounded umbel has the following measurements: total length, 188 »; depth of umbel, 80 »; greatest width of umbel, 128 yw; shaft, 18 » thick at the middle. A typical spicule with truncated and flattened umbel has the following measurements: total length, 210 »; depth of umbel, 95 1; greatest width of umbel, 152 »; shaft, 20 » thick at the middle. Mesamphidiscs (Fig. 7, Plate 2), 60-80 mw long, are present, but in small number, in the parenchyma. The shaft bears scattered, minute, sharp denticulations, and often but not always a circle of somewhat larger similar protuberances at the middle. The umbels are deep bell-shaped, slightly deeper than wide, and a little more than 4 the total length. The umbel rays are 8 in number, rather narrow, and with pointed free end. A typical spicule measures: total length, 80 ~; umbel depth, 36 w; umbel width, 30 wp. Smaller amphidiscs, micramphidises, agreeing in shape with the mesam- phidiscs, are scantily present. They range down to a total length of 28 pu. Hyalonema, species diversae. Plate 2, Figs. 12-16. Under this heading I briefly describe several specimens so fragmentary that the shape of body cannot be inferred with any approach to certainty. The skeletal elements of them all, especially of the forms designated Hyalonema 1 and Hyalonema 2, offer close resemblances to those of H. bianchoratum. Hyalonema 1. Station 3381, a fragment, apparently from the lower end of the sponge, including a part of the bundle of root spicules and surrounding tissue. The bundle of root spicules in its thickest part has a diameter of 5 mm., and the larger spicules a diameter of 700 yp. 4 THE SPONGES. bo oO The micro-oxyhexacts agree with those of HH. lianchoratum. The dermal pinules also resemble those of the latter species, differing only in that the upper spines on the distal ray are somewhat longer, the ray thus appearing more bushy. —— There are two types of macramphidise essen- tially like those of H. bianchoratum. In the case of the larger type, the total length reaches 510 »; the shaft bears 3 or 4 scattered sharp micro- tubercles; the depth of the umbel is 4 the total length; spicule fairly abundant. In the case of the shorter type, the umbel rays commonly reach equator of spicule, often slightly interdigitating ; total length reaching 255 »; larger sizes of the spicule truncated at the poles and with flattened sides, as in Fig. 1, Plate 2,——- The smaller amphidiscs, 24—70 mw long, do not differ from those of H. bianchoratum. ent and include stauractines and diactines. Spicules are spinose only at the ends of the rays; ends often rounded and enlarged, also pointed. The spiculation indicates that the fragment belongs to H. bianchoratum, perhaps Acanthophorae are pres- representing a local variety of this species. Hyalonema 2. Station 3414, a fragment from the peripheral part of the sponge includ- ing both dermal and gastral surfaces, and about 50 mm. wide with a greatest thickness of 20 mm. On the gastral surface are the apertures of several canals about 7 mm. wide. The micro-oxyhexacts agree with those of H. bianchoratum, but forms with straight rays predominate. The pinules of the dermal and gastral There are two surfaces agree closely with those of H. bianchoratum. types of macramphidisc, shown in Figs. 13 and 16, Plate 2, which differ only in minute details from the macramphidises of Z. bianchoratum. In the larger type, Fig. 13, the total length is 250-320 »; shaft smooth or with a few scattered sharp tubercles and often with 4 or 6 protuberances at the middle. A typical spicule measures: total length, 280 «; depth of umbel, 90 ~; width of umbel, 130 ». In the smaller type, Fig. 16, the total length is 100-150 yw; rays of opposite umbels often slightly inter- digitating at the equator of spicule. The umbels are never conspicuously truncated and flattened as in many of the spicules of H. bianchoratum (Fig. 1, Plate 2). A typical spicule measures: total length, 144 »; depth of umbel, 70 4; width of umbel, 100 p. In addition, amphidises of the type shown in Figs. 14 and 15, Plate 2, are fairly abundant. The total length is 150-180 »; umbel depth varying THE SPONGES. 27 from 4 nearly to } total length; shaft with a few scattered tuberosities, and frequently but not always with a ring of tuberosities at the middle. The umbel is‘often evenly rounded, as in Fig. 15, but as frequently some- what truncated at the poles and flattened along the sides, as in Fig. 14. These spicules, which owing to their size may be regarded as a third type of macramphidisc, pass through intermediate stages into very similar mesamphidiscs, 60-80 » long, in which the umbel depth somewhat exceeds } the total length. The mesamphidises are connected by transitional forms with micramphidises, 18-20 p» long, of the character usual in LHyalonema. The skeletal resemblances to H. bianchoratum are extensive and close. The only important point of difference is afforded by the third type of macramphidisc, which might properly be designated as an enlarged mesamphidise, since it is connected by an unbroken series with the latter. Hyalonema 3. Station 3414. Three fragments, all including the lower end of the sponge with the root tuft. The largest piece is a laterally compressed triangular mass 45 mm. wide by 50 mm. high. The root tuft where it adjoins the sponge is 5 mm. thick, and includes about 25 spicules varying in thickness from 200 p to 730 w. The spicules are all broken off below, the fragment of root tuft measuring 100 mm. in length. The tuft bears no anemone. At the base of the body there is a conspicuous firm and dense collar-pad surrounding the root spicules. The other two fragments are likewise laterally com- pressed triangular masses, but from smaller sponges. The upper diameter of the root tufts is something less than 2 mm., and the spicules do not exceed 250 » in thickness. Again the lower ends are all broken off, although the tuft in one case is 200 mm. long. Both tufts at the upper end are surrounded by small Pulythoa colonies, each including two indi- viduals. In these specimens the basal collar-pad has not developed. —— It is noteworthy that in all three specimens the lower end of the sponge body has a triangular outline, and is laterally compressed to a marked degree. The surface is so injured that it is not possible to reach a conclu- The micro-oxyhexacts, sion as to the character of the dermal pinuli. which are abundantly scattered throughout the parenchyma, are smail slender forms (Fig. 12, Plate 2) with rays 30-36 pw long by 2 yp at the 28 THE SPONGES. base, tapering gradually to the point. Rays are faintly roughened, almost smooth; slightly curved or straight, both types common. Two types of macramphidise quite similar to the spicules of Hyalonema 2, which are shown in Figs. 15 and 16, Plate 2, are sparsely present. A third type of macramphidise, quite like the spicule of Hyalonema 2, which is represented in Figs. 14 and 15, Plate 2, is present in great abundance. The spicule measures 150-180 yw in length, and is connected by abundant transitional forms with similar mesamphidises 40-60 » long. Micram- phidiscs, of the usual character, 16-20 » long, are abundant. The similarity to Hyalonema 2 afforded by the amphidises is striking. Mention may here be made of imperfect tufts of root spicules taken at Stations 3381 and 3399. The spicules resemble those of the specimen of Hyalonema pateriferum from Station 3376. Hyalonema pateriferum, sp. nov. Plate 1, Figs. 1-13. Diagnosis. Body obconical to saucer-shaped, the upper surface approximately flat, the under surface very or moderately convex. No main gastric cavity, but instead a comparatively large number of efferent canals covered in by the gastral membrane. Root spicules thick, about 1 mm. in diameter. Characteristic micro-oxyhexacts with slender, curved, and minutely denticulate rays 30-45 » long. Dermal pinules, slender forms; distal ray, 150-200 » long, ending in a slender point. Characteristic macramphidises 100-200 » long, with smooth shaft and umbels which have the shape of wide, shallow bowls; depth of umbel about } the total length, or shallower. Station 3376, 1 specimen; Station 3363, 1 specimen and 3 fragments; Station 17 of the “ Albatross” 1900 cruise, 2 specimens. In the specimen from Station 5376 (Figs. 12 and 13, Plate 1), the body is irregularly obconical, and is somewhat compressed in one of the morpho- logical horizontal diameters. The root tuft emerges from one side of the base, and exhibits an open spiral curvature. The general asymmetry is doubtless an individual feature, associated with the position assumed by the animal in its habitat. If in the natural position the root tuft was verti- cal, then the body of the sponge probably lay under a stone or some such object, the gastral face looking out from under this protection. The sponge body is 65 mm. high and has a greatest width of 90 mm. The a wees THE SPONGES. 29 root tuft is 330 mm. long, 6 mm. thick, near the body, and includes about 50 spicules, most of which are thick, having a greatest diameter of about 1 mm., although much slenderer ones are intermingled. At its base the tuft is embraced by a single Actdwa-like anemone of a yellow-brown color and firm leathery consistency. A marginal ridge is sharply defined round the greater part of the gastral surface. The membrane covering the gastral surface is in general a typical reticulum, but in places near the periphery of the surface the reticulum is not developed, the membrane here appearing to the eye dense and per- forated only by scattered oscula about 1.5 mm. in diameter. The mem- brane covers in a relatively large number of efferent canals, the transverse diameter of which at the surface is 5-10 mm. These canals extend more or less vertically into the body, and largely excavate it. Some of them are easily traceable from the gastral surface to the base of the body. The dermal membrane, covering the lateral surface of the body, is a reticulum with a somewhat finer mesh than the gastral membrane. Over a considerable part of the surface the dermal membrane has been abraded, disclosing the canals beneath. These, presumably afferent, canals are in general smaller than the canals opening on the gastral surface. At the base of the body there is a rather inconspicuous collar-pad, surrounding the root tuft and containing the characteristic acanthophorae. The entire specimen from Station 5563 is a saucer-shaped mass, with a flat upper and convex lower surface. The root tuft has been torn out, leaving an evenly bounded aperture, 6 mm. wide, about in the centre of the lower surface. The sponge body has a depth of 40 mm., the hori zontal diameters being respectively 85 mm. and 65 mm. _ In other respects, the description given of the specimen from Station 3376 applies to this specimen. | The better specimen from Station 17 (1900 cruise) resembles in essential shape the one from Station 3363, but is even flatter. The depth is 15 mm., the horizontal diameters respectively 80 and 60 mm. The root tuft has again been torn out, leaving a somewhat irregular aperture 7 mm. in diameter. The gastral membrane, covering the upper surface, appears continuous and not reticular, the probable explanation being that the apertures are closed. The dermal membrane, covering the under surface, is reticular as in the other specimens. Toward the periphery the body thins away to a sharp margin, which is somewhat injured. The other 30 THE SPONGES. specimen from this station is fragmentary, but appears to have belonged to a much-flattened sponge similar to the one just described. It is of course possible that in the packing these specimens have been artificially flattened. There is, however, every indication that the present shape is approximately the natural one. The parenchymal macroscleres are oxydiacts and oxyhexacts. The oxydiacts are very abundant, scattered in all directions through the par- enchyma, chiefly in tracts, but also singly. They are smooth, with or without a median swelling, the rays tapering evenly to points. The size varies from 500 » long by 8 p» thick, near the middle, to 3 mm. long by 28 p thick. smooth, straight, and tapering, measuring in a typical case 700 p long The oxyhexacts are few in number. The rays are by 48 p thick at the base. The hypodermalia and hypogastralia are alike, and are smooth oxypentacts and oxydiacts. The oxypentacts are abundant, with no trace of the distal ray, the other rays tapering to points. They vary in size from small ones with tangential rays 150 » x 12 » to large ones with tangential rays 600 » x 48 pp. The proximal ray is in general longer than the tangential rays, but on the gastral sur- face it is sometimes shorter. The diacts are abundant, sharing with the tangential rays of the pentacts in forming the supporting reticulum of the surface. They are similar to the diacts of the interior. The common and characteristic form of parenchymal microsclere is a micro-oxyhexact with slender, curved, and minutely denticulate rays, 30-45 p long and 2 yp thick at the base (Fig. 9, Plate 1). The denticula- tions are sharp and just perceptible, with a power of 600 diameters. The curvature is well marked. Mingled with these are a good many similar oxyhexacts, in which the ray is either straight (Fig. 8, Plate 1) or shows only a very slight curvature. These average a larger size, the ray length. being 40-60 ». Oxyhexacts and oxypentacts with straight, minutely den- ticulate rays, 60-80 » long, are found sparsely in the parenchyma, and in or near the walls of some of the canals. They are probably in all cases canalaria. The dermal pinules (Fig. 3, Plate 1) are slender spicules, in which the distal ray ends in a pretty long slender point (“endspitze’’), and bears sharp spines, which are not thickly crowded. The lower part of the distal ray, z to } the total length, is smooth. The tangential rays are smooth, or nearly so, and pointed. There is no trace of the proximal ray. The distal — THE SPONGES. 31 ray is commonly 150-200 » long, with a basal thickness of 5 »; tangential rays 30-40 p long. Spicules in which the distal ray is only 100 yp long, or as long as 220 p, occur. The pinules are abundant, and their tangential rays rest upon the reticulum formed by the hypodermal diacts and the tangential rays of the hypodermal pentacts. The gastral pinules are in general like the dermal pinules; but long, very slender forms of the type shown in Fig. 2, Plate 1, occur. In these spicules the spines on the distal ray are so reduced in size as to be mere prickles. The distal ray is 300-400 » long; tangential rays about 40 yp long. Transitional forms between these and the ordinary type of pinulus occur. The gastral pinules are arranged in the same way as the dermal. Canalar pinules are rather sparsely scattered over the walls of the larger efferent canals. They are similar to the common type of dermal and gastral pinulus, but are somewhat shorter. The distal ray is 110-150 pu long; tangential rays 40 w long. Only a comparatively small percentage of the spicules rest upon the underlying diacts. The marginal ridge is well preserved in only one specimen, that from Station 3376. Oxydiacts (marginalia) in considerable number project radially from the ridge. Many are like the common hypodermal diact ; ?. e., smooth and enlarged at the middle. Others are specialized marginalia, in which the outer projecting half is covered with small, sharp spines, while the inner half is smooth (Fig. 1, Plate 1). An occasional prickle is found in some spicules on the inner half. A typical spicule is 700 » long and 12 p» thick near the middle. This form of marginal diact is shorter and slenderer than the common sizes of the smooth form. The characteristic macramphidisc of the species is shown in Figs. 5, 6, and 7, Plate 1. The shaft is ordinarily smooth, very rarely with one or two rounded protuberances at the middle. The umbels have the shape of wide shallow bowls; depth of the umbel about { total length of the spicule, or less; umbel rays 8 in number, broad and leaf-like, pointed. In the specimens from Stations 3376 and 3363 the macramphidiscs are alike. In these specimens the total length of the spicule is 100-200 yp; thickness of the shaft at the middle 8-16 »; depth of the umbel varying somewhat, but close to 1 the total length. One of the smaller sizes is ‘shown in Fig. 5, one of the larger in Fig. 6. In the specimens from Station 17 (1900 cruise), macramphidises occur that are similar to those of the other specimens, but in most of the spicules the umbels are very 32 THE SPONGES. hel shallow, and with noticeably flat tops. The umbel depth is commonly about 1 the total length. The total length is the same as in the other specimens. A typical macramphidise of this specimen is shown in Fig. 7, Plate 1. gastral membranes of all specimens. A few are found in the internal The macramphidiscs are very abundant in the dermal and parenchyma, but this position may not be natural. Mingled with the characteristic 8-rayed macramphidiscs are a consider- able number of amphidises of the type shown in Figs, 10 and 11, Plate 1. The umbel is usually 4-rayed, but umbels with 5 and 6 rays occur. The shaft is smooth, and the umbel rays very similar to those of the 8-rayed form. The spicule varies somewhat as regards the precise shape of the umbel, which in some spicules (Fig. 10) is deeper than in others (Fig. 11). The total length is 60-100 p. It is possible that these spicules represent young stages of the 8-rayed form. Mesamphidises (Fig. 4, Plate 1), 50-80 uw long, are abundant. The shaft bears scattered small tubercles in varying number. The umbel is deep bell-shaped, with 8 pointed and rather narrow rays. The umbel depth is slightly more than } the total length. Micramphidises of the common flyalonema type, 20-25 p ie are abundant. Intermediate forms between these and the mesamphidises are common, and especially abundant in the walls of the main efferent canals. Pathological amphidiscs of small size are occasionally observed, similar to those described by Marshall and Meyer, 1879, p. 261, Taf. XXV Figs. 19, a, 6, ec. In one such, which measured 72 p» in total length, 3 rays of the hexact had developed umbels which were not quite alike, one ray was club-shaped at the end, and a fifth ray had the form of a short spine. The lower end of the body is well preserved in only one specimen, that from Station 3376. The collar-pad here bears on its outer surface the ordinary dermal skeleton, but round the root spicules there is a thick dense layer of acanthophorae. In these, only the ends of the rays are spinose. The ends are frequently but not always slightly enlarged. The spicules include diacts, tauacts, stauracts, pentacts, and hexacts. Stauracts with subequal or very unequal rays are the commoner forms. In a repre- sentative diact, the total length is 900 p. Ina typical stauract the length of the longest ray is 250 p. The basalia vary in greatest diameter from 150 » to 1 mm. The spicules in their lower portions taper rather rapidly and over a consid- THE SPONGES. 33 erable distance, 60 mm. or so, exhibit the well-known annular ridges. The lower ends are broken off. Of the known species of Hyalonema, the form uere described stands nearest H. conus F. E. Sch. (Schulze, 1887, p. 209; 1893, p. 35). A precise point of difference concerns the micro-oxyhexacts which in 7. conus have straight and distinctly roughened rays 50-60 » long. The pinules and macramphidiscs are very similar in the two species, although in WZ. pateriferum the macramphidisc umbels are wider and shallower, especially in the specimens from Station 17 (1900 cruise). HEXASTEROPHORA F. E. Schulze. EUPLECTELLIDAE Gray. Euplectella Owen. 1841. Huplectella Owen, 1841, pp. 3-5. 1887. Huplectella Owen, Schulze, 1887, p. 53. 1901. c «© Tjima, 1901, pp. 37-58. Euplectella, sp. Plate 4, Fig. 4. Station 3404, 6 imperfect specimens. The specimens include only the dictyonal framework, with no free spicules, and thus a closer identification is impossible. The specimens represent in all cases the lower part of the body, which is somewhat curved and tapers to the inferior apex. The paragastric cavity is limited below by a “bottom plate.’ The extreme lower end of the body is worn in most of the specimens, but in one the longitudinal bundles of the wall are continued below the body, curving toward one another so as to form a conspicuous cone, which extends 6 mm. below the bottom plate. The wall is made up of the longitudinal, transverse, and oblique silicious strands characteristic of Huplectella, all firmly united together by exceed- ingly numerous synapticula. The longitudinal tracts on the outer surface are strongly, whereas the transverse tracts on the inner surface are feebly, developed. Along the former, and also along the oblique tracts, are here and there developed isolated protuberances, which in places are united to form parietal ridges. The parietal apertures are rounded and for the most part arranged in fairly regular transverse or oblique rows. 5 34 THE SPONGES. The specimens are of interest in that the cavity of the sponge is crossed by one or more horizontal or oblique septa. The septa (Fig. 4, Plate 4) are obviously similar structures to the well-known “bottom plate,’ which has been observed in several species of Huplectella (Schmidt, 1880, p. 60; Schulze, 1895, p. 17; Schulze, 1902, p. 53; Ijima, 1901, pp. 40, 92, 103, 207). The septa are spongy, easily broken, and com- posed of a reticular tissue formed by the continued deposition of silica chiefly round small diacts, with the development of synapticula. The diacts discernible in the beams of the reticulum have rounded and slightly _ enlarged ends. The proper skeletal strands of the sponge wall take no share in forming the septa. In the specimen figured the septa are arranged one above the other. In the other specimens there is but one septum. The septa are thicker at the periphery, thinning out toward the centre. Except in one case they are perforated by several rounded apertures, about the size of the parietal apertures. In the case referred to, the septum is imperforate, but is exceedingly thin in spots. On its upper surface the reticular beams have a predominantly radial disposition, and moreover are vaguely divided into groups, each of which radiates from a particular part of the body wall. This arrangement suggests that the septum arises as a number of centripetal outgrowths from the wall, the outgrowths meeting and coalesc- ing. And in fact, in the specimen figured, three independent outgrowths of this character, all lying in the same transverse plane, exist above the uppermost septum. The reticular tissue of the septa, at the periphery of the latter, is prolonged as a thin unevenly developed layer over the adjacent parts of the sponge wall. In some of the specimens, in which there are no septa, this reticular layer is developed over extensive areas of the gastral sur- face, and in a less degree over the outer surface. In such places the tissue is thick enough to hide completely from view the proper composi- tion of the wall, passing over and so closing the parietal apertures. The union of the smaller parenchymalia into a finely reticulate tissue, which in the one case spreads over the surface of the sponge, and in the other crosses the paragastric cavity, would seem to be everywhere essentially the same phenomenon, perhaps having a definite physiological function. The structure of the “bottom plate” in the specimens studied is similar to that of the septa. OO EE THE SPONGES. © Or Regadrella O. Schmidt. 1880. Regadrella phoenix O. Schmidt, 1880, p. 61. 1887. Regadrella O. Schm., Schulze, 1887, p. 84. 1901. Se i Tjima, 1901, p. 220. Regadrella, sp. Plate 9, Fig. 9. Station 3380, two fragmentary specimens including only the macerated skeletal framework. In both specimens the base is preserved with the lower part of the tube wall. The better specimen is figured (slightly above the natural size, owing to an accident in the taking of the photograph). In the other specimen the base is smaller, and less of the wall is included. The base in each is a nearly flat plate with few irregularities. The skeletal strands forming the wall are cemented together at the points of crossing, and the wall as a whole is somewhat flexible and elastic. I have had for comparison specimens of R. okinoseana Tjima and R. phoenix O.Schm. As compared with the former species, and in a less degree as compared with the latter, the “ Albatross” specimens are remarkable for the thin character of the parietal strands and the consequent large size of the meshes. As compared with the only other species of Regadrella the whole body of which is known, &. kameyamai Tjima, it would seem from Tjima’s description (1901, p. 257) that the parietal strands in my specimens undergo a more extensive fusion. These specimens again differ from the described species of Regadrella in the greater regularity of arrangement displayed by the skeletal strands. On the gastral surface the obliquely transverse beams, which are very strongly developed, are arranged parallel to one another. Crossing them at about right angles are ascending bundles. The meshes would thus be squarish, but oblique fibres extending both to the right and left cross the meshes usually at the corners, thus giving rounded apertures. The oblique fibres may cross the middle of the mesh in such a way as to obliterate the aperture. The arrangement of the skeletal strands thus approaches the regularity found in Luplectella. The coarser skeletal strands are made up, each, of one or a few large diacts surrounded by very slender diact comitalia, all united by and covered with cement. The principal diacts are 120 « or somewhat less in 36 THE SPONGES. transverse diameter, tapering gradually toward the ends, which are apt to break off or remain concealed in the cement. The comitalia are very slender, often only 6 or 8 yw thick. The length of the principalia may exceed 20 mm., and they are frequently though not always bent at the middle, as in the similar spicules of A. phoenix (Schulze, 1899, p. 21). The reticulum of silicious beams constituting the basal plate presents no peculiarly characteristic features, agreeing in general with the description given by Ijima (1901) for &. okinoseana and Lt. kameyaman. Regadrella delicata, sp. nov. Plate 3, Figs. 7,8. Plate 4, Figs. 2,11. Diagnosis. The marginal spicules round the sieve-plate are stauracts, the longitudinal axes of which are not included in the skeletal bundles of the body wall. The superior rays of the stauracts are accompanied by slender parenchymalia, not by the parenchy- malia principalia. Species close to Regadrella phoenix O. Schm., agreeing with it in general spiculation. Station 3404, one specimen. The specimen is fragmentary, including a part of the lateral wall of the sponge, which had been ripped open longitudinally. In packing, the piece was compressed so that it reached me in the shape of a folded plate, the line along which the folding had taken place corresponding with the long axis of the sponge. This plate-like fragment had a length of 150 mm. and a greatest width of 50 mm. Its upper edge is formed by the margin of the sieve-plate area, and seems to be uninjured. The dermal and gastral sur- faces of the lower part of the piece are shown in Fig. 11, Plate 4. When the two halves of the specimen, which had been folded together, were separated, some idea of the natural shape of the sponge was obtain- able, and it could be seen that the sac tapered markedly toward its upper end. The upper edge is 25 mm. long. How much of the natural periphery of the sieve-plate area this represents could not be ascertained, although the curvature of the wall indicates that it represents a very large part. | The wall of the sac is thin, about 1.5 mm. thick. The rounded parietal apertures are mostly 2 to 1.5 mm. in diameter, except in the extreme upper part of the sponge, where the diameter is very commonly about 1 mm., although much smaller apertures are here present. The apertures exhibit an imperfect arrangement in oblique or nearly transverse rows, and are LOE THE SPONGES. 37 mostly 3-5 mm. apart; more closely crowded in the upper part of the sponge, where the interval is 2 mm. to 0.75 mm. On both dermal and gastral surfaces, the small apertures of the numerous canals, 0.5 mm. and less in diameter, give the sponge a porous appearance. The gastral surface is smooth, and the dermal surface, which is without prostalia, exhibits no elevations, except that as in other species of the genus the coarse skeletal bundles project. These latter bundles pursue a nearly longitudinal or somewhat oblique course, converging toward the lower end of the fragment, where there are some cases of anchylosis. Elsewhere there is no, or only the feeblest anchylosis. The obliquely transverse bundles on the gastral surface are not large enough to cause elevations of the gastral membrane. In the uppermost part of the body the transverse arrangement of these bundles is more marked than elsewhere. The spiculation agrees with that of &. phoenix as described by Schulze (1887, p. 84; 1895, p. 34; 1899, p. 20) and Ijima (1901, p. 265), except in a few details. And the close resemblance to &. phoenix possibly indi- cates that the specimen represents a late stage in the development of that species. The principal diacts, smooth and tapering to a point at each end, bent at the middle or more evenly curved in bow-like shape, reach a diameter of 300 » and a length of 30 mm. There are abundant smaller sizes down to 12 mm. x 170 yu, below which still smaller spicules with the character- istic shape down to 6.5 mm. x 90 p» are found. The slender cylindrical comital diacts of various lengths are mostly 6-8 p» thick. They are swollen in the middle, and usually with subtermi- nally roughened ends, which are often but not always enlarged. Similar diacts, 6-30 yw thick, constitute the smaller bundles or lie loose in the par- enchyma. Other small parenchymalia, intermediate between the diact and hexact condition, with cylindrical rays rounded at the ends, are also found. The tauact and stauract forms are the commonest. The comital diacts are not cemented together over the principalia, and the latter are only loosely combined. to form bundles. Many of the prin- cipal diacts lie scattered through the parenchyma, unassociated in bundles. Some of these are without comitalia, although in general so provided. To form a long bundle the ends of succeeding diacts overlap, and become covered with a continuous sheath of comitalia. Very commonly 4 or 5 38 THE SPONGES. diacts, each with its own sheath of comitalia, lie side by side, but separated by considerable intervals, thus forming a tract, but not a single bundle. Frequently such diacts in one part of their course will be closely bound together, while separate elsewhere. This is the commonest arrangement, although there are bundles consisting of 2 to 4 parallel diacts, which are closely bound together by comitalia throughout their length. The principal diacts in the extreme upper part of the body average a smaller size than elsewhere, the larger ones here measuring only 10-13 mm. in length by 200-220 w in thickness. Some of them exhibit the characteristic bend, which is frequently not at the middle, while others are gently curved or nearly straight. They are in general surrounded by relatively few comitalia and lie separately or in approximately longi- tudinal bundles composed of a few (2-4) spicules. In one. part of the marginal region, about 5 mm. wide, the parietal apertures are scarcely developed, being here few and small and not rounded but irregularly elongated gaps. In this small region the principal diacts show what is probably the general arrangement before the apertures develop. They are here arranged close together in an approximately longitudinal direc- tion, radiating toward the upper margin, reaching or nearly reaching or projecting slightly beyond this margin, and are without proper comitalia, although the slender diacts are present in abundance between them. The dermalia are slender hexacts, which agree with the description given by Tjima (1901, p. 273) for &. phoenix. The short distal ray is cylindrical or only very slightly expanded, 6-8 » thick, rounded at the end and with “ obsolete microtubercles.’ Measurements of a characteristic spicule are: distal ray, 80 » long; tangential rays, 160 » long; proximal ray, 200 » long; tangentials and proximal, about 8 » thick at the base. Meshes of the dermal network formed by the tangential rays are in general square, 160-250 p» on the side, including 3 or 4 pores which are 50-110 p» in diameter, In the upper part of the body, within a distance of 15 mm. from the margin, the dermal hexacts are much less uniform in structure than else- where. While spicules occur, like those which are common lower down, most of the hexacts are larger, stouter forms, in which the short distal ray is either rounded or pointed at the end, and the tangential rays taper conspicuously to points. Many sizes are found, ranging up from spicules, in which the rays are about as long but twice as stout as in the common THE SPONGES. 39 dermalia, to large forms with tangential rays 600 » ong and 65 yp thick. In the larger ones the tangential rays are commonly of unequal lengthis. Some of these stout spicules are found here and there over the general surface of the body. The gastralia are scattered. Pentacts answering to the description given by Ijima (1901, p. 275) for R. phoenix are common. The tangential rays are equal or unequal, and mostly 150 to 250 » long by 10 to 12 u thick. The proximal ray is longer, frequently 500-750 mw by 10-12 p. Similar tauactines and stauractines are common. Hexact forms also are frequent, like the pentacts except that the small rounded boss is repre- sented by a ray, which is usually much shorter than the tangentials, and is smooth, terminally rounded, cylindrical, or slightly enlarged, the greatest thickness sometimes reaching 16 pw. Gastralia similar to those here described are present in a specimen of R. phoenix from the Museum of Comparative Zodlogy. Onychasters are abundant. The principal rays measure 4-6 p, the terminals about 30 w in length. The terminals are slender and tapering, usually 3 to a principal. The spicules differ from those of R. phoenix (Schulze, 1899, p. 21, Plate IIL; Ijima, 1901, p. 216, Plate X.) in the size of the claws. In R. phoenix, according to Schulze’s and Ijima’s figures, the claws are fully 2 » long. In my specimen the claws are 1 p» long and exceedingly fine. The difference, although one of degree, is easily noticed. In a preparation of R. phoenix (specimen in Museum of Com- parative Zodlogy) I can observe the claws with a power of 300, and can study them very well with a power of 600. In my specimen, with the former power, all the spicules look like oxyhexasters, and even with 600 most of them present this appearance. To make sure of their general presence, it is necessary to use an immersion objective. In the spicule, only two claws can be made out on each terminal, and these project forwards. A floricome is found in the immediate neighborhood of the distal ray of each dermal pentact. The spicules are like those of R. phoenix (Schulze, 1887, p. 85, Plate XIII. ; Ijima, 1901, p. 276, Plate X.), but the size is some- thing smaller than that given by Ijima, the diameter being about 80 x. Graphiocomes must be rare. I do not find any in my preparations, and only a very few rhaphides. Nor do I find any spicules peculiar to the border of the parietal apertures. 40 THE SPONGES. Over a part of the surface, as in the specimen of R. phoenix studied by Ijima (1901, p. 269), the hydranths of a commensal hydroid cause minute elevations. The elevations are not abundant nor conspicuous, although the opaque body of the hydroid catches the eye. As in Ijima’s specimen very large and modified dermal hexacts, mingled with the common dermal hexacts, are found round the hydrozoan body. The spicules differ in some details from those found by Ijima (1901, p. 274, Plate X. Figs. 25-27). The distal ray is not club-shaped but cylindrical, beginning to taper near the upper end and running out to a point. The tubercles spread over its distal half. The tangential rays are sometimes short, as in Ijima’s spicules, again almost as long as the distal ray, and they vary in length in the same spicule. The proximal ray is generally shorter than, but sometimes as long as, the distal ray. A fairly character- istic such spicule has the following measurements: Distal ray, 1 mm. x 85 p; proximal ray, 500 » x 50 p; tangential rays, 500-4 x 50 pw to 200 wp x 50 pw. The spicules in a single clump vary in absolute size and in proportions of parts. For instance, in some cases a tangential ray is much the longest of all. Many intermediate sizes between the dimensions above given and the ordinary dermal hexact are to be seen. The preservation of the hydroid itself is very imperfect. But it can be seen that the form is a tubularian hydroid, that the hydranths have several tentacles, and are borne upon a slender branching stolon. The sieve-plate region presents a simpler structure than in &. phoeniz. The sponge ends above in a thickened margin which contains 6 large stauractines arranged in a ring (Fig. 7, Plate 3; Fig. 2, Plate 4). The superior rays of the stauractines project obliquely upward and centripetally, as if to form the radial beams of a sieve-plate. With the exception of one stauractine the superior ray of which is bare (Fig. 7, Plate 3), these rays are densely covered with the smaller parenchymalia, chiefly slender, cylindrical diacts 8-20 w thick, mingled with which are small tauacts, stauracts, pentacts, and hexacts. These latter spicules, like the slender diacts, have smooth cylindrical rays, usually of unequal lengths in the same spicule, rounded or round-pointed at the ends, where they are frequently enlarged and subterminally roughened ; rays, 34-225 p long, 8-20 p thick. The tip of the large stauract ray emerges from its covering. Unlike the adult &. phoenix, no principal diacts accompany the superior rays of the large marginal stauracts. The inferior rays which extend THE SPONGES. 4] longitudinally down into the sponge wall are unaccompanied by comitalia, and with the exception of one case are not in intimate association with particular principalia. In the case alluded to, Fig. 2, Plate 4, the tip of the inferior ray is well overlapped by the two principal diacts of a longitudinal skeletal bundle. There is but one case of anastomosis between the bundles which are supported by the superior rays of the stauracts (Fig. 2, Plate 4), and in only one of these bundles is there any evidence of an existing or beginning transverse connection. This bundle bears a laterally projecting small hexact (Fig. 2, Plate 4). Whether the central ends of the bundles, which are supported by the superior rays of the stauracts, were connected in the living sponge, must remain an open ques- tion. It may be added, however, that the upper margin of the sponge and the surfaces of the bundles are smooth, and there is no indication in the specimen itself that anything has been torn away. The marginal stauracts are 1.5 mm. to 3 mm. apart. The two rays corresponding to the proximal and distal of the dermal hexact are reduced to conical bosses. (This condition, instead of the pentact, is occasional in hk. phoenx according to Ijima, 1901, p. 271.) The superior ray is always long, nearly cylindrical, tapering eventually to a point, smooth or with a surface made undulating by scattered low and rounded tubercles. The three other rays are smooth and taper gradually to a point. They vary greatly in absolute and relative lengths. The inferior ray may be the longest or much the shortest of the four developed rays. The lateral rays, lying parallel to the sieve-plate margin, may be equal or unequal in length, and very much shorter than or nearly as long as the superior ray. In the largest spicules the superior ray is 4.5 mm. to 4.8 mm. long and 250 p to 300 » thick at the base; the other rays having about the same basal thickness. Two of the marginal stauracts, one of which is shown in Fig. 7, Plate 3, are much smaller than the others, the superior ray measuring about 2.4 mm. x 120 w. The marginal stauractines, it will be seen, are small as compared with the larger of the corresponding spicules in fk. phoenix (Ijima, 1901, p. 272, gives the combined length of the superior and inferior rays as reaching 30 mm.), and the rays vary more in relative length than in the spicules examined by ]jima. The thickened margin representing the cuff of some other species is a band about 1 mm. wide. It consists chiefly of closely packed diacts, mostly slender forms 10 to 30 » thick, with some larger ones up to 60 pw thick, 6 ‘ 42 THE SPONGES. arranged in large measure parallel to the free margin of the sponge. Mingled with these are some of the other common, small parenchymalia, intermediate between the diact and hexact conditions. The band projects on the gastral surface, and while larger, is essentially similar to the slender skeletal bands which elsewhere project on this surface. The marginal stauractines lie on the dermal side of the band. In R. phoeniz Schulze describes the margin of the sieve-plate area (1887, p. 84, Plate XIII.) as surrounded by large dermal hexacts, the long spinous distal rays of which project radially from the surface to a distance of about 5 mm. Jjima also finds enlarged dermal hexacts, which “lie crowded on the cuff edge” (1901, p. 275), but the spicules are much smaller, the distal ray measuring 1 to 1.25 mm. in length. Large dermalia of this character are not found in the marginal region of &. delicata, unless the single spicule shown lying on the marginal band in Fig. 7, Plate 3, belongs in this category. In the extreme marginal region of R. delicata many of the dermalia have probably been lost, but groups remain here and there. Among these are hexacts which do not differ from the common forms of dermalia found in the upper part of the sponge. Several such are shown in Fig. 8, Plate 3. In the upper left corner a fairly typical dermal hexact is figured. Lower down lie other hexacts differing considerably in size and detailed character. In some of the larger dermalia, near the margin, the distal rays are reduced to conical bosses, and the proximal rays are but little longer. A group of three such spicules is shown in Fig. 8, Plate 3 (to the right). The tangential rays here, as in the other enlarged dermalia, are subequal or strongly unequal. The condition of these spicules is not far from that of the marginal stauractines, and it seems probable that it is from them that the stauractines are recruited as the sponge increases in size. That is, the condition of the marginal region of this specimen suggests that an enlarged dermal hexact, in which one of the tangential rays is especially elongated, is from time to time pushed to the edge, the tangen- tial ray in question becoming the projecting superior ray of the stauract. THE SPONGES. 43 ASCONEMATIDAE F. E. Schulze. Caulophacus F. E. Schulze. 1887. Caulophacus ¥. K. Schulze, 1887, p. 124. 1897. Caulophacus ¢ 1897, p. 6. 1903. Caulophacus ¥. BK. Sch., Ijima, 1903, pp. 85, 112. Caulophacus schulzei, sp. nov. Plate 4, Figs. 1, 3, 5-10; Plate 5, Figs. 1-6, 8-10. Diagnosis. Body of the usual character, and gray-brown in color. Dermal and gastral pinules are hexacts, which differ only in that the dermalia are slightly shorter and stouter than the gastralia. The usual spinose discohexacts and discohexasters are present. In the smooth discohexaster the terminals are commonly 5-10 in number, arranged in a whorl, and considerably longer than the principals. Station 3382, 10 specimens; Station 3599, 4 specimens. The body (Fig. 3, Plate 4) varies from a disc shape to a distinctly calyculate shape, and has a diameter varying from 22 mm. to 50 mm., with a thickness, taken midway between the attachment of the stalk and the edge, of 3 to 6 mm. The thickness diminishes toward the edge, which is sharp. The dermal surface of the body —that to which the stalk is attached —is in several specimens distinctly convex, in other specimens flat or slightly concave. The opposite, or gastral surface, is in general slightly concave, but in some of the specimens it is slightly convex. In all cases the stalks are broken off near the body. There can be no doubt, however, that five of the six stalks that were in the same jars with the sponge bodies, belong to them. This is demonstrated by the agreement in spiculation, and in diameter and appearance, between the upper end of the detached stalk and the lower end of the fragment that is united with the body. The stalk is more or less curved, slender, the diameter in the middle region ranging from 2 to 3 mm.; about cylindrical, but enlarging above and below. Below, the stalk makes an angle with its narrow, elongated base, the precise shape of which varies, although the surface of attachment is in all cases flattened. In the natural condition the base is evidently attached to the root spicules of Hyalonema, round which it grows. Frag- ments of some of the Hyalonema spicules remain, perforating the base in the direction of its long axis, also the parallel impressions left by others of these spicules on the attaching surface of the base. 44 THE SPONGES. The stalk enlarges at its upper end, where it passes into the body. Its connection with the body is always excentric, and except in two speciinens oblique, as shown in the figure. In the two cases referred to, the rem- nant of the stalk projects vertically from the body. The body itself is heavy, and because of the numerous canals perforating it, 1s easily torn. The stalk is firm and hard, except in its uppermost region. Here, where there are no synapticula between the principal supporting spicules, it is comparatively soft and easily broken, although in the living specimen doubtless flexible. The attachment of the stalk to the Hyalonema root spicules shows that the body itself cannot, in the natural position, be far from the surface layer of mud. And this fact, taken together with the character of the uppermost part of the stalk, suggests that the disc rests upon the surface mud, something after the fashion of a Renilla, instead of projecting freely in the water. Both surfaces of the body exhibit the apertures of very numerous canals, which pass vertically into the interior. They are about equal in abundance and size on the two surfaces, being smallest in the peripheral region. The diameter of the apertures ranges from less than 1 mm. to 2mm., or in some individuals to 4 mm., and on both surfaces they are covered in by the dermal and gastral membranes respectively. The axial canal in the stalk is 3 to ; the diameter of the stalk, widening greatly above where the stalk passes into the body, and opening into several smaller canals. In two specimens examined these canals pursued so intricate a course that I could not trace them to their openings. In another speci- men the canals were larger and opened on the gastral surface, as described by Schulze for Caulophacus latus (1887, p. 124) and C. agassizit (1899, p. 37). As in the other species of the genus, the principal parenchymalia are hexacts and diacts. The hexacts are distributed through the body of the sponge, where they are abundant but not crowded (Fig. 10, Plate 5). In the stalk they are few in number, and those seen lay outside the main diact skeleton. The hexacts are rather slender, with straight or gently curved rays which are often accompanied by a few diacts. The rays are smooth, taper evenly to rounded points, and in general are equal or subequal, measuring 700-1200 » x 28-48 ». Rarely hexacts are found in which all the rays are covered with sharp microtubercles. Such spicules seem to be modifications of the hypodermal pentacts. The diacts of the body are slender, straight, or slightly curved; in ee — THE SPONGES. 45 general cylindrical, or somewhat thicker in the middle region and taper- ing toward the ends, which are enlarged, rounded, and subterminally roughened with microtubercles. The ends may not be swollen and may be smooth. Often, though not always, a trace of the lost rays is retained in the form of a slight annular thickening containing an axial cross, The spicules vary in length from 1 to 4mm., in thickness from 8 to 12 p. Lengths of 1.5 to 2.5 mm. are the commoner sizes. Exceptionally the diact is thicker and perceptibly fusiform, tapering evenly from the middle to the rounded smooth points. A typical spicule of this character measures 1700 » x 24 ». Bundles of diacts and, less commonly, separate diacts run in all directions through the sponge body (Fig. 10, Plate 5). The wall of the stalk is largely occupied by diacts, which run for the most part longitudinally. In the upper part of the stalk these spicules are free. Elsewhere they are connected by abundant synapticula, a con- tinuous framework thus being produced. Scattered diacts protrude radially from the surface of the stalk to a distance of from 1 to5 mm. The diacts as a class are similar to those of the body, but longer and thicker, many reaching a size 7-8 mm. x 24-32». The extremities may be entirely covered with sharp microtubercles, or the tuberculation may be subter- minal. In the lowest part of the stalk some diacts are met with which have smooth, pointed extremities. The dermal and gastral pinules (Figs. 7 and 10, Plate 4) are much alike. They are both hexacts in which the proximal and tangential rays are about equal in length and thickness. These rays are pointed and tapering, with very small, sharp microtubercles near the end, elsewhere smooth or with only a few scattered tubercles; about 100 » x 8-10 ph. In both pinules the distal ray is covered with overlapping upwardly projecting narrow scales, which have a greatest length of 16-20 ». Near the base the scales degenerate into small prickles, projecting at about right angles to the axis of the ray, and at the extreme base the ray is smooth. The ray ends above in a terminal cone, not in a long point. This in the slenderer spicules is commonly longer than wide, but in the stouter ones is as wide as long and is nearly concealed by the uppermost scales, its tip not infrequently being rounded instead of pointed. Except in two specimens the dermal and gastral pinules differ slightly as regards the length and thickness, and consequently the outline, of the distal ray. In the two specimens referred to, measurements failed to 46 THE SPONGES. show a constant difference between the pinuli of the two surfaces. In the other specimens the dermal pinuli have distal rays which are slightly shorter and thicker than those of the gastral pinuli. Measurements show that this relative difference between dermal and gastral pinul exists, although in some specimens both kinds of pinuli are perceptibly stouter than in others. Thus m a number of specimens the gastral pinuli were like the one shown in Fig. 10, Plate 4, where the distal ray is so slightly swollen in the middle as to be almost cylindrical in outline. In the same specimens the dermal pinuli were like the one shown in Fig. 7, Plate 4, where the distal ray is suffi- ciently swollen in the middle for the outline to be distinctly fusiform. In other specimens the gastral pinuli were quite as stout and fusiform as Fig. 7, Plate 4, and the dermal pinuli still somewhat stouter and more fusiform. Thus while the individual sponges differ among themselves, within narrow limits, to be sure, in respect to the precise shape of the distal ray, the relative difference between the two surfaces is usually maintained. This generalization is illustrated by the following tabular statement, show- ing the common range of variation among the spicules of two individuals, the one with pinuli as slender as in any of the specimens, and the other with pinuli as thick as in any of the specimens. Distal Ray of Dermal Pinule, Distal Ray of Gastral Pinule. Length. Greatest Thickness, Length. Greatest Thickness, 1. Sponge with slender pinules, 240-320 p 36-40 p 260-360 p 32-36 p 2. Sponge with stout pinules, 210-240 p 44-56 p 280-320 p 36-40 p On both surfaces the following uncommon types of pinuli make their appearance. In one type, Fig. 8, Plate 4, the distal ray is conspicuously shortened but not very swollen. Much less frequent is the type shown in Fig. 1, Plate 4, in which the distal ray is very short and greatly swollen. The latter spicule is similar to the dermal pinuli of C. datus F. E. Sch. and C. elegans ¥. K. Sch. The general dermal covering of the stalk is in all cases lost, but in one of the specimens some of the pinules on the upper part are preserved. These are smaller than the pinules of the body, the proximal and tangential rays measuring about 80 » x 6-8 yp, the distal ray about 200 p x 24 yp. The covering spines on the distal ray are not so closely set as in the pinules of the body. The hypodermal and hypogastral pentacts (Fig. 9, Plate 4) are alike. All the rays taper to rounded points, and there is no trace of the THE SPONGES. 47 distal ray. The proximal ray is ordinarily longer than the tangentials, but occasionally is very short, especially in the case of pentacts lying over the main canals. The proximal ray is roughened with sharp microtubercles in its upper part. The tangential rays commonly show a few sharp micro- tubercles near the point of intersection, but may be smooth, or on the other hand extensively covered with microtubercles in this region. The tangential rays are straight or very slightly incurved ; exceptionally some- what outcurved. The spicules vary in size in the same individual, the tangential rays measuring 400-750 w x 36-48 py, the proximal ray measur- ing commonly 780-1000 » x 50-60 p. The tangential rays of the pentacts overlap and give rise to a mesh- work, the meshes of which are very commonly square or squarish. The size of the mesh varies considerably in different regions of the same indi- vidual as well as in different individuals; diameter commonly 340-680 y. Where the pentacts are crowded, some lie at a slightly lower level than others, and so interfere with the regularity of the meshwork. On the upper part of the stalk a few of the hypodermal pentacts remain. Some are like those of the body; others differ in that they are quite smooth. The spinose microsclere (Figs. 1, 4, 5, 9, Plate 5) found in all the species of the genus is here present in the greatest abundance, everywhere fillmg the parenchyma. While the true discohexact, in which none of the rays are branched, occurs in all of the specimens, and in a few is the predominant form, it is in most of the specimens uncommon. The rays of the discohexact are 80-110 » x 8 yu, tapering strongly toward the apex, which is capped by a watchglass-shaped end-plate, 10-12 u in diameter, divided marginally into about 6 strong teeth. The rays except near the centre of the spicule bear strong recurving spines, which diminish in size toward the apex of the ray. In most of the specimens the great majority of these spicules are im- perfect hexasters (Figs. 1 and 4, Plate 5). Spicules in which 3 or 4 of the original hexact rays are branched, while the others remain single, are the commonest types, although perfect hexasters, in which all 6 original rays are branched, occur. In the hexasters, imperfect or perfect, the principal rays are smooth and short, and the combined length of princi- pal and terminal equals the length of the undivided hexact ray. The terminals are spinose and capped, as in the true hexact forms. The 48 THE SPONGES. principals may bear 2 or 3 terminals, but 2 is the commoner number. Exceptionally, as in one of the rays of Fig. 9, Plate 5, there is no proper principal, the branching occurring so close to the centrum of the spicule that the terminals are confluent with it. A detail of some interest as bearing on the mode of development of the hexaster form is indicated in Figs. 1 and 4, Plate 5. In spicules where the principal ray bears but two terminals, the latter commonly pass into the principal in an asymmetrical fashion. One of the terminals makes a bend at its lower end, thus becoming strongly convex on this part of its outer surface, while the corresponding surface of the other terminal and the adjoining part of the principal form a weakly concave surface. This fact, tegether with the angles which the several rays make with one another, often suggests that certain rays represent the primitive hexact rays and that other terminals are produced as lateral branches on these. Much less commonly the two terminals are symmetrically disposed on the prin- cipal, suggesting an early dichotomy, but in such cases the symmetry may have been superinduced on an earlier asymmetry. Where the principal bears 3 terminals, the arrangement is usually symmetrical and gives no hint as to whether the branching had been lateral or not. But exceptional spicules, like that shown in Fig. 5, Plate 5, occur which speak for the lateral origin of the branches, in that two terminals occupy a lateral position on the same side of what seems to be a primitive hexact ray. Very frequently, perhaps always, the opposite rays of a diameter branch in planes at right angles to one another, as shown in Fig. 4, Plate 5,— a phenomenon observed by Schulze in the hexasters of several species (1887, pS), A form of discohexaster, Fig. 5, Plate 4, very similar to the corre- sponding spicule of C. datus F. E. Sch. (Schulze, 1887, Plate XXIV.) and C. agassizu F. KE. Sch. (Schulze, 1899, Plate VI.) occurs with about the same distribution as in the latter species (Schulze, 1899, p. 38). It is most abundant near the gastral membrane and in the walls of the large efferent canals, less abundant near the dermal membrane and in the walls of the main afferent canals. In this spicule, the principals are smooth, and taper very slightly toward the apex, where they enlarge to form a base for the terminals. These commonly vary in number from 5 to 10, and are arranged in a whorl. Not infrequently, however, spicules are found with more numerous terminals, up to 16, which do not form a whorl but a brush, THE SPONGES. 49 some being surrounded by others, Fig. 6, Plate 5. The terminals are roughened and very slender, and taper toward the apex, where they bear small end-plates of a watchglass shape. In the larger spicules the end-plate is obviously divided into marginal teeth, and the ray in its distal half is not merely roughened but bears small recurving spines. The size of the spicule varies considerably in the same specimen. The terminals as a rule considerably exceed the principals in length, being from 1.5 to 2.5 the length of the latter. The principal measures 36-50 » x 6 yp, the terminal 60-100 » x 2 p. Occasionally this discohexaster exhibits an abnormality of some in- terest. One or several of the principal rays, in addition to bearing terminal umbels, bear one or in some cases two lateral branches, one above the other, Fig. 2, Plate 5. Such lateral branches resemble the terminal rays. Moreover, examination of the larger discohexasters shows that the base of the umbel is frequently asymmetrical. An extreme case of this kind is shown in Fig. 8, Plate 5. These appearances receive an explanation on the hypothesis that the umbel of terminals represents an aggregation of lateral branches, and that during the growth of the spicule some of the lateral branches may become separated from the main cluster. A good many small discohexasters occur, having a similar distribution to the large form just described. Some of these are doubtless stages in the development of the latter type, although the principal ray is often about equal in length to the terminals, as in Fig. 3, Plate 5. The principal may bear one or two lateral rays. In a selected spicule of this character the principal rays and the terminals are both 40 yw» long; in another such spicule the principal is 26 yp, the terminals 28 » long. Together with these spicules occur discohexasters of a different type, one of which is shown in Fig. 6, Plate 4. The principals and terminals in this spicule are com- monly subequal in length, 16 to 24 » long, but not infrequently the prin- cipal is perceptibly longer than the terminals. The brush-like clusters are relatively wide and include numerous, from 20 to 30, terminals. The principals are smooth, or bear one or two comparatively large tubercles, or sometimes a lateral ray. It is possible that this spicule is of foreign origin. But against this supposition speaks its distribution, as does also the fact that other small discohexasters occur, which are intermediate in structure between the types shown in Fig. 3, Plate 5, and Fig. 6, Plate 4. As an example of such intermediate forms I select a spicule in which the principal 7 50 THE SPONGES. rays are 24 » long; terminal rays 20 » long; terminals 10-15 in a cluster ; clusters intermediate in relative width between Fig. 3, Plate 5, and Fig. 6, Plate 4. Caulophacus, sp. Plate 5, Fig. 7. At Station 3414, the lower part of a stalk belonging to a species of Caulophacus, apparently not C. schulzei, was taken. The stalk (Fig. 7, Plate 5), which is attached to the root spicules of Hyalonema, is firm, hard, and of a dark-brown color. The length of the fragment is 40 mmn., the diameter of the upper broken end 4 mm. The axial cavity is very small, about 0.75 mm. in diameter. The base is an irregular mass elongated in the direction of the Hyalonema spicules, round which it has grown. Some of the Hyalonema spicules remain in situ, while others have been pulled out, leaving their impressions upon the Caulophacus base. The dermal covering has been lost. Whether the few pinules and large pentacts adhering to the surface belong to the specimen is questionable. The diacts forming the chief support are arranged for the most part lon- gitudinally, and are connected by synapticula. Scattered diacts protrude more or less radially from the surface to a distance of from 1 to 4 mm. The diacts taper slightly from the middle toward the ends, frequently exhibit an annular thickening in the middle, which is very slight in the large forms, but conspicuous in some of the smaller, and end in smooth pointed extremities. The diameter of the larger spicules is 24-30 p, th length reaching at any rate 4 mm, At the same station, two other Caulophacus stalks of a somewhat differ- ent appearance were dredged. Only the diact skeleton remains. One of the stalks is remarkable for its thickness, having a diameter at one end of 13 mm., the axial cavity being about 2 mm. wide and filled with mud. THE SPONGES. 51 ROSSELLIDAE F. E, Schulze. Bathydorus F. E. Schulze. 1887. Bathydorus ¥. E. Schulze, 1887, p. 150. 1897. ee F. E. Sch., Schulze, 1897, p. 14. 1898. a F. E. Sch., Tjima, 1898, p. 46. Bathydorus levis F. E. Schulze. 1895. Bathydorus laevis F. E. Schulze, 1895, p. 57, Taf. VI. Figs. 1-10. 1902. Bathydorus levis ES 1902, p. 78, Plate XIV. Figs. 1-10. Bathydorus levis spinosus, subsp. nov. Plate 5, Figs. 11-13; Plate 6, Figs. 1, 2. Diagnosis. Body calyculate. Both dermal and gastral surfaces with scattered pros- talia. Autodermal stauracts densely covered with sharp spines 1-2 » high. Distal ray of autogastral hexact longer, and with longer spines, than the other rays. Oxyhexasters 80-100 » in diameter. Station 5382, 2 specimens; Station 3599, 1 specimen and a fragment. Of the specimens taken at Station 3582, one is a thin, laterally com- pressed sac, Fig. 11, Plate 5, with a greatest horizontal diameter of 46 mm. and a depth of 25 mm. The wall is about 1 mm. thick, thinning toward the edge. The extreme lateral compression is doubtless unnatural. There is a marked concavity on one side of the sac, and the base of the sponge is somewhat pointed, projecting toward the concave side. The dermal sur- face of the basal portion is indented by an oblique furrow about 2 mm. wide and 6 mm. long, probably caused by the cylindrical body (Hyalonema root spicule?) to which the sponge was attached. The second specimen from Station 3382, Fig. 1, Plate 6, is much broken, but fortunately the base is preserved intact. As in the other individual, the sac is pointed below and concave on one side. Nearly the whole of one lateral wall of the sac has been torn off. The opposite lateral wall is about 60 mm. wide and 2 mm. thick, thinning toward the free edge, only a part of which is natural. The sac in the concave region to one side of the pointed base tightly grasps what is probably a Caulophacus stalk. ‘The stalk in question is a - fragment 30 mm. long and about 4 mm. thick. It is roughly cylindrical, hollow, and slightly curved. Only the skeletal framework remains, which consists of diacts, running for the most part longitudinally, and richly connected by synapticula. 52 THE SPONGES. The specimen from Station 3399 has the shape of a wide, shallow cup and is not laterally compressed. The cup, in which both base and edge are preserved, is 53-63 mm. wide and about 20 mm. deep; the wall 3 mm. thick near the centre and thinning out toward the edge. In the centre of the cup the wall is steeper than elsewhere, and thus an inner basin is marked off from a more peripheral region. The peripheral part of the wall flattens out somewhat, tending toward the horizontal plane, and in one region is recurved much as in Schulze’s figure of B. levis (Schulze, 1902, Plate XIV. Fig. 1). The inner surface of the cup, in the peripheral region, is undulating, and the edge likewise undulating, as in Schulze’s figure. Viewed from the under surface, the base of the cup forms a well- marked protuberance, to one side of which the sponge tissue has grown round three Hyalonema root spicules, remnants of which remain half buried in the wall. The fragments of root spicules lie close together, parallel to one another and about parallel to the horizontal axis of the cup. They are doubtless part of a Hyalonema root tuft to which the Bathydorus was attached. The fragment dredged at Station 3399 is a plate-like piece 30 mm x, 25 mm. and about 2 mm. thick, including a part of the natural edge of the sponge, 40 mm. long. In all the specimens both dermal and gastral surfaces exhibit fairly abundant although scattered prostalia, projecting obliquely or radially to a distance of from a few to 10 mm. The spicules are chiefly smooth — diacts, but in part large smooth hexacts with unequal rays, only one ray of which projects. The rays of a single hexact may vary in length from 2 to 10 mm., the protruding ray being long. In addition both surfaces are abundantly covered with the ends of slender diacts projecting 1-2 mm. _ Round the edge, numerous diacts project 1-2 mm., and _ scattered diacts protrude through all distances up to about 5 mm. These scat- tered spicules are pretty far, 5-10 mm., apart. The spicules project at all angles from the edge, and nowhere constitute anything so definite as a fringe. On both surfaces of the sponge, the rounded apertures of small canals are abundant and plainly visible. The diameter of the canals is in general less than, although reaching, 1 mm. The autodermal stauracts, Fig. 13, Plate 5, are abundant, the rays overlapping so as to form a meshwork. The rays are cylindrical, or taper THE SPONGES. 