‘ Vf eit daca qeasise Ped ae bed D RAV ED ted LW wren be ts uy “ ergot : degen wessich wd eaerdal we Pease n ean sitet Hae Pr Aue dodge theld tate sei aetionn Tb eed eee dedrece @ 0 da ile? “te SPO br tad me re aK itorgtanedepsatied: te de 1 Woanas ee ae ae ent ta Wee eRe Pe 4 det ede ted jth ane AG rrhtaet cand -g-4ed: de ee ehnten gy. 4-@ ase sree Nadie te a r) or) bb i ei ataeas Gee Med Y, datebed Linde? 4 terest Led qed Or seat da teidie ye eaees nee edad ante 44) ° ieeaa VA Wed de toto ” tad tae ioe Wa se ‘aaa et Te) or % tt yataid ou G Masten 4 vats de aK fay deed i928 ty aa a gedde ry tat we wear reer st 2 Speen was he A te Sab tay “1 Py oie OT oY 4 iat f a a ie: en alow a GW aw a 88 GANGA GoM ay Te NN ET . aie ay) shh ’ qint dane Weeastaetious bets ‘ VP ehh 4 woe aa Hu HRD orn ed Solna unin Ot btm nh 8 Mele Sta signs e's 40 ty 19% ian AO ey vs Ay bet HE ‘en Sethe ha ein Racer ASiclgneavaics® tw wind phadeqar at Genie oT han aye PM eeraseclertiss Hoek aS hat i “4 was alas avy es eS Si satitetavtatatie Slates nes fale betes Cities tana (on) Prats he ‘gaia aot rr eae pigs ea) tow rasan dail geen alt Vatatalely geen iV D kaalats isa oe ty hs one aril Le Ae AAV) oct fry 4" ty eat i Ht) rr “haath ora Het tat ah idysias aes jectes, en “04 iy 4 poe ate i) 4 ny ti signee , My ite cea aK $s Bah ed se ot Adah ies Apne: Soak a eeaLA he Arh ed Aids cata Littersent A ew ide hewn vent od ie Wire rtmidee an ii ¥ate iets at ets paren it He (A wie a A heh areas + i +) ri = ny 0 tstthot Wine Haein elhdt ” vateg wih Ae ta4 wad rr: Stk eres see Vga at 94 3 sera ayg , a2 ysibed 44 Leah ait id feats Nota na Aa @usaietiyee “ aaa ed NPCS to weeds m8 th ~ ‘ Poyteh i his oy re Aoeek whe mae tnt Taye eyed erro SO a Ok Mace an We Hen fier ‘ aradye wack year ae ban Abe) Dede bby shag Ao oon cat te Pn Febed ehh Hak dat se I: Me, deh an aentes wes Hegre as F8y vi eee sy Ae We ee eG Ss : ee Yoh ee Siete ate Ozma bell shi rates Rivasiiyss FAvgt Srl be be Tk y Lyakh ies Sd Reisen eH ve: Pater eat 44 zy Taisen bad weeds wi VS WE Ri Mb ekad ae Woe hen ayre vp ae ate wy my oy Wiededruantieg eine i oye ene rei “o USP UN ty evoke we at vara at Pawiprvets ¢ Wits sett MSU bat vate sowie nmi y bad 7m l oe Most Win se ‘ eee Se Cue to ‘ Sine Sep uSutba say nee SW GI Se or av yn ne, Baie ae. ae bg hey Hine acy aie Seep weil ay A Tecrker cen Nt, suas Reriee Bid ayy arr Ase Sata feo rosie aint ra ae a Fie ate vase «town i *, - ao verse ieee Sie : fone tee? Sis aatte rp oat wate a aha 5 iat ged st he siete tat 4 oh, Bel Ay ig sgh ot a ie ivi bh tek seeds Seite itu rect ae it me rio is apniotey sit i ee satiate een ah ne wo a Beats re Dandnininaly Dae Cs ip Hie erin) aii: Nb ANY SFO LIV AH) pees Sieoe aes ot ae uth: PASbP Eh ae A ots dic Lyng + be “ iar hbed ytd atte ‘ th ies a ia) Pastal ’ *" ee ya i Was tae as sone beats iS a a ! OY Asa! hy by Hi 5 lal i W Vo | DIF Hl i Wy Xt AA a sb i } ye ae incre > EARNER EX LIBRIS Wilham Healey Dall Division of Mollusks Sectional Library ews | os EDULIS LRA EREG Lai Sp + F 4, wt - - a Rs f 7 o ret ae Ma Coe EN, ES ae eT ee ee 8 yy ; ee oe i) A a i eh de an ay K- (di ehialt ial em in oe i ae a ‘shige ‘ya 5 ‘ 62 SIR CHARLES ELIOT ON : [May 6, [ From the PROCEEDINGS OF THE ZOOLOGICAL Society OF Lonpon, May 6, 1902.) On some Nudibranchs from Zanzibar. By Sir CHARLES Exrot, K.C.M.G., Commissioner and Consul-General in the British East-African Protectorate. (Plates V. & VI.’ and Text-figures 2-5.) During the last year Mr. Crossland has been most kindly investigating for me the fauna of the eastern and western coasts of Zanzibar. He has not only collected a large number of Opis- thobranchs, but also greatly increased the value of his collection by drawings of the living animals. The present paper contains some of the results of his labours in the shape of notes on three apparently new genera of Nudibranchs—Zatteria, Dunga (.Holidide), and Crosslandia (Scylleide), and on two interesting species on which little seems to have been written since the time of Alder and Hancock—Melibe fimbriata and Madrella ferruginosa. The Aolidide are already divided into forty or fifty genera, and it is with reluctance that I add to their number, believing that it would more properly be reduced. But as long as the definitions of the existing genera are so minute and narrow, they cannot be made to accommodate fresh forms, for which new, though probably only provisional, genera must be created. ZATTERIA BROWNI, gen. et sp. nov. (Plate VI. figs. 9-13.) Three specimens were found in seaweed collected on the reefs round Prison Island, in Zanzibar Harbour, in May 1901. The largest was ‘8 cm. long by *2 cm. broad. The body is long and narrow, and terminates in a peculiarly slender tail, which is nearly a quarter of the length of the whole animal. The cerata are arranged in eight transverse rows (PI. VI. fig. 10), each row contain- ing eight cerata, four on each side. The first two rows and the last four are crowded together, but the two series in the middle are separated one from another and from the anterior and posterior clumps by considerable intervals. The most distinctive character of the genus is the shape of the cerata (Pl. VI. fig. 11), which are not even but swell out into two or three projecting rings, the first 3 For explanation of the Plates, see p. 72. [2] *) 1902.] NUDIBRANCHS FROM ZANZIBAR. 63 a little below the tip. Above the mouth are two short, slender, erect tentacles. Behind them are the rhinophores, which are considerably longer and more slender than the cerata: they bear about eight bracket-like semicircles, which alternate with one another, so that there is not a complete circle round the rhinophore ; at the base of each rhinophore is a black eye-spot. The foot is rounded in front and the corners are not produced in tentacle-like expansions (Pl. VI. fig. 13). The body is translucent and colourless (the viscera being white) with a few blotches formed of opaque white dots. The cerata are also translucent, except at the rings, which are opaque white ; they bear a few orange spots or streaks. There is a long orange streak on each rhinophore. The jaws are small and the masticatory edge is finely denti- culate. The radula is short and uniseriate. Each tooth is shaped like a horse-shoe and bears on its anterior margin one large denticle with six small ones on each side (PI. VI. fig. 12). No trace of armature was discoverable in the reproductive organs. In many characters, in the disposition of the cerata, the rounded anterior margin of the foot, and the buccal parts, the animal appears allied to C ‘ratena ; but it differs in two points, the rudimentary perfoliation of the rhinophores and the rings round the cerata. The latter peculiarity is, so far as I am aware, unrecorded among the Adolids, but it almost entirely disappears in specimens preserved in alcohol, and it is therefore possible that it may really exist in other genera which have been described from such specimens. DUNGA NODULOSA, gen. et sp. nov. This animal is fairly common on colonies of Sertularia. The body and tail are both long. The cerata are easily detached and have then some power of independent movement. They are carried very erect in the living animal and are set in transverse rows varying from four to six in number. Behind the last transverse row is a clump of smaller cerata, also of varying number. Probably the caducous character of the appendages has something to do with these variations. Each transverse row consists of ten cerata, gradually increasing in size from the outside to the centre, the two middle ones being much larger than the others. The outer cerata are of the ordinary cylindrical shape; the middle ones are swollen and ovate, but terminate in a fine point. At the top of the broad part and at the base of this point are eight knobs. The rhinophores are very long and simple. The tentacles are moderately long, and the anterior angles of the foot are produced into processes of about the same length. The foot‘is narrow and without markings. The length varies from °5 to 1:2 cm. The coloration is very variable and ranges from clear light yellow to purplish brown. These differences may be partly due to two different colours of the liver diverticula seen in the [3] 64 SIR CHARLES ELIOT ON | May 6, transparent cerata, But in all cases the tips of the cerata are pink and the knobs of a brilliant white, with a white streak extending upwards and