! ^iisrCs' '# %^ >:J't^„:fC ^ .?^-s '•^: '"^'^!*b!^.. ■^ x-t ^)t.«^''#- ■V-. J ?*^ J' ^~' tr^ v»^-^t^- K^:-t ■■■^ v ^-f: ^^f^y^y --^ ■Fu- ^r '.> '■-' *-y*^ r*l^ ..^<^' >^ *^i iV^ r^ D m □ □ ^= DISCOVERY REPORTS VOLUME XVUI CAMBRIDGE UNIVERSITY PRESS LONDON- BENTLET HOUSE NEW rORK, TORONTO, BOMBAY CALCUTTA, MADRAS: MACMILLAN TOKYO: MARUZEN COMPANY LTD All rights reser-ved H^ ^r- DISCOVERY REPORTS ^^a.c^, Issued by the Discovery Committee Colonial Office, London on behalf of the Government of the Dependencies of the Falkland Islands VOLUME XVIIl CAMBRIDGE AT THE UNIVERSITY PRESS 1940 PRINTED IN GREAT BRITAIN BY WALTER LEWIS, M.A., AT THE UNIVERSITY PRESS, CAMBRIDGE CONTENTS COAST FISHES. PART III. THE ANTARCTIC ZONE (published 31st May, 1938) By J. R. Norman Introduction Systematic Part General Part pagez 4 87 Bibliography Plate I 102 following page 104 ON THE OPERATION OF LARGE PLANKTON NETS (published 25th August, 1938) By James W. S. Marr, M.A., B.Sc page 107 Plate II following page 120 CRINOIDEA (published nth November, 1938) By D. Dilwyn John, M.Sc. Introduction ftrifTO T 01 ERRATA DISCOVERY REPORTS, VOL. XVIII, 1939. "The Leopard Seal Hydmrga leptonyx (De Blainville)" p. 241, line 15 for no. 1901 read no. 1091 p. 24s, in the list of specimens under South Shetland, 1921 delete South Sandwich region Literature 238 THE LEOPARD SEAL HYDRURGA LEPTONYX (DE BLAINVILLE) (published 3rd October, 1939) By J. E. Hamilton, M.Sc. Introduction page 241 Historical 241 Distribution 242 53a^8 PRINTED IN GREAT BRITAIN BY WALTER LEWIS, M.A., AT THE UNIVERSITY PRESS, CAMBRIDGE 0 CONTENTS COAST FISHES. PART III. THE ANTARCTIC ZONE (published 31st May, 1938) By J. R. Norman Introduction Systematic Part General Part page 3 4 87 Bibliography Plate I 102 following page 104 ON THE OPERATION OF LARGE PLANKTON NETS (published 2Sth August, 1938) By James W. S. Marr, M.A., B.Sc page 107 Plate II following page 120 CR I NO IDEA (published nth November, 1938) By D. Dilwyn John, M.Sc. Introduction .... New Forms .... Care of the brood . Distribution and relationships Acknowledgements . Key to Antarctic species Systematic account Pentacrinoid larvae Antarctic Comatulids previously in the British References .... Plates III-VI .... Museum Collection page 123 125 125 128 131 131 133 199 220 221 following page 222 THORACIC CIRRIPEDES COLLECTED IN 1925-1936 (published 31st January, 1939) By C. A. Nilssen-Cantell, Sweden Introduction General Systematic Literature page 225 226 227 ■ 238 THE LEOPARD SEAL HYDRURGA LEPTONYX (DE BLAINVILLE) (published 3rd October, 1939) By J. E. Hamilton, M.Sc. Introduction page 241 Historical 241 Distribution 242 53a??8 CONTENTS Nomenclature Material and methods Physical characters Bionomics References Plates VII-XIII . page 242 ■ 243 • 243 • 259 264 following page 264 HYDROMEDUSAE FROM THE FALKLAND ISLANDS (published 29th December, 1939) By (the late) Edward T. Browne and P. L. Kramp Introduction . List of Species Systematic Account Anthomedusae Leptomedusae LiMNOMEDUSAE List of Literature . Plates XIV-XIX . page 267 270 271 271 291 302 319 following page 322 MADREPORARIAN CORALS, WITH AN ACCOUNT OF VARIATION IN CARYOPHYLLIA (pubhshed 5th December, 1939) By J. Stanley Gardiner, M.A., F.R.S page 325 Plates XX, XXI following page 328 [Discovery Reports. Vol. XVIII, pp. 1-105, Plate I, May 1938] COAST FISHES PART III. THE ANTARCTIC ZONE By J. R. NORMAN Department of Zoology, British Museum (Natural History) COAST FISHES PART III. THE ANTARCTIC ZONEi By J. R. Norman Department of Zoology, British Museum (Natural History) (Plate I; Text-figs. 1-62) INTRODUCTION THE collections dealt with in this, the third and final part of the report on the coast fishes, include about 1500 specimens, representing 42 species, of which five prove to be new to science,'- and five others were previously unrepresented in the National Collection. The fishes belonging to the division Nototheniiformes are particularly well represented, and, together with the fine series of specimens recently obtained by the B.A.N.Z. Antarctic Research Expedition, and the rich material already in the collection of the British Museum, form an excellent basis for a new revision of this interesting group. Of the 86 species of fish recorded from the Antarctic Zone, no less than 65 (or 75 per cent) belong to the division Nototheniiformes. It may be noted that of these 65 species, 60 (or 92 per cent) are represented in the British Museum: I have been able to examine examples of three more species in Paris and elsewhere, so that only two species are included solely on the evidence of published descriptions. One of these were obtained by the Australasian Antarctic Expedition (1911-1914), and is preserved in the museum at Adelaide, South Australia ; the other formed part of the collection made by the ' Gauss ' (Deutsche Siidpolar-Expedition, 1901-1903), and is presumably to be found in the Berlin Museum. A visit to the Museum National d'Histoire Naturelle in Paris enabled me to examine a number of specimens obtained by the two French Antarctic Expeditions. My thanks are due to the members of the Discovery Committee for permission to study these collections and to prepare this report. I am also indebted to Professor J. Pellegrin, Dr C. F. Angel, and Dr P. Chabanaud, for many courtesies during my short stay in Paris ; to Dr V. Van Straelen, for kindly allowing me to borrow the unique types of Gerlachea australis and Racovitzia glacialis preserved in the Musee Royal d'Histoire Naturelle de Belgique; and to Dr H. Rendahl, for the loan of the unique type of Chionodraco hamatiis preserved in the Naturhistoriska Riksmuseum at Stock- holm. To all these gentlemen my thanks are due and are gratefully tendered. I am also indebted to Mr G. C. L. Bertram for allowing me to study the small collection of fishes 1 For details as to the limits of this region see p. 87. ^ Preliminary descriptions of three of these new species were published recently, see Norman, 1937, Ann. Mag. Nat. Hist., (10) xx, p. 475. 4 DISCOVERY REPORTS made by the British Graham Land Expedition (1934-1937) and for permission to include the names in this report. The illustrations, as in the earlier parts of this report, are the work of Lieut. -Col. W. P. C. Tenison, D.S.O. SYSTEMATIC PART MYXINIDAE Myxine australis, Jenyns. 1842, Zool. 'Beagle', Fish., p. 159; Norman, 1937, Discovery Rep., xvi, p. 6. Hab. Coasts of Patagonia and southern Chile; Falkland Islands; South Shetland Islands. No specimens of this species were obtained by the expedition, but there is one from the South Shetlands in the British Museum collection. This is 230 mm. in total length, and was collected by Mr J. E. Hamilton in 1924. RAJIDAE Raja georgiana, sp.n. St. 39. 25. iii. 26. East Cumberland Bay, South Georgia. From 8 cables S 81° W of Merton Rock to 1-3 miles N 7° E of Macmahon Rock. Large otter trawl, 179-235 m.: i female specimen, 185 mm. (width of disc 140 mm.). Disc much broader than long, its width about | of the total length ; anterior margins a little undulated but not emarginate ; outer angles rounded. Vent a little nearer to tip of snout than to end of tail. Snout scarcely projecting, its length rather more than 1 width of disc ; interorbital width greater than longitudinal diameter of eye and about equal to length of eye + spiracle, which is about 2 in that of snout. Internasal width about I praeoral length of snout. Mouth nearly straight; about 24 rows of teeth in upper jaw. Upper surface of disc more or less evenly covered with small spinules; a single praeocular spine and two postocular spines, of which the hinder is the larger; a pair of strong curved scapular spines set one behind the other ; a single spine on the median line of the back midway between the second postocular and the first scapular spine, and another level with the posterior scapular spines ; a series of similar spines extending in a median line from just behind the suprascapulary region to the first dorsal fin. Lower surface quite smooth. Dorsal fins close to end of tail, their bases contiguous. Pale greyish-brown, mottled and spotted with yellowish-white; hinder margins of pectorals and edges of pelvics with a narrow pale border; lower surface yellowish-white, with a dusky tinge towards the hinder parts of the pectorals and pelvics. Hab. South Georgia. This species appears to be most nearly related to R. macloviana, Norman, from the Patagonian region, and R. miirrayi, Giinther, from Kerguelen, but differs from both in the shape of the disc, in the smaller eyes, in the greater internasal width, and in the RAJIDAE 5 form and arrangement of the spines. It may, of course, prove to be identical with R. arctowskii, Dollo, the only other species of Raja described from the Glacial District (see below). Fig. I. Raja georgiana. Holotype. x |. Raja arctowskii, Dollo. 1904, Res. Voy. ' Belgica' , Poiss., p. 52, pi. ix, fig. 10. St. 170. 23.11. 27. Off Cape Bowles, Clarence Island, 61° 25' 30" S, 53^ 46' 00" W. Large dredge, 342 m. : 6 egg-capsules. Hab. Bellingshausen Sea; between South Shetlands and South Orkneys. DoUo's original description of this species was based upon three egg-capsules, each 60 mm. in length (without the horns), from 70° 15' S, 84° 06' W; 70° 23' S, 82° 47' W; and 71' 19' S, 87° 37' W; at depths ranging from 400 to 569 metres. The fish itself has not yet been recognized. Raja sp. St. 599. 17. i. 31. 67° 08' S, 69° 06-5' W. Large dredge, 203 m.: i large egg-capsule with con- tained embryo. Hab. Graham Land. The length of this capsule, without the horns, is about 200 mm., the greatest width about 80 mm. The embryo is 160 mm. in total length and measures 70 mm. across the disc. In general form the capsule is very similar to that of the European Skate, Raja batis, Linnaeus. Other species of Raja from the Antarctic Zone are R. murrayi, Giinther and R. eatonii, Giinther, both from Kerguelen. 6 DISCOVERY REPORTS MURAENOLEPIDAE Muraenolepis microps, Lonnberg. Muraetwlepis jnarmoratus microps, Lonnberg, 1905, Wiss. Ergebn. Schioed. Si'idpol.-Exped., v (6), p. 43- Muraenolepis microps, Regan, 1914, Rep. Brit. Antarct. {'Terra Nova') Exped. 1910, Zool., i (i), p. I, pi. ii, fig. 2; Norman, 1937, Discovery Rep., xvi, p. 58. St. MS 10. 14. ii. 25. East Cumberland Bay, South Georgia. Small beam trawl, 26-iSm.: I specimen, no mm. St. MS — . 28. ix. 25. Grytviken, South Georgia. From kelp root. 2 specimens, 50, 52 mm. St. MS 68. 2. iii. 26. East Cumberland Bay, South Georgia. Large rectangular net, 220-247 m.: 1 specimen, 285 mm. St. 39. 25. iii. 26. East Cumberland Bay, South Georgia. Large otter trawl, 179-235 m.: 4 specimens, 105-260 mm. St. 42. I. iv. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 120-204 m. : 2 specimens, 130, 158 mm. St. 45. 6. iv. 26. 2-7 miles S 85° E of Jason Light, South Georgia. Large otter trawl, 270- 238 m.: 6 specimens, 150-335 mm. St. 46. 6. iv. 26. 51° 13' 00" S, 49° 50' 00" W. Large otter trawl, 270-238 m.: i specimen, 305 mm. St. 123. 15. xii. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 230- 250 m.: 4 specimens, 248-300 mm. St. WS 25. 17. xii. 26. Undine Harbour (North), South Georgia. Small beam trawl, 18-27 "i-- I specimen, 50 mm. St. 140. 23. xii. 26. Stromness Harbour to Larsen Point, South Georgia. Large otter trawl, 122-136 m.: I specimen, 120 mm. St. 142. 30. xii. 26. East Cumberland Bay, South Georgia. Large otter trawl, 88-273 m.: I specimen, 142 mm. St. 148. 9.1.27. Off Cape Saunders, South Georgia. Largeotter trawl, 132-148 m.: 2 specimens, 235, 290 mm. St. WS 424. 14. iv. 29. 54° 34' 00" S, 38 ' 49' 00" W. I m. tow-net, oblique, 95-0 m.: i specimen, 39 mm. St. — . 29.1.35. Grytviken, South Georgia. Large gauze fish trap, i specimen, 320 mm. St. 1660. 27. i. 36. 74° 46-4' S, 178" 23-4' E. Large otter trawl, 351 m.: i specimen, 220 mm. Depth of body 4I to 6 in the length, length of head 4I to $\. Snout i\ times to twice as long as eye, diameter of which is 4 (young) to 6 in length of head, equal to or less than interocular width, and greater than interorbital width. Barbel i or ^ length of head. Maxillary extending to below anterior part of eye in young, to middle of eye or beyond in larger specimens. Dorsal filament varying from f to about twice diameter of eye, inserted above or a little behind base of pectoral. Origin of anal only a little nearer to end of snout than to base of caudal. Length of pectoral \ to f, of pelvic | to I that of head ; pelvic inserted well in advance of base of pectoral. Uniformly greyish or brownish. Hab. Burdwood Bank, south of the Falkland Islands ; South Georgia ; South Sand- wich Islands; South Victoria Land. This species was originally described from 4 specimens, 140 to 335 mm. in total length, from Cumberland Bay, South Georgia, at depths of 20 to 100 metres. NOTOTHENIIFORMES 7 Examination of this fine series of specimens shows that M. microps is very closely related to M. tnormoratus, Giinther, from Kerguelen, and may even prove to be identical with that species. Comparison of examples of equal size, however, suggests that M. marmoratus has a smaller head, a slightly larger eye, and perhaps a shorter dorsal filament. Fig. 2. Muraenolepis microps. x |. The only other species of Muraenolepis from the Antarctic Zone is M. microcephaliis, Norman, known only from the unique holotype, 125 mm. in total length, from 63° 51' S., 54° 16' E (2000 metres of wire), collected by the B.A.N.Z. Antarctic Research Ex- pedition. This form appears to be most nearly related to M. orangiensis, Vaillant, from the Patagonian region. I have given a key to the four known species of Muraenolepis elsewhere.^ REVISION OF THE ANTARCTIC NOTOTHENIIFORMES Regan's revision published in 1913,'^ together with his subsequent report on the rich material collected by the 'Terra Nova',^ has placed the taxonomy of this group of fishes on a sound basis. In more recent years, however, the Australasian Antarctic Expedition of 1911-1914, and the British, Australian, New Zealand Antarctic Research Expedition of 1929-193 1, together with the important collections of the Discovery Committee, have brought to light a number of additional forms, and have added con- siderably to our knowledge of the fish fauna of the Antarctic Zone.* The problems raised by the study of this material have led me to undertake a new revision of these fishes, which, although it amplifies that of Regan, only modifies his conclusions in certain comparatively unimportant details. Division NOTOTHENIIFORMES Percoid fishes, normally without pungent fin-spines, with the spinous dorsal, when developed, shorter than the long soft dorsal and anal, the principal caudal rays reduced in number (usually 14), the pectorals typically broad-based, and the pelvics jugular, 1 Norman, 1937. Rep. B.A.N.Z. Antarct. Res. Exped., Ser. B, I, p. 69. ^ Trans. R. Soc. Edinburgh, XLix, pp. 249-289. ^ 1914, Rep. Brit. Antarct. (' Terra Nova') Exped. 1910, Zool. I (i), pp. 1-54. * This Zone includes the Glacial District and the Kerguelen District, see Regan, 1914, t.c, pp. 33-36. 8 DISCOVERY REPORTS separated by an interspace, and each formed of a spine and 5 branched rays. There is a single nostril on each side. In osteological characters the more generahzed forms are very similar to the Perciformes, but the structure and position of the radials of the pectoral fins are highly characteristic. The radials are 3 in number, rather large flat plates ; all or 2 are inserted on the hypocoracoid, and the lowest is the narrowest and has its lower edge in contact with the metapterygoid process. [After Regan.] Five families, of which one (Bovichthyidae) does not occur in the Antarctic Zone. Synopsis of the Antarctic Families I. Palatines and pterygoids normally developed; ribs present. A. Spinous dorsal fin present; mouth protractile. 1. Operculum normal; body scaly; gill-membranes forming a fold across the isthmus I . Nototheniidae. 2. Operculum hooked upwards posteriorly, its upper edge deeply concave; body naked; gill-membranes broadly united to isthmus 2. Harpaaiferidae. B. Spinous dorsal fin absent; mouth usually non-protractile ... ... 3. Bathydraconidae. II. Palatine in great part ligamentous; no mesopterygoid ; mouth not protractile; snout pro- duced and depressed; ribs not ossified ... ... ... ... ... 4. Chaenichthyidae . Family NOTOTHENIIDAE Body scaly. Snout not or only a little produced. Mouth protractile; palatines and pterygoids normally developed ; palate toothless. Gill-membranes forming a fold across the isthmus ; operculum normal. Spinous dorsal fin present. All three radials situated on the hypocoracoid. Ribs present. Five genera, of which one (Eleginops) does not occur in the Antarctic Zone. Key to the Antarctic Genera I. Foramen partly bordered by hypocoracoid (fig. 3 b) i. Notothema. II. Hypercoracoid enclosing its foramen (fig. 3 a). A. Scales not very thin, those of upper lateral line with tubules, those of lower lateral line with tubules or pits; two lateral lines; skeleton well ossified; most of the praecaudal vertebrae with parapophyses 1. Teeth in bands, villiform; snout not much longer than eye ... 2. Trematomus. 2. Teeth in upper jaw biserial, those of outer row enlarged, spaced, canine-like; a group of stronger teeth on each praemaxillary; teeth of lower jaw uniserial, spaced, canine-like; snout much longer than eye 3. Dissostichus. B. Scales very thin, cycloid, those of the 3 lateral lines each with a notched hinder margin and with a vertical row of 3 shallow pits; skeleton feebly ossified; only the posterior praecaudal vertebrae with parapophyses 4. Pleuragramma. Genus NOTOTHENIA Notothenia, Richardson, 1844, Zool. 'Erebus' and 'Terror' (Fish.), p. 5; Gunther, i860, Cat. Fish., n, p. 260; Regan, 1913, Trans. R. Soc. Edinburgh, XLix, p. 264. Type Notothenia coriiceps, Richardson. NOTOTHENIIDAE 9 Macronotothen, Gill, 1862, Proc. Acad. N.S. Philad. (1861), p. 520. Type Notothenia} rossit, Richardson. ? Pagothenia, Nichols and LaMonte, 1936, Amer. Mus. Novit., 839, p. 2. Type Pagothenia antarctica, Nichols and LaMonte. Body scaly ; two lateral lines with tubular or pitted scales. Mouth of moderate size or rather large ; teeth usually in bands, but sometimes irregularly bi- or tri-serial with some of the teeth of the outer series enlarged and canine-like. Snout not much longer than the eye. Gill-membranes united, free or forming a free fold across the isthmus. Pectoral fin rounded or vertically truncated. Skeleton well ossified ; vertebrae 44 to 54 (16-20 + 27-34); most of the praecaudals with parapophyses to which the ribs and epipleurals are attached ; hypercoracoid foramen margined below by the hypocoracoid. Coasts of the Antarctic Continent and northwards to Patagonia, Chile, the Falkland Islands, South Georgia, Bouvet Island, Marion Islands, Crozet Islands, Kerguelen, Macquarie Island, southern New Zealand, and Chatham Islands. Fig. 3. Pectoral fin-skeleton of A, Trematomus newnesi; B, Notothenia coriiceps. cl, cleithruin; sc, hyper- coracoid (scapula);/, foramen; cor, hypocoracoid; m, metapterygoid process; i, 2, 3, radials. (After Regan.) Key to the Antarctic Species I. Cheeks and opercles largely scaled, the scaling extending downwards for at least half the depth of the operculum; upper surface of head scaly (except in A'^. nudifrons). A. Pectoral with 21 to 25 rays; dorsal IV-VIII, 31-41 ; anal 30-40. I. Snout and praeorbital scaly (except in young); 12 to 17 gill-rakers on lower part of anterior arch. a. Upper lateral line of 39 to 47 tubular scales; similar scales generally present in lower lateral line; depth 4 to 4I in length. a. Eye 3 to 3.!, interorbital width 9 to 12 in head; dorsal IV-V, 36-37; anal 32; 15 to 18 tubular scales in lower lateral line ... ... ... i. squamifrons. j8. Eye 3, interorbital width 9 in head; dorsal VI, 34; anal 30; 4 to 8 tubular scales in lower lateral line ... ... ... ... ... 2. macrophthalma. y. Eye 2| to 3J, interorbital width 14 to 16 in head; dorsal V-VI, 36-38; anal 32-33; usually 9 to 17 tubular scales in lower lateral line ... ... 3. kempi. b. Upper lateral line of 46 to 58 tubular scales; no tubular scales in lower lateral line; depth 4 J to 5 J in length; dorsal I V-VI, 35-41 ; anal 36-40 ... ... 4. larseni. lo DISCOVERY REPORTS 2. Praeorbital naked; snout naked or scaly; 8 to 13 gill-rakers on lower part of anterior arch. a. 30 to 42 tubular scales in lower lateral line (except in young) ; pelvics shorter than pectorals, scarcely reaching as far as vent; dorsal VII-VIII, 31-34; anal 31-34 5- gibberifrotis. b. Lower lateral line without tubular scales; pelvics longer than pectorals, ex- tending as far as vent or anal fin. a. Upper surface of head scaly; dorsal IV-V, 34-37; anal 33-35 6. mizops. p. Upper surface of head naked ; dorsal I V-VI, 36-40 ; anal 34-36 7. midifrons. B. Pectoral with 19 to 21 rays; dorsal V-VII, 28-31; anal 27-31. 1. Opercles fully scaled; pelvics reaching as far as vent; 54 to 60 scales in a longi- tudinal series; upper lateral line of 36 to 39 tubular scales 8. acuta. 2. Opercles scaly above, naked below ; pelvics not reaching as far as vent ; 46 to 52 scales in a longitudinal series; upper lateral line of 26 to 35 tubular scales. a. Interorbital width 15 to 20 in head; dorsal V-VIII, 28-30; anal 29-31. 9. angustifrons. b. Interorbital width 10 in head; dorsal VII, 29; anal 27 10. marionensis. II. Opercles scaled only on upper part of operculum ; upper surface of head naked. A. Anal of 26 to 34 rays. 1. Interorbital width 5 to 6i in head; pectoral with 20 or 21 rays; cheek scaly below and behind eye, its lower \ (young) or \ (adult) naked; dorsal I V-VI, 33-36; anal 30-34; scales ctenoid on body 11. cyanobrancha. 2. Interorbital width 3I to 5 in head; pectoral with 17 or 18 rays; cheek scaly behind eye; dorsal III-VII, 35-40; anal 27-31; scales cycloid on body 12. coriiceps. 3. Interorbital width 2| to 3! in head; pectoral with 21 to 23 rays; cheek scaly behind and to some extent below eye; dorsal I V-VII, 32-36; anal 26-30; scales cycloid on body i3- ''o^«^'- B. Anal of 22 to 25 rays; interorbital width 2 1 to 4 in head. 1. Dorsal III- VI, 29-31 ; 50 to 60 scales in a lateral longitudinal series 14. macrocephala. 2. Dorsal VI-VIII, 26-27; 84 to 92 scales in a lateral longitudinal series ^^- '^^Ibecki. It has proved impossible to ascertain the true position oi Notothenia phocae, Richard- son (1844, Zool. 'Erebus' and ' Terror' (Fish.), p. 8), and this must remain a doubtful species. The examples mentioned by Richardson as having been preserved in spirits no longer exist, and I have been unable to trace the drawing said to have been made by Dr Hooker. The fish were all taken at 65° S, circa 155° W, from the stomach of a seal, and ranged in size from 65 to 170 mm. in length (without the caudal fin). Notothenia squamifrons, Giinther. Notothenia squamifrons, Giinther, 1880, Shore Fish. 'Challenger' , p. 16, pi. viii, fig. C; Regan, 1913, t.c., p. 270. Depth of body 4^ in the length, length of head 3I. Snout shorter than eye, diameter of which is 3 to 3^ in length of head ; interorbital width 9 to 12. Jaws equal anteriorly ; maxillary extending to below anterior J of eye ; upper surface and sides of head, in- cluding snout and praeorbital, scaly; about 5 rows of scales between the eyes; 14 to 16 gill-rakers on lower part of anterior arch. Scales on body cycloid or feebly ctenoid ; 55 in a longitudinal series from above base of pectoral to caudal, 44 or 45 in upper NOTOTHENIIDAE i, lateral line, which terminates below end of dorsal, or just behind it, 15 to 18 in lower lateral line. Dorsal IV-V, 36-37. Anal 32. Pectoral with 24 or 25 rays, f length of head, rather shorter than pelvics, which reach the anal. Caudal peduncle deeper than long. Body with irregular cross-bars; cheek with two oblique stripes; spinous dorsal partly blackish. Fig. 4. Nofothema squamifrons. x |. Hab. Kerguelen. Known only from the types of the species, two specimens, no and 150 mm. in total length. Notothenia macrophthalma, Norman. Notothenia macrophthalma, Norman, 1937, Discovery Rep., xvi, p. 68, fig. 30. Very closely related to the preceding species, but diameter of eye 3 in length of head; 3 rows of scales between the eyes; 12 gill-rakers on lower part of anterior arch; only 4 to 8 tubular scales in lower lateral line; dorsal VI, 34; anal 30. Fig. 5. Notothenia macrophthalma. Holotype. x; ,2 DISCOVERY REPORTS Hab. Near the Burdwood Bank, south of the Falkland Islands. A full description of this species, which is known only from the unique holotype (190 mm. in total length), has been given in a previous report, but, since it seems to be more nearly related to Antarctic than to Patagonian species of Notothenia, I have included it again here. Notothenia kempi, Norman. Notothenia ketnpi, Norman, 1937, Ann. Mag. Nat. Hist., (10) xx, p. 475. St. 180. II. ii. 27. 17 miles W of North point of Gand Island, Schollaert Channel, Palmer Archipelago. Large dredge, 160 m.: i specimen, 88 mm. St. 181. i2.iii. 27. Schollaert Channel, Palmer Archipelago. Large otter trawl, with nets (4 mm. and 7 mm. mesh) attached, 160-335 m.: 8 specimens, 80-235 mm. St. 182. 14. iii. 27. Schollaert Channel, Palmer Archipelago. Large otter trawl, 278-500 m.: I specimen, 210 mm. Holotype. Fig. 6. Notothenia kempi. Holotype. x |. Depth of body 4 to 4! in the length, length of head 3! to 3! . Snout shorter than eye, diameter of which is 2| to 3^ in length of head ; interorbital width 14 to 16. Jaws equal anteriorly ; maxillary extending to below anterior J of eye ; upper surface and sides of head, including snout and praeorbital, scaly in adult; snout and praeorbital partly or entirely naked in young; upper surface of eyeball scaly in adult; 2 or 3 rows of scales between the eyes; 12 to 14 gill-rakers on lower part of anterior arch. Scales on body ctenoid ; 55 to 59 in a longitudinal series from above base of pectoral to caudal ; 40 to 47 in upper lateral line, which terminates below posterior rays of dorsal or a little farther back, 9 to 17 in lower lateral line. Dorsal V-VI, 36-38. Anal 32-33. Pectoral with about 25 rays, | to | length of head, as long as or a little longer than pelvics, which reach to or nearly to vent in adults, sometimes beyond in young. Caudal subtruncate or a little rounded; caudal peduncle deeper than long. Body with broad irregular NOTOTHENIIDAE 13 cross-bars, which extend on to the base of the dorsal fin ; cheek with two obUque dark stripes; spinous dorsal with a black blotch; soft dorsal with a submarginal blackish band, which in adults covers the greater part of the fin ; anal fin blackish in adults, the margin narrowly white; the whole anal fin pale in young; caudal, pectorals and pelvics all more or less dusky in adults, with somewhat narrow pale margins. Hab. Palmer Archipelago, Graham Land. This well-marked species was named after Dr Stanley W. Kemp, F.R.S., the former Director of Research. Notothenia larseni, Lonnberg. Notothenia larseni, Lonnberg, 1905, Wiss. Ergebn. Schwed. Siidpol.-Exped., v (6), pp. 31, 46, pi. i, fig. 3, pi. ii, fig. 6; Regan, 1913, t.c, p. 272. St. MS 67. 28. ii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 38 m.: i specimen, 43 mm. St. MS 71. 9.111.26. East Cumberland Bay, South Georgia. Small beam trawl, 110-60 m.: 21 specimens, 60-80 mm. St. 39. 25. ill. 26. East Cumberland Bay, South Georgia. Large otter trawl, 179-235 m.: 6 specimens, 65-102 mm. St. 42. I. iv. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 120-204 m. : numerous specimens, 52-195 mm. St. 45. 6. Iv. 26. 27 miles S 85° E of Jason Light, South Georgia. Large otter trawl, 238- 270 m.: 135 specimens, 60-185 mm. St. 123. 15. xii. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, with tow-net attached, 230-250 m.: 14 specimens, 52-104 mm. St. 140. 23. xll. 26. Stromness Harbour to Larsen Point, South Georgia. Large otter trawl, 122-136 m.: 8 specimens, 60-123 rnm. St. 142. 30. xll. 26. East Cumberland Bay, South Georgia. Net (7 mm. mesh) attached to back of trawl, 8S-273 m.: 3 specimens, 115-145 mm. St. 144. 5. 1. 27. Off mouth of Stromness Harbour, South Georgia. Net (4 mm. mesh) attached to back of trawl, 155-178 m.: 3 specimens, 60-86 mm. St. WS 42. 7. i. 27. 54° 41' 45" S, 36° 47' 00" W. I m. tow-net, horizontal, 198 m.: i specimen, 163 mm. St. 148. 9. 1. 27. Off Cape Saunders, South Georgia. Large otter trawl, with net (4 mm. mesh) attached, 132-148 m.: 13 specimens, 58-104 mm. St. 149. 10. 1. 27. Mouth of East Cumberland Bay, South Georgia. Large otter trawl, 200- 234 m.: 4 specimens, 150-170 mm. St. WS 62. 19.1.27. Wilson Harbour, South Georgia. Small beam trawl, 26-83 m. 145 specimens, 57-64 mm. St. 156. 20. 1. 27. 53 ' 51' 00" S, 36° 21' 30" W. Large dredge, 200-236 m.: i specimen, 160 mm. St. 160. 7.11.27. Near Shag Rocks, 53° 43' 40" S, 40° 57' 00" W. Large dredge, 177 m.: 2 specimens, 70, 72 mm. St. 167. 20.11. 27. Off Signy Island, South Orkneys. Net (7 mm. mesh) attached to back of trawl, 244-344 m.: 13 specimens, 123-176 mm. St. 175. 2. ill. 27. Bransfield Strait, South Shetlands. Large dredge, 200 m.: 2 specimens, 170, 170 mm. St. 181. 12. iii. 27. Schollaert Channel, Palmer Archipelago. Nets (4 mm. and 7 mm. mesh) attached to back of trawl, 160-335 m.: 17 specimens, 100-153 n^"™- 14 DISCOVERY REPORTS St. WS 165. i.iii. 28. 53"44'S, 38" 46' W. 70 cm. tow-net, oblique, 133-0 m.: i specimen, 93 mm. St. WS 179. 7. iii. 28. 55' 08' 00" S, 35° 20' 00" W. I m. tow-net, oblique, 112-0 m.: i specimen, 72 mm. St. WS 358. 13-14.1.29. 55' 02' 30" S, 35" 18' 00" W. I m. tow-net, oblique, 85-0 m.: i specimen, 53 mm. St. 341. 5-6. ii. 30. 54°43'S, 36°42i'W. i m. tow-net, oblique, iii-om.: i specimen, 172 mm. St. 366. 6. iii. 30. 4 cables S of Cook Island, South Sandwich Islands. Large dredge, 155- 322 m. : 10 specimens, 52-1 10 mm. St. 371. 14. iii. 30. I mile E of Montagu Island, South Sandwich Islands. Large otter trawl, with nets (4 mm. and 7 mm. mesh) attached, 99-161 m.: 17 specimens, 65-195 mm. St. 456. 18. X. 30. I mile E of Bouvet Island. Large dredge, 40-45 m.: 40 specimens, 90- 190 mm. St. 474. 12. xi. 30. I mile W of Shag Rocks, South Georgia. Large dredge, 199 m. : 4 specimens, 63-70 mm. St. 476. 12. xi. 30. I mile N of Shag Rocks, South Georgia, i m. tow-net, oblique, 165 m.: 6 specimens, 63-72 mm. St. 1873. I3.xi. 36. 61' 20-8' S, 54° 04-2' W. Rectangular dredge, with Russell net, 117 m.: I specimen, 200 mm.; 210-180 m.: 4 specimens, 85-190 mm. Fig. 7. Notothenia larseni. x |. Depth of body 4^ to 5I in the length, length of head 3I to nearly 4. Snout shorter than eye, diameter of which is 2| to 3J in length of head; interorbital width 11 to 13. Jaws equal anteriorly ; maxillary extending to or a little beyond vertical from anterior margin of eye; upper surface and sides of head, including snout and praeorbital, scaly in adult, snout and praeorbital often partly or entirely naked in young; 14 to 17 gill-rakers on lower part of anterior arch. Scales on body ctenoid ; 58 to 65 in a longi- tudinal series from above base of pectoral to caudal ; 46 to 58 in upper lateral line, which extends beyond end of dorsal; lower lateral line without tubular scales. Dorsal IV-VI, 35-41. Anal 36-40. Pectoral with 23 to 26 rays, about | length of head, a little longer than pelvics, which just reach the anal. Caudal subtruncate or a little rounded ; caudal peduncle about as long as deep. Body with some irregular oblique dark cross-bars, which in the young tend to break up into 3 or 4 series of alternating spots or blotches ; distal part of spinous dorsal with a black blotch ; soft dorsal with oblique dark stripes or series of spots ; anal plain or similarly marked ; caudal often with indistinct dark cross-bars, sometimes dusky distally; pectoral more or less dusky, with pale posterior margin. NOTOTHENIIDAE 15 Hab. Palmer Archipelago, Graham Land; South Shetlands; South Orkneys; South Georgia; South Sandwich Islands; Bouvet Island. Lonnberg described this species from 8 specimens : i from the Shag Rocks, 5 from outside Cumberland Bay, South Georgia, and 2 from south of Snow Hill; at depths ranging from 250 to 310 metres. It was named after Capt. C. A. Larsen, the chief navigator of the Swedish Antarctic Expedition. Although nearly the commonest species obtained in this region by the Discovery Committee, N. larseni was not previously represented in the British Museum collection. Notothenia gibberifrons, Lonnberg. Notothenia gibberifrons, Lonnberg, 1905, Wiss. Ergebn. Schwed. Siidpol.-Exped., v (6), p. 33, pi. iii, fig. 10; Vaillant, 1906, Bull. Mus. Paris, 1906, No. 3, p. 139; 1906, Exped. Aniarci. Franf. (1903-1905), Poiss., p. 33; Lonnberg, 1906, K. Svensk. Vet.-Akad. Handl., xl (5), p. 93; Regan, 1913, t.c, pp. 240, 271. Notothenia acuta, Vaillant, 1906, t.c, p. 139; 1906, t.c, p. 31. Notothenia vaillanti, Regan, 1913, t.c, p. 272. St. MS 24. 13. iv. 25. East Cumberland Bay, South Georgia. Small beam trawl, 30 m.: 2 specimens, 33, 38 mm. St. MS 32. I. V. 25. East Cumberland Bay, South Georgia. Small beam trawl, 40 m.: 44 speci- mens, 39-210 mm. St. — . Feb., 1926. Cumberland Bay, South Georgia. Long line: i specimen, 223 mm. St. MS 63. 24. ii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 23 m.: i specimen, 145 mm. St. MS 67. 28. ii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 38 m.: 2 specimens, 65, 85 mm. St. MS 71. 9. iii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 110-60 m.: 3 specimens, 130-225 mm. St. MS 74. 17. iii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 22-40 m.: 2 specimens, 33, 38 mm. St. 39. 25. iii. 26. East Cumberland Bay, South Georgia. Large otter trawl, 179-235 m.: 23 specimens, 38-400 mm. St. 42. I. iv. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 120-204 m.: 30 specimens, 35-280 mm. St. 45. 6. iv. 26. 27 miles S 85° E of Jason Light, South Georgia. Large otter trawl, 270- 238 m.: 27 specimens, 36-335 mm. St. 123. 15. xii. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 230- 250 m.: 3 specimens, 61-300 mm. St. 140. 23. xii. 26. Stromness Harbour to Larsen Point, South Georgia. Large otter trawl, 122-136 m.: 4 specimens, 68-145 ™^^- St. WS 62. 19.1.27. Wilson Harbour, South Georgia. Small beam trawl, 26-83 m.: 72 specimens, 47-100 mm.; small beam trawl, 15-45 iri-- 3 specimens, 50-53 mm. St. 167. 20. ii. 27. Off Signy Island, South Orkneys. Net (7 mm. mesh) attached to back of trawl, 244-344 ™-- 16 specimens, 142-270 mm. St. 189. 23-25. iii. 27. Port Lockroy, Wiencke Island, Palmer Archipelago. Large fish trap, 70 m.: 6 specimens, 265-375 mm. St. WS 358. 13-14.1.28. 55° 02' 30" S, 35" 18' 00" W. I m. tow-net, oblique, 85-0 m.: i specimen, 57 mm. St. 371. 14. iii. 30. I mile E of Montagu Island, South Sandwich Islands. Large otter trawl, with nets (4 mm. and 7 mm. mesh) attached, 99-161 m.: 6 specimens, 170-210 mm. i6 DISCOVERY REPORTS St. 393 c. 8. V. 30. 54° 17' S, 35° 30' W. I m. tow-net, oblique, 121-187 m.: i specimen, 39 mm. No data. 2 specimens, 325, 330 mm. Fig. 8. Nolothenia gibberifrons . x |. Depth of body 5 to 5I in the length, length of head 3I to 3|. Snout as long as or rather shorter than eye, diameter of which is 3 (young) to 4I in length of head ; inter- orbital width 12 to 16. Jaws equal anteriorly; maxillary not or only just reaching vertical from anterior margin of eye ; cheeks, opercles, upper surface of head, and greater part of snout scaly in adult, but praeorbitals naked ; 8 to 11 gill-rakers on lower part of anterior arch. Scales on body ctenoid ; 54 to 60 in a longitudinal series from above base of pectoral to caudal ; 34 to 47 (5 1) in upper lateral line, which ends below posterior part of dorsal; 30 to 42 in lower lateral line, which generally extends forward to or beyond level of tip of pectoral ; lower lateral line usually without tubular scales in young. Dorsal VII-VIII, 31-34. Anal 31-34. Pectoral with 21 or 22 rays, | to | length of head, longer than pelvics, which reach vent or origin of anal in young, but not or scarcely as far as vent in adult. Caudal rounded or subtruncate ; caudal peduncle about as long as deep. Upper part of head and body irregularly spotted and blotched, but only rarely with traces of darker cross-bars ; young with irregular cross-bars, broken up into 3 or 4 series of alternating spots ; usually a more distinct Y-shaped mark at base of caudal ; dorsals, caudal, pectorals and sometimes the pelvics with series of dark spots forming bars ; anal pale. Hab. Palmer Archipelago, Graham Land ; South Shetlands ; South Orkneys ; South Sandwich Islands ; South Georgia. This species was originally described from numerous specimens, all from South Georgia, at depths ranging from 20 to 310 metres. According to Lonnberg, it is a sluggish bottom fish, but the Discovery Committee obtained some young examples in tow-nets. Lonnberg records a specimen 460 mm. in total length. The large specimen, 410 mm. in total length, described by Vaillant as Notothenia acuta, has been examined by me in the Paris Museum (No. 06—132), and undoubtedly belongs to this species. There are also two small specimens (No. 06—131), 53 and 56 mm. in total length, the larger of which is the holotype of N. vaillanti, Regan. These prove to be young A'^. gibberifrons, in which the snout is always more or less naked and the tubular scales in the lower lateral line are usually absent. The ' Fran9ais ' obtained these specimens from Booth Wandel Island and Wiencke Island, at depths varying from 20 to 40 metres. NOTOTHENIIDAE 17 Notothenia mizops, Giinther. Notothenia mizops, Giinther, 1880, Shore Fish. 'Challenger', p. 16, pi. viii, fig. D; Regan, 1913, i.e., p. 273; Norman, 1937, Rep. B.A.N. Z. Antarct. Res. Exped., Ser. B, i, p. 61. Depth of body 4 J to 4I in the length, length of head 3 (young) to 4. Snout much shorter than eye, diameter of which is 3 to 3 J in length of head; interorbital width about 15. Jaws equal anteriorly; maxillary extending to below anterior | or | of eye; cheeks, opercles occiput and interorbital region scaly; 9 to 13 gill-rakers on lower part of anterior arch. Scales on body ctenoid; 48 to 55 in a longitudinal series from above base of pectoral to caudal; 30 to 38 in upper lateral line, which ends below posterior part of dorsal ; lower lateral line without tubular scales. Dorsal IV-V, 34-37. Anal 33-35. Pectoral with 21 or 22 rays, | to | length of head, shorter than pelvics, which extend to, nearly to, or beyond origin of anal. Caudal rounded; caudal peduncle deeper than long. Body with two series of large, partly confluent, irregular blackish spots; cheek with two oblique dark stripes; a blackish spot on spinous dorsal; soft dorsal, caudal and anal with or without series of dark spots. Hab. Kerguelen. The above description is based upon the types of the species, 5 specimens, 70 to 170 mm. in total length, from off Christmas Harbour and Howes Foreland, 220 metres; and upon 10 specimens, 40 to 145 mm. in total length, collected by the B.A.N. Z. Antarctic Research Expedition, at depths ranging from 20 to 150 metres. Notothenia nudifrons, Lonnberg. Notothenia mizops var. tmdifrons, Lonnberg, 1905, Wiss. Ergebn. Schwed. Siidpol. -Exped., v (6), pp. 30, 45, pi. i, fig. 2; Lonnberg, 1906, K. Svensk. Vet.-Akad. Handl., xl (5), p. 92; Roule and Despax, 1911, Bull. Mus. Paris, 1911, No. 5, p. 278; Roule, Angel and Despax, 1913, Deux. Exped. Antarct. Franf. (1908-1910), Poiss., p. 4. Notothenia mizops, Vaillant, 1906, Bull. Mus. Paris, 1906, No. 3, p. 139; 1906, Exped. Antarct. Franf. (1903-1905), Poiss., p. 30. Notothenia nudifrons, Regan, 1913, t.c, pp. 240, 273. St. MS 6. 14. ii. 25. East Cumberland Bay, South Georgia. Small beam trawl, 24-30 m.: 4 specimens, 38-220 mm. St. MS ID. 14. ii. 25. East Cumberland Bay, South Georgia. Small beam trawl, 26-18 m.: II specimens, 47-133 mm. St. MS 12. 17. ii. 25. East Cumberland Bay, South Georgia. Small beam trawl, 25 m.: i specimen, 105 mm. St. MS 25. 13. iv. 25. East Cumberland Bay, South Georgia. Small beam trawl, 36 m.: 2 specimens, 50, 80 mm. St. MS 32. i.v. 25. East Cumberland Bay, South Georgia. Small beam trawl, 40 m.: 2 specimens, 94, 142 mm. St. MS 33. I. V. 25. I cable E of Hobart Rock, East Cumberland Bay, South Georgia. Small beam trawl, 40 m. : i specimen, 95 mm. St. MS 62. 24. ii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 31m.: 6 specimens, 40-84 mm. St. MS 63. 24. ii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 23 m.: 6 specimens, 37-67 mm. i8 DISCOVERY REPORTS St. MS 67. 28. ii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 38 m.: 8 specimens, 42-185 mm. St. MS 71. 9. iii. 26. East Cumberland Bay, South Georgia. Small beam trawl, with tow-net of coarse silk attached, 110-60 m.: 26 specimens, 42-150 mm. St. MS 74. 17. iii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 22-40 m.: 18 specimens, 14-158 mm. St. 39. 25. iii. 26. East Cumberland Bay, South Georgia. Large otter trawl, 179-235 m.: 3 specimens, 130-150 mm. St. 42. I. iv. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 120-204 m. : 3 specimens, 52-66 mm. St. WS 25. 17. xii. 26. Undine Harbour (North), South Georgia. Small beam trawl, 18-27 ^^'^ 45 specimens, 40-145 mm. St. 140. 23. xii. 26. Stromness Harbour to Larsen Point, South Georgia. Large otter trawl, 122-136 m.: 14 specimens, 45-155 mm. St. 141. 29. xii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 17-27 m.: 3 specimens, 85-195 mm. St. 144. 5. i. 27. Off mouth of Stromness Harbour, South Georgia. Net (4 mm. mesh) attached to back of trawl, 155-178 m.: 5 specimens, 53-90 mm. St. 145. 7. i. 27. Stromness Harbour, South Georgia. Net (4 mm. mesh) attached to back of trawl, 26-35 ^■'- ^4 specimens, 55-170 mm. St. 148. 9. i. 27. Off Cape Saunders, South Georgia. Large otter trawl, with net (4 mm. mesh) attached, 132-148 m.: 17 specimens, 47-160 mm. St. 149. 10. i. 27. Mouth of East Cumberland Bay, South Georgia. Large otter trawl, 200- 234 m.: 3 specimens, 163-176 mm. St. WS62. 19.1.27. Wilson Harbour, South Georgia. Small beam trawl, 26-83 m. : 6 specimens, 52-130 mm. St. 156. 20.1.27. 53° 51' 00" S, 36° 21' 30" W. Large dredge, 200-236 m.: 3 specimens, 62- 128 mm. St. 160. 7.11.27. Near Shag Rocks, 53° 43' 40" S, 40° 57' 00" W. Large dredge, 177 m.: 5 specimens, 73-153 mm. St. 163. 17. ii. 27. Paul Harbour, Signy Island, South Orkneys. Small beam trawl, 18-27 ^■'■ I specimen, 70 mm. St. 173. 28. ii. 27. Port Foster, Deception Island, South Shetlands. Small beam trawl, 5-60 m.: 3 specimens, 73-142 mm. St. 184. 15-16. iii. 27. Fournier Bay, Anvers Island, Palmer Archipelago. Large fish trap, 36 m. : I specimen, 100 mm. St. 190. 24. iii. 27. Bismarck Strait, Palmer Archipelago. Large dredge— large rectangular net, 93-130 m.: I specimen, 152 mm. St. 371. 14. iii. 30. I mile E of Montagu Island, South Sandwich Islands. Large otter trawl, with nets (4 mm. and 7 mm. mesh) attached, 99-161 m.: 4 specimens, 130-165 mm. Depth of body 4 to 4I in the length, length of head 3^ to 3I. Snout as long as or rather shorter than eye, diameter of which is 3 to 4 in length of head ; interorbital width 14 to 18. Jaws equal anteriorly; maxillary extending to below anterior J or ^ of eye; cheeks and opercles more or less fully scaled, but lower part of cheek and edge of praeoperculum naked; occiput, interorbital region, snout and praeorbital naked; 10 to 13 gill-rakers on lower part of anterior arch. Scales on body ctenoid; 55 to 62 in a longitudinal series from above base of pectoral to caudal; 35 to 42 in upper lateral line, which ends below posterior part of dorsal; lower lateral line without tubular NOTOTHENIIDAE 19 scales. Dorsal IV-VI, 36-40. Anal 34-36. Pectoral with 22 or 23 rays, f to | length of head, generally rather shorter than pelvics, which reach vent or anal. Caudal rounded or subtruncate ; caudal peduncle much deeper than long. Body with 2 or 3 series of large, partly confluent, irregular blackish spots and blotches, sometimes forming more or less definite oblique cross-bars ; cheek usually with two oblique dark stripes ; a round black spot on spinous dorsal; soft dorsal, caudal and anal with or without series of small dark spots, which, when present, may form stripes or bars ; pectorals and pelvics pale, sometimes with narrow darker cross-bars. Fig. 9. Notothetiia nudifrons. A, Dorsal view of head of N. midifrons; B, The same of N. mizops. x |. Hab. Graham Land and neighbouring islands; South Shetlands; South Orkneys; South Sandwich Islands ; South Georgia. This species was originally described from 7 specimens from Cape Seymour and south of Snow Hill, Graham Land, and from numerous specimens from South Georgia, at depths ranging from 12 to 250 metres. It is very common at South Georgia, at all depths from 10 to 250 metres. In life the fish appears to be reddish or reddish-brown on the back and sides and silvery white below ; the spots are dark brown. In general appearance this species is very similar to A'^. mizops, from Kerguelen, but may be readily distinguished by the naked occiput and interorbital region, the some- what more numerous dorsal and anal rays, and the smaller scales. I have examined 3 specimens in the Paris Museum, 62 to 118 mm. in total length, from the South Shetlands. Notothenia acuta, Giinther. Notothenia acuta, Giinther, 1880, Shore Fish. 'Challenger', p. 17; Pappenheim, 1912, Deutsche Siidpol.-Exped., xni, Zool. v, p. 171, pi. ix, fig. 3; Regan, 1913, t.c, p. 272, pi. viii, fig. 3; Norman, 1937, Rep. B.A.N.Z. Antarct. Res. Exped., Ser. B, i, p. 62. Depth of body 5I to 6 in the length, length of head 3 to 3! . Snout a little shorter than eye, diameter of which is 3 to 3! in length of head; interorbital width about 16. Jaws equal anteriorly; maxillary extending to below anterior part of eye; cheeks, opercles, occiput and interorbital region scaly; snout naked in young, but partially scaled in larger specimens; praeorbital naked; 12 or 13 gill-rakers on lower part of anterior arch. Scales on body ctenoid; 54 to 60 in a longitudinal series from above base of pectoral to caudal; 36 to 39 in upper lateral line, which ends below posterior 3-2 20 DISCOVERY REPORTS part of dorsal or extends to base of caudal ; 15 to 18 in lower lateral line. Dorsal VI-VII, 28-30. Anal 28 31. Pectoral with 19 or 20 rays, as long as or nearly as long as head, longer than pelvics, which reach the vent. Caudal rounded ; caudal peduncle somewhat deeper than long. Body marbled with darker, and with 3 or 4 short, broad dark bars on upper parts of sides ; dorsal rays sometimes with series of small spots ; caudal fin faintly barred. Hab. Kerguelen. This description is based upon the type, about 62 mm. in total length, and upon 5 specimens, 48 to 95 mm. in total length, collected by the B.A.N. Z. Antarctic Research Expedition. The species occurs in fairly shallow water, the greatest depth at which it has been taken being 30 metres. Notothenia angustifrons, Fischer. Notothenia angustifrons, Fischer, 1885, Jahrh. Hamburg. Wiss. Anst., n, p. 55; Regan, 1913, t.c, p. 274, pi. viii, fig. I. Notothenia marionensis, Lonnberg, 1905, Wiss. Ergebn. Schwed. Siidpol.-Exped., v (6), p. 27; 1906, K. Svensk. Vet.-Akad. Handl. XL (5), p. 93. St. MS 6. 14. ii. 25. East Cumberland Bay, South Georgia. Small beam trawl, 24-30 m.: 1 specimen, 64 mm. St. MS ID. 14. ii. 25. East Cumberland Bay, South Georgia. Small beam trawl, 26-18 m.: 12 specimens, 62-125 mm. St. MS 12. 17. ii. 25. East Cumberland Bay, South Georgia. Small beam trawl, 25-60 m.: 2 specimens, 70, 120 mm. St. MS 25. 13. iv. 25. East Cumberland Bay, South Georgia. Small beam trawl, 36 m.: i specimen, 70 mm. St. MS 32. I. V. 25. East Cumberland Bay, South Georgia. Small beam trawl, 40 m.: 2 speci- mens, 88, 182 mm. St. MS 33. I. V. 25. I cable E of Hobart Rock, East Cumberland Bay, South Georgia. Small beam trawl, 40 m.: i specimen, 86 mm. St. MS 62. 24. ii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 31m.: 2 specimens, 73, 77 mm. St. MS 63. 24. ii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 23 m.: 3 specimens, 68-100 mm. St. MS 65. 28. ii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 20-12 m.: 2 specimens, 120, 150 mm. St. MS 66. 28. ii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 18 m.: 3 specimens, 72-88 mm. St. MS 67. 28. ii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 38 m.: 2 specimens, 65, 85 mm. St. MS 71. 9. iii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 1 10-60 m.: I specimen, 60 mm. St. MS 74. 17. iii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 22-40 m.: 8 specimens, 55-205 mm. St. — . 2. iv. 26. King Edward's Cove, South Georgia. Hand line, 5 m.: i specimen, 68 mm., found in mouth of Notothenia rossii. St. WS 25. 17. xii. 26. Undine Harbour (North), South Georgia. Small beam trawl, 18-27 m.: 15 specimens, 35-155 mm. NOTOTHENIIDAE 21 St. 141. 29. xii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 17-27 m.: 14 specimens, 57-115 mm. St. 145. 7. i. 27. Stromness Harbour, South Georgia. Small beam trawl, 26-35 m.: 3 specimens, 103-180 mm. St. WS 62. 19. i. 27. Wilson Harbour, South Georgia. Small beam trawl, 26-83 m. : 2 specimens, 50, 52 mm. St. 370. 10. iii. 30. 2 miles NE of Bristol Island, South Sandwich Islands. Large otter trawl, 80-18 m.: 42 specimens, 75-148 mm. St. — . 25. iii. 30. Grytviken, South Georgia. Hand line, 5-5 to 10 metres: i specimen, 72 mm., from stomach of Notothenia rossii. Depth of body 5 to 5 f in the length, length of head 3^ to 3I. Snout as long as or rather longer than eye, diameter of which is 3 1 to nearly 5 in length of head; inter- orbital width 15 to 20. Jaws about equal anteriorly; maxillary extending to below anterior margin or anterior 5 of eye ; upper surface of head scaly to between nostrils ; cheeks and opercles in great part scaly, but naked below and at edge of praeoperculum ; snout and praeorbital naked; 10 to 12 gill-rakers on lower part of anterior arch. Scales on body ctenoid ; 46 to 52 in a longitudinal series from above base of pectoral to caudal ; 26 to 34 in upper lateral line, which ends below middle or posterior part of dorsal ; 15 to 26 in lower lateral line. Dorsal V-VIII, 28-30. Anal 29-31. Pectoral with 19 to 21 rays, f to f length of head, much longer than pelvics, which do not reach the vent in adult. Caudal rounded; caudal peduncle about as long as deep. Back with dark transverse bars, which break up into spots on the sides of the body ; often a bar through the spinous dorsal, connecting the bases of the pectorals ; dorsal, caudal and pectorals with series of small dark spots on the rays ; anal and pelvics pale, sometimes with some darker spots ; often a blackish spot on the upper part of the pectoral base. Hab. South Sandwich Islands; South Georgia. Fig. 10. Notothenia angustifrons . xf. A, Dorsal view of head of N. angiistifrons; B, The same of A^. marionensis. x i . This species, which is very common in shallow water round South Georgia, was originally described from 2 specimens, 82 and 88 mm. in length respectively, now preserved in the Zoological Museum at Hamburg (No. 3921). According to Lonnberg, this fish lives on the bottom among the algae, but is also often found " lying openly on the clayey bottom, inside the kelp". Unlike A'', gibberifrons, it is said to be very active and difficult to catch. 22 DISCOVERY REPORTS Notothenia marionensis, Giinther. Notothenia marionensis, Giinther, 1880, Shore Fish. 'Challenger', p. 17; Regan, 1913, t.c, p. 273, pi. viii, fig. 2. Very closely related to A'^. angustifrons, but the interorbital width is 10 in the length of the head. Dorsal VII, 29. Anal 27. Pectoral | length of head. Caudal peduncle a little deeper than long. Hab. Marion Island, Kerguelen District. This species is known only from the holotype, 82 mm. in total length, from 100 to 150 metres. Notothenia cyanobrancha, Richardson. Notothenia cyanobrancha, Richardson, 1844, Zool. 'Erebus' and 'Terror' (Fish.), p. 7, pi. iv; Studer, 1879, Arch. Naturg., xlv (i), p. 131; Pappenheim, 1912, Deutsche Siidpol.-Exped., xni, Zool. V, p. 171; Norman, 1937, Rep. B.A.N.Z. Antarct. Res. Exped., Ser. B, i, p. 62. Notothenia purpuriceps, Richardson, 1844, t.c, p. 7, pi. ii, figs. 3, 4; Gunther, i860. Cat. Fish., n, p. 262; Gill, 1876, Bull. U.S. Nat. Mus., iii, p. 41; Studer, 1879, t.c, p. 131. Notothenia cyaneobrancha, Giinther, i860, t.c, p. 261; Regan, 1913, t.c, p. 275. Notothenia coriiceps, Pappenheim, 1912, t.c, p. 170, pi. ix, fig. 2. Fig. II. Notothenia cyanobrancha. x |. Depth of body 4 to 5 in the length, length of head 3 to 4. Snout as long as or a little longer than eye, diameter of which is 4 to 6| in length of head ; interorbital width 5 to 6| . Jaws equal anteriorly ; maxillary extending to below middle or posterior part of eye ; upper surface of head naked except for a few temporal and post-temporal scales, which may be absent in the young; cheek scaly behind and below eye, the lower \ (young) or \ (aduh) naked; upper part of operculum scaly; 10 to 13 gill-rakers on lower part of anterior arch. Scales on body ctenoid ; 60 to 70 in a longitudinal series from above base of pectoral to caudal; 32 to 39 in upper lateral line which ends below posterior part of dorsal; lower lateral line, when developed, with 6 to 15 tubular scales- Dorsal IV-VI, 33-36. Anal 30-34. Pectoral with 20 or 21 rays, f to f length of head, as long as or longer than pelvics, which do not generally reach the vent in aduh. Caudal rounded; caudal peduncle much deeper than long. More or less uniformly brownish or black, but sometimes with traces of darker markings ; usually an oblique dark stripe from eye to angle of praeoperculum, and another below this ; all the fins more or less dusky or blackish, at least in adult. NOTOTHENIIDAE 23 Hab. Kerguelen. In addition to a number of specimens in the British Museum collection, 120 to 260 mm. in total length, including the type of the species (260 mm.) and the type of N . purpuriceps (128 mm.), the above description is based upon a large series of examples, 37 to 280 mm. in total length, collected by the B.A.N.Z. Antarctic Research Expedition. This is a shallow water and littoral species, which is abundant under stones between the tide-marks. Notothenia coriiceps, Richardson. Notothenia coriiceps, Richardson, 1844, Zool. 'Erebus' and ' Terror' (Fish.), p. 5, pi. iii, figs, i, 2; Giinther, i860, Cat. Fish., u, p. 261; Giinther, 1879, Phil. Trans. R. Sac, CLXvni (extra vol.), p. 166; Vaillant, 1906, Bull. Mm. Paris, 1906, No. 3, p. 139; 1906, Exped. Antarct. Franf. (1903-1905), Poiss., p. 24; Roule and Despax, 191 1, Bull. Mus. Paris, 191 1, No. 5, p. 278; Roule, Angel and Despax, 1913, Deux. Exped. Antarct. Franf. (1908-1910), Poiss., p. 3; Regan, 1913, t.c, pp. 240, 276; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, in (i). Fishes, p. 22; Norman, 1937, Rep. B.A.N.Z. Antarct. Res. Exped., Ser. B, I, p. 63. ? Pagothenia antarctica, Nichols and LaMonte, 1936, Amer. Mus. Novit., 839, p. 3, fig. i. St. — . 1922-23. Deception Island, South Shetlands, 10-30 m.: i specimen, 240 mm., collected by Mr A. G. Bennett. St. 154. 18. i. 27. Jason Harbour to Larsen Point, South Georgia. Large otter trawl, 60-160 m.: I specimen, 530 mm. St. — . 29. i. 27. Grytviken, South Georgia. Hand line, 3-4 m. : i specimen, 340 mm. St. 163. 17. ii. 27. Paul Harbour, Signy Island, South Orkneys. Small beam trawl, 18-27 m.: I specimen, 420 mm. St. 165. 18-19. ii- 27- Dove Strait, SE of Queen's Bay, Signy Island, South Orkneys. Large fish trap, 10 m.: 6 specimens, 330-370 mm. St. 173. 28. ii. 27. Port Foster, Deception Island, South Shetlands. Small beam trawl, 5-60 m. : 1 specimen, 72 mm. St. 174. 28. ii.-2. iii. 27. Deception Island, South Shetlands. Large fish trap, 5-10 m.: 5 specimens, 240-360 mm. St. 179. 10. iii. 27. Melchior Island, Schollaert Channel, Palmer Archipelago. Small dredge, 4-10 m.: I specimen, 140 mm. St. — . 13. iv. 27. Deception Island, South Shetlands. From trap on outer side of entrance: 2 specimens, 230, 235 mm. St. 1092. 18. i. 33. Signy Island, South Shetlands. Found in bird's nest: i specimen, 72 mm. Fig. 12. Notothenia coriiceps. 24 DISCOVERY REPORTS Depth of body 3I to 4I in the length, length of head 2| to 3I. Snout longer than eye, diameter of which is 4 (young) to 7 in length of head; interorbital width 3I to 5 (3! in large specimens). Jaws equal anteriorly; maxillary extending to below anterior part or middle (young) or posterior margin (adult) of eye; head naked, except for a few scales behind eye, on upper part of operculum, and on post-temporal region ; 10 to 14 gill-rakers on lower part of anterior arch. Scales on body cycloid; 54 to 68 in a longitudinal series from above base of pectoral to caudal; 34 to 49 in upper lateral line, which ends below posterior part of dorsal; 6 to 17 in lower lateral line. Dorsal III-VII, 35-40. Anal 27-31. Pectoral with 17 or 18 rays, from less than § (large specimens) to f length of head, longer than pelvics, which do not or scarcely reach the vent. Caudal subtruncate or a little rounded in adult and half-grown specimens, emarginate in young ; caudal peduncle nearly as long as deep. Coloration varying from dark greenish-black to a pale orange-yellow, with or without spots and markings; usually one or two oblique dark bars across cheek, sometimes broken up into spots; head sometimes with pale spots enclosed in dark rings; spots on body and dorsal fin sometimes large and tessellated, more often small and scattered, never very distinct and rarely uniting to form longitudinal stripes; dorsal and anal usually with pale edges, and sometimes with dark submarginal bands; pectorals and caudal more or less dusky; pelvics paler. Young dark on back, silvery on sides and beneath; a large black blotch on extremity of pectoral fin. Hab. Graham Land and neighbouring islands; South Shetlands; South Orkneys; South Georgia; Kerguelen; Crozets; Heard Island; Adelie Land; Victoria Land. This is the only species of Notothenia occurring on the coasts of the Victoria Quadrant of the Antarctic Continent. The type of the species from Kerguelen measures 255 mm. in total length. The silvery young, which are probably pelagic in habit, are very similar to those of the closely related N. rossii, but may be readily distinguished from them by the black blotch on each pectoral fin. Pogo- thenia antarctica is known only from the unique holotype, 68 mm. in length to base of caudal, from Echo Canyon, 16 miles south-west of Little America (78° 45' S, 165° 00' W). Several of these fish were seen alive in a seal hole — an opening in a water filled crevasse, about 20 feet deep. Their colour in life was seen to be "an iridescent blue like the general color of the crevasse, some having patches of coral pink near the head". It is possible that these were young examples of N. coriiceps, and it may be noted that in the young specimen obtained from a bird's nest in the South Shetlands by the Discovery Committee the tubules of the upper lateral line are somewhat indistinct. Fig. 13. Young of A, Notothenia coriiceps; B, A^. rossii. x i. NOTOTHENIIDAE 25 I have examined a poorly preserved specimen in the Paris Museum (No. 05 — 540), 198 mm. in length (without caudal). This bears a label "Orcades du Sud — Lahille". Notothenia rossii, Richardson. Notothenia rossii, Richardson, 1844, Zool. 'Erebus' and 'Terror' (Fish.), p. 9, pi. v, figs, i, 2; Giinther, i860, Cat. Fish., u, p. 263; Regan, 1913, t.c, pp. 240, 276; 1916, Ann. Mag. Nat. Hist., (8) xviii, p. 378; Norman, 1937, Rep. B.A.N. Z. Antarct. Res. Exped., Ser. B, I, p. 64. Macronotothen rossii, Gill, 1861, Proc. Acad. Nat. Sci. Philad., 1861, p. 521. Notothenia marmorata, Fischer, 1885, Jahrb. Hamburg. Wiss. Anst., 11, p. 53. Notothenia macrocephala subsp. marmorata, Lonnberg, 1905, Wiss. Ergebn. Schwed. Siidpol.- Exped., V (6), p. 34; 1906, K. Svensk. Vet.-Akad. Handl., xl (5), p. 94. Notothenia coriiceps var. macquariensis, Waite, 191 6, Austral. Antarct. Exped. Sci. Rep., Ser. C, ni (i). Fishes, p. 64, pi. v, fig. 3, text-fig. 16. St. MS 6. 14. ii. 25. East Cumberland Bay, South Georgia. Small beam trawl, 30 m. : 2 specimens, 115, 133 mm. St. MS 10. 14. ii. 25. East Cumberland Bay, South Georgia. Small beam trawl, 26-18 m.: 5 specimens, 65-68 mm. St. MS 25. 13. iv. 25. East Cumberland Bay, South Georgia. Small beam trawl, 36 m.: 6 specimens, 65-70 mm. St. MS 63. 24. ii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 23 m.: i specimen, 68 mm. St. MS 71. 9. iii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 110-60 m.: 2 specimens, 60, 69 mm. St. MS 74. 17. iii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 22-40 m.: I specimen, 68 mm. St. — . 2. iv. 26. King Edward's Cove, South Georgia. Hand-line, 5 m.: i specimen, 330 mm. St. 45. 6. iv. 26. 27 miles S 85° E of Jason Light, South Georgia. Large otter trawl, 270- 238 m.: I specimen, 420 mm. St. — . 6.1.27. Stromness Harbour, South Georgia. Hand-line, 9-10 m.: 2 specimens, 295, 355 mm. St. 145. 7. i. 27. Stromness Harbour, South Georgia. Small beam trawl, 26-35 m.: 3 specimens, 103-250 mm. St. — . ii.i. 27. Prince Olaf Harbour, South Georgia. Hand-line, 5 m.: 2 specimens, 195, 320 mm. St. WS 65. 22. i. 27. Undine Harbour (North), South Georgia. Hand-line, 5 m. : i specimen, 290 mm. St. 174. 28. ii.-2. iii. 27. Deception Island, South Shetlands. Large fish trap, 5-10 m.: i specimen, 245 mm. St. WS III. 25. V. 27. 53° 39' 00" S, 35 '34' 00" W. 70 cm. tow-net, horizontal, 0-5 m.: I specimen, 65 mm. St. — . 27. iii. 28. South Georgia, i specimen, 56 mm., picked up on whaling plan. St. — . 1-4. iv. 29. Government Jetty, Grytviken, South Georgia. 0-2 m.: 8 specimens, 62- 70 mm. St. WS 426. 15. iv. 29. 53° 34' 00" S, 40" 10' 00" W. Washed on board through scupper: I specimen, 60 mm. St. — . 25. iii. 30. Grytviken, South Georgia. Hand-line, 6-10 m.: 22 specimens, 60-80 mm., taken from stomach of Notothenia rossii. St. — . 30. iii. 30. Stromness Harbour, South Georgia. Hand-net, surface: 8 specimens, 65- 73 mm. 26 DISCOVERY REPORTS St. WS 534. 22-23. i- 31- 54° ^iV S, 35° 39' W. Hand-line (?), 5 m. (?): i specimen, 695 mm. St. — . 28. iii. 31. Jetty, South Georgia. Hand-line: 4 specimens, 142-168 mm. In addition to the above, the following specimens were obtained by S.S. ' Sitka': 7. i. 27. 30 miles NE of Cape Saunders, South Georgia. Hand-line, 1-2 m. : 2 specimens, 540, 590 mm., caught in a patch of "krill". Fig. 14. Dorsal view of head of A, Notothenia coriiceps; B, N. rossii. x |. Depth of body 3I to 4I in the length, length of head 3^ to 3I. Snout longer than eye (except in young), diameter of which is 3! (young) to 6| in length of head; interorbital width 2| (large specimens) to about 3! . Jaws equal anteriorly or lower a little projecting; maxillary extending to below anterior part or middle of pupil, sometimes a little beyond ; scales on upper part of cheek and operculum and on temporal region ; upper surface of head papillose; 11 to 13 gill-rakers on lower part of anterior arch. Scales on body cycloid ; 55 to 62 in a longitudinal series from above base of pectoral to caudal ; 45 to 52 in upper lateral line, which ends below posterior part of dorsal; 10 to 18 in lower lateral line. Dorsal IV-VII, 32-36. Anal 26-30. Pectoral with 21 to 23 rays, I to I length of head, extending to level of vent or beyond (not as far in large specimens) ; pelvics much shorter. Caudal subtruncate or a little rounded in adult and half-grown specimens, emarginate in young; caudal peduncle as long as deep or a little deeper than long. In young the back is dark, the greater part of the sides and the lower parts silvery ; the body generally becomes more or less marbled with age, the markings sometimes tending to form irregular longitudinal stripes ; head in adult and half-grown specimens often with dark spots; dorsal fins with 2 or 3 series of dark spots, usually united to form irregular vermiculations or longitudinal bands; anal usually with a broad dusky band, and with a pale edge ; other fins varying from pale yellow to more or less dusky, the caudal sometimes with a dark band. Hab. South Shetlands; South Orkneys; South Georgia; Kerguelen; Macquarie Island. The type of the species was a large stuffed specimen, 850 mm. in total length. The actual locality is unknown, but, as suggested by Regan, it was quite likely taken at Kerguelen. Richardson describes the first dorsal fin as "comparatively low. . .sup- ported by bluntish spines, destitute of flexibility, or of filamentous tips". Such a NOTOTHENIIDAE 27 spinous dorsal is often found in large Nototheniids, and has no taxonomic signi- ficance. The silvery young of this species are said to have the back "lustrous blue" in life, and the fins quite transparent. They are found in shallow water, and are frequently taken at or near the surface. The adults and half-grown individuals present some variation in coloration, this being apparently dependent to some extent upon the depth of water in which they are living. According to Lonnberg, "the large specimens caught in the open sea, lived near the surface so that they easily could be seen swimming hither and thither". One large specimen was captured by the 'Discovery', however, at a depth of from 238 to 270 metres. Notothenia macrocephala, Giinther. For a full synonymy and description of this species see Norman, 1937, Discovery Rep., xvi, p. 88, fig. 43. Hab. Patagonia ; Falkland Islands ; Straits of Magellan ; coast of Chile ; Kerguelen ; New Zealand ; Auckland Island ; Campbell Island ; Macquarie Island. Peters (1876, Monatsber. Akad. Berlin, p. 837) described this species from Kerguelen as Notothenia antarctica, and there are two very small specimens from the same locality taken by the ' Challenger ', which have been referred to this species by Regan (see Waite, 1916, Austral. Antarct. Exped. Set. Rep., Ser. C, m (i). Fishes, p. 69, footnote). I have given a full description and figure in my previous report. Notothenia colbecki, Boulenger. Notothenia colbecki, Boulenger, 1902, Rep. Coll. Nat. Hist. 'Southern Cross', p. 185, pi. xvi; Waite, 1909, Subantarctic Isl. N. Zealand, Pisces, p. 594; Regan, 1913, t.c, p. 278; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, ni (i), Fishes, p. 70. Hab. Auckland Island ; Campbell Island ; Macquarie Island. The types of this species, up to 385 mm. in total length, are from Campbell Island, Waite has recorded a single large individual, 685 mm in total length, caught with the hook amongst the kelp at Macquarie Island. Genus TREMATOMUS Trematomus, Boulenger, 1902, Rep. Coll. Nat. Hist. 'Southern Cross', p. 177; Regan, 1913, Trans. R. Soc. Edinburgh, xlix, p. 258. Type Trematomus newnesi, Boulenger. Differs from Notothenia only in that the hypercoracoid encloses its foramen. Vertebrae 52 to 56(17-21+32-35). Coasts of the Antarctic Continent ; South Shetlands ; South Orkneys ; South Georgia. Key to the Species I. Upper surface of head naked. A. Cheeks and opercles fully scaled. 1. Interorbital width 3^ to 5 in head; dorsal V-VIII, 32-38; anal 32-36 i. newnesi. 2. Interorbital width 8 or 9 in head; dorsal IV, 37; anal 32-35 ... ... 2. nicolai. 4-2 28 DISCOVERY REPORTS B. Cheeks and opercles scaly above, naked below. 1. Dorsal V-VI, 34-37; anal 31-33 3- borchgrevinki. 2. Dorsal IV-V, 30-33; anal 29-30 4. brachysoma. II. Occiput scaly; cheeks and opercles fully scaled. A. Interorbital region naked or incompletely scaled. 1. Eye 3 to 4i in head, which is 3$ to 4 in length (without caudal); interorbital region naked or with a median series of scales; 60 to 75 scales in a longitudinal series 5 . bernacchii. 2. Eye (3! young) 4I to 4! in head, which is 3I to 3 J in length (without caudal); interorbital region with 2 or 3 series of scales in the middle (naked in young); 55 to 59 scales in a longitudinal series 6. vicarius. B. Interorbital region fully scaled. 1. Snout and lower jaw naked; praeorbital naked (except in T. scotti). a. Upper lateral line with at least 30 tubular scales, extending well beyond ex- tremity of pectoral; praeorbital naked. a. 60 to 75 scales in a longitudinal series; 36 to 46 in upper lateral line; pectoral with 27 to 29 rays. * Dorsal V-VII, 36-41 ; anal 33-36 7. hmisoni. ** Dorsal V-VI, 31-33; anal 31-33 8. loennbergii. ;8. 52 to 56 scales in a longitudinal series; 30 to 39 in upper lateral line; pectoral with 24 or 25 rays; dorsal I V-VI, 32-36. * Dorsal spines flexible; eye 3 J to 3I in head 9. penneUii. ** Dorsal spines pungent; eye 2| to 3 J in head 10. centronotus. b. Upper lateral line with 10 to 23 tubular scales, rarely extending beyond ex- tremity of pectoral fin; praeorbital usually partly or entirely scaled 11. scotti. 2. Snout, lower jaw, and praeorbital scaled. a. Vent much nearer to tip of snout than to end of anal fin; 26 to 38 tubular scales in lower lateral line; dorsal VI- VII, 31-33; anal 34-36 ... 12. lepidorhimts . b. Vent nearly equidistant from tip of snout and end of anal fin; 10 to 22 tubular scales in lower lateral line; dorsal VI-VII, 34-37; anal 32-34 13. eulepidotus. Trematomus newnesi, Boulenger. Trematomtis newnesi, Boulenger, 1902, Rep. Coll. Nat. Hist. 