M. J. PENRITH

THE FISHES OF TRISTAN DA CUNHA, GOUGH ISLAND AND THE VEMA SEAMOUNT

March 1967 Maart Volume 48 Band

Part 22 Deel

ANNALS OF THE SOUTH AFRICAN MUSEUM

ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

Cape Town Kaapstad

The ANNALS OF THE SOUTH AFRICAN MUSEUM

are issued in parts at irregular intervals as material becomes available

Obtainable from the South African Museum, P.O. Box 61, Cape Town (Cash with order, post free)

Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

word uitgegee in dele op ongereelde tye na beskikbaarheid van stof

Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad (Kontant met bestelling, posvry)

OUT OF PRINT/UIT DRUK

1, 2(I, 3> 3D» 7-8), B25 5 is Dele)s 5(2, 5> 7-9), 6(1, t-pi.), 7(1, 3), 8, g(1-2), 10(1-3), ULES 250 75) tele) ual, 24. (2) Sana (8) un Osea (Ay)

Price of this part/Prys van hierdie deel

60c

Printed in South Africa by In Suid-Afrika gedruk deur The Rustica Press Pty., Ltd., Die Rustica-pers Edms., Bpk., Court Road, Wynberg, Cape Courtweg, Wynberg, Kaap

THE FISHES OF TRISTAN DA CUNHA, GOUGH ISLAND AND THE VEMA SEAMOUNT

By M. J. PENRITH* South African Museum, Cape Town

(With 2 text-figures and 1 plate)

CONTENTS

PAGE Introduction EMT ier ME 23 Material : : : : Bi) Gea Systematic account . BMD ea 20 Zoogeography . . ‘ fs Gyan Summary . : : 3 - 545 Acknowledgement ; : 2) 546 IReferencesiin yale ien : - 546

INTRODUCTION

In the south-east Atlantic Ocean there are three known areas of very shallow water (not counting the African coast). Two of these are around the shores of islands, the Tristan da Cunha group and Gough Island, while the third is the Vema Seamount.

Tristan da Cunha

The island of Tristan da Cunha, together with the two smaller islands of Inaccessible and Nightingale, lies almost midway between South Africa and South America (37° 05’ S., 17° 40’ W.). It is permanently inhabited and has been the subject of several scientific expeditions and reports.

Capt. Dugal Carmichael, F.L.S. (army surgeon to the garrison in the island, 1816-17) was the first to publish a report on the flora and fauna of Tristan. In this paper (1818) he described four species of fishes from the island and listed five others as being common to both Tristan and the Cape. Car- michael’s work was referred to by Regan (1913a) but no further fishes were added to the nine already known from the island until Barnard (1923, 1925) listed two more species which he considered to be conspecific with South African species. Norman (19352) described a further species believed by him to be endemic to the island group. After the Norwegian expedition of 1937-8, Sivertsen (1945) could add a further seven species to the fish fauna of the island, two of them being new species and believed by him to be endemic to Tristan da Cunha. Finally, Rowan & Rowan (1955) added three more records of oceanic fishes from the island.

* Seconded from the Council for Scientific and Industrial Research Oceanographic Unit, University of Cape Town.

923 Ann. S. Afr. Mus. 48 (22), 1967: 523-548, 2 figs., 1 pl.

524 ANNALS OF THE SOUTH AFRICAN MUSEUM

At present the known fish fauna of Tristan consists of 19 species which Sivertsen (1945) divided into an oceanic component and a coastal component. (In the list below, the first reference to the species occurring at Tristan is given in parenthesis after the author of the species.) The islanders’ common name for the species is also given where known.

(a) Oceanic species Prionace glauca (Linn.) (Sivertsen, 1945) . . . . Blue shark Maurolicus muelleri (Gmelin) (Barnard, 1925) sna a Myctophum humboldti (Risso) (Sivertsen, 1945) . . — Scomberesox saurus (Walbaum) (Sivertsen, 1945) . . — Exocoetus exiliens Bloch (Carmichael, 1818) . . . a Alepisaurus ferox Lowe (Rowan & Rowan, 1955) — Notopogon lilliei Regan (Sivertsen, 1945) . . . . —

Seriola lalandi Valenc. (Rowan & Rowan, 1955) . Yellowtail Polyprion americanus (Bloch) (Rowan & Rowan, 1955) Steambrass Thyrsites atun (Euphrasen) (Carmichael, 1818) . . Snoek

(b) Coastal species Ariosoma australis (Barnard) (Barnard, 1923). . . —_ 1Gaidropsarus insularum Sivertsen (Sivertsen, 1945). . _- Acantholatris monodactylus (Carmichael) (Carmichael,

TOES) te). tds ok I OSC a es ie eee hae ehimoer Bovichthys diacanthus (Carmichael) (Carmichael, 1818) Klipfish 2Decapterus longimanus Norman (Norman, 1935a) . . Mackerel Labrichthys ornatus (Carmichael) (Carmichael, 1818) Concha or Conger 3Helicolenus tristanensis Sivertsen (Sivertsen, 1945) . Soldier

Sebastichthys capensis (Gmelin) (Carmichael, 1818) . Soldier Sertolella antarctica (Carmichael) (Carmichael, 1818) Bluefish Seriolella christophersent Sivertsen (Sivertsen, 1945) . Stumpnose

The identity of the fish recorded by Carmichael as the ‘Roman fish’ Sparus , is not known. The South African ‘Roman’ is a Sparid (Chryso- blephus laticeps) but has not been recorded from Tristan.

Gough Island

Gough Island lies slightly farther south than Tristan (40° 20’ S., 9° 55’ W.), and is not permanently inhabited, although the South African Government has maintained a weather station on the island for several years.

Although several expeditions have called at the island, it has not been subject to the same degree of collecting as has Tristan. Only one expedition, the Gough Island Scientific Survey 1955-6, has spent much time on the island, but they have not published any data on the fishes obtained.

1 As G. capensis (Kaup) by Barnard (1925). * Probably the fish recorded by Carmichael as Scomber trachurus. 8 Recorded by Carmichael as Sebastichthys maculata.

THE FISHES OF TRISTAN, GOUGH AND THE VEMA SEAMOUNT 525

Accounts of fish collected by the Scotza, which called on the way back from the Antarctic (Regan, 1913b) and by the R.V. William Scoresby (Norman, 19350), have been published.

The fishes recorded are:

1Helicolenus tristanensis Sivertsen (Regan, 1935)) Sebastichthys capensis (Gmelin) (Regan, 19136)

Bovichthys diacanthus (Carmichael) (Regan, 19136) Caesioperca coatsi Regan (Regan, 19136)

Acantholatris monodactylus (Carmichael) (Norman, 1935))

Holdgate (1958), in a popular account of the Gough Island Scientific Survey, listed the following species of fish under their Tristan common names: five- fingers, soldiers, klipfish, congers.

Vema Seamount

This exceptionally shallow seamount was discovered in 1957 by the R.V. Vema, and traverses were made then as well as in 1959 by the Vema and in 1963 by the R.V. Robert D. Conrad. In November 1964 a detailed survey of the peak was undertaken by the South African National Committee for Oceanographic Research, using the diamond prospecting tug, Emerson K (Simpson & Heydorn, 1965). The peak was found to have a position 31° 38’ S., 8° 20’ E. and a least depth of only 14 fathoms, with a well-defined plateau about five square miles in extent and of approximately 35 fathoms depth. Subsequent to the visit by the Emerson K there have been many visits by com- mercial fishing vessels from Cape Town as well as a short visit by the R.V. Africana II in April 1965.

MATERIAL

The material used for the present paper was obtained from several sources. The South African Museum has been slowly building up a collection of Tristan fish for many years, mainly obtained as donations from Mr. Keytel (1907), the rock-lobster survey of 1949, and small but valuable donations from Tristan and Gough Islands by the Division of Sea Fisheries and Capt. M. T. Scott, master of the rock-lobster fishing vessel, Tristania.

Specimens of fishes from the Vema Seamount were obtained from the Division of Sea Fisheries (those from the Emerson K collection, and commercial fishing vessels, and a collection made from R.V. Africana II by Mr. A. C. Paterson), as well as directly from the fishing companies.

These collections have resulted in several further species being added to the fish fauna of Tristan da Cunha and Gough Island as well as adding to our knowledge of the species already recorded from the islands, and in addition have enabled their fauna to be compared with that from the Vema Seamount.

1 Recorded as Sebastes maculatus C. & V.

526 ANNALS OF THE SOUTH AFRICAN MUSEUM

SYSTEMATIC ACCOUNT Exocoetidae

Cypsilurus lineatus (Cuvier & Valenciennes) Cypsilurus lineatus (Cuvier & Valenciennes) Bruun, 1935: 47.

Only one specimen, of 382 mm standard length, was examined. The fish was found in a drum of unsorted material from Tristan in the Department of Ichthyology, Rhodes University.