sometimes with white spots below. The back, cerata, rhinophores, and tentacles are covered with small spots of the same colour as the body but darker. The rhinophores have usually, but not always, dark circular bands. The intestines, which are clearly visible, are light or dark yellow. The jaws are of moderate size; the masticatory edge is bluntly denticulate, but on the lower part only. The radula consists of a single row of teeth. The central part of each tooth is pro- longed into a short bluntish point; on either side are three denticulations. I could discover no armature in the reproductive system. The general characters and inflated cerata of this genus resemble the Tergipedine, and the figures of Capellinia ee (by Trinchese) and those of Tergipes (Capellinia) doric (by Vayssiere) represent the cerata of these species as nodulous. But the Tergipedine have the foot rounded anteriorly, and the arrange- ment of the cerata in this animal is peculiar; it therefore seems necessary to create a new genus for it. CROSSLANDIA VIRIDIS, gen. et sp. nov. (Plate V. figs. 1-8). In July and September, 1901, were captured at Zanzibar four specimens of a nudibranch closely allied to Seyllea, though strikingly different in external appearance. The four specimens seem to constitute a new genus and possibly two species, though one may prove to be merely a well-marked variety. The animal in question may possibly be a Verea, Lesson. I have not access to the original authorities, but Fischer’s ‘ Manuel de Conchyhologie,’ p. 536, says: ‘Le genre Verea, Lesson, 1830, a été placé dans le voisinage des Se ylloett Rhinophores courts, coniques, ciliés, visibles au dessus d’un petit voile frontal: téte courte, tronquée en avant; corps fournissant de chaque cédté deux lobes; branchies disposées en petites touffes sur les lobes latéraux et sur la queue.” From this description and from the fact that Bergh, in his ‘System der Nudibranchiaten Gasteropoden,’ takes no notice of Merea, it may be presumed that the characters are not sufficiently defined to constitute a valid genus. The length of a large specimen is nearly 5 centimetres, and the general appearance superficially resembles Hlysia and in no way recalls Seyllea, which, however, I have never seen alive, although I have examined numerous alcoholic specimens. The body is fairly long ; it 1s produced into a neck and tail and laterally into two wing-like lobes, one on each side, which are more or less distinctly bifid, but in no specimen can be compared to the two pair of cerata found in Scyllea, and are not constricted at the base. The animal when crawling generally directs them laterally. The colour is vivid green, harmonizing exactly with the young leaves of Zostera on which the animal was found. At the side of the body below the lobes is a row of irregular projecting sandy [4] 1902.] NUDIBRANCHS FROM ZANZIBAR. 65 markings. Round the edges of the lobes, the angles of the body, the ridge of the tail, the cups of the rhinophores, and the frontal veil runs a brown line. The surface of the body is covered with microscopic brown specks, which here and there are aggregated into spots just visible to the naked eye. There are also a few other spots. There are no anterior tentacles, but a small frontal velum. The rhinophores are perfoliate and set in little open cups on the top of fairly long pillars, which are usually held nearly horizontally and have not any process behind as in Scyllwa (see Pl. V. fig. 5). The back and inside of the wings are covered with co slourless, transparent contractile branchie similar to those of Scylleea (text-fig. 3, p. 66). Oa the tail is a shght ridge, not amounting to a crest. The foot is very narrow. Text-fig. 2. Digestive organs of Crosslandia viridis. @., esophagus. gl., gland. | st.pl., stomach-plates. The body-cavity is spacious anteriorly until the commence- ment of the liver. After this point it is quite narrow, owing to the thickness of the soft transparent body-wall. The j jaws (PLY. figs. 7, 8) are large and hinged dorsally. The masticatory edge is smooth and flexible, of a deep brown colour and bent outwards, the muscles being Aitached at the bend and covering the large stiff cheeks. The radula (PL. V. figs. 4, 6) is short ; each row con- sists of a median tooth and about 30 laterals on each side. The median tooth has a fairly large spine in the centre of the anterior margin and about 5 denticulations on each side, of which the pair nearest the centre are considerably larger than the others. The lateral teeth are also denticulate, but var y somewhat in form, and the corresponding teeth in the different rows do not always Proc. Zoou. Soe.— 1902, Vou. Il. No. V. 5 66 [May 6, SIR CHARLES ELIOT ON General view of the intestines of Crosslandia viridis.—The body has been opened by a cut made somewhat to the right and dividing the heart in two. a., auricle. $1, 62, $8, three hermaphro- | &., kidney. 6., buecal mass. dite glands. 1., liver. br., branchial tuft. gl., gland on esophagus. v., ventricle. em.s., cantral nervous system. i., intestine. One branchial tuft enlarged is shown separately. [6] 1902. | NUDIBRANCHS FROM ZANZIBAR. 67 ° i . agree in shape. Those nearest the rhachis are generally denti- culate on both sides: the ordinary teeth are denticulate only on the external side: those towards the end of the row are again denticulate on both sides but of a peculiar form; the outermost are degraded. On the cesophagus lies a large gland of apparently salivary functions. The esophagus, which is narrow, broadens at this gland, and the digestive tract continues of much the same length until near its termination. On slitting it open (text-fig. 2, p. 65) the traces of a stomach are found, and an internal con- striction is caused by the presence of a ring of large horny teeth. Just beyond this point is a large typhlosole with an irregularly laminated interior surface. The anus is lateral, beneath the right wing. The liver is in two compact masses, anterior and posterior; they send forth very slender light green diverticula, which until carefully examined have rather the appearance of veins, to the base of the wings and rhinophores. Text-fig. 4. Hermaphrodite gland of Crosslandia viridis. amp. § gl., ampulla of hermaphrodite gland. m.gl., mucous gland. Sgl., one of the three portions of the Sp., spermatotheca. hermaphrodite gland. v.d., vas deferens. The ganglia in the central nervous system are distinct, the pedal being ventral to the esophagus. m4 5* 68 SIR CHARLES ELIOT ON | May 6, The pericardium (text-fig. 3, p. 66) is embedded in the body- wall: its pulsations are visible externally. The kidney (text-fig. 3) is spread over the liver, and also on its ventral surface, as a number of distinct branching tubes, which continue in front of the liver, lying loosely in the body-cavity. The hermaphrodite gland (text-fig. 4, p. 67) consists of three granular, spherical bodies, somewhat on the right side of the liver, one at each end and one in the middle, but not fused with it or embedded in it. The ampulla is large. There is only one spermatotheca. Prostates are absent, and the penis is small and unarmed, It will thus be seen that in its internal structure this animal closely resembles Scyllea. The only important difference is that the hepatic diverticula are very small and extend only to the bases of the wings, whereas in Scyllea (? in all species) they are said to penetrate to the ends of the cerata and into the branchial tufts. I have wondered whether the creature could be a young Seyllea in which the bifid lobes would subsequently divide into two pairs of cerata, but the size, which is as large as that of most Scyllec, renders this improbable. Taken in conjunetion with the character of the liver, the external differences (the wings instead of two pairs of cer: ata, the absence of a caudal crest and of flaps behind the rhinophores) seem sufficiently great to warrant the creation of a new genus, which I have named Crosslandia after Mr. Crossland, who dredged the first specimen. One of the specimens (PI. V. fig. 3) showed marked peculiarities, and is certainly a well-defined variety if not of a distinct species. The body was stouter and the outline more wrinkled and indented. The colour was that of Fucus, with a few pointed sandy projections and coralline purple spots. If it proves to be a distinct species I would eall it C. fusea. Mere FimpriAta Ald. & Hance. Trans. Zool. Soc. vol. 111. pp. 1387— 139 (1864). A large number of specimens of this remarkable animal were captured on both the east and west sides of Zanzibar in 1901. Alder and Hancocek’s figure and description give a good idea of its external appearance; but the coloration is very variable, ranging from clear bright yellow to ashy grey. Sometimes the colour i is uniform, but more often the sur face of the body and of the papillee is mar rked with irregularly disposed spots and blotches, which may be black, white, grey, or sandy. These markings harmonize with the ordinary environment of the creature, and ‘ause it to closely resemble a piece of seaweed besprinkled with sand and partially encrusted with sponges and other animal growths. In full-grown and perfect specimens, which are six inches long or more, the number of papillae seems to be six or seven on each side of the body ; but they are very easily detached, and few individuals have the two series complete. T also found Alder and Hancock's description of the internal [8] 1902. | NUDIBRANCHS FROM ZANZIBAR. 