'Southern Cross', p. 177, pi. xi; Regan, 1913, Trans. R. Soc. Edinburgh, xlix, pp. 239, 259; Waite, 1916, Austral. Antarct. Exped. Set. Rep., Ser. C, iii (i). Fishes, p. 18; Norman, 1937, Rep. B.A.N. Z. Antarct. Res. Exped., Ser. B, i, p. 69. Notothenia cyaneobrancha, Vaillant, 1906, Bull. Mus. Paris, 1906, No. 3, p. 139; 1906, Exped. Antarct. Frang. (1903-1905), Poiss., p. 26. Notothenia microlepidota, Vaillant, 1906, t.c, p. 139; 1906, i.e., p. 35. Notothenia hodgsoni, Boulenger, 1907, Nat. Antarct. Exped. Nat. Hist., 11, Fish., p. 2, pi. i, fig. 2. St. — . 8. ii. 24. Deception Island, South Shetlands. 3 specimens, 98-110 mm., from stomach of Shag; collected by Mr J. E. Hamilton. St. 163. 17. ii. 27. Paul Harbour, Signy Island, South Orkneys. Small beam trawl, 18-27 m.: I specimen, 75 mm. St. 173. 28. ii. 27. Port Foster, Deception Island, South Shetlands. Small beam trawl, 5-60 m.: 7 specimens, 66-80 mm. St. 179. 10. iii. 27. Melchior Island, Schollaert Channel, Palmer Archipelago. Small dredge, 4-10 m.: 5 specimens, 30-70 mm. NOTOTHENIIDAE 29 St. 184. 15-16. iii. 27. Fournier Bay, Anvers Island, Palmer Archipelago. Large fish trap, 36 m.: I specimen, 87 mm. Fig. 15. Trematomus newnesi. A, Dorsal view of head of T. newnesi; B, The same of T. nicolai. x J. Depth of body 4 to 5^ in the length, length of head 3I- to 4^. Snout rather shorter than eye, diameter of which is 3 to 4^ in length of head; interorbital width 2I to 5. Jaws equal anteriorly or lower a little projecting ; maxillary extending to below anterior part or middle of eye (young) or beyond (adult) ; upper surface of head naked, cheeks and opercles scaly; 15 to 20 gill-rakers on lower part of anterior arch. Scales on body rather feebly ctenoid ; 68 to 86 in a longitudinal series from above base of pectoral to caudal; 40 to 52 in upper lateral line, which ends below posterior rays of dorsal; 3 to 19 in lower lateral line. Dorsal V-VIII, 32-38. Anal 32-36. Pectoral with 24 or 25 rays, | length of head or more, longer than pelvics, which reach the vent in young but not in adult. Caudal truncate (adult) or a little emarginate (young) ; caudal peduncle about as long as deep. Brownish, usually spotted or marbled, or with irregular cross- bars ; spinous dorsal blackish ; other fins dusky, often with small dark spots. Hab. Graham Land; South Shetlands; South Orkneys; Mac-Robertson Land; Queen Mary Land ; Adelie Land ; South Victoria Land. The numerous types of this species, ranging in length from 90 to 190 mm., are from Duke of York Island, near Cape Adare, 6 to 10 metres, and from Cape Adare, 8 to 16 metres ; those of A'^. hodgsoni, measuring up to 160 mm. in total length, are from the 'Discovery' Winter Quarters, Ross Island. This species seems to occur mainly in shallow water. I have examined 7 specimens in the Paris Museum (No. 06 — 119), 70 to 120 mm. in total length, identified by Vaillant as Notothenia cyaneobrancha. Specimens identified by the same author as A^. microlepidota cannot now be found. The British Graham Land Expedition, 1934-7, obtained 4 specimens (60 to 92 mm.), from approximately 65° S, 65° W, at a depth of about 12 metres. Trematomus nicolai (Boulenger). Notothenia nicolai, Boulenger, 1902, Rep. Coll. Nat. Hist. ^Southern Cross', p. 184, pi. xv. Trematomus nicolai, Regan, 1913, t.c, p. 260; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, in (i), Fishes, p. 18. 30 DISCOVERY REPORTS Depth of body nearly 4 in the length, length of head 3 4 to 3I. Snout shorter than eye, diameter of which is 2| (young) to 3^ in length of head; interorbital width 8 to 9. Lower jaw a little projecting; maxillary extending to below anterior J or ^ of eye; upper surface of head naked; cheeks and opercles scaly; 11 or 12 gill-rakers on lower part of anterior arch. Scales on body ctenoid; 58 to 62 in a longitudinal series from above base of pectoral to caudal; 39 to 43 in upper lateral line, which ends below posterior rays of dorsal, 8 to 18 in lower lateral hne. Dorsal IV-V, 37. Anal 32-35. Pectoral with 28 to 30 rays, | to | length of head, somewhat longer than pelvics, which reach vent or origin of anal. Caudal rounded; caudal peduncle nearly as long as deep. Brownish, with indistinct dark cross-bars, and sometimes with some scattered small dark spots ; fins dusky. Hab. Queen Mary Land ; South Victoria Land. The types of the species, 3 specimens 150 to 250 mm. in total length, are from Cape Adare, 10 to 16 metres, and from Duke of York Island, near Cape Adare, 8 metres. The two examples recorded by Waite from Queen Mary Land, each 115 mm. in total length, were from a depth of 643 metres. Trematomus borchgrevinki, Boulenger. Trematomus borchgrevinki, Boulenger, 1902, Rep. Coll. Nat. Hist. 'Southern Cross', p. 179, pi. xii; 1907, Nat. Antarct. Exped. Nat. Hist., n. Fish., p. i; Pappenheim, 1912, Deutsche Siidpol.-Exped., xni, Zool. v, p. 171 ; Roule, Angel and Despax, 1913, Deux. Exped. Antarct. Franf. (1908-1910), Poiss., p. 6; Regan, 1913, t.c, pp. 239, 260. Trematomus borchgrevinski, Roule and Despax, 191 1, Bull. Mus. Paris, 191 1, No. 5, p. 279. Fig. 16. Trematomus borchgrevinki. x |. Depth of body 4 to 5 in the length, length of head 3* to 4^. Snout as long as or a little longer than eye, diameter of which is 4 to 5 in length of head ; interorbital width 3 to 4. Jaws equal anteriorly or lower a little projecting ; maxillary extending to below anterior i of eye ; upper surface of head naked ; upper parts of cheeks and opercles scaly; 16 to 19 gill-rakers on lower part of anterior arch. Scales on body ctenoid; 78 to 96 in a longitudinal series from above base of pectoral to caudal ; lateral lines without or with only a few tubular scales. Dorsal V-VI, 34-37. Anal 31-33. Pectoral with NOTOTHENIIDAE 31 23 rays, | to | length of head, longer than pelvics, which rarely reach the vent. Caudal rounded or subtruncate ; caudal peduncle as long as deep or deeper than long. Yellowish, with dark spots or irregular cross-bars; dorsal and caudal sometimes with series of spots. Hab. Graham Land ; South Orkneys ; Wilhelm Land ; South Victoria Land. The types of the species, 120 to 275 mm. in total length, are from Cape Adare and from Duke of York Island, near Cape Adare. The species is known only from shallow water. The British Graham Land Expedition obtained 2 specimens (108 and 130 mm.) at the surface, and 7 specimens (c. 80 to 135 mm.) from stomachs of Pygoscelis, from approximately 65° S, 65° W. Trematomus brachysoma, Pappenheim. Trematomus brachysoma, Pappenheim, 1912, Deutsche SildpoL-Exped., xni, Zool. V, p. 172; Regan, 1913, t.c, p. 261; Regan, 1914, Rep. Brit. Antarct. {'Terra Nova') Exped. 1910, Zool. I (i), pp. 2, 3, pi. ii, fig. 3. Depth of body 4^^ to 4I in the length, length of head 3 to 3I. Snout about as long as eye, diameter of which is 3! to a little more than 4 in length of head; interorbital width 4 to 5|. Jaws equal anteriorly; maxillary extending to below anterior part or middle of eye ; upper surface of head naked ; upper parts of cheeks and opercles scaly ; 15 to 19 gill-rakers on lower part of anterior arch. Scales on body cycloid or feebly ctenoid; 55 to 60 in a longitudinal series from above base of pectoral to base of caudal; lateral lines without or with only a few tubular scales. Dorsal IV-V, 30-33. Anal 26-30. Pectoral with 23 or 24 rays, about | length of head, much longer than pelvics, which reach the vent or not quite as far. Caudal rounded ; caudal peduncle a little deeper than long. Yellowish brown, with a series of 6 dark spots on back below the dorsal fins, and 5 below them at level of base of pectoral ; a round black spot on spinous dorsal ; soft dorsal, caudal and sometimes the anal with series of small round dark spots forming irregular bars. Fig. 17. Trematomus brachysoma. x Hab. Wilhelm Land ; South Victoria Land. The above description is based upon a co-type of the species, 126 mm. in total length, taken at the surface at the Winter Quarters of the ' Gauss ' ; and another example, 33 DISCOVERY REPORTS 170 mm. in total length, stranded on an ice floe in 67° 24' S, 177" 34' W, taken by the 'Terra Nova'. This species is closely related to T. borchgrevinki, but may be readily distinguished by the larger scales and the lower number of dorsal and anal rays. Trematomus bernacchii, Boulenger. Trematomiis bernacchii, Boulenger, 1902, Rep. Coll. Nat. Hist. 'Southern Cross', p. 181, pi. xiv; 1907, Nat. Antarct. Exped. Nat. Hist., n. Fish., p. i ; Waite, 191 1, Brit. Antarct. Exped., 1907-1909, II, Biol., II, p. 12; Roule and Despax, 1911, Bull. Mus. Paris, 1911, No. 5, p. 279; Roule, Angel and Despax, 1913, Deux. Exped. Antarct. Franf. (1908-1910), Poiss., p. 7; Regan, 1913, t.c, pp. 239, 262, fig. 6 A; Regan, 1914, t.c, p. 3, pi. i, fig. i; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, III (i). Fishes, p. 19. Notothenia marionensis, Vaillant, 1906, Bull. Mus. Paris, 1906, No. 3, p. 139. Notothenia elegans, Vaillant, 1906, Exped. Antarct. Fratif. (1903-1905), Poiss., p. 28. Notothenia longipes (part), Boulenger, 1907, t.c, p. i. Notothenia nicolai, Boulenger, 1907, t.c, p. 2. Notothenia nicolai, Roule and Despax, 1911, t.c, p. 278; Roule, Angel and Despax, 19x3, t.c, p. 5. St. 178. 9-10. iii. 27. Melchior Harbour, Schollaert Channel, Palmer Archipelago. Large fish trap, 17 m.: I specimen, 137 mm. St. 189. 23-24. iii. 27. Port Lockroy, Wiencke Island, Palmer Archipelago. Large fish trap, 70 m.: I specimen, 147 mm. Depth of body 3 to 4^ in the length, length of head 3^ to 4. Snout as long as or rather shorter than eye, diameter of which is 3 to 4I in length of head ; interorbital width 5 to 9. Jaws equal anteriorly or lower a little projecting; maxillary extending to below anterior part or middle of eye ; cheeks, opercles and occiput scaly ; interorbital region naked or with a median series of scales; 13 to 15 gill-rakers on lower part of anterior arch. Scales on body ctenoid ; 60 to 75 in a longitudinal series from above base of pectoral fin to caudal ; 30 to 42 in upper lateral line ; lower lateral line usually without tubular scales. Dorsal IV-VI, 34-38. Anal 31-35. Pectoral with 23 to 25 rays, about I length of head, about as long as pelvics, which just reach anal in young but not in adult. Caudal rounded; caudal peduncle deeper than long. Brownish; large dark spots in 2 or 3 alternating series ; upper part of spinous dorsal blackish ; other fins often more or less dusky. Hub. Graham Land; South Shetlands; South Orkneys; Mac-Robertson Land; Queen Mary Land; Adelie Land; South Victoria Land. The types of the species, 120 to 250 mm. in total length, are from Cape Adare, 10 to 16 metres, and from Duke of York Island, near Cape Adare, 6 to 8 metres. Like the previous species, T. bernacchii usually occurs in rather shallow water, but the ' Aurora ' obtained one example from Commonwealth Bay at a depth of 635 to 732 metres. Three young specimens (45-47 mm.), from Deception Island, South Shetlands, collected by Mr A. G. Bennett in February 1918, appear to be referable to this species. I have examined the specimen in the Paris Museum (No. 06 — 126), 118 mm. in total length, identified by Vaillant as Notothenia elegans. The fish (No, ?) from Peterman NOTOTHENIIDAE 33 Island, at a depth of i metre, identified by Roule, Angel and Despax as A^. nicolai, is 135 mm. in total length, and is in very poor condition. Scales are apparently present on the occiput, but not on the interorbital region, and I have little doubt that the specimen should be referred to this species. The British Graham Land Expedition obtained 9 specimens (133 to 230 mm.) from fish traps at approximately 65° S, 65° W, and one specimen (95 mm.) thrown ashore at Barry Island (68° S, 67° W). Fig. 18. Trematomus bernacchii. A, Dorsal view of head of T. bernacchii; B, The same of T. vicarius. x |. Trematomus vicarius, Lonnberg. Trematomus bernacchii subsp. vicarius, Lonnberg, 1905, Wiss. Ergebn. Schwed. Siidpol.-Exped., V (6), p. 26. ? Notothenia dubia, Lonnberg, 1905, t.c, p. 28, pi. iii, fig. 9. Trematomus vicarius, Regan, 191 3, t.c, p. 261, fig. 6 B. Trematomus dubius, Regan, 1914, Rep. Brit. Antarct. {'Terra Nova') Exped., 1910, Zool. i (i), p. 2. St. 39. 25. iii. 26. East Cumberland Bay, South Georgia. Large otter trawl, 179-235 m.: i specimen, 180 mm. Very similar to the preceding species, but length of head 3 J to 3^ in that of fish (without caudal); diameter of eye (3I young) 4! to 4! in length of head; interorbital width about 5 (6^ in young) ; interorbital region naked in young, but with 2 or 3 series of scales in the middle in adult and half-grown specimens. 55 to 59 scales in a longi- tudinal series, (30) 34 to 37 in upper lateral line. Dorsal V, 33-38. Anal 31-32. Pectoral I to I length of head, about as long as pelvics, which just reach the anal. Hab. South Georgia. The type of T. vicarius, 240 mm. in length without the caudal fin, was taken outside Boiler Harbour, Cumberland Bay, South Georgia, at a depth of 30 metres. This species is very close to T. bernacchii, and may eventually prove to be no more than a local race. Notothenia dubia was based upon 3 specimens, 39 to 45 mm. in length without the caudal fin, from the same locality, at a depth of 20 metres ; one of these types was examined by Regan. There can, I think, be little doubt that this form is the young of T. vicarius. Regan has compared the example of 50 mm. with one of T. bernacchii of the same size, and finds that it has a smaller eye, broader interorbital region, and larger scales — exactly the characters used to separate T. bernacchii from T. vicarius. 34 DISCOVERY REPORTS The British Graham Land Expedition obtained one specimen (69 mm.) from Stromness Bay, South Georgia, where it was found to be food of the Tern. Trematomus hansoni, Boulenger. Trematomus hansoni, Boulenger, 1902, Rep. Coll. Nat. Hist. 'Southern Cross', p. 180, pi. xiii; Roule and Despax, 1911, Bull. Mus. Paris, 1911, No. 5, p. 279; Waite, 1911, Brit. Antarct. Exped., 1907-1909, n, Biol., n, p. 13; Roule, Angel and Despax, 1913, Deux. Exped. Antarct. Franp. (1908-1910), Poiss., p. 6; Regan, 1913, t.c, pp. 239, 262; 1914, Rep. Brit. Antarct. {'Terra Nova') Exped., 1910, Zool. i (i), p. 3, pi. i, fig. 2. Trematomus hansoni subsp. georgianus, Lonnberg, 1905, Wiss. Ergebn. Schwed. Sudpol.-Exped., V (6), p. 25, pi. V, fig. 17. Notothenia sirna, Vaillant, 1906, Bull. Mus. Paris, 1906, No. 3, p. 139; 1906, Exped. Antarct. Franf. (1903-1905), Poiss., p. 23. Notothenia lepidorhinus (part), Pappenheim, 1912, Deutsche Siidpol. -Exped. xiii, Zool. v, p. 169. St. MS 68. 2. iii. 26. East Cumberland Bay, South Georgia. Large rectangularnet, 220-247 m. : 4 specimens, 190-285 mm. St. 30. 16. iii. 26. West Cumberland Bay, South Georgia. Large dredge, 251 m.: i specimen, 170 mm. St. 39. 25. iii. 26. East Cumberland Bay, South Georgia. Large otter trawl, 179-235 m.: 22 specimens, 140-320 mm. St. 45. 6. iv. 26. 27 miles S 85° E of Jason Light, South Georgia. Large otter trawl, 238-. 270 m.: 2 specimens, 180, 205 mm. St. 123. 15. xii. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 230- 250 m. : 2 specimens, 240, 295 mm. St. WS32. 21. xii. 26. Mouth of Drygalski Fjord, South Georgia. Small beam trawl, 91- 225 m. : I specimen, 172 mm. St. 148. 9. i. 27. OtF Cape Saunders, South Georgia. Large otter trawl, 132-148 m.: i specimen, 250 mm. St. WS 62. 19. i. 27. Wilson Harbour, South Georgia. Small beam trawl, 26-83 m.: i specimen, 79 mm. St. 160. 7. ii. 27. Near Shag Rocks, South Georgia. Large dredge, 177 m.: i specimen, 138 mm. St. 167. 20. ii. 27. Off Signy Island, South Orkneys. Net (7 mm. mesh) attached to back of trawl, 244-344 m.: 2 specimens, 142, 150 mm. St. 178. 9-10. iii. 27. Melchior Harbour, Schollaert Channel, Palmer Archipelago. Large fish trap, 17 m.: i specimen, 135 mm. The following young specimens may also belong here : St. 42. I. iv. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 120-204 m. : 9 specimens, 41-44 mm. St. 45. 6. iv. 26. 27 miles S 85° E of Jason Light, South Georgia. Large otter trawl, 238- 270 m. : 6 specimens, 40-43 mm. Depth of body 3^ to 4! in the length, length of head 3i to 4. Snout as long as or a little longer than eye, diameter of which is 3 § to 5 in length of head ; interorbital width 5 to b\. Jaws about equal anteriorly; maxillary extending to below anterior part or middle of eye; occiput, interorbital region, cheeks and opercles scaly; 13 to 16 gill- rakers on lower part of anterior arch. Scales on body cycloid or ctenoid ; 60 to 68 in a longitudinal series from above base of pectoral to caudal; 38 to 46 in upper lateral line; lower lateral line usually without tubular scales. Dorsal V-VII, 36-41. Anal NOTOTHENIIDAE 35 33-36. Pectoral with 27 to 29 rays, f to | length of head, longer than pelvics, which just reach the vent in young but not in adult. Caudal subtruncate; caudal peduncle about as long as deep. Brownish, with large dark spots or cross-bars; head often spotted; spinous dorsal blackish; other fins usually barred with series of dark spots. Fig. 19. Trematomus hansoni. x |. Hab. Graham Land and neighbouring islands; South Shetlands; South Orkneys; South Georgia ; Coats Land ; Wilhelm Land ; Queen Mary Land ; South Victoria Land. The types of the species, 160 to 280 mm. in total length, are from Cape Adare, 8 to 16 metres, and from Duke of York Island, near Cape Adare, 6 to 8 metres. Regan has shown that Lonnberg's supposed subspecies from South Georgia is fully identical with the typical form. The specimen of 137 mm. from Booth Wandel Island (Paris Museum, No. 06 — 99), recorded by Vaillant as Notothenia sima, appears to belong here. Trematomus loennbergii, Regan. Notothetiia nicolai {non Boulenger), Lonnberg, 1905, Wiss. Ergebn. Schwed. Siidpol.-Exped., v (6), p. 45. Notothenia longipes (part), Boulenger, 1907, Nat. Atitarct. Exped. Nat. Hist., n, Fish., p. i. Trematomus loennbergii, Regan, 1913, t.c, p. 263, pi. viii, fig. 4; 1914, Rep. Brit. Antarct. (' Terra Nova') Exped., 1910, Zool. i (i), p. 3; Waite, 1916, Austral. Antarct. Exped. Set. Rep., Ser. C, in (i). Fishes, p. 20. Depth of body 4 to 5 in the length, length of head 3 to 3I. Snout rather shorter than eye, diameter of which is 3 to 3 1 in length of head; interorbital width 6 to 10. Jaws about equal anteriorly; maxillary extending to below anterior ^ or | of eye; upper surface of head to nostrils, cheeks and opercles scaly; 10 to 13 gill-rakers on lower part of anterior arch. Scales on body more or less ctenoid ; 60 to 75 in a longitudinal series from above base of pectoral to caudal ; 34 to 46 in upper lateral line ; lower lateral line, when developed, with 10 to 15 tubular scales. Dorsal V-VII, 31-35. Anal 31-35. Pectoral with 28 or 29 rays, nearly as long as head, longer than pelvics, which reach vent or origin of anal. Caudal rounded or subtruncate; caudal peduncle as long as or longer than deep. Brownish, with irregular darker cross-bars; inside of mouth and branchial cavities blackish ; sometimes an indistinct dark blotch on spinous dorsal. 36 DISCOVERY REPORTS Hab. Graham Land; Queen Mary Land; Adelie Land; South Victoria Land. This species was originally described from 3 specimens: i (132 mm. in total length) from south-west of the Balleny Islands, 476 metres ; i (65 mm.) from Tent Island, near Ross Island ; and i (about 65 mm.) from Seymour Island, collected in tide-pools. It is closely related to T. hansoni, but may be distinguished from that species by the smaller number of dorsal and anal rays, and by the presence (usually) of tubular scales in the lower lateral line. Further, Waite has pointed out that in life this species is reddish in colour, whereas T. hansoni is greenish or greyish. As a general rule, T. loennbergii appears to live in deeper water than T. hansoni. Fig. 20. Trematomus loennbergii. x J. Trematomus pennellii, Regan. Trematomus pennellii, Regan, 1914, Ann. Mag. Nat. Hist., (8) xin, p. 12; 1914, Rep. Brit. Antarct. (' Terra Nova') Exped., 1910, Zool. i (i), p. 4, pi. iii, fig. 2; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, iii (i), Fishes, p. 20. Depth of body 4I to 5 in the length, length of head 3^ to 3J. Snout as long as or rather shorter than eye, diameter of which is 3^^ to 3^ in length of head; interorbital width 8 to 10. Jaws equal anteriorly; maxillary extending to below anterior J of eye; occiput, interorbital region, cheeks and opercles scaly; 13 to 16 gill-rakers on lower part of anterior arch. Scales on body ctenoid; 52 to 56 in a longitudinal series from above base of pectoral to caudal ; 32 to 36 in upper lateral line, which ends below posterior part of dorsal; lower lateral line without tubular scales. Dorsal IV-VI, 32-35 ; spines flexible. Anal 29-31. Pectoral with 24 or 25 rays, J or | length of head, as long as or longer than pelvics, which reach origin of anal. Caudal subtruncate ; caudal peduncle as long as deep. Olivaceous, with two or three series of large dark spots, which may unite to form irregular cross-bars ; distal part of spinous dorsal blackish. Hab. Queen Mary Land ; South Victoria Land. The types of the species, 5 specimens 100 to 140 mm. in total length, are from off Cape Adare, 82 to 94 metres. Waite has recorded 10 specimens, up to 180 mm. in length, from off Drygalski Island, 112 metres. Trematomus centronotus, Regan. Trematomus centronotus, Regan, 1914, Ann. Mag. Nat. Hist., (8) xni, p. 12; 1914, Rep. Brit. Antarct. (' Terra Nova') Exped., 1910, Zool. i (i), p. 4, pi. iii, fig. i ; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, in (i), Fishes, p. 21; Norman,. 1937, Rep. B.A.N. Z. Antarct. Res. Exped., Ser. B, i, p. 70. NOTOTHENIIDAE 37 Depth of body 4 to 5 in the length, length of head 3! to 3! . Snout shorter than eye, diameter of which is af (young) to 3I in length of head; interorbital width 10 to 12. Jaws equal anteriorly; maxillary extending to below anterior I or 3 of eye; occiput, interorbital region, cheeks and opercles scaly; 13 to 15 gill-rakers on lower part of anterior arch. Scales on body cycloid or rather feebly ctenoid ; 52 to 56 in a longitudinal series from above base of pectoral to caudal ; 30 to 39 in upper lateral line, which ends below posterior part of dorsal; lower lateral line without tubular scales. Dorsal V-VI, 32-36; spines stiff, pungent. Anal 29-32. Pectoral with 24 or 25 rays, f to | length of head, a little shorter than pelvics, which extend to origin of anal. Caudal apparently subtruncate ; caudal peduncle about as long as deep. Brownish, with large dark spots, uniting to form irregular cross-bars ; a black spot on distal part of spinous dorsal. Fig. 21. Lateral view of head of A, Trematomus pennellii; B, T. centronotus. x \. Hab. Mac-Robertson Land; Queen Mary Land; Adelie Land; South Victoria Land. The types of this species, 2 specimens 175 and 210 mm. in total length, are from McMurdo Sound, 187 to 375 metres. It is very like T. pennellii, and, apart from the stiff and pungent dorsal spines, it scarcely differs from that species except in the larger eye. T. centronotus appears to be found in rather deeper water than T. pennellii. Trematomus scotti (Boulenger). Notothenia scotti, Boulenger, 1907, Nat. Antarct. Exped. Nat. Hist., 11, Fish., p. 2, pi. i, fig. i; Roule, Angel and Despax, 1913, Deu.v. Exped. Antarct. Fratif. (1908-1910), Poiss., p. 5; Regan, 1913, Trans. R. Soc. Edinburgh, xlix, p. 271. Trematomus scotti, Regan, 1914, Rep. Brit. Antarct. {'Terra Nova') Exped., 1910, Zool. i (i), p. 4, pi. iv, fig. 2; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, in (i), Fishes, p. 21 ; Norman, 1937, t.c, p. 70. St. 181. 12. iii. 27. Schollaert Channel, Palmer Archipelago. Nets (4 mm. and 7 mm. mesh) attached to back of trawl, 160-235 "i-- 6 specimens, 95-145 mm. St. 187. 18. iii. 27. Neumayr Channel, Palmer Archipelago. Large dredge, 259 m.: i specimen, 145 mm. St. 190. 24. iii. 27. Bismarck Strait, Palmer Archipelago. Large dredge, 315 m. : i specimen, 68 mm. St. 599. 17. i. 31. 67° 08' S, 69° o6-|' W. Large dredge, 203 m.: 3 specimens, 70-125 mm. St. 1652. 23. i. 26. 75' 56-2' S, 178° 35-5' E. Rectangular dredge, with Russell net, 567 m.: I specimen, 130 mm. St. 1660. 27. i. 36. 74° 46-4' S, 178° 23-4' E. Large otter trawl, 351m.: 11 specimens, 61- 145 mm. 38 DISCOVERY REPORTS Depth of body 4 to 5 1 in the length, length of head 3 to 3I. Snout shorter than eye, diameter of which is 2| to 3! in length of head; interorbital width about 12. Jaws about equal anteriorly ; maxillary extending to below anterior j of eye ; occiput, interorbital region, cheeks and opercles scaly; praeorbital sometimes scaly, sometimes incompletely scaled or nearly naked; 10 to 13 gill-rakers on lower part of anterior arch. Scales on body ctenoid ; 46 to 54 in a longitudinal series from above base of pectoral to caudal ; 10 to 23 in upper lateral line, which rarely extends beyond extremity of pectoral fin; lower lateral line without tubular scales. Dorsal IV-VI, 31-36. Anal 29-33. Pectoral with 20 to 22 rays, f to f head, about as long as pelvics, which reach vent or anal fin. Caudal rounded ; caudal peduncle as long as deep or rather deeper than long. Brownish, with dark spots or irregular cross-bars ; spinous dorsal with a deep black blotch ; soft dorsal and anal with a blackish band which is broadest and most intense posteriorly, that on the dorsal sometimes reduced to a deep black blotch on posterior part of fin ; caudal, pectorals and pelvics often barred. Hab. Graham Land and neighbouring islands; South Shetlands; Enderby Land; Mac-Robertson Land ; Princess Elizabeth Land ; Queen Mary Land ; South Victoria Land; near King Edward Land. Fig. 22. Trematomus scotti. x |. The type of the species, no mm. in total length, is from off the Ross Barrier at a depth of 562 metres. The above description is based upon a large number of specimens, up to 180 mm. in total length, obtained by the 'Terra Nova', 'Discovery' and the B.A.N.Z. Antarctic Research Expedition. The species seems to occur mainly in fairly deep water. Trematomus lepidorhinus (Pappenheim). Notothenia lepidorhinus (part), Pappenheim, 191 1, Ges. nat. Freiinde Berlin, No. 8, p. 382; 1912, Deutsche SUdpol.-Exped., xin, Zool. v, p. 169, pi. ix, fig. i, pi. x, fig. i ; Regan, 1913, t.c, p. 270; Norman, 1937, t.c, p. 71. Trematomus lepidorhinus, Regan, 1914, Rep. Brit. Antarct. {'Terra Nova'') Exped., 1910, Zool. I (i), p. 5, pi. iv, fig. I. St. 1660. 27. i. 36. 74° 46-4' S, 178° 23-4' E. Large otter trawl, 351 m.: i specimen, 280 mm. Depth of body 4 to 4I in the length, length of head 3I to 3I. Vent much nearer to tip of snout than to end of anal fin. Snout nearly as long as eye, diameter of which is NOTOTHENIIDAE 39 3 to 3 § in length of head; interorbital width about 6. Mouth nearly horizontal; jaws about equal anteriorly; maxillary extending to below anterior edge of pupil; upper surface and sides of head, including snout and praeorbital, scaly; 14 to 16 gill-rakers on lower part of anterior arch. Scales on body ctenoid ; 72 to 80 in a longitudinal series from above base of pectoral to caudal ; 43 to 56 in upper lateral Une, which ends behind dorsal or below its posterior rays; 26 to 38 in lower lateral hne. Dorsal VI-VII, 31-33. Anal 34-36. Pectoral with 27 to 29 rays, as long as or nearly as long as head, longer than pelvics, which reach vent or anterior rays of anal. Caudal rounded or sub- (^ /^v^^^^ truncate; caudal peduncle somewhat •-<''"'-^^^^-^^^^^^^ deeper than long. Brownish, with ir- ^:^ \ ^^^^^^ .V.V.::: )^^3 regular and somewhat indistinct darker ^"^~— =^^^J^fc^^^j^^ssv^^^^^^ ^^^5=5=^ cross-bars; spinous dorsal blackish; soft ^^^ dorsal with dark oblique stripes. The ,, , , , _. Fie. 23. Trematomus lepidorhinus. x 1. specimen collected by the Discovery Committee has the inside of the mouth and the branchial cavities blackish. Hab. Mac-Robertson Land ; Wilhelm Land ; South Victoria Land. The types, up to 240 mm. in total length, were taken at the Winter Quarters of the 'Gauss', at a depth of 385 metres. There are 3 specimens, 135 to 150mm. in total length, in the British Museum collection, taken by the 'Terra Nova' from off new land south of the Balleny Islands at a depth of 375 metres. Trematomus eulepidotus, Regan. Trematomus eulepidotus, Regan, 1914, Ann. Mag. Nat. Hist., (8) xni, p. 12; 1914, Rep. Brit. Antarct. (' Terra Nova') Exped., 1910, Zool. i (i), p. 5, pi. iv, fig. 3 ; Waite, 191 6, Austral. Antarct. Exped. Sci. Rep., Sen C, in (i). Fishes, p. 22; Norman, 1937, t.c, p. 71. St. 167. 20. ii. 27. Off Signy Island, South Orkneys. Net (7 mm. mesh) attached to back of trawl, 244-344 "!•• 4 specimens, 90-190 mm. St. 181. 12. iii. 27. Schollaert Channel, Palmer Archipelago. Large otter trawl, 160-335 m.: I specimen, 163 mm. St. 1660. 27. i. 36. 74° 46-4' S, 178° 23-4' E. Large otter trawl, 351 m.: i specimen, 255 mm. Depth of body 3I to 4I in the length, length of head 3^ to 3f. Vent nearly equi- distant from tip of snout and end of anal fin. Snout shorter than eye, diameter of which is 2,1 to 3I in length of head; interorbital width 4I to 5 J. Mouth oblique; jaws about equal anteriorly ; maxillary extending to below anterior \ of eye ; head covered with scales, only the lips naked; 12 to 15 gill-rakers on lower part of anterior arch. Scales on body ctenoid ; 67 to 76 in a longitudinal series from above base of pectoral to caudal ; 40 to 48 in upper lateral line, which generally ends below posterior rays of dorsal; 10 to 22 in lower lateral line. Dorsal VI-VII, 34-37. Anal 32-34. Pectoral with 27 or 28 rays, | to | length of head, longer than pelvics, which reach vent or origin of anal. Caudal subtruncate; caudal peduncle as long as or nearly as long as deep. Pale brownish; sides of body with dark spots connected to form a network; spinous dorsal more or less dusky or blackish distally ; soft dorsal with oblique dark stripes. 40 DISCOVERY REPORTS Hab. Palmer Archipelago, Graham Land; South Orkneys; Enderby Land; Mac- Robertson Land; Queen Mary Land; South Victoria Land. Fig. 24. Trematomus eulepidotns. x |. The types of this species, 140 and 165 mm. in total length, are from the entrance to McMurdo Sound, 300 metres, and from near Inaccessible Island, 416 to 452 metres. Trematomus eulepidotns is closely related to T. lepidorhinus, but differs in the shorter snout, broader interorbital region, more oblique mouth, shorter tail, more numerous dorsal rays, and much shorter lower lateral line. Genus DISSOSTICHUS Dissostichus, Smitt, 1898, Bih. Sv. Vet.-Akad. Handl., xxiv, iv, No. 5, P- 3; Regan, 1913, Trans. R. Soc. Edinburgh, xlix, p. 279; Norman, 1937, Discovery Rep., xvi, p. 91. Type Dissostichus eleginoides, Smitt. Differs from Trematomus in the longer snout and strongly projecting lower jaw; teeth in the upper jaw biserial, those of the outer row enlarged, spaced, canine-like; a group of stronger teeth on each praemaxillary ; teeth of lower jaw uniserial, spaced, canine-like. Two species ; one from the Patagonian region, one from the Antarctic Zone. Coasts of the Antarctic Continent and northwards to the Patagonian region and the coast of Argentina. Dissostichus mawsoni, Norman. Dissostichus eleginoides {nan Smitt), Vaillant, 1906, Bull. Mus. Paris, 1906, No. 3, p. 139; 1906, Exped. Antarct. Franf. (1903-1905), Poiss., p. 36. Dissostichus mawsoni, Norman, 1937, Rep. B.A.N.Z. Antarct. Res. Exped, Ser. B, i, p. 71, fig. 6. Depth of body 5* to nearly 6 in the length, length of head about 3. Snout ij to if times as long as eye, diameter of which is 4 to 5i in length of head ; interorbital width 4I to 5. Maxillary extending to below middle of eye or a little beyond; upper surface of head to below middle or anterior edges of eyes, cheeks and opercles scaly ; no elongate naked areas on upper surface of head; 12 or 13 gill-rakers on lower part of anterior arch. 135 to 150 scales in a longitudinal series from above base of pectoral to caudal; 90 to 96 in upper lateral line, which ends below posterior rays of dorsal; 35 to 38 in lower lateral line, which does not extend forward as far as level of middle of anal fin. Dorsal VIII-IX, 25-27. Anal 25-26. Pectoral with 27 to 29 rays, nearly | length of head, much longer than pelvics, which do not nearly reach vent. Caudal truncate or a little rounded; caudal peduncle longer than deep. More or less uniformly brownish, NOTOTHENIIDAE 41 or with 4 irregular and incomplete dark cross-bars and with some dark spots; spinous dorsal blackish distally, sometimes with a dark horizontal band near the base; traces of oblique dark stripes on soft dorsal and anal ; caudal pectorals and pelvics plain. Fig. 25. Dissostichus tnawsotii. x about |. Hab. Graham Land ; off Mac-Robertson Land. The type of the species, 370 mm. in total length, is from 66° 45' S, 62° 03' E, at a depth of 219 metres. The specimens collected by the 'Franfais' are from Booth Wandel Island, 20 to 30 metres. I have examined four of these in the Paris Museum (No. 06 — 140-143), 212 to 262 mm. in total length, and, although the skin has been largely rubbed off the head region, I have little doubt that they are referable to this species.^ Dissostichus mawsoni may be readily distinguished from the Patagonian D. eleginoides by the absence of the elongate naked areas on the upper surface of the head, the less completely scaled interorbital region, the somewhat smaller scales on the body, and the shorter lower lateral line. Genus PLEURAGRAMMA Plettragramma, Boulenger, 1902, Rep. Coll. Nat. Hist. 'Southern Cross', p. 187; Regan, 1913, Trans. R. Soc. Edinburgh, xlix, p. 264. Type Pleuragramma antarcticum, Boulenger. Gelidus, Whitley, 1937, Rec. Austral. Mus., xx, p. 19. Type Pleuragramma antarcticum, Bou- lenger. Body covered with very thin, cycloid scales ; 3 lateral lines ; each lateral line scale with a notched hinder margin and with a vertical row of three shallow pits. Mouth of moderate size, the lower jaw strongly projecting; jaws with bands of villiform teeth; one or two pairs of somewhat larger teeth at the praemaxillary symphysis. Snout flattened, but not spatulate. Gill-membranes narrowly united, free from the isthmus. Skeleton feebly ossified, with parapophyses developed only on the posterior praecaudal vertebrae; vertebrae 53 (19 + 34); hypercoracoid enclosing its foramen. A single species. Coasts of the Antarctic Continent. There would appear to be no justification for the introduction of a new name for this genus, since, under the International Rules of Nomenclature, Pleuragramma, Boulenger, cannot be regarded as invalidated by Pleurogrammus, Gill, Further, the '^ One of these specimens has now been received in exchange from the Paris Museum 42 DISCOVERY REPORTS meanings of the two words are quite different, Gill's genus being for a fish with 5 lateral lines on each side, Boulenger's for a fish believed to be without any lateral line. Mr Whitley has also erected a new family (Gelididae) for its reception, pointing out that Boulenger was in error in placing this fish in the family Leptoscopidae. Regan, however, in his monograph of the Nototheniiformes published in 191 3, has clearly demonstrated the close affinities of Pleuragramma and Trematomus. Pleuragramma antarcticum, Boulenger. Pleuragramma antarcticum, Boulenger, 1902, i.e., p. 187, pi. xviii; Lonnberg, 1905, Wiss. Ergehn. Schwed. Sudpol.-Exped., V (6), p. 49; Vaillant, 1906, Bull. Mus. Paris, 1906, No. 3, p. 139; 1906, Exped. Antarct. Frang. (1903-1905), Poiss., p. 48, figs. 3, 4; Boulenger, 1907, Nat. Antarct. Exped. Nat. Hist., 11, Fish., p. 4; Waite, 191 1, Brit. Antarct. Exped., 1907-1909, II, Biol., II, p. 14; Pappenheim, 1912, Deutsche Siidpol.-E.xped., xiii, Zool. v, p. 164; Regan, 1913, Trans. R. Soc. Edinburgh, xlix, p. 264; 1914, Rep. Brit. Antarct. (' Terra Nova') Exped. 1910, Zool. i (i), p. 6; 1916, ibid. Zool. i (4), p. 131, pi. ii, figs. 3-6; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, iii (i), Fishes, p. 22; Nichols and LaMonte, 1936, Amer. Mus. Novit., 839, p. i ; Norman, 1937, Rep. B.A.N.Z. Antarct. Res. Exped., Ser. B, i, p. 72. Gelidus antarcticus, Whitley, 1937, Rec Austral. Mus., xx, p. 20, fig. 5. St. 181. 12. iii. 27. Schollaert Channel, Palmer Archipelago. Nets (4 mm. and 7 mm. mesh) attached to back of trawl, 160-335 m.: i specimen, 160 mm. St. 182. 14. iii. 27. Schollaert Channel, Palmer Archipelago. Large otter trawl, 278-500 m.: I specimen, 200 mm. St. WS 506. 7. ii. 30. 70° 31' S, 81° 36' W. Picked up dead at surface: i specimen, 185 mm. St. — . 20.1.36. Bay of Whales (78° 24-85' 8, 164° 10-3' W). Picked up dead at surface: i specimen, 173 mm. Fig. 26. Pleuragramma antarcticum. x ^. A, Lateral line scales (enlarged). Depth of body 5 to 6 in the length, length of head 3^^ to 4. Snout as long as or longer than eye, diameter of which is 3 j to 4 in length of head ; interorbital width 5 to 6. Maxillary extending to below anterior h of eye ; upper surface of head naked ; cheeks and opercles scaly; 20 to 25 gill-rakers on lower part of anterior arch. 50 to 58 scales in a longitudinal series from above base of pectoral to caudal ; about 45 notched scales in upper lateral line, which ends below last rays of dorsal; about 35 in both middle and lower lateral lines. Dorsal VI-VII, 33-38. Anal 36-39. Pectoral with 20 or 21 rays, i to | length of head, as long as or longer than pelvics, which scarcely reach the NOTOTHENIIDAE 43 vent. Caudal slightly emarginate; caudal peduncle about as deep as long. Silvery, back darker ; back and sides powdered with blackish dots ; fins all pale. Hab. Circumpolar in the Glacial District (see p. 97). The types of the species, up to 165 mm. in total length, are from the ice barrier, Victoria Land, at 78° 35' S. Regan has described and figured several larval and post- larval specimens from the Ross Sea and the coast of Victoria Land, and points out that in these the eye is proportionately smaller than in the adult fish, no doubt because the latter descend to greater depths. He has shown that larval and post-larval examples have been taken not far from the coast and at depths not exceeding 80 metres ; young fish (30 to 35 mm.) in the Ross Sea at a depth of 356 metres. The B.A.N.Z. Antarctic Research Expedition obtained larger young (65 to 82 mm.) in 66° 36-5' S, 141° 44' E, with tow-nets hauled obliquely from o to 11 1 metres, and in 68° 14' S, 70° 11' E, with hand-nets at the surface (the fish being frozen and seen floating past the ship). The 'Aurora' obtained many young {c. 70 mm.) in 66° 8' S, 94° 17' E, at a depth of 20 metres. Adults have been taken at various localities, and generally at depths ranging from about 150 to 500 metres. A number of specimens have been recorded frozen on the ice barrier or at the surface of the sea, and others have been taken from the stomachs of seals or penguins. Family HARPAGIFERIDAE Closely related to the Nototheniidae, but with the body naked; two lateral lines, the upper always with tubules, the lower sometimes reduced to a series of pores; gill- membranes broadly united to the isthmus, not forming a fold across it; operculum hooked upwards posteriorly, its upper edge deeply concave. Five genera, all occurring in the Antarctic Zone, and one {Harpagifer) extending into the Subantarctic Zone. Key to the Genera I. A mental barbel present; opercles not spinate. A. Post-temporal not prominent; head not or scarcely broader than deep; interorbital region narrow. 1. Spinous dorsal fin above base of pectoral ... ... ... i. Artedidraco. 2. Spinous dorsal fin above operculum ... ... ... ... 2. Dolloidraco. B. Upper limb of post-temporal projecting as a prominent curved ridge; spinous dorsal fin above operculum. 1. Head longer than broad, scarcely broader than deep; interorbital region narrow. 3. HiSTIODRACO. 2. Head nearly as broad as long, much broader than deep; interorbital region wide or rather wide 4. Pogonophryne. II. No mental barbel; operculum and suboperculum each forming a prominent spine. 5. Harpagifer. Genus ARTEDIDRACO Artedidraco, Lonnberg, 1905, Wiss. Ergebn. Schwed. Siidpol.-Exped., v (6), p. 39; Regan, 1913, Trans. R. Soc. Edinburgh, XLix, p. 280. Type Artedidraco skottsbergi, Lonnberg. A comparatively short upper lateral line with tubules ; lower lateral line reduced to a series of pores. Head not or scarcely broader than deep ; snout shorter than eye ; 6-2 44 DISCOVERY REPORTS interorbital region narrow ; post-temporal not prominent ; opercles not spinate. Mouth of moderate size ; a mental barbel present ; teeth in villiform bands, without distinct canines. Spinous dorsal fin placed above the base of the pectoral. Vertebrae 38 (15 + 23)- Five species. Coasts of the Antarctic Continent, extending northwards in the Weddell Quadrant to South Georgia. Key to the Species I. Maxillary not extending beyond anterior ^ of eye; dorsal II-IV, 23-28; anal 16-21. A. Eye 3 A to 4 in head; barbel club-shaped, with papillose terminal knob, in the male, without distal expansion in the female, its length from less than I to nearly I that of head. 1. Depth 4 to 4^ in length; last ray of anal joined or nearly joined to base of caudal; pelvic fins not nearly reaching vent ... ... ... ... ... i- mirus. 2. Depth 5 to 5 i in length; last ray of anal separated by an interspace from caudal; pelvic fins reaching vent or origin of anal 2. orianae. B. Eye z% to 3I in head; barbel without distal expansion in both sexes, either smooth, finely papillose, or slightly fringed, its length never more than \ that of head. 1. Caudal fin rounded or subtruncate; dorsal II-IV, 24-27 3. skottsbergi. 2. Caudal fin slightly emarginate; dorsal II-III, 26-28 4. loennbergi. II. Maxillary extending to or nearly to below middle of eye; dorsal III-V, 27-30; anal 19-20; eye 3^ to 3I in head; barbel smooth, tapering, its length \ to more than \ that of head. 5 . shackletoni. Artedidraco mirus, Lonnberg. Artedidraco mirus, Lonnberg, 1905, Wiss. Ergebn. Schwed. Siidpol.-Exped., v (6), p. 40, pi. i, fig. 4, pi. iv, fig. 14; Regan, 1913, Trans. R. Soc. Edinburgh, XLix, p. 280. St. MS 71. 9. iii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 110-60 m.: I specimen, 73 mm. St. 39. 25. iii. 26. East Cumberland Bay, South Georgia. Large otter trawl, 179-235 m.: 1 specimen, 100 mm. St. 42. I. iv. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 120-204 m. : 1 specimen, 93 mm. St. WS 25. 17. xii. 26. Undine Harbour (North), South Georgia. Small beam trawl, 18-27 ^■'■ 2 specimens, 65, 70 mm. St. 140. 23. xii. 26. Stromness Harbour to Larsen Point, South Georgia. Large otter trawl, 122-136 m.: 12 specimens, 33-107 mm. St. 142. 30. xii. 26. East Cumberland Bay, South Georgia. Net (7 mm. mesh), attached to back of trawl, 88-273 "!•• I specimen, 105 mm. St. 148. 9. i. 27. Off Cape Saunders, South Georgia. Large otter trawl, with net (4 mm. mesh) attached, 132-148 m.: 2 specimens, 92, 100 mm. St. 149. 10. i. 27. Mouth of East Cumberland Bay, South Georgia. Large otter trawl, 200- 234 m.: 4 specimens, 95-115 mm. Depth of body 4 to 4-1 in the length, length of head 2| to a little more than 3 . Diameter of eye 3^ to 4, interorbital width about 9 in length of head. Maxillary extending to below anterior 1 or 1 of eye ; barbel from less than ^ to nearly i length of head, with a papillose terminal knob in the male, smooth and without distal expansion in the female; 6 or 7 gill-rakers on lower part of anterior arch. Upper lateral line short. HARPAGIFERIDAE 45 ending below anterior part of soft dorsal. Dorsal III, 23-25 ; middle rays of soft dorsal longest, the last joined to base of caudal; fins higher in adult males. Anal 16-18; last ray joined or nearly joined to base of caudal. Pectoral with 16 or 17 rays, if to if in length of head, extending to above vent or origin of anal ; pelvics a little shorter, not nearly reaching vent. Caudal subtruncate or rounded. Four dark cross-bars on back, extending upwards on the dorsal fins, the first at the base of the spinous dorsal, the second small and inconspicuous; a dark patch on praeorbital and another on cheek; lower parts of body irregularly spotted and marbled ; fins with series of spots, the dorsal sometimes more or less dusky, the anal blackish with a pale margin ; all the markings sometimes very faint. Fig. 27. Artedidraco mints, x i. Hab. South Georgia. This species was originally described from 4 specimens, 40 to 92 mm. in length without the caudal fin, from Antarctic Bay and Cumberland Bay, at depths varying from 22 to 250 metres. It was not previously represented in the British Museum collection. Artedidraco orianae, Regan. Artedidraco orianae, Regan, 1914, Ann. Mag. Nat. Hist., (8) xiii, p. 12; 1914, Rep. Brit. Antarct. (' Terra Nova') Exped., 1910, Zool. i (i), p. 7, pi. vi, fig. 2. Depth of body 5 to 5 1 in the length, length of head 2f to 3. Diameter of eye 3I, interorbital width 10 in length of head. Maxillary extending to below anterior \ of eye ; barbel \ length of head, club-shaped, with papillose distal expansion; 7 short gill- rakers on lower part of anterior arch. Upper lateral line of moderate length, extending to or nearly to below middle of soft dorsal. Dorsal III-IV, 25 ; rays of soft dorsal decreasing from third or fourth ; last adnate to caudal peduncle. Anal 17-18 ; separated by an interspace from caudal fin. Pectoral with 17 rays, f or f length of head, extending to above third or fourth ray of anal; pelvics as long, extending to vent or origin of anal. 46 DISCOVERY REPORTS Caudal subtruncate. Five dark bars on back, extending upwards on the dorsal fins, the first at base of spinous dorsal, the second and fourth stronger than the rest ; a spot on praeorbital, another on cheek; lower part of body irregularly spotted; fins with series of spots. Hab. South Victoria Land. Known only from the types, 2 specimens, each 80 mm. in total length, from off Cape Adare, 85 to 95 metres. Artedidraco skottsbergi, Lonnberg. Artedidraco skottsbergi, Lonnberg, 1905, t.c, p. 48, pi. iv, fig. 15; Regan, 1913, t.c, p. 280; Vaillant, 1906, Bull. Miis. Paris, 1906, No. 3, p. 139; 1906, Exped. Antarct. Fratif. (1903- 1905), Poiss., p. 46; Regan, 1914, Rep. Brit. Antarct. (' Terra Nova') Exped. 1910, Zool. i (i), p. 7, pi. V, fig. i; Norman, 1937, Rep. B.A.N.Z. Antarct. Res. Exped., Ser. B, i, p. 73. St. 599. 17. i. 31. 67° 08' S, 69° o6i' W. Large dredge, 203 m.: i specimen, 85 mm. Depth of body 4I to nearly 6 in the length, length of head 3 to 3I. Diameter of eye 3 to 3f, interorbital width about 14 to 20 in length of head. Maxillar}' extending to below anterior i or j of eye ; barbel less than i diameter of eye, simple, smooth or finely papillose; 7 gill-rakers on lower part of anterior arch. Dorsal II-IV, 24-27; middle soft rays longest. Anal 17-20. Last rays of dorsal and anal joined or nearly joined to base of caudal. Pectoral with 15 to 17 rays, i| to if in length of head, extending to above vent or origin of anal ; pelvics shorter, scarcely reaching vent in young, not as far in adult. Caudal rounded or subtruncate. Body with numerous irregular dark spots ; a series of blotches at base of dorsal fin sometimes continued on sides of body as bars ; vertical fins with series of spots on the rays, those on posterior part of dorsal and anal and near upper and lower edges of caudal deep black ; pectorals barred. Hab. Graham Land; South Shetlands; Enderby Land; Mac-Robertson Land; South Victoria Land. The type of the species, 57 mm. in length without the caudal fin, was taken south of Snow Hill, Graham Land, at a depth of 125 metres. In addition to the specimen collected by the Discovery Committee, the above description is based upon 18 specimens in the British Museum collection, up to 120 mm. in total length, and 13 others, 45 to 107 mm. long, collected by the B.A.N.Z. Antarctic Research Ex- pedition. I have also seen a specimen, 70 mm. in total length, in the Paris Museum (No. 06 — 152). This is from Booth Wandel Island, at a depth of 40 metres. Artedidraco loennbergi, Roule. Artedidraco loennbergi, Roule, 1913, Bull. Mus. Paris, 1913, No. i, p. 4. Artedidraco loenbergi, Roule, Angel and Despax, 1913, Deux. Exped. Antarct. Franf. (1908- 1910), Poiss., p. 13, pi. iv, fig. 4. Artedidraco loennbergii, Regan, 1914, Rep. Brit. Antarct. {'Terra Nova') Exped. 1910, Zool. i (i), p. 8, pi. V, fig. 2. Depth of body 5 to 6 in the length, length of head 2f to 4^. Diameter of eye 2| to 3 1, interorbital width 15 to 20 in length of head. Maxillary extending to below HARPAGIFERIDAE 47 anterior | of eye ; barbel | to J length of head, simple or slightly fringed distally ; 6 or 7 gill-rakers on lower part of anterior arch. Dorsal II-III, 25-28. Anal 18-21. Soft dorsal and anal usually highest posteriorly, their last rays joined or nearly joined to base of caudal. Pectoral with 14 to 16 rays, if to if in length of head, extending to above anal in young but not in adult ; pelvics as long as or a little shorter than pectorals, reaching vent in young but not in adult. Caudal slightly emarginate. A series of dark blotches at base of dorsal fin and sometimes at base of anal ; a dark lateral band, made up of irregular spots, from eye to caudal fin ; pale bands above and below it are con- tinued along the upper and lower margins of the caudal fin ; dorsal and middle of caudal with spots on the rays ; pectorals usually barred. Hab. Graham Land ; South Victoria Land. The type of the species, 93 mm. in total length, was collected by the ' Pourquoi Pas?' near Marguerite Bay (68° 00' S, 70° 20' W), at a depth of 230 metres. I have examined this type in the Paris Museum (No. 13 — 186). It has a somewhat larger eye than is seen in examples from Victoria Land, but appears to be otherwise identical. The 'Terra Nova' obtained 22 specimens, measuring up to no mm. in total length, from: Ross Sea, 74° 25' S, 179° 3' E, 295 metres; off new land south of the Balleny Islands, 375 metres; entrance to McMurdo Sound, 76° 56' S, 164° 12' E, 300 metres and 77 13' S, 164° 18' E, 380 metres; and near Inaccessible Island, 415 to 450 metres. Artedidraco shackletoni, Waite. Artedidraco shackletoni, Waite, Brit. Antarct. Exped. 1907-1909, 11, Biol., 11, p. 15, pi. ii; Regan, 1914, t.c, p. 8; Waite, 1916, Austral. Antarct. Exped. Set. Rep., Ser C, in (i), Fishes, p. 23; Norman, 1937, t.c , p. 73. Depth of body 4 to 5 in the length, length of head 2| to nearly 3. Diameter of eye 3i to 3I, interorbital width about 14 in length of head. Maxillary extending to below middle of eye or not quite as far ; barbel as long as or nearly as long as diameter of eye and I to more than I length of head, smooth, tapering ; 6 or 7 gill-rakers on lower part of anterior arch. Dorsal III-IV (V), 27-30; middle soft-rays longest. Anal 19 or 20. Last rays of dorsal and anal joined to base of caudal. Pectoral with 15 to 17 rays, if to if in length of head, extending to above vent or origin of anal; pelvics as long or shorter, not reaching vent (except in young). Caudal subtruncate. Pale brownish, uniform or with three dark brown bars on the back ; lower parts of hinder portion of body sometimes spotted with brown ; caudal and pectoral fins spotted, and sometimes traces of spots on the other fins; often a dusky band along base of anal fin. Hab. Enderby Land; Mac-Robertson Land; Queen Mary Land; South Victoria Land. The type of the species, 146 mm. in length, was from off Cape Royds, at a depth of 56 to 150 metres. Two specimens (132 and 142 mm.) in the British Museum col- lection, from the entrance to McMurdo Sound, and 6 specimens (65 to 120 mm.) collected by the B.A.N.Z. Antarctic Research Expedition are also included in the above description. 48 DISCOVERY REPORTS Genus DOLLOIDRACO Dolloidraco, Roule, 1913, Bull. Mus. Paris, 1913, No. i, p. 5; Roule, Angel and Despax, 1913, Deux. Exped. Antarct. Fraiif. (1908-1910), Poiss., p. 15. Type Dolloidraco longedorsalis, Roule. Differs from Artedidraco in having the spinous dorsal fin placed above the oper- culum. A single species. Coasts of the Antarctic Continent. Dolloidraco longedorsalis, Roule. Dolloidraco longedorsalis, Roule, 1913, Bull. Mus. Paris, 1913, No. i, p. 6; Roule, Angel and Despax, 1913, t.c, p. 16, pi. iv, figs. 1-3; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Sen C, III (i), Fishes, p. 24, pi. i, fig. 2, text-fig. 3. Dolloidraco longidorsalis, Regan, 1914, Rep. Brit. Antarct. {'Terra Nova') E.xped. 1910, Zool. i (i), p. 28. St. 1644. 16. i. 36. Bay of Whales (78° 24-85' S, 164'' 10-3' W). Russell Bottom Net, 626 m.: I female specimen, 115 mm. Depth of body about 5 in the length, length of head 2| to 2|. Diameter of eye ai to 2| in length of head ; interorbital space very narrow. Maxillary extending to below anterior margin of pupil or beyond ; barbel about as long as diameter of eye, simple (females?) or with a terminal expansion (males ?); 6 small, knob-like gill-rakers on lower part of anterior arch. Dorsal III 22-25. Anal 14-15. Last rays of dorsal and anal well separated from base of caudal. Pectoral with 16 or 17 rays, about | length of head, extending to above anterior part of anal; pelvics shorter, just reaching vent. Caudal subtruncate or a little rounded. Yellowish-brown, becoming darker on hinder part of body; a dark patch on cheek below eye, and a dusky vertical bar across the operculum; spinous dorsal dusky; soft dorsal mainly dusky; anal black in its basal half, white distally ; caudal white, with an incomplete oblique black bar across the base and a broader and more distinct oblique bar across its distal portion, the two being connected by black spots along the upper edge of the fin; pectorals with indistinct darker cross-bars, the lower rays more or less blackish ; pelvics dusky in their basal parts. Fig. 28. Dolloidraco longedorsalis. x Hab. Graham Land ; Queen Mary Land ; South Victoria Land. This species was originally described from 6 specimens, 72 to 89 mm. in length without the caudal fin, of which one is now in the British Museum collection, all from HARPAGIFERIDAE 49 Marguerite Bay, 230 metres^. In all these specimens the barbel is simple. Of the 7 specimens, 100 to 137 mm. in total length, taken by the 'Aurora' at Queen Mary Land, 610 to 670 metres, some have a simple barbel, but in others this has a distal expansion. In the specimen obtained by the Discovery Committee, which I have ascertained is a female, the barbel is simple. Thus, it appears probable that this is a secondary sexual character as in Artedidraco mirus. Genus HISTIODRACO Histiodraco, Regan, 1914, Rep. Brit. Antarct. {'Terra Nova') Exped. 1910, Zool. i (i), p. 9. Type Dolloidraco velifer, Regan. Differs from Dolloidraco in that the upper limb of the post-temporal projects as a prominent curved ridge. A single species. Coasts of the Antarctic Continent. Histiodraco velifer (Regan). Dolloidraco velifer, Regan, 1914, Ann. Mag. Nat. Hist., (8) xin, p. 12. Histiodraco velifer, Regan, 1914, Rep. Brit. Antarct. {'Terra Nova') Exped. 1910, Zool. i (i), p. 9, pi. V, fig. 3 ; Norman, 1937, Rep. B.A.N.Z. Antarct. Res. Exped., Ser. B, i, p. 74. Depth of body 4 to 4I in the length, length of head 2| to af. Diameter of eye 3 J to 3f, interorbital width about 12 in length of head. Maxillary extending to below middle of eye or beyond ; lower jaw rather strongly projecting ; barbel f to more than J length of head, fringed in its distal half; 7 very short, knob-like gill-rakers on lower part of anterior arch. Dorsal II-III, 26; spinous dorsal short and high; soft dorsal sometimes more or less elevated anteriorly, the highest rays f to ^ length of head. Anal 17. Pectoral with 18 or 19 rays, about | length of head, longer than pelvics, which do not reach vent. Caudal truncate above, rounded below. Body marbled or irregularly barred with darker brown ; fins more or less spotted ; a dark bar at base of pectoral and another across the caudal. Fig. 29. Histiodraco velifer. x |. Hab. Mac- Robertson Land; South Victoria Land. The types, 2 specimens 180 and 190 mm. in total length, are from the entrance to McMurdo Sound, 77° 13' S, 164° 18' E, at a depth of 385 metres. Two other examples 1 I have seen 4 types in the Paris Museum (No. 13 — 187-190), 100 to 120 mm. in total length. 50 DISCOVERY REPORTS (150 and 170 mm.) from 66° 45' S, 62° 03' E, 219 metres, collected by the B.A.N.Z. Antarctic Research Expedition, are also included in this description. Genus POGONOPHRYNE Pogonophryne, Regan, 1914, Ann. Mag. Nat. Hist., (8) xni, p. 13; 1914, Rep. Brit. Antarct. (' Terra Nova') Exped. 1910, Zool. i (i), p. 9. Type Pogonophryne scotti, Regan. Differs from Histiodraco in the wider interorbital region and the strongly depressed head, nearly as broad as long and much broader than deep; both lateral lines with tubules. Two species. Coasts of the Antarctic Continent and neighbouring islands. Pogonophryne scotti, Regan. Pogonophryne scotti, Regan, 1914, t.c, p. 13; 1914, t.c, p. 9, pi. vi, fig. i; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, ni (i). Fishes, p. 26. St. 1660. 27.1.36. 74° 46-4' S, 178" 23-4' E. Large otter trawl, 351m.: 2 specimens, 240, 250 mm. Fig. 30. Pogonophryne scotti. x i. Depth of body about 4 in the length, length of head 2| to 2|. Head li to i^ times as long as broad. Diameter of eye 5^ to 5I, interorbital width about 4^ in length of head. Maxillary extending to below middle of eye; lower jaw strongly projecting; barbel shorter than diameter of eye, blunt, papillose; 7 to 10 short, knob-like gill-rakers on lower part of anterior arch. Dorsal II-III, 25; spines low; anterior part of soft dorsal sometimes elevated. Anal 17-18. Pectoral with 19 rays, i to | length of head, extending to above origin of anal; pelvics short, rounded, not nearly reaching vent. HARPAGIFERIDAE SI Caudal a little rounded. Body faintly spotted and marbled with darker; a more distinct dark blotch on either side of base of anterior part of soft dorsal ; upper surface of head without markings ; dorsals, caudal and pectorals with series of dark spots on the rays. Hab. Queen Mary Land; South Victoria Land. The type of the species, 290 mm. in total length, is from the Ross Sea, 74° 25' S, 179° 3' E, at a depth of 295 metres. The 'Aurora' obtained a single example, 240 mm. in length, from off Drygalski Island at a depth of 112 metres. Pogonophryne marmoratus, sp.n. ? Artedidraco skottsbergi, Lonnberg, 1905, Wiss. Ergebn. Schwed. Siidpol.-Exped., v (6), pi. ii, fig-7- St. 1948. 4.1.37. 60° 49-4' S, 52°4o'W. Rectangular dredge, with Russell net, 490-610 m.: I male specimen, 175 mm. Holotype. Closely related to P. scotti, but head about if times as long as broad; interorbital width more than 7 in length of head ; barbel a little longer than diameter of eye, papil- lose, with terminal expansion. Dorsal II, 25. Anal 16. Upper surface and sides of head and body mottled and spotted with purplish brown ; lower parts of body mainly blackish ; soft dorsal and pectorals with series of dark spots, sometimes tending to form irregular bars; an irregular dark patch covering part of caudal; anal pale; inner rays of pelvics dusky. Fig. 31. Pogonophryne marmoratus. Holotype. xj. Hab. Near the South Shetland Islands. Regan [1914, Rep. Brit. Antarct. (' Terra Nova') Exped. 1910, Zool. i (i), p. 9] has suggested that the coloured drawing of a fish from the Bransfield Straits, 849 metres, reproduced by Lonnberg as Artedidraco skottsbergi, "seems rather to represent a 7-2 52 DISCOVERY REPORTS Pogonophryne", and it seems probable that the fish in question belonged to the species described above. The barbel is shown without a terminal expansion, but this may be due to the fact that the sketch was prepared in a hurry. Alternatively, the form of the barbel may differ in the sexes, as in Artedidraco mirus. Genus HARPAGIFER Harpagifer, Richardson, 1844, Zool. 'Erebus' and 'Terror' (Fish.), p. u; Regan, 1913, Trans. R. Soc. Edinburgh, XLix, p. 94; Norman, 1937, Discovery Rep., xvi, p. 94. Type Batrachus bispinis, Schneider. A comparatively short upper lateral line with tubules ; lower lateral line reduced to a series of pores ; head rather broader than deep ; snout about as long as eye ; interorbital region of moderate width; post-temporal not prominent; operculum and suboper- culum each forming a prominent spine. Mouth rather large ; no mental barbel ; teeth in villiform bands, without distinct canines. Spinous dorsal fin placed above the base of the pectoral. A single species. Patagonian region to Graham Land, Kerguelen and Macquarie Island. Harpagifer bispinis (Schneider). Batrachus bispinis, Schneider \ex Forster MS.], 1801, in Bloch, Syst. Ichth., p. 45. Callionymus bispinis, Forster, 1844, Anim. Mar. Atist., p. 360. Harpagifer bispinis, Richardson, 1844, Zool. 'Erebus' and 'Terror' (Fish.), pp. 11, 19, pi. vii, figs. 1-3, pi. xii, figs. 8, 9; Giinther, i860. Cat. Fish., u, p. 263; Gill, 1876, Bull. U.S. Nat. Mus., Ill, p. 41; Studer, 1879, Arch. Naturg., XLV (i), p. 130; Giinther, 1880, Shore Fish. 'Challenger', p. 17; Fischer, 188^, Jahrb. Hamburg. Wiss. Anst., 11, p. 57; Vaillant, 1906, Bull. Mus. Paris, 1906, No. 3, p. 139; 1906, Exped. Antarct. Franf. (1903-1905), Poiss., p. 45; Roule and Despax, 191 1, Bull. Mus. Paris, 1911, No. 5, p. 279; Pappenheim, 1912, Deutsche Siidpol. -Exped., xni, Zool. v, p. 177, pi. ix, figs. 5a, 5^; Roule, Angel and Despax, 1913, Deux. Exped. Antarct. Franf. (1908-1910), Poiss., p. 13; Regan, 1913, Trans. R. Soc. Edijiburgh, XLix, p. 280; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, in (i). Fishes, p. 70; Norman, 1937, Discovery Rep., xvi, p. 94, fig. 46; Norman, 1937, Rep. B.A.N.Z. Antarct. Res. Exped., Ser. B, i, pp. 59, 64. Harpagifer palliolatus, Richardson, 1844, t.c, p. 20, pi. xii, figs. 5-7. St. MS 33. I. V. 25. I cable E of Hobart Rock, East Cumberland Bay, South Georgia. Small beam trawl, 40 m. : i specimen, 65 mm. St. WS 25. 17. xii. 26. Undine Harbour (North), South Georgia. Small beam trawl, 18-27 '^••■ 1 specimen, 37 mm. St. 145. 7. i. 27. Stromness Harbour, South Georgia. Small beam trawl, 26-35 m.: i specimen, 95 mm. Depth of body 3! to 4'| in the length, length of head 2f to 3^. Diameter of eye 4 to 5 in length of head, equal to or rather greater than interorbital width. Jaws equal anter- iorly or lower a little projecting; maxillary extending to below middle or anterior part of eye ; gill-rakers reduced to a few rudimentary knobs near angle of arch. Dorsal III-V, 21-26. Anal 16-21. Pectoral with 16 to 18 rays, more than | as long as head, a little longer than pelvics, which do not reach the vent. Caudal rounded. Coloration HARPAGIFERIDAE S3 very variable; body usually with dark blotches or irregular cross-bars; fins mostly with small dark spots, often united to form narrow bars. Hab. Graham Land ; South Shetlands ; South Orkneys ; South Georgia ; Straits of Magellan; Patagonia; Falkland Islands; Marion Islands; Crozets; Kerguelen; Mac- quarie Island. Fig. 32. Harpagifer bispinis . xi. In addition to the specimens listed above, this description is based upon a large series of examples, up to 100 mm. in total length, in the British Museum collection, and a number of others in the collection of the B.A.N.Z. Antarctic Research Expedition. Schneider's Batrachiis bispinis was based upon the MS. notes and drawing of Forster (MS IV, 45) of a specimen from the Straits of Magellan. The type of Harpagifer pallio- latus, 70 mm. in total length, is from the Falkland Islands. This is mainly a littoral species, occurring in tide pools and under rocks and stones ; it is also found in the "kelp" in shallow water. Family BATHYDRACONIDAE Closely related to the Nototheniidae and Harpagiferidae, but with the head more or less depressed and the snout usually produced; mouth usually non-protractile; palatine elongate, without lateral ethmoid attachment ; gill-membranes attached to the isthmus or forming a fold across it. No spinous dorsal fin. Eight genera, all confined to the Antarctic Zone. Key to the Genera I. Body scaly or with serrated bony plates; snout spatulate; teeth villiform or cardiform, in bands, without canines. A. Body without serrated bony plates; lateral line usually complete, but always extending at least as far as middle of dorsal. 1. A single lateral line, running to or towards middle of base of caudal fin. I. Bathydraco. 2. Two or three lateral lines (at least in adult), the upper always present and running near base of dorsal fin. a. Two lateral lines, a long upper one, and a short lower one placed above base of anal fin ; snout narrow, more than twice as long as eye ; dorsal with 45 to 47 rays, anterior part of fin not elevated; caudal emarginate 2. Gerlachea. 54 DISCOVERY REPORTS b. Three lateral lines (at least in adult), a long upper one, a short middle one on side of caudal peduncle, and a short lower one above base of posterior part of anal fin; snout broader, not more than i\ times as long as eye; dorsal with 30 to 37 rays, anterior part of fin forming a fairly distinct lobe; caudal rounded or subtruncate 3. Racovitzia. B. Body quadrangular, with a series of V-shaped, serrated bony plates at each angle, each plate with a backwardly directed spine; a single short lateral line, ending below anterior part of dorsal 4. Prionodraco II. Body naked. A. Snout produced, spatulate; teeth in villiform bands, without canines; operculum with divergent ridges, ending in spinous points. 1. Upper lateral line ending below posterior part of dorsal; dorsal with 61 to 66 rays. 5. Cygnodraco. 2. Upper lateral line extending beyond end of dorsal; dorsal with 42 to 46 rays. 6. Parachaenichthys. B. Snout not spatulate; canine teeth present in one or both of the jaws; operculum with a strong spine terminating in a hooked process. 1 . Head and anterior part of body only slightly depressed ; no canine teeth in lower jaw; mouth protractile; three lateral lines 7. Psilodraco. 2. Head and anterior part of body distinctly depressed; strong canine teeth near symphysis of mandibles, exposed in front of snout; mouth not protractile; two lateral lines 8. Gymnodraco. Genus BATHYDRACO Bathydraco, Giinther, 1878, Ann. Mag. Nat. Hist., (5) 11, p. 18; Regan, 1913, Trans. R. Soc. Edinburgh, xlix, p. 281. Type Bathydraco antarcticus, Giinther. Body scaly; a single lateral line, running to or towards middle of base of caudal. Snout produced, spatulate; jaws with bands of small villiform teeth. Operculum with upper edge concave, but without hooked branch posteriorly ; gill-rakers well-developed and of moderate length; gill-membranes free from the isthmus, but slightly united anteriorly ; 7 branchiostegals. Five species. Deep water near the coasts of the Antarctic Continent. Key to the Species I. Depth 9 to 10 in length; eye 3! to 5 in head; dorsal 34-38, anal 29-31 ; sides of head more or less scaly. A. 1 1 to 16 gill-rakers on lower part of anterior arch; lateral line complete. 1. Only the upper half of operculum scaly; cheek completely scaled; 16 gill-rakers on lower part of anterior arch; about 140 scales in a longitudinal series. I . antarcticus. 2. Operculum nearly completely scaly; cheek naked below suborbitals; 11 gill-rakers on lower part of anterior arch; about 125 scales in a longitudinal series. 2. marri. 3. A small patch of scales on upper part of operculum; cheek naked below sub- orbitals; II gill- rakers on lower part of anterior arch; about 90 scales in a longi- tudinal series ... 3- macrolepis. BATHYDRACONIDAE 55 B. 19 to 22 gill-rakers on lower part of anterior arch; lateral line ending at some distance from caudal; a small patch of scales on upper part of operculum ... ... 4. scotiae. II. Depth 7 J in length; eye 3^ in head; dorsal 27, anal 22; sides of head naked; 17 gill-rakers on lower part of anterior arch .. . ... ... ... ... ... ... 5. nudiceps. Bathydraco antarcticus, Giinther. Bathydraco antarcticus, Giinther, 1878, Ann. Mag. Nat. Hist., (5) 11, p. 18; 1887, Deep-Sea Fish. 'Challenger", p. 47, pi. viii, fig. A; Regan, 1913, Trans. R. Soc. Edinburgh, XLix, p. 282. Depth of body 9 in the length, length of head 3. Snout i^ as long as eye, diameter of which is 4 in length of head ; interorbital width 20. Lower jaw projecting; maxillary extending to vertical from anterior margin of eye ; cheek completely scaled ; only the upper half of the operculum scaly; 16 gill-rakers on lower part of anterior arch. About 140 scales in a lateral longitudinal series, about 60 in the lateral line, which is com- plete. Dorsal 36. Anal 31. Pectoral with 22 rays, truncated, as long as head without snout, extending to above origin of anal; pelvics much shorter, not nearly reaching vent. Caudal subtruncate. Pale brownish, the lower parts somewhat darker; caudal, anal, pelvics and lower parts of pectorals dusky. Hab. South-east of Heard Island, 2300 metres. Known only from the type of the species, 260 mm. in total length. Bathydraco marri, sp.n. St. 1658. 26.1.36. 