The body proportions as thousandths of standard length are as follows: pectoral fin 675, pelvic fin 309, dorsal height 111, depth 157, head 199, pre- dorsal distance 728, preventral distance 560, preanal distance 754, snout 50, eye 65. Fin counts are dorsal 12, anal 10 and pectoral 15.

This is an extremely large specimen of the species.

Gadidae

Gaidropsarus Rafinesque

Kaup (1858) described a small gadid as Motella capensis but gave only a brief description and no locality, although on the basis of the trivial name it is believed to be South Africa.

Small rocklings were described from St. Paul Island as M. capensis by Kner (1868) and Sauvage (1879), while Barnard (1925) published a description of material in the South African Museum which he also assigned to Kaup’s species.

Sivertsen described specimens from Tristan and showed that while they were very similar to those described from St. Paul Island, they showed marked differences from the South African specimens as described by Barnard. Since South Africa was believed to be the type locality of capensis, Sivertsen proposed a new name, G. insularum for the Tristan and St. Paul Island specimens.

On re-examining the material used by Barnard, I find that only one of the two fish he examined exactly fits the description he published, and that the two fish he examined appear to belong to two distinct species. This view is shared by Mr. A. Wheeler of the British Museum (Natural History), who has also examined the South African Museum material, and Professor J. L. B. Smith, who informs me that he has reached a similar opinion from examination of material in his collection (Table 1).

TABLE I Comparison of Gaidropsarus

Table 1a gives body proportions of two species of Gaidropsarus in millimetres and as thousandths of standard length in parenthesis.

Table 15 compares body proportions of Gaidropsarus from various sources.

12719 G. insularum from False Bay

12528 G. capensis. Port Elizabeth 7858 G. capensis. East London LN G. insularum from Tristan da Cunha (from Sivertsen, 1945) B G. capensis from South Africa (Barnard, 1925)

THE FISHES OF TRISTAN, GOUGH AND THE VEMA SEAMOUNT 527 TABLE Ia 12719 12528 7858

Standard length 153 gI 156

Depth 22-5)(147)| U5 (165) 927 (473)

Head 29 (190) 2075 (226) 33 (212)

Eye . 4°6 (30) 3°38 (42) 5:3 (34)

Snout Se 3 See (52) 6:8 (75) 9°3 (60)

Interorbital . Gay Bes 6 (39) AN (42) 4°5 (29)

Maxilla . F 3 : 13 (85) ON (TO) 917722) (iT)

Basevor firstidorsall 2). 953) = 7(85) 2 r2) (132) 19 (122)

Base of anal . ; - 77 (500) 44 (485) 76-5 (490)

TABLE 15 A 12719 B 12528 7858

Depth/length 63-7 63 6 6 6 Head/length Je 0) AG SS 5 43 43 43 ye eaG meer nih ye OR 7, a 5 4 64 Eye/interorbital . . 1 -1t Tt 91 +Y 1} and Pelvic ray/Origin ray to cone 2 2 3 3 3 Caudal peduncle/head . . 22-24 2} — 34 34 Rays, first dorsal : 47-49 +45 55 65 56 Rays, second dorsal 62 — 64 62 45 44 43 Rays, anal fin 52-54 50 4.0 37 39

One specimen from Kalk Bay in the Cape Peninsula as well as some juveniles from Table Bay and Lambert’s Bay fitted the description of the Tristan fishes as given by Sivertsen and the St. Paul Island fishes as described by Kner and Sauvage, while the specimens from East London and Port Elizabeth are Clearly different, and are assumed to be G. capensis.

Gaidropsarus insularum Sivertsen

Motella capensis (non Kaup) Kner, 1868: 279. Sauvage, 1879: 42. Gaidropsarus capensis (non Kaup) Barnard, 1925: 323 (partim). Blanc, 1961: 145. Blane & Paulian, 1957: 327- Gaidropsarus insularum Sivertsen, 1945: 8. Recorded from Tristan da Cunha, St. Paul Island, the Cape Peninsula and west coast of South Africa. Body proportions and fin counts given in Table 1.

Gaidropsarus capensis (Kaup)

Motella capensis Kaup, 1858: 90. Gaidropsarus capensis (Kaup) Barnard, 1925: 323 (partim).

Depth 6 in standard length, head about 434. Eye equal to interorbital about 6 in head. Second pelvic ray 2 distance origin pelvic to vent. Caudal peduncle 34 in head. Fin counts: Free dorsal rays 55-65, Dorsal rays 42-45, Anal rays 36-40.

Description based on two specimens, one of 91 mm standard length

528 ANNALS OF THE SOUTH AFRICAN MUSEUM

deposited in the South African Museum (Reg. No. SAM 12528), collected in Algoa Bay at a depth of 25 fathoms by the Preter Faure, and a second of 156 mm standard length in the collection of the Department of Ichthyology, Rhodes University, Grahamstown (Reg. No. 7858), collected at East London.

Body proportions and fin counts are given in Table 1.

G. capensis differs from znsularum in having a relatively longer first dorsal base (12—13°5% of body length as opposed to about 8-5%), a longer head, larger eye, slimmer caudal peduncle and rather marked differences in fin counts.

Berycidae Beryx decadactylus Cuvier

Beryx decadactylus Cuvier. Fowler, 1936: 542.

One specimen obtained from Captain Scott, caught in “Tristan waters’. Tears in the mouth suggest that the fish was caught with hook and line but depth of capture is unknown.

The fish is 362 mm standard length and has the following body proportions (expressed as thousandths of standard length): head 359, depth 472, eye 152, maxilla 185, base of anal fin 348, base of dorsal fin 290. Fin counts are dorsal IV 19, anal IV 27, pectoral 15, pelvic I ro.

B. decadactylus has not been recorded from South Africa although the related B. splendens is trawled quite regularly. In the course of this work an unrecorded specimen of B. decadactylus was found in the collection of the South African Museum, trawled off Table Bay in 1948. It was 345 mm in standard length and had the following body proportions (in thousandths of Ls): head 322, depth 446, eye 125, maxilla 174, base of anal fin 359, base of dorsal fin 293. Fin counts are dorsal IV 20, anal IV 27, pectoral 15, pelvic I ro.

Macrorhamphosidae Notopogon lillier Regan

Notopogon lilliei Regan, 1914: 14. Mohr, 1937: 50.

Sivertsen (1945) records two specimens of Notopogon which were found washed up dead on Tristan. He assigned them to NV. lilliei, a species known from South Africa, Australia and New Zealand. It should be noted, however, that his published figure (fig. 5) resembles V. macrosolen Barnard rather than WN. lilliet, especially as neither the figure nor the text makes any reference to the dorsal brush which is a conspicuous feature of NV. lilliei but is absent in WN. macrosolen.

Chilodactylidae

Sivertsen (1945) recorded only one species of fish belonging to the family Chilodactylidae, Acantholatris monodactylus (Carmichael) ; called by the islanders the ‘Fivefinger’.

THE FISHES OF TRISTAN, GOUGH AND THE VEMA SEAMOUNT 529

He claimed that the islanders spoke of a second fish very like the fivefinger which they were said to call the ‘yellowtail’. Rowan & Rowan (1955) have shown that the fish called yellowtail is Sertola lalandi. No additional species of chilodactylids have been received from Tristan but a species which appears to be new has been obtained from Vema Seamount.

An additional species which appears to be undescribed has been obtained from Gough Island, but is not described here as only one specimen was obtained. It is known to fishermen as the ‘funny fish’.

Acantholatris monodactylus (Carmichael) (Plate XXIs)

Chaetodon monodactylus Carmichael, 1818: 500. Cheilodactylus carmichaelis Cuvier & Valenciennes, 1830: 360. Kner, 1868: go. Chilodactylus monodactylus (Carmichael) Sauvage, 1879: 23; Regan, 19134: 466. Norman, 1935):

ne monodactylus (Carmichael) Gill, 1863: 119. Sivertsen, 1945: 10.

This species is known from Tristan and Gough Islands and has also been found to be common at the Vema Seamount. Gunther (1860) recorded this species from Chile but Norman (19350) has shown this to be a misidentification, the Chilean species being the related A. gayi (Kner). In Table 2 below are given body proportions and counts of several examples of the species from Tristan da Cunha, Gough Island and the Vema Seamount.

