ater ss meV ae ae pope b's ee eee Om fle CS re be a =— ot 33.542 SESE sii ewes Sot PON SENT aasbpe warge ea ee eerieion-sives DP pepe a dnt sv Mareen eset wmare ie par sea ssa sr 7 a ene ena aE ta ae ee eR TE PO rocug ° . PS ee aaa tia ai iat ed CE SK CEC C ‘ Pr. oH a a ita 5 RD ve gah ty if ea ea ae Pyle u Raa) Lun \ t A G Maar ire sy vis a phony A Pais We ANNALS OF THE SOUTH AFRICAN MUSEUM VOLUME XT Fi ANNALS OF THE SOUTH AFRICAN MUSEUM VOLUME XIII We Gr o S¢ S | Ones Lill fo ON. F 7 es PRINTED FOR THE TRUSTEES OF THE SOUTH AFRICAN MUSEUM BY NEILL AND CO., LTD., EDINBURGH. 1913 — 1923. TRUSTEES OF THE SOUTH AFRICAN MUSEUM. The Right Hon. JoHn Xavier Merriman, P.C., M.L.A. Sir Toomas Murr, Kt., C.M.G., LL.D., F.R.S., F.R.S.E. The Hon. JoHN Witit1am JacGceEr, M.L.A., F.R.Stat.S. SCIENTIFIC STAFF OF THE SOUTH AFRICAN MUSEUM. Louis ALBERT PEéRINGUEY, D.Sc., F.Z.8., F.E.S., Director. KepreL Harcourt Barnarp, M.A., F.L.S., Assistant Director and Assistant in Charge of Fish and Marine Invertebrate Collections. ArtHuR Lewis Hatt, B.A., Hon. Keeper of the Geological and Mineralogical Collections. Stpney Henry Havueuton, B.A., D.Sc., Hon. Keeper of the Palaeontological Collections. REGINALD FREDERICK Lawrence, B.A., Assistant in Charge of Arthropoda (Insects excluded). Srar GARABEDIAN, B.A., Assistant in Charge of Botanical Collections. LIST OF CONTRIBUTORS. K. H. Barnarp. PAGE Diagnoses of New Species of Marine Fishes from South African Waters. 439 G. A. BoULENGER. A Revision of the Lizards of the Genus Nucras Gray . : : LOS Description of a New South African Lizard of the Genus Eremias a Pile Descriptions of Three New Fresh-Water Fishes from Northern Rhodesia 437 lel, 1b, Cinini< The Echinoderm Fauna of South Africa. : 3 : 5 2221 M. ConNoLLy. Notes on South African Mollusca, I-III. : : : : : 99 Notes on South African Mollusca, IV ; d : : F . 179 J. D. F. Gitcurist and W. W. THompson. Descriptions of Fishes from the Coast of Natal (Part IV). : 5 OF R. KirKPatTRICK. Note on the Occurrence of the Euplectellid Sponge, Regadrella ES O. Schmidt, off the South African Coast ; : + (iB W. MIcHAELSEN. Report upon the Oligochaeta in the South African Museum atCapeTown 43 W. R. OGILVIE-GRANT. On a New Lark from the Cape Province . ; : : 3 . 41 L. P&RINGUEY. Inscriptions left by Early European Navigators on their Way to the East 1 L. J. SHACKLEFORD. Two New Species of Marginella from South Africa. . . 5 «ON Two New Species of Marginella from South Africa. : ‘ : 193 vi INDEX OF NEW GENERIC NAMES INTRODUCED IN THIS VOLUME. PAGE Diastobranchus (Synaphobranchidae) BARNARD : 3 ; : . 441 Dictenophiura (Ophiolepididae) CLARK. : é ; : . OOL Perissasterias (Asteriidae) CLARK. : : : : ‘ : . 307 Spatagobrissus (Spatangidae) CLARK ; : : : : : . 402 Tropholampas (Nucleolitidae) CLark F : : : : ‘ . 395 DATE OF ISSUE OF THE PARTS. Part 1, July 24th, 1913. Part 2, October 16th, 1913. Part 3, May 7th, 1914. Part 4, April 8th, 1915. Part 5, October 6th, 1916. Part 6, May 30th, 1917. Part 7, May 1923. Part 8, September 1923. ;: 5 \ 77 ( : ' . ‘s e » iy _ Pi ae i Ni A [ 4 t) | iMtaa : MG f r ny i s - ? Pe ne i. : he tye oan - ; ; ae oi j we 4 J Pr i Pee yi ’ i (hep ie MPN Py), feet Wi We ‘ hak ce eg oes ’ ¢ : V aighy ‘ Vv bie faa afk Ut al hes 7 ‘ } 4 bee et ha arg a As nfl My m7 ee teh AN a mi ya ay has [ me Pepe ray ieee ee ee ee i Ay” Ta ain ix 1 i ‘idis b AY oh ave he Ph, at ni ‘a | ra D os j ¢ fe’ vi it: ‘ : ae, : : ry tie " q { 5 Pr aa : |? " c ie : i * 7 ; ; 3 nL : ee M mt od ; Ly : ie 1 Wil hAiaee tie) or vie ANY <7 at. ome i heeds as sous ¥ ae ; ; ay : J vi Ay 4 7 ; +e ye i i fc ‘ ; 5 vl Mi Z wit f a fy ; oe wm, ten / : el , nt : , : ‘ te hg \s i : ee PP) ea ima” : my ue Jy . Habe, UO) rin Hibs, P i" nee ai fpr Yee. 6 pee : ei asi a he He a ae : : a nee, Sav FT thee ay Pe, ae 7 2 Ap a? ; | : tee a > in emt (eel ee renee bak eS ee ai eee onal ie: Ht ’ CA Lela et eesti th lee = LIST OF PLATES. PLATE I. Regadrella phoenix O. Schmidt. Trigonephrus globulus (Mull.). Trigonephrus gypsinus (M. & P.). Trigonephrus rosaceus (Mull.). Il. . Trigonephrus porphyrostoma (M. & P.). Trigonephrus namaquensis (M. & P.). Trigonephrus lucanus (Mull.). 4 -‘Trigonephrus ambiguosus (Fer.). Dorcasia coagulum (v. Mts.). III. Dorcasia rogersi n. sp. Doreasia cernua (v. Mts.). Dorcasia alexandri (Gray). IV. Anatomy of the Dorcasiinae. VY. Anatomy of the Dorcasiinae. j Nucras delalandii (D. & B.). \ Nucras intertexta (A. Smith). VIL. j Nucras intertexta var. holubi (Steind.). \Nucras tessellata (A. Smith). Monachocrinus coelus n. sp. | Liparometra multicirra n. sp. Comanthus wahlbergi (J. Mull.). Chondraster elattosis n. sp. exe! Tosia tuberculata (Gray). | Pteraster capensis Gray. X. Hymenaster gennaeus n. sp. Xx Cryaster brachyactis n. sp. | Hymenaster lamprus n. sp. xq. / Pseudarchaster brachyactis n. sp. \Calliaster acanthodes n. sp. xu. J Cladaster macrobrachius n. sp. \ Plutonaster proteus n. sp. Ceramaster patagonicus var. euryplax n. XIV. l Ceramaster trispinosus n. sp. Ceramaster chondriscus n. sp. xy, | Echinaster reticulatus n. sp. | Poraniopsis capensis n. sp. Mediaster capensis n. sp. XVI. , Asterina gracilispina n. sp. - Asterina dyscrita n. sp. ix List of Plates. PLATE Asterina granifera (Gray). XVII. | Asterina granifera var. sporacantha n. Uanseronads habracantha n. sp. f Lophaster quadrispinus n. sp. \ Perissasterias polyacantha n.g. and sp. Dictenophiura anoidea n.g. and sp. XIX. eens nerthepsila n. sp. Ophiomitrella corynephora n. sp. { Ophiochiton australis n. sp. XVIII. XX. 4 Ophiactis carnea Ljung. Vastrothamans papillatus n. sp. xx, /Coenopedina capensis n. sp. \Coelopleurus interruptus Doderl. XXII. Paracentrotus grandis n. sp. XXII. Spatagobrissus mirabilis n.g. and sp. Acanthodrilus Achatina . Actinopyga Agriolimax Alepocephalus Alestes Amphioplus Amphipholis Amphiura. Anseropoda Arion Asterias Asterina Asteronyx Astrocladus Astropecten Astrophiura Astropyga Astrothamnus Austrofromia Aspidodiadema Bathybiaster Batrachus Benthodytes Bimastus . Bregmaceros Brisaster . Brisinga Brissopsis Bythocrinus Calandrella Calliaster . Centrechinus Centriscus Ceramaster Champsodon INDEX OF GENERA. A 264 372 258 Chilota Chondraster Cladaster . Clypeaster Cochlicella Coecilioides Coelopleurus Coenopedina Comanthus Cominia Congermuraena . Coronaster Coscinasterias Srossaster . Crotalometra Cryaster Cryptopelta Cucumaria Culcita Cyphosus . Dentex Diastobranchus Dictenophiura Diplopteraster . Dipsacaster Dorcasia . Dysomma. Echeneis . Echinaster Echinocardium . Echinocucumis . Echinocyamus . Echinodiscus Echinolampas Echinometra Echinosoma Echinostrephus . Echinothrix 161, 68 441 361 300 246 179 443 79 290 405 418 393 394 397 390 375 387 373 Xil Echinus Eisenia Eodrilus Epinephelus Eremias Eucidaris . Eulota Gobius ; Gorgonocephalus Helix Helodrilus Henricia Heptatretus Hippasteria Holothuria Hoplichthys Hymenaster Tsidora Kaliella Laganum . Lamprechinus Latilus Lepidotrigla Leptychaster Lethrinus Leucochiloides . Limax Limnaea Linckia, Liparometra Lophaster . Lophius Lovenia Luidia Luidiaster Lutianus . Marginella Marinula G M Index of Genera. PAGE | 98, 384 61 45 217 370 185 | 80 318 187 61 289 439 | 270 421 74 300 190, 193 102 Marthasterias Mediaster | Melania 66 Microchaetus Milax ; Monachocrinus . Myxus Nannocharax Nardoa Neobythites Nucras Ophiacantha Ophiactis . : Ophiarachnella . Ophichthys Ophiernus Ophiochiton Ophiocnemis Ophiocoma Ophiocten Ophioderma Ophiomastix Ophiomisidium . Ophiomitrella Ophiomusium Ophiomyxa Ophionereis Ophioplocus Ophiopsammium Ophioscolex ; Ophiothamnus . Ophiothela Ophiothrix Ophiura Ophiuropsis Oreaster Orechinus Ostracion . Pachylometra Paracentrotus Parasterina Parechinus Pectinaster Pelodrilus | Pentacta . Pentametrocrinus Periophthalmus. Perissasterias Phormosoma Phyllophorus Planktothuria Planorbis . Platyglossus Plutonaster Polita Poraniopsis Porcellanaster Pourtalesia Prionocidaris Pseudarchaster . Pseudocucumis . Psilaster Psolus Pteraster . Pupisoma Raia Regadrella Retaster Rumina Salenia Salmacis . Schizaster. Selachophidium . Spatagobrissus . Spatangus : Sphagebranchus Stereocidaris Stichaster Stichopus Index of Genera. PAGE 417 420 189 86 242 182 289 239 399 370 253 417 248 418 298 186 Stomopneustes . Subulina . Syngnathus Synodontis Temnopleurus Testacella Tetrodon . Thyone Tiara Tosia Toxopneustes Trigla Trigonephrus Tripneustes Tropholampas Tropiometra Trypauchen Tulbaghinia Urechinus. Vallonia Vitrea Zonitoides 398 186 182 183 et a. ; aan i ct reer a) a ert | : 1 S- a> : inte pie bedeer, WiLit Ty’ iP at ' LPs ie iy Path Thorn ere ie ¥ : st 7 ., an ae ae ANNALS OF THE SOUTH AFRICAN MUSEUM. (Von. XIIT.) 1.—Inscriptions left by Harly European Navigators on their Way to the Hast.—By L. Prrinaury, Director. IN a country as recently occupied permanently as 1652, relics of very ancient history cannot be plentiful. Yet, earlier European pioneers landed on its shores as far back as 1485; others even annexed Table Valley to the Crown of their King, before the Dutch founded their settlement; and these forerunners of the present Colonists have left a few relics which are the more precious because of their rarity. These are exhibited in part of the Entrance Hall. PoRTUGUESE COMMEMORATIVE PILLARS AND INSCRIPTIONS. Earliest in date is the pillar termed by the Portuguese ‘ padrao,” erected by Diogo Cao in 1485 at Cape Cross, the second being that erected by Bartholomew Diaz at Angra Pequena in the year 1487. Both these localities are now part of German South-West Africa. Within a short period, dating from the end of the fifteenth and the beginning of the sixteenth centuries, three geographical events of unique magnitude took place. Christopher Columbus discovered America on the 12th of October, 1492, some four years after Dias, whom he met in Lisbon on the latter’s return, had doubled the Cape. Also, five years later, 1497-1499, da Gama succeeded in reaching India, and Magellan in the first voyage round the world I 2 Annals of the South African Museum. connected the western and eastern extremities of the Old World. All these discoveries were the results of attempts to discover a sea-road to India. Even at his death, Columbus was persuaded that he had reached Asia. Bent upon discovering this sea-way, the Portuguese navigators, mostly hugging closely the coast of Africa, had persistently proceeded southwards, thus gradually going beyond Liberia, the extreme limit reached by -the Carthaginian Hanno in his celebrated voyage of discovery or settlement along the north-west coast of Africa, about five hundred years before the Christian era. His fleet is said to have consisted of sixty large vessels, on which were embarked thirty thousand persons of both sexes. But nearly a century before the Portuguese entered upon their grand career of discovery a chartered company of Dieppe and Rouen merchants did, between 1364 and 1413, in the reign of Charles the Fifth of France, send expeditions to the Gold Coast. When John the Second, King of Portugal, ascended the throne the efforts to reach this goal, India, were not relaxed. Diogo Cao and Affonso d’Aveiro were commissioned to go farther south; Cao reached the River Congo, or Zaire, and ultimately Cape Negro (Cape Cross) in 21° 41' §. There he erected a padrao, which, however, unlike those put up later on by Bartholomeu Diaz, seems to have had no special name. The scanty records of the period left it doubtful where this last pillar of Cao had been set; but in 1893 the captain of the German man-of-war Falke discovered this relic on Cape Cross. It bore two inscriptions in a sufficiently good state of preservation not to offer any difficulty in deciphering; and according to L. Cordeiro the two inscriptions are as follows :— «(A)mundi creatione fluxerunt anni 6684 et (a)Christi nativitate 148; ? 9(uum) (e)xcelenti(ss)imus (s)erenissi(mus)que Rex d. Johannes secundus portugal (iae) per ia(co) bum canum ejus militem colu(m) nam hic situari jus (s)it.”’ (Six thousand six hundred and eighty-four years had elapsed since the world was created, and 148? since the birth of Christ, when the most excellent and most serene King, D. Joao the Second of Portugal, ordered this column to be set up by Jacobus Canus, his knight.) The second inscription, which follows the turn of the upper cylindrical part of the column, is in Portuguese— Inscriptions left by Early European Navigators. 3 “Hra da creacao do mundo de bjMbjclxxxb e de X (to) de llllelxxxb o. eycelent(e) esclarecido Rei dom J s°. de portugal mandou descobrir esta terra e poer este padram por d(°c.) a0 eav’. de sua casa.” The translation into English would be— ‘Tn the year of the creation of the world, 6684, by the excellent and enlightened King Dom John the Second of Portugal, was Fie. 1. ordered the discovery of this land, and the erection of this pillar by Diogo Cao, a knight of his house(hold).” [For Diogo Cao had been made a knight in 1484 as a result of his first voyage, which led to the discovery of the River Zaire, now usually termed Congo. ] By order of the Emperor of Germany this pillar was removed to Germany, and a replica of the same erected where the original stood. Through the courtesy of the German authorities the 4 Annals of the South African Museum. Trustees of the South African Museum have the promise of a facsimile of the original padrao. In 1487 King John the Second of Portugal fitted out another expedition, consisting of three vessels, two of some fifty tons’ burthen, the other being smaller and used as a store-ship, and this new venture of discovery he placed under the command of Bartholomeu Diaz. It is during this reign that the discoverers were, for the first time, provided with commemorative pillars, or ‘‘ padrdes,’”’ to be erected at the farthest point reached, or to mark the progress of their journey. Cao is the first navigator who left Portugal with these regulation pillars. But the old chroniclers are certainly not clear about the number of padroes erected by that other bold navigator, Bartholomeu Diaz, who was to round the extreme part of Africa. Recent research has, however, brought forth conclusive evidence that he erected five pillars during his memorable journey, each having its respective name. The exact dates of the erection, corresponding mostly with his landings, are only approximate, but they are as follows, according to Codine :— Departure from Lisbon, August 1-14, 1487. Landing at Angra Pequena and erection of the pillar called Padrao Santiago, November 13-14, 1487. Landing at Angra das Voltas, November 19-24, 1487. Erection there of a third but nameless padrao. This bay, the translation of the name of which is “bay of tackings,”’ owing to Dias’s ships taking five days to reach the shore, was taken to be in 28° 44’ lat. south, and corresponds nearly to that of the present mouth of the Orange River (28° 58’ S.). At this place Dias left his store-ship with nine men. But as he called there on his return it is not certain if the pillar was erected at the above- mentioned date, or on his return (August 24, 1488). Arrival at the Bahia of Vaqueiros, or Bay of Cowherds, and at the Bahia of San Braz (Mossel Bay), apparently towards the end of January, 1488. He was certainly there on the 3rd of February of that year. It has been suggested that Cowherds Bay and San Braz Bay are identical, the date of his arrival corresponding to that dedicated to St. Blaise led to it being renamed as such. Reaches Algoa Bay, sets up a third pillar (Padrao da Cruz), on a small island of that bay, middle of February, 1488, Inscriptions left by Early European Navigators. 5 Reaches the River Rio Infante; thence retraces his way to Algoa Bay and erects a fourth pillar, Padrao San Gregorio, on Cape Padron, to the east of the bay, February, 1488. Discovery of the Cape of Good Hope (Cabo tormentoso),* and npr vommeesvemmmertaasntn serene Servet Mie. 2) 22 cm. x 58 cm. erection there of a fifth pillar, Padrao San Felipe, August 16-17, 1488. Exploration of St. Helena Bay, August 18, 1488. Return to Angra das Voltas, where the only survivor of the men he left there nine months previously expires at the sight of his comrades. Departure for Prince Island, and arrivalin Lisbon, December, 1488. * Dias, in spite of the legend, never named this part otherwise than Cape of Good Hope. 6 Annals of the South African Museum. Of the five padroes here mentioned as erected by Dias, only one of them has hitherto been found, namely, Padrao Santiago, The fragment exhibited is part of it. It was originally erected on the summit of a small granite eminence, and was discovered by Captain Owen in 1833, but “ cast down evidently by design as the part of the shaft that had originally been buried inthe rock had remained unbroken.”’ This pillar, includ- ing the part originally placed in the ground, would have been altogether 7 feet 9 inches in length, corresponding in height with that erected at Cape Cross by Diogo Cao, and “ was composed of marble rounded on one side, but left square on the other.’ The cross surmounting the pedestal was found at some distance. ‘It was sixteen inches square, of the same breath and thickness as the shaft, and had on the centre an inscription almost obliterated.” Three pieces of the original Padrao Santiago were received at the Museum in 1856, two of which, part of the rounded side, were sent to Lisbon; but a replica of the same has been made for us, presented by the Museum of the Geographical Society of Lisbon; a fourth fragment is to be found in Auckland, New Zealand, left there as a gift by a former Cape Governor, the late Sir George Grey. The cross itself has not been recovered. PORTUGUESE INSCRIPTIONS. Vasco da Gama was the next navigator who was to complete, ten years later, the exploration of Dias, and to reach India (1497-1499). It seems, however, that the erection of padrdes had then fallen into disuse, for there is, I believe, no record of any put up by this explorer, who, it is now almost certain, retraced the itinerary of his predecessor, Dias. Other expeditions were to follow on the way to the Hast Indies. Pedro Alvarez, better known under his nickname of ‘ Cabral,’’ in trying to double the Cape, discovered Brazil. One of his captains, Pedro de Atayde, separated from the fleet by a storm, reached the Bay of San Braz (Mossel Bay), and left there a letter in a shoe, placed, it is said, on the island in a conspicuous situation, and which was found by Joas da Nova, who had sailed from Portugal on the 5th of March, 1501, in charge of four vessels. Stone I.—The Mossel Bay Stone. The fragmentary inscription on a stone found in Mossel Bay, appears to be a record of the visit of one of these two expeditions. ~I Inscriptions left by Karly European Navigators. ‘“ At the demolition of the old Government House, there were found two stones on which were engraved, on the one a cannon, and onthe other Portuguese words, of which some were broken off, others were indistinct, yet without doubt are the names of a certain ship and its captain, also the time of the arrival here, being the year 1500 or 1501.” The stone with “ the cannon”’ seems to have disappeared, or it is no longer to be found. The inscription on the Museum specimen is, however, very baffling, owing to its present incompleteness. Hie. 3. fitcmyx< 2 em: It is skilfully graved, however; and if we assume, as is highly probable, that “DA NOVA” or ‘“ NUEVA” was spelt, by mistake or otherwise, Novoa, and that BRA stands for Braz (the Portuguese name Sao Braz), the inscription refers to the call at Mossel Bay of da Nova’s squadron in 1501; in spite of the graved date, which appears to be 1500. This, however, has no importance, because 1500 in the old style may partly correspond to 1501 new style, the year running then from end of March. If da Gama has left in Africa other inscriptions, and if they are found eventually, they would not be much more ancient than that of the Mossel Bay Stone. The latter can, therefore, be looked upon as the third most ancient European relic hitherto found in Southern Africa. 8 Annals of the South African Museum. Stone II1.—The Plettenberg Bay Stone. The second Portuguese inscription, known as the Plettenberg Bay Stone, was removed to Cape Town about the year 1860. It is stated that the stone originally stood on a sandhill about three miles south of the present village of Plettenberg, on the littoral of the Cape Province. TGA 52 cm. x 47 cm. This inscription reads, ‘‘ Here was lost the ship Sao Gonzales. Year 1630. They made two boats.”’ The Sao Goncalo, or Gonzales, reached India on the 24th of September, 1629, and left for the kingdom, 7.e. Portugal, on the 4th of March of the year following. Fernao Lobo de Menezes was then captain of the vessel, and the latter having sprung a leak, he made for the land, and ‘came to the bay called Fermosa, on the confines of the Cape of Good Hope.” But while they were endeavouring to pump the vessel dry and trying to right her, a storm came upon her while at anchor and she foundered. Inscriptions left by Early European Navigators. 9 All the people left on her (one hundred and fifty) perished, but one hundred of them were ashore at the time of the gale. They eventu- ally built two boats with the debris of the vessel, one of which set sail for Mozambique, where it arrived safely ; the other reached the Cape (Table Bay), and sighting there the Sao Ignacio de Loyola, of the fleet of 1630, were taken on board, but this vessel perished upon the bar of Lisbon. Such is the abbreviated history of the three Portuguese relics in the Museum. FRENCH, EnauisH, DutcH, AND DANISH INSCRIPTIONS. After de Gama’s discovery, Cabral, da Cunha, Albuquerque, Almeida, Sequeira, and other explorers, all Portuguese, visited the eastern seas and the Islands of Spices via the Cape. The Spaniards, by now a powerful maritime nation, did not follow on their track, because the Pope had arbitrated on the respective sphere of both Portugal and Spain. The ventures were very lucrative, as proved by the number of ships sent from Portugal: 507 from the year 1500 to 1550, and 264, of much larger tonnage, from 1550 to 1560. But a few years only after da Gama had opened the road to India other nations are found to have entered this newly discovered field. Privateers or merchantmen, or both together, began to operate in the Mozambique Channel and other regions, and these were French. In 1508 Queimado, commander of one of the ships of Tristan da Cunha’s fleet, was captured by Frenchmen in the Mozambique Channel, In 1560 Captain Bondard, from La Rochelle, was hanged at Mozambique for plundering Portuguese caravells in the Indian Ocean.* Of three French privateers that sailed from Dieppe in 1526, one is known to have stopped at Madagascar, and to have done some trading there. So that it is inexact to say that during the period 1500 to 1560 no Kuropean flag, other than the Portuguese, was seen in the Eastern seas. But their expeditions did, after a time, sail from the island of St. Helena without touching at Table Bay, and therefore left no inscriptions there. FRENCH INSCRIPTION. Stone ITT. On one stone is a French inscription with the date un- fortunately mutilated. This piece of rock has been badly used; * The Portuguese claiming a monopoly of their discoveries under a Papal Bull, the operations of any competitor were considered by them to be piratical. 10 Annals of the South African Museum. the date is missing, also the continuation of the four lines of letters. On the reverse of the thick slab is a Dutch record dated 1634. On examination it becomes apparent that the block on which the French inseription stood was pared or reduced so as to allow of the new one, which is entire, being graved on the reverse. But how long the first preceded the second, and whose record it is, remains, so far, a mystery. Paulmier de Gonville is believed, with good reason, yet without much documentary evidence, to have rounded the Cape in 1503, and to have reached Madagascar in that year. But he sailed from Qe, 57 cm. x 24 cm. ICY EST ARRIVE DAVID DIGAED DE DIEPPE 8 10 DE FEVRER lL (Here arrived David Digaed from Dieppe, 8-10 February, 1. . .) Honfleur, not Dieppe, in June. The inscription cannot be, there- fore, ascribed to him. There is no information about the vessels who were flying the French flag in 1508 in the Mozambique Channel, and even captured there, as stated before, one of the ships of Tristan da Cunha’s fleet. In 1527 a French vessel, one of a company of three, all from Dieppe, stopped at Madagascar, traded there, and left behind a sailor, whom Diogo de Fonseca picked up in 1531. The brothers Parmentier, also of Dieppe, following the Cape route, left with two ships, La Pensée (400 tons) and Le Sacre (120 tons), on Inscriptions left by Early European Navigators. 1 March 28, 1529, and reached Sumatra the same year, where one of the brothers, Jean, died in December. But apart from the name of the home port, there seems to be no connection with that of the ships or the dates. The only instance of French vessels being recorded near Table Bay, but not iz Table Bay as is generally believed, in these early times, is by the Dutch Commander Spilbergen, a record corroborated by the narrative of Francois Pyrard, from Laval, who was on board an expedition from St. Malo, consisting of two vessels, Le Croissant and Le Corbin, under the command of La Bardeliére. The accounts of both leave no doubt that the French vessels met the Dutch Commander sailing out of the Bay, and did not land. It does not, therefore, seem improbable that this French graved record is that of one of the vessels that left Dieppe in 1526; in which case this is the oldest inscription other than the Mossel Bay. Post OFFICE STONES. From the end of the sixteenth and the beginning of the seven- teenth century many vessels, other than Portuguese, touched at Table Bay in order to obtain fresh provisions from the natives. Anthony Marlow, on board the English ship Hector in 1602, says, speaking of Table Bay, ‘the best and cheapest place to refresh men in these voyages that ever ship can come to”; and it is recorded that on the first voyage of an English squadron to the Hast Indies begun under command of tae Admiral G. Raymond, who perished with the Flag-ship, and finished by Captain James Lancaster, they put into Table Bay where ultimately thirty natives brought forty bullocks and about as many sheep, of which the English got a good supply, giving two knives for an ox. ‘These vessels left graved on flat stones inscriptions recording the name of the ship and of her captain and the dates of arrival and departure. Letters were often also placed beneath these stones, as borne out by the inscriptions. In these early days a stream descending from Table Mountain ran to the sea, discharging its water into Table Bay near what is now Adderley Street, and there it was that the mariners landed, to fill their water-casks at or near the bottom of Strand Street, where was a large sand-dune continued as far as the present Green Point Common. Most of the inscribed stones recovered have been found near what is now Adderley Street, with the exception of two English which were used, intentionally or not, in building the walls of the ‘ Castle” 12 Annals of the South African Museum. with the inscription outwards. But as from 1602 the vessels of the English and Dutch Hast India Companies called nearly every year at the Cape, and as moreover the captains of the English vessels were instructed to leave such records, it is possible that grayed stones other than those now recorded will be found at some future time. While digging foundations for an extension of the present railway station in Cape Town in 1906, the old landing-place at the foot of Adderley Street was uncovered, and a number of graved stones that had evidently been collected and brought to the spot in former days were exposed to view. These stones, and others recovered before, form the series bearing the name of “ Post Office Stones.” An extract of a letter addressed by Edward Blitheman to Sir Thomas Smith (Hast India Company’s Records) leaves no doubt as to the object for which these stones were inscribed, and seems to explain also the presence of the two inscriptions in the Castle at some distance from the customary place : “And in the time of our being there (Table Bay, October, 1613) the Dutchman (also in the Bay) made known unto us a packet of letters which their company had found on the top of a hill. So our General sent myself and Mr. Millward for the fetching of them, being a place at least distant two miles from our tents. So finding them we perceived them to be the letters of the factors of Captain Downston’s fleet . . . and afterwards our General sealed them up again in a letter of his directed to your worship and buried them by the stone where he placed his name.” The French Commander Beaulieu, who sailed on the 2nd of October, 1619, from Harfleur, in Normandy, on a voyage to Bantam, vid Senegal and the Gold Coast where he traded, landed in Table Bay on the 16th of March, 1620, and he writes thus :— “Some of our men going ashore happened to light upon a great stone, with two little packets of pitched canvass underneath, which we afterwards found to be Dutch letters. When we opened them we found first a strong piece of pitched canvass, then a piece of lead wrapped round the packet; under that two pieces of red cloth, then a piece of red frieze, all wrapped round a bag of coarse linen in which were the letters very safe and dry. They contained an account of several ships that had passed that way ; particularly of an English advice boat that was gone to England to acquaint the Company with the injury the Dutch had done them in the East Indies. They likewise gave notice to ships that passed that way to take care of the natives who had murdered several of their crew, and stolen some of their water-casks. ’' Inscriptions left by Early European Navigators. 13 This narrative of the French Commander throws, in addition, a singular light on the dangers attending at that time landing in Table Bay, for he adds: “The next day I sent fifty men on shore with sails to make tents of; when the boat returned they told me they had found several corpses of dead men and clothes scattered up and down, and a small fortification of earth which we guessed to be built by the Danes, for one of the natives that spoke a sort of jargon of broken English gave us to understand more by signs than by his language that five ships had sailed from thence to the eastward about three months before.” ENGLISH INSCRIPTIONS. Stone IV. If the French followed very early in the wake of the Portuguese, such cannot be said of the English, for it is only in 1577 that the famous sea Captain Drake, and, nine years later, Thomas Candish, sighted the Cape; but they did not land. In July, 1591, however, the fleet of Admiral Raymond put into Table Bay, and on the 22nd of April, 1601, the first fleet fitted out by the ‘‘Governor and Company of Merchants of London trading to the East Indies,’ and commanded by Sir James Lancaster, sailed from Torbay. It consisted of the Dragon (600 tons); the Hector (300 tons); the Ascension (260 tons); and the Susan (240 tons). It is not known if Admiral Raymond, or Laneaster, left any inscribed stones to denote their landing in Table Bay, but the ship Hector of Laneaster’s fleet was again at the Cape homeward bound in 1605, i.e. 1606 present style, as proved by the graved stone No. 6. Antony Hippon, who was mate or master of the Hector, and had put into Table Bay in 1605, did again call at the Cape as mate or master of the Dragon in 1607. He looked for and found his first inscription, and added to it the date of his second arrival or departure. He was in charge of the Globe in the seventh voyage, and reached the Cape on May 21, 1611, sailing hence on June 6th. Possibly the name Anthony H. in smaller letters on the lower part of the slab is a record of this occasion. It is this Captain Hippon who planted the first English factory on the mainland of India (Masulipatan). He died on board the Globe one month after leaving Table Bay. The stone bearing this inscription was discovered lately, embedded in one of the walls in the Castle, and it is the oldest in date of the English records of call in Table Bay. 14 Annals of the South African Museum. These two vessels, the Dragon and the Hector, were often navigating in company, and numerous indeed have been their calls. They were in Table Bay from September 9, 1600, to October 29th on their outward voyage. Lancaster commanded the former, J. Middleton the latter. Both vessels were again in Table Bay in July, 1604, and also on their return voyage from December 27th to January 16th of the same year. PANTO OF HAH Ur MIRE i ANT.HI IPPOK eon — a8 CE CEeeENE O07 Fia, 6. 82 cm. x 80 cm. ANTO HIPON MA(STER) OF THE HECTOR BOUND HOME JANUARY 1605, ANTO HIPPON MA(STER) OF THE DRAGON 28 DECEMBER 1607, and in smaller letters, Anthony H. The Hector was in Table Bay in 1602, as shown by a letter of Anthony Marlowe already quotod. We find again the Hector in the Bay from December 18, 1607, to January 1, 1608; and the Dragon, possibly on some date prior, homeward bound. Captain John Saris, commanding the Clove, Hector, and Thomas, outward bound, sailed from Table Bay on his outward voyage on August 9, 1611. The Gift, Hector, and Hope were in Table Bay in 1614, Inscriptions left by Early European Navigators. 15 William Edwards, writing to the East India Company on June 28, 1614, says: ‘ Found, as the custom is, advice of divers of our ships that had lately touched there (Table Bay) both outward and home- ward (viz.) the Dragon arriving there the 20th February, 1613, departing thence the 6th March next, homeward.” The Dragon was one of the vessels of Keeling’s fleet of 1615, and had on board the factor, Thomas Arthington, who in a letter to the Company dated from Saldania, June 18, 1615, and sent by the home- ward-bound vessel the Hope, mentions that ‘Ten of the condemned men set willingly ashore at Saldania.” From 1601 to 1612 the shareholders or subscribers to the concern that became eventually the East India Company, founded on December 31, 1600, bore individually the cost of the voyage of each venture; and the profit or loss resulting therefrom was theirs. It is only in, or about, the year 1612 that these ventures were conducted on the joint-stock system, and it is on record that the Concord from London, the first ship of the Company sent on this new system, was met in Table Bay by Captain John Saris, homeward bound, on May 16, 1614. For this reason the records of ‘‘calls” anterior to 1612 are not very complete, and the dates are possibly somewhat inaccurate, owing to the then partly acceptance of the Gregorian computa- tion, 25th of March to 24th of March, instead of Ist January to 31st December. It seems, however, established that the Red Dragon, Hectcr, Ascension, and the Susan, under Sir Henry Middleton, anchored July 17, 1604. ‘‘Cole, master of the Hector, was drowned. Sailed August 20th.” This was the second voyage of the English Company. The upper inscription of Stone 6 records the return voyage of the Flector. In 1605 the English Captain Sir Edward Michelburne came to Saldania Bay (read Table Bay) with the Tigre and Tigre’s Whelp, and stayed there from April 9th to May 3rd. The natives brought him so much cattle that he took some to sea; for a piece of an old iron hoop he bought a large bullock. In 1607 William Keeling came to “ Salania,”’ and bought much cattle. He found the natives very bold and daring. On a rock he saw engraved the words: ‘“ Captain Middleton, in the Consent, 24th July, 1907.” In the same year Captain D. Middleton, a brother of Sir Henry of that name, called here also, and obtained much cattle. 