69 anatomy to be correct, particularly as regards the absence of jaws. They say: “In Melibe the buccal organ is provided with neither tongue, jaws, nor collar.” Bergh, in his monograph on the genus (in Malac. Untersuch. in Semper’s Reisen, Th. i. Bd. i. p. 363), thinks this statement will probably prove incorrect as other species of Melibe are provided with jaws, and he gives as a generic character: “ Bulbus pharyngeus cum mandibulis aliquantulum ut in Phylliroidis: margo masticatorius mandibule fortiter dentatus.” Text-fig. 5, poy My Lycos pus Ui iditkgg i Y, LY If Yh Wig , yy Uf yy \j (( ) y shy e az Melibe fimbriata. a., anus. pyl.st., pyloric portion of stomach. div., diverticulum. st., thin-walled stomach. f.ce., folds of esophagus. st.pl., belt of stomach-plates. h., hood. v., ventricle of heart. U., lip. Mr. Crossland and I have, however, dissected several specimens of Melibe fimbriata, and in all failed to detect any trace of jaws. Our drawing (text-fig. 5) will perhaps explain clearly the structure ot the digestive tract. In the centre and bottom of the hood is a protruding, circular, fleshy lip. This leads straight into the [Y] 70 SIR CHARLES ELIOT ON [May 6, cesophagus, which is provided with a series of folds, but no hard armature of any kind. The stomach is set with a belt of horny plates of two sizes and usually alternating regularly, the small being next to the large. The pyloric portion of the stomach below this belt is provided with muscular ridges, and passes almost imperceptibly into the intestine. At the point where it begins to be constricted is a pouch-like diverticulum with a laminated interior. With the exception of the absence of jaws, the other characters of this animal clearly connect it with Melibe, not Tethys. The foot is very narrow, the body rather high and compressed; the cerata are covered with knots; the buccal opening passes straight into the esophagus; the stomach is armed with plates; the liver is long and follicular and does not extend far into the cerata; the hermaphrodite gland is composed of many separate lobes at the side and under the liver. In Zethys, on the contrary, the foot is broad and the body flat; the cerata are smooth ; there is a division of the alimentary canal before the esophagus which may be called a buceal cavity; the liver is a compact mass sending diverticula to the ends of the cerata ; the hermaphrodite gland forms a thick covering over the liver. Fur ther, Tethys is described as possessing true branchie set at the base of the cerata. The back of Melibe jfimbriata is covered with branched papillze which bear a superficial resemblance to gills, but I could not discover that they have any special connection with the vascular system, and they seem ana- logous to the ramose appendages of Plocamophorus and some species of Votarchus. It would thus appear that Melibe fimbriata is intermediate between Zethys and the jaw-bearing species of Melibe. It does not, however, seem necessary to create a new genus, but rather to modify the existing description of the genus and say jaws present or absent. The shape of the foot, body, and cerata, the presence of stomach-plates, the absence of branchiz, and the character of the liver distinguish it sufficiently from Tethys. In spite of its want of j jaws, Melibe fimbriata is a most voracious animal, and I more than once found in the stomachs which I examined limbs of crustacea more than an inch long. The way in which it captures its prey is extremely curious. The circular oral veil acts as a net with an elastic rim. When seeking for food it expands the net and sweeps with it the surface over which it is crawling. The skin of the hood is stretched so tight as to be quite transparent and the marginal cirri are almost invisible. The moment a small crustacean or other prey is caught the net closes up, the cirri almost unite on the under surface, and the skin ceases to be perfectly transparent. Then the MMelibe tosses the hood, which has now practically become a closed sac, backwards, and creates a current of water with the cirri, which forces its prey towards its mouth. The movements of the animal are rapid and energetic, whether it crawls or swims. It can also float on the surface foot uppermost. [10] 1902.] NUDIBRANCHS FROM ZANZIBAR. @ MADRELLA FERRUGINOSA. (Plate VI. figs. 14-16.) Madrella ferruginea Ald, & Hance. Trans, Zool, Soe, iii. pp. 141-2 (1864), No fresh details have, I believe, been published respecting this genus since Alder and Hancock’s description. I have seen two specimens at Zanzibar, one about half an inch long and the other nearly double the size. ‘The colour of the body is a deep coppery red. Round the edge of the mantle, including the anterior margin, are transparent copper-coloured cerata, into each of which passes a very short diverticulum of the liver. The black or deep purple ramifications of the liver are visible through the dorsal integuments. There are many more cerata in the large than in the smaller specimen, and it is therefore possible that they increase with age. The middle of the dorsal area is bare, except that it carries several irregularly distributed tubercles or papillee. In the large specimen they pass between the rhinophores and form a sort of rudimentary crest, but in the smaller specimen, though they occur on the back, they do not pass between the rhinophovres. The large specimen had a white blotch between the rhinophores, the smaller none. The form of the rhinophores is somewhat unusual, They are not perfoliate, but there is a circle of papille round the top of the club, somewhat as in 7ritonia. There are no anterior tentacles, but the head is very broad and crescent- shaped, with produced ends. The front of the foot is wide and square, but the corners are not prominent. The mouth is ventral. Both the mantle-edge and the foot are wide, but between them is a deep groove. In crawling the foot projects beyond the mantle. The mantle overhangs the head and forms a wide frontal veil. The genital orifices are in the anterior part of the right-hand side, the anus in the posterior part, distinctly lateral and not dorsal. The internal anatomy, so far as I could examine it, agreed with the description of Alder and Hancock. I was unable to see any salivary glands. The jaws are very large, enclosing the buccal mass, but not denticulate. The radula (Pl. VI. fig. 16) is triseriate and long. The median tooth has a strong blunt spine in the centre and about 7 denticulations on each side. The laterals have a large spine on the outer margin and 8 or 9 denticulations on the inner side. These denticulations seem therefore somewhat less numerous than those described by A. & H. The ganglia in the central nervous system are very distinct. Madrella appears to be sluggish in its movements. In confinement it discharged some fluid, which imparted a ferruginous colour to the water round it. This discharge did not appear to proceed from any particular organ, but from the whole surface of the body. The genus forms an interesting connecting link between the Janide and other Alolide. The arrangement of the liver and cerata connect it decisively with the former, but in its lateral anus and triseriate radula it approaches the general characters of the group and departs from the exceptional peculiarities of the Janide, [11] 72 ON NUDIBRANCHS FROM ZANZIBAR. [May 6. which have a dorsal anus and a multiseriate radula. As a family the Janide are characterized by the presence of hepatic diverticula and of cerata in the anterior portion of the dorsal surface, in front of the rhinophores. Another character peculiar to the family, but not universal in it, is the crest between the rhino- phores. There are four well-marked genera :—Jadrella, with lateral anus, triseriate radula, papillous rhinophores, a rudimen- tary crest, and jaws without teeth. The other three genera have the additional peculiarities of a dorsal anus and a multiseriate radula. Proctonotus has simple rhinophores, no crest, and jaws without denticles. Janolus has perfoliate mandibles and very large jaws without denticles ; the foot is also exceptionally broad. Janus has perfoliate rhinophores, a toothed mandible, and a crest. EXPLANATION OF THE PLATES. Puate V. Figs. 1, 2. Crosslandia viridis, p. 64. 