53 very slightly toward the apex, there becoming suddenly rounded or pointed, and are densely covered with short, sharp spines. Basal diameter of the ray excluding the spines, which are 1-2 w high, is about 5 wp. The rays are equal or subequal in length; total diameter of the spicule, 100- 160 ». A direct comparison with preparations of B. /evis shows that in the latter species the stauracts are much less strongly spinose than in the form here described. In the autogastral hexacts, Fig. 2, Plate 6, the rays are straight or slightly curving, the distal ray commonly more distinctly curved than the others. The tangential and proximal rays bear very small, sharp spines. The distal ray, which is longer than the others, bears longer spines, many of which project upwards. The tangential and proximal rays taper evenly to points; the distal ray is cylindrical, then tapering. All rays have a basal diameter of about 4 ». Length of the distal ray, 110-140 »; length of the proximal and tangential rays, which are subequal, 60-90 yu. Schulze (1902, p. 80) mentions that in B. levis the spines on the distal ray of the autogastral hexact are often slightly different from those on the other rays. On the other hand, in the form here described, spicules occa- sionally occur in which the 6 rays are equally long. The oxyhexasters, Fig. 12, Plate 5, are 80-100 ww in diameter. The smooth principals are 4-6 » long. The delicate, roughened terminals, of which there are 2 or 3 to a principal, diverge strongly, are slightly curved, and taper evenly to points. In the hypodermal pentacts all rays are smooth, tapering to points which are not very sharp; no trace of the distal ray. The proximal ray, which passes more than halfway through, often nearly through, the sponge wall, is 1.0-1.7 mm. x 30 »; frequently accompanied by 2 or 3 diacts. The paratangential rays are 340-500 » x 24 mw, overlapping and forming a meshwork, with meshes 340-500 w in diameter. The ends of the diacts are pointed, or rounded and often enlarged; roughened with microtubercles, which may cover the entire end or be restricted to a subterminal area. The slender diacts are commonly pointed and not enlarged at the ends, and are nearly cylindrical. The larger diacts obviously taper from the middle toward the ends. The diacts vary in length from 1 to 15 mm., in thickness from 7 to 60 4. While most of them lie parallel to the sponge surfaces, numerous slender ones and the scattered large prostalia pass obliquely or radially through the wall. As 54 THE SPONGES. in B. lews, the largest diacts are in the neighborhood of the gastral surface, where they frequently form tracts. In addition to the large smooth hexacts already mentioned, one ray of which protrudes as a prostal spicule, smaller hexacts are occasionally found with equal or unequal rays, reaching 700 yw in length, spinose, and some- times curved. It may be questioned whether they belong to the sponge. Schulze’s specimens of the type were taken in the southwestern part of the Bay of Bengal on Globigerina ooze, at a depth of 1997 fathoms. It is very easy for scattered prostalia to be lost in the handling of a sponge, and as Professor Schulze has suggested, perhaps such spicules were originally present in his specimens. Staurocalyptus Ijima. 1887. Rhabdocalyptus F. KB. Schulze, pars, Schulze, 1887, p. 155. 1897. , “ | Schulze, 1897, p. 33. 1897. Staurocalyptus Tjima, 1897, p. 53. 1898. ss Tjima, 1898, p. 52. 1899. Staurocalyptus Tjima, Schulze, 1899, p. 47. Staurocalyptus, sp. Plate 6, Figs. 4-10. At Station 3370, a small sponge was taken, which is completely mace- rated, but in which the shape has been preserved, owing to the fact that the parenchymal diacts are so interwoven with one another. There are no discoverable autodermal or autogastral spicules distinguishable from the parenchymalia, and the probability is that they have been completely lost. The specimen differs from the described species of the genus, but may turn out to be a young form. Owing to the absence of the autoder- malia and autogastralia, the sponge cannot be adequately characterized, and I refrain from giving it a specific name. The body, Fig. 6, Plate 6, is a flattened sac 25 mm. high with a greatest transverse diameter of 15 mm.; wall about 4 mm. thick, gradually thinning out above to an oscular edge. Long prostal oxydiacts protrude from the upper end of the body, forming a collar. These spicules extend longitudinally through the lateral wall of the body, emerging above, at some distance below the oscular margin, which is thus left free of pro- jecting spicules, as in Aphorme horrida F. HE. Sch. (Schulze, 1899, p. 41). far THE SPONGES. Dd The prostalia are the large oxydiacts just referred to and pentacts, which properly are hypodermal, but which may protrude. The oxydiacts measure 2 cm. x 150 p» to 3.5 cm. x 225 y, are slightly curved, smooth, tapering evenly toward each end. The smaller ones run out to very fine points. In the larger ones the extreme ends are broken off. The prostal and hypodermal pentacts have paratropal, paratangen. tial rays, Fig. 5, Plate 6. The shaft is 6 mm. or more in length, 100 » thick at the upper end, tapers evenly to a point, and is very faintly tuberculate, appearing smooth. The paratangential rays measure 2.5 mm. x 75 » to 3.75 mm. x 85 p, and are nearly straight or slightly curved ; tapermg evenly to points; minutely tuberculate. Small forms of the The tubercles on the paratangential rays are fine, closely set prickles, which same spicule occur, with paratangential rays as short as 550 p. in general project toward the apex of the ray, Fig. 7, Plate 6. They diminish in number toward the proximal end of the ray, and here may be nearly or quite absent (Fig. 5, Plate 6). In some spicules the tubercles are so fine that the whole ray appears nearly smooth. The tubercles are outgrowths of the superficial silicious layer, which is fre- quently cracked, and may peel off, leaving the ray smooth. The parenchymal macroscleres are chiefly long slender diacts, more or less curved, extending in all directions through the body; many running parallel to the surface; others more or less radially to the surface and protruding slightly. Length extremely various, 4 mm. to about 1 em.; diameter, 8-24 yw. The two ends of a spicule are unlike: one end sharp-pointed ; the other end blunt-pointed, or rounded without enlargement, or dilated. Both ends are minutely spinose, the shaft smooth. Smaller oxydiacts, about 0.56 mm. x 8 yp, are also common in the paren- chyma; with median enlargement and axial cross; both ends running out to fine points; feebly spinose along the whole length. In the parenchyma the following additional macroscleres occur rarely: hexacts, with rays about 0.5 mm. x 30 yp, smooth, tapering evenly to points; tauactines and stauractines, rays smooth, tapering evenly to points, about 250 w x 8 p. Discoctasters are abundant in the parenchyma. Many conform to the type, having 8 principal rays, each of which bears usually 3 slender There are no discoverable hypogastralia. terminals; terminals minutely enlarged at the end; centrum with a rounded protuberance in the centre of each set of 4 rays. The principal fr 56 THE SPONGES. ray is about 45 » long; terminals about 60 p» long; centrum with a diameter of 20 wp. That the octaster rays are secondarily produced, as Schulze (1893) has demonstrated, by the fusion of components, which themselves are derived from the rays of an original hexact, is well shown both in the typical and “abnormal”’ spicules of this species. The various facts described by Schulze (J. ¢.), such as the trilobed transverse section of the octaster rays, the ridges passing from the central protuberances out upon the rays, and the delicate row of lacunae, extending lengthwise through the latter, may all be observed in the typical spicules. The axial cross is plainly visible in the centrum, as Ijima (1897, p. 44) pointed out. It frequently happens that, as in some of the species described by Schulze, e. g. Rhabdocalyptus mirabilis (Schulze, 1899, Taf. XIII.), one or several accessory rays which end like the terminals are developed as outgrowths from the central protuber- ances (Fig. 8, Plate 6). It also happens not infrequently that the fusion of the originally separate components is very incomplete, in which case the octaster ray appears split to its very base, as in one of the rays of Fig. 10, Plate 6. In some cases there is almost no fusion, as in some of the rays of Fig. 4, Plate 6. Where the fusion is imperfect, and in addition several of the protuberances on the centrum are directly pro- longed into rays, a very irregular spicule is the outcome, in which the octaster character is not conspicuous (Fig. 4, Plate 6). Irregular spic- ules of this kind are referred to by Schulze in several places (1887, p: 1o7; 1893, p. 6): Oxyhexasters are abundant in the parenchyma. ‘The principal rays are smooth, cylindrical, about 3 p long. There are two terminals to each principal, nearly straight, strongly diverging, slender, and tapering evenly to the point, about 35 » long. Oxyhexasters are common in which on one or more of the principal rays only one terminal is present. The remaining terminal and the principal ray may or may not make an angle with each other. In the latter case the division point between principal and terminal is not recognizable. When all six rays are of this character, a small hexact is produced such as has been designated by Ijima (1897, p. 45) a hexactin-shaped oxyhexaster. —-- Microdiscohexasters were carefully looked for, but none were found. THE SPONGES. 57 ‘FARREIDAE F. BE. Schulze. Farrea Bowerbank. 1864. Larrea Bowerbank, 1864, p. 204. 1887. Farrea Bwk., Schulze, 1887, p. 266. 1899. 5 < * 1899, pp- 106-109. Farrea occa (Bowerbank) Carter. 1864. Farrea occa Bowerbank, 1864, p. 204. 1885. Farrea occa Bwk., Carter, 1885, p. 387. 1887. Farrea occa (Bowerbank) Carter, Schulze, 1887, p. 277. 1895. “ 73 “ec cé p- 67. 1899. “cc “e “cc é Dp: 68. Farrea occa claviformis, subsp. nov. Plate 6, Figs. 3, 11-14; Plate 7, Figs. 1-3, 6. Diagnosis. Habitus like that of the type. With oxyhexasters. Characteristic dermal elavulae, with smooth ovoidal heads. Characteristic gastral clavulae umbellate, with few (6-9) teeth. Station 3425. Two fragmentary specimens. The habitus, Fig. 3, Plate 6, is like that of the type, but the projecting ends flare more. The tube diameter is 10-14 mm. The dictyonal frame- work is single-layered, and the radial tuberculate processes vertical, or nearly so, to the framework. The specimens show some vaguely marked elevations, which are prob- ably comparable to the tubular ridges described in this report for Luwrete erectum, and interpreted as having been produced during the growth and division of the cup-like apertures. In spite of the individual variation among the spicules, it may be seen that the dermal pentacts (Fig. 2, Plate 7) are somewhat smoother than the gastral (Fig. 6, Plate 7). Otherwise the two classes are alike. The five rays are of about the same size, and a rudiment of the distal ray ordi- narily persists as a tubercle, which is pointed in some spicules, rounded in others. The precise degree and character of the curvature of the tan- gential rays, and of the tuberculation, and the shape of the ends of the tangential rays, vary slightly. On the outer surface of the tangential rays the tubercles are well developed, elsewhere nearly absent, except at the end of the ray. A common size of ray measures 280 px 12 p. The spicules closely resemble those described by Schulze for the “ Challenger” 8 58 THE SPONGES. specimens of F. occa, differing in some details from the pentacts present in the specimen of F. occa which Schulze had from the Bay of Bengal (1895, p. 67). The uncinates, Fig. 11, Plate 6, commonly show a difference between the two ends. At the external end, as over the middle part of the shaft, the spines are pretty long and nearly parallel to the shaft. Toward the inner (gastral) end they become minute sharp denticulations. The length of the spines varies considerably on different uncinates, and they may degenerate all over the spicule into denticulations. The uncinates may even become smooth, in which case I have found them to be slenderer than the common forms. A common size is about 600 p long by 5 p thick, excluding the spines. But spicules up to twice this size occur. Oxyhexasters are abundant, and similar to those of the type (Schulze, 1887, Plate LX XI. Fig. 7). The principal ray is 20-24 uw long, and bears 3, 4, or 5 terminals, which are about half as long as the principal. Oxyhex- asters occur here and there which differ from the common form in that the terminals are as long, or nearly as long, as the principals. Such spicules are somewhat larger than the common form, and usually there are only two or three terminals to a principal. The dermal clavulae, Fig. 2, Plate 7, with rare exceptions, have smooth ovoidal heads, and are about 300 » long. The stalk is slender and smooth, except near the point, where it is roughened. ‘The spicules occur in the usual position, in groups surrounding the proximal rays of the pentacts. The number in a group is inconstant, always small (3-5); but some spicules may have fallen out. The exceptional dermal clavulae, which are very rare, are like the form common on the gastral surface. In the type (Schulze, 1887, p. 283) the shape of the upper end of the dermal clavula varies from a many-toothed umbel, or a tuberculated swelling, to a smooth club. The first-mentioned shape is the predominant form. The gastral clavulae are arranged, like the dermal, in small groups of 3 to 5, round the proximal rays of the corresponding pentacts (Fig. 6, Plate 7). The common form, Figs. 12, 13, 14, Plate 6, has an umbel with 6-9 teeth, which overarches a smooth swelling. Umbels with as many as 16 teeth occur, and occasionally a spicule is found in which the swelling is minutely tuberculate. In the latter cases observed the umbel had 14-16 teeth. The stalk is like that of the dermal clavula, but is THE SPONGES. 59 shorter, — common length about 240 yw. Spicules are quite frequently found in which the umbel is degenerate, the teeth remaining as minute structures, Fig. 1, a, b, c, Plate 7. Sometimes the merest rudiments of the teeth, so small that they are apt to escape notice, remain on the otherwise smooth head, and very occasionally a clavula with a perfectly smooth head is found. In certain cases all the teeth degenerate except one, which is fairly well developed. A very few gastral clavulae of the type shown in Fig. 3, Plate 7, were found. The umbel teeth are long and with a distinct spiral twist. The stalk is much longer than in the common gastral clavulae, and is slightly curved. In the spicule figured, the stalk bore a lateral spine which was absent in the others found. Only half a dozen of these spicules were found in as many preparations, and they would naturally be looked upon as foreign, were it not that they occupy the same position with respect vo the pentacts as do the other clavulae. This spicule is closely similar to the peculiar clavula of Farrea convolvulus F. E. Sch. (Schulze, 1899, Plate XVI.), the stalk of which bears 3 to 5 lateral spines. Its very occa- sional occurrence here has the greater interest for the existence of a species in which the characteristic spicule is so closely similar. In the type (Schulze, 1887), the predominant form of gastral clavula has a smooth anchor-like head with 4-8 long teeth. But forms having a terminal umbel with numerous teeth, which overarches a swelling, occur and may predominate. Farrea clavigera F. K. Sch. (1887, p. 287, Plate LX XV.) resembles the subspecies here described in having smooth club-shaped dermal clavulae, but differs from it in habitus, in having anchor-like gastral clavulae, and in having a second peculiar form of dermal clavula. Farrea occa is a widely distributed species. J. convolvulus F. E. Sch. was taken 32° 49° N., 117° 27° 30° W., at a depth of 656 m. 60 THE SPONGES. Farrea mexicana, sp. nov. Plate 7, Figs. 4, 5, 7, 8, 10, 11. Diagnosis. Habitus like that of Farrea occa. With oxyhexasters. Characteristic dermal clavulae umbellate, with 12-16 teeth in the umbel. Gastral clavulae in part umbellate like the dermal; in part anchor-like with 4—5 teeth, the stalk usually with 2 lateral spines. Species close to #. aculeata F. E. Sch. Station 8430, 1 fragmentary specimen. The habitus, Fig, 7, Plate 7, is similar to that of Farrea ocea. The tube diameter is 8 mm. The dictyonal framework does not differ from that of F’. occa. In places it is one-layered, in other places two- layered. Small oxyhexacts with rays 60-80 » long are in places fused with the framework. The fragment is obviously* from the older part of a stock. The dermal and gastral pentacts are alike (Fig. 8, Plate 7). The proximal ray is slightly longer than the tangentials, which commonly measure about 280 ~x 12 yp. The rudimentary distal ray forms a small rounded tubercle. In the majority of the spicules, the rays are smooth except at the ends. In others the tangential rays bear very weak tubercles over their whole surface. Occasionally spicules are found, in which the tangential rays on their outer surface bear the well-known strong spinous tubercles. The uncinates commonly exhibit the same difference between the two ends which I have mentioned under Farrea occa claviformis. The spicules vary greatly in size. Forms, 720 » x 5 p, and larger ones up to twice this size, occur. The oxyhexasters resemble those of F’. occa, but are larger; and the terminal rays, of which there are usually 2 or 3 to a principal, are as long or nearly as long as the principal. The length of the latter is about 40 p. There is occasionally a surprising difference in length between the prin- cipals of the same spicule. In Farrea aculeata F. KE. Sch., Schulze (1899, p. 70) says the oxyhex- asters do not essentially differ from those of F. occa. He mentions that the number of terminals to a principal is 3, 2, or 1; and in the spicule figured, the principal ray measures 35 m in length, while the terminals are about as long. Thus in the minute points of difference which the + x THE SPONGES. 61 oxyhexasters of /. mexicana exhibit toward those of F. occa, there seems to be an agreement between it and /’. aculeata. The dermal clavulae are with few exceptions umbellate forms (Fig. 8, Plate 7), having commonly 12 to 16 teeth in the umbel, occasionally less than 12 or more than 16. The stalk including the enlarged upper end is smooth except near the point, where it is roughened. The length of the spicule is about 320 p. On the dermal surface are found a very few of the anchor-like clavulae which are common on the gastral surface. Both kinds of dermal clavulae are in place, in the usual position round a pentact, the number actually present in a group being small, 3-6. The gastral clavulae are in part umbellate, in part anchor-like forms. The former, which perhaps are the more abundant, do not differ from the type which is found on the dermal surface. The anchor-like forms occur in the usual position with respect to the pentacts. Frequently such a spicule is found alone, sometimes together with a few of the umbellate spicules. Where the spicule is alone, it is of course probable that the other members of the group have fallen out. In this spicule (Figs. 4, 5, 10, 11, Plate 7), the stalk is about 500 w long, smooth, practically straight, or slightly or strongly curved, becoming very slender and ending below in a point. It terminates above in a rounded knob, below which there is no conspicuous bulb-like swelling. The knob bears 4 or 5 curved strong teeth, which in some spicules, but not in all, have a slight spiral curvature (Fig. 11, Plate 7). A short distance below the head, the stalk bears usually two curved lateral spines, sometimes only one, and occasionally none. The curvature of the spines themselves varies, sometimes being simple, again feebly spiral, while rarely the spines are straight. While the head ordinarily bears 4 or 5 teeth, spicules are occasionally observed with but 3 teeth (Fig. 5, Plate 7). As regards length of the teeth and width of the umbel (anchor), the anchor-shaped spicule varies considerably. While always larger than the many-toothed umbellate spicule, it is some- times only about twice as wide, and again fully four times as wide as the latter, across the umbel. The actual length of the teeth varied, in the numerous spicules measured, from 24 p» to 48 pw, and the width of the umbel (anchor) from 32 p» to 88 uz. In the obviously closely related species Farrea aculeata F. E. Sch. (Schulze, 1899, p. 70), the dermal and gastral clavulae are alike, and essen- 62 THE SPONGES. tially similar to the anchor-like forms above described. The number of lateral spines is 2-5, and they may be straight or curved. The teeth of the anchor never show a spiral curvature, as they do in some of the spicules of F. mexicana. Similar anchor-like clavulae with lateral spines are de- scribed by Topsent (1901, p. 466) for Farrea weltnert Tops., in which the hexaster is a discohexaster. Farrea aculeata F. KE. Sch. was taken 47° 29’ N., 125° 33’ 30" W., at a depth of 1163 m. Farrea, sp. At Station 3425, 8 specimens, and at Station 3430, 2 specimens, of Furrea were taken. In these only the dictyonal framework is preserved, and a closer identification is therefore impossible. Schulze (1887, p. 278) mentions that in F. occa he has occasionally found the terminal openings of the tubes covered in with porous plates. In one of the specimens here recorded, I find that two of the openings are nearly closed in by such reticula. In each case a small aperture, which is rounded and apparently natural, has been left near the edge of the original opening. EURETIDAE F. E. Schulze. Eurete Semper. 1868. Eurete Semper, 1868. 1887. Hurete (Semper) Carter, Schulze, 1887, p. 289. 1899. Eurete Semper, Schulze, 1899, pp. 106-109. EKurete erectum F. E. Schulze. 1899. Lurete erectum Schulze, 1899, p. 72, Taf. XVII. Figs. i=: At Stations 3358, 5859, and 3380, 19 specimens were taken referable to this species. While they were all partially macerated, the free spicules had been retained. These specimens differ from the type and from one another in details, and fall into three well-marked groups which I designate as subspecies. The specimens from Station 3380 constitute one of these groups. Lach of the other two groups includes specimens from both of the other two stations. A dozen completely macerated specimens, agreeing with the above in habitus and dictyonal framework, were obtained at the above stations and THE SPONGES. 63 at Station 3370. These probably also belong to Lurele evectum, which is obviously an abundant species in these waters. Schulze’s specimens came from the same general locality. They were taken in the neighborhood of the Galapagos Islands, 0° 24’ S., 89° 06’ W., at a depth of 717 m. on a sandy bottom. Eurete erectum tubuliferum, subsp. nov. Plate 7, Figs. 9, 12; Plate 8, Figs. 1-3, 6. Diagnosis. Sponge body differs from that of the type in that the axis is not dichoto- mously prolonged into branches at its upper end. Dermal pinules, gastral hexacts, and gastral pentacts resemble those of the type. With onychasters. Characteristic dermal scopulae are small forms with 3-4 distal rays, which are denticulate and which terminate in small smooth heads. The gastral scopulae resemble those of the type, but the distal rays are not covered with recurving spines, but are either smooth or minutely denticulate. Qa Station 5358, 3 specimens ; Station 3559, 7 specimens. As in the type there is a basal plate from which the slender hollow axis arises, but the plate is not included in all of the specimens. The axis, which at first is only about 7 mm. wide, gradually enlarges, becoming about 15 mm. wide, and soon acquires the spiral curvature characteristic of the species (Fig. 1, Plate 8). In several of the specimens the upper end of the sponge is preserved, and it may be seen that the axis is not dichoto- mously prolonged into branches, as in the type, but remains single. The degree of development of the lateral branches varies considerably. They may appear as simple cups with a slightly or considerably flaring wall. This is especially the case on the lower part of the stem, and all of the branches have this character on some of the fragments, which yet are large. The cup wall is often broken off short. Very commonly, however, the branches are incompletely or completely divided, but not more than once. Various stages in the division are present. The opposite parts of the cup edge may simply project toward one another, as in Schulze’s figures of Farrea (1887, Plate LXXII. Fig. 3). Or the growth may have continued until the opposite lips of the original cup are apposed, as in Fig. 12, Plate 7, which represents in apical view the upper end of the specimen shown in Fig. 1, Plate 8. With continued growth the apposed edges coalesce, and we then have a branch which opens to the exterior by two separate and distinct apertures. The shape of such 6-4 THE SPONGES. branches is tubular with an aperture at each end. The character of the branches is well shown in Fig. 1, Plate 8, and in Fig. 9, Plate 7, which represents a detail of the opposite side of the sponge shown in Fig. 1, Plate 8. A comparison of different branches shows that before the lips of the original cup fuse, they become folded outward so as to present toward each other parallel faces (Fig. 12, Plate 7). Fusion then takes place between these folds along their outer edges, and later, often imperfectly, along their inner edges. In this way hollow ridges or tubular structures open at the opposite ends, and at first communicating with the cavity of the branch, are formed. Such structures, once formed, indicate the line along which the cup lips have concresced (Fig. 9, Plate 7). The fusion between the outer lips of the folds may take place at first in spots, and thus for a time the cavity of the ridge or tube communicates with the exterior through slits or a series of rounded pores, as in the ridge extending between the two concrescing cups of Fig. 12, Plate 7. The sponge at its upper end terminates in an expanded cup, which in the different specimens shows different stages in division. The division of the terminal cup (Fig. 12, Plate 7) takes place in the same way as that of the lateral cups: the concrescing lips become apposed along a straight line and then bend outward. . All along the sponge, extending from the lateral branches on to the axis, are tubular structures or ridges similar to those formed in the division of the cups. One such is shown in Fig. 9, Plate 7, extending at about right angles from the tubular ridge which itself extends between the terminal apertures of the branch. On the lower part of the sponge these structures are often insignificant in size, as if in process of disappearance, but in general they are conspicuous. They extend sometimes from the edge of a flaring cup on to the axis, and again as in the figures they lie at a distance from the cup edges. They are arranged spirally along the axis of the sponge, their long axes coinciding with those of the turns of the sponge. Obviously these structures are relicts left at successive stages of growth by the continually dividing terminal cup. The latest-formed such structure in one specimen is shown in Fig. 12, Plate 7. It here has the character of a two-walled ridge which extends between the two cups in process of division, and about at right angles to the lines along which the lips of these cups are conerescing. The ridge, as I interpret it, marks THE SPONGES. 65 the line along which the lips of the terminal cup have coalesced, in what may be spoken of as the preceding stage of growth. The arrangement of the tubular structures or ridges indicates that when the terminal cup, which is inclined obliquely to the stem (Fig. 1, Plate 8), divides, the lower half remains as a lateral cup, while the upper half, now the terminal cup, grows so as to add to the stem of the sponge. It moreover grows in the direction in which the cup lips last fused, but the next line of such fusion occurs at about a right angle to the last. The several steps in the gradual building up of the whole sponge from a cup-like young stage thus seem to be marked out by these relicts. The structures in question are indicated in Schulze’s figure of this species (Schulze, 1899, Taf. XVII.). This method of continuous separation of lateral cups from the terminal with formation of a seam is probably universal in the family, and in my macerated specimens of Lurete, sp. traces of such seams are found. Also in some of the Fuarrea specimens, ridges are observable which correspond in position to the structures just described for Eurete, but they are vague and of themselves would be incomprehensible. If Hurete and Farrea colonies are ontogenetically developed from cup- like young stages, as general considerations and the special structures above described suggest, then the small cup-like species of Farrea that have been described may be merely stages in the growth of larger complex colonies. Bowerbank (1875, p. 273, Plate XX XIX. Figs. 1, 4, 5) describes and figures several small cup or vase shaped Farvea skeletons from the West Indies. To these he gives the name of F. pocillum. In this subspecies, as in Schulze’s specimens, there is very little anasto- mosis between the lateral branches. ‘The few instances involve a fusion between the oral lips of adjacent and bifurcated branches. Schulze observed that in his specimens the dictyonal framework of the terminal cups for a certain distance from the edge was Furrea-like, con- sisting of but one layer of beams. In my specimens, the dictyonal frame- work of the terminal cups, like that of the lateral cups, includes in general two or more layers. And in parts of the extreme periphery, where the edge seems to be unbroken, I find two layers. I have also, however, found in the terminal cups quite small tracts, passing irregularly into the two-layered condition, in which the framework consisted of but one layer. Again in some of the lateral cups, in the stage of division corresponding 9 66 THE SPONGES. to Schulze’s figure of Farrea occa (1887, Plate LXXII., Fig. 3), I find that the rounded tongue-like outgrowths, which project toward one another, consist in part of but one layer of beams. The free edges of the lateral cups, in these dried specimens, have commonly a conspicuously thickened appearance, due to the great number of dermal pinuli which are here massed together. In my specimens the dictyonal framework shows the usual difference between the gastral and dermal surfaces, and the radial processes on the latter surface are frequently bifid. The dermal pinules (Fig. 8, Plate 8) closely resemble the corre- sponding spicules of the type. Very commonly all six rays are about equal in length, although the proximal ray or more rarely the distal ray may be the longest. The bushy distal ray has a relatively long, bare basal portion which is cylindrical and is always slightly thicker than the corresponding parts of the other rays. The proximal and tangential rays taper evenly toward the pointed ends, where they are roughened, else- where smooth. The measurements of a characteristic spicule are: distal ray, 160 » long with a greatest thickness of 36 yw, and a bare basal part 50 » long and 10 p» thick; tangential and proximal rays, 160 » x 8 p. The rays may only be 100 w long, proximals and tangentials then having a basal thickness of 6-7 p, the distal a basal thickness of 8 ». In some spicules the bushy distal ray, the characteristic appearance of which is given in the figure, may be thin and nearly cylindrical, bearing spines which are considerably shorter and sparser than in the typical form. The gastralia, Figs. 2 and 6, Plate 8, include both pentact and hexact forms. The former in most of the specimens are much the more abun- dant, while in two of the specimens they are scarcely more abundant than the hexacts. The two forms are alike except as regards the distal ray. In the pentacts, the distal ray is represented by a boss which is small and of an irregular, angular shape. In the hexacts the distal ray, which is of varying length, up to 120 ~, may be nearly cylindrical or consider- ably swollen. Characteristic conditions are shown in Figs. 2 and 6, Plate 8.—— The tangential rays, which measure about 250 » x 16- 20 », have large blunt or rounded teeth on the distal surface and sides, while the proximal surface is nearly smooth, having only a very few such teeth. The rays curve in very slightly, are often nearly straight, taper evenly and slightly toward the end, which is blunt or rounded and not THE SPONGES. 67 enlarged. The proximal ray as a rule is shorter than the tangentials, often about 200 ~ long, of about the same thickness as the tangentials, tapering evenly to a point above which it is roughened, elsewhere smooth or with a few scattered minute prickles. The uncinates vary greatly in size, and exhibit the same difference between the two ends which has been described for Fuarrea occa claviformis. They extend radially or obliquely, often through the entire tube wall. In the wall of the cups numerous large uncinates, commonly about 2 mm. long, are found running parallel to the surface and at right angles to the cup edge. The discohexasters are of the onychaster type, and are scanty or only fairly abundant. Principal ray is 4-6 » long; terminal rays, 24-30 pw long. The principal rays are smooth. The terminal rays are slender, taper toward the apex, and are roughened; capped by a minute disc about 3 w in diameter, which is divided into 4 or 5 claw-like teeth. Spicules occur in which the roughening on the terminal ray is represented by exceedingly minute prickles. Other spicules occur in which one or several of the principal rays, or even all, bear but one terminal each. In such spicules, a “knee” usually marks the passage of the principal into the terminal, but this may not be present. The common form of dermal scopula, Fig. 3, Plate 8, has 3 or 4 distal rays, which are cylindrical, curved very slightly, covered with minute sharp denticulations, and which terminate in very small, smooth and rounded enlargements. The shaft at its upper end has a definitely circumscribed enlargement on which the rays rest; tapering thence to the point, above which it is roughened; elsewhere smooth. The shaft is 200-240 » long, and 4 w thick just below upper enlargement; rays, 40 pu mo pe. Larger spicules are present in some abundance, in which the number of distal rays varies from 4 to 10. The rays measure 60 p x 2 uw to 100 w x 3 pw, and terminate in rounded heads which are usually small, about 5 w in diameter, but sometimes large, about 8 w in diameter. The rays are covered with very small sharp denticulations, which enlarge upon the head, sometimes sufficiently to appear as recurving spines. The shaft has a thickness of 6 » and is somewhat larger than, although otherwise like, that of the typical dermal scopula, The gastral scopulae, Fig. 2, Plate 8, have 4-6 slender distal rays, 70-80 p» long, which terminate in spheroidal heads. The heads bear, round 68 THE ‘SPONGES. their equator and over their under surface, recurving spines. The upper surface of the head is smooth, or covered with minute prickles into which the recurving spines gradually pass. The rays usually ascend obliquely, and then diverge rather suddenly, but the precise curvature varies. The shaft is like that of the dermal. scopulae, and about 300 » long by 5p thick below the upper enlargement. The gastral scopulae fall into two classes between which there are transitions. In the one the distal rays are smooth, and gradually enlarge from the base upwards, being 2 p thick below, 4 w thick above, then expanding into a large head, 12 w in diameter, which bears strong recurv- ing spines. In the other form, the distal rays are nearly cylindrical ; roughened with minute denticulations; and provided with heads which are small, 8 » in diameter, and feebly spinose. The first form predomi- nates in the specimens from Station 3359, the second form in those from Station 3358. It will be seen that in this subspecies there are two extreme types of scopulae, the small dermal form (Fig. 3, Plate 8) and the gastral scopula, with smooth distal rays (Fig. 2, Plate 8). The larger scopulae on the dermal surface, and the gastral scopulae with roughened rays, constitute intermediate forms. In the type (Schulze, 1899, p. 75) the dermal and gastral scopulae are alike, and resemble the gastral scopulae of this subspecies, but have spinose rays, Eurete erectum mucronatum, subsp. nov. Plate 8, Fig. 7. Diagnosis. Like Eurete erectum tubuliferum, but with oxyhexasters instead of ony- chasters. Station 3358, 4 specimens; Station 3359, 1 specimen. In this subspecies there are oxyhexasters instead of onychasters. The oxyhexasters are abundant in some specimens, only fairly so in others. The difference is doubtless due to the extent of maceration. The spicules, Fig. 7, Plate 8, vary somewhat in size in the different specimens. In the spicules of one specimen the principal ray is 4-6 p long, the terminals 40-50 ». In another specimen the principal ray is 4-6 p long, terminals 32-40 w. In a third specimen the principal ray is 6-8 p long, termi- : : THE SPONGES. 69 nals 36-48 p. In yet another specimen, along with spicules in which the principal ray is 4-6 w long and the terminals about 40 yp long, are many spicules with relatively long principals. In these the length of the There are 2 or 3 terminals to a principal. The principal is smooth, the termi- principal reaches 8 yw, while the terminals are 28-32 yp long. nals faintly roughened. The terminals diverge considerably, and are nearly straight, or only slightly outcurving, delicate, and tapering to fine points. Oxyhexasters occur in which some, or all, of the principals bear but one terminal each, a slight “knee” usually marking the passage of the principal into the terminal. These sponges resemble in all other respects the specimens described as LKurete erectum tubulferum, and which were taken at the same two stations as the above. While I have separated the two groups of individuals, and designated them as subspecies, it seems to me quite possible that they are merely classes of individuals which differ in respect to a quality of individual variability. Their detailed resemblance in respect to the other structural features suggests that this is the case. Eurete erectum gracile, subsp. nov. Plate 8, Figs. 4, 5, 8, 9; Plate 9, Figs. 1, 3, 5. Diagnosis. Axial part of sponge body forms a very elongated spiral. Distal ray of the dermal pinules very thick. Tangential rays of the gastral pentacts and hexacts beset all over with minute sharp prickles. With onychasters. Characteristic dermal scopulae are small forms, in which the distal rays are minutely denticulate, and taper from the base to the apex, which is smooth and not enlarged. Gastral scopulae are large forms, 600-1500 uw long; distal rays slender and with large heads, or stout and tapering from base to apex, and without heads. Station 5380, 4 specimens. In these specimens the upper and lower ends of the sponge are not present. The axis, Fig. 5, Plate 9, forms a very elongated spiral, which varies but little in diameter, being 8-10 mm. thick over a length of 150 mm. The cups are much broken, but it may be seen that they are mostly undivided, and that they have a flaring wall. There are, however, some cases in which the cups are partially or completely bifurcated, and it may be seen that in the bifurcation of the cups the same peculiar ridges, or tubular structures, are produced that have been described for Lurete erectum tubuliferwn. And extending from the cups onto the axis similar 0 THE SPONGES. structures are found, as in the other subspecies. In the shape and appear- ance of these structures, however, this subspecies differs somewhat from the others, in that the structures are here strongly compressed ridges, which in side view appear triangular (Fig. 5, Plate 9). The longitudinal char- acter of their arrangement, so conspicuous in the figure, is obviously correlated with the very elongated spiral character of the sponge body. The dermal pinuli, Fig. 5, Plate 8, differ from those of the other sub- species, .¢ubuliferum and mucronatum, in respect to the distal ray. This is noticeably thicker, with a more rounded outline, and the lower bare part of the ray is very short. The ray, which may be of the same length as the others, but is often shorter, is commonly about 50 pw thick, and has a bare basal portion about as long as thick, measuring 16 » x 16 p to 20 » x 20 p. The tangential rays are in general pointed, but exceptionally are rounded and enlarged at the ends, and like the other rays are some- what thicker than in the two other subspecies. A characteristic spicule has these measurements: distal ray, 120 » x 50 »; proximal and tangential rays, 150 » x 10 p. Exceptional forms are found in some of which the development of spines on the distal ray is so great as to obliterate the lower bare part; and others of an opposite character, in which the lower bare part of the ray is nearly as long as in the other subspecies. The gastralia include pentact and hexact (Fig. 8, Plate 8) forms, which are alike except in respect to the distal ray. The tangential rays differ noticeably from those of the type and subspecies tubuhferum and mucronatum in that they are beset all over with minute sharp denticulations. This denticulation may be so fine that the ray appears nearly smooth. The proximal ray is also beset throughout its length with similar minute prickles. The tangential rays, which are very slightly incurved, commonly end, as does the proximal, in points, but exceptionally they are rounded and enlarged at the ends. In a characteristic spicule the tangential rays measure 200 » x 12 pw, the proximal ray 240 p x 12 p. In the pentact forms the distal ray is represented by a small rounded or angular boss. In the hexact forms (Fig. 8, Plate 8), the distal ray, which is 100-150 p long, is in general more bushy than in the other subspecies, although it varies to a nearly cylindrical spinose shaft, somewhat thickened at the top. In one of the specimens the hexacts greatly predominate, in the others the pentacts and hexacts are about equally abundant. In one of the latter specimens the spicules vary toward the condition of the type and the ————- - - - —————— a THE SPONGES. 71 other subspecies, in that the tubercles on the tangential rays are strong prickles and are more abundant on the distal surface of the ray. The uncinates do not differ from those of the other subspecies. The hexasters are onychasters, Fig. 4, Plate 8. The claws at the ends of the terminal rays are more distinctly developed than in subspecies tubuliferum, the diameter of the whole disc being about 4 ». The principal ray is commonly 8 p» long, the terminals 28-32 » long. Smaller sizes, with principal ray 6 » and terminals 20 y» long, are present. The characteristic dermal scopulae are small forms (Fig. 9, Plate 8). The distal rays, 4 in number, are roughened with minute denticulations, are practically straight and taper conspicuously from the base to the apex, which is smooth, rounded, and not enlarged. The divergence of the rays varies, being sometimes so slight that the rays look nearly parallel. In different spicules the size of the terminal ray varies from a length of 50 u with a lower diameter of 4—5 » and an upper diameter of 2 p, to a length of 70 » with a lower diameter of 6 » and an upper diameter of 3 wp. The shaft, about 300 w long by 6-8 pw wide above, is smooth or nearly so, straight or slightly curved, and tapers evenly to the point. Above it The spicule varies. The tapering of passes gradually into the rays. the rays may be slight, and a small head may be developed, in which case it is only the slight difference in the shape of the upper end of the shaft which distinguishes the spicule from the dermal scopula of subspecies tubuliferum. Mingled with the dermal scopulae just described, and quite as common, are scopulae 600-700 » long, with 4-6 nearly cylindrical roughened rays 70-100 » long by 3-5 » thick. The rays terminate in small and feebly spinose heads, 6 yw in diameter, or in large and strongly spinose heads up to 12 » in diameter. This type of scopula is similar to the smaller forms found on the gastral surface. The characteristic dermal scopula (Fig. 9, Plate 8), also resembles in shape one of the gastral forms (Fig. 1, Plate 9), although it is very much smaller. The gastral scopulae, Figs. 1, 3, Plate 9, are large forms 600-1500 u long, the commonest sizes ranging between 600 and 800 w. The larger forms frequently penetrate the entire tube wall, even where the latter consists of 5 or 6 layers of beams. The shaft is 8-16 mw thick, pointed, smooth, or with a few denticulations. Above it has not a definitely cir- cumscribed enlargement, but passes gradually into the rays. The rays, yo THE SPONGES. 3 to 6 in number, 100-120 p long, are nearly straight and are covered With regard to the precise with exceedingly small denticulations. shape of the distal rays, the spicules vary between two extremes. At one end of the series are found scopulae (Fig. 3, Plate 9), in which the rays are cylindrical, 4-5 pw thick, passing above into large rounded heads 12 p in diameter, which bear recurving spines. This type is like the corresponding scopula of subspecies tubuliferum, except in its greater size and in the comparative straightness of its terminal rays. The other extreme is represented by a scopula (Fig. 1, Plate 9), usually one of the longest, in which the terminal rays are very stout and taper con- spicuously from below upward. The lower diameter of the ray is about 12 pw, the upper diameter 6 pw, and the ray ends in a smooth and not enlarged, rounded, or conical extremity. Immediately below the extremity the denticulations are enlarged and form short recurving spines. Between these two extreme types of gastral scopulae are found intermediate forms in which the distal rays taper slightly or considerably from the base upward, and end in strongly or very feebly spinose heads which vary from a very small size, 6 pw in diameter, to a large size, 12 w in diameter. The extreme apex of the head may be smooth, or denticulations may be here developed. Eurete, sp. At Stations 3370 and 3350, 4 specimens belonging to urete were obtained. The specimens are completely macerated, the dictyonal framework alone remaining, and thus do not admit of a more precise identification. THE SPONGES, 73 MELITTIONIDAE Zittel. Aphrocallistes Gray. 1858. " Zi * —— Topsent, 1901 2, p. 11, Pl. IL, Fig. 9. Petrosia variabilis crassa, subsp. nov. Plate 17, Figs. 6, 9, 12; Plate 21, Figs. 2, 3. Diagnosis. Form variable, subcylindrical, and branching, or more or less plate-like and partially incrusting. Body stony; interior dense. Surface smooth to the eye. Oscula, 0.7 to 1.0 mm. in diameter; rather numerous and scattered. Pores in the meshes of the dermal skeleton, one to a few in the mesh. Oxea, 510 u x 32 p. Spiculo-fibres of the main skeletal reticulum 300-600 p thick, consisting of many rows of spicules; superficial spicules of the fibre only loosely combined with the body of the fibre; meshes rounded and of a diameter about equal to thickness of the fibres. Dermal reticulum merely the outermost part of main skeleton, and not differing essentially from it. Station 3405, four specimens. Two of the specimens are subcylindrical sponges broken off below (Fig. 9, Plate 17), branching above, the branches rounded off at the free ends. The two specimens are much alike, save that in one the elsewhere solid body is excavated for a length of 35 mm. in its lower part by an axial cavity, which has probably been bored out by the crustacean found therein. In the specimen figured, the length is 60 mm., diameter at the lower end 10 mm. The other two specimens are of a very different habitus. One is an undulating plate which was obviously attached over a part, at any rate, of its smooth lower surface. The plate has a greatest length of 45 mm. and greatest thickness of 5 mm., thinning away toward the edges. The second specimen (Fig. 12, Plate 17) starts from a similar plate-like ex- pansion, with a smooth under-surface. It then becomes incrusting upon the branching cylindrical skeleton of an alcyonarian, creeping over the THE SPONGES. 119 latter in several directions in the shape of narrow elevated bands, which are confluent with one another. The color is a light yellowish-brown. The body is stony and incom- pressible, and the interior is very dense. The surface appears smooth to the eye, but is rough to the touch. With a lens the points of barely pro- jecting radial spicules may be seen, distributed generally over the surface, in some places thickly enough to form a nearly continuous furze, in other places very scantily. Where the surface is uninjured it appears to the eye, or with a lens, perforate with abundant minute rounded apertures 0.3 mm. to 0.5 mm. in diameter. These which at first sight seem to be simple apertures, are the more conspicuous areas of membrane occupying the meshes of the dermal skeletal reticulum. They are more distinct in Fig. 9, Plate 17, than in Fig. 12, Plate 17, simply for the reason that the photograph from which the former figure was made, was taken from the partially dried sponge, while in the other case the photograph was taken from the sponge in alcohol. The pores are rounded, 50-80 » in diameter. They lie in the meshes of the dermal skeleton, one to a few in the mesh. In some of the meshes, the dermal membrane exhibits no pores. In the cylindrical specimens, rather numerous small rounded oscula, 0.7 to 1.0 mm. in diameter, are scattered over the body, showing a partial arrangement in longitudinal rows. In the two more or less plate-like forms, oscula similar to those on the cylindrical specimens are scattered here and there over the surface without regularity of arrangement. Spicules. Oxeas, Fig. 6, Plate 17. Spicule smooth, slightly curved, cylindrical and then tapering at each end to a sharp point. Size, about 510 p x 32 x. Spicules are sometimes found divided at one end into three short diverging branches. Rarely the spicule assumes the shape of a strongyle or style. Skeletal Arrangement. The main skeleton (Fig. 3, Plate 21, a section vertical to the surface) is a reticulum of thick spiculo-fibres which are frequently indistinctly outlined. The fibres are 300-600 p» thick, consisting of many rows of spicules, arranged for the most part about lengthwise in the fibre, and packed together in a fairly close fashion. The spicules are united by a small amount of spongin, which is insufficient, however, to give the fibre a compact character. The superficial spicules of the fibre are only loosely 120 THE SPONGES. combined with the body of the fibre, and there are always some free spicules in the meshes. Individual fibres traceable only for short distances. Reticu- lum undeveloped in spots, such places being occupied by a confused mass of spicules. Meshes more or less rounded; diameter of the meshes about equal to thickness of the fibres. In the cylindrical specimens longitudinal fibres are vaguely discernible in the axial region. In all the specimens fibres directed more or less radially to the surface, with tangential connectives, may be distinguished in the superficial region of the body. ‘The dermal membrane is supported by a reticulum (Fig. 2, Plate 21), which is merely the outermost part of the main skeleton, and does not differ essentially from the latter. It consists of rather poorly defined fibres 400-600 » thick, enclosing rounded meshes, the diameter of which about equals the thickness of the fibres. Many of the meshes are nearly free of spicules. Others are crossed by numerous scattered spicules, and partly by spicule tracts. Although there are scattered over the surface radially projecting spicules (only the points of which emerge), there are no projecting tufts of spicules, such as in many Peérosia species are pro- duced by the continuation of the radial fibres. On the smooth under surface of the two more or less plate-like forms, the spicules of the dermal skeleton are thickly and irregularly scattered, and not so arranged as to form a reticulum. In spots, however, this con- tinuous layer of irregularly strewn spicules is interrupted by areas of dermal membrane free of spicules, and containing, each, one or a few pores. Comparative. In the type specimen of P. variabilis taken by the “ Alert,” near Port Darwin, North Australia (Ridley, 1884, p. 415), the oxeas meas- ured 4004x19p. In the “Challenger” specimen of P. variabilis var. from the Philippine Islands (Ridley and Dendy, 1887, p. 13), the oxeas were 450 » x 22 ». In the “L’Hirondelle” specimens of P. vartabilis, from the Azores (Topsent, 1892, p. 68), the oxeas were 530 w x 33 yp. In the “Belgica” specimens of P. variabilis, from the Antarctic Ocean (Topsent, 1901 a, p. 11), the oxeas were 535p x 23 p. As to the skeletal arrangement of P. variabilis Ridley (1884) says, “ Main skeleton — very loose primary lines of spicules, about three spicules broad, running irregularly towards surface, crossed by secondary tracts of similar character, 2 or 3 spicules broad, at right angles to the primaries and about er THE SPONGES. 121 4 millim. apart. Dermal skeleton — extremely loose tracts of irregularly par- allel spicules, 3 or 4 spicules broad, surrounding roundish or polygonal areas from .18 to .28 millim. in diameter.” I have examined the “Challenger” specimen of P. variabilis var., and add the following note. The body of P. variabilis crassa is much harder and much less cavernous, and the fibres of the main skeleton are much thicker than in the “Challenger” specimen. The surface, in general, of the “ Challenger” sponge is distinctly reticulate to the eye. The obvious character of the reticulum, as compared with the condition in P. variabilis crassa, is due to the relative slenderness of the fibres and large size of the meshes. Over some parts of the surface, however, the reticulum is very indistinctly de- veloped, the fibres being thicker and the meshes smaller than over the general surface. In such places fibres about as thick, and meshes about as small, as in P. variabilis crassa may be found. Petrosia similis Ridley and Dendy. 1887. Petrosia similis Ridley and Dendy, 1887, pp. 9-12, Plates IT. III. Petrosia similis densissima, subsp. nov. Plate 17, Figs. 7, 10; Plate 21, Figs. 4, 5. Diagnosis. Sponge irregularly lobate; lobes short, subcylindrical branches, or merely rounded protuberances. Hard, almost stony. Surface appears finely reticulate and smooth to the eye. Oscula, 2-3 mm. in diameter, at or near the free ends of the lobes, leading into cloaca-like cavities. Pores in meshes of the dermal reticulum, 1 to 5-6 pores in a mesh. Oxea, 220 » x 16y. Skeletal fibres compact, consisting of many rows of spicules with but little spongin; 80-180 » thick. Very few spicules in the meshes of the skeletal reticulum. In the superficial region of the body, main skeletal reticulum regular, consisting of radial fibres with connectives; meshes here 170-250 » in diameter. In the interior, main skeletal reticulum irregular, with meshes 200-500 » in diameter. Dermal reticulum merely the outermost layer of the main skeleton; fibres, 170-250 yw thick; meshes, 170-250 uw in diameter. Station 3405, 1 specimen. The specimen is an irregular lobate mass, attached below to white con- glomerate, upon which for a short distance it spreads out in an incrusting fashion. The mass is 40 mm. high, somewhat flattened; width of the flattened faces about 40 mm. In Fig. 7, Plate 17, one of the flattened surfaces is shown. Some of the lobes are well marked, though short, subcylindrical branches; others, mere rounded protuberances. The trans- verse diameter of the branches, and the thickness of the whole mass in 16 122 THE SPONGES. the short horizontal diameter, are about equal, and 7-8 mm. The sponge is doubtless to be looked on as a ramifying form with subcylindrical branches, the growth in this particular specimen being predominantly, although by no means exclusively, in one vertical plane. The color is brown, passing here and there into terra-cotta, as if that were a remnant of the natural color. The sponge is hard, almost stony ; the canals large enough to givg the interior a lacunose appearance when cut across. To the eye or with a lens the surface appears finely reticulate, the reticulum most evident over the larger canals. ‘To the eye and the touch the surface appears smooth, and in reality over much of it there are no projecting spicules. Nevertheless there are plenty of places where spicules project radially for a short distance in considerable number (Fig. 4, Plate 21). Rounded oscula, 2-3 mm. in diameter, are found at or near the ends of the branches and on the protuberances. They are the apertures of cylindrical cloaca-like cavities, the inner face of which both laterally and at the bottom shows the openings of efferent canals. The cloaca-like cavities are pretty deep, extending 4-8 mm. into the body of the sponge, but are not continuous with one another. The pores are rounded, 60-80 » in diameter, and lie in the meshes of the dermal reticulum; 1 to 5-6 pores in a mesh. Pores are closed in some regions, but even then perceptible as rounded darker spots, the rest of the pore area appearing as lighter-colored trabeculae between the closed pores. This condition of the closed pores is sufficiently distinct to appear in a photograph (x 30). The flagellated chambers (Fig. 10, Plate 17) are somewhat flattened, about 40» x 32 p, and eurypylous. Spreules. Oxea, Fig. 10, Plate 17; smooth, slightly curved, cylindrical, and then tapering at each end to a point. Size, 220 x 16 p. Skeletal Arrangement. Main skeletal reticulum (radial section, Fig. 4, Plate 21), in immediate neighborhood of the surface of the sponge, is regular, with radial and tangential fibres. In the interior, reticulum is irregular, and with larger meshes. Fibres of the main skeleton are in general compact, although in spots they lose their sharp boundaries and fade into one another. In such a spot there is no reticulum, merely a mass of thickly scattered spicules. Fibres, 80-180 p» thick, averaging a somewhat smaller size in the superficial region than in the interior. In the superficial region the connectives are sometimes as thick as the radial THE SPONGES. 123 fibres, more often somewhat thinner. Meshes of the reticulum, rounded at the corners; in superficial region, 170-250 » in diameter ; in interior, 200-500 » in diameter. Spicules of the spiculo-fibres are closely packed in many rows, arranged lengthwise and cemented together by a very small amount of spongin, which does not form a coating over surface of the fibre. Spicules also project irregularly from surface of the fibres. There are almost no free spicules in the meshes. Many radial fibres are prolonged a short distance beyond the surface, for the length or less than the length of a spicule, thus forming very small projecting tufts. In addition a few separate spicules project radially or obliquely, at points between the ends of the radial fibres, 7. e. from the fibres of the dermal reticulum between the nodes. But over much of the surface these minute projections are lacking, surface being quite smooth. The dermal membrane is supported by a reticulum (surface view, Fig. 5, Plate 21) which is merely the outermost layer of the main skeleton. The fibres measure 60-120 yw in thickness; meshes, rounded-polygonal and 170-250 » in diameter. As in the main skeleton, the fibres in general are compact and sharply outlined. Here and there spicules project from the fibres well into the meshes, or cross them, but in the meshes in general there are almost no free spicules. Usually in the nodes the crossing of spicule tracts is discernible. Comparative. I have examined the type specimens of Petrosia similis Ridley and Dendy, and find that P. similis densissima stands closest to var. compacta (Ridley and Dendy, 1887, p. 12). The skeletal fibres of this variety are not nearly so compact as in P. similis densissima. This statement applies both to the surface reticulum and to the main skeleton as well. Ridley and Dendy (1887) correctly say that “the skeleton fibre is by no means so com- pact and well developed” as in P. dwra (Nardo), whereas in P. similis den- sissima the fibres are fully as compact as in P. dura (Specimen No, 1818 in Berlin Museum f. Naturkunde, from Rovigno), and appear more so because there are fewer spicules scattered in the skeletal meshes than in the Medi- terranean species. Owing to the comparatively indistinct outlines of the fibres in P. similis var. compacta, preparations of the dermal membrane and radial sections present a marked difference to corresponding preparations of P. similis densissima. Especially the internal skeletal reticulum, as seen in radial section, is confused and indistinct as compared with subsp. densissima. 124 THE SPONGES. Previously known distribution of P. similis. Petrosia similis: South of Cape of Good Hope (Lat. 35° 4’ S., Long. 18° 87’ E.) at depth of 150 fath.; be- tween Kerguelen and Heard Island (Lat. 52° 4’ S., Long. 71° 22’ E.) at 150 fath. BP. similis var. massa: between Strait of Magellan and Falkland Islands (Lat. 51° 35’ &., Long. 65° 39’ W.) at 70 fath. P. similis var. com- pacta: Philippine Islands (Lat. 11° 37’ N., Long. 123° 31’ E.) at 18 fath. Pachychalina O. Schmidt. 1868. Pachychalina O. Schmidt, 1868, p. 8. 1887. Pachychalina QO, Schm., Vosmaer, 1887, p. 342. 1887. Y «© Ridley and Dendy, 1887, p. 19. 1890. se es Dendy, 1890, p. 353. 1894. 93 ss Dendy, 1894, p. 240. 1902. a «* — Lundbeck, 1902, p. 5. Pachychalina acapulcensis, sp. nov. Plate 16, Figs. 7, 8; Plate 17, Figs. 1-5, 13. Diagnosis. Sponge body an erect lamella, not simple, but a complex composite of erect lobes, many of which are flattened. Lobes intimately connected below, becoming more free and projecting above. Conuli, 3-6 mm. high on the upper portions and projecting edges of the lobes, nearly absent elsewhere. Dermal membrane finely reticulate to the eye. Color yellowish-gray. Sponge very compressible and flexible, yet firm and elastic. Oscula, 2-4 mm. in diameter, over the upper ends and projecting edges of the lobes. —— Oxea very commonly 85-90 » x 3-4 p, although larger (100 » x 5 ») and smaller (60-85 uw x 2 ») forms are abundant. Skeletal bundles 0.5 to 1.0 mm. thick, formed of closely interlacing spiculo-fibres, ascend more or less vertically through the sponge body and give off oblique branches which terminate as axial bundles in the conuli. Skeletal network extending between the vertical bundles, on the whole irregular, although fibres directed more or less radially to the surface are everywhere distinguishable. Radial fibres in parts of the body, invariably in the conuli, extensively developed; in such places, forming with the approximately transverse connectives a fairly regular skeleton. Stronger fibres of reticulum, 50-80 »; connectives, 15-30 p» thick; ultimate meshes often about 300 » wide. Larger fibres well filled with spicules; spongin nevertheless forming a distinct sheath round the fibre. Spongin relatively more abundant in the connectives; spicules here forming from 1 to about 6 rows. Dermal reticulum composed of fibres 40-60 » thick, forming meshes subdivided by fibres 15-30 p» thick; ultimate meshes, 150-350 p» in diameter; fibres like those of main skeleton. Abundant villi commonly about 120 » high, made up of spicules and spongin, project from dermal reticulum. Station recorded as “ Acapulco,” one specimen. The sponge body is essentially an erect lamella, which is, however, curved so that the two ends of the lamella, shown at the right in Fig. 8, Plate 16, are brought close together. Possibly the entire lamella in the natural state encircled some slender upright object. The lamella is by no THE SPONGES. 125 means simple, but may be regarded as composed of numerous erect lobes, many of which are flattened, while others approach the cylindrical shape. The lobes are intimately connected below, becoming more free and _ pro- jecting above. Many of them appear, moreover, as buttresses projecting from and only partially free from the faces, both inner and outer, of the general lamella. The result of this complex order of growth is that the underlying lamellate character of the body is made less distinct, the sponge appearing at first sight as a fruticose mass. The whole mass in the natural position is 110 mm. wide, with a greatest height of 170 mm.; the thickness of the constituent lobes varying, but in the neighborhood of 12 mm. Numerous long conuli, covered with villi which are minute but distinct to the eye, are present on the upper portions and projecting edges of the lobes, nearly absent elsewhere. They are 3-6 mm. high, tapering to a point, slenderly conical, or somewhat flattened and spatula-like. The dermal membrane appears to the eye as a fine reticulum, which only indistinctly allows the arrangement of the internal cavities and paren- chyma to: be seen. When cut across, the interior appears porous, with very numerous small canals mostly 0.5 mm. or slightly over in diameter. The color isa yellowish-gray. The sponge is very compressible and flexible, yet firm and elastic. Oscula are scattered in some abundance over the upper ends and pro- jecting edges of the lobes. They are rounded, 2-4 mm. in diameter, and for the most part lead very quickly each into several efferent canals. Pores rounded, 85-250 »# in diameter, in the meshes of the dermal reticulum ; mostly one or two in each mesh. Interior of sponge is macerated, but in places the size of the flagellated chambers can be made out; chambers measuring 24 x 20 p. : Spicules. Oxea, smooth, slightly curved, cylindrical, not suddenly pointed but tapering gradually at the ends (Figs. 3, 4, 5, Plate 17). The vari- ability in the size of the spicule is considerable. Much the commonest size is 85-90 » x 3-4. Longer and stouter spicules up to 100 u@ x54 are, however, not uncommon. Smaller slender forms 60-85 p x 2 mw are found in the fibres and also scattered sparsely in the meshes of the skeletal reticulum, perhaps representing stages in the development of the larger spicules. Often the spicules in a particular fibre are of nearly the same size, thus in one fibre measuring mostly 85 » x 4 uw, in another fibre mostly 85 px 2 p. 126 THE SPONGES. Skeletal Arrangement. Coarse columns or bundles of spiculo-fibres (/. d. in Fig. 7, Plate 16, and Fig. 2, Plate 17) extend more, or less vertically through the body, branching as they go. The branches extend obliquely upward and outward, and terminate as axial bundles in the conuli (Fig. 13, Plate 17). The bundles are well seen in a piece of sponge that has been macerated for some hours in cold caustic potash. The skeletal network, connecting the columns together and extending between them and the surface, is on the whole irregular. But more or less radially directed fibres, extending out from the branching columns to the surface, may everywhere be distinguished. In some parts of the body (Fig. 2, Plate 17) and invariably in the conuli (Fig. 13, Plate 17) they are conspicuously developed, with connectives commonly about at right angles; the skeleton here becoming fairly regular. Hlsewhere the radial fibres are sparsely developed (Fig. 7, Plate 16), and are accompanied by or pass into other strong fibres, which pursue a very oblique or tangential course. The connectives in such regions are without order, and the network is irregular. The vertical columns have not clearly defined boundaries, but their thickness is in the neighborhood of 0.5 to 1.0 mm. They are composed of coarse spiculo-fibres, 70-110 , thick, together with finer fibres, all interlacing to form a close irregular network (J. 6. in Fig. 7, Plate 16, Fig. 2, Plate 17). Here and there in the constituent fibres of the column, the spicules are less compactly arranged than elsewhere, and the fibres themselves merge into one another. In such spots the reticular nature of the columns is only vaguely apparent. The stronger fibres of the general skeletal reticulum of the body, whether radial, oblique, or tangential, are 50-80 » thick. The finer fibres or connectives measure 15-30 ,» in thickness. In the conuli the radial fibres are slenderer than in the body, about 30-40 p» thick. The ultimate meshes of the skeletal network, as may be seen from the figures, vary a good deal in diameter ; a common width being in the neighborhood of 300 uw. The strong fibres of the general network (Fig. 3, Plate 17) and of the vertical columns are well filled with spicules; spongin, however, abundant and forming a distinct sheath for the fibre. In the connectives (Figs. 4, 5, Plate 17) the spongin is relatively more abundant and the spicules are not as closely packed as in the larger fibres, sometimes forming only an axial core. In the finest connectives, the spicules are —— THE SPONGES. 