76" 09-6' S, 168° 40' E. Rectangular dredge, with Russell net, 520 m.: I specimen, 112 mm. Holotype. Fig. 33. Bathydraco marri. Holotype. x i|. Depth of body about 9 in the length, length of head 3I. Snout slightly longer than eye, diameter of which is 3I in length of head ; interorbital width about 12. Lower jaw a little projecting ; maxillary not reaching vertical from anterior margin of eye ; cheek naked below the suborbitals, but a large patch of scales on the praeoperculum and operculum; 11 gill-rakers on lower part of anterior arch. About 125 scales in a lateral longitudinal series, about 55 in the lateral line, which is complete. Dorsal 36. Anal 29. Pectoral with 22 rays, as long as head without snout, extending to a little beyond origin of anal; pelvics shorter, not reaching vent. Caudal a little rounded (.?). Pale 56 DISCOVERY REPORTS yellowish-brown, with about 6 irregular dark patches on the back, of which 2 extend downwards on the sides and reach the anal fin ; a number of irregular dark spots and blotches on sides of body and on the head ; a dark longitudinal band near base of dorsal ; caudal blackish, except at its base ; a dusky patch on upper part of base of pectoral ; pelvics a little dusky in their distal parts ; anal pale. Hab. Ross Sea. This species is apparently most nearly related to B. macrolepis, Boulenger. I have much pleasure in naming it after Mr J. W. S. Marr, a member of the Discovery Com- mittee's scientific staff. Bathydraco macrolepis, Boulenger. Bathydraco macrolepis, Boulenger, 1907, Nat. Antarct. Exped. Nat. Hist., 11, Fish., p. 4, pi. i, fig. 3; Regan, 1913, i.e., p. 282. Depth of body 9 in the length, length of head 3. Snout if as long as eye, diameter of which is 4I in length of head ; interorbital width 14. Lower jaw projecting ; maxillary reaching vertical from anterior margin of eye ; cheek naked below the suborbitals ; a small patch of scales on upper part of operculum; 11 gill-rakers on lower part of anterior arch. About 90 scales in a lateral longitudinal series, about 55 in the lateral line, which is complete. Dorsal 34. Anal 29. Pectoral with 23 rays, as long as head behind middle of eye, extending to above origin of anal ; pelvics shorter, not reaching vent. Caudal subtruncate. Brownish; posterior part of caudal, distal parts of pelvics, and margin of anal more or less dusky. Hab. South-west of Balleny Islands, 460 metres. Known only from the type of the species, 210 mm. in total length. Bathydraco scotiae, DoUo. Bathydraco scotiae, Dollo, 1906, Proc. R. Soc. Edinburgh, xxvi, p. 65; Regan, 1913, t.c, p. 282, pi. ix, fig. 4; Norman, 1937, Rep. B.A.N.Z. Antarct. Res. Exped., Ser. B, i, p. 74. Depth of body 9 to 10 in the length, length of head 3^ to 3^. Snout if as long as eye, diameter of which is about 5 in length of head ; interorbital width 12 or 13. Lower jaw projecting; maxillary not reaching vertical from anterior margin of eye; cheek naked below the suborbitals; a small patch of scales on upper part of operculum; 19 to 22 gill-rakers on lower part of anterior arch. About 100 scales in a lateral longi- tudinal series, 36 to 43 in the lateral line, which ends at a distance from the caudal equal to J or | its own length. Dorsal 38-40. Anal 30-31. Pectoral with 22 or 23 rays, as long as head without snout, extending to above origin of anal or a little beyond; pelvics shorter, not reaching vent ; caudal subtruncate. Pale brownish, the lower parts somewhat darker ; caudal, anal, pectorals and pelvics more or less dusky. Hab. Off Coats Land ; off coast between Queen Mary Land and Wilkes Land. The types of the species, 2 specimens 133 and 145 mm. in total length, were taken by the ' Scotia' off Coats Land (71° 22' S, 16° 34' W), at a depth of 2645 metres. The BATHYDRACONIDAE 57 B.A.N. Z. Antarctic Research Expedition obtained a single specimen, i6o mm. long, from 64° 21' S, 116" 02' E, at a depth of 2260 metres. Bathydraco nudiceps, Waite. Bathydraco nudiceps, Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, in (i), Fishes, p. 27, pi. i, fig. 3, text-fig. 4. Depth of body 7^ in the length, length of head 3. Snout a little longer than eye, diameter of which is 3^ in length of head ; interorbital width 12. Lower jaw projecting ; maxillary extending to below anterior margin of eye ; head entirely naked ; 17 gill-rakers on lower part of anterior arch. Body covered with cycloid scales ; lateral line forming a low arch over the operculum, consisting of 5 or 6 large pores ; a series of imperforate depressed scales forms a line from above the pectoral, thence along the middle of the side to the base of the caudal rays. Dorsal 27. Anal 22. Pectoral with 22 rays, extending to above the third anal ray; pelvics about as long, reaching vent. Caudal truncate. Pale brownish, the lower parts darker ; dorsal fin pale, all other fins dark. Hab. Off Queen Mary Land. The type of the species, 144 mm. in total length, was from off Shackleton Ice-shelf (65° 20' S, 95° 27' E), at a depth of 450 metres: the 'Aurora' obtained two other examples in the same neighbourhood at a depth of about 600 metres. Genus GERLACHEA Gerlachea, Dollo, 1900, Bull. Acad. roy. Belg. {Classe Set.), No. 3, p. 195; 1904, Res. Voy. 'Belgica', Zool., Poiss., p. 24; Regan, 1913, t.c, p. 283. Type Gerlachea australis, Dollo. Differs from Bathydraco in having two lateral lines, the long upper one running near the base of the dorsal fin, the short lower one above the anal fin. Snout narrow, more than twice as long as diameter of eye. Dorsal with 45 to 47 rays; anterior part of fin not elevated ; caudal emarginate. A single species. Deep water off the coasts of the Antarctic Continent. Gerlachea australis, Dollo. Gerlachea australis, Dollo, 1900, t.c, p. 196; 1904, t.c, p. 25, pi. ii, fig. i, pi. v, fig. 2; Roule and Despax, 1911, Bull. Mus. Paris, 1911, No. 5, p. 280; Roule, Angel and Despax, 1913, Deux. Exped. Antarct. Franf. (1908-1910), Poiss., p. 10, pi. i, fig. i, pi. ii, figs. 1-2, pi. iii, fig. 5; Regan, 1913, t.c, p. 283; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, in (i). Fishes, p. 29. Depth of body 7I to more than 12 in the length, length of head 3^ to nearly 3 J. Snout more than twice as long as eye, diameter of which is ^l to about 5!- in length of head; interorbital width 11 to 14. Lower jaw projecting; maxillary not nearly reaching anterior margin of eye ; cheek scaled ; a patch of scales on upper part of oper- culum; gill-rakers represented by about 8 low dentigerous patches on the first arch. 85 scales in upper lateral line, which extends to below last ray of dorsal fin or a little beyond; length of lower lateral line variable. Dorsal 45-47. Anal 34-35. Pectoral 58 DISCOVERY REPORTS with 26 to 28 rays, f to nearly f length of head, extending to above vent or anterior rays of anal ; pelvics much shorter, not nearly reaching vent. Caudal emarginate, its angles rounded. Yellowish brown; body with dark blotches or irregular cross-bars; a dark line from the snout to the eye, thence obliquely downwards across the opercles ; all fins more or less dusky, except the anal, which is pale ; distal parts of pelvics black. Fig. 34. Gerlachea australis. Holotype. x f . Hab. Graham Land; Charcot Land; Queen Mary Land: in deep water. Through the courtesy of Dr V. Van Straelen, I have been able to examine the holotype of this species preserved in the Musee Royal d'Histoire Naturelle de Belgique. This is 180 mm. in total length and is in excellent condition. It was taken by the 'Belgica' at 71° 14' S, 89° 14' W, at a depth of 450 metres. The 'Pourquoi Pas?' obtained 4 examples^ from 70° 10' S, 80° 50' W, at a depth of 450 metres, one of which (in a poor state of preservation) has been examined by me in the Paris Museum. The 'Aurora' obtained a single example, 208 mm. in length, from 64° 44' S, 97° 28' E, at a depth of 670 metres. The extent to which the lower lateral line is developed seems to be somewhat variable. In the holotype it appears to be represented by about 18 scales on the left side, whereas, on the right side it is longer and is represented by at least 30 scales. Roule, Angel and Despax remark that in the specimens obtained by the 'Pourquoi Pas?' this lateral line is unequally developed on the two sides of the same fish, and Waite records the same condition in his example. Three of the specimens taken by the 'Pourquoi Pas?' in January were mature females, and contained large eggs, varying in size from 2-5 to 3 mm. Genus RACOVITZIA Racovitzia, Dollo, 1900, Bull. Acad. roy. Belg. (Classe Set.), No. 3, p. 317. Type Racovitzia glacialis, Dollo. Racovitzaia, Dollo, 1904, Res. Voy. 'Belgica', Zool., Poiss., p. 28; Regan, 1913, t.c, p. 283 [emend, pro Racovitzia]. Aconichthys, Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, iii (i), Fishes, p. 30. Type Aconichthys harrissoni, Waite. Related to Bathydraco, but with three lateral lines (except in the young), the upper always present and running near the base of the dorsal fin, the short middle one on the side of the caudal peduncle, and the short lower one above the base of the posterior 1 One of these specimens has now been received in exchange from the Paris Museum : this is also in a very poor state of preservation. BATHYDRACONIDAE 59 part of the anal fin. Snout rather broad, spatulate, not more than i| times as long as the eye. Dorsal with 30 to 37 rays, the anterior part of the fin forming a fairly distinct lobe ; caudal rounded or subtruncate. One or two species. Coasts of Antarctic Continent. Comparison of the unique type specimen of Racovitzia glacialis^ with examples of Aconichthys harrissoni shows that the two fishes are almost certainly congeneric and may eventually prove to be identical. The body of Racovitzia was described by Dollo as having scattered groups of scales, but this character seems to be due to damage to the fish in capture. A specimen of Aconichthys (98 mm. in total length) shows a very similar condition, but a close examination of the skin reveals that many of the scales have become detached, leaving isolated groups on various parts of the body. In the type of Racovitzia only the upper lateral line is present, and this runs near the base of the dorsal fin. In a specimen of Aconichthys, 80 mm. in total length, the same condition is found, and the single lateral line ends below the posterior rays of the dorsal fin. In the specimen of 98 mm. mentioned above, this upper line extends a little beyond the end of the dorsal fin, and there is an indication of the middle line on the side of the caudal peduncle; there is, however, no trace of a lower lateral line. In all the larger examples of Aconichthys examined by me three lines are present, although in one example of 195 mm. on one side of the body the lower line is reduced to 2 or 3 scales, and in another of about the same size this line is present on one side of the body but entirely wanting on the other. I am unable to confirm the presence of the "in- cubatory pouch" between the pelvic fins and the vent described by Dollo, but am inclined to think that a post-mortem shrinkage in this region accompanied by folding of the skin may have produced the pouch-like condition. Racovitzia glacialis, Dollo. Racovitzia glacialis, Dollo, 1900, t.c, p. 318. Racovitzaia glacialis, Dollo, 1904, t.c, p. 29, pi. ii, figs. 2, 3, pi. v, fig. 3; Regan, 1913, t.c, p. 283. Depth of body 12 in the length, length of head 35. Snout longer than eye, about 2| in length of head; diameter of eye about 3I, interorbital width 25. Lower jaw slightly projecting; maxillary not reaching vertical from anterior margin of eye. Dorsal 30. Anal 27. Pectoral with 25 rays, about | length of head, extending as far as vent; pelvics shorter, not reaching vent. Body with dark spots and blotches forming very irregular cross-bars; an irregular black spot on anterior part of dorsal; anal pale; a black spot on upper edge of caudal ; pectorals blackish at the base and with a black spot on upper edge ; pelvics with a black spot distally. 1 This specimen, which is 82 mm. in total length, is in very poor condition, the caudal fin being largely wanting and a portion of the caudal region of the body broken off. It is very difficult to handle the specimen or to make any close examination of it, for fear of causing further damage. 8-2 6o DISCOVERY REPORTS Hab. 71° 19' S, 87° 37' W; 435 metres. Known only from the unique holotype, 82 mm. in total length, in the Musee Royal d'Histoire Naturelle de Belgique, Brussels. As pointed out above, I have re-examined this specimen, but, owing to the poor state of preservation, have been able to add very little to the original description. Fig. 35. Racovitzia glacialis. Holotype. x i. (After Dollo.) Racovitzia harrissoni (Waite). Aconichthys harrissoni, Waite, 1916, t.c, p. 30, pi. ii, fig. i, text-fig. 5; Norman, 1937, Rep. B.A.N.Z. Antarct. Res. Exped., Ser. B, i, p. 74. St. 600. 17. i. 31. 67° 09' S, 69° 27' W. Large dredge, 487-512 m.: i specimen, 80 mm. St. 1652. 23. i. 36. 75° 56-2' S, 178° 35-5' E. Rectangular dredge, with Russell net, 567 m.: I specimen, 293 mm. Depth of body 7I to 9^ in the length, length of head 3 to 3^. Snout i\ to i| times as long as eye, 3 to 3^ in length of head; diameter of eye 3 1 to 4 J in length of head; interorbital region very narrow. Lower jaw projecting ; maxillary extending to or nearly to level of anterior margin of eye ; a large patch of scales on the cheek and a small patch on upper part of operculum; 10 to 12 short gill-rakers on lower part of anterior arch. About 140 scales in a lateral longitudinal series; 82 to 91 in upper, 9 to 13 in middle, and 5 to 13 in lower lateral line (when this is present). Dorsal 34-37. Anal 28-31. Pectoral with 22 or 23 rays, about f length of head, extending to above origin or anterior rays of anal ; pelvics much shorter, not nearly reaching vent. Caudal rounded or sub- truncate. Pale yellowish brown, spotted and variegated with darker; 6 black blotches on the back, the 4 below the dorsal fin most conspicuous and extending on to the fin ; lower parts more or less dusky; dorsal fin with small dark spots on the rays, and sometimes with a dark blotch covering the posterior rays ; anal fin pale or with a dusky longitudinal submarginal band, and with a white margin; caudal with its distal part dusky and its upper and lower edges black; always a distinct black spot on proximal part of upper edge; sometimes a narrow white posterior edge to the fin; pectoral usually with cross-bars ; pelvics more or less dusky. Hab. Off Graham Land ; Mac-Robertson Land ; Queen Mary Land ; South Victoria Land. This species was originally described from 5 specimens taken off Queen Mary Land at depths varying from 450 to 610 metres. The holotype was 278 mm. in total length. BATHYDRACONIDAE 6i In addition to the 2 specimens listed above, the description includes 4 examples (98 to 195 mm.) obtained by the B.A.N.Z. Antarctic Research Expedition. It seems possible that this species will prove to be identical with R. glacialis, but, as described by DoUo, the latter has fewer dorsal and anal rays, a more slender body, and a rather longer snout. It may be noticed that the intense black spot on the upper edge of the caudal fin characteristic of this species is shown by DoUo in his figure of the type of R. glacialis. Fig. 36. Racovitsia harrissoni. x|. Genus PRIONODRACO Prionodraco, Regan, 1914, Ann. Mag. Nat. Hist., (8) xni, p. 13; 1914, Rep. Brit. Antarct. {'Terra Nova') Exped. 1910, Zool. i (i), p. 10. Type Prionodraco evansii, Regan. Body elongate, compressed, quadrangular, with a series of V-shaped, serrated, bony plates at each angle ; each plate with a backwardly directed spine ; lower series of plates ending in a group of nearly normal serrated scales behind pectoral fin ; usually a series of similar scales along middle of side ; body otherwise naked. Lateral line single, in- complete. Snout spatulate; mouth slightly protractile; teeth small, villiform, in bands. Operculum hooked upwards posteriorly, its upper edge deeply concave ; gill-membranes narrowly united to the isthmus anteriorly. Vertebrae 50 (16 + 34). A single species. Coasts of the Antarctic Continent. Prionodraco evansii, Regan. Prionodraco evansii, Regan, 1914, t.c, p. 13; 1914, t.c, p. 10, pi. vii, fig. i ; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, ni (i), Fishes, p. 35; Norman, 1937, Rep. B.A.N.Z. Antarct. Res. Exped., Ser. B, i, p. 75. St. 167. 20. ii. 27. Off Signy Island, South Orkneys. Net (7 mm. mesh) attached to back of trawl, 244-344 "^•- 3° specimens, 65-162 mm. Depth of body 7 to 8 in the length, length of head 3 to 3|. Snout as long as or a little longer than eye, diameter of which is 3 to 3Mn length of head ; interorbital width 14 to 16. Lower jaw a little projecting; maxillary extending to or nearly to level of anterior margin of eye ; cheeks and opercles entirely naked ; 15 to 18 gill-rakers on lower part of anterior arch. About 50 plates in the upper series; 12 to 17 tubular scales in 62 DISCOVERY REPORTS lateral line, which ends below anterior part of dorsal fin. Dorsal 34-38. Anal 29-33. Pectoral with 21 or 22 rays, f to f length of head, extending to above anterior rays of anal; pelvics shorter, not reaching vent. Caudal rounded or subtruncate. Head and body with numerous dark spots ; usually a well-defined lateral series of large oblong or squarish spots ; a deep black spot on base of anterior part of dorsal fin and sometimes 2 or 3 smaller spots more posteriorly ; some small spots on the rays of the dorsal fin ; caudal and pectorals with spots or irregular cross-bars; anal generally with a dark longitudinal band along its margin, becoming narrower posteriorly, the rays tipped with white ; pelvics pale or more or less dusky. Fig. 37. Prionodraco evansii. x Hab. South Orkneys; Mac-Robertson Land; Queen Mary Land; South Victoria Land. The types of the species, 11 specimens, measuring up to 132 mm. in total length, are from the Ross Sea (74° 25' S, 179° 3' E), 290 metres ; and the entrance to McMurdo Sound (76° 56' S, 164° 12' E and 77° 13' S, 164° 18' E), 300 and 380 metres. In addition to the types and the examples obtained by the Discovery Committee, the above description includes a number of specimens, up to 145 mm. in total length, collected by the B.A.N. Z. Antarctic Research Expedition. Genus CYGNODRACO Cygnodraco, Waite, 1916, Austral. Antarct. Exped. Set. Rep., Ser. C, in (i), Fishes, p. 32. Type Cygnodraeo mawsoni, Waite. Body elongate, compressed, naked; two lateral lines, the upper terminating below posterior part of dorsal fin, the lower represented by a broken series of depressed, non- perforate scales, running along middle of hinder part of body. Head scarcely depressed ; snout produced, spatulate; interorbital region narrow; jaws with bands of small, villiform teeth. Operculum with upper edge concave, armed with divergent ridges ending in spinous points; gill-rakers represented by very small, flat, dentigerous knobs; gill-membranes free from the isthmus but slightly united anteriorly. Dorsal with 61 to 66 rays; anal with 35 to 38 rays. Pelvics rather long and narrow, as long as or longer than the pectorals. A single species. Coasts of the Antarctic Continent. BATHYDRACONIDAE 63 Cygnodraco mawsoni, Waite. Cygnodraco mawsoni, Waite, 1916, t.c, p. t,2, pi. iii, fig. i, text-fig. 6; Norman, 1937, t.c, p. 75. Depth of body yf to nearly 10 in the length, length of head 2| to 3. Snout nearly \ length of head; diameter of eye 5I (young) to nearly 6, interorbital width 31 to 33 in length of head. Lower jaw a little projecting; maxillary not nearly reaching vertical from anterior margin of eye; 14 or 15 gill-rakers on lower part of anterior arch. About no scales in the upper lateral line, which follows the dorsal profile. Dorsal 61-66. Anal 35-38. Pectoral with 22 or 23 rays, length 2 to zl in that of head, not nearly reaching vent; pelvics as long as or longer than pectorals (much longer in young), not reaching vent. Caudal rounded. Yellowish or greyish-brown; traces of a dark line on each side of the snout, a dark spot below the eye, and a dark bar running back- wards from the eye towards the suboperculum ; body with about 7 rather indistinct dark cross-bars on the back and upper parts of the sides; dorsal, anal and caudal more or less dusky, at least in adults, the dorsal with two narrow longitudinal dark stripes running for the whole length of the fin ; pectorals and pelvics pale. Fig. 38. Cygnodraco mawsoni. x\. Hab. Mac-Robertson Land ; Queen Mary Land. The type of the species, 466 mm. in total length, was taken off Drygalski Island (65° 42' S, 92° 10' E), at a depth of 112 metres. The above description is based mainly upon 6 specimens, 125 to 360 mm. in total length, obtained by the B.A.N.Z. Antarctic Research Expedition, at a depth of 219 metres. Genus PARACHAENICHTHYS Parachaenichthys, Boulenger, 1902, Rep. Coll. Nat. Hist. "Southern Cross', p. 176. Type Chaenichthys georgianus, Fischer. Very closely related to Cygnodraco, but with the head and anterior part of the body more or less depressed and the interorbital region broader. The upper lateral line extends beyond the end of the dorsal fin. Dorsal with 42 to 46 rays ; anal with 30 to 33 rays. Pelvics broader, generally a little shorter than pectorals. Two species. Graham Land and neighbouring islands ; South Georgia. 64 DISCOVERY REPORTS Parachaenichthys georgianus (Fischer). Chaenichthys georgianus, Fischer, i8S ^, Jahrb. Hamburg. Wiss. Anst., ii, p. 50, pi. i, figs. 1-2. Parachaenichthys georgianus, Boulenger, 1902, i.e., p. 176; Lonnberg, 1905, Wiss. Ergebn. Schwed. Siidpol.-Exped., v (6), p. 36; 1906, K. Svensk. Vet.-Akad. Handl., XL (5), p. 95; Regan, 1913, Trans. R. Soc. Edinburgh, XLix, p. 287. St. MS 32. I. V. 25. East Cumberland Bay, South Georgia. Small beam trawl, 40 m.: 2 speci- mens, 84, 185 mm. St. 39. 25. iii. 26. East Cumberland Bay, South Georgia. Large otter trawl, 179-235 m.: 8 specimens, 260-470 mm. St. 42. I. iv. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 120-204 m.: 2 specimens, 83, 92 mm. St. 45. 6. iv. 26. 27 miles S 85° E of Jason Light, South Georgia. Large otter trawl, 270- 238 m. : 4 specimens, 280-480 mm. ; large otter trawl, 238-270 m. : i specimen, 82 mm. St. 140. 23. xii. 26. Stromness Harbour to Larsen Point, South Georgia. Large otter trawl, 122-136 m.: I specimen, 160 mm. St. — . 6. i. 27. Stromness Harbour, South Georgia. Hand line, 9-10 m.: i specimen, 455 mm. St. 148. 9. i. 27. Off Cape Saunders, South Georgia. Larger otter trawl, 132-148 m. : i specimen, 170 mm. St. 149. 10. i. 27. Mouth of East Cumberland Bay, South Georgia. Large otter trawl, 200- 234 m. : 4 specimens, 175-435 ■^"^• St.WS62. 19.1.27. Wilson Harbour, South Georgia. Small beam trawl, 26-83 m. : 1 1 specimens, 65-115 mm. St. WS 65. 22. i. 27. Undine Harbour (North), South Georgia. Hand line, 5 m.: 2 specimens, 450, 475 mm. St. — . Nov. 1930. I specimen, 133 mm., taken from stomach of Chaenocephalus aceratus. Fig. 39. Parachaenichthys georgianus. x J. Depth of body 7 to 10 in the length, length of head 2J to nearly 3. Snout nearly \ length of head; diameter of eye 5 (young) to yj, interorbital width 13 to 17 in length of head. Lower jaw a little projecting ; maxillary not nearly reaching vertical from anterior margin of eye; operculum with 3 radiating ridges terminating in spinous points; 16 to 19 gill-rakers on lower part of anterior arch. 102 to no tubular scales in upper lateral line, which extends beyond last ray of dorsal fin and joins the lower lateral line on the caudal peduncle ; this part of the lower lateral line with 4 to 9 tubular scales, in front of which is an irregular series of depressed non-perforate scales. Dorsal 43-^46. Anal 31-33. Pectoral with 21 to 23 rays, length 2 to 2\ in that of head, not or scarcely reaching vent; pelvics a little shorter, not reaching vent in adult. Caudal rounded or BATHYDRACONIDAE 65 subtruncate. Yellowish-brown, spotted and variegated with darker, the markings being more distinct in young and half-grown specimens; dorsal pale in young, with series of spots on the rays forming two longitudinal dark stripes, becoming more or less dusky posteriorly ; in half-grown and adult specimens the whole dorsal fin is dusky, with a narrow white margin ; anal pale anteriorly, dusky posteriorly (young), or wholly dusky or blackish, the rays sometimes tipped with white (adult); caudal dusky or blackish ; pectorals pale ; pelvics blackish on the membrane between the inner rays. Hab. South Georgia. This species was described by Fischer from a specimen 490 mm. in total length. He mentions 2 whole specimens and 4 heads in the Zoological Museum at Hamburg (Nos. 3916, 385s). Parachaenichthys charcoti (Vaillant). Chaenichthys charcoti, Vaillant, 1906, Bull. Mus. Paris, 1906, No. 5, p. 247; 1906, Exped. Antarct. Frarif. (1903-1905), Poiss., p. 39, text-fig. i. St. 1873. 13. xi. 36. 61° 20-8' S, 54° 04-2' W. Rectangular dredge, with Russell net, 200-137 m. : I specimen, 290 mm. Close to the preceding species, but with a somewhat shorter snout, and the inter- orbital width about ii| in length of head. Maxillary extending to or nearly to level of anterior margin of eye ; operculum with 5 or 6 radiating ridges ending in spinous points, the uppermost with an antrorse hook; 13 gill-rakers on lower part of anterior arch. 114 tubular scales in upper lateral line; no tubular scales in lower lateral line, which is represented by a series of depressed, non-perforate scales running along middle of hinder part of body. Dorsal 42. Anal 30. Pectoral with about 21 rays. Yellowish- brown; head and anterior part of body with numerous small dark spots and rings; hinder part of body with incomplete and irregular dark cross-bars; dorsal and anal plain, the latter with a pale margin ; caudal blackish, with a narrow pale hinder edge ; small dark spots at bases of pectorals ; pelvics dusky. Fig. 40. Parachaenichthys charcoti. x i. 66 DISCOVERY REPORTS Hab. Graham Land and neighbouring islands. The type of this species consists of a head, 145 mm. in length (measured to tips of opercular spines), and a portion of the tail-end of the body, the latter very poorly preserved, in the Paris Museum (No. 06 — 144). As the jaws and opercles of this specimen are fully expanded, as shown in Vaillant's figure, the general appearance of the head appears at first sight unlike that of a Parachaenichthys. I have carefully examined the type, however, and have little doubt that it belongs to the same species as the much smaller fish obtained by the ' Discovery II '. The type is from Booth Wandel Island, at a depth of 30 metres. Genus PSILODRACO Psilodraco, Norman, 1937, Ann. Mag. Nat. Hist., (10) xx, p. 475. Type Psilodraco breviceps, Norman. Body naked, subcylindrical anteriorly, compressed posteriorly; three lateral lines, consisting of series of imperforate scales more or less embedded in the skin. Head not much depressed; snout not produced, bluntly pointed; mouth moderately pro- tractile; a band of villiform teeth in the upper jaw; each praemaxillary with a group of canine teeth anteriorly; teeth in lower jaw in several rows anteriorly, becoming uniserial laterally. Gill-membranes forming a fold across the isthmus; 6 branchio- stegals ; gill-rakers reduced to a few vestiges near the angles of the arches ; operculum with a fairly strong spine with a hooked branch; suboperculum without spine. No spinous dorsal fin. Pelvic fins short, with the middle rays longest. Vertebrae 48 (16 + 32). A single species. South Georgia. This genus is clearly related to Gymnodraco, and examination of a skeleton prepared for me by Mr W. R. Sherrin reveals no important osteological differences. Psilodraco differs from Gymnodraco mainly in the shape of the head and body, the form of the jaws and teeth, and in the presence of three lateral lines. Psilodraco breviceps, Norman. Psilodraco breviceps, Norman, 1937, t.c, p. 476. St. MS 68. 2. iii. 26. East Cumberland Bay, South Georgia. Large rectangular net, 220-247 i"-- I specimen, 128 mm. St. MS 71. 9. iii. 26. East Cumberland Bay, South Georgia. Small beam trawl, 110-60 m.: I specimen, 36 mm. St. 39. 25. iii. 26. East Cumberland Bay, South Georgia. Large otter trawl, 179-235 m.: 22 specimens, 36-195 mm. (holotype, 190 mm.). St. 42. I. iv. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 120-204 m.: 12 specimens, 48-137 mm. St. 45. 6. iv. 26. 27 miles S 85° E of Jason Light, South Georgia. Large otter trawl, 238- 270 m.: 22 specimens, 90—180 mm. St. 123. 15. xii. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 230- 250 m. : 6 specimens, 70-115 mm. BATHYDRACONIDAE 67 St. 142. 30. xii. 26. East Cumberland Bay, South Georgia. Large otter trawl, 88-27301.: 3 specimens, 137-185 mm. St. 149. 10. i. 27. Mouth of East Cumberland Bay, South Georgia. Large otter trawl, 200- 234 m.: 2 specimens, 123, 145 mm. Fig. 41. Psilodraco breviceps. Holotype. x J. Depth of body 4! to 5I in the length, length of head about 3. Snout about as long as eye, diameter of which is 3 to nearly 4 in length of head ; interorbital width 8. Lower jaw projecting; maxillary extending to below anterior | or middle of eye. Upper lateral line extending to below last rays of dorsal or not as far; middle lateral line extending forward to level of middle or posterior part of pectoral ; lower lateral line of varying length, running parallel with the base of the anal fin. Dorsal 27-30. Anal 27-29. Pectoral with 25 to 27 rays, nearly as long as head, extending to above anterior part of anal ; pelvics much shorter, not reaching vent. Brownish above, more or less silvery on sides and beneath, without any definite markings, but with the silvery parts speckled with close-set dark dots; distal part of snout and tip of lower jaw usually darker ; fins usually all pale, the caudal and occasionally the other fins sometimes a little dusky. Hab. South Georgia. Genus GYMNODRACO Gymnodraco, Boulenger, 1902, Rep. Coll. Nat. Hist. 'Southern Cross', p. 186; Regan, 1913, t.c, p. 283. Related to Psilodraco, but with the body depressed anteriorly, compressed posteriorly. Two lateral lines. Head depressed; snout produced, pointed; mouth non-protractile; jaws with curved compressed teeth, close-set in a single series and with large anterior canines, those of the mandible exposed in front of the snout. Operculum with a strong spine with a hooked branch; suboperculum with a short spine. Vertebrae 48 (20 + 28). A single species. Coasts of the Antarctic Continent. Gymnodraco acuticeps, Boulenger. Gymnodraco acuticeps, Boulenger, 1902, t.c, p. 186, pi. xvii; Pappenheim, 1912, Deutsche Siidpol.-Exped., xni, Zool. v, p. 176, pi. ix, fig. 4; Regan, 1913, t.c, p. 284. Depth of body 8 in the length, length of head about 3. Snout as long as postorbital part of head ; diameter of eye 5 to 6 in length of head, interorbital width 6 to 7. Lower 9-2 68 DISCOVERY REPORTS jaw Strongly projecting; maxillary extending to below anterior margin of eye. Upper lateral line extending from the gill-opening to about the vertical of the origin of the dorsal, lower lateral line from below the extremity of the upper to the base of the caudal fin, running along the middle of the caudal region. Dorsal 28-30. Anal (22) 24-26. Pectoral with 21 or 22 rays, truncated, J as long as head; pelvics much shorter, not nearly reaching vent. Caudal truncate. Brownish-olive, the lower parts whitish ; large dark spots on head and body; fins dusky. Fig. 42. Gymnodraco acuticeps. x \. Hab. Wilhelm Land ; South Victoria Land. The types of the species, 5 specimens up to 300 mm. in total length, are from Cape Adare, 7 to 14 metres: two of these (200 and 300 mm.) are in the collection of the British Museum. The ' Gauss ' obtained a further example, 200 mm. long, from the Winter Quarters of the expedition in Wilhelm Land. Family CHAENICHTHYIDAE Body naked. Snout produced and depressed, spatulate. Mouth non-protractile; palatine in great part ligamentous; pterygoid slender; no mesopterygoid. Gill-mem- branes very narrowly united to isthmus or forming a fold across it ; operculum usually with radiating ridges ending in simple or branched spines. Spinous dorsal fin present. Ribs not ossified. Nine genera, all of which occur in, and eight of which are confined to, the Antarctic Zone. Key to the Genera I. Middle rays of pelvic fin longest. A. Lateral line without bony plates. I. Spinous dorsal of 8 to 15 spines; operculum with at least 3 spines. a. Two lateral lines (upper and middle); dorsal fins subcontinuous. a. Dorsal IX-X, 32-40; anal 31-39; no rostral spine; gill-rakers short, but well- developed ... ... I- Champsocephalus. p. Dorsal (IX) XII-XV, 26-31 ; anal 24-27; a rostral spine; gill-rakers vestigial or absent 2. Pagetopsis. CHAENICHTHYIDAE 69 b. Three lateral lines; dorsal fins well separated; dorsal VIII-X, 29-31; anal 28-31 ; gill-rakers short, but well-developed ... 3. Pseudochaenichthys. 2. Spinous dorsal of 3 spines; operculum with a single broad, flat spine; two lateral lines (upper and lower) 4. Dacodraco. B. Lateral line with bony plates; a rostral spine; dorsal VII-VIII, 30-34; anal 28-33; dorsal fins separated by an interspace; two lateral lines (upper and middle). 5. Chaenichthys. II. Two outer soft rays of pelvic fin longest. A. Sub- and inter-operculum not spinate; rostral spine reduced or absent. 1. Pelvic fins comparatively short, with the rays branched or bifid (except in young); spinous dorsal of 7 or 8 spines, separated from soft dorsal by a short interspace; two lateral lines 6. Chaenocephalus. 2. Pelvic fins long, with the rays simple; spinous dorsal of 3 to 5 spines, separated from soft dorsal by a long interspace; three lateral lines ... ... 7. Cryodraco. B. Sub- and inter-operculum bearing a pair of spines behind angle of praeoperculum; rostral spine well-developed; three lateral lines. 1. Pelvics I 5; gill-rakers reduced to a few vestiges near the angles of the arches. 8. Chionodraco. 2. Pelvics I 4; gill-rakers developed as dentigerous knobs or patches 9. Chaenodraco. Genus CHAMPSOCEPHALUS Champsocephaliis, Gill, 1862, Proc. Acad. N.S. Philad., (1861), p. 509; Regan, 1913, Trans. R. Soc. Edinburgh, XLix, p. 285. Type Chaenichthys eso.x, Giinther. Body naked, elongate; two lateral lines (upper and middle), without bony plates. No rostral spine ; eye nearly in middle of length of head. Jaws with rather narrow bands of small, sharp teeth, forming only two series laterally ; lower jaw not projecting. Gill- rakers short, but well developed on all the branchial arches, dentigerous; operculum with 3 radiating ridges terminating in spinous points. Spinous dorsal well developed, its base less than | that of the soft dorsal, with which it is almost continuous; pelvics comparatively short, with the rays normally branched, the middle ones the longest. Two species, of which one occurs in the Antarctic Zone. Patagonian Region; South Georgia; Kerguelen. Champsocephalus gunnari, Lonnberg. Champsocephaliis gtinnari, Lonnberg, 1905, Wiss. Ergebn. Schwed. Siidpol.-Exped., v (6), p. 37; 1906, K. Svensk. Vet.-Akad. Handl., XL (5), p. 96; Regan, 1913, t.c, p. 285, pi. x, fig. 2; Norman, 1937, Rep. B.A.N.Z. Antarct. Res. Exped., Ser. B, i, p. 65. St. 39. 25. iii. 26. Cumberland Bay, South Georgia. Large otter trawl, 179-235 m.: 14 specimens, 235-275 mm. St. 42. I. iv. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 120-204 m. : 2 specimens, 250, 325 mm. St. 45. 6. iv. 26. 2-7 miles S 85" E of Jason Light, South Georgia. Largeotter trawl, 270-238 m.: 2 specimens, 205, 258 mm. St. 123. 15. xii. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 230- 250 m.: 5 specimens, 220-270 mm. 70 DISCOVERY REPORTS St. 140. Stromness Harbour to Larsen Point, South Georgia. Large otter trawl, 122-136111.: I specimen, 155 mm. Depth of body 6 to yl in the length, length of head 3^ to 3I. Snout about as long as postorbital part of head; diameter of eye 4^ to 5^, interorbital width 3I to 4I in length of head. Maxillary extending to below anterior J or ^ of eye ; upper and middle opercular spines only free distally, appearing as a single bifid (sometimes trifid) spine. Dorsal IX-X, 37-40. Anal 36-39. Pectoral with 25 to 27 rays, about | length of head, extending to above anterior rays of anal ; pelvics about f length of head. Yellowish- brown or plumbeous, with a series of darker cross-bars ; dorsal and anal fins usually more or less blackish ; other fins pale or dusky. Fig. 43. Champsocephalus gunnari. x J. Hah. South Georgia; Kerguelen. This species was originally described from 9 specimens, 322 to 436 mm. in total length, all from Cumberland Bay, South Georgia, 75 to 195 metres. C. gunnari may be readily distinguished from the Patagonian C. esox by the deeper body, rather shorter head, larger eye, shorter maxillary, form of the opercular spines ; and higher number of dorsal and anal rays, as well as by the difference in coloration. Recently, the B.A.N.Z. Antarctic Research Expedition obtained 2 examples (235 and 300 mm.) of this species from Kerguelen at a depth of 91 metres, as well as remains of other specimens from the stomachs of fishes and penguins. Genus PAGETOPSIS Pagetopsis, Regan, 1913, t.c, p. 286. Type Champsocephalus macropteriis, Boulenger. Body naked, moderately elongate; two lateral lines, without bony plates. An antrorse curved rostral spine; eye behind middle of length of head. Teeth in jaws small, sharp, biserial; lower jaw slightly projecting; gill-rakers vestigial or absent; operculum armed with 3 or 4 radiating ridges terminating in spinous points, the uppermost with an antrorse hook. Spinous dorsal well developed, its base more than ^ that of soft dorsal ; pelvics rather long, the rays bifid or sUghtly branched. A single species. Coasts of Antarctic Continent. CHAENICHTHYIDAE 71 Pagetopsis macropterus (Boulenger). Champsocephalus macropterus, Boulenger, 1907, Nat. Antarct. Exped. Nat. Hist., 11, Fish., p. 3, pi. ii; Pappenheim, 1912, Deutsche SiidpoL-Exped., xiii, Zool. v, p. 174; Roule, Angel and Despa.x, 1913, Deux. Exped. Antarct. Frang. (1908-1910), Poiss., p. 13. Pagetopsis macropterus, Regan, 1913, t.c, p. 286; 1914, Rep. Brit. Antarct. ('Terra Nova') Exped. 1910, Zool. I (i), p. 11; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, III (i), Fishes, p. 37, fig. 8; Norman, 1937, t.c, p. 76. St. 1660. 27. i. 36. 74° 46-4' S, 178° 23-4' E. Large otter trawl, 351m.: 3 specimens, 185- 235 mm. Depth of body 4 to 5I in the length, length of head 2f to 2|. Snout nearly i length of head; diameter of eye about 5, interorbital width 4 to more than 4! in length of head. Maxillary extending to below anterior | or | of eye. Dorsal (IX) XII-XV, (26) 27-31. Anal 24-27. Pectoral with 22 or 23 rays, | to f length of head, extending to above vent or origin of anal ; pelvics a little longer, reaching vent or not quite as far. Caudal subtruncate or a little rounded. Dark spots and vermiculations on head; irregular double cross-bars on body; spinous dorsal more or less blackish, the spines tipped with white; soft dorsal dusky; anal and caudal mostly pale; pectorals pale; pelvics blackish. Fig. 44. Pagetopsis macropterus. x J. Hab. Graham Land ; Enderby Land ; Mac-Robertson Land ; Wilhelm Land ; Queen Mary Land ; Adelie Land ; South Victoria Land. This species was originally described from 1 1 specimens, 65 to 240 mm. long, obtained from the stomach of a Weddell Seal near Cape Armitage, Ross Island: of these, 6 examples (160 to 250 mm.) are in the collection of the British Museum. Genus PSEUDOCHAENICHTHYS Pseudochaenichthys, Norman, 1937, Ann. Mag. Nat. Hist., (10) xx, p. 476. Type Pseudochae- nichthys georgianus, Norman. Body naked, moderately elongate; three lateral lines, without bony plates. An antrorse curved rostral spine ; eye a little behind middle of length of head. Jaws with bands of small, curved, sharp teeth; lower jaw projecting; gill-rakers very short, knob- 72 DISCOVERY REPORTS like, dentigerous; operculum with a group of 4 or 5 spines, the uppermost with an antrorse hook. Spinous dorsal well developed, its base not more than h that of soft dorsal, from which it is separated by an interspace ; pelvics rather long, with the rays normally branched, the middle ones the longest. A single species. South Georgia. Pseudochaenichthys georgianus, Norman. Pseiidochaenichthys georgianus, Norman, 1937, t.c, p. 476. St. 39. 25. iii. 26. East Cumberland Bay, South Georgia. Large otter trawl, 179-235 m.: 2 specimens, 285, 350 mm. St. 45. 6. iv. 26. 27 miles S 85° E of Jason Light, South Georgia. Large otter trawl, 270- 238 m.: 4 specimens, 350-480 mm. (holotype, 480 mm.). St. 123. 15. xii. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 230- 250 m. : 6 specimens, 155-175 mm. St. 142. 30. xii. 26. East Cumberland Bay, South Georgia. Large otter trawl, 88-273 m.: I specimen, 153 mm. St. 148. 9.1.27. Off Cape Saunders, South Georgia. Large otter trawl, 132-148 m.: 3 specimens, 430-465 mm. Depth of body 4I to nearly 6 in the length, length of head 2\ to 2|. Snout nearly \ length of head ; diameter of eye 5 (young) to 7, interorbital width rather more than 4 in length of head. Supraorbital edges roughened (except in young), but only slightly raised ; upper surface of head not rugose ; maxillary extending to below middle of eye ; about 18 gill-rakers on lower part of anterior arch. Dorsal VIII-X, 29-31; sixth or seventh spine longest. Anal 28-31. Pectoral with 23 or 24 rays, | to | length of head, extending to above origin or anterior rays of anal ; pelvics as long as or rather longer than pectorals. Caudal subtruncate or a little rounded. Brownish, the upper parts of the body irregularly marked with darker in the young, but these markings tend to disappear in the adults, which are nearly uniformly coloured; spinous dorsal and pelvics blackish ; other fins pale or a little dusky. Fig. 45. Pseudochaenichthys georgianus. Holotype. Hab. South Georgia. CHAENICHTHYIDAE Genus DACODRACO 73 Dacodraco, Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, iii (i), Fishes, p. 35. Type Dacodraco hunteri, Waite. Body naked, moderately elongate ; two lateral lines (an upper and a lower), without bony plates. No rostral spine ; eye behind middle of length of head. Jaws each with a single row of small teeth, with an inner series of large spaced canines ; lower jaw slightly projecting; no gill-rakers; operculum with a single broad, flat spine. Spinous dorsal with only 3 spines, well separated from soft dorsal ; pelvics of moderate length, with the rays normally branched, the middle ones the longest. A single species. Coasts of the Antarctic Continent. Dacodraco hunteri, Waite. Dacodraco hunteri, Waite, 1916, t.c, p. 36, pi. ii, fig. 2, text-fig. 7. } Pagetodes antarcticm (part), Pappenheim, 1912, Deutsche Sudpol. -Exped., xni, Zool. v, P- 175- Depth of body 6| in the length, length of head zf . Snout about \ the length of the head ; diameter of eye \\, interorbital width 5I in length of head. Maxillary extending to below middle of eye. Upper lateral line originating above the angle of the oper- culum, rising to below the interdorsal space, and following the profile to the base of the caudal ; lower lateral line commencing above the twelfth ray of the anal and running near to the ventral profile to the base of the caudal. Dorsal III, 32; spines subequal, low, as long as diameter of eye, not higher than the rays of the soft dorsal. Anal 29. Pectoral with 24 rays, extending to above the sixth ray of the anal ; pelvics about as long as pectorals, just reaching the anal. Caudal truncate. Reddish brown, with 7 darker cross-bars on body, the first two of which are slightly oblique ; spinous dorsal black ; other fins dusky, the pectorals being black proximally and the pelvics distally. Fig. 46. Dacodraco hunteri. Holotype. x \. (After Waite.) Hah. Wilhelm Land (?); Queen Mary Land, 610 metres. Known only from 2 specimens, both obtained by the 'Aurora' off Shackleton Ice- shelf (65° 6' S, 96° 13' E). The type was 211 mm. in total length. The specimen of 69 mm., identified by Pappenheim as Pagetodes antarcticus, from the Winter Station of the ' Gauss ', may belong here. 74 DISCOVERY REPORTS Genus CHAENICHTHYS Channichthys, Richardson, 1844, Ann. Mag. Nat. Hist., xiii (June), p. 461. Type Channichthys rhifiocerattis, Richardson. Chaenkthys, Richardson, 1844, Zool. 'Erebus' and ' Terror', (Fish.), P- 12 [emend, pro Chan- nichthys). Chaenichthys, GiJnther, i860, Cat. Fish., 11, p. 249; Regan, 1913, Trans. R. Soc. Edinburgh, XLix, p. 286 (emend, pro Chaenicthys). Related to Chompsocephaliis, but there is a rostral spine, the teeth are in broader bands, and the lateral line is provided with bony plates. Gill-rakers short, dentigerous. Dorsal fins well separated. Two species. Kerguelen. Chaenichthys rhinoceratus, Richardson. Channichthys rhinoceratus, Richardson, 1844, Ann. Mag. Nat. Hist., xin (June), p. 461. Chaenicthys rhinoceratus, Richardson, 1844, Zool. 'Erebus' and 'Terror', (Fish.), p. 12, pi. vi, figs. '1-3; Gill, 1876, Bull. U.S. Nat. Mus., in, p. 41. Chaenichthys rhinoceratus (part), Giinther, i860. Cat. Fish., 11, p. 249. Chaemethvs rhitioceratus, Studer, 1879, Arch. Naturg., xlv (i), p. 131. Chaenichthys rhinoceratus, Giinther, 1880, Shore Fish. 'Challenger', p. i6; Pappenheim, 1912, Deutsche Siidpol.-Exped., xiii, Zool. v, p. 173; Regan, 191 3, Trans. R. Soc. Edinburgh, XLix, p. 286; Monod and DoUfus, 1932, Bull. Soc. Zool. Fr., lvii, p. 74; Norman, 1937, Rep. B.A.N.Z. Antarct. Res. Exped., Ser. B, i, p. 65. Fig. 47. Chaenichthys rhinoceratus. x J. Depth of body 5I to more than 7 in the length, length of head 2| to nearly 2f. Diameter of eye 5J to 7^, interorbital width 5 to b\ in length of head. Maxillary ex- tending to below middle of eye or beyond in adults. Head moderately or rather strongly rugose, the supraorbital edges usually not much raised. 69 to 84 plates in upper lateral line; a'few plates on middle of side. Dorsal VII-VIII, 32-34; second and third spines longest, thence decreasing rapidly in length. Anal 30-33. Pectoral with 19 to 22 rays, extending to above vent or origin of anal; pelvics as long as or a little longer than pectorals, not reaching vent. Caudal rounded. Brownish, with darker spots and reti- culations ; spinous dorsal blackish ; other fins pale or dusky. CHAENICHTHYIDAE 75 Hab. Kerguelen. The type of the species, 450 mm. in total length, was taken in the kelp close to the shore, the fish being caught with the hook. In addition to this specimen, and a smaller one (175mm.) collected by the 'Challenger', several examples (300 to 510mm.) collected by the B.A.N. Z. Antarctic Research Expedition are included in the above description. Chaenichthys rugosus, Regan. Chaenichthys rhinoceratiis (part), Giinther, i860, t.c, p. 249. Chaenichthys rhinoceratiis (non Richardson), Giinther, 1879, Phil. Trans. R. Soc. London, CLXVin (extra vol.), p. 166. Chaenichthys rugosus, Regan, 1913, t.c, p. 287; Norman, 1937, t.c, p. 66. Very closely related to C. rhinoceratus, but diameter of eye y^ to 8 in length of head ; maxillary extending to or nearly to below middle of eye. Head rougher and supra- orbital edges more elevated. 62 or 63 plates in upper lateral line; a nearly continuous series of plates on middle of side. Dorsal VIII, 30-31 ; third and fourth spines longest, fifth about as long as first. Anal 28-29. Pectoral with 18 rays, scarcely extending to above vent ; pelvics longer. Nearly uniformly brownish. Hab. Kerguelen. The type of the species, 400 mm. in total length, has the head and body much distorted. The B.A.N.Z. Antarctic Research Expedition obtained a second example, 245 mm. in total length, which has been included in the above description. A stuffed specimen in the British Museum collection may also belong to this species. Genus CHAENOCEPHALUS Chaenocephalus, Regan, 1913, t.c, p. 287. Type Chaenichthys aceratus, Lonnberg. Body naked, elongate; two lateral lines, without distinct bony plates. A small prominence at anterior end of ethmoid ; eye somewhat behind middle of head. Jaws with small sharp teeth forming rather broad bands, there being several series even at the sides; lower jaw not projecting; gill-rakers absent except for 3 or 4 very short ones below the angle of the first arch ; operculum with 3 or 4 radiating ridges ending in spines, the uppermost with two or more points. Spinous dorsal well developed, its base about \ that of the soft dorsal, from which it is separated by an interspace; pelvics comparatively short, with the two outer rays the longest, enveloped in thick skin, but bifid, the others normally branched. A single species. Palmer Archipelago ; South Orkneys ; South Georgia. Chaenocephalus aceratus (Lonnberg). Chaenichthys aceratus, Lonnberg, 1906, K. Svensk. Vet.-Akad. Handl., xl (5), p. 97. Chaenocephalus aceratus, Regan, 1913, t.c, p. 288, pi. xi. St. MS 68. 3. ii. 26. East Cumberland Bay, South Georgia. Large rectangular net, 220-247 m. : I specimen, 515 mm. 76 DISCOVERY REPORTS St. 39. 25. iii. 26. East Cumberland Bay, South Georgia. Large otter trawl, 179-23501.: 5 specimens, 255-420 mm. St. 42. 1. iv. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 120-204 m. : 2 specimens, 255, 290 mm. St. 45. 6. iv. 26. 27 miles S 85° E of Jason Light, South Georgia. Large otter trawl, 270-238 m.: I specimen, 495 mm., with cranium of another. St. 123. i5.xii. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 230- 250 m.: 5 specimens, 147-160 mm. St. 149. 10. i. 27. Mouth of East Cumberland Bay, South Georgia. Large otter trawl, 200- 234 m. : 8 specimens, 135-160 mm. St. 167. 20. ii. 27. Off Signy Island, South Orkneys. Net (7 mm. mesh) attached to back of trawl, 244-344 m.: ' specimen, 210 mm. St. 181. 12. iii. 27. Schollaert Channel, Palmer Archipelago. Large otter trawl, 160-335 m.: I specimen, 580 mm. Fig. 48. Chaenocephalus aceratus. x J. Depth of body 5 to 7 in the length, length of head 2J to 2f . Snout a little less than I length of head ; diameter of eye about equal to interorbital width, and 5 to 6 in length of head. Maxillary extending to below middle of eye or beyond (not so far in young). Supraorbital edges raised and more or less roughened. Dorsal VII-VIII, 38-40. Anal 37-39. Pectoral with 25 or 26 rays, | to | length of head, extending to above origin or anterior rays of anal; pelvics in young longer than head and reaching to beyond middle of anal, in adult much shorter, subequal to pectorals, not reaching vent. Caudal a little rounded. Greyish or pale brownish, with 4 or 5 dark cross-bars, the first from spinous dorsal to base of pectoral, the second downwards from origin of soft dorsal, the others less regular and sometimes with narrower bars developed between them; spinous dorsal blackish or dusky ; other fins more or less pale. Hab. Palmer Archipelago ; South Orkneys ; South Georgia. The type of this species was 522 mm. in total length, and was caught in Cumberland Bay at a depth of about 5 to 7 metres. Genus CRYODRACO ? Pagetodes, Richardson, 1844, Zool. 'Erebus' and ' Tenor' (Fish.), p. 15. (No type.) Cryodraco, Dollo, 1900, Bidl. Acad. roy. Belg. {Classe ScL), No. 2, p. 129; Regan, 1913, t.c, p. 288. Type Cryodraco antarcticus, Dollo. CHAENICHTHYIDAE 77 Related to Chaenocephaliis, but the rays of the pelvic fins are simple, the two outer ones enlarged and prolonged ; an additional lateral line at the base of the anal fin. Two species. Coasts of the Antarctic Continent. Cryodraco antarcticus, DoUo. ? Pagetodes, Richardson, 1844, Zool. 'Erebus^ and ^ Terror^ (Fish.), p. 15, pi. viii, fig. 3. Cryodraco antarcticus, Dollo, 1900, Bull. Acad. roy. Belg. (Classe Sci.), No. 2, p. 129; 1904, Res. Voy. 'Belgica', Zool., Poiss., p. 20, pi. i, pi. v, fig. 7; Regan, 1913, Trans. R. Sac. Edinburgh, XLix, p. 288; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, ni (i). Fishes, p. 39; Norman, 1937, t.c, p. 76. Cryodraco atkinsoni, Regan, 1914, Ann. Mag. Nat. Hist., (8) xni, p. 13 ; 1914, Rep. Brit. Antarct. {'Terra Nova') Exped. 1910, Zool. i (i), p. 11, pi. vii, fig. 2. St. 167. 20. ii. 27. Off Signy Island, South Orkneys. Net (7 mm. mesh) attached to back of trawl, 244-344 rn. : 2 specimens, 220, 230 mm. St. 1660. 27. i. 36. 74° 46-4' S, 178° 23-4' E. Large otter trawl, 351m.: 2 specimens, 245, 250 mm. Depth of body 7J to 8| in the length, length of head 3! to 3^. Snout | or nearly ^ length of head ; diameter of eye 4 to 5 J, interorbital width 4I to more than 5 in length of head. Lower jaw not projecting; maxillary extending nearly to below middle of eye. Supraorbital edges not raised ; operculum with two divergent ridges, each ending in a spine, the upper one being bi- or tri-fid. Dorsal III-V, 40-44; the membrane between the spines sometimes missing (? torn away) in young; first and second spines subequal or first a little longer than second. Anal 42-46. Pectoral with 23 to 25 rays, about f length of head, extending to above anterior rays of anal ; pelvics from f (adult) to twice (young) length of head, extending in adults to below middle of dorsal fin or not as far, in young nearly to end of dorsal. Caudal a little emarginate. Greyish or pale brownish, with dark spots on head and with 6 or 7 dark cross-bars on body, sometimes incomplete and sometimes with irregular bars or blotches between them; spinous dorsal blackish ; pelvics and distal part of caudal dusky ; other fins pale. Fig. 49. Cryodraco antarcticus. x |. Hab. Graham Land; South Orkneys; Mac-Robertson Land; Queen Mary Land; South Victoria Land. The type of the species, 200 mm. in total length, was from 71° 18' S, 88° 2' W, at a depth of 450 metres. The type of C atkinsotn, 293 mm. in total length, is from the 78 DISCOVERY REPORTS Ross Sea (74° 25' S, 179° 3' E), at a depth of 296 metres. I have elsewhere given reasons for following Waite in regarding these species as identical. The name Pogetodes was given by Richardson, without trivial name, to a fish 160 mm. in total length, which was "thrown up by the spray in a gale of wind, against the bows of the Terror, and frozen there", at 77° 10' S, 178^ W. The specimen was unfor- tunately appropriated by the ship's cat before a description or detailed drawing could be made. The rough sketch published by Richardson depicts a fish not unlike a Cryo- draco, but, as Regan has pointed out, in some characters it shows more resemblance to Pagetopsis. The name Pagetodes should, therefore, be eliminated. Cryodraco pappenheimi, Regan. Pagetodes antarcticus {non Dollo), Pappenheim, 1912, Deutsche Siidpol.-Exped. ,xin, Zool. v, p. 175. Cryodraco pappenheimi, Regan, 1913, t.c, p. 289. Closely related to C. antarcticus, but length of head 2f in the length of the fish. Snout \ length of head ; diameter of eye 5, interorbital width 4 in length of head. Dorsal V, 45. Anal 39. Pelvics reaching fourteenth ray of dorsal (the prolonged rays perhaps not entire). Hab. Wilhelm Land. The type of the species, 168 mm. in length to base of caudal, was obtained from the stomach of a seal at the Winter Station of the ' Gauss'. Genus CHIONODRACO Chionodraco, Lonnberg, 1906, K. Svensk. Vet.-Akad. Handl., xl (5), p. 99; Regan, 1913, t.c, p. 288; 1914, Rep. Brit. Antarct. {'Terra Nova') Exped. 1910, Zool. i (i), p. 12. Type Chaenichthys rhinoceratus hamatus, Lonnberg. Body naked, elongate; three lateral lines, without bony plates. A rostral spine; eye somewhat behind middle of head ; supraorbital ridges crenulated. Teeth bi- or tri-serial ; gill-rakers absent except for a few vestiges near the angles of the arches ; operculum with a group of 3 to 5 spines, the uppermost with 2 to 4 points ; sub- and inter-oper- culum bearing a pair of spines just behind angle of praeoperculum. Spinous dorsal well-developed, of five to seven spines ; pelvics comparatively short, of a spine and five branched or bifid rays, the two outer the longest and (in adult) enveloped in thick skin. Skeleton essentially similar to that of Champsocephaliis. Vertebrae 64. Two species. Coasts of the Antarctic Continent. Chionodraco kathleenae, Regan. Chionodraco kathleenae, Regan, 1914, Ann. Mag. Nat. Hist., (8) xni, p. 13; 1914, Rep. Brit. Antarct. {'Terra Nova') Exped. 1910, Zool. i (i), p. 12, pi. viii; Waite, 1916, Austral. Antarct. Exped. Set. Rep., Ser. C, ni (i), Fishes, p. 41 ; Norman, 1937, t.c, p. 77, figs. 7 b, 9. St. 1658. 26. i. 36. 76° 09-6' S, 168° 40' E. Rectangular dredge, with Russell net, 520 m.: 1 specimen, 270 mm. Depth of body 5 to 6 J in the length, length of head 2f to 3. Snout a little less than 1 length of head; diameter of eye 4I to 6^, interorbital width 3^ to about 4 in length CHAENICHTHYIDAE 79 of head. Rostral spine erect or retrorse. Jaws equal anteriorly or lower jaw projecting ; maxillary extending to below anterior part or nniddle of eye. Dorsal V-VII, 38-45 (51) ; third or fourth spines generally longest ; spinous dorsal separated from soft dorsal by an interspace which varies from less than | to about | of the length of the base of the former. Anal (33) 34-38. Pectoral with 22 to 24 rays, | to nearly f length of head, extending to above vent or anterior rays of anal; pelvics I 5, | to | length of head, reaching vent, origin of, or anterior rays of anal in young, not reaching vent in adult. Caudal subtruncate or rounded. Pale brownish or greyish ; head and body with dark spots and bars ; spinous dorsal blackish ; other fins pale or more or less dusky. Fig. 50. Chionodraco kathleenae. x J. Hub. Mac-Robertson Land; Queen Mary Land; South Victoria Land. The types of the species, 250 to 500 mm. in total length, are from the Ross Sea (74° 25' S, 179° 3' E), 296 metres, and McMurdo Sound, 187 to 375 metres. The B.A.N.Z. Antarctic Research Expedition obtained 16 examples, 275 to 470 mm. in total length, from off Mac-Robertson Land at a depth of 219 metres: these specimens have been included in the description of the species. Two young specimens (70 and 71 mm.) were collected by the same expedition with mid-water nets hauled obliquely from 169 metres to the surface. The eye is proportionately smaller in these young specimens, the diameter being nearly 6 in the length of the head, the pelvic fin-rays are simple, and the pelvic fins themselves are blackish in their distal parts. Chionodraco hamatus (Lonnberg). Chaenichthys rhinoceratus subsp. hamatus, Lonnberg, 1905, Wiss. Ergebn. Schwed. Siidpol.- Exped., V (6), p. 47. Chionodraco hamatus, Lonnberg, 1906, K. Svensk. Vet.-Akad. Handl., XL (5), p. 99; Regan, 1913, Trans. R. Soc. Edinburgh, xlix, p. 288; Norman, 1937, t.c, p. 77, figs. 7 A, 8. Closely related to C. kathleenae, but with a somewhat larger eye, diameter of which is 4§ in length of head; interorbital region broader, width 3f in length of head. Dorsal VII, 37. Anal 33. Pelvics shorter, \ length of head. Hab. Graham Land; South Shetlands. The type of the species, 330 mm. in total length, was from south of Snow Hill, Graham Land, at a depth of 125 metres. Through the kindness of Professor Dr H. 8o DISCOVERY REPORTS Rendahl, I have been able to examine this specimen, and to compare it with examples of C. kathleenae of similar size. Two damaged specimens, 125 and 135 mm. in total length, from Deception Island, South Shetlands (taken from the stomach of a Blue Whale) ; and two others, 67 and 69 mm. in total length, from the stomach of a Fin Whale, captured in the Weddell Sea (65° 04' S, 54° 12' W), appear to belong to this species. As in the young of C. kathleenae, the eye is proportionately smaller in these small examples, the diameter being about 5 in the length of the head in specimens of 67 and 69 mm. The pelvic fin-rays are simple, and the pelvic fins are entirely pale. Genus CHAENODRACO Chaenodraco, Regan, 1914, Ann. Mag. Nat. Hist., (8) xiii, p. 13; 1914, Rep. Brit. Antarct. {'Terra Nova') Exped. 1910, Zool. i (i), p. 12. Type Chaenodraco wilsoni, Regan. Closely related to Chionodraco, but the supraorbital ridges are not crenulated, the gill-rakers are developed as dentigerous prominences, and the pelvic fins are each formed of a spine and only four rays. A single species. Coasts of the Antarctic Continent. Chaenodraco wilsoni, Regan. Chaenodraco wilsoni, Regan, 1914, t.c, p. 13; 1914, t.c, p. 12, pi. ix, fig. i; Norman, 1937, t.c, p. 79. ? Chaenodraco fasciatus, Regan, 1914, t.c, p. 14; 1914, t.c, p. 13, pi. ix, fig. 2. Fig- 51- Chaenodraco wilsoni. x \. Depth of body 6 to 7 in the length, length of head z\ to 3^. Snout less than \ length of head; diameter of eye 4 (young) to 4^, interorbital width 3f to 4I in length of head. Rostral spine antrorse. Jaws equal anteriorly or lower a little projecting; maxillary extending to below anterior J or ^ of eye. Operculum with a group of 4 or 5 spines, the uppermost bi- or tri-fid; 12 to 14 gill-rakers on lower part of anterior arch. Dorsal VI-VIII, 38-42; spinous dorsal separated from soft dorsal by a narrow interspace (young), or continuous with it at the base (adult). Anal 32-35. Pectoral with 22 to 24 rays, f to nearly | length of head, extending to above anterior rays of anal ; pelvics I 4, as long as or nearly as long as head, extending to anterior part of anal or beyond. Caudal subtruncate or a little rounded. Pale brownish, with dark spots on the head CHAENICHTHYIDAE 8i and with 5 blackish cross-bars on the body; sometimes some additional blotches or bars between the cross-bars ; a large black spot on the spinous dorsal ; other fins more or less dusky. Hab. Mac-Robertson Land ; South Victoria Land. The type of the species, 250 mm. in total length, is from McMurdo Sound, 187 to 375 metres ; the type of C. fasciatus, 92 mm. in total length, is from McMurdo Sound (77° 13' S, 164° 18' E), 380 metres. I have elsewhere given reasons for regarding these species as probably identical. A number of specimens, 205 to 268 mm. in total length, collected by the B.A.N. Z. Antarctic Research Expedition, are included in the above description. ZOARCIDAE Key to the Antarctic Genera^ I. Pelvic fins present; gill-opening cleft downward at least to middle of base of pectoral. A. Origin of dorsal well behind base of pectoral ... ... ... i. Lycenchelys B. Origin of dorsal above base or anterior part of pectoral 1. Mouth large, with wide lateral cleft; anterior canines in upper jaw and lateral canines in lower; palate toothless 2. Lycodichthys. 2. Mouth moderate, with short lateral cleft; no well-marked canines; teeth on palate. 3. AUSTROLYCICHTHYS. II. No pelvic fins; gill-opening small, above base of pectoral ... ... 4. Melanostigma. Genus LYCENCHELYS Lycenchelys, Gill, 1884, Proc. Acad. N.S. Philad., p. no; Regan, 1913, Trans. R. Soc. Edin- burgh, XLix, p. 242. Type Lycodes muraena, Collett. Form elongate, with the tail long and slender, body naked. Upper jaw projecting; teeth in jaws slender, uni- or bi-serial; teeth on vomer; palatine teeth uniserial. Gill- opening rather wide, cleft downward to lower end of base of pectoral. Origin of dorsal fin well behind base of pectoral ; pelvic fins present. Several species from deep water north of the Equator ; one species from the Antarctic. Lycenchelys antarcticus, Regan. Lycenchelys antarcticus, Regan, 1913, t.c, p. 242, pi. ix, fig. 3. Depth of body 16 in the length, length of head 6. Head as broad as deep, its breadth a little more than \ its length. Snout twice as long as eye, diameter of which is 6 in ^ In 1885, Fischer [Jahrb. Hamburg. Wiss. Anst., u, p. 60, pi. ii, fig. 9) described and figured a new genus and species of Zoarcid fish, Gymnelichthys antarcticus. The description was based upon 4 specimens, no to 220 mm. in total length, said to be from South Georgia. Through the kindness of Miss Erna Mohr, I have been able to examine three of these types, now preserved in the Hamburgisches Zoologisches Museum und Institut: No. 3902, 2 specimens, no and 120 mm.; No. 3942, i specimen, 183 mm. As was the case with Fischer's genus of Cottidae, Sclerocottus schraderi (see Norman, 1935, Copeia, No. 3, p. 141), it seems likely that the type locality is incorrect, and that the fish were never collected at South Georgia. Comparison of these type specimens with material in the British Museum shows that they are referable to a species of Zoarcid common in Arctic seas, namely, Gymnelis viridis (Fabricius). g2 DISCOVERY REPORTS length of head; interorbital width about i6. Muciferous channels of sides of head and lower jaw with large pores. Teeth in jaws rather slender and obtuse, uniserial, biserial near symphysis of lower jaw; teeth on palate acute, wide-set. About no rays in dorsal fin, 9 in caudal, 103 in anal ; origin of anal at a distance behind head equal to the length of the latter, 1 as distant from vertical through origin of dorsal as from that through base of pectoral; pectoral a little more than darker ; fins brownish grey. as long as head. Bluish-grey; head Fig. 52. Lycenchelys antarcticus. Holotype. xi. Hab. Near the South Orkneys. The type of the species, 128 mm. in total length, was obtained by the 'Scotia' at Station 313 (62° 10' S, 41° 20' W), at a depth of 3225 metres. The specimen is pre- served in the Royal Scottish Museum (Bruce collection), and I am indebted to the authorities of the museum for the loan of this fish for examination. Genus LYCODICHTHYS Lj;co&/?%^, • Pappenheim, 191 1, Sitzungsh. Ges. Naturf. Freunde, p. 382; 1912, Deutsche Sudpol.-Exped., xni, Zool. v, p. 180; Regan, 1913, t.c, p. 244. Type Lycodichthys antarcti- cus, Pappenheim. Form moderately elongate; body with rounded scales embedded in the skin. Upper jaw projecting; mouth large, with wide lateral cleft; teeth conical, uniserial; anterior pair in upper jaw enlarged; lateral teeth of lower jaw spaced, canine-like; palate tooth- less. Gill-opening rather wide, cleft downward almost to lower end of base of pectoral. Origin of dorsal fin just behind head, above anterior part of pectoral ; pelvic fins present. A single species from the coasts of the Antarctic Continent. Lycodichthys antarcticus, Pappenheim. Lycodichthys antarcticus, Pappenheim, 191 1, t.c, p. 383; 1912, t.c, p. 180, pi. ix, fig. 6, pi. x, fig. 4; Regan, 1913, t.c, p. 244; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, ni (i). Fishes, p. 14. Depth of body 8 or 9 in the length, length of head 5 to 5*. Diameter of eye 5 to 6 in length of head. Maxillary extending to below posterior margin of eye or a little beyond. Dorsal with 85 to 90 rays, anal with about 65, caudal with about 10; origin of anal i to i\ head-lengths behind head; pectoral I length of head. Yellowish brown; head, body and fins spotted or marbled with darker brown. Fig. 53. Lycodichthys antarcticus. ZOARCIDAE 83 Hab. Wilhelm Land. This species was originally described from 25 specimens, no to 200 mm. in total length, from the Winter Station of the ' Gauss', depth 385 metres: 2 co-types, 160 and 200 mm. in total length, are in the collection of the British Museum. Genus AUSTROLYCICHTHYS Austrolycichthys, Regan, 1913, t.c, p. 244. Type Lycodes brachycephalus, Pappenheim. Form elongate, compressed ; body with rounded scales embedded in the skin. Jaws level anteriorly or upper somewhat projecting; mouth moderate, with short lateral cleft ; teeth conical, in two or more series in both jaws, at least anteriorly ; no well- marked canines; teeth on vomer and palatines. Gill-opening cleft downward at least to middle of base of pectoral. Origin of dorsal fin just behind head, above or a little behind base of pectoral ; pelvic fins present. Three species from Antarctic Seas. Key to the Species I. Gill-opening cleft downward nearly to lower end of base of pectoral; tail from less than i| to about 1 1 times as long as rest of fish ; anal with about 68 to 70 rays ; inside of mouth pale. A. Pores on head not very conspicuous; origin of anal if to if head-lengths behind head; upper jaw a little projecting i- brachycephalus. B. Pores on head very conspicuous; origin of anal a little more than i head-length behind head ; upper jaw strongly projecting ... ... ... ••• ••• 2. concolor. II. Gill-opening cleft downward to a little below middle of base of pectoral ; tail about 2^ times as long as rest of fish; anal with about 90 rays; inside of mouth blackish. 3. bothriocephalus. Austrolycichthys brachycephalus (Pappenheim). Lycodes brachycephalus, Pappenheim, 1912, Deutsche Sudpol.-Exped., xni, Zool. v, p. 179, pi. X, fig. 3. Austrolycichthys brachycephalus, Regan, 1913, i.e., p. 244, text-fig. 2; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, ni (i), Fishes, p. 15, pi. i, fig. i, text-figs. 2, za. St. 1652. 23.1.36. 75" 56-2' S, 178" 35-5' E. Rectangular dredge, with Russell net, 567 m.: I specimen, 280 mm. Depth of body 8 to loi in the length, length of head 5I to 6|. Tail from less than ii to if times as long as rest of fish. Pores on head not conspicuous. Diameter of eye 5 to nearly 6 in length of head. Maxillary extending to below anterior part of eye ; upper jaw a little projecting ; teeth in 2 or 3 series anteriorly, usually becoming uniserial laterally ; vomerine teeth in a group, palatine teeth in 2 rows. Gill-opening cleft down- ward to or nearly to lower end of base of pectoral. Dorsal with about 90 rays, anal with about 70, caudal with 10; origin of anal if to if head-lengths behind head; pectoral | to | length of head. Greyish or brownish; inside of mouth pale. Hab. Wilhelm Land ; Queen Mary Land ; South Victoria Land. This species was originally described from 58 specimens from the Winter Station of 84 DISCOVERY REPORTS the ' Gauss', depth 380 metres, of which 2 examples, 150 and 155 mm. in total length, are in the collection of the British Museum. The ' Aurora ' obtained other examples from off Shackleton Ice-shelf (65° 6' S, 96° 13' E), 600 metres, and off the Western Base (66° 18' S, 94° 58' E), 494 metres. Austrolycichthys concolor (Roule & Despax). Lycodes concolor, Roule and Despax, 1911, Bull. Mus. Paris, 191 1, No. 5, p. 279; Roule, Angel and Despax, 1913, Deux. Exped. Antarct. Franp. (1908-1910), Poiss., p. 19, pi. ii, fig. 3, pi. iii, fig. 6. St. 195. 