16 Annals of the South African Museum. Captain Robert Cavertree also came in the same year to Saldanha, where he bartered many sheep and oxen for old iron. Alexander Sharply was in Table Bay on July 4, 1608. He took in about 400 (?) head of cattle. In 1608, July 13th, the Ascension and Union anchored in Table Bay, and left on September 25th. ‘«« Viewing over the stones where the ships that are bound outward and homeward do use to set their names, when we found Captain Keeling, Captain Hawkins, Captain Middleton and divers others, being passed towards the Indies, vizt., Capt. Middleton in July 1607, and Captain Keeling the month of December ditto anno.”’ In the year 1609 Captain Keeling on his home voyage took some sheep, ‘“‘the fattest he had ever seen,’ from Robben Island, and left lean ones, obtained on the mainland, in their place. In 1610 Captain Nicholas Downton, with the Peppercorn and the Darling, and Sir Henry Middleton in the Trades Increase, arrived in Table Bay. Sir Henry Middleton found a letter buried underground, according to agreement made in England with his brother, but so soiled by damp that he could ‘not read any part thereof.” Downton touched again, outward bound, on June the 15th, 1614; but he was not so successful this time in revictualling his ship ‘‘ Refreshing of flesh we had in a manner none; we had some little fish by our continual endeavours in the river. Set sail and put to sea, the 2nd of July.” Keeling’s fleet of 1615 was also in Table Bay outward bound in that year. From it were landed at Robben Island ten convicts from the Old Bailey, to make a settlement. These are the men alluded to by Thomas Arthington. But no inscription left by the commanders of the ships above mentioned has been as yet recovered, with the exception of one of Sir Henry Middleton’s ships, the Hector, and of Saris’ vessel, the Thomas. Stone V. In preparing the foundations for one of Mr. Garlick’s stores, at the foot of Adderley Street, the workmen found a fragment of stone bearing an inscription on each side. The original stone is in the possession of the Cape Town Corporation. The date of the arrival and departure coincide with that of the outward-bound fleet commanded by John Saris, of which the Thomas was one, together with the Hector and Clove, and it is therefore to be assumed that the inscription is that of the Thomas. It will be noted that in one corner of the stone there is what Inscriptions left by Early Huropean Navigators. 17 Fig. 7. ‘ 32 cm. x 21 cm. On the obverse the words read as follows :— THOMA Mr. M. Saee TH. BARN. Mr. MAT. JULY YE auGusT 1. (?)D. c. #. Iniel, {8 30) Gms ems WILLIAM PAGE, OUTWARD BOUND, 1618. 2 18 Annals of the South African Museum. appears a fairly successful attempt to reproduce the “square and compasses ’’ in a position in which these implements are usually associated with Freemasonry; certainly the first figures of their kind revealed in South Africa. On the reverse of the stone the impression is very plain. This is the first record of call of this vessel. ) Another will be found further on bearing the date 1628. Stone VI. The next inscription in point of date is that of the arrival and departure of the fleet commanded by Charles Clevenger (the C.L.E. of the legend), who flew his flag on the Palsgrave. The latter, 1,083 tons, and Elizabeth, 978 tons, were new ships built by the East India Company in the previous year. Fia. 9. 68 cm. x 51 cm. CHARLES CLE CHEIFE COMAD O(F) PALSGRAVE ELIZABETH AND HOPE ARR(IVED) Y® XXIIII JUNE AND DEP FOR BANTA Y" XX JULY 1619. THOMAS BROCKEDOM CAPE MERCHANT OF Y= PALSGRAVE. This inscription records for the first time the presence on board the Commodore’s ship of a ‘ factor’’ styled there ‘‘ Cape Merchant.” The status of these supercargoes had at last been established, by Inscriptions left by Early European Navigators. 19 order doubtless of the East India Company. That the relations between the commanding officers of the vessels and the factors who had really charge of the commercial side of the ventures were not always of the most amicable kind would seem to be proved by the following extracts: the complaints to the Company emanating from the factors (Hast India Records, 1613-15) :— “Tt is good to distinguish or limit the officers and charges of captain and cape merchant, for the captairs do arrogate all authority to themselves from your merchants.” Again. Robert Gipps, Cape Merchant, in the Peppercorn, 19th June, 1615, in Saldania, to the Hast India Company.—Discord between Captain Harris and Robert Gipps. The Captain reviled the Merchant and threatened him. ‘The Captain arrogates much over the Merchants. He brings £100 to sea for private purposes.”’ This falling out between Captain Harris and the Cape merchant was, however, smoothed over, and the two men reconciled by the Council of the fleet. Thomas Brockedon, the Cape merchant on board the Palsgrave, subsequently became the chief agent of the British Hast India Company at Batavia. Although the inscriptions give the date of departure of vessels, this does not necessarily imply the absolute date when the ships sailed. Thus, a factor named Mills, writing from Tiku, in Sumatra, which was reached on August 23rd, says :— “The 38rd of August before day we sett sayle from the Cape where we were 16 dayes wynde bound.’ The vessel had evidently been retained in the roadstead by adverse winds for fourteen days. At the Cape they had met the Lesser James, homeward bound. This No. 9 inscription, cut into solid rock, was found a few feet below the surface, when Messrs. Wilson, Miller, & Gilmore, of Adderley Street, were rebuilding their premises in Adderley Street, opposite the Post Office. Stone VII. Four months after the departure of Clevenger outward bound, and of the Lesser James, returning to England, another English vessel recorded its arrival and departure from Table Bay. This stone bears an inscription on each face. The oldest in date is that of the Bull of 400 tons, which touched at Table Bay in 1619, but not for the first time, because one reads in the “ English Factories in India’’ that the ship was sent home in 1618. On the reverse of this stone are graved letters which, judging ’ 20 Annals of the South African Museum. from their shape or style of cutting, are parts of four different inscriptions. All four, however, are very obscure. Of the first inscription on the upper side of the stone, the figures 161 alone are distinct, and these seem to be completed by the figure 8, which would thus give the year as 1618. i E My | De OPO BE: | Fic. 10. 40 cm. x 38 cm. RO" ADAMS COM?®: OF THE. BUL ARIVED 29 OF NOVEM & DEP®” THE 12 oF DECc® 1619 FOR BANTAM 1O*® COCKRAM. CAPE MARCH. LETTERS UNDER. Below this the second inscription, which is very distinct in parts, reads as above. Below this are two lines graved in a different style and undecipher- able, and lastly, on the lower end of the block, and executed again by another hand, a very rough graving— WILL VOR LETTERS 1629. The ship William had returned to England in December, 1628 ; but she might have touched again here in 1629. Inscriptions left by Early European Navigators. 21 CAE. 42 18! ,O\ PAI ee i N ARP THE 23 Ol "DFP FOR: SVRRAT T Hi {YVNE *: as DANMIL Ht CEIEE, BIKE NArG Ere. 1 42 cm. 37 cm IDANG INL ogous apeocde Vipy, auEe0n) Drisccosoovcan. DEP FOR SURRAT THE............ MOBS OA Iiososgacnmeac XII IIIT GEORGE PIKE MARC As to the better-preserved inscription, in which the name of George Pike figures plainly, a person of this name was a Factor of the Company. Stone VITTI. The stone bears two inscriptions. (See next page.) The slab is deposited in the vestibule of the General Post Office, Cape Town. The specimen on exhibition in the Museum is a cast of the same. The London, a vessel of 800 tons, Captain Richard Blyth, with the Jonas and the Lion, sailed from Surat, December 18, 1622, anchored in ‘“‘Saldania Bay, March 10, 1622-3; left again March 23rd, not 20 as inscribed, reaching the Downs, July 18, 1623.” The second inscription would appear to be a Dutch one. This stone was found in 1897, when an excavation was being 22 Annals of the South African Museum. made in the ground immediately in front of the then recently com- pleted offices of the Union Castle Company in Adderley Street. It had, however, been discovered before, but was evidently left in s7tu until it became again hidden from sight, for we read in the Cape Town Gazette of Friday, August 17, 1827: “On removing the earth to make some repairs to a drain in the Heeregracht (now Adderley Street) a large stone was uncovered, upon which the following inscription was traced,’ and a correct transcription is there given. FF LONDON*cARINED FE el 0 e OEM LERFo FROM SVRATe BOVNDEOR 2° ENGLAND AND» DE PAR*EEoD1O DICTO RICHA) BLYH © CAPTAN, 16 i Ma LEARE VNDER LOOK! FOR LET TERS 1GZ9 JauRey"C LOCK SAS V BERIVGH 7 | Mire. 127, 105 em. « 60 cm. The upper inscription reads :— THE LONDON . ARIVED THE. 10. OF M(ARCH) HERE . FROM SURAT. BOUND . FOR . ENGLAND AND . DEPAR(TED) THE . 20 piTTro . 1622. RICHARD BLYTH CAPTAIN. HERE UNDER LOOK FOR LETTERS. Below this— 1629. JAN. REY™ CLOCK (OR CLOOK) GASP V BERING™=" H, His Stone LX. The London a few years later met the Blessing and the William at the Cape both inward bound, and a record of this call exists in an Inscriptions left by Early European Navigators. 23 inscription on a very uneven rock, and graved by a hand but little acquainted with cutting letters on stone :— THOMAS WALIS WILLIAM HARVEY MYSMATES OF THE ONDON 8 OF MAY 1627 JOHN SHORT A M. The same ship was in Table Bay in 1631; as shown by the inscription graved on the obverse of a flat stone bearing on the other side a Dutch inscription dated 1632 (Stone XVII.) RICHARD AL NVT EF Com OF LOND OMAR VED.MLY 4 ee? Dies ANS}. 56 cm. x 33 cm. RICHARD ARNOTT COM, OF LONDON ARRIVED JULY 4. AN 1631. DEP. XXII. ? MIIH 24 Annals of the South African Museum. Stone X. has three English inscriptions cut on a large slab found built, topsy-turvy, in one of the outer walls of the Castle in Cape Town, about ten feet above the glacis, and not far from the main gateway. There seems to be little doubt that this stone was lying close by, and was utilised for the original building of the IOMN-ROBERTS COMMAVNDER OF Tas ae SSER: AMES ARY8 - DECEM-pES§ 261622 »LOVKWITHAHS LINE FOR LETERES : HE NR’MA Aye SER a. 1910 si x » Hedley & Suter, Proc. Mal. Soc.ix.p.151. L. Shell small, elongate-ovate, imperforate, solid, epidermis purple- black, interior and columella paler. Spire moderately produced, about two-thirds the length of the aperture; sides rather convex, apex minutely pointed. Whorls 3%, not very convex, rapidly increasing, with fine, irregular striation following the lines of growth. Suture very shallow. Aperture acuminate-ovate, outer Notes on South African Mollusca. 109 lip thin, simple, inner furnished with three white dental processes— (i) a prominent, linguiform tooth parallel to, and about 1:7 mm. from the outer lip; (ii) a short, rather conical tooth about half-way between the upper one and the base; (iii) a minute columellar fold rather nearer the second tooth than the base. Long. 10-4; lat. 5:5; apert. 67x 4:7; last whorl 9:6 mm. MaRINULA TRISTANENSIS, Sp. NOV. Epidermis purple-black ; lines of growth eroded white. Mr. G. C. Robson has very kindly examined the anatomy, on which I am therefore enabled to present the subjoined report :— On the Anatomy of Marinula tristanensis.—By G. C. Rosson, B.A. (Published by permission of the Trustees of the British Museum.) It was hoped that the ample material representative of this species, received for examination from Dr. Péringuey, would have afforded an excellent opportunity for giving a complete account of the anatomy of this interesting form, the generic position of which is now treated for the first time from the anatomical standpoint. It is to be regretted, however, that the specimens obtained were in such a state of contraction as to render dissection of anything more than the gross anatomy an impossibility. It was possible to make out a certain amount by sections stained with Haematoxylin and Methylene Blue, but the time available for continuing this method not being forthcoming the author is compelled to publish only such descriptions as may serve to indicate approximately the systematic position of the genus, together with a few other notes. The radula (Fig. 1) differs in many well-marked characters from those of the genera usually regarded as closely akin to Marinula. There are in each row numerous teeth—about 230—the general form and disposition of which readily place it among the Awriculidae. The lateral teeth present a simple rounded main cusp with a slightly hollowed interior edge. The basal plate is produced very far beyond 110 Annals of the South African Musewm. the upper edge of each tooth and in a slanting direction. A transi- tion to the marginal type of tooth is effected by the appearance on or about the thirty-fifth tooth of a small interior cusp on the acutely KAW ah sy i rar i Go $5 90: Y Jeet, Th Marinula tristanensis. Radula. (4 x 3; Homog. imm.) pointed second cusp. The latter gradually divides into two and eventually more cusps appear until at about the 90th tooth four or five small, equal cusps are found arising from a somewhat amorphous basal plate. The radula thus described clearly is distinguishable from that of Melampus and Phytia, although a family likeness is obviously present in all three. The jaw (Fig. 2) is a typically Auriculid structure consisting of numerous loosely arranged fibres. It differs from that of Melampus, however, in being of a regular gently arcuate form slightly upturned at the extremities. At the latter the fibres are better developed and more regularly arranged than in the median area of the jaw. The genitalia (Fig. 3). The organs of reproduction do not offer any singular departure from the usual Auriculid type although they are substantially different in disposition and form from those of Melampus, Phytia, and Pythia. Fig. 2. Marinula tristanensis. Jaw. (4 X 6 obj.) The male genitalia do not differ very markedly from those of the three genera alluded to above in the presence or absence of any essential part. On the whole it is more like the slender delicate Notes on South African Mollusca. iB system found in Melampus. It differs, however, from all three in having a number of muscle fibres inserted into the penis half-way down from the external aperture. The female reproductive system appears to be considerably simpler than in any of the three other genera, and, if the author's Hig, 3: Marinula tristanensts. Genitalia. (2 x a3 Binoc. Zeiss.) , a. Penis. b. Penis muscle. c. Vas deferens; c’, continuation of same. d. Accessory penis muscles. e¢. Vagina. f. Hermaphrodite duct. g. Gonad. h. Spermatheca (?). interpretation is accepted, somewhat specialized, as the vas deferens, instead of splitting off the spermoviduct, has become completely separated from it and issues from the base of an accessory organ * * This organ may possibly be an accessory glandular structure. If it is ulti- mately proved to be the spermatheca, then the absence of other accessory organs in the female generative system compares strangely with the condition seen in Melampus, Phytia, &c. 112 Annals of the South African Museum. a which appears to represent the spermatheca. If this interpretation is ultimately found to be correct the spermatozoa must be carried down the common duct to the vagina where they are separated from the ova and ultimately find their way to the exterior after traversing part of the spermathecal duct. It remains to make one or two disconnected observations upon isolated points of anatomy. (a) Portions of the resorbed shell septa are found in a superficial position in various parts of the body enclosed in epithelial pockets. (>) The anterior edge of the mantle is supplied on the right-hand side with hard semi-calcified pads which no doubt act as protection against the rather formidable teeth of the shell aperture. The anus opens in a groove in one of these pads. (c) The animal is apparently omnivorous, as the lower part of the intestine contained a variety of foraminiferan skeletons together with sand granules and fragments of a red crystalline substance. Hab. InaccessiBLE Istanp, Tristan D’AcuNHA (Keytel (Type) and others). GoucH Is~anp (Scotia Expedition, fide Melvill & Standen). SourH Arrica. Camps Bay (Dale; McBean); Dassen Island (Lightfoot). Type in South African Museum. This species has long been known in collections under the name of Marinula nigra (Phil.), but as the last is merely a synonym of M. pepita, from Chiloe, a new name is necessary for the Tristan form. The shell very closely resembles that of pepita, but the early whorls are rather less convex and the small columellar fold is, on the average, slightly more pronounced than in King’s species, which, from its widely different locality, should surely be distinct. I have not seen the examples mentioned by Melvill & Standen, but have no reason to doubt their identity with trzstanensis, while their appearance on Gough Island furnishes a stepping-stone between Tristan d’Acunha and South Africa. Only three specimens of Marinula, probably referable to tristan- ensis, have so far been collected in the last-mentioned sub- continent. Two of these, picked up at Camps Bay, near Cape Town, are somewhat beach-rolled, while the third, taken alive below high-water mark on the rocky shore of Dassen Island, is so Notes on South African Mollusca. 113 eroded as almost to have lost its shape; but taken together the shells appear to be a little longer in the spire and less solid than the generality of Tristan examples, and the interior is bright corneus instead of purple-brown. However, the difference in colour may be due to the effect of alcohol on the Tristan shells, while the length of spire is not a constant feature among specimens from the last-named locality, so that there are no good grounds for regarding the South African form as distinct from M. tristanensis, unless a larger series can be collected and proves to be specifically different. MARINULA VELAINI, nom. nov., 1915. 1877 Marinula mgra, Phil., var. minor, Vélain, Arch. Zool. Exper. Vine omele DV £25. Ue: Shell small, ovate-acuminate, imperforate, not very solid, dark corneous. Spire short, acute, sides very slightly convex. Whorls 34, very rapidly increasing, very slightly convex, with no sculpture except an occasional growth-line. Aperture large, acumi- nate-ovate, outer lip thin, simple. Columella furnished with 3 dental processes—(i) a long, curved tooth near, and parallel to, the outer lip; (1) a shorter, curved tooth, at right angles to the columella and half-way between the upper tooth and base; (iii) a minute projection just below the foregoing. Dimensions of a specimen in the British Museum: Long. 6:8; lat. 4:2; apert. 5:0 x 2:2; last whorl 6°3 mm. Animal unknown. Hab. Sv. Pavuu’s and AmsrerpAm Isnanps (Vélain ; McGillivray). Type whi ? The Type measured 7:5 x 4°5 mm., and the author states that the maximum size observed was 9:0 x 5-0 mm. Vélain also remarked: ‘This Marinula is identical with that described from the Island of Tristan d’Acunha by Philippi under the name of M. nigra, but it is of much smaller size and must be considered as a var. minor of this species.” But even if M. tristanensis has reached the Atlantic coast of Africa, it is most unlikely to have found its way round the Cape and half- way to Australia ; in addition to which, the disparity in size between it and Vélain’s var. minor is constant and considerable. On both geographical and conchological grounds, therefore, the latter should be separated from the Tristan shell, and, if really distinct from IZ. maindroni, it appears fully worthy of specific rank. I venture to name it in honour of its original describer. 114 Annals of the South African Musewm. MARINULA MAINDRONI, Vélain. 1877 Marinula maindrom, Vélain, Arch. Zool. Exper. vi. p. 126. Lees iia oat ig! Dee 1901 fe 93 " Kob., Conch. Cab. p. 284. Pl. ORO f. 24. DF. I have been unable to discover the whereabouts of the Type or a single example of this species. Vélain’s diagnosis runs :—“ Shell thin, short and globular, semitransparent, coloured bright brown ; surface smooth and shining. Spire small, very acuminate, almost completely concealed by the last whorl, which is rounded and greatly developed. Aperture large, swollen towards the base ; columellar margin not thickened, marked with much sharper folds than in the preceding species; outer margin thin and sharp, not sinuous. The shell measures 4 x 3+ mm.”’ Animal unknown. Hab. AmstTERDAM IsLANp. ‘“ Dans les vacuoles des laves, assez rare.” I cannot help suspecting that the description might possibly refer to the young of the preceding species, but if such were the case Vélain would hardly have separated them, and it is only just to accept his verdict. MaRINULA PARVA (Swainson). 1855 Cremnobates parva, Swains., Pap. & Proc. Roy. Soc. V.D. Land, nie 4 Pl Vo DEE. 1901 Ophicardelus parvus, Swains., Tate & May, Proc. Linn. Soe. N.S.W. xxvi. p. 420. L. 1910 Cremnobates parva, Swains., Hedley & Suter, Proc. Mal. Soc. ix. pp-.1o1, 152. DUAR. 1913 - ‘3 a Suter, Man. N.Z. Moll. p. 594. DAUR: Shell small, imperforate, prolonged ovate-acuminate, rather thin, brown, corneous. Spire moderately produced, about one-third the length of the aperture ; apex acute. Whorls 4, rather convex, very rapidly increasing, showing no sculpture except faintest irregular growth-marks. Suture fairly deep. Aperture ovate, peristome whitish, outer lip simple, acute, making almost a right angle, and thus appearing perceptibly swollen, at its junction with the paries. Columella white, thickened, with two dental processes ;—the upper sharp and prominent, jutting out at right angles, equidistant between the apex of the aperture and the second tooth, which is shorter and Fae Notes on South African Mollusca. 115 thicker, situate a little less than half-way from the base to the apex of the aperture. Dimensions of a typical specimen from Tasmania, in the British Museum: Long. 9°2; lat. 5:3; apert. 6x3°6; last whorl 8:4 mm. There is no epidermis, the shell appearing to be a little beach- rolled. Hab, Tasmanta. Oyster Cove, near Hobart Town (Milligan) ; near Pirates Bay ; Tasman’s Arch (May). Antipodes Islands (Bollons). Type whi? Hedley and Suter selected this species as the Type of Swainson’s genus Cremmnobates, and published a few notes on its anatomy. This does not appear to differ to any considerable extent from that of M. tristanensis, as diagnosed by Robson in these pages, except that in parva ‘‘a central tooth could not be distinguished,” while in tristanensis a central tooth is shown, but an infinitesimally small one. Without knowledge of the power of magnification employed, it is hardly safe for me to suggest that this feature may possibly exist and have been overlooked in the Tasmanian animal; but Mr. Robson informs me that it took a 4, Homog. imm. objective to enable him to trace it in tristanensvs. The shell of parva is certainly that of a true Marinula, and, on the whole, there seems hardly sufficient cause for regarding Cremno- bates as other than synonymous with King’s genus, although, on the other hand, Hedley and Suter’s description of the anatomy gives us no definite clue for uniting the two genera. MaRINULA FILHOLI, Hutton. 1878 Marinula filholi, Hutton, J. de C. xxvi. p. 42. D. 1880 5 a » Man. N.Z. Moll. p. 34. D. 1913 5; s » suter, Man. N.Z. Moll. p. 591. D. Shell small, prolonged ovate-acuminate, imperforate, pale cor- neous. Spire somewhat produced, nearly as long as aperture, sides convex, apex mamillate. Whorls 34, convex, rapidly in- creasing, finely microscopically striate parallel to the lines of growth. Aperture ovate, outer margin thin, simple; inner bearing three prominent white dental processes ;—a long, thin tooth about equidistant between the outer lip and the next process, a smaller, conical tooth, which is situate a little more than half-way down from the apex of the aperture to the base; below the second, and parallel to it, is a still smaller tooth or fold, half the size of the central one. 10 Oe te RI — wel see _ —————— i ee 116 Annals of the South African Museum. Dimensions of a specimen from Chatham Island: Long. 7:4; lat. 4:5; apert. 4:55x2°8; last whorl 63 mm. unknown. Hab. New Zeauanp. Massacre Bay (Filhol) ; Auckland (Cheese- man); Banks Peninsula; Lyall Bay; Otago, alive on coast above high water (Iredale) ; Chatham Island (in coll. Connolly). MARINULA XANTHOSTOMA, H. & A. Adams. 1821 2? Pedipes ovulus, Fér., Tabl. Syst. iii. p. 109 (or 118). D. 1832 2? Melampus patulus, Lowe, Zool. Journ. v. 19. p. 289. D. 1854 Marinula xanthostoma, Ads., P.Z.S. p. 35. D. a4 ~ cymbaeformis, Récl., Ads., P.Z.S. p. 35 (without characters). 1855 Cremnobates solida, Swainson, Pap. & Proc. Roy. Soc. V.D. Dand eried p44. IPL VI, fo.) DEE 1856 Marinula patula, Lowe (= ovulus? and «xanthostoma), Pfr. Mon. Auric. p. 62. D. g cymbaeformis, Récl., Pfr., Mon Auric. p. 63. aD: re i Cat. Auric. p. 47. D. ‘a Pe patula, Lowe, Pfr., Cat. Auric. p. 46. D. 1867 5 xanthostoma, Ads., Angas, P.Z.S. p. 231. WN. 1901 os patula, Lowe (= pellucida, xanthostoma, and solida}, Tate & May, Proc. Linn. Soc.N.5.W. Oa )acaleh | Ip KS ed ye » (=cymbaeformis), Hedley, Proc. Linn. Soc. N.S. W. xxvi. p. 704. Pl. XXXIV, fig Woe NPIL Before describing the shell, careful attention must be paid to the synonymy. Férussac’s diagnosis of Pedipes ovulus runs: ‘‘ More elongated than afra, smooth and polished, and has not the internal rib on the external border of the aperture. Hab. ?” This might equally refer to a Phytia or Marinula, but it seems doubtful whence Férussac could have obtained an example of the latter genus. The Type of ovulus is not forthcoming, and in its absence the name is obviously worthless. I do not think that patula can stand. It has been impossible to trace the Type or a single authentic example of Lowe’s species, which was unfigured and insufficiently described ; but the following facts throw grave doubt on its being an Australian shell :— Notes on South African Mollusca. iby When describing Marinula pepita in 1832, King mentioned that he had submitted his entire American collection to Sowerby, who had already obtained several of the species from other members of King’s expedition. Now in the same year Lowe described Melampus patulus from a shell received from Sowerby, and gave its measurements as 75 X 7'¢ une., which are identical with those given by King for pepita. Errors of locality were frequent in those early days, and the translation of America into Australia would be a mere nothing, due to some one’s misreading four letters of the name. There is no proof that Pfeiffer was acquainted with originals of patula, and the fact of his connecting it with xanthostoma is easily accounted for by the locality. Lowe’s name of patula would take precedence of either pepita or xanthostoma, but, while the circumstantial evidence is strongly in favour of its application to the former rather than to the latter species, it may be advisable to retain the names which admit of no doubt, and to regard patula as a lost species pending the rediscovery of the missing type. M. xanthostoma was founded by the brothers Adams on a set of well-preserved Queensland shells in the Cuming collection, while C. solida was described a year later by Swainson, from beach-rolled examples collected by Milligan on Flinders Island off the north coast of Tasmania. The Type set of zanthostoma and good Co-types of solida are in the British Museum. M. cymbaeformis was first mentioned in print by H. & A. Adams as having been described by Récluz in the Mag. de Zool. I can find no mention of any such name in Récluz’ writings, and the species appears to have been first characterized by Pfeiffer from a shell in his own collection, purporting to come from the Island of Ticao, Philippines. What this was I cannot say, but the Adams’ set is probably still to be seen in the British Museum. It consists of 6 shells on one tablet labelled both “‘ patula”’ and ‘ cymbeformis”’ and assigned to Ticao (in Cuming’s writing) and also to Tanhay, I. of Negros, in the Philippine Islands. These are simply beach-rolled examples of typical Australian wanthostoma, and failing further proof to the contrary, the Philippine locality may, I think, be accepted as erroneous. Pfeiffer also mentions a var. 3, gracilior, from Australia. This appears to be represented on another British Museum tablet, which 118 Annals of the South African Museum. contains 4 smaller shells collected by Strange at Moreton Bay. They are a slender form of xanthostoma, and show that it is a variable species. The following description is from one of the Adams’ Type set. Shell acuminate-ovate, imperforate, not very solid, apparently somewhat beach-rolled, pale brown. Spire rather short, with nearly straight sides; apex very sharp, slightly produced. Whorls 5, rapidly increasing, the apical smooth and glossy, remainder marked with occasional faint, almost invisible striation in the lines of growth. Suture very shallow. Aperture a little longer than spire, ovate, base rounded, apex acuminate, outer lip sharp, simple; columella thickened, furnished with three white plaits; the upper- most linguiform larger and further, in proportion, from the others ; the second, straight and much shorter, about half-way between the foregoing and the base of the columella; while the third, close below the second, is merely a small projection. Long. 9:8; lat. 6:0; apert. 6:1 x2°5; last whorl 8-6 mm. There is an appearance of a faint interior rib, parallel to and just inside the outer lp, but this is probably an accident, due to the thinness of the edge of the lip owing to the shell not being quite mature, rather than a characteristic feature. Animal unknown. Hab. East anp SoutH Coasts oF AUSTRALIA; TASMANIA. Type in British Museum. The typical form of xanthostoma is, in itself, on the extreme borderland of Marinula, having flatter whorls and a more acuminate spire than any of the preceding species; but in almost any large series intermediates may be found, which link it insolubly to the more normal form described by Swainson under the name of solida. The latter differs markedly from the Adams’ Types in its more elongate and convex spire and rather more pronounced upper tooth. It may be regarded, on the whole, as representing the southern form of xanthostoma, which appears to undergo a gradual transition in shape between its extreme limits of geographical distribution ; but as neither form is constant, even among specimens from the same locality, and as solida is by no means an appropriate name for a form which is in reality less solid than the Type, it appears in- advisable to perpetuate it. So far as I have been enabled to examine specimens from different localities, the typical form, as well as Pfeiffer’s slender var. 3, hail from Moreton Bay, Queensland: the extreme form of Swainson’s so- called solcda is in the great majority in sets from Flinders Island Sa a ee Se Notes on South African Mollusca. 119 (Milligan) ; Portland, Victoria (Cox); North coast of Tasmania (Brazier) and Cossacks, West Australia (Woodward) ; while examples almost exactly intermediate between the type and the variety have been found in New South Wales at Freemantle (Kemp) and Coogee Bay (Brazier). The species is further recorded from King Island (Tate & May), and Shoalhaven and Port Jackson (Angas); but I cannot tell which forms are to be referred to these localities. There is also a typical set in the British Museum labelled ‘‘ New Zealand” (Hancock, 1856), but, although the authority is said to be reliable, it is very doubtful indeed whether living specimens have ever reached the Dominion. It is earnestly to be hoped that the whole question of the distribution and varieties of M. xanthostoma will be taken up by some of the great Australian conchologists, and the anatomy examined with a view to the determination of its true generic position. Distribution of Marimula. It would hence appear that Marinula is a truly Antarctic genus, originating in the Australasian region, with a circumpolar range extending eastward through New Zealand to Chiloe, and westward through St. Paul’s and Amsterdam Islands and South Africa to Gough Island and Tristan d’Acunha. In the sole case where it extends northward, e.g. M. xanthostoma, the shell tends to lose its distinctive shape and to merge into Phytia ; while in Africa and America it is confined to the extreme south of the continent and retains its normal form. Its habit seems to be entirely marine, and it is questionable whether the genus would not be better included among marine, rather than non-marine Mollusca. It now only remains to deal with the mis-named Cumingian shells which have been the cause of so much confusion. I have already mentioned that these occupy three tablets in the British Museum ; one contains three larger shells from Tumaco Island, Colombia, and each of the others two smaller specimens, from ‘Chile”’ and an unknown locality respectively, which appear to be inseparable from the Tumaco examples. We know that the Tumaco shells constitute the Type set of Auricula recluziana, Petit; but Pfeiffer considered them to be identical with Conovulus triplicatus, Anton, which was described from the River Maule, Chile, in 1839. This appears to be correct, age en ee et ——— eee 120 Annals of the South African Museum. in which case Anton’s name has priority, and the synonymy is as under :— PHYTIA TRIPLICATA (Anton). 1839 Conovulus triplicatus, Anton, Verzeichniss, p. 48. D. 1842 Auricula recluziana, Petit, P.Z.S. p. 201. OD. 1844 - pepita, “ King,” Kiist., Conch. Cab. p. 35. Pl. V (1843), fod DE: 1855 Marinula ,, a H. & A. Adams, Gen. rec. Moll. 11. PI GexeRe et Da. Er. 1856 - * mn Pires MoneAurie sp: 09. =D. 1857 . ne ae » Cat..Auric. p. 44. D. Hab. Cutur. River Maule (triplicata, Anton). Cotomsia. ‘Tumaco Island (recluziana, Petit). An altogether coarser form than P. acuta (d’Orb.) described from Guayaquil in 1835, which much resembles this species. The Types of recluziana and acuta are in the British Museum ; that of tréplicata was in coll. Anton. III.—A MonoaraprH oF THE Dorcasiinae. Until recent years, comparatively few shells and only three or four live specimens of the desert snails of South Africa have been available for scientific examination. Inferences regarding the species were therefore often based on somewhat insufficient material, while, although the genera Dorcasia and Trigonephrus were established and the anatomy of D. alexandri and T. globulus published, considerable doubt existed as to which genus some of the allied species belonged. Since 1908, however, thanks to the researches of Dr. A. W. Rogers and other able explorers, several comparatively large series of shells have been gathered, including a most gratifying number of live specimens. The result is that, through the kindness of Drs. Péringuey and Thiele, Miss Wilman, and Messrs. Gude, Natermann, and Ponsonby, I have been privileged to examine the most extensive assemblage of these shells that has ever been got together; in addition to which Mr. Hugh Watson, of Cambridge, has dissected the spirit specimens, and his reports, embodied in the following pages, are of the utmost OO Notes on South African Mollusca. 