127 arranged uniserially; in the coarser, in several series up to about 6. There are a few free spicules, all of the smallest size given above, scattered in the meshes of the skeletal reticulum. The supporting reticulum of the dermal membrane (Fig. 1, Plate 17) consists of stouter and finer fibres, like those of the main skeleton. Stouter fibres, 40-60 yw thick, form coarse meshes which are subdivided by finer fibres 15-30 w thick; ultimate meshes squarish or polygonal with diameter 150 » to 350 w. In the larger fibres the spicules are arranged polyserially (about 10 rows); in the finer fibres uniserially or in two to a few rows. From coarse and fine fibres alike project abundant villi, many of them about 120 yw high and consisting of a bunch of spicules, 4 or 5 spicules thick, with considerable horny matter; others consisting of only 2 or 3 spicules. Some of the villi are prolongations of radial skeletal fibres; others are independent projections from the dermal reticulum. Comparative. The species above described resembles in the lamellate character of its growth, and in some other respects as well, Pachychalina spuulamella Dendy (Dendy, 1889), a type specimen of which I have ex- amined. The lamellate character, which is disguised in P. acapulcensis, is pronounced in Dendy’s. The spicules in P. spinilamella are exceedingly slender, measuring about 0.126 by 0.0017 mm. Oceanapia Norman. 1869. Oceanapia Norman, Rep. Brit. Ass., 1868 (1869), pp. 334-35 (generic diagnosis here given is quoted in Bowerbank, 1882, p. 171). 1870. Rhizochalina O. Schmidt, 1870, p. 35. 1882, Phloeodictyina Carter, 1882, p. 117. 1884. Rhizochalina Schmidt, Ridley, 1884. 1887. Rhizochalina Schmidt + Oceanapia Norman, Ridley & Dendy, 1887, pp. 32, 36. 1894. Oceanapia Norman, Dendy, 1894, p. 248. 1894. Rhizochalina Schmidt + Oceanapia Norman, Topsent, 1894 a, p. 10. 1902. Phloeodictyon Carter + Rhizochalina Schm. + Oceanapia Norman, Lundbeck, 1902, pp. 55-56. Ridley (1884) merged Phieodictyon Carter in Rhuzochalina Schmidt. Ridley and Dendy (1887, p. 32) suggest that Rhizochalina (+ Phloeodictyon) and Oceanapia Norman should be united, and Dendy (1894) combines the two under Oceanapia, while Topsent (1894 a) retains the separate genera. Lundbeck (1902) thinks that the group Phloeodictyina (= Oceanapia, sensu Dendy) includes three separate genera, Rhizochalina Schm., Phloeodictyon 128 THE SPONGES. Carter, Oceanapia Norman, and that the group, moreover, is heterogeneous. He proposes therefore to give up the group, and to assign Lhizochalina to the Chalininae, Phlocodictyon to the Renierinae, Oceanapia to the Gelliinae. In this report I conceive the genus in the sense of Dendy. Oceanapia bacillifera, sp. nov. Plate 17, Fig. 8; Plate 18, Figs. 2-4. Diagnosis. Only the fistulae known. These are yellowish-brown tubes, 50-80 mm. long and 8-12 mm. in diameter, with unobstructed cavity. Wall of fistula dense antl firm, 0.5 to 2.6 mm. thick, with smooth outer surface and nearly smooth inner surface. Spicule, a smooth, cylindrical, distinctly curved strongyle, 360-380 p x 24 yp. Wall of the fistula almost entirely filled with a dense skeleton, the greater part of which forms a vague reticulum, consisting of wide, loose spicular tracts, which bound small, rounded meshes. Spicules united by considerable spongin, and arranged tangentially to the sur- face of the sponge. In the innermost layer of the wall a few long spicular tracts occur. These give rise to a reticulum with long, narrow meshes,— meshes elongated in the direction of the long axis of the fistula. At the outer surface skeleton not reticulate, spicules here lying side by side, in any particular region parallel to one another. Station 3404, two specimens. Both specimens are fragments, including only the fistulae. These are yellowish-brown tubes (Fig. 2, Plate 18), somewhat curved, and show- ing here and there low irregular protuberances, or ridges. The tubes are open at both ends, the larger measuring 80 mm. in length, with a trans- verse diameter of about 12 mm., the smaller 50 mm. in length, with a transverse diameter of about 8 mm. The wall is very firm and dense. Throughout the greater part of the larger tube it is extremely thin, 0.5 mm. thick, although in spots, especially near one end, it attains a thickness of 2.5 mm. In the smaller specimen the wall is thicker, the thickness ranging from 0.75 mm. to 2.0 mm. In both specimens the cavity of the tube is unobstructed, the outer surface quite smooth, the inner surface somewhat less so, and showing closely set whitish lines which course longitudinally, and, anastomosing, form a reticulum with narrow, elongate meshes. These lines represent the innermost layer of the skeletal reticulum. No dermal membrane is present, the superficial layer of spicules being quite bare (surface view, Fig. 8, Plate 17), except in spots, where they are covered by exceedingly thin patches of an incrusting species of eo oe —— THE SPONGES. 129 Hymeraphia Carter (the spicules of which are shown in Fig. 3, Plate 18), differing in some few details from LHymeraphia minima Topsent (1892, p- 114, Plate 11, Figs. 2-3). Spicules. Strongyle (Fig. 8, Plate 17), smooth, cylindrical, distinctly curved, ends evenly rounded off. Size, 360-380 p x 24 yp. Skeletal Arrangement. The wall of the fistula is supported by a dense skeleton, which appears as a confused mass of strongyles (Fig. 4, Plate 18, a tangential section), arranged for the most part tangentially to the surface of the sponge, but lying in all (tangential) directions, united by considerable spongin, and interrupted here and there by rounded gaps. This mass of spicules may be regarded as forming a reticulum, which consists of vaguely outlined fibres, or tracts bounding small rounded meshes. The tracts are often 150-200 pw thick, the meshes somewhat less in diameter than the thickness of the tracts. Individual tracts are trace- able only for short distances. Spicules of a tract arranged loosely and about lengthwise; united together with considerable spongin. The dense skeleton occupies most of the thickness of the wall, extending nearly to the inner surface of the fistula. The innermost layer — about 100 » thick — of the fistular wall contains a good many canals and granular cells, and comparatively little skeleton. What skeleton there is has the shape of vaguely defined tracts about 200 yw thick, which resemble those of the skeleton in general, but most of which pursue an approximately longitudinal course for considerable distances. The longitudinal tracts, together with similar connectives, give rise to a reticulum, the meshes of which are long and narrow. The dense skeleton extends quite to the outer surface of the fistula, but in its outermost layer the reticular character is lost, the spicules here lying side by side, in any particular region parallel to one another (Fig. 8, Plate 17, surface view). In the region occupied by the dense skeletal reticulum, only the spicules and spongin are discernible in the present state of the sponge, and although maceration has undoubtedly gone on, there cannet in the natural condition be much soft tissue in this part of the fistula. Comparative. As regards the shape of the spicules the fistulae here described most closely approach Oceanapia singaporensis Carter. In the type of this species (Phlocodictyon singaporense Carter, 1883, p. 326) the spicules are of two forms: oxeas and strongyles, the latter about $ the length of 17 130 THE SPONGES. the former. The oxeas are chiefly confined to the spiculo-fibre of the interior, the strongyles to the surface layer. Habitat, Singapore. Under the head of Rhizochalina singaporensis Carter var., Ridley (1884) describes specimens “in which a large proportion of the (usually acerate) spicules have both ends more or less rounded.” “The largest adult spicules have nearly the same size as the acerates of R, fistulosa, viz., .38 x .0127 mm., but they vary immensely in length.” Habitat, Prince of Wales Channel, West and Alert Islands, Torres Straits, 7 fathoms. Ridley and Dendy (1887, p. 34) record under Rhizochalina singaporensis Carter a fistula taken by the “Challenger” (locality uncertain). I have examined this specimen, and in my preparations the spicules were nearly all strongyles. In the skeletal arrangement and color this specimen differs from O. bacillifera. The sponge identified by Lindgren (1898, p. 297, Taf. 19, Fig. 11 a-b) as Rhizochalina singaporensis Carter must, from the present standpoint of classification, be placed in another genus, since it has chelae. In the color and general appearance the fistulae, here described as O. bacillifera, are very similar to fistulae taken off Bahia by the ‘“ Chal- lenger,” and referred, with a query, by Ridley and Dendy (1887, p. 34) to Rhizochalina putridosa (? Lamarck), but in these specimens the spicules are oxeas, and the reticulate character of the skeleton is strongly marked. Gellius Gray. 1867. Gellius Gray, 1867, p. 538. 1887. Gellius Gray pars, Vosmaer, 1887, p. 349. 1887. Gellius Gray, Ridley & Dendy, 1887, p. 37. 1894. af “ Topseut, 1894 a, p. 8. 1894, S55 “Dendy, p. 247. 1902. “« “« Lundbeck, p. 62. Gellius perforatus, sp. nov. Plate 17, Fig. 11; Plate 18, Fig. 1; Plate 21, Fig. 6; Plate 22, Fig. 1. Diagnosis. Body may appear as a flattened plate-like mass, perforated by spaces which pass through from one surface to the other ; or as an amorphous mass excavated by spaces which pass through the body in several planes, and divide it into a number of anastomos- ing lobes. Color ashy gray. Sponge firm and of cartilaginous consistency. Upper and lower surfaces differentiated. Upper surface roughened with closely set minute projections which reach 1 mm. in height, and consist of, or are supported by, tufts of spicules projecting from the dermal reticulum. Under surface as compared with the d E ‘ 4 F ; J THE SPONGES. wt upper is smooth. —— Oscula small and scattered over upper and outer surfaces. Pores abundant in meshes of dermal reticulum of upper surface; scattered over lower surface, in places abundant. Spicules : Oxea, 320 » x 20 w, with smaller sizes. Sigmatai8 » long, abundant. Main skeleton a confused irregular reticulum of spiculo-fibres, with abundant free spicules scattered between the fibres. Fibres consist chiefly of spicules, with only a very small amount of spongin, and in general are not sharply separated from the scattered spicules. On the upper surface is a dermal reticulum of spiculo-fibres from which single spicules and tufts of spicules everywhere project. On the lower surface a dermal reticulum is developed in places, the membrane elsewhere containing only scattered spicules. Station given as Panama, 4 specimens. The largest specimen, Fig. 1, Plate 18, is an irregular plate, the under surface of which has apparently been moulded over several rounded objects. The plate is 5 to 10 mm. thick, with a greatest width of 95 mm. Perforating spaces 4 or 5 mm. in diameter pass through the body, from the upper to the lower surface. The upper surface, which is the one figured, is very uneven, and from it project numerous lobes having the shape of low rounded or irregular elevations, often with a subterminal osculum. It is apparently the case that such elevations are primarily simple and independent, but in some instances in the course of lateral expansion they meet and fuse with one another, thus roofing in tunnel- like spaces which come to lie between the body of the plate and the fused lobes. On the under surface, too, especially near the periphery, some similar tunnel-like spaces have been formed, apparently by the fusion of lobes growing out from this surface. Of the other specimens, two are fragments of similar plate-like masses, and may indeed have been broken off from the larger piece just described. In one of them the perforating spaces, passing through the body from upper to lower surface, are large, reaching 15 mm. in diameter. And the lobes projecting from the under surface, and fusing in the manner described, give the mass a thickness of 30 mm. The remaining specimen has a different shape. It is an amorphous mass about 40 mm. in diameter, and consists of a few irregular but in the main subcylindrical lobes, 5 to 15 mm. thick, anastomosing in several planes and thus enclosing spaces which continue to pass quite through the body of the sponge, and have a diameter themselves of 5 to 15 mm. The upper and lower surfaces of the whole mass are readily distinguish- able, resembling the corresponding surfaces of the flattened plate-like specimens. 132 THE SPONGES. On comparing the different specimens, and making some use of hypoth- esis, it becomes possible to describe the habitus of the species in the following way. The sponge has differentiated upper and lower surfaces. Primarily solid, in the course of growth it develops lobes, which may spread and fuse, not only in the horizontal plane, but in planes above and below the level of the original body. Thus spaces are partially enclosed which continue to pass through the body in various planes. The enclosed spaces may be small or large, and thus the whole mass comparatively compact or very cavernous. If the growth be predominantly in the horizontal plane, a flattened plate-like body results. When the growth is not predominantly in one plane, a labyrinthine mass of anastomosing lobes results. Except in shape, the several specimens agree. The color is a light ashy gray. The sponge is firm, only very slightly compressible, and of a marked cartilaginous consistency. The upper surface is in general roughened with closely set minute projections which vary a good deal in character. In regions (part of Fig. 1, Plate 18), the projections are fittingly designated as villi, and consist of small tufts of more or less radially disposed spicules, a tuft including only 2 or 3 spicules. Round the base of such tufts the sponge tissue forms an elevation, and thus the projection is strictly conical, although very slender. Such villi are exceedingly abundant over parts of the The villi intergrade with the larger elevations with which the greater part of the surface, especially in the peripheral region of the sponge. surface is thickly covered, and which reach 1 mm. in height. These elevations, which may be referred to as conuli, taper toward the apex. In their distal portions at any rate, they are supported by the larger dermal tufts of spicules, and (hence) frequently appear branched. In places the conuli are especially low and feebly developed, such regions being almost smooth. The under surface, as compared with the upper, appears in general smooth to the eye and the touch, noticeable villi or conuli appearing only here and there. Fairly numerous round oscula, 1.5 to 2 mm. in diameter, are scattered over the upper and outer surfaces. They lead into deep main canals, and are especially developed on the prominences. To the eye the dermal membrane of both surfaces appears imperforate and opaque. In the - k f : THE SPONGES. 153 meshes of the dermal reticulum of the upper surface numerous afferent canals the larger of which measure 80-120 yw in diameter, and which in many places are separated only by thin partitions 10-50 yw thick, abut against the dermal membrane. The pores on this surface, which in some places are closed, but in many regions are open, measure 20-40 w in diameter, and lie thickly crowded in the meshes of the dermal reticulum. Those overlying the more conspicuous afferent canals seem with a low magnification to form rounded pore areas, which have about the same diameter as the canals. The dermal membrane of the under surface is in places riddled with pores, but elsewhere pores appear only here and there, probably owing to closure. Spicules. Megascleres. 1. Oxea (Fig. 11, Plate 17), smooth, slightly curved; points usually sharp, occasionally rounded off, one end rarely strongylate. Spicule abundant. Common size is about 320m x 264, although smaller sizes occur grading down to spicules only 150 p x 2-3 p. The latter are probably young stages in the development of the character- istic oxea, Microscleres. 2. Sigmata (Fig. 11, Plate 17), 18 w long by something less than 2 p thick, are abundantly scattered through the parenchyma and in the dermal membrane of both surfaces. Skeletal Arrangement. In the main skeleton the oxeas are arranged in spiculo-fibres, and are also scattered without order between the fibres. The spiculo-fibres, which are often very loose and best described as tracts, form a confused irregular reticulum, the meshes of which differ greatly in size (Fig. 1, Plate 22, a section vertical to surface). The fibres are of varying thickness, from 200 pu, representing about 10 rows of spicules (as seen in optical section), to 80 p, or even thinner. They consist of spicules arranged for the most part lengthwise in the fibre, and held together by a very small amount of spongin. In the body of the fibre the spicules are pretty densely packed, becoming loosely arranged at the surface. Thus the fibres in general are not sharply separated from the scattered spicules. The fibres may be fairly compact, and the spicules which are scattered in the meshes rather few in number. In such spots the reticular nature of the skeleton is obvious. In other spots the fibres are looser and the scattered spicules more abundant, and the reticular nature is obscured. In still other spots the reticular nature is practically lost, there being in 134 THE SPONGES. such places only a confused mass of spicules. In the superficial region of the sponge, fibres may be distinguished which extend radially to the surface. The dermal membrane of the upper surface is supported by a reticulum of tangential spiculo-fibres, 70-175 mw thick, which produce ridges on the surface. Meshes irregularly polygonal or rounded, and varying greatly in size, from 1 mm. to 200 » in diameter (Fig. 6, Plate 21, surface view). The meshes contain, as a rule, no or only a few free spicules. In the larger meshes some free spicules are generally present, and they are frequently combined to form slender tracts, 1 to about 3 spicules thick, which more or less perfectly subdivide the mesh (as in the centre of Fig. 6, Plate 21). The fibres resemble the more compact fibres of the main skeleton. While the dermal reticulum on this surface is in general well developed, there are small areas here and there in which the fibres merge into one another, thus obscuring or obliterating the reticular character. From the dermal reticulum of the upper surface, single spicules and tufts of spicules everywhere project freely. The latter range from very sinall tufts including only 2 or 3 spicules to tufts formed by the prolonga- tion of radial fibres of the main skeleton (Fig. 1, Plate 22), or by the oblique prolongation of dermal fibres (Fig. 6, Plate 21), and which at the base have about the thickness of the fibre. The larger tufts are abundant, measure 350-500 » in length, and commonly split distally into branches, each branch including from 1 to 2 or 3 spicules. On the under surface a dermal reticulum similar to that of the upper surface is developed in places. Elsewhere such a reticulum is absent, the membrane containing only scattered spicules lying tangentially and crossing at all angles. The non-reticular condition seems to predominate at this surface. In the non-reticular regions there may be no projecting spicules, the membrane being quite smooth. Or single spicules, and less often small tufts of 2 or 3 spicules, project. Comparative. The species here described resembles some others, assigned to Gellius by recent writers, in that the spicules are in parts of the sponge ~ combined to form spiculo-fibres, which nevertheless are poor in spongin, and do not form a continuous fibrous skeleton as in Gelliodes Ridley. Among such species may especially be mentioned Gellius flagellifer Ridley & Dendy (1887; Lundbeck, 1902). —_—— eee —_— sa oe THE SPONGES. li cso or POECILOSCLERIDAE Topsent (1894). Tylodesma Thiele. 1870. Desmacella pars O. Schmidt, 1870, p. 53. 1880. Desmacodes O. Schm. pars Vosmaer, 1880, p. 104. 1885. Gellius Gray pars Vosmaer, 1885, p. 28. 1887. Gellius Gray pars Vosmaer, 1887, p. 349. 1887. Desmacella Schmidt, Ridley & Dendy, 1887, p. 58. 1892. Bremma Gray + Desmacella O. Schm. pars Topsent, 1892, p. 80. 1894. Biemma Gray, Topsent, 1894 a, p. 11. 1902. Biemma Gray, Lundbeck, 1902, p. 82. 1903. Zylodesma Thiele, 1903, p. 944. Thiele (1903, pp. 943-44) remarks that Schmidt, when he established the genus Desmacella (1870, p. 53), included in this genus, along with his several new species, D. (Hymedesnua) johnsoni (Bwk.), which was the type of Hamacantha Gray (1867, p. 538); that Desmacella is therefore a synonym of Hamacantha, and must be cancelled, as Vosmaer has already pointed out. Thiele further calls to mind that Gray, in establishing the genus Biemna (1867, p. 538), names as the only certain species LBiemna (Desmacidon peach Bwk.); that this species is commonly assigned at the present time to Desmacella; and that the name Desmacella as used in the customary sense of to-day should be replaced by Liemna. For the group of species included in the genus Liemna (sensu Topsent, 1892, 1894 a), Thiele proposes the name of Zylodesma. Thiele’s contention that Desmacidon peachii Bwk., which has been commonly assigned since 1870 (O. Schm. 1870, p. 77) to Desmacella, and similar sponges should again be designated as species of Biemna Gray, and that Biemma sensu Topsent (1892, p. 80) cannot be maintained, is, I think, incontestable. (Rules for Zool. Nomenclature in Verhdlg. v. Intern. Zool.- Congress, p. 966, Art. 2.) The name Desmacella is strictly a synonym, and Vosmaer (1885, p. 28; 1887, p. 221) and Thiele (1903) are therefore justified in cancelling it. The new species described by Schmidt under this name (1870, p. 53) neverthe- less constituted a new group, which was homogeneous and is identical with Biemna as defined by Topsent in 1892, and recently again defined by Lundbeck (1902). Ridley and Dendy (1887, p. 58) regard these species as “the types of the genus Desmacella,’ and retain the name. Vosmaer (1880) designates them Desmacodes, but this name cannot be retained for 156 THE SPONGES. them, since in the type of Desmacodes (D. subereus O. Schm., 1870, p. 54) the predominant megasclere is an oxea (Spindelnadel). Vosmaer later (1885, pp. 28-29; 1887, p. 349) merged Desmacodes Schm. in Gellius Gray, and accordingly one of Schmidt’s species was designated by him Gellius vagabundus. But in Gellius (Gray, 1867, p: 538) the megascleres are diactinal, while in Schmidt's species they are tylostyles (Stecknadeln). In Gellius, therefore, the sponges will not go. There is the more reason to follow the example of Ridley and Dendy, and retain the name Desmacella for the group, of which D. vagabunda Schm. and D. pumilio Schm. serve as types, since Schmidt himself later (1880, p. 82) removed D. (Hymedesmia) yohnsoni (Bwk.) from the genus. Moreover, Schmidt’s generic diagnosis was obviously made especially to fit his new species, and he refers to D. yohnsoni as “ein sich isolirt habender Nebenzweig von Desmacella” (1870, p. 54). Nevertheless, the case is one in which the rules of nomenclature demand a new generic name, and I have adopted that proposed by Thiele. Tylodesma alba, sp. nov. Plate 18, Figs. 5-7; Plate 22, Figs. 2, 3. Diagnosis. Sponge body massive or lamellate. . Surface differentiated into pore and oscular regions, these regions occupying opposite surfaces when the body is lamellate, intermingling to some extent when the body is massive. In the oscular regions the dermal membrane is smooth, and imperforate save for scattered small oscula, which occur singly or in groups of 2 to 4. In the pore regions the dermal membrane is rough and exhibits numerous pore membranes perforated by 1 to a few pores. Sponge firm; color of surface white. Spicules. Tylostyles, 1275 x 36 » to 290x 8 pw. Sigmata, 64 to 22 » long. Main skeleton loose, consisting of tracts of spicules and scattered spicules.. Dermal membrane of the smooth nonporous regions densely filled with tangentially disposed tylostyles. In the rough porous regions more or less radially disposed tracts expand to form superficial brushes of small tylostyles, which project beyond the surface. Station 3405, one entire specimen and a fragment. The entire specimen has roughly the shape of a truncated pyramid, inverted so that the base of the pyramid is represented by the upper surface of the sponge, the truncated apex by the lower surface, which is attached to conglomerate. In Fig. 7, Plate 18, the sponge is viewed obliquely so that the upper surface is plainly seen. This surface is poly- gonal with six sides, longer in one direction; unevenly concave, rising THE SPONGES. 137 gradually toward the edge which is rounded and projects outward. The lateral surface of the sponge is divisible into six uneven faces, which slope suddenly away from the upper edge toward the contracted base. The height of the mass is 50 mm., its greatest width 75mm. Behind the main body, when the latter is seen in the position of Fig. 7, Plate 18, the sponge extends for a short distance in incrusting fashion over the con- glomerate, and then rises up in the shape of a small nearly vertical lamella which is partially divided into two lobes and is about 5 mm. thick. The upper surface of the main body is covered with a smooth dense membrane, which is quite imperforate save for the oscula. A similar mem- brane covers the ridges and prominent parts of the lateral surface and one surface of the lamellate continuation. The more depressed parts of the lateral surface of the main body, comprising the greater part of this surface, appear to the eye rough and comparatively porous, and one surface of the lamellate continuation has this appearance. Microscopic examination shows that in the rough regions the dermal membrane is plentifully perforated with pores. Thus the surface of the sponge is differentiated into pore and oscular regions, these regions occupying opposite surfaces where the body is lamellate, but intermingling to some extent where the body is massive. The oscula measure 1 to 2 mm. in diameter, and are found scattered irregularly over the smooth regions, sometimes singly, but more often in small groups of two to four. In the rough regions numerous pore mem- branes roofing in canals are distributed irregularly. The membranes are rounded or irregularly shaped, perforated by one to a few pores, and measure from 2 to 3 mm. to a fraction of 1 mm. in diameter. The indi- vidual pores are mostly about 200 » in diameter, with larger ones occurring less frequently. The color of the surface is white, that of the interior light brown. The sponge is firm. The flagellated chambers are rounded, 28 to 36, in diameter. The arrangement of the chambers and canals indicates that the chambers are eurypylous, although the actual openings cannot be made out. Spicules. Megascleres. 1. Tylostyle, Fig. 5, a-d, Plate 18; smooth, sharp-pointed, slightly curved; head well marked; tapering toward head end as well as toward point. Size varies from 1275 w x 36 pw to 290 ux 8x. Rarely the spicule appears strictly diactinal, bearing an enlargement at some point along its course, whence it tapers to a point at each end. Microseleres. 2. Sigmata, Fig. 6, a-g, Plate 18, scantily distributed 18 138 THE SPONGES. through parenchyma, becoming fairly abundant in places. They vary greatly in size, ranging from 64 to 22 p» in length. The microscopic pictures afforded by the spicules differ a good deal in appearance, but the spicules all have essentially the same shape. Skeletal Arrangement. The main skeleton (Figs. 2 and 38, Plate 22, sections radial to the surface) is loose, consisting of irregularly disposed tracts of spicules and scattered spicules. The larger-sized spicules, 800 « and upward in length, predominate. In the superficial region numerous tracts, extending more or less radially to the surface, are distinguishable. The spicules are cemented together here and there by very small amounts of yellowish and distinctly stratified spongin. The dermal membrane of the smooth, non-porous regions is densely filled with tylostyles, disposed tangentially or slightly obliquely to the surface, and forming several layers (Fig. 3, Plate 22; lower part of Fig. 2, Plate 22). As in the main skeleton the larger-sized spicules predominate. In the rough, porous regions, the radial or obliquely radial tracts expand to form superficial brushes of small tylostyles, which project beyond the surface (upper part of Fig. 2, Plate 22). The spicules of the brushes measure for the most part 500 to 290, in length. Some of the brushes project radially from the surface, but many project so obliquely as to lie almost flat. The flat brushes, in which the spicules diverge widely, and which consequently present a fan-like appearance, point in all (tangential) directions, and in places cross one another to some extent. Comparative. Tylodesma alba resembles in different points several of the species from the Florida coast briefly described by Schmidt (1870, p. 53) under the name of Desmacella. Thus, as in Desmacella pumilio O. Schm., the tylostyles are “theils geschichtet, theils in Fasern und ragen mit den Spitzen hervor.” The smooth dermal membrane found over a large part of the surface corresponds to that described for D. vagabunda O. Schm. Flat- tened brushes of spicules similar to those present over the rough parts of the surface in T’ylodesma alba are mentioned by Schmidt as characteristic of Desmacella vicina : “ mit flachen, oft ficherigen Ziigen von Stecknadeln.” In the fact that the pore- and osculum-bearing surfaces are differentiated, T’. alba resembles 7’. (Biemma) grimaldii Topsent (1892), which apparently lacks the striking peculiarity common to J. alba and TJ. vagabunda O. Schin., viz. the smooth dermal membrane filled with tangentially arranged tylostyles. » - @ - THE SPONGES. 139 Tylodesma vestibularis, sp. nev. Plate 18, Figs. 8, 9; Plate 19, Fig. 1; Plate 22, Fig. 4; Plate 23, Figs. 1-3. Diagnosis. Sponge primarily incrusting, but it may so grow as to completely incorpo- rate the substratum, thus appearing massive. Surface exhibits numerous vestibular spaces, appearing as elongated cavities extending tangentially, and separated from the exterior only by the dermal membrane; opening at one end by an osculum. ‘Transverse diameter of such spaces, 1 to 4 mm.; length, frequently 10 to 20 mm. Dermal membrane in general riddled with pores. Color, light yellowish-brown. Sponge moderately firm, but very brittle. Spicules. 'Tylostyles, 630 » x16 pw to 240 px 8p. Sigmata com- monly 36 to 12 p» long. Main skeleton consists of scattered tylostyles and irregularly disposed short tracts of same spicule. In the superficial region numerous radial or ob- liquely radial tracts are distinguishable, ending at the surface in projecting brushes. The adjacent obliquely radial tracts, with their terminal brushes, are prolonged into the vestibular membranes, there occupying an approximately tangential position. Station 3405, one specimen. The sponge (Fig. 1, Plate 19) is incrusting, below upon conglom- erate, above upon the dictyonal framework of a Hevactinella, which agrees, in regard to the framework, with fH. labyrinthica mihi, and very probably is this species. The thickness of the incrusting sponge, external to the conglomerate or Hexactinellid support, is about 1 mm., or often less. The conglomerate is in part a firm, solid mass; in part, of a very loose composition. Where the mass is loose, consisting of bits of shells, spines, annelid tubes, and Polyzoa, the sponge has grown into all the crevices between the component particles, and aids in holding them together. Above, the sponge does not form a mere incrustation upon the surface of the Hexactinella skeleton, but has incorporated the latter, having so grown through its interstices that the Hexactinellid framework is now found in the interior of the sponge, along with the proper Monactinellid skeleton (vide Fig. 4, Plate 22, and Fig. 1, Plate 23, sections vertical to the surface). Over a part of the surface the lobes of the supporting Flexactinella remain distinct, although they have been individually incor- porated by the Tylodesma. But over most of the surface the primitively incrusting sponge has filled up the gaps between the Hexactinella lobes, thus assuming the character of a continuous amorphous mass. This mass is, however, excavated internally by some large cavities, which probably represent spaces between the Hevactinella lobes. The Hexactinella skeleton is in a measure disintegrated, and along with it the massive part of the Tylodesma has incorporated other fragments of a stony nature, most of 140 THE SPONGES. which seem to be particles of echinoderm spines. The massive part of the specimen is 70 mm. wide, 30 mm. thick, and 40 mm. high. The surface exhibits numerous vestibular spaces, appearing as elongated, irregularly tubular cavities, often branching, extending tangentially beneath the surface, and separated from the exterior only by thin dermal membrane. One of the largest of these spaces is shown in Fig. 1, Plate 19, to the left. The transverse diameter of such spaces varies from about | to 4 mm. The length, which is often difficult to measure, owing to the meandering course of many of the spaces, is frequently 10 to 20 mm. At one end many, probably all, of the spaces communicate with the exterior through an osculum 1 to 4 mm. in diameter. (The surface of the sponge is injured here and there, and the natural apertures are not everywhere discernible with certainty.) The membranous covering of the spaces is moreover perforated here and there by apertures 85 to 200 » in diameter, scattered singly, or in small groups. At the non-oscular end the vestibular spaces lose themselves in the more solid sponge tissue. The spaces are larger and comparatively far apart in the massive part of the sponge body, smaller and much more abundant where the sponge is spreading over a loose, broken substratum. The surface of the sponge between the vestibular spaces appears to the eye dotted with small, round areas, about 0.5 mm. in diameter. These vary greatly in abundance, being in places 1 to 2 mm. apart, but again only scantily scattered. They are perforated membranes roofing in canals of corresponding size, which pass radially into the interior. The mem- branes for the most part contain several apertures, but sometimes only one, which probably are to be regarded as oscula. The dermal membrane in general is riddled with thickly strewn pores, which vary considerably in size, the diameter ranging at any rate from 85 to 220 ». Small subdermal cavities everywhere underlie the dermal mem- brane. The flagellated chambers are 32-36 mw in diameter, and are crowded together in regions which are separated by collenchymatous tracts, the latter traversed by the larger canals. The arrangement of the chambers in the trabeculae of the sponge indicates that they are eurypylous, The color is alight yellowish-brown, the membranes roofing in the vestibular spaces appearing translucent and darker than the general surface, when the body is immersed. The sponge, while moderately firm, is exceedingly fragile, owing to its great brittleness. THE SPONGES. * 141 Spicules. Megascleres. 1. Tylostyle, Fig. 8, a—c, Plate 18; smooth, very slightly curved, with small head. Spicule tapers slightly toward tylote end as well as toward pointed end, but in the smalier sizes the tapering toward the tylote end is scarcely perceptible. Size ranges from 630 p x 16 » to 240 %x 8p. The smaller sizes— 240 to 350 » in length — pre- dominate in the surface brushes and the vestibular membranes; the larger, in the radial tracts and the loose skeleton of the interior. Microseleres, 2. Sigmata, Fig. 9, a-c, Plate 18. Length ranges from 45 to 10%; common sizes from 36 to 12 yw in length. The sigmata are abundant in the parenchyma and general dermal membrane ; only scantily present in dermal membrane over the larger vestibular spaces. Skeletal Arrangement. In the deeper parts of the sponge which are occupied by the Hexacti- nellid skeleton, tylostyles are scattered separately and in slender short tracts, without arrangement. In the superficial region numerous radial or obliquely radial tracts extend toward the surface, there ending in pro- jecting brushes composed of diverging short tylostyles (Fig. 4, Plate 22; Fig. 1, Plate 23). Spongin appears to be absent. At any rate, it was not to be observed either in balsam or glycerine sections or teased prepa- rations. While the surface in general is covered with the projecting brushes, between which small subdermal cavities very commonly lie, in the dermal membrane covering the vestibular spaces the surface skeleton has a different character. The obliquely radial tracts which are adjacent to such a space extend out into the covering membrane (Fig. 4, Plate 22, section ver- tical to the surface; Fig. 3, Plate 23, surface view of comparatively large vestibular space with some of the surrounding area), thus coming to occupy a tangential or nearly tangential position. In the case of the smaller and medium-sized spaces the tangential tracts, as they pass from the margin toward the middle of the vestibular membrane, preserve their individuality (Fig. 5, Plate 23). In some cases, the whole tract occu- pies a tangential position, the terminal spicules spreading out fan-wise in the horizontal plane. In other cases, while the body of the tract lies tangentially, the terminal spicules form a diverging bunch which points obliquely upward much like the bunches of spicules found over the general surface. Both conditions appear in Fig. 3, Plate 23. In the tracts of spicules which extend out into the membranes covering 142 THE SPONGES. the larger vestibular spaces, all of the spicules lie in a tangential or nearly tangential position. These tracts lose to a greater or less extent their individuality (Fig. 2, Plate 23, surface view of part of a large vesti- bular membrane. The upper, left, and lower margins of the figure repre- sent cut edges. The right curved margin represents part of the edge of an osculum. The left and lower margins are not far from the periphery of the entire membrane), in that they become loose and fray out terminally into free spicules, which are scattered in moderate number through the membrane. Some of the tracts are prolonged for considerable distances through the membrane as narrow stream-like bands, which eventually break up into free spicules. Comparative. The spicules in 7’ vestibularis are pretty close to those of T. corrugata (Bwk.) (Biemma corrugata, Topsent, 1892), a parasitic form. Moreover, Topsent says the spicules at the surface are arranged in diverg- ent bunches (‘en bouquets divergents”). But Bowerbank describes (1866, pp. 242-3) and figures (1874, Plate XLIII., Fig. 3) the dermal membrane of this sponge (Halichondria corrugata Bwk.) as strongly reticulated, and vestibular spaces such as occur in 7’ vestibudaris are not mentioned by either writer. Lundbeck (1902, p. 82) describes in detail a Tylodesma (Biemma rosea - Frst.) known in plate-like fragments, which bear the pores on one surface, the oscula on the other. With this well-marked species, which he so excel- lently describes, Lundbeck thinks it possible to identify another specimen of a very different habitus. This is a little sponge occurring as a thin incrusta- tion on a Hexactinellid skeleton, and which Lundbeck regards as a young individual. In the description of this specimen Lundbeck does not go into details, and it may be questioned whether it belongs to 7. rosea. Lund- beck’s description of 7’. rosea in general would indicate that this particular specimen and 7’. vestibularis have some points of resemblance in addition to the parasitic habit. ; ( | r 1 ; : THE SPONGES. 143 Tophon Gray. 1867. Lophon + Alebion Gray, 1867, p. 534. 1887. Lophon Gray pars, Vosmaer, 1887, p. 354. 1887. ILophon Gray, Ridley & Dendy, 1887, p. 116. 1892. Dendoryx (Iophon) Gray, Topsent, 1892, p. 96. 1894. Lophon Gray, Topsent, 1894 a, p. 14. Iophon chelifer Ridley and Dendy. 1887. Lophon chelifer Ridley & Dendy, 1887, p. 119, Plates XVI., XVII. 1893. Iophon chelifer R. & D., Lambe, 1898, p. 30, Plate IL, Figs. 7, 7,a 1896. es ss se Lambe, 1896, p. 191. 1900. = s s$ Lambe, 1900, p. 23. Iophon chelifer ostia-magna, subsp. nov. Plate 20, Figs. 2, 4,10, 11; Plate 24, Fig. 1. Diagnosis. Body plate-like, 5 to 8 mm. thick, with rounded free edge bearing large oscula 4 to 6 mm. wide, which are the openings of correspondingly wide efferent canals. Color dark brown. Upper and lower surfaces alike. Small oscula mostly 300-500 pw in diameter, but reaching diameter of 2mm., scattered abundantly over both surfaces, and also present at the free edge. Pores abundant, scattered throughout dermal membrane. Spicules. 1. Style, 440 by 20 pw, sparingly spinose. 2. Subtylote, 315 by 8 yp, ends feebly spinose. 3. Chelate bipocillus, 16-20 » long; axis terminating at one end in 2 or 3 pointed teeth, at other end in a curved plate divided into 2 or 3 lobes. 4. Anisochela, 12-20 » long, palmate. Main skeleton a loose reticulum with squarish meshes, the side commonly formed by a small fascicle of spicules. Continuous bundles more or less radial to the surface, distinguishable as in type. Station 5584, two specimens. The body (Fig. 4, Plate 20) is plate-like, 5 to 8 mm. thick, but not very flat; the plate somewhat bent here and there, and with both surfaces made uneven by irregular depressions and elevations. Both specimens are fragmentary, but include a part of the natural free edge of the sponge. This is rounded off alike toward the two surfaces, and bears several large oscula, which are somewhat elongated in the horizontal plane of the sponge body, measuring in this plane 4 to 6 mm. These large oscula lead into efferent canals of corresponding width, which are about 10 mm. deep, passing inward in the horizontal plane of the body. Other smaller, rounded oscula, 1.5 to 2 mm. in diameter, are also present on this edge. There is no discoverable difference between the two surfaces, which to the eye appear porous. The color is dark brown, and the sponge very fragile, owing to its great brittleness. . 144 THE SPONGES. Collenchyma is found in some abundance at the surface in the shape of small, irregular, and vaguely defined areas, which in the uninjured sponge are inconspicuous, although evident in preparations. As in the other species of the genus, the skeletal reticulum is absent from such areas. In the intervening regions constituting the greater part of the surface, the skeletal reticulum lies beneath the dermal membrane in the usual way. Preparations of the surface show that there is no constant relation between the superficial collenchymatous areas and the presence of oscula. The latter, measuring commonly 300-500 » in diameter, and occasionally reaching a diameter of 2 mm., are scattered abundantly over the surface, ~ and occur both in the collenchymatous areas and in the intervening regions. Smaller apertures of all sizes, from 35 pw to 300 pw in diameter, are also scattered abundantly over the whole dermal membrane. The smaller are doubtless pores. It is not, however, possible in this sponge to distinguish, by their morphological characteristics alone, the smallest oscula from the larger pores, since there is such a perfect intergradation in size, and since the oscula’and pores are both irregularly scattered. The flagellated chambers in the present condition of the specimens vary in diameter from 24 to 32 ». Some are spheroidal, others markedly com- pressed. Their arrangement indicates them to be eurypylous. Spicules. Megascleres. 1. Style, Fig. 10, Plate 20. Spicule about cylindrical, slightly curved, very sparingly spinose throughout its length; pointed end sometimes rounded. Size, 440m x 20 p. 2. Subtylote, Fig. 11, Plate 20. Spicule very slightly, sometimes not, enlarged at the ends ; smooth, except at the extreme end, where it is feebly spinose. Size, 315 x 8p. ; Microseleres. 3. Bipocillus 16-20 pw long, Plate 20, Fig. 2. a, C, a. The curved axis terminates at one end in 2 or 3 pointed teeth, which project toward the opposite extremity. At the other end the axis ter- minates in a thin plate-like expansion with spherical curvature, divided by one or two narrow incisions into 2 or 3 lobes. Axis has a ventral keel, which disappears toward the toothed end. On each side of the keel, axis thins away, forming a lateral flange, which is sharply marked off from _ the terminal lobe of that side by a rounded incision. 4. Anisochela 12-20 w long, Plate 20, Fig. 2b. Spicule of the pal- mate type, with a little spine at the small end. Skeletal Arrangement. The main skeleton (Fig. 1, Plate 24, a section : : ee THE SPONGES. 145 vertical to surface, and extending from the surface to one of the large efferent canals. The right margin of figure represents surface of sponge. The left margin represents the canal wall) is a reticulum, formed of spinose styles. The reticulum is loose; meshes commonly squarish, though often subdivided obliquely into triangular meshes. Side of mesh equals length of spicule, and is commonly formed by several spicules (2, 3, 4, or even more), making a loose bundle. Continuous bundles, or tracts, more or less radial to the surface, are distinguishable. In places the reticulum might be described as made up of these tracts, with transverse connectives. At the angles of the meshes the. spicules are united by spongin. The dermal skeleton consists of the superficial layer of the main skeletal reticulum, and of abundant subtylotes. The latter are scattered without order, singly and in loose fascicles, both in the collenchymatous areas of the ectosome and in those parts directly supported by the skeletal reticulum. The microscleres occur in the dermal membrane, in the walls of the larger canals, and in the parenchyma in general. They are only fairly abundant. Comparative. The sponge just described closely resembles Jophon chelifer Ridley and Dendy (1887, p. 119). The bipocilli are not only chelate, but in general shape and in size are nearly identical with those of the latter species. The skeletal reticulum in both forms shows vaguely developed fibres, which extend more or less radially to the surface. The spinose ‘styles and tylotes are of about the same size in the two forms. Ridley and Dendy describe Lophon chelifer as “ amorphous, massive, honeycombed,” and add, “ Exact form uncertain, specimen fragmentary.” The “Challenger” specimens were taken lat. 35° 4’ S., long. 18° 37’ E., off the Cape of Good Hope; lat. 46° 41’ S., long. 38° 10’ E., off Prince Edward Island; lat. 46° 55’ S., long. 51° 52’ E., between Prince Edward and Ker- guelen Islands; the depth varying from 150 to 550 fath. I have examined the type specimens of J. chelifer, and I find that although they are amorphous there is some reason for regarding them as thin plates which, because of the irregular character of the growth, have assumed an amorphous character. Actually, however, they differ markedly in appearance from J. chelifer ostia-magna. While the skeletal re- semblances between my subspecies and the type are very close, the chelate 19 146 THE SPONGES. character of the bipocillus cannot be regarded as a feature indicative in itself of species-relationship, since in the very different Jophon lamella, the bipocilli are also chelate. Lambe has recorded Jophon chelifer R. and D. from the Pacific coast of Canada (1893, p. 30), and from several localities off the Atlantic coast of Canada (1896, p. 191; 1900, p. 23). Lambe’s specimens differ in hab- itus from mine, being “amorphous and honeycombed,” or massive, per- forate, and consisting “of an inosculation of short, stout, irregularly shaped, nodose branches, which coalesce, frequently to such an extent as to become amorphous ”’ (1893, p. 30). Iophon lamella, sp. nov. Plate 20, Figs. 3, 7-9, 12, 13. Plate 24, Figs. 2-4. Diagnosis. Body lamelliform, 5 to 12 mm. thick. Efferent canals, 1-2 mm. in diameter, open in abundance over both surfaces. The surfaces, upper and lower, are in a measure differentiated. Pores irregularly scattered throughout the dermal membrane, wherever it overlies the skeletal reticulum. Spicules. 1. Spinose style, 210-220 w x 12-16 4y.° 2. Tylote, 220-240 » x 7-8 », ends minutely spinose. 3. Chelate bipocillus, 12-16 p» long; terminating at small end in two pointed teeth, at larger end in a bilobed plate. 4. Ani- sochela, 14-28 » long, palmate. Main skeleton a uniform reticulum of spinose styles. Meshes commonly triangular. Side of mesh formed by 1, 2, 3, or occasionally more spicules. Station 3405, five specimens. The sponge body is lamellar but irregularly thickened, and sometimes considerably curved; the free edge not possessing special characters dis- tinguishing it from the rest of the surface. The thickness varies from 5 to 12 mm.; greatest width, about 50 mm. The sponge is firm, and while easily broken is not especially brittle. The color is a light yellowish-brown. The upper surface of a specimen is shown in Fig. 12, Plate 20, and the lower surface of the same specimen in Fig. 13, Plate 20. The surfaces of the plate-like body are in a measure differentiated. The one surface, designated as the upper, is more even and in general of a lighter color. The other surface, regarded as the lower, exhibits shallow, irregular, and large concavities, as if here moulded over an underlying object. The appearance of the surface is extremely variable, although there is an underlying uniformity of character. This appearance is largely ) > > 4 ; THE SPONGES. 147 determined by the character of the main efferent canals, which conditions the arrangement of the oscula and superficial collenchyma. The main efferent canals are numerous, cylindrical, and pass radially into the body from both surfaces. The diameter does not exceed 2 mm., and commonly is 1 to 2 mm. The canals penetrate deeply into the body, and in many cases pass completely through the body from one surface to the other. The oscula are sometimes single apertures, but often the end of the canal is covered in by a fenestrated membrane, including a few, 3 or 4, apertures. When the canal passes completely through the body, at least one end seems always to be covered in with a fenestrated membrane. Near the lower surface of the sponge several canals, which open inde- pendently on the upper surface, may unite and thus produce a vestibular space which is separated from the exterior only by the dermal membrane of the lower surface. Such vestibular spaces are abundant in some speti- mens on those parts of the lower surface which seem to have been moulded over an underlying object (Fig. 13, Plate 20), and here appear as de- pressed membranous areas, which are usually elongated, often somewhat meandering. (In the figure they appear darker, the more solid sponge tissue between them reflecting the light better —the sponge being im- mersed.) They open by oscula, in the case of the larger spaces by several, which range from a small size to a diameter of 2.5 mm. In these specimens the vestibular spaces are found only on the lower surface. The common type of osculum, represented by the single apertures or fenestrated membranes of the canals which open independently on the surface, is in some regions not surrounded by collenchyma (Fig. 12, Plate 20), in other regions is so surrounded (Fig. 3, Plate 20). The oscula again may not be depressed (middle part of Fig. 12, Plate 20), or may be markedly depressed (Fig. 3, Plate 20). The oscula, and associated canals, may in one portion of a specimen be so numerous as to honeycomb the sponge (left of Fig. 12, Plate 20), and in another region (middle of same figure) be comparatively far apart. The oscula are in general more abundant on the upper surface. In two of the specimens large parts of both surfaces present a striking modification, which may be referred to as the reticulate modification. In these regions the surface is comparatively smooth and exhibits collen- chymatous areas of a rounded, polygonal shape and 1 to 2 mm. in diameter, separated by narrower tracts of the more solid sponge tissue (Fig. 2, 148 THE SPONGES. Plate 24). A small osculum, about 0.5 mm. in diameter, lies in the centre of the area, and this is surrounded by a few other, usually smaller, apertures, or by the ends of canals abutting against the dermal membrane and appearing as apertures. The oscula lead into canals of corresponding size which penetrate, radially or obliquely, deep into the body of the sponge, where they continue to be surrounded with collenchyma as at the surface. The mass of collenchyma surrounding the main efferent canal, as may be seen in sections taken vertically to the surface of the sponge (Figs. 3 and 4, Plate 24, sections passing entirely through the lamellate body), passes through the body from one surface to the other, and is honeycombed by numerous smaller canals. The main efferent canal itself, on the other hand, which is well shown in the middle of the micro- photograph, Fig. 3, Plate 24, does not appear to pass through the entire thickness of the body. The collenchymatous tracts both at the surface (Fig. 2, Plate 24) and in the interior (Figs. 5 and 4, Plate 24) lack the skeletal reticulum, which everywhere permeates the intervening regions. The kind of structure, which has just been described, is obviously to be regarded as a modification brought about by the excessive development of collenchyma round the main efferent canals, coupled with the diminution in diameter of these canals. The specimens exhibiting the reticulate modification are elsewhere like the other individuals, the body being penetrated by the common larger type of efferent canal, the surface appearing uneven, uniformly dense, without obvious collenchyma, and showing irregularly scattered oscula about 1 mm. in diameter. The pores, measuring 60-150 pw in diameter, are scattered irregularly but thickly on both surfaces of the body over the solid tissue intervening between the oscula, vestibular spaces, or the reticulately arranged collen- chymatous areas. The flagellated chambers have a shrivelled appearance due doubtless to the faulty preservation. They now measure about 20 pu in diameter, and their arrangement indicates them to be eurypylous. Spicules. Megascleres. 1. Style, 210-220 » x 12-16 yp, Fig. 8, Plate 20. Spicule nearly cylindrical, slightly curved, spinose with small sharp prickles. The prickles are more abundant near the ends, less abundant in the middle. The extreme point is smooth. Rounded end and spinose region near the point sometimes slightly dilated. 2. Tylote, 220-240 » x 7-8 yp, Fig. 9, Plate 20. Spicule slightly ’ THE SPONGES. 149 thicker in the middle, tapering toward each end. Heads small, minutely spinose over distal half. Frequently one or two prickles on shaft, close to ends. Precise character of end varies: end commonly enlarged and rounded, but sometimes enlarged and irregular; sometimes not enlarged. Microscleres. 3. Bipocillus, 12-16 » long, Fig. 7 @ and ¢, Plate 20. Curved axis shows a thickened median keel, which disappears toward small end of spicule. On each side of keel, axis thins away, forming a lateral flange which is sharply separated by a rounded incision from the terminal plate. Axis terminates at one end, the larger, in a thin plate-like expansion which has a spherical curvature, and is divided by a narrow median incision into two lobes. At the other end axis terminates in two pointed teeth, which project toward the larger end. 4. Anisochela, 14-28 » long, Fig. 7 6, Plate 20. The smaller sizes are the commoner. Spicule of palmate type; a little spine at the smaller end. Skeletal Arrangement. Main skeleton a uniform reticulum of spinose styles. Meshes are commonly triangular, but the shape may be construed as due to the fact that a spicule or a small fascicle of spicules extends obliquely across a squarish mesh, acting perhaps as a brace. Side of mesh equal to length of a spicule and formed by 1, 2, 3 or occasionally more spicules. At the corners of the meshes the spicules are united by masses of spongin, which is colorless. Dermal skeleton consists of the outermost layer of the skeletal reticulum, and of scattered tylotes. The latter are frequently found in loose fascicles or tracts, and occur throughout the dermal membrane. . The microscleres are present in the dermal membrane, and in the paren- chyma. They are very abundant in the walls of many of the canals. Iophon lamella indivisus, subsp. nov. Plate 20, Figs. 14-16. Diagnosis. Sponge distinguished from the type by the character of the bipocillus, which is not chelate. Bipocillus, 8-10 » long, terminating at the large end in a curved plate of rounded outline, which is ordinarily not divided, terminating at the other end in a smaller plate with denticulate margins. Station 3405, 6 specimens. Along with the specimens of Jophon lamella were taken six other speci- mens, four of which are fragmentary, having the same plate-like habitus 150 THE SPONGES. and the same general arrangement of the canals. The skeleton too is similar to that of J. damella except in the matter of the bipocilli. The upper surface of a specimen is shown in Fig. 16, Plate 20, and the lower surface of the same in Fig. 14, Plate 20. The efferent canals as in the type pass into the body from both surfaces, sometimes passing through from one surface to the other. As in the type the upper surface is lighter in color, and bears more numerous oscula than the lower surface. On the lower surface elongated vestibular spaces are extensively developed, appearing as furrows lined with smooth membrane. The flagellated chambers are of the same size as in the type. The two good specimens differ from the type as regards the detailed appearance of the upper surface. The point is doubtless one of individual difference, and in the remaining specimens could not be determined. In these two specimens the upper surface bears abundant depressions, many of which are furrow-like. The efferent canals open in the depressions (Fig. 16, Plate 20). Here and there several efferent canals, instead of opening separately, unite beneath the dermal membrane of this surface to form a vestibular space, essentially similar to those which are more con- spicuously developed on the lower surface. Skeleton. The megascleres, styles, and tylotes, are like those of the type, and the skeletal arrangement offers no points of difference. Microseleres. 1. Bipocillus, 8-10 p» long, Plate 20, Fig. 15, a, ec, d, e, f. The curved axis terminates at one end in a thin plate-like expansion having a spherical curvature and a rounded outline. This is usually undivided, but occasionally spicules are found in which it is divided by a deep median incision into two lobes, as in J. damedla. At the other end the axis terminates in a smaller curved plate, which is pointed, and in which the free edge on each side of the terminal point is minutely denticu- late. Axis itself, near the larger end of the spicule, flattens out on each side, forming a thin lateral flange, which is separated from the terminal plate by a rounded incision. The spicule is scantily present in the paren- chyma, more abundant in the dermal membrane. . 2. Anisochela, 12-28 » long, Plate 20, Fig. 15%. Spicule does not differ from anisochela of the type, and is scantily present in the dermal membrane and parenchyma. Comparative. The chelate character of the bipocillus makes a striking point of resemblance between Jophon lamella and Iophon chelifer R. and D. ——— : THE SPONGES. 151 But when a comparison is made between the three forms J. chelifer ostia- magna, I. lamella, and I. lamella indivisus, it becomes obvious that the chelate character is in itself not a guide to relationship. On the one hand, two sponges (J. chelifer and J. lamella) may occur which differ widely in most respects, but agree in having the chelate bipocillus. While on the other hand two sponges (J. lamella and I. lamella indivisus) occur agreeing in most particulars, but having the one chelate, the other non-chelate bipocilli. Iophon indentatus, sp. nov. Plate 19, Fig. 6; Plate 20, Figs. 1, 5, 65 Plate 23, Fig. 4. Diagnosis. Sponge incrusting, 2-3 mm. thick, fragile, of brown color. Surface in- dented with polygonal collenchymatous depressions 0.5 to 1 mm. in diameter, separated by narrower ridges of more solid skeletogenous tissue. Oscula, 150-200 » in diameter, occupy the centres of the depressions. Pores, 75 u in diameter, scattered over the ridges. Spicules. 1. Spinose style, 220 w x 14-16 p. 2. Subtylote, 220 « x 8p, minutely spinose at extreme ends. 3. Bipocillus, 8 » long; smaller end, a curved plate with den- ticulate margins; larger end, an undivided curved plate of a rounded outline. 4. Ani- sochela, 14 » long, palmate. Main skeleton a uniform reticulum of spinose styles. Side of the squarish or triangular mesh formed by 1, 2, or occasionally 3-4 spicules. Station 8405, 3 specimens. The sponges are all incrusting upon a Gorgonia. The incrustation is 2 to 3 mm. thick, extending in places in the shape of sheets which occupy the axils of the Gorgonia branches. The color is a rather light brown. Sponge fragile, easily torn and broken. The surface (Fig. 1, Plate 20) is indented with collenchymatous de- pressions of a polygonal or rounded polygonal outline, 0.5 to 1 mm. in diameter. These depressions, which appear to the eye translucent and gelatinous, are separated by considerably narrower ridges composed of the more solid sponge tissue. A small osculum, 150 to 200 » in diameter, lies about in the centre of an area, and in some areas is surrounded by a few other smaller apertures. Between the collenchymatous areas, over the surface of the ridges, abundant pores measuring about 75 mw in diameter are irregularly scattered. The surface resembles that of the smooth, reticulate portions of Jophon lamella’ (Fig. 2, Plate 24), but the collen- chymatous areas are considerably smaller and less sharply limited than in the latter species. 152 THE SPONGES. The oscula lead into main efferent canals which penetrate deeply into the interior of the sponge, where they continue to be surrounded by a thick layer of collenchyma. The canals are of about the same diameter as the oscula. Some of them are, throughout the thickness of the sponge, radially directed to the surface, but more are obliquely inclined, often curving so that a section which is vertical to the surface of the sponge cuts them transversely. This is the case in Fig. 4, Plate 23, which repre- sents such a section taken through the Jophon and a part of the underlying Gorgonia axis. The-collenchyma surrounding a main canal in the sponge interior forms a roughly cylindrical tract traversed by the canal, and preserving approxi- mately the diameter which it has at the surface of the sponge. These tracts cut the sponge body up into intervening regions permeated through- out their extent by the skeletal reticulum, which does not extend into the collenchymatous tracts themselves. The parts of the body permeated by the skeletal reticulum may be thought of as partitions between the collen- chymatous tracts. These partitions more commonly have a thickness less than the diameter of the collenchymatous tracts, appearing in sections as thin trabeculae, as in Fig. 4, Plate 23. Elsewhere, however, in the same specimen the skeletogenous partitions may appear as thick masses. As regards the arrangement of the main canals and the relative disposition of skeletogenous and collenchymatous tracts in the sponge interior, there is much resemblance between this species and the reticulate specimens or parts of specimens of Z. Jamella, but in the latter the main canals are more frequently radially directed, and both the collenchymatous tracts and inter- vening skeletogenous portions are thicker and probably on this account appear better defined in sections (Plate 24, Figs. 3 and 4). Spicules. Megascleres. 1. Style, 2204 x 14-16 p, Plate 20, Fig. 5. Spicule spinose with small, sharp prickles, which are stronger and more numerous near the ends. Extreme point smooth. Slightly curved, nearly cylindrical, very slightly enlarged at rounded end and near the point. 2. Subtylote, 220 » x 8 p, Plate 20, Fig. 6. Very slightly, scarcely at all, curved. Tapering a little from the middle toward ends, which are scarcely enlarged and most minutely spinose. Shaft in general smooth, but near the ends are a few scattered prickles. : Microscleres. 3. Bipocillus, 8 » long, Plate 19, Fig. 6d, c, d, e. Curved axis at the smaller end terminates in a spoon-shaped expansion with den- = EEE ee THE SPONGES. 153 ticulate margins. This expansion at the extreme end sometimes appears rounded, and sometimes angular. The difference in appearance is probably due to a difference in position, and the end is probably always angular. Toward the other end the axis develops the usual thin lateral flange, beyond which there is the usual incision separating the flange from the large terminal plate with rounded outline and spherical curvature. The denticulate plate is sometimes nearly equal in size to the larger plate. In minute details the spicule differs from the very similar bipocillus of Z. lamella indivisus (comp. Plate 20, Fig. 15). 4, Anisochela, 14 » long, Plate 19, Fig.Ga. Spicule, of the palmate type common in the genus, with a little spine at the smaller end. Skeletal Arrangement. Main skeleton consists of a uniform reticulum of spinose styles. Meshes squarish or triangular. The side of a mesh is equal to the length of a single spicule, and is formed by one, two, or occasionally three or four spicules. Spicules at the corners of the meshes are united by spongin. Dermal skeleton consists of the superficial layer of the skeletal reticulum and of abundant subtylotes, which are scattered irregularly, often in loose tracts. | The microscleres are abundant in the dermal membrane ; also present in considerable abundance throughout the parenchyma, especially in the walls of, and in the tissue immediately surrounding, the larger canals. Comparative. Ridley and Dendy combine (1887, p.-117), under the name of Iophon pattersoni (Bwk.), a number of previously described species, and record under this head specimens taken by the “ Challenger” off the coast of Patagonia and Tristan da Cunha. All of these forms have palmate anisochelae with pointed smaller ends, up to 30 u long, and minute bipocilli. Lophon indentatus must be very similar, judging from Bowerbank’s figures, in surface appearance to one of the species combined, vz.: Halichondria nigricans Bwk. (Bowerbank, 1866, pp. 266-68; Bowerbank, 1874, Plate XLY. Fig. 25) which occurs as a “massive” body and also incrusting. I have examined type specimens in the British Museum of this and the other Bowerbank Jophons, but the specimens are dried and old, and no longer permit the character of the surface and the canal arrangement to be studied. According to Ridley and Dendy (4 ¢. p. 118) in Halichondria nigricans the spined styles measure 218 p x 8 p, the tylotes 195 p by 3 pw, and are thus much slenderer than in J. indentatus. 