ite tw) aang de toe the Hew Petiek re ; : TL bo het Galt eae F ! vie . nA o) Beas be hiaser se OW BOER a : oe Pt be Qe esis cae teat was payee 3 Ae ‘hal oh bathe feet a}: uv ote ” ane . ve ch Ratt ® aly wr) i cet wey ; Bits ee * - nad 7 Say a fy ei PPT ag LAP EL RE te fie gl ot ocd aed Ha Frye aie wasse ind r : ee : sts ‘ Foy ge ~ ies ; a haved ok} te set sab eR te AY od aM dove & ; hat ene te tela aa ree fee Fame et he GLEE spit th ot a * } ie 1d trie aie a Peon “3 , ” FE Pn Cy ae + 7 ‘ . Sent bj wo a on Bhd +s . ‘ : .! ; be “Hun bide n ey A Pasig Sel pate fe dey ceed et lado pin hoatey i a 3 : ae degrade ‘ Mt Yi aad 4 “ a phys Se ee ert - a 5 + rh @ \ a w= 5 <9" Abid Cea Rit thee Reena, au rte tk acme rot . 2 ews send ~ ay wi aese iv cieeiets Sone eae” he ere oe fate e ae vi awaneey G00 Ses <9 forte} ° “se ao ik a ae dtereresciperets : H i ye : wv iiriye a A ce wteneit Ht \4 ; : i. Saar Va die wa * uy bp ps | ee lek ede a ~* re wt wee giert tan» Bae Sm aR vee Leaky el ees) ee \ l H, eh al | Pin ee ann iit SSS. ( : =e ‘ Ee & ee ce Nes Ge 4 | | i ] 3 — Se? ig sg “ iM roe age y bi ah SEY OEM ae a es oe = Rip eacac (lil ork ‘ Oe - rie i Be oneeeae iaY) Pi ec ee sentry ate a HEPA COE nen” “Siireraatseueeeeo sein qt ae a, ; eae “4 Th, EY, Ht fi bait ane Ea f ease dl Ea as Le aga a OL meres ma ard Fag TEED Ee ratselid | 4 hats | | Se 7S Sener oe fragt aie micttinabiay U ’ ¥ 7 hace 1 ie A 1 % P = L - i 4 i os ot ‘ =) 7 fe y i ioe ‘ Se) 1 — t ~ y 7 x - 2 ? te i i - my ANNALS OF THE SOUTH AFRICAN MUSEUM ae - VOLUME XLV ANNALS OF THE SOUTH AFRICAN MUSEUM VOLUME XLV _PRINTED FOR THE TRUSTEES OF THE SOUTH AFRICAN MUSEUM 1959 — 1960 pes | ey ¥ Wy 5 ¢ ( 7 * a | : ae | Le rs 2 @ @ PRINTED IN THE |THE RUSTICA PRE: fod i V7 Free ALS ee ban a : * a 1507.62 v.45 IGFS94-6CO LIST OF CONTRIBUTORS BARNARD, K. H. Contributions to the knowledge of South African marine mollusca. Part II. Gastropoda: Prosobranchiata: Rhachiglossa (published June 1959) .. Bone, E. L. See SINGER, R. Brown, A. C. A new Solifugid Arachnid from Table ee oe es eee sp. n. (published July 1960) : : DAY, J. H. The Polychaet Fauna of South Africa. Part 5. Errant species eae ecg off sa Coasts (published May 1960) Juss, R. A. . Note on locality records of Freshwater Fishes presented by F. D. McKean to the South African Museum (published April 1960) . Minrarp, N. A. H. Hydrozoa from ships’ hulls and eee i in ite Town Docks ae July 1959) Sincer, R., and Bonk, E. L. Modern Giraffes and the Fossil Giraffids of Africa (published July 1960) Warnor, F. H. Additions to the South African Museum Collection of Marine Fishes eee April 1960) : = - ae as é sae Ta.sor, F. H. Notes on the Biology of the Lutjanidae (Pisces) of the East African Coast, with special reference to L. bohar (Forskal) (published July 1960) SMii }HSOT NUAN INSTITUTION Page DD 261 260 239 375 257 549 JUN 7 1961 ‘i oa ALL i. ae iy We. fo IN THIS VOLUME . : Page s (Eunicidae), Day, 1960 va oy: oe Sy, ae Be iar vio) syllis (Syllidae), Day, 1960... ne eg ite ot as be Be) Adinopsis Afritrophon Afrivoluta Afrocominella Aglaophamus Allothunnus .. Amblyosyllis .. Anaitis .. Anaitides Ancilla Antinoe Aphareus Aphrodita Aprion Arabella _ Aspella Autolytus Babylonia Barbourisia Bhawania Bougainvillia .. Bramatherium Bullia .. Burnupena Callithea Campanularia .. Cancellaria Cantharus Charitodoron.. Chloeia Clavisyllis Clytia .. Columbarium Columbella Cominella Coralliobia Coralliophila .. Cronia Ctenopoma Demoulia Dibaphus LIST OF GENERA AND SUBGENERA Page 142 207 23 152 327 258 313 208 296 61 281 568 274 551 363 232 317 147 257 294 242 519 123 158 50 248 12 150 i44 295 320 248 234 180 169 193 189 229 260 122 RE Diopatra Dorsanum Dorvillea Drilognathus .. Drilonereis Drupa.. Drupella Ehlersia Engina Enipio. . Ephesia Epidiopatra Epinephelus .. Eteone Eulagisca Eulalia Eumida Eunice Euphione Euphrosyne Eurythoe Euthalenessa .. Euthria Exogone Fasciolaria Fulgoraria Fusivoluta Fusus .. Galeodes Genetyllis Giraffa Giraffokeryx .. Clycera Glycinde Goniada Gonothyraea Griquatherium Grubea Halosydna Haplochromis 466 et passim 143 298 375 et passim 519 328 331 333 249 315 282 260 Haplosyllis Harmothoe Harpa.. he Helladotherium Hemilepidia ‘Hindsia Hololepida Honanotherium Hyalinoecia Hydaspetherium Hypoeulalia Imbricaria Indratherium Kefersteinia Kirchenpaueria Labrorostratus Laeonereis Lagisca Ae Lamellisyllis .. Langerhansia Laomedea Latiaxis Latirus Leiodomus Leodice .. f Lepidasthenia Lepidonotus .. Leptoecia Libytherium .. Lovenella Lumbrineris .. Lutjanus Lysidice Malmgrenia . Marginella Marphysa Mastigonereis .. Melapium Melongena ied Micronephthys .. Mitra . Murex.. Myrianida Mysta .. Nassa .. Nassaria Neanthes Nephthys 1A LIST OF GENERA AND SUBGENERA 498, -- 491,495, Neptuneopsis . . Nereis... Nothria Notocirrus ; Notophyllum .. Stauronereis Obelia. . ae Odontosyllis .. Okapia Oliva . Onuphis 2 Ophioglycera.. Orangiatherium Orasius Orthopyxis Palaeotragus .. Paleanotus Pareurythoe Pelmatochromis Perigonimus Perinereis Peristernia Pharyngeovalvata Pholoe. . Phyllodoce Phyllosyllis Pionosyllis Platynereis Plumularia Polyeunoa Polynoe Pristipomoides Procerastea Progirafia Propalaeomeryx Protomystides Psammolyce .. Pseudonereis .. Pterocirrus Pterosyllis Pterothrissus .. Pusia .. Pusiostoma Pyrene Rapana ; Rhamphobrachium .. Rhizorhagium Samotherium. . Sarsia .. 3 330 -+ 447,491, 525 515, 519 248 515, 519 294, 295 LIST OF GENERA AND SUBGENERA Page Page 281 ritonalia) 4 52 Bs re sta wig 212 260 Trophon an we wis set Od 290 Trypanosyllis . . ee aed eh wees 303 Tubulara -. :. oe ae sewn, 240 417 et passim Turricula a he we Se 52 a 8595 Typhis es a ks Sig J 2TO ee 320 - Sphaerosyllis . . 316 U Staurocephalus .. 371 Urosalpinx .. as G te ragt Stauronereis .. ee Sees < . *Steggoa.. a es e Sur GOO My _ Stegolophodon ae Bs Se. RBA Vasum a vi Re os 36 -Sthenelais .. ia ae 25 |: 289 Vexillum as aye ae ae 52 _ Syllides oe i ue se REO Vishnutherium ans as ye ARNO - Syllidia ss ue wa al ee SOm Volema bs be afi gel an TAS “Syllis .. me oY ia ee SOT Voluta Ms ze as M, 18 Sylvanocochlis ue ‘ As 60 Volutocorbis .. “es ss, ae 24 ; Syncoryne CONTRIBUTIONS TO KNOWLEDGE OF S.A. MARINE MOLLUSCA 7 34° 33'S. 18° 21’ E., 318 metres (von Martens). Cape Point, N. 50° E., 180 fathoms, two (Shackleford). The Pieter Faure obtained no other examples of this species. Marginella augusta Thiele 1925. Thiele. loc. cit., p. 193, pl. 33 (21), fig. 28. 95° 11'S. 23° 2’ E., 500 metres (Thiele). One specimen, 17 X 6-5 mm. (outer lip broken), without precise locality (S. Afr. Mus. P.F. coll.). Remarks. Thiele’s figure shows 4 columella pleats, but in his description he does not reckon the anterior fold of the columella as a pleat. Very like brocktoni, but with 4 columella pleats. Marginella neglecta Sow. 1846. Sowerby. Thes. Conch., i, p. 390, pl. 76, figs. 135, 136. 1852. Krauss. Arch. Naturg., i, p. 38 (reevei). 1853. Gaskoin. Ann. Mag. Nat. Hist. (2), xi, p. 359 (rufula). 1903. Von Martens. D. Tiefsee Exp., vii, pl. 3, fig. 3 (reevei) (not the specimens = clara). 1917. Tomlin. loc. cit., p. 283 (neglecta), p. 294 (reevet), p. 295 (rufula). No specimens definitely referable to this species were obtained by the Pieter Faure. Marginella clara ‘Thiele 1903. Von Martens. D. Tiefsee Exp., vii, p. 35 (not the fig. = reevei = neglecta). moet Einele. loc. cit., p. 192, pl. 33 (21), figs. 21, 22. 35, 16'S. 22° 26’ E., 155 metres (von Martens, Thiele). Very like neglecta, or possibly a not fully mature atractus. Marginella bensoni Rve. 1865. Reeve. Conch. Icon., pl. 27, fig. 158. 1904. Smith. 7. Malac., xi, p. 32, pl. 2, fig. 20 (dulcis). ? 1925. Thiele. loc. cit., p. 195, pl. 33 (21), figs. 35, 36 (laetztia). A faint spiral band below the suture and another on lower part of base (Thiele). Thiele said his species was near adela, but smaller. It certainly seems near bensont. By 10'S. 22° 26’ E., 155 metres (Thiele). P.F. specimens from the following localities seem referable to bensonz; they are smaller than Thiele’s specimens. Off Umkomaas River (Natal), 40 fathoms; off Cape Morgan, 47 fathoms; 34° 55'S. 25°55 E., 67 fathoms; False Bay, 22 fathoms (S. Afr. Mus. P.F. coll.). 8 ANNALS OF THE SOUTH AFRICAN MUSEUM Marginella differens Smith 1892. Sowerby. Mar. Sh. S. Afr., p. 20 (bulbosa, non Reeve). 1904, Smith. 7: Malac.. xt p. 32, pi--2, fig..19: 1916. Shackleford. Ann. S. Afr. Mus., xiii, p. 194, text-figs. 3, 4 (taylort, non Olsson). 1919. Tomlin. Proc. Mal. Soc., xiii, p. 65 (barnardi, nom. nov. for taylori, preocc.). The type and cotype of taylort in 8. Afr. Mus. do not seem distinguishable from other specimens identified by Shackleford as differens. differens forma eugenes n. Shape of differens, but larger and slightly less obese. 4 whorls. Columella pleats 4 with traces of 1-4 additional pleats posteriorly. Outer lip not reaching suture above, scarcely shouldered, not varicoid, internally with 18-20 well- developed denticulations or plicae, slightly inflected at upper end, aperture indented anteriorly. 7-5 x 4:8 mm. (Type), another 8 x 5 mm. Uniform pale buff. Off Cape Natal (Durban), 85 fathoms, one; off Umkomaas River (Natal), 40 fathoms, one; 34° 27'S. 25° 42’ E., 256 fathoms, 5; 34° 26 sympa 124 fathoms, 6; Gericke Point (Knysna area), 42 fathoms, one; off Cape St. Blaize, 37 fathoms, 2; the same, 125 fathoms, 11; off Cape Infanta, 46 fathoms, one (5) Air Mins: weaker colle Remarks. ‘The 8 xX 5 mm. specimen from off Gericke Point was seen by Tomlin and Shackleford, and considered to be a n. sp. but ‘too poor’ for description. Actually it seems in quite fair condition, though the surface is not glossy. Therefore the glossy, but slightly smaller, specimen from off Umkomaas River is taken as the type (S. Afr. Mus. no. A8786) of this large form. Only further research will show whether a name is really necessary; the difference in size between this form and normal differens (5-8 mm.) is no greater than that found in musica (15-25 mm.). Marginella burnupi Sow. 1897. Sowerby. Append. Mar. Sh. S. Afr., p. 10, pl. 6, fig. 35. 1929. Dautzenberg. Faune Col. Franc., 111, 4, p. 381. 1932. Turton. Mar. Sh. Port Alfred, p. 42. Not Thiele. 1925. loc. cit., p. 192, pl. 33 (21), figs. 24, 25. Conical, widest slightly above middle, narrowing anteriorly, spire flat. Columella pleats 5 (Sowerby), 6-8 in larger specimen. Outer lip incrassate (when mature), internally plicate. 4 42 4 : “| a yh coneneieag ii. Spiral sculpture irregular, coarser and finer lirae more or es alter- Mating. 2 eee, ee ee ee og awe b. Profile angular, with shoulder knobs. i. Protoconch small, diam. 1:5 mm. Shell usually with dark ee lines, usually in pairs, sometimes unicolorous. . . . . trapezium ii. Protoconch large, diam. (2°5 mm. worn) 3°5-5 mm. Shell uni- colorous 2 ee ee ee ee Fasciolaria filamentosa Lam. 1880. Von Martens. Mauritius G Seychellen, p. 245. 1911. Strebel. loc. cit., p. 34, pl. 6, figs. 33, 34; pl. 7, figs. 35-37; pl 15, ieee 1929. Thiele. Handbuch, i, fig. 377 (radula). 1952. Satyamurti. Bull. Madras Govt. Mus., n.s. 1, 2, pt. 6, p. 185, pl. 17, figs. 9 a-c. 1952. Braga. Anais Est. Zool. Invest. Ultramar., vii, 3, p. 73, pl. 2, fig. 2. No parietal callus. Aperture longer than spire. Profile convex, sometimes with slight shoulder. Protoconch 14 whorls, small (see Strebel, pl. 7, fig. 37), smooth. Postnatal whorls 7; 1st with 7 axial ribs, increasing to 11 on middle whorls, on later whorls forming only slight swellings on the shoulder; crossed by 4 spiral lirae on 1st whorl, increasing in number on middle and later whorls, more or less regular above shoulder, but alternately broad and narrow below, the larger ones becoming better marked on the base. Columella pleats distinct, glaze narrow; outer lip internally plicate. 172 X 71 mm. (Strebel). Radula, lateral plate with 13 cusps (Thiele, fig. 377). Natal (Sowerby), Mozambique (Braga). Distribution. Red Sea, Zanzibar, Réunion, Mauritius, Seychelles, Madagascar and Indo-Pacific to Japan. Remarks. A variable species (Strebel), but distinguished by the absence of the parietal callus. Turton’s record from Port Alfred (1932. Mar. Sh. Pt. Alfred, p. 49) is probably not filamentosa. Sowerby’s Natal record was probably a dead shell. Fasciolaria rutila Watson Figs. 18(a), 19(c) 1882. Watson. 7. Linn. Soc. Lond., xvi, p. 335. 1886. id. Challenger Rep., xv, p. 242, pl. 13, fig. 6. 1903. Sowerby. Mar. Invest. S. Afr., ii, p. 227, pl. 3, fig. 2 (juv. shell and radula). 1903. Von Martens. D. Tiefsee Exp., vii, p. 30. ? 1923. Odhner. Géteb. K. Vet. Handl., xxvi, p. 6 (= Med. Géteb. Mus., xxiii). ? Tomlin. Ann. Natal Mus., v, p. 290. Thin-walled, profile of whorls convex, without any shoulder. Protoconch (uncorroded) large, subglobular, 1 or 14 whorls, diam. 5:5—7 mm., alt. 5-6 mm. (corroded: diam. 3°5, alt. 3 mm.). Postnatal whorls 5, smooth in appearance, but with numerous fine spiral lirae closely set, c. 11-12 on 1st whorl, becoming CONTRIBUTIONS TO KNOWLEDGE OF S.A. MARINE MOLLUSCA 75 a little stronger and more widely spaced on later whorls. Parietal callus present; columella with 3 pleats, but posterior one feeble and obscure in large specimens. Canal long, narrow, sigmoid; outer lip thin, not plicate internally. Periostracum thin. 132 X 51 mm. Operculum 35 X 19 mm. in 132 mm. shell. White with pale yellowish brown periostracum, operculum dark brown. Radula with 230-250 rows, usually an accessory denticle externally on both sides on the central plate; lateral plate with 8 cusps in juv. (figured by Sowerby), in adult 13 on one side and 14 on the other (2 specimens), or 15 and 16 resp. (one specimen) (excluding the tiny internal denticle). Fre. 1G: (a) Fasciolaria rutila Watson, protoconchs of two shells. (6) F. lugubris Rve., two views of protoconch extracted from (c) egg-capsule. (d) F. heynemanni Dnkr., protoconch. imine: 95°45. 18° 37 E., 150 fathoms (Watson); 33° 41 S. 18° E., 178 metres (von Martens); off Cape Peninsula, 154 fathoms (Sowerby: juv.); off Cape Point, 85-145 fathoms (S. Afr. Mus. P.F. coll.); 33° 11'S. 17° 57’ E., 77 metres; 31° 39’ 8. 16° 55’ E., 287 metres (s.s. Africana) ; ? off Umhloti River (Natal), 40 fathoms (Sowerby). Dead? St. Sebastian Bay, 40 fathoms (Odhner); off Cape St. Francis (Tomlin, coll. Burnup). Remarks. ‘The Burnup shell recorded by Tomlin was 5} in. (133 mm.) long. The Pieter Faure obtained only one shell off the Natal coast (if the label was correct); all the others were from off the Cape Point. The Africana has shown that the distribution extends to 31° S. on the west coast. The large Natal 76 ANNALS OF THE SOUTH AFRICAN MUSEUM specimen, now in the British Museum, should be re-examined on account of its alleged locality. More probable misidentifications are Odhner’s 150 mm. and Burnup’s 133 mm. shells. I consider that both of these are more likely to be the off-shore form of lugubris, which has stronger and more irregular spiral liration than rutila; the latter also has a more sigmoid canal. Fasciolaria lugubris Rve. Figs. 16(0, ¢), 19(@, 0) 1847. Reeve. Conch. Icon., sp. 2. 1848. Krauss. Stidafr. Moll., p. 110, pl. 6, fig. 12 (badia). Toit. Strebel: loc: cit.,.p./ 9, pl) 6, tias.30,, 304, 31- 1932. Tomlin. Ann. S. Afr. Mus., xxx, p. 157, fig. 1 (agulhasensis). 1932. Haughton. Tr. Geol. Soc. S. Afr. (1931), pp. 34, 49- Aperture longer than spire. Profile of whorls evenly convex, but sometimes with faint indication of a blunt shoulder, rarely a definite shoulder with low knobs. Protoconch (from egg-capsule) 14 whorls, diam. 3—3°5, alt. (i.e. apex to rostrum) 5 mm., smooth, later part of whorl with 6-8 feeble axial ribs, crossed by spiral lirae, c. 20 from suture to lower end of outer lip. Postnatal whorls 7; 10 axial ribs on 1st whorl, on 2nd and 3rd and sometimes 4th a similar number of slight undulations or low rounded knobs, obsolete on later whorls; crossed by numerous spiral lirae, varying in size, usually coarser and finer alternating, the larger ones sometimes with one or more striae, the peri- pheral lira sometimes distinctly stronger than its neighbours, almost a costa (agulhasensis), sometimes a pair of peripheral lirae stronger than the others. Parietal callus present, but weak in young and half-grown shells; columella pleats distinct. Outer lip evenly convex, not so distinctly incurved at beginning of canal as in heynemanni, plicate internally, edge at some stages of growth denticulate. Periostracum fibrous-fimbriate. 183 (tip of canal broken, about 5 mm. missing) * 75 mm. (type of agulhasensis); two False Bay examples (tip of canal broken in both) 145 « 52 mm. and 142 x 55 mm. Largest littoral example 81 (tip of canal broken) x 43 mm. Operculum 28 X 14 mm. in 81 mm. shell. Protoconch and first 2 (24) whorls white, rest fulvous or chestnut-brown, operculum dark brown. Animal sealing-wax red with white dots on sides of foot (ie:EL. Ba) Radula: from protoconch in egg-capsule, 40 rows, lateral plate with 6 cusps; from 18 mm. shell 135 rows, lateral with 8 cusps; from 33 mm. shell 190 rows and 11—12 cusps resp.; from 85 mm. shell 225 rows and 12 cusps resp. ; from 142 mm. shell 265 rows and 14-16 cusps resp. (in all cases excluding the tiny internal denticle); central plate quadrangular, with well-developed cusps, the middle one stronger than the side cusps. Egeg-capsule club-shaped, 20 mm. high, distally elliptical in section, major diam. 9 mm., minor diam. 8 mm., horny; only one specimen seen, containing _ CONTRIBUTIONS TO KNOWLEDGE OF S.A. MARINE MOLLUSCA 77 4 embryos, but the apex was open and some of the embryos had probably escaped. Fossil, late Tertiary; Saldanha Bay (Haughton). Dead. Natal (Krauss); False Bay (Strebel, coll. Fritsch); Cape Hangklip (S. Afr. Mus.); between Cape St. Blaize and Flesh Point, 28 fathoms (S. Afr. Mus. P.F. coll. type of agulhasensis). Living. Steenberg Cove (St. Helena Bay) and Langebaan (Saldanha Bay), littoral (U.C.T.); Oudekraal, west coast of Gape Peninsula, littoral (U.C.T.); Sea Point, Cape Town, littoral (S. Afr. Mus.); Kalk Bay (False Bay), littoral (S. Afr. Mus.), and 4-5 metres (U.C.T.). Remarks. The striae on the larger spiral lirae are not (in the specimens I have seen) so prominent as Strebel described in his specimens, and cannot be regarded as a specific character. I have seen a 51 mm. shell from Cape Hangklip (S. Afr. Mus. no. 14052), and a 95 mm. shell from False Bay (U.C.T.) with distinct angular shoulders and two peripheral lirae. Another False Bay example 142 mm. long (U.C.T.) has g peripheral swellings on the 6th whorl, but none on the preceding whorls. The resemblance of these specimens to heynemanni is strong, but the latter species has finer and more regular lirae above the shoulder, and the shoulder pro- jections are definitely knobs, not mere swellings or undulations; and the 2 peripheral lirae are situated in the middle of the whorl, not as in the dunkert form of heynemanni adjoining the suture of the following whorl. The contrast between the smooth slender examples and, e.g. the 95 X 46 mm. shouldered specimen is very striking, and is paralleled with the adamsu and ocelliferus forms of Fusus verruculatus (p. 90). Littoral examples are smaller and stouter than off-shore forms (cf. heynemannt). The 145 X 52 mm. shell from False Bay, 23 fathoms, was identified by Tomlin as filamentosa in spite of its having a parietal callus. It bridges the gap in size between the common littoral form and agulhasensis. The latter was contrasted by Tomlin with scholvient, an obviously different species, but he did not mention any differences between his n. sp. and lugubris. In fact there are no differences. Krauss said his badia was very like Fusus mandarinus, and also bore resem- blance to Fasciolaria fusiformis and filamentosa.* ‘There are no later records of lugubris from Natal, or indeed from east of Cape St. Blaize. Fasciolaria trapezium Linn. 1880. Von Martens. Mauritius G Seychellen, p. 245. 1895. Cooke. Cambr. Nat. Hist., ii, fig. 121 (radula). 1903. Smith. Proc. Mal. Soc., v, p. 368 (heynemanni, non Dnkr.). * See also Fusivoluta pyrrhostoma (Volutidae) for the Gazelle shells reported to be Fusus mandarinus (p. 30). 78 ANNALS OF THE SOUTH AFRICAN MUSEUM 1911. Strebel. loc. cit., pp. 40 sqq. and vars. pls. 7-10, 13, 14, figs. 38-45, 48, 49, 61, 62. 1930. Fulton. Ann. Mag. Nat. Hist. (10), vi, p. 685, pl. 18, figs, 2, 2a (strebelz). 1952. Satyamurti. Bull. Madras Govt. Mus., I, 2, pt. 6, p. 186, pl. 17, fig. 10. 1952. Braga. Anais Est. Zool. Invest. Ultramar, vii, 3, p. 73, pl. 2, fig. 1. Aperture longer than spire. Profile of whorls with angular shoulder. Protoconch 14 whorls, diam. 1-5 mm., smooth (Strebel, figs. 38, 42). Postnatal whorls 9 (? 10); spiral grooves only on early whorls but indicated on later whorls by dark lines; on 1st and and (—3rd) whorls 8-9 axial ribs extending from suture to suture, but on later whorls gradually restricted to the periphery and forming knobs on the shoulder, which becomes definitely marked from 4th whorl onwards: 5-7 in forma typica, 8-11 in varieties. Parietal callus present, but weak in juveniles. Columella with 3 pleats, and often additional wart-like nodules anterior to the main pleat; canal straight or slightly sigmoid; outer lip internally with numerous plicae, usually a smooth zone between the plicae and the edge; the latter in some stages of growth denticulate. Up to 215 mm. long (f. typica); var. ponderosa 230 mm. (Strebel). White or pinkish, with narrow brown spiral lines usually in pairs, plicae in aperture orange-brown, periostracum yellowish-brown. Radula (only a portion removed from a 153 mm. shell); lateral plate with 30 cusps. Cooke’s figure shows 22 cusps. Dead. Durban (Sowerby, Smith); Natal (Fulton, and S. Afr. Mus.); Inhambane and Mozambique Island (Braga). Living. 28° 28’ S. 32° 25’ E. (off Cape St. Lucia), 27 metres (s.s. Africana). Distribution. Red Sea, Zanzibar, Querimba Is., Mauritius, Réunion, Seychelles, Madagascar, Indo-Pacific. Remarks. The relatively small protoconch distinguishes this species from heynemanni, apart from other characters. Smith’s mention of the small proto- conch indicates that his two specimens were really young trapezium. A Natal specimen, 53 x 24 mm. (S. Afr. Mus.) might certainly be regarded as heynemanni because it is unicolorous without any trace of the dark spiral lines so characteristic of trapezium; but it has the small protoconch (and narrower ist and 2nd whorls) of trapezium. The living specimen, 153 mm. long, dredged by the s.s. Africana has indications of double lines only on the latest part of the outer lip, and only feeble colourless plicae within; the base is quite smooth, without costae or lirae; the protoconch is broken off, but was probably small. That the characteristic dark spiral lines are not always developed in juveniles and half-grown trapezium is shown by Smith’s (size not given), Fulton’s 69 mm., and the 8. Afr. Mus. 53 mm. examples. In addition to the one Natal specimen, 5. Afr. Mus. also has examples from extra-African localities. Even the 153 mm. Africana specimen was only just beginning to develop these dark lines. Fulton’s reasons for separating his strebeli from heynemanni were its more slender shape and absence of well-marked spiral lirae; Smith also mentioned CONTRIBUTIONS TO KNOWLEDGE OF S.A. MARINE MOLLUSCA 79 these characters. But neither Smith nor Fulton mentioned the size of the proto- conch, which, according to Fulton’s photographic figure, was small; nor did they compare their specimens with trapezium. A further difference between trapezium and heynemanni is: the former develops six whorls against the latter’s four, in half-grown specimens of approximately equal size. Fasciolaria heynemanni Dnkr. Fig. 18(d), 19(d) 1876. Kobelt in Kiister. Conch. Cab., p. 139, pl. 28, fig. 5. 1903. Von Martens. D. Tiefsee Exp., vii, p. 30. 1g11. Strebel. loc. cit., p. 28, pl. 5, figs. 27, 28; and p. 31, pl. 6, fig. 29 (scholvienc) ; and p. 33. pl. 6, figs. 32, 32 a, b (dunkeri). 1915. Bartsch. Bull. U.S. Nat. Mus., 91, p. 46, pl. 4, figs. 3, 3a, 3b (alfredensis). 1923. Odhner. Géteb. K. Vet. Handl., xxvi, p. 6 (alfredensis). 1932. Turton. Mar. Sh. Pt. Alfred, p. 49; and p. 49, pl. x1, no. 362 (juv. dunkeri, non Strebel) ; and p. 49 (scholviert, typ. err.); and p. 49 (alfredensis). Not: Smith. Proc. Mal. Soc., v, 368 (= trapezium). Spire 14-2 times in aperture. Profile of whorls with angular shoulder. Protoconch 14 whorls, large, diam. 3°5—4°5, alt. 3-4 mm., smooth, last half whorl with about 6 axial ribs, from suture to suture, continued on Ist postnatal whorl, but restricted to the periphery and forming knobs, 5-6 spiral lirae above the knobs, increasing to 8 or more on later whorls. Postnatal whorls 6, with peripheral knobs varying in size, and in number from 9-14. Spiral lirae over whole whorl, including the knobs; on base 8-10 more or less conspicuous very flat costae, each traversed by about 3—5 lirae; usually only one, but sometimes 2 (scholuient) of these costae appear between the knobs and the suture of the following whorl. On the rostrum the spiral lirae become almost axial in direction (parallel with the canal), distinct in examples with straight rostrum but obscured by the growth-lines in those with costate rostrum. Parietal callus present. Columella with 3 pleats. Canal long, narrow, straight or slightly sigmoid, especially when left side of rostrum is thickened and costate. Outer lip slightly flattened in the middle and incurved below, not plicate internally, in some examples with denticulate edge. 203 (protoconch and tip of canal broken) xX 80 mm.; slenderest specimen 160 X 55 mm. Operculum 62 X 33 in 203 mm. shell. Pale buff or white, unicolorous, periostracum usually chestnut-brown, but sometimes olivaceous golden-brown (Cape St. Blaize example) or golden- brown (False Bay example); protoconch white; operculum dark brown. Radula with c. 230-280 rows, lateral plate with 15 cusps on one side, 16 on the other, in another specimen 17 and 18 (excluding the tiny internal denticle) ; middle cusp of the central plate asymmetrically bifid in one specimen. Dead. Natal (Kobelt); Port Elizabeth, Port Alfred (Sowerby, Strebel, Bartsch, Turton, $8. Afr. Mus.); Elim (i.e. Bredasdorp coast, Gape Agulhas area) (Strebel: dunkeri); ‘Cape’ (Strebel: scholuvieni); off Glendower Beacon . 80 ANNALS OF THE SOUTH AFRICAN MUSEUM (Port Alfred area), 66 fathoms (S. Afr. Mus. P.F. coll.). Delagoa Bay S. Afr. Mus. coll. K.H.B. 1912). Living. Between Plettenberg Bay and St. Francis Bay, 100 metres (von Martens); off Cape Infanta, 34-40 fathoms (Odhner) (also dead examples) ; off Kowie, 40-43 fathoms; Algoa Bay, 40 fathoms; off Cape St. Blaize, 46 fathoms; False Bay (S. Afr. Mus. P.F. coll.). Cape Agulhas, littoral CURE FE. Remarks. A variable species, as was recognized by Strebel. He figured the ‘deep-water’ and the ‘coastal’ forms. The former has a straight non-costate — rostrum, the latter a costate rostrum. But the present series shows that this difference is not due to habitat, because both were obtained together in one haul off the Kowie (P.F.); nor does it seem to be sexual because out of 5 animals 2 gg and 1 @ are non-costate, 2 99 costate. This costate thickening of the rostrum seems to appear only in specimens approximately 100 mm. in length upwards. At the Kowie locality plump and slender examples were taken in the same haul: width approx. 2-2—2-8 in the length, spire varying from 14, 132, to 2 times in aperture (incl. canal), knobs g (in smallest specimen 58 mm. long) to 14 (the 2 largest examples 203 and 187 mm. had 11 and 13 knobs resp.). The peripheral knobs are usually in the middle of the whorl, but may be lower (more anterior); in dunkeri they adjoin the suture, except on the last preserved whorl. In some juveniles 2 subequal series of knobs are developed (cf. dunker?). It is difficult to decide where the protoconch ends and the ist postnatal whorl begins. The apex of the subglobular protoconch (14 whorls) is white, the 1st postnatal whorl brown, but there is no sharp division between the two. The axial ribs begin on the last part of the white region and are continued, without any obvious interruption of growth, on the brown Ist whorl. One specimen of protoconch is very large: diam. and alt. both 4:5 mm. The curve of the outer lip helps to distinguish this species from lugubris, but not from trapezium. Strebel suggested that scholuient might possibly be a large form of the dwarf heynemanni; but the present material negatives this; the differences are merely individual. Bartsch’s alfredensis was based on a worn slender specimen. Although the plump, short-spired forms with strong knobs look different from the slender, high-spired forms with weaker knobs, the present S. Afr. Mus. material, meagre as it is, exhibits transitional forms. F. dunkert, based on one young specimen (41:5 X 19:3 mm.), I regard as an aberration. In some young examples of typical heynemanni the knobs on the early whorls are very close to the suture of the following whorl; and in some very young specimens a double row of feeble knobs occurs; but I have seen no specimen exactly corresponding with dunkeri, i.e. with a double peripheral shoulder adjoining the suture. ~ 2) 2 i 5 ar = CONTRIBUTIONS TO KNOWLEDGE OF S.A. MARINE MOLLUSCA 81 In spite of Kobelt’s record from Natal, the Delagoa Bay locality is remark- able. I cannot doubt the provenance because I myself found the specimen. One would expect to find trapezium there, but the specimen is a typical heyne- mann without a single one of the features distinguishing the Indo-Pacific species. It retains the periostracum, and is thus unlikely to have been carried very far by currents; moreover the Mozambique current sets in the wrong direction to have carried this shell from Natal or the Port Alfred area. Gen. Latrirus Mont. 1810. Montford. Conch. Syst., u, p. 530. 1840. Swainson. Treat. Malac., xxviii, p. 304 (Plicatella). 1891. Melvill. Mem. Manch. Philos. Soc., n. ser. 4, vol. 4, pp. 365-411 (Latirus + Peristernia). 1911. id. 7. Conch., xiii, p. 164 (Latirus + Peristernia). Shell ovate or fusiform, usually ribbed. Canal long or short, columella usually with pleats, aperture sometimes internally plicate. Operculum ovate, apex curved inwards, nucleus apical, internal surface as in Fasciolaria (? all species). Radula central plate tricuspid, lateral plate with several cusps without intervening smaller denticles. The South African species fall into two groups: Whorls with peripheral knobs— abnormis, subcontractus. Whorls with axial ribs and spiral irae— polygonus, clausicaudatus, bairstowr, rousi, alboapicatus, turritus. No living examples of bazrstowi Sow. 1886 or roust Sow. 1886 have yet been found. The only South African record of turritus (Gmelin) is a dead shell from Durban (Sowerby 1897); the species occurs at Mauritius, Seychelles, Ceylon, etc. Latirus abnormis Sow. 1894. Sowerby. 7. Conch., vii, p. 6. 1897. id. Append. Mar. Sh. S. Afr., p. 8, pl. 6, fig. 7. 1902. id. Mar. Invest. S. Afr., ii, p. 96, pl. 2, fig. 1 (émbricatus). Fgoz. id. -ibid., p. 2277. tgit. Melvill. 7. Conch., xiii, p. 165. Aperture (incl. canal) a little longer than spire. Protoconch 2 whorls, diam. 1-7, alt. 1-3-1-5 mm., smooth, junction with Ist postnatal whorl distinct. Postnatal whorls 7, profile concave above the periphery, in some specimens rather strongly so, almost forming a subsutural groove. Axial ribs 10 on Ist whorl, well marked on periphery but feeble above and below, decreasing to 7-8 peripheral knobs on later whorls; crossed by 4—5 spiral lirae on 1st whorl, increasing by interpolation on later whorls, but evanescent on 6th and 7th whorls; 2 of the peripheral lirae on whorls 2—5 stronger than the others, making the knobs subcarinate, but obsolete on knobs on 6th and 7th whorls; spiral “' vt @ % 82 ANNALS OF THE SOUTH AFRICAN MUSEUM lirae continued on base, 2 of them in upper third (at posterior end of aperture) stronger than the others and forming a series of feeble subcarinate knobs cor- responding in position with the peripheral knobs. Parietal callus weak; columella without pleats. Canal not abruptly separated from rest of aperture, which is posteriorly more or less indented; columella glaze in large specimens discrete from rostrum forming a narrow umbilicus; outer lip not plicate within. Periostracum thin, fibrous, imbricate-scaly. 72 X 29 mm. (S. Afr. Mus.); Brit. Mus. specimen (when perfect) probably 75 mm. long (Smith). Operculum ovate, gently curved, 13 X 6:5 mm. in 49 mm. shell, internal surface as in Fasciolaria. Ochraceous salmon, or orange-brown, periostracum amber-brown, operculum brown. Living. Off Durnford Point (Zululand), 13 fathoms (S. Afr. Mus. P.F. coll.). Dead. Natal (probably from fish stomachs) (Smith); off Tugela River, 46 fathoms (Sowerby), and 14 fathoms (S. Afr. Mus. P.F. coll.). Latirus subcontractus (Sow.) 1902. Sowerby. Mar. Invest. S. Afr., li, p. 97, pl. 2, fig. 2 (Fusus s.). 1932. Tomlin. Ann. S. Ajy. Mus., xxx, p. 158, fig. 2 (mosselensis). Aperture a trifle longer than spire. Protoconch 2 whorls, diam. 1:3, alt. 1 mm. smooth, beginning of Ist postnatal whorl distinct (type of mosselensis). Postnatal whorls 8; 9 (? 10) (subcontractus) or 8 (mosselensis) axial ribs on each whorl, from suture to suture on 1st-3rd whorls, but from 4th whorl restricted to the periphery; crossed on 1st whorl by 5 spiral lirae, increasing on following whorls, but from about the 5th whorl evanescent, except the peripheral lira which persists on the knobs (not in the intervals) making these subcarinate; on base a second series of weaker subcarinate knobs, and a few weak lirae on rostrum. Parietal callus weak, columella with 2 obscure pleats (mosselensis) ; canal rather abruptly demarcated from rest of aperture, narrow (subcontractus) . A narrow umbilicus (mosselensis) ; outer lip not plicate internally. Periostracum thin, smooth. 40 * 18 mm. (subcontractus); 53°5 X 23 mm. and 60 X 25 mm. (type and figure of paratype of mosselensis). Operculum ovate, gently curved (Tomlin). Pale pinkish, periostracum yellowish-brown (mosselensis type). Dead: off Cape Natal (Durban), 200 fathoms (Sowerby). Living: Mossel Bay, 27 fathoms (Tomlin) (both S. Afr. Mus. P.F. coll.). Remarks. Has a strong resemblance to a Fasciolaria, cf. heynemanni or trapezium. ; Type of subcontractus ? in British Museum. Tomlin figured the paratype (? in coll. Tomlin) with operculum, and the back view of the type of mosselensts. The latter is in S. Afr. Mus. without operculum. The radula was not described. CONTRIBUTIONS TO KNOWLEDGE OF S.A. MARINE MOLLUSCA 83 Tomlin compared his species with L. armatus A. Adams 1854 (see 1886. Challenger Rep., xv, pl. 13, fig. 1, Fasciolaria armata), but made no reference to _ the truly remarkable resemblance to Sowerby’s species. Except that subcontractus appears (from Sowerby’s figure) to have one (? 2) more axial ribs (knobs) than mosselensis, and that the latter is rmmate and has a slightly curved columella, there are no differences between Sowerby’s figure and the type of mosselensis. The lower part of the canal and outer lip are broken in the latter, which may account for the umbilicus being visible. In spite of the distance apart of the two localities, and the difference in depth, one cannot accept more than the one species. Latirus polygonus (Linn.) 1816. Lamarck. Tabl. Encyclo., pl. 423, fig. 1 (Fusus p.). 1859. Chenu. Man. Conchyl., i, fig. 908. 1903. Smith. Proc. Mal. Soc., v, p. 369 (var.). The only South African record is from ‘Durban, deep water’ [probably from fish stomach] (Smith). The species occurs at Mauritius, and other localities in the Indian Ocean. Specimens in S. Afr. Mus. from ee (coll. Rawson W. Rawson). The coloration appears to be distinctive: buff or ochraceous, with brown axial ribs divided into oblong patches by the pale spiral lirae. Latirus clausicaudatus (Hinds) 1844. Hinds. ool. Voy. Sulphur. Moll., p. 13, pl. 1, figs. 10, 11 (Fusus c.). 1892. Sowerby. Mar. Sh. S. Afr., p. 3 (Fusus c.). Elongate-fusiform, turreted, aperture (incl. canal) 14—1} times spire. Profile of whorls evenly convex. Protoconch 24 whorls, diam. and alt. 2 mm., smooth, junction with rst postnatal whorl! distinct. Postnatal whorls 7; 7 axial ribs on Ist whorl, increasing to 8 or g on last whorl, from suture to suture but becoming weaker above periphery on later whorls, and evanescent on lower two-thirds of base; sometimes indistinct on last part of 7th whorl, but reappearing as a well-marked rib on outer lip; crossed by spiral rae, 7-8 on Ist whorl increasing to 16-20 on later whorls, 35-40 additional on base. Parietal callus bluntly dentiform; columella with 2 pleats, the anterior one distinct, the other obscure; canal abruptly marked off from rest of aperture, very long, nearly twice the rest of aperture, nearly closed throughout its length by the discrete edge of the columella glaze, no umbilicus. Aperture internally without plicae except one at base of canal opposite the columella pleat. Perio- stracum thin, smooth. 51 X 15°5 mm., smallest specimen in 8S. Afr. Mus. 31 mm. long. Hinds’s figure 58 xX 16-5 mm. Operculum oval, apex incurved, 7 X 3°5 mm. in 45 mm. shell. 84 ANNALS OF THE SOUTH AFRICAN MUSEUM White or pale buff, periostracum pale greyish brown, operculum dark brown. : Dead: Agulhas Bank, 50-60 fathoms (Hinds). Off Cape Natal (Durban), 54 fathoms; off Cape Morgan, 77 fathoms; off Nahoon Point (East London), 45 fathoms; off Hood Point (East London area), 49 fathoms; off Nanquas Peak (eastern part of Algoa Bay), 63 fathoms; Algoa Bay, 37 fathoms; (S. Afr. Mus. P.F. coll.). Living: off Riet Point (east of Algoa Bay), 23 fathoms (S. Afr. Mus. PE, jcoll:): Remarks. These specimens are clearly referable to Hinds’s species, which Sowerby (1892, also 1903. Mar. Invest. S. Afr., 1, p.. 97) said was represented. by the unique type in the British Museum. In one specimen the protoconch and 3 whorls are slightly curved to the right, and in another the protoconch is curved to the left as in Hinds’s figure. The animal of the only specimen taken alive was not preserved. Latirus alboapicatus Smith 1902. Smith. 7. Conch., x, p. 250, pl. 4, fig. 5. 1906. Smith. Ann. Natal Mus., i, p. 34, pl. 7, fig. 7 (burnupt). 1911. Melvill. 7. Conch., xii, p. 166 and p. 168 (burnupt). 1931. Tomlin. Ann. Natai Mus., vi, p. 433. Fusiform. Profile of whorls slightly angularly convex (fig. of alboapicatus), slightly concave above, then gently convex (burnupi). Protoconch 2 whorls, diam. and alt. ¢. 1-5 mm., smooth. Postnatal whorls 6; axial ribs c. 12 (albo- apicatus) on 1st whorl (in burnufi corroded), 8 on 2nd and following whorls, from suture to suture, but more prominent on the periphery, and evanescent on base; crossed by spiral lirae (number not stated in alboapicatus) 5 on and whorl increasing to 8-10 on last whorl, the first one or first 2 or 3 below suture granu- lose, 9-11 additional on base, of which the 4th or 5th is stronger than the others, covered posteriorly by the parietal callus and forming a small denticle on outer lip; columella with 3 weak pleats, canal straight (slightly recurved in alboapicatus), distinctly but not abruptly marked off from rest of aperture, umbilicus slight or absent, outer lip sometimes plicate internally. Periostracum thin, smooth. 28 X 11°5-12 mm. Operculum (burnupi) ovate, apex incurved, 6 X 3 mm. in 28 mm. shell. Rufous with white apex, and a pale band below centre of body whorl, aperture rufescent within (alboapicatus); white with brown periostracum, the strong lira on base showing as a pale line, aperture rosy or purplish within, operculum dark brown (burnup) (see Remarks). Dead: (alboaficatus) Durban (Tomlin). Living: (burnupi) Port Shepstone (Natal) (Smith, also S. Afr. Mus. Ross-Frames coll. ex Burnup). Remarks. Although Smith must have had his alboapicatus for comparison when he described burnupi, and Melvill accepted both species, I strongly suspect CONTRIBUTIONS TO KNOWLEDGE OF S.A. MARINE MOLLUSCA 85 that the two are conspecific. As I have seen no specimens of the former species, the description is based mainly on specimens of burnupi collected in the type locality by Burnup. Like Smith’s specimens, all these are more or less corroded at the apex, so that the true size of the protoconch or the sculpture of the 1st whorl cannot be determined. Tomlin (1931) records (fide Burnup) a specimen of alboapicatus 44 (45) X 16-5 mm. (!), of which Burnup said the ribs and growth-lines were paler than the intervening spaces, and the spiral grooves much darker than the lirae. This applies also to Burnup’s specimens of burnupi in S. Afr. Mus. Gen. PERISTERNIA Morch 1852. Morch. Cat. Conch. Yoldi., i, p. 99. 1891. Melvill. Mem. Manch. Philos. Soc. (4), iv, p. 365 (Latirus part). 1911. id. 7. Conch., xiii, p. 164 (Latirus part). 1935. Yen. Notes Malac. Chinoise, i (2), p. 41 (Peristerina emend.). Shell broadly fusiform, more or less distinctly axially ribbed and spirally lirate, the lirae often scabrous or squamose; not or only slightly umbilicate, canal moderate, slightly recurved, columella usually with 2 pleats, aperture internally plicate. Operculum as in Latzrus. Radula: central plate piriform, narrowed in front, with 3 feeble cusps, and lateral plate with denticles between the main cusps. Included conchologically in Latirus, by Melvill, but can be distinguished by the radula. Peristernia leucothea Melv. Fig. 19(e) 1891. Melvill. loc. cit., p. 399, pl. 2, fig. 15. 1900. Sowerby. Proc. Mal. Soc., iv, p. 1, pl. 1, fig. 2 (Euthria eburnea). ? 1932. Turton. Mar. Sh. Pt. Alfred, p. 54, pl. 12, no. 400 (Euthria ordinaria). 1937. Peile. 7. Conch., xx, p. 300, fig. 1 (radula). Length about twice breadth. Aperture subequal to spire. Protoconch 14 whorls, diam. 0:8, alt. 0-75 mm., smooth, 2 riblets close together before junction with Ist postnatal whorl. Postnatal whorls 7; axial ribs g-10 on 1st whorl, increasing to 11-12 on last whorl, from suture to suture but more prominent peripherally, evanescent on base; crossed by spiral lirae, 3 on Ist and 2nd whorls, 4 on 3rd, 5-6 on 4th, increasing to 10-11 on last whorl, often with intermediaries, 10-12 additional on base, also with intermediaries, usually one or two at about the middle of base stronger than the others; fine close-set growth-lines producing punctae in the sulci between the lirae. Parietal callus present, columella with 3 pleats but usually only 2 or one distinct, forming a short keel at beginning of the short canal. Outer lip internally plicate, the 1st 86 ANNALS OF THE SOUTH AFRICAN MUSEUM (posterior) and last (opposite columella pleat) plicae larger than the others, more or less dentiform. Columella glaze sometimes rimate anteriorly forming a feeble umbilicus. Periostracum very thin. 25 X 12 mm. Operculum ovate, apex incurved, 4°5 < 2°5 mm. in 22 mm. shell, internal surface as in Fasciolaria. Creamy-white or pale buff, unicolorous; or orange-brown with markings, the colour when present is mostly around the suture, between the ribs, and in one or two bands on base, aperture internally white or pale brown, or pale violaceous, operculum chestnut-brown, periostracum yellowish-brown. Radula with 250-270 rows, lateral plates with g-11 (12) cusps and denticles, not always symmetrical and the sequence of cusps and denticles varying from one part of the radula to another. Dead: Port Natal (Durban) (Melvill), Isipingo and Umkomaas (Natal) (Smith), Tongaat (Natal) (S. Afr. Mus.); Pondoland (Sowerby: eburnea) ; Port Alfred (Turton: ordinaria). Living: off Durnford Point (Zululand) 13 fathoms (S. Afr. Mus. P.F. coll.); Scottburgh (Natal) littoral (S. Afr. Mus. coll. K.H.B.); Umpangazi, Umbhlali, Durban, Umtwalumi, Port Edward and Port St. Johns (U.C.T.). Remarks. The apex is usually corroded in littoral specimens; only one of the numerous Scottburgh specimens had a complete unworn protoconch. Appears to be an ‘albino’ form of nassatula, as there is no conchological difference between the two. Of the specimens I have seen, those most strongly marked with orange-brown come from Umblati, Tongaat, and Durban. Others from the last locality are uniform white or buff; one of the Umpangazi " shells has a pale violaceous aperture. For the present retained separate from nassatula. One specimen with an aberrant operculum: oval, nucleus intramarginal in apical third of length. A series of water-worn specimens (S. Afr. Mus. locality?) is interesting. The effect of wear is to broaden the ribs and reduce the intervening grooves in — which the spiral lirae, or their intervening sulci, persist, though much reduced. Even when completely worn away at the upper part of a whorl, they usually persist in the lower part adjoining the suture of the following whorl. The final result—a perfectly smooth shell—appears to be represented by Turton’s ordinaria from Port Alfred, the locality farthest removed from the habitat of the living animal. Even badly worn specimens, however, are too broad to be mistaken for Suscotincta. Fic. 19. ST Central and lateral radula plates of (a), (b) Fasciolaria lugubris Rve. from juvenile from egg- capsule, and from 85 mm. shell; (c) F. rutila Watson; (d) F. heynemanni Dnkr.; (e) Peristernia leucothea Melv.; (f) P. fuscotincta (Sow.); (g) Fusus verruculatus Lam.; (h) F. faurei n. sp.; (1) F. rubrolineatus Sow., two variants of lateral plates; (j) F. colus Linn., half-grown and adult, the latter with abnormal 4-cuspid central plate; (k) F. africanae n. sp. 88 ANNALS OF THE SOUTH AFRICAN MUSEUM P. incarnata Desh., recorded from Natal (Sowerby 1892) (occurs also at Mauritius, Red Sea, and Indo-Pacific) differs from Jleucothea in having fewer spiral lirae on last whorl (6-7 in Philippine specimens in S. Afr. Mus.), and in coloration, which is yellow or orange with brown intervals between the ribs. Peristerna nassatula (Lam.) 1859. Chenu. Man. Conchyl., i, fig. gto. 1880. Von Martens. Mauritius G Seychellen, p. 246 (Plicatella n.). The description given for leucothea will apply to this species which is, however, more brightly coloured. Cream, upper half of whorls and the grooves between the ribs brown, shading off into orange-brown, base with a pale spiral band at level of top of aperture, followed by a dark brown band and then another pale band, rostrum orange-brown, aperture violaceous. Radula (one specimen from Delagoa Bay examined) incomplete but with at least 180 rows, lateral plate with (8) 9-10 cusps and denticles, varying in size and sequence as in leucothea. Natal (Krauss, ? dead). Delagoa Bay, living (U.W.). Distribution. Mauritius, Réunion, Seychelles, East Indies. Peristernia fuscotincta (Sow.) Fig. 19(f) 1886. Sowerby. 7. Conch., v, p. 2 (Euthria f.). 1689, id. “abid, wi, pli, fies 18. 1892. id. Mar. Sh. S. Afr., p. 4, pl. 1, fig. 13 (Huthria f.). 1938. Peile. Proc. Mal. Soc., xxiii, p. 99, fig. 36 (radula). 1947. Stephenson. Ann. Natal Mus., xi, pp. 271, 273 (Cominella f.). Length distinctly more than twice breadth. Aperture shorter than spire. Protoconch 2 whorls (all specimens worn). Postnatal whorls 5; faint indications of weak ribs on upper 2 or 3 whorls; all whorls with spiral grooves, 4 on 2nd whorl increasing to 10-11 on last whorl, usually in pairs, and punctate where crossed by the growth-lines, g-10 additional on base, some in pairs. Parietal callus present, columella with 3 pleats, but only one or two distinct, forming a short keel at beginning of the short canal. Outer lip internally plicate, 1st and last plicae larger than the others. Sometimes a feeble umbilicus. 20 x 8 mm. Operculum oval, apex incurved. White with irregular brown markings, either as patches or axial flames, usually a more or less continuous brown band. below periphery. Radula with 140-160 (Peile: 167) rows, lateral plate usually with 6 major cusps with intervening denticles (1 or 2); the arrangement varies in successive rows, and is not always symmetrical on the two sides (cf. leucothea). CONTRIBUTIONS TO KNOWLEDGE OF S.A. MARINE MOLLUSCA 89 P.. fuscotincta: dead; Port Elizabeth, Port Alfred (Sowerby, Bartsch, Turton); Port Shepstone (Natal) (S. Afr. Mus. coll. Burnup). Living: Port St. Johns to Richmond (Alexandria Division) (Stephenson) ; East London to Kleinmond (Bathurst Division) (U.C.T.). Remarks. In fuscotincta the spiral grooves were described as very obscure (they certainly are in beach-worn specimens!). Some of the beach-worn specimens in S. Afr. Mus. are almost wholly white, but even so the subperipheral band and a few brown marks are just visible. Stephenson (1947 p. 271 footnote) transferred this species to Cominella (Afrocominella) seemingly at Tomlin’s suggestion—a clear case where con- chological guessing proved wrong —although Peile’s examination of the radula had already (1938) put the species in its correct genus. Gen. Fusus Brug. 1789. Bruguiére. LEncycl. Meth. (1), xv. Shell more or less elongate-fusiform spire often high. Canal moderately or very long; columella without pleats but sometimes rimate. Operculum oval or piriform apex blunt more or less incurved or sharply pointed, nucleus apical, internal surface with marginal thickening as in Fasciolaria. Radula, central plate subtriangular, narrowed in front, tricuspid, lateral plate with several cusps, without intermediate denticles. Remarks. F. radialis Watson has proved, not unexpectedly, to be a Columbarium (p. 234), and F. speratus a Tritonalia (p. 215). Fusus africanus (Sow.) 1897. Sowerby. Append. Mar. Sh. S. Afr., p. 1, pl. 6, fig. 19. 1903. Smith. Proc. Mal. Soc., v, p. 368, pl. 15, fig. 19. 1932. Turton. Mar. Sh. Pt. Alfred, p. 50, pl. xi, no. 370 (kowzensis). Piriform. Aperture (excl. canal) a little longer than spire (incl. canal: 24-24 times spire). Profile of early whorls nearly straight, of later whorls angular. Protoconch 2 whorls, smooth (but no perfect specimen seen). Post- natal whorls 6; on first 3 whorls 10-11 very obscure axial ribs, reduced on following whorls to peripheral rounded knobs, which become stronger on last 2 whorls: on 4th whorl near suture, on 5th and 6th nearer middle of whorl; crossed by spiral lirae, 5 on 1st whorl, increasing to 8-g (10) on 4th whorl but thereafter evanescent; on base 12-15 low blunt costae, evanescent towards end of rostrum. A blunt parietal callus, columella curved, glaze discrete forming an umbilical rimation; canal long, straight, narrow, distinctly marked off from go ANNALS OF THE SOUTH AFRICAN MUSEUM rest of aperture; outer lip not plicate within. Periostracum thin. 104 X 52mm. (Turton); 79 (protoconch missing) xX 38 mm. (S. Afr. Mus.). Operculum and radula unknown. Creamy, buff, pale orange-brown, with darker marks between the knobs and sometimes axial flames, grooves between costae on base orange-brown, periostracum brown. Port Elizabeth (Sowerby); Port Alfred (Turton: kowiensis); off Durban [from fish stomachs] (Smith, also S. Afr. Mus.). Remarks. Only dead specimens known. Except for Sowerby’s original young specimens, and Turton’s specimen, this species has only been obtained from fish stomachs in Natal waters. Not taken by the Pieter Faure. Fusus verruculatus Lam. Figs. 19(g), 20(a) 1816. Lamarck. Tabl. Encycl. Meth., p. 429, fig. 7, and Liste, p. 7 (ocelliferus, name and figure only). 1822. Id. Anim. sans Vert., vii, p. 129. 1870. H. Adams. Proc. Zool. Soc. Lond., p. 110, text-fig. (ventricosus, non Gray). 1876. Kobelt. Conch. Cab., p. 152, pl. 47, fig. 3 (adamsit). 1886. Watson. Challenger Rep., xv, p. 195 (references). 1892. Sowerby. Mar. Sh. S. Afr., p. 3 (robustior). 1925. Thiele. D. Tiefsee Exp., xvii, p. 184, pl. 32 (20), fig. 19 (juv. referred with ? to capensis Thiele). 1932. Turton. Mar. Sh. Pt. Alfred, p. 50 (ocelliferus and robustior), pl. xi, no. 369 (robustior juv.). Aperture (incl. canal) 14 to nearly 14 times spire. Profile of early whorls slightly angular, of later whorls angular (with knobs) (verruculatus), convex (adamsi). Protoconch 24 whorls, diam. and alt. 1-5-1:75 mm. (rarely perfect except in juv.), smooth but with some irregular plicae before the abrupt junction with 1st postnatal whorl. Postnatal whorls 8; axial ribs 9-10 on Ist whorl, increasing to 11, usually from suture to suture on first 3 whorls, but thereafter only at periphery where they become blunt, more or less complanate knobs, 11-15 in number, continuing on to 8th whorl forming a prominent shoulder in typical verruculatus, but petering out on 7th or 8th whorl in adamsiz ; crossed by spiral lirae, 3 (4) on 1st whorl, increasing to 5 (6) on 2nd and grd whorls, end and 3rd lirae strongest, thereafter the 3rd strongest, on the periphery and carrying knobs, i.e. 2 lirae above and 2 (sometimes 3) below the peripheral keel; 15-20 additional lirae on base, with intermediaries; also over the whole whorl fine spiral striae. Parietal callus not strong, sometimes in addition with 3-4. (up to 6) plicae; columella curved, rimate in half-grown to adult examples, forming a deep but narrow umbilicus. Canal nearly straight in juv., usually more strongly curved in older examples, subequal to and not sharply marked off from rest of aperture. Outer lip at some stages of growth plicate within. Periostracum thin, fibrous and imbricate, especially near the suture, fimbriate. CONTRIBUTIONS TO KNOWLEDGE OF S.A. MARINE MOLLUSCA QI (verruculatus) 150 (protoconch and tip of rostrum broken, say: 153) X 70 mm.; 119g X 45 mm.; (adamsiz) 133 X 48 mm.; 125 X 50 mm. Operculum oval, apex incurved, 35 < 18 mm. in 125 mm. shell. Cream or buff, the peripheral knobs usually reddish-brown, sometimes also indications of orange flames, periostracum greyish or yellowish-brown, operculum horny or reddish-brown; juveniles seem to. be more brightly coloured (at least in some beach-worn examples the coloration shows better), with darker knobs and flames. Animal bright red with minute white specks (K.H.B.). Radula in 7 mm. shell with 110 rows, lateral plate with 7 cusps, in 23 mm. shell resp. 165 and 8, in 25 mm. shell resp. 170 and 8, in 30 mm. shell resp. 205 and 9, in 38 mm. shell resp. 215 and g, in 62 mm. shell resp. 230 and 11-12, in 114 mm. shell resp, 285 and 12, in 150 mm. shell resp. 325 and 13 (the tiny denticle at inner end excluded in all counts). Dead and beach-worn specimens recorded from Port St. Johns, Port Alfred, Port Elizabeth, Agulhas Bank, Still Bay, False Bay (Sowerby, Adams, Bartsch, Turton and 8. Afr. Mus.). 5° 16’ S. 22° 26’ E., 155 metres (Thiele, juv.). Living: Simon’s Bay (False Bay), 15-20 fathoms (Watson); Algoa Bay and Agulhas Bank to mouth of False Bay, 10-66 fathoms (S. Afr. Mus. P.F. coll.). Sea Point (Cape Town), low tide (S. Afr. Mus.), Knysna, low tide (S. Afr. Mus. P.F. coll.). Both the latter adamsii form). 33° S. 28° 11’ E. (off East London), 31 fathoms (verruculatus form) (U.Q.T.). Saldanha Bay, 10-14 fathoms (S. Afr. Mus. P.F. coll.). 34° 35'S. 19° 14’ E., 66 metres; False Bay, 3-24 metres; west coast of Cape Peninsula, intertidal; Langebaan (Saldanha Bay); off Lambert’s Bay, 66 metres (U.C.T.). ) Remarks. The East London locality bridges the gap between Port St. Johns and Port Alfred. The verruculatus and adamsi forms are not restricted to separate areas. Juveniles from 5 mm. long (protoconch plus 1st whorl) examined. While the early whorls show little variation, the later whorls show marked dimorphism: the typical verruculatus with strong shoulder knobs, and adamsii with evanescent knobs and evenly convex whorls. The institution of adamsii as a distinct species is not surprising when only the extreme forms were available. But they are connected by transitional forms. Plump and slender examples occur in both verruculatus and adamsii, though the latter in general is the more slender. The most slender specimen I have seen is one (adamsi) taken in False Bay measuring 118 x 40 mm.; it is not scalariform but the spire is elongated to such an extent that its length equals the length of aperture (incl. canal). In the early whorls the axial ribs are usually well developed (see Thiele’s figure), and the bicarinate periphery on the 2nd and 3rd (sometimes also but less conspicuous on 4th) whorls is very characteristic. Q2 ANNALS OF THE SOUTH AFRICAN MUSEUM One specimen (S. Afr. Mus. no. A4661, off Cape St. Blaize) 82 mm. long, and a juvenile (locality ?) 20 mm. long, have unusually large protoconchs: diam. and alt. almost 2:3 mm. Another specimen, also from off Cape St. Blaize, 147 x 57 mm., has peripheral knobs extending on to the 8th whorl but the profile is convex, not shouldered; and the canal is markedly sigmoid. One specimen (S. Afr. Mus. no. A4662, off Cape St. Blaize) 137 (tip of canal broken, probably 140-142 when perfect) x 66 mm., is subscalariform, with strongly convex ventricose whorls, and deeply sunken sutures. It is a question whether the name ocelliferus in Lamarck’s Liste des objets (sometimes attributed to Bory, but see: Sherborn & Woodward. Proc. Zool. Soc. Lond., 1893, p. 584) should be used for this species. The figure is recog- nizable as a representation of this species, but is it adequate to i it from other species? Fusus colus Linn. Figs. 19(j), 20(d) 1816. Lamarck. Tabl. Encycl., pl. 423, fig. 2, and Liste, p. 6 (longicauda). 1859. Chenu. Man. Conchyl., i, fig. 597. 1876. Kobelt. Conch. Cab., p. 146, pl. 30, fig. 3, pl. 47, fig. 1. 1942. Gravely. Bull. Madras Govt. Mus., V, 2, p. 62, fig. xi, i (longicauda). 1952. Satyamurti. ibid., n.s. I, 2 (6), p. 187 (longicauda). Protoconch 24 whorls, diam. 1, alt. 1-3 mm., smooth, with a dozen or more fine axial ribs in last half whorl, which is more or less sharply demarcated from ist postnatal whorl. Radula in 25 mm. shell with 80 rows, lateral plate with 8 cusps, in 40 mm. shell resp. 140-170 and 8-9, in 97 mm. shell c. 220 and 11 (excl. the minute inner denticle); central plate broader than long, narrowed in front, tricuspid (in one specimen 4-cuspid). Living: off Umhloti and Umvoti Rivers (Natal), 25-27 fathoms; off Amatikulu River (Zululand), 24 fathoms (S. Afr. Mus. P.F. coll.). Delagoa Bay (S. Afr. Mus. coll. K.H.B., and U.W.). Inhambane (U.C.T.). Dead: off Tongaat and Umhlanga (Natal), 22-36 fathoms (S. Afr. Mus. P.F. coll; Remarks. F. toreuma (Martyn), recorded from Natal by Smith (1903), is distinguished by the angular profile of the whorls. If the Natal specimen was taken from a fish stomach, the species is probably living in South African waters. Fusus torulosus Lam. 1816. Lamarck. Tabl. Encycl. Meth., pl. 423, fig. 4, and Liste, p. 6. Specimen in S. Afr. Mus., probably from Ceylon or Indian Ocean. Aperture (incl. canal) longer than spire. Protoconch missing. Postnatal CONTRIBUTIONS TO KNOWLEDGE OF S.A. MARINE MOLLUSCA 93 whorls 10; profile convex with crenulations due to the spiral lirae. Axial ribs 5-6 on 1st and 2nd whorls, increasing to 10-11 on last, broad and rounded, from suture to suture; crossed by sharp spiral lirae 4—5 on 1st whorl, increasing to 10-11 on last, with an intermediary between the peripheral pair; c. 35 additional lirae on base and rostrum, some of them feebler than others. Growth- lines distinct immediately below suture, more or less so between the lirae on rest of whorl. Sutures deep. 91 (protoconch missing) X 24 mm. A fragment of two half whorls exactly agreeing with the sculpture of the above described specimen: off Cape Natal, 85 fathoms; off Umhloti River, 40 fathoms, 2 worn fragments; off O’Neil Peak (Zululand), 90 fathoms, one worn fragment (S. Afr. Mus. P.F. coll.). Fusus rubrolineatus Sow. Fig. 19(2) 1870. Sowerby. Proc. Zool. Soc. Lond., p. 252. 1880. id. Thes. Conch., iv, p. 80, pl. 411, fig. 68. 1903. id. Mar. Invest. S. Afr., ii, p. 228. 1903. Von Martens. D. Tiefsee Exp., vii, p. 30. 1903. Thiele. ibid., p. 169, pl. 9 (4), fig. 60 (radula). 1915. Bartsch. Bull U.S. Nat. Mus., 91, p. 47. Aperture (incl. canal) 1-1} times spire. Protoconch 24 whorls, diam. and alt. 1:2 mm., smooth, with 4-6 axial ribs on last half whorl, junction with Ist postnatal whorl distinct. Postnatal whorls 7; axial ribs 14-15 on 1st whorl, increasing to 17-18 on last whorl, from suture to suture, but evanescent towards suture on last whorl and on lower half of base, from about the 4th whorl slightly curved. below the suture; fine close-set growth-lines; crossed by 3 spiral lirae which appear alone on 1st whorl and continue as 3 main lirae on all following whorls but with added intermediaries (e.g. on 6th whorl 6—7 between suture and Ist main lira, 3 between Ist and 2nd, and between 2nd and 3rd, 3-4 below grd lira); 15-20 additional lirae on base, those on upper half with inter- mediaries; main lirae on the whorls and upper half of base forming horizontal complanate nodules at intersections with axial ribs. No parietal callus, colu- mella slightly curved, in some specimens with a slight swelling (scarcely a pleat) at the bend. Canal shorter than, but not marked off from rest of aperture, slightly curved. Periostracum thin, smooth. 38 x 13 mm., a plump specimen 32 X 13 mm. Operculum broadly oval, apex rounded, slightly on outer side of median line (von Martens said toward inner side), 6 X 4 mm. in 35 mm. shell, internal surface as in Fasciolaria. - Pale buff, main lirae on whorls and base, and also some of the more prominent intermediaries orange-brown, forming spiral lines, usually con- tinuous but often more intense on the axial ribs, producing an effect of series of spots or axial flames. Q4 ANNALS OF THE SOUTH AFRICAN MUSEUM Radula with 150-180 rows, central plate longer than wide, narrower in front, with 3 rather strong cusps, lateral plate with 5-6 cusps (excl. inner denticle), varying in size, and often asymmetrical. Agulhas Bank (Sowerby 1870); 35° 16'S. 22° 26’ E., 155 metres (von Martens); off Cape St. Blaize, 53 fathoms (Sowerby 1903). Living and dead: Agulhas Bank, from approximately 22° E. to 274° E., and southwards to Brown’s Bank approx. 363°S., 63-124 fathoms (S. Afr. Mus. 'P:F. coll:). Remarks. Von Martens (loc. cit. 1903. p. 103, pl. 2, fig. 10) described rufinodis from off Zanzibar and Sumatra, and compared (p. 104) it with rubro- lineatus. But he seems to have compared his specimens with Sowerby’s figure instead of with actual specimens of rubrolineatus which were available to him. The differences are not very convincing and the two species are obviously closely allied. F. rufinodis, however, has only 10 axial ribs on the 7th (pen- ultimate) whorl and 11-12 on the 8th whorl; and the central plate of the radula (Thiele, loc. cit., fig. 59) is nearly square. An even more closely allied species is F. libratus Watson 1886 from Fiji Islands, 315 fathoms. Most of the Pieter Faure examples are, like the Valdivia specimens, covered with an Alcyonarian. Fusus faurei n. sp. Figs. 19(h), 20(e) Aperture (incl. canal) 14 times spire. Profile of whorls convex, slightly biangulate. Protoconch 2 (24) whorls, diam. and alt. 2 mm., smooth (but all specimens more or less corroded). Postnatal whorls 6; axial ribs 11-12 on Ist whorl, increasing to ¢. 18-20 on last whorl (irregular and obscure on last half whorl), low, rounded not prominent except at the 2 peripheral lirae, slightly retractive below suture; crossed by 2 main peripheral costae from 1st whorl onwards, with finer lirae above and below, on last whorl 6-7 above and 3-4 below the peripheral costae, varying in strength, the lowest one (next the suture) the strongest, 2-3 between the 2 costae, 15-20 additional lirae on base, usually an intermediary between each pair. No parietal callus. Columella gently curved, no pleats, not rimate. Canal a little shorter than, and distinctly but not sharply marked off from rest of aperture, straight or very slightly reflexed at tip, open. Periostracum thin. 50 (protoconch and tip of canal broken) X 22 mm.; 33 X 15 mm. Operculum oval, apex blunt, incurved, 9 X 5 mm. in 38 mm. shell. Creamy-white, periostracum grey, operculum amber-brown. Radula with at least 180 rows, central plate narrowed in front, tricuspid, lateral plate with 11 cusps (excl. inner denticle). Living and dead: off Cape Point, 300-560 fathoms (S. Afr. Mus. A4581 (Type), Aq58e. PF. coll). CONTRIBUTIONS TO KNOWLEDGE OF S.A. MARINE MOLLUSCA 95 Fusus bonae-spei n. sp. Fig. 20(c), (f) 1925. Thiele. D. Tiefsee Exp., xvii, p. 183, pl. 32 (20), fig. 18 (capensis, non Dunker). Not the juv., p. 184, fig. 19 = verruculatus. Aperture (incl. canal) a little longer than spire. Profile of whorls convex but weakly biangulate. Protoconch 24 whorls, when not corroded diam. 2 mm., alt. 2-5-3 mm., smooth, junction with 1st postnatal whorl abrupt. Postnatal whorls 7; feeble axial ribs 8—g on Ist whorl, increasing to g-10, but not traceable beyond 4th whorl; growth-lines distinct; crossed by spiral lirae, on Ist whorl 3, the lower two stronger than the upper one, continued on following whorls, the two stronger ones forming the periphery, with weaker intermediaries producing a fine cancellate sculpture, 20-25 additional lirae on base also with intermediaries. No parietal callus, columella slightly curved, Fic. 20. (a) Fusus verruculatus Lam., protoconch, with detail of one lira. (6) F. colus Linn. protoconch. (c) F. bonae-spei n. sp. protoconch. (d) F. africanae n. sp., protoconch. (e) F. faurei n. sp., profile. (f) F. bonae-spet n. sp., profile. 96 ANNALS OF THE SOUTH AFRICAN MUSEUM canal 13-14 times, but not very sharply marked off from rest of aperture, straight or slightly curved and reflexed, narrow. Outer lip not plicate inter- nally. Periostracum thin, fibrous-fimbriate. 102 X 34 mm. Operculum oval, apex incurved, 20 X 12 mm. in 102 mm. shell, internal surface as in Fasciolaria. Creamy-white or pale buff, periostracum yellowish-buff, operculum brown. Dead: 34° 33’ S. 18° 2’ E., 318 metres (Thiele). Living and dead: off Cape Point, 85-256 fathoms (S. Afr. Mus. A4622, A4628-32, Type A4632. P.F. coll.). Remarks. I have not seen Dunker’s description or figure, but it seems most improbable that Thiele’s capensis is the same. Krauss gave the dimensions of Dunker’s species as 10 X 5:2 lines. Although there are eight specimens in S. Afr. Museum with their opercula, no animal has been preserved. Fusus africanae n. sp. Figs. 19(k), 20(d) 1925. Thiele. D. Tiefsee Exp., xvii, p. 184, pl. 32 (20), fig. 20 (‘Fusus n. sp.’, sine nom.). The Valdivia specimen from 35° 16’ S. 22° 26’ E., 155 metres, measured 22°5 < g mm. and consisted of a large protoconch and 3 whorls. The profile of the whorls is evenly convex, without the bicarinate periphery of juvenile verruculatus and bonae-sper. Thiele was therefore correct in regarding it as a distinct species. Unfortunately he did not state the number of ribs and spiral lirae; judging by his figure there might be 13 or 14 ribs and 4 or 5 lirae on the end and 3rd whorls. The Pieter Faure obtained two smaller specimens, 15 X 6:5 mm.; one in 1900 from off west coast of Cape Peninsula, 131-136 fathoms, and another in 1906 from Brown’s Bank (approx. 363° S. 21° E.), 80-100 fathoms (S. Afr. Mus. nos. A8826 and A8610 Type). Aperture (incl. canal) a little longer than spire. Protoconch 24 whorls, diam. 3, alt. 2°75-3 mm., smooth, with 2-3 feeble growth-lines (scarcely ribs) on outer lip. First postnatal whorl beginning abruptly, and not quite flush with outer lip of protoconch; 14 low feeble axial ribs on 1st and 15 on 2nd whorl (slightly irregular in A8826 owing to injury), 3rd whorl incomplete; crossed by 4 spiral lirae of nearly equal strength (the uppermost one slightly weaker) on ist and 2nd whorls, a finer one between 1st lira and suture, and indications of a fine intermediary between each pair of main lirae (more distinct in A8826 than in A8610), also a lira below the 4th but partly obscured by suture of following whorl; intersections very slightly nodulose; on base of last whorl c. 10 main lirae, with finer intermediaries. Columella curved, no visible pleats, canal slightly curved, a little longer than rest of aperture. Operculum oval, apex incurved, 3:5 X 2 mm. 6 CONTRIBUTIONS TO KNOWLEDGE OF S.A. MARINE MOLLUSCA 97 The Fisheries Survey vessel Africana obtained (1948) a larger example Peet, mm., from 34° 35'S. 19 14°. E., 66 metres (AFR. 864 E.). Protoconch 24 whorls, diam and alt. 3 mm., smooth, with 2-3 feeble lines of growth on outer lip. Postnatal whorls 5; 1st beginning abruptly and not quite flush with outer lip of protoconch; axial ribs on 1st whorl obscure, possibly 14, on 2nd whorl 15 (16) also feeble, on 3rd whorl 16 (17), on 4th 17 (18), thereafter evanescent; crossed by predominant spiral lirae, 5 on Ist whorl, the lower 3 strongest, a finer one between Ist lira and suture, and below the 5th another partly obscured by suture of following whorl; on 2nd and following whorls 6 lirae, the 3 lowest strongest on 2nd and 3rd whorls, the 4 lowest on 4th and 5th whorls, one fine intermediary between each pair of stronger lirae; on base of last whorl 12 (and some obscure ones on rostrum) additional lirae with fine intermediaries; lirae on later whorls distinctly flattened. A small dentiform parietal callus, no visible columella pleats. Operculum oval, apex incurved, 9 x 6 mm. Protoconch glistening white, shell greyish-white, operculum horn coloured. A very thin periostracum obscured by a thin layer of sponge; and with numerous oval horny capsules (maj. diam. 0-5 mm.) glued firmly to the shell (not belonging to this species, or any other Fasciolariid as they are not stalked). Radula with c. 190 rows, central plate tricuspid, lateral plate with 11 cusps (excl. the tiny inner denticle), the innermost 2 strong, the others alternately smaller and larger, the 10th cusp rather strong and the outermost one the smallest; the arrangement is not quite symmetrical on the two sides, nor on successive plates on the same side, particularly so on the left side; the right side as shown in fig. 19(k) seems to be the most usual arrangement. Radula of the two juv. P.F. specimens with 105-110 rows, lateral plate with 5 unequal cusps. Remarks. The Africana specimen is separately described. to show the slight difference in detail of the spiral lirae; viewing it side by side with the other two there is no doubt they are conspecific; and there seems little doubt that they are the same as the Valdivia example. The 25 mm. shell in bad condition figured by Thiele (loc. cit., fig. 19) from the same locality has a smaller protoconch and seems to be referable to verruculatus. The four known examples are all from moderate depths on the southern and south-western slopes of the Agulhas Bank, and it is surprising that no others have been obtained. From the size of the protoconch one would suspect an adult at least as large as verruculatus. The radulae of all three specimens are remarkable for the inequality of the cusps on the lateral plate, especially noticeable in the large Africana specimen, thus resembling the radula of Peristernia more than that typical of Fusus. 98 ANNALS OF THE SOUTH AFRICAN MUSEUM Fusus albinus A. Adams 1855. A. Adams. Proc. Zool. Soc. Lond., p. 222. 1880. Sowerby. Thes. Conch., p. 80, pl. 7 (411), fig. 72. 1903. Von Martens. D. Tiefsee Exp., vii, p. 9, pl. 2, fig. 9 (appressus). 1903. Thiele. ibid., p. 169, pl. 9 (4), fig. 61 (radula) (appressus). 1912. Dautzenberg. Ann. Inst. ocean., V, pt. 3, p. 28. Adams described ‘a large white solid species with moderately long beak, and with longitudinal rounded rib-like plicae which are obsolete at the sutures’ from Ichaboe Island, north of Liideritzbucht. Von Martens compared his species, 101 X 40 mm. from Great Fish Bay (Angola), 16 fathoms, with albinus, but concluded that the two were distinct. Dautzenberg recorded albinus from Mossamedes, 15-20 metres, and Praya Amelia, 15-35 metres, but made no mention of appressus. ? Fasciolaria holcophorus n. sp. Fig. 21 Fusiform, aperture subequal to spire. Profile of whorls convex, without shoulder. Protoconch 14 whorls, diam 0-6, alt. 0-5 mm., smooth, with 4 or 5 faint axial ribs on last half whorl, junction with Ist post- natal whorl indistinct. Postnatal whorls 5; axial ribs 10-11 on ist whorl, 8 on and, 7 on each of the 3rd, 4th and 5th whorls, from suture to suture and strong on early whorls, but from later part of 4th whorl becoming feeble and causing mere undulations on the periphery of 5th whorl, obsolete on base; crossed by spiral striae, 4 on 1st whorl, 7 on and, 8 on 3rd, 9 on 4th and 10 on 5th, about 16 additional on base, those on rostrum almost parallel with the canal, all striae regularly spaced except 2 or 3 fine intermediaries on base. Columella gently curved, with 3 pleats, the lowest one feeble. Canal straight (tip broken). Periostracum thin, smooth. I1 X 4 mm. White, periostracum pale buff. Off Cape St. Blaize, 125 fathoms, 1 dead (S. Afr. Mus. no. A881Q9. P.F. coll.). Remarks. In the absence of the radula the generic position of this pretty little shell is quite uncertain. There is a somewhat fanciful resemblance to Ptychatractus, which Fic. 21. . : te iprtiy stat Thiele (1929) removed from the Fasciolariidae to the Fasciolaria ? holco- : phorus n. sp. Vasidae. Fam. NASSIDAE Gen. NassA Lam. 1799. Lamarck. Mem. Soc. H. N. Paris, p. 71 (non Nassa Bolten, 1798). 1806, Duméril. ool. Analyt., p. 166 (Nassarius). a CONTRIBUTIONS TO KNOWLEDGE OF S.A. MARINE MOLLUSCA 99 1916. Tredale. Proc. Mal. Soc., xii, p. 82 (Nassarius). 1928. Tomlin. Ann. S. Afr. Mus., xxv, p. 313 (Nassarius). 1929. Thiele. Handbuch, 1, p. 322 (incl. Desmoulea [sic]). 1931. Lebour. 7. Mar. Biol. Assoc., xvii, p. 797 (eggs and larva). 1936. Peile. Proc. Mal. Soc., xxii, p. 139 (radula). 1939. id. ibid., xxii, p. 276 (radula). Remarks. Neither Nassarius Duméril 1806 nor Nassaria Link 1807 (Bucci- nidae) are to be rejected on account of the similarity of their termination (Rules Zool. Nomencl. Art. 36. Rec.), though Iredale thought otherwise; but as they are very liable to confusion, no apology is made for reverting to Nassa Lam; Thiele accepted it, several years after Iredale’s proposed alteration; and everyone knows the distinctive facies of a “Nassa’. As regards subgenera, until some agreement is reached on their definition, they are better ignored, at least so far as South African species are concerned. These are all typically ‘Nassa’, with the one exception of kraussiana, the radula of which is also distinctive. South African ‘beach-conchology’ has run rampant in this genus; and when once a species has acquired synonyms it seems to attract more synonyms, e.g. capensis and kochiana (cf. Tomlin). Most of Turton’s ‘n. spp.’ will probably prove to be synonyms, but without actual examination of his material one can only suggest possible or likely synonymy. He took no notice of Tomlin’s 1928 paper. : Nassa analogica Sow. Fig. 22(a) 1853. A. Adams. Proc. Zool. Soc. Lond. (for 1851), p. 113 (trifasciata, non Gmelin). 1903 (8th July). Sowerby. Mar. Invest. S. Afr., uu, p. 219, pl. 4, fig. 3. 1903. id. ibid., p. 228, pl. 4, fig. 2 (trifasciata Adams). 1903 (18th Dec.). Von Martens. D. Tiefsee Exp., vii, p. 27, pl. 3, fig. 18 (circumtexta). 1903. Thiele. ibid., p. 167, pl. 9 (4), fig. 52 (radula) (circumtexta). 1906. Smith. Ann. Natal Mus., 1, p. 36 (circumtexta and analogica). 1910. Dautzenberg. Act. Soc. Linn. Bordeaux, p. 55 (gallandiana Fischer). 1912. id. Ann. Inst. ocean., vol. V, fasc. 3, p. 33 (trifasciata Adams). 1923. Odhner. Géteb. K. Vet. Handl., xxvi, p. 6 (trifasciata Adams). 1925. Thiele. D. Tiefsee Exp., xvii, p. 182 (circumtexta). 1928. Tomlin. loc. cit., p. 316 (circumtexta). 1932. Turton. Mar. Sh. Pi. Alfred, p. 59 (trifasciata Adams). Radula. ‘Thiele gave the number of cusps on the central plate of a radula 4°75 mm. long as 8. In S. Afr. Mus. specimens of radulae 3:5-4:5 mm. long, there are 65-75 rows, central plate with 11-13 (rarely 10) cusps, slightly variable in size znier se, with a minute one externally on both sides, no inter- mediate plate, outer cusp of the lateral plate with a slight bulge (scarcely a denticle) on inner margin, but sometimes obscure, inner cusp rather slender. No difference in the radulae of the ‘trifasciata’ and analogica forms. The 1st whorl always, and usually also the 2nd whorl, appear to have only spiral lirae; but axial ribs may be developed on the 3rd and 4th whorls, or on the 4th and 5th, or on the 5th and 6th, or on all these whorls (3-6). The number of ribs is 14-15 on 3rd whorl, c. 20 on 5th, and 22-24 on 6th. oO ANNALS OF THE SOUTH AFRICAN MUSEUM Sl A) el WH) wi A — Ae. SI ceasing to be ribs, becoming nodules around gees WE : . : : FERRET the shoulder, increasing in prominence on later ages whorls; crossed by spiral lirae 4 on Ist whorl producing a clathrate sculpture, on 2nd only 2 lirae but numerous fine striae, which become very fine on 4th or 5th whorl and visible only above the shoulder, finally petering out, obscured by the growth-lines; base smooth, with growth-lines but only very faint indications of spiral lirae; shoulder at middle of whorl, profile above concave, below straight or slightly convex. Parietal callus dentiform, canal longer than rest of aperture, narrow, with nodule or angular ridge at its base, recurved, outer lip internally smooth. 55 X 24 mm. Smallest specimen seen 17°5 X 8-5 mm. Operculum broadly oval, nucleus apical, incurved, 14 X 8 mm. in 55 mm. shell. Creamy-white, irregularly suffused with Fic. 33. brown, and with brown flames, shoulder nodules Eythria queketti Smith, juvenile pale, operculum deep amber-brown. 17°5 mm., with protoconch of ‘ . £. ponsonbyi Sow. for comparison. Radula with c. 160 rows, central plate tri- ee y angular, length a little less than basal width, with 3 cusps, the middle one longer than its neighbours, lateralplate with outer cusp larger than the inner. Off Durban, 40 fathoms [from fish stomachs] (Smith; also $. Afr. Mus. Ross-Frames coll.). Off Umhlangakulu River (Natal), 50 fathoms, 1 dead, 1 living; off Umbhloti River (Natal), 40 fathoms, 1 juv.; off Cone Point and off O’Neil Peak (Zululand), 34 and 55 fathoms, 1 each, dead; Algoa Bay, 25 fathoms, 1 living (S. Afr. Mus. P.F. coll.). Remarks. The size and shape of the protoconch and the sculpture of the early whorls are important features distinguising this species from ponsonbyt. Also, in queketti the ribs decrease in number, whereas in ponsonby: they increase. Cooke’s figure of a radula ascribed to queketti is due to some mistake; it is 172 ANNALS OF THE SOUTH AFRICAN MUSEUM certainly that of an Afrocominella. The true radula is very like that figured by Cooke for E. cornea, but with the central plate more like that figured for linea, and the lateral plate with a slightly larger outer cusp. It is curious and suggestive that Thiele (loc. cit., p. 312) gave the distri- bution of cornea as ‘Mittelmeer bis Siidafrika’; I am not aware that cornea has ever been recorded from South Africa. Compared with a shell of cornea in S. Afr. Mus. (ident. Tomlin) queketti is broader and less fusiform owing to the stronger nodulose shoulder, and has a longer canal; but Smith (loc. cit., p. 111) said guekettt was more ‘slender’! The presence of this species in Algoa Bay is rather unexpected; the P.F. specimen is the only one, dead or alive, known from west of Natal. The juvenile, 17°5 < 8:5 mm. with protoconch and 4 whorls (fig. 33), appears at first sight very different from the adult, because the prominent nodules around the shoulder are developed only on the last two, or three, whorls. The size and squatness of the protoconch, and the sculpture of the early whorls agree exactly with older examples and leave no doubt as to its identity. There is just a hint on the back of the outer lip of the nodules to come. Euthria filmerae Sow. 1900. Sowerby. Proc. Mal. Soc., iv, p. 1, pl. 1, fig. 3. 1932. ‘Turton. Mar. Sh. Pt. Alfred, p. 54. Profile of whorls straight, without shoulder. Protoconch 14 whorls, diam. not quite 2 mm., smooth. Postnatal whorls 8; 1st whorl with 12-14 axial ribs, decreasing to 7 on last whorl; crossed by numerous spiral grooves, 4 on Ist whorl increasing to 11-13 on last whorl, 13-15 additional ones on base. Parietal callus nodiform, canal rather long, open, oblique, outer lip thin, internally smooth. 40 X 14 mm. (living); 47 * 16 mm. Operculum unguiform, nucleus apical, incurved, 13°5 X 4 mm. in 40. mm. shell. Lower part of body whorl below periphery white, rest of shell brown. Dead: Pondoland (Sowerby, also S. Afr. Mus.); Port Alfred (Turton). Off Itongazi River (Port Shepstone area, Natal), 24 fathoms, one living; off Amatikulu River (Zululand), 25 fathoms, one dead (S. Afr. Mus. P.F. coll.). Remarks. The living and dead P.F. specimens both have a very thin shiny periostracum (like a coating of white-of-egg). The former retains the operculum, but most unfortunately the animal was not preserved. Fam. PYRENIDAE 1902. Pace. Proc. Mal. Soc., v, pp. 36-154 (Columbellidae, list of species). 1929. Thiele. Handbuch, i, p. 302 (Columbellidae). 1931. Tomlin.” Ann. Natal Mus., vi, p. 436. Tomlin correctly takes the oldest genus to form the family name. ‘... it is a matter of considerable difficulty to satisfactorily subdivide the CONTRIBUTIONS TO KNOWLEDGE OF S.A. MARINE MOLLUSCA 173 Columbellidae into genera and subgenera; and the difficulty has been much increased by the misdirected efforts of the mere conchologist . . . whatever value a section may have originally possessed, its true significance has in most cases been entirely lost sight of by subsequent authors, and species have been scattered about among the various genera and subgenera in an amazingly haphazard fashion’ (Pace, p. 40). For example, apicata was put into Nitzdella by Smith, but into Alza by Bartsch. Thiele admits four genera, each with a distinctive radula. So far as is known the South Africa species fall into Pyrene and Columbella. The lateral plate of the Pyrene radula is tricuspid: one cusp about midway on the plate, varying slightly according to the species (e.g. filmerae, burnupi, albuginosa, fig. 34), and separated by a varying distance from the two apical falcate cusps. In the available South African material the only radula of the Columbella form is that of fulgurans, the lateral plate of which is also tricuspid, but only the apical one is falcate, the other two being broad, like shark’s teeth. Species dealt with here, but whose radulae are unknown, are included in ‘Columbella’. The identification of the smaller species, e.g. Thiele’s species from the Valdivia collection, is often difficult; and suggested synonymies are provisional. Pyrene albuginosa (Rve.) Fig. 34(f) 1859. Reeve. Conch. Icon., sp. 223. 1921. Sowerby. Proc. Mal. Soc., xiv, p. 126, fig. (approximata). 1926. Tomlin. Ann. Natal Mus., v, p. 291, pl. 16, fig. 5 (Mitrella natalensis). 1931. id. ibid., vi, p. 436 (Mitrella approximata). 1932. Turton. Mar. Sh. Pt. Alfred, p. 72 (var. major, nom. preocc.). 1933. id. 7. Conch., xix, p. 370 (nom. nov. var. rietensis). Protoconch 24-3 whorls (junction with Ist postnatal whorl indistinct), alt. 0-75, diam. 0-5 mm., smooth. Postnatal whorls 7; no axial or spiral sculpture, except c. 10-11 lirae on lower half of base and rostrum. Outer lip thickened, c. 6 plicae within; columella with 4-6 granules. Periostracum thin, crinkly, scarious. 12 X 4°5 mm. Pale corneous, uniform, or mottled or reticulated with fawn or orange- brown, usually an interrupted subsutural band, and 2 spiral series (one peri- pheral, one infraperipheral) of opaque white spots enclosing between them a pale non-reticulate band; some specimens uniform with only a pale peripheral band. Periostracum usually pale, but sometimes amber-brown. Radula with c. 225 rows, central plate very delicate, lateral plate apically bifalcate, proximal cusp rather strong. False Bay to Algoa Bay, Port Alfred, East London, and Natal (auct. and S. Afr. Mus.). Natal (natalensis). Off Cape Vidal (Zululand), 80-100 fathoms, one; off Morewood Cove, Tongaat, Umhloti, and Umhlanga (Natal), 22-36 fathoms, 14; off Umkomaas 174 ANNALS OF THE SOUTH AFRICAN MUSEUM River, 40 fathoms, 5 juv.; off East London, 20 fathoms, one; all dead but more or less fresh (S. Afr. Mus. P.F. coll.). 29° 38'S: 31° E., 49 metres Weta, Living: False Bay and East London (U.C.T.). Remarks. The juveniles collected at Still Bay by Dr. Muir clearly show that natalensis is a synonym. Specimens in 8S. Afr. Mus. labelled albuginosa and floccata Rve. do not seem very different, except the latter are larger and broader with less tapering spire, and are also more strongly marked with orange-brown blotches and reticulation; both show on the body-whorl 2 spiral series of opaque white spots, enclosing between them a pale non-reticulated band, also usually a series of white spots below the suture (on the upper whorls the peripheral spots just show above the suture of following whorl). Some specimens have more or less regularly spaced, straight or crinkly, axial bars between suture and periphery; some are orange-brown with a pale peripheral band, others uniform orange- brown. Von Martens (1903. D. Tiefsee Exp., vii, pp. 56, 106) mentions uniform ‘scarlet-red’ examples of floccata from Pondoland; and a beach example from Natal in 8. Afr. Mus. is reticulate with brown, but has the peripheral band bright pink. The same mottled and reticulate pattern is found in C. seychellarum von Martens (loc. cit., p. 105, pl. 5, fig. 17) which, however, is an even more broadly oval shell than floccata. 7 SING, a Fic. 34. Protoconchs of (a) Pyrene kraussii (Sow.); (6) P. burnupi (Smith); (c) P. dianae (Thiele); (2) ‘Columbella’ hella Thiele. Central and lateral radula plates of (e) P. burnupi (Smith) ; (f) P. albuginosa (Rve.), lateral plate only; (g) P. filmerae (Sow.), lateral plate only; (h) Colum- bella fulgurans Lam. A CONTRIBUTIONS TO KNOWLEDGE OF S.A. MARINE MOLLUSCA 175 A comparison with apicata Smith is also not irrelevant. The colour- pattern is very similar, and although the colour is never such a deep brown as in apicata, living and fresh albuginosa often show a very clear pale brown mottling. P. albuginosa, however, does not have a bulbous protoconch, and I have not seen a trace of spiral striae on the last whorl in any specimen. Pyrene filmerae (Sow.) Fig. 34(g) 1892. Sowerby. Mar. Sh. S. Afr., p. 21 (sagena, non Rve.). mgopsuid.. Proc. Mal. Soc., iv, p. 3, pl. 1, fig. 8. 4 Protoconch 2 whorls, alt. and diam. o-8—o-g mm., smooth, slightly lop- sided. Periostracum forming a crinkled and scarious band below the suture, giving a somewhat coronate or turreted appearance to the spire. The characteristic (as Sowerby said; but see also splendidula Sow.) dark brown band against (below) the suture is usually interrupted by white spots, which may be so large as to disrupt the band into a series of alternating brown and white areas; this is particularly noticeable in the Natal shells, some of which also show triangular brown marks on a white ground instead of the more usual white spots on a brown ground. The dark sutural band is not very obvious in the fresh specimen from Durnford Point, which is cream with, on the body-whorl, a peripheral series of orange-brown spots through which runs a continuous thin white line, a less distinct series of spots on middle of base, protoconch pinkish. Aperture in living examples violaceous. Periostracum dull brown, obscuring the bright pattern seen in beach examples; in the Durnford Point example pale amber. Radula with c. 200 rows, central plate very delicate, lateral plate apically bifalcate, the proximal cusp small. Port Elizabeth and Pondoland (Sowerby). Port St. Johns, and Natal (between Durban and Port Shepstone) (S. Afr. Mus.). Off Durnford Point (Zululand), 13 fathoms, 3 dead, but one with perio- stracum (S. Afr. Mus. P.F. coll.). Living: Umgazana (south of Port St. Johns), littoral (U.C.T.). Remarks. ‘Turton obtained no examples at Port Alfred, and consequently one suspects that the Bairstow and Filmer shells originally came from farther north. The scarious band of the periostracum is seen in the Umgazana shell, but is more strongly developed in the fresh specimen from Durnford Point. The latter was identified by Sowerby (3rd) as ‘probably splendidula’ (see Sowerby: 1847. Thes. Conch., i, Columbella, pl. 37, figs. 65, 66). If he had seen the other two shells from the same haul, which are obviously worn filmerae, he might not have suggested the Philippine splendidula. Some specimens are very similar in pattern to the illustration of tringa Sow. (loc. cit., pl. 37, fig. 62). 176 ANNALS OF THE SOUTH AFRICAN MUSEUM Pyrene pura (von Martens) Fig. 35(a) 1903. Von Martens. D. Tiefsee Exp., vii, p. 25, pl. 2, fig. 14 (Euthria p.). 1925. Thiele. ibid., xvii, p. 173, pl. 30 (18), fig. 21 (Columbella helena). 1925. id. ibid., p. 180 (Euthria p.). Protoconch 2 whorls, alt. and diam. 1 mm., smooth (but all specimens worn). Postnatal whorls 5; spire subtending an angle of 35°, profile of whorls gently convex. Growth-lines but no axial sculpture. Fine spiral striae over the greater part of whorl, but (in the present more or less corroded specimens) variable, when traceable c. 8 on 3rd whorl, increasing to ¢. 12 on 5th; additional striae on base ¢. 20 (scarcely traceable on rostrum). Outer lip thickened. Periostracum thin. 14 X 6 mm.; 17 X 7°5 mm. Fic. 35. (a) Pyrene pura (von Martens). (5) P. parhelena n. sp. (c) ‘Columbella’ polyarosus n. sp., apex. (d) ‘Columbella’ confertilirata n. sp., apex and operculum. White; two shells with very faint indications of subsutural brown spots and even fainter peripheral marks. Periostracum pale buff or yellowish. Radula with c. 200 rows, central plate very delicate, lateral plate apically bifalcate, proximal cusp moderately strong, separated rather widely from the apical cusps. 34° 31'S. 23° 2’ E., 500 metres (von Martens: one); 35° 16’ S. 22° 26’ E., 155 metres (Thiele: one pura, 2 helena). Cape Point NE. 4 N. 18 miles, 135 fathoms, one living; Vasco da Gama Peak N. 71° E. 18 miles, 230 fathoms, one dead; Lion’s Head N. 67° E. 25 miles, 131-136 fathoms, 2 dead; 10 dead without precise locality (S. Afr. Mus. P.F. coll.). Remarks. These shells are clearly referable to yon Martens’s species. Thiele separated as helena two out of three shells taken in the same locality, seemingly because these two showed spiral striae only on the lower part of the whorl. His figure of helena resembles that of pura in shape. 1l CONTRIBUTIONS TO KNOWLEDGE OF S.A. MARINE MOLLUSCA ey ig Pyrene parhelena n. sp. Fig. 35(6) Protoconch 2 whorls, alt. 1:25, diam. 1 mm., smooth, junction with Ist postnatal whorl distinct. Postnatal whorls 5; spire subtending an angle of 30°, profile of whorls nearly straight; growth-lines but no axial sculpture. Extremely fine spiral striae over whole whorl, from 3rd whorl stronger striae appear on lower half of each whorl, 4—5 on 3rd whorl, increasing to 6—7 on 5th; additional striae on base c. 20. Outer lip thickened submarginally. Periostracum thin. 14 X 5 mm. Pale fawn with faint white spots; periostracum pale. Radula with c. 200 rows, central plate very delicate, lateral plate apically bifalcate, proximal cusp moderately strong, well separated from apical cusps. Cape St. Francis NE. x E. 4 E. 36 miles, 70 fathoms, one dead; Cape St. Blaize N. x E. 73 miles, 125 Se one living, 3 eee (S. Afr. Mus. A886, moe70 (lype). P.F. coll.). Remarks. Similar in shape to barbara ‘Thiele 1925, but with different sculpture; sculpture similar to that of helena, but shape different. Pyrene kraussiu (Sow.) Fig. 34(@) 1844. Sowerby. Proc. Kool. Soc. Lond., p. 53 (Columbella k.). 1847. id. Thes. Conch., i, Columbella, sp. 99, p. 144, pl. 40, figs. 180, 181. 1848. Krauss. Sidafrik. Moll., p. 109, pl. 6, fig. 11 (Mangelia fulgurans). 1848. id. ibid., p. 122, pl. 6, fig. 17 (cereale Menke in litt.). 1860. Gould. Proc. Bost. Soc. Nat. Hist., vii, p. 334 (fulminea). 1894. Sowerby. 7. Conch., vii, p. 7 (kitchingz). 1897. id. Append. Mar. Sh. S. Afr., p. 10, pl 6, fig. 3 (kitching?). 1915. Bartsch. Bull. U.S. Nat. Mus., 91, p. 56, pl. 37, fig. 5 (alfredensis). 1931. Tomlin. Ann. Natal Mus., vi, p. 436. 1932. Turton. Mar. Sh. Pt. Alfred, p. 70, pl. 17, no. 500 (var. albanyana). 1932. id. ibid., p. 70, pl. 17, no. 504 (helena, non Thiele). Protoconch 24-3 whorls, alt. and diam. 0-5 mm., smooth, last whorl feebly keeled, with minute spiral striae below the keel (seen only in unworn examples), the keel runs down obliquely at junction with Ist postnatal whorl. Postnatal whorls 4-5; axial ribs 10 on each whorl. Radula with 110-150 rows, central plate very delicate, lateral plate apically bifalcate, proximal cusp rather strong (cf. albuginosa). (krausswu and kitchings forms examined.) Table Bay, and False Bay to Durban and Tongaat (S. Afr. Mus.). Off East London, 32 fathoms (S. Afr. Mus. P.F. coll.). Living: Port Nolloth, Langebaan (Saldanha Bay), False Bay, and Knysna (GEE. 