30. iii. 27. Admiralty Bay, King George Island, South Shetlands. Net (7 mm. mesh) attached to back of trawl, 391 m.: i specimen, 125 mm. Depth of body 9I in the length, length of head about 5|. Tail about if times as long as rest of fish. Pores on head large and conspicuous. Diameter of eye nearly 5 in length of head, nearly equal to interocular width. Maxillary extending to beyond anterior margin of eye; upper jaw strongly projecting; several rows of small teeth in each jaw ; a small group of teeth on the vomer and a single tooth on each palatine. Gill-opening cleft downward to lower end of base of pectoral. Dorsal with about 75 rays, anal with about 70, caudal with about 10; origin of anal a little more than a head-length behind head; pectoral about | length of head. Pale yellowish-brown, the upper parts of the sides irregularly marbled with darker. Fig. 54. Austrolycichthys concolor. x |. Hab. South Shetlands. The above specimen agrees well with the description of the type, which was 175 mm. in total length. I am indebted to Dr C. F. Angel for information concerning the type specimen preserved in the Paris Museum (No. 11 — 99). Austrolycichthys bothriocephalus (Pappenheim). Lycodes bothriocephalus, Pappenheim, 1912, t.c, p. 178, pi. x, fig. 2. Austrolycichthys bothriocephalus, Regan, 1913, t.c, p. 245; Norman, 1937, Rep. B.A.N. Z. Antarct. Res. Exped., Ser. B, i, p. 80, fig. 10. Depth of body 11 in the length, length of head a little more than 6. Tail about 2 J times rest of fish. Pores on head very large and conspicuous. Diameter of eye about 5 in length of head. Maxillary extending to below anterior margin of eye ; jaws level anteriorly; teeth small, in 2 or 3 series anteriorly in both jaws, becoming uniserial laterally ; a small patch of teeth on the vomer and a pair of teeth (placed one behind the other) on each palatine. Gill-opening cleft downward to a little below middle of base of pectoral. Dorsal with about no rays, anal with about 90, caudal with about 10 ; origin of anal about one head-length behind the head; pectoral about | length of head. Brownish, irregularly and faintly variegated with paler; inside of mouth and branchial cavities blackish. ZOARCIDAE 8s Hab. Enderby Land ; Wilhelm Land. The type of the species, i8i mm. in total length, was from the Winter Station of the ' Gauss ', depth 380 metres. The above description is based upon another specimen, 114 mm. long, obtained by the B.A.N. Z. Antarctic Research Expedition off Enderby Land (66° 12' S, 49° 37' E), at a depth of 300 metres, which appears to belong to Pappenheim's species. Genus MELANOSTIGMA Melanostigma, Giinther, 1881, Proc. Zool. Soc, p. 21; Regan, 1913, t.c, p. 248; Norman, 1937, Discovery Rep., xvi, p. 109. Type Melanostigma gelatinosum, Giinther. Form elongate, compressed; body naked. Mouth terminal; teeth uniserial in both jaws and on vomer and palatines. Gill-opening small, above base of pectoral. Origin of dorsal fin just behind head ; no pelvic fins. Several species from deep water in the Atlantic and Pacific ; one from the Subantarctic and Antarctic Zones. Melanostigma gelatinosum, Giinther. Melanostigma gelatinosum, Giinther, 1881, t.c, p. 21, pi. ii, fig. A; Regan, 1913, t.c., p. 248; Norman, 1937, t.c, p. 109. St. 1082. 29. xii. 32. 53' 44' S, 38° 30-9' W. I m. tow-net, oblique, 290-120 m.: i specimen, 205 mm. Depth of body 7 to 10 in the length, length of head 6. Diameter of eye 3^ to 4 in length of head, interorbital width 8 to about 12. Mouth oblique; maxillary extending to below middle of eye. Origin of anal a head-length behind head; pectoral nearly \ length of head. Pale yellowish-brown, the sides in young spotted and marbled with purplish-grey ; end of tail blackish ; inside of mouth and gill-openings black. Fig. 55. Melanostigma gelatinosum. x |. Hab. Straits of Magellan ; north-west of South Georgia. The type of the species, 140 mm. in total length, was captured at a depth of about 45 metres. This represents one of the very few species found both in the Antarctic and Subantarctic Zones. CONGIOPODIDAE Zanclorhynchus spinifer, Giinther. Zanclorhynchus spinifer, Giinther, 1880, Shore Fish. 'Challenger', p. 15, pi. viii, fig. A; Waite, 1916, Austral. Antarct. Exped. Sci. Rep., Ser. C, in (i), Fishes, p. 72; Norman, 1937, Rep. B.A.N.Z. Antarct. Res. Exped., Ser. B, i, p. 59, fig. 4. Hab. Kerguelen ; Macquarie Island. The type of the species from Kerguelen is 90 mm. in total length. Waite recorded two specimens (202 and 238 mm.) from Macquarie Island, and the B.A.N.Z. Antarctic Research Expedition obtained 7 specimens (50 to 175 mm.). 86 DISCOVERY REPORTS LIPARIDAE Paraliparis gracilis, Norman. Paraliparis gracilis, Norman, 1930, Discovery Rep., 11, p. 352, fig. 42. St. 39. 25. iii. 26. East Cumberland Bay, South Georgia. Large otter trawl, 179-235 m.: 15 specimens, 40-115 mm. St. 42. I. iv. 26. Off mouth of Cumberland Bay, South Georgia. Large otter trawl, 120-204 m.: 2 specimens, 103, 115 mm. St. WS 32. 2i..xii. 26. Mouth of Drygalski Fjord, South Georgia. Small beam trawl, 91-225 m.: 2 specimens, 45, 58 mm. St. 142. 30. xii. 26. East Cumberland Bay, South Georgia. Net (4 mm. mesh) attached to back of trawl, 88-273 ™-- 5 specimens, 52-95 mm. St. 154. 18. i. 27. Jason Harbour to Larsen Point, South Georgia. Net (4 mm. mesh) attached to back of trawl, 6o-i6o m. : 9 specimens, 42-50 mm. Depth of body 5 to 5J in the length, length of head 4I to nearly 6. Diameter of eye about 3 in length of head. Dorsal 58-66. Anal 56-63. Pectoral with an upper portion of 10 to 12 rays. Fig. 56. Paraliparis gracilis. Holotype. x ij. Hob. South Georgia and neighbourhood. The holotype of the species, 70 mm. in total length, is from 53° 48' 00" S, 35° 37' 3° " ^' 72^ metres. The other Antarctic species are P. antarcticus, Regan [? = P. zvildi, Waite] ; and P. terrae-novae, Regan. I have given a key to these species in my report on the oceanic fishes (1930, i.e., p. 352). Careproctus georgianus, Lonnberg. Careproctus georgianus, Lonnberg, 1905, Wiss. Ergebn. Schwed. Siidpol.-Exped., v (6), p. 41, pi. iii, fig. II ; Burke, 1930, Bull. U.S. Nat. Mas., CL, p. 122. St. MS 68. 2. iii. 26. East Cumberland Bay, South Georgia. Large rectangular net, 220-247 m.: I specimen, 55 mm. St. 142. 30. xii. 26. East Cumberland Bay, South Georgia. Nets (4 mm. and 7 mm. mesh) attached to back of trawl, 88-273 m. : 7 specimens, 27-90 mm. St. 144. 5. i. 27. Off mouth of Stromness Harbour, South Georgia. Large otter trawl, 155- 178 m.: 3 specimens, 12-14 mm. Depth of body 4! to about 5 in the length, length of head 4 to 4I. Diameter of eye \ interorbital width, about 4 in length of head. Teeth simple. Gill-opening above base of pectoral. Dorsal 45-52. Anal 42-46. Pectoral with about 30 rays. Hab. South Georgia. LIPARIDAE 87 This species, which was not previously represented in the British Museum col- lection, was originally described from 2 specimens, 63 and 105 mm. in total length, from South Fjord, Cumberland Bay, depth 195 metres. Liparis steineni, described by Fischer (1885, Jahrb. Hamb. Wiss. Anst., 11, p. 63) may belong to this genus. The types, 2 specimens ^'S- 57- Careproctus georgiams. xf. 65 and 70 mm. in total length, were said to be from tide pools or shallow waters, Royal Bay, South Georgia. BOTHIDAE Mancopsetta maculata (Giinther). Lepidopsetta maculata, Giinther, 1880, Shore Fish. 'Challenger', p. 18, pi. xxx, fig. C; Norman, 1930, Discovery, Rep., n, p. 361. Mancopsetta maculata, Norman, 1934, Syst. Monogr. Flatfish., i, p. 248, fig. 189. St. 158. 21. i. 27. 53° 48' 30" S, 35° 57' 00" W. Large dredge, 401-411 m.: i specimen, 270 mm. Hab. South Atlantic, north of the Falkland Islands ; north-east of South Georgia ; southern Indian Ocean, near Prince Edward's Island. The type of the species, 134 mm. in total length, was collected by the ' Challenger' at Station 145, at a depth of 580 metres. The 'William Scoresby' obtained another example, 238 mm. long, from 45° 45' 00" S, 59° 35' 00" W, at a depth of 311 to 247 metres. This represents another species found in both the Antarctic and Subantarctic Zones. GENERAL PART THE ANTARCTIC ZONE In his report on the fishes of the 'Terra Nova' Expedition, Regan (19146) has dis- cussed in detail the distribution of Antarctic and Subantarctic fishes, and has defined three zones of distribution south of the Tropical Zone, namely, the South Temperate Zone, the Subantarctic Zone, and the Antarctic Zone. This last region was said to include "the coasts of the Antarctic continent and the islands that lie to the south of the isotherm of 6° C, with the possible exception of Macquarie Island". This zone was further subdivided into two districts. Glacial and Kerguelen. The Glacial District includes the coasts of the Antarctic continent and neighbouring islands, together with South Georgia, the South Sandwich Islands, and Bouvet Island, all situated within the extreme limit of the pack-ice. The Kerguelen District includes the island of Kerguelen and also Heard Island, the Crozet Islands, and Marion and Prince Edward Islands : as will be seen, there appears to be good reason for including Macquarie Island within this district. The mean annual surface isotherms of 6° C. and 12° C. correspond on the whole fairly closely to the Antarctic and Subtropical Convergences as delimited by Deacon^ in his valuable paper on the hydrology of the Southern Ocean. The following classifi- 1 1937) Discovery Reports, xv, pp. 1-124, pis. i-xliv. gg DISCOVERY REPORTS cation for the distribution of Antarctic Fishes, therefore, follows almost exactly that proposed by Regan : Antarctic Zone. The area bounded to the north by the Antarctic Convergence. Glacial District. Antarctic Continent; South Shetlands; South Orkneys; South Sandwich Islands; South Georgia; Bouvet Island. Kerguelen-Macquarie District. Kerguelen; Heard Island; Crozets; Marion and Prince Edward Islands; Macquarie Island. Fig. 58. Map showing the principal localities mentioned in the text. ANTARCTIC ZONE 89 HISTORICAL! The voyages of the 'Erebus' and 'Terror', under the command of Sir J. C. Ross, during the years 1839 to 1843, resuhed in the first specimens of fishes from the Antarctic Zone being brought to this country, examples of four species being collected at Ker- guelen. The fishes were reported upon by Richardson (i 844-1 848) who gave the following names to the species from Kerguelen : Notothenia cyanobrancha, Richardson. A'^. purpuriceps, Richardson (N. cyanobrancha). N. coriiceps, Richardson. Chaenichthys rhtnoceratus, Richardson. The following species was described from an unknown locality, but was quite likely captured at Kerguelen: Notothenia} rossii, Richardson. In addition, Richardson mentions three other fishes in his report, which appear to have been obtained in the true Glacial District:^ Sphyraena. (?? Gymnodraco acuticeps.) Notothenia phocae, Richardson. (An unidentifiable Nototheniid.) Pagetodes, Richardson (? Cryodraco antarcticus). During the years 1874 to 1875 two expeditions visited Kerguelen to observe the "Transit of Venus", both of which collected a few specimens of fishes during their stay on the island. The fishes collected by the American expedition were dealt with by Gill (1876), who recorded the following species from the Antarctic Zone for the first time : Harpagifer bispinis (Schneider). Giinther (1879) reported upon the specimens obtained by the British expedition, and described a new species of Ray from this district : Raja eatonii, Giinther. ! A valuable historical summary of early Antarctic voyages, commencing with that of Captain Cook in the years 1772-1775, has been given by Dollo (1904, pp. 5-11). ^ Concerning the first two of these, Richardson (p. 8) writes: "On the 14th of January 1842, when the ships were embayed among ice on the 65th parallel of south latitude and about the 155th west meridian, a seal was taken with twenty-eight pounds of fish in its stomach. The fish were of two kinds, one a Sphy- raena, the other a Notothenia, of which there were many mutilated individuals. Dr Hooker made a careful drawing of the most perfect, and put several examples in spirits. ..." I have been quite unable to trace either the drawing or the spirit specimens, so that Notothenia phocae must be regarded as unidentifiable, since the description is quite inadequate. No description of the Sphyraena is given, but Dollo (1904, p. 9) has suggested that this may have been a specimen of Gymnodraco acuticeps, Boulenger. Concerning Pagetodes, he writes: "When the ships were in the high latitude of 77° lo' S, and long. 178^° [W], a fish was thrown up by the spray in a gale of wind, against the bows of the Terror and frozen there." Unfortunately, this specimen, which was 160 mm. in total length, was appropriated by the ship's cat before a detailed drawing or description could be made, and the rough sketch reproduced by Richardson (pi. viii, fig. 3) is not sufiiciently accurate, to admit of positive identification. It seems probable, however, that this fish was of the same species as that described later by Dollo as Cryodraco antarcticus. 90 DISCOVERY REPORTS The S.M.S. ' Gazelle' (1874-1876), under the command of Captain F. von Schleinitz, visited Kerguelen during her voyage. The zoological results were edited by Studer (1889), but the single new fish obtained from this zone had been previously described by Peters (1876): Notothenia antarctica, Peters (A'^. macrocephald). The ' Challenger', under the command of Sir G. S. Nares, made valuable collections at Kerguelen, Marion Island and Prince Edward's Island, which were reported upon by Giinther (1880, 1887), and the following additions made to the fish-fauna of the Kerguelen District: Raja murrayi, Giinther. Muraenolepis marmoratus, Giinther. Notothenia acuta, Giinther. A'^. marionensis, Giinther. A'^. mizops, Giinther. A'^. squamifrons, Giinther. Bathydraco antarcticiis, Giinther. Zanclorhynchus spinifer, Giinther. Lepidopsetta maculata, Giinther (Maticopsetta maculata). The 'Deutschen Polarkommission ' (1882-1883) obtained a few fishes from South Georgia, which were described by Fischer (1885), who added the following species to the list of Antarctic fishes : ^ Notothenia marmorata, Fischer (A^. rossii). N. angustifrons, Fischer. Chaenichthys georgianus, Fischer {Parachaenichthys georgianus). Liparis steineni, Fischer (? Careproctus steineni). The voyage of the S.Y. 'Belgica' (1897-1899), under the command of A. de Gerlache de Gomery, with Dr E. G. Racovitza as naturalist, led to the addition of the following species to the list. The collection, which included the first fishes actually obtained from within the Antarctic Circle,^ was reported upon by Dollo (1900, 1904): Raja arctowskii, Dollo. Gerlachea australis, Dollo. Racovitzia glacialis, Dollo. Cryodraco antarcticiis, Dollo. The ' Southern Cross' (1898-1900), under the command of Mr C. E. Borchgrevink, obtained a valuable collection of fishes, working mainly in the Victoria Quadrant. These were described by Boulenger (1902), and the following added to the list: Notothenia nicolai, Boulenger (Trematomus nicolai). Trematomiis newnesi, Boulenger. T. borchgrevinki, Boulenger. T. hansoni, Boulenger. T. bernacchii, Boulenger. Pleuragramma antarcticum, Boulenger. Gymnodraco acuticeps, Boulenger. 1 See footnote on p. 81 of this report. 2 This expedition worked mainly in the Weddell Quadrant, but one or two specimens were obtained from the Ross Quadrant. ANTARCTIC ZONE 91 The Swedish South Polar Expedition, with the 'Antarctic', under the direction of Dr O. Nordenskjold, visited South Georgia, and also made collections of fishes at Graham Land and the neighbouring islands (Weddell Quadrant). The fishes were reported upon by Lonnberg (1905), who added the following to the Antarctic list: Miiraenolepis marmoratus microps, Lonnberg {M. microps). Notothenia dubia, Lonnberg (? Trematomus vicarius). N. misops nudifrons, Lonnberg (A'^. nudifrons). N. larseni, Lonnberg. A'^. gibberifrotis, Lonnberg. Trematomus hansoni georgianus, Lonnberg {T. hansoni). T. bernacchii vicarius, Lonnberg {T. vicarius). Artedidraco mirus, Lonnberg. A. skottsbergi, Lonnberg. Champsocephalus gunnari, Lonnberg. Chaenichthys rhinoceratus hamatus, Lonnberg {Chionodraco hamatus). Careproctus georgianus, Lonnberg. The German South Polar Expedition (1901-1903), with the 'Gauss', under the leadership of Erich von Drygalski, visited Kerguelen and also collected near the boundary between the Victoria and Enderby Quadrants of the Antarctic Continent. The fishes were dealt with by Pappenheim (1911, 191 2), who added the following new species : Notothenia lepidorhinus, Pappenheim (Trematomus lepidorhinus). Trematomus br achy soma, Pappenheim. Lycodes bothriocephalus, Pappenheim [Austrolycichthys bothriocephalus). L. brachycephalus, Pappenheim [Austrolycichthys brachycephalus). Lycodichthys antarcticus, Pappenheim. The National Antarctic Expedition (1901-1904), with the ' Discovery' under Captain R. F. Scott, obtained a few specimens in the Victoria Quadrant, which were described by Boulenger (1907). The following species were added to the Antarctic list: Notothenia scotti, Boulenger (Trematomus scotti). N. hodgsoni, Boulenger (Trematomus newnesi). Bathydraco macrolepis, Boulenger. Champsocephalus macropterus, Boulenger (Pagetopsis macropterus). The Scottish National Antarctic Expedition (i 902-1 904), with the 'Scotia', under the leadership of Dr W. S. Bruce, made important collections in the neighbourhood of the South Orkneys and Coats Land in the Weddell Quadrant. Dollo (1906, 1907, 1908, 1909) presented several preliminary descriptions of fishes obtained by this expedition, but the complete report was prepared by Regan (1913). A complete monograph of the Nototheniiform fishes was included in this report, and the following species were added to the list : Notothenia vaillanti, Regan (A^. gibberifrons). Trematomus loennbergii, Regan. Bathydraco scotiae, Dollo. Chaenichthys rugosus, Regan. Cryodraco pappenheimi, Regan. 92 DISCOVERY REPORTS The French Antarctic Expedition (1903-1905), with the 'Fran9ais', under the command of Dr Jean Charcot, collected off Graham Land and the neighbouring islands (Weddell Quadrant). The fishes were described by Vaillant (1905, 1906), who added the following new names to the faunal list : Dissostichiis eleginoides Smitt {D. mawsoni). Chaenichthys charcoti, Vaillant {Parachaenkhthys charcoti). Lonnberg (1906) described some fishes from South Georgia, collected by Mr Erich Sorling, and added one new species to the list of those known from the island : Chaenichthys aceratus, Lonnberg {Chaenocephalus aceratus). The British Antarctic Expedition (1907-1909), with the 'Nimrod', under the command of Sir Ernest Shackleton, collected a few fishes in the Victoria Quadrant, which were reported upon by Waite (191 1), who added one new species: Artedidraco shackletoni, Waite. The Second French Antarctic Expedition (1908-1910), with the 'Pourquoi-Pas? ', under the command of Dr Jean Charcot, made valuable collections in the Weddell Quadrant of the Antarctic. These were studied by Roule and Despax (191 1) and by Roule, Angel and Despax (1913), who described three new forms: Artedidraco loennbergi, Roule. DoUoidraco loiigedorsalis, Roule. Lycodes concolor, Roule and Despax [Austrolycichthys concolor). The British Antarctic ('Terra Nova') Expedition (1910), under Captain R. F. Scott, made very rich collections of fishes in the Victoria Quadrant. These were reported upon by Regan (1914, 1916), who described the following new species: Trematomus pennellii, Regan. T. centronotiis, Regan. T. eulepidotus, Regan. Artedidraco orianae, Regan. Histiodraco velifer, Regan. Pogonophryne scotti, Regan. Prionodraco evansii, Regan. Cryodraco atkinsoni, Regan (C. antarcticus). Chionodraco kathleenae, Regan. Chaenodraco wilsoni, Regan. C.fasciatus, Regan (? C. wilsoni). Paraliparis antarcticus, Regan. P. terrae-novae, Regan. The Australasian Antarctic Expedition (1911-1914), with the 'Aurora', under the leadership of Sir Douglas Mawson, obtained a number of fishes from Macquarie Island and from the Victoria Quadrant of the Antarctic continent. These were described by Waite (1916), who added the following names to the Antarctic list: Notothenia coriiceps var. macquariensis, Waite (A'^. rossii). Bathydraco ntidiceps, Waite. Aconichthys harrissoni, Waite [Racovitzia harrissoni). ANTARCTIC ZONE 93 Cygnodraco mawsoni, Waite. Dacodraco hunteri, Waite. Paraliparis wildi, Waite (? P. antarcticus). Finally, the British, Australian, New Zealand Antarctic Research Expedition (1929- 193 1), with the 'Discovery', under the leadership of Sir Douglas Mawson, made extensive collections in the Victoria and Enderby Quadrants of the Antarctic, and also obtained fishes from Kerguelen, Heard Island, the Crozets, and from Macquarie Island. The fishes were reported upon by Norman (1937), who described two new forms : Muraenolepis tnicrocephalus, Norman. Dissostichus mawsoni, Norman. DISTRIBUTION OF ANTARCTIC COAST FISHES It has been pointed out in the introduction to this report that, of the 86 species of fish recorded from the Antarctic Zone, 65 (75 per cent) belong to the Percoid Division Nototheniiformes. Taking the Glacial District alone, the percentage rises to 76, and for the coasts of the Antarctic Continent it is nearly 85. It follows, therefore, as Regan (19146, p. 26) has pointed out, that this group of fishes is of paramount importance in delimiting an Antarctic region or zone. The table on page 95 shows the distribution of those species of Nototheniiform fishes occurring within the area bounded by the Antarctic Convergence (fig. 59). The Glacial District is dealt with first, the four quadrants of the Antarctic Continent preceding the more outlying islands ; the localities of the Kerguelen-Macquarie District follow, and finally the Subantarctic districts — the Patagonian region and the Antipodes. A glance at the two right-hand columns shows at once the importance of the Con- vergence as a boundary for Antarctic coast fishes. Of the 65 species listed, only three {Notothenia macrocephala, N. colbecki, Harpagifer bispinis) occur also in the Sub- antarctic Zone, and only one of these {H. bispinis) is common to the Glacial District and the Subantarctic Zone. Of the 26 genera listed, 21 are peculiar to the Glacial District, i is peculiar to the Kerguelen-Macquarie District, and 4 are found also in the Subantarctic Zone. Of these four genera, Harpagifer is monotypic, Dissostichus and Champsocephaliis each have one species in the Antarctic and one in the Sub- antarctic Zone, and the large genus Notothetiia is well represented in both zones. Regan {t.c, p. 29) has shown that the species of Notothenia form five natural groups, and the species occurring in the Antarctic Zone may be distributed in these groups as follows : "tessellata" group "squamifrons" group "acuta" group " marionensis" group "coriiceps" group squamifrons kempi larseni mizops nudifrons gibberifrotis acuta angustifrons marionensis coriiceps rossii macrocephala colbecki 94 DISCOVERY REPORTS APPROXIMATE POSITION OF THE BOUNDARY BETWEEN THE EA5TAND WEST WIND DRIFTS IN THE ANTARCTIC ZONE ♦ * + * * NORTHERN BOUNDARY OF THE WEDDELL SEA CURRENT 120 150° West 180 East 150° Fig. 59. The Antarctic and subtropical convergences, the northern boundary of the Weddell Sea current, and the approximate position of the boundary between the East and West Wind Drifts. (After Deacon.) ANTARCTIC ZONE Table I. 95 Notothenia squamifrons Notothenia ke?npi Notothenia larseni Notothenia gibberifrons Notothenia mizops Notothenia tiudijrons Notothenia acuta Notothenia angiistifrons Notothenia marionensis Notothenia cyanobrancha Notothenia coriiceps Notothenia rossii Notothenia macrocephala Notothenia colbecki Trematomus newnesi Trernatomtis nicolai Trematomus bordigrevinki Trematoynus brachysoma Trematomus bernacchii Trematotmis vicarius Trematomus hansoni Trematomus loennbergii Trematomus pemiellii Trematomus centronotus Trematoynus scotti Trematomus lepidorhinus Trematomus eulepidotus Dissostichus mawsoni Pleuragramma antarcticum Artedidraco mirus Artedidraco orianae Artedidraco skottsbergi Artedidraco loennbergi Artedidraco shackletoni Dolloidraco longedorsalis Histiodraco velifer Pogonophryne scotti Pogonophryne marmoratus Harpagifer bispinis Bathydraco antarcticus Bathydraco tnarri Bathydraco macrolepis Bathydraco scotiae Bathydraco nudiceps Gerlachea australis Racovitzia glacialis Racovitzia harrissoni Prionodraco evansii Cygnodraco mawsoni Paracliaenichthys georgianiis Parachaenichthys charcoti Psilodraco breviceps Gymnodraco acuticeps Champsoceplialus gunnari Pagetopsis macropterus Pseudochaenichthys georgianus Dacodraco hunteri Chaenichthys rhinoceratus Chaenichthys rugosus Chaenocephalus aceratus Cryodraco antarcticus Cryodraco pappenheimi Chionodraco kathleenae Chionodraco hamatus Chaenodraco wilsoni Antarctic Zone Glacial District o 5 0-0 .S3 X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X lU -O X X X X X X X X X X X X X X X X ' 3 -■o X X X X X X X X X X X X X X X X X X CO X X X X o X X X J3 u '% -d s a Cfi Cfi X X £? o u c Kerguelen-Macquarie District SUB- ANTARC- Tic Zone U 3 1-1 3 IT C O ■d o X X X X X 96 DISCOVERY REPORTS It will be seen that the " tessellata " group, so characteristic of the Patagonian Region, is altogether wanting in the Antarctic Zone. Another interesting fact clearly shown by this table as well as by the map (Plate I) is that, although the genus Notothenia is well represented by six species on the coasts of Graham Land and at the neighbouring South Shetlands, it is represented in the Victoria Quadrant of the Antarctic Continent only by A^. coriiceps, and has not been found at all in the Enderby and Ross Quadrants. It may be noted that A^. coriiceps is probably not so demersal or littoral in its habits as many other species of the genus, and the silvery young may be mainly pelagic. From Graham Land species of Notothenia spread eastwards to the South Orkneys, South Sandwich Islands, South Georgia and Bouvet Island in the Glacial District, and other species occur in the Kerguelen-Macquarie District. In contrast to Notothenia, the genus Trematomus occurs in all four quadrants of the Antarctic Continent, extending northwards in the Weddell Quadrant to the South Orkneys and South Georgia, but not reaching Bouvet Island ; it is absent altogether in the Kerguelen-Macquarie District. With regard to the Zoarcidae, the group next in importance to the Nototheniiformes, I have little to add to the remarks on their distribution made by Regan {t.c, pp. 31-33). Two of the Antarctic species are the southern representatives of the northern deep- water genera Lycenchelys and Melanostigma : Melanostigma gelatinosum is found in both Antarctic and Subantarctic Zones. The remainder are all generically distinct from the northern members of the family. The two genera Lycodichthys and Austrolycichthys (the first monotypic, the second with three species) are peculiar to the Antarctic and are confined to the Glacial District. THE ANTARCTIC ZONEi This includes the coasts of the Antarctic Continent and all the islands lying on or to the south of the Antarctic Convergence (fig. 59). To quote Regan {t.c, p. 33), "it is characterised by the complete absence of South Temperate types, by the absence of Bovichthyidae [the most generalised family of Nototheniiformes] and by the great development of other Nototheniiformes". The percentage of genera peculiar to this region is high (about 67 per cent), and the per- centage of peculiar Nototheniiform genera very high (85 per cent). The percentage of species ranging beyond the limits of the Antarctic Zone is very low indeed (less than 10 per cent). Glacial District This includes the coasts of the Antarctic Continent and neighbouring islands, together with South Georgia, the South Sandwich Islands, and Bouvet Island, all found within the extreme Umit of pack-ice. NoTOTHENiiDAE. Notothenia (7 species); Trematomus (13 species); Dissostichiis (i species); Pleuragramma (i species). 1 The data which follow are arranged exactly as in Regan's report (19146, p. 33) to facilitate ready re- ference and comparison. ANTARCTIC ZONE 97 Trematomiis and Pleiirogramma are restricted to this district, and of the 7 species of Notothenia, 5 are peculiar, only A'', coriiceps and A^. rossii ranging to the Kerguelen- Macquarie District. A^. angustifrons is confined to South Georgia and the South Sand- wich Islands. Dissostichus is represented by D. mawsoni in this district, and there is one other species {D. eleginoides) in the Patagonian Region of the Subantarctic Zone. Fig. 60. Known distribution of Harpagiferidae. A i, Artedidraco minis; A 2, A. orianae; A 3, ^. skottsbergi; A 4, A. loennbergi; A 5, A. shackletoni; D, Dolloidraco longedorsalis; V, Histiodraco velifer; P I, Pogonophryne scotti; P 2, P. marmoratus; H, Harpagifer bispinis. 13 98 DISCOVERY REPORTS Fig. 6i. Known distribution of Bathydraconidae. B i, Bathydraco antarcticus; B 2, B. marri; B 3, B. macroleph; B 4, B. scottae; B 5, B. nudkeps; G, Gerlachea australis; R i, Racovitzia glacialh- R 2, R. harrissoni; Pr, Prionodraco evansii; C, Cygnodraco mawsoni; Pi, Parachaenichthys georgianus; P 2, P. Charcot! \ Ps, Psilodraco breviceps; Gy, Gymnodraco acuticeps. ANTARCTIC ZONE 99 Harpagiferidae. Artedidraco (5 species); Dolloidraco (i species); Histiodraco (i species); Pogonophryne (2 species); Harpagifer (i species). With the sole exception of Harpagifer bispinis, which occurs also in the Kerguelen- Macquarie District, as well as in the Patagonian Region of the Subantarctic Zone, all the genera and species of this family are peculiar to the Glacial District. Bathydraconidae. Bathydraco (5 species); Gerlachea (i species); Racovitzia (2 species); Prionodraco (i species); Cygnodraco (i species); Parachaenkhthys (2 species); Psilodraco (i species); Gymnodraco (i species). All the members of this family are confined to the Glacial District. According to Regan {t.c, p. 33), Bathydraco must be regarded as oceanic, but, in spite of their being taken mainly in deep water, it seems likely that the related genera Gerlachea, Racovitzia, and perhaps Prionodraco, should be reckoned as coast fishes. None of the specimens of Bathvdraco, however, has been captured far from the coast of the Antarctic Continent. Cygnodraco, Parachaenichthys, Psilodraco and Gymnodraco are clearly to be looked upon as coast fishes. Chaenichthyidae. Champsocephalus (i species); Pagetopsis (i species); Pseudo- chaenichthys (i species); Dacodraco (i species); Chaenocephalus (i species); Cryodraco (2 species); Chionodraco (2 species); Chaenodraco (i species). The members of this family are highly characteristic of the Glacial District, only three species being found outside the area. Of these, the two species of Chaenichthys are confined to Kerguelen, and the second species of Champsocephalus, C. esox, occurs in the Patagonian Region of the Subantarctic Zone. Zoarcidae. Lycenchelys (i species); Lycodichthys (i species); Austrolycichthys (3 species); Melanostigma (i species). The species of Maynea from South Georgia mentioned by Regan {t.c, p. 34) must be removed from the list of Southern Zoarcidae (see footnote on p. 81). The species of Melanostigma is found also in the Subantarctic Zone. The four species of Zoarcidae from the Antarctic Continent belong to two peculiar genera, Lycodichthys and Austro- lycichthys, related to the Patagonian genera Iluocoetes and Austrolycus respectively. Muraenolepidae. This family includes a single genus, of which four species have now been described, one from the Patagonian Region, one from Kerguelen, and two from the Glacial District. One of these last (M. microps) has also been recorded from the neighbourhood of the Burdwood Bank, just north of the Antarctic Convergence, and, thus, outside the Antarctic Zone. The additional collections obtained since the publication of Regan's report (1914^) have increased the number of species known to have a circumpolar distribution, but our knowledge of the range of many species in the Glacial District is still far too incomplete to make possible any subdivision of this area. Of the 6 species of Notothenia occurring on the coasts of Antarctica, only one (A^. coriiceps) has a circumpolar dis- tribution, whereas, of the 12 species of Trematomus, 7 have been recorded from both the Weddell and Victoria Quadrants. Pleuragramma antarcticum and Dissostichus mawsoni are likewise known to be circumpolar. Of the Harpagiferidae, 3 out of 13-2 DISCOVERY REPORTS 9 species are known from both sides of the Antarctic, of the Bathydraconidae, 3 out of 12 species, and of the Chaenichthyidae, 2 out of 8 species. As Regan has pointed out {t.c, p. 36), the affinities of the Nototheniiform fishes found at South Georgia are predominantly Glacial, and the presence of Notothenia larseni near Bouvet Island suggests that this island is also rightly included within this area. Kergiielen-Macquarie District This district includes Marion and Prince Edward Islands, the Crozets, Kerguelen, Heard Island, and Macquarie Island. I have dealt with the fish-fauna in some detail in my report on the fishes obtained by the B.A.N. Z. Antarctic Research Expedition, but it may be of interest to summarize the conclusions here. The following table shows the distribution of the species of coastal fishes so far recorded from this district : Table II. Raja eatonii Raja miirrayi Muraenolepis marmoratus Notothenia squamifrons Notothenia mizops Notothenia acuta Notothenia marionensis Notothenia cyanobrancha Notothenia coriiceps Notothenia rossii Notothenia macrocepha/a Notothenia colbecki Harpagifer bispinis Champsocephalus gunnari CItaenichthys rhinoceratus Chaenichthys nigosus Zanclorhynchus spinifer Mancopsetta maculata 3 -0 0 a, '5 0 2 'Si 0 0 ai u u u c < X X X X X X X X X X X X X X X X X X X X X X > > < < X X X X X X X X X X X Related to R. scaphiops (Patagonia) Related to R. macloviana (Patagonia) Related to M. microps (mainly Antarctic) Related to N. macrophthalma (Birdwood Bank) and N. larsoni (Antarctic) Related to A^. nudifrons (Antarctic) Related to A'', angustifrons (Antarctic) Related to A'^. aiigustifrons (Antarctic) Related to N. coriiceps (Antarctic) C. esox occurs in Patagonia It is clear that, although the coastal fish-fauna of this district shows certain features of resemblance to that of the Patagonian Region and the Antipodes, its affinities are mainly with that of Antarctica. Of the Nototheniiform fishes, the characteristic Antarctic genus, Trematomus, is absent from the Kerguelen-Macquarie District, but in the genus Notothenia the " tessellata'' group so characteristic of the Patagonian Region is also wanting. Five species of Notothenia are peculiar to the district, all of which appear to be most nearly related to species occurring in the Glacial District, and mostly to species characteristic of the Weddell Quadrant. Of the remaining species, A^. coriiceps and N. rossii are found elsewhere only in the Glacial District. N. macrocephala, however, occurs in both the Patagonian Region and the Antipodes, but not in the Antarctic, and N. colbecki in the Antipodes. Harpagifer bispinis occurs in the Patagonian Region as well as in the Glacial ANTARCTIC ZONE Fig. 62. Known distribution of Chaenichthyidae. Cs. i, Champsocephalus esox\ Cs. 2, C. gunnari; P, Pagetopsis macropterus; Ps, Psetidochaetiiclitliys georgiamis; D, Dacodraco hunteri; C. i, Chaenichthys rhinoceratus; C. 2, C. rugosus; Cn, Chaenocephalus aceratus; Cr. i, Cryodraco antarcticus; Cr. 2, C.pappen- heimi; Ci. i, Chionodraco kathleenae; Ci. 2, C hamatus; Ch, Chaenodraco wilsoni. I02 DISCOVERY REPORTS District. Champsocepholus giimiari is found also at South Georgia, and a related species, C. esox, in the Patagonian Region. Chaenichthys (with two species) is a genus peculiar to the island of Kerguelen, but belongs to a family, most of the members of which occur in the Glacial District. Thus, the only way to show the dissimilarity of the fish-fauna of the Kerguelen- Macquarie District from those of the Patagonian Region and the Antipodes, and to express its affinity to that of Antarctica, is to include it in the Antarctic Zone as a separate district (Regan, 1914^, p. 36). I do not propose to discuss the question of the origin of the Antarctic fish-fauna in this report, and have nothing to add to the general conclusions of Regan {t.c, p. 40). As he points out, "the distinctive features of the fish-fauna. . .are that nearly all the genera and species are peculiar and that they nearly all belong to a single group, the Nototheniiformes, which is characteristic of and almost restricted to the Antarctic and Subantarctic Zones. In the Antarctic Zone this group has developed into a large number of types that diiTer greatly in structure, appearance and habits. These facts seem to point to the conclusions that Antarctica may have long been isolated and that its coasts may have been washed by a cold sea for a long time, probably throughout the Tertiary Period." Regan has also shown that neither the fresh-water fishes, nor the marine fishes, whether Antarctic or South Temperate, support the theory that the Antarctic Continent connected South America with Australia during the Tertiary Period. BIBLIOGRAPHY List of the principal memoirs and papers dealing with the marine fishes of the Antarctic Zone from 1844 to 1937, arranged in chronological order 1844-48. Richardson, J. The zoology of the voyage of H.M.S. 'Erebus' and ' Terror', under the command of Captain Sir James Clark Ross, R.N., F.R.S., during the years 1839 to 1843. Edited by J. Richard- son and J. E. Gray. II. Ichthyology. London, 4°, 139 pp., 60 pis. 1876. Gill, T. N. Fishes. In Kidder, J. H. Contributions to the natural history of Kerguelen Island, made in. . . 1874-75, II- In Bull- U-S- N^^- ^u^-' '"• PP- 41-42- ■ 1876. Peters, W. C. H. Vbersicht .der ivdhrend der von 1874 bis 1876 unter dem Commando des Hrn. Capitdn z. S. Freiherrn von Schleinitz ausgefuhrten Reise S.M.S. 'Gazelle' Monatsber. Akad. Berlin, 1876, pp. 831-854. 1879. GtJNTHER, A. An account of the petrological, botanical, and zoological collections made in Kerguelen's Land and Rodriguez during the Transit of Venus Expeditions. . .in the years 1874-75. Fishes. Phil. Trans. R. Soc. London, clxviii (extra vol.), p. 166. 1879. Studer, T. Die Fauna von Kergueknsland. Arch. Naturg., xlv (i), pp. 114-141. [Fishes, pp. 130- 131-] 1880. Gunther, a. Report on the shore fishes procured during the voyage of H.M.S. ' Challenger' in the years 1873-1876. Challenger Reports, i (6), 82 pp., 32 pis. 1885. Fischer, H. G. Ichthyologische und herpetologische Bemerkungen. Jahrb. Hamburg. Wiss. Anst., 11, pp. 49-121, 4 pis. {Ober Fische von Siid-Georgien, pp. 49-62, pis. i, ii, fig. 9.] 1887. Gunther, A. Report on the deep-sea fishes collected by H.M.S. 'Challenger' during the years 1873- 1876. Challenger Reports, xxii (57), 268 pp., 68 pis. BIBLIOGRAPHY 103 1889. GuNTHER, A. Report on the pelagic fishes collected by H.M.S. 'Challenger' during the years 1873- 1876. Challenger Reports, xxxi (78), 47 pp., 6 pis. 1889. Studer.T. Die Forschungsreise S. M.S. 'Gazelle' in den jfahren i8j^ bis i8y6 3. Theil. Zoologie u. Geologic. Berlin, 4°. 1900a. DoLLO, L. Cryodraco antarcticus , poisson abyssal nouveau, recueillipar VExpidition Antarctique Beige {Communication preliminaire). Bull. Acad. roy. Belg. (Classe Sci.), 1900, No. 2, pp. 128-137. 19006. DoLLO, L. Gerlachea australis, poisson abyssal nouveau, recueilli par I'Expedition Antarctique Beige {Communication prehminaire). Ibid., No. 3, pp. 194-206. 1 900c. DoLLO, L. Racovitsia glacialis, poisson abyssal nouveau , recueilli par I'Expedition Antarctique Beige {Communication preliminaire). Ibid., No. 4, pp. 316-327. 1900^. DoLLO, L. Macrurus Lecointei, poisson abyssal nouveau, recueilli par I'Expedition Antarctique Beige {Communication preliminaire). Ibid., No. 6, pp. 383-401. 1902. BouLENGER, G. A. Pisces. In Report on the collections of natural history made in the Antarctic regions during the voyage of the 'Southern Cross'. London, 8°, 344 pp., 52 pis. [Fishes, pp. 174-189, pis. xi-xviii.] 1904. DoLLO,L. Expedition Antarctique Beige. Rhultats du voyage duS.Y .' Belgica' en 1 897-1 898-1 899 Rapports scientifiques . . .Zoologie. Poissons. Antwerp, 4°, 240 pp., 12 pis. text-figs. 1905. LoNNBERG, A. J. E. The fishes of the Swedish South Polar Expedition. Wissensch. Ergebn. Schwed. Siidpol.-Exped., v (6), 69 pp., 5 pis. 1906a. DoLLO, L. Bathydraco Scotiae, poisson abyssalnouveau,recueillipar I'Expedition Antarctique Nationale £cossaise. Note preliminaire. Proc. R. Soc. Edinburgh, xxvi, pp. 65-75. 19066. DOLLO, L. Neobythites Brucei, poisson abyssal nouveau, recueilli par V Expedition Antarctique Nationale Ecossaise. Note preliminaire. Ibid., pp. ijz-iSi. 1906. LoNNBERG, A. J. E. Contributions to the fauna of South Georgia. I. Taxonomic and biological notes on vertebrates. K. Svensk. Vet.-Akad. Handl., XL (5), 104 pp., 12 pis. [Fishes, pp. 99-100.] 1906a. Vaillant, L. Sur les poissons recueillis pendant I'Expedition Antarctique Franfaise commandee par le Dr Jean Charcot. Note preliminaire. Bull. Mus. hist. nat. Paris, xii (3), pp. 138-140. 19066. Vaillant, L. Sur une nouvelle espece de Chaenichthys provenant de I'Expedition Antarctique Franfaise sous le commandement du Dr Jean Charcot. T.c, pp. 246-247. 1906c. Vaillant, L. Expedition Antarctique Franfaise (1903-1905) commandee par le Dr Jean Charcot. Poissons. Paris, 4°, 52 pp., 4 text-figs. 1907. BouLENGER, G. A. National Antarctic Expedition. Natural History, u {iv). Fishes, 5 pp., 2 p\s. 1907. DoLLO, L. Prymnothonus Hookeri, poisson pelagique de la 'Erebus' et de la ' Terror' retrouve par I'Ex- pedition Antarctique Nationale Ecossaise. Note preliminaire. Proc. R. Soc. Edinburgh, xxvii, pp. 35- 45- 1908. DoLLO, L. Notolepis Coatsi, poisson pelagique nouveau, recueillipar I'Expedition Antarctique Nationale Ecossaise. Note preliminaire. /6z(f., xxviii, pp. 58-64. 1909a. DoLLO, L. Cynomacrurus Piriei, poisson abyssale nouveau recueilli par I'Expedition Antarctique Nationale £cossaise. Note preliminaire. Ibid., xxix, pp. 316-326. 19096. DoLLO, L. Nematonurus Lecointei, poisson abyssal de la 'Belgica ' retrouve par I'Expedition Antarctique Nationale Scossaise. Note preliminaire. Ibid., xxix, pp. 488-498. 191 1. Pappenheim, P. Neue antarktische Fische. Nach dem Material der deutschen SUdpolar-Expedition, 1901-1903. Sitzungsber. Ges. naturf. Freunde, Berlin, 1911, pp. 382-383. 191 1 . RouLE, L. Sur quelques particularites de la faune antarctique d'apres la collection de poissons recueillie par I'Expedition Franfaise du ' Pourquoi-Pas?' . C.R. Acad. Sci. Paris, CLIII, pp. 80-81. 191 1. RouLE, L. and Despax, R. Expedition Antarctique Franfaise du 'Pourquoi-Pas?', dirigee par le Dr J.-B. Charcot (1908-1910). Note preliminaire sur les poissons antarctiques . Bull. Mus. hist. nat. Paris, 191 1, No. 5, pp. 276-280. 191 1. Waite, E. R. British Antarctic Expedition 1907-1909 under the command of Sir E. H. Shackleton, C.V.O., II Biology, Antarctic Fishes, pp. 11-16, pi. ii. 1912. Pappenheim, P. Die Fische der deutschen SUdpolar-Expedition i<)oi-\()Ot,. I. Die Fische der Antarktis und Subantarktis. Deutsche Siidpol.-Exped., xill, Zool. V, pp. 161-182, pis. ix-x, 10 te.xt-figs. los DISCOVERY REPORTS 1913. Regan, C. T. The Antarctic fishes of the Scottish National Antarctic Expedition. Trans. R. Soc. Edinburgh, XLix, pp. 229-292, 11 pis. 1913. RouLE, L. Description d'un genre nouveau et de deux especes nouvelles de poissons antarctiques . Bull. Mus. hist. nat. Paris, 1913, No. i, pp. 4-7. 1913. RouLE, L., Angel, F. and Despax, R. Deuxieme Expedition Antarctique Frangaise (1908-1910) coni- mandee par le Dr Jean Charcot. Poissons. Paris, 4°, 24 pp., 4 pis. 1914a. Regan, C. T. Diagnoses of new marine fishes collected by the British Antarctic (' Terra Nova') Ex- pedition. Ann. Mag. Nat. Hist., (8) xiii, pp. 11-17. 19146. Regan, C. T. British Antarctic {'Terra Nova') Expedition, 1910. Fishes. Terra Nova Reports, Zool. I (i), pp. 1-54, 13 pis., text-figs. 1916a. Regan, C. T. Antarctic and Subantarctic fishes. Ann. Mag. Nat. Hist., (8) xviii, pp. 377-379. 19166. Regan, C. T. British Antarctic {'Terra Nova') Expedition, 1910. Larval and post-larval fishes. Terra Nova Reports, Zool. I (4), pp. 125-156, 10 pis. 1916. Waite, E. R. Australasian Antarctic Expedition, 1911-1914. Under the leadership of Sir Douglas Mawson, D.Sc, B.E. Fishes. Austral. Antarct. Exped. Sci. Rep., Ser. C, in (i), 92 pp., 4 pis., 2 maps, 16 text-figs. 1932. MoNOD, R. and Dollfus, R. P. Sur quelques animaux rapportes par E. Aubert de la Rue des lies Australes {Kerguelen et St Paul). Bull. Soc. zool. France, lvii, pp. 55-76. [Fishes, pp. 74-76.] 1937. Norman, J. R. B.A.N.Z. Antarctic Research Expedition, 1929-1931, under the command of Douglas Mawson, Kt., O.B.E., B.E., D.Sc, F.R.S. Fishes. Rep. B.A.N.Z. Antarct. Res. Exped., Ser. B, I (2), pp. 49-88, II text-figs. oijiQ. fan a) a'waA, A ."/i ,a ; A\ ,A 1 tMj\twUOZlKi PLATE I Known distribution of Notothenia (red), Trematonius (blue) and Dissostichus (black). A, Notothenia squamifrons ; B, N. kempt; C, N. larseni; D, A^. gihberifrons; E, A^. mizops; F, A^. nudifrons; G, N. acuta; H, N. angtistifrom; I, N. marionensis ; K, A^ cyanobrancha; L, A'^. coriiceps; M, A'^. rossii. The plain red circles show the kno\vn range of the Subantarctic species of Notothenia. A, Trematomus newnesi; B, T. nicolai; C, T. borchgrevinki; D, T. brachysoma; E, T. bernacthii; F, T. vicarius; G, T. hansoni; H, T. loennbergii; I, T.penneUii; K, T. centronotus; L, 7". scotti; M, T. lepidorhinus ; N, T. eulepidotus . A, Dissostichus eleginoides; B, £). mawsoni. DISCOVERY REPORTS XVIII PLATE I [Discovery Reports. Vol. XVIII, pp. 105-120, Plate II, August, 1938.] ON THE OPERATION OF LARGE PLANKTON NETS By JAMES W. S. MARR, M.A., B.Sc. ON THE OPERATION OF LARGE PLANKTON NETS By James W. S. Marr, m.a., b.Sc. (Plate II; Text-figs. 1-4) FOR the exploration of the deeper layers of the sea and the capture of the bigger and more active pelagic animals such as fishes, squids and the larger decapod Crustacea, which through their strength and speed can readily elude the ordinary tow- net, a net of exceptional size is necessary. The larger the apparatus the better, especially if investigations are to be made in deep water, for there are many indications that the ocean depths, while predominantly the haunt of relatively large and swiftly-moving creatures, are on the whole more thinly populated than the upper layers, so that the chances of their successful exploitation by ordinary nets are incUned to be small. However large the net may be it must be possible to close it, otherwise it can only be assumed, and, as will be shown later, assumed on inadequate grounds, that the species captured were taken from the particular level at which the net was fishing and not during its ascent to the surface. Although large plankton nets have for long been in general use, until within com- paratively recent years they were employed for the most part as simple open nets. The problem of closing them satisfactorily first appears to have received serious attention during the early days of the Discovery Committee's work when, between 1925 and 1927, closing experiments with several types of large mid-water net were conducted from the auxiliary barque ' Discovery ' and formed an important part of her scientific programme. These experiments were concerned chiefly with two types of net — one a pelagic trawl of the well-known Petersen young-fish type, the other a tow-net of graded mesh, 60 ft. long and mounted on a ring 450 cm. (nearly 15 ft.) in diameter. The latter, usually referred to as the 4I m. net, or N 450, is believed to be the largest plankton net that has yet been fished. A third net, also of graded mesh and very similar in design to the N 450 but with an opening of 200 cm. diameter, was used for some time in the ' Discovery ' ; but after a short trial it was discarded in favour of the all-stramin pelagic trawl, which, mounted on a 200 cm. frame and used as an ordinary tow-net, proved in many ways a more efficient apparatus, and, being of stouter construction, better suited to the boisterous conditions of the Southern Ocean. The dimensions and structural details of the N 450 are shown in Fig. i . Since the first description of this net was published one important structural alteration has been made. As recommended by Kemp,i the longitudinal support roping (Fig. i,^), originally 1 Kemp, S., and Hardy, A. C, 1929, Discovery Investigations: Objects, Equipment and Methods, Part II, Discovery Reports, I, p. 187. io8 DISCOVERY REPORTS of Stout log-line and later of soft 2 in. circumference tarred hemp known in Admiralty dockyards as "Rumbo", has been replaced by a flat braided sennet, or webbing, of exceptional strength. This material has given every satisfaction, never taking turns in the water and interfering with the efficiency of the net as the " Rumbo" line was inclined to do. In the ' Discovery ' experiments the N 450 was towed horizontally,^ at that time the common practice with most if not all types of towed net, particularly those of larger size. Although it had worked well when fished open, yielding substantial catches in first-rate condition, as a closing net the N 450 met with little success. For closing, the familiar Nansen method was used, but even when all difficulties in the actual operation had eventually been overcome the results were never satisfactory : the volume of the catch was in most instances disappointingly small, the material often in poor condition and occasionally damaged so badly as to defy identification. Fig. I. Plan of 4I m. net showing principal measurements and main structural features (after Kemp and Hardy). A, cylindrical non-catching part of i| in. mesh netting, 50 ft. in circumference. BCD, conical catching part tapering to a circumference of 3 ft. at e, with meshing as follows: B, | in., C, 7 mm., D, 7 mm., Hned at the cod-end with finer 4 mm. material, e, canvas cylinder to slip over bucket; /, canvas band; g, longitudinal webbing supports; /;, bridles; r, net ring, 450 mm. in diameter, 50 ft. in circumference. The curious poverty of material in the closing net gave rise to the somewhat dis- turbing view that a large proportion, perhaps the majority, of the animals captured in a big open net fished horizontally might be taken not from the restricted horizon at which the net was fishing, but during its ascent towards the surface when the speed of towing as a rule is augmented by hauling. In other words it began to be suspected that the N 450, big as it was, was not being towed fast enough to capture the strong and actively-moving species for which it was designed — it seemed exceedingly likely in fact that during the horizontal tow all but the weakest were contriving to make their escape from the net, if they were not indeed eluding it altogether. It was for this reason that in later experiments the old method of slow horizontal fishing was abandoned and the faster method of the oblique tow, which for some time had been employed to advantage with smaller nets, took its place. The poor condition of the material is not surprising, for so large an apparatus when closed on the Nansen principle would appear to behave as Kemp has put it " like an ill- 1 At a speed which is stated to have been about 2 knots. See, however, p. 115. LARGE PLANKTON NETS 109 constructed kite, the bucket lashing from side to side and causing serious damage to the contents". Such erratic movement is doubtless aggravated by the behaviour of the heavy iron ring, which, no longer towed squarely by the bridles or buoyed up and held in steady control by the open net, is free to surge about at random. But that is not all that may happen : if, for instance, the moment of closing should escape notice, so that there is some delay in hauling after the net has actually been closed, or if some hitch should occur while hauling necessitating a stoppage of the winch, there is danger of the collapsed net with its cumbrous unsupported frame sinking and fouling the warp. Such a possibility cannot be overlooked, for it has been seen time after time when working flights of small silk and stramin nets that, for all their lightness, they must be hauled at fair speed as soon as they are closed, otherwise in their collapsed and unbuoyed state the mere force of towing alone is not enough to keep them from sinking and getting round the warp. In view of its far greater bulk and weight a mishap of this nature is more liable to happen with the N 450 than with smaller nets, especially in a very deep haul when the exact moment of closing may readily pass undetected by those on deck. Owing to the failure of the Nansen principle the original idea of employing the N 450 as a routine deep-water net was temporarily abandoned, and for a long time during the earlier voyages of the 'Discovery IT it was put only to rare and occasional service, mainly as an open net. During the fourth voyage of the 'Discovery IT, however, in response to increasing demands for more accurate information regarding the depths frequented by the larger pelagic creatures, the various problems involved in working this large apparatus began again to receive attention. The chief problem, that of closing, had already been brought much nearer solution as a result of experiments with the young-fish trawl conducted towards the end of the third voyage of the 'Discovery 11'. The Nansen principle had been abandoned and a system of internal throttling introduced in which the heavy net ring was so effectively controlled that all possibility of surging after closing was eliminated. Under the new system, which was recently described by Mackintosh,^ the throttling rope operates on the inside of the net and when the bridles are slipped the net falls back bodily until the fore part, tightly strangled, is drawn in a cone up through the mouth of the ring. The latter does not capsize and turn end on as it does in the Nansen method, but falls back fairly until arrested, still maintaining its original fishing position, by a stray line or preventer, one end of which is attached to the release gear and the other to the towing shackle at the apex of the bridles. Thus, after closing as before, the ring is held in steady control by the bridles from start to finish of the haul. The length of stray line required, which is the distance the net must be allowed to fall back, depends on the circumference of the net at the point where it is to be throttled. If the throttling rope works as a running noose the net must be dropped a distance at least equal to the circumference at the point of closing. On the other hand if the throttling rope is arranged to work " on the bight " the distance need only be half 1 Mackintosh, N. A., and Ardley, R. A. B., 1936, The Royal Research Ship 'Discovery //', Discovery Reports, xiii, pp. loo-ioi. > Cl. .-> CO Cl. J3 O o c J5 M E c o M to u "« E u OJ to to to 4J 0^ c 3 OJ 'C DX] ^ l^ _« '^" £^ b£ 'C 60 -M _C u to C _o M 13 ^C ui 'S tS !l to ■a 4-* o J3 0) g s f ^M « « Cfl c O Ui 1-. OJ u ,C 0" -d z s t4-l o c ^ o ~o ^^ « u g C ^ ' ' 6 C9 o Ui U s N « M s LARGE PLANKTON NETS iii the circumference, although in practice, to make sure that the net is completely strangled, a somewhat longer drop is required. Equally effective throttling can be obtained by either arrangement, but of the two the second is much to be preferred, since with its shorter drop it is less inclined to strain the gear — a matter of major importance where heavy deep-sea apparatus is concerned. Until it has come into play, that is while the net is still fishing, the stray line is coiled small behind the release gear and stopped off with sail twine, otherwise its loose bight might become involved in the twisting warp, or by fouling the closing mechanism precipitate release. The arrangement of the stray line is shown in Figs. 2 and 3. Fig. 3. Large closing nets: details of the arrangement of the gear behind the closing mechanism, a, stray line ; b, throttling rope ; d, bridles ; h, hauling rope to net ring ; r, closing mechanism ; w, warp. It is interesting to recall that an internal closing arrangement, acting in the opposite direction but essentially similar in principle to that which has just been described, had been employed by Kemp^ some ten years previously for sending down large mid-water nets closed and opening them under water. So far the new internal throttling arrangement had only been employed with the young-fish trawl, mounted as already described on a 2-metre ring: whether it could be applied as it stood to the much larger and heavier N 450 with an opening five times the area was a matter for some concern, for to close the N 450 as it would normally be closed, that is by throttling at the canvas band (Fig i ,/) where the circumference is 50 ft., ^ Kemp, S., loc. ctt., p. 202. 112 DISCOVERY REPORTS a drop of at least 30 feet, and probably more, would be required. This was felt to be an undesirably long drop, the shock of which sooner or later might result in the collapse of the heavy frame, if not in the loss of the entire apparatus. With a view therefore to keeping the drop within reasonable limits it was eventually decided to throttle the net in the region of the narrow cod-end, at a point only 15 ft. away from the bucket. By closing at this point the backward fall of the net could be reduced to as little as 12 ft. There was another and equally important reason for choosing such an unusual closing position. With such an arrangement a large part of the net, the part forward of the closing point, must obviously continue to fish after throttling had taken place, and it was hoped that the steadying influence thereby exerted on the bucket behind would contribute materially towards preserving the catch from damage. To allow such a large apparatus to fall back open and then to arrest it, still fishing, in its backward path, would it was realized tax the gear severely ; by the sudden jerk bridles and stray line in particular would be put to an unusually severe test, while the unwieldy ring might well be strained to the point of collapse. To provide against mishap the heavy ring, which if suspended in air from a single point is inclined to buckle under its own weight, was specially stiff^ened by the introduction of a system of four thin steel wires stretched taut across the mouth of the frame, radiating in the form of a cross from a small steel ring in the centre which was to act as a fair-lead for the closing rope (Fig. 2, c, and Plate II). For the cross-wires the ordinary 4 mm. diameter plankton wire was used. It is of exceptional strength but small enough not to cause any serious obstruction in the mouth of the net. Further protection against buckling was provided by employing six bridles of 6 mm. diameter wire instead of the usual four. Obstruction in the mouth of a tow-net, especially that caused by bridles, should of course be reduced wherever possible, and in this respect the efficiency of the N 450 would doubtless be increased by the introduction of a heavier frame than that at present in use ; with a sufliiciently rigid frame there would be no need for so many bridles — three would be enough — and at the same time it would be possible to dispense altogether with the stiffening cross-wires. For the stray line, on the strength of which so much would depend, i J in. circumference wire was used, and a special preventer was provided, connecting the eye of the bridles with the after end of the stray line, in case the towing shackle, which is of rather light construction, should carry away when the whole weight of the fishing apparatus fell suddenly on to it. These arrangements, although not altogether satisfactory at first, with gradual modification eventually achieved their purpose. With the throttling rope rigged to operate internally and so near the bucket, the fore part of the net, as anticipated, con- tinued to fish after the narrow cod-end had been throttled. The net in consequence became a double-acting one, traversing in its path two distinct horizons — a lower prior to closing and an upper during its ascent to the surface — and yielding therefrom a catch in the bucket and another in the bag formed by the constriction of the net 1 5 ft. above. The condition of the material in the bucket, the major concern of the experi- ment, was excellent, there being no sign of the buffeting to which it had previously LARGE PLANKTON NETS 113 been subjected when the net was being closed by the Nansen method. Some of the animals, though taken below 1000 m., were brought to the surface alive and vigorous, a circumstance for which the steady fishing of the forward open end of the net, through controlling the collapsed portion behind, was doubtless largely responsible. There was, however, a serious objection to the internal method of closing when applied so far down the net, for no matter how taut the throttling rope might be drawn a narrow passage would always be left in the constriction through which, it was strongly suspected, part of the catch in the fore part of the net was being carried by the pressure of water back into the bucket behind. This objection was overcome by adopting a closing arrangement which, while retaining the same throttling position 15 ft. away from the bucket, was a compromise between the internal and Nansen methods. The throttling rope, as before, was led inwards through the mouth of the net by way of the fair-lead on the cross-wires, then out again through a small hole in the canvas band, down through a series of rings sewn in one of the webbing supports, and so round the outside of the net. The throttling rope, originally of li in. circumference manilla, was now replaced by a 6 mm. diameter wire — not for the sake of its strength, but because it renders so much more readily round the net than manilla. Thus rigged the net was paid away to a depth of about 1000 m., closed immediately, and hauled rapidly to the surface. The constriction was evidently a good one, blocking all access to the rear, for the whole of the catch was confined to the fore part of the net and the bucket was empty. With the net in this rig a large number of closing hauls, both horizontal and oblique, were conducted without hitch at all depths between 3500 and 400 m. The new closing arrangement, however, simple as it was and thoroughly reliable mechanically, did not meet with complete success, for although the condition of the material in the bucket continued to maintain a consistently high standard, the actual quantity taken, as in the early experiments, was still far from satisfactory. More than once during these opera- tions the bucket came back practically empty, and never under any circumstances — regardless that is to say of the speed, depth, or duration of the haul — was the catch more than a fraction of what would normally be expected in such a large apparatus. On the other hand the catch in the forward open section of the net was always substantially large, generally from ten to twenty times greater than that in the bucket behind. In view of the persistent failure to obtain satisfactory results a number of tests were carried out in which the throttling rope was moved successively farther away from the bucket, and as a result it was shown that the net was being throttled too far aft, part of the catch which ought to have found its way to the cod-end apparently being cut off to appear instead in the open section in front. The explanation of this is not quite clear, but it is possible among other things that the animals these nets are designed to capture, being big and powerful and exceedingly active swimmers, may contrive for a time at least to remain far forward in the net, especially if it is being towed slowly, with the result that if throttling takes place too far to the rear some of them may never reach the bucket. In the arrangement finally adopted the net was throttled at a point half-way between 114 DISCOVERY REPORTS the bucket and the ring and in this position large catches in first-rate condition were obtained in the bucket at all depths down to 2000 m. The modified net in its open and closed positions is shown in Fig. 2. With the rope working externally, as described, this is probably the most suitable arrangement that can be contrived for closing such a large net. Where throttling now takes place the circumference of the net is 32 ft., and a drop of 20 ft., which is not unduly large, is required to effect a satisfactory constriction. There is still, after closing, 30 ft. of the forward coarse-meshed end to keep up the even fishing strain, and this not only steadies the collapsed portion tailing out a further 30 ft. astern, but is enough to "float" the heavy ring so that even with the winch stopped there is little fear of its sinking and fouling the warp. The net, of course, is still a double-acting one, yielding large catches in the I in. netting of the forward section. In general condition the front section material is naturally somewhat inferior to that in the bucket, yet the bulk of it is in a very passable state. The throttling rope it is true might be moved still farther up the net to operate at its normal position, the canvas band, but for various reasons such an arrangement might not prove altogether satisfactory. In the first place it would be necessary to fall back on an internal rope, and it is rather doubtful if the 50 ft. of material composing the canvas band could be constricted internally without leaving, as already explained, an undesirably large hiatus. Further, apart from the much longer drop that would be required, and the increased risk of straining the gear that this must entail, there would be greater difficulty after closing in controlling the heavy ring, unless it were being hove very fast, for the net behind, being completely collapsed, would no longer impart to it the necessary buoyancy. In recent practice the faster oblique method of towing, in which the speed of the ship can be augmented by that of the winch, has superseded the horizontal. The importance of rapid fishing, already suspected, was established at an early stage in these experi- ments, when oblique nets travelling more quickly through the water, but covering com- paratively little ground, produced consistently better results — bigger catches and larger individual animals — than horizontal tows of twice the duration. In spite of its advantage in speed, however, the oblique method has certain obvious limitations, for whereas a horizontal net can be towed indefinitely within the limits of the same restricted horizon, the oblique net, as long as it remains open, must unavoidably widen its horizon, slowly or rapidly according to the speed at which it is being hauled. At first sight therefore it would appear impossible with a fast-moving oblique net to examine a sufficiently narrow horizon without a drastic shortening of the haul — a serious disadvantage, especially when working in deep water, where, since the animal community is so widely scattered, the time of fishing should be as long as possible. The tendency of a large fast-moving oblique net to rise through too wide a horizon can however be effectively controlled without unduly curtailing the actual period of the haul. At normal towing speeds of from i to i| knots the N 450, if hauled obliquely, LARGE PLANKTON NETS 115 will rise through a vertical height approximating to half the amount of wire hove in. For example, if 2000 m. be hauled, the vertical ascent is likely to be in the neighbour- hood of 1000 m. It is possible, however, hauling the same amount of warp, to fish a much narrower horizon than this if the speed of the ship be kept down to about half a knot while the net is coming in. With the ship moving so slowly ahead, the net, for lack of sufficient forward force to give it the necessary lift, is strongly inclined to sink- especially in deep water work when long lengths of wire up to several tons in weight are used— and this downward tendency, acting against the oblique upward pull, greatly reduces the angle of ascent, the horizon in consequence being kept within more reason- able limits. It is of course obvious that oblique fishing from a very slow-moving ship can only be carried out with fast-moving nets, travelling at speeds far in excess of normal, for the usual oblique hauling rate of 10 m. per min. is too slow to overcome the downward tendency of the warp and net, the latter in consequence sinking faster than it can rise. In practice the speed of the ship is reduced as soon as the wire is all out, and hauling begins after a lapse of about a minute, or if necessary several minutes, when it has been ascertained from the steepening of the warp that the net has a strong downward tendency. When working in deep water, 2000 m. is hauled before closing, the rate of haul depending a good deal on the state of the weather. A maximum speed of 40 m. per min. has been reached in calm conditions, but as a rule the average rate is about 35 m. per min., a speed that can be maintained in winds up to force 3 or 4 without straining the gear. At this rate the net can be hauled for an hour, long enough to produce an ample quantity of material from the greatest depth at which the apparatus in its present form has so far been used. The horizon examined rarely exceeds 600 m., for deep-sea work not unduly wide, but with experience it can be kept down to as little as 500 m. or even less (see p. 117 and Fig. 4). Although excellent results have been obtained by this new method of rapid oblique hauling the actual rate of fishing does not at first sight appear to have undergone any marked increase. At its fastest, that is, allowing 40 m. per min. for the net and two- thirds of a knot at most for the ship, it cannot be higher than 60 m. per min., or about 2 knots — no more in fact than the speed at which the net is stated to have been fished in the original horizontal experiments. The success that has attended the recent operations, however, and the repeated failure of the early horizontal closing hauls can only have one satisfactory explanation, that even at 2 knots the net travels more quickly now than it did in the past, when in all probability the horizontal speed was less than it is claimed to have been. It is impossible otherwise to account for the vastly increased efficiency of the net. The increased catching power of the net is perhaps best illustrated by the presence in recent hauls of unusually large deep-sea forms, including among the pelagic prawns species of Acanthephyra 20 cm. long. So far it is true the occurrence of such forms has been rare, yet by employing still faster nets there is no reason why they should not be captured consistently, along with other and even larger animals which doubtless exist. ii6 DISCOVERY REPORTS In oblique fishing, provided the ship is moving ahead slowly enough, practically the only limit to the rate of haul is the bursting point of the net itself. By further strengthening the gear therefore at all vulnerable points, and if necessary constructing the meshing of the net of stronger material, it should be possible to haul at speeds up to 70 or 80 m. per min., giving a total fishing speed of about 3 knots. The main objection to such a big increase in the oblique rate is that the time of fishing, already short enough, would have to be still further reduced; at a rate of 70 m. per min., hauling 2000 m. which is the most that can be hauled without interfering unduly with the vertical horizon, the fishing time would be barely half an hour. For this reason, then, high-speed horizontal towing, if it could be accomplished, would be preferable. Without much alteration of the existing gear, however, there seems to be little hope of obtaining any marked improvement in the present horizontal rate. A much heavier and stronger warp would certainly have to be used, and consequently a bigger and more powerful winch, for the strain of towing such a large net at 3 knots, in view of the enormous weight that would be required to keep it down, would probably be more than tlie existing warp could stand. In the original horizontal experiments a 100 lb. sinker was used at the end of the warp. If double that weight could be employed with safety no doubt some increase in horizontal speed could be obtained without resorting to a larger warp. At the same time various devices might be employed to improve the catching power of the net itself. Non-return pockets might be provided as in the commercial otter trawl, while, as already remarked, much of the existing obstruction in the mouth could be avoided by the introduction of a heavier and more rigid frame. It might reasonably be objected that any further increase in the fishing rate will result in serious damage to the catch, but in view of the admirable condition of the material that has recently been captured there seems to be little ground for anxiety on this account. Slow fishing might well have a more detrimental effect, especially upon the larger and more vigorous animals, which in their efforts to escape from a slow- moving apparatus must no doubt suffer considerable damage through dashing them- selves violently against the sides of the net. On the other hand in a net travelling at high speed even the most active species must tend to be swept rapidly back to the bucket, where, should they escape being stunned or killed outright, they must at least be rendered immobile. Much of the recent experimental work could not have been undertaken without the help of a reliable depth gauge. The instrument now in use, the improved deep-water pattern described by Mackintosh, ^ has given excellent service throughout a long series of hauls extending down to 3500 m. From the records obtained it appears that when a large net is hauled rapidly from a slow-moving vessel it is inclined at first to move almost horizontally through the water when the downward tendency of the warp and net is more or less evenly balanced by the oblique upward force. As the haul progresses the angle of ascent becomes gradually more accentuated, but it is only in the latter stages shortly before closing is about to take place that the net begins to rise steeply. 