3. a » _-var.(?) fusca. 4. - » outer teeth of two consecutive rows of the radula. 5. - of rhinophore cup, one side removed. 6. ms 55 central and inner teeth of radula. ie 5, jaws, from in front. 8. a 3) jaws, from the side. PratE VI. Fig. 9. Zatteria browni, p. 62. 10. * c aie ammatic view showing position of cerata, wale 3 3 one of the cerata. 12. i » tooth of radula. 13. e view of head from below. 14, 15. Madrelia ferruginosa, p. 71, dorsal and ventral view. 16. 3 ss one ew of radula—a, median; 4, 5, lateral teet [12] mie oe Loe, vol. lela. 2 & Danielsson LY lith. NUDIBRANCHS FROM ZANZIBAR. ———— ——— a6 Ce ne nati ~~ oa - ee : dm Al oaae ; * janes - ae Ley Ngee ‘ me 5 ihe Ms xem, oe Commissioner for the East Africa Sp E.Z8. [From the PROCEEDINGS or tHE Zoonoarcat Socrery or Lowvow, : 1908, vol. 1.] pate [Published August 6, 1903.] if ©- So J 3 . hi wu ‘ Meee (nei § yas ad a - <= - =] { é =) % be if ae i 4 ae he a ' j - ; ai ‘ 2 y, 250 SIR CHARLES ELIOT ON NUDIBRANCHS [Mar. 17, [From the ProcenpinGs OF THE ZoonoaicaL Socrpry or Lonpon, 1908, vol. i. | ' [Published August 6, 1903. On some Nudibranchs from Hast Africa and Zanzibar. Part IJ. By Sir C. Euior, K.C.M.G., H.M. Commis- sioner for the East Africa Protectorate, F.Z.8. CERATOPHYLLIDIA AFRICANA, gen. et sp. nov. One specimen from near Wasin, EK. Africa, in 9 fathoms. The living animal was described by Mr. Crossland, who dredged it, as of a light greenish-yellow colour on the upper surface, but with the foot, branchie, and under side of mantle, white. The back was very hard and smooth, but its most remarkable character- istic was the presence of a number of papille, consisting of round or pear-shaped bodies set on stalks. The stalks as well as the base and tip of these globes were white, but the middle part was black, owing toa dense aggregation of black spots, which, however, can be seen to be separate under a lens. The globes were quite soft and the stalks flexible; they shook when the animal was moved, but were not observed to execute any spontaneous move- ments. The mantle-edge was wavy. The alcoholic specimen is of a uniform pale lemon-yellow, the black bands of the globes being, however, preserved. The breadth across the middle of the back is 1:9 centim. Unfortunately the animal is contracted almost into a circle, but apparently the length, when stretched out, must have been about 2°2 centim. The consistency of the body is like hard wax, and fragments of the mantle, which is ample, could easily be detached with the forceps. The whole dorsal surface is a mass of closely packed spicules. It bears about a hundred of the stalked globes. They ‘p of very varying size; many are quite minute, but the largest 1 For Part I. see P. Z.S. 1902, vol. i. p. 62. [2] 1903. ] FROM EAST AFRICA AND ZANZIBAR. 251 rhinophore-pockets and the anal papilla project; the edges are smooth. In the preserved specimen the rhinophores are grey. The branchiz are arranged in a circuit interrupted only by the head and genital papilla. They vary in size, but though in places long and short branchiz seem to alternate, this cannot be said to be the general rule. Over the mouth are two tentacles each about 1:5 millim. long, and 1 millim. broad at the base. They are not directed sideways but straight forward, and being set close together so that the division is not visible, they appear to form a sort of head. They are united at their bases. The mouth is larger than is usual in this order, and though it is suctorial is hardly poriform, Though the animal was dissected only three months after capture, the internal organs were already much dried and shrivelled, the spirit having apparently been unable to penetrate the hard integument. It was clear, however, that the buccal organs are of the type of Phyllidiopsis rather than Phyllidia. The buccal opening led into a sausage-shaped tube about 6 millim. long and 2 millim. broad, with muscular walls transversely striped. This passed into a long, narrow, coiled tube, which preserved the same calibre until it dilated into the stomach, ‘Two ample glands (salivary ?) entered the larger part of the tube on either side, but were not in any way fused with it. The liver was large and undivided behind. The central nervous system was enclosed in a thin capsule and somewhat concentrated, the cerebral and pleural ganglia being hardly distinguishable and the pedal ganglia lying beneath them. The eyes were large, black, and distinct. The genital mass was much hardened, but the two spermatothecz, one white and empty and the other black and full, were quite distinct. It was impossible to ascertain whether the glans was armed with hooks and whether the folds on the dorsal wall of the pericardium (sometimes called the pericardial gill) were present, but it is highly probable that the species possesses these family characteristics. In virtue of its buccal apparatus this animal belongs to Bergh’s genus Phyllidiopsis, although the tentacles are not attached through their whole length and are rather larger than is usual in the Phyllidiade. It is remarkable that the genus Phyllidiopsis contains one species, Ph. papilligera, which has also black papillee on the back. To me, the presence of these dorsal papille seems a peculiarity sufficiently marked to merit generic rank. If Hehino- doris is a genus, why should not Phyllidiadz which have the same peculiarity enjoy the same distinction? I would propose to call the genus Ceratophyllidia, and its characters will be :—Back studded with papille ; buccal apparatus in the known species similar to that of Phyllidiopsis. PLEUROPHYLLIDIELLA HORATII, gen. et sp. nov. One specimen from Wasin, East Africa. Mr. Crossland, who captured it, gives the following notes on the living animal :— “ Three inches long. Mantle edged with light salmon- colour ; it [3] 252 SIR CHARLES ELIOT ON NUDIBRANCHS [ Mar. 17, ends in two ridges near rhinophores, similarly edged; front edge of velum similarly edged. Back grey, mottled with a darker shade, the top of the numerous longitudinal ridges being sprinkled with clear black spots. Underside white. Gills also white. Rhinophores longitudinally lamellated, grey in colour, and can be contracted or retracted, though the pockets do not seem very definite or complete.” The somewhat bent alcoholic specimen measures 3 centim. from head to tail, but would be at least ‘5 longer if it were straightened out. The breadth of the back is 1:5 centim., of the foot ‘6. The foot is long and narrow, pointed behind and truncate in front, the corners not projecting. No caruncle or nuchal papille are visible, but it is somewhat difficult to reconcile the head-parts of the preserved specimen with the description quoted above. According to a rough drawing made from the living animal, it would seem that the mantle-edge passes between the rhinophores and forms a sort of velum in front of them; but in the preserved specimen it appears to he behind them as in an ordinary Pleurophyllidia, and not to pass through them at all. The salient character of the genus is that there are no branchie and no trace of a branchial cleft. Lateral lamelle are, however, present. They are situated exclusively on the under edge of the mantle, and not on the sides of the body. They extend from the head to the tail, and are about 30 in number on each side. They are irregular in size and shape. Some terminate before they reach the mantle-edge, and some run from the mantle-edge only halfway to the body. The genital papilla is 4 millim. and the vent 1°2 centim. from the anterior end of the body. Cnidopfores are distinctly visible round the edge of the mantle. The mouth is ventral, and forms a large transverse slit, with slight indications of a T-shape. The jaws are yellow, narrow, and united so as to form a shape like a boat. They bear no denticles, but there are a few irregular coarse indentations of the edge, due apparently to its being jagged by use. The radula consists of about 30 rows, the formula for each of which is about 50+1-+ 50. The central tooth consists of a squarish basal plate with a long cusp, which bears about 10 denticles on each of its sides, The first lateral is larger than the others and resembles the central tooth, except that the denticles are only on the internal side. The second and third laterals are also denticulate; the rest appear to be simply hamate. This form appears to be intermediate between Pleuroleura, which has neither branchial clefts nor side lamelle, and Pleuro- phyllidia, which has both. I have indicated its affinities by the name Pleurophyllidiella. BOLIDIA MAJOR, 0. Sp. (Bergh, in Semper’s Reisen im Arch. Philipp., Malac. Untersuch. vol. iil. pp. 778-780, 1880.) [4] 1903. | FROM EAST AFRICA AND ZANZIBAR. 