'E.). Remarks. Sowerby (1892, p. 4) doubted whether fulgurans was a Pleuro- tomid, and Turton (1932) listed it definitely as a Columbella, probably after con- 178 ANNALS OF THE SOUTH AFRICAN MUSEUM sultation with Tomlin. I have seen examples from Knysna (Krauss’s type locality), including plump (kraussi) and slender (fulgurans) forms. Turton’s name helena is preoccupied by Thiele, but the two are quite different species (cf. pura). ‘Turton said his helena had no zigzag lines, but his photograph shows them. Although the characteristic zigzag lines are usually present and easily visible, in some specimens they are obscured by a uniform chestnut-brown coloration, with sometimes a series of pale spots below the suture and another below the periphery (kztchingz). The Durban specimens, taken alive by Burnup, are more delicate and translucent than specimens from other parts of the coast, especially the uniform brown kztching: form from False Bay and the west coast. aberr. zo Bartsch 1915. Bartsch; loe.1cit:; jo. 57,137. SA: 1931. Tomlin. loc. cit., p. 436 (as kraussi aberr.). Tomlin regarded 20 as an abnormal kraussw. It is more slender even than the fulgurans form. Although I have seen no similar aberration among the numerous examples from False Bay and Still Bay, there is one specimen taken together with two kraussi at ‘Tongaat which agrees with Bartsch’s description and figure. Protoconch 2 whorls (but worn), diam. 0-5 mm., smooth. Postnatal whorls 5; 1st whorl worn, 2nd and 3rd each with to axial ribs, but evanescent towards end of 3rd whorl, protractive as in kraussu (Bartsch said retractive, but see his figure), obsolete on 4th and 5th whorls. 7 x 2°5 mm. Translucent, with orange-brown zigzag axial lines. Port Alfred (Bartsch; one specimen; Turton: ‘rare’). Tongaat (Natal), one specimen (8S. Afr. Mus.). Pyrene burnupi (Smith) Fig. 34(d), (e) 1901. Smith. J. Conch., x, p. 112, pl. 1, fig. 2 (Columbella b.). 1932. Turton. Mar. Sh. Pt. Alfred, p. 67, pl. 16, no. 488 (Columbella kowiensis). Protoconch 3 whorls, alt. 0-6, diam. 0-5 mm., smooth, last whorl with feeble peripheral keel which runs down obliquely at junction with Ist postnatal whorl. Postnatal whorls 3-4; axial ribs 12-13 on 1st whorl, 13-14 (15) on end—4th; crossed by spiral lirae 4 on 1st whorl, 5 on 2nd, 6-7 on 3rd and 7-8 on 4th, additional lirae on base 9-10; intersections forming rounded granules or beads. 4°5 X 1°5-1°75 mm. Translucent yellowish, 3-4 red-brown interrupted lines (on the spiral lirae) around middle of whorl, lower part of base also with dark lines or spots. Radula with c. 200 rows, central plate delicate, lateral plate apically bifalcate, proximal cusp well separated from the apical cusp. CONTRIBUTIONS TO KNOWLEDGE OF S.A. MARINE MOLLUSCA 179 Natal (Smith); Port Alfred (Turton). Living: Durban and Scottburgh (S. Afr. Mus. coll. Burnup). Pyrene langleyt (Sow.) 1897. Sowerby. Append. Mar. Sh. S. Afr., p. 10, pl. 8, figs. 8, 9 (Columbella I.). Protoconch 2 whorls, alt. 0-6, diam. 0:5 mm., smooth. Postnatal whorls 4. (natalensis: 5); growth-lines but no axial ribs, occasionally a thickened growth-line simulates a rib; no spiral sculpture except some (c¢. 10) obscure striae on base. 4°5 X 2 mm. Corneous-brown, uniform or with 2 series of white spots, one infrasutural, one from top of aperture. Radula with c. 200 rows, central plate very delicate, lateral plate apically bifalcate, proximal cusp strong, well separated from apical cusps (cf. burnupr). Port Elizabeth (Sowerby); Port Alfred (Turton); Kalk Bay and Buffels Bay (False Bay) (S. Afr. Mus.). Living: False Bay (U.C.T.). Pyrene lightfoott (Smith) 1901. Smith. 7. Conch., x, p. 112, pl. 1, fig. 3 (Columbella l.). 1932. Turton. Mar. Sh. Pt. Alfred, p. 67, pl. 16, no. 487 (var. assimilans). Protoconch 2 whorls, alt. and diam. 0:8 mm., smooth, junction with Ist postnatal whorl distinct. Postnatal whorls 3. Axial ribs, when present, c. 15 on Ist whorl, ¢. 17 on 2nd, ¢. 17-19 on 3rd; crossed by spiral lirae 7 on Ist whorl, 8 on 2nd, 9-10 on 3rd, additional lirae on base 18-20; lirae flattened, broader than the sulci. 7 en - - * Out of print: Vols. I, II beck I-33 55 75 3), III (Ste ‘is bi r (Pate 4, 1,2 Index), VII (Parts 1, 2, 3, 5), VIII, 1X (Parts 1, 2), X- a 2), XI ar 2, _ (Part 1), XXIV (Part 2), XXXI (Parts I-3). “Vol. 1g00-1902 1903-1905 1903=1908 1906-1910 1908-1910 1908-1913, IQII-1918 IQII-1914 I1g91tI-1918 1913-1924 1913-1923 1915-1924 1914-1916 191977933 1917-1920 1921 1924-1925 1924-1926 1925-1927 1925-1928 1925-1926 | 1929-1938 1927-1928 1928 1929 1929-1932 Palaeontology 1929-1931 Zoology .. bee uy 1931-1935 Zoology .. of papers, authors, and muljeees ubledtiod in "AFols. 1934-1950 Palaeontology : 1935-1940. Zoology’, So ae pen 1939 Zoology .. .. a 1938 Zoology... ite we 1956 Zoology .. Ne A 1942-1948 Zoology .. ae a 1947-1952 Archaeology ty ue 1950 Zoology .. ee ee 1952 Zoology .. rae kiran 1952-1956 Botany We aM ok 1952-1955 Zoology 1953-1956 1955-1957 1957-1959 Zoology and Geology Zoology .. Palaeontology Geology, ican, Zocleey, Anthropology. ake, Parts 1, 2,5, 8,9) ..- Zoology : Palaeontology a oe Botany A Pg Wear ai Zoology .. ou Zoology . Palaeontology and fos Se and tie! Zoology a Zoology .. wes oe at Botany .. 4 se Zoology .. Ne fy, POGIOEY 7 0 2B na Zoology .. els te Zoology «ae sis Zoology . bie eats vias Plecauiclagy wand Geile | Zoology .. Anthropology and Enology Zoology .. ale ‘ Zoology .. ate ae Anthropology Bi o. Palaeontology Ms Zoology and Palasoniolaey Zoology and Palaeontology Copies may be obtained from— ee * ¥ . A i Oe pe ele 2: “(Oka Part Ly Index) . (Parts 4 and 6 only) — “(exc Parts 1 and 2) . (excl. Part : (excl. Parts Ce Index) ee ee ee e ee es ee ee (excl. ‘Part » (exc ‘Part 2) I-XXX ‘soph ie 4 eee The LIBRARIAN, Sourn Arrican Museum, Care Town. mH N OWN OAD < o Seaita a De we DY we DOB OO OO 7 8b OD 16 * oa ae By Bis ei . 2 9 10. a ey 13° 5 bed eee of 19.0% 4177 3. ° © wo oo ~ 2 ? a ate © 69 25 Bers eine. Oo Oo wt BSSoGoaBuUs aa i -) ca -y ° < 7. She no oO 4 ad co coone ¥ hu OO ~ ‘ eee =“ oe , Ce] Gowoe On a DAO ON stb ANNALS OF THE SOUTH AFRICAN MUSEUM VOLUME XLV PART II, containing: — Additions to the South African Museum Collection of Marine Fishes. By F. H. Tarsot, M.Sc. Note on Locality Records of Freshwater Fishes presented by F. D. McKean to the South African Museum. By R. A. Juss. ISSUED APRIL 1960 PRICE od. PRINTED FOR THE TRUSTEES OF THE SOUTH AFRICAN MUSEUM BY THE RUSTICA PRESS (PTY.) LIMITED, COURT ROAD, WYNBERG. CAPE Ny NR Baca DARIUS f { \ 5) r \ ‘ ri A = i a eee ‘ ¢ Py ci = thie ‘ ey Gee Saks be iee Nop & she rey ae mie. ‘ 4 ; sans 4 ihc Bean . x) aes ml: ADDITIONS TO THE SOUTH AFRICAN MUSEUM COLLECTIONS OF MARINE FISHES By F. H. Tarsot, M.Sc. ra . . Barbourisia rufa Parr* Parr, A. E., Barbourisidae, a new family of deep-sea fishes, Copeta, 1945, 127-129, 1 pl. Two specimens, the locality of one ‘off Table Bay 1956’, no depth given, and the other 32°10'S./16°15’E., 285 fathoms, 1959. (Registered numbers S.A.M.19967 and S.A.M.22893 respectively.) As the only record in literature of Barbourisia rufa is a single specimen taken from the Gulf of Mexico in 1937, these two examples from the South Atlantic off Cape Town are a most interesting find and greatly extend the distribution of this species. The present two specimens differ slightly from the 180 mm. holotype, the main differences being: slightly smaller orbit, slightly shorter jaws, and shorter gill rakers. In all other measurements and characteris- tics they show very close similarity (see Table 1) except that in Parr’s description of the genus he states ‘7 soft branchiostegal rays’, and the present specimens have eight. As the thick skin covering the rays and their soft nature make counting difficult it is probable that this will prove to be a miscount. A pseudobranch of a few gill tufts is present. Lateral line pores between 6 and 16 shows the occasional raised pores mentioned by Parr. All pores are small and without flaps, excepting the last which is very large, being the full width of the lateral line tube. Post-mortem colour is a brilliant scarlet-red (Ridgeway), which fades rapidly on preservation to an off-white. Pterothrissus belloct Gadenat Cadenat, J. 1938. Note sur deux poissons nouveaux de la céte occidentale d’Afrique. Bull. Mus. Hist. nat., Paris (2), 10, 361-369, 2 pl. Poll, M. 1953. Poissons. III. Téléostéens Malacoptérygiens. Résult. sci. Expéd. océanogr. belge Eaux cot. afr. Atlant. S., 4 (2), 1-258. Four specimens trawled off Walvis Bay, 22°2’S./13°26’E., in 80 fathoms (registered number $.A.M.21706). Donated by the Director, Division of Fisheries, Department of Gommerce and Industries. Poll (1953) has found this species to be abundant on the deeper portions of the continental shelf off tropical West Africa. The present specimens are the most southerly record. *Since the above was written R. R. Rofen, Galathea Rep., 1, pp. 255-260, September 1959, records and plates a specimen of B. rufa, from near Madagascar, mid-water, 450-700 metres. This extends the range of the family to the Indian Ocean. 257 ITHSONIAN Se STITUTION JUN 9 1953. 258 ANNALS OF THE SOUTH AFRICAN MUSEUM Type S.A.M. S.A.M. 19967 22893 Standard length ai i a4: nf .. 180 mm. 297 mm. 294 mm. Headii) ce 54 ye on ee Be 35°6 31 32 Snout .. ie ae a ee He ae Wiley Lion 11°6 Orbit diameter ap ts ay a Bi 4 pai! Py) Upper jaw length a Se # ne af 25°6 22 23 Low jaw length che Ae he the ae 28 21 Ba Greatest width of skull ae oP de oF 15 12°8 15 Interorbital distance .. =i Ay. KY Ge. 12" 10°4 LE=G Snout to D. origin... Ns wie 3 st 64 61 61 Snout to A. origin... oe ae fs sas 69 68 71 Snout to V. origin... ne dik gs is 54-55 50 50 Dorsal fin base. . m7 a WY dibs aah 24 26 2 Anal fin base $A A Ene ue oa 18 17°5 16°3 Width of V. base a BS ae aah oe eae Ig 199 Widih of pectoral base es ats Ns aad ehab, Vee 2°4 Length of pectoral rays si as kes .. 6:°5-7 7° 4. 6-1 Length of ventral rays aie ane 1 .. 66°5 6-4 Bek Longest dorsal ray .. A Mt, a es 1 8-4 8-2 Longest anal ray , sie dvs ail ee 10°5 8-4 9°5 Depth at shoulder sails we ae ae a 25 20 23 Least depth caudal peduncle oe oes ane 5 oon 8-2 D. to procumbent caudal rays ii oe bie 8 won| 8-2 D. to mid-base of caudal rays 6 a ~ 14-5 1471 15*3 Longest gill rakers .. se me a Bi Are 2°4 257 Dorsal count .. a se i aa as 20 20 20 Anal count a : ie ish ns 16 17 16 Number of lateral ee Sones set ia: Ss 29 29 30 Branchiostegal rays .. Bs st a a 7(?) 8 8 Gill rakers 5 me Lo Be ae . 64 14 6+ 15 6+ 15 TasLe I. A comparison of proportions of the type of Barbourisia rufa (from Parr, 1945) with two South African specimens. Lengths are expressed as a percentage of standard length. Allothunnus fallai Serventy Serventy, D. L. 1948. Allothunnus fallai, a new genus and species of tuna from New Zealand. Rec. Canterbury (N.&.) Mus., 5, 131-135. One specimen, fork length 835 mm. taken by spear gun in 3 fathoms off Millers Point, Cape Peninsula, on 8 April 1958, and donated by Mr. D. Hammond (registered number S.A.M.21546). The Slender Tunny, with its reduced dentition and high gill-raker count, is stated by R. A. Falla (in a note appended to Parrot, 1958, Rec. Dom. Mus., Wellington 3, p. 119) to be ‘not uncommon in southern New Zealand seas, but rarely caught’. Serventy’s original 3 specimens were from South Island, New Zealand, south of 43°, and Falla mentions sight and photographic records from the Auckland Islands (50° S.), so it can presumably be considered a cold-water tunny. This is the first South African record. The present specimen shows some small differences from the original description (see Table II), but there seems no doubt that it is conspecific. ADDITIONS TO S.A. MUSEUM COLLECTION OF MARINE FISHES 259 Scaling is not as complete as in the type. Behind the distinct corselet of larger scales the body is covered in fine scales on its upper half to below the lateral line and the lower half of the sides and the belly are naked. The vomer and palatines are slightly rough to the touch, being covered in microscopic granular teeth. Serventy doubtfully gives a vertebral count of 41. X-ray photographs of the present specimen show 39 vertebrae however. Fin counts: Dorsal XVII 12 plus 7; anal 13 plus 7. Gillrakers 23 plus 48 left, and 22 plus 53 right. * The specimen is a mature male. Post-mortem colour was steely blue above shading to silver below, and with no distinctive markings. A photograph taken immediately after death shows a dark line from the pectoral tip, running longitudinally and upwards to the 3rd dorsal finlet. Type S.A Fork Length (snout to caudal fork) a ah ae .. 616 mm. 850 mm. Diameter of eye - a Life a sub We is Q°7 3°6 Head Length .. of; Bee aie Se ef, = sie 26 26 Snout to origin of pectoral .. me Sp yp te as 27 28 Snout to origin of first dorsal or MA st se ba 31 30 Snout to origin of second dorsal .. ss a st ae 63 59 Snout to origin of ventral .. st An ee a ce 28 31 Snout to vent .. : 64 61 Depth of body at Sees of Pa D. te = Speck greatest t depth) a1 20 Depth of body at vent ae ‘ ; 18-8 17°4 Length of pectoral .. : aif Ae oi és 10°5 11-2 Length of pectoral pared. pode aN Semen cre a 12°2 12-2 Inter-orbital distance .. ae Be sy o Ae ea 7°8 a2 Length of maxillary .. : : ae me wf bie 9°3 9°3 Snout to hinder edge of gece is sts ae ate 20°1 19°8 Height of first dorsal . oY Ke am oy oie ay 10°4 8-1 Height of second Heed 7°9 7°5 Height of anal 5 ie ae 7°8 7°78 Snout to anterior nostril ae Sue ee is ms ‘e 4°6 4°9 Snout to posterior nostril wg aS af be An 6-8 6-9 Longest gill raker ue ee be ue ue ie ie 4°5 452 Tas_e IT. A comparison of the proportions of the type of Allothunnus fallai with the South African specimen. Fork length is in mm. and all other measurements are expressed as a percentage of fork length. ~NOTE ON LOCALITY RECORDS OF FRESHWATER FISHES PRESENTED BY F. D. McKEAN TO THE SOUTH AFRICAN MUSEUM By R. A. Juss Department of Ichthyology, Rhodes University, Grahamstown Gilchrist and Thompson (1917) have recorded species of fishes presented by Mr. F. D. McKean and have given the locality as Sawmills, Bulawayo, Rhodesia. Sawmills is actually a railway station approximately 60 miles north of Bulawayo, on the Bulawayo-Victoria Falls railway line, and situated on the Umgusa River. The Umgusa is a tributary of the Gwaai River which belongs to the Middle Zambezi River system. The fish fauna of the Upper Zambezi system above the Victoria Falls differs from that of the Middle Zambezi system (Jubb, 1958, p. 178). Except for Tilapia melanopleura, which is widely distributed, the species represented in McKean’s collection are found only in the Upper Zambezi system and the Kafue River, and not in the Middle Zambezi River system; as a typical example the predatory Serranochromis, much sought after by anglers, are conspicuous by their absence from this section of the Zambezi River. It would appear that Mr. McKean’s address, prior to 1917, has been recorded as the locality from which the specimens were collected. It has not been possible to get any information about Mr. McKean. The names listed are in accordance with Barnard’s (1949) revision of the South African Cichlidae. I have personally examined the Serranochromis specimens in the South African Museum which were presented by Mr. McKean. It is highly probable that these fishes actually were collected in the Zambezi above the Victoria Falls. . 548 Ctenopoma multispinis Peters. . 538 Recorded as Pelmatochromis robustus, but is Haplochromis smith Castelnau. . 526 Three specimens recorded as Serranochromis thumbergi Castelnau. . 525 Two specimens recorded as Serranochromis angusticeps Boulenger. 499 One specimen Tilapia mackeani Gilchrist & Thompson, 1917, which is a synonym of Tilapia melanopleura Dumeril. p. 482 Two specimens Tilapia intermedia Gilchrist & Thompson, 1917, which are synonyms of Tilapia macrochir Boulenger. p- 481 Tilapia kafuensis Boulenger. ode oo BARNARD, K. H. 1949. Revision of South African Cichlidae. Rep. Inland Fish. Dept., C.G.H., No. 5 (1948), 48-61. Giucurist, J. D. F. & THompson, W. W. 1917. The freshwater fishes of South Africa. Ann. S. Afr. Mus., 115 465-575. Juss, R. A. 1958. A preliminary report on the collections of freshwater fishes made by the Bernard Carp expeditions to the Caprivi Strip, 1949, the lower Sabi River, 1950, and to Barotseland, 1952. Occ. Pap. nat. Mus. S. Rhod., No. 22B, 177-189. 260 ANNALS SOUTH AFRICAN MUSEUM VOLUME XLV PART III, containing :— The Polychaet Fauna of South Africa. Part 5. Errant Species dredged off Cape Coasts. By J. H. Day. (With 14 figures in the text.) (a din T ea vin ie oi i qs cw NS iy Vege! ISSUED MAY 1960 ~~ PRICE 16s. PRINTED FOR THE TRUSTEES OF THE SOUTH AFRICAN MUSEUM BY THE RUSTICA PRESS (PTY.) LIMITED, COURT ROAD, WYNBERG, CAPE vrs THE POLYCHAET FAUNA OF SOUTH AFRICA PART 5. ERRANT SPECIES DREDGED OFF CAPE COASTS by J. H. Day Professor of Zoology, University of Cape Town CONTENTS page page Introduction .. “2 as = 201 Systematic account ... ae soe 27K! List of collecting stations .. ne) 204), Weferences ot a i ila ie af INTRODUCTION Earlier papers in this series have dealt with the estuaries and intertidal fauna of Mogambique, Natal and Cape Coasts. This is the first account of species which occur below tide-marks and it will be evident from the long lists of collecting stations that a great deal of field-work has been done since the work first started in 1947. Actually the prime object of this dredging survey is to determine distribution patterns around the southern part of Africa and not only the Polychaeta but the whole benthonic fauna is being surveyed. Many systematic reports on the other groups have been published and further work is in progress. All the records are being entered on card catalogues and when the area has been adequately covered a biogeographic analysis will be attempted. Meanwhile many new species are being discovered. The material has been obtained from many sources. I am indebted to the Director of the Division of Fisheries for dredgings brought in by R.S. Africana II and Palinurus and I also wish to thank Messrs. Irvin and Johnson for allowing students to collect specimens on board commercial trawlers. The Hydrographic Section of the South African Navy generously provided facilities for dredging on board S.A.S. Natal during three cruises between Port Elizabeth and Durban and Dr. Nafe of the Lamont Geological Observatory kindly allowed me facilities during the cruise of the Vema between Cape Town and Durban. To all of these organizations I tender my thanks but the bulk of the material has been collected by members of my own department working on small boats _ during university vacations. In this way rich hauls were made in Lamberts Bay, Saldanha Bay, Table Bay, False Bay, Mossel Bay and Algoa Bay. The expenses of such trips were covered partly by grants from the Staff Research Fund of the University and partly by grants from the South African Council of Scientific and Industrial Research. The latter organization has also paid the 261 SMITHSONIAN . INSTITUTION JUN 9S 1868 262 ANNALS OF THE SOUTH AFRICAN MUSEUM salary of my research assistant for many years and Dr. John Croil Morgans made some very valuable collections by diving in False Bay during his tenure of the post. A full report of his diving survey has recently been published elsewhere (Morgans 1959). Earlier records of errant polychaets dredged off the Cape coasts will be found in McIntosh (1885 and 1904), Ehlers (1908 and 1913), Ramsay (1914), McIntosh (1925), Monro (1930 and 1936), Day (1934) and Treadwell (1943). The species recorded by Augener (1918 and 1931) from dredgings off South West Africa must be included for there can be no doubt that this area has a similar fauna to the Western Cape. Altogether these earlier workers recorded 74 species of errant polychaets from below tide-marks. The present paper contains new records, notes or full descriptions of 171 species. 61 of them are species known from the earlier dredgings, 42 are species previously known only from the shore, 34 species are new records for South Africa, 22 are new species and 12 are doubtful species or new varieties. The full total of species now known from dredgings around the Cape or South West Africa is 184 and for the sake of convenience, the 13 species recorded by earlier workers and not included in the systematic section of this paper is given below. Aphroditidae Eunoe assimilis McI.—MclIntosh 1925 Eunoe macrophthalma Mc.1.— McIntosh 1925 Lagisca hubrechtt Mcl.—Monro 1930 Macellicephala mirabilis Mc1.—MclIntosh 1905, 1925 Polynoe caput-leonis Mc1.— McIntosh 1925 Panthalis oerstedi var. capensis Mc1.—MclIntosh 1925 Leama hystricis Ehl.— McIntosh 1925 Hesionidae Magalia (=Syllidia) capensis (McI.) —MclIntosh 1925 ?Irmula spissipes Ehl.—Augener 1918 Syllidae Sphaerosyllis perspicax Ehl.—Augener 1918 Nereidae Nereis pelagica Linn.— Ramsay 1914 Eunicidae Eunice grubei Grav.—Ehlers 1908a Onuphis quadricuspis Sars— McIntosh 1925 THE POLYCHAET FAUNA OF SOUTH AFRICA 263 The other names which occur in the literature are synonyms of species described in this paper. Many of them are misidentifications of European species and one of my main tasks has been to eliminate these names from the South African faunistic lists. Unfortunately the descriptions are often incomplete and it has been necessary to examine the original material. During 1952 the C.S.1I.R. provided me with funds to visit the British Museum and examine the South African material housed there and to compare my own collections with the types. I wish to thank the Director of the British Museum and Mr. Norman Tebble of the Annelid Section for their kindness and help on this occasion and during a subsequent visit in 1958. I also wish to thank the Directors of the Scottish National Museum, the Swedish State Museum, the Hamburg Museum, the Berlin Museum and the U.S. National Museum for sending me South African specimens lodged in their respective institutions. One of the most important results of this sort of work was a general review of the genus Diopatra, which has been reported, not only from South Africa, but also from many other parts of the world under the name Diopatra neapolitana. An examination of material from the type locality (Naples) showed that the great majority of the records are misidentifications. A general discussion of Diopatra will be found on p. 338; the point which is stressed here is that similar work on difficult genera such as Eulalia, Exogone, Autolytus, and Lumbrineris suggests that a re-examination of type material or, where this is lacking, of material from the type locality is well worth while. It will lead to the solution of many anomalies of distribution and it now seems very probable that species of Polychaeta are by no means as widespread as has been supposed. Distribution patterns in this group will probably be found to follow the lines which Ekman (1953) has proposed for the bulk of the marine fauna. There is no doubt that many new species of errant polychaets await discovery in Cape waters. The University now has its own research vessel and almost every dredging brings up species new to the area or new to science and the deeper waters off the Eastern Cape have hardly been explored. For this reason no systematic key is included in this paper although one has been produced and is constantly being revised. Further work on errant species must wait until the bulk of the sedentary species has been described. This will form the subject of the next paper in this series. STATION Lists I must apologize for not giving the full station data below each species. It is realized that this causes a certain amount of inconvenience but space does not permit the full collection data to be repeated in this way. Full details for each dredging or diving station are given below, and under each species will be found only the station number with the number of specimens in brackets, e.g. AFR.728(1) under Aphrodite alta means that 1 specimen was obtained by R.S. Africana II and reference to the station list will give full details of date, 264 ANNALS OF THE SOUTH AFRICAN MUSEUM position, depth and type of bottom in the conventional abbreviated form. The sequence of the station lists is from the west or Atlantic coasts around the Cape towards the Eastern Province and Natal although some of the Africana (AFR) and Trawler (TRA) stations cover a wide range of coastline. In some cases the number of specimens obtained at a particular station was not accurately counted but was noted as common indicated as (c), fairly common (fc), or merely present (p). These letters in brackets are thus shown against the relevant dredging stations. The types described in this paper will be deposited in the South African Museum, Cape Town. 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 DATA LAMBERT’S BAY Drepcinc (LAM) No. Date Position Depth Bottom (Metres) LAM. 1 16.1.57 32.04.39/18.18.2E 15 Ss. LAM. 4 do. do. do. do. LAM. 5 L757 32.04.59/18.18.E 17 Sh.R. LAM. 6 17.1.57 32.04.79/18.18.5E 9 Serie LAM. 8 18.1.57 32.059/18.17.9E 23 S. Sh. R LAM. to 17.1.57 32.04.79/18.17.7E 23 S. Sh. LAM. 11 18.1.57 32.059/18.17.7E 29 S. Sh. LAM. 13 19.1.57 32.049/18.18.1E 18 R. LAM. 15 18.1.57 32.059/18.17.7E 17 S. Sh. R LAM. 16 Iigoliay 32.04.85/18.18.2E II Ss. LAM. 17 afta a) 32.05.39/18.17.4E 23 S. LAM. 18 18.1.57 32.04.859/18.17.8E 17 R: LAM. 19 do. do. do. do. LAM. 22 17.1.57 32.07.59/18.17.6E 20 S. R. LAM. 23 17.1.57 32.04.15/18.18.6E 15 S. Sh. LAM. 24 16.1.57 32.04.68/18.18.2E 17 R. LAM. 25 1 seis 32.04.25/18.18.4E 8 S. Sh. R. LAM. 26 18.1.57 32.04.99/18.17.5E 27 S. Sh. R LAM. 27 16.1.57 32.04.19/18.18.4E 16 R. LAM. 31 19.1.57 32.05.19/18.17.7E 20 R. LAM. 33 19.1.57 32.05.29/18.17.5E Plankton sample. LAM. 35 19.1.57 32.05.59/18.17.7E 27°5 Sh. R. LAM. 38 19.1.57 32.05.49/18.17.7E 27 S. sh: LAM. 39 19.1.57 32.05.49/18.17.6E 30 S. Sh. LAM. 40 19.1.57 32.05.59/18.17.6E 28 S. Sh. LAM. 41 21.1.5 7 32.059/18.17.7E 20 S. Sh. LAM. 43 21.1.57 32.04.99/18.18.2E 13°5 Sis LAM. 44 21.1.57 32.04.75/18.17.6E 20 R. LAM. 45 21.1.57 32.059/18.18.2E 8 DB. Ike LAM. 47 21.1.57 32.04.49/18.17.7E 23 R. LAM. 48 22.13.57 32.045/18.17.9E B27 S. Sh. LAM. 49 21.1.57 32.04.85/18.18.1E 10 S. R. LAM. 51 23.1.5 7 32.08.59/18.17.7E 16°5 ae LAMBERT’s Bay Drepcinc (LAM) No. - Date 21.1.57 23.1.57 22.1.57 23.1.57 23.1.57 23-1.57 23.1.57 23.1.57 23.1.57 23.1.57 19.1.57 Position 32.04.79/18.18.2E 32.099/18.17.8E 32.04.25/18.17.7E 32.115/18.18.1E 32.109/18.18.1E 32.095/18.18E 32,02S/18.18E 32.12S/18.17.9E 32.01.59/18.18E 32.01.59/18.17.8E 32.05.59/18.17.7E SALDANHA Bay Drepcinc (SB) No. SB.113 SB.114 SB.115 SB.116 SB.117 SB.118 SB.119 SB.120 SB.121 SB.122 SB.125 SB.127 SB.129 SB.130 SB.132 SB.135 SB.136 SB.143 SB.144 SB.145 SB.173 SB.175 SB.177 SB.179 SB.180 SB.181 SB.183 SB.184 SB.189 SB.193 SB.195 SB.197 SB.199 SB.202 SB.203 SB.205 SB.207 SB.208 Date Position 33.00.75/17.59.8E 33.00.459/17.57.5E do. 33.00.1S/17.59.2E 33.00.39/17.58.5E 33.01.59/17.58E 33.02.85/18.01.2E 33.03.45/18.01.8E 33.059/18.01.4E 33.04.95/18.00.4E 33.01.49/17.57-7E 33.04.6S/17.59.8E 33.04.59/18.00E 33.04.6S/18.00.6E 33-045/17.59.3E 33.035/17.58.6E 33.035/18.00.5E 33.05.15/18.01.2E 33.05.39/18.01E 33.04.85/18.00.5E 33.05.19/18.01.5E 33.02.85/18.00.6E 33.035/18.00.gE 33.03.6S/18.00.4E 33.03.59/17.58.5E 33.01.65/17.59.3E 33.02.59/17.58.7E 33.01.59/17.58.8E 33.01.19/18.00.3E 33.00.79/17.58.4E 33.03.59/17.59.2E 33.04.45/17.50.4E 33.01.79/18.01.4E 33-03-59/17.57-5E 33-05.59/17.55-5E 33.03.6S/17.56.4E 33.02.59/17.57.5E 33.01.99/17.56.3E Dept (Metres) Depth (Metres) ~ [o) Qu La] OND HUNTOW OANA OM -_ m ND Oe WN on THE POLYCHAET FAUNA OF SOUTH AFRICA - Bottom DDD win N in Bie Bottom wn do. APnnn cn re) DnnAY Ss 265 266 ANNALS OF THE SOUTH AFRICAN MUSEUM LANGEBAAN LAGOON DREDGING (LB) No. LB.155 LB.158 LB.159 LB.160 LB.161 LB.169 LB.190 LB.191 LB.239 LB.299 LB.300 LB.323 LB.363 LB.364 LB.38o0 LB.382 LB.391 LB.456 LB.472 LB.496 Date 15.7.46 do. do. TasLe Bay Drepcinc (TB) No. TB.3o1 TB.302 TB.303 TB.304 TB.305 TB.306 TB.307 TB.308 TB.309 TB.310 TB.312 TB.313 TB.314 Date 4.8.46 11.2.4.7 do. 15.12.57 TB.315-333 do. Position Depth Bottom (Metres) 33.07.15/18.02.4E 2 f. S. 33.099/18.04.2E 4 do. 33.105/18.04.8E 4°5 f. S. M. 33.06.48/18.01.9E 3 S. Sh. 33.05.6S/18.00.8E 5 Sh. R. 33.10.59/18.03.8E 2 i. 0. 33.11.39/18.05.5E 0-2 f. S. M. do. do. do. 33.079/18.02.7E 2 f. S. 33.06.85/18.01E 2°5 S. Sh. 33.07.69/18.02.3E a Ss. 33.06.8S/18.01E 2°5 S. Sh 33.07.15/18.02.7E 4 S 33.05.99/18.01.7E 5 S. Sh. 33.06.3S/18.01E 4°5 do 33.05.59/18.01.6E 12°5 do. 33.07.99/18.02.1E 2°5 f. S. 33.07.79/18.02.4E 4 S. Sh. 33.07.4.9/18.02.5E 3°5 do. 33.05.79/18.01E 5 Gr. R Position Depth Bottom (Metres) 33.49.59/18.27.5E 12°8 S. Sh. R. 33.48.35/18.24E II S. St. 33.4.7.59/18.24.3E 19°5 S. Sh. St. 33.485/18.24.3E 16 Sh. Gr. 33.52.79/18.29.7E 9 S. St. 33.50.19/18.27.8E V7 Sh. R. 33.50.39/18.28E 15°5 R. Sh. 33.51.25/18.27.3E 23 S.R. 33.52.79/18.26.8E 20°5 R. 33.485/18.21E 16°5 Ss. 34.059/18.21E 11 do. 33.529/18.28E 1 S. Sh. 33.48.6S/18.24.6E 15 Sh. St. do. do. do. Suips’ HuLts AND EXPERIMENTAL PLATES SUBMERGED IN TABLE Bay Docks (SH) Code No. SH. 69 SH. 71 SH. 74 SH.134 SH.168 SH.204 SH.277 SH.324 SH.366 SH.376 SH.393 SH.400 Date 12.11.46 17.4.46 18.4.46 21.1.47 1.4.47 27-5-47 4-9-47 6.2.48 2.12.48 26.1.49 16.3.49 do. Remarks. Norfolk from India and east coast of Africa. Natal—from India and east coast of Africa. Windward—local wooden yacht 9 months in water. Empire Liddell—local ship 1 month in water. Experimental plate submerged for 120 days. Experimental plate submerged for 175 days. Experimental plate submerged for 275 days. Barge working in Table Bay. Wooden Teredo trap submerged for ? days. Experimental plate submerged for 96 days. Wooden frame submerged for 94 days. Experimental plate submerged for 7 months 2 days. THE POLYCHAET FAUNA OF SOUTH AFRICA 267 Suip’s HuLts AND EXPERIMENTAL PLATES SUBMERGED IN TABLE Bay Docks (S.H.) Code No. SH.415 SH.427 SH.428 SH.430 Date 14.2.49 15.12.49 29-7-53 28.7.58 Remarks. Wooden frame submerged for over 1 year. Experimental barge 14 months in water. Scraping from submerged caisson. Leeukop—local wooden trawler. 84 months in water. West Coast Drepcinc (WCD) No. WCD. 3 WCD. 5 WCD. 8 WCD.13 WCD.15, WCD.19 WCD.21 WCD.23 WCD.26 WCD.28 Date 25.2.59 do. 24.3-59 do. 24-459 29-4-59 30.4.59 1.5.59 do. 2.5.59 Position 34.09.8S/18.16.5E 34.099/18.14.8E 34.09.359/18.17.5E 34.09.45/18.16.5E 33-04.39/17-54-7E 33.05.68/17.54.5E 33-04.59/17.55-5E 33.00.48/17.53.7E 33.06.5S/17.55.4E 33.05.59/17.50.4E FatsE Bay Drepcinc (FB) No. FB.301 FB.302 FB.305 FB.306 FB.307 FB.308 FB.309 FB.310 FB.311 FB.312 FB.313 FB.314 FB.316 FB.317 FB.318 FB.319 FB.320 FB.321 FB.322 FB.323 FB.324 FB.325 FB.326 FB.327 ‘FB.328 FB.329 FB.330 FB.331 FB.332 FB.333 FB.334 Date 8.7.46 8.9.46 12.11.46 24.11.46 22.2.47 do. do. 21.447 28.4.47 Position 34.085/18.27E 34.08.59/18.26.5E 34.08S/18.27E 34.09.39/18.27.7E 34.07.59/18.31E 34.08S/18.31.5E 34.07.59/18.29.3E 34.085/18.32E 34.109/18.27.8E 34.09.59/18.27E 34.085/18.29E 34.098/18.27.7E 34.099/18.28E 34.09.59/18.28.3E 34.10.25/18.27E 34.09.25/18.26.8E 34.085/18.29.6E 34.085/18.31E 34.079/18.29E 34.109/18.29.5E 34.099/18.29.5E 34.08.59/18.27E 34.08.99/18.27.4E 34.09.6S/18.26.6E 34.09.85/18.26.1E 34..10.25/18.26.2E 34.10.19/18.26.1E 34.109/18.26.1E 34.09.39/18.26.4E Off St. James 34.07.59/18.29E Depth (Metres) 78 Bottom Bottom NNNN nF nn 49 Tm Fs i ee os pe > 9 ee) a S. lithoth. do. (e) ANBANNNA Eee Cx f a n 268 FAtsE Bay Drepcinc (FAL) No. Date 22.2.52 do. Windmill B Position 34.09.59/18.35E 34.08.35/18.35.3E S.E. Oatland Pt. do. do. 34.135/18.28E (plankton) do. do. 34.139/18.29E do. 34.139/18.29E 34.125/18.29E do 34.059/18.44E 34.09.6S/18.49.2E do. 34.09.39/18.49.6E do. 34.09.49/18.50.8E do. 34.09.45/18.50.4E 34.17.59/18.49.2E 34.17.39/18.48.7E 34.17.25/18.49.4E do. 34.16.59/18.49.5E do. do. 34.10.65/18.4.7.3E 34.08.35/18.35.3E 34.09.49/18.50.4E 34.135/18.29E 34.10.65/18.47.3E 34.09.49/18.51.7E 34.09.39/18.51E do. do. Simons Bay Glencairn (diving) do. do. do. Oatland Pt. (diving) do. do. do. Gordons Bay Quay (diving) Oatland Pt. (diving) do. do. do. do. do. do. each (diving) ANNALS OF THE SOUTH AFRICAN MUSEUM Depth Bottom (Metres) 35 S. Sh 24. R. 8-9 R.S do. do. do. do. 11-12 S.Sh do. do. do do. 15-21 §.Sh.R do. do. 28 S. Sh 33-36 Sh do. do. 7 S.R Q21°5 do. do. do. 18 R.S do do. 8 R. do. do. 12 S.R 22 S. 37-38 S.Sh 16-19 iR. do. do. 14-17. do do do. do do. 36 S. Gr. R 24. R. 12 S.R. 15-21 S.Sh.R. 36 S. Gr. R. 1+5-2°5 5S. 8-12 R.S 45 R. do. do. do. do. 23°5 Lithoth oF R. 2-4 do. 2-7 do. do do. I-2 do. 0-2 do do. do. do. do. o-4 do 0-5 do fe) do 4°5-5°5 S.R Oo. oO 0-3 R do. do do. do FatsE Bay Drepcinc (FAL) Date 17.11.53 do. 21.11.53 Position Oatland Pt. (diving) 34.22.18/18.35.2E do. 34.12.85/18.36.5E 34.07.19/18.35.6E 34.17.6S/18.39.2E do. 34.09.99/18.42.4E 34.06.8S/18.40.3E 34.07.15/18.35.6E 34.12.45/18.43.5E do 34.079/18.32.5E 34.13.99/18.31.6E do. 34.10.59/18.32.4E 34.20.39/18.31.8E do. 34.17.49/18.31.4E 34.15.39/18.44.8E 34.18.25/18.45.8E 34.21.19/18.46.8E 34.22.79/18.43.1E 34.20.65/18.39.4E 34.23.79/18.40.9E 34.18.59/18.34.2E do 34.22.59/18.37.3E Oatland Pt. (diving) do. 34.079/18.32.5E 34.07.15/18.35.6E 34.079/18.32.5E 34.22.79/18.43.1E 34.23.79/18.40E Oatland Pt. (diving) do. do. Noah’s Ark (diving) do Windmill Beach (diving) Roman Rock (diving) do. 34.239/18.36E (S.A. Museum material) Depth Bottom (Metres) 0-2 S.R. do. do. do. do. 24 do. 4-6°5 do. do. do. 4 do. 25 do. 73 S. Sh. Gr do. do. 46 S. Sh. (larva trap, night) 62 S. Sh. do. do. 36°5 Soak 29 S. Sh. oy S. Sh. R 42 S2 ik: do. do. 18 S. Sh. R 40 co. S. Sh do. do. 36 S. 64 do. do. do. 49 S. R. 48 S. Sh. Gr. R 55 R. 64 do. 79 gn. M. 82 S. gn. M 88 gn. M. 64 S. M. Sh do. do. 80 S. Sh. 0-3 R. 4 do. 18 S. Sh. R Ba Tas, do. 18 do. 79 gn. M 88 do. 4575 RK: 2°5-5°5 do. 10°5 S. Sh. 14 S. Sh. M. 11-14 R. 4-5 do. 14-17 do. do. do. 12-14 do. do. S. Sh. do. do. 35 R. THE POLYCHAET FAUNA OF SOUTH AFRICA 270 ANNALS OF THE SOUTH AFRICAN MUSEUM Farsz Bay Drepcine (FAL) No. Date Position Depth Bottom (Metres) FAL.303 9.10.02 34.265/18.37E 73 R. (S.A. Museum material) FAL.304 15.10.1897 34.09.459/18.49.5E 18 do. (S.A. Museum material) FAL.306 11.9.53 34.22.15/18.35.2E 73 S. Sh. Gr FAL.314 19.4.55 34.09.65/18.27.4E 26 Ss. FAL.324 6.10.1898 34.185/18. ?E ? ? (S.A. Museum material) FAL.327 10.9.57 Kalk Bay 3-4 R. FAL.328 31.1.59 34.199/18.34.6E 40 S. Sh. FAL.334 do. 34.159/18.36E 51 co. S. Sh FAL.338 do. 34.139/18.35E 44 Ss. FAL.341 do. 34.118/18.35.5E do. f. br. S FAL.345 do. 34.11S8/18.33.5E 38 f. S. FAL.347 do. 34.10.85/18.31E 35 do. FAL.349 do. 34.08.75/18.31.6E 27 w. 9. FAL.352 1.2.59 34.23.39/18.40.3E 88 gn. M. FAL.355 24.2.59 34..23.39/18.39.4E 97 S.R. FAL.357 do. 34.18.85/18.39E 73 co. S. Sh FAL.359 do. 34.16.85/18.40.9E 62 S. Sh. FAL.365 252.59 34.09.25/18.46.6E 30 R. FAL.367 do. 34.11.15/18.46.9E 37 S.R. FAL.371 do. 34.12.65/18.46.7E 40 R. FAL.373 do. 34.15.19/18.44.8E 54 S. Sh. FAL.375 do. 34.16.88/18.42.8E 60 gn. S. Sh FAL.376 do. 34.18.75/18.37.2E 72 do. FAL.378 do. do. do. do MATERIAL From COMMERCIAL TRAWLERS (TRA) No. Date Position Depth Bottom (Metres) TRA. 20 5.5.46 33.48S/17.35E 311 gn. M. TRA. 21 4.9.46 34.259/18.10E 301 M. R. TRA. 25 8.4.48 34.305/20.54E 66 S. M. TRA. 27 21.7.48 34.48S/20.20E 67 do. TRA. 30 9.11.47 34.499/20.21E 86 M. R. TRA. 33 20.7.49 34.559/21.10E go S. R. TRA. 36 21.1.50 34.359/20.50E 73 M. St. TRA. 40 —.7.50 34.309/20.57E do. do. TRA. 41 26.7.51 34.315/20.50E 66 S. M. TRA. 43 ? 29.499/31.48E 770 M. TRA. 46 24.9.52 31.259/16.20E 366 gn. M. TRA. 48 do. 33.159/16.00E 415 M. S. TRA. 52 do. 32.125/16.38E 394 M. TRA. 54 28.11.52 34.409/21.35E 73 S. R. TRA. 55 do. do. do. do. TRA. 56 do. do. do. do. TRA. 58 26.11.52 34.285/21.45E 70 S. St. TRA. 62 = 25.11.52 34.309/21.15E 62 S. M. TRA. 63 3928.11.52 34.268/21.50E 64 S. M. R. TRA. 68 6.2.53 32.245/18.07E 69 gn. M. TRA. 69 do. (plankton) 32.455/18.00E 15 S. R. TRA. 70 do. 32.295/18.02E 27 M. eA a7 52.53 32.059/18.14E 66 R. S. THE POLYCHAET FAUNA OF SOUTH AFRICA 271 MATERIAL From COMMERCIAL TRAWLERS (TRA) No. . Date Position Depth Bottom (Metres) TRA. 73 3.2.53 32.068/16.37E 311 gn. M TRA. 74 52-53 32.059/17.52E 123 do. TRA. 75 do. do. do. do. TRA. 77 6.2.53 32.419/18.03E 27 S. M. TRA. 80 4.2.53 32.239/17.48E 143 gn. M. TRA. 84 = 13.11.51 32.379/18.17E 6 Ss. TRA. 85 22.3.53 32.445/18.02E 18 do. TRA. 86 233.53 32.485/17.58E 9 do. TRA. 88 do. 32.445/18.00E II do. TRA. 89 do. 32.459/18.03E 9 S.R. TRA. 91 15.7.53 33-519/25.50E 46 M. TRA. 93 —.1.54 35.035/21.50E 110 S.R. TRA. do. do. do. do. TRA.102 —.3.56 34.259/21.30E 55 S. R. Polyz TRA.104 6.8.56 34.31S/19.21E 22 Ss. TRA.106 do. 34.339/19.19E a7 do TRA.107 7.8.56 34.105/18.48E 29 f (surface plankton) TRA.108 6.8.56 34.339/19.19E 37 Ss. TRA. 110 8.9.56 34.199/18.32E 58 S. Sh. R TRA.112 do. 34.199/18.33E 60 S. R. TRA.113 do. 34.199/18.32E 58 do. TRA.114 do. 34.199/18.33E 62 do. TRA.I15 = 29.11.56 34.159/18.43E 54 Ss. TRA.116 do. 34.115/18.39E 44 do. TRA. 121 25.1.57 34.125/18.44E 37 S. R. TRA.122 do. 34.13.59/18.45E 44 S. TRA.123 do. 34.125/18.45E 40 S. R. TRA.127 22.2.7 34.199/18.30E 51 do. TRA.132 —.2.57 34.205/18.30E 55 Phyllochaetopterus tubes S. Sh. R. TRA.133 do. do. do. do. TRA.135 23.2.57 34.195/18.30E 52 do. TRA.143 27.3.57 34.189/18.31E 51 do. TRA.151 6.3.58 34.51S/19.55E 22 R. TRA.152 do. do. do. do. Drepoinc By S.A. FisHERIES RESEARCH VESSEL Africana II (AFR) No. Date Positton Depth Bottom (Metres) AFR.689 ? 32.36.68/16.44E 391 gn. M. AFR.691 8.5.4.7 32.385/16.52E 347 Cl. S. AFR.707 26.5.47 31.409/16.55E 287 d. gn. M AFR.718 19.6.47 32.099/18.06E 108 do. AFR.723-5-7 10.8.47 31.309/17.00E 366 ? AFR.728 15.8.47 31.149/16.36E 272 Polyz. R AFR.730 do. 31.30S/16.03E 459 y. Cl S.R AFR.736 17.8.47 30.429/15.59E 201 co. gn. S. Sh AFR.761 10.9.4.7 30.139/15.18E 260 Gr. S. R. AFR.773 14.9.47 28.525/14.50E 194 Cl. S. AFR.775 15.09.47 29.16S/14.48E 238 Cl. S.R AFR.783 24.9.4.7 32.439/17.31E 222 S. M. AFR.789 28.9.47 33.059/17.27E 408 bl. S.R 272 ANNALS OF THE SOUTH AFRICAN MUSEUM Drepcinc BY S.A. FIsHERIES RESEARCH VESSEL Africana II (AFR) No. AFR.790 AFR.791 AFR.801 AFR.830 AFR.831 AFR.835 AFR.842 AFR.882 AFR.945 AFR.9g50 AFR.957 AFR.967 AFR.994 AFR.995 AFR.1224 AFR.1335 AFR.1529 AFR.1532 AFR.1535 AFR.1544 AFR.1545 AFR.1554 AFR.1576 AFR.1578 AFR.1579 AFR.1581 Date 28.9.4.7 4.10.47 7.10.47 19.11.47 do. 20.11.47 25.11.47 10.2.48 19.3.48 20.3.48 22.3.48 23.3.48 19.4.48 do. 15.10.48 13.11.48 4.6.49 do. 9-7-49 23-7-49 do. 28.7.49 9-9-49 do. do. 10.9.49 Position 33.125/17.40E 32.419/17.18E 32.348/17.52E 32.125/18.42E 35-159/18.39E ?35.099/19.02E 34.359/19.18E 34.399/18.42E 36.255/21.08E 36.44S/21.18E 35.139/21.19E 35-079/20.49E 34.359/21.26E 34.299/21.26E 26.345/15.04E 25.519/14.51E 32.409/17.43E 33.125/17.58E 29.099/16.45E 29.179/16.42E 29.099/16.37E 32.059/18.17E 32.285/18.06E 32.305/17.49E 32.259/17.42E 32.225/17.59E Mosse, Bay Drepcinc (MB) No. Date Position 34.099/22.07.1E 34.04.25/22.13.8E do. 34.119/22.10.1E 34.08.59/22.07.2E 34.08.85/22.07.3E 34.11.19/22.09.9E 34.08.35/22.09.4E 34.09.39/22.10E 34.10.19/22.07.8E 34.10.19/22.08E do. 34.08.59/22.08.8E 34.11.39/22.10E 34.115/22.09.gE 34.10.75/22.09.6E do. 34.04.39/22.13.5E 34.04.15/22.13.9E 34.04.35/22.14.2E 34.04.85/22.13.1E do 34.09.15/22.07.3E Depth (Metres) Io 19 do. 3 < Ske ~ PP Za. [e) Q. aa N ks S me nm: mn nn! —N TN a re) BRB mw ae ye) THE POLYCHAET FAUNA OF SOUTH AFRICA Mosset Bay Drepcinc (MB) Date Position 19.1.56 34.08.6S/22.07.3E do. 34.09.1S/22.07.2E do. 34.08.75/22.07.4E 20.1.56 34.11.35/22.06.3E do. do. do. 34.059/22.11.8E do. 34.06.25/22.10.9E 21.1.56 34.11.48/22.10.1E 17.1.56 34.11.35/22.10E do. 34.11S/22.09.9E 18.1.56 34.04.85/22.13.1E ALGoA Bay Drepcinc (LIZ) No. No. Date Position 54-54 33-55-79/25-37-2E do. do. 5-4-54 33.56.1S/25.40E 6.4.54 33.58.1S/25.38.9E do. do. do. 33.58.25/25.38.8E 7-4-54 33-58.45/25.40.5E do. 33.58.59/25.42E 8.4.54 33-585/25.43E 11.4.54 34.00.45/25.44.5E do. do. do. 34.00.85/25.42.4E do. do. Drepcinc (SCD) Date Position 15.4.58 35.279/20.10E 18.4.58 34.205/24.40E do. do. 19.4.58 34.159/25.05E 20.4.58 32.525/28.12.5E 26.5.58 34.07.39/23.23.8E do. 234.465/23.27E 24.5.58 34.02.59/25.46.5E 23.5.58 33.4.79/26.04E 22.5.58 33.38.65/26.54.7E 21.5.58 33.039/27.50.2E 19.5.58 32.15.29/28.57.7E do. 31.38.85/29.34.4E 20.8.58 34.018/25.45.5E 19.8.58 33.379/26.56.6E 16.8.58 33.025/27.56.2E 14.8.58 31.57.29/29.36E 5-7-59 33-315/27.14.5E 15.7-59 31.4.1.79/29.33.5E Depth (Metres) Depth (Metres) Bottom pp a TN FERRED fe) (e) nn ep) = I Bottom Bottom BY provers ge ele nin ne D NM 5 glutinous br. M. grass. 273 and 274 ANNALS OF THE SOUTH AFRICAN MUSEUM SoutH Coast Drepcinc (SCD) No. Date Position Depth Bottom (Metres) SCD. 74 16.7.59 32.335/28.38E 55 S. M. SCD. 78 do. 32.379/28.31E 49 br. S. SCD. 80 do. 32.439/28.28E 58 St. Sh. SCD. 82 17.7.59 27.549/33.03E 51 br. S. Sh. SCD. 89 do. 33.039/27.55E 27 R. SCD. 94, 96 20.7.59 34.215/25.41E 110 Sh. SCD. 99 21.7.59 34.339/24.01E 130 R. SCD.100 do. do. do. do. SCD.103 22.7.59 35.079/22.15E 120 mS, SCD.105 23.7.59 34.339/21.28E 67 co. S. br. Sh. SCD.106 do. 34..359/21.10E 67 S. M. SCD.109 do. 34.359/21.11E 75 co. S. Sh. St. Family APHRODITIDAE Subfamily HERMIONINAE Aphrodita alta Kinberg 1855 Aphrodite alta. Kinberg 1857, p. 2, pl. 1, fig. 1 a-g. Monro 1930, p. 36, fig. 5 a-l. Aphrodita A near alta McIntosh 1925, p. 18. Records: AFR.728(1), 835(1). Notes: In this species the dorsal setae do not project through the felt, the eyes lack pigment, the ceratophore of the median antenna is short and stout but bears a very fine ceratostyle which is three-quarters the length of the prosto- mium. The ventral setae have curved, bearded tips. Monro (1930) states that the median antenna is short and stout but this does not agree with Kinberg’s pl. 1, fig. 15, and it is probable that Monro was describing the ceratophore from which the ceratostyle had fallen. Again Monro’s fig. 5a shows the stout dorsal setae as quite smooth. In mine they are covered with minute hairs. McIntosh’s specimens in the British Museum have been checked as identical with my own. Subfamily PoLyNOINAE Harmothoe aequiseta (Kinberg) 1855 ?Lagisca extenuata Ehlers 1913, p. 446. Harmothoe aequiseta. Augener 1918, p. 137. Day 1953, p. 400. Harmothoe crosetensis (non McIntosh) Monro 1930, p. 57 (partim). Records: SB.118(1); TRA.71(1); FAL.16(1), 30(1), 44(5), 51(p), 56(4), 69(p), 80(p), 223(p); MB. 20(1), 57(1), 77(2); KNY.6(1), 11(1), 21(2). Notes: The specimens reported by Monro (1930) from Simonstown as as H. crosetensis are definitely H. aequiseta. McIntosh described H. crosetensis as having fringed elytra but an examination of his type in the British Museum THE POLYCHAET FAUNA OF SOUTH AFRICA 275 shows that the surface of the elytron is densely covered with soft papillae right to the edge but the margins are not fringed. Monro’s specimen of H. crosetensis from the South Shetland Islands agrees with MclIntosh’s type. ‘The specimen recorded by Ehlers (1913) as Lagisca extenuata was a juvenile and should probably be referred to H. aequiseta. Harmothoe africana Augener 1918 Harmothoe africana Augener 1918, p. 139, pl. 2, figs. 15-19, text fig. 6. Records: FB.316(1). Notes: Both H. africana and H. goreénsis (recorded below) are very closely allied to H. aequiseta if they are not mere varieties of the latter. In H. aequiseta the larger tubercles on the elytra look like straight dark thorns. In H. africana the larger tubercles are almost cylindrical and end in 2-4 points but there is considerable variation among the smaller ones. Harmothoe goreénsis Augener 1918 Harmothoe goreénsis Augener 1918, p. 142, pl. 2, figs. 4-6; pl. 3, fig. 42, text-fig. 7. Records: LAM.8(1), 22(3), 31(2), 44(1), 47(1), 57(1); SB.180(1), 183(3), 184(3), 189(1), 207(2); T.B.go1(1), 306(1); WCD.8(1), 13(1); FAL.8(1), 58(P), 113(1), 134(5), 144(5), 149(2), 156(6), 216(1), 238(1), 280(4), 338(1), 341(1), 359(1), 375(1); MB.9(1), 13(2), 16(1), 27(3), 41(7), 49(12), 53(12), 56(2), 67(10), 69(1), 74(2); 77(6), 85(2); LIZ.2(6), g(x), 18(4), 35(4), SCD. 40(1), 54(5), 58(5), 74(1)- Notes: This is a small species first recorded from shallow waters off Angola and Senegal. This is the first record from South Africa. It differs from H. aequiseta in having short, blunt, cylindrical or crown-shaped tubercles on the elytra instead of long sharp ones. Harmothoe fraser-thomsont McIntosh 1897 Harmothoe fraser-thomsoni. Fauvel 1923, p. 68, fig. 25 a—-e. Day 1953, p. 400. Records: SB.129(1), 132(1); LB.456(2); LIZ.6(1); SCD.58(1). Notes: As stated earlier the South African material is a little different from typical European forms. The elytra have more numerous and crowded tumid papillae and the notosetae have long naked tips instead of short tips. Harmothoe gilchristi n.sp. (Fig. 1 a) Records: AFR.835(2); FAL.355(1); SCD.22(1). Description: The holotype is the single specimen from SCD.22. It is complete and measures 16 mm. by 3:5 mm. with 38 segments. The body is pale but there are brown markings on the elytra, antennae and dorsal cirri. 276 ANNALS OF THE SOUTH AFRICAN MUSEUM Fic. 1. Harmothoe gilchristi: a elytron; 6 details of tubercles and papillae; ¢ posterior view of 8th parapodium; d head; e, e! notoseta; f, f1 neuroseta. Harmothoe agulhana: g posterior view of 8th parapodium; h, h1 notoseta; 7, 71 neuroseta; j head; k elytron; / details of ‘tubercles’. aL THE POLYCHAET FAUNA OF SOUTH AFRICA 277 The prostomium (fig. 1d) is almost square, slightly broader than long with very sharp prostomial peaks. The anterior pair of eyes are large and placed well back on the sides of the prostomium. The three antennae arise from short swollen cirrophores and all are brown, densely clad with long papillae and tapering. The median antenna of the holotype is missing but a specimen from AFR.835 shows that it is twice the length of the prostomium. The laterals on account of the very large prostomial peaks are markedly ventral in origin and each is well tapered and three-quarters the prostomial length. The body is slightly tapered posteriorly and the last 6 segments lack elytra. The dorsal cirri are tapered, densely papillose and exceed the length of the neuropodia but not the neurosetae. Each has two dark bands of pigment. The elytra (fig. 1a) are large, circular, mottled with purplish-brown and cover the back except for the last few segments. Each has a very small fringe of unicellular papillae on the external margin and the surface (fig. 1b) is densely covered with short cylindrical or bollard-like tubercles plus a few elongate papillae. There are no large posterior vesicles. . The notopodium (fig. 1c) is well developed with a radiating tuft of numerous notosetae (fig. 1e). Each is stout and strongly serrated to the blunt grooved tip. The neuropodium is well developed with a pointed presetal lobe containing the aciculum and a truncate postsetal lip. The neurosetae (fig. 1f) are bidentate and have fairly long blades with about 15 rows of well-developed spinules and long naked tips (fig. 1f). The terminal tooth is broad with a curved point and the secondary tooth which lies in line with the shaft is exceedingly long and slender. Its length is almost twice the width of the shaft at the origin of the tooth and three-quarters the length of the terminal tooth. H. gilchristi comes fairly close to H. goreénsis Augener but the prostomial peaks are better developed, the tubercles on the elytra are larger and most of them are swollen distally instead of being sculptured. Again, the notosetae have shorter, grooved tips which are rounded not pointed. The secondary tooth of the neuroseta is also distinctly longer. Type locality: Agulhas bank. Harmothoe agulhana n.sp. (Fig. 1 g—l) Records: ?AFR.7070(1); FAL.365(1); MB.67(1); LIZ.25(1). Description: The holotype is the single complete specimen LIZ.25 dredged in Algoa Bay. It is 12 mm. long by 2 mm. excluding setae and has 36 setigers. The body is narrowly oblong, hardly tapered posteriorly and is pale in alcohol with a faint network of brown on the exposed parts of the elytra. The prostomium (fig. 1j) is about as broad as long with poorly marked frontal peaks. The eyes are rather small and the anterior pair is laterally situated almost half-way back. The tapered median antenna is as long as the prostomium but the laterals are very short and stumpy, barely a quarter of 278 ANNALS OF THE SOUTH AFRICAN MUSEUM the prostomial length. All antennae and cirri are sparsely beset with small papillae. ' The dorsal cirri are tapered and on all except the last few segments tee are shorter than the neurosetae. The ventral cirri are very small and distinctly tapered. The 15 pairs of elytra cover the body except for the last two segments. Individual elytra (fig. 1k) are large and oval and so thin that the edges tend to crumple. The colour is generally pale but there is a faint speckling or network of brown pigment over most of the surface. There is a patch of small rounded chitinous tubercles on the antero-medial margin, and a scattering of similar tubercles (fig. 1/) over the surface which do not have chitin-thickened walls and for this reason are not very obvious. The notopodium (fig. 1g) is normally developed and the notosetae (fig. 1) are fairly numerous but rather short and stout and strongly serrated to their blunt tips (fig. 1h!). The neuropodium is a truncate cone with a small presetal projection covering the end of the aciculum. The neurosetae (fig. 12) have blades of normal length with about 10 rows of spinules. Apart from 2-3 superior neurosetae which are unidentate, the tips (fig. 17!) are bidentate with a strong hooked terminal tooth and a fine secondary tooth. A fragment of what may be the same species was obtained from station AFR.707. While generally similar, there are signs of prostomial peaks, the notosetae are very stout and have grooved tips and the neurosetae have more rows of spinules, so that both types of setae are very similar to those of H. gilchristi. But the short lateral antennae and the elytra are very like those of H. agulhana with only tiny hemispherical tubercles and entire margins. This species appears to be related to H. ljungmam. The notosetae are similar and the elytra are not very different though the tubercles are not so well developed. The neurosetae, however, are quite distinct. Harmothoe corralophila n. sp. (Fig. 2 a) Records : AFR.g50(1) ; WCD.3(1), 13(3); FAL.378(1) ; SCD.100(1), 103(1). Description: The type was selected from WCD.13. It is 15 mm. long with 37 segments. It is quite white in alcohol with rather glassy setae. The prostomium is bilobed and broader than long with obvious frontal peaks and rather large eyes. The anterior pair are half-way back and much wider apart than the posterior pair. The median antenna is twice as long as the prostomium and is mounted on a stout ceratophore from which an obvious ridge extends back along the dorsal surface of the prostomium for half of its length. The lateral antennae are tapered, markedly ventral in origin and equal to the prostomial length. All antennae and cirri are smooth. The palps are fairly large. _ The elytra cover the whole length of the body. All of them are white and have entire margins without a sign of a fringe but the surface of the first few THE POLYCHAET FAUNA OF SOUTH AFRICA 279 Fic. 2. Harmothoe corralophila: a first elytron; a’ chitinous tubercle; 5 posterior elytron; 5° soft papilla; c posterior view of parapodium; d notoseta; ¢ neuroseta; el tip of superior neuroseta; é* tip of middle neuroseta. Malmgrenia purpurea: f neuroseta; f 1 tip of neuroseta; g head; h notoseta; h! tip of notoseta; Jj elytron; k posterior view of parapodium. 280 ANNALS OF THE SOUTH AFRICAN MUSEUM differs from more posterior pairs which at first sight appear to be quite smooth. The first pair (fig. 2a) are smaller than the rest, circular in outline and most of the surface is thickly chitinized and densely beset with numerous highly chitinized conical tubercles (fig. 2a') plus a few indistinct soft cylindrical papillae. On some specimens there is only a narrow marginal belt around the elytron which is not thickly chitinized and free from tubercles but in others there is a relatively broad naked margin. The second pair of elytra has a more restricted area of thick chitin and tubercles and a larger area which is naked apart from the soft digitiform papillae. In the third and fourth pairs there are even smaller chitinized and tuberculate patches. All succeeding elytra (fig. 2b) are thin, entirely devoid of tubercles and have only a few indistinct panies (fig. 251) on the otherwise smooth surface. The dorsal cirri are smooth and tapered. The first few barely reach the ends of the neurosetae but further back they are longer. The notopodia are rather small and the notosetae (fig. 2d) are not very numerous but each is stout and well developed with widely spaced rows of large serrations or cusps preceding the long naked tip. Most of the notosetae are sharply pointed but some of the longer ones have grooved tips. The neuropodium (fig. 2c) has a pointed presetal lip and a fan of long, clear, rather slender setae (fig. 2e). The spinules are poorly marked. Although there are actually 20 rows of spinules which is more than usual, the inferior neurosetae may at first appear to be smooth because the spinules lie so close along the blade. The terminal tooth is always well developed but the secondary tooth is very variable. In superior neurosetae (fig. 2¢!) it is shorter but almost as broad as the terminal one so that the tip of the blade appears to be bifid. In middle neurosetae (fig. 2e%) the base of the secondary tooth is stout but the point is exceedingly fine and often bent or broken leaving a stout stump. No truly unidentate setae have been seen. The first specimen obtained (AFR.g50) has lost the anterior elytra and was provisionally identified as Harmothoe joubint Fauvel 1914 which it resembles in many respects apart from the anterior elytra and the tips of the neurosetae. Later several stylasterid corals (Allopora bithalamus) were obtained from SCD.100 with galls in the shape of open tunnels on their sides. H. corralophila was found in these galls and is obviously the cause of their formation. Since both ends of the tunnel are open it is probable that the worm can move in or out at will and uses the tunnel for protection. Harmothoe lunulata (Delle Chiaje) 1841 Harmothoe lunulata. Fauvel 1923, p. 70, fig. 26. Records: FAL.285(1). Notes: This is a new record for South Africa. THE POLYCHAET FAUNA OF SOUTH AFRICA 281 Harmothoe saldanha Day 1953 Harnothoe saldanha Day 1953, p. 401, fig. 1 a—d. Records: SB.202(1); SCD.32(1). Harmothoe (Lagisca) waahli (Kinberg) 1855 Harmothoe waahli Monro 1933, p. 489, figs. 1-3. Records: SB.121(4), 127(2); LB.155(2), 161(4); WCD.19(1); FB.302(1), g11(1), 319(1); FAL.43(2), 51(1), 58(1), 80(1), 235(1), 249(1), 269(1), 367(1); MB.16(1); LIZ.6(3). Notes: Several authors have recorded this species under the generic name Harmothoe but it should be noted that the last 12 segments are narrowed and not covered by elytra. However I do not feel that this character merits full generic distinction. Scalisetosus pellucidus (Ehlers) 1864 Scalisetosus pellucidus. Fauvel 1923, p. 74, fig. 27 a-—f. Records: LAM.1(1), 8(3), 25(1), 31(2), 63(1); SB.118(1), 119(1), 179(1)5 LB.161(1); TRA.71(1), 102(1); FB.302(1); FAL.8(p), 30(1), 95(P)> 113(p), 122(1), 134(4), 166(1), 184(1), 275(1), 327(1), 338(1), 367(2); MB.4o(1), 74(1), 86(1); KNY.6(1), go(1); LIZ.g(1), 29(1), 35(1); SCD.9(2), 22(2), 100(1). Antinoe lactea Day 1953 Antinoe lactea Day 1953, p. 403, fig. 2 a-g. Records: 1.B.299(3), 300(1), 364(5). Notes: It is surprising that this species has never been recorded outside Langebaan Lagoon. Malmgrenia purpurea n. sp. (Fig. 2 f-k) Records: WCD.5(1); FAL.229(1), 359(1), 375(1).- Diagnosis: A purple species with very stout antennae and cirri. Description: The type material consists of two complete but broken specimens dredged with Spatangus capensis from stations FAL.359 and FAL.375. The larger specimen measures 17 mm. and has 38 segments; the smaller specimen measures 10 mm. and has 37 segments. Both are purple in alcohol but many of the dorsal cirri and elytra are missing and some of the antennae as well. For this reason the description is based on both specimens. The prostomium (fig. 2g) is rectangular and longer than broad with rather small eyes which are not easily distinguished against the purple background. 282 ANNALS OF THE SOUTH AFRICAN MUSEUM The anterior pair are set half-way back on the sides of the prostomium. At first sight it was thought that the insertion of the antennae was harmothoid and then it was noticed that prostomial peaks are absent and the lateral antennae are not ventral but subterminal in origin. When the worm was turned over it was further noted that there is a well developed facial tubercle and that the bases of the cirrophores which bear the lateral antennae are fused to the lower side of the prostomium but their distal ends incline upward so that the antennae appear to be almost terminal. This type of insertion is best termed.subterminal but it should be noted that it is quite distinct from the so-called subterminal insertion of Halosydna where the lateral antennae arise from a lower level than the median antenna only because the latter is actually dorsal in origin. The median antenna here is terminal and it isa large dark almost club-shaped organ about as long as the prostomium. Its surface is quite smooth. The lateral antennae are similar in shape but only half as long. The palps are rather short and the tentacular cirri are rather stout. The tentacular segment bears a single seta. There are 15 pairs of elytra which cover the dorsum. Each is broadly oval in shape (fig. 27) and quite smooth except for a small patch of minute rounded tubercles on the anterior margin. There is no trace of a marginal fringe. The anterior half of each elytron is colourless where it is overlapped by the preceding one but the exposed posterior half is dark purple with clear cells here and there. The dorsal cirri (fig. 2k) are similar to the antennae, being dark in colour, quite smooth and swollen distally before the tip. The notopodium bears about a dozen stout notosetae and the neropodium which has a pointed or triangular presetal lip and a shorter, more rounded post-setal lip bears some 20-30 rather short neurosetae. The ventral cirrus is smooth, evenly tapered and extends almost to the end of the neuropodium. The notosetae (fig. 2) are stout, very lightly serrated and end in abruptly pointed tips. Under high power the tiny close-set serrations produce a herring- bone pattern on the surface of the seta but individual serrations cannot be seen. The neurosetae (fig. 2f ) have more slender shafts than the notosetae but the blades are of normal length and bear about 25 rows of very fine, transparent spinules. The tips (fig. 2f1) are short, the terminal tooth is sharp and well hooked but the secondary tooth is minute or even absent on some of the inferior neurosetae. The pygidium bears a pair of dark sausage-like anal cirri. Reference to Fauvel (1923) and Monro (1936) shows that the distinction between the genus Malmgrenia McIntosh 1876 and Eulagisca McIntosh 1885 is not clear. Malmgrenia as defined by Fauvel (1923) is generally similar to Harmothoe but differs in having the lateral antennae subterminal in origin, the notosetae very stout and faintly spinulose and the neurosetae either unidentate or with a minute secondary tooth. Fauvel in his definition states that the insertion of the antennae is similar to that of Halosydna but his figure of the head of Malmgrenia castanea shows that it is not similar to that of Halosydna THE POLYCHAET FAUNA OF SOUTH AFRICA 283 but the same as M. purpurea described above. Eulagisca was not defined by McIntosh (1885) but has been defined by Monro (1936) in terms which suggest that it is synonymous with Malmgren. However the figures of Eulagisca corrientis (the type species of Eulagisca) given by Monro (1930) show that the insertion of the antennae is similar to that of Halosydna, the notosetae strongly serrated and the neurosetae unidentate. The present species M. purpurea may be distinguished from M. castanea by the position of the eyes, the possession of shorter and much stouter antennae and dorsal cirri and by the fact that the neurosetae are mainly bidentate. M. castanea is known to be commensal of the echinoid Spatangus purpureus ; M. purpurea is probably a commensal of Spatangus capensis. Lepidonotus durbanensis Day 1934 Lepidonotus durbanensis Day 1934, p. 18, fig. 1 a—c. Day 1951, p. 9. Records: MB.86(1). Notes: Only a single small specimen was obtained but it is quite distinct from the common L. clava var. semitecta. This is the most southerly record of this Natal form. Lepidonotus clava (Mont.) var. semitecta Stimpson. 1855 Lepidonotus clava var. semitecta. Willey 1904, p. 256, pl. 13, fig. 4. Day 1953, p. 399. Records; SB.207(1); LB.161(3); TB.302(1), 305(2),. 309(1),. 313(1), 317(1); TRA.122(1); WCD.8(3); False Bay—5o records on rock o—73 metres, common; MB.9(2), 16(1), 23(1), 40(12), 49(12), 53(15), 56(2), 59(1), 62(6), 67(x), 85(3); KNY.21(1), 22(1); LIZ.2(2), 6(4), 18(c), 27(3); SCD.89(1). Polynoe erythrotaenia (Schmarda) 1861 Hemilepidia erythrotaenia Willey 1904, p. 258, p. 13, figs. 6, 26. Records: SB.179(1); TB.324(1). Polynoe scolopendrina Savigny 1820 Polynoe scolopendrina. Fauvel 1923, p. 80, fig. 30. Day, 1953, p. 406. Records: LAM.8(1), 13(1), 25(1), 31(4), 59(common), 63(1); LB.299(2); TRA.135(1); FAL.r1o(1), 122(f), 219(1), 345(1); MB.41(1), 53(1), 56(1), 67(1); LIZ.2(3), 6(1), 29(2); SCD.58(1). ?Polynoe capensis McIntosh 1885 Polynoe capensis McIntosh 1885, p. 114, pl. 4, fig. 4; pl. 15, fig. 1; pl. 19, fig. 4; pl. ga, figs. 4 and 5. Notes on the type material. No new specimens have come to hand, but the type specimens dredged from 98 fathoms off the Cape of Good Hope and 284 ANNALS OF THE SOUTH AFRICAN MUSEUM now in the British Museum (register number 1885 12.1.94) were re-examined. The type material consists of blackened fragments of two specimens and several loose elytra. The median antenna is missing and the lateral antennae which are terminal in origin, are two-thirds the length of the prostomium. The anterior pair of eyes are slightly larger than the posterior pair and are dorsal in position. The total number of elytra originally present cannot now be determined but one posterior fragment shows 8 posterior segments without elytra scars. The loose elytra are oval, one half brown and one half white. The pale half has a triangular patch of small chitinous tubercles. The parapodia have a fair number of notosetae, each of which is weakly spinulose with an abruptly tapered tip. The neurosetae are also weakly spinulose and the tip at first appears to be unidentate, but careful examinations show a small, blunt secondary tooth. Although the colouration of the elytra is reminiscent of Polynoe erythrotaenia, the other characters differ and the terminal insertion of the lateral antennae shows that this species should be removed from the genus Polynoe. Fresh material is required before its generic position and exact characters can be determined. Lepidasthenia elegans (Grube )1840 Lepidasthenia affinis Horst 1917, p. 85, pl. 19, fig. 8. Lepidasthenia elegans. Fauvel 1923, p. 88, fig. 23 a-g. Records: TRA.133(1); SCD.58(1). Notes: These two specimens from dredgings off the Cape and another that I have seen from the shore of Inhaca Island agree very well with Fauvel’s description apart from the minor points noted below. They have rather fewer elytra (23 as against 30-36 for the Mediterranean specimens) and individual elytra are rather larger though they leave the central third of the back bare except at the anterior end. Each elytron is speckled with dark pigment except for a white spot which marks the area of attachment. The prostomium, eyes, antennae and dorsal cirri are identical. As in the Mediterranean form the notopodia usually lack setae but careful search revealed that a few feet have a single minute notoseta with poorly marked serrations. The neurosetae are variable both along the length of the body and within a single fascicle. The first few feet have 2-3 slender superior setae with long, coarsely spinulose blades. The remaining setae of an anterior foot are stouter with short spinulose blades and strongly bidentate ends though the secondary tooth is markedly smaller than the terminal one. Further back along the body the slender superior setae are lacking and a giant brown superior seta appears. Moreover the secondary tooth is reduced and may be completely lacking so that the giant seta becomes unidentate, with only a few rows of rather worn spinules. L. elegans has already been recorded from Zanzibar by Potts (1910) but he makes no mention of the slender superior setae of anterior feet. Horst’s description of L. affinis from Lombok in the East Indies leaves no doubt that THE POLYCHAET FAUNA OF SOUTH AFRICA 285 his specimen is conspecific with mine though it had 40 pairs of elytra. He separates L. affinis from L. elegans as described by Potts on the relative size of the elytra and minor differences in the disposition of the setae. I feel, however, that Potts’s description was too brief and that further specimens have shown these characters are too variable to warrant specific distinction. Lepidasthenia brunnea n. sp. (Fig. 3 a-d) Records: FAL.352(2). Description: ‘The type material consists of two broken specimens, the largest anterior fragment measuring 40 mm. by 5 mm. (including parapodia) and having 48 segments. The body is light brown in alcohol with colourless parapodia and the large deciduous elytra are half brown and half transparent. The prostomium (fig. 3a) is bilobed and almost twice as broad as long. The anterior pairs of eyes are larger and wider apart than the posterior pair. The three long smooth antennae are very alike. All arise from short cerato- phores on the anterior margin of the prostomium and taper slowly towards the final slender tip. The median, which is a little longer than the laterals is 5-6 times the length of the prostomium and roughly equal in length to the width of the body. The tentacular cirri are both about as long as the lateral antennae and each arises from a stout base with a projecting aciculum. The palps are stout and equal the length of the tentacles. The body is long and flattened, brown dorsally and pale ventrally. The elytra are large and overlap to cover the back They are inserted as usual on segments, 2, 4, 5, 7,9... and alternating segments anteriorly but on every third further back and, since the body is broken, it is not possible to say how many there were, but as the anterior fragment has 21 elytra scars there must have been many more on the complete worm. Each elytron is oval in shape, very thin and quite smooth. The anterior half which lies under the preceding elytron is colourless but the exposed posterior half is pale brown. The dorsal cirri arise from short broad cirrophores and extend outwards well past the tips of the setae. Each one is quite smooth and colourless, rounded in section and tapers evenly to a slender tip. The first few are considerably longer than those from the middle of the body. The notopodia (fig. 3b) are reduced to tiny pointed lobes lying on the dorsal surfaces of the neuropodia. Each contains an aciculum but there are no notosetae. The neuropodia are stout fingerlike organs projecting from the sides of the body and posterior ones are longer than the body is broad. Each has a long pointed presetal lip, a rather shorter postsetal lip and between these two issues a fan of long setae. The ventral cirrus is short and between it and the ‘body the ventral surface of the parapodium bears a single row of about 8 elongate papillae. 286 ANNALS OF THE SOUTH AFRICAN MUSEUM Fic. 3. Lepidasthenia brunnea: a head; 6 posterior view of parapodium; c¢, c) superior neuroseta; d, d' inferior neuroseta. Sthenelais papillosa: e posterior view of parapodium; f head; g elytron; / neuroseta. THE POLYCHAET FAUNA OF SOUTH AFRICA 287 The neurosetae are characterized by the possession of a superior group of setae (fig. 3c) with long blades feathered to their fine hairlike tips which end in minute knobs. A few such setae are present in the anterior feet of many species of the genus Lepidasthenia but in this species there are many of them in all feet. The inferior setae (fig. 3d) are stouter and have much shorter feathered blades which end in bidentate tips. The secondary tooth is half the size of the terminal one but the feathering comes so close to the end that the secondary tooth may be obscured. | This species differs from L. elegans in many respects particularly the colouration, the persistence of fine unidentate setae, the lack of a giant seta and the possession of papillae on the base of the neuropodia. In L. maculata Potts, which has papillae on the ventral surface of the parapodium, the fine superior setae are very few and only present in anterior segments and the bidentate setae have very few rows of serrations. L. berkeleyae Pettibone (1948) is close but the antennae are shorter, there are fewer fine neurosetae and there are no ventral papillae on the neuropodia. Lepidasthenia sp. Records: AFR.790(1). Notes: An anterior fragment of 32 segments measuring 18 mm. was obtained. It is completely colourless with rather glassy setae, and large, delicate translucent elytra. The general shape of the body, the head and the appendages is similar to L. brunnea but there are two important differences. There are no papillae on the ventral surface of the neuropodia and there are no bidentate neurosetae, only fine setae with long blades feathered to their slender hairlike tips. Since the fragment consists of only 32 segments with 15 pairs of elytra it is impossible to say how many elytra are present in the entire worm. Further, the bidentate type of neurosetae might appear in posterior feet. For these reasons this is not described as a new species. Euphione elisabethae (McIntosh 1885) Euphione elisabethae McIntosh 1885, p. 62, pl. 9, fig. 3; pl. 17, fig. 7; pl. 18, fig. 10; pl. 8A, figs. 3-6. Records: AFR.691(1), 707(1), 791(1), 1529(1), TRA.21(1). Hhololepida australis Monro 1930 Hololepida australis Monro 1930, p. 93, fig. 9 a—h. Records: AFR.707(1). Notes: The present material has been compared with Monro’s type in the British Museum and found to be conspecific. The nuchal flap is triangular. The elytra which are fragile and deciduous bear numerous three-pronged 288 ANNALS OF THE SOUTH AFRICAN MUSEUM tubercles exactly as described by Monro. They are reminiscent of the tubercles of Halosydna and indeed the genus Hololepida is closely related to Halosydna. The notosetae are not smooth as stated by Monro for under high magnification they show steplike serrations. The superior neurosetae are as shown by Monro but his intermediate type was not found. The inferior neurosetae are bidentate and the secondary tooth, though finer than the terminal one, is epeh oe: as long so that the ends of these setae seem to be split. This is a new record for South Africa. Polyeunoa laevis Mcintosh 1885 Enipio rhombigera Ehlers 1908, p. 47, pl. 4, figs. 1-12. Polynoe agnae McIntosh 1925, p. 21, pl. 2, figs. 3 and 4. Hemilepidia erythrotaenia (non Schmarda) McIntosh 1925, p. 26, pl. 2, figs. 9 and 1o. Records: AFR.789(1), 831(1); TRA.48(1); WCD.3(4). Notes: The present specimens, like those described by Monro (1936), have elytra which are smooth apart from a triangular patch of minute hemispherical tubercles near the point of attachment. These papillae are absent from MclIntosh’s type. The superior neurosetae have markedly stronger spinules than the inferior ones. Small specimens of 18 and 20 mm. have a minute but distinct secondary tooth on the neurosetae but large specimens are usually unidentate though vestiges of the secondary tooth may occasionally be found among the inferior setae. The identity of Ehlers’ Enipio rhombigera with P. laevis has long been recog- nized. A recent examination of the type of Polynoe agnae (also called Eunoa agnae by McIntosh 1925) which is now in the British Museum (registered number 1924:7:21-27) shows that this is also P. laevis. The specimen recorded by McIntosh (1925) as Hemilepidia erythrotaenia is in a very poor condition but re-examination again shows that it is also P. laevis. Subfamily SIGALIONINAE Pholoe minuta Fabricius var. inornata Johnston 1865 Pholoe minuta var. inornata. Fauvel 1923, p. 120, fig. 44 a—h. ?Pholoe minuta Ehlers 1913, p. 450. Augener 1918, p. 118. Records: SB.189(1); FAL.22(1), 152(1), 314(1); SCD.26(1). Notes: The median antenna has a stout base and a slender tip, the whole equalling the length of the prostomium. The eyes are coalescent. The elytra are rounded to reniform and the margins carry soft papillae which are not annulated. The papillae on the neuropodia are not obvious and the shaft-heads of the neurosetae are lightly serrated. Ehlers (1913) recorded P. minuta from False Bay and Augener (1918) who recorded the variety inornata from South West Africa has discussed its distribution and affinities, but the distinctions between the various species and varieties given by Fauvel (1923) are not very convincing. THE POLYCHAET FAUNA OF SOUTH AFRICA 289 Sthenelais boa (Johnston) 1833 Sthenelais boa. Fauvel 1923, p. 110, fig. 41 a-l. Day 1953, p. 406. Records: LB.299(3); TRA.88(1), 133(1). Sthenelais limicola (Ehlers) 1864. Sthlenelais limicola. Fauvel 1923, p. 113, fig. 42 a-g. Records: AFR.736(1); ?TRA.106D(1); FAL.184(3), 206(1), 228(1), 237(1), 238(3), 242(1), 341(1), 352(7), 375(3), 376(4), 378(1); ?>MB.79(1); SCD.109(1). Sthenelais papillosa n. sp. (Fig. 3eJ) Records: FAL.223(1), 334(1), 341(1). Diagnosis: A species with specked elytra lacking sumple serrate neurosetae and having a papillose ventral surface. Description: The type material consists of two fragmentary specimens dredged in False Bay. The larger specimen (FAL.223) is an anterior end of 40 segments and the smaller one FAL.334 is in three fragments but possesses elytra. It is estimated that the larger specimen might have measured 40 mm. by 3 mm. when complete. The prostomium (fig. 3 f) is ovoid, a little longer than broad, with a stout median ceratophore and two pairs of well-marked eyes. The ‘ctenidia’ on the ceratophore are rather small and the ceratostyle is slightly longer than the prostomium. The tentacular segment has a flattened presetal lip, a bundle of simple notosetae, a short dorsal cirriform appendage (which, according to Fauvel (1923), corresponds to the lateral antenna) a large dorsal cirrus (or ? postsetal lobe) and a long ventral cirrus arising from the base of the foot. The body is elongate and the whole of the ventral surface including the midventral line and the bases of the parapodia is densely covered with small spherical papillae. Anterior elytra are not known but those from the middle of the body (fig. 3g) are reniform without any external notch and the margins bear minute unicellular papillae which are elongate laterally and spherical posteriorly. The surface of each elytron is speckled with brown and studded with tiny, transparent, very lightly chitinized and flattened tubercles or cushion-like papillae. The notopodium (fig. 3e¢) has about 6 short stylodes and the usual bundle of long notosetae. The neuropodium has 2-3 short stylodes at the apex of the acicular lobe and a low presetal lip edged with about 8 elongate papillae. The ventral margin of the parapodium as mentioned above, bears numerous spherical papillae and a single short ventral cirrus. The notosetae are typical. The neurosetae (fig. 3j) lack superior simple serrate setae and the compound setae are all very similar. The shafts are stout, 290 ANNALS OF THE SOUTH AFRICAN MUSEUM the triangular shaft-heads are lightly serrate and most of the blades are long and simple though some of the short inferior ones have 2-3 poorly marked articulations. 7 S. papillosa has been compared with the types of S. zeylanica Willey and also S. variabilis, S. orientalis and S. foliosa Potts (1910) all of which have a papillose ventral surface but in each case other characters differed. S. papillosa also approaches §. minor in the lack of simple serrate neurosetae but the latter may be distinguished by the possession of elongate and _pluriarticulate neurosetae and the lack of papillae on the ventral surface. Sigalion squamatum Delle Chiaje 1841 Sigalion squamatum. Fauvel 1923, p. 104, fig. 39 m-—o. Records: FAL.243(1), 357(1). Notes : wo anterior fragments were obtained belonging to large specimens probably exceeding 100 mm. Four tiny well-separated eyes are just visible through the skin of the prostomium. The dorsal cirrus of the first setiger is 4—2 the length of the ventral cirrus but dorsal cirri are absent from the second and subsequent setigers. A clavate presetal papilla appears at the end of the notopodium from the 5th foot onwards. Each elytron is rectangular and the external margin bears simple papillae on its upper surface and bipinnately branched papillae along its external margin, each of these having 7—10 pairs of branches. Rudimentary cirriform gills appear on the medial and lateral margins of the elytrophore of the second or third foot and by the 6th foot the gills are well developed. Later the medial gill decreases in size but the lateral one remains large. The anterior face of the notopodium of each foot has a patch of conical tubercles near its base. The notosetae are numerous and minutely serrate. The neurosetae are of 5 types: (a) about 6 simple bipectinate setae in the superior group above the aciculum. (5) About 2 compound setae with coarsely serrate shafts and tapered, pluriarticulate blades. (c) About 4 compound setae with swollen, closely serrate shaft-heads and pluriarticulate blades. (d) About 6 compound falcigerous setae with fine serrations on the shaft-heads and simple bidentate blades. (e) Very numerous inferior compound setae with very lightly serrate shaft-heads and long pluriarticulate blades. All the various types of compound setae of this and all other species of Sigalion which have been examined have bidentate — tips. Statements to the contrary are probably due to the examination of broken-tipped setae. This South African material has been compared with specimens of S. squamatum from Naples, which is the type locality. The European material shows that the development of the superior ‘stylode’ (=presetal lobe) of the neuropodium is variable and not of much value in separating S$. squamatum from S$. mathildae. Again the pennate marginal papillae on the elytra are not normally as widely different as figures 39 c and m in Fauvel (1923) would THE POLYCHAET FAUNA OF SOUTH AFRICA 291 suggest. I agree with Fauvel that the main difference lies in the setae, though the tubercles on the face of the notopodium are more poorly developed in S. mathildae. Incidentally these same tubercles are exceedingly long and occasionally branched in S. buskiz McIntosh (1885) and suggest that this is a valid species which lacks falcigers with simple bidentate blades. It is also to be noted that Fauvel makes no reference to the medial gill on the elytrophores, and this character was not checked on the Naples material. Sigalion capense n. sp. (Fig. 4 af) Records: FAL.237(1), 375(1); MB.4(1). Description: FAL.237 an ovigerous female, is selected as the type. It is 16 mm. long by 2 mm. wide and is broken at the 6oth segment. It is quite white in alcohol. The prostomium (fig. 4a) is almost oblong, a little longer than wide and somewhat rounded posteriorly. The two pairs of small eyes are visible through the skin about the middle of the prostomium. The anterior and posterior pair are close together on either side. The antennae are small cylindrical papillae arising from the prostomium at its junction with the forwardly projecting tentacular segment which bears two pairs of subequal tentacular cirri. The palps are long and slender. The ventral cirrus of the second foot is about the same length as the tentacular cirri. A single cirriform branchia arises from the lateral side of the elytrophore of the 4th and all succeeding segments and in anterior segments there is also a small branchia (or ctnidium) on the medial side of each elytrophore. The posterior elytrophores are swollen with developing eggs. The elytra themselves (fig. 4c) are rounded to rectangular with smooth surfaces and bear 10-12 bipinnate papillae on the external margin; each of these has 4-8 pairs of branches (fig. 4c). The notopodium (fig. 4b) is swollen distally with a single large presetal papilla (stylode) at its end and three glandular cushions on its superior margin. It bears a bundle of long slender setae with hairlike tips. The stouter ones are serrate on the inferior margin. The neuropodium is obliquely truncate with a bluntly conical acicular lobe, a vestigial presetal lip, a well-developed, trangular postsetal lip and a long tapered ventral cirrus. There are 4 types of neurosetae. The supra-acicular neurosetae include: (a) 3-6 simple bipinnate setae (fig. 4d); (b) 4 fairly stout compound setae (fig. 4) with long pluriarticulate blades and serrated shaft- heads; (c) about ro fairly stout compound setae with long pluriarticulate blades and smooth shaft-heads (fig. 4¢) The infra-acicular setae include: (d) about 4 setae similar to group (c) and (e) very numerous fine compound setae with long pluriarticulate blades and smooth shaft-heads. It is emphasized that 292 ANNALS OF THE SOUTH AFRICAN MUSEUM h Fic. 4. Sigalion capense: a head; b parapodium; ¢, c! elytron and details of marginal papillae; d simple serrate neuroseta; e smooth-shafted neuroseta; f serrate-shafted neuroseta. Psammolyce articulata: g head; h neuroseta; j parapodium; k, k! 6th elytron and details of margin; / posterior elytron. 2 ‘ THE POLYCHAET FAUNA OF SOUTH AFRICA 293 all compound setae have long pluriarticulate blades with minutely bidentate tips’: . S. capense may be distinguished from S. sqguamatum and S. mathildae by the lack of neuropodial setae with simple blades. In this and other respects it resembles S. ovzgerum Monro (1924) but the latter carries eggs in its swollen elytra and has supra-acicular neurosetae with excavated shaft-heads as figured by Monro (1936) fig. rad. Psammolyce articulata n. sp. (Fig. 4 g-l) Records: FAL.117(1), 211(2), 214(1). Description: Four broken specimens were obtained and specimen FAL.117 which is in two fragments totalling 50 mm. by 4 mm. for 100 segments was selected as the holotype. It is sandy brown in alcohol with darker setae. The elytra and the uncovered part of the back between them is covered with sand- grains and foraminiferan shells and the ventral surface is completely covered with small rounded tubercles. The head (fig. 4g) is protected between, but not fused to, the forwardly directed first pair of feet which bear the tentacular cirri. The prostomium itself is as broad as long and bears two pairs of large eyes and three antennae. The first pair of eyes are anterior and ventral and concealed beneath the large ceratophore of the median antenna. The posterior pair of eyes is on the sides of the prostomium immediately behind the short stumpy lateral antennae which are quite separate from the tentacular segment. The median antenna is borne on a large ceratophore which curves down like an elephant’s trunk. The first setiger or tentacular segment has well-developed dorsal and ventral rami and two bundles of setae. The notopodium has a rather short dorsal tentacular cirrus and a stumpy setigerous lobe. The neuropodium has a presetal bract, a bundle of neurosetae and a large cirriform postsetal lobe below which is the long tapering ventral tentacular cirrus. The second segment bears the first pair of elytra, a short cirriform gill, two setigerous lobes and a long ventral cirrus. There is no dorsal cirrus. The third segment bears a long stout elytrophore extending almost to the end of the neuropodium and a shorter cirrus. The elytra (figs. 4 k-l) vary along the length of the body but all of them are covered with sand grains. The anterior pair are pear-shaped and extend forward over the head. The margins appear smooth and there are certainly no incisions or large projections. The second pair are reniform. The margins bear minute papillae and the surface is studded with microscopic tubercles. The next few elytra are more rounded (fig. 4k) straight in front and with two short lobes on the medial margin and small pear-shaped lappets posteriorly. The surfaces are covered with microscopic tubercles and the papillae are better developed, both on the surface and the margins. All of them are clearly jointed. 294 ANNALS OF THE SOUTH AFRICAN MUSEUM Further back the two antero-medial lobes tend to disappear (fig. 4/) and the 4-5 posterior lappets become irregular. Most of the elytra are roughly triangu- lar and fairly small so that the middle third of the back is uncovered. The feet (fig. 43) show little change from the 4th onward. The notopodium is well developed. The neuropodium is obliquely truncated with a presetal row of elongate papillae and tufts elsewhere. The lower surface is covered by the rounded tubercles which extend on to the parapodia from the ventral surface. The ventral cirrus is long and tapered and 2-3 very long and slender papillae arise from its base. The notosetae are numerous and very fine. Each is minutely serrated to its hairlike tip. The neurosetae (fig. 4h) are all compound and falcigerous. Those of the second setiger have densely serrated shafts but in later feet the serrations are reduced to 3—5 rows on the shaft-head. The blades are all bidentate but never pluriarticulate. In his diagnosis of the genus Psammolyce, Fauvel (1923) states that the lateral antennae are fused to the first setiger or possibly absent in P. inelusa. Monro (1936) makes the same remark. In the present species as stated, the lateral antennae arise from the prostomium which is quite separate from the first setiger. P. articulata is also characterized by the possession of jointed papillae on the elytra. Faint indications of jointing were seen in the papillae of P. semiglabra Monro (1936), the type of which was examined, but the two species differ in many characters. Possibly P. articulata comes closest to P. zeylanica Willey (1905) in the shape of the elytra but the latter again has the lateral antennae fused to the first setiger. Thalenessa oculata (Peters) 1854 Euthalenessa dendrolepis (Clap.) Fauvel 1923, p. 114, fig. 42 h-o. Euthalenessa oculata Day 1953, p. 407. Records: SB.207(1); TB.301(1), 304(4); FB.312(1), 321(1); FAL.233(1), 238(1), 349(3); SCD.105(2 juveniles). Family CHRYSOPETALIDAE Bhawania goodei Webster 1884 Bhawania goodei. Augener 1918, p. 98, pl. 2, figs. 1-2, text-fig. 1. Day 1953, p. 407. Records: False Bay —21 records on rocky or shelly bottoms between o and 36 metres. MB.13(1), 53(1), 56(1), 77(1), 85(1), 86(1); LIZ.9(1), 27(1), 33(1); SCD.58(1), 89(1). Paleanotus chrysolepsis Schmarda 1861 Paleanotus chrysolepis Schmarda 1861, p. 163, pl. 37, figs. 326-9. Ehlers 1913, p. 450. Day 1957, p. 66. Records: False Bay—15 records on rock or shelly bottoms between o and 40 metres. THE POLYCHAET FAUNA OF SOUTH AFRICA 295 Family AMPHINOMIDAE Chloeia inermis Quatrefages 1865 Chloeia gilchristi McIntosh 1925, p. 15, pl. 1, figs. 7-8. Day 1934, p. 27, fig. 4 a-b. Chloeia inermis. Monro 1936, p. 80. Records: AFR.801(1); TRA.20(1). Notes: The specimens were recovered from the stomachs of fish and are rather soft but they still retain rather vague purplish markings along the middorsal line and the dorsal cirri are purple. Monro states that the gills begin on setiger 5 but here they begin on setiger 4. The setae agree perfectly with Monro’s description. A spur is virtually absent from most of the setae though a minute one can be seen on some of the ventral ones. All but a few of the harpoon setae are smooth. Euphrosyne capensis Kinberg 1857 Euphrosyne capensis. McIntosh 1885, p. 1, pl. 2, fig. 5, pl. 1A, figs. 1-3. Day 1953, p. 408. Records: LAM.8(1), 15(1), 31(8), 35(2), 47(1), 51(3), 57(1), 59(3), ©3(1) ; TB.305(1), 323(1); False Bay—17 records between o and 42 metres on rock or broken shell; MB.16(2), 49(2), 67(1), 77(1), 78(2); LIZ.18(1) Euphrosyne myrtosa Savigny 1818 Euphrosyne myrtosa. Gravier 1901, p. 254, pl. 10, figs. 147-9. Records: TRA.132(1), 135(1); SCD.40(1). Notes : Ehlers (1913) who previously recorded this well-known species from the Cape was doubtful whether it was distinct from E. capensis. Many specimens of varying size were therefore examined and it appears that there is a constant difference in the branchiae. In E. myrtosa there are 6-8 branchial trunks and the tips of the branches are blunt and not expanded. In E. capensis there are 10-11 branchial trunks and the tips of the branches are swollen and pointed rather like acorns. Eurythoe chilensis Kinberg 1857 Pareurythoe chilensis Hartman 1948, p. 45, pl. 5, fig. 11. Eurythoe chilensis Kinberg 1857, p. 13. Monro 1930, p. 28, fig. 1 a—e. Records: FAL.209(1). Notes: The single specimen is 20 mm. long. The caruncle is attached to the dorsum as far back as the second setiger, but a free posterior projection extends back to the fourth setiger. This species is easily distinguished from the tropical E. complanata which extends down the Natal coast by its smaller size and the fact that all the spurred setae are lightly serrated in E. chilensis and smooth in E. complanata. This is a new record for South Africa. 