1 Mackintosh, N. A., and Ardley, R. A. B., 1936, loc. cit., p. 102. LARGE PLANKTON NETS 117 In oblique work of this kind, where the vertical ascent of the net is inclined to vary considerably from one haul to another, it is most important that the exact depth of closing should be known. In recording this depth the depth gauge has proved indispensable, the shock of closing causing the pen to swing violently and so produce a vertical stroke on the paper across the line representing the path of the net. One of the more recently obtained depth gauge charts illustrating the features described above is reproduced in Fig. 4. The chart represents the path of the N 450 during the operation of a deep-water haul in which 4000 m. of wire was paid away Fig. 4. Reproduction of a depth gauge chart showing the path of the N 450 during a 2000 m. obhque haul (i actual size), abc, the path of the net while fishing; c, closing point; s, zero line. For further explanation see text. in all and 2000 m. hauled before closing. The path of the net while fishing is indicated by the letters abc, a being the point where fishing actually began, and c the closing point clearly shown by the vertical stroke of the pen. It will be seen that in this particular haul the net for more than half the time (from a to b) was travelling almost horizontally through the water. Another feature especially well illustrated in Fig. 4 is the remarkably narrow horizon that can sometimes be examined by these large oblique closing nets in spite of the great length of wire that is brought in. In this particular instance the depth gauge reading was 1950-1550 m., showing a vertical ascent for the net of only 400 m. In the ' Discovery II ' the N 450 is worked over the stern through the fair-lead in the rail on the port side. When in daily service it is carried, ring uppermost, flat on the spacious poop, with the closing mechanism attached and the gear all clear for running —the stray line coiled and stopped off, the throttling wire carefully slacked away to ensure that the net will open to its full capacity on entering the water. When shooting, the weight^ is attached to the end of the warp and 20 or 30 m. paid away. The warp, 1 No weight other than the depth gauge, which weighs nearly a hundredweight, is used. ii8 DISCOVERY REPORTS with the stop and heavy closing mechanism attached inboard, is now Hfted out of the fair-lead with a snatch-block and tackle from the poop derrick plumbed directly over- head, and the gear is passed clear over the side and held fast about a foot abaft the rail. The wire is then slacked back into the fair-lead. In passing the gear over the rail the snatch-block is placed in front of the large mechanism, from which it must be backed well clear by means of a whip to the rigging, otherwise by riding down the warp it will come in contact with the apparatus and bring about a premature release. The derrick is now swung amidships and the frame, suspended from a slip-hook provided with a tripping line, is hoisted until it clears the rail when the bucket and net are put over and allowed to drift astern. When all is clear the frame is lowered to the water's edge and tripped, preferably when about three parts submerged to avoid an unnecessary jerk on the release. It helps greatly when shooting if the throttling wire and all six bridles be gathered together into a single bundle and the whole stopped off with a few turns of sail twine. This procedure is a safeguard against mishap as the bridles and wire, if flying about loose, are apt to foul the release while the frame is being lowered into the water. The stops, of course, part readily as soon as the strain comes on the bridles. Throughout this operation the ship is kept moving very slowly ahead, at little more than half a knot, but as soon as the net has entered the water and is streaming well away speed is increased considerably and the warp allowed to run out as rapidly as possible. In a deep haul requiring 4000 or 5000 m. of wire the first 2000 m. as a rule is paid away at from 60 to 80 m. per min., the ship making from 2 to 3 knots. The remainder, however, is run out at an ever decreasing rate as the increasing load over the stern automatically retards the forward speed of the ship. Particular care must be taken in paying out the last 1000 and the last 500 m., for if then allowed to run out too fast, the wire, by virtue of its greatly increased total weight, may over-run the net, and, becoming slack, may kink or fly into a bight round the release. In shallow water work, when the time occupied in paying out is merely a matter of minutes, nets shot open catch next to nothing on the way down. Deep nets on the other hand, taking anything from an hour to an hour and a half to reach their required level, are naturally far more liable to error through fishing on the way out. It is therefore of the greatest importance that in practice they be paid away at a rate which as nearly as possible is just a fraction under that at which the ship is travelling. When the wire is all out and the necessary reduction in the ship's forward speed has been made, hauling begins and is continued without a break to the surface, the mes- senger, as in vertical hauling, being despatched in time to meet and close the ascending net at the moment when the metre-recording dial shows that the required amount of warp has come in. The winch may be stopped for a few seconds while the messenger is being attached to the wire, but it is better not to do so, for even a momentary slackening of the fishing speed might be enough to allow the more active individuals in the catch to escape. In timing the despatch of the messenger allowance is made for its own speed and for the rate at which the net is approaching from below. At the steep warp angles employed in rapid oblique fishing the 22 lb. messenger in use travels about 2000 m. LARGE PLANKTON NETS iig in lo min., the net, hauled at 40 m. per min., travelUng 400 m. in the same time. Thus, if closing be desired on 2000 m. of wire, the messenger must be released at the surface when the recording dial reads 2400 m. Corresponding dial readings can be calculated for any length of wire and for any warp or messenger speed. The moment of closing, even in deep water, can generally be detected on deck either by feeling the wire or by watching the drum of the winch which revolves for about a second at high speed as the net, falling back, momentarily ceases to fish. When it has been ascertained that the net is throttled, the open-meshed forward section only con- tinuing to fish, the hauling rate may safely be increased to between 45 and 50 m. per min. It should not however be more. When bringing the net on board the release gear, hauled slowly up to the rail, is stopped just outside the fair-lead, and when the warp has been raised as before, the large mechanism is hoisted on board with the winch. The warp, slacked back into the fair-lead, is now hove up until the eye of the hauling rope (Fig. 3, /?), with which all large net rings are provided, can be cast adrift outboard, passed through the fair-lead, and hooked on to a light purchase on deck. With the help of the purchase the heavy frame is lifted out of the water and hauled close up to the rail when the large tackle can be hooked on and the frame hoisted to the derrick head. It is now a simple matter to get the net on board. While the frame is being hauled and hoisted the throttling wire must be brought in simultaneously, hand over hand, otherwise it will slack off and allow the catch in the fore section of the net to drop back and mingle with that in the bucket behind. For easier handling the wire is provided at its forward end with a "tail" of 1 1 in. manilla. Throughout the entire operation, while shooting, paying out and hauling, the ship from start to finish is kept dead before the wind. It is impossible in practice to work for long on any other course, for at the low forward speeds employed when hauling all steerage way is lost, and the ship, virtually anchored by the stern, is inclined all the time to come stern on to the wind. When operating long wires under very high tension such a turning movement is apt to have serious consequences, for the warp, leading across the stern, is liable to score deeply into the fair-lead cheeks, and even although the cheeks are recessed, may in a very short time cut its way through and jam on the axle of the horizontal roller. !T 3TA.1'I i«ob lo'i PLATE II Shooting the 4^ m. net. Note the stiffening wires stretched across the mouth of the frame and the central fair-lead to take the throttling wire. At the time when this photograph was taken the net was not rigged for closing but was being used to tow a large and heavy tank specially designed for the capture of live pelagic fish by Prof. A. C. Hardy of Hull. DISCOVERY REPORTS VOL. XVIII PLATE II John BaJe, Son5 A. Cumow. L^ London , THE 4!^- METRE NET [Discovery Reports Vol. XVIII, pp. 121-222, Plates III-VI, October 1938.] CRINOIDEA By D. DILWYN JOHN, M.Sc. Assistant Keeper, Department of Zoology, British Museum (Natural History) CONTENTS Introduction P'^S^ 123 New forms 125 Care of the brood 125 Distribution and relationships 128 Acknowledgements 131 Key TO Antarctic species 131 Systematic account i33 Antedonidae i33 Heliometrinae i33 Zenometrinae 152 Bathymetrinae 164 Isometrinae i74 Notocrinidae i93 Comasteridae i99 Pentacrinoid larvae 199 Antarctic comatulids previously in the British Museum collection . 220 References 221 Plates III-VI following p. 222 T CRINOIDEA By D. Dilwyn John, m.Sc. Department of Zoology, British Museum (Natural History) (Plates III-VI ; Text-figs. 1-24) INTRODUCTION HE Discovery collection of crinoids is entirely of comatulids; no stalked form was taken. It was made between the years 1935 and 1937. There are three hundred and twenty-six specimens of which by far the greater number come from depths of between 100 and 600 m. Two hundred and sixty-nine are from the Antarctic region, the majority from the South American sector, a small number from the Ross Sea. Fifty-four are from the Burdwood Bank and the Patagonian shelf; they are of one species, Isometra vivipara, common to that region and the Antarctic. The remaining three specimens, of Comanthus novae zealmidiae, are from New ERRATA DISCOVERY REPORTS, VOL. XVIII p. 126, footnote: for pp. 40-1 read pp. 44-5 p. 153, eighth line from bottom: for each syzygial pair read syzygial pairs p. 201, fifth line: for fig. 90 read fig. 920 I Phrixometra longipinna var. antarctica n.var. 5 Phrixometra nutrix (Mortensen) i Phrixometra rayneri n.sp. i Subfamily ISOMETRINAE Isometra vivipara Mortensen 55 Isometra flavescens n.sp. 12 Isometra graminea n.sp. 8 Isometra hordea n.sp. 12 Family NOTOCRINIDAE Notocrimis virilis Mortensen 26 Notocrimis mortenseni n.sp. 10 CONTENTS Introduction New forms Care OF THE BROOD Distribution AND RELATIONSHIPS jjg Acknowledgements Key to Antarctic species Systematic account Antedonidae Heliometrinae page 123 • 125 131 131 133 133 133 CRINOIDEA By D. Dilwyn John, M.Sc. Department of Zoology, British Museum (Natural History) (Plates III-VI; Text-figs. 1-24) INTRODUCTION THE Discovery collection of crinoids is entirely of comatulids; no stalked form was taken. It was made between the years 1935 and 1937. There are three hundred and twenty-six specimens of which by far the greater number come from depths of between 100 and 600 m. Two hundred and sixty-nine are from the Antarctic region, the majority from the South American sector, a small number from the Ross Sea. Fifty-four are from the Burdwood Bank and the Patagonian shelf; they are of one species, Isometra vivipara, common to that region and the Antarctic. The remaining three specimens, of Comanthiis novaezealandiae, are from New Zealand. The following is a complete list of the unstalked crinoids known from the shallower waters of Antarctic seas. The names of the two of which no specimens were secured are included in brackets. The number of specimens taken of each of the others is shown. Family ANTEDONIDAE Subfamily HELIOMETRINAE Promachocrimis kergiielensis Carpenter 163 specimens Florometra mawsoni A. H. Clark 8 „ Florometra antarctica n.sp. 3 „ Anthometra adriatii (Bell) 14 ,, {Solanometra antarctica (Carpenter)) Subfamily ZENOMETRINAE Eumorphometra aurora n.sp. i ,, Emnorpfiometra fraseri n.sp. 1 ,, Eumorphometra marri n.sp. i ,, {Eumorphometra concinna A. H. Clark) Kempometra grisea n.g. and sp. 2 ,, Subfamily BATHYMETRINAE Phrixometra longipinna var. antarctica n.var. 5 ,, Phrixometra nutrix (Mortensen) i „ Phrixometra rayneri n.sp. i „ Subfamily ISOMETRINAE Isometra vivipara Mortensen 55 ,, Isometra Jlavescens n.sp. 12 ,, Isometra graminea n.sp. 8 ,, Isojnetrahordea n.sp. 12 ,, Family NOTOCRINIDAE Notocrinus virilis Mortensen 26 ,, Notocrinus mortenseni n.sp. 10 ,, J24 DISCOVERY REPORTS Nearly two-thirds of the specimens from the Antarctic are of Promachocrinus kerguelensis, the largest and most common of Antarctic crinoids. The bigger of the remaining species — those of the family Notocrinidae and the subfamilies Heliometrinae and Isometrinae — are represented by moderate or large numbers of specimens. Of the smaller species belonging to the subfamilies Zenometrinae and Bathymetrinae there are only one or two specimens, often imperfect, of each, despite the numerous dredgings made by the Discovery Committee's vessels. It seems probable that many more small species remain undiscovered in moderate depths in Antarctic seas. I have included notes on the two species, Solanometra antarctica and Eiimorphometra conchma, which are known to occur in the Antarctic but which are not represented in the Discovery collection; and on Eiimorphometra hirstita from Marion Island. I have also partially redescribed from their types in the British Museum collection three non-Antarctic species, Isometra lineata, I. ongustipinna and Phrixometra longipinna, to which certain of the Antarctic species are related. How well many of the specimens are preserved may be seen by referring to the plates of this report. The larger were fixed by gently inducing them while still alive into tubes just wide enough to hold them, and then by adding strong spirit. Myzostomtim was found on Promachocrimts kerguelensis, the two species of Florometra, Aiithometra adriani, Isometra flavescens and Notocrinus mortenseni. In the lists of stations under each species the gear in which the specimens were caught is shown by means of symbols: OTL, DLH, etc. These symbols are explained in the introductions to the Station Lists in this series of reports (Vol. i, pp. 3-5, 1929 ; Vol. in, p. 4, 1930; Vol. IV, pp. 3-4, 1932) except for the following: DS, small dredge; DRR, a large dredge bag attached to the frame of a Russel bottom net. I have examined the Antarctic comatulids, those collected by the National Antarctic ('Discovery') Expedition, 190 1-4 and the British Antarctic (' Terra Nova ') Expedition, 1910, in the British Museum collection. They were reported on by Bell in 1908 and 1917 respectively. He recorded Antedon antarctica {= Solanometra antarctica) which was not represented and did not recognize two species, Florotnetra mawsoni and Noto- crinus virilis, which had not been described at that time but which were represented; he gave, too, only a few of the localities at which specimens were taken. I give at the end of this paper (p. 220) a complete record of the Antarctic comatulids taken by those expeditions. The collection contains about fifty-eight pentacrinoid larvae. About twenty of them are of Isometra vivipara and are attached to the upturned cirri of the females. The development of this species has been described by Mortensen (1920) and nothing is added here. Clark (1921) has described a long series of pentacrinoid larvae of Promachocrinus kerguelensis. There are three in the present collection (p. 200). The remaining larvae are of much greater interest, for they were not previously known. They are described in detail (pp. 202-219). There is a series of fifteen, the youngest NEW FORMS 125 a prebrachial stage, the oldest with three whorls of cirri, which I believe to be of the new species Isometra hordea. There are twenty, the youngest with the beginnings of arms, the oldest with only two whorls of cirri, which are of Notocrinus virilis. The fully- formed larva of this species has been described by Mortensen (1920). NEW FORMS The proportion of new forms is very large : I have described ten new species and one new variety. One of the new species, Notocrinus mortenseni, belongs to the peculiar genus for which Mortensen found it necessary to establish a new family, the Noto- crinidae (Mortensen, 1918); only the type species, A^. virilis, also from the Antarctic and abundantly represented in the present collection, was previously known. Clark (1937) has recently described a new species of Florometra, F. mawsoni, from the Ant- arctic. It was taken by the Discovery vessels and they also found three specimens of a new species, F. antarctica. There are three new species of Isometra. Two resemble /. vivipara but differ too strongly from it to be described as identical ; the third, /. hordea, is undoubtedly a distinct form. The other five new species are of smaller forms, each represented by a few specimens only. For one of them, unique among Antarctic comatulids in that it lacks oral pinnules, I have had to create a new genus, Kempometra. Three appear to be new species of Eiimorphometra of which two species, one from the Antarctic and one from Marion Island, were previously known. I have described a new species and a new variety of Phrixometra, a genus previously known from one species taken off the River Plate. CARE OF THE BROOD The main interest of the collection is that it contains so many species which care for the brood ; which are, in other words, viviparous. The vast majority of comatulids shed their eggs directly into the sea. Only three viviparous species were previously known. In this paper I add to that number five new species and one new variety from the Antarctic, and two described species, one from the Antarctic and one from off the River Plate, which their authors had not recognized as viviparous. Thus there are now eleven known viviparous comatulids. The three previously known species were Isometra vivipara from the Patagonian shelf, the Burdwood Bank and the Antarctic; Phrixometra nutrix from the Burdwood Bank ; Notocrinus virilis from the Antarctic. All three were taken by the Swedish South- Polar Expedition and they and their viviparous habits were described by Mortensen (1918, 1920).^ 1 The care of the brood in Isometra vivipara had been described long before by K. A. Andersson (1904), who did not however recognize the species as new but regarded it as Jlntedon hirsuta { = Eiimorptiomcfra hirsuta). 126 DISCOVERY REPORTS The Discovery collection contains many specimens of Isometra vivipara and Noto- crinus virilis ; there is one male of Phrixometra nutrix from the Bransfield Strait in the Antarctic. The following table includes these three species and the seven species and one variety which are shown in this paper to be viviparous ; it is a complete list of the known viviparous crinoids.' The localities from which each is recorded are shown. Ajitarctic Outside the Antarctic Burdwood Bank Patagonian Shelf Off River Plate Eumorphometra concinna A. H. Clark + _ _ _ Kempomeira grisea n.sp. PhrixomeUa longipinna (Carpenter) P. longipinna var. antarctica n.var. P. nutrix (Mortensen) Isometra vivipara Mortensen + + + + + + + + + /. flavescens n.sp. + - - — 1 /. graminea n.sp. /. hordea n.sp. Notocrinus virilis Mortensen + + + - - - A', mortenseni n.sp. + — - - The table includes two species, Eumorphometra concinna and Phrixometra longipinna, which were not taken by the Discovery vessels. I owe my knowledge that the former is viviparous and the opportunity of describing its brood-pouches to Mr A. H. Clark: when he heard from me how many viviparous forms I was finding he re-examined Eumorphometra concinna, saw that the female had brood-pouches and immediately sent me some. Carpenter had described " much swollen ovarian " sacs in Phrixometra longi- pinna ; I found them to be brood-pouches. The table shows that of the eleven viviparous forms eight are known only from the Antarctic. One is known from the Antarctic and the Burdwood Bank, another from the Antarctic, the Burdwood Bank and the east coast of South America as far north as the River Plate. One is known from off the River Plate only. If the table be compared with the list on p. 1 23 it will be seen that of the nineteen comatulids known from the Antarctic ten, over 50 per cent, are viviparous. More than 600 species of unstalked crinoids are known from other seas and only five of them, two of which also occur in the Antarctic {Isometra vivipara and Phrixometra nittrix), are viviparous ; the other two are P. longipinna and the two viviparous forms from southern Australia. The percentage of viviparous forms from the Antarctic may be considerably higher. All the Antarctic species of the Notocrinidae and the Isometrinae, but only four of the eight species of the Zenometrinae and Bathymetrinae, are viviparous. But the re- ^ Since this was written two species from southern Australia have been described as viviparous (H. L. Clark, 1938, Echinoderms from Australia, Mem. Mus. Comp. Z06I., LV, pp. 40-1). CARE OF THE BROOD iz7 maining four species of the Zenometrinae and Bathymetrinae are known from single specimens and each is a male. It may be that when the females are found some or all of these species will be discovered to be viviparous. The five large species of the subfamily Heliometrinae are not viviparous. Dr Mortensen has shown in a recent Discovery Report (1937) that there is in Ant- arctic ophiuroids a similar very high percentage of viviparous forms. The Antarctic ophiuroid fauna is much richer in species than the crinoid fauna. About 50 per cent of them are viviparous. The highest percentage elsewhere is 15, in New Zealand. When Mortensen (1918, p. 2) made known the first three viviparous forms he re- marked that it added to their interest that each had a separate way of caring for the brood ; he was indeed fortunate, for he had three most interesting species before him. The methods of the eleven viviparous forms now known are compared below. In only one group, the species of Isometra, does the viviparous habit appear to have led to a modification of the hard parts. The segments of the genital pinnules of the females are strongly expanded to arch over and protect the ovary and brood-pouch. There is a much smaller expansion of the segments carrying the testes in the male — so much smaller that whereas in almost all comatulids males and females are indis- tinguishable to the naked eye, in the species of Isometra they may be recognized at a glance. In all the viviparous species there is in the female a brood-pouch beside each ovary. In Notocrinus virilis and N. mortensefii the gonads have a unique position. They do not lie, as in all other crinoids, along the pinnules, but in the axils between the pinnules and the arms, largely on the arms. The brood-pouches of the females are distal to the ovaries. The two species form an interesting contrast in the extent to which they protect the brood. A brood-pouch of a big A^. mortenseni may contain as many as ninety-two embryos in all stages of development and varying in size from 0-25 to 0-48 mm. The largest have five broad bands of cilia ; they presumably go through a free-swimming stage before metamorphosing into pentacrinoids. The brood-pouches of A^. virilis each contain only one to three, usually two, embryos, all at the same stage of development. They are up to 2 mm, in length, that is, four times as long as those of A^. mortenseni, and have no trace of ciliated bands. It must be supposed that they drop out of the brood- pouches to the sea floor and change into pentacrinoids there. A series of pentacrinoids is described on pp. 210-219. In the other viviparous crinoids the brood-pouches lie alongside, or beyond, the ovaries on the pinnules. They lie alongside the ovaries and always on their distal sides, i.e. on the side of the pinnule nearest to the arm from which it springs, in Eumorpho- metra concinna, Phrixometra longipinna, P. longipinna var. antarctica and the four species of Isometra. In these species each brood-pouch contains a fair number of small embryos which possess ciliated bands. (I can see no ciliated bands in those of Phrixometra longipinna, but I think that is because they are too young.) It is probable that in these species, as in Notocrinus mortenseni, the embryos pass some time in a free-swimming stage before 128 DISCOVERY REPORTS metamorphosing into pentacrinoids. Mortensen has described how short the free- swimming stage is in Isometra vivipara : the embryos travel no farther than the upturned cirri of their mothers before settling down and turning into pentacrinoids. This has not been observed in any of the other species. In Phrixometra lo/igipi/ina and one of the females of the var. antarctica all the embryos in the brood-pouches are at the same stage of development. In the other species and the other female of P. longipinna var. ant- arctica they are, as in Notocrinus mortenseni, at various stages of development. In the single known female of Phrixometra nutrix the brood-pouches lie on the oral side of some pinnules, on the aboral side of others. In Kempometra grisea the brood- pouches are on the aboral side of the pinnules, but they lie for the most part beyond, not alongside, the ovaries. In these two species the care of the brood is carried even farther than in Notocrinus virilis. Phrixometra nutrix protects the young throughout their larval existence so that they leave the parent as young comatulids. It seems probable that Kempometra grisea does the same. The larvae of Phrixometra nutrix change into pentacrinoids in the brood-pouch. The stalk is attached to the wall of the brood-chamber or to the pinnule segment on its floor ; the head projects through a slit in the wall. In the single known female the brood- pouches hold one or two pentacrinoids each ; where there are two they are at the same or at diff'erent stages of development. No brood-pouch contains developing embryos (Mortensen, 1918). In the two females of Kempometra grisea in this collection I found one large embryo in one pouch, two in another, a pentacrinoid larva in a third. The pentacrinoid is still completely enclosed in the pouch, which is not ruptured. No pentacrinoid emerges from any brood-pouch as in Phrixometra nutrix, nor is one attached to any part of the body. The ovary contains few, four to nine, very large eggs; they are up to 0-6 mm. long. It is interesting to see that in those species in which the care of the young is carried the farthest — Notocrinus virilis, Phrixometra nutrix and Kempometra grisea — the number of young produced is the smallest. DISTRIBUTION AND RELATIONSHIPS The table below shows the localities from which the Antarctic comatulids are known. The localities within the Antarctic are divided into "continental coasts and adjacent islands" and "outlying islands". In the former category the term "and adjacent islands" has an application only to the Weddell Sea sector in which the South Shetland Islands, separated from the shores of Graham Land by neither great distance nor deep water, are grouped with it. No crinoid has yet been named from the South Orkney Islands though the ' Scotia ' took feather-stars in Scotia Bay (Wilton, Pirie and Brown, 1908, p. 21). The South Sandwich Islands, South Georgia and the Shag Rocks are outlying islands in the Weddell Sea, or South American, sector. Heard Island and Kerguelen are in the Indian Ocean sector. Kerguelen lies on the extreme northern edge of the Antarctic area : the Antarctic convergence, the boundary between the sur- DISTRIBUTION AND RELATIONSHIPS 129 face waters of the Antarctic and sub-Antarctic zones, passes through it (Deacon, 1937, p. 23 and fig. 5). Kerguelen and the adjacent islands may, however, be regarded as a sub-Antarctic district (see pp. 1 30-131). Antarctic localities Non-Antarctic localities Continental coasts and adjacent islands Outlying islands South America Weddell Sea sector Indian Ocean sector Ross Sea sector South Sandwich Islands South Georgia Shag Rocks Heard Island Kerguelen Burd- wood Bank Patagonian Shelf p. kergtielensis F. mawsoni F. antarctica A. adriani S. antarctica E. aurora E. fraseri E. jnarri E. concinna K. grisea P. longipinna var. antarctica P. nutrix P. rayneri I. vivipara I. flaztescens I. graminea I. hordea N. virilis N. mortenseni + + + + + + + + + + + + + + ^ H H H 1- + + + • + + H H h + + + + + + + + + The table shows that seventeen of the nineteen Antarctic comatulids occur in the Weddell Sea sector (in the wider sense, i.e. including outlying islands) ; six in the Indian Ocean sector, including Kerguelen and Heard Island ; five in the Ross Sea sector. Four species, Promachocriniis kergtielensis, Florometra mawsoni, Anthometra adriani and Notocrinus virilis, occur in all three sectors and may therefore be presumed to be circumpolar. They are all large species and therefore the most likely to be taken by collecting expeditions. This and the fact that more collecting, and more careful col- lecting, has been done in the Weddell Sea sector than in any other makes me think that many of the species now known from there alone may later be found elsewhere around the continent ; in other words I think it premature to discuss the distribution of the species within the Antarctic. Of the species that are known to be circumpolar Pro- machocrinus kergtielensis has been found at all the Antarctic localities where collecting has been done except the Shag Rocks ; the other three are, so far as is known, confined to the coasts of the continent and the adjacent islands. The most interesting feature of the table is that it shows that two Antarctic species do occur outside the Antarctic and that they occur towards or along the east coast of South America: Phrixometra nutrix on the Burdwood Bank, Isometra vivipara on the Burdwood Bank and on the Patagonian shelf as far north as the River Plate. There are, I think, other reasons for believing that the shallower water crinoid fauna I30 DISCOVERY REPORTS of the Antarctic is related to that of the extreme south of South America. The Phrixo- metra, of which the Discovery vessels took five specimens, one at South Georgia and four in the Bransfield Strait, is so like Carpenter's Phrixometra longipinna from off the River Plate that I have described it as a variety. The only other species of the genus are P. nutrix, and P. rayneri from the Shag Rocks. There are four species of Isometra in the Antarctic, one of which also occurs on the east coast of South America. Two other species are known, each described from a single Challenger specimen taken off the River Plate. These facts point to a relationship between the shallow-water crinoid faunas of the Antarctic and of the east coast of the extreme south of South America. There appears to be a relationship with the west coast of South America too. Of the four genera of the Heliometrinae represented in the Antarctic three, Pro- machocrinus, Anthometra and Solanometra, are monotypic and the species are confined to the Antarctic. The fifth, Florometra, of which there are two species in the Antarctic, has ten other species ranging northwards from Cape Horn along the west coasts of South and North America to the Behring Sea, westwards along the Aleutian Islands, and southwards to the Pacific coast of Japan. It appears from these facts that the littoral crinoid fauna of the Antarctic is related to that of the south of South America. Workers on the other four classes of echinoderms have arrived at the same conclusion for those classes. Mortensen (1910, p. 100) wrote of echinoids: "The South American (Patagonian-Chilean) littoral region must be re- garded as the centre of the whole Antarctic — sub-Antarctic region ; a very great propor- tion of the echinoids occurring in the Patagonian region are known from this region alone, while others are also known from the Antarctic region." Koehler (191 2, p. 242), after comparing the littoral asteroid, ophiuroid and echinoid faunas of the Antarctic, the Magellan region of South America, the Kerguelen district, New Zealand and the Cape concludes: "II ressort clairement de cette discussion que la faune antarctique des Asteries, Ophiuries et Echinides, — du moins en ce qui concerne les especes littorales, — a son point de depart dans la faune subantarctique des cotes de la pointe de I'Amerique du Sud. Les comparaisons que nous avons faites, la composition et les rapports des faunes etablissent ce fait d'une maniere indiscutable." Ekman (1925, p. 187) writing of holothurians concludes: "Hier mochte ich nur die Meinung aussprechen, dass die von anderen Forschern hervorgehobene, verhiiltnismassig nahe Verwandtschaft der ant- arktischen Litoralfauna mit derjenigen des Magellangebietes auch betreffs der Holo- thurien Giiltigkeit hat, und dass allem Anschein nach beide aus einem gemeinsamen Entwicklungszentrum herstammen." There is only one other species of Eiimorphometra than those listed above; it is E. hirsuta (Carpenter) from Marion Island. Marion Island belongs to the Kerguelen group of islands comprising Kerguelen itself. Heard Island, the Crozets and Marion Island. It possesses in more groups of animals than one a characteristic littoral fauna, partly related to, or having species in common with, those of two other areas, the KEY TO SPECIES 131 Antarctic and the Magellan region of South America. So far as three classes of echino- derms — asteroids, ophiuroids and echinoids — are concerned the fauna of the Kerguelen group is much more nearly related to the fauna of the Magellan region than to that of the Antarctic (Mortensen, 1910, p. 100; Koehler, 1912, pp. 221-7). But Regan (1914, p. 36), having studied the fish faunas of the three regions, describes the Kerguelen group as an Antarctic district; and Norman in two papers on more recent collections of fishes confirms his views (1937, 1938). He says (1938, p. 100): "it is clear that, although the coastal fish fauna of the Kerguelen district shows certain features of resemblance to that of the Patagonian region and the Antipodes, its affinities are mainly with that of Antarctica." Since no crinoid is known to occur both in the Kerguelen and Magellan districts, whereas two {Promachocrinus kerguelensis and Solanometra antarctica) are known both from Kerguelen and Heard Islands and the Antarctic, I have treated those islands, with the above reservations, as outliers of the Antarctic. No other species of Notocriniis than virilis and mortenseiii are known. ACKNOWLEDGEMENTS It is a great pleasure to acknowledge in the first place the constant and generous help I have received from Mr A. H. Clark. When he knew that I was finding many viviparous species he re-examined his co-types of Emnorphometra conciniia, found that it too was viviparous, and sent me brood-pouches so that I might describe them. He has kindly compared my single specimen of E. aurora with E. concinna and has found that they are not, as I thought they might be, identical. He examined for me a specimen of Kempomeira grisea, suggested the diagnosis of the genus which I have used, and helped me to decide upon the position of the genus. Dr Mortensen has helped me by comparing a specimen of my hometra graminea with one of his specimens of /. vivipara, taken by the Swedish Antarctic Expedition, from the Graham Land region; and Professor Sixten Bock of Stockholm kindly sent me, for examination and comparison, all the remaining specimens of /. vivipara taken by the Swedish Expedition. I take this opportunity of thanking them both. The drawings for the text-figures were made by Miss E. C. Humphreys. KEY TO THE COMATULIDS KNOWN TO OCCUR IN DEPTHS OF o-iooo METRES IN ANTARCTIC SEAS A Oral pinnules long and whip-like; P^ of 30-60 or more segments; P, usually as long; large species. B 6-1 1 rays and 12-23 arms— usually 10 rays and 20 arms Promachocrinus kerguelensis. BB 5 rays and 10 arms. C Ossicles of the division series and the brachials with a conspicuous and usually high narrow median keel; cirri of large specimens of 60-80 segments; P, much shorter than Pi Antiiometra adriani. 132 DISCOVERY REPORTS CC Division series and brachials not as above; cirri of not more, usually much less, than 40 segments ; P, about as long as Pj . D Pj of 28-40 segments; cirri of not more than 31 segments; lower brachials with an abruptly elevated spinous patch in the distal portion of the dorsal surface or with the distal edge everted at right angles and strongly pro- duced ... ... ... ... ... ... ... Florometra mawsoni. DD Pj of more than 40 segments; lower brachials smooth or, exceptionally, elevated into small and low spinous patches. E Pj of about 60 segments; cirri of 25-40 segments; lower brachials smooth; brachials beyond third syzygy all short, much broader than long ... ... ... ... ... ... ... Solanometra antarctica. EE Pj of 44-50 segments; cirri of 23-32 segments; lower brachials may have small spinous patches; brachials beyond third syzygy not much broader than long ... ... ... ... ... Florometra antarctica. AA Pj of less than 20 segments, usually short. B Third and some of succeeding segments of genital pinnules strongly expanded to cover the gonads; oral pinnules short with their basal segments attached to the disk or to the arm by webs of tissue. C Pg shorter than Pj . D Ventral surface of disk plated; P3 is first genital pinnule Isometra flavescens. DD Ventral surface of disk naked; P4-P6, usually Pg, first genital pinnule Isometra vivipara. CC Pg as long as or longer than Pj. D Longest cirri of up to 70 segments; large robust species Isometra hordea. DD Cirri of 28-43 segments; fragile species of medium size Isometra graminea. BB No segments of genital pinnules strongly expanded. C Gonads not on pinnules but in axils between pinnules and arms; perisome plated; oral pinnules short with their basal segments attached to the disk or to the arms by webs of tissue. D Cirri long, of 36-76 short segments of nearly uniform length Notocrirms virilis. DD Cirri shorter, of 21-32 segments of which some of the proximal are longer than the remainder ... ... ... ... ... Notocrinus mortenseni. CC Gonads in normal position on genital pinnules; perisome not plated; oral pinnules (where present) entirely free. D Pj and Pa absent ; P3 a genital pinnule ... ... Kempometra grisea. DD Oral pinnules present. E Cirri of less than 20 (12-19) segments. F First genital pinnule (Po or Pg) not much more than half as long as Pj ... ... ... ... Phrixometra longipinna var. antarctica. FF First genital pinnule (Pg or P3) about as long as Pj . G Cirri about XLV ... ... ... Phrixometra nutrix. GG Cirri about XXX ... ... ... Phrixometra rayneri. EE Cirri of more than 20 segments (of up to 40 segments). F Cirri of 30-40 segments ... ... ... Eumorphometra fraseri. FF Cirri of less than 30 segments. G Cirri in 2 or 3 closely placed alternating rows; Pj longer and more massive than Pg; in the lower genital pinnules the segments carrying the gonad are slightly expanded Eumorphometra marn\ HELIOMETRINAE i33 GG Cirri in lo slightly irregular columns; Pj not more massive than Pgi segments of genital pinnules not expanded. H Elements of division series and brachials with central portion abruptly elevated and prominently spinous; axillaries shorter (broader than long) Eumorphometra concinna. HH Elements of division series and brachials not as above; axillaries longer (longer than broad) Eumorphometra aurora. SYSTEMATIC ACCOUNT Family ANTEDONIDAE Subfamily HELIOMETRINAE Genus Promachocrinus Carpenter Promachocrinus kerguelensis Carpenter (Plate III, fig. i) Promachocrinus kerguelensis Carpenter, 1888, p. 350, pi. i, figs, i a-d, pi. Ixx; Clark, 1915 a, pp. 128-34, pis. iii-v, and list of earlier references with synonymy; Clark, 191 5 &, many references, fig. 505; Bell, 1917, p. 2; Mortensen, 1918, pp. 18-20, fig. 16; Clark, 1921, numerous references including pp. 530-57, figs. 881-937 (on pentacrinoid young), pi. iv, figs. 1001-2, fig. 807; Mortensen, 1925 b, p. 2; Gislen, 1928, p. 11 ; Grieg, 1929 a, p. 3; 1929 b, p. 3; Clark, 1929, p. 662; Bernasconi, 1932 a, pp. 29-35, %s. 1-3, pi. i; 1932 b, pp. 86-7, fig.; John, 1937, p. 9; Clark, 1937, pp. 8-9. Promachocrinus vanhoffeniamis Minckert, 1905, pp. 496-501, figs, i and 2. Promachocrinus joubini Vaney, 1910, pp. 158-62, figs, i and 2. The stations at which it was taken are divided into five groups, according to locality, below. South Georgia St. MS 14. 17. ii. 25. East Cumberland Bay. 190-iiom. Gear DS. Two small specimens. St. MS 71. 9.111.26. East Cumberland Bay. 1 10-60 m. Gear BTS. Twenty-eight specimens, mostly large. St. MS 74. 17. iii. 26. East Cumberland Bay. 22-40 m. Gear NCS-T. Fragments. St. 27. 15. iii. 26. West Cumberland Bay. iiom. Gear DL. Bottom: mud and rock. One specimen; one pentacrinoid larva. St. 39. 25. iii. 26. East Cumberland Bay. 179-235 m. Gear OTL. Bottom: grey mud. One specimen. St. 42. i.iv. 26. Oflt mouth of Cumberland Bay. 120-204 m. Gear OTL. Bottom: mud. Eight specimens. St. 123. i5.xii. 26. Off mouth of Cumberland Bay. 230-250 m. Gear OTL. Bottom: grey mud. One specimen. St. 144. 5.1.27. Off mouth of Stromness Harbour. 155-178 m. Gear OTL. Bottom: green mud and sand. Six specimens. St. 148. 9. i. 27. Off Cape Saunders. 132-148 m. Gear OTL. Bottom: grey mud and stones. Four specimens. St. 149. 10. i. 27. Mouth of East Cumberland Bay. 200-234 m. Gear OTL. Bottom: mud. Five specimens. St. 152. 17.1.27. 53° 51' 30" S, 36° 18' 30" W. 245 m. Gear DLH. Bottom: rock. Two specimens. 134 DISCOVERY REPORTS St. 156. 20.1.27. 53° 51' S, 36° 21' 30" W. 200-23601. Gear DLH. Bottom: rock. Eight specimens. St. 345. 8. ii. 30. 55° 20' S, 34° 47' 30" W. 180 m. Gear N 70 V. Bottom: small stones and shells. One specimen. South Sandwich Islands St. 363. 26. ii. 30. 2-5 miles S, 80° E of south-east point of Zavodovski Island. 329-278 m. Gear DLH. Bottom: scoria. Three specimens. St. 366. 6. iii. 30. 4 cables south of Cook Island. 77-152 m. Gear OTL. Bottom: black sand. Six specimens; i pentacrinoid larva. St. 371. 14. iii. 30. I mile east of Montagu Island. 99-161 m. Gear OTL. Five specimens. Bransfield Strait region St. 170. 23.ii. 27. Off Cape Bowles, Clarence Island. 61° 25' 30" S, 53° 46' W. 342 m. Gear DLH. Bottom: rock. Six specimens. St. 172. 26. ii. 27. Off Deception Island. 62° 59' S, 60° 28' W. 525 m. Gear DLH. Bottom: rock. Two specimens. St. 175. 2. iii. 27. Bransfield Strait. 63° 17' 20" S, 59° 48' 15" W. 200 m. Gear DLH. Bottom: mud, stones and gravel. Three specimens. St. 177. 5. iii. 27. 27 miles south-west of Deception Island. 63° 17' 30" S, 61° 17' W. 1080 m. Gear DLH. Bottom: mud, coarse sand and stones. Eight specimens; one pentacrinoid larva. St. 1948. 4.1.37. 60° 49-4' S, 52° 40' W. 490-610 m. Gear DRR. One specimen. St. 1952. II. i. 37. Between Penguin Island and Lion's Rump, South Shedands. 367-383 m. Gear DRR. Bottom: soft mud. Forty-three specimens. St. 1955. 29. i. 37. 61° 35-1' S, 57° 23-3' W. 440-410 m. Gear DRR. Two specimens. St. 1957. 3. ii. 37. 7 miles east of Cape Bowles, Clarence Island, South Shctlands. 830 m. Gear DRR. Bottom: rough, stony. One specimen. 785-767 m. Gear DRR. Bottom: stones. One specimen. West coast of Graham Land St. 180. II. iii. 27. 1-7 miles west of north point of Gand Island, SchoUaert Channel, Palmer Archipelago. 160-330 m. Gear OTL. Bottom: mud and stones. Five specimens. St. 181. 12. iii. 27. SchoUaert Channel, Palmer Archipelago. 64° 20' S, 63° 01' W. 160-335 m. Gear OTL. Bottom: mud. Three specimens. St. 182. 14. iii. 27. SchoUaert Channel, Palmer Archipelago. 64° 21' S, 62° 58' W. 278-500 m. Gear OTL. Bottom: mud. Six specimens. St. 599. 17.1.31. Adelaide Island. 67° 08' S, 69° 06' 30" W. 203 m. Gear DLH. One specimen. The Ross Sea St. 1644. 16.1.36. Bay of Whales. 78° 24-8' S, 164" 10-3' W. 626 m. Gear BNR. Bottom: rocks and mud. One specimen. St. 1652. 23. i. 36. Ross Sea. 75° 56-2' S, 178° 35-5' W. 567 m. Gear DRR. Nineteen speci- mens. St. 1658. 26. 1. 36. Off Franklin Island, Ross Sea. 76° 09-6' S, 168° 40' E. 520 m. Gear DRR. Four specimens, one very young. Historical. Carpenter (1888, p. 348) included in his definition of the genus Pro- machocrinus the words "ambulacra. . .not provided with any definite skeleton". Minckert (1905, p. 496, figs, i, 2), on the basis of two fully grown specimens from the collection of the Deutsche Siidpolar-Expedition, described the new species P. van- HELIOMETRINAE i35 hoffeniamis differing from P. kerguelensis in having: (i) longer cirrus segments, (ii) long and slender axillaries and second brachials, (iii) no lateral notch between the radial and the costal, and (iv) well-developed side-plates along the pinnule ambulacra. Vaney (1910, pp. 158-62, figs, i, 2) described the new species P. joubini, from one specimen from the west of Graham Land. It is described as being distinguished from P. kerguelensis by having clusters of spines on the lower brachials, by the position of its syzygies, and by the possession of a naked dorsal pole. Like P. kerguelensis it has rhombic axillaries and lacks an ambulacral skeleton ; its cirri are tolerably like those of P. vanhoffeniamis. Vaney considered it to be nearer the former than the latter. Clark (191 5) regards both Minckert's and Vaney 's species as invalid, believing the specimens upon which they were based to have been immature P. kerguelensis. Mortensen (1918, p. 19) is not satisfied that Clark is correct, more particularly in re- garding P. vanhoffeniamis as identical with P. kerguelensis ; nor is he convinced that the presence of an ambulacral skeleton is a sign of immaturity, a doubt shared by Grieg who is otherwise inclined to accept Clark's opinion (Grieg, 1929 a, p. 4; 1929 b, p. 4). My examination of the present collection of 163 specimens from low and high lati- tudes in the Falkland sector of the Antarctic zone, and of thirty-eight specimens from the Ross Sea, and of one or two taken by the ' Challenger ' near Kerguelen and Heard Island, has convinced me that there is only one, very variable, species. Description. In the great majority of specimens there are 10 rays and 20 arms but individuals with 12, 16, 17, 18, 19, 22 and 23 ^ arms occur. The arms of large specimens are commonly 180-200, and may be up to 250, mm. long. The centrodorsal is usually a large high rounded cone, but it is variable. It may be more sharply conical; in smaller specimens it is often a moderately low cone. It is closely covered with cirri which may, exceptionally, number 200 or more. The cirrus sockets are arranged in slightly irregular alternating rows, so that those of alternate rows make slightly irregular columns. The dorsal pole is usually of medium size, smooth and rounded. In large specimens it is sometimes sunken and rough. It may, in either large or small specimens, be a sharply triangular or a rough truncated pillar-like pro- jection. It may be very small or absent : though this is most often so in small specimens it may be so in large. Cirri : up to CC or more. There is always a sharp contrast between the apical cirri and those of the more ventral part of the centrodorsal : the latter are long, of up to 47 or, exceptionally, 65 segments ; the former very much shorter and of fewer segments. The contrast is greatest in large specimens ; the longest peripheral cirri may be three or more times as long as the shortest apical cirri. They may be up to 130 mm. long. The segments of the cirri of different specimens vary greatly in length. I have seen none in which those of one cirrus are so nearly equal to one another as in Carpenter's figure. The range of variation is illustrated by the extremes shown in Fig. i ; ^i and a^ are a long and a short cirrus of a specimen with cirri of short segments, and a^ and a^ are 1 Mortensen, 1918, p. 19. 136 DISCOVERY REPORTS those of a specimen with cirri of long segments. The cirrus Ui is of 52 segments and is nearly 75 mm. long. The first three segments are short, the fourth is longer than broad ; the fifth and sixth are nearly twice as long as broad, the seventh to about the twentieth or twenty-fifth twice as long as broad. The remainder gradually decrease in length until the most distal are as broad as long. In the segments following the first ten or so the distal edge projects slightly on the dorsal side ; the projection is most marked, but not Oo Fig. I. Piomachociiims kcrguelensis. aj_2, long and short cirri of a specimen with cirri of short segments, X 1-25. a3_4, the same of a specimen with cirri of long segments, x 1-25. b, distal segments of a short cirrus, X9. Strong, in the distal segments. All but the basal segments are compressed from side to side, the distal most strongly. The small apical cirrus of the same specimen, a, , is of 24 segments and 1 5 mm. long. The basal segments, the first two or three, are broader than long; the proximal are longer than broad but not nearly so much so as in the long cirri. The distal segments are as broad as long and have the dorsal projection more strongly developed and more keel- like than do those of the larger cirri. The more elongated cirrus, from another specimen, shown in Fig. i (jg, is of 51 seg- ments and nearly 100 mm. long. The first two segments are broader than long, the third HELIOMETRINAE 137 about as long as broad. The fourth is longer than broad ; the fifth is nearly, the sixth more than, twice as long as broad. The seventh and eighth are about three times as long as broad. The ninth to about the fourteenth are the longest: more than three times as long as broad. ^ Beyond the fourteenth the segments gradually decrease in length but the most distal are nearly twice as long as broad. They have a less well-marked projec- tion on the dorsal side than those of cirrus a^. The whole cirrus is less robust; it is slightly compressed from side to side. The shortest apical cirrus from the same specimen, ^4 , is of 26 segments and about 30 mm. long. The first two segments are broader than long, the third nearly twice as long as broad. The fourth is 2\ times as long as broad ; the fifth and sixth are the longest : nearly three times as long as broad. The segments beyond gradually decrease in length to the end, the most distal being, however, nearly twice as long as broad. The segments of this cirrus, particularly the outer and especially on the dorsal side, overlap one another much more strongly than do those of the long cirrus from the same specimen. The terminal claw is mostly curved, claw-like and hyaline and there is usually a small opposing spine or the trace of one ; but some of the longer cirri of younger specimens, or even all the cirri of older specimens, may entirely lack the opposing spine, and at the same time the terminal claw may be straight, not claw-like, and it may not be hyaline. The opposing spine may, however, be strong, especially on small apical cirri (Fig. i b). The radials, primibrachs and first and second brachials are very variable in shape (Fig. 2 a). They differ, though not constantly nor in the same way, in different speci- mens, with age. Since very few preserved specimens have complete arms it is not possible to use arm length as a rough criterion of age. I have used instead the number of segments in the longer peripheral cirri : in Fig. 2 a the number beside each separate drawing is that of the segments in the longest peripheral cirrus of that specimen. In very young specimens in which the interradial rays are still smaller than the others the alternate radials are of a diff'erent shape and size (flj): those of the radial ray are com- paratively wide, wider at the base than distally; the interradial radials are longer, narrower plates, wider distally than at the base. The distinction in size between the alternate rays does not appear to last long. In older, but still small immature specimens, the radials and succeeding ossicles are of the shapes shown in a.^ or of similar shapes. The radial is long; the costals are not in opposition laterally. The axillary is shield- shaped, longer than broad ; it makes no shoulder-like projection by its incision of the costal. In older specimens the radials are always shorter, and in large specimens they may be very short (Fig. 2, a^ and ^4). In some, but by no means all, old specimens in which the radials are very short the costals are in partial lateral opposition, as if they had been brought down against one another by the reduction of the radials {a^). The axillary may be of any of the shapes shown in Fig. 2 ai-a^ , or of a shape intermediate between them. It may be elongated and longer than broad, or it may be broader than long ; the second brachials vary with it. In the biggest specimens its incision of the costal and the incision of the first brachial by the second make shoulder-like projections which may ^ I have seen these segments four times as long as broad in the cirri of another specimen. 138 DISCOVERY REPORTS \ 1 r-r ^r^' =\. V. 7, -3^ -^ m^ I Fig. 2. Promachocrimis keiguelemis. a, radials, costals, axillaries and lower brachials of specimens of different ages; the numbers beside the figures are those of the segments in the longest peripheral cirrus of each specimen; i and 2, x 13; 3 and 4, xy. b, fourth to fifteenth brachials showing the dorsal surfaces raised into spinous patches, x 11. c, largest side-plates found in the distal pinnules of specimens from the South Sandwich Islands, X47. HELIOMETRINAE 139 be faint or exceedingly strong: they are faint when the axillary and second brachial are broad and short, strong when they are elongated (compare a^ and a^). The partial lateral apposition of the costals referred to above appears to come about only in those old specimens in which the axillaries are comparatively wide and form a faint, or no, shoulder with the costals. The first syzygy is between the third and fourth brachials ; the second and third are usually between the ninth and tenth and fourteenth and fifteenth respectively, though irregularities occur. The syzygial pairs beyond the third are separated by one to five, usually two or three, brachials. The brachials between the first and third or fourth syzygies are rectangular or quadrate, usually broader than long, but sometimes as long as broad or even slightly longer than broad. Beyond the third or the fourth syzygy the brachials are triangular or wedge-shaped, broader than long or as long as broad. The distal brachials are rect- angular, as long as or longer than broad. The brachials of young specimens are pro- portionately longer than those of older specimens. The lower brachials may be smooth or spiny ; the distal edges of the outer brachials are nearly always slightly raised and produced into spines. The brachials between the first and second syzygies may have more than the distal halves of their dorsal surfaces raised into rectangular patches of strong spines standing out at right angles to the surface, conspicuous in profile (Fig. 2 b) ; the patches become smaller in area and more triangular in shape between the second and third syzygies; beyond the third syzygy they become reduced to a row of strong forwardly directed spines along the distal edge of the segment, and they persist as such to the end of the arm. The spine patches of the lower brachials may be smaller and of a diff'erent shape, and they may not start till the sixth or seventh brachial or beyond, and the spines of the outer brachials may be correspondingly smaller ; or the lower brachials may be quite smooth and those beyond may have slightly produced edges which are smooth or finely or strongly thorny. Pi is long and whip-like, of 40-60 or, exceptionally, of up to 75 segments; it is usually 15-22, sometimes as much as 26, mm. long. The first 6-12 segments are usually broader than long, or as long as broad, a little stouter than the following and roughly diamond-shaped. Their short and narrow dorsal surfaces may be raised into thorny crests. The following segments are of a regular shape and longer but not elongated: they are not so much as twice as long as broad; the more proximal may be raised dorsally into a fine thorny crest. P2 is usually of fewer segments : those examined were of 26-46 except for one which was of 62. The number is usually between five and ten less than that of Pj of the same specimen. Pg is of about the same length as Pj: it may be shorter or longer. The first 4-5 segments are usually more massive, compared with the others, than the basal seg- ments of Pi ; the others are slightly more elongated, though rarely so much as twice as long as broad. The first genital pinnule is usually P3 or P4. P3 is commonly of 16-40, sometimes of up to 60, segments. Whether it is an oral or a genital pinnule it may be of roughly 3-2 I40 DISCOVERY REPORTS the same length as Pj, but sometimes, when it is a genital, it is very much shorter and of segments which diminish rapidly in stoutness from the base to the end. P3 is usually of anything from 5 to 20 fewer segments than P2 of the same specimen ; all but its basal segments are more elongated, being as much as, or more than, twice as long as broad. The middle genital pinnules are of about 20-26, exceptionally more, segments ; they are 10-20 mm. long, sometimes considerably longer. The outer pinnules are of about the same length or longer. In both the genital and outer pinnules the first segment is very short, the second longer but not so long as broad, the others considerably longer than broad. The way in which the primary ambulacral furrows on the disk divide so as to provide the arms is very variable though certain arrangements appear to be more constant than others (see below). Sacculi are abundant on the pinnules. Along the sides of the pinnule ambulacra there may be a single continuous series of large plates, three or four to each pinnule, the distal edge of one overlapping the proximal of the next. Each is curved in the axis of the pinnule and has its outer part curved over the ambulacral furrow. Clark (1921, p. 268, fig. 378) has described them at their highest development. When they are present the tentacles contain numerous knobbed spicules. The plates may be smaller, and therefore not touching one another, or very much smaller and fewer: in some specimens the pinnules have only one or two extremely simple small plates near the end (Fig. 2 c). They are often entirely absent. In some specimens with reduced plates the spicules are few ; if the plates are very reduced they may be absent : this is so in over twenty-five specimens of the present collection. On the other hand, it is only in one specimen that there are spicules (and they are few) and no plates. (It was, of course, only three or four pinnules of any specimen that were examined.) The presence of plates and spicules has been regarded as a sign of im- maturity. The examination of the present collection has shown that it is not so, but that there does appear to be some correlation between the degree of development of the plates and spicules in a number of specimens and the locality in which they occur (see below). The colour is very variable. The entire specimen may be straw-coloured, very light grey, flesh-coloured, pale or brilliant yellow, orange, brown or purple; the greater part of it may lack stronger colours except for bands of dark brown or purple on the cirri, or the arms (when the bands often coincide with the syzygial pairs), or the pinnules, or all three ; or portions of a specimen, such as the pinnules or the distal parts of the arms, may be brown or purple or yellow. There may be great variation in the colour of the specimens of one colony as is shown by a note, made at the time of capture, describing the eight specimens from St. 42 : " The animals showed a gradation of depth of colouring ranging from cream to very light grey or flesh, to pale or deep chocolate brown, to dark purple. The colour was usually least developed on the cirri, and— grading through the arms — most enforced on the pinnules; it was sometimes equally developed on all." The gonads were bright orange. HELIOMETRINAE 141 The specimens from Sts. 1 948-1 957, from the Bransfield Strait region, were fixed and preserved by Mr J. W. S. Marr who made many notes concerning them. He writes, from his observations on captured specimens, that the species is "very powerful and active and swims with a remarkable grace of movement". Because of its activity it breaks itself into pieces if killed in a confined space. The best results were obtained by fixing in fresh water in ample space. The species has previously been taken from depths between 10 and 650 m.; the majority of the present collection come from between 22 and 525 m. But eight speci- mens were taken from 1080 m. in the Bransfield Strait (St. 177); they are all small but they possess the usual twenty arms and do not appear abnormal. A pentacrinoid larva was also taken at this station. Number of arms. The number of arms possessed by the specimens of this collection are shown, under the localities from which they come, in the following table : No. examined Arms 22 20 19 18 17 16 12 South Georgia South Sandwich Island Bransfield Strait Graham Land region Ross Sea 60 14 62 15 23 I 59 14 41 9 22 I I 16 2 I I 3 I 2 The single abnormal specimen from South Georgia, with 22 arms, has eleven radials. The twenty-one specimens with 16, 17, 18 and 19 arms from the Bransfield Strait all come from one of the eight stations made in that region (St. 1952). Twenty-two speci- mens with 20 arms were taken with them. They are much younger and smaller than those with abnormal numbers of arms: their longest cirri consist of 25-38, mostly 30, segments; those of most of the specimens with 16-19 arms are of 50-65 segments. Five of the six specimens with abnormal numbers of arms from the Graham Land region are of medium size, their longest cirri of 39-45 segments ; one is small, its longest cirrus of 28 segments. In the 19-armed specimen from the Graham Land region (it is from St. 180) the single arm arises from a normal radial which is followed by a regular ossicle slightly longer than the costals of the other rays. Next is a still longer ossicle with a pinnule arising from either side of it. It is succeeded by a syzygial pair with a pinnule arising from the epizygal and beyond this the arm is normal. The single arms of the 19-armed and 17-armed specimens from the Bransfield Strait are different. The second ossicle beyond the radial is shaped like an irregular axillary, and a pinnule arises from one side of it. It is followed by a very short ossicle with no pinnules, which appears to correspond with the first brachial of normal arms; after it comes a longer ossicle shaped like the second brachial of normal arms, with a pinnule on the opposite side to the first. A syzygial pair 142 DISCOVERY REPORTS with a pinnule arising from the epizygal, on the same side of the arm as the first pinnule, comes next and thereafter the arm is normal. Dorsal pole. In four of the specimens from the South Sandwich Islands I could see no dorsal pole ; it is very small and difficult to detect in the other specimens, large and small, from the same locality. I failed to see a dorsal pole in one small specimen from South Georgia and in one from the Bransfield Strait. It was present in all the other specimens. Cirri. Fig. i a shows extreme types of cirri : in a^ and a-^ the segments are as short as in any full-grown specimen I saw ; a^ and o^ are of cirri of very elongated segments, though I have seen others slightly longer. The specimens from the South Sandwich Islands and South Georgia have cirri of the first type though the segments are usually a little longer than in the figure ; those of the South Sandwich specimens appear to be the shorter. Two specimens from South Georgia, by no means the largest, have cirri of the second, long, type. The specimens from the Bransfield Strait and the west coast of Graham Land have cirri of the long type, though not usually so long as those figured ; those from the latter locality appear to have the longer. All but two of twenty- three Ross Sea specimens have cirri of the long type. In Fig. i a, i and 2 are from a specimen from the South Sandwich Islands, 3 and 4 from one from the west coast of Graham Land. Division series and lower brachials. None of the bigger specimens from South Georgia have elongated axillaries and second brachials forming strong shoulders with the costals and first brachials; in the biggest the shoulders are but slight. On the other hand, most of the bigger specimens from the South Sandwich Islands and the Bransfield Strait and all from the west coast of Graham Land have axillaries and second brachials which are elongated and form strong shoulders with the costals and first brachials respectively. Among the Ross Sea specimens there are none so large and old as in the collections from the Falkland sector ; nevertheless, most of them have moderately long axillaries and second brachials, but in only five do they form strong shoulders with the costals and first brachials respectively. Ambulacral furrows on the disk. Mortensen (1918, p. 19) describes how variable is the arrangement of the ambulacral furrows on the disk, "hardly two specimens (of his collection) being alike"; Bernasconi {1932b, pp. 33-5, fig. 3a-e) gives a figure showing the different arrangements in five specimens, though the differences are slight between each of two pairs. Most of the specimens of the present large collection are well pre- served with their arms bunched together so that the disk is hidden. But in ten from one locality (St. MS 71) which are preserved otherwise the disk can be seen, and in each one of them the arrangement of the ambulacral furrows is different from that in any other ; in one it resembles that of two (a and b) in Bernasconi's figure. Mortensen writes : "The normal condition evidently is that each primary ambulacral furrow divides so as to provide four arms ; but often one or two of them (mostly the left posterior and the right anterior) divide so as to proceed to six arms." In seven of my ten specimens the left posterior divides so as to provide six arms, but the right anteriors are very variable. HELIOMETRINAE 143 Bernasconi's five drawings show the anterior primary furrow supplying only one pair of arms: it does so in seven of my specimens. It appears that some of the primary ambulacral furrows divide in a less variable way than others. Side plates and spicules. Clark(i9i5 «, p. 132) thinks thatthe presence of well-developed plates is a sign of immaturity, and that poorly developed plates, or the absence of plates, is characteristic of mature individuals. I do not find it so in this collection. But there is a relation between locality and the presence or absence of well-developed plates : it is shown by the following table: Locality No. of specimens examined No. with both plates and spicules No. with plates only No. with spicules only West coast of Graham Land (Sts. 180, 182, 599) Bransfield Strait (Sts. 170, 175, 177, 1948, 1952, 1955, 1957) South Sandwich Islands (Sts. 363, 366, 371) South Georgia (Sts. 39, 42, 123, 144, 148, 149, 152, 156, MS 71) II 37 13 34 7 29 2 0 0 2 4 S I 0 0 0 It is not only a question of numbers : the degree of plating varies as well. Of the seven specimens with plates from the west coast of Graham Land five are heavily plated ; of the twenty-nine from the Bransfield Strait at least thirteen are heavily plated. The six specimens having plates from the South Sandwich Islands all have very small ones (the size of the largest is shown in Fig. 2b): m three of them the greatest number of plates in one pinnule is two, and they are minute and near the tip. In the five specimens from South Georgia possessing plates they are as small as, or smaller than, those of the South Sandwich specimens. In three of them there are two or three very reduced plates near the tips of some pinnules but none in others. So it is probable that if a larger number of pinnules of every South Georgia specimen were examined the proportion of indi- viduals with plates would be found to be higher: but they would be very small plates, few in number and unevenly distributed. It is, at least, quite certain that none of the South Georgia specimens has even moderately developed plates ; nor were spicules seen in the tentacles of any. Although the South Sandwich specimens have plates nearly as small as those from South Georgia, two of them do have spicules in the tentacles as well and so, in this way, stand intermediately between the South Georgia population and that of the Bransfield Strait and the west coast of Graham Land where most of the specimens with plates have spicules too. From these facts it would appear that in this sector of the Antarctic the Promacho- criiius kerguelensis living in lower latitudes are most often without plates and always devoid of spicules, but that a small proportion have very reduced plates, few in number, on at least some of their pinnules ; that the majority of those living in high latitudes have 144 DISCOVERY REPORTS big plates along the entire lengths of the pinnules and many spicules in the tentacles ; and that those living intermediately are intermediate in character. But Mortensen (1918, p. 19) found no plates, not only in four specimens from South Georgia, but in sixteen from the east side of Graham Land. On the other hand, Grieg (1929a) described the pinnules of seven specimens from the Bransfield Strait as having well-developed cover-plates. The pinnules of thirty-eight Ross Sea specimens, twenty from the present collection and eighteen from the National Antarctic and British Antarctic Expeditions, were examined. Thirteen have plates along the ambulacra and spicules in the tentacles ; in eight the plates are strongly developed, in five they are small and scattered. Another thirteen have small and scattered plates but no spicules. One specimen has spicules in the tentacles and no plates along the pinnule ambulacra ; eleven have neither plates nor spicules. Two pinnules of a small Challenger specimen from Kerguelen were ex- amined. One has a single diminutive plate near the tip, the other none. That the presence of highly developed plates and many spicules is not a sign of immaturity is shown by the following table (p. 145), in which forty-six specimens from the Bransfield Strait and the West Graham Land region are arranged in order of age, so far as that is shown by the number of segments comprising their longest cirri. The table shows that it is not generally the immature but the larger that have the most highly developed plates. Specimens taken from one place may vary in the degree of their plating: for example, the smaller of the three specimens from St. 175 has highly developed plates and many spicules while the other two have none. Although Minckert (1905, p. 499, fig. 2) described P. vanhoffenianus as being dis- tinguished from P. kerguelensis by, among other things, the possession of a skeleton along the pinnule ambulacra, his drawing, to illustrate the plating, shows no plates but a reticulation in the ambulacral lappets such as I have seen in a large number of specimens. It is formed of pigment. Hartlaub (1912, p. 485) describes the same appearance. I have not noted it as present in specimens from South Georgia, bvit as being of frequent occurrence in those from the South Sandwich Islands, the Bransfield Strait, the region west of Graham Land, and the Ross Sea. It may occur together with, or in the absence of, plates ; if the former, in such a way as to suggest to me that it follows a reduction in the size of the plates. Florometra mawsoni A. H. Clark (Plate III, fig. 2) Solanometra antarctica (part) Clark, 1913, p. 61. Promachocrimis {Promachocrimis) kerguelensis (part) Clark, 19150;, p. 130 (bottom of page), pi. iv, figs. I a, b. Florometra mawsoni Clark, 1937, pp. 10-14. St. 180. II. iii. 27. Schollaert Channel, Palmer Archipelago. 