20 154 THE SPONGES. Tophon indentatus also resembles in surface appearance Alebion proximum Ridley (Ridley, 1881, p. 114), another of the species combined by Ridley and Dendy. But the styli here are 158 » x 9, and thus much smaller than in my sponge. Moreover, the skeletal reticulum is described as composed of primary fibres, five to six spicules thick, running from the base to the surface and crossed by secondary bars approximately at right angles, — an arrangement not found in J. indentatus. The “Challenger” specimens of Jophon patterson’ R. and D., which I have examined, differ markedly in surface appearance from my form. They are, as Ridley and Dendy describe them, massive and amorphous. I may add that they are honeycombed with comparatively large canals, which open over the whole surface. The spines and tylotes (R. and D., 1887) are considerably slenderer than in J. indentatus. On the other hand, I find the bipocilli are of about the same size as in my form, and have a similar shape, the smaller end being denticulate. But this point of resémblance probably means little, since minute bipocilli with denticulate small end also oecur in Lophon radiatus Topsent (Topsent, 1901 a, p. 22, Plate III. me and in Lophon lamella indivisus. Topsent (1892 under Dendoryx (Lophon) nigricans Bwk.) criticises Lophon pattersoni sensu Ridley and Dendy, and is disposed to regard it as a hetero- geneous group, on the score that some of the forms which Ridley and Dendy combine, and which have been taken several times, are readily dis- tinguishable. I must say that I find the published data for the union of these several forms inadequate. I therefore designate my sponge as a new species, although Jophon pattersoni as conceived by Ridley and Dendy is probably comprehensive enough to include it. aE —— ee —— = - 77). 2 Or Qn THE SPONGES. 15! AXINELLIDAE Ridley and Dendy. Phakellia Bwk. 1864. Phakellia Bowerbank, 1864, p. 186. 1880. Phakellia Bwk., O. Schmidt, 1880, p. 81. 1887. ae ae) FAgingodoity i ert . iY a aqpovoiiy #: 4034 Puate 1. Spinose diact from marginal fringe; x 400. Pinule from oscular (gastral) surface; x 400. Dermal pinule; x 400. Mesamphidise; x 600. Macramphidise; x 400. Macramphidise; x 400. Macramphidise; x 400. Micro-oxyhexact; x 600. Micro-oxyhexact; x 600. Amphidise with 4-rayed umbels; x 400. Amphidise with 4-rayed umbels ; x 400. From a photograph; }. From a photograph ; 45. vA ry on et F Peay, ¢ CP i Leake. © » we > a athe. 006 x ; ci idfeyistary tonlf - . . a Jarrod osrastod . % y A » " : F ? * . ; \ i id iu - > i’ a 4 4 4 a wae : ihe Baie Oo as Hi » » F008 x ; oni bile ares O08 x ; cetbrdqur _.=— = a is =~ 2 PLATE 2. Hyalonema bianchoratum. Macramphidise ; x 300. Sp. « Dermal pinule; x 100. Dermal pinule ; x 100. Dermal pinule; x 100. Macramphidise; x 300. From a photograph; }. 5 Mesamphidise; x 300. Macramphidise; x 300. Micro-oxyhexact; x 300. Micro-oxyhexact; x 300. Micro-oxyhexact; x 300. Micro-oxyhexact; x 300. Macramphidisc; x 300. Mesamphidise; x 300. Mesamphidise; x 300. Macramphidise; x 300. Sh. Ex. 18 ALBATROSS SS ee ms _~ 4 - SSS =p \ H.Y.Wilson and EwRubsamen, del ae aa? 2 arata 000 = ‘ tonxodyx0-0n9il/ et dev “i navy J AOD. x, , veibidquaxronl! OO x ;s0exadyx0-079i 17 ° OO) x 5 ht i Pelpeers ot | - isisintd .digal uf or | ; obi : 08 a cy er J. Bis) qa luoe gerbi sP¥aas sob Tyg far y\ .' Batt § law ¥lnod 4189 ol . inermdiqe: oie: alosth lsurydensieg t6bsals Mid gar? encoun veol an exoniedl Yo ¥ r y! . > - ‘ Ba ditw sigisia sqqu to Imeq sation s. . i Boiss shila. soetawe Inno geste: molt buy area ised tqooK9 Ob m3 ele on or ned’ uo 4 ee x ys r b ‘ — Se ' . > 2 s! . , 7 a Xs ; y gh y : , otf * 4 ’ a " 4 Me ae on 5 4 ’ 6 . a * Fad “, t ' rh hb a sah Ba ; thi \ , ey . ] 4+ 4 w ‘ was ' ng } - 7 | ‘ t " . ~ f oR wt PLATE 3. Hyalonema pedunculatum. Micro-oxyhexact; x 600. Macramphidise; x 400. Micro-oxyhexact; x 600. Dermal pinule; x 400. From a photograph. Lateral part of body sliced away on one side; }. Mesamphidisec; x 600. Regadrella delicata. Part of margin bounding oscular (sieve-plate) area, with adjoining lateral body wall. Dermal surface. The slender parenchymal diacts are represented, for the sake of clearness, as less numerous than in nature; x 14. Another part of upper margin, with adjoining lateral wall. Dermal surface. Slender, parenchymal diacts omitted except near margin, and there represented as less numer- ous than in nature; x 45. iin we. ALBATROSS Ex. 1891. SPONGES.PL.3 H.V.Wilson and Ew.Rubsamen, del ae. 2 a sliew wbod-dpyets! witintayy ae. e8Ig31 Netw ybod ons 1 RV Sue = BF om ; ota ich seta a80%9y Con beaodest inge: hse Ydod Ton : a Tf padqoa woe int f / Ih i> : rhasw ! “ : ; ; aa Bs ‘ a 7 : 2 os _ — e o.; tae ’ < ira H eee: us') ‘i stew ‘ ' - - ai"’ E Ted nd } : é& Oat x": codtitiaiinrier.: fuerti an rus athe Es » + $ - ul } * t aie > « a 7 ,U) : ‘ = -a9wol to Boonie Abe wre tha lnwentl = Seite lettety | Wiehe! of deine! %3 ® VP. ae n . ma & ts > ’ * B.S) : ~~. . 4 7 as =, .. " i * = t ¥, 4 ~ gd at Fah Py ee | n 7 _ r — — ft | - = “5 a 4 ine « ‘ ~ “ 4 4 4 " ‘As > i i { U ? ws “< * ie cay .~ > ing - 5 .,)"* £% , - Foes 5 , eo iv ars. Ah ‘> ~ 5 c - s , | ' a Sr PRE AN ee eR Me = a ase Bey PLATE 4. Caulophacus schulzei. Unusual type of pinulus; x 150. Regadrella delicata. Part of upper margin with adjoining lateral body wall. The slender parenchymal diacts of the body wall repre- sented as less numerous than in nature; x 14. Caulophacus schulzei. From a photograph. Upper body and stalk fastened together with pin; }. Euplectella, sp. From a photograph. Lateral wall sliced away to show septa; }. Caulophacus schulzei. Discohexaster; x 250. S oe Small type of discohexaster; x 250. sf . Pinulus from dermal surface; x 150. . sy Uncommon type of pinulus from gastral surface; x 150. ‘ ¢ Pentact ; x 70. 3 4 Pinulus from gastral surface ; x 150. Regadrella delicata. From a photograph. Dermal and gastral surfaces of lower part of body ; dermal surface to the left; gastral surface to the right ; 4. ~ Mi ip ae, H.V.Wilson and Ew.Riibsamen, fec oe. oer ey —_ - aio 7 te eusoe isiw Jouaiotl ooai le seo nig a ea wan) + iF @ AWS x pfpstidos to ‘paver dl ee ernilsieial pails p21 A%lquin x rota + . ‘ i : Loney | 2 4yh. >) 1 € OES x i 19temeodacetl ~~ , , . j Ylraen Jolt t93anxlimedlossih seonigt ¥ nt x ; aimexollonih Yasir lo d ; ites ; reienouliniggd ay 2. Mammen ; “sa OGE x -eslousd ede kaw git ; feo ono bite ovvdiiyo ; Dadepanmttoeg tl —_ * 5 2 ; eeesweleys Jo eolivije d601 os Letlonfin Aleit oy saniiohwd i A+ Falqary be ; : Ayes pariwotie yer Welgniyy. A ; tiedredoseil ealivias aio lte dna) a, = ° - : 1 : is Ue OSS x 4 dese alesiinied otis sailing wedd fo! q ; ope abat Dey abe ieee “ Coe eo wy aed a. =e 1 \Rereeas Ri rw ~ c< yhia to — om Sn a Fan é « ca , Ss we avereh (in “ : Le ' .* Paige ; we pa ; ie ao Yr. 4) F, - are 4 44 | : aA oy . bo PLATE 5. Caulophacus schulzei. Spinose discohexact with some bifid rays (Discohemi- hexaster of Schulze) ; x 250. Discohexaster; a complete ray showing lateral branches ; x 250. e i Discohexaster ; x 250. Spinose discohemihexaster that nearly reaches condition of perfect discohexaster; x 250. Spinose discohemihexaster; centrum and one principal ray with its branches; x 250. Discohexaster; centrum and one complete ray; x 250. Caulophacus, sp. Stalk attached to root spicules of Hyalonema. From a photo- graph; }. Caulophacus schulzei. Discohexaster; a complete ray, showing asymmetrical base on which the terminals rest; x 250. Spinose discohemihexaster ; x 250. 3: 34 Radial section through body showing skeletal arrange- ment. Thickness of section permits only larger spaces to be seen. Gastral surface uppermost; x 20. Bathydorus levis spinosus. From a photograph ; }. fs . * Oxyhexaster; x 300. 2 s i: Stauracts; x 300. HV.Wilson and EwRiibsamen. fec © a ale a ’ mp dontnog Setrom, to ; eters ty ee: Pate 6. Ss . . ' 7 . Ob x coed; nit) t thy ee 0G ae? ptt moyen 3 7 ee at Rie x BYR a - ‘ Y og . ; aes 7 MG) x = olimiqa smna to. ow) old padetizen > oh 3 - : i ee Pes . in rite 1a ¢ : i an , aq “ Deb 2 “my : ti tee fi igsices ae GR 2 Woiv Lin : FPR) OGh.2 sweaty is PLATE 6, Bathydorus levis spinosus. From a photograph; }. “ “ : Autogastral hexact; x 300. Farrea occa claviformis. From a photograph; }. Staurocalyptus, sp. « « Staurocalyptus, sp. Irregular discoctaster; x 300. Paratropal, paratangential rays of prostal pentact; from above; x 100. From a photograph; }. Part of a paratropal ray of prostal pentact; x 400. Discoctaster with two accessory rays; x 300. Oxyhexaster; x 600. Discoctaster; centrum and three rays; x 450. Farrea occa claviformis. Uncinate; the two ends of same spicule ; x 450. Gastral clavula; from above; x 450. Gastral clavula; lateral view; x 450. Gastral clavula; lateral view; x 450. ALBATROSS Ex. 1891 SI IGE H.V.Wilson and Ew. Rubsamen, del. +ddsat wlersooyeh idiw onlovelo. lavteni ae 7 fpalines! Witiw Jogdoey Lamas - 4 S ba i ‘< '? O82 «atoll vun (luda Lager Pr ” Pe Zz ren + z ‘ } (md x : = ~ F 4. * Uas HOY EID iW GOTO “ ' . 1) dae rne OG& » + eRtovelio etkiisdaw at ie - his ods ser rt 1as t aystoye to wer? conlthe i dh 9g moa A Agvite MH -tuods oguage mites Bo ei at Lan 9 iw Saba sirirt on shies ilyin toi ‘ i: to pei anit Hlofdiw-oneln eoatl sid eteeut in imie A’ ew? doagy white pei ue % * i. ai 7 7 ee - pes | ’ »e 2 : , * 4 +e Ae an | y . ‘ _ a a 5 ae oe ; 7. a ae s > ir a i. Wel 2 t ' ~ , j J : 1 ty ' ; y ; j a 4 \ 2” H 4 } i : ’ a at we or > : i OP we Oe a . a ’ 1 : Mir» Ree ae ‘an - % my ety | ¢ gtd Sc arene Linoted ah er . 3 aid ; on otygn sh % f ee jHonaraqesG an : pi ; Sas bea! PF TBlngoos Haute WE y, ox a a eT Tl a ah “1 Sy P ‘ x ; T0tenzotoositl Pate l1 a aN i - - 00 x i rap dl 2 an oo => mae yO ee tee is) Jonbasq Ina Ve ee a Aa ae ox Par ¢(o,5) atoning wilto owe 7 he if f » es j ne bus es16q yoiworle eussd mgs” Sxurv19 =A ee 4 Ses 5 ) I ; Siete toning, . a. SS : : Sod lnvitror otinoa “ ” ¥i 0) ee 7? mptal 6 o1n onsgadsiteat is Te See west) Hobad@ elaano ke | ry a Blsme nk edawsilo betalls ‘J wn = * De gta a i ERE IF ot eos NN Hexactinella labyrinthica. “ “ PuatTe 11. Lateral surface of skeletal plate. Vertical beams in figure are radial to sponge surface; macerated preparation; x 70. Small scopula; x 600. Discohexaster; x 600. Dermal oxydiact; middle region and the two ends; x 600. Dermal pentact (a), and aberrant proximal rays of two other pentacts (0, c); x 170. Dermal membrane, showing pores and supporting pentact rays ; x 100. Section, vertical to surface. Just beneath dermal membrane are a large, and two small, afferent canals. Shaded tissue densely filled with flag- elated chambers and small canals; x 70. ALBATROSS EX. 1891. H.V.Wilson and Ew.Rubsamen, del. ~ . SS . s* - - e 2 , ~ a os » 7 oD : PLATE 12 . oe bo 10. At: 12. 13. Thenea echinata. Sa a 2 PLATE 12. Upper surface; from a photograph; }. From a radial section; upper surface of sponge and adjoin- ing parenchyma; x 20. Spirasters; x 600. Protriaene; x 30. Dichotriaene, modified toward protriaene; x 30. Spicule from root; clavate at one end; x 100. Cladome of dichotriaene; x 30. Somal anatriaene; x 100. From a photograph; from the side; }. Thenea lamelliformis. Anatriaene, somal; x 150. Anatriaene, radical; x 150. Anatriaene, somal; x 40. Upper surface; from a photograph; }. F io Naa T~ 1QgQ7 ALBATROSS Ex. 189] i ANT ) Np i iy H.V.Wilson and Ew. Rubsamen, fec 13 o My STs a i ; ee | Ry “sibs feats NY tet 3° ” PT) pouen ere a et Se Ah ae say ‘ fit I LBeMETE Dc yaa Nt 2 a “orton ster LS SHOP AD syrvsenret\ ania 4 Sitiborts ssuasintosloit | j “# z ’ : ‘ : ee, : F Seats GOTT pilqergodoily’ rH 107 ae ANVOLI AL, enaaee ont ; Aqatygotordq: sfmor >. tihod WY) ainda, sendy BBD x » ensloaoroint Inn glories MATIC Vy Hasta VT a Mody ade miflee. 0 MO ttoy Seton Waset att ae yndode ; SO SO gt Se OU re Gels RO vies ey ey ete FweP RS PLATE 13. Thenea lamelliformis. Spirasters; x 600. Thenea fenestruta O. Schm. From a photograph ; from above; }. a e - Spirasters; x 600. . . Re Protriaene; x 100. Thenea pyriformis. From a photograph; from the side; }. Thenea fenestrata O. Sch. Protriaene; x 100. x x * Dichotriaene, modified toward protriaene ; x 100. Thenea pyriformis. From a photograph; from above; }. Thenea fenestrata O. Schm. From a photograph; from the side; 4. Thenea pyriformis. Parenchymal microscleres ; x 600. : %: Microscleres of dermal membrane; x 600. Poecillastra tricornis. Annulated microxea; x 150. S z Triaene with reduced rhabdome; from the side; x 30. “ « Microxea, nearly smooth; x 400. ALBATROSS Ex. 189]. H.Y.Wilson and EwRubsamen fec. Pes a D adidein: Bealiue 112; ? Petwarig feat eds nsing MY , testo: Lamy Febive0 yittwode -oiioen lasitiey vtuibsr « : My ei fis. & Yo Llaidle wo doidw to modk vais ous: teqa bes saewdarsa isluseo woloth oo “08 x slaseo od eaowe ebnotxe uniqee DF : gaivgils oie St ices beset Ry LIAISE ples 74v & Io Motiad whiarbd>,onetidtern sol ps ong oils Bint eo odit Tes: hod Pou Vai By rSdes-ii 18 i" biirita be te ; . - “ nas EM: 49 Soeth i 2) opt 1 lane’) Pinan Peon: iit G burnt inl bd ie ea! CT ~~ iy Silo fitint strtreda linnte Ss vdvhbaime? yitt te: witdeb ip aodsaiies «21. dotiby-jo ‘ “ 0 i Muiaa1 a dw worts srod 4: (hecarn emo ‘ a - ts yaks 173 Siw Gio wedtmdnodianigseo os’ Aa rita sun [ity 2 Be i MoWdorngos] ontine, PE GOT S19076 ~ ee gil woe ei Fs: fy Ira¥ it; LEY ~ a dy us) . : wns & 8 ts Mae TOF Lies Wythe A Widistxe Bont ne 20 (Sh w of fine) BAAOTS ci, lige caodosscad fo Bey te Li a e AAR k oie » fp edge Peotdd “a Howl & ooetins vale ipvods MOF) 8oOon. 7 iIPhiLP _- (ye (vee Je. > Mer TY Hoa ai Ue. ws Bia ieteyh le etree : Hleeed ete eshte bak | OOD Y - pore? Share tae i Z = e : ao * A - a o« i 1 J 7 “P » "Mea ae a . - “? \ * ~ si A ae > bil * ae reer | giwdTaqA BH) algpordd woizesiqgel ‘welvse0 og uti of ‘- < PLATE 14. Poecillastra tricornis. From a section tangential to surface, showing flagellated chambers with Sollas’s membrane; c¢. w., canal wall; x 250. Streptasters ; a, spiraster from dermal membrane; ce, paren- chymal metaster; 6, d, parenchymal spirasters; x 600. From a radial, vertical section, showing oscular surface, with an oscular depression, through the aperture at the bottom of which a shell of a Foraminifera is partially protruded. Below oscular membrane and aperture an incomplete septum extends across the canal; x 30. From a radial, vertical section, showing pore surface. Pore membrane, forming bottom of a very shallow con- cavity, is produced into a plug-like process. Axis of this is marked by dense streak of spirasters, representing the closed pore. Side wall of canal, into which pore opens, is shown in perspective and exhibits apertures leading into branch canals. Canal wall bears some incomplete septa. Lowest and thickest septum is per- forated by a small aperture, in immediate neighborhood of which is a collection of débris, largely shells of Foraminifera; x 30. Pore surface. One pore area with nearly closed pore shown. Surface preparation; x 30. Oscular surface. Three oscular:membranes, one with open aperture, shown. Surface preparation; x 30. From a radial, vertical section, showing oscular surface. A single oscular depression, floor of which constitutes the oscular membrane, is shown. Osculum itself closed. Canal to which osculum belongs exhibits on its lateral wall apertures of branches, and is crossed by perforated septa; x 30. . ’ Oscular surface; from a photograph; }. Poecillastra cribraria, Cladome of triaene, from above; two rays branched; x 70. ‘ Characteristic triaene, from the side; x 70. Microxea; x 250. - sp Streptasters: a, spiraster from dermal membrane; 2, plesiaster from parenchyma; c¢, d, plesiaster-metasters from parenchyma; x 600. 1b ' | | | wn N f j , <-* yt ~ X ‘ . za. wi * ee el ax Ly < ; H.V.Wilson and EwRubsamen, fec worl oiioos laeriey ilnibert ec arvrr'l ae AgNO Michaud eisiv siuyionera' : esi ," w Hi] ie a Wolew see. .notdose «£ osor'l cubiived yop yndintayly! bere roe | ; J arent halons ylonit Uo a! zai 99. oul) Jeon 2 uO = .eotiabawod-lis Bh gating’ e ,wlloo asllos sil? Yo (515 "000! x :euplogin o guivolone dopo opean 7 petra Medion tee odie welyoeD oy ‘ *% Te $ ret § mott een sit ms eg #4 mad} :) Orrainis dd urs Pilati, ‘erat ple) Pe eeratr dss agen py. mi yon savtebiis' ast, anal Toe ov he Wua oR ena 1 sé ownlt Mpokiy Ges) ais PLATE 15. in, a % LOL Se bebulou hue a! Apgiexs wees § ied Pie eM Aeedyst oni! Dein j yrds vt Mah? 3 1) don Sy Reina 3 wuitdiud suid Egat sit? wma? @otigu alece dnvretiv Heo a dxour a ai sth ceed © oriyt oft to wens t |; ne9q velomovansn wd suroe .slinine “‘tolt) Jauira Wo. ;alanso doviolie neds Io a 4. ; Abe ApG'S BOSUs hey 8 ator) s sont th , OCS

~~ gis } a3, ae mia whit ObrUA Re ie Adio Biciedroroi ssnoaisd os v fey f Wee firs hel lit yisareh ui sna! if oid soo gious an lew en aomohalo 08 RU Meimsqmg sonlhwe = j-bettiio her 9 } more + i. oe 10. 11. Poeec illastra cribraria. Penares foliaformis. “ “ PLATE 15. From a radial vertical section, showing oscular surface. Parenchyma with abundant conspicuous granular cells ; x 100. From a section. Posterior wall of flagellated chamber showing several chamber-pores. Boundary membrane of chamber is a thin, finely granular membrane show- ing no cell-boundaries. On it rest the cell-bodies (wide apart) of the collar cells, appearing as dense angular masses, each enclosing a nucleus; x 1000. Oscular surface ; surface preparation; x 30. Pore surface ; *t ss x 30. From a section. A flagellated chamber opens to the right into its efferent canaliculus ; afferent canaliculus abutting against boundary membrane of chamber, to left; Sollas’s membrane in perspective. Two other flagellated chambers are cut. In right lower corner, a part of an ovum (ov.) with part of its nucleus is included; x 1000. From a section vertical to surface of sponge, showing choanosome with a number of small efferent canals cut lengthwise; some canals leading in the section to the flagellated chambers from which they start. The small efferent canals unite to form the two larger canals of the figure. These unite in the next section. Other canals, some cut transversely, appear; some of them, afferent canals; ¢.w., canal wall; x 70. Surface; from a photograph ; }. Microrhabds ; x 250. Vertical section showing ectosome and adjoining choano- some. Two radial pore canals open, each into a sub- dermal chamber; x 30. Oxyasters ; x 600. Surface, showing arrangement of pores and cladomes of the triaenes. Microrhabds with which dermal mem- brane is densely filled, and which extend over the cladomes as well as throughout the pore areas, are omitted. Surface preparation; x 30. a Geek . oe Pe a Vago i , a Ae \2as ’ H.Y.Wilson and EwRiibsamen, del. 9 oe View lianso wo ; oitdoe o nt0TT, ortirvirw Misses faa old od abort t1 .alnoodat wid? pm eviquose tail god el Jales suH19Ro Oink ote TEgR oli Ih guise a yd -baimoptieg «i Mol eit Of Sundin anod neewind baetxe efleoralloD .ovog sodmary g- iptial “Sis pened toad Maxiio# brn gre moot ~ WOOL 4°: s¥ieoqereq bas, sorsooe i '< ears | Ohie aiden? ;dqergatodg 2 .07f .nidonnsnst nrhansmgio’l © S le : f ;oontive walgoeo ; dqetgotaly 2 mot = sind steno allio sglvoso to flaw igudgd? soitose petereasyP braspnt bsomylok | A : nla Ty pe on ovodA. iause Inixn Yo !law . » pede ae a owl — PLare 16. — AN eon Oy ye by OT x jduo 918 & bad baie . + OT a ;40lnniq? » ; hy a mm fenit19% aarasa “ 2 a. ines ost wotmiedialen Daal il nyse : edly ! 990)! est eater fleare. to plkol Hoste sn alts ido 7 me ite b duns tea Aasgqs aolh nd ye Gar Spite” oollsredissotg peohysonr Rafer 8 thyvords ; bine ne eeane e104, difw eeosqe } gi OE % ueodiorg yrallimearsm to » @, Iabitieynoitoss Lanibutigaod stanon\ugaanesbodolon Dien orld o spnoqe to eostwA .opuoqe bole Sve os ba jexaoqqs SS elband Intolode le nibusipaol ames — s a 3 patiuse ews pualegor! Ipisuee fit ow fen ae? Li ‘ yabia os mont sigernotodg & mor — . Ci ae ve ey ts ie e: 4 oo. 3 gird “ PLATE 16. = 1. Poecillastra cribraria. From asection; ¢.w.,canal wall. A flagellated chamber occupies a thin trabecula. It opens, to the right, by a wide aperture into efferent canal. Its boundary membrane, to the left, is perforated by a single chamber pore. Collar cells extend between boundary membrane and Sollas’s membrane; the latter shown in section and perspective; x 1000. 2. Polymastia maeandria. From a photograph; from the side; }. oe Poecillastra cribraria. From a photograph; oscular surface; }. 4. Polymastia maeandria, Transverse section through wall of oscular papilla; c. w., wall of axial canal. Above are the surface brushes of small tylotes. Two longitudinal spicular bundles are cut; x 70. 5. = - Spicules; x 70. 6. * = Section vertical to surface of sponge and through a 7. Pachychalina acapulcensis. mammillary protuberance. Beneath superficial layer of small spicules lies layer of more or less tan- gentially arranged tylostyles. Five radial skeletal bundles appear. Main afferent canal in mammillary protuberance connects with a larger internal canal through a chone-like structure. Small subdermal spaces with pore canals are seen, especially in region of mammillary protuberance; x 10. Longitudinal section, vertical to surface, of mace- rated sponge. Surface of sponge to the right. A longitudinal skeletal bundle, 7. b., appears; skeletal network in general irregular; s. v., surface villi; x 20. From a photograph; from the side; }. H.Y.Wilson and EwRubsamen, del * a oe A Pe ii agrgie d}is ano getodifiado hataiiougk : ed a . eriennih .davciilt otto fs Per Melon oui wohivily v sone ; OOr "(ieee sf ' ; ' f teroodtorla “a= 2 i r ae : “ rf a ¥ oe ’ ul ; ree ty f SB ol apiqa te offitued Jer: tile 7 e' Pik pHoUss ayo . at on Ba wwontes wt lwstix0¢ \aoitov [riagxiiols PA avods oenors 1 bonti4 BtT Ot ~~, 5 , ‘ ; ok Relieves 9 of ahs) olbnud ltelod , ve 7 j ee Sain vioviderpe: foo Sewis es 4 be a sme} Of x ’ BOE x ; o1dR Liao! I OO & 1 oyitoonMs Hane ATE 17. Con » { t] , j bd ? - rior} tT 1 ; oO. Hh Bort) ; Wo” at 46/5 JONTeRe ofa! a r 4 r 4 VHIEZO LRIOLODE Lit ncamitatino cert. - ail ° eT Oey & moi 1 : : J Ae ee ibirdrodet ae ” ~ YY , : \ enone ‘ - ) Q i : By)" 4 ) SMTP ARIK \y ARE 10 pad mr) 7c i “4935005 el it e* OS AIC WD ) DWI * ll salad Suree so 90 10. 1, 12. 13. PLATE 17, Pachychalina acapulcensis. Dermal membrane. Pores and dermal reticulum with villi. Surface preparation; x 70. Horizontal section, vertical to surface, of macerated sponge. Surface of sponge above. A longitudinal skeletal bundle, /. }., is cut transversely. Skeletal network comparatively regular; s. v., surface villi; x 20. ° * = Large radial skeletal fibre; x 300. : * Small connective; x 300. . - One of the finest fibres, subdividing a skeletal mesh ; x 300. Petrosia variabilis crassa, Oxeas, with spongin; x 100. Petrosia similis densissima. From a photograph; }. Oceanapia bacillifera. Surface of fistula; x 100. Petrosia variabilis crassa. From a photograph; 4. Petrosia similis densissima. Oxea and flagellated chambers, one of which opens into efferent canal; x 250. Gellius perforatus. End of skeletal oxea, with sigmata; x 450. Petrosia variabilis crassa. From a photograph; }. Pachychalina acapulcensis. Longitudinal section through macerated conulus. An axial bundle of spiculo-fibre gives off, and ultimately divides into, radial fibres; s. v., surface villi; x 20. eel. 7 @ A — eo — Roary “ton Peed 1 pw eg e yy). ft eit gp oieaaie qa mow ;xalotel “HWS ES Khel) OSE x. ; ;eeluoig& SY Wind Manno muritewvioni shi uve 3h : CH aaa to [Lew dhgrsonsts noijose Initndgosl ame\Piens vigommsO | 4 »* - SitOcy & eolagiqa latelasdte 163 MPAIS RIT Bxae ak stoi iow Saust -Piate 18. ecpuliive allie oi datoay bs eae I pid are eau ft 2fRole sai» panes) wit 4 > ) : r tr j Airey 4 e yRity git F ~ at : A Sage diqurmadgiy a 107" » , a piidde At soghue reggae Vials 02 ‘ geen OR Peiaolwannell arerewiblas® aac bowl ra - = oo ORh x; PI nus gic - “ & & : av q Aatoios3 Yo - vorhive eas aregblonad aaledeWd cor * ofl 788 ni Tats: 199K Mee i= | . Z 7 See in ee 10. PuaTE 18. Gellius perforatus. Upper surface; from a photograph; }. Oceanapia bacillifera. Fistula; from a photograph; }. Hymeraphia, sp. incrusting Oceanapia bacillifera. Spicules; x 450. Oceanapia bacillifera. Tangential section through wall of fistula, showing arrangement of skeletal spicules and spongin. Meshes in the skeletal framework, which in the figure are represented as vacant, are filled with parenchyma; x 70. Tylodesma alba. Megascleres; x 70. ; sf Sigmata; x 450. x ‘ From a photograph. Sponge viewed obliquely from the side, so that upper surface is seen; }. Tylodesma vestibularis. Megascleres; x 70. a ;: Sigmata; x 450. Phakellia lamelligera. Outer surface of skeletal framework. Piece has been macerated in potash. From a photograph; x 5. aA L fh 6 6 \ Sa SAS ASA 77 ZN S N Nw 44 / QV SADA vif Vass! / fm \ Ss ve > ae ei &! ra oy a eg te meme... oD i ae AO Sty a me ee Ae “Y P . Outs Aven at Mr reipvaemtijtei ol ea 2 SoM aO tase “tot a ie - : ' ; i = : Wy: ain | . >» . oil: gioit yloupil ido a9 ‘ ——. - ae ie > = eo 7] Yi Reseed 011!) ee : tit) EINE eke . _ PuatTE 19. op 1933 : we ve r a ff ie Tiny ‘ ah wdtes (! ' 7 ¥ bu { maris si AD . mf bo waty is Rs auth ogid’» ;abie 99 fu thee Poiogi % ; abie mayne mK we , ‘ i : r glanpiido Javenor no02 og “we a} ~ Ww Tylodesma vestibularis. Phakellia lamelligera. “ <3 Auletta dendrophora. Lophon indentatus. Auletta dendrophora. PLATE 19. From a photograph. Sponge viewed from the side; whitish conglomerate below. Vestibular spaces ap- pear as darker areas. The one best brought out in the figure is well to the left, and has indented mar- gins. The small darker area at extreme left of this space is an osculum ; x 4. Spicules; x 70. From a photograph. Sponge seen obliquely from the side; xX 75. Inner face of sponge wall, lining paragastrie cavity. Apertures of radiating canals appear; /. b., longitu- dinal skeletal bundles; x 100. From a transverse section through a“ person.” Surface of sponge above. A radiating canal opens into a subdermal cavity, s. c. Part of the wall of the latter is seen in perspective. Wall of another cavity in lower left corner seen in perspective. On each side of the radiating canal, section strikes a radiating bundle of spicules (styles); x 100. Microscleres; a, anisochela in lateral view; 6, bipocillus from concave side; ¢, bipocillus, apical view of larger end; d, bipocillus from convex side; e, bipocillus in lateral view ; x 1000. From a photograph; sponge seen somewhat obliquely from the side; x 4. ( e a A i ;.: a °c : Le | > * ttt moo mov? arilioogid nlodooaina > ; eh 1 it : i = - ’ ‘ t ; Puate 20. . ree} | 2 1) : 1 e.4.F8 e ‘ 4 : + fs TH ; 4 pit r y f f eo ; ol : 4 ne | Z ; r “J Soreintes HUM MEMS MONO) : . Pepostrte tobas ; dqergotodd 2 mao avin oo wiis sogok ht as sv? ii}} rat x : ri! ‘ : ; ii . J ; y > \) } fir hia "3 ’ rs i ' | p> f , Z - . nif ae aa ay . a io ia a hata!) Pe ; sf | ‘Obie oveonos mot enifioogid , Ea teiv Isistcl ‘nt alodoonin Ly Wsiv Exiaisl onpildo. yltiiyi L eiioog id y ol 1OyIBL. 202 ys rf ‘ipl rf - {tiw bebukut i tre: (8 : li ‘ tel div SILO iH) “AA : PLATE 20. "tie ui plodoostna A p06? ror { Pha 20. sj 5 OOOL 63 Ver! “7 & » : city ren abies aoe 8 ~ < ‘ aqid ne ih Tie Ly 3) 9 em ee Pm Lye sieg as Pr tts tis! e DORR My iit B0° voi taigs: cul . arioonye | a0 ple TIA, | ere yode dG pl ae ed rs 2 Pah as 6a § 1 Aor > ADO 16. PLATE 20. Iophon indentatus. From a photograph; x }. Iophon chelifer ostia-magna. Microscleres; a, bipocillus from concave side and smaller end; 0, anisochela in lateral view; c, bipocillus in slightly oblique lateral view; d, bipocillus from concave side, larger plate in optical section ; x 1000. Iophon lamella. From a photograph. Oscula depressed and surrounded with collenchyma; x }. Iophon chelifer ostia-magna. Froma photograph. Margin of sponge, with large oscula, to the right; x }. Tophon indentatus. Skeletal style; x 250. a r Tylote; x 250. Iophon lamella. Microscleres; a, bipocillus from convex side; 6, anisochela in lateral view ; ¢, bipocillus in lateral view; x 1000. “ _ Skeletal style; x 250. = .. Tylote; x 250. Tophon chelifer ostia-magna. Skeletal style; x 250. : 4 a Tylote ; x 250. Iophon lamella. From a photograph; upper surface; x }. « if9 From a photograph. Lower surface of the specimen shown in Fig. 12; x4. Lophon lamella indivisus. From a photograph ; under surface; x }. “ “ “ Microscleres; a, bipocillus from convex side; 4, anisochela in ventral view; ¢, bipocillus from concave side; d, bipocillus in lateral view; e, bipocillus, apical view of larger end; f, bipo- cillus, apical view of smaller end; x 1000. is . From a photograph. Upper surface of specimen shown in Fig. 14; x }. oO , = yy. = i —— Segsmouainieletcladte eniwiile emvidgrnsgyry Lowintersy ois 3 | me ' tae j oy ; ) ‘ ‘ is ime nay ea vein . at . » 7 ae he VIN ate . tenvrotl : 4 . re, . ‘ =. atta ti a Sl lle inde PuatTE 21. espe of thick unstained preparations, shee skel Fig. 1. Polymastia maeandria. pau surface; x 50. — “« 2. Petrosia variabilis crassa. JIS atxat * 20. Lee.) ~ ss “ Section vertical to surface, 1 upper margin of figure; x oie “4. Petrosia similis densissima. Section vertical to surface, v upper margin of figure; = ly a g i, = Dermal surface; x 20. “6. Gellius perforatus. Dermal surface; x 25. ‘" Vil ¢ “ : ar 3 BS Duta be str ae - 4 ‘i : re * rd A hatnazenyst | et M cadatruz Of-lpoiiiey wih fray aq al sont tire lye! MAL) .gtusil to wisp “rik tl ihe ¥° gone. - PLare 22. uf wits ; dédbi d ’ 5 thtawinrs, a spp lua ee fo igen ses Frias? Sie! onity ( ew (-,08 eines ‘ wat ‘oa Vv hOMU 4 r rad bataomrge ai iloid w ose OF Insitiey corioue H BpRge eel idecsy A. .enomit 2o wasps ashen sigent ai notolode Dilber d os 5 work 5 KONDHE ort to ying ne ; %, i nen ee 7. 4 : rie ne le ea) ; ee fern Jorolite rintoits Bode bouineanp faseanqo% en iiotigramaiiue 6) Tridisy coljese Dai doit Sea Lspvia (rol Fs PLATE 22. Microphotographs of thick unstained preparations showing skeletal arrangement. Fig. 1. Gellius perforatus. Section vertical to surface, which is represented by right margin of figure; x 15. “ 2. Tylodesma alba. Section vertical to surface, which is represented by right margin of figure. Dermal surface in upper part of figure is roughened with projecting spicular tufts, elsewhere smooth; x 15. . bores a‘ Section vertical to surface, which is represented by right margin of figure. Surface is smooth, although it scarcely appears so, owing to the fact that it is partially seen in perspective ; x 15. ce 4. Tylodesma vestibularis. Section vertical to surface, which is represented by right margin of figure. A vestibular space is cut across. The hexactinellid skeleton is conspicuous in the deeper part of the section; x 15. e, photogr ht en. yn and Z 10) } i i.V.Wi l i a as BY 2 ata l . Pia to ou bash Sh) Ikoiltaey jojtes% summ\nliimn as ee 2 sige: Fiyin, yd hetuse . oh i ete ATE). Tae i alien Siaiid us9cs Inario & . - ade rer ES ddgit req 5 ws s 7 os ~ . ns us ‘ : al zx B yainiophs hiss sul gtizvves otimiditent Leyriotl OE x ; ytives teludiveoy beste : 4 #eobnloti uolosk ©. tnyit to niwract — ee Mol ils 03) cixe vinoywd Bintoloz: yi iba anietqoiy boarbatir Aoidt toes loro 38 tot el ‘daisiw eontuwe 02 Ipaitiey v0 MASE wisn wtols (=i... @ i of Gah ee 1 1 ee. PuaTE 23. Microphotographs of thick unstained preparations showin Fig. 1. Tylodesma vestibularis. Section vertical to s “ 9 ¢ “ “ . “ By 114 “ “4. TIophon indentatus. margin of figure. ion includ fe Gorgonia axis (to the left) ; ee Pais hoa —— Y ALBATROSS Ex. 189] H.Y. Wilson and Zehle, photogr 3 in Dania ob Me Sraci we id PUG, asic ys oo Huy aa: sairgi Yo > ei iguanas fal ea ne ohed PLATE 24. Microphotographs of thick unstained preparations showing skeletal arrangement. Fig. 1. JIophon chelifer ostia-magna. From a section vertical to surface and radial to margin of sponge, and passing through a large marginal osculum and efferent canal (ef. Fig. 4, Pl. 20). Figure includes thickness of sponge lying between surface and canal. Right margin of tigure represents surface of sponge; left margin, canal wall; x 15. “ce 2. Iophon lamella. Surface; reticulate condition; x 15. MS A Section vertical to surface, and through entire thickness of lamellate body. Right and left margins of oo repre- sent surfaces of body; x 15. A: . fl Idem ; x 15. @ co E 4 —1./% ALBATROSS H.V. Wilson and Zehle, photogr : aotaet Meg 09. gotenne pe Pat se Sih ot fnonay udites= ye oeusld + Inosiiod itt [BW othllonial dquo ptf fronee Taathiudivao! wether s worl to anstane flasasiqor Simi -PLate: 25, y “yt RO. giiternttay bh sorte teotam 7 A it +6 Jake Apsibutiznol Pp Siasqys ittwian: sstis! f aoe _ ies, om 2 x <. , aUG ehiis has puis midoe trois gory ior S Baio vai iiort «tse willomat clatolod2 anigsnct Yel bas adpin pools evlosiga gaitostorg ) 101 “paitosjory xolvoiqe ot xa ousgit nA 6 braav0 Detosesih : abw silemiml « .oginn ae eaters aoa Now jon eaw ional asobl vibat 1s aot lunige (Tat bite tl 1) RUIIO4% PLATE 25, Microphotographs of thick unstained preparations showing skeletal arrangement. Fig. 1. Phakellia lamelligera. Section vertical to surface, and passing completely through lamellate wall; in horizontal plane of sponge; x15. . “ 2. Auletta dendrophora. From a median longitudinal section through a “ person.” Left margin of figure represents surface of sponge; right margin, surface of paragastric cavity. Along latter margin appears a longitudinal skeletal fibre, from which short echinating spicules project; x 30. if 3. Phakellia lamelligera. Skeletal lamella, seen from its flattened face. Long projecting spicules along right and left margins of figure are the spicules projecting from surfaces of sponge. Lamella was dissected out and cleaned with potash; x 10. he ta @ - Idem. Figure was not well reproduced, and the pro- jecting (prostal) spicules are indistinct; x 10. 3 otoar ehle pr 7 mn and Z Wile WiiSO H.Y.| . > Aa oe) Sa = al ad - e& “lye ee 4 i La 7 * \ d 4 e aM ™e _ fe Ne ae » ee ae Aste s goitiont ot Ae ool oldqergorhyit adore Seaodsd Ld, ‘e quld 26 stony? ads iin val . ae ae n E = o 3 a) a. S HX 3 =) fis| ~ wi t y > Pande = 38°. SPER ge Sis see . we os Ti oll OX ee fathoms Memoirs of the Museum of Comparative Zoilogy AT HARVARD COLLEGE. Vor ea ks No; 2. REPORTS ON AN EXPLORATION OFF THE WEST COASTS OF MEXICO, CENTRAL AND SOUTH AMERICA, AND OFF THE GALAPAGOS ISLANDS, IN CHARGE OF ALEXANDER AGASSIZ, BY THE U. S. FISH COMMISSION STEAMER ‘ ALBATROSS,” DURING 1891, LIEUT. COMMANDER Z. L. TANNER, U. S. N., COMMANDING. XXXII. THE RADIATING ORGANS OF THE DEEP SEA FISHES. By ROBERT VON LENDENFELD. WITH AN APPENDIX ON THE STRUCTURE OF THE BUD-LIKE ORGANS OF MALTHOPSIS SPINULOSA GARMAN. By EMANUEL TROJAN. WITH ELEVEN PLATES, AND A CHART OF THE ROUTE. [Published by Permission of MarsHati McDonatp and Grorce M. Bowers, U. S. Commissioners of Fish and Fisheries. ]} CAMBRIDGE, U.S. A.: Printed for the Museum, Avueust, 1905. olay ¥ : peat tie wee tate rg Hiss ba, Ral oe iy estedhes hie ea ys ee y ; bed hs SAA aaa" pa Teh, ¥E . . | atl as Wa a8 ee a oy . } ri ns el; ' . . te =f x ane ; , : ‘ + e cy ae in? fy ee ae . eR: fh ok = , >: mg = rn oy Wed FD. FAH asl ; we hg oul ae AIAG et TABLE OF CONTENTS. PAGE PAGE Pete RODUCTION ... « ... « 169 (2) THe Fisnes witnh Rapratine 2. DESCRIPTION OF THE RADI- Sein ele leds ATING ORGANS ..... 170 Ceieeseai wiiads (el 4, es 14.17) 197 Leucicorus lusciosus (Pl. 3, Figs. (1) Tue FisHes with OcELLAR Rapi- WORDS) ioe og 8 198 ATING ORGANS . . . 170 Halosaurus radiatus (PI. 1, Figs. Argyropelecus lychnus ve 6, | Figs. SA 0) 8 se 8 199 io) Paes I) Macrurus canus (PI. 5, Figs. 18-20) 202 Sternoptyx obscura (Pl. 6, “Fins. Ipnops agassizti (Pl. 1, Figs. 3-5) 202 26,27). . . - WS (3) Tue Fisnes wits Tusurar Ra- Lychnopoles Pretrentas. a ae. DIATING ORGANS: . 203 ee) o. - a Malthopsis spinulosa (PL 1, Figs. Cyclothone acclinidens (PI. 6, Figs. 1,2; Pl. 2, Figs. 6-9) 203 Popeye. Le 180 2 hte. Myctophum pe eleieratten vr 5 8. CONCLUSIONS .. . . . . 205 Figs. 21-23) . . . ao2 | AEP LERATHWRE, ©... ss s- 207 Chauliodus barbatus (Pl. 8, Figs. 38-40; Pl. 9, Figs. 43-46) . . 183 APPENDIX. Idiacanthus antrostomus (Pl. 8, THe STRUCTURE OF THE BUD-LIKE Figs. 41, 42) . ae 191 OreGans oF MALtHopsis sPINU- Stomias eats ‘aah 10, Figs LosaA. By E. Trogan . . . 209 i 38) eee ae eke ieee - 192 POVERAVURG GS) see eo a” a AIS gi ala < 4 eis mate ~ Kae “it f i THE RADIATING ORGANS OF THE DEEP SEA FISHES. 1. INTRODUCTION. Tue fishes collected by the “Albatross” in 1891 have been carefully described by Mr. Samuel Garman (99). A considerable number of them possess radiating, or, as they were formerly termed, phosphorescent organs. Mr. Alexander Agassiz was so kind as to place many of these at my disposal and to permit me to study the minute structure of their radiating parts. For this valuable material and his subsequent furtherance of the work, I express my sincerest thanks. The results of the examination of these organs are detailed in this paper. The fishes sent to me belong to the following species : — Argyropelecus lychnus Garman. Lychnopoles argenteolus Garman. Bassozetus nasus Garman. Macrurus canus Garman. Chauliodus barbatus Garman. Malthopsis spinulosa Garman. Cyclothone acclinidens Garman. Mixonus caudalis Garman. Halosaurus radiatus Garman. Myctophum aurolaternatum Garman. Idiacanthus antrostomus Gilbert. Sternoptyx obscura Garman. Ipnops agassizii Garman. Stomias hexagonatus Garman. Leuciocorus lusciosus Garman. All, with the exception of Mixonus caudalis, were sufficiently preserved in alcohol to allow of the examination of their radiating organs. Since Garman did not pay particular attention to these organs, his state- ments concerning their arrangement and number are not always sufficient. This applies to the figures even more than to the descriptions. For this reason the entire fishes are here figured again. The minute structure of the organs was studied by means of paraffin and celloidin sections. Various stains were employed. Van Gieson’s haematoxylin-picric acid-fuchsin gave the best results. Most of the sections 170 THE RADIATING ORGANS OF THE DEEP SEA FISHES. photographs, and drawings were made by Mr. EK. Trojan while work- ing in my laboratory. I must express my thanks for the great pains Mr. Trojan has taken with this work; and for his paper on the structure of the bud-like organs of Malthopsis spinulosa which I add as an appendix to this memoir. Three types of radiating organs are represented by the fishes ex- amined : — ocellar radiating organs, radiating discs, and tubular radiating organs. 2. DESCRIPTION OF THE RADIATING ORGANS. d) THE FISHES WITH OCELLAR RADIATING ORGANS. Argyropelecus lychnus Garman. Plate 6, Figs. 24, 25. This species has been described by Garman (’99, p. 234, Plate J, Fig. 1, 1b). The radiating organs are of considerable size, possess exceedingly brilliant reflectors and are surrounded by a layer of dark pigment, so that they are very conspicuous. There are on each side 1 anteorbital (Plate 6, Fig. 25, ao), 1 postorbital (po), 2 opercular (op), 6 branchiostegal (br), 6 pectoral (pe), 12 ventrothoracic (vt), 2 anterior lateral (al), 6 medial lateral (ml), 4 ventromedial (ve), 6 anal (a), and 4 ventrocaudal (vc), radiating organs. Besides these there also exists, according to Garman ('99, p. 234), a frontal. The axis of all the organs lies very obliquely and nearly parallel to the surface in such a way that it is directed vertically downwards. The anteor- bitals, the postorbitals, and the opercular are simple. ‘The others are joined in longitudinal rows in which they lie side by side. The right and left. ventrothoracic organs form a double row which appears as a medial crest, protruding ventrally. All the organs with the exception of the anteorbitals, in which the sheaths are absent, are similar in structure, and size, and the differences in their appearance are due chiefly to the different degrees of coalescence. This is ni in the three cephalic organs mentioned, and greatest in the ventrothoracics. The organs are 2 to 3 mm. long and 1 mm. ora little more broad. Each radiating organ (Fig. 24) is composed of a slender elongated ellipsoidal proximal (A), and a narrow funnel-shaped distal part (D). These parts are : { 4 j ; THE RADIATING ORGANS OF THE DEEP SEA FISHES. 171 separated by a clearly marked constriction (C). The proximal oval part has in the single organs a circular transverse section throughout; in the joined organs it is not so regular. The distal funnel shaped part is in life probably a rotation paraboloid, but on account of shrinkage appears somewhat irreg- ular in the sections (Fig. 24). In accordance with the smallness of the angle enclosed between the axis of the organ and the surface of the fish, the funnel abuts very obliquely on the latter; it is long on the inner, and very short on the outer face, a very oblique truncated cone with bulging parabolic sides. The whole organ with the exception of the funnel mouth, which is closed by a transparent membrane, is surrounded by a thin layer of cells containing a dark brown pigment. These form an outer pigment sheath (p). Inside this lies another sheath composed of colorless, highly refractive fibres. This (r) is exceedingly thick at the proximal end of the inner, oval part of the organ and gradually thins out from here to the margin of the funnel. It is most highly developed in the joined ventrothoracic organs, where it forms a continuous mass. The fibres composing this layer are parallel, pretty stout, lie close together, and extend longitudinally from the proximal apex of the elliptical part to the distal margin of the funnel. In life these fibres probably follow the curvature of the layer which they form, but are not otherwise bent. In sections (Fig. 24) they often form zigzag lines in the distal thin part of the reflecting layer (B). This also is probably attributable to shrinkage. In the tissue occupying the proximal and central part of the interior space surrounded by the fibrous layer three regions can be distinguished : an inner (g), a middle (1), and an outer (s). The limits between these regions are well defined. The inner region extends to the constriction between the oval and conic part of the organ and there ends with a slightly convex sur- face vertical to the axis. In the joined organs, particularly in the ventro- thoracics, the inner regions of adjacent organs coalesce to form extensive cellular masses surrounded by common fibre and pigment sheaths. The tissue occupying the inner region (g) has in all sections, whatever their direction may be, an alveolar or reticulate appearance ; in the alveoles or meshes residual masses of protoplasm and small spherical nuclei, both readily staining with haematoxylin, are clearly visible. Very rarely a slight indication of an arrangement of the meshes of this network in rows is met with. The meshes or alveoles are about 20 » wide. Occasionally one sees 172 THE RADIATING ORGANS OF THE DEEP SEA FISHES. a capillary vessel traversing this tissue but these appear to be rare. The residual protoplasmatic masses and the nuclei appear attached to the strands of the network, or the walls of the alveoles. The structure of this tissue is pretty much the same throughout the whole region it occupies. The middle region (1) occupies the lower portion of the conic part of the organ. Its limit towards the inner region is, as we have seen, concave and smooth. ‘The outer limit is not well defined, the middle region gradually passing into the outer, which is composed of small digitate papillae lying close together. The middle region is composed of cells, which, in longi- tudinal, axial sections (Plate 6, Fig. 24) appear transversely arranged and mostly long, slender, and spindle-shaped. On the whole this tissue takes. stains, particularly haematoxylin, somewhat less readily than the parts below and above it. It is traversed by numerous capillary vessels. Whether the cells here in the proximal part of the middle region are really spindle-shaped or lamellar, overlymg each other like the cells in the epithelium of the human skin, could not be decided. Distally, towards the papillae, the arrangement of these cells becomes less regular and the cells themselves become smaller, the nuclei lying much closer together. The papillae forming the outer region are composed of similar cells; these are often much curved, but also on the whole arranged paratangentially. This outer region is poorer in blood vessels than the inner. The distal part of the cone beyond the outer region contains only a few doubtful traces of cells and chiefly consists of a gelatinous substance, which takes haematoxylin and some other stains pretty readily and which in some sections exhibits a very strongly marked stratification, the layers being slightly oblique to the outer surface. The structure of the radiating organs in the genus Argyropelecus was first examined in A. hemigymnus by Ussow (79, pp. 103-104, Plate 1, Fig. 5). According to his description these organs of A. hemigymmus are similar in appearance and arrangement to those of A. lychnus described above. Ussow states that the internal tissue of the inner region is composed of glandular cells supported by a scaffolding (Geriist) of connective tissue. Leydig (’81, pp. 26-39, Plate 1, Fig. 5; Plate 1, Figs. 13, 14; Plate 3, Figs. 19, 20, Plate 4, Figs. 22-27; Plate 5, Figs. 28-31; Plate 6, Fig.°35) has de- scribed these organs in the same species. So far as their arrangement and appearance are concerned, this description accords with that of Ussow. The anteorbital organ is innervated by branches of the nervus trigeminus, which THE RADIATING ORGANS OF THE DEEP SEA FISHES. 173 enter it at various points. The stoutest nerve branch enters the organ at the constriction between the oval and conic parts. The cells composing the inner region are described as arranged in columns, Leydig suggests (1. ¢., p. 38) that the fish may swim on its side like a pleuronectid, and not verti- cally. He supports this view by stating that the pigment is not equally developed on the right and left side and by quoting a statement of Valenci- ennes to the effect that the similar Sternoplyx hermanni “ flotait renversé sur le cOte.” Garman (1. c.) does not mention the occurrence of any difference in the coloration of the two sides in A. lychnus, nor have I noticed it. However the case may be with A. henugymnus I do not think it likely that A, lychnus swims on its side. In a previous paper (’87, pp. 303-507) I have described the radiating organs of A. hemigymnus and other fishes, in which they have a similar structure. The fibrous layer in the organs of A. henu- gymnus was found to contain parallel and longitudinally arranged fibres of considerable length; so there is no difference between these species and A. dychnus in this respect. The statement there made that the inner region is composed of tubes med with glandular cells does not apply to A. lychnus, where such tubes could not be made out; there is hardly even a trace of the columnar arrangement described by Leydig. Brandes (99, p. 470), who also described A. hemigymnus, says that the anteorbital organs are connected with muscles and can be moved about at will so as to throw their radiation in any direction the fish may choose. This statement has been contradicted by Handrick (’01, p. 54). The fibre sheath consists of long and slender tissue cells containing guanin and lime. The tissue of the inner region is, according to Brandes (’99, p. 484), composed of roundish gland cells and is without efferent ducts. These cells produce a secretion which, according to Brandes, disintegrates chemically and thereby . emits light. The light passes the middle region, considered by him as a : biconcave lens, and illuminates the gelatinous mass which occupies the outer region, and which is backed by the inner face of the distal part of the reflecting fibre layer. Thus the luminosity would be visible chiefly from points at the side of the fish. Chiarini (’00, pp. 7-10, Figs. 1, 4, 5) has likewise studied the organs in A. hemigymnus. He reproduces a figure (1) of an axial longitudinal section of one of them, which—apart from its being upside-down — closely corre- sponds to the figure of a similar section of the anterior lateral organ in A. lychnus here given (Plate 6, Fig. 24). According to Chiarini the pigment 174 THE ‘RADIATING ORGANS OF THE DEEP SEA FISHES. layer of the organs of A. hemigymnus is not quite continuous, but inter- rupted by longitudinal lines free from pigment. The fibre layer is ex- ceedingly thick. The internal tissue has no special covering membrane and is in immediate contact with the fibre layer. The inner region of the internal tissue is composed of polygonal cells. Each contains granules and one or two nuclei. These cells are not regularly arranged; there is, however, an indication of a concentric stratification near the outer limit (towards the middle region). The cells of this inner region are supported by reticulate connective tissue. The middle region is composed of cells with a polygonal contour, arranged transversely. In the outer region the cells are arranged more or less longitudinally. Also in this middle region a supporting reticulation of connective tissue is met with. The distal portion of the cone is occupied by gelatinous tissue. Handrick (01) has published a detailed account of the nervous system and the radiating organs of the same species. He appears to have been unacquainted with Chiarini’s paper cited above. Handrick (01, p 58, 59) carefully worked out the innervation for most of the radiating organs (01, pp- 98, 59). The anteorbital is —innervated by the trigemius, the postorbital, opercular, and branchio- as Leydig had previously stated (see p. 173) stegal by the facialis, the others by spinal nerves. The radiating organs are not movable relative to the body (versus Brandes, see p. 173) but the con- cavity of the sheath can be changed by muscular action (01, p. 54). A few very elongated nuclei belonging to the fibre cells were observed in the in- ner sheath. The tissue of the inner region, the “Leuchtkérper,” is very minutely described (01, p. 55). It is not in direct contact with the fibre sheath (as stated by Chiarini), but separated from it by a special membra- nous layer of connective tissue. _From this, strands of connective tissue extend inward. These are much branched and anastomizing form a net- work along which nerves and blood vessels extend, and in the meshes of which large round or polyedric cells lie. The columnar arrangement de- scribed by Leydig (81, p. 31, see p. 173) cannot be made out. Handrick | C01, p. 56) gives a detailed description of the internal structure of these cells and distinguishes two varieties of them. According to him they al- ways have only one nucleus, not sometimes two as stated by Chiarini (see above). The tissue of the middle region is, according to Handrick (’01, p. 57), also supported by a reticulation of connective tissue. He considers the cells forming its proximal portion as spindle shaped, those forming the distal por- THE RADIATING ORGANS OF THE DEEP SEA FISHES. 175 tion as “ polygonal,’ meaning perhaps polyedric. The latter are, as it has been described above of A. dychnus, also in A. hemigymnus, smaller than the former. Handrick found blood vessels but no nerves in the middle region. The ex- tent of the gelatinous tissue occupying the distal part of the cone is dif ferent in the different organs; it is absent altogether in the anteorbitals. In this region a loose network of connective tissue with large nuclei ex- tends, the interstices being occupied by a gelatinous substance (’01, p. 58). According to this, the latest paper on A. hemigymnus, the radiating organs of this fish seem to have very much the same structure as those of A. lychnus. Sternoptyx obscura Garman. Plate 6, Figs. 26, 27. , This species was described by Garman (’99, p. 232, Plate 53, Fig. 1). The radiating organs are, as in Argyropelecus lychnus, very conspicuous. There are on each side 2 opercular (Plate 6, Fig. 26 op), 5 branchiostegal (br), 10 ventrothoraic (vt), 3 ventral (ve), 3 anterior lateral (al), 3 poste- rior lateral (pl), 3 anal (a) and 4 ventrocaudal organs (vc). As in Argyrope- lecus lychnus all these groups form rows in which they lie close together. The 20 ventrothoracic organs of the two sides are joined to form a medial crest protruding from the ventral side of the fish. The inner regions of the internal parts of these organs are joined and form a continuous mass. i RR i The axes of all the organs enclose very small angles with the outer surface and thus lie nearly paratangentially. The 2 opercular of each side are oblique, directed downwards and backwards, all the others more or less vertically downwards. | Besides these organs observed by me, Garman (99, p. 233) mentions a silvery area below the eye which he thinks may be a rudimentary suborbital radiating organ. > dee ee ee el The structure of the organs is apparently the same as in Argyropelecus lychnus. In some of them, however, chiefly in the ventrothoracic organs (Plate 6, Fig. 27), the proximal part of the pigment sheath (p) is thicker and the middle and outer region together are much less extensive than in that fish. These appear here (1 and s) as a rather thin, oblique, lenticular body separating the extensive gelatinous outer mass (gl) from the cellular inner region (g). Traces of stratification could also be detected in the outer region. « 176 THE RADIATING ORGANS OF THE DEEP SEA FISHES. I have described (’87, p. 303, Plate 70, Figs. 15-24) the radiating organs of Sternoplyx diaphana. In that species extensive radiating organs are met with on the lips, and I think that such may also be present in S. obscura ; but as that portion of the specimen at my disposal is somewhat injured I cannot express myself with certainty on this point. The glandular tubes as figured (87, Plate 70, Fig. 20) in the inner coalesced regions of the ventrothoracic organs cannot be made out in S. obscura; in this fish they appear to consist, as in Argyropelecus, of an alveolar network containing large, roundish cells. Lychnopoles argenteolus Garman. Plate 7, Figs. 31-37. This species has been described by Garman (99, p. 244, Plate 53, Figs, 4, 4a). The radiating organs appear as spots, 700 » in diameter, with a silvery lustre, and are pretty conspicuous. ‘There are on each side 1 ante- orbital (Plate 7, Figs. 31, 32 ao), 1 mandibular (Figs. 32, 33 ma), 11 bran- chiostegal (Fig. 33 br) 9 guttural (Figs. 33 g), 2 opercular (Figs. 31, 32 op), 14 ventrothoracic (Figs. 31, 32 vt)— Garman (’99, p. 245) counts 15 of these — 19 ventromedial (Figs. 31, 32 ve), 22 anal (Figs. 31, 32 a), 11 anterior lateral (Figs. 31, 32 al), 10 mediolateral (Figs. 31, 32 ml), and 22 posterior lateral (Figs. 31, 32 pl) radiating organs. Garman (99, p. 245) states that there are besides these two parallel rows of radiating organs on the sides of the body. In the partly injured specimen at my disposal these could not be made out. With the exception of the anteorbital organ which is somewhat peculiar and which will be described below, all the radiating organs mentioned have the same external and internal structure. They consist, as in Argyro- pelecus and Sternoptyx, of two parts, a proximal and a distal. The proxi- mal part (Plate 7, Fig. 837 A) is approximately spherical and forms about three quarters of a sphere. The constriction (C) between the proximal and distal part is well defined but shallow, so that the “neck ”’ of the organ has a considerable breadth. The distal part (D) appears as a broad and some- what short rotation paraboloid, obliquely cut off where it abuts on the surface of the body. The blood vessels in the subcutaneous connective tissue surrounding the organ (b) are more numerous close to the pigment sheath than elsewhere. The pigment sheath (Plate 7, Figs. 35, 37 p.) is stout, THE RADIATING ORGANS OF THE DEEP SEA FISHES. 177 dense, and of uniform thickness in the inner spherical part and the proxi- mal portion of the outer parabolic part. Towards the surface it thins out. The reflecting layer (Figs. 35, 37 r) is about as thick as the pigment layer and composed of longitudinally arranged fibrous cells. It does not extend to the surface and terminates some distance within the margin of the pigment layer. On the inner side of the reflecting layer a conspicuous connective-tissue membrane (Figs. 55, 37 c) about as thick as the two other layers, is met with. Distally this membrane passes into the gelatinous portion of the outer region of the organ. Three regions (Plate 7, Fig. 37 g, l, s), can be distinguished in the tissue forming the internal part of this organ in Lychnopoles as in Argyropelecus and Sternoptyx; in Lychnopoles, however, these are built on a somewhat different plan. From the connective-tissue membrane (Plate 7, Figs, 35, 37 c) mentioned above membranous diverticula extend inward which join to form facets arranged radially. In the proximal spherical part of the organ these mem- branes are straight and each facet encloses a regular pyramidal space. Its apex lies near the centre of the sphere, and its base abuts on the connective-tissue membrane overlying the reflecting layer (Plate 7, Fig. 37). The radial membranes (facets) do not extend quite to the axis. A some- what irregular tubular cavity (Plate 7, Fig. 37 cc) occupies that portion of the axial part of the organ which lies between the centre of the proxi- mal sphere and the focus of the distal paraboloid. The margin of this tube is so well defined that it looks as if a special limiting membrane were present.- Its cavity appears empty, neither blood corpuscles nor other bodies were found in it. In the pyramidal spaces enclosed by the facets, nuclei and residual protoplasmatic masses highly stainable with acid-fuchsin are met with. These adhere chiefly to the connective-tissue membrane at the base of the pyramid and to the adjoining parts of the radial membranes: the central parts of the facets (Plate 7, Fig. 37 ci) are pretty free from them. The facets in the distal paraboloid are not so regular as those in the proximal sphere. The inner (proximal) part of the paraboloid is occupied by pretty narrow, more or less curved facets (Plate 7, Fig. 37, em) extend- ing from the sides towards the distal end of the tubular cavity above described. These facets are pyramidal, distally broad and proximally con- tracted. The outer (distal) part of the paraboloid is occupied by regular 12 178 THE RADIATING ORGANS OF THE DEEP SEA FISHES. straight, prismatic facets (Plate 7, Fig. 87 co) which are parallel to the axis of the organ; they are rounded off distally and drawn out proximally to points converging towards the distal end of the tubular cavity. ‘The distal ends of these facets lie in a surface convex outwardly and i dolo vertical to the axis of the whole organ. The triangular space outside these facets, between them and the surface, is occupied by gelatinous tissue (gl). The small curved facets and the proximal portions of the large straight ones which together occupy the inner part of the paraboloid and form the middle region of the organ (Plate 7, Fig. 37) are clothed with a single layer of cylindrical cells which form a kind of simple epithelium (Plate 7, Fig. 36). | These cells take up the greater part of the space within the facet and only leave an exceedingly narrow central canal (Plate 7, Fig. 36 en) free. This canal extends longitudinally in the axis of the facet towards the tubular space in the interior of the organ. Probably it opens into it, as I have seen many indications of such a connection, but Iam not certain about it. The cylindrical cells forming the epithelial clothing of these facets and surround- ing the central canal are composed of two kinds of protoplasm. The proxi- mal three quarters of each cell (lying next the canal) (Plate 7, Fig. 36 ph) are transparent and only slightly stainable; the distal quarter (lying next the facet wall) (Plate 7, Fig. 56 pt) is granular and takes stains very readily, so that it appears dark and not transparent in the stained sections. In this highly stainable, basal portion of the cell the nucleus (nu) is situated. It is probable that the facets of the proximal spherical part of the organ described above contain the same kind of tissue, but the condition of the material does not allow of this being definitely determined. The tissue occupying the distal parts of the large straight cylindrical facets and forming the outer region of the organ (Plate 7, Fig. 57 s) is of an entirely different character. Here we find elongated, slender cells, which are rather indifferent to most stains, but become intensely yellow after treatment with picric acid. In the distal basal part of each of these facets these cells are very slender, spindle-shaped, and arranged longitudi- nally, parallel to the axis of the facet (Plate 7, Fig. 34 sp). Towards the proximal part they become stouter and shorter and here they often appear club-shaped. The cells appear to rise from the walls of the facet. Those joined to the terminal face of the facet arise more or less vertically ; those joined to the sides arise very obliquely and extend nearly parallel to the facet wall to which they are attached. The cells in the distal part of the THE RADIATING ORGANS OF THE DEEP SEA FISHES. 179 facets are nearly straight throughout the whole of their length, As we approach the proximal part of the facet we find these cells curving away more and more from the walls of the facet so that their free ends come to lie more or less obliquely (Plate 7, Fig. 54 sc). Finally they are situated nearly transversely, the cells arising from the different sides interlacing axially. In the sections one often finds a circular structure of considerable size (Plate 7, Figs. 34, 35 0) in the centre of the facet. Sometimes it appears as an empty space, and is perfectly colorless and structureless; sometimes it has—in sections stained with acid-fuchsin —a bright red color and then one occasionally notices traces of a granular structure within it. The tissue surrounding the circle often shows a fine radial structure (Plate 7, Figs. 34, 35). Ihave not been able to ascertain the real nature of these circles and their coronas with any degree of certainty. The circles seem to be the expression of spherical spaces lying between the cells, and the radial structure surrounding them might be due to the fact that the slender inner terminations of the spindle cells abut vertically on this sphere. The organ is rich in blood vessels. These follow the facet walls and can often be traced for a considerable distance. The gelatinous tissue (Plate 7, Figs. 34, 37 gl) mentioned above is com- posed of a gelatinous ground substance traversed in all directions by exceed- ingly slender spindle cells. It passes into the connective-tissue membrane overlying the reflecting layer. A thick granular lamella or cushion (Plate 7, Fig. 37 gr) underlies this gelatinous tissue and separates it proximally from the reflecting, and distally, where this is absent, from the pigment layer. As mentioned above, the anteorbital radiating organ differs in. some respects from the others. It appears to be much larger and measures 1.5 mm. in diameter. Closer examination shows that it is a duplex organ composed of two, each similar to the other radiating organs of the fish. These two organs point different ways and are enclosed in a common pig- ment sheath which has the shape of a curved tube open at both ends. The two orifices of this tube lie in the surface; one is oblique, directed upwards and forwards towards the eye, the other vertical, directed straight down- wards. The spherical parts (inner regions) of the two organs occupy its central, the paraboloidal parts (middle and outer regions) and the gelatinous and granular tissue the terminal portions of the tube. These two joined organs throw their radiation in different directions, one into the field of 180 THE RADIATING ORGANS OF THE DEEP SEA FISHES. vision and the other downward, neither inward nor backward, as Garman (99, p. 245), who also noticed that this organ differs in structure from the others, seems to think. Cyclothone acclinidens Garman. Plate 6, Figs. 28-30. This species has been described by Garman (’99, p. 247, Plate J, Fig. 4), whose statements concerning the radiating organs are, on the whole, in accordance with my observations; in his figure (4), however, 2 opercular and 5 posterior lateral organs are drawn which do not exist and which Garman himself does not mention in his description. The other organs are repre- sented in the figure as lateral whilst they are in reality situated on the ven- tral side of the body. The radiating organs are very small, only 700, in diameter, and are destitute of reflectors. On account of their insignificant size and want of lustre, they are far from conspicuous and not easy to make out. Thére are on each side 10 branchiostegal (Plate 6, Fig. 29 br), — Garman (’99, p. 247) gives their number as 13 — 13 ventrothoracic (Figs. 28, 29 vt), 3 ventrome- dial (Figs. 28, 29 ve), and 16 ventroanal (Figs. 28, 29 a) radiating organs. All the organs seem to have the same structure. They are approxi- mately spherical and enclosed in a pigment sheath which forms over three quarters of a sphere and encloses them on all sides except where they abut on the surface of the fish. A straight line drawn through the centre of the sphere and the centre of the more or less circular superficial orifice of the pigment sheath is vertical to the surface of the body. The pigment sheath (Plate 6, Fig. 30 p) is composed of cells containing dark brown pigment granules. It is very stout. Its inner surface is pretty smooth, the outer covered with irregular protuberances. Within the pig- ment sheath a connective-tissue membrane (Plate 6, Fig. 30 c) is met with. Near the surface this splits up into two membranes, one covering the mass of radiating cells proper on the outer side (Plate 6, Fig. 30 cr) the other forming a thick superficial cell layer (Plate 6, Fig. 50, 1 and cd). Three zones can be distinguished in the organ: an inner, occupying about three quarters of the whole sphere, an intermediate, and an outer. The inner zone is composed of a mass of large polyedrical cells and enclosed in a connective-tissue capsule. The part of this capsule which covers the radiating cell mass inwardly and laterally and separates it from . THE RADIATING ORGANS OF THE DEEP SEA FISHES. 181 the pigment sheath (Plate 6, Fig. 30 c) is the connective-tissue membrane mentioned above. The part of the capsule covering the cell mass outwardly and separating it from the intermediate zone (Plate 6, Fig. 50 cr) is the innermost of the two layers into which the connective-tissue membrane splits up near the surface. From this outward part of the covering membrane lying near the surface diverticles extend inwards dividing the outer two thirds of the cell mass into somewhat irregularly pyramidal facets, the (distal) bases of which are closed by the outer capsule membrane whilst the (proxi- mal) apices are wide and open. The innermost part of the cell mass is not thus divided into facets. The facets are occupied by large polyedrica! cells (Plate 6, Fig. 30 r) arranged in radial rows. Lach cell contains a large and conspicuous spherical nucleus; its protoplasm is readily stainable with acid- fuchsin. The same cells are met with on the inward side of the organ, but here they form layers parallel to the surface of the organ and are not arranged in radial rows. In the interior, a little below the centre, a mass of much smaller cells (Plate 6, Fig. 30 m) is observed. These’ look like closely packed blood corpuscles. Possibly there exists here a blood sinus; | have not been able, however, to make out clearly a wall separating these smaller internal cells from the larger ones around them. The intermediate zone (Plate 6, Fig. 380 t) has in sections the appearance of a nearly empty space. It is in life perhaps occupied by a gelatinous sub- stance with few slender cells. It may, however, be a mere artifact pro- duced by shrinkage. The outer zone is composed of two layers, an inner layer of cylindrical cells arranged vertically to the surface (Plate 6, Fig. 30 1) and an outer layer of ordinary connective tissue composed of slender cells arranged para- tangentially (cd) which forms the superficial covering of the organ in the orifice of the pigment sheath. The protoplasm of the cylinder cells of the inner layer is concentrated at their proximal ends which abut on the inter- mediate zone. Here also the nucleus is situated. The rest of the cell body is very transparent and the cell walls are difficult to make out. These cells do not take stains readily and thus essentially differ from the polyedric cells of the inner zone. In their microchemical behavior and their general ap- pearance they resemble the cells of the middle region of the radiating organs of Argyropelecus and Sternoptyx. A blood vessel, artery, (Plate 6, Fig. 30 b) of considerable size entering the organ at the innermost point was observed in several sections. 