296 ANNALS OF THE SOUTH AFRICAN MUSEUM Family PHYLLODOCIDAE Phyllodoce (Anaitides) madeirensis Langerhans 1879 Phyllodoce (Anaitides) africana Augener 1918, p. 171, pl. 2, fig. 25; pl. 3, figs. 49-51; text-fig. 11 (partim). Phyllodoce (Anaitides) madeirensis. Fauvel 1923, p. 150, fig. 53 a-d. Phyllodoce patagonica (non Kinberg) Monro 1930, p. 72 (partim). Records: AFR.707(1); TRA.56(1), 133(1); FB.316(1); FAL.241(1), | 373(1). Notes: By the kindness of the Director of the Hamburg Museum I was | able to examine Augener’s specimens of P. africana from Goree (number V.1986). There are two specimens, one with the proboscis extruded and one with the proboscis retracted; both were quite pale in alcohol, the pigmentation to which Augener refers having faded. The former specimen with the extruded proboscis is clearly on Anaztides with 6 lumpy ridges on the distal part of the proboscis and 6 lateral rows each with 8-10 compressed papillae basally. The prostomium is cordate with 4 normal antennae, a pair of large eyes and an occipital papilla in the posterior notch. The first tentacular segment is invisible dorsally but the second and third are distinct. There is no parapodium or : I I setae on the third tentacular segment, the formula being 1 + o—_+o0—. | I N Anterior dorsal cirri are broadly lanceolate but later ones are obliquely truncate near the tip and rhomboidal. The ventral cirri are pointed and longer than the setigerous lobes. This specimen seems to me to be a typical P. madeirensis. The second specimen with the retracted proboscis was dissected and its proboscis proved to be irregularly covered with large pointed papillae except » at the base where the dorsal wall was bare. The prostomium is elongate and deeply incised posteriorly forming a pair of lateral lobes which extend back to segment 2. No occipital papilla was seen. The first tentacular segment is not visible dorsally, the second is partly visible between the posterior lobes of the prostomium but the third segment is fully visible. A small setigerous lobe bearing 2-3 setae is present on tentacular segment 3, the formula thus being I I ce 1 + o— + S—. Anterior dorsal cirri are broadly lanceolate, almost cordate, I N but later ones are longer and more asymmetrical with broad cirrophores. The ventral cirri are pointed and about as long as the setigerous lobes. ‘This specimen does not belong to the sub-genus Anaitides and should be named Phyllodoce (Phyllodoce) africana. I have also examined the specimen reported by Monro (1930) from Simonstown under the name of P. patagonica. It is quite clearly P. madeirensis and may be distinguished from P. patagonica by the absence of setae on tentacular segment 3. THE POLYCHAET FAUNA OF SOUTH AFRICA 297 It may also be mentioned that some of my own specimens recorded above from deeper dredgings are blotched with dark pigment and some of the dorsal cirri are black and others white. Otherwise they are indistinguishable from the normal green variety of P. madeirensis. Phyllodoce sp. Records: ‘TRA.133(1). Notes: Specimen TRA.133.L certainly belongs to a species which has not been recorded from South Africa before but the preservation is not too good and its exact determination is doubtful in consequence. The body is brownish blotched with darker pigment. The proboscis appears to be diffusely papillose. The prostomium is cordate and there is a small occipital button in the posterior notch. The first tentacular segment is fused to the prostomium but the second and third are distinct. The tentacular cirri are unusual for they are stout, sausage-shaped and constricted basally. : I I The tentacular formula is 1 + 0— +S—. The dorsal cirri are ovoid and I I swollen and the oval ventral cirri are a little larger than the blunt setigerous lobes. The setae have oval shaft-heads and short blades. Phyllodoce macrophthalma Schmarda 1861 _ Phyllodoce macrophthalma. Fauvel 1923, p. 146, fig. 51 f-g. Records: MB.20(t). Notes: The single 12 mm. specimen is referred to P. macrophthalma with some hesitation. The body is slender and dark green. The prostomium is cordate with a very small posterior notch and the presence of an occipital button is doubtful. The proboscis on dissection seems to be lightly papillose. The first tentacular segment is fused to the prostomium but the second and third are distinct. All the tentacular cirri are well developed and cylindrical, the formula being: 1 + ele = ey I I The dorsal cirri are cordate, possibly a little longer than broad and the ventral cirri are ovoid. The setigerous lobe has a notched presetal lip and bears numerous setae with oval shaft-heads striated distally and fairly short blades. | Schmarda (1861) stated that the dorsal cirri are rhomboidal but both Ehlers (1913) who recorded this species from Simonstown and Fauvel (1923) follow Saint-Joseph (1888) and describe cordate dorsal cirri. I have not seen the latter paper. 298 ANNALS OF THE SOUTH AFRICAN MUSEUM Phyllodoce (Anaitis) capensis n. sp. (Fig. 5 a-c) Records: FAL.316(1); TRA.133(1). Description: Specimen FAL.316 which measures 35 mm. by 1:5 mm. for about 100 segments was chosen as the holotype. It is depressed and tapered posteriorly and is creamy white in alcohol. The prostomium (fig. 5a) is broadly rounded anteriorly and produced posteriorly. The frontal antennae are well developed, the eyes are large and there is a small occipital button. The first and second tentacular segments are fused and form a sort of transparent shield which grows forwards over the sides of the prostomium to cover part of the eyes but the occipital button in the mid-dorsal line is not covered. The third tentacular segment is distinct. All the tentacular cirri are cylindrical and tapered but T, and V, are shorter than T, and T,. There is a small setigerous lobe on the third tentacular segment : I I so that the tentacular formula is: 1 + o— + S_. I The proboscis was not extruded and dissection was not entirely successful. The distal part definitely has 6 longitudinal rows of large soft papillae or rugosities but the oral end is indistinct. There were no obvious rows of papillae and it may be smooth. The papapodia and dorsal cirri (fig. 55) are essentially similar throughout the length of the body. The dorsal cirri are fairly large and rounded to broadly cordate but they do not cover the back. The ventral cirrus is oval and larger than the setigerous lobe which has a notched presetal lip and bears about 12 fine heterogomph spinigers. The shaft-head (fig. 5c) is asymmetrical with a large curved tooth accompanied by 3-4 smaller denticles on one side and a smaller tooth on the other side. The finely serrated blade is long, fairly broad basally and tapers gently towards the tip. This South African species is obviously related to P. (A.) kosteriensis Malmgren from northern Europe but differs in several respects. The shield formed by the fusion of T, and T, is better developed, the proboscis is different, the dorsal cirri are proportionately longer and the setae have slightly different shaft-heads. P. (A.) wahlbergi Malmgren from the Arctic is, according to Bergstré6m (1914) a much broader worm with a single blunt tooth on the shafthead of the seta. Phyllodoce castanea Marenzeller 1879 — Genetyllis castanea Bergstrom 1914, p. 158, pl. 3, fig. 4, text-fig. 53. Phyllodoce rubiginosa Augener 1918, p. 168. Records: SB.120(1); FB.302(2), 307(3); FAL.128(1), 266(1); MB.23(1). 299 THE POLYCHAET FAUNA OF SOUTH AFRICA — Fic. 5. Phyllodoce (Anaitis) capensis: a head; 6 anterior view of parapodium; ¢ seta. Eulalia bilineata: d head; e anterior view of parapodium; / seta. Eulalia macroceros: g head; h anterior view of parapodium; 2 seta. 300 ANNALS OF THE SOUTH AFRICAN MUSEUM Notes: The Director of the Hamburg Museum very kindly sent me the specimen V.8731 recorded by Augener 1918 from Swakopmund under the name of Ph. rubiginosa. The proboscis was retracted but when dissected proved to be very long and covered with irregularly arranged conical papillae. The whole specimen is rusty red all over with an ovoid prostomium not notched posteriorly and fused to tentacular segment 1. The antennae are very small. All the tentacles are short, tapered and rounded. Tentacular segment 1 is not visible dorsally but 2 and 3 are distinct and bear setigerous lobes with numerous setae the formula being 1 + Gib + s—. Dorsal cirri are broadly cordate and I almost symmetrical. Ventral cirri are ovoid to reniform and extend beyond the rounded setigerous lobes. P. castanea and P. rubiginosa are very alike but the latter is described by Fauvel as having large not small antennae and he figures the dorsal cirri as asymmetrical. Certainly Augener’s specimen from Swakopmund is very like mine from the western and southern coasts of the Cape. Eulalia (Steggoa) capensis Schmarda 1861 Eulalia capensis Schmarda 1861, p. 86, pl. 29, fig. 231. Willey 1904, p. 259. Eulalia viridis var. capensis McIntosh 1904, p. 34. Day 1953, p. 30. Eulalia viridis (non Muller) Ehlers 1913, p. 455. Day 1934, p. 30. Steggoa magalhaensi (non Kinberg) Augener 1931, p. 284. Records: LAM.18(1), 59(3), 63(1); FB.302(1), 326(1); FAL.23(1), 29(1), 44(2); 69(1), 106(1), 149(1), 245(1), 334(1); TRA.123(r). Notes : This species although generally similar to E. viridis differs in having a more flattened ventral tentacular cirrus on the second tentacular segment and in lacking setae on the same segment so that the formula is 1 + 07 “ Si", I It thus belongs to Bergstrém’s genus Steggoa here relegated to a sub-genus. On the other hand I cannot agree with Augener (1931) that it is identical with Steggoa magalhaest Kinberg which has spear-shaped dorsal cirri three times as long as broad. Eulalia (Hypoeulalia) bilineata (Johnston) 1840 (Fig. 5 df) Records: SB.197(1); FAL.371(1); MB.66(2), 87(2); SCD.40(3). Description: These South African specimens differ in several important respects from the descriptions given by Bergstré6m (1914), p. 165, text-fig. 57, and Fauvel (1923), p. 162, fig. 58 a—e, which themselves differ in minor respects. For this reason a full description is given below. The largest of the 3 specimens from SCD.40 measures 20 mm. by 1 mm. for 155 segments. It is a slender yellowish worm with two dark green stripes along its back just above the parapodia. These colours persist in alcohol. THE POLYCHAET FAUNA OF SOUTH AFRICA 301 The prostomium (fig. 5d) is rounded in front and almost straight posteriorly. The frontal antennae are well developed and the median antenna arises well in front of the eyes which are relatively large. The proboscis is covered with small conical papillae. The relation between the prostomium and the first tentacular segment is difficult to ascertain. In contracted specimens the two seem to be fused but in expanded specimens the first tentacular segment seems to be reduced dorsally but not actually fused to the prostomium. The second and third tentacular segments are definitely distinct. The three dorsal tentacular cirri are cylindrical and those on the second and third segment are rather long; on the other hand V, is short and flattened though not bladelike as in the sub-genus Sige. The second and third tentacular segments have : I I setigerous lobes with setae so that the tentacular formula is 1 + S— + S—. I The body segments do not change appreciably along the length of the worm. The dorsal cirrus (fig. 5¢) is bluntly lanceolate in adults but distinctly broader, almost cordate in juveniles. The setigerous lobe is bluntly rounded and, as usual, the presetal lip is deeply notched. The ventral cirrus is ovoid. ‘The 12- 16 setae are heterogomph spinigers (fig. 5f) with ovate shaft-heads striated distally and lightly serrated blades of normal length. According to Bergstrém, Hypoeulalia bilineata has the first segment fused to the head, V, is cylindrical and the dorsal cirri are ovoid. According to Fauvel (1923) Eulalia bilineata has the first segment narrowed but fairly short and the dorsal cirri are stout and oval-obtuse. Judged by these descriptions the most ‘marked difference between South African and European specimens concern the shape of V,. This is a new record for South Africa. Eulalia (Eumida) sanguinea (Oersted) 1843 Eulalia (Eumida) sanguinea. Fauvel 1923, p. 166, fig. 59 f-k. Records: LAM.22(1); LB.161(2), 382(1); FB.316(1); FAL.70(3), 376(1); MB.86(1); SCD.9(3), SCD.109(1). Eulalia (Pterocirrus) macroceros Grube 1860 (Fig. 5 g-4) non Sige macroceros Bergstrém 1914, p. 136, text fig. 40. Eulalia (Pterocirrus) macroceros Fauvel 1923, p. 167, fig. 60 d—g (partim). Eulalia (Pterocirrus) ?macroceros Day 1953, p. 411. Records: MB.66(1), 86(1). Description: The discovery of two further specimens allows me to confirm my previous identification, and to complete the description. The body is broad and short and dark green in life but brownish in alcohol. The prostomium (fig. 5g) is cordate with a large posterior excavation containing a dark cushionlike lobe which may represent the dorsal remnant of the first 302 ANNALS OF THE SOUTH AFRICAN MUSEUM tentacular segment. The eyes are large and there are usually pigment granules behind them. The two pairs of frontal antennae are surprisingly long and the dorsal antenna arises slightly in front of the eyes. The proboscis has not been seen everted but on dissection the base proves to be smooth but further along there are large soft rugosities. It is certainly not densely covered with cylindrical papillae as in Eulalia viridis or E. capensis. The first tentacular segment is fused to the prostomium and bears a pair of cylindrical tentacular cirri. The second segment is distinct and bears a pair of long cylindrical dorsal tentacular cirri on swollen cirrophores, a pair of flattened, often blade-like ventral cirri but there is neither setigerous lobe nor setae. The third tentacular segment bears a long cylindrical dorsal cirrus, a normal foliaceous ventral cirrus but no setigerous lobe or setae. The tentacular formula is thus 1 + lia + Ges. I N The parapodia (fig. 52) are very similar throughout the length of the body. The dorsal cirri are elongate-cordate and pointed, the setigerous lobe has the usual bilobed presetal lip but here the larger superior lobe is pointed and the ventral is best described as orbicular with a pointed end. The setae are very numerous (c.40). Each has a very slightly expanded shafthead faintly striate distally (fig. 52) and a blade of normal length which is broad and almost smooth basally and then suddenly narrows and becomes strongly serrated. As will be seen, the above description does not agree with that given by Bergstr6ém (1914) and differs in several respects from that given by Fauvel (1923). In a private communication Dr. K. Banse has informed me that my (1953) description agreed almost exactly with his specimen from Naples, and went on to say that EF. macroceros from the Mediterranean (type locality Quarnero in the Adriatic) is not synonymous with the boreal species E. (Sige) fusigera Malmgren (1865) (described from Koster-Inseln and Skelderviken in Sweden and Drobnak in Norway). Bergstrém’s description is based on the Swedish material and should be referred to E. (S.) fusigera. Fauvel presumably had more than one species before him. Some doubt remains regarding the subgenus to which EF. macroceros should be referred. It does not fit exactly into any of the numerous genera used by Bergstrém, and certainly cannot be referred to Sige which has setae on the second and third tentacular segments. It is suggested here that it should be referred to Claparéde’s Pterocirrus established for P. velifera Claparéde (1865) from Naples which according to Grube (1880) is synonymous with E. macroceros. If this be accepted Pterocirrus should be defined as a subgenus of Eulalia with the first tentacular segment fused to the prostomium which has a posterior excavation containing a cushion-like lobe. The second and third tentacular segments are distinct, and tentacular cirrus V, is flattened. There are no , j ! I setae on any of the tentacular segments, formula being: 1 + o__ + 0—. I N THE POLYCHAET FAUNA OF SOUTH AFRICA 303 Eulalia (Sige) falsa n. sp. (Fig. 6 a—c) Records: FAL.187(2), 338(1); MB.87(1). Description: ‘The type material consists of the two specimens from FAL.187 dredged in False Bay. One is a regenerating individual 17 mm. long by 1-5 mm. wide with 60 segments and an everted proboscis. It is brownish in alcohol. The other is a juvenile 9 mm. long by 1 mm. wide with 60 segments. A fresh specimen (FAL.338) has a brown dorsum with a paler head and para- podia. The body is elongate and of the usual proportions. The prostomium (fig. 6a) is cordate with a pair of large eyes, subulate frontal antennae and a median antenna which arises between the eyes. The first tentacular segment is fused to the head but there is an area between the posterior lobes of the prostomium which may represent the dorsal remains of this segment, or an anterior projection of segment 2. The second and third tentacular segments are distinct and bear long, dorsal, tentacular cirri. Tentacular cirrus V, is flattened and may even be blade-like in juveniles. Both the second and third tentacular segments bear setae but those on the second segment are very few and arise from the ventral cirrophore, there being no separate setigerous lobe on this segment. A small setigerous lobe with numerous setae is present on the third segment and below this is the foliaceous ventral cirrus, the tentacular formula being: 1 + ou + S_". The extruded proboscis is faintly hexagonal I N with 6 low longitudinal ridges. The surface is covered with very small and poorly marked papillae so that on first inspection it appears to be smooth. The opening is encircled by 20 large rounded papillae. The parapodia of the adult (fig. 65) are very similar throughout. The dorsal cirrus is elongate-cordate with a pointed apex, and in specimen MB.87 there is a dark central spot. The setigerous lobe has a bilobed presetal lip, the superior lobe being large and pointed and the inferior lobe being small and blunt. In the juvenile specimen these characters are more marked and the superior lobe has a very long pointed lobe. The setae (fig. 6c) are numerous fine heterogomph spinigers with very slightly expanded shaft-heads bearing about 4. small denticles at the distal end. The blades are smoothly tapered and finely serrated. This species is rather similar to Bergstrém’s description of Sige macroceros which, as shown above, really refers to E. (S.) fusigera. However there are differences in the shape of the prostomium, the nature of the proboscis, the shape of the cirri and the structure of the setae. Later work may show that this South African material is conspecific with E. (S.) fusigera from northern Europe, but to avoid further confusion in the synonymy it is as well to keep them separate at present. 304 ANNALS OF THE SOUTH AFRICAN MUSEUM ove ig tee RRO E HAIN Pine ; Fic. 6. Eulalia falsa: a head and proboscis; 6 anterior view of parapodium; ¢ seta. Protomystides capensis: d seta; e anterior end; f anterior view of parapodium. Syllis trifalcata: g head; h parapodium; i seta. THE POLYCHAET FAUNA OF SOUTH AFRICA 305 Eulalia trilineata Saint Joseph 1888 Eulalia trilineata. Fauvel 1923, p. 162, fig. 57m. Eulalia near albopicta Day 1951, p. 20. Eulalia near trilineata Day 1953, p. 410. Records: LAM.5(1), 10(1), 22(1), 35(1); FAL.22(2), 27(3), 81(1), 103(2), 113(7), 128(1), 131(1), 145(2), 152(1), 219(1), 245(1), 260(2), 262(2); © LIZ.29(1). Notes: The numerous specimens now available allow me to confirm the identity of this common South African species with Fauvel’s brief description. Apart from the colour pattern the setae are quite characteristic. There are relatively few setae (10—15), the shaft-heads are markedly swollen and the blades are very short and strongly tapered. In 1951 I suggested that a specimen from Port St. Johns was close to E. albo-picta Marenzeller from Japan. Since then I have been able to consult Izuka (1912) who gives an excellent description of this species. It has tentacular cirrus V, flattened and there are setae on the second tentacular segment. In F. trilineata V, is almost cylindrical and there are no setae on the second tentacular segment. Notophyllum splendens (Schmarda) 1861 Notophyllum splendens. Day 1953, p. 4.08, fig. 2 h-k. Records: LAM.44(1); TB.302(1), 303(1), 309(1), 310(1); TRA.122(1), 143(1); FAL.27(1), 31(p), 56(1), 80(p), 149(1), 162(1). Eteone foliosa Quatrefages 1865 Eteone foliosa. Fauvel 1923, p. 174, fig. 62 g-k. Day 1953, p. 411. Records: SB.183(3), 189(1), 195(1); LB.323(1); TB.go1(1); FB.302(1); FAL.110(1); FAL.113(3); FAL.228(1); ?FAL.375(2). Eteone (Mysta) syphodonta (Delle Chiaje) 1822 Eteone (Mysta) siphonodonta. Fauvel 1923, p. 178, fig. 63 e-A. Records: FAL.349(1); TRA.113(1); SCD.61(1). Notes: The body is brown to mauve dorsally and pale ventrally. The prostomium is white, bluntly triangular and depressed with a pair of eyes which are visible through the skin. The two pairs of antennae are subequal and rather slender. ‘The tentacular cirri are equal. The first normal segment lacks a dorsal cirrus but has a small ventral cirrus and a setigerous lobe with several setae. The dorsal cirri are 1-5-2 times as long as broad and borne on rather long broad cirrophores. The ventral cirri are bluntly pointed and a little longer than the blunt setigerous lobes. There are 15-20 setae with fairly long, evenly tapered blades and shaft-heads which are asymmetrical having one large tooth and 3-5 denticles. The proboscis when dissected proved to have two ventro-lateral 306 ANNALS OF THE SOUTH AFRICAN MUSEUM rows of large triangular papillae, a broad brownish dorsal band of minute flattened and denticulate papillae and a much narrower ventral band of slightly larger globular papillae. This is the first record from South Africa but the above description agrees very well with that of Fauvel with the exception that the dorsal cirri are a little longer. Eteone sp. Records: TRA.108(1). Notes: Specimen ‘TRA.108.K is an unidentified species of Eteone new to South Africa. It is dirty white in alcohol and 15 mm. long. The prostomium is unusually long and slender and somewhat reminiscent of a Glycerid. It is tapered and over twice as long as the basal breadth with two pairs of stumpy antennae one behind the other and a pair of well-developed eyes posteriorly. Behind the eyes the head swells out to encompass the proboscis which had been lost. The two pairs of tentacular cirri are very small; the dorsal pair is no more than an elongate papilla and the ventral is only one-third the breadth of the tentacular segment. The next segment has only a ventral cirrus there being neither a dorsal cirrus nor setigerous lobe nor setae. All the parapodia are small. The dorsal cirrus is roughly semicircular and no broader than its cirrophore. The setigerous lobe is rather elongated with a blunt apex and the ventral cirrus is ovoid. There are about 10 setae per bundle each having an asymmetrical shaft-head with a large tooth on one side and a minute one borne on a projecting lobe on the other. The blade is broad basally but tapers rapidly to a slender tip. This species is definitely new to South Africa and the slender prostomium and tiny tentacular cirri suggest that it may be a new species but until the nature of the proboscis is known it is not advisable to give it a specific name. Protomystides capensis n. sp. (Fig. 6 d-f) Records: TRA.86(1); WCD.28(1). Description: The holotype is a slender orange worm from TRA.86 richly speckled with red. It is 17 mm. long by 0-7 mm. wide with 110 segments, and is well tapered at each end. The prostomium (fig. 6¢) is small and cordate, a little longer than broad with two pairs of slender antennae and a pair of laterally placed eyes. Dissection showed that the proboscis had been lost. There are 3 pairs of small subulate oS, I I tentacular cirri on three segments according to the formula 1 + S— + S_. N The first tentacular segment is fused to the prostomium and its cirrus is cylindrical and markedly tapered distally. The second segment is broad and THE POLYCHAET FAUNA OF SOUTH AFRICA 307 distinct, its dorsal cirrus is oval in section and about 1-5 times as long as the prostomium. The setigerous lobe is well developed and bears several setae. The ventral cirrus (V,) is quite definitely similar to those of normal body segments. The third tentacular segment is also distinct but rather narrow and its cylindrical dorsal cirrus is markedly tapered and only two-thirds the length of that on the second segment. The setigerous lobe, setae and ventral cirrus are similar to those of the succeeding body segments. Normal body segments are depressed and the dorsal cirri are well to the sides so that most of the back is uncovered. Each dorsal cirrus (fig. 6f) is symmetrically cordate and about as broad as long. The setigerous lobe is rather long and has a simple blunt apex, the presetal lobe not being notched as is the case in most Phyllodocids. The ventral cirrus is ovoid and possibly a little longer than the setigerous lobe. There are about 12-18 compound setae (fig. 6d) with swollen, symmetrical and almost truncate shaft-heads which bear a series of very fine subequal teeth distally. The blade is short and strongly tapered. Protomystides is a rare genus and as far as I am aware no species has been described from the southern oceans, certainly none has been described from South Africa. Bergstrém (1914) describes P. bidentata from the North Atlantic and Mediterranean as having all the tentacular segments fully developed, and free from the prostomium. Family HESIONIDAE Syllidia armata Quatrefages 1865 Magalia perarmata Mar. et Bobr., Fauvel 1923, p. 246, fig. 92. Records: SB.115(1), 183(3), 184(1), 207(1); TRA.86(1), 88(1); FAL.31(1), 43(2), 136(6), 145(1), 164(2), 174(1), 266(1), 275(1), 283(1); MB.88(1); LIZ.9(2). Kefersteinia cirrata (Keferstein) 1863 Kefersteinia cirrata. Fauvel 1923, p. 238, fig. 89 a-e. Records: FAL.283(4). Notes: This is a new record for South Africa, but the characters agree very well with Fauvel’s description. Family SYLLIDAE Syllis (Haplosyllis) spongicola Grube 1855 Syllis (Haplosyllis) spongicola. Fauvel 1923, p. 257, fig. 95 a-d. Records: AFR.707(1), 842(1); TRA.151(1); MB.16(2); SCD.54(1). 308 ANNALS OF THE SOUTH AFRICAN MUSEUM Syllis (Haplosyllis) trifalcata n. sp. (Fig. 6 g—z) Records: FAL.216(1). Diagnosis: The dorsal cirri have 8-12 joints, and the setae have 3 falcate teeth. Description: ‘The holotype is 9 mm. by 0-4 mm. with 88 segments. There are no colour markings. The head (fig. 6g) is broader than long with rather flattened palps bent ventrally but not united at the base. The antennae are subequal and rather short. There are 4 eyes. The pharynx has an anterior dorsal tooth and extends back to setiger 9 and the cylindrical proventriculus with 40 rows of points then extends on to setiger 16. The tentacular cirri and dorsal cirri are short, tapered and twist like pigs’ tails; they have 9-12 well-marked joints. The setigerous lobes (fig. 6h) are obliquely truncate cones and the ventral cirri are small. There are 2 acicula with blunt tips. Each parapodium contains 3-6 simple setae (fig. 62) which are all similar, each having an expanded end (corresponding to the shaft-head) bearing 3 claw-like teeth of about the same size. The common S. (#.) spongicola Grube has dorsal cirri with more joints and the setae are roughly like boat-hooks with 2 small teeth above a large triangular rostrum. S. (H.) depressa Augener 1913 from Australia has setae with only 2 teeth rather like the open beak of a bird. S. (H.) abberans Fauvel 1919 from Indochina is fairly close but the dorsal cirri are long and apparently not jointed; moreover the setae are narrowed before the apex which has teeth approaching those of S. (H.) spongicola. Syllis vittata Grube 1840 Syllis vittata. Fauvel 1923, p. 263, fig. 98 7-1. Day 1953, p. 412. Records: LAM.22(1); FAL.82(1), 134(1), 171(2); MB.66(r). Notes: Specimens LAM.22.W and FAL.171.Z are doubtfully referred to S. vittata. The body is creamy white without markings and rather stout. The pharynx is short. Dorsal cirri have about 20 joints. The setae always have a very small secondary tooth and in the middle of the body they tend to be short and hooked. Syllis variegata Grube 1860 Syllis variegata. Fauvel 1923, p. 262, fig. 97 h—n. Day 1953, p. 412. Records: LAM.22(1); SB.189(1), 197(2); WCD.8(2), 19(4); FB.307(4); FAL.8(p), 113(2), 128(1), 131(1), 134(1), 145(2), 156(5), 162(12), 174(2), 178(2), '302(1),'303(1);) MB.86(1)5 LIZ.2(1) (ny 20K n)). 3 THE POLYCHAET FAUNA OF SOUTH AFRICA 309 Syllis prolifera var. zonata (Haswell) 1886 Syllis zonata Augener 1918, p. 236, pl. 4, fig. 86; pl. 5, fig. 107, text-fig. 19. Records: LAM.25(1), 31(1), 44(1), 47(1), 57(1), 59(1); SB.197(1); TB.332(1); FAL.31(8), 81(p), 103(1), 110(p), 128(1), 134(1), 145(2), 149(3), 156(3), 162(20), 166(2), 171(19), 219(1), 275(2), 280(1); LIZ.18(1). Notes: Like S. variegata this species has dorsal cirri with 25-35 joints and strongly bidentate setae; it differs in having a short pharynx and in having two narrow black lines across the anterior segments where S. variegata has a pattern of broken brown bars. Syllis armillaris Miller 1776 Syllis armillaris. Fauvel 1923, p. 264, fig. 99 af. Day 1953, p. 412. Records: Lamberts Bay 13 records from 17-23 metres on rock. Common. SB.207(1); TB.305(6), 306(1), 308(2), 309(1), 331(4); SH.168(2), 204(1), 366(1), 415(2); WCD.8(2); False Bay—27 records from o—33 metres on rock (common). AFR.835(1), 967(1); MB.13(1); LIZ.18(1), 36(2), 37(2); SCD. 9(2), 22(1), 106(1). Syllis gracilis Grube 1840 Spllis gracilis. Fauvel 1923, p. 259, fig. 96 f-i. Day 1953, p. 412. Records: FAL.17(2), 50(1), 166(2), 275(2), 280(1); MB.86(1); LIZ.2(1). Syllis hyalina Grube 1863 Syllis capensis McIntosh 1885, p. 193, pl. 33, figs. 8-9; pl. 15A, fig. a1. Syllis hyalina. Fauvel 1923, p. 262, fig. 98 a—b. Records: MB.57(1), 86(1). Notes: The type of Syllis capensis from the Cape is now in the British Museum. It is a small worm and probably immature. The dorsal cirri are cylindrical not fusiform and have 13 joints anteriorly, 11 in the middle of the body and 10 posteriorly. There are 6-10 strongly bidentate setae per foot. The specimens from Mossel Bay (MB.57 and 86) are referred to S. hyalina with some hesitation for the dorsal cirri have 15-20 joints. The setae are strongly bidentate but in superior setae the two teeth are very close together and project at right angles to the blade somewhat in the fashion shown by Gravier for S. bouviert. Syllis cf. trapobanensis Willey 1905 (Fig. 7a) Typosillis taprobanensis Willey 1905, p. 268, pl. 3, figs. 77, 78. Records: TRA.55(4); SCD.54(1). 310 ANNALS OF THE SOUTH AFRICAN MUSEUM Notes: There are faint transverse bars across the anterior segments when fresh but these soon fade in alcohol leaving the body white. The body is of the usual size (16 mm.) and shape. The palps are rather flattened and short but separate at the base. The pharynx is strongly chitinised and the dorsal tooth is rather small. There is no sign of an occipital flap. The dorsal cirri are rather long and markedly tapered with 20-30 joints. The setigerous lobes are stout and the ventral cirri slender. The setae (fig. 7a) are characteristic and all similar. There are 10-12 per bundle and each has a swollen shaft-head with a short almost triangular blade with two large blunt teeth. Syllis (Langerhansia) anops Ehlers 1897 Syllis (Ehlersia) anops Ehlers 1897, p. 40, pl. 2, figs. 40-45. Records: FAL.248(1). Notes: A single incomplete specimen was obtained measuring 16 mm. for 45 segments. It is a threadlike worm with slender dorsal cirri. The prostomium lacks eyes, has large palps fused at the base and short antennae. The tentacular cirri are a little longer but still shorter than the anterior dorsal cirri. The pharynx has the dorsal tooth at its anterior margin and stretches back to setiger 9. The proventriculus is also long and extends back from setiger 9 to setiger 18. The uniformly slender dorsal cirri are about as long as the body is broad and have 20-25 joints. The setigerous lobe is in the form of an obliquely truncate cone and the ventral cirrus is slender and a little longer than the foot. The setae are similar throughout. In each foot there are 2—3 superior setae with very long tapering swordlike blades which give the impression of having minutely knobbed tips. Below these there are about 12 setae with unidentate blades of normal length. The South African specimen agrees with Ehlers’ species from the Magellan area in all respects save one. Ehlers states that the anterior setae are bidentate and his figure (pl. 2, fig. 44) shows both the Ehlersza type and the normal setae each with a small secondary tooth. In my specimen all setae agree with his pl. 2, fig. 45, which shows the blades of both types of setae with unidentate tips. This is a new record for South Africa. Syllis (Langerhansia) ferrugina Langerhans 1881 Syllis (Ehlersia) ferrugina. Fauvel 1923, p. 269, fig. 100 k—u. Records: SB.183(3), 189(1); FAL.149(1). Notes: This is the first record from South Africa but Augener (1918) recorded it from Angola. Syllides longocirrata Oersted 1845 Syllides longocirrata. Fauvel 1923, p. 284, fig. 108 a-g. Records: FAL.65(1), 82(1). Notes: The body is small and the pharynx lacks teeth. The antennae, tentacular cirri and the first two pairs of dorsal cirri are unjointed but the THE POLYCHAET FAUNA OF SOUTH AFRICA 311 Fic. 7. Syllis cf. taprobanensis: a seta. Trypanosyllis ankyloseta: b seta; ¢ entire animal; d anterior end; e middle parapodium. Lamellisyllis comans: f anterior end; g head with left antenna and dorsal cirri removed; h parapodium; 7 seta. 312 ANNALS OF THE SOUTH AFRICAN MUSEUM remaining dorsal cirri are very long and deeply annulated with 12-15 joints. The setae have long tapered blades with indistinct tips which are possibly unidentate. This is a new record for South Africa. Trypanosyllis zebra Grube 1860 Trypanosyllis zebra. Fauvel 1923, p. 260, fig. 101 a-e. Day 1953, p. 413. Records: FB.322(1); LIZ.o(1). Trypanosyllis gemmulifera Augener 1918 Trypanosyllis gemmulifera Augener 1918, p. 278, pl. 5, figs. g9-101, text-fig. 27. Day 1953, p. 413. Records: LAM-4(¢), 1o0(1), 15(2), 18(3), 22(3), 33(1)5)@ 9) ayaa 5B.132(2); LB.161(1); TB.306(2); SH.324(1); AFR.842(1); TRAR@os(3) r10(1); WCD.8(3); FAL.57(p), 103(p), 106(p), 145(1), 140(5)aumy olay. 184(1), 216(2); MB.g(1), 13(1), 16(1), 53(2), 67(3), 77(1), 86(1); LIZ.29(3). Trypanosyllis prampramensis Augener 1918 Trypanosyllis prampramensis Augener 1918, p. 276, pl. 4, figs. 91, 92, text-fig. 26. Day 1953, p. 414. Records: FAL.156(1). Trapanosyllis ankyloseta n. sp. (Fig. 7 b-e) Records: FAL.216(1). Diagnosis : A short broad body, dorsal cirri with 6-8 joints, simple setae with the blade fused to the shaft-head. Description: ‘The holotype is the single specimen dredged in False Bay at 32°12-4'5/18°43°5'E at a depth of 42 metres on a sand and rock bottom. The specimen (fig. 7c) is very broad and short and markedly flattened. It measures 8 mm. by 1°8 mm. and is roughly oval with about 120 segments. The colour is yellowish white in alcohol. The prostomium (fig. 7d) which is sunk in between the anterior segments, is rectangular with 4 large eyes, a pair of ovoid palps directed ventrally and 3 short antennae. The lateral pair are anterior in origin and have 6 joints while the median arises from the centre of the prostomium and has 8 joints. The tentacular cirri are borne on anteriorly directed projections arising between the prostomium and the first pair of parapodia and the tentacular segment is not visible dorsally. The dorsal pair of tentacular cirri are longer than the ventral pair and are about equal to the dorsal cirri of setiger 1. The mouth is ventral and the long pharynx is folded on itself in the dorso-ventral plane. The trepan has about 10 teeth. The proventriculus which has about 40-50 rows of points extends from setiger 16-26. THE POLYCHAET FAUNA OF SOUTH AFRICA 313 Anterior segments increase in width until an average segment in the middle of the body is about 20 times as broad as long. Posterior ones decrease again as the body tapers to an oval anterior end. The parapodia (fig. 7e) are similar throughout. Each has a short dorsal cirrus of 6 to 8 joints borne on a broad projecting cirrophore. Below this is the setigerous lobe with a terminal papilla and below this again is the somewhat shorter ventral cirrus. The 4-5 setae (fig. 7b) are not compound but simple since the falcate unidentate blade has become fused to the shaft-head. The posterior end tapers to a bilobed pygidium bearing a pair of ovoid anal cirri with 3-4 joints. Odontosyllis polycera (Schmarda) 1861 Odontosyllis polycera. Augener 1918, p. 283, pl. 5, fig. 97. Day 1953, p. 415. Records: LAM.31(2); FAL.23(p), 31(1), 50(3), 82(2), 104(p), 113(2), 164(1), 365(1); MB.57(2), 87(4). Odontosyllis ctenostoma Claparede 1863 Odontosyllis ctenostoma. Fauvel 1923, p. 277, fig. 104 f-l. Records: ‘TRA.121(1). Notes: ‘This species has been recorded from Angola by Augener (1918) but this is the first record for South Africa. Pharnygeovalvata natalensis Day 1951 Pharyngeovalvata natalensis Day 1951, p. 26, fig. 4 e-j. Records: FAL.171(1). Notes: ‘The single specimen is incomplete but measured 9 mm. for 48 segments and is thus larger than the type. It has a general resemblance to Odontosyllis ctenostoma but the structure of the pharynx is characteristic. Amblyosyllis lineolata (Costa) 1864 Pterosyllis formosa Clap. Fauvel 1923, p. 280, fig. 105 h-n. Amblyosyllis lineolata Day 1953, p. 415. Records: FAIL.136(2), 159(1), 162(1), 171(1). Pionosyllis ehlersiaeformis Augener 1913 Pronosyllis ehlersiaeformis Augener 1913, p. 225, pl. 3, fig. 32; text-fig. 31 a-e. non Pionosyllis ehlersiaeformis Day 1953, p- 415, fig. 3d. Records: WCD.13(8); FAL.269(4). Notes: The material consists of four ovigerous females 4-8 mm. long living in mucus tubes attached to hydroids. Two of them had developing embryos on their backs. 314 ANNALS OF THE SOUTH AFRICAN MUSEUM The prostomium bears 3 antennae on its anterior margin and the median, which is the longest, is twice the length of the prostomium. There are 4 large eyes and broad palps which are united basally and bent ventrally. The tentacular segment is short and distinct from the prostomium. It bears two pairs of tentacular cirri, the longer dorsal pair being equal to the median antennae. The dorsal cirri are smooth and tapered. The pair on the first setiger are about 1°5 times the breadth of the body but succeeding ones are shorter and over most of the body the dorsal cirri are only two-thirds the breadth of the body. Some of the cirri are wrinkled but none are annulated. There are 10-12 setae per foot. Those of the first foot are stouter than the rest and at certain angles the blade almost appears fused to the shaft but in the following feet the majority of the setae are clearly compound with bidentate blades. A single slightly curved, simple needle-like seta appears on the roth foot and a little later two Ehlersia-type compound setae with very long slender blades. Beneath these are several compound setae with shorter blades with two large terminal teeth. An inferior simple seta was not seen but in two specimens natatory setae with immensely long blades were found in posterior feet. Augener’s descriptions of the dorsal cirri are not quite consistent for in the first account (Augener 1913) he states that the dorsal cirri are not ringed and in the second (Augener 1918) he says that they are. The tendency for transverse wrinkling or indistinct ringing is common in the genus and it is possible that the different descriptions are due to variations in methods of preservation. Augener (1918) suggested that P. malmgreni described by McIntosh (1904) from False Bay as having 20-30 annulations to the dorsal cirri was synonymous _with his P. ehlerstaeformis. In Day (1953) I followed Augener but noted that my specimen did not have true Ehlersia-type setae. I now feel that they should be kept separate for the Cape form of P. malmgreni lacks the Eflersia-type superior setae, has 20-30 indistinct joints to the dorsal cirri and grows to a much larger size (40 mm. as against 8 mm.). Pionosyllis cf. longocirrata St. Joseph 1887 Pionosyllis cf. longocirrata. Fauvel 1923, p. 288, fig. 110 h-l. Pionosyllis sp. Day 1953, p. 418, figs. 3 ef. Records: SH.204(2), 415(1). Notes: These specimens show many similarities to Fauvel’s description but the identification remains uncertain. All three individuals are soft and fragile and have broken in several places but might measure 15 mm. if complete. The prostomium has 2 pairs of eyes and the anterior pair are larger and further apart. The palps are fused at their bases and bent ventrally. The pharynx is short with a smooth rim and a fairly large anterior tooth. The proventriculus THE POLYCHAET FAUNA OF SOUTH AFRICA 315 has about 30 rows of points. The dorsal cirri are very long, smooth, tapered and over twice the breadth of the stout body. The anterior ventral cirri are large and triangular but not lamellar and further back they become more digitiform. The setae are very long and fine. The blades themselves are slender, and not obviously tapered. The tips are bidentate and as shown in Day (1953) figs. 3 e-f, there is a strong hooked terminal tooth with a slender tooth directed obliquely towards it. There are also faint indications of a hood over the terminal tooth. No simple setae were seen. The European P. longocirrata has even longer dorsal cirri, up to 4 times the body breadth but Fauvel’s figure 110h does not suggest that they are tapered and his figure of the seta (110/) does not suggest a hooked terminal tooth. Pionosyllis magnidens Day 1953 Pionosyllis magnidens Day 1953, p- 416, fig. 3 a—c. Records: FAL.132(2), 174(3), 178(1). Grubea furcelligera Augener 1913 Grubea furcelligera Augener 1913, p. 256, pl. 3, figs. 20, 21; text-figs. 39. Records: FAL.275(2). Notes: The material consists of two females with natatory setae carrying developing embryos on their backs. They are colourless in alcohol. The best preserved measures 5 mm. in length and has 40 segments. The prostomium is rounded with a pair of palps which are square in front, fused for most of their length but with the distal ends free. There are 3 antennae all arising from the anterior margin of the prostomium. All are bottle-shaped with tapered ends and the median is twice the length of the laterals which are equal to the width of the body (not including the parapodia) at the level of setiger 1. There are 4 eyes. The pharynx extends back to setiger 4 with the dorsal tooth near the anterior end. The barrel-shaped proventriculus extends over a further 2 segments. The tentacular segment is clearly marked off from the prostomium but is very narrow and not very distinct from setiger 1. It bears the usual 2 pairs of tentacular cirri, of the same elongate subulate shape as the antennae. The dorsal pair which is twice as long as the ventral pair is equal to the median antenna. The dorsal cirri are similar to the antennae and the tentacular cirri but vary in length; the dorsal cirri of setiger 1 are ? the length of the antennae, those of setiger 2 are very short and hardly exceed the length of the setigerous lobe, those of setigers 3 and 4 increase again and the dorsal cirri of setigers 5 and subsequent segments are about half as long as the tentacular cirri or a little over half the width of the segment that bears them. The setigerous lobes are short truncate cones and the ventral cirri are rather stout and ovoid. The normal compound setae have swollen shaft-heads and very small, unidentate dagger-like blades. There is also a single pointed simple seta in the 316 ANNALS OF THE SOUTH AFRICAN MUSEUM superior part of each bundle. The natatory setae appear in setiger 9. Each has a long slender shaft from which arises a very fine hair-like tapered blade. This South African material lacks the small anterior third pair of eyes described by Augener and the antennae, tentacular cirri and dorsal cirri of setiger 1 are considerably longer than Augener describes, possibly due to differences in preservation. This is a new record for South Africa. Grubea rhopalophora Ehlers 1897 Grubea rhopalophora Ehlers 1897, p. 53, pl. 3, figs. 66-70. Augener 1918, p. 295, pl. 4, fig. 94. Records: FAL.17(1), 246(2); MB.85(1). Notes: The Cape material agrees perfectly with Ehler’s description and I can confirm that the compound setae are unidentate. ‘This species has previously been recorded by Augener (1918) from the shore at Swakopmund and shallow dredgings at Liideritzbucht in South West Africa. In his notes on individuals from Liideritzbucht Augener describes two with minute and truncate dorsal cirri containing fibrillar structures. These agree very closely with the description of Fauvel (1923) of Grubea pusilla (Dujardin). Grubea rhopalophora is generally similar to G. limbata Claparéde but differs in the fact that the antennae, and dorsal cirri are shorter with more swollen bases, the tentacular segment is more completely fused with the prostomium and the palps separate towards their extremities. Sphaerosyllis sublaevis Ehlers 1913 Sphaerosyllis sublaevis Ehlers 1913, p. 482, pl. 32, figs. 10-15. Records: SB.167(1); FAL.82(1); TRA.113(1). Notes: These specimens agree very well with Ehlers’ description. The body surface is smooth and there is no dorsal cirrus on setiger 2. ‘The antennae, tentacular cirri and anterior dorsal cirri are all small and flask-shaped with swollen bases and tapered ends but further back the dorsal cirri become longer and more bottle-shaped. ‘The setae have unidentate blades. Ehlers states that there is a third minute pair of eyes on the anterior margin of the prostomium. Their absence in the present specimens is not regarded as important as it has been noted that such eye-specks are often invisible in individual specimens of Syllids. S. sublaevis is close to S. claparedit Ehlers (1864) but the latter is reported to have a dorsal cirrus on setiger 2. Sphaerosyllis hystrix Clap. var. capensis Day 1953 Sphaerosyllis hystrix Day 1953, p. 420, fig. 4 g-l. Records: SB.183(2). THE POLYCHAET FAUNA OF SOUTH AFRICA 317 Exogone clavator Ehlers 1913 Exogone clavator Ehlers 1913, p. 485, pl. 33, figs. 1-6. Day 1953, p. 418. meoras =) WOOD .5(1) >, 5b.183(4); FAL.17(7), 110(1),. 131(6), 152(1), 159(2), 266(1); MB.57(1); LIZ.20(1). Exogone gemmifera (Pagenstecher) 1862 Exogone gemmifera. Fauvel 1923, p. 305, fig. 117 a-d. Exogone verugera (non Claparéde) Day 1953, p. 418. moms oh.107( 2); FAL:29(7), 82(2), 103(3), 128(1), 178(10), 280(1). Notes: In 1953 I referred several specimens to E£. verugera though it was noted that they lacked a dorsal cirrus on setiger 2. Examination of a great deal more material from both South Africa and Europe and the discovery of the typical E. verugera in South Africa has shown that the presence or absence of a dorsal cirrus on setiger 2 is a constant and important character. It is now possible to summarize the main differences between the three closely related species E. verugera, E. gemmifera and FE. heterosetosa, all of which occur in the Southern hemisphere. In £. gemmifera the palps are short and broad, the three antennae are of equal size and about the same length as, or a little longer than, the prostomium. The proventriculus is short, extending over 1-2 segments, and has 10-12 rows of points. ‘he superior compound seta has a long dagger-like blade. There is no dorsal cirrus on setiger 2. In E. heterosetosa the palps are short and the three antennae are about the same length as the prostomium. The proventriculus is rather long, extending over 3 segments and has 15 or more rows of points. The superior compound seta has a characteristically swollen shafthead and a short broad blade. There is no dorsal cirrus on setiger 2. In E. verugera the palps are rather long and tapered and the three antennae are equally minute and much shorter than the prostomium. The proventriculus is fairly long, extending over 2—3 segments and has 25-30 rows of points. The superior seta has a dagger-like blade. There is a dorsal cirrus on setiger 2. Exogone verugera Claparéde 1868 Exogone verugera. Fauvel 1923, p. 307, fig. 117, figs. m—r. Records: SH.400(1); FAL.162(13); SCE.54(1). Notes: The diagnostic characters of this species are given above. Autolytus charcott Gravier 1906 Autolytus charcoti Gravier 1906, p. 7, pl. 1, figs. 1, 2. ? Autolytus afer Ehlers, 1908b, p. 46. iecords.; VB.30e(1yinFAL.o497(1)5(LIZ.29(1). 318 ANNALS OF THE SOUTH AFRICAN MUSEUM Notes: Specimen TB.312 measures 15 mm. for 75 segments. The body has conspicuous black bands at the intersegmental junctions starting at setiger 1/2 then every junction to setiger 6/7, then misses 7/8 and 14/15 and thereafter is present on every fourth junction to the end of the body. On specimen LIZ.29 every intersegmental junction from 2/3 onwards is banded. Diverging nuchal epaulettes extend from the back of the prostomium to setiger 2. The trepan has 10 equal teeth. The antennae, tentacular cirri and the dorsal cirri of setiger 1 are stout and just longer than the width of the body, but thereafter the dorsal cirri decrease and in the middle of the body they are rather less than one-third the body width. The specimen described by Ehlers (1908) from Liideritzbucht as A. afer agrees in general characters, but as it was preserved in osmic acid no details of colour pattern are available. Autolytus tuberculatus (Schmarda) 1861 Autolytus tuberculatus. Augener 1918, p. 307. Day 1953, p. 421. Records: FAL.22(1), 82(7), 103(3),.113(1), 122(1), 145(2); 150( pees 171(2), 247(1), 290(5) 3-2 LIZ.56(1): Notes: Further material has shown that the length of the nuchal epaulettes is variable. They may reach setiger 6 or hardly reach setiger 4. Anterior dorsal cirri are unequal; those of setigers 1, 2, 4 and 6 are much longer than those of setigers 3, 5 and 7 or subsequent segments. In the middle of the body they are only one-third to one-half of the width of the body. Autolytus prolifer (Miller) 1788 Autolytus prolifer. Fauvel 1923, p. 311, fig. 119. Records: FAL.334(1); MB.58(1), 69(1); SCD.40(1). Notes: The pharynx is S-shaped and crowned with 10 large triangular teeth. Indistinct nuchal epaulettes are present on setiger 1 but do not extend on to setiger 2. The antennae, tentacular cirri and dorsal cirri of the first setigers are long but the remaining ones are only one-third to one-quarter the width of the body. Autolytus maclearanus McIntosh 1885 Autolytus maclearanus. Ehlers 1913, p. 488, pl. 33, figs. g—-11. Autolytus inermis (non St. Joseph) Ehlers 1913, p. 488. Records: SH.430(20); FAL.43(1); SCD.61(3). Notes: This South African material agrees very well with the species described by Ehlers from Kerguelen under the name of A. maclearanus. McIntosh’s original description is so vague that it might refer to any species of Autolytus, but Ehlers’s description and figures are clear. THE POLYCHAET FAUNA OF SOUTH AFRICA 319 The diagnostic features are the long antennae, dorsal tentacular cirri and dorsal cirri of setiger 1. All of these greatly exceed the width of the body and are often so wrinkled as to give the impression of being annulated which they are not. The dorsal cirri of normal body segments are about one-third of the body width. The tentacular segment is much shorter than setiger 1 and on some specimens it has vague indications of small nuchal epaulettes as stated by Ehlers. The pharynx has 6 rounded lappets instead of sharp chitinous teeth at its entrance. These were only seen when the pharyngeal sheath was dissected away and the pharynx was first thought to be unarmed as in A. inermis. The latter species, however, has a doubly convoluted pharynx while the present species has a single large loop. Ehlers stated that the proventriculus of his Kerguelen specimen lay in the 7th segment. Here it is in the 4th and has 30 rows of points. In fresh material it is faintly greenish. Myrianda phyllocera Augener 1918 Myrianida phyllocera Augener 1918, p. 301, pl. 4 figs. 87-89, text-fig. 30. Day 1953, p. 421. Records: FAL.178(1); LIZ.2(1). Lamellisyliis gen. nov. Prostomium with 3 lamellar antennae. Palps united at their bases. Pharynx straight with an anterior dorsal tooth. Prominent nuchal epaulettes. A single pair of cylindrical tentacular cirri. Dorsal cirri lamellar, setae compound, ventral cirri on all segments. Type species L. comans. Lamellisyllis comans n. sp. (Fig. 7-1) Records: FAL.110(1). Description: The holotype is a pale, flattened worm, roughly Harmothoid in outline and measures 8 mm. for 50 segments. The prostomium is sunk back between the anterior segments which project forwards and outwards so that the front end of the body (fig. 7/) appears rounded. The palps are normal and united only at their bases. The small rounded prostomium (fig. 7g) has 4 eyes set in a rectangle and 3 subequal foliaceous antennae. The lateral pair arise from the anterior margin while the median arises from the centre of the prosto- mium. The mouth is ventral and the straight, weakly chitinized pharynx bears a single dorsal tooth near its anterior margin. It extends back to setiger 7 and the barrel-shaped proventriculus with 20 rows of points extends on to setiger 12. ‘T'wo grooved, finger-like nuchal organs diverge from the posterior margin of the prostomium towards the sides of setiger 3. They were first thought to lie freely on the dorsum but attempts to move them showed that they are 320 ANNALS OF THE SOUTH AFRICAN MUSEUM attached throughout their length. A single pair of tapered and cylindrical tentacular cirri project forwards on either side of the prostomium. This pair corresponds to the ventral cirri of a normal segment for each arises from a lobe of the tentacular segment which is wedged between the prostomium and the first setiger and above it there is a lump which seems to correspond to a dorsal cirrophore. The normal body segments (fig. 7h) are all similar. Each is about 20 times as broad as long and has a dorsal cirrus, a setigerous lobe and a ventral cirrus on its lateral margin. The dorsal cirrus is borne on a stumpy cirrophore placed well above the setigerous lobe and the cirrus itself is flattened, oval to circular in outline, and is attached to the cirrophore by its edge. Alternate dorsal cirri are more medial and more lateral in origin. ‘The setigerous lobe is an obliquely truncate cone with a minute papilla at its apex. There is a pointed aciculum and about 20 compound setae (fig. 77) whose blades are strongly bidentate and ‘hairy’. The ventral cirrus is conical and slightly shorter than the setigerous lobe. The posterior end of the worm is markedly tapered and the pygidial segment bears a pair of foliaceous anal cirri. The possession of foliaceous appendages is unusual in the family Syllidae. The genus Myrianida has flattened head appendages and dorsal cirri but the completely fused palps, the sinuous pharynx crowned with a trepan of teeth, the minutely-bladed compound setae and lack of ventral cirri immediately place it in the sub-family Autolytinae. Phyllosyllzs Ehlers (1897) from South Georgia obviously belongs to the Autolytinae as well although it has setae on the tentacular segment. Knox (1957) has recently described Clavisyllis from New Zealand with inflated, ovoid dorsal cirri, normal ventral cirri, large nuchal epaulettes, palps united only at their bases and a straight pharynx with an anterior tooth. In these characters it is very similar to the present Lamellisyllis, but it differs in having cylindrical antennae, a nuchal cirrus between the nuchal epaulettes and two pairs of tentacular cirri. Possibly Clavisyllis should be included in the sub-family Eusyllinae. Lamellisyllis with its single pair of tentacular cirri shows resemblances to Sphaerosyllis, Exogone and Spermosyllis all of which belong to the sub-family Exogoninae. But all the members of the latter sub-family, have palps which are fused throughout their length so it might be better to place Lamellisyllis in a sub-family of its own. The characters of Clavisyllis and Lamellisyllis show that the family Syllidae is closer to the family Phyllodocidae than had previously been realized. Procerastea perriert Gravier 1900 Procerastea perriert. Fauvel 1923, p. 327, fig. 126 a-c. Records : SH.430(4); SB.167(LC). Notes: ‘This is a new record for South Africa but the specimens agree perfectly with Fauvel’s description. This species seems to feed on hydroids growing just below low-tide mark. ae a ee THE POLYCHAET FAUNA OF SOUTH AFRICA 321 Family NEREIDAE Laeonereis ankyloseta Day 1957 Laeonereis ankyloseta Day 1957, p. 83, fig. 5 a-y. Records : FB.302(3), 307(6); FAL.8(1), 43(1), 50(1), 80(p), 110(2), 126(1 I71(1), 225(4), 249(1), 275(1), 304(1), 345(1); MB.13(1), 16(1), a 56(heteronereid), 62(1 juvenile), 86(1), 87(1), 88(3); LIZ.1(1), 9(1); 8 32(3), 54(2)- Notes: MB.62 is a juvenile which lacks the characteristic ankylosed setae but the other characters are typical. MB.56 is a heteronereid. b) ) dh Nereis (Neanthes) operta Stimpson 1855 Nereis (Neanthes) operta. Day 1934, p. 38, fig. 5. Day 1951, p. 28. Day 1953, p. 424. Records: LAM.4(1), 8(1), 16(1), 22(2), 25(1), 44(t), 51(1), 59(1)s 6t(x), 63(1); SB.118(1), 175(2), 179(1), 181(1), 189(15); SH.366(1), 428(1), TB. 302(1), 325(1), WCD.23(1); TRA.69(2 planktonic heteronereids), False Bay: 24 records from 2-38 metres on sand and rock. MB.49(1), 62(1), 87(6), 88(1) ; LIZ.6(6). Nereis (Neanthes) willeyi Day 1934 Nereis (Neanthes) willeyi Day 1934, p. 38, fig. 6 a—c. Day 1951, p. 28. Day 1953, p. 424. Records : False Bay: 22 records from o—22 metres on sandy rocks. MB.40(4), 49(t), 59(1), ?71(1). Nereis (Neanthes) cf. kerguelensis McIntosh 1885 Nereis kerguelensis. Ehlers 1897, p. 65, pl. 4, figs. 81-93. Records: AFR.g50(1); 994(1); TRA.143(1). Notes: ‘The largest of the three specimens measures 22 mm. It is pale brown in alcohol with a more intense bar across setiger 2. The proboscis has group I = o; II = a wedge of 8-9 points; III = 5-6; 1V = a wedge of 10 points; V = 0; VI = 2-5 in a close set group; VII and VIII = a single row of 3-5. Anterior feet have three notopodial lobes and a rather longer dorsal cirrus. In posterior feet there are only two notopodial lobes. There are no notopodial falcigers. The neuropodial falcigers have straight blades with a tendon towards the tip. This South African material differs from published descriptions of WV. keurguelensis in having more denticles on group VI. In this and in the brown bar on setiger 2 it resembles WV. unifasciata Willey but the latter only has two notopodial lobes on anterior feet. Nereis (Neanthes) succinea Frey & Lueckart 1847 Nereis glandulosa Ehlers 1908a, p. 74, pl. 8, figs. 1-6. Records: TRA.91(3); LIZ.1(7), 3(10), 38(1). 322 ANNALS OF THE SOUTH AFRICAN MUSEUM Nereis (Nereis) lamellosa Ehlers 1868 Nereis lamellosa Ehlers 1868, p. 564, pl. 22, figs. 10-17. Fauvel 1936, p. 36. Records: TRA.33(3), .91(1); FB.306(6); FAL.184(1), 167(@) Seema: 209(1), 223(4), 240(2); MB.4(3), 34(2), 75(2); SCD.20(1), 61(1), 94(3), 105(1). Notes: As shown by Fauvel (1936) this species is very close to WV. succinea but has dorsal homogomph falcigers in posterior feet. All neuropodial falcigers have feathered blades with a tendon from the apical tooth. Nereis (Nereis) zonata var. persica Fauvel 1911 Neireis zonata var. persica Fauvel 1911, p. 385, pl. 19, figs. 10-16, 18-23; pl. 20, figs. 24-25. Records: SH.71(1); SCD.50(1), 63(1). Notes: This species which is well known from the tropical Indian Ocean has been found as far south as Mogambique (Day 1957). The present records show that it extends even further south in dredgings off the eastern Cape Province. It has also been found on the hull of a ship visiting Table Bay from the Indian Ocean. Nereis (Nereis) jackson’ Kinberg 1866 Nereis (Nereis) jackson. Fauvel 1932, p. 97. Records: L1Z.27(2); SCD.100(1). Nereis (Nereis) falsa Quatrefages 1865 Nereis falsa. Fauvel 1923, p. 337, fig. 129 e—m. Neries callaoana (non Grube) Augener 1918, p. 174 (partim). Records: SH.71(1). Notes: ‘The specimen is typical. Though known from the Natal coast it has not been recorded from the Cape. The present record is from the hull of a ship from India. Augener’s 4 specimens from Swakopmund labelled WV. callaoana (V.8782) were kindly sent to me by the Director of the Hamburg Museum. Three proved to be WV. falsa and one Platynereis dumerilit. Nereis (Nereis) eugeniae (Kinberg) 1866 Nereis eugeniae. Ehlers 1897, p. 67, pl. 4, figs. 94-105. Monro 1936, p. 136. Records: TRA.74(1), 80(3). Notes: The anterior part of the prostomium is free from the bases of the palps; the eyes are small and the tentacles are short. ‘The dental formula of the largest specimen is: I is 0; II is 3-4 in a line; III is 0; IV is 8-9 in a single to double row; V is 0; VI is a transverse group of 3; VII and VIII consist of 4 THE POLYCHAET FAUNA OF SOUTH AFRICA 323 widely separated points in a single row. All paragnaths are minute and as shown they are not numerous in these small specimens. Anterior feet have two dorsal lobes. In the posterior feet all lobes are pointed and the dorsal lobe is much larger than the rest. By contrast the dorsal cirrus is thin and delicate. Most posterior notopodia include 1-2 stout homo- gomph falcigers with very short conical blades hardly longer than the width of the shaft. Ventral falcigers of anterior feet have long straight blades; in posterior feet there are fewer but larger ones. Nicon eugeniae was first described by Kinberg from a specimen collected off Argentina opposite the Rio de la Plata. Most workers however refer to the description given by Ehlers (1897) for Nereis eugeniae based on specimens collected in the Magellan area and compared with Kinberg’s type. It is significant that neither author refers to homogomph falcigers in the posterior notopodia, and that Ehlers describes and figures the tentacular cirri of his specimen as being annulated and says that the tentacular cirri of Kinberg’s specimen were ‘gegen die Spitze hin deutlich gegliedert’. I have examined specimens in the British Museum identified by Monro (1930) and (1936) as Nereis eugeniae. None of these show annulated tentacular cirri and all have homogomph falcigers with short conical apices in the posterior notopodia. They are identical with my South African material. Kinberg’s types must be examined before the identity of Monro’s specimens and mine can be firmly established. NV. eugeniae is closely related to WN. trifasciata Grube 1878 from the Phillipines. In the original description there was again no reference to homogomph falcigers in posterior notopodia but Augener (1922) who states that he was not able to see Grube’s type, describes WN. trifasciata from Juan Fernandez Island as having notopodial falcigers in posterior feet with long, almost straight blades some 4-5 times as long as the width of the shaft. Fauvel (1932) refers to similar setae in his description of material from the Indian Ocean. Here again the type must be examined before the naming is certain, but it will be obvious that the main difference between WN. eugeniae (sensu Monro 1936) and WN. trifasciata (sensu Augener 1922) lies in the structure of the notopodial falcigers. Nereis (Nereis) sp. (Fig. 8a) Records: TRA.62(1). Notes : The material consists of a single complete specimen 8 mm. long with 40 segments. The prostomium is broadly rounded in front and the palps are large with small palpostyles; the tentacles are normal and rather short. There are no colour markings. The dental formula is I = 1; II is a wedge of close-set points; III is a few scattered points; IV is like II; V = 0; VI is a rosette of 8-10 points; VII and VIII is a continuous band consisting of 2-3 irregular rows. 324 ANNALS OF THE SOUTH AFRICAN MUSEUM Anterior feet have two notopodial lobes. In posterior feet the superior notopodial lobe is enlarged and flattened and carries the cirrus at its apex; the inferior notopodial lobe is a slender cone. Anterior notosetae are homogomph spinigers with short blades. Anterior neurosetae include homogomph spinigers with short blades and heterogomph spinigers. In middle feet there is a gradual change in the setae; the notopodial spinigers decrease in number and some of the neuropodial falcigers lose the articulation between shaft and blade. In posterior feet the notosetae include 2-3 homogomph falcigers with long straight blades and the neuropodial setae include 1-2 stout simple hooks (fig. 8a) formed by fusion of the straight pointed blade with the shaft, 2-3 normal heterogomph falcigers with fairly straight blades and 1-2 fine homogomph spinigers. As far as I can ascertain no species with similar setae has been described but the single specimen is a juvenile of 8 mm. and for this reason is not named as a new species. Nereis (Neanthes) caudata (Delle Chiaje) 1841 ? Nereis cricognatha Ehlers. Augener 1913, p. 163. Knox 1951, p. 217, pl. 45, figs. 6-8. Nereis (Neanthes) caudata. Fauvel 1923, p. 347, fig. 135 a-e. Day 1953, p. 425. Records: LB.161(1); SH.204(1), 366(1); MB.16(4). Notes: After studying the descriptions of Augener (1913) and Knox (1951) I see no reason why N. cricognatha Ehlers (1905) from New Zealand should not be included in the synonymy of N. caudata. Perinereis capensis (Kinberg) 1865 Perinereis capensis. Monro 1933, p. 495, figs. 7-11. Day 1934, p. 42, figs. 8 a-e. Records: SB.177 (1 juvenile); FAL.80(p), 122(p), 127(1), 159(2), 171(2), 174(1), 245(1); MB.o(1), 13(1), 16(2), 20(2), 40(4), 49(21), 53(18), 56(2), 62(1), 67(2), 74(2), 78(1), 85(1); LIZ.1(1), 6(4), 18(fc), 27(4). Pseudonereis variegata (Grube) 1856 Nereis (Mastigonereis) variegata McIntosh 1904, p. 37, pl. 1, figs. 6-10. | Records: FAL.126(1); TRA.69 (1 swimming in plankton). Notes: ‘This species which is so common between the tide marks has only been found on one occasion below low water. Platynereis dumerili (Aud. and M.-E.) 1833 Platynereis dumerilii. Fauvel 1923, p. 359, fig. 141 a—f. Day 1953, p. 429. Records: Lamberts Bay 14 records from 10-23 metres on rock (common) ; SB.114(c), 116(c), 122(p), 127(2), 129(3), 136(1), 180(8), 181(2), 183(3), 184(2), 194(4), 195(c); LB.155(c), 160(5), 161(4), 190(2), 380(4); SH.74(1); 4 THE POLYCHAET FAUNA OF SOUTH AFRICA 325 k Fic. 8. Nereis sp.: a simple seta from posterior neuropodium. Glycera benguellana: b jaw support; c papilla from proboscis; d head of compound seta; e posterior view of foot. Glycinde capensis: f4—> proboscidial papillae (f} plan view of row /, f?-> lateral view of rows 2-5); g anterior view of anterior parapodium; A anterior view of posterior parapodium. Glycinde kameruniana: j4—*> proboscidial papillae (j1 plan view of row, /, 72-5 lateral view of rows 2—5); k notoseta of posterior foot; / anterior view of anterior foot; m anterior view of posterior foot. 326 ANNALS OF THE SOUTH AFRICAN MUSEUM TRA.69 (abundant in plankton); TRA.107(1); False Bay: 33 records from o-36 metres on Algae or hydroids; (MB.4(1), 20(2), 27(4), 38(1), 40(6), 56(8), 50(1), 74(13), 86(2), 87(2); LIZ 1(2), 6(6), Watr)ieaiay: SCD.20(1). : Platynereis australis (Schmarda) 1861 Platynereis magalhaensis Kinberg. Fauvel 1916, p. 434, pl. 8, figs. 21, 22. Platynereis australis Day 1953, p. 429. Records: SB.130(7); LB.155(1), 472(3); SH.366(1); 415(1). Platynereis calodonta Kinberg 1866 Platynereis hewitti Day 1934, p. 44, fig. 9 af. Platynereis calodonta Day 1953, Pp. 429. Records: FAL.134(1); MB.40(1); LIZ.6(1), 27(1). Family SPHAERODORIDAE Ephesia gracilis Rathke 1843 Ephesia gracilis. Fauvel 1923, p. 377, fig. 148 a—f. Records: LAM.8(2), 31(1), 54(1), 59(2); SB.119(1), 183(1); TB.305(1), 330(1); TRA.58(1), ?102(3), 143(2); FAL. 223(1), 371(1), SCD.54(1). Notes: This is a new record for South Africa and the specimens have been checked as identical with European material. Family NEPHTHYDIDAE Nephthys (Nephthys) capensis Day 1953 Nephthys (Nephthys) capensis Day 1953, p. 431, text fig. 5 g—m. Records: LAM.52(3); LB.364(5); FAL.107(1), 209(1); TRA.104(1). Notes: In these larger dredged specimens the gills are often cirriform throughout so that there is a strong superficial resemblance to WV. hombergi. However, it lacks a bilobed presetal lamella having at most a rudimentary presetal lamella in the notopodium, and always has short, saw-edged geniculate setae in the posterior row as well as the usual elongate capillaries. This species is close to WV. graviert Augener (1913) from Fremantle, and may indeed be identical. However, Augener’s figures of the anterior setae, the shape of the parapodial lamellae and the gill suggest that there are important differences. Moreover, he only mentions two types of setae and gives the impression that there are only saw-edged geniculate setae in the posterior row. THE POLYCHAET FAUNA OF SOUTH AFRICA 327 Nephthys (Nephthys) ? paradoxa Malmgren 1874. . Records: AFR.1578(1). Notes: The material consists of the anterior half of a large worm. The prostomium is pentagonal with 4 subequal antennae and a pair of colourless eyes. The proboscis has 22 rows of papillae with 5-6 papillae per row. The ventral cirrus of setiger 1 is a little shorter than the antennae and the dorsal cirrus shorter still. Branchiae first appear on setiger 9 as foliaceous organs but become stout and cirriform towards the end of the fragment. In anterior feet the notopodium has an oval setigerous lobe, a rudimentary presetal lamella and a rounded posterior lamella a little larger than the setigerous lobe. The gill is flattened and bears a small dorsal cirrus. The neuro- podium has a bluntly conical setigerous lobe, a rudimentary presetal lamella and a rounded postsetal lamella only slightly larger than the feet. The ventral cirrus is normal. Later feet are essentially similar with slightly more divergent and pointed rami. Anterior setae are normal laddered capillaries and posterior setae are long and bear transverse bands of conical teeth. These are deciduous and where they have dropped off transverse marks are left. More material is required before this determination can be confirmed. Nephthys (Nephthys) hombergu Savigny 1818 Nephthys hombergii. Fauvel 1923, p. 367, fig. 143 a-d. Day 1953, p. 431. Records: SB.184(3), 202(8), 203(13), 205(1); LB.3o0o(c), 391(2); TRA. 85(1); TB.gor(1); WCD.15(6), 19(3), 21(7), 23(1), 28(6); FAL.187(2), 341(5), SCD.25(1), 63(8), 109(1). Nephthys (Nephthys) tulearensis Fauvel 1919 Nephthys tulearensis Fauvel 1919, p. 422, pl. 16, figs. 31-39. Records: MB.81(1). Nephthys (Micronephthys) sphaerocirrata Wesenberg-Lund 1949 Nephthys sphaerocirrata Wesenberg-Lund 1949, p. 294, figs. 24-26. Day 1953, p. 431. Records: SB.175(1), 199(2), 202(4), 203(16), 205(2); LB.323(1); WCD. Hato). 49(20), 21(38), 23(18), 26(14), 28(12); FB.3916(2); FAL.229(1), 250(2), 328(3), 338(1), 352(3), 375(7), 376(3), 378(3); L1Z.25(2) ; SCD.82(2), 109(3). Nephthys (Aglaophamus) macroura Schmarda 1861 Aglaophamus macroura Hartman 1950, p. 118. Records: AFR.783(1); AFR.1578(1); SCD.9(4). Notes: ‘These specimens agree very well with Hartman’s description with the exception that the branchiae start on setiger 2 not 3 or 4 as stated. It might 328 ANNALS OF THE SOUTH AFRICAN MUSEUM also be added that the presetal lamella (not described by Hartman) is bilobed. The lamella is well developed in the first few feet but soon becomes reduced. It disappears in the notopodia but in the neuropodia minute lobes persist above and below the setigerous lobe. Family GLYCERIDAE Glycera convoluta Keferstein 1862 Glycera convoluta. Fauvel 1923, p. 383, fig. 150 h—n. Records: LAM.45(2), 49(1), 52(4); SB.125(8), 175(1), 177(14), 179(1), 183(1), 184(2), 189(6), 193(1); LB.158(6), 159(10), 160(2), 169(1), 239(5), 299(c), 300(c), 363(4), 364(6), 380(c), 382(c), 391(6), TB.go1(1); AFR. 1224(1); TRA.77(c), 143(1); WCD.15(5), 19(4), 21(4), 23(5), 26(1), 28(9); FB.323(7), 330(1); FAL.58(1), 187(1), 205(1), 209(1), 219(1), 225(1), 226(r), 240(9), 241(1), 243(1), 338(1), 341(c), 345(3), 347(2), 376(1), 378(2); KNY. 61(1); LIZ.24(1). Glycera alba Miller 1788 Glycera alba. Fauvel 1923, p. 385, fig. 150. Records: ‘TRA.108(2). Notes: The first gill is on setiger 33. I doubt very much whether this species is really separate from G. convoluta. Glycera parashadi Fauvel 1932 Glycera parashadi Fauvel 1932, p. 126, pl. 5, figs. 1-8. Day 1957, p. 86. Records; FB.317(1); FAL.211(1), 245(1). Glycera unicornis Savigny 1818 Glycera unicornis. Fauvel 1923, p. 389, fig. 153 e-7. Day 1953, p. 430. Records: AFR.935(1); TRA.27(1), 91(1); FB.323(3); FAL.117(1), 229(1), 341(1); LIZ.24(1); SCD.18(1). Glycera papillosa Grube 1857 Glycera papillosa. Kinberg 1857-1910, p. 58, pl. 21, fig. 3. Augener 1922, p. 203, text-fig. 9 a—c. ?Glycera kerguelensis McIntosh 1885, p. 344, pl. 35A, figs. 3-4. Records: FB.307(1); MB.74(1); LIZ.25(3), 38(1). Notes : The prostomium has about 8 rings; the proboscidial papillae include a few ovoid forms, but the majority are slenderly conical, being five times as long as the basal breadth; they are not ringed. The jaw supports are deeply forked, the shorter limb being almost half the length of the longer one. There are no gills. The superior presetal lobe is minute but the inferior presetal lobe > ee THE POLYCHAET FAUNA OF SOUTH AFRICA 329 _ is large and pointed. There is a single, broadly rounded post-setal lobe which reaches the same height as the ventral cirrus. This species has longer proboscidial papillae and a much smaller superior presetal lobe than G. capitata. It has been suggested that Glycera kerguelensis McIntosh 1885 is identical with G. papillosa and as Augener (1922) says, the description and figures of the proboscidial papillae agree very well. The type now in the British Museum has had its head cut off but the remainder of the worm shows that the original was much larger than Augener’s specimen or mine. In the feet the superior presetal lobe is much smaller than the inferior one but not quite so minute as in the present specimens. Otherwise the feet are very alike. G. kerguelensis is certainly much closer to G. papillosa (type locality Valparaiso, Chile) than it is to G. capitata (type locality Denmark). Glycera benguellana Augener 1931 (Fig. 8 b-e) Glycera capitata v. benguellana Augener 1931, p. 303, text-fig. g. Records: LAM.26(2), 35(2), 40(2), 41(3), 49(1), 52(1); TB.304(1), 322(1); FAL.51(1), 63(1), 65(1), 206(1), 209(1), 233(3), 250(1); 349(1), 365(1), 378(1); MB.62(1); SCD.109(1). Notes: The body is tapered at both ends, slightly swollen anteriorly and the segments are biannulate. The prostomium is very long with numerous ( ?30) indistinct rings. There are no visible eyes. The jaw supports (fig. 85) have only one fork developed, the other limb being reduced to a mere expansion of the base. The proboscidial papillae include numerous elongate forms (fig. 8c) with 14—16 distinct V-shaped marks on one side and a few stout forms which are essentially similar in structure. The parapodium (fig. 8e) has a fairly large dorsal cirrus, two subequal presetal lobes, a single low rounded postsetal lobe and a broad pointed ventral cirrus. There are no gills. As shown above this species is quite distinct from G. capitata in the shape of the jaw supports, proboscidial papillae and even the feet for the superior presetal lobe of G. capitata is distinctly smaller than the inferior one. Augener’s record was from deep water off South West Africa but the present records show that it is common all round the Cape. Glycera longipinnts Grube 1878 Glycera longipinnis. Fauvel 1932, p. 125, pl. 4, figs. 11-14. Records: FAL.211(1). Notes: The prostomium has about 12 poorly marked rings and rather long terminal antennae. There are no visible eyes. The proboscidial papillae are elongate and conical but not ringed. Gills start on the 2oth setiger as simple 330 ANNALS OF THE SOUTH AFRICAN MUSEUM filaments arising from the dorsal edge of the foot at the same level as the presetal lobes. The gill filament is longer than the presetal lobes. The dorsal cirri are relatively large and arise from the body wall well above the foot in anterior segments but just at the base of the foot in the posterior part of the body. There are two long presetal lobes with pointed ends, and the superior one is distinctly shorter than the inferior. There is only one low rounded postsetal lobe; the ventral cirrus is short and pointed. This is a new record for South Africa. Glycera roux Aud. & M.-E. 1834 Glycera sagittariae McIntosh 1885, p. 346, pl. 42, fig. 8; pl. 22A, fig. 10. Glycera rouxii. Fauvel 1923, p. 389, fig. 153 a—c. Glycera goesi Malmgren, McIntosh 1925, p. 69. Records; TRA.G0(3))727(1); PAL 2973). Notes: The proboscidial papillae are simple blunt cones some of which show a trace of two rings. There are also a few spherical papillae. The jaw supports lack one prong of the complete Y shape. Branchiae start on the 18th foot as long, single, retractile filaments arising from the anterior face of the parapodium. There are two, unequal, pointed postsetal lobes, the superior being slightly longer. An examination of the type of Glycera sagittariae McIntosh 1885 which is now in the British Museum shows that it has feet with single retractile gill filaments arising from the anterior face of the parapodium and that the postsetal lobes are triangular, the superior one slightly longer than the inferior. Both postsetal lobes are shorter than the presetal ones. 3 Fauvel (1932) has recorded G. sagittariae but his description and figures do not agree with the type specimen seen by me. Also MclIntosh’s pl. 42 fig. 8 shows no gill on the dorsal edge of the foot, and his remark that the gill arises ‘from the upper and anterior part of the foot’ agrees with his specimen. Hartman (1950) suggests that Fauvel’s specimen is really a G. tesselata, but G. tesselata lacks gills. It differs in the proboscidial papillae from G. alba or G. prashadt. Hartman suggests that G. sagittariae McIntosh is G. gigantea but this is incorrect, since the latter has globular gills and rounded postsetal lobes. Ophioglycera eximia (Ehlers) 1900 Goniada eximia Ehlers, Monro 1936, p. 141, fig. 25 a-j. Ophioglycera eximia Hartman 1950, p. 38. Records: AFR.1579(1); FAL.233(1). Notes: These specimens are identical with Monro’s specimens in the British Museum. The largest specimen has 4-5 teeth on the macrognaths, 30 dorsal and 19 ventral micrognaths. Parapodia 1-57 are uniramous with a dorsal cirrus which is notched on the dorsal edge; at the 58th foot the dorsal cirrus is THE POLYCHAET FAUNA OF SOUTH AFRICA 331 accompanied by a smaller inferior lobe; on the 60th notosetae appear between _ these two lobes and on succeeding feet the inferior notopodial lobe grows larger and by the 7oth the two lobes are equal in size. This is a new record for South Africa. Glycinde capensis n. sp. (Fig. 8 f-A) Records: FB.306(1), 316(2); FAL.209(2), 250(1), 375(1 juv.); MB.81 (1 juv.); SCD.61(1), 63(1), 82(1). Description: The holotype is the largest of the six incomplete specimens from False Bay and comes from FB.306. It measures 40 mm. for 112 segments and is pale yellow in alcohol. The tapered prostomium is 8-ringed with 2 pairs of eyes and 4 minute terminal antennae. The basal pair of eyes is internal and the distal pair just below the terminal antennae, is external. The long proboscis is covered with the usual rows of papillae and is armed with a pair of ventral _ macrognaths each with 5 teeth and a dorsal ring of 24 micrognaths (one of the paratypes has only 15). According to Hartman’s formula the proboscidial papillae (fig. 8f) consist of 6 groups. Group I which lies along the median dorsal line is formed by a sparse double row of minute tubercles each with a single point (fig. 8f1). Group II (fig. 8f) consists of 6 oblique rows of much larger tubercles running along the dorso-lateral surface. IIa the (dorsalmost) is small and the apex of each tubercle ends in 2 points. IIb and IIc are the largest and each tubercle is curved with a single apical point. IId, e and f decrease in size and the apex of each ends in two points like the open beak of a bird. Group III (fig. 8f?) consists of a row of minute, oval tubercles each ending in double points. Group IV (fig. 8f4) is a row of slightly larger tubercles, each with a 3-pointed apex. Group V (fig. 8/5) is a row of large, soft, bluntly conical papillae each with a small pore at the apex. Group VI, as usual, is absent. The anterior region of the body consists of 28 uniramous segments, each with a straplike dorsal cirrus (fig. 8g), a compressed and tapered setigerous lobe and a ventral cirrus similar in shape to the dorsal but a little shorter. All along the length of the body there is a tendency for the parapodial lobes to become shorter and broader and for the ventral cirrus to be attached more distally. This is true both for the anterior as well as the posterior region. Moreover, the rudiment of what, in the posterior region behind segment 29 will become the notopodium becomes evident about the 20th segment as an inferior thickening and later expansion of the dorsal cirrus. The setigerous lobe consists of the fused presetal and postsetal lobes with the setae issuing as superior and inferior fans. — | _ The posterior, biramous region starts at setiger 29. Here, as the notopodium develops, the dorsal cirrus becomes shorter and moves into a postero-dorsal 332 ANNALS OF THE SOUTH AFRICAN MUSEUM position while an anteroventral lobe grows out and soon reaches the same size. The notosetae issue from a slit between them. The neuropodium is homo- logous with, and essentially similar to, the setigerous lobe and ventral cirrus of the anterior region. A typical foot of the posterior region (fig. 8h) has a broadly triangular notopodium with a small flattened dorsal cirrus and a larger but still essentially triangular neuropodium with a ventral cirrus. In the anterior region the setae are all compound with spinigerous blades. In the posterior region, the notosetae are 2-3 acicular setae with bluntly hooked ends protected by a spike-like guard. The neurosetae are similar to the setae of the anterior region. The South African material is obviously close to G. nordmanni (Malmgren) and G. wirent Arwidsson. However both of these are northern forms, the type locality of the former being Norway and of the latter, the Behring Straits. Moreover there are several small differences from both these species. G. nordmanni has 36-37 anterior segments and examination of several specimens of various sizes in the British Museum showed that this figure is surprisingly constant. G. wireni has 31 anterior segments, only a single pair of eyes at the base of the prostomium and few (17) micrognaths. It seems safer to distinguish the South African material as a separate species until more is known about the variation and distribution of these rather rare worms. Glycinde kameruniana Augener 1918 (Fig. 8 j-m) Glycinde kameruniana Augener 1918, p. 398, pl. 4, fig. 93; pl. 7, fig. 211. Records: FAL.237(1), 341(1), 375(1); SCD.109(4). Description: As far as I am aware this species has not been recorded since Augener’s original description of a 10 mm. ovigerous female from the Cameroons in tropical West Africa. Augener’s description is very incomplete and his figures add nothing to the text. For this reason a full description of the South African specimen is given below. The specimen dredged in False Bay is a complete, ovigerous female measuring 39 mm. with 100 segments. It is pale in alcohol. The prostomium has 8 indistinct rings, 4 minute terminal antennae and a pair of eyes embedded in the basal ring. A distal pair of eyes is lacking. The proboscis has a pair of ventral macrognaths each with 4 teeth and a dorsal arc of 4 micrognaths. The papillae on the proboscis (figs. 8j 1-5) are of the usual type. Those of the mid-dorsal row (group I (fig. 871) are minute oval tubercles with a single point. Group II (fig. 87?) is a dorsolateral band formed of 5 large falcate tubercles in oblique rows; IIa is small, stout and has a single point; IId is very large, with a curved single-pointed tip; IIc is similar but the tip is not so sharp and IId and ¢ are progressively smaller and have two points curved like the open beak of a bird. Group III (fig. 87%) is a row of minute THE POLYCHAET FAUNA OF SOUTH AFRICA 333 tubercles each with 3 points. Group IV (fig. 8j*) is a row of slightly larger tubercles whose oblique tops have 3 points dorsally and a sort of prow ventrally. Group V (fig. 875) is as usual, a row of large soft papillae with faintly bilobed apices. The anterior uniramous region consists of 21 segments. Each parapodium (fig. 8/7) consists of a strap-like dorsal cirrus, a setigerous lobe with a single tapering presetal lobe, a similar, subequal postsetal lobe and a ventral cirrus essentially similar to the dorsal one. The first few parapodia (e.g. the 8th) have the dorsal cirrus as a simple strap but from about the 15th it becomes obvious that the inferior side of the cirrus is bulging so that such dorsal cirri could conceivably be described as having a notch below the tip. In the posterior region the body is a little flatter and the biramous para- podia are relatively larger. The notopodium, formed as a ventral outgrowth from the dorsal cirrus, is at first a simple bilobed structure but further back the dorsal cirrus becomes relatively shorter and posterodorsal in position. The setigerous lobe itself develops a minute second lobe (see fig. 8m) The neuropo- dium remains essentially similar to the setigerous lobe of the anterior region but here it is evident that the presetal lobe is distinctly longer than the postsetal lobe. The setae are of the usual types. In the notopodium there are usually 2 acicular setae (fig. 8k) with bluntly curved apices and pointed guards. The setae of the anterior region and those of posterior neuropodia are slender-bladed spinigers. I have not been able to see the detail given by Hartman (1950). The above description agrees with Augener’s brief account in regard to the number of anterior segments, setae, shape of parapodia and eyes. Of the micrognaths he says he was only able to see ‘einige ganz feine schwarze Punktchen’. ... | take this to mean that there were only a few micrognaths. Augener gives no account of the proboscidial papillae. _ G. solitaria (Webster) from the Atlantic coast of U.S.A. has been re- described by Hartman (1950) and is obviously close to the present material. The number of anterior segments is a little larger (24), a distal pair of eyes is present and there are more micrognaths (10). On the other hand the papillae on the proboscis seem generally similar though groups IV and V are a little different. The parapodia are very alike. It is possible that further collecting on the tropical West African coast will show that G. kameruniana is a synonym of G. solitaria. Goniada maculata Oersted 1843 Goniada maculata. Fauvel 1923, p. 392, fig. 154 a-g. Hartman 1950, p. 20, pl. 1, figs. 7-8. Records: WCD.5 (1 juv.), 26(1 juv.); TRA.1rro(1 juv.); FAL.240(3), 250(1), 352(1). Notes: South African specimens are quite typical. The papillae on the proboscis are as figured by Hartman and the feet agree with Fauvel’s figures. In juvenile specimens it was noted that while the number of micrognaths 334 ANNALS OF THE SOUTH AFRICAN MUSEUM remains 3—4 dorsally and 3-4 ventrally, the number of teeth on the macrognaths may be as low as 4; again the change in the parapodial structure occurs a little earlier, on the 35th as against the 39—41st foot as usually quoted. Specimens of all sizes lacked eyes and from the 30th foot on, where there are two fingerlike presetal lobes to the neuropodium, the superior one is always a little longer than the inferior one. Family EUNICIDAE Subfamily EunicinAE Eunice vittata (Delle Chiaje) 1828 Eunice vittata. Fauvel 1923, p. 404, fig. 158 h—-n. Day 1953, p. 433. Records: AFR.691(p), 801(p), 994(p); TRA.56(p), 58(p), 71(a), 152(p); FB.302(1); FAL.26(1), 29(1), 184(1), 223(2), 243(1), 328(3), 334(7), 338(1), 341(1); MB.9(2), 67(4), 78(1); LIZ.18(4); SCD.22(1), 82(1), 99(3), 100(4), 109(3). | Eunice floridana (Pourtales) 1869 Eunice floridana. Fauvel 1923, p. 402, fig. 157 a-g. Records: APR 7G © a7 30) aa ely Eunice pennata (Miiller) 1776 Eunice antarctica Baird 1869. Eunice savignyi (non Grube) Ehlers 1g08a, p. 88, pl. 11, figs. 7-13. Hartman 1956, p. 283. Eunice pennata. Fauvel 1923, p. 400, fig. 156 h-o. Monro 1930, p. 118, fig. 42. Leodice langi Treadwell 1943, p. 3, figs. 14-18. Records: AFR.691(1), 707(1); TRA.115(1); FAL.375(2). Notes: Baird’s type of Eunice antarctica from ‘Antarctic seas’ which is in the British Museum has been compared with £. pennata from Europe identified by Fauvel and E. pennata from ‘Tristan de Cunha identified by Monro (1930) and the present material from South Africa. All are identical. Baird states that in E. antarctica the gills start on the 8th foot, but examination of the type shows that there are small single filaments from the third foot. E. pennata has fairly short tentacles with faintly marked annulations towards the tip. It is close to E. savignyi Grube from the Philippines but the latter is a tropical shallow water species with distinctly jointed or even moniliform tentacles. Moreover in E. savignyi the median tentacle is very long and reaches the 16th setiger whereas in E. pennata it only reaches the 3rd to 8th foot. Ehlers’s record of E. savignyt from the Agulhas Bank obviously refers to FE. pennata. By the courtesy of the U.S. National Museum I was able to examine Treadwell’s type of Leodice lang: collected by H. Lang from 160 fathoms off Cape Town. The material, Ref. No. A6o099, consists of a single specimen from which the jaws have been removed. The gills begin on the grd setiger as simple filaments, reach a maximum of 8 filaments on the 15th foot and end on THE POLYCHAET FAUNA OF SOUTH AFRICA 30) the 37th foot. The setae also agree with those of Eunice pennata. Hartman (1956) has referred Treadwell’s Leodice langi to E. savignyi following Ehlers’s record as cited in Fauvel (1932). Eunice aphroditois (Pallas) 1788 Eunice rousseaui Quatrefages, Fauvel 1923, p. 403, fig. 158 a—g. Eunice aphroditois Day 1943, p. 432. Records: LB.157(2), 161(1), 299(c); FAL.51(1), 159(1); LIZ.35(1). Eunice australis Quatrefages 1865 Eunice australis. Fauvel 1932, p. 139. Day 1953, P- 432- Records: TRA.135(5), 152(p); FAL.30(1), 302(1), 334(2); MB.49(1), 67(1); LIZ.18(1); SCD.9(1), 32(2), 40(4), 54(3), 58(1), 89(9). Marphysa sanguinea (Montagu) 1815 Marphysa sanguinea. Fauvel 1923, p. 408, fig. 161 a—h. Records: LB.299(fc). Marphysa capensis (Schmarda) 1861 Marphysa capensis. Augener 1918, p. 332, text-fig. 33. . Records: TB.318(4). Marphysa depressa (Schmarda) 1861 Marphysa depressa. Day 1953, p. 434, fig. 5 n, p. Records : LB.299(fc), SCD.63(1). Marphysa purcellana Willey 1904 Marphysa purcellana Willey 1904, p. 263, pl. 13, fig. 17. Day 1953, p. 435- Records: FAL.219(1), 223(1); MB.20(1); LIZ.18(1); SCD.58(1). Notes : This is the second record of this rare species. The head and anterior segments are reddish brown and the falcigerous setae have long blades. Marphysa sp. Records: L1Z.35(1). Notes: A single specimen of Marphysa which is new to South Africa was dredged in Algoa Bay (LIZ.35.T). It is 20 mm. long by 0-6 mm. with about 100 segments. The tail end is missing. This small, slender worm may be a juvenile and for this reason is not described as a new species. 336 ANNALS OF THE SOUTH AFRICAN MUSEUM The diagnostic characters are as follows: Prostomium bilobed, eyes present, antennae faintly annulated, a little longer than the prostomium. Gills appear far back (approximately the 6oth setiger) and never develop more than a single filament. Setae include superior capillaries and inferior bidentate falcigers in all feet. Posterior feet have in addition comb-setae with about 12 teeth and bidentate acicular setae with guards. The acicula are bluntly pointed and all setae and acicula are pale. Lysdice natalensis (Kinberg) 1865 Lysidice capensis Grube, McIntosh 1905, p. 40, pl. 3, fig. 13. Lysidice natalensis Day 1951, p. 40. Day 1953, p. 435. Records: AFR.967(1); FAL.16(4), 21(p), 44(3), 80(p), 113(p), 122(p), 126(1), 134(1), 156(2), 219(1), 269(1), 371(1); MB.49(1), 56(2), 67(1), 78(1); LIZ.6(1), 18(1), 27(1); SCD.89(1). Subfamily ONUPHIDINAE Onuphis emerita Aud. & M.