121 value in confirming, or correcting, deductions drawn from concho- logical features alone. Even now, however, more than one of the species, of which animals have been obtained, is represented by only a single spirit specimen, and others are not in the best con- dition, so that the results of their examination are not quite as satisfactory as would have been the case had a more complete representation been available. It will be seen that Watson states that Dorcasia and Trigonephrus are nearly allied to each other, but differ considerably from other genera whose anatomy is known. They are rightly placed by Pilsbry in the Family Acavidae, and a new Sub-family, Dorcaswinae, may conveniently be founded to contain them. The Dorcasiinae are distributed in all shapes and sizes along the south-western seaboard of Africa, from Algoa Bay to Damaraland, and in some instances also to a considerable distance inland. I must confess that when I first undertook this paper, it was with the idea that many existing species could be shown to be hardly varietally distinct, and that, owing to the presence of connecting links, their names would be scarcely worth preserving. Extended study, however, tends to prove that most of the named forms exhibit quite sufficient and constant difference to entitle them to specific rank; while, in other cases, shells from the same locality constitute a local race of some species, consistently well defined, and so widely divergent from the Type that a name is advisable in order to prevent confusion. From an anatomical standpoint, Watson places the genus T'rigo- nephrus first and Dorcasia after it, and as his arrangement of the species, based on their anatomy, is perhaps the most convenient for comparison of the shells, I follow it in the ensuing pages. The history of the Acavidae and all subsequent references to anatomical details are in the exact words of Mr. Watson, who has also furnished the figures of the animals, and my warmest thanks are due to him for his kind co-operation. I may here mention that I have made no attempt to make my measurements, taken from the actual shells, agree with the figures, all of which are the natural size. Hardly any two persons will be found to measure a globular helicoid in exactly the same fashion, and a very slight accidental difference in the angle from which the shell is viewed or measured will account for a propor- tionate difference in the dimensions recorded. As full references to every species were given in my ‘“ Revised Reference List of South African Non-marine Mollusca,” published 122 Annals of the South African Musewm. two years ago in these ANNALS, they are omitted in the present instance. Faminry ACAVIDAH, Pilsbry, 1900. Proc. Acad. Nat. Sci. Phila., p. 564. It is generally believed that millions of years ago a great southern continent stretched westward from South Africa as far as the Andes, and eastward through Madagascar and Southern India to the Australian region. The climate of this continent was very cold towards the end of the Carboniferous period; but later it became warmer, and the snails that dwelt there flourished exceedingly, several new families being evolved. And one of these families seems to have been the Acavidae. Now it is not difficult to form some idea as to what the first members of the Acavidae were like ; for it is reasonable to suppose that they would possess any primitive characters that are still retained by some of their descendants, as well as such other features as are found in all the modern members of the family, or at least in all the more archaic genera. We may suppose, therefore, that the eggs of these snails would be larger than usual, and that the animals themselves would be above the average size. The shell would have a pointed spire and laterally compressed whorls; that is to say, both the shell and its aperture would be higher than broad. This is the form of shell found in the most primitive Euthyneura, such as the Actaeonidae, the Chilimdae, and the Awriculidae, as well as in many of the more ancient families of the Stylommatophora, and it is still retained by several members of the Acavidae. There would be the usual two pairs of tentacles, and the labial lobes would be well developed. The foot would have neither a definite peripodial groove nor longitudinal grooves on the sole. No caudal mucous pore would be developed. The lung would be rather short, and the pulmonary veins would branch over the whole of its roof, the first branch of the pericardial vein being nearly as large as the principal pulmonary vein. The excretory system would be unusually simple, the kidney being broad and probably without even a primary ureter. The nervous system would be of the type found in most of the more primitive snails and slugs, there being eleven separate ganglia. The cerebral ganglia would probably be rather near together, but the connectives uniting them with the remaining ganglia would be long. The jaw would be without ribs, but would probably have faint vertical striae and a Notes on South African Mollusca. 123 slight median projection. The radula would have nearly straight rows of “quadrate’’ teeth of the ordinary Helicid type, and the alimentary canal would pursue the usual course, though the differ- ence in size between the two divisions of the liver would be less marked than is the case in most snails at the present day. The reproductive organs would be of a primitive type. A well-developed vesicula seminalis would be present, and there would be a swelling near the junction of the free oviduct and the duct of the sperma- theca—a swelling which may have grown out into an appen- diculum ; but the spermathecal duct would have no diverticulum, and there would be no other accessory outgrowths of any kind. Though the penis and vagina would open together, a common vestibule would not have been developed. A considerable portion of the vas deferens, instead of lying freely in the body-cavity, would be still partially embedded in the invaginated portion of the skin which forms the wall of the penis. The retractor of the right upper tentacle would pass between the penis and the vagina. Such we may suppose to have been the main characters of the earliest members of the Acavidae. It will be seen that they were more primitive than many modern snails, but probably at the beginning of the Mesozoic era they were both larger and more highly organized than most of the other families then living. And the family increased and multiplied, and spread far and wide through that ancient southern continent. Moreover, both their multiplication and their spreading led to the further evolution of these snails. As the individuals increased in number under the favourable conditions, they would become more crowded, and the competition between them would become more severe. This would lead to the evolution of higher types: unfavourable varieties would be elimi- nated; progressive mutations would alone survive. Thus the eggs grew larger and larger, or, in an extreme case, the animal became viviparous, the embryo reaching a relatively enormous size. The adult also became larger and broader, the whorls very much wider, and the apex more and more obtuse, until the form of the shell became globular, and finally depressed. The lip also became more expanded. A closed primary ureter was developed, and eventually the beginning of a secondary ureter.* ‘The fine striae on the jaw either developed into ribs, or, more frequently, disappeared altogether. Similarly, the comparatively useless small secondary cusps on the teeth of the radula were generally eliminated, the main cusps becoming larger and broader at their expense. The reproductive * See Randles, Proc. Mal. Soc., 1900, iv. Pl. IX, fig. 1. 124 Annals of the South African Musewm. organs also became simplified by the disappearance of any appen- diculum or swelling at the junction of the free oviduct and the duct of the spermatheca; and, at the same time, the vas deferens gradually emerged from the wall of the penis and came to lie freely in the body-eavity. Now evolution of this kind would proceed most rapidly in the central area where the family first arose, which was probably nearer Ceylon than South Africa; for, unless the local conditions were unfavourable, the individuals would be most densely crowded towards the centre, and there the competition would consequently be most severe. Towards the periphery of distribution, on the other hand, the individuals would be more sparsely scattered, and the primitive forms would therefore survive, as Taylor has main- tained in dealing with other groups. There is, however, another kind of evolution, which proceeds more rapidly towards the limits of the distribution of a group than in the centre; for as the animals spread in different directions they encounter new conditions to which they have to adapt themselves. This may lead to a profound modification of some of the organs ; and if there are great differences in the environment, the various peri- pheral genera may differ more from the ancestral form than do those which still inhabit the central area. But the conditions seem to have been fairly uniform throughout that ancient southern continent; and although the peripheral members of the Acavidae developed differences as they spread in various directions, they remained more primitive than those in the centre of evolution. Until about the middle of the Mesozoic era the distribution of the family was probably continuous. And then there came the sea. First in one region and then in another, partly by denudation but chiefly by subsidence, that ancient continent gradually disappeared beneath the encroaching waves. The region extending from the south of India to Madagascar was cut off from Australia on the one hand and from South Africa on the other, and, later, Africa was separated from Brazil. Thus the old home of the Acavidae was divided into four large islands. Henceforth each of the divisions of the family would develop independently, their separation accen- tuating their differences, until the Acavidae were divided into four subfamilies, each inhabiting a different area. The subfamily inhabiting the most westerly region, which included Brazil and the adjacent parts of South America, would be on the whole the most primitive, as it was furthest from the centre of evolution. So far as is known, none of the Heliciform members Eee ee Notes on South African Mollusca. 125 of the family ever reached South America. Only the Bulimiform genera, Strophocheilus (including Borus) and Gonyostomus, are found in that region, and Plate has emphasized the very primitive character of these snails.* The researches of Semper,} von Ihering}, Plate,s and Pilsbry || have shown that, besides retaining an elongate shell, these South American genera have a very simple kidney with no ureter, a radula in which the ectocones are usually still present on the marginal teeth, and a free oviduct which is swollen or pro- vided with an appendiculum at its junction with the duct of the spermatheca. But although these snails retain many primitive characters, they have undergone a considerable development under the favourable climatic conditions of Brazil. In some species the jaw has developed stout ribs. Many of the forms attain a great size, and have very big eggs, A large number of species have arisen, and they have spread over the greater part of South America, though they are still most abundant to the east of the Andes. To this subfamily Pilsbry has given the name of Strophocheilinae. The subfamily which inhabits Australia and some of the neigh- bouring islands contains far fewer species; but so diverse are these that they are assigned to at least twice as many genera as the South American forms. Probably the group was once much commoner in the Australian region than it is at the present day. The south of Australia is nearly as far from the centre of evolution as is Brazil, and, with the exception of Anoglypta launces- tonensis, all the species from this region are more or less Bulimiform. Further north, however, we find the greatly depressed genus Pedinogyra in Queensland and New South Wales; while in the most northerly species of Hedleyella, instead of the spire having been flattened, the entire shell has become reduced, so that it can no longer contain the animal. The teeth of the radula have almost entirely lost their secondary cusps in this group,‘ and a primary * Sitz.-Ber. Ges. naturf. Freunde, Berlin, 1896, p. 149. + Reis. im Archip. Philippin., 1874, iii. p. 150. Pl. XIV, f. 10, Pl. XVI, f. 25, PI SexVillestoel Pl By (S94) sto: t Zeitsch. f. Wissensch. Zoologie, 1884, xli. p. 270. Pl. XVII, f. 6, and Bull. Scientifique, 1891, ii. p. 213. Pl. V, f. 11. § Op. cit. || Man. Conch., 1895, x. p. 1, and 1902, xiv. pp. iv—viii. PJ. XLIX, f. 8, IPM, IGM, vy UG, a, BIL, 1, PA P7 Teale OMMIES ai, Bis), Ie, IGIAY, ti, Sf, JBL INO, se, Ng) 65-68. ‘| The ectocones, however, can still be distinguished in the marginal teeth of some specimens of Anoglypta launcestonensis. 126 Annals of the South African Museum. ureter seems to have been developed, though further information is needed about the pallial organs of these snails. On the other hand, the fine striae on the jaw are retained in Pedinogyra and Anoglypta, and the vas deferens is still partially embedded in the wall of the penis in the latter genus. Moreover, the lip of the shell is unusually simple in this subfamily, the appendiculum is well developed, and the egg, though large, does not attain the enormous size found in some of the more highly specialized members of Acwvidae. The penial retractor arises from the columellar muscle, or from far back on the floor of the lung, instead of near the front. Hedley, to whom we owe so much of our knowledge of these snails, was the first to demonstrate that the Australian genera Hedleyella,* Caryodes, Anoglypta, and Pedinogyra, were related to one another, notwithstanding their striking external differences. | Semper had already remarked on the resemblance of Hedleyella to Acavus and Helicophanta from Ceylon and Madagascar,{ but after comparing the anatomy of Hedleyella and Caryodes with that of Acavus, Helicophanta, and Ampelita, Pilsbry states that the relation- ship between the Australian and Indo-Madecassine genera is by no means intimate §. The Australian region has long been cut off from the other three areas in which the Acavidae are found, and contains a separate branch of the family, less primitive in some respects than the Strophocheilinae, but not so highly organized as Acavus and its allies. This subfamily may be named the Caryodinae. The region extending from the south of Madagascar to the south- western part of the peninsula of India remained a single large island, or a closely connected chain of islands, long after it was separated from the Australian region on the east and South Africa on the west; and it was probably not until Tertiary times that a series of faults, accompanied by subsidence, sent the greater part of that land beneath the Indian Ocean. The genera Acavus found in Ceylon, Stylodonta in the Seychelles, and Helicophanta and Ampelita in Madagascar, must therefore be regarded as_ the surviving remnants of a group which formerly also inhabited the intervening areas. This group is that which lies nearest to the original centre of evolution of the family, and accordingly it includes the most highly * Hedleyella, Iredale, 1914 (Proc. Mal. Soc., xi. p. 174) = Panda, Albers, 1860, nec van Heyden, 1826. + Rec. Australian Mus., 1892, ii. p. 29. t Reis. im Archip. Philippin., 1873, iii. p. 104. § Man. Conch., 1895, ix. p. 164. © Notes on South African Mollusca. 127 organized members of the Acavidae. As in the Caryodinae, the most primitive forms are found in the south of the region, that is to say among the Madagascan species. Some of these have Bulimiform shells, while in the genus Ampelita the egg is only moderately large ; but in all the genera of this subfamily, excepting Ampelita, the embryo attains a relatively enormous size before it is hatched or born (for Stylodonta is viviparous), and most of the Madagascan species, and all those from Ceylon and the Seychelles, have Helici- form or depressed shells. It is true that in some of the species of Acavus the spire is rather high, but this is almost certainly a case of reversion due to their having acquired arboreal habits. There is always a tendency among snails that live on trees or cliffs to become elongated, and the broad aperture of the shell in Acavus suggests that it was originally depressed. Moreover, in the two species of Acavus that are not arboreal, the spire is nearly flat. The shell is often highly coloured, and the lip, unlike that in the Caryodinae, is usually broadly expanded. Excepting in Stylodonta, all the teeth of the radula are without secondary cusps, and the jaw is without striae. An ureter is present. The reproductive system differs from that of the Caryodinae in being without an appendiculum; but in the Malagassy genus . West, Newman del.et lith PATE ADE. Fig. 1.—Dorcasia coagulum (vy. Mts.); from a specimen in the South African Museum. 2.—D. rogersi, nov.; from the Type in the South African Museum. 3.—D. rogersi, forma maxima; from a specimen in the South African Museum. 4.—D. cernua (vy. Mts.); from the Type in Berlin Museum. 5.—D. alexandri (Gray); from the Type in British Museum. 6.—D. alexandri, var. minor, Bttg.; from a specimen in the South African Museum. 7.—D. alexandri, var. rotundata, Mouss.; from a specimen in the South African Museum. 8.—D. alexandri, var. perspectiva, nov. ; from the Type in Kimberley Museum. XII. Acne S Ate Mais Vol. a 4 ri ee ee a Se > 4 ; ; Trachycystis (Endodontide) 187, 188 namaquensis (Trigonephrus) . 154 | Tyjo onephrus (Acavide) . 121, 140 triplicata (Phytia) . : 106, 120 O tristanensis (Marinula) .. . 108 trivia (D. alexandri, var.) . elie octona (Subulina) : : . 189 | tropica (Isidora) : : _ Io orcula (Pupisoma) . 186, 187 | truncatula (Limnea) . = L889 ovampoénsis (Ceecilioides) . . 189 | tuberculata (Tiara) . ; > 89 ovulus (Pedipes) . 5 : . 116 | Tulbaghinia (Acavide) . 5 172) umlaasianus (Limneeus) — . . 189 P : V parva (Marinula) : , . 114 patulus (Melampus) . . 107,116 | Vallonia (Helicide) . : a lS6 pellucida (Auricula) . ‘ . 106 | variegatus (Limax) . : . 184 pepita (Marinula) . 102,107 | velaini (Marinula) . : Slits perspectiva (D. alexandri, var.). 172 | vespertina (Phytia) . : = 105 Phytia (Auriculide) . . 103,120 | Vitrea (Zonitide) : ; 5 ae pisana (Helix) . : . 187 | vulcani (Auricula) . . 105 Planorbis (Limnide) : » ge Polita (Zonitide) : : . 182 x ponsonbyi (Doreasia) . 159 porphyrostoma (Trigonephrus) . 152 | xanthostoma (Marinula) . 5 Ke procerus (‘T. namaquensis, var.). 155 pulchella (Vallonia) . : 5 dbs{s) Z, punctata (Ovatella) . 5 . 105 Pupisoma (Helicide) . : . 186 | Zonitoides (Zonitide) . : 5S Se (198 ) 9.— Two New Species of Marginella from South Africa. —By Lewis J. SHACKLEFORD. MARGINELLA TOMLINI Sp. Nov. Shell. Four whorls rather obtusely conical, smooth and very polished, pale straw-coloured, with no markings except two rows of irregularly J. S. Gladstone, photo. J. 8. Gladstone, photo. Fie. 1.—Marginella tomlini. x 2. Fie. 2.—Marginella tomlini. x 2. oblong black spots on the body-whorl, the upper of which is continued round the upper whorls, the spots becoming rounder and smaller as they approach the apical whorls, which are plain and glassy. The lower begins near the margin and ascends spirally on to the upper- most plait. There are ten spots in this row, three of which are on the plait itself. Spire raised only about 3 mm. above the summit of the outer lip. Suture not impressed. Spire moderately convex. Aperture long.,15 mm.; lat.max.3 mm. Margin moderately thickened. Columella with four well-defined plaits, the upper two being nearly straight and rather far apart, the lower two oblique. The outer lip 194 Annals of the South African Musewm. is white and smooth within and considerably curved. The plicae and margin are also white. Long. 18mm. ; lat. 9mm. Hab. Cape St. Blaize (S. Africa) N. by E. 3 E., distant 68 miles— 105 fathoms. s.s. ‘‘ Pieter Faure.” Type unique in the South African Museum. MARGINELLA TAYLORI Sp. NOV. Shell. Subtriangular ovate, smooth, shining, colour pale cream with a faint yellow band round the base; spire blunt, the apex glassy ; suture slightly impressed; whorls 4, the last whorl rising distinctly towards the aperture. J.S. Gladstone, photo. JS. Gladstone, photo, Fie. 3.—Marginella taylori. x 4. Fig. 4.—Marginella taylori. x 4. Columnella covered with a thin callus, with seven plaits, somewhat oblique, the three uppermost almost obsolete, the others well defined and rather far apart, the penultimate the largest, the last very oblique. Aperture narrow for two-fifths of the upper part, thence widening as far as the base. Length of aperture 4 mm.; greatest width 5mm. Labrum moderately curved, thickened, finely irate within, minutely denticulated along its entire length and inflected backwards in the lower part. Long. 5mm. ; diam. max. 2 mm. Hab. Cape St. Blaize (S. Africa) N. by E. } E., distant 68 miles— 105 fathoms. s.s. “ Pieter Faure.” Two specimens, one broken, in the South African Museum. Named after Mr. J. Kidson Taylor (Buxton, Eng.), who has made Marginella a special study. The apparent marking shown on the figures, especially the back view, is due to the partial erosion of the shell. 7 | | | (195) 10.—A Revision of the Lizards of the Genus Nucras, Gray.—By G. A. Boutencer, LL.D., D.Sc., F.R.S. (Published by permission of the Trustees of the British Museum. ) (With Plates VI-VIL.) INTRODUCTION. When, some twenty-five years ago, the late Dr. R. Klebs submitted to me the oldest known Lacertid (Oligocene) with the lepidosis preserved through imbedding in amber, a careful comparison with recent lizards led me to refer it to the genus Nuweras, although the essential generic character of the position of the nostril could not be ascertained, my conclusion being based on an examination of the digits and of the scaling of the gular and pectoral regions, which agreed better with Nucras tessellata than with any other lizard with which I was acquainted.* Since then I have made a more thorough study of the Lacertidae from the point of view of their probable evolution, and independent, correlative reasons have confirmed my provisional identification so far that, quite apart from any palae- ontological consideration, I am now inclined to regard Nucras as, on the whole, the most primitive genus of the Lacertidae. At the time I examined the lizard in amber, the representatives of the genus were believed to be confined to Africa no further north than the Zambesi Basin, and my suggested identification may there- fore have seemed somewhat risky from the standpoint of zoogeo- graphy. We must, however, bear in mind that, in Tertiary times, the general character of the reptile fauna of the northern parts of * Of. R. Klebs, Schrift. Phys.-oek. Ges. Kénigsberg, li., 1910, p. 227. As this lizard has not received a name I propose to designate it as Nucras succineus. 15 196 Annals of the South African Museum. what is now the Palaearctic Region differed strongly from that of the present day. Iguanidae, now confined to the New World, Fiji, and Madagascar, occurred in the Miocene of Europe, and the Pleurodiran Chelonians, at the present time found only in Tropical and South Africa, Madagascar, and South America, were represented in the Eocene as far north as England. Within the last fifteen years the range of Nucras has been ascertained to extend further to the north in Africa (Lake Victoria), and, in accordance with the view of the probable origin of these lizards, the northernmost species (N. emini) has every claim to be considered, from the morphological standpoint, as the most primitive of the genus. I therefore believe that Nucras had a northern origin, an opinion further supported by the fact that the Lacertidae, like the Agamidae, being absent from Madagascar, must have extended their range towards the south only after the connection of Africa with that island had been severed, whilst the presence of Iguanidae, Gerrhonotidae, and Chamaeleontidae may be explained by these having reached Madagascar from Africa at a period previous to the southern extension of the Lacertidae and Agamidae. The reasons for regarding the genus Nucras as the most primitive of the Lacertidae are the same as set forth in my recent paper on the derivation of the species of Lacerta,* in which L. agilis is held to be the surviving representative of the ancestor of most, if not all, of the species of the genus Lacerta with which we are at present acquainted. Of the ten characters, or sets of characters, there mentioned nine are in accordance with this view, the only two (7, 9) in which Nucras is not so primitive being the reduction of the dorsal lepidosis to smooth granules and the long tail, in which all the species at present known agree. Otherwise we find (1) constant presence of teeth on the palate; (2) a non-depressed or feebly depressed skull of moderate ossification (no supraorbital fontanelle, no dermal ossifications in the temporal region), although less primitive than that of L. agilis, owing to the narrower internarial space (comparable to ZL. vivipara in N. delalandii, to L. muralis in N. tessellata); (3) presence, in some forms at least, of the foramen parietale; (4) nostril between two or three nasals, the first upper labial being well separated from it, and absence, in some species, of small scales between the supraoculars and the superciliaries ; (5) lower eyelid without transparent disc; (6) no * Tr, Zool. Soc. xxi., 1916, p. 1. + Unless it be true that the tail of N. boulengeri is only 14 to 11 times the length of head and body, as stated in the description ; but it is not improbable that the fact of the organ being in a regenerated condition has been overlooked. A Revision of the Lizards of the Genus Nucras, Gray. 197 denticulation in front of the ear-opening; (8) cylindrical or feebly compressed digits with smooth lamellae inferiorly ; (10) the ideal type of primitive markings in some forms, no vivid colours on the head and body. The main principles of the evolution of markings, as held by me, are well supported by a study of the genus Nucras, which embraces striated, ocellated, and barred forms. The most primitive pattern, with 11 light longitudinal streaks, at least anteriorly, occurs in N. intertexta, var. holubi, and in N. tessellata (taeniolata, Smith). In the latter species the markings may vanish towards the posterior part of the body, and the streaks on the sides break up into spots and, further, rearrange themselves into cross-bars, as happens also in N. intertexta, var. holubi. 'The dorsal striation may disappear and lead to ocellated forms (N. intertexta, typica). The most pronounced ocellar pattern, accompanied by the loss of the longitudinal streaks, is exemplified by N. delalandii, in which, further, the ocelli may disappear, to be replaced by black cross-bars. As a rule the white longitudinal streaks are more numerous on the nape than on the body, but I find one individual exception in a typical N. tessellata, which shows three dorsal streaks on the nape and four on the body. There are two important points in which the striation in Nueras differs from that in Lacerta: (a) The outer light dorsal streak, instead of starting from the superciliary edge, originates on the border of the frontal shield and then follows the supraorbital border and the parietal shield ; (b) the vertebral streak, instead of ending on the base of the tail, may be continued a considerable distance along that organ ; evidently a primitive condition in accordance with Himer’s law. The distinction of species in this genus has always been a matter of difficulty, all the greater for the small number of specimens which most authors have had at their disposal. I have been so fortunate as to be able to compare large series, preserved partly in the South African Museum, partly in the British Museum. Not long ago* I attempted a rearrangement into varieties from the South African material entrusted to me by my friend Dr. L. Péringuey, but I have since made a more profound study of the subject, resulting in the present monograph, in which detailed descriptions are given of the species and varieties. So much doubt still exists as to the value of certain characters of lepidosis and coloration, that the minute analysis of individual variations, as presented in this paper, will prove useful to * Ann. 8. Afr. Mus. v., 1910, p. 473. 198 Annals of the South African Museum. those who might feel inclined to pursue further the investigation of this difficult group of lizards. NUCRAS. Nucras, Gray, Ann. N. H.,i, 1838, p. 280; Lataste, Ann. Mus. Genova (2), ii, 1885, p. 124; Bouleng., Cat. Liz., ili, p. 52 (1887). Lacerta, part., Dum. & Bibr., Erp. Gén., v, p. 174 (1839); Bedriaga, Abh. Senck. Ges., xiv, 1886, p. 24. Nucras, part., Gray, Cat. Liz., p. 33 (1845). Zootoca, part., Gray, op. cit. p. 27. Bettaia, Bedriaga, t.c., p. 435. Head-shields normal. Nostril well separated from the labials, pierced between two or three nasals. Lower eyelid scaly. Collar well marked. Dorsal scales small; ventral shields feebly imbricate, smooth. Digits cylindrical or very feebly compressed, with smooth lamellae inferiorly. Femoral pores. Tail long, cylindrical. Synopsis of the Species. I. No small scales between the supraoculars and the superciliaries, or one or two exceptionally present ; head not or but little broader than deep ; 16 to 20 lamellar scales under the fourth toe. Head 4 to 43 times in length to vent*; foot as long as or a little longer than head; parietal foramen present ; 40 to 51 scales across middle of body; ventrals in 28 to 382 transverse series ; transversely enlarged plates under the fore-arm 1. N. emini. Head 43 to 5% times in length to vent; foot not longer than head; parietal foramen usually absent ; 34 to 41 scales across middle of body ; ventrals in 32 to 87 transverse series; trans- versely enlarged plates underthe fore-arm absent or small ; : : : ‘ Head 43 to 5 times in length to vent; foot shorter than head; 45 to 538 scales across middle of body ; ventrals in 27 to 34 transverse series 2. N. delalandii. 3. N. boulengeri. II. A series of 2 to 7 small scales between the supraoculars and the super- ciliaries ; 40 to 60 scales across middle of body; transversely enlarged plates under the fore-arm. * The head is measured to the posterior border of the ear-opening, the skull being considerably longer than the pileus, which accounts for Bedriaga’s state- ment that the ear-opening is situated further back than in Lacerta, “ on the side of the neck.” A Revision of the Lizards of the Genus Nucras, Gray. 199 Head 35 to 43 times in length to vent, not or but little broader than deep ; foot not or but slightly longer than head; parietal foramen usually present’; ventrals in 27 to 34 transverse series; 20 to 26 lamellar scales under the fourth toe . 4, N. intertezxta. Head 4 to 42 times in length to vent, considerably broader than deep ; foot considerably longer than head; parietal foramon usually absent ; ventrals in 25 to 33 transverse series; 25 to 31 lamellar scales under the fourth toe ‘ ‘ : . 5. N, tessellata. 1. NucrAs EMINI. Nucras delalandii, Tornier, Zool. Jahrb., Syst., xii, 1900, p. 593. Nucras emini, Bouleng., Ann. and Mag. N. H. (7) xix, 1907, p. 488; Nieden, Mitt. Zool. Mus., Berl., viii, 1913, p. 76. Nucras ukerewensis, Bolkay, Archivam Zool. (Budapest), 1, 1909, p- 13, figs. Nucras tessellata, Sternf. in Schubotz, Wiss. Ergebn. Deutsch. Z.-Afr. Exped. iv, ii, p. 222 (1912). But for the longer tail, proportions much as in L. agilis. Head a little broader than deep, 12 to 12 times as long as broad, its length (to posterior border of ear-opening) 4 to 44 times in length to vent ; snout obtuse; cheeks swollen in the male. Pileus twice as long as broad. Body scarcely depressed. Hind limb reaching the wrist or the axil; foot as long as the head or a little longer; digits feebly compressed. Tail nearly twice as long as head and body. Nasals in contact behind the rostral; frontonasal broader than long ; praefrontals forming a median suture, or frontal narrowly in contact with the frontonasal; frontal about 1% times as long as broad, as long as its distance from the end of the snout; fronto- parietals much shorter than the frontal; parietals 14 times as long as broad ; interparietal 2 to 21 times long as broad ; occipital very small. Four supraoculars, first and fourth small, first im contact with the frontal; 4 or 5 superciliaries, in contact with the supra- oculars. Two superposed postnasals ; anterior loreal not half as long as second; subocular not or but little narrower beneath than above, between the fourth and fifth upper labials ; two large upper temporals, first much longer than the second and in contact with the fourth supraocular ; lower temporal scales rather large, upper smaller and granular ; a large tympanic shield. Five large and vertically elongate scales in the middle of the lower eyelid. 200 Annals of the South African Museum. Parietal foramen and pterygoid teeth present. Gular scales much enlarged towards the collar, 18 to 21 between the symphysis of the chin-shields and the median collar-plate ; no gular fold. Collar serrated, composed of 7 to 9 plates. Seales granular, round or squarish, smooth, enlarged on the lower part of the side and passing gradually into the ventral plates; 42 scales across the middle of the body. Ventral plates in 6 or 8 longitudinal and 28 or 29 transverse series.* ‘Two large praeanal plates, one before the other, and a single semicircle of smaller plates. A series of 7 or 8 transversely enlarged plates on the lower surface of the forearm. Scales on upper surface of tibia smaller than dorsals. 10 to 12 femoral pores on each side. 17 or 18 lamellar scales under the fourth toe. | Upper caudal scales rather broad, slightly oblique, strongly keeled, posterior border very obtusely pointed or rounded. The two specimens examined differ in the coloration. The smaller, the type of N. emini, appears to be a half-grown male. Pale reddish brown above, with very small black dots and a faint trace of a light vertebral streak; a black lateral band from the eye to the root of the tail, edged with a light streak above and beneath and bearing a series of white spots with a tendency to run together into a streak; limbs with small dark brown spots or vermiculations ; tail reddish ; lower parts white. The larger specimen, an adult male of stouter habitus, agrees better with Bolkay’s figure of N. wkerewensis. Dark brown above, with an interrupted white vertebral streak ; five series of small, white, black- edged spots on each side, the upper and lower corresponding to the light streaks of the preceding specimen ; hind limbs with small white, black-edged ocellar spots ; lower parts white. Measurements (in millimetres) : From end of snout to vent. : 5 . 