red eet, ea es ee et ae: ee ee Pe ae Tein PEPER | RHE eheleie| 5 | ENS POL, teal pi Se 4 Oe Bagh At 3 SH eek oat ae ye! a aN HT ae Seat ee i recon aa oo mee eee Pe) at SS = O eee! ‘S ete a alee Z i c cart y 0 3 — 3 ‘y = Pe 7 2p ye aa ii S Ht wie ees SS. eg Se SRS, _— ee fits Opes fe | OR a ey de f a o a, i" eS (Pesta pred Pipers cor Ces ei, thee yee aes i ch ehey any if Y i Der) Ae spat aed Da AP a oat ait war Fiata) iy) isis = ca iia MG i ny MN "Ni Wt Dan pts oh att ese actgubb bi dara, ANNALS OF THE SOUTH AFRICAN MUSEUM VOLUME XLVI With the Compliments of The Director IMPERIAL WaR MUSEUM, LAMBETH RoOaD, LONDON, S.E.1. ANNALS OF THE SOUTH AFRICAN MUSEUM VOLUME XLVI PRINTED FOR THE TRUSTEES OF THE SOUTH AFRICAN MUSEUM 1961-1963 @ PRINTED IN SOUTH AFRICA BY THE RUSTIGA PRESS, PTY., LTD., WYNBERG, CAPE Fis! OF CON TEN ES BARNARD, K. H. Deep sea Mollusca from west of Cape Point, South Africa (published April 1963) Boonstra, L. D. The dentition of the titanosuchian dinocephalians (published January 1962) Crompton, A. W. On the dentition and tooth replacement in two bauriamorph reptiles (published February 1962) Crompton, A. W. Tooth replacement in the cynodont Thrinaxodon (published September 1963) Day, J. HH. Polychaete fauna of South Africa: part 7. Species from depths between 1,000 and 3,330 metres west of Cape Town (published March 1963) FREEDMAN, L. New cercopithecoid fossils, including a new species, from Taung, Cape Province, South Africa (published April 1961) GRINDLEY, J. R. The Pseudodiaptomidae (Copepoda: Calanoida) of Southern African waters, including a new species, Pseudodiaptomus chartert (published March, 1963) Haropine, J. P. Some South African Cladocera collected by Dr. A. D. Harrison (published April 1961) Eiesse, A. J. Supplementary contributions to the revision of the Bombyliidae (Diptera) of Southern Africa: the genus Systropus (published March 1963) Mitiar, R. H. Further descriptions of South African ascidians (published January 1962) Mitxiarp, N. A. H. The Hydrozoa of the south and west coasts of South Africa. Part I. The Plumulariidae (published June 1962) scorr, K. M. F. Some new caddis flies (Trichoptera) from the western Cape Province—III (pub- lished April 1961) Sseorr, K. M. F. Some Ecnominae from the Transvaal and South West Africa (Trichoptera: Psychomyidae) (published June 1963) Secon, i. M.-F. Some new caddis flies (Trichoptera) from the western Cape Province—IV. Some Hydroptiliidae (published June 1963) SmiTH, J. L. B. Nomenclatorial change for a long-known South African fish (published February 1962) SOUTHWARD, E. C. Pogonophora from South Africa (published April 1961) Stock, J. H. South African deep sea Pycnogonida, with descriptions of five new species (published March 1963) TREWAvAS, E. A new cichlid fish from the Limpopo Basin (published June 1961) Unmann, E. Dactylispa capicola (Péringuey) und Verwandte. (203. Beitrag zur Kenninis der Hispinae (Coleoptera, Chrysomelidae) ) (published February 1962) WINTERBOTTOM, J. M. The South African sukspecies of the buffy pipit, Anthus vcelensis Shelley (published March 1963) 469 ha Pes ‘ ERR CIA a 9 Rh CY arma Fg alc x : Mies Ny Pa PALO i esting AOTC RET nek ie nee ris Las wae BALD S92 * NEW SUB-FAMILY PROPOSED IN THIS VOLUME Halopterinae Millard, 1962 (Plumulariidae) 268, 272 NEW GENERIC NAMES PROPOSED IN THIS VOLUME Chetia Trewavas, 1961 (Cichlidae), 53 Corhiza Millard, 1962 (Plumulariidae), 272, 275 Pantopipetta Stock, 1963 (Colossendeidae), 334 INDEX TO GENERA AND SUBGENERA (SYNONYMS IN ITALICS) A Abra, 413, 418, 450 Abyssochrysos, 410, 414, 417, 439 Acladia, 268, 286, 289 Acropterus, 35, 44 Admete, 4.00, 415, 429 Aglaophamus, 354, 362 Aglaophamus, 362 Aglaophenia, 273, 274, 303 Aglaophenia, 269, 312, 313, 315 Agnesia, 174, 218 Agnosaurus, 101 Alloeocarpa, 182, 218 Alona, 35, 44 Amaroucium, 119, 120, 122, 124, 125, 130 Amphitrite, 354, 368 Anabolia, 18 Anoplodactylus, 321, 340 Antenella, 269, 270, 273, 274, 281 Antennopsis, 271, 318 Antennopsis, 275, 281 Antennularia, 272 Antennularia, 299 Anthus, 341 Aphrodita, 354, 357 Aplidium, 119, 139, 217, 218 Archaeosuchus, 69 Arg yrobothrus, 476 Aricidea, 354, 364 Ascidia, 119, 172, 217, 218 Ascidia, 171 Ascorhynchus, 321, 323 Astyris, 434 Athripsodes, 15 Atlanta, 411, 415 Australopithecus, 1, 8, 14 B Basilissa, 411, 414, 441 Bauria, 231, 232, 233, 241, 249 Bauroides, 232 Belonogaster, 397, 398 Boltenia, 119, 198, 218 Bosmina, 35, 43 Botrylloides, 177, 217 Botrylloides, 175 Botryllus, 119, 175, 218 Botryllus, 177 C Cadulus, 412, 413, 446 Calliostoma, 411, 416, 442 Calliotropis, 410, 411, 414, 415, 416, 417, 440 Camptocercus, 35, 44 Cancellaria, 409, 417, 429 Captorhinus, 253 Cardiapoda, 411, 414 Cavolinia, 411, 413, 414, 415, 416, 417 Cercopithecoides, 4. Ceriodaphnia, 35, 42 Cerithiella, 410, 414, 439 Charitodoron, 409, 413, 414, 415, 416, 417, 432 Chauvetia, 433 Chetia, 53 Chrysodomus, 433 Chydorus, 35, 45 Ciona, 171, 218 Cladocarpus, 273, 274, 304 Clathurella, 428 Clathurella, 426 Clavatula, 408, 413, 414, 415, 416, 417, 420 Clavelina, 139, 218 Cnemidocarpa, 119, 187, 218 Colella, 162 Colossendeis, 321, 326 Columbarium, 410, 413, 416, 438 Columella, 434 Conops, 395 Corella, 174, 218 Corhiza, 272, 273, 275, 291, 318 Cricodon, 516 Cryptodon, 450 Ctenicella, 203 Cuspidaria, 413, 415, 418, 450 Cuvieria, 411 Cynthia, 190, 193 Cynthiopsis, 195 Cystodites, 141, 14.7, 217 Cythara, 408, 414, 415, 416, 417, 422 D Dactylispa, 223 Daphnella, 409, 413, 415, 427 Daphnia, 35, 36 Dentalium, 411, 412, 413, 414, 415, 416, 417, 418, 444 Deuterosaurus, 60 Dextrocarpa, 180, 218 Diacria, 411, 414, 417 Diadectes, 252 Diademodon, 233, 251, 252, 253, 480, 502, 515, 519 Diaphanosoma, 35 Diarthrognathus, 231, 251 Didemnum, 113, 119, 163 INDEX TO GENERA AND SUBGENERA Dimetrodon, 232, 233 Diplosoma, 167, 171, 218 Distaplia, 149, 218 Distoma, 143, 147, 154, 157, 158, 160 Doridoxa, 411, 416, 4.43 E Echinisca, 35, 44. Ecnomus, 454 Epidiopatra, 363 Epizoanthus, 436 Ericiolacerta, 231, 232, 233, 234, 249, 480, 482 Eucoelium, 145 Eudistoma, 113, 119, 139, 154, 155, 217 Eudistoma, 162 Eugyra, 119, 206 Eunoe, 354, 359 Euphione, 354, 355 Euryalona, 35, 44 Eurycercus, 35, 44 Euthelepus, 370 F Falsilunatia, 410, 415, 439 Fluxina, 441 G Gamaster, 209 Gastropteron, 411, 416, 417 Gattya, 268, 270, 273, 281, 295 Goodsiria, 178 Guivillea, 408, 409, 415, 416, 417, 429 Gymnobela, 409, 413, 416, 428 Gynandrocarpa, 178, 218 Ja Halicardia, 413, 450 Halicordia, 450 Halicornaria, 273, 274, 307 Halocynthia, 119, 201 Halopteris, 268, 269, 270, 273, 280, 281, 282, 295, 303 Haplochromis, 53, 55 Harmothoe, 354, 357 Helicopsyche, 23 Heptabrachia, 48 Hermonia, 354 Herse, 411, 417 Heterocalanus, 374 Heteroplon, 267, 268 Heteroplon, 285, 286 Heterotheca, 267, 268 Heterotrema, 128, 130 Hispa, 223, 227 Hyalinoecia, 354, 363 Hydroptila, 475 Ilyocryptus, 35, 43 4) Jonkeria, 76 K Kirchenpaueria, 273, 274, 292 Kirchenpaueria, 295 L Labrus, 258 Lachesis, 433 Laetmonice, 354, 355 Lamellaria, 410, 415, 439 Lamellisabella, 47 Lechia, 412, 414, 446 Leanira, 354, 360 Leda, 412, 414, 416, 417, 448 Leptecho, 15, 23, 32 Leptoclinides, 167, 170, 217 Leptoclinum, 171 Leydigia, 35, 45 Lichnochromis, 54 Limopsis, 413, 414, 415, 416, 417, 418, 449 Lissoclinum, 119, 168 Lumbriclymene, 354, 366 Lumbrinereis, 354, 363 Lutjanus, 258 Lytocarpus, 273, 274, 312 M Macellicephala, 354, 355 Macroclinum, 135 Macrothrix, 35, 43 Maldanella, 354, 355, 365 Malletia, 412, 416, 417, 418, 447 Mangilia, 409, 413, 415, 416, 427, 428 Melinodon, 232 Metandrocarpa, 119, 183 Metrocarpa, 177 Microcosmus, 196, 217, 218 Microgomphodon, 232 Microhelodon, 232 Mohnia, 433 Moina, 35, 43 Molgula, 119, 203, 211, 218 Moniliopsis, 408, 415, 417, 421 Monoculus, 43 Monospilus, 35, 45 Monostaechas, 269, 270, 272, 273, 278, 280, 2Q1 Monotheca, 301 N INemertesia, 271,273,274, 202. 207 Neanthes, 354, 360 Neosabellides, 354, 367 Nephthys, 354, 362 Neptunea, 409, 415, 416, 417, 432 Nereis, 354, 360 INDEX TO GENERA AND SUBGENERA Nothosaurus, 252 Notomastus, 354 Nucula, 412, 418, 446 Nux, 409, 414, 416, 432 Nymphon, 321, 322 O Octopus, 412, 413, 414, 416, 417 Oligokyphus, 233, 518 Odcorys, 410, 414, 415, 416, 417, 439 Orthotrichia, 469, 470 Orthotrichiella, 474 Oswaldella, 273, 274, 2905 Oxyethira, 469, 476 IP Pantopipetta, 321, 334 Papio, I Paragattya, 270, 281, 318 Paraonis, 354, 363 Parapapio, 2 Parapapio, 1 Parascapanodon, 69 Pista, 354, 369 Pipetta, 336 Phalacrostemma, 354, 355, 367 Philbertia, 409, 415, 426 Physocephala, 395 Pleisiosaurus, 252 Pleurotoma, 423 Pleurotomella, 427, 429 Pleuroxus, 35, 45 Plumularia, 268, 269, 273, 274, 285, 286, 289, 293, 300 Plumularia, 285, 286, 292 Poecilochaetus, 354, 365 Polyandrocarpa, 180, 186, 217 Polycarpa, 186, 217 Polycarpa, 186 Polycitor, 145, 147, 155 Polycitor, 160, 162 Polycitorella, 119, 143 Polyclinum, 113, 135, 217 Polynices, 410, 413, 414, 416, 417, 439 Polysyncraton, 119, 165 Pomadasys, 257 Prionium, 16, 18, 30 Prionospio, 354, 365 Pristipoma, 258 Pronucula, 446 Pronucula, 446 Prosipho, 409, 415, 416, 417, 433 Psammaplidium, 124, 127 Pseudodiaptomus, 373 Pseudodistoma, 138, 217 Pseudodromia, 143 Pycnotheca, 273 Pyrene, 410, 415, 417 Pyura, 193, 218 S Sabellides, 367, 368 Sarepta, 412, 418, 449 Scala, 410, 415, 439 Scalenodon, 233, 480, 502, 518 Scaloposaurus, 250, 251 Scapanodon, 63, 69 Schaphander, 411, 414, 415, 416, 417 Schapholeberis, 35, 40 Schizotrichia, 268, 269, 270, 273, 292 Schmackeria, 374, 381, 387 Scirpus, 472, 473, 474 Serranochromis, 53, 55, 56 Serranochromis, 54 Sertularia, 272 Sesamodon, 232 Setodes, 15, 27 Sigillina, 119, 152 Silicula, 448 Simocephalus, 35, 41 Stmosa, 41 Sipho, 433 Solariella, 411, 414, 442, 443 Solariella, 440 Spergo, 428 Spiophanes, 354, 365 Stomazoa, 139 Streblosoma, 354, 369 Strepsiceros, I Struthiocephalus, 103 Styela, 119, 190, 217, 218 Surcula, 408, 413, 421, 428 Surcula, 423 Sycozoa, 147, 218 Symplegma, 217 Syndesmya, 450 Synoicum, 119, 131 Systropus, 393 ar Terebellides, 354, 371 Terebra, 408, 417, 418 Terebratula, 413 Tethyum, 187, 190 Tetrazona, 119, 146 Thalassocyon, 410, 414, 415, 416, 439 Thecocarpus, 273, 274, 310, 312, 318 Thecocaulus, 268, 269, 270 Thecocaulus, 286 Thrinaxodon, 232, 239, 252, 254, 479 Thyasira, 413, 450 — Tribolodon, 232, 252 Trididemnum, 170, 217 Trirachodon, 516 Trochus, 4.43 Trophon, 410, 413, 414, 415, 416, 417, 435 Turbonilla, 410, 413, 439 Turris, 418, 419 Turris, 420 INDEX TO GENERA AND SUBGENERA Typhlomangelia, 408, 416, 422 Typhlosyrinx, 409, 414, 416, 417, 423 Watsoniella, 232 U Weismanella, 374 ill Ulemosaurus, 60 Wyvillea, 429 V Ventromma, 295 Yoldia, 448 foo e ANNALS OF THE Pew lr AFRICAN MUSEUM VOLUME XLVI PART I NEW CERCOPITHECOID FOSSILS, INCLUDING A NEW SPECIES, FROM TAUNG, CAPE PROVINCE, SOUTH AFRICA By LEONARD FREEDMAN Department of Anatomy, University of the Witwatersrand, Johannesburg ISSUED” APRIL 1061 PRICE 36c¢ PRINTED FOR THE TRUSTEES OF THE SOUTH AFRICAN MUSEUM BY THE RUSTICA PRESS (PTY.) LIMITED, COURT ROAD, WYNBERG, CAPE INSTRUCTIONS TO AUTHORS MSS. submitted for publication must be typewritten, double spaced with good - margins, typed on one side of sheet only and sheets numbered. Two MSS. and one set of illustrations must be submitted. Papers must include a Table of Contents and a Summary containing a succinct account of the subject, results obtained and conclusions. The position of the text-figures and tables must be clearly indicated. ILLUSTRATIONS. Please must be kept to a minimum and made up to occupy 5” x 7” (full plate excluding captions) or in direct proportion to these measure- ments if they are to be reduced. A scale (metric system) must appear with all photographs. : REFERENCES. Harvard system—authors’ names and dates of publication given in the body of the text; references arranged at the end of the paper in alphabetical order of authors’ names. Bibliographical arrangement of references at the end of the paper must give: 1. Name of author, followed by his initials. Names of joint authors connected by ampersand (&), and not ‘“and”’. 2. Year of publication. If several papers by the same author in one year are cited, suffixes a, b, etc., are used. 3. Full title of the paper. Initial capital letters only for the first word and for proper names, except in German, where the usage of the language is followed. 4, Title of the journal, abbreviated according to the World list of scientific periodicals, and underlined to indicate italics. 5. Series number, if any, in parenthesis, e.g. (3), (n-s.), (N-F-)2 Be 6. Volume number in arabic numerals (without prefix ‘vol.’), with wavy underlining to indicate black type. 7. Part number, only if the separate parts of one volume are independently numbered. 8. Page numbers, first and last, in arabic numerals (without prefix ‘p’). When reference is made to a separate book, give in the order listed: Author’s name; his initials; date of publication; title, in italics; edition, if any; volume number, if any, in black type arabic numerals; pagination, if reference is to part of book only; place of publication; name of publisher. When reference is made to a paper forming a distinct part of another book or collection, give: | Name of author of paper; his initials; date of publication; title of paper; ‘In’, italicized; name of author of book; his initials; title of book, in italics; edition, if any; volume number, if any, in black type arabic numerals; pagination of paper; place of publication; name of publisher. [| CONTINUED ON. INSIDE BACK COVER NEW CERCOPITHECOID FOSSILS, INCLUDING A NEW SPECIES, FROM TAUNG, CAPE PROVINCE, SOUTH AFRICA By LEONARD FREEDMAN Department of Anatomy, University of the Witwatersrand, Johannesburg (With 2 figures in the text and plates I-X) INTRODUCTION In the collection of the Department of Anatomy, University of Cape Town, Professor L. H. Wells recently found a number of blocks of breccia and partially developed fossils from ‘Taung, Cape Province, South Africa. These had been collected by Professor M. R. Drennan in 1925, a few months after the discovery of the type skull of Australopithecus africanus, and are almost certainly derived from the same cave deposit as the Australopithecus skull. Professor Wells arranged for some of this material to be further developed by Mr. J. W. Kitching of the Bernard Price Institute for Palaeontological Research, Johannesburg. The only identifiable non-primate fossil recovered from this material by Professor Wells is a juvenile lower jaw of a large antelope most nearly resembling the kudu (Strepsiceros strepsiceros). Five cercopithecoid specimens from this collection which have been transferred to the South African Museum, Cape Town, have been handed on to me for description. These constitute a significant addition to the cercopithecoid material available from Taung. The new cercopithecoid specimens comprise a fairly complete skull, with some upper teeth, of a female (S.A.M.11728); most of the right half of a juvenile skull, with two molars (S.A.M.11729); the damaged lower part of a calvaria and the left posterior part of a palate and maxilla, with one molar, of an individual of unknown sex (S.A.M.11730) ; a bilaterally crushed but almost complete immature female mandible with teeth (S.A.M.11731); three upper incisors in a small anterior fragment of premaxilla (S.A.M.11732). Two of these specimens (S.A.M.11728 and S.A.M.11730) appear to represent a hitherto unrecorded species. A reconsideration of the material from Taung previously studied (Freedman, 1957) has led to the conclusion that two other specimens (T.10 and T.13, Transvaal Museum, Pretoria), may now be assigned to this species. SPECIMENS S.A.M. 11728 anp S.A.M. 11730 Papio wellsi sp. nov. Parapapio antiquus (non Haughton) Freedman, 1957 (partim: T.10). p. 174. Papio izodi (non Gear) Freedman, 1957 (partim: T.i3). p. 180, pl. 27, fig 56. Holotype. A fairly complete female skull (S.A.M.11728), with only the v right P§, P4, M? and M3 present. I Ann. S. Afr. Mus. 46 (1), 1-14, pls. 1-10, 2 figs. BTN UN? 06T 2 ANNALS OF THE SOUTH AFRICAN MUSEUM Repository of type. South African Museum, Cape Town, South Africa. Type site. Limestone Deposits, ‘Taung, Cape Province, South Africa. Referred Material. ‘The following specimens from Taung are referred to P. wells: Specimen T.13: The left half of a skull with M1-M®, and also the right M® in a small piece of maxilla, probably of a female. Specimen T.10: A damaged calvaria and posterior part of the muzzle, with the left P*-M? and right M?, of a juvenile—probably male. Specimen S.A.M.11730: The badly damaged base of a calvaria and left maxilla with a molar (probably M?) of unknown sex but most likely a male. Diagnosis. P. wellst is a small species of Papio which in norma lateralis shows a deep concavity between glabella and rhinion. The muzzle is characterized by fairly well-developed maxillary ridges and large maxillary fossae of variable depth. The calvaria is considerably flattened and there appears to be appreciable sexual dimorphism in skull size. The teeth are of the Papio-Parapapio type; all female teeth are smaller than their male equivalents. M3 shows some reduction of the disto-buccal cusp (metacone). Description. For purposes of description, skulls have been oriented so that the occlusal plane is horizontal. The type specimen of P. wells: consists of a fairly complete female skull of a small baboon-like species. ‘The skull lacks the anterior tip of the muzzle, the left orbit and both zygomatic arches. Also, the muzzle is bent at an angle to the calvaria. Much bone has been lost from the calvaria—particularly the posterior, left parietal and basal portions—but a good endocranial cast has been exposed in most of these areas. From the almost unworn state of M?, it would appear as if this tooth had only just come into functional position and the individual had thus only just become fully adult. ‘The other good female specimen, T.13, consists of most of the left half of the skull but lacks the anterior tip of the muzzle, most of the nasal bones and the top of the calvaria. Viewed in profile (Plate I) the muzzle of S.A.M.11728 is seen to be elongated and of about the same length as the calvaria. The region between glabella and rhinion forms a deep concavity and the reduction in height is particularly rapid in the inter-orbital region. The portion of the nasal profile just posterior to the nasal aperture is almost horizontal and slopes inferiorly only very slightly. Specimen T.13 (Plate II) has lost most of both nasal bones, but it seems clear that the profile shape was very similar to that of the type specimen. The muzzle dorsum of the type specimen is short antero-posteriorly and slopes down steeply from either side of the rounded nasal bones to the maxillary ridges. From these ridges, which are not particularly well developed in this specimen, the maxillae slope down almost vertically to the alveolar margin, forming large but only slightly excavated bilateral maxillary fossae. In T.13 NEW CERCOPITHECOID FOSSILS INCLUDING A NEW SPECIES 3 the shape of the muzzle dorsum is similar to that of the type but, on the sides of the muzzle, there are very large, deeply excavated maxillary fossae. Both $.A.M.11728 and T.13 present fairly large orbits. The supra-orbital tori are only slightly developed in the type specimen but well developed in T.13. In $.A.M.11728, the palate is relatively long and narrow, with fairly deep sides and a flat roof (Plate III) ; the posterior palatine foramina are large. The palate in T.13 is damaged but its shape appears similar to that of the type (Plate IV). The calvaria of the type specimen is considerably flattened (Plate I). The supra-orbital tori are slightly raised anteriorly and there is a slight ophryonic groove running across the calvaria posterior to them. The post-glenoid process of this specimen is broad mesio-laterally. Anterior to it there is a prominent groove which is deep laterally but becomes shallower medially (Plate III). Anteriorly this groove is bounded by the posterior edge of the inferiorly projecting root of the zygomatic process of the temporal bone. The post-glenoid process, the root of the zygomatic process of the temporal bone and the glenoid fossa are all situated relatively high above the level of the occlusal plane (Plate 1). In T.13, the calvaria also appears to have been flattened (Plate I1) but the glenoid fossa region is not present for comparison with that of $S.A.M.11728. The nasal bones of the probable male, T.10, are considerably damaged but its profile contour (Plate V) must have been of the Papio type (see ‘Discussion’ below) and similar to that of the type specimen. In spite of considerable damage to this general region, the muzzle dorsum would appear to have been flattened and have had almost vertical lateral surfaces. On the left side, the vertical surface of the maxilla appears to have been pushed in, but from the small posterior part of the right side preserved, it seems clear that an excavated maxillary fossa was present. ‘The calvaria is flattened supero- inferiorly and of very similar shape to that of the type specimen. Although immature, T.10 is considerably larger than the females $.A.M.11728 and T.13. The other probable immature male, S.A.M.11730, has only a small portion of the palate and left maxilla present (Plate VI) but the very deep maxillary excavation and the tooth size (see ‘Discussion’ below) suggest that it belongs in this species. In the females $.A.M.11728 and T.13 the dental arcade is horseshoe- shaped but elongated antero-posteriorly (Plates III and IV); the male shape cannot be determined. The male P* of T.10 (Plate VII) is very well preserved and almost unworn. It is a typical bicuspid cercopithecoid premolar and has a faint disto-lingual vertical groove. The two female premolars of the type specimen are considerably worn and damaged and little can be seen of their structure. The male and female molars are all typical bilophodont teeth of the Papio-Parapapio type and several have grooves marking off mesio-lingual cuspules (Plates III, IV and VII). The disto-buccal cusps of M? on both female specimens show signs of reduction. (No male M? is known). On size 4. ANNALS OF THE SOUTH AFRICAN MUSEUM (table 3) the molars of the female, T.13, are considerably larger than those of the other female, S.A.M.11728. Nevertheless, it is apparent that the female teeth of P. wellsi are considerably smaller than those of the males (table 2), the difference being of the order usually found in the genus Papio. The question of possible mandibular fragments and teeth of this species is dealt with in the ‘Discussion’ below. Measurements. The measurements of the female skull and the male and female upper teeth of this species are given in tables 1, 2 and 3 respectively. Discussion. In previous studies two baboon-like species, Parapapio antiquus (Haughton) and Papio izodi Gear, and possibly a third, Parapapio jonesi Broom, have been recorded from Taung, together with a fourth species of considerably different character, Cercopithecoides williamsi Mollet. The material assigned to P. wellsi must therefore be compared with the material which has been referred to the first three of these species (Freedman, 1957). The two known male individuals of P. wells: are immature as well as damaged and fragmentary. They appear to agree with the females in general | . Be ms. j { % 5S ems. J FiGeas Ries FicureE 1. Pantograph tracings of the muzzle profile Figure 2. Pantograph tracings of the shapes of females of (from above downwards): muzzle coronal section shapes (taken Parapapio antiquus (Tvl.639), Papio izodi (AD.g92), midway between glabella and rhinion) Papio wellsi (S.A.M.11728) and Parapapio jonesi of females of (from above downwards) : (S/ES-565). (Xa. Papio izodi (AD.992), Parapapio jonest (STS.568), Parapapio antiquus (Tvl.639) and Papio wellsi (S.A.M.11728). X I. NEW CERCOPITHECOID FOSSILS INCLUDING A NEW SPECIES 5 morphology but (as described above) they differ in size to the degree usual in Papio. Insufficient is thus known of them—and also of the males of some of the species referred to— to make comparisons between the males of value at present. The comparisons made below between P. wells: and other known cercopithecoid species will thus depend mainly on comparisons between females. Both Parapapio antiquus and Papio 1zodi are larger than P. wells in most skull and some dental measurements (tables 1 and 3), P. zzodi being the largest. In morphology P. wellst shows important resemblances to both P. antiquus and P. izodi. Thus, in profile shape of the muzzle P. wells: is clearly of the Papio type, with a steep drop in the inter-orbital region similar to that seen in P. izodi (fig. 1). On the other hand, the muzzle shape of the new species in coronal section (fig. 2) is very similar to that of P. antiquus. P. wellsi and P. antiquus both have definite maxillary ridges and maxillary fossae, whereas neither is present in P. 2zodi. The new species differs from both P. izod: and P. antiquus in its generally smaller skull and dental size, its flatter calvaria and the shape and situation of the structures in the glenoid fossa region as described above. On muzzle profile shape and sexual dimorphism the group of specimens referred to P. wellsi (S.A.M.11728, T.13, T.10 and S.A.M.11730) must clearly be placed in the genus Papio. However, to suggest including them in the species P. 1zodi would be to increase the range of variation, particularly morphological, beyond reasonable limits. P. wellsi is similar to Parapapio jonesi in skull size (table 1) and in muzzle coronal section shape (fig. 2). However, it differs considerably from that species in muzzle profile shape (fig. 1) and also in the shape and relative height of the calvaria. In the teeth (table 3), the morphology of the two species is similar but on size, particularly in the length dimensions, the P. wells1 molar measurements are considerably larger. Before this new species was established, T.13 was referred to P. tzodt (Freedman, 1957). The presence, in a single specimen, of very large, deep maxillary fossae and a slightly different muzzle dorsum shape, were not felt to be sufficient grounds for the creation of a new species. Despite the size and morphological differences from the type specimen, described above, it now appears that this specimen is a female of the new species, P. wellsi. The main difference between T’.13 and the type specimen of P. wellsi lies in the size and depth of the maxillary fossae. In a large number of skulls of Papzo ursinus and Papio cynocephalus which were examined for this feature, maxillary fossae were found to be of very variable size and depth. In that same study (Freedman, 1957) it was also suggested that the muzzle shape of T.10, in profile, was probably that of a Parapapio. The profile shape appeared to be of the Papio type but the inter-orbital region was considerably damaged and the individual from which it came, immature. On the shape of the muzzle in coronal section this specimen agreed with the Parapapio described from Taung (P. antiquus) and disagreed with P. 1zodi, the only Papio then known from Taung. It was thus decided that the shape in an undamaged adult of this 6 ANNALS OF THE SOUTH AFRICAN MUSEUM type might well be similar to that of a Parapapio and the specimen was referred to P. antiquus. The new species now being described is a Papio but has a muzzle which, in coronal section, is similar to that of Parapapio antiquus. The apparent contradiction of profile and coronal section muzzle shapes is now resolved and T.10 can, with reasonable certainty, be referred to P. wellsi. On the size of the skull and teeth the specimen seems clearly to be from a male. TABLE I. Comparison of the skull measurements (in mm.) of Papio wellsi females with those of similar female cercopithecoid specimens.* Para- Papio Papio papio | Parapapio Papio angust- wellsa jonest antiquus ~zodt ceps S.A.M. STS vn 11728] T.13 | 565 | 639 | T.17 |AD.992/AD.946/K.A.194 General Greatest length | (134) | (130) | (135) | (143) (138 ) 139 Basal length (QI) 99 Calvaria Height Basion-bregma 54 58 60 (64) Basion-glabella 64 70 72 Breadth Greatest temporal | (65) 68 69 72 67 Length Inion-glabella (82) (88) | (go) (go) Inion-basion (50) 47 53 48 Foramen magnum Length (18) Muzzle Height Anterior to P? 16 22 20 17 a0 Breadth Anterior to M?| 43 A5 46 46 42 47 45 Anterior to P? | 29.5 35 36 Dorsal to M? 2 30 35 a1 Length Muzzle (67) (65) 78 78 Palate (58) 51 63) (62) Orbit Inter-orbital (9) 10 8 Nasal Breadth 15.5 ig 15.4 * The comparative measurements in this table, and also those in the succeeding tables, are all from Freedman (1957). Professor L. H. Wells has pointed out to me that the figure for the muzzle length of AD.946 was erroneously given in that study as 88 mm. instead of 78 mm. NEW CERCOPITHECOID FOSSILS INCLUDING A NEW SPECIES Vi The new specimen, S.A.M.11730 is only a small maxillary fragment (Plate VI). However, the deeply excavated maxillary fossa is strongly suggestive of P. wellsi and, when the single M? is compared to its equivalents in tables 2 and 3, it is again seen to fit well as a male in that species. Because of the resemblances and the small number of specimens of the four baboon-like species from Taung, and particularly with the overlap of characters between P. wellsi, and the previously described species Parapapio TABLE 2. Comparison of the dimensions (in mm.) of the upper teeth of Papio wellst males with those of similar male cercopithecoid specimens. (no males known) Papio angusticeps Range* * The ranges given in this and the succeeding tables are based on very few (mostly 3-6) specimens. jonesi, Parapapio antiquus and Papio izodt, it is difficult to identify positively a number of specimens. For example, C.T.5360 and T.22, previously referred to P. antiquus (Freedman, 1957), consist of upper teeth only and could, on size, equally well belong in P. wellsi. Similarly, T.16 was previously described as a female P. antiquus and could now also be referred to P. wellsi. The possibility of even C.T.5356, previously considered a male of P. antiquus, being a badly distorted male of P. wells cannot now be ruled out completely. The distal reduction of M? of this specimen is mainly on the buccal side and very similar to that seen in the female P. wellsi specimens. If this specimen were referred to P. wellsi, this would resolve the apparently anomalous situation of a female P. antiquus having an M2? very considerably reduced 8 ANNALS OF THE SOUTH AFRICAN MUSEUM distally, and, the male of the same species having the distal portion of that tooth only slightly reduced. However, it does not appear as if the specimen had a Papio-like muzzle, although this region is badly distorted and damaged, possibly sufficiently so to be misleading. Further, the diagnostic features of the above mentioned four species lie mainly in the skull and few specimens are known which include diagnosable skull fragments associated with mandibles with teeth. In particular, no good P. wellsi skull is associated with lower teeth. Hence, with the description of the new species P. wellsi, the mandibles referred to P. antiquus (T.12 and T.23— males; T.18, T.21 and T.27—females; A.D.944— ? sex) could all now equally well be referred to P. wellsi. AD.943, a male specimen formerly referred to P. antiquus, on tooth size and probable muzzle shape, the latter as deduced from the case of the internal surface, should probably now be referred to P. wellsi. It is therefore now apparent that the status of a number of specimens from Taung, including almost all of the known mandibles, is best left in abeyance until more material, especially associated skulls and mandibles, is discovered. The presence at Taung of two similar species of Papio might at first glance appear improbable, but it should be recalled that the limestone deposit at this site covers a considerable period of time and the two species might not necessarily have been in the area at the same time. The material described in this paper almost certainly comes from the Australopithecus africanus type site, whereas at least some of the material previously described from Taung came from other parts of the deposit and may thus possibly be of different age. Further, a fairly large number of different fossil cercopithecoids (fifteen species of five genera), dating from roughly the same time period as that covered by the Taung deposit, have been described from Southern Africa. From one of these sites (Sterkfontein), three species of one genus (Parapapio jonesi, P. broomi and P. white) have been recorded. With regard to the other species of Papio described from Southern Africa, P. wellsi shows resemblance to Papio angusticeps (known from Kromdraai and Cooper’s Quarry, near Krugersdorp, Transvaal, South Africa) in the depth and extent of the maxillary fossae. However, it is a smaller species, has less well-developed maxillary ridges and a flatter calvaria. The teeth of the two species are similar in size and morphology. The new species is very considerably smaller than Papio robinson. (known mainly from Swartkrans, Transvaal, South Africa) and Papio ursinus (the extant South African baboon), and also differs from these two species in a number of morphological characters. Of all the cercopithecoid species known from Southern Africa, P. wells: appears to have its closest affinities with P. angusticeps. These two Papio species are of generally similar morphology and are both characterized by large and usually deep maxillary fossae. P. angusticeps is somewhat larger than P. wells but it occurs at sites considered to be of more recent date than Taung. However, P. wells: shows differences even to P. angusticeps, e.g. in its flattened calvaria, which would appear to rule it out as a direct antecedent of that species. NEW CERCOPITHECOID FOSSILS INCLUDING A NEW SPECIES ‘UMOUYUN SI xos SHI fSuney, WOIZ UMOUY sotoeds sty} FO —_—— | | pe | VL 0.4 :ISUv YY c660°'qV 97o'dv : ISULYY g-4 L.9 : ISU YY aV IL, SIT, Lg | g2L41rww's | | Sh J | |e | EO I 0-6 0-31 0-g | &.o1 (S.6) | D.11 Sol, V-O1 1-9 I-OI Pel. 6.6 v.6 V.11 (V-g) | (4-6) Pq wq sW O-I1 | 9-11 V.6 V.OI C.6 C.11 0-6 L.Oo1 z-6 L.o1 ¢.9 5-01 V.6 L.6 G-Ob |eookt 0-6 | (0.01) Pq uq ZN ‘suouitoads prosayjtdoo.199 (2sauol ordvdvivg X98 ¢ pur) Q[eULIJ Ie]IUUIIs JO sso} YIIM sapeuray 27720 o1dvg JO 190} soddn ay} Jo (‘uu UT) sUOIsSUDWAIP 9Y4} Jo UOsIIedUIOT) *& ATAV J, uouttoods ATUO 9} ST SIUT, x sfarusnsup ogog Gaee ougvg snnbyup orgvdvivg asauol orgoqvivg asyjam dog IO ANNALS OF THE SOUTH AFRICAN MUSEUM SPECIMEN S.A.M.11729 Description. This specimen (Plate VIII) consists of the right half of the skull of a juvenile baboon of unknown sex. Most of the nasal and frontal bones are missing and there are a number of fractures and slightly weathered areas on the calvaria and muzzle. The base of the calvaria is considerably damaged but the palate is well preserved. Of the dentition, only the right M? and the buccal half of the right dm? are present. The muzzle dorsum, on each side, is triangular in shape and slightly concave. It slopes down rather steeply from the lateral edge of the rounded nasal to the clearly marked maxillary ridge. From the maxillary ridge, the maxilla slopes down even more steeply to the alveolar margin and a large, slightly excavated, maxillary fossa is formed. Features of note when this specimen is compared with Papio (P. ursinus and P. cynocephalus) and Parapapio (M.3004, Parapapio sp.) specimens of approximately the same dental age are: (1) The size of the orbit is relatively large. (2) The zygomatic arch is considerably better developed. (3) The anterior part of the temporal line is rather well defined. (The posterior part of this line is weathered away.) (4) The mastoid region is well developed. ) The supramastoid crest leads to an occipital crest. (6) The planum occipitale is more vertical than usual. These differences may, in some of the above instances, be more apparent than real as it is very difficult to get comparative specimens of precisely the same age. Considerable time periods elapse between the eruption of successive teeth in the cercopithecoids. However, morphologically the skull of this specimen does give the impression of being robust and adult for its age. The palate of S.A.M.11729 is fairly deep and smoothly concave from the alveolar margin to the midline, which is raised into a crest anteriorly. The posterior palatine foramen is situated in a deep recess. The fragment of dm? appears similar to its equivalents in Papio and Parapapio; the single M? is well preserved, almost unworn and of the Papio-Parapapio type. Measurements. As the specimen is juvenile, no skull measurements were taken. The dimensions of the single M! (in mm.) are: breadth (mesial) —8.5, breadth (distal)—8.0 and length—g.5. Discussion. On the shape in coronal section of the muzzle, S.A.M.11729 is similar tu Papio wells:, Parapapio antiquus and Parapapio jonesi. Reconstructions of the frontal and nasal regions in plasticine to give the probable profile shape did not give wholly satisfactory results, as the damage is so extensive that there is a measure of variation in the shape which can be produced, depending on how far forwards the supra-orbital tori and glabella are produced. However, 2B NEW CERCOPITHECOID FOSSILS INCLUDING A NEW SPECIES Tok the angle at which the remaining small portion of the nasal bones lies, and the shape of the small piece of frontal lying at the supero-lateral corner of the orbit, do seem to favour at least a shallow concave type of nasal profile. The various notable features of $.A.M.11729, to which attention was drawn above, cannot be compared fairly with the equivalents in the three species under consideration. Firstly, usually only female skulls of those species are available and secondly, the known specimens are often damaged in the relevant areas. Further comparative material is necessary for the evaluation of those features in $.A.M.11729. However, the general robustness of the specimen may well be due to the specimen being a juvenile male and may not have any taxonomic significance. The dimensions of the single M! of S.A.M.11729 are similar to those in the males and females of P. jones: (tables 2 and 3). When compared to their equivalents in P. wells: and P. antiquus, the dimensions of this tooth are almost all smaller than those of the females of both of those species (table 3), but, of the two, they are appreciably closer to P. wellst. Compared with the male equivalents in those two species (table 2), the tooth is considerably smaller than those of P. wellsi (in which there is size sexual dimorphism of the molars) but there is no known male M! of P. antiquus for comparison. However, as there is little or no sexual dimorphism in the molars of any of the known Parapapio species, the size difference should be about the same as for the female of that species. It is most difficult to come to any definite conclusion about the affinities of S.A.M.11729. From the muzzle coronal shape, the specimen could belong in P. wellsi, P. antiquus or P. jonest. ‘The reconstruction of the muzzle profile shape is slightly more suggestive of a Papio but P. jonesi does show some concavity in this region and the juvenile shape of Parapapio spp. generally is not well known. The size of the single M! is most similar to the equivalent tooth in P. jonest but, it is also not very much smaller than the equivalent tooth in the P. wellsi female (T.13) and it should be remembered that only a few specimens are involved in this comparison. On balance of evidence, it would appear that S.A.M.11729 might be a juvenile female of P. wells: or, more likely, a juvenile male or possibly female of P. jonesi. However, the provenance of P. jonesi at Taung depends on a single specimen, consisting of a small piece of maxilla and damaged teeth, and the possibility of the specimen belonging even in P. antiquus cannot be ruled out completely. SPECIMEN S.A.M.11731 Description. S.A.M.11731 is an almost complete mandible of an immature female cercopithecoid. The specimen has been fractured slightly to the right of the symphysis, and the right half has been pushed to the left and lies close up against, and slightly above, the left half of the mandible (Plate IX). I,—-M, are present on the left side. The premolars and molars are very IS J, snnbiup QI], odvdviwg asauol rosuey ordodvivg SLI VW'S ANNALS OF THE SOUTH AFRICAN MUSEUM ‘sune], Wo1 UMOUY soroads «e[TUIIS Jo sayeuIay UT s}UdTeAINba Toy? ym “ISLI y's “[qrpuew syeway ay) Jo YyI991 JaMoT 9y} Jo (‘uu UT) SUOISUSUIp 94} Jo uOstIvduIOD “by aTavy, 12 NEW CERCOPITHECOID FOSSILS INCLUDING A NEW SPECIES 13 well preserved but the incisors and canines are cracked and damaged. On the right, the incisors are also present but very badly damaged. The only other tooth present on this side is M, and it is slightly damaged distally. M, has not yet erupted on either side but the tooth can be seen, on both sides, lying in its crypt. The tips of the cusps of that tooth are at about the level of the buccal alveolar margin. Anteriorly, the symphysis slopes down steeply and is flattened. A foramen symphyseosum is present in the midline. The flattened anterior surface of the symphysis is demarcated by a prominent oblique ridge on either side. The body of the mandible is low and thin but shows some thickening posteriorly in the region of M;. A large but very shallow mandibular fossa is present, on the antero-inferior margin of which is the mental foramen. The ramus is about as vertical as that of Papio ursinus but does not rise as high. The gonial angle is inverted. The coronoid process lies slightly above the level of the condyle; the sigmoid notch was probably not very deep. The teeth are all typical of the female Papio and Parapapio and show no unusual features (Plates IX and X). A small mesio-buccal cuspule is present on both M, teeth. Measurements. As the specimen is immature, no measurements of the mandible itself were taken. The dental dimensions and comparisons are given in table 4. Discussion. On molar size, which has often proved a most useful taxonomic character in the Cercopithecoidea, $.A.M.11731 is smaller than the few known female specimens previously referred to P. izodi and P. antiquus, the difference being greater in the former case. In the incisors and premolars, the dimensions of S.A.M.11731 and P. antiquus females are similar but there are no comparable teeth of P. izodi. As compared to the female P. jones: teeth, the incisors of S.A.M.11731 are larger but the premolars and molars are of about the same size. No lower teeth or mandibles have so far been assigned to P. wellsz. It is not possible to assess accurately what the adult dimensions of the mandible of S.A.M.11731 would have been but, as far as can be judged, the size would probably have been about the same as in P. jonesi females, slightly smaller than in P. antiquus females (assuming T.18 and T.27 to have been correctly assigned) and considerably smaller than in P. zzodi females. As concluded above, the Taung mandibles are best left in abeyance until more and better associated skulls and mandibles are recovered. However, it would appear probable that, on size and the presence of a mandibular fossa, the new immature, female, mandibular specimen, $.A.M.11731, might well belong in P. wellst, but it could also reasonably be referred to P. jonest. SPECIMEN S.A.M.11732 This specimen consists of a small portion of the left premaxilla with I? and I? and an associated right I?, almost certainly of the same individual. 14. ANNALS OF THE SOUTH AFRICAN MUSEUM The teeth are of the typical cercopithecoid pattern and show no unusual features. Insufficient upper incisors are known from Taung for the affinities of the specimen to be determined. SUMMARY 1. Five new cercopithecoid fossils are described from Taung. ‘They almost certainly come from the Australopithecus africanus type site. 2. Two of these specimens, plus two specimens previously described from Taung, form the basis of a new species, Papio wellst. 3. Because of an overlap in size and morphological characters between the four baboon-like species recorded from Taung, the identity of several specimens, notably mandibles, cannot at present be determined. ACKNOWLEDGEMENTS My thanks are due to Professor L. H. Wells for handing me this material for description and also for his most helpful criticisms of the manuscript. Professor P. V. Tobias, Anatomy Department, University of the Witwatersrand, Johannesburg, and Dr. J. T. Robinson, Transvaal Museum, Pretoria, also kindly read through the manuscript and offered useful suggestions. Messrs. A. R. Hughes and R. Klomfass, both of the Anatomy Department, University of the Witwatersrand, Johannesburg, took the photographs which illustrate this paper. This paper has been published partly with the aid of a grant from the University Council Research Committee, University of the Witwatersrand. REFERENCES FREEDMAN, L. 1957. Ann. Transv. Mus., 23, 121—262. Ann. S. Afr. Mus., Vol. XLVI Plate I Papio wellsi, female (S.A.M.11728). Skull, lateral view. Ann. S. Afr. Mus., Vol. XLVI Plate II Papio wellsi, female (T.13). Skull, lateral view. iy .' , q Plate III Ann. S. Afr. Mus., Vol. XLVI ] Papio wellsi, female (S.A.M.11728). Skull, basal view Plate IV NMG Ammen os) Adi. Mus... Vol; XacV: lew. asal \ b 3). Skull, elit Papio wellsi, female (° fe weds =“ ——— "MOIA [R19}P] -Olojue UolUSseay =[[Nnyg *(o€ZL11 ‘W'V'S) 9yew s[qeqoad ‘2sjjam o1vg {Al ING “MOIA [PIO] “TTNAS (its L) 2]ew aTqeqoaid “asjjam oidvg A aN | IATX “TPA “Sny “sty Ann. S. Afr. Mus., Vol. XLVI Papio wellsi, probable male (T.10). Skull, basal view. Ann. S. Afr. Mus., Vol. XLVI Plate VII Plate VIII Juvenile cercopithecoid (S.A.M.11729). Skull, lateral view. Anne oe Airs Mus... Vole SaaVI Plate IX Immature female (S.A.M.11731). Mandible, lateral view. Ammweoe Ati Mus, Vol: SV Plate X Immature female (S.A.M.11731). Mandible, occlusal view. References thus appear as follows: AUGENER, H. 1913. Polychaeta. Jn Michaelsen, W., ed. Beitrdge zur Kenntnis der Meeresfauna Westafrikas. 2, 67-625. Hamburg: Friederichsen. EKMAN, S. 1953. 1933 1917-1920 1921 1924-1925 1924-1926 1925-1927 1925-1928 1925-1926 1929-1938 1927-1928 1928 1929 1929-1952 1929-193! L952 s1935 Zoology and Geology (excl. Parts 1-3, 5, 7, 8) Zoology (excl. Part I) Palaeontology .. Geology, Ralacontoléey: Zoology. Anthropology (exc. Parts 1.72), 5.ys09) Zoology (excl. Part I, Index) Palaeontology (excl. Parts 1-3) Botany (excl. Parts 1, 2) Zoology (excl. Part 2) . Zoology (excl. Parts 2, 7, Inde) Palaeontology and Geology Archaeology and Zoology Zoology Zoology Botany Zoology Zoology Zoology Zoology 8 Zoology (excl. Part 1) Palaeontology .. Zoology Anthropology sna Etbnalage (eee: Part aie Zoology Zoology Anthropology Palaeontology .. Zoology Zoology of papers, authors and subject published! in viele: ras XXX 1934-1950 1935-1940 1939 1938 1956 1942-1948 1947-1952 1950 1952 1952-1956 1952-1955 1953-1956 1955-1957 19657-1959 1959-1960 The LIBRARIAN, SoutH AFrican Museum, Care Town. Palaeontology (excl. Part 1) Zoology Zoology Zoology Zoology Zoology Archaeology Zoology Zoology Botany Zoology Palaeontology Zoology and Palaeontology Zoology and Palaeontology Zoology and Palaeontology Copies may be obtained from— R 2.75 4.38 © 8.08 1.97 5.48 2.41 6.03 11.27 6.73 rz 8.55 8.04. 10.26 8.57 8.45 10.33 8.10 6.27 4.33 4-85 4-25 6.00 4.70 3.60 3.60 6.30 5.80 8.88 0.23 6.13 7.62 5-05 5:75 14.40 6.13 4-33 g.00 6.55 3.18 9.60 8.10 4.80 6.64 ii fo 1,68 BNINALS OF THE eOowWIm AFRICAN MUSEUM VOLUME XLVI PART II SOME NEW CADDIS FLIES (TRICHOPTERA) FROM THE WESTERN CAPE PROVINCE — III By Ke Mer. Scorr, Pa. D., F-R.E:S. National Institute for Water Research South African Council for Scientific and Industrial Research ISSUED APRIL 1961 PRICE 33¢ PRINTED FOR THE TRUSTEES OF THE SOUTH AFRICAN MUSEUM BY THE RUSTICA PRESS (PTY.) LIMITED, COURT ROAD, WYNBERG, CAPE INSTRUCTIONS TO AUTHORS MSS. submitted for publication must be typewritten, double spaced with good margins, typed on one side of sheet only and sheets numbered. Two MSS. and one set of illustrations must be submitted. Papers must include a Table of Contents and a Summary containing a succinct account of the subject, results obtained and conclusions. The position of the text-figures and tables must be clearly indicated. ILLUSTRATIONS. Please must be kept to a minimum and made up to occupy 5” x 7” (full plate excluding captions) or in direct proportion to these measure- ments if they are to be reduced. A scale (metric system) must appear with all photographs. REFERENCES. Harvard system—authors’ names and dates of publication given in the body of the text; references arranged at the end of the paper in alphabetical order of authors’ names. Bibliographical arrangement of references at the end of the paper must give: 1. Name of author, followed by his initials. Names of joint authors connected by ampersand (&), and not “and”. 2. Year of publication. If several papers by the same author in one year are cited, suiiixes a, bb; ete., are used. 3. Full title of the paper. Initial capital letters only for the first word and for proper names, except in German, where the usage of the language is followed. 4. Title of the journal, abbreviated according to the World list of scientific periodicals, and underlined to indicate italics. 3. Series number, if any, i parenthesis, e.g: (3), (is.), (N-E2)aGe 6. Volume number in arabic numerals (without prefix ‘vol.’), with wavy underlining to indicate black type. 7. Part number, only if the separate parts of one volume are independently numbered. 8. Page numbers, first and last, in arabic numerals (without prefix ‘p’). When reference is made to a separate book, give in the order listed: Author’s name; his initials; date of. publication; title, in italics; edition, if any; volume number, if any, in black type arabic numerals; pagination, if reference is to part of book only; place of publication; name of publisher. When reference is made to a paper forming a distinct part of another book or collection, give: Name of author of paper; his initials; date of publication; title of paper; ‘In’, italicized; name of author of book; his initials; title of book, in italics; edition, if any; volume number, if any, in black type arabic numerals; pagination of paper; place of publication; name of publisher. [ CONTINUED ON INSIDE BACK COVER SOME NEW CADDIS FLIES (TRICHOPTERA) FROM THE WESTERN CAPE PROVINCE — III By K. M. F. Scott, Pu.D., F.R.E.S. National Institute for Water Research South African Council for Scientific and Industrial Research (With 8 figures in the text) INTRODUCTION This is the third paper in the present series, in which caddis flies from the western Cape are being described. In it an account is given of the larval and pupal stages of the three new species of Leptoceridae described in Part II (Athripsodes prionii, A. bergensis and Leptecho helicotheca); the pupal skin of Athrip- sodes tuckeri (Barnard) ? var. is also described, and all stages of a small species of Setodes, S. barnardi, from the Great Berg River, which appears to be new. The immature stages have been described in considerable detail with the object of collecting sufficient information to build up keys for the identification of larvae and pupae collected in field surveys. The methods employed have been the same as in previous papers, and, as before, holotypes will be deposited in the South African Museum, and paratypes will be sent to the British Museum (Natural History). Terminology as before, except that Ross’s term ‘stridulator’ is used instead of the more cumbersome ‘plaque d’appui’ (Barnard), or ‘process of the trochantin’ (Ulmer). A full descrip- tion of the habitats of the species described will be found in Harrison and Elsworth (1958) and Harrison (1958). (See addendum to the latter paper for names used.) | The author would like to express her grateful thanks to Mr. D. E. Kimmins of the British Museum (Natural History) for much helpful advice, also to Dr. A. D. Harrison of the National Institute for Water Research, and to Professor J. H. Day of the Zoology Department, University of Cape Town, in whose department the work was carried out. The Trustees of the South African Museum gratefully acknowledge a grant towards the cost of publishing this paper made by the S.A. Council for Scientific and Industrial Research. DESCRIPTION OF IMMATURE STAGES Athripsodes prionii Scott Fig. 1 A-R; Fig. 5 K-Q Remarks: Larvae were collected from the Palmiet River at Elgin in October and November, and from the Great Berg River near Assegaibos, Ann. S. Afr. Mus. 46 (2), 15-33, 8 figs. T9 SMT SUNTAN institution YUN 7 Wet 16 ANNALS OF THE SOUTH AFRICAN MUSEUM Driefontein, Groot Drakenstein and Bridgetown between April and October. The only place at which they were common was at Groot Drakenstein in autumn (April and May). The larvae are swimmers and live mainly among the marginal vegetation, particularly the palmiet (Prionium serratum). Larva (Fig. 1 A-M): Described from entire specimens. Case-bearing swim- ming larvae; eruciform; length up to 8-0 mm. Head: hypognathous, oval, clypeus somewhat asymmetrical with 2 pairs of lateral indentations and 10 bristles, a suture-like line present paralleling the epicranial arms; gular sclerite pear-shaped, not fused with genae. Head golden brown with darker brown patterning as indicated in figure 1 A-E. Anteclypeus membranous, pale. Eyes large, black, set in a clear pale area, with 6 raised transparent cuticular plates over each forming hyaline lenses. Antennae conspicuous, with bulbous proximal segment and long distal segment terminating in a sensory bristle. Mouthparts: labrum armed with 4 blade-shaped spines and 6 long setae, there are also 2 pairs of very short spines just below the central emargination. Mandibles unequal in size, dark brown, heavily sclerotized; each mandible with 5 teeth, 2 small brushes of stout bristles on the inner face of the larger left mandible. Maxillary palps 3-segmented, 3rd segment bearing several small papillae; maxillary lobe bordered with long blade-like spines, smaller spines and setae. Labium a low cone, labial palps each tipped with a papilla. Thorax: pro- and mesonota sclerotized, patterned with darker brown and cream as indicated in figure 1 B; metanotum membranous except for 2 small oval patches bearing setae. The patterning on both head and thoracic nota is variable, as is the ground colour, which varies from dark brown, when the pattern scarcely shows, to pale yellow, when it shows up very clearly. Each half of the pronotum is subdivided by a groove. The mesonota are lightly sclerotized, with a pair of narrow dark bars. The metathorax bears what appear to be 2 small gills on each side (fig. 1 A). Legs: prothoracic legs short, stout; tibia and tarsus edged with spines ; bristles, spines and hair-like setae fringe the femur. Stridulator stout, with blunt apex. Mesothoracic legs armed with strong spines, interspersed with hair- like setae along edge of femur. Metathoracic legs long, strong, with heavy double fringe of long hairs used in swimming. All leg joints of one segment only except trochanters which have two; femora of mid- and hind-legs also appear to have two segments, though it is not clear whether this division is functional or merely incipient. Coxopleurites, trochantin etc. of legs heavily sclerotized. Abdomen: white; on segment 1 lateral humps covered with small scabrosities, and a pro- minent dorsal hump bearing a patch of spicules; on segment 2 a small sclerite on each side and 3 simple filiform gills (presegmental dorsal and ventrolateral gills, and a postsegmental lateral gill) ; on segments 3 and 4, 2 gills (presegmental dorsal and ventrolateral) ; no gills on segments 5-9. Lateral line extends from segment 3 to 7; on segment 8 it is replaced by a row of small bifid tubercles. Segment 9 is membranous but the tergite is bordered posteriorly with a row of 6 strong black bristles. The anal prolegs are fused basally to form a 1oth segment; each is short and bulbous with a lateral sclerotized patch and several SOME NEW CADDIS FLIES FROM THE WESTERN CAPE PROVINCE I7 ————— M eS = —_— =S i \ i ‘ i | 'Z Zag / Fic. 1. Larva of Athripsodes prionii Scott. A, lateral view of whole larva, with lateral tubercle enlarged. B, dorsal view of head with pro- and mesonota. C, antenna. D, clypeus. E, head from behind showing gular sclerite. F, right mandible. G, left mandible. H, labrum. J, posterior end of larva (right side). K, anal claw. L, prothoracic leg. M, metathoracic leg. N, case of early instar of larva. P, Q, cases of later instars. R, case of mature larva. 18 ANNALS OF THE SOUTH AFRICAN MUSEUM stout black setae; anal claws have 4 barbs (1 large and 3 small); anus flanked by patches of small spicules. Case (fig. 1 N-R): Tubular, wider at anterior end, with hood. Small larvae make their cases of sand grains, lining them with secretion; older larvae utilize small bits of vegetable debris, charcoal, etc., with a few sand grains at the posterior end; mature larvae make a very characteristic rounded case from 3 neatly shaped pieces of dead leaf (usually of Prionium serrata); one piece is always ionger than the others and forms a hood at the anterior end of the case and a projecting strip at the posterior end. The case is lined with secretion and may be neatened round the front and hind openings by the addition of a few small bits of leaf glued on with secretion; there is always a neat round hole at the posterior end. Pupa (fig. 5 K-Q): Labrum rounded, rising to a low point; setae compara- tively short, some recurved; mandibles large, strong, inner edge finely serrated, outer edge sinuous, ending in a strong condyle, one pair of lateral setae. Antennae much longer than body, distal ends coiled. Fore-tarsi slightly, mid-tarsi heavily, fringed. Dorsal plates present on abdominal segments as follows: on segment 1, 2 pairs of protuberances covered with small sclerotized points (these point in opposite directions in each pair); on segments 3 and 4, 1 pair of hook-bearing plates; on 5, 2 pairs; on 6, 1 pair. Segment g bears a pair of small lappets tipped with forwardly directed setae. The plates on 3 and 4 each have 3 hooks, the anterior plates on 5 and those on 6 each have 2 large hooks and 3-4 smaller ones. There are minute scabrosities scattered on the dorsal surface of the abdomen. Lateral lines extend from segment 3 to 8, where the two curl round and meet ventrally. Anal appendages are long and stout, with rounded tips; along the inner edges are small barbs and a few setae. Athripsodes bergensis Scott Fig. 2 A-L; fig. 5 A-J Remarks: The larvae are common in the Great Berg River near Drie- fontein and Groot Drakenstein in the late winter and spring (August to October). They are crawlers with long, slender, rough-looking cases of brown sand-grains to which pieces of stick or charcoal are often attached, and they are frequently to be seen crawling actively about on top of the stones seeking food. They are carnivorous and in captivity must be given insect larvae or other small animals to eat to prevent cannibalism. It is possible that the sticks etc. attached to their cases serve to discourage predatory fish, as in the case of Anabolia nervosa in Britain. Pupae were found in October, clustered in numbers under marginal stones; some were collected and bred out in the laboratory. The pupal cases are not larval cases sealed off, but each is constructed from small white sand- grains at the end of the old larval case, being first formed as an extension of it. When completed however the pupal case is closed off at each end by a neat lid SOME NEW CADDIS FLIES FROM THE WESTERN CAPE PROVINCE Serene rz iEsschpeiziue \\ er ei aigae uae eH T 8, 9); VI (Part 1, Index), VII (Parts 1-4), VIII, [IX (Parts 1, 2), X (Part 2), XI (Parts 2, 7, Index), SOK XO (Bart) OO (Bacio): Current prices: Vol iit. nite IV. AE XVI. XVIT. OWAO US XIX. DS. XXII. XXIII. XXIV. XXV. XXVI. XXVII. XXVIII. XXIX. XXX. INDEX XXXII. XXXII. XXXII. XXXIV. XXXV. XXXVI. XXXVI. XXXVITIT. XXXIX. XL. XLI. XLII. XLITT. XLIV. XLV. XLVI. 1900-1902 1903-1905 1903-1908 1906-1910 1908-1910 1908-1913 1911-1918 IQII-1914 1911-1918 1913-1924 1913-1923 1915-1924 1914-1916 1917-1933 1917-1920 1921 1924-1925 1924-1926 1925-1928 1925-1926 1929-1938 1927-1928 1928 1929 1929-1932 1929-1931 1931—1935 Zoology and Geology (excl. Parts 1-3, 5, 7, a Zoology (excl. Part I) Palaeontology .. Geology, Balaconeolony, Zoology, Anthropology (exc. Paris 1,2. 5,05)9) Zoology (aa, Part I, Index) Palaeontology (excl. Parts 1-4) Botany (excl. Parts 1, 2) Zoology (excl. Part 2) Zoology (excl. Parts 2, 7, Inde) Palaeontology and Geology Archaeology and Zoology Zoology Zoology Botany Zoology Zoology Zoology Zoology ap Palaeontology .. Zoology : Anthropology anal Ethnology eal Part ‘Sy Zoology Zoology Anthropology Palaeontology .. Zoology Zoology of papers, authors and sae TAPERS] H in aycls ie XXX 1934-1950 1935-1940 1939 1938 1956 1942-1948 1947-1952 1950 1952 1952-1956 1952-1955 1953-1956 1955-1957 1957-1959 1959-1960 1961— The LIBRARIAN, Sourn Arrican Museum, Carr Town. Palaeontology (excl. Part 1) Zoology Zoology Zoology Zoology Zoology Archaeology Zoology Zoology Botany Zoology Palaeontology : Zoology and Palaconi clea Zoology and Palaeontology Zoology and Palaeontology Zoology and Palaeontology (Parts a, Copies may be obtained from— R 2.75 4.38 8.08 1.97 5.48 1.26 6.03 MTG 6.73 11.71 8.55 8.04 10.26 8.57 8.45, 10.33 8.10 6.27 4.85, 4.25, 6.00 4.70 3.60 3.60 6.30 5.80 8.88 0.23 6.13 7.62 5-95 5-75 14.40 6.13 4-33 9.00 6.55 3.18 9.60 8.10 4.80 6.64 1t.29 1.03 07.68 ANNALS OF THE SOU (eearnRiCAN MUSEUM VOLUME XLVI PART VI THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS By LIEUWE Dirk BoonstTRA South African Museum, Cape Town er cH i ISSUED JANUARY 1962 PRICE 95¢ PRINTED FOR THE TRUSTEES OF THE SOUTH AFRICAN MUSEUM BY THE RUSTICA PRESS (PTY.) LIMITED, COURT ROAD, WYNBERG, CAPE INSTRUCTIONS TO AUTHORS MSS. submitted for publication must be typewritten, double spaced with good margins, typed on one side of sheet only and sheets numbered. Two MSS. and one set of illustrations must be submitted. Papers must include a Table of Contents and a Summary containing a succinct account of the subject, results obtained and conclusions. The position of the text-figures and tables must be clearly indicated. ILLUSTRATIONS. Please must be kept to a minimum and made up to occupy 5" x 7” (full plate excluding captions) or in direct proportion to these measure- ments if they are to be reduced. A scale (metric system) must appear with all photographs. REFERENCES. Harvard system—authors’ names and dates of publication given in the body of the text; references arranged at the end of the paper in alphabetical order of authors’ names. Bibliographical Arrangement of references at the end of the paper must give: 1. Name of author, followed by his initials. Names of joint authors connected by ampersand (&), and not ‘and’. 2. Year of publication. If several papers by the same author in one year are cited, suflixes a, b, etc., are used. 3. Full title of the paper. Initial capital letters only for the first word and for proper names, except in German, where the usage of the language is followed. 4. ‘Title of the journal, abbreviated according to the World list of scientific periodicals, and underlined to indicate italics. 5. Series number, if any, in parenthesis, e.g. (3), (n.s.), (N.F.), (B). 6. Volume number in arabic numerals (without prefix ‘vol.’), with wavy underlining to indicate black type. 7. Part number, only if the separate parts of one volume are independently numbered. 8. Page numbers, first and last, in arabic numerals (without prefix ‘p’). When reference is made to a separate book, give in the order listed: Author’s name; his initials; date of publication; title, in italics; edition, if any; volume number, if any, in black type arabic numerals; pagination, if reference is to part of book only; place of publication; name of publisher. When reference is made to a paper forming a distinct part of another book or collection, give: Name of author of paper; his initials; date of publication; title of paper; ‘In’, italicized; name of author of book; his initials; title of book, in italics; edition, if any; volume number, if any, in black type arabic numerals; pagination of paper; place of publication; name of publisher. [| CONTINUED ON INSIDE BACK COVER PAT DEM TITION GF THE ‘TITANOSUGHIAN DINOCEPHALIANS By LizuwE Dirk BoonstRra South African Museum, Cape Town [Accepted January, 1961] (With 1 plate and 41 figures in the text) CONTENTS PAGE PAGE INTRODUCTION . : E : . 57 DESCRIPTIVE (SPECIMENS ‘'TAXONOMIC- GENERAL d : : : Puss: ALLY) 3 : : eer 5) MATERIAL : : : : a 45O Titanosuchida <4, 69 DEscRIPTIVE (GENERAL). DE 'G@ Jonkeriidae . ; 76 Incisors : ; . €o0 Comparison wITH OTHER DINOCEPHALIA 97 Canines : . 63 COMPARISON WITH CONTEMPORARY Postcanines . , 3 ; : 63 ‘THERAPSIDS . : : : i ALOS ToorH REPLACEMENT ; ; . 65 OricIn oF DINOCEPHALIAN DENTITION . 104 Incisors. , ‘ . 66 ARTICULATION OF THE LOWER JAW 4108 Canines : : : . 67 ACKNOWLEDGEMENTS . : BA tits Postcanines . - 68 REFERENCES - : : Hag i INTRODUCTION In the South African Dinocephalia the teeth are but poorly known. The inadequate accounts hitherto given by the authors, who have studied this sub-order, are chiefly due to the nature of the preservation of the studied material and the rather rough methods of preparation hitherto employed. Many specimens are recovered only when already naturally weathered out of the entombing rock and lying exposed on the eroded surface of the rocks of the Yapinocephalus zone as broken-up pieces weathered to various degrees. In the cranial material thus collected the crowns of the teeth are seldom well preserved, but exceptionally such material has yielded isolated well-preserved crowns. Such is the case in specimens which when petrified had full sets of teeth, but it is evident that a large number of skulls had on death lain exposed before entombment and during this time many of the teeth had often either fallen out of their sockets or had their crowns damaged or wholly broken off. This is evident from the fact that in many cases we find empty alveoli filled by matrix and by the fact that loose-lying teeth occur lying near parts of the skull or even unassociated with any other skeletal material, the harder dental substance being preserved and the less hard bone being lost. During thirty-two years of collecting in the Tapinocephalus zone I have found that the whole aif Ann. S. Afr. Mus. 46 (6), 1962, 57-112, 1 pl., 41 figs. +} 2% AGS 58 ANNALS OF THE SOUTH AFRICAN MUSEUM teeth—crown plus root—fall out of the alveoli before petrifaction more frequently in the Tapinocephalia than in the other three infra-orders; and in the Titanosuchia the crowns are broken off at the level of the alveolar border either before petrifaction or during the later weathering more frequently than in the other infra-orders. I believe this to be due to the difference in implanta- tion and the mode of replacement of the teeth in these infra-orders. When crowns are preserved in situ the intractable matrix makes good preparation difficult with the older hammer-and-chisel technique employed. Improved newer mechanical methods—vibro-needles and sectioning with rotating diamond-studded saws—and chemical solution now give better results in suitable specimens. If available, radiography would be a further improvement. GENERAL The dentition of the sub-order Dinocephalia (Anteosauria, ‘Titanosuchia, Tapinocephalia and Styracocephalia) can be readily distinguished from those obtaining in all the other contemporary sub-orders of the order Therapsida by the following distinctive character: In occlusion the upper and lower incisors intermesh, whereas in all the other contemporary therapsids with incisors (thus excluding the anomodonts) the upper incisors pass labially of the lower incisors. This distinctive intermeshing is developed further in the Titanosuchia, where the lower canine passes in front of the upper canine to intermesh between the last upper incisor and the upper canine (fig. 1). This process is carried still further in the Tapinocephalia and the Styracocephalia where the whole battery of teeth intermesh. VAN ANAY Fic. 1. A diagram to illustrate the intermeshing of the incisors and canines in the Titanosuchia. | Abbreviations used in this and subsequent figures: 1-5 = functional incisors. 1’—5’ = first replacing incisors. 1”—5” = second replacing incisors. 1°-5° = pre- decessors to the functional incisors. An = angular. Ar = articular. C = canine. C! = replacing canine. Cor = coronoid. D = dentary. F = lingual flange of the dentary. F. Pt. = fossa for the quadrate ramus of the pterygoid. G = groove labial to the flange of the dentary. H = heel of the tooth. I.C. = inner condyle. L.C. = lower canine. M = maxilla. m = maxillary tooth. m1! = replacing maxillary tooth. N = nasal. O.C. = outer condyle. Pa = prearticular. Pal = palatine. PC = postcanine. PC! = replacing postcanine. PCla = alveolus for replacing postcanine. Pm = premaxilla. Proc = processus musculus pterygoidus pos- terius. Pt = pterygoid. Q = quadrate. Q.F. = foramen quadrati. QRPt = quadrate ramus of the ptery- goid. R = root of the tooth. Sa = surangular. Sm = septomaxilla. Sp = splenial. St = stapedial recess on the quadrate. T = talon of the tooth. U.C.. = upper canine. V = vomer. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 59 This intermeshing of the teeth is, I believe, a character of sufficient impor- tance, when taken in conjunction with other evidence available from the rest of the dinocephalian skeleton, to validate the old conception of grouping the four infra-orders together in one sub-order—Dinocephalia, and to show that the newer attempts in classification which are intended to refute this close consanguinity are ill-advised and contrary to the facts. Although the four infra-orders of the Dinocephalia are linked together by this common character of intermeshing, there are certain dental features (together with other skeletal characters) distinctive of each infra-order, but all derivable from one common ancestral condition. The dentition in the Titanosuchia can be readily distinguished from that of the Tapinocephalia and Styracocephalia by the presence of large specialized canines in both upper and lower jaws; from that in the Anteosauria, which also have large specialized canines, by the fact that in the Titanosuchia the lower canine is directed both outwards and forwards to intermesh between the upper canine and the fifth upper incisor, whereas in the Anteosauria the lower canine is directed upwards to lie inside the upper canine with its point housed in a special recess in the maxilla. The infra-order Titanosuchia has, on skeletal characters other than those shown by the dentition, been subdivided into the three families Titanosuchidae, Jonkeriidae and Dinartamidae. The dental material at my disposal has not enabled me to recognize differences in the dentition to substantiate this subdivision into families—a sub- division which is undoubtedly valid for the former two families. The family Dinartamidae founded on a single very poor specimen is of doubtful validity. MATERIAL This report on the dentition of the Titanosuchia is based mostly on the material in the collection of the South African Museum. This consists of 37 specimens, from isolated localities situated over the whole of the western fossiliferous part of the Tapinocephalus zone, in which adequate evidence of the dentition is preserved. Of these 37 specimens, 1 is from an unknown collector, 1 collected by Cloete, 1 by Cairncross, 1 by Hugo, 3 by Whaits, 3 by Haughton and the other 27 by the author. Although the majority of these specimens are poorly preserved, the total of the determinable features gives us a fairly complete picture of the nature of the dentition in the infra-order as a whole. The available facts are, however, insufficient as reliable characters employable for taxonomic purposes within the group. Such characters as have been used by previous authors for the establish- ment of genera and species do not at present appear to be sufficiently distinctive or constant and such classification must at the present stage be considered of doubtful validity. 60 ANNALS OF THE SOUTH AFRICAN MUSEUM DESCRIPTIVE General Before giving an account of the individual specimens, in an attempted taxonomic order, the general features of the dentition of the infra-order as a whole follows. In the Titanosuchia a full and mature set of teeth gives the formula: , 5 i peak) la, Cy, PC ye rer. Where previous authors have given different formulae for their specimens and in the specimens at my disposal, which give a different count, these differences should at present, without further confirmatory evidence, not be attributed to any taxonomic distinction. That the different genera and species Fic. 2. Jonkeria. The left second upper incisor S.A.M. E. x $. A = anterior or inner view. B = posterior or outer view. C = lingual view. D = labial view. may have had different numbers of post-canines cannot, of course, be discounted ; but in the light of our present knowledge such differences should rather be considered to be due to imperfect observation, imperfect preservation, age of the animal, or to the stage which the replacement of the teeth happens to have reached in each individual jaw. Noted differences in the left and right sides of the jaws of the same beast bear this out. The Incisors Both the upper and lower incisors have the structure which has been called ‘dinocephalian’ by various authors. What these authors really mean is that the incisors have the general structure of the teeth of the Tapinocephalia (especially the Russian Deuterosaurus and Ulemosaurus). Later in this paper it will be shown THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 61 that the titanosuchian incisor can be readily (figs. 2 and 3) distinguished from the tapinocephalian teeth. Essentially the distinctive structure of these teeth is the development in the crown of a piercing point separated from a cutting and/or crushing surface. Each tooth consists of a strong, long, curved, conical root; oval in section, and a nearly equally long, curved, claw-like point or talon with a convex labial face and a flattened lingual face with longitudinal striae and fairly flat sides. At the Fic. 3. Jonkeria. Lateral view of the upper jaw teeth. S.A.M. E. x4. Partly restored. junction of crown and root the crown has on its lingual or inner face an oblique step, heel or cingulum. This heel forms an oblique ledge with its outer part higher than its inner part in the upper incisors and its outer part lower and its inner part higher in the lower incisors. In the unworn tooth the lingual edge of this heel forms a fairly sharp ridge with coarse serrations; after some wear the edge loses its serrations and the sharp edge becomes rounded; when greatly worn the edge is lost and the whole ledge acquires a flat face. Concomitant with this wear of the heel the talon is also worn away lingually to become more and more slender and thereby increasing the size of the crushing flat face of the heel. 62 ANNALS OF THE SOUTH AFRICAN MUSEUM There are certain differences between the five upper inter se and between the four lower incisors (fig. 4). The first pair of upper incisors is more lightly built than all the others; this lightness is mainly due to the fact that the teeth are compressed from side to side, lying close together and, in occlusion, passing their talons in between the pair of lower incisors. The fifth upper incisor has the rear face of the talon modified to receive the lower canine with which it intermeshes. In the lower jaw the fourth incisor is somewhat weaker than the anterior ones. Other differences Fic. 4. Jonkeria. S.A.M. 9162. x 4. A = anterior view. B = lateral view. between the incisors can best be attributed to age or wear, but with more material available some of these differences may prove to be of taxonomic value. The length of the incisor series varies in length: upper 73-106 mm. and lower 68-94 mm., but the variation between the two sides of the jaw is such that these measurements have a doubtful taxonomic value. In occlusion the incisors of the two jaws intermesh (fig. 1) as follows: The talons of the upper No. 1 pair lie in between the talons of the lower No. 1 pair. | Upper No. 2 lies between lower Nos. 1 and 2. Upper No. 3 lies between lower Nos. 2 and 3. Upper No. 4 lies between lower Nos. 3 and 4. Upper No. 5 lies between lower No. 4 and the lower canine. The outer half of the heel of upper No. 1, abuts against the inner half of the heel of lower No. 1; the heel of upper No. 2 has its inner half abutting THE DENTITION: OF THE TITANOSUCHIAN DINOCEPHALIANS 63 against the outer half of the heel.of lower No. 1 and its outer half against the inner half of lower No. 2; similarly upper No. 3 abuts against the halves of lower 2 and 3; upper No. 4 abuts against the halves of lower 3 and 4; upper No. 5 abuts against the halves of lower 4. and the inner face of the lower canine. The inner face of the lower canine shows in some specimens an oval worn face where attrition by the upper No. 5 incisor takes place. In both upper and lower jaws the incisors do not stand -yenacsllbe ¢ in the jaws but are directed forwards so that in occlusion the tips of the talons protrude anteriorly. | The Canines Each maxilla and dentale bears a single robust canine (figs. 3. and aye | The upper canine has a large conical crown which curves moderately to . strongly backwards and is moderately flattened from side to side to present a — greater or lesser oval in cross section, its posterior face is rounded with no sharp edge or serrations. The root is large and long and its presence causes the maxilla’ to bulge both externally as well as internally; internally the root opens on to the inner face of the maxilla, where the large pulp cavity remains open for the passage of nerve and nutrient vessels (fig. 31). Where it leaves the jaw the canine diameters vary from 35 to 50 mm. antero-posteriorly and from.21 to 36 mm. from side to side. The crown length varies from 60 to 110 mm. Between the last upper incisor and the canine there is a moderate diastema +20 mm. which allows the lower canine to pass between these two teeth. In occlusion the upper canine passes behind the lower canine and fits in a hollow in the outer face of the dentale. In place of the lower canine lying lingually of the upper canine in occlusion there are here situated the first 4-5 lower post- canines. The lower canine is smaller than the upper, shorter and less pointed and curved, with its blunted point directed outwards to pass in between the upper canine and the 5th upper incisor and to pass externally of the edge of the upper jaw. Its intero-anterior face frequently shows an attritional face where the last upper incisor wears against it. The outwardly directed lower canine has by Broom been considered a distinctive character of Scapanodon, but it is a feature common to all the known Titanosuchia, which in this character differ from all other contemporary Therapsida. | | | The Postcanines The postcanines (fig. 5) form a long series in both jaws. The spacing is close and, except where replacement is taking place, regular. They differ from both the incisors and canines in that the teeth of the upper and lower jaws do not intermesh. In occlusion the lower set lies lingually of the upper set, as is normal, in the contemporary therapsids. The upper set is implanted close to the outer edge of the maxilla, whereas the lower set lies close to the inner edge 64. ANNALS OF THE SOUTH AFRICAN MUSEUM Fic. 6. Parascapanodon. S.A.M. 12213 X $4. A = right upper postcanine in labial and posterior view. B = a postcanine in labial view. C = the 11th right lower postcanine in labial and posterior view. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 65 of the dentale, with on their outside, a groove for the reception of the upper set. The upper postcanines are somewhat larger than the lower ones, but built on essentially the same plan. In both jaws the crowns lie obliquely to the long axis at an angle of about 40°, with the anterior edge directed inwards and the posterior edge outwards. The first postcanine in both upper and lower jaws is directed somewhat anteriorly and is usually the largest postcanine, the next four to five are of the same size, and the rest gradually decrease in size backwards. The postcanines in both jaws follow on their respective canines without a diastema. As the lower canine passes anteriorly of the upper canine, and the length of upper and lower series being approximately equal, it follows that the last 4-5 upper postcanines have no lower antagonists. Furthermore in occlusion the first 4-5 lower postcanines lie lingually of the upper canine. Each postcanine (fig. 6) consists of a long root more or less oval in section, with a spatulate crown roughly triangular in outline; the outer face is moderately convex and the inner face flat to slightly concave. Both inner and outer faces carry fairly coarse striae, which in unworn teeth extend to the edge which thus becomes serrate. In my specimens the length of the upper series varies from 169 to 188 mm. and the lower set from 165 to 210 mm. The full set of postcanines numbers 19-21 ; in those specimens with a shorter set and a lesser number of teeth it would be reasonable to assume juvenility, additional teeth being added posteriorly with increasing age. TootH REPLACEMENT In nearly every specimen replacement of teeth can be seen (figs. 7 and 8). Evidence of this is, first, irregularity in the series, the presence of empty alveoli, and, lastly, the presence of crowns in the process of eruption. Replacement occurs in incisors, canines and postcanines, but due to the great mass to be replaced replacement is less frequently seen in the large canines. There is evidence of at least two sets of canines and postcanines and of at least three sets of incisors. Replacement may thus be a continuous process— . there being no positive evidence of it ceasing. In all the teeth the immediate replacing tooth erupts lingually of its predecessor and is thus a younger member of the same tooth family. In the Titanosuchia functional distichism thus no longer occurs. Since in each tooth, generation follows generation independently, at any given time the functional teeth are, except initially, never all of the same age or generation. This is specially noticeable in the incisors, where in almost every set of functional teeth very distinct stages of development and subsequent wear can be seen. In many cases, from the stage of wear of the functional incisors, the order of replacement can be deduced and this can then be verified by noting the degree of development of their respective replacing teeth. 66 ANNALS OF THE SOUTH AFRICAN MUSEUM Replacement of the Incisors In the premaxilla (fig. 7) a cavity develops in the bone lingually of the alveolus which houses each functional incisor. This cavity houses the germ from which the replacing incisor develops. By the time the replacing crown reaches a certain stage of development the inner or lingual wall of this cavity develops, by resorption of the bony tissue, a perforation. This opening enlarges to form a long oval fenestra and through this fenestra the crown of the replacing tooth 908209 OQ S DP 22000 ee = Se Fic. 7. Jonkeria boonstrat. S.A.M.9079 x t. Ventral view of upper jaws. emerges. As the replacing tooth continues its growth it moves labially into the lingual wall of the alveolus of the functional tooth which is gradually reabsorbed, until the new alveolus coalesces with the old alveolus. The old tooth has con- comitantly suffered resorption and as it disappears the new tooth takes over its alveolus. In the meantime the fenestra through which the replacing crown can first be seen has closed up from above downwards so that the new tooth when arriving in the old alveolus has its own lingual alveolar wall. The talon of the young tooth, when just emerging, is flattened antero-posteriorly, roughly triangular in outline with its inner face coarsely striate and since these striae extend to the edges these are serrate. | THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 67 Noting the degree to which each replacing incisor crown has developed, I have tried to determine if there is a regular order in the replacement. (This can be controlled by noting the degree of wear in the crowns of the functional incisors.) But as could be expected this was not possible. In the dentary (fig. 8) the eruption of the replacing incisors appears to follow a different course than noted in the upper incisors. This is due to the fact that in the dentary there lies lingually to the erupting incisors a flange of bone with a free upper edge, which thus forms a groove labially and at first sight it appears as if the replacing incisors emerge from this groove. Fic. 8. Jonkerta boonstrai. S.A.M. 9079. x $. Three consecutive frontal sections through the anterior part of the dentary to show the replacement of the incisors and the canine. When seen in section (fig. 9), however, it is clear that the replacing teeth develop in cavities in the solid dentary labially of this groove. The mode of development is thus similar to that of the upper incisors, viz. lingually of each alveolus housing a functioning incisor a cavity is formed in which the tooth germ develops and the further course of development parallels that already described for the upper incisors. Replacement of Canines Little is known of the replacement of the canines (figs. 7 and 8). In the upper jaw there is frequently a small hollow in the lingual wall of the maxilla just above the inner alveolar edge of the functional canine. That this hollow indicates the point of eruption of the replacing canine as do the fenestrae lying 68 ANNALS OF THE SOUTH AFRICAN MUSEUM lingually of the incisors is proved by the one case (S.A.M. 11884) where on both sides it houses the tip of the replacing canine. In two cases we have an empty canine alveolus, but in-neither can a replacing tooth be determined with certainty. In the lower jaw a hollow lying lingually of the posterior end of the inner free-edged flange of the dentary is frequently present and this in four cases houses a small tip of the replacing canine. Cia Bao Ties we orn AT So ae 85 + ie aI . Q . ~ Gy Bs Lies Pop SNe ainAioss f- rid) Bog 0 0000 OO © 97°08 Fic. 9. Jonkeria. S.A.M. 9161. A = dorsal view of mandible x }. B = fracture through the jaw at the level of the first right incisor showing the dental flange (F) with its free upper border and the groove (G) lying labially. x 4. Replacement of the Postcanines In both upper and lower jaws the emerging cusps of replacing teeth can frequently be seen, each lying lingually of either a functional tooth of the same family or of an alveolus from which its predecessor has disappeared (fig. 7). The replacing crown-tips appear on the alveolar border without the develop- ment of a fenestra in the lingual wall of the jaws as is the case in both incisors and canines. Some, at least, of the replacing postcanines arise in separate alveoli lying lingually of those of the functional teeth and these alveoli coalesce on the loss of the functioning tooth. The erupting teeth appear as striate, serrate tips triangular in outline. Usually each series of postcanines has only a few replacing tips showing, but in a few specimens a larger number—up to 9—have erupted at more or less the same time. At the most 3 adjacent teeth have been seen erupting together, but mostly the erupting teeth appear at irregular intervals. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 69 DESCRIPTIVE (Specimens Taxonomically) Titanosuchidae Genera ARCHAEOSUCHUS and SCAPANODON The dental material which constitutes the types of these genera is so poorly preserved that no useful purpose will be served by attempting to add to Broom’s descriptions. Genus PARASCAPANODON Parascapanodon avifontis (Fig. 10) S.A.M. 9127. Type. Voélfontein, P.A. Collected Boonstra 1929. This specimen consists of parts of a skull including parts of the upper and lower jaws together with good bones of the postcranial skeleton. Apart from some roots of the upper incisors there is an incomplete right mandibular ramus showing the roots of most of the teeth. Of the incisors only the root of the fourth is present. The canine root is oval in section (diams. 28 and 36 mm.) and the lingual wall of the alveolus shows an opening for the replacing canine. The set of postcanines consisted of 21 teeth of which the first is much larger than the rest. Nos. 8, 10, 12 and 15 are replacing teeth lying lingual of alveoli from which the former set has fallen out. The length of the postcanine series is 180 mm. a i ' Re me we ’ ‘ ) ] ¢ é ‘ ¢ SS n SC) (aa Fae ee Ream, ah ear Fic. 10. Parascapanodon avifontis. Type. S.A.M. 9127. x. A = dorsal view of right mandible. B = lateral view of right mandible. 70 ANNALS OF THE SOUTH AFRICAN MUSEUM Parascapanodon sp. S.A.M. 12213. Kroonplaas, B.W. Collected Boonstra 1959. Under this number there are parts of at least three individuals found together. Besides some skull pieces, a number of isolated loose-lying teeth and a fibula, are preserved: A (fig. 11). A fairly good anterior two-thirds of a mandibular arch in which crowns of the teeth are preserved in a fair condition. The dental formula is i4, cl, pel19g: Incisors On the left side the crowns are broken off just below the heel. Lingually of No. 2 lies the erupted crown of its replacement. It lies in the groove just Fic. 11. Parascapanodon sp. S.A.M. 12213 A. Xx}. A = dorsal view of mandibles. B = lateral view of the right mandible. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 71 anterior to the flange of the dentary and against the lingual alveolar wall of the functioning tooth. Its inner face is striate. Lingually of No. 4 a pocket in the alveolar wall contains the crown of its replacement. In the right dentary the four incisors are well preserved—only the tips of the talons being lost. ‘The heels of all four are very well preserved, and are very little worn so that the lingual edge still forms a raised edge with a hollow between this edge and the posterior face of the talon; and in No. 4 the posterior surface of the talon still shows fairly coarse striae. Lingual of Nos. 2 and 4 replacing crowns are present in their respective pockets. The fact that replacing crowns are already showing, while the functioning crowns are still little worn, indicates that considerable time may elapse before the actual replacement takes place. Canines The left canine is broken off at the base of the crown, which is here oval in section with diameters 39 and 28 mm. Of the right canine the tip of the crown is lost. It is a moderately sized tooth (basal diameters 38 x 28 mm.) directed much outwards. On its antero-internal face there is an oblique groove in which the 5th upper incisor fitted in occlusion. Lingual to both canines there is a notch in the flange of the dentary and a pocket in the internal alveolar wall for the emergence of the replacing canines. The Postcanines On the left dentary only a few crowns are preserved — the rest being broken off below the spatulate crown. The presence of 19 postcanines can be deter- mined occupying 197 mm.—No. 12 is represented by an empty alveolus and there is a replacing crown lying postero-lingual to No. 17. No. 1 is large (diams. 9 x 7 mm.), whereas No. 5 is only 7x6 mm. and the others approxi- mately of this size. In the right dentary the crowns of the postcanines are better preserved. The original presence of 19 teeth occupying 188 mm. can be determined. No. 6 and No. 11 are represented by empty alveoli and lingual of alveolus 6 is a replacing crown. The crown of the 1st postcanine is widely spatulate, with a convex outer and a flattened internal face; the edges are sharp and coarsely serrate. This, as also the other postcanines, have their spatulate crowns lying obliquely to the long axis of the skull, with the anterior edge lingually and the outer labially directed. In lateral view they thus overlap where closely spaced. The other postcanines are very similar to No. 1, except that their crowns are smaller and the spatula more sharply triangular in outline. There is a gradual decrease in size posteriorly. The postcanines lie on the inner part of the upper dentary edge; external to the tooth row there is a longitudinal groove for the reception of the upper postcanines. ANNALS OF THE SOUTH AFRICAN MUSEUM ~I i) B (fig. 12). A mandibular arch in somewhat better condition than A. In the left dentary the four incisors and the canine have the major part of their crowns preserved; of the postcanines only the first five and last two are preserved and they occupied 194 mm. In the right dentary most of the crowns of the four incisors, canines and the original 18 postcanines are preserved. The postcanine series measures 199 mm. and lingual of No. 7 there is a replacing crown and No. 13 is repre- sented by an empty alveolus. Fic. 12. Parascapanodon sp. S.A.M. 12213 B. Xf. A = dorsal view of mandibles. B = lateral view of right mandible. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 73 Incisors The talons of the incisors are all lost except in No. 4 right. All the heels are very well preserved. The heel of No. 2 left has been worn fairly flat, whereas in all the others the lingual edge of the heel is still sharp; between this sharp ridge and the posterior face of the talon the face of the heel is hollowed out. In all, the inner corner of the heel is higher than the outer corner—the attritional face thus lies obliquely to the frontal plane. On the left side replacing crowns lie in pockets of the alveolar wall of Nos. 2, 3, 4, and there is an empty pocket behind No. 1. On the right side a strong replacing talon emerges obliquely from a pocket behind No. 2; the pocket behind No. 4 contains a tip, and empty pockets lie behind Nos. 1 and 3. Canines All except the tips of both canines is preserved. The canines are large teeth, oval in cross-section. The diameters at the crown base are for the left 37 x 28 mm. and for the right 43 x28 mm. Both canines are directed outwards. On the antero-internal face of the left canine there is a hollowed-out attritional face against which the upper 5th incisor worked. The Postcanines The spatulate crowns stand obliquely to the long axis of the skull; they are triangular in outline with sharp edges and are coarsely serrate. The outer face is convex and the inner fairly flat. The last five are appreciably smaller than their predecessors in the series. The postcanines are implanted on the inner upper edge of the dentary with a shallow groove externally for the reception of the upper postcanines. C (fig. 138). A weathered anterior part of the dentary shows the roots of all four functional incisors, but all four have reached the age at which they are to be shed. Lingually of each is a large empty alveolus from which the replacing 74 ANNALS OF THE SOUTH AFRICAN MUSEUM tooth has been lost. In No. 3 the alveolus, which housed the replacing tooth, has labially become confluent with the older alveolus housing the functional older tooth of the family. In the other three the alveoli of the replacing teeth are still separate from the older alveoli of the still functioning teeth. The canine crown is broken off at its base where the diameters are 44x27 mm. The pulp cavity is large. The inner alveolar wall has a pocket for the replacing canine. Evidence of nine postcanines is preserved; Nos. 4 and 5 are represented by empty alveoli and pockets for replacing teeth lie lingually of Nos. 1, 2 and 8. Parascapanodon sp. (Fig. 14) S.A.M. 12219. Bosluiskraal, Laingsburg. Collected Boonstra 1959. This specimen consists of a good left half of a snout and much of the post- cranial skeleton. The snout has its left half preserved in natural occlusion, but upper incisors Nos. 4. and 5 and the upper canine have fallen out of their respective sockets. The intermeshing of the incisors and canines is very clearly shown in this specimen. The first upper incisor is a very slender tooth. The lower canine is directed much outwards to pass outside of the maxillary edge. A well-developed striate replacing crown lies in a pocket lingual of the right upper No. 1 incisor and a less developed one behind the left No. 1 incisor. Fic. 14. Parascapanodon sp. S.A.M. 12219. X f. A = lateral view of snout. B = anterior view of the left half of snout as partly restored. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 75 Parascapanodon sp. S.A.M. 1204. Letjiesbos, B.W. Collected Whaits 1908. This specimen (fig. 15) consists of the mentum of the lower jaw in which the incisors are fairly well preserved together with part of a canine crown and three postcanine crowns. Incisors Most of the crowns of the eight incisors are present; the heel is imperfect in one but present in the other seven where only the talons are partly lost, but these can be restored with confidence as is shown in the figures. Fic. 15. Parascapanodon sp. S.A.M. 1204. X t. A = dorsal view of lower jaws. B = anterior view of lower jaws. C = lateral view of the left mandible. Partly restored. On the left side the four heels are excellently preserved in different stages of wear. No. I is worn smooth and nearly flat, in No. 4 there is still an indication of a lingual ridge, whereas in No. 3 and particularly in No. 2 the lingual ridge is still sharp and the face of the heel somewhat hollowed out. If the degree of wear is a criterion of age the order from oldest to youngest would be 1, 4, 3 and 2. In the lingual wall of the alveolus of each functioning incisor there is a pocket. In pockets Nos. 1, 3 and 4, talons of the replacing incisors are present, 76 ANNALS OF THE SOUTH AFRICAN MUSEUM whereas that of No. 2 is empty. The replacing talons are in outline isosceles triangles, with longitudinally striate faces. According to the degree of develop- ment the order of eruption would be 1, 4, 3 and 2 agreeing with the degree of wear of the functional incisors. On the right side the heel of No. 4 is damaged, that of No. 3 is worn flat, that of No. 1 fairly flat, and in No. 2 there is still a sharp lingual edge. The order of wear is thus 3, 4, I, 2. Lingually there are again four pockets in the alveolar wall. No. 2 is empty, whereas the others house replacing talons. The replacing order is 3, 4, 1 and 2. Lingual of the replacing teeth the free upper edge of the flange of the dentary forms a high sharp edge. Canines The base of the left canine is preserved, with basal diameters of 33 and 25 mm. Lingual of the canine, at the edge of the dentary flange, a pocket contains the replacing canine tip. Postcanines Only three spatulate postcanines have their crowns preserved, with a replacing crown lying lingually of No. 1. Jonkeriidae Genus JONKERIA Jonkeria cloetet S.A.M. 731. Type Gamka River. Presented by Mr. Justice Cloete. This type (fig. 16) consists of the anterior part of a left dentary, showing in section the roots of 4 incisors, I canine and the first four postcanines. The incisors were apparently fairly lightly built teeth. Lingually of each incisor the Fic. 16. Fonkeria cloetet. Type S.A.M. 731. Xt. A = dorsal view of left dentary. B = lateral view of left dentary. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS fey inner alveolar wall of the functioning incisors has been resorbed to form pockets through which the replacing incisors would in time have appeared. According to the development of these pockets the replacing order would have been 2, 4, ace The base of the canine is only slightly oval in outline with diameters of 31 and 25 mm. Lingually the alveolar wall of the functioning canine also shows a pocket and notch for the emergence of the replacing canine. The roots of the four postcanines are large, with the first—with diameters of 7 and 10 mm.—smaller than those lying posteriorly. The flange of the dentary with its free upper edge is well shown. Fic. 17. Fonkeria dubius. Type S.A.M. 2759. x. A = dorsal view of right dentary. B = lateral view of right dentary. Jonkeria dubius S.A.M. 2759. Type Platfontein of Abrahamskraal, P.A. Collected by Haughton 1913. This type specimen (fig. 17) consists of the anterior part of a right dentary in which only the roots of the teeth are shown in section. There are four roots of functional incisors, No. 1 shows only the tip of a root lying in an alveolus from which it is just about shed, and intero-lingually lies a well-developed replacing incisor already occupying the functional alveolus. The other three incisors, as well as the canine, have the inner alveolar walls strongly developed as encasing sheaths. The flange of the dentary extends high up to almost meet the inner face of the alveolar walls at a high level. The canine has a large root, which in outline is a long pinched-in oval] with diameters 47 and 27 mm. 78 ANNALS OF THE SOUTH AFRICAN MUSEUM Of the postcanines the first 13 have their roots preserved. The first is the largest (diameter 9 x 8 mm. whereas the others have on the average diameters of 7x7 mm.). The depression for the reception of the upper canine immediately behind the lower canine forms a deep anteriorly directed groove. On the inner face of the dentary, where the splenial has fallen away, there are embedded an incisor and two postcanine crowns, possibly from the upper jaw. Jonkeria haughton S.A.M, 4343. [ype Welgemoed, P.A. Collected Haughton 1916. This type (fig. 18) consists of a fair skull without mandible, but with much of the postcranial skeleton. Inctisors On the left the tip of the premaxilla has been lost and only sections of the roots of the five incisors can be seen at depth. On the inner alveolar face there are pockets for the replacements of Nos. 2 to 5 with crown tips of Nos. 3 and 4 lying in their pockets. The replacing order appears to be 3, 4, 5, 2, I. In the right premaxilla the crowns of Nos. 4 and 5 are very well preserved, whereas the other three are broken off at the alveolar border. As preserved the incisors are directed very much anteriorly, but most of this is due to the dorso-ventral crushing of the skull after death. The two preserved crowns differ from those known in other forms. In No. 4 the inner face of the talon runs obliquely from the heel to the tip without the usual distinctive face on the heel itself, which in other forms is either flat or hollowed out. In No. 5 this sloping inner face is also evident, but in this tooth it is bipartite —the outer and greater part is for the reception of the lower canine. Lingually of the right incisors there are four pockets in the alveolar wall for the replacements 2-5; in Nos. 2 and 4 there lie the striate tips of the talons of the replacing teeth. The replacing order appears to be 4, 2, 5, 3, I. Canines In the left maxilla there is an irregular empty alveolus from which the canine has dropped out and lingually of this there is a shallow pocket for its replacement. As this pocket contains no replacing crown it is reasonable to assume that the functional canine has not been naturally shed but lost post mortem, otherwise the animal would have been without a functioning canine for an inconveniently long time. The right canine has its crown very well preserved. As preserved, it is rather short and greatly recurved and directed much outwards—this may be unnatural and really due to the aforementioned dorso-ventral crushing the skull has undergone. At its base the crown has diameters of 42 and 24 mm. and a crown length of at least 60 mm. . THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 79 ae [ao wy we OO Blie © & Ronee inate at BEHANY 1B 42 E tH ‘ of vaste a Fic. 18. Fonkeria haughtoni. Type S.A.M. 4343. x4. A = ventral view of the anterior half of the palate. B = lateral view of the right side of the snout. So ANNALS OF THE SOUTH AFRICAN MUSEUM Postcanines On the right side some of the crowns are preserved, whereas on the left only the roots are to be seen in section. The backward slant of the preserved crowns is again due to the crushing the skull has undergone. On both maxillae there is evidence of 19 postcanines, occupying on the right 187 mm. and on the left 188 mm. On the right side no replacing teeth can be determined and No. 8 is represented by an empty alveolus. On the left there are replacing teeth lingual of Nos. 1 and 3 and lingual to the empty alveoli of Nos. 12 and 13 is another replacement. Alveoli Nos. 16 and 17 contain two small tips and alveoli 10, 12 and 13 are empty. Jonkeria boonstrat S.A.M. 9079. Rietkuil, B.W. Collected Boonstra 1929. This specimen (figs. 7 and 8), which has been referred to Janensch’s type, consists of a weathered disarticulated skull, with good upper jaws and part of the mandibular ramus. Inctsors In the right premaxilla the roots of five incisors are present, with pockets in the inner alveolar walls of Nos. 1, 3 and 5, in which lie the crown tips of replacements 1 and 3, whereas no crown tip has as yet been formed for No. 5. the replacement order is 1, 9515,.%, ¢- On the left side the roots of Nos. 1, 4 and 5 are definitely present, whereas those of 2 and 3 are indefinite. Four pockets in the inner alveolar wall can be seen, each housing a tip of the replacing tooth. The replacement order is 5, 2, 4, 1, 3. The lingual face of replacing No. 5, which is well developed, shows coarse striae, and the edge, which in outline is elongated triangular, is coarsely serrate with blunt rounded serrations. Canines The canine roots have the following diameters: left 40x39 mm., right 40 X30 mm. Lingually of each canine there is a pocket in the alveolar wall indicating where the replacing canines are to appear. Postcanines Both maxillae show a good series of roots of functional postcanines and crowns of the replacing set. On both sides the complete set consisted of 18 teeth, with the first larger than the rest, which gradually decrease in size posteriorly. On the right side the series measures 169 mm. and on the left 170 mm. in length. On the right side Nos. 9, 10 and 13 are represented by empty alveoli and lingual to Nos. 1, 2, 5, 8 and 11 are pockets for replacements and lingual to 6, 9, 10, 13 and 15 there are tips of the replacing teeth showing. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS SI On the left side No. 6 has an empty alveolus, linguo-anteriorly of No. 1 there is a fairly large pocket for a replacing tooth, and lingually of Nos. 1, 2, 3, 5, 6, 10 and 11 there are replacing crowns. The large pocket may be for a tooth anterior to the one I am considering the first postcanine, which would then be the second. The erupted crowns of the replacing postcanines are triangular to spatulate in outline with serrate edges composed of blunt rounded cusplets and the inner face is finely striate. The Lower Faw Of the anterior part of the left dentary (fig. 8) I have cut three slabs in frontal section to show the roots of the teeth in use and the stages of development of the replacing tooth. Section A shows the roots of the four functioning incisors, the canine and the first postcanine. Lingually there are cavities behind Nos. 2 and 4 which contain the beginnings of the cusps of the replacing incisors. Lingually of Nos. 1 and 3 there are empty cavities as yet not showing any dentine of replacing Nos. 1 and 3. Lingual of replacing No. 2 there is an empty cavity presumably for the successor to replacement No. 2. Lingually of the canine root is a pocket which contains the tip of the crown of the replacing canine lying loosely in the pocket. Section B still shows the functioning teeth roots; behind No. 2 a crescent of dentine in a large cavity is that of replacement No. 2; replacement No. 1 is represented by an empty cavity and lingual of replacement 2 the cavity for its replacement is also still empty. Two empty cavities situated far lingually may be for the second replacements of 3 and 4. The replacing canine is here seen to occupy the antero-labial part of a cavity. Section C. Besides the roots of the functioning teeth this section shows that labially of the root of the functioning No. 4 incisor there lies the still unresorbed root remnant of an older fourth incisor. The two empty cavities are presumably those of replacement No. 4 and the second replacement of No. 2. Jonkeria vanderbylr S.A.M. 11884. Skroefpaal, P.A. Collected by Boonstra and Rossouw in 1948. This fairly well-preserved and complete skull, with the lower jaws preserved in natural occlusion, has been referred to Broom’s type (figs. 19 and 20). In separating the lower jaws from the upper, the crown of the right lower canine has broken off and lies in its occlusional position, viz. between the upper canine and the 5th upper incisor; similarly the crown of the upper right canine is broken off near its base and lies in its occlusional position, viz. labial to the first 4-5 lower postcanines. 82 ANNALS OF THE SOUTH AFRICAN MUSEUM Incisors Both premaxillaries carry five functional incisors. The No. 1 pair are smaller teeth lying close together. On the left side the inner alveolar wall is fenestrated behind Nos. 1, 3, 4 and 5, and in each pocket lies the crown of a replacement tooth, triangular in outline and with its inner face striate. The replacement order is 5, 1, 3, 4, 2. In the dentaries four incisors are preserved on the left side and three on the right side—No. 1 being lost. No replacements have been exposed. Fic. 19. Fonkeria vanderbylt. S.A.M. 11884. x 4. A = ventral view of the palate. B = anterior view of the snout. This specimen again very clearly shows the intermeshing of the upper and lower incisors. The pair of slender upper Nos. 1 having their talons passing in between the lower Nos. 1, with the outer halves of their heels meeting the inner halves of the heels of the lower incisors; upper No. 2 passes between lower Nos. 1 and 2, and No. 3 between 2 and 3; No. 4 between 3 and 4, and No. 5 between lower 4 and lower canine. Canines The upper canines are large, pointed and curve backwards. The length of the crown is at least 90 mm. and the basal diameters are right 49 x 36 mm. and left 45 X 34 mm. In the pockets lingual of each canine lies the tip of the replacing canine. The lower canines are smaller than the upper and are directed forwards and outwards. The basal diameters are right 29 x 27 mm. and left 27 x 27mm.., and the lengths 50 mm. and 45 mm. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 83 The lower canines intermesh between the upper fifth incisor and the upper canine. Postcanines The upper postcanines are very poorly preserved —in fact, most have fallen out after death and their alveoli are difficult to determine. It is clear, however, that on both sides the first postcanine is much larger than those further back. Fic. 20. Fonkeria vanderbyli. S.A.M. 11884. x 4. A = dorsal view of mandibular arch. B = lateral view of the snout. In the dentaries the postcanines are fairly well preserved. On the right side 18 teeth occupy 180 mm.—Nos. g and 16 are represented by empty alveoli and lingual of alveolus No. g lies a replacing crown. In the left dentary 19 teeth occupy 180 mm. Nos. 18 and 19 are represented by empty alveoli. Lingual to Nos. 6 and g lie replacing crowns and lingual to No. Io is an empty alveolus. Jonkeria truculenta S.A.M. 12030. Bosluiskraal, Laingsburg. Collected Boonstra 1956. This specimen (figs. 5, 21, 22, 23, 24, 40 and 41) consists of a fairly good though somewhat distorted skull, which I am referring to van Hoepen’s species. 84 ANNALS OF THE SOUTH AFRICAN MUSEUM In both upper and lower jaws of the left side the dentition is particularly well preserved and on the right side sectioning shows complementary features. On the left side the dental formula is: i 2, c t, pc 73%. Inctisors In the left premaxilla (fig. 21) the root only of No. 1 is preserved, whereas the other four incisors have most of the crowns present. The heels of Nos. 3 and 4 are worn flat, whereas in No. 2 the heel has a fairly sharp ridge, bounding a hollow lying between it and the talon, and in No. 5 the inner face of the talon is bipartite with the outer part, which receives the lower canine, the greater. Lingual of each incisor there is a pocket in the inner alveolar wall of the functioning incisors of which Nos. 3 and 4 house replacing tips and the others as yet empty. The replacing order is T, 4) 3) 5. 2 In the right premaxilla I have cut two frontal sections (fig. 24B): Re -,* ee re LIS - Sieine<) (eae ree Qe ne 7, n't eras, PS . ee of = wo st. i‘ QU O > Sassen DOTA wa iy O. fess seen ess * “ Cr gk GOOQLDSGOO BON. Ses “29 ee LT oe ree 1. shows the roots of five incisors with No. 1 a slender tooth. Five lingual pockets are present, of which Nos. 2, 3 and 4 contain the talons of the replacing incisors and the other two are empty. ears = 7 arse 2. again has the roots of the five functional incisors; cavities 2 and 3 contain replace- ment teeth and Nos. 1 and 4 are empty. The replacing order is 2, 4, 3, 1 and 5. ~ In the left dentary (fig. 23) four incisors + are preserved; No. 4 has lost its crown; No. 3 F is well worn so that the occlusial surface of the IG. 21. Jonkeria truculenta. ‘ SAN neo ee heel is smoothly flat and the talon so worn on Ventral view of left half of snout. its lingual face that all that remains of the talon is a weak cone; No. 2 is much less worn and has a fairly strong talon; No. 1 just shows the tip of a talon which appears to be piercing a plug of cancellous bone filling the alveolus. Three frontal sections of the right dentary (fig. 244) show: 1. Just below the alveolar level four incisor roots; lingually of Nos. 2 and 4 there are pockets in whose anterior part lie sections across the talons of the replacing teeth; lingually of No. 3 is an empty pocket. The flange of the dentary lying lingually of these pockets is seen as a loose strip of bone. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 85 2. Lying 9 mm. deeper this section shows replacement No. 2 lying in its pocket, a cavity housed the replacement of No. 3 and cavity No. 4 is empty. Lingually of the functional canine root lies a pocket housing the replacing canine. 3. Here an empty cavity for replacement of No. 1 is seen and a replacement of the third postcanine. Bloc eo Fic. 22. Fonkeria truculenta. S.A.M. 12030. x t. Inner face of left mandible. Canines The left upper canine is beautifully preserved. The crown length is go mm. and at its base its diameters are 42 and 29 mm. It is thus a long, moderately strong tooth and has only a moderate posterior curvature; its posterior face is rounded with no indication of a cutting edge. The right upper canine has diameters of 50 and 24 mm. A small pocket for the left replacing canine lies linguo-posteriorly of the functional canine. The left lower canine is fairly well preserved; its length is 41 mm. and its basal diameters are 24 and 31 mm. It is thus quite a moderately sized tooth. It is directed both forwards and outwards. Just behind its tip there is an oval hole which is a result of resorption of the dentine. Its antero-dorsal face is excavated to form a heel bearing some similarity to that typical of the incisors. The fifth upper incisor fits with the postero-external part of its heel into this heel of the lower canine. The mode of intermeshing is thus as is typical of the incisors. Lingually of the canine base there is the usual notch in the flange of the dentary for the replacing canine. In the right dentary, sectioning shows a canine root with diameters 30 X 20 mm. and a pocket containing the tip of the replacing canine. 86 ANNALS OF THE SOUTH AFRICAN MUSEUM Fic. 24. Jonkeria truculenta. S.A.M. 12030. x 4. A = Three consecutive frontal sections through the right Fic. 23. Jonkeria truculenta. dentary. S.A.M. 12030. xX f. B = Two frontal sections through Dorsal view of left mandible. the right upper jaw. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 87 Postcanines In the left upper jaw (fig. 21) a well-preserved series of postcanines follows immediately on the canine without any diastema. ‘The length of the series is 179 mm. and there were 18 teeth. No. 1 is much larger (diameters 13 X 11 mm.) than its immediate successors (diameters 10X8 mm.) and then the teeth gradually decrease in size posteriorly (the last 5 x 4 mm.). No. 4 is represented by an empty alveolus and No. 11 by the crown of a replacing tooth. The bulbously spatulate crown is very well shown in a number of teeth. The crowns lie obliquely to the long axis, with the posterior edge directed outwards and the anterior edge inwards. The posterior edge is sharp and is separated from the main bulbously swollen part of the crown by a shallow longitudinal groove. The inner face of the crown is flattened. In the left dentary (figs. 22, 23) a good set of crowns has been exposed from the lingual side. The series of 18 teeth occupy a length of 179 mm. No. 8 is represented by an empty alveolus. No. 9 is represented by a nearly fully erupted crown of a replacing tooth and so is No. 16. The functional lower post- canines show on their lingual faces a moderately developed heel forming a cingulum at the junction of crown and root; above this the inner face of the crown is concave with longitudinal striae. In the unworn replacing teeth those striae are more definite and extend to the edges of the crown, which thus become serrate with rounded cusplets. The first 4-5 lower postcanines lie lingually of the upper canine in occlusion and the rest lingually of the upper postcanines with no indication of any inter- meshing. As the upper and lower series of postcanines occupy the same length of jaw, follow immediately on their respective canines, and the lower canine lies in front of the upper canine, it follows that the posterior upper postcanines (four of them) have no lower antagonists. The first upper postcanine lies labially of the 6th lower postcanine to constitute its antagonist. Jonkeria ingens S.A.M. 11573. Klein-Koedoeskop, B.W. Collected Boonstra 1940. S.A.M. 11574. Klein-Koedoeskop, B.W. Collected Boonstra 1940. S.A.M. 11575. Klein-Koedoeskop, B.W. Collected Boonstra 1940. Twenty years ago I excavated three quite good skulls lying together in a thin sandstone layer. In all three, few of the crowns of the teeth are preserved. Of 11573 (fig. 25) I have cut a frontal section across the premaxillaries and the other two have been serially cross-sectioned for the study of the internal structure which is being published shortly. What has been seen of the teeth confirms what has been described in the foregoing pages for other Jonkerta species, with little of taxonomic importance. 88 ANNALS OF THE SOUTH AFRICAN MUSEUM Fic. 25. Jonkeria ingens. S.A.M. 11573. X ¥. Frontal section through the premaxil- laries to show the replacing incisors. Fic. 26. Fonkeria ingens. Fic. 27. Jonkeria sp. S.A.M. 11486. x t. S.A.M. 12248. x 4. Lateral view of the right side of the snout. Dorsal view of right dentary. Upper Incisors In the right premaxilla the roots of the functional Nos. 2, 3, 4 and 5 are present with No. 1 represented by an empty alveolus. Lingually of these lie a well-developed replacement No. 1 and 3, the former in a separate alveolus and the latter growing into the old alveolus; replacement No. 2 is just developed and for replacements of Nos. 4 and 5 there are empty alveoli. In the left premaxilla there is no trace of the functional No. 2, and No. 4 is being resorbed, Nos. 3 and 5 have well developed roots and so has the slender No. 1. Replacement No. 1 is seen in its pocket and Nos. 2, 3, 4 in separate cavities, and for No. 5 there is an empty alveolus. 2 THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 89 S.A.M. 12248. Skoppelmaaikraal, Laingsburg. Collected Boonstra 1959. This specimen (fig. 26) consists of the weathered disarticulated bones of both jaws, of which the right dentary is almost complete and shows cross- sections of the roots of the whole set of teeth. Incisors There are four functioning incisors and lingual to No. 2 there is an erupting talon of a replacing tooth lying in a separate alveolus. Canines At its base the canine has the diameters 37 x 26 mm. and, lingual to it, above the flange of the dentary, lies the pocket for its replacement. Postcanines There were 19 postcanines occupying 189 mm. of the jaw. The first post- canine lies immediately behind the canine. No. 5 is represented by an empty alveolus and lingual to the empty alveolus of No. 14 lies its replacing crown. Jonkeria sp. S.A.M. 11486. Mynhardtskraal, B.W. Collected Boonstra 1940. This specimen (fig. 27) consists of a fair snout of which the left half has been serially cross-sectioned and the right half prepared externally to show the dentition. Incisors The crowns of the five right incisors are preserved in a fair state. The postero-lateral face of No. 5 receives the antero-internal attritional face of the lower canine, which clearly intermeshes between it and the upper canine. In the right dentary the first three incisors are functional, whereas No. 4 has fallen out, but has a well-developed replacement crown occupying the lingua] part of the alveolus. The tip of No. 2 replacement is also well erupted and pockets for Nos. 1 and 3 are present. The replacement order is 4, 2, 3 and 1. Canines The upper canine is a large conical tooth only slightly recurved. The crown is 82 mm. in length and the basal diameters 50 xX 22 mm. The root is very large with a roomy pulp cavity, its length is at least 100 mm. The lower canine is short (crown length 32 mm.) and weak with basal diameters 26 x 21 mm. It is directed outwards to lie between the upper canine and the upper fifth incisor against which it bears. Postcanines The upper postcanines follow on the canine without diastema. There is evidence of 19 teeth occupying 183 mm. This specimen clearly shows that the go ANNALS OF THE SOUTH AFRICAN MUSEUM first postcanine differs from the rest; it is directed parallel to the direction of the canine and it appears to be more pointed than spatulate. Its antero-posterior diameter is 14 mm. as against 11 mm. in the second postcanine and 7 mm. in the last postcanine. Nos. 14 and 18 are represented by empty alveoli with presumably replacements lying lingually. In the dentary there are also 19 postcanines occupying a distance of 176 mm. The postcanines have their crowns situated obliquely to the sagittal plane. ner, lee Lol oy; 2Ow?OPOO 2Qo0? o9 © * Fic. 28. Jonkeria sp. S.A.M. 5017. xt. Dorsal view of mandibular arch. This specimen again clearly shows that, since the postcanines follow immediately behind both upper and lower canines and the lower canine passes anterior to the upper one, the lower series cannot extend so far posteriorly as the upper series even if they were of the same length. In fact the lower series is 7 mm. shorter and ends 50 mm. anterior to the level of the last upper post- canine. The last 6 upper postcanines thus have no antagonists and the first 4-5 lower postcanines have the upper canine as antagonist. Jonkeria sp. S.A.M. 5017. Abrahamskraal, P.A. Collected Haughton 1917. This specimen (fig. 28) consists of a poorly preserved skull lacking upper jaws and two imperfect mandibular rami. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS OI Incisors On the right dentary there are roots of the four incisors and lingually of Nos. 1 and 2 replacing talons are developed. The flange of the dentary extends high up behind the replacing teeth as in Jonkeria dubius. Canines The base of the canine crown is seen in section and the diameters are 35 x 28 mm. Lingually there is a pocket in the alveolar wall and a notch in the flange of the dentary for the emergence of the replacing canine. Postcanines On the right side there is evidence of a series of 17 postcanines occupying 173 mm. Nos. 7 and g are represented by empty alveoli lingual to which are the crowns of their replacements. A further replacing crown lies lingual of No. 4. On the left side the dentary is incomplete anteriorly to the 3rd postcanine. Posteriorly there is evidence of 16 postcanines occupying 140 mm. Nos. 7, 9, 11 and 17 are represented by empty alveoli lingual to which lie erupting crowns of their replacements. Jonkeria sp. S.A.M. 9161. Wakkerstroom of Wolwefontein, P.A. Collected Boonstra 1929. The specimen (fig. 9) consists of parts of a broken-up skull, including pieces of the upper jaw and much of the mandibular arch. Incisors In the right dentary four functional incisors are present, but both Nos. 1 and 3 are on the point of being shed with their respective replacements already well developed. This is particularly so in No. 1, where a lucky fracture shows the mode of replacement very well (fig. 98). The root of functional No. 1 is nearly resorbed and in its alveolus lingually to the remains of the root lies the already well-developed replacement. The pulp cavity of functional No. 1 is closed, whereas that of its replacement is large and widely open. The talon of the replacing tooth is well shown as well as the hollowed face of the heel with its rounded lingual ridge. Lingual to the heel the inner alveolar wall is seen in section and is thus a fairly thin sheet of bone; lingual to this there is a deep groove and still further lingually the flange of the dentary with its free upper edge is seen in section. In the left dentary No. 1 incisor is represented by an empty alveolus in whose lingual part there lies the talon of the replacing tooth. Canines On the left a section across the base of the canine crown is seen, which is much laterally compressed with diameters 29 and 19 mm.; on the right there is a broken crown directed much outwards with basal diameters 32 x 24 mm. Q2 ANNALS OF THE SOUTH AFRICAN MUSEUM Postcanines On the right a series of 16 postcanines occupying 154 mm. is preserved. No. 1 is much larger (diameters 11 x8 mm.) than its successors in the series. Nos. 4, 11 and 13 are represented by erupting alveoli with replacing crowns lingual to Nos. 4 and 13. No. 12 is the remains of the root of a functional tooth with an erupting alveolus lingual to it. Jonkerwa sp. S.A.M. 9162. Wakkerstroom of Wolwefontein, P.A. Collected Boonstra 1929. This specimen (figs. 4 and 29) consists of the anterior half of a skull in which the bones of the jaws have become partially disarticulated and shifted from their natural position. The figures given here are reconstructions. qo00O0D ~. Fic. 29. Jonkeria sp. S.A.M. 9162. x4. Ventral view of the anterior part of the palate. Incisors In both premaxillaries the crowns of the incisors are considerably damaged; there are five pairs with the first pair slender and situated close together. This specimen shows the intermeshing of the upper incisors with the lower, as already described, very well. On both sides the five pockets for the replacing incisors have been exposed. In the right premaxilla the replacing crowns of Nos. 2 and 4 are visible in their pockets and they show the elongate inner face of the talons with the edges coarsely serrate. The replacing order is 2, 4, 5, 3, 1. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS £8; pode): [Ot 2 Iio).03 eee 2° Q ote 4 ‘ PCi—* Fic. 30. Jonkeria sp. x t. .M. 11980. Dorsal view of right dentary. M. 12021. Dorsal view of right dentary. 6) PC! 19090 VDO*O BiG 32) Jonkeria sp. Juv. Swale Wiig Sox Dorsal view of right dentary. 93 Fic. 31. Jonkeria sp. S.A.M. 12024. X {. A = dorsal view of the anterior end of the mandibular arch. B = a fracture through the maxilla gives a longitudinal section through the left canine. O4. ANNALS OF THE SOUTH AFRICAN MUSEUM In the left premaxilla the replacing crowns are visible in pockets Nos. 1, 3, 4 and 5, and the replacing order is 5, 1, 3, 4 and 2. On the right side the incisors in the dentary are in the natural occlusional position with the upper incisors, and the intermeshing of the four lower with the five upper incisors is very definite. Canines The lower canine in occlusion lies in front of the upper canine passing outwards in between this and the upper fifth incisor. It is a moderately sized tooth with its greater basal diameter 29 mm. On the right side there is a large upper canine very little recurved; the length of the crown is at least 80 mm. with basal diameters of 43 and 28 mm. Postcanines The crowns of the postcanines are only preserved in some of the more anterior teeth. ‘They have the usual swollen spatulate crowns lying obliquely to the long axis. On the right maxilla there is evidence of 17 teeth occupying 142 mm. Nos. g and 10 are represented by empty alveoli and lingual to Nos. 8, 9, 10, 12 and 13 replacing teeth are present. Jonkeria sp. S.A.M. 11980. Lammerkraal, P.A. Collected Haughton 1916. This right dentary (fig. 30A) shows the incisor roots of Nos. 2, 3 and 4 and an empty alveolus for No. 1, lingual of which is a well-developed talon of the replacing tooth; pockets contain the replacements of Nos. 2, 3 and 4. The canine at its base measures 34 X 26 mm. Nine closely packed postcanine roots are preserved with replacements lying lingually of Nos. 1 and 2. This may be a second specimen of Jonkeria cloetet. Jonkeria sp. S.A.M. 12021. Skoppelmaaikraal, Laingsburg. Collected Boonstra 1956. This specimen (fig. 308) consists of a left premaxilla and a right dentary in which the teeth are seen in cross-section. Four functioning incisors, a canine and 19 postcanines are evident with a replacing 18th postcanine. Jonkerva sp. S.A.M. 12024. Skoppelmaaikraal, Laingsburg. Collected Boonstra 1956. The specimen (fig. 31) consists of weathered broken pieces of both upper and lower jaws. HM ¥ , f , THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 95 The interesting features shown are: in the left dentary incisors Nos. 1 and 2 are functional, with well-developed talons of their replacements lying lingually; in Nos. 3 and 4 well developed replacing talons are visible lying loosely in the alveoli. In the right dentary No. 1 alveolus is occupied by a very well-developed replacing tooth, which is just coming into use. The inner edge of the heel is unworn and is sharp with fine striae and the face of the heel is deeply hollowed out. In a piece of the left maxilla fracturing has revealed the root of the canine in longitudinal section. The root is curved strongly posteriorly and has a roomy pulp cavity, which opens on to the inner maxillary face just behind the strong internal maxillary bulge. Jonkeria sp. S.A.M. 12151. Danskraal of Bloudraai, P.A. Collected Boonstra 1957. This small dentary (fig. 32) is of interest in that it appears to be that of a young reptile. The canine is small with basal crown diameters 29 and 22 mm. Notwithstanding its assumed youth there is already an empty postcanine alveolus (No. 5), but No. 3 is a newly erupted crown. No. 1 incisor has a pulp cavity widely open at its base, but lingually in the same alveolus there is already a replacing talon with an undeveloped root. One can thus with reason state that in youth the incisor replacement occurs rapidly and probably often. In alveoli Nos. 2 and 4 lie talons just appearing above the alveolar border, and in No. 3 there is a root lying loosely in the alveolus. Lingual to No. 4 there is already a pocket for the next tooth. Jonkeria sp. S.A.M. E. This specimen bears a field number E, but both the collector and the locality are unknown. The specimen (figs. 2 and 3) consists of a left premaxilla and maxilla not in articulation with four incisors broken away from the premaxilla and one root still in the premaxilla. Incisors Of the incisors there is preserved: the root of No. 5; No. 4 has part of the root and crown preserved and this is in position in the piece of premaxilla, which also holds Nos. 3 and 2 in natural relation and the latter two consist of good crowns with only the tips missing; No. 1 consists of a detached crown. All the incisors have the same general build although differing somewhat in size and proportions. 96 ANNALS OF THE SOUTH AFRICAN MUSEUM Measurements as reconstructed: ji 2 3 4 5 Total height of crown and root ? 1@ fe) 110 ? ig Height of crown 45 53 53 ? Ly Maximum width of talon 14 15 18 ? ? Maximum width of heel 13 16 17 ? ? Ant.—post. length over heel 31 30 29 24 ? Height of talon 26 32 30 ? ? No. 1 is thus shorter and more slender than the others. In the incisors the crown and root are roughly two curved cones of about equal size with their bases applied to each other, so that in side view they together have an outline resembling a sickle moon. At the junction of crown and root there is on the inner face of the crown a ledge, shelf or heel. The inner face of this shelf looks like a cingulum lying obliquely with the outer part situated higher than the inner. This inner face is convex and on the occlusal face presents a rounded edge lying behind the half-moon-shaped depression lying between it and the inner face of the talon. The roughly conical cusp lying below the heel is here referred to as the talon. The inner face of the talon is flattened with a low median and lateral ridges. Canines The canine is a large, somewhat recurved conical tooth, slightly laterally compressed so that its section is oval. The length of the crown is 110 mm. and at its base the diameters are 45 X 30 mm. From about half-way down the crown its postero-external face forms a fairly sharp edge but without any trace of serrations. It thus appears to be a remnant of a truly cutting edge. Postcanines The postcanines follow close on the canine without any real diastema. A series of thirteen teeth is preserved of which the first twelve occupy a length of 118 mm. They decrease in size from front to back. The first postcanine has the crown imperfectly preserved; it is directed anteriorly, whereas the others incline posteriorly; it is the largest postcanine and its greater diameter at its base is 12 mm. as against the 8 mm. of No. 2. The postcanines all have the same general build, viz. a quite long stem oval in section, then a neck followed by a swollen spatulate crown triangular in outline. The crowns lie obliquely to the long axis of the skull, with the anterior edge lying lingually and the posterior edge labially. The outer face of the spatulate crown is convex with longitudinal striae and sharp, somewhat everted external as well as internal edges which are serrate when unworn. The inner face is flattened to concave. No. 6 has fallen out, but lingually there lies the erupted crown of its 3 le Oe at! THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 07 replacing tooth with a flattish crown and serrated edges. No. 11 has its outer face coarsely striate and the edge serrate. COMPARISON OF THE TITANOSUCHIAN DENTITION WITH THAT OF THE OTHER DINOCEPHALIAN INFRA-ORDERS Earlier in this paper I have stressed a distinctive character common to all the infra-orders of the Dinocephalia and mentioned certain striking features in the dentition peculiar to each infra-order. A more detailed consideratior is, however, called for. ANTEOSAURIA In the mature anteosaurian skull the dental formula is i 4, c 4, pc &. The postcanine series thus has a small number of teeth. Incisors The incisors are long recurved conical teeth, the greatest length encoun- tered being 100 mm. The housing of these greatly lengthened teeth is made possible by a retreat of the premaxillary alveolar border. Although most incisors studied have a simple conical form with an oval cross-section, there is in the South African forms a faint indication in some teeth of a differentiation of a talon and a slight heel, which is apparently more pronounced in some Russian forms. A fact which would strengthen our argument for close consan- guinity with the other dinocephalians. The intermeshing of the incisors is similar to that of the titanosuchians, but the fifth upper incisor is an antagonist of the lower canine by passing labially to it in occlusion and not anteriorly and does thus not intermesh with it. The replacement of the incisors in the Anteosauria has not been adequately studied, but in a series of five frontal sections which I have cut in one specimen (fig. 33) it is evident that the replacing incisor arises lingually of the functioning tooth as it does in the Titanosuchia. The manner of eruption, however, shows certain differences. The development of pockets in the lingual face of the alveolar wall of the functioning incisors for the emergence of the replacing incisors, which is so typical in the premaxillaries of the Titanosuchia, has not been found in the Anteosauria. In fig. 33 the five sections show: Sections 1 and 2: The former section passes just below the alveolar border and the latter just above. On the right side the roots of Nos. 2, 3 and 4 are those of functioning incisors. On the left the functioning incisors Nos. 2, 3, 4 and 5 are seen and lingual of No. 5 lies a section cut through the tip of the lower canine. Section 3: On both sides the young developing crowns of pair No. 1 appear and on the left the remains of the root of an older member of the family of No. 2 incisor is seen as an as yet unresorbed remnant. 98 ANNALS OF THE SOUTH AFRICAN MUSEUM Section 4: The developing right No. 5 incisor can now be seen. Section 5: The replacing tooth of the as yet unerupted right No. 1 incisor is already commencing its development, i.e. before the other older incisors develop any replacements. Canines The upper canines are large slightly recurved pointed teeth with crowns up to 120 mm. in height, flattened from side to side with a sharp posterior edge in which no serrations have been seen. ‘The lower canines are not directed outwards to pass in between the upper fifth incisor and the upper canine as they Fic. 33. Anteosaurus abeli. S.A.M. 9123. XH. Five consecutive frontal sections through the snout showing the replacement of the incisors. do in the Titanosuchia, but are directed upwards and have their tips received in a hollow in the maxillary surface anterior to the base of the upper canine. Practically nothing is known of the replacement of the canines, but in one specimen an upper canine appears to have its crown about to be shed and lying posterior to it is a pointed tooth which may be its replacement. If this observation be correct then the replacement of the anteosaurian canine wou!d be different from that of the Titanosuchia and be similar to that known in the contemporary Therocephalia and Gorgonopsia. Postcanines In the Anteosauria the postcanines are inadequately known. They form a short series with at most 8 teeth, irregularly spaced. The crowns are spatulate, bluntly conical in outline and bulbously swollen from side to side. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 99 TAPINOCEPHALIA The dentition of the South African Tapinocephalia is as yet inadequately known (fig. 34). This is mainly due to the fact that in most of the jaws known the teeth have been lost before entombment, and in many cases where they have been petrified while still in the jaws they have on subsequent erosion dropped out or have had the crowns broken off or otherwise damaged. As the Tapinocephalia have in neither upper or lower jaw teeth specialized as canines, the usual connotation cannot be employed. The teeth in the pre- maxilla will be referred to as incisors and the others as maxillary and dentary teeth. A dental formula on this basis for the Tapinocephalia would thus read: Premaxilla 3-5, Maxilla 8-17 Dentary 14-19 Whether the differences in the count given here have much taxonomic value is to be doubted, and in absence of the necessary facts I am more inclined to suggest inadequate observation coupled with the probability that age is an important factor. It is more than probable that the incisor count of 3-5 may, as is the case in the Anteosauria, only signify a difference in age, with 5 incisors the count for the mature animal. All the teeth of the mature upper jaw intermesh with those of the mandible, but this intermeshing becomes gradually less definite posteriorly. All the more anteriorly situated teeth when mature have the same general build, with the crown consisting of a piercing pointed cusp or talon and lingually a cutting and/or crushing heel in general similar to those of the incisors of the Titanosuchia. As will be evident from figure 34 the anterior or incisor teeth of the Tapinocephalia differ, however, in detail from those of the Titanosuchia. Some of the differences in the incisors are here enumerated: (1) The roots are laterally compressed and the linguo-labial diameter greatly increased thus giving a very flat oval cross-section (diameters 24-40 by 13-19 mm.); (2) the crowns are similarly laterally compressed and the linguo-labial diameter over the heel greatly increased (diameters 31-43 by 16-20 mm.) ; (3) the talons are much shorter; in the Titanosuchia the crown is about the same length as the root, whereas in the tapinocephalian incisors it 1s about five-eighths ; (4) the nature of the heel shows considerable differences; in the Tapino- cephalia the occlusial face of the heel presents a fairly deep oval hollow bounded labially by the talon, anteriorly and posteriorly by a sharp edge with a slight spur lingually. The anterior (or inner) edge lies much higher than the posterior (or outer) edge. Further back in the series the tapinocephalian teeth gradually lose these incisor characters and begin to look more like titanosuchian incisors, except that the talons remain short. The labio-lingual diameter gradually decreases 100 ANNALS OF THE SOUTH AFRICAN MUSEUM Fic. 34. Tapinocephalus. S.A.M. K203. x $. Four anterior teeth of the right upper jaw shown: A = anterior (or inner) view. B = posterior (or outer) view. C = lingual view. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS IOI in length, the sharp ridges on the heel become lower, the face of the heel loses its concavity and even develops a median ridge. Still further posteriorly the linguo-labial character becomes still less and the lingual edge of the heel begins to look very titanosuchian-like, although much smaller in size. Near the posterior end of the series the teeth lose the heel altogether and the crown becomes spatulate and looks very like those of the postcanines of the Titanosuchia (fig. 35). We may summarize the above observations as follows: In a series of functional teeth in the tapinocephalian jaw the posterior teeth have spatulate crowns, which, as one proceeds anteriorly, step by step Fic. 35. Struthiocephalus whaitsi. S.A.M. 3012. x4. Six right upper teeth lying loose under a skull from which they have dropped. A = posterior (or outer) view. B = lingual view. develop a pointed talon and a heel, the heel commencing as a lingual thickening, becomes, firstly, a cingulum, then a step; this step develops a bounding edge with a face between it and the talon; the edge becomes sharper and the face hollower and this hollow increases its linguo-labial diameter. In one genus (Agnosaurus) evidence is preserved that a set of unerupted small spatulate teeth is being replaced by a younger set, which has crowns composed of a talon and heel. This replacement commences anteriorly and proceeds posteriorly, but is not evident anterior to the 6th or 5th tooth in the dentary. T apinocephalus sp. S.A.M. 12139. Rietfontein, P.A. Collected Boonstra 1957. In a weathered skull of Tapinocephalus (fig. 36), where only roots and empty alveoli are preserved, some features of the replacement of the teeth are shown. 102 ANNALS OF THE SOUTH AFRICAN MUSEUM In the left premaxilla the roots of five functioning incisors are preserved. Lingual to No. 5 there is a large replacing root lying in the same alveolus as the functioning root which has been partly resorbed. In No. 4 a functioning root lies loosely in its alveolus and lingual to it in a separate alveolus, a section across the replacing root is present. Lingual of No. 3 there is an empty alveolus and lingual of No. 2 there lies a root of its replacing incisor. In the right maxilla replacing roots are preserved lying lingual to maxillary teeth Nos. 2 and 1o. In the left maxilla the roots of the first two maxillary teeth are seen in section; Nos. 3 and 4 are represented by a large confluent alveolus; of No. 5 the small tip of a root is embedded in the bone. No. 6 is represented by an empty alveolus. No. 7 is again a root-tip embedded in the bone. No. 8 is represented by a small root-tip lying loosely in an alveolus confluent with the Fic. 36. Tapinocephalus sp. S.A.M. 12139. XH. Fic. 38. Struthiocephalus sp. S.A.M. 12050. x4. Ventral view of the anterior part of the palate A frontal section through the anterior end of the showing tooth replacement. upper jaws to show the tooth replacement. Fic. 37. Tapinocephalus sp. S.A.M. 12139. xX. Three frontal sections through the left dentary showing tooth replacement. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 103 empty alveolus of No. 9. No. 10 is again a small root-tip lying loosely in the alveolus. Nos. 11 and 12 are represented by empty alveoli. I have cut three frontal sections through the anterior end of the left dentary to determine the nature of the replacement (fig. 37). Section 1. The 2nd functioning tooth shows a root being resorbed, with the remnant of a still earlier root lying antero-internally. And lingual to both Nos. 1 and 2 lie sections through the talons of their replacing teeth. Lingually of Nos. 10 and 12 is the tip of the talon of the replacement of No. 11. Section 2 shows replacements of Nos. 1, 2, 3, 4, 6 and 8 and an empty cavity lingual to No. 5. Further lingually, cavities for a second set of replacements are seen for Nos. 2, 4, 5 and 6. Section 3 shows replacements of Nos. 1, 2, 4 and 8 and a cavity for No. 3 first replacement and a cavity for No. 3 second replacement. Struthiocephalus sp. S.A.M. 12050. Modderfontein, Laingsburg. Collected Boonstra 1957. I have cut some frontal sections through the snout of a good skull of Struthiocephalus, but as most of the teeth have been lost before petrifaction the evidence obtained of the replacement of the teeth is poor (fig. 38). In the right premaxilla can be seen: a narrow root of No. 1 with lingually to it the tip of the talon of its replacement. No. 2 shows the partially resorbed root of the functioning tooth with lingual to it a well-developed replacement. Of No. 3 there is a large oval alveolus from which the functioning tooth has dropped out, and in its lingual end lies the tip of the replacing talon. No. 4 shows a well-developed, but as yet unerupted, crown with talon and heel. In the maxilla large irregular alveoli can indistinctly be seen with a section through a replacing talon lying loosely in the first maxillary alveolus. In the right half of the upper jaw there is evidence of twelve teeth of which Nos. 1, 8, 9, 10 and 12 were functioning, No. 11 is represented by an empty alveolus and Nos. 3, 4, 6 and 7 are represented by replacing teeth. On this scanty evidence we may tentatively state that the replacement in the Tapino- cephalia takes place lingually of the functioning teeth as it also does in the Titanosuchia and the Anteosauria. The anterior (incisor) teeth do not emerge through pockets in the alveolar walls of the functioning teeth as they do so typically in the Titanosuchia. In this they agree with the condition, also ten- tatively, seen in the Anteosauria. COMPARISON WITH THE CONTEMPORARY T'HERAPSIDS The relevant comparative facts are known in a number of the contemporary Therocephalia and Gorgonopsia of the Tapinocephalus zone. 104 ANNALS OF THE SOUTH AFRICAN MUSEUM In these older therapsids there are usually more (1-2) incisors, but the number of postcanines is always much smaller than in the Dinocephalia. Primitively there were two functional upper canines, but in most the upper canines have been reduced to one only. All the teeth are simple, conical but with serrations developed, usually on the posterior edge, in the incisors and canines. The dentition in these two sub-orders was thus carnivorous. As in the Dinocephalia each tooth is replaced lingually by a member of its own family and this replacement occurs more than once. In the Dinocephalia there is a single alveolus for the upper canine, whereas in these Therocephalia and Gorgonopsia there are two alveoli. Primitively these two alveoli house functional canines at the same time, but in the other forms the functional canine appears alternately in the two alveoli. PELYCOSAURIA In the older and more primitive Pelycosauria there may be as many as g incisors and up to 45 maxillary teeth. In the later Sphenacodontia a pair of teeth becomes functional canines. All the teeth are simple, conical, but sometimes become serrate. Each tooth is replaced lingually by a member of its own family. Primitively the functional replacement of a tooth is by the replacing member of an adjacent tooth—thus distichially, but in the more advanced Sphenacodontia the functional replacement is by a member of its own family, except in the case of the upper canines which are replaced alternately. THE ORIGIN AND FURTHER DEVELOPMENT OF THE DINOCEPHALIAN DENTITION The story of the origin and further development of the dinocephalian dentition starts from some such point as seen in the less specialized Pelycosauria. Both upper and lower jaws have a large number of teeth —all being simple conical structures, with the upper set, in occlusion, passing labially of the lower set. The dentition is of an unspecialized ‘carnivorous’ nature. Each functioning tooth is during life replaced a number of times by a tooth, which arises lingually to it, but as each functioning tooth is lost its function is not immediately taken over by the tooth arising lingually to it but this function is assumed by the tooth arising in the adjacent alveolus (distichial replacement). This simple type of dentition is changed, when in the maxilla two of the more anterior maxillary teeth become larger than the rest and are then called ‘canines’. Not much later a further change takes place, when each of the teeth, not canines, is functionally as well as actually replaced by the tooth arising in the jaw lingually to it. The pair of canines is replaced alternately, i.e. when Pe THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 105 both are not functioning at the same time either the anterior or the posterior one assumes the function of both. This is the stage reached by the Sphenaco- dontia. The next stage is one to be postulated for the ancestral therapsid or pro-therapsid. Here the dentition would consist of about 7 upper incisors, a pair of canines, which are the first two teeth on the maxilla, numerous post- canines; one incisor less in the dentary, a single lower canine and a long row of postcanines. All the teeth would be simple conical teeth, with probably serrated edges in the incisors and canines. From some such pro-therapsid stage the various primitive therapsid sub-orders diverged. One of these branches represents the sub-order Dino- cephalia. In the ancestral dinocephalian the dentition must have been as follows: 5 upper and 4 lower incisors, a single upper and lower canine and numerous postcanines. All the teeth were still simple conical structures with or without serrated edges. Each functioning tooth was replaced by a tooth arising lingually to it. In occlusion all the teeth of the upper jaw passed labially of the teeth of the lower jaw and the dentition was definitely of a habit usually inter- preted as being ‘carnivorous’. In the next stage, this carnivorous dentition with upper teeth occluding labially of the lower teeth, changed to a carnivorous dentition in which the incisor teeth commenced to intermesh with each other. This development of intermeshing incisors was a step which definitely split the Dinocephalia off from the other Therapsida and set them off on their own separate course, and it is the possession of this distinctive character that we consider to be a major justification of our considering the Dinocephalia to constitute a distinct sub-order. Although continuing to retain this common character the Dinocephalia very soon started to develop divergent features. The first dichotomy appears to have split off the infra-order Anteosauria from the rest; from the latter branch a further dichotomy split off the Titanosuchia and still later the Tapinocephalia and Styracocephalia developed as two divergent branches. Anteosauria Titanosuchia Tapinocephalia Styracocephalia ANTEOSAURIA In the nature of its dentition the Anteosauria appear to have changed least from the above postulated condition. 106 ANNALS OF THE SOUTH AFRICAN MUSEUM The incisors have retained the simple conical structure, the canines are retained and the postcanines are still simple teeth. But certain specializations have taken place, viz. the original fairly small incisors have become greatly elongated, and this increase in length is made possible by a concomitant upward retreat of the premaxillary edge; the canines have greatly increased in size and provision for the reception of the tip of the lower canine is made by the development of a recess in the lower face of the maxilla antero-internally of the base of the upper canine; the postcanines have been reduced in number (maximum of 8) and the simple pointed crown has become spatulate with a bulbous outer face. These specializations have changed the original simple carnivores into anteosaurs, with a highly efficient grab, pierce-and-tear dental instrument. With a slinking habit (deduced from the structure of the limbs) such a dentition would allow of a sudden pounce on its prey, and the tearing out of a lump of flesh. The degeneration of the postcanines indicates that no shearing, cutting or grinding took place prior to the deglutition of the lump of flesh helped by the teeth on the palatine boss. TITANOSUCHIA In the Titanosuchia the development away from the ancestral condition was much greater and this can be described as follows: The incisors have completely lost their original simple pointed form and have become highly specialized piercing, cutting and crushing teeth. The strong point, cusp or talon situated labially is a highly efficient piercing instrument and it is held very firmly in the jaw by a large strong root. The well-developed heel situated lingually on the crown has a structure adapting it very well for cutting and crushing. Although such incisors could execute a bite into flesh of no mean order they can hardly be considered to be efficient piercing and cutting instruments adapted for biting into flesh and the crushing abilities of the heel would serve no reasonable purpose when applied to meat. For piercing, cutting and crushing hard fibrous vegetable tissue these teeth would be highly efficient and well adapted. The large canines can reasonably be considered a relict of former car- nivorous days, retained and even enlarged to be employed complementary to the incisors for piercing and severing fibrous material. This is most probably the reason why the lower canines have changed their direction to become teeth intermeshing between the last upper incisor and the upper canine and thus really forming part of the incisor-battery. The postcanines have not been reduced in number and may even have been increased and their serrate spatulate crowns forming a low row along the margins of both jaws could very well have been of use in vegetable feeding. TAPINOCEPHALIA In the ‘Tapinocephalia the development away from the ancestral condition has been greater than in both the Anteosauria and the Titanosuchia. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 107 In the former only the incisors intermesh and in the latter the canines are, as a further step, included in this intermeshing, and now as a third step the Tapinocephalia have the maxillary set of teeth intermeshing with the posterior dentary teeth, with a concomitant loss of the specialized canines. The inclusion of the maxillary teeth in this intermeshing commences anteriorly and then gradually extends to the end of the tooth rows. This process includes the change from spatulate teeth to teeth in which a talon and heel is developed. It has been indicated how the change from spatulate teeth to talon-and-heel teeth has taken place in the maxillary and posterior dentary teeth, but in no specimen have we seen any indication that premaxillary and anterior dentary teeth (incisors) were once spatulate and then became tooth-and-heel teeth. Though in some Russian anteosaurs Orlov (1959) found that the conical incisors develop an attrition face that may be the beginning of a heel. The point now seems to be whether from a simple conical ancestral incisor the Titanosuchia and the Tapinocephalia developed talon-and-heel incisors independently or whether it was the Titanosuchia who took this step and handed the condition on to the Tapinocephalia. If the latter alternative is accepted, then it must be maintained that more or less at the same time, the large titano- suchian canines also became talon-and-heel teeth fitting nicely into the tapino- cephalian series. We have noticed the development of attritional faces in the titanosuchian canines which can be interpreted as indicating such a develop- ment of talon-and-heel teeth. Certain differences in the structure of the incisors of respectively Titano- suchia and Tapinocephalia have already been pointed out. Now, do these differences indicate that talon-and-heel teeth have been developed on two separate occasions independently of each other, or do the similarities rather point to a common origin? The tapinocephalian incisor can be readily derived from the titanosuchian incisor. If in the titanosuchian incisor the length of the talon is reduced, the labio-lingual diameter of the heel increased and the inner edge of the heel raised to form a high and sharp cutting edge then we have a tapinocephalian incisor. If such a development took place it could have been caused by a change in the nature of the vegetable food. The reduction of the size of the talon of the incisor together with the loss of the specialized canines could be due to the food becoming less hardy and fibrous and thus not necessitating piercing, prising and tearing off. The improved cutting abilities of the heel would at the same time be a better adaptation for the cutting of leafy vegetation. I am thus of the opinion that the tapinocephalian dentition is a develop- ment of that present in the Titanosuchia and that the Tapinocephalia hived off from the Titanosuchia after these had acquired the typical talon and heel incisors. I have based the above account mainly on the South African material, but hope to correlate with the Russian material in the near future. 108 ANNALS OF THE SOUTH AFRICAN MUSEUM THE ARTICULATION AND MOVEMENTS OF THE LOWER JAW (Figs. 39-41) The quadrate has on its ventral surface a pair of convex condyles separated by a groove and these fit fairly accurately into two concavities separated by a ridge on the postero-dorsal face of the articular. The quadrate is firmly fixed in the skull, being firmly clasped by the squamosal, braced by the quadrato- Fic. 40. Jonkerta truculenta. S.A.M. 12030. xt. A = posterior view of the jaw articulation. B = posterior view of the lower jaw. jugal, butted against by the paroccipital and held by the quadrate ramus of the pterygoid. The articular is also firmly wedged in between the angular and surangular and the prearticular. The only possible movement is thus at the joint between quadrate and articular. Fic. 39. Ventral view of the quadrate. x 4. A = Parascapanodon avifontis S.A.M. 9127. B = Parascapanodon sp. S.A.M. 12026. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS 109 ith the quadrate = outer view. iculating w B f the left mandible art with the mouth closed. . Fic. 41. Jonkeria truculenta. S.A.M. 12030. }. Inner View O . A= IIO ANNALS OF THE SOUTH AFRICAN MUSEUM The two condyles on the quadrate do not lie in the same plane—the outer lying posterior to the inner. In section they present segments of circles with radii of (in $.A.M. 12026) 26 and 32 mm. which are thus not co-centric. The radii lie in a plane diverging from the sagittal plane in posterior direction. The result of this structure is that when the jaws are opened the movement of the jaws is a sliding one in posterior direction of at least 15 mm. and the backs of both jaws at the same time move outwards for at least the same distance. At the same time the lower edge of the jaw appears to move inwards. The latter two movements are rendered possible by the fact that the two rami of the jaws are not firmly fixed to each other at the symphysis. What effects do these movements of the jaws have on the bite of a titanosuchian ? The forward movement of the jaws when the mouth is closed would push the talons of the lower incisors from behind forwards between those of the upper incisors and the heels of the lower incisors would slide under those of the upper teeth thus executing a crushing and cutting action. The upper and lower canines would move relatively as do the talons of the incisors, and the lower postcanines would have a cutting action when moving past their upper antagonists. When the jaws are closed the articulars move towards each other with the result that the more posterior postcanines also move inwards and would thus pass inwards of the upper postcanines leaving some space between the upper and lower teeth. If the lower mandibular edge, when the jaw is closed, moves outwards, the alveolar edge would move inwards and this would still further increase the space between upper and lower postcanines. This would be functionally disadvantageous! Faunistic Hitherto the main elements of the fauna of the Tapinocephalus zone were divided into two groups representing the opposing herbivores and carnivores as follows: herbivores: Pareiasauria, Tapinocephalia and Anomodontia. carnivores: ‘Titanosuchia (including the Anteosauria), Therocepha- lia and Gorgonopsia. This presumed balance always appeared to be loaded in favour of the carnivores, for not only were the Titanosuchia known to be large reptiles but were in addition numerous, as judged by the frequency in which remains (mostly poor, it is true) are encountered in the field. In the new picture the herbivores are (and this is more natural) better represented, the largest forms of the times being the herbivores. In order of size the herbivores of this zone are: Titanosuchia, Pareiasauria, Tapinocephalia and Anomodontia, and the carnivores: Anteosauria, Therocephalia and Gorgonopsia. THE DENTITION OF THE TITANOSUCHIAN DINOCEPHALIANS IST The more marsh-dwelling herbivores were the Pareiasauria and the Tapinocephalia; the more upland-dwelling forms the Titanosuchia and Anomodontia. Of the carnivores the Anteosauria, with large heads, were more of a slinking habit; the Therocephalia more active upland reptiles of prey, and the Gorgonopsia just starting their role as important carnivores. ACKNOWLEDGEMENTS In the collection of the material here described I am pleased to be able to express my gratitude to all those farmers in the Karoo who have so kindly allowed me to work on their property, helped me in various ways, and always been most hospitable. Mr. H. Zinn, technical assistant of the Museum, has accompanied me on many of the collecting trips and has been most helpful. In recent years C.S.I.R. has given me a grant which covered the current expenses of the collecting trips undertaken. Mr. C. Gow has during the last year assisted in the preparation of many of the jaws studied. His painstaking work, chiefly with a vibro-tool, has produced some really adequately prepared specimens. The sections made of a number of specimens were rendered possible by the use of a diamond-studded rotating saw purchased with a grant made by Gs R. The Trustees of the South African Museum gratefully acknowledge a grant towards the cost of publishing this paper made by the South African Council for Scientific and Industrial Research. REFERENCES Boonstra, L. D. 1936. The cranial morphology of some titanosuchid deinocephalians. Bull. Amer. Mus. nat. Hist. 72, 99-116. Boonstra, L. D. 1953. A suggested clarification of the taxonomic status of the South African titanosuchians. Ann. S. Afr. Mus. 42, 19-28. Boonstra, L. D. 1953. The cranial morphology and taxonomy of the tapinocephalid genus Struthiocephalus. Ann. S. Afr. Mus. 42, 32-53. Boonstra, L. D. 1954. The cranial structure of the titanosuchian: Anteosaurus. Ann. S. Afr. Mus. 42, 108-148. Broom, R. 1923. On the structure of the skull in the carnivorous dinocephalian reptiles. Proc. zool. Soc. Lond. 1923, 661-684. Broom, R. 1929. On the carnivorous mammal-like reptiles of the family Titanosuchidae. Ann. Transv. Mus. 13, 9-36. Broom, R. 1932. The mammal-like reptiles of South Africa. London: Witherby. Erremov, J. A. 1940. Ulemosaurus svijagensis Riab- ein Dinocephale aus den Ablagerungen des Perm der UdSSR. Nova Acta Leop. Carol. 9, 155-205. Haueuton, 8S. H. 1915. On two new therocephalians from the Gouph. Ann. S. Afr. Mus. 12, ama Ee JANENSCH, W. 1959. Eine Jonkeria aus der Karru-Formation des Kaplandes. Paldont. £. 33, 22-49. Ortov, J. A. 1959. [Carnivorous Deinocephalia (Titanosuchia) from the Upper Permian deposits of the Middle Volga.] Trud. paleont. Inst. Acad. Sci. U.R.S.S. 72, 1-114 (in Russian). Owen, R. 1879. Description of fragmentary indications of a huge kind of theriodont reptile ( Titanosuchus ferox, Ow.) from Beaufort West tract, Cape of Good Hope. Quart. F. geol. Soc. Lond. 35, 189-199. I12 ANNALS OF THE SOUTH AFRICAN MUSEUM Van Hoepen, E. C. N. 1916. A new Karroo reptile. Ann. Transv. Mus. 5, No. 3, suppl. 3,1 p. Watson, D. M. S. 1914. The Deinocephalia, an order of mammal-like reptiles. Proc. zool. Soc. Lond. 1914, 749-786. Watson, D. M. S. 1948. Dicynodon and its allies. Proc. zool. Soc. Lond. 118, 823-877. Watson, D. M. S., & Romer, A. S. 1956. A classification of Therapsid reptiles. Bull. Mus. comp. Kool. Harv. 114, 37-89. - Ann. S. Afr. Mus. Plate XI d reconstruction of the whole -SIZE fe i ically in the South African Museum. is that of al 1oram The head ited d i 6 SS Gq ~*~ o 15) q or} SH e) Gq aus) CS} lo} oO ae, J ~~ fo) 19} o = o 12) e) g n os} = Photograph of the head of Parascapanodon avifont repti References thus appear as follows: AUGENER, H. 1913. Polychaeta. In Michaelsen, W., ed. Beitrdége zur Kenntnis der Meeresfauna Westafrikas. 2, 67-625. Hamburg: Friederichsen. EKMAN, 9S. 1953. ® re > } qu ISSUED JANUARY 10962 PRICE R1-63 PRINTED FOR THE TRUSTEES OF THE SOUTH AFRICAN MUSEUM BY THE RUSTICA PRESS (PTY.) LIMITED, COURT ROAD, WYNBERG, CAPE INSTRUCTIONS TO AUTHORS MSS. submitted for publication must be typewritten, double spaced with good margins, typed on one side of sheet only and sheets numbered. Two MSS. and one set of illustrations must be submitted. Papers must include a Table of Contents and a Summary containing a succinct account of the subject, results obtained and conclusions. The position of the text-figures and tables must be clearly indicated. ILLUSTRATIONS. Please must be kept to a minimum and made up to occupy 5" x 7” (full plate excluding captions) or in direct proportion to these measure- ments if they are to be reduced. A scale (metric system) must appear with all photographs. REFERENCES. Harvard system—authors’ names and dates of publication given in the body of the text; references arranged at the end of the paper in alphabetical order of authors’ names. Bibliographical Arrangement of references at the end of the paper must give: 1. Name of author, followed by his initials. Names of joint authors connected by ampersand (&), and not ‘and’. 2. Year of publication. If several papers by the same author in one year are cited, suffixes a, b, etc., are used. 3. Full title of the paper. Initial capital letters only for the first word and for proper names, except in German, where the usage of the language is followed. 4. Title of the journal, abbreviated according to the World list of scientific periodicals, and underlined to indicate italics. 5- series number, if any, im parenthesis, €.¢. (3), (n-s.), (N-Fjyiep 6. Volume number in arabic numerals (without prefix ‘vol.’), with wavy underlining to indicate black type. 7. Part number, only if the separate parts of one volume are independently numbered. 8. Page numbers, first and last, in arabic numerals (without prefix ‘p’). When reference is made to a separate book, give in the order listed: Author’s name; his initials; date of publication; title, in italics; edition, if any; volume number, if any, in black type arabic numerals; pagination, if reference is to part of book only; place of publication; name of publisher. When reference is made to a paper forming a distinct part of another book or collection, give: Name of author of paper; his initials; date of publication; title of paper; ‘In’, italicized; name of author of book; his initials; title of book, in italics; edition, if any; volume number, if any, in black type arabic numerals; pagination of paper; place of publication; name of publisher. [CONTINUED ON INSIDE BACK COVER FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS By R. H. Mitiar Marine Station, Millport, Isle of Cumbrae, Scotland [Accepted April 1959] (With 45 figures in the text) CONTENTS PAGE Introduction . : : a hrs Station list 3 : 4 fy ADORE List of new species and forms ._ 119g Description of species. § Lag Distribution of species . oo e2t7 Summary . ; ; ; A esi) References : ‘ : 220 INTRODUCTION In a previous paper (Millar, 1955) I described a number of ascidians from the coasts of South Africa, collected during an ecological survey carried out by the Zoology Department of the University of Cape Town. A continuation of this survey has produced more ascidian material, which forms a large part of the subject of the present paper. The principal areas from which this new material was collected (fig. 1) are: Langebaan Lagoon, Table Bay, False Bay, Mossel Bay, Algoa Bay, and Mozambique. Small amounts of material were also taken from several other places. The collecting was mainly from sub-littoral areas, and this makes difficult a strict comparison with the faunistic results obtained from the extensive surveys of Stephenson (1939, 1944, 1948), whose studies refer to littoral areas. Material is also included from the South African Museum, mainly belonging to the Preter Faure collection, and some South African specimens from the British Museum (Natural History), most of these latter being from Morten- sen’s Java—South Africa Expedition. Although most of the specimens formed the material of an ecological survey, the present paper deals with systematics. I have had certain difficulties owing to the still unsatisfactory state of the systematics in some ascidian genera; this is particularly true of Polyclinum, Didemnum and Eudistoma. As I have explained when discussing Dzdemnum stilense I believe that certain species will be defined adequately only when a large series of specimens is available from different areas and seasons, and when the biology is studied in addition to 113 Ann. S. Afr. Mus. 46 (7), 1962, 113—221, 45 figs. ANNALS OF THE SOUTH AFRICAN MUSEUM 114 “SoTBIO] SuNdsa][0I Moys 0} woLyW YyyNog jo yred yo deur yoxoyg “1 ‘OIG Asen\sq JaAry apaaig JOAlY S uewUsSNg uoogeq ueegasue] << Aeg euyuepyes ueqing Aeg 1SO> auaquinso/\| Aeg asje Aeg |asso ‘Secs | en ejnsuluag ade eushuy| Aeg eosiy y : a - Le celecl teas SE FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS PES preserved material. Such needs are, of course, difficult to meet, and impossible in the course of a general ecological survey. For these reasons some of the identifications are provisional, and some material has not been identified even provisionally, as that course might lead to greater confusion. I am indebted to Professor J. H. Day, of the Zoology Department of the University of Cape Town, and to Dr. J. F. Croil Morgans, formerly of that Department, for information regarding the specimens, and also to the Zodlogisch Museum of Amsterdam for the loan of type specimens of many of Sluiter’s South African species. The Trustees of the Museum gratefully acknowledge the grant in aid of publication of this paper received from the South African Council for Scientific and Industrial Research. STATION List Most of the material belongs to the ecological survey of the University of Cape Town, and to the Preter Faure collection of the South African Museum. Details of stations relating to these two collections are given below. Collecting details of other material mentioned in this report are given in the text. UNIVERSITY OF CAPE TOWN ECOLOGICAL SURVEY SALDANHA BAY U.C.T., SB 1, Station 3, 13—7—46. U.C.T., SB 91, Station 13, 14—7—46. U.C.T., SB 92, Station 8, 14—7—46. U.C.T., SB 94, Station 3, 13—7—-46. U.C.T., SB 128, Salamander Bay, 7 m., 9-4—53, mud weed and sponge. U.C.T., SB 130, off Donkergat Point, 7-9 m., 9-4—54. U.C.T., SB 134, 4-3-37, from bottom of Szgnis. LANGEBAAN LAGOON U.C.T., LB 352, Oosterval, balanoid zone, 4—5—53, rocks. U.C.T., LB 367, Skaapen Island, lower balanoid zone, 7—5—53, weed and stones. U.C.T., LB 374, south side of Skaapen Island, lower balanoid zone, 7—5-53. U.G.T., LB 375, Skaapen Island, upper balanoid zone, 7—5-53. U.C.T., LB 378, Langebaan jetty, 0-1 m., 7—-5—53, wooden pilings. U.C.T., LB 384, Oosterval, L.W.N.T., 7-5-53, sandy limestone. U.G.T., LB 472, 6-5-55. TABLE BAY U.C.T., TB 171, 33° 49°5’ S., 18° 27°5’ E., g-18 m., 4-8-46, shells or sand and shells with occasional rocks and mussels. U.G.T., TB 172, Station 3, 33° 50°35’ S., 18° 20’ E., 15:5 m., 3-7-47. U.C.T., TB 173, Station 5, 35° 52-7’ S., 18° 26-8’ E., 20-5 m., 3-7—47. 116 ANNALS OF THE SOUTH AFRICAN MUSEUM CAPE PENINSULA U.C.T., CP 393, Oatland Point, intertidal, 27—9-54. U.C.T., CP 425, Oatland Point, 14 inches above low water to below lowest Poiniinceres. 12-10-54. FALSE BAY U.G.T., FAL 5, Station 3, just south of Seal Island, 35 m., 22-2-52, sand. U.C.T., FAL 53, Gordon’s Bay, 18 m., 25-6—52. U.C.T., FAL 108, Gordon’s Bay, 34° 09-3’ S., 18° 51’ E., 7-11 m., 23-1-53, some sand. U.C.T., FAL 109, Gordon’s Bay, approximately 34° 09°3’ S., 18° 51’ E., 7-11 M., 23-11-53. U.G.T., FAL 117, Simon’s Bay, 23:5 m., 12-2—53, broken branched coralline. U.C.T., FAL 121, Glencairn power station, rock terraces, 6 m., 17—2-53, scat- tered rocks next to sand. U.C.T., FAL 130, Oatland Point, 1-2 m. below L.W.S.T., scattered rocks, 26-2-53. U.C.T., FAL 136, Oatland Point, o-2 m., 27—2-—53, vertical rock faces. U.C.T., FAL 137, Gordon’s Bay quay, 0-4 m., 4-3—53, rock. U.C.T., FAL 144, Oatland Point, bottom of Octomeris zone to 5 m. below, 9-3-53, rock face. U.C.T., FAL 145, details as FAL 144. U.C.T., FAL 152, Oatland Point, 4:5-5:5 m., 12-3—53, stones and rocks on sand. U.C.T., FAL 158, Oatland Point, 0-3 m., 22—4-—53, rocks. U.C.T., FAL 163, Oatland Point, S.W. face of Big Rock, 0-2 m., 23—5—53, area free of Pyura community. U.C.T., FAL 167, Oatland Point, S.W. face of Big Rock, area free of Pyura community, 2-4 m. below Echlonia zone, 10—-6-53. U.C.T., FAL 175, Oatland Point, S.W. face of Big Rock, 4—-6:5 m. below Echlonia zone, 10—6—53. U.C.T., FAL 176, Oatland Point, S.W. face of Big Rock, from top 2 m. of — flourishing Pyura community, 10-6-53. U.C.T., FAL 177, Oatland Point, 9-8-53. U.C.T., FAL 181, from bottom of ship General Botha, Simon’s Town, 6—3-37. U.C.T., FAL 182, Oatland Point, lagoon rocks, 7—9-53. U.C.T., FAL 208, Africana, 34° 0-9’ S., 18° 42°4’ E., 36 m., rock, 10—9—53. U.C.T., FAL 212, Africana, 34° 07:1’ S., 18° 35°6’ E., 21 m., 9—9—53, limestone reefs and sand. U.C.T., FAL 216, Africana, 34° 12°4' S., 18° 43°5' E., 42 m., 10-9-53, bottom rocky. U.C.T., FAL 221, Africana, N.W. of Seal Island, 34° 07-0’ S., 18° 32:5’ E., 18 m., 9Q-9—-53, limestone reefs and sand. U.C.T., FAL 225, off Miller’s Point, 34° 13-9’ S., 18° 31:6’ E., 40 m., 9-90-53, coarse sand and finely broken shell. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 117 U.C.T., FAL 2209, off Cape Point, 34° 20-27’ S., 18° 31-8’ E., 64 m., 9-9-53, coarse sand. U.G.T., FAL 232, Africana, 34° 17°35’ S., 18° 31-4’ E., off Buffels Bay, 50 m., nO oS: U.G.T., FAL 234, Africana, South Kogel Bay, 34° 15:35’ S., 18° 44:8’ E., 47 m., 10—9—53, coarse sand with shell, pebbles and stones. U.C.T., FAL 256, Oatland Point, 4:0-5:5m., 17-11-53. U.C.T., FAL 259, Oatland Point, about 80 yards E. of Big Rock, 10:5 m., 2I—11—53, coarse shelly sand. U.C.T., FAL 265, half-way between Seal Island and Strandfontein, 15-25 m., 4-4-54, rock. U.C.T., FAL 272, Roman Rock, 100 yards ESE. of lighthouse, 14-17 m., 18—9—-54, sloping granite rocks. U.C.T., FAL 277, Roman Rock, 14-17 m., 21-9—54, rocks. U.C.T., FB 1102, 21-4-47, 0°5 mile east of Seal Island, 27 m., rock. U.C.T., FB 1105, 28-4-47, 34° 10’ S., 18° 28’ E., 24 m., sand. U.C.T., FB 1106, 20-8—47, off Somerset Strand, 25 m. MossEL BAy Weer WMib 7. 34°04 17 9., 22> 13° 53” E., 19 m., 12-1—56, rock. U.C.T., MB 14, 34° 11’ 04” S., 22° 10’ 09” E., 16 m., 13—1—56, rock. U.G.T., MB 18, 34° 08’ 45” S., 22° 07’ 12” E., 13 m., 13—1—56, rock. U.G.T., MB 22, 34° 08’ 50” S., 22° 07’ 20” E., 12°5 m., 13—1—56, rock. U.C.T., MB 27, 34° 11’ 04” S., 22° 09 55” E., 19 m., 13—1—56, rock. U.C.T., MB 38, 34° 10’ 07” S., 22° 07’ 46” E., 8-5 m., 16—-1-56, sand. U.C.T., MB 39, 34° 10’ 08” S., 22° 08’ 00” E., 9 m., 16—1—56, rock. U.C.T., MB 48, 34° 11’ 19” S., 22° 09’ 58” E., 10 m., 17—1—56, rock. U.G.T., MB 53, 34° 10’ 57” S., 22° 09’ 55” E., 14. m., 17—-1-56, rock. U-@:T., MB 55, 34° 10’ 42” S., 22° 00’ 38” E., 9 m., 17—1—56, rock. U.G.T., MB 58, 34° 04’ 18” S., 22° 13” 32” E., 12-5 m., 18-1—56, rock. U.C.T., MB 59, 34° 04’ 08” S., 22° 13’ 52” E., 11-5 m., 18-1—56, rock. U.C.T., MB 60, 34° 04’ 18” S., 22° 14’ 10” E., 17-20 m., 18—1-56. U.C.T., MB 65, 34° 04’ 47” S., 22° 13’ 06” E., 26 m., 18—1—56, rock. U.G.T., MB 68, 34° 00’ 08" S., 22° 07’ 19” E., 13 m., 19—1—56, rock. U.G.T., MB 69, 34° 08’ 35” S., 22° 07’ 20” E., 13°5 m., 19—1—46, rock. U.C.T., MB 72, 34° 09’ 04” S., 22° 07’ 10” E., 12 m., 19—1—56, rock. U.C.T., MB 84, 34° 11’ 26” S., 22° 10’ 8” E., 29 m., 21-1—56, rock. ALGOA BAY U.C.T., LIZ 1, Station 1, 33° 55-7’ S., 25° 37:2’ E., 8-5-10°0 m., 5-4-54, mud. U.C.T., LIZ 2, details as Station 1. U.C.T., LIZ 3, Station 2, 33° 56-1’ S., 25° 40-0’ E., 17-18 m., 5-4-54, sand. U.C.T., LIZ 9, Station 3, 33° 58-1’ S., 25° 38-9’ E., 9m., 6—4—54, stones and rock. 118 ANNALS OF THE SOUTH AFRICAN MUSEUM U.C.T., LIZ 11, Station 5, 33° 57-2’ S., 25° 38-0’ E., 9-10 m., 6-4-54, lime- stone and grey clay. U.C.T., LIZ 19, Station 8, 33° 58:5’ S., 25° 42:0’ E., 27 m., 7-4-54, sand and shell. U.C.T., LIZ 25, Station 12, 34° 00-4’ S., 25° 44°5' E., 39 m., 11-4-54, coarse sand and shell. U.C.T., LIZ 28, Station 13, 34° 00-8’ S., 25° 42-4’ E., 5-7 m., 11-4-54, rock. U.C.T., LIZ 32, Station 3, 33° 58-1’ S., 25° 38-9’ E., 6—4—-54, stones and rock. U.G.T., LIZ 34, Station 7, 33° 58:4’ S., 25° 40°5’ E., near charted ‘coral’ patch, 14 m., 7-4-54, stones. U.G.T., LIZ 38, Station 8, 33° 58-5’ S., 25° 42-0’ E., 27 m., 6-4-54, sand and shell. MORRUMBENE ESTUARY U.C.T., MOR 43, 20-1-54, channel at Linga-linga, 6-9 m. U.C.T., MOR 50, 20-1-54, channel at mouth of Rio Coche, 3-5 m., sand. U.C.T., MOR 76, 23-1-54, sand at L.W.S.T., Mongue. U.C.T., MOR 77, 23-7-53, Linga-linga, 3 m., sand. U.C.T., MOR o1, 12—7—54, channel at Linga-linga. U.C.T., MOR 108, 14—7-54, edge of channel, Linga-linga, 1-5-4 m., sand. U.C.T., MOR 109; 14-7-54, channel at Linga-linga, 4 m., sand. U.C.T., MOR 122, 15—7—-54, off Linga-linga. U.C.T., MOR 132, 16—7—54, channel opposite Rio Coche, 2 m., sand and shell. U.C.T., MOR 187, 18—7—54, intertidal mud near L.W., N.W. of Mongue. U.C.T., MOR 188, 18—7—54, stones at L.W.S.T., Mongue. OTHER LOCALITIES IN MOZAMBIQUE U.C.T., PEA 1, 26—7—53, attached to iron railings in bathing-pool. U.C.T., PEA 3, Linga-Linga, 10 m., 23—7—-53, sand and weed. U.C.T., PEA 4, Maxixe, 24—7-53, low water, sandbank and piles. COMMERCIAL TRAWLERS U.C.T., TRA 60, Kaffirkuils Bay, 34° S., 21° E., 62—70 m., 25-11-52. U.C.T., TRA 62, 44 miles off Cape Barracouta, 63 m., 25-11-52, covering a dromiid crab. U.C.T., TRA 71, 32° 05’ S., 18° 14’ E., 66 m., 5—2-53, hard Polyzoa and rock. SOUTH AFRICAN MUSEUM Pieter Faure COLLECTION (denoted in text by PF) PF 11, Mossel Bay, 0-55 m. PF 586, 33° 50’ S., 25° 54’ E., 46-92 m. PF 673, 33° 45’ S., 26° 44’ E., near Kowie, 73-79 m. PE1704,°33° 58° 9.5125 51) Eo dom: FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS PF 739, between Roman Rock and Cape Recife, 31 m. PF 740, details as PF 739. PF 895, 32° 47’ S., 28° 16’ E., 50-54 m. PF 1095, between Bird Island and mainland, Algoa Bay, 18-29 m. PF 1711, Cape St. Blaize, N. by E 2? E., 6 miles, 64 m. PF 1864, Cape St. Blaize, N. 36 miles, 99 m. PF 2326, Lion’s Head, N. 67° E., 25 miles, 248 m. PF 2348, details as PF 2326. PF 2361, details as PF 2326. PF 2531, Lion’s Head, N. 63° E., 34 miles, 283 m. PF 10165, Cape St. Blaize, N. by W. 4 W., 5 miles, 62 m. PF 10477, Cape St. Blaize, N. by E., 8 miles, 72 m. PF 10722, Cape Natal, W. by N., 6 miles, 91 m. PF 12393, Itongazi River mouth, Natal, NW. 2 W., 3 miles, 46 m. PF 12943, East London, low tide. PF 13393, Cape Morgan, NW. £N., 11 miles, 160 m. PF 13432, Sandy Point, NE. by E. 6 miles, 94 m. PF 13481, Sandy Point, N. + E., 10 miles, 175 m. PF 14560, Cape Point, N. 50° E., 18 miles, 341 m. PF 14582, details as PF 14560. PF 15584, Rocky Bank, False Bay, 31 m. PF 15797, off Seal Island, False Bay, 18-19 m. PF 15984, off Zwartklip, False Bay, 18-22 m. PF 18785, Cape Seal, W. by N. 4 N., 7 miles, 72 m. LisT OF NEW SPECIES AND FORMS 119 Synoicum capense sp. n. Botryllus anomalus sp. n. S. australe sp. n. Metandrocarpa fascicularis sp. n. Polycitorella pallida sp. n. Cnemidocarpa psammophora sp. n. Tetrazona porrecta sp. n. Styela radicata sp. n. Sigillina vasta sp. n. Boltenia africana sp. n. Eudistoma digitatum sp. n. Halocynthia spinosa Sluiter f. defectiva n. Didemnum (Polysyncraton) magnilarvum sp.n. Molgula cryptica sp. n. Lissoclinum cavum sp. n. Eugyra myodes sp. n. Ascidia stenodes sp. n. Eugyra macrentera sp. n. DESCRIPTION OF SPECIES Family Polyclinidae Verrill, 1871 Genus APLIDIUM Savigny, 1816 As I have pointed out (Millar, 1960) there seems to be no good reason for separating the genera Aplidium Savigny 1816 and Amaroucium Milne Edwards 1842, and I am using Aplidium, which has priority. I20 ANNALS OF THE SOUTH AFRICAN MUSEUM Aplidium flavo-lineatum (Sluiter) (Fig. 2) Amaroucium flavo-lineatum Sluiter, 1898, p. 30, pl. 1, fig. 7; pl. 4, fig. 12. Amaroucium astraeoides Sluiter, 1898, p. 33, pl. 1, fig. 9; pl. 5, figs. 2-5. Hartmeyer, 1912, p. 351, pl. 44, fig. 5, text-fig. 10. Amaroucium simplex Sluiter, 1898, p. 35, pl. 1, fig. 10; pl. 5, fig. 6. Amaroucium erythraeum Michaelsen, 1934, p. 137, pl. 7, fig. 1. Millar, 1955, p. 270, fig. 2. Known distribution Cape Province. A <> OODOOD OC | EE © 00000 ee Oo “o ya AY : Se ae 8h ANIA Rad ; Ts a SNORE NI Le aoee tee: ee ao Fic. 2. Aplidium flavo-lineatum (Sluiter). A, larva; B, stomach of type specimen of Amaroucium flavo-lineatum Sluiter ; C, stomach of type specimen of Amaroucium astraeoides Sluiter; D, stomach of type specimen of Amaroucium simplex Sluiter. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS I2I Localities able bays O.C.0., FB 171 B. Cape Peninsula: U.C.T., CP 393 8; U.C.T., CP 425 R. Feber payenu.cut., PAL 1902; U.GeghAL 163 Ns U.G.0, BAL 167 Y; WEG yl PAL 175 K; U.Cov. FAL 162 f; Pl15797 A;. Mortensen’s Java-South Africa Expedition, Station 65, 19-12-29, B.M.(N.H.) TES. NO. 34.2.1 .4. iMesh bay: W.-C... MB 18 P;°U.C:1., MB 48°A; U.C.1., MB 6o P; Wee ahs MB 65 C. arcnmnay a). Col LEZ 28K: U.Cil., LIZ. 32° T;. PF 704. . Description This species is often red in life, the collectors of the present material having noted that the specimens were ‘red’, ‘blood orange’, or ‘scarlet’, but occasionally ‘colourless’. The variable amount of sand on the surface and within the colony no doubt determines to a large extent its colour in the field. After preservation the colour of the specimens may change, and the material in this collection shows zooids with red, orange-yellow, sulphur-yellow, yellow-green and, in one colony, bright blue colour in the body wall. Most of the specimens in the present collection have quite conspicuous small round or oval systems, each with a small common cloacal opening in the centre. I have examined many colonies and find that the number of folds on the wall of the stomach varies from 19 to 33, but is usually between 25 and 30. Larva Larvae were present in colonies collected in January, February, April, June, September, and December. They generally numbered from 1 to 4 ina breeding zooid. Most larvae measure 0:80—1.00 mm. in length from the end of the papillae to the base of the tail, but some were only 0-64 mm. The anterior papillae have long, slender stalks. There is a variable number, generally 4-6, of finger-like median ampullae, and many short lateral ampullae or large vesicles. Remarks There has been much confusion in the past over this species, largely owing to the inaccurate description of the type specimens given by Sluiter (1898). Sluiter (1898) also described two other species, A. astraeoides and A. simplex, the type specimens of which I have been able to examine along with those of A. flavo-lineatum. In some important points I find Sluiter’s descriptions wrong, and I have no doubt that the three species are synonymous, the name A. flavo- lineatum having page priority. In particular the number of folds on the wall of the stomach is not that given by Sluiter. Careful examination of the type specimens (fig. 2, B-D), particularly transverse sections of the stomachs, 122 ANNALS OF THE SOUTH AFRICAN MUSEUM showed that there are 23—26 narrow folds in each species, although Sluiter gave the following descriptions :— A. flavo-lineatum— to folds. A. astraeoides—areolated stomach. A. simplex—12 folds. The discrepancy is difficult to understand, even assuming that Sluiter made only a rough estimate of the total number from the number on one side of the stomach. This species has also appeared in accounts of South African ascidians under the name A. erythraeum Michaelsen (Michaelsen, 1934; Millar, 1955), but A. erythraeum was originally described from the Red Sea and I am now doubtful if it is the same as the South African species, although anatomically similar. Michaelsen (1934) also considers A. phortax (Michaelsen) from New Zealand to be a synonym. If A. flavo-lineatum, A. erythraeum and A. phortax are synonymous, then we are dealing with a species having an apparently very curious and discontinuous distribution, and I think it more likely that the three species are separate but anatomically similar. It might be possible to distinguish them if the larvae of all three were known. See below for remarks on the possible identity of A. circulatum (Hartmeyer) and A. flavo-lineatum. ? Aplidium circulatum (Hartmeyer) Amaroucium circulatum Hartmeyer, 1912, p. 349, pl. 39, fig. 2; pl. 44, fig. 6. Hartmeyer, 1913, p. 130. Michaelsen, 1934, p. 130. Millar, 1955, p. 171. Known distribution Cape Province. Locality False Bay: U.C.T., FAL 234 W. Description The single colony is approximately mushroom-shaped, about 2-2 cm. across the head, and provided with a short stalk. In the preserved state it is buff-coloured, but in life, according to the collector’s note, the zooids are ‘brick-red’. There is sand within the colony but not on the surface. The stomach of zooids in this colony has about 16 longitudinal folds, com- pared with about 24 in the specimen which I previously described from False - Bay (Millar, 1955). In his original account Hartmeyer (1912) described the folds as numerous but did not state the number. Remarks This specimen agrees in most respects with Hartmeyer’s account of A. circulatum. But, as Michaelsen (1934) has suggested, that species may be the FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 123 same as A. erythraeum (Michaelsen), and I have accepted A. erythraeum as a synonym of A. flavo-lineatum. The distinctions between A. circulatum and A. flavo-lineatum are not great, the main one being the presence of a stalk on the colony. There may also be fewer stomach folds. A much larger series of specimens will be needed, however, to decide whether A. circulatum is a distinct species. Aplidium pantherinum (Sluiter) (Fig. 3) Psammaplidium pantherinum Sluiter, 1898, p. 26, pl. 2, fig. 3; pl. 4, figs. 8, 9. Psammaplidium obesum Sluiter, 1898, p. 28, pl. 1, fig. 6; pl. 4, figs. 10, 11. Amaroucium unilarviferum Millar, 1955, p. 172, fig. 4. B C Fic. 3. Aplidium pantherinum (Sluiter). A, larva; B, stomach of type specimen of A. pantherinum (Sluiter); C, transverse section of stomach of type specimen of A. obesum (Sluiter). 124 ANNALS OF THE SOUTH AFRICAN MUSEUM Known distribution Cape Province and Isipingo, Natal. Localities Table Bay: WU. LB 172;,0.C. 1, 2B 173 Ae Off Cape of Good Hope: PF 2361 A. False Bay: U.C.T., FAL 136 L; U.CG.T., FAL 145 A; Us@ol aie 152 N; U.G.T., FAL. 175°G; U/C. 2.) FAL 256'O: Mossel Bay: U.C.T., MB 69 E. Remarks When I described a new species Amaroucium unilarviferum (Millar, 1955) from Cape Province I had not examined Sluiter’s type specimens of South African material. Having now done so I find such close agreement between A. unilarviferum and two species which Sluiter named Psammaplidium pantherinum and P. obesum that I must regard the three as synonymous. As Psammaplidium is a synonym of Aplidium, the specific name becomes Aplidium pantherinum which has page priority over A. obesum. Only an examination of Sluiter’s type specimens shows the identity of these species, as his descriptions are very misleading with regard to the stomach, a key character in the determination of species of Aflidium. Careful examination of isolated stomachs and of transverse sections of stomachs of the type specimens of A. pantherinum shows 5 folds (fig. 3, B), although Sluiter stated that there were 12 folds. Five or 6 folds are also present in the stomach of the type specimens of A. obesum (fig. 3 CG), which Sluiter stated had 8 folds. The type specimens of A. pantherinum and A. obesum differ from each other principally in external appearance, owing to the peculiar arrangement of sand grains on the surface of the former. This is a character of little or no systematic significance, and the identity of A. pantherinum, A. obesum and A. unilarviferum is shown by the structure of the zooid and of the larva. Larva (fig. 3 A) Larvae range in length, measured from the end of the papillae to the base of the tail, from 0-50 mm. (some larvae from type specimens of A. pantherinum from Isipingo, Natal) to 0-80 mm. (some larvae from type specimens of A. obesum from Sea Point, near Cape Town). The characteristic features of the larva, as shown in all specimens, are the absence of lateral paired ampullae, and the presence of a fringe of small epidermal vesicles round the anterior margin of each side. From 1 to 3 larvae are present in each breeding zooid. The larva which I figured under the name A. unilarviferum (Millar, 1955, fig. 4 D) was evidently not fully developed, as the fringe of epidermal vesicles is not shown; I have subsequently found the characteristic arrangement of vesicles in other larvae from the same colony. The larva can be used to distinguish between A. pantherinum and A. fuegiense Cunningham, a South American species in which the stomach also has 5 or FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 125 6 folds. In A. fuegiense the larval trunk is deep, with prominent paired lateral and unpaired median ampullae, but no vesicles (Millar, 1960), and is thus very different from that of A. pantherinum. Aplidium colelloides (Herdman) (Fig. 4) Amaroucium colelloides Herdman, 1886, p. 233, pl. 27, figs. 9-12. Known distribution Off Cape of Good Hope. Locality PF 18785 A. Description There are several colonies in the collection, all of a very characteristic shape. The colony is divided into an oval, rounded, or almost triangular head and a long narrow stalk (fig. 4 A). In one typical colony the head is 1-5 cm. long and the stalk 6-0 cm. The head is smooth, free of sand, and dull orange-grey in colour. The zooids are clearly visible and are arranged in long narrow oval systems orientated so that the zooids appear to be in vertical double rows. Common cloacal openings, which Herdman (1886) failed to find in the type specimen, are narrow slits placed among the rows of zooids. The stalk is sharply marked off from the head and has a uniform coating of sand grains. In most specimens the lower end of the stalk is attached to a narrow stem-like object which may be the tube of a polychaete. The zooids (fig. 4 B) have a short thorax and abdomen which together measure about 1-5 mm. in an average zooid. The post-abdomen is very long and narrow and extends far down into the stalk of the colony. There are 6 pointed lobes on the oral siphon and the atrial opening is surmounted by a triangular languet of moderate length. The body wall of the thorax is thin and transparent, allowing the 9 rows of stigmata to be seen. Muscles on the thorax are so poorly developed as to be scarcely visible. The abdomen is short and the gut has a correspondingly short loop. There are 12 or 13 entire longitudinal folds on the walls of the stomach. In some zooids the post-abdomen contains an accumulation of reserve material, and in others a long series of testis follicles. Although no ovary was seen several zooids were carrying larvae in the atrial cavity. Larva The larva (fig. 4 C) measures about 0-70 mm. from the end of the papillae to the base of the tail. The three papillae are borne on long narrow stalks. Large numbers of small epidermal vesicles lie round the anterior end of the 126 ANNALS OF THE SOUTH AFRICAN MUSEUM trunk in the region of the papillae. Both ocellus and otolith are present, the ocellus, at least in the preserved state, being red, and the otolith black. Remarks This is an interesting species and the only previous record is that of Herd- man (1886) who described the single type specimen. Herdman’s specimen was collected by the Challenger from a depth of 150 fathoms, a few miles south of the 0.5 mm. 0.3 mm. Fic. 4. Aplidium colelloides (Herdman). A, colonies; B, zooid; C, larva. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS P27) Cape of Good Hope. This colony was incomplete, and gave little idea of the relatively great length of the stalk which apparently is usual. A long stalked colony is exceptional in the genus Aplidium. The present specimens are identified as A. colellordes mainly by the shape of the colony, the proportions of the zooid, and the form of the larva, which Herdman illustrated. Aplidium retiforme (Herdman) (Fig. 5) Psammaplidium retiforme Herdman 1886, pp. 248-9, pl. 32, figs. 8-10. Wauygunndt hans ntuur’ prog Fic. 5. Aplidium retiforme (Herdman). A, colony; B, zooid. Known distribution mm.. Kerguelen. Localities False Bay: U.C.T., FAL 158 8; ?U.C.T., FAL 108 Z. ? Mossel Bay: U.C.T., MB 55 X. ? Algoa Bay: U.C.T., LIZ 9 J. Description The colony FAL 158 S consists of a group of slender club-shaped lobes united basally to a common stolon (fig. 5 A). Sand and broken shell completely cover the surface but do not penetrate within the lobes, where the test matrix is clear. The largest lobes are about 2:5 cm. long and o-8 cm. in greatest diameter. 128 ANNALS OF THE SOUTH AFRICAN MUSEUM Each lobe contains only a few zooids, probably constituting a single system, but common cloacal openings were not seen. The zooids (fig. 5 B) attain a length of 6 mm., or more if the post-abdomen is very long. A layer of quite strong muscles is present on the thorax, which consequently is usually contracted and bent, in the preserved condition. The atrial opening is small and is surmounted by a short languet which has 3 shallow terminal lobes. There are 9 or 10 rows of stigmata. The stigmata are moderately long and narrow, although Herdman (1886) describes those of his specimens as ‘not large’ and shows them in his figure to be short. There are about 16 undivided longitudinal folds on the wall of the stomach. The post- abdomen contains neither ovary nor testes in the present specimens. Remarks This is one of those species which present a problem to the systematist. On the one hand there is close agreement between the structure of the specimen from False Bay and the type specimen from Kerguelen as described by Herdman, | although it must be said that Herdman’s account lacks some important details. On the other hand there is the great difference in temperature which must raise doubts whether one species exists in both localities. The specimens from the three queried localities have smaller colonies, but otherwise are similar to the above description and may belong to the same species. Aplidium sarasinorum (Fiedler) (Fig. 6) Heterotrema sarasinorum Fiedler, 1889, pp. 859-78, pl. 25, figs. 1-14. Known distribution Ceylon. Localities False Bay: U.C.T., FAL 208 S. Mozambique: U.C.T., PEA 3 K; U.C.T., MOR 77 K; MOR 108 L. Locality unknown: S.A. Museum. Description 7 Most of the colonies are growing on and round the stems of algae or hydroids. The specimen from False Bay is 4:2 cm. long, narrow, and flattened to a leaf-like shape and divided for half its length into 2 broad lobes. The consistency is firm and the surface almost smooth, but raised into small round swellings which mark the positions of the zooids. The test is impregnated with sand. Other colonies are more fleshy and less subdivided into lobes, but the shape seems to depend largely on the form of the object to which the colony is 1 FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 129 attached. Sand is generally present in the matrix of the colony, but is sparse or absent on the surface. In structure the zooids (fig. 6 A) resemble the type specimen described by Fiedler (1889). The thorax is short, stout, and often bent in the preserved state, owing to muscular contraction. The oral siphon is short and 6-lobed. The atrial opening is a transverse slit with a wide 3-lobed languet (fig. 6 B). About 8 0.5 mm. Fic. 6. Aplidium sarasinorum (Fiedler). A, zooid; B, atrial languet. powerful longitudinal muscles are present on each side of the thorax and con- verge towards the ventral side of the abdomen. There is a band of shining pale gold cells along each side of the endostyle, and although this feature was not mentioned by Fiedler, it is present in all the South African specimens. The gut has the usual subdivisions found in Aplidium. The stomach has 5-8 wide and rather indistinct folds. ‘The post-abdomen is generally short, and only indistinctly marked off from the abdomen. In some zooids there are no gonads and in others a compact group of testis follicles immediately below the lower bend of the intestine. 130 ANNALS OF THE SOUTH AFRICAN MUSEUM Remarks Fiedler (1889) founded his genus Heterotrema for a colony collected from Ceylon, but Michaelsen (1923, 1924) pointed out the similarity between Hetero- trema and certain species of Aplidium (Amaroucium), and Pérés (1952) regarded H., sarasinorum as a variety of A. circumvolutum (Sluiter). A. circumvolutum was originally taken from New Zealand waters and later Pérés (1952) described specimens from Kerguelen as a variety of the same species, A. circumvolutum var. kerguelenense. Pérés (1952) also described a new species A, antarcticum from Kerguelen, but regarded it as very close to A. circumvolutum and possibly identical. I have no doubt that all the South African specimens listed here under ‘Localities’ are of the same species as Fiedler’s colony, but I do not follow Pérés in identifying Heterotrema sarasinorum with Aplidium circumvolutum. A. sarasinorum is characterized by the flat, expanded and lobed colony, with zooids on both surfaces, the wide 3-lobed atrial languet, the stomach with about 7 folds, the very compact testis, and the short post-abdomen which is not clearly marked off from the abdomen. In A. circumvolutum the colony is more massive and broadly based, the atrial languet long, the stomach smooth-walled (Sluiter, 1900) or with 5 or 6 folds (Michaelsen, 1934), and the post-abdomen more sharply constricted and longer than in A. sarasinorum. In addition to these distinctions, which are admittedly rather slight, except the different shape of the colony, there is the difference in distribution to separate the species. It seems unlikely that the same species should be found at Ceylon and at Kerguelen, and I believe that we are dealing with two species | of similar structure, one distributed from South Africa across the Indian Ocean (A. sarasinorum) and the other of wide distribution in southern waters including New Zealand, Kerguelen, and also the Patagonian shelf, Graham Peninsula, and South Georgia (Millar, 1960). ?Aplidium galeritum (Hartmeyer) Amaroucium galeritum Hartmeyer, 1912, p. 344. Known distribution St. Francis Bay, Cape Province. Locality False bay OU: Cie Paleo re Description The specimen is slightly lobed and rather flattened, and measures 9:5 cm. by 2:7 cm. by 1:0 cm. It is heavily encrusted on the surface and impregnated throughout with sand and broken shell, so that the consistency is very firm. There is no division of the colony into a softer head with zooids and firmer base FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 131 without zooids, which Hartmeyer (1912) found in his type specimen, but I do not know if this is an important difference. As regards the zooids, the structure agrees in most points with A. galeritum, but in the specimen from False Bay there is no very distinct atrial siphon, and there are only 10 to 12 folds on the stomach, whereas A. galeritum has a short atrial siphon and 14 to 18 folds. Larva From 3 to 6 larvae are present in the atrial cavity of breeding zooids. The trunk is about 0-4 mm. long, has the usual 3 papillae of the genus, and a fringe of small epidermal vesicles round the anterior end. Both ocellus and otolith are present. | Remarks I am not sure that this specimen agrees well enough with A. galeritum to be placed in that species, but it may be within the variation to be expected, and I hesitate to add another doubtful species to the South African list. Genus syNoicuM Phipps, 1774 Synoicum capense sp. n. (Fig. 7) Diagnosis of species Colony stalked, with a conical or hemispherical head. Surface of head and stalk coated with sand and shell. Inner test matrix clear, without sand. A few systems of zooids in each head. Zooids with 8-10 rows of stigmata, a very short atrial siphon with moderately wide and long upper lip or languet. Stomach smooth. Post-abdomen of variable length, with a long series of testis follicles. Holotype In the South African Museum, S.A.M. A25608 (U.C.T., FB. 1106C). Localities aiserpayen U,.G)t..F Bb ri0o,.Cs.U.C.L., FAL 53 L. Description The two colonies, which are stalked, are 1-3 cm. and 1°7 cm. long, the head occupying less than half of the total length. The stalk is narrowest at the base and gradually widens towards the upper end (fig. 7 A). The head is almost twice as wide as the upper part of the stalk, and the larger colony is o-6 cm. in greatest diameter. The head is widest at its junction with the stalk and is either hemispherical or conical. Sand and broken shell fragments cover the whole surface of the colony. Common cloacal openings are not visible, but several must exist on each head, since dissection shows the zooids to be arranged 132 ANNALS OF THE SOUTH AFRICAN MUSEUM common test, the inner test matrix of both head and stalk being clear. The zooids (fig. 7 B) may reach 12 mm. in length, but some are shorter. Only the thorax and abdomen are contained in the head of the colony and the post-abdomen extends some distance down within the stalk. . in several systems. Sand and shell are confined to the surface layer of the : 0.5 cm. Fic. 7. Synoicum capense sp. n. A, colony; B, zooid. mm LEK) S°OIMSRagG There are 6 short pointed lobes on the oral siphon. The atrial siphon is very short with the upper margin drawn out into a short languet, which is either quite simple or has 3 small distal lobes. About 6 slender longitudinal muscles pass down each side of the thorax. Eight to 10 rows of stigmata are present in the branchial sac, each row with about 12 stigmata. The oesophagus is of moderate length, and the stomach is ovoid with smooth walls. The post- stomach and the intestine are well-marked. There is little difference in diameter between the abdomen and the post-abdomen, which may however be slightly FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 133 narrower. Some zooids of the larger colony have a fairly long series of testis follicles in the post-abdomen, but none was found with an ovary. In the smaller colony the zooids were without gonads. Remarks This new species perhaps most resembles S. adareanum (Herdman), but is distinguished by the more slender club-shaped colony, the complete covering of sand and shell, and the smaller number of rows of stigmata. From S. arenaceum (Michaelsen) it differs in having several systems on each head and in lacking a conspicuous rim round the cloacal openings. Synoicum australe sp. n. (Fig. 8) Diagnosis of species Colony rounded, smooth and moderately firm. Zooids arranged in small round systems. Oral siphon with shallow lobes; atrial siphon tubular with oblique opening and the dorsal margin slightly produced as a short lip. Longi- tudinal muscles many and narrow. About 20 oral tentacles and 20 rows of stigmata. Dorsal tubercle with a simple transverse oval opening. Oesophagus longer than stomach. Stomach with smooth walls. Anus about half-way along the thorax. Post-abdomen about equal in length to the thorax and abdomen together. | Holotype | In the South African Museum. §8.A.M. A25609 (U.C.T., MB 27 H). Locality Mossel Bay: U.C.T., MB 27 H. Description There is only one specimen, a colony measuring 2-3 cm. by 1:6 cm. by 1-0 cm. The colony is slightly lobed, but there is no obvious area of attachment and it is not apparent how the specimen was fixed to the substratum during life. No foreign matter is present on the surface, which is quite smooth, and there is also none within the colony. The colour, in the preserved state, is translucent grey, with the systems of zooids showing through as paler stellate patterns. The zooids (fig. 8) reach 6 mm. in length and are divided into a thorax and abdomen of about equal length, and the post-abdomen which is as long as the thorax and abdomen together. The thorax is wider than the other two divisions of the body. The oral siphon is short with 6 shallow rounded lobes, and the atrial siphon tubular with an oblique opening. No well developed languet is generally present but the dorsal margin of the atrial opening is produced to form a rounded lip of variable length. Numerous slender longitudinal muscles pass along each side of the thorax, and circular muscle strands surround the 134 ANNALS OF THE SOUTH AFRICAN MUSEUM oral siphon. About 20 tentacles stand at the base of the oral siphon. The dorsal tubercle is small with a simple transverse oval opening. There are up to 20 rows of stigmata, but the number in each row is difficult to count owing to contrac- tion of the thorax. The oesophagus is longer than the stomach, laterally flattened, and slightly curved. It enters the oblique anterior end of the stomach. The stomach is rounded with quite smooth walls. The post-stomach and intestine are together 0.3 mm. Fic. 9. Polyclinum arenosum Sluiter. Larva. 1.5 mm. 0.2 mm. Fic. 8. Synoicum australe sp. n. Zooid. Fic. 10. Polyclinum istpingense Sluiter. Larva. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 135 slightly longer than the stomach. The rectum extends about half-way along the thorax, and ends in a 2-lipped anus. None of the many zooids examined from this colony had any trace of gonads. The post-abdomen contained numerous storage cells. Remarks Except for S. capense just described, no species of Synoicum has yet been recorded from the coast of Cape Province. Michaelsen (1914, 1915) has described Macroclinum angolanum from Angola, and this species should probably be referred to the genus Synoicum. It differs in many respects from S. australe, notably in its much smaller number of rows of stigmata, and much more numerous oral tentacles. Two species described by Sluiter (1898) as Polyclinum arenosum and P. insulsum were recorded respectively from Natal and Mozambique. These were both regarded by Michaelsen (1919) as belonging to the genus Macroclinum, which is now recognized as a synonym of Synoicum. I have examined the type material and have no doubt that Sluiter was right in placing both species in the genus Polyclinum. In appearance of the colony the present species resembles S. gzard: (Herd- man), an Antarctic species, but the conspicuously areolated stomach of the latter readily distinguishes it. The zooid of S. australe is perhaps most like that of S. adareanum (Herdman), a species of the high Antarctic, but the colony forma- tion is quite different. S. curcumvolutum Kott, from Kerguelen and Heard Island, is like S. australe, but the colony is sometimes stalked and the zooids have only 12 rows of stigmata. Another Antarctic species, S. minutum (Herdman), has a colony somewhat like that of S. australe, but the zooids are irregularly scattered and have usually only 7 or 8 rows of stigmata. Genus POLYCLINUM Savigny, 1816 Polyclinum arenosum Sluiter (Fig. 9) Polyclinum arenosum Sluiter, 1898, p. 20, pl. 4, figs. 1, 2. Millar, 1955, p. 174. Known distribution Isipingo, Natal; False Bay, Cape Province. Locality Algoa Bay: PF 1095 A. Description The deep division of the colony into long narrow lobes, which the specimen shows, is typical of the species. The colony is considerably more massive than Sluiter’s type specimen, as the lobes attain a length of 4 cm., compared with I°4. cm. in Sluiter’s material. 136 ANNALS OF THE SOUTH AFRICAN MUSEUM Larva In Sluiter’s type specimens, which I have examined, larvae are present in some zooids, but are in such a bad state of preservation that few details are visible. They measure from 0-32 to 0:36 mm. in length, from the end of the papillae to the base of the tail. Larvae (fig. 9) are also present in the new material from Algoa Bay. These vary in length from 0-46 to 0-60 mm. They have the usual 3 anterior papillae, an ocellus and otolith. Small epidermal vesicles are grouped along the dorsal and ventral parts of the trunk. Before the larva is fully developed there are 4 pairs of lateral ampullae flanking the papillae, but later they sub- divide to give rise to 8 pairs. Polyclinum tsipingense Sluiter (Fig. 10) Polyclinum isipingense Sluiter, 1898, p. 21, pl 2, fig. 1, pl. 4, fig. 3. Hartmeyer, 1913, p. 83. Millar, 1955, p. 175. Known distribution Isipingo, Natal; False Bay, Cape Province. Localities Morrumbene Estuary: U.C.T., MOR 50 J; MOR 76 F; MOR 108 M; MOR 122 F; MOR 132 G; MOR 187 A. Description The colonies are of a rather solid appearance, and often reach 2—5 cm. in greatest diameter. They are generally dull purple-brown in colour, and some have a coating of sand and broken shell, but the surface of other colonies is bare. Small common cloacal openings are scattered over the colony, but are not raised on swellings as they were in the type material. The zooids reach 5 mm. in total length. They are therefore shorter than the zooids of the type material which Sluiter (1898) gives as 9 mm. long, but the length of zooids in Polyclinum, as in most genera of the family, changes much during the life of a colony and is a character of little systematic value. | Larva Larvae (fig. 10) are present in several colonies, and there are usually 2 in each breeding zooid. They range in length from 0-32 to 0-40 mm., measured from the end of the papillae to the base of the tail. There are 3 anterior papillae, and 4 pairs of lateral ampullae. A dorsal and a ventral group of epidermal vesicles is present on each side of the trunk. Polyclinum neptunium Hartmeyer (Fig. 11) Polyclinum neptunium Hartmeyer, 1912, p. 331, pl. 38, fig. 9; pl. 44, figs. 2-4. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 137 Known distribution Off Cape of Good Hope. Localities PF 2326; PF 2428; PF 2531; PF 14560. Description There are many colonies of this species in the Pieter Faure collection, all of a dull purple-brown colour and fleshy appearance. Some are over 7 cm. in greatest diameter and are thus rather larger than Hartmeyer’s type specimens. The surface is generally smooth and free of encrusting matter, except on the basal area of attachment where sand or small stones adhere to the test, but a little sand may also be scattered on the surface. The zooids are arranged in oval or elongated systems, but these are only sometimes visible externally. In structure the zooids agree closely with Hartmeyer’s description. Fic. 11. Polyclinum neptunium Hartmeyer. Larva. Larva (fig. 11) Larvae are present in several colonies. In length, measured from the end of the papillae to the base of the tail, they range from 0:52 mm. to 0-80 mm. The trunk is rather deep, dorsi-ventrally. The larva has the structure usual in Polyclinum, with 3 anterior papillae, a sensory vesicle containing ocellus and otolith, and well developed siphonal rudiments. The ampullary tissue well illustrates the structure pointed out by Carlisle (1952) for Polyclinum aurantium Milne Edwards; on each side a lateral ridge arises from the anterior end of the trunk and sends one posterior branch dorsally and one ventrally. It is from these ridges and their branches that the larger anterior ampullae and the smaller dorsal and ventral vesicles arise; both ampullae and vesicles retain their connection with the ridges by slender stalks. 138 ANNALS OF THE SOUTH AFRICAN MUSEUM Genus PSEUDODISTOMA Michaelsen, 1924 Pseudodistoma africanum Millar (Fig. 12) Pseudodistoma africanum Millar 1954, p. 128. Known distribution Cape Province. Localities False Bay: U.C.T., FAL 158 L; U.C.T., FAL 175 F; PF 15584. Mossel Bay: U.C.T., MB 65 D. Algoa Bay: U.Q.T., LIZ 28 V. Natal: PF 12393 A; off Durban, collection 8.8. Pickle, B.M. (N.H.) reg. no. 1926., 6.1.4, 29° 91 077 5. 81,29 on) Eee am: 0.5 cm. Fic. 12. Pseudodistoma africanum Millar. Larva, possibly not fully developed. Description Most of the colonies agree well in shape and colour with the type specimens (Millar, 1954). In the present collection the largest specimens are one of 8 cm. length from Mossel Bay (U.C.T., MB 65 D), and a complex colony from Natal (PF 12393 A) in which the head and stalk may reach 8 cm. The colony from Natal is dull purple, unlike the other specimens which are some shade of yellow or grey. The collectors have described the specimens variously as colourless or pink in life. In structure the zooids are generally typical of the species, and many have a single very large egg in the post-abdomen. The zooids of a colony from Natal (B.M. (N.H.) reg. no. 1926.8.1.4) have up to 20 longitudinal muscles on each side, this being a greater number than in the type specimens. Larva Only one colony so far collected has larvae (fig. 12) and these were probably not fully developed. This was the colony from Natal (B.M. (N.H.) reg. no. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 139 1926.8.1.4). In some zooids of this colony there was, in the atrial cavity, one very large embryo or larva occupying the whole length of the thorax. The most advanced larvae were about 1 mm. long from the end of the papillae to the base of the tail. Unfortunately details were not clear, but there are 3 anterior papillae in a vertical row, and both ocellus and otolith in the cerebral vesicle. A series of ampullae is also present at the anterior end of the trunk. Remarks Pseudodistoma is somewhat intermediate between the families Polyclinidae and Clavelinidae, having an abdomen and post-abdomen of the Aplidium-type and a thorax of the Eudistoma-type. Such details of the larva as can be seen suggest Aplidium rather than Eudistoma, thus confirming the position of the genus in the family Polyclinidae. Family Clavelinidae Forbes and Hanley, 1848 Genus CLAVELINA Savigny, 1816 Clavelina roseola Millar Clavelina roseola Millar, 1955, p. 183. Known distribution False Bay, Cape Province; Red Sea (as Stomozoa murrayt, Kott 1957). Localities Mossel Bay: U.C.T., MB 60 Q; U.C.T., MB 65 E. Description One colony was collected at each station. The larger specimen is 5 cm. long, 3°5 cm. wide, and 2°5 cm. high. The base of the colony is coated with sand, but the heads are not. Some details may be added to those seen in the type specimen. The musculature consists of numerous quite strong longitudinal strands passing mainly from the intersiphonal region, down the sides of the thorax to concentrate in two wide bands on the abdomen. In the new specimens the folds on the stomach are few and have the appearance of artefacts. It 1s, therefore, probable that the 11 or 12 longitudinal folds described in the type material were also artefacts, and that the stomach in its natural state is smooth. The anus is 2-lipped and lies opposite the 12th or 13th row of stigmata. Sub- divisions of the gut below the stomach correspond closely to those in Kott’s (1957) figure 3 of Stomozoa murrayt. Remarks Stomozoa murrayi Kott, collected from the Red Sea by the John Murray Expedition, is evidently the same as Clavelina roseola and becomes a synonym. The resemblances are very close, in the general form of the colony and zooid, the gut, and particularly in the peculiar and characteristic siphonal lobes. I40 ANNALS OF THE SOUTH AFRICAN MUSEUM The only important difference is the large number of rows of stigmata in S. murrayt (26) compared with C. roseola (17), but the greater development of the branchial sac in the Red Sea specimens may be due to a more favourable environment in warmer waters. Kott (1957) accommodated her species in a new subfamily Stomozoinae, but the siphonal lobes constitute the only character differing markedly from the normal structure of Clavelina. The species seems to be no more than an aberrant member of the genus Clavelina. Clavelina steenbrasensis Millar Clavelina steenbrasensis Millar, 1955, p. 185. Known distribution False Bay, Cape Province. Localities False Bay: U.C.T., FAL 175 E. Mossel Bay: U.C.T., MB 18 Q; U.C.T., MB 38 D. Description There are several specimens in the collection, and these show that the lateral compression of the colony noted in the type specimen is not a common feature of the species. A well-developed colony is about 4 cm. in greatest diameter and consists of a number of upright lobes, the largest about 2 cm. long. The lobes are widest at the top and narrowest at the base where they are united to a small mass of common test. The test is transparent and firm, the dark blue zooids being visible from the outside. A section through the stomach shows only 5 indistinct longitudinal folds, although in the type specimen there were 8 folds. As in other species of Clavelina, however, it is often difficult to decide whether or not some of the folds in fixed material are artefacts. Some zooids have a slender posterior vascular process extending downwards from the end of the abdomen. Remarks I have already pointed out the general similarity between this species and C. gigantea Van Name, but the latter species even if it does not differ in the form of the stomach as I previously believed, is distinguishable by its dark pigmented test and apparently unpigmented zooids, and is recorded from the Gulf of Mexico. There are other specimens which I include in this species with some hesitation. A single piece of a colony (U.C.T., FB 1106 D, 20—8—47) is solid, smooth, and hyaline, and contains zooids like those of C. steenbrasensis but without any pigment. There are also colonies from False Bay (S.A. Museum Pieter Faure collection, no. 17458, 17-22 m.), and from an unknown locality FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS I41I (S.A. Museum Preter Faure collection), which, although having a form of colony resembling the typical specimens of this species, yet have a much softer and more flexible test, and have no pigment on the zooids. It is possible that the consistency of the test depends on the age of the colony, or the conditions of preservation which may also affect the colour of the zooids. Genus cystopiTEs Von Drasche, 1883 Cystodites roseolus Hartmeyer (Fig. 13) Cystodites roseolus Hartmeyer, 1912, p. 310. Known distribution Algoa Bay, Cape Province; Gulf of Guinea, west Africa; Seychelles. Localities Cape Peninsula: PF No. 2361 B. Algoa Bay: between Bird Island and mainland, 18-30 m. East of Algoa Bay: near Kowie, 33° 45’ S., 26° 44’ E., 73-78 m. Description The largest specimen is one from east of Algoa Bay, and measures 8 cm. high and about 6 cm. across the base. This colony and the other almost as large, from the same station, are roughly conical with a wide base of attachment (fig. 13 A). All specimens are rather soft, with a smooth surface, and of a dull pink or buff-pink colour. No common cloacal openings could be seen, and no definite arrangement of the zooids, except in one colony from east of Algoa Bay, in which the zooids appeared to be arranged in parallel spiral lines round the colony. Hartmeyer (1912) was uncertain of the arrangement of zooids in his type material, but described an apparently linear alignment of marks on the surface of the colony, which may correspond with the appearance of the present colony. Spicules are generally absent from the most superficial layer of the common test, but form a narrow layer a little below the surface, about the level of the base of the thoraces. Below this there are scattered irregular masses of calcareous matter, probably representing large groups of spicules. The spicules (fig. 13 B) are mainly of a crude stellate shape with few and rather stout rays. Both the central ball and the rays of the spicules are made up of many needle-like rods. The spicules may attain 130u in diameter, and are therefore larger than the spicules of Hartmeyer’s type specimen. There are also smaller lozenge-shaped or rod-like spicules. The zooids (fig. 13 C) are contracted and in this state measure about 2 mm. in length. In structure they are very like those of the type specimen, but 142 ANNALS OF THE SOUTH AFRICAN MUSEUM have well-developed testes in the form of a rosette of follicles at the posterior end of the abdomen. The branchial sac has narrow parastigmatic bars across the rows of stigmata. Larva Larvae (fig. 13 D) are present in colonies from Cape Peninsula and Kowie. They measure 0-8-1-2 mm. in length from the end of the papillae to the base of the tail. The 3 vertically arranged papillae arise from within an elliptical band of pO 0.2 mm. Fic. 13. Cystodites roseolus Hartmeyer. A, colony; B, spicules; C, zooid; D, larva. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 143 about 20 ampullae. In advanced larvae the rudiments of both oral and atrial siphons are present. The sensory vesicle contains a cup-shaped ocellus which is red in these alcohol-preserved specimens, and a spherical black otolith. Remarks This species is now known in South African waters from the Cape Peninsula to Algoa Bay. It has also been described from the tropical waters of the Gulf of Guinea and the Seychelles. The distribution thus given is a surprisingly wide one, and the species appears to exist under the very different temperature conditions of Cape Province and the Gulf of Guinea. In cases like this I wonder if we are dealing with two species which are so similar structurally that they cannot be distinguished, at least until many more specimens are available. Cystodites dellechiajet (Della Valle) Distoma dellachiajiae Della Valle, 1877, p. 40. Known distribution Very widely distributed in warm waters throughout the world, but apparently not known from the West Indian Ocean. Locality _Morrumbene Estuary, Mozambique: U.C.T., MOR 43 W. Description There are several colonies, each investing a specimen of the crab Pseudo- dromia integrifrons. The spicules form close-fitting capsules round the abdomens of the zooids, but not round the thoraces. The convex disc-shaped overlapping spicules of these capsules are about 200u in diameter and quite characteristic of the species. Genus POLYCITORELLA Michaelsen, 1924 Polycitorella pallida sp. n. (Fig. 14) Diagnosis of species Colony cushion-like, broadly based, and dull white. Zooids closely spaced and visible on the surface of the colony as small oval areas each with two siphonal openings. Stellate and irregular spicules present in all layers of the test. Zooids divided into thorax and abdomen. No lateral abdominal organs visible. Siphons each with 6 lobes. At least 40 oral tentacles. Dorsal tubercle simple. Thirteen to 15 rows of stigmata. Oesophagus long; stomach smooth- walled; anus 2-lipped. Holotype In the South African Museum. S.A.M. A25610 (PF 739). 144 ANNALS OF THE SOUTH AFRICAN MUSEUM Locality Algoa Bay: PF 7309. Description The single colony is about 12 cm. long, 6-5 cm. wide, and 3°5 cm. thick. It was evidently attached by most of the lower surface. In the preserved state the colour is dirty white. The whole upper surface is marked by many closely spaced low oval swellings, each marking the position of a zooid, and each having the openings of the oral and atrial siphon. There is no encrusting material on the colony. Fic. 14. Polycitorella pallida sp. n. A-C, spicules; D, zooid; E, dorsal tubercle; ht., heart. Spicules (fig. 14, A, B, C) are numerous throughout all parts of the test. Most of them are stellate with 5 to 10 rays in optical section. The rays are sometimes conical and fairly smooth, but usually can be seen to consist of many needle-like rods which sometimes produce an uneven or toothed end to the rays (fig. 14 C). Groups of rods or needles are also present, and form irregular spicules. In diameter the spicules range from 7p to 6op. The zooids (fig. 14D) are divided into a thorax 2-5 mm., and an abdomen about 10 mm. long. The oral siphon is FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS I45 short and 6-lobed, and the atrial siphon, which originates a short distance from the anterior end of the thorax, longer and more slender. Both siphons have powerful circular muscles. Narrow closely-spaced longitudinal muscles cover most of the thoracic region and, grouped together as a single narrower band, continue down each side of the abdomen. Circular muscles are not conspicuous on the thorax or abdomen. The zooids generally show two constrictions, one between the thorax and abdomen and the other about a quarter of the length of the abdomen behind this. The latter constriction corresponds in position with a specially dense accumulation of spicules and a region of tougher test forming a ring round the abdomen at that point. This is the part of the abdomen where Michaelsen (1924) found a pair of lateral abdominal organs in P. mariae, but I can find no similar structures in the present species, although it is likely that this is a region specially active in the formation of spicules. There are at least 40 simple oral tentacles of different sizes. The tentacles seem to be arranged in a single circle, but this is difficult to see. The dorsal tubercle (fig. 14 E) is small with a narrow oval slit placed transversely. There are 13 to 15 rows of stigmata in the branchial sac, each row with at least 12 rather wide and rectangular stigmata. The transverse bars are tall and narrow. The dorsal languets are triangular and quite large. The oesophagus is narrow and long, generally extending almost half the length of the abdomen. In outline the stomach is ovate or somewhat rectangular. Its walls are smooth. The remaining subdivisions of the gut are obscure, but behind the stomach there is a clearly defined section which is probably the post-stomach, and another one or possibly two indistinct chambers representing the mid-gut. The rectum is of uniform diameter and ends in a 2-lipped anus about the level of the 8th row of stigmata. None of the zooids examined showed any trace of gonads. The heart is a U-shaped tube beside the lower loop of the gut. Remarks This genus was established by Michaelsen (1924) for a species P. mariae from the extreme northern end of North Island, New Zealand, and differs from the genus Polycitor in having spicules in the test and lateral abdominal organs on the zooids. Apart from the New Zealand species the genus probably contains only one known species, described by Savigny (1816) as Eucoelium hospitiolum, from the Gulf of Suez. Savigny’s species differs from P. pallida in having few rows of stigmata, a very short abdomen, and spicules confined to the surface of the colony. P. mariae is much more like the South African species, but has a different form of colony and has zooids with lateral abdominal organs. ‘The wide geographical separation, with no intervening records, is also a reason for separating the species. 146 ANNALS OF THE SOUTH AFRICAN MUSEUM Genus TETRAZONA Michaelsen, 1930 Tetrazona porrecta sp. n. (Fig. 15) Diagnosis of species Colony variable but usually long, narrow and lobed. Zooids not in systems. Abdomen about 3 times as long as thorax. Both siphons short and 6-lobed. About 16 oral tentacles. About 30 stigmata in each of the 4 rows. Oesophagus long; stomach rectangular in outline. Testes below gut loop. Holotype In the South African Museum. 8.A.M. A25611 (PF 1095 B). Locality Algoa Bay: PF 1095 B. A | | 4 cm. mm. Fic. 15. Tetrazona porrecia sp. n. A, colony; B, zooid. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 147 Description The colonies vary in size and shape a good deal, but are usually long and relatively narrow with irregular lobes and swellings. As the colony generally seems to have grown on the stem of an alga these swellings give some of the specimens the appearance of a string of beads (fig. 15 A). Specimens of this kind reach a length of at least 24 cm. The colony is firm, and the surface rough owing to the presence of a uniform coating of sand grains. Fewer sand and shell particles are present within the test. The zooids (fig. 15 B) have an average length of about 4 mm., of which the thorax occupies 1 mm. and the abdomen 3 mm. Both siphons open directly to the surface, and both are short, with 6 sharp lobes. Muscles are not strong on the thoracic wall and the zooids are consequently little contracted. The oral tentacles, about 16 in number, are of moderate length, alternating in size, and arranged in a single circle. Each of the 4 rows of stigmata has about 30 rather narrow slits. The abdomen is only slightly narrower than the thorax. The oesophagus extends about half the length of the abdomen, and is quite wide, except where it narrows before entering the stomach. In lateral outline the stomach is rect- angular and in transverse section there is a suggestion of about 6 indistinct folds, but these, or some of them, may be artefacts. The post-stomach is narrower than the stomach, and the mid-gut, which forms the lower, horizontal part of the gut loop, is wider. The rectum is of nearly uniform diameter. Of the gonads, only the testis was present in the zooids examined, and this consisted of a quite massive group of pear-shaped follicles lying immediately below the lower bend of the gut loop, with the sperm duct passing straight forwards to the atrial cavity. Remarks I am following Huus (1937) in accepting Michaelsen’s genus Tetrazona for a small group of species differing from Cystodites essentially in lacking spicules. The type species was described by Sluiter (1906) as Distoma glareosus, and the genus evidently also includes Polycitor vitreus (Sars) and probably Polycztor magalhaensis (Michaelsen). Removal of this well-defined group of species with four rows of stigmata leaves the genus Polycitor with those species having a much larger, but variable, number of rows of stigmata, with the type P. crystallinus Renier. Genus sycozoa Lesson, 1830 Sycozoa arborescens Hartmeyer (Fig. 16) Sycozoa arborescens Hartmeyer, 1912, p. 316. Known distribution Off Cape Town, South Africa. 148 ANNALS OF THE SOUTH AFRICAN MUSEUM Localities False Bay: U.C.T., FAL 234 S. Mossel Bay: U.C.T., MB 48 B; U.C.T., MB 65 F. Algoa Bay: PF 586 A. Description The colonies which come closest to Hartmeyer’s (1912) description of the type specimen are two from Mossel Bay (MB 48 B). These colonies measure respectively 2-0 and 2-2 cm. in greatest diameter (fig. 16 A, B). They consist of a short basal column from the top of which spring several heads, each incom- pletely divided into two. The heads are fan-shaped and laterally flattened. In r52 99 9OO oe & -°9¢¢g00 00 e O55. emr: Fic. 16. Sycozoa arborescens Hartmeyer. A-C, colonies. colour the colonies are pale grey with the zooids showing as darker marks. The zooids are arranged, as in Hartmeyer’s specimen, in vertical double rows. They agree with the description of the type material, and like it, lack gonads. The three specimens from False Bay do not have the characteristically branched colony. Two of them are single stalked heads (fig. 16 C), and one is a pair of headless stalks joined basally. One of the stalked heads has the flattened fan-shape which is found in the heads of more typical specimens of the species and has the same double linear arrangement of zooids. Some of the zooids of this colony have brood pouches containing a number of developing eggs, but no larvae. Other specimens, from Mossel Bay (MB 65 F) and Algoa Bay, are less typical, but I am also including them in this species. They are single stalked heads resembling the small specimen shown by Hartmeyer (1912) in his text- figure 5. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 149 Remarks The form of branching, which was more clearly developed in the type specimen than in any of the present specimens, readily distinguishes this species from others of the genus. It is the only Sycozoa known from South African waters. Genus DISTAPLIA Della Valle, 1881 Distaplia capensis Michaelsen (Fig. 17) Distaplia capensis Michaelsen, 1934, p. 141. Known distribution False Bay and Still Bay, Cape Province. Localities Eemcenoy We eal., FAL 136M; U.C.T., FAL 158 N; U:C.T, FAL 175 NEU Cris BAL 182 W. ? Mossel Bay: U.C.T., MB 27 I. 0.5 mm. 0.3 mm. Fic. 17. Distaplia capensis Michaelsen. A, colony; B, zooid; C, larva. 150 ANNALS OF THE SOUTH AFRICAN MUSEUM Description The best-developed and most typical colonies are these from FAL 175 N (fig. 17 A). In these the basal part forms a short squat column and the head, which alone contains the zooids, is slightly wider than the column. The basal column, which Michaelsen (1934) designates the stalk, is white and tough, and the head has clear soft gelatinous test but is red-brown owing to the pigmentation of the thoraces. In the collector’s notes the colonies are described as ‘purplish, sloppy’, or ‘off-white, gelatinous’. I can add little to the detailed description of the zooids, as the present material confirms in every respect the structure of the type specimens. In all zooids the folding of the stomach was distinct. In those sexually mature, only the testis was developed; this consisted of a rosette of about 10 follicles and was situated beside the intestinal loop. The position of the gonad, entirely within the abdomen, is an important point of distinction between the present species and D. skoogi (see figs. 17 B and 18 B). There is one specimen from MB 27 I which I include, with some doubt, in D. capensis. This is a single colony of soft consistency and pink-brown colour, with small zooids having nothing to identify them, as the gonads are not developed. The thorax is orange in colour, in the preserved state. Many of the zooids have a single larva, and this seems to differ in some respects from that of D. skoogi, in which species the specimen otherwise might be placed. The length, from the end of the papillae to the base of the tail, is 0-8 to 0-9 mm. Compared with the larva of D. skoogi the trunk is deep. The arrangement of the papillae and their basal ampullae is similar. On each side of the postero- ventral part of the trunk there is a curved flat structure which partly conceals the oesophagus and rectum, only the stomach and part of the intestine being clearly visible. No structures of this kind were seen in the larva of D. skoogi. As far as could be determined these structures appear to be the thickened sides of the abdomen. It is therefore likely that this colony is not of D. skoogi and may be of D. capensis. Distaplia skoogi Michaelsen (Fig. 18) Distaplia domuncula Michaelsen, 1923, p. 15. Distaflia skoogi Michaelsen, 1934, p. 131. Known distribution Walker Bay and False Bay, Cape Province. Localities False Bay: U.C.T., FAL 232 E. Off Cape St. Blaize: PF 1864. Algoa Bay: U.C.T., LIZ 34 H. ? Morrumbene Estuary: U.C.T., MOR 43 X. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS I51 Description Some of the colonies of this species in the present collection invested the carapace of crabs, but others were attached to the stems of algae, and it appears that the association with a crab, although common, is not obligatory. The smallest specimen is one of 1:3 cm. diameter, attached to an algal stem, and the largest specimen is 5:2 cm. in length, and is part only of a colony. The colonies lcm o) Uw 3 3 C) 4, toaooooveve 90000890000" < P) eanngange? Vg ggagoll? ie: 0.5mm. Fic. 18. Distaplia skoogi Michaelsen. A, colony, B, abdomen of zooid; C, larva; D, ocellus and otolith of larva; l.c., lens cells; oc., ocellus; ot., otolith. are of fairly firm texture and all have clearly visible small round or oval systems of zooids, generally closely crowded together. The common cloacal openings are fairly conspicuous. The zooids have already been described (Michaelsen, 1923; Millar, 1955). There is rather more variation in the shape of the atrial opening and languet 152 ANNALS OF THE SOUTH AFRICAN MUSEUM than the previous accounts suggest. In some zooids of some colonies the atrial opening is relatively small and is surmounted by a long languet which has 3 distal lobes; in other colonies the zooids have a wide opening with a short, broad, undivided languet. The most characteristic feature of the zooid is the position of the gonad, which is accommodated in a small pouch projecting from the end of the abdomen (fig. 18 B). This pouch often projects even farther than is shown in Michaelsen’s figure. The colonies from FAL 232 E and LIZ 34 H had fully developed larvae. Not more than one larva was seen in each breeding zooid. The larva (fig. 18 C) measures 1:0 to 1:1 mm. from the end of the papillae to the base of the tail. There are 3 large papillae, two dorsal and one ventral, the ventral papilla lying to the right of the sagittal plane. Each papilla has two round ampullae at its base. The branchial sac and gut are well developed. Both ocellus and otolith are present in the sensory vesicle, but they are so close together that under low magnification it might be thought that only one pigmented organ is present. The ocellus has 3 lens cells, apparently in linear arrangement (fig. 18 D). Remarks Michaelsen (1923) described a new species from South Africa and named it Distaplia domuncula, and I used the same name for a later record (Muillar, 1955). The name, however, was pre-occupied by a species which Hartmeyer (1913) had described, and Michaelsen (1934) gave his own species the new name Distaplia skoogt. Michaelsen (1934) believed Dvstaplia skooge Michaelsen and Dhistaplia domuncula (Hartmeyer) to be separate. A comparison of the three species recorded from Cape Province is given in Table 1. TABLE I Colony Kooids Gonads Stomach D. capensis Slightly Unisexual. Beside gut loop; 14-16 folds. stalked. 10-14 male fol- licles. D. skoogi Not stalked; Hermaphrodite. In sac projecting Smooth externally; usually in- from abdomen; with lines _inter- vesting crabs. about 5 male nally. follicles. D. domuncula Not stalked; Probably uni- Incompletely Smooth externally; investing sexual. known. with lines inter- crabs. nally. D. capensis is evidently distinct from D. skoogi and D. domuncula, but the latter two species are similar and might eventually have to be united. Genus SIGILLINA Savigny, 1816 I am using the genus Szgi/lina in the wide sense employed by Michaelsen (1930) and Hastings (1931) for the following single species, which does not agree completely with any of the sub-genera which Michaelsen defined. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 153 Szgillina vasta sp. n. (Fig. 19) Diagnosis of species Colony massive, with a system of cloacal canals below the surface, and apparently a few common cloacal openings. Zooids with long oral and atrial Fic. 19. Sigillina vasta sp. n. A, zooid with testis; B, abdomen of zooid with ovary. mm. mm. 154 ANNALS OF THE SOUTH AFRICAN MUSEUM siphons, the latter at the posterior end of the thorax opening into a cloacal canal, and usually backwardly directed. Three rows of stigmata. Abdomen about the same length as thorax, and with a long narrow posterior vascular process. Gonad beside intestinal loop, with the ovary sometimes at the extreme posterior end of the abdomen. Holotype In the South African Museum. $.A.M. A25612 (U.C.T., TRA 60 E). Locality Kaffirkuils Bay: U.C.T., TRA 60 E. Description The specimen is part of a colony brought up in the catch of a commercial trawler, and the whole colony was, according to the note accompanying the specimen, a ‘large gelatinous mass approximately 3 by 8 by 10 inches (2-3 lbs.)’. In the preserved state the specimen is purple-brown, with a very smooth surface. There is no encrusting material. A few irregular slits were found on the surface, and these are taken to be the external openings of the common cloacal canals which lie a little below the surface. The oral openings are apparently scattered without order and the zooids are therefore apparently not arranged in regular systems. The zooids (fig. 19 A) are quite large, and may reach 7 mm. in length, with the thorax and abdomen of about equal length. The oral siphon is remarkably long, and may equal or even exceed the thorax in length. Its opening has 6 lobes. The atrial siphon arises from the posterior end of the thorax and is about the same length as the oral siphon. It is generally directed obliquely back, instead of towards the surface of the colony, and leads into one of the common cloacal canals. There are 6 lobes on the opening of the atrial siphon. At the base of the oral siphon there are about 4 large tentacles and many short ones occupying a zone anterior to the large ones. The dorsal tubercle is small and its opening could not be clearly seen. Three rows of long narrow stigmata are present with at least 12 stigmata per row. The oesophagus is quite long, the stomach almost globular, and the post-stomach and intestine much as in species of Eudistoma. Some zooids have a group of about 1o testis follicles arranged in a rosette beside the gut-loop (fig. 19 A), with the sperm duct passing straight forward. In other zooids the ovary is developed, consisting of a small group of oocytes, or sometimes a single large egg projecting from the posterior end of the abdomen (fig. 19, B). There is a posterior vascular process from the end of the abdomen of very variable length. Remarks This remarkable species is difficult to place in its right genus. It would fit Eudistoma were it not for the presence of common cloacal canals into which the atrial siphons open. Another species, originally described by Sluiter (1895) under the name Distoma deerratum and later by Hastings (1931) as Szgzllina FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 155 deerrata also has a long oral siphon, and a posterior atrial siphon opening into a common cloacal canal. This species, however, which Hastings identifies with Polycitor coalitus Sluiter, is known only from the tropical waters of Malaya and north and north-eastern Australia, and in spite of structural similarities it is very unlikely that the South African specimen is of that species. There may be a case for recognizing a new subgenus of Szgzllina to accom- modate the two species S. deerrata and S. vasta which differ from Eudistoma in having the atrial siphon opening to a cloacal system. Genus EUDISTOMA Caullery, 1909 Eudistoma digitatum sp. n. (Fig. 20) Diagnosis of species Fully grown colony massive and firm with solid base, and apex divided into a number of finger-like lobes which do not spread, but have a tendency to grow inwards. Test fibrous, with large bladder cells which are few and widely spaced. Zooids with a short wide thorax, a longer but narrower abdomen, and a long posterior process of variable width. Gonads beside intestinal loop. Holotype In the British Museum (Nat. Hist.). Bowerbank, ‘Cape of Good Hope’. Localities Mossel Bay: PF 11 A; U.C.T., MB 48 C. Algoa Bay: U.C.T., LIZ 32 P; B.M. (N.H.) collection, reg. no. 1852.3.12. 81, collected by J. Bowerbank; B.M. (N.H.) collection, reg. no. 1852.3.12.89, collected by J. Bowerbank. ‘Cape of Good Hope’: B.M. (N.H.) collection, no reg. no., collected by J. Bowerbank. Description The colony, when large, has a very characteristic appearance, best shown by the specimens collected by Bowerbank from ‘Cape of Good Hope’ (fig. 20 A), and Algoa Bay. The 3 largest specimens have the following dimensions: height (cm.) 140 7:6 10:0 width (cm.) 12°5 6-3 5:0 From a solid base there arises an upright portion which is usually stout and columnar, but may be very short. The apex of this portion is divided into a number, up to 12, of characteristic finger-like or wide tongue-like lobes. These lobes are grouped close together and instead of spreading outwards have a tendency to grow inwards at the top, towards the vertical axis of the colony. The whole colony is very massive and firm, and the surface smooth without 156 ANNALS OF THE SOUTH AFRICAN MUSEUM any encrusting matter. There is no sand within the colony and the common test is fairly homogeneous, except for fibres and a few scattered bladder cells. In well-preserved colonies the zooids are sometimes visible externally; they are mainly confined to the lobes. The zooids (fig. 20 B) consist of a short wide thorax often about 1 mm. long, and a narrower abdomen about 2 mm. long. A long vascular process A B 3 cm. mm. Fic. 20. Eudistoma digitatum sp. n. A, colony; B, zooid. projects from the end of the abdomen and extends, often for many mm., downwards towards the base of the colony. Generally this process is very slender, but in some colonies it is almost as wide as the abdomen and contains large quantities of opaque tissue, probably consisting of reserve food cells. The siphons are both short, with 6 indistinct lobes, and open close together on the surface of the colony. The thorax is usually widest at its anterior end. About 16 oral tentacles were counted, but others were present spread over a zone at FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 157 the base of the siphon. There are about 16 stigmata in each of the 3 rows. The length of the oesophagus varies, but is usually greater than that of the thorax. It is tapered at the lower end where it enters the round or wide smooth- walled stomach. A narrow post-stomach leads to the curved intestine. The rectum is of uniform diameter. In mature zooids there is a group of many small testis follicles situated beside the lower part of the gut loop and sometimes projecting slightly posterior to it. ‘The ovary is represented by one large oocyte, no doubt with other smaller ones not visible. Remarks This species, which should not have been mistaken for others, does not seem to have been described, although several species of Eudistoma are known from South African waters. It is remarkable for the large and massive appearance of the colonies. Eudistoma coeruleum (Sluiter) (Fig. 21) Distoma coeruleum Sluiter, 1898, p. 14. Known distribution Natal; ?Still Bay, Cape Province; Mozambique. Localities Natal: Bluff, M.S. Evans, British Museum (N.H.) reg. no. 1876.9.3.4; Scottburgh, S.A. Museum, low water level. Fic. 21. Eudistoma coeruleum (Sluiter). A, larva; B, C, sensory vesicle of different larvae to show variation in position of ocellus and otolith. 158 ANNALS OF THE SOUTH AFRICAN MUSEUM Description This species is easily recognized by the club-shaped colonies of dark slate- blue or purple colour. Zooids are confined to the expanded head of the colony, and in the present material measured from 1-5 to 2-5 mm. in length; Sluiter (1898) gives 4-5 mm. as the maximum length of zooids. In some of the new material and especially that from Scottburgh, there are moderately powerful longitudinal and transverse muscles on the thorax and the longitudinal ones pass down towards the ventral side of the abdomen. In the branchial sac at least 10 stigmata were counted in each row. The dorsal languets are long and slender. Larva Colonies from both stations had larvae, only one to three in the atrial cavity of each breeding zooid. The larva (fig. 21 A) measures 0-5 to 0-6 mm. from the end of the papillae to the base of the tail. The trunk is rather long and narrow, with the sensory vesicle about the middle of the dorsal side. Both a black otolith and a paler ocellus are present, but their relative position varies (fig. 21 B, C). There are three anterior papillae. Eudistoma modestum (Sluiter) (Fig. 22) Distoma modestum Sluiter, 1898, p. 18. Known distribution Durban, Natal. Localities Algoa Bay: U.C.T., LIZ 1 Z; U.G.T., LIZ 9 M; U.CY., LiZ sai PF 895. Description The colonies vary a good deal in appearance, but most are rather soft and gelatinous in texture, and flattened, the flattening perhaps being an artefact. The specimens do not have a definite stalk, as in Sluiter’s type material. As Sluiter’s specimens did not exceed 1 cm. in diameter, however, they may not have reached the full size of the species; it is possible that further growth would have brought about a change in shape. The colour is brown or semi-transparent with brown zooids showing through. The zooids (fig. 22) are 2-4 mm. in length, and have a shorter waist than is common in the genus. Although Sluiter described the waist as long and thin, I have found, on examining the type specimens, that in many zooids it is only slightly longer than the thorax. The longitudinal thoracic muscles in the present material are slender and number up to 18. In some colonies the siphons and the anterior end of the thorax are chocolate coloured, and brown spots also FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 159 mark the rest of the thorax. Zooids of the type material, which I have examined, have the same parts marked with pink. The difference in colour may be due to a difference in the method of preservation. In most other respects the structure of the zooids is like that of the type specimens, but the oral tentacles number at least 30, whereas Sluiter recorded only 16. | mm. Fic. 22. Eudistoma modestum (Sluiter). Zooid. Remarks This is one of those cases in which the new specimens agree fairly closely with type material but show such differences as might be expected to result from their having been collected at a different age, or stage of maturity. 160 ANNALS OF THE SOUTH AFRICAN MUSEUM Eudistoma reniert (Hartmeyer) (Fig. 23) Polycitor (Distoma) reniert Hartmeyer 1912, p. 309. Known distribution St. Francis Bay, Cape Province. Localities Noxth) of (Durban 2045974, 5..038 lo ake. 60 m., ‘Gilchrist Collection’, British Museum (N.H.) reg. no. 1926.8.1.5. Description The single example of this species is a colony of roughly cylindrical form, about 7-5 cm. long and 3:0 cm. in diameter. It is very firm in con- sistency, partly owing to the sand which encrusts and impregnates the test. There is a small area round each zooid free of sand, and the sur- face of the colony consequently has a mottled appearance. The zooids (fig. 23) are about 5 mm. long, and are similar in proportions to those described by Hartmeyer (1912). According to Hartmeyer the lower part of the gut has a twisted loop characteristic of the species. This feature was also noticeable in the material from the British Museum (N.H.). A narrow process of variable length arises from the lower end of the abdomen, but Hartmeyer did not find a similar process in his material. ] mm. Remarks The present record extends the known range Fic. 23. Eudistoma renieri of this species into the warmer south-eastern (Hartmeyer). waters off the coast of Natal. Zooid. Eudistoma illotum (Sluiter) (Fig. 24) Distoma illotum Sluiter, 1898, p. 16. Polycitor (Distoma) illotus (Sluiter). Hartmeyer, 1912, p. 303. Known distribution Cape ‘Town and Simon’s Bay, Cape Province. 3 i | | FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 161 Localities False Bay: FAL 175 L. Mossel Bay: ?MB7Q; ?MB 22 F; ?MB 27 J; ?>MB 48 D. Description Colonies are of rounded, lobed, upright or even slightly stalked, shape. No sand is present either on the surface or within the colony. The test is of a milky or semi-transparent appearance. Large vesicular cells are present in the test and these were noted by Sluiter (1898) as a distinguishing feature of the species. a 0.3 mm. Fic. 24. Eudistoma illotum (Sluiter). Larva. In the colonies from False Bay the zooids are 3-4 mm. in length, and slender with a long waist as in the type specimens. ‘The zooids in the colonies collected from Mossel Bay range from 2 to 6 mm. in length and have only a short waist. I am therefore a little doubtful of the identification of the material from Mossel Bay. Larva No larvae were found in the present material, but Sluiter’s type specimens, which I have been able to examine, have larvae (fig. 24). These range from 0-96 to 1:20 mm. in length, from the end of the papillae to the base of the tail. There are 3 anterior papillae with a pair of narrow lateral ampullae between each two adjacent papillae, on each side. Both ocellus and otolith are present. Only one larva was found in each breeding zooid, perhaps owing to the large size of the larvae. Reliable characters in the genus Eudistoma are few, there is considerable 162 ANNALS OF THE SOUTH AFRICAN MUSEUM similarity in the zooids of different species, and the colony is often variable. For these reasons the identification of a number of specimens is doubtful and I include the following two species with hesitation. ?Eudistoma angolanum Michaelsen Polycitor (Eudistoma) paesslerioides var. angolana Michaelsen, 1914, p. 430. Known distribution Gulf of Guinea (W. Africa); NW. and W. Australia. Localities False Bay: FAL 158 Q; FAL 208 T. Mossel Bay : MB 7 Es MB GoIR: Algoa Bay: LIZ 34 F. Description The colonies are usually pillow-shaped, rounded, or slightly lobed. They are coated and impregnated with sand and are consequently hard. The zooids seldom exceed 6 mm. in length, of which the thorax generally occupies 1°0-1°5 mm. Muscles are well developed, consisting principally of about 10 longitudinal strands on each side of the thorax, and a powerful ring of muscles on each siphon. The waist of the zooids is long, containing the correspondingly long oesophagus and rectum. A short vascular process some- times extends below the abdomen. Gonads, present in specimens collected in January 1958, consist of a group of 11-20 or more testis follicles beside the lower part of the intestinal loop, and a small ovary. Remarks E. angolanum has not been recorded from closer than tropical west Africa, but I can find no important differences between that species and the present specimens. The identification remains doubtful, however, owing to the diffi- culty of separating species of Ewdistoma, and to the wide geographical separation of previous records. ?Eudistoma mobiusi (Hartmeyer) Colella mobiusii Hartmeyer, 1905, p. 396. Polycitor mobiust Hartmeyer, 1912, p. 305. Known distribution Cape Province and Madagascar. Localities False Bay: FAL 208 R. Algoa Bay: LIZ 19 Z 1. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 163 Description The single colony from False Bay and the two from Algoa Bay are all about the same size, 1:-2—-1-4.cm. long. There is a rounded or flattened head, and a narrower but slightly longer stalk. Both head and stalk are firm and pale grey to dull yellow in colour. Zooids are visible on the head. The zooids usually attain a length of up to 7 mm. The oral tentacles are very small and arranged in several circles. There are at least 15 stigmata in each of the three rows. The waist of the zooid is rather short, so that the oesophagus and rectum are not so long as in some species of Eudistoma. This is a feature, however, which might change at different periods in the life of the colony. No gonads were present. Family Didemnidae Giard, 1871 Genus DIDEMNUM Savigny, 1816 The genus Didemnum has given systematists great trouble and even the most experienced workers have hesitated in deciding specific distinctions. Thus Van Name (1945, p. 80) wrote in his account of the American species “The treatment of the species that is possible with present information is necessarily somewhat provisional and will doubtless require future revision, but the evidence seems to be that instead of Didemnum being one of the largest genera of ascidians, the species are in reality rather few, though in some cases widely distributed geographically.’ And again in treating Didemnum candidum Savigny he states (p. 86) that ‘This is the most difficult of all the American ascidians to deal with from a taxonomic point of view. I am far from being able to overcome the fear that I am confusing more than one species, but after the examination of a large amount of material from various American localities I am at a loss to find a reliable basis for dividing it by studying museum specimens.’ The situation is similar with regard to the South African species. Michaelsen (1934, p. 147), in his description of D. stilense from Still Bay, Cape Province, had to admit that ‘Because of our uncertainty as to the limits of variability and because of the occurrence of so many species insufficiently described, it is hardly possible to say whether D. stzlensis is indeed a new species, or whether it should be identified with any older species.’ I believe that to try to identify preserved specimens, usually few in number, collected at different seasons and states of their annual cycle, and generally without any biological data, would only add further confusion. A proper taxonomic study of the genus should involve regular collection of specimens throughout the year, in large numbers, together with observation of biological characters such as breeding season. For these reasons I have made a very tentative identification of a number of specimens of the genus Didemnum, as D. stilense. 164 ANNALS OF THE SOUTH AFRICAN MUSEUM ?Didemnum stilense Michaelsen Didemnum stilense Michaelsen 1934, p. 146. Millar, 1955, p. 176. Known distribution Cape Province. Localities Langebaan Lagoon: U.C.T., LB 367 V. False Bay: U.C.T., FAL 121 B; U.C.T., FAL 136 N; U.C.T., FAL 152 M; U.G.T., FAL 158 X; U.C.T., FAL 163 P; U.C.T., FAL a aan. U.C.T., FAL 176 R; U.C.T., FAL 182 U; U.C.0:) BAISorGeeaS U.G.T.,. FAL 239 E; U.GT., FAL 272 F; U.C.1,, FAL Roy ae U.C.T., FB 1106 B. Mossel Bay: U.C.T., MB 18 R. Algoa Bay: U.C.T:, LIZ 9 K; U.C.T., LIZ 28 S; UC. Re izreae PF 586 B. Description Colonies are generally thin and encrusting but otherwise are of very variable appearance. In the preserved state the colour is white, grey, buff, pink or blue, and the collector’s notes show that living material is also of various colours, but that these may change considerably on preservation. Spicules are usually many, closely crowded, and present throughout the whole test. They are generally of a regular stellate form with 6-10 conical rays as seen in optical section. Sometimes the rays are more numerous, or small and rounded, or with blunt ends. The larger spicules are generally 20—30p in diameter and occasionally as large as 40 or even 70p. The thorax of the zooids is rather smaller than the abdomen, and possesses a slender retractor process which may be almost as long as the abdomen. The atrial opening varies in size and depends at least partly on the degree of con- traction of the thorax. The gut is divided as usual in the genus. A single large egg generally represents the ovary when this is well developed. The testis has a single large follicle with the proximal part of the sperm duct forming 8-10 coils on its surface. Larva Larvae were present in colonies from the following collections: False Bay, 17-2-53, 20-8-47, 18-9-53. Mossel Bay, 13-1-51. Algoa Bay, 7—4-54. The larva has the general structure commonly found in the family. There are 3 anterior adhesive papillae, 4 lateral ampullae on each side, and a sensory vesicle with ocellus and otolith. Larvae varied considerably in size, the smallest having a length of 0.4 mm. (False Bay, 18-9-53) and the largest a length of ie > FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 165 0:95 mm. (Algoa Bay, 7—-4-54). This is a large range within one species, but Carlisle (19544, 19540) has shown that didemnid larvae vary in size according to the size of the zooids, which itself is variable. Remarks It must be emphasized that the identification of the above specimens as D. stilense is tentative. Subgenus POLYSYNCRATON Nott 1891 Didemnum (Polysyncraton) magnilarvum sp. n. (Fig. 25) Diagnosis of species Colony flat but rather fleshy. Spicules few, stellate, up to 75 in diameter. Common cloacal openings numerous. Zooids with wide atrial opening, short atrial languet, slender retractor process, testis of 8 to 10 follicles, and sperm duct of 2 to 3 coils. Larva with trunk about 1-3 mm. long, having a collar of about 30 narrow ampullae surrounding the 3 papillae. Budding is precocious, advanced larvae each having 2 buds. Holotype In the South African Museum. S.A.M. A25613 (PF 12393 B). Locality Natal: PF 12393 B. Description The single colony is 6 cm. by 3:7 cm. in diameter and 0-7 cm. thick. It is flat but of a rather fleshy appearance. In the preserved state the colour is greyish-pink. The upper surface is smooth without any encrusting matter, and the whole colony is soft and flexible. Certain areas of the colony are paler owing to the presence of greater numbers of spicules, which, however, are scarce in most places. The common cloacal openings are numerous and scattered irregularly over the surface. Spicules are nowhere numerous in the common test, but are most plentiful near the upper and lower surfaces and scarcest in the central layers. ‘They are stellate and the largest ones reach 75u in diameter, but most are smaller. The zooids (fig. 25 A, B) do not show any obvious arrangement in systems. They are generally about 1 mm. long, the thorax being only slightly longer than the abdomen. The oral siphon is tubular, with 6 rather shallow lobes. Theatrial opening is wide and exposes quite a large part of the branchial sac. A short atrial languet is present; it has a narrow base and slightly wider free end which is sometimes indented, giving the languet two shallow lobes. No trace was found of lateral thoracic organs, and there were no accumulations of spicules beside 166 ANNALS OF THE SOUTH AFRICAN MUSEUM the thorax such as generally mark the position of these organs in species which possess them. There are 4 rows of stigmata, and a long slender retractor process from the base of the thorax. The oesophagus is of moderate length, leading to the ovate or almost globular stomach. There is a short post-stomach, and simple intestinal loop. The testis consists of a rosette of 8 to 10 follicles, and the sperm duct makes about 24 spiral turns before passing forward towards the rectum. Larva Numerous larvae (fig. 25 CG) were embedded in the common test. When fully developed these have a trunk about 1-3 mm. long, measured from the end of the papillae to the base of the tail. The 3 papillae project from the centre of a 0.5 mm. eed 0.5 mm. Fic. 25. Didemnum (Polysyncraton) magnilarvum sp. n. A, thorax of zooid; B, abdomen of zooid; C, larva. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 167 collar of about 30 narrow finger-like ampullae, formed by the subdivision of the paired ampullae generally present in didemnid larvae. The sensory vesicle has both ocellus and otolith. In addition to the branchial sac and gut of the oozoid there is, on each side of the larval trunk, an equally well developed branchial sac and gut representing the first two blastozooids. A similar precocious development was noted by Lahille (1890) in Dizdemnum (Polysyncraton) lacazi. In the Didem- nidae this is characteristic of Diplosoma but I do not know of other references to the phenomenon in Polysyncraton. ‘The larvae of only a few species are known, however, and I believe that the precocious appearance of blastozooids, together with the subdivided anterior collar, may be found characteristic of larvae in the sub-genus Polysyncraton. Remarks Few species of the sub-genus are known from the waters of South Africa or the western Indian Ocean. D. (Polysyncraton) spongiotdes (Hartmeyer) and D. (Polysyncraton) chuni (Hartmeyer), both from 318 m. off Cape of Good Hope, appear to be very closely related to each other, if distinct, but they are clearly distinguished from D. (Polysyncraton) magnilarvum by their small number (3 or 4) of testis follicles. D. (Polysyncraton) tubtporae (Michaelsen) from Zanzibar is distinguished by its characteristic lateral thoracic organs with adjacent masses of spicules. D. (Polysyncraton) paradoxum Nott var. mahenum Michaelsen is described from the Seychelles and the species originally from New Zealand. The variety seems to be rather like the present species, but has much smaller spicules and usually only 4 testis follicles. D. (Polysyncraton) amethysteum (Van Name) is known from tropical Atlantic America, tropical west Africa and the eastern Mediterranean; it is therefore a species of widespread occurrence in warm waters and might be expected to occur in the Indian Ocean. It resembles the new species but has smaller spicules, fewer testis follicles and a larva only half as large. Genus LEPTOCLINIDES Bjerkan, 1905 Leptoclinides capensis Michaelsen Leftoclinides capensis Michaelsen 1934, p. 151. Known distribution Preekstoel, near Still Bay, Cape Province. Localities able Bay WG eb 171 As Balsebayen) Cube EAL dig: Description The specimen from False Bay is a colony 1 cm. in diameter, which had been growing on the carapace of a small dromiid crab, according to the 168 ANNALS OF THE SOUTH AFRICAN MUSEUM collector’s note, but the specimen from Table Bay is incomplete and the sub- stratum unknown. In the colony from False Bay there is a single central common cloacal opening; Michaelsen (1934) also apparently found only one opening in his type specimen. I have found the structure of the common test to agree in general with Michaelsen’s description, except in the absence of bladder cells from the whole superficial layer, the type specimen having a narrow subcortical layer without bladder cells. Spicules are few, being concentrated almost entirely near the opening of the oral siphon and at the lateral thoracic organs. The zooids agree well with those of the type specimen, but have rather larger lateral thoracic organs. Six spiral turns of the sperm duct lie over the single testis follicle. The colony from Table Bay, collected on 8-4-1946, had developing larvae in the common test, but none fully formed. Genus LissocLinum Verrill, 1871 Lissoclinum cavum sp. n. (Fig. 26) Diagnosis of species Colony consisting of an upper and a lower layer of common test, separated by a large common cloacal space across which the zooids are stretched. Spicules up to 40u in diameter. Zooids with a very large atrial opening, and two groups of spicules over the postero-ventral corners of the thorax. Testis with about 5 follicles. Sperm duct not coiled. Holotype In the South African Museum. S.A.M. A 25614 (U.C.T., LIZ 32 V). Localities False Bay: U.G.T., FAL 234 Ut. Algoa Bay: U.C.T., LIZ 32 V. Description The specimens from both stations are thin irregular sheets, of a dirty yellow or brown colour. Small darker spots on the surface mark the oral openings. The quite numerous round or oval common cloacal openings are scattered over the surface. The colony consists of an upper thin layer of common test and a lower rather thicker layer. Between these layers is the very large common cloacal cavity. Spicules (fig. 26 A) are numerous and closely packed in the upper and lower layers of the colony, and are also present in the test which covers the abdomen. They are stellate but somewhat irregular with a tendency for the rays to be arranged in pairs, or to be reduced to round swellings. ‘The maximum diameter found was 40u. In addition there is an ovoid mass of spicules at the postero-ventral corners of the thorax, presumably marking the position of the FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 169 lateral thoracic organs. The zooids (fig. 26 B), each enclosed in a thin sheath of test material, are suspended between the two layers of test. Most zooids are between 1 and 2 mm. in length. The thorax is generally slightly longer than the abdomen. The oral siphon is tubular, of variable length, and ends in 6 indistinct lobes. Much of the dorsal surface and sides of the thorax are exposed Sy 30 u 0.5 mm. | a ee | 0.2 mm. Fic. 26. Lissoclinum cavum sp. n. A, spicules; B, zooid; C, larva. by the very large atrial opening, the anterior border of which has a short projecting lip or languet. There are about 24 long slender oral tentacles which meet across the base of the oral siphon. Each of the four rows has about 14 stigmata. The abdomen lies almost at right angles to the thorax, thus displacing the gut from the attitude normal in didemnids. The stomach, which is ovoid or cylindrical, is followed by a short post-stomach, and an intestine and rectum of almost uniform diameter. No ovary was present in any of the zooids examined. The testis consists of a rosette of 5 or 6 wedge-shaped follicles, and the sperm duct, originating from the centre of the rosette, does not make any spiral turns, but is gently curved near its point of origin. Larva The specimen from Algoa Bay, taken on 6-4-1954, contained larvae (fig. 26 C). The trunk of the larva is about 0-6 mm. long from the end of the papillae 170 ANNALS OF THE SOUTH AFRICAN MUSEUM to the base of the tail. The papillae project from the centre of a group of irregular ampullae. Both ocellus and otolith are present in the sensory vesicle. Remarks Lissoclinum bilobatum Millar is perhaps the South African species most like L. cavum, but is distinguished by having only 2 testis follicles. Genus TRIDIDEMNUM Della Valle, 1881 Trididemnum cerebriforme Hartmeyer Trididemnum cerebriforme Hartmeyer, 1913, p. 139. Known distribution Cape Province. Localities Saldanha Bay: U.C.T., SB 130 U. Langebaan Lagoon: U.C.T., LB 352 U; U.C.T., LB 374 F; U-CT EB Bye on OP Cal MA Heys Sms es UN Col bee Fis aig (Ge False Bay: U.C.T., FAL 136 8; U.C.T., FAL 144 Z; U.C-T,, FA or6us: U.G.T., FAL 225 P; U.C.T., FAL 234 U2; U.C.i., FA aoees Mortensen’s Java—South Africa Expedition 1929-30. Station 61, B.M. (N.H.) reg. no. 34.2.1.3; Mortensen’s Java—South Africa Expedition, 1929-30. Station 65, 19.12.29, B.M. (N.H.) reg. no. 34.2.1.5. Mossel Bays U.C.1T., MB 59 K; U:C/E. MBS. S: Algoa Bay: U.G.T., LIZ 9 L; U.C.T.; LIZ 28 T; U.C i) Eiaieeae UW. Gol EYZ oak Commercial trawlers: U.C.T., TRA 71 F. Description The largest of the three specimens on which Hartmeyer (1913) based this species had a much-folded surface but in the smallest the surface was almost plain. The brain-like folding cannot therefore be regarded as a diagnostic character. Among the numerous specimens in the South African collections that I have examined there are many colonies with a smooth surface and only a few which are much convoluted. One of the latter kind, taken by Mortensen’s expedition, is folded over so that the anatomically lower surfaces are almost in contact, leaving an extensive but narrow space into which open slits which I take to be the common cloacal openings. ‘These openings are therefore on the under surface of the colony, a condition thought to exist (Michaelsen, 1930) in Leptoclinides brasiliensis, but otherwise very rare in the Didemnidae. In all other respects this South African specimen agrees with more typical examples of T. cerebriforme. The zooids of various colonies show much variation in the FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS El amount of pigment on the thorax. In some there is no pigment, in others a small dark spot over the anterior end of the endostyle, and in many specimens the thorax has a great quantity of dark blue pigment. Genus DipLosomMA Macdonald, 1859 Diplosoma listerianum Milne Edwards Leptoclinum listerianum Milne Edwards, 1841, p. 84. Diplosoma listerianum Michaelsen, 1919, p. 42. Known distribution West coast of Europe; Mediterranean; Adriatic; Sargasso Sea; west Africa; Luderitz Bay, South West Africa to Durban, Natal. Localities Saldanha Bay: U.C.T., SB 92. Malse Bay: U.C.0., FAL 109 Y; U.C.T., FAL 272 F. misoa bays, U.C. 1. LIZ 1 Y. Remarks I can find no constant differences between these specimens and colonies of D. listerianum from British waters. There is some doubt, however, regarding the validity of certain species of Diplosoma, and it may prove that D. modestum described by Michaelsen (1920) from the Seychelles and Zanzibar is identical to D. listerianum. The Atlantic American D. macdonaldi Herdman also appears to be very similar anatomically, but Van Name (1945) was unwilling to identify it with D. listerianum without having examined European material. This is a genus in which anatomical features of colony and zooid are very uniform and specific differences may have to be sought in larval or biological characters. Family Cionidae Lahille, 1887 Genus cCIONA Fleming, 1822 Ciona intestinalis (Linnaeus) Ascidia intestinalis Linnaeus, 1767, vol. 1, pp. 2, 1087. Ciona intestinalis Fleming, 1822, p. 512. Known distribution West coast of Europe, Mediterranean, tropical west Africa, east and west coasts of north America, Malay Archipelago, Japan, Australia, South Africa. Localities Saldanha Bay: U.C.T., SB 94; U.C.T., SB 134 A. Description The South African specimens are typical of the species as found in European waters. I was able to confirm the distribution of the longitudinal muscles to the siphons, as previously described (Millar, 1953). 172 ANNALS OF THE SOUTH AFRICAN MUSEUM Family Ascidiidae Adams, 1858 Genus ascip1a Linnaeus, 1767 Ascidia stenodes sp. n. (Fig. 27) Diagnosis of species Anterior part of body narrow and elongated, with terminal oral siphon. Atrial siphon posterior to middle of body. Body wall with a sharply defined band of transverse muscles along the dorsal side. Gut small, mainly in the posterior part of the body, and forming a single narrow loop with the anus near the oesophagus. Holotype In the South African Museum. S.A.M. A25615 (U.C.T., FAL 259 Q). Localities False Bay: U.C.T., FAL 5 J; U.C.T., FAL 259 QO; U.G. 9.) FB vex B Fic. 27. Ascidia stenodes sp. n. A, specimen with test removed, seen from the left; B, dorsal tubercle; C, region of atrial opening, an., anus; atr.op., atrial opening; ovd., oviduct. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 173 Description There are 6 specimens in the collections, the largest being about 5:5 cm. long, and the smallest 1-0 cm. long. The surface of all specimens is thickly coated with small stones, pieces of shell and sand which obscure the shape of the body. The test, when freed from adhering particles, is transparent and not very thick. When the body is removed from the test it is seen to consist of a long tapering anterior part terminating in a short narrow oral siphon, and a wider posterior part which accommodates the gut (fig. 27 A). The short incon- spicuous atrial siphon is near the posterior end of the body and in young specimens very near it. Most of the body wall of the left side is thin and trans- parent, without muscles, but a thick narrow band of transverse muscles runs along the dorsal side and extends only a short distance on to the left and right sides; the margins of this band are sharply defined. The left side of large specimens also has a ventral band of transverse muscles, but this is absent in small specimens. The oral tentacles number about 30 in a specimen of body length 1-3 cm., about 40 in one of 1-7 cm. and 60 in one of 5:5 cm. In two small specimens examined, the opening of the dorsal tubercle is a slightly curved longitudinal slit with the concavity to the left (fig. 27 B). In large specimens it is an irregular C-shaped slit. The small ovoid ganglion is immediately posterior to the tubercle. The branchial sac extends back to the posterior end of the body. The dorsal lamina is a moderately wide smooth-edged membrane. The longitudinal bars are of uniform thickness and number about 40 in the specimen 1-7 cm. long. There are 2—7 stigmata in each mesh according to the size of the specimen. Rather small papillae are present at the intersections of transverse and longi- tudinal bars, but no intermediate papillae. The gut is rather small. It starts with the short curved oesophagus lying across the body and leading to the ovoid stomach which passes obliquely forward. A few longitudinal folds are present on the wall of the stomach. The intestine makes a simple loop and the rectum ends near the base of the atrial siphon, in a small two-lipped or simple anus. The gonad, as usual in the genus, is spread over a large part of the intestinal wall, and the ducts lie beside the intestine and rectum (fig. 27 C). Remarks The diagnostic features of this species have already been noted. To them might be added the covering of shell fragments or stones, which was found on all specimens, but this is a character which may depend on the nature of the substratum. According to the collector’s note the specimens FAL 259 Q were lying unattached on the bottom and this may be the usual condition in the species. Ascidia sydneiensis Stimpson Ascidia sydneiensis Stimpson, 1855, p. 387. For synonymy see Kott 1952, pp. 310, 311. LA. ANNALS OF THE SOUTH AFRICAN MUSEUM Known distribution West Indies; Hawaii; Malay Archipelago; Amboina; Japan; Australia; South Africa. Localities False Bay: U.C.T., FAL 158 V; U.C.T., FAL 175 Q; U.C.T., FAL 181 A; from catch of a trawler, S.A. Museum, April 1904. Mossel Bay: U.C.T., MB 18S. East London: S.A. Museum no. 12943, low tide. Algoa Bay: U.Q.T., LIZ 32 S. S.A. Museum (no details). Description These specimens agree closely with those which I have already described from South African waters (Millar, 1955). The largest of the present animals is 13 cm. long. Remarks Except for the South African records A. sydneiensis is known only from warm waters where, however, it is very widely distributed throughout the world. I am not sure that A. sydneiensis, A. incrassata Heller, and possibly A. multi- tentaculata Hartmeyer, all recorded from South African waters, might not prove to represent a single species. Family Corellidae Lahille, 1887 Genus CORELLA Alder & Hancock, 1870 Corella eumyota ‘Traustedt Corella eumyota Traustedt, 1882, pp. 271, 273. For synonymy see Kott, 1952, pp. 318, 319. Known distribution Antarctic; Subantarctic; South Africa; St. Paul (Indian Ocean); New Zealand; Australia. Localities Langebaan Lagoon: U.C.T., LB 367 U; U.C.T., LB 374 E; U.C.T., LB 375 X. Family Agnesiidae Huntsman, 1912 Genus AGNESIA Michaelsen, 1898 Agnesia glaciata Michaelsen Agnesia glaciata Michaelsen, 1808, p. 370. Agnesia krausei Michaelsen, 1912, p. 181, figs. 24, 25. Agnesia capensis Millar, 1955, p. 191, fig. 19. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 175 Remarks No specimens of A. glaciata are present in this collection, but material which I previously examined from South Africa contained one specimen of a species which I described as a new species A. capensis (Millar, 1955). During examination of the Discovery collections I found another specimen, also from False Bay, the same locality as that of the type specimen of A. capensis. The Discovery specimen was intermediate between A. capensis and A. glaciata, and I have given reasons (Millar, 1960) for regarding A. capensis as a synonym of A, glaciata. Family Styelidae Sluiter, 1895 Subfamily Botryllinae Adams, 1858 Genus BOTRYLLUs Gaertner, 1774 Botryllus magnicoecus (Hartmeyer) Botrylloides nigrum var. magnicoecum Hartmeyer, 1912, p. 271. Botryllus magnicoecus (Hartmeyer). Michaelsen and Hartmeyer, 1928, p. 331. Known distribution South Africa; Australia; New Zealand. Localities False Bay: U.C.T., FAL 109 Z; U.C.T., FAL 158 U; U.C.T., FAL 175 GC; U.G.T., FAL 225 M; PF 15797 B; PF 14582. Table Bay: Woodstock Beach, S.A. Museum, 19-6—-14, washed up on beach, collected by K. H. Barnard. Saldanha Bay: U.C.T., SB 128 P. Algoa Bay: U.G.T., LIZ 9 H. Botryllus anomalus sp. n. (Fig. 28) Diagnosis of species Atrial openings of the zooids lead directly to the surface of the colony, instead of into common cloacal chambers. Zooids with 8 rows of stigmata, 5 or 6 folds on the stomach, and a short straight pyloric caecum. A gonad on each side of the body, each gonad consisting of a group of 6 or 7 testis follicles and a single large ovum. Holotype In the South African Museum. S.A.M. A25616 (Saldanha Bay, 5-9-12). Localities Saldanha Bay: S.A. Museum, level of low water, 5—9-12, collected by K. H. Barnard. Table Bay, Woodstock Beach: S.A. Museum, washed up on beach, 19—6-14. collected K. H. Barnard. % - i a — 176 ANNALS OF THE SOUTH AFRICAN MUSEUM Description There are several colonies of this species in the collection. A large specimen is 9 cm. long, 3 cm. wide, and less than 1 mm. thick. All specimens have been | growing on thin flat algal fronds. ‘The surface of the colony is smooth and even, ‘ without any foreign matter, and the colour of the specimens, in the preserved state, a dull pink. There are no common cloacal cavities in the test and the Fic. 28. Botryllus anomalus sp. n. A, surface view of part of colony to show arrange- ment of zooids, and vascular ampullae; B, zooid, seen from the dorsal side; C, transverse section of stomach. 0.5 mm. zooids are not arranged in systems, but are closely and regularly spaced over the colony (fig. 28 A). Each zooid (fig. 28 B) has a pair of small openings on the dorsal side, marking the oral and atrial siphons, both of which open directly on the surface of the colony. Rounded vascular ampullae are present between the zooids. The zooids are small, rarely exceeding 1 mm. in length. Near the anterior end of the branchial sac is the round oral opening, through which can be seen the 8 tentacles. The branchial sac has 3 inner longitudinal bars on each side, and 8 rows of stigmata along the length of the sac. The digestive system consists of a short oesophagus leading from the posterior end of the branchial sac to the short barrel-shaped stomach. This has 5 or 6 complete folds on the walls 4 FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 7/7) (fig. 28 C), and a short straight pyloric caecum with slightly swollen distal end. The intestine makes a sharp bend to the left and upwards to the short rectum which ends in the simple anus below the atrial opening. On each side of the body is a single gonad, each gonad consisting of a group of 6 or 7 pear-shaped testis follicles and one large dark brown ovum. The left gonad is slightly anterior to the right one. No larvae were present in the colonies. Remarks This species is chiefly remarkable for the opening of the individual atrial siphons of the zooids directly on the surface of the colony, and the corresponding absence of common cloacal chambers. The same features characterize Botryllus primigenus Oka which, however, is distinguished by the presence of only 4 rows of stigmata in the branchial sac, and the rather larger, curved pyloric caecum. Genus BOTRYLLOIDES Milne Edwards, 1841 ?Botrylloides leacht (Savigny) Botryllus leachii Savigny 1816, p. 199. Known distribution Western Europe; Mediterranean; Australia; New Zealand; South Africa. Localities : Saldanha Bay: U.C.T., SB 91. Remarks This species has already been recorded by Michaelsen (1934) from Table Bay and False Bay. No gonads are developed in the present specimen, and I am identifying it as B. leacht because it agrees with that species in other characters. It must be admitted, however, that the distinctions are not entirely satisfactory in a number of species in both Boérylloides and Botryllus. A single colony (U.C.T., SB 89, 13—7—46) from the same area is similar but has more circular systems and I am not sure if it is of this species, or possibly Botryllus schlosseri (Pallas). Botrylloides mgrum Herdman var. giganteum Pérés Metrocarpa nigrum var. giganteum Pérés, 1949, p. 205. Known distribution (Of variety) Senegal, west Africa; Durban, Natal and Knysna, Cape Province. Locality Morrumbene Estuary: U.C.T., MOR 122 G. 178 ANNALS OF THE SOUTH AFRICAN MUSEUM Description The colony is dark purple in the preserved state. The zooids are large, with a long narrow thorax containing about 16 rows of stigmata. There is a small pyloric caecum. The gonad consists of an anterior testis with about 12 follicles, and a posterior ovary containing one large ovum. Subfamily Polyzoinae Hartmeyer, 1903 Genus GYNANDROCARPA Michaelsen, 1900 Gynandrocarpa unilateralis (Michaelsen) (Fig. 29) Gynandrocarpa placenta (Herdman) var. unilateralis Michaelsen 1900, p. 29. Gynandrocarpa placenta (Herdman). Michaelsen 19048, p. 30. [non] Goodsiria placenta Herdman, 1886, p. 328. Gynandrocarpa domuncula Michaelsen, 19044, p. 247. Known distribution South Africa; south-east Africa. Localities False Bay: U.G. 0), PAL 265, 8; U: Cale WANG oa): Mossel Bay: U.C.T., MB 53 Q; U.C.T., MB 65 G; PF 1711; PF 10165. Algoa Bay: PF 1095 C; Port Elizabeth, B.M. (N.H.) registered no. 1871- 5-12-1; U.C.T., LIZ 28 N. Description The colonies are of two types: (1) a flattened upright disc with basal stalk attached to one edge of the disc (fig. 29 A), and (2) a hollow dome-shaped plate with no stalk (fig. 29 B). The stalked type of colony appears to have been attached, during life, to a firm object on the bottom of the sea, but the dome- shaped specimens were taken from the carapace of crabs. ‘These latter colonies were originally considered to be a separate species, G. domuncula Michaelsen, but Michaelsen (1934) later regarded them merely as colonies which are modified in shape by their attachment to the crab. Having examined several specimens of both types I entirely agree with this view, as I find no differences in the zooids or larvae of the two types. Preserved specimens are generally dull buff, grey, or pink-grey in colour, but in some colonies the zooids are quite bright pink, which may have been the colour of the colony in life. One specimen is described in the collector’s note as ‘strawberry red’. Michaelsen (1904a) has dealt thoroughly with the structure of the zooids of this species under the name G. domuncula. The most interesting feature is the gonad (fig. 29 C), which is single and on the right side. Many zooids of the new material had a well-developed gonad. This consists of a sac-like ovary flanked FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 179 on each side by a rounded or ovoid testis. The oviduct is short and has its distal part expanded to form a brood pouch which lies against the outer surface of the branchial wall. The brood pouch opens into the cavity of the branchial sac by means of an oval slit in the branchial wall, which at this place is devoid of stigmata. In the present material the brood pouch of breeding zooids never contained more than 2 or 3 embryos or larvae. One zooid contained a straight- tailed fully developed larva in the branchial sac, thus confirming that larvae pass from the brood pouch into the branchial sac before escaping to the exterior. Each testis has many small pear-shaped follicles, the ducts from which unite to form one duct from each follicle, and these join the single sperm duct. The emb. _—————E— Eee ]} mm. Fic. 29. Gynandrocarpa unilateralis (Michaelsen). A, stalked colony; B, stalkless colony; C, gonad; br.p., brood pouch; emb., embryo; 1.b., longitudinal branchial bars; ovy., ovary; sp.d., sperm duct; te., testis; D, larva. 180 ANNALS OF THE SOUTH AFRICAN MUSEUM sperm duct opens into the peribranchial cavity. Michaelsen has pointed out that, with the oviduct opening to the branchial sac and the sperm duct to the atrial cavity, there may be obligatory cross-fertilization between zooids. Larva The larva (fig. 29 D) is large, having a trunk 1-0 to 1:1 mm. long and a tail of 2-0 to 2:1 mm. Epidermal ampullae almost completely cover the anterior two-thirds of the trunk. There are 2 dorsal and one ventral papillae. The rudiments of the oral and atrial siphons are already present in the larva. A sensory vesicle is present but contains neither ocellus nor otolith. In shape and organization this larva is very similar to that of Dextrocarpa solitaris (Millar, 1955), but is more than twice the size. Larvae obtained from the stalked type of colony and from the stalkless (“(domuncula’) type are indistinguishable. Remarks There has been confusion between this species and another South African species of similar shape, Polyandrocarpa placenta (Herdman), as I have already pointed out (Millar, 1955). Both the form of the colony and the general structure of the zooids are alike in the two species, which are most clearly separated by the gonads. In P. placenta there are several polycarp-type gonads, a condition very different from that described above in G. unilateralis. It is remarkable, however, that the larvae of P. placenta have, according to Herdman (1886), an ‘arrow shaped’ trunk about 1-0 mm. long and a tail 2-5 mm. long. They there- fore agree well with the larvae of G. unilateralis, and also appear to lack both a pigmented ocellus and otolith. But the larvae of P. placenta were present in the peribranchial cavities and not, as in G. unilateralis, in the brood pouch or branchial sac. Genus DEXTROCARPA Millar, 1955 Dextrocarpa solitaris Millar (Fig. 30) Dextrocarpa solitaris Millar, 1955, p. 200. Known distribution False Bay, Cape Province. Localities PF 18785 B. Description Each specimen consists of a group of pear-shaped individuals joined by a narrow stalk to a basal stolon (fig. 30 A). They therefore have a social rather than a colonial or solitary organization. The whole test is heavily coated with sand and shell fragments. The stalk uniting the individuals with the stolon varies in length, being in some cases much shorter, and in others longer, than FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 181 the body. Within the stalk there is a short extension of the innermost layer of test which surrounds the body (fig. 30 B, t), but no extension of the body wall, and the individuals therefore do not appear to be in organic connection with each other. In the type specimens I found no connexion between individuals, which were therefore regarded as quite solitary. The present specimens show a 0.3 cm. Fic. 30. Dextrocarpa solitaris Millar. A, a group of individuals; B, specimen with test removed, seen from the left; t., test. condition intermediate between a fully social organization with zooids in organic connexion, and a quite independent organization. It may be that in this species budding is confined to relatively young individuals, and in older ones complete separation may occur, or the apparently social organization may result from the fusion of the test of neighbouring but separate individuals. I have little to add to my earlier account of the species. A large individual dissected had about 38 oral tentacles, 19 folds on the stomach, and the following arrangement of longitudinal bars in the branchial sac: dorsal line o (8) 2 (7) 3 (8) 3 endostyle. ee 182 ANNALS OF THE SOUTH AFRICAN MUSEUM f Breeding individuals had up to 110 eggs each, in various stages of development, contained in the brood pouch. The larvae are as previously described. i Genus ALLOEOCARPA Michaelsen 1900 i Alloeocarpa capensis Hartmeyer (Fig. 31) Alloeocarpa capensis Hartmeyer, 1912, p. 261. é Known distribution : St. Francis Bay, Cape Province. Localities Algoa Bay: PF 740; PF 2348. Fic. 31. Alloeocarpa capensis Hartmeyer. A, gut; B, gonads; C, larva. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 183 Description The collection contains 4 colonies, or pieces of colonies, from Algoa Bay, the largest of which is 10 cm. by 4:5 cm. in diameter. In external appearance these specimens are very like the type material illustrated by Hartmeyer (1912, plate 38, fig. 7). The colonies are attached to shells or algal stems. They are of a dirty grey-brown colour and the surface has low but conspicuous swellings over the zooids. The specimen collected off Lion’s Head has zooids which are more or less free, being united only to a basal sheet of test by their narrow posterior ends. The zooids, when extracted from the common test, are dull red. Hartmeyer’s (1912) description of the zooids leaves little to be added, but I have found the arrangement of the gonads to be different. According to Hartmeyer the male gonads are on the left half of the body and the female on the right. In all the mature zooids which I have examined, from colonies at both collecting stations, the testes and ovaries were on the left, the testes in the anterior part and the ovaries far back in the posterior part (fig. 31 B). Itis a little difficult to determine the exact position of the ovaries relative to the middle line of the body, but careful examination shows them to lie to the left of the membrane which unites the endostyle to the ventral body wall. I have seen no gonads male or female to the right of this membrane, in any of the many zooids dissected. The position of the ovaries is confirmed by the location of larvae, which are in the left peribranchial cavity. Larva Only a few zooids had larvae (fig. 31 C), and there were never more than two in a breeding zooid. The larva is large, the trunk being 0-9 to 1:0 mm. long, and the tail about 1-3 mm. long. The trunk is brick-red. There are 3 anterior papillae and, a short distance behind these, a complete ring of about 24 narrow ampullae round the trunk. When the larva has been dehydrated and cleared a single spherical black body is seen, in the sensory vesicle, but is quite hidden by the red pigment of the trunk in the untreated larva. This body is presumably a ‘photolith’, the compound sensory structure of larvae in several genera of the subfamily Polyzoinae. Genus METANDROCARPA Michaelsen, 1904a Metandrocarpa fascicularis sp. n. (Fig. 32) Diagnosis of species Zooids separate, with a narrow stalk joining them to the basal common test. About 30 oral tentacles in two circles. Dorsal tubercle with a straight oblique slit. Branchial sac with 7-9 longitudinal bars on each side. Stomach with about 16 folds and a hooked pyloric caecum. Ten to 20 testes on each side, on the posterior part of the body wall. Two to 5 ovaries on each side, on the anterior part of the body wall. 184 ANNALS OF THE SOUTH AFRICAN MUSEUM Holotype In the South African Museum. S.A.M. A25617 (PF 10477). Locality Near Mossel Bay: PF 10477. Description There are 4 colonies the largest of which is 3:5 cm. by 2:0 cm. in diameter and 1-0 cm. thick. The colonies are attached to pieces of sand-encrusted tubes 2mm. | Fic. 32. Metandrocarpa fascicularis sp. n. A, colony; B, zooid, with test removed, seen from the left; C, oral tentacles; D, dorsal tubercle; E, a testis; F, larva. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 185 probably the tubes of polychaete worms. The zooids of a colony are closely crowded together, giving the appearance of a bunch of grapes (fig. 32 A). Large zooids (fig. 32 B) reach a length of 4 or 5 mm. The body of the zooid is pear-shaped, with the lower narrow end produced into a short slender stalk which unites the zooid with the basal creeping stolon of the colony. In some zooids the body is almost sessile on the stolon and in others the stalk is well developed. The oral and atrial openings are small and inconspicuous, situated quite close together on the upper end of the body. In the preserved state the specimens are pale brown, and have no encrusting material, except on the stolon. The test is thin but semi-opaque. The body wall is dull red, but sufficiently thin and transparent to allow the internal organs to show through. Muscles are poorly developed. The oral and atrial siphons are very short and end in simple round openings. There are about 30 oral tentacles arranged in two circles, the outer one consisting of about 16 large tentacles and the inner one of smaller tentacles (fig. 32 C). The dorsal tubercle is small and has a narrow straight slit- like opening which is obliquely placed (fig. 32 D). In the branchial sac of large zooids there are 9 longitudinal bars on each side and smaller zooids have 7 or 8. ‘There are no folds. The dorsal lamina is a narrow plain-edged membrane. The oesophagus is narrow. The stomach is of a long barrel-shape, with about 16 longitudinal folds and a narrow, hooked pyloric caecum. The intestine forms a flat bend parallel to the stomach, and the rectum ends in a smooth-edged anus. The female gonads are confined to the anterior part of the body and the male gonads to the posterior part. On each side of the endostyle there are 2 to 5 large spherical ovaries of a dark brown colour. These seem to produce only one mature egg at a time, since ovaries are often seen with a single developing embryo. Embryonic development takes place within the ovary or oviduct instead of freely in the atrial cavity. The testes are 10 to 20 small pear-shaped glands on the left and right sides of the posterior part of the body, each testis with a very slender sperm duct (figs92) 1): Larva Zooids were never found with more than two larvae. The larva (fig. 32 F) is reddish brown and has a trunk 0-7 to 0-8 mm. long and a tail of about 2:0 mm. There are 3 anterior papillae arranged in a triangle, and a circle of about 24 narrow ampullae round the trunk. From each of these ampullae a line of dark pigment extends back along the trunk. The single black sensory spot (“photo- lith’) can be seen only after dehydration and clearing. Remarks No species of Metandrocarpa has yet been described from the waters of South Africa or the western Indian Ocean. 186 ANNALS OF THE SOUTH AFRICAN MUSEUM There are other specimens in the collection which I cannot definitely identify, but which may belong to this species (e.g. PF 2348). Genus POLYANDROCARPA Michaelsen, 19044 Polyandrocarpa anguinea (Sluiter) Styela (Polycarpa) anguinea Sluiter, 1808, p. 52. Known distribution Knysna and False Bay, Cape Province. Localities Mossel Bay: U.C.T., MB 18 T; U.C.T., MB 22 G; U.C.T., MB 53 R; U.G.T., MB 55 U; U.C.T., MB 60 S. Algoa Bay: U.C.T., LIZ 28 Q. Morrumbene Estuary: U.C.T., MOR 108 N; U.C.T., MOR 132 H. Remarks This species has hitherto been placed in the genus Polycarpa and was considered as a simple form which grows in close aggregations of individuals. A study of the more numerous specimens now available in the present collection makes it clear that the species is a colonial styelid, to be placed in the genus Polyandrocarpa. It is very similar to P. lapidosa (Herdman) from south-eastern Australia, but appears to be specifically distinct, and I shall compare the two species in another publication. Subfamily Styelinae Herdman, 1881 Genus POLYCARPA Heller, 1877 Polycarpa rubida (Sluiter) (Fig. 33) Styela (Polycarpa) rubida Sluiter, 1898, p. 53. Polycarpa rubida (Sluiter). Hartmeyer 1909, p. 1364. Michaelsen, 1918, p. 37. Known distribution Mozambique. Locality Morrumbene Estuary: U.C.T., MOR 132 1. Description f The single specimen is 2:3 cm. long and 1-4 cm. wide. It had been basally attached, and is upright with the oral opening terminal and the atrial opening about one-third of the body length from it. Both openings are almost sessile, the siphons being very short. The test is grey with a pink-brown tinge, and has irregular furrows and mounds, but is not further divided into small fields as in the type specimen described by Sluiter (1898). Small rose-coloured marks on the surface of the test may indicate the colour in life; Sluiter’s specimen was FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS I 87 evidently red and the preserving alcohol had extracted this colour. The test is moderately thick and is tough. Within, it is pearly white. The body wall is dull red-grey, the red colour being most marked round the oral and atrial openings. Circular muscles form a continuous sheet and below these is a layer of longi- tudinal muscles, the whole set of muscles rendering the body wall thick and opaque. There are 24 oral tentacles of alternating lengths. The dorsal tubercle is a fairly prominent pad with an angular U-shaped opening facing obliquely forward and to the left (Fig. 33A). Four well- developed folds are present on each side of the branchial sac, with the following arrange- ment of longitudinal bars: dorsal line 2 (15) 2 (10) 2 (10) 1 (9) 1 endostyle. The dorsal lamina is a plain membrane. Most of the gut is situated in the posterior part of the body (fig. 33 B). ‘The oesophagus is curved and the stomach barrel-shaped with about 20 indistinct longitudinal or slightly oblique folds and a hooked pyloric caecum. The intestine and rectum form an S-shaped bend and the anus has about 15 shallow lobes In Sluiter’s specimen the anus was some dis- tance behind the atrial opening, but in the Fic. 33. Polycarpa rubida (Sluiter). present specimen it lies close to it. iy dorsal cubercles By eur: The gonads are small and numerous, about 40 being present on the left side of the body. They are of the type usual in the genus and are scattered over the inner surface of the body wall. Between them are numerous transparent flattened endocarps. A few endocarps are also present within the intestinal loop. Remarks This new specimen differs from the type specimen principally in having fewer longitudinal bars between the branchial folds, a character which may be related to the much smaller size of the specimen. The opening of the dorsal tubercle is also different, but as Sluiter remarked, the S-shaped slit of the type specimen is easily derived from a horse-shoe-shaped slit. Genus CNEMIDOCARPA Huntsman, 1913 Cnemidocarpa asymmetra (Hartmeyer) Tethyum (Styela) asymmetron Hartmeyer, 1912, pp. 253-257. Styela asymmetra (Hartmeyer). Michaelsen, 1915, pp. 394-398. Cnemidocarpa asymmetra Hartmeyer, 1926, pp. 180-183. [non] Styela asymmetrica Sluiter, 1904, p. 87. 188 ANNALS OF THE SOUTH AFRICAN MUSEUM Known distribution Cape Town, Cape Province and Luderitz Bay, South West Africa. Locality Langebaan Lagoon: U.C.T., LB 378 Q. Description The larger of the 2 specimens in the collection is 5-2 cm. high and 3:0 cm. wide, and the smaller specimen 4:0 cm. high and 3:0 cm. wide. The body is erect with both siphons projecting from the upper end. The test is slightly wrinkled, rather thin but tough, and semi-transparent. In the larger specimen the opening of the tubercle is horse-shoe-shaped with both horns turned out- wards, and in the smaller specimen similar but with only the right horn turned outwards. The following is the arrangement of longitudinal bars on one side of the smaller specimen: dorsal lamina o (8) 3 (11) 3 (10) 2 (8) 2 endostyle. There are 10-14 stigmata per mesh. The gonads form the most characteristic feature of the species. In the present specimens the shape of the gonads does not differ much from Hartmeyer’s (1912) original description. It is remarkable, however, that in the new specimens the testis lobes occupy a central position, being largely embedded within the ovary, whereas Hartmeyer (1912) gives the opposite arrangement (‘die innere Partie wird vom Ovarium, die aussere von den Hoden gebildet’). ‘The gonoducts are arranged in pairs and are distributed over the inner surface of the gonads. The larger specimen from Langebaan has a total of 7 of these pairs. Cnemidocarpa psammophora sp. n. (Fig. 34) Diagnosis Surface coated with sand or shell. Test with short hair-like processes over most of the surface and one or a few long root-like basal processes. Many closely spaced flat oral tentacles. Dorsal tubercle with a C-shaped or simple longitudinal slit. Branchial sac with 4 rather low folds. Stomach with about 20 folds and a short, curved pyloric caecum. Five to 8 gonads on the left side and 10 or 11 on the right. Holotype In the South African Museum. S.A.M. A25618 (U.C.T., LIZ 25 L). Localities Alsoa Bay: U-C.0* LIZ 19:2 90 Ci ei Zios ale: Description All three specimens in the collection have an almost globular body (fig. 34 A, B). In two the diameter is 10 mm. and in the third the body is FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 189 15 mm. long and 12 mm. wide. These measurements were made over the coating of sand grains and broken shell which completely covers the body. Two simple openings fairly close together on the upper side mark the positions of the oral and atrial siphons which do not project from the outline of the body. Among the sand grains and broken shell are many short tag-like processes of the test, which are generally distributed over the surface. At the base of the 5mm. Fic. 34. Cnemidocarpa psammophora sp. n. A, B, two intact specimens; C, specimen with test removed, seen from the right; D, E, dorsal tubercles of two specimens; F, part of stomach and intestine, to show pyloric caecum; G, gonads; ovd., oviduct; sp.d., sperm duct. body a few of these tags have been greatly developed as stout rooting processes. Generally about 6 of these are present, and one of them is longer and thicker than the rest. In the largest individual the main process is 24 mm. long. When the test is removed the body is seen to be almost globular or a little elongated, with the siphons projecting as short conical tubes with shallow ter- minal lobes (fig. 34 C). The body wall is semi-opaque although the muscles 190 ANNALS OF THE SOUTH AFRICAN MUSEUM are rather poorly developed. On the inner surface of the body wall are quite numerous small endocarps. The base of the oral siphon is provided with many closely spaced slender and flattened tentacles. In two specimens these numbered at least 60 and 80 respectively. ‘The opening of the dorsal tubercle in the largest specimen is a simple narrow longitudinal slit (fig. 34 D), and in one of the smaller specimens dissected a C-shaped slit facing forward, with slightly inrolled horns (fig. 34 E). The dorsal lamina is a moderately wide membrane with smooth margin. There are 4 rather low folds on each side of the branchial sac with the following arrangement of longitudinal bars in the largest specimen: dorsal lamina 1 (18) 2 (9) 3 (20) 3 (16) 2 endostyle. The gut consists of the curved oesophagus, the barrel-shaped stomach with about 20 folds and a short pyloric caecum (fig. 34 F), and the S-shaped intestine and rectum. The border of the anus is cut into a large number of very small lobes. Gonads are present on each side of the body and number on the left from 5 to 8, and on the right 10 or 11. The gonads are tubular and nearly straight, of the kind that characterizes the genus (fig. 34 G). Most of the gonad is occupied by a long tubular ovary, and closely applied to each side of this is a series of small pear-shaped testis follicles. The individual sperm ducts join the common sperm duct which passes along the mesial surface of the gonad. Both the sperm duct and the oviduct project slightly from the end of the gonad. Remarks There is no west African or South African species with which C. psammophora is likely to be confused. The only species of the genus recorded from the western Indian Ocean is C. madagascariensis Hartmeyer, a species incompletely described as the only specimen was damaged. From what is known, however, it is evident that several characters distinguish it from C. psammophora. C. madagascariensis has a test divided into areas separated by furrows, and lacks the coating of sand and the tags and rooting processes of C. psammophora. Hartmeyer’s species also has many fewer oral tentacles and fewer gonads. Several species of Cnemidocarpa are adapted to life on sandy bottoms. C. psammophora is one of these and shows the adaptive features of this ecological group, namely, small size, spherical form, and rooting processes of the test. Genus sTYELA Fleming, 1822 Styela angularis (Stimpson) Cynthia angularis Stimpson, 1855, p. 387. Tethyum costatum Hartmeyer, 1911, p, 564. Known distribution South Africa. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS I9gi Localities False Bay: U.G.T., FAL 158 W; U.C.T., FAL 216 G; U.C.T., FAL 225 L; Weir BAL 220M: Mossel Bay: UW:C.T., MB 7 R; U.C.T:, MB 14 Y; U.C.T., MB 18 U; UC. MBr438 By U-C:l) MB se7S; UC. 1, MB 55 Ve U.Grr: MB 68 F; U.C.T., MB 72 U. Cape Morgan: PF 13393; PF 13432; PF 13481. Description The characteristic shape of this species renders it unmistakable (see Hart- meyer, 1911, Millar, 1955). In the present collections specimens range from a very small size up to 9°5 cm. long. Styela marquesana Michaelsen Styela marquesana Michaelsen, 1918, p. 27. Known distribution Lourengo Marques, Mozambique and Durban, Natal. Localities Morrumbene Estuary: U.C.T., MOR 132 K. Cape Natal: PF 10722. Description Michaelsen’s (1918) detailed description leaves little to be added. In the present material specimens have from 22 to 26 folds on the stomach, and the branchial bars vary between the following limits: dorsal lamina 4 (10) 1 (10) 1 (10) 2 (6) 3 endostyle, and dorsal lamina 10 (12) 13 (16) 15 (15) 10 (9) 3 endostyle. Styela radicata sp. n. (Fig. 35) Diagnosis of species Surface coated with sand or shell. Test with one or a few basal processes. Dorsal tubercle with a simple transverse slit. Stomach with about 15 folds, and a small pyloric caecum. One gonad on each side, with tubular ovary and lobed testis follicles close to the sides of the ovary. Holotype In the South African Museum. 8.A.M. A25619 (U.C.T., MOR 132 J). Locality Morrumbene Estuary: U.C.T., MOR 1932 J. Description The single specimen (fig. 35 A) is almost spherical, but slightly narrowed at the base, and measures 1-1 cm. in height and 1-0 cm. in width. It is com- 192 ANNALS OF THE SOUTH AFRICAN MUSEUM pletely covered with sand grains and fragments of shell. The oral and atrial openings are on the upper side and not very far apart. They scarcely project from the surface. On the lower side there is a root-like process of the test about 4 mm. long, and a few much shorter tags. The test is thin, but very firm perhaps owing to the closely adhering sand and shell. Fic. 35. Styela radicata sp. n. A, intact specimen; B, specimen with test removed, seen from the left; C, dorsal tubercle; D, anus; E, gonad. Muscles are not strong and consist of radial strands surrounding the two siphons, and a series of circular muscles mainly round the siphons but also spreading downwards over the upper half of the body (fig. 35 B). The body wall is thin and delicate. The internal siphons are very short, and are not lobed. There are 32 oral tentacles, of which about 16 are quite small. The dorsal tubercle is small, with a simple transverse slit (fig. 35 C). The branchial sac is delicate, with 4 well developed folds having the following arrangement of longitudinal bars: dorsal lamina 2 (16) 4 (7) 4 (11) 5 (7) 2 endostyle. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 193 The dorsal lamina is narrow, with a plain margin. Each mesh of the branchial wall contains an average of 4 stigmata, which are long, narrow and regular. The short curved oesophagus leads to the barrel-shaped stomach which has about 15 undivided longitudinal folds and a very short rounded caecum. The intestine and rectum make a simple S-shaped bend ending in a narrow anus with minutely lobed margin (fig. 35 D). One gonad is present on each side of the body. On the left side it is anterior to the intestinal loop and directed obliquely back towards the atrial opening. The right gonad occupies a corresponding position. Each gonad (fig. 35 E) consists of an almost straight tubular ovary ending in a short simple oviduct, and a series of 8 to 10 lobed testis follicles along the whole length of each side of the ovary. No endocarps are visible on the inner surface of the body wall. Remarks This species is quite unlike any other known from the waters of west, South or east Africa. Its closest resemblance is to S. schmitti van Name from the southern part of the east coast of South America. ‘The small size, the shape, the coating of foreign particles, and the basal process are similar in the two species, but §. radicata differs from S. schmitt2 in having only one gonad on each side and in the structure of the branchial sac. The external similarities probably indicate adaptive convergence rather than systematic affinity. Family Pyuridae Hartmeyer, 1908 Genus PpyuRA Molina, 1782 Pyura stolonifera (Heller) (Fig. 36) Cynthia stolonifera Heller, 1878, p. 92. Pyura stolonifera (Heller) Hartmeyer, 1911, p. 554. Known distribution South Africa;? east and south Australia. Localities Hascbay Cle Wh AlsrAg Ye Go. FAL 163 R; U.C:i., FAL 177.N; U.C.T., FAL 221 8; PF 15984 A. Mossel Bay: U.C.T., MB 18 V; U.C.T., MB 53 T; PF 11 B. AlgoaiBay, ULGil Lie wx UE. l., LIZ 2) B; PE rog5 D: East London: S.A. Museum (no details). Description P. stolonifera is a species which may have to be split, if a critical examina- tion is made of enough material from different places. Most of the specimens 194 ANNALS OF THE SOUTH AFRICAN MUSEUM listed above are undoubtedly of this species, and have the dorsal tubercle basically in the form of a C opening posteriorly, but with the horns spirally inrolled and contorted to an extent that increases with the size of the animal (fig. 36 A-D). This is the arrangement found by Hartmeyer (1911) both in the specimens taken by the German South-polar Expedition in Simon’s Bay, and in Heller’s type specimens which Hartmeyer re-examined. Specimens from A B | G ————EeEe 0.5 mm. Fic. 36. Pyura stolonifera (Heller). A-D, dorsal tubercles respectively of small (A, B), medium-sized (C) and large (D) specimens; E, spicules. Australia, however, which have been assigned to this species (Hartmeyer and Michaelsen, 1928) have a dorsal tubercle facing forward. I believe that we may be dealing with two species which are structurally similar except in some characters like the dorsal tubercle. It would be surprising if P. stolonifera occurred in Cape Province, South Africa, and on the south coast of Australia but not between these places. Kott (1952), in dealing with the Australian records of the species, admits that ‘the nomenclature of the group is still very confused’. Some, but not all, of the specimens from False Bay and Mossel Bay have : : FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 195 spicules (fig. 36 E) in various body tissues, but not in the test. The spicules are most abundant in one large specimen from Mossel Bay (MB 18 V), in which the walls of the endostyle, the branchial folds, the oral tentacles and the dorsal tubercle have many closely packed simple or branched spicules. Spicules of this kind have been recorded in certain tissues of Cynthiopsis valdiviae Michaelsen and spicules of a simple shape in C. herdmanni (v. Drasche), and both of these species are regarded as representing P. stolonifera (Heller) (Hartmeyer, 1911). Pyura capensis Hartmeyer (Fig. 37) Pyura capensis Hartmeyer, 1911, p. 561. Known distribution False Bay, Cape Province. Fic. 37. Pyura capensis Hartmeyer. A, intact specimen; B, oral tentacle; C, dorsal tubercles of two specimens; D, gut; E, one of the gonadial sacs. 196 ANNALS OF THE SOUTH AFRICAN MUSEUM Locality Mossel Bay: U.C.T., MB 18 W. Description The largest of the 3 specimens is 2-5 cm. long and 1:6 cm. wide, and the smallest is 1-4 cm. long and 1:3 cm. wide. In all 3 the body is upright and the siphons fairly close together at the upper end. One of the characteristic features of this species, which readily distinguishes it from P. stolonifera, is the subdivision of the test into hexagonal or irregular plates, each plate having a red or brown central mark (fig. 37 A). This feature was noted by Hartmeyer (1911) and is distinct in the new specimens. The test is thin but tough. The oral tentacles (fig. 37 B) number about 16 in the largest specimen. They have slender primary branches with few and small secondary branches. In the two specimens dissected the opening of the dorsal tubercle is a U-shaped slit with a few short straight branches (fig. 37 C). The tubercle illustrated by Hartmeyer (Taf. 57, fig. 8) is more complex but has the same essential features. The 7 branchial folds have the following arrangement of longitudinal bars: dorsal line 1 (10) 3 (10) 2 (15) 2 (14) 2 (16) 2 (12) 3 (9) 2 endostyle-. A long row of dorsal languets is present. The gut of the new specimens agrees closely with Hartmeyer’s description. The margin of the anus, which Hartmeyer appears not to have seen clearly, is smooth. In most species of Pyura each gonad consists of a compact double row of sacs with central common ducts, but in P. capensis the individual gonadial sacs are numerous and scattered so that the gonads have a more diffuse appearance than is usual in the genus. Remarks This is a distinctive species, and the present material, the first since the type specimens were described, confirms the account given by Hartmeyer. Genus microcosmus Heller, 1878 Microcosmus oligophyllus Heller (Fig. 38) Microcosmus oligophyllus Heller, 1878, p. 101. Known distribution Cape Province. Localities Mossel Bay: U.C.T., MB 22 H. Saldanha Bay: Hoedjies Bay, low water, S.A. Museum. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 197 Description This species has already been adequately described (Heller, 1878; Hart- meyer, 1912; Millar, 1955), and only a few points about the new specimens need be noted. There are many specimens, 17 having been col- lected at low water from Hoedjies Bay alone. The dorsal tubercle shows some variation in the shape of its opening. In the type specimen re-examined by Hartmeyer (1912) the opening was U-shaped with the right limb longer than the left and slightly inrolled. Fic. 38. Microcosmus Another specimen from the same collection as the type eae had the right limb more markedly inrolled and the oe left one only slightly so. In the new material this difference in the limbs is sometimes even more obvious, and in some specimens results in the two limbs having well-formed spirals in opposite directions, giving the opening a modified horizontal S-shape (fig. 38). Microcosmus exasperatus Heller Microcosmus exasferatus Heller, 1878, p. 99. For synonymy see Van Name, 1945, pp. 346, 347. Known distribution Formosa; China Sea; Malay Archipelago; West Indies; Venezuela; Colombia; Natal; Mozambique. Localities Natal: PF 11001. Mozambique: U.C.T., PEA 1 B; U.C.T., PEA 4 Z; U.C.T., MOR 43 M; U.C.T., MOR 108 O; U.C.T., MOR 109 R; U.C.T., MOR 122 H; U.C.T., MOR 188 M. ? Mossel Bay: U.C.T., MB 22 I. Remarks This is a well-known species in many warm areas throughout the world. It seems to be common in coastal waters of Mozambique. There is one doubtful record from Mossel Bay; this specimen is covered with sand, and has only 7 folds on each branchial wail, but in other respects it agrees well with M. exasperatus. Maicrocosmus pedunculatus Pérés (Fig. 39) Maicrocosmus pedunculaius Pérés, 1951, p. 1062. Known distribution Senegal and Sierra Leone, west Africa. 198 ANNALS OF THE SOUTH AFRICAN MUSEUM Localities Mossel Bay: U.C.T., MB 72 V. Morrumbene: Estuary: U.Q.T., MOR 91 R; U.C.T., MOR 132 L. Description A specimen of average size is 1-6 cm. long and 1-4 cm. across, slightly compressed laterally and triangular in outline with the lower, narrow end produced into one or several root-like processes (fig. 39 A). These processes are whitish and free of sand but the rest of the body is completely coated with sand grains. A narrow ridge which joins the siphons is continued down the sides of the body towards the base. The internal structure (fig. 39 B) is in complete agreement with the descriptions already published of this species (Pérés, 1951; Millar, 1956). Ds 8 CS a ah es i : 4 4s oy paren SAE pt) VN? 2. - ba TS ee it o . S x = ase PONS - co Lhd Fic. 39. Microcosmus pedunculatus Pérés. A, intact specimen; B, specimen with test removed, seen from the left. Remarks This species apparently presents a remarkable case of discontinuous distribution, since the only previous records are from tropical west Africa north of the equator. The identification of the South African specimens, however, can scarcely be doubted on anatomical grounds. This is either a case of structural convergence of two species, or true discontinuous distribution, or continuous distribution of M. pedunculatus masked by the accidental absence of specimens from collections in intermediate areas. Genus BOLTENIA Savigny, 1816 Boltenia africana sp. n. (Fig. 40) Diagnosis of species Body erect, not stalked, completely covered with sand and shell. About 12 compound oral tentacles. Dorsal tubercle with a straight, C-shaped or FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 199 S-shaped slit. Branchial sac with transverse stigmata, and with 6 folds on each side. Stomach with longitudinal folds, and a compact lobed mass of glandular tubules. Intestine and rectum of uniform diameter. Anus with a plain margin. One gonad on each side, the left one in the secondary intestinal loop. Holotype In the South African Museum. 8.A.M. A25620 (PF 15984 B). 2 cm. ¥ ory , . . Kies 2 . 5 . 6 i NS ote So, o 6 v9 ae Y Brae Bes Of bi we wg 6" 4) < ae We rad ey, a : > ; Gs ~~ S one E by QAY 6, S, saath CROSG ~N I Se SH tun ava “Wy © Cir ott LS <> NS | a = ee fie on Fic. 40. Boltenia africana sp. n. A, intact specimen; B, specimen with test removed, seen from the right; C, dorsal tubercles of three specimens; D, part of branchial wall; E, gonad. Localities False Bay: U.C.T., FAL 212 T; PF 15984 B. Mossel Bay: U.C.T., MB 55 Y. Algoa Bay: U.C.T., LIZ 38 G. Description A number of the specimens collected from False Bay and from Mossel Bay had been growing closely together and formed compact masses. Large indivi- 200 ANNALS OF THE SOUTH AFRICAN MUSEUM duals reach 4 cm. in length and 3:5 cm. in breadth. The shape is variable, being cylindrical, squat or irregular (fig. 40 A). The two siphons generally project from the upper end but are sometimes scarcely visible. Sand, broken shell and other debris completely cover the surface and no doubt render the animals very inconspicuous in nature. his may explain why the species has not previously been described. The test is only moderately thick, but is tough. When removed from the test the body is seen to be red-brown. The siphons are quite long, and conical. A set of strong muscles radiates from each siphon over the sides of the body, and circular muscles pass round the siphons. The body wall is rather thin, so that the gut and gonads are visible (fig. 40 B). There are 11-14 oral tentacles, of alternating sizes. ‘The largest of these are tri-pinnate. The dorsal tubercle varies a good deal, the opening being a simple, almost straight longitudinal slit, or C-shaped, or in the form of a horizontal S (fig. 40 CQ). On each side of the branchial sac there are 6 folds, of which the ventral 2 or 3 are smaller than the more dorsal ones. A long row of dorsal languets is present. In a specimen of moderate size, and in a large one, the longitudinal bars of the branchial sac had respectively the following arrange- ment: dorsallme 4 (8). 1 (14) 3 (15) 3 (2)) 4 (6) 5 G6) S¥endoseyic dorsal line 15 (15) 10 (33) 8 (26) 12 (23) 12 (20) 12 (5) 12 endostyle The stigmata, as usual in the genus, are transversely placed in longitudinal rows (fig. 40 D). A longitudinal bar passes across the centre of each row of stigmata. The oesophagus is narrow, curved, and rather short. The stomach is wide at the oesophageal end and tapers towards its junction with the intestine, where there is a conspicuous lobed mass of glandular tubules. Numerous longitudinal folds are present on the walls of the stomach. The intestine forms a flat, hori- zontal, almost closed loop and the rectum bends upwards and forwards towards the base of the atrial siphon, where it ends in the plain-edged anus. There is little variation in diameter along the whole length of the intestine and rectum. One gonad is present on each side of the body, that of the left side lying just above the dorsal limb of the intestine, that is in the secondary gut loop (fig. 40 B). On the right side the gonad is in a corresponding position. Each gonad is a long, rather bulky, and slightly curved body (fig. 40 E). The ovary occupies the mesial side and the testis is confined to the parietal side, next to the body wall. Indentations on the margins of the ovary give it a slightly lobed appearance. The testis follicles are round bodies which form a white mass visible from the outside, through the body wall. Both the sperm duct and the oviduct are at the posterior end of the gonads, from which they project only very slightly. Remarks No species of Boltenia, as characterized by the transverse arrangement of stigmata, has been known hitherto from west, south, or east Africa. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 201 Genus HALOCYNTHIA Verrill, 1879 Halocynthia spinosa Sluiter f. defectiva n. (Fig. 41) Species Halocynthia spinosa Sluiter, 1905, p. 16. Diagnosis of form As the typical form of the species, but having 2 gonads on the left side and none on the right. Holotype In the South African Museum. 8.A.M. A25621 (U.C.T., MB 55 W). Known distribution (Of species) Gulf of Aden; Red Sea. Localities alse bay; W.C.W., FAL 137 N. Mossel Bay: U.C.T., MB 55 W; U.C.T., MB 68 G. Description The largest specimen in the collection is 4:5 cm. long and 5:0 cm. across. The body is rounded with a narrow base, or roughly hemispherical and broadly based (fig. 41 A). On the upper side the two siphons project prominently. In colour the specimens vary from pale grey to brown and generally have a pink tinge or quite bright red suffusion, especially on and near the siphons. Spines are present on almost all parts of the test usually in groups of one long central spine and several short encircling spines (fig. 41 B). The spines are larger on the siphons than elsewhere, but not so large as those shown by Sluiter (1905) in his figure of.the species. The test is fairly thick and is tough. Stout longitudinal and circular muscles render the body wall thick and opaque. The oral tentacles number 12 to 14 in these specimens, and are tri-pinnate. In form the dorsal tubercle is constant, having a C-shaped slit facing forward and with the horns spirally inrolled (fig. 41 D). The dorsal languets form a long series of slender projections from the roof of the branchial sac. Nine or 10 tall narrow folds are present on each side of the branchial sac, with up to 26 longi- tudinal bars on the largest folds. The gut forms a wide loop of almost uniform diameter, except for the stomach which is short and wide with longitudinally pleated walls and a lobed mass of glandular tubules (fig. 41 E). The gonads constitute the main peculiarity of these specimens. There are two on the left side (fig. 41 C) and none on the right. The gonads are large, and, instead of being parallel are curved so that they meet and fuse at their dorsal and ventral ends, forming what appears to be one ring-shaped body with two pairs of gonoducts. 202 ANNALS OF THE SOUTH AFRICAN MUSEUM Remarks In the type material Sluiter (1905) described the gonads as a lobed mass on each side, but Michaelsen (1919) noted, as an abnormality, the complete absence of gonads on the right side. It is interesting that Van Name (1921) found a similar condition in a specimen of H. pyriformis (Rathke) from the Bahamas, although that species normally has 4—7 gonads on each side. There Fic. 41. Halocynthia spinosa Sluiter f. defectiva n. A, intact specimen; B, spines of test; C, specimen with test and right body wall removed, seen from the right; D, dorsal tubercle; E, gut. seems to be a tendency to eliminate the gonads of the right side in Halocynthia, This tendency has apparently become fixed in the population of H. spinosa inhabiting the waters of the Cape Province, and it is perhaps best to recognize the situation by describing that population as a forma of the species. Neverthe- less, if further collections from intervening waters of the western Indian Ocean make it clear that there is discontinuous distribution of the species, it may be necessary to recognize f. defectiva as a new species. 6A ee FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 203 Family Molgulidae Lacaze-Duthiers, 1877 Genus MOLGULA Forbes and Hanley, 1848 Molgula falsensis Millar Molgula falsensis Millar, 1955, p. 217. Known distribution False Bay, Cape Province. Localities alse Bay 0.C.0., FAL) 136 R; U.C.T., FAL 158 Q; U.C.T., FB 1105. Description A few specimens of this species are present in the collections. They all confirm the characters which I have described in the type material (Millar, 1955). In some of the new specimens, however, the oviduct although turned up at right angles to the gonad, is quite short and in others it is not bent up from the gonad. In the type specimens I could not see the sperm duct, but in the new specimens it is situated about half-way along the gonad and is a distinct, free, finger-like projection. The test of some of the new specimens is thicker and much firmer than in the original specimens and approaches in appearance and consistency the condition found in many species of Ascidia. Short test projections or hairs are, however, present. Molgula conchata Sluiter Molgula conchata, Sluiter, 1808, p. 58. Ctenicella conchata (Sluiter). Hartmeyer, 1913, p. 128. Known distribution Knysna, Cape Province. Localities Mossel Bay: U.C.T., MB 7S; U.C.T., MB 39 M; U.C.T., MB 58 M; U.C.T., MB 60 T; U.C.T., MB 72 W. Alison bays (O.C.m., LIZ 11 Zi. Description Many specimens are contained in the collection, including large ones over 4 cm. long. All of them show the characters which readily distinguish this from other South African species of Molgula, namely: the coating of sand, slight lateral compression of the body, horizontal S-shaped slit of the dorsal tubercle, _and the form of the gonad and its position relative to the intestinal loop. Remarks The present records extend the known range of this species westwards to Mossel Bay and eastwards to Algoa Bay. The geographical distribution never- 204 ANNALS OF THE SOUTH AFRICAN MUSEUM theless seems to be very restricted, and M. conchata is replaced by allied species in neighbouring areas. Molgula scutata Millar Molgula scutata Millar, 1955, p. 215. Known distribution Langebaan, Cape Province. Localities Saldanha Bay: U.C.T., SB 1. PAlgoa Bay: U.C.T., LIZ 28 U. Description The dorsal tubercle of this specimen has an oblique and almost straight slit; the previous examples had a C-shaped or an S-shaped slit. One specimen from Algoa Bay, which I have provisionally included in this species, is similar in all respects to typical specimens except that each gonad is bent downwards at its dorsal end. This may be an individual peculiarity. Remarks This, the second record of the species, scarcely extends its known range. Molgula cryptica sp. n. (Fig. 42) Diagnosis of species Surface covered with sand. About 18 oral tentacles. Dorsal tubercle with C-shaped opening usually facing to the right. Seven branchial folds, each with 3 longitudinal bars. Primary intestinal loop closed; secondary loop semi- circular. Left gonad in secondary loop. Each gonad consisting of a sausage- shaped ovary and a testis of radially disposed follicles surrounding the blind end of the ovary. Holotype In the South African Museum. 8.A.M. A25622 (U.C.T., FAL 158 R). Locality False Bay: U.C.T., FAL 158 R. Description The body is generally ovoid, with the short siphons close together at the upper end. Most of the surface is covered with a thin layer of adhering sand grains. The largest specimens in the collection are about 3 cm. long, including the siphons. The test is rather thin and, but for the coating of sand, would be transparent. | FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 205 When the test is removed the body is seen to be elliptical or almost round in outline, and the tubular siphons close together (fig. 42 A, B). Each of the 6 oral lobes and the 4 atrial lobes has a series of narrow projecting teeth forming a fringe round the openings. The body wall in the preserved state is pale brown but is sufficiently transparent to allow the gut and gonads to be seen. Muscles are mainly confined to the siphons, where circular and longitudinal ones exist, and to the dorsal half of the body which has one set of longitudinal muscles radiating from the base of each siphon. The oral tentacles generally number about 18 and are of alternating sizes. They are tri-pinnate, but are not particularly bulky (fig. 42 C). The dorsal tubercle is small, with a C-shaped or V-shaped opening. The C-shaped slit is more common and usually faces to the right (fig. 42 D). Ossie Fic. 42. Molgula cryptica sp. n. A, B, specimens with test removed, seen from the right and left respectively; C, oral tentacle; D, dorsal tubercle; E, branchial fold in transverse section to show the three longitudinal bars (in black); F, gonad. 206 ANNALS OF THE SOUTH AFRICAN MUSEUM Seven low folds are present on each branchial wall, with a constant arrange- ment of 3 longitudinal bars on the folds; this appears to be characteristic of the species. On the ventral face of each fold near the base, there is a single bar, and two bars are present at the summit of the fold (fig. 42 E). The dorsal lamina is quite long and wide with a plain margin which is sometimes inrolled. The stigmata are subdivided into oval or long narrow straight slits which, as they tend to be arranged in regular transverse rows, obscure the spiral pattern. The oesophagus is short, and it narrows towards its junction with the stomach. A large mass of tubular glandular tissue surrounds and hides the stomach. The intestine and rectum are narrow throughout their length. The primary loop is quite closed and the secondary loop forms a semicircle (fig. 42 B). No incisions are present on the margin of the anus. On the left side the gonad lies in the secondary intestinal loop, which it largely fills, and on the right side it is immediately dorsal to the renal sac (fig. 42 A, B). The ovary is sausage-shaped or somewhat pear-shaped, with a rather short narrow oviduct directed towards the atrial siphon. Round the ventral, blind end of the ovary the many small testis follicles are radially arranged to form a compact crescentic or rosette-like testis (fig. 42 F). The common sperm duct is wide and extends about half-way along the inner face of the ovary to end in a remarkably large opening. The opening of the sperm duct is, in fact, wider than that of the oviduct. The renal sac is a large curved body in the lower half of the right side. Larva Larvae are present in the atrial cavity of some specimens. The larvae have a trunk about 0-2 mm. long and a tail of about 0-7 mm. There is a single black sensory pigment spot. Remarks M. cryptica is most like M. pulchra, which has been recorded from sub- Antarctic South America and adjacent parts of the Antarctic. M. pulchra, however, has a large number of oral tentacles, and the sperm duct has a characteristic convoluted shape. It seems likely that these two species have diverged from a common ancestral stock. I have brought together, in fig. 43, those characters of gut and gonad which best distinguish the four species of Molgula found in the present South African collections. Genus EUGYRA Alder and Hancock, 1870 Eugyra myodes sp. n. (Fig. 44) Diagnosis of species Test provided with fine processes, and covered with sand grains. Body wall thin and transparent. Muscles consisting mainly of (1) short transverse bands FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 207 D Fic. 43. Comparison of gut loop and gonads of A, Molgula falsensis ; B, Molgula scutata; C, Molgula cryptica; and D, Molgula conchata. ©). nak B G Fic. 44. Eugyra myodes sp. n. A, B, specimen with test removed seen respectively from the right; and left C, oral tentacle; D, dorsal tubercle; E, gonad. 208 ANNALS OF THE SOUTH AFRICAN MUSEUM in a line between the siphons and extending down the anterior and posterior margins of the body, and (2) a V-shaped series of bands on each side of the body. About 8 tri-pinnate oral tentacles. Dorsal tubercle with an almost straight oblique slit. Branchial sac with 7 longitudinal bars and regular double spiral stigmata. Primary intestinal loop open at the bend. Right gonad above renal sac. Left gonad in secondary intestinal loop. Gonad consisting of a central sac- like ovary surrounded by lobed testis follicles. Holotype In the South African Museum. 8.A.M. A25623 (U.C.T., LIZ 3 Q). Locality Algoa Bay: U.C.T., LIZ 3 Q. Description The larger of the two specimens is about 1:5 cm. in greatest diameter, and the smaller specimen about 1-1 cm. In outline both are almost circular, and at least in the preserved state the body is laterally compressed to some extent. The siphons are not visible externally and the whole body is covered with sand grains. Numerous fine hair-like processes project from the surface of the test. The test, when freed of adhering sand grains, is thin, flexible and transparent. The internal siphons are separated by a distance equal to about half of the body diameter (fig. 44 A, B). They are short and their margins bear narrow teeth, which on the oral siphon number about 24 and on the atrial siphon about 20. Circular muscles surround each siphon but do not extend beyond their bases. The remaining muscles of the body wall consist of 2 sets. The first set is of short transverse fibres arranged in a line which runs between the siphons and extends in the median plane down the anterior and posterior sides of the body (fig. 44 A, B). The second set is a V-shaped line, on each side of the body, of longer fibres. The body wall is very thin and transparent. There are 8 or 9 tri-pinnate oral tentacles of alternating sizes (fig. 44 CQ). The dorsal tubercle is small, with a straight or very slightly curved slit (fig. 44 D). Behind the dorsal tubercle and to the right of the ganglion lies the small branched neural gland. The branchial sac is delicate. There is quite a long dorsal lamina which increases in width towards the posterior end, and has the margin rolled in. Seven longitudinal bars, only slightly curved, are present on each side, and below each bar a row of stigmata in double spirals. Each component slit of the double spiral makes about 3 coils. The oesophagus is curved. The stomach is not large and has on its right wall a glandular diverticulum with longitudinally folded walls. The primary loop of the intestine is open in its anterior part but closed in the middle region. The rectum bends up and forward and ends near the base of the atrial siphon in the plain-edged anus (fig. 44 B). FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 209 The left gonad lies in the secondary intestinal loop and the right one above the renal sac. In each the ovary is sac-like with a short terminal oviduct. Round the whole margin of the ovary and applied to its mesial face is a series of about 20 lobed testis follicles (fig. 44 E). The ducts of these lead to a main duct which passes along the face of the ovary, to open by a few short slender sperm ducts projecting into the atrial cavity. On the posterior part of the right side is the small straight renal sac. Remarks Few species of Eugyra have been recorded from west, South or east African waters, and of those few none is likely to be confused with the present species. From Walvis Bay, South West Africa, Michaelsen (1914, 1915) has described E. woermanni which, having a gonad only on the right side, properly belongs to the sub-genus Gamaster, and is very distinct from E. myodes. No species of Eugyra is known from Cape Province or the African waters of the eastern Indian Ocean. FE. kerguelenensis Herdman is known from Kerguelen, the coastal waters of Argentina, and also the Patagonian Shelf and South Georgia (Millar, 1960). That species bears the closest resemblance to E. myodes. It also has bands of short muscles on the body wall but these are differently arranged. The position of left gonad mainly within the primary intestinal loop also distinguishes E. short muscles on the body wall but these are differently arranged. The position of the left gonad mainly within the primary intestinal loop also distinguishes £. kerguelenensis, as does the C-shaped opening of the dorsal tubercle. Nevertheless the similarities are enough to suggest a close relationship between the two species. Eugyra macrentera sp. n. (Fig. 45) Diagnosis of species Test with hair-like processes, and covered with sand grains. Muscles confined to the siphons, which have circular and radial strands. Eight bipinnate oral tentacles. Dorsal tubercle with a C-shaped opening. Branchial sac with 7 longitudinal bars. Stigmata in uninterrupted double spirals. Gut long, with the primary loop almost completely closed, and the secondary loop also very narrow. Left gonad in secondary intestinal loop. Right gonad oblique. Renal sac small. Holotype In the South African Museum. 8.A.M. A25624 (U.C.T., LIZ 11 Za). Locality Alcoa pays WG... IZ 11 2.2. Description There is only one specimen and this measures 1-2 cm. by 0-6 cm. The body is oval and is covered with a rather thick layer of sand grains. It appears 210 ANNALS OF THE SOUTH AFRICAN MUSEUM to have been attached to the frond of an alga, part of which is still adhering to the specimen. The test is thin and its surface has numerous fine hair-like processes. The body wall, in spite of some brown pigmentation, is sufficiently trans- parent to allow the gut and gonads to be seen when the test is removed. Narrow, closely spaced circular muscles surround the siphons and stouter radial muscles spread out for a short distance from their bases (fig. 45 A, B). The remainder of the body wall, however, lacks muscles. Narrow teeth form a fringe round the margins of both siphons. (ONS) inn Fic. 45. Eugyra macrentera sp. n. A, B, specimen with test removed, seen respectively from the left and right; C, oral tentacle; D, dorsal tubercle. Eight bipinnate tentacles are situated at the base of the oral siphon (fig. 45 C). The dorsal tubercle is small, with a C-shaped opening facing towards the right (fig. 45 D). The dorsal lamina is rather narrow but long, and has a plain margin. Seven tall longitudinal bars are situated on each side of the branchial sac. They are less curved than in many species of Eugyra. Below each of these bars is a row of infundibula, each with a pair of spiral stigmata. There are, in addition, some smaller accessory infundibula in the spaces between the main ones. Narrow transverse bars separate adjacent transverse rows of infundibula. FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 2II The oesophagus is short, narrow and curved, and the small stomach bears a lobed glandular diverticulum. The intestine is long and folded into a flat double loop, making an almost closed primary loop and a narrow secondary loop. The rectum bends forward and upward, to end in a plain anus near the base of the atrial siphon. In this specimen the gonads are not well developed, and seem to be in a post-spawning state. They are long and rather narrow, consisting of a central tubular ovary fringed by testis lobes. No further details of the gonads could be seen. The left gonad is in the secondary intestinal loop and the right gonad lies obliquely across the centre of the right side, pointing towards the atrial siphon. The renal sac is a small semi-circular body about the middle of the right side. Remarks E. macrentera is remarkable for the length of the intestine and the way in which it is folded. A few species of Molgula, notably M. platei Hartmeyer and M. contorta Sluiter, have a somewhat similar intestinal loop, an arrangement which allows a long intestine to be accommodated in a small space. ANNALS OF THE SOUTH AFRICAN MUSEUM 22 ae ar : ; : sisuasviquaajys “1 oF Si : 3 : pj0aSOL DUIIAND]') x JOATY 1ZVSu0}] * =e ae <5 ; : wnuvni4fo0 DULoJsipopnas + : : ; * wmunjqau *q oF ap Ar " asuadundisy “J a -t- a ; : * wmnsouas wnurj9 0g + : : : : a]DAISND * + g : : asuaqo) wngouds y Avg slouvsy 3S * + : : ; ' wnpuaps “ye ‘uoyAar) ae + : S ; * WNALOUISDIDS “WW ‘UdTONSIIV] é é — : : : - guiofyjas “W , Avg sIouely 1S ¥ = 2 : ; " wnyojnoi9 “Ww as Ar : ; : GB) + + +}+]+ Se aineguog 7, + ap Se aie) ae " wngvauy-oanyf umpydy Q ee) bito) ele eee EAS, eae leo cele ie |e - wel os | S ee Siege aioe eee | oe) S| 6 dear |) ol ot a8 58 Os Eras aS deo es ye EE ee et) ee es: eee moO! sy | » eo) nw Ey ise) ay = Q Opn ’ << pe) me < 5 26. o a 2 E ci |< gk re ea 3 ap 03 Bog oq 4 fe) a & 8 4 [ueg seypnsy ‘purleo7 MON] ‘erperisny ‘odoing ‘purlve7 MOY feIpersny ‘o]9 ‘puryeazZ MON forjoseyuRqng fono1ejUy 4, UOpuoTT \sveq f°OJo “elperjsny ‘soIpuy “MA ‘oj ‘eotmowuy ‘odoin | “eolpy “MM ‘odoin ANNALS OF THE SOUTH AFRICAN MUSEUM x JOATY IZeSU0}T (QuaqUuINIIO;y Aeg [assojy JIATY apoo1g Aeg asjeq Aeg 214%], Aeg eyuepres SOT}ITBIOT IYO, SUIPNjOU!) vorTIpy (uo1]es SUNDITI[OO JSOIVIU SO}VOIPUT y yseq osonsnz10g JOATY Ss ueuysng elnsutusg oden “eolIpy YINOG UT jt) uoosey uveqosuey 214 S1sDI10S DGLVION XAT sypsaqojiun Ddinz01puoudyy wungqUDsIs “IVA WNAsU “gg 1y9a] sapropjcuog é sSnjpuouD “gq * snoaonusvU snjcuog " pjoduina 17)a40/) sisuaraupas * fy "sapouas DIpLIsy SupUlsajut DUuor’) *wnupisajsy] DULOSO]GIy sULsofiagasaa wnuUapIprs J, wnavd WNULLIOSSYT * sasuadva sapiurj90jqaT wunosopuspu (uojoizudsdjog) * 215 FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS -- + ; pauysafap *} vsourds vryjudz0j0F7 = + = : : punnifo v1uazj0g "BOlIFY ISOM + “- : : ; snqojnounpad "JW ‘TeveN ‘suorsor weM Aueypy a é : : ; * snyosagsoxa “Ww + +]°- * — snqjdygosrjo snwusoz0sny 4. + ; : * , sisuaqvo *q “eljelsny SYIOTON Wog ‘ye1eN +/+] + - — ; ; : p1afiuojojs Dandy aL . ° . . DIDIIPDL ‘$ ‘Teen aden a a : : : * pupsanbivu *¢ ‘Aeg ziJopn'y — + : : 3 “sipynsuv D1akq¢ + : : : pioygouups¢ °F “Aeg ziJopn'y +) + : : DAJQULULASD DGADIOPIWIAU’) -|- : ; : ppiqns vgivrog ~ +y4+ t+ -- : : pauinsuv DgivI01pubdd] 0g +r : * srapjnaiasvf vgivI01puvjayyy x Ab SIOURIT “3S x + : $ * sisuadv2 vd1v9090/) 7 Oo Are ae) or] Hh) eel] 482 Co o = x 2) iS S | & 4 _ al) Se o os oe aaa e®elel sisi |e1eislol oi cles BS Bee | S| ee | ey aS 3 a Canes ® FES SEES s.| 3 el ee |e] F|e|s eel fer ies geo ‘es < x 5 =) | es PP ws of: fw Ze) 5 c fe | eh 8 > @) —< ion tev) ai — © = 3 re ta 53.4 o = 0 te GLAS) fe) 5 Q a ® g 216 ANNALS OF THE SOUTH AFRICAN MUSEUM (uOl}e}S SUTJDOT[OO JSOIVIU SO}VIIPUT y. “Bory YING Ur jr) SOTPTCIOT JOYIO SuIpnpoul) voLyy (Quoquind10yy yseq osonsnj10g keg 1soyy | ueqing | JOATY S ueuysng | keg voaty | | + o |+]+| somes SENG ELS ene i Ea Arenysy JIA opooig | | | sera] | LL einsutusg aden | | | | | | | ee) uooseT] uvegosury | | | ab | | | | a S| 8 sae ies SS Sine eats reiiecy| ee aes as SPS PSS Ra FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 24107) DISTRIBUTION OF SPECIES Geographical components of the ascidian fauna Stephenson (1944) has analysed the littoral fauna and flora of the South African coast between Port Nolloth on the west and Durban on the east, and recognizes the following components: (1) the warm-water component, (2) the cold-water component, (3) the ubiquitous component, (4) the south coast component, (5) the more local components. The ascidians described in the present paper (Table 2) and in a previous one (Millar, 1955) can be grouped into the first four components defined by Stephenson. Excluding doubtful records and insufficiently known species, the ascidians can be divided as follows: The warm water component: Aplidium sarasinorum Polyclinum isipingense P. arenosum P. constellatum Pseudodistoma africanum Cystodites dellechiajer Eudistoma coeruleum E.. modestum E. reniert Ascidia pygmaea Botrylloides nigrum var. giganteum Symplegma viride Polyandrocarpa anguinea Polycarpa rubida Styela plicata S. aequatorialis S. marquesana Microcosmus exasperatus M. pedunculatus The cold water component: Aplidium flavo-lineatum A. pantherinum Polyclinum neptunium Cystodites roseolus Trididemnum cerebriforme Leptoclinides capensis 218 ANNALS OF THE SOUTH AFRICAN MUSEUM Corella eumyota Botryllus magnicoecus Cnemidocarpa asymmetra Microcosmus oligophyllus Molgula scutata The ubiquitous component: Diplosoma listerianum Ciona intestinalis The south coast component: Aplidium circulatum A. retiforme Clavelina roseola Sycozoa arborescens Distaplia capensis Ascidia sydneiensis Agnesia glaciata Gynandrocarpa unilateralis Dextrocarpa solitaris Alloeocarpa capensis Styela angularis S. pupa Pyura capensis P. stolonifera Boltenia africana Molgula falsensis M. conchata Table 3 shows the number of species in each of the four main components, as a percentage of the total number of ascidian species collected, and also shows the comparable percentages obtained from Stephenson’s (1944) figures of all species of littoral animals. TABLE 3 Percentage of total species Ascidians All littoral animals Warm water component 39 52 Cold water component 22 18 Ubiquitous component 4 8 South Coast component 35 13 FURTHER DESCRIPTIONS OF SOUTH AFRICAN ASCIDIANS 219 It should be pointed out that the figures are not strictly comparable because the ascidians include both littoral and sublittoral material. The ascidian fauna compared with the littoral animals as a whole, shows a warm water component almost as important, and a relatively more important south coast component. The cold water component and the ubiquitous com- ponent are about the same relative size in ascidians as in the total of littoral animal species in South Africa. There is therefore quite good agreement between ascidians and other groups. The distribution of South African species in other regions shows some interesting features. Particularly notable is the high proportion of the ascidian fauna which appears to be endemic. Care must be used in comparing the results in the present paper with those of other authors dealing with different areas, owing to the persisting difficulties of ascidian systematics, and the con- sequent possibility of confusion of species. Nevertheless it appears that more than half of the species described in the present account are endemic. As pointed out by Ekman (1953), however, the faunas of the neighbouring regions, South West and south-east Africa, are very imperfectly known; it is therefore likely that the list of endemic species will shorten as knowledge of these areas increases. The wider affinities of the South African ascidian fauna appear to be with the adjacent areas of the warmer Indian Ocean. Only Apflidium retiforme, Corella eumyota, and Agnesia glaciata are found also in any of the cooler waters of more southerly regions. In this respect the South African ascidian fauna con- trasts markedly with that of southern South America where many species are shared with the Subantarctic and Antarctic regions (Millar, 1960). The tem- perature regimes of southern South America and southern South Africa are, of course, very different, but it seems likely that the greater geographical isolation of the latter may also have been important in preventing fusion of the faunas of South Africa and more southerly areas. In other respects also the ascidians of South Africa and South America are dissimilar. Several South American ascidians reach a very great size, as for instance Distaplia cylindrica (Lesson), and the molgulids Ascopera gigantea Herd- man and Paramolgula gregaria (Lesson). No species of a comparable size is known from South Africa, except Szgillina vasta sp. n., which is known only from the incomplete type specimen. Differences in the abundance of phytoplankton may be responsible for the different sizes of the ascidians in the two areas. SUMMARY 1. A systematic account is given of the ascidians from South African waters contained in several collections. 2. Sixty-four species are identified, and five are given provisional identifi- cations. Of the total, seventeen are described as new species, and one as a new form of a known species. 3. The distribution of the species is briefly discussed. 220 ANNALS OF THE SOUTH AFRICAN MUSEUM REFERENCES Car.isLeE, D. B. 1952. On ampullary tissue in the larva of Polyclinum aurantium Milne Edwards. j. Mar. biol. Ass. U.K. 31, 63-64. Car.isLeE, D. B. 1954a. Notes on the Didemnidae (Ascidiacea). II. The number of rows of stigmata in Didemnum gelatinosum Milne Edwards and in Didemnum maculosum (Milne Edwards). 7. Mar. biol. Ass. U.K. 33, 27-32. Car isLe, D. B. 19545. Notes on the Didemnidae (Ascidiacea). III. A comparison of Didemnum maculosum, D. candidum, D. helgolandicum and Trididemnum alleni. F. Mar. biol. Ass. U.K. 33, 313-324. DELLA VALLE, A. 1877. Contribuzione alla storia naturale delle Ascidie composte del Golfo di Napoli. Neapel. EKMAN, 8S. 1953. i a Y WW —AIMAX CS €\ 7 : Int.c ay Ws “@O. NS Fic. 5.—Bauria cynops. Ventral view of palate x 14. (Key to lettering, p. 255.) movement of the lower jaw would account for their shape. It is also unlikely that the abrasive action of food particles would produce such constant and sharply defined wear-facets. It is possible, therefore, that in Ericiolacerta more complex jaw movements were possible. Watson (1931) has described upper cheek teeth as . transversely widened, the high original cusp passing on its lingual side into a ridge which separates two concave areas into which grind the posterior and anterior cusps of the two lower molars between which it fits’. The present study does not confirm this arrangement. The wear patterns clearly show that the labial cusps of the uppers did not occlude outside the opposing lowers. On the contrary, the apices of the crowns when present must have met one another. The teeth are not transversely expanded to the extent suggested by Watson, and at their bases the crowns are circular in cross-section. 6 DENTITION AND TOOTH REPLACEMENT IN TWO BAURIAMORPH REPTILES 241 THE DENTITION OF BAURIA CYNOPS The palate (Plate XIII A and fig. 5) has been well described by Boonstra (1938) and Broom (1937). The secondary palate (sec. p.) is formed only by the premaxillae and maxillae. It is narrowest across the posterior border which lies adjacent to the 5th postcanines. Anteriorly to the posterior border of the secondary palate the maxillae curve sharply upwards to their contact with the premaxillae (P.M.). The premaxillae have a broad tongue-shaped process which extends backwards in the mid-line to meet the maxillae. Two ovoid foramina (p. for.) are present between the lateral edges of this structure and the maxillae. These foramina were not reported by Boonstra or Broom. From the anterior edge of these foramina an interdigitated suture between the premaxillae and maxillae passes outwards through the diastema between the last incisor and the canine. No evidence for Broom’s view that part of the vomer is visible between the premaxillae and maxillae could be found. Upper dentition (Plates XII B, XIII, and figs, 5, 6, 8) There are 4 incisors, 1 canine and 10 postcanines in the upper jaw. This is in agreement with the findings of Broom (1909), Boonstra (1938) and Brink and Kitching (1953). The roof of the diastema between the incisors and canines has the form of a shallow pit and receives the lower canine. A thin strip of the maxilla forms a labial border to this pit and therefore the tip of the lower canine was not visible when the jaws were closed. The crowns of all the incisors and both canines were damaged, but the postcanines are fairly well preserved, except for damage to their outer surfaces. Because of this damage it was essential to prepare both left and right dentitions. The postcanines of each side are arranged to form a curve. In the region of the 5th postcanines the opposing curves are closest to one another, and anterior and posterior to this point they diverge away from each other (fig. 5). Figures of the length of the postcanine series are of little value unless it is clearly stated how the measurements were taken. Measured along the lingual surface of the upper postcanine series they measure 4-1 cm. In crown view (fig. 6) the postcanines are oval with the lingual side of the crown considerably wider than the labial. Brink and Kitching (1953) claimed that in the third specimen of Bauria the crowns of adjacent teeth expand so that they come into contact with one another, but that their necks are separated from one another. It is clear in the type, however, that the majority of the postcanines were separated by narrow gaps with an average width of 0-6 mm., although in some cases teeth have been slightly displaced with the result that consecutive crowns touch one another. The form of the crowns of the teeth is of great interest because of the marked changes they undergo as a result of wear. These changes are essential for the interpretation of jaw action and tooth replacement, and will therefore be described in detail. Interpretation of the wear pattern is complicated by 242 ANNALS OF THE SOUTH AFRICAN MUSEUM Fic. 6.—Bauria cynops. A, B and C, reconstructed sections through the 4th, 5th and 7th right upper postcanines and D, crown view of the upper postcanine dentition. All x 3. (Key to lettering, p. 255.) ORR le mse ce ee ae ee ee DENTITION AND TOOTH REPLACEMENT IN TWO BAURIAMORPH REPTILES 243 the fact that the occluding tooth rows do not lie parallel to one another, but at a small angle to one another. The majority of the upper postcanines lie slightly further lingually to the lowers, but the anterior upper postcanines and canines lie slightly labially of the corresponding lower teeth. In order to understand the shape of the crowns, outline drawings of the posterior view have been given in addition to the crown views (fig. 8). The 5th postcanine (fig. 6 B, rep. t. 5) on the right side was preserved shortly before it was due to be replaced, and both the replacing and functional teeth are in good condition (Plate XII B). Contrary to all previously published reports, the crown of an unworn upper tooth is not flat. On the contrary, it supports a prominent cusp (lab. c.) on the narrow labial edge of the crown while the wide lingual edge of the crown supports a crenulate ridge. In posterior view the crown has a concave centre, the centre point of the concavity lying towards the labial side of the crown. A broad, ill-defined transverse ridge extends a short distance across the crown from the apex of the main cusp. On either side the crown falls slightly away. On the left side the replacing tooth of the 5th postcanine was preserved shortly after it had fully erupted and become the functional tooth; it has the same crown structure as the incompletely erupted counterpart on the right. The oth, 7th and 4th are little worn, and a labial cusp and prominent lingual edge can still be recognized. The 2nd upper postcanine is also little worn and consists of a high labial cusp and small, ill-defined lingual ridge at a much higher level than the labial cusp. Wear is greater on the lingual side in this tooth, whereas in the posterior teeth wear is greatest on the labial side. This is due to the crossing of upper and lower postcanine rows. In the 8th and 6th the crowns show advanced stages of wear. Two prominent wear-facets are visible. The labial cusp has been worn away completely to a flat wear-facet extending lingually and slightly upwards to the centre of the crown where it meets a flat oblique wear-facet extending upwards and inwards from the lingual edge of the crown. The orientation of these wear-facets is best seen in the posterior profiles and stereo-photographs. In the functional 5th postcanine on the right, wear had proceeded further. The labial wear-facet extends further across the crown. The prominent ling11 edge and oblique lingual wear-facet have been considerably reduced in size. In the 3rd tooth the crown is worn to a practically horizontal plane. The ist and 1oth postcanines are considerably smaller than the other postcanines and consist of a high labial cusp lingually of which the crown falls away rapidly. The measurements of the individual postcanines in mm. are as follows: WIGH SANE: POSte | 2-2 VR 2eAy OR WOO) | 90 | 29) | 321 || 2:7 | 297 2 Wax vine apy o°On SO) Asti) 4e7 | 5S. lea-7 | 51 14s | 2:9 2 244 ANNALS OF THE SOUTH AFRICAN MUSEUM Fic. 7.—Bauria cynops. A, crown view of lower postcanines and B and C, sections through mandible to show the roots of the 6th and 7th postcanines. All x 3. (Key to lettering, p. 255.) ee ee ’ eS a on < ee en ee eee ee eee Oe ee) DENTITION AND TOOTH REPLACEMENT IN TWO BAURIAMORPH REPTILES 245 Lower dentition (Plate XIV; figs. 7, 8, 9) There are three lower incisors. This confirms the findings of Brink and Kitching (1953). It is easy to understand why Broom (1909) considered that the type had four. A narrow diastema separates the 3rd lower incisor from the canine. The crown of the 4th upper canine fits into this diastema. The anterior region of the snout is damaged in the type in such a way that the crown of the upper 4th incisor is preserved fitting into this diastema, and is broken off at the UPPER POSTCANINES. hoes Gr, LA | a fa LOWER POSTCANINES. (oobi aa aemm Fic. 8.—Bauria cynops. Posterior views of upper and lower postcanines. All x 3. same level as the remaining lower incisors so that it can easily be mistaken for a lower incisor. The postcanine series follow immediately after the canine without an intervening diastema. As in the case of the upper jaw the postcanines of both sides are arranged to form an arc, and the two arcs diverge away from one another in front and behind the 5th postcanine. Eleven postcanines are present on both sides. In crown view the postcanines are transversely ovate but have a greater eccentricity than the corresponding uppers. The roth left postcanine shows the least signs of wear (Plate XIV B; fig. 8). It consists of a prominent labial cusp. The remainder of the crown is flat except 246 ANNALS OF THE SOUTH AFRICAN MUSEUM for a faint ridge along the lingual edge and a prominent medium transverse ridge. This ridge is drawn upwards to form an ill-defined cusp in the centre of the crown. Faint ridges extend from the apex of the main cusp along the anterior and posterior borders of the crown. On the right side the 1oth post- canine is an old tooth which had not been replaced before death, and is greatly worn. In the oth left postcanine the main cusp has been lost, but the medium transverse ridge, although slightly worn, can still be recognized. In the 4th, 5th and 7th postcanines initial wear stages are present. The central ridge has been worn away and a basin is present in the centre of the crown lingually to the main cusp. This basin is flanked in front and behind by the two ridges extending inwards from the apex of the main cusp. In the 8th tooth wear is further advanced and this basin has deepened considerably. Consequently the main cusp is more prominent in this tooth than in unworn teeth. In the 6th postcanine wear has advanced further. The main cusp is low and an oblique wear-facet is present on the lingual surface of the crown. The crowns of the 1st and 3rd teeth are smaller than adjacent teeth and are worn to a flat horizontal surface. The 11th postcanine is much smaller than the remaining postcanines; it consists of a prominent labial cusp lingually of which the crown falls rapidly away. As in the upper postcanines, a small gap separates consecutive crowns. The gaps are larger anteriorly than posteriorly. The measurements of the individual postcanines, given in mm., are as follows: Max. ant.-post. | 2-2 | 2-5 >|) 2:2.) 9°@ | 8:0) | 9-9 | 9-4) 9-0 || 9-9 sia males Max. ling.-lab. | 3-2 | 4-2 | 3-4.) 4:8 | 4:8 | (4°83 | 5-4 | 4:85) 5-1 eee Relationship between the wear pattern of upper and lower postcanines An outstanding characteristic of the dentition is that in the initial wear stages the upper teeth are subject to wear mostly on their labial edge and centre of the crown, whereas in the lower postcanines wear is greater in the centre of the crown and on the lingual edge. It is only in very advanced wear stages that postcanines tend to acquire nearly flat occluding surfaces. Although an attempt has been made to illustrate (fig. 10) the three progres- sive stages of wear of Bauria postcanines, it is naturally difficult to determine from one specimen the exact changes the postcanine teeth underwent as a result of wear. Careful investigation of other bauriamorph specimens will be necessary to confirm these findings. The initial wear (fig. 10 A, I) in the centre of the lower postcanines, and wearing away of the upper labial cusps, is apparently due to the upper labial cusp meeting the lower postcanine lingually of the lower labial cusp. The lingual edge of the upper tooth occludes lingually of the la DENTITION AND TOOTH REPLACEMENT IN TWO BAURIAMORPH REPTILES 24°7 lingual edge of the lower teeth and this results in the formation of the oblique wear-facet extending upwards and inwards from the lingual edge of the upper postcanines. In a more advanced stage (II) an oblique wear-facet is developed on the lingual edge of the lower teeth where the oblique wear-facet of the uppers meets the lowers. Consequently both upper and lower are worn away on their lingual sides during this phase. In an advanced wear stage (III) all B unworn Fic. 9.— Bauria cynops. A, medial view of the right lower postcanine series and B, graph to illustrate decrease in extent of wear of odd and even-numbered post- canines. Wear stages I, II, and III correspond to those shown in Fig. 10. A X 3. (Key to lettering, p. 255.) the prominent features of the crown tend to be worn away so that only flat occluding surfaces remain. The height of the crown above the alveoli borders varies greatly in consecu- tive postcanines. As a general rule the height is greater in unworn teeth than in worn teeth. The result is that wear-facets of consecutive teeth are seldom on the same horizontal plane (fig. g A). The correct interpretation of the degree of wear is of importance in determining the order of tooth replacement in the postcanine series. Roots of postcanines and alveoli Several teeth are loosely held in enlarged alveoli whereas others are tightly held by their alveoli. In the latter the alveoli are built up on the lingual 248 ANNALS OF THE SOUTH AFRICAN MUSEUM side of the tooth in such a way that a horizontal shelf is formed at the base of the lingual side of the postcanines. This is more marked in the upper jaw than in the lower. As a general rule worn teeth are tightly held and unworn teeth loosely. A fracture through the right ramus of the lower jaw has exposed the roots of the 6th and 7th postcanines (fig. 7 B, C). The single root of the 7th is substantial and tapers gradually to a point. There is no indication of any division of the root. The alveolar bone is closely apposed to the root. A narrow unworn ul Mt A Fic. 10.—Bauria cynops. A, occlusal relations of unworn and three wear stages typical postcanine teeth and B, three wear stages compared with unworn upper and lower postcanines. All x 3. bony transverse septum separates successive teeth. The remnants of the septum between the 6th and 7th obscure part of the root of the 6th postcanine. Preserva- tion of the dentary surrounding the root is not exceptional, but it is possible to recognize large cavities and canals lingually to the apex of the root. In addition the root is slightly resorbed on its lingual side. Unfortunately no sections through the root of an upper postcanine exist. Reconstructed sections based on external features only are given in fig. 6 A, B, C. Features that could not be observed are shown in dotted lines. Posterior to the secondary palate a thick mass of bony tissue constituted by the maxilla and a thin plate of the palatine (PAL) lies lingually to the last six teeth (fig. 6 C). The medio-ventral edge of this bony mass is drawn downwards to form a prominent ridge so that a narrow groove separates the outer surface of this ridge from the teeth. A longitudinal section was cut through the bony mass. Numerous cavities and canals are present in it, but their shape and course could only be accurately determined if serial sections were made. Anteriorly this lingual mass of bone is obscured by the presence of secondary palate, but DENTITION AND TOOTH REPLACEMENT IN TWO BAURIAMORPH REPTILES 249 the ridge on the lingual edge of the lingual mass continues forward close to the lingual surface of the teeth. Tooth replacement Only in one case are a functional tooth and its successor preserved and visible in a single alveolus. This supplies important information on how tooth replacement actually takes place, and details on the structure of an unworn crown. The tooth preserved in the act of replacing is the 5th right upper (fig. 6 B). Its alveolus is greatly enlarged. The functional tooth lies hard against its labial edge. The crown of a replacing tooth is situated lingually and slightly behind the functional tooth. The lingual edge of the replacing tooth is preserved at the same level as the edge of the alveolus. The replacement tooth is obliquely orientated and does not, as would be expected, erupt vertically downwards, but rather downwards and inwards. The lingual edge of the root of the functional tooth has been resorbed and the labial edge of the replacing tooth is closely apposed to it. After the functional tooth was lost, the crown of the replacing tooth would presumably have moved in a labial direction to the vertical position. On the left side replacement was in advance of that on the right because the old functional tooth has been shed and an unworn replacement tooth is present. Although the alveolus is still greatly enlarged the new tooth is vertically orientated. Alveoli size is therefore important in determining the order of tooth replacement. Although horizontal sections through the bony mass lingual to the last 7 upper postcanines revealed no additional replacing teeth, the wear patterns of adjacent teeth appear to indicate that active replacement of the postcanines was probably still taking place in both upper and lower jaws although in no case in the lower jaw are both the replacing and old functional teeth preserved in a single alveolus. A probable explanation for this is that the replacing teeth developed initially deep in the maxilla or dentary lingually to the apex of the roots of the functional teeth. An empty cavity in this position alongside the root of the greatly worn 6th lower postcanine appears to confirm this view. Replacing teeth would therefore only become visible lingually to the functional teeth shortly before eruption. It was considered inadvisable to section the type to confirm this view. Boonstra (1938) mentioned a series of replacing teeth lingual to the functional postcanines exposed by grinding in the American Museum specimen, but did not give a detailed description. In Ericiolacerta the replacing teeth are visible throughout their develop- ment, and the order of replacement is easily determined. In the presentspecimen of Bauria the order of replacement has to be determined by other criteria such as the degree of wear of the postcanines, size of alveoli, and size of individual teeth. These alone are insufficient because, for example, if replacement in both jaws is not accurately synchronized, occluding teeth of different ages may well produce different wear patterns in teeth of similar age. It seems clear from the presence of unworn teeth in the postcanine series 250 ANNALS OF THE SOUTH AFRICAN MUSEUM and the actual replacement observed, that tooth replacement is still taking place throughout the postcanine series. The available evidence suggests that teeth replace alternately in Bauria as in Ericiolacerta. For example, in Bauria, in the right mandible (fig. 9 A), the 6th, 8th and roth postcanines are slightly smaller, are more tightly held by their alveoli, and are lower in height above the alveolar border than the 5th, 7th or gth. Wear is certainly further advanced in the former group. Similar conditions can be observed in the left mandible. The gth, 7th and 5th postcanines are larger, more loosely held in their alveoli, and higher than the 6th and 8th. In the left mandible, replacement is in advance of that of the right. In the left maxilla the 5th and 7th are large, unworn, loosely held in their alveoli, and high-crowned in contrast to the 6th and 8th. Replacement in the right maxilla also lagged slightly behind that in the left, and the replacing tooth for the 5th had not yet erupted before death. Alternate teeth throughout the series did not erupt simultaneously. For example, the right 7th upper postcanine already showed signs of wear although the 5th had not yet erupted. On the left the 5th alveolus is larger than the 7th. The 10th postcanine in the left mandible had been replaced only shortly before death, yet the 8th is well worn and in life would presumably have been the next to be replaced. This evidence suggests that a wave of replacement moved forwards along the odd- and even-numbered postcanines. In the case of the maxilla, the wave that affected the odd-numbered postcanines was in advance of that which affected the even-numbered, but the waves are spaced so that alternate replacement is present over short distances. An expected result of replacement waves moving in an anterior direction would be that the wear of alternate teeth would diminish in an anterior direc- tion. There is some evidence of this (fig. 9 B). In the right mandible the 1oth postcanine is greatly worn and was about to be replaced. The 8th and 6th teeth are well worn but the 4th is less worn. The 2nd tooth appears to have erupted shortly before death and shows few signs of wear. The gth, 7th and 5th, on the other hand, appear also to have erupted shortly before death and are not much worn, whereas the 3rd is small, greatly worn and loosely held in its alveolus, and presumably would have been the next to be replaced. The three wear stages upon which fig. 9 B is based correspond to those shown in fig. 10. Similar facts were observed in the maxilla. Tooth replacement in occluding upper and lower jaws appears to have been fairly closely correlated, with the result that a recently erupted and unworn tooth of the upper jaw would have met a recently erupted and unworn tooth of the lower jaw. For example, in both upper and lower jaws on the right side the gth, 7th and 5th postcanines appear to have been replaced shortly before death. DIscussION In both Ericiolacerta and Scaloposaurus there was a marked reduction in the size of the canines. The postcanine teeth of both genera are similar. Both are DENTITION AND TOOTH REPLACEMENT IN TWO BAURIAMORPH REPTILES 251 circular in cross-section at the alveolar border and in both there is a tendency to develop small cusps although these are limited to the lower jaw in Ericiolacerta. In Scaloposaurus the shearing action of the postcanine teeth has been retained, but in Ericiolacerta the reduction of the canines has apparently made it possible for the apices of the upper and lower postcanines to meet despite the limited area of contact, and the apices of the crowns are completely worn away in old teeth. It is not possible to determine whether the shearing action was retained in Ericiolacerta or whether a limited amount of lateral and propalinal jaw move- ment was possible. One of these three movements may account for the wear- facets on the postcanines of Ericiolacerta which could not be accounted for only by a simple dorso-ventral movement of the lower jaws relative to the uppers. In Bauria and all bauriamorphs from the Cynognathus zone the teeth have expanded transversely to increase the occlusal areas. The crushing action of Bauria is therefore much more efficient than that of Ericiolacerta, and the latter appears to be very close to ancestral forms which deviated away from a simple sectorial action of apposing dentitions. A similar shift from a sectorial dentition to a crushing dentition took place in some early mammals (Patterson, 1956), e.g. the dryolestids. Here, too, there has been a reduction in the size of the canines. The development in the bauriamorphs paralleled that of some of the later mammals and consequently there is no question of the known bauriamorphs being ancestral to any known Jurassic or Rhaetic mammal. A possible exception is Dvarthrognathus. ‘This form may have developed from early bauriamorphs, and because it possesses a squamoso-dentary articulation may therefore be classified as a mammal. The late cynodonts also developed a crushing dentition, but a reduction of the size of the canines does not appear to have triggered off this development. In cynodonts such as Diademodon the sectorial action was retained and a crushing part of the tooth added lingually to the original labial shearing cusps, the apices of which never met. It has been tacitly assumed that the flat crowns of Bauria are the result of a grinding action of the jaws. Because of the mode of replacement in Bauria the crowns of the postcanine teeth are of varying heights. Consequently, for individual teeth of the upper and lower jaws to have met it is necessary for the teeth to intermesh. This would appear to have ruled out any significant propalinal movement of the mandible. The large upper canine, the pit for the lower canine, the large lingual cusps of the lower postcanines, and the transverse processes of the pterygoid would probably eliminate any significant lateral movement of the mandible. For this reason it is unlikely that true grinding was developed in Bauria but that the upper and lower teeth could only crush food as the result of a simple pounding action. Unfortunately details of the nature of the mandibular joint in Bauria are not available. Recent work to be published shortly on the jaw action of gomphodont cynodonts has indicated that it is unlikely that any propalinal movement was present in these forms and that the 252 ANNALS OF THE SOUTH AFRICAN MUSEUM wear pattern of the postcanines which suggest propalinal movement are due to other factors. In Ericiolacerta and Bauria it has been shown that the teeth replaced alternately. There is no question, therefore, that a mammalian type of dental succession was present in either of these bauriamorphs. In Friciolacerta it has been shown that the replacing teeth developed initially in distinct pits in the dentigerous bones and lie lingually to the functional dentition. The pits were connected by a shallow longitudinal groove, which presumably housed the dental lamina. As the replacing teeth increased in size their pits enlarged and became confluent with the alveoli of the functional teeth, and the replacing teeth moved labiad. Recent work to be published in the near future has shown that in the Cistecephalus zone cynodonts, in Thrinaxodon and in Tribolodon, the replacement teeth developed in a similar fashion. This type of replacement was not confined to mammal-like reptiles. In his excellent résumé on tooth replacement Edmund (1960) has cited the relevant literature and has pointed out that in Wothosaurus and Pleisiosaurus replacing teeth developed in an almost identical fashion to that described above. Edmund (1957) has shown that in ornithischian dinosaurs a groove for the dental lamina connected a series of foramina which penetrated the dentigerous bone and opened into the alveoli of the functional teeth. Tooth germs budded off from the dental lamina and passed through these foramina into the base of the alveoli. In Bauria the teeth did not develop in superficial pits, but apparently deep in the dentigerous bones lingually to the base of the roots of the functional teeth. No indication of a groove for the dental lamina or any foramina (guber- nacular canals) which connected the developing teeth with the oral epithelium could be observed, but better material is necessary to confirm this point. This may be interpreted to mean that tooth replacement has ceased in the particular specimen of Bauria described above, but this is unlikely. To obtain the type of replacement found in Bauria and in mammals it was necessary for the dental lamina or remnants of it to lie not superficially but to be buried within the dentigerous bones. A similar development has taken place in the crocodiles (Edmund, 1960), where remnants of the dental lamina remain in each alveolus and retain their capacity to form teeth. The reason for the shift in Bauria to a deep site for the development of the replacing teeth was presumably correlated with the transverse widening of the teeth in this form. If the develop- ing teeth were to have developed in superficial pits they would have projected into the oral cavity and either been damaged by or hindered mastication. A similar shift from a superficial to deep-lying position took place in other groups which have transversely widened teeth, e.g. Diadectes, Watson (1954). In Diademodon replacing teeth also developed deep in the dentigerous bones below the functional teeth, but the groove for the dental lamina remained in the superficial position lingually to the bases of the postcanines. A similar shift from a superficial site for replacing teeth to a site deep in the dentigerous bones probably took place in the ancestors of mammals. In mammals, in contrast to DENTITION AND TOOTH: REPLACEMENT IN TWO BAURIAMORPH REPTILES 253 Bauria and Diademodon, developing permanent teeth retain their contact with the oral epithelium through their gubernacular canals. The origin of the mammalian dental succession from the types of dental succession reported in mammal-like reptiles remains an unsolved problem. Bauria shows that alternate replacement was present in advanced mammal-like reptiles and that teeth did replace alternately even though expanded transversely. It is pertinent to mention that alternate replacement was also still present in the advanced mammal-like reptile (or early mammal) Diarthrognathus from the late Trias. This does not support the view of Bolk (1922) that as the postcanines became more complex they would no longer erupt alternately but that one of the alternate-numbered series would actually replace the other series, one becoming the deciduous and the other the permanent. The specialized condi- tions in gomphodont cynodonts do not help to solve the problem of the origin of the mammalian succession. An alternate explanation for the mammalian dental succession has been suggested by Edmund (1960). Edmund has shown that in early tetrapods such as Captorhinus the teeth of a single zahnreithe erupt as a unit and are ankylosed to the dentigerous bones. Edmund has suggested that in mammals the deciduous teeth represent the remnants of one zahnrethe and permanent dentition a second anterior zahnreihe. This theory also explains the front-to-back order of replacement in mammals and, because of its simplicity, is extremely attractive. But certain problems present themselves. Edmund made no attempt to correlate his theory with palaeontological evidence. Eruption and functioning of teeth in a complete zahnrethe is known only in a few specialized groups such as the captorhinids. In mammal-like reptiles alternate tooth replacement similar to that of typical reptiles is present except in the specialized gomphodont cynodonts, where the position is difficult to ascertain. Edmund’s theory would imply that in the late mammal-like reptiles there should be a tendency to reduce the number of teeth to two zahnreihen, but no tendency in this direction has been reported. Although these facts alone do not disprove Edmund’s view they certainly do not confirm it. An alternative explanation for the origin of the mammalian conditions is worth while considering. It is possible to derive the mammalian succession from forms in which the time lag between the budding-off of the even- and odd-numbered tooth germs from the dental lamina is reduced so that odd- and even-numbered teeth can erupt consecutively and function simultaneously. Even in mammals consecutive teeth in the series do not erupt one after the other in an orderly fashion; e.g. in man the first deciduous molar erupts before the deciduous canine, and the first molar erupts before the second premolar. Compared with the pelycosaurs there appears to be a tendency to reduce the relative time-interval between the eruption of alternate teeth in later mammal-like reptiles. In pelycosaurs there is a tendency for the odd- and even-numbered teeth not to function simultaneously. This has been termed ‘functional distichial replacement’? by Kermack (1956). He has pointed out 254 ANNALS OF THE SOUTH AFRICAN MUSEUM that functional distichial replacement is not present in the therocephalians and gorgonopsians studied by him. On the other hand, although alternate tooth replacement is present in Thrinaxodon and Bauria, the functional periods of odd- and even-numbered teeth overlap considerably. A reduction of the time-lag between the formation of odd and even teeth together with a lengthen- ing of the time between eruption and replacement of individual teeth would be expected if the form were moving towards the mammalian type of replace- ment. It is difficult because of lack of reliable evidence on tooth replacement in mammal-like reptiles to determine a trend of this nature, but tooth replacement was apparently far less frequent in a form such as Bauria than in Enriciolacerta. A similar reduction of the frequency of replacement can be traced in the later cynodonts compared with those of the Cistecephalus and Lystrosaurus zones. What is urgently required is further growth series of individual mammal-like reptiles. SUMMARY (1) The dentition, tooth replacement and wear stages of the postcanine teeth of the type specimens of Ericiolacerta parva and Bauria cynops are described. (2) In both genera alternate tooth replacement was present. In Enciolacerta the replacing teeth developed initially in pits in the base of the superficial groove which supported the dental lamina. In Bauria replacing teeth appear to have developed deep in the dentigerous bones. The origin of the mammalian order of replacement is briefly discussed. (3) Ericiolacerta represents an early stage in the transition from a sectorial to a pounding or crushing dentition. This transition was probably correlated with reduction of size of the canines in this form. In Bauria a crushing dentition is better developed than in Ericiolacerta. (4) It is doubted whether a true grinding action was present in Bauria. Wear of the postcanines could be accounted for solely by a simple pounding action. REFERENCES Bok, L. 1922. Odontological essays. Fifth essay. 7. Anat. Lond. 57, 55-75. Boonstra, L. D. 1938. On a South African mammal-like reptile, Bauria cynops. Palaeobiologica 6, 163-183. Brink, A. S. 1955. A study of the skeleton of Diademodon. Palaeont. Afr. 3, 3-39. Brink, A. S. 1956. Speculations on some advanced mammalian characteristics in higher mammal-like reptiles. Palaeont. Afr. 4, 77-06. Brink, A. S. & Kircuine, J. W. 1953. On some new Cynognathus zone specimens. Palaeont. Afr. 1, 29-48. Brom, F. & ScHRO6DER, J. 1935. Beobachtungen an Wirbeltieren der Karrooformation. VII. Ein neuer Bauriamorphe aus der Cynognathus-Zone. S.B. bayer. Akad. Wiss. 1935, 21-36. Broom, R. 1905. Preliminary notice of some new fossil reptiles collected by Mr. Alfred Brown at Aliwal North, S. Africa. Rec. Albany Mus. 1, 269-278. Broom, R. 1909. Notice of some new South African fossil amphibians and reptiles. Ann. S. Afr. Mus. 7, 270-278. Broom, R. 1911. On the structure of the skull in cynodont reptiles. Proc. zool. Soc. Lond. 1911, 893-925. DENTITION AND TOOTH REPLACEMENT IN TWO BAURIAMORPH REPTILES 255 Broom, R. 1913. On evidence of a mammal-like dental succession in the cynodont reptiles, Bull. Amer. Mus. nat. Hist. 32, 465-468. Broom, R. 1932. The mammal-like reptiles of South Africa. London: H. F. & G. Witherby. Broom, R. 1937. On the palate, occiput and hind foot of Bauria cynops Broom. Amer. Mus. Novit. No. 946, 1-6. Crompton, A. W. 1955a. A revision of the Scaloposauridae with special reference to kinetism in this family. Navors. nas. Mus., Bloemfontein 1, 149-183. Crompton, A. W. 1955b. On some Triassic cynodonts from Tanganyika. Proc. zool. Soc. Lond. 125, 617-6609. Crompton, A. W. 1958. The cranial morphology of a new genus and species of ictidosaurian. Proc. zool. Soc. Lond. 130, 183-216. Epmunp, A. G. 1957. On the special foramina in the jaws of many ornithischian dinosaurs. Contr. R. Ont. Mus. Zool. Palaeont. No. 48, 1-14. Epmunp, A. G. 1960. Tooth replacement phenomena in lower vertebrates. Contr. R. Ont. Mus., Life Sci. Div. No. 52, 1-190. Kermack, K. A. 1956. Tooth replacement in mammal-like reptiles of the suborders Gorgonopsia and Therocephalia. Phil. Trans. (B) 240, 95-133. Ktune, W. G. 1956. The Liassic therapsid OLt1cokyPHus. London: British Museum. ParRRINGTON, F. R. 1936. On tooth replacement in theriodont reptiles. Phil. Trans. (B) 226, 121-142. PATTERSON, BRYAN. 