160-330 m. Gear OTL. Bottom: mud and stones. One specimen. St. 181. 12. iii. 27. Schollaert Channel, Palmer Archipelago. 64° 20' S, 63° 01' W. 160-335 m. Gear OTL. Bottom: mud. Two specimens. HELIOMETRINAE 145 Longest cirrus Well- Specimen developed plates, Small plates. Small plates, no Spicules, no No plates, no No. of Length, many spicules spicules plates spicules segments mm. spicules 177 S2 22 — — + — — — 177 S4 c. 22 <20 - + - — - 175 Si 24 20-25 + - - - - 1957 24 — — — — — + 1952-8 25 — + — — — — 177 S5 c. 25 25 - + - — - 1955 26 — + — — — — 1957 27 — - - - — + 182 Mi 28 >4o - - - — + 180 Mi 28 45 - 4- - - - 1952-7*" 28 — - + - - - 1952-7A 28 — — + - — — 180 M 2 29 — — — - - + 1948 30 — - - - — + 1952-2 3° — — + — — — 1952-66 30 — — + — — — 1952-6C 30 — - + - - — 1952-6 3i >55 — — — — + 1955 32 — + — — — — 170 S3 33 38 + — — — — 170 S4 33 35 + — — — — 175 S2 c- 33 '^■45 - — — — + 170 Si 34 3° + — — — — 1952-4 34 — + - — — — 180 M4 36 55 - - - + - 170 LM 38 60 - - + — - 1952-6^ 38 — + — - — — 182 M2 39 c. 70 + - - — — 182 M3 39 c. 70 + - - — — 180 M3 42 55 - + — - — 599 Mi 42 c. 50 + - - — — 182 Ms 46 c. 80 + — - - — 182 Li 54 no + — - — - 1952-1 58 — + - - - — 1952-3 « 60 — + - - — — 1952-36 61 — + - - — — 1952-5 61 c. 130 + - - - — 1952-9 63 >ii5 + - — — — 1952-76 64 > 1 10 + — — ~ 'Note. Each specimen appears in the table under the number assigned to it in the study of this collection. In each the first number (e.g. 177 in 177 S2 or 1952 in 1952-7^) is the number of the station from which the specimen came. 146 DISCOVERY REPORTS St. 190. 24. iii. 27. Bismarck Strait, Palmer Archipelago. 64° 56' S, 65° 35' W. 130-100 m. Gear DLH, NRL. Bottom: mud, stones and rock. One specimen. St. 599. 17. i. 31. Adelaide Island. '67° 08' S, 69° o6|' W. 203 m. Gear DLH. Bottom: no data. Four specimens. St. 1652. 23. i. 36. Ross Sea. 75° 56-2' S, 178 ' 35-5' W. 567 m. Gear DRR. One very young specimen. St. 1660. 27. i. 36. Ross Sea. 74° 46-4' S, 178" 23-4' E. 351 m. Gear OTL. Bottom: mud. One very young specimen. I had written a full description of this species before knowing that Mr Clark had found it in the Australasian Antarctic collection and described it. I think it useful to add the following remarks and to publish the figures I had prepared. The arms of the eight specimens vary in length from 50 to no mm. I find the cirri to be XL-LXV, 16-31, but mostly 24-26. The most elongated seg- ments, the fourth to seventh, are not quite so elongated as in Clark's description; the most distal are slightly longer : they are longer than their distal widths. Clark describes how the distal edges of the radials are considerably produced out- wardly at the sides. In all but one, a small specimen, of the present collection the distal edge of the radial is produced into a thin lip-like frill which may stand out at right angles or be curved farther backwards. It has the appearance of being formed of a row of spines all but the points of which are connected by a web, and it is stronger on the sides than in the mid-line. The distal edge of the costal is produced into a similar thorny frill. Fig. 3 b-e shows the shapes of the brachials. The spinous processes on the distal edges of the lower brachials give the animal a very characteristic appearance. They may be small on, or absent from, the first and second brachials. They are strongest between the first and second syzygial pairs where they stand out at right angles to the arm. On the brachials beyond the third syzygy they are forwardly directed but still coarse and strong. (They are stronger than appears in Fig. 3 d which, being a dorsal view, does not show how strongly the edge of one segment overlaps the beginning of the next.) They persist to the end of the arm. On one arm of one specimen there is only a single ossicle, larger than the normal first or second brachial, between the axillary and the first syzygial pair. In the smallest adult specimen Pi is of 25 segments, in the others of 28-39 segments and up to 12 mm. long. P2 has from 24 to 38 segments. P3 may be a whip-like oral pinnule, with heavier basal segments than the others, but in most of the specimens it is a genital pinnule. The earlier genital pinnules may be considerably shorter than the orals. The number of segments in the genital pinnules varies from 14 to 23, the proximal usually having a smaller number than the distal. All but the first two segments are longer than broad. The distal edge of each is produced into a row of spines. The distal pinnules are like the genitals but slightly longer, of 20-30 segments. The disk is naked. Sacculi are abundant. In some of the specimens there is a single row of small fenestrated plates along each HELIOMETRINAE H7 side of the ambulacral furrow in a few segments near the tips of the pinnules. They vary in number from about two to ten and when the bigger number is present the more f Fig. 3. Florometra mawsoni.- a, cirrus, xy. b, centrodorsal and proximal part of a ray, x 7. c, first to eleventh brachials in profile and P3, xy. r/, brachials beyond third syzygy, xii. e, distal brachials, xii. /, Pj, XII. g,^i2, XII. A, side-plates of a distal pinnule, X53. proximal are very small, the distal larger (Fig. 3 h). In some specimens there are spicules in the tentacles. Both plates and spicules may be absent or one or both may be present. There is no information as to the colour of the specimens in life. In spirit they are 4-2 148 DISCOVERY REPORTS pale straw with a dark bluish tinge in places. In one specimen the bluish tinge is con- fined to the outer pinnules. In the others it occurs on some of the brachials and the basal segments of some of the pinnules : it is always strongest on the brachials of the syzygial pairs and the bases of the pinnules arising from them and is sometimes present only on these ossicles. In one specimen the syzygial pairs stand out as dark bands along the whole length of the arm, darkest along the middle arm. There is no blue coloration on the calyx, the centrodorsal or the cirri. In one specimen the centrodorsal and radials are yellow. One of the specimens was infested with two cysts of Myzostomum cysticoliim. Distribution. There are in the British Museum collection eight specimens of this species from the Ross Sea which had been labelled Anthometra adriani or Solanometra antarctica. One is from Discovery Winter Quarters and another two are Discovery specimens labelled "Mt Erebus and Terror". The remaining five are Terra Nova specimens (see p. 220). The species is then known from the coast of the continent in the Ross Sea, Indian Ocean and Weddell Sea sectors of the Antarctic ; it is not known from any of the out- lying islands. Florometra antarctica, n.sp. (Plate III, fig. 3) St. 180. ii.iii. 27. SchoUaert Channel, Palmer Archipelago. 160-33001. Gear OTL. Bottom: mud and stones. Two specimens. St. 599. 17.1.31. Adelaide Island. 67° 08' S, 69° o6i' W. 203 m. Gear DLH. Bottom: no data. One specimen. Description. One of the three specimens is nearly complete and has arms about 95 mm. long. In another, less complete but more massive, the arms must have been over 100 mm. long. The third specimen is smaller, with arms so broken that it is im- possible to estimate their length when complete. The centrodorsal is a large rounded cone closely covered with cirrus sockets arranged in regular or fairly regular columns (Fig. 4 o). The dorsal pole is rounded and rough in two of the specimens, sunken in the third. Cirri LXIX-LXXVII, 23-32, usually 27-32. The cirri are long and strong, much longer in proportion to the length of the animal than in Florometra mawsoni, particularly so in the largest specimen. The first two segments are short, the third is about as long as broad. The fourth is twice as long as broad and fairly strongly waisted. The fifth to about the ninth are about two-and-a-half times as long as broad and faintly waisted. Beyond the ninth the segments gradually decrease in length though all are longer than broad. The distal segments possess a strong dorsal spine, the apex of which is sub- terminal. The opposing spine and the terminal claw are strong (Fig. 4 b). In the specimen with arms 95 mm. long the basal rays are visible externally as small triangular plates (Fig 4 a). In the other two specimens, the largest and the smallest, I can see only four basal rays. The radials are short, especially in the mid-line, because their distal edges are strongly HELIOMETRINAE 149 concave. They are in apposition for the greater part of their length but the distal corners are free. The distal edge may be smooth or raised into a low finely-thorny ridge. The costals are longer than the radials, not in apposition laterally. The distal edge of each is raised on either side of the incision by the axillary into a thorny lip at right angles to the surface of the ossicle. Fig. 4. Florometra antarctica. a, centrodorsal and parts of three rays, X7. b, cirrus, xy. c, Pj, xii. d, spicules from tentacles of distal pinnules, x 1 14. The axillaries and the first and second brachials are of the shapes shown in Fig. 4 a. The axillary is wider than the costal and forms a shoulder where it incises it ; its distal edge is smooth but there are small spines on the free outer portions of its proximal edges. The inner edges of the first brachials are much shorter than the outer; they do not meet, igo DISCOVERY REPORTS or at most their corners meet, above the axillary ; the distal edges are smooth or finely thorny. The second brachial forms a slight shoulder where it incises the first ; there may be fine spines along part of its proximal edge but its distal edge is smooth. The first syzygy normally occurs between the third and fourth brachials but in one specimen it is between the fourth and fifth brachials on one arm, between the tenth and eleventh on another. The second syzygy is usually between the ninth and tenth brachials but it also occurs between the eighth and ninth and the tenth and eleventh. Syzygies are numerous beyond the second with one to four brachials between the pairs. The brachials between the first and third syzygial pairs are nearly rectangular, wider than long ; those beyond are at first wedge-shaped and then triangular, a little broader than long ; farther out on the arm they become rectangular, at first broader than long, then as long as broad. In the two large specimens the distal edges of the brachials between the first and second syzygies are smooth. The first one to three of these brachials are smooth in the small speci- men but the others are raised distally, in the mid-line only, into a group of spines much smaller and lower than those of F. mawsoni but, like them, standing out at right angles to the arm. The distal edges of the brachials beyond the third syzygy are the same in all three specimens: they are produced into a row of strong forwardly directed spines. The oral pinnules are longer and composed of a greater number of segments than in F. mawsoni and the first genital pinnule is farther out on the arm. Pi is of 44-50 seg- ments, 16 mm. long (Fig. 4 c) ; P2 of 38-45 segments, 15-16 mm. ; Pg of 35-43 segments, about 13 mm. None of the segments of the oral pinnules is much longer than broad; the proximal have spiny dorsal carinations which become reduced to low rounded pro- tuberances on the segments at the tip of the pinnule. The first genital pinnule is P4 or, more often, P5. P5 is of about 23 segments and 11 mm. long; the gonad lies along the fourth to the tenth or eleventh segments. The other genital pinnules are of between 23 and 30 segments and up to 16 mm. long; the gonads usually lie along the third to eighth segments. All but the first two segments of the genital pinnules are considerably longer than broad. Their distal edges are beset with fine spines. The disk is naked. Sacculi are abundant. None of the specimens shows any trace of an ambulacral skeleton. In the largest specimen there are many spicules, some smooth and rod-like and others branched, in the tentacles of the distal segments of the outer pinnules (Fig. 4 d). The specimens are of a pale straw colour in spirit, darker on the pinnules than else- where. There is a yellowish tinge on the distal parts of the cirri and pale brown spots on the gonads and the tissues on the inner sides of the arms. Some of the brachials have a bluish tinge along the sides. One of the specimens has a large cyst of Myzostomiim cysticoliim v. Graff lying be- tween the bases of a pair of its arms with its lower extremity resting on the disk. Florometra antarctica is distinguished from F. mawsoni by its considerably longer oral and genital pinnules which are composed of a greater number of segments. The HELIOMETRINAE iSi distal edges of the lower brachials are not everted into strong spiny ridges at right angles to the dorsal surface as in F. mawsoni. The cirri, although they do not comprise more or many more segments than those of F. mawsoni, are much heavier and longer. Genus Anthometra Clark Anthometra adriani (Bell) (Plate III, fig. 4) Antedon adriani Bell, iqocS, p. 4, pi. ii; 1917, p. 2. Promaclwcrimis {Atithoiiietra) adriani Clark, 1913, p. 60; 1915 a, pp. 135-7, pls- vi-vii. Anthometra adriani Mortensen, 191 8, p. 18; Clark, 192 1, many references including description of side plates (p. 270) and of pentacrinoid young (pp. 557-9, fig. 938); Mortensen, 1925 b, p. 2; Gislen, 1928, p. 11; Clark, 1929, p. 662; Grieg, 19290, p. 4; John, 1937, p. 10; Clark, 1937, pp. 14-16. St. 190. 24. iii. 27. Bismarck Strait, Palmer Archipelago. 64° 56' S, 65° 35' W. 130-100 m. Gear DLH, NRL. Bottom: stones, mud and rock. Two specimens. St. 1660. 27.1.36. Ross Sea. 74° 46-4' S, 178" 23-4' E. 351m. Gear OTL. Bottom: mud. One broken specimen. St. 1952. II. i. 37. Between Penguin Island and Lion's Rump, South Shetlands. 367-383 m. Gear DRR. Bottom: soft mud. Nine specimens. Fig. 5. Anthometra adriani. a, a male genital pinnule from the side, b, a female genital pinnule from the side, c, the same obliquely from above. All x 7. It appears from my own counts and those of others that the cirri of large specimens may be numerically described as follows : L-LXX, 60-86, up to 90 mm. long. Gislen says that some of the segments of the proximal pinnules are expanded and that it thus "forms a transition to the Isometrinae". It is the segments of the genital pinnules along which the gonad lies that are expanded, and they are a little more ex- panded in the female than the male. Figures are given (Fig. 5): the female pinnule is from a well preserved Terra Nova specimen with nearly complete arms which are 152 DISCOVERY REPORTS 250 mm. long, a much greater length than any previously recorded; the male pinnule is from a smaller specimen. One of the specimens from St. 190 is infested with a Myzostomum on the disk. Genus Solanometra A. H. Clark Solanometra antarctica (P. H. Carpenter) Antedon antarctica P. H. Carpenter, 1888, p. 144, pi. i, figs. 6 a-d, 7 a, b, pi. xxv. Antedon australis P. H. Carpenter, 1888, p. 146, pi. xxvi, figs. 4, 5, pi. xxvii, figs. 14-20. Promachocrinus {Solanometra) antarctica Clark, 1915 a, p. 135— list of earlier references and synonymy. Solanometra antarctica Clark, 1937, p. 9. There are no specimens of this species in the present collection. Bell records it (as Antedon antarctica) from the Ross Sea both in his Discovery (1908) and Terra Nova (19 17) reports. I have re-examined the specimens and find that not one of them is Solanometra atitarctica. The species is known only from the Challenger specimens taken near Heard Island (of which there are six in the British Museum collection, three each from Sts. 150 and 151), and from the Australian Antarctic Expedition's collection from near the coast of Adelie Land on the Antarctic continent. Subfamily ZENOMETRINAE Genus Eumorphometra A. H. Clark Eumorphometra aurora n.sp. (Plate IV, fig. i) St. 160. 7. ii. 27. Near Shag Rocks. 53° 43' 40" S, 40° 57' W. 177 m. Gear DLH. Bottom: grey mud, stones and rock. One specimen. Description. All the arms are broken off at what I judge to be about three-quarters of their length; the parts remaining are composed of about 30 brachials and are nearly 20 mm. long. The centrodorsal is a rounded cone not so high as it is broad at the base. The ventral edge is produced into very low and wide corners interradially. The cirrus sockets are arranged in ten columns, one close against another, the arrangement of which is regular but for the most ventral circle which is composed of fourteen. Cirri ca. XLI, 17-28. They increase in size from the apex of the centrodorsal to the edge. Those around the dorsal pole are of about 17-19 segments and short; the next circle are longer, of 21-23 segments; the outermost are of 27 or 28 segments and up to 10 mm. or more in length, about twice as long as the apical (Fig. 6 a). The first three segments of the cirri are broader than long, the fourth is as long as broad, and the fifth to the eighth or ninth are sHghtly longer than broad. All these have the distal end slightly flared, more strongly on the dorsal than the ventral side, so that they are a little ZENOMETRINAE 153 wider distally than at the base. The more distal segments are about as long as broad. Their dorsal sides are raised into low keel-like protuberances highest near the end of the segments, making them wider distally than proximally. The terminal claw is short and the opposing spine small. The radials are nearly rectangular and fairly long, about half as long as broad (Fig. 6 b). The distal half is half-cup shaped. In profile the proximal half makes only Fig. 6. Eumoiphometra aurora, a, cirrus, b, proximal part of a ray. c, fourth to tenth brachials and Pa. d, Pg from the side. All x 13. a slight angle with the dorsoventral axis but the distal half bends sharply outwards. The costals are not quite so long as the radials; they are widely separated from each other and are not deeply incised by the axillaries. The axillaries are slightly longer than broad. The first and second brachials are of the shape shown in Fig. 6b\ the former are widely separated. The distal edges of the radials, costals, axillaries and the first and second brachials are raised into very fine spines. Syzygies occur between the third and fourth, ninth and tenth, and fourteenth and fifteenth brachials, and thereafter with intervals of two brachials between eachsyzygial pair. The fifth to eighth brachials, between the first and second syzygies, are nearly rectangular, about one-and-a-half times as broad as long (Fig. 6 c). Those between the second and third syzygies are wedge-shaped and longer than broad. The more distal brachials are longer on one side than the other, slightly longer than broad. The distal edges of all the brachials beyond the first syzygy are faintly raised and produced into short spines; they are most conspicuous on the distal brachials, but are much shorter and less conspicuous than those of E. hirsuta. 154 DISCOVERY REPORTS There is no strong contrast between the different kinds of pinnules. In all, the distal edges of the segments are faintly flared out and produced into spines ; the flare is not strong enough to give the joints a swollen appearance. Pj is slender and tapers evenly from the base to the tip. It is 4-5 mm. long, of 13 rounded segments all but the first two or three of which are longer than broad, the distal more so than the proximal ; the distal are up to three times as long as broad. Po is similar to Pj but of fewer segments and shorter, of 8-g segments, about 3 mm. long. It tapers more rapidly and its distal segments are more elongated. In the genital pinnules the first two segments are short, the others longer than broad, the middle segments being more elongated than in the oral pinnules. P3 is the first genital pinnule; it is of 8-9 segments like P2 but is slightly longer, 3-4 mm. long. It bears a large egg-shaped gonad on the third to fifth segments : a testis, for the specimen is a male. The genital pinnules extend to P7 or Pg (Fig. 6 d). They are all of about the same number of segments as Pg , but the distal are a little longer, up to 5 mm. long; their gonads are more fusiform being along the third to sixth segments. The pinnules immediately beyond the genitals are about 5 mm. long and of about 12 seg- ments, the first two short, the others considerably longer than broad. There are no complete outer pinnules left. The disk is naked. Sacculi are few and inconspicuous. Along the pinnule ambulacra there are small rods, three to each segment, which are perforated, forked or branched at the end. They are very reduced side-plates and are better developed along the middle than the proximal segments. At the ends of some, but not all, there are equally reduced cover-plates. Both are similar to, but more reduced than, those of Phrixometra mitrix (Fig. i\e). The specimen is white in spirit. Mr A. H. Clark has kindly made for me a direct comparison of this species and Emnorphometra concinna. E. aurora is considerably larger, its centrodorsal is lower and broader with more numerous cirrus sockets more irregularly arranged ; its longer cirri have more segments ; its radials and axillaries are longer ; the elements of its division series and its brachials do not have their central portions abruptly elevated and prominently spinous as in E. concinna. I have been able to compare E. aurora directly with E. hirsuta. E. aurora is far less spinous ; the shapes of the elements of the division series and of the first two brachials are quite different in the two species. The cirrus sockets of E. hirsuta are in about fifteen irregular columns. E. aurora is easily distinguished from E. fraseri by its smaller number of cirrus segments; and from E. marri by the facts that its cirrus sockets are in columns not alternating rows, that P^ is not much more massive than P, , and that the segments of the lower genital pinnules carrying the gonads are not expanded. ZENOMETRINAE iSS Eumorphometra fraseri, n.sp. (Plate IV, fig. 2) St. 1955. 29. i. 37. North of South Shetland Islands. 61° 35-1' S, 57° 23-3' W. 440-410 m. Gear DRR. One specimen. Description. No arm is complete ; the longest is of 44 brachials, 28 mm. long. The centrodorsal is a high and pointed cone with a sharp and rough dorsal pole. The ventral edge of the centrodorsal is produced into low corners interradially. It is thickly covered with cirrus sockets. Fig. 7. Eumorphometra fraseri. %§. 922-3). The crown is 1-3 mm. long. The stem is incomplete, of 23 columnals, the longest of which PENTACRINOID LARVAE 201 are four to five times as long as broad. The radials are just in contact. The only complete arm is of two brachials. I see no side-plates or sacculi. The pentacrinoid from the Bransfield Strait region (St. 177, 1080 m.) resembles, so far as the shape and proportions of the radianal, radial and basal plates are concerned the specimen shown in Clark's fig. 90 (p. 545). But it is older and larger: the crown is 1-8 mm. long; the stem, of 26-30 columnals and a terminal plate, is nearly 7 mm. long. The arms are of three to six brachials. The South Georgia pentacrinoid closely resembles the oldest of Clark's series, his pentacrinoid No. 43 (pp. 551-3). The crown is 9 mm. long but the arms are not com- plete. Only 12 columnals remain. The centrodorsal is as high as it is broad and truncated distally. It has four whorls of cirrus sockets. Those of the most apical whorl are radial in position, the next are interradial, and so on. The cirri of the lower interradial row are the longest, being of 15 segments and reaching to about the fifth brachial. Those of the more apical radial whorl are shorter and of fewer, 11-12, segments but they are more mature in appearance. Both rows of cirri are in general similar to those described by Clark. The next, inter- radial, row are of 12-15 segments and reach to the first or second brachial but they are of immature form. The peripheral, radially situated, cirri are mere rudiments. The columnal next to the centrodorsal is short, about three times as broad distally as it is long. Its width is equal to that of the base of the centrodorsal, greater than that of its truncated tip. It is wider distally than proximally for the distal half is raised into five forwardly projecting rounded lobes which are in contact only on the distal edge of the columnal. They perhaps represent the beginnings of the five plates that Clark describes as arising from the corresponding columnal of his pentacrinoid No. 43 ; but they are in the distal half of the columnal whereas Clark's arose from the proximal edge. Two narrower discoidal columnals follow and they are succeeded by a still narrower columnal which is half as long as broad. The remaining columnals are elongated. Short basals are visible. The radial radials approach the adult form. The interradial radials are much narrower but they reach nearly as far forward. Each carries a small costal and axillary, small oblong plates, the latter reaching nearly as far forward as the radial axillaries ; there are no brachials on the interradial rays. The radial arms are long and well developed but none is complete; the longest re- maining is of about 20 brachials. They resemble those of Clark's specimen but I cannot see side-plates; it is certain that there are not large conspicuous plates such as Clark describes. This is interesting: the specimen comes from South Georgia where few adults have side-plates and those that do possess them have but few and small and scattered plates; Clark's specimen came from near Gaussberg, and it has been shown (pp. 143-4) that adults from such high latitudes have on the whole more and better developed side- plates. From the eleventh brachial onwards there are long but incomplete pinnules. Pj, P^ and P2 are also present but very small, of 1-3 segments. 202 DISCOVERY REPORTS Large triangular oral plates are present near the mouth. The posterior interradius is damaged and too broken to show if the radianal plate remains. Pentacrinoid larvae of Isometra hordea n.sp. There are fifteen pentacrinoid larvae from St. 170 (23. ii. 27. Off Cape Bowles, Clarence Island, 61° 25' 30" S, 53° 46' W, 342 m.) which form a series. The youngest is a prebrachial or cystid stage; the oldest has three whorls of cirri on the topmost columnal of its stem, and arms of about thirty brachials. From the same station there are eighteen pentacrinoids of Notocrinus virilis which are described hereafter (p. 210). The adult crinoids taken at this station were: Promachocrimis kerguelensis 6 Isometra hordea 3 Notocrinus virilis 17 A'^. mortenseni 2 It does not follow that the pentacrinoids of the present series belong to any one of these species but it is probable that they do. I believe them to be Isometra hordea. They are not of Promachocrimis kerguelensis, the stages of which are known (see above). They do not closely resemble the pentacrinoids oi Notocrinus virilis as one would expect those of A^. mortenseni to do. On the other hand, they do appear to show some resemblances to Isometra hordea. The shapes and proportions of the primibrachs and brachials of the oldest larvae and of /. hordea are similar. The distal brachials of the oldest pentacrinoid have a slight zigzag character suggestive of that of the middle part of the arms of /. hordea. The pinnule ambulacra are lined by side-plates and the tentacles have numerous spicules, the former not unlike, the latter resembling, those of /. hordea. 1. This is the only specimen of a cystid or prebrachial stage (Fig. 21 a). The crown is 0-9 mm. long. The column is of 26 segments and 3-4 mm. long. The distal part of the column is narrower than the proximal. The first four or five columnals are very short and discoidal, much shorter than wide, but increasing in length. The middle columnals are nearly as long as wide, the distal are slightly longer. All are con- siderably wider medianly, where they are encircled by a narrow girdle, than at either end ; they are evenly rounded off towards each end. The terminal plate is thick and rounded. The base of the basal cup is considerably wider than the topmost columnals. The sides are faintly convex. The height of the cup is about two-thirds of the distal width. The orals form a cup slightly higher than the basals. The lateral edges of the orals are strongly bent outwards to give a high double ridge along each of the sutures between them. When the crown is seen with these ridges in profile the oral cup has more strongly convex sides than the basal. 2. In this stage the five radials and the radianal plate are present (Fig. 21 b). The crown is about i mm. long, the column 5-8 mm. PENTACRINOID LARVAE 203 The column consists of 28 columnals and a terminal plate. The first six columnals are of roughly equal lengths, short and discoidal. The remainder are somewhat barrel- shaped; each is encircled by a narrow median girdle and, with the exception of a small number following the discoidal columnals and a small number at the distal end, is slightly longer than broad. The articular faces of the longer columnals are broadly oval, the long axes of those of the two ends of one columnal being at right-angles to one another. The long axes of the opposing faces of two contiguous columns coincide. For these reasons the alternate articulations of a part of the column appear, from some angles, to be of different kinds. The terminal plate is thick and rounded. The base of the basal cup is not wider than the proximal columnals. Its height is equal to that of the oral cup ; it is considerably less than its distal diameter. The lateral edges of the oral plates are more strongly bent outwards, especially proximally where a furrow lies between those of adjacent plates, than in the pre-brachial stage. A small rounded radial plate is present in each of the angles between the basals and the orals. To the left of and a little below one of them which is smaller than the others is a slightly smaller plate, the radianal. In this radius the suture between the basals and that between the orals are not in line with one another as in other radii : the former is slightly to the left of the latter with the anal plate lying directly in hne with it ; the radial is in line with the suture between the orals. 3. Length of crown i-o mm. (Fig. 21 c); length of column 5 mm. The column consists of 25 segments and a terminal plate and is similar to, though shorter than, that of the previous stage. Most of the columnals are encircled by a strong median girdle. The terminal plate is small, round and simple. The basal cup is about two-thirds as long as its distal diameter. Its base is not wider than the proximal columnals ; its sides are only slightly convex. The oral cup is a little longer than the basal. The radials are very much bigger than in the previous stage; the corners of the basals are cut away to receive them. The right posterior radial is strongly asymmetrical being undeveloped on the left side where the much smaller radianal plate lies. The right distal angle of the posterior basal is strongly cut away to accommodate the radianal plate. Each of the radials except the right posterior bears a small costal, and a smaller axillary ; they lie in the furrow between the strongly turned out lateral edges of the oral plates but are massive enough to project beyond the edges of them in profile. The right posterior radial bears no ossicles. The suture between the posterior and the right posterior basals and that between the corresponding orals are only a little out of line with one another. 4. Length of crown i-o mm. ; length of column 4-6 mm. The column is of 27 segments and a terminal plate and is similar to that of the previous stage. The first seven columnals are short and discoidal ; the second is slightly longer and wider than the others which are of equal size. The terminal plate is thick and circular. The proximal edges of the basal plates are slightly rounded. The base of the basal cup is wider than the proximal columnals. It is shorter than its distal diameter and 204 DISCOVERY REPORTS shorter than the oral cup and its sides are nearly straight. The radials are slightly smaller than in the previous stage. The radianal is similarly placed. Each of the radials, in- cluding the asymmetrical right posterior, bears a costal and an axillary smaller than those of the previous stage. There is a strong contrast between the nearly straight-sided basal cup and the oral cup which has a broadly-rounded dome-like profile because of the strongly everted edges of its plates. Fig. 21. Pentacrinoid larvae of hornet la hordea. a, crown and column of No. i. b, crown and part of column of No. 2. c, crown and proximal columnals of No. 3. d, same of No. 5. e, same of No. 8. All X 26. 5. Length of crown 1-2 mm. (Fig. 21 d)\ length of column 5-4 mm. The column is of 28 segments and similar to that of the previous stage ; as in that stage the second of the proximal discoidal segments is larger than the others, which are of equal size. The terminal plate is large and irregular. The sides of the basal cup are only very slightly convex. Its base is of the same width as the proximal columnal with which it is closely associated. It is about two-thirds as long as its diameter at the basi-oral suture, a little longer than the oral cup. The lateral PENTACRINOID LARVAE 205 edges of the orals are still more strongly everted than in the previous stages ; the radials are considerably, the costals and axillaries very much, larger. The axillary is as large as the costal and carries two very small first brachials. The radianal causes the posterior basal and the right posterior radial to be asym- metrical. The suture between the posterior and the right posterior basal plates is out of fine with, to the left of, that between the corresponding oral plates; the axis of the ray coincides with the latter. 6. Length of crown 1-3 mm. ; length of column 67 mm. The column is of 33 segments. The first is very short and closely in contact with the basal cup. The next five are short and discoidal and become progressively and gradually smaller: that is, the second is shghtly longer and wider than the third, and so on. The remaining columnals are similar to those of the stages already described : all but two or three following the discoidal proximal segments and three or four near the terminal plate are longer than broad, somewhat barrel-shaped, each with a narrow projecting girdle ; the articular faces are broadly oval and the long axes of the two faces of each columnal are at right angles to one another. The terminal plate is simple. The base of the basal cup is a little wider than the topmost columnal. Its distal diameter is nearly twice its length ; its sides are nearly straight. The proximal portions of the lateral edges of the orals are widely separated by the well-developed radials, costals and axillaries against which they are turned up; the distal portions are in contact with one another. The radials are five-sided, about as long as broad; they do not meet one another. The posterior basal and the right posterior radial are made asymmetrical by the radianal ; the axis of the right posterior ray is not in line with, but considerably to the right of, the suture between the posterior and right posterior basals. The radianal appears to override the posterior oral. The costals and axillaries are strongly developed ; the latter reach as high as the apex of the oral dome and each carries a pair of first brachials, and some a pair of second brachials, which reach farther. This is the first stage in which the radial structures reach beyond the orals. 7. The crown is damaged, about 1-3 mm. long; length of column 7 mm. The stem is of 26 columnals and a small thick terminal plate. The first five columnals are short; the first and second are wider than the third to fifth. The remaining columnals, except for two or three following the short proximal columnals and three or four before the terminal plate, are longer than broad, the longest being one and a half times as long as broad. In other ways they resemble those of the stages already described. The basals and radials together form a straight-sided cup (as in the next stage, see Fig. 21 e). The radials are pentagons with rounded angles; they do not meet. The right posterior radial is made asymmetrical by the radianal which rests closely against it, and the ray it bears lies to the right of the line of the suture between the posterior and right posterior basals. The radianal is nearly in contact with the left posterior radial; it rests on the posterior oral. 2o6 DISCOVERY REPORTS Some of the arms are broken ; others are of three brachials curhng in over the apices of the oral plates (which appear to be in contact). 8. Length of crown 1-4 mm. (Fig. 21 e); length of column 77 mm. There are 31 columnals and a small thick round terminal plate. The first five columnals are short and discoidal. The first is as wide as the base of the basal cup, the second is as wide but a little longer, the third resembles the first ; all three are a little irregular. The fourth and fifth are narrower, shorter and more regular. The remainder, but for three or four following the discoidal segments and three or four adjoining the terminal plate, are considerably longer than broad — nearly one and a half times as long as broad. They are in other ways similar to those of the earlier stages described in this series; the median girdle is very faint and does not project. The sides of the basi-radial cup are nearly straight. The radials almost meet in all interradii except the posterior where they are widely separated by the radianal. The radi- anal still causes a strong asymmetry in the right posterior radial ; it is also in "contact with the left posterior radial. It rests on the posterior oral. The anterior arms are of four brachials and curl in over the orals. No other ossicles than the costals remain on the posterior radials: that on the left is smaller than that on the right. The orals rest against the lower edges of the costals ; they turn in under the axillaries and brachials and where the latter are lost it may be seen that the edges of the orals are in contact (Fig. 21 e). They are strongly everted. 9. Length of crown 27 mm. (Fig. 22 a-c); length of column 9 mm. The stem is of 30 columnals and a round and simple terminal plate. The first five segments are short and irregularly discoidal. The first three are wider and slightly longer than the fourth and fifth. The sixth is longer, but broader than long; the seventh is almost as long as broad, the eighth longer than broad. The ninth to about the sixteenth are considerably longer, nearly twice as long as broad. The remainder gradually decrease in length, the last two or three being shorter than broad. Most of the longer segments have a very faint median girdle which does not project. The sides of the basi-radial cup are straight. The radials, with the exception of the posterior pair, are in broad contact with one another. The posterior are separated by the radianal which lies in the midline of the posterior interradius, in contact with both radials and making each of them equally asymmetrical : whereas the other radials are seven-sided they each have six sides. The radianal rests on the posterior oral and ex- tends as far forward as the distal edges of the radials. The orals, which are in contact with the inner distal edges of the radials, are broad plates with the lateral edges of the distal portion, which bends strongly in over the disk, strongly everted and in contact with one another. The arms are of ten or eleven brachials. I can see no sacculi and side-plates along the ambulacra. There are no pinnules. 10. Length of crown 3-2 mm. ; length of column 10 mm. There are 34 columnals in the stem which is similar to that of No. 9. There are six short irregularly discoidal proximal columnals. The first is very short and closely PENTACRINOID LARVAE 207 attached to the basal cup. The second to fourth are wider and more robust than the first and fifth and sixth ; the second is wider than the third and the third than the fourth. None of the columnals has a median girdle. The terminal plate is large. The sides of the basi-radial cup are straight. All but the posterior radials are in broad contact. The proximal corners of the posterior radials meet, cutting off^ the radianal plate from contact with the basal for the first time ; the distal portions of the posterior radials are separated by the radianal. Its centre lies to the right of the midline of the posterior interradius and it is in contact with the costal of the right posterior ray as well as with the two radials; its centre is in line with the articulation between the radial and the costal. Both the posterior radials are made asymmetrical by the radianal, the right far more so than the left, for its left distal corner is cut away to accommodate the radianal. The radianal overlaps the posterior oral. The bases of the orals are in contact with the inner distal edges of the radials. They are flat plates with slightly out-turned edges which bend in over the disk, where their edges appear to meet, and are evenly rounded distally: they do not narrow to bar-like projections. The costals and axillaries are considerably stouter than in the previous stage (No. 9). The arms are of 12 or 13 brachials. There are no pinnules and I can see no sacculi or side-plates along the arm ambulacra. 11. Length of crown 3-5 mm.; length of column 11 mm. There are 36 columnals and a large lobed terminal plate. The first five columnals are short and irregular. The first and second are as wide as the base of the basal cup, the second a little longer than the first. The third and fourth are slightly, the fifth con- siderably, narrower. The remainder of the stem is similar to those of Nos. 9 and 10. None of the segments has a median girdle. The sides of the basi-radial cup are straight. All the radials are in broad contact though that of the posterior radials is incomplete distally because of the radianal which lies in contact with both of them and the costals they bear. It is a nearly circular plate lying in the mid-line of the posterior interradius, its proximal edge opposite the articulations between the radials and the costals, its centre opposite a point a third of the way along the costals. Its contact is closer with the right than the left posterior ray. It lies much nearer to the outside than the oral. The orals lie deep within the arms, their bases in contact with the inner distal edges of the radials, their lateral edges underneath the edges of the arms on either side. They bend in over the disk, gradually narrowing as they do so : their edges probably do not meet over the disk. The costals, axillaries and brachials are considerably more robust than in No. 10. The arms are of about 14 brachials; the ninth and tenth brachials of some arms bear the beginnings of the first pinnules to be formed. I see no sacculi and no side plates along the arm ambulacra. 12. Length of crown 5-6 mm. (Fig. 22 d); length of column ca. 12 mm. There are 36 columnals and a large lobed terminal plate. The first columnal is short 208 DISCOVERY REPORTS and as wide as the base of the basal cup to which it is closely attached. It bears the beginnings of five radially situated cirri, each, except for the left posterior which is broken off, of about three to five segments. The proximal border of the second columnal is deeply notched radially opposite each of the developing cirri. The third to fifth colum- Fig. 22. Pentacrinoid larvae of Isometra hordea. a, crown and proximal columnals of No. 9. b, tenth to fourteenth columnals of No. 9. c, last columnals and terminal plate of No. 9. d, proximal portions of crown and column of No. 12. e, same of No. 14. /, same of No. 15. All x 18. nals are as short as the first and second; they are irregularly discoidal. The sixth to eighth are longer, the ninth slightly longer than broad. The remainder of the stem is similar to that of No. 1 1 ; none of the columnals has a median girdle. The radials are longer than the basals. They are all in broad and complete lateral contact. The right distal corner of the left posterior radial extends farther forwards than the contiguous corner of the right posterior radial ; the radianal plate is in contact with PENTACRINOID LARVAE 209 the former but not with the latter. It is a long narrow plate lying between the posterior rays, just below the level of their edges. The oral plate can be seen deep within the arms behind it (but not in such a way as to be shown in the figure). The proximal border of the radianal plate is half-way along the costal, its distal border opposite the beginnings of the first brachials. In two of the remaining interradii the arms are too closely pressed together for the oral plate to be seen. In each of the other two the oral is visible, its base opposite the middle of the axillary, i.e. for the first time out of contact with its radials, its distal portion sloping steeply inwards. I can see no plates in the perisome separating it from the radials. The costals, axillaries and brachials are more massive than in the previous stage. The lateral edges of the axillaries are parallel. The arms are of about 20 brachials; from the ninth brachial onwards there are pinnules of up to eight segments. There are a few small irregularly arranged sacculi on some arms. There appear to be small side-plates along the arm ambulacra. 13. Length of crown 5-4 mm. ; length of column 12 mm. The column is of 34 columnals and a large lobed terminal plate. It is very similar to that of No. 12. The first columnal bears five slightly longer cirri, the longest as long as the side of a basal plate. The crown is very similar to that of No. 12. The sides of the basi-radial cup are straight in the proximal basal portion, bulge slightly outwards in the radial portion. The right distal corner of the left posterior radial extends a little farther forward than the contiguous corner of the right posterior radial — not nearly so much so as in No. 12 : the radianal is far removed from it. The proximal edge of the radianal plate is opposite a point half-way along the costal ; the proximal edges of the orals are opposite the distal edges of the costals. There are a few sacculi and small side-plates along the arms. 14. Length of crown 6-8 mm. (Fig. 22 e); length of column 13 mm. There are 35 columnals and a large lobed terminal plate. The first columnal is longer and more massive than in the previous stages. It bears two whorls of cirri, a radial whorl of five larger cirri arising from large sockets occupying nearly the entire length of the segment, and an interradial whorl of five very small cirri arising from the proximal half of the segment. The longest of the radial cirri consist of more than 12 segments and are longer than the basal plates. The second columnal is discoidal, of the same diameter as, but shorter than, the first ; its proximal edge is not notched opposite the cirri of the first. The third is shorter and narrower than the second and irregularly discoidal. The fourth and fifth are still narrower but slightly longer. The remainder of the column resembles that of the stages described above. The contiguous distal corners of the posterior radials are unequal as in Nos. 12 and 13. The radianal plate is more distant from them than in those younger specimens: its proximal edge is opposite a point a third of the way along the axillary ; it extends to a third of the way along the first brachial. It is smaller than in specimen 13. It rests on the anal tube which ends opposite the distal half of the first brachial. The oral plate may be seen deep behind the end of the anal tube. In other interradii it may be seen that the 2IO DISCOVERY REPORTS Other oral plates are pushed farther distally than in the younger specimens; their proximal edges are level with those of the first brachial. The costals and axillaries are as wide as the radials and of the shapes shown in Fig. 20^. The arms are of 22-24 brachials with pinnules arising from the tenth and succeeding brachials. The longest pinnules are of ten or more segments. There are side-plates and a few sacculi along the arms. 15. Length of crown ca. 11 mm. (Fig. 22/). This is a far bigger and older specimen than the last (No. 14). The stem is of 32 columnals and a very large and lobed terminal plate. The first columnal is considerably longer than in No. 14, and tapers somewhat distally. It bears three whorls of cirri. The largest is a radial whorl arising from the distal half of the columnal, one in each radius ; the longest are of 27 segments and extend as far as the seventh brachial. They terminate in claws and their penultimate segments carry small opposing spines. A whorl of much smaller cirri arises interradially from half-way along the columnal. There are five cirri in this whorl, one in each of three interradii, two in the fourth and none in the fifth which is contiguous with it. The beginnings of a third, radial, whorl arise from the proximal border of the columnal. The second columnal is short and discoidal and slightly wider than the tapered end of the first ; its proximal border is incised opposite each of the large radial cirri, so deeply opposite some cirri as to extend for the length of the segment and so cut its periphery into lobes. The third columnal is short and discoidal and very slightly narrower than the second. The fourth and fifth are irregular and narrower and longer. The remainder of the columnal is similar to the same parts of those of the stages already described. The basal plates are very reduced ; in the radial mid-line they appear to be less than one-sixth as long as the radials. The radials are long and in close contact with one another. The costals and axillaries are as wide as the radials. The arms are too closely pressed together for the oral plates or the radianal to be seen. The arms are of about 30 brachials. Syzygies occur between brachials 3 + 4, 9 + 10 or 10+ II, 13 + 14, 18+ 19. Long pinnules or 12 or more segments arise from the tenth and succeeding brachials. The beginnings of the first pinnules (Pj) are present on the second brachials. There are fairly regularly arranged small dark sacculi along the arms and small side- plates along the arm ambulacra. There are small side-plates, one or two to each segment, along the pinnule ambulacra and numerous spicules in the tentacles. A few of the spicules are nearly smooth but most are strongly thorny and some are branched. The side-plates are not unlike, and the spicules resemble, those of adult Isometra hordea. Pentacrinoid larvae of Notocrinus virilis Mortensen St. 170. 23. ii. 27. Off Cape Bowles, Clarence Island. 61° 25' 30" S, 53° 46' W. 342 m. Gear DLH. Bottom: rock. Eighteen specimens (Nos. 1-7, 10-20). St. 175. 2. iii. 27. Bransfield Strait. 63° 17' 20", 59° 48' 15" W. 200 m. Gear DLH. Bottom: mud, stones and gravel. Two specimens (Nos. 8 and 9). PENTACRINOID LARVAE 211 That these twenty pentacrinoids from the Bransfield Strait region are of N. virilis is shown by the fact that they all carry in superficial pits in the orals, and sometimes in other adjacent ossicles, glandular sacs similar to those described from the posterior end of the embryos by Mortensen (1920) who has described the eggs and embryos of the species. One of the larvae (No. 7) occurred on a cirrus of an adult A^. virilis. The last three of the series (Nos. 18-20) have small plates in the naked perisome which separates the oral plates from the radials : the perisome of adult A^. virilis is strongly plated. The oldest larva has along the arm ambulacra strong side- and cover-plates of the same nature as those of N. virilis. The series is not so complete as that of Isometra hordea. There is no prebrachial stage : the youngest larva has arms of two or three brachials. The oldest has only one whorl of cirri on the topmost columnal of its stem; its arms are of nineteen brachials. The adult crinoids taken at St. 170 are shown on p. 202. Those at St. 175 were: Promachocrinus kerguelensis 3 Phrixometra mitrix i Isometra vivipara I Notocrinus virilis 7 I. Length of crown ca. 1-4 mm. (Fig. 23 a); length of column ca. 6-4 mm. The column is of 34 columnals and a thick roughly circular terminal plate which appears to be simple: I cannot see any supplementary plates like those known in the larva. The first seven columnals are short and discoidal. The most proximal is in close contact with the basal plates. From the second to the seventh there is a gradual decrease in width. The eighth is about one-third, the ninth about a half, as long as broad ; the proximal half of each is encircled by a narrow projecting girdle. The remain- ing segments are about as long as broad except for five or six at the distal end which are shorter. They are rounded off to, and narrower at, each end than in the middle, where they are encircled by a narrow projecting girdle. They are somewhat barrel-shaped. The articular faces are very broadly oval, the long axes of those at the opposite end of each columnal at right angles to one another. The columnals are of a coarse texture like the ossicles of the crown. The sides of the basal cup are strongly rounded. Its height is equal to about half its diameter at the distal end. The distal corners of the basal plates are deeply cut away to receive the radials. The crown is damaged on one side and one ray has been broken away: it must be the right posterior ray, for there is no sign of the radianal plate against any of the other radials. In three rays that are complete the arms are of two, or some- times three, brachials. They curl in above the apices of the orals. The surfaces of the lower part of the oral plates are only a little sunken below the level of those of the basals, the radials, the costals and the axillaries with which their edges are in contact. The texture of all the plates is coarse but that of the orals coarser than any ; the surface of the proximal part of each is deeply pitted and some of the pits carry pale yellow spherical bodies which must be the glandular sacs known from the posterior end of the embryo (Mortensen, 1920, p. 52, pi. xxiv, fig. 2). In preparations of embryos slightly older than that figured by Mortensen I have seen the sacs lying close 212 DISCOVERY REPORTS against the oral plates, and although they were not partly enclosed in their stereom it seemed that they might easily become so as growth went on. Such sacs occur in all the older larvae to be described in this series. They are some- times abundant and may occur in superficial pits in the surfaces of the basals, the radials, the costals and the axillaries as well as in the orals. 2. Length of crown i-6 mm. (Fig. 23 b); length of column 67 mm. The stem is of 36 columnals and similar to that of the previous stage. A larger num- ber, about 12, of the distal columnals are shorter than broad and they are slightly wider than the other columnals; they become progressively wider as they approach the ter- minal plate. The terminal plate is simple, thick and slightly irregular. The radials are widely separated from one another. The rays are bigger than in No. i ; the arms are of five brachials and extend well beyond the orals. The proximal edges of each oral rest against the basal, the radials and the costals of its interradius. Opposite the axillaries it narrows into a strap-like projection with slightly out-turned edges which curves in over the disk. The projections of the five orals do not enter into close contact. The orals are of very coarse texture. The surface of the proximal portion of each is pitted and some of the pits carry glandular sacs of diverse sizes. One or two of the radials and costals carry similar sacs. There is a single sacculus on one of the arms which has exactly the same appearance as the glandular sacs. The radianal plate occupies an unusual position : its centre lies to the left, not right, of the mid-line of the posterior interradius and it lies close against, and causes asym- metry in, the left, not the right, posterior radial. The corner of the right posterior radial is just in contact with it. The distal half of the radianal rests upon the posterior oral. 3. Length of crown 1-9 mm. (Fig. 23 c); length of column 8 mm. The stem is of 39 columnals and a terminal plate which is large and lobed but does not appear to be composed of more than one element. The column is generally similar to that of the younger stages but there are slight diff^erences : the discoidal proximal seg- ments, of which there are nine, are even shorter, whereas the majority of the middle segments are slightly longer, being a little longer than broad. The radials are widely separated. Three of the rays are stronger than in the previous stage, the arms of about six brachials. The anterior and the left anterior rays are small, bearing only the first brachials; they are shorter than the orals. The lateral edges of the orals become free opposite the costals, beyond which the plates rapidly narrow to in- wardly curved strap-like ends. Some of the orals are in contact with the basal plates of their interradius, others separated from it by a very narrow strip of perisome. A wide area of perisome separates the posterior oral from its basal. On it lies the radianal plate with its distal end overlapping the oral. The centre of the radianal plate lies to the right of the mid-line of the posterior inter- radius but it is not in contact with the right posterior radial, which, nevertheless, is strongly asymmetrical. The axis of its ray lies far to the right of the suture between the posterior and right posterior basals. The proximal left-hand corner of the radianal touches the left posterior radial. Strong side-plates are developed along the ambulacra of the arms. PENTACRINOID LARVAE 213 4. Length of crown 2-1 mm. ; length of column 7 mm. The stem is of 41 columnals and a thick lobed terminal plate which appears to be simple; it is generally similar to that of the previous stage. There are ten very short Fig. 23. Pentacrinoid larvae of Notocrinus virilis. a, crown and proximal columnals of No. i. b, same of No. 2. c, same of No. 3. d, same of No. 6. All X27. discoidal proximal columnals, those nearer the basal cup wider than the others. Most of the remaining columnals in the first half of the stem are nearly as long as broad but those of the distal half become progressively and gradually shorter towards the terminal plate. 214 DISCOVERY REPORTS The basal cup has strongly rounded sides. The radials are widely separated from one another. The arms of only two rays are complete ; they are of five brachials. The radianal plate is in broad contact with the right posterior radial which it makes asymmetrical; it is in contact with the proximal half of the left posterior radial. It overrides the oral ; no naked perisome is to be seen in this interradius. In other interradii the orals may be separated from the basals by a narrow strip of perisome. The edges of the orals become free opposite the costals, beyond which the plates narrow as they turn in over the disk. There are side-plates along the ambulacra of the arms. 5. Length of crown 1-9 mm. ; length of column 9 mm. The column is generally similar to that of previous stages. There are 43 columnals and a round terminal plate which appears to be simple. The first seven columnals are short and discoidal and become progressively more narrow from the first to the seventh. The tenth to about the twentieth are about as long as broad, the remainder broader than long. The crown is smaller than in the previous stage and the basal cup is not so strongly rounded. Some of the radials are widely separated, others less so, and one pair, the right posterior and the right anterior, are in contact. The centre of the radianal plate is to the right of the mid-line of the posterior interradius ; it is in broad contact with the right posterior radial which it makes asymmetrical. It is in less complete contact with the left posterior radial. None of the orals is separated from the more proximal ossicles of its interradius by naked perisome. They are shaped as in the previous stage. The arms are of about six brachials. 