-E. 1834 Onuphis emerita. Fauvel 1923, p. 415, fig. 163. Monro 1930, p. 128, fig. 47. Records: TRA.41(3); FB.321(3); FAL.205(2), 211(1), 228(2), 238(7), 242(4), 341(3), 352(2); MB.67(1), 81(1); LIZ.19(1), 24(1), 31(2 juvs.); SCD.96(1). Notes: The specimens recorded above vary from juveniles of 25 mm. (LIZ.31) to well-grown individuals of 75 mm. They agree very well with Fauvel’s description. Some individuals are pale but others have well-developed brown pigment pattern identical with that on preserved specimens from Naples now in the British Museum. This species was first described from South Africa by Monro (1930) but his account is slightly inaccurate. I have checked his specimen from False Bay (British Museum No. 1930.10.8.1792) and find that the ‘little conical tubercle’ which he states is restricted to the first 3 or 4 setigers is actually the setigerous lobe. The structure which Fauvel 1923, p. 415 and fig. 163. C, calls ‘un petit tubercle conique entre le mamelon sétigére et la base du cirre dorsal’ is a superior projection of the presetal lobe. This is well marked on the 5th—1oth foot on a 100 mm. specimen from Naples but is not present on the smaller 23 mm. specimen described by Monro nor on the present material. Onuphis (Nothria) holobranchiata Marenzeller 1879 Onuphis (Nothria) holobranchiata. Izuka 1912, p. 106, pl. 11, figs. 10-12. Records: LAM.11(5), 26(5), 35(2), 39(c), 40(2), 41(2); LB.382(7); TRA.110(1), 113(1), 135(1), 143(3); FAL.375(5), 376(c), 378(3). Notes: Most of the specimens were broken but the average size when complete was probably 4—5 cm. Only faint traces of pigment remain between THE POLYCHAET FAUNA OF SOUTH AFRICA 337 the parapodia and some worms are completely pale. Eyes are present just external to the inner lateral occipital antennae. The ceratophore of the median antenna has 10-12 rings but the other antennae are longer and have cerato- phores with about 14 rings. The jaws are weakly chitinized and the formula starting with the main fangs is Mx.I = 1 (left) + 1 (right); Mx. II = (5-7) + (6-8); Mx. III = (6-7) +0; Mx. IV=7+ 10; Mx. V=1+1. The tentacular cirri are slightly longer than the peristomial segment. The dorsal cirrus is cirriform throughout. Gills are present as a single filament on the first and all succeeding feet to near the posterior end of the body. As usual the first 5 feet have a cirriform ventral cirrus, and a prominent setigerous lobe and a cirriform ‘post-setal lobe’. From the 6th foot onwards the ventral cirrus is represented by a glandular pad, the setigerous lobe becomes inconspicuous and a presetal swelling appears, though the presetal lobe is never well marked. The ‘post-setal lobe’ diminishes in size and from about the 20th foot it is an inconspicuous conical projection of the foot partially enclosed by a dorsal arc of setae. This seems to be what usually happens in the Onuphidinae. The first 5 feet bear about 6 pseudocompound setae with bivalve hoods. These setae are all tridentate but the third tooth may be so minute that unless it is seen in profile, it may be thought to be absent. There are also 2-3 simple capillaries. On the 6th foot the pseudocompound setae disappear and the capillaries develop narrow wings. Two bidentate acicular setae appear in the 10th foot and 2 very fine comb-setae with about 18 teeth are present in the 20th foot although they may be present before this. An average foot from the middle of the body thus contains 2 pale acicula with very slender tips projecting from the surface, about a dozen narrow-winged capillaries, 2 fine comb-setae and 2 bidentate acicular setae. O. holobranchiata first described from Japan has been recorded from the Indian ocean by Crossland (1904) and Fauvel (1930, 1932). The pseudo- compound setae are variously described as bidentate and tridentate. Hartman (1944) has made a very thorough study of the Onuphidinae of the western hemisphere and under the name Wothria describes two species NV. elegans (Johnson) and WN. iridescens (Johnson) which, like O. holobranchiata, also have gills as simple filaments from the first setiger. She discusses the differences between the three species on p. 88. Onuphis (Nothria) geophiliformis (Moore) 1903 Nothria geophiliformis Moore 1903, p. 445, pl. 25, figs. 57-59. Onuphis geophiliformis Izuka 1912, p. 103, pl. 11, figs. 8-9. Records: FB.311(10), FAL.219(1), 328(1 juv.). Notes: All the specimens were broken but the largest was approximately 30 mm. long when complete. Most of the specimens are quite pale in alcohol, 338 ANNALS OF THE SOUTH AFRICAN MUSEUM but two show brown markings between anterior parapodia and one has the anterior dorsum uniformly brown. The tube is unknown. Frontal antennae are ovoid, and the occipital antennae have ceratostyles 3-4 times longer than their ceratophores. The median antenna which reaches back to the 6th setiger has a ceratophore with 8-9 rings. Eyes are present external to the inner lateral ceratophores. The tentacular cirri are a little less than the length of the peristomial segment. The jaws are very pale and soft. The mandibles have the usual form and the maxillary formula is Mx. I = (left) 1 + (right) 1; Mx. II = 8 + 9; Mx. III = 8+ 0; Mx. IV. =649; Mx. V=141. The parapodia are of the usual (igi Gills as single filaments appear on the 4th or 5th setiger, and are always longer than the dorsal cirrus from whose base they arise. They persist over most of the body but are absent from the last 40-50 feet. Each of the first 5 feet has a prominent setigerous lobe but thereafter the setigerous lobe becomes reduced and is hidden between the presetal swelling and the cirriform post-setal lobe. The latter also decreases and at the 12th foot becomes a mere papilla partially encircled by setae. The ventral cirrus is a tapered cirriform organ for the first 5 feet but thereafter becomes a glandular pad and merges with the setigerous lobe at about the 15th foot. The hooded pseudocompound setae of the first few feet are tridentate with the terminal tooth much longer than the second and third. The winged capillaries are simple, never compound. Hooded and bidentate acicular setae appear in the 9th—1oth foot and 2-3 minute comb-setae with about 12 teeth further back. The above description agrees well with that of Izuka except that he states that eyes are absent and comb-setae have 16 teeth. According to Moore (1911) and Hartman (1944), O. geophiltformis is distinguished from O. pallida Moore Ig11 mainly by the shape of the pseudocompound setae. In the latter species the terminal tooth is no larger than the others. Onuphis (Nothria) conchylega Sars 1835 Onuphis conchylega. Fauvel 1923, p. 415, fig. 164. Records: TRA.33(1). Genus DIOPATRA Although Diopatra has been regarded as one of the most clearly defined genera of the Onuphidinae, distinguished by the spiral arrangement of the fila- ments on the branchial trunks and the possession of tentacular cirri, it is shown below in the discussion of Dzopatra dubia that the separation of Diopatra from Epidiopatra is by no means simple. Moreover the question as to whether there is one species of Diopatra or several species is a matter of controversy. Important discussions will be found in Augener (1918), Fauvel (1933), Hartman (1944) and Rullier (1958). Augener divided the genus into two main groups of species THE POLYCHAET FAUNA OF SOUTH AFRICA 339 based on the number of teeth on the comb-setae. Fauvel who gives parallel lists of species with few or many teeth on the comb-setae from similar regions, concludes that there is one species with minor variations. Hartman used various characters to distinguish several species in the western hemisphere. Rullier agrees with Fauvel. , I have examined the South African material reported below and material in the British Museum which includes 30 samples labelled Dzopatra neapolitana from various parts of the world and samples labelled with 10 other specific names. This has shown beyond doubt that there ave several species and that the synonymy of Diopatra neapolitana in particular is very confused. However, once the diagnostic characters are recognized, the specific distinctions are fairly clear. The following characters seem to be of value. PIGMENT PATTERN While the intensity of the pigmentation varies from specimen to specimen and the fainter markings fade with age, the pattern is surprisingly constant and the more intense diagnostic marks have persisted in spirit in specimens of D. neapolitana collected in the last century. PROSTOMIAL APPENDAGES These are often called tentacles but are here referred to as antennae as they arise from the prostomium. The frontal antennae are presumable affected by the method of fixation but in one species at least (D. dubia) they are charac- teristic, shovel-like expansions while in others they are sausage-like or subulate. The ceratophores of the five occipital antennae are ringed and the number of rings seems to have a limited variation within a species e.g. 3-5, 6-8, 9-12, 15-20. In D. dubia and in Epzidiopatra hupferiana the rings develop lateral projections so that the ceratophores may be said to be branched. Hartman (1944) has also drawn attention to ‘glandular structures on antennae’ but unfortunately she does not describe her methods of preparation, and the picture seen varies with the method used. However, if the clear cuticle is stripped from the ceratostyle, a number of depressions formed by oval cells will be found projecting into the thickness of the cuticle. These are usually arranged in 15-25 longitudinal rows or may be irregularly placed. JAws Most workers are agreed that the maxillae are of no diagnostic value and though this may not be the case, there is variation both in the number of well- formed teeth and in the interpretation of what are teeth and what are not. Incidentally there is some confusion in the numbering of the maxillary plates since the left side has one more plate than the right. Dissection shows that Mx. IV (which are curved plates) correspond, also Mx. I (the main fangs), Mx. IT and Mx. V, so that it is Mx. III which is missing from the right side. The mandibles too are usually disregarded, but in at least one species (D. monrot described below) the mandibular shafts are characteristically swollen. 340 ANNALS OF THE SOUTH AFRICAN MUSEUM PARAPODIA These provide useful characters but the structure of the feet changes over the length of the body and the real nature of the different parts has not been realized. Thus in the first 4 feet, a presetal lobe is not developed, but the setigerous lobe which is oval and compressed has been called a presetal lobe. Further back in the branchiferous region, a true presetal lobe is formed and may be symmetrical or asymmetrical with a marked inferior projection. Towards the end of the branchiferous region the presetal lobe is again reduced to an insignificant swelling in front of the setae. In the anterior region there is what has become known as a cirriform postsetal lobe. This later decreases is size, fuses with the now insignificant setigerous lobe and becomes partially surrounded by a dorsal arc of setae—it becomes, in fact, a small conical setigerous lobe but for convenience, the term postsetal lobe will be retained for anterior feet, and in D. chiliensis it is quite characteristic for there are two cirriform postsetal lobes instead of the usual one. A somewhat similar arrange- ment is described by Willey (1905) for Diopatra amboinensis from Ceylon. The ventral cirrus which is cirriform for the first 4 or 5 feet later becomes a ventrolateral glandular pad. The dorsal cirrus is always cirriform but diminishes in size on posterior segments and develops a dorsal branchial trunk on the 4th, 5th or 6th foot. The structure of the gills has been studied by several workers but it seems that the method of preservation has so much effect that only the most marked differences remain constant for a species. Thus the number of branchiferous segments (usually 40-50) may be greater in large specimens and smaller in juveniles. The largest gill is usually the 3rd—6th but in juveniles it may be the first. In well-preserved specimens the branchial trunk is itself spiral but this is not true of compressed specimens removed from a tube. In most species the filaments have a length equal to 3-4 times the thickness of the trunk but in one species the basal filaments are hardly longer than the thickness of the branchial trunk. SETAE The pseudocompound setae of the first four feet provide characters of the greatest importance—they may be unidentate, bidentate or tridentate and the guards (or hoods) may be well or poorly developed. The winged capillaries of later feet seem to vary in breadth of wings but the degree of serration at the base of the wing seems to change along the length of the worm. No significant departure from the norm has been noted in the structure of the four tapered acicula which just project from the setigerous lobe. The shape of the bidentate acicular setae seems to be very constant but the number of teeth on the comb- setae and the angle at which they are set seems to be diagnostic within limits. Thus in D. neapolitana there are a few large teeth (4-10), in D. cuprea there are numerous (15-25) small teeth and in D. musseraensis the blade is rolled up like a paper trumpet and at certain angles appears to have a stout central tooth as figured by Augener (1918) and Tebble (1955). 5 THE POLYCHAET FAUNA OF SOUTH AFRICA 34.1 TuBE In most species the projecting end of the tube is beset by shell fragments or other foreign objects such as leaves set edgeways on. In some, however, the tube is composed of hardened mud or sand without shells and Hartman has described one species with a ringed tube. As mentioned earlier many records have been referred to the species ‘Diopatra neopolitana which was regarded by Fauvel as being widely distributed in warmer waters. Among others, most of my own records from South Africa are incorrect and it is convenient before describing new material to redescribe a specimen of D. neapolitana from Naples and discuss the tangled synonymy. Diopatra neapolitana Delle Chiaje 1825 (Fig. 9 a-g) Diopatra neapolitana. Claparéde 1868, p. 122, pl. 6, figs. 4 a-h. Ehlers 1868, p. 285, pl. 12, figs. 6-20. non Diopatra neapolitana Crossland 1903, nec Day 1934, nec Day 1957, nec Tebble 1955, nec Monro 1936, nec Fauvel 1932, nec Wasenberg-Lund 1949. Material (from Naples) British Museum numbers 1898: 5: 6: 137-9; 1919.11.6.25/26; 1921.5.1. 1873/74; 1876:10:4:41; 1890.6.7.9.13 other samples in the British Museum labelled D. neapolitana from South Africa, West Africa, West Indies, various parts of the Indian Ocean and Australia were not this species. Diagnosis: Pseudocompound setae with a very small secondary tooth and well-developed guards; comb-setae with 4-10 teeth; a dark spot in the middle of the back on branchiferous segments. Description: The following description is based on three British Museum specimens 98.5.6.137/9 from Naples. The specimens are very large, measuring 24 cm. by 5 mm. with over 250 segments. The general colour is brown, darker anteriorly and on the inner sides of the ceratophores of the occipital antennae. There is a short, dark, transverse, mid-dorsal bar on the anterior margin of each of the first 10 branchiferous segments (see fig. ge). The frontal antennae are tapered and fairly long. The ceratophore of the median occipital antenna has 9-11 rings and is about a third the length of its ceratostyle which has 20-25 broken longitudinal rows of clear cells projecting into the cuticle. The tentacular cirrus is three-quarters the length of the median ceratophore. The mandibles have calcareous bilobed cutting edges and well chitinized, straight, dark, tapered shafts. The maxillae are also well chitinized and dark. The supports are heart-shaped and the main fangs (Mx. I) are strong. The dental formula is Mx. I (left) = 1 + (right) 1; Mx. IT = 8 + 8; Mx. III = Be Os ix TV — 7 178; Mx. V — 1 -+ 1. In one large specimen Mx. V = Seo) 342 ANNALS OF THE SOUTH AFRICAN MUSEUM The first few feet tend to be bent forward and downward, and the first three (fig. 9a) are of the usual form described earlier for the genus Diopatra. The first gill occurs on the 4th or 5th foot. The largest gill is on the 7th or 8th foot and it extends two-thirds of the way across the dorsum; it has about 10-12 whorls of branchial filaments. Individual filaments of the basal whorls have a length equal to 3 to 4 times the thickness of the branchial trunk. Succeeding gills decrease slowly in size to end about the 50th foot. The presetal lobe is well marked from the 5th to the 2oth foot and in these well-preserved specimens there is an obvious ventral projection (see fig. 9b). The first four feet have 2 superior capillary setae and a fan of 5-6 hooded pseudocompound setae. The latter (fig. 9d) usually possess a very small secondary tooth but this may be absent. Pseudocompound setae are absent from the 7th foot and are replaced by winged capillaries which later develop serrations on the base of the wings. Comb-setae (fig. 9g) appear on the 8th foot and, tested over the whole length of the body, always have truncate ends with 5-10 coarse teeth. Two bidentate acicular setae (fig. of) appear on the 15th foot and persist over the rest of the body. Discussion: Neither of the two descriptions given by Delle Chiaje (1825) and (1841) are sufficient for more than generic diagnosis, though the notes on pigmentation vaguely suggest D. neapolitana. Quatrefages (1865) described D. gallica which according to most authorities is synonymous with D. neapolitana, but not having seen the type, I must rely only on the published description which is not sufficiently detailed for D. neapolitana. On the other hand both the description of Claparéde (1868) and Ehlers (1868) are very good. Claparéde’s figure 4D of the pseudocompound setae probably represents a broken seta but 4£ shows an unidentate seta and the text states of the pseudocompound seta: ‘son extremité se recourbe de maniére a constituer une veritable serpe (4E). Chez quelques individus cette serpe est bidentée.’ If he had added that the secondary tooth is usually rudimentary the matter would have been clearer. His figures 4H and 47 show comb-setae with 7 and 1o teeth. Ehlers (1868) gives a detailed description of pigmentation and of the mid- dorsal spot on branchiferous segments. His figures show comb-setae with 8 teeth but the pseudocompound setae were presumably broken off short as often occurs for he took them to be acicula and does not describe pseudocompound setae at all. Crossland (1903) who describes a different pigmentation, strongly bidentate pseudocompound setae and comb-setae with numerous teeth obviously had a different species in front of him. Fauvel (1923), whose figure 166d shows a comb-setae with few teeth and figure 166f shows a strongly bidentate pseudocompound seta presumably had more than one species before him. Fauvel (1930) and (1932) had specimens with numerous teeth to the comb-setae but the structure of the pseudocompound setae is not stated. Subsequent authors such as myself Day (1934) and (1957), Monro (1937), (1938), Tebble (1955) and Wesenberg-Lund (1949) have followed Crossland and Fauvel. THE POLYCHAET FAUNA OF SOUTH AFRICA 343 Fic. 9. Diopatra neapolitana: a anterior view of 2nd foot; 5 anterior view of 12th foot; ¢ anterior view of posterior foot; d, d1 pseudocompound seta of 2nd foot and an unidentate variety; e 10th segment showing pigmentation; / bidentate acicular seta; g comb-seta. Diopatra neapolitana var. capensis: h, ht pseudo-compound seta and variation; 7 anterior end showing pigmentation; j bidentate acicular seta; k comb-seta; / anterior view of 2nd foot. 344 ANNALS OF THE SOUTH AFRICAN MUSEUM It would appear then that the true.D. neapolitana which occurs at Naples and presumably elsewhere in the Mediterranean has often been recorded in error from the Indian Ocean, Australia, South Africa, Pople West Africa and the West Indies. NOTE ON SPECIMENS FROM DuRBAN BAy Record: DBN.26(6). After the diagnostic characters of the genus Diopatra had been determined, all earlier collections from South Africa were re-examined, and among others the material from Morrumbene Estuary, Inhaca Island and Durban Bay reported by me (Day 1957) as D. neapolitana. The Morrumbene and Inhaca specimens proved to be D. cuprea but the Durban Bay specimens were D. neopolitana. The latter had the characteristic pigmentation, comb-setae with 5-9 teeth, the inferior projection from the presetal lobe of the 6th—12th foot but the pseudocompound setae were more variable than usual. Some were unidentate, some minutely bidentate and some had a well-developed secondary tooth. Diopatra neapolitana var. capensis nov. (Fig. 9 h-l) Records: LAM.17(1), 48(4), 55(1), 56(5), 60(3), 63(1); SB.135(1), 181(1), 202(2); AFR.1535(1), 1544(1), Senta TRA.g1(1); FAL.209(1); MB.33(2), 34(1), 37(2), 81(4); LIZ.19(p), 23(3), 24(4); SCD.20(4), 63(10). Description: The type is one of the 3 specimens from station LIZ.23 dredged in Algoa Bay at 33°58'S/25°43’E in 38-5 metres on a mud and clay bottom. It is an anterior half of a worm and measures 50 mm. by 4 mm. for 50 segments. The tube is of the usual form with broken shells attached edgeways. The differ- ences between this variety and the stem form concern the colour pattern, the setae, the shape of the gills and the presetal lobe of branchiferous feet. The central area of the prostomium between the occipital antennae is touched with brown and both the inner and outer faces (but not the sides) of the ceratophores are brown. The peristome (fig. 92) has a continuous transverse bar and the first five setigers have 5 dorsal marks, 3 on the anterior margin and 2 posteriorly thus _ _ . Within the next few segments the 3 anterior marks fade but the 2 posterior marks persist to the middle of the branchiferous region. The cuticle of the occipital antennae has 25 rather irregular rows of clear oval cells projecting into it. The only obvious difference in the shape of the feet between the Cape and the stem form from Naples, concerns the presetal lobe which, in the stem form, is well developed between the 6th and 15th feet. In this Cape material the inferior projection of the presetal lobe is poorly marked or absent so ae te lobe is symmetrical. THE POLYCHAET FAUNA OF SOUTH AFRICA 345 The gills appear on the 4th or 5th foot, soon reach a maximum size and thereafter decrease gradually to end about the 50th foot. Each gill has a characteristically long stout trunk and short filaments; thus the basal filaments first appear half-way up the trunk and are not longer than twice the width of the trunk. The pseudocompound setae of the first feet (fig. 92) are almost always unidentate and only on one specimen was a small rudiment of a secondary tooth found. Moreover this Cape material usually lacks hoods over the apices of the pseudocompound setae and only in a few cases have vestiges of hoods been seen. Comb-setae (fig. 9k) with 9-12 rather fine teeth appear about the 12th foot, and bidentate acicular setae (fig. 97) on the 18th foot. Juvenile and Epidiopatra stages. Five young stages were found which varied from 15 mm. to 25 mm. in length. Three of them lacked tentacular cirri but the cirrophores of the occipital antennae were quite normal, 7-ringed and without lateral branches such as commonly occur in Epzdiopatra species. Pseudocompound setae were unidentate and provided with small hoods. No comb-setae had been developed but bidentate acicular setae were found in posterior feet. Gills were present from the 4th or 5th foot to the gth, roth or r1th. The first gill was not only larger than succeeding ones but also better developed. The only markings were brown spots in the intersegmental junctions above the feet in the middle of the body (cf. D. punctifera Ehlers). Two other specimens of the same size possessed tentacular cirri and had better-developed gills from the 4th to the 16th foot. In all other respects they agreed with the Epidiopatra stages. This evidence, together with that of Monro 1924A who described a juvenile Diopatra cuprea from Madeira which also lacked tentacular cirri and the note below (p. 350) which describes the Epidiopatra stages of Diopatra dubia, shows that the juveniles of certain species of Diofatra lack tentacular cirri. However, this is not always the case. Among empty Diopatra tubes dredged from False Bay was a mud cocoon containing 6 post-larval Diopatras of 8-15 mm. Each had a normal pair of tentacular cirri, obvious spiral gills and bidentate pseudo- compound setae. Diopatra monroi n. sp. (Fig. 10 a) Diopatra cuprea (non Bosc) Augener 1918, p. 530, text-fig. 39 (partim). Diopatra punctifera (non Ehlers) Monro 1930, p. 124, fig. 44 a-b; 1936, p.-147. Records: SB.124(1), 208(1); LB.162(3); WCOD.26(1); AFR.718(7); TRA.68(abundant), 70(c), 77(c), 88(8), 89(fc); FAL.352(2). Diagnosis: Pseudocompound setae strongly bidentate; mandibles stout and thicker in the middle than at the ends; tube usually of compacted mud. Description: The holotype is one of 7 anterior fragments from AFR.718.D. It is 28 mm. long by 4 mm. wide with 66 segments. The complete worm was 346 ANNALS OF THE SOUTH AFRICAN MUSEUM probably twice this length and not fully grown. The general colour is brownish anteriorly and pale posteriorly, but the important features are that there is a dark spot on the prostomium behind the median occipital antenna and the peristome and anterior segments have brown cross bars across the back which are best marked in the branchial region (see fig. 10¢). The frontal antennae are ovoid and the five occipital antennae are borne on ceratophores with 6-8 rings. The median antenna is 4-8 times the length of its ceratophore. When the cuticle is skinned off the antenna no clearly marked rows of cells were found projecting into it. If such cells are present they must be poorly developed. No eyes are visible on the prostomium. The mandibles (fig. 105) are quite characteristic. The cutting edges are normal and white but the supports are black, sausage- or spoon-shaped and curved. Even in juveniles where the mandibles are pale they are thicker in the middle than at the ends. The maxillae on the other hand are always weakly chitinized and brown. The supports are heart-shaped and the main fangs (Mx. I) are soft. The dental formula of the type is: Mx. I = 1 (left) + 1 (right); Mx. Il=5+ 7; Mx. II]=6+0; Mx. IV=5+7; Mx. V= I+ 1. Dissection of other specimens shows that the number of teeth is not constant so that the average formula is: Mx. 1 = 1 (left) + 1 (right); Mx. II = (5-7) + (6-8); Mx. III = (6-8) + 0; Mx. IV = (5-10) + (7-10); Mx. V = 1+ 1. The fifth pair of plates are pale with a single dark tooth on the edge of each. ! Anterior feet (fig. 10a) each have an ovoid and compressed setigerous lobe, tapered dorsal and ventral cirri and a tapered post-setal lobe. The presetal lobe is not developed in the first few feet. The dorsal cirrus persists throughout the branchiferous region but gradually becomes more slender and reduced in size. The ventral cirrus is reduced over the first 5 feet and from the 6th onwards, it is represented by a ventro-lateral glandular pad. The setigerous lobe soon becomes less obvious and on the 6th foot it is concealed behind a low, swollen presetal lobe. The postsetal lobe is gradually reduced in size to a conical papilla. Gills appear on the 5th foot, reach a maximum size by the 8th foot and continue to the 43rd foot. Gill filaments are of normal length and appear fairly near the base of the trunk which is not particularly stout. The first four feet bear a fan of setae including 1-2 superior simple capil- laries and 4-5 pseudocompound setae (fig. 10e) each with a strongly bidentate tip covered by a pointed hood. Only the simple capillaries persist and these become more numerous in posterior segments; the blades remain narrow and do not develop serrations at the base. From the 6th or 8th foot comb-setae with 15-25 teeth appear. Bidentate acicular setae with guards (fig. 10/) appear about the 16th foot. The tube is of hardened mud or occasionally of sand and no sign of shells have been seen in many hundreds of specimens examined, even in those from shallow depths where the substratum is sand and shells. This species occurs on THE POLYCHAET FAUNA OF SOUTH AFRICA 347 Fic. 10. Diopatra monroi: a anterior view of 2nd foot; 6 ventral view of mandibles; c anterior end; d comb-seta; e pseudocompound seta from 2nd foot; f bidentate acicular seta. Diopatra dubia: g anterior view of 2nd foot; A anterior view of 8th foot; 7 anterior end; Jj anterior view of prostomium with ceratostyles omitted; k pseudocompound seta from 2nd foot. 348 ANNALS OF THE SOUTH AFRICAN MUSEUM muddy bottoms along the west coasts of South and South West Africa in enormous numbers. ‘The main banks are 60-100 fathoms deep and fishermen report that at times the mud tubes may clog a trawl so that the gear has to be cut away. , I have examined the specimens recorded by Monro (1930) and (1936) as D. punctifera and find that they are identical with the specimens described here; equally certainly they are not the Diopatra punctifera which Ehlers described from the Agulhas Bank. Monro himself drew attention to the difference in the mandibles which are so characteristic. By the courtesy of the Director of the Hamburg Museum I was able to examine the three specimens V.8718 recorded by Augener 1918 from Swakop- mund under the name of D. cuprea. According to Augener they were obtained from mud tubes, and they still retain a dark spot behind the ceratophore of the median antenna and brown cross-bars across anterior segments. ‘The mandibles have stout black spoon-shaped shafts. They are in fact typical examples of D. monrot. Diopatra dubia n. sp. (Fig. 10 g—k) Records: TRA.73(2), 80(2), 110(1), 113(1), 143(3); FAL.237(3), 240(3), 352(3)s 376(3). Diagnosis: A small species with poorly developed gills, flattened, shovel-like frontal antennae and side branches on the ceratophores of the occipital antennae. Description: The holotype was selected from a collection of 7 specimens from FAL.237.H. It is a well-preserved specimen measuring 23 mm. by 1:8 mm. with 50 segments. The tail end is missing. The tube is rather fragile and com- posed of flocculent debris, small sand grains and calcareous fragments. The worm is mainly pale but there are golden brown markings anteriorly though in other specimens these may be faint or lacking. The type has flecks forming a rough circle on the slightly concave area of the prostomium between the antennae and 3 paratypes each have a brown spot in the ceratostyle of the occipital antennae. The anterior segments are ringed with brown with a break over the ventral nerve cord and the glandular ventral cirri. The markings are strongest between the parapodia and persist there after they have faded elsewhere. The palps are oval cushions projecting outwards below the frontal antennae. The frontal antennae (fig. 107) are unusual. Instead of being stout and cylindrical as in most species, they are flattened, much broader than long and rather spade-shaped. In large specimens they are fused for most of their length. | The occipital antennae (fig. 107) all have 5-ringed ceratophores with side projections on each ring except the enlarged terminal one. The median antenna has branches on either side of its ceratophore but the inner laterals and outer’ THE POLYCHAET FAUNA OF SOUTH AFRICA 349 laterals only have branches on their medial sides. The ceratostyles are relatively short, and the whole median occipital antennae only reaches back to setiger 3. In fresh specimens an internal brown spot may be seen in the ceratostyle. The jaws are very weakly chitinized. The mandibles have calcified cutting edges but their straight, tapered shafts are pale except for narrow black lines on the inner edge of the dorsal surface. The maxillae are equally soft and only the teeth, outer margins and a conspicuous crescentric area medial of Mx. IV is blackened. The supports as usual are heart-shaped and starting with the main fangs (Mx. I) the formula of the holotype may be expressed as: Mx. I = (left) 1 + (right) 1; Mx. II1=9-+ 9; Mx. II]=9-+0; Mx. IV=6+09; Mx. V = 14+ 1. Dissection of other specimens has shown that this formula is higher than most and the range is shown by the figures in brackets: I = 1 fea) © (right); Il — (5-9) + (6-9); ITT — (6-9) + 0; IV = (5-6) + 9; Ve r+ 1. The peristomium bears a pair of rather short but quite obvious tentacular cirri showing clearly that the type material must be included in the genus Diopatira as at present defined. The anterior feet (fig. 1og) are of the usual fom, There is an ovoid setigerous lobe with a rather short cirriform post-setal lobe behind it. The cirriform dorsal cirrus above is considerably longer and the cirriform ventral cirrus below is almost as long. The dorsal cirrus develops a dorsal gill on the 5th foot. However, the gills (fig. 10h) are never well developed. The largest gill is the 3rd on the 7th foot and this has only 2—3 whorls of filaments and just reaches across the mid-dorsal line. ‘Thereafter the gills diminish rapidly and in most cases there is only a small dorsolateral tuft of filaments. The last gill occurs on the goth foot. The postsetal lobe of anterior feet diminishes in size posteriorly, fuses with the setigerous lobe and disappears about the 25th foot. The cirriform ventral cirrus of the first 3-4 feet becomes a ventrolateral pad on later segments. No presetal lobe was distinguished. Apart from the 4 colourless and tapered acicula present in all feet, the anterior feet each have one superior capillary seta and about 5 pseudocompound setae (fig. 10k) with long pointed hoods over the usual strongly bidentate tips. The pseudocompound setae are replaced by winged capillaries on the 6th foot. These increase to 12 and then decrease again to about 4. Bidentate acicular setae with guards appear about the 1oth foot and fine comb-setae with 18-20 teeth appear on the 15th. One specimen (not the holotype) has a complete posterior end which bears two fine cirri below the anus. The above description of a small species of Diopatra with shovel-shaped frontal antennae, branched ceratophores to the occipital antennae and poorly developed gills would seem at first to be an unusually well-defined species of Diopatra. However, further samples revealed very similar worms of the same size which lacked tentacular cirri. Careful study revealed certain small differences 350 ANNALS OF THE SOUTH AFRICAN MUSEUM detailed below but it was obvious that all these forms were very closely related if not mere sexual differences or growth forms of the same species. The question immediately arose as to whether the genera Diopatra and Epidiopatra are distinct. All the present material was re-examined as well as specimens of Epzdiopatra hupferiana and E. drewinensis from tropical West Africa and Monro’s material from False Bay which has been referred by me (Day 1957) to E. hupferiana var. monrot. It should also be noted that the discovery of a cocoon of Diopatra neapolitana var. capensis containing newly hatched juveniles 10-20 mm. in length with well-developed tentacular cirri shows that in one species of Diopatra at least, these structures are present from the earliest stages. I have finally decided to leave the matter open for the present and merely give the characters of the form with tentacular cirri below. NOTEs ON AN ‘EPIDIOPATRA FORM’ OF DIOPATRA DUBIA Records: TRA.73(3), 74(1), 80(2). Notes: This is a form of Duopatra dubia which agrees with the above description with the following exceptions. Tentacular cirri are absent. Occipital antennae have ceratostyles which are 2-3 times the length of their ceratostyles. The hoods of the pseudocompound setae are considerably longer and the gills are even more poorly developed. These ‘Epidiopatra forms’ differ from Epidiopatra hupferiana var. monrot in pseudocompound setae, in tube formation, in pigmentation, in number of gills and in the shape of the frontal antennae which are broad and flat instead of elongate and cylindrical. Again the ceratophores of the occipital antennae are much more richly branched. The gills are very similar to E. hupferiana from tropical West Africa but the other characters are still distinctive and the tubes are not plastered with shell fragments. Diopatra cuprea (Bosc) 1802 Diopatra cuprea. Augener 1918, p. 350, text-fig. 39 (fartim). Hartman 1944, p. 54, p. I, figs. 9-14. Diopatra neapolitana (non Delle Chiaje) Crossland 1903, p. 132, pl. 14, fig. 1. Day 1934, p. 54. Day 1957, p. 92 (partum). Notes: I have re-examined my own specimens from Portuguese East Africa and have compared them with specimens in the British Museum from the Gold Coast. They all agree with Hartman’s description of D. cuprea collected on the Atlantic coast of the U.S.A. near Bosc’s type locality (Charles- ton). The pigmentation is diffuse brown and no clear pattern could be distinguished, except that there is a dark internal spot at the base of the dorsal cirri of the first two feet. Tentacular cirri are rather longer than usual, almost as long as the ceratophores of the long occipital antennae. When the cuticle was removed from an occipital antenna, 15-20 broken longitudinal rows of clear cells are seen projecting into it. Pseudocompound setae are strongly bidentate and have well-developed hoods. Comb-setae have 18-25 teeth. The presetal lobe of the 5th—15th foot is small and symmetrical. THE POLYCHAET FAUNA OF SOUTH AFRICA 351 It is very probable that many of the records of D. neapolitana from the Indian Ocean which refer to specimens with bidentate pseudocompound setae and comb-setae with numerous teeth really refer to D. cuprea. Diopatra cuprea var. punctifera Ehlers 1908 Diopatra punctifera Ehlers 1908a, p. 78, pl. 10, figs. 1-11. Records: ‘TRA.143(2); False Bay: 25 records from 7—64 metres on shelly sand; MB.81(3); KNY.6(1); LIZ.3(c); SCD.1(2), 25(c), 26(6), 33(1), 50(c), 58(1), 61(c), 63(3), 74(1), 78(2), 80(1), 94(2). Description: This Cape material differs from tropical specimens of D. cuprea in several minor respects but these differences are constant and for this reason it is as well to refer them to a separate variety for the present. The differences concern the pigment pattern, the cuticle of the occipital antennae and the comb-setae. The colour pattern though often faint shows a transverse row of 4 brown spots across the back of each of the first 4-8 segments near the posterior margin thus - - - -. In the anterior branchiferous region the two inner spots on either side of the mid-dorsal line spread and fuse with the outer spots so that the whole back is brown with a white streak down the middle. When the clear cuticle is removed from one of the occipital antennae, scattered clear cells are found. The comb-setae have rather fewer teeth than in the stem form. In three speci- mens comb-setae appeared on the 6th, 8th and oth foot and in different parapodia the number of teeth on the comb-setae was 9, 14, 13, 14, 9, 10, 15, 14, 11, 15, 18 giving a range from g—18. The feet are of the usual form without any distinctive features. In the first five the setigerous lobe is oblique and in setigers 8-15 the presetal lobe is symmetrical without any inferior projection. Ehlers (19084, p. 78) has included D. neapolitana as a synonym of D. cuprea and records specimens from 16°36’S/11°46’E (off South West Africa), 33°50'S/25°48’E (Port Elizabeth). He states that the colour pattern is variable and his synonymy suggests that he had more than one species in his collections. Diopatra punctifera Ehlers (1908a) recorded from the Agulhas Bank at 35°19'5/20°15’E is not clearly described. Setae of the first 4 feet are stated to be *hellgelebraun, zusammengesetzt: der dunne Schaft lauft mit einem wenig hakenformig gekriimmten, 0-06 mm. langen Endglied aus, das mit einfachen gedeckten Zahn endet, die Deckplatte ist tiber den Endzahn hinaus verldngert ; vereinzelt stehen daneben feine einfache Capillarborsten (Taf. X, fig. 7)’. Reference to fig. 7 shows neither a simple capillary nor a unidentate pseudocompound seta but a strongly bidentate seta. The comb-seta is shown with numerous teeth. Dr. Hartwich of the Berlin Museum very kindly sent me the type material of D. punctifera for further examination. There are two specimens, both with mud tubes without any shell fragments. The colour is faint and although the 352 ANNALS OF THE SOUTH AFRICAN MUSEUM anterior branchiferous region is brownish, no colour pattern remains nor are there any ‘eye spots’ between the parapodia to which Ehlers refers. The median ceratophore has 8 rings and the median occpital antenna is a little shorter than the inner laterals with a cuticle which shows a few scattered cells. The tentacular cirri are longer than usual, almost as long as the cerato- phores. The mandibles have straight tapered shafts and both the mandibles and the maxillae are weakly chitinized. The maxillary formula is given by Ehlers. Anterior parapodia are of the usual shape without any peculiarities; there is no presetal lobe other than a swelling from the 6th setiger onward. The gills start on the 5th foot, have fairly slender trunks and filaments whose length is 4-5 times the diameter of the branchial trunk. The first few feet have 1-2 capillaries and about 4 pseudocompound setae with well-developed hoods and bidentate tips; the secondary tooth is well developed but distinctly smaller and more slender than the terminal one. Bidentate acicular setae of the usual shape first appear on the 15th foot and comb-setae are present on the same foot. The comb-setae are fairly long and have 15-18 fine teeth set on the very slightly oblique blade. Ehlers’s type and my specimens from the same locality agree very well though the colour pattern of the type has faded and the comb-setae have rather more teeth than usual. Epidiopatra gilchristi n. sp. (Fig. 11 af) Records: SCD.33(1), 100(10), 103(2). Diagnosis: A long slender species with a tough, translucent tube which is sometimes annulated (fig. 11a). Description: ‘The holotype is an incomplete specimen from SCD.33 Bagi removed from its annulated tube. It is 55 mm. long by 0-5 mm. wide with more than 150 segments. There appear to be large eggs at the posterior end but the tube is so tough that it is impossible to remove them without damage. The body is creamy brown in alcohol with faint brown spots in the occipital antennae. The head (fig. 11d) bears a pair of subspherical palps, stout subulate frontal antennae and five occipital antennae born on 5 ringed ceratophores. ‘The inner laterals have one or two blunt projections on the basal rings of the ceratophores. The ceratostyles of the median and inner lateral antennae are three times as long as their ceratophores and extend back to setiger 5 but those of the outer laterals are considerably shorter. The head is generally pale in alcohol but there are brown flecks on the prostomium between the antennae and a row of ocular specks may be discerned outside the bases of the inner lateral antennae. The peristome is longer and darker than the succeeding segments. There are no tentacular cirri. The jaws are pale and very weakly duaehee The mandibles have white cutting edges and their shafts are so pale that they are hardly distinguishable THE POLYCHAET FAUNA OF SOUTH AFRICA 353 from the muscle. They appear to be straight and tapered. The maxillae are mainly colourless but the teeth are tinged with brown and Mx. V are toothless crescents of a darker brown. The dental formula is: Mx. I =1 (left) + 1 fee) ix) Ti — 38 -- 7; Mx. Il = 7 + 0; Mx. IV = 6+ ?10; Mx. V = o + o. The first three feet are ventro-lateral in origin and are directed forwards. All the lobes are flattened against the side of the body probably due to compres- sion inside the tube. The first foot (fig. 11e) has a subulate dorsal cirrus, a low presetal fold, then the setae and then a conical postsetal lobe and an inferior rounded papilla. The ventral cirrus is short and blunt. The next two or three feet are generally similar but thereafter the ventral cirrus is reduced to a low glandular pad and the postsetal lobe decreases in size. The first gill (fig. 11f) appears on the fourth foot as an outgrowth from the small dorsal cirrus. It has a stout trunk bearing 6 finger-like filaments in an open spiral. Succeeding gills are smaller and the last arises from setiger 7 so that there are four pairs of gills in all. Specimens dredged off Natal (station NAD.10) also had four pairs but in this case they started on setiger 5. Specimens from SCD.100 and SCD.103 had only one gill on setiger 5. The post-branchial feet each consist of a fingerlike dorsal cirrus, a rounded setigerous lobe with a sheaf of setae and minute post-setal papilla and below this a cushion-like ventral cirrus. The setae of the first three feet consist of about 8 stout pseudocompound setae (fig. 115) with strongly bidentate tips and well-marked hoods. The dorsal cirrus contains 3 fine acicula and the setigerous lobe has 3 stout ones whose pointed tips just pierce the surface. In the fourth or fifth foot broad-bladed capillaries appear and by the 6th foot the pseudo-compound setae have gone and the first comb-seta appears. It has about 12 teeth. Two bidentate acicular setae appear in the post-branchial feet. This species is easily distinguished from E. hupferiana v. monroi by the character of its tube (fig. 11a). Epidiopatra hupferiana Augener var. monrot Day 1957 Epidiopatra hupferiana Augener var. monrot Day 1957, p. 92. Records: FAL.219(1), 245(1 juv.), 328(1), 365(1), 376(1), 378(2); MB. 88(1). Notes: ‘The tubes are fragile and covered with debris,—hydroid stems, pieces of alga, shell fragments, mud and flocculent matter. The worms are typical and most have only 3 pairs of gills but a large 60 mm. specimen had 4 pairs. In fresh specimens 4 broad brown streaks extend along the back but these tend to fade to a uniform brown on preservation in alcohol. I have examined Monro’s specimen in the British Museum (registered number 1930:10:8:1372) and find it agrees with my type, but the colour has faded and the ceratophores of the occipital antennae lack lateral branches. The latter character is variable. 354 ANNALS OF THE SOUTH AFRICAN MUSEUM Fic. 11. Epidiopatra gilchristi: a Tube; b pseudocompound seta from the first foot; c bidentate acicular seta; d anterior end; e anterior view of first foot; f anterior view of 4th foot. Leptoecia antarctica: g pseudocompound seta; h bidentate acicular seta. Rhamphobrachium capense: i, anterior end; j anterior view of first foot; k, k1 pseudocompound seta and unidentate variation (LIZ.25); 1 comb-seta. THE POLYCHAET FAUNA OF SOUTH AFRICA 355 A juvenile with 42 segments measuring 9 mm. (FAL.245.J) was obtained which probably belongs to this species. The frontal antennae are ovoid, the ceratophores of the occipital antennae have small lateral branches, tentacular cirri are absent and no gills have been developed. It is interesting to note that eye specks are still present behind the inner lateral antennae and the ventral cirri are cirriform on only the first 3 instead of the first 5 feet as in the adult. Rhamphobrachium capense n. sp. (Fig. 11 7—l) Records: TRA.152(1); FB.307(c), 322(1); FAL.58(p), 117(1), 219(c), 378(1); MB.62(1); ?LIZ.25.T(1); SCD.89(2). Diagnosis: Gills as single filaments from 30th—4o0th foot; anterior pseudo- compound setae tri- or bidentate with clawed hoods. Description: ‘The holotype is a well-preserved specimen with 78 segments measuring 34 mm. The tail end is missing. It was selected from among 30 specimens dredged in False Bay. The tube is weakly constructed of mucus with adherent fragments of shells, coralline algae and a few sand grains. Fresh specimens show dark marks on the prostomium, palps, ceratophores of the antennae, parapodia and there are two rows of spots on the dorsum. All of these markings fade in alcohol and the type is colourless apart from tiny black eyespots at the base of the inner lateral antennae. The prostomium (fig. 112) is ovoid with cushion-like palps, ovoid frontal antennae and 5 short stout occipital antennae which just reach back to setiger 1. Each of these antennae consists of a stout bulbous ceratophore with 2 rings at the base and a slightly longer subulate ceratostyle. The peristomium is narrow and the tentacular cirri when laid forward do not reach the bases of the occipital antennae. The first few segments are usually tilted upward so that the first two feet point forward, the 3rd—5th obliquely downward and the rest are normally lateral. Anterior feet (fig. 117) have subulate dorsal and ventral cirri, a low presetal lip and two unequal conical postsetal lobes of which the superior is much the larger. The inferior postsetal lobe disappears after the first 3—4 feet. Then the superior postsetal lobe is reduced and on the 12th foot it is no more than a low rounded cone. The dorsal cirrus retains its structure throughout but becomes reduced in size. Between the 30th and goth segment a single branchial filament grows out from the dorsal cirrus and soon greatly exceeds it in length and continues to near the end of the body. The short ventral cirrus is conical for the first 4 feet, reduced on the 5th and on subsequent segments it forms a ventral glandular pad. The first three feet each have a few very small capillaries and 12-18 long pseudocompound setae (fig. 11k) projecting forward. The apex of each is bidentate or occasionally tridentate though the third tooth is minute and the 350 ANNALS OF THE SOUTH AFRICAN MUSEUM bivalve hood which covers the apex terminates in tiny claws. The shaft has two rows of spinules along the inferior margin. In the single specimen from Algoa Bay (LIZ.25.T) which is otherwise similar to the Cape material, the pseudo- compound setae were unidentate or minutely bidentate, the shafts had finer spinules and the ends of the hoods though bent, were not clawed. The next few feet (e.g. the 8th) lose the pseudocompound setae and have about 8 winged capillaries while 2 acicula with long tapered tips project from the surface. In posterior feet either the capillaries are modified to develop long slender tips like the acicula, or the acicula themselves become more numerous and project further from the parapodium. Apart from these very tapered capillaries or acicula there are 2-3 fine comb-setae (fig. 11/) with about 12 teeth and 1-2 brown acicular setae with the usual bidentate apex and guards. This species differs from those previously described by the position and nature of the gills and the structure of the pseudocompound setae. FAlyalinoecia tubicola (Miller) 1788 Ayalinoecia tubicola. Fauvel 1923, p. 421, fig. 166 7-g. Records: AFR.831(1). Leptoecia antarctica Monro 1930 (Big ree.) Leptoecia antarctica Monro 1930, p. 133, fig. 50. Records: FAL.131(2 juvs.), 159(2); ?SCD.3(8). Notes : The False Bay material agrees well with Monro’s specimens dredged off the South Shetland Islands in 1,080 metres. The present specimens are rather smaller, the only complete individual measuring 23 mm. by 0:8 mm. for 70 segments. Tubes are missing. The worms are uniformly pale in spirit, but small eyes are visible external to the bases of the inner lateral occipital antennae. The frontal antennae are ovoid to sausage-shaped, the occipital antennae have ceratophores with 4 rings and the ceratostyles are at least 5 times the length of their ceratophores. The median antenna which is shorter than the inner laterals reaches back to setiger 8. Tentacular cirri are absent and the peristome is about the same length as succeeding segments. The first three feet project outwards and downwards, but succeeding ones change until over most of the body the parapodia are dorsolateral. Dorsal cirri are always cirriform. Anterior ones are approximately equal to the segmental length but over the rest of the body they are much shorter and roughly equal to the setae. The setigerous lobe of the foot is never prominent. 6 THE POLYCHAET FAUNA OF SOUTH AFRICA 357 The postsetal lobe is cirriform for the first 3 feet, then decreases and is not distinguishable after the 8th foot. A presetal lobe is not developed. The ventral Cirrus is cirriform on the first three feet and thereafter becomes a glandular pad which becomes continuous with the setigerous lobe from about the roth foot. The first 3-4 feet contain about 4—6 pseudocompound setae (fig. 11g) with bivalve hoods and bidentate apices, the second tooth being much smaller and more slender than the terminal one. Winged capillaries appear about the 4th foot and by the 8th foot there are 4 capillaries with blades well marked off from the shafts. Two bidentate acicular setae (fig. 114) appear in the oth foot and comb-setae with about 14 teeth further back. An average foot in the middle of the body has 4 winged capillaries, 1-2 fine comb-setae, 2 stout acicula with fine tapered and projecting tips, and 2 bidentate acicular setae with guards. The posterior end of the body bears 2 pairs of anal cirri which are a little shorter than neighbouring segments. The above description reveals several minior differences from Monro’s types with which the present specimens have been compared. In particular, South African specimens have more elongate frontal antennae, they have eyes and the secondary tooth of the pseudocompound seta is distinctly smaller and smore slender than the terminal one. The 8 specimens from station SCD.3 are doubtfully referred to L. antarctica. The material consists of a number of fine horny tubes up to 40 mm. in length and 0-5 mm. in diameter attached to a stone. The basal parts of the tubes have sand grains attached to them but the distal parts are naked and erect. Moreover several of them are twisted into a fine spiral. The worms inside are of course more slender than the False Bay specimens but seem to agree in structure. Monro’s specimens had mud tubes; the tubes of the False Bay specimens are unknown, and until more is known the identification of the SCD.3 material is doubtful. Subfamily LumMBRINERINAE Lumbrineris albidentata Ehlers 1908 (Fig. 12 a—b) Lumbrinereis albidentata Ehlers 19082, p. 97, pl. 13, figs. 7-13. Eeccoras ABN. 736(p); TRA.41(7), 74(2), 80(6), 110(1); 11g(c), 116(1), 143(c), TB.303(1), 309(1); FAL.g5(1), 117(1), 206(1), 228(2), 238(2), 241(1), 243(1), 250(1), 251(10), 328(1 juv.), 345(1), 349(1), 375(1 juv-), 378(1); SCD.105(1). Notes: Ehlers’s type was small and incomplete and the present material which includes numerous specimens of all sizes, now makes it possible to supplement the original description. The prostomium is conical and there is a dorsal slit at its junction with the peristomium containing nuchal sense organs. The jaws are large and in juvenile specimens the first 3-4 segments are expanded 3 58 ANNALS OF THE SOUTH AFRICAN MUSEUM to accommodate them; in adult specimens this swelling is not noticeable. The mandibles are heavily calcified and the maxillae are quite characteristic, the formula being: Mx. I] = 1 + 1; Mx. IT = (2-3) + (2-3); Mx. III = 2+ a; Mx. IV = 0 + 0; the teeth of Mx. II are often edged with white and in one juvenile (AFR.736.Q) gave the impression of having a double row of teeth. Mx. III are very small; Mx. IV are very large plates with a black margin in which a distinct tooth is not differentiated, but the whole forms a cutting edge, thus it is represented in the formula as o + o. The maxillary supports are long and triangular without marked notches at their bases. Anterior feet (fig. 12a) have lamellate lobes which are longer than deep. The presetal lobe is at first small, but in middle feet it is considerably larger, and in posterior feet (fig. 125) it is almost as long as the postsetal lobe. Both lobes project outwards and upwards but are never as long as the setae, and much shorter than the posterior lobes of L. bifilaris or L. meteorana. Ehlers’s type lacked posterior segments so he does not describe this bilabiate condition. There are long compound hooks from the first setiger changing to simple hooks at the 30th setiger and persisting to the end of the body. Winged capillaries are also present from the first foot but these decrease posteriorly so that at the 50th foot there is only one, but thereafter there are usually one or two in most posterior segments. The foot contains four yellow acicula. Lumbrineris cf. meteorana Augener 1931 ? Lumbrinereis meteorana Augener 1931, p. 300, fig. 8. Records: SB.177(1), 199(1); WCD.23(1), 26(2). Notes: Augener described an anterior and a posterior fragment as follows: Body very slender. Prostomium conical. Mandibles very pale with a tooth- shaped process near the symphysis. Maxillary formula: Mx. I=141; Il=5 +5; W1 = (1 or 2) + (1 or 2); IV = 1+ 1. Third maxillary pla. with indistinct teeth—possibly 1 or possibly 2. Anterior feet with low presetal and postsetal lobes. Posterior feet with long threadlike presetal and postsetal lobes of equal length. Winged capillaries restricted to anterior feet. Hooded compound hooks in the first 20 feet but replaced by simple hooks over the rest of the body. Type locality: 17°13'S/11°43’E, off the coast of Angola. My specimens agree perfectly with the above description except in regard to the maxillary formula. In my specimens the formula is 1 + 1; 3 + 3; ?2 + Pa;1 + 1. Mx. II have three very stout almost bilobed teeth and Mx. III is a cutting plate which in some cases shows no teeth at all and in others shows two small projections. Mx. IV is a relatively large plate with a pale centre and a dark edge from which a single tooth projects. It may also be added that the prostomium is oval rather than conical, that the postsetal lobe of anterior feet is only slightly shorter than the postsetal one, that both lobes increase slowly in size over the middle of the body but in the last 20 segments or so both lobes increase enormously to form long threadlike projections. The acicula are pale THE POLYCHAE1 FAUNA OF SOUTH AFRICA 359 Fic. 12. Lumbrineris albidentata: a anterior view of anterior foot; 56 anterior view of posterior foot. Lumbrineris heteropoda var. atlantica: c anterior view of 12th foot; d anterior view of posterior foot. Lumbrineris brevicirra: e anterior view of anterior foot: f anterior view of middle foot; g anterior view of posterior foot. Lumbrineris magalhaensis: h anterior view of anterior foot; 7 anterior view of middle foot; 7 anterior view of posterior foot. 360 ANNALS OF THE SOUTH AFRICAN MUSEUM and the hooks change from compound to simple at the 14th foot which is a little earlier than stated by Augener. Since all the other characters agree so well, the difference in Mx. II is surprising. Augener’s type should be re-examined and it may be that the anterior end he described does not belong to the posterior region with bilabiate feet. | My specimens differ from L. bifilaris Ehlers in the structure of the maxillae and in having jointed hooks anteriorly. In these characters it is closer to L. albidentata Ehlers described above but a side-by-side comparison proves that they are distinct. L. albidentata grows to be a much larger, stouter worm but even when compared with a juvenile it is evident that L. meteorana is longer and more slender, that Mx. III are larger plates with less distinct teeth, that the postsetal lobe of anterior feet is conical not lamellar and that both lobes of posterior feet are much longer and more threadlike than they are in L. albidentata. Lumbrineris heteropoda Marenzeller var. atlantica nov. (Fig. 12 c-d) Lumbrinereis heteropoda Monro 1930, p. 137. Monro 1936, p. 154. Records: LAM.44(1); SB.136(1), 179(1), 183(1), 202(4), 203(10); AFR. 882(1), 1224(1), 1535(1), 1545(1), 1554(1), 1576(1), 1335(1); TRA.68(c), 70(c), 71(fc), 77(c), 80(c); WCD.15(6), 19(3), 21(4), 23(9), 26(45), 28(6). Notes: These are very large worms, a complete specimen being over 300 mm. long, 5 mm. wide and iridescent reddish brown in colour when alive. The prostomium is short and conical and the maxillary formula as usual is Mx. T=1+1; Mx. I1=4-+5; Mx. il=2-+2; Mx. IV =f but the secondary tooth on Mx. III is poorly developed. The maxillary supports are heart-shaped and nearly as broad as long. In anterior feet (fig. 12c) the presetal lobe is short and swollen while the postsetal lobe is ear-shaped and as deep as long. The feet soon increase in length and the postsetal lobe becomes relatively longer until at about the 6oth foot it reaches the tips of the capillaries as a finger-shaped organ. In the posterior part of the body (fig. 12d) it is much longer than the setae and often three times as long as the basal part of the parapodium. In anterior feet the setae are all winged capillaries with brown acicula embedded in the flesh. The superior group of capillaries have very long slender blades. Short-bladed simple hooks appear about the 4oth foot and for the next 30 segments, both capillaries and hooks are numerous. Thereafter both types of setae become less numerous and in posterior feet there are about 4 hooks and one capillary. It is stressed that capillaries may be found even near the end of the body. The present material has been checked as identical with specimens described by Monro (1930) from Tristan da Cunha and, as has been THE POLYCHAET FAUNA OF SOUTH AFRICA 361 shown earlier (Day 1957, p. 94) there are small but constant differences from material recorded in the intertidal zone of the tropical Indian Ocean. Lumbrineris cavifrons Grube 1869 Lumbrinereis cavifrons. Day 1953, Pp. 437, text-fig. 6 a—d. Records: TRA.152(1); False Bay: 30 records from o-35 metres on rock, gravel and shelly sand; MB.49(5), 57(1), 85(1), 87(4); LIZ.2(1), 18(3), 29(2), 35(2); SCD.40(1), 89(1). Lumbrinerts latreilli Aud. & M.-E. 1833 Lumbrinereis latreilli. Fauvel 1923, p. 431, fig. 171 m-r. Records: FB.307(1), 319(1); FAL.334(1); TRA.132(5); SCD.50(1). Lumbrinerts tetraura (Schmarda) 1861 Lumbrinereis impatiens Claparéde, Fauvel 1923, p. 429, fig. 171 a-i. Lumbrinereis tetraura Day 1953, P- 435- Records: LAM.22(1), 35(10), 38(1), 49(1), 52(1); SB.189(10) ; LB.300(c); SH.204(1), 415(1); TB.320(1); WCD.21(1); FB.331(1); FAL.58(p); LIZ.2(6), =7(1)- Lumbrinerts coccinea (Renieri) 1804 Lumbrinereis coccinea. Fauvel 1923, p. 432, fig. 172 g—n. Day 1953, p. 436 with synonymy. Records: TRA.122(1); WCD.8(1); FAL.8(p), 113(2), 126(3), 127(1), 144(1), 156(7), 162(3), 171(4), 182(1), 214(2), 269(2), 275(4), 327(1), 371(1). Lumbrineris harimani Day 1953 Lumbrinereis harimani Day 1953, p- 437; fig. 6 e—m. Records: FAL.245(1); MB.23(1), 88(1); LIZ.19(7), 25(3); SCD.58(4), 89(1). Notes: Some of these specimens are much smaller than the holotype and the simple hooks appear as early as the 23rd segment as against the 45th in the type. Lumbrineris brevicirra (Schmarda) 1861 (Fig. 12 e-g) Notocirrus brevicirrus Schmarda 1861, p. 117. Lumbriconereis brevicirra Ehlers 1904, p. 35, pl. 4, figs. 13-20; pl. 5, figs. 1-2. Records: TRA.73(1); FAL.359(1). Notes: The present specimens lack a posterior end. The prostomium is short and conical with a nuchal pocket at the junction with the peristome. The 362 ANNALS OF THE SOUTH AFRICAN MUSEUM maxillary formula is 1 + 135 + 5; ?2 + ?2;1 + 1. Mx. III are cutting plates with 1-2 indistinct teeth. ‘The maxillary supports are heart-shaped. Anterior parapodia (fig. 12e) are small and each has a poorly developed presetal lobe and a well-developed postsetal lobe which is compressed, deeper than long and roughly triangular with a superior point, rather like a dog’s ear. Towards the middle of the body (fig. 12f) the whole foot grows longer, the presetal lobe becomes obvious and the post-setal lobe is reduced; further back still it is similar to, and not much longer than the small pointed presetal lobe (fig. 12g). The tail end of the worm is missing. The anterior setae include both simple hooks and winged capillaries. The capillaries which have very long slender blades, start on the first foot and continue to the middle of the body (about segment 50). The simple hooks appear about the 12th foot and continue to the posterior end (segment 120). The blade is at first very long so that the anterior hooks give the impression of being capillaries with broken tips, but further back the blade decreases in length until it is only 2-3 times as long as broad. The acicula are pale throughout. This South African specimen has been compared with a New Zealand specimen in the British Museum (No. 1928.2.29.156) identified by Benham (1927). Unfortunately the anterior setae are broken. Ehlers (1904, p. 36) states that Mx. II have 5 teeth on the left and 7 on the right, an unusually high number. However, his figure (p. 5, fig. 1) suggests that the number is smaller as do Schmarda’s original drawings. Lumbrineris magalhaensis Kinberg 1864 (Fig. 12 h-7) Lumbrinereis pettigrewi McIntosh 1885, p. 239, pl. 36 figs. 7-9; pl. 17A, figs. 11--15, text-figs. 4-6. Lumbrinereis magalhaensis. Ehlers 1897, p. 74. Monro 1930, p. 135. Hartman 1948, p. 93, pl. 14, figs. 1-3. Records: FAL.352(1). McIntosh’s record of L. pettigrewt is station 141 dredged in 98 fathoms off the Cape at 34°41’S/18°36’E. Notes: McIntosh’s type of L. pettigrewi is in the British Museum. An examination showed that the prostomium is long and conical; the dental formula is 1 + 1;4 + 4;1-+ 1; 1-+ 1. The maxillary supports are short and broad with practically no notch at the base, Mx. III are curved cutting plates without a distinct tooth, and are best represented in the formula as I + I. The presetal lip of anterior feet (fig. 12h) is a low ridge. The postsetal lobe of anterior feet is flattened, has a rounded edge and is deeper than long, but further back (fig. 122) it becomes more regularly digitiform. Even in the posterior part of the body (fig. 127) it is much shorter than the setae. The presetal lobe increases in size but remains a little shorter than the postsetal lobe throughout. The majority of the setae are broken, but one or two com- THE POLYCHAET FAUNA OF SOUTH AFRICA 363 pound hooks remain in the 12th foot of one specimen and a few simple hooks without swollen ends were found in the posterior feet of another specimen. The acicula are yellow. According to McIntosh the capillaries which are very long and slender are restricted to the anterior part of the body. However, his account is very confused for he figures simple hooks in anterior foot of ‘a variety’, black acicula in one specimen and pale ones in another. Monro 1930 recorded L. magalhaensis from South Georgia and his specimens which were examined in the British Museum were found to be practically identical with McIntosh’s L. pettigrewt. However, the following minor differences were noted. The prostomium is conical but short, the dental formula is the same, but the maxillary supports are slightly longer being 1-5 times as long as broad. In these complete specimens it may be seen that compound hooks are present from the first setiger. The postsetal lobe of anterior feet is again very deep but here not deeper than long. Specimen FAL.352 is an anterior half of a worm which has been compared with the type of L. pettigrewt and appears to be identical. In this fresh specimen the setae are unbroken and it can be seen that the compound hooks start in the first foot and persist to the 19th where they are replaced by simple hooks. All hooks have short blades. The parapodial lobes are short throughout and in posterior segments the presetal lobe is only slightly shorter than the postsetal one. Ehlers (1897) described two forms of prostomium, one long and one short but both conical. This type of variation, probably due to the method of preservation, is quite common in the genus. Ehlers however has made one error in his description. He states that Mx. IV has two teeth. Both McIntosh’s specimen of L. petiigrewi from the Cape and Monro’s specimen of L. magalhaensis from South Georgia have Mx. IV in the form of a cutting plate with an undulating edge, but not two distinct teeth. _ Hartman (1948) has redescribed Kinberg’s type material of L. magalhaensis which consists of several specimens. One was without jaws and obviously dissected by Kinberg. In other characters however, this specimen agrees with the description given above, and Kinberg stated that L. magalhaensis has Mx. III with one tooth, and this is the interpretation accepted by Ehlers and Monro. Hartman dissected other specimens of the type material and found that in these Mx. III has 2 teeth and the maxillary supports are twice as long as broad. I suggest that these specimens are different from the one dissected and described by Kinberg and seem closer to L. latreilli. Arabella tricolor var. caerulea (Schmarda) 1861 Arabella iricolor var. caerulea. McIntosh 1904, p. 46, pl. 4, figs. 16-17. Day 1953, p. 439, fig. 6n. ihecords:, ¥B.305(1), . 319(3); FAL44(1),. 51(1), 58(2), 69(1), 80(2), 235(1), 245(1), 249(2); MB.23(1), 41(1), 42(1), 49(2), 56(5), 67(1), 85(1); LIZ.29(1); SCD.40(2), 89(6). 364 ANNALS OF THE SOUTH AFRICAN MUSEUM Arabella mutans (Chamberlin) 1919 Arabella mutans. Monro 1933, p. 501. Records: FAL.184(1), 229(p). Notes: These worms have slender bodies with the segments as long as broad. The prostomium is large and oval with four eyes in a line just in front of the prostomium/peristomium junction. The mandibles are strong and black; the maxillae have long filiform supports and a dagger-like median appendage. The first pair of maxillae do not form strong hooks, but all maxillary plates have the anterior tooth stronger than the succeeding ones, the formula being: Mx. 1=8+ 8; Mx. T=7+ 7; Mx. TIl=6+6; Mx) 1Vi22e Mx. V=1-+1. Parapodia have a pimple-like dorsal cirrus and no ventral cirrus. The foot has a low rounded presetal lobe and a fingerlike postsetal lobe. Between these are 3-4 winged capillaries with serrations at the base of the blade and two acicula with projecting filiform tips. Drilonereis falcata Moore 1911 (Fig. 13 a-e) Drilonereis falcata Moore 1911,.p. 298, pl. 20, figs. 150-154. Hartman 1944, p. 179. Drilonereis filum (non Clap.) Monro 1936, p. 158. Records: FAL.219(1), 352(1). Notes: A single anterior fragment of 66 segments was obtained. The prostomium (fig. 13a) is depressed, oval in plan and lacks external eyes though there is a faint suggestion that internal eyes may be present. The mandibles (fig. 13c) are stout, black and roughly triangular with a short hinge line. The maxillae (fig. 13b) have long filiform supports which are very narrow where they join the main fangs (Mx. I) and a dagger-shaped median piece which is blackened throughout. Mx. I are stout hooks with toothed bases, Mx. II are rectangular with the first tooth rather larger than the rest, Mx. III have an anterior large fang-like tooth and small denticles behind, Mx. IV and V which are very close together, each consist of a single fang. In the following dental formula the difference in size of teeth is not indicated as is sometimes done. Mx. I=8+6; T=8+ 6; Wl=4 +3; IV =1+4 1; VV —= ee The first of the two achaetous segments is largely fused with the prostomium but leaves a crescentric depression on the dorsal surface. Anterior feet (fig. 13d) are small and may be partially retracted but succeeding ones rapidly increase in size. There is no dorsal cirrus. The presetal lobe of the foot is a low semi- circular ridge at the base of the bluntly conical postsetal lobe. From the groove between the lobes project a fan of about 6 winged capillaries and the filiform tips of fine acicula may just be seen. A stout yellow aciculum appears on the 18th—24th foot. The parapodia elongate posteriorly (fig. 13e) and the aciculum THE POLYCHAET FAUNA OF SOUTH AFRICA 365 (sometimes 2) projects further until it almost reaches the end of the conical and rather short postsetal lobe. The above description agrees with Moore’s figures and description apart from minor details. There are more teeth on the main fangs though Moore mentions some obscure crenulations ‘as well as 3-4 distinct small teeth’. Hartman states that there are numerous teeth at the base of the forceps. Again Moore shows one large tooth only on Mx. III whereas here there are 2-3 denticles as well, but Hartman states that Mx. III have 4-5 smaller teeth as well as the longer one. Monro (1936) described a specimen dredged off the Falkland Islands which he referred to D. filum while noting the differences from typical European forms of that species. I have examined the specimen in the British Museum, whose registered number is 1936:2:8:2355, and find it to be D. falcata so that the range of this species is now from California (o—-172 fathoms) to 46°18’S/ 65°02’W (100 m.) and False Bay (18-88 m.). Drilonereis monroi n. sp. (Fig. 13 f-2) Drilonereis sp. Monro 1930, p. 142. Records: AFR.718(1), 801(1), 1319(1), 1535(1), 1544(1), 1545(1), 1554(1), 1576(1), 1581(1); TRA.68(2), 70(1), 77(2), 88(1), 89(1); FAL.237(1), 240(1). Description: ‘The type was obtained from station AFR.718. It is 220 mm. long by 3 mm. broad for 250 segments. It is rusty red in colour and extremely tough and wiry. The prostomium (fig. 13f) is depressed, broadly triangular and lacks eyes. The pharynx usually protrudes slightly. The mandibles are lacking, there being merely toughened skin on the floor of the mouth. The maxillae (fig. 13g) have long filiform supports and an unusually short, shield- shaped median piece. Mx. I are stout hooks not denticulate at the base and Mx. V are missing, the formula being Mx. I =1-+ 1; Mx. II = (6-8) + (6-8); Mx. III = 3 + 3; Mx. IV = 1 + 1. On Mx. II and III the first tooth is much larger than the others. Anterior parapodia (fig. 13h) are small but the feet increase in size posteriorly. There are no dorsal or ventral cirri. The presetal lobe of the foot is a low curved ridge and the posterior lobe is a short blunt cone. Between these project a sheaf of winged capillaries and a stout, blunt, yellow aciculum which is evident even on the first foot. Posterior feet (fig. 137) are longer, the superior part of the presetal lobe being more expanded, and about half as long as the postsetal lobe. The Drilonereis sp. described by Monro 1930 from Tristan da Cunha has been compared with these South African specimens and is identical. This species is related to D. nuda Moore described by Hartman (1944) from California and D. major Crossland (1924) from Suez and Zanzibar, both 366 ANNALS OF THE SOUTH AFRICAN MUSEUM Fic. 13. Drilonereis falcata: a anterior end; b maxillae; c mandibles; d anterior foot; é posterior foot. Drilonereis monroi: f anterior end; g maxillae; A anterior foot; i posterior foot. THE POLYCHAET FAUNA OF SOUTH AFRICA 367 of which lack distinct mandibles. It differs from D. nuda in not having teeth at the base of the pincers (Mx. I) and in these respects it is closer to the type of D. major which I have dissected and whose dental formula is Mx. I = 1+ 1 (main fangs only); Mx. II = 6 + 6; Mx. III = (3-4) + (3-4); Mx. IV = 3 + 3. However, in D. major the teeth on Mx. II are of fairly even size while in D. monroi the first tooth is much larger than the rest. There also tend to be fewer teeth on Mx. III + IV but these plates are more variable in all species. Other differences are in the setae which are always longer in D. monroi. Moreover the projecting aciculum of D. monroi appears in the first foot and in D. major in the 15th. It may also be noted that in the posterior feet of D. major the presetal lobe remains rudimentary while in D. monrot it becomes enlarged, — nearly as long as the postsetal lobe. Notocirrus australis n. sp. (Fig. 14 a—d) Records: FB.306(1); FAL.22090(1). Description: ‘The type material from False Bay includes five fragments, two anterior and three posterior ends. It is estimated that a complete worm would measure about 100 mm. by 1 mm. with about 200 segments each about three times as broad as long. The colour is uniformly pale in alcohol. The prostomium (fig. 14c) is conical with 4 eyes in a transverse row at the posterior margin. The outer pair is larger than the inner pair. ‘The jaws consist of well-developed mandibles and maxillae. Each mandible (fig. 14a) is strong and triangular and the two are narrowly joined in the median line. ‘The maxillae (fig. 14b) consist of five pairs of toothed plates so closely crowded together that it is difficult to distinguish the teeth of one plate from those of the next; in fact plates I and II overlie one another in one specimen and in the other, plates I and II are fused on the left side. The anterior tooth of each plate is hooked and slightly larger than succeeding ones but not to the extent seen in the genus Drilonereis. The supports (or ‘carriers’) are long and slender and the median piece is very faint and pale. ‘The maxillary formula in one specimen is Mae F — 7 (left) + 7 (right); I= 7+ 8; WW =7+6;1IV—=5+4; V= 1 + 1 and in the other where Mx. I and II are fused on the left side: Mx. I ee — i: (left) + 7-and 9 (right); II] = 9 + 6; IV = 6 + 4; ee ales ae The first two apodous segments are rather shorter than the subsequent ones, all of which bear well-developed parapodia of increasing size. Each parapodium (fig. 14d) has a minute, pimple-like dorsal cirrus above the setigerous lobe. The presetal lobe is rudimentary and the thumb-shaped postsetal lobe is at first almost ventral in position, but in posterior segments it moves round to the normal posterior position. Each parapodium bears about 3 winged capillaries 368 ‘ANNALS OF THE SOUTH AFRICAN MUSEUM and a stout yellow needle-like aciculum which projects almost as far as the postsetal lobe. The capillaries have rather broad wings which are smooth except for a few serrations at the base. The genus Wotocirrus has most recently been reviewed by Hartman (1944). The present species is closest to NV. lorum Ehlers from the Magellan area and N. californiensis Hartman from Southern California. By the courtesy of the director of the Hamburg Museum I have been able to examine Ehlers’s type. Unfortunately the jaws have been removed and I have nothing that I can add to the original description. The main difference between the three species lies in the dental formula of the maxillae. As has been mentioned, these plates are small, crowded together and overlapping, and thus difficult to read. NV. lorum is reported to have only 4 pairs of maxillary plates but Ehlers’s figure 125 suggests that the 4th and 5th dental plates have not been separated. Again it may be that the left MX. II on which three teeth are shown is partially covered by (or possibly fused to) Mx. I. WV. californiensis is very close to the present species except that Mx. II has 13 teeth on the left side and the usual minute dorsal cirrus is not figured above the parapodium. It is probable that further work will reveal that the maxillae are more variable than has been suspected and a number of species will be sunk in the synonymy. Fauvel (1923, p. 451) regarded WV. scoticus McIntosh 1869 as a doubtful species but a re-examination of the type material now in the British Museum (registered numbers 1921-5-1-1681-86) shows that it is definitely a WNotocirrus though MclIntosh’s description of the structure of the feet is very confused due to his having inverted his preparation. ‘Thus what he described as a dorsal cirrus is really the postsetal lobe, and what he referred to as a ventral cirrus is really the dorsal cirrus. According to Ehlers (1875, p. 55), V. scoticus isa synonym of N. tricolor (Johnston) 1865. A brief summary of the characters of the type of WN. scoticus may now be given. Body brown, rather small for the genus and segments markedly moniliform for the segments are about as broad as long with deep intersegmental constrictions between one and the next. Prostomium conical with two pairs of eyes which fade in alcohol. Jaws consist of a pair of well-developed, triangular mandibles and 4-5 maxillary plates with the usual long supports. The dental formula is doubtful. McIntosh (1910) gives a drawing (p. 62, fig. ga) which shows a number of larger and smaller teeth which may be variously interpreted. One interpretation is Mx. I = (left) 7 + (right) 8; Mx. II = 12+ 7; Mx. III] =6-+ 7; Mx. IV = 5 + absent; Mx. V = 1 + absent. The dental apparatus on which his drawing was based has not been preserved; indeed all the jaws of the type material are missing except one from the Porcupine Expedition (registered number 1921-5-1-1685). ‘These jaws are very small and, as usual, difficult to read. My reading is Mx. I = (left) 7 + (right) 8; Mx. I] = 6 + 7; Mx. WI = 5 + 5; Mx. IV =3 +4; Mx. V = doubtful. It will be obvious that the dental formula quoted depends on the inclusion or omission of minute or partially formed denticles on the maxillary plates quite apart from individual variation. The distinction between THE POLYCHAET FAUNA OF SOUTH AFRICA 369 oe Fic. 14. Notocirrus australis: a mandibles; b maxillae; c anterior end; d middle foot. Drilognathus capensis: e mandibles; f vestiges of maxillae; g anterior end; A middle foot; 7 posterior end. 370 ANNALS OF THE SOUTH AFRICAN MUSEUM individual species of the genus Wotocirrus rests largely on the dental formula so that one becomes sceptical as to whether there really are several species and not one world-wide one. Drilognathus capensis n.g. et sp. (Fig. 14 e-1) Records: 5 specimens found in the body cavity of Onuphis holobranchiata dredged in Lamberts Bay 18.1.57 (station LAM.11). The holotype is a complete specimen which is twisted, but if straight would measure about 3 mm. by 0-3 mm. with about 60 segments. The whole worm is creamy white, and is tapered posteriorly. The prostomium (fig. 14g) is well marked off from the succeeding segment. It is ovoid, somewhat tapered anteriorly and lacks appendages. No eyes are visible on the surface but when cleared in glycerine two eyes are visible posteriorly. Dissection of one specimen showed that the mandibles (fig. 14¢) are well developed and black. They are of the usual Drilonereis type and there is no sign of the recurved rostra which has been described for Labrorostratus parasiticus. The maxillae (fig. 14f) are represented by a blacked cuticular ridge on the dorsal wall of the pharynx. The length of this black cuticular streak is reminiscent of the long maxillary supports of Drilonereis and Arabella but no fangs or distinct maxillary plates were seen. The first two segments lack parapodia. Each of these achaetous segments is about 4 times as broad as long. Succeeding segments up to the middleof the body have well-developed parapodia (fig. 14h) of the usual Lumbrinereis type, though the presetal lobe is rudimentary and even the postsetal lobe is no more than a blunt cone. From the middle of the body onwards the parapodia are progressively reduced, first to mere lateral papillae and eventually over the last 10-15 segments they are entirely lacking. The pygidium however is well developed (fig. 147) and has a pair of large ventro-lateral lobes projecting outwards at right angles to the body. Each parapodium is supported by a stout aciculum of the Drilonereis type. It is a bluntly pointed yellow needle which in most parapodia seems not to pierce the skin, but in some it does, and then just projects in front of the post- setal lobe. There are no other setae, a fact which immediately distinguishes this species from related genera. Pettibone (1957) gives a most useful key to endoparasitic members of the Arabellidae. It is with considerable hesitation that I name this as a new genus, for according to Pettibone the young stages of Notocirrus occur as parasites in the Onuphidae and Pettibone’s figures 5L and M of the jaws of Notocerrus ? spiniferus are not unlike those of the present specimens. However, all the other genera that have been described have setae in the parapodia whereas Drilognathus has merely a well-formed aciculum. THE POLYCHAET FAUNA OF SOUTH AFRICA 371 Subfamily DorvILLEINAE Dorvillea neglecta (Fauvel) 1923 Staurocephalus neglectus Fauvel 1923, p. 447, fig. 179 1-g. Dorvillea neglecta Day 1953; p. 439- Records: SB.183(1); SH.366(1). Dorvillea egena (Ehlers) 1913 Stauronereis egena Ehlers 1913, p. 501, pl. 35, figs. 1-6. Augener 1918, p. 377, pl. 5, fig. 102, 103, text-fig. 40. Records: FAL.284(1). REFERENCES AUGENER, H. 1913. Polychaeta I,—Errantia. In Michaelsen, W. & Hartmeyer, R., eds. Die Fauna Stidwest-Australiens, 4, 65-304. Jena: Fischer. AUGENER, H. 1918. Polychaeta. In Michaelsen, W., ed. Beitriége zur Kenntnis der Meeresfauna Westafrikas, 2, 67-625. Hamburg: Friederichsen. AUGENER, H. 1922. Litorale Polychaeten von Juan Fernandez. Jn Skottsberg, C., ed. The natural history of Juan Fernandez and Easter Island, 3, 161-218. Uppsala. AUGENER, H. 1931. Die bodensassigen Polychaeten nebst einer Hirudinee der Meteor-Fahrt. Mitt. zool. St. Inst. Hamb., 44, 279-313. Bairp, W. 1869. Remarks on several genera of Annelides belonging to the group Eunicea, with a notice of such species as are contained in the collection of the British Museum, and a description of some others hitherto undescribed. 7. Linn. Soc. (Zool.), 10, 341-361. BenuaAM, W. B. 1927. Polychaeta. Nat. Hist. Rep. Terra Nova Exped., Zool. 7, 47-182. BERGSTROM, E. 1914. Zur Systematik des Polychaeten Familie der Phyllodociden. Zool. Bidr. Uppsala, 3, 37-224. CLAPAREDE, E. 1868. Les annélides Chétopodes du Golfe de Naples. Mém. Soc. Phys. Genéve, 19, 313-584. CrossLanp, C. 1903. The marine fauna of Zanzibar and British East Africa from collections made by Cyril Crossland in the years 1901 and 1902. Polychaeta. Part II. Proc. zool. Soc. _ Lond., 1903, 2, 129-144. CROSSLAND, C. 1924. Polychaeta of tropical East Africa, the Red Sea and the Cape Verde Islands collected by Cyril Crossland, and of the Maldive Archipelago collected by Professor Stanley Gardiner, M.A., F.R.S. Proc. zool. Soc. Lond., 1924, 1-106. Day, J. H. 1934. On a collection of South African Polychaeta, with a catalogue of the species recorded from South Africa, Angola, Mosambique and Madagascar. 7. Linn. Soc. (Zool.), 39, 15-82. Day, J. H. 1951. The polychaet fauna of South Africa. Part 1: The intertidal and estuarine Polychaeta of Natal and Mosambique. Ann. Natal Mus. 12, 1-67. Day, J. H. 1953. The polychaet fauna of South Africa. Part 2: Errant species from Cape shores and estuaries. Ann. Natal Mus. 12, 397-441. Day, J. H. 1957. The polychaet fauna of South Africa. Part 4: New species and records from Natal and Mocambique. Ann. Natal Mus. 14, 59-129. DELLE Curaje, S. 1825. Memorie sulla storia e notomia degli animali senza vertebre del regno di Napoli. Naples: Stamperia. DELLE Curaje, 8. 1841. Descrizione e notomia degli animali invertebrati della Sicilia citeriore osservati vivt negli anni 1822-1830. 3. Naples. Envers, E. 1864-68. Die Borstenwiirmer (Annelida Chaetopoda) nach systematischen und anatomischen Untersuchungen dargestellt. Leipzig: Engelmann. Enters, E. 1875. Beitrage zur Kenntniss der Verticalverbreitung der Borstenwiirmer im Meere. Kk. wiss. Kool. 25, 1-102, 372 ANNALS OF THE SOUTH AFRICAN MUSEUM En ers, E. 1897. Polychaeten. Hamburger magalhaensische Sammelreise. Hamburg: Friederichsen. EHLERS, E. 1905. Neuseelandische Anneliden. Abh. Ges. Wiss. Géttingen, Math.-phys. Kl. (N.F.) 3, 1-80. EHLERS, E. 1908a. Die bodensdssigen Anneliden aus den Sammlungen der deutschen Tiefsee- Expedition. Wiss. Ergebn. ‘Valdivia’ 16, 1-167. Euters, E. 1908). Polychaete Anneliden der Angra Pequena-Bucht. Jn Schultze, L. Zoologische und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-Afrika. 1, 43-50. Jena: Fischer. Denkschr. med.-naturw. Ges. Jena 13, 43-50. EHLERS, E. 1913. Die Polychaeten-Sammlungen der deutschen Siidpolar-Expedition 1901-1903. Dtsch. Stidpol.Exped. 13, 397-598. EKMAN, S. 1953. 407mm Fic. 8. The ranges of variation of the transverse diameter (breadth) of the permanent dentition. | = mean value. in graphic form (figs. 7, 8 and g). It may be observed from the mean and from the relative dimensions of the teeth (fig. 10), that there is a decreasing gradation of size for the postcanine teeth from back to front, except for M2 which is in general longer (A-P) in the upper jaw, and broader in both jaws, than M3. The decrease in size of the teeth in both upper and lower jaw is gradual, except for P2, which is relatively 396 ANNALS OF THE SOUTH AFRICAN MUSEUM much smaller than P3 (more in the lower than in the upper). In the family Giraf- fidae, P1 has disappeared, while the unusual decrease in relative size observed in Pa, linked with the fact that in some specimens it is also absent (see chapter 3), may suggest an eventual disappearance of Pa. elralte) ec pmelonaiaelie ULTS Nt ae Tr: A-P Index 30. 40 50° GO. 7085580 190), 100", NO 120 G0. 40 Zo Fic. 9. The ranges of variation of the transverse/A—P index of the permanent dentition. { = mean value. Both the premolars and the molars of the upper jaw tend towards a square shape, the breadth being normally even greater than the A—P length (fig. 10). In table 7 (and fig. 9) this is indicated by the fact that the mean of the transverse/A—P 397 MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA ‘siyopsngopauvs vffpii4y y[Npe ULOpoU Jo y199} Jo suoisusuTIq *L ATAV], by.g | 6.9 | V.cz1—G.16 | 629. FPP-Lo1 | tat | g6-9 | gz.s | 6.gE—L.E% | Loz. Fzg.cE | az1 | SP.L | gz.a | 1.66 —G.Ez | ggr.F1F-08 | gz1 | ew ¥z.G |09-G | 0.gz1—g-16 | G6. F6L.gor1 | of: | Hg. | 16-1 | S.gE—g.La | Lor. $g.8E | 181 | gS.9 | 60.2 | F.SE—0.gz | 91. FaL.18 | LE1 | Ww 1Z.G | ¥6. v.611—0.98 | LoS. —16.€01 | 9&1 | Pz-g | 66.1 | 9.VE—G.Lz | ggt- 16.08 | 9&1 | $9.9 | L6.1 | S.E6—z.gz | gg1-F 14.62 | 161 | {IN gi-g | 41.9 | ¥.gS1—o.F11 | SOL. Fog-c&1 | Grr | E.G | gS.1 | 1.66—8.92 | Lb1.FS9.6z | G11 | 8.9 | 1b.1 | §.S2—P.91 | gai. Fhz.3% | Sar | $2.9 | ob-g | g-QS1—o.101 | oLL. Far-Ger) gii |G1-9 | &4.1 | ¥.1€—1.€% | ogi. s1.gz | git | Lo.4 | 09-1 | 6.40—€.L1 | 1h1.F09.%z | gar ed oG-L |Gz.6 | 0.6v1—0.o1 | 9Sg. Fgg-se1 | 411 | Gg-4 | 96.1 | L.o€—G.1z | 191. 6G.G% | Z11 | 09-4 | 9S.1 | 9.-¥a—g-Gr | 1b1.FQl.08 | gai ed 00-g | 06.6 | %.94 —g.Lb | Ghb. Fia.19 | oar | 6g-4 | Hg.1 | 1-93—L.L1 | gg1-F08.Ez | 01 | oL.9 | 9G. | &.bh—1.2€ | gaz. Q1-gE | gar | *W 6z.G | z1.V | 6.06 —L.69 | GSGE. FLL.LL | GE1 | 10.9 | ab.1 | F.gs—E.g1 | Sa1.F19.% | SE1 | 6.9 | 09.1 | 6.EE—L.bz | GG1.FzE.0€ | gf1 | Ww GG.9 | S1.G | 1.96 —6.4g | c&b. —6S.g/ | Ebi | ob.g | Sb.1 | €.G%—1.L1 | 031. FLz.z% | sbr | Lag | gl.1 | £.€6—1.Gz | Shr. Fo9€.92 | gh1 | TW OG.h | GL. 6.go1—L.69 | so0g. —s0.6g | gz1 | o9-g | gg-1 | S-Sz—o.b1 | 991.F 91.1%] gz1 | 09-9 | 2g-1 | 1.g2—0.16 | bh1. + PS.bz | gai "d L9.6 | 06.9 | 0.S11—9.99 10g. -%6.16 | var | £E.o1 | €1.6 | G.€s—6.c1 | 161.F&S.0% | Par | LV.L | Gg.1 | g.a—F.g1 | 6F1. Fg0.c2 | Gat a 13-61 | GG.11 | 0.v31—SG.9G | oo1-1F1b-Lg | Gor | 60.G1 | GE.z | F.61—¥.g | Gaz. FLG.G1 | Gor | 6.11 | a1.% | 9.23—§.E1 | Gos. F0g-41 | Gor oF v9.61 | 90-g | ¥-0L —g.g2 | gf1-120.1b | LV | 0G.c1 | gsr | G.11—P.G Igi---¥c.01 | oS | 96.91 | ag-V | L.16—G.6 | 999. 6&.Gza | 6h 8) LL.€1 | 06.6 |0.€01—L.0G | ogh.1- 16.14 | gb | Gz.11 | Qr-1 | 0o.c1—z%.G | ogi. +gh.o1 | EV | gz.g1 | HP.2 | G.61—G.g | gLE.F96.E1 | ab oT £g-01 | 9%-g | 0.46 —6.09 | o1c-169.g4 | OG | gz.11 | Gz.1 | g-g1—0.G | gL1.+ 0.11] 0G | GE.Gr | G1.z | §.g1—g.g | 108.00.F1 | 1S aT 90-31 | 99-8 | 0-46 —1.cG | glz.1-bE.EL | gb | 00-61 | zb.1 | g.31—6.0 | cov. 26.01 | 6V | gh.G1 | 18.6 | o-61—9.8 | gz€.00.S1 | 0G | UuOT]eLIe A, ae UOT}eLIV A ean uOTIeLIE A sas A > jo osury as aN | N A 4 jo osury eae N A jo osury mae | N p00], xopuy g — V/‘Asuriy, ("WIUI) dSIOASURIT, (wu) I—V 398 ANNALS OF THE SOUTH AFRICAN MUSEUM index is greater than 100. On the other hand, in the lower jaw, the A-—P length is always greater than the breadth, so that the teeth tend to be rectangular in their longitudinal axis. ae Giraffa Giraffa_camelopardalis 304 x 32.6 31.7x 33.8 267x309 222x296 226x281 207x255 5 263x242 iJ 3x24. ' a. sxe te 3x77 ra ane @& = 2 ~ = DM4 DM3 DM2 Cc ig in ty JS € & S = ae eae 18.5 x66 206x137 x93 329 x 171 oc = 381 x 23.3 303x236 283x222 245x217 220x205 178x155 = ste a ae ee = 253 x 102 M3 M2 M1 PS) PSY Spo c |. Bee Fic. 10. Graphic representation of the relative ranges of dimensions (A—P length and breadth) of adult and milk (immature) dentitions. Mean dimensions are indicated. Although similar in breadth to the incisors, the canine is approximately twice the length of each incisor, probably because of its bilobate nature. The incisors show a decreasing length (A—P) from front to back, although they have approximately the same breadth. Index N Range of Variation M s Vv p2 Length = 122 73°8— 109°2 91°940°527 5°86 6:37 P2 Breadth pe 106 74°0—104°8 91°28-+0°429 4°38 4°79 tnd Transv./A—P P? 3 8-1 3 ee Bee nde ScanEnEienenraraietiemtien 10 ‘O— 131° : 0°841 : : i Transv./A—P P% ? : ate cae : ? ? P3 Length =i 121 84°9—124°0 101°60+0°572 6°30 6-20 ps | Breadth par 109 79°6—105°9 95°02-+0'373 3°88 4:08 Ind Transv./A—P P? 3 6 PAN aye 6-8 a ndex = 4—————___—__ 4 oye . ‘19-0: ‘B80 : Transv./A—P P4 Me ge ay os 4 ae q P Length ar 109 56-2— 94°8 80:40+0°746 7°76 9°65 3 MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 399 Index N Range of Variation M S V P, Breadth = 110 57°7— 98:8 76-70-+0°665 6-92 9°02 3 “2 Transv./A-P P, : : ndex Tea eas 109 74°2—130°0 95°84+1:042 10°84 11°31 P Length = 123 77°8— 103°1 90°23-+0°396 4°36 4°83 a P, Breadth = 116 68-8—105:0 93°17-0°570 6-10 6°54 4 meee Es 8-8—128 64-£0-78 8-6 8 ndex aeeiney. ALP SP) I2!I 7 FS MELO! 103 40°7 I O 29 Be la Length Ve 127 80°5— 101°6 93°12+0°348 3°90 4°18 Ml! Breadth ME iO 81:3— 98:8 91°44-+0°320 3°52 3°84 = Transv./A—P M1}! ee P 3 3 ndex The eae 120 4°3—108°9 QS°12+0°412 4°50 4°5 M2 Length a 122 84°6—140°8 104°13+0-°665 7°32 7°02 M2 ; Breadth Me 116 92°2—119°0 104°12+0°413 4°41 4°23 ret Transv./A—P M2? 