TABLE 2 Body proportions and fin counts of Acantholatris monodactylus in thousandths of standard length. Origin of the specimen is: Cat. No. SAM 10500 Tristan da Cunha 10501 ‘Tristan da Cunha 11764 No history, but register says ‘Very likely Tristan’ 21284 ‘Tristan (perhaps!!) 22890 Gough 24288 Vema Seamount 24291 Vema Seamount Cat. No. SAM 11764 11784 11764 10500 10501 10501 21284 10501 24291 24288 22890

Standardlength. 1391 153 153 #175 227 4238. 301. 353'' 338 443 444

ead chen STA 220m, STON NS0QRagoT 1° 996n) Neb GNaeh Ti) Son Toon og6 Depth. .... 405 385 405 389 379 391 396 391 370 381 405 EV CMe et ani) OA: 78 82 75 75 76 67 62 62 59 56 Interorbital. . . 69 78 76 75 84. 92 93 88 74 79 88 Base of dorsal. . 656 620 634 657 625 630 638 636 636 663 °~ 605 BASeIOmAHAlba my LO3/ O39) al7 70 072/154, TOO) 1153 147) ) | TAG ny US Shin 9109

Length of pectoral 351 340 336 349 365 353 349 © 306 328 309 ©6320 Snout to dorsal

origin .... 351 334 329 315 339 344 379 360 343 350 334 Snout to anal

Sliciteem re O70) 700.) 715. "680". 693), 705 687 97679) 674) 665) 736 Snout to pectora

origin... . 313 308 323 303 334 332 342 340 332 332 . 356 Dorsal spines . . 17 16 17 17 16 17 17 17 17 17 17 Dorsalrays. . . 25 25 25 2 26 25 26 25 2 25 22 Anal spines... 3 3 3 3 3 3 3 3 3 3 3 AABN ee 12 II 12 II II II II II II II 10

Gillrakers . . .7+14 8+15 84+15 8+14 8415 7+15 7+15 Has)

530 ANNALS OF THE SOUTH AFRICAN MUSEUM

One specimen (SAM 22890) obtained by Captain Scott at Gough Island has been assigned to A. monodactylus although the fin counts are lower than normal for the species. The body proportions fit within the range for A. mono- dactylus however and coloration in formalin is identical.

Acantholatris vemae sp.n.

(Plate XXIa, Table 3)

MATERIAL

Two specimens from the Vema Seamount collected by the Emerson K, November 1964, donated by the Division of Sea Fisheries, Sea Point (SAM 24290, 24292).

Five specimens (all gutted) from the Vema Seamount collected by com- mercial fishing vessels and donated by Messrs. Friedman & Rabinowitz (Pty.) Ltd. (SAM 24289, 242093, 24296-8).

Two specimens from the Vema Seamount collected by Mr. A. C. Paterson, April 1965, and donated by the Division of Sea Fisheries (SAM 24525, 24527).

Depth of body about 3, length of head about 3? in standard length. Eye diameter 5 in length of head, 2 in snout and 1} in interorbital width.

Dorsal fin XVII 24-26, the eighth spine longest, about equal to snout, 4 in body depth. Last spine 24 in first ray. Base of dorsal a little over 14 in body length, distance snout to origin dorsal 3} in body, about equal to length of pectoral.

Anal fin III 10-11, base short only } of dorsal base. Second spine very stout, only a little longer than third, shorter than first ray.

Pectoral having g branched, 6 simple rays, first simple ray elongate, + body length, free portion 3} in length of ray.

Forehead scaled to level of posterior nostril, both nostrils of similar size, round.

Scales fairly large, very thin, lateral line 55-58, transverse 7g, cheeks scaled. Gill rakers, upper 6—7, lower 14-16, total 21-23.

Colour: in formalin, dark blue-black dorsally; fading to silver below.

Edge of fins dark, caudal with a few dark bands.

In life said to be bright blue.

Detailed body proportions of the seven specimens are given in Table 3.

Catalogue No.: 24289 24290 24292 24293 24296 24297 24298 24525 24527

Standard length . . 327 320 385 249 372 338 392 384 383 IBKSEXG! Gio olde Since 288 271 260 277 255 264 263 266 253 Depth ...... 345. 347 343 333 352 329 349 3597 agon EY CRIM sy-N--occt che 58 57 52 56 54 53 51 50 55 Imterorbitall = 2. 80 75 75 76 75 71 79 78 78 Base of dorsal . . . 624 660 673 650 683 663 666 690 682 Bascohanalies een LOS 161 174 169 172 178 168 174 175

THE FISHES OF TRISTAN, GOUGH AND THE VEMA SEAMOUNT 531

TABLE 3 (continued)

Catalogue No.: 24289 24290 24292 24293 24296 24297 24208 24525 24527 Length of pectoral . 327 328 294 321 328 34.0 319 354. 329 Snout to dorsal origin 306 306 288 297 272 278 283 294 287 Snout to anal origin 686 627 607 650 650 642 654.

Snout to pectoral. . 303 278 299 273 253 255 268 263 264. Dorsal spines ... 17 17 17 107 17 17 17 Worsalerayss — es 2) 24: 25 26 24. 25 24. 25 Amalespines) soe). | 3 3 3 3 3 3 3 AVAUTENE! Gol & ola | UO 10 Il II II II II

Body proportions (in thousandths of standard length) of Acantholatris vemae sp.n. from the Vema Seamount.

Dates of collection of the various specimens given in the text.

SAM 24527 the Holotype.

The specimen deposited in the South African Museum (Reg. No. SAM 24527) is designated as the type.

A. vemae is very close to A. monodactylus (both having 9 branched and 6 simple pectoral rays) which is also found on the Vema Seamount. It differs, however, in colour and colour pattern, lacking the six dark vertical bars found in monodactylus; also it has a markedly shorter head and slightly less deep body when compared to monodactylus. As a result of the shorter head, the distances of the snout to the origin of dorsal fin, the origin of anal fin and the origin of pectoral fin are proportionately less. The slope of the forehead in vemae too is steeper, and whereas it has the upper corner of the pectoral fin angular it is rounded in monodactylus. Finally it has much thinner lips than monodactylus.

It differs from A. gayi and A. aspersus in the pectoral not reaching the soft anal.

Latris lineata (Bloch & Schneider)

Cichla lineata Bloch & Schneider 1801: 342. : Latris hecateia Richardson, 1839: 99. Richardson, 1842: 106. Gunther, 1860: 86. Kner, 1868: 95. Sauvage, 1879: 17. Angot, 1951: 19. Blane & Paulian, 1957: 331. Blanc, 1961: 152.

Two specimens, SAM 22623, 22891, both collected by Captain Scott at Gough Island. The former is a skin with standard length of about 900 mm and the latter a complete specimen of 904. mm.

The body proportions of the 904 mm specimen (as thousandths of L’s) are as follows: head 324, depth 303, pectoral fin 192, snout 149, base of dorsal fin 638, base of anal fin 207.

Fin counts are dorsal XVIII 35, anal III 26, pectoral 9 + 9.

Counts of the skin are dorsal XVIII 36, anal III 26, pectoral 9 + 9, gill rakers 8 + 16.

L. lineata was originally described from Australia and New Zealand, but has been recorded from St. Paul Island by Kner (1868), Sauvage (1879) and Angot (1951). This is the only species that is without doubt common to the Atlantic islands and Australasia.

532 ANNALS OF THE SOUTH AFRICAN MUSEUM

Serranidae Polyprion americanus (Bloch & Schneider)

Amphiprion americanus Bloch & Schneider, 1801: 205. Polyprion americanus (Bloch & Schneider) Barnard, 1925: 488. Smith, 1949a: 199. Blanc, 1961: ay ae Gunther, 1859: 169. Angot, 1951: 15.

As noted above, Sivertsen listed two species of fish known to the Tristan islanders by common names that he was unable to identify due to lack of specimens. The fish called by the islanders the steambrass was said to be a very large fish, nearly as large as a shark (+ 2 metres) ‘but much broader and with soft rays in the fin’. Rowan & Rowan (1955) record an 8-foot long speci- men of Polyprion americanus from Tristan and state that it is the species the islanders refer to as the steambrass.

Specimens were received from the Vema Seamount. Two examples from the Seamount of 465 and 602 mm standard length were found to have the length 22— 3 times depth and 22— 23 of head length. In both the pectoral fin was slightly greater than half the head, while the longest dorsal spine went 24— 24 into head. The species occurring at Vema is thus clearly P. americanus rather than the related P. prognathus.

Ephinephelus aeneus (Geoffroy St. Hilaire)

Serranus aeneus Geoffroy St. Hilaire, 1809: 317. Fowler, 1936: 756. Epinephelus aeneus (Geoffroy St. Hilaire) Norman, 1935): 9.

One specimen (SAM 24299) presented by Messrs. Friedman & Rabino- witz, Ltd., one of several landed by commercial fishing vessels from the Vema Seamount.

Body proportions as thousandths of the standard length of 465 mm are as follows: head 402, depth 331, length pectoral fin 219, height longest (fifth) dorsal spine 123, eye 62, interorbital 64, snout 114, pelvis 172.

The cheeks are scaled with cycloid scales, as is body above lateral line; scales ctenoid below lateral line. Caudal slightly rounded. Bottom opercular spine posterior to top spine. Pre-operculum serrated, angle produced with very long serrations, operculum acutely pointed.

Mandible with two rows of teeth.

Colour in formalin: Dark brown above shading to lighter below. Sparse scattered very dark brown spots. Unpaired fins dark edged, paired fins light grey-brown. Three dark stripes on face; one from posterior of eye to lowest opercular spine, one from below eye to edge of operculum opposite base of pectoral fin and the third above maxilla to pre-operculum below the angle.