1956. Early Cretaceous mammals and the evolution of mammalian molar teeth. Fieldiana, Geol. 13, 1-105. Romer, A. S. & Price, L. W. 1940. Review of the Pelycosauria. Spec. Pop. geol. Soc. Amer. No. 28, i-x, 1-538. SEELEY, H. G. 1895. Researches on the structure, organization and classification of the fossil reptilia. Part ix. On the Gomphodontia. Phil. Trans. (B) 186, 1-57. Watson, D. M. S. 1914. Notes on some carnivorous therapsids. Proc. zool. Soc. Lond. 1914, 1021-1038. Watson, D. M. S. 1931. On the skeleton of a bauriamorph reptile. Proc. zool. Soc. Lond. 1931, 1163-1205. Watson, D. M. S. 1954. On Bolosaurus and the origin and classification of reptiles. Bull. Mus. comp. Zool. Harv. 111, 299-449. Key To LETTERING at. bo. Attachment bone. c. Canine. can. Canal in dentary. DEN. Dentary. Lt25 Functional 5th postcanine. lab. Labial. lab. c. Labial cusp. Jeena Nh oF Pit for lower canine. ling. Lingual. long. gr. Longitudinal groove. MAX. Maxilla. p. Pit for replacing tooth. PAL. Palatine. p. for. Palatal foramen. rem. f. 11 Remnant of 11th functional tooth. rep. t. Replacing tooth. rep. t. 11 11th replacing tooth. sec. p. Secondary palate. V: Vomer. w. fac. Wear-facet. w. fac. h. Horizontal wear-facet. w. fac. o. Oblique wear-facet. Ann. S. Afr. Mus., Vol. XLVI Plate XII A (top): Ericiolacerta parva. Left maxilla. B (bottom): Bauria cynops. Detail of replacement of 5th postcanine. fl ; ; a: i ; i = — ™ i i < 5 } Tie % ¥ = ~ z “— Le Aba ; 2 ink Sar ok SE he ia ee 1 r a ” ~ y Ann. S. Afr. Mus., Vol. XLVI Plate XIII A (top): Bauria cynops. Ventral view of palate. B (bottom): Bauria cynops. Right postcanine dentition. Ann. S. Afr. Mus., Vol. XLVI Plate XIV A (top): Bauria cynops. Crown view of left postcanine series. B (bottom): Bauria cynops. Detail of posterior three postcanines. References thus appear as follows: AUGENER, H. 1913. Polychaeta. In Michaelsen, W., ed. Beitrége zur Kenntnis der Meeresfauna Westafrikas. 2, 67-625. Hamburg: Friederichsen. EKMAN, 8. 1953. 71 15.7.47 7] Sh, S, M KNY 127 15.4.49 O-I M KNY 165 Q.7.50 littoral R, Buoy 264 L. Littoral material from East London, on the south coast. Date: July 1937. ANNALS Position: 33°1'S./27°54’E. LAM. Dredged in Lambert’s Bay, west coast. LAM 2 LAM 7 LAM 9 LAM 13 LAM 16 LAM a1 LAM 26 LAM 27 LAM 30 LAM 35 LAM 43 LAM 45 LAM 46 LAM 51 LAM 59 LAM 62 126 1asy 190 257 367 378 392 4.72 51 I 513 LIZ 7 LiZn3 LIZ 16-17 LIZ 27 MB. Dredged in Mossel “sf wy NM san NM w APLPPARRPRMAAALRLAA NA AA wn wo OF THE SOUTH AFRICAN MUSEUM & = S 3 aw ww nny o> NM [ar > Bottom f£S, M Sys bd R wooden piling S, Sh R R Bottom 3 iw) Wn DH WnNnnDMDWH v << N= w ) NM 2 bd 2 ~m = Date Position Depth (m.) 16.1.57 32°4:5'S./18°18-3’E. 17 18.1.57 32°5’S./18°17°9’E. 29 Dele ay 32°4°7'S./18°17°7' EB. 23 19.1.57 32°4/S./18°18-1E. 18 17.1.57 32°4:8'S./18°18-2’E. II 17.1.57 32°7-5/8./18°17-6'E. 20 18.1.57 32°4°9/S./18°17°5E. 27 16.1.57 32°4°1'S./18°18-4’E. 16 19.1.57 32°5°1'S./18°1 7-7 E 20 19.1.57 92°5:5/9./18°17°7' E 27°5 Q1.1.57 32°4:9'S5./18°18-2’E 13°5 21.1.57 32°5/S./18°18-2’E 8 CIEE | B9°APAS./ TO yey ku. 23 23.1.57 32°6°5 S./1O 17-7 E- 16°5 23.1.57 32°9’S./18°18’E. 16 23.1.57 32°1°5/S./18°18’E. 25 LB. Material from Langebaan Lagoon, on west coast. Date Position Depth (m.) 15.7.46 BQC7 hos) LO a1 ki. 4 16.7.46 33°6"3'9./16 3:3. 7 26.4.49 33°11°3/S./18°5°5/E. O-2 2.5.51 Bin 7 Teh Ike) Gh lee littoral 7.5.53 33°6/S./18°1°5’E. littoral 7.5.53 33°6/S./18°9’E. O-I 8.5.53 93°7-O1S.|1O 2° 1 Le. 2°5 6.5.55 33-7°4'S./18°2°5’E. 3-4 23.9.57 33°6/S./18°1'5/E. littoral 5.1.58 33°6-5/S./18°2’E. littoral LIZ. Dredged in Algoa Bay, Port Elizabeth, south coast. Date Position Depth (m.) 6.4.54 33°58°1’S./25°38-9’E. 9 6.4.54 33°58°2’S./25°38-8’E. 7-8 7-454 33°58°4'S./25°40°5’E. 14 11.4.54 34.°0°8'S./25°4.2°4'E. 5-7 Bay, south coast. Date Position Depth (m.) 12.1.56 34.°8:0'S./22°8°5’E. QI 12.1.56 BY PGy STIG ECY 1B 19 13.10.56 BAL l= 95/225 LOntgis. 16 13.1.56 BA iE (94) 2270;0781. 19 13.1.56 34°11.1'S./22°10°1'E. QI 15.1.56 34.°8°3/S./22°9°4’E. 31 16.1.56 34°9°3/S./22°10:0'E. 31 16.1.56 34°10°1'S./22°8-0’E. 9 16.1.56 34°8-5/S./22°8-8’E. 25 17.1.56 34°11°3/9./22°10°:0 EB. 10 17.1.56 34°11:0'S./22°9-9'E. 14 17.1.50 34°10°7'S./22°9'6’E. 9 18.1.56 34°4°3/S./22°13°5’E. 12°5 18.1.56 34.°4:1'S./22°13-9'E. 115 18.1.56 34.°4°3/S./22°14°2’E. 17-20 19.1.56 34°8-9’S./22°7-0’E. 18 19.1.56 84 °O°1'S./22° 7°21. 12 20.1.56 34°5:0'S./22°11°8’E. 19 20.1.56 34.°6-2’S./22°10°9’E. 27°5 HYDROZOA OF SOUTH AND WEST COASTS BAett-619./22> 10-2) 1. 29 9A°T1-4/9./22°10°1 BE. 29 MB 83 21.1.56 MB 84 21.1.56 MB 88 18.1.56 34.°4°8/S./22°13°1’E. 26 OF SOUTH AFRICA S, Sh, R R co S, Sh, R 265 N. Littoral material from Port Nolloth, west coast. Date: 11.11.35. Position: ZOO.) TO 2ues NAD. Dredged off Natal, east coast. NAD 1 NAD 9 NAD a1, 22 Date 17.5.50 23.4.50 12.8.58 29°46’S./31°17’E. Position Depth (m.) Bottom 30°47°1'S./30°29°1’E. 44. St 110-130 49 29°58’S./31°2’E. PP. Littoral material from Paternoster, west coast. Date: 24.9.57. Position: 32°43'5./17°55 E. S. Littoral material from Still Bay, south coast. Date: January, 1932. Position: BAe 2 038/21. 20) Er. SAMH. Material dredged by the s.s. Pieter Faure and lodged in the South African Museum. (The positions were given in the original records as compass bearings off salient points on the coast, and were probably not very accurate. These have been converted into latitude and longitude and given to the nearest minute.) Date Position Depth (m.) Bottom SAMH 148-153 15.7.1898 34°8’S./22°16’E. St SAMH 163 28.10.1898 33°54'S./25°47'E. 40 d 8 SAMH 165-169 11.11.1898 33°4.9'S./25°56’E. SAMH 190-194 3=.28.12.1898 33°9/S./28°3’E. 86 S, Sh, R SAMH 195, 196 15.2.1899 33°53'S./25°45 E. 33-42 SAMH 197-201 7.3.1899 33°59/S./25°51 EE. 24-27 SAMH 210 15.3.1899 33°4'7'S./26°19'E. 18-29 8S, Sh, St SAMH 212 24.3.1899 33°50'S./26°35’E. gI M SAMH 215-225 19.6.1899 34°26'S./21°42’E. £S SAMH 238-241 ~—=11.10.1900 34°8/’S./22°59°5'E. 73 S, Sh, Crl SAMH 248 24.10.1900 94°29°5./22°01E:: ai R SAMH 249 15.7.19O1 33°13°5/S./27°58’E. 89 brk Sh SAMH 254-277 17.7.1901 9307S (27° 47°5 E. {S$ SAMH 278-280 25.7-1901 32°50'S./28°18°5’E. 86 brk Sh SAMH 284 6.8.1901 33°5'S./27°50’E. 7 R, brk Sh SAMH 285-203 13.8.1901 32°45/S./28°26’E. 53 brk Sh, St SAMH 297-208 13.8.1901 32°4.7'S./28°28’E. 82 brk Sh SAMH 299-304 15.80.1901 32°42’S./28°26’E. 31 R SAMH 305-309 10.9.1901 33°54’S./26°51’E. 120 brk Sh, St SAMH 319, 320 23.9.1901 34°5'S./26°34’E. 115 S, Sh, bk Spks SAMH 322-333 22.9.1904 34°12’S./22°15°5’E. 51 £S SAMH 344-347 19.10.1904 34°15'5/9./22°14’E. 64 M SAMH 349 15.8.1905 34.°8’S./23°32’E. 73 M SAMH 353 28.8.1906 33°54°5'9./26°28’E. 104 M SAMH 355 —.1.1913 29°55/S./31°14’E. littoral SAMH 360 25.7.1901 32°50’S./28°18-5/E. 86 brk Sh SB. Saldanha Bay, west coast. Date Position Depth (m.) Bottom SB 141 2.5.57 33°2°5'S./18°2’E. littoral R SB 150 22.09.57 33°2°5S./18°2’E. littoral R SB 168 25.9.57 Coie Dixie fl eine a oe littoral R SB 178 28.4.59 33°3°6/S./18°0-4’E. 15 Sh, kh 8, R SB 194 1.5.59 33°3°5'S./17°59°2’E. 20 Rea SB 196 1.5.59 33°4°4S./17°56-4E. 35, 266 ANNALS OF THE SOUTH AFRICAN MUSEUM SCD. Dredged off the south coast. SCD 5 SCD 20 SCD 22 SCD 26 SCD 29 SCD 33 SCD 36, 37 SCD 50 SCD 52 SCD 61 SCD 67 SCD 69 SCD 75 SCD 79-81 SCD 82 SCD 84, 85 SCD 94 SCD 96 SCD 103 SCD 104 SCD 108 SCD 117 SCD 122 SCD 126 SCD 138 SCD 141 SCD 152 SCD 153 SCD 169 SCD 175 SCD 186-8 SCD to1 SCD 195 SCD 199 SCD 201 SCD 206 SCD 219 SCD 234 SCD 239 SWA. Littoral material from rocky shore at Luderitz, South West Africa. Date 19.4.58 26.5.58 26.5.58 23.5.50 22.6.58 21.5.50 19.5.58 18.5.58 20.8.58 15.08.58 4-7-59 5:7-59 16.7.59 16.7.59 17-7-59 17-7-59 20.7.59 20.7.59 22.7.59 23-7-59 23-7-59 14.2.60 14.2.60 3.6.60 28.6.60 28.8.60 2.6.60 25.11.60 24.11.60 30.11.60 30.11.60 29.11.60 29.11.60 30.11.60 29.11.60 30.11.60 29.11.60 30.11.60 29.11.60 Position 34°15'S./25°5'E. 34°9°3'S./23°23°8'E. 34°26-7'S./23°26-0’E. 33°47'5./26°4’E. 33°38-6’S./26°54°7’E. 35°3'9./27°56:2'E. 32°15:2'S./28°57°7'E. 31°38:8'S./29°34°4’E. 34° 1'S./25°45°5 E. 32°17°7'S./28°54°5'E. 34°20'S./24°22’E. 33°31'S./27°14°5E. 32°39'S./28°38’E. 32°43'S./28°28’E. Bo 82709. /2q 540 gts. 33°3'9./27°55 E. 33°55°5 9-/25°51'E. 34°21'S./25°41'E. 35-7'9./22°15/E. 34°26°5/S./21°48’E. 34°35'S./21°56’E. 34°46’S8./22°5/E. 34°55'S./21°26’E. 34°3’S./25°59'E. 33°58°9/S./25°41°4’E. 34°20'S./23°31’E. 34°10'S./23°32 EB. 34°4°3'S./23°25°8'E. 34°7°5'S./23°31°7'E. 34°10'S./23°32’E. 34°5'8'S./23°23°2’E. 34°51’S./23°41'E. 34°2:0'5./23°28-4’E. 34°51'S./23°41E. 34°2:0’8./23°28-4’E. Date: —.7.57. Position: 26°33'S./15°9’E. TB. Material dredged from Table Bay. TRA 35 TRA 36 TRA 38 TRA 42 Depth (m.) 49 Date Position Depth (m.) 11.2.47 33°47°5'S./18°24°3’E. 19-20 25.10.46 33°52°5'9./18°26’E. 22 4.8.46 33°49°5'9./18°27°5/E. 9-18 11.2.4.7 33°48'°3'S./18°24’E. 10-12°5 Tle 33°4.7°5 9./18°24°3’E. 19-20 26.6.4.7 33°52°7'9./18°28-7’E. 9 BOTA 33°52°7'9./18°26°8’E. 20°5 15.12.57 33°48-6/S./18°24-6’E. 15 TRA. Material collected by commercial trawlers. Date Position Depth (m.) 21.1.50 34°34’S./20°50’E. 70 21.1.50 34°34’S./20°50’E. 73 —.7.50 34°30'S./20°56’E. 73 —.7.51 34.°30’S./20°55’E. 70 24.9.52 31°50’S./16°50’E. 330 TRA 47 Botiom R Sh M, Sh R S, Sh R R R 2S 2S Ss, M St, Sh br S, Sh R bk M, S Sh S co S, brk Sh co S, Sh, St R kh § Bottom Sh. ot iy DAD DNA RNANNN GP jor Ge fer DN ct DN St ior a) ~ DN Se DN HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 267 TRA 56 28.11.52 34°40’S./21°35’E. 70-77) Ree S TRA 86 23.3.53 32°41°7'S./17°58°5 E. 9 ) TRA 92 —.1.54 35,°3’8./21°50'E. 110 Ss TRA 99 18.1.56 34°25°5'S./21°50°2’E. 60 S TRA 134 23.2.57 34°19'S./18°30’E. 52 S, Sh, R TRA 151 6.3.58 34°51'S./19°55E. 22 R TRA 156 15.10.58 BA 1249. / lo. 220E, 18-27 Cable TRA 159 6.7.58 33°560'S./25°36’E. Turtle’s back WCD. Dredged off west coast. Date Position Depth (m.) Bottom WCD 7 24.359 SABO P3095) LO ale]: by bu 43 R WCD i2 24.3.59 34°9°4'S./18°16°5’E. 75 R WCD 18 ~—_.29..4.59 33°5°618./17.54°5 E. 64. kh M WCD 20 30.4.59 eGR ts) iyo ay ae 86 R WCD 32 15.12.59 34°10'S/18°16’E. 03 R WCD 34 ~—:115.12.59 34°11-2’S./18°20°2’E. 27 R WCD 56 21.9.60 32°4:6'S./18°18’E. 18 Saple WCD 81 15.9.49 34°5/S./18°21’E. II N) REMARKS ON CLASSIFICATION In the taxonomy of the Plumulariidae three problems are of recurrent interest, namely: (1) The validity of fixed or movable nematothecae as a basis for subdivision; (2) the presence or absence of cauline hydrothecae as a useful diagnostic character, and (3) the value of the method of branching in the limitation and grouping of genera. In 1883, Allman, using the type of nematotheca as a basis, divided the Plumulariidae into two large sections: the Eleutheroplea and the Statoplea. However, the presence of certain genera such as Heteroplon Allman 1883 and Heterotheca Stechow 1921, with two kinds of nematothecae, made this simple division impracticable. In 1923b Stechow made use of 4 sub-families, but also used as his basis the type of nematotheca. Stechow’s subfamilies included: Kirchenpaueriinae Stechow 1921. Nematothecae reduced. 2. Plumulariinae Kiihn 1913 (= Eleutheroplea). Nematothecae 2-chambered and movable. Acladiinae Stechow 1923b. Intermediate forms. 4. Aglaopheniinae Stechow 1911 (= Statoplea). Nematothecae 1-chambered and immovable. eS The Kirchenpaueriinae and the Aglaopheniinae appear to be valid sub- families representing natural assemblies of genera (in fact the latter is raised to family rank by some systematists), but the other subfamilies need further consideration. The Acladiinae was created for intermediate forms with fixed mesial nema- tothecae and fixed or movable lateral nematothecae. It was an unsatisfactory group, as Stechow himself realized, for he abandoned it in 1925, adding its genera to the Plumulariinae. It has in fact been increasingly recognized that 268 ANNALS OF THE SOUTH AFRICAN MUSEUM in this group the type of nematotheca does not form a good basis for family, or even generic, diagnosis. ‘This was pointed out as early as 1913 by Billard and in 1915 by Bale, and most systematists have dropped the genera Heteroplon, Heterotheca and Antennellopsis Jiderholm 1896 for this reason. There is, however, a group of genera which are linked by the presence of cauline hydrothecae, including Thecocaulus Bale 1915, Halopteris Allman 1877, Gattya Allman 1886 and Schizotricha Allman 1883, genera which Stechow dis- tributed between the Plumulariinae and Acladiinae. Bale was the first to realize the importance of cauline hydrothecae as a diagnostic character when he established the genus Thecocaulus in 1915 for ‘forms in which hydro- thecae are borne on the rachis as well as on the pinnae’. Such genera possess stem internodes which are essentially similar to those of the hydrocladia, and thus differ from the ‘Plumularia’ type where the stem internodes lack hydro- thecae and thus differ from those of the hydrocladia. This difference is a funda- mental one exhibited in the earliest stages of growth, influencing, as it does, the nature of the first upright stem. It is a sound character for generic diagnosis, and under no circumstances should species with cauline hydrothecae be included in the same genus as species without, as has been done by Billard (1913) (i.e. in the genus Plumularia). It is also considered to be a sound character on which to group genera, and a new subfamily, the Halopterinae, is proposed for forms with cauline hydrothecae. (Stechow’s Acladiinae cannot be retained since Acladia is a synonym for Halopteris.) The Plumulariinae can be retained in Stechow’s sense, though excluding certain genera dealt with below. Stechow implies that the Kirchenpauerinae is the most primitive subfamily of the Plumulariidae because of the poorly developed nematothecae, but it is maintained here that the arrangement of hydrothecae and hydranths is of far greater fundamental significance. It is rational to assume that a primitive form would be one in which a functional feeding unit (i.e. a portion of stem and a hydranth) is established as rapidly as possible in early development and before budding commences, as in fact occurs in the sympodial growth of primitive Calyptoblast families such as the ‘Campanulinidae’ and Haleciidae. This arrangement is possible in the Halopterinae but not in the other Plumu- lariidae where a stem must develop hydrocladia before the formation of hydranths is possible. The Halopterinae is thus considered to contain the primitive stock of the Plumulariidae. Genera of Halopterinae The genus Halopteris was established by Allman in 1877 for H. carinata, which is the type species. Totton in 1930 showed that Plumularia catherina Johnston 1833 (the type species for Thecocaulus) is congeneric, and thus sinks Thecocaulus in the synonymy of Halopteris. Although Allman’s conception of Halopteris (and also Stechow’s) was the presence of fixed nematothecae, its main diagnostic characters are now con- sidered to be the presence of cauline hydrothecae, a pinnate stem with HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 269 unbranched hydrocladia and an untoothed thecal margin. The genus as thus envisaged includes the following species: Plumularia alternata Nutting 1900, P. buski Bale 1884, P. campanula Busk 1852, Halopteris carinata Allman 1877, Plumularia catherina Johnston 1833, P. concava Billard 1911, Halopteris constricta Totton 1930, Plumularia crassa Billard 1911, P. diaphana Heller 1868, Halopteris gemellipara n. sp., Aglaophenia glutinosa Lamouroux 1816, Thecocaulus heterogona Bale 1924, Plumularia liechtensternit Marktanner-Turneretscher 1890, Thecocaulus minutus Trebilcock 1928, Plumu- laria polymorpha Billard 1913, P. sulcata Lamarck 1816, P. tuba Kirchenpauer 1876, Heteroplon valdiviae Stechow 1923, Plumularia zygocladia Bale 1914b, and possibly also Plumularia diaphragmatica Billard 1911 and P. jedani Billard 1Q13. In addition to the characters mentioned above, these species usually have other features in common. Thus, the colonies as a whole tend to resemble Aglaophenia in their general appearance rather than Plumularia, with robust stems and close-set hydrocladia and hydrothecae. Most of the species possess hinge-joints near the base of the stem or its branches. There is a tendency towards the production of opposite hydrocladia—in H. gemellipara, H. catherina and H. zygocladia the hydrocladia are opposite throughout all or most of the length, and in practically all other species opposite hydrocladia are borne by the first or second thecate internodes. Internodal and intrathecal septa are typically absent. Each hydrocladial internode typically has one median inferior nematotheca and one or two pairs of laterals, and sometimes one or more supracalycines. The cauline internodes have a similar arrangement, though they may possess supernumerary supracalycines as well. The nematothecae may be I- or 2-chambered, movable or immovable. The gonothecae, where known, are dimorphic, typically with the female seated on the stem and the male on the hydrocladia of the same colony, and they bear nematothecae, at least in the female. It is not suggested that all these characters are essential for the inclusion of a species in Halopteris (in fact there are exceptions as regards most of the characters listed), but that the common possession of most of them links together a group of species which are closely related and can be considered as constituting a single genus. Few hard and fast rules can be laid down, as many of the species appear to be in a state of flux and show surprising variability of structure, even within a single colony. It has already been demonstrated how the segmentation of the hydrocladia, and the structure and arrangement of nematothecae, may vary in H. glutinosa (see Millard, 1958, and also Broch, 1933). It is proposed to include in the same subfamily (i.e. Halopterinae) the genera Antennella, Schizotricha, and Monostaechas, which are clearly related to Halopteris, and differ only in their method of branching. 270 ANNALS OF THE SOUTH AFRICAN MUSEUM The genus Antennella Allman 1877 was created for A. gracilis Allman 1877, a form with simple stems similar to hydrocladia, and this is the type species. Schizotricha Allman 1883 was created for forms with pinnate stems and branch- ing hydrocladia, and the type species is S. unifurcata Allman 1883 according to Totton 1930. Finally Monostaechas Allman 1877 was created for M. dichotoma Allman 1877 (= M. quadridens (McCrady) 1857) in which the hydrocladia branch sympodially from the posterior surface, the resulting ‘stem’ being formed from the proximal parts of successive hydrocladia (cf. also Billard, 1913, p. 14). The type of branching, however, forms no inflexible demarcation between the four genera so far mentioned. It is known, for instance, that many species of Halopteris may exist in a simple form, and some species of Antennella may produce pinnate stems. Thus Broch (1933) has shown that Antennella secundaria can produce pinnate stems although the simple form is the more common, and that Halopteris diaphana can exist in the simple form although the pinnate form is more common. H. campanula, H. catherina and H. constricta can all exist in the simple form (Bale, 1913, Hincks, 1868, and Millard, 1957, respectively). Accordingly Broch has sunk Thecocaulus (= Halopteris) in the synonymy of Antennella, and this practice has been followed by most Continental syste- matists. But it has been shown recently (Millard, 1958) that Monostaechas natalensis, which has a completely different method of branching, can also exist in the simple form, and Billard 1913 has shown that A. secundaria can branch in the Monostaechas way. To follow Broch’s system we should then sink Monostaechas also in the synonymy of Antennella. It appears to the present author that within this group of species the simple form is the most primitive type and that from it have led two main evolutionary lines, one leading to Halopteris with the tendency towards pinnate branching, and one to Monostaechas with branches arising from the posterior surface of the stem, and that since these genera represent grades of evolution it is wise to keep them separate. The placing of intermediate types will always be a matter of opinion, though in general a species should be named according to its dominant form, thus Antennella secundaria but Halopteris diaphana. Passing on to forms with branching hydrocladia, these surely represent a further stage of the evolutionary line leading through Halopteris, and if we are to retain Halopteris as separate from Antennella then we must also retain Schizotricha, in spite of intermediate forms. Another genus to be included in the Halopterinae is Gattya Allman 1886 (= Paragattya Warren 1908, see p. 281), which was created for Gattya humilis Allman 1886, a form with a toothed thecal margin. Whether a toothed hydro- theca is a sufficient character to distinguish a genus is a matter of opinion. In most of their characters the species are closely related to the Antennella- Halopteris line (for instance, G. humilis possesses hinge-joints and occasionally branching hydrocladia, and the first thecate internode of the stem bears a pair of opposite hydrocladia). Possibly they represent a deviation from the Halopteris stock, which is already foreshadowed in H. carinata and H. valdiviae which have HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 271 a sinuated hydrothecal margin. At any rate the character is a definite one and useful in distinguishing species, and the genus may be allowed to stand. The genera NEMERTESIA and ANTENNOPSIS The genus Antennopsis was founded by Allman in 1877 for A. hippuris, the features distinguishing it from Nemertesia Lamouroux 1812 being the ‘scattered disposition’ of the hydrocladia, and the uncanaliculated coenosarc. The former character has generally been accepted as worthless, and the latter character has also been discredited by many authors (see Bedot, 1921). The SCHIZOTRICKHA SA GATTYA HALOPTERIS MONOSTAECHAS =e Lat: ANTENNELLA Fic. 1. Diagram showing possible evolutionary trends in the Halopterinae. 25 7/2 ANNALS OF THE SOUTH AFRICAN MUSEUM author considers that the type species of Antennopsis (A. hippurts Allman 1877) is congeneric with the type species of Nemertesia (Sertularia antennina Linnaeus 1758) and consequently sinks Antennopsis in the synonymy of Nemertesia. There are, however, certain species (Antennularia fascicularis Allman 1883 and Antennopsis scotiae Ritchie 1907) included by Bedot in Antennopsis which differ markedly from other species of the composite genus Nemertesia and from the type species of this genus. In Nemertesia the hydrocladia arise from an unfas- cicled stem, or from the central tube of a fascicled stem, either irregularly or in verticils. In the two species mentioned the hydrocladia arise, quite irregularly, from the peripheral tubes of the stem, and there is no recognizable central or main tube. It is just as though the hydrorhizal tubes of a colony of Antennella had become bound together to form a fascicled stem, each branching or separating from the others independently to give hydrocladia. This character is, I feel, of generic value, and I propose a new genus Corhiza for the reception of the two species mentioned and two new species described below, with Antennopsis scotiae Ritchie 1907 as the type species (i.e. Corhiza scotiae (Ritchie) 1907). | These four species are also linked by certain less important characters which do not occur in Nemertesia and which emphasize the cleavage between - the two genera, namely the large hydrothecae, the absence of a mamelon on the base of the hydrocladium, the position of the lateral nematothecae on long supporting processes at the sides of the hydrotheca, and the gonothecae which are borne on the hydrocladia and themselves bear nematothecae. Further, the author feels that the sum total of the characters of the genus Corhiza indicates a closer relationship with the Halopterinae than with the Plumulariinae, and proposes to place it in the former subfamily. The stem of course bears no cauline hydrothecae, but if the stem is considered as a con- glomeration of hydrorhizal tubes this becomes understandable. The genus Corhiza appears to be closely related to Antennella and Mono- staechas. In A. africana the hydrorhizal tubes often rise up from the substratum in bunches, giving rise to clumps of simple stems (i.e. hydrocladia). A further development of this process would produce a colony of the Corhiza type. A fascicled stem similar to that of Corhiza is found in some members of Mono- staechas (M. natalensis Millard 1958 and WM. faurei Millard 1958), and MM. natalensis appears to occupy an intermediate position between the two genera, where the branching of the hydrocladia has not developed to the same extent as in other species of the genus. The evolutionary relationships between the genera of Halopterinae as visualized by the author are illustrated in figure 1. The subfamilies of the Plumulariidae are summarized as follows. Subfamily 1. Halopterinae nov. subfam. Hydrocladia arising independently from hydrorhiza, from a pinnate stem, or from the superficial tubes of a HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 273 compounded stem. Pinnate stems bearing cauline hydrothecae. Paired lateral nematothecae present, of variable structure. South African genera: Antennella Allman 1877 Corhiza nov. gen. Gattya Allman 1886 Halopteris Allman 1877 Monostaechas Allman 1877 Schizotricha Allman 1883 Subfamily 2. Kirchenpauerinae Stechow 1921. Stem simple or pinnate, the latter without cauline hydrothecae. No paired lateral nematothecae. Median nematothecae poorly developed, seldom 2-chambered, often rudimentary and sometimes represented by naked sarcostyles only. South African genera: Airchenpaueria Jickeli 1883 Oswaldella Stechow 1919 Pycnotheca Stechow 1919 Subfamily 3. Plumulariinae Kihn 1913. Stem upright, giving off hydro- cladia pinnately or in verticils. No cauline hydrothecae. Paired lateral or supracalycine nematothecae present. | Nematothecae 2-chambered, movable, free from hydrotheca. South African genera: WNemertesza Lamouroux 1812 Plumularia Lamarck 1816 Subfamily 4. Aglaopheniinae Stechow 1911. Stem upright and pinnate. No cauline hydrothecae. Hydrocladia with close-set hydrothecae generally facing towards the anterior surface. Paired lateral nematothecae normally present. Nematothecae 1-chambered and immovable, usually adnate to hydrotheca. South African genera: Aglaophenia Lamouroux 1812 Cladocarpus Allman 1874 Halicornaria Allman 1874 Lytocarpus Allman 1883 Thecocarpus Nutting 1900 List OF SPECIES Subfamily 1 Halopterinae Antennella africana Broch 1914 Halopteris gemellipara n. sp. Antennella secundaria (Gmelin) 1788-1793 Halopteris glutinosa (Lamx.) 1816 Corhiza bellicosa n. sp. Halopteris tuba (Kirch.) 1876 Corhiza pannosa n. sp. Halopteris valdiviae (Stechow) 1923 Corhiza scotiae (Ritchie) 1907 Monostaechas natalensis Millard 1958 Gattya humilis Allman 1886 Schizotricha simplex Warren 1914 Halopteris constricta Totton 1930 274 ANNALS OF THE SOUTH AFRICAN MUSEUM Subfamily 2 Kirchenpaueriinae Kirchenpaueria pinnata (Linn.) 1758 Oswaldella nova (Jarvis) 1922 Kirchenpaueria triangulata (Totton) 1930 Subfamily 3 Plumulariinae Nemertesta ciliata Bale 1914b Plumularia pulchella Bale 1882 Nemertesia cymodocea (Busk) 1851 Plumularia setacea (Ell. & Sol.) 1755 Nemertesia ramosa Lamx. 1816 Plumularia spinulosa Bale 1882 Plumularia filicaulis Kirch. 1876 Plumularia wasini Jarvis 1922 Plumularia lagenifera Allman 1886 Subfamily 4 Aglaopheniinae Aglaophenia late-carinata Allman 1877 Halicornaria exserta n. sp. Aglaophenia pluma (Linn.) 1758 Halicornaria ferlusi Billard 1901 Cladocarpus leloupi n. nom. Lytocarpus filamentosus (Lam.) 1816 Cladocarpus lignosus (Kirch.) 1872 Thecocarpus flexuosus (Lamx.) 1816 Halicornaria arcuata (Lamx.) 1816 Thecocarpus formosus (Busk) 1851 SYSTEMATIC SECTION Subfamily Halopterinae Antennella africana Broch, 1914 Antennella africana. Millard, 1957, p. 226. Records. West coast: TB 1A, 17B. TRA 86N, 156E. LAM 30], 35D. SB 196L. CP 650F. WCD 56K. South coast: AFR 994L. TRA 38N. MB 8R (pp), 12X, 47S, 52E, 58D, 59E, 60K. LIZ 7V. SCD 5J, 22B, 36U, 81Q, 96G, 152Y, 153E, 175F. SAMH 241, 279; 293, 309. Remarks. This material includes one specimen (SAMH 309) in which all the dimensions are about double the normal. The species is common on the west and south coasts of South Africa (from Luderitz Bay on the west (Broch, 1914) to approximately 29°E. on the south) and occurs in the intertidal region down to a depth of 120 m. Antennella secundaria (Gmelin), 1788-1793 Antennella secundaria. Millard, 1958, p. 199. Records. South coast: MB 8R (pp), 47S. SAMH 153, 165, 191, 212, 277, 360. LIZ 16G, 27K. Description. Colonies reaching a height of 1:2 cm. Stems and hydrothecae sturdy and of greater dimensions than the material from Inhaca (Millard, 1958). Athecate internodes comparatively short and only rarely exceeding the neigh- bouring thecate internodes in length, bearing one or two nematothecae. HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 275 Measurements (mm.) * Inhaca Mossel Bay Hydrocladium, length of basal part 0°99-2°75 1°73-3-27 thecate internode, length 0-26-0°34. 0°32-0°43 | athecate internode, length 0:26-0:38 0°22-0°45 athecate internode, maximum diameter 0-04—0°05 0:08—0-12 Hydrotheca, length abcauline O-15-0'19 0°20—0:225 diameter at margin O-15-0'19 0°27—0°32 ‘Gonotheca, male, length (without Dedicely 0°32—0°40 maximum diameter te 0°22—0°29 Gonotheca, female, length Guithout Sedicel) 0-61—0-66 maximum, diameter 0°41—-0°48 Remarks. In poor material it is difficult to distinguish A. secundaria from A. africana. ‘The number of nematothecae on the intermediate internodes and the proportionate length of the latter are no criterion, as both are variable charac- ters. The only sure criterion is the presence of supplementary lateral nematothe- cae in A. africana and of a median supracalycine nematotheca in A. secundaria, and these are often lost. There appears to be no difference in the structure of the gonophores of the two species. In South Africa both species are known only in the simple form. A. secun- daria occasionally gives off one or two branches, but only from the back of the basal part of the stem, as mentioned also by Billard (1913, p. 8). True pinnate branching never occurs. A. secundaria occurs on the south and east coasts of South Africa (from approximately 19°E. on the south coast (Stechow, 1925) to Inhaca on the east coast), in the littoral region, and down to a depth of 100 m. (Stechow, 1925). Genus CorHIZA nov. gen. Type species: Antennopsis scotiae Ritchie, 1907 Diagnosis. Plumulariidae with an upright and fascicled stem composed of a number of interwoven and intercommunicating tubes of equal diameter and importance. Hydrocladia arising from the component tubes in a completely irregular fashion, not rebranching. Hydrotheca with untoothed margin. Nematothecae (so far as is known) all bithalamic and movable, including paired laterals. Gonothecae borne on the hydrocladia, bearing nematothecae. Corhiza bellicosa nov. sp. (Fig. 2 A—E) Types and records. Holotype: SCD 84S (South African Museum registered num- ber = SAMH 364). Paratype: SCD 153J. (Both from south coast.) * Including material described by Millard, 1958. 276 ANNALS OF THE SOUTH AFRICAN MUSEUM Description of holotype. Several short and thick fascicled stems reaching a maxi- mum height of 5:0 cm., a few of which branch irregularly, giving rise to numer- ous hydrocladia in an irregular fashion from the component tubes in the distal region. The individual tubes of the stem are connected to one another by communicating pores and branch irregularly. In the distal region the tubes usually separate from one another in clumps (the branches), and each one terminates in a hydrocladium. The hydrocladia of a clump or branch tend to face in the same direction, giving a ‘Monostaechas’ appearance to the colony. Hydrocladia unbranched, with proximal region athecate, divided by an irregular number of transverse internodes and bearing a double series of nematothecae. Remainder divided into thecate internodes by oblique nodes. Hydrotheca borne on a projection near base of internode, with depth exceeding diameter, swollen in basal half, then narrowed slightly and expanding again to margin. Margin smooth, at right angles to internode. Mostof adcauline wall adnate, but a small distal portion free. Nematothecae generally 13 to an internode, including: 1 median inferior borne close to proximal end of internode and not reaching base of hydrotheca; 1 pair lateral inferior borne on base of internode slightly above level of median inferior and just reaching base of hydrotheca; 1 pair lateral borne on long processes arising at the side of the hydrotheca and not quite reaching thecal margin; 1 pair lateral borne on outer surfaces of bases of above processes; 1 pair lateral borne at the side of the hydrotheca about midway between above- mentioned process and thecal margin, overtopping hydrotheca; generally 2 pairs supracalycine borne above thecal margin (1 pair close to the midline and another pair further to the sides and at a slightly higher level). There is, however, much variation in the number and position of the supracalycine nematothecae, and as few as 2 and as many as 9 have been observed, generally in pairs but often quite irregularly arranged. All nematothecae bithalamic and movable. Nematotheca large, expanding towards margin, with sides of distal chamber cut right away to level of dia- phragm resulting in a margin of 2 flaps—a high abcauline (i.e. outer) one, and a lower adcauline (i.e. inner) one. Gonophores absent. > A-E. Corhiza bellicosa n. sp. A. Two stems from the holotype, SCD 84S. There has probably been some distortion of the hydrocladia during preservation. B-D. Two lateral views and an anterior view of portions of hydrocladia, from the holotype. B is taken from near the proximal end, and C from near the distal end, of hydrocladia. E. Different views of the lateral nematothecae, the top one from SCD 153J, the other two from the holotype. F. Monostaechas natalensis Millard. Different views of the nematothecae from SCD 153D (large form). HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 7] Fic. 2. Corhiza bellicosa and Monostaechas natalensis. 278 ANNALS OF THE SOUTH AFRICAN MUSEUM Measurements (mm.) SCD 848 SCD 153] (Holotype) Internode length (on posterior surface) 0°4.9—0°94. 0°55-0°80 diameter (above hydrotheca) 0-13-0'18 0-14—0°20 Hydrotheca, height .. 0°22—0'27 0°23-0°31 diameter at margin O°1'7—0°22 0°16—-0:24, diameter /height 0:67-0:85 0-70—-0'84. Nematotheca, height 0:09—-0°15 O-11-0-16 Remarks. This species is very close to Monostaechas faure: Millard 1958, differing from it in the unbranched and longer hydrocladia, in the presence of 2 extra pairs of nematothecae to each internode, in the greater measurements of all individual parts, and in the shape of the hydrotheca where the proportion of diameter/height is greater. ‘The shape of the nematotheca is very characteristic of the species. Corhiza pannosa nov. sp. (Fig. 3 A-B, D-G) Types and records. Holotype: SCD 5H (South African Museum registered num- ber = SAMH 362). Other records: SCD 61H, 67M, 186%. SAMH 292. WCD 12G. Description of holotype. A bushy colony reaching a height of 5:3 cm. Hydrorhiza a matted reticulum. Stem strongly fascicled, branching irregularly, and giving off hydrocladia from its component tubes in a completely irregular fashion. Tubes of stem not canaliculated. Hydrocladia borne on short apophyses of stem, reaching a maximum length of 2 cm. and bearing up to 20 hydrothecae. Basal 4 or 5 mm. devoid of hydrothecae, but divided by transverse nodes into 2 or more rather long inter- nodes bearing a double series of nematothecae. Remainder of hydrocladium divided into thecate internodes by oblique nodes, though in some parts the distal end of an internode may be cut off by a transverse node. Hydrotheca large, with almost parallel sides, standing away from hydro- cladium at an angle of about 50°, with about half adcauline wall adnate to internode. Nematothecae generally 7 to each internode, namely: 1 medial inferior; 2 pairs laterals (1 seated on a process arising next to the hydrotheca and not > A-B, D-G. Corhiza pannosa n. sp. A, the whole colony, and B, the lateral and supplementary lateral nematothecae in adcauline view, from the holotype SCD 5H. D-G, portions of hydrocladia from different samples to show variation. (D, the holotype. E, WCD 12G. F, SAMH 292. G, SCD 61H). F shows a gonotheca. C. Corhiza scotiae (Ritchie). The lateral and supplementary lateral nematothecae from TRA 35B as a comparison with B. HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 279 Fic. 3. Corhiza spp. 280 ANNALS OF THE SOUTH AFRICAN MUSEUM quite reaching margin, and 1 seated in the angle between this process and the internode) ; 2 (or occasionally 1 or 3) median superior. Nematothecae all bithalamic and movable. Lateral nematotheca large and of characteristic shape: terminal chamber cup-shaped and deep, and with nearly parallel sides which turn in slightly at margin, with a distinct, deep excavation in adcauline (i.e. mesial) wall. Median nematotheca smaller, with very short adcauline wall. Gonophores absent. Measurements (mm.) SCD 61H SAMH 292 SCD186X WCD12G SCD5H (holotype) Internode length eons distal end when cut off). . . 0°62-0:88 0:62-0:68 0:77-0:95 0:80-1:09 0°70—-1°17 diameter, maximum .. O'0Q-O°12 O*1I—O°13 009-0714 O°13-O0°17 O°18—0:23 Hydrotheca, height abcauline . 0°26-0°38 0:34-0:38 0°35-0°46 0:30-0:36 0-34-0°46 height adcauline, free pans . O°16—0°245 0:20-0:24. 0°20—-0°:22 0°14-0°20 0-16—-0°25 diameter at margin .. 0°32-0°34 0°345-0°39 0°31-0°43 0:36-0:39 0°37-0-41 free part/abcauline height | . 0°58-0°71 0°54-0°66 0°50-0°57 0°42-0°58 0°41-0°555 diameter/abcauline height . 0°84-1°31 O°QQ-1'I5 0°72—-1:06 1:00-1:23 0°87—1-09 Nematotheca height, lateral . O1O-O'II OtII-O'I14 O'II-O°13 OTI-O'16 O°15—O°19 supplementary lateral . 0°03-0°08 0:045-0:06 0:07-0:09 0°07—0°12 0*09-0°15 Gonotheca, height 0:80 0:96-1:07 breadth 0°33 0:36-0:59 Remarks. In this species the form of the colony resembles closely that of Corhiza fascicularis (Allman) 1883 from Tristan da Cunha, the type material of which apparently no longer exists. It can be distinguished from other species of the genus in particular by its distinctive lateral nematothecae, which resemble most those found in Halopteris glutinosa (cf. Millard, 1958, figure 10). ‘The details of the hydrocladia resemble those of Monostaechas natalensis Millard 1958, from which it is distinguished by its larger hydrothecae which are less adnate, and by the nematothecae. Remarks on ‘other records’. 1 have included with this species one stem fragment bearing hydrocladia (SCD 67M) exactly like the holotype, and 4 samples of solitary hydrocladia which appear to represent simple forms of the same species. The latter agree in all important features and particularly in the shape of the lateral nematothecae which are identical. A few minor differences do, however, occur. Thus the transverse nodes separating off intermediate internodes are better defined and almost invariably present, and in 3 of the samples (SCD 61H and 186X, and SAMH 292) the processes supporting the lateral nematothecae arise at a slightly higher level on the internode. Also in these simple forms most of the individual parts are smaller, particularly the lateral and supplementary lateral nematothecae. Yet the range of variation in the measurements of the different samples, and the overlapping which occurs, make it impossible to distinguish two species on these grounds. Two of the simple forms (SAMH 292 and SCD 186X) bear male gonothe- cae. Each gonotheca arises immediately below a hydrotheca, is roughly pear-shaped but curved towards the hydrocladium, and bears a round HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 281 operculum at the wide distal end and two large nematothecae near the base. These simple forms can always be distinguished from South African species of Antennella by the characteristic shape of the lateral nematothecae. Corhiza scotiae (Ritchie), 1907 (Fig. 3 C) Antennopsis scotiae Ritchie, 1907, p. 543, pl. 3 (fig. 3). Ritchie, 1909, p. go, fig. 8. Millard, 1957, p- 235. . Records. West coast: WCD 7L, 12F, 18M. South coast: TRA 35B, 38C, 92D. SCD 26B, 67L, 84R, 153G, 175D, 191Q, 195K. SAMH 307, 323, 331, 345: Remarks. ‘This endemic species has a somewhat restricted distribution round the South African coast, being found between the lines of 17° and 28°E. longitude and in a depth-range of 18-120 m. It has never been found in the littoral area, and is typical of the Agulhas Bank. Gattya humilis Allman, 1886 Gattya humilis Allman, 1886, p. 156, pl. 24 (figs. 5-7). Paragattya intermedia Warren, 1908, p. 323, pl. 47 (fig. 27), fig. 16. Mi€ullard, 1957, p. 230. Millard, 1958, p. 209. Records. West coast: PP 4Z. LAM 27G, 43H. SB 168D (recorded by Day, 1959, as Paragattya intermedia). CP 646D, 650E. South coast: MB 15D, 52G, 60M, 88K. LIZ 27G. SCD 36R, 84Q. SAMH 275. Remarks. Examination of Allman’s type material of Gattya humilis in the British Museum of Natural History (reg. no. 86.2.19.41), and a comparison with Warren’s cotypes of Paragattya intermedia, show that these species are synony- mous. The material figured by Allman (pl. 24, fig. 6) includes 3 young stems, but in the same colony branching stems are also present similar to those described by Warren. Allman failed to observe the median superior nemato- theca. Allman’s material is labelled ‘Port Elizabeth (?), Cape of Good Hope’, and is almost certainly South African in origin. The genus Paragattya Warren 1908 thus becomes sunk in the synonymy of Gattya Allman 1886, with Gattya humilis Allman 1886 as the type species. The following points may be added to the published descriptions. Firstly the first two oblique nodes of the stem are extra well defined as hinge-joints, and between them arise not one, but a pair, of opposite hydrocladia. Secondly the hydrocladia occasionally rebranch, the branch being of similar construction to the hydrocladium and bearing 1-3 hydrothecae. Both these features are also characteristic of the closely related genus Halopteris. 282 ANNALS OF THE SOUTH AFRICAN MUSEUM Halopteris constricta ‘Totton, 1930 (Hing G)), Halopteris constricta. Millard, 1957, p. 227, fig. 14 A. Records. South coast: LIZ 27J. MB 55H. (Additional from False Bay: FAL 326D.) Description. Two small colonies including simple and pinnate forms. Structure as previously described, except that in some stems the first (and sometimes also the second) segment after the hinge-joint gives rise to a pair of hydrocladia arising one on each side of the cauline hydrotheca. Female gonothecae (not previously described) borne on thecate inter- nodes of stem, each on a very short pedicel which arises on one side of the internodal projection which supports the hydrotheca. Gonotheca strongly curved towards the stem, with a wide distal aperture facing the stem, bearing 2 bithalamic nematothecae on basal part. (These gonothecae were not present in the samples from the south coast, but on new material from False Bay.) Halopteris gemellipara nov. sp. (Fig. 4. A-F) Types and records. Holotype: SAMH 308. Other records: (NAD 22B). SCD 36P, 153L. Description. Stem unfascicled, reaching a maximum height of 5:5 cm., bearing pinnately arranged hydrocladia. Basal part without hydrothecae or hydro- cladia, divided by irregular transverse nodes, and bearing a median series of nematothecae on the anterior surface. No hinge-joints. Remainder divided into rather long internodes by oblique nodes; each internode bearing a hydro- theca near the base on the anterior surface, and a pair of opposite hydrocladia arising one on each side of the hydrotheca. Hydrocladium borne on a short apophysis of the stem and separated from it by a transverse node, which, however, is often not clearly visible. Remaining nodes oblique. First internode athecate, bearing one median nematotheca near distal end. Remaining internodes thecate, but with the distal end sometimes cut off by a transverse node situated just above the attachment of the hydro- theca. Hydrothecae not exactly on the distal surface of the hydrocladium, but shifted slightly towards the anterior surface. > A-F. Halopteris gemellipara n. sp. A and B, two hydrocladial internodes from SAMH 308 (the holotype) and NAD 22B respectively. C, the whole stem; D, a cauline internode in anterior view; E, a lateral nematotheca; F, the median inferior nematotheca, all from the holotype. G. Halopteris constricta Totton. A portion of the stem in lateral view showing female gonothecae, from FAL 326 D. H, J. Halopteris valdiviae (Stechow). H, part of a hydrocladium with 2 male gonothecae; J, a female gonotheca. Both from SAMH 349. K. Halopteris glutinosa (Lamx.), female gonotheca, from SAMH 278. HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA ee A—.B Dl! : 5S mm. Fic. 4. Halopteris spp. 283 284 ANNALS OF THE SOUTH AFRICAN MUSEUM Hydrotheca cylindrical, expanding towards margin; adnate for about 2 height, then free; abcauline height exceeding (or rarely equal to or less than) width at margin; margin forming an angle of 55° or more with internode. Each thecate internode bearing 3 nematothecae, one median inferior seated well below the base of the hydrotheca, and 2 laterals seated on short processes one on each side of the hydrotheca and not reaching margin. A single median supracalycine nematotheca sometimes present on hydrocladial internodes, and 2-4 on cauline internodes. _ Median nematotheca short and broad, probably immovable, 2-chambered ; with small basal chamber and larger distal chamber separated by transverse septum; with distal chamber cut away on adcauline side and margin facing towards internode. Abcauline wall strongly convex in supracalycine nemato- thecae, the nematotheca of the first hydrocladial internode, and the median inferior nematothecae on the stem; only slightly convex in the median inferior nematothecae on the hydrocladia. Lateral nematotheca minute, 1-chambered, immovable, with adcauline wall cut away and margin facing towards internode. Gonothecae absent. Measurements (mm.) SAMH 308 NAD 22B- SCD 36P Stem, internode length .. se .. 1°24-1°69 1:18-1:96 1-00—1°63 diameter ais he be .. 0°28-0'41 0°21-0°32 =0°27—0°365 Hydrocladium, internode length, first athecate Ags be: i .. 0°34-0°55 0°43-0°48 0°39-0°53 normal thecate ms Hes .. 0°55-0°71 0:66-0-80 0°53-0-70 Hydrotheca, height abcauline .. .. 0°29-0:38 0°40-0°50 024-050 height, adcauline, free part .. O'10-0°16 = =0°13-0:18 —0:08-0°15, diameter at margin .. yl .. 0°29-0°33 0:24-0:29 0°26-0°295 free part/abcauline height .. .. 0°29-0°46 0:30-0:36 0:24-0°47 diameter/abcauline height .. .. 0°82-1:02 0°51-0°60 0°54-1°17 Nematotheca, height, median cauline .. 0:095-0:18 0:13-0:16 0-12—0°16 hydrocladial, median inferior .. O'0Q-O'II O'115—-0°135 0:09—-0°13 lateral Sif a we .. 0°035-0:05 0:05-0:06 0:03-0:06 Remarks. In the measurements and the figure a sample from Natal (NAD 22B) is included, as it illustrates the variation possible in the species. It differs from the holotype in several particulars, mainly in the proportions of the hydro- thecae, which are deeper and narrower. As a result the diameter at the margin is little more than half the abcauline height, whereas in the holotype the diameter is practically equal to the height. Further, the axis of the hydrotheca is more nearly parallel to the hydrocladium so that the margin forms an angle of about 70° with it as against about 55° in the holotype (fig. 4B). Intermediate internodes are also more common on the hydrocladia, especially towards their HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 285 distal ends. Sample SCD 36P contains both forms, though on separate stems. Since the points mentioned are all differences of proportion and degree, and all structural details are exactly the same, and bearing in mind the tendency towards variation in Halopteris, I hesitate to separate these two forms as separate species, at least until more material is available. One stem of the holotype shows a rather unusual abnormality and bears a pair of hydrocladia on the posterior surface in addition to those on the anterior surface. H. gemellipara resembles H. catherina (Johnston) 1833 and H. zygocladia Bale 1914a in the possession of opposite hydrocladia, but differs from both in the monothalamic lateral nematothecae. In this respect, and in other structural details, it is very similar to Antennellopsis integerrima Jaderholm 18096, from Japan, differing from it in the longer free part of the hydrotheca and the oblique hydrocladial internodes. FHlalopteris glutinosa (Lamx.), 1816 (Fig. 4 K) Heteroplon pluma Allman, 1883, p. 32, pl. 8 (figs. 1-3). Plumularia glutinosa. Billard, 1910, p. 36, fig. 16. Stechow, 1925, p. 502. Plumularia alternata (Nutting), 1900. Jarvis, 1922 (pp), p. 345, pl. 25 (fig. 16). Halopteris glutinosa. Millard, 1958, p. 200, fig. 10 A—-D. Records. South coast: LIZ 16E, 27H. MB 8S, 15E, 24V, 47M, 52D, 6o0J. SCD 36Q, 84P, 153K, 169U. SAMH 103, 197, 225, 276, 278, 298. TRA 92G. Description. Colonies altogether more robust than those on the Natal coast (Millard, 1958), most of them reaching a height of 2-3 cm., but three large colonies reaching 10, 12 and 18 cm. (TRA 92G, SCD 84P, and SAMH 278 respectively), thus resembling more closely the colonies described by Allman and Stechow. In the larger colonies the lower regions of the stem have no visible nodes, the hydrocladia are usually lost, and the hydrothecae tend to be eroded and shallow. All measurements exceed those of the east coast material, though the proportions and appearance of the hydrothecae are similar. Hydro- cladia may bear up to 15 hydrothecae. Lateral nematothecae with terminal chamber deeply cut away on adcauline (1.e. inner) side. Stem internodes may bear as many as 3 extra pairs of nematothecae above the margin of the hydrotheca, and occasionally the 1-chambered supracalycine nematotheca is paired in the lower part of the hydrocladium. Female gonothecae (not previously described) borne on the stem, each on a short pedicel of one segment arising next to a hydrotheca. Very large, flattened, bluntly rounded at distal end, bearing 2-5 2-chambered nematothe- cae near base, reaching a length of 3-74 mm. and a diameter of 1°70 mm. Remarks. The material described by Jarvis (1922) from East Africa, and assigned by her to Plumularia alternata Nutting, is present in the British Museum of Natural History. On examination the samples from Zanzibar and Wasin 286 ANNALS OF THE SOUTH AFRICAN MUSEUM / both proved to be mixed, with some plumes (those illustrated in pl. 25, fig. 16, of Jarvis) resembling exactly Halopteris glutinosa and attributable to this species. In other plumes, however, the stems were more slender and geniculate and possessed athecate intermediate internodes, and the hydrothecae had sinuated margins, each with one distinct anterior lobe and 2 lateral lobes. It was undoubtedly these plumes which led Jarvis to assign the material to P. alternata, although the stems do not possess the delicacy of Nutting’s material. They possibly represent a separate variety of H. glutinosa or a separate species altogether. Halopteris tuba (Kirch.), 1876 (Fig. 5) Plumularia tuba Kirchenpauer, 1876, p. 44, pl. 1 (fig. 2), pl. 4 (fig. 2). Acladia africana Marktanner-Turneretscher, 1890, p. 261, pl. 5 (figs. 11, 11a). Heteroplon jaederholmi Stechow, 1912, p. 366, figs. F, G. Plumularia (Heteroplon) africana. Stechow, 1925, p. 500, figs. 44, 45. Thecocaulus tuba. Leloup, 1939, p. 12, fig. 8. Records. South coast: SCD 5M, 52L, 79F, 82K, 84N, 1175, 153F, 175A, 186W, 219U, 234Q, 239A. SAMH 194, 238, 249. AFR 865.0.G. Description. Colonies of unfascicled, unbranched stems reaching a maximum height of 22 cm. Basal part of stem without hydrothecae or hydrocladia, with a few irregular transverse nodes and a double row of movable nematothecae on the anterior face. ‘This part usually terminated by 1 or 2 oblique hinge-joints. Where 2 hinge-joints occur the segment between them bears a single median hydrotheca on the anterior face and a pair of hydrocladia arising one on each side of it, also 1 medial inferior nematotheca and 4-8 movable nematothecae rather irregularly arranged above the hydrotheca. Remainder of stem bearing hydrothecae and hydrocladia on anterior surface, unsegmented except for extreme distal end, where regular, slightly oblique nodes may occur. The first hydrotheca is median, and from the same level arises a pair of hydrocladia, one on each side of it. Thereafter the hydro- thecae are displaced alternately to the right and left of the mid-line and the hydrocladia arise singly, alternately from the right and left sides of the hydro- thecae (from the right side of a hydrotheca displaced to the right and vice versa). The hydrothecae and hydrocladial origins thus form 2 longitudinal rows on the anterior surface of the stem. Very rarely a pair of hydrocladia arises from the same level. Corresponding to each hydrotheca are 5 cauline nematothecae: 1 inferior, 2 lateral, and 2 minute supracalycine. Of these the inferior nema- totheca is not situated immediately below its hydrotheca, but is displaced towards the opposite side of the stem. Hydrocladia borne on short apophyses of the stem arising next to the cau- line hydrothecae (the hydrothecae thus appear to be seated on the apophyses), divided into internodes by distinct oblique nodes. Apophysis without nema- tothecae. First internode short, athecate, with 1 median movable nematotheca. a HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 287 o:-S5mm. as Se i eae Fic. 5. Halopteris tuba (Kirch.). A. The lower part of the stem in anterior view, showing hinge-joints and arrangement of hydrocladia and cauline nematothecae. From SAMH 238. B, C. Cauline hydrothecae from SCD 5M to show shape of hydrotheca and supracalycine nematothecae—B from upper region, and C from lower region, of stem. Aperture for lateral nematotheca shown on left of diagram. D. The first 2 hydrocladial internodes, from SCD 5M. E-G. Stages in the development of the female gonophore, from SCD 5M. 288 ANNALS OF THE SOUTH AFRICAN MUSEUM Remaining internodes (up to 12 in number) each with 1 hydrotheca, 1 median inferior nematotheca, 2 lateral nematothecae, and 1 supracalycine naked sarcophore. Old colonies may have weak internodal septa at the proximal and distal ends of the internodes. Hydrotheca shallow, with width exceeding (or, rarely, equal to) depth, and with untoothed margin. Hydrocladial hydrotheca adnate to internode to within a short distance of margin, then free. Cauline hydrotheca with abcauline wall greatly thickened, and margin eroded, particularly on adcauline and lateral sides, so that no free part remains. Hydropore very small, circular, close to abcauline side. Nematophores of 4 types: (i) Lateral nematothecae: slender, elongated, and movable. Bithalamic, with funnel-shaped distal chamber which is not cut away on adcauline (i.e. inner) side. Those on the hydrocladia arise from short processes of the internodes next to the hydrotheca and usually reach well above the margin. Those on the stem are much larger, and are not seated on processes, but emerge directly from oval openings in the perisarc. They are not con- tiguous to the hydrotheca but separated from it by a distance equal to about half the thecal width. The median nematotheca of the first hydro- cladial internode is very similar, but smaller and with a wider base— it appears to be intermediate between this and the following type. (ii) Median nematothecae: short, hook-shaped and immovable. With a transverse septum and thus bithalamic. Distal chamber with adcauline wall completely cut away. Situated well below hydrotheca and not reach- ing to base. (iii) Supracalycine nematothecae: immovable, hook-shaped, minute, mono- thalamic. Situated immediately above the eroded adcauline wall of the cauline hydrothecae. (iv) Naked sarcophores: situated in the angle behind the free part of the adcauline wall of the hydrocladial hydrothecae. It is possible that a nematotheca may be present and fused with the thecal wall, but it is not visible as a separate entity. Male and female gonothecae (not previously described) borne on the same colony and often on the same hydrocladium with the male more distal than the female. Female gonotheca arising from a small triangular pedicel attached to the first thecate internode of hydrocladium just below the hydrotheca and slightly to one side. Very large, flattened, rounded distally and tapering proximally. Bearing 3 nematothecae near base on surface facing hydrocladium, these being similar in size and appearance to the laterals of the stem. Containing a single embryo between the arms of a bifurcating blastostyle. Male gonothecae borne on internodes of hydrocladia other than the first. Smaller than female, elongated-oval, with abcauline side more convex than HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 289 adcauline, with distal aperture. (Appearance similar to male gonothecae of H. glutinosa (see Millard, 1958, fig. 10).) Measurements (mm.) SCD 5M SCD 79F Hydrocladial internode, length on posterior sur- face, athecate 0-2I—0°30 0°27—0°30 EhWecate .. 0-50—-0'64 0°43-0°53 Hydrotheca, cauline, heen abeauliive 0-19—0°22 width at mouth ee 0°22-0°27 hydrocladial, height abeaulinies 0-12—0°16 0:16-0:20 width at mouth 0°20-0°27 0-20-0'25 Nematotheca, cauline, lateral 0-26—0°30 median inferior O-12—0°13 supracalycine 0:04-0:06 hydrocladial, lateral 0-16—-0-21 O-1I-O°17 median inferior (first) O-II-O'15 O-1I-O'14 median inferior (normal) 0:08-0°15 0:065-0°12 gonothecal : 0°24—0°30 0-19—-0°26 Gonotheca, female, length Patou pedicel 2°95-3°36 2°83-3-11 maximum diameter I*51—-1'Q4. 1°40—-1°54. Gonotheca, male, length 0°36-0°44. maximum diameter O°15-0°25 Remarks. 1 am satisfied that Marktanner-Turneretscher’s species Acladia africana (redescribed by Stechow, 1925) is a synonym for Kirchenpauer’s species Plumularia tuba, and have accordingly combined the two. Kirchenpauer’s material was re-examined by Leloup in 1939, who testifies to the excellence of Kirchenpauer’s description and adds a more accurate diagram of a hydrocladial internode. The only features distinguishing the two species are (i) the segmentation of the stem which is said to be absent in A. africana and present in P. tuba, and (ii) the hydrotheca, which is said to be completely adnate in A. africana and only partially so in P. tuba. Kirchenpauer’s type material has unfortunately been lost, but I have seen a prepared slide of Stechow’s material of A. africana, and have found that the upper part of the stem is in fact segmented, and that the hydrocladial hydrothecae do have a short region which is not adnate. There is an apparent variability in the length of the lateral nematothecae on the hydrocladia, which sometimes just overtop the thecal margin (as in Leloup’s figure) and sometimes greatly overreach it (as in Stechow’s figure, 1925). This appears to be a result of erosion of the thecal margin in the latter case (in deep hydrothecae the margin is sinuated and in shallow hydrothecae the margin is even and worn). Both types may occur on the same stem, but the lateral nematothecae always reach approximately the level of the next node. 290 ANNALS OF THE SOUTH AFRICAN MUSEUM A full description of the species has been given, since none of the published descriptions is quite complete. The arrangement of the cauline nematothecae was found to be exactly similar in Stechow’s material. It is evident that, as in all typical species of Halopteris, the arrangement of the cauline segments is essentially the same as that of the hydrocladia, differing only in the tendency for nodes to disappear, the lateral displacement of the hydrothecae and the nature of the supracalycine nematophores. This endemic South African species occurs on the Agulhas Bank east of 19°E., extending round the coast to Natal waters. It is found in depths of II-130 m. Halopteris valdiviae (Stechow), 1923 (Fig. 4 H, J) Thecocaulus(?) valdiviae. Stechow, 1925, p. 495, figs. 42-43. Halopteris valdiviae. Millard, 1957, p. 228, fig. 14 B. Records. South coast: SAMH 1096, 248, 324, 332, 346, 349. SCD 79G, 169W. West coast: AFR 736X. WCD 7M. Description. ‘Two magnificent tree-like colonies reaching 35 and 37 cm. in height (SAMH 248 and 349), and a number of smaller colonies and detached portions. A few details may be added to previous descriptions. Hydrorhiza forming a dense mat 5 cm. wide by 3 cm. high. Stem strongly fascicled, reaching 1 cm. in diameter near base, branching irregularly. Stem and branches giving off hydrocladia-bearing pinnae from the deeper tubes in an irregular fashion and from all surfaces. Pinnae unfascicled, or very lightly fascicled near base, with basal part terminated by 2 oblique hinge-joints, but occasionally as many as 4, possibly due to regeneration. Segment between hinge-joints bearing 1 median cauline hydrotheca, and 1 pair of opposite hydrocladia on anterior surface. First segment beyond hinge-joints similar, with median hydrotheca and 1 pair of hydrocladia. Remaining segments each bearing 1 cauline hydrotheca, and 1 hydrocladium arising alternately from the right and left of the hydrotheca. Hydrocladia as described by Stechow (1925) and Millard (1957), bearing up to 9 hydrothecae. Nematothecae as previously described, except that it is now possible to distinguish a pair of minute, monothalamic supracalycine nematothecae behind the free part of the adcauline wall of each cauline hydrotheca. Gonothecae (not previously described) dimorphic, male and female on same colony, each borne on a small triangular pedicel of 1 segment which arises immediately below a hydrotheca and slightly to one side. Male gonothecae borne on thecate internodes of hydrocladia, pear-shaped but slightly asymmetrical with the greatest curvature on side away from hydro- cladium, with flattened distal end, bearing 2 bithalamic nematothecae near base on side facing hydrocladium. HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 291 Female gonothecae borne on internodes of pinnae, very much larger than male, flattened in a plane at right angles to pinna, expanding distally to a wide aperture, bearing 3 or 4 bithalamic nematothecae on side facing pinna. Measurements (mm.) Gonotheca, male, length ne Ae ae aH, a .. 0°58-0°83 maximum diameter a = ne a .. 0°20—0°:39 Gonotheca, female, length .. ae Ss Sy af 2. 1°24-1°55 maximum diameter ae e. bs mS eis -2) W:O1—1-22 The measurements of the trophosome agree well with those given by Millard (1957), except that the range is in some cases greater. Remarks. It is now obvious that the material described by Millard (1957) consisted of 2 detached pinnae bearing hydrocladia. This endemic South African species occurs on the west and south coasts as far east as 29°E., and in depths of 33-201 m. It is remarkably similar to H. tuba in the arrangement and structure of its hydrocladia, differing from it in its fascicled stem, and proportions of hydrotheca in which the depth is almost invariably greater than the width. Monostaechas natalensis Millard, 1958 (Fig. 2 F) Monostaechas natalensis Millard, 1958, p. 206, fig. 12. Records. South coast: GPR 46J. SCD 153D. Description. The first colony is of small size and similar to the holotype in its appearance and measurements, but the second colony (SCD 153D) is altogether of larger proportions. The fascicled stems branch irregularly and reach a maximum height of 7-6 cm., and individual hydrocladia reach a length of 3-4 cm. All the minute measurements are larger. The proportions of the parts are the same except that a greater part of the adcauline thecal wall is free from the internode. Towards the ends of the hydrocladia there is a tendency for the distal parts of the internodes to be cut off by transverse nodes. No branching hydrocladia were observed—any branching which may occur is obscured in the fasciculation of the stem. The general appearance of the colony is very similar to that of Corhiza pannosa n. sp., but it is clearly distin- guished from it by the nature of the lateral nematothecae. With the larger size the structure of the latter could be more exactly ascertained. Each nema- totheca has the sides deeply cut away to the level of the diaphragm, with a high, bonnet-shaped abcauline (i.e. outer) wall and practically no adcauline (i.e. inner) wall at all. It resembles that of Corhiza bellicosa n. sp., except that the adcauline wall is more reduced. In the holotype the structure of the nema- tothecae was not easy to determine because of the small size and the opacity of the contents, but a re-examination of the material with a phase-contrast 292 ANNALS OF THE SOUTH AFRICAN MUSEUM microscope shows that they are of a similar nature and that the diagram given by Millard in 1958 (fig. 12E) was not strictly accurate. The measurements of the larger colony are given for comparison with the holotype. Measurements (mm.) SCD 153D Internode length (on posterior surface) .. aoe ue .. O'7I-1°05 diameter (above hydrotheca) - ie HY .. OTI4-0°19 Hydrotheca, height .. vs ie Br oe ae ». 0°34-0°45 height of free part a: a: ae si x .. | OFFG=0794 free part/total length .. ae a oS if ..> 0°44-0°65 diameter at margin... oe a oe ne ~. 0°27-0:34 Nematotheca, length ae in es a: wr .. 0°09—0°17 Remarks. In view of the larger measurements sample SCD 153D might well be included in a separate variety, but one is reluctant to establish a new variety on size alone when little is known about the variation within the species. Schizotricha simplex Warren, 1914 Schizotricha simplex Warren, 1914, p. 83, pl. 6 (figs. 1-19), figs. 1-4. Records. S 65B (recorded by Stephenson, Stephenson and du Toit, 1937). L 430. CPR OA. Description. Colonies reaching a maximum height of 1-4 cm., agreeing exactly with the beautiful description by Warren, bearing male and female gonophores. In the larger stems practically all the ‘pinnae’ bear 3 thecate internodes and sometimes even 4, and many of the ‘pinnules’ bear 2. Subfamily Kirchenpaueriinae Kirchenpaueria pinnata (Linn.), 1758 Kirchenpaueria pinnata. Millard, 1957, p. 233. Millard, 1959b, p. 252. Records. West coast: CPR 44A. SB 141U. LB 126, 190Q, 257P, 367X, 378E, 392R, 472H, 513J (recorded by Day, 1959). WCD 18N. South coast: S 65A (recorded by Stephenson, Stephenson and du Toit — (1937) as K. unilateralis). KNY 710, 165A (recorded by Day, Millard and Harrison, 1952). E 83. LIZ 7X, 27F. E54, 317, 3360. (CPR 7Daoms Kirchenpaueria triangulata (Totton), 1930 (Fig. 6 E-J) Plumularia triangulata Totton, 1930, p. 225, fig. 61. Records. South coast: SCD 175G. Description. Colony growing on the stems of Nemertesia ciliata Bale, and including both simple and pinnate forms. 2 HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 293 Pinnate stem reaching a length of 1 cm., and consisting of long internodes separated by transverse nodes which may be indistinct in the lower region. Each internode giving rise to a hydrocladium-bearing apophysis from its distal end. Each internode bearing a well-developed mamelon on the upper surface of the apophysis, and 2-4 cauline nematothecae. The latter including 1 on proximal end immediately above apophysis of last internode, rarely 1 near distal end below the apophysis, and 1 (or rarely 2) on upper surface of apophy- sis distal to the mamelon. Hydrocladia alternate, the two rows in one plane or shifted very slightly to one side of the stem. Hydrocladium normally homomerous and consisting of long thecate internodes separated by transverse nodes, but intermediate athecate internodes commonly occur and in quite an irregular fashion. Thus there may be 1-4 or none at all before the first thecate internode, and 1 or 2 of irregular length between successive thecate internodes. Hydrotheca borne on the proximal half of the internode, which thus has a long distal region; flaring slightly to margin, completely adnate or with a very short free portion, with diameter exceeding depth. Nematothecae: 1 median inferior and one median supracalycine; with convex abcauline wall and practically straight adcauline wall and oblique margin; movable; monothalamic, or with a very thin septum close to base. Simple stem reaching a length of about 4 mm., similar to hydrocladium, borne on a long apophysis of hydrorhiza. Measurements of parts slightly smaller than corresponding ones of pinnate stem. Gonothecae (male) borne by pinnate stem on hydrocladial apophyses to one side of mamelon, elongated, tapering below, truncated above, triangular in section. Pedicel of 2 segments. Measurements (ram.) Pinnate Simple stem stem Stem internode, length 0°48—-0°75 diameter bs iy Re: 0:07-0:16 Hydrocladium, thecate internode, length 0:66-0:85 0°59-0°81 diameter at distal end 0:055-0°075 0:05-0:07 Hydrotheca, height abcauline 0:07—0:095 0:07-0:08 height of free part 0:00-0:01 0:00-0:01 diameter at mouth O-10-0'12 0:095-O0'1I Nematotheca, length 0:05-0:08 0:05-0:07 Gonotheca, male, length 1°74-2°46 maximum. diameter 0:80-0:85 Remarks. The author agrees with Bedot (1923) that the main diagnostic charac- ter separating Azrchenpaueria from Plumularia should be the absence of paired nematothecae. Although the nematothecae in KAirchenpaueria as a whole are characteristically reduced, any attempt to separate the two genera on the 294. ANNALS OF THE SOUTH AFRICAN MUSEUM W O'lmm,. 0°'3 mm. Fic. 6. Oswaldella nova (Jarvis) and Kirchenpaueria triangulata (Totton). HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 295 structure of the nematothecae alone will fail because of the presence of inter- mediate forms. Thus a group of species with but one superior nematotheca (including P. halecioides Alder 1859, P. bonneviae Billard 1906, P. inermis Nutting 1900, and others) has been assigned by Bedot (1923) to Kirchenpaueria, by Stechow (1923b) to a new genus Ventromma, and by most other authors to Plumularia. To this category also belong Plumularia irregularis Millard 1958 and P. triangulata Totton. Most of these species have monothalamic superior nematothecae, and are thus intermediate between a typical Plumularia and a typical Kirchenpaueria. In K. triangulata, however, a poorly developed septum is often visible in the nematotheca, and in K. irregularis it is quite distinct, so that in these genera the nematothecae are strictly speaking bithalamic and verging on the Plumularia type. The South African material of K. triangulata agrees closely with Totton’s description and resembles it in the epizootic habit of growth. The material shows much evidence of regeneration after injury, and this probably accounts for the presence of the many intermediate internodes, which were not men- tioned by Totton. In one case the end of a stem has regenerated as a hydrocladium. The species is closely related to K. irregularis (Millard) 1958, differing from it in the longer distal part to the hydrocladial internode and in the arrangement of the cauline nematothecae. It is the first record of the species from South Africa. Oswaldella nova (Jarvis), 1922 (Fig. 6 A-D) Plumularia nova Jarvis, 1922, p. 347, pl. 26 (fig. 20). Kirchenpaueria adhaerens Millard, 1958, p. 203, fig. 13 F, G. Records. South coast: MB 47Z. SCD 84T. TRA o2H. West coast: WCD 7N. (Additional from Natal: NAD 1P.) Description. A number of samples of epizootic colonies growing on other hydroids (Halopteris glutinosa (Lamx.)), H. valdiviae (Stechow) and Gattya humilis Allman). coed A-D. Oswaldella nova Jarvis. A. Part of colony on back of stem of Halopteris glutinosa, from TRA 92H (hydrocladia of host cut off short). B-D. Portions of hydrocladia from different colonies to show variation in size and shape: B from TRA 92H, C from NAD 1P, and D from MB 47Z. E-J. Kirchenpaueria triangulata (Totton), from SCD 175G. E. A colony including a pinnate stem bearing gonothecae, and several simple stems, removed from host. F. A portion of a pinnate stem showing arrangement of nematothecae. G, H. Hydrocladial internodes from a simple stem and a pinnate stem respectively. J. A single nematotheca with diaphragm. 296 ANNALS OF THE SOUTH AFRICAN MUSEUM Hydrocladia arising direct from adherent stem, usually unbranched, but in one sample (TRA 92H) often forked. The division occurs immediately above the origin of a hydrotheca; both limbs are of equal thickness, and their hydro- thecae do not face one another, but lie in the same plane as those of the original hydrocladium. Unbranched hydrocladia bearing up to 14 hydrothecae. Hydrotheca variable in size and shape, with diameter generally exceeding, but occasionally equal to, height; usually free from internode for a small proportion of its length, but occasionally completely adnate. Gonothecae absent, but scars probably of old ones present below hydro- thecae and slightly to one side. Measurements (mm.) Type MB 47Z NAD 1P. TRAo2H — material Internode length (posterior surface) if .. 0°29-0°31 0°35-0°41 0:42-0:57 0:°39-0°48 diameter near distal end Me .. 0:06-0:07 0:065-0:08 0:08-0:11 0:04-0:05 Hydrotheca, length abcau- late see a .. 0°05-0:08 0:075-0'10 0:08-0'12 0:06—-0:075 length of free part .. 0:01-0:03 0:00-0:02 O-0I-0:04, *O-01 ~ free part/abcauline length. .. O°125-0°43 0:00-0:25 O°10-0°40 *0-125 diameter at mouth .. 0:°105-0°135 O*10-O‘II O:II-O°13 0:07-0:08 Remarks. Examination of the type material of Jarvis’s species Plumularia nova in the British Museum of Natural History (slide no. 23.2.15.283 from Zanzibar) has established the fact that this species is synonymous with Kzirchenpaueria adhaerens Millard. The correct measurements are quoted above and replace the incorrectly calculated ones given in 1958. The type material is scanty and damaged, and no hydrocladia bear more than 2 hydrothecae. The median inferior nematothecae are missing, though their orifices remain. In the only hydrotheca lying in a suitable position there is a short free part to the adcauline wall. The remains of one very damaged gonotheca are present, but it is impos- sible to determine the shape. The discovery of forked hydrocladia in the new material necessitates the transfer of the species from Kirchenpaueria to Oswaldella, to which genus all its structural details indicate a closer affinity. Further knowledge of the growth- habits may indicate that it is merely an epizootic form of O. bifurca (Hartlaub) or O. antarctica (Jaderholm). * One measurement only. HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 297 Subfamily Plumulariinae Nemertesia ciliata Bale, 1914b (Fig. 7 E-G) Nemertesia ciliata Bale, 1914b, p. 170, pl. 36 (fig. 1). Bale, 1915, p. 298. Briggs, 1914, p. 307, pl. 10 (fig. 3). Jaderholm, 1919, p. 23. Records. West coast: TRA 47N. AFR 729H. South coast: TRA 92F. AFR 945L. SCD 5N, 175C. SAMH 3209. Description. Several colonies, the largest reaching a height of 24:5 cm. Stem and larger branches fascicled, smaller branches unfascicled, branching always in one plane. Hydrocladia given off in whorls of 2, 3, or 4, but irregular in lower parts of branches; consisting of alternate thecate and athecate internodes, of which the first is always thecate. Appearance and structure as described by Bale and Briggs, with the following minor points of difference. The branching is not so luxuriant as in the specimen figured by Briggs and the branches are more widely spaced, with intervals in the neighbourhood of 5-10 mm. between them. In branches with two hydrocladia to a whorl it sometimes happens that the alternate pairs are not set at right angles to each other but compressed in one plane, forming angles of 40-60/120-140° with each other, and giving a flattened appearance to the branch. The same ten- dency may be found in stems with 3 hydrocladia to a whorl, as in NV. japonica (Stechow) 1909, but it is not constant within a colony. The hydrocladia are often longer than in the type material, reaching 5-7 mm., and the internodes are in general somewhat longer, but the measurements of the hydrothecae are very similar. Gonothecae as described by Bale. Measurements (mm.) Hydrocladium, internode length, thecate oe Af .. 0°29-0°51 athecate ate in he ae = a .. 0*14-0:28 diameter at distal end of internode .. ot ao .. 0°04—0°095 Hydrotheca, height abcauline a Ne ot a .. 0°02—0:07 diameter at margin .. ie bs M nie .. 0°06-0:085 Nematotheca, lateral, height a of ae ia -. 0°05-0-10 Gonotheca, length .. ie $4 i a reaching 0°49 maximum diameter .. oy - oh reaching 0°24 Remarks. This is the first record of the species from South Africa. So far it is known only from Tasmania and Japan. I have not separated a var. cruciata as was done by Bale (1915), although both this and the normal form are present, as the tendency to produce different numbers of hydrocladia to a whorl almost certainly depends on the growth and age of the colony as has been shown for other species of the genus. 