68 46 53 - fore limb. ; oi Oe uli: Length of head. : : ; ; Se Lon gel Width of head , , : : LONE. Depth of head ‘ : 3 A : : I 6 Fore imb A . : : ; i 20 als Hind limb ; ? F . ~ \ 26) 224 Hooti) : ‘ ; : : A sean aloe melee Tail ; : : : : ; . . 184 — * 32 in the type of NV. ukerewensis, according to Bolkay. A Revision of the Lizards of the Genus Nucras, Gray. 201 Particulars of Specimens Examined. I 25. SAC: 6. Ue 6, Loika . 63 42 297 TO 2 aA lO eu7 Her. 3, type. . 46 42 23) SoSH iets 1. Length from end of snout to vent (in millimetres). 2. Number of scales across middle of body. 38. Transverse series of ventral plates. 4. Plates in collar. 5. Gular scales in a straight line between symphysis of chin-shields and median collar-plate. 6. Femoral pores (right and left if differing). 7. Lamellar scales under fourth toe. The type is from the southern shore of Lake Victoria, from the collection of Emin Pasha. A second specimen, from Loika, British East Africa, from the collection of Mr. W. P. Lowe, is preserved in the British Museum. A third, from Shirati, east shore of Lake Victoria, has been described and figured by Bolkay as N. uwkerewensis. The lizard from Katoma, German East Africa, shortly noticed by Tornier under the name of N. delalandii, is referred to this species by Nieden. It is stated to have 40 scales across the body, 30 trans- verse series of ventral plates, 14 femoral pores, and the foot a little longer than the head. Specimens from Ukamba and Taita, also referred to this species by Nieden, have 42 to 51 scales across the body, 26 to 31 transverse series of ventrals, and 12 to 14 femoral pores on each side. Sternfeld’s “ N. tessellata” is from the Kili- manjaro district. 2. NucRAS DELALANDII. Lacerta lalandii, M.-Edw., Ann. Se. Nat., xvi, 1829, pp. 70, 84, pl. v, fig. 6. Hremias (Nucras) lalandii, Gray, Ann. N. H., i, 1838, p. 280. Lacerta delalandii, vars. a, c, Dum. & Bibr. Erp. Gén., v, p. 241, pl. xlvii (1839). Nucras lalandii, Gray, Cat. Liz., p. 33 (1845). Bettaia delalandii, Bedriaga, Abh. Senck. Ges., xiv, 1886, p. 435, pl. —, figs. 1, 5, 35. Nucras delalandii, Bouleng., Cat. Liz., ii, p. 53 (1887); Werner, Jahrb. Nat. Ver. Magdeb., 1896-98, p. 141; Bouleng., Ann, 8S. Afr. Mus., v, 1910, p. 475. Nucras delalandii, var. bedriagai, Werner, l.c. Head small, not or but slightly broader than deep, 12 to 13 times as long as broad, its length 44 to 51 times in length to vent in males, 5 to 5} times in females ; snout very obtuse. Pileus 1} to 2} times as long as broad. Body much elongate, cylindrical. Limbs short, just 202 Annals of the South African Museum. meeting (males) or more or less widely separated when adpressed in the adult, overlapping in the young; foot $ to once length of head ; digits short, not or but scarcely compressed. Tail thick, swollen for some distance behind the base in males, 15 to 27 times as long as head and body. Nasals in contact behind the rostral; frontonasal broader than long, as broad as or broader than the internarial space ; praefrontals forming a median suture which may be very short; frontal as long as its distance from the rostral or the end of the snout, 12 to 1% times as long as broad, usually narrower, behind, than the supraoculars ; frontoparietals much shorter than the frontal, usually shorter than their distance from the posterior extremity of the pileus; parietals 14 to 1} times as long as broad, outer border sometimes feebly emarginate for the accommodation of the upper temporal; interparietal narrow, 2 to 4 times as long as broad, sometimes divided into two ; occipital very small, sometimes pushed back behind the pileus, or separated from the interparietal by the parietals forming a short median suture. — Four supraoculars, first and fourth small, first usually * in contact with the frontal; 5 to 7 superciliaries, rarely 4, in contact with the supraoculars, or one or two granular scales intervening between them Two superposed postnasals, rarely onet ; anterior loreal 4 to } length of second ; 4 upper labials { anterior to the subocular, which is usually a little narrower beneath than above; an elongate upper temporal, some- times very broad, sometimes narrow, often divided into two or three, nearly always in contact with the fourth supraocular§ ; temple covered with small hexagonal or granular scales, which are not or but little larger than the dorsals ; a round or oval tympanic shield, rarely absent. 5 or 6 vertically enlarged scales in the middle of the lower eyelid. Parietal foramen usually absent. Pterygoid teeth present. Gular scales small, juxtaposed, increasing in size and imbricate towards the collar, 23 to 30 in a straight longitudinal series; no cular fold. Collar feebly serrated, composed of 7 to 14 scales (usually 8 to 10). Scales on body round or oval, juxtaposed, smooth, 34 to 41 across the middle of the body, 2 or 3 on the side corresponding to a ventral plate. Ventral plates in 8 longitudinal series, those of the second * In 17 specimens out of 24. + On both sides in a male from Van Reenen, on the right side in a male from Damaraland. + Five on both sides in a male from Peri Bush, on the right side in a female from Lessouto. § Exceptions in a male from Peri Bush and in another from Barberton. A Revision of the Lizards of the Genus Nucras, Gray. 208 series from the median line the broadest, the outer small, and 32 to37 transverse series. Praeanal region covered with irregular small plates or with a large plate, which may be longer than broad or broader than long, bordered by one or two semicircles of smaller plates ; sometimes with two subequal broad plates, one in front of the other. No transversely enlarged plates on the lower surface of the fore limb, or a series of a few feebly enlarged plates. Scales on upper surface of tibia smaller than dorsals. 10 to 15 femoral pores on each side. 16 to 20 lamellar scales under the fourth toe. Caudal scales forming whorls of nearly equal length, upper moder- ately broad, the median pair broader, strongly keeled, obtusely pointed behind, with distinct sensory pits. Young pale brown above, with 8 or 10 longitudinal series of white black-edged ocellar spots, with traces of a white vertebral streak on the nape and anterior part of the back (Pl. VI, fig. 1); sides of head and sometimes of neck with black and white vertical bars ; hind limbs and tail orange or reddish. The ocelli usually persist in the adult (Pl. VI, fig. 2) which is greyish or reddish brown above, and the black rings may expand into large spots with a tendency to run together into cross-bars (Pl. VI, fig. 3), in which case the white eyes may become much reduced or disappear, leaving nothing but more or less regular black cross-bars (Pl. VI, figs. 4,5); head with black spots above, with black spots or vertical bars on the sides, the latter sometimes extending to the throat ; these black bars may be accompanied by five very conspicuous white bars between them, the first behind the eye, the fourth behind the tympanum ; tail with black spots, or with ocelli as on the body. Lower parts white, uniform or with more or less numerous round black spots. Measurements (in millimetres) : of 2 its 3onnan 4. 5. 6. From end of snout to vent. . 96 94 8a ss9> ss - Ps » forelimb. 30 27 24 385 26 24 Length of head . : : . 19 Wor Re 20r 16-16 Width of head . : : » 12. 2A AO Depth of head . ; A eS Oe rk Fore limb . 4 : ; , (252 2a Ee? 7A 23 28 Hind limb . : : 4 . 85 sor toh) 39 929) 29 Foot . : , é ; . SEAS eee nS. PS Pail | : : ; . 215 195 155 225 145 155 1. Port Elizabeth. 2. Damaraland. 3, 5. Krugersdorp. 4. S. Africa, 6. Pirie Bush, near King Williamstown. 204. Annals of the South African Museum. The specimens in the South African Museum are from the following localities: Knysna, Kentani, Uitenhage, Port Elizabeth, Burghersdorp, Hast London, Encobo, and West Pondoland, in Cape Colony, Umvoti in Natal, Morija in Basutoland, and Johannesburg in the Transvaal. The following particulars are taken from specimens in the British Museum : Particulars of Specimens Examined. ly 28 iS: A. ib: 6. le 6 S. Africa (A. Smith). eeOe236 36, (9. 29. Soars, ” 99 65 : - 86° 38 34° 12 29 ? ? ” 9 - : » 00h e3e) B20 “925 13: 1 2 x a : - 1205740" 37 9) 30) 1814 19 3 1 aie : : d So 5387 "34 90380) IASI as ? rare : : ; 2 28 39> 35 1A 27s ToS S Port Elizabeth (Teslie) . eo Por 1365 3 oll Ae ele x se (Moorhouse). 75 41 34 10 27 12-13 19 2 a (Drege) . = 86. 36 35) “9° 795= E12 as 9 ss 7 aes el AD Son ie9 15 20 » Pirie Bush (Trevelyan) » . 88 35 385 9 24122100 ay 3 a (Stenning) . . 02" B6 S77 10s 2A aloes 2 East London : : . 106 35 “37 8 23 42 16 g E. Cape Colony . ; | |) gde 30 936, Se eeb ee 20 ,, Van Reenen, Natal : . ' 88437 "Sa eel0) 223 19 Natale i : ; . 102 34 32° 87126 ae 18 ? Sibudeni, resiteleral : » 74. FAO Fe Sbine oo) ao. eee » Lessouto, Basutoland (Lataste coll.) ; » 90535 386 10 25 WsSl4 7 g Krugersdorp, Transvaal eel ge) hap 7°29) aes alg ©) 5 a ; Gaeoo oD! eo. 9: oO meee 3 Barberton, 3 4 IS Ot. or. aes ie 20 » Damaraland . : , 94 37 34 10 28 14 49 Columns as in the riteoaiiina § species. Habitat.—Eastern parts of Cape Colony, Natal, Basutoland, Trans- vaal. As observed by Hewitt, Ann. Transv. Mus., ii, 1910, p. 114, the occurrence of this eastern species in Damaraland is doubtful and rests only on the specimen recorded above, which forms part of a series of Reptiles purchased in 1865, without any indication of the collector ; but there is this to say in favour of the correctness of the locality, that the other specimens associated with it belong to species known to inhabit S.W. Africa, A Revision of the Lizards of the Genus Nucras, Gray. 205 3. NUCRAS BOULENGERI. Nucras boulengeri, O. Neumann, Ann. and Mag. N. H. (7), v, 1900, p. 56; Sternfeld in Schubotz, Wiss. Ergebn. Deutsch. Z.-Afr. Exped., iv, i, p. 222 (1912) ; Nieden, Mitt. Zool. Mus., Berl., vii, 1913, p. 76. “ Body elongate; head not depressed, its length (to ear-opening) con- tained 43 to 5 times in the length from snout to vent ; two postnasals ; no granules between the supraoculars and the superciliaries ; inter- parietal not so long and narrow as in N. tessellata and N. delalandii ; occipital very small; subocular bordering the lip between the fourth and fifth upper labials; two supratemporals bordering the parietals ; tympanum half as large as the ear-opening. Dorsal scales small, pointed behind, larger on the sides of the body ; 45 to 53 scales round the body ; ventrals in 6 longitudinal and 27 to 30 transverse series. Femoral pores 11 or 12. Foot much shorter than the head. Tail thinner than in N. tessellata and N. delalandii, 13 to 12 as long as head and body.* Colour brown above, with small indistinct blackish spots ; bluish white boneath.”’ Distinguished from N. delalandii by the smaller size, the smaller and pointed dorsal scales, fewer ventrals, and the shorter foot. Lubwas, Usoga, British East Africa (two specimens). This species, which is only known to me from the above description, appears to be perfectly distinct. A third specimen, a male 63 mm. long from snout to vent, from Lake Victoria, has since been described by Sternfeld. 51 scales round the body, ventrals in 8 longitudinai and 34 transverse series, 12 femoral pores on each side. A fourth, from the Eldama River, British East Africa, with 10-11 femoral pores, has been noticed by Nieden. 4. Nucras INTERTEXTA. Forma typica. Lacerta intertexta, A. Smith, Mag. N. H. (2), 11, 1838, p. 93.+ Lacerta delalandii, var. b, Dum. & Bibr. Erp. Gén., v, p. 243 (1839). * Tail probably regenerated. In the specimen noticed by Nieden, it is nearly twice as long as head and body. + The type specimen, described by A. Smith and by Duméril and Bibron, was presented to the British Museum by the former author in 1865, under the name of L. delalandii, along with the types of the other Nucras in his private collection, and its absolute concordance with the original description was over- looked by me, when, following Smith himself, I placed L. intertexta in the synonymyof N. delalandii. Although not labelled as such, the specimen is cer- tainly A. Smith’s type. It was referred by me to N. tessellata, 206 Annals of the South African Museum. Nucras tessellata, part., Boulenger, Cat. Liz., iii, p. 52 (1887) ; Hewitt, Ann. Transv. Mus., ii, 1910, p. 112. Nucras tessellata, var. ocellata, Bouleng., Ann. S. Afr. Mus., v, 1910, p. 475. Nucras delalandii, part., Hewitt, t.c., p. 111. Head small, slightly broader than deep, 12 to 1% times as long as broad, its length 4 to 42 times in length to vent; snout obtuse. Pileus 21 times as long as broad. Body feebly depressed. Limbs moderate, the hind limb reaching the wrist or the elbow; foot as long as the head; digits feebly compressed. ‘Tail tapering from the base, 1} to 22 times as long as head and body. Nasals forming a short or very short suture behind the rostral ; frontonasal broader than long, broader than the internarial space ; praefrontals forming a short or very short suture; frontal as long as its distance from the end of the snout, 12 to 14 times as long as broad, narrower, behind, than the supraoculars ; frontoparietals much shorter than the frontal or than their distance from the posterior border of the pileus ; parietals 13 times as long as broad, outer border some- times emarginate for the accommodation of the anterior upper temporal; interparietal narrow, 3 times as long as the occipital, which may be broader or rudimentary and pushed back behind the pileus; parietals and interparietal shorter in proportion to their width in the very young. Four supraoculars, first and fourth small, first narrowly in contact with the frontal; 5 or 6 superciliaries ; 2 to 4 small scales between the supraoculars and the superciliaries. Two superposed postnasals; anterior loreal barely half as long as second ; 4 upper labials anterior to the subocular, which is usually narrower beneath than above; an elongate upper temporal, in contact with the fourth supraocular, followed by 2 or 3 smaller shields; temple covered with small hexagonal or granular scales, which are about as large as the dorsals ; a round or oval tympanic shield. Lower eyelid with 5 or 6 vertically enlarged scales in the middle. Parietal foramen and pterygoid teeth present. Gular scales small, juxtaposed, increasing in size and imbricate towards the collar, 27 to 36 in a straight longitudinal series ; no gular fold. Collar even-edged or feebly serrated, composed of 10 to 13 plates. Scales on body oval, juxtaposed, smooth, 40 to 44 across the middle of the body, 2 and 3 on the side corresponding to a ventral plate. Ventral plates in 6 or 8 longitudinal series, those of the second series from the median line the broadest, and 29 to 34 transverse series A Revision of the Lizards of the Genus Nucras, Gray. 207 Praeanal plate large, with a smaller one on each side and a large pair in front, or two subequal praeanals, one in front of the other. A series of 4 to 7 transversely enlarged plates on the lower surface of the fore limb. Scales on upper surface of tibia smaller than dorsals. 11 to 14 femoral pores on each side. 20 to 25 lamellar scales under the fourth toe. Caudal scales forming whorls of nearly equal length, upper rather narrow, the median pair sometimes broader, rather strongly keeled, truncate behind, with distinct sensory pits. The type specimen, a female from Latakoo, near Kuruman, now rather bleached, has the markings well preserved, although the black has turned to a pale brown, and answers to A. Smith’s diagnosis : “Colour above, reddish brown, with two rows of circular white spots, discontinued about half-way between the anterior and _ posterior ~ extremities, each spot surrounded by a black ring; sides chequered, black and white, the latter colour disposed in narrow vertical stripes. Tail light brown, with a dotted black line on each side, and the space between them above marked with small black spots. Under parts white.” This description is supplemented by a very good account of the same specimen by Duméril and Bibron, of which this is a translation : Instead of a great number of small black spots with white pupils (as in L. delalandii), there are only two series, but a little larger, on each side of the back. Two or three irregular blackish spots on the upper lip. Two vertical blackish stripes on the temple, which is white ; a third above the ear, and three or four on the neck. Others along the flanks, but shorter; on examining them carefully, one may guess how they were formed. It is probable that, in early youth, white spots encircled with black existed on the flanks ; gradually, as they enlarged, the black circle opened above and beneath; then each of the two portions became raised and fused with the other, whilst simul- taneously the white central spots enlarged vertically, thus producing alternating black and white vertical bars. Upper surface of hind limbs with some white spots incompletely surrounded with blackish. Here and there some black spots on the upper surface of the base of the tail; others, smaller, are present on the sides, so regularly arranged and so crowded as to form a longitudinal stripe (Pl. VI, fig. 8). The interpretation given to the markings by the authors of the ‘ Erpétologie Générale’ is fully confirmed by the examination of the young, with which we are now acquainted (var. ocellata, Blgr.). Very young specimens (37-40 mm. to vent), from Pietersburg, Transvaal, are dark brown above and blackish on the sides, with 208 Annals of the South African Museum. numerous white ocelli in three or four series on the back and three series on each side; a white vertebral streak on the nape, which may be continued, interrupted, on the body ; sides of head and neck with black and white vertical bars ; upper orbital border whitish ; a white streak on each parietal shield, continuous with the outer dorsal series of ocelli; tail coral-red. In a larger young (43 mm.), from Kokong. Bechuanaland, the dorsal markings are the same, but the ocelli on the sides of the body have fused to form vertical bars (Pl. VI, fig. 7). A half-grown female, from Rustenburg, Transvaal, is reddish brown above, with an interrupted light, black-edged vertebral streak, a dorso- lateral series of ocelli, and three series of ocelli on each side, the lower of which are more or less confluent into a light longitudinal streak from the shoulder to the root of the hind limb; head and neck as in the preceding (Pl. VI, fig. 6). A half-grown male, from Rustenburg, Transvaal, is similar to the preceding, but the white eyes of the ocelli on the nape are in the form of longitudinal lines, whilst the black borders of the ocelli run together to form cross-bands on the back, as is frequent in N. delalandii, from the young of which it is hardly to be distinguished, so far as the coloration is concerned. Measurements (in millimetres) : I. 2. 3. From end of snout to vent . : ») 804 163°. 55 Pe ‘ fore limb . bean 22iy Well, Length of head . ; : camel fay eal 2, Width of head . : : . ee, el: 9 8 Depth of head . : 2°75 7 Fore limb . : ; 2 : cue deen! 205046 Hind limb . ; ‘ ; : fi BA QS OG Foot . : : 4 3 ade! Wim Gas § 9213} Htanilayes : : ; ‘ NOG ae l. 9, type, Latakoo. 2. 2, Pietersburg, Transvaal. 3. ¢, Rus- tenburg, Transvaal. Under the name of var. holubi, Stdr., I group together a number of specimens which, whilst agreeing essentially in structure with N. intertexta, differ from the type in the back being striated through- out life. Var. holubi. Lacerta tessellata, part., Peters, Reise Mossamb., iui, p. 44 (1882). Eremias holubi, Steind., Sizb. Ak, Wien, lxxxvi, i, 1882, p. 83, pl. —. ae Ger Se in, A Revision of the Lizards of the Genus Nucras, Gray. 209 Lacerta cameranoi, Bedriaga, Abh. Senck. Ges., xiv, 1886, p. 378 pl. —, figs. 2, 9, 11, 31. Nucras tessellata, part., Bouleng., Cat. Liz., iii, p. 52 (1887). Nucras tessellata, Bouleng., in Distant, Nat. Transv., p. 174 (1892). Nucras tessellata, var. taeniolata, Bocage, Herp. Ang., p. 30 (1895). Nucras tessellata, var. ornata, Bouleng., Ann. Natal Mus., i, 1908, p. 225. Nucras tessellata, vars. holubi, ornata, Bouleng., Ann. S. Afr. Mus., v, 1910, p. 474. Nueras holubi, Sternf. in Schubotz, Wiss. Ergebn. Deutsch. Z.-Afr. Exped., iv, ii, p. 222 (1912). Head 32 to 43 times in length to vent, sometimes as deep as broad, sometimes a little broader, the cheeks often swollen in the males, Pileus usually twice as long as broad. The hind limb reaches the wrist or the elbow, rarely the axil* or just overlaps the fore limb f ; foot as long as or slightly longer or slightly shorter than the head. Lepidosis as in the typical form, but suture between the praefrontals sometimes longer, frontal sometimes nearly twice as long as broad. interparietal often broader (2 to 3 times as long as broad), first supra- ocular often extensively in contact with the frontal, the fourth some- times separated from the anterior upper temporal {; 2 to 6 small scales between the supraoculars and the superciliaries, of which there may be 7; anterior loreal sometimes more than half as long as second § ; tympanic sometimes very small, rarely absent. 25 to 33 gular scales in a longitudinal series ; collar composed of 7 to 14 plates. 44 to 60 scales across the middle of the body. Ventral pilates in 27 to 34 transverse series. A large praeanal bordered by 4 or 6 smaller shields, or 2 large praeanals, one in front of the other, or 3 forming a triangle bordered by a semicircle of small plates. 11 to 20 femoral pores on each side. 20 to 26 lamellar scales under the fourth toe. * Male and young from Bulawayo. + Female from Lake Nyassa. { Males from Vredefort Road and Rustenburg, females from Lydenburg and Kimberley. The upper temporal is then entirely on the temple. Bedriaga observes, @ propos of his L. cameranoi, that the wpper temporals are on the upper surface, forming part of the pileus, in the South African species (my Nucras). The series of specimens here referred to N. intertexta shows this character to be by no means a constant one, as these shields may be lateral and perpendicular to the parietals. There is thus in Nucras the same amount of variation with respect to this feature as in L. muralis, in which Méhely has used it for the distinction of his Archaeolacertae and Neolacertae. § A single postnasal on one side in a young from Bulawayo. 210 Annals of the South African Museum. Varies much in markings. The principal variations may be arranged as follows, starting with the most primitive. A. (N. tessellata, var. taeniolata, Bocage.). Four or five* white dorsal streaks separated by wider dark brown interspaces, and three white streaks on each side, the upper (proceeding from the temple above the ear-opening) broken up, anteriorly, into a series of round spots ; on the posterior part of the body, these markings fade into a pale buff colour, which also occupies the upper surface of the limbs and tail. The coloration is thus very similar to that of Smith’s L. taeniolata.—Dongwenna, Mossamedes. (PI. VII, fig. 1.) B. (E. holubi, Stdr., l.c., lower figure). Three white dorsal streaks separated by broader black or dark brown interspaces, and 2 (some- times broken up into spots) along each side; the white vertebral streak continued for a short distance on the tail, which bears 3 dark longitudinal streaks; the outer dorsal light streak extends on the parietal shield, where it joins the hght supraorbital border.—Limpopo Valley, Transvaal (Steindachner) ; Rustenburg, Transvaal ; Vredefort Road, Orange River Colony ; Kimberley, Burghersdorp, Cape Colony. (Bl. Vid, fig:.2). C. (#. holubi, Stdr., 1c., upper figure). Back reddish brown, with 3 dark-edged light streaks; a broad dark brown or black lateral band from the temple to above the hind limb, bearing 1, 2, or 3 series of roundish white spots, and edged below by a white streak which may be broken up into spots.— Limpopo Valley (Steindachner) ; Zoutpansburg, Transvaal; Lydenburg, Transvaal; Vredefort Road, Orange River Colony; Bulawayo; Port Elizabeth. (PL VIL, fig. 3). D. As in the preceding, but temple and side of neck with black and white vertical bars.—Umfolosi River, Natal; Pretoria; Bindura, S. Rhodesia. (Pl. VII, fig. 4). E. The black and white vertical bars are continued, more or less distinct, on the flanks.—Umfolosi River. (Pl. VIL fig. 5). F. Back reddish brown with black dots and mere traces of the 3 light streaks ; a blackish lateral band with very numerous small round white spots; sides of head with black and white vertical bars, tail with numerous small dark and light spots.—Lake Nyassa. (PI. VI, fig. 10). This form appears to represent Bedriaga’s L. cameranoi, from Tette, Mozambique, but the fingers are not quite so short +, the figure accompanying the description showing them to be very similar to those of N. delalandia. Four in the male, five in the female; only two specimens examined. + They are shorter and thicker in the female than in the male. A Revision of the Lizards of the Genus Nucras, Gray. 211 G. As in £, but without the light vertebral streak, and with black dots on the back and on the sides of the belly —Umfolosi River. H. As in D, but no light vertebral streak, and the light dorso-lateral streak ending midway between the fore and hind limbs; black dots on the sides of the belly. This variation forms a complete connection with the typical N. intertexta, the only difference being that the light ocellar spots on the nape and anterior part of the back have fused to form a dark-edved lateral streak.—De Kaap Goldfields, Transvaal. (PIS Vir fies 9): All the young specimens examined have 3 or 5 light dorsal streaks and the tail is of a coralline red. The var. holubi must be regarded as more primitive than the typical form, and the pattern described under a, along with the taeniolata form of N. tessellata, as the original from which all others in the genus ean be derived without the least difficulty. Measurements (in millimetres) : From end of snout to vent 80 96 95 5 86 55 62 58 70 sy 5 comore limbs. 28°29 34 25) 330 1920223 Head . < . 5 : , 197 20) 245 i ZO Eanes 15 Width of head . ; : a TP BS is, SL Ie 8 10 Depth of head. ; ‘ 5 | A ah ss NO’ 1® 7 87) 4835 Fore limb . 5 : : eeZon 826) (299225 ee mol ae peel oO Hind limb . ; F f oS Ol | 45° S33 SG Z6eeZO eo eSs Foot . : : : 3 to) 20) 21 ai, See ISi See AeG Aten ; ‘ : lO — 215; 1123) S80 Geel35 1. g, Nyassa. 2. 2, Nyassa. 3. 6, Umfolosi R. 4 9, Zoutpansberg. 5. 2, De Kaap Goldfields. 6. g, Vredefort Rd. 7. 9, Vredefort Rd. 8. 3, Dongwenna. 9. 2, Dongwenna. Particulars of Specimens Examined. ere Ob 42 8, le & @) Forma typica. 2 Latakoo (type) . ; : . 80 40 8 32 12 34 11 24 48 Her. Kokong, Bechuanaland . . 44 42 6 31 13 33 %18 24 438 g Rustenburg, Transvaal : . 55 44 6 29 18 3834 13-14 22 3 2 Pietersburg, - f : ; 63) 44) (8) 3451236 14. 25 4 ice 55 ; : : : - 40) 43) Sers0s OMS is—12) 23) 3 i" 7 ‘ : ; : - 88 42°58) 23 1e aie st 12 20 2 var. holubi. g Port Elizabeth ‘ j 3 O00) Bis} =) RS 11 21 45 ,, Burghersdorp, Cape Col. : - OO) AO NGHeZSeaiin2s 1 24 65 ? Kimberley (8S. A. Mus.) F = OZ) 44 SRS alle 29) 1 2—13)9 922) 4, & Vredefort Road, O.R. Col. . 2 05 | 460 9Sie29Ne 12828 12 24 4-5 i 53 % 53 é . 52-48 8 28 10 26 12-14 21 5 212 Annals of the South African Museum. eae eS: As OSA: tf: 8. 9. var. holubi. 2 Vredefort Road, O.R. Col. . . 62 48 8 380 12 26 1413 21 A ss ae 5 5 F So0 mp2 18 30 100270) 12 ies e * - * » . 85 8 29 88. a8" 20s & Umfolosi R., Zululand . : 5 63) (BO) ich SH IPA eA ibe) 22) D2) x ek, gh ORT) Me GIGS mae ang "87 (418 (San Ee eon meer We, [3 Bs é , 46 5746 932° 14) 33 1G ebb 4 3 De Kaap Goldfields, Transvaal . 86 55 6 30 12 32 15-16 23 48 Ye. Pretoria, Ss - 44 48 6 30 10 28 14 26 3-4 6 Rustenburg, 35 . b8 45 8 29 14 31 13 21 5-6 @ Lydenburg, 25 5 OO BY Se sil Oy ese Weiss 25 ie if ¥ i. 1764.49 8 31 “11 30 1721624 » Zoutpansberg es . 75 55 8 382 11 382 20-19 24 38 6 Bulawayo : 5 : , BbO 50) 8ie27 28) AsSb 22 0G Nr rshey; , j g en 42 58.6529 ) 205888 hy Ob we oueey ? Bindura, 8. Rhodesia (S.A. Mus.) 49 46 6 28 10 25 15-17 25 4 6 L. Nyassa 5 5 é . ; 80 44 6 30 10 30 16-15 255 4 2 a UE) i eG) AG. G o31. (46eeSiI Miele ee 3 Dongwenna, Mossamedes_. es, 45) 8) 928) 1027 43 bees 9 i i‘ 370. 45 8 84° 12 30) 1a bees 1. Length from snout to vent. 2. Number of scales across middle of body. 3. Longitudinal series of ventral plates. 4. Transverse series of ventral plates. 5. Plates in collar. 6. Gular scales in a straight line between symphysis of chin-shields and median collar-plate. 7. Femoral pores (right and left if differ- ing). 8. Lamellar scales under fourth toe. 9. Scales between supraoculars and superciliaries. Habitat.—The range of N. intertexta extends from Portuguese Hast Africa, Nyassaland, and Angola to the northern and eastern parts of Cape Colony. The specimens in the South African Museum are from Burghersdorp, Little Namaqualand, and Kimberley in Cape Colony, Smithfield in the Orange River Colony, Barberton in the Transvaal, Bindura and Bulawayo in Southern Rhodesia. 5. NUCRAS TESSELLATA. Lacerta tessellata, A Smith, Mag. N. H. (2) ii, 1838, p. 92; Dum. & Bibr. Erp. Gén., v, p. 244 (1839); Bedriaga, Abh. Senck. Ges., xiv, 1886, p. 374. Lacerta livida, A. Smith, Le. Lacerta elegans, A. Smith, l.c. Lacerta taeniolata, A. Smith, t.c., p. 98; Dum. & Bibr., t.c., p. 247 Bedriagia, t.c., p. 381. Zootoca taeniolata, Gray, Cat. Liz., p. 29 (1845). Nucras tessellata, Gray, op. cit., p. 33; Werner, Jen. Denkschr., iv, p- 829 (1910). A Revision of the Lizards of the Genus Nueras, Gray. 213 Teira ornata, Gray, Proc. Zool. Soc., 1864, p. 58. Lacerta tessellata, part., Peters, Reise Mossamb., iii, p. 44 (1882), Lacerta tessellata, subsp. pseudotessellata, Bedriaga, t.c., p. 377, pl. —, figs. 8, 21. Nucras tessellata, part., Bouleng., Cat. Liz., iii, p. 52 (1887). Nucras tessellata, vars. elegans, livida, taeniolata, Bouleng., Ann. 8S. Afr. Mus., v, 1910, p. 474. Head small, considerably broader than deep, 12 to 14 times as long broad, its length 4 to 43 times in length to vent; snout obtuse. Pileus 2 to 2¢ times as long as broad. Body feebly depressed. Hind limb reaching the elbow, the axil, or the shoulder; foot considerably longer than the head ; digits slender, feebly compressed. Tail flattened and widened at the base in males, nearly 2 to 2? times as long as head and body. Nasals forming a short suture behind the rostral; frontonasal broader than long, broader than the internarial space; praefrontals forming a short suture; frontal as long as or a little longer than its distance from the end of the snout, 13 to 2 times as long as broad, narrower, behind, than the supraoculars ; frontoparietals much shorter than the frontal or than their distance from the posterior border of the pileus ; parietals 14 to 1? times as long as broad, outer border some- times emarginate for the accommodation of the anterior upper temporal; interparietal narrow, 2} to 4 times as long as_ broad ; occipital very short, sometimes broader than the interparietal. Four supraoculars, first and fourth small, and sometimes divided into two, first extensively in contact with the frontal; 7 or 8 superciliaries ; 2 to 7 small scales between the supraoculars and the superciliaries. Two superposed postnasals, rarely one * ; anterior loreal } to ¢ times as long as second; 4 upper labials anterior to the subocular, which is a little narrower beneath than above; an elongate anterior upper temporal, often in contact with the fourth supraocular f, followed by 1 or 2 smaller shields; temple covered with small hexagonal or gran- ular scales, which are about as large as the dorsals or smaller; tym- panic shield roundish, often small or absent. Parietal foramen usually absent. Pterygoid teeth present. Gular scales small, juxtaposed, increasing in size and imbricate * Types of L.taeniolata. Also in a young from Clanwilliam which, in its markings, agrees with the typical L. tessellata. + Not in contact in five specimens: one of the types of L. taeniolata, two of the types of L. livida, male from Deelfontein, and female from Little Namaqua- land. 214 Annals of the South African Museum. towards the collar, 25 to 33 in a straight longitudinal series; no gular fold. Collar even-edged, composed of 8 to 13 plates. Seales on body roundish or oval-hexagonal, smooth, 40 to 60 across the middle of the body, 2 and 3 on the side corresponding to a ventral plate. Ventral plates in 6 or8 longitudinal series, those of the second series from the median line the broadest, and 25 to 34 transverse series. Praeanal region covered with several irregular shields, or with two large shields one in front of the other. A series of 6 or 7 transversely enlarged plates on the lower surface of the fore limb. Scales on upper surface of tibia smaller than dorsals. 11 to 16 femoral pores on each side. 25 to 31 lamellar scales under the fourth toe. Caudal scales forming whorls of nearly equal length, upper rather narrow, the median pair often broader, rather strongly keeled, truncate or very obtusely pointed behind, with more or less distinct sensory pits. As in the preceding species, the markings differ very strikingly according to individuals, and some at least of the different patterns, on which species have been founded, perhaps indicate local forms or varieties. JI here enumerate those with which I am acquainted, be- ginning with the most primitive : A. (L. taeniolata, Smith).—EHight white streaks on the back and sides, sometimes nine on the nape and anterior part of back, separated by black streaks ; the outer dorsal light streak extending to the fourth supraocular, the upper lateral, originating just above the ear, some- times broken up into spots. Posterior part of back and tail brown above, the latter inclined to red near the extremity and with a blackish lateral streak. Lower parts white.—‘“ Grassy districts of Cape Colony,” Smith; Little Namaqualand; Pine Town, Natal (South African Museum). (Pl. VII, fig. 6.) B. (L. livida, Smith).— Back with light and dark streaks as in the preceding, or pale buff behind with black vermiculations; sides black with numerous small white spots, which form irregular vertical bars on the temple and neck.—‘ Northern parts of Cape Colony,” Smith ; Little Namaqualand ; Deelfontein. (Pl. VII, figs. 7, 8.) C. (ZL. tessellata, Smith; T. ornata, Gray).—Neck and anterior part of back black, with 3 or 4 white lines above and very regular white vertical bars on the sides ; posterior part of body grey or pale buff, with more or less distinct black bars on the sides. Feet and tail coral-red or reddish, at least in the young.—* Hastern parts of Cape Colony,” Smith ; Clanwilliam, Calvinia, Worcester, Klipfontein, in Cape Colony; Zambesi (Sir J. Kirk). (Pl. VII, figs. 9, 10.) A Revision of the Lizards of the Genus Nucras, Gray. 215 In the var. pseudotessellata, Bedr., from Mozambique, there are 5 white lines on the nape. D. (ZL. elegans, Smith).—Pale reddish brown above and on the sides ; two white, black-edged streaks on the neck.—‘ Little Namaqua- land and the country towards the Orange River,’ Smith; Smithfield, Orange River Colony. In the following tabulation of specimens examined the same arrange- ment is adopted : Measurements (in millimetres) : 3 2 => SSS 1., “2S eee From end of snout to vent ~ 10. Go eae: 80 62 3 a » forelimb. 