Carangidae Seriola lalandi Cuvier & Valenciennes

Seriola lalandi Cuvier & Valenciennes, 1833: 208. Barnard, 1925: 555. Smith, 19494: 221. Rowan & Rowan, 1955: 129. Seriola pappei (Castelnau) Smith, 1959: 256.

THE FISHES OF TRISTAN, GOUGH AND THE VEMA SEAMOUNT 533

One specimen was donated by the Division of Sea Fisheries, caught at the Vema Seamount. It appears to be a very common fish there, as one com- mercial fishing vessel returned with over a thousand large specimens.

It has also been recorded from Tristan da Cunha (Rowan & Rowan, 1955).

Decapterus longimanus Norman Decapterus longimanus Norman, 1935@: 255.

Two specimens from the Vema Seamount (SAM 24294) presented by the Division of Sea Fisheries. The species was formerly considered endemic to Tristan da Cunha.

The two specimens are the smallest yet recorded being 215 and 225 mm in standard length. Body proportions of the 215 mm specimen in thousandths of standard length are: head 288, depth 204, pectoral fin 260, base of soft dorsal 400, base of anal 335, gillrakers 44, scutes 42.

Measurements could not be made on the other specimen as it was rather damaged, having been obtained from the gut of a larger fish.

Emmelichthyidae

Plagiogeneion rubiginosus (Hutton)

Therapon rubiginosus (Hutton) 1876. Plagiogeneion rubiginosus (Hutton) McCulloch, 1914. Barnard, 1927. Smith, 1949a.

A single specimen was donated by Messrs. Friedman & Rabinowitz (Pty.) Ltd. The fish was about 330 mm in standard length and was caught by handline on the Vema Seamount. Unfortunately the fish had been cleaned and somewhat distorted in freezing so that accurate measurements were not possible. It could readily be distinguished from the related P. macrolepis by its lateral line count of 68 as opposed to about 50 in macrolepis.

P. rubiginosus 1s recorded from Australasia and St. Paul Island, is not uncommon off the east coast of South Africa (Smith, 1949) and. a single specimen has been recorded off Table Bay (Barnard, 1927).

Pentacerotidae

Pentaceros richardsoni Smith

Pensxercs richardsoni A. Smith, 1844 (Plate XXI and 2 pages unnumbered text). Follett & Dempster, 1963: 315. Smith, 1964: 572.

Pentaceros knert Steindachner, 1866: 208.

Pseudopentaceros richardsoni (Smith) Barnard, 1927: 621. Smith, 1949a: 242.

Griffinetta nelsonensis, Whitley & Phillips, 1939: 233.

Only a partial bibliography is given here; a more complete one is given by Follett & Dempster (1963). It should be pointed out, however, that in Follett & Dempster the references to Andrew Smith’s original description are incorrect. The plates each with a page or more of unnumbered text were

534 ANNALS OF THE SOUTH AFRICAN MUSEUM

published separately between 1838 and 1847 but were not dated while only the index and title page were dated and published in 1849. (See Waterhouse, 1880, and Barnard, 1956.)

Pentaceros richardsoni was for long thought to be an extremely rare species of fish, being known from the type, an adult caught off Cape Point, South Africa, on a deep handline, and a few juveniles from Australia and New Zealand. Recently, however, many medium-sized fish have been obtained from the North Pacific (Follett & Dempster, 1963, give a complete set of published records) while Smith (1964) has described in detail a second adult fish from the Cape of Good Hope, and pointed out that the fish described from Cape Horn by Steindachner as P. kneri is clearly a juvenile of P. richardsoni.

A juvenile of P. richardsoni obtained at Tristan was presented to the Museum by Captain Scott. Unfortunately there are no data as to its method or depth of capture. Measurements and counts of this juvenile are given in Table 4, and compared with published measurements and with those of two specimens in the South African Museum; a large adult specimen trawled off Cape Town in 300 fathoms during March 1961 (SAM 23076) and a slightly smaller speci- men trawled off Cape Columbine in April 1965 (SAM 24541) both presented to the Museum by Messrs. Irvin & Johnson, Ltd.

TABLE 4 Standard length . : , , SN 2o0r 24.07 254? 808 4514 358° Head F : : : : ‘ - 325 325 319 338 321 319 Depth) ‘ : ; : : - 413 433 422 4.25 418 4.22 Eye : 3 . : : : Sy 79 83 100 gI 76 Interorbital distance. : : 5 ee TL 118 135 109 106 Least depth caudal peduncle : 5 OB 100 102 112 106 109 Dorsal spines . : : : ‘ : 14 14 14 14 14 14 Dorsal rays. : : : : . 8-9 9 9 10 9 10 Anal spines. : : : : ; 4 4 4 4. 4 4 Anal rays 2 : : : : E 7 7 7 8 8 7 Pectoral . : : . : . 17-18 17 17 13 17 18 Gillrakers total 4 : : : a 103 22-25 22-94 = = — ateralsline:scalesie. |). ae 68-69 71 76 —= 86 =

Body proportions (in thousandths of standard length) and counts of Pentacerosrichardsoni from various localities.

1 Japan (Abe, 1957).

2 North Pacific (Welander et al., 1957).

5 Tristan da Cunha (Reg. No. SAM 21791).

4 Cape of Good Hope (Reg. No. SAM 23076). ° Cape of Good Hope (Reg. No. SAM 24541).

P. richardsont is probably found throughout the world in the warm temper- ate waters, but the published records of its occurrence are confined to a few scattered localities. At most of these places it has been found more than once. Both Follett & Dempster (1963) and Smith (1964) suggested reasons for this, but it seems likely that the known distribution of the species bears no relation

THE FISHES OF TRISTAN, GOUGH AND THE VEMA SEAMOUNT 535

to the true distribution but is due mainly to the widely scattered areas of intensive fishing and the introduction of new fishing techniques in various areas.

Thunnidae

Thunnus spp.

Several species of the genus Thunnus have been landed by commercial fishing vessels from the Vema Seamount. Unfortunately all specimens have been brought ashore gutted and without heads or fins. Many small tuna (about 50—75 lb. whole weight) had bright yellow peduncle keels which is characteristic of the southern bluefin (7. maccoyi Castelnau) (known from South Africa and Australia), while others were stated by the fishermen to be yellow- fins (T. albacares (Bonaterre) ) and bigeyes (7. obesus (Lowe) ). Mr. A. Heydorn (Division of Sea Fisheries) (fers. comm.) informs me that he saw a yellowfin while diving on the Vema Seamount.

Until whole examples are obtained nothing definite can be stated con- cerning the species of tuna congregating about the Seamount.

Stromateidae Palinurichthys antarcticus (Carmichael)

Perca antarctica Carmichael, 1818: 501.

Seriollella antarctica (Carmichael) Regan, 19134: 467.

Seriolella antarctica (Carmichael) Sivertsen, 1945: 21.

Palinurichthys porosus (Carmichael) Barnard, 1948: 395. Smith, 1949a: 304. Mupus perciformis (Mitchill, 1818: 244). Smith, 1949): 843.

The Tristan bluefish was first described and figured by Carmichael (1818). It was not discussed by Regan in his review of the Stromateidae (1902), but in a later paper Regan (1913) reviewed the fishes described by Carmichael and suggested, but without giving reasons, that antarctica was ‘rather closely related to Sauvage’s Seriollella velaini from the Island of St. Paul’.

Sivertsen (1945) placed antarctica in the genus Seriolella, but again no reasons for doing so were given; nor was the generic definition modified. According to Regan (1902) the genus Seriolella had inter alia the following characters: a fin formula of VIII 27-40; anal III 19-24, and a lateral line concurrent with the dorsal profile. The fin formula of antarctica on the other hand is dorsal TX 18-20, anal JII 14-15, and the lateral line becomes straight before the caudal peduncle.

The correct genus for antarctica would thus appear to be Palinurichthys Bleeker 1859.

A specimen of P. antarcticus was donated to the Museum by the Director, Division of Sea Fisheries, after it had died while on display in the Sea Point Aquarium. The fish was unquestionably a Tristan bluefish although actually obtained from Gough Island. When compared with other Palinurichthys material in the Museum collection it showed a marked resemblance to two similar sized fish in the collection recorded by Barnard (1948) as P. porosus (Richardson).

536 ANNALS OF THE SOUTH AFRICAN MUSEUM

TABLE 5 24532519573 al 955 een 2 33 ,

Standard length’). =) 7: 884 436 453 432 35 34. Head Hane aoa 339 351 356 358 370 380 Deyo Cabra etn ect cet: 347 344, 335 358 400 380 Length anal fin base. . IQI 222 212 226 230 235 Diamecteney.ch-mlnem enue 75 66 75 2 103 103 Length pectoralfin . . — 266 274 306 260 265 Wengthypelvicitim se) 152 160 166 158 245 235 Snout TL PAD NATE) He 105 88 95 102 86 88 IVGascil ae Siete tay pc 153 138 150 150 145 147 Pectoral eraysites ae ey ae QI QI QI 20 — —

IDYoall Gomavesg Logo Bape VSSew Ser Baer —Oapwe! Ob a Dorsalirayseee enn 19 19 18 20 19 19 Amiallespimesimivn aicmvue ent 3 3 3 3 3 3 Anal rays stg Seen oa 14 15 14 14 14 14

Body proportions in thousandths of standard length and counts of P. antarcticus from the Cape and Gough Island and P. porosa from the Kermadac Islands.