298 ANNALS OF THE SOUTH AFRICAN MUSEUM rl F i 50 5mm. Fic. 7. Nemertesia spp., portions of hydrocladia showing variation. A-D. Nemertesia ramosa Lamx. A. The typical appearance from TRA 125K. No superior medial nematotheca. B. An example with smaller hydrothecae, longer internodes, and superior medial nema- tothecae, from NAD oK. C, D. Hydrocladial internodes from old and young regions respectively of the same colony, from TRA 1348S. C with internodal septa and thickened perisarc. E-G. Nemertesia ciliata Bale. E. The typical appearance, from TRA 47N. F, G. Hydrocladial internodes from old and young regions respectively of the same colony, from SAMH 329. HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 299 Nemertesia cymodocea (Busk), 1851 Nemertesia cymodocea. Billard, 1910, p. 39. Millard, 1957, p. 234. Millard, 1961, p. 207. Records. West coast: SB 178B. TB 3, 4, 5, 18, 21A. TRA 156A. South coast: MB 81N. AFR 865R. TRA 92K. SCD 26C, 36T, 52M, 153H, 175E, 219V. SAMH 148, 240, 347. Description. The length of the hydrocladial internodes is variable, sometimes being almost double that quoted by Billard (1910, p. 40), though the size of the hydrotheca remains fairly constant. Observations suggest that during the growth of a colony the stems first produce alternate hydrocladia, one to an internode, and that later the decussate arrangement develops at the tip of the stem, and still later the hexastichous arrangement. Thus young colonies bear only alternate hydrocladia and stems may reach a length of 10 cm. (and bear gonophores) before the decussate arrangement starts. The hexastichous arrangement is characteristic of the distal ends of old colonies, though by this time the proximal hydrocladia may be lost or buried by the peripheral tubes of the stem. In young colonies, too, the peripheral tubes bear a thick coating of nematothecae, which appear to fall off with age. Nemertesia ramosa Lamouroux, 1816 (Fig. 7 A—-D) Antennularia ramosa, var. plumularioides Billard, 1906, p. 215. Nemertesia ramosa. Millard, 1957, p. 235. Millard, 1961, p. 206. Nemertesia ramosa, var. plumularioides (Billard), 1906. Vervoort, 1959, p. 293, figs. 46 b, 47. Records. West coast: WCD 12D, 32H. South Coast: TRA 92E. SAMH 333. SCD 5P, 96A, 153N, 175B. (Natal: NAD 9K.) (Additional from False Bay: TRA 1348.) Description. A number of colonies of varying size, the largest reaching 28-2 cm. Considerable variation in appearance and structure occurs, much of which is probably due to growth and ageing of the colony. The number of hydrocladia produced in a verticil appears to increase with age. In young colonies the hydrocladia have an alternate arrangement, I or 2 pairs to an internode. As growth proceeds the young shoots produce 2, 3, 4, or even 5 hydrocladia to a verticil, members of one verticil alternating with those of the next resulting in as many as 10 longitudinal rows. In old colonies only the verticillate arrangement is visible since the lower hydrocladia with alternate arrangement tend to fall off and their apophyses to be buried by the peripheral tubes of the stem. The thickness of the perisarc also appears to increase with age and is possibly influenced by other factors as well. In young colonies the perisarc is thin, with the hydrocladia white and the stem a pale yellow or straw colour. 300 ANNALS OF THE SOUTH AFRICAN MUSEUM In old colonies the thickened perisarc gives to the stem and bases of the hydro- cladia a dark brown colour. Associated with the thickening of the perisarc the nodes of the stem become indistinguishable and the hydrocladial apophyses very thick and long, athecate internodes become more numerous in the hydrocladia, and in the apophyses and basal parts of the hydrocladia appear internodal septa (up to 12 to an internode). The appearance of these regions is thus very different from that in the growing tips of the stems and the distal ends of the hydrocladia where the structure is similar to that usually described for the species. Measurements (mm.) Hydrocladium, internode length, thecate oo ae .. 0°34-0°86 diameter at distal end of internode .. as ais .. O°0G=O-%2 Hydrotheca, height abcauline ve “ ae ce .. 0:06-0°14 diameter at margin... = es xe at .. 0°07-O°14, Nematotheca, lateral, height ay ae oe a .. O:07-O785 Gonotheca, length .. x an Be fy reaching 0:94 maximum diameter .. ae of f, reaching 0°43 Remarks. Billard (1906) has created a var. plumularioides for forms of NV. ramosa with alternate hydrocladia, but if, as seems practically certain, the arrangement of hydrocladia is a matter of age, the retention of this variety is unnecessary. Vervoort (1959) has described specimens with alternate hydrocladia reaching lengths of up to 25 cm. from tropical west Africa, but he himself mentions that the hydrocladia may be arranged in whorls of 3 in the upper parts of the larger colonies. Plumularia filicaulis Kirchenpauer, 1876 Plumularia filicaulis. Millard, 1958, p. 209, fig. 13 D, E. Records. West coast: WCD 34D, 81K. Plumularia lagenifera Allman, 1886 Plumularia lagenifera. Millard, 1957, p. 230. Records. West coast: TRA 156D. LAM 32], 7L, 9V, 13H, 26Q, 46N, 590A, 62T. PP 1W. LB511E (recorded by Day, 1959). A382. WCD 34C, 81H. CP 650C. South coast: KNY 300 (recorded by Day, Millard and Harrison (1952) as P. setacea). Plumularia pulchella Bale, 1882 Plumularia pulchella. Millard, 1957, p. 232. Records. South coast: SCD 37S. SAMH 290, 302. MB 8Q, 47Q, 52F, 55G, 60L. KNY 71B, 165B (recorded by Day, Millard and Harrison, 1952). LIZ 27C. HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 301 Description. Colonies reaching a maximum height of 1-1 cm. (KNY 71B), but usually between 0-3 and 0-7 cm. Gonothecae ovate, truncated distally, with wide distal aperture, 5-54 times height of hydrotheca. Plumularia setacea (Ellis and Solander), 1755 Plumularia setacea. Millard, 1957, p. 232. Millard, 1958, p. 212. Millard, 1959b, p. 252. Records. West coast: WCD 18R, 2o0F, 34E. ‘TRA 156B. B 114D. LAM arT, 27Z, 35C, 45X. SB 150A. LB 152, 367W (recorded by Day, 1959). TB 2, 21H. CP 640C. South coast: SCD 5K, 52S, 169V. TRA 159A. SAMH 108, 224. CPR 1, 46H. MB 817, 24X, 47R, 55F. E 170 (recorded by Stephenson, Stephenson and Bright, 1938). LIZ 17H, 27E. L 57, 145 (recorded by Eyre, Broekhuysen and Crichton, 1938). Plumularia spinulosa Bale, 1882 Plumularia spinulosa. Bale, 1884, p. 139, pl. 12 (figs. 11-12). Bale, 1888, p. 783, pl. 19 (figs. 11-13). Warren, 1908, p. 320. Monotheca spinulosa var. obtusa Stechow, 1923b, p. 224. Plumularia spinulosa var. obtusa. Millard, 1957, p. 232. Plumularia spinulosa var. typica Stechow, 1923b. Millard, 1958, p. 212. Records. South coast: SCD 29G, 36W, 75G, 85A. SAMH 199, 291, 303. MB 47P, 58C. LIZ 13D, 27D. L 111 (reported by Eyre, Broekhuysen and Crichton, 1938). Remarks. I have no longer attempted to distinguish var. obtusa from var. typica as I have found that the length of the terminal spine of the hydrocladium varies not only in different samples but also within the same colony. The origin of the hydrocladium, whether from the middle or distal end of a stem internode, is also a variable feature. Plumularia wasim Jarvis, 1922 (Fig. 8) Plumularia wasini Jarvis, 1922, p. 349, pl. 26 (fig. 23), fig. 1. Records. South coast: CPR 46G. Description. Stem reaching a height of 1-6 cm., unfascicled, unbranched, divided into internodes of irregular length by transverse nodes; basal part short and without hydrocladia; no hinge-joints; distal part giving rise to alternate hydro- cladia, 1-3 per internode, the 2 rows in the same plane. Intervals between hydrocladia not regular, some being short and others long. Stem bearing 2 rows of cauline nematothecae (2-14 per internode), sometimes on the lateral sides but often shifted on to the anterior or even the posterior surface. Hydrocladia borne on very short apophyses of stem, divided into alternate 302 ANNALS OF THE SOUTH AFRICAN MUSEUM athecate and thecate internodes terminated by oblique and straight nodes respectively. Apophysis without nematothecae or mamelon. Hydrotheca large, with even or slightly sinuated margin; free from inter- node for more than half height; with diameter always exceeding depth. Hydro- thecae not strictly on upper surface of hydrocladium but shifted slightly onto anterior surface. Nematothecae: 1 (or occasionally 2) on each athecate internode; 4 on each thecate internode, including 1 median inferior not quite reaching base of hydrotheca, 1 pair of laterals seated on processes next to hydrotheca and not quite reaching its margin, and 1 superior in angle behind adcauline thecal wall. All bithalamic and movable except superior which is monothalamic. Gonothecae borne on thecate hydrocladial internodes, each on a short pedicel of 2 segments, and each bearing 2 or 3 large nematothecae on basal region. ‘wo types present: the larger (probably female, but empty) on first thecate internode of hydrocladium only, pear-shaped, with large operculum at distal end; the smaller (male) usually on thecate internodes other than the first, more rounded and without operculum. Measurements (mm.) Stem internode, length se te x oA 7 .. 0°34-0°99 diameter bye ate af: ck ie ie .. O°145—-0°24 . Hydrocladium, internode length (on posterior surface), thecate 0-16—-0-21 athecate Me ie ie se N m .. OFL7—-OF Hydrotheca, height abcauline ne Bre Avs sf .. O0°12—0°185 height adcauline, free part .. igh on sts .. O°10-0°13 free part/abcauline height .. ae is ane .. 0°65-1-:00 diameter at margin: .. a - ai aig . 227 (O02 O=aean Nematotheca, lateral a re Mi. 4g =e .. 0:06—0:085, inferior median. . *: os a at 2A .. 0°07—0:09 superior By yd ag an. ots a .. 0°03-0'045 cauline .. 2 aK. se ae ~ ae .. .0*0Q—O'II Gonotheca, female ?, length without pedicel he reaching 0-69 maximum diameter a Re a reaching 0-61 male, length without pedicel, He se reaching 0°35 maximum diameter ie oe wy reaching 0-24. Remarks. This material was compared with a slide of Jarvis’s type in the British Museum of Natural History. In the latter the athecate internodes normally. bear 2 nematothecae, but the number is variable and some have only one. In the present material 1 is the normal condition and 2 occur only rarely. In the type material the superior median nematotheca is 1-chambered and not 2-chambered as figured by Jarvis (1922, pl. 26, fig. 23 A). Certain minor variations sometimes occur. Thus in CPR 46G one hydro- cladium bears a branch arising from its first thecate internode and bearing one hydrotheca. HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 303 (ee ee ee ty foddtesthalinal Fic. 8. Plumularia wasini Jarvis. A. A portion of stem in anterior view showing male (small) and female (large) gonothecae and origins of hydrocladia. Gonothecae removed on right side. B. A portion of a hydrocladium. Both from CPR 46G. This speciés is in some ways reminiscent of Halopteris, as instance the large hydrothecae which are not completely adnate, the shape and arrange- ment of the gonothecae, and the presence of nematothecae on them. The species is so far only known from Wasin, East Africa. It is a new record from South Africa. Subfamily Aglaopheniinae Aglaophenia late-carinata Allman, 1877 Aglaophenia late-carinata. Millard, 1958, p. 213, fig. 14. Vervoort, 1959, p. 309, fig. 54. Records. South coast: CPR 46F. 304. ANNALS OF THE SOUTH AFRICAN MUSEUM Aglaophenia pluma (Linn.) 1758, subsp. dichotoma (M. Sars), 1857 eee eee var. dichotoma. Millard, 1957, p. 239, fig. 15 B, CG. Vervoort, 1959, p. 308, g. 52 b. Records. West coast: LAM 16W, 35B, 51A. WCD 32G. South coast: MB 5E, 34F, 37C, 42A, 79E, 81M, 84A. KNY 22F, 28C, 70B, 71D, 127A (recorded by Day, Millard and Harrison (1952) as A. ?dicho- toma). ‘TRA 35Q, 36A, 38A, 42A, 56W, 99M. SAMH 149, 163, 195, 215, 322, 330, 344. SCD 20C, 26A, 29D; 690A, 94C, 103D, 104F,, 108], 120) eiesee 152X, 153B, 191P, 195J, 199U, 230B. Aglaophenia pluma (Linn.) 1758, subsp. parvula Bale, 1882 Aglaophenia pluma, var. parvula. Millard, 1957, p. 239, fig. 15 D-F. Millard, 1958, p. 215. Ver- voort, 1959, p. 307, figs. 52 a, 53 b. Records. West coast: SWA 5Y. N 116. LAM 45Z. B 114A. SB 141T, 194]. LB 511D. CP 17, 328, 335, 650G. TRA 156K. WCD 34F, 81A. South coast: E 133. L 149. SAMH 201, 304. Cladocarpus leloupi nov. nom. Cladocarpus flexilis Leloup, 1939, p. 14, fig. 10. Records. South coast: TRA 92J. SCD 96B, 122V, 141K, 145A, 153M, 188R. Description. A number of colonies and fragments without root-stock, the longest stem reaching 22 cm. Stem fascicled but unbranched, bearing alternate hydro- cladia. Stem internodes each bearing 3-6 median nematothecae, one in the axil of the hydrocladium and the rest below it. Cauline nematotheca with 1, 2 or 3 openings. Hydrocladium and hydrotheca as described by Leloup, but intrathecal ridges tending to disappear towards the distal ends of the hydrocladia. Lateral nematotheca usually with 3 distal openings (but occasionally with 1, 2 or 4) and 1 lateral opening on mesial surface. Gonophores absent. Measurements (mm.) Hydrocladial internode, length ne ae Br ae .. 0°70-0:87 diameter at distalend .. ae a ay ne .. 0°08-0°13 Hydrotheca, depth to median tooth .. Ss camuiiger: ys .. 0°45-0°62 diameter at margin me Ns se 5 oes .. O*19—-0°26 Remarks. In view of the fact that the name Cladocarpus flexilis is preoccupied by an American species, C. flexilis Verrill 1883, I assign to Leloup’s species the new name C’ leloupi. 5 ie sai HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 305 Cladocarpus lignosus (Kirchenpauer), 1872 (Fig. 9) Cladocarpus lignosus. Stechow, 1923b, p. 243, fig. J*. Stechow, 1925, p. 505. Millard, 1961, p. 206. Records. South coast: TRA 38H, 56S, 151E. SAMH 167, 280, 284, 297, 306. Sie 26), 52H, 84A, 153A, 201X. (Natal: NAD aiF.) Fic. 9. Cladocarpus lignosus (Kirch.). A, C. Phylactocarps bearing gonothecae, from NAD 21F and TRA 38H respectively. B. The distal part of the hydrocladium-bearing tube of the stem to show segmentation and cauline nematothecae, from NAD alF. D. Two nematothecae from a phylactocarp, one with 2 apertures and the other with 3. Description. Colonies large, reaching a maximum height of 34:5 cm., easily recognizable by the woody nature of the stem and the branching, which is always in one plane. Stem dividing and subdividing in a very irregular fashion, some of the smaller branches showing a tendency towards an alternate arrangement, but frequently opposite or unilateral—the final result is a fan-shaped colony with well-dispersed branches. Hydrocladia borne on the smaller and medium-sized branches, absent 306 ANNALS OF THE SOUTH AFRICAN MUSEUM on the larger ones. Hydrocladia arising alternately from a single tube of the fascicled branch, this tube, where it is free from the peripheral tubes at the extremity, being regularly segmented, each internode giving rise to a long hydrocladium-bearing apophysis at about the middle of its length. Hydro- cladium with first internode short and athecate, with 1 median nematotheca and 2 internodal septa; remaining internodes longer and thecate, each with a median and 2 lateral nematothecae, and 3 or 5 internodal septa. Three of the internodal septa situated as described by Stechow, but frequently an additional one at the extreme proximal and another at the extreme distal end. Hydrotheca with depth about 14 times width at mouth, slightly compressed laterally, with 1 anterior and 2 lateral marginal teeth, which may all be of equal size or with the anterior tooth slightly exceeding the laterals in length. Median nematotheca arising from the basal swelling of the internode on which the hydrotheca is seated, very short, quite free from hydrotheca, with 2 distal, tubular apertures. Lateral nematotheca with 3 apertures, one on the mesial surface, and 2 distal and tubular, of which the anterior one is long and reaches beyond thecal margin, and the other is shorter and reaches to the margin or just below te Two cauline nematothecae on one surface of each internode of the hydro- cladia-bearing tube, 1 near base on opposite side to apophysis, and 1 in axil of apophysis, both with 2 distal apertures. Two rows of small, single nemato- thecae on each peripheral tube. Phylactocarps present in the older parts of the colony, 1 or a pair, on each hydrocladium arising one on each side of the median nematotheca of the first thecate internode. Phylactocarp unbranched, curving towards hydro- cladium, bearing up to 8 pairs of nematothecae. Nematothecae decreasing in size from base to tip, basal ones stout and with 3 apertures—2 terminal and I on mesial surface, distal ones more slender and generally with 2 apertures— I terminal and 1 on mesial surface. One to four gonothecae (not previously described) borne on each phylacto- carp, 1 between the members of each pair of nematothecae on surface facing hydrocladium. Gonotheca elongated, with more or less rectangular aperture near distal end facing towards phylactocarp. Measurements (mm.) Hydrocladium, first internode, length on posterior surface .. 013-0°18 remaining internodes, length on posterior surface .. .. 0°33-0°42 Hydrotheca, depth to lateral tooth ae au 1h _. > (O23 o ag width at mouth ae be Be au at .. O*15—-O0°19 Median nematotheca, length of free part be ee .. 0°025-0°05 Lateral nematotheca, length ae fe wf ay .. | \0°L2=016 Gonotheca, length .. ah ans i a reaching 1°78 WAIL INE eet a PN a ay a: reaching 0:68 HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 307 Remarks. This endemic South African species occurs on the Agulhas Bank from 19°E. eastwards, extending round into Natal waters as far as 31°E. It occurs in depths ranging from 22 to 120 metres. The plentiful material available makes possible the amplification of Stechow’s description. Halicornaria arcuata (Lamouroux), 1816 (Fig. 10 E) Halicornaria arcuata. Millard, 1958, p. 218, fig. 15 D, E. Records. South coast: MB 24U, 47G. KNY go0T, 71F. CPR 46A. SAMH 257, 300. SCD 20B, 84C, 117P, 169R. Description. Rich colonies of unbranched stems reaching a maximum height of 12°8 cm. 2mm. TRAN oe . site Lipkwoerer rept eepe SS Se Sd Oesimim: Fic. 10. Halicornaria arcuata (Lamx.). A-D. Subsp. epizootica Millard. A. Part ofa colony from SCD 84D growing on the back of the stem of the normal subspecies. Hydrocladia of host cut off short. B-D. Examples of reduced hydrothecae, B from SAMH 255, C and D from the upper and lower parts respectively of the same hydrocladium, from SAMH 355. E. Subsp. arcuata, two gonothecae from MB 47G. 308 ANNALS OF THE SOUTH AFRICAN MUSEUM Hydrothecal margin with an anterior tooth, a posterior tooth and usually 3 pairs of lateral teeth. Of the latter the posterior pair is often poorly defined in the proximal region of the hydrocladium, and the middle pair in the distal region. Median nematotheca usually closed at the tip, but often open in the proximal regions of the hydrocladia. Lateral nematothecae with 2 openings, of which 1 may be produced into a tube or into a long closed spine near the distal ends of the hydrocladia. One stem bears several branching hydrocladia, possibly the result of injury. Each branch arises from the posterior surface of the hydrocladium and bears hydrothecae on the surface facing it. Female gonothecae borne on the hydrocladial apophyses, 1 to each, and forming a double row on the anterior surface of the stem. Each is rounded below and truncated above, with the width slightly exceeding the height. Remarks. Female gonophores are said to have been mentioned by Krauss (1837) as occurring in the axils of the hydrocladia, but this is the first description of these structures. Halicornaria arcuata epizootica Millard, 1958 (Fig. 10 A-D) Halicornaria arcuata var. epizootica Millard, 1958, p. 218, fig. 15 F. Records. South coast: CPR 46C. SAMH 255. SCD 75B, 84D. (Natal: SAMH 355.) Description. Colonies epizootic on Thecocarpus formosus (Busk) and Halicornaria arcuata (nominate subspecies); consisting of solitary hydrocladia and upright pinnate stems. Pinnate stems reaching 12-5 mm. in length and bearing hydrocladia with as many as 14 hydrothecae. Solitary hydrocladia bearing as many as 13 hydrothecae. The hydrothecae and nematothecae are very variable in structure, those on the distal ends of the hydrocladia of pinnate stems resembling those of the nominate subspecies (see Millard, 1958, fig. 15 F), but those on the proximal ends of the same hydrocladia and those on solitary hydrocladia being successively reduced towards the base of the colony (fig. 10 B-D). This reduction affects: (a) the paired lateral and the posterior thecal teeth which become ‘smoothed out’ until the margin is merely sinuated; (b) the median inferior nematotheca, which becomes shorter and no longer reaches the level of the thecal margin; thus it may have no free portion and only one opening; and (c) the lateral nematothecae which become reduced, with the two apertures combined into one. Gonophores absent. HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 309 Remarks. The same modifications were present in the holotype (Millard, 1958), but owing to the poor condition of the colony and the covering of silt it was not possible to determine whether the condition was normal or due to damage. Hfalicornaria exserta nov. sp. (Fig. 11 A-E) Types and records. Holotype: SAMH 169 (South African Museum registered number). Additional records: SCD 36N, 153Y. All from the south coast. Description (holotype). One upright stem 8-3 cm. in length and bearing gono- phores. Hydrorhiza of same diameter as stem, with thick perisarc. Stem unfas- cicled and unbranched. Proximal region (about 1:5 cm.) unsegmented and without hydrocladia or nematothecae. Middel region (about 1 cm.) divided by transverse nodes into rather irregular internodes which tend to become pro- gressively shorter, and which bear 1 or 2 median nematothecae each. Distal region divided into short hydrocladia-bearing internodes by slightly oblique nodes which slope in opposite directions so that the side of an internode bearing the hydrocladium is longer than the other. Hydrocladia alternate, borne on short apophyses of the stem, one to an internode, reaching a maximum length of 1-75 cm. Each divided into thecate internodes by transverse nodes. No internodal septa. Hydrotheca expanding to margin, with adcauline wall free from internode for a short region distally, and with abcauline wall bent strongly forwards. Margin with 1 prominent anterior tooth, 3 pairs of well-defined lateral teeth, and a low, rounded, posterior tooth. Hydropore with margin inturned. Median inferior nematotheca short (about 4 length of hydrotheca), adnate to hydrotheca for most of its length, with the terminal opening extending along all of the short free part, with no communication with thecal cavity. Lateral nematotheca ovoid, not reaching thecal margin or even beginning of free part, with 1 broad distal aperture. Cauline nematothecae similar to laterals: 1 inferior, 1 axillary anterior, and 1 axillary posterior. Gonophores borne on hydrocladial apophyses and forming a double row down front of stem, bowl-shaped, with very short pedicels and broad, truncated, distal ends. Measurements. See subspecies epizootica. Remarks. This species is related to H. sibogae Billard 1918 and H. expansa Jaderholm 1903, but differs in the unbranched stem, shorter hydrothecae, the presence of a posterior marginal thecal tooth, and other minor details. Halicornaria exserta epizootica nov. subsp. (Fig. 11 F—H) Holotype: SAMH 320 (South African Museum registered number). From south coast. 310 ANNALS OF THE SOUTH AFRICAN MUSEUM Description. Hydrorhiza epizootic on back of stem and hydrocladia of Thecocarpus flexuosus (Lamx.) umbellatus, giving off solitary hydrocladia and upright pinnate stems. Solitary hydrocladia arising alternately, reaching a maximum height of o-5 cm. and bearing up to 10 hydrothecae, forming an angle of about 45° with hydrorhiza. The two rows of hydrocladia not in the same plane but shifted towards posterior surface of host. All hydrothecae facing same direction as those of host. Pinnate stems reaching a maximum height of 0-9 cm. and bearing alternate hydrocladia as in nominate subspecies. Structural details similar to those of normal form except that the hydro- cladial internodes tend to be longer and narrower and the hydrothecae are not so strongly bent forwards. Median and lateral nematothecae tending to be reduced in size towards the proximal region of solitary hydrocladia, so much so that the median nematotheca may not even reach the base of the hydrotheca. Gonophores absent. Measurements (mm.) Halicornaria subsp. exserta eprzootica SAMH 169 SCD36N SAMH 320 Hydrocladium, internode length 0°24-0°39 0°29-0°37 0°33-0°45 diameter at distal end O°15—0°23 O°19-0°26 =0:09-0°15 Hydrotheca, length adcauline 0°24-0°34. 0°30-0°36 0°25-0°35 length adcauline, free part .. 0:03-0:06 0:03-0:06 0:04-0:08 length abcauline 0:28-0°34 0:34-0:36 0°22—-0°33 diameter at margin . 0:22-0'27 0:22-0:25 0:18-0-26 Median nematotheca, length O'10-0°I2 O°10—-0°12 0:06—0-09 Lateral nematotheca, length 0:09-0'I12 O°I0-0'12 0:05-0:09 Gonotheca, length 0:26 maximum diameter 0°42 Remarks. This subspecies has a similar growth-form to that of H. arcuata epizootica, and shows the same tendency for reduction of nematothecae in solitary hydrocladia. A-E. Subsp. arcuata. A, B. The stem, and the gonotheca, from the holotype, SAMH 169. C. Anterior view of stem showing the cauline nematothecae and scars for gonothecae (striped), from SCD 36N. D, E. Anterior and lateral views respectively of part of a hydrocladium from the holotype. F-H. Subsp. epizootica, from the holotype SAMH 320. F. Part of a colony on the back of a stem of Thecocarpus flexuosus umbellatus showing solitary hydrocladia and 2 pinnate stems. Hydrocladia of host cut off short. G, H. Anterior and lateral views respectively of part of a hydrocladium. eet 311 HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA q{ "wu G-O Hohe Gd =a SS BRS. Vee aI Le we 0 O—————OoO Fic. 11. Halicornaria exserta n. sp. 312 ANNALS OF THE SOUTH AFRICAN MUSEUM Flalicornaria ferlusi Billard, 1901 Halicornaria Ferlusi. Billard, 1907, p. 370, pl. 25 (fig. 8), fig. 14. Records. South coast: CPR 46B. (Additional from Natal: NAD 1R.) Description. Stem reaching a maximum height of 3 cm. Structural details exactly similar to those described by Billard. Male gonothecae present (in NAD 1R only), similar to those described by Jarvis (1922) for var. brevis, except that there is no aperture, and no ‘refrac- tive discs’ are visible. Each containing one gonophore. Remarks. This species is known only from Madagascar, and var. brevis Jarvis from Wasin, East Africa. This is the first record from South Africa. Lytocarpus filamentosus (Lamarck), 1816 Lytocarpus filamentosus. Millard, 1957, p. 241. Millard, 1958, p. 220. Records. South coast: MB 15A, 24S, 26D, 39V, 47J, 52B, 55E, 58B, 70F, 72F, 83A. LIZ 7M, 13B, 27B. CPR 46E. TRA 42B. SAMH 150, 259: SCD SAE; o4k; 117, 169); Thecocarpus flexuosus (Lamx.), 1816 (Bigey 2) Aglaophenia flexuosa Lamouroux, 1816, p. 167. Kirchenpauer, 1872, p. 25. Aglaophenia plumifera Kirchenpauer, 1872, p. 31, pl. 1 (fig. 6), pl. 3 (fig. 6). Stechow, 1923b, Pp- 255, fig. K1. Thecocarpus giardi Billard, 1907, p. 381, pl. 25 (figs. 9-10), pl. 26 (figs. 11-17), fig. 21. Vervoort, 1946, p. 335. Millard, 1957, p. 240. Millard, 1958, p. 221, fig. 16 A-C. Thecocarpus flexuosus. Billard, 1909, p. 330. ? Aglaophenia(?) bifida Stechow, 1923a, p. 117. Stechow, 1925, p. 515, fig. 53. Diagnosis. Stem thick, woody and fascicled, subdividing and giving rise to ‘branches’ formed from sympodially branching pinnae. Each pinna giving rise to a subsequent one from its anterior surface, the whole ‘branch’ twisted in a spiral manner and with a geniculate main axis formed from the proximal parts of successive pinnae. Each pinna consisting of a basal part bearing 1 large median nematotheca on each internode, and a distal part bearing 1 hydrocladium and 3 nematothecae on each internode. Hydrocladia alternately arranged, divided into internodes by slightly oblique nodes, each internode typically with 2 internodal septa. Hydrotheca adnate, deep, expanding to margin, with an intrathecal septum near base, and with 9 marginal teeth (1 median and 4 pairs lateral). Median nematotheca short, reaching approximately to level of intrathecal septum, with about + length free. Lateral nematotheca reaching to thecal margin. Corbula replacing hydrocladium, with pedicel of 1-6 thecate internodes, elongated, closed. Each rib of corbula bearing a series of nematothecae along outer edge, a hydrotheca near the base, and a crested HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 313 process of variable length below it. Base of corbula with a rounded projection facing towards pedicel. Remarks. This species is endemic to southern Africa, where it appears to be widely distributed around the coast extending as far north as Madagascar on the east. It is extremely variable in its growth form and in its microscopic structure, necessitating the establishment of a number of subspecies. The charac- ters of a subspecies are generally constant in a sample from a particular locality, though the occurrence of occasional intergrading specimens shows that all the material is conspecific. The species can always be recognized by the spiral arrangement of the pinnae, by the method of branching (described in detail by Billard (1907), and Millard (1958) under the name of T. gzardi) and by the structure of the corbula. It has already been shown (Millard, 1958) that the keel-like ridge along the abcauline wall of the hydrotheca (i.e. the outer point of the median tooth) may be hollow or solid. Material from the Agulhas Bank now shows that this ridge may be very poorly developed (so that the outer point of the median tooth may be shorter than the inner one), or absent altogether. Material with the latter arrangement has been placed in a separate subspecies, but amongst it rare hydrothecae occur showing the beginnings of a ridge, i.e. the first bifur- cation of the median tooth. A series of hydrothecae can thus be arranged (fig. 12, C-M) showing the progressive development of this structure culminat- ing in the large hollow spine of subsp. flexuosus and subsp. umbellatus. The structure of the hydrotheca in the ‘keelless’ variety (fig. 12 C) closely resembles Stechow’s diagram (1923b, fig. Kt) of Kirchenpauer’s type material of Aglaophenia plumifera. The preserved material of the type of A. plumifera has unfortunately been lost, but Stechow states that the pinnae arise alternately and Kirchenpauer that they arise irregularly. Probably both are correct, for a series of alternate pinnae twisted in a spiral manner would give an irregular appearance. I have seen whole mounts of A. plumifera prepared by Stechow (loaned from the Munich Museum) including a complete pinna, and this shows absolute agreement with the ‘keelless’ variety including the presence of hinge-joints and the large median nematothecae on the internodes below them. The two are thus considered to be synonymous, and are given the status of a subspecies, under Kirchenpauer’s name plumiferus. Thecocarpus flexuosus plumiferus (Kirch.), 1872 (Fig. 12 €, D) Aglaophenia plumifera Kirchenpauer, 1872, p. 31, pl. 1 (fig. 6), pl. 3 (fig. 6). Stechow, 1923b, p- 255, fig. K?. Diagnosis. Pinnae arranged in a loose to tight spiral around a ‘main axis’ which is seldom definitely geniculate. Intervals between pinnae regular and approxi- mately equal. Basal and distal regions of pinnae separated by hinge-joints. 314 ANNALS OF THE SOUTH AFRICAN MUSEUM SS S\ GH ey \ LAF = SS a oy — Fic. 12. Thecocarpus flexuosus (Lamx.). HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 315 Hydrocladia short, 3-6 mm. in length. Hydrotheca with no abcauline ‘keel’, but abcauline wall somewhat thickened near margin. First and second pairs of marginal thecal teeth not bifid. Records. South coast: SCD 36L, 79H, 104G. SAMH 353. Description. One of the colonies (SAMH 353) very large, reaching 46 cm., and with base of stem 12 mm. thick. The spiral twisting of the ‘branches’ is sometimes difficult to recognize in the lower, heavily fascicled regions of the colony, where also the geniculate appearance is lost. It is, however, quite distinct in the distal regions. Unlike other subspecies a number of internodes below the hinge-joints may bear hydrocladia. Hinge-joints distal to origin of next pinna. Corbulae sometimes very long, reaching a maximum length of 20 mm. and bearing up to 39 pairs of ribs. pce also Pp. 313. Measurements. See under subsp. umbellatus. Thecocarpus flexuosus flexuosus (Lamx.) 1816 (Fig. 12 A, J-L) Thecocarpus giardi Billard, 1907, pp. 381-385, pl. 25 (fig. 9), pl. 26 (figs. 1 1-16), fig. 21. Vervoort, 1946, p. 335. Millard, 1957, p. 240. Millard, 1958, p. 221, fig. 16 A. ? Aglaophenia(?) bifida Stechow, 1923a, p. 117. Stechow, 1925, p. 515, fig. 53. Diagnosis. Pinnae arranged in a loose to tight spiral, around a geniculate ‘main axis’. Intervals between pinnae regular and approximately equal. No hinge-joints. Hydrocladia short, about 6 mm. in length. Hydrotheca with abcauline ‘keel’ extended beyond median tooth as a hollow spine. First and second pairs of marginal thecal teeth typically bifid. Records. South coast: KNY go0R. ie A. Subsp. flexuosus. Part of colony showing a spirally twisted ‘branch’ and its pinnae, from 30R. B. Subsp. umbellatus n. subsp. Part of colony showing one ‘branch’, from the holotype SCD 36M. C-M. Selected hydrothecae showing variations, particularly of the anterior ‘keel’. C, D. Subsp. plumiferus (Kirch.), a typical hydrotheca with no ‘keel’ from SCD 36L, and a rare hydrotheca with a slight indication of a ‘keel’, from SAMH 353. E-H. Subsp. solidus Millard, a rare terminal hydrotheca with the ‘keel’ barely distinct from the median tooth, from TRA 92C; a hydrotheca with well-developed ‘keel’ but restricted abcauline thickening, from TRA 92C; a typical hydrotheca with abcauline thickening reaching to base, from AFR 1028A (described by Millard 1958); and a rare hydrotheca with a hollow ‘keel’, from AFR 1028A (described by Millard, 1958). J-L. Subsp. flexuosus, a rare hydrotheca with a solid ‘keel’, from PF 396A (described by Millard, 1957); a typical hydrotheca with a hollow ‘keel’ and a bifid lateral tooth, from PF 12308B (described by Millard, 1958); and a hydrotheca with an extra long ‘keel’ pierced at distal end, from KNY 30R. M. Subsp. umbellatus n. subsp., a typical hydrotheca, very similar to K except that the lateral teeth are not bifid, from the holotype SCD 36M. 316 ANNALS OF THE SOUTH AFRICAN MUSEUM Thecocarpus flexuosus solidus Millard, 1958 (Fig. 12 E-H) Thecocarpus giardi Billard, var. solidus Millard, 1958, p. 222, fig. 16 B, C. Diagnosis. Pinnae arranged in a loose to tight spiral around a geniculate ‘main axis’. Intervals between pinnae regular and approximately equal. Basal and distal regions of pinnae separated by hinge-joints. Hydrocladia short, about 6 mm. in length. Hydrotheca with abcauline ‘keel’ forming a solid spine. First and second pairs of marginal thecal teeth not bifid. Records. South coast: TRA 92C. SCD 29F, 36K, 84E. SAMH igo. Remarks. One of these samples (TRA 92C) shows characters tending towards subsp. plumiferus (fig. 12 E, F). The abcauline ‘keel’ of the hydrotheca is poorly developed and restricted to the margin, the typical chitinous thickening which usually extends down to the junction of the median nematotheca being absent. In the same sample there is a number of branching hydrocladia. A hydro- cladium may produce as many as 7 subsidiary hydrocladia, and in rare cases the subsidiary hydrocladia branch too. The hydrocladial internodes which give rise to branches are without hydrothecae and similar in every respect to those of the distal region of the pinna. Thecocarpus flexuosus umbellatus nov. subsp. (Fig. 12 B, M) Diagnosis. Pinnae (usually 5) arranged in a very close spiral forming an umbel- like cluster at the end of a long ‘stem’ formed by the proximal region of the first pinna. Hinge-joints present on first pinna only. Hydrocladia long (reaching 15 mm.). Hydrotheca with abcauline ‘keel’ extending beyond median tooth as a hollow spine. First and second pairs of marginal thecal teeth not bifid. Types and records. Holotype: SGD 36M (South African Museum registered num- ber — sa Wie o2)e Further records, south coast: SCD 33B, 117M, 153C. SAMH 168, 258, 286, 305, 3109. Description of holotype. Colony reaching a height of 12:3 cm. Stem fascicled and dividing near base to give the branching pinnae or ‘branches’. First pinna with basal part long (several cm.), usually lightly fascicled (though the peripheral tubes do not extend on to the distal part), divided into internodes each with 1 large median nematotheca with several (3-6) openings. This part terminated by 2 oblique hinge-joints, the internode between them bearing 2 large median nematothecae. Distal part bearing 2 hydrocladia on the first internode and 1 on each of the rest, and giving origin to the next pinna from the anterior surface of the first to fifth internode. HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 317 Second pinna facing towards the first, seated on a short apophysis, with no basal part and no hinge-joints, consisting only of hydrocladia-bearing inter- nodes, and giving rise to the third pinna from its second internode. Subsequent pinnae (usually 3) as second, and the ‘main axis’ (formed by the origins of consecutive pinnae) twisted in a spiral manner. Hydrocladia alternate, long, reaching 15 mm. Hydrotheca similar in structure to subsp. flexuosus except that the first and second lateral teeth are not bifid. Hydrothecal depth and hydrocladial internode length varying from that found in subsp. flexuosus to that in subsp. perarmatus Billard 1907. Corbula (not from holotype) with up to 20 ribs and reaching 12 mm. in length. In some cases there is an enormous development of the terminal parts of the ribs and the ‘crests’ arising from their bases, giving to the whole a spidery appearance similar to that sometimes encountered in T. formosus. Measurements (mm.) subsp. subsp. Aglaophenia umbellatus plumiferus plumifera* _Hydrocladium, internode length .. O*19-0°32 0:28-0:37 0:28-0°30 Hydrotheca, height (to tip of inner point of median tooth) .. si .. 0°22-0°30 0°23-0°31I 0°24-0°27 diameter at margin (inside) .. O°16-0-2I 0-14-0°205 0°13-0°18 Median nematotheca, length .. .. O°14-0°24 O°13-0°2I 0°15-0°18 length of free part .. a: .. 0°03-0:06 0:03-0:04 0:03-0:04 Corbula, length .. ay, reaching 8-50 20°00 diameter . ae reaching 1-40 ge Remarks. Although the arrangement described above is remarkably constant, variations do occur (e.g. longer intervals between origins of pinnae) which show that this form is not worthy of specific rank. The top part of the colony breaks easily at the hinge-joints and many specimens brought up by the dredge consist of crowns only. The pinnae often terminate in long tendrils. Thecocarpus formosus (Busk), 1851 Thecocarpus formosus. Millard, 1958, p. 221. Maillard, 1961, p. 206. Records. South coast: MB 24Y, 47F. KNY 30S, 57K, 71A (recorded by Day, Millard and Harrison, 1952). LIZ 7N, 27A. BMR 12K. L 55. CPR 46D. SCD 20A, 29E, 36H, 50B, 75A, 84B, 94D, 117N, 1698S. SAMH 200, 210, 254, 285, 299. * Measurements from Stechow’s slides borrowed from the Munich Museum. 318 ANNALS OF THE SOUTH AFRICAN MUSEUM SUMMARY A total of 35 species of plumularian hydroids is described from the south and west coasts of South Africa. Among these are 4 new species and 4 new records from South Africa. A new subfamily, the Halopterinae, is created for genera with cauline hydrothecae and related forms, and possible evolutionary trends within it discussed. The genus Antennopsis is sunk in the synonymy of Nemertesia, and a new genus, Corhiza, created. The genus Paragattya is sunk in the synonymy of Gattya. Various problems of nomenclature are dealt with, including that of Thecocarpus flexuosus, and the subspecies of the latter are revised. REFERENCES ALLMAN, G. J. 1877. Report on the Hydroida collected during the exploration of the Gulf Stream by L. F. De Pourtalés, assistant United States Coast Survey. Mem. Mus. comp. Kool. Harv. 5, 1-66. AtuimaNn, G. J. 1883. Report on the Hydroida dredged by u.m.s. Challenger during the years 1873-76. Part I. Plumularidae. Challenger Rep. Zool. 7, 1-54. ALLMAN, G. J. 1886. Description of Australian, Cape and other Hydroida, mostly new, from the collection of Miss H. Gatty. 7. Linn. Soc. (Zool.) 19, 132-161. Bate, W. M. 1884. Catalogue of the Australian hydroid zoophytes. Sydney: Australian Museum. Bate, W. M. 1888. On some new and rare Hydroida in the Australian Museum collection. Proc. Linn. Soc. N.uS.W. (2), 3) 745-799- . Bate, W. M. 1914a. Report on the Hydroida coHected in the Great Australian Bight and other localities. ool. Res. Fish. Exp. ‘Endeavour’ 2, 3-62. BALE, W. M. 1914b. Report on the Hydroida collected in the Great Australian Bight and other localities. Part II. ool. Res. Fish. Exp. ‘Endeavour’ 2, 166-188. Bate, W. M. 1915. Report on the Hydroida collected in the Great Australian Bight and other localities. Part III. Zool. Res. Fish. Exp. ‘Endeavour’ 3, 241-336. Bepot, M. 1921. Notes systématiques sur les plumularides. 2me Partie. Rev. suisse Zool. 29, I-40. Bepot, M. 1923. Notes systématiques sur les plumularides. 3me Partie. Rev. suisse Zool. 30, 213-243. BILLARD, A. 1906. Hydroides. Expéd. sci. ‘Travailleur’ et du ‘Talisman’ 8, 153-244. BILLARD, A. 1907. Hydroides de Madagascar et du sud-est de l’Afrique. Arch. Zool. exp. gén. (4), 72 335-396. BILLARD, A. 1909. Revision des espéces types d’hydroides de la collection Lamouroux, con- servée a l'Institut botanique da Caen. Ann. Sci. nat. zool. (9), 9, 307-336. BILLARD, A. 1910. Revision d’une partie de la collection des hydroides du British Museum. Ann. Sci. nat. zool. (9), 115 1-67. BILLARD, A. 1913. Les hydroides de l’expédition du Siboga. I. Plumulariidae. Siboga Exped. monogr. 7a, I-I15. BILLARD, A. 1918. Notes sur quelques espéces d’hydroides de l’expédition du ‘Siboga’. Arch. Kool. exp. gén. 5°75 21-27. Briccs, E. A. 1914. Notes on Tasmanian Hydrozoa. 7. roy. Soc. N.S.W. 48, 302-318. Brocu, H. 1914. Hydrozoa benthonica. Beitr. Meeresfauna Westafr. 15 19-50. Brocu, H. 1933. Zur Kenntnis der adriatischen Hydroidenfauna von Split. Skr. norske Vidensk- Akad., Mat.-natur. Kl. 1933, no. 4, I-115. Day, J. H., Mittarp, N. A. H. & Harrison, A. D. 1952. The ecology of South African estuaries. Part III. Knysna: a clear, open estuary. Trans. roy. Soc. S. Afr. 33, 367-413. Day, J. H. 1959. The biology of Langebaan Lagoon: a study of the effect of shelter from wave action. Trans. roy. Soc. S. Afr. 353 475-547- HYDROZOA OF SOUTH AND WEST COASTS OF SOUTH AFRICA 319 Eyre, J., BRoEKHUYSEN, G. J. & Cricuton, M. I. 1938. The South African intertidal zone and its relation to ocean currents. VI. The East London district. Ann. Natal Mus. 9, 83-111. Hincxs, T. 1868. A history of the British hydroid zoophytes. 2 vols. London. JADERHOLM, E. 1896. Ueber aussereuropaische Hydroiden des zoologischen Museums der Universitat Upsala. Bzh. svensk. VetenskAkad. Handl. 21, 3-20. JADERHOLM, E. 1903. Aussereuropaische Hydroiden im schwedischen Reichsmuseum. Ark. Zool. I, 259-312. JADERHOLM, E. 1919. Zur Kenntnis der Hydroidenfauna Japans. Ark. Zool. 12, no. 9, 1-34. Jarvis, F. E. 1922. The hydroids from the Chagos, Seychelles and other islands and from the coasts of British East Africa and Zanzibar. Trans. Linn. Soc. Lond. Zool. (2), 18, 331-360. KiIRCHENPAUER, G. H. 1872. Ueber die Hydroidenfamilie Plumularidae, einzelne Gruppen derselben und ihre Fruchtbehalter. 1. Aglaophenia Lx. Abh. Naturw. Hamburg 5, 1-52. KiIRCHENPAUER, G. H. 1876. Ueber die Hydroidenfamilie Plumularidae, einzelne Gruppen derselben und thre Fruchtbehalter. II. Plumularia und Nemertesia. Abh. Naturw. Hamburg 6, 1-59. Lamouroux, J. V. F. 1816. Histoire des polypiers coralligénes flexibles, vulgairement nommés zoophyies. Caen. Leloup, E. 1939. Notes sur quelques hydropolypes exotiques. Bull. Mus. Hist. nat. Belg. 15, 1-19. MARKTANNER- TURNERETSCHER, G. 1890. Die Hydroiden des k.k. naturhistorischen Hofmuseums. Ann. naturh. (Mus.) Hofmus. Wien 5, 195-286. Miiiarp, N. A. H. 1957. The Hydrozoa of False Bay, South Africa. Ann. S. Afr. Mus. 43, 173-243- Mitiarp, N. A. H. 1958. Hydrozoa from the coasts of Natal and Portuguese East Africa. Part I. Calyptoblastea. Ann. S. Afr. Mus. 44, 165-226. Mitrarp, N. A. H. 1959a. Hydrozoa from the coasts of Natal and Portuguese East Africa. Part II. Gymnoblastea. Ann. S. Afr. Mus. 44, 297-313. Mirarp, N. A. H. 1959b. Hydrozoa from ships’ hulls and experimental plates in Cape Town docks. Ann. S. Afr. Mus. 45, 239-256. Mitxarp, N. A. H. 1961. A report on Busk’s collection of South African hydroids. Ann. Mag. nat. Hist. (13), 4, 203-208. Nuttinec, C. C. 1900. American hydroids. Part I. The Plumularidae, Spec. Bull. U.S. nat. Mus. 4, 1-285. Ritcur, J. 1907. The hydroids of the Scottish National Antarctic Expedition. Trans. roy. Soc. Edinb. 45, 519-545. RitcutE, J. 1909. Supplementary report on the hydroids of the Scottish National Antarctic Expedition. Trans. roy. Soc. Edinb. 47, 65-101. StecHow, E. 1912. Hydroiden der Miinchener zoologischen Staatssammlung. Zool. Fb. 325 333-378. STECHOW, E. 1923a. Ueber Hydroiden der deutschen Tiefsee-Expedition, nebst Bemerkungen liber einige andre Formen. ool. Anz. 56, 97-119. StEcHow, E. 1923b. Zur Kenntnis der Hydroidenfauna des Mittelmeeres, Amerikas und anderer Gebiete. II Teil. Zool. 7b. 47, 29-270. STECcHOW, E. 1925. Hydroiden der deutschen Tiefsee-Expedition. Wiss. Ergebn. ‘Valdivia’ 17, 383-546. STEPHENSON, T. A., STEPHENSON, A. & pu Torr, C. A. 1937. The South African intertidal zone and its relation to ocean currents. I. A temperate Indian Ocean shore. Trans. roy. Soc. S. Afr. 24, 341-382. STEPHENSON, T. A., STEPHENSON, A. & Bricut, K. M. F. 1938. The South African intertidal zone and its relation to ocean currents. IV. The Port Elizabeth district. Ann. Natal Mus. 9, I-19. Totron, A. K. 1930. Coelenterata. Part V. Hydroida. Nat. Hist. Rep. Terra Nova Exped. 5, 131-252. VerRvoorT, W. 1946. Exotic hydroids in the collections of the Rijksmuseum van Natuurlijke Historie and the Zoological Museum at Amsterdam. Zool. Meded. 26, 287-351. VeERVooRT, W. 1959. The Hydroida of the tropical west coast of Africa. Atlantide Rep. 5, 211-325. Warren, E. 1908. On a collection of hydroids, mostly from the Natal coast. Ann. Natal Mus. 1, 269-355. WarREN, E. 1914. On the development of the planula in a certain species of plumularian hydroid. Ann. Natal Mus. 3, 83-102. References thus appear as follows: AUGENER, H. 1913. Polychaeta. In Michaelsen, W., ed. Bettrdége zur Kenntnis der Meeresfauna Westafrikas. 2, 67-625. Hamburg: Friederichsen. EKMAN, S. 1953. 2. Anthus vaalensis daviest Roberts Anthus daviesi Roberts, 1914, 172— Matatiele. Larger than vaalensis and with less sexual dimorphism of size. Range: ‘The higher parts of the Cape Province, from Matatiele to Nieu- woudtville and the Cold Bokkeveld, north to Philipstown and Colesberg. Occasional birds from as far north as Ngoma, in Northern Rhodesia, are inseparable. Measurements: 6 3s; wing 108-113 mm., av. 112°0 mm.; tail 70-83 mm., av. 75°3 mm.; culmen 18-20 mm., av. 19°5 mm.; tarsus 23-30 mm., av. 27°6 mm.; hind-claw 9-10 mm., av. 9°6 mm. I 9; wing 105 mm.; tail 71 mm.; culmen 20 mm.; tarsus 26 mm.; hind- claw II mm. Material examined: South African Museum 4 (Nieuwoudtville, Cold Bokkeveld, Hanover, Glen Grey); British Museum 1 (Deelfontein) ; Transvaal Museum 1 (Matatiele (type, daviesi)); Albany Museum 1 (Colesberg) ; Durban Museum 1 (Steynsburg); total, 8. 3. Anthus vaalensis clanceyt subsp. nov. Decidedly greyer than vaalensis or daviesi; intermediate in size; virtually no sexual dimorphism of size. 348 ANNALS OF THE SOUTH AFRICAN MUSEUM Type: Ad 4, in Durban Museum: ‘Glen Lyon’, Glen, Modder River, Bloemfontein; collected by Durban Museum personnel, 8 June, 1960. I name this subspecies after Mr. P. A. Clancey, who first drew my attention to it. Range: The Central Orange Free State, from Bloemfontein to (?) Vrede- fort Road.* Measurements: 8 gs; wing 103-109 mm., av. 105°7 mm.; tail 72-79 mm., av. 74°7 mm.; culmen 18-19 mm., av. 18:5 mm.; tarsus 25-30 mm., av. 27:2 mm.; hind-claw 8-10 mm., av. 9:4 mm. 5 Qs; wing 102-111 mm., av. 106-8 mm.; tail 70-83 mm., av. 74:4 mm.; culmen 18-19 mm., av. 18-4 mm.; tarsus 25-28 mm., av. 26:4 mm.; hind-claw g-10 mm., av. 9°4 mm. Material examined: Durban Museum 6 (Glen (type, clanceyz)); British Museum 3 (Vredefort Road); Albany Museum 2 (Vredefort Road, Bloemfon- tein); East London Museum 2 (Glen); total, 13. 4. Anthus vaalensis exasperatus subsp. nov. Smaller than vaalensis or any other race (w. normally less than 100 mm.); little sexual dimorphism in size (w. 9, 99% of 3). Type: Ad 3, in National Museum of Southern Rhodesia, Bulawayo, No. 29147, 3 m. S. of Nata, Bechuanaland Protectorate; collected 28 January, 1957, collector’s number M.118. | Range: North-eastern Bechuanaland, from the Nata River to Francistown. Occasional males from Matabeleland and Northern Rhodesia are inseparable; and females are only distinguishable in series. Measurements: 4 gs; wing 91-98 mm., av. 95:0 mm.; tail 64-69 mm., av. 67-0 mm.; culmen 18 mm., av. 18-0 mm.; tarsus 25-29 mm., av. 26-7 mm.; hind-claw 10 mm., av. 10:0 mm. 7 Qs; wing QI-102 mm., av. 94:4 mm.; tail 60-71 mm., av. 67-9 mm.; culmen 17-19 mm., av. 18-1 mm.; tarsus 24-28 mm., av. 26-7 mm.; hind-claw G-1T im AV. Ora mate Note that females of this race are less obviously smaller than those of neighbouring populations (S.W.A., w. av. 94:7 mm.; Caprivi Strip, w. av. 95 mm.) than are the males, in correlation with the lesser sexual dimorphism of size shown by this race. Material examined: National Museum, Bulawayo 12 (Mache, Nata (type, exasperatus), 4m. 8S. of Francistown, Makarikari); total, 12. 5. Anthus vaalensis chobiensis (Roberts) Meganthus vaalensis chobiensts Roberts, 1932, 29— Kabulabula. Anthus leucophrys marungensis Chapin, 1937, 342—Kaschi, Marungu, Belgian Congo. Anthus vaalensis muhingae White, 1944, 6— Muhinga plain, Kasendoge, Katanga. * Since the above was written, Mr. Clancey has found specimens of this race in Griqualand West. S.A. SUBSPECIES OF THE BUFFY PIPIT ANTHUS VAALENSIS SHELLEY 349 Similar to A. v. vaalensis but smaller; larger than A. v. exasperatus, especially in the males. Range: From the Limpopo to western Nyasaland, Northern Province of Northern Rhodesia, Katanga and eastern Angola, except the north-eastern corner of Bechuanaland. Measurements: 99 3s; wing 90-109 mm., av. 101-9 mm.; tail 60-79 mm., av. 72°0 mm.; culmen 15-22 mm., av. 18-9 mm.; tarsus 21-29 mm., av. 24:2 mm.; hind-claw 7-12 mm., av. 9:1 mm. 78 Qs; wing 90-105 mm., av. 96-4 mm.; tail 60-78 mm., av. 69:4 mm.; culmen 16-21 mm., av. 18-4 mm.; tarsus 2I-30 mm., av. 25:7 mm.; hind-claw 7-1) mm, av. Ol mim. Material examined Southern Rhodesia: Durban Museum 1 (Tjolotjo); National Museum, Bulawayo 104 (Gazuma Pan, Wankie Game Reserve, Kana R., Dorset Siding, Hopefountain, Nyamandhlovu, Copleston, Tjolotjo, Lupane, Bulawayo, Shangani R., Matopos, Bembezi, Inseze, Limpopo R., Khami, Sebungwe, Umvuma, Darwendale, Headlands, Salisbury, Rusape, Inyanga, Sabi-Lundi confluence, Hunyani R., Selukwe); Albany Museum 2 (Matopos); total, 107. Northern Rhodesia: South African Museum 2 (Balovale, Lusaka) ; National Museum, Bulawayo 33 (Ngoma, Mazabuka, Choma, Lochinvar, Chuunga, Chilanga, Lusaka, Lukanga Swamp, Luanshya, Ntemwa-Lufupa confluence, Moshi-Lufupa confluence, Mweru Marsh, Mpika, Mwinilunga, Kasanga, Mporokoso, Kasusu, 8 m. N. of Nangweshi, Matabele Plain, Chiolola, Chunga Pools, Luete, Senenga) ; total, 35. Caprivi Strip: Transvaal Museum 3 (Kabulabula (type, chobdiensis), Kabuta); total, 3. 6. Anthus vaalensis neumanni Meinertzhagen Anthus leucophrys neumanni Meinertzhagen, 1920, 23—Ambaca, Angola. Decidedly redder than any other race. Range: South West Africa and western Angola. Some specimens from eastern Angola, the Rhodesias and Nyasaland are indistinguishable from this form. Measurements (S.W.A. birds only); 3 gs; wing 101-103 mm., av. 102:0 mm. ; tail 73-76 mm., av. 74:7 mm.; culmen 17-18 mm., av. 17°7 mm.; tarsus 24-26 mm., av. 25:0 mm.; hind-claw g-12 mm., av. 10-3 mm. (Hall, 1959, gives w. 93-105 for Angola birds, and White, 1948, 95-106 mm. for S.W.A. birds.) 3 9s; wing 92-98 mm., av. 94:7 mm.; tail 63-73 mm., av. 67:7 mm.; culmen 17-18 mm., av. 17:7 mm.; tarsus 27-28 mm., av. 27:7 mm.; hind-claw g-I2 mm., av. 10°3 mm. (Hall gives w. 89-102 mm. for Angola birds and White 94-99 mm. for S.W.A. birds.) 350 ANNALS OF THE SOUTH AFRICAN MUSEUM Material examined: Durban Museum 4 (Omaruru); British Museum 2 (Kamanjab, Franzfontein) ; total, 6. SUMMARY The forms of Anthus vaalensis are discussed, and six subspecies: vaalensis Shelley, daviest Roberts, clanceyt subsp. nov., exasperatus subsp. nov., chobiensis Roberts, and newmanni Meinertzhagen, are recognized. A statistical analysis is made of the measurements of the specimens. ACKNOWLEDGEMENTS I am deeply indebted to Mr. A. N. Rowan for statistical analysis of the measurements; to Mr. P. A. Clancey for drawing my attention to the peculiar series from the Orange Free State; and to Miss A. Schweitzer for the map. REFERENCES Amapon, D. 1949. The Seventy-five per cent Rule for Subspecies. The Condor 51, 250-8. Benson, C. W. 1953. A Check List of the Birds of Nyasaland. Blantyre. Benson, C. W., AND Wuirte, C. M. N. 1957. A Check List of the Birds of Northern Rhodesia. Lusaka: Government Printer. Cuapin, J. P. 1937. The Pipits of the Belgian Congo. Rev. Zool. Bot. afr. 29, 336-45. Hatt, B. P. 1959. The Plain-backed Pipits of Angola. Bull. Brit. Orn. Cl. 79, 113-16. Hoescu, W., AND NIETHAMMER, G. 1940. Die Vogelwelt Deutsch-Siidwestafrikas. 7. Orn., Lpz. 88, Sonderh. Macpona.Lp, J. D. 1957. A Contribution to the Ornithology of Western South Africa. London: British Museum (Natural History). McLacuian, G. R., AND LIVERSIDGE, R. 1957. Roberts’ Birds of South Africa. Cape Town: South African Bird Book Fund. MEINERTZHAGEN, R. 1920. New subspecies of birds. Bull. Brit. Orn. Cl. 41, 19-25. RosBeErts, A. 1914. Notes on birds in the collection of the Transvaal Museum with descriptions of several new sub-species. Ann. Transv. Mus. 4, 169-79. Roserts, A. 1932. Preliminary descriptions of sixty-six new forms of South African birds. Ann. Transv. Mus. 15, 21-34. Roserts, A. 1940. The Birds of South Africa. London: Witherby. ScLATER, W. L. 1930. Systema Avium A&thiopicarum. London: British Ornithologists Union. SHELLEY, G. E. 1900. The Birds of Africa, vol. 2. London: R. H. Porter. SMITHERS, R. H. N. 1959. Notes on the Avifauna of the South-West Kalahari Area of Bechuana- land Protectorate. Proc. Pan. Afr. Orn. Cong. Ostrich, Suppl. 3, 126-32. SMITHERS, R. H. N., Irnwin, M. P. S., anp PATERsON, M. 1957. A Check List of the Birds of Southern Rhodesia. Bulawayo: Rhodesian Ornithological Society. SmiTHERs, R. H. N., AND Paterson, M. 1959. Notes on the Avifauna of the Eastern Sector of the Great Makarikari Lake, Bechuanaland Protectorate. Proc. Pan. Afr. Orn. Cong. Ostrich, Suppl. 3, 133-43. VauriE, C. H., Wuits, C. M. N., Mayr, E., AND GREENWAY, J. C. 1960. Family Motacillidae. In Check List of Birds of the World. g (ed. Mayr and Greenway). Cambridge, Mass.: Museum of Comparative Zoology. VINCENT, J. 1952. A Check List of the Birds of South Africa. Cape Town: South African Bird Book Fund. Wuirte, C. M. N. 1944. New Races of Lark, Pipit and Cordon Bleu from Northern Rhodesia, Angola, and the Belgian Congo. Bull. Brit. Orn. Cl. 65. Wuire, C. M. N. 1948. The African Plain-backed Pipits—a Case of Sibling Species. The Ibis go, 547-53: Wurite, C. M. N. 1961. A Revised Check List of African Broadbills, Pitias, Larks, Swallows, Wagtatls and Pipits. Lusaka: Government Printer. S.A. SUBSPECIES OF THE BUFFY PIPIT ANTHUS VAALENSIS SHELLEY 351 APPENDIX STATISTICAL ANALYSIS OF THE NUMERICAL DATA By A. N. Rowan In his study of Anthus vaalensis, Winterbottom measured the lengths of wing, tail, culmen, tarsus and hind claw in 258 museum specimens, of which 150 were males. These data, arranged geographically, were submitted for statistical analysis. As a first step, male wing lengths were tested for homo- geneity using the F test, and were found to be very significantly heterogeneous, thus indicating that there were statistically valid differences between various populations. A preliminary study of the figures suggested that male wing lengths tended to decrease with decreasing latitude, and the possible existence of a cline was examined by plotting these two measurements against each other. However, the correlation coefficient was not significantly different from zero, showing that a cline did not exist. From a further inspection of the data it appeared (1) that males from northern Bechuanaland had smaller wings and tails than other groups; and (2) that the entire series could be separated into two groups, one with larger dimensions south of the Limpopo River, and the other comprising smaller birds north of that river. Accordingly, the ‘t’ test was employed to test for significant differences between the postulated northern and southern groups and between them and the birds of northern Bechuanaland. Tables A and B give the results for wing, tail, tarsus and culmen lengths. The dimensions of the hind claw were omitted from this analysis since they showed no clear relation to other measurements and the over-all differences in length were too small in relation to the units of measurement employed. From Table A it is clear that southern males are very significantly larger than their northern counterparts in all measurements and that southern females are significantly larger in all measurements except culmen length. TABLE A: COMPARISONS BETWEEN NORTHERN AND SOUTHERN BIRDS AVERAGES Southern Northern Character Sex group CB: group t Probability mm. mm. Wing length Male 105-80 0-60, 101-76 2°58 0:01-0:02 Female 101 °33 0:70 96°53 >6 Far less than 0-001 Tail length Male 74°50 0°45 71°64 3°40 Less than 0-:01% Female 71°60 0°25 69°78 2°24 0:*02-0°05 Length of tarsus Male PG 5 0°33 26-06 3:26 Less than o-o1 Female 26°63 0°34 25°59 2°53 O°OI-0'02 Length of culmen Male 19°23 0:27 18-72 2°47 O°OI-0°02 Female 18-63 a-12 18-39 1-11 Not significant 352 ANNALS OF THE SOUTH AFRICAN MUSEUM These results were further examined using the Coefficient of Difference, as recommended by Mayr et al. (Methods and Principles of Systematic Xoology, New York, 1953: 146). Table A shows the C.D. for various combinations. The conventional level of subspecific difference is 1:28. None of the coefficients of difference approaches this figure. Thus, using this criterion, it would appear that no single character is sufficient to indicate subspecific difference, but if all of them are considered together a good case might be made for subspecific difference between the populations north and south of the Limpopo. Table B gives similar figures for northern Bechuanaland birds and those of the northern groups. It can be seen that while the northern males are larger than the northern Bechuanaland birds in wing, tail and culmen length the difference is highly significant only in wing length and moderately significant in tail length. The females of the two groups are not significantly different. In no case does the coefficient of difference reach the conventional level for sub- specific difference. In view of the few specimens of northern Bechuanaland birds examined and because only male wing length is very significantly larger, it is considered very doubtful whether they should be separated on the mensural data alone. TABLE B: COMPARISONS BETWEEN NORTHERN BIRDS AND NORTHERN BECHUANALAND BIRDS AVERAGES Northern Character Sex Northern Bechuanaland t Probability birds birds mm. mm. Wing length Male 101+ 76 95°00 3°51 Less than 0-01 Female 96-53 94°23 1:76 Not significant Tail length Male 71°64 67:00 2°58 0-'OI-0:02 Female 69:78 67-86 1:50 Not significant Length of tarsus Male 26-06 26°75 0:82 Not significant Female 25°59 26°71 1°41 Not significant Length of culmen Male 19°72 18-00 1:24 Not significant Female 18-39 18-14 0°65 Not significant [ CONTINUED FROM INSIDE FRONT COVER References thus appear as follows: AUGENER, H. 1913. Polychaeta. In Michaelsen, W., ed. Beitrdége zur Kenntnis der Meeresfauna Westafrikas. 2, 67-625. Hamburg: Friederichsen. EKMAN, S. 1953. » se J. D. Gilchrist WCD 73-75 ~=8/10/60 =34°25’S./17°36’E. 1240 on. S.M. 393 Ann. S. Afr. Mus. 46 (14), 1963, 353—371, 3 figs. eo || 354 ANNALS OF THE SOUTH AFRICAN MUSEUM These are not the first deep-water collections made in this area. In November 1921 the late Professor Gilchrist occupied several abyssal stations off Cape Town. The Polychaeta from these stations were described by McIntosh (1925). The total list of species recorded from depths exceeding 1,000 metres in this area is shown below. Species McIntosh 1925 Present records Aphrodita alta Kbg. x Hermonia hystrix (Sav.) x Laetmonice benthaliana Mcl. x = L, filicornis (Mcl. non Kbg.) x Eunoe assimilis Mcl. Eunoe macrophthalma Mcl. x Harmothoe profunda n. sp. x Harmothoe sp. Euphione elizabethae Mcl. x Macellicephala mirabilis Mcl. Leanira hystricis Ehl. x Leanira tetragona (Oersted) Nereis (Neanthes) papillosa n. sp. Nephthys (Aglaophamus) macroura Schm. Nephthys hystricis Mcl. Nephthys paradoxa Malmgren Hyalinoecia sp. Lumbrineris brevicirra (Schm.) Lumbrinerts magalhaensis Kbg. Paraonis gracilis oculata Hartman Aricidea suecica simplex nov. Prionospio steenstrupi. Malmgren Spiophanes soderstromt Hartman Poecilochaetus serpens Allen Notomastus latericeus Sars. x Maldanella fibrillata Chamberlin Lumbriclymene minor Arwidsson Phalacrostemma elegans Fauvel Neosabellides elongatus (Ehl.) Amphitrite cirrata Miller Streblosoma abranchiata n. sp. Streblosoma chilensis (MclI.) Terebellides stroemi Sars Pista cristata (Miiller) x x x x Wee Xe Oe DS Pe ee He OG Ow eS eK OG XK x The list includes 31 identified species of which 4 are new. A further 9 species are new records for South Africa; in fact only 4 of the 31 species listed are known to occur at depths less than 100 metres in South African seas. Many POLYCHAETE FAUNA OF SOUTH AFRICA 355 of the others such as Laetmonice benthaliana, Euphione elizabethae, Macellicephala mirabilis, Maldanella fibrillata and Phalacrostemma elegans are abyssal species from such widely separated regions as the sub-Antarctic, the North Pacific, New Zealand, the Gulf of Panama and Madeira. This does not indicate any. affinity between the faunas of these areas, merely that abyssal species are very widely distributed and poorly known. It also emphasizes the importance of indicating the depth when listing distribution records. The study of this collection was facilitated by side-by-side comparisons with the rich collections housed in the British Museum. I wish to thank my many friends at the British Museum for their help while I worked there and the Oppenheimer Memorial Trust which financed my research. The material described in this report has been deposited in the South African Museum, Cape Town, and the registration number of each species is given in the text. SYSTEMATIC SECTION Family Aphroditidae Sub-family Hermoninae Genus Laetmonice Kinberg 1855 Both L. filicornis Kbg. and L. producta var. benthaliana MclI. have been recorded from South Africa but the published descriptions do not indicate clearly the differences between them. Before the new material was identified specimens of Laetmonice in the British Museum were examined. The results are set out below. Laetmonice filicornis Kinberg 1855 (type locality Sweden) Material examined: British Museum specimen 1865 :9 : 23:13 from Bohuslan, Sweden, and 1921 : 5 : 1 : 94-100 from the Shetland Islands. The two samples are conspecific and the following description is based on the Bohuslan specimen. Length 30 mm. for 32 segments. Ommatophores not pigmented. Tentacular segment without purse-shaped lobes at the sides of the prostomium. 15 pairs of elytra. A felt present. Harpoon-setae numerous and much larger than the notopodial spines of cirrigerous feet. Ventrum covered with rounded papillae. Ventral cirri well developed, their tips reaching the bases of the inferior neurosetae. Laetmonice producta Grube 1877 (type locality Kerguelen) Material examined: British Museum specimens 1941 : 3 : 3 : 439-50 from Kerguelen. Length 40-100 mm. for 42-46 setigers. Each ommatophore with a small eye-spot. Tentacular segment with a pair of purse-shaped lobes at the sides of the prostomium. 18 pairs of elytra. No felt. Harpoon-setae rare and the notopodial spines of cirrigerous feet as strong as those of elytrigerous feet. Ventrum covered with rounded papillae. Ventral cirri are well developed and reach the bases of the inferior neurosetae. 356 ANNALS OF THE SOUTH AFRICAN MUSEUM Laetmonice producta var. wyville: McIntosh 1885 Material examined: British Museum type 1885: 12:1 : 34 dredged off Prince Edward Island, and type 1885 : 12:1 : 35 from Challenger station 157 in the Antarctic. The first is a specimen of L. producta differing from Kerguelen specimens only in the possession of smaller lobes on either side of the prostomium and indistinct eyes. All other characters agree. The second, chosen as the lectotype of L. producta var. wyvillei, may be described as follows: Length 40 mm. for 42 segments. Ommatophores not pigmented. Tenta- cular segment without lobes at the sides of the prostomium. 18 pairs of elytra. Felt poorly developed. Harpoon-setae well developed and stronger than the notopodial spines of cirrigerous feet. Ventrum with a few papillae. Ventral cirri are small and do not reach the bases of the inferior neurosetae. Laetmonice benthaliana McIntosh 1885 Material examined: ‘Types of L. producta var. benthaliana in the British — Museum numbered 1885 :12:1:1 from Challenger station 147 off Prince Edward Island (lectotype). 1885 : 12:1 :36 from Challenger station 244 in the Pacific. 1885 : 12: 1:37 from Challenger station 241 in the North Pacific. Types of L. producta var. willemoest in the British Museum numbered 1885 : 12:1 : 31 from Challenger station 169 off New Zealand. 1885 : 12 : 1 : 32 from Challenger station 146 off Prince Edward Island. The lectotype measures 35 mm. for 32 segments. Ommatophores not pigmented. Tentacular segment without lobes at the sides of the prostomium. 15 pairs of elytra. Felt present but scanty (? lost). Harpoon-setae common and markedly larger than the notopodial spines of cirrigerous feet. Ventrum smooth. Ventral cirri are very small and do not reach the bases of the inferior neurosetae. Key To Laetmonice producta, producta wyvillet, filicornis AND benthaliana 1. Body with 42 or more segments; 18 pairs of elytra Li s) 3 ae oe hee Body with about 32 segments; 15 pairs of elytra 2. Purse-shaped lobes at the sides of the prostomium. No felt. Notopodial Seiiaes ae cirri- gerous feet as strong as those of elytrigerous feet ie ie .. L. producta No purse-shaped lobes. A scanty felt. Notopodial spines of sinieecous feet weaker than those of elytrigerous feet .. : Ae L. producta wyvillei 3. Ventrum covered with rounded papillae. Tree of ventral cirri reach bases of inferior neurosetae : si .. L. filicornis Ventrum smooth or ih a aur aerate at the pee ot the panned. Ventral cirri small and not reaching the base of the neurosetae .. a A a L. benthaliana ReEcorps oF Laetmonice FROM SoUTH AFRICA Laetmonice benthaliana McIntosh 1885 Laetmonice producta var. benthaliana McIntosh 1885, p. 45, pl. 8 figs. 4-5, pl. 4 fig. 12. ? Laetmonice filicornis McIntosh 1925, p. 20. Records: Station A.1g1 (2) and A.193 (9)—S.A. Museum Register No. A.19778. POLYCHAETE FAUNA OF SOUTH AFRICA 357 Station A.316 (15), A.317 (2), A.318 (2), A.gig (4) and A.322 (1)— Register No. A.19959. Notes: This. species seems to be one of the commonest polychaetes at abyssal depths west of Cape Town, and these South African specimens agree closely with the type described above. McIntosh (1925) recorded L. filicornis from five stations in this area and from 4-10 fathoms in Saldanha Bay. The only two specimens from this collection, which are now in the British Museum (numbers 1924: 7 : 21 : 21-22) are Hermonia hystrix; possibly the missing specimens were L. benthaliana. At any rate the record of L. filicornis must be eliminated. Laetmonice producta wyville: McIntosh 1885 Laetmonice producta var. wyvillei McIntosh 1885, p. 44. Laetmonice producta (non Grube) Day 1934, p. 18. Notes: The two specimens recorded by me (Day, 1934) from deep dredgings off Portuguese East Africa have been re-examined. They are L. producta wyviller with 18 pairs of elytra, a smooth ventral surface and very small ventral cirri. Aphrodita alta Kinberg 1855 Aphrodita alta Kinberg 1857, p. 2, pl. 1 fig. 1 a-g. Day, 1960, p. 274. Records: Station A.319 (1 juvenile) —S.A. Museum Register No. A.19964. Sub-family Polynoinae Harmothoe profunda n. sp. (Fig. 1 a-e) Records: Station A.193 (2)—S.A. Museum Register No. A.19780. Stations A.316 (2) and A.318 (1) —Register No. A.1g9960. Description: The holotype from station A.193 is the only complete specimen. It is 16 mm. long for 33 segments. The dorsum is black or possibly purple when fresh but the elytra and parapodia are pale. The prostomium (fig. 1a) is bilobed with obvious frontal peaks and the anterior pair of eyes is large and half-way back on the sides of the head. The median antenna is missing and the laterals are small, about half as long as the prostomium, markedly tapered and obviously ventral in origin. The dorsal cirri are about as long as theneuro- | setae and like the antennae they are sparsely beset with long papillae. The whole length of the body is covered with 15 pairs of elytra. Individual elytra (fig. 1b, 6’) vary from rounded to broadly oval and each is densely covered with chitinous tubercles which increase gradually in size from the anterior to the posterior margin. The small anterior ones are bluntly conical, those near the centre are larger and stouter, and a few near the posterior margin are almost spherical. Some are smooth, others rugose and some even have small blunt projections. A few small soft papillae are to be found near the external margin and it might be said that the elytra are minutely fringed. 358 ANNALS OF THE SOUTH AFRICAN MUSEUM The notopodium is well developed (fig. 1c) and bears numerous sabre- shaped notosetae much stouter than the neurosetae. Each notoseta (fig. 1d, d’) bears numerous rows of strong serrations and has a short, flanged and bluntly pointed tip. The neuropodium is rather short with a triangular presetal lobe and rounded postsetal lip. The neurosetae (fig. 1e, ce!) have rather long blades with 15-20 rows of well-marked spinules. The tips are of normal length and bidentate, but the relative size of the two teeth varies. In superior setae the secondary tooth is not much smaller than the terminal one, but in the middle of the series the secondary tooth decreases and inferior setae have a strong terminal tooth and a weak secondary one, only half the length of the terminal. There are many resemblances between this species and H. exanthema var. bergstromt Monro 1936 from the Magellan area. An examination of the type in the British Museum proved that the main difference is in the character of the tubercles on the elytra. In Monro’s species there is a scattering of uniformly small, slightly curved conical tubercles and a few very large, soft, pedunculate vesicles but no intermediate forms. Moreover the antennae and cirri are more densely papillose, the notosetae have finer serrations and the neurosetae have longer naked tips and a smaller secondary tooth. Ss SSS = = PG EG Fic. 1. Harmothoe profunda. (a) Head. (6, 5’) Elytron and details of marginal tubercles. (c) Parapodium. (d, d’) Notoseta and details of tip. (e, e’) Neuroseta and details of tip. POLYCHAETE FAUNA OF SOUTH AFRICA 359 Harmothoe sp. Records: Station A.190 (1)—S.A. Museum Register No. A.19771. Notes: Although the specimen is too poorly preserved to give it a specific name it is obviously different from any other recorded from South Africa. The characters may be briefly summarized as follows. Antennae and cirri smooth. Elytra with minute conical tubercles anteriorly and 4-5 large cylindro- conical soft papillae on the external margin. Notosetae with small, close-set serrations and strongly flanged tips. Neurosetae with long, weakly spinulose blades and minutely bidentate or unidentate tips. Eunoe assimilis McIntosh 1925 Eunoe assimilis McIntosh 1925, p. 21, pl. 2 figs. 1-2, pl. 3 fig. 3. Records: Station A.193 (2)—S.A. Museum Register No. A.19782. Station A.316 (1) —Register No. A.19961. Station A.318 (1) —(doubtful identification) Register No. A.19962. Description: MclIntosh’s original description is incomplete and a re-examination of the type in the British Museum and the discovery of a complete specimen with elytra allow me to give a summary of the diagnostic characters. Body 20-30 mm. long with 36-37 segments, rather broad but tapered posteriorly. Body completely covered with 15 pairs of elytra. Colour generally pale, but the ventral cirri and outer margins of the elytra are tinged with purple. Prostomium with small frontal peaks but without eyes. Median antenna three times the prostomial length; laterals ventral in origin and about 1°5 times the prostomial length. Antennae and cirri appear smooth but are really clad with short papillae. Dorsal cirri are tapered and reach the ends of the neurosetae. Elytra large and delicate with smooth margins; the surface is glabrous apart from a small patch of tiny rounded tubercles near the anterior margin. Notosetae fairly numerous, each stout with close-set rows of weak serra- tions and a long, narrowly flanged and bluntly pointed tip. Neurosetae few and unidentate, the swollen blades having about 20 rows of short spinules and a long hooked and sharply pointed tip. There is no trace of a secondary tooth. E. assimilis is close to E. abyssorum McIntosh 1885 dredged in 2,600 fathoms south of Australia. The type was examined in the British Museum. Like E. assimilis it lacks eyes and has papillose antennae and dorsal cirri. The elytra are missing. The setae are not markedly different, but the notosetae are more strongly serrated and the tips are quite smooth, not flanged. On the other hand each neuroseta does have a flange preceding the strong terminal tooth. £. assimilis may prove to be a synonym of E. abyssorum, but until the elytra of E. abyssorum are known the two should be kept separate. 