5 oe Ewe 6 e ——— EEE 7 . . ° A ° ° ndex Transv./A-P M3 114 C71) Wao) 100°15+ 0°61 5 55 M, Length Ni. 135 74°8— 106-0 93°7640°415 4°82 5°14 2 M, Breadth “Ta 134 83:9—101°8 94°38-+0°276 3°18 3°36 2 Ind Transv./A—P M, ae 3 ie 3 nae SS SE Rr eae I “Q—I ° ° ° ° ° 2 x Transv./A-P M, 33 9—127 100°99= 0°520 on 59 M Length aa 125 69°8— 95°4 _— 79°58-+0°421 4°68 5°88 3 M, Breadth a 120 87-1 —133°4 101°34+0°372 4.06 400 3 os Transv./A—P M, ReCaEAG 6 6-68 eee 109°0— 150° 127° oO: : . et Transv./A—P M, ue 2 oma? ales ane 2° . Ps Breadth il 108 73°6—106-2 91 -03-++0°469 4°84 5°31 Pt Breadth Mi 108 83°0—109°5 95°740°458 4°72 4°93 P, Length —— 124 69:0— 89-6 80-96-0388 4°31 5°32 M, B Breadth a 121 76-6— 104-2 92°62+0°485 5°34 5°76 2 TABLE 8. Dental Index for lengths and breadths, as well as the index for the transv./A—P ratios of successive permanent teeth of G. camelopardalis. 400 ANNALS OF THE SOUTH AFRICAN MUSEUM The ratios of the length and of the breadth of successive postcanine teeth (P2/P3, P3/P4, M1/M2, M2/Ms3) are expressed as the ‘Dental Index’ for premolars and molars. For both their lengths and breadths, the range of variation, mean, standard error, standard deviation and coefficient of variation were determined (table 8, fig. 11). In addition, the ratios of the transverse/A—P dimensions have been compared and expressed as an Index. This data has proved particularly valuable in assisting the authors to determine the diagnosis of isolated fossil teeth. In this respect, the ratios of the breadths of P?/M!, P4/M}, P IM and of lengths of P,/M, have also been helpful (table 8). B. DEctipuous DENTITION The measurements of 331 milk-teeth are summarized in table 9 and figs. 10, 12. It is clear that the front teeth (incisors and canines) and the first molar, both in the mandible and the maxilla, have the largest coefficient of variation, while the postcanine teeth show much less variation of both length and transverse breadth. In most of the cases, the breadth seems to be slightly less variable than the length. A-P (mm.) Transverse (mm.) Tooth Range of Range of N ~— Variation M o Vv N_ Variation M a Vv 1, 23 «9th —13°5) «-11°53+°225 1°08 9:36] 22 6:0— 8-2 +7-48+-127 o-60 8-02 le 24 7°2—13°5 «11°27+°308 «61°51 13°399| 23 5°7— 8:2 7:22+-133 0°64 8-86 1g 27 7:°7—18:8 10°85+:-219 1°14 10°50] 26 5:°0— 7:7. 6:67+°149 0°76 11°39 Cc 30 13°8—21-4 18°54+°285 1°56 8-41| 30 4°7— 84 6:°64-°135 0°74 11-14 DM, | 37 *12°2—19°8 “14°7532201 81°77 12:00] 397. «7-2 11-7 «= -: 395-12) ae eg DM, | 38 18°0—24°:1 20°64+:230 1°42 6:87] 36 11°7—15°7 13°72+°155 0°03 6-77 DM, | 39 28°0—35°9 32:99+°227 1°42 4°30] 37 15°71—18-9 17°14+°171 1°04 6-06 DM? | 36 16-0—20°4 18:37+:240 1:44 7:83] 36 16:3—19°5 17°79+°193 1°16 6°44 DM? | 38 22-2—27-2 24°47+°209 1°29 5°27| 38 19°3—23°1 91-29---169)) 7 -Oo mage DM? | 38 24:3—29:0 26°33+-191 1°18 4°48] 38 21°-3—26°3 24°:28+-189 1°17 4°82 TABLE 9. Summary of measurements of 331 milk teeth. (N = number of specimens; M = mean given with standard error; o = standard deviation; V = coefficient of variation). The breadth/length index has also been calculated (table 10). Here again, the front teeth and first lower molar show a very high degree of variation, reaching 17°38%. The data indicate that all the milk teeth are longer than broader, the ratio being close to 100% for the upper molars. They are all relatively longer than the corresponding permanent teeth (fig. 10), and this feature is responsible for the lower index (table 10). MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 401 DENTAL- INDEX LENGTH BREADTH B/L INDEX M D mis: 79.58 Sait: Weil ‘ Sos Tees eaves DENA SUIS TS M3 py, eee aged ee Be Rete eee ee Fic. 11. Ranges of variation of the Dental Index for length and breadth and of the breadth/length index of permanent teeth in Giraffa camelopardalis. Mean values are indicated. ANNALS OF THE SOUTH AFRICAN MUSEUM 402 “syopingojauvs vffvuy ut yI294 (snonproap) ¥IIw JO suoIsuswAIp asroAsUe) PUL d-V 2} jo (402) uoNMerIeA Jo JUSTOYJI09 pur ‘( +) UOTIeIAap prepurys ‘(@) uvow a ) uoryeIreA Jo sa8uvs ay} Jo uoNeIUasoidos otydeIn “ZI ‘Oly ~ wu 0& 76 gl cl 9 wu ¥€ 82 co ce] OL —@q——_ 28-b ge | pW] ge TI) a iad . ge ewd 8E cas ' 9¢ | cWN0| g¢ 90-9 LE Ywa 6€ LL-9 * 9& fwa ge £8: +? Le Cw LE pill a ee of ) 0€& 6E-tL +? 92 e Le 98:8 +? EZ C1 bo ak a2; = C6 =| al he ee TDA N N wre ASNVUL MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 403 Range of M in corresponding Tooth N Variation M o Vv permanent tooth i 22 52°2— 8or1 66-14+ 1-149 6-80 10°28 72-34 ly ag 48°7— 97:0 65°26-+2-187 10°48 16°05 78:69 Ig 26 41°3— 814 62°62+ 1-603 8-16 13°03 71°QI c. 29 24°2— 44°38 35°78-+ -680 3°66 10°22 41°02 DM, 37 41°Q— QI-2 64°44 1-894. 11°52 17°87 78°59 DM, 36 52°6— 78-2 66-58-+ 1-086 6-52 9°79 Daaly| DM, 37) 46°:2— 58-0 52°30-+ +509 3°10 5°92 61-21 DM? 35 86-5— 106-6 96-68 -+ 1°035 6-12 6-33 103°QI DM? 38 78-8— 96-0 86:96-+ -699 4°31 4°95 106-79 DM? 38 82:0—I101°0 92°28-— -741 4°57 4°95 107°44 TaBLeE 10. Breadth/Length (Transverse/A—P) Index of 331 milk teeth, compared with the same index in the corresponding permanent teeth. (N = number of specimens; M = mean, given with standard error in milk teeth; o = standard deviation; V = coefficient of variation). The ‘Dental Index’ has also been determined for the immature teeth (table 11, He. T 2). DENTAL INDEX LENGTH BREADTH B/L INDEX DMo DM3 DM3 DM4 Fic. 13. Ranges of variation of the Dental Index for length and breadth and of the breadth/length index of deciduous teeth in Giraffa camelopardalis. Mean values are indicated. 4.04. ANNALS OF THE SOUTH AFRICAN MUSEUM Range of : Index N Variation © M s Vv DM? . Length ue 36 62:7— 81°9 75°38+0-811 4°87 6-46 DM? : . Breadth == 36 (7551023 83°94+0-678 4°07 4°84 = Transv./A—P DM? F ndex Transv./A_P DM 306) 94°0— 127°4 110°342-0°629 3°72 3°37 DM? Length DM: 38 85°7— 102°2 92°93+0°545 3°36 3°61 DM? Breadth DMA 38 82°4— 92°7 87°47+0°472 2°91 Bae rade Transv./A-P DMS , 3 ie 3 ndex Transv./A-P DM# 3 51D TOL] 94°11 0°735 4°53 4°01 DM, ae . Lengtn OM 37 54°6— 96:3 71°05+0:983 5°98 8-41 3 Breadth = 35 56-2— 80-2 68-95-- 0-991 5°86 8°49 3 A Transv./A-P DM, 2 : : an e€ ———— . —= . . . . ° n ji Transy (AcP ODM, 25 3°7—134°2 90°544 3°113 10°40 i 7 Length awe 38 55°6— 72°3 63:11-+0°581 3°58 5°67 4 DM, Breadth = 36 7FO-2— O21 79°73+0°821 4°93 6-18 4 Transv./A-P DM, Index ——————____ 36 98°7—144:°0 123°52+1°760 10°56 8°54 Transv./A-P DM, TABLE 11. Dental Index for lengths and breadths, as well as the index for the transverse/A—P ratios of successive immature teeth of Giraffa camelopardalis. CHAPTER 6 ROOT ANOMALIES Root anomalies are by no means uncommon among recent giraffes. They will be dealt with in the adult dentition first, and then in the milk dentition. I. ADULT DENTITION Three main types of anomalies have been observed: accessory roots, root reduction and root fusion. A. ACCESSORY ROOTS They are by far the commonest abnormality. Among the 62 adult specimens analysed, not less than 15 individuals (24%) show distinct accessory roots in a total MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 405 . of 38 teeth (table 12). The data indicate that the accessory roots are found exclusively in the last two lower premolars and in the lower molars, the incidence being much higher in M, and M3. ———$_— |, | ————————_ ———————————————————— — P; P, M, M, M, SS ee | LeR ey ee eens | Deal ped ss L Xs x De De x 24291 x x 27137 x x 27752 x x x x 34425 x x 34424 x x x x x 34426 ? ? ? ? 35536 x x x x 51198 x x x x 69403 > 4 x x 83460 x 139696 x x 154033 ? 251797 x TOTAL 5 =I 4 14 14 e's ff TABLE 12. Distribution of accessory roots in adults. (R = right, L = left. In two specimens, it has been impossible to tell from the records whether the anomaly affects the particular teeth bilaterally or not: a ‘?’ indicates that the side is unknown). The most common type is that of a small, thin and short root located between the two major anterior and posterior roots, either on the buccal or the lingual side. It never develops the flattened plate shape (side-to-side) of the two main roots. It is rather circular, hardly extending into the bony socket of the mandible. In some cases it is poorly developed, resembling a very small tubercle and projecting 2 or 3 mm. from the root base, forming a small depression in the alveolar border of the mandible. In this series the accessory root is always situated on the lingual side of the premolars and on the buccal side of the molars. On the premolars it occurs unilater- ally, whereas—with very few exceptions—the accessory root is bilaterally situated on the molars when present. In these cases, the right molar presents the anomaly as the mirror image of the left corresponding tooth. It is a debatable point whether this root is really an accessory root or whether it is just the result of some division or splitting off of one of the main roots. On one specimen only (27137), there isa definite indication of a bilateral division of the anterior root of M,: on both teeth the anterior root is bifid, the distal branch being - clearly situated under the protoconid. In all the other cases, however, the accessory root is central, situated under the protoconid-hypoconid junction, quite separate from the anterior and posterior main roots. _ It is not possible to link in any way the molar accessory root with the presence of an ectostylid: among the 21 cases of molar accessory roots, only 13 correspond to specimens presenting any form of ectostylid. This does not indicate a significant 406 ANNALS OF THE SOUTH AFRICAN MUSEUM correlation between accessory root and ectostylid, especially when one considers that in 86 molars presenting an ectostylid (table 6) there are only 13 which have an accessory root. , B. Root REDUCTION Root reduction has been observed in three cases only, all P,. In these, the tooth has only one single root, round in section and rather conical in appearance, instead of the two normal (anterior and posterior) roots. The specimens are 34423, 34424 and 83460. On the former two, the reduction has only been observed unilaterally, the corresponding opposite tooth root being perfectly normal. Specimen 83460 has only one P, (right), the corresponding left premolar being congenitally absent. The measurement of the teeth concerned in the three specimens are compared with the mean and range of variation of 109 other P, specimens (table 13): Transv./A—P Specimen Taker td Transverse Index 34423 R (reduced root) | 13-0. OG, 88-4 L (normal root) 18°9 15°9 84:0 34424 R (normal root) 14°] 17°5 120°4 L_ (reduced root) 15°5 16-4 105°9 83460 R (reduced root) 13°6 13°5 99°2 L_ (tooth absent) = — — Series of 109 teeth Mean 17°8 15°5 87°4 Range of variation 13°3—22°6 8°4—19°4 56-5— 124-0 TABLE 13. From these figures it may be inferred that, in two of the three observed cases, reduction of the root corresponds to an appreciable reduction in A—P diameter (viz. 22°5% and 23:6% in 34423 and 83460 respectively). ‘The reductions in trans- verse diameter are 21:7% and 12:9% respectively: it is probably not very significant, especially in view of the negative reduction (—o-5) in 34424. Similarly, in these three specimens, the A~P measurement is situated at the lower limit of the range of variation, while the transverse breadth falls within the range of one sigma from the mean value. Comparing further the teeth with the reduced root with the opposite ones which have the normal root, there is a distinct correlation between reduction in root and in size of the crown in one case only (34423). On the other hand, it is remarkable that, with the exception of the above case, a reduction in size of the tooth does not seem to correspond to a reduction of the root. C. RooT FUSION One single case of root fusion has been noted: the M? in 53546 has (bilaterally) a bridge completely fusing the posterior buccal root with the lingual one along the total height of the roots. 2 MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 407 II. MILK DENTITION Root variation is frequently observed in the third lower milk molar DM,. A total of 20 immature skulls, in which this tooth was still bilaterally in situ and sufficiently well preserved to allow observations, were studied. All the 40 DM, have three roots. The two main roots (anterior and posterior) are well developed, extending bucco-lingually, being rather flattened from side-to-side, and corre- sponding to the two extreme pillars of the deciduous molar, but they do not show a very high degree of variability. However, situated between them, there is con- stantly a third root, corresponding to the intermediate pillar. It is rather buccally located, circular in shape, and is shorter than the two main roots. The development of this central root is quite variable. The normal appearance (so-called because it was observed in 28 out of 40 cases) is that of a cylindrical root, sufficiently long to penetrate the alveolar socket of the tooth where it is really ‘rooted’, although less than the other two main roots. This is the general rule for the left DM, (18 cases), and shows in only to of the cases on the right side of the mandible (table 14). Variation Right Left Total ‘Normal’ 10 18 28 Reduced single central root I I 2 Divided central root: 2 normal rootlets I normal and I truncated 2 truncated 3 truncated tm NOD iS~) TOTAL 20 20 40 TABLE 14. Variability in the central root of DM,, and distribution in right and left teeth. In one case, this central root was so small and reduced that it hardly indented the mandible, and could scarcely be called a real root, being not very unlike some of the accessory roots observed on the permanent molars or premolars (vide supra), although it was much thinner. In ten cases, however, the central root was bifid or multiple, possessing two or even three small thin rootlets parallel to each other and originating from the crown-root junction. Here again there appears to be a high degree of individual variation, and one could schematically summarize the obser- vations by distinguishing three main types of ‘multiple’ central root: (i) a double central root, each of the two parallel rootlets being as long and as large as the normal single ones: (ii) in addition to the normal central root, and parallel to it, the second accessory root is much shorter and reduced, hardly reaching the bone where it is never well ‘rooted’; (iii) two or even three small reduced rootlets are found, truncated and just reaching or hardly penetrating the alveolar border. 408 ANNALS OF THE SOUTH AFRICAN MUSEUM CHAPTER 7 APPEARANCE AND VARIATIONS OF THE HORN-CORES Lydekker (1904) attempts to classify Giraffa camelopardalis, distributed over the vast geographical area from the Cape to the Sudan and Ethiopia, according to the colour and the blotching of the skin (see Introduction, page 375) and to the variation of the horns. He describes a gradual transition from south to north from a two- horned animal into one (so far as the males are concerned) with three horns, but he believes that this is by no means regularly progressive, for one finds in the eastern districts of the Continent a five-horned and even a six-horned ‘race’. On this basis, Lydekker finds it possible to distinguish: (a) seven subspecies of the Northern and Eastern giraffes which are all characterized by the development of a large frontal unpaired horn, namely, typica (in Nubia), reticulata (in Somaliland), antiquorum (in Kordofan), cottoni (in South Lago, Uganda), rothschildi (Baringo), tippelskirchi (Kilimanjaro) and congoensis (Congo) ; (b) four other subspecies of the Western and Southern giraffes, whose frontal horn is rudimentary or even aborted, namely, peralta (Nigeria), angolensis (Angola), ward. (northern Transvaal), and capensis (Cape) (fig. 14). The main paired horns are always present, but particularly well developed in rothschild: and still better in wardi. Posterior or occipital horns are very well developed in these two subspecies, but less in cotton. No reliable information on these posterior formations was available for tippelskircht and angolensis. An unusual ‘azygous’ orbital horn was described in cotton: and rothschildi, projecting horizontally outwards from the middle of the frontal border of the right and left orbit respectively. Lydekker’s observations are summarized in the following table. Main pair of Additional or Subspecies Anterior horns horns Posterior horns - ‘azygous’ typica x x — — antiquorum 54 x — — reticulata Bre x ? — cottont x x x x smaller than in smaller than in on right orbit rothschildi rothschildt rothschildi x Ke x sometimes; then well developed well developed on left orbit tippelskirchi x x e, — smaller than in rothschildi congoensis a x ? — angolensis x x ? — wardi x x x — large, aborted well developed stronger than in rothschildi capensis x x — — rudimentary peralta —- x ? — TABLE 15. Schematization of Lydekker’s subspecific classification and horn distribution. (x = present; — = absent; ? = doubtful.) MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 4.09 TYPICA (Nubian desert) ANT/QUOQRUM ( Sudan ) PERALTA ( Nigeria) RETICULATA (Somaliland ) COTTONI \ (Uganda) ROTHSCHILD! (Kenya ) TIPPELSKIRCHI CONGOENSIS (Tanganyika ) ( Congo) ANGOLENSIS (Angola) WARD! (Transvaal) CAPENSIS (Cape) Fic. 14. Map of Africa indicating the distribution of the subspecies of Giraffa camelopardalis, according to Lydekker’s classification (1904). Because of the large amount of material available to the authors, the following observations indicate that Lydekker’s useful information must be somewhat revised. The sexual difference is very marked in the horns which are usually poorly developed in females (fig. 15): thus it is hardly possible to make any subspecific classification on the basis of the horns. The median frontal horn does not exist in females, there being only a slight swelling on the frontal without development of the ‘ossicone’. The main horns, which are always much smaller than in the males of the same catalogued subspecies, measure 8-15 cm. in length from tip to base, the dia- meter not exceeding 3 cm. They are cylindrical and smooth, pointed at their tip, and 410 ANNALS OF THE SOUTH AFRICAN MUSEUM \ x! x2 20884 34423 eee ey — 53546 53549 82002 83458 83459 162988 163324 270594 Fic. 15. Sketches showing portions of the skulls of female giraffes (Giraffa camelopardalis) indicating the major variations in the development of the horns. Lateral views, except X! which is norma verticalis. usually constricted at their base. Posterior occipital horns and azygous horns were not found in the females. ode eee AK ee TN A 8377 b 24292 Fy Die we 34422 94425 34426 53550 53765 oe 57675 pais Fic. 16a. Sketches of portions of skulls of adult male giraffes (Giraffa camelopardalis) showing major variations in the development of the horns. Views are either norma lateralis or verticalis. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA AII Males show a better development of their horns (fig. 16a, b) and it is possible to notice important differences. But does the wide range of variation permit sub- specific distinctions? In order to answer this question, each type of horn will be discussed separately. Se 81821 81822 82001 8B 83460 / MY oa ea adh 2) ' B 139696 154033 155438 163312 165051 200151 251799 279405 Fic. 16). Further sketches of portions of skulls of adult male giraffes (Giraffa camelopardalis) showing major variations in the development of the horns. Views are either norma lateralis or verticalis. A. SINGLE MEDIAN FRONTAL HORN It varies from a simple swelling of the frontal (251799) to a true, very strongly developed horn (155438). The intermediate stages are observed as either a horny plate (154033), a slight eminence (279405), a globular formation (54123), or a rather large and regular cone (200151). Usually the anterior part of the horn has a more gentle slope (155438) which gradually extends on to the nasal bones. Very often on the median sagittal line, anteriorly, are one (155438), two (81821), three (27752), four (81820) or more (27137) smaller nodules; one or two may even be found on the side of the horn (139696) or on the top (81822 and 53765). The profile of the posterior part of the horn is mostly S-shaped (163312), so emphasizing its knob-like appearance. The whole formation is generally symmetrical on both sides of the sagittal plane. In one case however (27137), its transverse horizontal section looked like a crescent open anteriorly and to the right, with both extremities of the crescent converging towards a completely isolated smaller knob. 412 ANNALS OF THE SOUTH AFRICAN MUSEUM B. Parr OF MAIN HORNS All the male giraffes have these two horns very well developed. The shape is commonly that of a rounded column, measuring 15-23 cm. in length, 4-6 cm. in breadth, and the diameter at the tip is 4-7 cm. A difference in size between right and left horn is by no means rare, the one being sometimes 35% longer than the other. Both horns are usually parallel: if not, their divergence from the base is hardly noticeable, so that the external interhorn breadth measured at their tips is always smaller than the external biorbital breadth. The angle formed by the horns with the Frankfort plane is very constant and in all the specimens where this could be measured, it ranged from 48° to 58°. The general shape of the horns is that of a column regularly cylindrical from the base upwards, and at the tip it forms a rounded knob (83460 and 24292). The tip is sometimes flattened (27137), the medial edge of which then projects more than the lateral one. In other cases the column maintains its regular circumference from base to tip (53550, 1396906). These fronto-parietal main horns are usually relatively smooth, having small, shallow vertical grooves, but they may possess a few marked axial ridges (163312 and 8371b), or display small knobs, in series of 2-5, forming irregular crests on the anterior (57675), medial (27752) or lateral (54123) margin of the horn. C. PosTERIOR OCCIPITAL HORNS The posterior ‘horns’ are much more a type of occipital crest or exostosis, con- densed in two parallel eminences, than real horns. The crest is also very variable and may be completely absent. D. SUPPLEMENTARY KNOBS Smaller isolated supplementary knobs are not infrequently developed either on the lateral aspect of the frontal, between the posterior supra-orbital border and the base of the main horns (155438) when there are two or three concentrated in the same area, or on the orbital border itself (54123) when it is comparable to the azygous orbital horn described by Lydekker. But in the specimens studied it was distinctly bilateral and it is a knob rather than a ‘horn projecting outwards horizontally’. On the basis of the above results, the conclusions of Lydekker relative to the taxonomical significance of horn-shape and disposition are questionable. It is not impossible to find typical examples displaying the subspecific characteristics proposed by Lydekker: male 165051, for instance, from South West Africa, with big main horns (23 cm. in length) and a very rudimentary ossicone on the frontal. On the other hand, it is not difficult to find specimens which do not fit in his descriptions, €.g. 251799, registered as G.c. tippelskircht, which certainly originated from ‘Tan- ganyika, has no anterior horns and shows just a slight swelling of the frontal. Similarly specimens 24292, 34422, 34425, 34426 and 83460 which are all males from Bechuanaland (Mababe Flats) definitely have a very well developed anterior horn even though they geographically belong to the group of Angola, northern Transvaal or Cape, which—according to Lydekker’s description—should have no such a large MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 413 horn. Specimens from British East Africa, identified on the basis of their skin as G.c. rothschildi, have very strong main horns (155438 and 200151) whose length, general pattern or morphological appearance could not differentiate them from tippelskirchu. Among both groups one finds long, regularly cylindrical and smooth horns, or shorter, plumper horns with knobs (163312 and 27752; 54123 and 251799). Lateral ‘azygous’ knobs are found in both types (27752, 54123 and 155438), but on the one skull from Uganda, which should be G.c. cottont, no such knobs were found, as should have been expected from Lydekker’s description. Consequently, it would be reasonable not to attempt any subspecific determi- nation of modern African giraffes on the basis of the horn-cores. This principle is extremely important, as will especially be indicated in the discussion on the horn- cores of the extinct giraffids. SECTION II THE FOSSILIZED GIRAFFIDS INTRODUCTION Fossilized giraffid material has been discovered at various sites in Africa, namely: Nort ArricA Oran (St. Charles) (Pomel 1892) ; Bou Hanifia, Chaachas (Reygasse, 1919-20) ; St. Arnaud (Arambourg, 1934, 1948) ; Garet (Garaet) Ichkeul (Arambourg, 1949) ; Douaria (Roman and Solignac, 1934); __ Djebel M’dilla (Arambourg, 1952). EGYPT Wadi Natrun, Garet-el-Multk (Stromer, 1907); SUDAN Abu Hugar (Bate, 1951). East AFricA Omo (Arambourg, 1947); Serengeti (Dietrich, 1937, 1942); Olduvai and Kagua (Hopwood, 1934, 1936); Olorgesailie, Kanam and Rawi (Leakey, 1951). SouTH AFRicA Vaal River (Haughton, 1922); Florisbad (Dreyer and Lyle, 1931); Cornelia and Tierfontein, near Port Allan (van Hoepen, 1932) ; Makapansgat (Cooke and Wells, 1947; Broom, 1948) ; Hopefield (Singer, 1954); some unknown place, probably from the Vaal River gravels (Cooke, 1949). Nearly all the above-mentioned material have been either assembled in Cape Town on loan for study, or examined in the respective Museums where the specimens are housed. In addition, for comparative purposes, an extensive survey of the giraffid material from the Siwaliks was carried out in the American Museum of Natural History (New York) and the British Museum (Natural History), London. It was not possible to obtain the original specimens previously described from the following sites outside Africa: GREECE Veleés (Schlosser, 1921); Pikermi (Gaudry, 1861, 1867); Samos, Salonique (i.e. Bounardja and Vateliik) (Arambourg and Piveteau, 1929). U.S.S.R. Taraklia (Khomenko, 1913). Huncary Baltavar (Peth6, 1885); Polgardi (Kormos, 1911). | FrANcE Mont Lébéron (Vaucluse) (Gaudry, 1873). Turkey Adrianople (Abel, 1904); Maragha (de Mecquenem, 1924). 414 MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 415 In respect of the latter group of specimens all the available literature was studied. All the original material (from Africa) examined will be described according to geographic distribution. CHAPTER I OLDUVAI (OLDOWAY) GORGE, TANGANYIKA A. GEOLOGY The history and the geology of the enormous Olduvai Gorge (fig. 17(a), (b)) have been recorded by Leakey and by Reck, respectively (Leakey, 1951). Five beds are described: Bed I was deposited at the beginning of the East African Middle Pleistocene (early Kamasian), Beds II and IV correspond to the Kamasian and Kanjeran pluvials respectively, the Beds III and V represent the dry inter-pluvial and the post-Gamblian periods respectively. Giraffid remains have been recovered from Beds I-IV, although most of the material comes from Bed II (vide infra). ATL. ALBERT 5 L.EVASI_? 6 Fic. 17a. Map of Central-East Africa indicating some of the more important fossil sites. 416 LETTERS REFER TO AREAS OF SITES NOTE: ANNALS OF THE SOUTH AFRICAN MUSEUM 08 HW PEK +O en “xO ¢ 2 s¢ Leen e &— AY \ MILE Fic. 175. Map of Olduvai Gorge indicating designations of various sites (from Leakey, 1951). MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 417 B. THe MATERIAL The material recovered consists of loose teeth, or fairly complete dentitions, postcranial bones (mainly incomplete) and so-called ‘antlers’ (horn-cores). 1. Previously described. Hopwood (1934) described a right fourth lower premolar (M 14200) as the holotype of Helladotherium olduvaiensis sp. n., and as paratypes he mentioned a rolled fragment of a right mandibular ramus with P,-M, (M 14686, horizon unknown) and a partial hind-limb (M 14687, Bed I). With the discovery of Sivatherium-like ‘palmated antlers’ (inverted commas, ours) from Bed II Olduvai (M 14954-14955) and from Kagua (M 14956), Hopwood (1936) referred the above material to Sivatherium oldu- vaiense. However, Dietrich (1937) stated that on the basis of a metacarpal (E. 122) which he had studied, he had already referred the Olduvai material to the Sivatheriinae (although he does not state whether this was ever published). 2. Observations on material housed in the Coryndon Museum (Nairobt). Material was kindly sent on loan by Dr. L. S. B. Leakey, Curator of the Coryndon Museum, and in addition horn-cores and postcranial remains from Olduvai were also studied in the British Museum (Natural History) and in the Coryndon Museum (Nairobi). Each specimen is described in detail, the specimen being numbered according to the museum registers, and the information given below with each number is found written on each specimen. The inscriptions have the following interpretation: Midge :s 2.2: .. OlduvairGorze. Oldy. .. .. Olduvai (Oldoway) Gorge. Be =... , ..- Indication.of the site (sec-fig. 17(b)). Rey ere Soy Bed ir. 1951, 1952, etc. .. Collected during those seasons. a .. ... The specimen has been found on the surface of the Bed referred to, but it does not necessarily belong to that Bed. M ... .. .. Before a number indicates registration by British Museum (Nat. Hist.). Marsabit Road .. A site about 100 miles east of the southern point of Lake Rudolph in Kenya. Coryndon Museum Guraffid Material A. Horn-cores 86 ns .. Old. BK II 1952. Old. 1.53 .. (Provisional number given by Mrs. Coryndon, in 1956.) Old; ot5s\ 2. (ddem.) Old. 1952 .. SHK II BK II base (S). M 149545 .. Oldy. BK IIS (13.V.935). Old. 3.53... (Provisional number given by Mrs. Coryndon in 1956.) M 17027. .. Oldy. BK IIS (14.V.95). M 17028 .. Oldy. BK IIS (14.V.35). M 17029.) «. Oldy. BK ITS (14.V.935). M 14955 ~.. + (At present in British Museum (Nat. Hist.).) B. Fragments of jaws Upper: F 3655 .. Right maxilla, M!—M4&, Old. II, 1941, in situ. ANNALS OF THE SOUTH AFRICAN MUSEUM Lower 6 .. .. Right mandible, P,—M;, Old. BK II. F 3656 .. Left mandible, M,—Ms, Old. II. I .. .. Right mandible, P,—M,, Old. BK II. 365.. .. Right mandible, M,—P;, Old. SKH II, 1953. g92.. .. Right mandible, M., Old. SHK II, 1953. 3 0.. “2 Right mandible M5. Old2BK IL 93 .. .. Left mandible, M,—M,, Old. BK II, 1952. g2.. .. Left mandible, M,—Ms3, Old. BK II, 1952. gI . Left mandible, M,—M,, Old. BK II, 1952. Neguntiri site. Left mandible, M,— Mg, Old. BK II East 1953. g21.. .. Left mandible, M,—M,, Old. BK II, 1953. C. Isolated teeth Upper: premolars.. F 2989 ?P%, Oldoway 1941 II, in molars .. F 2993 M?, Old. II. 4 M$, Old. BK II. 109)=—s- M, Old. BK II, 1955 F 2992 ?, Old. II in situ. Lower: canine, 2. 97 Old. BK II, 1952. premolars... F 2991 P,, Old. IS, 1941. 2 P,, Old. BK II. 5 P,, Old. BK IL; 7 Pi Old BK AL. 96 P,, Old. BK II, 1952. molars .. 105 M,, Old. BK II, 1953. 132 M,, Old. BK IT, 1953. 95 Mz, Old. BK II, 1952. 120 Ms, Old. BK II, 1955 — M,, Marsabit Road. Fragments of molars, further unidentifiable 166 Old. BK II, 1955. 116 Old. BK II, 1953. 8 Old. too )=6©- Old. BK _IIT,, 1953. g8-—99 Old. BK IT, 1952. = Marsabit Road. — Olduvai surface. D. Post-cranial skeletal remains Forelimb: 116p.. .. Proximal end of ulna. Old. BK II, 1952. 115.2 ..».) Distal end of radius? Old. BK II, 1952. 941 =... ~ Os magnum. Old. BK IT, 1952. 114... .. Distal end of metacarpal. Old. BK II, 1952. Hindlimb: 100" )..5.. - Distal end of feniur- Old. BK IT, 1952. 101, 112 .. Distal end of tibia. Old. BK IT, 1952. 103, 108 .. Calcaneum. Old. BK IT, 1952. 102,107 .. Astragalus. Old. BK IT, 1952. 104, 109, 110 Cubonaviculare. Old. BK II, 1952. 105, 110A .. Cuneiform. Old. BK II, 1952. 106, 111 .. Proximal end of metatarsal. Old. BK II, 1952. Blanes .. Distal end of metatarsal. Old. BK II, 1953. M 14687 .. Almost complete articulated hindlimb, in B. M. (N.H.). Proximal phalanges: 113 .. Old. BK II, 1952. 185 »»,, Old. BR IT. 1953; MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 419 B:g097 :... Old. SII. Peob4. s.. Old. Sif, 1948. E965) J. ew, ic» Old. surface. Sesamoid bone: gaa Sh.” ‘Old’ TI, 1952. DESCRIPTION OF SPECIMENS A. Horn-CoRES 86. Old. BK II 1952 (plate 1(d), (f)). This is a right horn-core. The base is hollow, extending about 170 mm. in depth from the broken edge. In section it is roughly triangular, the rounded apex being posterior. This posterior border becomes increasingly rounded towards the tip, while the more anterior region, rounded at the base, narrows to a well-defined border towards the apex. The anterior surface has knobs on it. The medial surface is convex, both transversely and medio-distally, and presents deep grooves. Three of these are parallel to each other and to the convex anterior border, along which they run, even on to the knobs, following their profiles. Two other grooves are some distance from the anterior border, passing parallel to each other upwards and towards the posterior border. There is also a short, very broad, oblique groove which leaves the base at the anterior convex border and runs obliquely upwards and backwards on the medial aspect. The general curvature and the rotation (twist or torsion) of the horn-core are not marked: it is almost straight. No flange* is visible. This portion has been reconstructed in plaster but it would appear unlikely that a flange would have been present. The first visible knob has been called knob 2. This specimen is very similar to the horn-core from Garet Ichkeul (see Section ITI, chapter 1). Length along posterior (concave) border: 710 Length along anterior (convex) border: 810 Height of the knobs: 20-40 Circumference A-P Breadth* At base 415 150 114 Just under knob 2 350 128 At knob 2 145 89 Just under knob 3 330 11g At knob 3 130 83 Just under knob 4 300 113 At knob 4 135 78 Just under knob 5 280 106 At knob 5 110 64 At ‘tip’ (broken) 75 63 (* Side-to-side diameter.) TABLE 16. Measurements of Old. 86 (mm.). : * The term ‘flange’ is applied to the appearance often presented by the anterior border just above the base when it narrows, flattens and sweeps (flanges) outwards. 420 ANNALS OF THE SOUTH AFRICAN MUSEUM Old. 1.53 and 2.53 (plate 1(c), (e); 1 (a), (d)). 1.53 1s a right horn-core and 2.53 a left. They are very similar to one another, although they may not belong to the same individual because the rotation (torsion) of the horns differ from each other, 1.53 rotating go° anti- clockwise, while 2.53 rotates clockwise (normal) through 140° from base to tip. This indicates the tremendous variation in torsion between horns which are otherwise very similar. Both horns have a marked flange (267 mm. and 155 mm. in length respectively) which terminates superiorly in knob 2 above which there are three other knobs (3-5) along the antero-medial border. The flange of 2.53 is shorter than that of 1.53. On the flange of 1.53 there are four small knobs, the highest of which corresponds to knob 2. These are not present on the flange of 2.53. On both horns there is a knob opposite the middle of the flange, but it is situated on the posterior border somewhat medially. This knob is designated ‘P’ and may be compared with the blade-like projection which led Falconer (1868) and Abel (1904) to term the horn a tri-radiate antler. In both specimens the anterior border first passes upwards and laterally from the base (where it is ill-defined) and at about the middle of the flange it swings towards the medial side. In this way the lower part of the horn develops its rotary twist, which is more obvious in 2.53. Thus the anterior border which at the base is really antero-lateral tends to twist to the antero-superior aspect of the horn, while the posterior border—which is postero-medial at the base— tends to become inferior at the tip. Furthermore, as a consequence of this torsion, the antero-medial surface (which has the marked grooves) comes to lie on the posterior aspect of the horn above the flange region. In specimen 1.53 the postero-lateral surface has no grooves at all, while the antero-medial surface presents two grooves near the anterior border, both parallel to it, and both broad but not deep. The cranial fragment at the base of 1.53 contains a few shallow sinuses but does not display any sutures and it is not possible to assess the correct orientation of the horn to the skull. It would have been valuable to obtain an understanding of the position of the horn on the skull so as to determine whether the criteria the authors have utilized for ‘siding’ the horns are correct or not (see also Section III, chapter 1).* The base of 2.53 is hollowed out to a depth of approximately 300 mm. from the broken edge. There are some sinuses present. The grooves resemble those of 1.53 but they are more numerous, there being four on this specimen. Distances between the knobs 1253 2.53 2—3 141 196 54 110 164. 4—5 Uc k4o 146 * At the galley-proof stage, Dr. L. S. B. Leakey reported to one of the authors (R. S). that a Sivatherine skull complete with attached horn-cores had just been discovered at Olduvai and that a note on it would soon be published by him. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 421 Length of the flange 267 155 Length along anterior (convex) border 950 IOIO Length along posterior (concave) border 840* 690 * A portion of skull attached to the horn is included in this measurement. Circumference A—P Breadth* I-53 2293 £53 2-53 T5535: 2°53 At base 400 390 159 145 127 cere) Middle of flange 410 420 154 172 79 81 At knob 2 380 410 145 165 84. 89 Just under knob 3 300 290 105 IOI At knob 3 330 290 129 104. 76 75 Just under knob 4 280 250 106 go At knob 4 290 260 rie: 97 67 63 Just under knob 5 270 250 99 97 At knob 5 290 290 bry 115 59 58 Tip 170 170 52 59 52 47 (* Side-to-side.) TABLE 17. Measurements of 1°53 and 2°53 (mm.). Old. 3.53 (plate 2) This is a fragment of a left horn-core with part of the proximal portion missing. The tip and distal end are practically intact. At the base the flange is still present, ending in knob 2. Above this is a large knob 3, the only other knob on the anterior border. The hollowed-out portion of the base is very narrow and extends about 130 mm. from the broken edge. On the antero-medial surface there are two deep grooves parallel to the profile of the knobs. There are also two more regular parallel grooves near the posterior concave border. They all extend to the tip. In addition, there are a number of small grooves at the proximal end which peter out in the region of the upper end of the flange. On the lateral surface there are some broad, but less marked, grooves. As in most of the other specimens (except that from Tierfontein) the edges of the grooves are irregular and heaped-up in parts. The fragment, in general, is very similar to the corresponding portion of the Tierfontein horn-core (C 431). The curvature and rotation are rather poorly expressed. The region above knob 3 is almost straight. Length along anterior (convex) border 700+ Length along posterior (concave) border 430-+ Circumference A—P Breadth* At base 360 138 — Middle of flange 360 146 77 At knob 2 380 155 73 Just under knob 3 310 122 70 At knob 3 360 14! 68 100 mm. above knob 3 250 89 67 At tip 150 52 43 (*Side-to-side.) TaBLE 18. Measurements of 3:53 (mm.). 422 ANNALS OF THE SOUTH AFRICAN MUSEUM | Old. 1952 SHK II BK II base (S) and M 14954b Oldy. BK IIS. (Plate 3(c), (d)) These are two closely fitting fragments belonging to the same left horn- core. M 14954b noted by Hopwood (1936), but he did not include the ‘b’. The flange exhibits a small knob 1 at its centre along the anterior border, Beyond knob 2 (at the upper end of the flange) are knobs 3, 4, 5 and 6. Distances between the knobs (mm.) I j=- (2: 83 2 taceral io 185 3 — 4: 140 4 — 5 Ta 1 5 — 6: Sel 6 — tip: 174 There are a number of nutrient foramina (as in 1.53 and 2.53) which are not very large (2-3 mm. in diameter) and not specifically related to the grooves. It exhibits the same type of curvature and torsion as 1.53 and 2.53 but they are much more twisted and curved. In turn, it shows more torsion than Old. 86 and 3.53. Thus it occupies an intermediate position and this series indicates a further grade of variation of the torsion of the Sivatherine horns. As in 1.53 and 2.53, there is a medio-posterior knob (‘P’) on the inferior portion near the base and opposite the flange. The grooves on the antero-medial surface are large, similar to those in the Hopefield specimens. The base contains a few sinuses and is hollowed out to a depth of 80 mm. The horn closely resembles M 14955 (vide infra). Length along anterior (convex) border 840 Length along posterior (concave) border 580 Circumference A—P Breadth* At base 330 117 100 Flange (knob 2) 350 143 78 At knob 3 360 14! 66 Just under knob 4 270 102 aa At knob 4 290 112 62 Just under knob 5 240 go _ At knob 5 260 94 64 Just under knob 6 220 80 ~ At knob 6 240 85 59 Tip 150 49 46 (*Side-to-side. ) TABLE 19. Measurements of M 14954 (mm.). M 17027 Oldy. BK IIS. The tip and distal portion of a right horn-core. The only interesting features are that the extreme tip is somewhat recurved posteriorly and that the grooves on the antero-medial surface have extended to the base of this recurved portion where they end abruptly. The extreme tip is rather rough and irregularly rounded. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 423 Measurements of M 17027 (mm.) Length of fragment 164 Circumference A—P Breadth* Broken end 180 63 59 Tip — 36 36 ( * Side-to-side.) M 17028 A fragment of bone. It is difficult to be certain whether it is of a horn-core or not. The inner, smooth surface may represent a portion of the endocranial occipital region. M. 17029 Oldy. BK IIS. This is the anterior border of the flange with two knobs. There is also a smaller knob (P) opposite the flange. A fairly deep groove is visible near the knobs. Measurements of M 17029 (mm.) Distance between knobs 1-2 95 Total length of fragment 228 Breadth of flange (side-to-side) 51 M. 14955 (housed in the British Museum (Natural History) (plate 3(a), (0)) This is a left horn-core, the knobs pointing antero-laterally and the grooves being on the antero-medial surface. Part of the base is broken away, so that the central (hollowed) part formed by the cranial sinuses is not dis- tinguishable. The posterior portion of the base is rounded, while the anterior flange portion narrows to a ridge. Just above the lower end of the horn-core there is a broad flattened flange, the upper portion of which is broken away in a V-shaped notch just below a knob-like projection on the antero-lateral border. This flange has an S-shaped, curved anterior border which arches forwards, upwards and laterally; then upwards, medially and forwards; and then upwards, medially and backwards. The antero-lateral border of the horn also has two other knobs, and there is a small rough projection just above the highest knob. Just above knob 2 the horn becomes more circular and tends to narrow (A-—P) rapidly. The tip of the horn is broken off; but it appears to have pointed horizontally, medially and backwards. This specimen is noted by Hopwood (1936). There are deep grooves on the antero-medial surface, which sweep up from the base more or less parallel to the anterior and posterior borders. On the postero-medial surface of the horn-core there are less marked grooves present, some of them being along the axis of the horn, while others are transverse. 4.24. ANNALS OF THE SOUTH AFRICAN MUSEUM Length along the posterior border .. ENS 2¢ (4 568 A-P length across the broadest part of the flange Bye nt LGO A-P length at the highest part of the flange (knob 2) oh OE Breadth (side-to-side) at highest eae of paeee ee Bi ay 85 A-P length at knob 3 te i Breadth (side-to- cg} at knob 3 ae oe ie uy, 54. A-P at knob 4 . Aye Ne ia She 88 Breadth (side-to- side) at knob 4 a te his 23 53 A-P at distal extremity of the horn .. oo 46 Breadth (side-to-side) at distal extremity of the horn s 40 TABLE 20. Measurements of M 14955 (mm.). B. FRAGMENTS OF JAWS F 3655 Oldoway II: in situ 1941, with others marked ‘A’ (plate 4) It consists of a large fragment of a right maxilla, a portion of the right palatine bone also being intact. It contains M3, M? and M!. The teeth are in a very worn state; M! in particular has been worn almost to the crown-root junction on its anterior pillar. Buccal surface. ‘The cingulum is rather marked especially in M*. In M?, however, the buccal surface is almost completely broken away. Continuous with the rounded cingulum in M? and Mz? is a very prominent parastyle, a mesostyle and a metastyle, the parastyle being particularly large and the metastyle being rather rounded. ‘The metastyle of M? is almost absent, being heavily impacted against the anterior pillar of M%, and the enamel in that region is also worn away. The prominence of these styles is most marked on the third molar. Lingual surface. Rather coarse rugosity. On each tooth there is a cingulum below which there is a slight bulge. Occlusal surface. ‘The central pits of M? and M? are U-shaped, although the enamel on the buccal aspect of the pit tends to be rather V-shaped. In M? the enamel of the central pit of the posterior pillar is irregular on the lingual side. In M?, where there is more wear, the limbs of the pits which sweep towards the styles are more or less obliterated; and this is even more obvious in M}! where the central pit is completely obliterated in the anterior pillar, while the posterior pillar has a small irregular island of enamel remaining, representing the central pit. On the posterior surface of the posterior pillar of M3, there is a sharp indentation of the enamel which is also reflected on the root. The posterior buccal root of M? is massive and triangular. The buccal roots of the teeth can be seen as the bone is broken away over them, while only the base of the lingual root can be seen, and there are no significant differences in appearances between these and the usual giraffid pattern (vide supra). 6 Old. BK II (plate 7). The body of a right mandible, containing M. g-P,. The vertical axes of M, and M, are tilted forwards. ‘The enamel of all the teeth is fairly rugose (except for P,). MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 425 M, has a rounded cingulum. The rugosity cannot be determined as the enamel is covered by a thin layer of breccia. There is a rudimentary ectostylid and protostylid similar to those of Old. 1 (vide infra). The talonid is slightly angulated to the A—P axis of the pillars in a buccal direction. M,, M, and P, have features similar to those of Mg. P, is similar to P, of specimen 365 (vide infra), except that the paraconid is more separated from the parastylid by a ‘pinching in’ of the enamel. P, Buccal surface. Finely rugose. There is a rounded cingulum. Lingual surface. There is a slight cingulum. The parastylid and the meta- stylid are present. Occlusal surface. The tooth-shape is triangular. The buccal surface is slightly rounded and the apex of the tooth points anteriorly. The posterior surface of the tooth is markedly worn away. Wear. ‘The teeth are in advanced wear. On the anterior surface of Ms, the enamel is thinned by ‘impaction’ of the posterior surface of M,. On the anterior surface of M, the enamel is almost completely worn away; a similar thinning is seen on the contiguous surfaces of M, and P, where the enamel is completely absent. The contiguous surfaces of P, and P, show loss of enamel, as does the posterior surface of P,, but the anterior surface of P, has its enamel only slightly thinned. Maximum wear on the premolar is found on the posterior occlusal surface on the buccal side. Roots. ‘Those of P, are broken away on the buccal side, but there are two roots rounded in shape and they are continuous with each other on the area visible above the alveolus, which is about 10 mm. F 3656 Oldoway 1941, S II (plates 5(b), (d); 6(a)) A portion of a fragmented left mandible, containing M,, M, and a part of M,. M; | Buccal surface. There is a cingulum on the posterior pillar and talonid, but on the anterior pillar there is a slight depression in the cingulum region and there is an unusual ridge of enamel a few millimetres above the crown-root junction. The rugosity is coarse. The tooth is in a rather advanced stage of wear. Lingual surface. Rather rugose. Marked cingulum formation. On the posterior surface of the talonid the ‘abnormal cingulum’ arrangement is also present. Leading up from the cingulum there is a slight entostylid and a slightly more marked parastylid, while the metastylid cannot be detected. The median costa of the entoconid is more obvious than that of the metaconid. The talonid is rounded and its axis is deviated buccally at an angle of about 40° to the A-P axis of the two pillars. Occlusal surface. ‘The two sides of the central pit of the anterior pillar are in close contact, except anteriorly where they are separated by a small triangular 426 ANNALS OF THE SOUTH AFRICAN MUSEUM space. The sides of the central pit of the posterior pillar are slightly more separated than in the anterior pillar, and there is a protrusion of enamel in the direction of the talonid. ‘The anterior pillar is more rounded than the posterior pillar, the peripheral enamel of the posterior pillar being more V-shaped than that of the anterior pillar, which is rather U-shaped. Enamel of the anterior surface, where in contact with the posterior pillar of Mg, is almost completely worn away. M, Buccal surface. ‘There is a cingulum with a rounded bulge above it. Coarse rugosity. Its characteristics are generally similar to those of M;. The posterior surface of M, is worn away where it meets M,, and the anterior surface where it meets M, is very worn, as is the posterior surface of M,, at least half of the surface of which has no enamel whatsoever. M,. A part of the anterior pillar is missing. 1 Old. BK II (plate 6(c), (d), (e)) This specimen consists of a portion of the body of a right mandible con- taining M,—P, in a fairly advanced stage of wear. M; Buccal surface. ‘The enamel is coarsely rugose. There is a rounded cingulum which is particularly marked ‘on the anterior pillar and the talonid. There is a very large ectostylid; and there is a slight bulge above the cingulum of the anterior pillar, but the posterior pillar does not show this and near the occlusal surface of both there is a slight concavity of the enamel surface. Lingual surface. ‘The enamel is coarsely rugose. The cingulum is present, forming a rounded bulge in the region of the base of the entoconid. ‘The features are similar to M, of Old. 6. The lingual surface is cracked and the whole crown has been forced lingually. Occlusal surface. No additional features, except that the enamel is very thick. M,. On the buccal surface the features are similar to those of M;. On the lingual surface the enamel of the posterior pillar is broken away. The appearance otherwise is similar to that of any M, described here. M, Buccal surface. There is a great similarity to M,, except that the bulge above the cingulum is more rounded. A distinct nodular ectostylid is present. The other aspects of the tooth are similar to those of any M, described here. P,. The general description is identical to that of P, of specimen 365. 365 Old. SHK II, 1953 (plate 8) Fragment of a right mandible containing M,, P, and P,;. The teeth are in a very advanced stage of wear. M, Buccal surface. ‘The enamel is rather rugose, and the cingulum is present. On the posterior surface of M, all the enamel has been worn away, the contact MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 427 surface with M, being formed by dentine, which itself has been worn away, forming a notch in this surface. Lingual surface. The enamel has a rather rugose appearance. There is a small cingulum present. No features can be identified because of the marked wear and fragmentation. Occlusal surface. ‘The central pits are completely worn away, except for a small island of enamel in the posterior pillar. ‘The buccal enamel of the posterior pillar is U-shaped, while that of the anterior pillar is V-shaped. Py Buccal surface. Rather rugose enamel. A cingulum is present, being particularly marked on the hypoconid. A small ectostylid is present between the hypoconid and the protoconid. Lingual surface. Rather rugose. Cingulum present. The base of the entostylid can be seen. There is a rounded bulge forming the base of the metaconid. The enamel of the posterior surface is worn thin, while that of the anterior surface is worn away completely. Occlusal surface. ‘The central pit of the anterior pillar is V-shaped, while between the anterior and the posterior pillars there is a small triangular pit. The buccal enamel of the anterior pillar is convex and wide, “boat-shaped’, while the buccal enamel of the posterior pillar is typically in the form of a ‘U’ compressed from front to back. Roots. On the lingual aspect, the bases of the roots can be seen. Thus two triangular roots are visible, the anterior being slightly larger than the posterior and they are separated by an inverted V-shaped interval. Ps Buccal surface. Coarsely rugose. The cingulum is present, being particularly rounded in the region of the hypoconid. Lingual surface. Rather rugose. There is a marked rounded cingulum, being particularly marked at the base of the entostylid and not quite as marked at the base of the parastylid. There is a distinct bulge in the region leading up to the metaconid. Occlusal surface. ‘The whole tooth gives the appearance of an irregular triangle, the buccal surface being more convex than the lingual surface, the apex pointing anteriorly. The posterior surface has its enamel worn away completely, and the dentine has been hollowed out by P,. The posterior ‘pillar’ has an irregularly shaped triangular central pit, while all that remains of the central pit of the anterior ‘pillar’ is a small circular island of enamel surrounding a small pit at the posterior end. Roots. The root of the anterior ‘pillar’ has a broad triangular base. It is larger than the root of the posterior ‘pillar’ which is separated from it by an interval shaped like an inverted V. Mandible. Because it is highly fragmented and almost completely filled with plaster, it is considered advisable not to take any measurements of the mandible. 428 ANNALS OF THE SOUTH AFRICAN MUSEUM 392 Oldoway SHK II E, 1953 (plate 9) Fragment of a right mandible containing M, and the sockets of M, and M,. M,. In an advanced stage of wear. Buccal surface. ‘The enamel is rather rugose. The cingulum is present, and there is a rounded bulge above it. Lingual surface. Rugosity is not clear because the tooth is worn smooth on this surface. The cingulum is present. The entostylid and parastylid are present and are continuous with the cingulum at their bases. A metastylid is not present. Occlusal surface. ‘The lips of the central pit are more widely separated in the posterior pillar than in the anterior. Roots. ‘The shallow sockets indicate that the roots were short. 3 Old. BK II (plate 10) Fragment of a right mandible with M;. The tooth is in a fairly advanced stage of wear. Buccal surface. ‘There is a marked rounded cingulum. Enamel rather rugose. Hypoconulid and ectostylid are present as small nodules. The enamel of the anterior surface of the tooth is broken off, as also the enamel of the lingual surface of the anterior and posterior pillars. The talonid is rather rounded and its axis is in direct line with the longitudinal axis of the anterior and posterior pillars. ‘There is quite a marked V-shaped groove between the enamel surface of the pillars, and also between the posterior pillar and the talonid. Lingual surface. The enamel of the talonid is rather rugose, and if the talonid is looked at from the posterior aspect the enamel can be seen to bulge and drape down on the buccal aspect in the typical ‘apron’ effect. The lingual surface of the talonid slopes rather markedly buccalwards and upwards towards the apex from the bulge above the cingulum. The buccal surface has a con- cavity just above the rounded cingulum. Occlusal surface. The central pit of the anterior pillar is irregularly U- shaped and closed, while the central pit of the posterior pillar is closed off anteriorly by the enamel of the posterior surface of the anterior pillar, and posteriorly the pit is open and continuous with the central pit of the talonid. Mandible. Breadth opposite talonid: c. 42 mm. 93 Old. BK II, 1952 (plates 5(a), (c); 6(d)) Portion of the left mandible, containing M, and M, and a piece of M,. The teeth are in a most advanced stage of wear, and the wear is extremely irregular in that the lingual cones of the pillars of M, and M, have been com- pletely worn down beyond the crown-root junction, while on the buccal aspect a fair amount of enamel is still present. Because this type of wear is the reverse of the usual, it is probable that the cause is a pathological one. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 429 VE Buccal surface. The enamel is rather rugose, although it is worn smooth in parts. There is a bulge above the cingulum on the pillars and on the talonid. Lingual surface. Only the enamel of the talonid and that part where it communicates with the posterior pillar is visible. The talonid is placed at a peculiar angle to the rest of the tooth. Not only is its A—P axis angulated buccally in relation to the mesio-distal axis of the anterior and posterior pillars, but it is also tilted upwards so that its occlusal plane is angulated at 20° to the occlusal plane of the pillars. Occlusal surface. The central pits of the anterior and posterior pillars are still visible. M, Buccal surface. ‘The buccal portion of the posterior pillar and the roots are broken away, and the surface of the anterior pillar has a rather rugose appearance and a small cingulum. Lingual surface. The shape of the tooth has been completely disfigured by the abnormal mode of wear, as in the case of M3. Occlusal surface. There is a small island of enamel projecting above the surface in the region of the parastylid. | M,. A portion of the buccal surface of the posterior pillar remains, this pillar being in extreme wear. The rest of the tooth is fragmented and broken away. 92 Oldoway, 1952, BK II (plate 11) A fragment of a left mandible, containing M, and M, and a piece of the posterior root of M,. The teeth are in early wear and the crown-root junction has not yet appeared above the alveolar surface. M; Buccal surface. ‘This surface is coarsely rugose. The median part of the hypoconid and of the paraconid is rather angulated, especially near the occlusal surface. Lingual surface. A portion of the mandible is broken away uncovering the crown-root junction where a cingulum can be seen. Just above it there is a bulge, and leading up from the cingulum on the anterior pillar there is a marked narrow parastylid, which does not quite reach the occlusal surface of the tooth. About half-way up from the crown-root junction, the base of the -metastylid commences and near the occlusal surface it becomes a prominent ridge overlapping the anterior portion of the entoconid. The median ridge of the metaconid is prominent for about the same distance as the metastylid. The median ridge of the entoconid is prominent, but the entostylid is much less marked. The lower % of the lingual surface of the enamel of the anterior and posterior pillars are fused in the region of the metastylid, but the upper 4 is separated by a groove. This surface of the entostylid is continuous with that of the talonid. 430 ANNALS OF THE SOUTH AFRICAN MUSEUM Occlusal surface. ‘The tooth is in early wear which is most marked along the anterior portion of the protoconid. The central pits are V-shaped, their hollowed portions being continuous with each other in the region between the two pillars, while the central pit of the posterior pillar is continuous with a small central pit of the talonid. The central pit of the anterior pillar is closed off anteriorly where the enamel of the paraconid fuses with that of the proto- conid. M,. The general characters of M, are similar to those of Ms, except that the entostylid and parastylid are more marked than in M,, and the metastylid presents a vertical groove which kinks it posteriorly. g1 Old. BK II, 1952 This specimen consists of the posterior portion of the left body of a mandible including the angle, and contains M,, a part of M,, and the roots of M3. Owing to previous inaccurate reconstruction M, appears to be lower than M,. M, Buccal surface. Coarsely rugose. Cingulum present. The enamel is thickened above the cingulum of the posterior pillar. A small ectostylid is present. The tooth is in a medium stage of wear. Lingual surface. Finely rugose; cingulum present. It has the general characters of M, described previously. A portion of the metaconid is absent. Occlusal surface. Shows two central pits; their lips are fairly widely separated, the anterior pit broadening out anteriorly and the posterior pit broadening out posteriorly. The enamel of the anterior surface of the anterior pillar is worn rather thin due to contact pressure of M,. M,. The general description is similar to that of M,, but it is markedly impacted against M,, so that its entostylid appears rather squashed. Oldoway BK II—East Neguntini site, 1953 (plate 12) Fragment of an unnumbered left mandible containing M,, M, and the anterior pillar of M,. It is in an advanced stage of wear. M; Buccal surface. There is a small rounded cingulum above which the buccal surface is flat. Fairly rugose. Lingual surface. 'The enamel is chipped away. Occlusal surface. The anterior pillar has a rounded appearance. The parastylid is small. M, Buccal surface. Rounded cingulum above which the enamel bulges fairly markedly forming a ridge. Fairly rugose enamel. Lingual surface. It has the general characters of other M, described previously. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 431 - Occlusal surface. ‘The central pits of M, are linear. The posterior pit has an extension up towards the entostylid. M,. Rather similar in general appearance to M,. 321 Old. BK II Ex. 1953 (plate 13) Fragment of a left mandible containing a portion of an anterior pillar of M, and a fragmented M,. It is at the same stage of wear as the Nguntiri specimen. M,. The buccal surface of the anterior pillar is rather flattened out and rounded. It has a finely rugose enamel. M,. It is in rather advanced wear. Buccal surface. ‘There is a sma]l cingulum which has a small rounded bulge above it on the anterior pillar. Finely rugose enamel. The anterior surface of the anterior pillar is broken away. Lingual surface. Fairly rugose. Cingulum present leading up to reach the parastylid and entostylid. The parastylid is prominent. Occlusal surface. It presents an identical appearance to any other molar at a similar stage of wear. The enamel of the posterior surface of the posterior pillar is slightly thinned out due to contact wear by M3. C. IsoLATED TEETH F 2989 Oldoway, 1941, II, in situ with ‘A’ (plates 15(e); 16(e); 17(e)) This is an isolated upper left premolar, probably P® (vide infra), in an extremely advanced stage of wear which is reminiscent of that of M! of F 3655. The lingual surface is completely worn away, while the buccal surface presents a fairly rugose enamel. Just as the wear is more marked on the lingual surface, so it is more marked on the anterior surface. The base of a rather large and rounded metastyle is present. Occlusal surface. Only a small slit remains of the central pit, as well as a small island of enamel near the metastyle. Roots. ‘The lingual root, which is broken off near its tip, is enormous with a rather convex lingual surface and a flattened buccal surface. There are two buccal roots, the anterior one being oval, the posterior one being triangular. The anterior root is broken off near its tip. The measurement of the mesio-distal length of the root, just above the crown-root junction is 27-1 mm. which compares favourably with the length of the root of the Hopefield P* (4025), viz. 29°7 mm. Determination of the diagnosis of F 2989 In order to determine which upper premolar it is, the premolar index (see Section I) has been calculated. Morphologically, specimen F 2989 seems to belong to the same individual maxilla as F 3655. Both were found in 1941, ‘with A’, and they look very 432 ANNALS OF THE SOUTH AFRICAN MUSEUM much the same, i.e. degree of wear, type of fossilization, colour of tooth and breccia. 43 a) The premolar index for the transverse breadth is — — one 47 In G. camelopardalis, the same premolar index for the transverse breadth is Range of N IMD (Sc sce; G Vv Variation P3 aE 108 91:03-+ -469 4°84 5°31 73°6—106-2 pt a 108 95°74 458 4°72 4°93 83°0— 109°5 - According to these figures, the premolar index of F 2989 is more closely related to that of P?/M? than to that of P4/M!, and it should rather be considered asayb?: However, it should be noted that 91-48 is different from the mean value of P4/M! by less than one sigma (95:74—4:72 = 91:02) so that no conclusive significance could be attached to this mathematical approach of the problem of determination, especially in view of the large range of variation. F 2993 Oldoway II, with ‘A’ (plate 14(a), (0d), (c)) Isolated upper right molar, probably M?. It is in a fairly advanced stage of wear, and has a rather rugose enamel, and a cingulum which is very marked on the anterior pillar. There is a small entostyle and hypostyle. Buccal surface. The paracone is missing, but the metacone shows the usual characters except that the cingulum is a very marked ridge, and that the central costa of the metacone is not quite confluent with the cingulum at its base. There is a heaped-up ridge of enamel just above the cingulum. Lingual surface. Below the rolled lower edge of the cingulum, the enamel appears to be thrown into folds. Occlusal surface. It is typical of the upper molars with no unusual features, except that the enamel surface of the hypocone is quite widely separated from that of the protocone where they tend to come together between the two pillars. 4 BK II (plates 15(b); 16; 17) Isolated right M%, with the metacone broken away. It is in a most advanced stage of wear. Buccal surface. There is a marked cingulum. The typical formation of the styles are seen here again, but the ridge of enamel leading to the paracone is not very marked. Lingual surface. The enamel is fairly rugose, but is thrown into ridges forming an entostyle and a small protostyle. There is a cingulum and the tooth bulges below it. Here, as in all other M2’, the posterior pillar projects less in a MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 433 lingual direction than does the anterior (cf. Hopefield 4024, Old. 3655). In all other upper molars the posterior pillar projects more lingually. There is a typical marked angulation of the lingual surface from the base towards the apex in a buccal direction. Occlusal surface. The central pits are most irregular on the lingual side, the buccal enamel lip being V-shaped, while the lingual enamel is irregularly U-shaped. The pits of the two pillars are in continuity in the region of the mesostyle. Roots. ‘Typical formation of the lingual root with its more massive anterior portion formed by a vertical depression between the anterior and the posterior portions of the lingual aspect. The buccal roots are roughly triangular in shape, the posterior one being larger, and its tip curves posteriorly. 109 Olduvai BK II (1955) (plate 14(d), (e), (f)) Isolated upper left M? in early wear, with a portion of the hypocone broken off. The enamel is finely rugose. There is a cingulum present and a very small entostyle, while a ridge of enamel represents both the hypostyle and the protostyle. There is a slight bulge below the cingulum and the lingual surface slopes at a marked angle towards the apex in a buccal direction. Buccal surface. 'The cingulum and its styles are typical of the teeth described previously, except that the styles are not as obvious as, for example, in F 3655. Occlusal surface. The central pits are in continuity with each other. The paracone is rather widely separated from the protocone, but the metacone is closer to the hypocone, and in fact the apex of the metacone tends to be twisted in a lingual direction. Roots. ‘The roots are broken off at the base. F 2992 Olduvai II, in situ with ‘A’ This is an isolated upper molar in a very fragmented state. It is not possible to determine which molar it is and to which side it belongs. It shows characteristics similar to analagous portions of upper molars previously described. It is in a fairly advanced stage of wear. 97 Old. BK II, 1952 Isolated right canine with the mesial portion of its enamel broken away. The tooth is in an advanced stage of wear. Buccal surface. Marked cingulum with a rounded bulge above it. The enamel is coarsely rugose. The tooth is partially bilobed, the outer lobe pro- jecting beyond the true transverse plane of the longest axis of the root. Lingual surface. ‘The dentine is hollowed out laterally. The occlusal edge shows a broad rim of wear. The occlusal edge of the tooth slopes outwards and backwards. Crown height ¢. 32 mm. Maximum breadth 15°O mm. Length 28+ mm. Tooth height c. 60 mm. 434. ANNALS OF THE SOUTH AFRICAN MUSEUM F 2991 Old. IS, 1941 (plates 18(a); 19(a); 20(a)) Isolated right P,, in a fairly advanced stage of wear. Buccal surface. Coarsely rugose; marked cingulum especially at the base of the hypoconid. A nodular ectostylid is visible. Lingual surface. There is a cingulum. Rugosity is coarse, although the tooth is worn smooth over the major part. The parastylid and the metastylid are prominent, the parastylid rising up as a distinct ridge from the cingulum. The median ridge of the metaconid is present. Occlusal surface. The central pit of the anterior pillar is rather wide anteriorly; it narrows in the centre, and then forms an open slit between the metastylid and the entoconid. The posterior pillar has a central pit which has a rounded formation anteriorly and which narrows posteriorly in the region of the entostylid. Roots. ‘They are broken off near the base, the posterior root appearing large and quadrangular-shaped, the smaller anterior root being somewhat oval with a concavity on its posterior aspect. On viewing the posterior aspect, the ‘apron’ effect of the enamel is visible. 2 Old. BK II (plates 15(a); 16; 17) Isolated right P,. It has the general characteristics of F 2991, although it is in a slightly less advanced stage of wear. The roots have been repaired and replaced at an abnormal angle. 5 Old. BK II (plates 15(d); 16; 17) Isolated left P,, with the roots broken away. The tooth has just com- menced wear. Buccal surface. It shows a coarse rugosity and a cingulum which is particu- larly accentuated at the base of the hypoconid. An ectostylid is present on the posterior part of the protoconid. Lingual surface. Cingulum present. The entoconid shows a marked bulge about half-way up the posterior aspect of the tooth. The median ridge of the metaconid is prominent and there is a distinct metastylid which overlaps the lingual surface of the entoconid and is separated from it by a wide interval near the occlusal surface, but is fused with it near the base of the crown. Occlusal surface. ‘Two wide central pits are present. The hypoconid is separated from the protoconid by a wide V-shaped interval in the upper 3 of the tooth, but is fused with it in the lower 4 (as in F 2991 and 4). 7 Old. BK I Isolated right P,. The roots are absent. Part of the base of the anterior pillar is missing, and part of the lingual surface is absent. Buccal surface. Coarsely rugose, and there is a thick cingulum at the base of the posterior pillar. Very similar to 5, except that the hypoconid is not separated from the protoconid by a wide interval, but only by a furrow, and MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 435 that the ectostylid appears here only as a minute nodule. The tooth has just erupted and is not in wear at all. 96 Old. BK II, 1952 Isolated left P,, with a part of the buccal surface, the base and the roots missing. Lingual surface. Very similar to that of F 2991. Occlusal surface. Is also similar to F 2991 except that the central pit of the anterior pillar is now closed off and there is a small island of enamel near the parastylid, and the enamel of the entostylid has fused with that of the metastylid. 105 Old. BK II ex. 1953 (plates 18(c); 19(c); 20(c)) Isolated right M,. Part of the anterior root is broken off and a portion of the buccal enamel of the posterior pillar is chipped off. Buccal surface. A small cingulum is present. The enamel is thickened in the region of the ectostylid. The hypoconulid is formed by a marked vertical ridge, although it is worn away at its base. The enamel is rather rugose. Lingual surface. Small cingulum. It leads up to a parastylid on the anterior pillar, and on the posterior pillar the entostylid seems to have a rather broad base, and it is separated from the median ridge of the entoconid by a slight vertical furrow. The base of the mesostylid can be identified and it is continuous with the metaconid. Occlusal surface. ‘The pillars have a rather circular shape. The central pits are slightly curved with their convexities facing buccalwards, the anterior portion of the anterior pit being larger than the posterior portion, while the posterior part of the posterior pit is triangular and its lips are more widely separated than those of the anterior portion of the pit. The buccal enamel of the central pit of the anterior pillar is continuous with the lingual enamel on the entoconid. There is a slight kinking of the enamel in the region of the metastylid. Roots. ‘The posterior root is broad and curves posteriorly at its tip in an abnormal fashion and on its posterior surface it has a prominent ridge which appears to be continuous with the unusually large hypoconulid. On the posterior aspect of the base of the anterior root there is an aberrant root nodule which has been broken off. 132 Old. BK II, 1953 Isolated right M,, in a very advanced stage of wear. Both roots are broken off at the base. Buccal surface. Rather rugose. Very slight cingulum. The crown is rounded just above the cingulum. A small ectostylid is present. The anterior surface of the tooth has a piece missing. 436 ANNALS OF THE SOUTH AFRICAN MUSEUM Lingual surface. ‘The posterior pillar has almost all its enamel worn down by abnormal wear (cf. specimen 93). The cingulum is barely recognizable on the anterior pillar. ‘The enamel is rather rugose. The base of the central ridge of the metaconid is just recognizable where it fuses with the metastylid. Occlusal surface. ‘The central pits are irregular in shape. ‘The central pit of the anterior pillar presents a bow-tie effect. ‘The posterior surface has its enamel worn away completely. 95 Old. BK II, 1952 (plates 18(b), 19(d); 20(d)) A right M,. The roots are broken off at the base. Buccal surface. There is a fairly well-defined cingulum on the anterior pillar which is less marked on the posterior pillar, and a prominent protostylid leads up from the cingulum. There is a small nodular ectostylid prolonged on to the posterior pillar while there is an elevated ridge (hypostylid) on the posterior aspect of the pillar. There is a small hypoconulid present. The enamel is rather rugose. Lingual surface. ‘There is a smal] cingulum. The parastylid is broken off and the median ridge of the metaconid is prominent and rounded and is hardly separated from the metastylid, so that the general impression of the anterior pillar is that there is a very broad convex portion on the lingual surface (cf. Old. 120, infra). On the posterior pillar, the entostylid is a small narrow ridge and hardly separable from the entoconid. Seen from the anterior or the posterior aspect, the buccal enamel presents the ‘apron’ effect. Occlusal surface. The central pit of the anterior pillar is broad anteriorly, while posteriorly it tends to slope towards the metastylid and in the central portion the two enamel surfaces almost approximate each other. The central pit of the posterior pillar presents a U-shaped appearance of the buccal lip of enamel, and a V-shaped one of the lingual lip. 120 Old. BK II, 1955 (plates 18(e); 19(e); 20(e)) Isolated left M,, with most of the roots missing. It has a rolled appearance and is in an advanced stage of wear. Buccal surface. The pillars have a marked cingulum; on the anterior portion of the posterior pillar, the cingulum is less marked, while it is most marked on the talonid. A short ectostylid is present. The surface is rather rugose. The buccal surface slopes quite markedly towards the apex in a lingual direction. There is no or very slight rounding above the cingulum. A slight cingulum is present on the anterior surface of the anterior pillar. Lingual surface. Cingulum is marked on the pillars, but least marked on the talonid. Leading up from the cingulum of the anterior pillar, there is a rather marked parastylid; the metastylid is absent, and there is only a minute entostylid. However, the central ridge of the metaconid is extremely broad and rounded while the central ridge of the entoconid is also very large, but it is smaller than the metaconid. Between the entoconid and the metaconid the MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 437 enamel folds rather deeply in a buccal direction emphasizing the metaconid and the entoconid even more. The metastylid has fused with the metaconid and thus produces this large bulge, and similarly part of the entostylid has probably joined the entoconid to produce the latter’s large size (cf. specimen 95). The talonid is rather large and rounded; a hypoconulid is present. The buccal enamel of the posterior pillar is continuous with that of the anterior pillar but the two pillars are rather separated, which is a general variable feature of the lower molars. Occlusal surface. ‘The central pit of the anterior pillar is shaped like a bow- tie, in that the anterior portion is triangular in shape, and the central pit is still obvious, as also the posterior portion, but in between the two enamel lips are lying against each other. In the posterior pillar, the central pit has a similar appearance except that the anterior portion is more oval, the posterior portion is narrower and is continuous with the central pit of the talonid, which is fairly large. The enamel is generally very thick in this tooth. Marsabit Road (plate 14(g), (4), (z)) This specimen has no number. It is a left third lower molar. Part of the talonid is broken off, as well as a part of the anterior surface. Buccal surface. The pillars have a marked cingulum and it is coarsely rugose. Lingual surface. Idem. The tooth is in advanced wear. The ridge of the parastylid can just be made out, while the metastylid has fused with the metaconid. Occlusal surface. ‘The anterior central pit is L-shaped. The posterior central pit is broadly U-shaped, and just behind it there is a rather wide central pit on the talonid. Although the roots are broken, they appear to be very short. 166 Old. BK II, 1955 (plates 18(d); 19(b); 20(b)) Isolated right lower molar, probably M,, in early wear. Buccal surface. It has a distinct rounded cingulum, more prominent on the anterior pillar than on the posterior one. The enamel shows a coarse rugosity. The ectostylid is represented by a small ridge of enamel. There is a small parastylid and a protostylid present. No hypoconulid. The metastylid is very prominent. Lingual surface. A well demarcated cingulum leads up to a marked para- stylid and a slightly less marked entostylid which however is broken off near the occlusal surface. The central ridge of the metaconid is well defined, and is separated from the metastylid by a slight depression. The entoconid is broken off. Occlusal surface. ‘The central pit of the anterior pillar is slightly curved and is wider anteriorly than posteriorly. The central pit of the posterior pillar appears to be V-shaped: a part of the pit is broken away. 438 ANNALS OF THE SOUTH AFRICAN MUSEUM 116 Old. BK II Ex. 1953 (plates 15(c); 16(c); 17(c)). Isolated right lower molar, just commencing wear, with the hypoconid missing ‘and a portion of the entoconid broken. The roots are broken. It is probably a M,. Buccal surface. ‘The cingulum cannot be observed because the tooth has been broken away. Surface coarsely rugose. Lingual surface. The parastylid is a marked ridge. The central ridge of the metaconid has a ribbed appearance, and a prominent metastylid is on the upper half of the tooth; the central ridge of the entoconid also has a ribbed appearance. 8 Old. An isolated pillar of a left lower molar in early wear. Buccal surface. There is a rounded cingulum present. Small ectostylid. Rather rugose. Lingual surface. Small cingulum leading up to a small ridged parastylid. The central ridge of the metaconid is only obvious near the occlusal surface. Occlusal surface. ‘The central pit shows an anterior widening and a posterior narrow part, and the central portions of the enamel surfaces are continuous. too Old. I Ex. 1953 This is an isolated entoconid of a lower right molar. Wear is just com- mencing. It presents the general features of entoconids previously described. 98 Old. BK II, 1952 Broken isolated fragment of the entoconid of a lower right molar. 99 Old. BK II, 1952 An isolated fragment of a lingual pillar of an upper molar. Marsabit Road This specimen has no number. It is the posterior pillar of a right lower M, with a smal]] fragment of its root. It is in advanced wear and has the general characteristics of M, specimens previously described, except that its buccal pillar is markedly twisted posteriorly. The central pit is V-shaped and it has an irregular infolding of enamel posteriorly. There is a marked ecto- stylid present. The only measurements that can be taken are: Maximum breadth cc. 32 mm. Occlusal length 24:0 mm. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 439 Oldoway Surface (plates 18(/); 19(f); 20(/)) It is a right lower molar, probably M, in an intermediate stage of wear. Most of the posterior pillar and root are absent, and the anterior portion of the anterior pillar is broken away. Buccal surface. Enamel rugose; cingulum is present and there is a slight bulge above it. Lingual surface. Fairly rugose, but worn smooth. The parastylid and the metastylid are well defined. Occlusal surface. Central pit has a typical bow-tie appearance. Note.— The measurements of all the teeth are given in table 40 at the end of section 2. D. PosTcRANIAL SKELETAL REMAINS 116p Old. BK II, 1952 (plate 23(d)) The letter ‘p’ has been added to this number by the authors so as to differentiate it from the dental fragment with the same number (supra). Proximal end of a right ulna. This presents a massive olecranon process which is separated from the articular facets by a massive rectangular ‘slab’ of bone. The shaft of the ulna is broken off. Olecranon process (posterior extremity) to articular facet along superior border .. ee Ee we r .. 184 mm. Olecranon process, maximum height .. Ae ae Lea Meritise Olecranon process, maximum side-to-side breadth .. fo) 8 7G ER. Breadth at centre of the ‘slab’ .. a se ee ae eG) | ROT: Maximum breadth of articular process. . - i sO mm. Most of its articular surface (i.e. about 4) is for articulation with the humerus; only two small facets below this are for articulation with the radius. The surface area of the radial articulation is relatively less than in the modern giraffe. 115 Old. BR II, 1952 (plate 21(b)) Distal epiphysis of a right radius presenting a marked inferiorly projecting tuberosity. Maximum breadth at radial tuberosity See ee ae, ,b22 smi. Maximum breadth at proximal end of fragment ~: 2) 195 mim. Maximum A-P length .. “i ae = . Jc), 92: Tare. 341 Old. BK II, 1952 (plate 21(e)) Os magnum of the left carpus: very similar to that of modern giraffe. Maximum length A-P .. ats ip oe ee relay igus sisi Maximum breadth 4 < mat ae ae ie a Ge: onan: Maximum thickness of postero-lateral side... ue Si f- SOy aaa: 440 ANNALS OF THE SOUTH AFRICAN MUSEUM 114 Old. BK Il, 1952 (plate 22(a)) Distal end of a metacarpal and a piece of the distal shaft: fragments have been broken off, and the distal end has been chipped and rolled. In com- parison with the distal end of the metatarsal (vide infra, specimen 314) the shaft presents a definite flattened appearance, convex anteriorly and scooped out posteriorly. It is almost identical in appearance to a specimen described by Dietrich (1937), E 122 from Oldoway (his text-figs. 1 and 2, and table VI, fig. 1), and another from Serengeti (Garussi-Korongo 1.39) also described by Dietrich (1942) (his table XXII, fig. 187). The latter specimen is still a young individual with an epiphysis; specimen 114 from Olduvai has a fused epiphysis and. the distal end appears to be broader. Maximum breadth across condyles as .. LOh yuan Maximum breadth of lateral condyle, taken anheniorly .. ) RP aataa Maximum breadth of medial condyle, taken aa «| 2 pana A-—P length of the condyles ue a .. 54+ mm. 60 mm. above the distal end: A—P of the ee oe .. | 42 diag Breadth of the shaft .. MNRAS (0920958011 100 Old. BK II, 1952 (plate 23(c)) Distal extremity of a right femur, consisting of 2 condyles and the patellar condyle. Maximum breadth across the condyles. . ae 3 .. SOT mags Maximum breadth across medial condyle aie ie «| Ge staal Maximum breadth across lateral condyle ve Me .. , 55), mama. Maximum breadth across patellar condyle... ae +s 7G. aaa Cord length of the patellar condyle in the centre... .. ‘TOG y mama tor Old. BK II, 1952 (plate 21(c)) Distal end of a left tibia which articulates with numbers 102, 103. Adult. It presents a marked bulge above the medial malleolus and another large rough tuberosity on the antero-lateral aspect just above the articular surface. Leading down from the shaft to this tuberosity there is a large linear ridge. Distal extremity: Maximum A-—P bye Me up .»») (OR tomaame Maximum breadth .. By Ne .» 20. ial About 80 mm. from distal end: A-P_ .. ps “ . |) ear ca Breadth Ai Wi, A) Gap aaa 112 Old. BK II, 1952 Distal extremity of a right tibia, which is a in many respects to IOI, except that it is smaller. Distal extremity: Maximum A-—P di a ii li CG, Mae Maximum breadth .. oy ni a) TOG) maa MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 441 103 Old. BK II, 1952 (plates 23(a), (b); 24(c)) Left calcaneum, articulating with 102 and 104. It is a massive bone with a markedly prismatic proximal tuberosity (tuber calcis). The body of the bone is broad from front to back, and constricted from side to side. The facet articulating with the posterior surface of the astragalus (talus) is irregular in shape, and has a narrow downward and posterior projection, while it is broad and quadrangular above this. The fibular facet, for articulation with the fibular sesamoid, is quadrangular in shape, convex from front to back and is angulated from behind forwards in a medial direction. At the distal extremity, on the inferior aspect of the lateral side, there is a concave articular facet arched upwards for articulation with the cuboid; it has a lateral convex border, and medially it has a concave border. 108 Old. BK II, 1952 A left calcaneum, shorter than 103, presenting roughly the same features, except (1) that the tuberosity is more rounded and massive, (2) that the body is shorter, and (3) that the fibular facet is smaJler and more angulated medially. 103 108 mm. mm Maximum length of calcaneum . oy 216 198 Maximum breadth (side-to-side) of tuberosity . . we 69 67 Maximum height (A—P) of tuberosity. . te 69 €. 62 Maximum length (A—P) opposite the fibular facet .. 89 gI Body length from the superior border of astragalus facet (along the anterior border) .. AP ai ‘oe 131 CEO Minimum body breadth . F fis 39 42 Fibular facet: A—P length (on the convex portion) . 39 33 Breadth .. ae ue ar ay 28 24 TABLE 21 102 Old. BK IT, 1952 Astragalus (talus), belonging to the left side, articulating with the distal end of the tibia No. 101 proximally, and with No. 104 distally, and with No. 103 posteriorly. The proximo-lateral articular ridge, for articulation with the lateral fossa of the tibia, is large and wide, whereas the medial articular ridge is narrow and has a large articular surface on its medial aspect for the medial malleolus of the tibia. On the lateral aspect of the bone, just behind the mid- point, there is a big oblique quadrangular-shaped surface for the articulation of the calcaneum, and at the anterior end there is a small irregularly rounded facet for articulation with the anterior extremity of the calcaneum. Between these two surfaces there is a rough, hollowed-out region for the interosseous ligament. On the anterior aspect, the fossa for reception of the lower border of the tibia is long and saddle-shaped. 107 Old. BK II, 1952 (plate 21(a)) Right astragalus, with features similar to those previously described in specimen No. 102. Because of the proximity of their discovery, they probably belong to the same individual. Articulates with No. 110. 442 ANNALS OF THE SOUTH AFRICAN MUSEUM 102 107 mm. mm Maximum proximo-distal length .. Bs 113 112 Maximum A—P diameter, medially Me 73 va Maximum A—P diameter, laterally Be 64 63 Maximum breadth, proximally .. 4 87 86 Maximum breadth, distally a ae ¢. 75 76 Maximum articular breadth proximally .. 74 79 Maximum articular breadth distally ay ¢. 75 76 TABLE 22 104 Old. BK II, 1952 (plates 21(d); 24(g)) A left cubonaviculare, articulating proximally with astragalus No. 102 and calcaneum No. 103, distally with cuneiform No.:105 and metatarsal No. 106. Proximally, the medial facet for articulation with the astragalus is much broader than the lateral one, the two being typically separated by a ridge. The tuberosity of the naviculare is short but very broad and massive. Laterally, on the proximal surface, is the bean-shaped facet for articulation with the calcaneum. On the distal surface, the articular facet of the cuboid which articulates with the upper surface of the lateral metapodial is longer and broader than the medial articular facet of the naviculare for the fused cuneiforms. Posteriorly to this facet and continuous with it, there is a small rounded facet for the external cuneiform. Posterior to the cuboid articular facet for the metatarsal, there.is a well-defined groove which runs transversely. Posterior to this groove there is a tuberosity which does not have an articular facet for articulation with the metatarsal. This facet is present in the modern giraffe. 109 Old. BK II, 1952 Isolated right cubonaviculare. The features are very similar to those described in the previous specimen (104). 110 Old. BK II, 1952 Right cubonaviculare, articulating with No. 107, 110 A and 111. It has identical features to No. 104. Probably belongs to the same animal. 104 110 args mm. mm. mm. Maximum breadth (side-to-side) across the centre 110. 108.) Ton Maximum A-—P length across the tuberosity of naviculare 106 ~=6106 95 Maximum length of naviculare articulating facet for cuneiform 55 55 48 Maximum breadth of naviculare articulating facet for cuneiform a7 oF 35 Maximum length of cuboid articulating facet for metatarsal 61 61 55 Maximum breadth of cuboid articulating facet for metatarsal 45 48 42 TABLE 23 105 Old. BK II, 1952 (plate 21(g)) Left cuneiform consisting of the fused I and II cuneiforms. It has an oval shape. It articulates with No. 104 and 106. The proximal surface is concave from front to back. The distal surface is slightly convex in the central portion. Medially and posteriorly there is an irregular prominence, the upper border MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 443 of which is in the same plane as the articular facet. On the surface of this prominence are a number of almost parallel grooves which run from above in a downward. and forward direction: they are probably grooves formed by the astragalo-metatarsal ligament. On the antero-medial part of the superior border, the articular surface has a small lip which projects downwards. It articulates with a similarly shaped small projection on the cubonaviculare, at the junction of the cuboid with naviculare, at the antero-medial part of the opposing facet. r10 A Old. BK II, 1952 A right fused cuneiform, articulating with No. 110 and 111. It is identical in appearance to No. 105, but its posterior portion is broken off. 105 110A Maximum A-P length Me oe ig sq 4 000mm. 59 mm. Maximum breadth side-to-side... ae / o> 4am: 40 mm. Maximum thickness (postero-lateral) a! se) Os Srna 200 mana. 106 Old. BK II, 1952 (plate 24(a), (b), (f)) This is a proximal end of a left metatarsal and a piece of the shaft. It articulates with Nos. 104, 105. The anterior median groove is extremely wide with marked ridges on each side. Proximally are three articular facets—two are kidney-shaped facets for the cubonaviculare, while in between them and slightly medially and posteriorly is the facet for the small rounded external cuneiform. 