182 THE RADIATING ORGANS OF THE DEEP SEA FISHES. al Myctophum aurolaternatum Garman. Plate 5, Figs. 21-23. This species has been described by Garman (’99, p. 254, Plate 55, Fig. 3). His statements concerning the distribution of the radiating organs agree with my observations, but his figure (99, Plate 55, Fig. 3) differs in several respects from what I find in the specimen examined. The radiating organs are for the most part nearly equal in size about 600 » in diameter. They are provided with reflectors, lustrous, and there- fore, although so small in size, quite conspicuous. There are on each side 3 mandibular (Plate 5, Figs. 21,22 ma), — these are mentioned by Garman (’99, p. 265), but not represented in his figure (’99, Plate 55, Fig. 3),— 4 branchiostegal, 2 opercular (op),— the lower one, which is smaller than the upper one, is not represented in Garman’s figure (99, Plate 55, Fig. 3),— 3 pectoral (pe) forming an arc just behind the gill slit, 5 ventrothoracic (vt), over the last of these 1 anterior lateral (al), 4 ventromedials (ve), 3 mediolaterals (ml) forming an oblique row, 10 anterior ventroanal (va), 1 posterior lateral (pl), over the last one of the former 6 posterior ventroanal (vp) — according to Garman’s figure (99, Plate 55, Fig. 3) 7, — and 2 ventrocaudal radiating organs (vc) at the base of the tail. All the organs examined appear to have the same structure. Viewed from the surface each appears as a lustrous spot enclosed above and at the sides by a dark strip, which has the shape of a horseshoe open below. Kach radiating organ of the body lies between two scales. The outer one which covers it (Plate 5, Fig. 23 ss) is convex and possesses near the centre an H-shaped thickening (t) produced by a local accumulation of calcareous laminae, which make the scale very brittle. On treatment with acid the thickening disappears. The horseshoe-shaped pigment stripe mentioned above is attached to the inner side of this scale. An axial section of the organ vertical to the medial plane of the fish passes through the apex of the horseshoe (p). The lower scale (si) is concave; on its inner side a continuous layer of pigment (pp) is observed. The lenticular space be- tween the convex outer and the concave inner scale is for the most part occupied by connective tissue (c). This is composed of a ground substance traversed by slender, elongated cells which are somewhat irregularly curved. THE RADIATING ORGANS OF THE DEEP SEA FISHES. 185 This tissue is rich in blood vessels, The radiating organ proper, that is, the mass of cells (pe) which probably emit the radiation, lies in the middle of this connective tissue. It is pear-shaped, the thinner end extending outward and downward. The cells composing it are flattened and extend parallel to the surface of the body. The protoplasm is quite transparent and each cell contains a large granular nucleus. The blood vessels form a fine-meshed capillary network (b) on the surface of this cell mass, but do not penetrate into its interior. Radiating organs of this kind seem to be rare in fishes. Somewhat simi- lar ones have been described by Emery (’84, p. 473) from the ventral series of Scopelus elongatus. Chauliodus barbatus Garman. Plate 8, Figs. 38-40; Plate 9, Figs. 43-46. This species has been described by Garman (’99, p. 271, Plate K, Figs. 2, 2a). The radiating organs are very numerous. Three different kinds of them can be distinguished. The differences between these are considerable and clearly discernible with the naked eye. Some of the organsare simple, some composed, as in the fishes described above, of an inner spherical and an outer paraboloidal part. The simple ones again are either uncovered or provided with a pigment sheath. I have designated these three kinds of organs accordingly as simple organs without pigment sheath, simple organs with pigment sheath, and compound organs with pigment sheath. The simple radiating organs without pigment sheath appear as very small whitish dots. They are arranged in groups of from 4 to 16. These groups are scattered over the medial part of the ventral surface (Plate 8, Fig. 39 u) and also occur in great numbers on other parts of the surface, chiefly on the back of the fish. The simple organs with pigment sheath appear as dark spots 300 » in diameter. They are scattered in great numbers over the whole of the sur- face and seem to be most numerous on the sides of the head just behind the eyes (Plate 8, Fig. 38 x). In the barbel there are 8 of these organs (Plate 8, Figs. 38-40 Ba). The compound organs with pigment sheath are lustrous, elongated, about Imm. long and 600 » broad. Of these organs there are on each side | anteorbital (Plate 8, Fig. 38 ao), 1 suborbital (Fig. 38 so), 18 branchio- 184 THE RADIATING ORGANS OF THE DEEP SEA FISHES. stegal, 8 guttural (Plate 8, Fig. 39 g), 19 ventrothoracic (Figs. 38, 39 vt), 22 ventromedial (Figs. 38, 39 ve), 11 ventroanal (Figs. 38, 39 a), 17 ante- rior lateral (Figs. 38, 39 al), and 22 medial lateral (Figs. 38, 39 ml). The simple radiating organs without pigment sheath are not in a partic- ularly good state of preservation. So far as they could be made out, their structure corresponds with the same organs of Chauliodus sloani described by Chiarini (’99, pp. 16, 17). ‘They are spherical or oval, enclosed in a capsule of connective tissue and composed of a layer of superficial, radial, cylin- drical elements surrounding a group of central polyedrical cells. The simple radiating organs with pigment sheath (Plate 9, Figs. 43, 44) have an oval (Fig. 43) or irregular (Fig. 44) shape. They seem, however, always to be elongated in the direction of their axis. The angle between the axis and the part of the surface of the fish where they are situated is very variable. It may be 90°, the organ then being vertical to the surface (Fig. 43), or it may be smaller, sometimes so small that the axis of the organ is nearly parallel to the surface of the fish (Fig. 44). It appears that this angle is correlated to the shape of the organs, the regularly oval ones being usually vertical to the surface, the irregular ones inclined to it; and it seems that the organs are the more irregular the more obliquely they are situated. These organs are not, as is generally the case, imbedded entirely in the body of the fish, but protrude considerably over the surface, forming rather conspicuous rounded protuberances (Plate 9, Figs. 45, 44). They are covered distally by a layer of considerable thickness (Plate 9, Figs. 43, 44h) which appears structureless and transparent in the sections. Laterally and proximally they are enclosed in a pigment sheath (p). In the regu- larly oval organs of this kind the axis of which is vertical to the surface (Fig. 43) the pigment sheath extends only over the part of the organ imbedded in the body of the fish. In the irregular ones (Fig. 44) with oblique axis it covers also the proximal portion of the protruding part. The pigment sheath is thickest at the proximal apex of the organ and gradually thins out towards the margin. Its distal orifice is of considerable extent, and occupies about a third of the whole surface. In this a fine connective-tissue membrane (Plate 9, Figs. 43, 44) is extended which here replaces the pigment sheath. In this thin layer which separates the radiat- ing cell mass proper from the structureless outer covering mentioned above, nuclei, appearing paratangentially elongated in axial sections of the organ, are observed. The interior of the organ is occupied by a mass of poly- ; ; : THE RADIATING ORGANS OF THE DEFP SEA FISHES. 185 edrical cells in the centre, and cylindrical elements on the surface. The latter (Plate 9, Figs. 43, 44 pe) resemble a cylinder epithelium clothing the pigment sheath and distal connective-tissue membrane internally. They are not vertical to this surface however, but so arranged as to converge to a point within the organ very near its proximal apex. In the distal part of the organ these cells are short, hardly longer than broad. Inwardly they increase very rapidly in length and those situated at the sides are nearly half as long as the organ itself. The nucleus is conspicuous, spheri- eal, or radially extended, oval, and always situated at the distal end of the cell. The protoplasm is granular. The limits between these cells are distinct enough distally, but become very indistinct centrally and finally vanish altogether; in the proximal part of the organ to which these cells converge, no trace of cellular division is discernible. This part of the organ is occupied by a granular protoplasmatic mass (Plate 9, Figs. 43, 44 m) in which neither cell walls nor nuclei are met with. The proximal apex itself is in the regular oval organs free from these cells, or cells of any kind, the granular mass (m) extending here right up to the pigment sheath (Plate 9, Fig. 43). In the axial sections of irregular organs, however, a net- work of cell limits with nuclei in the meshes is usually observed in the part appearing as the proximal apex (Plate 9, Fig. 44). It seems to me, however, that this point does not exactly correspond to the proximal apex of the regular organs and that the “cells” seen there are transverse sections of the same kind of superficial cylinder cells as appear cut longitudinally else- where. These organs being so very irregular in shape, an irregularity in the arrangement of the superficial cylinder cells, giving rise to such an appearance of axial sections, would be not at all surprising. The inner mass of polyedrical cells (Plate 9, Figs. 43, 44 1) has the shape of a bee-hive or truncate cone resting with its broad base on the distal cylinder cells and extending inward a good way beyond the centre of the organ. The distal cells of this mass are smaller and arranged irregularly, the proximal ones larger and situated in more or less clearly pronounced longitudinal rows. In their appearance these cells differ essentially from the cylindrical elements surrounding them. Their nuclei are pretty large, spherical, or flattened radially, and their protoplasm is transparent and free from granules. Picric acid stains them a bright yellow. The superficial cylindrical cells are comparable to the cells forming the inner region of such radiating organs as those described above of Argyropelecus and Ster- 186 THE RADIATING ORGANS OF THE DEEP SEA FISHES. noptyx; the inner polyedrical cells on the other hand resemble the cells of the middle region of the organs of these fishes. Three of the simple radiating organs with pigment sheath in the barbel are joined to form a composite mass. Two of these are irregular and lie symmetrically right and left opposite each other. The third is a regular one and lies between the other two. The axes of the two former lie in a plane vertical to the axis of the barbel; the axis of the latter is ori- entated longitudinally, parallel to the axis of the barbel. A close inspection of a continuous series of sections through this organ reveals the remarkable fact that the central one of these three organs. is completely enclosed by the pigment layer on all sides, a thing [ have not observed in any other radiat- ing organ, although Leydig (’79, p. 369) mentions having occasionally seen simple organs with pigment sheath of C. sdoani thus entirely enclosed by pigment. Supposing, as one surely has a right to do, that the pigment sheath intercepts the radiation, we would have in that organ of the barbel, and in the similar ones observed by Leydig, radiating organs unable to radiate. A careful examination of these organs completely surrounded by pigment in Chauliodus may be warmly recommended to future investigators. According to Chiarini’s description (00, p. 16, Plate 5, Fig. 6) the simple radiating organs with pigment sheath of Chauliodus sloani are very similar ; his figure (6) however differs somewhat from mine (Plate 9, Figs. 43, 44). Chiarini states that in staining the superficial cylinder cells with haema- toxilin-eosin the nucleus absorbs the former, the protoplasm the latter. The compound radiating organs with pigment sheath (Plate 9, Figs. 45, 46) on the body are directed downward. Their axis is vertical, and parallel to the median plane of the fish. Thus the angle between it and the adjacent part of the surface is much larger in the ventral rows than in the lateral. In the latter it is very small, often appearing to be nearly ni, the axis then being approximately parallel to the adjacent surface. In their external shape these organs resemble the compound organs of Lychnopoles argenteolus described above. They are composed of a proximal spherical (Plate 9, Fig. 45 A) and a distal paraboloidal part (D). These parts are separated by a sharp and well-defined stricture (C) of considerable depth. The proximal part forms about three quarters of a very regular sphere. The distal part is short, broad, and cut off obliquely at the base, where it abuts on the sur- face. With the exception of the oblique terminal face of the distal para- boloidal part, the whole organ is enclosed in a stout pigment sheath (Plate THE RADIATING ORGANS OF THE DEEP SEA FISHES. 187 9, Figs. 45, 46 p). The cells composing the sheath contain brown pigment. They are very closely packed towards the inner well-defined and _ perfectly smooth surface. Towards the outer side they are farther apart and this surface is not so well defined, the pigmented tissue passing here somewhat gradually into the unpigmented tissue surrounding the pigment sheath. In the distal paraboloidal part of the organ a lustrous reflecting layer (Plate 9, Fig. 45 r) composed of slender fibres, is found within the pigment sheath. In the proximal spherical part no such reflecting layer could be made out. A connective-tissue membrane (Figs. 45, 46 c) clothes the whole organ. In the proximal spherical part it appears to rest immediately on the pigment sheath, in the distal paraboloidal part it rests on the reflecting layer. Distally this membrane merges into the transparent connective tissue (Fig. 45 d) which occupies the triangular space between the distal limit of the radiating cell mass proper and the outer surface of the fish. From the whole of the connective-tissue capsule fine radial septa extend inward. The radiating cell mass proper is composed of three regions, an inner (Fig. 45 i), a middle (m) and an outer (0). The inner region occupies the whole of the proximal spherical part of the organ with the exception of a distal conic sector. It is composed of large, elongate pyramidal cells (Figs. 45,46 pc). These are arranged radially in a very regular manner and nearly vertical to the connective-tissue capsule, on which their bases rest. They do not quite reach the centre to which they converge. ‘This is occu- pied by a plug-shaped mass of small cells (Plate 9, Figs. 45, 46 a) which appears to be connected with the radial septa extending inward from the connective-tissue capsule enclosing the radiating cell mass. These septa can often ‘be made out between the pyramidal cells. In sections large blood vessels (Plate 9, Fig. 46 b) are frequently seen extending radially inward from the surface to the central cell mass. Here they continue their course between the latter and the proximal ends of the pyramidal cells, and seem to give off exceedingly fine branches which extend radially outward between the pyramidal cells (Plate 9, Fig. 46) along the connective-tissue septa, separating the latter from one another. The pyramidal cells themselves (Plate 9, Figs. 45, 46 pe) are occupied at their distal end by protoplasm staining deeply with haematoxylin (Plate 9, Fig. 46 pt) whilst their central and proximal parts (ph), about five sixths of the whole, show no great affin- ity for haematoxylin and other stains. The protoplasm occupying this central part of the cell is, however, not hyaline but distinctly granular. In 188 THE RADIATING ORGANS OF THE DEEP SEA FISHES. the distal highly stainable terminal part of each cell one rather small nucleus (nu) is situated. The middle region (Plate 9, Fig. 45 m) occupies the distal sector of the proximal spherical part of the organ and the proximal portion of the distal paraboloidal part. The proximal limit is conic, the distal limit very regularly convex. The cells occupying the space between these limits (1) are well defined and polyedric in shape. The distal ones are small. From here they increase in size inwardly, the innermost ones being four to five times as large as the outermost ones. The protoplasm of these cells is homogeneous, not granular, and hardly stainable with haematoxylin. It shows a great affinity to picric acid and the whole cell mass of this region appears bright yellow in sections stained with picric acid-haemotoxylin. The nucleus is nearly spherical and larger in the large cells than in the small. The distal limit of the outer region is nearly a plane vertical to the axis of the organ, and so, its inner limit being concave and some distance away, the outer region (Plate 9, Fig. 45 0) has the shape of a thick plane concave lens. From the mass of connective tissue which abuts distally on the outer region, septa, nearly vertical to the limiting surface and parallel to the axis: of the organ, extend inward, dividing the outer region into prismatic facets rounded off distally. These facets are occupied by small elongated cells, the protoplasm of which behaves towards stains in a similar manner as the protoplasm of the cells of the middle region. The compound radiating organs of Chauliodus were first studied in C. slam by Leuckart (64, pp. 153-155). According to his description these organs of C. sloani are similar to those of C. barbatus. Leuckart chiefly studied the branchiostegal radiating organs (1. ¢, p. 154). These are cylinders, 1 mm. long and 500 p» broad. Their distal portion is occupied by a structure Leuckart considers as a lens. From the proximal face of this a conic protuberance arises, penetrating some distance into the inner cell mass; as mentioned above such a cone is also observed in C. barbatus. The reflecting layer consists of hexagonal pavement cells occupied by numerous crystals. In the outer part of the distal mass, that is, in the “lens,’ Leuckart claims to have seen fibres, whilst the inner part, and the proximal cone are composed of closely packed rods 5 w thick radiating from the apex of the cone. The substance composing these rods is highly refractive. No fibres or rods of this description occur in the middle and outer regions, which cor- respond to Leuckart’s “lens,” in the compound organs of C. barbatus. The THE RADIATING ORGANS OF THE DEEP SEA FISHES. 189 proximal mass which Leuckart terms “ Glaskérper,” and which corresponds to the inner region, is composed of large crystal-clear radial cones similar to cells; Leuckart, however, does not consider them as such, not having been able to find a nucleus in them. Each organ is supplied by a slender nerve. Leuckart considers these organs as accessory eyes, “ Nebenaugen.” It is noteworthy that in the discussion following the reading of this paper Claus (Leuckart, ’64, p. 155) drew attention to the similarity between these organs and the well-known structures of Euphausia, which were at that time also considered as accessory eyes. Ussow (’79, pp. 94-97, Plate 1, Figs. 1, 2, Plate 4, Figs. 25, 26) has also described the structure of the compound radiating organs of C. sloani. He finds them composed of two semispherical parts, the outer one of which is occupied by a gelatinous mass, colorless and transparent in the living fish, but coagulating and becoming opaque on being treated with the reagents used for preserving the specimen. Below this gelatinous mass a mushroom- shaped “lens” (79, Plate 2, Fig. 6 Kl) is seen, the cylindrical stalk of which extends inwards, whilst the semispherical upper part, com- posed of conic elements, protrudes into the outer gelatinous mass. This “lens” appears to correspond to our middle region. The conic elements of its distal, semispherical part are according to Ussow attenuated proxi- ‘ mally to fine threads, which extend inward parallel to the axis of the organ < and form the stalk of the mushroom, in which they are connected with “se granular multipolar cells. Concerning the inner region Ussow merely 7 corroborates the previous statements of Leuckart. These, however, he does not seem correctly to have understood. Ussow also states that a slender nerve leads to each organ. His description and also his figures are very different from what subsequent authors have seen in these organs of C. a sloani and show very little similarity to the structure of these organs in C. barbatus described above. Leydig (’79, pp. 365-382, Plate 15) examined the radiating organs of C. slam simultaneously with and independently of Ussow. His material, a specimen preserved in spirits for several years, was not good, but nevertheless his results were more valuable than those of Leuckart and Ussow. He also chiefiy studied the compound organs; his notes on the others have been teferred to above. The compound organs are surrounded by annular blood vessels and a slender nerve branch leads to each. The distal portion of the compound organ is composed of two parts, an outer cap and a lower mass. 190 THE RADIATING ORGANS OF THE DEEP SEA FISHES. The proximal cone described by Leuckart is part of this latter. In the distal cap club-shaped cells are met with, the distal end of which is the thicker one. It is laterally enclosed by stout cell walls, but destitute of a terminal membrane. The thin proximal end passes into a fine thread of considerable length. The nucleus is situated in the lower end of the thick part of the cell. The cells of the lower mass of the distal portion are of greater size, more finely granular, and have a larger nucleus. The proximal cone appears to be composed of processes of these cells which converge towards its apex. This cell mass is traversed by fibrous strands, extending inward from the surface and forming a kind of scaffolding which has the shape of a honey comb. The cells of this honey comb evidently correspond to the connective-tissue facets described above of C. barbatus. The cones occupy- ing the proximal part (inner region) of the organ are cells. Leydig found a nucleus in their granular outer end. The centre to which these. conic cells converge is occupied by a finely granular substance (‘‘ Punktsubstanz”’). Leydig thinks that the nerve leading to the organ terminates in this “ Punkt- substanz” and that the cells both above (cone of the middle region) and below (inner region) are connected with it. A portion of the reflector form- ing half a ring outside the stricture has a distinctly golden lustre. Leydig calls these organs accessory eyes, but uses this term (‘‘ Nebenaugen”) with a certain amount of mistrust. From this and from a quotation of Willemoes- Suhm, which he publishes, one can conclude that he was already at that time, although calling these organs eyes, quite open to the view that they had in fact a radiating function. | The descriptions given by these authors are based on material insuffi- ciently preserved and studied without the application of modern methods. They can therefore claim hardly more than an historical interest. Now, however, we have to review the papers by Brandt (1899) and Chiarini (1900) in which the compound radiating organs of Chauliodus sloani are described in a far more exact manner. ~~ According to Brandt's description (99, pp. 447-450) these organs of C. sloani seem to be identical in strueture with those of C. barbatus. He considers the cells of the inner region as gland cells in which a secretion is formed and then oxidized whereby light is produced, whilst the middle and outer region have no other function than that of conveying the ra- diation produced to the outer world. Chiarini (00, pp. 14-17, Figs. 6-7) also describes the radiating organs of THE RADIATING ORGANS OF THE DEEP SEA FISHES. 191 C. shanti in detail. To his statements concerning the simple organs we have f 4 v, : referred above. ‘The compound organs of this fish (01, pp. 14, 15, Fig. 7) consist, according to him, as in @. darbatus, of a proximal, nearly exactly spherical part, to which a funnel-shaped neck, separated by a distinct incision, is attached. In the compound organs of the body the funnel is cut off obliquely where it abuts on the surface of the body. In the branchiostegal organs transversely. Each compound organ consists of an outer pigment layer, a reflector, a connective-tissue capsule, and a central cell mass, in which three regions, an outer, middle, and inner can be distin- guished. These regions correspond to the three regions in the organs of C. barbatus. Beyond the outer region gelatinous tissue is met with, occupy- ing the space between its outer face and the surface of the fish. The outer region is divided by connective-tissue membranes into very narrow facets di- verging distally. ‘The middle region is clearly divided into an outer portion, occupied by polyedrical cells, and an inner portion divided into radial facets. The inner region consists of radially arranged conic cells, the larger central parts of which are fine grained and stainable with eosin but not with haema- toxylin, whilst the much smaller peripheral portions forming about a quarter of the length of each cell are readily stainable with haematoxylin. Each of these cells contains one or two nuclei situated in the peripheral portion. It will be seen from this that the structure of these organs in C. sloani is very similar to that in C. barbatus. The chief differences between them are the following: in C. sloani the facets of the outer region are narrow, the middle region is distinctly divided into an outer and an inner portion, and the cone cells of the inner region contain one or two nuclei; in C. barbatus, on the other hand, the facets of the outer region are broad, the middle region is not divided into two clearly distinct parts, and the cone cells of the inner region always contain a single nucleus. Idiacanthus antrostomus Gilbert. Plate 8, Figs. 41, 42. This species was first described by Gilbert (’90, p. 54), later more in detail by Garman (99, p. 280). It possesses two kinds of radiating organs, one without, the other with, a pigment sheath. The former are exceedingly small and arranged in two large elongate groups (Plate 8, Fig. 41 u) on the 192 THE RADIATING ORGANS OF THE DEEP SEA FISHES. sides of the head under the large suborbital organ with pigment sheath, and in six longitudinal rows, three on each side, of smaller groups on the body. The intervals between these small groups on the body are equal in extent to the intervals between the larger organs with pigment sheath, and the former alternate with the latter. The radiating organs with pigment sheath have a silvery lustre. Those of the body measure about 300m in diameter; the suborbitals are larger. There are on each side 1 suborbital (Plate 8, Fig. 41 so), 13 branchiostegal, 10 guttural (Fig. 42 ¢), 25 ventrothoracic (Figs. 41, 42 vt), 20 ventromedial (Figs. 41, 42 ve), 35 ventroanal (Figs. 41, 42 a), 25 anterior lateral (Fig. 41 al), 20 medial lateral (Fig. 41 ml), and 35 posterior lateral (Fig. 41 pl), radiating organs with pigment sheath. The small organs without pigment sheath are spherical, enclosed in a connective-tissue capsule, and composed of radially arranged cells, which usually enclose a central cavity. The organs with pigment sheath closely resemble the compound organs of Chauliodus barbatus described above. The suborbital organs are so dis- posed as to throw their radiation into the field of vision; the organs of the body are directed downward. Stomias hexagonatus Garman. Plate 10, Figs. 47-51. This species has been described by Garman (99, pp. 276, 277, Plate 56, Fig. 5). It possesses five different kinds of radiating organs. An organ in the barbel, a pair of suborbital organs, and on the body simple organs without pigment sheath, simple organs with pigment sheath, and compound organs with pigment sheath. The radiating organ in the barbel (Plate 10, Figs. 47, 48, 50), lies in a thickening 1.5 mm. long at the end of the barbel (Fig. 47,481). From its distal part three short terminal threads arise. The radiating organ itself (Plate 10, Fig. 50), is oval and placed transversely in the swelling of the barbel, the long axes of the two crossing nearly at right angles. It is enclosed in a stout connective-tissue sheath (Plate 10, Fig. 50 c), and com- posed of two parts, an upper, dorsal, smaller, dorsoventrally compressed (A), and a lower, ventral, spherical part (B). The smaller upper part is enclosed in a somewhat loose pigment sheath (p), lying within the upper part of the THE RADIATING ORGANS OF THE DEEP SEA FISHES. 193 connective-tissue sheath enclosing the whole organ. The upper portion of this pigment sheath forming the roof of the upper part is dome-shaped, the lower portion, forming its floor and separating it from the lower spherical part, is flat or even a little concave. At the circular line where the upper part joins the lower a conspicuous ring-shaped thickening of the pigment layer (pt) is observed. From this a small annular ridge extends outwards and downwards, forming the rudiment of a cup enclosing the basal, upper portion of the lower, spherical part of the organ. The interior of the upper part of the organ is traversed by a horizontal, strongly pigmented membrane (pm), which divides it into a larger upper and a smaller lower chamber. In both these chambers large cavities (cv) are seen. In the upper one there seems to be only one such cavity. This is situated ventrally and surrounded by a special endothelial cell layer, separating it from the transparent connective tissue occupying the dorsal and lateral parts of the upper chamber. In the lower chamber more cavities than one are seen. These lie dorsally. Below them in the ventral part of this chamber large sinuous blood vessels (b) extend. The lower part of the organ (B) is covered dorsally by the pigmented floor of the upper part of the organ and the annular rudiment of a pigment cup referred to above. Laterally and ventrally it is surrounded by the transparent connective-tissue sheath only. From the middle of the roof of this lower spherical part of the organ a cylindrical thread composed of connective tissue (z) arises which extends vertically downward to its centre. This thread is composed of longitudinally arranged, spindle-shaped connec- tive-tissue cells with oval nuclei. Along this thread blood vessels (b) and probably also a nerve extend from above down to the centre of the sphere. The lower end of the thread is thickened to form a terminal knob in which a sinuous cavity, filled with blood corpuscles, is observed. Apart from this thread with its terminal knob the whole of the sphere is occupied by large radially arranged conic cells (pce) equal in length to its radius. In arrange- ment and structure these cells are similar to the conic elements in the inner region of the compound organs of Chauliodus barbatus.described above. They are, however, more slender. The inner seven eighths are occupied by finely granular, transparent protoplasm not readily stainable; in the outer eighth of the length coarser grained protoplasm showing great affinity to hematoxylin and other stains is observed. In the peripheral part of each cell one small, spherical nucleus is situated. 13 194 THE RADIATING ORGANS OF THE DEEP SEA FISHES. The suborbital radiating organs (Plate 10, Fig. 47 so) are represented by Garman in his figure (99, Plate 56, Fig. 3), but not mentioned in his description (’99, pp. 276, 277). The suborbital organ lies below and behind the eye; it appears nearly spherical, is 800 pin diameter, and enclosed in a pigment sheath and a connective-tissue capsule. The pigment sheath forms a sac, the opening of which abuts on the surface of the fish. It is composed of concentric layers of rather loosely scattered cells containing brown pigment. Within this pigment sheath lies the connective-tissue cap- sule. The structure of the inner cell mass could not clearly be made out, as these organs were somewhat injured in the specimens at my disposal. The simple radiating organs without pigment sheath appear as small white dots 100 in diameter. They are scattered in great numbers over the ventral side of the fish and also occur in the three terminal threads of the barbel mentioned above. They are more or less spherical, enclosed in connective-tissue capsules and composed of pretty large radially arranged cells with conspicuous nuclei. The simple radiating organs with pigment sheath are met with on the hexagonal scales of the body and also on the head of the fish, They appear as dark spots and measure 300 in diameter. On each ventral scale a group of 7 of these organs occurs, upwards they gradually become scarcer, and near the dorsal medial line of the fish only one such organ is found on each scale. In their structure these organs resemble the simple radiating organs with pigment sheath of Chauliodus barbatus described above. The hyaline mass which in the latter covers the organ outside, is, however, absent in these organs of Stomias hexagonatus. The compound radiating organs with pigment sheath (Plate 10, Figs. 47, 48, 49, 51) appear as conspicuous dots with a silvery lustre and are 300- 500 in diameter. On the whole they are largest in front and decrease in size backwards. ‘There are on each side of the body 12 guttural (Plate 10, Figs. 47, 48 ¢), 16 branchiostegal, 38 ventrothoracic (Figs. 47, 48 vt), above these in a parallel row 87 anterior lateral (Figs. 47, 48 al), 10 ventromedial (Kigs. 47, 48 ve), 10 medial lateral (Figs. 47, 48 ml), and 19 ventroanal (Figs. 47, 48 a) compound radiating organs with pigment sheath. Two different kinds of these organs can be distinguished. The first somewhat larger kind is represented by the anterior compound organs of the body which form the ventrothoracic and anterior lateral rows. All the other compound organs belong to the second, smaller kind. { THE RADIATING ORGANS OF THE DEEP SEA FISHES. 195 The compound organs of the first kind (Plate 10, Figs. 49, 51) are spherical and have a short process directed downwards towards the outer surface. They might be compared to spherical bottles with small necks placed upside down. The axis of the organ is vertical, parallel to the median plane of the fish.. The ventral organs, that is, those belonging to the ventro- thoracic row, therefore abut nearly vertically on the surface, whilst the lateral ones, that is, those belonging to the anterior lateral row, are very oblique and enclose but small angles with the surface of the fish. In the former the “neek” appears cut off transversely ; in the latter, obliquely. The pigment sheath (Plate 10, Fig. 51 p) enclosing the organ is rather thin and composed of cells containing brown pigment. It surrounds all parts of the organ with the exception of the terminal face of the “ neck.” As in other compound radiating organs we also find in these below the pigment layer a reflecting layer. This is however not at all extensive, being confined to the distal part of the neck-shaped portion, where it forms a ring (Plate 10, Fig. 51 r) surrounding the “neck.” This-reflecting layer is com- posed of highly refractive threads. Within the pigment sheath and, where this is developed, the reflecting layer, a connective-tissue membrane is met with which entirely surrounds the inner cell mass and forms a perfectly closed capsule, from which diver- ticula extend inwards. The greater part of the interior of the proximal, spherical part of the organ is occupied by large radially arranged conic cells which converge towards the centre of the sphere. ‘These cells are very similar to the conic elements in the inner region of the compound organs of Chauliodus barbatus and like them consist of a long, proximal, finely granular part (ph) not readily stainable, and a short, distal part (pt) showing great aflinity to haematoxylin. The nucleus is remarkably small and situated in the distal staining part of the cell. These cells are divided from each other by fine connective-tissue membranes extending from the outer connective-tissue capsule radially inward towards the centre. In these membranes small radial blood vessels (b) can be made out. The “neck” of the organ (Plate 10, Fig. 51 cy) above referred to is a stout cylinder, circular in transverse section, the axis of which coincides with the axis of the whole organ. This cylinder is surrounded by an extensive annular cavity (s) dividing it from the connective-tissue capsule forming the innermost layer of the outer covering of the organ. The outer, distal, ter- 196 THE RADIATING ORGANS OF THE DEEP SEA FISHES. minal face of the cylinder is rounded off, simple, convex; the inner, proxi- mal face is drawn out to form a regular cone about as high as broad, which penetrates the mass of radial conic cells occupying the proximal spherical part of the organ. The apex of this cone lies a little way above the centre of this sphere. The whole appears as a cylindroconic plug (Plate 10, Fig. 51). From that part of the connective-tissue capsule of the organ which covers the terminal face of this plug, diverticula extend inwards parallel to the axis. ‘These form parallel facets. The central (axial) facet (cc) is the longest; it extends to the apex of the cone, whilst towards the sides of the cylinder the facets become shorter. Transverse (Plate 10, Fig. 49) and longitudinal axial (Plate 10, Fig. 51) sections of the plug-shaped outer part of the organ show that the contents of the central facet (Plate 10, Figs. 49, 51 cc) are different from those of the other facets (pe). In staining with | Van Gieson’s haematoxylin-picric acid-fuchsin one finds that the latter take up haematoxylin and acid-fuchsin very readily, whilst the substance in the central facet absorbs only the picric acid. It also stains with eosin. In the distal parts of the outer facets nuclei and cell limits can be distinctly seen ; here pretty large elongate cells are situated. In transverse sections of the distal part of the plug one sees two or three cell sections lying side by side in each facet (Plate 10, Fig. 49). In the proximal parts of the outer facets and throughout the whole length of the central facet cell limits cannot be made out and nuclei are hardly to be seen. Ussow ('79, p. 91) has described these organs in S/omias barbatus. He says that they are similar to those of Chauliodus, but that the middle region, which he represents as a mushroom-shaped body (see above) in Chauliodus sloani, is absent in Stomias barbatus. Organs similar to these have been described by Chiarini (’00, p. 12, Fig. 3) from Maurolicus poweriae. He says that each of the cells of the inner spheri- cal part and also of the cylindroconic plug, that is, the “neck” of the organ, often contain two nuclei. These cells are always mononuclear in Slomias hexagonalus. In longitudinal sections the limits between the cells are often so indistinct that one can indeed be easily misled on this point and imagine that the two or three nuclei seen imbedded in an apparently continous pro- toplasmatic mass belong to one and the same cell. Transverse sections, how- ever, show, at least in the distal part, the cell limits well enough and make it clear that here each nucleus belongs to a separate cell. The compound radiating organs of the second kind which are met with — THE RADIATING ORGANS OF THE DEEP SEA FISHES. 197 on the hinder part of the body have the same structure as the compound organs of Chauliodus barbatus described above. It is therefore not necessary to give a detailed account of them. (2) THE FISHES WITH RADIATING DISCS. Bassozetus nasus Garman. Plate 4, Figs. 14-17, This species has been described by Garman (’99, pp. 159, 361, Plates 77, 78). Its head shows a number of deep depressions (Plate 4, Fig. 14). On remov- ing the semitransparent skin from it a number of subdermal, clearly defined, white, somewhat elongated, rhombical patches 1-3 mm. long are exposed to view. ee ae of ix ASCE a > oe : > CE = > 4 J < = ~ — a 5 i) Oo G a 28 a a I mM } I i] ea f Wy ies th fe : ay, PLATE 2. Figs. 6-9. Malthopsis spinulosa Garman. . The head in three quarter profile. Drawing; natural size ; b, doubtful bud-like marginal organs; te, frontal tentacular papilla. Part of a transverse section of the interior of one of the ventral lobes of the distal trilobed part of the frontal tentacular papilla. Stained: haematoxylin-picric acid-fuchsin. Drawing. ¢, connective-tissue septa between the longitudinal chambers; cv, lumina of the glandular tubes; pc, walls of the glandular tubes composed of cylindrical cells. Transverse section of the distal trilobed part of the frontal tentacular papilla. Stained: haematoxylin-picric acid-fuchsin. Photograph, magnified 1: 41. b, peripheral blood vessels; b/, large central sinuous blood vessel; ¢, connective- tissue septa between the longitudinal chambers; cv, lumina of the glandular tubes; m, muscles; p, groups of pigment cells; pe, walls of the glandular tubes composed of cylindrical cells ; s, longitudinal strands of semicartilaginous tissue. Transverse section of the distal trilobed part of the frontal tentacular papilla. Stained : haematoxylin-picric acid-fuchsin. Drawing, magnified 1:41. b, peri-_ pheral blood vessels; b/, large central sinuous blood vessel; c, connective-tissue septa between the longitudinal chambers; cv, lumina of the glandular tubes; m, muscles; p, groups of pigment cells; pe, walls of the glandular tubes composed of cylindrical cells ; s, longitudinal strands of semicartilaginous tissue. “ALBATROSS EX 1891. FISHES -RADIATING ORGANS SS) I A222 cn (SY? "90 J 6-9 MALTHOPSIS SPINULOSA Garm a pe cr ‘ a Ages a i. Pi - ‘ ly t rf - 7 er £ arar'T te sTyessi ot Ha neghabest B4e0L anta j - ~ dabbp rbd pie? poos defteatl .Ob aid suits m seh vallaibm A tl aT UF ant fs } Sige ff fosasy Be init Pid ai> Mutitaton: A EL pil eens? vee Bree oma Hee at iaeere Pbissyses :tinnint ssn Sipe pcm amt SF pil ae rg a 7 a> «s* , * pba eieied 01501 ETD hoger a1 Poitin oo 0% a Aa wo A ; 4 - yoo ; ; at Se ut jaibecs ould Aapigaee © ae gallos re oe be ee I ae ~: | o> ee ee “ r or % ad cae as be ig? "2 . x ey — 40 Pod . 4 + fie eS oe ~ ie - - ji s _— fs | dA A o . < as Aes oe ; pees = vg a a ‘ = < <> % s ; a ) Se Asal Fig. 10. Fig. 11. Fig. 12. PLATE 38. Figs. 10-13. Leucicorus lusciosus Garman. The fish seen from the side. Drawing; natural size. A radiating dise in transmitted light. Photograph; magnified 1:41. b, blood vessel; n, main dise nerve; n’, smaller accessory disc nerves. A radiating dise in transmitted light. Photograph, magnified 1: 21. b, blood vessels; n, main disc nerve; n’, smaller accessory disc nerves. . Transverse section of a radiating disc. Stained: haematoxylin. Drawing; mag- nified 1: 42. b, blood vessels; c, paratangentially arranged connective-tissue cells; cv, cavities in the basal part of the radiating disc; n’, small accessory disc nerve ; p, pigment cells. ay | — — lusctosus Garman Leucicorus ) 10 -1. ‘i i a obt iT Ee apa laa itt Shhh bwin i pet Saves Ak ‘ 4 ys wat * 3 oat Dowyed nobshisitet (¢ yore Panis Hib Sisileac & ono Setevenny'T va Ea ok, away ac Oo pelsaedy heolif do Sd eb bellin is 5) Risen einen 4g povapucosih nisin mary eliva Sie Tren. iedbviad ~ sides mort poe aes AT BT Sif Te thas) frags isis td ii yeibens silanes eri: Saved: & PLATE 4. Fig. 14-17. Bassozetus nasus Garman. The fish seen from the side. Drawing; natural size. . Aradiating disc in transmitted light. Photograph; magnified 1: 58. n, main dise nerve; p, reticulation formed by the pigment cells with their long processes. . A radiating disc in transmitted light. Photograph; magnified 1: 27. n, main dise nerve; p, reticulation formed by the pigment cells with their long processes. . Transverse section of a radiating disc. Stained: haematoxylin. Drawing; mag- nified 1: 42. b, blood vessels; c, paratangentially arranged connective-tissue cells; n, main disc nerve; p, basal pigment cells. “Ki ‘ OA) ileal pee oat y © & 2 Su iy Stee (By why / at ine Foe “a eey ane ws OR OS es tale. 5 SH A nd RSET eee A TS S, ~~ + bl ee: i 5 3) © 08 ~ , 5 Es ie = RS = = ——~ = 3 = : re as 3 \ . ~ - ——— “at — a =X }. — SS ee —— - ~~ —_—_ = —_ NS 4A 47 Rp 5 a ’ 14 lf DASSOZ@ULS ASUS Garman SERN? F SIRT ATSB TO) £4 baitiart bagpibas 1h. SUS ihiLetey tus ee han: that g Aga } Bites tyre af batuabiivinsle: ue jit -wermank th osineic rich wo - en bedieer enistdilctign fy walt uF i Fetes at D jets Bevtihrery armel .OC-21 ft . -* - > aa a a 8 | e. MIRA (TSR De . A. wiipal Lie A teehee a ike = aaa Bysittah ae in atl G vie ees AE TS a Boinaenkl wha off ney Ts: + f PLate 5. ie ¢ od oo ina ah tes [Ts item . ayy We wr tue fs + SURI) Ty 4. f thi ? Bie MeL tiie. tte) Piri Hinil Hie? Bist oes esos iin a Dy! MVOE? bc as in : ytats hay 70? AOE i i AT ty me . we a eaibeariO & *TRiaEL vE ayats ‘ay tp oawlh omotod yy dpe tial rat di P ont 5 fy Thesmteragtinyy uy oe, cal Ott Sev ! Patty vais haa, Lash wie els sti thi at. ond wy Litay tase nes We teiaisiul fei . orf epic s:it : be pani) | ree? 0 Cele on A word Pebnyis cae: coryfle A -sais Sy, tira tun! > xemesrt Tink Hpourong li ipveirt tid fait 2 ane, emo? Be ee ’ 40) ncarbenetey mit it: Bi git 3 ett8 ot Se yee 14} Laret) Siar ais 2 RS yriltiHe rly fas Stes i” | a fades ob). is r Weg « a + ’ try ay td Pua s rr ihet eh] ang er thse i> iv] : P ly a4 ° Sree a. ie sah ioe y > al - : | aif t ~ Oe 7?) . 5 - }4 | rd vc * $4 , ¥ 2 . 2. wy i) « PuATE 5. Figs. 18-20. Macrurus canus Garman. Fig. 18. The fish seen from the side. Drawing; natural size. Fig. 19. A radiating disc in transmitted light. Photograph; magnified 1:41. n, main dise nerve; p, branched pigment cells. Fig. 20. A radiating disc in transmitted light. Photograph; magnified 1:17. n, main dise nerve; p, branched pigment cells. Figs. 21-23. Myctophum aurolaternatum Garman. Fig. 21. The fish seen from the side. Drawing; natural size. al, anterior lateral radiating organs; ma, mandibular radiating organs; ml, mediolateral radiating organs ; op, opercular radiating organs; pe, pectoral radiating organs; pl, posterior lateral radiating organs; va, anterior ventroanal radiating organs; ve, ventro- caudal radiating organs; ve, ventromedial radiating organs; vp, posterior ven- troanal radiating organs. Fig. 22. The fish seen from below. Drawing; natural size. al, anterior lateral radiating organs; ma, mandibular radiating organs; ml, mediolateral radiating organs ; op, opercular radiating organs; pe, pectoral radiating organs; pl, posterior lateral radiating organs; va, anterior ventroanal radiating organs; ve, ventro- caudal radiating organs; ve, ventromedial radiating organs; vp, posterior ventroanal radiating organs. Fig. 23. Axial section vertical to the median plane of the fish through a posterior lateral radiating organ. Stained: haematoxylin-picrie acil-fuchsin. Drawing; magni- fied 1: 140. b, blood vessels ; ¢, connective tissue of the interior of the organ; p, Section through the outer horseshoe-shaped pigment cell band; pe, radiating cells; pp, internal pigment cell layer; si, scale underlying the radiating organ; ss, scale covering the radiating organ; t, section through the H-shaped thicken- ing of the covering scale. — =e J F ; 7 7 yf FISHES —RADIATING S yh ) 5) i } Q 4 pee, 18-20 MACRURUS CANUS Garm. 21-23 MycTOPHUM AUROLATERNATUM wn aandT aii ainey| eievlqosyywik OF Fe aghd pois tin 116 dpnoattd awit wilt bveriadil anion ott od diegkhey avdtoad tity J wy i: 1 Dohingaw 5 itive ar eae | rrsiey hosts 7 ayo Suitcthes f yoitocnet ef ho POPS NY Sari BIN 1 jeer ah) Wo Fieg live lenin org ‘Insc abs 7 Rye jit. 0 jue Leth bus Jominor ot dipowied! aviturieuus at ; Borgen eee AS ae vain 2 pee off Te Mag jnisotod. Ass TRO 6 , Vv al Eitilooltet x ifange a0 waite leiden pomwerll .alie odd sucvl neck aed ad) 62 vit -PRidieshic a8 Ggreweo witethi Iratal oman . fe thir hiiaiss init joi peavey yuileites laguteokioinyd suey Phiniyte matin tarcdesy 9G .anaygorynisihns whis MOF) PHULE RIO urtitnitios febw erred: .o sahtneriza sioirie waivers qiarenonnay 59) Sein? inhomlo xy hjormd | Th wy 28? Swim igiits jective § wobie ot otott ooye eh oT 2S ofl Sanigeal rel Pisa’ ene oe irda olny .|) ; 6 isla ifs Bret TOPT REE «hey rie guplaiticr talus oy ta aa silinibes levtnse 47. penne sypsidALLys; 7 Lishas Sd ES _Wetistty eet e “en tit lo “ira isthonrould oF fio ticg Nolin tntxh. -<% wid tine hiv Honmpiilyzotememl «howe SBI waite Sieger) Svea evoitichos . ang mu oO} tortiniy| isensv ben Ail tt OES | bralt ’ mw) ures GYidOaMNGS Litres bo ) hha diwiikge pat ahd embtiw seta duperyigy isha To soho silt ti oes I te auirevirs a ’ do vers oid witht soo aie > UO TER YS tet eth) To ? 0 old Bo.S789 Lyrstin: mt pode peo) vie Gis 3) co shal F 2a og meptinly 4 ¢ alivo lo: pir qj i grit i ihin« = thi Reacr ist “PoanPns Alles Yo.! VOHOUETID, Lt butin wht. a4 tte tonilere : rN PLATE 6. Figs. 24, 25. Argyropelecus lychnus Garman. . 24. Axial section vertical to the median plane of the fish through an anterior lateral radiating organ. Stained: picric acid-carmin. Drawing; magnified 1: 84. A, proximal oval part of the organ; B, zigzag fibres of the reflecting layer; C, constriction between the proximal and distal part of the organ; D, distal para- boloidal part of the organ; g, inner region; 1, middle region; p, pigment sheath; r, reflecting layer; s, outer region. Fig. 25. The fish seen from the side. Drawing; natural size. a, anal radiating organs ; al, anterior lateral radiating organs; ao, anteorbital radiating organs; br, branchiostegal radiating organs; ml, medial lateral radiating organs; op, oper- cular radiating organs; pe, pectoral radiating organs; po, postorbital radiating organs; ve, ventrocaudal radiating organs; ve, ventromedial radiating organs ; vt, ventrothoracic radiating organs. Figs. 26, 27. Sternoptyx obscura Garman. Fig. 26. The fish seen from the side. Drawing; natural size. a, anal radiating organs; al, anterior lateral radiating organs; br, branchiostegal radiating organs; op, opercular radiating organs; pl, posterior lateral radiating organs; ve, ventro- caudal radiating organs; vé, ventral radiating organs; vt, ventrothoracic radi- ating organs. Axial section vertical to the median plane of the fish through a ventrothoracic radi- ating organ. Stained: haematoxylin-picric acid-fuchsin. Drawing; magnified 1:42. g, inner region; gl, gelatinous tissue occupying the distal part of the organ; 1, middle region; p, pigment sheath; r, reflecting layer; s, outer region. Lp 9) bas ga bo owt Figs. 28-30. Cyclothone acclinidens Garman. Fig. 28. The fish seen from the side. Drawing ; natural size. a, ventroanal radiating organs; ve, ventromedial radiating organs; vt, ventrothoracic radiating organs. Fig. 29. The fish seen from below. Drawing; natural size. a, ventroanal radiating organs ; br, branchiostegal radiating organs; ve, ventromedial radiating organs; vt, ventrothoracic radiating organs. Fig. 30. Axial section vertical to the median plane of the fish through a ventrothoracic, radiating organ. Stained: haematoxylin-picric acid-fuchsin. Drawing; magni- fied 1: 520; b, blood vessel leading to the organ; c, connective-tissue capsule within the pigment sheath; cd, transparent connective tissue forming the super- ficial covering of the organ in the orifice of the pigment sheath; cr, distal part of the connective-tissue capsule covering the mass of radiating cells outside ; |, cylindrical cells of the outer zone; m, central part of the organ composed of small irregularly arranged polyedrical cells; p, pigment sheath; r, outer part of the organ composed of cells arranged in regular rows; t, transparent inter- mediate zone. ‘ALBATROSS EX. 1891. FISHES-RADIATING ORGANS y 26 om 4 rod : ia ee Hie SST — SSS BS | tes SO — Sie Weiser Se ee tp SSShagaae H af M > SSS . } XN ‘ = 24,25 Argyropelecus lychnus Garm., 26,27 Sternoplyx obscura Garm., — 28-30 Cyclothone acclinidens Garm. ETrojan del. Lith Ansty Wemer&Winter Frankfurt? = SION dT bilan i XY avaad Sane) RoOstosyiB Sooqgontiond 2b TR ap ii 8 .b-.0Nth Leswdda 5 Botwea ohie ails, Mood poon dod ont tf ov a lmtihoolim on + anh Daidnibs lyiotnl sere le Aq hire. QNtriter. yblprogo a0 + aie. ARI Der Looted Lobo wey Jy win wibbin debate ear: 4 silane walivibet bev! . AURUNO Tykes Then" oo men A esis lodan : ealveetl Wolad “vit poe de oT O28 ah ro Baidnibies hiiimeiie on ; eiegO Bethe: Uestel tolatan, jn a Peaaio piney levsel ilecs Ath oF att AD: Rather velidibarn <4 EUByY yiitebat Inpodah sobre dq flere Briieibns “ ad =a Apo Rey Oi odape I7-F anayio ysrk te tian Tiled inet tor wel aawel hates iw sieht odd Yo bes oT G8 utd @ yltiben leregrst sy, SM seeg0 unveihe lcylanilowmd ad Joxie Lean ; serie Suteibes sabudilmem ass wn ihSevsts Hen od) Io saciq usgihom ott.ot leiiey aolfeoe leieeouu wo hot oe aol Paotenan sboniath .udives tine silt guinute POM wibsting le atel uh HURL oo - WA: | bottinasn :yaiwaG ee or) Dion “at Peeee NV GOLe Sines mlonvcloy de peibasl odf Yo alle ol onengy OTS, de Beavis 58. pried oli do rie Woo Gh ab ydiveo .O peerw 9 4 ‘iP LAE By lemixorg odd she eliia”, . .% Shik’ y- 3 I a ae i uf? wi uiles ’ ah, wD Bi nalso ci lo notion oepdiiu’k ot Bhi P =] Abou -hiee® Shialhy-ahyzotamend : cae le allew ol? oalemot susidamat suentt (igor Mot 7 Mtevle Iuomygiq .q 346081» To dang a os Ves oti; lmtnih odd ui #floa beijaila- aliarige lan we Sheiy cxiicom edt 03 leolipy nodtoos cir git esau et . tab 36 dann) uw Pad Sc wOUe (iro Quin ; ‘ ty Tbinn ohoig-wiiyzceniGhd sisal aw ed y eM odd paired : stmiiiaioniaareen ln vid: mum = =wPOT ’ ‘ a Palin “0% lor Da they) Geld ai bard: >» lntzn db Myo « hed ie Mnttinde Miiyet tg pelos Gi Yo naihioy Lerst xerty ey | allow i) oft 2 onely neihbow ol oh tooikiny polos betel 08 wT es otreic-uilerotcdiil shang orem ytd eller ‘7 : br mgys Bet Do Jung. tinisortge lemdzorg A 04 03 t hon ; 7 ine fos a holinisienas (0) saeods tnoitany tub) ; eet 7: oe ae Seabee 2 ;uegio oft Yo Jind lnhiolodiiiq leteih f | | as. + “vigetg Meee nl yore setuid 9 55 synl yaidooliat od) j : oe Re claps Toon! on) Yaudeon eortather jo + aamte ot Yo ae FT | | ee eanOL OOF coOrgo1 Olhbing ol! VO trang vont ol? To wdooal t f Wo uly ‘ aeald-ayidoun aon wing gis ;poton. sik an}. 20 tag weve 2 = Wo nol: soul 2; campo Rie Hag cases heme ) iq ladaily od Raiygiose gual) ol} Vomiois ny Unies’) saggy one ay Inset GHD BarQaoH oveett sity dodo lnemd anlodey py byt ihonia Javiireiy pn Hi Qo nuiget atid tf vg wie Biv iis ZS ny ARE, : e ; Ea i) Yo suely acibow oft of leider aoltoon JaizA (Lb vail Pree Ohsiq-nilyzothamel ih han lee ivy silt ioowiad notlsinanos : ‘ » Fw feeb 0: esite Aneta Ay oid nittiw oluxqes onwall fri tsdeil ond PEEKS meen Geno ebonialon ,b ; amr nigow wont pene relied sit Yo ebooe? ad) ni alioe 3 bi Ht yHoinn ist aS o ;aomet ofhhin or ; aotget albhin oni pevievel saliostet 1 pues seni ol? Yo eilss ones syral - a sehcye Hite By Zour ety FO cee al) douowls uotines saievanm I ob sort : BF arcinesien (heim angie soiteiier havoqaos bette) wae PRS” ' / Oneal alae tleiie to rm et SOM) | bétiwesm :valwesG ers 9d) ahaa) Roolit A nario! mifato soe Inoirariys lei oonttrt es aan ; a ed oa ayes} avi iponfue +> ; odds ald to onaary att a. gitifgile vine Tnaizony dq 5 icote Inourgiy Fd i al E wt Youaie aidaiiet vite Triitqiney §q pAlep Sao Sie Vile Sa , ' . pare = PLATE 9. Figs. 43-46. Chauliodus barbatus Garman. Fig. 48. Axial section vertical to the median plane of the fish through a regular simple radiating organ with pigment sheath. Stained: haematoxylin-picric acid-fuchsin. Drawing; magnified 1: 280. ¢, connective-tissue membrane forming the cover- ing of the mass of radiating cells on the outer side; h, outer apparently struct-_ ureless covering of organ; 1, central mass of polyedrical cells; m, granular protoplasmic mass at the base of the organ; p, pigment sheath; pe, peripheral cylindrical radiating cells. Fig. 44. Axial section vertical to the median plane of the fish through an irregular simple radiating organ with pigment sheath. Stained: haematoxylin-picric acid-fuchsin. Drawing; magnified 1: 280. ec, connective-tissue membrane forming the cover- ing’ of the mass of radiating cells on the outer side; h, outer, apparently structureless covering of organ; 1,, central mass of polyedrical cells; m, gran- ular protoplasmatic mass at the base of the organ; p, pigment sheath; pe, peripheral cylindrical radiating cells. . 45. Axial section vertical to the median plane of the fish through a medial lateral compound radiating organ. Stained: haematoxylin-picric acid-fuchsin. Draw- ing; magnified 1: 140. A, proximal spherical part of the organ; a, mass of small cells in the central portion of the proximal spherical part of the organ ; C, constriction between the proximal and distal parts of the organ; ¢, connective- tissue capsule within the pigment sheath; D, distal paraboloidal part of the organ; d, gelatinous connective tissue occupying the distal part of the organ; f, cells in the facets of the outer region; i, inner region; 1, cells composing the middle region; m, middle region; 0, outer region; p, pigment sheath; pe, large conic cells of the inner region; r, reflecting layer. Fig. 46. Transverse section through the centre of the proximal spherical part of a medial lateral compound radiating organ. Stained: haematoxylin-picric acid-fuchsin. Drawing; magnified 1: 340. a, mass of small cells in the central portion of the proximal spherical part of the organ; b, blood vessels extending radially to the centre of the sphere; ce, connective-tissue capsule; nu, nucleus of the conic cells; p, pigment sheath; ph, proximal only slightly stainable part of the conic cells; pt, peripheral highly stainable part of the conic cells. 5) oq. 43-46 Chauliodus barbatus Garman , + <4 ’ a « ‘ i j a | OL aread an Be Ah) aninnook ead ender 1-3) .notT Inhussoviisy sh) osie lewten SeniwevG able of mon moos en oY TI ni ‘ Be PROS BuiIaier Hrvoginos beredel doiIAE ole eum unitnibes ap eee me to cey7o uoltsiies ,l p ertenie enitatbes batons v HgTe Biaidvsto: t daittroilae , On ; PAIL BIO wehbe iis brundpsioo liietel nits 1 baxptoo Biot poipaie hy dy cases gestion heveqiuoy isiboin : : } : ni o Laie tty 2 .osttolirntigu peetwese - owoled tio? aese Hea afl .Fh coi a pao sidebar bertrocpniey Liotnl ainetan las attiewio saidai bay ym Baines Damo qgntoy byrtech Tid sears Ans: eaago guiierbr havo Fine demo niey ty cape malig Spueqios beihomonay ; ie ALLE a0 pitts brut! yom To. iey Lae gee APA MeR yO: okeveoRTF .OE 279 aetead beni jpeyio ywillrilua Serpe oie aleeogs © lo sueg one ‘ Yo door! fetuses on URS 2 1 baleen gone fit t-bine oroig — 16 Btosnt tobe ant oi ae ame 10. syed py ‘ala aa ‘OIE } Qo weno guitsilun ot Agiont Sein! vis of eesoyedie potdaen fnixk 06 ui : fT hoilingam :yoiver vis dan. bing ‘oi mieniligeodiaes wil: boninie a) to eq (iatiuiey) asvwol a 7 ORS tubs Te reap (awieh) {nyqo 1 Pui alssagy boold eaounia ben pbeoudlt Ietus ond guiarollol Jonesy dood TRE © sya ot jo diag Iseoholt Io rediasto 7 lnssob- fj ttemo ot Wo dieq tegqo od pi eottives j¥oO {REE oT - » othe to Suan Hering yavrol oil) 1é allso sine beifet atal nq jalteode sano 69 Yo Meg egy ont yuinwvent batons belsomelgy uy . weaved oth} tii ioaisedt3 nize x. ; wae domi “ofl gf eniteaniy ‘a > : bbs ose | 8) asl} a Z Herowls Wi oat Wr aise lq Matis of} Ob Inothier uote tee A 4 8i4 j Alar arinic-itiiy pedesiroed + boris SeMip Hite: bed foririnmy. Reswsad xiSeseodt ni eloeesy hook! b - oer altima Oueed- sByitarue pag oil to Meg deohalge tenizeig™ ~ paliial lirdiee wo obs 1stuo add ne net. ei oto Pa oo Hrawineon aii 9 (sey oft te Haq Mae off seinied y eseyite wantey Asia title Fo | firyiO mit To Mey hegaile.”. , , ; 0 silt ci Hb iieutoto yuciinniganl jy pales Jiromsiq RT 7 oe een yin: ‘Snail aq; wage gill Yo Jveq betel hoquil- wolg - “Fe tag teolisiqe lective of oi ile ofneey ‘wolkeiiasr ct ; 4 te Jipaifs Deysiieceulky off yolbtporine oamuinrtot yaiisyltey scuba iutsif f ejnde-uoly vill waihunoiure vives telecon ne Tics. = 5 i ¥ ‘ an f * F ae a ae eg Fig. . 48. ig. 50. PLATE 10. Figs. 47-51. Stomias hexagonatus Garman. . The fish seen from the side. Drawing; natural size. a, ventrocaudal compound radiating organs; al, anterior lateral compound radiating organs; g, guttural compound radiating organs; L, radiating organ of the barbel; ml, medial lateral compound radiating organs; so, suborbital radiating organ; ve, ventro- medial compound radiating organs; vt, ventrothoracic compound radiating organs. The fish seen from below. Drawing; natural size. a, ventrocaudal compound radiating organs; al, anterior lateral compound radiating organs; g, guttural compound radiating organs ; ml, medial lateral compound radiating organs; ve, ventromedial compound radiating organs; yt, ventrothoracic compound radiating organs. Transverse section through the outer distal part of the eylindroconic plug-shaped part of a ventrothoracic compound radiating organ. Stained: haematoxylin- picrie acid-fuchsin. Drawing; magnified 1: 280. ce, central facet of cylindro- conic plug; pe, transverse sections of cells in the outer facets of the cylindroconic plug. Axial section transverse to the barbel through the radiating organ of the barbel. Stained: haematoxylin-picrie acid-fuchsin. Drawing; magnified 1 : 120. A, upper (dorsal) part of the organ; B, lower (ventral) part of the organ; b, blood vessel following the axial thread, and sinuous blood vessels in the lower chamber of the dorsal part of the organ; c, connective-tissue sheath inclosing the organ; cv, cavities in the upper part of the organ; p, dorsal pigment sheath; pe, large radial conic cells of the lower spherical part of the organ; pm, pigmented membrane traversing the upper part of the organ; pt, annular thickening of the pigment layer; z, axial thread in the lower spherical part of the organ. Fig. 51. Axial section vertical to the medial plane of the fish through a ventrothoracic com- pound radiating organ. Stained: haematoxylin-picric acid-fuchsin. Drawing ; magnified 1: 160. ad, blood vessels in the septa between the conic cells of the proximal spherical part of the organ; c, connective tissue covering the distal part of the organ on the outer side; cc, central facet of cylindroconie plug forming the distal part of the organ; co, conic proximal end of the outer plug- shaped part of the organ; cy, cylindroconic plug-shaped outer part of the organ ; p, pigment sheath; pe, longitudinally elongated cells in the outer facets of the piug-shaped distal part of the organ; ph, inner only slightly stainable part of the radiating conic cells in the proximal spherical part of the organ; r, annular reflecting membrane surrounding the plug-shaped distal part of the organ; 5, annular cavity surrounding the plug-shaped distal part of the organ. i) IRGAN S Cc FISHES —-RADIATING Beis ” “ALBATROSS ml Se gle: : ‘4 fe te ie ie i) 4 « ° 2EI9 9 B. Meisel. lith. Boston 47-51 STOMIAS HEXAGONATUS Garm. eat } a AL atasd stig eMaibial anopaolal 34-S0 cyl Be oy Lesing °5 obi nib incr nope iat onl .24 ee Yt betiergam pp ea shih te nei olladqas wstiteih eg 60 ai ) ‘gamma Jn jovTes oil niet wo ponib of) at wnol lozaoy beoold spyal o werion seth yainanen ip? a i Boititeent siqetiedil = oodtyll Pesiipietiow selb.cilndqeo yuieibad .b0 git i fe wt povaew Seth atime a josih nid oF guibesl aloe? boold spr a I suangine BAT youre ay Sf befiimgamn jilqergood% ily), behing aa > ait agigqan Hake 26.91% a ® pas sevion ovib dina Oe emis <_ ‘. Fig. . 2. . 56. PLATE 11, Figs. 52-57. Halosaurus radiatus Garman. The fish seen from the side. Drawing; natural size. Radiating cephalic dise in transmitted light. Photograph; magnified 1: 17. Db’, large blood vessel leading to the disc; n, main disc nerve; n/, smaller ac- cessory dise nerves. Radiating cephalic dise in transmitted light. Photograph; magnified 1 : 41. b/, large blood vessels leading to the disc; n, main disc nerve; n’/, smaller ac- cessory disc nerves. Radiating cephalic dise in transmitted light. Photograph; magnified 1 : 17. b’, large blood vessel leading to the disc; n, main disc nerve; n/, smaller ac- cessory disc nerves. Transverse section through a cephalic radiating dise along its longer axis. Stained: haematoxylin-picric acid-fuchsin. Drawing ; magnified 1:42. b, close capillary network of blood vessels underlying the cylinder cell layer; b’, large blood vessel leading to the disc; ¢, thickened connective tissue of the inner wall of the slime canal on which the cylinder cell layer rests; cv, cavity of slime canal; d, eyl- indrical cells of the radiating disc; m, marginal zone of the radiating disc; n, main dise nerve; s, outer wall of slime canal. . Part of a transverse section through the cylinder cell layer of a cephalic radiating dise. Stained: haemotoxylin-picric acid-fuchsin. Drawing ; magnified 1: 200. b, close capillary network of blood vessels underlying the cylinder cell layer: ce, thickened connective tissue of the inner wall of the slime canal on which the cylinder cell layer rests; d, cylindrical cells of the radiating discs. “ALBATROSS EX. 1891. FISHES-RADIATING ORGANS PLATE Oo O Per E Trojan del. lith AnstvWemer& Winter, Frankfurt 9M. 52-57 Halosaurus radiatus Garman. ta Agha an a jy * » a a 7% ~~ - a. . 4 { i ; s ‘ & ‘ a 7 re # ’ RE a7and = ee ee SPOT BI we. es - HOM oot seinvoily’) Yop) yd ays iOrt ORinn'T oft Yo dated oidgergorby Ft Pipher. - URI eS ding A € Cf pravsdel “watrhed| A od Yo domT ai) © SS ae CORR. - or ? aay meet Fe ee ee oe nk: ie ern F | - ane j , : : 2 g: ie = ) ae fathoms | 1500 < ae Pread, ot be ' ES - % amedrs® ia * is CNET TH er NOTRE EO a PEL ta = - >? rue « - eA - a ab Paine a+-290_fathome PP Sto, * Ps airissh sce Lm et ny a OTe q it ne ee Lf Rad Bad Panes st OPP POs YE STE He rt ’ ; ‘ hmpimmepntase en: < Memoirs of the Museum of Comparative Zoology AT HARVARD COLLEGE. Vou; XXX, No. 3. REPORTS ON AN EXPLORATION OFF THE WEST COASTS OF MEXICO, CENTRAL AND SOUTH AMERICA, AND OFF THE GALAPAGOS ISLANDS, IN CHARGE OF ALEXANDER AGASSIZ, BY THE U.S. FISH COMMISSION STEAMER ‘“ ALBATROSS,” DURING 1891, LIEUT. COMMANDER Z. L. TANNER, U. S. N., COMMANDING. i a6 a fk ie | AXXVI. | EIN BEITRAG ZUR MORPHOLOGIE SS DES TIEFSEEFISCHGEHIRNES. | Von EMANUEL TROJAN. MIT SECHS TAFELN. [Published by Permission of MArsHALL McDonaLp and Grorcs M. Bewrrs, U. 8. Commissioners of Fish and Fisheries. ] CAMBRIDGE, U.S. A.: Printed for the Museum, OctToBER, 1906. INHALTSVERZEICHNIS. PAGE Pace merle oe ke ee ee TD Dog Mitelira og. ke y 24D MeN, fo> > | ee Cw 220 Das Hinterhirn.. . 20%. % «| DAA MEPHODEN. ........ . 220 Das Nackhirn= oo. ee a o BAL Das GEHIRN VON Bassozetus nasus 245 Das GEHIRN VON LEUCICORUS LUS- Das Vorderhirn .... . QYA45 TEE RNa ere! GS Pat ng ge Ok ; : c Das Zwischenhirn. . . . . 246 Das Vorderhirn . . « « « 228 aie Peed ; Jas ‘e a Releren ae Gd Das Zwischenhirn . . . .« « 225 D ‘ Ses vibe Oe cb Sere me tewle a Das Mittelhirn . . . . 1. 229 a see ae Das Hinterhirn. . . . .-. 281 Ae aera Fig ce ae = Das Nachhirn . . . . . . 288) Bee at Hee F200 USA : an a ke Das GEHIRN VON MIXONUS CAUDALIS 235 > praca 250 Das Vorderhan. os ew = 288 LITERATURVERZEICHNIS .. . 253 Das Zwischenhirn. . . . 287 Der Pinealapparat. . . . . 287| TAFELERKLARUNG ... .. 