253 One specimen from Chuaka, HK. coast of Zanzibar, under a stone between tides. The living animal was about 4 centimetres long. The body and appendages were of a uniform greyish white, with spots of a dull opaque white. The whole animal closely resembled a kind of detachable sea-anemone which is very common at Chuaka, and appears to be sometimes almost free-swimming. The alcoholic specimen is 3 centim. long and 1 centim. broad at the widest part, including the cerata. The foot is moderately broad, and has fairly long tentacular expansions in front; but its most remarkable character is the size and distinctness of the anterior groove, which measures 2 millim. across. The upper lip is separated into two parts bya deep cleft. The oral tentacles are large and very thick. The rhinophores are shorter and studded with minute knobs, which, in the preserved specimen at any rate, appear not to be set in rings. The cerata are much flattened and almost leaf-like, and the hepatic diverticula within them are ramified. They begin at the anterior end of the large peri- cardial prominence, and are arranged in about 17 groups on each side, each containing about 10 cerata. There are very distinct gaps between the anterior groups, and a broad bare space runs down the middle of the back, but towards the end of the body the cerata are huddled together and continue until the extreme tip, there being no tail. The outermost cerata of all the rows are smaller, and the inner considerably larger, but at the base of the innermost are frequently quite small, some hardly larger than tubercles. The genital orifice is below the first group of cerata, and the lateral vent behind the second. The specimen was only partially dissected. The jaws are very large, colourless and transparent, with a perfectly smooth edge. The radula consists of 32 pectinate teeth, very similar to those of B. mebii (see Bergh, J. c. pl. lxxix. fig. 16), with striations under each denticle. They are, however, very much broader, the widest measuring 2 millim., and the denticles are more irregular in shape, being probably worn by use. There are about 150 of them on the broader teeth. The three or four central denticles are generally, but not always, smaller than the others. This specimen is clearly referable to Bergh’s genus Beolidia, and the difference between it and the type is mainly one of size, B. mebii being only 8 millim. long. The similarity of habitat makes one think that this may be merely a full-grown individual of the same species ; and we know so little of the variations which the radula and arrangement of cerata may present in Afolids at the different periods of their growth, that Iam not prepared to reject this hypothesis. Still, the single specimen examined by Bergh appears to have been sexually mature, aud this being so the two animals each present peculiarities amounting to specific differences :—(1) In B. mebii the tentacles are said to be “ tibge- plattet fingerformig” ; the cerata begin behind the rhinophores and are set in-rows: in 4. major the tentacles are stout and round ; [5] 254 SIR CHARLES ELIOT ON NUDIBRANCHS [Mar. 17, the cerata begin further back and are set in groups. (2) In B. mebii the reproductive orifice is under the third row of cerata, and the vent between the sixth and seventh rows, almost dorsal : in B&B. major the reproductive orifice is under the first group of cerata, and the lateral vent after the second. (3) In B. major the basal part of the teeth is proportionally narrower than in Z. meebii, and the denticles are more irregular. Var. ORNATA. One specimen captured at Chuaka, May 1902, seemed to be a typical Bwolidia major, except for a somewhat more ornate coloration. The ground-colour was of a yellowish-white with a yellowish-brown pattern, consisting of a series of irregularly shaped lozenges containing white spots, down the middle of the back as in B. major. The oral tentacles were white with green stripes. The cerata were white with yellow tips, below which was a bright blue band. CERBERILLA AFRICANA, 0. Sp. (Bergh, Journ. Mus. Godef. Heft i, 1873, . Heft vi. 1874; Beitr. zur Kenntniss der Aeolidiaden, Theil 11.; Semper’s Reisen im Arch. Philipp., Malac. Untersuch. 111. p. 879.) One specimen from the reef Jembiani, Zanzibar, 3°5 centim. long and 1°3 broad in life. The back was almost entirely covered with cerata, so that the body-colour was hardly visible. Most of the cerata were very dark green with a bright yellow ring, but the innermost were white with bluish tips, with only a ving of dark green. Theoral tentacles were dark blue, with green bases. The rhinophores had four bands of colour, which were, starting from the base, greenish brown, white, blue, white. On the head were two yellow lines, extending from the oral tentacles to the rhino- phores, and the margins of the head and foot were also of a bright light yellow. The animal was stoutly built. The foot projected considerably beyond the body on either side. Its anterior angles were expanded into long tentacular processes. The head had also two lateral expansions, from which projected at right angles the very large and conspicuous oral tentacles. In life the. rhinophores were quite simple and fairly long. In the alcoholic specimen they were contracted and somewhat wrinkled, The numerous and thick-set cerata were disposed on peduncles. There was a bare triangular patch behind the rhinophore, and a narrow bare space down the middle of the back, but the cerata folded over the latter so that neither it nor the transverse bare areas were visible. There were about twenty transverse rows of cerata. The first row at the side of the rhinophores consisted of about 10 cerata, much smaller than the rest. After the third row was a distinct gap, and a smaller gap after the fourth. After that the rows were so close together that they could not be distinguished superficially. The innermost cerata were larger than the. others, and sometimes bifid: small cerata extended almost to the end of the very short tail. [9] 1903. ] FROM EAST AFRICA AND ZANZIBAR. 