111 Old. BK II, 1952 Proximal end and portion of the shaft of a right metatarsal, showing similar features to 106. It articulates with Nos. 110, 110 A. 106 yigey | mm. mm. Maximum length (A—P) at proximal end 81 84 Maximum breadth at proximal end .. i pas Hi 93 95 About 60 mm. below: A-—P of shaft * a at 67 65 Breadth is a a oe 65 GP Maximum A-—P of medial articular surface .. i “fi 59 59 Maximum A-P of lateral articular surface .. Mi aye 64. 58 Maximum breadth across centre of medial articular surface 33 36 Maximum breadth across centre of lateral articular surface 40 c. 50 TABLE 24 314 Old. BK II, Ex. 1953 (plate 24(d), (e)) This is the distal third of a right metatarsal. It presents a deep and wide central groove anteriorly, of which the lateral lip is more prominent and higher than the medial. The anterior surface has a general convex appearance, while the posterior surface is flattened. Posteriorly a shallow and ill-defined groove can be seen centrally, leading down to the space between the trochleae (‘Rollen’, Dietrich, 1942). On the outer aspect of each trochlea, there is a 444 ANNALS OF THE SOUTH AFRICAN MUSEUM fossa for the attachment of collateral ligaments, and the lateral one is deeper and larger than the medial. Above the medial fossa there is a rough tuberosity which is larger than the Jateral. The trochleae are big and separated by a deep groove; at the base of the groove, where fusion has occurred, there is an extensive central pit extending upwards. The grooves for the sesamoid bones are shallow, the deepest one being the most lateral. Maximum breadth at the trochleae .. bie He . 2) OG Breadth of the lateral trochlea .. ae it iy el aie Breadth of the medial trochlea .. bs oh se ty 5 A-P length across the trochlea, medially Ps oe 2) OT laterally zs MAMem onan bu. (Oe, Breadth of the distal extremity across the tuberosity .. ie Oe Shaft some 120 mm. above the distal extremity maximum breadth os ie Be iy 2 a maximum A—P as ty He My, Me - 1. 6.00 M 14687 mm. mm. mm. mm. mm. mm. mm. my. This is an articulated hind limb in the British Museum (Nat. Hist.) and Hopwood’s paratype (1934). ‘The femur, proximal end of the tibia and the and and 3rd phalanges are missing. The measurements (mm.) are: Tibia Total length . sik ba bel . | AOOre Length: crest af tibia — distal estiveanite dl, ul 2 r23e Distal extremity: Maximum A-—P ae a 2 a Maximum breadth yi oe +. 4 BOS Metatarsal Maximum length .. aN a se Lo) AE Proximal extremity: Maximum A-P oe cp ape oy Maximum breadth si. a MOS Distal extremity: Maximum A-P .. Me a stay ge Maximum breadth ae ie ahh AGO Phalanx I Maximum length .. si Be 1 Deg '1 0) Proximal extremity: Maximum A-P srk a py sane 5,0) Maximum breadth .. mn 5 Distal extremity: Maximum A-P .. a hee Ne pec}, Maximum breadth ayy fe anne 113 Old. BK II, 1952 185 Old. BK II, Ex, 1953 (plate 22(c)) F 364 Old. S I, 1941 (plate 22(b)) Old. ‘Surface’ (plate 22(d)) F 3297 Old. II, 1941, with ‘A’ MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA A445 F 365 These are all proximal phalanges and it is not possible to say with any degree of certainty whether a particular phalanx is a medial phalanx of the right limb or whether it belongs to the lateral side of the left limb; or whether a lateral phalanx of the right may belong to the medial side of the left. Further- more, it is not possible to state whether an individual phalanx belongs to the fore or hind limb; but we have observed that in one particular extant animal the proximal phalanx of the fore limb is more massive than that of the hind. limb. On this basis it is suggested that specimen F 3297 and ‘Oldoway surface’ probably belong to fore limbs, while the other four specimens probably belong to hind limbs. ay 185 113 HOLA StinfacthimiaZ297. F365: Maximum length .. Ho P07 108 113 120 — 114 Breadth (side-to-side) at base as 50 ¢. 51 44 60 60 52 A-P at base. ae We 54 C. 51 48 - 58 57 51 Minimum breadth, ‘shaft .. ks 46 43 Ba 47 — 39 Distal extremity: Maximum breadth (side-to-side) 48 49 46 58 — 48 Maximum A-P length .. oe 33 34 29 38 — 31 TABLE 25 342 Old. BK II, 1952 (plate 21(/f)) Sesamoid bone articulating with the head of the middle phalanx and the base of the distal phalanx. Maximum A-—P 53 mm. Maximum breadth 36 mm. Maximum thickness 34 mm. CHAPTER 2 ORANGE FREE STATE (UNION OF SOUTH AFRICA) A. LOCALITIES Fossil Sivatheriinae have been recovered from five different localities of the Orange Free State which extend over a large area (the furthermost points being about 150 miles apart), but they all belong to the Vaal River basin (fig. 18). Consequently, in spite of the fact that some of the specimens have been described by different individuals, that they have been found at various stratigraphical levels and that they are housed in different museums, the general geological picture of the Vaal River basin (Cooke, 194.9) provides good reason to consider them in one group. The fossil specimens are recorded as being derived from the following Sites: 1. MMK 3685 (McGregor Memorial Museum, Kimberley) is stated as aw 446 ANNALS OF THE SOUTH AFRICAN MUSEUM gs PRETORIA £7) m{LICHTENBURG we Pi JOHANNESBURG VEREENIGING Cornelia @ KROONSTAD @ Hoopstad Viakkraal and A Florisbaa \? Douglas ; Soe : a ~ 5 Se =o to] BLOEMFONTEIN ai- Waldeck’s Plant ac Gong-gong a3- Barkly-West Fic. 18. Map of Vaal River basin indicating major fossil sites (A) (modified after Cooke, 1949). coming from an ‘unknown locality of the Vaal River basin’ (Haughton, 1922). Assigning this tooth to a new genus and species, Griquatherium cingulatum, Haughton stated that it came from the collection of Mr. A. Grumpelt at Barkly West, and Cooke (1949) stated that on the basis of the other specimens in this collection, it is quite likely that the Griquatherium specimen came from the 60-foot terrace at Waldeck’s Plant or Gong-gong. 2. In 1926, three Sivatheriinae teeth were recovered from the upper layers at Florisbad (Dreyer and Lyle, 1931). The actual teeth were sent in 1932 to scientists in Europe for study, but unfortunately they were not described and it is now not possible to trace them. The only remnant of the specimens is a poor plaster-cast of the crown of a lower molar (C 1492) which is housed in the Nasionale Museum, Bloemfontein. Notes on the geology of Florisbad have been published on various occasions (Dreyer and Lyle, 1931; Dreyer, 1938; Hoffman, 1953; Oakley, 1954a; Singer, 1956, and Meiring, 1956). Oakley (1954a, page 84) states: ‘The Florisbad deposits consist of sands, intermittently ejected by springs of gaseous water during Pleistocene and recent times, alternating with seams of peat formed by salt marsh vegetation which spread across the area when the springs were quiescent. There are two parallel lines of spring centres (or MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 447 eyes’) which have become sealed off progressively in an easterly direction. Thus the “‘eyes’’ are fossilized on the western side of the site but still active on the eastern side. ‘Fossil mammalian bones and teeth and Stone Age implements occur in the beds of sand formed by the spring waters during Pleistocene times. Owing to the occasionally disruptive action of the springs, when old “eyes”? are reopened, it is not always possible to be sure of the original stratigraphical position of specimens found in these deposits. The most reliable finds are those from below uninterrupted seams of peat.’ Despite deficient data (Dreyer and Lyle, 1931, p. 5, and the Nasionale Museum Register), Oakley (1954a) seems to have obtained information some- where that a tooth of the extinct Sivatherine (Orangiathertum) was found between Peat II and Peat III. 3. A horn-core fragment, the type specimen of Orangiatherium vanrhyni v. Hoepen, was merely mentioned by van Hoepen (1932). He also mentioned ‘a terminal fragment of a large antler and a series of large teeth, which were probably associated’. These specimens, presently housed in the Nasionale Museum and registered under the numbers C 431A, C 431B, and C 426 respectively, are derived from the farm Tierfontein, on the Vet River, 9 miles from Port Allan (personal communication from the Director of the Museum, Dr. A. C. Hoffman) (fig. 18). 4. Specimen F 39 (Archaeological Survey of the Union of South Africa, Johannesburg) is stated by Cooke (1949) to be derived from ‘an unknown locality in the Vaal River basin’. This is the type specimen of Griguatherium haughtont Cooke. 5. Two Sivatherine upper milk molars were identified in the collection of specimens from Cornelia, in the Nasionale Museum, Bloemfontein, by one of us (R.S.): these specimens are not registered and are allocated B! and B? by the authors. The Cornelia site consists of a large erosional area adjacent to a small branch of the Vaal River. Specimens are found on the floors and on the sides of “dongas’ (eroded clefts) which are washed by seasonal rains and partly covered by flooding of the nearby river, and consequently there are numerous redepositions. The surface consists of hardened calcited sand which rests on varying projections of Karoo formations. A typical view of the site is shown in plate 5 of Oakley’s above-mentioned paper (1954a). B. List oF MATERIAL AND DESCRIPTION MMK 3685 : Isolated left M*— Vaal River (? Waldeck’s Plant). 1492 : Cast of occlusal portion of left M,—Florisbad. Cc 426 : 4 isolated upper molars (right M? and M®, left M, and M?)—Tierfontein. C 431 : Left posterior horn-core and a right (?) anterior horn-core (ibid.). F 39 : Stated to be an isolated anterior pillar of a lower left M, or M,;— Vaal River (unknown locality). B! and B? : Isolated right DM® and DM*— Cornelia. 448 ANNALS OF THE SOUTH AFRICAN MUSEUM I. HORN-CORES C 431 There are two specimens marked C 431 from Tierfontein. The one is an almost complete posterior horn-core with the tip missing, and this is here designated C 431 A. The other, C 431 B, is a fragment of an anterior horn, and not a part of A. C 431 A (plate 29(¢), (d)) This is a left posterior horn-core (vide infra, ‘Discussion’), the base of which is pear-shaped, the narrower portion being anterior. The apex of the hollowed-out portion of the base, formed by sinuses, is 27 cm. distant from the broken edge of the base. From the base, the anterior border runs out- — wards in a gentle arc towards the first knob, then upwards and slightly medially and then laterally, giving the broad surface of the horn a double twist. The horn is flattened from side to side opposite the first and second knobs, but above these it tends to bulge where the anterior border becomes rounder. Grooves on antero-medial surface (cf. Old. 3.53) From the base there are three fairly deep and broad grooves starting near the front at almost a single point and running obliquely up and back at an angle of about 45°. The most medial groove ends in a trifurcation at the posterior border opposite the first knob. Along the antero-lateral border, at the base of the broken-off flange, three deep grooves pass vertically up parallel to the border. Just below the first knob they tend to deviate from each other, the anterior one running along the anterior border and, passing behind (lateral to) the second knob, it divides into a number of smaller grooves running up almost parallel to each other towards the tip. Between the levels of the first and second knobs there are four other deep main grooves on the convex surface which run up this surface fairly parallel to each other. The posterior one divides into three just below the level of the second knob, and these then pass towards the posterior border and run laterally. The anterior border in the region of the flange has two irregular rough tuberosities with numerous vascular foramina. There are, in all, three knobs, very crinkly in appearance and fairly evenly spaced from each other. Between them are two small raised irregular tuberosities. Circumference at base a ie (a ett Ne! 360 Circumference between flange and knob 2 at e..8Alh Circumference 100 mm. above knob2 .. 258 Circumference at knob 3 .. a Se ey 255 Circumference at tip ay alls ave Total length hh Ne a af 5 570+ MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 449 A-P Side-to-side Base .. cp Le Ria galt Ge oF Between flange and knob 2 es oe ee 137 79 Above knob 2 (100 mm.) .. ap uy a, 93 7p At knob 3°... fA ae se eile ae 97 65 At tip a; ne i Re ai ae — — TABLE 26. Measurements of C 431 A (mm.). C 431 B (plate 20(a), (d)) This is an anterior horn, probably right (if the grooves are taken to be on the medial side). The base is flat from side to side and rather triangular with the broad end posterior. Hollowing-out the base are three cranial sinuses. The medial surface is extremely irregular with numerous ridges formed by deep grooves. The grooves are deeper anteriorly and commence at the base of the anterior border and pass in an inverted triangular fashion, the anterior grooves being vertical and the posterior oblique. The anterior groove branches about half-way up. The posterior part of the medial surface is roughened by small, shallow grooves. ‘The anterior border has a knob just above the base— it has a cauliflower appearance and is distorted by (? vascular) grooves. ‘The top of the anterior border has another similar knob. The superior surface is very irregular and grooved, the central portion being smooth. The posterior border is concave, the upper end passing backwards and there is one small protuberance at the base and one at the upper end. The outer surface is rather smooth with a single deep groove near and parallel to the anterior border. Circumference at base ae ae Hs 280 Circumference at tip ae oe: Be 320 Total length: Anterior .. ae tr 62220 Posterior .. Fe Vo neGn LO A-P Stde-to-side Base... Me a oe Be WNT c. 70 Middle (opposite knob) ee Tol 55 ALT OMNP oe Be ae 123 61°5 TABLE 27. Measurements of C 431 B (mm.). 2. TEETH MMK 3685 (plate 25) An isolated upper molar in a medium stage of wear, with the roots broken off near the base. This is the type specimen of Griquatherium cingulatum Haugh- ton. Although it was originally described as a second molar (Haughton, 1922), and stated by Cooke (1949) to be either a M? or a M3, it is here considered to be unquestionably a third molar because of the relative decrease of the lingual- ward projection of the posterior lingual pillar compared to the anterior one (vide infra). Buccal surface. Each pillar has a marked cingulum, that of the anterior pillar leading to the broad base of the rounded marked parastyle. The central median costa of the paracone is fairly distinct, the bulge commencing about half-way to the occlusal surface. 450 ANNALS OF THE SOUTH AFRICAN MUSEUM The mesostyle is distinctly prominent, projecting out markedly in an anterior direction from the surface of the pillar and having a broad base continuous with the cingulum of the posterior pillar on the one side and with the cingulum of the anterior pillar on the other side, though in the latter there is a slight vertical groove partly separating them. The vertical median costa of the metacone is narrow and hardly prominent, commencing near the base just above the rolled edge of the cingulum. Posteriorly, the cingulum of the posterior pillar is continuous with the bulging base of the short metastyle. Lingual surface. ‘The enamel is rather rugose. On the anterior pillar there is a marked rounded cingulum, the lower border (the one towards the occlusal surface) of which increases in thickness on the anterior surface producing an unusually long, rolled and ridged protostyle. The cingulum of the posterior pillar is relatively small and aimost absent in the region of the entostyle. On the posterior surface of the posterior pillar the actual cingulum remains small and ridged but a distinct elongated crest forms an unusual hypostyle. The lingual surface slopes towards the occlusal surface at an acute angle from the cingulum, but near the occlusal surface the angulation changes and the lingual surface tends to become slightly more vertical. The slope of the buccal surface on the other hand is almost vertical. Occlusal surface. The central pits have widely separated enamel surfaces, the pit of the anterior pillar being acutely V-shaped and closed anteriorly by the anterior limb of the protocone. The central pit of the posterior pillar has a more obtuse V-shape, being closed posteriorly in the region of the metastyle, but open anteriorly and continuous with the space between the enamel surface of the contiguous side of the protocone and the hypocone. The occlusal surface of the posterior pillar is abnormally longer than that of the anterior, the ratio being 114.4 (table 28). Stvathertinae Measurements (mm.) and indices M? M? MME (Mean) (Mean) 3685 Occlusal length posterior pillar ee eee Bao 1144 Occlusal length anterior pillar ry Maximum breadth posterior pillar ele 99°1 gt git Maximum breadth anterior pillar Maximum tooth length if a ae 47:6 515 53°8 Maximum tooth breadth .. is re 49°7 48°7 54°6 TABLE 28 From the table, the ratio of the breadth of the posterior pillar relative to that of the anterior pillar corresponds closer to that of four other Sivatherine M? than to that of two Sivatherine M?. From observations on the extant material this ratio in M$ is a very constant one, and consequently, despite the fact that the ratio for the relative occlusal length appear nearer to that of M?, the authors consider that MMK 3685 is a M3. The length and the breadth of the tooth are at the outer limits of the range of Sivatherines (tables 26, 40). MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 451 C 1492 (Cast) This is a plaster reproduction of the occlusal portion of the crown of a left M,. Observations on the cingulum are impossible and what can be made out from the enamel of the cast, it appears to be fairly rugose. The tooth does not appear to be in an advanced stage of wear. Lingual surface. There is a prominent metastylid and the buccal enamel of the metaconid is continuous with the lingual enamel of the entoconid. The parastylid is slightly prominent. The central costa of the metaconid appears rather flattened. Occlusal surface. ‘The anterior pillar is more rounded than the posterior pillar on the buccal aspect; the central pit of the anterior pillar is ill-defined, but that of the posterior pillar shows a fairly wide central portion, narrow anteriorly and broad posteriorly. C 426 This number is given to four upper molars. The authors have sub- divided them into A, B, C and D which will be entered in the register of the Nasionale Museum, Bloemfontein. A and B belong to the same individual because of the obvious contact surfaces. C 426 A (plates 26(a); 27(a); 28(a)) This is a right M? in an extreme degree of wear. Part of the paracone is missing. Buccal surface. A cingulum can be seen, especially in the region of the base of the metastyle where it is considerably heaped up. There is no cingulum in the region of the base of the mesostyle and the surface between meso- and meta- style has been hollowed out, in the same ‘W’ formation as other teeth described previously: there is a slight bulge in the region which leads up to the apex of the metacone. The paracone is almost completely absent. Lingual surface. Finely rugose. On the hypocone the rugosity is extremely fine with additional transverse striations. On the anterior aspect of the proto- cone the enamel is raised slightly at one spot, but the rest of this surface is extremely smooth due to contact pressure, and this wear has even extended on to the base of the root in this region. On the hypocone there is a very slight cingulum. On the protocone, the cingulum is slightly more marked, but still negligible. On the protocone too, the lingual surface bulges slightly above the crown-root junction, and it can be seen to slope towards the apex in a buccal direction. The posterior surface of the hypocone is worn right down to the crown-root junction, only a small piece of enamel being visible here. Here also, the wear has extended on to the root. Occlusal surface. Despite the marked wear, the central pits are obvious, the enamel edges of each pit of the anterior and posterior pillars being separated to a fair degree. In the anterior pillar, the pit has an L-shape, the upright of the ‘L’” extending right up to the parastyle, which is broken off. However a small 452 ANNALS OF THE SOUTH AFRICAN MUSEUM island of enamel belonging to this pit can be seen extending towards the meso- style, but it is separated from the L-shaped portion by dentine which is hollowed out by wear. The pit of the posterior pillar is irregular in shape, the lingual aspect having its enamel thrown into two folds which project into the centre of the pit, and the extremities of the pit extend towards metastyle and mesostyle. The dentine is particularly hollowed out between the enamel of the lingual surface and that of the central pit. On each occlusal surface can be seen striations which in some places are almost distinct scratches indicating a side- to-side chewing movement of the jaws. The shape of the enamel of the lingual surface viewed from the occlusal side is arc-shaped for both cones, the arc of the hypocone being more flattened than that of the protocone. Roots. Despite the marked fragmentation, it is possible to identify three roots. The lingual root being composed of two massive pillars joined by a plate, and the anterior pillar being the larger. The posterior buccal] root is triangular in shape, while the anterior is broken off at its base and fragmented so that its shape cannot be identified. C 426 B (plates 26(b); 27(b); 28(5)) This is a right M? which is very fragmented, so that only a portion of the enamel of the hypocone on the lingual surface is present, while that of the protocone is absent. Both the metacone and the paracone are missing. Lingual surface. Rugosity is fairly marked in parts, and there is a small cingulum which tends to have a rounded bulge below it. The anterior surface of the protocone shows some ridging of the enamel just below the small cingu- lum (protostyle) while the rest of this surface, which is only represented by a small fragment, is very smooth due to contact wear. A small piece of the central pit of the anterior pillar is present, the enamel of the two sides of the pit being separated to a fair degree. The roots are represented only by a portion of the lingual root. C 426 C (plates 26(d); 28(c)) A left M? which is a badly fragmented tooth with only protocone and hypocone and a portion of the lingual root present. Lingual surface. Rather rugose, with a poorly defined cingulum on the hypocone and a well-defined cingulum on the protocone. On the lingual surface of the protocone, near the occlusal surface, is a ridge of enamel parallel to the cingulum. There is no bulge above either cingulum. The lingua] surface of the protocone slopes only slightly towards the apex, while that of the hypo- cone seems to have a slightly more marked slope. The shape of the lingual enamel of the occlusal surface is arc-shaped for both cones. On the anterior surface of the protocone the enamel shows a slight horizontal ridge (proto- style) but most of the surface is smooth by contact wear. The enamel of the hypocone ends abruptly at the posterior surface and this surface is hollowed MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 453 out and appears to be worn smooth by contact with the adjacent tooth. This appearance is not uncommon in giraffids. Occlusal surface. ‘The hollowed-out pits are irregular in shape, the enamel of the lingual and buccal lips of the pit being fairly separated. Because of the absence of parastyle, mesostyle and metastyle, it cannot be determined whether there are any separate islands of enamel in those regions which may have been linked to the central pit at an earlier stage of wear. Roots. Most of the lingual root is present and it has a similar appearance to that previously described for 426 A and B. C 426 D (plates 26(c); 28(d)) A left M! which is very incomplete, only a portion of the protocone remaining below a very fragmented lingual root. There is a small cingulum and a very slight thickening of the enamel just below the region of the cingulum where the tooth tends to bulge. The general shape and appearance of the protocone is identical to that of C 426 C. F 39 (plate 27(c), (d), (¢)) This is the type specimen of Griquatherium haughtont Cooke 1949. It has been described as an isolated anterior pillar of a lower left molar, either M, or M, (vide infra). It is in an early stage of wear. Although the evidence for the viewpoint that this is half a lower molar is reasonable, peculiarly enough the specimen also presents a number of features which raise considerable doubt of the accuracy of this diagnosis and which could support the proposition that this specimen is an upper premolar. For this purpose it is necessary to compare it directly with Hopefield 4025 and Olduvai F 2989 (plates 45, 15-17). This matter is discussed in detail below, but the authors consider the evidence to be in favour of a left upper premolar and will describe the specimen on that basis as follows. Buccal surface. ‘The tooth is broken probably just below the crown-root junction, and a portion of the posterior aspect is broken away. The rugosity is fairly coarse. There is a marked rounded parastyle which is separated from the prominent median costa by a broad groove, and the costa in turn is separated from the prominent metastyle by a narrower groove. Near the occlusal surface, the metastyle swings posteriorly in an arc. Lingual surface. It is coarsely rugose. ‘The tooth is fractured at the crown- root junction and a small rolled cingulum can be seen at the base of the posterior lingual aspect. This appears to be sufficiently localized to warrant being called an entostyle. From the anterior aspect, it is obvious that the enamel of this surface presents a marked ‘apron’ effect. The lingual surface slopes markedly downwards in a buccal direction to a point approximately at the junction between middle and lower 4 of the tooth. Then the surface tends to slope more vertically to the occlusal surface, so that the general effect is that the upper 2 of the tooth has a marked lingual bulge. This lingual bulge is the most promi- 454 ANNALS OF THE SOUTH AFRICAN MUSEUM nent observed in the whole series of African Sivatheriinae. This matter will be more fully dealt with in the discussion, but it is here noted that an X-ray of the tooth (plate 27(c)) indicates internal vertical fracture lines, which may have been caused by intrusive compressing breccia. This may be an explanation for some of the excessive bulging observed. Occlusal surface. ‘The central pit is fairly wide and open posteriorly, while anteriorly the two enamel surfaces are in continuity with each other. Posterior surface. A large piece of the enamel near the buccal aspect is broken away, especially from the upper 2 of the tooth. But sufficient of the enamel is present near the occlusal surface to ‘close’ the lingual and buccal enamel crescents of the protocone. ‘There is also sufficient enamel on the lingual surface to indicate that it is extending without interruption well towards the buccal aspect: this is further back than would be expected if the enamel were to be continued on to the contiguous surface of another pillar (on the alternative supposition that this may have been a lower molar). The enamel is heaped up slightly in the region of the protostyle. Morphologically it may however be compared favourably with the anterior pillar of M, and M, of Olduvai 92 (BK II, 1952). On the other hand, if the probability of the dimensions of F 39 is calculated by the ¢ test in respect of its being a M, or Msg, the following is obtained: On the basis Value of P of P39— iM, F39—2, Occlusal length of anterior pillar .. 05 “I Maximum breadth of anterior pillar “4 =) Occlusal breadth of anterior pillar. . zo) 6 Buccal height of anterior pillar... Ol Ol Because of the large dimensions of this ‘pillar’ (on the alternative supposi- tion that it is a lower molar), the available measurements have been utilized to reconstruct the whole tooth by a comparison with M, or M, of other known Sivatheriinae. Calculated on this basis, the Tr./A—P index of F 39 would be: M, M; F 39 de ee ih 641 46-1 Mean of Sivatheriinae .. 70:0 49°7 . Occlusal length of anterior pillar and the ratio ——————s—— —____—_—_—_ i VOOn: Maximum breadth of tooth F 39 oe site ake aE 812 Mean of 12 Sivatheriinae M, .. 74°5 Mean of 10 Sivatheriinae M, .. 752 Consequently, it is reasonable to state that one cannot assign F 39 as a definite M, or M;: it is furthermore clear that from the point of view of the length and breadth of the tooth (table 29), it is unlikely to be an anterior pillar of a lower molar. Its height is greater than that of any tooth in the available series (see also ‘Discussion’). 5 MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 455 Anterior pillar: Maximum tooth Anterior pillar: Maximum tooth Stvatheriinae occlusal length length maximum breadth breadth M, (mean of 12 specimens) ne. 26-0 512 34°9 35°7 M, (mean of 10 specimens) Bi 25°0 68-6 33°3 33°6 F 39 Actual measurements Ae 32°5 c. 40 Inferred measurements: M, si ue 64:0 41-0 If M, ma *, 87°5 40°3 TABLE 29 B! and B? These two teeth are both from one site, probably Cornelia, but there are no records available concerning their discovery. The one tooth B! has three distinct roots, one lingual and two buccal, and is therefore an upper molar. In B? the roots are broken away and the tooth is only just commencing wear on proto- and paracone. It is also obvious that both teeth belong to the same jaw and the same side because of the exact fit of their contiguous surfaces. Because of, first, the small size of the root in B!, secondly, of the hollowness of the root and the tooth, thirdly, of the thinness of the cement and enamel, and fourthly, of the general appearance of B! and B?, it is considered that these teeth are deciduous. Consequently, B! is diagnosed a right DM? and B? is a right DM+. B* (plate 30(2), (¢); (¢)) Right DM. It is in early wear Buccal surface. Small cingulum present. Rugosity very fine. Metastyle, mesostyle and parastyle have a very obvious rib-like effect, the parastyle being the largest of the three styles and being decisively rounded and separated from the lingual surface of the paracone by a distinct cleft. The ridge leading up to the apex of the paracone is more obvious than that leading up to the apex of the metacone, which is almost absent. The appearance is typical of the Family, the surface having an undulating effect, especially when seen from the lingual or the buccal aspect, the crests being formed by the apices of the paracone and protocone, and of the hypocone and the metacone, while the depressions are opposite parastyle, mesostyle and metastyle. Lingual surface. ‘The protocone has an obvious cingulum, while that of the hypocone is less obvious. On the antero-lingual aspect the enamel is heaped up and forms a ridge (protostyle). Similarly on the postero-lingual aspect of the hypocone, there is another ridge of enamel (hypostyle), but this is very small and less extensive than the protostyle. The surface is finely rugose. Just below each cingulum, there is a bulge which is less marked on the hypocone than on the protocone, and from this bulge the lingual surface slopes fairly acutely towards the apex. From the anterior and posterior aspects, the crown- root junction has an irregular line which is ‘apron-like’, as in the Hopefield upper teeth. Occlusal surface. ‘The protocone is quite separate from the hypocone but the central pits of the two pillars are continuous opposite the mesostyle, and it 456 ANNALS OF THE SOUTH AFRICAN MUSEUM can be seen that the separation of the pits would have been caused at a later stage of wear by the hypocone going to meet the paracone to form the meta- conule. The lingual and buccal lips of the pits are widely separated, maximally at the centre of each pillar. The meeting of the protocone and the paracone takes place at the parastyle where their dentine and enamel surfaces are continuous. But the meeting of the hypocone and metacone is only by contact of their enamel surfaces. Roots. The lingual root is rather small and oval-shaped, the separation into the two thickened portions of the root being hardly noticeable. The cement is very thin, the root and the tooth being completely hollow. The two buccal roots are thin and hollowed although the cement is thicker than in the lingual root; the posterior one is triangular and the anterior one is oval in shape. B® (plate 30(b), (4), (f)) Right DM?. Buccal surface. ‘Vhere is a slight cingulum present on the hypocone and it leads up to a very marked mesostyle and a slightly Jess marked metastyle. The parastyle is prominent and rounded, and it has a small nodular excrescence on its buccal aspect. ‘The mesostyle is somewhat split by a vertical furrow. The buccal aspect of the paracone is much more marked than that of the metacone ‘The buccal surface of the metacone is distinctly spatulate-shaped as is commonly found in milk dentitions. Lingual surface. Finely rugose. The hypocone has a slight bulging cingu- lum from which the lingual surface slopes rather sharply to the apex. Although the protocone is rather fragmented at its base, there seems to have been no cingulum, and the lingual surface also slopes rather sharply towards the apex. On the anterior surface, the enamel is thrown into a small ridge (protostyle) just above the crown-root junction. Occlusal surface. ‘The hypocone and metacone are just beginning to wear, whereas the protocone and paracone are worn to a more marked degree. All four cones are separated, their fusion only occurring between the hypocone and metacone near the metastyle, and between the protocone and paracone near the parastyle, while paracone and metacone are almost completely fused on the buccal surface medial to the mesostyle. Consequently, the central pits are ‘wide open’ although the two cones of each pillar have fused at the base of each pit at the base of the tooth. Roots. No roots present. They have been broken off. The tooth is hollowed out. Shape. The protocone is V-shaped and the hypocone is slightly more arched and broader. The paracone is almost a straight line while the metacone is spatulate-shaped. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 457 CHAPTER 3 MAKAPANSGAT (NORTHERN TRANSVAAL, UNION OF SOUTH AFRICA) The various discoveries in and the geology of the Limeworks Cave in the Makapan Valley have been fully described by Dart (1954), Oakley (1954b), Brain, van Riet Lowe and Dart (1955), Howell (1955), Brain (1957, 1958), and Wells and Cooke (1957) in whose publications further references may be obtained. All the material described here is on loan from the Bernard Price Institute for Palaeontological Research (Johannesburg), and is derived from the Lime- works dumps wherein out of 1,862 skull remains already recovered some 30 belong to Giraffids (Dart, 1957). However one specimen, M 553 B}, is from an unknown locality in the Makapan Valley, ‘and probably from one of the limeworks’ (Cooke and Wells, 1947). Some of the material is still heavily filled with calcified grey cave breccia. The material received for study may be divided into two major groups: the first group consists of giraffid teeth referable to Giraffa, the determination of the species of which will be mentioned later. The second group is composed of fragments of the dentition and of the jaws of Sivatheriinae. Both groups were recovered from the same deposits, and from the nature of the breccia it is probable that they were contemporaneous in so far as the limits of the formation of the breccia are widely separated by a large period of time. List oF MATERIAL A. GIRAFFA (1) Upper dentition (a) Milk M 646 ae Pa) aright Py Nie M 944 es. o2, «left DM? M 536 we os leit: DM® M 263 - .. fragment of right maxilla with DM?— M1? M 533 a3 oe IOI M 535 =: ~ fete DM? MELE TS 9 5. a= left. DM? (b) Adult M 532 se seit lett Re M 531 De a! lett Re M 264 ae “2 tebe PS Merri. (o.. 20 gishit Pe M 552 i gee 5 Eiht Ve M551 at 6. dere Vit M 550 Se .. fragment of maxilla with left M2? M 528 Bie .. fragment of maxilla with left M?— M3 (2) Lower dentition (a) Milk M 939 ey, .. fragment of left mandible, with DM, and DM, M 540 a .. fragment of right mandible, with DM,— DM, (6) Adult M 936 right M, M 942 and Mit 13 joined fragments of right M, M 938 3 .. fragment of right mandible with M, 458 ANNALS OF THE SOUTH AFRICAN MUSEUM B. SIVATHERIINAE (1) Upper dentition Milk M 937 eh .. isolated left DM? or DM?* M 524 bis .. right, probably DM? M 941 i .. right, DM? or DM? (2) Lower dentition (a) Milk M 525 ae ie ete va. M539B .. .. fragment of right mandible with DM, (b) Adult M 527 of ond) LMICISOF, Miri. se in) MCISOL M553A .. .. fragment of left mandible with P, M553B OL. .. fragment of right mandible with P,— P, M553 BI .. .. fragment of left mandible with M,—M, M 943 om isolated pillar of right M, (See also Appendix. ) DESCRIPTION A. GIRAFFA The general appearance of these teeth is the same as that of the extant Giraffa camelopardalis and nothing can be gained by describing each specimen in detail. The only variations that may be noted are minor individual ones, namely, those of irregularity of enamel ridges present in some and not in others, the presence of entostyle and parastyle, of ectostylid and parastylid and other variations which are also found in the non-fossilized extant material. OM2 Makapansgat es COPE C immature) pM3 om4 DMo DM3 DM, 5 10 15 20 25 30 35 mm Fic. 19. Dimensions of deciduous teeth of Makapansgat Giraffa plotted against ranges of variation in modern Giraffa camelopardalis. —— = A-P length. ------ = Transverse (breadth). . = Makapansgat specimens. (1) UPPER DENTITION (a) Milk teeth M 646 and M 944 (plate 31(a), (7), (9); 31 (6), (4), (7) These are deciduous upper second molars, 944 being left and 646 being MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 459 right. They are in the same stage of early wear and are identical to each other. It is considered that they probably belong to the jaw of the same individual. M 536 (plate 31(¢), (4), (s)) An isolated left upper deciduous third molar, the posterior buccal root of which is broken off at the base, while the other roots are broken about half- way down. The tooth is in fairly advanced wear. M 263 (plates 32(g); 33(d); 34(d)) This is a fragment of the right upper jaw containing DM’, DM? and M1. The teeth are in early stage of wear, DM? being in a more advanced stage than M?! whose posterior part of the posterior pillar is just commencing wear. In spite of their similar appearance and their similar degree of wear, it is unlikely that 944 and 536 (left DM? and DM respectively) on the one hand, and 646 and 263 (right DM? and DM? respectively) on the other hand belong to a single individual. The Dental Index has been calculated (see Section I, chapter 5) for both pairs of teeth, and compared with the corresponding range of variation for the extant G. camelopardalis. As seen in the following table, all three ratios (length, breadth and Tr./A—P index ratio) fall outside the respective ranges for the extant material: Dental index Tr.|A-—P Length index Breadth index index 944/536 .. Se Me es - 83°3 5 hig | 85°7 646/263... AS ea ae . gg'0 74:8 75°7 Range of DM?/DM3 in G. camelopardalis 62-7-81°9 75°5-92°3 94°0-127°4 TABLE 30 M 533 (plate 31(@), (/), (¢)) It is an isolated right upper deciduous fourth molar, with the roots broken off at the base. The anterior pillar has just commenced wear. M 535 (plate 31(e), (m), (w)) It is an isolated left upper deciduous fourth molar, with the anterior buccal root broken off at the base, while the two other roots are broken about half-way down. M 1115 (plate 31(f), (n), (2) It is an isolated left upper deciduous fourth molar, whose roots are broken off near the base. It is in an early stage of wear. (b) Adult teeth M 532 (plates 33( f); 35(@), (¢)) This is a left upper premolar, P?, in early wear. Buccal surface. It has the typical appearance of a paracone of an upper premolar. Although there is a vertical furrow on the buccal surface between 460 _ ANNALS OF THE SOUTH AFRICAN MUSEUM paracone and metacone, there is no separation on their lingual surface. Between the apex of the paracone and the parastyle, there is a smal] enamel tubercle. The central pit is irregularly V-shaped being open anteriorly, but closed by a ridge of enamel posteriorly between the protocone and the metacone, while the enamel] in the region of the protocone invaginates into the central pit as a double fold. Lingual surface. ‘The surface is fairly rugose. There is a ridge of enamel opposite the centre of the protocone, while this ridge increases posteriorly to form a broad entostyle. M 531 (plates 33(2); 35(d), (#)) A left P?. This is a very worn tooth. Be Makapansgat maa an oy ar Rey Cadults) p3 m1 aa Me “ . Me baPiS 2) ele S ine UNOS eee 8 ae Mo e ° M3 PEPPR il emera ye Sten se 15 20 25 30 She, 40 45 mm Fic. 20. Dimensions of adult teeth of Makapansgat Giraffa plotted against ranges of variation in modern Giraffa camelopardalis. —— = A-P length. ------ = Transverse,(breadth). . = Makapansgat specimens. Lingual surface. A cingulum is present; rugose lines may be seen scattered all over the enamel, but the surface has been smoothed. The enamel bulges just below the cingulum and there are marked bulges at the bases of the parastyle and the metastyle. Occlusal surface. The centra] pit is almost obliterated, the enamel of the two sides being in close contact or even fused, and there is a small island of enamel cut off from the central pit in the direction of the metastyle. The roots are very similar in shape to M 1114 (vide infra), but they are ever shorter, their tips tapering to a sharp point very rapidly. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 461 M 264 (plates 33(2); 35(4), (F)) A right P?, which is in very advanced wear. The general appearance is that of M 531, except that the rugosity is more widespread and that the small island of enamel on the occlusal surface which leads up to the metastyle is in contact with the enamel of the rest of the central pit. The parastyle is markedly curved and twisted posteriorly on its own axis, as in M 531 and M 1114. As in M 531 too, the entostyle is not present because the tooth is worn above the level at which it projects in M 1114 (vide infra). Roots. They are rather smaller than in other teeth and there appears to be a constriction at the base of the posterior buccal root. M 1114 (plates 33(g); 35(¢), (g)) A right upper ? third premolar, fragmented and in advanced wear, with the metacone broken off. Lingual surface. A cingulum is detectable and there is a rounded bulge below it. There is a minute enamel nodule on the antero-lingual aspect in the region of the protostyle. It is fairly rugose. There is a marked entostyle on the posterior part of the lingual surface. Half the buccal surface is missing. The anterior half resembles the paracone of a premolar in that the enamel ridge passing to the apex is thick, rounded and triangular in shape and the enamel is folded to form a prominent parastyle. The central pit is irregularly V-shaped. It narrows towards the parastyle, but broadens out towards the metastyle where the enamel is partly broken away. As mentioned in M 264, M 531 and M 532, the buccal enamel of the protocone evaginates into the central pit and with wear it isolates (M 531, M 264) an island of enamel of the central pit in the region of the metastyle. This feature is common to the pre- molars in Giraffa and Sivatheriinae. Roots. ‘There are three roots. The lingual root is very broad at the base, triangular in shape, and the two buccal roots are smaller and separated by a V-shaped interval from each other and from the lingual root. The roots are all short. The two buccal roots are broken off half-way. M 552 (plate 31(g), (2), (w)) An incomplete, isolated right M1; the roots are broken off at the base. The tooth is in quite early wear. M 551 (plate 31(h), (£), (*)) It is an isolated left upper molar, M4, with the roots broken off at the base. Very slight degree of wear can be observed only on the anterior cones of the anterior pillar. M 550 (plate 34()) This is a fragment of a maxilla containing a left M?, while the sockets of the roots of M!—P? are visible. The tooth is very worn especially on the posterior 462 ANNALS OF THE SOUTH AFRICAN MUSEUM aspect of the posterior pillar where the enamel has disappeared completely. The dentine on the occlusal surface has been hollowed by an unusual type of wear. The maxilla is rather squashed and the empty root sockets are filled with calcite matrix. On the upper aspect of the specimen a portion of the maxillary sinus has remained. M 528 plates 32(/), 33(¢), 34(¢)) This is a fragment of maxilla containing left M* and M? which are in an advanced stage of wear. (2) LOWER DENTITION (a) Milk teeth M 939 (plate 37(2), (A), (-)) A fragment of a left mandible containing DM, and DMs, and part of the socket of DM, is present. M 540 (plate 37(4), (g), (2)) A fragment of a right mandible containing DM,—DM,. Measurements of mandible opposite talonid of DM, (mm.) Lingual height .. Ao Buccal height ies ears Thickness .. A io yO (b) Adult teeth M 936 (plates 32(¢); 33(4); 34(¢)) A fragmented lower right molar, probably M,. M 942 (plates 32( /); 33(¢); 34(2)) It is the posterior pillar of a right lower second molar. M 1113 (plates 32( 7); 33(¢); 34(4)) It is the anterior pillar of a right lower second molar. It and M 942 have identical contact surfaces and fit each other, forming a single tooth. M 938 (plates 32(¢); 33(4); 34(a)) It is a fragment of a mandible containing an erupting M, and there is also a socket for the posterior buccal side of the root of a M,: the posterior root of M 942 fits perfectly into this socket, and it is obvious that M 938 and M 942— M 1113 belong to the same individual. DETERMINATION OF THE SPECIES OF THE MAKAPAN GIRAFFA It has already been stated (see introduction to the description, chapter 3) that morphologically (non-metrically) these fossil specimens are indistinguish- DM2 DM3 DM4 M3 40 MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 463 Makapansgat Tr: A-P Index 50 60 70 80 90 100 0 120 130 140% Fic. 21. Transverse/A—P Index of Makapansgat Giraffa plotted against ranges of variation in modern Giraffa camelopardalis. @ = Makapansgat specimens. 464. ANNALS OF THE SOUTH AFRICAN MUSEUM able from the extant Giraffa camelopardalis. Statistical analysis of the dimensions of the specimens indicates that in general they fall well within the range of variation of Giraffa camelopardalis (figs. 19, 20, 21, table 39). However, it must be pointed out that in quite a number of cases, the dimensions tend to fall towards the lower limit of and even just outside the extant range; while in some cases, they fall beyond the upper end of this range (for instance, length and breadth of M 646, M 944; lengths of M 938, M 942 and M 11193). This distribution raises the possibility that the specimens may belong to another species or subspecies. Consequently the ¢ test was applied to both the: A-P and transverse dimensions of each of the specimens. The results of this analysis are given in table 31. It will be seen that the probability (P), i.e. of belonging to G. camelo- pardalis, is less than -o1 (1%) in only five cases. A-P Transverse Tooth Specimen t-value P smaller t-value P smaller than than 2 he M oat } 3°2156 ‘OI 2°1258 "05 M a f °5027 Wi "2974 8 DM* M 533 ) M 263 | MEd j 1°358 “2 1354 “2 i Ne 535 1 939 : . 3 , 2 M 540 -| 2663 8 1-464 2 DM M 939 | F M 540 -$ 9029 4 9443 4 DM, M 540 6874 5 ‘6801 5 ie M 532 "3984 7 "2550 8 Pps M 1114 | M 264 4°5118 ‘OI 1°5456 “2 M 53! | M!? M 263 M 552 "58 6 “791 5 SDE. , 2 ” NE PBB ir NAMNGEELE 9 2°3455 “02 M? M 528 19095 “I 2°846 ‘OI M, M 936 | M 942 ; 3°069 ‘OI 1°074 3 M 1113 | M,; M 938 1°64.54 ‘Ol "34.09 8 TABLE 31. Results of application of the ¢ test to the length and breadth measurements of the teeth of Giraffa from Makapansgat. (P = probability) A further step was to estimate the extent to which these ‘abnormal’ dimensions occur together in a single population. Because of a probable correlation between transverse and A—P measurements in individual teeth, the probability should only be considered for one or the other measurement, and consequently it has been estimated separately for each dimension. In respect of the A-P dimension, applying the binomia! distribution, the pro- bability that in 13 cases, 4 measurements occur on a less than -o1 chance is -09, which is not highly significant. It is obvious that P, estimated on the basis of MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 465, the transverse dimension, is still less significant. Therefore there is no basis for placing these specimens in a different species from G. camelopardalis. Furthermore, because the Makapansgat teeth are mainly isolated speci- mens and the measurements of the majority of the modern specimens were taken on intact jaws, slight differences in the actual measurements probably resulted. This would sufficiently account for the few isolated exceptional results. However it has also been established in other groups of the Hopefield fauna that the Pleistocene fossil representatives of modern species tend to have dimensions (of teeth and skeletal parts) towards the upper end of or just outside ranges of variation of the modern species (Ewer and Singer, 1956; Hooijer and Singer, 1960). B. SIVATHERIINAE UPPER DENTITION Milk M 937 and M 524 For the same reasons as given for Cornelia B! and B?, these two teeth are deciduous molars. (2) M 937 (plate 37(¢), (¢), (J)) It is either an upper left DM? or DM?. It is unworn. Buccal surface. ‘There is a cingulum on each of the cones which have a typical spatulated appearance, the paracone being less curved than the meta- cone. The general appearance is similar to the buccal surface of B? from Cornelia, except that the central ridges leading up to the apex of the paracone and of the metacone are much more prominent. The parastyle is broken off. Lingual surface. ‘The crown-root junction is fragmented, but a cingulum appears to have been present adjacent to this area, and there is a rounded bulge below the cingulum. The enamel is fairly rugose and is thrown into horizontal ridges on the anterior surface of the anterior cone. The lingual surface of the protocone has a thickened pilastering which leads up to the apex. This effect is slightly less marked on the hypocone. Occlusal surface. The hypocone has a V-shape, while the paracone tends to be slightly more U-shaped. The central pits are open in the region of the mesostyle. Roots. ‘The lingual root is broken off at the crown-root junction while the other two roots are embedded in a fragment of the maxilla and matrix. (>) M 524 (plate 37(¢), (), (4) A right ? DM®. Buccal surface. ‘There is a rounded cingulum. The metacone presents a typical spatulated surface, with the mesostyle curving outwards and forwards. 466 ANNALS OF THE SOUTH AFRICAN MUSEUM The paracone has a marked central costa which is very broad at the base and which has two vertical furrows near the apex. The parastyle is prominent with a thick rounded base. There are small nodular excrescences on the para- style which tends to recurve near the apex to give the cone its spatulated shape. Lingual surface. It is finely rugose. There is a rounded cingulum which is more marked on the protocone and which has a small denticulated ridge of enamel below it. There is a very marked protostyle on the anterior surface. On the posterior surface of the hypocone, the enamel ridge is much less marked and there is no cingulum on this surface. There is a small linear entostyle present. i Occlusal surface. ‘The central pits are wide and open, except in the region of the parastyle where the anterior part of the protocone fuses with the paracone. The hypocone is V-shaped while the protocone is flattened out. Degree of wear. The anterior pillar and the anterior part of the posterior pillar are slightly worn while the posterior part of the posterior pillar is just beginning to show signs of wear. M 941 This is either a right DM? or DM?+. The roots are broken away, and on the crown only the hypocone is more or less intact. It is in early wear. This tooth shows similar characteristics to those seen on M 937. The lingual surface slopes at a fairly marked angle towards the apex. The buccal surface also slopes but to a lesser degree. LOWER DENTITION (a) Milk teeth M 525 (plate 32(), (4), (¢)) This is a deciduous tooth and probably a left DM, (see Dental Index of M 525/M 539 B (DM,), vide infra). There is a small rounded cingulum which is particularly prominent on the hypoconid. The surface of the tooth is finely rugose and worn; there is a rather large bulge on the cingulum on the posterior aspect of the hypoconid. There is no cingulum in the region of the entostylid where the hypoconid fuses with the entoconid. Buccal surface. There is a fine groove between the hypoconid and the protoconid. The protoconid has a distinct bulge. Posteriorly the protoconid is fused with the enamel of the entoconid, and the metaconid juts out on the lingual aspect of this fused ridge. Anteriorly, the protoconid is continuous with the parastylid which also juts out at right angles from the enamel ridge, and this, in turn, fuses the protoconid with the parastylid. Consequently, there is a V-shaped depression between parastylid and paraconid on the lingual aspect of the tooth, although the two are fused on the lingual surface about half-way from the occlusal edge of the tooth. There is also a U-shaped depres- sion on the lingual side of the protoconid between the projections of paraconid MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 467 and metaconid, which are fused to each other about two-thirds of the way down from the occlusal edge of the tooth, so that there is a V-shaped interval _ between them on the lingual surface. There is a small depression between metaconid and entoconid near the occlusal edge. A linear central pit stretches between the hypoconid and the enamel ridge, which links entostylid—to— entoconid—to—protoconid. There is a small cingulum on the lingual surface. ‘There are two hollow roots, a posterior one below the hypoconid and an anterior root below the protoconid. A distinct similarity is observed between this tooth and P, of M 553 B. The only differences in appearance being that the metaconid pro- jects in a more lingual direction in M 525, and that the enamel ridge forming the central pit of the paraconid in M 553 Bis circular, while in M 525 it is more V-shaped, and that the occlusal edge of the lingual surface is not as high in M 525. M 539 B (plate 38) This is a fragment of a right mandible containing a DM, which has just erupted, and the tip of paracone of M, is visible because the surrounding alveolus is broken away. Buccal surface. ‘The enamel of DM, is fairly rugose. ‘There appears to have been a cingulum and there is a large hypostylid and a broken-off ectostylid. Lingual surface. ‘There is a rounded cingulum. Finely rugose. On the para- conid there is a broad, flattened central costa leading up to the apex of the pillar, while the parastylid and mesostylid are less marked; the parastylid curves slightly inwards at its anterior extremity, so that the surface presents a slightly spatulate appearance. The metaconid has a similar appearance, but anteriorly it fuses with the mesostylid of the anterior cone. The lingual surface of the anterior pillar overlaps that of the posterior pillar which in turn overlaps that of the talonid. The talonid has a very marked central costa leading up to its apex from the cingulum. This pillar is particularly broad at its base so that it appears to occupy most of the surface. The anterior portion of the surface curves slightly inwards, and comes to lie against the metastylid of the posterior cone. Occlusal surface. ‘The A—P axis of the lingual surface of the posterior cone is at a slight angle to that of the talonid; the anterior portion of this cone is angulated at an even greater angle to that of the posterior cone. The central pits are wide and open, and on the buccal surface of the paraconid the enamel is markedly ridged, so that it projects backwards into the central pit. The buccal cones are V-shaped in occlusal view, the central portion of each surface producing a ridged effect which leads up to the apex of each cone. Roots. Through the broken mandible the top of the short central root is visible. On X-ray (plate 38(c)) the outline of the developing tooth bud of P, may be seen, as well as the outline of the alveolar ‘pocket’ containing M, which is still embedded in the mandible. 468 ANNALS OF THE SOUTH AFRICAN MUSEUM Measurements of mandible (mm.) Breadth opposite the talonid .. 3.20 Breadth opposite M, .. Aenats ¥1-'0) Height opposite the talonid .. ¢. 47 The Dental Index (see Section I, chapter 5) has been calculated between left DM, (M 525) and right DM, (M 539 B), in order to determine whether they belong to the same jaw. All three ratios fall well within the range of variation of the corresponding index for the corresponding measurements in extant G. camelopardalis and are very close to their mean values. Length index Breadth index Tr.|A-—P index M 525 /M539B ae 60-0 75°6 124°0 G. camelopardalis : Mean oth sas 63°71 F@Qe7 123°5 Range) eae ~5 1 55:6-72"3 7029251 98°7-144 TABLE 32 (b) Adult teeth M 1116 and M 527 These are two incisors in very early wear; M 527 is at a slightly earlier stage than M 1116. Judging by the slope of the anterior edge of the enamel of the teeth, it would appear that M 1116 belongs to the left side and M 527 is right. M 1116 (plate 36(d), (e), (f)) Buccal surface. ‘The enamel is finely rugose and there is a distinct cingulum which bulges out on the lingual and the buccal aspect. However, the cingulum is absent on the sides of the tooth. In buccal view, the tooth is rather V-shaped, but the apex of the V (i.e. at the crown-root junction) is flattened, while the upper edges of the V are slightly flanged outwards. The outer surface is convex in a mesio-distal plane. Lingual surface. ‘The anterior edge of the enamel is spade-shaped, while the posterior edge is irregular ( | | ). The dentine is hollowed out, the edges of the dentine sloping towards a deep central portion; it 1s slightly worn at its anterior end. Roots. The root is almost complete and rather oval in shape, being slightly flattened anteriorly, resembling very much the root of Hopefield 4026. Viewed from the side, the crown-root junction has an inverted V-shape. M 527 (plate 36(a), (b), (c)) | This appears to be a right incisor, identical in appearance and shape to M 1116, but being a little broader from side to side and in a mesio-lingual plane at the crown-root junction. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 469 Measurements (mm.) M 1116 M 527 Total height : oe ie ye if Ce Gees — Crown (buccal) height St. ‘- bf ey e 37°4 38-4. Length om ae 22°0 22-7) Maximum height G just above the crown-root junction, i.e. at its highest point) sb - ee Re 17°6 17°4. TABLE 33 M 553 A (plates 39(2), (4); 40) This is a fragment of a left mandible on which the posterior part of the symphysis menti is present. It contains a P, about to erupt as well as the tip of the root and I, and the breccia-filled sockets for the tips of I, and I,;. There is no evidence of the canine. On the outer aspect, near the front of the fragment, a large foramen mentale is present; it is oval in shape, with rounded edges, and leading from it anteriorly there is a deep groove. The tip of the canine is approximately opposite the mental foramen which in turn is opposite the posterior end of the symphysis. The superior border of the fragment is very sharp, while the inferior border is rounded. The medial surface is smooth and convex in shape, while on the outer surface, there is a deep horizontal furrow about one-third of the height below the superior border. The symphysial region is markedly irregular and its posterior end is oval in shape and flattened from above downwards. P, Although partially embedded in the mandible, most of the crown is visible for description. Buccal surface. Fairly rugose. The crown-root junction is not visible and a cingulum formation cannot be commented on. But the hypoconid is rounded. and bulging above the crown-root junction, and slopes gradually towards the apex. It is split vertically by a deep groove which extends down half-way from the occlusal surface. On the posterior surface there is a thickened ridge of enamel, at the lingual end of which the medial portion of the entostylid is broken off. Lingual surface. ‘This is also fairly rugose, and its prominent feature is the metaconid which bulges out on the surface so that there is a hollow between it and the entoconid. But anteriorly there is only a slight depression between the metaconid and the paraconid; the parastylid is not visible. Occlusal surface of this unworn tooth is irregular in shape, the highest point being the apex of the protoconid which is still fused with that of the metaconid. The absent canine is broken off near the tip of the root. All that one can see is its circular shape. It almost occupies the whole breadth of the bone in this region; it is Ig mm. in diameter. Height opposite anterior end of P, (buccal 1a) sk as = 67°7 Height opposite mental foramen se ve fe 67:2 Maximum thickness opposite P, , 26:5 Distance (direct) between anterior edge of alveolus of P, and posterior edge of mental foramen g9°0 Distance between anterior edge of alveolus of P, “and Posterior border of symphysis ae Se 102°5 TABLE 34. Mekal of “ane an ys 470 ANNALS OF THE SOUTH AFRICAN MUSEUM M 553 B (plate 39(c), (d)) This is a fragment of the right half of a mandible containing P, and P,, and showing a part of the alveolar fossa for P,. Beyond P, the sharp anterior border continues for about 7 cm. to the broken end where there is a vertical fracture. On the buccal aspect of the mandible there is a broad and deep groove which almost disappears opposite P,: here the buccal surface tends to be almost straight in a vertical direction. The Jingual surface is convex and the inferior border is rounded, as in M 553 A. Teh, It is just erupting and partly broken. Part of its alveolar socket is broken away, so that one can ascertain that it is similar in appearance to P, of M 553 A. Ps It is still embedded in its bony alveolar socket and only its tip is erupting. Most of the lingual surface of the alveolus is broken off, so that a large part of the crown is visible for study. The buccal surface is fairly rugose, while the lingual surface is slightly smoother. | The highest point of the tooth is the protoconid which is shaped like an inverted V, and this cone occupies most of the buccal surface of the tooth. The hypoconid is small and in continuity with the protoconid. Half of the lingual surface is occupied by the entoconid and metaconid, the metaconid being almost at a right angle to the lingual aspect of the protoconid and is fused with it just posterior to the protoconid apex: in this way an irregular L-shaped central pit is formed. Between the metaconid and the paraconid there is a wide interval, the two being joined laterally by a part of the lingual aspect of the protoconid. The metaconid meets the paraconid just above the crown-root junction so that, viewed from the lingual aspect, there is a V-shaped interval between them. The paraconid is angulated at right angles to the A—P axis of the body of the protoconid, while the anterior part of the protoconid is also angulated at right angles to the rest (body) of its own axis. Thus an oval- shaped pit is formed between the protoconid and the parastylid. On the lingual surface, the parastylid is seen as a prominent bulge, the apex of which is at the occlusal junction between the recurved paraconid and the protoconid. Maximum breadth opposite P;_.. ae vs 35°0 Maximum breadth opposite P, is e. 27°0 Height opposite anterior end P, (buccal side) si 67°3 Height opposite anterior end P; (buccal side) .. 77°4 TABLE 35. Measurements of mandible (mm.). M 553 B} (plate 41) This specimen was described by Cooke and Wells (1947) as Griquathertum cingulatum neotype. It is a fragment of a left mandible containing M, and M,. Posterior to M,, where M, would have been, there is a mass of limestone matrix. Anterior 6 MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 471 to M, the posterior cones of P, are just erupting through the alveolar surface, and there is a vertical fracture through the anterior cones, indicating the early stage of eruption of P,. On X-ray, the socket for the anterior root of M, is seen filled with breccia. The roots of the other teeth are very long, penetrating more than two-thirds of the height of the body of the mandible. The crown- root junction is well below the alveolar margin. M, The tooth is in early wear, and hypsodont in appearance. It slopes slightly anteriorly from below upwards. The buccal surface is fairly rugose and there is no cingulum formation visible. ‘The enamel has slight irregularities on it, and there is a small protostylid. Looking at the posterior surface, it can be seen that the buccal surface has quite a marked slope towards the apex, and that, in similar fashion, the lingual surface has an even more marked slope, especially in the region of the entoconid. Lingual surface. The crown-root junction is not visible, but there appears to be no, or very little, cingulum formation. The surface is fairly rugose. The rugose lines of the enamel tend to radiate upwards from the crown-root junction and forwards and backwards from the central costa in a fan-shaped fashion. The entostylid is broken off, and the anterior portion of the entoconid is in contiguity with the metastylid at the occlusal surface, but the enamel on their lingual surfaces tends to fuse lower down. The metastylid is partly broken off but it is fairly marked, while the parastylid is short and blade-like, and it is sharply angulated anteriorly. The central costa passing from the crown-root junction to the apex of the metaconid is much thicker and more marked than that of the entoconid. Occlusal surface. The four cones are almost completely separated from each other. The protoconid is contiguous with the paraconid at the apex of the parastylid, while the hypoconid meets the entoconid in the region of the ento- stylid. But the central pits are not completely closed off where the anterior pair of cones meets the posterior pair of cones, and the medial and lateral lips of the central pits are rather widely separated. At this early stage of wear, the dentine appears as a narrow strip in each cone, and the protoconid and the hypoconid are V-shaped, while the metaconid and the entoconid tend to be more flattened. M, It is in a slightly more advanced stage of wear than M,, and has the identical appearance of M,, except that the dentine is thicker in each cone. Height opposite anterior pillar M, (lingual aspect) 78°5 Height opposite anterior pillar M, (buccal aspect) = 75°0 Breadth opposite anterior pillar M, (maximum). . a 49°7 Breadth opposite anterior pillar M, (maximum) .. se 55°5 TABLE 36. Measurements of mandible M 553 B! (mm.). 472 ANNALS OF THE SOUTH AFRICAN MUSEUM Note.—M 553 A and M553B must belong to opposite sides of the same lower jaw, because of: 5 (1) the proximity of the discovery; (2) the same dental age, judging by the stage of eruption; and. (3) the similar size and shape of corresponding regions of their fragments and of their teeth (P,). That portion of M 553 B which we consider to correspond to the posterior portion of A has been reconstructed (fig. 22). It appears obvious that M 553 A must belong to M 553 B?; this is in accordance with the degree of eruption of P, and the degree of wear of M, and M, in the latter fragment. Consequently these three fragments must belong to the same individual. (See also ‘Appendix’.) M 55381 Fic. 22. Reconstructed mandible of the Makapansgat Sivatherium olduvaiense. M 943 This is a fragmented posterior pillar of a right M,. It has all the features of the corresponding pillar of M, of M 553 B!, but it shows more clearly the groove separating the hypoconid and the entoconid on the posterior surface, just above the point where they fuse. The degree of wear and the general features are so similar to the above corresponding specimen that they could almost be considered to belong to the same jaw. The lingual surface has a large piece of enamel broken off; the whole of the buccal surface may be measured because the tooth is not embedded in its jaw (table 40). CHAPTER 4. HOPEFIELD (CAPE PROVINCE, UNION OF SOUTH AFRICA) The description of the geological features of the fossiliferous site on the farm ‘Elandsfontein’, near Hopefield, 60 miles north-west of Cape Town (fig. 23) has been recorded previously (Drennan, 1954; Singer, 1954; Mabbutt, 1956). All the material, referred to as ‘the Hopefield specimens’ have been recovered from the site and are now housed in the S.A. Museum. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 473 No stratification has been found on the site which consists of approximately 2 square miles (5 km.”). “The site lies in the sand veld between the Sout River and the Langebaan-Saldanha Lagoon. This site is 300 feet above sea-level and is divided into wind-scoured kloofs or depressions by sand-dunes which are either drifting or, where covered by vegetation, stationary. Ridges of SALDANHA Oo BAY HOPEFIELD A <9, ELANDSFONTEIN Ry , DARLING fe) CAPETOWN fe) e) English miles TOMS 20) Fic. 23. Map indicating the Elandsfontein fossil site, near Hopefield. ferricrete cut diagonally across the length of the site, and in places the dunes are capped by massive calcrete mounds or flat boulders of partly silicified surface limestone. Softer, cellular calcretes are found in certain places at the lowest parts of the depressions. The tortuous courses of the ferricrete ridges indicate that they are the indurated lower flanks of old sand-dunes* now * Recent research by Mr. D. Needham, University of Cape Town, tends to contradict that there were sand-dunes, in the geological sense, originally present. Convincing evidence indicates that the modern ‘dunes’ are the result of a recent ‘heaping-up’ of the tertiary marine- deposited sands. 4.74: ANNALS OF THE SOUTH AFRICAN MUSEUM stripped bare of the sand walls. This ferruginization is usually associated with moist ground conditions, a fairly high stable water-table and an abundance of vegetable acids in the soil. It seems that this fossil site was at one time a large vlei or lagoon along the edge of which animals roamed.’ (Singer and Keen, 1955): The movement of the present sand-dunes uncover the calcareous floors from which the fossils and artefacts are recovered. The majority of the giraffid teeth as well as the cranial fragments have been recovered from different localities on the ‘Main site’. However, fragments 4029, 4029 A, 4029 B and 4374 were discovered on a ferricrete deposit in site E-extension (Tex. 1). Further tooth fragments (4030, 4030 A and 4031) were found on and near a ferruginous plateau in ‘Buffalo Bay’, less than 50 yards from E-extension, but separated from it by a large sand-dune, and it is quite likely that the ferricretes of the two localities are in continuity through the sand-dune. | As a matte of interest, specimens 4029 (heavily encrusted with ferricrete) and specimen 4028 A (found on a calcareous floor and only slightly ferrugi- nized) were X-rayed together at a distance of 3 feet with 115 kv. for 2 seconds. The skiagram showed that 4029 was far denser than 4028 A, which possibly signifies that it may be more heavily fossilized. ‘The disparity in the skiagrams is also reflected in the Uranium (U,O,) content (by courtesy of Dr. K. P. Oakley, British Museum of Natural History), that of 4028 being 14+2 p.p.m. and that of 4029 being 4-+2 p.p.m. This indicates that the latter may have come from a higher level, and that conditions for fossilization may have been more favourable. The fluorine contents (by courtesy of Mr. H. E. Krumm of African Metals Corporation, Bellville, Cape) are 1-983°% and 1:663°% respec- tively, and the nitrogen contents are 0-11 and 0-085 respectively. List OF MATERIAL Hopefield S.A. Museum No. No. A. GIRAFFA 3345 11716 right P? B. SIVATHERINAE (1) Cranial fragments 4372B 11717 base of a skull 4372 ena] posterior horn-core 4372 A 11717 fragments belonging to 4372 4373 11718 posterior horn-core 4373 A 11718 anterior horn-core 4373 B, CG 18 yp te) fragments belonging to 4373 (2) Teeth Upper 4025 11719 right P® 4027 11720 right M+ 4023 11721 left M? 4024. 11722 left M? Lower 4026 11723 iNCisOr | 4374 11724 left M, or M, 4028-4028 A 11725 fragment of left mandible with M,—M, 4029— 4029 B 11726 fragment of left mandible with M,—M, 4.030 119727 right M, 4.031 11728 right M, 4030 A Tali 7207 fragments of roots and pillars MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 475 DESCRIPTION A. GIRAFFA 3345 (plate 45(c), (d), (2) This is a right P?. The tooth is in late wear. Buccal surface. There is a well-defined cingulum. The surface has a ‘W’ formation with marked parastyle and metastyle as well as a marked paracone. Between the latter and the two styles there are deep grooves. In anterior or posterior view, the lingual and buccal surfaces slope towards each other in the direction of the apex. Lingual surface. ‘The enamel is finely rugose and there is a well-marked cingulum. The lingual surface slopes sharply from the cingulum towards the apex in a buccal direction. There is a smooth surface on the posterior aspect of the tooth indicating contact with P?. Occlusal surface. ‘The enamel of the paracone is arc-shaped, while the occlusal edge of the lingual surface is more or less straight. The central pit is shallow (because of the marked attrition) and it is slightly curved, the ends pointing towards parastyle and metastyle. Roots. ‘There are three roots, one lingual and two buccal, the one buccal being broken off at its base, while the other is partly broken. The roots are very short and tend to be triangular in shape. Diacnosis. Its features and dimensions exclude it from the Palaeotraginae, such as Giraffokeryx and Okapia. The flat profile and slope of the lingual surface of the tooth are uncommon features among G. camelopardalis in which a rounded bell-shaped body is the typical shape, but they are observed in G. gracilis (Arambourg, 1947, pl. XXII, 1a). The only premolar known to represent G. gracilis is a P* which is smaller in its overall dimensions than that of G. camelo- pardalis. Although 3345 falls within the range of G. camelopardalis both for A-P and transverse dimensions it tends to be at the lower end of the range. As it is not possible to compare 3345 directly with a G. gracilis specimen, it is tenta- — tively assigned to G. camelopardalis. This view is strengthened by the fact that the Transverse/A—P index of 3345 is identical to the mean index (123) of P? in G. camelopardalis. Nevertheless it should be kept in mind that the non- metrical features of the Hopefield specimen are very suggestive of those of the G. gracilis premolar. This fact may be more significant than its metrical relationship to G. camelopardalis because of the extensive range of variation of the latter (table 7). B. SIVATHERIINAE (1) CRANIAL FRAGMENTS 4372 B (plate 42(a), (4)) Five fragments, found in situ associated with the horn-cores 4372, 4373 and 4373 A, were reconstructed to form the base of a skull in the region of the 4.76 ANNALS OF THE SOUTH AFRICAN MUSEUM foramen magnum. It contains portions of the two occipital condyles, most of the lower and most of the upper border of the foramen magnum and a portion of the occipital bone. On the inner aspect of the fragment, the bone structure consists of numerous large and small inter-connecting sinuses. The medial edges of the condyles project inwards slightly, decreasing the side-to-side breadth of the foramen magnum. The medial condylar surfaces are not com- plete, while the lateral articular portions are missing. The medial surfaces are slightly convex from side to side and fairly convex from front to back. There is a small groove on the postero-medial aspect of each: condyle where it is continuous with the base of the skull. The striking features of the foramen magnum, viewed from the inferior aspect, are its large size and its almost circular appearance. The postero-superior border of the foramen magnum forms a broad, flattened arc rather than the acute-angled arch of the modern giraffe. Although a portion of the bone is missing above this region, it is clear that the protu- berant mass of solid bone where the two exoccipitals and the supraoccipital fuse in a triradiate fashion, which is present in the modern giraffe, is absent in the fossil specimen. In fact it appears that there is a thickened pillar of bone (the region of the exo-supraoccipital suture junction) leading up to the occipital crest. This feature is very similar to that observed in Szvatherium giganteum (see Falconer and Cautley, 1846-9, plate XCII, 1c), except that in the Hopefield specimen the postero-superior border of the foramen magnum appears broader than in giganteum. Furthermore, the postero-superior portion of the occipital condyle is angulated quite differently to the foramen magnum than in the modern giraffe in that the plane of this region of the condyle is parallel to the occipital surface, while in the modern giraffe it is almost at 90.° This may indicate that the skull was balanced on the vertebral column at an obtuse angle, rather than the right angle between the vertical axis of the upper cervical vertebrae and the plane of the base of the skull of the modern giraffe. Mechanically this would be in accordance with the short neck of the Siva- theriinae, which probably contained a group of powerful extensor muscles to assist In maintaining the balance of the large skull with its massive, heavy, curving horns. Measurements 4372 B Sivatherium giganteum* mm. inches inches Minimum intercondylar breadth .. 56-0 ee 2°6 Foramen magnum: Maximum A-P, internal .. zh 51 2°0 Maximum A-P, external .. ate 58 2°3 Pe! Maximum breadth, external Me 58 2°3 Maximum breadth, internal st 61 2°4 TABLE 37 1 Falconer and Cautley, 1836. 4372 (plate 43) This is a Sivatherine posterior horn-core (so-called ‘antler’?) which was found to be fairly complete after reconstruction of the fragments. The base is MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 477 pear-shaped, narrow anteriorly, the posterior part being rather globular. The lower part of the horn-core, as high as the flange, is hollow, narrowing superiorly, and only at the lowest end (for c. 12 cm. from the broken edge) is sinus formation visible. The base is fragmented and incomplete, the anterior portion missing and the flange broken off at the anterior border. The body of the horn-core is fairly complete, but the tip is missing. Just above the base the core turns outwards, and at the base of the second knob it begins to twist posteriorly and laterally, so that the outer surface tends to face anteriorly. The posterior border is rounded; the anterior border at, and above, the flange narrows (and it is more angular than at the base) while nearer the tip, above the second knob, the anterior border becomes rounder. There are deep longitudinal grooves running up on the medial convex surface. There are three grooves very close together at the base of the flange apparently diverging up from a single point. At the upper end of the flange the posterior groove swings sharply towards the posterior border, while the anterior two slowly separate, and the posterior of these two swings more acutely backwards opposite the second knob. The anterior groove runs up parallel to the anterior border as far as the second knob where it crosses the medial surface posteriorly and upwards towards the tip. In addition, arising from the anterior part of the base, there are three other grooves which travel obliquely upwards (from a point behind that of the above-mentioned grooves) towards the posterior border opposite the middle of the flange, and there they tend to disappear towards the back. Measurements (mm.) : Circumference at base .. ar c. 4.00 Circumference between flange and knob 2 ae ¢. 330 Circumference 100 mm. above knob 2 .. a, c. 265 Circumference at tip ‘ ah ag Total length along posterior border... die c. 590 A-P Side-to-side Base ae aP eA Gat 70 G27 Between flange and knob 2 i. ae oie ¢. 130 82:0 Above knob 2 pee mm. ) ah a Nae QI‘O 72:0 At tip = i : oe on ae = = TABLE 38 4373 (plate 44) A posterior horn-core, the partner of 4372. No additional features are observed on it as it has the same general appearance of 4372. But, viewed from the front, the anterior border has a distinct sinuous (‘twisted’) appearance (see also Old. 86 (BK II)). It is rather incomplete, especially on the anterior border and convex surface. Therefore, the measurements would all be approxi- mate and as they would add nothing to what has been obtained from 4372, no measurements are given. 4373 A (plate 42(c), (d)) This is the posterior portion of an anterior horn-core found in association with 4372, 4373. One surface is markedly concave, leading from the base to an 478 ANNALS OF THE SOUTH AFRICAN MUSEUM irregular rounded knob (cf. C 431 from Orange Free State). No sinuses are visible at the base which is broken away. This is the first Sivatherine anterior horn-core to be described from Africa. Thus it becomes the paratype for Stvatherium olduvaiense (see ‘Discussion’, Section III). Height of fragment .. 160 mm. Tip of fragment .: 55 mm. (transverse) X 59 mm. (A-P). 4373 B A number of small fragments found with 4373 which cannot be fitted into the reconstruction. 4372 a, b, c, etc. Groups of fragments found in association with 4372. Some are tiny fragments of the horn-core, but there are also a few fragments of skull, too small for identification or description. (2) TEETH 4025 (plates 45(2), (/), (g); 50(2)) This is an isolated tooth, a right P3. Buccal surface. There is a rounded cingulum. The posterior half of this surface is missing. The parastyle bulges markedly below the cingulum, while a deep cleft separates the parastyle from the paracone. Lingual surface. ‘The enamel is fairly rugose and there is a marked rounded cingulum. Roots. There are three roots. The lingual root which represents the root of the protocone is massive and rounded, and it projects medially. The buccal roots are broken off, but the anterior appears larger than the posterior root. General shape. The tooth is very large, the buccal surface being flattened, while the lingual surface is arc-shaped. The central pit tends to be U-shaped, its enamel surfaces being widely separated in the centre; the ‘arms’ of the U taper towards the anterior and posterior ends of the buccal surface. The enamel on the lingual aspect of the pit is evaginated into it posteriorly, and the dentine here contains a rounded cone of enamel (cf. Makapansgat, M 1114). The paracone is wide from side to side, and its dentine tapers both anteriorly and posteriorly. 4027 (plates 46(a); 47(a); 48(a)). An incomplete isolated right upper molar, probably M1}, in rather advanced wear. Buccal surface. Marked, rounded cingulum. The tooth is rather fragmented and the paracone is missing. The metacone is similar to that of M?, 4023 (vide infra), and the mesostyle again is prominent. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 479 Lingual surface. The rugosity is finer than in the other teeth. The cingulum is marked and rounded. There is a bulge below the cingulum which is more marked in the anterior than in the posterior pillar. In the posterior pillar, this bulge gives way to a sharp outward angulation of the lingual surface in the direction of the apex, similar to 4024 (vide infra). The two pillars meet on the lingual aspect, similarly to those of 4024, 1.e. the folding of the two contiguous enamel sides of the pillars being in an anterior direction. Occlusal surface. The dentine is hollowed out on the 4 cones, forming an arc in a side-to-side direction. The central pits tend to be U-shaped: as in many of the other upper molars, the enamel on the lingual side tends to be U- shaped, while on the buccal side it is V-shaped. The U-shaped side however is more worn than the V-shaped one. The inner arms of the two pits are not continuous, although the enamel of the two arms are touching each other. As in M? (4023), the posterior part of the pit of the anterior pillar has an evagination directed buccally towards the mesostyle. Roots. ‘The lingual root is broad, thickened above each pillar and a vertical groove on the medial aspect demarcates the two parts. The posterior buccal root tends to be oval in shape and flattened from front to back in an antero- posterior direction. The anterior buccal root is missing. 4023 (plates 46(d); 47(a); 48(a)) An isolated left M?, found in association with 4024, with which it estab- lishes a contact. The tooth is in a rather advanced stage of wear. Buccal surface. ‘This is similar to 4024 (vide infra) but the mesostyle, which is here complete, is very marked and appears to be associated with the posterior rather than with the anterior pillar. In the central hollow of each pillar there is a slight ridge of enamel (the costa), stretching up from the cingulum towards the apex of the tooth. This is more marked on the anterior pillar. The para- style is broken off. Lingual surface. There is a marked cingulum below which there is a slight bulge in each pillar. The enamel is rugose. The enamel is raised in a marked ridge on the anterior aspect of the tooth (protostyle). Occlusal surface. ‘The anterior pillar is again larger than the posterior one, and both tend to be arc-shaped on the lingual aspect. The slope of the buccal and lingual aspects are similar to those of 4024 in that the lingual surface of the protocone has a more vertical slope than that of the hypocone, while this is reversed on the buccal surface where the metacone is more vertical than the paracone. The central pits are more V-shaped, although the apex of the V, which points lingually, is slightly flattened. The inner arms of the V of each pillar become continuous with each other in the region of the mesostyle. Just on the anterior side of this point the pit of the anterior pillar has a marked evagination, which almost reaches the enamel junction of the two pillars on the lingual aspect. 480 ANNALS OF THE SOUTH AFRICAN MUSEUM Roots. ‘They present an identical appearance to those of M? (4024), but the tip of the lingual root tends to be angulated in a vertical direction, and the lingual root is shorter than the buccal ones. The posterior buccal root is larger than the anterior one; both are triangular in shape. The tips of the anterior and posterior buccal roots curve anteriorly and posteriorly respectively. 4024 (plates 46(c); 47(¢); 48(¢)) An isolated left M?. It is fairly complete, rather cracked, with most of the buccal roots missing. Buccal surface. The enamel is rather rugose and not as smooth as in the lower teeth. The cingulum is well marked and rounded but in the centre of each pillar the area just below the cingulum tends to be scooped out. The cingulum becomes continuous with parastyle, mesostyle and metastyle, forming costae, so that viewed occlusally, the surface has a hollowed-out effect between these three costae. ) Lingual surface. ‘The enamel is rather rugose. Below the rounded cingulum there is a slight bulge which is more emphasized on the posterior pillar. The enamel is thrown into horizontal ridges on the anterior and posterior surfaces about 5 mm. from the crown-root junction (forming the protostyle and hypo- style respectively). ‘The lingual and buccal surfaces of the tooth tend to slope towards each other in the direction of the apex, as in 4023. On the posterior pillar, there is a marked depression between the hypocone and the metastyle, which tends to accentuate the metastyle. Similarly, there is a depression between paracone and parastyle. This ‘pinched’ effect emphasizing the meta- style and parastyle is found close to the crown-root junctions and is a common feature in Sivatherines (e.g. F 3655, C 426, MMK 3685). This effect may also be seen on the posterior side of the posterior buccal] root (e.g. C 426, 4024). The junction between the anterior and posterior pillars is not marked, the enamel surface of each pillar being contiguous for about 6 mm. towards the centre of the tooth. Occlusal surface. The shape of the lingual enamel edge of the two pillars is V-shaped, but the posterior is much narrower and smaller, the latter effect being produced by the more acute slope of its lingual surface towards the apex. The central pits are U-shaped, the central portion of the U being fairly wide, while the enamel sides of the ‘arms’ tend to come close together. The outer arms of each pit curve, one towards the parastyle and the other towards the metastyle, while the two inner arms are directed towards the mesostyle. The mesostyle is missing in part. The central portion of the U in the anterior pillar has a slight evagination posteriorly in a lingual direction. ‘The image of this in the central pit of the posterior pillar is broken. The tooth (as well as 4023) illustrates very well the typical occlusal wear. The enamel and the dentine are raised in a straight line from side to side in the centre of each pillar, while they are well worn along the anterior aspect MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 481 of the anterior pillar and the posterior aspect of the posterior pillar. Maximum wear has occurred along the contiguous surfaces of the two pillars. Roots: The lingual roots of each pillar are fused by a plate, so that in effect one only finds one broad root on the lingual side, with the anterior pillar reflecting the largest bulge on the root. The two buccal roots are separated, but the posterior one is broken off near the base and the anterior root is broken at the crown-root junction. The lingual root is angulated in a medial direction. The buccal roots are arranged so that the posterior root points backwards, and the anterior one forwards and outwards. 4026 (plates 45(4), (¢), (7); 50(4)) This is an incomplete isolated incisor. Assuming for the purpose of description that it is a right incisor, the anterior half of the crown is broken, mainly on the buccal aspect. The enamel is finely rugose. The enamel on the occlusal edge of the tooth on the buccal aspect is broader than it is at the crown-root junction. Looked at from the side, the ‘apron’ appearance of the enamel is again observed. On the posterior surface the enamel has a sharp ridge running up from the cingulum to the occlusal surface. The posterior surface on the lingual side of this ridge has no cingulum. Near the occlusal edge this ridge is worn, suggesting contact with a contiguous tooth. The lingual aspect of the tooth has no enamel, indicating that the tooth is in advanced wear, and the dentine is slightly raised at the crown-root junction. On the buccal aspect, in the centre of the tooth, the enamel presents a groove which extends from the occlusal edge to the bulge above the cingulum. Measurements of 4026 (mm.) Total height .. en A a ae ies 72+ Crown (buccal) height a a es Se BT) Length (A-P) .. ey ee: Be a Soha Maximum breadth at the crown-root junction .. 