255 + a fi ¥ se : >< t Me ss * , id . » * A4s.4 pices pt EIN BEITRAG ZUR MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES. EINLEITUNG. Nachdem die in Bd. 30 Nr. 2 dieser Memoirs veriffentlichte Beschreibung der Leuchtorgane der gelegentlich der Expedition des “ Albatross’? in 1891 unter A. Agassiz erbeuteten Tiefseefische vollendet war, iibergab mir mein verehrter Chef, fiir den ich die Vorarbeiten zu jener Beschreibung durchge- fiihrt hatte, mit Zustimmung des Herrn Professor A. Agassiz, diese Fische, und versetzte mich so in die Lage, ihr Gehirn und ihre Sinnesorgane, an denen mir bei jenen Vorarbeiten einige interessante Besonderheiten aufge- fallen waren, zu studieren. Ueber das Ergebnis der die Hautsinnesorgane von Malthopsis spinulosa betreffenden Untersuchungen habe ich bereits berichtet,' itiber Auge und Gehér werden im folgenden einige Angaben gemacht. Besondere Aufmerksamkeit verdiente das Gehirn der Tiefseefische, von denen mir Leucicorus lusciosus, Mixonus caudalis, und Bassozetus nasus zur Verfiigung standen, da bisher erst zwei Tiefseefischgehirne einer ge- naueren Untersuchung unterzogen worden sind. Die eine stammt von Handrick und betrifft Argyropelecus hemigymnus (1901), die andere von Gierse, Cyclothone acelinidens behandelnd (1904). Beide ausgezeichnete Ar- beiten haben viel Neues und Interessantes zutage geférdert. An diese zwei Gehirne reihen sich nunmehr die von mir untersuchten drei an, wodurch an sich eine bedeutende Bereicherung des einschligigen Materials eingetreten ist, was aber relativ freilich noch nicht allzuviel, namentlich im Hinblicke auf die Frage, ob, und wenn ja, in welcher Weise das Gehirn der Tiefseefische in seinem Baue von dem der Seichtwasserfische abweicht, besagen kann. 1 Sitzungsber. d. D. nat. med. Ver. “Lotos” i. B., Bd. XXIV, S. 99-105, m. 3 Fig. i. T., 1904. “The structure of the bud-like organs of Malthopsis spinulosa.” Mem. Mus. Comp. Zool. Harvard College, Vol. XXX, p. 209-214, 1905. 220 MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES Fiir die Ueberlassung dieses mir hochwillkommenen und wertvollen Materiales muss ich an dieser Stelle Herrn Professor A. Agassiz, und meinem hochverehrten Chef, Herrn Professor Dr. R. von Lendenfeld, Vorstand des Zoologischen Institutes, meinen ergebensten Dank aussprechen. Durch die zahlreichen Winke und Ratschlige, durch die Unterstiitzung, die er mir weiterhin wihrend meiner Arbeit in liebenswiirdigster Weise zuteil werden liess, wurde ich zu ganz besonderem Danke verpflichtet, dem hier Ausdruck zu geben mir mehr als Pflicht ist. Meinem Kollegen Dr. Ludwig Freund, Assistenten am hiesigen Tieriirztlichen Institute, danke ich fiir das der Arbeit entgegengebrachte Interesse. MATERIAL. Die obengenannten Fische entstammen der durch die Tiefsee-Expedition des “ Albatross” im Jahre 1891 zusammengebrachten Sammlung, und sind von Garman (1899) in den Ergebnissen jener Forschungsreise niher be- schrieben. Sie waren in Alkohol konserviert, der ihre Gehirne noch keines- wegs derart beeinflusst hatte, dass eine richtige Untersuchung derselben unméglich gewesen wiire. Freilich waren die Ergebnisse bei einer anderen Art der Konservierung und einer kiirzeren Zeit bis zur Bearbeitung reichere gewesen. METHODEN. Der kleinste von den drei Fischen, Bassozetus, wurde entkalkt. Obgleich ich mich hiebei bloss einer 1% Salpetersiure bediente, dauerte der Prozess nicht lange, da das Skelett dieses ausgesprochenen Tiefseefisches vorwie- gend aus Knorpeln besteht. Die beiden grisseren Exemplare, Leucicorvs und Mixonus, wurden derart behandelt, dass nach dem Oeffnen der Cranial- héhle und eines Teils des Riickenmarkkanals das Gehirn aus der Schiidel- héhle herausgehoben wurde, was mit einiger Gefahr betreffs der Erhaltung des Pinealapparates und der Hypophyse geschah. Der erste war bei dieser Art der Behandlung zum Teil verloren, die letzte keineswegs. Alle Gehirne wurden in Celloidin eingebettet und in Querschnittserien zerlegt. Die Dicke der Schnitte betrug 20 oder 40 p. Gefiirbt wurde entweder Schnitt fiir Schnitt oder in toto mit Delafieldschen Hematoxylin. Zur Anfertigung der Bilder fiir die Totalansichten der Gehirne dienten Photographien der herauspriiparierten Gehirne, Zeichnungen, und Plattenmodelle. MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES 221 Das Gehirn von Leucicorus lusciosus. Taf. 1, Fig. 1-8, Taf. 2, Fig. 9-13. Folgende Bemerkung G'armans war es (1899, S. 140), die mich auf diesen Fisch aufmerksam machte: “Eye rudimentary, apparently without pupil or iris and with the ball greatly reduced and covered with black pigment. The eye differs greatly in appearance from that of other species of Bro- tuloids and suggests a possible adaptation to sensation from phosphorescence, or perhaps a modification fitted for the production of phosphorescent light.” Ks war mir nun daran gelegen, festzustellen, welche von diesen beiden Vermutungen hier zutreffend ist, anderenteils interessierte es mich, inwie- weit die Reduktion der Augen den Bau des Gehirnes beeinflusst hat. Das Ergebnis meiner Untersuchungen war die Bestiitigung der ersten und die Widerlegung der zweiten Vermutung. Offenbar verleitete das Fehlen der Iris und der Pupille, eine Tatsache aber, die seit Lrauers wertvollen Untersuchungen “Ueber den Bau der Augen einiger Tiefseefische” (1902) gar nicht absonderlich erscheint, Garman zu der Annahme, die Augen von Leucicorus dusciosus konnten Leuchtorgane sein. Auf Grund meiner mikroskopischen Untersuchungen kann ich erkliren, dass diese von Garman angezweifelten Augen tatsiich- lich der Lichtperzeption dienen. Der Umstand, dass es sich um einen ausgesprochenen Tiefseefisch handelt—er wurde aus einer Tiefe von 3436 m hervorgeholt — sowie das Fehlen der Iris mit der Pupille riefen in mir den Gedanken wach, ob nicht in den vorliegenden Gebilden sogenannte Teleskopaugen zu suchen wiiren. Diese Annahme erschien aber alsbald widerlegt sowohl durch die Gestalt — sie sind ellipsoide, und nicht cylin- drische Gebilde —als auch durch die Lage der Augendffnung -—~ sie ist ausgesprochen lateral und nicht rostral wie bei den Teleskopaugen. Nur die stark vorgetriebene Cornea, die mich einen weiteren Schluss beziiglich der Gestaltung der Linse ziehen lasst, wiirde fiir ein Teleskopauge sprechen. Wenn auch Garman keine Erwiihnung von einer Linse tut, so zweifle ich ihr Vorhandensein nicht an, da ich auch das Ligamentum pectinatum vorgefunden habe. Sie diirfte sehr gross und kugelig sein und hat, wie es bei solcher Gestaltung Regel ist, mit ihrer Vorderseite die Cornea stark vorgetrieben, mit der Hinterseite die Entstehung der Iris unterdriickt. Im iibrigen sind jedoch die Bestandteile eines Auges auch hier vorhanden und in mikroskopischen Priparaten unschwer zu erkennen. Die Sklera als 222 MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES siusserste Wand verknorpelt an keiner Stelle wie es sonst bei Fischen der Fall zu sein pflegt. Der Chorioidea fehlt die Silberschicht. Geftiss- und Pigmentschicht sind vorhanden, die letzte sogar in sehr dicker Lage. Das Pigment befindet sich in Dunkelstellung, nachdem der Fisch wohl kaum je aus dem Dunkel der Tiefsee herauskommen diirfte. An der Retina fallt die tibermiissige Liinge der Stiibchen auf, die tibrigens von Brauer (1902, S. 44) auch nicht unberiicksichtigt gelassen wurde; Zapfen fehlen. Nicht festzustellen war, ob neben der Hauptretina auch eine Nebenretina vor- handen ist. Wenn ich also meine Betrachtungen uber das Auge von Leucicorus lusciosus zusummenfasse, so fuhren sie. zu dem Resultate, dass gener Fisch im Besitze wirk- licher Augen, die nut Leuchtorganen nicht das geringste gemeinsam haben, ist. Diese Augen verraten eine Anpassung an die Tiefsee, die sich auf die Cornea, Linse und Iris erstreckt und die von Brauer (1902, S. 44) fiir das Sehen in den Tiefen des Ozeans bei der durftigen Beleuchtung als Gusserst zeweckmissig gefunden wurde. Ob aber das Sehvermégen dieses Fisches speciell ein gutes ist, das ist eine andere Frage, die ich eher auf Grund der Untersuchung der nervésen optischen Leitungsbahnen und Zentren, als der Sehwerkzeuge beantworten méchte. Ueberraschend schwach sind die beiden die Augen versehenden Nervi optici in Anbetracht der Sehorgane, noch iiberraschender aber fiir jeden, der den Typus eines Fischgehirnes kennt, die Kleinheit der Sehherde, der Lobi optici, welch letzter Umstand dem Gehirn von Zeucicorus ein ganz eigentiimliches Gepriige verleiht. In dem Masse, in dem sich der Gesichtssinn riickgebildet hat, scheinen sich Geruch- und Gehorsinn héher entwickelt zu haben, was namentlich in der Dicke der Nervi olfactorii und der entsprechenden Bulbi, in den miichtigen Ursprungsgebieten der Nervi accustici im Bereiche der Medulla oblongata, und der zu ihren Seiten gelegenen, grossen Sacculi zum Ausdrucke kommt. Die Sacculi bergen Gehorsteine (Taf. 1, Fig. 8), deren Grésse mich nicht wenig tiberraschte. Diese Otholiten haben Bohnenform, 11 mm Linge, 7mm Breite und 2 mm Dicke, also fiir einen Fisch von kaum 3 dm Linge ganz bedeutende Dimensionen. Ihr Rand ist ziemlich scharf, die Aus- senseite ist glatt und die Innenseite zeigt eine konzentrische Riefung. Der grosse, Gehorstein steckt locker im Sacculus und schwebt natiirlich zu Lebzeiten des Fisches in der dicken Fliissigkeit des Sackraumes. Es ist eine bekannte Tatsache, dass die Sacculusotholiten der Knochenfische gross - MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES 223 sind. In dem umfassenden Werke “Das Gehérorgan der Fische” von Retzius (1881) finden sich genug derlei Beispiele vor, doch in Gestalt und relativer Grésse kommt kein anderer diesem gleich. Die michtigen Sacculi weisen auf ein hochentwickeltes Gehérorgan hin, das beziiglich der Austrittstellen einiger Nerven eine Erscheinung hervorruft, auf die bereits Sagemehl (1891, S. 559) mit folgenden Worten aufmerksam macht: “Die miichtige Ausbildung des Labyrinthes bei Knochenfischen ist die Ursache gewesen, dass die vor demselben liegenden Trigeminus und Facialis niher an einander geriickt sind und _ teilweise Verbindungen mit einander eingegangen sind.” Handrick (1901, S. 6) hatte die Gelegenheit, diese Erscheinung an dem Tiefseefisch Argyropelecus zu bestatigen und konnte tiberdies noch feststellen, dass sich zu den ver- schmolzenen Wurzeln des V. und VII. Nerven auch die des VIII. zu einem breiten Band zusammendriingen; auch der Ursprung des IX. und X. ist vereinigt. Ich fand alle diese Verhiltnisse an Leucicorus ebenso vor. DAS VORDERHIRN. (Telencephalon, Prosencephalon, Hemisphaeren, sekundiires Vorderhirn) (iat. 1, Pic. 1, 2; 3 VE). Schon die vielen Namen dieses Gehirnteiles verraten, dass er oft Gegen- stand eingehender Untersuchungen gewesen sein muss. Und in der Tat, es hat lange Zeit gedauert, bevor die Forscher dariiber ins Klare gekommen sind. Das, was die iilteren Autoren wie Camper, Cuvier, Treviranus, und Fritsch dabei irregefiihrt hat, war der Umstand, dass dem Fischgehirne die gewohnlich stark verdickte Grosshirnrinde abgeht. So suchten denn alle den vermissten Teil weiter hinten, wodurch keine geringe Konfusion hervor- gerufen wurde. Indessen gelang es Rabl-Ruckhard (1884, 8. 21), das Homologon der Grosshirnrinde zu finden. Hs ist das bei Fischen zeit- lebens auf embryonalem Stadium verbleibende Pallium. Grottsche (1835, S. 244) kannte zwar diese einfache Epithelschichte, ohne sie aber der Gross- hirnrinde gleichzusetzen. Andere Autoren iibersahen sie, weil sie sie beim Abtragen der Cranialhoéhle vielleicht zerstérten, was ja haufig bei dieser Art von Behandlung zu geschehen pflegt, wenn nicht genug Vorsicht angewendet wird. Mit dieser Entdeckung Rabi-Riickhards, die alsbald auch allgemein als richtig erkannt wurde, war der Bau des Knochenfischgehirnes analog dem der anderen Vertebraten aufgekliirt und es ist seither nicht schwer, an 224 MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES einem jeden Teleostier-Vorderhirn, so auch am vorliegenden, die einzelnen Teile zu unterscheiden. Um den als Ventriculus communis bekannten Hohl- raum (Taf. 1, Fig. 4 Vc) gruppieren sie sich in folgender Art: Den Boden und die Seitenwiinde bilden die Corpora striata (Taf. 1, Fig. 4, 5 Cs). Diese von Haller (1898a, 8. 617) als Basalganglien bezeichneten Gebilde nehmen hinten aus dem Thalamus des Zwischenhirns ihren Ursprung, nachdem dieser sich selbst stark verjiingt hat. Sie sind dort unscheinbar und durch einen diinnen Boden miteinander verbunden. Rostralwirts aber verbreitert sich der letzte, und dorsal lagern sich den basalen schwiicheren Teilen michtige Hirnmassen an, nicht hoch, aber derart breit, dass sie iiber die basalen heruber- hingen (Taf. 1, Fig. 5 Cs); das sind die eigentlichen Streifenhiigel. Sie sind von einander durch einen schmalen Spalt, die Fortsetzung des Ventriculus communis nach vorne getrennt, bis ungefiihr in der Mitte eine Verbindung durch die Commissura anterior oder interlobularis (Taf. 1, Fig. 5 Ca) hergestellt wird. Ein Chiasma zweier Blutgefiisse (Taf. 1, Fig. 5 ch) ist ebenda auch wahrzunehmen. Vor dieser Commissur sind die Streifenhiigel dorsal total miteinander verschmolzen, die basalen Teile haben hier ihr vorderes Ende. Die Verschmelzung muss dennoch ganz vorne einer Pialamelle und einem schmalen Spalt, der zur vollstindigen Trennung der Streifenhigel bei allmihlicher Abnahme an Hirnmasse fiihrt, Platz machen; sie gehen in die Tracts olfactoru (Taf. 1, Fig. 1, 2, 8 Trol) tiber. Die Linge der letzten ist unbedeutend, immerhin geniigt sie aber, um eine markante Grenze zwischen den Corpora striata und den Bulbi olfactori anzuzeigen. Die Tracti hiingen jedoch miteinander nicht zusammen, sondern divergieren. Dies hat zur Folge, dass die Bulbi (Taf. 1, Fig. 1, 2, 3 Bol) von einander getrennt sind, was in so hohem Masse noch an keinem Teleostier wahrgenommen wurde. Von obenher ist der Ventriculus communis durch das bereits erwihnte epitheliale Dach (Taf. 1, Fig. 4 Pa), das seitlich in die Streifenhiigel iibergeht, gedeckt ; rostral aber miindet er frei in die Cranialhéhle aus. “Es finden sich,” schreibt &abl-Riickhard (1883, S. 308) “ mit Bezug auf das Verhalten dieser Hirnteile zwei verschiedene Typen des Knochenfisch- gehirns: solche mit lang ausgezogenen getrennten Tractus und von einander abstehenden Bulbi, sowie kurzen Nervi olfactorii (Cyprinoidentypus), und solche mit sitzenden verwachsenen Tractus und einander geniherten Bulbi, bei langen Nervi olfactorii (Salmonidentypus). In ersterem Falle stellt jeder Tractus eine geschlossene Rihre dar, die vorn mit einem Hohlraum des Bulbus, hinten mit der Spitze des Ventriculus communis in Verbindung MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES 225 steht und deren dorsale Wand vorwiegend epithelial, die ventrale markig ist. Im zweiten Falle sitzen die Bulbi dem vorderen Teile des Grosshirns unmittelbar auf; die Tractus sind tiusserst kurz und bilden den Boden einer gemeinsamen Hohle, die eine unmittelbare Fortsetzung des Ventriculus communis nach vorne ist, und die ihrerseits in zwei seichte Divertikel iiber- geht, einen fiir jeden Bulbus. Letztere verbinden sich medial, ohne indess miteinander zu verschmelzen, indem eine senkrechte Pialamelle sich zwischen sie schiebt.” Demzufolge ist es unschwer, das Vorderhirn von Leucicorus als ein nach dem Salmonidentypus gebautes zu erkennen, das nur insofern eine Abweichung erleidet, als die Tracti olfactoru von einander getrennt, und die Bulbi einander nicht so nahe sind, wie wir es sonst zu finden gewohnt sind. DAS ZWISCHENHIRN. (Diencephalon, Thalamencephalon, Thalamus, Lobus opticus, primires Vorderhirn) (Taf. 1, Fig. 1, 3 ZH). Die Grenze zwischen Vorder- und Zwischenhirn ist zwar makroskopisch nicht sichtbar, wohl aber an mikroskopischen Priiparaten durch das Auf- treten eines Plexus chorioideus (Taf. 1, Fig. 4, 6 Plch) gut gekennzeichnet. Dieser letzte scheint, was seine Ausbildung betrifft, bei Fischen den mannig- fachsten Variationen zu unterliegen. So wurde seine Existenz bei Teleos- tiern tiberhaupt gelaugnet. Und wieder war es Rabl-Ruckhard (1883, S. 291) vorbehalten, sein tatsiichliches Vorhandensein sicherzustellen, obzwar er anfangs mit zu denen gehiorte, die es verneinten. Nach Darstellungen dieses Autors geniesst er aber weder beim Karpfen, noch beim Barsche eine beson- dere Ausbildung, und ist daher unvollkommen zu nennen. Wenn andere Forscher, die sich mit Fischgehirnen beschiiftigten, ihn stillschweigend tibergehen, kann man daraus folgern, dass sein Vorkommen eben nicht allgemein ist. Ich selbst muss gestehen an den mir zugebote stehenden Tiefseefischen Plexus chorioidei von verschiedener Grdésse bis zur verschwindenden Unschein- barkeit gefunden zu haben. Von allen aber musste der von Leucicorus lusciosus durch seine Dimensionen auffallen. Das einschichtige epitheliale Dach wird durch Blutgefiisse in so zahlreichen Falten in das Innere des Ventriculus tertius (Taf. 1, Fig. 4 Plch) und zum Teil bis in den Ventriculus communis (Taf. 1, Fig. 6 Plch) derart eingetrieben, dass es denselben nahezu ausfullt wie etwa bei Dipnoern und Amphibien. 15 226 MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES Unmittelbar hinter dem Plexus beginnen die Ganglia habenulae (Taf. 1, Fig. 7 Gh). Auch diese Gebilde lassen in ihrer Entwickelung nichts zu wiinschen iibrig, obgleich sie von aussen nicht sichtbar sind. Sie liegen am Zwischenhirn dorso-lateral und sind vollkommen symmetrisch. Beziiglich des letzten Punktes schreibt Hadler (1898a, 8. 574) tiber die Haubenganglien bei Fischen folgendermassen: ‘Nach Goronowitsch (1834, S. 436) soll auch bei Acipenser” (wie nimlich bei den Selachiern) “eine Assymmetrie der Haubenganglien sich vorfinden und das rechte Ganglion miichtiger entwickelt sein wie das linke. So weit meine Erfahrungen reichen, kommt eine solche Assymmetrie bei Knochenfischen nicht vor, sondern es verhalten sich diese Ganglien durchaus symmetrisch wie bei den Salmoniden.” Wenn auch diese Ansicht Huallers hier zutrifft, muss ich doch hinzufiigen, dass sie keine allgemeine Giltigkeit hat. Schon Handrick (1901, S. 7) hat auf die Assymmetrie der Ganglia habenulae bei Argyropelecus hemigymnus aufmerksam gemacht. ine weitere interessante Beobachtung diesbeziiglich machte Gierse (1904, S. 618-619) an Cyclothone acclindens. Seine Befunde an den ihm zur Verfiigung gestellten 10 Fischchen driickt er in folgenden Worten aus: “Bei zwei Exemplaren war das linke Ganglion habenulae stiirker entwickelt als das rechte, bel weiteren zwei Tieren waren beide Ganglien einander ungefihr gleich, wiihrend die anderen sechs deutliche Assymmetrie zeigten, indem das rechte Ganglion sich starker entwickelt erwies als das linke. Es kommen bei Cyclothone alle drei Méglichkeiten vor, doch iiberwiegt die Assymmetrie und zwar mit starkeren Ganglion dextrum habenulae.” Diese Mannigfaltigkeit dirfte meiner Ansicht nach auf versehiedene Altersstufen Jener zehn EHxemplare zuruckzuftihren sein, denn Haller (1898a, 8. 574) fand, dass sich die Symmetrie der Haubenganglien bei Salmo erst mit zuneh- mendem Alter einstellt. Nach dieser Erklirung kénnen von jenen zehn Kxemplaren acht jiinger und zwei alter gewesen sein. Dass die Assymmetrie dieser Gebilde namentlich dem Embryonalstadium angehért, dariiber schreibt bereits Goronowitsch (1888, S. 442): “Es ist aber von Interesse zu notieren, dass ich bei Salmonidenembryonen die Anlage des rechten Ganglions in einem gewissen Stadium immer grisser fand als die Anlage des linken. Es hat sich also im embryonalen Zustande des Gehirns der Knochenfische dieser primitive Charakter erhalten.” Bei niederen Fischen ist die Assymmetrie konstant, wie Ahlborn (1883, S. 227) an Petromyzon, G'oronowitsch (1888, S. 436) bei Acipenser und Amua (S. 442), MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES 227 Eddinger (1900, S. 140) bei anderen niederen Vertebraten nachgewiesen haben, Auch ich werde im Verlaufe meiner Arbeit Gelegenheit haben, einen neuen Tiefseefisch mit assymmetrischen Ganglien habenulae zu nennen. Auf der Kommissur, welche das Ganglion der rechten Seite mit dem der linken verbindet, sowie weiter vorne iiber dem Pallium, sah ich den Stiel des Pinealorganes (Taf. 1, Fig. 6, 7 Est) verlaufen. Lr ist ein solider Strang. Ein Parapinealorgan ist be Leucicorus lusciosus nicht vorhanden, Soviel iiber die Teile des sogenannten Epithalamus. Hinten lateral gehen die Ganglia habenulae in den Stammteil des Zwi- schenhirns, den Thalamus (Taf. 1, Fig, 6, 7, T), tiber. Dieser ist sehr volu- minds und lisst nur einen ganz schmalen Spalt ftir den Ventriculus tertius (Taf. 1, Fig. 6, 7, 8, V III) frei. Eine scharfe Abgrenzung dieses Teiles nach vorn oder hinten ist unméglich; dariiber lesen wir auch bei Haller (1898a, S. 591): “Rostralwiirts geht letzterer,’ er meint den Thalamus, “ganz kontinuierlich und ganz ohne jede Begrenzung in jenen Teil des Vorder- hirns iiber, der fiir gewéhnlich als Pedunculus cerebri bezeichnet wird.’ Und iiber das hintere Ende schreibt Edinger (1896, 8. 8): “ Hs gehen hier Zwischen- und Mittelhirn unmittelbar in einander iiber.” Zum Thalamus pflegen auch die Lobi laterales (Taf. 1, Fig. 3, Taf. 2, 9, 10 Ll) geziihlt zu werden; das sind Wiilste, die zu beiden Seiten desselben lon- gitudinal verlaufen. Sie werden bei Fischen mitunter sehr gross. Handrick (1901, S. 8) bringt dies mit dem Sehvermégen des Fisches in Zusammenhang und meint, “dass der Lobus lateralis ein Gebilde sui generis ist, ein Produkt der im Thalamus befindlichen Sehzentren, infolge der excessiven Ausbildung der Augen.” KEiner anderen Ansicht ist Haller (1898a, S. 598). Er be- trachtet den Lobus lateralis als einen dorsalen, differenzierten Teil des Lobus inferior. Beide Hypothesen werden von Gierse (1904, S. 626, 627) angefoch- ten. Dieser Autor fand an Cyclothone, wie Handrick an Argyropelecus, dass die Lobi laterales an Masse die Lobi inferiores iiberragen, und er schreibt daher: “Gegen die Ansicht Hallers spricht das ungleiche Gréssenverhiltnis der beiden Lappen, so dass man bei meinem Knochenfische eher daran den- ken kénnte, der Lobus inferior wire eine Differenzierung, ein ventraler Teil des Lobus lateralis als umgekehrt. Desgleichen kann dieser seitliche Lappen kein Produkt der Volumzunahme der im Thalamus befindlichen Sehzentren sein, denn Cyclothone hat verhiiltnismiissig kleine Augen und schwache Nervi optici.” Leucicorus lusciosus liefert hierin nun insofern ein interessantes Verhalten, als 228 MORPHOLOGIE* DES TIEFSEEFISCHGEHIRNES er von jedem Charakter etwas in sich vereinigt. LEvnesteils sind die Lobi laterales kleiner als die Lobi inferiores (Salmo), anderenteils die Augen gross ( Argyropele- cus), die Nervi optici schwach (Cyclothone). Allerdings ist eine solche Zusam- menstellung nicht darnach angetan, uns uber die ursachlichen Beziehungen der beiden Lappen niher aufzukliiren. Vielleicht wird die entwicklungs- geschichtliche Forschung einst mehr Aufschluss geben kénnen. Der basale Teil des Zwischenhirns, der Hypothalamus, besteht aus der Pars infundibularis mit der Hypophyse, den Lobi inferiores und dem Saccus vasculosus. Das Infundibulum von Leucicorus (Taf. 1, Fig. 4, 6, 7J) ist gross. Eine starke Hirnmasse umgibt die Infundibularhéhle, die eine Fortsetzung des Ventriculus tertius nach unten ist (Taf. 1, Figs. 4, 6, Jh). Sie hat zahlreiche Seitendivertikel, von denen die wichtigsten und liingsten die hinten in den Lobi inferiores sind. Nach vorne werden die Ausstiilpungen zahlreicher, aber auch kleiner, bis sie sich ganz verlieren (Taf. 1, Fig. 4 Jh). Hier hat die Infundibularhéhle die Form einer niedrigen, horizontalen Spalte mit glat- ten Innenwiinden, die sie umgebende Hirnmasse nimmt Trichterform an. Das ist auch die Regel, denn hiemit beginnt der Stiel des Hypophysenorganes (Taf. 1, Fig. 2, 3, Hst). Er ist bei verschie- denen Fischen verschieden lang, zum Beispiel bei Cyprinus unscheinbar, bei Gadus wohl merklich. Davon hiingt auch die Festigkeit des Zusammen- hanges zwischen Hypophyse und Gehirn ab, auf die schon Gottsche (1835, S. 435) aufmerksam macht; je linger der Hypophysenstiel, um so leichter reisst die Hypophyse bei Herausnahme des Gehirnes aus der Cranialhohle ab. Bei Leucicorus ist er missig lang, sokd und trigt an semem Ende eine ansehnliche Hypophyse (Taf. 1, Fig. 3 H). Die Lobi inferiores (Taf. 1, Figs. 2, 3; Taf. 2, Fig. 9, 10 Li) whertreffen die Lobi laterales so an Grosse, dass diese in der Ansicht von unten gdnzlich verdeckt werden. Ich konnte in ihrem Inneren Hohlriume konstatieren. Diese sind nichts anderes ais seitliche Ausstiilpungen des Ventriculus tertius. Als solche erkannte sie auch Edinger (1904, S. 151), denn er schreibt: “Zwischen den Ganglien des Lobus inferior ragt der von zentralem Héhlen- grau ausgekleidete Ventrikel des Zwischenhirns hinein, mit lateralen Ausstiil- pungen noch einmal Seitenventrikel des Hypothalamus bildend.” Wer nun die mannigfach gefaltete Innenwand der Infundibularhéhle kennt, wird leicht begreifen, warum jene Ventrikel in ihrer Grisse variieren, bisweilen auch ganz und gar fehlen. Sie gehdren eben nicht zu den notwendigen gi i ek MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES 229 Bestandteilen des Knochenfischgehirnes. Wenngleich Cuvier (1828, S. 427) von ihrem seltenen Auftreten spricht, fand sie Grottsche (1835, S. 290) an allen von ihm untersuchten Fischen vor. Handrick (1901, 8. 8) stellte sie an Argyropelecus fest, bei Cyclothone sind sie nach Gierse (1904, S. 630) iiber- haupt nicht vorhanden. Diese Hohlriume kommunizieren bei Lewcicorus wie bei Argyropelecus hemigymnus, Hsox lucius, Trigla gurnardus und anderen Teleostiern mit der Infundibularhohle, mcht aber durch diese mit der Hypo- physe wie bei jenen Knochenfischen. Unter allen Teilen des Hypothalamus ist der Saccus vasculosus (Taf. 1, Fig. 1-4, 6; Taf. 2, Fig. 9 Sac) der grésste. Lr ist der basale Teil des Infundibulums, aufs reichlichste mit Blutgefassen versorgt und in seinem mittleren Teile so breit wie der Thalamus. Die dusserst mannigfach ver- zweigten Hohlriiume in seinem Inneren hiingen mit der Infundibularhéhle und somit auch mit dem Ventriculus tertius zusammen. DAS MITTELHIRN. (Mesencephalon, Corpora bigemina) besteht aus den Pedunculi cerebri, den Lobi optici und den Tori semicirculares. Die Pedunculi cerebri beginnen im Vorderhirn, unmittelbar hinter der Commissura interlobularis, durchziehen unter steter Volumzunahme den Thalamus des Zwischenhirns und stellen bei Leucicorus die Hauptmasse des Mittelhirns vor. Die Lodi optic’ — sie werden auch Tectum mesencephali genannt — sind es eben, die jedem Fischgehirn das charakteristische Aussehen verleihen. So schreibt Edinger (1892, S. 21) iiber sie folgendermassen: “. . . wer die kleinen Vierhiigel des Menschen kennt, wird erstaunt sein, wenn er die ungeheueren Lobi optici eines Fisches oder Vogels sieht.” Und so ist es auch. Nicht wenig war ich daher uberrascht, im vortiegenden Fische ein Gehirn vorgefunden zu haben, dessen Lobi optict makroskopisch uberhaupt mcht zu erkennen sind. Erst das mikroskopische Priiparat zeigt ihr Vorhandensein an und zugleich ihre ausserordentliche Reduktion (Taf. 2, Fig. 9, 10 Lop.). Wenn schon Gierse iiber das Tectum opticum von Cyclothone acclinidens (1904, S. 630, 631) schreibt: “Im Vergleich zu der starken Ausbildung des Tec- tum opticum bei anderen Knochenfischen, z. B. Rhodeus amarus (dinger, 1900, S. 126, Fig. 183), Salmo fario (Hailer, B. 1898, Bd. XXVI, Taf. XIII, XIV), Cyprinus carpio (Mayser, 1882, Taf. XIX, Fig. 35, Taf. XXI, Fig. 48, 49) ist das Tectum mesencephali bei Cyclothone bedeutend schwacher 230 MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES entfaltet und ragt vor allem nicht so weit nach unten, dass es noch seitlich beiderseits die Mittelhirnbasis umgreift und dariiber herabhingt,”’ so muss ich dasselbe Verhalten bei Lewcicorus lusciosus in einem um vieles erhéhtem Masse in Anspruch nehmen, wovon auch die Darstellung der diesbeziiglichen Querschnitte (Taf. 1, Fig. 8, 9, 10) am besten Zeugnis ablegt. Grottsche (1835, S. 262) tiusserte die Meinung, dass die Grisse der Augen mit der der Lobi optici in geradem Verhiltnisse steht. Gegen diese Ansicht wandte sich Malme (1891, S. 10) und behauptete, dass Gottsche’s Regel durchaus keine allgemeine Giltigkeit habe. So weist er auf Ammodytes tobdianus (Taf. III, Fig. 44a) mit kleinen Augen und grossen Lobi optici, und um- gekehrt auf Sebastes viviparus (Taf. 1, Fig. 11) mit grossen Augen und kleinen Lobi hin. Gierse (1904, S. 630) stimmt Malme bei, denn seine Befunde an Cyclothone sprechen ebenfalls gegen Grottsche. Es ist klar, dass auch ich auf die Seite der zwei letzten Autoren treten muss, nachdem Leu- cicorus so unansebnliche Loben bei grossen Augen zeigt. Die Unvollkommenheit des Dachabschnities gibt sich auch nach einer anderen Richtung hin kund. Im hinteren Teile (Taf. 2, Fig. 12, Lop) ist seine mittlere Partie nimlich auf eine ganz niedrige Hirnmasse reduzert, die endlich auch verschwindet und einer einfachen Membran Platz macht. Dies ist offenbar auf das Hinterhirn zuriickzuftihren, das sich mit seinem Stamm- teil (Taf. 2, Fig. 12 Ctr) nach vorne umlegt. Hine derartige Erscheinung ist bereits von Alaatsch (1850, 8. 25) an Cyprinoiden vorgefunden worden. Jener Autor fiihrt sie dort auf die iibermissige Ausbildung der Valvula cerebelli zuriick. Auch Malme (1891, S. 10, 11) macht eine Erwihnung von der Unvollstiindigkeit des Tectum opticum “bei Fischen, deren Cere- bellum nach vorne gerichtet und an das Mittelhirn gedriickt ist.” Von einer Gusseren Liingsfurche zwischen dem rechten und linken Lobus opticus ist bei Leucicorus keine Spur vorhanden. Der Torus longitudinalis (Taf. 2, Fig. 10 Tol), der median innerhalb des intralobuliren Hohlraumes verliuft, besteht wie in der Regel aus zwei symmetrischen eng aneinander liegenden Wiilsten, die férmlich eingekeilt zwischen den beiden Lobi optici liegen. Haller (1898a, S. 503) fand ihn bei Salmoniden stark erhaben und weit in den intralobulairen Hohlraum hineinragend, bemerkt aber zugleich, dass dieses Verhalten nur bei sehr alten Tieren vorkommt. Bei Argyropelecus ist der Torus longitudinalis nach Handrick (1901, S. 9, Taf. 1, Fig. 4-6 Tlo) ein einfacher Wulst, bei Cyclothone fehlt er nach Gierse (1904, S. 631) tiberhaupt. 7 ; MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES 231 Vom Boden und den Seitenwiinden des intralobuliiren Hohlraumes er- heben sich hinten die Tori semicirculares (Taf. 2, Fig. 12 Tse) in der Weise, dass sie mit ihren konkaven Innenflichen die Valvula cerebelli (ibid. Valc) ganz einschliessen. Dieser Befund stimmt mit dem Stiedas (1868, S. 25) an Lota vulgaris und Handricks (1901, Taf. 1, Fig. 6-9) an Argyropelecus tiberein. Ungewéhnlich stark entwickelte Tori semicirculares fand Klaatsch (1850, Fig. 84, 85) an Cyprinus carpio, und Gierse (1904, S. 632, Taf. XIV, Fig. 1, 2, 3 Tse) an Cyclothone vor. Namentlich bei dem letzten Fische ist ihr Verhalten merkwiirdig, denn sie treten frei hervor, wihrend sie im allgemeinen von den unteren seitlichen Abhiingen des Tectum opticum vollkommen verdeckt werden (Mayser, 1882, bei Cyprinus carpio, Taf. XXT, Fig. 48 Tor. sem; Hdinger, 1900, S. 126, Fig. 83, bei Rhodeus amarus; und Haller, B., 1898a, Taf. XIII, XIV p, bei Salmo). Die Tort semicirculares von Leucicorus sind von aussen nicht erkennbar ; sie entfalten thre Grosse nur nach mnen. Der Ventriculus tertius, den wir als schmalen Spalt im Zwischenhirn ver- lassen haben, geht bei diesem Fisch merkwiirdigerweise auf eine kurze Strecke in einen Kanal von rundem Querschnitt tiber (Taf. 2, Fig. 9 Vuz), der sich alsbald unter dem Tectum opticum verbreitert und Divertikel nach beiden Seiten hin sendet. Die in seinem Boden median verlaufende Lings- furche (Taf. 2, Fig. 10 Lf) gewinnt nach hinten immer mehr und mehr an Tiefe. Sie ist jedoch nur von unbedeutender Lange. Denn durch den obenerwihnten Zusammenschluss der Valvula cerebelli und der Tori semi- circulares wird sie vom intralobuliren Hohlraume getrennt, tauscht alsbald ihre Rinnenform mit einer im Querschnitte rundlichen ein, und wird zu einem wohlausgebildeten Aquaeductus sylvu (Tat. 2, Fig. 12 AS). DAS HINTERHIRN (Metencephalon, Cerebellum, Kleinhirn) (Taf. 1, Fig. 1, 3 HH) fallt durch seine Grodsse auf und iibertrifft in dieser Hinsicht alle iibrigen Teile des Gehirns. Es besteht aus einem wnpaaren Slammtei (Taf. 1, Fig. 1, 3; Taf. 2, Fig. 11, 12 Ctr) und einem Paar Seitenteilen (Taf. 1, Fig. 1, 3; Taf. 2, Fig. 11 Tl). Gottsche (1835, S. 462) nennt ein solches Kleinhirn ein “ Cere- bellum trilobatum.” Aber auch vor ihm war bereits diese Dreiteilung des Kleinhirns bekannt. So bemerkt Camper (1774, 8S. 181): *‘ Le cervelet form une espéce de cone tronqué qui a deux tuberosités laterales unies avec le cervelet.”” Und Arsaky (1813, S. 19) schreibt: “Hoc monente Cuviero 232 MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES semper azygon est. Vereor tamen ne vir praeclarus hic egregie falsus fuerit. Quamvis enim eum recte pronuntiasse modus, quo cerebellum avibus et quadrupedibus sensim evolvitur probare videatur, tamen plurima in piscibus exstant exempla, quae cerebellum minime tantum ex impari tubercula, sed saepissime ex tribus componi, impari scilicet in medio posito aliquoque tuberculorum pari eiusdem lateribus adstante ostendunt.” Auch Klaatsch (1850, S. 19) liess diese Verhialtnisse nicht unerwihnt. Wiewohl im grossen ganzen das Kleinhirn von Lewcicorus sich diesem allgemeinen Typus unterordnet, muss ich dennoch niher darauf eingehen, da es im besonderen einigermassen von dem der bis jetzt untersuchten Tief- seefische abweicht. Der Stammteil (Taf. 1, Fig. 1, 3; Taf. 2, Fig. 11, 12 Ctr) hat die Gestalt eines Rotationselipsoids, das unter einem Winkel von 45° aufgestellt ist, von vorn nach hinten abfallend, und iiberragt die Seitenwiilste bedeutend an Grésse und zwar derart, dass er sie von oben ganz verdeckt. Bei Argyro- pelecus sind gerade umgekehrt die Seitenwiilste grésser als der Stammteil, und Handrick (1901, S. 9) meint, dass die Vergrésserung der ersteren auf Kosten der letzteren geschehen ist. Bei Cyclothone sind alle drei Abschnitte “ungefiihr gleich stark entwickelt ” (Gierse, 1904, S. 633). Dass das Vor- derende des Stammteiles am Mesencephalon aufruht, habe ich bereits oben erwihnt. Das Hinterende ist scharf abgesetzt und liegt frei auf den Lobi posteriores des Nachhirns. Die Settenwilste (Taf. 1, Fig. 1, 3; Taf. 2, Fig. 11 Tl), identische Bildun- gen der Corpora restiformia Haller’s, stehen mit der Mitte des Stammteiles in engstem Zusammenhange. Ich habe sie nur in der Seitenansicht darstellen kénnen, weil sie an Breite weder den auf ihnen lastenden Stammteil, noch das sie tragende Nachhirn iibertreffen und daher sowohl in der Ansicht von oben, als auch in der von unten verdeckt sind. In der Linge stehen sie Gebilden ihrer Art weit nach, denn sie héren noch vor dem Ende des Stammteiles auf. Die Oberfliche des Kleinhirnes von Leucicorus ist vollkommen glatt ; Langs- oder Querfurchen sind nirgends zu bemerken, obgleich sie am Teleostier-Kleinhirn nicht ausgeschlossen sind. So lesen wir bei (ottsche (1835, S. 459): “Gewéhnlich sieht man am Cerebellum eine mehr oder weniger deutliche Mittellinie, welche bei Gadus und Esox stark ausgedriickt ist, und hiiufig sieht man ebenfalls eine Querfaserung auf demselben, z. B. bei Gadus aeglefinus sehr deutlich, so dass man mit Recht behaupten kann, der MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES 233 Typus der héheren Tiere verliere sich bei den Fischen nicht giinzlich.”’ Auf Mayser’s Darstellungen (1864, Taf, 1, 2, ff) ist diese Liingsfurche auch mehr oder weniger ersichtlich. Sie entgieng ferner weder Haller (1898a, Taf. 12, Fig. 2) an Salmo fario, noch Malme (1891, S. 13) an Anguilla vulgaris, auch nicht Grerse (1904, S. 634, Fig. 1, 20, 21) an Cyclothone. Querfurchen sah bereits Cuvier (1774,8. 182) am Kleinhirn des Thunfisches, und Maine (1891, S. 13) bei anderen Scombriformes (Echeneis, Elacate, Coryphaena). Noch ein Bestandteil des Hinterhirns muss hier angefiihrt werden, der, da er an Querschnitten bereits im Bereiche des Mittelhirns sich zeigt, auch dort schon kurz von mir beriihrt wurde ; es ist die Valvula cerebelli (Taf. 2, Fig. 12 Valc). Sie gehért ihrem Ursprunge nach dem Kleinhirn an und soll daher hier niiher besprochen werden. Dieses Gebilde, das nach Haller (1898a, - §. 512) fiir das Teleostiergehirn charakteristisch ist, geniesst bei ver- schiedenen Fischen eine ungleiche Ausbildung. Eine wohlentwickeite Valvula, die weit unter die Lobi optici reicht, besitzt die Forelle (Radi Rickhard, 1883, Taf. 12, Fig. 1; Haller, B., 1898a, Taf. 16, Fig. 41); eine sehr kleine dagegen nach Malme (1891, 8. 14) Callionymis lyra und Agamus cataphractus, “ wo sie nur als zwei ganz schwach hervortretende Anschwellun- gen an dem Teil des Hinterhirns zu beobachten ist, der die Héhle des Mittelhirns nach hinten begrenzt.” Auch bei Cyelothone ist sie nach Gierse (1904, S. 635) sehr klein und “erreicht nicht einmal den hinteren Rand des Tectum mesencephali, sondern liegt unter dem Vorderteil der Seitenwiilste des Hinterhirnes auf dem hinteren Basalteil des Mittelhirnes, sowohl mit diesem als auch den Seitenwiilsten (Tl) verwachsen (Fig. 20, Valc).” Die Valvula cerebelli von Leucicorus ist gut entwickelt. Sie erreicht zwar nicht die Commissura posterior wie bei den Salmoniden, immerhin aber erfiillt sie den intralobuliren Hohlraum bis zur Mitte. Ihre Oberfliiche weist keine Faltung auf, sondern nur eine deutlich ausgepriigte Liingsfurche. Sonst pflegt des dfteren auch eine Querfurche mit Ueberresten der Pia mater vorgefunden zu werden; diese habe ich hier nicht gesehen., DAS NACHHIRN (Myelencephalon, Medulla oblongata, verliingertes Riickenmark) (Taf. 1, Fig. 1, 2, 3; Taf. 2, Fig. 11, 12, 13 NH) ist bei Leucicorus das stark ver- dickte vordere Ende des Riickenmarkes, in welch letztes es auch unmittelbar tibergeht. Dorsal sitzen ihm die beiden Lodi posteriores (Taf. 1, 1; Taf. 2, Fig. 13 Lpo) auf. Wahrend den Seiten des Basalteiles die meisten der 234 MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES Hirnnerven entspringen, birgt er in seinem Inneren den Ventriculus quartus, die Rautengrube (Taf. 2, Fig. 11 Viv), die in ihrem mittleren Teil durch die Lobi posteriores gedeckt wird, vorn und hinten aber frei liegt. Ueber die Lodi posteriores kann ich einen alten Gewihrsmann, Gottsche (1835, S. 464), hier sprechen lassen: “Gleich hinter dem kleinen Gehirn kommt ein Paar graulicher Lappen in den meisten Fischen vor, welche sich miteinander iiber dem vierten Ventrikel verbinden, indem sie in der Mittel- linie ein Thal zwischen sich lassen. Aussen ist graue, inwendig weisse Substanz; ... 7’ Diese Worte passen wohl wegen ihrer Allgemeinheit auf jedes Fischgehirn, so auch auf dieses. Genaueres dariiber erfahren wir schon durch Gierse (1904, 8S. 636). Bei Cyclothone entstehen sie “schon in der vorderen Region, bald nach dem Auftreten der Seitenwiilste des Hinter- hirnes” und “bilden daselbst eine Art Schaufel oder Schépfléffel, in dessen Hvhlung der hintere Abschnitt des Stammteiles vom Kleinhirn seinen Platz findet (Taf. 16, Fig. 21).’ Wie die Abbildungen desselben Autors zeigen (Taf. 14, Fig. 1, 3) sind sie auch deutlich abgesetzt und gut ausgebildet. Dies kann ich von Leucicorus nicht sagen: hier. gehen sie unmittelbar ohne diussere Begrenzung aus den Seitenwulsten des Klemhirns hervor und nehmen ungefiihr dort ihren Anfang, wo das Hinterende des Kleinhirn-Stammteiles aufruht. Das “Thal,” von dem Gottsche spricht, ist eine longitudinal ver- laufende Rinne, die durch die Neigung der beiden Loben gegeneinander hervorgerufen wird und so lang ist wie diese. Auch bei Argyropelecus (Handrick, 1901, 8. 10) beginnt sie bei der Verschmelzungsstelle der Loben und diirfte auch ihr Ende erreichen. “Bei Cyclothone dagegen,’ schreibt Gierse (1904, S. 636), “ist die Liingsfurche nur an der Verschmelzungsstelle der Lobi posteriores eine kurze Strecke deutlich (Fig. 23 Lp) und ganz am Ende derselben, wo sich infolgedessen eine deutliche Zweiteilung der Lobi nachweisen lisst (Fig. 1 Lp). Zwischen diesen beiden Hinkerbungen bilden die beiden Lobi miteinander einen einheitlichen Wulst, ohne jegliche Fur- chung (Fig. 1, 22, 23 Lp).” Ks eriibrigt noch einige Worte iiber die Lodi vagales zu verlieren. Sie sind bei Leucicorus tiusserlich tiberhaupt nicht zu sehen und es verriat nur der Ursprung des Nervus vagus, wo wir sie zu suchen hatten. Hin Querschnitt an jener Stelle hat viel Aehnlichkeit mit dem von Salmo fario (Haller, B., 1898a, S. 399, Textfig. 10). Auch Handrick erwiihnt sie an Argyropelecus mit keinem Worte, wohl aber Guerse (1904, S. 636) an Cyclothone; und wenn er sie auch in der Beschreibung “im Vergleich zu den Lobi posteriores schwach und a MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES 235 unansehniich entwickelt” nennt, zeigen doch die Abbildungen (Taf. XIV, Fig. 1, 3 Lv), dass ihr Vorhandensein schon idusserlich deutlich erkennbar ist. Ebensowenig wie die Lobi vagales war der Lobus impar, der am Boden der Rautengrube verlaufen soll, aufzufinden. Er ist eben nur ein sekundiires Gebilde (Goronowttsch, 1896, S. 4) und fehlt bei anderen Fischen auch, was tibrigens schon Malme (1891, S. 17) konstatierte. Das Gehirn von Mixonus caudalis. Taf. 3, Fig. 1-7, Taf. 4, Fig. 8-13. Entsprechend der Griésse des Fisches — das mir zugebote stehende Exem- plar misst ein halbes Meter— ist auch das Gehirn voluminiser als das des vor- hergehenden. Auf den ersten Blick erkennt man, dass auch hier die Gestalt etwas Kigentiimliches an sich hat. Der Grund diirfte wohl folgender sein. Die Anforderungen, die an die Augen gestellt werden, sind nicht geringer Art, denn die Sehzentren, namentlich der Thalamus und auch das Tectum opticum, sind sehr gross und verleihen dem vorderen Teil des Gehirnes ein eigentiimliches Geprige. Hinten sind die grossen Labyrinthe auf die Gestal- tung der Gehirnteile von demselben Einfluss gewesen, wie ich es bereits auf Seite 223 bei Leucicorus geschildert habe. Trotzdem aber dieser Fisch den vorigen um mehr als eine Korperlinge iibertrifft, messen die Otolithen hier kaum zwei Drittel der ersten (Taf. 3, Fig. 7). In Gestalt kommen sie ihnen nahezu gleich. | DAS VORDERHIRN Taf. 3, Fig. 1, 2, 3, VH * von Mizonus lisst makroskopisch zwei Teile, die Corpora striata und die Bulbi olfactorii leicht, den dritten Teil, die Z’ractus olfactori, weniger gut erkennen. Dieser Umstand, den auch die mikroskopische Untersuchung bestiitigt, weist darauf hin, dass wir es nach Rabl-Riickhard (1883, S. 308) “ mit sitzenden, verwachsenen Tractus und einander geniiherten Bulbi, bei langen Nervi olfactorii (Salmonidentypus)” zu tun haben. Die Bulbi olfactorw (Taf. 3, Fig. 1, 2, 3, 5 Bol) sind im Verhiltnis zur Gehirngrisse klein, gleichsam unscheinbare Anschwellungen der aus ihnen hervortretenden starken Nervi olfactori (Taf. 3, Fig. 1, 2,3, 1). Die kurzen Tracti olfactorii (Taf. 3, Fig. 4, Trol) bilden eine verschmolzene Masse, in deren basalem Teile ein Kanal, die Fortsetzung des Ventriculus communis, nach vorne verliuft und frei in die Cranialhéhle endet. Ganz dem Salmonidentypus entsprechend gehen die letzteren nach hinten unmittelbar ohne jede besondere Begrenzung in die 236 MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES Corpora striata (Taf. 3, Fig. 6, Taf. 4, Fig. 8 Cs) iiber. Diese haben die besondere Evigentiimlichkeit, dass die von Haller (1898a, 8. 617) unterschiedenen Teile sich in Grisse gerade umgekehrt verhalten als sonst ; es ist ndmlich hier der basale Teil der breitere, wiihrend der dorsale thn an Grosse nicht errecht. Lateral verliuft eine longitudinale Furche (Taf. 3, Fig. 6 F) als Grenze zwischen beiden. Beziig- lich des Bodens, der das rechte Corpus mit dem linken verbindet, und der Commissura interlobularis verweise ich auf meine Schilderung auf Seite 224, um sie nicht wiederholen zu miissen, nachdem die Verhaltnisse hier gleich denen dort liegen. Die hintere untere Grenze des Vorderhirnes ist durch das Chiasma nervorum opticorum gegeben. Dieses zeigt aber bei Mixonus ein ganz eigenttimliches Verhalten. Die Sehnerven entspringen bekanntlich zum grossen Teile aus dem Tectum opticum (Hdinger, 1904, 8. 127). Ihre Faserbiindel umziehen den Thalamus in der Richtung von hinten oben nach vorn unten, um an der Hirnbasis vor dem Infundibulum auszutreten und das bekannte Chiasma zu bilden. Die Art der Nervenkreuzung ist bei verschie- denen Tieren auch sehr verschieden. Es handelt sich zuniichst darum, ob sie total oder partiell ist. Zdinger (1904, S. 157) meint, dass sie wohl mit der Stellung der Augen in innigem Zusammenhange steht; bei frontaler Lage der Augen (Mensch, Affe) ist eine totale Kreuzung nicht unbedingt notwendig, da die Bilder der medialen Teile zusammenfallen; daher tritt in diesen Fiillen nur partielle Kreuzung ein. Bei Fischen, wo die Augen voll- kommen lateral stehen, sind die Gesichtsfelder total verschieden und daher die Kreuzung iiberall total. Es ist nun Regel bei den Knochenfischen, dass der Nervus opticus der linken Seite sich einfach iiber den der rechten hinii- berlegt. Gleichwohl macht aber Wéedersheim (1898, 8. 235) auf Harengus und Hngraulis aufmerksam, wo der eine Nerv durch einen Schlitz des anderen hindurchtritt. Auch Edinger (1904, S. 158) lasst diese Erscheinung nicht ausser acht. Im allgemeinen gehen die Sehnerven nach der Kreuzung unmittelbar auseinander. Bei Mizonus caudalis machte ich aber folgenden Befund. In dem Chiasma der Nervi optici (Tat. 3, Fig. 3, 6 Ch) tut sich eine derartige Verflechtung der beiden Nerven kund, dass ntihere Verhiiltnisse der Art von Verkreuzwng sich absolut nicht erkennen lassen. Als einheitlicher Stamm, und zwar ganz plattgedriickt, nicht von rundem Querschnitt, verlassen die Sehnerven die Hirnbasis. Diese Nervenplatte begleitet das Vorderhirn vom Infundibulum bis unter die Bulbi olfactorii. Erst nahe unter dem Vorderende der letzten erfolgt die Auflisung in die beiden Nervi optici (Taf. 3, Fig. 1, Il). ine Bindegewebslage von ziemlicher Dicke umgibt das lange Chiasma von seinem, Ursprunge an MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES 237 allseits. Die Sehnerven gehen also hier nicht “gleich nach der Kreuzung”’ (Stanmus, 1849, 8. 13) auseinander, sondern verlaufen ein Stick des Weges gemeinsam am Boden der Schiidelhohle und verlassen sie erst spiter. Kinen iihn- lichen Fall beschreibt Gierse (1904, S. 639) bei Cyelothone. Allerdings hat das Chiasma dort nicht die Dimensionen und die Gestalt des vorliegenden. DAS ZWISCHENHIRN. Taf. 3 Fig. 1, 2, ZH. Der Lpithalamus des Zwischenhirnes besteht aus dem Plexus chorioidei (Taf. 4, Fig. 8 Pleh) und den Habenularganghen (Taf. 4, Fig. 9 Gh). Der erste erreicht bei weitem nicht jene Dimensionen wie bei Leucicorus. Die Blutgefiisse treiben die einfache Epithellage vor sich hin und erzeugen auf diese Weise Falten von unbedeutender Héhe. Derartige Verhiltnisse, die auch bei anderen Fischen vorgefunden werden, diirften einst manche Autoren glauben gemacht haben, dass es wahre Plexus chorioidei bei den Fischen nicht gebe. Der zweite Teil, die Gangla habenulae, nehmen die normale Lage ein, “am hinteren Seitenrand der dorsalen Zone” ( Wredersheim, 1902, S. 211; Ahlborn, 1883, 8. 228; Haller, B., 1898a, Taf. XVI, Fig. 41 gh). Es sind zwei vollkommen symmetrische, lobulire Gebilde von bedeutender Grdsse, die auch makroskopisch dem Beobachter nicht entgehen kénnen. Nur im hinteren Teile ist ihre Lage eine versteckte. Der Grund hiefiir liegt aber nicht darin, dass das Zwischenhirn in die Tiefe geriickt und durch das iiber- missig ausgebildete Mittelhirn von oben her verdeckt wird —dies kommt nimlich bei Fischen sehr hiaufig vor (vergl. Stieda, 1873, S. 446) — nein, hier ist es vielmehr der Thalamus (Taf. 2, Fig. 9 T), der die Gangha habenulae umgreift, indem er mdchtig anschwillt, sie in die Tiefe drdngt und schliesslich, nachdem sie selbst an Volumen zugenommen haben, hinten in sich aufnimmt. Die Commissur (Taf. 4, Fig. 9, 10 Com), welche das rechtseitige Ganglion mit dem linkseitigen verbindet, bildet nur vorne eine kurze Strecke eine Furche, ist aber sonst erhaben. DER PINEALAPPARAT von Mizonus setzt sich aus einem /Pimeal- und einem Parapinealorgan zusammen. Es ist bis jetzt tiberhaupt nicht bekannt, welche Bedeutung diese Organe haben sollen. Leydig gebiihrt das Verdienst, das erstere entdeckt zu haben ; 238 MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES Graaf und Spencer haben es eingehend untersucht, und Alinckowstroém seinen nervisen Zusammenhang mit dem Gehirne und zwar den Habenularganglien festgestellt. Wenn Haller (1898a, S. 569, Taf: XVI, Pig, 41) Tat, aoe Fig. 80) wie aus seinem Texte und namentlich den Abbildungen hervorgeht, den Ursprung der Epiphysendriise bei Fischen hinter die Basalganglien, an die Grenze zwischen Dien- und Mesencephalon setzt, so weicht hierin Mixonus insofern ab, als seine Epiphyse vorne aus den Ganglien habenulae entspringt, also eine vorgeschobene Stellung einnimmt. Hier sind vielmehr die Worte Fritsch’s (1878, 8.19) am Platze, der diesbeziiglich von der Zirbeldriise schreibt: ‘Sie bezeichnet gewissermassen die Grenzmarke zwischen dem Vorder- und Zwischenhirn.” Ehlers (1878), der als erster die Zirbel bei Haien und Rochen einer ein- gehenden Bearbeitung unterzog, unterscheidet an ihr drei Teile, die auch von anderen Forschern nach ihm stets genannt werden; es sind dies ein proximaler oder cerebraler, ein medialer, und ein distaler oder cranieller. Der erste pflegt “schwach kegelférmig gegen die mittlere Strecke sich ver- jiingend”’ zu sein, der zweite repriisentiert den Stiel, der dritte das Ende des Pinealorganes. Diese Teile werden auch bei Mixonus caudals unter- schieden, wenngleich sich das Ende des Pinealorganes meiner Untersuchung entzog. Die Epiphyse beginnt im rechten Ganglion habenulae wie sonst mit einer kleinen Anschwellung, die sich alsbald in einen Stiel (Taf. 3, Fig. 1, 2; Taf. 4, Fig. 8 Est) verwandelt. Dieser bewahrt bei Acanthias (Ehlers, 1878, S. 61) gleiche Dicke im Gegensatz zu Raja, wo der Stiel die Verjiingung der Ursprungsstelle bis zum feinsten Faden fortsetzt. In semem Anfang ist er solid, wohl aber nur eine kurze Strecke ; die weitaus lingere ist er hohl, also ein Schlauch. Ob seine Cavitiit auch auf den distalen Teil iibergeht, konnte ich nicht feststellen. Sollte dies aber der Fall sein, dann hatten wir dieselben Verhaltnisse wie sie Hill (1894, S. 248) bei Salmo fario schildert : “The distal part and the distal portion of the stalk retain the cavity, but the proximal portion of the stalk (not shown in the figure) has become solid.” Gerade umgekehrt fand es Kénckowstrém (1893, S. 563) bei Cal- lichthys (asper und litoralis), wo das proximale Ende des Stieles schlauch- formig, der distale Teil des Pinealorganes solid ist. Ein massiver Stiel mit solidem distalen Ende pflegt bei Fischen 6fter vorzukommen, wofiir Cattie’s (1882) Untersuchungen sprechen. Auch an dem Tiefseefische Cyclothone ist es so (Gierse, 1904, S. 623), wihrend Argyropelecus einen soliden Stiel, im Endteil aber eine Cavitit zeigt (Handrick, 1901, S. 37). MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES 239 Das Parapinealorgan von Mixonus (Taf. 3, Fig. 2, 6; Taf. 4, Fig. 8 P) ist, wenn auch bedeutend kiirzer als das Pinealorgan, doch gut ausgebildet. An der Ursprungsstelle jenem im Querschnitte gleichend, verjiingt es sich nicht, sondern schwillt keulenformig gegen sein Ende an. Es reicht nicht iiber die Corpora striata hinaus. Adbweichend von den bisherigen Befunden musste ich konstatieren, dass dieses Organ bei Mixonus durchwegs solid ist. Es fiele mir auf Grund dieser Untersuchung schwer, es als das zu erkennen, was es wirklich sein soll und ist, nach Ai// (1894, S. 209) nimlich eine Ausstiilpung des Ven- triculus communis. Auf eine solche konnte es auch Gierse an Cyclothone zuriickftihren, denn “es ist ein in seiner ganzen Ausbildung von einem zentralen Lingskanal durchzogener Schlauch, dessen Lumen beim Austritt vor das Gehirn ungefihr kreisrurfd ist und sich in seinem distalen Endteil in einen schlitzf6rmigen Spalt verengt. . . . Das Parapinealorgan von Cyclo- thone ist also wirklich eine Ausstiilpung des Ventriculus communis (Gi%erse, 1904, S. 624).”” Handrick (1901, S. 39) beschreibt es bei Argyropelecus mit folgenden Worten: “ Es ist im wahrsten Sinne des Wortes ein kolbenformig gestaltetes hiiutiges Blaschen (End), das caudalwirts in einen réhren- formigen hiutigen Stiel (E’st) ausliuft.” Daraus geht aber hervor, dass fiir manche Fische eine dorsalwirts gerichtete Ausstiilpung — Paraphyse — fest- steht, wie sie tibrigens bei allen Vertebraten vorhanden ist. dinger hiilt sie noch 1904, S. 137, bei Knochenfischen fiir unsicher, wohl aber nur in dem Sinne, als ihr Vorkommen bei dieser Vertebratenreihe kein allgemeines ist. Pineal- und Parapinealorgan von Mixonus verlaufen auf der Habenularcom- missur nicht iibereinander, sondern nebeneinander. Dieses Verhalten ist neu und aiusserst wichtig fiir die Hypothese Hi//’s. Jener Forscher, dem die Ent- deckung des Pinealorganes bei Knochenfischen vorbehalten war, kam auf Grund entwicklungsgeschichtlicher Studien an Salmo fontinalis, purpuratus, und fario zu dem Schlusse, dass die Epiphysen urspriinglich nebeneinander lagen (1894, S. 259). Handrick konnte sich fiir diese Hypothese nicht entscheiden (1901, S. 39), da am Argyropelecus eine etwas verschobene Lage des Parapinealorganes nur an einem Exemplare vorgefunden wurde. Noch weniger vermochte Gierse (1904, S. 625) mit Cyclothone Hil’s Ansicht zu bestiitigen, da das Pinealorgan streng median iiber dem Parapinealorgan ver- lauft. Die extramediane Lage bezieht sich nur auf einen Teil des Stieles. fier ist die Nebeneinanderlage, wenn auch nicht bis an das Ende, doch zum grossen Teile und zwar an einem entwickelten Tier, nicht am Embryo vorgefunden worden (Taf. 4, Fig. 8 P, Est). 240 MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES Wenn Edinger (1904, S. 186) den Thalamus als die Hauptmasse des Dien- cephalon bezeichnet, so liefert hiefiir auch Mixonus ein gutes Beispiel. Ich habe bereits bei den Gangliae habenulae (vergl. 8. 238) auf seine tibermissige Ausbildung und die Folgen derselben hingewiesen. Trotz seiner Grisse ist es aber auch an diesem Fische unméglich, ihn nach vorn oder hinten deutlich abzugrenzen. Ja gerade im hinteren Teile steigt er zu so einer Hohe heran, dass er unmittelbar in das Tectum opticum tibergeht. In seinem Inneren dehnt sich der Ventriculus tertius (Taf. 4, Fig. 9, 10, Vix) aus, der durch Verschmilerung aus dem Ventriculus communis hervorgegangen ist und mit Ausnahme des oberen erweiterten Teiles Spaltform besitzt. Er kommuniziert unten mit der Infundibularhéhle. Die Aussenseite weist die Lobi laterales (Taf. 3, Fig. 1, 3; Taf. 4, Fig. 11 Ll) und die dem Hypothalamus bereits angehérenden Lobi inferiores (ibid. Li) auf. Rostralwiirts sind die beiden’ kaum von einander geschieden und erscheinen als einziger michtiger Wulst. Caudalwiirts tritt zwischen beiden eine Liangsfurche auf, die zwar in der Mitte genug tief ist, noch vor dem Ende der Loben aber verschwindet. An Grosse stehen die Lobi laterales den inferiores nach; es wire somit Haller’s Hypo- these (vergl. S. 227) hier zutreffend, ier speziell, da man die ersten aus den letzten vorne entstehen und hinten in thnen aufgehen sieht. Der basale Teil des Zwischenhirns, der Hypothalamus, ist bei Mixonus sehr breit und ausserordentlich lang; reicht er doch tiber das gewohnte Mass hinaus und endet erst unter dem Hinterhirn. Seine Bestandteile sind das Infundibulum mit der Hypophyse, der Saccus vasculosus, und die Lobi inferiores. Das Infundibulum (Taf. 3, Fig. 1; Taf. 4, Fig. 9, 10 J) bildet die Seiten- winde und den Boden der mit zahlreichen Divertikeln versehenen Infun- dibularhéhle (Taf. 4, Fig. 10 Jh), welche natiirlich als Ausstiilpung des Zwischenhirnventrikels mit diesem zusammenhiingt. Median findet dieser Hohlraum seine Fortsetzung nach vorn unten in dem Hypophysenstiel, zu dem sich das Infundibulum verjiingt. Der Hypophysenstiel (Taf. 3, Fig. 1, 3; Taf. 4, Fig. 8, 9 Hst) cst sehr lang. Kin solcher kommt bei Fischen, wie ich schon bei Leucicorus erwihnte, seltener vor. Gottsche (1835, 8. 433, 434) war er bekannt; er fand ihn am lingsten an Lophius piscatorius und Clupea alosa, verglich ihn mit einer *‘Nabelschnur ” und nannte die an seinem Ende hingende Hypophyse eine “‘ Hypophysis pedunculata.” Andere Vertreter mit langem Hypophysenstiel fiihrt Malme (1891, 8. 29, 42, 43) in den Familien der A therimidae, Cyprimdae, und Characinidae an. Von den bis jetzt untersuchten Tiefseefischen ist MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES 241 Cyclothone hier zu nennen. Auffilliger aber als die Linge ist fiir den Hypo- physenstiel von Mixonus der Umstand, dass er hohl ist. So wurde er nur noch an Lophius piscatorius gefunden (Fritsch, 1878, 8. 23, Fig. 17, 18 Taf. Il). Wiah- rend sich Mritsch tiber die Natur des Schlauches nicht niher ausdriickt, muss ich bemerken, das der vorliegende durchaus nervdsen Charakter aufweist. Die Hypophyse (Tat. 3, Fig. 1 H) liegt zwischen Basi- und Praesphenoid. Diese Lage ist bei Fischen, ob mit kurzem oder langem Hypophysenstiel, allgemein. Haller (1898, S. 60) fand sie so urspriinglich an Selachierem- bryonen, konstatierte aber dann, dass dieser Zustand verloren geht und die Sphenoidalknorpel miteinander verwachsen. Bei den Salmoniden bleibt jedoch nach der Angabe desselben Autors die embryonale Lage der Hypo- physe zeitlebens erhalten. Das Griibchen, das die Hypophyse aufnimmt, ist beii:den Characinidae seicht, bei den Cyprinoidae tief (1891, S. 43). Mia- onus gleicht in dieser Beziehung den ersten. Seine Hypophyse hat iiberdies nicht die gewohnte Bohnenform, sondern die einer flachgedriickten Linse. Der Saccus vasculosus (Taf. 3, Fig. 1, 3; Taf. 4, Fig. 11 Sac) ist unbeden- tend und scheint zwischen den miichtigen Lobi inferiores nahezu ganz unterdriickt. Das Lumen des Infundibulum findet in ihm zentral seine Fortsetzung; die Wiande sind reich gefaltet und mit vielen Blutgefiissen versorgt. Noch vor dem Ende der Lobi inferiores, die medial bis zur Bertihrung aneinander treten, in einen ganz schmalen Spalt eingeengt verschwindet er. Die Lobi inferiores (Taf. 3, Fig. 1, 3; Taf. 4, Fig. 10, 11 Li) von denen ich schon gelegentlich der Lobi laterales und des Saccus vasculosus einiges vorausgeschickt habe, biegen aus ihrer seitlichen Lage im hinteren Teile gegen die Mediane unter das Mittelhirn ein. Sie sind sehr voluminis, reichen hinten bis unter das Hinterhirn und iibertreffen den Sacecus an Linge. Dies letztere trifft. bei Knochenfischen vielleicht ebensohiufig zu wie das umgekehrte; wenigstens gehéren von den 27 Familien, die Malme (1891) in seiner Arbeit anftihrt 10 der ersten und 17 der anderen Art an; bei den iibrigen halten sich Lobi inferiores und Saccus ungefiihr das Gleich- gewicht. Im Inneren eines jeden Lobus ist ein kleiner Hohlraum, der in Anbetracht der grossen Hirnmasse seines Triigers leicht tibersehen werden kann. Das Ende der Loben weist dorsal und ventral eine kurze Lings- furche auf. Noch eines muss hier in Bezug auf die Lobi inferiores und laterales erwihnt werden, dass sie niimlich trotz ihrer guten Entwicklung keine 16 242 MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES scharfe Abgrenzung nach vorne oder hinten zeigen; dadurch aber biissen sie den gewohnten Jobuliiren Charakter, an dessen Stelle der ausgesprochen wulstige tritt, vollkommen ein. Ich argwéhnte anfangs, ob dieser Umstand nicht eine Folge der Schrumpfung sei, iiberzeugte mich alsbald von der Unzulinglichkeit meiner Annahme dadurch, dass ich nach dem Oeffnen der Cranialhiéhle anderer Fische ebendieser Sammlung jene obenerwiihnten Teile deutlich lobulir ausgepriigt vorfand. DAS MITTELHIRN. Von einem normalen Teleostier-Mittelhirn schreibt Hdinger (1904, S. 121), dass es enorm gross ist und dass nur die Oblongata, welche alle die gewaltigen Hirnnerven des Fisches aufzunehmen hat, ihn an Ausdehnung erreicht. So trifft es auch bei Mizonus caudalis zu. Namentlich sind es die Lobi optici (Taf. 3, Fig. 1, 2,3; Taf, 4, Fig. 11 Lop), die diesem Gehirnteile jene Grisse verleihen. Das Gehirn erreicht in ihnen das Maximum an Breite, so dass sie auch in der Ansicht von unten vorragen. Auffallend an ihnen ist, dass sie jeder kugeligen Ausbildung, ebenso wie die bereits genannten Loben des Zwischenhirns entbehren und ganz flach sind. Dies gilt von ihrer ganzen Linge, namentlich aber vom hinteren Teil, wo auch der Grund sehr nahe liegt: es ist der von oben her auf ihnen lastende Stammteil des Kleinhirns, der sie so flachdriickt, Die mediane Longitudi- nalfurche ist deutlich ausgebildet. . Der Torus longitudinalis (Taf. 4, Fig. 11 Tol) verléiuft median in der Decke des intralobuliren Hohlraumes. Hr ragt ganz frei hinein wie bei Salmo fario und besteht aus zwei Wiilsten, denen wiederum lateral aussen je ein sekundirer Wulst aufsitzt. Hin solcher werwulstig r Torus ist an Fisch- gehirnen bis jetzt mcht beobachtet worden. Die Hirnmasse des Tectum opticum nimmf von vorn nach hinten ab. Ihr gewaltiges Volumen im vorderen Teile wiirde ich ebenso wie Handrick (1901, S. 8) “lediglich auf die excessive Ausbildung der Augen” zu- riickfiihren, wenn nicht GYerse (1904, S. 631) gezeigt hitte, dass auch bei kleinen Augen viel Hirnmasse im Tectum opticum vorkommen kann. Die Abnahme der Hirnsubstanz im riickwirtigen Teile des Tectum habe ich wie bei Leucicorus als Unvollkommenheit desselben bei Fischen mit vorgestiilpten Stammteil des Cerebellums erkannt. Hier ist tibrigens die Reduktion nicht so stark, da die Lobi optici der Valvula cerebelli und dem Kleinhirn nach rechts und links ausweichen, um nicht ganz unterdriickt zu werden. ! MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES 243 Wenn ich erkliren muss, dass das intralobuliire Lumen trotz der Grisse des Mittelhirndaches gering ist, so ist dies folgenden Faktoren zuzuschreiben : ausser der bereits erwihnten miichtigen Hirnmasse der Lobi optici, der stark gehobenen Pars peduncularis, und der Valvula cerebelli. Die Valvula cerebell (Taf. 4, Fig. 12 Vale) erstreckt sich bis in die Mitte des Hohlraumes und wird seitlich von den Tori senucirculares (Taf. 4, Fig. 12 Tse) zur Halfte eingeschlossen. Diese letzten sind schwach entwickelt und zeigen ebensowenig wie bei Leucicorus eine Spur dusserer Umgrenzung. Es ist wohl anzunehmen, dass sie in den Seitenteilen des Tectum opticum auf- gehen. Der Boden der Hoéhlung ist, wie oben erwihnt, durch die Pars peduncularis (Taf. 4, Fig. 10) gehoben. Eine median in ihm verlaufende Furche (Taf. 4, Fig. 11 Lf) ist vorhanden. Ueber eine solche schreibt Hand- rick (1901, S. 9): “Im Boden des durch die Lobi optici gebildeten Hohl- raumes, tiber der Pars peduncularis des Mittelhirns beginnt gleichfalls eine Lingsfurche, welche bei Argyropelecus wiihrend ihres Verlaufes nach hinten zuniichst mit dem hinteren Teile des Ventriculus tertius kommuniziert und dann als Aquaeductus unter dem Kleinhirn verlaufend in den Ventriculus quartus miindet (Taf. 1, Fig. 6-10; Lf).” Ebenso fand es Guerse (1904, S. 632) bei Cyclothone. Beide Autoren weisen auf Stieda (1868, S. 25) hin, der bei Gadus lota auch von jener Furche spricht ; sie soll dort aber bereits unter der Valvula cerebelli verschwinden. Durch diese Worte wire man zu der Annahme verleitet, dass hiedurch die Verbindung zwischen dem dritten und vierten Ventrikel aufgehoben wird. Dem ist nun nicht so; denn derselbe Autor schreibt eine Seite friiher (S. 24): “Unter diesem mit einer Liings- furche versehenen Kérperchen,” er meint die Valvula cerebelli, ‘‘ wird die Verbindung des vierten Ventrikels mit der Héhle des Sehlappens vermittelt.” Ich verstehe die Worte Steda’s dahin, dass gerade so, wie bei Mizenus cau- dalis, die wenn auch tiefe Lingsfurche im Boden des Ventriculus tertius ver- schwindet, nur aber als solche, wahrend das iibrige Lumen des Ventrikels auf einen schmalen unscheinbaren Kanal von rundem Querschnitt, den Aguae- ductus Sylvii, reduziert wird. Die Abbildungen der Querschnitte seitens Handrick’s (1901, Taf. 1. Fig. 7, 8, 9) und seitens Gierse’s (1904, Taf. XVI, Fig. 20) weisen auch eine Furche im Boden des Aquaeductus auf, die bei Mixonus nicht vorhanden ist. 244 MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES DAS HINTERHIRN Taf. 3, Fig. 1, 2, HH von Mixonus ist der miichtigste unter allen Teilen des Gehirnes. Es vermag uns aber weniger Interesse abzugewinnen, da es nach dem gewohnten Typus ” eines “Cerebellum trilobatum Abweichungen zu zeigen. Die schon Camper bekannten drei Teile, der gebaut ist, ohne irgendwelche besondere unpaare mittlere und die paarigen seitlichen, stehen im umgekehrten Ver- hiltnisse als bei Leucicorus, denn die letzten tibertreffen den ersten an Masse und Linge. Der Stanmteil (Taf. 3, Fig. 1,2; Taf. 4, Fig. 12, 13 Ctr) hat regelmiissige Halbkugelform. Sein Vorderende dringt sich (vergl. 8. 242) zwischen die auseinanderweichenden hinteren Partien der Lobi optici hinein. Das Hin- terende ist nicht verschmiilert, sondern rund, und geht in den Lobi poste- riores (Taf. 4, Fig. 18 Lpo) auf, denn diese “schieben sich,’ um Gierse’s Worte zu gebrauchen, “zwischen Stammteil und die in ihrem vorderén Abschnitte in der Mitte zusammenhiingenden Seitenteile und stellen so eine Verbindung zwischen den Teilen des Kleinhirns her.” Die Seitenwiilste (Taf. 3, Fig. 1. 2, 3; Taf. 4, Fig. 13 Tl), gehen vorne allmihlich aus dem Stammteil hervor und sind daselbst nicht gentigend abgegrenzt. Dennoch kann man annehmen, dass sie ihren Anfang unge- fihr in der Mitte des Stammteiles haben. Sie biegen seitlich nach unten um und reichen weit tiber das Ende desselben hinaus. Lings- oder Querfurchen sind an der Oberflache des Kleinhirns nicht wahrzunehmen. Die Valvula ceredelli (Taf. 4, Fig. 12 Vale) gehért zu den mittelmissig entwickelten und iiberschreitet nicht das hintere Drittel des Tectum opticum. Dort, wo ihr die Lobi optici Platz machen (vergl. 8. 243) wiirde sie eigentiim- licherweise fre: liegen, wenn nicht das Cerebellum hier den Abschluss bildete. DAS NACHHIRN Taf. 3, Fig. 1, 2, 3; Taf. 4, Fig. 13 NH von Mixonus hat viel Aehnlichkeit mit dem des Leucicorus lusciosus. Es umgibt mit seiner Hirnsubstanz, die ganz gewaltig ist, den vierten Ventrikel (Taf. 4, Fig. 12, 13 Vy). Von seinen besonders differenzierten Teilen sind nur die Loli posteriores (Taf. 3, Fig. 1, 2, Taf. 4, Fig. 13 Lpo) zu nennen. Sie setzen bereits zwischen dem Stammteil und den Seitenwiilsten des Cere- MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES 245 bellums ein. Eine innere Abgrenzung dieser drei Teile gegeneinander ist nicht zu erkennen ; sie sind zu einer einheitlichen Masse verschmolzen und nur dussere Anschwellungen zeigen, wo sie beginnen. Unmittelbar hinter dem Stammteil des Kleinhirns éffnet sich der geriumige Ventriculus quartus nach oben und wird nur durch ein Choroidalgeflecht bedeckt. Die Lobi posteriores verlaufen zu seinen beiden Seiten und man sieht, dass sie mehr als mit der Halfte der gesammten Hirnmasse des Nachhirns an dem Aufbaue des letzten beteiligt sind. Nach hinten nehmen sie an Hohe ab, gehen dafiir mehr in die Breite und erlangen schliesslich die Ausbildung von dicken gegeneinander geneigten Platten, die nach der Mediane bis zur Beriihrung zusammenrticken, ohne indessen miteinander zu verschmelzen. Das “Thal” Gottsche’s ist demzufolge wihrend ihres ganzen Verlaufes schén zu beobach- ten. Die geriiumige Rautengrube, die sich unterhalb der Beriihrungstelle der Lobi posteriores in einen schmalen Kanal verwandelt hat, erweitert sich hinten nochmals und zwingt jene Loben zum Auseinanderweichen ; kurz vor den Ursprungsstellen des Vagus verschwinden sie giinzlich. Lobi vagales und ein Tuberculum impar sind bei Mixonus nicht vorhanden. Das Gehirn von Bassozetus nasus. Taf. 5, Fig. 1-8, Taf. 6, Fig. 9-17. Auch dieses Gehirn zeigt sich in der Totalansicht (Taf. 5, Fig. 1-3) weniger gut mit dem normalen Physostomentypus iibereinstimmend. LEiner- seits muss die langgestreckte Form, dabei aber andererseits die Zusammen- schiebung der einzelnen Teile auffallen. Auch hier verteilen sich die Ursprungsstellen der zehn Fischgehirnnerven auf folgende Weise: I, I, III, IV, V+VII+VIII, VJ, IX+X. Auf die Ursache dieser Verhiltnisse hinzuweisen hatte ich bereits bei den vorhergehenden Fischen Gelegenheit und fiige hier die Fig. 8 auf Taf. 5 bei, welche die geriiumigen Sacculi zu den Seiten der Medulla oblongata deutlich zeigt. Die Cranialhthle dieses Fisches bietet dem Gehirne bei weitem mehr Platz als dieses namentlich mit dem vorderen und mittleren Teile in der Tat einnimmt. DAS VORDERHIRN Taf. 5, Fig. 1, 2, 3 VH ist verhiltnismissig gross. Die Unterscheidung seiner Bestandteile, die Tracti olfactorii ausgenommen, liesse sich wohl auch mit der Lupe bereits durchfiihren. 