24 24 18 25 28 20 Head . , : : ‘ S ley al ale nf al} Width of head . : ; ~ 9 S10 iO 12 8 Depth of head . : : oe ikl OE iOMmES 6 Fore limb . - : 3 . 20 22 16 22 22 18 Hind limbrwik ea 4. + 84.87%) S2sowtONe 29 Foot . 3 . . F ; 19> 2079 owezill 21 16 dig . — — 125 — — 120 1. Type of L. tessellata. 2,4. Little Namaqualand. 3. Klipfontein. 5. Type of L. livida. 6. Type of L. taeniolata. Particulars of Specimens Examined. 1D). } A. i. 2, “Sh Age oneo: Te Sino: 2? Type of L. taeniolata . : . 62 48 6 32 11 28 15-14 26 67 You |. if we, . 40° 45. Ga 27 isOnS=16neo an 85 ? Little Namaqualand. < > Sf 52 8) 2OR1ON 29 3) 228) 4 B. 3 Type of L. livida . : 5 - 04 AGS 1G RZ5e Ge 29 13 27 43 f . - es), 47 «(SGP 1G 208 wees de =15) 2208 62 2 oS 5 5 . : , 80 47 8) 730) iss ial 28 5 $ Deelfontein, Cape Colony . « (2 48) 78ee Zeer st 13 26 43 Cc. 3 Type of L. tessellata : 7 (0) At NOR SIS Sees leo —3 e270) iG » Little Namaqualand . i patie etsy (248) 8) BY) 15 27 4 a es + : 5 . (4 42 6 381 12 30 15 27 5-4 3g Guires, Little Namaqualand . 58 A445 Greeieeloee it 15 28 4 Yg. Clanwilliam, W. Cape Colony . 40 41 6 81 9 29 14-15 30 17-6 s i A . 40, 40 GMESIONOueS2 a4, 9G) (54 3 Klipfontein, E. Cape Colony . At 455 ORSON 88 15-16) 27 G=4 Yg. Zambesi, type of T. ornata - 32) ADRORRCOMP One 26) 05 1405 82604: D. 2? Type of L. elegans . j : 2 CO) 4ORISEoomacu ns § 16—14 ail 2 xD as d , ‘ «G2; “AGeeShese oS. 2 14-15-80) 4, » Smithfield, O.R. Col. (S.A. Mus.) . 63 45 6 32 8 33 14. 29 3 (1) Tabulated as in the preceding species. 216 Annals of the South African Museum. I have examined in addition 19 specimens preserved in the South African Museum. Scales across the body 40 to 60; femoral pores 12 to 16. One specimen, from Little Namaqualand, with a single postnasal. The habitat of N. tessellata is a wide one, extending from Great Namaqualand to the Karroo and Natal, and the species being also on record from Mozambique (Berlin Museum) and the Zambesi (Sir J. Kirk), it will probably be discovered in Southern Rhodesia. The species of Nucras appear to be of very local occurrence, and much more collecting will have to be done before their distribution can be properly mapped out. It is hoped that this contribution to the knowledge of them may be an incentive to the collecting and study of further material. EXPLANATION OF THE PLATES. . Prats VI. Fig. 1. Nucras delalandii, yonng. East London, 2. DEN © ips os 6. Van Reenen, Natal. &. 3 » . Krugersdorp, Transvaal. lee ie Ss 6. Barberton, Transvaal. Be 5; 53 2. Krugersdorp, Transvaal. Game. intertexta, 2. Pietersburg, Transvaal. 4%. HEM. tess 45 young. Kokong, Bechuanaland. 2. Sie es y, type. Latakoo, near Kuruman. Owes ms var. holubi. 92. De Kaap, Transvaal. HO Sines a 35 uy 6. Lake Nyassa. Priarn VII. Fig. 1. Nucras intertexta, var. holubi. 9. Dongwenna, Mossamedes. §&. 2. » bai ep 6. Burghersdorp, Cape Colony. 2. BB » oy er 35 6. Barberton, Transvaal. 2. 4 + ve a S 6. Umfolosi, Natal, 42. 5. ” ” 29 2” 3 7 2 2” Bs Goes tesseliata, young, type of L. taeniolata. Cape Colony. 2. ees a 3. Deelfontein, Cape Colony. 32. shu) doy a 3, type of L. livida. Cape Colony. 12. Cet 7 3, type. Cape Colony. #. ) LOS Gis 44 young. §8. Africa. 2. Nucras intertexta. Nucras delalandii. 0) 10 NIN ELD imp N eV 18) Nucras te a?) yyy La | Np acs at JD Oa 24 F ~~ ane ay To aeete —o= sr * o 2 @F ~papeae : 9 Pao merely to ac RRAN YONI TEBE Seat: fo <= Nucras intertexta. cp) Adlard &West Newman mp J.Green. delet lith, i } } . (217) 11.— Description of a New South African Lizard of the genus Eremias. —By G. A. Boutenerr, LL.D., D.Sc., F.R.S. (Published by permission of the Trustees of the British Museum.) EREMIAS ASPERA, Sp. 0. Head and body rather strongly depressed. Head 1% times as long as broad, its length 4 to 4¢ times in length to vent, its depth equal to the distance between the centre of the eye and the tympanum; snout obtusely pointed, with the nasals very feebly swollen, as long as the postocular part of the head; a feeble concavity in the middle of the upper surface of the snout, extending to the middle of the frontal shield. Pileus nearly twice as long as broad. Neck a little narrower than the head. Hind limb reaching the collar or a little beyond ; foot 14 times as long as the head; toes rather slender, feebly com- pressed. ‘Tail twice or a little less than twice as long as head and body. Upper head-shields flat, very rugose, with granular asperities. Nasals in contact with each other behind the rostral, the suture between them + to ; the length of the frontonasal, which is a little broader than long; praefrontals as long as broad or longer than broad, separated by an azygos shield ; frontal a little shorter than its distance from the end of the snout, 13 to 1% times as long as broad, separated from the supraoculars by a series of small scales ; parietals as long as broad; interparietal larger than the frontoparietals, in contact with or narrowly separated from a very small occipital. Two supraoculars, together hardly 2 the length of the supraocular region, the first shorter than the second, the space in front of them filled by 4 or 5 transverse series of granules ; 6 or 7 superciliaries, separated from the supraoculars by 2 or 3 series of granules. Three nasals, lower in contact with the rostral and the first upper labial; anterior loreal as long as deep or a little deeper than long, shorter than the second ; 4 or 5 upper labials anterior to the subocular, which is keeled below the eye and largely borders the mouth, its lower border much narrower than the upper. Temporal scales granular, keeled; no 218 Annals of the South African Museum. tympanic shield; no auricular denticulation. Lower eyelid with a transparent disc formed of two large black-edged scales. A pairs of chin-shields, the first 3 in contact in the middle. 31 or 32 gular scales in a straight line between the symphysis of the chin-shields and the median collar-plate; no gular fold. Collar straight, free, composed of 11 plates. Seales granular and keeled on the nape, rhombic, feebly imbricate, and strongly keeled on the body, graduating into the caudals, larger and smooth towards the ventrals; 51 to 56 scales across the middle of the body. Ventral plates in 12 or 14 rather irregular longitudinal and 33 to 37 rather angular transverse series, the plates as long as broad or a little broader than long. Praeanal region covered with small irregular plates. Scales on upper surface of fore hmb moderately large, keeled. Scales on upper surface of tibia rhombic, feebly imbricate, keeled, as large as the posterior dorsals; one series of very large and one of small plates on the lower surface; 15 or 14 femoral pores on each side. Subdigital lamellae bicarinate, spinulose, 21 under the fourth toe. Upper caudal scales oblique, truncate behind, strongly and diagonally keeled, basal subcaudals feebly keeled ; 24 or 26 scales in the fourth or fifth whorl. Pale greyish-brown above, with 2 dorsal and 2 lateral darker bands, and 4 series of whitish, brown-edged ocellar spots on the back; a white streak below the dark lateral band, from below the eye, through the tympanum, to the groin; tail yellowish with a brown lateral streak in the anterior half. Lower parts white. Measurements : From end of snout to vent : : . Ol millim. - a fore limb ; rAd Olas Head 5 : ; ‘ ; : we AD Width of head . : ; 3 : : Oe Depth of head . : ; ; : : GO ass Fore limb. : : : : ; |, GLOWS aie Hind limb ‘ , : : : ty Toots Foot. ; ‘ : . : : atu Alene Taal. : é : a VOR ak. Habitat.—Bechuanaland. Described from two female specimens from Mochudi, received by the South African Museum from Mr, J. Fenton, along with examples of Hremias lugubris and Ichnotropis capensis from the same locality. This species is closely related to the typical EH. lineo-ocellata, differing in the very rugose upper head-shields, the absence of projecting scales Description of a New South African Lizard. 219 in front of the ear, and the lower number of subdigital lamellae (21 instead of 24 to 28). The number of scales across the body is low (51 or 56 as against 53 to 73), as compared with H. lineo-ocellata, in which I have counted the scales in 64 specimens and find only 8 with less than 57. x. a] 3 The Echinoderm Fauna of South Africa. 227 indicates the lapse of a long time since this deep water fauna entered the South African region. Perhaps the conclusion is more sweeping than the facts warrant but the impression left by the study of the South African fauna is that the region south of Delagoa Bay now forms a very distinct zoogeographical region, only superficially connected with the Indo- Pacific region to the north and east, and quite isolated from any other region; that its original echinoderm fauna was common to a large continental area to the northwest in the Atlantic and to the southward; and that its present day littoral fauna has moved in from the northeast under the influence of the Agulhas current, but res- tricted by the cold winter water from the southeast. SEA-LILIES. CRINOIDEA. The crinoids form a very insignificant part of the Echinoderm fauna of South Africa. They were listed in 1915 by Mr. Austin Hobart Clark (Deutsche Siid-Polar Exp.: Zoologie, vol. 8, p. 163) who gives three species as occuring along shore in 0-30 fms. and two species as occuring in deep water, 250-450 fms. The col- lection of the South African Museum (45 specimens) contains four of these five species and also four species not known hitherto from the South African region. Of these, one is from comparatively shal- low water (90 fms.) but the other three were taken by the PIETER Faure only in depths of 900-1000 fms. It is interesting to see therefore that the South African crinoids fall into three groups of three species each, an ‘“abyssal” group of two stalked forms and a five-armed comatulid, a “continental” group of comatulids and a “littoral” group of comatulids. Of the abyssal group, one (Monachocrinus coelus) appears to be a new species of a genus previously known from both the Atlantic and Indian Oceans. A second species (Bythocrinus chun) was hitherto known only from the western part of the Indian Ocean near the Somaliland coast in something over 900 fms. The third species (Pentametrocrinus varians) was hitherto known only from the northeastern Indian Ocean, the vicinity of the Philippine Islands and southern Japan, in 361-1050 fms. It is interesting to note that the VaLprviA took an as-yet-undescribed species of Pentametrocrinus in the same region where Bythocrinus chuni was taken, but in slightly shallower water. The Prerer Faure found the two genera at the same station. Of the three continental comatulids, one (Liparometra multicirra) appears to be an undescribed species of a wide-spread East Indian 228 Annals of the South African Museum. group, while the other two are also apparently endemic species of Kast Indian genera. The three littoral species are of particular interest in connection with questions of geographical distribution. One ( Cominia occidentalis) is a peculiar, endemic species of a genus known otherwise only from Korea Strait in 170 fms. while a second (Comanthus wahlbergii), also endemic, finds its nearest relative in a South Australian species. The third South African littoral comatulid is the wide-ranging T'o- piometra carinata, which occurs from Zanzibar, the Seychelles and Mauritius, southward around the Cape of Good Hope and thence northwestward to St. Helena, the coast of Brazil and the extreme southeastern West Indies. The genus is otherwise distinctly East Indian. It is evident therefore that the South African crinoid fauna is essentially Kast Indian in its relationships and no doubt in its origin also. The only exception is the Comanthus which is closely allied to a species known only from southern Australia in shallow water. This clearly hints at a common origin for the two and suggests interesting speculations. In the following key to South African crinoids, | have used only the simplest and most obvious characters. There are two reasons for this: first, Mr. Austin Hobart Clark, in his most useful work on the Crinoids of the Indian Ocean (1912, Echinoderma of the Indian Museum, pt. 7) has given admirable keys to the families and genera and it is therefore quite superfluous for me to repeat his work; second, the South African species of Cominia is so unlike the Coman- thus that it is not feasible to fit them into the same section of a brief, artificial key and | have therefore ignored their family relationship. Consequently the following key is absolutely artificial and does not give the natural sequence of the species, a sequence which is followed in the subsequent pages. The number of species involved is, however, so small that little inconvenience will result from the inconsistency. Key to the South African Species of Crinoids. Stalk present. Only one or two short discoidal segments at top of stalk, immediately below calyx : ‘ : 3 Bythocrinus chunt. More than a dozen short discoidal segments at top of stalk Monachocrinus coelus. Stalk wanting (Comatulids), Arms 10 or more. Cirri numerous, 35—40. Cirrus segments few, 14—18 . : . Cominia occidentalis. Cirrus segments many, 30—36 . 3 LInparometra multicirra. suseenpeeeneneneenaeeeiteaeenaieeaamaeaminmaatl The Echinoderm Fauna of South Africa. 229 Cirri not so numerous, 12—30. Cirrus segments few, 12—20. Mouth excentric; basal pinnules of arms with terminal comb Comanthus wahlbergu. Mouth central; no terminal combs on pinnules. Arms 10, more or less compressed; arm-segments beyond middle of arm, short, the distal portion more or less rough and projecting . 3 Tropiometra carinata. Arms 10, or more in adults not compressed; arm-segments beyond middle of arm not short with the distal portion more or less rough and projecting Pachylometra sclatert. Cirrus segments numerous, 50—63 : Crotalometra magnicirra. Arms only five. . : : : . Pentametrocrinus varians. BOURGUETICRINIDAE. ByTHOCRINUS CHUNI. Rhizocrinus chuni Déderlein, 1907, Stpoga Stalked Crinoids, feet tics 6 5 pl. A figmo: Rhizocrinus (Bythocrinus) chuni Déderlein, 1912. VaLpivia Stalked Crinoids, p. 14, pl. 3. The specimens of Bythocrinus in the collection all lack the arms, and only two have the calyx still intact. The best has the stalk 47 mm. long and half a millimeter thick at the top; the calyx is 3 mm. high and rather more than 1°5 mm. in diameter at the top. All the specimens are white. They answer so well to Déderlein’s descrip- tion and figures of his specimens from off the Somaliland coast, that, in spite of their imperfect condition, their identity seems sure. Pierer Faure. 17350. Cape Point. N. 86° E., 43 miles. 900—1000 fms. Gray mud. 4 specimens. MONACHOCRINUS COELUS, * sp. nov. Plate VIII. Fig..4. Fragment of upper part of stem present, not quite 7 mm. long, about ‘-30 mm. in diameter at broken end, a very little thicker where it joins calyx; it is made up of 29 segments of which the topmost 42 are very low and discoidal, the height about one-fifth or one-sixth the diameter; the next six are discoidal but successively higher; the nineteenth is nearly, and the twentieth quite, as high as thick, and * xovdoc = hollowed, in reference to the slightly concave lines of basals and radials. 230 Annals of the South African Museum. the remainder are much higher than broad (the 29th is three times as high as thick), smooth and cylindrical. Basals completely fused into a truncated cone, about half a milli- meter high, nearly half a millimeter in diameter, where it joins the radials, and about one-third of a millimeter where it joims the stem. Seen from the side, the lateral margins of this cone are distinctly though very slightly concave. Radials 5, about °75 mm. high; the upper (distal) diameter of the cup they form is one millimeter. Seen from the side, the lateral margins of this cup are distinctly though slightly concave. I Br, about 1140 mm. long and ‘85 mm. wide, very little wider distally than proximally. The lateral margins are very slightly thinned and flaring. The median line is not at all carmate but is barely elevated on the distal two-thirds of the plate. I Br,, the axillary, is remarkably low and wide; it measures about °85 mm. in width, but is only about ‘60 mm. high, even in the median line where it is slightly higher than at the sides. The lateral portions are flat, in contrast to the middle, but are hardly flaring. The brachials are about twenty in number; the lowest is about ‘40 mm. wide where it joins the axillary but is only about -35 mm. at the distal end, and that is the approximate width of the following segments. The brachials are arranged in pairs, *75—"80 mm. long, the total length of the arms, from axillary to tip being about 8 mm. The latero-distal margin of the distal brachial of each pair is slightly projecting and overlapping, first on the outer side of the arm ‘second brachial), then on the inner (fourth), and thus in regular alternation, but the projection is much too slight to give the arm a serrate or even a rough appearance. Colour, nearly white. P.F. 17350. Cape Point N. 86° E., 43 miles. 900-1000 fms. Gray mud. 4specimenonly. Holotype South African Museum, no. A 6434, This is a most interesting little crimoid, clearly a Monachocrinus, but differing from all the previously known members of the genus in the very wide, low axillaries, and in the slightly concave radials and basals. These two characters taken in connection with the large number of discoidal columnars and the structure of the arms, make the species easily recognizable. As the genus is known from both the Atlantic and Indian Oceans, its occurrence off South Africa is quite natural. The Echinoderm Fauna of South Africa. 231 COMASTERIDAE. COMINIA OCCIDENTALIS. A. H. Clark, 1915. Deutsche Siid-Polar Exp.: Zool., vol. 8, p. 164; pl. 10. This little comatulid is a most perplexing form, for while the general appearance is quite like a comasterid, the central mouth and excentric anal tube combined with the absence of terminal combs on the basal pinnules completely conceal the family relation- ship. Mr. Clark in his original description says: “Die Zihne des Endkammes sind so wenig entwickelt dass sie bei gewohnlicher Untersuchung nicht auffallen.” I have failed to detect the combs even with the aid of a magnification of 70 diameters, in either alcoholic or dry material. It is true that with high magnification, on dry pinnules an uneven margin can be found at the tip, but it is not enough to consider even as a rudimentary comb. In view of this absence of combs and the central position of the mouth, it is hard to see why this species should be considered one of the Co- masteridae but in deference to Mr. A.H. Clark’s much wider expe- rience and greater knowledge of the group, I leave it where he has placed it. Some of the Pirrer FAURE specimens are a little larger than those of the Gauss and there are some trivial differences. The cirri are about XL, 416-18, and the longer ones measure 12-15 mm. The dorsal interradial perisome has calcareous plates more or less abun- dant but it is not “heavily plated”. The three lower pairs of pinnules are approximately equal. Genital glands occur out as far as the twentieth pinnule. The color is yellow-brown with no trace of olive. The Gauss specimens were taken in False Bay (west side, Simon’s Bay) while those of the Pirrer Faure, it is imteresting to note, were collected well up on the Atlantic coast of Cape Colony, P.F. 14905. Saldanha Bay, Cape Colony, 10-14 fms. Sand and mussel-beds. 46 specimens. CoMANTHUS WAHLBERGII. Plate VII. Hiosges Alecto wahlbergii J. Miller, 1843. Arch. f. Naturg., Jahrg. 9, vol. 1, p.181. Comanthus wahlbergit A. H. Clark, 1914. Proc. U.S. Nat. Mus., vol. 40, Dee a We Actinometra parvicirra Bell, 1905. Mar. Inv. South Africa, vol. 4, p. 441. It is interesting to note, although the fact may not be of any significance, that the distribution of this characteristically South 232 Annals of the South African Museum. African species is from Simon’s Bay, eastward to the Tugela River, Natal, while the preceding species seems to range rather from Simon’s Bay westward and northward. This apparent difference of distribution may however be quite unreal and due only to our present ignorance. P.F. 18282. Simon’s Bay, False Bay, Cape Colony, 8-10 fms Rocks. 3 specimens. MARIAMETRIDAE, LIPAROMETRA MULTICIRRA,* Sp. nov. Plate Vis hig 2: Disk about 23 mm. across, very deeply incised; arms about 85 mm. long but they are not quite equal and some scarcely exceed 75 mm. Disk membrane full of crowded, small, calcareous plates. Centro-dorsal large, thick, dorsally flat or a little concave, 6 mm. in diameter; bare dorsal area, nearly 4 mm. across. Cirri XLIII, 30-36, cylindrical at base, but compressed distally; the segments 7-10 have the length about equal to or even a little exceeding the diameter, but elsewhere the greatest diameter exceeds the length; beginning usually with the tenth or eleventh segment, but on some cirrl further out, there is a median, dorsal elevation, at first rather blunt but soon with a short compressed tip or even a sharp point; on the last segment this becomes an opposing claw as long as half the diameter of the segment; terminal claw longer than last segment, very sharp. Arms about 50, all but two broken and detached from disk at or near base; arm-segments numerous, exceeding 150, the distal ones being quite short. Division series all 2, well-separated, rounded and smooth. First syzygy between brachials 3 and 4 of the free arm; second syzygy far out, usually after an interval of more than 20 segments and often 30-40, rarely before segment 20; subsequent syzygies few and at very wide intervals. Low and relatively incon- spicuous synarthrial tubercles occur on all the division series. Lower pinnules not noticeably larger on outer side of arm than on inner. P, (P. similar) about 9-10 mm. long, consisting of 17-21 segments, all but the basal three longer than wide and all but the basal five or six, cylindrical. P, and Py very similar but noticeably larger, 12-13 mm. long, with 24-26 segments. P, and P, similar to P, and approximately equal, or a little smaller and with 1-3 fewer segments. P, and Py, distinctly smaller, about equal to P,. * Multicirrus = haying many cirri. The Echinoderm Fauna of South Africa. 233 Succeeding pimnules somewhat smaller, about 7 mm. long. All the basal pinnules are moderately stout at base but taper to a slightly flagellate tip, which is not however very slender. Colour, pale fawn with the oral surface of disk and arms very dark brown, almost black; margins of food grooves on disk, black. P.F. 12157. Durnford Point, Zululand, N.W. 3/, W., 42 miles. 90 fms. Broken shells. 1 specimen. Holotype, South African Museum, No. A 6435. It is with no little hesitation that I put this fine new comatulid in the genus Liparometra, but as P, and Ps, are of approximately equal size, it seems to me clear that it cannot be placed in either Dichrometra or Lamprometra, as those genera are diagnosed by their founder, Mr. Austin H. Clark. I am somewhat inclined to question the desirability of recognizing these three very closely allied genera, but here again I must defer to the much wider experience of my friend. -The present species is, I think, quite distinct from any pre- viously known form, as the large number of arms and cirri, with their numerous segments, are quite characteristic. The few and widely spaced syzygies is also a noticeable feature. TROPIOMETRIDAE. TROPIOMETRA CARINATA. Comatula carinata, Lamarck, 1816. Anim. s. Vert., vol. 2, p. 534. Tropiometra carinata, A. H. Clark, 1907. Smithson. Misc. Coll., vol. 50, p-. 349. Antedon capensis, Bell, 4905. Mar. Inv. South Africa, vol. 4, p. 139; pl. 2. The distribution of this species is of considerable interest. It ranges from the Seychelles, Réunion, Mauritius and Zanzibar southward to the Cape of Good Hope and thence northwestward to St. Helena, Brazil and the southernmost West Indies. It is true that Mr. A. H. Clark considers the specimens from the latter regions specifically distinct from those taken on the east coast of Africa, but a prolonged comparison of specimens from Tobago, B. W.1L, with individuals of the same size from Zanzibar has satisfied me that the supposed differ- ences do not exist. The specimens from the South African Museum are not notable, except that the smallest (12405-c) has the arms only 20 mm. long, and, like specimens from Tobago of a similar age, the colors are pale yellow and pink-purple. . Mozambique; low tide. Nov. 1912. K. H. Barnard coll. 2 specimens, 234 Annals of the South African Museum. Delagoa Bay, Portugese East Africa. Oct. 1912. K. H. Barnard coll. 1 specimen. P.F. 12405-c. Itongazi River, Natal, N.W. %/, W., 3 miles. 25 fms. Sand and stones. 4 young specimen. | THALASSOMETRIDAE. * CROTALOMETRA MAGNICIRRA. Antedon magnicirra, Bell, 1905. Mar. Inv. South Africa, vol. 4, p.141; pl. 4. Crotalometra magnicirra, A. H. Clark, 1909, Proc. Biol. Soc. Washington, vol. 22, p. 80. This species is not now in the South African Museum, the original specimens having apparently all been retained at the British Museum. It was taken in 300-450 fms., 45-20 miles off the coast of Cape Colony, near Kast London. PACHYLOMETRA SCLATERI. Antedon sclateri, Bell, 1905. Mar. Inv. South Africa, vol. 4, p. 140; pl. 3. Pachylometra sclateri, A. H. Clark, 1909. Proc. Biol. Soc. Washington, vol. 22, p. 21. This characteristic species is represented in the present collection by an armless adult specimen and a number of quite young indivi- duals. The latter were rather puzzling owing to the small centro- dorsal and the relatively long I Br series, and the presence in every case of just ten arms. On the other hand, the cirri are essentially like those of the adult (XV-XVI, 15-17) and the I Br series and lower brachials are distinctly wall-sided and in close apposition. The radials are conspicuous, the height beimg equal to half the breadth, while in the adult specimen they are not only completely concealed but even | Br, is barely visible. The adult specimen has the calyx about 12 mm. in diameter and the cirri 18-20 mm. long, while the young ones are only 2 mm. in diameter through calyx and the cirri are but 4-6 mm. long. It is to be regretted that the condition of the adult does not permit of a full description for Bell’s account is utterly inadequate. In the young specimens, P, is stiff, erect with 7 seg- ments and P, is similar. P, is a little longer, with 9 segments; Py, the same. Ps; (and P.) is a little longer, with 44 segments and is more flagellate at the tip. Subsequent pinnules are shorter. P.F. 12872. East London, Cape Colony, N. 15 miles. 340 fms. Mud. 4 adult specimen, with arms all broken off. * Those species marked with an asterisk are not represented in the South African Museum collections. The Echinoderm Fauna of South Africa. 235 P.F. 12884. East London, Cape Colony, N. 15 miles. 340 fms. Mud. 1 young specimen. P.F. 43227. Cove Rock, near East London, N.W. 3), W., 43 miles, 80-130 fms. Coral rock. 413 young specimens. PENTAMETROCRINIDAE. PENTAMETROCRINUS VARIANS. Eudiocrinus varians, P. H. Carpenter, 1882. Jour. Linn. Soc., Zool., vol. 16, p. 496. 1888, CHALLENGER Comatulae, pl. VIII, figs. 3-7. Pentametrocrinus varians, A. H, Clark, 14908. Proc. Biol. Soc. Washington, vol. 21, p. 435. Although this specimen lacks cirri and has all the arms broken, it is so similar to specimens of varians from southern Japan, with which I have compared it, that I do not doubt it belongs to that species. The nearest point to South Africa at which the species has previously been taken is near the Andaman Islands in the north- eastern Indian Ocean. P.F. 17351. Cape Point, N. 86° E., 43 miles. 900-1000 fms. Grey mud. 1 specimen. SEA-STARS. ASTEROIDEA. * The sea-stars form a very large and important part of the South African Echinoderm fauna. They were listed in 1910 by Déderlein (Schultze’s Zool. Anthrop. Ergeb. Forschungsr. Sudafrica, vol. 4, pt. 1, p. 246) but he did not include species occurring only at depths over 278 fms. (500 m.) nor did he extend the South African region to include Mozambique. His list includes 30 species, but two are synonymous (Astropecten capensis and pontoporueus) and one (1.e. Sladen’s record of Asterina gunnic) is probably due to a mistaken identification or a misplaced label. The collection sent me from the South African Museum contains 51 species but of these only 14 are in Déderlein’s list. There are however 9 additional species pre- viously recorded from Mozambique or from deep water off South Africa and hence not listed by Diéderlein which fall within the scope of the present report, which thus includes 74 species; 18 seem to be new to science and are here described for the first time. * After this section was ready for the press 1 had the pleasure of a visit from Dr. W. K. Fisher, the well-known authority on sea-stars, who very kindly ex- amined many of the specimens and permitted me to profit by his wide knowledge and sound judgement. For this help I beg to offer him herewith my best thanks. 236 Annals of the South African Museum. Of these 74, 35 are truly littoral occurring in water less than twenty fathoms deep, while 9 are strictly abyssal, occurring only in depths beyond 600 fms. The remaining 30 species may be classed as continental. Of the 35 littoral species, 143 are endemic so far as our present knowledge goes; as 10 of these have been known for a considerable time and have not yet been reported from elsewhere, it is probable that they are truly characteristic forms. Of the remaining 22, 18 are East Indian or Indian Ocean species, of which 15 were previosuly known from the east coast of Africa, north of Mozambique. There are two littoral species (Asterina calcarata, Henricia ornata) which occur on the shores of the southern end of South America, but both these cases require further investigation; each belongs in a genus in which specific limits are ill-defined. There are also two littoral species known from the coasts of southern Australasia; one of these ( Coscinasterias calamaria) is a well-defined species and its occurrence at Mauritius has long been known, it is unquestionably a valid link between the littoral faunas of Australasia and Africa; the other species however is the dubious Henricia ornata, a name under which several species are perhaps involved. Of the two remaining South African littoral species, one is the rare and little known Culcita veneris, originally from St. Pauls Island, southern Indian Ocean, and since recorded by Bell only, from Cape Colony; the other is the northern starfish, Marthasterias glacialis, whose occurrence at the Cape no longer admits of doubt. It is of importance to note that 12. of the littoral seastars here listed as South African, are not known from south of Mozambique and there are two or three others whose occurrence south of that point is known from only a single record. Of the 30 Continental species, 20 appear to be endemic, but 12 of these are here described as new and may later be found elsewhere. Nevertheless the Continental fauna is very characteristic for in addition to the endemic forms, three are known only from the Kerguelen region. There are three species hitherto known from the Atlantic, two from the East Indian region and one from Australia. The thirtieth, one of the most remarkable members of the Conti- nental fauna is Ceramaster patagonicus, which occurs not only in South American waters but along the Pacific coast of North America to the region of the Commander Islands in Bering Sea. One of the Atlantic members of this fauna (Diplopteraster multipes) has an equally remarkable range, as it occurs in the North Atlantic from about 35° North to Barents Sea and Norway and in the North Pacific The Echinoderm Fauna of South Africa. 937 from San Diego, California, and Suruga Gulf, Japan, to Bering Sea. Of the 9 abyssal species occurring in the present list, only 3 are endemic, while 5 are already known from the deep waters of the Atlantic and one is Antarctic. The endemic species are all new to science. It is of interest to note that none of the abyssal species seems to have come from the east, whereas the littoral fauna has nearly all come from the Indian Ocean. In conclusion then, we may say that so far as our present know- ledge goes the sea-star fauna of South Africa is highly characteristic. Nearly half (56) of the species are endemic and several others occur only in the region of Kerguelen or St. Pauls Island. Of the non- endemic forms, 20 are from the Indo-Pacific region and 410 from the Atlantic, while the remainder are Australian or South American. The affinities of the littoral fauna are distinctly Indo-Pacific, but if the tropical species, not known from south of Mozambique, are left out of account, it is evident that most of the littoral starfishes of South Africa have become specifically differentiated. On the other hand the continental and abyssal faunas, while perhaps equally well differentiated and as characteristic, have slight East Indian but rather strong Atlantic affinities. The impression made by the study of the South African sea-stars is that the shallow-water forms are of Indian origin and the deeper-water forms are from the Atlantic. There is very little evidence of an Australian or South American influence in the composition of the fauna. It is true that Coseinas- terias calamaria is a characteristic Australian species, but it seems to be very rare in South African waters. As already stated no reli- ance can be placed on evidence offered by such forms as Henricia ornata and Asterina calcarata. The occurrence of the characteristically Antarctic genus Cryaster in Algoa Bay is worthy of more than pas- sing notice, since the entire family is otherwise unknown outside of the Antarctic region. The 74 species included in this report belong to no fewer than 16 families. They can be most easily recognized if these families are first differentiated from each other, Under each family will be found the necessary key to the species included in it, which occur in South African waters. Key to the South African Families of Asteroidea. Marginal plates large, defining the contour of the body; abactinal skeleton never reticulate or imbricated but made up of plates, which often bear paxillae or granule-bearing tabulae. 