: ' : : Ns b . ©. ae eae c* eet . “4, 4 seas ' ‘ . ® bd ae 7 a ita 1 = ‘ . > . v . . . = a N oe = ‘ nd sd % eo a Se ad : . : ‘ EE er : . . I won . : 2 oan) a > 4 . : eee ¥ Les ’ Ar ae... ie : pag = ‘ : : . : : ‘ A a] 4 ’ . 4 : en} 2 . “Aa 5 F wW i | . s) v4 . . ms A . . u a aa | vor . . . . . . ‘ “ 1 OP a ea : $ ‘ - . “¢ i J a . af 8 . we n 2 . 7 % . , . eee . a RA ts \ on ae 3 . . ts a ‘ io. tees nr . J > or ‘ . an Sg, e"s . 4 . : . aor Sad Sean, " + 4 ‘ vee h A ; “4 : ’ ar F 5 . ‘ ” Me © WreeQsh = «. (fg = a = oc c = yf UY c < = < 4 «x GY 4 = = ee 5 Be es wz 5 m Fa 7 ty z al > $0 oe ISTITUTION NOILALILSNI NVINOSHLINS S3I14¥VYSIT SMITHSONIAN = - = = r = ca S : z — - a he - Se) = > vie > ~ is > = oe _ 2 E “a Ee es < Es = = o z OD a ty o LIBRA INSTITUTION NOILALILSNI NVINOSHLINS S31IYNVUE = ” Zhe ” z n a = . = =a = < = = = = - zl S z 8 2 2 3 y 8 ‘2 2 E Wo 2 - z = 2 ~~ = = = = SS > = OS = s = NOILALILSNI_ NVINOSHLIWS Saiuvag iol BRARIES SMITHSONIAN us Z wy z 2 Fa = ae ” w = + =< = fad - co = oO aus < = < = < ia a 4 a = ce = = a z ae J z ar = Zz aluvug ryt BRARIES _oMITHSONIAN INSTITUTION NOILNLILSNI~ NVINOSHLIWS S31yVvy = Be = i z 9) = : -e] = re = = es 5 E a = | — Fe pa = ar = m ee ee z : 2. NSTITUTION NOILALILSNI NVINOSHLINS S3IYVYSIT LIBRARIES ~ INSTITU K 2 < = 42 Ge eS ZX. NAS par z = nf z = Zz SS ~ 2 , a = | a 2 a aiuvyug Pia LIBRARI ES SMITHSONIAN _INSTITUTION NOILNLILSNINVINOSHLINS $3 1yvy a . . ve VLILSNI_ NVINOSHLINS S3IYVYSIT_ LIBRARIES SMITHSONIAN \ _————— N\ Sh S! ew es ARIES SMITHSONIAN NOILNLILSNI NOILNLILSNI LIBRARIES INSTITUTION NOILNLILSNI NVINOSHLIWS S3INVUdIT LIBRARIES INSTITUTION INSTITUTION ALILSNI NVINOSHLINS S31YVUGIT_ LIBRARIES SMITHSONIAN oe NVINOSHLINS S31u¥Vvasli SMITHSONIAN NVINOSHLIWS SMITHSONIAN a aie SS ARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINWS NVINOSHLINS LIBRARIES LIBRARIES Sa3iuvugdit LIBRARIES 4 INSTITUTION NOILNLILSNI INSTITUTION NOILNLILSNI saiuvudl S3Iuvuaiy INSTITUTION NOILALILS S3IUVUGIT LIBRARI N' NI ES SJlyvagl SMITHSONIAN NSTITUTION NOILNLI: LIBRAR VYSIT LIBRARIES SI INSTITUTION. NOILNLI NVINOSHLINS S314& saiuvugit NOILALILSNI INSTITUTION NOILNLI SMITHSONIAN INSTITUTION NOILOALILSNI NVINOSHLINS SAIYVUSIT_LIBRAF INSTITUTION = z =| ra = = we 8 . ig = 5 N\ fe) =m - Oo = e) = NX = Le z ad ca i A > = ad \N z a hil = = = = os _NVINOSHLINS S31YVUGIT_ LIBRARIES SMITHSONIAN INSTITUTION | NOILNL é 0 a : 5 z qc WN Ft c cS: ‘ce = + RS NN & ox ec ., par 7 ao. . 2 3° 2 5 = = = ; cee 3 2 : a= > mr ZARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS Sa1uvudi7 LIBRA E 2 = 2 NS OE = =: > = > WS 2 = alg 2 = Fe) = ed w = — x Z o zZ o ee ye & pod! UVUGITLLIBRARIES SMITHSONIAN INSTITUTION, = = = : = < W& = = : : : Z : 9 = : re wo ° wn = O 3 x ve) = fe) = 2 iE 2 i z a 2 a ae 2 2 RARIES SMITHSONIAN _ INSTITUTION ” “n = ees e am ASaag SIN W CK 2 >- . Liideritz Krater \. ° Granitberg @ Elizabeth Bay Fe Pomona \.Bogenfels m \- eEisenkieselklippenbake \.-Advokat Bake © \.