246 MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES Die Hauptmasse stellen die deutlich lobulir ausgebildeten Corpora striata (Taf. 5, Fig. 6; Taf. 6, Fig. 9 Cs) vor, wihrend die Bulb: (Taf. 5, Fig. 5 Bol) und die Zracti olfactorii (Taf. 5, Fig. 6 Trol) ganz in den Hintergrund treten. Die letzten sind iiusserst kurz, nur an mikroskopischen Priparaten erkennbar, Aafiir aber durch eine deutliche Grenze von den Streifenhiigeln geschieden. Nach hinten nehmen aus ihnen die Pedunculi cerebri ihren Anfang, nach vorne gehen sie ohne Begrenzung in die unansehnlichen Bulbi olfactora tiber, die sich alsbald zu den Riechnerven (Taf. 5, Fig. 1, 2, 3, I) verjiingen. Fiir den Ventriculus communis (Taf. 5, Fig. 5, 6, 7 Ve) bleibt im Vorderhirn nur wenig Platz in Form der bekannten T-férmigen Spalte tibrig. Dadurch, dass das ihn deckende Pallium vorne frontal mit den Bulbi verwiichst, findet er einen vollkommenen Abschluss, ohne etwa in die Cranialhéhle frei auszu- miinden (Argyropelecus, Leucicorus, Mixonus). Das Vorderhirn von Bassozetus gehort unzweifelhaft dem Salmonidentypus an ; eine kleine Abweichung zeigt es insofern, als die Bulbi olfactoru dorsal mit einander verschmolzen sind (Taf. 5, Fig. 5). DAS ZWISCHENHIRN Taf. 5, Fig. 1, 2 ZH liegt wie iiberall dort, wo nicht michtige Lobi optici darauf lasten, ganz frei (Taf. 5, Fig. 1, 2 ZH). Der Epithalamus besteht aus dem Plexus chorioide und den Gangha habe- nulae. Der erste (Taf. 6, Fig. 9 Plch) kommt an Grisse dem von Leucicorus nahezu gleich, nimmt viel Raum fiir sich in Anspruch, weshalb die letzten zuriicktreten miissen. Die Ganglia habenulae (Taf. 5, Fig. 7 Gh) sind iiusser- lich kaum angedeutet, an Querschnitten im Mikroskope sehr leicht erkenn- bar. Das rechte liegt weiter hinten als das linke und ist auch etwas kleiner als dieses. Die Habenularganglien sind also hier assymmetrisch. Die mediane Kommissur, welche die Haubenganglien miteinander verbindet, bildet eine seichte Furche, die nach hinten an Tiefe abnimmt und endlich dort, wo die Ganglien in die Seitenteile des Thalamus tibergehen, sich ganz ausflacht. DER PINEALAPPARAT von Bassozetus weist ein Pineal- und ein Parapinealorgan auf. An dem ersten lassen sich die gewohnten drei Teile unterscheiden: 1. Der proxi- male Endteil (Taf. 5, Fig. 7 E’st); er geht als ein kegelférmiger Zapfen aus dem rechten Haubenganglion nahe der Mediane hervor, ist massiv, histolo- MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES 247 gisch analog seinem Ursprungsherd zusammengesetzt. 2. Der Stiel (Taf. 5, Fig. 1, 2, 3, 4, 6; Taf. 6, Fig. 9 Est); er ist diinn aber lang, verliuft median, reicht weit iiber die Bulbi olfactorii hinaus und ist dorsoventral zusammengedriickt, Kr zeigt durchwegs nervésen Charakter und schwillt an seinem Ende an. 3. Der distale Endteil (Taf. 5, Fig. 1, 2,3; Taf. 6, Fig. 11, 12 E); er ist infolge des langen Stieles weit vorgeschoben und an die Innenseite des Schideldaches in der Gegend der sogenannten Epiphysar- spange befestigt und hat eirunde Form. Sein histologischer Bau lisst zwei von einander verschiedene Zonen, eine grob- und eine feinkérnige, deutlich unterscheiden. Die erste (Taf. 6, Fig. 11, 12 Z) bildet den Kern des End- teiles; in ihr sieht man den Stiel des Pinealorganes (Taf. 6, Fig. 11 Est), der hinten unten eingetreten ist, beinahe bis zur Spitze verlaufen. Gang- lienzellen (Taf. 6, Fig. 12 G) treten in ihrem Inneren spirlich auf, grup- pieren sich dagegen schén an der Peripherie. Hier sind sie sehr zahlreich, schliessen die zentrale Partie gleichsam in eine einschichtige Kapsel allseits ein und senden Fortsitze nach innen. Die tiussere Zone (Taf. 6, Fig. 11, 12 z) ist nicht arm an Nervenelementen. Bipolare Ganglienzellen zu langen Ketten aneinandergereiht findet man daselbst vor. Wenn andere Autoren den nervisen Charakter des Pinealorganes bei manchen Fischen durch den drusigen verdriingt sehen, so muss ich von Bassozetus das Gegenteil behaupten. Von enem Lumen im Innern ist keine Spur vorhanden; es ist eben durch die zentrale Partie, die tibrigens viel Aehnlichkeit mit der sonstigen Hirnsubstanz hat, ausgefullt. Ks freute mich, den fiir Teleostier neuen Befund Gierse’s (1904, S. 623), den er an Cyclothone machte, auch an Sassozetus konstatieren zu koénnen. Jener Autor fand nimlich, dass an dem Endteil des Pinealorganes, dort, wo es in den Stiel iibergeht, zwei feine Nervensttimmchen abgehen und Jiings des Stieles nach dem Gehirne ziehen. So ist es auch hier (Taf. 5, Fig. 4 En). Die Hauptsache aber, um die es sich in beiden Fiillen handelt, ist die Zweizahil und die Symmetrie dieser Pinealnerven, denn sonst ist bei Teleostiern nur von einem die Rede. Fiir Saurier sind wohl zwei von Kiinckowstriéim (1893, S. 266, 267) nachgewiesen; in einem Falle fand er noch iiberdies einen unpaaren TZractus pinealis vor. Das Parapinealorgan des Bassozetus (Taf. 5, Fig. 4, 7; Taf. 6, Fig. 9 P) ist stark verkiirzt, unscheinbar; erreicht es ja nicht einmal das Vorderende der Streifenhiigel. Dennoch gelang es mir aber, jene gewagte Behauptung Handrick’s, mit der er Hill’s Erklirung des Parapinealorganes auch fiir 248 MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES Argyropelecus bestiitigt (vergl. S. 239) und fiir die ich mich auf Grund der Tatsachen bei Mzonus nicht recht entschliessen konnte, hier klar und deutlich verwirklicht zu sehen. Dort, wo der kegelformige Anfangsteil des Pineal- organes in den Stiel iibergeht (Taf. 5, Fig. 7), gerat er unter das Parapineal- organ, denn dieses vollkommen gleich einer Falte des Plexus Chorioideus — es liegt auch in dessen Bereiche — stiilpt sich nach oben vor und nimmt alsbald die Form eines Schlauches mit einschichtiger, epithelialer Wand an (Taf. 5, Fig. 4 P). So kann auch ich fiir Bassozetus erklitren, dass das Para- pinealorgan eine Ausstulpung des Veniriculus communis ist. Die kurze Strecke, welche Pineal- und Parapinealorgan gemeinsam ver- laufen, sind sie itibereinander gelagert, das erste median tiber dem zweiten. Der Thalamus, seiner Hauptmasse nach aus den Pedunculi cerebri beste- hend, ist gut entwickelt und engt den Ventriculus tertius auf einen noch kaum merklichen Spalt ein. Die ihm seitlich aufsitzenden Lobi laterales (Taf. 5, Fig. 1; Taf. 6, Fig. 10 Ll) halten in Grisse den inferiores (Taf. 5, Fig. 1,3; Taf. 6, 10, 14 Li) das Gleichgewicht. Querschnittbilder sprechen eher fiir enge Zusammengehorigkeit der beiden, wenngleich das Aeussere — sie sind schiirfer voneinander abgegrenzt als bei den vorhergehenden Fischen — mehr auf ihre Selbstiindigkeit schliessen lisst. Der Hypothalamus steht an Volumen dem Thalamus nicht nach. Waren es jedoch bei Mixonus die Lobi inferiores, die an seiner Zusammensetzung den Hauptanteil hatten, so ist és hier der Saccus vasculosus (Taf. 5, Fig. 1, 3; Taf. 6, Fig. 10, 14 Sac). Seine vordere Grenze, die in der Regel durch das Chiasma Nervorum opticorum gegeben ist, wird hier durch eine solch unscheinbare Anschwellung des Infundibularteiles angedeutet, dass man seine Not damit hat, sie tiberhaupt zu erkennen, dazumal auch die Sehnerven iiusserst zart sind. Deutlicher setzt der Hypophysenstiel (Taf. 5, Wig. 1,3; Taf. 6, Fig. 13 Hst) gleich hinter dem Chiasma am Infundibulum an; er ast lang, solid, und von nervdsem Charakter. Die in der Pituitargrube sitzende kleine Hypophyse (Taf. 5, Fig. 1; Taf. 6, Fig. 13 H) wird von Bindegewebe umgeben. Ihr Stiel durchsetzt diese Hiille, sowie auch ihre periphere Zone und lést sich im Inneren in zahlreiche Fasern auf (Taf. 6, Fig. 13). Der Ventriculus tertius erweitert sich im Infundibulum zu einer Héhlung / mit glatten Wiinden, die nur drei Divertikel aufweist: die ersten zwei gehen in die Lobi inferiores, der dritte setzt sich in den Saccus vasculosus fort. Ich habe bereits oben die gewaltige Grésse dieses Zwischenhirnteiles MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES | 249 erwihnt; die beste Vorstellung von ihr gewinnt man beim Anblicke des Gehirnes von unten (Taf. 5, Fig. 3); es fehlt gar nich viel, so verdeckt er, wie es bei Leucicorus der Fall ist, die iiber ihm gelagerten Teile des Gehirns, zu denen auch der breiteste, das Tectum opticum, gehdrt. DAS MITTELHIRN. Taf. 5, Fig. 1 MH Ich bin iiberzeugt, dass mich diese Gehirnpartie von Bassozetus, wenn ich nicht entkalkt, sondern die Schiideldecke gedffnet hatte, ebenso iiberrascht hatte, wie es bei Leucicorus der Fall war. Wenn auch das Tectum opticum hier (Taf. 5, Fig. 1, 2, 3; Taf. 6, Fig. 10,14 Lop) um ein weniges héher und differenzierter erscheint, ist es doch nicht das den Fischen typische. Seine Hirnmasse besitzt nur miissige Dicke, ist hinten unvollkommen, auf eine diinne Membran reduziert, was nicht ausser acht zu lassen ist, da, wie ich schon oben bemerkte, auch die Sehnerven von einer unvergleichlichen Zart- heit sind. Die schwach konvexe Wolbung der Loben bringt es auch mit sich, dass die diussere mediane Lingsfurche sehr seicht ist. Der unter ilr innerhalb des intralobuliren Hohlraumes verlaufende TZorus Jlongitudinalis (Taf. 6, Fig. 10 Tol) wt auch unansehniich. Er besteht aus zwei Wiilsten. Die Pedunculi cerebri erreichen die Medianebene bei weitem nicht, und lassen auf diese Weise viel Raum fiir die Lingsfurche im Boden des Hohlraumes (Taf. 6, Fig. 14 Lf). Sie ist breit und bleibt auch unter der Valvula cere- belli, im Aquaeductus Sylvii erhalten, bis sie in der Gegend der Tori semi- circulares verschwindet. Diese letzten (Taf. 5, Fig. 1; Taf. 6, Fig. 14 Tse) sind leicht als das, was sie Mayser (1882, S. 271) mit Recht nennt, nimlich “partielle Verdickungen der Basis und der hinteren Seitenwand des Mittel- hirns in seinen hinteren Abschnitten”’ zu erkennen. Sie umgreifen die Valvula cerebelli nahezu vollkommen, treten auch ein wenig frei hervor, dazumal die Lobi optici nicht hinreichen, um sie zu decken. DAS HINTERHIRN Taf. 5, Fig. 1, 2 HH ist ein Cerebellum trilobatum, bestehend aus dem unpaaren Stammteil (Taf. 3, Fig. 1, 2; Taf. 6, Fig. 15 Ctr) und den paarigen Seitenwiilsten (Taf. 5, Fig. 1, 2, 8; Taf. 6, Fig. 15 Tl). Beide Teile sind ungefihr gleich stark entwickelt und iiusserlich von einander zu unterscheiden. Der Stammteil wilbt sich rostralwirts vor, caudalwirts verschmiilert er sich unbedeutend, senkt sich aber nicht wie sonst in die Tiefe, sondern ragt 250 MORPHOLOGIE DES TIEFSEHFISCHGEHIRNES frei hervor. igentiimlich fand ich die daselbst befindliche Hohlung (Taf. 6, Fig. 15 Cav). Die Seifenwiilste weisen nichts besonders Erwiihnenswertes auf. Die vom Stammteil nach innen entspringende Valvula cerebeli (Taf. 6, Fig. 14 Valc) ist gross. Vorne ragt ein schmaler Spalt zwischen ihre sym- metrischen Hilften hinein und wiirde sie vollkommen voneinander trennen, wenn nicht eine schmale Verbindung im dorsalen Teil bestiinde. DAS NACHHIRN Taf. 5, Fig. 1, 2,3; Taf. 6, Fig. 17 NH zeichnet sich wie bei allen Fischen durch seine Lange aus. Dem basalen Teile liegen dorsal die Lobi posteriores (Taf. 5, Fig. 1, 2; Taf. 6, Fig. 15, 16 Lpo) an. Sie zeigen insofern ein von den bis jetzt untersuchten Tiefsee- teleostiern abweichendes Verhalten, als sie vorn mcht das Hinterende des Kleinhirnstammteiles zwischen sich aufnehmen, da jenes bei der schon erwiihnten Lage mit ihnen gar nicht in Beritihrung kommt. Sie sind hier weit auseinander geriickt. Nach hinten kommen sie einander niher, bis sie total verschmelzen. Von da an beginnt auch ihre michtige Hirnmasse, die in der ganzen Linge ein seichtes “Thal” aufweist, einen Wulst gegen den basalen Teil der Medulla vorzutreiben. Ganz hinten trennen sich die Lobi posteriores abermals und lassen in die Tiefe der Rautengrube hineinblicken. Im Boden der letzten gewahrt man zwei Wiilste, die sich an mikroskopischen Priparaten als Lodi vagales erweisen (Taf. 6, Fig. 17 Lv). Sie riicken allmah- lich aus der Tiefe hervor, bis sie die Oberfliiche erreichen, ohne sich aber tiber sie zu erheben. Zusammenfassung. Wenn ich nun die von mir an dem Gehirne der drei Tiefseefische Leu- cicorus lusciosus, Mixonus caudalis, und Bassozetus nasus gemachten Befunde iiberblicke, so ergibt sich folgendes: Alle drei Gehirne zeigen Formen, die vom Physostomentypus abweichen. Dies gibt sich einesterls darin kund, dass die Vorder- und Zwischenhirnteile eine vollkommene, normale Entfal- tung nicht erreichen, obgleich ihnen hiezu in dem langgestreckten und geriumigen Cranium genug Platz geboten wird; so geht den Lobi optici, laterales und inferiores tiberall die gewohnte lobulire Ausbildung ab. Die miachtigen Labyrinthe zu den Seiten der Medulla oblongata verursachen weniger die Zusammenschiebung der Hinterhirnteile, als die der Ursprung- MORPHOLOGIE DES TIEFSEEFISCHGEHIRNES 251 stellen der Nerven. Alle drei Gehirne haben langgestielte Hypophysen ; die Lobi vagales sind so gut wie nicht entwickelt. Im besonderen ist an Leucicorus erwihnenswert: Die Tracti und die Bulbi olfactorii gehen stark auseinander. Die dorsalen Teile der Corpora striata sind so gross, dass sie seitlich iiber die basalen heriiberhiingen. Der ungewohnlich hoch ausgebildete Plexus chorioideus fiillt den Ventriculus tertius und communis nahezu vollkommen aus. Trotz der Grisse der Augen sind die Nervi optici nur unscheinbar, das Tectum opticum auf eine niedrige, flache Hirnplatte reduziert. Die Seitenwiilste des Kleinhirns hingen mit den Lobi posteriores zusammen, so dass die letzteren als direkte Fortsetzung der ersteren nach hinten erscheinen. Mixonus caudalis zeigt trotz starker Riechnerven verhiltnismiissig kleine Bulbi olfactorii. In den Corpora striata ist abweichend von der Regel der basale Teil der breitere und grdssere, der dorsale dagegen nur ein niedriger Wulst. Das Chiasma Nervorum opticorum zieht als breite Nervenplatte unter dem Vorderhirn bis an das Vorderende der Streifenhiigel, wo erst die Auflésung in die beiden Sehnerven erfolgt. Das Parapinealorgan ist solid. Pineal- und Parapinealorgan verlaufen nebeneinander. Der Thalamus ist so stark entwickelt, dass er die Ganglia habenulae in ihrem riickwiirtigen Teile ganz umschliesst. Die Lobi laterales sind hier sicherlich nicht selb- stindige Bildungen, sondern differenzierte dorsale T'eile der Lobi inferiores. Bassozetus nasus weicht insofern von dem reinen Salmonidentypus im Vorderhirn ab, als die Bulbi olfactorii dorsal miteimander verschmelzen. Das Pinealorgan ist ein echt nervises Organ, das mit einer Driise nicht die geringste Uebereinstimmung hat. Es zeigt in seinem Inneren keinen Hohl- raum. Zwei symmetrische Pinealnerven sind hier vorhanden. Die nied- rigen Lobi optici erheben sich nur wenig tiber das Zwischenhirn. Ich habe alle diese Besonderheiten leichterer Uebersicht halber hier zusammengestellt. Ihre Anzahl] ist eine ziemlich grosse und zeigt zugleich, dass weitere Forschungen am Tiefseeteleostiergehirn nicht erfolglos sein diirften. Es wird noch groésserer Tierreihen bediirfen, um die einzelnen Resultate verallgemeinern und von Folgen der Anpassung an das Leben in der Tiefsee sprechen zu kénnen. Jetzt wiiren derlei Schliisse vorzeitig. 1883. 1884. 1813. 1887. 1902. 1774. 1890. 1882. 1883. 1828. 1892. 1904. 1878. 1878. 1899. 1898. 1904. 1888. 1835. 1886. 1898. 1898a. 1901. 1891. 1894. LITERATURVERZEICHNIS 253 LITERATURVERZEICHNIS. Auporn, F., Untersuchungen iiber das Gehirn bei Petromyzonten. Zeitschr. f. wiss. Zool. Bd. XXXIX. Ueber die Bedeutung der Zirbeldriise. Ibid. Bd. XL. Arsaky, A., De piscium cerebro et Medulla spinali. Bearp, J., The parietal eye in fishes. Nature, July 14. Braver, A., Ueber den Bau der Augen einiger Tiefseefische. Verh. d. deutsch. Zool. Ges. Camper, Mémoires de mathematique. Tome IV. Paris. CarrizrE, Neuere Untersuchungen tiber das Parietalorgan. Biolog. Zentralbl. Bd. IX. Cartiz, J., Recherches sur Ja gland pinéale (Epiphysis cerebri) des Plagiostimes, des Ganoides et des Teléostéens. Archiv. d. Biolog. Vol. III. Ueber das Gewebe der Epiphyse von Plagiostomen, Ganoiden und Teleos- tiern. Zeitschr. f. wiss. Zool. Bd. XXXIX. Cuvier, G., et Valenciennes, A., Histoire naturelle des poisson. TomelI. Paris. Epincer, L., Zwolf Vorlesungen iiber den Bau der nervosen Zentralorgane. 5. Aufl. Leipzig. Vorlesungen iiber den Bau der nervésen Zentralorgane des Menschen und der Tiere. 6. Aufl. Leipzig. Euters, E., Die Epiphyse am Gehirn der Plagiostomen. Zeitschr. f. wiss. Zool. Bd. XXX, Suppl. Fritscu, G., Untersuchungen iiber den feineren Bau des Fischgehirns. Berlin. Garmay, S., Reports on an Exploration off the West Coasts of Mexico, ete. XXVI. The Fishes: Mem. Mus. Comp. Zool. Harvard College, Vol. XXIV. Gaver, E., Zirbel, Parietalorgan und Paraphysis. Ergebnisse der Anatomie und Entwicklungsgeschichte. Bd. VII. Giersz, A., Untersuchungen tiber das Gehirn und die Kopfnerven von Cyclothone acclinidens. Morph. Jbch. Bd. XXXII. Goronowitscu, N., Das Gehirn und die Cranialnerven von Acipenser ruthenus. Morph. Jbech. Bd. X. Gortscur, M., Vergleichende Anatomie des Gehirns der Gratenfische. Miiller’s Archiv. Ginruer, A., Handbuch der Ichthyologie. Wien. Hatter, B., Untersuchungen iiber die Hypophyse und die Infundibularorgane. Morph. Jbch. Bd. XXV. Vom Bau des Wirbeltiergehirns. Morph. Jbch. Bd. XXVI. Hanpricx, K., Zur Kenntnis des Nervensystems und der Leuchtorgane von Argy- ropelecus hemigymuus. Zoologica, Bd. XIII., 32. Heft., 1. Lief. Hi, Ch., Development of the epiphysis in Coregonus albus. Journal of Morph. Vol. V. The epiphysis of Teleosts and Amia. Ibid. Vol. IX. 1893a. 1891. 1864. 1882. 1888. 1882. 1883. 1883-4. 1885. 1886. 1881. 1884. 1885. 1891. 1849. 1868. 1873. 1893. 1893. 1894, 1899. 1832. 1898. 1902. LITERATURVERZEICHNIS Kxaatscu, H., De cerebris piscium ostacanthorum aquas nostras incolentium. Dis- sertatio. Halis Saxonum. 1850. Kurncxowstrom, A. von, Beitrige zur Kenntnis des Parietalauges. Zool. Jbch. Bd. VII, Abt. f. Anat. u. Ontog. der Tiere. Die Zirbel und das Foramen bei Callichthys (asper und littoralis). Anat. Anz. Bd. VIII. Jena. Mame, G., Studien itiber das Gehirn der Knochenfische. Bihang Till K. Svenska Vet. Akad. Hadlingar. Bd. XVII, Afd. IV, Nro. 3. Mayer, F., Ueber den Bau des Gehirns der Fische in Beziehung auf eine darauf gegriindete Einteilung dieser Tierklasse. Verh. d. kais. Leopoldino-Carolinischen deutsch. Akad. der Naturf. Dresden. Bd. XXX. Maysrer, P., Vergleichend-anatomische Studien tiber das Gehirn der Knochenfische mit besonderer Beriicksichtigung der Cyprinoiden. Ztschft. f. wiss. Zool. Bd. XXXVI. Owss4nnrkow, Ph., Ueber das dritte Auge bei Petromyzon fluviatilis nebst einigen Bemerkungen iiber dasselbe Organ bei anderen Tieren. Mem. de l’Acad. imper. des sciences de Petersbourg. VII. Serie, Tom. XXXVI. Rasi-Rickuarp, H., Zur Deutung und Entwickelung des Gehirns der Knochen- fische. Arch. f. Anat. u. Physiolog. Anat. Abt. Das Grosshirn der Knochenfische und seine Anhangsgebilde. Ibid. Weiteres zur Deutung des Gehirns der Knochenfische. Biol. Zentralbl. Bd. 2H, Das Gehirn der Knochenfische. Ibid. Bd. IV. Zur Deutung der Zirbeldriise. Zool. Anz. Bd. IX. Rerzius, G., Das Gehororgan der Wirbeltiere. I. Gehérorgan der Fische und Amphibien. Stockholm. SacEMEHL, M., Beitrage zur vergleichenden Anatomie der Fische. I. Das Cranium von Amia calva. Morph. Jbch. Bd. IX. Beitriige zur vergleichenden Anatomie der Fische. III. Das Cranium der Characiniden nebst einigen Bemerkungen iiber die mit einem Weber’schen Apparate versehenen Physostomenfamilien. Ibid. Bd. X. Dasselbe. IV. Das Cranium der Cyprinoiden. Ibid. Bd. XVII. Srannius, H., Das peripherische Nervensystem der Fische. ~ Rostock. Strepa, L., Studien tiber das zentrale Nervensystem der Knochenfische. Zeitschr. f. wiss. Zool. Bd. XVIII. Ueber die Deutung der einzelnen Teile des Fischgehirnes. Ibid. Bd. XXIII. Srupnicka, F., Sur les organes pariétaux de Petromyzon planeri. Prague. Zur Morphologie der Parietalorgane der Cranioten. Referat von Vejdovsky im Zool. Zentralbl., Jhg. I. Zur Lésung einiger Fragen aus der Morphologie des Vorderhirns der Crani- oten. Anat. Anz., Jhg. 9. XXXVII. Ueber den feineren Bau der Parietalorgane von Petromyzon marinus L. Sitzungsber. d. kgl. béhm. Ges. d. Wiss. Math. naturwiss. Klasse. Prag. TreyrraNnus, Die Erscheinungen und Gesetze des organischen Lebens. Wiepersuem, R., Lehrbuch der vergleichenden Anatomie der Wirbeltiere. Vergleichende Anatomie der Wirbeltiere. 5. Aufl. Jena. TAFELERKLARUNG. 255 TAFELERKLARUNG. IN ALLEN FIGUREN AUF TAFEL 1-6 BEDEUTEN: AS .. . Aquaeductus Sylvii. ee Blutgefass. Bd . Bindegewebe. Bol. . Bulbi olfactorii. Ca . Commissura anterior. Cav . Hohlraum. Com . Kommissur. Cs . Corpora striata. Ctr . Stammteil des Kleinhirns. roe Pinealorgan. En . Pinealnerven. Est . Stiel des Pinealorgans. E’st . Stiel des Parapinealorgans. _ oe Furche. Diep e< Ganglienzellen. Gh . . Ganglia habenulae. H .... Hypophyse. Hst . Hypophysenstiel. Ch . Chiasma nervorum opticorum. ch . . Chiasma der Blutgefisse. aa Infundibulum. Jh . . Infundibularhohle. Lf . . Longitudinalfurche. Li . Lobi inferiores. Ll .... Lobi laterales. Lop . Lobi optici. . Lobi posteriores. Ly . . Lobi vagales. MH ... Mittelhirn. NH . Nachhirn. Eee iacas Parapinealorgan. Pa . Pallium. Plieh . Plexus chorioideus. Sac . Saccus vasculosus. sac . Sacculus. i eee ee Thalamus. T! . Seitenwiilste des Kleinhirns, Tol. . Torus longitudinalis. Trol . Tracti olfactorii. Tse. . Tori semicirculares. Vale . . . Valvula cerebelli. VH . Vorderhirn. + Ve . Ventriculus communis. VE cu cio Ventriculus tertius. Meet era: Ventriculus quartus. YAR BR a aie Zentrale Zone des Pinealorgans. TS a SRT Periphere Zone des Pinealorgans. ZH . . Zwischenhirn. | tap eae Nervus olfactorius. a . . Nervus opticus. Wiens aes Nervus trigeminus. VII . Nervus facialis. VII . Nervus accusticus. x . Nervus vagus. pele os h in ss A aataT sloeS oys¥ .nédo ov eomidoD-eyb ddeianaintol iT dionts wie ¥ betas sv sontides) exh tdoleualeoT = .£ .yil ih ‘ 0% morte?) sab divinaads — Ap) aie Le ike 2 me 53 ibe e ae J "ie ss OED vift jit; nee LEHANE) wo BAA ie ity M4¢2 wis eo 4 ; * jt A Nien) ka wire a | > ONT sk v7 De, fat) aM TTY 1, Og wey i ie he fe Bee z ee ie = ed se = a mt air ur y ; Bs HY) shih pie vy J re d si ed , WN PARK JH See hylete ay PON» op B. AY A ywenious) anartiied =i aisatS syiwV .nodo tev esmideD gh dfolanelstol = ois «wie .ontan ov puatidod) esb Jitolausgiasod A ftosS spyts ¥ Tare os Ov egavide)) coh sdvieneisicT . WonI0S wage | asidowdosiwX but -1stio V nov esas) tab 1s Hindyeron) _.+ oi ane g06 go V awards ermyio2 sib alymh HindsewaP -.8— sod o. £ Som WEsrgeV .antoy wsiitnodosiwS web owb dines <0 pra : oe Monl0® rye of rob ui mtidasipsterS eab dor atingomanys 7 yi a do nov ( 0 HSebeR Woy ( d nodmi soy ( & enlnoor® cals bun AsiitO resto) = pil 1 * ine 43 ei . : 31 3m8¥ “Ge | CONT OH TAFEL 1, Leucicorus lusciosus. Totalansicht des Gehirnes von oben. Vergr. 2fach. Totalansicht des Gehirnes von unten. Vergr. 2fach. Totalansicht des Gehirnes von der Seite. Vergr. 2fach. Querschnitt an der Grenze von Vorder- und Zwischenhirn. Vergr. 30fach. Querschnitt durch die Corpora striata. Vergr. 30fach. Querschnitt durch das Zwischenhirn vorne. Vergr. 30fach. Querschnitt durch das Zwischenhirn in der Mitte. Vergr. 30fach. Grosser Otolith aus dem Sacculus a ) von innen b ) von aussen c ) von oben Vergr. 13. TIEBFSEEFISCHGEHIRNES HH a 1 Bol VH ZH lop CtrTL Lpo NH : j ‘ rot scl | Hust Lt li LEUCIGORUS LUSGIOSUS Garman B. Meise! jth Seswe- ~~ ta ‘ais ie Ve ar TAFEL 2. Leucicorus lusciesus. Querschnitt durch das Zwischenhirn hinten. Vergr. 20fach . Querschnitt durch das Mittelhirn. Vergr. 20fach. Querschnitt durch das Hinterhirn. Vergr. 20fach. Querschnitt durch das Mittelhirn (hintere Partie). Vergr. 20fach. Querschnitt durch das Nachhirn. Vergr. 20fach. TIEFSEEFISGHGEHIRNES PLATE C2 B. Meise! lh Basan. “LEUGICORUS LUSGIOSUS Garman SS oJ e , re ee ie wre € silabses sano domtS rye .stie®@ tsb nov wetmisioD ash tdoianaleloT 1 .gfF dogté rgi9¥ “asdo nov searitiod aoh ddoizaniatoy <£ .3ff Asnté «yi TarEL 3. botified sob tdislaunietol 2.3 doxt08 ayrwY inodpstlg Woxtl otto) ttindowmu) 2 olf Honi08 “179 ¥_ digg on te idis® ob dew iidtoweu & gil rita Oa G40 » 102 sib domth MintoeiD. 3 HOME AHO MAG Wi oi) werivh sioioeho ‘ays ie BIHIOD sib sloyuh siintoeani) 20 colmunpeleont'l agb pollo wee zdun gy’ Iaith dauih Mtindezianyy 7 Hodte get, “ebony” deck inal Hinkenuy) da. da’ da’ da’ da" da da 4 ies — gS Soo PF ON a TAFEL 5. Bassozetus nasus. Totalansicht des Gehirnes von der Seite. Vergr. 8fach. Totalansicht des Gehirnes von oben. Vergr. 8fach. Totalansicht des Gehirnes von unten. Vergr. 8fach. Querschnitt durch das Parapinealorgan. Vergr. 520fach. Querschnitt durch die Bulbi olfactorii. Vergr. 20fach. Querschnitt durch die Corpora striata. Vergr. 20fach. Querschnitt durch die Ursprungsstellen des Pinealapparates. Querschnitt durch das Cranium. Vergr. 4fach. Vergr. 260fach. TIEFSEEFISGHGEHIRNES PLATES MIT HH VH EstGh ZH Lop Te Cir Tl kyo Hyp Hst It Lt Sue HH VHP GhiH Lop CtrTl’ Lpo NH B. Meisel lh Bosion BASSOZETUS NASUS Garman 9 wa4aTl nea aodesosanth fone avr ¥ .vidastoxiwX exh domih diimdoerant) doe yess'Y Auioy midlet sab doarh Niniowiend domes .wzieV «tepiolaoni eb doih Witoeyenyp Peionr00e. aniay .. Soll ilietilk eab doh dtinidpetonyy Ait: yap a niizoiaihi sah comb tisisaeu)) pe Wontat- ozo! anor aiiddaa4: aah tomb tincomeays eshi08 cit aotal wtiddosA ash dowh Mindearonp ol fe TAFEL 6. Bassozetus nasus. Querschnitt durch das Zwischenhirn. Vergr. 30fach. Querschnitt durch das Mittelhirn vorne. Vergr. 20fach. Querschnitt durch das Pinealorgan. Vergr. 50fach. Ein Teil des in Fig. 10 dargestellten Querschnittes. Vergr. 300fach. Querschnitt durch die Hypophyse. Vergr. 30fach. Querschnitt durch das Mittelhirn hinten. Vergr. 20fach. Querschnitt durch das Hinterhirn. Vergr. 20fach. Querschnitt durch das Nachhirn vorne. Querschnitt durch das Nachhirn hinten. Vergr. 15fach. Verg. 30fach. i" = _ . “Aupatross’ Ex. 1891. TIEFSEEFISCHGEHIRNES PLATE 6 B. Meisel lah. Basin. * BASSOZETUS NASUS Garman. ~~ ae al Memoirs of the Museum of Comparative Zodlogy ae AT HARVARD COLLEGE. Vout. XXX. No. 4. THE GALAPAGOS TORTOISES. BY SAMUEL GARMAN. WITH FORTY-TWO PLATES. CAMBRIDGE, U. S. A.: — Printed for tbe Museum. Janroany, 1917. TABLE OF CONTENTS. INTRODUCTION. ‘ (he loco orn aeate : ee : ; : a TESTUDO TABULATA, PuratE 2... : ‘ , : z i>. 270 _ TESTUDO MACROPHYES, Puatss 3-5 rn : : : : : . 273 _ TESTUDO VICINA, Puatss 6,7 . 7 Feber : é ; : . ate TESTUDO NIGRITA, Puiates 8-10 : 2 , : : : : : “216 TESTUDO MICROPHYES, Puates 11-20; Puate 38, Fic. 3 : : . . 280 TESTUDO CLIVOSA, PuiatTE 21... ; ‘ : , : : : 283 TESTUDO TYPICA, Puates 22, 34 ' 2 ; : ‘ : c~° (SBS | TESTUDO NIGRA, Puiates 22-33, 35, 36. : ; ; : . 286 TESTUDO ELEPHANTOPUS, Ptates 37, 38, Fic. 1, 2; Prates 39-42. . 290 EXPLANATION OF THE PLATES. » : v : THE GALAPAGOS TORTOISES. INTRODUCTION. A warm interest of the Museum authorities in the Giant land tortoises has led to the acquisition of so many notable specimens, through liberal exchanges and purchases, as to demand a revision of portions of the collection, especially of that portion directly pertaining to the Galapagos Archipelago. This forms the reason for the following article. It is based mainly on partial specimens, 1. €., carapaces and sternums, but it is thus that these tortoises are most generally known, thus that they are most widely, and most commonly represented in descriptions, figures, and collections. A few attempts have been made at complete characterization of the species by including the anatomical features; these were founded on single specimens, and the individuals of the species are found to differ too much to admit of accurate distinctions unless confirmed by averages secured from repeated dissections, for which much of the material and the labor has yet to be supplied. ; The Galapagos Islands form an isolated group in the eastern Pacific on the equator about 6° west of Ecuador, or in other words, they are situated between 89° and 92° of west longitude and between 1° 30’ south and 2° north latitude. The largest of them is about eighty miles long and at its widest is about fifty miles wide; from this the sizes vary to some that are mere points of rock or shoals. They are separated from the mainland by more than four hundred miles of deep sea, a thousand fathoms or more in depth. The wide separation from the continent, their considerable distances from one another, with great differences in altitudes and consequent variations in climate and fertility give them exceptional attractions in the eyes of naturalists. Here if anywhere they might hope to find the species of the flora and fauna distinct from those of the world around them and here it might be possible to trace their development and derivation. Questions of origin go back to the advent of the islands themselves; neither in case of lands, plants, nor animals have the questions been answered with any great degrees of satisfaction. Some authorities have decided that the islands are oceanic, that they never were connected with the continent, but were pushed up from the sea-bottom by the numerous volcanoes they contain, 262 THE GALAPAGOS TORTOISES. — This accounts for the islands alone. There are two methods of explaining the presence of the organisms: one by origin in place, another by accidental importa- tions. If originating in place similarities in climate and other conditions might be adduced to account for affinities, but accidental importations could hardly be relied on for the development of parallel harmonious series such as are now in place on the Galapagos. The accidents would be heterogeneous, and no reason appears for limiting them in time or numbers. Theodore Wolf, at one time Geologist of Ecuador, published an account of the Archipelago from actual study, adopted the oceanic theory and placed the appearance of the islands in the Tertiary and later, thereby enhancing the scientific interest because of the comparatively short period elapsing since the uplifts. He recognized affinities between the insular and the continental organisms, and found the rocks of the Galapagos basaltic and those of the highlands of Ecuador trachyte and andesite. A most able recent advocate of the continental theory of origin was George Baur, 1891. For months he made collections on a majority of the islands in preparation for a comparative account of the life and conditions. His conclusion was that the Archipelago included the tops of volcanic moun- tains of a greater area of land at one time part of the continent, later sunken below the sea-level. All the islands were formerly connected in a single large one which by continued subsidence divided into a number, the highlands that remained above the sea. Each of the latter in a long course of time developed peculiar races, eventually species, in its plants and animals, because the condi- tions were not identical. As proof especial stress is laid on the harmonious distribution of the organisms. Nearly every island has its own races, and their affinities commonly turn toward the continent. In this the idea of communica- tion and transportation between the islands is not considered. Discovery of the Archipelago is said to have been made by Berlanga, 1535. The Giant land tortoises, then found in immense numbers suggested the name Galapagos, previously applied in Spain to fresh-water tortoises, a designation which does not appear to have been entered on the maps as a name for the islands till nearly fifty years later. The Spaniards paid little attention to the territory. For a couple of centuries it was merely a place of call for a supply of fresh meat. Among the earliest visitors were the buccaneers, Dampier and Cowley, 1684— 97, who furnished accounts of portions of the group. Cowley published a map on which Spanish names were displaced by English, now perhaps the more widely known. Ecuador, the present owner, clings to the Spanish American THE GALAPAGOS TORTOISES. 263 names and the prospect is that they will ultimately prevail. The Archipelago was surveyed by Colnett, 1793. The chart, Plate 1, with the positions of the islands, directions of the currents, and the two series of names is sketched from that published under the direction of the Secretary of the Navy by the United States Hydrographic Office, 1915. From the directions of the currents it will be seen that affinities with South American organisms would be of the most natural imaginable, if dependence for origins-of the flora and the fauna were placed upon the marine drift. The Spanish names of the islands with their English equivalents are as follows: — San Cristobal (Chatham).. Espanola (Hood). Santa Maria (Floreana) (Charles). Santa Fé (Barrington). Santa Cruz (Chaves) (Indefatigable) (Porter’s). Tortuga (Brattle). Pinzon (Duncan). Isabela (Albemarle). Fernandina (Narborough). Rabida (Jervis). San Salvador (James). Marchena (Bindloe). Pinta (Abingdon). In what ever way the balance of the fauna reached the islands, immediate concern here is with the tortoises, and there is a possibility that they may have been introduced by men. No one would care to assert that they developed from birds or even from marine chelonians. There appears to be a sort of general agreement that they reached the Archipelago as tortoises not very different from what they now are. There is no evidence as yet, in the way of fossils, that they established themselves in the Tertiary or other formations earlier than the most recent. Their affinities are so close to living species on the mainland there is hardly room for doubt their ancestors were the same if indeed a species of the continent was not the direct progenitor brought, possibly, in the times of the Incas or still earlier by the aborigines. Because of the heavy and solid structure, one would not risk the suggestion that the Jaboty had drifted across the sea; but there is greater likelihood that island forms may have been drifted from one island to another after finding lodgment in the Archipelago. Whether it was 264 THE GALAPAGOS TORTOISES. cut off with the islands, from the mainland, or transported by other means it is very likely that the ancestral form is the widely distributed Jaboty, Testudo tabulata (Plate 2) of South America. This is supported by the numerous features possessed by the Galapagos in common with it, and by the comparatively slight divergence. Distribution from island to island was easier in the early days; as the lands sank the distances between the islands increased. Again there is no -proof whatever of inability to swim on the part of the island tortoises. They are not so different from those of the Seychelles which were proved to be good swimmers as long ago as 1801, by Grandpré, Voyage dans l’Inde et au Bengal. If animals like those of the islands of the Indian Ocean were able to swim from shore to shore leagues apart there is positively no reason for denying similar ability to wander to those on the other side of the earth, equally strong and per- haps more buoyant. Porter’s testimony, concerning the tortoises thrown over- board by the crews of vessels preparing for action, in which he says ‘‘ A few days afterward at daylight in the morning we were so fortunate as to find ourselves surrounded by about fifty of them which were picked up and brought on board, as they had been lying in the same place where they had been thrown over, incapable of any exertion in that element except that of stretching out their long necks”’ is proof only of great buoyancy, a quiet sea and absence of land in sight to direct and stimulate exertions. It certainly is no proof of inability to swim or exert themselves in the water. What would have happened with favoring winds and currents and shores in view may be readily surmised. Being thrown into the sea is not the only way of going adrift for a creature, fond of soaking in the water and of wallowing in the mud like swine, known to fall from the cliffs and to roll down the declivities. Transportation by men has certainly played a large part in the history of these creatures. By different vessels they have been carried to the Juan Fer- nandez, Chile, Peru, the eastern and the western United States, Europe, the Hawaiian, Marquesas and Tonga Islands, Australia and China; though there is no evidence at hand of actual colonies being established. There is no very direct assertion of transportation of tortoises from one island to others, except such as were to be used for food, yet in the face of the evidence of mixed breeds or hybrids in collections made in comparatively recent times it is difficult to convince one that such transportations did not occur. The testimony of Wolf, 1879, concerning the cattle lessens the doubt of actual occurrence: — ‘“‘ Das Rind lebt in grossen Heerden auf den Hochplateaus und Bergen von Floreana und Chatham und seit einigen Jahren traf man auch einige Stiicke auf dem THE GALAPAGOS TORTOISES. 265 Gebirge von Siid-Albemarle, ohne das man wiiszte wie sie dorthin gekommen sind” (Kin Besuch der Galdpagos = Inseln, p. 31). Floreana was the name given by Villamiel to Santa Maria (Charles). Young tortoises of which one might carry a number in his pocket presented no such difficulties as cattle in way of distribution. Cattle like most tortoises are tolerably expert at swimming. Sailors are fond of pets and a young tortoise or even a giant is an unfailing attraction to them. Undoubtedly they have helped to bring about the puzzling mixtures now gathered in various museums. The islands were frequently visited before the year 1800, but no attempt was made at colonization. Watkins, a sailor, was said to have been marooned for nearly a year, 1808, on Santa Maria. Villamiel’s colony of 1832, on the same island, numbered several hundred people in 1835 and at this time they had continued the destruction of the multitudes of tortoises so effectively that the settlers were drawing supplies from other islands. The BraGuxr, 1835, got specimens from three of the islands, Santa Maria (Charles), San Cristobal (Chatham) San Salvador (James), mostly young, apparently of a single species or so very young experts were unable to distinguish the species. This colony on Santa Maria was a penal colony. Wolf, 1879, says it soon melted away until finally ‘“‘der Rest der zuletzt tbriggeblie- benden Rauberbande rieb sich zum Theil selbst auf, zum Theil entwich er auf den anlegenden Schiffen der Wallfischfainger’’ (Ein Besuch, p. 4), and for a long time no traces of the colony had existed. Darwin, 1839, p. 457 says of Charles (Santa Maria), ‘‘the main article of animal food is derived from the tortoise. Their numbers in this island have of course been greatly reduced, but the people yet reckon on two days’ hunting supplying food for the rest of the week.’’ In the seventies Baldisan established a small colony on Santa Maria. He was killed by the colonists about eight years later; after his death this island was deserted. In 1865 Cobos landed a party on San Cristobal (Chatham) to gather Orchilla, a lichen used as a dyestuff; these remained till 1869, when they left the island. Ten years later Cobos returned to San Cristobal with more than a hundred men and founded a colony which appeared prosperous in 1891, and which may yet be in existence. The work of the colony was not limited to the island upon which it settled; it drew supplies from the other is- lands. The meat hunters, the oil collectors, and the orchilla pickers passed from island to island so frequently that it is not to be expected that any of the islands has its own unmixed race of tortoises, unaffected by mixtures from others. The older specimens, those secured before the exploitation of the islands, are perhaps the least likely to be suspected of being hybridized or mongrelized by 266 THE GALAPAGOS TORTOISES. importations, and a multitude of tortoises already established on an island might not be perceptibly influenced by the advent of a few new additions from elsewhere. The effect of the latter would be infinitesimal, but a species re- duced in number of individuals, near extinction, or not yet firmly established, on another island might lose its identity through the advent of one or a few new arrivals, as may have been the case on San Salvador (James) where the species no longer agrees with Porter’s description, or even on Santa Cruz (Indefatigable) in consequence of the rumored importation by Baur, himself one of the most earnest advocates of the opinion that each of the islands is inhabited by a distinct race. The few individuals that could produce an entirely distinct race in one locality might be unable to produce any effect in another. In specimens of T. vicina recently collected there is evidence of considerable mixtures, so also in those of 7. nigrita and of 7’. elephantopus. The BraGLE may or may not have secured one species on three islands; four species are located on Isabela (Albemarle). The Hassimr, 1872, obtained four species on Santa Maria that may be supposed with some confidence to have originated in four different localities, on three distinct islands. Comparatively little definite observation by trustworthy observers has been made in regard to either length of life, rapidity of growth, or rates of in- crease. Waite, 1899, brought together some notes of importance in the Records of the Australian Museum, 3, p. 95-103, pl. 20-22, in regard to a male T. nigrita taken by Porter to the Marquesas, thence to Tonga, thence to Sydney and thence to London where it died in 1898. This follows it nearly a century, without determining its age in 1813. In 1896 the length of the carapace was 4 ft. 23 inches, its width 2 ft. 11 inches and its weight 575 Ibs. (p. 98). Roths- child gives the length of this specimen after its death as 483 inches. Waite also notices another of this species which weighed 56 Ibs. when brought to Sydney in 1853. In 1893 it weighed 368 pounds. In 1896 according to Waite it was an egglaying female, had a length of 3 ft., a width of 2 ft. 5 inches and weighed 3203 pounds, having lost 473 pounds in three years. Quite recently very definite information concerning rate of growth appeared in Science, December 31, 1915, p. 933, in a note by Messrs. Daggett and Heller. The specimen of 7. vicina had been secured, by the latter, at Iguana Cove, Isabela, June, 1899, when it weighed 29 pounds and was supposed to be not much over a year old; it doubled its weight annually. Daggett says ‘‘At the time of its death [April 18, 1914] it weighed 450 lbs. and its carcass measured 41 inches long, 31 wide and 21 high.” In about sixteen years the individual had attained the bulk of speci- THE GALAPAGOS TORTOISES. 267 mens commonly said to be 400 years of age. Further observations are needed concerning rates of increase. Colnet is to be ecreditec with the statement that the nests never contain more than three eggs. Porter says the females without exception were full of eggs of which generally from ten to fourteen were hard, ready for extrusion. Beck says from ten to twenty eggs are ready for extrusion together, while twenty or thirty more were from one half to two thirds the normal size. The number of eggs laid by a single female in a season of course depends on her size and age; at any rate the evidence indicates she might densely populate a given locality, if beyond interference from enemies, in a very few years. Neither rate of growth nor abundance of progeny favors the conclusion that the tortoises have been on the islands from the very earliest times. Yet in estimating the numbers of the tortoises Baur quoting from Reynolds says that between October 13, 1832 and August 30, 1833, thirty-one whaleships reported at Santa Maria; he adds that if each vessel carried away but 200 it would make 6,000 from this island alone in less than a year. In alater MS. he states there is little doubt that about 10,000,000 tortoises were taken from the islands since their discovery. The factors of the greatest importance in the differentiation of species and varieties are the differences in the altitudes, which in the various islands range to 4,000 feet more or less, with the consequent differences in temperature, moisture, dryness, food and feeding habits, soil, etc. The variations in rapid- ity of growth, sizes attained, increase in numbers and the like are readily traced to one or several of these agents. In some of the most superficial characters their efficiency is quite perceptible; thus for example in the epiderm, the slough, which grows in correspondence with the skin and the bones beneath it. Sloughing is a process undergone by reptiles in general. It is part of their method of renewing and enlarging the epidermal covering. The new epiderm grows under the old one, the slough, between it and the balance of the skin. In the majority the discarded epiderm is thrown off at particular seasons. On some forms it is retained in one way or another and made to serve useful purposes, as protecting the skin or bones, or as claws or spines, and in a few it is so greatly modified as to serve as rattles. There are differences among tor- toises in regard to the habit. Most of the marine forms slough early in life and subsequent sloughs are less noticeable. Soaking in the water aids in slough- ing, but on the other hand a dry skin appears to be a more effectual preventive of loss and welds the various sloughs together one after another, cementing then firmly so that instead of a single thin horny layer of epiderm, of a single 268 THE GALAPAGOS TORTOISES. season, there may be a thick and strong covering that is increased in thickness by successive growths in its duration. The earliest slough in some marine tortoises is well described in Fry’s remarks on Chelonia depressa in, 1913, Records of the Australian Museum, 10, p. 162, ‘“‘Chelonia depressa then, emerges from the egg with each scute covered by a ‘larval shield’ which, as the animal grows, becomes an areola almost identical with that found in land tortoises; this is finally shed before the turtle reaches maturity, leaving the smooth scutes described by Garman and figured on Pl. XXI-XXII. As far as I can ascer- tain these areolae are unique amongst Marine turtles.” Giinther, 1877, dis- cussing the land tortoises of the Galapagos, says, p. 18, as long as the Tortoises are young, growth, as far as it is externally visible, proceeds along the margins of all the scutes; the sutures get broader, appearing as whitish seams, soft and very sensitive. After some time the young portion of the epidermis becomes horny, and is raised in a line (stria) running along each side of the suture. Ata later period this increment takes place only (at least only conspicuously) in certain portions of the carapace.” Marine and all land tortoises are hatched with the larval shield on each scute; it forms the areolar space on the scute which in the land species may or may not be shed, but which appears to be shed in marine forms at an early date. If not shed all the concentric striae remain, unless possibly affected by wear, each successive stria being the index to the amount of surface enlargement or growth, beyond the stria immediately preceding it. If there were no lateral growth, from starvation or disease, the scute, if there were no slough, might thicken by successive accretions beneath but possibly might not increase the number of concentric striae around its edges. Some specimens from dry localities, have retained the larval shield and have never sloughed; year after year they have increased by one or more the record of the striae on the scute. Other specimens appear to have kept the sloughs and striae for long periods then suddenly by a slough have lost the entire record of the series at once and from the striate and grooved condi- tion have become smooth and polished, to begin at the edges of each scute another striated record, see Plate 27. Testudo clivosa, Plate 21, a twenty-five © inch specimen may have an entire record. On the youngest specimens, a year or more of age, of most if not all of the species in the collection, no slough has occurred. The larval shield and all of the striae are in place. Larger speci- mens of some of the same species show plainly that through a slough the larval shield and the striae have been carried away leaving the carapace smooth; still larger ones testify to more or less regularity in sloughing and to consequent 5 5 3 THE GALAPAGOS TORTOISES. 269 smoothness: such species are 7. nigra, T. microphyes, and T. macrophyes. Testudo vicina, T. nigrita, and T. elephantopus and its varieties 7’. abingdonii, T. becki, and T. duncanensis all slough when young and afterward retain the striae as if no sloughing took place. Evidence that may be adduced in regard to the shortness of time since separation from one another is seen in the affinities of the tortoises; likewise in this connection there is no lack of confirmation for the statement that the species of various islands have been modified by importations from others. The early specimens now in museums, nearly all of them without known localities, are very difficult to place even with the aid of the considerable numbers in recent collections from certain islands. Changes have occurred in the last century that make some of the descriptions quite contradictory. Porter, 1815, in his Journal describes the tortoises from James Island (San Salvador) as round, plump, and black as ebony, their shells ‘‘sometimes remarkably thin and easily broken but more particularly so as they become advanced in age; when, whether owing to the injuries they receive from their repeated falls in ascending and descending the mountain, or from injuries received otherwise, or from the course of nature, their shells become very rough, and peel off in large scales, which renders them very thin and easily broken.”’ Van Denburgh, 1914, in his mono- graph, p. 321, says ‘“‘The James Island tortoise is a very large, heavy, thick- shelled species which resembles most closely the tortoise of Jervis Island [Ra- bida] and the Testudo vicina of southern Albemarle. It is somewhat intermedi- ate between the saddle-backed and dome-shaped races. The front of the carapace is high, but the middle of the back rises still higher. There is but little narrowing of the front of the carapace.’ Porter’s description was made a century earlier than Van Denburgh’s. Porter’s description of the tortoises of Santa Maria (Charles) and Espanola (Hood) applies a little better to 7. elephantopus Harlan than to T. nigra Dum. & Bib. now known to be the Charles Island species, of 1835. ‘‘The form of the shell of the latter is elongated, turn- ing up forward in the manner of a spanish saddle, of a brown color and con- siderable thickness.’’ How much the differences are is apparent on comparing with the descriptions and plates below. The disagreements may be accounted for by very rapid differentiation, or by modifications or replacements by im- portations. The following descriptions are made for most ready comparison with those in the majority of the literature. Percentages are not given as they do not lend themselves readily to visualization, an absolute necessity in descriptions 270 THE GALAPAGOS TORTOISES. and comparisons; they are too abstract and vary too much with age and sex to be really practicable. Excepting in the synonymy and the direct references the bibliography is not repeated; it has been worked out by Ginther, Baur, and Van Denburgh. TrEsTuDO TABULATA Walbaum. Plate 2. Testudo tabulata WaLBAUM, 1782, Chelonographia, p. 122; ScHorrrr, 1792, Hist. Test., p. 56, 62, pl. 12, fig. 2, pl. 13, 14; Dauprn, 1805, Hist. rept., 2, p. 242; Wimp., 1825, Beitr., 1, p. 51; ABBILD., pl. —; Betx, 1835, Monogr. Test., pl. —; Dumérit ET Brsron, 1835, Erpétol. génér., 2, p. 89; Gray, 1844, Cat. tort., p. 5; 1855, Cat. shield rept., p. 5; Srraucu, 1862, Chelon. stud., p. 80; 1865, Verth. schildkr., p. 25; Gray, 1870, Suppl. cat. shield rept., p. 4; BouLencEr, 1889, Cat. Chelon., p. 157; Srraucu, 1890, Bemerk. schildkr., p. 12; Gorn, 1904, Chelonios do Brazil, p. 14. Testudo denticulata ScuoEPFr, 1792, Hist. Test., p. 119, pl. 28, fig. 1. Testudo tessellata SCHNEIER, 1792, Schr. Berl. naturf. freunde, 10, p. 262. Chersine tessellata MERREM, 1820, Tent., p. 31. Testudo hercules Spix, 1824, Test. Bras., p. 20, pl. 14. Testudo sculpta Spx, 1824, Test. Bras., p. 21, pl. 15. Testudo carbonaria Sprx, 1824, Test. Bras., p. 22, pl. 16; Bru, 1835, Monog. Test., pl. —; Dumirit et BrBron, 1835, Erpétol. génér., 2, p. 99; Straucu, 1862, Chelon. stud., p. 80; 1865, Verth. schildkr., D2: Testudo cagado Sprx, 1824, Test. Bras., p. 23, pl. 17. Chersine tabulata GRAVENHORST, 1829, Del. Mus. Vrat. Rept., p. 19. Testudo boiei WAGLER, 1829, Icon. Amph., pl. 13. Chelonoides tabulata GRAY, 1873, Proc. Zool. soc. London, p. 724, pl. 60, fig. 3. The conclusion reached in this study of the Galapagos tortoises is that they were derived in comparatively recent time, much later than the Tertiary, from species of the nearer lands of the continent of South America. How their transportation was effected may not be determined at present. In order that the closeness of the relationships with one of the most widely distributed con- tinental species may be made the more evident the following description and illustrations of a specimen of Testudo tabulata from Porto Rico are introduced here. They are taken from a fair representative of the species and will be useful in comparisons. The measurements in inches of the specimen are: — Direct Curved Sternal M.C.Z. length Width length Width Height length Width 12050 22% 132 29 254 103 193 i The shape of the carapace of this species is more elongate than that of 7’. vicina; it has a circumference of forty-one inches and approaches the subcyclindrical in shape. The sides are steep; they are nearly parallel in an upper view; longitudinally the back makes a long, low arch, which does not rise in the middle as in T. ngrita. The front declivity resembles that of the latter. A majority | : ; THE GALAPAGOS TORTOISES. 271 of the scutes have little convexity; they are flattened, without prominent areolar spaces. The striae persist near the areolae, except in the older speci- mens. On the upper half of the fifth vertebral scute there is a rounded promi- nence or boss and backward from it a steep nearly vertical descent. The upper and the hinder edges of the third marginal meet in a sharp angle on the present individual. The outline of the eighth marginal bends abruptly outward behind the femoral notch and then continues in a regular curvature to the caudal. The caudal is large; its lower edge is strongly convex below the contiguous marginals and its surface is prominently convex. Faint scallops are formed by the outward edges of the marginals. A faint keel is between the humeral and the femoral notches on the fourth to the eighth marginal scutes. The front edge of the fourth vertebral is less than twice as wide as the hinder. The single axillary scute on each side has faint indications of having been formed by the fusion of two; the lower is the smaller and solidly united with the upper. The inguinal scales are single. On this specimen the sternum extends farther forward than the carapace about one and one fourth inches; it is deeply concave behind the middle and has a rounded prominence along each side of the lower surface. The nuchals are moderate, rounded in front, and have no traces of the lateral angles so notice- able on young specimens. The bones included in the humeral extension in front of the bridge are thick, strong, and curved outward on their outward margins. Compared with the swollen gulars the anal plates are thin; the deep notch between them is crescent-shaped, concave; the outer angles are pro- duced and blunted. The scales on the exposed portions of the legs and feet are large, somewhat imbricate and pointed. Each arm has two larger scales in front. Behind the hand there is another, and a short distance from this a smaller one about half as large. Enlarged scales cover the tail, and the exposed portions behind the thighs, similar to those on large specimens of 7’. microphyes. Carapace black, with a small spot of yellow on the areola of the second vertebral scute. Lower surfaces and head yellow mottled with brown or black. Skin between scales brown on neck and legs. Of this species, in Brazil, Goeldi, 1904, says ‘‘Testudo tabulata, 0 nosso jaboty, é animal imponente, cuja casca dorsal por si sé péde attingir de 55 ate 70 cm. de comprimento.” The largest at hand is that described above, 57.15 cm. Among the small specimens, from numerous localities between Dominica, Trinidad, Surinam, and southern Brazil, there is the same dissimilarity in the 272 THE GALAPAGOS TORTOISES. young as compared with the old of this species, or with the Galapagos tortoises of whatever ages. For example, the majority of the young of 7. tabulata have two axillary scales on each side. They are not as seen in T.. argentina Sclater, T. chilensis Gray, shown by Siebenrock, 1912, fig. 1, where the lower is the larger. The lower is much the smaller; in cases it is absent, in others, it evi- dently has fused with the upper, thus bringing about the condition obtaining in the Porto Rican specimen, normal in the Galapagos. There is evidence that the single axillary is not excessively rare: it was figured by Schoepff, 1792, Pl. 13-14, as 7. tabulata Walb., by Spix, 1824, under the synonym 7’. sculpta, Pl. 15, and by Bell, 1835, Mon. Testud., Pl. —, Sowerby and Lear, 1872, Pl. 14, under 7’. carbonaria Spix, Pl. 16, another synonym. The young of 7’. tabulata vary greatly in color, from yellow to black. On the back, whether light or dark the areolae are commonly yellow to orange, the color being more limited on the black individuals. On the majority the lower surfaces are yellowish, as to a considerable extent on the Galapagos. Specimens of less than six inches have marginal denticles on a thick swollen sternum, more or less produced, notched, and angled in front, etc., and they differ in some of the same features from the large or the aged of the species and from the Galapagos of whatever size, age, or species. The typical forms described as species of the latter have most often been chosen from the adult or the aged and these have provided the distinguishing characters from those induced by age. Among the old the greater difference exists and from them the more one approaches the newly hatched the more alike the specimens appear. This is what should be expected in cases of close genetic relationship. Comparing the tortoise of northern South America with those of the Galapagos it is found that the nearer approaches from the one to the others are the farther from the egg and mainly made by T. tabulata. Yet it is very doubtful if such close affinities would have obtained without the aid of a common ancestor. The results of all the comparisons made in this study tend to the conclusion that the origin of the Galapagos tortoises is directly connected with the species 7. tabulata of northern South America. : . 4 THE GALAPAGOS TORTOISES. 273 TESTUDO MACROPHYES, Sp. Nov. Plates 3-5. Testudo microphyes GUNTHER, 1877, Gigantic land-tortoises, p. 78, pl. 32-36, 38, 45, fig. A-C (part); Beck, 1903, 7th Ann. rept. N. Y. zodl. soc., p. 170; Smzpenrock, 1909, Zool. jahrb. Suppl., 10, p. 534; Van Densurecu, 1914, Proc. Cal. acad. sci., ser. 4, 2, p. 329, pl. 70-83 (part); Roruscuip, 1915, Nov. zool., 22, p. 406, 409 (part). Testudo macrophyes is the name here applied to a tortoise inhabiting the section of Santa Isabela Island (Albemarle) near Tagus Cove. The tortoise was first made known by Giinther, in 1877, who described and figured it from a number of the Cookson specimens but he did not recognize the species as dis- tinct. He made it identical with that he had characterized in 1875 under the name 7’. microphyes from a specimen of unknown origin, said at the time to be “a fully adult male” representing the ‘“‘smallest of the Galapagos Tortoises”’ and supposed to have come from Espajiola Island (Hood’s). Testudo macrophyes is one of the largest species. Compared with that of T. microphyes, Plate 11, the carapace appears more elongate, narrower across the humeral region, broader across the femoral plates, and somewhat higher in the arches across the middle of the back, over the third and the fourth verte- brals. Viewed from above the outlines in some degree resemble those of 7. ephippium. 'The convergence toward the front and the incline of the flanks are greater than on 7’. microphyes and consequently the opening in front between the carapace and the plastron approaches an angular in the nuchal section. The notch at the eighth marginal on each side appears more decided because of a slightly greater spread of the marginals over the femoral arches and farther back. The striae of growth are present on the younger; on the old they are more effaced. The straight width is about three fourths of the straight length; the curved width is little greater than the curved length. The differences in the sternum are even more patent. The humeral extension from the front of the bridge is narrowed forward; the gular plates are reduced, somewhat pointed, and are partly separated by a notch, Plate 5, fig. 3. The femoral extension from the bridge backward is narrowed toward the anal plates, which appear small, though larger than the gulars, and are rounded to meet in front of the shallow notch. Plate 3 shows the outlines of the carapace of the half grown specimen figured in Giinther’s, 1877, Plate 38, Plate 5 those of the 27 inch female on his 274 THE GALAPAGOS TORTOISES. Plates 35, 36, and on Plate 4 are those of the thirty-three and a half inch male drawn on his Plate 34 and fig. A of Plate 35. The following measurements, in inches, were given of the Cookson specimens, all supposed to be from Tagus Cove. Curved Curved Sternal Sternal Length Width length width length width 25 193 323 343 203 164 3334 243 41 42 244 23 334 25 44 4G O71 95 20 22 35 38 213 20 TEsTuDO vIcINA Giinther. Plates 6, 7. Testudo vicina GUNTHER, 1875, Philos. trans. Royal soc. London, 165, p. 277, pl. 35 A, 40, fig. B, pl. 41, fig. A, C, pl. 45, fig, C, C’, D; 1877, Gigantic land-tortoises, p. 73, pl. 31, 46, fig. B, pl. 47, fig. A, C, pl. 54, fig. C, C’, D; BouLenanr, 1889, Cat. Chelon., p. 170; Roruscuitp, 1902, Nov. zool., 9, p. 448; Heuer, 1903, Proc. Wash. acad. sci., 5, p. 54; Becx, 1903, 7th Ann. rept. N. Y. zodl. soc., p. 164; Srepenrock, 1909, Zool. jahrb. Suppl., 10, p. 354; Van Densurau, 1914, Proc. Cal. acad. sci., ser. 4, 2, p. 344, pl. 93-110; Roruscuiip, 1915, Nov. zool., 22, p. 406. Testudo elephantopus Baur, 1889, Amer. nat., 23, p. 1044; Lucas, 1891, Smithsonian rept., pl. 104, fig. —; Gapow, 1901, Cambridge nat. hist., 8, p. 378. Testudo nigrita CopE, 1889, Proc. U. 8. N. M., p. 147; Lucas, 1891, Smithsonian rept., pl. 104, fig. —. In the lot of young tortoises purchased by Prof. Louis Agassiz on Santa Maria Island (Charles) in 1871 there is a fourteen inch specimen of Testudo vicina. How it came to be on that island is not known. From the young it is very evident that this species is most closely allied to T. nigrita, Plate 10. Testudo vicina differs from the latter in being longer, narrower, and higher, and there are other differences of which the measurements give no hint. The length over the curvature at this stage of growth about equals the width over the curvature in each species. The carapace of 7’. vicina is the more flattened, it is more depressed, that is, it has less of the dome shape on the middle of the back. The arches between the areolae of opposed costal plates are lower and broader. The curves from the nuchal notch to the areola of the fourth vertebral scale are broader and the descent from that point to the lower edge of the caudal scute is less steep. In the sternum, the two species are similar in the gular and the anal scales. The areolar spaces appear to be smaller on the back of T. vicina, but the striae are equally distinct. A photograph of the specimen is shown in Plate 6. The outlines of the type, as drawn by Ford for Gunther, THE GALAPAGOS TORTOISES. 275 1875, Plate 35A, from a thirty-three inch specimen, are included in Plate 7. Specimens at hand are more curved from the middle of the fourth vertebral ' scale to the anterior marginals than in this drawing. In the fourth vertebral and backward the curve is sharp, but the marginals are less declivous, which is also true of the forward marginals. On both front and back the marginals are somewhat reverted, concave on their upper surfaces, in the older specimens. A number of the specimens in the M. C. Z. were secured by Webster near Iguana Cove in the southern part of Santa Isabela (Albemarle); these are compared with the figures of those secured by Van Denburgh in the same locality. The declivity from the middle of the fourth vertebral backward is usually greater than that forward from the second. The caudal plate is directed downward, and slightly forward at the lower edge. A large shallow notch partly separates the marginals of the foremost pair. The indentation at the fourth, and that at the eighth marginal, on each side are shallow, as also the grooves, compres- sions, extending from them. Except perhaps on specimens of greater age, the concentric striae are present. The areolar spaces vary in convexity; in eases they are prominent. The carapace is broader posteriorly; it appears subtruncate, the caudal scale being indented between the marginals at each side of it. In front the margin is the more rounded, subacuminate. Above the humeral and the femoral arches the marginals form scallops. The sternum of the type is longer than broad; the humeral extension, in front of the bridge, is broad at the end and is somewhat concave on its lateral margins in the large specimens. The femoral extremity is shorter, broader, and is convex on the lateral margins of the abdominal plates. In both young and old the gulars thicken upward; on the aged they have a comparatively small amount of the downward swelling. The anal scutes are broader than the gulars and have an angular notch between them; they thicken and curve down- ward with age. The sternal concavity is deepest below the hinder portions of the abdominal scales; it is deeper on the old males. In recent collections of specimens of this species there are appearances of considerable mixtures by transportation from one island to another. The earlier captures appear more distinct than some of the later ones. Undoubt- edly the oil collectors, the orchilla pickers, the meat hunters and others were us quite ready to contribute to a distribution that has left traces in various direc- : tions, and latterly the young specimens from different islands have been on sale within reach of collectors at particular localities, 276 THE GALAPAGOS TORTOISES. MEASUREMENTS. (In inches). Direct Direct Curved Curved M. C. Z. length width length width Height 4482 144 102 183 183 7 11076 27 21 344 354 144 11077 30 24 36 403 153 11078 2832 223 343 37 143 11087 39 283 474 50 18} TESTUDO NIGRITA Duméril et Bibron. Plates 8-10. * The Great Gallapago-Tortoise Mitcuri1, 1815, Med. repos., ser. 2, 2, p. 309, 402. Testudo nigrita DumMéRr1L ET BrBRon, 1835, Erpétol. génér., 2, p. 80; Srraucu, 1865, Verth. schildkr., p. 25; GinTuEr, 1875, Philos. trans. Roy. soc. London, 165, p. 267, pl. 33, fig. B, pl. 35, fig. C, pl. 37, fig. B, pl. 38, fig. D, pl. 39, fig. D; 1877, Gigantic land-tortoises, p. 69, pl. 30, fig. B, pl. 31b, fig. C, pl. 42-44, figs. D; BovuLencrr, 1889, Cat. Chelon., p. 169; Srraucu, 1890, Bemerk. schildkr., p. 52; Warrs, 1899, Records Austr. mus., 3, p. 95, pl. 20-22; HexiEr, 1903, Proc. Wash. acad. sci., 5, p. 50; Srespenrock, 1909, Zool. jahrb. Suppl., 10, p. 531; Roruscmrip, 1915, Nov. zool., 22, p. 407. Testudo indica Gray, 1831, Syn. rept., p. 9 (part); 1844, Cat. tort., p. 5; 1855, Cat. shield rept., p. 6, part; 1870, Suppl. cat. shield rept., p. 5; Sowrersy & Lxar, 1872, Tort., pl. 6. Testudo planiceps GRAY, 1855, Cat. shield rept., p. 6, pl. 34; 1870, Suppl. cat. shield rept., p. 5. Testudo elephantina StravcH, 1862, Chelon. stud., p. 83. Testudo elephantopus Gray, 1870, Proc. Zool. soc. London, p. 708, pl. 41 (part); 1872, Appendix cat. shield rept., p. 3. Testudo wallacei VAN DenBURGH, 1914, Proc. Cal. acad. sci., ser. 4, 2, p. 351, pl. 111, 112. Testudo porteri ROTHSCHILD, 1903, Nov. zool., 10, p. 119; StzBENRocK, 1909, Zool. jahrb. Suppl., 10, p. 532; Van DenBuras, 1914, Proc. Cal. acad. sci., ser. 4, 2, p. 354, pl. 113-121. Testudo darwini VAN DenBurGH, 1907, Proc. Cal. acad. sci., ser. 4, 1, p. 4; SreBENROCK, 1909, Zool. jahrb. Suppl., 10, p. 533; Van Denpures, 1914, Proc. Cal. acad. sci., ser. 4, 2, p. 319, pl. 56-63; RoruscHILpD, 1915, Nov. zool., 22, p. 405, pl. 36. Testudo sp. VAN DenBuraH, 1914, Proc. Cal. acad. sci., ser. 4, 2, p. 362, pl. 122. A young specimen of this species, measuring eleven and one half inches in direct length, Plate 10, does not differ greatly in its shapes from the smaller of the types originally described, which measured twenty-two inches, and of which the outlines are sketched on Plate 9 from the plate by Giinther, Trans. Zool. Soc. London, 1875, 170, pl. 35, fig. C. The carapace is a short broad oval, rounded in front and over the femora, and subtruncate across the caudal section. Its height is about half of the length; the back is broadly arched. At the nuchal notch the height is about half of that in the middle of the body. The strongest vertebral declivities occur from the areolar space of the first vertebral scale forward and from that of the fourth vertebral backward. The flanks are nearly straight and bear a low keel below the areolar spaces of the : : THE GALAPAGOS TORTOISES. 