SAM 19573 NW. of Cape Town 19557 NW. of Cape Town 23311 Gough Island 24532 Vema Seamount

! Kermadec Islands. Collection of the British Museum (Natural History).

Body proportions and counts are given (Table 5) for the Gough Island fish, the two Cape fish and a larger specimen caught at the Vema Seamount by Mr. A. C. Paterson and presented by the Division of Sea Fisheries. As far as proportions and fin counts are concerned it is clear that there are no signifi- cant differences between the fish that could be construed as being of a specific nature.

In all four examples the pectoral fin is long and falcate, reaching the anal origin. The preoperculum has a curved scaled patch and the operculum is completely scaled. There is also a small patch of scales on the head above the preoperculum, as shown in Barnard (1948, fig. 14.). The mouth is large, reaching the posterior edge of the pupil. The origin of the dorsal fin is posterior to the pectoral fin base.

P. porosus is known from the coast of Australia and the Kermadec Islands. Seven juveniles from the latter locality, kindly lent by the British Museum (Natural History), were compared with P. antarcticus. Several differences in body proportions were apparent (Table 5), as well as the shape of the pectoral fin. Regan (1902), however, mentioned that the stromatid fishes changed considerably in body shape during growth, especially in that the pelvic fin becomes shorter and the pectoral fin becomes falcate in shape.

Although there are marked differences in the size of the fishes examined (over 430 mm as opposed to about 35 mm in total length), the similarity in pattern of the scales of the head suggest that P. forosus is at least very closely related to P. antarcticus.

In recent years large numbers of stromatid fishes assigned to the species

THE FISHES OF TRISTAN, GOUGH AND THE VEMA SEAMOUNT 537

P. porosus have been reported from Australian waters (Cowper, 1960) but no redescription or figure of a specimen larger than the type of 140 mm has been published.

Two stromatids have been recorded from St. Paul Island, Serzolella velaint which has a fin formula of dorsal VIII 27 and anal III 20 (Sauvage, ° 1879), and P. porosus, which was recorded and briefly described by Angot (1951). His brief description could apply equally to porosus or antarcticus, but the sketch shows a fish with a long falcate pectoral fin characteristic of antarcticus rather than the short rounded fin as shown in Richardson’s figures of the type of porosus.

In the North Atlantic there is another related species, P. perciformis (Mitchill) which Smith (19493) recorded from the west coast of South Africa.

This species appears to be generally smaller than antarcticus; it is said to be dark green in colour (Jordan & Evermann, 1896) whereas antarcticus is blue; the dorsal spines after the fourth are subequal (Regan, 1902) whereas in antarcticus and porosus they decrease in size and, finally, the pectoral fin in perciformis (as figured by Merriman, 1945) is markedly shorter and less falcate than in antarcticus, even at the largest recorded size (284 mm L’s).

It would appear therefore that there are two closely related species of Palinurichthys in the Atlantic Ocean, a northern species P. perciformis (Mitchill, 1818) and a southern species P. antarcticus (Carmichael, 1818). P. antarcticus has been found in South Africa, at Tristan and Gough Islands, the Vema Sea- mount and also appears to be present at St. Paul Island in the southern Indian Ocean. If adult specimens of porosus and antarcticus are compared and found to be conspecific P. porosus (Richardson, 1845) must be relegated to the synonomy of antarcticus.

Mupus imperialis Cocco

Mupus imperialis Cocco 1833: 20. Smith, 1949): 843. Mupus ovalis Smith, 19494: 303. Lirius ovalis Barnard, 1948: 392.

This species seems very common on the Seamount and grows to a large size, probably at least to 1 metre in length. In view of its rounded head it is called ‘stumpnose’ or ‘biskop’ by South African fishermen visiting Vema. ‘Two specimens have been examined. In thousandths of standard lengths of 590 mm and 762 mm the present specimens have the following respective body pro- portions: head 289-297, depth at pelvic origin 334-343, max. depth 393-370, snout 76-79, eye 68-66, pectoral fin 263-244, base of dorsal fin 728. Counts were dorsal VI-VII 28, —28, anal III 20-21.

M. imperialis has been recorded from wide areas of the Atlantic and the Mediterranean and appears to grow to a large size. Sivertsen (1945) described as new a stromatid fish collected at Tristan which he named Sertolella christo- pherseni. From the description I can find no significant characters whereby this species differs from M. imperialis. Both have similar body proportions, especially

538 ANNALS OF THE SOUTH AFRICAN MUSEUM

the large eye and rounded head, rather small mouth and a body which is deepest posterior to the pelvic fin origin. The counts for the unique specimen of christopherseni are all within the range for M. imperialis.

Coryphaenidae Coryphaena hippurus Linnaeus

Coryphaena hippurus Linnaeus, 1758: 261. Fowler, 1936: 649. Gibbs & Collette, 1959: 117.

Two small specimens of this species were collected at the Vema Seamount by Mr. Paterson of the Division of Sea Fisheries. Both specimens (SAM 24531) are immature, being only 352 and 375 mm in standard length, but are clearly hippurus rather than the related equiselis, having total gillraker counts of 9-10 and 62-63 dorsal rays.

Scorpaenidae Scorpaena scrofa Linnaeus Scorpaena scrofa Linnaeus, 1758: 266. Smith, 1957: 51 Scorpaena natalensis Regan, 1906: 5. Barnard, 1927: 902. Smith, 19494: 371.

One specimen from a commercial fishing vessel at the Vema Seamount. Body proportions as thousandths of standard length (313 mm) are as follows: head 418, depth 313, eye 80, interorbital 73, snout 130, dorsal base 534, anal base 125, pectoral 265.

Congridae Ariosoma australis (Barnard) Congermuraena australis Barnard, 1923: 442. Barnard, 1925: 190. Sivertsen, 1945: 4.

Arissoma balearica (non de la Roche) Smith, 19494: 393 partim. Congermuraena ?habenata (Richardson, 1845: 109). Sivertsen, 1945: 4.

Barnard (1923) described two species of congrid eels as Congermuraena australis and C. albescens, the former from shallow water at the Cape of Good Hope and Tristan da Cunha and the latter from deep water at the Cape. Sivertsen (1945) obtained one specimen at Tristan, and, although he assigned it to C’. australis, he stated that there were minor differences between his specimen and Barnard’s description, and suggested that the Tristan species may be C’. habanata Richardson, known from New Zealand and St. Paul Island.

Smith (1949a) reduced both australis and albescens to the synonymy of the Mediterranean species Ariosoma balearica (de la Roche) but gave no reasons for doing so.

A careful examination of A. australis from shallow water at the Cape and Tristan and A. albescens (from deep water off the Cape) has shown the two eels to have clear differences. Body proportions are given in Table 6. There is little difference between the fish in so far as the body proportions are concerned except that the eye in albescens is proportionally smaller and the interorbital width greater than in australis from the Cape and with the Tristan fish having an intermediate eye size but a narrow interorbital width.

THE FISHES OF TRISTAN, GOUGH AND THE VEMA SEAMOUNT 539

In general appearance, however, the fish from Tristan are very similar to australis, which differs from albescens in the following characters: in australis the rays in the vertical fins are clearly visible, the fin has a dark edge, the lateral line has one tiny pore to each myotome, the upper dentition has a rather

L

Fig. | Upper tooth patterns of a, Ariosoma australis syntype; b, Ariosoma albescens holotype. (Diagrammatic and not to scale.)

elongate, slender pattern (fig. 1a), the body is well formed and firm, and I can find no trace of the epidermal processes stated by Castle (1960) to occur on the body of A. balearica. In albescens the vertical fins are enclosed in a thick gelatinous envelope, the fins are pale edged, the lateral line is a distinct ridge and has a

540 ANNALS OF THE SOUTH AFRICAN MUSEUM

large lower pore and a small upper one to each myotome, the pattern of the upper dentition is shorter and heavier (fig. 15), there are distinct epidermal processes present and, finally, the body is soft and flabby. It is unlikely that the differences are due to changes taking place during growth, as a specimen of australis at 280 mm was found to be fully mature and to have ripe ovaries containing large eggs.

Smith’s (1949a) combining of australis and albescens is therefore considered to be incorrect.

A. balearica is known from the tropical Atlantic and the Mediterranean (Fowler, 1936) and is very similar to A. australis. It, too, has a very indistinct lateral line which will separate it from A. albescens. There is also very little difference between australis and A. habenata from New Zealand and St. Paul Island as pointed out by Sivertsen (1945) and between the New Zealand deep- water species Pseudoxenomystax hirsutus Castle (1960) and the South African deepwater species A. albescens. Castle divides the two species by the position of the origin of the dorsal fin, stating that it lies over the gill slit in hirsutus and at a level half-way along the pectoral in Smith’s specimens of ‘balerica’. In the type of albescens (designated in the museum catalogue as SAM 12775) the origin of the dorsal is above the pectoral origin.