16 238 Annals of the South African Museum. Cribriform organs * present in each interradius . . Porcellanasteridae. No cribriform organs. Marginal plates very spiny, more or less alternate; papulae restricted to special areas at base of rays . ; : Benthopectinidae. Marginal plates opposite; papulae nat restricted to special areas at base of rays. Abactinal surface covered with paxillae. Superomarginal plates well developed j Astropectinidae. Superomarginal plates aborted ‘ A . Lududae. Abactinal surface not covered with paxillae. Disk large with big actinal interradial areas, but no actinal papulae. Marginal plates large and conspicuous; disk more or less flat; papulae single or a few together Gonrasteridae. Marginale plates not conspicuous; disk elevated or at least very thick; papulae numerous in large groups. Marginal plates large; abactinal skeleton more or less conspicuous. : : . Oreasteridae. Marginal plates small and with abactinal skeleton covered and concealed by a thick skin . . Poranudae. Disk small with very small actinal interradial areas, or if the latter are well developed there are actinal papulae; marginal plates small; tegumentary developments, granulate (rarely wanting) Ophidiasteridae. Marginal plates small or wanting; abactinal skeleton more or less imbricated or reticulate. Disk not circular and sharply set off from long, more or less terete, and readily detachable arms; marginal plates small but regularly present (except Cryasteridae). Pedicellariae rare or wanting, never pedunculate forcipiform; ambulacral ossicles rarely crowded; pedicels usually in two series. Oral plates rather small, not shovel- or plowshare-shaped; ambu- lacral furrows narrow. Marginal plates conspicuous; actinal plates regularly radiatingly arranged f : : : : . Ganerudae. Marginal plates quite inconspicuous. Abactinal skeleton formed of closely imbricated plates, bearing very small spinelets . : . Asterinidae. - Abactinal skeleton not abaeias Abactinal skeleton more or less reticulate Echanasteridae. Abactinal skeleton entirely aborted Cryasteridae. Oral plates big and shovel- or plowshare-shaped; ambulacral furrows wide. Abactinal skeleton with paxillae or pseudopaxillae, not concealed by a supradorsal membrane ; ‘ . Solasteridae. * Technical terms used in this or subsequent keys are fully explained and Illustrated in Sladen’s CHALLENGER report (1889) or in Fisher's North Pacific Asteroids (Bull. 76 U.S. Nat. Mus, 1911). The Echinoderm Fauna of South Africa. 239 Abactinal skeleton with paxillae concealed, more or less, under a remarkable supradorsal membrane é . Pterasteridae. Pedicellariae abundant, especially forcipiform; ambulacral ossicles crowded ; pedicels in four series 3 ; 5 : : Asterudae. Disk circular sharply set off from the long; more or less terete and readily detachable arms; marginal plates microscopic or wanting . Brisingidae. PORCELLANASTERIDAE. This deep water family is represented in the South African region by only a single species. PORCELLANASTER CA&RULEUS. Wyville Thomson, 1877. Voy. Challenger: Atlantic, vol 1, p. 378; Ke a 2 figs, D798: The specimens are all small, with R=7-9 mm. They are too young to make their specific identity certain but comparison with somewhat larger specimens of ce@rulevs, taken by the CHALLENGER and the BLake, indicates that they are immature examples of that species. The only noteworthy differences are the absence of spines on the superomarginal plates and the incomplete calcification of the inter- brachial areas below. Both these however are easily accounted for as evidence of immaturity. On account of the locality, it would be natural to refer these specimens to P. eremicus Sladen but I am myself satisfied that the specimen on which that species is based, is a young ceruleus. P.F. 46905. Cape Point, N.E. by E.!/, E., 40 miles, 800-900 fms. Green mud. 1 specimen; young. P.F. 17351. Cape Point, N. 83° E., 43 miles. 900-1000 fms. Gray mud. 3 specimens; young. BENTHOPECTINIDAE. This family of deep-water starfishes was not known from the South African region hitherto, but the Prerer FAuRE has found two species, each representing an interesting genus. One of these forms was known only from near Kerguelen while the other is a widely distributed Atlantic species. They may be separated from each other by the characters given in the following key, but it is evident that each is somewhat variable and does not conform exactly to a strict specific description. 240 Annals of the South African Museum. Key to the South African Species of Benthopectinidae. Papularium a small, circular elevated area; one large spine on actinal surface of adambulacrals . : : ; ; : Pectinaster filholi. Papularium V-shaped; three fares spines on actinal surface of each adambul- acral. ; : : : : : . Luidiaster hirsutus. PECTINASTER FILHOLI. Perrier, 1885. Ann. Sci. Nat. (6), vol. 19, no. 8, p. 74. Sladen, 1889, CHALLENGER Ast., pl. 8, figs. 3, 4 (as forcipatus). The South African specimens show slight, but obvious differences from a cotype of filholi with which I have compared them, but agree very closely with a cotype and other specimens, from the north- western Atlantic, of Sladen’s Pontaster forcipatus. From the geogra- phical point of view they would naturally, and I think correctly be referred to Sladen’s variety echinata (sic) but Ludwig considers forci- patus a synonym of filholi and after a comparative study of the material in the M.C.Z., I believe he is right. The species has a wide range from near Nova Scotia in the northwest to the vicinity of Marion Island in the southeast, but it is always an abyssal form, ranging from 699 fms. down to 1700. The specimens taken by the PIETER FauRE are of varied size, the smallest havmg R=8 mm. and r==2(R=4r), while the largest has R59 mm. and r= 11 (R=5-4r); the body form is thus assumed very early in life. In spinulation, the smallest specimen is surprisingly like the largest, the only difference of importance being the presence, in the adult, of two spines on many inferomarginal plates. The youngster has only a very minute madreporite, scarcely distinguishable, and the papularia are each represented by a single pore, or two, but in the largest specimen there are only 10-12 pores in each papularium. The number and distribution of the pedicellariae shows great diversity in this species; in the PreTER FAuRE specimens they are rather numerous but are confined to the actinal surface. P.F. 16902. Cape Point, N.E. by E. '/, E., 40 miles. 800-900 fms. Gr. m. 3 specimens; young. P.F, 16905. Same station. 4 specimen; very young. _P.F. 17332. Cape Point, N. 86° E., 43 miles. 900-1000 fms. Gray m. 6 specimens; adult and young. P.F, 17351. Same station. 4 specimen; young. iw) — >» The Echinoderm Fauna of South Africa. LUIDIASTER HIRSUTUS. Studer, 1884. GazELLe Ast., p. 47; pl. 4, figs. 7a—d. This species was originally found northwest of Kerguelen, on sandy bottom, in 130 fms. of water. Its occurrence in South African waters is thus of much interest, though not surprising. The individ- uals before me show a range in size from R = 20 mm. to R = 65 mm. but the growth changes are trivial between these two extremes. In the largest specimen, there are not infrequently 3 large ee on the actinal surface of the adambulacral plates and there are 2 large inframarginal spines. It is remarkable that Ludwig in his hehe useful key to the species of Luidiaster (1910. Sitz. K. Preus, Acad. Wiss. Berlin, p. 453) says of hirsutus ‘“untere Rand- platten mit einem Stachel”, when Studer distinctly says they bear two long spines. Even in the smallest specimen at hand, there are two such spines on the basal inferomarginals. P.F. 18904, 36° 40'S x 24° 26’ E, 200 fms. Gr. s. 3 young specimens. P.F, 18913. Same station. 2 adult specimens. ASTROPECTINIDAE. This family is represented in South African waters by 10 species, most of which are however continental rather than truly littoral forms. One, apparently new species, is distinctly abyssal. They may be distinguished from each other as follows: Key to the South African Species of Astropectinidae. No specialized spines or spinelets on either series of marginal plates Leptychaster kerquelenensis. More or less conspicuous spines or spinelets on inferomarginals and often on superomarginals as well. Actinal interradial areas more or less extensive; madreporic body hidden by paxillae on its surface. Inferomarginals, and often superomarginals also, with single large spinelets. No large spine on actinal surface of adambulacral plates Plutonaster intermedius. A large erect spine on actinal surface of each adambulacral plate, at least distally : : : . Plutonaster proteus. Inferomarginals with a few racrantilonit spinelets on each, none on superomarginals ; c . Dipsacaster sladeni. Actinal interradial areas small; rieidiee ponte body small not hidden by paxillae on its surface. Marginal plates, especially inferior, more or less vertical, at least at base of ray, the vertical height of ray at base being approximately equal to combined height of both series of marginals. LS = bo Annals of the South African Museum. Papillae of marginal plates squamiform, and spinelets short and very flat . 3 . Bathybraster robustus. Papillae of marginal lates not a all squamiform ; spinelets of infero- marginals slender and rather long : . Psilaster acwminatus. Marginal plates, at least inferior, oblique or nearly horizonal; vertical height of ray at base not remarkable. Large sharp spines present on superomarginals Astropecten polyacanthus. Spines on superomarginals small or wanting. Small spinelets on at least some superomarginals. R = 2°5—3-5 r; radial paxillar areas much wider than com- bined marginal plate series . Astropecten pontoporaeus. R= 4—5r; radial paxilar areas narrower than combined marginal series . : Astropecten hemprichi. No superomarginal spinels : Astropecten granulatus. * LEPTYCHASTER KERGUELENENSIS. E.A. Smith, 1876. Ann. Mag. Nat. Hist. (4), vol. 17, p. 410. Sladen, 4889. CHALLENGER Ast., pl. 31, figs. 41, 2. Although Bell (1905, Mar. Inv. South Africa, vol. 3, p. 242) records this starfish from three stations, there are now no specimens in the South African Museum, and [ include it here solely on the strength of Bell’s identification. PLUTONASTER INTERMEDIUS, Goniopecten intermedius Perrier, 1881. Bull. M. C. Z., vol. 9, p. 25. 1884, BLAKE PASE. pla figs. de, 2: Plitonaster intermedius Perrier, 1894. ane et TAL. Ast., p. 016. Comparison of the South African specimens with others from off the east coast of the United States fails to reveal any differences worthy of note. The African specimens are adult, the greater radius being 53-75 mm. P.F. 17394. Cape Point E. !/, N., 34 miles. 500-550 fms. Green mud. 1 specimen; adult P.F. 18110. Cape Point N.E. 3/, N., 46 miles. 760 fms. Green mud, 4 specimen; adult. PLUTONASTER PROTEUS * sp. nov. Plate XIII. Figs. 3-7. R= 58 mm. © = 145 mms RSs Be 1m, ha Disk moderately large, rather flat; rays narrow, flat, tapering, at * Ifomreic = Proteus, in reference to the remarkable change in appearance during growth. The Echinoderm Fauna of South Africa. 243 first very slightly, but distally more abruptly, to a somewhat blunt tip. Disk and rays, within area bounded by superomarginal plates, covered by numerous low, rounded pseudopaxillae, each “10-25 mm. in diameter, the height, little, if any, greater; each carries about ten (6-16) short slender spinelets, some of which form a slightly radiating marginal circle; these are rather longer than those within it and the latter may be scarcely more than rounded granules; the pseudopaxillae show no regular arrangement. Papulae small, single, numerous. Madreporic body large, nearly 4 mm. across, concealed under some 14 pseudopaxillae of varied size; the outer margin of the madreporite is less than 5 mm. from the inner margin of the superomarginal plates. The latter are 29 in number on each side of the ray; interradially they are nearly square but conspicuously swollen or elevated at center; distally they soon become longer than wide and less swollen and on the distal half of the arm they are scarcely swollen at all but are evidently wider than long; each plate bears a single large spinelet, which is, in the interradial regions, 4 mm. high and basally ‘5 mm. in diameter and occupies the center of the plate but becomes smaller and smaller distally and is placed more and more near the outer (lower) edge of the plate; on most of the proximal plates a second, but much smaller spinelet occurs on the inner (upper) margin of the plate; the rest of the surface of each superomarginal is covered by a fairly uniform but well-spaced coat of low spinelets or spiniform granules, longest and most numerous along the lateral margins of the plates. Terminal plate moderately large; it has all its spinelets rubbed off in the only instance where the plate itself is not missing. Inferomarginals almost exactly like the superomarginals in_ all particulars, except that the large spinelets are rather longer, and the second spinelet on the inner edge of the interradial plates is larger and so is quite conspicuous. The two series of marginal plates form a vertical wall for each side of the ray, about equally in evidence above and below; the fasciolar channels between the plates are moderately developed more particularly in the imterradii. Actinal intermediate plates wanting at tip of ray and indeed on the entire distal half; the first one adjoins the sixth inferomarginal (counting from interradial line) and there are rather more than a dozen, lying next to the adambulacrals, between that point and the oral plate; a second series begins at the fourth marginal and contains nine or ten plates; some 25-30 smaller plates fill up, more or less irregu- larly, the remainder of the notably small actinal interradial area; all the intermediate plates are covered, but not very thickly, with 244 Annals of the South African Museum. short, well-spaced, rough spinules; a few of these are enlarged here and there into short, thick spinelets and rarely a little group make up a pedicellaria of a rudimentary sort. Adambulacrals about 37 on each side of the furrow; except the first two or three and the distalmost half dozen, they are longer than wide; furrow-margin of each plate with about 8 conspicuous spinelets, the middle ones 1°5 mm, long, the adoral one shortest; outside this series, on the actinal surface of the plate, near its distal margin, is a single large spinelet, nearly equal to those on the marginal plates; the rest of the surface of each adambulacral plate is sparsely covered by spinelets like those on the actinal inter- mediate plates. Oral plates rather large, swollen; each bears a marginal series of a dozen spinelets, of which the first (inner) two are the largest (about 2 mm. long), the others being gradually smaller; surface of plate rather thickly covered with spinelets, of which those near the interradial margin are largest, particularly. those at inner end of plate. Color, dull brownish-yellow, in the present condition, dried from alcohol. Cutting through and laying back the skin of one ray reveals large double ampullae, the complete absence of dorsal muscle bands, and the genital glands confined to the interradial regions. Seen from within the plates of the dorsal skeleton are circular and isolated, but seemingly more crowded along the sides of the ray. P.F. 16743. Cape Point, N.E. by E. 3/, E., 38 miles. 755 fms. Gr. m. 14 specimen; adult. P.F. 16902. Cape Point, N.E. by E. 3/, E., 40 miles. 800-900 fms. Gr. m. 5 specimens; young. P.F. 16931. Same station. 1 specimen; adult. P.F. 16944. Same station. 4 specimens; young. P.F. 17351. Cape Point, N. 86° E., 43 miles. 900-1000 fms. Grey m. 2 specimens; young. Holotype, South African Museum no. A 6427, P.F. 16934. This species is undoubtedly near to P. bifrons (Wyv. Th.) but it is at once distinguished from that species by the absence of large spinelets on the actinal interradial areas and the presence of a second series of spines on both sets of marginal plates in the inter- radii. Another very marked difference is that in very young bifrons, the infero marginal spines are well developed while in much larger specimens of proteus, they are lacking or just beginning to appear. The series before me affords opportunity for a very interesting study of growth changes which are of more than ordinary interest The Echinoderm Fauna of South Africa. 245 in this species. The smallest specimen (R = 6 mm.) hasR= 1‘5r and is thus somewhat pentagonal with deeply concave sides; the pseudopaxillae are similar to those of the adult and the madreporite is completely concealed; there are 8 marginal plates in each series, on each side of a ray and they are quite uniformly covered with minute rough spinules; the actinal interradial plates are few and covered like the marginals. In two particulars this youngster is quite different from the adult; the large spimes and spinelets of the adambulacral and marginal plates are wanting and there are distinct, though simple, pedicellariae on the abactinal surface and between the marginal plates, as well as actinally. On the most interradial of the inferomarginals, one spinule is distinctly larger than the others and may be considered the first indication of the spine, later so prominent. The terminal plate of each ray is relatively very large; on each side of the tip, near the oral surface is a large spinelet and back of this (orally) are two smaller spinelets. The next larger specimens have R = 7:'5 mm., r = 3:5; hence R=21r. The spinulation of these individuals is exactly like that of the smallest, except that on some of the distal adambulacral plates, one of the actinal spinelets is noticeably bigger than the others; pedicellariae are very noticeable, especially among the marginal plates. A specimen with R — 40 mm. is not essentially different in any way. A specimen with R=12°5 mm. and r= 45mm. (R= 2°75 r) has the large spinelet indicated on most of the inferomarginal plates, quite distinct on nearly all the adambulacrals, and evident on the interradial superomarginals; there are no pedicellariae except on the actinal interradial areas. A specimen from the same station as this one, with R=13 mm. and r=5'5 mm. (R = 2°377r) has distinctly wider rays and there are many pedicellariae, chiefly of two spinelets, all over the abactinal surface; large spinelets are indicated only on the interradial inferomarginals and doubtfully on a few distal adam- bulacrals. The largest of the young individuals, from the same Station as. the holotype, has R= 419 mm. and, — 7) fy — 2:77: there are no pedicellariae, the large spinelets of the inferomarginals are conspicuous while those of the superomarginals are evident; the proximal adambulacrals show no large actinal spinelet but on all of those on the distal half of the arm it is perfectly distinct. The adult specimen from 16743 has R= 48 mm., r= 14 and hence R=35r but in only one other particular does it show any notable difference from the holotype; there is no second large spinelet on any margi- nal plate. To sum up the growth changes of this species then we may say 246 Annals of the South African Museum, that it changes from a nearly pentagonal form, uniformly covered with pseudopaxillae and minute rough spinules, with no large spine- lets whatever, into a stellate form with moderately long rays, having conspicuous spinelets on all adambulacral and marginal plates. During this change pedicellariae are wholly lost, at least abactinally. It is worthy of special note that the large spinelet of the adambula- cral plates appears first on the distal part of the ray and occurs proximally only after the individual is half grown. The second set of spinelets on the marginal plates appears only im what is apparently the fully grown individual. DIPSACASTER SLADENI. Alcock, 1893. Ann. Mag. Nat. Hist. (6), vol. 14, p.87; pl.5,. figs. 3, 4. These specimens answer so well to Alcock’s description that I feel satisfied they should be referred to sladeni, but in two particulars they are different; the pedicels of the paxillae are certainly not “long, slender’, as I understand those terms, and the adambulacral spines are not what I should call ‘needle-like’. Such terms ought not however to be construed too rigidly. The adambulacral armature of the South African specimens is almost exactly like that of laetmophilus Fisher, and the only point in Fisher’s description to which the present specimens do not answer is the covermg of the inferomarginal plates, in describing which Fisher uses the word ‘‘squamiform”, There is nothing ‘‘squamiform” in the spinelets covering the inferomarginals of the African specimens. Comparison of Fisher’s description and figures of laetmophilus with Alcock’s of sladeni certainly suggests the identity of the two, but oddly enough Fisher makes no reference whatever to sladeni. * The present series reveals some very interesting growth changes in this starfish. The smallest specimens have R= 15 mm, and r =7, while the rays are nearly 10 mm. across at their very base; thus R = 2r and about 15br. A somewhat larger specimen has R—26 mm... 711 and 6,—43: thus Ri = 2367 andead or. ne next larger specimen has R = 45 mm., r = 17 and br = 18; thus * After critical examination of the South African specimens of sladeni, Fisher finds at least half a dozen differences between them and laetmopnilus. Of these the most obvious, if not the most important, is in the spinelets of the infero- marginals, which are distinctly syuamiform in the Alaskan species and spiculiform in the African. Other important differences are to be found in the form of the inferomarginals, in the plates and fasciolar channels of the actinal intermediate areas, and in the mouth plates. The two species, although nearly allied, seem to be perfectly distinct. —_ a | The Echinoderm Fauna of South Africa. 24 R=2:65r and 2‘56r. In the largest specimen, R = 78 mm., r= 26 and br = 28; thus R=3r and 2-78 br. It is thus obvious that the larger the specimen of this species, the longer and proportionately narrower are the rays. The number of superomarginal plates on each side of a ray in these four specimens is 16, 22, 28 and 39 respectively. The number of marginals is relatively greater there- fore in proportion to the length of the ray in young specimens than in adults; thus, while the length of ray imereases five times the number of marginals is increased only two and a half times. It will be noticed that the largest African specimen has several more supero- marginal plates than much larger specimens of sladeni and laetmophilus, but I think this is merely a matter of individual, or possibly, geo- graphical variation. In the smallest specimen, the enlarged spinules on the outer ends of the inferomarginal plates are barely recognizable and then only in the interbrachial arcs. They are more pronounced but are not at all conspicuous in the specimen with R=26 mm. The adambu- lacral armature shows very little change with growth; in the smallest specimen there are 5 and often 6 adambulacral spines and they are somewhat compressed, especially near base; in the largest specimen, there are 7, occasionally 8, adambulacral spines and they are mar- kedly compressed at base. Colour in life: upper surface reddish orange, lower surface pale. P.F, 2285. Lion’s Head, Cape Town, N. 67° E., 25 miles. 131-136 fms. Black specks. 2 specimens; adult. P.F. 2330. Same station. 2 specimens; young. P.F. 2798. Vasco de Gama‘ Peak, N. 74° E., 18 miles. 230 fms. St. 1 specimen; adult. P.F. 47604, Cape Point, E. by N., 30 miles. 345 fms. Green sand and mud, 4 specimens; very young. BATHYBIASTER ROBUSTUS. Archaster robustus Verrill, 14884. Amer. Journ. Sci. (3), vol. 28, p. 383. Bathybiaster robustus Verrill, 1894. Proc. U.S. Nat. Mus., vol. 417, p. 256. These specimens range in size from R = 7 mm. to R= 80mm. The largest has been critically compared with similar specimens of robustus from off the Eastern coast of the United States and there is no doubt of their identity. The growth changes of this species are very interesting. Small individuals were described by Sladen (1889, CHALLENGER Ast., p. 236, pl. 40, figs. 3-6) as Phoraster pumilus, supposedly representing a new genus, distinguished from Aathybiaster 248 Annals of the South African Museum. by the presence of an epiproctal cone and the absence of pedicellariae, but Verrill has shown that both these features are youthful and quite unreliable. In the present series, there is no epiproctal cone in the large specimen, but it is obvious in all the small ones; it is however smailest in the smallest specimen (1 mm. high) and largest (3 mm.) in a specimen with R = 17 mm. Apparently therefore it reaches its fullest development in late youth and then disappears, but is still evident in specimens one-third grown. The terminal plate is but very little larger in the big specimen than in the smallest and has entirely lost the three conspicuous spines which it bears in youth. The adambulacral plates are relatively considerably longer in the adult but the adambulacral armature changes but little, as there are 3 spines in the smallest specimen and only 5 in the big one. There is no indication of a superomarginal spinelet in the smaller specimens but in the largest it is evident on a dozen plates or more in each series; it is however remarkably low and squamiform. P.F. 16742. Cape Point N. E. x KE. 3), E., 38 miles. 755 fms. Green mud. 1 specimen; adult. P.F. 16902. Cape Point N. E. x E. 3/, E., 40 miles. 800-900 fms. Green mud. 2 specimens; young. P.F. 17351. Cape Point N. 83° E., 43 miles. 900-1000 fms. Green mud. 2 specimens; young. PSILASTER ACUMINATUS. Sladen, 1889. CHALLENGER Ast., p. 225; pl. 40, figs 4, 2. It is not without some hesitation that I refer these specimens to Sladen’s species, for in one particular they are very different from his description. He says the marginal plates are more or less bare (lower part of superomarginals, upper part of inferomarginals) and covered by a membrane, while in the African specimens, papillae cover the plates; along the margins the papillae are slender but on the surface of the plates they are quite squamiform. In one specimen, the lower portion of the largest superomarginals is only sparsely covered with papillae so perhaps if the specimens were larger these plates would be bare. But in these specimens, R = 60 mm. + and Sladen’s type had R only 65 mm. Another difficulty is that these specimens are so unlike a much larger Psilaster from Australia which in my ENDEAvouR report I have called acwminatus, that it is hard to believe they are the same species. Sladen however called attention to differences between the African, Australian and New Zealand specimens of the CHALLENGER The Echinoderm Fauna of South Africa. 249 collection, but he felt that more material was necessary before it could be conclusively determined whether all were the same species or not. I certainly have not sufficient available material to enable me to satisfy myself in the matter, so I follow Sladen’s example and let all remain under the name which he gave. P.F. 2330. Lions Head, Cape Town, N. 67° E., 25 miles. 431-136 fms. Black specks. 1 specimen; adult ? P.F. 14976. Lions Head, Cape Town, S.E. !/, E., 47 miles. 175 fms. Green sand. 4 specimens; adult ? ASTROPECTEN POLYACANTHUS. Miller and Troschel, 1842. Syst. Ast., p. 69. The occurrence of this species south of Zanzibar is noteworthy and its presence on the coast of Natal is really remarkable. The present specimen (R = 70 mm.) though the rays are somewhat broken, is in admirable condition for study. The superomarginal spines are unusually small and slender, the largest (those on the imterradial pair of plates) being less than 3 mm. high and about two-thirds of a millimeter in diameter at base. The paxillae bear many spinelets ; those on the convex surface are very low and papilliform while those on the margin are relatively long and slender. The oral surface is much less spiny than in typical examples of polyacanthus, this ap- pearance being due to the somewhat squamiform spinelets and the absence of large spines on the adambulacral end of the inferomarginal plates. The species is so widespread and so diversified that local races will probably be recognized ultimately, and when that is done the South African form will probably be given a subspecific name. The more typical form is well figured by Savigny, 1803. Pl. d@’Ech. Egypte, pl. 4, fig. 4. ; P.F, 12516. Off Umhlanga River, Natal, 2'/. miles. 22-26 fms. Fine sand. 4 specimen; adult. Delagoa Bay. K. H. Barnard. ASTROPECTEN PONTOPORAUS. Sladen, 1883. Jour. Linn. Soc. Zool., vol. 47, p. 259. 1889, CHALLENGER, Ast., pl. 35, figs. 4, 2. Astropecten capensis Studer, 1884. GAzELLE Ast., p. 44. The present specimens (R = about 35 mm.) are a trifle smaller than Studer’s but they leave no doubt in my mind as to the identity of capensis and pontoporeus. The differences mentioned by Studer are trivial. The relatively longer arms in Sladen’s specimens are 250 Annals of the South African Museum. due to their larger size, while thé degree of projection of the infero- marginal plates and the exact form of their spines is a matter of in- dividual diversity. Bell (1905, Mar. Inv. South Africa, vol. 3, p. 243) records pontoporeus from 21 stations and capensis from one, but he does not hint at the means by which he distinguished them. P.F. 15835. Cape Poimt, N.W. 5 miles. 47 fuss. Sand and rocks. 4 specimens; adult ? * ASTROPECTEN HEMPRICHII. Miller and Troschel, 1842. Syst. Ast., p. 74. De Loriol, 1885, Cat. Rais. Ech. Mauritius: Stellérides, pl. 24, figs. 7-8. This species is reported by Peters (1852) from Inhambane, P.E.A. | y , and by Bell (1884) from Mozambique. I have not myself seen specimens from the African coast, south of Zanzibar. ASTROPECTEN GRANULATUS, Miiller and Troschel, 1842. Sys. Ast., p. 75. Ddéderlein, 1896. Jena Denksch., vol. 8, lief. 3, pl. 18, figs. 30, 30a. These South African specimens were at first identified with mona- canthus Sladen but in the larger specimens the paxillae always show several to many central granules, and Sladen emphasizes the single central granule as an important species character. Koehler however has stated that the number of central granules on the paxillae is ¢ matter of age and examination of these specimens satisfies me that he is correct. Careful study of his text and figures, and those of Déderlein, with Sladen’s, convinces me that monacanthus is identical with granulatus. The only point on which I am doubtful is the coloration, some specimens (none from South Africa however) showing a conspicuous mottling of the upper surface. This mottled form is figured by Sladen as granulatus and Koehler says his specimen from the Aru Islands is exactly like it in color. On the other hand he says his specimen is identical with that figured by Déderlein from Torres Strait and Déderlein’s specimen is unicolorous. Probably the coloration is more or less subject to individual diversity. The length of the superomarginal plates and the extent to which they occupy the dorsal surface of the arms is a matter of age; they are longest and dorsally most conspicuous in the smallest individuals before me (R75 mm); they are relatively shortest and Jeast visible from above in the largest specimens (R= ¥8 mm). These large specimens are just the size of Koehler’s from the Aru Islands, and considerably larger than those seen by Sladen and Doéderlein, but they are smaller The Echinoderm Fauna of South Africa. 251 than some which Koehler has had from India. One of the smaller African specimens shows six superomarginal spines and similar spines occur in one of the larger specimens of the Indian Museum. It is interesting to note that the proportion of R to 7 is practically the same in the smallest and largest specimens, namely R=4r, but the arms are broadest in the smallest specimens, are 2:7 br; in the large individuals, R = 3°5 br. In spite of a deficiency of material which is much to be regretted, I think we may say then that granulatus is a small species of Astropecten with unarmed superomarginal plates, which ranges from India to South Africa on the west and to Torres strait on the east. One of the specimens here referred to granulatus (18904) may perhaps represent .a different species. The colour is a noticeably deeper brown, there are usually two and often three infero-marginal spines, and the spinules everywhere, but especially on the oral surface, appear to be more or less sacculate. This imdividual is obviously immature (R = 19 mm.) and comes from deeper water than the others, so that the probability of its not being granulatus is rather strong. P.F. 10975. Tongaat River, Natal, N. W. by N. 4/, N., 5 miles. 36 fms, Sand and rocks. 2 specimens, very young. P.F. 12516. Off Umhlanga River, Natal, 2!/, miles. 22-26 fms. Fine sand. 9 specimens; adult? and young. P.F. 18904. Cape Agulhas, Cape Colony, N. W. 175 miles (36° 407 S., 24° 26’ E.). 200 fms. Green sand. 4 specimen, Young and dubious. LUIDIIDAE. It is not certain whether two or three species of this family are found on the coast of South Africa, but it is likely that at least three occur and not improbable that others will be found when the marine fauna is better known. The species recorded from the region may be distinguished from each other as follows; Key to the South African Species of Luidiidae. Rays 5; no enlarged central spinelet on paxillae : Inidia africana. Rays 7 or more. No enlarged central spinelet on any paxillae; iatter with quadrate tabulum Inuidia maculata. An enlarged central spinelet on many paxillae; latter with a stellate crown TInudia savignyi. 252 Annals of the South African Museum. LUIDIA AFRICANA. Sladen, 1889. CHALLENGER Ast., p. 256; pl. 44, figs. 1 and 2. [ have not seen this species but Sladen records it from Simon’s Bay, Cape of Good Hope and Bell lists it from four stations in 85—90 fms. - * LUIDIA MACULATA. Miller and Troschel, 1842. Sys. Ast., p. 77. H. L. Clark, 1916. EnpEavour Kch., pl. 5. This species is recorded by Peters from Mozambique (4852, Monatsb. Berlm Akad., p. 178) but de Loriol thinks he probably had L. savignyi. While this is quite possible, it does not seem to me unlikely that maculata occurs as far south as Mozambique and I therefore let Peter’s record stand. LUIDIA SAVIGNYI. Asterias savignyi Andouin, 1826. Expl. som. des pls. Echinod. de Egypte pub. par Savigny, p. 208; Rayonnés, pl. 3. Luidia savignyi Gray, 1840. Ann. Mag. Nat. Hist. (1), vol. 6, p. 183. This fine ZLaidia, origimally noted from the Red Sea, was known only as far south as Mauritius and Zanzibar. In the PreETER FAURE collection however, I find a badly broken specimen with R = 170 mm., which is undoubtedly this species, thus greatly extending the known range to the southward. It would be interesting to know by what characters Sladen distinguished his CHALLENGER species aspera from savignyi, for they seem to me identical, but he makes no reference to the old species. P.F. 410833. Natal: Umbhloti River, N. W. by W..%, W.,)23), miles, 25 fms. 1 specimen; adult. GONIASTERIDAE. Up to the present time only three species of this large family had been taken in South African waters. All of these are in the collection at hand and in addition the PieTeR FAuRE captured eight species, six of which seem to be new to science. Nearly all of the eleven species are deep water (85-500 fms.) forms and none seems to be common. Indeed not a species of Goniasteridae is represented in the collection by more than four specimens, and of four species there is but a single example of each. Unfortunately two at least The Echinoderm Fauna of South Africa. 