255 The jaws were large, with a smooth cutting-edge. The radula consisted of a single series of 13 transparent yellowish teeth of the shape usual in the genus, viz., pectiniform, with large irregular denticles and small accessory denticles. It was sometimes hard to decide whether the rather small denticles should be considered main or accessory; but the average number of main denticles on each tooth was 10, and the largest number (in one case only) 13. The central nervous system was somewhat concentrated. The specimen was only partly dissected. A new species must, I think, be provisionally created for this animal, though the discovery of intermediate forms may perhaps render {ts retention unnecessary. In some ways it is itself a con- necting-link between C. longicirrha and C. annulata, for the former is said to have 7, and the latter 5 denticles on each side of its teeth, whereas C. africana has 4, 5, 6, or 7 indifferently. It can hardly be C. annulata, for the difference in colour is too great, and besides there is much less bare space on the back. Neither can it be C. longicirrha, because (1) the coloration, though similar, is still distinct ; (2) C. longicirrha has the back bare up to the 7th row of cerata, and some of the cerata are very long, which is not the case here ; (3) the rhinophores are not perfoliate. This last point is of some importance for the characterisation of the genus. In the present animal the rhinophores were undoubtedly quite simple in life, and in alcohol they are wrinkled, though it is still possible to see that they are not really perfoliate. In C. longicirrha, Bergh says the perfoliations are 14 or 15 in number, and not deep. Of C. annulata he says that the rhino- phores have 12-14 well-marked perfoliations, and that Garrett has wrongly represented them as simple. But in Semper’s ‘ Reisen,’ xvii. he states that C. annulata var. affinis has simple rhinophores, and gives as a generic character: “Die Rhinophorien scheinen nicht perfoliirt zu sein.” I have not access to part ix. of his ‘ Beitriige zur Kenntniss der Aecolidiaden,’ which perhaps explains the matter ; but it looks as if the rhinophores are simple, but have a tendency to simulate perfoliations when preserved. PTERAEOLIDIA SEMPERI. (Bergh, Beitr. zur Kennt. der Aeolidiaden, iii. p. 22, and in Semper’s Reisen, Malac. Untersuch. vol. i. p, 18 (1870); under Llabellina.) Four specimens, which seem probably referable to this species, were dredged from 38 fathoms near Chuaka in July 1901. The body is very long, narrow, and vermiform, the largest individual being 5:5 centim. long and only 3 millim. broad. The ground-colour of the body in the living animal is brown, with opaque markings of very light green on the sides and back. The cerata are also dark brown, with numerous thin lines of the same green. The top of the head and the ends of the oral tentacles are opaque yellowish white. The lower part of the tentacles brown, with three rings of Proc. Zoot, Soc.—1903, Vou. I. No. XVII. 17 [7] 256 SIR CHARLES ELIOT ON NUDIBRANCHS [Mar. 17, vivid crimson-lake. The body of the rhinophores is brown and inconspicuous, but the tips are whitish and have a crimson ring. Although the colour of the animal when analysed is as described, the general effect in most lights is that it is purple with a silvery glaze. The specimens were active in their movements, and in particular the long oral tentacles were very mobile. The anterior margin of the foot was produced into two deeply-grooved processes, and its sides projected in two lamine along the whole length of the body. The cerata, which were not at all caducous, were set on fan-like projections of the dorsal margin, from 13 to 20 in number, on each of which were from 10 to 20 cerata. The largest individual appeared to have about 640 cerata in all. None of the cerata were large, but they became more crowded and smaller towards the end of the body, where they extended to the tip of the tail. In three of the preserved specimens there was a very distinct raised border on each side of the back, connecting the fan-like projections, and almost resembling the mantle-edge of Dorids; but in the fourth this remarkable character was not visible. The oral tentacles were very large and long; the rhinophores, whicgh stand between the first clumps of cerata, were small and lamellated, except the tips, which were narrow and smooth. In the individual dissected the radula consisted of a single series of 18 yellowish teeth. The separate teeth were much as represented in Bergh’s plate (J. ¢. pl. ui. figs. 5, 6, 7), but somewhat more regular in shape, having nine denticles very symmetrically arranged on each side of a central cusp, but not taking the form of indentations of the cusp itself. The cutting-edge of the jaw bore a row of fine but very distinct denticles, at the side of which were several rows of less distinct accessory denticles. No spine or other armature was discovered in the reproductive system, These specimens seem to clearly belong to Bergh’s genus Péeraeo- lidia. There are some small differences (such as the shape of the rhinophores and of the teeth and the lateral ridges) between them and his description and plates of Pteraeolidia (Flabellina) semperi, but the ridges were absent in one specimen and the other characters were slight. It is possible, however, that a comparison of the living animals might show a specific difference. ERCOLANIA ZANZIBARICA, N. Sp. (For the genus Hrcolania see Trinchese, ‘“ Un nuovo genere della fam. degli Holididei,” Ann. del Mus. di Stor. Nat. di Genova, ii. 1872; id. ‘* Aeolide del Porto di Genova”; Bergh, Beitr. z. Kennt. der Aeolidiaden, v. pp. 9-18, Wien, 1878; Vayssiére, Recherches sur Moll. Opisthobranches, Marseilles, 1888, pp. 121-128.) Two specimens captured at Chuaka, East Coast of Zanzibar, in February 1901. The living animal was rather more than 2 centimetres long, very fragile and delicate, transparent and of a uniform bright green, the hepatic diverticula in the cerata not [3] 1903. ] FROM EAST AFRICA AND ZANZIBAR. 257 having been distinguished by any difference or intensity of colour. It exactly resembled a piece of the racemose seaweed (Caulerpa) on which it was found. The length of the alcoholic specimens is 2 centim.; the extreme breadth of the back with cerata 8 millim., and of the foot 2°3 millim. The rhinophores are long and distinctly canaliculate. There are no oral tentacles, but two lobes over the mouth. Behind the rhinophores are two very distinct black eyes. The cerata are club-shaped as in G'alvina, of varying size, the largest inside. On each side of the back are four clumps of about nine cerata each, and there is a thick bunch on the tail, which, however, projects a considerable distance behind the last cerata. Down the centre of the back is a broad bare space, in the anterior portion of which is the very large, elongated (not oval) pericardial prominence. In front of this and fused with it is the vent, a large and conspicuous tube. The foot is rounded in front. I dissected one specimen, but was unable to obtain a clear view of either the central nervous system or the reproductive organs. The latter, as usual in this family, were extremely complicated, both the prostate and albumen-gland appearing to be extensively ramified. The verge was armed with a small spine. The hepatic diverticula in the cerata, being colourless, were not easy to dis- tinguish, but appear not to be much ramified and to resemble the figure of those in Hrcolaiia siottii in Trinchese, pl. ix. fig. 2. The mouth-parts, buccal muscles, radula, &c. are of the usual ascoglossan type. The teeth are not unlike those of Hrcolania viridis (v. Bergh, J. ¢. pl. xii. figs. 3 & 4), but the dorsal surface is a simple curve and does not show any depression. The upper portion of the radula contains 6 teeth, the lower 27, the last members being arranged ina spiral like that depicted in Trinchese’s plate of #. siotti (pl. x A. fig. 1), from which it may be concluded that the individual is young. As the specimen presents all the characters of the genus Hrco- lania, I describe it under that name, but I feel very doubtful if the genus is valid. The only characters which differentiate it from Stiliger, viz. that the rhinophores are canaliculate and the pericardial prominence elongate and not oval, are surely very slight. Vayssiére (/. c. p. 122) referred to the genus a species (funerea) with entire rhinophores, which is probably in any case a Stiliger. The animal is not likely to be specifically identical with E. viridis Bergh, for the coloration is not really the same, the size is much larger, and the shape of the teeth somewhat different. 17* [9] PO tOniy feed. 1 eae Cn. é ; ON SOME NUDIBRANCHS FROM EAST AFRICA AND ZANZIBAR. “Part Ill. py smo ELIOT, PR OMG.) EM: Commigignes for) the “Bast Africa Protectorate, F.Z.S.—Doridide Cryptobranchiate, I. ce Ma [From the PROCEEDINGS oF THE ZOOLOGICAL Socrery oF Lonpon, 1903, vol. ii. [Published April 1, 1904.] 3o4 SIR C, ELIOT ON NUDIBRANCHS [Dec. 1, [From the PROCEEDINGS OF THE ZooLoGicaL Socrery oF Lonpon, 1903, vol. ii.] [Published April 1, 1904.} On some Nudibranchs from East Africa and Zanzibar. Fart (11.* By Sir C. Hor, K.C.MG.) EM Com- missioner for the East Africa Protectorate, F.Z.8.— Dorididze Cryptobranchiate, I. (Plates XXXII.-XXXIV.t) This paper contains the results of an examination of a number of Cryptobranchiate Dorids from Zanzibar and the East Coast of Africa referable to Bergh’s subfamilies Archidoridide, Disco- dorididze, Diaululide, Kentrodoridide, and Platydoridide. They include the following species :— Archidoris Bergh. ] _ africana, sp. n. 2. 