TABLE 6

Catalogue No. SAM: 12781 12781 12780 12781 18097 12782 24551 12775 18097 18097

Standard lensth. :. . 206 925 238 o74 ) 2981 | 186 368 668) 737imoro Deon 5) o 5g ao 8 | RR 67 51 55 47 56 65 88 73 74. Headey) 4 SLGO} 160M eGo: OFT OON On 160) 0157) a5 Headitovent = = = 308: 301 | Aobwisoq2— 403 376) 405) 454 oonmeog Venttotail . . . . 606 608 592 607 597 635 595 550 596 587 EV Cin ee ror mrsnweks Mestre Tho Q 38 38 40 2 37 33 Q 26 Q

Interorbital) 2) 7 4 = 14 16 13 16 II 16 18 37 26 26 Snout'toiorigin dorsal . 180 182° 289) 4727182) 9178 188° 178) Toomer Snout to origin anal 3» AZ 405) 420) 405 403. — 481 — =

Height cillislit = 9-0) 17 18 07) 18 18 QI 19 25 — —

Body proportions (in thousandths of standard length) of Ariosoma australis from South Africa and Tristan da Cunha and A. albescens from South Africa.

SAM 12780 A. australis paratype, collected Kalk Bay 12781 A. australis, Kalk Bay 18097 A. australis, Kalk Bay 12782 A. australis, Tristan 24551 A. australis, Tristan 12775 A. albescens, holotype, deep water off Cape Town 23189 A. albescens, deep water off Cape Town

The systematics of this group of eels is in a very confused state, there being little agreement even at the generic level, and until such time as a careful revision of the family is made on a world-wide basis it will be impossible to decide whether A. habenata, A. australis from the Cape, A. australis from Tristan and several other species are in fact distinct species or are only geographical variants. It is considered advisable therefore that the existing species be retained

THE FISHES OF TRISTAN, GOUGH AND THE VEMA SEAMOUNT 541

in the meantime. It would be surprising, however, if the Tristan and St. Paul Island eels are specifically different.

Lagocephalidae Sphaeroides cutaneus (Gunther)

Tetrodon cutaneus Gunther, 1870: 287. Barnard, 1927: 971. Sphaeroides dubius von Bonde, 1923: 40. Sphaeroides cutaneus (Gunther) Smith, 1949a: 417.

This large species of puffer fish is found throughout the warm, temperate and tropical seas. It is known from the Cape and St. Helena Island.

Three specimens from Vema were collected and presented by Mr. Paterson (SAM 24524). All three were a dirty grey-green dorsally with a white belly. Dorsal, anal and pectoral fins were pale yellow. The pectoral fin has a charac- teristic shape, being broad, rounded on the lower distal corner, and pointed at the upper. The species is capable of very extensive inflation and all three examples examined were much distorted due to this.

ZOOGEOGRAPHY

The relative geographical positions of the islands of Tristan da Cunha, Gough, St. Paul, Marion and Crozet, the Vema Seamount and South Africa are shown in figure 2. Tristan, Gough and St. Paul islands lie on very similar latitudes, whereas the Vema Seamount is roughly as far north of Tristan as the Marion and Crozet groups are south. The Vema Seamount, however, has several fish in common with Tristan whereas the Marion and Crozet groups have no fish in common with Tristan or St. Paul. This, as was suggested by Sivertsen (1945), is almost certainly due to hydrographic conditions, mainly water temperatures.

e@ Mauritius Walvis Bay ® Reunion

Vema e

Cape Town “Port Elizabeth

Amsterdam,

St. Paul +

: z.. Crozet Marion e* eee

Fig. 2 Shallow-water areas in the southern Atlantic and Indian Oceans.

542 ANNALS OF THE SOUTH AFRICAN MUSEUM

There are two areas of sudden temperature change in the Southern Ocean, the Sub-Tropical Convergence, where the surface temperature rapidly changes from about 13°C to about 9°C and a more southerly line, the Antarctic Con- vergence, where the surface temperature again drops, from 6°C to about 2°C. The Antarctic Convergence is quite narrow but the Sub-Tropical Convergence is much broader and Sverdrup eé al. (1942) refer to it as a Region of Conver- gence. The Sub-Tropical Convergence fluctuates in latitude, at times the whole zone is 5-6° South of the Island and at other times Deacon (1937) suggests that it may be north of Tristan. Temperatures taken at Gough Island at a time when he suggested this do not support it, however.

Gough Island is approximately 3° 20’ south of Tristan, thus it may lie north or south of the Convergence according to the conditions prevailing in the ocean. Both, however, probably normally lie north of or within the Con- vergence zone, while Vema always lies well to the north and the Marion and Crozet group well to the south.

There is little published data available giving water temperatures at Tristan and even less for Vema or Gough. Christophersen & Schon (1942) published detailed sea surface temperature records taken during the Nor- wegian expedition to Tristan. The main features were:

Temperatures in °C

Month Average Lowest Highest December 1937 Sate 15°7 15°0 16-4 January. e103 G0 sen lee 16-7 14°8 17-9 Hebruary, 193Gm su a 18+4 17°6 19°8 March 1938 om, 18, eee 18-9 18-0 20°0

The above are all surface temperatures taken from the shore. ‘Two stations worked by R.V. Discovery gave surface temperatures of 14:59°C (Station 4) in January and 12-95°C (Station 397) in May. Station 4 was taken close inshore where the bottom was only 37 metres. A subsurface sample at 35 m gave a temperature of 13:64°C.

Gough Island too appears to have a very deep thermocline, stations worked by the R.V. William Scoresby in June 1927 showed the following:

Temperatures in °C at Sample Depth

Station Date om 100 m 150 m W/s 122 7-8/6/27 11°59 11°69 9°89 125 9/6/27 11°63 11°49 10°99 126 10/6/27 11°79 11*70 9:69 130 12/6/27 12-09 12°09 10°71

In April 1965, the supply ship R.S.A. recorded surface temperatures at Gough Island ranging from 11-8°C to 13:7°C over two days (Crawford, pers. comm.).

The main fishing areas at Tristan and Gough are in shallow (under 50 metres) water. At this depth it would seem that there is normally but little difference between surface and ambient temperature of the bottom living species.

THE FISHES OF TRISTAN, GOUGH AND THE VEMA SEAMOUNT 543

The surface temperatures at Vema are significantly higher, the Emerson Kk found surface temperatures of between 18-3°C and 18-8°C in November 1964 (Simpson & Heydorn, 1965) and in April 1965 the R.V. Africana IT obtained

the following temperatures on and about the Seamount:

Sounding Sampling depth Temperature Station (metres ) (metres) EG A.3617 go to) 20°44 50 20°13 80 18-50 A.3620 goo fe) 20°37 50 20°35 100 16-15 A.3624 480 0) 20°40 50 20°38 100 16-63 A.3625 73 (e) 20°36 50 20°33 A.3626 goo fe) 20°35 50 20-28 100 15°51

There is a more rapid decrease in water temperature in the first hundred metres at Vema than at the two southern islands, but the ambient temperature of the bottom living species (main fishing areas are 70-80 metres deep) is still several degrees higher than at Tristan or Gough.

It is probable that this bottom temperature is close to the maximum for the fish at Vema which are more normally associated with cooler water (Acantholatris spp. and Palinurichthys antarctica) and yet still suitable for sub- tropical species (Seriola lalandit and Epinephalus aeneus).

The surface water temperature at Vema throughout the year, and at Tristan occasionally in the summer is suitable for tropical and subtropical surface pelagic species (Coryphaena hippurus, Sphaeroides cutaneus and Exocoetidae) . Cool water surface species, however, such as Thyrsites atun, common at Tristan, appear to avoid the warm water and have not been found at Vema.

TABLE 7 South Tristan Gough Vema Africa St. Paul (a) Coast fishes

Gaidropsarus insularum. . + af a5 Acantholatris monodactylus - + ob Acantholatris vemae . . . + Latris lineata ah es =F Bovichthys diacanthus . . 4 + Decapterus longimanus . . Labrichthys ornatus . Felicolenus tristanensis . Sebastichthys capensis Scorpaena scrofa ... . ar a Epinephaius aeneus . . . a= Caesioperca coatsi. . . . a Ariosoma australis. . . 2. + in

++++4

544 ANNALS OF THE SOUTH AFRICAN MUSEUM

South Tristan Gough Vema Africa St. Paul (6) Oceanic fishes

Prionace glauca Alepisaurus ferox . Maurolicus muelleri . Myctophum humboldti . Scomberesox saurus Exocoetus exsiliens . Cypsilurus lineatus Beryx decadactylus . Seriola lalandi . : Plagiogeneion rubiginosus . Pentaceros richardsont Polyprion americanus Notopogon ?macrosolen . Thunnus spp. Thyrsites atun . ye. Mupus imperialis . . 2. 2.) + Palinurichthys antarcticus . — 4 Coryphaena hippurus . Sphaeroides cutaneus .