277 costal scales from the fourth to the eighth marginals, continuous with the edges of the carapace in front and behind. This keel is not retained on old specimens to such an extent as on 7’. clivosa, Plate 21. There is a weak notch on each side at the fourth marginal and another at the eighth; the grooves from these notches extending toward the nuchal notch and toward the caudal scale are shallow but distinct. In front of the humeral notches, and behind the femoral, the marginals form scallops. Anteriorly the marginals have a slight incline downward. All of the scales are strongly marked by striae. The amount of _ prominence in the areolar spaces varies; on specimens of a length of two feet or thereabout the areolars are decidedly prominent on the first and on the fourth and the fifth vertebrals. The descent from the fifth is quite steep. The striae persist on some; on other species they are nearly or completely lost, Plate 36 (T. nigra). The bones are light. On the sternum the concavity is absent or shallow on the young, of moderate depth on specimens more than half grown. The gular scutes are narrow and swollen above the ends which are slightly turned downward below the edge at each side of a shallow notch. The anal scutes are longer than wide and the pair are separated behind by a moderate notch; the angle on a scute is thin, sharp, and curled upward somewhat. The caudal notch persists on large specimens. Plate 8 contains the outlines of the larger of Bibron’s types, a forty-one inch specimen, taken from Giinther’s figure B of his Plate 33, apparently a 9. Males of more than thirty inches have served as types for 7’. wallacei, T. porteri, and T. darwin. This sex is commonly the more elongate and the flanks are less full and rounded. Occa- sionally the curved width is less than the curved length, as in case of the type of T. darwint but in most cases the curved length and the curved width are nearly equal, while the direct width is two thirds to three fourths or more of the direct length. Specimens identified with this species have been collected on Santa Maria (Charles), Santa Cruz (Indefatigable), Rabida (Jervis), San Salvador (James), and on Middle and on South Isabela (Albemarle). The color is a uniform dark brown or black, commonly without yellow markings on the lower surfaces. In all likelihood a note by Mitchill, 1815, was the first mention of this species. The measurements he gives are impossible when applied to any known tortoise of the Galapagos, but they make their nearest approach to TJ. nigrita. Since the article Description of the great GALLAPAGO-TorToIsE. From Dr. Mitchill’s Lectures on Natural History, contains interesting matter, and has been ignored heretofore, it is reprinted from the Medical Repository, 2, p. 309 and 404. 278 THE GALAPAGOS TORTOISES. “ About the middle of July, 1814, the ship Essex Junior, Lieutenant Downes, of the U. S. Navy, arrived in New York. He had been on a cruise, by order of government, along the coast of Brazil, and round Terra del Fuego, and off the land of Chili and Peru, in quest of British traders and whale-men. He served under Commodore David Porter, of the frigate Essex, a vessel of war which had almost broken up the enemy’s navigation and commerce, in the tract of ocean lying between Cape Horn and the Gallapagos Islands. After visiting Valparaiso and Lima, in March, 1813, Capt. Porter proceeded to the neighborhood of this group, and cruised there between April and October, for English vessels, where he captured twelve, which were chiefly occupied in the chase of the spermaceti whale. He describes the Gallapagos Islands as “being perhaps the most barren and desolate of any known,” and so utterly destitute of fresh water, that he was obliged to touch on the coast of America, during the time, to procure a supply of that necessary article. They are chiefly volcanic piles, and the water that condenses on their summits is absorbed by tufa, slag, and ashes, before it can reach the sea. From the Gallapagos the crew took a number of the native tortoises for food. These creatures are very large, and frequent there. They inhabit the land, and seldom or never enter the water from choice. Two of them were brought alive to New York. They bore the voyage of between two and three months without taking any food. They have been carefully examined, and described. Both were females. The larger had the following characters. The colour of the buckler and skin was a deep and uniform black. The head was rather small in proportion to the body, and at pleasure could be drawn out of sight, and concealed behind the fore legs, approximated for its protection. The back was very convex. The sides prominent and capacious; but the gibbosity was without knobs, asperities, or processes; and merely marked by dividing lines, among the pannels. There were five of these pannels along the back, four on each side, and twenty- three in the circumference, making thirty-six in the whole. The length, measured over the elevation of the buckler, between head and tail, was about two feet and a half. The distance from side to side over the back was almost as great, or nearly twenty-nine inches. The height, as the animal rested on the belly or sternum, was about two feet. The weight, when she arrived, poor, lean and famished was eighty pounds. The fore part of the legs was covered with a thick and hard skin, that by deep indenta- tions resembled the scales of an alligator’s hide. Each of her fore feet had five claws; of the hinder, four, and the balls of her feet were prominent and puffed, as if for walking over the ground, and not for creeping, or crawling. Such is the length of her legs, that her erect posture adds about a foot to her stature. This individual, weak and exhausted as it was, could move with the weight of a man on its back. The fore part of the crown of the head was rough, like the legs. It arrives in its native region, to the magnitude of three hundred pounds, and even more. When full grown and strong, it can travel away with the weight of three or four men. It is very prone to accumulate fat. In cooking the tiesh there is no need of employing butter. It can live, as is said, a year, without food or drink. The sailors travelled two miles and more inland upon the Gallapagos Islands in search of these tortoises, or furpins as they called them. When they catch the animals, they carry them in their arms, or on their shoulders, to the boat. There were more than two hundred on board the Essex. The English whaling vessels that were captured, mostly had some of them. Navigators prize them highly for food, and esteem them as savoury and wholesome. One of the men told me he had seen the same sort of tortoises on the Isles Tristan d’Acunha THE GALAPAGOS TORTOISES. 279 and Bourbon. Like the camel, the turpins have a stomach or reservoir in which they pre- serve water to the amount of several quarts for a long time. Voyagers sometimes kill them for the purpose of procuring this water to drink, which they pronounce to be cool and sweet. Commodore Porter told me he had repeatedly tasted it, and could bear witness how good and potable it was. The water the stomach contains is sufficient for cooking the flesh. The Gallapagos are stated to abound in volcanoes, and subterranean fires. They are rocky, peaked and forbidding. There are few springs or brooks of water. With great difficulty and exertion the Essex collected about half a dozen casks; and then sailed for the continent to obtain a further supply. There are no settled or stationary human inhabitants. The seas abound in excellent fish and green turtle. Cocoa-nuts may be found in some places on shore. And the Guanos lizard may be catched for eating. But it must be remem- bered that this is the Sea-Guanos, a species of lacerta, entirely different from that of the West-Indies. The Sea-Guanos of the Gallapagos, swim and feed in the ocean, and go ashore to rest and breed.” The following occurs on p. 404 of the same volume: “On the 13th of February, 1815, I examined the body of the female Gallapagos tortoise. I found the alimentary canal to be exceedingly large and capacious. The whole length of this tube, from the throat to the anus, was about thirteen feet. Of this the gullet and stomach were twenty inches; the small guts five feet, and the large ones six feet and a half. The cecum had no appendages; the colon had faint and ‘weak muscular bands; and the rectum communicated with the uterus and bladder a few inches before the posterior outlet. They are all united with one common cloaca. The bladder contained a considerable quantity of urine. It was remarkably large, and capable of holding four quarts of water, as we found by experiment. The creature, when alive, voided naturally great quantities of urine. The animal is said to hold within it, when in health, a plenty of potable water. I found none in this individual; though the stomach, colon, and bladder could each have contained a large supply. The reason probably was, that the creature had been for a long time under artificial restraint, and had been crammed to death, through kindness, by Indian meal (meal of maize). The uterus contained two eggs almost ready for exclusion, the weight of one alone was six ounces. These had beautiful calcarious shells, that were rough, white, round, and about the size of a one pound shot. It was divided into two parts, and the ova were very numerous, and of different sizes. Near the junction of the two cornua uteri with the strait intestine, were the two kidneys of a triangular figure, and of a convoluted structure. Their extreme length was four inches, and the breadth of the widest part two and a half. The trachea divided into two branches, one of which entered each lung. The cells of this organ were open, large, and distinct, as usual in these amphibious creatures. There were two large muscles parallel with the back, for retracting the neck. One of them arose from each side of the cervical vertebrae; they were of extraordinary length, and were inserted in the shell towards the rump. The outer coat of the shell looked as if it was sufficiently beautiful for manufacture. The heart consisted of two auricles and one ventricle; the auricles were separated by a septum. The pulmonary veins emptied into one, and the vena cava into the other. There was but a single ventricle; and two fleshy valves, in shape somewhat like the epiglottis, opposed the return of the blood from the ventricle into the auricles. From the ventricle proceeded three arteries; two of which soon divided into two branches each, making five in the whole, soon after leaving the heart. The heart was oblong and kidney shaped. These arteries had appropriate valves at their origin.” 280 THE GALAPAGOS TORTOISES. MEASUREMENTS. (In winches). Straight Straight Curved Curved Sternal length width length width Height length T. portert + 363 3l 503 55 . 223 312 T. darwini * 38 28 49 46 20% 324 T. darwini * 483 344 563 62 25.7 36 Posp? 262 225 344 37 16} 2235 M.C.Z. 11079 243 193 30 32 123 11080 213 163 27 274 11} 11088 27% 214 37 383 11091 22% 18 28 304 123 “ 42049 113 9 15 15 53 103 TESTUDO MICROPHYES Giinther. Plates 11-20, 38, fig. 3. Testudo elephantopus GUNTHER, 1875, Philos. trans. Roy. soc. London, 165, p. 261, pl. 33, fig. A, pls. 37-39, fig. A, pl. 40, fig. A, C, D., pl. 41, fig. B, pl. 42, fig. A, pl. 43, pl. 44, fig. A-A’”’, C, D; 1877, Gigantic land-tortoises, p. 63, pl. 30, fig. A, pl. 42-44, fig. A, pl. 46, fig. A, C, D, pl. 47, fig. B, pl. 51, fig. A, pl. 52, pl. 53, fig. A-A’”, C, D; Roruscuitp, 1902, Nov. zool., 9, p. 448, 618; Hr uumr, 1902, Proc. Wash. acad. sci., 5, p. 53; SteBppnrock, 1909, Zool. jahrb. Suppl., 10, p. 532. Testudo microphyes GUNTHER, 1875, Philos. trans. Roy. soc. London, 165, p. 275, pl. 36-39, fig. B; 1877, Gigantic land-tortoises, p. 78 (part), pl. 37, pl. 42-44, fig. B; Roruscuitp, 1915, Nov. zool., 22, p. 406, 416. Testudo giintheri Baur, 1889, Amer. nat., 23, p. 1044; Van Drenpureu, 1914, Proc. Cal. acad. sci., ser. 4, 2, p. 335, pl. 84-92; RoruscuiLp, 1915, Nov. zool., 22, p. 406, 410. Testudo galapagoensis ROTHSCHILD, 1915, Nov. zool., 22, p. 404 (part), pl. 27, 28. Testudo chathamensis VAN DENBURGH, 1907, Proc. Cal. acad. sci., ser. 4, 1, p. 4; 1914, 2, p. 323, pl. 64— 69; RoruscHILp, 1915, Nov. zool., 22, p. 406, 409, 416, pl. 29, 30. The species Testudo microphyes was founded on a specimen twenty-two inches and a half in direct length, at the time in the possession of the Royal Institution of Liverpool. It was said to be ‘‘a fully adult male,” ‘‘ the smallest of the Galapagos Tortoises,’ and was supposed to have come from Espafiola Island (Hood). Later this type was secured by the British Museum and in the publication on the Gigantic Land-Tortoises, 1877, after comparison with Cookson specimens it was said to be ‘‘undoubtedly a female” from the north of Isabela Island (Albemarle). The conclusions here recorded after a detailed 1 Rothschild, Nov. zool., 1903, 10. 2 Van Denburgh, Proc. Cal. acad. sci., 1907, 1. 5 Rothschild, Nov. zool., 1915, 22. 4Van Denburgh, Proc. Cal. acad. sci., 1914, 4. THE GALAPAGOS TORTOISES. 281 study are that the type is abnormal, somewhat aged, probably a dwarf, and differs so much from the specimens obtained by Cookson as to prevent retention in the same species. The Tagus Cove species is considered a new one, and, being one of the largest found on the islands is named 7’. macrophyes (Plate 4, 5). The type of 7. microphyes Ginther, 1875, was probably not an average individual of the species. The outlines of the original figures by Wesley, are sketched on Plate 11. As seen from above the shape is subelliptical, slightly irregular, and slightly narrower forward. The body is depressed and has a rather low arch on the back. The outline, from the side, is broadly curved from front to rear. Anteriorly in the first and the second vertebral plates there is some descent and posteriorly from the middle of the third vertebral the curve becomes steeper and sharper. The curved width is little if any greater than the curved length, certainly not so much greater as in the nearly allied species, T. nigra. Compared with that species the bones of the shell are thicker and heavier, the back is not so high posteriorly. The scales are smooth, the marginals appear to be much worn, the edges of the carapace are thickened and rounded at the edges. The sternum bears more resemblance to that of 7’. elephantopus than to that of 7’. macrophyes; it is broad and broadly rounded in front of the humeral extension, and behind the femoral extension, across the anal scutes, it becomes when old nearly or quite truncate. The skull differs from that of 7. macrophyes from Tagus Cove; it agrees with that of 7. giintheri from Villamiel. Young individuals of about fifteen inches in direct length, Plates 17, 19, purchased on Santa Maria Island (Charles) appear rather smooth, though the striae are sharper in the younger stages. The gular plates are rounded and ~ not separated by a sharp notch. In the specimen, Plate 19, the bone in the anal scutes is nearly truncate and the notch is shallow but the horny scales extend beyond and turn up in points behind it. Specimens of this size have the shallow early indications of the sternal concavity. Plates 12, 20 represent specimens of about two feet in length of cara- pace. In this size, with the exception of the blister-like pits, the scales are smooth, the gulars, the anals, and the scallops of the edges are much changed. The back is depressed to different degrees in different individuals and varies in curvature. The swollen flanks of the females indicate that they are fully adult. In cases the notch between the gulars, or that between the anals is obsolete; in others these scutes have suffered less. The gulars thicken up- ward; in some examples the anals have hardly changed. 282 THE GALAPAGOS TORTOISES. Specimens of from thirty-one to thirty-eight inches, Plates 14, 16, show the edges of the shells to be more modified and the backs to be more depressed, especially in the male, in which they have become subquadrate. The sternum is nearly truncate, is deeply concave and the slight constriction at the fourth and the eighth pairs of marginals is evident in both sexes. None of the speci- mens show decided bosses on the areolar spaces of the dorsal scutes. The specimen, Plate 12, was received in exchange from Van Denburgh, from Villamiel, southeastern Isabela (Albemarle); its affinities are evident on comparisons with his Plates 66, 68, 83, 84, 90, 92, and others. Plates 14, 16 of the present work are photographed from large specimens brought by Baur from the same locality. Plate 15 shows outlines of 7. elephantopus Giinther, 1875 and 1877 (not of Harlan, 1827), from drawings by Ford of a specimen the history of which was unknown. Much dependence was placed on the skull of the same specimen for distinction from other species. The differences appear to be individual rather than specific. The occipital crest affords the greatest variance; that crest, however, figured in Giinther’s, 1875, Plate 38, fig. A, and 1877, Plate 43, fig. A, had been broken or mutilated and in its repair took on a peculiar shape not occurring on other specimens. Young to medium sized specimens favor 7. microphyes Giinther, 1875; the old ones agree better with T. elephantopus Giinther, 1875,= T. giintheri Baur, 1889. On Plates 17 and 18 are figures of No. 4479 M. C. Z. and on Plate 19 is a photograph of No. 4476. These are the smallest specimens of this species in the collection; they were purchased by Prof. Louis Agassiz on Santa Maria. A description of No. 4476 will answer about as well for No. 4479, they are so nearly of the same size and appearance. No. 4476 has a rather plump appear- ance, being well and smoothly rounded on the sides and back. The indentation or concavity on the first to the third marginals is shallow and these scutes are but little turned upward. The indentation of the eighth marginal is not very distinct. 'The bosses on the vertebrals are low, the most marked being that on the first of the series; there are no bosses on the second and third, and none on the costals. Instead of a sharp ridge along the flank as on 7. wicina this portion is rounded and smooth, the carina making its appearance far in front, in the fourth marginal, and far back, in the seventh. The declivity in the first vertebral scute, the descent to the neck, is a little steeper than on older speci- mens; the greatest height is attained in the third vertebral. There is a large area in the central portion of each scale on which the striae are effaced, either by sloughing or by scouring; nearer the edge there are seven or eight more % THE GALAPAGOS TORTOISES. 283 distinct striae. From the fourth vertebral to the caudal there is a moderately sharp descent but the first of these scales has not a marked boss. The lines of junction between marginals and costals are moderately straight. The caudal does not extend as far backward as the marginal at either side of it, but curves downward somewhat lower. The sternum is broad; the abdominals make a long bend from their lowest portion to meet the marginals. The anals form a shallower notch than that of T. vicina; their hinder angles are more rounded off. The sternal concavity is shallow, not yet reaching the depth of half an inch in the hind parts of the abdominals. A deep groove near the edges of the scutes (stria of growth) shows that the abdominals have made a greater enlargement than the other scales (Plate 19, fig. 3). MEASUREMENTS. (In inches). Straight Straight Curved Curved Sternal Sternal length width length width length width Height T. elephantopus' 31 26 37% 40 243 23 T. microphyes * 224 153 26 29 18 14 10 T. chathamensis? 354 241 394 38 264 14 T. chathamensis? 223 17 281 26 183 113 M.C. Z. 11065 373 303 433 50 284 273 173 11066 313 243 364 40 23% 22 14 11067 253 196i) 904- oes 208 191 162 113 11071 238 High FORE, = RS 183 164 113 11085 42 31 503 303 283 17 11086 411 321 51 553 4476 153 123 193 193 123 7h 4479 153 113 192 193 123 103 73 TESTUDO CLIVOSA, sp. nov. Plate 21. This tortoise (Type M. C. Z. 11075) is described from a bony carapace and plastron on which there are seven or eight entire scutes in place. Of its history nothing definite is known. It was supposed to have come from the Mascarenes. Possibly a name and date, ‘“‘Narraga 1861,” may yet help to 1 Gunther, Philos. trans., 1875, 165. 2 Van Denburgh, Proc. Cal. acad. sci., 19C7, 1. 284 THE GALAPAGOS TORTOISES. determine the habitat. There is nothing in its structure that would preclude | an eventual determination of an origin, either indirect or direct, among the Galapagos. It has close affinities to their species though in some respects its peculiarities do not permit identification with any of which descriptions or figures have been published. Its outlines are suggestive of the wider and shorter individuals of 7’. elephantina Duméril et Bibron. It has two gulars but no nuchal plate. Direct Direct Curved Curved Circum- length width length width Height ference 253 in. 20 in. 304 in. 323 in. 122 in. 514 in. Height to nuchal notch 10, to middle of back 12}, and to lower edge of caudal scale 14 inches. The bones are thin and light, the entire weight, with attached scales, being less than five pounds. All sutures are complete and firmly joined excepting those at the inner edge of the marginals from the first vertebral to the edges of the supracaudal. The bones of the back at the lower edges of the costals are separated from those under the marginals by a space and their only connections are made by the ribs, except at the nuchals and at the supracaudals where they are rigidly united by broad anchyloses. The conditions along the flanks are like those existing on the very young of other species. The scutes are thin and fragile, the areolar spaces are small; the concentric striae are numerous, narrow, and sharp. The carapace is short and broad and is broadly arched on the back. In front from the middle of the second vertebral, through the first, the descent is about as in 7’. vicina Giinther. Fourth vertebral nar- row behind, hind edge about 32 times in front edge. Seen from above the shape approaches the subquadrangular, owing to a considerable spread and recurvature of the humeral and the femoral marginals. The anterior marginals extend forward and upward, and the posterior extend farther back than the caudal, which is narrow, convex, and curves down and forward at the lower edge. The humeral indentation, on the third marginal, is shallow but the groove from it to the middle of the first vertebral is decided, because of the recurvation of the marginals behind it similar to 7’. elephantopus. The verte- bral shields are broader than the costals; each has a high convex boss on the middle, that on the fifth vertebral being especially prominent. Including the areolar space on each costal there is a similar boss. On the plastron the gulars are broad and subtruncate in front, swollen and rounded above the edge and rounded on the angles. The pectorals are short, broad and in contact. THE GALAPAGOS TORTOISES. 285 The concavity is rather deep, deepest in the hinder halves of the abdominals. The anals are shorter and broader than the gulars; they meet in a shallow notch behind, below which the edge is swollen, thickened and bent downward. TESTUDO TYPICA, Sp. nov. Plate 22, 34. This species (Type M. C. Z. 11072) is one of the nearest allies of T. nigra. Looking downward upon it the carapace appears subquadrangular and some- what narrower across the humeral section than across the mid length or farther back. In the middle of the first vertebral scute there is a prominence and from it a considerable declivity forward. Behind this vertebral the dorsal arch is low and regularly curved to the areolar space on the middle of the fourth verte- bral, which is prominent and from which the descent is steeper to the space on the fifth vertebral, which latter is yet more prominent, almost acuminate. From this boss the descent to the caudal is nearly vertical. The upper margin of the fifth vertebral is narrow, being less than one fourth of the width of the lower edge. The convexity of the other vertebrals and of the costals is low. The marginals are comparatively large; the anterior three pairs and the posterior four pairs are reverted and form scallops along the margins. Marked grooves occur from the humeral and the femoral notches along the inward edges of the marginals. Along the flanks the costals and the marginals are sufficiently convex to give a full and plump appearance to the body. The lower edge of the caudal is not bent forward; it inclines straight back and downward. Axil- lary and inguinal scales are single. In the femoral section the marginals are somewhat wide and pass back and upward; the lower edge of the caudal de- scends sharply below the marginals. This form is distinguished from all its nearest allies by the prominent subacuminate boss of the fifth vertebral. Sternum subtruncate in the gulars with the angles rounded off. On the lower surface the concavity is deepest at the junction of the femorals with abdominals; it shallows gradually forward and more rapidly outward and backward. The anal scales are hardly swollen at.all; between them the notch is of moderate depth. In front the gulars are swollen upward. Solidity of bones and anchylosis of sutures are probably the effects of a considerable age. The color was nearly a uniform black. 286 THE GALAPAGOS TORTOISES. MEASUREMENTS. (In inches). Direct Curved Sternal M.C. Z. length Width length Width Height length Width Gulars Anals 11072 Type 24° 17} _ 38,° 30) Wig 5260 29 221 34 363 138 23 20 The type, Plate 22, differs from specimens of 7’. nigra, equal in size, in being less convex on the back, the arch from the first to the fifth vertebral scale being lower, as also that across the back; the marginals above the legs are entire, not worn and thickened, and form greater arches, the eighth marginal is wider on its upper edge; the bones of the sternum above the gulars and the anals are not swollen downward; the anal notch is a little more than 90°; and the fifth vertebral scale is narrow on its upper edge and bears in its upper half a steep and sharp prominence. On similar comparison with 7. microphyes it is seen that nearly all of these differences are pertinent. Specimen described is from the Boston Society of Natural History. Dr. C. F. Winslow collection (See Proc. Boston Society Natural History, 1861, 10 p. 59). , TESTUDO NIGRA Quoy et Gaimard. Plates 23-33, 35, 36. Testudo nigra Quoy ET GatmaRD, 1824, Voyage Uranie et Physicienne zool., p. 172, pl. 40; Cuvimr, 1829, Regne anim., 2, p. 10; Voret, 1832, Thierreich., 2, p. 9; Dumérm eT Brsron, 1835, Erpétol. génér., 2, p. 115; Wireemann, 1835, Nov. act. Leop.-Carol., 17, p. 118, pl. 138; Srraucn, 1862, Mem. Acad. St. Petersb., ser. 7, 5, p. 85; 1865, Verth. schildkr., p. 29; BouLencsEr, Cat. Chelon., p. 170; Srraucn, 1890, Bemerk. schildkr., p. 53. Testudo coliforniana Quoy ET GAIMARD, 1824, Bull. sci. nat., 1, p. 90, pl. 11. Testudo elephantopus JAcKsoN, Boston journ. nat. hist., 1, p. 4438, pl. 10, 11; Van Drensureu, 1914, Proc. Cal. acad. sci., ser. 4, 2, p. 245, 316, pl. 55, fig. 2, pl. 56, fig. 1. Testudo galapagoensis Baur, 1889, Amer. nat., 23, p. 1044; Gtnruer, 1902, Nov. zool., 9, p. 184, pl. 16-21; HewuEr, 1903, Proc. Wash. Acad. sci., 5, p. 53; Srspenrock, 1909, Zool. jahrb. Suppl., 10, p. 553; RoruscuiLp, 1915, Nov. zool., 22, p. 404, pl. 23-26. Testudo wallacei Rotuscuitp, 1902, Nov. zool., 9, p. 619; HetiEer, 1903, Proc. Wash. acad. sci., 5, p. 54; Srepenrock, 1909, Zool. jahrb. Suppl., 10, p. 533; RoTuscurip, 1915, Nov. zool., 22, p. 407, pl. 31, 32. The type described and figured by Quoy and Gaimard was said to be twelve inches long and eight inches wide. It was obtained by Freycinet at the Sand- wich Islands (Hawaiian) from the Captain of a vessel that had recently (about 1818) arrived from California. There is no mention of the original habitat. The figures published by the describers are sketched in outline on Plate 23, fig. 2, 3. Plate 23, fig. 1, is from a drawing published by Wiegmann of a THE GALAPAGOS TORTOISES. 287 specimen weighing a hundred and twenty-five pounds, after losing forty pounds of its weight in a year’s voyaging, before being bought by Meyen from a Galapagos Whaler at Honolulu. The more important items in the description are in the words of the au- thors: — ‘‘Testudo, toto corpore nigro; testa gibba, scutellis dorsalibus priori posteriorique altius in medio elevatis, cunctis loricae margine striatis, lateribus subcarinatis. Les enveloppes de cette tortue ont douze pouces de longueur et huit de largeur. La carapace est trés bombée, arrondie, et le disque composé de treize ecailles; des cing qui forment la rangée du milieu, deux ont un diametre transversal plus considérable que les autres; plusieurs sont protubérantes a leur centre, mais surtout l’antérieur et la postérieur....Le plastron se compose de seize piéces, dont huit en avant, une paire beaucoup plus large au milieu, et six en arriére: les deux premiéres sont arrondies et courbées en bas; les postérieurs assez profondément echancrées. Toutes offrent des stries concen- triques et paralléles entre elles.” Duméril et Bibron, Erpétol. Génér., 1835, 2, p. 118, give the length of the type as 34’’, and those of a larger carapace as ‘‘ Long. (en dessus) 71’’; haut. 28”; larg. (en dessus) 86’’.” In a description taken from a specimen of about eleven inches in length (Plate 24, probably an average individual), the form of the species is approxi- mately a short oval in which the ends do not converge enough to render them at all pointed. In fact the shape would not be badly described as subtruncate with the front a little the more rounded and the opposite extremity, across the caudal scale somewhat more truncate. The proportions of an individual of about this length have a width of nearly three fourths and a height of about one half of the straight length. A flattening on the back is usually most appar- ent across the third and the fourth of the vertebral plates. The descent from this portion is more gradual forward in the first and the second vertebrals and more rapid backward through the fourth and the fifth vertebrals and the caudal. The arch across the middle of the back is low and broad. The entire series of the marginal plates forms scallops the more prominent of which are the first to the fourth and the eighth to the hindmost, inclusive, at each side of the median line. The areolar spaces on the costal scales, and on the vertebrals are a low convex, those of the first and the fifth vertebrals being most prominent. The general outline is rather smooth or even. The concentric striae are strongly ‘marked on the scales of all the young. At the fourth and at the eighth of the marginals, on each side, the notches at the outer angles of the bridge are 288 THE GALAPAGOS TORTOISES. apparent, shallow, but distinct. The second, third, and the fourth marginals, toward their outward ends, are slightly curved upward; the first pair above the neck incline downward slightly; the hindmost one, the caudal, extends down abruptly and curves forward at the lower edge; and on the fourth to the eighth marginals, along the flanks, there is a low blunt keel. The width of the plastron is about nine elevenths of its length; the con- cavities on its lower surface are so shallow at this age they may be overlooked. The gular extremity is subtruncate, rounded, and a trifle swollen upward at the end in front. There is a weak notch between the gulars; they turn downward slightly at the front edge. The lateral outline of each humeral scale is convex in its hinder half and a very little concave anteriorly, forming a feeble sigmoid; and in the femoral section the lateral outline of each femoral scale is strongly convex. The anal scutes are separated by a deep notch, and each scale is longer than wide. In the young as in the old there is a considerable amount of individual variation, in outlines, prominence of areolar spaces, smoothness, — and in measurements. The specimens in Plates 24, 25, 26 are a little larger than the type to which the name 7. nigra was originally applied; they are part of the lot secured by the HassLer Expedition on Santa Maria Island (Charles). Plates 32 and 33 represent a group of medium sized specimens, about twice the length of the type. They exhibit considerable changes worthy of note in a specific description. There has been an increase in direct length as compared with the direct width and height, and an increase in the width across the curvature beyond that of the curved length as the width and bulging round- ness of the body increases, posteriorly there is a comparative increase in the height near the middle of the fourth vertebral scale and, in males especially a widening or flattening of the arches of the back without corresponding increase in height. The nuchal height has become greater, the humeral and the femoral arches have become higher and wider proportionally, and some of the marginal plates are curved and thicker at the free outer edges. The caudal plate has acquired a greater prominence as the marginals at its sides have worn away. The notches on the sides at the fourth and the eighth marginals, near the angles of the bridge, and the grooves above them are deepened; compression of the flanks anteriorly has steepened the sides and the saddle-shape, so called, has been increased thereby. From the scales the striae have disappeared in great measure, as also has the low ridges from the fourth to eighth marginals. In both sexes the plastron has become deeply concave; its ends are wider THE GALAPAGOS TORTOISES. 289 and more rounded. Between the anal plates the notch is more shallow; gulars and anals are widened, the former have thickened upward, the latter are swollen downwards. Along each side of the medial concavity there is a great rounded prominence. An extreme form of the female in medium size is that shown on Plate 33, a more common form on Plate 32. Specimens of about three times the length of the types are figured in Plates 29, 30, 31, 35. They emphasize the tendencies prominent on those of the medium sizes besides indicating others acquired on approach of maturity or depending on age, together with those dependent wholly or partly on sex. In the majority these features will be sufficiently evident from the illustrations. The large females on Plates 31, 35 as compared with those of the smaller group, Plate 32, have more fullness or roundness in the vertebral and in the costal plates and have much greater concavity in the sternum. They are even mcre concave than in the male, Plates 29, 30, fig. 3; the latter is more elongate and more depressed on the back and a trifle narrower across the humeral region. Being less deep in the sternal concavity may be a peculiarity of this specimen as the difference is not great. Plate 28, fig. 1 and Plate 31 pertain to the female described in Dr. Jackson’s article, The anatomical description of the Galapagos tortoise, Boston Journal of Natural History, 1837, 1, p. 443. Plate 29 is the male discussed in the same article; it is outlined on Plate 30, from Giinther’s figures in the Nov. Zool., 1902, 9, Pl. 16 and 17. These specimens were secured by the U. 8. 8. Potomac. Santa Maria (Charles) was the only island of the group visited. In the latter part of May, 1834, the vessel was at Boston, and in June the donation of “two gigantic Galapagos tortoises (living) weighing near three hundred and twenty pounds each, by Capt. John Downes (U. S. Navy) is recorded by the Boston Society of Natural History. The Potomac was at Santa Maria from August 31 to Sep- tember 10, according to Reynolds’s account, p. 547, ‘“‘a large nungber of the crew were daily on shore after terrapin, and frequently exposed throughout the day to a hot sun, with these immense animals on their backs, travelling over the broken lava.”’ The male of the Downes and Jackson specimens was examined by Baur and became the type of his 7’. galapagoensis. Afterward it crossed the Atlantic and was described and figured by Giinther, 1902. The -photograph, Plate 35, a fine specimen inscribed with the legend ‘“Suip ApicaiL 1835 Bj. Crark Master,” is probably from a native of the same island, Santa Maria. The ABIGAIL was a whaling vessel from New Bedford, Mass. The specimen is No. 11064 M. C. Z., received from the Boston Society of Natural History in exchange. 290 THE GALAPAGOS TORTOISES. MEASUREMENTS. (In inches). Straight Straight Curved Curved Length of | Width of Wi LOR AL length width length width Height sternum sternum 4477 102 8 133 133 4i 93 73 4480 122 104 163 173 63 112 gt 4478 15} li; 193 193 73 123 103 11074 243 153 303 32 124 193 18 11070 384 28 45 463 174 283 26 11069 423 324 523 56 203 33 30 1905 28 21% 323 363 13} 211 193 1904 273 213 35 38 14} 214 20 11064 41i 30 02 59 224 313 29 4668 26 203 33 34 13 20 18 For the type of Testudo wallacei these measurements are given by Roths- child, 1915: Straight Straight Curved Curved Sternal length width length width Height length 31.25 23.50 39 37.50 15.65 24.25 The straight length in the original description was 32.25 inches. TESTUDO ELEPHANTOPUS Harlan. Plates 37, 38, fig. 1, 2, 39-42. Testudo elephantopus HARLAN, 1827, Journ. Acad. nat. sci. Phil., 5, p. 284, pl. 11. Testudo ephippiwm GUNTHER, 1875, Philos. Trans. Roy. soc. London, 165, p. 271, pl. 34, 35, fig. B, pl. 37, fig. C, pl. 38, fig. C, pl. 39, fig. C, pl. 42, fig. B, pl. 44, fig. B, pl. 45, fig. A, B; 1877, Gigantic land-tortoises, p. 81, pl. 31, 32, fig. B, pl. 39, 42-44, fig. C, pl. 51, 58, fig. B, pl. 54, fig. A, B; Baur, Amer. nat., 23 p. 1040; Gtnruer, 1896, Nov. zool., 3, p. 329, pl. 20-22; Liprxu, 1898, Notes Leyden mus., 20, p. 126, pl. 3, 4; Gapow, 1901, Cambridge nat. hist., 8, p. 378; Hmiumr, 1901, Proc. Wash. acad. sci., 5, p. 57; Brcx, 1903, 7th Ann. rept. N. Y. zodl. soc., p. 172; StspENROCK 1909, Zool. jahrb. Suppl. 10, p. 534; Van Densureu, 1914, Proc. Cal. acad. sci., ser. 4, 2, p. 306 pl. 39-52; RoruscuiLp, 1915, Nov. zool., 22, p. 404. Testudo abingdonii GinTHER, 1877, Proc. Zool. soc. London, p. 66; Gigantic land-tortoises, p. 85, pl. 40, 41, 45, fig. D, F, pl. 48-50; Heturr, 1903, Proc. Wash. acad. sci., 5, p.57; Van DenspureH, 1914, Proc. Cal. Acad. sci., ser. 4, 2, p. 296, pl. 24-29; Roruscuixp, 1915, Nov. zool., 22, p. 408, pl. 21, 22. Testudo becki Roruscuttp, 1901, Nov. zool., 8, p. 372; HmtiEr, 19038, Proc. Wash. acad. sci., 5, p. 59; SrEBENROCK, 1909, Zool. jahrb. Suppl., 10, p. 530; Van Drensureau, 1914, Proc. Cal. acad. sci., ser. 4, 2, p. 308, pl. 31-388; Roruscuip, 1915, Nov. zool., 22, p. 404. Testudo bedsi HetuErR, 1903, Proc. Wash. acad. sci., 5, p. 59. Testudo hoodensis VAN DenBurRGH, 1907, Proc. Cat. acad. sci., ser. 4, 1, p. 3; 1914, ser. 4, 2, p. 313, pl. 52-55; SIEBENROCK, 1909, Zool. jahrb. Suppl., 10, p. 535. Testudo phantastica VAN DENBURGH, 1907, Proc. Cal. acad. sci., ser. 4, 1, p. 4; StpBENROCK, 1909, Zool. jahrb. Suppl., 10, p. 535; Van Densurau, 1914, Proc. Cal. acad. sci., ser. 4, 2, p. 299, pl. 30. | | Goel , y % THE GALAPAGOS TORTOISES. 291 What vessel carried Harlan’s tortoises to Philadelphia may not be known. They may have been brought by the same expedition bringing those to New York, under Capt. David Porter, U. 8. N., though no doubt there were _ numerous opportunities for transportation in the ten years or more elapsing after the visit of Porter and before Harlan’s date of publication. The references in the latter, mainly to Porter’s narrative, do not fix upon the donor, the time of arrival, or the island from which specimens were brought. MHarlan’s descrip- tion applies to such forms as are compressed and narrowed above and in front of the shoulders, rather than to any of the others. The dimensions given correspond somewhat closely with those of the specimen figured below on Plates 87, 38. The yellow blotches on the ‘“‘under jaw and upper part of the throat”’ do not distinguish from the form described as Testudo ephippium Giinther, 1875, and its closest allies. Harlan’s measurements in inches, are: Length of the back-plate, following the curvature 21.6 Breadth of the back-plate, following the curvature 22.6 Vertical diameter, or height of the animal 9 - Lateral diameter 14 Circumference of the body . 36 Evidently Dr. Harlan had specimens other than his type under examination. For he says, p. 287, ‘‘ The animal which is the subject of the present observations, is no doubt young, although larger than a similar species which lately lived for several months in the Philadelphia Museum. If we are permitted to judge from the shortness of the tail, and still less certain sign, the planeness of the sternum, our specimen is a female. Its weight is forty pounds.”’ On page 292 he remarks ‘‘The present specimen is living in the possession of Mr. Whitton Evans. For the drawings which accompany the description we are indebted to Dr. 8. G. Morton.” February 27, 1827, some months after the reading of his paper, September 5, 1826, the Museum recorded mention of the gift of a Testudo elephantopus from Richard Harlan, M. D., but did not state whether it was the type specimen, whether it was the one that had been in the possession of Mr. Evans, or whether it had been drawn by Dr. Morton. After the efforts of Giinther, Baur, Rothschild, Van Denburgh, and the authorities of the Phila- delphia Academy there are few grounds remaining for hope of tracing the type from recorded history. The only attempt in this article is made in comparisons of a specimen, probably of the same species, as nearly as may be of the size of the type with the description and the drawings. The specimen selected is 292 THE GALAPAGOS TORTOISES. No. 11063, M. C. Z., an exchange from the Boston Society of Natural His- tory of which nothing more of the history is now known. It is suggested that it may have served as Harlan’s type for description or Morton’s for drawing. The direct length of the type would be somewhat near seventeen inches. As for the drawings Dr. Morton drew them from a living individual which probably made accurate measurements difficult. The outlines on Plate 37, fig. 2, 3, are from the original illustration. The vertebral series of plates was about twice as wide as it should be; only half the width could be seen from the artist’s point of view. The scales nearest the feet also bore evidence of much uncertainty. On the same Plate 37, fig. 1, are the outlines of a carapace making the nearest approach of those in the collection to what Harlan’s type may have been. Plate 38, fig. 1, 2 show this carapace, with a sternum, fig. 3, that was fastened to it, by mistake no doubt as it belongs to a different species. In direct length this carapace is more than sixteen inches. Compared with T. vicina it lacks the dome-shape, is more elongate, is less broadly curved on the back and is narrower and higher in front. Placing this figure side by side with that of Plate 39, from a specimen of thirty-three inches, of the same species, brings out the close correspondence of the two. Seen from above the main difference in outline is due to wear of the reverted marginals on the older indi- viduals. On the smaller one the small areolar spaces are yet indicated and the striae are yet present, but have begun to show signs of wear. In connection with Harlan’s paper another series of measurements is given on page 292 of the Journal of the Philadelphia Academy of Natural Science, and on page 196 of the Med. Phys. Researches, in a note by DeKay: these are: Length 25 inches Breadth 19.5 inches Height 13 inches These dimensions may have been taken from specimens of either of several species; they were not taken from Harlan’s type and have no importance in the history of the species. Evidently the length and the width were taken directly instead of over the curvature. As seen from above, the carapace of a specimen of about two feet in length, Plate 41, has an outline that is oblong with rounded ends. It has a width of about three fourths of the length; in front of the mid length the shape is nar- rower, and more rounded; behind the middle it widens somewhat and above the tail it becomes subtruncate. The greatest height is nearly one half of the length and the outline curves with some regularity from the third vertebral THE GALAPAGOS TORTOISES. 293 scute back and downward to the lower edge of the supracaudal, which last extends back about as far as the marginal at each side of it. In the forward half of the body the back retains its height, even rises in males or some old females, or curves down toward the nuchal notch in most females and the young. Across the first pair of costal scutes the body is compressed into a blunt angle with its apex in the foremost vertebral; from this scale a groove descends along the hinder edges of the reverted marginals of the foremost three pairs and ends at or in the fourth marginal at each side. In front the opening into the carapace is nearly two thirds as high as it is wide, narrowing upward. The marginals form scallops along the edges. Similar to that at the fourth there is another indentation in the eighth marginal of each side, from which on the upper sides each of the posterior three pairs of marginals forms a groove to the caudal scale, by the recurvature of the scales. The concentric striae are present except on some old individuals; the areolar spaces are moderately larger and more or less convex and prominent; they are smooth except perhaps on scales of the very young. On the sternum the gulars are broad, about five sevenths of the width of the anals; they are rounded in front and have a slight notch between them. The outer edges of the humerals are concave, in large specimens; they are longer than the femoral scutes. The pectorals are short, but meet on the median line. The anal extremity of the sternum is short, broad and subtruncate behind. The anal scutes are broader than long, are thickened and swollen downward, and are partly separated by a notch below, much like that on specimens of T. vicina. The sternum is concave from the anals to the gulars; the concavity is greatest from the femorals through the hinder halves of the abdominals, farther forward it becomes shallow and disappears; it is less marked in females and is absent in young. The specimen to which particular reference is made, Plate 41, was secured by Dr. Baur on Pinzon (Duncan) Island. “The lateral ridges on the marginals of the flank are low; they become weaker with age. Females of twenty-one inches from the same island are higher and more arched on the back, less broadened posteriorly, more declivous in the first vertebral shield and are less concave in the sternum. With age the length becomes greater as compared with the width; this with the worn and broken scales of the second and the third pairs of the anterior marginals induces indi- vidual differences which might lead to selection as types of new species. The color of T. elephantopus in general is a black or blackish brown, marked with yellow in places on the lower surface. This species, including its varieties, is cone of the most distinct of the Galapagos tortoises. The variations of indi- 294 THE GALAPAGOS TORTOISES. viduals at different ages, or in different sexes, are such that measurements taken of specimens and reduced to percentages of the direct length parallel those from the other localities to so great an extent that they are of no real assistance in determining varieties or species. Testudo elephantopus Harlan has under various names been reported from five of the islands:— Espafiola (Hood), Pinzon (Duncan), North Isabela (Albemarle), Fernandina (Narborough), and Pinta (Abingdon). As will be seen by reference to the chart, Plate 1, this distribution might have been effected entirely by the currents after the species had secured a location on Espajfiola or it may in part have been brought about by the intervention of man. The agency of the currents, from the directions indicated by the arrows on the chart, is entitled to much favor. For it is sheer assumption that denies the possibility of drift and establishment by its agency on either of these localities to an animal of such buoyancy, as also that denying ability to swim or otherwise exert itself in the water to a creature possessed of so much muscular ability, endurance, and persistence in walking and climbing. While the buccaneers, the whalers, and the explorers have had something to do with distributing the tortoises since the discovery of the islands by the whites there is a possibility that earlier introductions perhaps including the earliest of all were due to aboriginal navi- gators of the times of the Incas or farther back. Considering the possible means of distribution, the amount of differentiation from such a species as 7’. tabulata, Plate 2, the rapidity of growth the variability of the individual it appears unnecessary to go back to the Tertiary or perhaps more than a few generations for the first appearance of the tortoises in the Archipelago. Among five so-called species placed under 7’. elephantopus there may be one or more entitled to varietal rank. All of them have been styled races, whatever that may have meant, but the individual descriptions given do not supply the details necessary for characterization. The original measurements in inches, of several so-called species are: — Straight Straight Curved Curved Sternal Sternal length width length width length width Height T. ephippium Ginther, 1875 33 234 40 40 24 214 17 is 5 95 83 43 63 33 T. abingdonii Giinther, 1877 38 22: 40% . 38 26 213 193 T. beckt Rothschild, 1901 314 2134 402 36 254 1634 T. hoodensis van Denburgh, 1907 22k 17%. 27 24%; 195 114 T. phantasticus van Denburgh, 1907 344 24 42 Al 24 18 THE GALAPAGOS TORTOISES. 295 MEASUREMENTS. (In «wnches). Straight Straight Curved Curved M. C. Z. length width length width Height 11063 164 113 193 192 ( 11068 25 183 29 30 123 11083 22 153 24 254 11084 16 112 183 182 11081 21 15 25 242 103 11082 21 15 25 253 1G! 11089 283 21 353 33 Number 11063, Plate 38, fig. 1, 2, was received in exchange from the Bost. Soc. Nat. Hist. Numbers 11081-11084, and 11089 were purchased of F. B. Webster. Plate 37, fig. 2, 3, are outlines from Harlan and fig. 1 is a lateral sketch from No. 11063. Plate 38, fig. 3, is a view of the sternum of 7. microphyes. Plate 39 gives the outlines of Giinther’s figures of the type of his Testudo ephippium. Plate 40 the outlines of Giinther’s figures of Testudo abingdoni from Pinta (Abingdon), Plate 41 represents a 25 inch specimen (M. C. Z. 11068) brought by Baur from Pinzon (Duncan). Plate 42, fig. 2, 3, presents outlines of a cast by Webster of a specimen from Pinzon, and figure 1, a dorsal view of another cast by Webster from Isabela (Albemarle) said to be from 7’. beckv. The type of 7. ephippium Ginther, 1875, was a specimen in the Museum of Science and Art, Edinburgh. Nothing was said to be known of its history, but Porter’s, 1822, description of the tortoises of Charles Island (Santa Maria) applied so well the author had no doubt the type came from that island. In 1877 in the Gigantic Tortoises this is repeated on p. 81, but on p. 11 and on Plate 39 it is ascribed without question to Indefatigable (Santa Cruz). It appears to have been the only tortoise from the Archipelago in the Edinburgh Museum; had there been another it would not have escaped mention. Baur, 1889, found that Captain Basil Hall had visited Abingdon (Pinta) (the only island of the group landed upon) in January, 1822, and remarked, in his Journal, concerning the tortoises ‘‘we took some on board, which lived for many months, but none of them survived the cold weather off Cape Horn. I preserved one in a cask of spirits, and it may now be seen in the Museum of the College at Edin- burgh: it is about medium size.’”’ The measurements given are those of a specimen of about the same size as the type of 7. ephippiuwm. Baur took the 296 THE GALAPAGOS TORTOISES. matter up with Dr. Traquair who replied ‘‘I have to say that I have had the records of the old College Museum searched for information as to the specimen of Testudo ephippium figured and described by Dr. Ginther, and the only entry which we can find which can possibly have any reference to that specimen is one in the year 1822-23 of a ‘Large Turtle from South Sea — Captain Basil Hall.’’’ From this Baur reached a most logical conclusion that the type of T. ephippium, 1875, and that of 7. abingdoniw, 1877, were both from Abingdon island. By comparisons of other specimens with the descriptions and figures he decided that they belonged to a singie species. Revision of the matter tends to confirm this, adding both names to the synonymy of 7’. elephantopus Harlan, 1827, and including as synonyms of the same species 7. becki, T. hoodensis, and 7’. phantastica. The types of T. elephantopus, T. ephippium, T. beckr and 7. hoodensis represent young to medium ages; 7’. abingdonw and T. phan- tastica the old and very old. " aoe ea ee eg FS ee ee eee 947) EXPLANATION OF THE PLATES. 3 are from drawings of Samuel Garman. The photographs are from the negatives of wl rt r 4 F a ‘ : ; tY en vir okt " | Ne Pecas ee | = Ls a" ect vi ¥ . oe - Reduced from chart currents. 1798 U.S. Hydrographic Office, 1915. "The arrows i : F > : one ea) a - i 7h 4 | oe i «= a ' = . » \ * i . | ~~ - « rain j = A ; Us. . " i € ‘ a . . — ~ ~~ ) . ..= - s —-o = ie - : LJ x cet.’ a ‘ id ers. Sipe Tee” |) ere Tee ee eee MEM. MUS, COMP. ZOOL. GALAPAGOS TORTOISES PLATE 1 Ya oS ; Marchena ie ( Bind foe OD eee CP ew Pin: 90° oe SAN SALVADO : iN (JAMES ) 5 ee ¢ FERNANDINA semis Rabidef) (NaRBoRoUGK) ‘ (Jervis) = | (puncan} SANTA CRUZ (INDEFATIGABLE ) #,) i i : SAN CRISTOBAL Santa Fé (CHATHAM ) ISABELA ; sm Crass Cea (ALBEMARLE ) eee ce 2 ried Sree cee ee ie . Sm oe C,Rose —qatil \e SANTA MARIA : (cnarves*"*) ; ; 4 ° ‘ oe ae : Esparola > Gardner GALAPAG OS ISLANDS 7. oe’ eee ee M. C. Z. 12050. Porto Rico. Length 223 inches. MEM. MUS. COMP. ZOOL. GALAPAGOS TORTOISES PLATE From Giinther, 1877, Gigantic tortoises, Plate 38. (7. British Museum, Cotyrr. North Isabela (Albemarle). eee 2 ty 1 ventral, 2 dorsal, 3 lateral. From Giinther, 1877, Gigantic tortoises, Plate 34, 35, fig. A. (7. micro British Museum, Cotyrr. North Isabela (A 240 pounds. t $ j “ r;, i iF? te ae we a - ‘ CZ i — 4) > , eas ve ae Se 7s ES | eee” 4 eee PLATE 5. 1 ventral, 2 dorsal, 3 lateral. From Gimther, 1877, Gigantic tortoises, Plate 36, 36, fig. B. (’. microphyes Giint! British Museum, Corypr. North Isabela (Albemarle). Adult female. Length 27 * * ork ‘ wal by ae 7) at) i ey 1 i) hit: , J — ie p ae, hat. Dekh ee ee ee o [ GALAPAGOS TORTOISES PLATE 6 COMP. ZOOL. MEM. MUS _ From Giinther, 1875, Philos. Trans. Royal Soe. Londo, 385, Pate, Z e be re: PLATE <2) eae Trsrupo viciwa Giinther, 1875. 7 1 ventral, 2 dorsal, 3 lateral. British Museum. Typr. Male. Lai wears ak ~- i * o' = . - > /. 4g : " 7 a = A . 4 e eal ae . 7 5 aa: 3 sal E ~* s 4 = “~~ . * dl > = ¥ ba a J -“ > — =e YS ? hs Pe ye ae -s = 7 & 1 ventral, 2 dorsal, 3 latera From Giinther, 1875, Philos. Trans. Royal Soc. London, 166, Plate 36, fig. C. Royal College of Surgeons, London. Length, 22 inches. + ae ’ é “+ «< GALAPAGOS TORTOISES PLATE 8 - “i From Giinther, 1875, Philos. Trans. Royal Soc. London, 166, Plate 33, fig. B. Royal College of Surgeons, London. Length 41 inches. _ e : = tied m ~ . ‘ ll = 2aie, ayes ie mo > Pa —- = . = 7 ee a : os. PP MEM. MUS. COMP. ZOOL. GALAPAGOS TORTOISES PLATE 9 PLATE 10. GALAPAGOS TORTOISES PLATE 10 MEM. MUS. COMP. ZOOL. 1 ventral, 2 dorsal, 3 lateral. MEM. MUS, COMP. ZOOL. GALAPAGOS TORTOISES PLATE 11 PLATE 12. — Ray GALAPAGOS TORTOISES PLATE 12 MEM. MUS. COMP. ZOOL. PLATE 43. , 4 = eee ee ee Y te 5 a — Lat ‘ - Ax MEM. MUS. COMP. ZOOL. GALAPAGOS TORTOISES PLATE 13 PLATE 14. Testupo mickorHvEs Giinther, 1875. 1 dorsal, 2 lateral, 8 ventral. M.C. Z. 11066. Villamiel, Southern Isabela (Albemarle). Googe Baur LOOL. GALAPAGOS TORTOISES PLATE 14 MEM. MUS. COMP. PLATE 16. From Giinther, 1875, Philos. Trans. Royal Soc. London 5 nec Harlan). ah : - University Oxford Museum. Length 31 inches. 7 id * MEM. MUS, COMP. ZOOL. GALAPAGOS TORTOISES PLATE 15 % Mm) ea) SU ee OP ee Be 2 eee eae ne <<. MEM. MUS. COMP. ZOOL. GALAPAGOS TORTOISES PLATE 16 a ae 7 ee) eee ee M. C. Z. 4479. Santa Maria (Charles), Hassler Exped. MEM. MUS. COMP. ZOOL. GALAPAGOS TORTOISES PLATE17 PLATE 18. hs eft > wari fi tt | ; ind yt ee AL ‘ ep Fides > a > ‘ Le tr if Lae Je { A wton tthe on ee *s ; ) ee Ae 7 oe PS 04 » « =—% 9 o \ : & W dh, —— a. cn o Wey 4 Py 4 «! * 4 er 7) i - = Lad | ? cr ¥ 5a) . * - * 4 PLATE 18. TESTUDO MICROPHYES Giinther, 1875. 1 ventral, 2 dorsal, 3 lateral. M. C. Z. 4479. Santa Maria (Charles). Hassler Exped. Length 15} inches. s 44 4 MEM. MUS. COMP. ZOOL. GALAPAGOS TORTOISES PLATE |8 | PLATE 18. TESTUDO MIcROPHYES Giinther, 1875. 1 ventral, 2 dorsal, 3 lateral. M. C. Z. 4479. Santa Maria (Charles). Hassler Exped. Length 154 inches. * ‘ . ~ m ' "iyo Ni? oe oh mel i oF ~ WAY ay , i rhs AP i CT beg : { ' iW “age \ a iy Je} , Ne p el Be |) . be ) ae . wn . e *- oT) > ; PLATE 19. ‘Testono atenopirns Gunther, 1876 1 dorsal, 2 lateral, 3 M. C. Z. 4476. Santa Maria (Charles). Louis Agassiz. MEM. MUS. COMP. ZOOL. GALAPAGOS TORTOISES PLATE 19 MEM. MUS. COMP. ZOOL. GALAPAGOS TORTOISES PLATE 20 PLATE 21. M. C. Z. 11075. Type. a : oT 1 dorsal, 2 lateral, 3 ventral. Boston Society of Natural History. Length 25§ inches. td « a. —_— a ; * 94 iE MEM. MUS. COMP. ZOOL. GALAPAGOS TORTOISES PLATE 21 PLATE 22. 1 dorsal, 2 lateral, 3 ventral. M. C. Z. 11072. Typx. Boston Society of Natural History. Length 24 inches = % s * . C ie : - 5 ; “ { . « at , ee i ‘ fl +. - -_ . . + = e - re a J, . a th, > aan = oa MEM. MUS. COMP. ZOOL. GALAPAGOS TORTOISES PLATE Gome 1 lateral, 2 ventral, 3 lateral. = 1 from Wiegmann, 1835, Nov. Acta Acad. Leop.-Carol., 17, Plate 13. Weight 125 } “ 2, 3 from Quoy et Gaimard, 1824, Voyage Uranie et Physicienne. Zool., Pla MEM. MUS. COMP. ZOOL. GALAPAGOS TORTOISES PLATE 23 a. J PLATE 24. TEsTuDO NIGRA Quoy et Gaimard, 1824. 1 dorsal, 2 lateral, 3 ventral. M. C. Z. 4477. Santa Maria (Charles). v dl » . es : NN eee ‘ pent ae < ty FF al 7 a La ay COMP. ZOOL GALAPAGOS TORTOISES PLATE 24 MEM. MUS. 4 j| ' sobvclw { A y y “A PLATE 26. - th 13 ine 4480. Santa Maria (Charles M. ©. Z. MEM. MUS. COMP. ZOOL. GALAPAGOS TORTOISES PLATE 25 ' - > . ~ = . 4, M.C. Z. 4478. Santa Maria (Charles). Hassler Exped. Length 15 inches a GALAPAGOS TORTOISES PLATE 26 MEM. MUS. COMP. ZOOL. 7, o) oe PLATE 27. —o — wT ———— a ——— MEM. MUS. COMP. ZOOL. GALAPAGO S TORTOISES PLATE 27 M. C. Z. 11070 @. Santa Maria (Charles). Boston Society of Natural | MEM. MUS. COMP. ZOOL. GALAPAGOS TORTOISES PLATE 28 M. C. Z. 11070 3. Santa Maria (Ch a MEM. MUS. COMP. ZOOL. GALAPA GOS TORTOISES PLATE 29 ‘ ‘ ‘ VAN A 7 v iw PLATE 30. TESTUDO NIGRA Quoy et Gaimard, 1824. ee a if 1 dorsal, 2 ventral, 3 lateral. ae at. le i PLATE 31. ww M. C. Z. 11069 9. onl OL. GALAPAGOS TORTOISES PLATE 31 PLATE 32. of: o PLATE 32. 1 dorsal, 2 lateral, 3 ventral. 1, M. C. Z. 1905 9. Santa Maria (Charles). Hassler Exped. Length 27} inches 2, 3 from Giinther, 1902, Nov. Zool., 9, Plate 17, fig. B (7. galapagoensis Baur t 4 * - : . Zs ain Ra + oe i ‘pe ” By: yn lp ib ree Pas “> * 4 A i g 7 S —@ ‘ "2 PLATE 338. v= FG ide ea RT aw) oa a. Fa -% 4 ) ‘ Q M.C. Z. 1904 9. Length 273 inches. © MEM. MUS. COMP. ZOOL. GALAPAGOS TORTOISES PLATE 33 PLATE 34. TrsTupo Typica Garman, 1917. 1 ventral, 2 dorsal, 3 lateral. M. C. Z. 6260. H. A. Ward Coll, Lengths) incies of PLATE 365. : : 1 dorsal, 2 lateral, 3 ventral. — M. ©. Z. 11064. Boston Society of Natural History. Length 41} inches. ‘ , . PLATE 35 GALAPAGOS TO RTOISES MEM. MUS. COMP. ZOOL. o. MEM. MUS. COMP. ZOOL. 36 GALAPAGOS TORTOISES PLATE ————<<—— -—S—i‘“ ECM! PLATE 37. , a a 3 oA » - . j =e : Soe | = ha 4; eae -s a 1M. C. Z. 11063. Bastdin Bosieby of Natural cuaieeye 2, 3 from Harlan, 1827, Journ. Acad. = Sci Philadelphia 8, Plate 24, “Med. 1835, p. 190. * “se ° a w =a iu“ a . a ‘ e * ae - _ a - ad Ds oP —s ry & _— SS ia y me 9 aah ‘ _ as a —— “= ? - _ = Pie ; : _ - is _ GALAPAGOS TORTOISES PLATE 37 well f > PLATE 38. PLATE 38. r 1 dorsal, 2 lateral, 3 ventral. — 1,2M.C. Z. 11063. Boston Society of Natural History. Length 164 inch 3 M.C. Z. 11073. Boston Society of Natural History. Length 12 inches. 4 s * " fig ~ > ? < ._ «sh ane ale Fe sal ao ; fe Ps eae vs we ‘ Es 7 t Na —< ad > Baal | 12 Pc 5 —— - roa MEM. MUS. COMP. ZOOL. GALAPAGOS TORTOISES PLATE 38 a ep, oa - ~~ a PLATE 39. '. aro 1) A eee ~ - ie ee i Sn . | P te 9 ‘ 2PHAD ft US - 18275 1 ventral,.2 dorsal, 3 eel From Giinther, 1875, Philos. Trans. Royal Soc. London, 165, Plate 34, 35, fig. B a. Giinther). Mus. Science and Arts Edinburgh. og Type (7. ephippium 33 inches. wy, bee 5 a0 Be - Fe ’ cated "a e,. r . t - * s - ‘ ' r * a -. : | ¢ - = _<¢ * a “o ~ s wae. ‘ 0 - aa = > _— Pig < ¥< . : -— ” » * . : =. gd eee ee om - A AG ee fs am x. ae Fe ee ¢ y ~~f tn ATs. a aa > Al ha > > PLATE 40. TrSTUDO ELEPHANTOPUS Harlan, 1827. i ventral, 2 dorsal, 3 lateral. From Giinther, 1877, Gigantic tortoises, Plate 40, 41 (7. abingdoni Giinther) British Museum Typ. (7'. abingdoni Giinther). Length 38 inches. ?, * ‘, kee = tine a \ ~ 1 } i E M. C. Z. 11068. GALAPAGOS TORTOISES PLATE 41 MEM. MUS. COMP. ZOOL. mi Py ad € iwe™ — - lo a . é a at a =. = s fe " one ah a ea Soph a Wy > cr a | ~~ ma 7 7 7 _ © = : at Pe — a ees Se = -- y > . = a <> 7 a i ad " Pan? an) - 4 We - . wt -s (eee a i ww re PLATE 42. - TESTUDO ELEPHANTOPUS Harlan, 1827. 1, 2 dorsal, 3 lateral. — rd ¥ 1M. C.Z. 11084. Cast (7. becki Rothschild). F.B. Webster. Length, 16 inches. 2,3M.C. Z. 11083. Casr. F.B. Webster. Length 213 inches. ’ ’ » ~~ me +: Sea ee « 7 < se "ib ee yot Am td fay ve ‘am : Se | , ad - . nd , ; eye fad “od “ a * a t A 5 einen " BROS, Ps § PUBLICATIONS OF iia i RAIS : “bees oe wine ae ae Vols. LYV., rent LXE, 4 | LXIL of the Buus geass XXXIX.,. ae XLY. to XIX. of the: Mass. ' BINDING SECT. JON6 1988 ay ene oe eed ay eee ty erie tie oe th al Px Tl shreepsee ee eee Seat kesittas SIRE a AT gy SS Stpre es aie ewes ete Pash Shoe eles arn eed Oe is kre ers : crepe ¢ — Resryeceticint Se ced Eamets Steen Srkearre: srllonets ry Oe “ v- : " erate Ir eet - 5 29-4 a POL Aenea oes hicabre tt) > 4 weit. vate Seren bob Pee oe “ae ay merle sere ees pliwe! Sao ‘pe Se 2 " alte — Pastrt sat be oh be! ae we steer bybreetirt 44 Peete ate stp os oe AseU ghey vets renee He oa Reb p Sab 276 bi rote be ive fi Erwin ” Le) aweeaee ree re La fy ttn dhe rhgie 4 Seki tae Testers we ke ible pnube wives wo b Vie be Weal teeth >) bebe tt petty bedew hh Pte hrs Sr bele eye Oe eke ot me vey © oT Seaghayy . 4 Oe ayer ed Ae steanesct wait Gees { fe cher SRS dat : eee * cores aes * * Seksten eatyore hy cot Sehee Saari oaeksits silt Wrist) bee Be (hs Bering. ita) ¥ CPAs ereherriets oe pan Bega em ave Pearce a + epee & ren gens AAG Larabee i iples. Siehe ve Shere y= - : x , Die Ue hep eek by 9b) year Fate at) foie ae , pen S03 ~~ Wrasse Shp nhewss shots ren ¥ A hee bee ubfecpeheaate he 4 pebehht tf ; eit wh es be nf asieetnie Pete tet cy pa hee ee Shae h ry Rr — be a ne ro ee sot ST ress I; seperate, Say tease te Ly MORI cehakee -- , Ye: Volek ner Srgalooearigiay — Sa lepwlypereres er merce ee end On ATS we ety * ste erties The sow! tehrerte absent wiping lenrxetnel ay ra eee 2 paar ia ts 2 ¥ ianehe' pares feneney ity Ue iscpetar; ise SUT ae foes pereoke a me Soe paket me taeny cared Rei pmo. oe pn Yateley de Sagat wepesceel baits a ve 4 ee hed a =: tt ‘ we reece the new : oe £00 ye et 4 ye Sakis We bem TELS ee aioe Ctra ime meus cgeatva tial aaa iy assy tatty \F oeleareh a tary S484) oF Pity Debebe te bab of acupiet ee Be spikes i PAW ryt’ ate igh ld fn be ve be at : { . gta ube Bi herb Velen sf - t Dike ab hate ‘ 4 Tumse A bed ; hy F heeh hey ints wf» te Oe ihe ake Ch red oleae 7 es Pky yn eve ye eeiits tee it pdytg hs a ? ; ge ‘ pee ba ope iby Spottt bay iene +42 a ie 8 ie, a pe We hone be ‘ iad “ Ws st rle Fy re ee phshes: Pes! lad aes? wie heh! inant owe ; a seh ee faa) a ed) bo (3 hittin cities sree ye bees Writin te ter pbad yes vols tek avitae Wp -tchder Sy > dab; Alter ike tes Nise ie Palit eee A th aad he erryh n ay in nyse ian { | et i Povorp aber ye! eeeree Wade Cts th hohed Cakes Bien Paya St eres uae re ¥r CO EILE thie o. pe 4 Thi bes epee - ya tan ’ ede baheds Sete Bi it Pld dead? Jebigs Tete pir: , We oe ub atc! or) lle wep: t mat) paths Gaia if Sire by f / whee asta ‘ St he te Vn edhcpeichlg yeoe > ie by bie Hs tt play Eeaes hh Dabs a hot + Arey TP howe Pe, Fe WEbL ta pepe Te te orp yeh lichs heaal Athos n 844: Th. We pple te aphe vite Vib arbew bees oe : thd ara on heirs revere rpayit 3 itegee: pee eleadeianrs \ so Fg bs Reds) pwr me! PMP / pentoepect] ‘ “ : ; 5 ; ba fry SAY Areca gor ei evet Ts phage we eet het - cri et es Neri ersesitee ee 4e8 ruebelb isk 1 (pris pie Ad Pray saa Ace haleueh * ¢ al bev] see she hoe pre deb pie BE. Weert ad 4 ej mmo ppechbaalper -= treat sie Tape he aad bps ops tabs oplrite er or Maes An at Threat pane ee Tes rater) cr rink dee sNppabecwneeeny iret’ pitts, ary ise chehe aig Cine ved areP Oh bm has Le te rele Shh eppeaile up rer oie - bi kee mite Bises) heoeie eect “or rae certs ue Marana! eetatiis * sa eed Fea Sparen: a a Teak tis, Mivacet ine etopend Ae anheieogety vee Ih ia ie feet sires eee tetee: ; Becht yr bess) cs herete ish’ dae noone $- bnnc hs pubs othe Lelippn oe fe beleive sete re Rh yet n | mwprn Ip) pps veh raetyy A ns o Mey rieeeb ’ oh oy pf eee Jeg. z0ne toes! rrr a Maat oer st tity aed as isd pe reest vay eaten thay tied pdydare align De ol ht pian kag oe) ame Gadi Vadadsnerdteke teeeie thee et rebrtee icky teh iebeiaaeesaae Fela wo Ft pe veh ghe ; oes ae pe Tee ya 1 ; t epmtnn aye vy . Net e'gtpss Dear beg! stare Seercgas ei tes betyeseteers eee Hy eabee ee at 4d “apr Petersen iow cesert ieee enh AULA Ae El ola be aye woke viene gp eh ee ~ jartmaielbeaey| Ape Deed Vio Wet) bad ne be re Dame ede os igs ststrh y rey + ae nos, escretesety > " ise A Aiatts ris See ie Fpoeries heb ong wet peter sgetane wel TTS her shy 9) Pa eesurbtigerares sre tetes be py bare 4 SPs te SL atl ddeenallt | lee ay ve em minn vote ys Or ied 1g BE uote pen ye beenp bree # errser re a> phe et or my ~% oi ipa petn mere ne hehehe pb enh en patie bee emir bese Aneel 1) a+ hieweiee br org gigs kine mepe Ze tp J Hy ta D penta serge en Sitecssteratiba gr tetee heel ieees of . ee itedets - bata be Vr apte sre i (ot) oer Thighs Sabbacp oes ne hres parse oy oT ar Coens wocay = Apes > othe be ok oe emne Pee the pe Joly ebthe ih Rott abtes ve ae “re thats Shemp heme b: byt ta (2) as ae bt