253 of these four appear to be new. The following key shows how easily the South African goniasterids can be distinguished from each other. Key to the South African Species of Goniasteridae. Abactinal surface of disk covered with pseudopaxillae or granule-bearing tabula. Rays more or less elongated; R more than 2r. Each inferomarginal with 1—3 small, more or less appressed spinelets; no true (alveolar) pedicellariae present. R = 3—4r; superomarginals occupy less than 1/,7r;.paxillar area at base of arm about 60 of arm-width . Pseuwdarchaster tessellatus. R = 2—2'5r; superomarginals occupy '/;7; paxillar area at base of arm about 40 of arm-width . Pseudarchaster brachyactis No spinelets on inferomarginals; at least a few true pedicellariae present Mediaster capensis. Rays short, form more or less pentagonal; R less than 27. Interradial superomarginals squarish, often longer than wide, but occa- sionally wider than long; distal subambulacral. spines not conspicuously enlarged. Inner ends of interradial superomarginals distinctly squarish; their length equals or exceeds width; paxillae granules very close set, the marginal series with vertical outer sides Ceramaster chondriscus. Inner ends of interradial superomarginals markedly rounded; their width exceeds length; paxillae granules rounded and not close-set Ceramaster trispinosus. Interradial superomarginals nearly twice as wide as long; distal subambu- lacral spines conspicuously enlarged Ceramaster patagonicus, var. euryplaz. Abactinal surface of disk with no pseudopaxillae or distinct tabula. Actinal intermediate plates, each with a heavy spine, more or less elongated. Pedicellariae wanting; adambulacral furrow series of 3 or 4 stout spines Calliaster baccatus. Pedicellariae present; adambulacral furrow series with 6—9 slender com- pressed spines . 3 : ; ‘ Calliaster acanthodes. Actinal intermediate plates with granules, tubercles and pedicellariae, but no spines. No marginal plates with spines or conspicuous tubercles Tosva tuberculata. Many marginal plates with tubercles or stout spines. No disk plates with stout capitate spines or big central tubercles Cladaster macrobrachius. Many disk plates with-stout capitate spines or big central tubercles Hippasteria phrygiana. PSEUDARCHASTER TESSELLATUS. Sladen, 1889. CHALLENGER Ast., p. 142; pl. 47, figs. 3, 4. The specimens at hand (R= 32-50 mm.) are about the same size as Sladen’s (R = 48 mm.) and answer very closely to his description. There is however a median unpaired spine at the tip of the jaw Ah 254 Annals of the South African Museum. which is not mentioned by Sladen. Bell (1905, Mar. Inv. South Africa, vol. 3, p. 242) lists the species from five stations but gives no data about the specimens. It way be mentioned here that he, consistently and erroneously, throughout his report gives the date of Sladen’s CHALLENGER report as 1887. P.F. 15436. Cape Point, N.E. by N. 73/, miles, 85 fms. F. gn.s, 4 specimens; adult. PSEUDARCHASTER BRACHYACTIS *, Sp. nov. Plate XI. Figs. 4. 2. R00 mm.; 7=——domm,, R227.) Be — 139 mm... with) paxdlag area 6 mm. wide at same point. Disk large, flat, about 6 mm. thick. Arms also flat and nearly as thick as disk, except distally; they taper rapidly from the wide base to the bluntly pointed tips. Inter- brachial arcs broadly curved. Abactinal area of disk and rays, within the boundary of superomarginals, covered by low pseudo- paxillae which typically bear one central granule and a marginal series of 6-8; the granules are large, somewhat angular, rather close- set and more or less nearly subequal; near the superomarginals the granule-bearing plates lose their tabulate form and the granules are arranged more or less evidently in rows parallel to the margin. Madreporite small but distinct, about half way between the inner end of the superomarginals and the center of the disk. Supero- marginal plates very oblique, approaching the horizontal, in position, about 22 on each side of each ray but the distalmost three are very small, with their inner ends abutting on the somewhat swollen but not large terminal plate; the two plates on either side of the inter- radial line are about 5 mm. wide, but only 4 mm. long at the outer end and less than 1:5 mm. at the inner; the succeeding plates gra- dually become longer and narrower but even near tlie tip of the ray they are twice as wide as long; each plate is closely covered by granules like those on the pseudopaxillae but more rounded; the largest granules (‘25-30 mm. across) are at the outer (lower) end of the plate while the smallest are along the inner margin; there are no spinelets or tubercles on any of the plates. Inferomarginals exactly like those of upper series, with granulation and end-width reversed; on all however, one or more (sometimes as many as four) of the median granules is, or are, enlarged, lengthened and flattened to form a small and appressed but distinct spinelet; the largest of these however rarely exceeds half a millimeter in length and they * Pouxts = short -+- axtig = ray, in reference to the relatively short arms, The Echinoderm Fauna of South Africa. 255 are only bluntly pointed. Actinolateral plates about 40 in each area ; the series next to the adambulacrals extends out only as far as the fifth inferomarginal; beyond that the inferomarginals abut directly on the adambulacrals; actinal areas covered so closely with coarse granules like those on the inferomarginals that it 1s almost impossible to make out the plates; near the oral plates are two or three very simple and slightly differentiated pectinate pedicellariae, formed by the marginal granules of adjoining plates. Adambulacral plates 28-30 in each series, about as long as wide or longer, markedly convex on inner margin and slightly swollen on the oral surface. Each plate, on proximal half of ray at least, carries a marginal series of 6, rarely 7, spines, subequal as to length (about 4 mm.) or the first and last shortest, the middle pair most slender and distal pair evidently the stoutest; on the oral surface of each plate are two or three slightly oblique series of 2-4 blunt well-spaced spinelets or granules; those nearest the furrow margin are most spine- like, while those of the opposite margin are only granules; one spinelet of the series nearest the furrow or of the next series, is somewhat enlarged and distally becomes conspicuous as a_ thick, blunt but not very long subambulacral spine; not rarely two such spines occur on a plate, especially near tip of arm. Oral plates not much swollen; each plate carries two series of 8-10 spinelets, one along the sutural margin, the other followimg the outer margin; in each series, the longest spines are proximal and they become shorter and stouter distally quite rapidly; sometimes there is an isolated spine between the two series. Whether an unpaired median spine is present at the tip of the jaw is not easy to determine in the holotype as the jaws are turned upward into the mouth. But on at least one jaw it seemed to be present while on another it was almost certainly wanting. Colour, uniformly brownish-yellow. P.F. 17965. Cape Point, N. 44° E., 38 miles. 315-400 fms. S., blk. sp. 3 specimens; very young. P.F. 18904. 36° 40’ S., 21° 26’ E., 200 fms. Gr.s. 1 specimen, adult. Holotype, South African Museum, no. A 6430, P.F. 18904. The specimens from 17965 are not only young but are in very poor condition and it is not impossible that they are the young of tessellatus or even that they represent some other species. The tips of the arms are missing and the granules are largely rubbed off from both surfaces. The holotype however is in good condition and I have little doubt that it is quite a different species from any as yet described. The short wide rays with the almost horizontal marginals give it a very characteristic appearance. In the young 256 Annals of the South African Museum. specimens, the median, unpaired spine at the tip of the jaw is very conspicuous in every case, so there is reason to believe it is nor- mally present in the adult. The smallest specimen has R = 7-5 mm., r=45 mm., R=—1-66r; the unpaired spine on the jaw is perhaps 35 mm. long by -25 mm. thick. In the largest of the young specimens, r= 5'5 mm. while Kk was certainly more than twice as much; the unpaired jaw-spine is about *70 mm. long by 30 mm, wide. ‘There is no indication of spimelets on any of the inferomar- ginal plates. MEDIASTER CAPENSIS *, sp. nov. Plate XVI. Figs. 1, 2. R= 53 mms 19 mm. R= 2872 Be 20emm. at emoddle of ray, 8 mm.; at tip, 25 mm. Disk large, somewhat swollen in the radial regions; arms wide at base, narrowing rapidly at first and then, on distal half of arm, very gradually to the blunt tip. Abactinal plates of disk and base of rays, tabulate, more or less paxilliform, crowned with a marginal series of 42—15, slightly angu- lar, blunt spinelets or coarse granules and within this circle 3—8 similar and scarcely smaller granules; in the interradial regions, near the superomarginals and on the distal part of the rays, the plates are less paxilliform and carry 5-10 small granules, variously arranged; occasionally one of the granules, on the larger plates, is replaced by a small 2-jawed pedicellaria, but these are neither numerous nor conspicuous. Papulae numerous, large, arranged quite regularly, so that around each plate, there are six, but around any two plates there are ten and around any four plates only sixteen. Madreporite small, rounded triangular, about as large as one of the larger abactinal plates, only half as far from centre of disk as from disk-margin. Superomarginal plates about 29 on each side of each ray, all wider than long, the interradial ones almost twice as wide as long; they are closely covered with granules, almost exactly like those on the adjoining abactinal plates; there are 50-60 granules on one of the interradial superomarginals; occasionally a pedicellaria replaces a granule, Terminal plates small, slightly swollen, almost circular or rounded hexagonal. Inferomarginals apparently one fewer than superomarginals on each side of each ray; the series alternate more or less clearly at least at the middle part of the arm; the covering * In reference to the geographical occurrence, the region being a new one for the genus. The Echinoderm Fauna of South Africa. 257 of the inferomarginals is like that of the upper series. Actinolateral plates in about eight series; the first (next the adambulacrals) extends from the oral plates to about the fourteenth inferomarginal (counting from interradius); the second series extends to the eighth inferomarginal and the third reaches the sixth; remaining series irregular and made up of somewhat smaller plates than the first three; each actinolateral plate carries a marginal series of 7-9 angu- lar granules, more widely spaced than on the abactinal plates, and a single central granule, or rarely two; there seem to be no pedi- cellariae on these plates. Adambulacral plates about 56 on each side of the furrow; they are distinctly wider than long and their armature is in three very sharply defined parallel series; the furrow series consists of 4, rarely 5, subequal, almost cylindrical, blunt spines, over a millimeter long; the second series consists of 3, rarely 4, very similar but somewhat more prismatic spines of about the same size; the third and outer- most series is made up of 3 angular spinelets not much larger than the granules on the adjoining plates. Oral plates not at all conspic- uous and little swollen; their outlines are quite indistinct; proximally there are 5 spines on each side, the ones at tip of jaw longest (about 2 mm.); these spines are very strongly compressed, with widened and rounded tips; on the surface of each plate are a dozen or more smaller and more prismatic spines, the distalmost much like the actinolateral granules. Colour, brownish-yellow. P.F. 18483. Cape Point, N. by E., 9 miles. 84-87 fms. Gr. m. and s. 2 specimens; adult. P.F. 18230. False Bay, 24 fms. Fne. s. 2 specimens; adult. Holotype, South African Museum, No. A 6422, P.F. 18230. Examination of the internal anatomy confirms the evidence of the external characters, and proves this to be a true Mediaster: The internal radiating ossicles of the abactinal skeleton are well developed and rudimentary superambulacral plates are present. As regards the latter feature, however, | do not place very much confidence in its value, for unless these plates can be shown to have a real morpho- logical value in some group of sea-stars, | must doubt their phylo- genetic significance, and their presence in a rudimentary condition, or their absence, would not seem a matter of any real importance. Their position is such with reference to the ambulacrals and adam- bulacrals that their independent origin in totally unrelated groups would appear to be highly probable. There is no doubt that Mediaster capensis is very nearly related to M. australiensis H. L.C. but I think the differences in the paxilliform t ore) Annals of the South African Museum. plates of both surfaces justifies regarding them as different species. Abactinally these plates in capensis are noticeably larger, especially in the midradial line, and they carry more granules within the marginal circle, than in australiensis, while actinally the reverse is true, the actinolateral plates of capensis rarely having more than one central granule while in awstraliensis there are almost always 2-5. The papulae in capensis are noticeably larger and more regularly arranged than in australiensis. In this particular, capensis is more like ornatus Fisher of Hawaii, but the differences in the actinolateral plates and adambulacral armature prevent any confusion with that species. The specimens from 18183 are smaller than those from 18230 but they are like them in all essentials and call for no special comment. CERAMASTER CHONDRISCUS *, sp. nov. Plate XIV. Wigs: 9550: R = 52 mm.; r= 30 mm.; R. =1-7 r. Interbrachial ares very broadly round; the interradial margins of the body are almost per- fectly straight; rays well marked and rather abruptly projecting. Abactinal plates tabulate, completely granulated; the six primary plates are easily seen as the largest tabulae; otherwise the largest tabulae are at the center of the disk and on the median line of the basal half of each ray; these larger tabulae are more or less perfectly hexagonal, but the plates of the proximal part of each interradial area are more rhomboidal (in the holotype, they are perfectly rhom- boidal) or pentagonal or irregular; distally in the interradii the plates become very small, and are roughly oblong or hexagonal; the sides of the tabulae are very straight, their marginal granules being sharply cut vertically on the outer side; on the larger tabulae there are about 20 marginal and about 25 central granules, all closely crowded. In the holotype and the smallest specimen, one or several of these granules are, on a few tabulae, replaced by large, bivalved, often excavate pedicellariae; on the third specimen, these are remarkably abundant. Superomarginal plates 16-18 (17 in the holotype) on each side of each ray. Those in the interradi are nearly or quite square and there are only 6 or 7 on the basal balf of the ray, as against 10 or 41 on the outer half; the distalmost three or four are however very short and this increase in number is no doubt correlated with the relatively long rays. The bare area, which in some species of Ceramaster may occupy the whole abactinal surface of the plate, is * yorvdoioxoc = a granule, in reference to the numerous abactinal granules. The Echinoderm Fauna of South Africa. 259 greatly reduced, and is entirely wanting on the large plates of the interradial region of the larger specimens; it Is evident on all the plates of a specimen with R= 42 mm. On the inferomarginals, the bare space is present though small on 5 or 6 plates on each side of the interradius in this small specimen but is wholly wanting in the larger specimens. It looks therefore as though with increasing age and size, the marginal plates tend to become wholly covered with granules. The number and distribution of the pedicellariae is very variable; in the small specimen they are very few but in the larger ones they are more abundant; in one of the latter, they are present on a large proportion of the dorsal tabulae, and on all the supero- marginals, except those near tip of ray, there is at least one, often there are two and not infrequently, three; on most of the infero- marginals too they are present, and even on the actinolateral plates a few are to be found; in the region just back of the oral plates are 3 or 4 pedicellariae notable for their large size, fully twice that of those on the abactinal surface; on the adambulacrals, there seem to be no pedicellariae. Actinolateral plates rather numerous and crowded but their out- lines are very distinct at the center of each area, less so near mouth and least so on the outer part of each ray; except near the mouth and distally, the two series adjoining the adambulacrals are wider than long and oblong; those at center of area are rhomboidal; else- where they are irregularly polygonal or rounded; the granulation is much coarser than abactinally and there is no obvious distinction between the marginal and central granules; even the largest plates have only 20-25 altogether. The series adjoining the adambulacrals extends out to about the sixth or seventh inferomarginal from the tip; the next series reaches only to the ninth. At the middle of each interradial margin there are about 3"), actinolateral plates abutting on each inferomarginal. Adambulacral plates 40-45 on each side (in the holotype) wider than long at first but becoming squarish distally. The armature consists of a furrow series of 4, blunt, thick, somewhat prismatic or slightly flattened, subequal spines about 15 mm. long; back of these is a nearly parallel series of 3 similar but shorter spines and the outer end of the plate is occupied by 2-4 still smaller, but yet some- what similar spinelets; these last are distinctly larger than the biggest granules of the adjoining actinolateral plates. Oral plates large but flat and not at all swollen; the armature is almost exactly similar to that of the adjoining adambulacrals; there are about 8 large spines on each free margin and a series of about 8 prismatic 260 Annals of the South African Museum. granules along each of the opposed margins. — Colour in alcohol, pale brown, becoming brownish-white on drying. P.F. 15147. Table Mountain, E. by S. !/, S., 25 miles, 190 fms. Gr. s. and bl. sp. 3 specimens. Adult. Holotype, South African Museum, no, A 6444. | had determined to call these three specimens patagonicus but Fisher thinks they are nearer to his recently described C. smithi from the Philippines, in 554 fms. He says that the South African specimens differ from smithi in the clean cut hexagonal tabulae of the mid-radial areas, the more numerous abactinal granules (only 10-15 central granules on largest tabulae in smithi), in the smooth tips of the subambulacral and furrow spines, in the lower abactinal pedicellariae, and in the larger oral plates. From patagonicus (of which I have seen no specimens) Fisher tells me the South African species differs “in having narrow, sunken, wholly granulated mar- ginal plates, broader abactinal radial plates with more crowded, numerous granules, large instead of small plates in center of disk, a different sort of actinal pedicellaria, etc.” It seems to me very clear that patagonicus, smithi and chondriscus are very closely related forms and that we shall not know the true interrelationship until | we have far more material. | CERAMASTER TRISPINOSUS *, sp. nov. | Plate XIV. Figs. 3, 4. | R=4A1 mm.; r=21 mm.; R=1:95r. Interbrachial arcs broad- ly rounded; rays bluntly pointed. Abactinal plates tabulate, poly- gonal, of diverse sizes and closely crowded; most of the plates are rather large with a marginal series of 10-20 coarse, rounded gra- nules and 10-20 similar, not crowded, granules within the marginal series; smaller plates have 6-12 marginal granules and 410 more on the top; the five basal plates are easily distinguishable, as one is somewhat crescent-shaped and encloses the madreporite on its outer side, while the other four have more numerous and smaller granules than the other tabulae, about 30 in the marginal series and about 35 within. Superomarginal plates 43 or 44 on each side of each ray or 26 or 28 on each side of the pentagon; the interradial pair are, each 4 mm. wide and 3 mm. long, with the inner end so curved as to be almost a semicircle; they are fully covered by about 150 granules, of which the largest are on the lower margin, next * trispimosus = having three spines, in reference to the armature of the adam- . bulacral plates. The Echinoderm Fauna of South Africa. 261 the inferomarginals; there are 8-10 on that margin, 8 or 9 in the marginal series up each side and 48-20 on the semicircular inner (upper) margin; the second, third and fourth superomarginals are similar but progressively slightly smaller and with more square cut inner ends; on the fifth plate is a small bare area and this increases in size on the succeeding plates, until on the distal plates only a marginal series of granules remains; the last three superomarginals of the two sides meet in the midradial line, so the abactinal plates do not reach the terminal plate; the latter is of moderate size, rounded triangular or pentagonal and decidedly swollen. Madreporite small, only 15mm. in diameter, its outer margin 12 mm. from edge of disk. Inferomarginals 14 or 15 on each side, always one more and sometimes two more than the superomarginals of the same side ; in the neighborhood of the sixth superomarginal there are two infero- marginals and at the tip of the ray another extra inferomarginal is often to be found; the inferomarginals are very similar in form and granulation to the adjoining superomarginals. Actinolateral plates numerous, but so crowded and so closely granulated that the series can be made out only with difficulty ; that adjoming the adambulacral plates extends to the eighth inferomarginal while the next series reaches only to the sixth; the granulation is much coarser than that on the marginals or abactinal plates and is well-spaced; there are rarely as many as 20 granules on a plate, Adambulacral plates about 50 in each series, short and crowded, much wider than long except distally where the length nearly equals the width. Each plate carries a series of 3 (or rarely 2) stout spines on the furrow margin; these spines are a millimeter long, subequal, blunt, cylindrical or more or less compressed; back of this series, there are on the oral surface of each plate, three pairs of spines; the first (innermost) of these is much stouter and a little shorter than the furrow-spines, and the distal spine is larger than the proximal; on the terminal part of the arm, this larger spine becomes quite conspicuous as relatively the biggest adambulacral spine; the other two pairs of spines are much smaller, and the outer one is scarcely larger that the granules of the actinolateral plates; on some adambulacrals, one (or even two) of these six surface spines is wanting. Oral plates not at all swollen; on each free margin is a series of-5 or 6 stout, more or less prismatic, subequal spines; just back of these is a series of 5 similar but shorter spines, and on the distal part of each plate are about 5 still shorter spines or coarse prismatic granules. There seem to be no pedicellariae anywhere. Color of dried specimen, uniformly dingy, brownish-yellow. 262 Annals of the South African Museum. P.F. 2798. Vasco de Gama Peak, N. 71° E., 18 miles, 230 fms. Stones. 1 specimen, adult. Holotype, South African Museum, no. A 64145. This species has a very characteristic appearance due to the form of the marginals, the absence of pedicellariae and the crowded adam- bulacral plates with their furrow-series of three spines. The form is distinctly less pentagonal than in most members of the family, the tips of the rays being markedly prolonged. The granulation both above and below is noticeably coarse, but it is especially so on the actinolateral plates. CERAMASTER PATAGONICUS var. EURYPLAX * var, nov. Plate XIV. Figs. 4, 2. R = 32 mm.; r = 20 mm.; R=167r. Form nearly pentagonal but the sides are slightly concave. Abactinal plates tabulate, poly- gonal, of diverse sizes and closely crowded, so that the sides are very straight and clear cut, as in C. patagonicus; radially the plates are perfectly hexagonal and interradially they are rhombic; they are smallest at center of disk and near the marginal plates; the larger plates have a marginal series of 12-14 coarse granules and 10-18 similar but slightly smaller granules are within the marginal series ; the latter have their outer sides quite vertical and the adjoining angles sharp; a central plate and the five basals are distinguishable by their smaller granules. Superomarginal plates 10 or 41 on each side of each ray or 20-22 on each side of the pentagon; the inter- radial pair are each 4 mm. wide and 2°5 mm. long, approximately rectangular, with nearly straight edges; succeeding plates similar but progressively shorter; there is little change in width until very near the tip of the ray; the central abactinal part of each plate is slightly tumid, bare and smooth; this bare area is largest distally and smallest on the interradial pair; elsewhere the plates are closely covered with a coat of granules of very uniform size, of which there may be more than 200 on a plate. Terminal plate of moderate size, very tumid, pentagonal, smooth. Madreporite small, wider than long, 1°75 mm. across, its outer margin 13 mm. from edge of disk. Infero- marginals of the same number as the superomarginals; the interradial pair underlie the interradial superomarginals but each succeeding plate lies progressively more distal so that near the tip of the ray the two series alternate; in granulation the inferomarginals resemble * évets = wide + wias = plate, in reference to the very wide interradial superomarginals. The Echinoderm Fauna of South Africa. 263 the upper series exactly except that the bare area is smaller, while in form they are perfect complements of the adjoining superomarginalss Actinolateral plates numerous and crowded, arranged in about eight series parallel to the adambulacrals; first series extends from oral plates to sixth inferomarginal and is made up of about 21 plates, which, excepting 2 or 3 at each end, are distinctly wider than long; succeeding series very crowded and hard to distinguish, the component plates about square; all the plates are covered by a close granulation like that on the inferomarginals but becoming coarser on the series near the adambulacrals. Adambulacral plates about 53 in each series, not much wider than long (if any) and not specially crowded. Each plate carries a series of 4 or 5 stout spines on the furrow margin; these spines are about a milli- meter long, blunt and thickened at tip, more or less compressed ; when 4are present, the middle pair are a trifle longer than the others; if a fifth spime occurs it is proximal in position and much smaller than the others; back of this marginal series, there are, on the oral surface of each plate, parallel with the furrow, three series of spinelets, of which two have three spinelets each and the outermost usually has four; the outermost series is no larger than the adjoining granules of the actinolateral plates, while the other series are slightly more spine-like; near the mouth, the outer series merges with the third or disappears altogether; distally the number of spinelets in each series is reduced. Just beyond the middle of the ray the distal spinelet of the second series is somewhat larger than its fellows: this disproportion increases as the tip of the ray is approached and the number of spinelets decreases, until, on the last ten or a dozen adambulacral plates, this spinelet is a conspicuous subambulacral spine, about a millimeter long and half a millimeter thick. Oral plates not at all swollen; on the free margin is a series of 9 stout, prismatic spines, the innermost stoutest; parallel to the sutural line between the two plates is a series of 8 crowded spinelets, of which the distal ones are scarcely larger than the granules of the adjoiming actino- lateral plates; a secondary series of 6 smaller spinelets runs irregularly parallel to this sutural series and there are 2 additional spinelets between it and the marginal series. There seem to be no pedicellariae. Colour of dried specimen, dingy brownish-yellow, P.F. 15366. Cape Point N. 16° E., 10 miles, 85 fms. Gm. m. 1 specimen; adult. Holotype, South African Museum, no, A 6413. This handsome goniasterid is very near patagonicus of the same size from Alaska. Dr. Fisher has kindly compared them and finds 264 Annals of the South African Museum. so little difference that he advises considering this specimen, for the present, as only a variety of patagonicus. He says the abactinal plates are larger than in patagonicus, being more as in granularis. It is possible that in larger specimens, the bare area on the marginal plates would disappear, at least interradially. CALLIASTER BACCATUS, Sladen, 1889. CHALLENGER Ast., p. 280; pl. 56, figs. 4-4. The Prerer FAuRE specimens agree well with Sladen’s description and figures. The larger has R= 44 mm. and the smaller, 40 mm.; the former is thus just the size of the original specimen. The Mossel Bay specimen is somewhat larger as R = 52 mm. On a single actinal plate of this specimen is an indubitable pedicellaria and there are several of the pits where pedicellariae have been. The pedicel- lariae are thus not invariably wanting in this species. Their usual absence is however one of the many good species characters which baccatus possesses. The single pedicellaria seen has unequal, asym- metrical, non-denticulate valves; the larger valve is scarcely higher than wide and is a little bent sideways; the smaller is more decidedly bent and is distinctly narrower. P.F. 1173. 34°48’ S., 22°43’ E., 38 fms. 4 specimen; adult? P.F.-1710. Cape St. Blaize, N. by E. */, E., 6), miles, 35) fms. M., s. 1 specimen, adult ? Mossel Bay. C. W. Black, 1913. 4 specimen, adult. CALLIASTER ACANTHODES * sp. nov. Plate XII. Figs. 3, 4. R= 790m. 3 7 — 27 mes he — nearly or, Br —=30) mmewab fifth superomarginal, br = 14 mm. and at 12th, bb’ = 9 mm. _ Disk large, slightly tumid but with depressions near interradial margins. Rays tapering at first abruptly but beyond fifth superomarginal, very gradually. Abactinal surface of disk covered with irregularly circular plates, which are more or less tumid and bare, though there is a marginal series of coarse, flat, irregular granules around each one; the median radial series comprises the largest plates and runs almost to the tip of the ray but the distalmost plates are separated from the terminal plate and from each other also, by the meeting in the midradial line of the distal superomarginal plates; the series of plates on either side of the radial runs as far as the 42th superomarginal ; * axav§odns = full of thorns, in reference to the numerous abactinal spines. The Echinoderm Fauna of South Africa. 265 all the larger abactinal plates and many small ones too, bear a single, central blunt spine, 1-3 mm. long and about }/, mm. in diameter; not rarely the spine, on the smaller plates, is replaced by a large non-denticulate spatulate-jawed pedicellaria; on the larger plates, spine and pedicellaria may both occur. Madreporite large, tumid, about 3 mm. in diameter and 12 mm. from the disk margin. Superomarginal plates 16 on a side, bare and tumid; the proximal are squarish and about as long as wide but distally the plates become much wider than long; each plate (except near tip of ray) carries 2, and sometimes 3, stout spines like those on the abactinal plates; these are placed one above the other; besides these spines one or more coarse granules or small tubercles may be present or, occasion- ally, one or even two pedicellariae occur instead of the tubercles; the usual series of marginal granules surrounds each of the plates. Terminal plate quite small, swollen and with no spines or tubercles whatever; it is possible that these may have been present in life and have since been knocked off but if so they have left no scars. Infero- marginal plates 17 on each side, the basal ones longer than wide and longer than the corresponding superomarginals, but distally they decrease in length rapidly and an extra one is intercalated below the twelfth of the upper series, or thereabouts; these plates carry 2-5 spines in a central group, or in a vertical or horizontal series; the spines are similar to those of the upper plates, and like them may be accompanied by pedicellariae. Actinolateral plates im six or seven series, the first parallel to the adambulacrals and reaching as far as the seventh inferomarginal; the second series does not quite reach the fifth inferomarginal; the remaining series are confined to the disk; each plate is surrounded by the usual marginal granules and these also occur more or less abundantly on the surface of the larger plates, especially near the mouth; each plate, excepting only the small ones, carries a large, central spine, similar to those of the abactinal surface but perhaps a little bigger; on some of the plates, the large characteristic pedicellariae occur. Adambulacral plates 57 in each series but 21 of these are on the last 148 mm. of the arm; there are 6-9 (usually 8 or 7) slender compressed spines on the furrow margin, which are subequal or the end ones may be much the smallest; on the surface of the plate are 2 large spines, placed one behind the other, and on the adoral, inner corner there is usually a big pedicellaria; the plates are surrounded by the usual marginal granules and a number of these occur on the face of the plate, particularly around the base of the outer spine. Oral plates long and narrow, but not swollen; on the free margin 266 Annals of the South African Museum. is a series of 8 or 9 long, blunt, compressed or prismatic spines, the innermost largest; on the face of each plate is a single big spine, between which and the tip of the jaw are three or four sharp, angular spinelets; distally a series of 10 or 14 granule-like spines runs along the outer margin, and 5 or 6 much coarser granules lie along the sutural margin. Colour of holotype, in alcohol, yellow- brown; of paratype, dull brownish-red above, more or less irregularly bleached; lower surface, nearly white. P.F, 12834. Buffalo River, N.N.E. 17 miles, 195 fms. St., r. 1 specimen; small adult. P.F. 14232. Cape St. Francis, N.E. 29 miles, 75 fms. S., sh., r. 2 specimes; adult; one very poor. Holotype, South African Museum, no. A 6424, P.F. 14232. This fine species is quite different from baccatus but is very near corynetes Fisher and spinosus H. L. C. It is readily distinguished from the former by the spiny upper surface and the pedicellariae on the adambulacral plates, and from spinosus by the bare abactinal plates and the presence of only one large spine on each oral plate. I was at first inclined to consider these specimens as adult baccatus but careful comparison shows that this idea is absurd. The differences in the adambulacral armature are fundamental and cannot possibly be construed as growth stages, and the same must be said of the condition of the marginal plates. One of the specimens from 14232 was evidently dried directly from salt water, perhaps with the laud- able purpose of preserving the colour, but unfortunately, with the passage of time, it has disintegrated sadly and is now of little value. It was somewhat larger than the holotype, as 7 = 30 mm. The present colour is deep red brown, the marginals being darker than the abactinal plates. TosIA TUBERCULATA. Plate IX. Figs. 1, 2. Astrogonium tuberculatum Gray, 1847. Proc. Zool. Soc. London, p. 79. 