6. Length of crown 2 mm. (Fig. 23 d) ; length of column 9 mm. The stem is similar to that of the previous stage. There are 42 columnals and a large, rather thin and slightly lobed, terminal plate which does appear to be made up of more than one element. The sides of the basal cup are strongly rounded ; its height is about half its distal diameter. The radials, except for the right and left posterior, are nearly or quite in lateral contact. By their lateral growth they have pushed the orals out of contact with the basals ; where they are not quite in contact small naked areas of perisome separate them from one another. The right and left posterior radials are separated by the radianal plate which lies in broad contact with the former and in contact with the proximal third of the latter. It causes the right posterior radial to be asymmetrical and the longitudinal axis of its ray lies far to the right of the suture between the basals on which it rests. The radianal plate rests on the oral. The lateral edges of the orals become free opposite the end of the costals ; beyond that point they narrow and bend in over the disk. They are of the usual coarse texture, deeply pitted proximally with some of the pits carrying glandular sacs. The left posterior ray is broken off" at the costal. The arms of the other rays are of about eight brachials. There are a few small sacculi of the same appearance as the glandular sacs. The arm ambulacra are lined by large side-plates. PENTACRINOID LARVAE 215 7. Length of crown 2-2 mm. ; length of column 10 mm. There are 45 columnals and a large faintly lobate terminal plate which is single. The first eight columnals are short and discoidal, of roughly equal length but gradually decreasing in diameter from the first to the sixth. The remainder of the column re- sembles those of previously described stages. The radials are widely separated. The radianal is in contact with the right posterior but not the left posterior radial. Its proximal half lies on the naked perisome which separates the posterior basal from the oral ; its distal half lies on the oral. The orals of the other interradii are separated from their basals by very narrow strips of perisome. They are shaped as in previous stages and are richly supplied with glandular sacs : there are as many as eight on one plate. Three of the rays have arms of eight brachials. The costals and axillaries of the right posterior ray are smaller than those of others and its arms are of no more than three or four brachials; the right anterior ray has equally short arms but they appear to be broken, not undeveloped. The right posterior radial is made asymmetrical by the radianal, and the whole of the ray, but for a corner of the radial itself, lies to the right of the line of the suture between the basals upon which it rests. There are a few irregularly placed sacculi, of a similar appearance to the glandular sacs, along the arms, and strong side-plates along the ambulacra. 8. Length of crown 2-4 mm. ; the column is incomplete. Only 13 columnals remain. The first five are short and discoidal but longer than those of previous stages. The sixth is about half as long as wide, the remainder about as long as wide ; they are encircled by a narrow projecting girdle which is in the proximal half of the sixth columnal but median in all the others. The crown is very similar to that of the previous stage except that : the radianal plate lies entirely to the right of the mid-line of the posterior interradius ; all the arms are of equal size, of about six brachials ; the basals and one of the radials — as well as the orals — carry many glandular sacs in pits. 9. Length of crown 3 mm. (Fig. 24 a); the column is incomplete. Only 29 columnals remain. The first is short and closely associated with the basal cup. The second to fifth are short and discoidal. The sixth is discoidal but longer. The seventh is about half as long as broad. The eighth to near the twentieth are as long as broad, the remainder slightly broader than long. The crown is similar to, though larger than, those of the two previous stages. Each of the basal plates is swollen so that the sutures between them run along depressions. The distal half of the radianal plate lies, not on the oral, but on a lobe of tissue at a higher level than, and overlapping the proximal edge of, the oral plate. I assume it to be the beginnings of the anal tube. As in the previous stage from the same station (St. 175) there are glandular sacs on other ossicles than the orals. The arms are of seven or more brachials. 2l6 DISCOVERY REPORTS There are five specimens of roughly the same age as those described as numbers 3 to 9 above, though for the most part they are shghtly larger. They do not call for individual description. The terminal plate appears to be single in all, though it is difficult Fig. 24. Pentacrinoid larvae of Notocrinus virilis. a, proximal portions of crown and column of No. 9, X 18. b, same of No. 15, x 14. c, tenth to thirteenth columnals of No. 15, x 26. d, last columnals and terminal plate of No. 15, x26. e, a few brachials of No. 15 showing size of side- and cover-plates, X26. /, proximal portions of crown and column of No. 18, x 14. g, same of No. 20, x 14. to be certain that it is. The longest columnals of the stem of one or two are longer than any yet described. In some the orals are in contact with the basals, in others separated from them by narrow strips of perisome. Most of the radials are nearly in contact with one another in all. In none of the specimens does the anal tube show behind the radianal plate as in number 9. PENTACRINOID LARVAE 217 The measurements of the specimens and the number of columnals in the stem and of brachials in the arms are as follows : 10. Crown 2-9 mm.; stem: 39 columnals, 9-6 mm.; arms of about 8 brachials. 11. Crown 2-5 mm.; stem: 43 columnals, 10 mm.; arms of about 8 brachials. 12. Crown 2-5 mm.; stem: 34 columnals, 11 mm.; arms of 8-10 brachials. 13. Crown 2-9 mm.; stem: 39 columnals, 10 mm.; arms of 8-10 brachials. 14. Crown 2-7 mm.; stem: 44 columnals, 10-4 mm.; arms of 10 brachials. The next specimen is considerably older than any of the preceding : 15. Length of crown 4 mm. (Fig. 24 b-e); length of column 10-5 mm. The stem is of 39 columnals and a thick slightly lobed terminal plate. The first columnal is incomplete so that the second, although for the most part in touch with it, is in touch with the basal cup too; it is not shown in the figure, in which the most proximal columnal is the second. The second to seventh columnals are short. The third is narrower than the second. The fourth to seventh are of the same diameter as one another, narrower than the third. There is a gradual increase in length from the fourth to the seventh columnal, the seventh being half as long as wide. The tenth is as long as broad; the eleventh to the fourteenth are slightly longer than broad (Fig. 24 c). The remaining columnals gradually decrease in length : the most distal are wider than long (Fig. 24 d). Only faint traces of the encircling girdles of the columnals remain. The articulating surfaces of the two ends of each columnal, or at least of the longer columnals, are broadly oval with the longer axis of one end at right angles to that of the other ; this is only just perceptible. The sides of the basi-radial cup are nearly straight ; its height is about half its distal diameter. The basals are considerably longer than the radials. The radials, except for the posterior pair, are in broad contact with one another; beyond the parts of the lateral edges which are in contact they narrow a little. The middle of the wide distal edge is indented for the articulation of the costal ; it occupies about a half of the distal edge of the radial. The lateral edges of the posterior radials meet proximally, but are cut away beyond to allow room for the radianal plate: it follows that both the radials are asymmetrical, and they are equally so. The radianal plate is diamond-shaped and longer than broad. Its proximal half is in contact with the radials and costals ; the distal half lies on the anal tube which is much wider than it and has gently converging lateral edges and a straight distal margin. It is very near the edge of the disk : the level of the radianal plate is only a little below that of the radials. The narrow end of the posterior oral may be seen beyond the anal tube, far below its level, curving over the disk (Fig. 24 b). The other oral plates bend sharply in over the disk and rapidly narrow in their distal halves ; they are separated from the radials by a narrow naked strip of perisome. The surface of the wider proximal part of each oral is pitted and some of the pits carry glandular sacs. The arms are of 15-16 brachials with large side- and long cover-plates (Fig. 24 e) and irregularly arranged sacculi. On the distal portions of some arms there is one sacculus D .Will '3 2i8 DISCOVERY REPORTS to each segment. The rudiments of pinnules arise from the eleventh and succeeding brachials. i6. Length of crown 3-5 mm.; length of column 11 mm. The column is similar to that of No. 15. There are 40 columnals and a damaged terminal plate. The crown is similar to that of No. 15. The radials are longer, being only slightly shorter than the basals ; the dorsal surface of each is strongly rounded. The posterior radials are in contact for more than half their length, beyond which they are cut away to accommodate the proximal part of the radianal. The other radials are in complete lateral contact. The radianal plate rests on the anal tube, completely covering it. There is no naked perisome in the posterior interradius, and the orals of other interradii are either in contact with the radials or separated from them by an extremely narrow strip of perisome. 17. Length of crown 5 mm. ; length of column 9 mm. There are 37 columnals and a large lobed terminal plate. The first columnal is short and it bears the rudiment of a cirrus in each of the radii except the right posterior. The remainder of the column resembles that of earlier stages ; each segment is encircled by a faint but distinct median girdle. The crown is generally similar to, though smaller than, those of Nos. 15 and 16. The radials, including the posterior pair, are in complete lateral contact. The dorsal surfaces of the ossicles of the division series and of the brachials are strongly rounded. The arms are of seventeen brachials of which the tenth and succeeding carry pinnules. 18. Length of crown 5 mm. (Fig. 24/). The stem is of 40 columnals and a large lobed terminal plate. The first columnal is short and it bears the rudiment of a cirrus in the anterior, the left anterior, and the left posterior radii. The second columnal is as wide as the first and longer. The third to fifth are narrower and discoidal. The remainder of the column resembles those of younger stages. The basal plates are slightly longer than the radials. The radials are in complete lateral contact. The radianal plate is far out of contact with the posterior radials, being separated from them by naked perisome ; its proximal edge is opposite the distal end of the costals. It is an oval plate resting on the anal tube which extends some way beyond it_to opposite the end of the first brachial. The posterior oral plate is clearly visible behind the anal tube. The proximal borders of the orals of other interradii are opposite a point half way along the axillaries so that they are widely separated from the radials by an area of perisome; a number of plates are present in it. The oral plates are of coarse texture with their proximal surfaces pitted as in younger specimens. The distal end of each narrows to a bar which bends in over the disk. The rays resemble those of No. 17. A few large sacculi occur at irregular intervals along the arms. 19. Length of crown 8 mm. ; length of column 13 mm. There are 41 columnals and a large terminal plate which is damaged. The first PENTACRINOID LARVAE • 219 columnal is much bigger than in any of the younger stages : it is more than one-third as long as broad. Long cirri of about 15 segments, which reach the ends of the radials, arise from it in all radii but the posterior in which there is only the beginning of a cirrus, about one-quarter the length of the basal plates. The second segment is as wide as the first but it is short. The third and fourth are narrower and they are discoidal. The column is otherwise similar to those of younger stages. The basals are longer in proportion to the radials than in No. 18. In other ways the crown resembles that of No. 18. The only complete arm is of 22 brachials with pinnules arising from the eleventh and succeeding brachials. There are very strong side- and cover-plates along the arm and pinnule ambulacra and a few irregularly arranged sacculi along the arms. 20. Length of crown 9 mm. (Fig. 24^); length of column 13 mm. The stem is of 38 columnals and a large lobed terminal plate. The first columnal is not so big as in No. 19 but its cirri, of which there is a single radial whorl, are longer though they are of unequal lengths. The anterior and left anterior are of about 22 seg- ments and reach to the first brachial ; the right anterior is somewhat shorter ; the left posterior is broken ; the right posterior is a rudiment one-third as long as the basal plate. The cirri arise from sockets which incise the columnal for its entire length. The second columnal is as wide as, and longer than, the first. There are indentations in both its proximal and distal margins opposite the larger cirri of the first columnal. The remainder of the column resembles that of slightly younger specimens. The columnals of the middle part of the stem are each encircled by a narrow median girdle. The crown resembles those of Nos. 18 and 19. A glandular sac occurs in a pit in the surface of one of the axillaries as well as in the orals. The arms are of 19 brachials. Pinnules of up to nine or ten segments occur on the ninth and succeeding brachials ; the lower pinnules are not yet beginning to form. There are strong side- and cover-plates, of the same nature as those of the adult Notocrimis virilis along the ambulacra. There are a few irregularly arranged sacculi along the arms. 13-2 320 DISCOVERY REPORTS ANTARCTIC COMATULIDS PREVIOUSLY IN THE BRITISH MUSEUM COLLECTION From the National Antarctic (^Discovery') Expedition, 190 1-4 Promachocrinus kerguelensis Off Coulman Island 100 fathoms 4 specimens Winter Quarters I ,, )> No. 10 Hole 4 •>-i East end of Barrier 100 fathoms I >» Mount Erebus and Terror 500 I J» Anthometra adriani Winter Quarters 124 fathoms 2 >) »> 130 2 ■>■> j» No. 10 Hole 3 j> Mount Erebus and Terror 500 fathoms 6 >> Florometra mawsoni Winter Quarters 178 fathoms I »j Mount Erebus and Terror 2 1) From the British Antarctic (' Terra Nova') Expedition, 1910 The localities are all in the Ross Sea. Their positions are given in Harmer and Lillie, 19 14, I-ist of Collecting Stations, Natural History Report, Vol. 11, No. i of the Expedition's series of reports. Promachocrinus kerguelensis St. 194 St. 294 St. 314 St. 316 St. 317 St. 339 2 specimens 3 I ,, 1 ,, 2 ,, 5 St. 340 St. 341 St. 349 St. 355 St. 356 4 specimens I 9 6 I Anthometra adriani St. 194 St. 225 I specimen I St. 316 Off Barne Glacier 5 specimens St. 314 3 »» Florometra mawsoni St. 314 4 specimens St. 316 I specimen Notocrinus virilis St. 295 St. 341 2 specimens I St. 314 fragments REFERENCES azi REFERENCES Andersson, K. a., 1904. Brutpflege bei Antedon hirsuta Carpenter. Wissensch. Ergebnisse d. Schwed. Siidpolar-Exped. 1901-3, v, pp. 1-7, pis. i-ii. Bell, F. Jeffrey, 1908. Echinoderma. I. National Antarct. Exped., iv, 16 pp., 5 pis. 1917. Echinoderma. I. Actinogonidiata. Brit. Antarct. ('Terra Nova') Exped., 1910. Nat. Hist. Rep., Zool., IV, No. i, 10 pp., 2 pis. Bernasconi, I., 1932 a. Notas sobre un crinoideo de Sud Georgia. Physis, Buenos Aires, xi, pp. 86-7, fig. 1932 b. Notas sobre un crinoideo de Sud Georgia. An. Mus. Argent. Buenos Aires, xxxvu, pp. 29-35, figs. 1-3, pi. i. Carpenter, P. H., 1888. Crinoidea. II. The Comatulae. Challenger Reports, Zool, xxvi, 400 pp., 70 pis. Clark, A. H., 1908. New genera of unstalked crinoids. Proc. Biol. Soc. Wash., xxi, pp. 125-36. 1913. Notes on the recent crinoids in the British Museum. Smithson. Misc. Coll., LXI, No. 15, pp. 1-89. 1915 a. Die Crinoiden der Antarktis. Deutsche Sudpolar-Exped. 1901-3, Zool., viii, pp. 103-210, pis. ii-x. 1915 b. A Monograph of the Existing Crinoids. Vol. i. The Comatulids. Part i. Bull. U.S. Nat. Mus., Lxxxii, 406 pp., 513 figs., 17 pis. 1917. A Revision of the Crinoid Family Antedonidae, with the Diagnoses of Nine New Genera. J. Wash. Acad. Sci., vii, No. 5, pp. 127-31. 1918. A new Crinoid from New Zealand, and another from Tasmania. Proc. Biol. Soc. Wash., xxxi, pp. 41-3. 192 1. A Monograph of the Existing Crinoids. Vol. i. The Comatulids. Part 2. Bull. U.S. Nat. Mus., LXXXII, 796 pp., 949 figs., 57 pis. 1923. Crinoidea. Danish Ingolf-Expedition, Vol. iv. Part v, 60 pp., 58 figs. 1929. On Some recent Crinoids in the Collection of the British Museum. J. Linn. Soc. Lond., Zool., XXXVI, pp. 635-64, pis. 40-4. 1931. A Monograph of the Existing Crinoids. Vol. i. The Comatulids. Part 3. Bull. U.S. Nat. Mus., LXXXII, 816 pp., 82 pis. 1937. Crinoidea. Sci. Rep., Sen C, Zool. and Bot., viii. No. 4, Australasian Antarct. Exped. 1911-14, 18 pp. Deacon, G. E. R., 1937. The Hydrology of the Southern Ocean. Discovery Reports, xv, pp. 1-124, pis. i-xliv. Ekman, Sven, 1925. Holothurien. Further Zool. Res. Swedish Antarct. Exped. 1901-3, I, 6, 194 pp., 37 figs. GiSL^N, T., 1928. Notes on some Crinoids in the British Natural History Museum. Arkiv for Zoologi, 19, A, No. 32, Stockholm, 15 pp. Grieg, James A., 1929 a. Some Echinoderms from the South Shetlands. Bergen Museum Arbok, iii, 10 pp. 1929 b. Echinodermata from the Palmer Archipelago, South Shetlands, South Georgia and Bouvet Island. Sci. Res. Nonveg. Antarct. Expeds. 1927-8 and 1928-9. Oslo. 16 pp. Hartlaub, C, 1912. Die Comatuliden. Reports of the Results of Dredging in the Gulf of Mexico, etc. Mem. Mus. Camp. Zool., xxvii. No. 4, pp. 277-491, 18 pis. John, D. Dilwyn, 1937. Crinoidea. Res. du Voy. de la Belgica 1897-99. Rapp. Scientif., Zool., 12 pp., 3 figs. Koehler, R., 1912. Schinodermes {Asteries, Ophiures et Echinides). Deux. Exped. Antarct. Fran?., 1908-10. 270 pp., 16 pis. Minckert, W., 1905. Das Genus Promachocrinus, zugleich ein Beitrag sur Faunistik der Antarktis. Zool. Anz., xxviii, pp. 490-501, 2 figs. MoRTENSEN, Th., 1910. The Echinoidea. Wissensch. Ergebnisse d. Schwed. Sudpolar-Exped. 1901-3, vi, 4, 114 pp., 19 pis- igi8. The Crinoidea. Wissensch. Ergebnisse d. Schwed. Siidpolar-Exped. 1901-3, vi, 8, 24 pp., 5 pis- 222 DISCOVERY REPORTS MoRTENSEN, Th., 1920. Studies in the Development of Crinoids. Pap. Dept. Mar. Biol., Carnegie Inst. Wash.. XVI, 94 pp., 28 pis. 1925 a. Echinoderms of New Zealand and the Auckland-Campbell Islands. III-V. Asteroidea, Holotlmroidea and Crinoidea. Vidensk. Medd. fra Dansk. naturh. Foren., lxxix, pp. 261-420, pis. xii-xiv. 1925 b. On a small collection of Echinoderms from the Antarctic Sea. Arkiv for Zoologi, xvii, A, No. 3 1 Stockholm. 12 pp., 8 figs. 1937- Echinoidea and Ophiuroidea. Discovery Reports, xii, pp. 199-348, pis. i-ix. Norman, J. R., 1937. Fishes. Reports — Ser. B. (Zool. and Bot.). B.A.N.Z. Antarct. Research Exped. 1929-31. PP- 51-81. II figs. 1938. Coast-Fishes. Part III. The Antarctic Zone. Discovery Reports, xviii, pp. 1-104, i pi., 62 figs. Regan, C. T., 1914. Fishes. Brit. Antarct. Exped. 1910, Zool., i. No. i, 54 pp., 13 pis. Vaney, C, 1910. Une nouvelle espece de Promachocrinus. Bull. Mus. d'Hist. Nat., Paris, pp. 158-162, 2 figs. Wilton, Pirie and Brown, 1908. Zoological Log. Sci. Res. Scott. Nat. Antarct. Exped., iv, Zool., i, 106 pp. 33 pis. and 2 maps. PLATE III Fig. I. Promachocrinus kerguelensis. Fig. 2. Florometra mawsoni. Fig. 3. Florometra antarctica. Fig. 4. Anthometra adriani. All figures natural size. DISCOVERY REPORTS VOL XVIII PLATE III inBaleSons &. Cumow. V^^ London. CRINOIDEA PLATE IV Fig. I. Eutnorphometra aurora, x 2. Fig. 2. Eutnorphometra fraseri, x 2. Fig. 3. Eutnorphometra marri, x 2. Fig. 4. Kempometra grisea, x 2. Fig. 5. Phrixometra longipmna var. antarctica, St. 1948, x 4. Fig. 6. Phrixottietra loiigipinna var. atitarctica, St. 156, x 4. Fig. 7. Phrixometra nutrix, x 4. Fig. 8. Phrixometra rayneri, x 4. DISCOVERY REPORTS VOL . XVUI PLATE IV John Bala Sol IS 4. Cumow. L'-'* Limdun. CRINOIDEA r PLATE V Fig. I. Isometra vivipara, S- Fig. 2. Isometra vivipara, ?. Fig. 3. Isometra flavescens, (J. Fig. 4. Isometra flavescens, ?. Fig. 5. Isometra graminea, 3. Fig. 6. Isometra graminea, $. Fig. 7. Isometra hordea, (J. Fig. 8. Isometra hordea, ?. All figures natural size. DISCOVERY REPORTS VOL . XVll] PLATE V John fiaJe. Sons i. C.jrnow V-^ London CRINOIDEA PLATE VI Fig. I. Notocrintis virilis, ?, natural size. Fig. 2. Notocrinus mortenseni, $, natural size. DISCOVERY REPORTS VOL. XVIII PLATE VI JohnBale Sons i Cumovw L*^ Laitdoii CRINOIDEA f Discovery Reports. Vol. XVIII, pp. 223-238, January 1939.] THORACIC CIRRIPEDES COLLECTED IN 1925-1936 By C. A. NILSSON-CANTELL, SWEDEN THORACIC CIRRIPEDES COLLECTED IN 1925-1936 By C. A. Nilsson-Cantell, Sweden (Text-figs. 1-5) INTRODUCTION IN an earlier paper (Nilsson-Cantell, 1930) I have worked out the Cirripedes brought home by the R.R.S. ' Discovery ', the R.R.S. ' William Scoresby ', and the staff of the Marine Biological Station at South Georgia during the years 1925-7. For the oppor- tunity to study further material from the same expeditions I have to express my best thanks to the former director of the expedition, Dr Stanley Kemp. The greater part of the material was collected during the years 1928-36, but some specimens from the earlier expeditions are also included here, since they were not at my disposal earlier. Stations made by the ' Discovery ' are entered first in the list of localities under the species headings and have no letters prefixed to the numbers. Those of the ' William Scoresby ' follow and are distinguished as WS 80, etc. Where the material was collected from whales, the serial number (No. 87, etc.) assigned to it by the Discovery observers is given. A number of the specimens were taken from whales for which such data are not available. The following symbols are used for nets, apparatus, etc., in accordance with Discovery Reports, i, Station List: BNR A 100 cm. net attached to a frame 1-47 m. long and 38 cm. wide with skids on either side which keep the mouth of the net 15 cm. above the sea bottom. BTS Small beam trawl. Beam 8 ft. in length (2-45 m.): mesh at cod-end I in. (12-5 mm.). All measurements are taken from knot to knot along one side of the mesh, not diagonally with the mesh stretched. DC Conical dredge. Mouth 16 in. in diameter (40-5 cm.), with canvas bag. DLH Large dredge. Heavy pattern, 4 ft. in length (1-2 m.). DRR A Russell net frame with rectangular dredge attached instead of a tow-net. N 100 I m. tow-net. Mouth circular, i m. in diameter: mesh graded, with the cod-end made of stramin, with 11-12 meshes to the linear inch. ,^^ ^ rJr Nets with mesh of 4 mm. or 7 mm. (o-i6 or 0-28 in.) attached to the back of trawl. N 7-Tj N-T A fine meshed net attached to the back of the otter trawl. NCS-T Tow-net of coarse silk, with 16 meshes to the linear inch, attached to a trawl. NH Hand net. NR Rectangular net. OTC Commercial otter trawl. Head rope 80 ft. (24-5 mm.): mesh at cod-end i\ in. (3-8 cm.). OTL Large otter trawl. Head rope 40 ft. long (12-2 m.): mesh at cod-end i^ in. (3-2 cm.). RM Mussel rake. 226 DISCOVERY REPORTS The abbreviations used in denoting the nature of the bottom are : b. blue d. dark gy- grey r. rock sh. shells br. brown f. fine m. mud s. sand sk. specks c. coarse gn. green p. pebbles sft. soft St. stones The depths of the water at the beginning and at the end of trawhng stations are shown, as, e.g., St. 144, 155-178 m. GENERAL The number of species in this collection is shown in the following table : Genus Total number of species New species Species not repre- sented in my earlier Discovery Report Scalpellum Lepas Conchodenna Balanus Coronula 8 I 2 3 2 2 4 Total number 16 2 4 This table shows that most of the species are represented in my first paper dealing with Cirripedes from the Discovery expedition (Nilsson-Cantell, 1930). Most species belong also in this collection to the genus Scalpellum, only a few of which are known to inhabit deeper water. The identification of young individuals of this genus is always rather questionable, because many species are described from small specimens. It will be seen in this paper, as in the former, that the shape of the valves of the capitulum are rather different in individuals of diff^erent age. Bathymetrical distribution Name of species Depth at which taken Lepas australis Floating Conchoderma auritutn „ virgatum i. hunteri Coronula diadema reginae Balanus algicola Scalpellum gibberum 0-253 m. Balanus laevis 12-78 m. „ maxillaris 27-64 m. Scalpellum bouveti n.sp. 40-45 m. ,, rathbunae 99-107 m. ,, convexum 120-383 m. „ scoresbyi n.sp. 146-242 m. ,, discoveryi 155-342 m. „ angulare 342 m. „ compactum 567 m. THORACIC CIRRIPEDES COLLECTED IN 1925-1936 227 The first five specimens in this table are typical pelagic forms. One of the Balanus species, B. algicola, is here found fixed to a fish from the surface. It was at first de- scribed as growing on algae, but later on I have found it fixed to the shells of mussels from the littoral. The two remaining Balanus species are typical littoral forms. As regards the bathymetrical distribution of the Scalpelliim species, it may be mentioned that Scalpellum compactum and Scalpellum convexiim are here recorded from deeper water than previously. Since our knowledge is incomplete, it is not possible to draw any conclusions as to the bathymetrical distribution of the Scalpellum species here dealt with. As regards the distribution of Cirripedes this material gives some results beyond those arrived at in the former paper dealing with the Discovery material (Nilsson-Cantell, 1930). Thus Scalpellum convexum is represented not only from South Georgia (where it was previously known to occur) but also from other parts of the Atlantic Ocean, viz. the South Sandwich and South Shetland Islands. Scalpellum rathbimae is now known from several localities east of Argentina from La Plata down to Patagonia. This characteristic species is likely to be found later on in new localities along the east coast of South America. Scalpellum compactum seems to be known from the southern part of the Pacific and Atlantic Oceans, the Ross Sea and the Graham Land region. The specimens in this collection are taken from a locality situated near the type-locality : McMurdo Sound. Scalpellum discoveryi, earlier noted from the Pacific, is, according to this collection also found in the South Atlantic, thus indicating a circumpolar distribution. Two species of Scalpellum are new : S. scoresbyi and S. bouveti, the former from the waters round the Falkland Islands, the latter east of Bouvet Island. Lepas australis, a typical pelagic species, is here represented from a locality (St. WS 237) for which the depth is given as 150-256 m. The specimens have certainly been caught by the net near the surface. Balanus maxillaris, earlier only known from South African waters, is here recorded from the southern part of the Indian Ocean. Nothing of special interest need be added here regarding the distribution of the other species in the collection. SYSTEMATIC Genus Scalpellum Leach, 1817 Scalpellum gibberum C. W. S. Aurivillius, 1892. For synonymy see Nilsson-Cantell, 1921, Zool. Bidrag Uppsala, vii, p. 178. St. 53. 12. V. 26. Port Stanley, East Falkland Island. Hulk of 'Great Britain'. 0-2 m. GearRM. One small specimen. St. 652. 14. ill. 31. Burdwood Bank, 54° 04' 00" S, 61° 40' 00" W. 171-169 m. Gear OTL, Several specimens on Hydroids. 228 DISCOVERY REPORTS St. WS8o. 14. iii. 27. 50° 57' 00" S, 63° 37' 30" W. 152-156 m., f.d.s. Gear OTG. Some small specimens on coral. St. WS 85. 25. iii. 27. 8 miles S 66° E of Lively Island, East Falkland Island. 79 m. Gear OTG. Some small specimens on Hydroids. St. WS225. 9. vi. 28. 50° 20' 00" S, 62° 30' 00" W. 162-161 m., gn.s. sh. p. Gear OTG. Several specimens from Hydroids, several specimens from carapace of the crab Euryfodius sp. St. WS231. 4. vii. 28. 50° 10' 00" S, 58° 42' 00" W. 167-159 m., f.gn.s. Gear N 4-T. One specimen. St. WS 243. 17. vii. 28. 51° 06' 00" S, 64° 30' 00" W. 144-141 m., c.d.s. Gear N 4-T. Some specimens on Hydroids. St. 244. 18. vii. 28. 52° 00' 00" S, 62° 40' 00" W. 253-247 m., f.d.s. m. Gear N 7-T. Some specimens on Hydroids. St. WS761. 13. X. 31. 44° 22' 00" S, 63° 02' 00" W. 97 (-0) m., gn.s. m. GearBNR. St. WS782. 4. xii. 31. 50° 29' 00" S, 58° 23' 45" W. i4i-i46m., gn.s. Gear OTG. One specimen. St. WS 787. 7. xii. 31. 48° 44' 00" S, 65° 24' 30" W. 106-110 m., c.br.sk. s. Gear OTG. Two small specimens on algae. St. WS811. 10. i. 32. 5i°27'45"S, 68°oi' 30" W. 98m.,s. st. GearOTG. One large specimen. Distribution. Pacific Ocean (according to Aurivillius uncertain), Atlantic Ocean south of La Plata, Patagonia, Magellan Strait, round the Falkland Islands, where the species seems to be richly represented, 0-253 ™- Discussion. To the supplementary description of ScalpeUum gibberum given by Nilsson-Cantell, 1930, in Discovery Reports, vol. 11, nothing need here be added. The localities given above may complete my earlier paper from this expedition. ScalpeUum angulare Nilsson-Cantell, 1930. Nilsson-Cantell, 1930, Discovery Reports, n, p. 239. St. 170. 23. ii. 27. Off Cape Bowles, Clarence Islands. 61° 25' 30" S, 53° 46' 00" W. 342 m., r. Gear DLH. Four specimens on Colossendeis. This specimen is taken from the same locality as the holotype. ScalpeUum convexum Nilsson-Cantell, 1921. Nilsson-Cantell, 1921, Zool. Bidrag Uppsala, vn, p. 194. Nilsson-Cantell, 1930, Discovery Reports, 11, p. 244. St. 42. I. iv. 26. Off mouth of Cumberland Bay, South Georgia. From 6-3 miles N 89° E of Jason Light to 4 miles N 39° E of Jason Light. 120-204 m., m. Gear OTL. Some specimens from Hydroids. St. 140. 23. xii. 26. Stromness Harbour to Larsen Point, South Georgia. 136-122 m., gn.m. st. Gear OTL. One specimen on Hydroids. St. 371. 14. iii. 30. I mile E of Montagu Island, South Sandwich Islands. 99-161 m. Gear OTL. Some specimens on Hydroids. St. 1952. II. i. 37. King George Island, South Shetland Islands. 367-383 m., sft.m. Gear DRR. One specimen on a polychaet tube. Distribution. South Georgia, South Sandwich and South Shetland Islands, 110-383 m. THORACIC CIRRIPEDES COLLECTED IN 1925-1936 229 Discussion. This species is here taken again in four locaUties. As I was previously able (Nilsson-Cantell, 1930) to study individuals younger and older than the type specimens, it was possible to state that the shapes of the valves of the capitulum are rather different in individuals of different age. The specimen here illustrated from St. 371 must be intermediate between a specimen figured by the author in 1930, text-fig. 9 d, and the type specimen (Nilsson-Cantell, 1921, text- fig. 29 //). The shape of the carinal latus especially is different in individuals of different age. There are certainly many doubtful species of Scalpellum described from young individuals. Thus, for instance, the specimen from St. 371 (Fig. i) shows a superficial resemblance to Scalpellum dubiiitn Hoek, 1883. The mouth parts and caudal appendages are here figured for this in- dividual to show the resemblance to the same parts of the type. Scalpellum rathbunae Pilsbry, 1907. Pilsbry, 1907, Bull. U.S. Nat. Mus. No. 60, p. 40. Nilsson-Cantell, 1921, Zool. Bidrag Uppsala, vn, p. 184. St. WS776. 3.xi.3i. 46°i8'i5"S,65°02'i5"W. 110-9901., gn.m. s. Gear OTG. Two specimens from Chaetopterus tubes. Fig. I. Scalpellum convexiim {Si. 271). «. Mandible. b. Maxilla I. c. Maxilla IL d. Cirrus VI and caudal appendage, e. Female{total length 9-4mm.), lateral view. /. Rostrum and rostral latera. Distribution. Off the east coast of Pata- gonia. The distribution of this collection is more westerly than Pilsbry 's. Discussion. This well-defined species, earlier mentioned by Pilsbry (1907), and Nilsson-Cantell (1921), is here represented by one small and one full-grown specimen, the later one characterized by its long peduncle (comp. Nilsson-Cantell, 1921). The peduncle is about twice as long as the capitulum. To the description given earlier nothing need be added here. Scalpellum compactum Borradaile, 1916. Borradaile, 191 6, Brit. Mus. {Nat. Hist.), Brit. Antarct. (' Terra Nova') Exp. 19 10, Nat. Hist. Rep. Zool. Ill, No. 4. London. Nilsson-Cantell, 1921, Zool. Bidrag Uppsala, vn, p. 198. St. 1652. 23. i. 36. 75° 56' 00" S, 178° 35-5' 00" W. 567 m. Gear DRR. Some specimens on Hydroids. Distribution. Graham Land region, Ross Sea, McMurdo Sound, 91-567 m. 230 DISCOVERY REPORTS Scalpellum discovery! Gruvel, 1906. Gruvel, 1906, Bull. Mus. Hist. Nat., t. 12, Paris, p. 271. Gruvel, 1907, Nat. Antarct. Exped. 1901-1904 {'Discovery'), Nat. Hist, iii, Crust, vi, p. 2, London. Borradaile, 1916, Brit. Mus. {Nat. Hist.) Brit. Antarct. {'Terra Nova') Exp. 1910, Nat. Hist. Rep. Zool. Ill, No. 4, p. 128, London. St. 170. 23. ii. 27. Off Cape Bowles, Clarence Island. 61° 25' 30" S, 53° 46' 00" W. 342 m., r. Gear DLH. One specimen on Colossendeis. St. 144. 5. i. 27. Off mouth of Stromness Harbour, South Georgia. From 54° 04' S, 36° 27' W to 53° 58' S, 36° 26' W. 155-178 m., gn.m. s. Gear NGS-T. One young specimen. Distribution. McMurdo Sound, east of Solomon Islands (7° 56' 00" S, 164° 12' 00" E), Clarence Island, South Georgia. 61-342 m. Discussion. The two specimens dealt with here agree well with the specimen of Scalpellum discoveryi Gruvel, 1906, figured by Borradaile, 1916. His specimen and the one illustrated here (Fig. 2) are of about the same size and must be considered as young specimens. It is of great value to illus- trate specimens of different age for further investigators. Larger specimens of the same species as figured by Gruvel, 1907, are quite different, owing to imperfectly calcified plates on the capitulum. I have found similar differences before (Nilsson- Cantell, 1930) as regards the species Scalpellum convexum. Internal parts are almost totally un- known. Of the mouth parts only palpus and mandible could be traced (Fig. 2). Fig. 2. Scalpellum discoveryi Gruvel, 1906 (St. 170). The palpus is typically conical. The - Pf P-- J^- Mandible .. Specimen, lateral view t^ ^ Jt" J _ (total length 9-5 mm.). » — Sir G. Grey (damaged) 3252 f. — 406 Lord Howe Island — J. MacGillivrav, H.M.S. 'Herald' 3257 — — — New Zealand — Wellington Museum 325^ — — — Kerguelen — ' Transit of Venus ' Expedition 325/ — — — Falkland Islands — Coppinger, H.M.S. 'Alert' 325 w m. — 313 ») — Holmstead 325 « f. — 402 „ — *) 325/' — — — No data — Purchased 93.9.14.1 f. — 398 Antarctic Seas — Captain D. Gray 1.1.4.1S m. — 367 Ross Quadrant — ' Southern Cross ' Expedition 8.2.20.54 m. 269 371 »» i. 04 ' Discovery ' Expedition 8.2.20.56 f. 333 — »» i. 04 »» 14.1.29.1 — — 313 »» — ' Terra Nova ' 14. 1. 29. 2 m. — 354 ,, — ,, 38.12.3.2 f. — 392 Weddell Sea 19. iii. 22 G. H. Wilkins 34. 1 2.4. 1 m. — 394 South Georgia — Cmdr. J. Chaplin Royal College of Sur geons 1090 m. 338 Tasmania No data 1091 — — 345 South Georgia Kearns, type 1092 f. — 422 Falkland Islands — No data 1093 — — 338 ,, — )) 1094 — — 336 New Zealand — »» 1095 f. — 383 — — yy 1095-3 m. — 386 Falkland Islands — Mrs Holgate 1095.4 — — »» — R. Vallentin (fragments) 1096 — — 321 — — Chevalier A. 175.1 m. — 405 South Georgia — Sir J. Middleton A. 175.11 f. — 406 " — " British Graham Land Exped ition, 1934- 6 B. 14 f. 345 4°3' 30- >v- 35" B.15 f. — 406 3- iv. 35 B. 16 f. 274 385 5- VI. 35 B. 17 f. — 425 22. ui. 35 B.55 f. 343 — 23- "i- 35 B. 81 f. 251 379 . Islands of coast of 22. xi. 35 ^ G. C. L. Bertram B. 126 f. 358 431 Graham Land 17. in. 36 B. 127 f. 420 28. iii. 36 B. 128 f. 292 392 21. iii. 36 B. 142 f. 292 397 27- i- 37 B. 144 f. 305 387 17- "• 37 B.I4S f. 300 417. 9. iii. 37 B. 146 m. 272 386 South Georgia 10. X. 36 B. B. Roberts 1 THE LEOPARD SEAL 245 Table I (contd.) Ref. no. Sex Body length Skull length Locality Date Collector South Shetland, 1921 Sh. ix m. — 398] February Sh. X f. — 405 Sh. xvi m. — 364 Pack ice, South J. E. Hamilton Sh. xviii f. 395 Sandwich region Sh. xix f. 351 403 Sh. XX f. 402' I Discovery Committee S.S. I f. 305 39n 18. i. 28 S.S. 2 m. 273 377 22. i. 28 S.S. 3 S.S. 4 m. m. 270 294 384 416 Pack ice. South 3. ii. 28 6. ii. 28 J. E. Hamilton S.S. 5 f. 221 310 Sandwich region 21. ii. 28 S.S. 6 f. 252 373 21. ii. 28 S.S. 7 m. 248 sse-l 21. ii. 28 W.S. 574 f. — 408" i-3i (presented) W.S. 651 m. 236 340 22. ix. 36 W.S. 653 m. 221 324 22. ix. 36 W.S. 654 m. 284 385 22. ix. 36 W.S. 655 m. 297 396 3-x. 36 W.S. 656 f. 284 378 3-x. 36 W.S. 657 m. 244 357 6. X. 36 W.S. 661 f. 226 310 9.x. 36 W.S. 662 m. 297 391 ■ Falkland Islands 23. X. 36 J. E. Hamilton W.S. 663 f. 250 339 iS.x. 36 W.S. 664 m. 282 380 22. X. 36 W.S. 665 m. 292 38s 22. X. 36 W.S. 666 m. 288 385 22. X. 36 W.S. 673 f. — 377 1936 (presented) W.S. 679 m. 403 1937 (presented) D.L6I6 f. 411 30- X. 34 (found dead) D.LII62 m. 284 384J 8.x. 31 - Note. The initials "W.S." and "D.L" only indicate different series of bone labels and have no other significance in the present connexion. They are therefore omitted in textual references. dimensions are about 100 cm. long and 50 cm. wide at the broadest part, just behind the flippers. The back is rather pale grey with a broad darker stripe down the middle line; the head is grey dorsally but with pale marks round the eyes; the lips and chin are white. The flanks and belly are almost white, sharply divided off from the grey of the back, but there is an irregular and asymmetrical sprinkling of dark spots and streaks on the flanks; the vibrissae are almost black. The colour distribution and markings are practically the same as in the adult, but the second coat, which can be seen already in the middle of the back, is darker than the first (Plate VII). 246 DISCOVERY REPORTS BODY FORM AND SIZE The immature seal of about a year old is very attenuated and almost cylindrical in shape, and with increasing age the principal change is in the development of the thorax and shoulders. This development is considerably more marked in the female than in the male which retains a form reminiscent of the juvenile throughout life, whereas in the large females the thoracic development is very conspicuous indeed. It may be added that the very large head combined with a thin body in youth, or the thoracic development of older animals, greatly facilitates the identification of this species at a distance, since its outline is so different from all the other Antarctic species. It was first observed by Bruce, during the voyage of the ' Balaena ', that the female leopard seal is larger than the male, and Barrett-Hamilton (1902) comments on this. Bruce's statement is fully confirmed by the measurements now available. The length from the tip of the nose to the tip of the tail is known for eighteen males and ranges from 221 to 305 cm., and of these only one, that is 5-6 per cent, exceeds 300 cm. Twenty females range from 221 to 358 cm. and ten of them, 50-0 per cent, exceed 300 cm. At birth the length is about 120 cm. (Matthews, 1929), and growth is evidently very rapid during the first six months; Valette (1906) secured a male and female of small size on 29 December and they were already 189 and 194 cm. respectively, and S.S. 5 of 21 February is 221 cm. During the second half-year of life the rate of growth appears to be less, since the average of four, of both sexes, killed at the age of about a year is only 233 cm. The averages for body and skull length for the four age groups distinguishable are shown below (Table II). The correlation between body and skull length in large animals is not close. For example, the body length of B. 144 is 305 cm. and the skull length is 387 mm., whereas B. 145 has a body length of 300 cm. but a skull of 417 mm., and B. 16 with a body of 274 cm. has a skull of 385 mm. Table II Body length (cm.) Skull length (mm.) Males Females Males Females First year Second year Third year Fourth year and over 228-5 2537 280-8 289-4 232-3 279-3 303-7 328-8 332 361-5 381-5 394-4 332 3877 402-9 411-5 It is appropriate to emphasize here that since the leopard seal is an animal of solitary habit it has to be collected as opportunity offers ; there can be no question of picking the largest specimen, and the series of the two sexes here considered may therefore be properly taken as representative. (For discussion of age grouping see pp. 248, 250-256.) Weight. Ross's record of his heaviest specimen, 850 lb., has already been noticed (p. 241). Apart from this the only definite information is that contained in the ' Scotia ' THE LEOPARD SEAL 247 Reports (1915), where the weights of two males are given — no. XLII, 305 cm., 606 lb. and no. XLV, about 4 cm. longer, 630 lb. The ' Scotia' seals were apparently weighed piecemeal, and for no. XLII only 3 lb. were allowed for loss of blood which seems very little indeed : no details of weights are given with reference to no. XLV. DEVELOPMENTAL STAGES OF THE SKULL The skull of Hydriirga is remarkable for its length, that of the male being slightly less attenuated than that of the female, though the latter has wider zygomatic arches. The female skull also attains to a greater size and more massive development than the male, and in the majority of specimens distinctions are visible which, although they may appear small when represented by measurements, nevertheless produce substantially different forms. In order to facilitate a comparison of the differences between the sexes two series were selected in which the condylo-basal lengths of one sex were matched as closely as possible in the other. There are fifteen of each. Table III. Cotidylo-basal lengths of skulls Males Females Ref. no. Length (mm.) Ref. no. Length (mm.) 651 340 661 339 325^ 346 663 347 8.20.2.54 371 S.S. 6 373 S.S. 22 377 656 378 664 380 B. 81 379 654 385 B. 16 385 665 385 B. 144 387 B. 146 386 325^ 389 666 389 S.S.I 391 662 391 38.12.3.2 392 34.12.4.1 394 B. 128 392 655. 396 Sh. xviii 395 Sh. ix 398 B. 142 397 679 403 B. 14 403 S.S. 4 416 B. 145 417 In dorsal view the frontal bones are greatly compressed from the point at which they leave the cranium to form the interorbital bar, but they expand once more to join the maxillae; the transition from the cranial to the orbital part is gradual in the male but rather abrupt in the female, and the bar itself is narrower in the latter. The maxillary expansion is wider in the male, but the preorbital region of that sex tapers rather more rapidly, so that the male has a more pointed snout than the female. In ventral aspect the palate of the female is more rounded anteriorly and less wedge-shaped than that of the male, and the walls of the nasopharynx are closer together. These walls are composed of the vertical parts of the palatines and pterygoids, and in life support the soft palate which in the female therefore is narrower than in the male. 248 DISCOVERY REPORTS The width of the nasopharynx was measured at the point where the pterygo-palatine sutures most nearly approximate ; such measurements give an indication of the width of the nasopharynx but naturally fail to show that in the female the ventral edge of the pterygoid turns inwards towards the middle line, and, as one might say, rolls round on itself, as age increases, whereas in the male the reverse happens, since the edges turn outwards. The width of the soft palate thus decreases in the female but increases in the male, relatively as well as absolutely, with advancing age. These characters are subject to a considerable degree of individual variation, but in spite of this it is possible to sex well-grown skulls. With juvenile specimens, however, it is difficult and often impossible, at any rate in the present state of knowledge. In very young skulls there is the roundness common to all young animals, and the sexual characters are very naturally not developed. The smallest male, no. 653, an animal about a year old, shows a proportionately normal interorbital measurement but the zygomatic, anterior frontal, maxillary and premaxillary measurements are all rather high compared with the length of the skull (Table IV). In the juveniles of both sexes the pterygoids are turned outwards, and in the females in proportion to the length of the skull the interorbital bar is widest in the youngest specimens. As in body length so also in skull development the female departs farther from the juvenile condition than the male; the former has typically larger crests than the latter (this applies particularly to the sagittal crest), a more conspicuous narrowing of the interorbital bar, and more massively developed jaws which produce a blunter muzzle, and, finally, it is only in the large females that there is complete fusion of the palatine sutures and of much of the nasal (vertical) wing of the premaxilla to the adjacent part of the maxilla. Even in the largest males fusion of the premaxillae to the maxillae only takes place in the alveolar region. The osteological changes due to increasing age are analogous in the two sexes, except that, as stated above, they progress farther in the female. The specimens of each sex are divisible into four groups on grounds of size and osteological characters, and since the former shows a progressive increase and the latter a progressive ossification and development of crests, etc., it is inevitable that the four stages represent stages in the development of the animal. Since it is established that this species has a definite and limited pupping season it is reasonable to conclude that all the seals of a given year will show common characters at any age up to sexual maturity at least, and such definite groups as those mentioned above must be taken to represent the first three years of life ; but in the absence of a much larger series than is at present available it is not possible to subdivide the fourth, oldest, group which therefore represents the fourth year and over. The following measurements were taken on each skull (see Figs, i and 2): (i) Total length, occipital condyle to tip of premaxillae. (2) Maximum zygomatic width (outside). (3) Minimum width of frontal bones (interorbital bar). (4) Maximum anterior width of frontals. THE LEOPARD SEAL 249 (5) Width of maxillae at level of upper canine teeth. (6) Width of the anterior margin of the premaxillae. (7) Width of nasopharynx near ventral end of the pterygo-palatine suture. (8) Alveolar length from the point of bone between the upper canine and the first cheek tooth to the posterior rim of the alveolus of the fifth cheek tooth, usually on the right side. IDom. Fig. 10 cm Fig. 2. Fig. I. Dorsal view of skull, the numbers refer to the numbered measurements in the text. Fig. 2. Nasopharyngeal area of skull, the number refers to the measurement bearing the same number. MALE SKULLS (Plates VIII, IX and X) The twenty-eight male skulls vary in length from 313 to 416 mm., and include repre- sentatives of the stages from the first year to full growth and maturity. Ossification never reaches the advanced condition found in some Pinnipedia where the skull may become an almost solid mass of bone. It is no more developed in S.S. 4 of 416 mm. than it is in 662 which is 25 mm. less. The observable fusions are those of the cranium, in the region of the alveolar parts of the maxillae and premaxillae where they are in contact, and in the maxillo-palatine sutures, but the last are fused in only two male skulls and both of them of the largest size. 250 DISCOVERY REPORTS With increasing age there is a general increase in smoothness and density of bone, most noticeable perhaps in that of the cranium. In the youngest skulls the lambdoidal suture is only partly fused, but with increasing age it is obliterated, as are the following sutures in the order given: sagittal, squamosal, coronal and occipito-sphenoidal. There are developed very large and thick lambdoidal crests and a less conspicuous sagittal crest which diminishes anteriorly and divides into two low, bilaterally symmetrical ridges marking the temporal lines. As has been stated on p. 248, four age groups may be distinguished in each sex, and the last undoubtedly comprehends all seals of the fourth year and over. In the following descriptions average measurements are given unless otherwise specified, and figures in brackets indicate the number of specimens on which was based the immediately succeeding figure : First year (5). Body length (2) 228-5 cm. ; skull length 332 mm. ; zygomatic width 172-6 mm.; alveolar length (4) 90-4 mm. The cranium is rounded and the anterior fonticulus may not be quite closed. The lateral parts of the lambdoidal suture can be seen and the occipito-sphenoidal suture is open; the sagittal suture is rather complex, the coronal sutures are unfused and the squamosal bone is not fused to any other except the periotic and the tympanic. The cheek teeth are crowded, the first in the upper jaw almost touching the canine. Second year (6). Body length (3) 253-7 cm.; ^^^^^ length 361-5 mm.; zygomatic width 187-6 mm.; alveolar length 96 mm. The lambdoidal crests are beginning to develop and signs of the sagittal crest give the dorsal profile a straighter line. The lambdoidal suture is obliterated and the sagittal fused, as is also a considerable length of the dorsal part of the coronal sutures. The squamosal suture is fused and the occipito- sphenoidal may be partly closed but is still conspicuous. The cheek teeth are better spaced. Third year (4). Body length 280-8 cm. ; skull length 381-5 mm.; zygomatic width 197 mm.; alveolar length 98-8 mm. The cranium is still smoother and better ossified, the crests are larger and the lambdoidal crest developed to such an extent that the dorsal profile of the skull may be almost saddle-shaped. The occipito-sphenoidal suture is partly fused and there may be fusion of the maxillae and premaxillae on their palatine faces. Fourth year and over (13). Body length (8) 289-4 cm.; skull length (12) 394-4 mm.; zygomatic width 207-3 mm.; alveolar length (12) 101-9 mm. This final group contains all the largest and most highly ossified skulls; the crests are invariably large and the occipito-sphenoidal suture not only fused but usually obliterated. The premaxillae and maxillae are fused in the alveolar part as well as the palatal. THE LEOPARD SEAL 251 Table IV. Measurements arranged in age groups {individuals in order of actual skull length) Males Body length Skull Zygo- Inter- Anterior Upper Pre- Width of Alveolar Ref. no. length matic width orbital width frontal width canine width maxillar width naso- length cm. mm. pharynx mm. mm. mm. mm. mm. mm. mm. First year 1 325»! — 313 169 38 — 6i-5 36 38 86 W.S. 653 221 324 171 37-5 76 65 39 36-5 — R.C.S. 1090 — 338 173 35 62-5 6i-5 38-5 34 95-5 W.S. 651 236 340 175 37 70 6i-5 36-5 39 91 325^ — 346 175 38 70 60 36 39 89 Second year 14.1.29.2 — 354 188 37 70-5 61 36-5 41 95 S.S. 7 248 356 193 43-5 78 69 43 40 90-5 W.S. 657 244 357 182 39 73 63 37-5 38 93 Sh. xvi — 364 183 46-5 8i-5 65 43 36-5 93 1.1.4.15 — 367 189 42-5 77-5 65 4' 48 95 8.2.20.54 269 371 191 41-5 73 68-5 41-5 47-5 100 Third year S.S. 2 • 273 377 196 42-5 75 69-5 43 47 98 W.S. 664 282 380 193 42 79 71 38 50 98 D.I. 1162 284 384 197 43-5 83 66-5 41-5 49-5 lOI W.S. 654 284 385 202 41 84 68 44-5 46-5 98 Fourth year and over S.S. 3 270 384 206 43 84 71-5 45 52 98 W.S. 665 292 385 204 43-5 82 73-5 45 58 104 B. 146 272 386 208 45 84 69 43 — 96-5 R.C.S. 10915.3 — 386 210 41-5 89 74-5 45-5 52-5 102-5 W.S. 666 288 389 191 45 80 71 44 57 100 W.S. 662 297 391 214 44 92 71 42 46-5 94 34.12.4.1 — 394 212 45 87 75-5 46 — lOI W.S. 655 297 396 209 44 79 73 44 51 104 Sh. ix — 398 215 46 87-5 74-5 48-5 49-5 i°5 W.S. 679 — 403 205 42-5 85 69 46 56 102 A. 175.1 — 405 199 40-5 83-5 73 44-5 52 109 S.S. 4 294 416 214 42 88-5 77-5 48 54-9 107 Additional lengths of males, not referr ed to skul s XLII 305 From Scientific Results of the ' Scotia ', vol. IV 231 G. C. L. Bertram, MS. (not used in tion of average length) calcula- 252 DISCOVERY REPORTS FEMALE SKULLS (Plates VIII, IX and X) The identifiable female skulls are thirty-two in number, and like those of the males display four age groups. On the whole the female displays rather greater individual variation than the male. First year (3). Body length 232-3 cm.; skull length 332 mm.; zygomatic width 174 mm.; alveolar length 89-8 mm. At this stage the skulls closely resemble those of 320- / / / / / / / 300- / / / / ^^ / ^^-^ / ^^-^ E2ao- f / u 1 / c 1 / 1 / _c 1 / "^260- 1 / 1 / QJ _J 1 / 3- 1 / T) 1 / 0 cG 240- 1 / 1 / 1 / 220- Averages of Ag'e Groups BodLj Length S 9 ZQQ- I n m EZ 420 400 380- c QJ 360 340 320 Years of Ag"e Fig- 3. 1/ Averages of Age Groups Skull Length <5_ 9- — I n I m Years of Ag"e Fig- 4- ]¥ Fig. 3. Graphical representation of growth stages of body, both sexes. Fig. 4. Graphical representation of growth stages of skull, both sexes. the males and exhibit the rounded form, comparatively poor ossification and conspicuous sutures of the immature animal. The vertical plates of the pterygoids are frequently inclined outwards — a juvenile character which is found in the male skull throughout life — and the occipito-sphenoidal suture may begin to fuse. Cheek teeth crowded. Second year (11). Body length (5) 279-3 cm.; skull length 387-7 mm. ; zygomatic width 199-3 mm.; alveolar length 101-5 mm. The bone is denser, the coronal and squamosal sutures are almost entirely fused, the occipito-sphenoidal is usually fused and may be obliterated, and the alveolar length has increased. THE LEOPARD SEAL 253 Table V. Proportions arranged in age groups Males Percentages of skull length Ref. no. Body length Skull length as % of body Zygo- matic Inter- orbital Anterior frontal Upper canine Pre- maxillar Width of Alveolar length length width width width width width naso- pharynx First year 22Sm — — S4-0 I2-I — 19-6 II-5 12-14 27-5 W.S. 653 221 147 52-8 II-8 23-5 20-1 12-0 II-3 — R.C.S. 1090 — —- 51-2 10-4 i8-s l8-2 II-4 lOI 28-3 W.S. 651 236 14-4 51-5 10-9 20-6 i8-i 107 II-5 268 325^ — — 50-6 II-O 20-2 17-3 IO-4 11-3 257 Second year 14.1.29.2 — — 53-1 10-5 199 17-2 10-3 II-6 26-8 S.S. 7 248 14-4 54-2 12-2 219 19-4 I2-I 11-2 25-4 W.S. 657 244 14-6 51-0 10-9 20-6 i8-i 107 II-5 26-1 Sh. xvi — — 5°'3 12-8 22-5 17-9 II-8 lo-o 25-5 1.1.4.15 — — 51-5 II-6 21-2 177 II-2 13-1 25-9 8.2.20.54 269 13-8 51-5 II-2 197 1 8-5 II-2 12-8 27-0 Third year S.S. 2 273 13-8 52-0 11-3 19-9 i8-4 II-4 12-5 26-0 W.S. 664 282 13-5 50-8 III 20-8 187 lo-o 13-2 25-8 D.I. 1 162 284 13-5 51-3 II-3 21-6 17-3 10-8 12-9 26-3 W.S. 654 284 13-6 52-5 10-6 21-8 177 II-4 I2-I 25-5 Fourth year S.S. 3 270 14-2 53-6 II-2 21-9 186 117 13-5 25-5 W.S. 665 292 13-2 53-0 "•3 21-3 191 117 is-i 27-0 B. 146 272 14-2 53-9 117 21-8 17-9 III — 25-0 R.C.S. 1095.3 — — 54-4 10-8 23-1 19-3 II-8 13-6 26-6 W.S. 666 288 13-5 49-1 II-6 20-6 i8-3 II-3 14-7 257 W.S. 662 297 13'2 547 II-3 23-5 l8-2 107 II-9 24-0 34.12.4.1 — — 53-8 II-4 22-1 19-2 117 — 25-6 W.S. 655 297 13-3 52-8 III 19-9 18-4 ii-i 12-9 26-3 Sh. ix — — 54-0 II-6 22-0 187 12-2 12-4 26-4 W.S. 679 — — 50-9 10-5 2I-I 17-1 1 1-4 13-9 25-3 A. 175.1 — — 49-1 lO-O 20-6 i8-o II-O 12-8 26-9 S.S. 4 294 14-1 51-4 lO-I 21-3 i8-6 II-5 13-2 257 254 DISCOVERY REPORTS Table VI. Measurements arranged in age groups Females Ref. no. Body length cm. Skull length mm. Zygo- matic width mm. Inter- orbital width mm. Anterior frontal width mm. Upper canine width mm. Pre- maxillar width mm. Width of naso- pharynx mm. Alveolar length mm. First year S.S. 5 W.S. 66i W.S. 663 221 226 250 310 339 347 178 173 171 41 45 40 65 70 70-5 63 64 63 41 40-5 39 39 42-5 42 83-5 91 95 Second year S.S. 6 W.S. 67^ W.S. 656 B. 81 R.C.S. 1095 252 284 251 373 377 378 379 383 204-5 208 200 187 i92'5 41-5 37-5 40 38-5 37-5 78 81 80 70-5 65 73-5 73-5 69 68-5 68-5 45 415 41 41 42 52 41-5 48 47 41-5 100 lOI 101-5 99 98 B. 144 305 387 202-5 44 86 70 42 47 ICO S.S. I 305 391 200 37 73-5 70 41-5 45-5 lOI 38.12.3.2 — 392 208 42-5 80 74-5 44 49 103 93.9. 14. 1 — 398 191 39 79 74 41 49 103 Sh. XX Sh.x — 402 405 191 208 41 44 74 85 70 76-5 47 52 43-5 49 108-5 IOI-5 Third year B. 16 325* 274 38s 389 205 210-5 41-5 38 87 73-5 73 49-5 45-5 43 46-5 97 99 B. 128 292 392 218 42-5 85 78 49-5 — 105 325" — 402 223 42-5 92 76 50 42 102 B. 14 B.iS 34S 4°3 406 207 206 42 43-5 91 84 79 52 46-5 47 45 104 log 325« — 406 221 42-5 78-5 82 — 43-5 103 325'^ — 410 210 42 92 77-5 45-5 49 108-5 D.I. 616 — 411 202 39 80 72 47 — 102-5 B.17 — 425 232 43 92-5 79 49-5 108 Fourth year and over Sh. xviii — 395 201 36-5 70 71-5 41-5 46 100 325/' — — 220 — — — — — — B. 142 292 397 241 40 80 78 46 48 105 Sh. xix W.S. 574 351 403 408 200 230 40 40 79 92 72-5 77-5 46-5 50-5 51 46-5 100-5 103 B.55 343 — 234 44 87-5 85 52-5 52-5 no B. 14s 300 417 252 39 92 77 50-5 34-5 99 B. 127 — 420 247 39-5 91 82 50 45 108 R.C.S. 1092 B. 126 358 421 431 253 248 42-5 41-5 100 lOI 86 90 55 52 41-5 43 112-5 109 Additional lengths of females, not refer red to sku Is III 350 226 From Scientific Results of the ' Scotia ' G. C. L. Bertram, MS. (not use d in calcL lation of average — 315 length ) — 338 - THE LEOPARD SEAL Table VII. Proportions arranged in age groups Females Ref. no. Body length Skull length as % of body Percentages of skull length Zygo- matic Inter- orbital Anterior frontal Upper canine Pre- maxillar Width of Alveolar length width width width width width naso- pharynx length First year S.S. 5 221 14-0 57-4 13-2 2I-0 20-3 13-2 12-6 26-9 W.S. 66i 226 15-0 51-0 13-3 20-6 18-9 II-9 12-5 26-8 W.S. 663 250 13-9 49-3 "•5 20-3 l8-2 II-2 I2-I 27-4 Second year S.S. 6 252 14-8 54-8 ii-i 20-9 197 121 13-9 26-8 W.S. 673 — — 55-2 99 21-5 19-5 II-O II-O 26-5 W.S. 656 284 13-1 52-9 10-6 21-2 i8-3 10-8 127 26-9 B. 81 251 15-1 49-3 10-2 i8-6 i8-i ic-S 12-4 26-1 R.C.S. 1095 50-3 9-8 17-0 17-9 II-O 10-8 25-6 B. 144 305 127 52-3 II-4 22-2 i8-i 10-9 I 2- 1 25-8 S.S. I 305 — 51-2 9-5 i8-8 17-9 10-6 II-6 25-8 38.12.3.2 — 53-1 10-8 20-4 19-0 II-2 12-5 26-3 93.9.I4.I — 48-0 9-8 19-8 i8-6 10-3 12-3 25-9 Sh. XX — 47-5 10-2 i8-4 17-4 117 10-8 27-0 Sh. X — 51-4 10-9 21-0 i8-9 12-8 I2-I 25-1 Third year 1 B. 16 274 14-1 53-2 10-8 22-6 19-1 12-9 11-2 25-2 3256 — — 54-1 9-8 — i8-8 II-8 12-0 25-4 B. 128 292 13-4 55-6 10-8 217 19-9 12-6 — 26-8 325" — — 55-5 10-6 22-9 18-9 12-4 10-4 25-4 B. 14 345 1 1-7 51-4 10-4 22-6 19-6 12-9 117 25-8 B.iS — — 507 107 207 — II-5 I i-i 26-8 3252 — — 54-3 10-5 19-3 20-2 — 107 25-4 325^^ — — 41-2 10-2 22-4 i8-9 ii-i I2-0 26-5 D.I 616 — — 49-1 9-5 19-5 17-5 II-4 — 24-4 B.17 — — 54-6 lO-I 21-8 i8-6 II-6 — 25-4 Fourth year Sh. xviii 325^ B. 142 — — 50-9 92 177 i8-i lo-S II-6 25-3 292 13-6 607 lO-I 20-2 19-6 II-6 I2-I 26-4 Sh. xix 351 II-5 49-6 99 19-6 i8-o 1 1-5 127 24-9 W.S. 574 — — 56-4 9-8 22-5 19-0 12-4 II-4 25-2 B.55 343* — — — — — — — B. 14s 300 13-9 60-4 9-4 22-1 i8-5 I2-I 8'3 237 B. 127 — — 58-5 9-4 217 19-5 II-9 107 257 R.C.S. 1092 — — 6o-i lO-I 23-8 20-4 13-1 9-9 267 B. 126 358 I2-0 57-5 9-6 23-4 20-9 I2-I lO-O 25-3 Badly damaged. 256 DISCOVERY REPORTS Third year (lo). Body length (3) 303-7 cm.; skull length 402-9 mm. ; zygomatic width 213-5 mm.; alveolar length 103-8 mm. The palatal sutures are beginning to fuse, most frequently starting with the maxillo-palatine followed by the median palatine and the palatine part of the maxillo-premaxillar. Fourth year and over (9). Body length (5) 328-8 cm.; skull length (8) 41 1-5 mm.; zygomatic width (10) 232-6 mm.; alveolar length 105-2 mm. The palatine sutures are fused and obliterated, the premaxillae are fused to the palatal and alveolar parts of the maxillae, and this fusion is continued along the vertical suture between these bones as far as half-way up, but the upper parts of these sutures are not fused in any skull so far examined. OS PENIS (Plate XI) The penis bones were not examined until it had been decided that there were four age groups in the skulls, but when they were, they clearly provided an additional means of distinguishing age groups. The smallest bone is that belonging to 325 ni, which is the smallest male skull, 313 mm. (first year). This os penis is 96 mm. long and about 2-5 g. in weight; it is curved in a dorsal direction with the maximum flexion about 34 mm. from the distal end, but beyond this point the axis of the bone returns to approximately its original line. The ventral surface has a narrow urethral groove for most of its length and there is a perceptible dorsal keel. The bone belonging to a second year animal, no. 657, is 132-5 mm. in length and 7-2 g. in weight; it is straighter than the bone of 325 m, but the distal part bends slightly downwards from about 38 mm. from the tip; the dorsal keel is less sharp, and about 40 mm. from the proximal end there are two elevations which mark the insertion of the ischio-cavernosus muscles. The OS penis of the third year, no. 664, exhibits a very marked increase in size since it is 180 mm. long and weighs 27-5 g. and is altogether a more massive structure. For most of its length the urethral groove is represented only by a flattening, but the dorsal keel is well marked. The sudden increase in size and weight compared with the bone of the preceding year — an increase of almost 36 per cent in length and of 380 per cent in weight — supplies adequate reason for belief that the third year is that which marks the onset of sexual activity, and this is in accordance with the structure exhibited by the pertinent testicular sections, that of no. 657 being juvenile in character but no. 664 showing definite signs of the adult condition. The remaining four bones are the massive structures of the adult, although there are indications that with sufficiently long series at least one additional growth stage could be distinguished ; the bones of nos. 662 and 655 are deeper and more compressed than those of nos. 665 and 666, with the ventral surface edged rather than flattened as it is in the last two. To summarize, the os penis shows a continuous increase in size up to that of the adult, and there is a sudden enlargement in the third year when the animal first becomes sexually active ; there is also a progressive alteration in shape from one which is curved THE LEOPARD SEAL 2S7 and almost cylindrical in section with a perceptible urethral groove to one which is nearly straight, sharply elliptical in section and almost devoid of a groove. The bone in the adult is not quite bilaterally symmetrical and exhibits conspicuous individual variation. Table VIII. Growth of the os penis Ref. no. Length Weight mm. g- First year ?,2Sm 96-0 2-5 Second year 657 132-5 7-2 Third year 664 i8o-o 27-5 Fourth year and over 665 225-5 39-5 666 238-0 51-5 662 240-0 59-0 655 225-0 6i-5 TEETH 2 I C As is well known the dentition of the leopard seal is — =32, and compares favour- ably with that of any other carnivore: no variations in number have been found in the seventy-two specimens examined. It is a matter for regret that there is no material for a description of the milk dentition. The frequency of broken teeth is commented on (p. 259), but as an example of an advanced state of injury the very large female no. 1092 in the Royal College of Surgeons may be described. The second lower left incisor is lost and the canine broken; the injury to the canine and the displacements are illustrated in PI. XIII. Of the ten cheek teeth in the upper jaw only four are complete, the remaining six being broken in varying degrees ; some indeed are destroyed. There is only one broken cheek tooth in the mandible. PATHOLOGICAL CONDITIONS (Fig. 5) In skull no. B. 55, about the beginning of the ascending ramus of the right mandible there are two bony nodules, one inside and the other outside and also numerous similar nodules of varying size on the anterior and ventral parts, there being at least ten on the palate alone. Five of these structures have been sawn through and some show a dense outer layer of bone with more cancellous structure within, but others are composed of dense bone throughout. In the left ethmoturbinal of the same skull there are two irregularly shaped nodules which are perfectly free in the dried specimen; they are both extremely dense. Skull no. 656 has two partly healed wounds in the bones of the face, and like similar lesions seen in other skulls they have apparently been caused by another leopard seal, as they are about the size one would expect to result from the use of the canines of 258 DISCOVERY REPORTS Hydnirga. There are also signs of inflammation in the alveolar region of this skull. It is a matter for speculation whether channels for infection may not have been opened either by the breaking of the teeth or the reception of wounds. It is well known that seals' wounds are almost invariably septic. B. Fig. 5. Pathological nodules on skull of B. 55. A, on the zygomatic process of the left maxilla and B, on the right pterygoid. FEMALE GENITALIA The genitalia of no. 656 were preserved ; this animal showed only early and micro- scopical signs of sexual activity, so that the following is a description of a young virgin (PI. XII). The ovaries are enclosed in almost complete capsules and are of a flattened bean- like form; the right organ is 2-8 by 1-4 cm. The cornua are rather short (13-2 cm.), and Aperture of Vagina Urethral Papilla Fig. 6. Dorso-posterior view of the vaginal aperture of no. 656, showing the hymeneal band. there is no projecting cervix. The vagina is i6-2 cm. in length. Its opening projects into the vestibule immediately before the urethral papilla and has a strap-like hymeneal band in a more or less median position. When fully straightened this band is 14-2 mm. long (Fig. 6). The vestibule, to the tip of the clitoris, is 9 cm. in length, so that the total length of the passage, excluding the Fallopian tube, is 38-4 cm. The hardness of the THE LEOPARD SEAL 259 specimen did not permit a dissection of the last-named portion. Thie clitoris is small and surrounded laterally and posteriorly by a low pad, presumably the homologue of the labia minora; pigmentation of the integument begins about the level of this organ. BIONOMICS HABITS AND FOOD Outside the breeding season Hydnirga appears to spend most of its time feeding, or sleeping off its meals on shore or on pieces of ice. A most striking characteristic is the solitariness of its life ; even if more than one be present on a beach the observer receives a strong impression of each seal being independent of the others and exhibiting the most complete indifference to them, in this contrasting very conspicuously with many other Pinnipedes. In water or on the ice it is much the same, and ordinarily these seals occur widely scattered. In thirty or forty miles steaming in open pack one may meet only three or four. The ten recorded in one day by Worsley (MS.) is quite an ex- ceptionally high number. It is only in the rutting season that they have been observed to take any interest in one another. Although like most other Pinnipedes this species tries to retreat when approached by man, it is more resentful of molestation than the other southern species, particularly in the water. Pouting (1922) describes how he irritated a leopard seal to such an extent that it pursued him over the snow until he was able to call in the assistance of E. A. Wilson who collected the specimen thus offered to hfm. Its reaction to interference in the water has earned it, in South Georgia, a reputation for ferocity which is scarcely justified. In the Falklands, reports occur from time to time of Hydnirga "attacking boats" or even leaping into them and attacking the occupants, but since it is extremely likely that the humans concerned took the first step by assaulting the seal the greatest caution must be observed in accepting such evidence. Shackleton (1920) writes that "a huge sea leopard climbed on to a floe and attacked one of the men"; since it is known that other seals are seized in this manner it may be that the leopard seal mistook the man on the ice for a Pinnipede (see p. 260). A Falkland Island report that a sheep dog has been killed by one of these animals is of course entirely credible, but one can hardly on that account support the thesis that Hydnirga is an evil creature. It seems probable that there is a good deal of fighting among these seals. Tooth marks on the skulls are to be found from time to time, some of them very severe, and broken teeth are fairly common, particularly in the large females, but whether these are due to accident, combat or courtship cannot be stated. The leopard seal has no natural enemies except perhaps the killer whale; Wilson (1907) mentions a seal which had been badly torn, and my specimen, no. 666, had on the neck two large old scars which had the appearance of marks left by the killer whale. The skull of the young female no. 661 had the right frontal and maxilla damaged to such an extent that a large fragment composed of parts of these two bones is detached 26o DISCOVERY REPORTS in the dry skull ; the pointed posterior end of the broken frontal projected from the line of the skull when fresh, and the edges of the bones show signs of regeneration ; there is a place on the crown where a fragment of the left parietal has been pushed down below the level of the rest of the bone and has fused in that position, but there were no skin markings to indicate the exact nature of the agency which produced such injuries. Even if it were necessary it would be very difficult to avoid the belief that the ferocity of Hydrurga is one of the reasons for its solitary mode of life, but it may also be a re- flection of another of its characteristics, one unique among the Pinnipedia, that of bemg a carnivore of large appetite and catholic taste. Besides being an eater of fish and cephalopods it is well known as a principal enemy of the penguins. Levick (191 5) has well described the seal's habit of lying in wait at penguin jumping-off places. R. N. R. Brown (1906) records the snatching of a penguin from a piece of ice, and I have myself seen it run down and catch a ringed penguin in open water, an impressive demon- stration of the speed of the mammal. Penguins are roughly shaken from their skins, but a great deal of skin and feathers is swallowed ; the faeces of animals which have been feeding on penguins are almost entirely composed of feathers. Other birds, e.g. Macronectes and Pelecanoides, are taken if opportunity offers. This species has frequently been seen feeding on the carcasses of whales and seals killed in the south. If large fragments are attacked pieces are grasped with the teeth and torn off by rotary movements. Hydrurga quite commonly feeds on other seals but not, so far as is known, on its own kind; Mawson (191 5) records having seen a Weddell seal eaten, and the chase of a crab-eater, which after several narrow escapes took refuge on an ice-foot. This was not necessarily safe : I was once called on deck in the pack to see a leopard seal which had jumped on to an ice-cake and attacked a crab-eater there. When I came out the leopard seal had disappeared but the prey had been torn open and killed, the ice as well as the carcass being smothered in fresh blood. A preference is shown for the softer parts of other seals, blubber, guts, etc. The results of the examination of the stomach and gut contents of thirty-two animals are recorded by the following: Gray, 1844, i; Ross, 1847, i ; E. A. Wilson, 1907, 2; Bruce, 191 5, 7 ; G. C. L. Bertram, MS., 6 ; J. E. Hamilton, subter, 15. They are as follows: Nature Occurrences Penguin Carrion 8 5 ; seal 4, whale i Squid Seal, fresh Fish 4 4, phocid and Otaria pup. 4, one contained 28 lb. Crustacea 3 ; unnamed 2, Euphausia superba i Pelecanoides I Ascidians, composite Alga Empty I 2, probably accidental 4 THE LEOPARD SEAL 261 SEXUAL MATURITY Males. Pieces of testis and epididymis from the following have been sectioned: nos. 651, 653, 654, 655, 657, 662, 664, 665 and 666. Of these nos. 651 and 653 are the two smallest animals and the sections are obviously those of juvenile testes: the lumina of the tubules are not yet developed, and there is of course no trace of spermatogenesis. In no. 657 the organ is somewhat more advanced, since lumina are beginning to appear in the form of irregular central spaces in the tubules, but spermatogenesis has not begun. Of the remaining six, three, nos. 654, 664 and 665, are in a state of almost complete inactivity: a few sperms only may be found by searching and large syncytial cells are present in all, although fewest in no. 664 which is the least active. No. 654 has a very few sperms in the epididymis and no. 664 none. The last three testes are those of nos. 655, 662 and 666; no. 662 shows a slight amount of general spermato- genesis, and a mass of sperms which is of limited size appears in three or four contiguous sections of the epididymial tubule. The other sections of it in the same slide are empty. Nos. 655 and 666 are both more active than no. 662, but neither of them appears to be fully so. Sperms are present in the ducts of both and there are syncytia in the testis of no. 666. From the evidence available, derived from the condition of the os penis and the testes, it must be concluded that sexual activity does not begin until the third year, and that animals of that age are adolescent, the fourth year being that of sexual maturity. There does not appear to be a definite anoestrous season in the male of Hydrurga as a species, but this does not exclude the possibility of a resting period for individuals such as no. 665. Females. Ovaries of six non-pregnant animals have been examined. The two first- year animals, nos. 661 and 663, have characteristically juvenile organs, although under the microscope a very early stage of a Graafian follicle was found in the ovary of no. 663. In the second-year seals, no. 656 has many primordial follicles and numerous developing Graafian follicles of microscopic size. This animal was killed in October and was a virgin. S.S. i, killed 18 January, has folhcles up to 4-5 mm. in diameter and a corpus luteum of 26 mm. longest diameter, and B. 144, killed 12 February, has numerous follicles up to 7 mm. and two corpora lutea, one recent and one older. Ovaries of the third year are all of pregnant animals and are three in number, but the one non-pregnant ovary, B. 142, from a mature seal (fourth year or over) was collected on 27 January and shows fairly numerous follicles with a maximum of 6-5 mm. and four corpora lutea of different ages, the largest being 30 by 18-5 mm. It is a matter for regret that the evidence is not more conclusive, but I do not consider that one may safely go further than to state that it seems highly probable that this species may begin ovulating in its second year, but does not always do so, and that it commonly becomes pregnant in its third year, perhaps for the first time. The third year therefore should be considered as the year of sexual maturity at present. 