Distribution of fishes, known from the Islands of Tristan da Cunha and Gough and the Vema Seamount.

++++4+++4++4+

++ +4 + ++ b+t+++++++4+44

|

+ +

++4++4

++++ +

Sivertsen felt that there was a large number of fish endemic to Tristan (33% of the known fish) and a greater similarity with St. Paul Island than South Africa.

Table 7, which lists all the fish known from Tristan, Vema and Gough, is divided into oceanic and coastal species in much the same manner as arranged by Sivertsen. It is never easy, however, to distinguish between oceanic and coastal species and some, considered to be coastal by Sivertsen, are here listed as oceanic. Thirteen species are considered coastal, of these eight are found at Tristan, seven at Gough and six at Vema. Three of these species also occur definitely at St. Paul Island and four in South Africa. None of the species definitely occur at all five localities and only one at Tristan, Gough and Vema.

It is not agreed with Sivertsen that Tristan has a large proportion of endemic fish, in fact it is a surprising conclusion that this isolated island has no endemic fish species. Vema, however, appears to have one, as does Gough. It is also surprising that such an apparently successful species as A. monodactylus has reached Vema but not South Africa.

Almost all the oceanic species listed in Table 7 are known from Tristan, only the tropical species being absent. All species recorded from Tristan are also known from South Africa, a few from Vema and almost none from Gough. This pattern is almost certainly due more to the varying intensities to which the various areas have been fished than to any real differences.

Some of the species recorded from Tristan are deep-living bathypelagic species that have either been caught on the surface at night or have come into shallow water due to the abrupt elevation of the ocean floor. Such species are Maurolicus muelleri, Myctophum humboldti, Alepisaurus ferox, and Pentaceros rich-

THE FISHES OF TRISTAN, GOUGH AND THE VEMA SEAMOUNT 545

ardsoni. All are species which have been found in widely scattered parts of the world.

Other fish again are surface-living pelagic forms not normally found close to land but presumably attracted by the rich food available at the edges of the oceanic islands and in the shallow water over the Vema Seamount. Fish of this type recorded are Coryphaena hippurus and Thunnus spp. from Vema, and Scomberesox saurus, Prionace glauca and Exocoetus exsiliens from Tristan.

Another group of pelagic fishes are also wide ranging but normally congre- gate close to land. Seriola lalandi, Thyrsites atun and Sphaeroides cutaneus are of this type. The first has been taken at both Vema and Tristan, the second only at Tristan and the third only at Vema. All three, however, are known from South Africa and the east coast of America. Palinurichthys antarcticus is possibly a similar fish but has a deep pelagic habitat rather than occurring close to the surface. It appears to have a very wide distribution, being known from all three areas discussed, South Africa, probably St. Paul Island and possibly Australia and New Zealand.

Some of the species which have a wide geographic distribution are small, weakly swimming, species, e.g. Labrichthys ornatus which is known from Tristan and St. Paul islands, and probably also Gough (Holdgate 1948) and Vema (Heydorn, fers. comm., underwater sighting). Other widely distributed, but weakly swimming species are Ariosoma australis which occurs in South Africa and Tristan, and possibly St. Paul as well and Gaidropsarus insularum known from South Africa, Tristan and St. Paul. Both the latter two species have pelagic larval stages.

Rowan & Rowan (1955) mention large fish seen jumping in the vicinity of Tristan. At times, in summer, the surface water temperature at Tristan reaches a temperature quite suitable for Tuna, Blue and White Marlin and Broadbill Swordfish. Isurus glaucus has been recorded from St. Paul (Blanc, 1961).

It is interesting that no elasmobranchs other than Prionace glauca have been recorded from the areas discussed, although ‘spiny dog fish’ Squalus fernandinus, are common at St. Paul (Blanc & Paulian, 1957).

None of the gadids, thought by Rowan & Rowan (1955) to belong to at least two species, have been obtained. They thought that one belonged to the genus Haloporphyrus but. as they obtained only damaged specimens washed ashore by storms, identification was not possible.

SUMMARY

The fishes known from the newly discovered Vema Seamount are described and compared with those found off Tristan da Cunha and Gough Island. A new species of chilodactylid, Acantholatris vemae, is described from Vema. The genus Gaidropsarus from Tristan and South Africa is discussed and the South African material shown to belong to two species, G. insularum, also known from

546 ANNALS OF THE SOUTH AFRICAN MUSEUM

Tristan and St. Paul Island, and G. capensis from the south-east coast of South Africa.

An attempt is made to relate the fish fauna to what is known of the hydro- graphic conditions around the islands.

ACKNOWLEDGEMENTS

For the material on which this paper is based I am indebted to the Director and staff, Division of Sea Fisheries, Sea Point (especially Messrs. Heydorn, Nepgen and Paterson), Capt. M. T. Scott, Messrs. Friedman & Rabinowitz (Pty.) Ltd., and Messrs. Irvin & Johnson (Pty.) Ltd., all of Cape Town. Material was lent by the Trustees of the British Museum (Natural History) and Professor J. L. B. Smith of Rhodes University, Grahamstown.

Unpublished data on water temperatures were supplied by Mr. A. Crawford, Mr. G. Stander and Mr. M. Orren, while Mr. A. Wheeler of the British Museum helped with much useful information.

Finally I am grateful to my wife, Dr. Mary-Lou Penrith, for checking the manuscript, and to the Librarian, Council for Scientific and Industrial Research, who obtained photocopies of several papers not available to me.

The Trustees of the South African Museum are grateful to the Council for Scientific and Industrial Research for a grant to cover the cost of publishing this paper.

REFERENCES

Ase, T. 1957. New, rare and uncommon fishes from Japanese waters. VI. Notes on the rare fishes of the family Histiopteridae. Jap. 7. Ichthyol. 6: 35-40, 71-74.

Ancort, M. 1951. Observations sur la fauna marine et Ja peche aux Iles Saint-Paul et Amster- dam. Mem. Instit. Sci. Madagascar (A) 6 (1): 1-51.

BARNARD, K. H. 1923. Diagnoses of new species of marine fishes from South African waters. Ann. S. Afr. Mus. 13: 439-445.

BARNARD, K. H. 1925. A monograph of the marine fishes of South Africa I. Ann. S. Afr. Mus. 21: 1-418.

BARNARD, K. H. 1927. A monograph of the marine fishes of South Africa II. Ann. S. Afr. Mus. 21: 419-1065.

BARNARD, K. H. 1948. Further notes on South African Marine fishes. Ann. S. Afr. Mus. 36: 341-406.

BARNARD, K. H. 1956. The dates of issue of the ‘Illustrations of the Zoology of South Africa’ and the ‘Marine Investigations in South Africa’, 7. Soc. Bibl. nat. Hist. 2: 187-189.

Bianc, M. 1961. Les poissons des terres australes et antarctiques Francaise. Mem. Inst. Sci. Mada- gascar (F) 4: 109-159.

Bianc, M. & PauLian, P. 1957. Poissons des Iles Saint-Paul et Amsterdam. Mem. Inst. Sci. Madagascar (F) 1: 325-335.

Biocu, M. E. & ScHNEwER, J. G. 1801. Systema Ichthyologicae. Saxony, Berolini.

Bruun, A. F. 1935. Flying fishes (Exocoetidae) of the Atlantic. Dana Rep. 6: 1-106.

CarMICHAEL, D. 1818. Some account of the Island of Tristan da Cunha and of its natural resources. Trans. Linn. Soc. Lond. 12: 493-513.

Caste, P. H. J. 1960. Two eels of the genus Pseudoxenomystax from New Zealand waters. Trans. roy. Soc. N.&. 88: 463-472.

Cocco, A. 1833. Su di alcuni pesci de mari di Messina. Giorn. Sci. Lett. Arti. Sicilia 42: 9-21.

Cowper, T. R. 1960. Occurrence of Pyrosoma on the continental slope. Nature Lond. 187: 878-879.

Cuvirr, G. L. & VALENCIENNES, A. 1830. Histoire naturelle des Poissons 5. Paris: F. G. Levrault.

THE FISHES OF TRISTAN, GOUGH AND THE VEMA SEAMOUNT 547

Deacon, G. E. R. 1937. The hydrology of the southern ocean. ‘Discovery’ Rept. 15: 3-124.

Discovery Committee. Discovery Investigations Station-List, 1925-1926. ‘Discovery’ Rept. I: I-140.

Foutett, W. J. & Dempster, L. J. 1963. Relationships of the percoid fish Pentaceros richardsoni Smith, with description of a specimen from the coast of California. Proc. Calif. Acad. Sci. 32: 315-338.

Fow ter, H. W. 1936. The marine fishes of west Africa. Bull. Amer. Mus. nat. Hist. 70: 1-1493.