45 minor, Sp. N. Staurodoris Bergh. 3. i depress, sp. Ni. A. BS calva, sp. n. Discodoris Bergh. 5. at boholiensis Bergh. 6. 6 cerulescens Bergh, variegata, subsp. n. Peltodoris Bergh. is Bs angulata, sp. n. 8. ms aurea, sp. n. Thordisa Bergh. 9. + villosa (A. & H.). 10. stellata, sp. n. Il. 5 crosslandi, sp. n. * For Part II. see P. Z. S. 1903, vol. 1. p. 250. + For explanation of the Plates, see p. 385. 1903.] FROM EAST AFRICA AND ZANZIBAR. 355 Trippa Bergh. 12: » areolata (A. & H.). 13. 5» monsont, sp. n. Fracassa Bergh. 14, 5 tuberculosa. Halgerda Bergh. 15. » willeyt, sp.n. (From the Loyalty Islands.) 16. By wasinensis, Sp. D. Kentrodoris Bergh. 17. BA rubescens Bergh. Platydoris Bergh. 18. : eurychlamys Bergh. 19. +5 scabra (Cuvier). 20. - Jormosa ? (A. & H.). 21. ellioti? (A. & HL). 22. 5 pulchra, sp. n. 23. er incerta, sp. Nn. 24. papillata, sp. n. Sclerodoris Eliot, gen. n. 25. * osseosa (Kelaart). 26. 5 tuberculata, sp. n. 27. 7 minor, sp. Nn. 28. 4 rubra, sp. 1. AAS) =n coriaced, sp. n. Asteronotus Ehrenberg. 30. 5 hemprichi Ehrenberg. The Cryptobranchiate Dorids are distinguished by having branchiz which are completely retractile into a permanent pocket. The rhinophores also are retractile into permanent pockets and are perfoliate. A radula is always present. In number of species they form one of the richest groups of the Nudibranchiata, but they show less variety of form than the Doridide Phanero- branchiatz, and it is not easy to find good generic characteristics, The most distinctly marked group is that consisting of the large genus Chromodoris (with more than 100 species) and some allied genera such as Casella and Ceratosoma. ‘These possess a charac- teristic shape, lip-plates,a radula with differentiated teeth, and simply pinnate branchie. Well characterised genera are also found in the Miamirade (MZiamira, Orodoris, Spherodoris) and the Cadlinide (Cadlina, Tyrinna). There remain the five sub- families mentioned above, which agree in being usually flat and oblong in form and in having a radula without a central tooth, consisting of numerous laterals which are generally hamate and uniform, though often smaller near the rhachis and degraded or denticulate at the outer end of each row. Anyone who describes new forms of these Dorids, however much he may disclaim any such ambitious task as a revision of 23% [3] 356 SIR C. ELIOT ON NUDIBRANCHS [ Dec. 1, the group, is bound to consider the value of the existing generic distinctions. The five subfamilies are divided into about 30 genera all created by the great master of the Nudibranchiata, Prof. Bergh, and several of them consisting of a single species. He himself observes (on Zhordisa in Semper’s ‘ Reisen’*, xv. p- 666) that the definitions of many of these genera seem to be too precise, and as new forms are discovered the old divisions are found to melt away. The discoverer of a new form often finds that it does not accurately fit in to any of the existing genera, and must ask himself whether he should create a new genus or enlarge the definition. I have little doubt that in most cases the latter is the preferable course. If animals are not divided by natural differences, there is no object in emphasising the im- portance of minute peculiarities. If Chromodoris is allowed to consist of 100 species showing a considerable range of variation, including the presence or absence of median teeth, there seems to be a want of proportion in splitting up the other Cryptobranchiata into so many genera. The genus Staurodoris offers a good example of the difficulty of classifying new forms. Taken by itself, the typical species St. verrucosa is remarkably well characterised in both the scien- tific and popular sense. Anyone could recognise it at first sight. The back is covered with club-like tubercles, the rhinophorial and branchial pockets are protected by similar tubercles which act as valves, and the branchie are simply pinnate. On the other hand, St. pseudoverrucosa (von Ihering) has no tubercles on the branchial pocket and has bipinnate branchiz. Bergh also refers to this genus the Doris pustulosa of Abraham, which has bipinnate branchiz and small, but apparently not valve-like, tubercles. I have specimens from the Indian Ocean which have the dorsal surface tuberculate and the following additional characteristics :— A. has the rhinophores arising among tubercles and simply pinnate branchie, but no tubercles on the branchial pocket (Staurodoris pecten). B, has five pinnate gills, the anterior margin of the foot entire, and small tubercles set on the edges of the gill-pocket and partly closing it (Stawrodoris calva). C. has tubercles round the rhinophores, none on the branchial pocket, and thin bipinnate branchie (Archidoris africana). D. has tubercles on the rhino- phore pockets but not around the branchial opening, and tri- pinnate gills (Archidoris minor). Of these, I think we must admit A. and B, to be Staurodoris, if we accept St. pseudoverrucosa, But Archidoris kerguelensis, A. australis, A. rubescens, A. incerta, and A, nyctea are all described hy Bergh as having tubercles on the edge of the branchial and rhinophorial pockets, and must come very near the less typical members of Stawrodoris. It is hard therefore to say whether ©, should be classed as Stawrodoris or Archidoris, There seems to be a complete series of links between the two genera, and, this being so, we must either unite * The letters S. R. in this paper refer to Prof. Bergh’s “ Malacologische Unter- suchungen,” published in ‘ Reisen im Aychipel der Philippinen,’ von Dr. C. Semper. {4 1903. ] FROM EAST AFRICA AND ZANZIBAR. 357 the species in question, or draw an arbitrary dividing line. In the latter case, I think we must say that the typical Stawrodoris has simply pinnate branchie and valve-like tubercles closing the rhinophorial and branchial pockets. One or other of these features may be absent or obscure in a species which otherwise possesses the generic characteristics ; but when both are absent, as in C., the animal, I think, must be referred to Archidoris. Again, Staurodoris pseudoverrucosa has the dorsal tubercles some- times connected by ridges, and the same phenomenon is found in Garstang’s Doris maculata (which appears to be a Staurodoris), and, sporadically, in the true S¢. verrucosa. But, as will be seen from a species described below, Halgerda wasinensis, this character brings Stawurodoris very near to Halgerda. It will be well to examine the value of the chief points by which the subfamilies and genera under consideration can be differentiated. They are as follows :— (a) The dorsal surface and general texture. The back is rarely quite smooth, as it is in Chromodoris; Halgerda and Asteronotus have the skin smooth but raised into ridges or lumps. As a rule, the surface is covered with projections which may be either minute granulations (Platydoris, Discodoris, &ec.), papille (Z'hor- disa, &c.), tubercles (Archidoris, &e.), clavate tubercles or warts (Staurodoris), compound tubercles (Zrippa, Fracassa). There is sometimes a ridge down the centre of the back. Some genera, notably Platydoris, are exceedingly hard ; others, such as 7'rippa, are so soft as to be almost gelatinous. On the whole, these external characters of the skin and texture form a fairly good indication of relationship. Platydoris, in which I should be disposed to include Hoplodoris, forms a distinct natural group, and the warty or tuberculate forms (Archidoris, Staurodoris) also hang together *. (6) Rhinophores and branchie: Neither the rhinophores themselves, nor the pockets into which they are retractile, seem to offer good generic characters, though they may often serve to distinguish species. But even within a species there may be variety: Archidoris tuberculata has the rims of the rhinophore pocket sometimes smooth and sometimes tuberculate. The branchiz also are disappointing as a means of classification. For instance, it does not seem possible to unite Staurodoris with the other genera having simply pinnate branchie (Chromodoris, Casella, Ceratosoma, Spherodoris, Halla, Thorunna, Rostanga), and these simple branchie often show a tendency to divide at the tip and become, strictly speaking, bipimnate. Similarly, we cannot bring together bipinnate forms and oppose them to the tripinnate. Perhaps the distinction between ample and scanty branchiz will prove to be of generic importance. The branchial pocket is of various shapes: round, crenulate, or stellate. However, * JT find it hard to agree with Prof. Bergh’s criticism of his own family Archi- doridide that it will prove “ ganz unhaltbar und kinstlich.” With the exception of Peltodoris the other forms seem to hang well together. 