360 ANNALS OF THE SOUTH AFRICAN MUSEUM Sub-family Sigalioninae Leanira tetragona (Oersted, 1845) Leanira tetragona (Oersted), Fauvel 1923, p. 117, fig. 43 a-g. Records: Station A.189 (2)—S.A. Museum Register No. A.19768. Notes: This is a new record for South Africa but although the specimens are in poor condition and lack elytra their other characters agree with speci- mens from northern Europe (the type locality is Norway). In particular the prostomium lacks eyes, the median antenna is long and unjointed with a pair of flaps at its base, there is a dorsal cirrus on setiger 3 and there are several slender stylodes on the parapodia; the notopodium has 5-7 and the neuro- podium a series of 10-12. Cirriform branchiae appear on setiger 5 or 6, and 3 ciliated swellings or ‘cupuliform branchiae’ are present between the noto- | podium and the elytrophore. There are no simple setae in the neuropodia— all are compound with smooth or weakly serrated shaft-heads and laddered blades. It may be added that European specimens have a few delicate papillae on the external margins of the elytra, a character which distinguishes this species from L. hystricis and L. incisa which have both been recorded from southern Africa. Both of them also lack a dorsal cirrus on setiger 3. Fauvel (1923) was doubtful of the presence of a dorsal cirrus on setiger 3 of any species of Leanira, but an examination of the types of L. magellanica McIntosh 1885, L. areolata McIntosh 1885 and L. japonica McIntosh 1885 shows that all three have a dorsal cirrus on setiger 3. A revision of the genus is necessary. Leanira hystricis Ehlers 1874 Leanira hystricis Ehlers, Fauvel 1923, p. 118, fig. 43 h-m. L. hystrix McIntosh 1925, p. 38. Records: Stations A.315 (1), A.317 (4) and A.gi9 (1)—S.A. Museum Register No. A.19963. WCD 73 (1). Notes: McIntosh’s specimen in the British Museum agrees with the present material. There are no eyes, the median antenna lacks ctenidial flaps, there is no dorsal cirrus on setiger 3, and cirriform gills appear between the 24th and 30th foot. The stylodes are all sausage-shaped; there is one on the notopodium and 2—5 on the neuropodium. There are no simple neuropodial setae. Family Nereidae Nereis (Neanthes) papillosa n. sp. (Fig. 2 a—g) Records: Station A.1g1t (1)—S.A. Museum Register No. A.19779. Description: ‘The holotype is a complete specimen 25 mm. long with 58 segments. It is pale in alcohol without any colour pattern. It is an incipient heteronereid female with enlarged eyes and swimming setae developing in setigers 16-35. POLYCHAETE FAUNA OF SOUTH AFRICA 361 The prostomium (fig. 2a) is broadly oval with short broad palps but the antennae and tentacular cirri do not present any features of special interest. The anterior pair of eyes is greatly enlarged but the posterior pair is normal. The proboscis (fig. 2b and c) when dissected proved to have very pale, poorly chitinized paragnaths difficult to see; in particular some doubt must remain as to the presence of paragnaths on groups VI, VII and VIII. Group I = 1, mie lp — 4 IN — 5) im a wedge,, Vi — 0, Vi = 20r' 3 (very faint), VII + VIII = a single row of about 4 very pale points. Fic. 2. Nereis (Neanthes) papillosa. (a) Head. (4, c) Dorsal and ventral views of proboscis. (d) 10th foot. (e) 25th foot. (f) 45th foot. (g) Posterior neuropodial falciger. Anterior feet (fig. 2d) have conical lobes and a rather short dorsal cirrus. There are only two notopodial lobes but the notosetae arise from a dorsal bulge on the inferior lobe so that the notopodium might be said to have 24 lobes. The neuropodium is normal. In the middle of the body (fig. 2e) there is a marked gap between the notopodium and the neuropodium, and all the lobes bear numerous club-shaped papillae. It is unlikely that these papillae are modifications due to the heteronereid phase for this does not occur in other species and in any case the papillae continue on the posterior feet where swimming setae are not developed. The posterior feet (fig. 2f) have long dorsal 362 ANNALS OF THE SOUTH AFRICAN MUSEUM cirri without any basal flaps, the parapodial lobes are slender and the papillae are less numerous. The notosetae are homogomph spinigers throughout, there being no notopodial falcigers. ‘The neurosetae include the usual homogomph and hetero- gomph spinigers and heterogomph falcigers (fig. 2g). These have rather long straight blades. In middle segments the tips of the developing swimming setae project from both setigerous lobes. There is a general similarity between WN. kerguelensis, N. unifasciata and N. papillosa. All have two notopodial lobes in anterior feet and a single row of paragnaths on groups VII + VIII. WN. papillosa is unique in having papillae on the parapodial lobes. Family Nephthydidae Nephthys (Aglaophamus) macroura Schmarda 1861 Aglaophamus macroura (Schmarda), Hartman 1950, p. 118. Nephthys (Aglaophamus) macroura Schmarda, Day 1960, p. 327. Records: WCD 732 (1). Notes: The single specimen is a juvenile 25 mm. long and its identification is therefore doubtful. The ventral cirrus on setiger 1 is large but the dorsal cirrus is rudimentary. Cylindrical, involute branchiae appear on setiger 3. An average parapodium from the middle of the body has conical setigerous lobes but lacks presetal lamellae on both rami. The postsetal lamella of the notopodium is divided into a large rounded superior part and a small inferior part. The notopodial cirrus is well developed, tapered and rounded in section, not flattened and blade-like. The postsetal lamella of the neuropodium is auricular and just exceeds the setigerous lobe. Anterior setae are normal laddered capillaries and posterior setae are long, fine and minutely denticulate on one side. The absence of a dorsal cirrus on setiger 1 and the early appearance of the branchiae suggest V. macroura, but the structure of the feet, particularly the notopodial cirrus, is closer to WN. malmgreni, which is also a deep-water species. Possibly the early appearance of the branchiae is a juvenile character. Nephthys hystricis McIntosh 1900 Nephthys hystricis McIntosh, Fauvel 1923, p. 373, fig. 146 a-e. Records: WCD 73 (1). | Notes: The single specimen is a juvenile and is referred to WN. hystricis with hesitation. Nephihys (Nephthys) ? paradoxa Malmeren 1874 Nephthys (Nephthys) ? paradoxa Malmgren, Day 1960, p. 327. Records: Station A.315 (1)—S.A. Museum Register No. A.19965. Notes: The specimen is broken and badly preserved. While it is similar to the one reported by Day, 1960, the identification must remain uncertain. A large parasitic nematode lies in the body cavity. POLYCHAETE FAUNA OF SOUTH AFRICA 363 Family Eunicidae Sub-family Onuphidinae Ayalinoecia sp. Records: WCD 73—15 specimens and 3 juveniles. Notes: These specimens approach H. bilineata var. rigida as described by Fauvel, 1923, p. 424, but differ in several respects. They are encased in fragile tubes covered with detritus quite unlike the tough quill tubes of H. tubicola, the only species of the genus previously recorded from South Africa. The two frontal antennae are rather broad and the five occipital antennae are mounted on 4-ringed ceratophores, each ring having a projecting lateral lobe. There are no eyes and no tentacular cirri. Conical ventral cirri are present on the first 3 setigers. Gills start on setiger 9 and continue for about 12 segments. Each gill is a small inconspicuous filament no larger than the dorsal cirrus. The hooded hooks of setigers 1 and 2 are unidentate and very faintly pseudo-compound. I know of no species of Hyalinoecia with branched ceratophores to the occipital antennae, but similar structures do occur in the genus Epzdiopatra and it is just possible that all these specimens may develop spiral gills at a later stage. For this reason they have not been named as a new species. Sub-family Lumbrinerinae Lumbrineris magalhaensis Kinberg 1864 Lumbrineris magalhaensis Kinberg, Day 1960, p. 362, fig. 12 hj. Records: Station A.189 (1)—S.A. Museum Register No. A.19769. Lumbrineris brevicirra (Schmarda 1861) Lumbrineris brevicirra (Schmarda), Day 1961, p. 361, fig. 12 e-g. Records: Station A.193 (1)—S.A. Museum Register No. A.19781. Family Paraonidae Paraonis gracilis oculata Hartman 1957 Paraonis gracilis oculata Hartman 1957, p. 331, pl. 44, figs. 1-3. Records: WCD 74 (3 juveniles). Notes: All three specimens are broken but probably measured less than 8 mm. when complete. The prostomium bears a pair of subdermal eyes and there are only 3 pairs of gills on setigers 6, 7 and 8. Posterior neurosetae include 1-2 capillaries and 2-3 short stout acicular hooks with curved, sigmoid shafts and blunt unidentate tips. There is no trace of an arista. Ehlers (1913) recorded Aonides gracilis (now recognized as a synonym of Paraonis gracilis) from False Bay but did not mention whether his specimen had eyes or not. 364. ANNALS OF THE SOUTH AFRICAN MUSEUM Aricidea suecica simplex var. nov. (Fig. 3 a—b) Records: WCD 73 (3). Description: All three specimens are broken and only one is sufficiently long to identify the posterior neurosetae. It has been chosen as the holotype. It is 8 mm. long with 65 segments and probably represents about half of the original worm. The prostomium (fig. 3a) is short and broad with well-marked nuchal slits, a very short median antenna, but no eyes. The anterior region is flattened dorsally and the posterior region rounded. Branchiae start on setiger 4 and continue to setiger 15 so that there are 12 pairs in all. Each gill is stout with a pointed tip and just meets its fellow in the mid-dorsal line. All gills are subequal except for the last 2-3 which are small. ‘The postsetal notopodial lobe of setiger 1 is a small papilla but later ones soon increase in size and in the branchial region they are 2/3 the length of the gills. The neuropodia are merely low CIAAIE SH Cd ZA 1 ; 2 j i ! Pye v [es \ \ \ 3 : \ i c epee IN : eee ies Fic. 3. Aricidea suecica var. simplex. (a) Head and anterior segments. (b) Posterior neuropodial hook. Streblosoma abranchata: (c) Lateral view of head and anterior segments. (d, ¢) Lateral and edge-on views of thoracic uncinus. (f) Thoracic notoseta. POLYCHAETE FAUNA OF SOUTH AFRICA 365 lateral ridges. Anterior notosetae and neurosetae are similar. Each is a curved capillary with a slender blade. The posterior region as defined by the appearance of the neuropodial hooks starts on setiger 24. The segments are rounded in section, each with a slender postsetal notopodial lobe. The posterior notosetae are 2-3 fine capil- laries but the neurosetae are more numerous. Each neuropodium bears about 5 long slender capillaries and an increasing number of hooks; at first there are only 1-2 but later as many as 10. The hooks (fig. 3) are all similar from segment 24. onwards. Each is sigmoid in shape with a shaft which is constricted where it leaves the surface and a curved unidentate tip which never has any prolonga- tion or sign of an arista. The stem form A. suecica Eliason 1920 from northern Europe is described as having an arista or slender blade projecting beyond the apex of the neuro- podial hook in the middle of the body, but this is lost later and posterior neuro- podial hooks are sigmoid and unidentate. There is also a pair of eyes. This is a new record for South Africa. Family Spionidae Prionospio steenstrupi Malmgren 1867 Prionospio steenstrupi Fauvel 1927, p. 60, fig. 21 f-1. Prionospio malmgreni var. dubia Day 1961, p. 489, fig. 3 j—n. Records: WCD 73 (1). Spiophanes soderstromt Hartman 1953 Spiophanes soderstromi Hartman 1953, p. 41, fig. 14 a-c. Day 1961, p. 484. Records: WCD 74 (1 juvenile). Family Disomidae Poecilochaetus serpens Allen 1904 Poecilochaetus serpens Allen, Fauvel 1927, p. 67, fig. 23 a-m. Day 1961, p. 497. Records: WCD 73 (1 juvenile). Family Maldanidae Maldanella fibrillata Chamberlin 1919 Maldanella fibrillata Chamberlin 1919, p. 413, pl. 72 figs. 1-6, pl. 73 figs. 1-2. Records: Station A.190 (5)—S.A. Museum Register No. A.19773. Description: The 5 specimens were encased in fragile tubes covered with grey silt. Their bodies were broken and the number of segments is unknown. Head with an oval cephalic plate inclined at 45° to the axis of the body and surrounded by a high, smooth rim which is continuous apart from a small anterior gap occupied by the prostomium. No eye spots. Nuchal grooves straight and half the length of the cephalic plate but their anterior ends are 366 ANNALS OF THE SOUTH AFRICAN MUSEUM continuous with a streak which curves sharply back around the inside of the rim. Anterior segments short, hardly longer than broad, and without collars but with glandular rings on the first 7. Middle segments long but the posterior ones shorter and sausage-shaped. 3 achaetous preanals which decrease in length. The last is not very distinct and is situated at the base of the pygidial funnel which is rimmed with 30 equal anal cirri. Anus on a pleated cone with- out a marked ventral valve. Setiger 1 with notosetae only. Setiger 2 with notosetae and 15-20 neuro- podial hooks. Later segments similar but with more hooks. The notosetae are all narrow-winged capillaries, some long and some short. The hooks on setiger 2 have a vertical series of 2-3 teeth above the main fang while those of later segments have 3—4 teeth. These characters agree with Chamberlin’s description and it may be added that his original specimen was complete and measured 70 mm. by 5°5 mm. and had 109 setigerous segments. M. fibrillata is a rare abyssal species from the Pacific off the coast of Panama. The only species of Maldanella known from South Africa is M. capensis Day which has 4 achaetous preanals (not 3) and only 2-3 neuropodial hooks on setiger 2 (not 12 or more). Lumbriclymene minor Arwidsson 1906 Lumbriclymene minor Arwidsson: Fauvel 1927, p. 196, fig. 68 k—q. Records: Station A.315 (1)—S.A. Museum Register No. A.19966. Notes: The tube is free and encrusted with foraminiferan shells. The body is 20 mm. long with 19 setigers and 3 indistinct preanals in front of the pygidium. The anterior end is bluntly rounded without a well-defined prostomium, cepalic plate or crest. Even the curved nuchal grooves are poorly marked. The 19 setigers do not differ greatly in length, each being about three times as long as broad. The anterior margins of anterior setigers are slightly glandular but the intersegmental constrictions are not deep and there is no sign of collars. The three achaetous preanal segments are short and poorly marked with indistinct lateral tori. The pygidium is blunt and slightly swollen with a dorsal anus. The ventral valve is very large and is continuous with the bluntly conical posterior end. The ventral surface of the pygidium is somewhat flattened and slopes upwards. There are no anal cirri. The notosetae are all winged capil- laries, some smooth-edged and some striated towards the distal end. Each of the first 4 setigers bears a single stout, smoothly pointed, acicular seta. Subse- quent neurosetae are rows of about 6 hooks, each with a vertical series of 4 teeth above the rostrum and a few lateral denticles. Below the rostrum is a short ‘neck’ and then a marked swelling in the shaft. This is a new record for South Africa. Its pygidium differs from L. cylindri- cauda Sars which has recently been found in shallow water in this area (unpub- lished record). In L. cylindricauda as the name implies the posterior end is POLYCHAETE FAUNA OF SOUTH AFRICA 367 cylindrical with a terminal anus; in L. minor the anus is dorsal and the ventral surface of the pygidium is flattened and slanting. Family Sabellariidae Phalacrostemma elegans Fauvel 1911 Phalacrostemma elegans Fauvel 1914, p. 270, pl. 24, figs. 1-16. Records: Station A.315 (1)—S.A. Museum Register No. A.19967. Notes: The single specimen is poorly preserved and lacks a tube, but the available characters, in particular the setae, agree completely with Fauvel’s description. The opercular lobes are separate and each bears a single ring of long, tapered and spirally serrated ‘paleae’ with about 3 stout acicular setae at the base. At the junction of the opercular lobes there is one pair of hooks dorsally and a single median tentacle ventrally. The buccal tentacles and palps have disintegrated. There are 4 parathoracic segments with oar-shaped setae. There are at least 12 abdominal segments bearing uncini and capillaries, but the gills have been lost and the posterior end has disintegrated so that the exact Tae? of segments and details of the caudal region are unknown. This is a new record for South Africa. Fauvel’s original specimen came from 1,968 metres off Madeira. Family Ampharetidae Neosabellides cf. elongaius (Ehlers) 1913 [?] Sabellides elongatus Ehlers 1913, p. 551, pl. 42, figs. 1-6. [?] Neosabellides elongatus (Ehlers), Hessle 1917, p. 104. Monro 1936, p. 175. Records: Station A.319 (3)—S.A. Museum Register No. A.19968. Description: ‘The 3 specimens are encased in characteristically slender and closely ringed brownish mud tubes. The body is slender, slightly swollen anteriorly, about 30 mm. long and 1-2 mm. wide at the broadest part of the thorax. The prostomium is bluntly spade-shaped without lateral grooves but has a single pair of eyes. The tentacles are largely retracted and lateral pinnules are poorly marked. Segment II is narrow but unusually distinct. Segment III lacks paleae and is fused to segment IV which bears a small bundle of noto- podial capillaries. Segments V and VI also bear notopodial capillaries but no uncini. Segment VII and the next 10 segments bear both notopodial capil- laries and neuropodial uncini so that there is a total of 14 segments with noto- setae of which the first three lack uncini. Segment III bears two groups of 3 smooth, tapered gills which project well beyond the prostomium. The two groups of gills are well separated in the dorsal median line and each group is arranged in a transverse row, but there is no obvious branchial ridge nor is there any sign of nephridial papillae between the two groups of gills. The first 368 ANNALS OF THE SOUTH AFRICAN MUSEUM 5 thoracic segments are short, each being about three times as broad as long, but subsequent segments become much longer until in the middle of the thorax each segment is three times as long as broad. The length of the segments affects the glandular ventral pads which are recognizable on all except the last thoracic segment. Anterior ones are well marked and contiguous, but posterior pads are poorly developed and well separated from one another. The abdomen consists of 32 segments. Each uncigerous pinnule is a small, roughly square lateral lobe with a small papilla above the row of uncini. The pygidium bears a circle of 6-8 tapered anal cirri and a larger pair of ventral lobes. The notosetae are smooth-winged capillaries. Thoracic uncini bear 8 teeth in two vertical rows of 4. These South African specimens agree with Ehlers’s original description of Sabellides elongatus from Antarctica in regard to most characters, but the first three setigers lack uncini whereas Ehlers states that only two anterior setigers lack uncini. Moreover Ehlers found only 19 abdominal segments or less, whereas these South African specimens have over 30. It may be that Ehlers (1913), Hessle (1917) and Monro (1936), who all describe material from the Antarctic or sub-Antarctic, missed the first bundle of notosetae, which is small and very close to the second, but for the present the identity of the South African speci- mens must remain doubtful. Family Terebellidae Amphitrite cirrata Miller 1771 Amphitrite cirrata Miller, Fauvel 1927, p. 251, fig. 86 7-0. Records: Station A.319 (6)—S.A. Museum Register No. A.19971. Notes: The present material agrees perfectly with Fauvel’s description. There is a general resemblance to a species of Thelepus since the three pairs of gills are filamentous, with the right and left tufts well separated. Each tuft arises from a basal stump. There are 17 thoracic segments starting on segment 4 (3rd branchiferous), each bearing a bundle of notopodial capillaries with minutely denticulate tips. There are small lateral lobes on segments 2, 3 and 4, and 12 ventral pads. Uncini start on segment 5 (setiger 2) and are arranged in double rows after the first few. Each uncinus is avicular with a close-set cap of denticles above the main fang. The denticles are irregularly arranged but approximate to the formula: MF : 4-5 : 5-6 : 8-10 : 10-15. The abdomen is broken in every specimen but has at least 20 segments with square uncigerous pinnules. This species is well known from high latitudes in the North Atlantic. Fauvel (1914) recorded Amphitrite cirrata profunda from abyssal depths off the Azores, but as he says himself it is very doubtfully distinct from the stem form. This is the first record of the genus Amphitrite from South Africa. POLYCHAETE FAUNA OF SOUTH AFRICA 369 Pista cristata (Miller 1776) Pista cristata (Miller), Fauvel 1927, p. 266, fig. 93 a-g. Records: Station A.317 (1), A.319 (6)—S.A. Museum Register No. A.19970. Notes: These specimens agree well with Fauvel’s description. There are 2 pairs of gills, each with a long stem and a terminal pompon of spirally arranged filaments. There are 3 pairs of lateral lobes; the pair on segment 2 is small and continuous across the ventrum, the pair on segment 3 is large and lateral in position, and the pair on segment 4 is small. 17 thoracic segments bear bundles of smooth-winged notopodial capillaries. The first four rows of uncini have a close-set crest of denticles, short ‘necks’ below the main fang, rounded bases and well-developed shafts; later uncini have no shafts. P. cristata var. capensis, reported by McIntosh (1925) from Portuguese East Africa, was said to have a single tooth above the main fang. The type has been lost. P. brevibrancha Caullery, reported by me (Day, 1951 and 1957) from Portuguese East Africa, may be distinguished from P. cristata by the short- ness of the branchial trunks and the uncini. Only the first row of uncini has well-developed shafts, the neck of the uncinus is longer and the base larger and more triangular. Terebellid— ? gen. et sp. Records: Station A.316 (12)—S.A. Museum Register No. A.19972. Notes: The specimens are very soft and poorly preserved, and some of the characters are doubtful. They are obviously different from the other species recorded here but cannot be identified with certainty. An average specimen is about 50 mm. long and 4 mm. wide across the anterior thorax. The collar-shaped tentacular lobe bears numerous orange tentacles. No gills were found; it is possible that they have been lost, but there was no clear sign of scars. No lateral lobes were visible. ‘There are 17 bundles of notosetae starting on segment 3 or possibly 4. The notosetae are of two lengths, both with smooth, broad wings and very long attenuated tips which appear to be quite smooth. Uncini start on setiger 2 (i.e. segment 4 or 5), and on the posterior thorax they are arranged in double rows. The thoracic uncini are all similar, each being avicular with a crest of numerous irregularly arranged denticles above the main fang. There are at least 12 ventral pads. The abdomen consists of 40 or more segments bearing short uncigerous pinnules. These specimens belong to the sub-family Amphitritinae as shown by the arrangement and structure of the uncini. I know of no species which has smooth-bladed notosetae and lacks gills. However, a better preserved specimen is necessary to confirm that gills are really absent. Streblosoma abranchiata n. sp. (Fig. 3 cf) Records: Station A.190 (numerous tubes and specimens) —S.A. Museum ' Register No. A.19770. 370 ANNALS OF THE SOUTH AFRICAN MUSEUM Stations A.3g15 (2), A.g19 (Common) and A.321 (1)—Register No. A.19969. Description: ‘The type material is No. A.19770 from Station A.igo. The tubes are long, fragile and heavily encrusted with foraminiferan shells. The worms themselves are poorly preserved, but judged by the size of the larger fragments the body was cylindrical, rather slender and about 30 mm. long by I°5 mm. wide with over 60 segments. The tentacular lobe (fig. 3c) is short and collar-like, with about a dozen long, grooved tentacles. There are no eye-spots. An oral hood overhangs the ventral mouth but the lower lip is small. There are no lateral lobes on anterior segments and no sign of gills although all specimens were carefully examined. The first bundle of notosetae is on segment 2 and there are at least 19 segments with notosetae. ‘The longest anterior fragment had disintegrated at this point but posterior fragments of 20 or more uncigerous segments are without noto- setae. The notosetae (fig. 3f) are smooth-winged capillaries. Uncini appear on setiger 4 (segment 5). They are arranged in single rows throughout and borne on poorly marked uncigerous ridges even in abdominal segments. Each uncinus (fig. 3d and e) has a close-set cap of denticles above the main fang. These are irregularly arranged but approximate to the formula MF: 4-5 :ca. 8: ca. 12. The base of the uncinus has the characteristic clog-shape of the sub-family Thelepinae with a forwardly produced rounded prow and a well-marked dorsal button. All other species of Streblosoma have at least 2 pairs of filamentous gills, and the present species was assigned to the genus with some hesitation; but it obviously belongs to the Thelepinae, and the segmental arrangement of the setae and the lack of lateral lobes on anterior segments agree with Streblosoma. Streblosoma chilensis (McIntosh 1885) Euthelepus chilensis McIntosh 1885, p. 467, pl. 51 figs. 4-5, pl. 28A figs. 14-15. Records: WCD 73 (1 juvenile). Notes: The genus Euthelepus was erected by McIntosh 1885 for two species, E. setabulensis and E. chilensis. The types of both have been re-examined in the British Museum. FE. setabulensis, which has page priority, has been accepted as the type species by both Fauvel (1927) and Hartman (1959). It has 3 pairs of filamentous gills on segments 2-4 and notosetae from segment 3 (not segment 2 as suggested by McIntosh). Uncini start on segment 5 (setiger 3). Lateral lobes are present on segments 2-4. F. chilensis has 2 pairs of simple filamentous gills on segments 2 and 3. Notosetae start on segment 2 (the first branchiferous) and uncini on segment 5 (setiger 4). There are no lateral lobes on anterior segments. It thus differs from EF. setabulensis in the possession of notosetae on segment 2 and in the absence of lateral lobes. It should thus be transferred to the genus Streblosoma. The single South African specimen is a juvenile but agrees in all essential respects with McIntosh’s type from abyssal depths off Chile. POLYCHAETE FAUNA OF SOUTH AFRICA 371 Terebellides stroemt Sars 1835 Terebellides stroemi Sars, Fauvel 1927, p. 291, fig. 100 7-@. Records: WCD 73—2 juveniles. SUMMARY Twenty-eight species of Polychaeta are described from abyssal dredgings west of Cape Town. The collection includes 4 new species and 11 new records for this area. Only 4 of these species are known from depths less than 100 metres; the rest are widespread at abyssal depths. There is no obvious Antarctic component. ACKNOWLEDGEMENTS The Trustees of the South African Museum wish to acknowledge a grant from the Council for Scientific and Industrial Research towards the purchase of the deep-trawling equipment with which the specimens reported on were collected, and a grant for the publication of this paper. Grateful acknowledge- ment is also due to the Director, Division of Fisheries, Department of Commerce -and Industries, for making the collecting possible. REFERENCES CHAMBERLIN, R. V. 1919. The Annelida Polychaeta. Mem. Mus. Comp. Zool. Harvard, 48, 1-514. Day, J. H. 1951. The polychaete fauna of South Africa. Part 1: The intertidal and estuarine Polychaeta of Natal and Mocambique. Ann. Natal Mus. 12, 397-441. Day, J. H. 1957. The polychaete fauna of South Africa. Part 4: New species and records from Natal and Mocambique. Ann. Natal Mus. 14, 59-1209. Day, J. H. 1960. The polychaete fauna of South Africa. Part 5: Errant species dredged off Cape coasts. Ann. S. Afr. Mus. 45, 261-373. Day, J. H. 1961. The polychaete fauna of South Africa. Part 6: Sedentary species dredged off Cape coasts with a few new records from the shore. 7. Linn. Soc. (zool.), 44, 463-560. En ers, E. 1913. Die Polychaeten-Sammelungen der deutschen Siidpolar-Expedition 1g01— 1903. Deutsch. Siidpol.-Exped. 13, 397-598. FAuvEL, P. 1914. Annélides polychétes non-pélagiques provenant des compagnes de l’ Hirondelle et de la Princesse-Alice (1885-1910). Result. Comp. Sci. Monaco. 46, 1-432. FAUVEL, P. 1923. Polycheétes errantes. Faune Fr. 5, 1-488. FAUVEL, P. 1927. Polycheétes sedentaires. Faune Fr. 16, 1-494. HARTMAN, O. 1950. Goniadidae, Glyceridae and Nephtyidae. Allan Hancock Pacif. Exped. 15 (1), 1-181. Hartan, O. 1953. Non-pelagic Polychaeta of the Swedish Antarctic Expedition 1901-1903. Further Zool. Res. Swed. Antarct. Exped. 4 (2), 1-185. Hartman, O. 1957. Orbiniidae, Aspidobranchidae, Paraonidae and Longosomatidae. Allan Hancock Pacif. Exped. 15 (3), 211-392. Hartman, O. 1959. Catalogue of the Polychaetous Annelids of the world. Part II. Allan Hancock Publ. occ. paper 23, 355-628. Hesste, C. 1917. Zur Kenntnis der terebellomophen Polychaeten. Zool. Bidr. Uppsala 5, 39-258. MclIntosu, W. C. 1885. Report on the Annelida Polychaeta collected by H.M.S. ‘Challenger’ during the years 1873-1876. Challenger Rep. Zool. 12, 1-554. MclIntosu, W. C. 1925. A second contribution to the marine Polychaeta of South Africa. Rep. Fish. Mar. biol. Surv. S. Afr. 4, spec. rep. 4, 1-93. Monro, C. C. A. 1936. Polychaete worms. II. Discovery Rep. 12, 59-198. he ros iba fA es are EE = - _ rye Wik dee. hb ee ets ) has i ; ey i ‘ "at Wee Be , ¥ 4 ry og , ' ¢ i f ii , enn j = J ’ P . } Ni fi Wy ® ere [ CONTINUED FROM INSIDE FRONT COVER References thus appear as follows: AUGENER, H. 1913. Polychaeta. In Michaelsen, W., ed. Beitrdége zur Kenntnis der Meeresfauna Westafrikas. 2, 67-625. Hamburg: Friederichsen. EKMAN, S. 1953. VI (Part 1, Index), VII (Parts 1-4), VIII, IX (Parts 1-2), X (Parts 1-3), XI ipa 25 7 Index), XXI, XXIV (Part 2), XXXI (Parts 1-2). Current prices: Vole i: III. IV. XX. XXII. XXIII. XXIV. XXV. XXVI. XXVIII. XXVIII. XXIX. XXX. INDEX XXXI. XXXII. XXXII. XXXIV. XXXV. XXXVI. XXXVII. XXXVITI. XXXIX. >. 6 XLI. XLII. XLIII. XLIV. XLV. XLVI. 1900-1902 1903-1905 1903-1908 1906-1910 1908-1910 1908-1913 IQII—19198 IQII-I19I14 Ig11—-1918 IQI13—-1924 IQ13—1923 IQI5—-1924 1914-1916 1917-1933 1917-1920 1921 1924-1925 1924-1926 1925-1928 1925-1926 1929-1933 1927-1928 1928 1929 1929-1932 1929-1931 1931-1935 Zoology and Geology (excl. Parts 1-3, 5, 7-8) Zoology (excl. Part 1) : ie Palaeontology .. Geology, Ralacontolocy, Zoology, Anthropology (exc. Parts 1-2, 5, 7-9) Zoology (excl. Part 1, Index) Palaeontology (excl. Parts 1-4) Botany (excl. Parts 1-2) Zoology (excl. Parts 1-3) ae Zoology (excl. Parts 2, 7, Index) Palaeontology and Geology Archaeology and ee Zoology Zoology Botany Zoology Zoology Zoology Zoology Ge Palaeontology .. Zoology Anthropology nd Ethnology (excl Part pel Zoology , Zoology Anthropology Palaeontology . Zoology Zoology of papers, authors and wubines mubenedat in ile! 12 XXK 1934-1950 1935-1940 1939 1938 1956 1942-1948 1947-1952 1950 1952 1952-1956 1952-1955 1953-1956 1955-1957 1957-1959 1959-1960 1961- The LIBRARIAN, SoutH Arrican Museum, CAPE Town. Palaeontology (excl. Parts 1—2) Zoology Zoology Zoology Zoology Zoology Archaeology Zoology Zoology Botany Zoology Palaeontology : Zoology and Palaeontology Zoology and Palaeontology Zoology and Palaeontology f Zoology and Palaeontology (Parts 1-1 8) Copies may be obtained from— R 2.75 4.38 8.08 1.54 5.48 1.26 6.03 10.54 6.73 Lip Garp 8.55 8.04 10.26 8.57 8.45 10.33 8.10 6.27 4.85 4-25 6.00 4.70 3.60 3.60 6.30 5.80 8.88 0.23 4.20 7.62 5:05 5:75 14.40 6.13 4-33 9.00 6.55 3.16 9.60 8.10 4.80 6.64 11.33 8.04 ANNALS OF THE EOU Py APRICAN MUSEUM VOLUME XLVI PART XX TOOTH REPLACEMENT IN THE CYNODONT THRINAXODON LIORHINUS SEELEY By A. W. Crompton South African Museum, Cape Town ISSUED NOVEMBER 10963 PRICE 750. PRINTED FOR THE TRUSTEES OF THE SOUTH AFRICAN MUSEUM BY THE RUSTICA PRESS (PTY.) LIMITED, COURT ROAD, WYNBERG, CAPE INSTRUCTIONS TO AUTHORS MSS. submitted for publication must be typewritten, double spaced with good margins, typed on one side of sheet only and sheets numbered. Two MSS. and one set of illustrations must be submitted. Papers must include a Table of Contents and a Summary containing a succinct account of the subject, results obtained and conclusions. The position of the text-figures and tables must be clearly indicated. ILLUSTRATIONS. Please must be kept to a minimum and made up to occupy 5” x 7” (full plate excluding captions) or in direct proportion to these measure- ments if they are to be reduced. A scale (metric system) must appear with all photographs. REFERENCES. Harvard system—authors’ names and dates of publication given in the body of the text; references arranged at the end of the paper in alphabetical order of authors’ names. Bibliographical arrangement of references at the end of the paper must give: 1. Name of author, followed by his initials. Names of joint authors connected by ampersand (&), and not ‘and’. 2. Year of publication. If several papers by the same author in one year are cited, suffixes a, b, etc., are used. 3. Full title of the paper. Initial capital letters only for the first word and for proper names, except in German, where the usage of the language is followed. 4. Title of the journal, abbreviated according to the World list of scientific periodicals, and underlined to indicate italics. 5. Series number, if any, in parenthesis, e.g. (3), (n.s.), (N.F.), (B). 6. Volume number in arabic numerals (without prefix ‘vol.’), with wavy underlining to indicate black type. 7. Part number, only if the separate parts of one volume are independently numbered. 8. Page numbers, first and last, in arabic numerals (without prefix ‘p’). When reference is made to a separate book, give in the order listed: Author’s name; his initials; date of publication; title, in italics; edition, if any; volume number, if any, in black type arabic numerals; pagination, if reference is to part of book only; place of publication; name of publisher. When reference is made to a paper forming a distinct part of another book or collection, give: Name of author of paper; his initials; date of publication; title of paper; ‘In’ italicized; name of author of book; his initials; title of book, in italics; edition, if any; volume number, if any, in black type arabic numerals; pagination of paper; place of publication; name of publisher. [CONTINUED ON INSIDE BACK COVER TOOTH REPLACEMENT IN THE CYNODONT THRINAXODON LIORHINUS SEELEY By A. W. Crompton South African Museum, Cape Town (With 17 figures in the text) CONTENTS PAGE Introduction , ; Amy ie Material . ; . 480 Replacement of individual teeth . . 481 Structure of the crowns of the postcanines . 483 Discussion Extension of growth series of Thrinaxodon . 497 Posterior migration of the postcanine row. 501 Replacement cycle > , . 502 Differentiation of the postcanine row eG Notes on the postcanine row of gomphodont cynodonts . ; » 515 Summary . : : , a RTO Acknowledgements . 520 References . . , . 520 Key to lettering : . 52x INTRODUCTION The very mammalian appearance of the dentition of such theriodont reptiles as the cynodonts invited the assumption that these reptiles replaced their teeth as mammals do and that they had milk and permanent dentitions (Broom, 1913) rather than the indefinite series of replacing teeth occurring in typical reptiles. But Parrington (1936) has shown that in the cynodont Thrinaxo- don alternately numbered postcanines were replaced at different times. In the eight skulls at his disposal Parrington could show that each of the postcanines was replaced at least once. He divided the material into a younger group in which postcanines nos. 1, 3 and 5 had been replaced shortly before death or were actually undergoing replacement, and an older group in which post- canines 2, 4 and 6 had been replaced shortly before death, and he assumed that the tooth position of the Ist postcanine in the younger group was homologous with the tooth position of the Ist postcanine in the older group. Parrington noted that some of the specimens have six upper postcanines whereas others have seven, but he gave no explanation for this. He claimed that the 7th post- 479 Ann. S. Afr. Mus. 46 (20), 1963, 479-521: 17 figs. a BST TUIIOH JAN J 194 480 ANNALS OF THE SOUTH AFRICAN MUSEUM canines were unlikely to have had predecessors and were erupting for the first time. Also Parrington described the mandibular postcanines in one specimen in which he noted that the 7th and 8th postcanines had five cusps and the 5th four cusps, instead of the customary three. Subsequently Brink (19556) has described a very young specimen of Thrinaxodon, found in close association with a large specimen which has seven teeth. He formed the opinion that the postcanines were replaced several times during life because the teeth of the young specimen were so much smaller than those of the large specimen. I have (Crompton, 1955) discussed tooth replace- ment in Cynognathus zone and Middle Triassic cynodonts, notably Scalenodon. In Scalenodon postcanine teeth were found to have been added at the back of the postcanine row during life and a few were lost in front. Fourie (1963) has obtained similar results in Diademodon. Recently I (Crompton, 1962) showed that in the bauriamorph E£riciolacerta the replacing teeth developed in pits in the alveolar bone lingually to the roots of the functional postcanines and that alternately numbered teeth were replaced at different times. MATERIAL Four specimens of Thrinaxodon have recently been beautifully prepared with the aid of acetic acid by the staff of the British Museum (Natural History). These are B.M.N.H. R3731, R5480, R511a and R511. The first three were studied by Parrington (his specimens E, G and I respectively). I studied three of the British Museum specimens, viz. R3731, R511a and R511. Several specimens in the collections of the South African Museum have also been prepared with the aid of acetic acid. Consequently much additional information on the structure of the tooth crowns and tooth replacement has come to light. For the purpose of this paper a small skull of Thrinaxodon housed in the Universitats Institut fiir Palantologie u. historische Geologie in Munich and the very young specimen of Thrinaxodon housed in the Bernard Price Institute for Palaeontological Research in Johannesburg were studied. For convenience the specimens studied have been lettered A to I, as follows: A = Bernard Price Institute for Palaeontological Research Catalogue No. 274. B = Universitats Institut fiir Palaontologie u. historische Geologie, Munich. = B.M.N.H. R3731 = Parrington’s Specimen E. 1D. SA. RS 77 E =S.A.M. K 380. F = B.M.LN.S. Rais, G=sA.M YE. 296; H = B.M.N.H. R511a = Parrington’s Specimen I. Po = Sea ee SG: The structure of the postcanine teeth indicates that this sample may contain more than one species of Thrinaxodon. This is discussed on pp. 511. TOOTH REPLACEMENT IN THRINAXODON LIORHINUS SEELEY 481 REPLACEMENT OF INDIVIDUAL TEETH Postcanines: A shallow longitudinal groove (long.g., fig. 2 A and 6 A and D) is present in the maxilla and mandible lingual to the alveolar borders of the postcanines and the canine. This groove presumably housed the dental lamina. Replacement teeth commenced their development in shallow pits (p) in the floor of this groove in the dentary and in the roof of this groove in the maxilla. Initially these pits were separated by a layer of bone from the alveoli of the A Rear ee dost Reacts Fic. 1. Yhrinaxodon liorhinus. A, right maxillary dentition of Specimen A. B and C upper dentition of Specimen B. Key to lettering p. 521. 482 ANNALS OF THE SOUTH AFRICAN MUSEUM functional postcanines which lie labial to them. These pits are well preserved in the acid-prepared specimens and in many cases replacing teeth are preserved in the pits (rep.t., fig. 2 B). The functional teeth are usually firmly held in place by a ring of attachment bone around the neck of the tooth. The attach- ment bone (a.b.) is well preserved around the neck of several postcanines studied and is especially well preserved around the 2nd and 4th upper post- canines of specimen E (fig. 5 A). A ring of small foramina indicates the Jine of fusion between the attachment bone and the tooth itself. The various stages of the growth of a replacing tooth between its initial development in a pit and its eventual eruption are well shown in the acid-prepared specimens. As the replacing tooth increased in size the pit containing it also increased in size and opened into the alveolus of the functional tooth. The replacing tooth migrated labially and the root of the functional tooth was resorbed to make space for it. Consequently, shortly before the replacing tooth erupted little remained of the root of the functional tooth. At this stage the remaining part of the functional tooth was secured solely by the ring of attachment bone around its neck at the alveolar border. In many specimens the functional tooth was lost at this stage by absorption of the neck of the tooth in a horizontal plane in line with the outer surface of the ring of attachment bone so that nothing visible remains externally to indicate the previous presence of a functional tooth except a ring of foramina in the attachment bone around the remnant of the tooth root. (See fig. 1 B.) This phenomenon accounts for the gaps in the tooth row in the right maxillae of Parrington’s specimens A and F. Before the replacing tooth erupted the alveolus was considerably enlarged by the resorption of the attach- ment bone and the new tooth erupted into an alveolus far larger than the neck of the replacing tooth. Consequently the younger teeth that erupted shortly before death were loosely held in their respective alveoli, e.g. 1st, 3rd, and 5th in the maxilla of specimen C and 3rd, 5th and 7th in the mandible of specimen C (fig. 2). The older teeth, e.g. 2nd and 4th in the maxilla of specimen C (fig. 2 A and B), were firmly held in place by attachment bone. A replacement cycle identical to that of Thrinaxodon appears to have been present in the bauriamorph Ericiolacerta (Crompton, 1962) and was also present in Cistecephalus zone cynodonts that have been studied by the author. The material studied for this paper supports Parrington’s (1936) conclu- sion that the odd- and even-numbered postcanines erupted at different times. This is clearly seen in most specimens, e.g. in the maxilla of specimen C (fig. 2 A) the 1st, 3rd and 5th erupted shortly before death and were loosely held in large alveoli whereas the 2nd and 4th were tightly held and have pits for replacing teeth in the maxilla lingual to them. However, it appears that all the odd- numbered (or even-numbered) teeth did not erupt simultaneously; but that the posterior replacing teeth erupted before the anterior teeth of the same set. This is well shown in the right mandible of specimen H (fig. 10 D). The 3rd, 5th and 7th teeth are loosely held in the alveoli and erupted shortly before death. The 7th tooth was fully erupted, the 5th partially erupted and the 3rd TOOTH REPLACEMENT IN THRINAXODON LIORHINUS SEELEY 483 was still completely contained in its alveolus. In terms of alternate replacement a replacing tooth would be expected to be present in the first alveolus. How- ever, the old functional tooth is still present in this position, but a deep pit for a well-developed replacing tooth is present adjacent to the old 1st functional postcanine. In the left mandible (fig. 10 C) the condition is similar to that in the right, except that the old functional Ist postcanine has been lost and a partially displaced replacing tooth is present in the first alveolus. This evidence suggests that the odd-numbered replacing postcanines erupted in the following order: 7th, 5th, 3rd, 1st. In the maxilla of specimen F (fig. 6 A) the replacing tooth for the 6th postcanine (r.pc.6.) is larger than that for the 2nd (r.pc.2.) and that of the 4th is intermediate in size between the 2nd and the 6th. The younger postcanines, i.e. 1st, 3rd and 5th of this maxilla, that are loosely held in their alveoli appear to have erupted in the following order: 5th, 3rd, 1st. A similar phenomenon can be observed in most of the specimens studied with the exception of the lower postcanines of specimen G (fig. 9) where the rule of alternate replacement seems to break down over a short distance. In most reptiles that actively replace their teeth a wave of replacement passes along each alternately numbered tooth series either from back to front or front to back (Edmund, 1960, 1962). In Thrinaxodon the wave passes from back to front. Canines: Replacing canines are present in both the upper and lower jaws of most specimens of Thrinaxodon studied. The replacing canines commenced their replacement in distinct pits in the maxilla and dentary. In the specimens studied the pits in the maxilla (specimen E, fig. 5 A) are always found antero- lingual of the functional canine whereas the pits in the dentary (specimen E, fig. 5 B) are always found postero-lingual of the functional tooth. As the replac- ing canine enlarged, the pit containing it fused with the main alveolus (speci- men D, fig. 3 A and B) and the replacing canine migrated to lie in front of the functional canine in the upper jaw and behind the functional canine in the lower jaw (specimen F, fig. 6 B and D). Fourie, however (1963), has described a specimen of Thrinaxodon in which the upper replacing canine lies behind the functional canine. Incisors: The incisor region is not well preserved in most of the specimens studied. The replacing incisors appear to have developed in pits behind the functional ones. These pits, where preserved, are confluent with the alveoli of the functional teeth. In the premaxilla of specimen F (fig. 6 A and B) replacing teeth are present behind all four incisors. It is not possible to determine the order of replacement of the functional incisors. STRUCTURE OF THE CROWNS OF THE POSTCANINES In a few instances Parrington (1936) was able to describe additional cusps on the lingual side of the crowns of upper postcanines. In the acid-prepared specimens the whole structure of the crowns of the teeth can be studied, and it 484 ANNALS OF THE SOUTH AFRICAN MUSEUM Fic. 2. Thrinaxodon liorhinus, Specimen C. A and B upper dentition, C and D, lower dentition and E, lingual view of the lower 6th postcanine. Key to lettering p. 521. TOOTH REPLACEMENT IN THRINAXODON LIORHINUS SEELEY 485 has been found that in many cases quite complex patterns are present which differ considerably from the simple crown pattern usually associated with Thrinaxodon. On the basis of the pattern of the crowns of the lower posterior postcanines, it is possible to divide the specimens studied into three ill-defined groups. Specimens A, B or I, however, cannot be assigned to any of these groups as the structure of their lower postcanines is not known. The members of the three groups form a consecutive series of increasing size (see table I, page 500). Group I consists only of specimen C. Group II consists of specimens D, E, F and G. Group III consists of specimen H. Group I Lower Postcanines. Specimen C' (fig. 2 C, D and E) Seven lower functional postcanines are present. A remnant of a tooth (r.pc.) is present anterior to the 1st functional postcanine and an unerupted postcanine (a.pc.) which does not appear to have had a predecessor is present behind the last functional postcanine. The crown structure of the more recently erupted 3rd, 5th and 7th post- canines is different from that of the older 2nd, 4th and 6th. In each series, i.e. the odd-numbered, 3rd, 5th, 7th and the even-numbered 2nd, 4th and 6th, the crown pattern becomes increasingly complex in a posterior direction. Of great interest and importance is the fact that each of the more recently erupted teeth has crowns with a simpler pattern than the older teeth anterior or posterior to it. The tooth with the most complex crown is the 6th postcanine (fig. 2 E). It consists of three cusps aligned antero-posteriorly; a main cusp (m.c.), an anterior accessory cusp (a.a.c.) and a posterior accessory cusp (p.a.c.). Because the anterior accessory cusp is larger than the posterior acces- sory cusp the anterior margin of the main cusp is shorter than its posterior margin. Anterior to the anterior accessory cusp and posterior to the posterior accessory cusp two smaller cusps are present. These have for convenience been designated as anterior and posterior cingular cusps (a.c.c., p.c.c.) respectively. It is, however, extremely difficult to determine whether the anterior and pos- terior cingular cusps do in fact develop upon a definite cingulum. In some cases this appears to be the case, but in other cases these cusps appear to be quite distinct from the cingulum. A well-developed cingulum is present on the lingual surface of the crown. It supports a series of small cusps. ‘The largest are twin cusps that lie either side of the midline. A deep valley separates the anterior and posterior twin cusps. Two small cuspules are present anterior to the twin cusps and one is present posterior to them. The crown pattern of the 4th and 2nd postcanines is similar to that of the 6th except that accessory and the anterior and posterior cingular cusps and the cusps on the lingual cingulum are not so well developed. The crowns of the 486 7 ANNALS OF THE SOUTH AFRICAN MUSEUM younger 7th, 5th and 3rd postcanines in the odd-numbered series also consist of three cusps (anterior accessory, main and posterior accessory) aligned antero- posteriorly. The anterior and posterior cingular cusps are situated low down on the lingual surface of the anterior and posterior accessory cusps and they do not lie anterior or posterior to the latter as they do in the 6th postcanine. In the odd-numbered teeth no cusps appear to be present on the lingual side between the anterior and posterior cingular cusps. Upper postcanines (fig. 2 A and B) As in the lower postcanines there is a great difference between the older and younger postcanines, but this is not as marked as in the lower jaw. The crowns are in many cases damaged. Six functional postcanines are present. The even-numbered erupted later than the odd-numbered. The 6th postcanine has apparently been shed and an unerupted postcanine is visible above it. The crown of the 4th postcanine, which is a member of the older series of teeth, consists of three antero-posteriorly aligned cusps, anterior and posterior cingular cusps and a series of cusps upon a lingual cingulum. The cingulum of the crown of the second postcanine is very poorly developed and supports only an anterior cingular cusp. In the younger 5th and grd postcanines anterior and posterior cingular cusps are present but no cusps are present on the lingual cingulum. Group II Lower postcanines. Specomen D (fig. 3 B and C and 4B) Seven functional postcanines are present. A remnant of a tooth root (r.p.c.) is present anterior to the 1st. A small, apparently non-functional tooth is present behind the 7th postcanine (pc.8.). This tooth is completely covered laterally by the anterior border of the coronoid process. As in specimen C, the crowns of the older teeth are different from those of the younger teeth. The sixth postcanine (fig. 4B), belonging to the older series, has three cusps (anterior accessory, main and posterior accessory) aligned antero- posteriorly. The main cusp and the anterior accessory cusp are about the same size and the posterior accessory is slightly smaller. The posterior cingular cusp lies directly behind the posterior accessory cusp whereas the anterior cingular cusp lies slightly antero-lingually of the anterior accessory cusp. As a result the crown, when viewed from above, has a wide, blunt anterior region and a tapering posterior region. A large and prominent cingular cusp (c.c.) is present on the lingual surface below the junction of the main and anterior accessory cusps. A small poorly defined cuspule is present posterior to this cusp. A poorly developed cingulum (cing.) is present between the two lingual cingular cusps (c.c.) and the posterior cingular cusp. In the fourth postcanine the main cusp is considerably larger than the accessory cusps. The anterior cingular cusp is fairly well developed, but the posterior cingular cusp is small and the cingulum TOOTH REPLACEMENT IN THRINAXODON LIORHINUS SEELEY 487 and cusps upon the cingulum are very poorly developed. When viewed from above, the crown has a wide anterior region and a tapering posterior region, as in the 6th postcanine. The 4th postcanine is shorter but wider than the 6th. In the second postcanine the crown is dominated by the main cusp and the anterior and posterior accessory cusps and the anterior and posterior rep.c. p. A a rep.t. rep.c. . ( PAKS: Fic. 3. Thrinaxodon liorhinus, Specimen D. A, lingual view of the upper postcanines, B, lingual and C, labial view of the lower postcanines. Key to lettering p. 521. cingular cusps are very small. The lingual cingulum is absent. In crown view the tooth is more circular than the 4th postcanine. Within the even-numbered postcanines there is, therefore, in a posterior direction a progressive increase in the size of the accessory and cingular cusps and a progressive decrease in the size of the main cusp relative to the size of the accessory cusps. In addition posteriorly the teeth become progressively more longitudinally ovate. Similar tendencies can be observed in the odd-numbered teeth of the same specimen, but the structure of the crown is slightly different. Unfor- tunately the crown of the 7th postcanine is partially destroyed. The remaining portion is similar to the 5th except that the cingulum and cingular cusps are absent and it is more longitudinally ovate. ANNALS OF THE SOUTH AFRICAN MUSEUM 488 DpD sed 489 TOOTH REPLACEMENT IN THRINAXODON LIORHINUS SEELEY "1aG *d Surszayj0] 0} Aoxy ‘soutueoysod saddn 9y} Jo smaIA UMOJO put [enSury ‘g pue souruvsysod soddn ay} Jo smart uMOJD pue [ensury “vy "d wourtsedg ‘snuiysoy uopoxouuyy “b -o1g hs” Sara 490 ANNALS OF THE SOUTH AFRICAN MUSEUM In the 5th postcanine the crown is dominated by the main cusp and the accessory cusps are small. In contrast to the even-numbered teeth the posterior cingular cusp does not lie directly behind the posterior accessory cusp, but postero-lingually to it. A well-defined anterior cingular cusp is absent; in its place is a series of small cuspules arranged to form a short ridge (a.cing.) running downwards in a posterior direction. Because the posterior cingular cusp is well developed and the anterior cingular cusp poorly developed, the crown has a wide (transverse) posterior region and a slightly tapered anterior region, i.e. Opposite to the conditions in the even-numbered teeth. A small cuspule (a.l.c.) called the anterior labial cusp is present on the labial surface of the tooth below the junction of the main and anterior accessory cusps. The 3rd postcanine is similar to the 5th except that the main cusp is relatively larger and the remaining cusps relatively smaller. The cuspule on the labial surface is absent. It is clear from the above description that the crowns of the odd-numbered series (3, 5 and 7) have the same basic structure and that this is quite distinct from the structure of the crowns of the even-numbered series (2, 4 and 6). In each series, however, the teeth become more complex in a posterior direc- tion. Thus not only do the odd- and even-numbered teeth replace at different times in specimen QC, but in addition the crown patterns of the two series are distinct from one another. The 8th postcanine is an enigma. It is smaller than any of the other lower postcanines with the exception of the first. ‘The main cusp dominates the crown. The only part of the cingulum developed is the anterior cingular cusp. It is discussed on p. 509. The last upper postcanine of this specimen is also of a simpler construction than the penultimate tooth. This is the only specimen in the series in which this phenomenon was observed. Specimen E (fig. 5 B, GC, D) This specimen is larger than D, but the individual teeth are smaller than those of D. Seven functional lower postcanines are present. The remnant (r.pc.) of a tooth is present anterior to the Ist postcanine. Partially erupted 8th post- canines were present on both sides, but have been lost from the specimen. These teeth appeared to have had no functional predecessor as there are no indications of a root remnant in their alveoli. Unfortunately only the 6th and 7th postcanines are well preserved. The crowns (fig. 5 CG and D) of these two teeth are characterized by a mixture of the features present on the crowns of the 6th and 7th postcanines of specimen D. In both, the accessory cusps are nearly as large as the main cusps. In the 6th postcanine the anterior cingular cusp lies almost directly in front of the anterior accessory cusp, but the posterior cingular cusp lies slightly postero-lingually to the posterior accessory cusp. Two well-developed cingular cusps (c.c.) are present on the lingual surface at the base of the anterior accessory cusp. These are followed posteriorly by three "16S *d Sursazj0] 0) Aoy 491 ‘souruvoysod yiZ puv y19 94} JO MOIA UMOID :soutuvoysod yiZ puv yig ‘YS oy} JO MOIA [eNsuTy ‘Fy *S9UIUVI}SOd IIMO] BY} JO MIIA [VNsSUT] ‘g ‘souiuvo}sod soddn oy} JO MOIA yensuly ‘VW ‘q uouwtvedg “ds wopoxpumyy *S ‘O1g iG TOOTH REPLACEMENT IN THRINAXODON LIORHINUS SEELEY Se OEE 492 ANNALS OF THE SOUTH AFRICAN MUSEUM minute cuspules. A well-developed cusp is present on the lingual surface of the base of the posterior cingular cusp. This tooth is almost identical with the 6th postcanine of specimen D. In the 7th postcanine the anterior cingular cusp is presumably represented by a well-developed cusp antero-lingual to the anterior accessory cusp. A single large cingular cusp (c.c.) is present below the junction of the anterior cingular cusp and anterior accessory cusp. No further cusps are 1CM. pc.1. Fic. 6. Thrinaxodon sp., Specimen F. A and B, upper dentition; C and D, lower dentition. Key to lettering p. 521. TOOTH REPLACEMENT IN THRINAXODON LIORHINUS SEELEY 493 present behind this cusp between it and the posterior cingular cusp. A well- developed cusp is present on the labial surface (a.l.c.) antero-labial to the apex of the anterior accessory cusp when seen in crown view. The 5th postcanine apparently had a high dominant main cusp. The anterior cingular cusps are arranged to form a ridge not unlike a similar structure of the 3rd and 5th postcanine of specimen D. Specimen F (Fig. 6C and D and fig. 7) The lower postcanines of specimens F and G are more similar to one another than they are to those of specimens D and E, but because of numerous similarities all four specimens have been placed in Group II. Gicic ~~ \t AC aCCa-a Ss : Wn ‘ ul Ay (RAW fc Fic. 7. Thrinaxodon sp., Specimen F. A and B, lingual and crown views of the 7th lower postcanine. Key to lettering p. 521. Seven functional postcanines are present in the right mandible (fig. 6 D). They are preceded by the remnant of a tooth (r.p.c.) and an unerupted and partially formed tooth (a.p.c.) is present in a crypt behind the last postcanine. The crowns of the individual postcanines are not sufficiently well preserved to permit a detailed comparison of the crown structures of odd and even post- canines, but the differences do not appear to be as marked as in specimens C and D. As in other specimens there is a progressive complication of the crowns in a posterior direction. The crowns are characterized by well-developed cingular cusps. These are particularly well preserved in the 7th functional postcanine (fig. 7). In this tooth the accessory cusps are large (a.a.c., p.a.c.), but not as large as the main cusp. An anterior cingular cusp is present antero- lingual to the anterior accessory cusp. It is followed by two large cingular cusps (c.c.). The posterior portion of the cingulum consists of a series of six cuspules which, with the exception of one, increase progressively in size in a posterior direction. Well-developed cuspules are present on the labial surface of the crown at the base of the anterior and posterior accessory cusps (a.l.c., p.l.c.). With the exception of the cusps on the posterior portion of the cingulum this tooth is almost identical to the 7th postcanine of specimen E (fig. 5 C and D). ANNALS OF THE SOUTH AFRICAN MUSEUM 494 *10G ‘d Sursa}30] 0) Ady ‘soutuvoj}sod 1taddn ay) Jo sMoIA UMOID “D pur [ensuly “g ‘soutuvoysod saddn oy} Jo Mara Tensuly “Vy *s) usumtvadg “ds uopoxnuuyy *g “SIA TOOTH REPLACEMENT IN THRINAXODON LIORHINUS SEELEY 495 Specimen G (fig. 9) The jaw contains eight functional postcanines. The 8th, 5th and 3rd appear to have erupted shortly before death. The 6th tooth is missing. The remaining teeth are all badly worn. This is the only specimen in which post- canines show advanced signs of wear and in which replacement does not appear to be strictly alternate. The crown of the 5th postcanine (fig. 9 B and C) is almost identical to the 7th functional postcanine of specimen F (fig. 7) except that the posterior cingular cusp (p.c.c.) lies lingual to and not posterior to the posterior accessory ‘cusp. The 8th postcanine of specimen G is identical to the 5th except that it lacks the posterior labia] cusp and is more longitudinally ovate. The anterior part of the cingulum of the grd postcanine does not support the three large cusps that are present in this position in the 5th and 8th postcanine, but consists of a series of six small cusps arranged to form a ridge which terminates anteriorly at the base of the anterior accessory cusp. The posterior cingulum supports a single well-developed cusp (p.c.c.) postero-lingual to the posterior accessory cusp. This tooth is similar to the 5th functional postcanine of specimen D (fig. 4B). In postcanines nos. 8, 5 and 3 the main cusp becomes progres- sively larger in an anterior direction and the accessory cusp progressively smaller. The 2nd and 4th postcanines although badly worn are characterized by large anterior cingular cusps and poorly developed posterior cingular cusps. In this respect they are not unlike the 6th postcanines of specimen D (fig. 4 B) and Specimen E. Upper postcanines, Specimen D (figs. 3 A and 4 A) Without exception the upper postcanines have crown patterns less complex than those of the corresponding lower teeth. Six functional postcanines are present in specimen D. The 6th is preserved partially erupted. The 1st and grd appear to have erupted shortly before death. The 5th already has a large pit for a replacing tooth at its base. The Ist postcanine (fig. 4 A) consists of a large main cusp, which is slightly recurved; the anterior accessory cusp is missing, but a small posterior accessory cusp (p.a.c.) is present. A cingulum and cingular cusps are absent. The 3rd postcanine has a large main cusp which is flanked by smaller anterior and posterior accessory cusps. A small posterior cingular cusp is present on the lingual surface at the base of the anterior accessory cusp. No other cingular cusps are present. The 5th postcanine is similar. The 6th post- canine is almost identical to the Ist postcanine except that it possesses a very small anterior accessory cusp. The 2nd postcanine appears to have both anterior and posterior cingular cusps. Specimen E (fig. 5 A) Six functional postcanines are present. The 1st, 3rd and 5th are only partly erupted indicating that they erupted shortly before death. The structure of the crowns of the postcanine teeth are identical to those of specimen D. ANNALS OF THE SOUTH AFRICAN MUSEUM 496 ‘IND. "12G ‘d Sutsa}j,0] 0) Avy ‘squIueo}sod YIg 0} PUS 9Y} JO SMITA UMOI “) . ‘squruvojsod 1g pus wy sy} JO MoIA [ensuly “gq ‘sautuvoysod 19MO] oY} JO MATA [eNnsur] “V “sy uowtvedg “ds uopoxnuuyy, *6 “og T3D\Gl ‘530 ‘a DD rou) UW ro G 7 5 5 oe Moe yl tal DS ; ‘SDD . \ ‘oped af ) ‘god TOOTH REPLACEMENT IN THRINAXODON LIORHINUS SEELEY 497 Specimen F (fig. 6 A and B) Six functional postcanines are present in specimen F.. The Ist, 3rd and 5th erupted shortly before death and well-developed replacing teeth are present in pits at the bases of the older 2nd, 4th and 6th postcanines. The crypt for an additional tooth is present behind the 6th postcanine. The structure of the crowns of the upper postcanines is identical to those of specimens D and E. Specimen G (fig. 8) Six functional teeth are present. In the even-numbered series the 2nd and 4th have well-developed pits for replacing teeth above them, but the replacing tooth for the 6th had already erupted before death. The 2nd and 3rd teeth are almost identical in structure and possess well-developed anterior and posterior cingular cusps. The 4th is badly worn. The 5th and 6th lack well-developed posterior cingular cusps. The 6th has two well-developed cusps on the labial side of the tooth (p.l.c., a.l.c.) one above the anterior and one above the posterior accessory cusp. Group III Lower postcanines. Specimen H (fig. 10 G and D) The lower postcanines of this specimen are characterized by poor develop- ment of the cingulum and cingular cusps. In this respect they are quite different from the lower postcanines of Groups I and II. The lower postcanines of specimen H closely resemble the upper postcanines of this and other specimens. Seven functional postcanines are present. The crown of the 7th consists of the characteristic main cusp, anterior and posterior accessory cusps, and anterior and posterior cingular cusps. In the 5th and grd the crown structure is similar except that the main cusp becomes progressively more dominant in an anterior direction. Upper postcanines. Specimen H (fig. 10 A and B) Six functional postcanines are present in the right maxilla. The cingular cusps are very poorly developed and appear to be entirely absent on most of the postcanines. Discussion Extension of Growth series of ‘Thrinaxodon’ The present study has confirmed Parrington’s (1936) conclusion that alternate replacement took place in the postcanine teeth of Thrinaxodon. Parring- ton showed that it was possible to divide the specimens of Thrinaxodon that he studied into two groups; a younger group in which the Ist, 3rd and 5th maxillary postcanines were younger than the remaining ones and an older group in which the 2nd, 4th and 6th maxillary postcanines were younger than the others. 498 ANNALS OF THE SOUTH AFRICAN MUSEUM The skull lengths of Parrington’s younger group vary between 61 mm. and 79 mm. (see table I). One of Parrington’s specimens, viz. his E (C in the present investigation), was prepared in acid and it has been confirmed that the Ist, 3rd and 5th maxillary postcanines were the younger. Four additional skulls falling within this size range were studied. These are specimens D, E, F and G. In all of them the Ist, 3rd and 5th postcanines were the younger. The skull lengths of these specimens are given in table I. Parrington concluded that the 1st, 3rd and 5th teeth were replaced at least once and that the and, 1CM A ol pc.2 rep.c. 6 CO 2. nem 1CM pers: Fic. 10. Thrinaxodon sp., Specimen H. A and B, upper and C and D, lower dentitions. Key to lettering p. 521. TOOTH REPLACEMENT IN THRINAXODON LIORHINUS SEELEY 499 4th and 6th were also replaced at least once. Parrington’s two stages are diagrammatically shown in figure 11 A. In this figure they are labelled II and III. Fully erupted teeth are drawn in black below the alveolar border and -unerupted replacing teeth in dotted lines above the alveolar border. Only the portion below the alveolar border of partially erupted teeth is shown in black. . Parrington’s younger group Parrington’s older group Fic. 11. Tooth replacement in Thrinaxodon. A, diagram to illustrate simple alternate replacement concluded by Parrington to have been present in Thrinaxodon. Additional growth stages I and IV have been added. B, diagram to illustrate a possible way in which posterior migration of the post- canine row could have been achieved. In figure 11 A, stage II, the 1st, 3rd and 5th are partially erupted and in figure 11 A, stage III, the 2nd, 4th and 6th of stage II have been replaced and partially erupted teeth are present in these alveoli. Specimens both larger and smaller than those available to Parrington were available for the present study. See table I. 500 ANNALS OF THE SOUTH AFRICAN MUSEUM TABLE I Designation of Length of skull Designation of specimens in (tip of snout Youngest teeth specimens in Parrington’s to posterior end present paper (1936) paper of sagittal crest) I 2 A 5 Omen 36 mm. x x x B 58 mm. x x x B 61 mm. x x x C 63 mm. x x x Cc E 68 mm. x x x F 70 mm. x « x D 73 mm. < x x E 78 mm. x x x G 78 mm. x x x F 78 mm. x x x G 80 mm. x x x H 82 mm. x x x H I 85 mm. x x x I 86 mm. x x x In specimen A (fig. 1 A) the end, 4th and 6th postcanines were younger than the Ist, 3rd and 5th. This skull is incomplete, but as reconstructed by Brink (19550) it could not have had a skull length of much more than 36 mm. It is the youngest specimen available and in comparison with older specimens the teeth are minute. It has seven functional postcanines. This fact is discussed on p. 502. In specimen B (fig. 1 A and B) (skull length 58 mm.) the 2nd, 4th and 6th postcanines were also the youngest. The difference in the size of the corre- sponding teeth in A and B suggests that several growth stages intervened between specimens A and B. In specimens with skull lengths between 61 mm. and 80 mm. the ist, 3rd and 5th teeth were the younger. From 80 mm. to 85 mm. the end, 4th and 6th were the younger. In a damaged Thrinaxodon skull (referred to as specimen I) in the collection of the South African Museum with a skull length of approximately 86 mm., the Ist, 3rd and 5th teeth appear to have been the younger. The additional material shows that the series available to Parrington covered only part of the growth series and that the individual postcanines were replaced several times. The additional growth stages are shown in figure 11 A, stages I and IV. Stage I represents specimens A and B in which the end, Ath and 6th are the youngest. The older teeth in stage I have been numbered 1 and the younger teeth 2. By the time stage IV is reached the teeth numbered 1 have been replaced twice and, therefore, the teeth in the corresponding positions have been numbered 3. The teeth numbered 2 in stage I are replaced once and the teeth in the corresponding position in stage IV are, therefore, numbered 3. It is, however, apparent from an analysis of the postcanine teeth of the specimens available that in Thrinaxodon the replacement pattern is more complex than the simple alternation shown in figure 11 A. Further detailed consideration of the specimens is therefore necessary. TOOTH REPLACEMENT IN THRINAXODON LIORHINUS SEELEY 501 Posterior migration of the postcanine row The number of postcanines in the jaws studied varies slightly. In many cases remnants of postcanines are present in front of the functional postcanines and in some cases unerupted or partially erupted teeth are present behind the functional row. In some cases these teeth do not appear to have had predeces- sors. This information is summarized in table II. The number of functional teeth appears to be fairly constant. All the maxillae, with the exception of specimen A, have six functional teeth and all the mandibles with the exception of specimens D and G have seven functional postcanines. TABLE II UPPER POSTCANINES Rare LOWER POSTCANINES Anterior No. Unerupted Anterior No. Unerupted tooth functional posterior _ Total tooth functional posterior Total remnants p.c. p.c. remnants p.c. p.¢. A 7 7 B 6 6 C 6 ORT renin) 7 (1) 9 D 6 6 (1) 8 9 E 6 LO (1) 7 (1) 9 [8 6 (1) 7 (1) 7 (1) 9 G 6 (1) 7 8 8 H 6 (1) 7 7 7 The constant number of functional postcanines in most specimens, the frequent occurrence of a tooth remnant anterior to the functional row and the presence behind the postcanine row of an unerupted postcanine which does not appear to have had a predecessor, suggest that during growth the anterior postcanines were lost and that new postcanines were added posteriorly. In mandibles of specimens C, E, F and G (figs. 2, 5, 6 and g) it would be expected in terms of alternate tooth replacement that the 1st functional tooth would be a younger tooth erupted shortly before death or at least that it would have a well-developed replacing tooth in a pit at its base. This is not the case. In all four specimens the first functional tooth is firmly held by attachment bone and there is no indication of a replacing tooth at its base. The fact that the first functional tooth was not about to have been replaced in these four specimens suggests that in a more advanced growth stage the first functional tooth became the tooth remnant anterior to the postcanine row which is preserved in some specimens. This fact supports the conclusion that postcanines were lost anteriorly and added posteriorly during growth. In other words, the functional row of postcanines migrated backwards during growth. This appears to be so arranged that the number of functional postcanines remained roughly constant during growth. An explanation of this nature could account for the varying postcanine count attributed to Thrinaxodon in the literature. For example, if in the lower jaw the new tooth behind had erupted sufficiently to be considered part of the 502 ANNALS OF THE SOUTH AFRICAN MUSEUM functional row the number would be given as eight, if not as seven. In figure 11 B, a series of growth stages are shown in which teeth are progressively lost in front and added behind. This has been superimposed upon the simple alternate replacement shown in figure 11 A. Although this model appears to be theoretically sound it only agrees with the state of eruption of the post- canines in some of the specimens, to wit those in which the 2nd, 4th and 6th postcanines are replacing. If teeth were added and lost in the manner suggested in figure 11 B, the state of the postcanine row at each growth stage would appear to be identical, i.e. it would always appear that the 2nd, 4th and 6th teeth were replacing because teeth are added and lost at the same rate. This is clearly not the case as specimens are also available in which the Ist, 3rd and 5th teeth are replacing. Posterior migration of the postcanine row appears to be a characteristic of many of the advanced cynodonts, e.g. Scalenodon (Crompton, 1955); South American gomphodont cynodonts (Patterson and Olson, 1961) and Diademodon (Fourie, 1963). There is ample evidence that the canines replaced several times. As the replacing canines were larger than their predecessors the new teeth invaded the region occupied by the anterior postcanine. It appears that in order to compensate for the loss of anterior postcanines further teeth were added at the back of the tooth row. Replacement cycle In figure 12 an attempt has been made to fit the postcanine rows of the specimens studied into an orderly sequence. This series indicates how posterior migration of the postcanine row was achieved in such a way that in some speci- mens the odd-numbered teeth are replacing and in others the even-numbered teeth are replacing. The upper postcanines are shown in figure 12 A and the lower postcanines in figure 12 B. In this figure teeth drawn in dotted lines above the alveolar border in the upper jaw and below it in the lower jaw had not yet erupted. The first growth stage (I) in the series in figure 12 A is a diagrammatic representation of the upper postcanine row in specimen H (fig. 10). The end, 4th and 6th teeth are in the process of erupting, but the 6th is further erupted than the 2nd or 4th (replacement wave). Replacement teeth are present above the grd and 5th postcanines, but not above the 1st. An unerupted tooth is present behind the 6th in position 7. There was no functional predecessor for this tooth. It is larger than the replacing tooth at position 5. In specimen A (stage 1 A on the right of fig. 12 A) seven teeth are present, of which the 2nd, 4th and 6th are the youngest. It is possible that the state of the postcanine row in specimen A is similar to that of H except that the unerupted postcanine in position 7 has erupted and is functional. In a later growth stage (fig. 12 A, II) the 1st functional postcanine which had no replacement tooth in stage I has been shed. The partially erupted 2nd, TOOTH REPLACEMENT IN THRINAXODON LIORHINUS SEELEY 503 4th and 6th postcanines of stage I are now fully erupted. Because one post- canine has been lost the 2nd, 4th and 6th of stage I become the ist, 3rd and 5th of stage II. Consequently the replacing teeth that were developing in stage I above the grd and 5th replace the previous teeth and erupt in stage II as the end and 4th. The unerupted additional tooth (position 7) of stage I erupts as the 6th in stage II. Consequently although in both stages I and II the and, 4th and 6th teeth are the youngest, the teeth referred to are not the same in both stages. Replacing teeth are developing above the rst, 3rd and 5th teeth in stage II. Stage II appears to correspond with specimen B. In this specimen there is no indication of an unerupted tooth behind the functional postcanine row. As the functional 1st, 3rd and 5th in specimen B are considerably larger than the corresponding teeth in specimen A, several growth stages probably separated these two specimens. This point will be considered again later. In stage III the replacing teeth developing above the Ist, 3rd and 5th postcanines in stage II have erupted. In stage II replacing teeth are erupting in the 2nd, 4th and 6th alveoli. In stage III these are fully erupted. An additional tooth is developing behind the functional row and the development of a replacing tooth for the 5th has commenced. The state of eruption and replacement of the postcanines in this stage corresponds exactly to that present in specimens E and F (figs. 5 and 6). Specimens C, D and G appear to be slightly in advance of stage III because the replacing tooth for the 6th postcanine has erupted and replaced the old functional tooth in this position. This is illustrated in stage III a on the right of figure 12 A. Stage III is followed by a stage IV. Apart from the fact that the first functional tooth is now one position further back, this stage is identical to I, thereby completing the cycle. The replacement cycle (fig. 12 B) in the lower jaw is basically the same as that concluded to have been present in the upper jaw. The only real difference is that the number of functional teeth is seven instead of six. In growth stage I (fig. 12 B) the 1st, 3rd, 5th and 7th teeth are in the process of erupting. Replacing teeth are developing below the 2nd, 4th and 6th postcanines. This stage corresponds with the postcanine row of specimen H. In the right mandible of this specimen the old functional first tooth had not been shed, but it was shed on the left. This tooth has been indicated by a dotted outline in stage I. In specimen H there is no indication of an unerupted post- canine behind the last functional tooth. Specimen D (fig. 3 B) corresponds with stage I in regard to the state of eruption of the first seven postcanines, but a small 8th tooth is present behind the postcanine row. The state of eruption of the postcanines in specimen D is shown diagrammatically in stage I A (on the right-hand side of fig. 12 B). This specimen is discussed in more detail below. In stage II the replacing teeth which were developing below the and, 4th and 6th in stage I are erupting and replacing teeth are developing for the 3rd, 5th and 7th postcanines, but not for the first postcanine. An unerupted tooth is developing behind the 7th postcanine. The 1st, 3rd, 5th and 7th teeth that 904: ANNALS OF THE SOUTH AFRICAN MUSEUM DVS TOOTH REPLACEMENT IN THRINAXODON LIORHINUS SEELEY ‘IIMOT “gq pue szoddn ‘vy ‘souruvojsod oy} [Je Jo s[9Ao yuous0¥"]doy ds UOpOXDULLY [. 9) & 506 ANNALS OF THE SOUTH AFRICAN MUSEUM were partially erupted in stage I have in stage II erupted further. The 7th is fully erupted, but the Ist is only partially erupted. This stage is not repre- sented in the material studied. In stage III the partially erupted 1st postcanine of stage II is fully erupted and retained. The other odd-numbered, functional teeth (3rd, 5th, 7th) of stage II have, however, been lost in stage III and their replacing teeth which were developing in stage II are now erupting. The retention of the Ist postcanine in stage III is presumably correlated with this tooth having no successor and being younger than the other functional odd- numbered teeth in stage II. Replacing teeth are developing below the end, 4th and 6th postcanines and a partially erupted tooth is present behind the functional row. This stage corresponds with the postcanine row in specimens C, E and F (figs. 2, 5 and 6). In these specimens the remnant of a tooth is visible anterior to the functional postcanine row. This is discussed below. Specimen G (fig. 9 A) is similar to stage III in the state of eruption of the first five post- canines, but the partially erupted 8th postcanine shown in stage III is fully erupted. The next stage, IV, is identical to I, except that the first functional tooth is one position further back. The ist postcanine of stage III which has no replacing tooth, is retained as a tooth remnant. This remnant would presumably have been retained in the next three stages of the replacement cycle. Edmund (1960, 1962) has made a thorough study of tooth replacement in reptiles. He has shown that the spacing between hypothetical impulses which initiate the development of teeth and which travel along the free margin of the dental lamina can account for the fact that in some reptiles the replacement pattern is strictly alternate, whereas in others waves of replacement pass from back to front or front to back along alternately numbered tooth series. Back to front waves are present in all the specimens of Thrinaxodon studied. Edmund has shown that the basic units in the dentition of all reptiles are