1866. Syn. Starfish, p. 10; pl. 4, fig. 2. Tosia tuberculata Verrill, 1899. Trans. Conn. Acad., vol. 410, p. 164. Although Bell (14905, Mar. Inv. South Africa, vol. 3, p. 246) recog- nized the fact that this species is very little known, he does not give one word of information about the numerous specimens he had be- fore him, except that the species is now ‘found to grow to a good size”. What ‘a good size’ may be each reader must decide for himself! However, two of Bell’s specimens are now in the collection The Echinoderm Fauna of South Africa. 267 of the Museum of Comparative Zodlogy and have been examined by Fisher, who has published some notes on them (1911, Bull. 76 U.S. Nat. Mus., p. 166). In the Prerer Faure collection, I find a single starfish (P.F. 18154. Cape Point, N.E. by E. 3/, E., 28 miles. 300 fms. Fne.s.) which is undoubtedly identical with these M.C. Z. specimens (as comparison side by side shows) but it is considerably larger and differs in certain details. Its most striking feature is the abundance of large bivalved, and often excavate, pedicellariae all over the ab- actinal and marginal plates; they are rather infrequent on the actinal surface and seem to be wholly lacking on the adambulacral plates, the only plates on which they are to be found in the M. C. Z. spe- cimens, one would infer from Fisher's notes (op. cit. p. 167). However Fisher probably does not mean to imply that, for there are numerous pedicellariae on the abactinal surface of both these specimens, while the adambulacral pedicellariae occur only in the Jarger. Judging from the three individuals at hand, in which R= 42, 48 and 54 mm. respectively, one would say of this species: large, bivalved, often excavate, pedicellariae occur commonly and even abundantly on the abactinal and superomarginal plates, but are less frequent and may be wanting on the inferomarginal and actinal plates; their occurrence on the adambulacrals is unusual and when present there, they are strictly bivalve and have high, rather narrow jaws. Both Verrill and Fisher put this species in Tosta but it would seem to be nearer to Plinthaster. Verrill apparently had not seen any specimens but, except for the large size of the pedicellariae, the individuals at hand, answer well to his diagnosis of Plinthaster. They also run down to Plinthaster most naturally and without question in Fisher’s admirable key to the genera of Goniasteridae (op. cit., pp. 169—174); here again the only difference is in pedicellariae. On the other hand the obvious presence of secondary plates in the radial areas seems an obstacle to putting this species in Tosta, and the. general facies is quite as unlike that genus as it is that of Piinthaster. Dr. Fisher thinks that the species these South African specimens represent might well be made the type of a new genus but I think it will be well to wait until more material is available and further study has been made of Gray’s type material in the British Museum. The Pieter Faure specimen has much longer rays relatively than either of the M. C. Z. specimens, so that the body form is quite different. This can best be shown by the following comparison. In the larger M. C. Z. specimen, R = 48 mm.; r = 28 mm.; Or half- way to tip of ray, 22 mm.; br three-quarters of the way to tip, 8 mm. ; 268 Annals of the South African Museum. thus R=1:77r; or 2°2br at half-way point; or 6br at three-quarters point. In the PreTerR Faure specimen, R =54 mm.; r=26 mm.; br at half-way point, 13 mm.; 6r at three-quarters point, 7 mm.; thus R=2-1r; or 44 br at half way point; or 7°7 br at three-quarters point. Probably a large series of specimens would show that there is considerable individual diversity in these proportions, and very likely, an increasing ray-length, with age. Colour in life: upper surface reddish orange, lower surface pale. Perhaps it ought to be added that it is not certain that the spe- cimens identified by Bell are really éaberculata; he does not say whether he compared them with the type or not. Certainly Gray’s figure does not resemble at all closely any one of the three specimens at hand. CLADASTER MACROBRACHIUS* sp. nov. Plate it. ihgss 42: R= 40 mm.; r=16 mm.; Ro==2:57- br 418 mm. but-at halt- way to tip it is only 9 mm. Disk large, somewhat convex but only about 8 mm. thick, even at center. . Rays flat, tapering, at first rapidly, then gradually to the blunt tip. Abactinal plates moderate in both size and number, irregularly polygonal, with rounded corners, thick and close together; papulae few, single, typically six about any one plate on center of disk or base of rays but usually one or more of the six, lacking. Each plate, in life, was evidently surrounded by a marginal series of small, well spaced granules and bore on top, several larger, more widely spaced granules, one of which was here and there replaced by a large bivalved, more or less excavate pedi- cellaria; in the preserved specimen (dry) all the top granules, some pedicellariae and many marginal granules have been rubbed off but each has left a shallow pit to indicate its location. Median radial series of plates shut off from terminal plate by the meeting of the five distal pairs of superomarginals; series of plates adjoining radials only extends as far as the fourth or barely to the fifth superomarginal. Madreporite small (less than 2mm. in diameter), pentagonal, situated about 10 mm. from the disk margin. Superomarginals 13 or 14 on each side of each ray, wider than long, more or Jess markedly tumid ; like the abactinal plates, each is surrounded by a marginal series of small granules, and in life was very sparsely covered by much coarser and more widely spaced granules; on the upper end of each plate, where it is most markedly tumid, there are two or three (distally one or none) large, shallow scars, which indicate that in life rather * waxoos = long + Poazyiwy = arm, in reference to the relatively long rays. The E'chinoderm Fauna of South Africa. 269 coarse granules or big tubercles were present. Terminal plate small and swollen; there are indications that in life it may bear 1—3 tubercles. Inferomarginals agreeing with superomarginals in number, form, size, position and granulation, except that the large, shallow scars are as a rule less well-marked and often seem to be wanting. Actinal plates few, irregularly arranged (except for series adjoining adambulacrals), of diverse sizes; the smaller ones are pretty well covered by the very large marginal granules, but all the larger plates show a bare central area on which is a big, wide-valved pedicellaria, and rarely a single big granule also; the series adjoining adambulacrals extends out only as far as the fourth inferomarginal. Adambulacral armature conspicuously heavy; the plates themselves are numerous, about 45 in each series, crowded, much wider than long proximally, but squarish distally; each plate carries a series of 3 (rarely 2) furrow spines, about a millimeter long near middle of arm (longer proximally, shorter distally) subequal, or middle one longest, markedly compressed at right angles to furrow and more or less conspicuously widened at tip; back of these is a second series of which the adoral is very small, the middle one is much larger and the aboral is a stout, somewhat capitate subambulacral spine, the largest spine on the plate; on the outer margin of the plate is a third series of three spines of which the middle one is much the largest; the two small ones are hardly bigger than the marginal granules of the adjoining actinal plates; proximally all the adambul- acral spines are longer, heavier and more conspicuous, while distally they decrease in number as well as in size. Oral plates not swollen, their outlmes hard to determine; each has a marginal series of strongly compressed spines, about 2 mm. long, with much widened tips; there is also a series along the sutural margin consisting of 5 or 6 spines of which the first is small and pointed, the second is a long heavy spine like those of the free margin, the third is like it but a little smaller and the remainder are successively shorter and smaller in every way. Color of dried specimen, light yellowish-brown. P.F. 2798. Vasco de Gama Peak, N. 71° E., 48 miles. 230 fms. Stones 1 specimen; adult? P.F. 17998. Cape Point, N.E. 3/, N., 39 miles, 310-500 fms. Gn. m. 1 specimen; adult ? Holotype, South African Museum, no. A 6429, P.F. 17998. Aside from these interesting individuals, which differ little from each other, only two specimens of Cladaster are known; one, the holotype of C. validus Fisher with R=17 mm. was taken near the Aleutian Islands; the other, the type of C. rudis Verrill with 18 270 Annals of the South African Museum. R=25mm., was taken in the West Indies. The present individuals are thus much larger and it is noticeable that they have clearly the longest arms; validus is most nearly pentagonal. Probably the rela- tive length of the rays increases with age. Perhaps the number of spines in the furrow series also increases with age, for the South African form has three as against two in the other species. Whether these South African specimens are adult seems doubtful and it is probable that a fully grown specimen would throw much light on the relationships of the genus. If it is true that the superomarginal plates in macrobrachius bear coarse tubercles, the definition of the genus will need some modification. HIPPASTERIA PHRYGIANA. Asterias phrygiana Parelius, 1768. K. Norske Vid. Sels. Skrift., vol. 4, p. 423; pl. 14, figs. 1, 2. Hippasteria phrygiana Verrill, 1885. Rep. U.S. Fish Comm. for 1883, p. 942. Up to the present time only a single specimen of Hippasteria has been recorded from the southern hemisphere. This was from the Strait of Magellan and was first described by Perrier as H. hyadesi, later as H. magellanica, and subsequently he used either name, ap- parently interchangeably. Verrill adopted magellanica but hyadesi seems to have priority, if the species has any validity. Perrier him- self says it is very difficult to distinguish from phrygiana, and the differences which he points out are no greater than are to be found between two specimens of phrygiana from the New England coast. He gives no measurements and no figures so that there is no way of determining whether his specimen was adult or young. The two specimens in the PieTER FAuRE collection only add to the difficulty; they are quite unlike each other and neither is like Perrier’s specimen. But I am quite unable to estimate the value of the characters they show, for while they seem like representatives of two different species, they are not so unlike each other as are two specimens of phrygiana from the north-eastern coast of America, which he before me. All four specimens are young, not half grown, but their peculiarities are not to any great degree due to their youth, I feel quite sure. I am forced to conclude that either all four represent one species, or each one represents a separate species. The former seems to me the more probable alternative and | am therefore referring the PirrerR FAURE specimens to phrygiana. It is quite likely however that a good series of adult Hippasteria from Sey) ai EE ei RON mie Ne, 95) The Echinoderm Fauna of South Africa. 271 either South Africa or the southern part of South America will show some constant specific characters. Meanwhile it may be well to record briefly the chief peculiarities of each of the PirreR FAuRE specimens. Speamen A. KR == about 52 mm; 7 = 245 mime: = 257:; br = 21 mm. but at half-way point is only 414 mm. Disk large, rather flat; rays tapering rapidly to an almost pointed tip. There are no large spines on the abactinal surface but each of the larger plates carries a big pedicellaria or a short spine or a high tubercle. Superomarginal plates, each with one or interradially two rather stout spmes; on the interradial plates there are some large granules in addition. Inferomarginals with a shorter and thicker spine and 2-10 coarse granules in addition; the imterradial plates have the most granules. Actinal plates usually with a big central pedicellaria and a marginal series of few very coarse granules; often a big granule or two replaces the pedicellaria. Adambulacral armature usually of a single large spine on the furrow margin, a similar but shorter subambulacral spine and 3 or 4 granules on outer end of plate; proximally there are 2 and rarely 3 spines on the margin, but they are more slender, and compressed, and there is no conspi- cuous subambulacral spine. Oral plates with only 4 or 5 marginal spines, but they are big, somewhat compressed and blunt; there are no big spines on the oral surface of plates. P.F. 2798. -Vasco de Gama peak, N. 71° E., 18 miles. 230 fms. Stones 1 specimen; young? Specemen. Ba) Re 750 mm.; r = 25 mm:; R= 273 be 30mm but at half-way poimt is 17 mm. Disk large, slightly tumid; rays broad, rather flat, taperimg uniformly to a blunt point. There are no spines at all on the abactinal surface; many plates carry a pedi-. cellaria or a single large granule at center but some are quite bare; the result is an unusually smooth surface for a Hippasterva. Super- omarginal plates, each with a single, short thick spme; on the interradial pair, a second shorter spine is below the first; on a few plates a large granule accompanies the spine. Inferomarginals with a single large tubercle or thick spinelet; interradially, several gra- nules accompany this tubercle. Actinal plates as usual with a big central pedicellaria or occasionally a large tubercle. Adambulacral armature made up of a furrow series of two stout, bluntly pointed spines, the aboral the larger, a very stout sugar-loaf shaped sub- ambulacral spine with one or two granules adoral to it, and about 4 coarse, angular granules on the outer end of the plate. Oral plates forming a rhomb, on each side of which are 3 stout spines; 272 Annals of the South African Museum. those of the inner sides are quite markedly compressed; on the sur- face of each plate is a single, stout spine. P.F. 17997. Cape Point, N. E. %, N., 39 miles, 340-500 fms. Gn. m. 1 specimen; young? A specimen of phrygiana taken by the CHALLENGER on La Have Bank, south of Nova Scotia, in which R = about 57 mm. is much like A in form and proportions but in its adambulacral armature it is much like B. On the other hand, a specimen with R= 48 mm., collected near Grand Manan, has so many big nearly spherical tubercles on the abactinal, marginal and actinal plates that its general appearance is quite different from any of the others; the adambulacral armature approaches that of A but the big furrow spine usually has a very small spine adoral to it and sometimes an aboral one is present also. There is little question that Hippasteria phrygiana is very varia- ble. Possibly more than one species is now included under that name or it may be that varieties or subspecies should be recognized. But until the growth changes are known and a large series of spe- cimens from many localities has been gotten together and studied, seems to me best to let a single name cover all the Atlantic forms of Hippasteria. OREASTERIDAE. There is only a single specimen in the South African collection to represent this well-known tropical family of big sea-stars. Four other species have been reported from South Africa however, so the family is better represented there than the present collection indi- cates. Nevertheless it must be granted that South Africa is a little too far outside the tropics for even such a ubiquitous warm-water genus as Oreaster to flourish, and probably south of Mozambique, the Oreasteridae are represented chiefly by stragglers. It is an easy matter to distinguish the few species that have been recorded hitherto. Key to the South African Species of Oreasteridae. Rays well developed. One or two distal superomarginals on each side of each ray bear a very big spine, while the remaining marginal plates are merely a little tumid and carry no spines’. é : : Oreaster linckar. Superomarginals without spines or aio nail or moderate ones on many plates, especially in interradiu—. Oreaster mammillatus. Rays very short or apparently wanting, as ane aha is thick and cushion like, and pentangular or roughly circular. The Echinoderm Fauna of South Africa. 273 Furrow-series of adambulacral armature with 5—7 spinelets; papulae confined to special areas above the margin. Papular areas with little spinelets; tubercles of dorsal side rather small and more or less pointed . : : . Culcita novaeguineae. Papular areas without spinelets; tubercles of dorsal side, big, scattered and blunt : . Culcita schmideliana. Furrow-series of aaeabalecral aiiaiare with only 2 or 3 spinelets; papulae all over back, clear to the margin : : : Culeita veneris. * OREASTER LINCKII. Asterias linkii de Blainville, 1830, Dict. Sci. Nat., vol. 60, p. 219. See also Linck, 1733, De Stell. Mar., pl. 7, no. 8. Oreaster linckii Liitken, 1864. Vid. med., p. 156. Linck’s figure gives a very good idea of a typical specimen of this species, which is common at Zanzibar and has been reported from Mozambique by both Peters and Bell. OREASTER MAMMILLATUS. Asterias mammillatus Audouin, 1826. Expl. som. des pls. Echinod. de Egypte pub. par Savigny, p. 209: Rayonnés, pl. 5. Oreaster mammillatus Miller and Troschel, 1842. Syst. Ast., p. 48. This is a very variable species and the growth changes and limits of variation need very much- to be worked out. In some spe- cimens, spines and even the big tubercles are nearly or quite lacking while at the other extreme, every big dorsal or super- omarginal plate carries a small or moderate spine. Peters re- ported the species from Mozambique and it is not recorded from south of there, but in the present collection is a specimen from Mossel Bay, Cape Colony, which thus extends the known range of the species many hundreds of miles to the south. The specimen is a small one (R = 60 mm.) and lacks one ray, which is however beginning to regenerate. There are no spines anywhere but many of the marginals, especially of the lower series, and a number of abactinal plates bear more or less elevated tubercles. The dry spe- cimen is light yellowish-brown. Mossel Bay. C. W. Black. 4944. * CULCITA NOVAEGUINEAE. Miller and Troschel, 1842. Syst. Ast., p. 38. Déderlein, 1896. Jena Denkschr., vol. 8, pls. 19 and 20, figs. 1—9. This widely distributed Indo-Pacific species is very variable and has been described under a number of names. It has been reported from Mozambique and there is a young individual from that place 274 Annals of the South African Museum. in the Museum of Comparative Zodlogy, identified and labelled by Perrier as “‘Randasia granulata Gray. jeune Culcita areolata E. Per.” CULCITA SCHMIDELIANA. Asterias schmideliana Retzius, 1805. Diss. Ast., p. ?* Culcita schmideliana Gray, 1840 Ann. Mag. Nat. Hist , vol. 6, p. 276. Déderlein, 1896. Jena Denkschr., vol. 8, pl. 20, figs. 10-15. This species seems to be common at Zanzibar and has been reported from Mozambique by Bell. A specimen is in the 8. A. Museum from Mozambique collected by K. H. Barnard, 1912. * CULCITA VENERIS. Perrier, 1879. Arch. Zool. Exp., vol. 8, p. 48; pl. 4. This species then known only from the holotype, taken at St. Paul Island in the southern Indian Ocean, was recorded by Bell (41905, Rep. Mar. Inv. South Africa, vol. 3, pag. 248) from near the Cape of Good Hope, in 23—37 fms. In spite of the extraordinary inter- est attaching to the rediscovery of so remarkable an animal, Bell does not give a single bit of information in regard to his specimen and we can only surmise that it was so much like the holotype in size and appearance that he felt no doubt of their identity. PORANIIDAE. This family was not hitherto known from the vicinity of South Africa, and it is represented in the PreTer Faure collection by only a single specimen. This however is of very great interest as it proves to be an undescribed species of a little-known, and hitherto monotypic genus of the North Atlantic. CHONDRASTER ELATTOSIS **, sp. nov. Plate VIII. Fig. 4. Re=— Mo mmijor — 75 mms = 1-539 Wisk elevated avande at center, 40 mm. Whole animal covered with a thick, smooth, fleshy skin. Abactinal skeleton wanting or greatly reduced; mar- ginal plates present but very spongy and without spines or tubercles. ' Papulae numerous, arranged in two parallel series, 20 mm. wide and 5 mm. apart, along the median, radial area; a few small groups of papulae, occur irregularly near the center of the disk. Anal opening * Few writers have been able to consult this paper and no one has given the page reference. The paper is not accessible to me. ** z)iacrow = to lessen, in reference to the reduction of the skeleton. Ae Leth llth: ela ane PRY ON eecay The Echinoderm Fauna of South Africa. 275 evident. Madreporite distinct but small, 3 mm. across; rather spongy. Actino-lateral areas with numerous parallel furrows running to margin and even over the margin onto the upper side; no spines or tubercles anywhere. Adambulacral plates with an inner series of 3 or often 4 sharp spines, 1-3 mm. long, sacculate, the saccules ex- tending far beyond the spine-tip; and an outer series of 3, rarely 4, similar but stouter spines, enclosed in a thick, fleshy sack and forming a low, racquet-shaped appendage, 3-4 mm. high and 2°5 mm. wide. Oral plates very thick but flat with no superoral spines or tubercles; at the inner tip of each plate is a rather stout, sacculate, nearly horizontal spine; along the free margin of each plate is a series of similar but longer and stouter, vertically placed spines, united with each other and with the plate itself by skin. Feet large in two series. Colour uniformly dull, deep pink; feet brown. P.F. 19003. South from Cape Infanta, Cape Colony, 36° 49’ 5., 21° 14’ E., 560 fms. Gn. s. 4 specimen; adult. Holotype; South African Museum No. A 6448. This remarkable sea-star was unfortunately preserved in formalin and it is evident that some decalcification has taken place. It is however impossible to determine now how much of the sacculate appearance of the adambulacral and oral spines is due to decalcifi- cation and how much is natural. It is also uncertain how much of the absence of a dorsal skeleton, and to what degree the sponginess of the marginal plates, is artificial. There is however little doubt as to the generic position of this notable specimen, as it agrees so well in its main features with Chondraster grandis Verrill, which occurs in the northern Atlantic, southeast of New England, in 220-538 fms. The South African species differs from the genotype however in the wider papular bands, the greater reduction of the skeleton, the absence of marginal tubercles, and particularly in the armature of the adambulacral plates. In grandis there are only two spines in the inner series. The two species apparently differ also in colour, as the northern form is red above and yellow beneath, while the southern species seems to be unicolorous. This may of course be only an individual matter. OPHIDIASTERIDAE. This is another tropical family and its inclusion in the present report is due chiefly to the fact that four species are listed by Bell in the ALERT Report (1884) as having been taken at Mozambique. One of these is represented in the South African collection before 276 Annals of the South African Museum. me by two small specimens but these also are from Mozambique. The only truly South African species is the interesting Austrofromia from False Bay. Key to the South African Species of Ophidiasteridae, Papulae on actinal surface; adambulacral armature spiniform. Papulae single; rays 3 or 4 times as long as wide at base Austrofromia schultzet. Papulae in areas; rays 5 or 6 times as long as wide at base Nardoa variolata. No papulae on actinal surface; adambulacral armature granuliform. Inner (furrow) series of adambulacral spines with spines separated from each other by vertical series of little granules. Colour blue; arms relatively short and wide, R= 5 or 6 br Linckia laevigata. Colours orange and green; arms relatively long and slender, R = 7—12br LInnckia multifora. Inner (furrow) series of adambulacral spines with small spines alternating with larger and no vertical series of little granules between Linckia diplac. * AUSTROFROMIA SCHULTZEI. Fromia schultzei Déderlein, 4910. Jena. Denkschr., vol. 16, p. 249; pl. 4, figs. 3-3b. Austrofromia schultzei H. L. Clark, 1921. Echin. Torres Strait, p. 49. This interesting species is based on a single specimen from False Bay, Cape of Good Hope. Its nearest ally, A. polypora H. L. C., occurs on the southern and western coasts of Australia. No nearly related forms are known from the African coast. * NARDOA VARIOLATA. Asterias variolata Retzius, 1805. Diss. Ast., p. 19. See Linck, 1733, De Stell. Mar., pl. 8, no. 10. Nardoa variolata Gray, 1840, Ann. Mag. Nat. Hist., vol. 6, p. 286. This well-known Indo-Pacific species is recorded by Bell from Mozambique, whence Peters also reported it half a century ago. Curiously enough, it has never been well figured, for Linck’s figures while recognizable are far from good. * LINCKIA LAEVIGATA. Asterias laevigata Linné, 1758. Syst. Nat. ed. 10, p. 662. See H. L. Clark, 19214. Echin. Torres Strait, pls. 9 and 26. Linckia laevigata Nardo, 1834. Oken’s Isis, p. 717. This handsome sea-star, notable for its colour, so unusual among echinoderms, has been reported from Mozambique by both Peters and Bell. The Echinoderm Fauna of South Africa. 277 LINCKIA MULTIFORA. Asterias multifora Lamarck, 1816. Anim. s. Vert., vol. 2, p. 565. Linckia multiforis von Martens, 1866, Arch. f. Naturg., Jhrg. 32, Bd. 4, p. 65. See de Loriol, 1885. Mém. Soc. Phys. Hist. Nat. Genéve, vol. 20) pero: This species is reported by Bell from Mozambique, and two young individuals from the same place, taken by K. H. Barnard in Novem- ber, 1912, lie before me. Some years ago (1908, Bull. M. C. Z. vol. 51, p. 283) I expressed the opinion that multifora could only be considered a variety of laeviyata. Since then I have collected and examined hundreds of laevigata near the Great Barrier Reef of Australia and I find its specific characters are very constant. I am inclined to think therefore that multifora is probably entitled to rank as a valid species, but its characters still need elucidation. * LINCKIA DIPLAX. Ophidiaster diplax Miiller and Troschel, 1842. Syst. Ast., p. 30. Linckia diplax Liitken, 1871. Vid. Med., p. 269. This species is reported by Bell from Mozambique. Its status is dubious. It is very near the species so beautifully figured by de Loriol (1885, Mem. Soc. Phys. Hist. Nat. Genéve, vol. 29, pl. 10) as L. ehrenbergii M. & T., while Ludwig ranks it only as a variety of L. pacifica Gray. In my opinion, pacifica is identical with the West Indian species, guiidingii Gray; at any rate, I have not been able as yet to find any tangible difference between them. The Indo-Pacific Linckias are badly in need of a careful revision based upon fieldwork, as the study of museum material alone proves very unsatisfactory. GANERIIDAE. The presence of this family in this report is due to a very young sea-star which [ am unable to refer to any known genus but which seems to belong in the Ganeriidae. The specimen was sent to Dr. W. K. Fisher for his examination and he writes: ‘My guess would be Cycethra or a close relative. ..... If the tube-feet have true disks, I think the Ganeriidae will be a safe assumption.“ The tube-feet appear to have true sucking-disks, so I am listing the family Ganeriidae in the present report. Most of the members of the family occur in the vicinity of the Straits of Magellan and the Falkland Islands. The important features of the youngster before me may be listed 278 Annals of the South African Museum. as follows: Disk and rays flattened, the general form being distinctly star-shaped but with very obtuse rays. R = 55 mm.; r = 3 mm.; R=18r; br = 3mm. Abactinal plates relatively few, tabulate, with well-spaced, short rough spines; under a magnification of 40 diameters, they thus appear paxilliform. No madreporite is visible but in each interradius is a small, bare, depressed area, covered only by thin skin. Marginal plates 5 on each side of each ray, in each series; all very much alike; they bear short, rough spinelets, well-spaced as on the abactinal plates. Terminal plate short but wide, roughly kidney-shaped, covered with little spinules, like the abactinal plates. Actinal plates small and rather numerous, each with 2—5 (usually 3) rough spinelets similar to those of the abactinal plates but rather longer; the series next to the adambulacrals runs nearly to the tip of the ray and the second runs to the fourth inferomarginal; the remaining two are very short and carry only 4 (or 3) and 2 (or 4) plates respectively. Adambulacral plates very wide and short and very characteristic; the adoral marginal corner extends inward half-way across the furrow, and at the tip curves abrubtly aborally, thus half-way encircling a large tube-foot with a fairly well-developed sucker; on this furrow-projection of the plate are three relatively long, rough spines, of which the middle one is slightly largest and stands at the bend in the plate, another is at the tip of the plate and the third is between the largest and the furrow-margin; on the actinal surface of each plate is an oblique series of 3 or 4 spines the largest being nearest the furrow and farthest from the mouth; the largest is equal to, or a trifle larger than, the one on the bend of the plate, while the smallest is about equal to the spines on the actinal plates. Oral plates of moderate size, flat but distally rather abruptly raised; on each free margin are four spines, the one at the tip of the jaw, much the largest, flat, wide and truncate, the others progressively smaller, more cylindrical and more slender; on the distal angle of each plate are two spines like those on the actinal plates and proximal to them is a single slightly larger spine. P.F. 13240. Cove Rock, near East London, N. 3/, E., 5 miles. 43 fms. St., brk. sh. 4 specimen; very young. I know of no sea-star with the ambulacral furrow guarded as in this specimen and I have little doubt it represents an undescribed genus. But it is conceivable that with growth the adambulacral armature would become more like that of Cycethra, and in any case it seems unwise to base a new genus on so obviously immature a spe- cimen, And in this opinion, I am glad to say, Dr. Fisher fully concurs. a fa? La a ery ae ae) Pat — The Echinoderm Fauna of South Africa. 279 ASTERINIDAE. This is the best represented in South African waters of any of families of sea-stars, although the present collection contains but seven forms. Eleven of those here listed are Asterinas in the wide sense of that term and several of them are very imperfectly known The group was revised by Verrill in 1913 (Amer. Jour. Sci., vol. 35, p. 477) but owing to an unfortunate mistake one or more para- graphs of his “key failed to be printed and as a consequence, it is quite useless. Some of his statements also are very summary and many species are not even mentioned. I have not found it prac- ticable therefore to adopt his proposed new genera, though I have no doubt they are destined to come into use when the numerous spe- cies of Asterinidae are carefully revised. Meanwhile I use Asterina in its old broad sense. I[ regret to have to add two new species and a new variety to this mass of undigested material but there seems to be no other course open. Another new species is a small but interesting Anseropoda. Sladen (1889, CHALLENGER Ast., p. 390) records Asterina gunnii Gray from the Cape of Good Bode but I feel sure this is a mistake and I therefore omit that species from the present list. The fourteen forms included are: separable as follows : . Key to the South African Species of Asterinidae. Body not very flat and thin; r = 1-25—2 v.d. at center of disk. R=18r or more, usually more than 27. Actinal intermediate plates, at least near mouth, each with 5 or more spines. Abactinal plates not imbricated, covered with spines Parasterina bellula. Abactinal plates more or less imbricated. Actinal intermediate plates, each with a cluster of 8—15 spines Asterina peneillaris. Actinal intermediate plates, each with 5—11 spines in a single, or rarely double, transverse series. Abactinal plates closely covered with minute, crowded spinelets . : : . Asterina granifera. Abactinal plates relatively bare, the spinelets scattered, frequently marginal or in a single transverse series Asterina granifera var. sporacantha. Actinal intermediate plates, each with 1—4 spines. Abactinal plates with 5 or more spines and often in addition a tuft of 2—4 stouter spines having a common base Asterina coronata. Abactinal plates not as above . : . Asterina burtonir. R = 1-256—1°8 7, only very rarely 27. 280 Annals of the South African Museum. Adambulacral spines 2 (or sometimes 3). No big subambulacral ne actinal intermediate plates with 3—6 spinelets . : 3 : : Asterina coccinea. A big subambulacral spine on the surface of each adambulacral plate; actinal intermediate plates with only 1 or 2 spinelets. Abactinal spinulation, granuliform. Abactinal granules coarse; many actinal intermediate plates with 2 spines each; subambulacral spine very large, blunt or truncate : : Asterina dyscrita. Abactinal granules fetes fine; actinal intermediate plates nearly always with 1 spine each; subambulacral spine not disproportionately big, pointed. ; Asterina exigua. Abactinal spinulation spiniform, the spinelets rather long but stout and blunt; actinal intermediate plates usually with 1 spine each ; subambulacral spine very large, blunt or truncate Asterina calcarata. Adambulacral spines 3 or 4, with 2 or more spines on the surface of each plate. Abactinal spinelets thick, blunt, crowded; actinal spines relatively long, blunt ; : Asterina liideritziana. Abactinal spinelets short, delreater sharp, well-spaced; actinal spinelets very similar : : Asterina gracilispina. Body very flat and thin; r= 25—5v. ue at Coniot of disk. Rays 9 : : : : : . Anseropoda novemradiata. Rays 5 J : : ; : Anseropoda habracantha. * PARASTERINA BELLULA. Patiria bellula Sladen, 1889. CHALLENGER Ast., p. 385; pl. 63, Jie 4, 2. The original specimens of this species were taken by the CHALLENGER in shallow water, Simons Bay, Cape of Good Hope. So far as I know it has not been met with since except by the ScoriA which took one specimen in Saldanha Bay. Fisher (1908. Smiths. Misc. Coll., vol. 52, p. 90) called attention to the error in using the generic term Patiria and suggested Parasterina, but he did not publish the combination of the latter name with bellula. Sladen emphasizes the non-imbri- cation of the abactinal plates, using that as the one distinctive cha- racter in his key. Fisher does the same in his key to the genera of Asterinidae (1911, Bull. 76 U.S. Nat. Mus., p. 253) and as I have never seen an authentic specimen of Parasterina, 1 can only follow in the steps of these eminent predecessors. I may add however that I am not convinced of the great importance of imbrication as a generic character; for the degree of imbrication is subject to indivi- dual diversity, especially in the long-rayed