5: [9] 358 SIR C. ELIOT ON NULIBRANCES [Decm, it does not seem possible to make any classification according to this shape, and to unite, for instance, all the stellate forms. In most of the Cryptobranchiata the tentacles appear to be digitate, but the Archidoridide have a tendency (not without exceptions) towards a flattened and furrowed form. Spherodoris has no tentacles, and the same is said of Halla and Hchinodoris. No part of preserved Nudibranchs is more liable to distortion than the tentacles, which may be variously contracted, flattened, or crumpled by the action of the preserving fluid or the pressure of the adjacent parts. (c) The foot appears to offer good characters ; as a rule, but not invariably, the forms which have a broad foot and narrow or moderate mantle- margin belong in other respects to groups different from those which have a narrow foot overhung on every side by a wide mantle-margin. The Archidoridide and Disco- dorididee have both, as a rule, broad feet, but the body of the former is plump, and of the latter flattened. The Platydoridide have flat bodies and narrow feet. In most forms the anterior margin of the foot is deeply grooved, and the upper lamina notched so as to form two flaps, which in the Kentrodoridide are very ample. Many (though not all) of the Archidoridide have the anterior margin simply grooved and otherwise entire. In a few genera(Zrippa, Halla *, Spherodoris) the divided upper lamina is attached to the sides of the head, but it does not seem possible to bring together the forms which present this peculiarity. (2) In the internal anatomy, the mouth-parts are perhaps the most important for classification. It is clear that considerable structural differences in other organs are generally accompanied by a difference in the radula. For instance, Acanthodoris and Lamellidoris, which bear a strong superficial resemblance to the Cryptobranchiata but have no permanent gill-pocket, have also a totally different radula, and most of the larger divisions of the Nudibranchiata have a characteristic arrangement of teeth. But it is hazardous to conclude from this that small differences of the radula have a generic value. One common variation from the ordinary hamate type of radula is the serrulation of the outer teeth. This may be present or absent in the same genus (e. g. Staurodoris verrucosa and St. bicolor), and in some species (Platy- doris argo and Halgerda formosa t) is only found in some of the rows of teeth. Asa general rule the innermost and outermost teeth are smaller: the latter often rudimentary or degraded. But it appears that these characters are only of specific, not of generic Importance. * [I venture to point out that the generic name Halla is preoccupied by a Poly- chaete (Lumbriconereid) worm named by A. Costa in 1844 (cf. Ann. Acad. d. Aspiranti Naturalisti Napoli, 11. p. 63 (1844).—C. CrossranD. + Ina specimen of Halgerda formosa from the Berlin Museum, given me by the courtesy of Prof. Martens, I found at the end of some but not all of the rows small rudimentary teeth, some but not all of which bore a few irregular serrulations. This is not quite the same arrangement as observed by Prof. Bergh. [6] 1903.] | FROM EAST AFRICA AND ZANZIBAR. 359 Another point of importance is the presence or absence of a labial armature, that is to say, of a ring or plates on the labial cuticle composed of a compact mass of minute hooks or rods. There is some reason for dividing the group into those which possess and those which do not possess this armature, but still I think that an absolute dichotomy of this kind presents difficulties. Firstly, a rudimentary labial armature is of some inconvenience in classification. Bergh has described such an instance in Platy- doris variegata; and the animal described below as T'hordisa crosslandi has two small areas on the labial cuticle which cannot be called plates, but consist of a loose collection of minute rods. In Platydoris pulchra the labial cuticle is strengthened with similar rods, distributed through its extent, but not collected into rings or plates. Secondly, though Discodoris B. is a well-defined genus, the same can hardly be said of the family Discodoridide, the genera of which mostly agree only in having a labial armature. Thus Fracassa and Carminodoris appear to be akin to Trippa, Hoplodoris to Platydoris, and Halla to Chromodoris. It would seem that the more primitive forms of the Cryptobranchiata are those which have a labial armature and some differentiation of the teeth, such as occurs in Chromodoris and Cadlina, and that the forms with no labial armature and uniform hamate teeth are more recent developments. If this be so, it is easy to under- stand that in many groups a few forms have survived in which the labial armature has persisted. Thus Fracassa is practically a Trippa which has preserved this character, and it appears to me that its analogies to 7rippa are greater than those to Discodoris. (e) Considerable use has been made by Prof. Bergh of the reproductive organs as a character for classification*, That great weight must be attached to important variations in these organs —such as the presence of one or two spermathecee—is obvious ; but while fully admitting the necessity of examining the internal anatomy and the futility of describing only the external characters of Nudibranchs, it must also be admitted that it is not con- venient to found genera of fairly large animals upon minute internal characteristics which can only be discovered by an expert microscopist, and by him only in a well-preserved and mature specimen. Such a criterion seems desirable only if it is of great anatomical importance. But what are the variations presented by these organs in the Cryptobranchiata? (i.) Accessory organs are sometimes present. These are generally accompanied by other characters of systematic importance as in Kentrodoris and Asteronotus. (ii.) The male branch of the hermaphrodite system sometimes broadens and sometimes does not into a dilatation called the prostate. It does not appear that this difference is generally accompanied by other characters of importance: e.g. a * But see System. d. Nud. Gast. p. 1088. “Bei dem jetzigen Stande unserer Kenntniss .... ist es sehr gewagt eine systematische Gruppirung derselben zu versuchen. Es ist es um so mehr als der generische Werth der bei der hier ver- suchten Gruppirung dem genital System, besonders den verschiedenen Conforma- tionen seiner Ausfiihrungsgange, beigelegt ist, als solcher kaum sicher gestellt ist.” Lv] 360 SIR C. ELIOT ON NUDIBRANCHS [ Dec. 1, prostate is present in both Discodoris and Platydoris, which are otherwise so distinct, and is absent in Archidoris but present in Anisodoris, which are otherwise identical. (iii.) In many genera is present an armature of the reproductive organs, that is, one or many spines, hooks, or plates generally only on the male branch. The value of this feature, as estimated by the con- comitance of other important characters, varies. In the families of Doridopside and Phyllidiade and in the well-marked genus Platydoris of the Cryptobranchiata an armature is, so far as we know, characteristic. On the other hand, we find two genera like Diaulula and Gargamella apparently identical but for its presence or absence. In the Aeolidiadz also a hook or spine is present or absent in closely allied genera. It appears to me that such an armature is not sufficient to constitute generic rank without other characters. (7) The other internal organs do not offer many features which serve our purpose of classification. It does not appear that we can unite the forms in which the stomach is enclosed in the liver, or in which there is only a single instead of a double blood-gland, or in which tke nervous system is very concentrated. As a result of these considerations, I think that the most pro- fitable way of classifying new Dorids of the group treated of here, is to refer them when possible to the following genera taken in a wide sense :—