GrOFFROY-SAINT-HILLAIRE, I. 1809. Histoire naturelle des Poissons de la mer Rouge et la Mediterraneé. Descrip. de l’Egypte Hist. nat. 1: 311-343.

Gress, R. H. & CoLt.ette, B. B. 1959. On the identification, distribution and biology of the dolphins Coryphaena hippurus and C. equiselis. Bull. mar. Sci. Gulf Caribb. g: 117-152.

Gut, T. 1863. Synopsis of the family of Cirrhitoids. Proc. Acad. nat. Sci. Philad. 1862: 102-124.

GunTHER, A. 1859. Catalogue of the Acanthopterygian fishes in the collection of the British Museum 1. London: Taylor & Francis.

GuNTHER, A. 1960. Catalogue of the Acanthopterygian fishes in the collection of the British Museum 2. London: Taylor & Francis.

GUNTHER, A. 1870. Catalogue of the Acanthopterygian fishes in the collection of the British Museum 8. London: Taylor & Francis.

Hotpeate, M. 1958. Mountains in the Sea. London: Macmillan.

Hutton, F. W. 1876. Contributions to the Ichthyology of New Zealand. Trans. Proc. N.Z. Instit. 8: 209-218.

Jorpan, D. S. & Evermann, B. W. 1896. The fishes of north and middle America. Bull. U.S. nat. Mus. 47: 1-1240.

Kaup, J. 1858. Ubersicht der Family Gadidae. Arch. Naturgesch. 24: 85-93.

Kner, R. 1867. Zool. Theil Fische. Reise der osterreichen Fregatte Novara um die Erde in den Jahren 1857, 1858, 1859 3: 275-433. Vienna: Staatsdrucken.

Mircue t, S. L. 1818. Memoirs on Ichthyology. The fishes of New York described and arranged. Amer. Mon. Mag. 2: 244-245.

Monr, E. 1937. Revision der Centriscidae. Dana Rept. 13: 1-69.

Norman, J. R. 1935(a). The Carangid fishes of the genus Decapterus Bleeker. Ann. Mag. nat. Hist. (10) 16: 252-264.

Norman, J. R. 1935(5). Coast fishes I. The South Atlantic. ‘Discovery’ Rept. 12: 3-58.

REGAN, C. T. 1902. A revision of the family Stromateidae. Ann. Mag. nat. Hist. (7) 10: 115-1323 194-207.

REGAN, C. T. 1913(a). The fishes described by Carmichael from Tristan da Cunha. Ann. Mag. nat. Hist. (8) 11: 498-504.

Recan, C. T. 1913(b). The Antarctic fishes of the Scottish National Antarctic expedition. Scot. nat. Antarct. Exp. Sci. Res. Scotia. 4: 311-374.

Recan, C. T. 1914. Diagnoses of new marine fishes collected by the British Antarctic (Terra Nova) expedition. Ann. Mag. nat. Hist. (8) 13: 11-17.

RicHarpson, J. 1839. Account of a collection of fishes from Port Arthur, Van Diemens Land. Proc. Zool. Soc. Lond. 7: 95-100.

RIcHARDSON, J. 1842. Description of Australian Fish. Trans. Zool. Soc. Lond. 3: 69-185.

RicHARDsON, J. 1844-1848. Ichthyology of the voyage of H.M.S. Erebus and Terror. London.

Rowan, M. K. & Rowan, A. N. 1955. Fishes of Tristan da Cunha. S. Afr. 7. Sct. 52: 129.

Sauvace, H. E. 1879. La faune ichthyologique de I’Ie Saint-Paul. Arch. Zool. exp. gen. 8: 1-46.

Simpson, E. S. Q. & Heyporn, A. E. F. 1965. Vema Seamount. Nature Lond. 207: 249-251.

SIVERTSEN, E. 1945. Fishes of Tristan da Cunha with remarks on age and growth based on scale readings. Results Norweg. Exped. T. da Cunha 2 (12): 1-44.

SmitH, A. 1844. Expedition into the interior of South Africa in 1834-36.

Smitu, J. L. B. 1949(a). The Sea Fishes of Southern Africa. Cape Town: C.N.A.

SmirH, J. L. B. 1949(b). The Stromateid fishes of South Africa. Ann. Mag. nat. Hist. (12) 2: 839-851.

Smitu, J. L. B. 1957. The fishes of the family Scorpaenidae in the Western Indian Ocean I. Ichthyol. Bull. Rhodes Univ. 4: 49-69.

Smiru, J. L. B. 1959. Serioline fishes (Yellowtails: Amberjacks) from the Western Indian Ocean. Ichthyol. Bull. Rhodes Univ. 15: 255-261.

SmitH, J. L. B. 1964. Fishes of the family Pentacerotidae. Ichthyol. Bull. Rhodes Univ. 29:

567-578.

548 ANNALS OF THE SOUTH AFRICAN MUSEUM

STEINDACHNER, F. 1866. Ichthologische Notizen 3. Uber einige neue Fisch-Arten aus Siid- america. S.B. Akad. Wiss. Wien 53: 208-214. SverpRuP, H. U., Jounson, M. W. & FLeminc, R. H. 1942. The Oceans. New Jersey: Prentice

eeiall: Von BonbeE, C. 1923. Shallow water fishes procured by the S.S. Pickle. Fish. Mar. Biol. Surv.

Rept. 3: 1-40. WatTeERHOusE, F. H. 1880. On the dates of publication of the parts of Sir Andrew Smith’s Illustrations of the Zoology of South Africa. Proc. Zool. Soc. Lond. 1880: 489-491. WELANDER, A. D. ef al. 1957. Occurence of the boar fish Pseudopentaceros richardsoni and the zeid Allocyttus verrucosas in the North Pacific, Copeia 1957: 244-245. Wauuitr.ey, G. P. & Puiturps, W. J..1939. Descriptive notes on some New Zealand fishes. Trans.

roy. Soc. N.Z. 69: 228-236.

Ann. S. Afr. Mus., Vol. XLVIII Plate XXI

A. Acantholatris vemae sp.n. Type of 383 mm. standard length.

B. Acantholatris monodactylus. Specimen of similar size from YVema seamount.

INSTRUCTIONS TO AUTHORS

MANUSCRIPTS

In duplicate (one set of illustrations), type-written, double spaced with good margins, including Taste or Contents and Summary. Position of text-figures and tables must be indicated.

ILLUSTRATIONS

So proportioned that when reduced they will occupy not more than 43 in. = 7 in. (7 in. including the caption). A scale (metric) must appear with all photographs.

REFERENCES

Authors’ names and dates of publication given in text; full references at end of paper in alphabetical order of authors’ names (Harvard system). References at end of paper must be given in this order:

Name of author, in capitals, followed by initials; names of joint authors connected by &, not ‘and’. Year of publication; several papers by the same author in one year designated by suffixes a, b, etc. Full title of paper; initial capital letters only for first word and for proper names (except in German). Title of journal, abbreviated according to World list of scientific periodicals and underlined (italics). Series number, if any, in parenthesis, e.g. (3), (n.s.), (B.). Volume number in arabic numerals (without prefix ‘vol.’), with wavy underlining (bold type). Part number, only if separate parts of one volume are independently numbered. Page numbers, first and last, preceded by a colon (without prefix ‘p’). Thus:

SmirH, A. B. 1956. New Plonia species from South Africa. Ann. Mag. nat. Hist. (12) 9: 937-945.

When reference is made to a separate book, give in this order: Author’s name; his initials; date of publication; title, underlined; edition, if any; volume number, if any, in arabic numerals, with wavy underlining; place of publication; name of publisher. Thus:

Brown, X. Y. 1953. Marine faunas. and ed. 2. London: Green.

When reference is made to a paper forming a distinct part of another book, give: Name of author of paper, his initials; date of publication; title of paper; ‘In’, underlined; name of author of book; his initials; title of book, underlined; edition, if any; volume number, if any, in arabic numerals, with wavy underlining; pagination of paper; place of publication; name of publisher. Thus:

SmitH, C. D. 1954. South African Plonias. In Brown, X. Y. Marine faunas. and ed. 3: 63-95. London: Green.

SYNONYMY

Arranged according to chronology of names. Published scientific names by which a species has been previously designated (subsequent to 1758) are listed in chronological order, with abbreviated bibliographic references to descriptions or citations following in chronological order after each name. Full references must be given at the end of the paper. Articles and recommendations of the /nternational code of zoological nomenclature adopted by the XV International congress of zoology, London, July 1958, are to be observed (particularly articles 22 and 51).

Examples: Plonia capensisSmith, 1954: 86, pl. 27, fig. 3. Green, 1955: 23, fig. 2.

When transferred to another genus: Euplonia capensis (Smith) Brown, 1955: 259. When misidentifiedas another species: Plonia natalengs (non West), Jones, 1956: 18. When another species has been called by the same name: [non] Plonia apensis: Jones, 1957: 27 ( = natalensis West).

"Wiig

3 9088 01206