18-4 4374 (plates 46(d); 47(¢); 48(d)) An isolated, very fragmented lower left molar, M, or M,, found near 4029, 4029 A and B. It consists of the posterior root, part of the hypoconid and entoconid and a small piece of the protoconid. The tooth is very worn. The buccal enamel is rather rugose, the region of the cingulum is broken away and the lingual surface is also broken. In occlusal aspect, the central pit is visible, rather flattened in shape, with its anterior portion curving towards the metastylid. : Measurements of 4374 (mm.) Buccal height a eo Pe oy Length of the hypoconid .. c. 28 482 ANNALS OF THE SOUTH AFRICAN MUSEUM 4028 (plates 46(e); 47(e); 48(e)) A fragment of left mandible, containing M, and M, (4028 A), in advanced wear. Ms; Buccal aspect. ‘The tooth is hypsodont and the enamel is fairly rugose. The cingulum is marked, more particularly on the posterior pillar and on the talonid; it is more rounded on the buccal than on the lingual side where it is only represented by a faint ridge. At the anterior part of the posterior pillar, the cingulum bulges in the region of the ectostylid. Similarly on the talonid, at the anterior end of the cingulum, there is a nodular hypostylid. The protoconid tends to be angular on the buccal side, and even presents a slight vertical central ridge, which is not observed on the hypoconid. Lingual surface. ‘The rugosity is smoothed and even disappears opposite the centre of each pillar where there is a slight bulge (costa). At the base of the crown there is a bulge on each pillar, the posterior pillar having the larger bulge. The bulge on the posterior pillar, which is unusually prominent and extensive, tends to resolve itself in an upward direction in three ill-defined ridges: the posterior ridge travelling in the direction of the entostylid, the central ridge (costa) travelling towards the apex of the tooth, and the anterior ridge in the direction of the metastylid. The enamel of the crown of the anterior pillar and of the talonid is broken off. About 30 mm. above the cingulum of the posterior pillar, on the lingual surface, the enamel forms a horizontal ridge which extends across the anterior pillar and towards the top of the talonid. There are also a few other irregularities of the enamel giving the appearance of ridges. The buccal surface of each pillar looked at from the side is vertical, whereas the lingual surface is slightly angulated laterally towards the apex. Occlusal surface. The protoconid and hypoconid form a triangle, but the apex on the buccal side is an obtuse angle of about 110°. The talonid is large; its occlusal view presents an oval surface, flattened from side to side, and its A-P axis is at an angle of 45° outwards to that of the other two pillars. The central pit is irregular and flattened from side to side so that in the centre of each pillar the enamel of the two sides of the pits is almost touching. The pit is confined to two pillars and only extends very slightly on the occlusal surface into the talonid. The enamel on the buccal side is fairly regular and wavy, being indented at the junction between the two pillars, whereas the enamel at the lingual side gets broader and evaginates towards the parastylid and also towards the metastylid. Roots. There are two roots: the anterior pillar having a single root which tends to bulge on both the buccal and lingual sides, these bulges being connected by a short flattened plate. The root tends to be vertical and it is separated by a triangular interval from the root of the posterior pillar. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 483 In the region of the talonid the root is also thickened and continuous with the thickened portion of the posterior pillar by means of a hollowed-out plate, so that in effect one observes one massive root. The tips of the roots tend to curve buccally. M, Buccal surface. ‘There is a cingulum which is not as rounded as in Mg, but on the lingual surface it is slightly more exaggerated than in M;. Also on the anterior surface of the anterior pillar, the cingulum is well marked. In the region of the ectostylid the cingulum tends to bulge somewhat. Lingual surface. There is a slight bulge above the base of the posterior pillar, and the resolving upward ridges are similar to those in Mg, but less distinct. The bulge above the cingulum of the anterior pillar is less marked, and is really just a thickening of the cingulum; it appears to lead up to the broken parastylid. ‘Towards the posterior part of the anterior pillar, another ridge (costa) stretches upwards from the cingulum towards the apex of the tooth, and between this ridge and the parastylid there is a slight depression. The enamel forms numerous horizontal ridges. Occlusal surface. Each pillar has a separated central pit: the pits are only very slightly separated in the region of the metastylid; the posterior one tends to be the more irregular, because each end of its arc sweeps towards the metastylid and the entostylid respectively. ‘The central pit of the anterior pillar broadens out anteriorly. The enamel surfaces of the two pillars sweep in towards each other at a sharper angle, so that they meet in a deep wedge between the two pillars. Once again the protoconid and hypoconid form angular apices on the buccal side; the apex of the posterior pillar being slightly more rounded. On the occlusal surface of the posterior pillar, the dentine has a shallow pit on the hypoconid. This is probably a post-fossilization artifact. Looking from the anterior aspect, the crown-root junction presents a horizontal line from the lingual surface towards the buccal, and just near the buccal surface the enamel projects down sharply like an apron. This is the best example of this ‘apron’ effect which is, in general, better displayed in the lower than in the upper teeth. Roots. The posterior root is broad when viewed from the posterior aspect and towards the tip it curves posteriorly. ‘The anterior root has a similar appearance, but the tip is shorter and less curved. On both roots, the buccal and lingual surfaces bulge more than the central portion of the root. At the crown-root junction, between the two pillars, an accessory rootlet projects down from the anterior pillar for a short distance. The tip is broken. 4029 (plate 49) Part of the horizontal ramus of a left mandible with M,, M, (4029 B) and part of the root of M,. The teeth are in advanced wear. 484 ANNALS OF THE SOUTH AFRICAN MUSEUM M; Almost complete, deeply stained by ferricrete (see Introduction to Chapter 4). Buccal surface. It is fairly rugose with a markedly rounded cingulum. The ectoconid is marked and pillariform although broken off just below the occlusal level. Just above the cingulum each pillar is rather rounded. Because of the bulge, that part of the buccal surface above it tends to slope lingually, whereas the lingual surface appears to be at about the same angle as in 4028. The talonid appears to be angulated to the A—P plane of the pillars, but less than in 4028. It has a rounded appearance because of a large bulge of the crown above the cingulum. Lingual surface. Rather broken. The cingulum is not rounded and not as pronounced as on the buccal side. ‘The surface presents similar features to 4028. Occlusal surface. The central pit is very similar to 4028. Despite a greater degree of wear, the central pits of the anterior and posterior pillars are not confluent. The ridges of enamel of each pit lie in contiguity. The buccal enamel of the pillars is arc-shaped rather than triangular. Roots. Appear to be the same as in 4028. M, Incomplete and rather cracked, but most of the features can be recognized. The general appearance is similar to M, of 4028. Buccal surface. Marked, rounded cingulum with a bulge above it on each pillar. There is an obvious ectostylid. | Lingual surface. The crown-root junction between the pillars appears to be more angulated upwards than in 4028. Occlusal surface. The central pit of the anterior pillar is small and asym- metrically placed towards the anterior end of the tooth; the central pit of the posterior pillar is very compressed in its anterior half; the posterior half presents a small triangular pit, and is very similar to the same region in M, of 4028. Anteriorly the pit does not extend beyond the junction of the anterior and posterior pillars, as it does in 4028. These differences however are probably due to wear. The general shape in occlusal view is the same as in 4028. Roots. Viewed from the front, the apron effect of the crown-root junction is not as marked as in 4028. M, The posterior root is present. It is broad from side to side, and flattened antero-posteriorly. Measurements of mandible (mm.) Breadth opposite posterior pillar of M, .. c. 37 Breadth opposite posterior pillar of M, .. c¢. 47 4030 This is a fragment of a right Mg, the roots of which are completely absent. Most of the posterior pillar has been reconstructed, while only a small piece MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 485 of the buccal surface of the anterior pillar remains. Although the tooth was found in a number of fragments, these have been rejoined at only the obvious contact surfaces. Because of the angulation of the fragment, the central pit could be wider which would give the tooth an increased breadth, but it was decided to replace the fragments at their minimum breadth. Although the tooth is in fairly advanced wear, it is hypsodont. From a rounded cingulum the buccal surface slopes medially at a fairly acute angle, and just near the occlusal edge the surface tends to be more vertical. ‘The lingual surface slopes upwards and laterally from the base, so that the width of the tooth at the crown-root junction is far greater than at the occlusal edge. Buccal surface. ‘The enamel is rather rugose and the buccal surface is irregular. The posterior surface shows a marked indentation at the crown- root junction. There is a small ectostylid present leading from the rather rounded cingulum of the anterior pillar. There is a small bulge above the cingulum of the anterior pillar. Lingual surface. Only a small fragment of the entoconid remains, but it can be seen from this that the central costa is flattened; there is a slight groove separating the upper portion of the costa from the fragmented remains of the entostylid. Occlusal surface. A small portion of the central pit remains. It can be ascertained that the occlusal edge of the buccal surface has a broad U-shape. 4031 This is a fragmented portion of an isolated right M, in which the talonid is fairly complete, but only a portion of the posterior pillar remains. Most of the posterior root is present. It is most probable that this tooth and 4030, found near it, belong to the same jaw. The tooth appears to have been in fairly advanced wear. It is markedly hypsodont. Buccal surface. ‘The talonid is very large, presenting a marked cingulum above which the crown surface is not regular. There is a broad flattened bulge above the cingulum and an indentation (convexity towards the lingual surface) just above this at about two-thirds of the distance up from the crown-root junction. The enamel is rather rugose. The posterior half of the buccal surface of the hypoconid is present and has a round cingulum; the surface of the pillar has two indentations, one just above the cingulum and one at about two-thirds of the height from the cingulum. These abnormalities of the surface of the tooth appear to be a purely local phenomenon (? ecological or dietary variation), because anomalies of the enamel are also observed on Hopefield 4028 and 4029. Lingual surface. ‘The cingulum is represented only by a thin ridge above which there is a bulge on the talonid, and although most of the lingual surface of the entoconid is missing, sufficient remains to indicate that there is a rounded bulge above the cingulum similar to that in 4028 A. ANNALS OF THE SOUTH AFRICAN MUSEUM 486 : : 13s 9:06 «&&% £.0¢ 0-31 v.16 I.QI V.c& : g-a& : P.G1 0-11 ¥.62 S.6z Son w : : V.tS V.1% figtre Lae O-11 OLS 6.71 L.gz : L.9z * I-VI Q-a1 9.92 8-92 “6G W . : F.g1 a-VI ¢.6z : 0-GI Qe 9 : : . 6z °9 . P19 gi ‘9 €.9% 9-92 + 266 Ty W G.z1 £.01 : : : ; ; ; ; : g-g2 = L.6z : . ; OSE. Gua: =: bos WH O-I1 9.6 5 0 . . : : 6 c ; C.o§ . . - 0.92 0.9% <= ce W 61 “9 Cin oi e . . . ° e ° e 0.62 . ° . . 6z “9 nee bry d Gore ole c "> Gr ge t= iiss ee C.€1 5-8 ; : : : ; : : : 6.41 6.6% : 6.91 6.91 s CHES zd Linady ‘g . . : : c.PI o.G1 s ; : - : 9-L1 er I.OI C.6 . o.b§ -- ob Ww P Wd z.O1 9-6 . . . e ° ° . . ° L.S1 . . . 6 6.32 656 W 9-8 0.6 : : : : P.a1 : ie . -- ofS € Wd €.9 v.6 : : . : : : ° : . v.9 : a-GI L.Gr ' 666 Ww 1-6 6.6 . e e e ° . . . e 9:8 . . . C.F1 . obS W ‘Wa . : C.L1 o-GI 6-91 ¥.GS1 L.6 Q-¥z = g- G1 L.va : C.9z c-S1 0-01 tlc” Sits SES Ww . ; L.gi 0.91 P.gt I-91 O-II 6.3% o-VI 6.0% : o-Gs : €.g1 C.a1 PEs SVE '" GEC Ww . : C.gI P.G1 o-GI §.V1 6.31 0.b% P.G1 L.Ve L.Ve : C.a1 9-01 0-93 1-Qo ’ SIIIW : : o-GI €.01 C.1 O-1 V.St 0-&% €.G1 1.6% 0.0% : Qg-I1 Sil P.bz C.Fz E92 W ; ; ie ; 5.0% 0.9% 0.9% 98S : : : : O-II Q-O1 : : : ; : L.0% ; : 0-35 0.8% Egz W Wd vEr Ogt ° : = 6st Liz -* Hey Q:31 0.01 : : : : ; : : : : G.S1 3 : : ; B:16 oro W eWd NW V poomg jonsury joomng jpondsury jo2ng jonowT "7990 “xv 9Q) «XD P99Q. xD VSN}29Q-VSN]29Q.-PSN}99IQ.—*]99Q. “KDI WY SIOFT Tel[td Tel td Tel|td Tell Yoo], pruojey, 4d Fellid Yoo L NAWIOadS {00 ], IOLIIJSOg IOIIO}UV IOLIN}SOg IOLIOJ UY IOTIO}SOg IOLIOJUY LHOIAY HLGVaIUg HLONGY 487 MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA z-6 8°8 ‘PPyedoy woy st yorym SHEE ydooxa 3eSsuede *(urur) vffrury TIssoy ueoLITYy Jo a ee 0.43 €.V1 1.13 9-8 I-II 9-4 6.g C.E% 0-o1 3 0-01 L.az : Ge .g1 9-6 §.0§ 9-6z : Vs °9 FW Woy ore suswutoads 94} [Tv 4199} JO syusuaInseayy *6 aTav J, 9-61 o-L1 oL1 1 °9 (bie) o.GI O-d1 8&6 W "Ww O60 9&6 W ob6-111 W "Ww ee 96S W sW “ o0SS Ww Ser} 92S W 2WV . . 9-16 gL : 8:6 . . ese Gn 2 : G.E% : ; 00S $9 oLb 9g.0P . . : g-11 1-3 = Q- G1 , : 61°) 0.11 £.g1 8-8 : . ; ort : : : ‘ : 9-9 ; ‘ . . P.LE 0.0% €.a& 1-€3 . : Sb -9 0&6 9 : 0.6% s . . 3-QI C.6 6.91 C.11 2) fe . . . . : 0-11 g : : 6.26 0.06 : C.33 . . F.9 L.L : : Z 3 . . ° . Lo 7) ° o-§1 o : : ; : a 6.65 G.cz : : : : qi BH @) 2 : : gz 9 0. La : 5-QG & i I-QI b.g1 ie) = : G:6GesOSe9= 106-9 Ts 79 a : gob 646 2&6 9.62 S : : 0-66 9-16 6.06 §.1% < . . c06 6.16 4.96 &.bz joomg jondury joomg jonsury jyosng jonsurT USI qel Td JETT 100 T, IOII9}SOg IOIINUY & he D1 a Yi + Lob —~L-L0 0-g3 9-6& Il ‘2 0.9P o& 9 ~=—s«1.0V O16 3&.&P Ga9—- aLv 3L1E 9.LP ¢.LE ~—g.0G 1 ath ee By 2a) 0-61 0.b% L.G1 0.96 ¢.91 0-68 gS F.aE 6.91 C.aE 1990 “XDIV TeTTId IOTI9}SOq gb 9 ~— 6.1 VE -9 =0.8P 6.93 G.oG £66 699 aL&E o.9F 3-96 bho g-1b = LLP 6.S6% obs 6.01 9-02 o-g1 68 °9 ag1 = L.g& o-GI 8-Qa 0-03 §8G.1§ 199Q = “XDI JETT JOI9]UY HLGVauUg OG °2 C.% G.Lz 9-06 0-08 L.3% C.Lz 9-91 V.1G L.61 9-02 V.3& 9-bz Qs °9 0.Gz 0-16 3-SG Go °9 0@ °9 ¢.61 oe : 0% °9 6% °9 Nits, : C.z& 9-F1 L.Lt G26 95 S659 6.3% 9-gI g-&z v.61 €.0% 9-81 I-91 9.3% 725N]9IQ. 12SN]IIO 7252]9IO pruoye Ud Jel JOMI4sOg IOMo]Uy HLONGY 099. 09 9. -** gsr 0-66 0-&S a vzov 1.67 0.16 II Sq 601 tho oer °* Tray €.6P C.67 -* GCoS ¥ eW ° Sie 927 y : “* ggzv DO 6h=o 67 =3-— — Ezov . QP ‘9 - €662 4 €.9P &.9V CGoS 4 2 : ; * qagztD CP -9 : Lzov : G9 = = Gguam IW : "+ (¢) 68 ¥ VE-9 GS-9 -** 69604 6§ °9 : CzoV ed Linay “¢ GG--9 -GGi9 gq 665 -w "Wa o-€& 1-66 > eS Ww "Wa Veo. - = “reo py ee eG LE6 W L.ob = L.g¥ eae: | +N P.sP L.sv . beS W 9-0b 0.17 Bae sWd WI “V 199Q *XDIN q200 J, N@WIOddS 489 MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA +€.5 GG °9 C.1S 1-0G g-Lz G.ga €.L1 LS o-G1 rat 20) 9-91 oa 9 +1P 1.9% 0:0& Cz '9 lz -9 0.Va [alt C.3% Iv ‘9 9-16 Me So) +3.98 6.0% 9-96 9-3 S-GG 6.9% Coat CE +2 L.&& gv °9 0-38 z-61 V.Ss SE Mats L.gt V.0% 3-O1 1-03 VE °9 9-Sz ie Tou ee Go& ** TWUNSNT o- oe 16 ANNALS OF THE SOUTH AFRICAN MUSEUM 490 nllke L.o§ V.gz V.Gs 0.GZ 3-QS 9-0& Tels 0-1% v.66 4100 T, "T1290 [12L °T 10 : O18 joomg jonsury joong jondury jvong jondurT SIO yoo], gv 9 : o-Q1 9.6z 6-18 V.96 0-9% Qo °9 Gel 6§ -9 6.62 2.96 C.o1 O L.Gz ; o& 9 =... 6 Q-VZ 0.9% CSo-9 : 0.18 6c °9 ° oS ‘9 o-.SP : +6P : tA lilral JOTIO SOG HO. yy ‘peoy iqesieypy fOaI S01 “mMuMeN s16 ‘99905 q ia 516000 4 if SS f1 ‘6019 £SQ906 SININ ‘ager Sp92z75 "6h6 We OSqgSSS Ww fg@6SS Ww SWESS WS q6ES WH “10h {ooh ‘g6zoh ‘6zob Syezob ‘gzob ‘bzob ‘Szob {Lzob Sea QI ‘9 ¢.Gz Qo ‘9 2 4G eo ae 6z0P Or 2 1-VG 0-62 oho LL 9-4 =seeer 1-L1 6-3S 0-Gz G9 ‘9 = 3-89 ae 1-0G V.1G 0-62 9:Sg 14.99 near as) 3 é C.2z 5 ; Se rmunsSN L1G C.bz Qz °9 69°. bhi) Sse €.31 G.1z 0.Gz 9-95 0.66 = Strece 0-02 V-1% v.Gz 0-89 3-99 -- Sof ¥ "NW ; 1-3S L.Lo C.oG &.0S 3 9gI : g-&% ¢.Gz 6S374 os? Se -- G6 : ; ; : : ae Sul ; 6a 9 : : ’ : oor . . . 8-09 . a. OVI ry : C.bz , : : -* €b6 WW josnjz9Q).-1SN}99Q-yVSNNIIQ =—*}99Q. “XO puoey, seytg Ill y100 J, NAWIOTdS JOWa}sOg IO1M2}Uuy HLONGY MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA A4QI Occlusal surface. ‘The occlusal surface is extremely broken but it can be seen that the talonid is angulated slightly to the longitudinal axis of the tooth in a lateral direction. The talonid is large and oval. 4030 A These are a number of small fragments found in the vicinity of 4030 and 4031, which cannot be included in their structure. They consist of the tips of two lower molar roots and many small fragments of cones. SECTION III GENERAL DISCUSSION AND CONCLUSIONS SUB-ORDER Ruminantia FamiLy Giraffidae SUBFAMILY Sivatheriinae Genus SIVATHERIUM Falconer & Cautley 1835-6 (Syn. InDRATHERIUM Pilgrim, G. E., 1910. Rec. Geol. Surv. India, 40, pt. 1, 63) Stvatherium olduvaiense Hopwood. 1934* * Hopwood’s specific name is retained, but the masculine form olduvaiensis is here corrected to the neuter olduvaiense, a form which Hopwood used in 1936 but dropped in 1937. 1934 Helladotherium olduvaiensis Hopwood, Ann. G Mag. Nat. Hist. (10) 17, 546. 1936 Sivatherium olduvaiense Hopwood, Ann. G Mag. Nat. Hist. (10) 17, 636. 1937 Sivatherium olduvaiensis Hopwood. Dietrich, W. O., Wissensch. Ergebnisse der Oldoway Exped. 1913. Dr. H. Reck, edit., p. 106. 1942 Sivatherium olduvaiense Hopwood. Dietrich, W. O., Palaeontographica, 94, A: 43. 1947 Griquatherium cingulatum Haughton. Cooke, H. B. S., & Wells, L. H., S. Afr. F. Scz., 43, 232. 1948 Sivatherium olduvaiense Hopwood. Arambourg, C., Mission Scientifique de l’Omo (1932-3). T. 1, Fasc 3, Paléontologie, Paris, Mus. d’Hist. Nat., p. 376. 1948 Libytherium maurusium Pomel. Arambourg, C., C.R.S. Soc. Géol. France, Paris, p. 178. 1953 Sivatherium sp. Dreyer, T. F., Res. Nas. Mus., Bloemfontein, J, pt. 3, p. 74. Sivatherium olduvaiense vanhoepeni subsp. nov. 1932 Orangiatherium vanrhyni van Hoepen, Pal. Nav. Nas. Mus., 2, pt. 5, 63. Sivatherium olduvaiense haughtoni subsp. nov. 1949 Griquatherium haughtoni Cooke, Geol. Surv. Mem. No. 35, pt. 3, 58. Sivatherium olduvaiense subsp. indet. Two milk molars from ? Cornelia ? Florisbad. See this paper, p. 526. Sivatherium cingulatum (Haughton) 1922 Griquatherium cingulatum Haughton, Trans. Geol. Soc. S. Afr., 24, 11. Genus LinyTHERIUM Pomel 1892 1892 Libytherium maurusium Pomel, C.R. Acad. Sci. (Paris), 115, 100. 492 ANNALS OF THE SOUTH AFRICAN MUSEUM CHAPTER I SUMMARY OF OBSERVATIONS ON AND VARIATIONS IN SIVATHERINES A. CRANIAL FRAGMENTS (1) SKULL The only fragment of a skull known from Africa is specimen 4372 B from Hopefield. ‘The predominant features are the great circular size of the foramen magnum, the absence of the supraoccipital bulge and the orientation of the occipital condyles (see page 475). (2) HORN-CORES There are two pairs of horns, the major features of which are: (a) Anterior horn-cores. There are two African specimens from which conclusions may be drawn, namely, at Hopefield and at Tierfontein (O.F.S.). The Hopefield specimen is incomplete, but the available fragment corresponds closely to the corresponding region of the Tierfontein specimen. The con- clusions are based on the pooling of observations on both specimens. The horn is short, with cranial sinuses projecting slightly into the base. The horn is flattened from side to side, presenting two surfaces and two borders. The anterior border is rounded and has a peculiar cauliflower-like prominence superiorly. The posterior border is also rounded medio-laterally, but from above downwards it presents a concavity facing posteriorly. The one surface is particularly marked by deep grooves, especially near the front. As will be demonstrated in the discussion below, the Hopefield and Tierfontein material are to be referred to S. olduvaiense. The Hopefield specimen 4373 A becomes the paratype, because the authors identified it as an anterior horn-core before studying the Tierfontein specimens. Van Hoepen (1932) mentioned only the posterior horn-core definitively, but did not recognize the other specimen (van Hoepen’s ‘terminal fragment’) as an anterior horn-core. He also did not indicate that either specimen belonged to the giraffids. (b) Posterior horn-cores. Posterior horn-cores have been recovered from North Africa—St. Arnaud (Arambourg, 1948), Ain Hanech (near St. Arnaud) and Garet Ichkeul (Tunisia) (Arambourg, 1949)—East Africa (Olduvai Gorge), and South Africa (Tierfontein and Hopefield). The features common to all of these are: (i) Posterior horn-cores are of great length, averaging 640 mm. along the posterior border (range 560-840), and of great size, averaging 390 mm. circumference at the base (range 330-410 mm.) (table 41). (ii) The base is rounded or pear-shaped and is hollowed out by the cranial sinuses. 493 ‘asuaiwanpjo winisayjoas JO ("“WUL) $9109-UIOY JOTI0)}Ssod jo suotsuswiq, *1) a1AVvJ, MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA oS9 ue ae d-V 2seq ‘% qouyy 06 “ ie "* Yysus] ‘oe qousy (sasnurs Aq poMmoy[OoFy{) 0g ob1 00% oLt OSI oGI te "* gseq jo JYUsIOY [eUID}UT CGr Loz Gh a ce "* -yysdusy oSurpy ob eb LY oS £9 ob Gv ob ch iy Je dn yw 99 89 GZ gf &8 €S e et "* & qouy tv ol bS G9 G9 1 oh "* & qoUy dAOqe "UTUT COI LL 1g 64 6g +6 GL oL 6£ 28 % qouy pur osury uwomiog gS "* gsury jo wed jsoysiy ry oot 06 Lot Pr Ol g6 Lg Lot ao Mg a: "* gseg SPs a P US. 6P oG 6S aS GL gb cS ob ee as es dn iy ivr Iv Poi 6z1 of1 88 L6 ae Ce "* § qouy ty ool 101 Gol 611 Lit 26 g6 £6 16 ** 8 qouy dAOqe ‘uIUT OOI Gh gor Gir orl LE of! % qouy pure osury usomjog GgI ‘* oSuvy jo javed ysoysty tv obt wht PG gol 061 = "* gSuvy jo aseq 1yv Li1 go Ghi 6G1 OSI obi O11 GE “11 os ay BM "* oseg snounip qJ—-V oGt oG1 oL1 oL1 og! LE1 rs - se dn vy ole oSs oSs ORs oof "' puw € sqouy usomjog 09% 09% 06% ob GGz es Me "* § qouy 1Vv o1& 06% oof ob& 0Gz GLz gSz Ggz "* — % qouy dAoqe ‘uIUT OOT oG& og9& osl oP oS O0gz oof 06% GUE of &% qouy pue odury usoMjog o£& 096 o6§ oot Gib oth 066 098 oob ph ny a ‘* oseg guar funasary ogS o&b 069 obg o1L 09S 099 og4 oLS 06S ri "* gOpsIOq IOTII1SOg obg ool O10! 0S6 01g 006 ft “2 ** QAIND JOINO quaudosf fo yjsuay 700 7, Be pi id - a oe oF Ke See 5 SS fee ao a ; oD © ! ~ ; we) Oo iat ae) Seah eee Ree RS epi ee i Eo oes ieee a Lea om eae ge aig wet SE Ss Sog | OP e Z wy Be ee acto G B. a On a. A. B Bing e Bt ue & 494 ANNALS OF THE SOUTH AFRICAN MUSEUM (ii1) The body of the horn-core is hollow for a varying extent. (iv) Quite distinct from Cautley’s specimen of the “‘palmated antler’ of S. giganteum (Falconer, 1868, I, pl. 21, fig. 3),* the African specimens generally are rounded or oval-shaped. The African specimens present a narrow convex anterior border and a rounded concave posterior border which also has a spiral twist, so that the anterior border becomes superior near the tip and the posterior becomes inferior, i.e. anti-clockwise (right horn) or heteronym (following the terminology of Lydekker, 1913). The horns are characterized by a flange-like projection (which varies in extent) of the anterior border just above the base, and knob-like projections (usually at least three) of varying size along the anterior border above the flange. It has been clearly demonstrated in Section 1, chapter 7, that in the modern Giraffa camelopardalis there are tremendous sex and individual variations within the species. It has been indicated that the previous sub-specific ecological criteria and horn variations are not valid. There can be no doubt that the dentitions of the Sivathertum specimens presently found at Olduvai belong to a single species, olduvaiense, in common with those from Hopefield, Port Allan (Tierfontein), Makapans- gat, Omo, and Garet Ichkeul. The horn-cores assigned to this species from these sites show a range of variation as wide as that in the modern giraffe. There are degrees of curvature in an antero-posterior direction and also of twisting or torsion. The torsion varies from nothing to almost 90°. This enormous range is found at Olduvai where most of the specimens are curved in an A-P direction with only a slight or medium degree of lateral twist, while a few are twisted markedly so that the planes of the surfaces alter and the horn develops an ‘antler’ effect. The twist usually occurs just at or above the flange. Unfortunately there are no associations with these horns to indicate the sex of the animals; possibly the males, as in the modern giraffe, show the most marked. variations. The few Stvatherium giganteum horn fragments found may well be at this extreme range of (? male) variation of torsion, giving a fairly constant ‘antler-like’ appearance, as described previously. There are also variations in the number, position and size of the knobs. In some specimens they are evenly distributed along the anterior border. In the horns with marked torsion one or more of the lower knobs get shifted towards the back. ‘The posterior ‘branch’ or knob on Sivatherium giganteum may well again be an extreme example of this. In other specimens the knobs are unevenly distributed, and in a few they tend to be clustered close together near the base of the horn, thus leaving a long smooth anterior border leading up to the tip. * However, fragment AMNH 19774 (from the Siwaliks), referred by Colbert (1935) to ? Sivatherium, is distinctly rounded and is characterized by features observed in the African specimens. 4 MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 495 There is no sex dimorphism in the dentition of the modern Giraffa camelopardalis. ‘The teeth of Stvatherium olduvaiense from each site tend to fall within a single wide range of variation. Based upon the above discussion, it is unreasonable to separate the horn-cores at Olduvai into two species merely because of the fact that some are more markedly twisted than others. The small differences between individuals produce a gradation where the specimens at each end of the range appear to differ fairly widely from each other. But the same, but not correlated, difference can be obtained for the size and the shape of knobs and the A-P curve. Arambourg (1948) states that Lzbythertum horn-cores differ from other Sivatheriinae, and in particular from the genus Sivathertum, because they are not so branched. He refers to the number and size of the knobs, and the torsion of the horns. However, the question of the ‘branching’ has been discussed above and it, as observed in the African specimens, does not provide a generic or specific criterion of distinction. Furthermore, the North African (St. Arnaud) horn-cores* do not differ essentially from those of the other African Sivatheriinae. In actual fact, it will be demonstrated later in the discussion that for many reasons (one feature being the horn-cores), part of the Libytherium maurusium material must be referred to Sivatherium. (v) On the antero-medial convex surface} characteristic and more or less parallel grooves of varying length and depth sweep upwards from the anterior part of the base. The significance of the grooves is not clear. Because of the fact that they appear to radiate from a single point of origin at the base, one gains the impression that they may be of vascular origin (one could compare these with the appearance of the middle meningeal vessels on the interior of a skull). This view had previously been accepted (Abel, 1904; Colbert, 1935; Arambourg, 1948). However, first, most of the grooves do not decrease in size propor- tionately to their distance from the origin, which one would expect if they were vessels. Secondly, one could surmise that a massive horn-core would require a large blood supply, but this argument falls away when one considers that the small anterior horn-cores (about one-quarter of the length and * These have been studied at the Muséum National d’Histoire Naturelle, Paris, by kind permission of Professors Arambourg and Lehman. The descriptions are not included in this paper as they are to be given by Professor Arambourg in a paper now in preparation. However, he has kindly permitted us to publish the measurements and their general appearance. + It is extremely difficult to determine to which side a horn-core belongs because of the absence of cranial attachment. However, following careful consideration of the Olduvai specimens 1.53 and 2.53 and of the S. giganteum specimens (Falconer & Cautley, 1846-49; Falconer, 1868; Abel, 1904; Colbert, 1935), it has been decided by the authors that the posterior horns are probably directed laterally and backwards, and set at an angle of about 45° to a vertical and horizontal plane, when looked at from the front. In this way, the surface of the horn-core exhibiting the marked grooves becomes the antero-medial surface. 496 ANNALS OF THE SOUTH AFRICAN MUSEUM less than one-tenth of the mass of the posterior horn-cores) present equally large or larger grooves. Furthermore, along the line of these grooves, one does not find particularly significant nutrient foramina. In fact the foramina are very small and are found scattered over the whole surface of the horn-core, without any specific relationship to the grooves. Thirdly, it would appear to be an unnecessary hazard for Nature to have exposed large vessels of great length in such an unprotected region as these curved horns. The only apparent alternatives are ligamentous or cartilaginous attachments, and on the basis of the large grooves found in other bovids such as Bubalus (Homotoceras) where cartilage is known to surround the horn-core, it would appear reasonable to assume that these grooves are mainly for cartilaginous attachment, and possibly some of the smaller grooves (those posteriorly) may be for blood-vessels. This is contrary to Colbert’s belief (1935) that the Giraffidae never developed a horn sheath during their evolutionary history, and that the hairy horns of the modern giraffe denote a primitive character that typified the various fossil forms. (3) MANDIBLE Making use of Makapansgat specimens M 553A, M 553B and M 553B?, of which A and B? belong to the left side, and B to the right side of the same mandible (see Section II, chapter 3), the left half of the mandible was recon- structed by using the mirror image of M 553B. This resulted in a rather robust mandible (fig. 22) which was compared with the Hopefield fragment 4029, the Olduvai specimens 1, 6, 91, 92 and 392, the type specimen of Libytherium maurusium (Pomel, 1892) and the North African specimen 1950—I-I (from Garet Ichkeul, Tunisia; in the Muséum d’Histoire Naturelle, Paris). These were further compared with the juvenile mandible M 539B from Makapansgat, and with data on modern giraffes. From this study the following observations and inferences were made: (a) The total length of the jaw from the anterior alveolar border at symphysis menti to the posterior border of the alveolar socket of M, is not much longer in Sivatherines than in modern giraffes (5-15 % difference). (b) However, it is in the relative proportions of various significant regions of the mandible that Sivatheriinae differ from G. camelopardalis. (i) The modern giraffe shows an absolute reduction of the length of the premolar-molar series (fig. 24) (for reduction of individual tooth sizes, see “C. Dentition’, infra). | The length of the premolar-molar series in proportion to the length of the jaw (as determined above) ranges from approximately 60 per cent in the fossil Sivatherines to 40 per cent in the modern giraffe. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 497 (ii) The distance from the anterior border of P, to the posterior border of the symphysis menti is greater in the fossil Sivatherines than in the extant group: SIVATHERIINAE 51.5x48.7 476 x 49.7 41.5x 50.9 467x297 >390x 44. 410 x 329 UPPER DM 4 DM 3 LOWER 331x179 §5.0x 240 686x 336 512 x 35.7 458x336- 426x304 354x259 260x182 M 3 M 2 M 1 P 4 P3 P 2 Fic. 24. Graphic representation of the relative ranges of dimensions (A—P and breadth) of adult and milk dentitions in the African Sivatheriinae (excluding F 39, MMK 3685). Mean dimensions are indicated. Comparable ranges of the modern Giraffa camelopardalis are superimposed on those of the Sivatheriinae. mm. Percentage of total length 1950-1-1 (North Africa) c. 146 32% Libytherium maurusium type ¢. 100 M 553A €. 103 G. camelopardalis eaeks: LO (iii) The distance between the anterior border of P, and the posterior border of the foramen mentale is much greater in the modern giraffe: mm. M 553A 99 Libytherium maurusium Pomel—type specimen 100 G. camelopardalis 145 498 (iv) ANNALS OF THE SOUTH AFRICAN MUSEUM Comparing the data of (ii) with that of (iii) it is clear that foramen mentale ‘s considerably anterior in relation to the position of the posterior border of the symphysis menti in the modern giraffe. In the Sivatherines, the foramen mentale is opposite the posterior border of the symphysis. This is confirmed by direct measurements: in Pomel’s specimen of Libytherium, they are opposite each other; in Makapansgat 553A, the posterior border of the foramen is c. 9 mm. posterior to symphysis menti, while in G. camelopardalis this distance is c. 67 mm. In the modern giraffe, the symphysis menti is considerably longer than that reconstructed in the Makapansgat specimen 553A when the measure- ment is taken from the position of the tip of the roots of the incisors. The measurements are 152 and approximately 100 mm. respectively. Following on observation (iv), if, in 553A, one considers the combination of the thickness of the symphysis (41 mm. at the anterior broken extremity ; compared with that of G. camelopardalis at the same point, viz. c. 15 mm.), the shortness of the symphysis and the height of the body of the mandible (c. 65 mm.) at the level of the root of I,, it would seem that the anterior canine-incisor alveolar arch must have been rather cup-shaped in the Makapansgat specimen, so that the teeth would have been embedded more vertically in the jaw than in the modern giraffe. ‘This conclusion is supported by the appearance of the same region in Hydaspitherium sp. (A.M.N.H. 19684) (plate 50(e)), although it is unlikely that the Makapansgat specimen’s teeth were as vertically embedded in the bony alveolus as those of HAydaspitherium. The Makapansgat Sivatherium canine-incisor row was probably in an intermediate position. Further- more as the canine has not yet appeared and the canine-incisor arch appears to have ‘no space’ for it, considerable lateral expansion and increase in depth of the arch would have to take place to accommodate all the teeth. This specimen also indicates that the canine is the last tooth to appear in the dentition, as in the modern giraffe. (See also ‘Appendix’.) (c) Certain features concerning the breadth and height of the Sivatherine mandibles are also important: (i) Breadth. The maximal breadth is found in the region between the . a anterior pillar of M, and the anterior pillar of M,. Anterior to this region, the breadth decreases in a regular progressive fashion. In the juvenile mandible fragment M 539B, the maximal breadth is found opposite the ‘talonid’ of DM,. This corresponds to the region opposite P, in the adult. If one now compares this with the breadth at this point in the adult, then it appears that a growth increase of about 30 per cent occurs between the two phases. : Height. The maximal height is (as expected) opposite the talonid of Ms, and anterior to this there is a progressive decrease towards the anterior (ii) MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 499 extremity of the jaw. Once again, if one compares the height in the juvenile at a point analogous to that in an adult specimen, then it appears that approximately a 100 per cent increase in growth occurs between the two phases. The ratio of height to breadth is expressed as the index of robusticity. This is maximal in Sivatheriinae in the region of M,—P,, but the range varies tremendously between different specimens (e.g. 48:3 in Libytherium (=Stvatherium) 1950-1-1; 69:1 in Olduvai 91). B. PostcRANIAL FRAGMENTS Postcranial fossil giraffid remains from Africa are very scarce. All the material studied by the authors comes from Olduvai.* Further specimens from various sites have been described by: (2) (4) (d) Stromer (1907): a portion of the proximal extremity of a femur, a meta- carpal, a fragment of calcaneum showing the fibular facet, and one phalanx: from Wadi Natrun, Egypt. Dietrich (1937): a metacarpal E. 122, from Olduvai gorge; and (1942): from Garussi-Korongo, a metacarpal numbered 1.29, a tibia fragment, a calcaneum, a few first phalanges, astragali numbered Vo 330.1 and Gar.K. 1.29, and 3 cubonaviculare numbered Vo 330. Arambourg (1947): a distal fragment of a scapula from Omo, Ethiopia. Bate (1951): a distal end of a radius and a distal portion of a humerus, from Abu Hugar, Sudan. The NON-METRICAL features of the Olduvai specimens which are distinctly different from those of extant giraffes, are: and the first: in the cubonavicular, the posterior articular facet which articulates with the metatarsal is absent; secondly, in the metatarsal the vertical anterior groove is much deeper and more scooped out; thirdly, Arambourg (1947) noted that the spine of the scapula was attached to its dorsum nearer the glenoid cavity; fourthly, Bate (1951) stated that the breadth of the shaft of humerus and radius relative to the distal end of those bones was much greater in Sivatherines than in the living giraffes. Comparisons of the MEASUREMENTS of various bones available for study, of those in the literature, produce a number of interesting features: 1. Scapula. There is an overall increase in the size of the scapula (table 42), maximal breadth at the base showing the most marked differences from * See also Appendix. 500 ANNALS OF THE SOUTH AFRICAN MUSEUM the modern Giraffa. There is also a difference in the distance between the root of the spine of the scapula and the glenoid fossa (which has already been noted as a non-metrical feature). It may be noted that the data for S. giganteum is very approximate to that of S. olduvaiense from Omo. 2 K S Giraffa' G. camelopardalis? Sous ee Measurements! 20 S eee Se SER Mn Er <3 § 33 2° 5 3 SO) ie 2 fi nae SG 2G oS cP 6S LS LY d- ro) V a Yoys-PIV = PGI GE gol Vol g&1 EVI eV Yi pesiq WMNUITxeyAy B CJIULaLIXI JVUIXOLG fy 2) a ip eae So Seen Seo sees oO 4 = nx a nN Bs re od — SES 5. S a: < x NI =) iS) © ee) es = S : Z 2) re & SS Se Z = alee nee: S. M4 < A as aS yy a 9 8 SIR S. R 2 os S So SS) (5) a syppavgojauva vffvsr5) 5 502 5°38 MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA "uaINnAIay, ‘asjag Osuor) np pedoy agsny (9) ‘sumnasnyy “W'S'f) Woy viep inc (S) ‘Joquinu sures oY} YIM spedreovjour yo] OM} oI OQhEI FY :910N "gg ‘Arpnesyy (2) "GE61 “a0qqor (1) *Lo61 ‘1aut0Ng (PF) eb61 ‘yororq (£) ‘Tedivowoy ‘oF alavy, SS nnn nn 999 4S9 trL ooL LGL 6S9 o9& ghE O62 fom LS “fg7 6g) GL wyas oy 9V 6Y Vis ov, or OY £6 6g. ZOl OV, OF. 995 96 co “ZY oF GVee Qu. 865 0og6 6S —gr SS ob Leb 966 o&P GGg =: ets Be Tedivovjoul Jo yySuz] [20 7, VGie “She SaTApuoo jo y18ua], q-V Oches: os DovpINS IJYNINIE JOLazUI “YpLoI g& 6S G&G eS GG -- (Ajrorsajue) sayApuod yerpour sapere Cann ee 1G "* (Ajsorsa}ue) safApuoo [eso}e] ‘yypeoiq WINUITXY A] 611 III gir oor 96 g6 oor Sor = "+ so[Apuoo ssosoe yIpoiq UINUTXLI fe es Mquuasxa 104StCT 6L Yppesiq :Aj1UI8I}x9 [eIsSIp 2Aoqe “wIUI Og NoGe 1eYS ol gS 19 gS ot ae 26 tineouy 1foys-pr 69 d-V s OZ! a ypeoiq : 611 O11 601 96 £01 c6 ie ue a ace Aqguasyxa 1DULXOL | je) aN . a $F 8 2 SOE ~ SR PE OL EE LEE LESSEE O PE as nS ns yn TG PS SY S «8 FS 5 = Sy SS Ss & Zs “I N o oan ioe = S ops ¢S § VS Ws Bs ee ee x Oo >~ oa Z = & NQF BFA Se sss Shee ee Si Dp (op) wa DO V4 > a. eh Mos SS Sl Sores S Ss 3 3.0 = OO wm wm wm lo) (= & § wS8S AS “SF 8 S5 Sees Sas oe 2@ S- SS BS wa SSO S86 S856 aS eis 3 Oe 2a sale Use SS Go 29 Be 8 ee ete ees deg lies z : Ss 5 eS s = S = 8 S as 3 5 =F > Se cus — — = x g o" S° - - - Sy s aes ~~ ~~ ~— 5 syppsngojauips offs) 3 ewe s CAN MUSEUM ANNALS OF THE SOUTH AFRI 504 ‘udINAIOT, ‘asjagq OSuO'D np [edoyYy a9sn7;q AP) Cov 66S 06S 66S oGI &h1 Lv1 Gor 50% 602 ae) c6 c6 08 <8 if) av 6S cG oy 09 Lo ol gf oS avi PGI ILI 67 9S LG rb 19 Lo of 1 £9 G9 v6 19 1 €Z 19 v6 CE 691 D oN rs) Qn nS ns mo of >a i) © (ec) O° I -_— I= Ca n < No iS eS = Se = suppsvgojauns offi) s6P (1) 611 £61 LbS6E Wa wnaqupsis wnisayjoa1gy men "W'S 1% poqda]JOo wIVp INGE (1) “INUIT Jao alav J, oG1 QZ Foi tL zg wnajuvsis wnisayjoary . . . ee . INUII} JO YISUZ] [ZO T, Aj[euszoyxo A][eus9}UI :1o39UIeIp 1O01190}sod-o13}uy 913U99 oy} UI sJApUOD IeTJoIed Jo YISUsT p10D afApuoos sreyjoyed ssoioe YIpesIq WNUWIxeyy a[APUOd [e972] SsOINv YIPVIIG WINUITXLI] a[APUOD [VIPeUl ssoioe YIpvIIq WNWIXe/y : *+ safApuoo ssoioe YIPRIIG UINUITXL]Y (1uasxa 10S . ee ee ee ISIDASUBLT, d-V :yfoys-pipy poy jo s0vjins JepNoNIe JO JoJQUIeIp asIOASUeIY, Pray je ooejans TepNIyIe JO Jo]9WIvIp Jo11a}sod-o19}Uy : come yaa Jo}e918 JO Ia}OUIvIp JOIIo}sod-o19}Uy ghS6& Wa 601 9L cc ol IQI 96 ol ZOI o61 Kelics) ©) a. = Sis = 3 iS iS} = S 3 Jayueyo0N Surpnpur yproig (J1WAs]Xd JDUIXOLT 595 ‘usInAIOT, ‘aspog OsuoZD np yedoy agsnyy (%) ‘sunasnyy “W°S'-), Woy IVP NO (1) *(-STET “2N]) UNIsNYY Ysilig sojOUuDp "J » ‘VIqhL, “gh alavy, MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 909 62S gro 089 €S9 oLG GzS x ee ‘+ yQBuay [210], GL GL CL £9 98 99 ol zQ £9 Som er: " q-Vy wnuirxepy vol ZOI OIl 131 QZ Lot Lot Lt 601 Os! ro Yi pesiq wWNUIxepy (Imad xa DISTT v bP 697 oS ¢ ee ee ee oe 7 at 99 69 38 ig G = a ee urpeond : pua [eIsIp aAOqe “uIW Og 7f/vy/5 PS ov G9 z9 zQ 6P et ee ‘0 q-y wnuixepy 89 £9 &L 0g 1g zQ 69 oi YIpeoiq UINUIIxepy ; : 7 Yoys-pIN 601 6 eo ee es Ey er Cer 691 ann a oe ie 0 epee Kee AJIUIALIXA [DUK OAT rs D aS Nn QMH => by es} bo bo ee © O O 1S) ice) Vad (To) : : oe) M aS S ron ees = 38 ¢ \¢ 386 82 ee fe) O > O TRAt oH + HOH Maximum proximo-distal length 113 112 130 112 124 99 93 A-P maximum length, medially 73 71 63 A-P maximum length, laterally 64 63 60 Maximum breadth, proximally.. 87 86 65 86 73 81 Maximum breadth, distally .. ¢. 75 76 72 Maximum articular breadth, proximally .. ae Mel 73 Maximum articular breadth, distally .. ive Eng anes WG 76 gI 76 TABLE 49. Astragalus. 1 Dietrich, 1942. * Colbert, 1935. * Musée Royal du Congo Belge, Tervuren. 8. Calcaneum. No difference is noted between the extant giraffe and the extinct forms (table 50). MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 507 = Ss be a = fe) 4 ~ Cr) = * = & S mS or) gS ee fe ge ee o foo) Os 8 Sa o-s oO 3 Oo (o) [e) ees Ss = o ons im S & x i = = S So = 3 fe) = 5, Skee eS Loe ee eee CMe Ml araieae ee (2. Sao Maximum length .. ae at 216 198 183 202 221 Maximum breadth .. ot ie 69 67 72 54 64 Maximum height, A—P tuberosity .. 69 =e. 62 72 62 72 Maximum length, A—P apres the fibular facet .. 89 gI Body length from the superior fodder of the astragalus facet .. Se Lot 6. L1G 119 133 Minimum body breadth .. a 39 42 31 42 A-P length of fibular facet ee its convex portion) 39 33 40 31 Breadth of fibular facet S93 Sie 28 24 25 26 Projection of heel .. bc ie 157 TABLE 50. Calcaneum. 1 Colbert, 1935. *Stromer, 1907: ? Libytherium discovered in Garet el Muluk; fibular articulation of a right calcaneum. 3 Musée Royal du Congo Belge, Tervuren. Q. Metatarsal. As in the metacarpal, there is a tremendous reduction in the length of the metatarsal as compared with G. camelopardalis. In all other aspects, this bone of the Sivatherines is similar to that of the modern giraffe (table 51, page 508). 10. Other tarsal and carpal bones. From the data available no distinct differences may be observed between Sivatheriinae and Giuraffa (tables 52, 53, 54)- Olduvai 34] Maximum length ne a a 75 Maximum breadth -. 3h Se 70 Maximum postero-lateral thickness .. 39 TABLE 52. Os magnum. ANNALS OF THE SOUTH AFRICAN MUSEUM 508 -uamare L ‘o819g osuor) np [ehoy s9snyzJ (9) ‘suInasnyyy “YW"S"(] Woy eIep no (S) ‘yedivovjopy] *1S alavy, ‘Lo61 ‘12uI0Ng (PF) “G61 “aaqtor (&) (9) 8h6F 3S 88 FTE “PIO zgg og SLL of9 zog 698 oSh i18b Eb obb Sib GSE 1S LG 09 69 . ° . . . . ° zc 6 gG—&Q te OV — OV Oa. 6 20 : : ; ; 19 6g 06 Gols 32 = Gb 69 = 69 9& +h gh LV OS S96 €¢ 16 $& Gb og 96 oF 2S —&Y Oe lS Sigs GV 16 : : : : 3 = £71 ~961 == V61 o¢ 66 ZG G9 wz €9 8 GQ LL Gc Looe Qo “Ga 6¢ of Gg ab Gee Oa Ll : : cL = Ge- 68 TOI --99 = 105 a9 — OF, 390. 06 eet Uge GL v6 : ; 99 «+6 or) - ~ OQBg2& baer Bar SoS QR 2 S$ 2 Ss geese ise iseise = <2 nm ater Ce Se. Fou Fon FS. ge oe ae HK See See esesao S| as fer) lor) lor) oFb > 3. 3.2 SS. 2 SS. 00 = RS ee aR SSR EST SSR ESR SF Se Bose se ae SO 8 We fe eh ce fe epee ae wees ae Med Se Se Le Se ® suppivdojaups vffpsixy (7) wm1ayqharT é TEE “PIO 90T “PIO “g261 “urpyog (2) “ob6r “yornjaIq (1) yjsua) 7070 7, Kaarpouns Ba[YOOI} ssoloe ySuU2] q-V Aj[219j}e] VapyooIy ssoioe YISUZT gq-V : evo[YooI} Terpoul ‘yIpRoig **VaTyooI) Tes1o}z] “yIpeoig vaTyoo1} ‘YI pvsig WMNUITXepy (qiwasjxa 10IS1q d-y Wee YIpesiq WNUIIXe/y a Giuesiie ]D1Sip anogv “wu OZT WIpeors yisus] d-V oud IOFUINIII') Ufoys-pLy TA REse oso ASies, yisus] q-V Gymoux TEE: mojag “uu 29 1foys : DOVJANS IL] SWE | ESE soayueo 9 SSOIDU Ysp¥esq VaNUETE JoRjANS IV] -Nd1}Ie TeIpour 691]UI9 sso19V epee crcaesey DOBJANS IV[NIVIV [e191] ‘d-V UMUWITXV JJ PET SNL Tk 21 ‘ad-V nae : aoejAMs IvNoIe ‘yIpeoig do¥"jins Iepnonsie “ySsus] gJ-V AJ1UL9L]X9 JDULIXOLT MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 509 ei = S) H So gery ae s Hens ais Eee Mi ap pen RE REEL) § Ly eg ry eS ee | hee 2) UP BN PS HENS. (OS OS Ela an Pi rey oN necle eete oad ee ey SAS Sachem eee 8. 295 SS e) O O po Po Poe MAH HNO +O Maximum breadth cia to side) across centre EEO) STOO, - TOO, Well, E20. T2h 7128 109 96 Maximum A-P length across tuberosity of navicular .. 106 106 95 127 98. = 98 Maximum length of navicu- lar articular facet for ~ cuneiform be 55 ag 50 ; 69 | Maximum breadth idem .. 37 37 35 38 3I Maximum length of cuboid facet for metatarsal of 61 61 55 Ole. Maximum breadth idem .. 45 48 42 46 38° TABLE 53. Cubonaviculare (Scaphocuboid). 1 Dietrich, 1942. 2 Musée Royal du Congo Belge, Tervuren. Olduvai Olduvai 105 110A Maximum A-P length .. Me 68 59 Maximum breadth (side to side) oe 41 40 Maximum postero-lateral thickness ists 23 23 TABLE 54. Cuneiform. C. DENTITION A large series of teeth provides ample opportunity for determining typical features and variations of specimens from a particular site, and also for deter- mining comparisons between dentitions from different sites. The characteristics of the Sivatheriinae teeth studied are: I. NON-METRICAL FEATURES (a) Wear. Those teeth which are either unworn or in the early stages of wear provide evidence that the teeth of the lower jaw are hypsodont, while the teeth up the upper jaw are, in general, at most mesodont, although F 39 is hypsodont. Asin all other palaeontological studies on dentitions, the appearance of the various stages of wear may be most misleading if one has not had the unworn teeth available. In the Sivatherine dentition particularly, because of the great breadth of the tooth near the crown-root junction and consequently because of the varying amount of dentine present, and because of the fusion of the cones at different stages, the diagnosis of the teeth must be approached with caution. The large collection of Sivatherine teeth assembled from various 510 ANNALS OF THE SOUTH AFRICAN MUSEUM sites made it possible to establish a range of the stages of wear of the upper and lower dentition from the unworn teeth to teeth worn right down to the crown- root junction. (b) Slope and bulge of tooth surfaces. Another important factor which requires careful consideration is that of the various degrees of slope of the buccal and lingual surfaces of the teeth from the crown-root junction towards the occlusal aspect. Furthermore, in relation to this slope, it has been observed that the profile of the enamel surface of the tooth may be either straight or it may have a rounded bulge in the region of the crown-root junction, quite apart from the cingulum. These variations are not only observed in teeth from the same site, but there may be variations in teeth of the same jaw. These features are seen particularly on the proto-hypocone (-id) enamel surface of the teeth. In Old. 1, for example, the profile of the buccal surface on the posterior pillar of M, tends to be vertical and flat, while in the talonid it has a rounded bulge. Similar rounded bulges are observed in M, and P,. Maximal bulging on the buccal surface is observed on the molars of Old. 6 and Hopefield 4029 (M,). On the other hand, Hopefield 4028 tends to show a flattened vertical buccal surface, while another variation of the flattened effect is observed in M 553 Bt where the profile is flat, but it is also sloping in a lingual direction towards the occlusal surface. The same effect is seen in Hopefield 4030. In the opposite direction, Old. 92 shows a very slight bulge above the cingulum, while most of the rest of the buccal surface is recurved so that a slight concavity faces buccally. Similar variations in the slope and bulging of the enamel on the lingual surface are seen in the upper teeth, except that, where the surface tends to be straight, it slopes in a buccal direction and is not vertical. For example, ~ Old. tog has a slight irregularity of the enamel just below the cingulum, and the lingual surface slopes buccalwards towards the occlusal surface. In Hope- field 4023 there is a slight bulge below the cingulum and the lingual surface has a slight concavity as it slopes towards the occlusal plane. In Hopefield 4024, there is a rounded bulge below the cingulum and on the posterior pillar the concavity is quite marked. The maximum degree of this is seen in MMkK 3685 where, immediately below the prominent cingulum, the lingual surface slopes at an acute angle markedly towards the occlusal surface in a buccal direction and also has a slight concavity facing towards the lingual side. Specimen F 39 (of the Vaal River site) slopes at a marked angle from the cingulum towards the apex with a slight convexity lingualwards, but just beyond half-way towards the present occlusal edge, the surface tends to become slightly more vertical. It has been mentioned that the entire interior of the tooth is completely filled with breccia, and on radiographic examination vertical fracture lines can be seen, apparently being caused and filled by compressing breccia. This could be an explanation for some of the excessive bulging of the upper part of the lingual surface of the tooth. However, there are no distinct MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 51! cracks on the surface of the enamel, except for a small one on the anterior edge of the cingulum, and for some small cracks on the occlusal surface. This explana- tion is only a tentative one, and has not been used as a diagnostic criterion for discussion. (c) Central pit. ‘The variations observed in various teeth are purely due to the different degrees of wear and should be regarded as individual variations rather than in the light of taxonomical differences. Similar shapes and variations have been observed in specimens from all the sites in Africa. (d) Enamel. The characteristic features and variations of the enamel of the Sivatherine dentition may be described as follows: (i) Rugosity As already mentioned, in the whole family of giraffids the enamel is rugose, and in the Sivatherines the rugosity is as variable as in other genera, namely from a fine pattern to a gross appearance. These variations are observed in the same specimen or in different specimens from the same site. Old. 6 for instance is finely rugose, while Old. 1 (from the same BK II locality) shows a coarse rugosity. Hopefield 4027 is finely rugose, while Hopefield 4023 (of the same jaw) is coarsely rugose. However, in the Sivatherine teeth in general, the enamel is thrown into far more numerous, closer-packed, vertical ‘ridges’ (each of which consists of a number of overlapping vertical spikes) than in G. camelo- pardalis, so that they tend to have a rougher appearance of the enamel than the modern giraffe. Even when the authors have described teeth as having a ‘fine rugosity’, the ridges—although less prominent—were still numerous and closely packed, quite distinct from the modern Giraffa, where they are fewer, shorter and separated from each other. The enamel pattern of the African fossil teeth ‘is more akin to S. giganteum than the available examples of the other Sivatherii- nae, in that in the latter the individual ‘ridges’ are less spiked. (ii) Cingulum A cingulum is practically always present on the buccal and lingual surfaces, varying from a thin, linear ridge to a rolled edge. The latter is usually found, when present, on the proto-hypocone (-id) aspect, while the former is typically found on the other aspect of the tooth. On the anterior and posterior surfaces, the cingulum is usually absent or deficient. Furthermore, there may or may not be a bulge related to the cingulum. On the other hand, just above (in lower) or just below it (in upper teeth) there may even be a concavity of the tooth surface (vide supra, (b)). (111) Styles Elevations of the cingulum are commonly present in the form of inter- pillaric styles (entostyles and ectostylids), but this varies from tooth to tooth and jaw to jaw. The recording of the presence or absence of the styles may be 5i2 ANNALS OF THE SOUTH AFRICAN MUSEUM difficult because of the late stage of wear. Indeed, the styles vary in that they may or may not reach the crown-root junction. When they do not, they are usually present near the occlusal aspect of the tooth, so that in late wear the tooth presents the appearance of not possessing that particular style. This inference is drawn from the presence of the style on the unworn teeth. The median costa is better developed on the anterior than on the posterior pillar, and in the lower teeth it seldom reaches the cingulum, occasionally ending as a rounded bulge above the cingulum. This effect is noticed in teeth from all the sites. The base of the metastylid usually commences about half-way up from the crown-root junction; consequently in late wear of the lower teeth, with the ‘absence’ of metastylid and median costa, the lingual surface of the tooth presents a rather flattened appearance. The mesostyle is always present as a persistent, well-marked, rounded ridge extending from the occlusal edge to the cingulum. The parastylid is usually ridged and slanting, being con- tinuous at its base with the cingulum. The parastyle is always rounded, prominent and vertical, with its base continuous with the cingulum where it forms a bulge, and its apex tapering off to a point beyond the midpoint of the tooth. ‘The entostylid is usually poorly represented, if present. However, the metastyle presents a fairly prominent bulge in the region of the cingulum, but it is relatively much smaller than the mesostyle or parastyle. In comparison with the modern (brachydont) G. camelopardalis, the African ‘Sivatheriinae show marked differences in respect of the relationship of the styles (-ids) to each other on the buccal (upper) and lingual (lower) aspects of the tooth. In the modern specimens, on the buccal aspect of the anterior pillars of the upper jaw, the parastyle meets the ridge-like prominent costa at an acute angle at its base. Posterior to the median costa the cingulum runs horizontal to the base of the prominent mesostyle. Posterior to the mesostyle, the buccal surface of the tooth is flattened and seldom presents a visible median costa or a metastyle. Consequently, the appearance of the buccal aspect of the upper tooth is that of an inverted trident. In the lower teeth, each pillar presents a similar appearance, the anterior overlapping the posterior, the median costa being rather rounded, not only in an A-P convexity but also in a superior-inferior convexity, so that the occlusal tip of the metaconid and entoconid tend to bend over in a buccal direction. Behind each median costa there is a slight groove separating it from its respective stylid which is continuous with it about half-way up from the crown-root junction. Below that level, the lingual surface presents a rather continuous smooth bulging appearance. In the Sivatheriinae, the styles and costae of the buccal surface of the upper teeth tend to be rather parallel to each other, the meeting of the base of the styles with the cingulum being in a broad U-shaped fashion. In the lower dentition, as mentioned above, the entostylids are poorly marked and the parastylids are prominent, which is the reverse of the situation in modern giraffes. Because the parastylid reaches the cingulum near the base of the MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 513 median costa, the lingual surface has a rather scalloped appearance and is not unlike the general appearance of the buccal aspect of the modern upper teeth. | | (e) Orientation. In the Sivatheriinae, the lingual (lower) and buccal (upper) surfaces of the pillars present a generally smoother, flatter plane than those of the modern giraffes, because the surfaces of the Sivatherine pillars tend to lie more parallel to the longitudinal axis of the jaws than in the modern Giraffa, where the surfaces of successive pillars tend to overlap each other more acutely so that each pillar has the appearance of being independent of the next. The talonid of the lower M, tends to be set at a variable angle to the longitudinal axis of the tooth: the axis of the talonid is nearly in line with that of the tooth (180°) in Old. 1, 3, 92, 120, Hopefield 4029, or at an angle of about 130° in Olduvai 6, F. 3656, Hopefield 4028A. 2. METRICAL FEATURES A-P Transverse Index Teg = N M Range of N M Range of N M Range of variation variation variation DM? 41°0 I.) 592-0 OO: DM+* tr 46-7 We 307 I 85:0 Ps 1 44: M? 2 41°5 38: —45- 2 509 47- —54-8 2 122°5 121°6-123°5 M? 3. 47:6 46:3-49- 3 49°7 47:0—-54- 3 10471 Q6-1-112°4 M® Bile 595) 440-60; 5 48:7 46:0-51°9 5 ORD. ape aren DM, De oer a) i) Byipi\ DM, I 55'0 T2450 1 ASF Pace 3. 26-0 3 18-2 17--19°8 3 70:2 65:4—76-2 Ps 3 354 34°0-36°3 3 259 23:-29°5 Sie o2) 04:0-81-3 Py 8 42:6 37°0-47°4 7 30°4 28°3-32°4 (eet Ooo TiS M, Tei 5 One) aka eos 9 336 27°9-39°5 Ty G2sOr 52-4897 M, 12 51:2 486-55: 10 35°7 33°2—40°2 10 70:0 62:8—76-0 M,; 9 68:6 59:0-77- 106 333°6 Ss 3I- —30°7 9 49:6 41-9-60°1 I? I Zlko~ ; 1 184 1 87:6 I? 222-39 22-0227 2 17°5 17°4-17°6 2 883 76:6-80-0 - TABLE 55. Transverse / A—P Index in the Sivatheriinae. The data exclude the dimensions of F.39, MMK 3685 and those teeth not definitely diagnosed as either M? or M?. (Italicized figures indicate approximations. ) _ The absolute A-P and transverse dimensions (table 55) in the Sivatheriinae are greatly in excess of those for modern G. camelopardalis. The calculation of the Transverse/A—P Index in both cases produces interesting observations: (a) The African Sivatheriinae have a constantly (except for M+) smaller ~Transverse/A—P Index than G. camelopardalis for the milk and the adult, in both upper and lower, dentitions. This indicates that the A—P length is relatively more reduced than the transverse breadth in the modern giraffe (fig. 24). . 514 ANNALS OF THE SOUTH AFRICAN MUSEUM Upper Lower Tooth Series ——_—_—__—————_ Mean Range of Mean Range of variation variation DM3]| Giraffa camelopardalis! 86-9 78-8— 96-0 66°6 52-6— 78-2 African Sivatheriinae? 80:2 54°1 DM4| Giraffa camelopardalis 92°3 82:0-101°0 52°3 46:2— 58-0 African Sivatheriinae 85°0 43°7 P2 Giraffa camelopardalis 122°9 104°0—-149'0 ; 87°4 56°5-124°0 Stvatherium olduvaiense*® I110°O0 100°0—-119'0 78-0 African Sivatheriinae 70°2 Sivathertum giganteum* 113°0 65°4— 76:2 P3 Giraffa camelopardalis 125°l1 101°0-157°0 91g §=-: 688-1150 Stvatherium olduvaiense 128-0 126:0-130°0 74:0 65:0— 83:0 African Sivatheriinae 737% 64-0— (61-9 Stvatherium giganteum 120°0 P4 Giraffa camelopardalis 132°9 = 114"0-158°4 89:0 63-7-106°9 Stvatherium olduvaiense 130°0 =110°0—-142°0 76:0 67:0-— 88-0 African Sivatheriinae 972°7 65°5— 77°3 Stvatherium giganteum 129°0 Mr Giraffa camelopardalis 103°9 88-0-119°4 78-6 67:9— 981 Stvatherium olduvaiense 102°0 89-0-116°0 74:0 64:0— 81°0 African Sivatheriinae 122°5 = 121°6-123°5 72°0 58:4— 89°7 Stvatherium giganteum 107°0 100°0—115'0 M2 Giraffa camelopardalis 106:8 g1‘6-128-0 77°8 69°7— 90°9 Stvatherium olduvaiense 100-0 71:0 65:0— 83°0 African Sivatheriinae 104°1 96-1-112°4 70°0 62:8— 76-0 Stvatherium giganteum 97:0 94°0—100°0 69:0 66-0— 72:0 M3 Giraffa camelopardalis 107°4 Q1°5—-122°4 61:2 47°6— 76-4 Stvatherium olduvaiense 106:0 102:0—108:0 50°0 =49°0— 51°0 African Sivatheriinae Q5°0 85°3-111'1 49:6 41°9g- 601 Stvatherium giganteum 93°0 92°0— 94°0 49°0 48-0- 51°0 TABLE 56. Tr. / A-P index of the different types of teeth in the extant Giraffa camelopardalis, and in several series of Sivatheriinae. 1 Calculated from data from U.S.A. Museums; see Section I, chapter 5. 2 From the African Sivatheriinae presently described. 3 Calculated from data, according to Dietrich, 1937, 1942. 4 From data according to Colbert, 1935. (b) These observations have also been made for other Sivatheriinae on the basis of the data of Dietrich (1937, 1942) for material from Olduvai, and of Colbert (1935) for S. giganteum from the Siwaliks. Their results fall into the range of variation of the African Sivatherine material here described. (table 56). However, the index for Dietrich’s and for Colbert’s M1? fall outside the range: this peculiarity, linked with the exceptional result obtained for M?! from the other African material, suggests that the sampling of the M! presently described has not been representative. Table 57 illustrates, for the various collections, the MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA aD difference of their respective index from that of G. camelopardalis: it averages 10 units in the case of the African Sivatheriinae. Tooth Sivatherium — 12°9 +29 == 2:9 st I°9 —6°8 —_ I*4 UPPER African —6°7 ae +18°6 = PG) —12°4 Stvathertum olduvaiense! Sivatheriinae* giganteum® ae, Tae =39 +3°1 —9°8 — 14-4 moe LEG — 13°0 — 4:6 —6:8 —I11-2 Stvatherium olduvaiense1 Sivatheriinae® giganteum® LOWER — T2°5 —8-6 = F°2 —18°8 —16+3 —6-6 —7°8 —11°6 African Sivatherium —8-8 —'2-0 TABLE 57. Absolute difference between the respective Tr. / A—P index of several collections of Sivatheriinae and that of G. camelopardalzis. 1 Data from Dietrich, 1937, 1942. 2 African material here described. 3 Data from Colbert, 1935. (c) The significance of the difference of means between the fossil Sivatherines and G. camelopardalis has been statistically tested, and has proved ‘highly significant’ for DMsg, P,, Ps, Py, M,, M2, M3, M3: ‘significant’ for DM,; ‘not significant’ for DM’, DM?# and M?. (d) A further step was to estimate the Tr./A—P Index in other fossil genera and subfamilies, namely, Palaeotragus, Honanotherium and Orasius (data according to Bohlin, 1926). A similar low index was obtained (see table 58). Con- sequently it appears that the same evolutionary trend of a reduction greater for length than for breadth has been demonstrated by different phyla in the giraffid family. Giraffa Tooth camelopardalis1 DM? __ 9668 (35) DM? 86-96 (38) DM? g2°28 (38) DM, = 64°44 (37) DM, 66-58 (36) DM, = 52°30 (37) PA 122-86 (117) p> 125712 (120) Pp 132-86 (116) M? 103°91 (138) M? 106-79 (130) M3 107°44 (121) P, 87-41 (109) P; 91-95 (124) r. 89-02 (127) M, 78°59 (143) M, 77°77 (135) M; 61-21 (121) Palaeotragus* 70°84 (2) 78°51 (6) go-82 (7) 63°14 (5) 66-15 (6) 60-03 (6) 100°08 (7) 107°73 (9) 124°35 (9) 108-37 (13) 10501 (13) 101-62 (12) 83°90 (5) 84°47 (9) 82-03 (10) 82°35 (4) 79°45 (13) 50°41 (9) Honanotherium? 71-91 (4) 80°65 (2) 95°53 (2) 58-82 (1) 61-90 (1) 59°37 (1) 102°54 (2) 115°41 (3) 134°77 (4) 106-11 (4 109°38 (8) 108-39 (6) 77°77 (1) 77°59 (2) 82-97 (2) 74°28 (1) 87°09 (1) 55°23 (3) Orasius? 80-95 (1) 97°95 (2) 109°76 (3) 101°85 (2) 97°73 (3) 91-63 (3) 70°00 (1) 77°27 (1) 65°38 (1) 69°23 (1) 52°77 (1) Sivatheriinae® 80-2 (1) 85:0 (1) 122°5 (2) 104°1 (3) 95°0 (5) 70°2 (3) 73°1 (3) 72°7 (7) 72°0 (7) 70°0 (10) 49°6 (9) TABLE 58. Transverse / A—P index in fossil and extant Giraffid teeth. Between parentheses ( ) is the number of specimens from which the mean has been calculated. 1 Specimens from U.S.A. Museums (see Section 1, chapter 2). 2 Bohlin, 1926. 3 African material here described. ANNALS OF THE SOUTH AFRICAN MUSEUM - 516 (z) o0-LL (z) aie (z) 09.26 (1) 09-L6 (11) 00-29 (1) 06.28 (1) 09-88 gdCUllLIgy} VAG uUvoLy (1) 00.001 (1) 0g-%g (1) 08.96 (1) 06.26 (1) 08.Sg (1) 00.011 (1) 00.Sg (1) oF-S6 POD 6) (9) g1-42 (G1) 14.€01 (31) SP.96 (11) 1.26 (1) Sh.6g (6) 99-%8 (01) SP.101 (8) 01-98 (01) 69-86 (6) 00-08 (g) 56.06 (9) $g-S2 (€) €g.LL g5N3D140I0] 0g (1) 08-94 (601) oF-0g (11) 10-611 (S) gt-Por (911) a1-¥or (z) SS.Po1 (9) 1-99 (a1) &1.Por1 (3) S6.99 (€) 1.76 (1a1) bP.16 (Z) 01-99 (9) 1&-S6 (4a1) a1.6 (3) oF-06 (z) 06.09 (601) z0.S6 (z) S6.For (z) 6.16 (161) 09-101 (1) 06.68 (3) 08-88 (901) gz-16 (2) 09-96 (z)Sg.L6 (a1) 46.16 (S) bS.6Z (1) 09-94 (g&) LP-Lg (V) of.99 (1) 0S.L9 (@&) £6.26 (€) g1-9/ (1) 06.69 (9) 16.9 (€) 06.9 (1) 09.S6 (g&) g&-SZ glunrsayzopopjaFT wiarsayjounuofy ,S4jvp1vgojauva vffo.iy *d W1sue'] *d YisueT eI WIPeeig JN WIpeesig eI Y18u0'] IN W807 eA WIpeerg TN Mpeesg WW Y1Su0T iW 4isu2T rd Ypeorg ed YIpeosig rd YSU] ed Y1Su9'T ed UPeoig ed YIpeoig ed Y}SU0'T ed Y1Su9T WN Mpeeig aN peerg *W Y}8u2'T aN Yi8u0T eING Wiper eC Wpeeig eNC Y8U2'T aN Y8u2T xopuy BP MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA "QS 3IGeT, 29S ¢ «2 4 ‘paye[noyvo useq sey UvoUT oy} YOIYM UO’ suowTIOods jo saquinu 9} st ( ) sasayjUosed UsIMJog ‘sdnois pyyesis JOUT)xXo [eIDAVs JO yey} YIM poseduro0o syvpsvdojawv9 -4H yUeX2 UI xopuy [ejueq ‘6S atavy, (b) oz-For (L) ol. (9) 09-g6 (¥) 0g-06 (z) 00.06 (%) 04.98 (z) 0S.L9 (1) o£.46 (1) 02.22 (1) 02.06 (1) 00.001 (1) 02.88 (1) 09-06 (z1) L&.6o1 (11) Fz-L49 (a1) 76.96 (S) 96.46 (o1) 4.48 (11) 06.49 (9) 99-4 (1) 00.g01 (1) 00-99 (1) 0S.%6 ) 02-98 (z) SE.z6 (1) 09-42 (01) VE.101 (Gz1) 96-64 (PS1) 9&-¥6 (S61) 9.86 (gi1) 41-86 (z1) $2.06 (o11) 04.94 "WW UpeeIg WA wpe *W Wsu0'T “WAL YSU] 7 Wpesrg ‘W peer WW Ysua'y TW WsueT "d wiped *d Mpeg "d yysueT *d W)8u0] ad Wpeoig *d Wperig 518 ANNALS OF THE SOUTH AFRICAN MUSEUM Two hypotheses may be advanced to account for the relatively increased breadth of the modern teeth and the reduction in length of the total molar- premolar series relative to the length of the jaw (see Section III, chapter I, A): (1) The maximal contact wear between contiguous teeth was observed in the fossil lower jaws, especially in M, and P,, far more than in the modern jaws where it is only occasionally observed. It is also in the lower jaw that the greater length reduction has occurred: this may have been an attempt to compensate for the ‘impacted’ effect of the over-crowded teeth and in order to accommodate them. | (ii) ‘The fact that the teeth decreased more in length than in breadth may be explained by the fact that there is a selective advantage in a broader grinding surface for the side-to-side masticatory movements. Owing to the lack of a complete maxilla, one is confronted with a problem to which an answer cannot yet be supplied, namely, the disproportion between the respective length reduction in the total upper and lower dentitions. (e) Dental index: The dental index (Section I, chapter 5) is calculated to be usually smaller in the fossil genera (Sivathertum, Honanotherium, Helladotherium, Palaeotragus, Orasius) for length and breadth (with constant exceptions in the dental breadth index for M1/M?, M,/M, and M,/M, (table 59). Consequently, in the dental series, a particular posterior tooth, if compared with the tooth immediately anterior to it, is relatively longer in the fossil genera than in the modern giraffe material. This indicates that the reduction in the length of individual teeth has been greater in respect of the more posterior teeth. CHAPTER 2 DIAGNOSIS OF AFRICAN FOSSIL GIRAFFID GENERA AND SPECIES Colbert (1935) has modified the classification of the family Giraffidae of Pilgrim (1911), Bohlin (1926) and Matthew (1929) along sound lines. Because of the limitation of the African material, it is not the purpose of this paper to criticize this classification which is generally acceptable as a basis of discussion: GIRAFFIDAE Large, ruminating artiodactyls, with heavy, rugose cheek teeth. The skull may or may not have horn-cores, but if they are present they show a great variety of development. Bones of cranial roof pneumatic. Lateral metapodials and digits atrophied. Palaeotraginae Primitive, medium-sized giraffids, having as a rule one pair of supra- orbital, frontal horn-cores. There may be a second pair of horn-cores at the 9 MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 519 anterior extremities of the frontal. Horn-cores in the form of simple tines, well developed in the males, feebly developed or absent in the females. Skull usually elongated, dolichocephalic. Cheek teeth brachydont, with moderately coarse sculpture on the enamel. Neck and limbs slightly elongated. Genera: Palaecotragus Achtiaria (syn. with Palaeotragus). Giraffokeryx Okapia Samotherium Alcicephalus, Chersonotherium, Shanshitherium (syn. with Samotherium). ae are thor doubtful status; placed here provisionally. Guiraffinae Large giraffids, with a moderately brachycephalic skull. Horns variously developed, being on the parietals and frontals, and in Giraffa a single median horn is also present, located on the nasals. Horn-cores rounded or flattened on the ends and covered with hair. Skull roof with highly developed sinus cavities. Cheek teeth brachydont, with heavily rugose enamel. Limbs and neck greatly elongated. Genera: Guraffa Orasius Honanotherium Stvatherinae Gigantic giraffids, with large, heavy, brachycephalic skulls. Horns variously developed, being of frontal and parietal origin. Skull roof with large sinus cavities. Cheek teeth moderately hypsodont, with heavily rugose enamel. Limbs not elongated but very heavy. Body heavy. Genera: Stvatherium. Indratherium (syn. with Sivatherium) Bramatherium Aydaspitherium Helladotherium Vishnutherium Griquatherium Libytherium On the basis of the above definitions, the Hopefield giraffids which form a homogeneous group (except for one tooth, 3345, which has already been referred to G. ?camelopardalis, see Section II, chapter 4) may be assigned to the Siva- theriinae. Although descriptions and diagnosis of the different genera of Sivatheriinae have been published, no diagnosis (per se) is available in the literature for Vishnutherium, Griquatherium or Libytherium, yet the features of the generic types are briefly commented on. Helladotherium is hornless and may be excluded from consideration of the Hopefield material. 520 ANNALS OF THE SOUTH AFRICAN MUSEUM Colbert (1935) proposes the diagnoses for Sivatherium, Bramatherium and Hydaspitherium as follows: Sivatherium A gigantic Pleistocene giraffid, with four horns in the male, an anterior conical pair, arising from the frontals, and a posterior, palmate pair situated on the parietals. As in the other gigantic Siwalik giraffids there are deep pits in the temporal fossa for the temporal muscles, and on the supraoccipital for the neck muscles. The face is very short, the nasals being retracted and strongly curved. The teeth are large, with rugose enamel. Body and limbs heavy, limbs not elongated. Bramatherium A gigantic Upper Tertiary giraffe having four horns, two of which grow up from the fronto-parietal region, and two of which extend laterally from the parietals. Face short, with nasals considerably retracted. A large groove occupies the parietal region just below the horn-core bases as an accommodation for the temporalis muscles. Deep pits are located in the supraoccipital for the heavy neck muscles. Teeth large and heavy, with rugose enamel. Limbs and body presumably heavy and massive. Aydaspitherium A gigantic giraffid with two horn-cores, fused at their bases into one solid mass, on the frontal-parietal region. The face is short, the nasals retracted. There is a large parietal or temporal groove below the horn-core for the accommodation of the temporal muscles. Teeth large, quadrate, with rugose enamel. Limbs massive and not extraordinarily elongated. From the above diagnoses it can be seen that these forms overlap con- siderably, except in respect of their horns. ‘The brief descriptions, however, are not found sufficient for positive determination of the Hopefield material, and consequently various other details of the material were compared with other collections, and it became obvious that the Hopefield specimens had greatest affinity to Sivathertum olduvaiense, to which the homogeneous Hopefield group is assigned without hesitation. It has already been stated (Section II, chapter I) how Hopwood (1934) first diagnosed his original Olduvai material as Helladotherium olduvaiensis on the basis of some teeth and a partial hind-limb. Later (Hopwood, 1936), with the discovery of ‘palmated antlers’, the material was referred to Stvatherium: Hopwood stated however that ‘the antlers are not so widely palmated as in S. giganteum, and terminate in a recurved point’; hence, the specific determina- tion of S. olduvaiense. It is now necessary to discuss and compare the material from the various sites in Africa in regard to this diagnosis. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 521 I. OLDUVAI AND OTHER EAST AFRICAN SITES All the specimens from Olduvai form a homogeneous group, both from a non-metrical and metrical point of view. As a result of the fact that Orangia- therium vanrhyni, which was discovered and actually named earlier than Sivatherium olduvaiense, has now been invalidated and referred to a subspecies of S. olduvaiense (vide infra), the question of generic or specific priority does not arise any more. Consequently all the Olduvai material and the specimens related to it are assigned to S. olduvaiense. To this group are also referred the two specimens marked ‘Marsabit Road’, from Kenya. Il. LIBYTHERIUM MAURUSIUM POMEL, 1892 On the basis of a fragmented mandible containing M,—P, and a part of P, (plate 53), Pomel proclaimed a new genus and species which he envisaged to be within the Sivatheriinae, but he does not state clearly his reasons for creating the new genus. In 1947, Arambourg assigned giraffid teeth and a scapula from Omo to Sivatherium olduvaiense, and in re-discussing Pomel’s material, he gave as reasons for separating Libytherium from Siwvatherium: (1) the premolar series is reduced; (2) P, has an open inner wall where para- conid and metaconid remain separated; (3) the parastylid is very developed in P, and in the three molars. Furthermore, in a right mandible (1950-1-1, Muséum d’Histoire Naturelle, Paris) from Garet Ichkeul, St. Arnaud (Tunisia) (plate 52(a)), diagnosed as Libytherium maurusium, the above features are absent, and in all _ respects this mandible is identical to specimens at a similar stage of wear from Olduvai (Old. 6, for instance). Arambourg (1948(a)) stated that excavations at St. Arnaud provided new fragments of Libytherium, and in particular characteristic ‘antlers’ (“ramures’) of the Sivatheriinae. Comparison of the non-metrical and metrical features of these horn-cores (1948-1-2, 1948-1-1) (plate 51), as well as of a third (unnumbered) specimen (plate 52(b)) (the cast of which has been made available by the kindness of Professor C. Arambourg and Professor J. P. Lehman) with posterior horn-cores from Olduvai, Hopefield and Tierfontein, indicates a distinct similarity of all the specimens and it is considered that they must be referred to Sivatherium olduvaiense. However, on the basis of Arambourg’s criteria for the Libytherium maurusium dentition, there is no doubt that the specimen on which he based these facts is decidedly different, not only from Sivatherium olduvaiense, but also from specimen 1950-I—1 from Garet Ichkeul. Furthermore, these criteria are so distinctive, especially the primitive nature of P,, that it is necessary to divide the North African material into Libytherium maurusium for the type specimen, and Siva- thertum olduvaiense for the horn-cores, mandibles and other similar specimens (e.g. specimens 1948-1-1, 1948-1-2, 1949-2-937, 1949-2-938, 1949-2-725, 1931-45, 1931-8, 1931-8-110, 1950—1—90, 1950—-1-1 in the Muséum d’Histoire 522 ANNALS OF THE SOUTH AFRICAN MUSEUM Naturelle, Paris) mostly as yet unpublished but examined by one of us, and to be published shortly by Professor C. Arambourg. However, it may yet become necessary to equate Libythertum with the rather widely varying genus Sivatherium. Ill. GRIQUATHERIUM Three specimens have been assigned to this genus: MMK 3685 and M 553B? are described as G. cingulatum, and F.39 as G. haughtoni. (a) ‘The Makapansgat specimen M 553 B* has been referred to G. cingulatum by Cooke and Wells (1947). They state that from an examination ‘it would appear that [these] lower teeth are of the correct size and form to belong to Haughton’s species though probably to a smaller individual’. From the range of variation obtained by the authors (see tables 39, 40) it is quite clear that the Makapan teeth fall within the range of the East African and the Hopefield material. In fact the breadth of M, of M 553 B! is the smallest in the range. The general form and character, which Cooke and Wells considered to be distinctive of the specimen, are identical to Sivatherium specimens at a similar stage of wear, e.g. Olduvai, specimen 92. Consequently, there is no basis, either metrical or non-metrical, for separating M 553 Bt from Sivatherium olduvaiense; this conclusion is further confirmed by the other Sivatherine material from Makapansgat described above, which are also referred to this genus and species. Although the lingual surface of P, of M 553 B is ‘open’ as in Pomel’s P, (Libythertum maurusium, type), there is still a considerable amount of the crown hidden by the bone, the tooth having just started to erupt. Because this ‘open’ appearance is confined to the upper portion of the tooth, it is estimated that it would resemble Old. 6 for instance, at an equivalent stage of wear. Further- more, the teeth are much larger than those of the Pomel specimen, and are identical to S. olduvaiense. ‘The former point mentioned was the only possible doubt the authors had concerning the determination of the Makapan M 553B specimens as S. olduvatense. (b) MMK 3685 (G. cingulatum Haughton 1922): ‘The designation of this tooth presents a number of problems. From the description of this specimen (vide Section II, chapter 2), the following characteristics appear to differ from those of other African specimens: (1) great breadth of the tooth; (2) relative difference in breadth of the two pillars; (3) marked cingulum forming an unusually developed proto- and hypostyle; (4) prominence of the meso- and parastyle; (5) the acute angulation of the lingual surface (in profile) to the crown-root junction; (6) the abnormal length of the posterior pillar relative to the anterior. Some, if not all, of these features may be found individually in single specimens, often in a modified form. But the above features appearing together produce this peculiarly gross form. It is certainly a Sivatherine, and it has been compared to both Hydaspitherium and Sivatherium, and it has been suggested that it is possibly even nearer Hydaspitherium (Bohlin, 1926; Colbert, MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 523 1935; Cooke, 1949; Singer, 1954). Nevertheless, those features in common with Hydaspitherium are also exaggerated and it is more massive. On this basis, and because Hydaspitherium has not yet been discovered in Africa, it is considered reasonable to refer this single, very worn tooth to Szvatherium with which it essentially shares most features. As regards the specific designation, although it bears a resemblance to F 2993, Old. 4 and tog from Olduvai, the latter specimens are closer to Hopefield 4027, 4023 and 4024 and share the same range of variation. MMK 3685 tends to fall just outside this group, but it is a single specimen and no other material from Africa can yet be referred to it. Furthermore there are no associated skeletal remains of the same animal. It is probable that if Cooke and Wells had all the comparative materia] from Olduvai and Hopefield available in 1947, they would not have referred M 533B! to G. cingulatum. It could almost be considered as a nomen vanum, but in the above discussion its generic nature is established. Consequently in the present state of our informa- tion it would be preferable to retain cingulatum as a species of Sivatherium. Fic. 25. Superimposition of dio- ptographs of the lateral aspects of F.39 ( Ver2Q80r Cite toe Ne and Hopefield 4025 (------ 22 ibhe probable crown-root junction of F.39 is superimposed on that of the other two premolars. The crown of F 2980 is worn right down to the root and beyond the crown-root junction except for a minute fragment buccally (on the left). The buccal border of 4025 is broken away so that the overlap of F.39 is only apparent. 1 Crown-root t: fe ce ogee bs aN o e e Levi Me JUNCTION Nz N \ : Se \; q: 2 : al Se \ “, areas are a re °° peo f (c) F 39 (G. haughtont Cooke 1949): In the description of the tooth (Section II, chapter 2), statistical data was presented to support the hypothesis that this specimen is not a lower molar. This may be strengthened further by comparing the specimen with known upper Sivatherine premolars, for example F 2989 and 4025 (fig. 25). It would not be difficult to reconstruct three roots on the specimen; its base fits perfectly and directly on F 2989, which is worn right down to the crown-root junction. The A—P axis and the breadth of F 39 are smaller than those of F 2989 and 4025. Comparison with other upper 524 ANNALS OF THE SOUTH AFRICAN MUSEUM premolars indicates that from the points of view of other dimensions and of general appearance, F 39 may be considered as an upper premolar, probably P?. ‘The major objections to this diagnosis are: (1) the extreme hypsodonty of the specimen, and (2) the bulge on the lingual surface just below the crown-root junction. The latter may be partly explained by the X-ray appearance (vide supra) and partly by individual variability. In connection with the hypsodonty, it is unfortunate that the available upper premolars of all the Sivatherine specimens are so few, and also that they are in advanced stage of wear. Consequently an adequate range of variation cannot be obtained for comparison. However, if one attempts to reconstruct available specimens such as 4025 then it would appear that the hypsodonty of F 39 may be equated. An unsatisfactory feature of this diagnosis is that the upper molar Old. 109—an almost unworn M?— measures only 47-0 mm. Even if this difference of 20 mm. (the teeth are at approximately the same stage of wear) were decreased by a greater number of specimens widening the ranges of variations, there would still be a significant difference in height. However, it is important to consider the dentition as a whole in the skull. If one examines the fairly complete S. giganteum skull (15283) in B.M.N.H., it is seen that there is a distinct downward convexity in a mesio-distal direction between P? and M3. Furthermore it is seen that P* is approximately at the summit of the convexity and consequently it is expected that P? and P* should be more hypsodont than M3. It is thus reasonable not to exclude this specimen from S. olduvaiense only on the basis of its hypsodonty, particularly because it is a single specimen and the comparable series is small. However, it is felt that there are sufficient grounds to tentatively place the specimen in a subspecies, namely haughtont. Because Cooke (1949) compared this specimen with Sivathertum (Griqua- therium) cingulatum, it is necessary to comment on his remarks. The bases of distinction drawn by Cooke for separating F.39 from Sivathertum (Griquatherium) cingulatum are not considered to be sufficient for a specific difference. First, the central pit of a tooth is so variable that it cannot be used for differentiating two species. Secondly, the small piece of cingulum remaining visible on the tooth indicates that the cingulum zs marked. For these reasons, and because of the fact that these two specimens come from the same area, although the exact localities of their discovery are unknown, the probability that these two unusually gross teeth belong to the same species is not insignificant. This independent conclusion supports the view stated above that S. cingulatum could be considered, on a subspecific level of S. olduvaiense. Further discoveries are required to resolve this problem. IV. MAKAPANSGAT It has been stated above (111 a) that specimens M 553 B! (G. cingulatum) and M 553 B have now been referred to Sivatherium olduvaiense. ‘This also holds MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 525 good for M 553 A and the other Sivatherine material described and discussed above (Section 11, chapter 3). V. ORANGIATHERIUM The horn-cores mentioned by van Hoepen (1932) show all typical features of those of S. olduvaiense, from North Africa, Olduvai and Hopefield. However, it has been mentioned that it is distinctive only for its extremely grooved appearance on its antero-medial convex surface. There are slight variations in regard to the position of the knobs, but these are known to be highly variable features. The four teeth from the same site (C 426 A, B, C and D)—two of them showing contact surfaces, and all four presenting the same degree of wear— almost certainly belong to one individual. Lack of accurate information con- cerning their discovery makes it impossible to associate definitely the teeth with the horn-cores. However, it would seem likely that they belong to the same species, if not the same individual, because they are derived from a single farm. Furthermore, van Hoepen (1932) stated that the teeth most likely belong to the horns. From a morphological point of view, C 426 A, B, C and D are identical to those extreme stages of wear in the S. olduvaiense material, but there are some important metrical differences. C 426 A is longer and broader than the M? of either Hopefield or Olduvai. Although the length of the anterior and posterior pillar do not differ much from S. olduvaiense, the great enlargement of the metastyle in C 426 A provides it with a great increase in length. The metastyle has been noted to be a well-developed character: but it is even more prominent in this specimen than in MMK 3685. Owing to the fragmentary nature of B, C and D, nothing can be said of the styles, but the length of the pillars, as for C 426 A, fall beyond the dimen- sions of the few M available. Unfortunately, the teeth are in extreme stages of wear, even the bases of the roots showing signs of attrition, and little more can be said of the distinctive features of these teeth. Because the horn-cores are so similar to those of S. olduvaiense (in which some specimens have marked grooves) and because the few morphological features of the dentition are also similar to S. olduvaiense, the authors hesitate to provide a new species only on the basis of dental size, especially in the light of the fact that so few M® are available and that the growth pattern in these few teeth appears to show a marked tendency towards ‘abnormalities’. It would appear that, because of the wide range of variation within the species, gross dental metrical exaggerations should be considered on a sub- specific level until clear criteria for specific differences can possibly be elicited in specimens yet to be discovered. These dental maxima may only be ecological variations, but, on the other hand, future evidence may indicate these maxima to be of a specific nature. 