262 DISCOVERY REPORTS BREEDING HABITS As might reasonably be expected from the sohtary habit of this animal, information regarding its breeding is scarce. From the previous section it appears that the ovaries of cows killed in the first three months of the year are active, and the only month in which I have seen leopard seals keeping company is February, when as many as three were observed swimming together. Wild's description (1923) of a "fight" when one seal was seen to leap repeatedly from the water to a height of 6 ft. was much more likely to be a record of courtship. The date was 9 February. The cow, S.S. i, killed on 18 January had a single corpus luteum and no trace of others, so that this was probably the first ovulation of the season. Bertram's no. 142 of 27 January has four corpora lutea in the two ovaries, the largest being 30 mm. in longest diameter and the other three in difTerent stages of degeneration, and his seal no. 144 has a corpus luteum in each ovary. I consider that it may be safely assumed that the first two months of the year, and perhaps the third as well, are the season of mating, and from the evidence of the ovaries it is highly probable that this species ovulates several times at short intervals. Additional evidence of a rather prolonged mating season is supplied by records of foetuses. On 16 February a foetus of 16 mm. was obtained by the Discovery Committee's stafT, and in late February 1921 I found foetuses about 30 mm. Bertram records the following in March and April : Table IX Ref. no. Date Length (mm.) 1572 9. iii. 78 1451 17. 111. 118 IIOI 22. 111. 42 1 103 22. iii. 77 1461 28. iii. no 1459 28. iii. 112 354 30. 111. 250 1121 3. IV. 157 1137 30. IV. 360 It will be observed that within nine days there were collected foetuses of 42, 112 and 250 mm., and that for March there is recorded a foetus of 118 mm., nearly three times the length of one secured five days later (no. i loi, 42 mm.), this difference being analogous to that of human foetuses of four and two and a half months. Records of newly born pups are very scarce. Bruce states that in the South Orkney Islands only one was seen, but not secured, in November; Matthews saw one about 107 cm. in September, and the specimen in the British Museum was killed, according to the label, at Christmas. The extended pupping season implied by these observations is quite in keeping with the prolonged mating season suggested above (January to March). It follows that the period of gestation is about eight months. Since females THE LEOPARD SEAL 263 killed in the mating season do not have milk it is to be presumed that Hydrurga, like the other Phocids, has a short period of lactation. The breeding habits of the leopard seal are comparable with those of Lobodon, which mates in the southern autumn and brings forth the young in the following spring, rather than with those of Macrorhimis, Leptonyx and some of the northern seals which come on heat very rapidly after parturition and have a period of gestation of eleven months or even more, and in this resemble some at least of the Otariidae. It is interesting to recollect here that in the multiplication of cusps of the teeth Hydrurga and Lobodon approach one another more nearly than they do to the other southern seals. COMMERCIAL VALUE A number of leopard seals are killed annually in South Georgia (Matthews, 1929) but this is only incidental to the killing of the elephant seal, and the number is so small that, coupled with the relatively low oil production of this species, it is a matter of no commercial importance. I consider that an animal of such scattered distribution and solitary life is most unlikely to be the object of commercial interest, unless indeed its skin became fashion- able for some purpose, and that does not seem very likely, since although it is well coloured and marked the hair is stiff and rather thin. PARASITES Infernal By the courtesy of Dr H. A. Baylis I am able to publish the following list. Nematoda : Contracaecum osadatiim {Kud.) = Ascaris rectangida (von Linstow, i ()oy) = As car is stenocephala Raillet and Henry, 1907. Contracaecum radiatum (von Linstow, 1907) = ^5ram/fl/a]^era Raillet and Henry, 1907. Porrocaecuni decipiens (Krabbe, 1878). Anisakis similis (Baird, 1853). Cestoda : Diphyllobothrium quadratnm (von Linstow, 1892) = ^. resimiim Raillet and Henry, igiz = Dibothriocephalus coatesi Rennie and Reid, 1912. Phyllobothrium sp. [P. de/phmi (Bosc, 1802) fide Southwell and Walker, 1936]. Acanthocephala : Corynosoma hamamii (von Linstow, 1892). Siphunculata: External Antarctophinis ogmorhini Enderlein. I am indebted to Miss Theresa Clay of the British Museum for the information that this parasite has been recorded from Hydrurga from Victoria Land, Booth Wandel Island and Macquarie Island. 264 DISCOVERY REPORTS REFERENCES AiNSWORTH, G. F. See Mawson (1915). Amundsen, R., 19 12. The South Pole, trans. Chater, London. Anderson, K. A., 1905. Wiss. Ergebn. schwed. SiidpolarExped. v (2). Barrett-Hamilton, G. E. H., 1901. Seals, in Voyage du Belgica. Anvers. 1902. ^ Southern Cross' Collections, I. London. DE Blainville, H. M. D., 1820. Sur quelques Cranes de Phoques. J. Physiol. Path. gen. xci. Brown, R. N. Rudmose, Mossman, R. C. and Pirrie, J. H. H., 1906. Voyage of the 'Scotia'. Edinburgh. Bruce, W. S., 1894. See Burn-Murdoch (1894). 1915- Scientific Results of the 'Scotia', Pt. xi, reprinted from Trans. Roy. Soc. Edin. 1913. Burn-Murdoch, W. G., 1894. From Edinburgh to the Antarctic. London. CuviER, F., 1824. Mem. Mus. Hist. nat. Paris, xi. 1826. Dictionnaire des Sciences naturelles, xxxix. 1829. Dictionnaire des Sciences naturelles, L. Ferris, G. F., 1936. Contribution towards a Monograph of the Sucking Lice, Pt. vii. Stanford University, California. Gaimard. See Quoy. GiSTL, J. von N. F. X., 1848. Naturgeschichte des Thierreichs. Stuttgart. Gray, J. E., 1844. Zoology of the Voyage of the 'Erebus' arid ' Terror'. London. 1862. Catalogue of Bones of Mamrnalia. London. 1866. Catalogue of Seals and Whales. London. 1874. Hand List of Seals, etc. London. Home, Sir E., 1822. Philos. Trans. Jacquinot, H., i844(?). Voyage au Pole Sud, Atlas. Paris. Jacquinot, H. and Pucheran, — , 1853. Voyage au Pole Sud, in, Mammiferes. Paris. Levick, G. M., 1915. British Antarctic {'Terra Nova') Expedition, 1910. Natural History Report, Zoology, I, 2. London. MacGillivray, John, 1853-4. Private Journal of Voyage of H. M.S. ' Herald'. Admiralty MS. Matthews, L. H., 1929. The Natural History of the Elephant Seal, etc. 'Discovery' Rep. i. Mawson, Sir D., 1915. The Home of the Blizzard. London. Milne-Edwards, a., 1891. Mission scientifique du Cap Horn, vi. MosELEY, H. N., 1892. Notes by a Naturalist. London. Peters, W., 1875. Mber. k. preuss. Akad. Berlin. Ponting, H. G., 1922. The Great White South. London. QuoY, J. R. C. et Gaimard, J. P., 1830-33. Voyage de I'Astrolabc. . .pendant 1826-29. Zoologie, vol. i, 1830-33- Paris. Ross, Sir J. Clarke, 1847. A voyage to the South. London. Shackleton, Sir E., 1909. The Heart of the Antarctic. London. 1920. South. London. Southwell, T. and Walker, A. J., 1936. Notes on a larval cestode from a fur seal. Ann. trop. med. Parasit. xxx, i . Trouessart, E. L., 1907. Exped. Antarct. Eratifaise, 1903-5. Mammiferes Pinnipedes. Paris. Valette, L. H., 1906. Viaje a las Islas Orcadas Australes. Ann. Minis. Agric. Zool. in, 2, Buenos Aires. Walker, A. J. See Southwell (1936). Weddell, J., 1825. A Voyage towards the South Pole. London. Wild, F., 1923. Shackleton's Last Voyage. London. Wilkins, G. H. Shackkton-Rowett Expedition, Mammalogical Report, ig22. Brit. Mus. MS. Wilson, E. A., 1907. National Antarctic Expedition, 11, i. London. Worsley, F., 1914-16. Shackleton's Second Expedition. MS. journal. A few additional references will be found in the ' Southern Cross ' and ' Scotia ' Reports. PLATE VII Fig. I. Leopard seal. Fig. 2. Skin of pup in first coat. DISCOVERY REPORTS VOL. XVUl PLATE VII JohnBaltSons & Cumow L.*^ London. THE LEOPARD SEAL PLATE VIII Skulls in dorsal view, to show successive growth phases (see pp. 250, 252, 256). Male. Fig. I. No. 653. First year group. Male. Fig. 2. 1.1.4. 15. Second year group. Male. Fig. 3. S.S. 2. Third year group. Male. Fig. 4. No. 679. Fourth year group. Female. Fig. 5. No. 663. First year group. Female. Fig. 6. B. 144. Second year group. Female. Fig. 7. 325 •.■'';..irj-jirn'v3.'i 'lUOV £ ')(, ■.^\■Jv\i\\[_ [■.yi ibdj pfir/J'jrie gabt PLATE XV Fig. I . Sarsia gracilis Browne. Lateral view of a tentacle belonging to an adult. Figs. 2-3. Hybocodon unicus (Browne). Fig. 2. Lateral view of the medusa. Fig. 3. Diagram of the margin of the umbrella. Oral view. Fig. 4. Staurocladia vallentini (Browne). Lateral view of a tentacle; type-specimen. Figs. 5-6. Podocoryne tenuis (Browne). Fig. 5. Lateral view of medusa, with medusa-buds. Fig. 6. Mouth expanded. Figs. 7-14. Bougainvillia macloviafia Lesson. Fig. 7. Lateral view of an intermediate stage. Fig. 8. Lateral view of a very early stage. Fig. 9. A transverse optical section of the umbrella of a young stage, showing the interradial and perradial furrows in the umbrella, and the internal adradial ridges. Fig. 10. A compound basal bulb, with a single row of tentacles. Inter- mediate stage. Oral view. Fig. II. A compound basal bulb, with a double row of tentacles (which are vigorously contracted), of a fully grown adult. Oral view. Fig. 12. The oral tentacles of an early stage. Fig. 13. An oral tentacle of a fully grown adult (umbrella 12 mm. in length and width). Fig. 14. Lateral view of the oral tentacles showing their relation to the mouth. DISCOVERY REPORTS VOL. XVII) PLATE XV 0- ■■> 0=-' ! J l^li 13 v.- ■ j.-i 14- JohnSaie 5ons A Cumow, L^*! Lmidon MEDUSAE FROM THE FALKLAND ISLANDS .(onwoid) j^siijii. ^JfTWOlfl ' i' -l -W' PLATE XVI Figs. 1-2. Halitholus intermedius (Browne). Fig. I . Lateral view of an early stage. Fig. 2. A tentacle, lateral view, contracted. Figs. 3-5. Laodicea pulchra Browne. Fig. 3. Oral view of an adult, showing the margin of the umbrella, the stomach with lobes, and the mouth. Fig. 4. Aboral view of an adult, showing the gonads on the lobes of the stomach. Fig. 5. The tentacles with the marginal clubs (cordyli) and ocelli (velum removed). Inner (adaxial) side, re, ring-canal. Figs. 6-8. Obelia multida Browne. Fig. 6. Oral view of an adult, male. Fig. 7. A tentacle, showing the outer (abaxial) side. Fig. 8. An ovar}'. Figs. 9-12. Obelia sp. Fig. 9. Oral view of an adult, female. Fig. 10. A tentacle, showing the outer (abaxial) side. Fig. II. An ovary, lateral view. Fig. 12. The base of the stomach, aboral view. DISCOVERY REPORTS VOL . XX'IIl PLATE XVI A / ' y ii 12 \'j\l .T. V \\ "O -^ 8 / '^- / E T.Brov/ne del / ::::K^^ '^'//y;. •iV 6 JohnSaleSons 4. Cumow. L*^ Louden. MEDUSAE FROM THE FALKLAND ISLANDS •'orig (Sllafidmu srlJ lo n .01 .gil .£1 .^I'i PLATE XVII Fig. I. Cosmetirella daviii (Brovfne). Lateral view of an adult. Figs. 2-4. Phialella falklandica Browne. Fig. 2. Lateral view of an adult. Fig. 3. An intermediate stage. Fig. 4. A statocyst. Figs. 5-9. Phialidium simplex Browne. Fig. 5. Intermediate stage, oral view. Fig. 6. An adult, lateral view. Fig. 7. Margin of the umbrella, showing tentacles and statocysts, outer (abaxial) view. Fig. 8. The mouth of an adult. Fig. 9. Diagrams showing the growth of the ovaries. A, umbrella 3 mm. in diameter; B, 4 mm.; C, 6 mm.; D, 10 mm.; E, 15 mm. ; F, 20 mm. Figs. 10-12. Witlia mutabilis Browne. Fig. 10. The clusters of nematocysts on the exumbrella, near the margin. Fig. II. The mouth expanded, of an early stage, showing the lips, and commencement of the folds. Fig. 12. Transverse section of ovary. DISCOVERY REPORTS VOL.XVUI PLATE XVII Y N? 1 li r-'' 10 E X BrowfTi« d«l )5 A B A J c o 9 t !.■' John Bak Sons A Cumow L*** Laiidan MEDUSAE FROM THE FALKLAND ISLANDS iirrK 3! •anv/oifl wtaiinwa w3V>i>inc'. .ilubs an ' as Ivj V/siv ItiaJiiJ .i .VI ' .t.3 (IB io y.'-' ' " ' .W3IV j^ eadol ^uo'i ■,.■'./- A yj'q .-j;;/;:;^. o; .Ij-j!mri-|7'. Q0!i7/ noiJiaoq :;> ^^ .nigifim ^rfi ^o aolioae iBob&cnmKr^Htb f Aw. ;s[f9id,niizi .xs .J8ydo}bJ8 axlr bne eabisJnaJ ■ .alsBJnaJ b lo noiJioq IsninnaJ sdT .di .j^il .aamhsqc. >vj^j io war .^i .-^H -ji'i "r:h/aih nigifim ifbidmu drfj '1 ■ ■■ ■' ,,,, .8i .siH Fig. I. Fig. 3. F'g- 3- Fig. 4. Fig- 5- Fig. 6. PLATE XVIII Figs. 1-6. Tiaricodon coeruleus Browne. Lateral view of an adult. Lateral view of an intermediate stage. Lateral view of an early stage. A tentacle, contracted, belonging to an adult, outer (abaxial) view. The mouth and its folded margin. Adult. Aboral view of the stomach, showing the four lobes bearing gonads, upon the peduncle. Adult. Figs. 7-16. Aglauropsis conantii Browne. Fig. 7. Lateral view of a large adult. Fig. 8. Lateral view of an intermediate stage. Fig. 9. Lateral view of an early stage. Fig. 10. A gonad of an adult. Fig. II. The base of the stomach when expanded. Fig. 13. The base of the stomach when contracted. Fig. 13. Radial section of a statocyst. Fig. 14. A diagrammatical section of the margin, showing the position of the tentacles and the statocyst. ex, exumbrella; sub, subumbrella ; re, ring canal ; v, velum. Fig. 15. The basal portion of a tentacle. Fig. 16. The terminal portion of a tentacle. Figs. 17-19. Vallentinia falklandica Browne. Fig. 17. Lateral view of type specimen. Fig. 18. Diagram of one quadrant of the umbrella margin, showing the position of the tentacles. Fig. 19. The adhesive disks of the perradial tentacles. DISCOVERY REPORTS VOL . XVin PLATE XVIII 1^1 ex \ ^Su6 fN / i-^ J-- i\ J — r 6. . Ring canal ms. m n 4- A- J^ 13 14 -,X~. 10 19 15 fWi il . k ^3 J ^--- — ' r^y/ N E.T, Brov.Tie del JohnBakSonsi Cumow. L'^ London. MEDUSAE FROM THE FALKLAND ISLANDS gnivv^ai yjA slTJ-d". .30/ .iy>l .■ii'ji:. M\Z .£ .uri ■laT • E -U' ■■ BnimioT 4 -i^i ■:. » JfioijqO ■I -8''^ • -' S'-i .r .y'll ;3ulb30 ,:ii> to.l .ft •T .01 .."ji'l •irr?r;:;-r j; ';n sfToiri' - V5T .-r .yi'^ PLATE XIX Fig. 1. Rathkeafor?nosissim a (Browne). Lateral view of the manubrium; specimen with medusa-buds, showing the relation of the oral arms to the mouth. Fig. 2. Staiirocladia vallentini (Browne). A tentacle of a specimen collected in 1901. Figs. 3-5. Phialella falklandica Browne. Fig. 3. Tentacle and basal bulb, outer (abaxial) view. Fig. 4. Terminal portion of a tentacle, expanded. Fig. 5. Optical section of basal bulb. Figs. 6-7. V allentinia falklandica Browne. Fig. 6. Upper part of young specimen, lateral view. Fig. 7. Young specimen, aboral view. Figs. 8-1 1. Tiaricodon coeruleus Browne. Fig. 8. Longitudinal section of basal bulb, ex, exumbrella; oc, ocellus; rad.c. radial canal; v, velum. Fig. 9. Transverse section of basal bulb at the level of the ocellus. Fig. 10. Transverse section of basal bulb, the thickened part beyond the ocellus. Fig. II. Transverse section of tentacle, median portion, showing clusters of nematocysts. Fig. 12. Willia mutabilis Browne. Six portions of a tentacle, from the basal to the distal part, showing the clusters of nematocysts. ab, abaxial side ; ad, adaxial side. DISCOVERY REPORTS VOL . XVIII PLATE XIX El r rad.c t ^*«» II 12 P I, Kramp del. \-ex 8 JohnB*lfcSons A, Cumow L*^ LoaAan MEDUSAE FROM THE FALKLAND ISLANDS [Discovery Reports. Vol. XVIII, pp. 323-338, Plates XX-XXI, November 1939.] MADREPORARIAN CORALS, WITH AN ACCOUNT OF VARIATION IN CARYOPHYLLIA By J. STANLEY GARDINER, M.A., F.R.S. MADREPORARIAN CORALS, WITH AN ACCOUNT OF VARIATION IN CARYOPHYLLIA T By J. Stanley Gardiner, M.A., f.r.s. (Plates XX-XXI) HE corals described in this report were collected by the scientific staff of the Dis- covery Committee between the years 1926 and 1933. The following localities are represented : Gough Island, St. 399. Tristan da Cunha and Inaccessible Island, Sts. 6, 1187. Cape Lopez, French Congo, St. 279. Annobon, Gulf of Guinea, St. 283. New Zealand — North Island, St. 939. South Georgia and Shag Rocks, Sts. 152, 160. Clarence Island, St. 170. Palmer Archipelago, Sts. 181, 182, 190. Cape Horn, St. 388. Patagonian Shelf, Falkland Islands, and Burdwood Bank, Sts. WS 76-99, WS 210- 250, WS 792-871. Further information is given in the Station Lists issued by the Discovery Committee. Corals were obtained in 34 dredgings and these have been referred to 14 species in 9 genera. In each of 29 dredgings only a single species was obtained, in 4 two species and in i three species. On the other hand, Flabellum curvatum was obtained in 17 dredgings, 79 to 278 m., and Balanophyllia cor?iu in 6 dredgings, 75 to 404 m. ; these are Moseley's Challenger species, which were obtained respectively from off Rio de la Plata, 600 fm., and off Ki Islands, 129 fm. Of the rest 4 species were each obtained from two dredgings and i from three. I can recall from no expedition any analogy to the wide distribution in its area of Flabellum curvatum, of which 158 coralla were ob- tained. A search for characters in these which might be supposed to be related with a low temperature, as the one non-fluctuating feature in their localities, proved unavailing ; indeed, they presented no differences in texture compared with F. rubrum andpavoniimm, of which I have had large numbers of specimens from the Indian Ocean. Of other species Moseley's CaryophylUa profunda is at present only known from the Antarctic, as are two species of Gardineria. The identification of the specimens of CaryophylUa led me to a consideration of taxonomic characters in this genus. The examination of the collection in the British Museum was of great value, for which I thank Captain Totton. Co-types of Pourtales were kindly supplied to me by the Agassiz Museum for which I have to thank Dr Barber, the Director; they will be added to the British Museum collections; they cause me to admire and feel confidence in Pourtales' work. 326 DISCOVERY REPORTS The Director of Research has kindly allowed me to insert in this Report an account of Turbinolia australiensis n.sp., the first species of the genus obtained alive. LIST OF SPECIES Fam. FLABELLIDAE. Fam. TURBINOLIDAE. Flabellum harmeri Gard. Caryophyllia mabaliiUii Gard. and Waugh. F. ciirratum Moseley. Tiirbmolia australiensis n.sp. Gardineria antarctica Gard. Sphenotrochus intermedins (Munster). G. lilliei Gard. Deltocyathus lens Alcock. Fam. TURBINOLIDAE. Fam. STYLOPHORIDAE. Desmophxllum capense Gard. Madracis decactis (Lyman). Caryophyllia cyathtis (E. and S.). Fam. EUPSAMMIDAE. C. profunda Moseley. Dendrophyllia oahensis Vaughan. C. arcuata Ed. and H. Balanophyllia cormi (Moseley). Genus Flabellum^ The corals of this genus have no theca and hence no costae, no pali and no columella, any central filling in being by trabeculae from the larger septa. Its epithecal wall is little thickened, except where the internal polyp is stimulated by destruction of the same to deposit additional corallum to heal the wounds. In growth it precedes the forma- tion of septa and, externally, commonly shows rather irregular transverse lines of growth. Some species remain attached, others become detached, while in a third group (cp. nibnim) the coral breaks off above its basal attachment leaving a distinct scar showing I or 2 cycles of septa ; very exceptionally the third method may be found in a species which normally is attached or free (cp. curvatum below). The number and size of the septal cycles vary, but hexamery is universal, no variation having been found either in young or old. The septa increase by the formation of new cycles during growth, these appearing as two new septa on either side of each septum of the last formed cycle, viz. alternating with the pre-existing septa. Each species aims at a definite number of cycles, but this is occasionally passed in the species with flattened calices by the forma- tion of new septa on either side of the terminal directive septa, these being necessary for the filling up of gaps in the calices, in which the septa are always regularly spaced. Flabellum harmeri Gardiner. Brit. Ant. (' Terra Nova') Exp. 1910, Zool. v, 122, I 19-20 (1929). St. 170, 342 m. A single dead specimen, 49 mm. high, calice 56' 25 mm., summits of side walls 12 mm. higher than ends, 24 larger and equal septa and cycles I-V complete with at the ends of the calice some of cycle VI. No trabeculae in the axial fossa are visible from the surface. The base is broken across and there are 12 septa in the section. The strongest features of separation from airvatn?n below lie in the size, the equality of septa I-III and 1 For a general consideration of the genus and some species I may refer to Marine Inv. S. Africa, 11, ^-^^^ i_iv, 1902 and Rec. bid. Mus. pp. 301-16 (1929). I have examined the polyps of several species without finding specific characteristics in the polyps other than those indicated by their coralla. MADREPORARIAN CORALS 327 the greater compression. The outside of the column is largely covered by growths of Polyzoa and one end of the caHce has been broken and regenerated. There are about 16 growth rings on the epitheca. Flabellum curvatum Moseley (PL XX, figs, i, 2). 'Challenger' Rep. p. 174, VI 2 (1881). The series comprises 159 specimens dried, and a number in spirit, well preserved. The polyps show a deep purple rim round the peristome where the tentacles rise, the other parts being less coloured ; the pigment is gradually soluble in alcohol so that the whole polyp may be lightly coloured. The tentacles are tightly set with nematocysts at their ends and are found over all septa except those of the last formed cycle. The corallum is a flattened cone with no side wings; it becomes detached from its original support and lies freely on the bottom. The calicular opening is about 100 to 65, length to breadth. The epitheca usually shows transverse growth ridges. The rim of the epitheca forms a slight upstanding wall above the edges of the septa, which never rise above it. In the smaller specimens 12 equal septa regularly fuse by trabeculae to fill in the central valley of the calice at about half its depth. In larger specimens septa III are usually joined to these, but are narrower and thinner, their trabeculae extending inwards at a deeper level. In median-sized forms four cycles of septa are present, but in larger specimens septa V appear and, when this cycle is com- plete, 1-4 or more pairs of VI may form, usually near the directives. Growth ridges are distinct on the septal sides and are set with spines ; the result is a rough appearance of the septa, the edges of which may show fine serrations. Every individual coral varies in its axial valley, the trabecular floor of which usually lies at a depth of one to two-thirds the height of the corallum. Septa I and II slope to this level almost perpendicularly, and the false columella formed by their trabeculae varies greatly in breadth and coarseness of structure ; its formation is very little assisted by extensions from septa III. The extreme height of the specimens is 48 mm., calicular length 36 mm. While plenty of specimens show curving, this is never as extreme as in Moseley's eight specimens which were dredged at 600 fm. off the Rio de la Plata. The external decompo- sition of the epitheca is as Moseley described, and this makes visible the septa below as ridges down the same. The wide distribution of these corals shown below suggests that we are dealing with a definite species, the variation of which will now be clear. There may be several synonyms, previously described species, or it may be a synonym of one of these, but no such relationship is clear. Localities: St. WS 76, 207-205 m., i, 20 mm. high by calice opening 22 x 15 mm., 12 septa solidly fused in centre, cycle V almost complete: St. WS 80, 152-156 m., I large and 2 small: St. WS 83, 137-129 m., 3, with calices 17, 16 and 12,^ in larger cycle V considerably developed: St. WS 85, 79 m., 3 young: St. WS 92, 145-143 m., 1 Such single measurements are the maximal along the calicular opening. 328 DISCOVERY REPORTS I, i6mm. long: St. 182, 278-500111., 18 and 15 mm., long, rather thin coralla: St. WS 210, 161 m., I young: St. WS 212, 242-249 m., 25, 21 and 11 mm., delicate: St. WS 216, 219-133 m., 26, half vary 22-30 mm., rest grading up in size from 3-5 mm. with 12 septa not yet joined and 5 mm. 20 septa: St. WS 237, 150-256 m., i dead and I young: St. WS 243, 144 m., i young: St. WS 244, 253-247 m., i of 18 mm., 11 small grading from 18 septa, these coralla open in centre, 6 still attached: St. WS 247, 172 m., heavy massive form, 27 mm. high, 27 < 17 mm. calice, septa I and II alone fused in centre, this most exceptional for the size: St. 652, 164m., 69, very uniform between i2and 19mm. in calicular length, 12 septa fusing, 4 cycles complete and some of V, many attached to pebbles, all somewhat eroded outside, most 5 or 6 growth rings: St. WS792, 102-112 m., I large of 33 mm.: St. WS 795, 157-61 m., 6 large open forms up to 30 mm. with 3 attached young, smallest 3-5 m., 16 septa not joined: St. WS 839, 403-434 m. i of 46 mm. high, calice 28 x 21 m., epitheca markedly upstanding, 5 young corals attached: St. WS 867, 150 m., 27, similar to 652, more graded and open caHces, smallest 5 mm. with 12 septa fusing and 12 of III. Genus Gardineria Vaughan Bull. U.S. Nat. Mus. no. 59, p. 65 (1907). Duncania, Pourtales, Metn. Mus. Comp. Zool. Harvard, iv, 44 (1874). I have expressed the opinion that this genus should be absorbed into Haplophyllia on somewhat slight grounds.^ Vaughan and Wells are considering the question of coral genera, and, to avoid confusion, it is better that I should retain the genus at present. Gardineria antarctica Gardiner. Brit. Ant. (' Terra Nova') Exp. 1910, Zool. v, 124, I 11-18 (1929). A single specimen, calice 17 x 15 mm. across, was obtained from St. 190, 90-130 m. ; it is closest to Fig. 14 above, the columella-pillars being even more developed. In comparison with G. capensis'^ only 12 septa fuse to form the columella, and the septa do not rise above the rim of the calice. There are also two large dead specimens, some- what decayed, from St. 152, 245 m. Gardineria lilliei Gardiner. Brit. Ant. {'Terra Nova') Exp. 1910, Zool. v, 125, I 3-10 (1929). St. 140, 122-136 m., I, 20 mm. high, broken calice with about 50 septa: St. 160, 177 m., (a) 2, large, 32 mm. high by 16 mm. across mouth of cahce and 24 by 17 mm., (b) 2, about 13 and 10-5 across, (c) 12, smaller, averaging 24 by 9 mm., and (d) 20, still smaller forms: St. 181, 160-335 '^•' 2, 42 by 17 and 31 by 15 mm., one much curved: St. 190, 90-130 m., I, 52 by 20 mm. The series contains 40 specimens, which were mostly attached to pebbles. They are upstanding cones, some a little bent, generally almost circular in section. 1 'John Murray' Exp. v, 173 (1938). - Duncania capensis, Marine Inv. S. Africa, iii, 120, I 6a-c (1904). MADREPORARIAN CORALS 329 The original description is correct for the six corals collected by the ' Terra Nova ', but requires qualifications if the present series is referred to lilliei, the only other course being to describe 4 new species on characters, which are believed to be dependent on growth as governed by environment. The thin epitheca is a marked feature, but, as in the smaller forms from St. 160, there may be external erosion or the attachment of sedentary animals, either of which may induce a thickening of the epitheca by the internal deposition of corallum ; these may also inhibit the broadening of the calice. The result of erosion is to show longitudinal striae corresponding to the internal septa. These cross the circular growth-markings, which for the intermediate series from St. 160 suggest 7-9 growth periods. The epithecal growth precedes the septal, so that the septa are never exsert. Septa I and II (12), as a regular feature, meet deep down in the calice in a mass of twisted trabeculae. Then there are septa III and IV; the latter are incomplete only in the smaller forms, and there are as many septa of V as the available interseptal gaps allow, the earlier septa showing little thickening. In the smaller forms the edges of the septa are untoothed and the sides almost smooth, but in the two larger from St. 160 they are studded with low spines. The size and depth of the trabecular, and hence false, columella depends largely on the shape of the corallite, in tall forms of small diameter being scarcely visible. The two low forms of St. 160 {b) are relatively shallow with 3 and 2 twisted trabecular pillars standing up on the false columella ; these have 4 cycles of septa. The specimen from St. 190 is a very clean and rapidly grown form; its septa are thin and tend to be wavy, while their fusion by trabeculae is so deep down as to be almost invisible. These characters are to a lesser degree seen in two similar specimens from St. 181. Genus Desmophyllum (Ehr.) Ed. and H. Desmophyllum capense Gardiner. Marine Inv. S. Africa, iii, 96 (1904). St. WS 99, 251-225 m., 7 large specimens; St. WS 250, 257-313 m., i, 45 mm. high by 32 mm. along the length of the caHce, only commencing to show the vertical ridges characteristic of the larger forms. Being doubtful as to the range of a species which may include Desmophyllum christa- galli Ed. and H.,i D. itigens Moseley^ and the present forms I must retain this species. The outer surfaces of the 7 coralla from St. 99 are swollen out into rounded ridges between the chief septa (I, II and III). Between these larger septa, in accordance with the interseptal gaps that these ridges may provide, are found septa of cycles IV, V, VI, VII and occasionally VIII. A specimen (co-type) of the D. capense before me has similar but lower ridges; its calicular opening measures 38 by 25 mm. and has a rather open 1 Vide Duncan, Trans. Zool. Soc. 1873, p. 321, XLI 10-16 and Marenzeller, ' Valdivia' Exp. p. 267, XV 2 (1904); the resemblance of the present species to the forms figured in 2 and 26 is obvious. 2 'Challenger' Rep. p. 160 (188 1). 330 DISCOVERY REPORTS centre. Four of the larger specimens above, 58 by 35 mm., 48 by 30 mm., 75 by 45 mm. and 58 by 35 mm., have 15, 18, 20 and 18 ridges; all are much closed in the centres of their calices by the inner edges of the packed and upstanding septa. The ridges extend down nearly to the bases of all four specimens. All are broken off by the trawls above their attachments, providing transverse sections; these show no endothecal trabeculae like dissepiments, except in one specimen, which is decayed in its lower parts, and here they are most irregular but useful in strengthening the corallum. Trabeculae may join together the larger septa in the centre of the corallum, thus producing a certain obliteration of such axial space as exists. The coralla would seem to form clusters, but there is no suggestion of other than an adventitious association of polyps. There is little internal thickening of the theca, but the edge zone of the polyp, as far down as it persists, would seem to be depositing corallum on the same and thus filling in the intercostal spaces. The specimens are much overgrown by worm tubes, by hydroids, etc., and the edge-zones of the polyps do not exist outside the theca for more than about 10 mm. I have specimens oi DesmophyUum before me identified by Pourtales as Desmophyllum christi-gaUi Ed. and H., Desmophyllum nisei, Desmopliyllmn cailleti both Duch. and Mich., from the West Indies, also Desmophyllum vitreum Alcock ; all of these on the material before me, I regard as good species. The constant difficulty in identification in genera such as this, with a growth to a large size, is a lack of knowledge as to whether any speci- men is adult or immature. Judging by the ways of nature, there must be an immense mortality in the young, few surviving to the adult stage, one perhaps in 1000 or more. Genus Caryophyllia Lam. Caryophyllia cyathus (E. and S.). Madrepora cyathus Ellis and Solander, Nat. Hist. Zooph. p. 150, xxvni, 7 (1786); Ed. and H., Ann. Set. Nat. ser. 3, ix, 287, iv, i (1848); Lacaze-Duthiers, Arch. Zool. exp. gen., ser. 3, xv, 12-91 (1897); Marenzeller, ' VaUivia' Exp. p. 295, XVI 6 (1904); Doderlein, Mit. Zool. Sta. Neapel, xxi, 117, VII I, 2 (1913). There is much confusion in literature between three of the earlier described, larger species. The aim in growth of clavus is 16 pali, while in cyathus and smithi it is 24, the same number as is sought for in graitdis,^ which is founded on a series of 71 specimens, all markedly cornute and mostly lying unattached upon the bottom. The specimens of smithi that I have examined do not exceed 15 mm. in diameter, whereas cyathus is much larger. C. profunda, also with 24 pali, differs in the great thickness and exsertness of septa I-III and in its columella being more deeply situated and in consequence smaller. It is just conceivable that these three species are environmental variations of cyathus, but details of surface structure do not support this suggestion. All the present specimens are attached forms. Each has a relatively broad base, an even column for about two-thirds of its height ending in a spreading out, slightly 1 'John Muiray' Exp. v, 117, I 2 (1938). MADREPORARIAN CORALS 331 compressed and shallow calice; the thick column is due to an extensive deposit of corallum outside its theca. Low costae, their surface finely granular, corresponding to the cycles of septa, extend down as far as the polyps, viz. to the part where the calice enlarges. The septa vary in thickness and exsertness according to their cycles, of which I to III are approximately equal in the larger specimens; all are relatively little exsert, the largest seldom extending more than 2 mm. above the theca. All the septa are relatively smooth and thin. The trabeculae joining I-III to the columella are generally clearly visible in surface view. The aim in growth is 24 pali situated before septal cycle IV, plates very slightly thicker than the septa and relatively smooth; they are joined by trabeculae to the columella. The columella consists of relatively thin, twisted ribbons, the number of which varies with the compression and depth of the calice but is always large. The surface of the columella is upstanding and rather domed. St. 190, 315 m., largest 33 mm. high and 24:' 21 mm. in diameter of cahce, 20 pali, 80 septa, septa and costae less differentiated than in the figures of Ed. and H. but very closely resembling those of Marenzeller. The next is 34 and 22 x 18 mm. with 24 pali, and there are 7 smaller specimens and many fragments. St. 399, 141-102 m. ; a specimen 4 cm. high, calice 32 by 24 mm. much thickened by worm tubes at its base. Caryophyllia profunda Moseley. Moseley, 'Challenger' Exp. p. 138, 16 (1881); Gardiner, Trans. Roy. Soc. Edinb. xlix, 688 (1913)- St. 6, 80-140 m., 12 corallites. The largest specimen is 37 mm. high and 25 :< 19 mm. in diameter of calice, 12 septa especially prominent both in thickness and in exsertness, 20 thin plate-like pali, septa to correspond ; one specimen has calice twice as long as broad (20 x 10) with 16 pah and all costae well marked and the smallest calice is 9 mm. in diameter with 12 pali. All show a similar differentiation in septal size and the costae are in prominent cycles; the smaller specimens have relatively thicker pali. The remarks in the second reference above on the specimens from the Scottish National Antarctic Expedition apply equally well to the present specimens. All have relatively smooth septa, not heavily spined, and trabeculae joining septa I-III to the columella are seldom visible in surface view. St. 1 187, 135-134 m. A corallum which had become attached to some sohd body and grown up for 5 cm. into a bent horn, its calice 27 x 19 mm. ; it has 21 pali. A second has become attached to the side of the same and grown up for 37 cm., its calice nearly round, 17 mm. across, septa thinner and less exsert, 24 pah; were the two not joined this latter corallum might well have been ascribed to C cyathus above. Also a dead specimen attached to a pebble which has become overgrown by a polyzoon. Caryophyllia arcuata (Ed. and H.). Ed. and H., Ann. Sci. Nat., ser. 3, ix, 290 (1898); Cor. II, 16 (1857). Lacaze-Duthiers, Arch. Zool. Exp. gen. ser. 3, xv, 91-107, v (1897); Gravier, Res. Camp. Sci. Monaco, lv, 25, II 26-9 (1920). 332 DISCOVERY REPORTS St. 190, 315 m. Two broken off specimens fall within the redescription of this species by Lacaze-Duthiers, who had a suite of over 20 specimens. They are nearly round with rather shallow calices, closed by a large number of columella-rods. The paU in each number 12. The shape is cornute, and the polyp tissues extend for 3 to 4 mm. only down the outside of the corallum. Septa I and II are equal and markedly more exsert. The details of structure, especially of the pali, almost exactly resemble Lacaze-Duthiers' fig-7- Caryophyllia mabahithi Gardiner and Waugh. 'John Murray' Exp. v, 178, III 6 (1938). St. 182, 278 m., 3 specimens, one being dead and one with regenerated calice. Caryophyllia sp. St. 279, 58-67 m., one specimen the shape and size of the last. It would seem to be a stage in the growth of a form that will have 12 pali. Genus Turbinolia (Ed. and H.) Ed. and H., British Fossil Corals, xvi, 13-21 (1840); Ann. Sci. nat., ser. 3, pp. 235-40 (1848); Cor. II, 60-65 (1857). Quayle, San Diego Sac. Nat. Hist, vn, 91-110 (1932). ^ The genus is best known from the Eocene and Oligocene, the only possible recent species being T. corbicula Pour.^ dredged in the Gulf of Mexico from 100 and 220 fm. and characterized as "perhaps fossil (?)". Most species are well illustrated and show a tendency or an actual fusion of septa III to septa I ; in T. costata we suspect that Ed. and H. in the Annales (p. 239) has confused the cycles of septa, their description in Cor. II, 64, omitting all reference to this matter. The European species show the primary septa fusing with the columella, which ends in an upstanding column. The American species have these septa rising higher and running into the columella, to the surface of which they give a stellate appearance (cf. Quayle). The species below has a rather depressed columella with no style, but the primary septa rise well above the level of the theca and fall to their fusion with the columella ; it is thus almost intermediate between the two groups. In all species there is a tendency for the septa of the last-formed cycle to be represented at first only by their costal portions, a character especially marked in the species below. Turbinolia australiensis n.sp. (PI. XXI, figs, i and 2). Three specimens, 3-4 mm. high, nearly round with calices about 2 mm. in diameter, were presented to the museum at Cambridge in 1 892 together with a collection of Polyzoa dredged in or near Port Jackson, New South Wales. They have now been presented to the British Museum. About 12 costae at the base, additional costae forming higher up and corresponding to the new septa ; no special costae over the directives and all relatively smooth, thick and well marked with deep intercostal valleys. 1 See here for references to Vaughan and others up to date. 2 Bull Mus. Comp. Zool. Harvard, v, 203 (1878). MADREPORARIAN CORALS 333 Six septa (I) thicker and fused in the centre of the caUce. The next six (II) much thinner, extending about two-thirds of the way to the centre. Twelve (III) still thinner and joining septa I at about two-fifths of their breadth. Lastly, cycle IV generally represented only by costae, alternating with those of I, II and III. Upper edges of all septa rising above the theca evenly. Septa of I with markedly upstanding edges, extending from the fusion with septa III almost to the depressed columella in the centre of the calice. Genus Sphenotrochus Ed. and H. Sphenotrochus intermedius (Munster). Ed. and H., Ann. Set. nat. ser. 3, ix, 243 (1848); Cor. II, 68 (1857). Duncan, Trans. Zool. Soc. vni, 320, XLI i-s (1873).! St. 388, 121 m., 7, one alive and the rest recently dead; St. 934, 100 m., one living, 1 1 mm. high with calice 7 by 3 mm. A series of much larger coralla than those of the 'John Murray' Expeditmi (p. 192), the Hving form from St. 388 being 7 mm. high, calicle 6 by 2-5 mm., the dead forms from same larger. There are three cycles of septa complete and in the larger forms the directive septa at the ends of the elongated calice may have pairs of septa IV on either side, in one corallum even some of septa V while at the sides of the calice there are only septa I-III. The costae are well marked ; new costae may look as if they arose by branching of the lower costae, viz. those near the base of the corallum, but the terminal costae belonging to the directive septa at the ends of the flattened calices never branch and are very conspicuous. The columella is a central plate along the calice and is joined by the ends of septa I and 11; it is capped by a series of upstanding, rather rough pillars, shown best in the two living forms. Genus Deltocyathus Ed. and H. Deltocyathus lens Alcock. ' Siboga' Exp. p. 19, II 16 (1902). A single specimen from St. 939, 87 m. It is a corallum that has regenerated from a broken fragment with 7 costae. The chevron-formation of the septa is little marked. Genus Madracis Ed. and H. The genus may be branching, or incrusting, or a mixture of both. The surface is covered with calices which are closely packed and small. The septal number is ten of equal size, all fusing internally with a styliform-columella. The calices are closed in below by tabulae regularly arranged as transverse partitions, and there is no other filling up of their cavities. The above characters apply to the following species: ospenila Ed. and H., hellana Ed. and H., decactis (Lyman) Verrill, pharensis Doderlein and singidaris Rehberg, if my 1 Duncan's figures illustrate the shape and texture well. I expect Marenzeller's aurantiacus (' Valdivia ' Exp. p. 280, XVIII 15) is a synonym, but I have not a long enough series to be assured. The same remark applies to S. auritus Pourtales {' Hasslar' Exp. p. 37, 1874). 334 DISCOVERY REPORTS identifications are correct, and, as far as they are described, to various fossil species of Vaughan and others. M. kauaiensis Vaughan has lo smaller septa of a second cycle, not reaching the columella. The cyclical number of ten is most important, the newly formed calices having no intervening stage with six septa so far as can be seen from an examination of budding. The genus varies in growth-form from fine branches to upstanding nodules or branches on a completely incrusting base. The latter may show no such upgrowths and may reach a considerable thickness so that the coral would be termed massive. Finally, in the lower parts of colonies the calices may be separated relatively widely by coenen- chyma, this studded with low spines. The calices vary up to 3 mm. diameter. They may be divided by a thin sharp-edged upstanding theca or this may be quite thick and rounded, the whole perhaps flattened so as to give a depressed appearance. The presence of low, projecting rods around the top of the theca is usually a specific character, but an appearance of a few such may occur in any species. The genus, as defined above, does not allow the retention of Madracis scotiae which I described in the account of the corals of the Scottish Antarctic Expedition in 1913; this has 8 septa, but definitely passes through a six stage. Probably it should be referred to Axhelia Ed. and H. Madracis decactis (Lym.) Ver. Pourtales, ///. Cat. Mus. Comp. Zool. Harvard, iv, 67, VII 1-3 (1871); Verrill, Zool. Bermudas, I, 108 (1903). A single mass with irregular surface 6 by 4-5 cm. and 4-5 cm. high from St. 283, near Annobon Island, is referred here. Below its surface for 4 to 6 mm. the corallum is intact ; this sits upon several layers of skeleton much decayed owing to the action of boring organisms. The calices are seldom more than 2 mm. in diameter; they may be crowded together and smaller where growth is active, these having thin, thecal walls. In hollows the calicular centres are somewhat separated, the theca greatly thickened, generally marked on the surface by a single row of blunt spines. Budding occurs on the edge and on actively growing knobs where three caHces meet ; in these it is very rare to find less than 10 septa if the theca is perfect. The columella is always present, but it is seldom surrounded by a circle of paliform-teeth. The species described by Rehberg {Abh. Nat. Ver., Hamburg, xii, 10, I 3, 4, 1892) as M. singularis seems to fall within the range of decactis; it was obtained at 80 fm., at Fiji. Some calices in the present specimen are identical with those figured by Rehberg. Genus Dendrophyllia (Blainville), Ed. and H. Dendrophyllia oahensis Vaughan. Bull. U.S. Nat. Mus. no. 59, p. 154, XLVI i (1907). St. WS 244, 253-247 m. Seven small colonies are here assigned, largely on the faith of Vaughan's photographs, of which one shows the costae and another brings out the very characteristic kind of network fusion together of the septa to form a thick wall. All have a semi-upright MADREPORARIAN CORALS 335 growth ringed in places and some colonies show thin epithecal coverings. The columella varies greatly ; it may be small and situated rather deeply, especially where the edge of the calice is bent in, or may be quite shallow, extending for half the length of the slightly compressed calice where the edge is straight or spreading. Dead calices with budded off polyps on their sides, still alive, may obtain a diameter of 13 mm., but the ordinary large calice is not more than 9 mm. Genus Balanophyllia Ed. and H. Balanophyllia cornu (Moseley). 'Challenger' Rep. p. 192, text-fig. and XII 11-15 (1885). St. WS 83, 75 m., 2 attached coralla with broad bases; St. WS 93, 133-130 m., 6 coralla all living and attached, largest 15 mm. long diameter of calicular opening; St WS 243, 144-141 m. 3 cornute forms, unattached, varying up to 13 mm. diameter; St. WS 246, 267-208 m., 5 cornute forms, 3 free up to 17 mm.; St. 839, 404-424 m., I bent, free, cornute form of 22 m.; St. WS 871, 336-341 m., 4 forms, 2 free, largest 23 mm. We have before us 21 coralla of the present collection from 6 stations, 3 from the 'John Murray' collection 229 m., and the 3 ' Challenger ' types from Ki Islands, 129 fm. The latter have a dense heavy corallum with well-marked costae from calice to base and the largest, a dead specimen, has distinct trace of an epitheca. All except three of our specimens show an epitheca and some would be described as of the Thecopsammia- facies. It is best marked, thicker and with transverse growth-hnes, in the largest speci- mens, on which it most closely approaches to the calicular margin. There is a less definite edge to the theca in some specimens than in the types due to a spongy arrange- ment of trabeculae. Generally the costae are not well-marked, in no specimen as promi- nent as in the types, but then the present specimens have not the same firm coralla. The important, specific character lies in the arrangement of the septa, which Moseley has shown in his text-figure and which we have not found in other large Balanophyllia. There are no prominent exsert septa or groups of septa. Septa I and II are quite separ- able and at a deep level run into the columella. Septa III, IV and V are as shown by Moseley and any variation in any other septa is in correlation with the depth of the upstanding columella in the axial fossa, a matter which every growth-form settles for itself. It is especially deep in the two small type-specimens and is always a more or less upstanding mass of twisted ribbons, its size variable. All the polyps of our specimens as preserved were of a deep brown colour. Their surfaces and internal anatomy were irretrievably damaged. Those from St. WS 93 are recorded as having a yellow corona when alive which we think refers to the lighter peristome, rather more marked around the edge of the stomodoeum. Whether the ascription of our very variable set of specimens to B. cornu is correct or not we cannot say, but they draw attention to an arrangement of septa that may form a sub-section of the genus. We have only found it recorded elsewhere in "Doderlein's figs. 81, 82",^ 1 Mitt. Zool. Sta. Neapel, xxi, IX (19 18). 336 DISCOVERY REPORTS reputed B. italica, which must be contrasted with his figs. 79, 80 and 84 which show the more usual fusing of septa ; this species is markedly compressed and the edge of its calice arched by the exsert ends of septa I and II. Variation in the Genus Caryophyllia Lam.^ Corals of this genus may be regarded as central to the Turbinolidae, since species are described in the reports of nearly every expedition that has indulged in deep-sea collecting in whatever seas it has worked. The genus is hence cosmopolitan, and its species can obviously withstand large ranges of temperature and salinity. Fossil forms are widely distributed in suitable tertiary deposits, but have been described also from the cretaceous. While the existing species are mainly inhabitants of waters deeper than 50 fm., which is about the limit of plant life, a few species occur up to the tidal areas. Now, all modern reports on corals and many other marine animals are based on a supposition of gradual migration from shallow to deep seas. If this be the case, some changes common to the skeletons of corals might be expected, the coralla perhaps becoming thinner and more delicate while still quite suitable to the calmer, deeper and colder waters. Examining living and fossil forms, I have not found such changes within genera, perhaps owing to my having had too frequently to draw my own conclu- sions as to the ecological conditions, which the living and now fossil corals had to face. Indeed, when I study reports on the species of Caryophyllia, I find that my difficulties as to the determination of specific characters and as to the range of variation in fossil and living forms are precisely the same. The study of the oecological conditions of living forms does not help much, for the information as to the nature of the bottom where they lived, and of the temperature, salinity and movements of the overlying waters is scanty. Furthermore, the study of shallow-water forms gives the impression that corals which are detritus-feeders and possess no symbiotic algae, have a wide tolerance of fluctuation in these. Using, as we must, the criteria of anatomy, especially that of living forms, it would seem abundantly clear that many, perhaps most, fossil forms of Caryophyllia still continue to exist to-day, in spite of the millions of generations that may have intervened since their deposits were laid down. The skeleton of a coral gives no suggestion as to whether its polyp was adult, viz. 1 The reader inter alia should consuh the "John Murray' Exp. Rep. v, no. 7, 1938 in which 55 out of 511 Caryophyllia, belonging to 6 species, are illustrated. Some remarks in this report (p. 169) are here more carefully discussed after a renewed examination of the same specimens, helped by the specimens below and many types and other supposed species in the British Museum. The Agassiz Museum, Harvard University, has most generously sent me co-types of Pourtales' species. Among Duncan's types of corals, which Capt. Totton has discovered to have been deposited in the British Museum under a different name from Duncan, those from the 'Porcupine' Expedition [Trans. Zool. Sac. 1873 and 1878) include 10 "species" of Caryo- phyllia each determined on a single specimen and 8 of these described as new. From the discussion below, it will be clear that none of these can in my view be accepted. I have found Seguenza's plates of his 47 fossil forms {Mem. Real. Acad. Set., Torino, ser. 11, 1864) and certain other palaeo-zoological literature most illuminating; Ceratocyathus is a synonym. C. cyathus and profunda have been left in this Report as two distinct species, because I have no series from the original locality of the former and no information as to the environment from which my specimens were taken. MADREPORARIAN CORALS 337 reproducing or not, though with a long series a size is reached beyond which there is Httle further growth.^ Caryophyllia starts as a polyp of even diameter, developing below it a " basal plate " of corallum on which a cycle of six radiating plates, the primary septa, are deposited. These grow in height, while their sides thicken, along a circle close to their external edges, to produce a ring-like theca, the whole standing up into the base of the polyp. I have found it to be a rule in Caryophyllia — and indeed in most or all coral genera — that, in growth, spaces between septa, if sufficiently wide, are invariably filled by the formation of more septa, which at once connect with the theca, indeed often seeming to be formed on its upper edge. Thus there are progressively added, as the polyp grows, the septa of cycles II (6), III (12), IV (24) and so on. Each species of Caryophyllia aims at some definite number of septa and, as the septa at all levels of the skeleton are about equidistant from one another, the corallum has to increase its diameter. This is usually met by the assumption of a horn-like (cornute) shape, but the young polyp may have broadened its basal attachment, so as to allow gaps for septal cycles II and III before upright growth is established.^ In many species this cornute-growth may be masked by the broadening of the whole stem. This is accomplished by the deposition of corallum between and over the outside edges of the septal plates — these are the costae — so as to form a thick wall of solid corallum on the outer side of the circular thecal growth, which is otherwise unaffected; thereby the diameter of the column may be increased several times. For this deposition the polyp tissues outside the corallum are responsible. Such soft parts are easily killed, especially by silt, with the cessation of further skeletal formation, and the naked column then becomes subject to solution and the ravages of boring and other animals. The result may be that a species, which normally grows a broad, straight stem, may become quite cornute, but of course in truly cornute species there is no such thickening. Irregularities in growth may take place in all corals and a common development is the assumption of an oval or compressed shape of calice, which in Caryophyllia may be fixed as a specific character. In these species or specimens, as the septa do not become curved in any way, spaces may be left at the ends of the compressed cup. A method here is to form 2 pairs of septa of a new cycle on either side of each terminal or directive septum of cycle I. All the septa here placed then gradually approximate in size to those of existing septa and may be classified into their cycles so that the hexasymmetry is altered ; this is well seen in C clavtis (Scacchi) which usually has 16 systems.^ C. octopali 1 Cf. Flabellum in "Protandry and Senescence in Corals" {Proc. Camb. Phil. Soc. xi, 463-71, 1902) and Marine Inv. in S. Africa, 11, 115-54, 1902. F. curvatum in the present report lends support to the "size" question. ^ A growth vertical to the base of attachment is found in the early stages of all corals and is presumably correlated with feeding in the detritus-layer; it is well seen in shallow- water Paracyathus. Disc-like genera later spread horizontally, but the correlation of septal formation to available gaps formed between septa applies to all. ' The greatest irregularities occur in the details of this growth in corals, whose compression seems to be a matter of environment rather than of descent and is hence not a specific character. Even in the latter two new systems may form first at one end of a calice or even one at each end, an additional two systems appearing more slowly in later grovrth. 338 DISCOVERY REPORTS Vaughan and C. mabahithi G. and Waugh have 8 systems and no conspicuously com- pressed shape of cahce, this character here appearing to be genetical. In this consideration the existence of the pali is helpful, the rule in the genus being a series of larger septa (6, 8, 12, 16 or 24 in different species) without pali, each separated from its neighbour by three septa, the middle one of which has a palus internally — each system has four septa, viz. the dividing septum and these three. There is no great varia- tion in growth,^ but it would seem that the numbers of septa and pali are genetical in each species and that the polyp will build to this end, overcoming by a variety of means all difficulties of growth. Of the other characters of Caryophyllia the columella, which exists in all stages of growth, is formed usually of twisted, somewhat flattened rods often termed ribbons, varying in their shape and number with the form and size of the central space between the inner ends of the septa. Sections suggest that there is the formation of one column on the basal plate, additional ribbons appearing as space allows. The presence or absence of costae together with their size and spininess, the exsertness and surface characters of septa, etc., scarcely require comment as these vary greatly with the environment and consequently with the rate of growth of each corallum. An interesting feature of growth is that, with increase of size of the corallum, there is not a proportionate thickening of septa and other parts of the skeleton, although these are covered by the formative tissues. Making certain generally accepted assumptions, such as (i) an increased metabolism of the polyp in proportion to size, (2) the formation of the skeleton in proportion to such metabolism, and (3) the coupling of reproductive activities with a considerable lessening of growth, it is so difficult to understand this cessation of skeleton-deposition that our somewhat mechanical assumptions would appear unjustified. 1 Sometimes a palus fuses with its septum which then may be classified as the septum of a previous cycle ; the old system divides symmetrically and the two new systems are perfected by the formation of a pair of septa in each; the middle septum of each acquires a new palus. PLATE XX Flabellum curvatum Moseley. x i. Fig. I. Surface view. Fig. 2. Side view. Top row from St. 652. Middle row, specimen to left St. 216 and to right St. 247, the two central St. WS 795. Lower row, specimen to left St. 792, rest St. 795. DISCOVERY REPORTS VOL. XVIII PLATE XX JohnSalcSonsACumow. L^LondOTi. FLABELLUM CURVATUM PLATE XXI Turbinola australiemis, n.sp. Fig. I. Side view. Fig. 2. Surface view. DISCOVERY REPORTS VOL XVIII PLATE XXI en 35 Z u < H D < < O z S a: H DISCOVERY REPORTS Vol. XVIII, pp. 1-105, plate I Issued by the Discovery Committee, Colonial Office, London on behalf of the Qovemment of the Dependencies of the Falkland Islands COAST FISHES PART HI. THE ANTARCTIC ZONE by J. R. Norman CAMBRIDGE AT THE UNIVERSITY PRESS 1938 Price fifteen shillings and sixpence net ^\s LONDON Cambridge University Press BENTLEY HOUSE, N.W.I NEW YORK • TORONTO BOMBAY • CALCUTTA • MADRAS Macmillan TOKYO Maruzen Company Ltd All rights reserved Cambridge University Press LONDON BENTLEY HOUSE, N.W.I NEW YORK • TORONTO BOMBAY • CALCUTTA • MADRAS Macmillan TOKYO Maruzen Company Ltd All rights reserved PRI NTED [ N GREAT BRITAIN B Y WALTER LEWIS MA A T THE CAMBRIDGE UNIVERSITY PRESS DISCOVERY REPORTS Vol. XVIII, pp. 121-222, plates III-VI Issued by the Discovery Committee, Colonial Office, London on behalf of the Qovernment of the Dependencies of the Falkland Islands CRINOIDEA by D. Dilwyn John, M.Sc. CAMBRIDGE AT THE UNIVERSITY PRESS 1938 Price sixteen shillings net LONDON Cambridge University Press BENTLEV HOUSE, N.W.I NEW YORK • TORONTO BOMBAY • CALCUTTA • MADRAS Mactnillan TOKYO Maruzen Company Ltd All rights reserved P RI NTED I N G REAT BRITAIN B Y WALTE R LEWIS MA AT THE CAMBRIDGE UNIVERSITY PRESS DISCOVERY REPORTS Vol. XVIII, pp. 223-238 Issued by the Discovery Committee, Colonial Office, London on behalf of the Qovernment of the Dependencies of the Falkland Islands y- THORACIC CIRRIPEDES COLLECTED IN 1925-1936 C. A. Nilsson-Cantell, Sweden CAMBRIDGE AT THE UNIVERSITY PRESS 1939 Vrice two ^T}i\Xng% and sixpence net LONDON Cambridge University Press BENTLEY HOUSE, N.W.I NEW YOHK • TORONTO BOMBAY • CALCUTTA • MADRAS Macmillan TOKYO Maruzen Company Ltd Atl rights reserved P RI NTED 1 N GREAT BRITAIN B Y W ALTE R LEWIS MA AT THE CAMBRIDGE UNIVERSITY PRESS DISCOVERY REPORTS Vol. XVIII, pp. 239-264, plates VII-XIII Issued by the Discovery Committee, Colonial Office, London on behalf of the Qovemment of the Dependencies of the Falkland Islands THE LEOPARD SEAL HYDRURQA LEPTONYX (DE BLAINVILLE) J. E. Hamilton, M.Sc. CAMBRIDGE AT THE UNIVERSITY PRESS 1939 Price six shillings and sixpence net LONDON Cambridge University Press EENTLEY HOUSE, N. W. 1 NEW YORK • TORONTO BOMBAY • CALCUTTA • MADRAS Macmillan TOKYO Maruzen Company Ltd Alt rights reserved PRI NTED I N GREAT BRITAIN B Y WALTE R LEWIS MA AT THE CAMBRIDGE UNIVERSITV PRESS DISCOVERY REPORTS Vol. XVIII, pp. 265-322, plates XIV-XIX Issued by the Discovery Committee, Colonial Office, London on behalf of the Qovemment of the Dependencies of the Falkland Islands HYDROMEDUSAE FROM THE FALKLAND ISLANDS (the late) Edward T. Browne and P. L. Kramp CAMBRIDGE AT THE UNIVERSITY PRESS 1939 Price twelve shillings net LONDON Cambridge University Press EENTLEY HOUSE, N.W.I NEW YORK • TORONTO BOMBAY • CALCUTTA MADRAS Macmillan TOKYO Maruzen Company Ltd All rights reserved PRI NTED 1 N G REAT BRITAIN B Y W ALTE R LEWIS MA A T THE CAMBRIDGE UNIVERSITY PRESS DISCOVERY REPORTS Vol. XVIII, pp. 323-338, plates XX-XXI Issued by the Discovery Committee, Colonial Office, London on behalf of the Qovernment of the Dependencies of the Falkland Islands MADREPORARIAN CORALS, WITH AN ACCOUNT OF VARIATION IN CARYOPHYLLJA by J. Stanley Gardiner, M.A., F.R.S. \A/C300© MOLE. CAMBRIDGE AT THE UNIVERSITY PRESS 1939 Price three shillings net LONDON Cambridge University Press BENTLEY HOUSE, N.W.I NEW YORK • TORONTO BOMBAY • CALCUTTA MADRAS Macmillan TOKYO Maruzen Company Ltd Alt rights reserved P RI NTED I N GREAT BRITAIN B V W ALTE R LEWIS MA AT THE CAMBRIDGE UNIVERSITY PRESS DISCOVERY REPORTS Issued b^ the Discovery Committee, Colonial Office, London on behalf of the Qovemment of the Dependencies of the Falkland Islands Vol. XVIII, pp. i-vi TITLE-PAGE AND LIST OF CONTENTS CAMBRIDGE AT THE UNIVERSITY PRESS 1940 Frice ninepence net \ma^^ CAMBRIDGE UNIFERSITT PRESS LONDON.- BENTLEY HOUSE NEW YORK, TORONTO, BOMBAY CALCUTTA, MADRAS: MAC MILL AN TOKYO: MARVZEN COMPANY LTD All rights reserved PRINTED JN GREAT BRITAIN BY WALTER LEWIS MA AT THE CAMBRIDGE UNIVERSITY PRESS '4'Vt '^ . ' 'V^'- -f. ^ > 4;- t^f-^'^ ^'•^•: 1 -■■■■■ V*».-'«M, ■ V - » f . \^- ■■■>-. c>- * ,* ;,^--^j;- .h. % <-f .t^' .,4 •■ c.:^,^**t"-^^<; .'K >