526 ANNALS OF THE SOUTH AFRICAN MUSEUM Consequently, even though criteria for a ‘subspecies’ in fossil material are difficult to elicit and clarify, it is proposed to refer the specimens of Orangiatherium vanrhyni to a subspecies of Sivatherium olduvaiense. ‘The question of priority which would normally arise both as regards genus and species, falls away as van Hoepen’s terminology is invalidated under Article 25(c) of the International Code of Zoological Nomenclature. To link the specimens with discoverer, it is proposed to designate them Sivatherium olduvaiense vanhoepent. The plaster cast representative (C.1492) from Florisbad appears to fall within the range of variation of Sivatherium olduvaiense. VI. ?CORNELIA, ?FLORISBAD The two milk molars marked B, and B, are Sivatherine in form and size but nothing is recorded of the site of their discovery. Because of lack of com- parative deciduous teeth and because of the fact that more than one subspecies might be represented in the Orange Free State, the authors have decided to refer these two teeth to an indeterminate subspecies of Szvathertum olduvaiense. ConcLuDING NoTE on TAXONOMY It is now necessary, on the basis of the above discussion, to modify Colbert’s classification (1935) of the genera of the Sivatheriinae as follows: Stvatherium syn. Indrathertum Bramatherium Hydaspitherium Helladotherium Vishnutherium Libytherium As far as the African fossil Sivatherine material is concerned, the following genera, species and subspecies are recognized from 22 sites extending from North Africa to the southern tip of South Africa: Libytherium maurusium Pomel 1892 Stvathertum olduvaiense Hopwood, 1934. Stvatherium olduvaiense haughtoni Sivatherium olduvaiense vanhoepent, subsp. nov. Stvatherium olduvaiense subsp. indet. Stvathertum cingulatum CHAPTER 3 THE FAUNAL RELATIONSHIPS AT THE AFRICAN SIVATHERINE SITES All the available data concerning the fauna identified at the various African sites (figs. 17a, 18, 23) where Sivatherines are known to have been 527 MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA i x x “mK OM “RRRHKRR oR eo a 78 e xX ° x 7 x oe oe oe x Muopy me ; "* DLOANIaIN : saqjn4 : uovIW'T ‘(uoteaqoi01g 14) u0dI01Q ; sup’) $@uoj0o') SNULOJSOIIDS snwospuaq? SIHIMAONYIO T, ‘UuoSs applinyy saqapag snsay SNYDMYLOSYAX) 4 snjoydar01ajaey "MISC xIspsCyouayy : sngaT ‘uas appiLodaT SNGDIVYABUILIG ye Bop ‘uos *a19asuT snouns! "* waLosonjny Sh se ie owory : SNIBYINGOJDAISNG ye pSNIIYIIGOYJDUTAYIDAT od = = snzayjidosa0y") STOR OUT ‘ ododvivd SEpLODEANFONE oie aa * §nga9094a') Spas ODEOS) *JOUL SNIAYI1G OWNS snaayngaqrT eI[IUIO‘) Peqsioy.y preyedoH yessuedeyeyyy | aT]Tes2sS10[O aSuaIa¢ | PIO III reanplo II reanp[o- I reanpio WeULyy OuIQ | | pneury 3S | [Neyo yoIeVsy UnNeN IpeM | ® VUANAYD ANNALS OF THE SOUTH AFRICAN MUSEUM 528 eljoulor) aa * peqst1of, ae oa ad 2 R yessuedeyeyy dI]Tesos10[O tas i 1tasusIIS Ie Pi@ III [24"PIO “SM MR "SR LEE MANE G) “83MM R "RR I FeanplO weuey OUWIC) - RM pneury 19 [NYxYIT yoIeVy unIyeN PEM, ISG Es teal Lig ynubodiaar SNLBOYIVALG yp snpojaay : os a aa $0121 NOD == = Ayqeqoad) SOL] AFUALAG : wnr41aYyjoqD1a") WN1ABYJOZIYISDIAIY Ale | LEEW ESE 0.0 oe ee oe o. snnby (sndqiyodjoy ‘sndqiyoyyousding =) siusyodqizy “= (uomandqnyoras BALIN) "Jour) uowvdqGiy] » "+ p1nDI0Lg Be : WN1LIYJOULIC] ye : ‘(uopoysipryI4p7 *[IUI) SsnyjnuuDpy Se ik { eporpewy ae ioe ise an see re He = UOPOTANG x UOpoYyo}052I5'~ "uoS plIsogolg | Ceres — ) InuwUDypy x "* SNIUDUP sngo1sagocic "TOUT “UdS appyay WNIAIYJOWO FT x SNPOLDYIVDIN eY 2s * xduoup -(oinyniog ‘ninpoudry ‘xud] *[OUL) SUaq = "* puavdey a a (aed ul niente *JouT) 7nI04/) sosunyy xDIUT : synud?y) VUANA 529 MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA ~* “mM RR R DIAZ Dx) cme Ras "+ sosankday oe oe avs snuanyqoy 2 x SNGDAZOSAN SNFDLJOILE) SDLBI0JOUNT x saqavyz0uuory "+ snopuyvag DIUDGDYDIN ; SDLIIOLO]IT (sypang— ) snpoy deny oy "* U0kL0) ze og He SNILDINUAD ET yp snasyoupq o° o- Oe o- xC1OQ ee me snsv.ogquyy mad Daag pounpa ee ea UL *[OUT) iene (orga yaed ul “[OuL) snqoy ox DQuioJUD]IY J snyqojoyqar) (smpoqng) snjoqng (Gaed UL SNIDGNG "[OUL) SVAI0L0UWOFT 4. (aed ur smjogne je) SNLBIULS * SNYIADING » za SNIDAJOAND J, nee ee snyqojasv4 J, Hi ee $04a91s Jays SnjaUD’) offs) oe “te snupjogodquyy = snsaoyrourgn J, » (SN4a0YIOJUOLIL) *JOUL) SNLBOYIOZONS 5, (SHuso1oopouoly ey) SNLBOYIONPIM AYN 4 ‘ SNLOOYIOSIYN’ 4. a a ce snsaoy ro CF] rhe ue ss SNLBOYIOU() » (snsa0yro4fyy *[QUL) SN4a0YyI0IvY ANNALS OF THE SOUTH AFRICAN MUSEUM Jo” "(961 ‘roftoofy 29 sa8UrIg) uopoyssprysp ATQeqord ¢ -(LG61 0} dn) saqis suTayAIG UeOTIFY Ye [VONUIPT S[Issoj JO UOTINQIASIp WIUs) “09 2FeL, *(9G61 ‘ueuIpser.g) snzayprdowigy Ul papnpoul MON _ ‘snugs JOUI]XxO = x pue , eljaui0‘r) jes) iGl preyedoy x yessucdeyry] dI[IVsas10[O 198Ua13G x Ne @® III [24 Plo JUG eu oy) . x xX x 3 vx x Orr =O Q sae aes ae ga = RR = i) (e) > @ < 5 ice Bo iS = =: & O S — “7 UNIZN IPE { sog SOLIIOLIIDYT $12040] 9g sapuosDnyqryjnd » MESA ¢ * * spo4opyup ‘uos adojuUup Gade ul = ) snévujospuayd VUINA MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 531 discovered is summarized in table 60. No material from the Vaal River deposits is included in this table because the Sivatherine specimens MMK 3685 and F 39 were recovered from unknown localities, and it would be impossible to associate them with any particular faunal group from the Vaal River as the fauna are derived from series of gravels of varying geological periods, the sequences of which have not been finalized. The data concerning the O.F.S. sites is incomplete, because in the case of Tierfontein very little is known of the associated fauna, while the mass of material collected from Florisbad and Cornelia has not been identified. It will be noted that only the genera are given in the table, because in very many cases there is dubiety about the specific identification. Furthermore, in some cases there is controversy concerning the genera themselves and some have been referred to or included in others (Hopwood and Hollyfield, 1954). Arambourg (1947) and Leakey (1958) differ, for example, as to the generic determination of the East African suidae. Purely on a statistical basis, the genera common to two or more sites have been selected and expressed in a table of correlation (table 61). ( ~ ee al ree: emits ep Se ae ee i ye ee eee MOmiee aye, oe Rh SS iy ion eS) Ge g 4 a 3s Oe Oe, te Se CunemnOn © HOMO May. Miweh ik Oo Garet Ichkeul 9 4 6 Gi 4 7 3 6 4 4 5 Sa ope T St. Arnaud 2 ene 7 mea 40) 6) 11 5 7 7 5 5) Eye ish 1 te Wadi Natrun 3 2 2 2 2 3 2 2 3 Teer otras I Olduvai I 2Oie) Ui LF OR eure) | TORN ary: On 2a Olduvai II Lea 2 2y hy RO Mh Say HP Say GLA ee tO” |) 4c 4 Olduvai ITI 12 8 Gi 6 6 8 Ria eal o Olduvai IV Ieee Ay eS Fe 2Oun EAN 5. =e A Olorgesailie 10 8 6 8 Sy gSte a Ee} Omo II a | TA ORAL Ges Serengeti 8 12 SOAP ato Kanam 6 (i GIR) Makapansgat OES ee ow Hopefield 6 2 Florisbad 3 ‘TABLE 61. Genera common to two different African sites. (See also Appendix.) Since this table was drawn up a number of extinct genera have been identified at Hopefield and at Olduvai, but they have not been included in this table. The interpretation of such a table must be undertaken with great care because of the varying amount of material collected from different sites. This table does not express a relative time correlation. However, if one takes Olduvai, Omo, Makapansgat and Hopefield, where large numbers of specimens have been recovered and identified, the correlation becomes more representative, but the actual figures do not indicate which genera are in common to all sites. A similarity between two sites is indicated: Hopefield for example appears to 532 ANNALS OF THE SOUTH AFRICAN MUSEUM have more in common with Olduvai IV than with Omo. Furthermore, if one estimates the extinct forms and the extant forms at each site, and works out the index of the relationship between extinct forms and the total, and between extinct and extant forms, one obtains an interesting gradation of series; especially if one selects arbitrarily those sites where there are more than a total of 15 recognized genera (table 62). For both indices one obtains the same gradation of increasing indices, namely, Hopefield, Olduvai IV, Serengeti, St. Arnaud, Makapansgat, Olduvai II and Olduvai I. It must be emphasized that this is not an attempt at an age sequence, but merely indicates the propor- tions of the extinct and extant fauna. Nevertheless there is some evidence that Hopefield overlaps the Olduvai IV period, and some of the faunal evolutionary sequences (e.g. higher crown of Mesochoerus lategani (Singer & Keen, 1955) compared with Mesochoerus olduvaiensis) and the more evolved human- manufactured stone tools (Singer & Crawford, 1958a), tend to corroborate this and even suggest that a portion of the ‘Hopefield period’ extends to slightly more recent times than the period indicated by Olduvai IV. Number of recognized genera : Extinct Extinct Total Extinct xX 100 Total Extant Garet Ichkeul 13 3 Dat 30:0 St. Arnaud 16 4 25°0 BE) Wadi Natrun 9 4. 44°5 80-0 Olduvai I 28 IO 35°7 55°6 Olduvai II 28 9 32-0 47°5 Olduvai III 13 5 38°4. 62°3 Olduvai IV 35 7 20:0 25°0 Olorgesailie II 4 36-4 57°0 Omo 26 9 34°6 53°0 Serengeti 33 8 24°3 32°1 Kanam 14 6 42°9 75°0 Makapansgat 48 15 Big 45°5 Hopefield 30 5 16°7 20°0 Florisbad 14 I 7% 749 Cornelia 7 4. 57°2 132°6 TABLE 62 However because all the material from Olduvai, Makapansgat and Hope- field has not yet been definitely identified, these indices may have to be altered in time (see ‘Appendix’). The data indicate clearly how, throughout Africa, a large percentage of the fossil material is extant and how these extant forms are widespread throughout the African Pleistocene. Consequently, it is not surprising that Giraffa has been recovered from Upper, Middle and Lower Pleistocene sites, and that Sivatherium has been recognized in each of the four Beds at Olduvai and even at Omo. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 533 Confined to Confined to Confined to Stage I Stage IT Stage III — Common to —————— Commonto =—————— Omo- I-II Olduvai I-II II-III Olduvai III-IV Kanam Serengeti Olorgesailie Dinopithecus Crocuta FAystrix Stmopithecus Otocyon Lepus Anancus Aonyx Canis Bubalus Syncerus Deinotherium Acinonyx Mesochoerus Nesotragus Aepyceros Sus Chalicotherium Bularchus Philantomba Kobus Metridiochoerus Pultiphagonides Adenota Phenacotragus Antidorcas Mammuthus Parmularius Hippotragus Redunca Menelickia Serengeticeros Gorgon Pelorovis Omochoerus Serengetilagus Beatragus Damaliscus Stegodon Heterocephalus Felis Thaleroceras Archidiskodon AXerus Notochoerus Homotherium Pedetes Choeropithecus Tachyoryctes Equus Mungos Orycteropus Metaschizotherium HAylochoerus Okapia 16°:2% 8-8% 250% 176% 13°2% Common to Stages I, II and III Hyaena Phacochoerus Hipparion Hippopotamus Ceratotherium Giraffa Diceros Alcelaphus Potamochoerus Oryx Gazella Strepsiceros T aurotragus 1G. 5 Total: 68 genera. TABLE 63. Correlation of East African Pleistocene Fauna. A correlation of the East African fauna as it is presently described, according to the stages recognized (table 63), produces a number of genera (20°) common to all three stages, which is almost identical to incidences in the Hopefield material. ‘This indicates that throughout Africa about 20 per cent of approxi- mately 70 genera persisted (though possibly undergoing specific determination) over a great length of time, despite changing ecological and climatological conditions. Furthermore, in respect of a single genus, namely, Giraffa, and one single species, camelopardalis, which has been recovered from the Lower Pleisto- cene (at Omo, see Arambourg 1947), there has been no evolutionary change, despite its wide dispersal—chronological, climatological and spatial. This fact demonstrates the tremendous adaptability of G. camelopardalis. However, during this period, a much smaller species, G. gracilis, became extinct, as well as two genera (Sivatherium and Libytherium) of another subfamily (Sivatheriinae). 534 ANNALS OF THE SOUTH AFRICAN MUSEUM APPENDIX After the MS. had been completed, three series of giraffid material were made available to the authors. Newly discovered specimens from the Lime- works breccia at Makapansgat were kindly sent by Mr. J. W. Kitching from the Bernard Price Institute for Palaeontological Research, Johannesburg. The Curator of Vertebrate Palaeontology (Dr. A. J. Sutcliffe) of the British Museum (Natural History) informed us that some more East African fossil giraffid specimens had been found in storage which had previously not been known to be available. The third series, one specimen, was discovered by one of the authors (R.S.) at Baard’s Quarry, Langebaan (Cape Province), which is about 10 miles NW. of the Elandsfontein site at Hopefield. This specimen, an astragalus, had been recovered, with other bones, from the layer of phosphatic nodules about 5-10 feet below the surface. Most of the bones have been identi- fied and they belong to animals similar to those found at Elandsfontein. Only one identifiable specimen, Stegolophodon sp., belongs to a much earlier horizon (Singer & Hooijer, 1958). | A. MAKAPANSGAT MATERIAL I. Giraffa M 2085: Left M, or M,. M 1801: Right canine with tip of root broken off. M 1798: Portion of left ramus of mandible with M, and fragmented Mg. M 1800: Fragment of maxilla with portions of right M}!, P4. IT. Sivatheriinae M 2087: Fragment of mandible containing unerupted I,, I,. M 2086: Fragmented right M3. M 539 A: Now in British Museum (Natural History) and numbered M 16729. Jaw fragment of juvenile. DESCRIPTION I. Giraffa M 2085 This is a left lower molar, probably M,, but the possibility of its being M, cannot be ruled out. It is a complete tooth with a fragment of mandible between its roots. The crown is identical in appearance to M 942 —M 1113, except that (i) M 942 has a minute ectostylid and M 2085 has no trace of it; (ii) the stylids on the lingual surface of M 2085 are less marked than those of M 942—M 1113; 7 (iii) M 2085 is in a more advanced stage of wear; and 10 MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 535 (iv) the posterior pillar of M 2085 is more rounded at the base of the buccal surface than M 942—M 1113, and M 2085 does not present the same marked indentation of the crown on the posterior surface that M 942 — M 1113 has. The roots of M 2085 are very robust, being more massive than those of the modern G. camelopardalis, although they are as short. Height CROWN Length Breadth Lingual Buccal (a) Anterior pillar 15°5 25°5 19°6 16°3 Posterior pillar E77 25°0 18°5 14°3 Whole tooth 34°7 25°5 (d) Roots Breadth Height Anterior 24°4 28°8 Posterior 23°7 28-0 Table 64. Measurements of M 2085 (mm.). These dimensions fall within the range of the other Makapansgat fossil Giraffa; the breadth falls within the range of variation (at upper end) of the modern Giraffa camelopardalis while the length falls just outside the range of the modern species. This, like the other Makapansgat Guraffa specimens, is included in G. camelopardalis. M 1801 (PI. 23, e, f) This is a rather worn right canine with most of the root intact. It presents no features not present in a canine of a modern G. camelopardalis. Its dimensions fall within the range of the modern species: Crown: Length: 21-5 (mm.) Breadth: 10-2 Height: 18-0 Root: Base-tip: 33+ M 1798 A portion of the left horizontal ramus of a mandible with a complete M, and a portion of the anterior pillar and the entoconid of M,. The buccal portion of the mandible is partly broken away exposing the anterior root of M, and the sockets of the roots of P,. The teeth are in a fairly advanced stage of wear, intermediate between the stages of M 942—M 1113 and M 2085. The general appearance of M 17098 is similar to that of the other Makapansgat fossil Giraffa M, specimens, except that it has a prominent ectostylid and only a slight hypostylid which has been worn away by the abutting anterior pillar of M,. The anterior root of M, tapers towards the tip, in contrast to that of M 2085 which is rectangular. 536 ANNALS OF THE SOUTH AFRICAN MUSEUM This specimen is the only M, of the fossil G. camelopardalis in the present survey. The remaining portion of the M, of M1798 resembles the other M, Makapansgat specimens, being tightly wedged against the posterior pillar of M,. On the lingual surface of the entostylid the median costa is dimpled by a V-shaped vertical depression. (2) ManprsLe: Breadth opposite M,/M,: c. 34 Height opposite P,/M,: ¢. 54 Height M, Length Breadth Lingual Buccal (6) Anterior pillar 15°5 22°79 c. 14 14°9 Posterior pillar 16°3 23°4 c. 15 14°2 Whole tooth 31°6 22°77 M, Anterior pillar 16°3 G29 c. 20 18-0 Posterior pillar _ — 20°6 — TABLE 65. Dimensions of M 1798 (mm.). M 1800 A fragment of right maxilla containing portions of P* and M!. They are in a very advanced stage of wear, the most worn down of the Makapansgat Giraffa series. M, Length Breadth Height Lingual Buccal Anterior, pillar. 3411-5 29:0 7:0 6:5 Posterior pillar... — — - 8-2 Whole tooth .. 25+ — —- — Dimensions of M 1800 (mm.) IT. Stvatheriinae M 2807 A fragment of the right body of a Sivatherine mandible, the symphyseal aspect of which fits that of M 553A perfectly (Pl. 40). A small portion of the posterior border of the body is present forming an arch with that of M 553A, but an anterior directed fracture has separated the body from the horizontal ramus about I cm. along the posterior border to the right of the symphysis. The anterior half of the mental foramen is present, as well as a small accessory foramen 11 mm. beyond the anterior border of the mental foramen with which it is continuous by a canal in the bone. Anteriorly the unerupted crowns of I, and I, are visible where the outer bony alveolus has been broken away and MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 537 the socket for the root-tip of I, is visible. It is evident that I, is at a more advanced stage of the process of eruption than I, which is ‘impacted’ against I, and is partly overlapping it. The anterior enamel edge of the crown of each tooth is at an angle of about 45° to the superior surface of the body of the mandibie, the medial edge of each enamel border being nearer the surface. Furthermore it is obvious that I, considerably overlapped the plane of I,, indicating that during subsequent growth considerable lateral expansion of the body was still to occur so as to accommodate all the teeth (see also p.498). It is curious that there is no sign of the canine, although the broken edge of the mandible just lateral to I, appears to be the socket for the root. On the left this region is obscured by breccia. The order of eruption, as regards the premolar-incisor-canine teeth, then, simulates that in modern Giraffa camelo- pardalis. ‘The enamel of the incisor is fairly rugose and the crown enamel is markedly convex, the occlusal edge of the enamel projecting vertically. If one mentally reconstructs the anterior portion of the body one obtains the conviction that the incisor would not be in the same plane as the superior surface of the body but rather at about 45° toit. This tends to confirm our opinion stated on page 408 that the incisor-canine row of the Makapansgat Sivatherium olduvaiense would be at an intermediate position between Hydaspitherium sp. (AMNH 19684) and Giraffa camelopardalis. It is now possible to assess more accurately the length of the body of the mandible, namely, about 120 mm. This specimen also indicates that the height of the body of the mandible relative to the breadth of the body is greater in the immature individual. With maturation of the individual the depth decreases and the breadth increases. This principle also pertains to the modern G. camelopardalis. The only measurement that can be taken on the teeth is the breadth of the crown of I,, namely, 23-7 mm. M 2086 This is a right M;, broken off at the crown-root junction with the anterior pillar almost complete, the posterior pillar having the occlusal portion of the hypoconid broken off, and the talonid having most of the buccal cone broken away. The slightly rolled edge of the enamel of the occlusal surface indicates that the tip of the crown of the tooth was only just above the alveolar margin of the mandible, so that most of this specimen must still have been embedded in the mandible. This very early stage of eruption is in conformity with that at which a M, of M 553B* would have been expected to be. Furthermore, as was found with M 2087 which belongs to M 553A, it would be reasonable to con- clude that M 2086 is the M; of the right side of the individual to which M 553B}, M 553B, M 553 and M 2087 belong. It can also be concluded that this specimen supports the view that the order of eruption of the molars of Sivatherium is the same as that in Giraffa camelopardalis, namely, M,-M,—M, in that order. 538 ANNALS OF THE SOUTH AFRICAN MUSEUM It is interesting to note that in M 2086 the angularity of the buccal cones seen in M 553B? is obvious only in the second pillar of M 2086 while its anterior pillar is more rounded. | In general appearance, M 2086 closely resembles M, of Old. SK. II, 92, except that the parastylid of M 2086 is a more marked ridge, while the median ridge (costa) of each pillar of M 2086 is not yet as prominently developed as those of the Olduvai specimen (which is at a much later stage of eruption). The finely rugose buccal surface of M 2086 has three horizontal ridges on the enamel on the anterior pillar and one on the posterior pillar—these are more marked than in Olduvai 92. Measurements of M 2086 (mm.) : Length Breadth Height Lingual Buccal Anterior pillars. |) 26-0 33:0 49+ 49+ Posterior pillar .. 22-0 29+ 5r+ — Talonid .. .. 14+ 19+ 31+ — Whole tooth .. 66+ 33-0 — = It is clear that this specimen falls within the range of the measurements of Sivatherium olduvaiense, and, being the least worn of the Msg series, it extends the upper range of variation of the height of the tooth to 51-++ mm. M 539A (PI. 36, g) This is the number given by the Bernard Price Institute for Palaeonto- logical Research, Johannesburg, but it isnow permanently in the British Museum (Natural History) where it has been given the number M 16729. It is a fragment of a left mandibular ramus of a juvenile and definitely belongs to the opposite side of the Sivatherine specimen M 539B. It contains the two posterior pillars of DM4, the most anterior pillar being broken away. Measurements of mandibular fragment (mm.) : Menetiiat een aoe: inickitessi ace ies B. ADDITIONAL East AFRICAN SPECIMENS The additional giraffid specimens in the British Museum (Natural History) are mainly fragmentary and are derived from Olduvai Gorge (Tanganyika), Laetolil beds (Vogel River, Tanganyika) and Broken Hill (Northern Rhodesia) (figs. 17a, 17b). Brief descriptions of the Sivatherine specimens only are given. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA (a) OLpuvart GorGE MATERIAL I. Giraffa M 14778: M 14781: M 14702: M 14793: M 14794: M 14795: M 14706: M 14797: M 14708: IT. Sivatheriinae M 14535: M 14779: M 14780: M 14701: Axis. Bed I. Cervical vertebra (7th). Bed I. Calcaneum. Bed I surface. Metacarpal, distal end. Bed I. Metatarsal, distal end. Bed I. Metacarpal, distal end. Bed IT. Metatarsal, distal end. Bed II. Metatarsal, distal end. Bed II, surface. Metacarpal, proximal end. Bed I. Horn core fragment. Bed I. Cervical vertebra (6th). Bed IV. Cervical vertebra (6th). Bed I, surface. Tibia. Bed ITI. M 17024-6: Three molar teeth. ?: Unnumbered left lower molar. G. RK III. III. Giraffid (no generic distinction possible) No numbers on specimens. Letters refer to sites. Astragalus .. Oldy. FLK II 8 — 10] Cited! ~ HEK II 8S a HEK II S — GTS IVS —~ GRK II S Phalanx! ). )...:) Oldy. GHJK IIS — EHK I _ DK. 15s _ HEK II Calcaneum Dy 1 GRK eS Tibia Oldy. FC II S_ (distal end) — EK I S_ (proximal end) _ Sc II S_ (proximal end) — GRK II S_ (proximal end) — DK. ;.1 base Cubonaviculare (Scaphocubeid)), Oldy.. THC 1, (y) — VEK I = SHK II 8 Sas) oa° Metapodial ANNALS OF THE SOUTH AFRICAN MUSEUM Oldy. FC II S_ (distal end) — DET Ss tdistalend} ~ HEK I (distal end) (b) Lartoxit BEps I. Stvatheriinae M 15088: Lower molar (or premolar). M 15089: Upper premolar. M 15090a, b: Lower premolars. IT. Giraffid Cubonaviculare (Scaphocuboid) 1710.5 Lit.A.S. Phalanx 1’ MnxX.S. LIT.AS. Astragalus .. 17108 LIT.A. (c) BRokEN Hii I. Giraffa Ni 2126, ts Il. Sivatherwinae M 12128 .. M 12129 .. III. Giraffid Tibia, distal end. Unnumbered horn-core, cervical vertebra and distal end of a femur. Metacarpal, distal end. (Same number as Giraffa tibia.) Astragalus. Radius, distal end. DESCRIPTION OF SIVATHERINE SPECIMENS M 14535 This fragment of horn core is entered in the record book as being derived from ‘Middle Pleistocene Bed I Olduvai, Tanganyika, Leakey collection 1932". One side is slightly convex in an A’-P direction and has 3 deep, wide grooves (7-8 mm. in breadth) more or less parallel to each other and to the length axis of the horn. The other surface is concave and relatively smooth. Although it is a small fragment it is very similar to the South African specimens, It is typically Sivatherine. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 541 M 14791 This is a complete tibia. Its dimensions are compared with the average of 5 Giraffa camelopardalis from Central Africa in the Musée Royal du Congo Belge in Tervuren. Giraffa M 14791 camelopardalis Total length Ks : 473 639 Proximal end: Transverse 153°5 151°8 A-P bie 138-0 Q2°4 Mid-shaft: Transverse 64°3 70°6 1 a Rds 49°4 55°8 Distal end: ‘Transverse IO1°O 109:0 A-P ee 75°4 76:2 TABLE 66. Measurements (mm.) of M 14791 compared with Giraffa camelopardalis. This again bears out the contention expressed above that Sivatherium olduvaiense is much shorter and has more massive extremities than the modern giraffe. M 17024-26 These 3 teeth of Sivatherium olduvaiense are all derived from Bed II. M 17024 This is a right lower molar, either M, or M,. The posterior end of the posterior pillar has been broken away. It is in early wear and the enamel is very rugose. It is markedly hypsodont and the cingulum bulges slightly. The costae are very prominent, as are also the protostylid and metastylid. The roots are broken off. Measurements (mm.) of M 17024: Maximum breadth .. ise SA ey PS Occlusal breadth, anterior pillar .. 17:8 posterior pillar .. 17-9 Height, lingual ye i) ta AD buccal a Me Se) ne Ord: M 17025 This right upper molar is probably M?. The anterior pillar is partly broken away on the buccal side. The enamel is coarsely rugose. The lingual surface slopes markedly from the base in a buccal direction. The styles are very prominent. The roots are broken off. Measurements (mm.) of M 17025: Maximum length .. 49 Maximum breadth .. 43 Occlusal length SH Wheel oho 542 ANNALS OF THE SOUTH AFRICAN MUSEUM M 17026 This is a left M! or M? in advanced wear. There is no cingulum. There is a marked protostyle and the median costa of the anterior pillar is flattened. Most of the posterior pillar is missing and the 2 buccal roots are broken off. The lingual root forms a broad plate. M 15088-15090 These Sivatherine teeth are from the Laetolil Beds (Vogel River) and recorded as ‘Pleistocene Tanganyika. Leakey collection 1935’. M 15088 and 15090a, b are lower P, and M 15089 is an upper premolar. M 15089 This right upper premolar has its crown fairly well preserved. It is finely rugose and. the lingual surface has a marked slope. Wee median costa is very prominent. The root is broken off. M 15090a Its anterior root is broken away and belongs to the side opposite that of 15090b. M 15090b The anterior pillar is worn down and it projects much more than the posterior pillar. The enamel is very finely rugose. The buccal surface of the anterior pillar is missing and the roots are broken off. Length (A—P) Breadth Height Max. Occlusal Max. Occlusal Buccal Lingual M 15088 BO 36-2 33°0 20°0 30°6 32°5 M 15089 33°2 29'°8 46-7 33°3 29°8 26-0 M 150904 37°8 3555 27°38 — 13-1 21°4 M 150906 36:3 36-2 32°0 28-7 24:6 30°1 TABLE 67. Measurements (mm.). Unnumbered GRK II Olduvai This is a lower left molar, probably M,. It is still embedded in matrix. The stylids are very prominent. Measurements (mm.) : Maximum length .. 47 Maximum breadth .. 49 Occlusal breadth, 7") 38 Height si et a Monee M 14778 This is an axis. The total height is 172 mm. and the total A—P length is 160 mm. The neural canal measures 40 mm. transversely and 38 mm. A—P. The inferior articular surface measures 65 mm. transversely and 77 mm. A—P. MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 543 M 14779 This is a 6th cervical vertebra. The maximum breadth of the body is 132 mm., and the inferior articular surface measures 74 X57 mm. M 14781 This is a 7th cervical vertebra from Bed I at Olduvai. The maximum breadth of the body is 115 mm., the neural canal is 45 mm. in diameter and the inferior articular surface is 104. mm. broad and 76 mm. A-P. The condyle (on superior aspect of the body) measures 67 X 51 mm. M 12128 and M 12129 These are specimens from Broken Hill, Northern Rhodesia. It is not certain whether these are Sivatherine. M 12128. This is a distal end of a metacarpal and measures 117 X 74. mm. M 12129. This is an astragalus and measures 81 X 122 mm. The following giraffid postcranial specimens are from the Olduvai Gorge. They are not numbered and no generic distinction is possible. Only relevant measurements (mm.) are given. (i) Astragalus Max. length A-P ‘Transverse INovsite ).. oe 106 60 70 Oldy FLK IIS... 103 61 65 tl DOs i Weare IOI 54 63 » HEKOILS .. Veit 67 73 oe el EIS Te Sic, 102 63 65 PG Gi UN Sales IOI 62 63 5 GRE IPS... 106 62 70 (ii) Phalanx I : Maximum Shaft Prox. end Distal end length A-P Transv. A-P Transv. A-P Transv. Oldy GHJK IIS .. — 104 44 54 B58 Bus 55 Pee lis * ie 121 — — 56 59 37 53 Pe DI TS. 114 AP 52 Som 59 Ba) 55 Peedi DEN 109 44 46 54 554 a7 50 (111) Calcaneum GRK IIS Maximum length .. 222 Body) tansy) iach height ue aes OO oeee ANNALS OF THE SOUTH AFRICAN MUSEUM (iv) Tzbza Distal end fragment: Oldy FC IIS ‘Transv... 97 ASP hit 86g Proximal end fragment (without fibular style) : Transverse A-—P EK. ALS oy 110 74 SC ELS 105 69 GRKIIS 87 77 DK I Base 87 81 (v) Cubonaviculare (scaphocuboid) Thickness A-P Transverse Oldy, THG, 1.W). 42 76 103 sie MKT Hie 43 77 IOI 3 be FS haar 36 64 64 (vi) Metapodial Distal end fragments: A-P ‘Transverse EQ Se 50 93 DIRS a Sne a Aiea 65 112 HEK I Me Me 57 105 The following giraffid postcranial specimens are from the Laetolil Beds (Vogel River, Tanganyika): (i) Astragalus Max. length A-P ‘Transverse LO ee 113 58 62 PAS AY 8G. 102 58 62 (ii) Phalanx I Maximum Shaft Prox. end Distal end length A-P Transv. A-P Transv. A-P Transv. MnX S_.. 97 39 35 48 49 43 30 LIT.AS .. 104 bie, 445 47 49 39 27 (* denotes exostosis included) (iii) Cubonaviculare (scaphocuboid) Thickness A-P ‘Transverse 1710.5 a 4.7 89 107 LT. Ae: 40 82 105 C. LANGEBAAN (CAPE PROVINCE) S.A.M. 11715 (Pl. 50, c, d) | This is an almost complete right astragalus (talus) now housed in the South African Museum, Cape Town. It is longer than Olduvai 102 and 107, MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 545 but its general appearance is identical to theirs. It belongs to Sivatherium olduvaiense. Maximum proximo-distal length .. erie ae =e Maximum A-P length medially .. “ 71 Maximum A-—P length laterally .. eo 26/67 Maximum breadth proximally... és 84+ Maximum breadth distally . . as ee 74. Maximum articular breadth proximally .. ial Maximum articular breadth distally f. 74 ACKNOWLEDGEMENTS A. FINANCIAL AID One of us (E.L.B.) received a travel grant from the United States Educational Foundation and the Fulbright Plan Organization, and a research grant from the Belgian Fonds National de la Recherche Scientifique (Brussels). The other author (R.S.) received grants from the South African Council for Scientific and Industrial Research and the Dr. C. L. Herman Research Fund, University of Cape Town. In addition, each of the authors received research grants independently from the Wenner-Gren Foundation for Anthropological Research, Inc., New York. We are exceedingly grateful to these organizations for the generosity which enabled all the fossil and recent material to be studied and collected. Furthermore, the Land Rover donated by the Wenner-Gren Foundation to the Hopefield Research Project of the University of Cape Town was extensively used for field work. We wish to record our thanks to the Director (Dr. A. W. Crompton) and Trustees of the South African Museum, Cape Town, who have not only accepted our paper for publication in a special issue of the Annals, but have also subscribed a portion to its cost of publication. The major publication costs have been met by special grants: (a) by the South African Council for Scientific and Industrial Research and by the Council of the University of Cape Town, and (0) by the Belgian ‘Fondation Universitaire’ to the other author (E.L.B.). For this assistance we are extremely grateful. B. Loan oF MATERIAL The following very kindly released material (including type specimens) on loan so that we could study most of the material in Cape Town: Dr. L.S. B. Leakey, Curator of the Coryndon Museum, Nairobi (to whom we are particu- larly grateful for his co-operation in sending all the East African fossils) ; Professor R. A. Dart, University of the Witwatersrand; Dr. A. S. Brink and Mr. J. W. Kitching, Bernard Price Institute for Palaeontological Research, University of the Witwatersrand; Dr. A. C. Hoffman, Director of the Nasionale Museum, Bloemfontein; Mr. B. D. Malan, Director of the Archaeological Survey of South Africa; Mr. J. H. Power, recently Director, and Dr. R. Bigalke, present Director of the McGregor Memorial Museum, Kimberley; Professors 546 ANNALS OF THE SOUTH AFRICAN MUSEUM C. Arambourg and J. P. Lehman of the Muséum National d’Histoire Naturelle, Paris; Dr. E. Colbert of the American Museum of Natural History, New York; and Dr. A. J. Sutcliffe of the British Museum (Natural History). C. DEPARTMENTs VISITED, TECHNICAL Alp, PHOTOGRAPHS The individuals mentioned in ‘B’ are also thanked for giving permission to study material (mainly recent) in their Institutions. In this respect we are also grateful to Professor A. S. Romer and Dr. B. Lawrence of the Comparative Zoology Museum at Harvard University (Cambridge, Mass.); Drs. G. G. Simpson, J. Anthony and D. Carter of the American Museum of Natural History; Dr. A. W. Crompton, South African Museum, Cape Town; Dr. W. D. Turnbul of the Natural History Museum, Chicago; Dr. R. Kellogg, U.S. National Museum, Washington; Dr. E. Schouteden, Musée Royal du Congo Belge, Tervuren. We thank Mrs. R. H. Nichols of the American Museum of Natural History for her assistance and kindness; also Mr. Kenneth Abrahams, a medical student at the University of Cape Town, who, with other students, greatly assisted on field collecting trips. Radiographs were kindly made by Drs. B. and H. Hirschon, Cape Town. We are extremely grateful to Mr. G. McManus, Surgery Department, University of Cape Town, who patiently made most of the magnificent photo- graphs of the specimens, and to Miss L. A. Abrahams who typed the manuscript. Mr. J. N. Darroch, Department of Mathematics, University of Cape Town, gave helpful advice with some of the statistical analyses. Miss A. Schweizer, South African Museum, kindly assisted with the preparation of the plates. REFERENCES Abel, O. 1904. Ueber einen Fund von Sivatherium giganteum bei Adrianopel. Sitzb. Kaiser]. Akad. Wiss. (Wien), Math-Naturw. K1., 113 (1), 639-653. Arambourg, A. 1934. Un nouveau gisement de Libytherium. C.R. Ass. frang. Avanc. Sct. (Paris), 58, 124. ——. 1947. Mission Scientifique de ’Omo (1932-1933). J, fasc. 3, Paléontologie. Paris, Mus. d*Histoire Naturelle. —. 1948. Un Sivathériné nord-africain: Libytherium maurusium Pomel. C-.R.S. Soc. Géol. France, séance du 10 Mai 1948, 178-179. ——. 1949. Les gisement de Veriébrés villafranchiens de l'Afrique du Nord. Bull. Soc. Géol. France, 5 éme série, 19, 195-203. ——. 1952. La Paléontologie des Vertébrés en Afrique du Nord frangaise. XIX Congrés Géologique International. Monographies régionales. Alger 1952, 1-64. —, & Piveteau, J. 1929. Les Vertébrés du Pontien de Salonique. Ann. Paléont. (Paris), 18 (2), 1-82. Bate, D. M. A. 1951. The Mammals from Singa and Abu Hugar. In: The Pleistocene fauna of two Blue Nile sites. Fossil Mammals of Africa, No. 2. London: Brit. Mus. (Nat. Hist.). Bohlin, B. 1926. Die Familie Giraffidae. Palaeont. Sin., Ser. C, 4, fasc. 1. Brain, C. K. 1957. New evidence for the correlation of the Transvaal ape-man bearing cave deposits. Third Pan-African Congress on Prehist., Livingstone, 1955. London: Chatto & Windus, 143-148. ——. 1958. The Transvaal ape-man bearing cave deposits. Transv. Mus. Mem., No. 11. or WW MODERN GIRAFFES AND THE FOSSIL GIRAFFIDS OF AFRICA 547 Brain, C. K., van Riet Lowe, C., & Dart, R. A. 1955. Kafuan stone artefacts in the post- australopithecine breccia at Makapansgat. Nature, 175, 16. Broom, R. 1948. Some South African Pliocene and Pleistocene Mammals. Ann. Transv. Mus., 21 (1), 1-38. Colbert, E. H. 1935. Siwalik Mammals in the American Museum of Natural History. Trans. Amer. Phil. Soc., N.S. 26, 401. Cooke, H. B. S. 1949. Fossil Mammals of the Vaal River deposits. Dept. of Mines, Geol. Survey of the Union of S. Afr., Mem. No. 35 (pt. 3). , & Wells, L. H. 1947. Fossil Mammals from the Makapan Valley, Potgietersrust. III. Giraffidae. S. Afr. 7. Sct., 43, 232-235. Dart, R. A. 1954. The significance of Makapansgat. <. Morph. Anthrop., 46 (2), 119-123. ——. 1957. The osteodontokeratic culture of Australopithecus prometheus. Transv. Mus. Mem., No. tro. Dietrich, W. O. 1937- Die Pleistozane Giraffiden und Bovinen aus Oldoway, Deutsch- OstAfrika. In: Wissenschaft. Ergebnisse der Oldoway Exped. 1913, Dr. H. Reck edit. ——. 1942. Altestquartare Sdugetiere aus der siidlichen Serengeti, Deutsch-OstAfrika. Palaeontographica, 94, A, 43-133. Drennan, M. R. 1954. Saldanha Man and his Associations. Amer. Anthropologist, 56, 879-884. Dreyer, T. F. 1938. The Archaeology of the Florisbad deposits. Arg. Nav. Nasion. Mus., Bloemfontein. Pt. 1 (8), 65-77. —, & Lyle, A. 1931. New Fossil Mammals and Man from South Africa. Dept. Zool., Grey Univ. Coll., Bloemfontein, 60 pp. Ewer, R. F., & Singer, R. 1956. Fossil Carnivora from Hopefield. Ann. S. Afr. Mus., 42 (pt. 4), 335-347- Falconer, H. 1868. Palaeontological Memoirs and Notes. Vol. I, Fauna Antigua Sivalensis. Edit. Ch. Murchison. London, Hardwicke. , & Cautley, P. T. 1836. Sivatherium giganteum. A new fossil ruminant genus, from the valley of Markanda, in the Siwalik branch of the Subhimalayan Mountains. Asiatic Res., 19 (pt. 1). . 1846-49. Fauna antiqua sivalensis (Plates). (Description of the plates: 1868, in Palaeont. Mem., 1). London. Freedman, L. 1957. The Fossil Cercopithecoidea of South Africa. Ann. Transv. Mus., 23 (pt. 2), 121-262. Gaudry, A. 1861. Note sur la Giraffe et l’Helladotherium trouvés a Pikermi (Gréce). Bull. Soc. Géol. France, (2) 18, 587. . 1867. Animaux fossiles et Géologie de I’ Attique. Paris, 2. . 1873. Animaux vertébrés fossiles du Mont Léberon (Vaucluse). Paris. Haughton, S. H. 1922. A note on some fossils from the Vaal River Gravels. Trans. Geol. Soc. S. Afr., 24, 11-16. (The volume relates to 1921, but was published in 1922.) Hooijer, D. A., and Singer, R. 1960. Fossil Rhinoceroses from Hopefield. Zool. Mededelingen Rijksmus. Nat. Hist., Leiden. (In the press.) Hopwood, A. T. 1934. New fossil mammals from Olduvai (Tanganyika Territory). Ann. @ Mag. Nat. Hist. (10) 14, 546-550. —. 1936. New and little-known fossil mammals from the Pleistocene of Kenya Colony and Tanganyika Territory. Ann. @ Mag. Nat. Hist. (10) 17, 636-641. —, & Hollyfield, J. P. 1954. An annotated bibliography of the fossil Mammals of Africa (1742-1950). Fossil Mammals of Africa, No. 8. London: Brit. Mus. (Nat. Hist.). Howell, F. C. 1955. The age of the Australopithecines of Southern Africa. Amer. 7. Phys. Anthrop., 13 (4), 635-662. Khomenko, I. 1913. La faune méotique du village Taraklia district de Bendery. Ann. Géol. et Miner. de la Russie. 15, livraison 4-6. Kormos, T. 1911. Der Pliozane Knochenfund bei Polgardi. Féldtani Kézléni (Budapest), 41 (1-2). Lankaster, E. R. 1907. The origin of the lateral horns of the giraffe in foetal life on the area of the parietal bones. Proc. Zool. Soc., London, 1, 100-115. Leakey, L. S. B. 1951. Olduvai Gorge. Cambridge Univ. Press. . 1958. Some East African Pleistocene Suidae. Fossil Mammals of Africa, No. 14. London, Brit. Mus. (Nat. Hist.). 548 ANNALS OF THE SOUTH AFRICAN MUSEUM Lydekker, R. 1904. On the subspecies of Giraffa camelopardalis. Proc. Zool. Soc. London, 1, 202-227. ——. 1913. Catalogue of ungulate mammals in the British Museum (Nat. Hist.). Vol. I, Artiodactyla, family Bovidae. London, p. 249. Mabbutt, J. A. 1956. The physiography and surface geology of the Hopefield fossil site. Trans. Roy. Soc. S. Afr., 35 (1), 21-58. Matthew, W. D. 1929. Critical observations upon Siwalik Mammals. Bull. Amer. Mus. Nat. Hist., 56, 437-560. Mecquenem, R. de. 1924. Contribution a l’étude des fossiles de Maragha. Ann. Paléont. (Paris), 23, 135-160. Meiring, A. J. D. 1956. The macrolithic culture of Florisbad. Res. Nas. Mus. (Bloemfontein), I (9), 205-230. Oakley, K. P. 1954a. Study tour of early hominid sites in Southern Africa, 1953. S. Afr. Archaeol. Bull., 9 (35), 75-87. —. 1954). The dating of the Australopithecinae of Africa. Amer. 7. Phys. Anthrop., 12 (1), 9-28. Osborn, H. F. 1892. Nomenclature of mammalian molar cusps. Amer. Natur., 26, 436-437. . 1907. Evolution of mammalian molar teeth to and from the triangular type. New York, Macmillan. Owen, R. 1840-45. Odontography, or a Treatise on the comparative anatomy of the teeth. London, Bailliere. . 1849. Notes on the birth of the Giraffe at the Zoological Society’s Gardens, and descrip- tion of the foetal membranes and of some of the natural and morbid prea observed in the dissection of the young animal. Trans. Zool. Soc., 3, 21-28. Pethé, J. 1885. Uber die fossilen Saugethier-Uberreste von Baltavar. Jahresh. K.U. Geol. Anst. f. 1884 (Budapest). Pilgrim, G. E. 1911. The fossil Giraffidae of India. Palacont. Ind., N.S. 4 (1), 1-29. Pomel, A. 1892. Sur le Libytherium maurusium, grand ruminant du terrain pliocéene plaisancier de l’Algérie. C.R. Acad. Sci. (Paris), 115, 100-102. Reygasse, M. 1921. Etudes de Palethnologie maghrebine (nouv. série). Ree. Not. Mem. Soc. Archéol. Constantine, 5 sér., 9 (52) (1919-20), 513-570. Roman, F., & Solignac, M. 1934. Découverte d’un gisement de Mammifeéres pontiens a Daren (Tunisie septentrionale). C.R. Acad. Sci. (Paris), 199, 1649-1650. Schlosser, M. 1921. Die Hipparionenfauna von Veles in Mazedonien. Abhdl. Bayer. Akad. Wissen, 29, 4. Singer, R. 1954. The Saldanha skull from Hopefield, South Africa. Amer. 7. Phys. Anthrop., n.s. 12 (3), 345-362. . 1956. Man and Mammals in South Africa. 7. Palaeont. Soc. India, 1, 122-130. —, & Keen, E. N. 1955. Fossil Suiformes from Hopefield. Ann. S. Afr. Mus., 42 (pt. 3), 169-179. ——, & Crawford, J. R. 1958a. The Significance of the Archaeological Discoveries at Hope- field. 7. Roy. Anthrop. Inst., 88 (pt. 1), 11-19. , & Hoojer, D. A. eee A Stegolophodon from South Africa. Nature, 182, 101-102. Stromer, E. 1907. Fossile Wirbeltier-Reste aus dem Uadi Faregh und Uadi Natraim in Aegyp- ten. Abh. Senckenb. naturf. Ges. (Frankfurt a.M.) 29 (2), 99-132. Van Hoepen, E. C. N. 1932. Voorlopige beskrywing van Vrystaatse Soogdiere. Paleont. Nav. Nas. Mus., Bloemfontein, Dl. 2, 63-65. Wells, L. H., & Cooke, H. B.S. 1957. Fossil Bovidae from the Limeworks Quarry, Makapans- gat, Potgietersrus. Palaeont. Afric., 4, 1-55. (The volume relates to 1956 but was published in 1957.) Ann. S. Afr. Mus., Vol. XLV Plate I a—Old. 2.53, antero-medial aspect. b—Old. 86, antero-medial aspect. c—Old. 1. 53, antero- medial aspect. d, e, f—Old. 2.53, 1.53, 86 respectively, illustrating the described features on their postero-lateral aspects. No scale. Ann. S. Afr. Mus., Vol. XLV Plate II Old. 3.53: a, c—postero-lateral and anterior views, respectively, illustrating features described. No scale. b—antero-medial aspect. Ann. S. Afr. Mus., Vol. XLV Plate III a, b—postero-lateral and antero-medial aspects of M 14955 (Olduvai), respectively. c— Old. 1952 SHK II BK II base (S) plus M 149546, antero-medial aspect. d—idem, postero-lateral aspect. No scale. Ann. S. Afr. Mus., Vol. XLV Plate IV scale em 1 2 Sonennssoncenannaciitennansccneenn tn tis erccnt TPO scale em 1 a Os erennnnnnninntnannnnanrcserennarnnnantennnnatt F 3655 (Olduvai): a—buccal aspect. b—lingual aspect. c—occlusal aspect. Ann. S. Afr. Mus., Vol. XLV Plate V a— Old. 93, buccal aspect. 5—Old. F 3656, buccal aspect. c—Old. 93, lingual aspect. d—Old. F 3656, lingual aspect. Ann. S. Afr. Mus., Vol. XLV Plate VI scale ver Bi a—Old. F 3656, occlusal aspect. b—Old. 93, occlusal aspect. c—Old. 1, buccal aspect. d—Old. 1, lingual aspect. e—Old. 1, occlusal aspect. Plate VII Ann. S. Afr. Mus., Vol. XLV scale om } 2 a Old. 6: a—lingual aspect. b—buccal aspect. c—occlusal aspect. Plate VUI Ann. S. Afr. Mus., Vol. XLV a—buccal aspect. b—occlusal aspect. Old. 365 Ann. S. Afr. Mus., Vol. XLV Plate IX seale cm 1 Old. 392. a—buccal aspect. 5—lingual aspect. c—occlusal aspect. Ann. S. Afr. Mus., Vol. XLV Plate scale cm 1 ae anannnconsccscssespaciods Old. 3. a—buccal aspect. b—lingual aspect. c—occlusal aspect. Ann. S. Afr. Mus., Vol XLV Plate XI Old. 92: a—buccal aspect. b—lingual aspect. c—occlusal aspect. *yoodse yesnjo00 —9 ‘yoodse tensutt—q *joodse yeoonq—v :126 ‘plo TIX #°1d ATX TA “SHY “APY *S “UU *yoodse pesnjo00 — 9 ‘yoodse jensulj—q ‘joodse peoonq—p ‘ows 1WUNSNy TTX #4°1d ATX TSA “SAIL “FV “S “UUYy Ann. S. Afr. Mus., Vol. XLV Plate XIV a— Old. F 2993, buccal aspect. b— Old. F 2993, lingual aspect. c—Old. F 2993, occlusal aspect. d—QOld. 109, buccal aspect. e—Old. 109, lingual aspect. f—Old. 109, occlusal aspect. g—Marsabit Road, buccal aspect. h—Marsabit Road, lingual aspect. i—Marsabit Road, occlusal aspect. Ann. S. Afr. Mus., Vol. XLV Plate AV a—Old. 2, buccal aspect. b—Old. 4, buccal aspect. c—Old. 116, buccal aspect. d—Old. 5, buccal aspect. e—Old. F 2989, buccal aspect. Ann. S. Afr. Mus., Vol. XLV Plate XVI FS BS scale em As in plate 15; lingual aspect. Ann. S. Afr. Mus., vol. XLV Plate XVII As in plates 15, 16; occlusal aspect. Plate XVII Ann. S. Afr. Mus., Vol. XLV . 105, buccal aspect. . surface’, buccal aspect. . b—Old. 95, buccal aspect. c—Old . e—Old, 120, buccal aspect. f—‘Old a—Old. F 2991, buccal aspect d—Old. 166, buccal aspect Ann. S. Afr. Mus., Vol. XLV Plate XIX scale em a—Old. F 2991, lingual aspect. b—Old. 166, lingual aspect. c—Old. 105, lingual aspect. d—Old. 95, lingual aspect. e—Old. 120, lingual aspect. f—‘Old. surface’, lingual aspect. Ann. S. Afr. Mus., Vol. XLV Plate XX GY yy scale cm — 1 2 3 hoertesmerscommmmnastinrimemeascencsemeomnnmtoneritsusecsmmmmed, As in plate 19: occlusal aspect. Ann. S. Afr. Mus., Vol. XLV Plate XXI a—Old. 107, astragalus, anterior aspect. b—Old. 115, radius, articular surface (<— indicates posterior border). c—Old. 101, tibia, anterior aspect. d—Old. 104, cubonaviculare, proximal articular surface. e— Old. 341, os magnum, distal articular surface. {— Old. 342, sesamoid bone, side view. g—Old. 105, cuneiform, proximal articular surface. Ann. S. Afr. Mus., Vol. XLV Plate XXII a—Old. 114, metacarpal, anterior (dorsal) surface. )—Old. F 364, proximal phalanx, posterior surface. c—Old. 185, proximal phalanx, side view. d—‘Old. surface’, proximal phal_nx, side view. Ann. S. Afr. Mus., Vol. XLV Plate XXIII Bale Co 4 z ceanerennneatnncnnnencteccomermecmteh seale enn a—Old. 103, calcaneum, medial aspect. b—Old. 103, calcaneum, lateral aspect. c— Old. 100, femur, distal articular surface. d—Old. 116p, ulna, anterior aspect of proximal end. e, {—Makapansgat M 1801; buccal, lingual aspects respectively. Ann. S. Afr. Mus.,"Vol. XLV Plate XXIV = a—Old. 106, metatarsal, posterior aspect of proximal end. b—Old. 106, metatarsal, anterior aspect of proximal end. c—Old. 103, calcaneum, anterior aspect. d—Old. 314, metatarsal, posterior aspect of distal end. e—Old. 314, metatarsal, anterior aspect of distal end. f—Old. 106, metatarsal, proximal articular surface. g—Old. 104, cubonaviculare, distal articular surface. Ann. S. Afr. Mus., Vol. XLV Plate XXV MMK 3685 (O.F.S.): a, b—buccal and occlusal aspects respectively. Ann. S. Afr. Mus., Vol. XLV Plate XXVI O.F.S.), lingual aspect. O.F.S.), lingual aspect. F.S B .F.S.), lingual aspect. d—C. 426C 4 c—C. 426D ( Ann. S. Afr. Mus., Vol. XLV Plate XXVII a—C. 426A, buccal aspect. b—C. 426B, buccal aspect. c—F. 39 (O.F.S.), X-ray (antero- -) posterior). d—F. 39 (O.F.S.), posterior aspect. e—F. 39 (O.F.S.), buccal aspect. Ann. S. Afr. Mus., Vol. XLV Plate XXVIII seale em ‘1 2 6 a—C. 426A, occlusal aspect. b)—C. 426B, occlusal aspect. c—C. 426C, occlusal aspect. d—C. 426D, occlusal aspect. Ann. S. Afr. Mus., Vol. XLV Plate XXIX Gy Z a—C. 431B, lateral aspect. Note cranial sinuses in base. b—C. 431B, medial aspect. c—C. 431A, anterior view. d—C. 431A, antero-medial aspect. Ann. S. Afr. Mus., Vol. XLV Plate XXX scale em 1 2 3 ‘eeecteeenmmnonieiteteenonsenannneenenneetennnnnndonneemmemeceneemeteell scale em 1 2 4 ‘remceceneeeennensinsnaenecsost enennetnarensereorsasuentnenenecieonearenenmeeetcoesennanaet (O.F.S.), buccal, lingual, occlusal aspects respectively. b, d, f—B? (O.F.S.), buccal, lingual, occlusal aspects respectively. Ann. S. Afr. Mus., Vol. XLV Plate XXXI Makapansgat specimens: a, i, g—M 646, buccal, lingual, occlusal aspects respectively. b, j, r—-M 944, buccal, lingual, occlusal aspects respectively. c, k, s—M 536, buccal, lingual, occlusal aspects respectively. d, 1, t—M 533, buccal, lingual, occlusal aspects respectively. e, m, u-M_535, buccal, lingual, occlusal aspects respectively. jf, n, v—M 1115, buccal, lingual, occlusal aspects respectively. g, 0, w—M 552, buccal, lingual, occlusal aspects respectively. h, p, x -M 551, buccal, lingual, occlusal aspects respectively. Ann. S. Afr. Mus., Vol. XLV Plate XXXII seale em 1 2 & J rvreerorrssnannnseslatsseneonomeneisensiencenr reso Makapansgat specimens: a, b, c—M 525, lingual, buccal, occlusal aspects respectively. d—M 938, buccal aspect. e—M 936, buccal aspect. f—M 1113 plus M 942, buccal aspect. g—M 263, buccal aspect. h—M 528, buccal aspect. Ann. S. Afr. Mus., Vol. XLV Plate XXXII scale em Makapansgat specimens, lingual aspect. a—M 938. b—M 936. c—M 1113 plus M 942. d—M 263. e—M 528. f—-M 532. g—M 1114. h-M 531. i—M 264. Ann. S. Afr. Mus., Vol. XLV Plate XXXIV Makapansgat specimens, occlusal aspect: a—M 938. b—M 1118 plus M 942. c—M 936. d—M 263. e—M 528. f—M 550. Ann. S. Afr. Mus., Vol. XLV Plate XXXV scale em 1 2 5 Makapansgat specimens: a, e—M 532, buccal and occlusal aspects respectively. b, f—M 264, buccal and occlusal aspects respectively. c, g—M 1114, buccal and occlusal aspects respectively. d, h—M_ 531, buccal and occlusal aspects respectively. Ann. S. Afr. Mus., Vol. XLV Plate XXXVI g scale in cm = OE Ds Be Makapansgat specimens: a, b, c—M 527, buccal, lingual, occlusal aspects respectively. d, e, f—M 1116, buccal, lingual, occlusal aspects respectively. g—cast of M 539A, now in B.M.N.H. and numbered M 16729. Bucco-occlusal view. Ann. S. Afr. Mus., Vol. XLV Plate XX XVII scale em 4 Zz g YG UG Vy Makapansgat specimens: a, h, k—M 939, buccal, lingual, occlusal aspects respectively. b, g, i—-M 540, buccal, lingual, occlusal aspects respectively. c, e, j—M 937, buccal, lingual, occlusal aspects respectively. d, f, /—M 524, buccal, lingual, occlusal aspects respectively. Ann. S. Afr. Mus., Vol. XLV Plate XXXVIII “4 scale in cm. SS ee ee Se Makapansgat M539B: a—bucco-occlusal aspect. 5—lingual aspect. c— X-ray. Ann. S. Afr. Mus., Vol. XLV Plate XXX1IX seale em 1 a 4 Makapansgat specimens: a, b—M 553A, buccal (lateral), lingual (medial) aspects respectively. c, d—M 5538, lingual, occlusal aspects respectively. Ann. S. Afr. Mus., Vol. XLV Plate XL scale em 1 os 4s 553A plus M 2087 (joined): a—shows buccal aspect of M 2087 and lingual aspect of 553A (inferior border adjacent to ‘a’). 5—inferior aspect. c—superior aspect. 419, J Pp Pp Pp Ann. S. Afr. Mus., Vol. XLV Plate XLI scale em Leccnsretminininrcencnttarsisastorancnoeiptesatstet Meee M 553B!: a—bucco-occlusal aspect. b—buccal aspect. | Ann. S. Afr. Mus., Vol. XLV Plate XLII << Hopefield specimens: a—4372B, viewed from endocranial aspect. b—4372B, inferior aspect: shows incomplete occipital condyles. c—4373A, side view. d—4373A, opposite side to (c). Ann. S. Afr. Mus., Vol. XLV Plate XLUI Hopefield 4372. a—postero-lateral aspect. b—anterior aspect. c—antero-medial aspect. Ann. S. Afr. Mus., Vol. XLV Plate XLIV Hopefield 4373: a—antero-medial aspect. b—anterior aspect. c—postero-lateral aspect. Plate XLV Ann. S. Afr. Mus., Vol. XLV Hopefield specimens: a, f, g—4025, buccal, lingual, occlusal aspects respectively. b, e, i—4026, buccal, lingual, occlusal aspects respectively. c, d, hR—3345, buccal, lingual, occlusal aspects respectively. Ann. 8S. Afr. Mus., Vol. XLV Plate XLVI Hopefield specimens: a— 4027, buccal aspect. b—4023, buccal aspect. c—4024, buccal aspect. d—4374, buccal aspect. e—4028 plus 4028A (considered as 4028), buccal aspect. Plate XLVII Ann. S. Afr. Mus., Vol. XLV 1 aspect. ingua b) As in plate 46 Ann. S. Afr. Mus., Vol. XLV Plate XLVUI SS \ As in plates 46, 47; occlusal aspect. Ann. S. Afr. Mus., Vol. XLV Plate XLIX scale em A 2 3 bneesemeeeestnncetmineeesenteenanannensbeereentmnmanantenad, scale em bette etree nena never! Hopefield specimens: a, b, c—4029 plus 4029B (considered as 4029), buccal, lingual, occlusal aspects respectively. Ann. S. Afr. Mus., Vol. XLV Plate L Ss = 3 scale in cm. pS Le ee See a—Hopefield 4025, posterior aspect. b—Hopefield 4026, mesial aspect. c—S.A.M. II715 (Langebaan), astragalus, anterior aspect. d—S.A.M. 11715 (Langebaan), astragalus, posterior aspect. e—A.M. 19684 (Siwaliks, India). Mandible of Hydaspitherium sp. By courtesy of the American Museum of Natural History, New York. Ann. S. Afr. Mus., Vol. XLV Plate LI cm. 0 5 10 15 St. Arnaud, Algeria: a—1948-1-2, Mus. d’Histoire Naturelle, Paris. Antero-medial aspect of posterior horn-core of Sivatherium olduvaiense (described by Arambourg, 1948, as Libvtherium maurusium). b—1948-1-1, Mus. d’Histoire Naturelle, Paris. Medial aspect. (By kind permission of Professor C. Arambourg.) Ann. S. Afr. Mus., Vol. XLV Plate LI a—cast of 1950-1-1 from Lac Ichkeul, Tunis. Buccal aspect. (Original in Mus. d’Histoire Naturelle, Paris.) 6—cast of posterior horn-core from St. Arnaud (Ain Hanech). (Original in Mus. d’Histoire Naturelle, Paris.) TTT 1d ‘2691 ‘Jawog wos poonpoidoy ‘joWMOg wmsninpu wnuayidqvT Jo UsuAtDads 9dAq Jo smara (gq) [eon pue (p) yesnjo9O \ WN ST ATX TPA OSU “IV WS ‘UU P's Mal OEE: HY: Ve) VI. Vil. BS Bee XL XI. Rory XIV. XVI. XVII. XVIIT.. SEN OK: KOT: MXIT. XXIII. XXIV. BY) yy SNL. XXVII. XXVIII. XXIX. Kee: - INDEX gs. XX XII. XXXIIL. KTV, 5 B40 XXXVI. XXXVII. XXXVIII. XEN. Laon 4 We XLI. 04 9 OF G A bet Te XLIV. LV. - Current prices: ys 1900-1902 1903-1905 1903-1908 — 1906-1910 1908-1910 ‘1908-1913 “Igii-1918 ‘IQII-1914. 1911-1918 1913-1924 1913-1923 1915-1924 1914-1916 1917-1933, 1917-1920 1921. 1924-1925, 1924-1926 1925-1927 1925-1928 1925-1926 1929-1938 1927-1928 af 1928 1929 1929-1932. 1929-1931 193% 7935 _ of papers, authors and subjects. duane in Nol E-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- Da ek Part D Palacentology — eee ea een Geology, Palaeontology, Zoology, Ani ‘Parts 1,2, 5, 8, Oh: Shi manip Zoology (excl. Part ne Index). ob : i Palaeontology (excl. Parts — Was os Botany (excl. Paris re ae - ¥ itt Dh Daag Zoology (excl. Part 2) rea, pees Zoology (excl. Parts 2, 7, Tey Ra . Palaeontology and Set. vee ‘Archaeology ca Zoology — en de Lee Zoology heater ee ae Oy dc ae . Zoology NOWNOD ATYIV4 LN3Saud (xOuddv) 3DNVONNEY dO S31NO9D3LVD ee, Tvv09 AA ads VIDO0OWNAD r) + S3AOYDNVAN Alam (‘S99 wey Aiferupy YsHtig Uo paseg) Jooy purysy oquioyy) ye fouULYD IeqizueZ oy) Wo Udye) OW DAVM JUIIOIA 0} posodxo jOU wore [e109 B pUL SoAOISULUT UT FULpUNG? IANLIPI Jo uoT}voIpul ue puv spruef}ny jo uonnqmysiq bv ‘og ONVISE UVEIZNVZ aS sziuag ~ ae * “ay Saiua S3a70IW p/\ z er MO138 VV HONOYHL NOILD3S SN3SIS3YIA I YVHOS8 1 sneaaion VNSILSONOW 1 vess 1 SISNSISIVAT ‘Anr SNANINONVS “1 VAWYV13S1A1NI 1 SNLVINDVNILNSOYV 7 1SY3EN3uHa "7 -——— ‘b ‘Sy ut se Ady ‘Jey JouUT oY) JO IVY} O} IeTIUIIS SI VUN] YS s}I se ‘OID POPN[OUI JOU sI JoUULYD 7vOq JY} PUY 4SoID JooI 941 UZIMIOq Jey Joor 193no uy ‘vore [SUIS OU UO poseq pUv II}eUIUIeISEIp sI BJYoId oY], ‘NuIe'T 12 syueq 190j3eM-doop 9y} pu ‘[auULYD }ZOq pur jeP Joor pesayays sit YIM ‘ueIO uUEIpuy usdo oy} 0} pasodxe jJoor [e109 & UO sdUpUNqe 2AT}¥IOI Jo UOTZoIpUr Ue pue spruefynT Jo UOTINIZISIG 9) tom Ton) G -olg ee $I SZ Ce aw eI ———— ° " SWOHLV2 Ni Hid3G o eooOn 79 T3aNNVHD LvOe 49309 43338 “oo - < J a w w w e —=- SNVTILAY'V SNdAL'd EEE Sid 1O¥ IW J 70556000000 SOO RRTTSSSRARSRSCSOSOOISOOOTTOD 3va3as"1 1 SNANINONVWS “1 SSSI BISA ASIST SSCS DOR SRS IE AARAARRT CSR OSTSSE > | SNSDIS3YIA Vv VUINSVH 1 NOTES ON THE BIOLOGY OF THE LUTJANIDAE SNLVANAIY “7 uvHoOe "7 SNLVIO3NI7'7 “ant "Ant , snesig’ VNDILSONOW'T [co RNR OIS IL MLKEEA ERI AN ‘ ’ JO ' ' q ‘ SISN3IDIVA'7 ee ore TOSI REMR APRESS , Vv W W Vv = | 4 | AWN 4 ie | ‘ant SNAVINIVNILN3SSYY ‘I EOD SESS LISSSS STOLE OL LLL OOS OTS IIRL R EOS ISI OO TOOT OTL LE, 564. ANNALS OF THE SOUTH AFRICAN MUSEUM Ripe females were found in March, August, October and December, suggesting an extended breeding season mainly in the north-east monsoon period. Lutjanus gibbus (Forskal) Lutjanus gibbus is a small species, seldom reaching more than 4 lb. in weight, occasionally seen in local fish markets, but seldom in large numbers. One hundred and twenty-one specimens were taken by handlining and trammel-nets, always on the bottom and never in mid-water with L. bohar and A. virescens. The size range of fish taken was from 170 to 355 mm. This species was usually taken at night. It was only found in shallow water of from 3 to 8 fathoms. Underwater observations showed that this species keeps a few inches above the coral, often sheltering in the branches of ‘stag coral’ (Acropora formosa [Dana]), and in the leaves of ‘platform coral’ (Acropora hyacinthus [Dana]). It was seldom seen singly but usually in dense, closely knit shoals, typically roving over the bottom in a single layer, closely following the bottom contours. Numbers ranging from one to fifty fish and over were seen, but usually shoals were from ten to twenty-five. The species is common on exposed and sheltered coral reefs with rich coral growth, and was seen at Tutia, Vulture, and Cumulus banks in the Mafia area, and on the fringing reef outer slope at Zanzibar. It was also taken on Latham bank and at Lamu. Adults were never seen in the fringing reef channels. Juveniles of about 50 mm. long were seen on the reef flat, and are occasionally taken in beach seine hauls on Zanzibar Island. FEEDING Foods eaten were mainly crustaceans, including crabs and Penaeid prawns. Small coral fishes were also occasionally taken. Coral and sand were sometimes present in small quantities. MATURITY AND SPAWNING The smallest mature females with gonads approaching spawning condition (Stage IV) were found at 223, 235, 240 and 245 mm. standard length. The smallest mature males were 240 and 280 mm. Ripe fishes (either sex) were found in March, November, September and December, i.e. the north-east monsoon period. Lutjanus sanguineus (Cuvier and Valenciennes) One hundred and two specimens were taken by handline and basket-trap, ranging from 170-650 mm. (13 lb.). Juveniles were taken in basket-traps in the Zanzibar channel and in the Mafia area, in 6 to 7 fathoms on coral and ; ; | 2 NOTES ON THE BIOLOGY OF THE LUTJANIDAE 565 Cymodocea bottoms, and were also found in the Chukwani fish-ponds (Zanzibar _ Island), a mangrove area. Adults were not taken by basket-trap, or by hand- lining in daylight over coral reefs. On one bank in the Mafia area (Snapper Knoll, Niororo Island) adults were taken on four occasions by handlining at night. Adults and juveniles were also taken by trammel-nets overnight in shallow water in Lamu Harbour. Adults of Z. sanguineus were common in _ 25-47 fathoms off Lamu, and were one of the dominant species in these deep- water catches (mentioned on p. 555). At certain times of year (January, February and March) this species occasionally floods local markets (Zanzibar Island), being taken off the southern tip of Zanzibar Island in 40 fathoms. Off Shimoni (Kenya) it is also taken in quantity at certain times of the north-east monsoon (November to April). L. sanguineus was never observed underwater. When it is caught it is seldom taken singly, but usually a number within a few minutes suggesting a shoaling habit. FEEDING Fishes were the commonest food (including Syngnathus biaculeatus, Monacanthids and Apogonids) but Penaeid prawns, crabs, stomatopods, cephalopods and plankton (salps, doliolids, pteropods and medusae) were also found. No change in diet with size was found. MAtTuRITY AND SPAWNING The smallest males and females with ripe gonads were 480 mm. and 505 mm. respectively. Stage IV females were found in March, August, September and November and Stage V females in April and August. Stage III males were found in March, April, August and November. Lutjanus kasmira (Forskal) Seventy-seven specimens were caught by trammel-net, underwater spearing, and handline, ranging from 125-205 mm. L. kasmira is a small species seldom reaching 3 lb. in weight, and is not caught unless very fine lines and hooks are used. Underwater it is seen to be abundant, often in dense shoals of 25 fish and more, never singly, and usually about actively growing coral in exposed areas. It has been seen off Ras Nungwe (Zanzibar), Tutia Reef, and outside the fringing reef on the Zanzibar east coast, in 2-5 fathoms. One specimen taken in deep water (35 fathoms) off Malindi (Kenya) differs slightly in coloration, scaling, and the number of dorsal spines from the shallow water specimens, and is probably a deep water race. This specimen has been described in a previous communication (Talbot, 1957). The species is not found in sheltered mangrove areas, and has not been seen about the Zanzibar channel reefs. It is occasionally seen in local markets. 566 ANNALS OF THE SOUTH AFRICAN MUSEUM FEEDING Crustaceans were the predominant food and included crabs and amphipods. Squid, fish remains and algae were also found. MATURITY AND SPAWNING Females were mature at the smallest sizes taken, i.e. 125 mm. Males were first seen with sperm at 155 mm., and first recorded as ripe at 165 mm. Ripe fishes were found in March and November suggesting pies in the warm water north-east monsoon period. | Lutjanus sebae (Cuvier) Twenty-seven specimens were taken by handlines (adults), basket-traps and spear-guns (immature fish), ranging from 128 to 665 mm. The juveniles of this species occur in shallow water (5-10 fathoms) and fish up to 360 mm. (34 lb.) have been taken on banks in the Mafia archipelago (Snapper: Knoll near Niororo Island and Sefo Reef). Larger specimens have been found to be fairly common in deeper water of 20-48 fathoms and were taken off Tutia Reef and on the Lanu Banks. Occasional specimens of up to 60 lb. have been seen at Ras Kizimkazi (Zanzibar Island) taken on handlines in 40 fathoms by local fishermen. L. sebae occurs regularly in small quantities in Zanzibar markets. It was never seen underwater by the author but has been seen in 5 fathoms on Sefo Reef by Dr. J. F. C. Morgans (personal communication). FEEDING Stomachs contained fish, stomatopods, crab and cephalopod remains. MATURITY AND SPAWNING Females with developed gonads (Stages IV and V) were only found above 490 mm. standard length. Insufficient data were obtained to estimate size at which males mature. Breeding is in the north-east monsoon period, ripe fish having been found from November to March. Lutjanus monostigma (Cuvier) Eighteen specimens were taken, 15 by underwater fishing and 3 by hand- lining in 1-4 fathoms. Size range was from 275 to 420 mm. This species is very rare in the local markets. Underwater observation, however, showed it to be common in areas where large coral growths form deep shelter, and it seems completely limited to this type of habitat. It was common about Tutia Reef, Latham Island, in the Zanzibar Channel, and was seen occasionally inside the fringing reef on the Zanzibar east coast. It was never seen in shoals although ; NOTES ON THE BIOLOGY OF THE LUTJANIDAE 567 two or three were often seen under one coral shelter. Individuals often remained under one coral for the whole period of observation (up to two hours). FEEDING Fish remains (including one Mullid and one Labrid) were present in most stomachs, and Penaeid prawn remains were also found. MATuRITY AND SPAWNING Ripe or nearly ripe females (Stages [V—V) of 395 mm., 390 mm., 420 mm., and 400 mm. were taken. No ripe males were caught. From the meagre data the fish appear to mature at over 350 mm. (2 lb.). Ripe females were found in November and February. Lutjanus argentimaculatus (F orskal) iEhistcen specimens of 300 to 630 mm. (15% lb.) were taken from abeléewed reef areas in up to 7 fathoms by handlining or set nets at night. This species is abundant in East African coastal waters, and is an important market species. It does not. occur on exposed coral reefs, however, and is therefore not well represented in the E.A.M.F.R.O. catches. It is very common in. shallow mangrove areas and estuaries, and common in sheltered waters such as the Zanzibar channel. It was commonly seen during underwater observation inside the fringing reef off the Zanzibar east coast, in the semi-estuarine waters of Chwaka creek (Zanzibar) and about large, and often dead, coral growths in the Zanzibar channel. It may occur singly or in shoals of up to 20 fish. Juveniles are fairly common in sheltered mangrove areas. | L. argentimaculatus was never seen during underwater observation on exposed coral reefs. MATuRITY AND SPAWNING Males were found with testes containing sperm at 330 mm., 410 mm. and 460 mm., and one ripe male of 515 mm. was taken in November 1958. One female was considered mature at 460 mm., and one De female of 630 mm.. was taken in November 1957. | Lutjanus ehrenbergi (Peters) Six specimens were taken from mangrove pools on Zanzibar Island (Chukwani fish-ponds) from 44 to 98 mm. standard length. Two of the. specimens were fully mature females of 75 mm. These are the smallest mature Lutjanids found during the survey. .L. fulviflamma, also maturing at a relatively small size, was first found mature at 150 mm. L. ehrenbergi was never seen during underwater observation. 568 ANNALS OF THE SOUTH AFRICAN MUSEUM Lutjanus lineolatus (Ruppell) Five individuals of this small species were taken (standard lengths 120-175 mm.) but no biological data were obtained from them. Underwater observation showed this species to be often present in large shoals (30 to over 100 fishes) on exposed coral reefs, often in conjunction with L. kasmira. It is also fairly common in East African markets, and its rarity in E.A.M.F.R.O. catches are due to catching methods. Lutjanus vaigiensis (Quoy and Gaimard) Six specimens were taken by underwater spearing and handlines from 200 to 250 mm. This small species has very occasionally been seen in local markets, and underwater observations shows that it is present, although not common, on shallow coral reefs (Ras Nungwe, Tutia Reef, inside the Zanzibar fringing reef), usually singly, but occasionally in pairs. It is always in or near coral shelter. It is often present in the same areas as L. monostigma. Shoaling in this species has never been seen. Juveniles have been taken in a mangrove area (Chukwani fish ponds, Zanzibar Island). Fully mature males were found at 190 and 205 mm. Pristipomoides microlepis (Bleeker) Twelve specimens were obtained by deep lining (47-67 fathoms) ranging from 260 to 645 mm. (14 Ib.). This species has been taken off Malindi, Pemba, and Tutia Reef. Smith (1954) records it as a major component of the Shimoni (Kenya) deep-water fishery. The species lives in the cold sub-surface water below the thermocline, and has not been taken in shallow water. Pristipomoides typus (Bleeker) One specimen of 525 mm. (84 Ib.) was taken off Tutia Reef in deep water (about 55 fathoms). Smith (1954) records it as occurring in the deep-water Shimoni fishery. Aphareus rutilans Cuvier One specimen, a mature female of 785 mm., was taken at 60 fathoms off Lamu. Water temperature at that depth was 19°C. from a bathythermograph reading. Lutjanus duodecimlineatus (Valenciennes) One specimen was purchased in the Zanzibar fish market (150 mm.). NOTES ON THE BIOLOGY OF THE LUTJANIDAE 569 CONCLUSION Few Lutjanids are restricted to the East African coast. Of the seventeen species recorded here, all except two (L. ehrenbergi and L. duodecimlineatus) reach the Australo-Pacific region. In the reverse direction distribution of this family is not so uniform, however. Many Lutjanids found in the Parific Ocean and the Eastern Indian Ocean are not found in the Western Indian Ocean. For example, in the genus Lutjanus 32 species occur in the Hawaiian Islands (Herre, 1953), excluding the freshwater LZ. maxwebert and the doubtful L. philippinus. Of these only 13 are found in East Africa. This suggests a centre of origin, or at least strong adaptive radiation in the Australo-Pacific region. At present there is clearly difficulty of dispersal in an east-west direction for many species, but whether this is due to paucity of suitable environments in the north and western Indian Ocean, or to some physical barrier to migration is not obvious. Temperature, which limits the southerly distribution of this typically warm-water family down the coast of Africa, does not operate as a barrier to its spread along the Indian and Iran coasts. Lutjanid distribution bears out Ekman’s statement that ‘the rich Indo-Malayan fauna is distributed over a large part of the Indian Ocean, but the number of species constantly decreases as we proceed in a westerly direction’. Down the African coast some species disappear at Delagoa Bay (26°S.), with the last coral reefs, and then there is a steady decrease in species to East London (33°S.), no members of the genus Lutjanus, and only the genera Aprion and Etelis (Indo-Pacific, not recorded during this survey) reaching farther south, to Knysna (34° 5/S.). (Note: Two specimens of L. sanguineus have been taken in Algoa Bay, and one at Plettenberg Bay, 34°S., during 1958-9.) The distribu- tion of this family is paralleled by that of the reef-building corals. Although no coral reef growths have been found south of Delagoa Bay, reef-building genera are found to just north of East London (Stephenson, 1947). Coastal temperatures drop rapidly between Port St. Johns (31° 40'S.) and East London (33°S.) due to an outward turning of the Agulhas current. The latter port coincides with the 20°C. isotherm in winter (Sverdrup et al., 1942). Vertical distribution is very limited in most of the family (see fig. 4). Of the seventeen species taken eight species were never caught deeper than 14 fathoms. Of these most were common about coral reefs, and their deepest limits coincided with the limit of vigorous coral growth, which was usually between 10 and 15 fathoms. It is possible that the same factors might be the cause of both the limit of distribution of the fishes and the end of reef growth, but more likely that the fishes are limited to the coral habitat. The thermocline, varying from 25-50 fathoms, is deeper than the foot of the actively growing coral area, and at 10-15 fathoms water conditions differ little from the surface. Temperatures at this depth were not found below 24°C. by Newell (1957) and in summer are very much above this, so temperature does not seem to be a barrier in this connection. Suggested reasons for the downward limit of coral growth will be discussed in a later paper; here it will 570 ANNALS OF THE SOUTH AFRICAN MUSEUM be sufficient to state that for L. vaigiensis, L. monostigma, L. lineolatus, L. gibbus, L. fuloiflamma and the shallow-water form of L. kasmira (referred to on p. 565), the foot of the living coral reef (excluding tallus slopes) is the downward limit. Although some physical environmental factor or combination of factors may be the cause of this, the abrupt ending coincident with that of the coral suggests that in a downward direction the fish distribution is determined by the latter. L. fulviflamma was not restricted to coral areas but was also abundant in sheltered areas such as mangrove swamp channels and Cymodocea beds. L. argentimaculatus, also limited to shallow water, was found in more sheltered habitats only, including mangrove areas, the boat channel of fringing reefs, and sheltered coral. L. ehrenbergi was taken only in a mangrove swamp. L. bohar and L. rivulatus were for a long period (1951-5) considered to have only a shallow-water distribution, never being taken below about 14 fathoms. In 1956, however, when the deep-water (25-65 fathoms) banks off Lamu, Kenya, were fished it was discovered that these two species were often abundant at much greater depths, reaching 46 and 51 fathoms respectively. In these areas the shelf is unusually wide for the East African coast (considered by Morgans, 1959, to be due to the deposition from an old river delta), and rich feeding-grounds are present. It seems clear that this favourable habitat is the reason for the presence of L. bohar and L. rivulatus in deeper water. On these banks the bottom temperature may vary from 22°C. to 29°C. Newell also gives one reading of 18°C. at 50 fathoms. (Newell 1959, Morgans in unpublished E.A.M.F.R.O. reports.) From the foods taken by these members of the genus Lutjanus it appears that they are all bottom feeders, and it is probable that temperature becomes an important factor in the distribution of these members of this genus to the colder deeper portions of these banks, if we may judge by their distribution down Africa, referred to above. It is quite clear from the great deal of fishing that has now been done in this area that the genus Lutjanus is only found in water of the East African coastal current, and not below the thermocline. No members of this genus have been taken in water below 23°C. where a bathythermograph has been used in conjunction with fishing. In addition to L. bohar and L. rivulatus this also applies to L. sebae and L. sanguineus whose adults were common on these banks. In the former only juveniles and young fishes up to 360 mm. were taken in shallow water, and adults in water of 30-49 fathoms. Adults of the latter extended from the shallow water of Lamu Harbour (a mangrove area of 4 fathoms) to 47 fathoms, and were more abundant in deep water. Juveniles of these species were only taken in sheltered water. In contrast to the genus Lutjanus the two species of the genus Pristipomozdes and Aphareus rutilans were taken in 47-67 fathoms and 60 fathoms respectively and never in shallower water. Also found on the Lamu Banks, these species are usually present below the thermocline. A. furcatus, not taken during this survey but recorded by Smith (1954) in East Africa, seems to be a shallow-water species, being referred to by Randall (1955) as fairly commonly seen underwater NOTES ON THE BIOLOGY OF THE LUTJANIDAE 571 in shallow coral-rich areas of the Gilbert Islands, and taken on the surface on lures by Schultz (1953) on Bikini Atoll. Smith’s record is of one specimen from 20 fathoms off Pemba Island. The genus Aprion has a pelagic habit and a depth range to about 50 fathoms, but is more abundant in shallow water over the reefs. Clear differences were found between the catches of Lutjanids from coral areas facing the open Indian Ocean and exposed to violerit wave action, and those of sheltered coral islands in the 15-mile channel between the African mainland and Zanzibar Island or from the inner Mafia Archipelago (see fig. 1). L. argentimaculatus was taken both in the boat channel and also from sheltered coral reefs, but never on the outer exposed reef slope. L. kasmira, common on the outer slope, was never taken in sheltered water. Numbers of species common to both types of coral area also differed. L. bohar was one of the dominant species of the outer slope, but although present in the more sheltered water it was very much less common. Conversely L. gibbus was common and sometimes abundant in sheltered areas, but on the outer slope it was poorly represented in the catches. The reasons for the patterns of distribution described in the above para- graphs are obviously complex, and not within the scope of this work, but many of the problems here raised could be approached experimentally. Temperature, salinity, light, O, concentration and turbidity preferences of the juveniles and adults of different species could be tested with carefully designed aquarium equipment, especially for the smaller species, and would undoubtedly give valuable clues to the reasons for their distribution. On a typical East African reef the Lutjanids are a major component of the fish fauna. They and the Serranidae form in general the bulk of the non- pelagic predators, in contrast to the Carangidae, Scomberomoridae and Sphyraenidae. They differ from the Serranidae in that while members of this family are usually often solitary and many are more or less stationary for long periods most Lutjanids tend to school and move actively over the coral. The commonest line-caught fish over exposed coral was the large, mainly fish-eating, L. bohar, with A. virescens common in mid-water and at the surface in the same areas. The commonest smaller species, L. fulviflamma, L. kasmira and L. lineolatus, are often present in shoals numbering 50 and more. These species swim close to the coral and are predominantly crustacean feeders. On sandy bottoms and Cymodocea beds their place is taken by the Lethrinidae. In the deeper Lamu Banks off the Kenya coast this family forms the bulk of the predators, if we may judge by the quantities of line-caught fish. Lutjanids, mainly L. bohar, L. rivulatus, and L. sanguineus formed 54% of the fishes taken, the Serranidae 26°% and the Lethrinidae 15%, with 5% of sharks and other species (Williams, 1958). No sharply marked breeding seasons were found in any of the species studied, although sometimes a single sample would contain many ripe fishes of both sexes. In general all species seemed to breed over a large part of the year, but mostly in the warm north-east monsoon period. 572 ANNALS OF THE SOUTH AFRICAN MUSEUM ACKNOWLEDGEMENTS My thanks for help and encouragement are due to the Director, Dr. J. F. G. Wheeler and Mr. Frank Williams, Dr. J. F. CG. Morgans and Mr. B. S. Newell of the East African Marine Fisheries Research Organization, Dr. P. H. Greenwood of the British Museum, and to Professor J. H. Day, University of Cape Town, for helpful criticism of the manuscript. SUMMARY Seventeen species of the family Lutjanidae from the East African coast are discussed, and notes on their distribution, feeding, spawning seasons and shoaling habits are presented. Eight species of the genus Lutjanus were found only in shallow water, and were never found below the limit of active coral reef growth (approximately 15 fathoms). The major thermocline at from 25-50 fathoms is suggested to be a barrier to deeper distribution of the five species (four of the genus Lutjanus and Aprion virescens) found from shallow water to below the living coral reefs. Three species (two of the genus Pristipomoides and Aphareus rutilans) were found only in deeper water, and never above the major thermocline. No evidence of migrations was found. Sheltered coral, and coral exposed to violent wave action were found to differ in the presence or absence of some species of the family, and also in the numbers of species common to both habitats. All the species studied were euryphagous predators. Details of feeding are given. No sharply defined breeding seasons were found, but the extended periods in which breeding took place were mostly in the warm months, November to April. Regular growth rings were found on the scales of Lutjanus bohar. Checks were formed at different times of year in different fishes. It is suggested that these are related to spawning. Growth increments of from 70 mm. (3rd—4th year) to 35 mm. (1oth—11th year) were estimated. REFERENCES ALLEE, W. C., & others. 1949. Principles of animal ecology. Philadelphia: Saunders. Bowers, A. B. 1954. Breeding and growth of whiting (Gadus merlangus L.) in Isle of Man waters. JF. Mar. biol. Ass. U.K., 33, 97-122. CrarKk, J. R. 1958. Consistency of scale reading. Spec. Publ. int. Comm. N.W. Atlantic Fish, 1, 191-192. EKMAN, S. 1953. VI (Part 1, Index), VII (Parts 1-3), VIII, TX (Parts 1, 2), X (Part 2), XI (Parts 25 77 imdex), XM (Part.1), XOSTV (Part 2), SOCK I (Part 1)e Current prices: Vol. HH: III. LN DOI XVI. XVII. XVIII. XIX. XX. XXII. XXII. XXIII. XXIV. XXV. XXVI. XXVIT. XXVIII. XXIX. XXX. INDEX XXXI. XXXII. XXXII. XXXIV. XXXV. XXXVI. XXXVIT. XXXVITI. XXXIX. XL. XLI. XLII. XLITI. XLIV. XLV. 1900-1902 1903-1905 1903-1908 1906-1910 1908-1910 1908-1913 IQI1I—-1918 IQII-1914 IQII—1918 1913-1924 1913-1923 IQI5—-1924 1914-1916 LOI aS 1917-1920 1Q21 1924-1925 1924-1926 1925-1927 1925-1928 1925-1926 1929-1938 1927-1928 1928 1929 1929-1932 1929-1931 193151959 Zoology and Geology (excl. Parts 1-3, 5, 7, 8) Zoology (excl. Part I) Palaeontology Geology, Palaeontology, Zoology, Anthropology (excl. Paris 1,;25.5, 6:59) Zoology Goat Part I, Index) Palaeontology (excl. Parts 1-3) Botany (excl. Parts 1, 2) Zoology (excl. Part 5) Zoology (excl. Parts 2, 7, ican Palaeontology and Geology Archaeology and ae Zoology Zoology Botany Zoology Zoology Zoology Zoology Zoology excl Part 2) Palaeontology Zoology : Anthropology and Ethnology feel Part 2 Zoology : Zoology Anthropology Palaeontology Zoology Zoology of papers, authors and subjects publishes in Volk: I-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- The LIBRARIAN, South Arrican Museum, Care Town. Palaeontology (excl. Part 1) Zoology Zoology Zoology Zoology Zoology i Avehteotone | Zoology Zoology Botany Zoology Palaeontology Zoology and Palueomiplogs Zoology and Palaeontology Zoology and Palaeontology Part 1 Part 2 Part 3 Part 4 Copies may be obtained from— £1 zy nN PNO FH HH NO HH NHWOHRoaA BOA AHO OM OO H ND NWN BHP HF OO KP DYN OO WT DN DO} OO 12 _ Ln | _ OD BIN” SoCo Co) (6) Cnn OO! N) Coico Ny ei Goes Gi No) (SO) Or a a _ DM HNO OT OKO CONC OO) O78) Oy! 10) (C0) (Co 1) Gn Gn Gol (Go) "Got O 6 EG 6) © GC) Gia © © co Gui GGG Go Cyd) 6 OS Cc — ANNALS OF THE SOUTH AFRICAN MUSEUM VOLUME XLV PART VI, containing :— A new Solifugid Arachnid from Table Mountain, Cape. Solpuga grindleyi, sp.n. By A. C. Brown. (With 2 figures in the text.) SUED JULY 1960 _—~PRICE 1s. LIBRARY PRINTED FOR THE TRUSTEES OF THE SOUTH AFRICAN MUSEUM BY THE RUSTICA PRESS (PTY.) LIMITED, COURT ROAD, WYNBERG, CAPE ¢ £ meen i | re r’ sae) ay 7 Zz Rae), “* Wa BNI Ml ly yea 7 y r y 4 1 ” ’ ae ' ; i , A ; , Pa A e 6 ; 14° ; ae CT Are. youre | Lee it) ae a haf id eye , ‘ i 7 j - ‘ r. { i hi hit J n ‘ ' re | x * HY F. a) oor - ~ : “ ‘ x i 4 i j we, y ac i] a - i ; ‘ * f t ; G ~ —_ - in ‘ 4 ‘ ¢ / 1 . . / é 4 ’ i rs , £ a ‘ ; ' — ¥ i} 7 * \ 4 ‘ E- | ’ ‘ i fr / i i i ‘ ‘4 i at f ; rie: , ial i tA url j i = ras Me Un 4 s? % ’ ’ J ' ‘ A - A } A NEW SOLIFUGID ARACHNID FROM TABLE MOUNTAIN, CAPE. SOLPUGA GRINDLEY], SP. N. By A. GC. BRown Koology Department, University of Cape Town Though the Solifugae constitute but a small order of the Arachnida, the group is extremely well represented in the South African region, 196 species being now recorded from the area. In particular southern Africa abounds with species of the family Solpugidae; notably the genera Solpuga and Solpugema, which account for no less than sixty-five of the South African species. It is no surprise, therefore, to find that a new species of Solifugid, recently taken on Table Mountain, in the Cape Peninsula, also belongs to the genus Solpuga. In contrast to the rest of South Africa, Solifugae are rare in the Cape Peninsula and only five species, excluding the new one, are recorded from the area. ‘They are Solpuga fusca Koch (1842), Solpuga monteirot Pocock (1895), Solpugema vincta (Koch, 1842), Blossiola litoralis (Purcell, 1899) and Toreus capensis Purcell (1899). The new species, Solpuga grindleyi, is quite easily distinguishable from each of these, not only in the laboratory but also in the field. In size and in general appearance it is similar to Solpuga fusca, but whereas the latter is a black-legged species, in S. grindleyi the legs are coloured yellow- ochre. The species is very much smaller than Solpuga monieirot and the head- plate is light yellow-ochre whereas in the latter species the cephalic plate is very dark and may be almost black. The new species also lacks the median black abdominal band of Solpugema vincta. Examination of the chelicerae (mandibles) of both male and female shows Solpuga grindley to be distinct from any other species of the genus so far recorded from southern Africa. The dentition of both the upper and lower jaws—not only the position of the teeth but also their number—at once separates it from all the other Cape species including S. ferox Pocock (1895), S. schlectert Purcell (1899) and S. bovicornis Lawrence (1929) as well as the species already mentioned. The new species closely resembles Solpuga fusca but is distinct from that species not only in coloration and in the dentition of the mandibles, but also with regard to the flagellum of the male, that of S. fusca being bzfurcate while in the new species the flagellar apex is entire. The author has two specimens of Solpuga grindleyi in his possession. A male specimen bearing the label ‘Solifugid— Table Mountain’ was discovered among the class-material used in the Zoology Department for teaching purposes, while 575 S.A.M. VOL. XLV. PT. VI. CART ISON) r} AIT HSONIAN Att r g f At ANNALS OF THE SOUTH AFRICAN MUSEUM 576 Fic. 1 Solpuga Grindleyi, sp. n. Holotype (2) x 10. A NEW SOLIFUGID ARACHNID FROM TABLE MOUNTAIN, CAPE 577 an adult female was taken complete with nest on the Gamps Bay side of Table Mountain by Mr. J. Grindley on 20 June 1958. As the female specimen was studied alive and as the male, though well preserved, does not bear an adequate label, it has been decided to designate the female as Holotype for the new species. The male therefore becomes the Allotype. Solpuga grindleyi sp. n. DEsCRIPTION OF HOLOTYPE (Female) Living coloration. Head-plate yellow-ochre with jet-black eyes; chelicerae light yellow-ochre over most of their surface but giving way to dark brown towards their tips. Abdomen dull grey at the sides and between segments, tergites greyish brown; uniformly light grey ventrally. Ventral surface of thorax light yellow, malleoli yellow with dark brown distal margins. Legs for the most part yellow-ochre, tending towards pale yellow on the tibiae and tarsi of the first three pairs. Pedipalps dark yellow-ochre. Setae in general agreeing with the colour of that part of the body on which they occur, but setae on the chelicerae dark brown against their light yellow-ochre background. Apparent proportions. Cephalic plate as broad as long, slightly shorter than chelicerae. Abdomen approximately four times length of cephalic plate and twice as broad, evenly rounded. Palp with both tibia and tarsus club-shaped; setae on pedipalps not arranged in a definite pattern, but several very long setae on each of the last three segments. First pair of walking-legs slender, two-thirds length of palp. Second walking-leg somewhat more robust but shorter than first leg. Third leg stouter than both preceding appendages and equal in length to first leg. Fourth pair of legs slender except for expanded femur, longer than all other appendages, including the pedipalp. Measurements. ‘Total length, 12-4 mm.; greatest (abdominal) width, 3°5 mm. Length of chelicerae, 2:4 mm.; length of head-plate, 2:0 mm.; greatest width of head-plate, 2.0 mm. Length of pedipalp, 9:2 mm.; femur, 3°0 mm.; tibia, 2-5 mm.; tarsus, 2-6 mm. Length of first walking-leg, 6-6 mm.; second leg, 5:0 mm.; third leg, 6-8 mm. Length of last leg, 9:8 mm.; femur, 3-0 mm.; tibia, 2-6 mm.; tarsus plus metatarsus, 3-2 mm. Remarks. The Holotype was taken alive from under a small stone, com- plete with nest. The latter consisted of a shallow depression lined with soft chips of wood and a few small pieces of bark. The specimen was kept with part of its nest in a glass tube for some weeks without feeding. It was then given a number of small beetle larvae, which it readily devoured. Vision is apparently poor, as the animal was not disturbed by movements 15 cm. away from it, except when these movements cast a shadow over it. In the latter case, and when a finger was moved to within 5 cm. of the animal, it reared up into 578. ANNALS OF THE SOUTH AFRICAN MUSEUM the defence attitude common among the Solifugae. The entire body is raised from the ground, the back legs being bent so that the body acquires an angle © : of some 30° from the horizontal, the pedipalps are elevated and held far apart, the first pair of legs is lifted from the ground and pointed forwards and the chelicerae are held agape. The animal orientates itself so as to face continually the source of potential danger. THE MALE SPEcIMEN (Allotype) Colours similar to but somewhat darker than those of the female. Legs and pedipalps tending towards light brown on tibiae and tarsi of all appendages | Fic. 2 Solpuga Grindleyt sp. n. Male Chelicera x 40. and also on the femur of the hind-most leg, though darkish yellow-ochre still predominates. Proportions of parts as for female except that the abdomen is slightly narrower; total body length, 13:6 mm. Body and appendages more setose than in Holotype and the setae themselves very much darker, varying from light brown on the legs to very dark brown on the head-plate and abdomen. Male chelicera as shown in figure 2. Flagellum very long and slender, being approximately twice the length of the chelicera itself; flagellar shaft tapering evenly, without spicules or denticles. Ventral jaw with a large blunt tooth proximally, followed by two less prominent rounded processes; a convex surface leads to the curved and tapered fang. Dorsal jaw with two large rounded teeth preceding the fang-tip. Between the proximal tooth and the articulation A NEW SOLIFUGID ARACHNID FROM TABLE MOUNTAIN, CAPE 579 of the lower jaw occurs, after a short adentate region, a double row of proceses - between which the lower jaw fits when at rest. The inner row of processes consists of a flat-topped projection distally to which are attached four stout elongate plumose setae pointing forwards, followed by three sharp spines. The outer row projects somewhat lower than the inner row and consists of two naked flat-topped processes, the first (distal) process being much larger than _ the second, followed by three sharp spines. Five extremely long, stout, plumose setae are attached immediately behind the base of the flagellum and lie over it, pointing forwards. The distal halves of both jaws are without setae but the proximal part of the lower jaw is quite heavily setose, the setae on the ventral side being simple, those towards the upper surface plumose. Both types have been deposited in the South African Museum. REFERENCES Koch, C. L., 1842. Die Arachniden. Arch. f. Naturg., 8 (1). Lawrence, R. F., 1929. New South African Solifugae. Ann. S. Afr. Mus., 29 (1), pp. 153-179. Lawrence, R. F., 1955. Solifugae, Scorpions and Pedipalpi in South African Animal Life, 1, pp. 152-262 (Uppsala). : Pocock, R. I., 1895. Notes on some of the Solifugae contained in the Collection of the British Museum, with Descriptions of New Species. Ann. Mag. nat. Hist. (6), 16, pp. 74-97. Pocock, R. I., 1897. On the Genera and Species of Tropical African Arachnida of the order Solifugae, with Notes upon the Taxonomy and Habits of the Group. Ann. Mag. nat. Hist (6), 20, pp. 249-272. Purcell, W. F., 1899. New and little-known South African Solifugae in the collection of the South African Museum. Ann. S. Afr. Mus., 1 (3), pp. 381-432. Re AB erik uk Lf Ay eve WAN) aE . er wists “a on, aed TEP on woury inden ie ae : a4 ‘gone es . Shy vine ahinene! pit on ahha thei)! ae et tats SAM Sek ae a netlist Se eULA'E uw Ha: wate vg tne eee Dee vit oh aa ype 8 “ve ee a OTN ELS ET: ieee Clastatn) fh Ht ech + sy Pde thu bes bs th MPR ner prieita Sap { ¥ & ’ : bi + FQN EMT ie ingen, Tg Ate ‘tonal othe, tak | Pied (pgarty aj ie “ 4 b oa OP y ’ ms ; ; Bei dts) EDA ea Oe i NRT 4 | ve at TE aay, - : wa | an it eae a Li 23-7 aheed aa ‘ ‘el % 100 Af a ¥ Bie per PAROS i ‘ irr i yee ’ F me - hae = Bi. 4 De \ } irs t “A : hele =) > i teeth + The ANNALS OF THE SOUTH AFRICAN MUSEUM are issued in parts at irom 4 - intervals as material becomes available. Out of print: VI (Part 1, Index), VII (Parts 1-3), VIII, IX (Parts 1, 2), Vol, 1, “UL: (Parts) 1-955; 92\0), Leb (Part 1) x7 XXI (Part 1), XXIV (Part 2), XXXI (Part 1). Current prices: Vol. II. Dit IV. XXII. yo, Ole XXIII. XXIV. XXV. XXVI. XXVII. XXVIII. XXIX. XXX. INDEX XXXI. XXXII. XXXII. XXXIV. XXXV. XXXVI. XXXVITI. XXXVIII. XXXIX. ole XLI. XLII. XLIII. XLIV. XLV. 1900-1902 1903-1905 1903-1908 1906-1910 1908-1910 1908-1913 IQiI-1918 IQII—1914 IQII—-1918 IQI3—-1924 1913-1923 1915-1924 1914-1916 1917-1933 1917-1920 1921 1924-1925 1924-1926 1925-1927 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, 8) Zoology (excl. Part I) Palaeontology Geology, Palaeontology, Zoology, Anthropology (excl. Parts 4,2,)5, 05.0) Zoology (eter Part I, Index) Palaeontology (excl. Parts 1-3) Botany (excl. Parts 1, 2) Zoology (excl. Part 5) Zoology (excl. Parts 2, 7, indeon Palaeontology and Geology Archaeology and Zoology Zoology ss oe Zoology Botany Zoology Zoology Zoology Zoology Zoology eet Part o Palaeontology Zoology ; Anthropology and Fitnoto ey geet Part be Zoology : if = ig Zoology Anthropology Palaeontology Zoology Zoology of papers, authors and subjects pnblichea in Wolls. I- 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- The LIBRARIAN, Soutn Arrican Museum, Care Town. Palaeontology (excl. Part 1) Zoology Zoology Zoology Zoology Zoology ; Agchatoloiny Zoology Zoology Botany Zoology Palaeontology Zoology and Palacomiglony. Zoology and Palaeontology Zoology and Palaeontology Part 1 Part 2 Part 3 Part 4 Part 5 Copies may be obtained from— Li 2 EN mNHOe ne NDOONDNHDNHOHROaOA A OA A OOO UO © HD NWN DB DP HO HD ND OO SI DN ND OO OO ee) sariee _ et a Mere Nn WDANHN AAT ON CO NwHwWHoND COUN Lol » 25 5, 8, 9), X (Part 2), XI (Parts 2, 7, Index), lop ops.) SOoeoOonwdoddods MOWWTDAMWWWHOODDDDADAMHMWHO OW AHO AWW ow Aw ow Mm Wy ie “ “yy | ie Mt F to ct: IE | 4 I ‘ik “oh ‘u ill il / J | ad ae } “ai mG : « oN “ = fl “Hl oS i a ers: aS é ee < (4 Heb ws: ih 1 W | “i \ Pee catal Ea 1 Pe aire y — a z ales Ps eutat ae Shed AT ~ <4 ‘s veoh Tp! “salt route? & SS a iy Hp’ yi Acta Vy sen sith el ich av y nn ol = Ae yy, t 1s ; Hii tit i" Hy He hy il I “il Bsc f {-+ ll alk” ( Jot ye iy Ann ni He Pan se Py cS Mi “Ar sea EES g, ’ QO, of) ae hit a pate Ge. foes : oy (isos ita” he , / fae " i a Q nN ae A ae "i s i hc sf < fe F 7 . ' ‘ i ¢ oe f hy ; y iF . E eet my rs : “ by x > 4 at - vai Ma Tht th i, ; oe be < Hi iY “ontiior se ATEN me : | é Canaan INSTITUTION LIBRARIES VIII le 3 9088 01206 5793