-*Buntfeldschuh = — a om an Fig. 1. Map showing the locations of Tertiary marine outcrops mentioned in the text. 1] i2 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 2. Pebble lens in Buntfeldschuh sandstones. Pebbles are predominantly vein quartz, chalcedony, agate and jasper. Fig. 3. Photomicrograph of a greywacke sandstone from Buntfeldschuh. Groundmass is comminuted mica and clay; grains are mostly quartz, with some altered feldspar and rock fragments. EOCENE MARINE SEDIMENTS IN THE SPERRGEBIET 13 Fig. 4. Cross-bedded aeolian sandstones overlying ?upper Paleocene-lower Eocene marine beds at Buntfeldschuh. ‘Diagenetic’-ooid-bearing calcrete caps the section. of calcrete, containing the same unusual diagenetic ooids and intraclasts (Fig. 5) first reported by Siesser (1973) from ?Pliocene—Pleistocene calcretes of the Cape Province. Further details of the Tertiary section at Buntfeldschuh are given by Beetz (1926). Unpublished work by geologists of Consolidated Diamond Mines, South West Africa suggests that the Buntfeldschuh-escarpment section may be more stratigraphically complex than previously realized. Local faulting and tilting in the section allows recognition of at least two marine members. The lower member overlies a remnant of the Pomona beds (here capped by Tafelberg Quartzite) in a small depression at the northern end of the escarpment (C. G. Stocken pers. comm. 1979). 14 ANNALS OF THE SOUTH AFRICAN MUSEUM ye Tis Wim. ‘Diagenetic’ ooids and intraclasts set in an intergranular fabric of microspar and micrite at Buntfeldschuh (see Fig. 4). “SS SS S SAGs SS Fig. 6. Burrow in sandstone at Buntfeldschuh. EOCENE MARINE SEDIMENTS IN THE SPERRGEBIET itp) PALAEONTOLOGY The only macrofossils found in the marine section were burrows (Fig. 6). Fish teeth have been reported by B6hm (1926) and C. G. Stocken (pers. comm. 1979). All samples collected were barren of microfossils. BOGENFELS Tertiary remnants are preserved 3-5 km north and north-east of Bogenfels village in two depressions on either side of the Langental (Fig. 7). The larger exposures are south and east of the Langental. The smaller, but better-known deposits, are on the northern side of the Langental, adjacent to Wanderfeld IV (Fig. 7). Most attention in the past has focused on the smaller exposure, which will be referred to here as the Wanderfeld IV locality, since it is the most fossiliferous exposure and also overlies the only known marine Cretaceous deposit between the southern Cape Province and the Kunene River. The following lithological and micropalaeontological descriptions of this section refer specifically to the Wanderfeld IV outcrops. LITHOLOGY Tertiary outcrops rest on a platform carved into the late Precambrian Bogenfels Formation (mostly dolostone and phyllitic schists in this area), at an elevation of about 70 m. Beetz (1926) gives general characteristics for the Tertiary beds. Klinger (1977) presents a diagrammatic stratigraphic section, and the following description is taken partly from his section, and partly from our own observations. The marine beds are only about 4 m thick. They consist of a pebble layer at the base, which is overlain by concretionary calcareous sandstones and marly sandstones and siltstones. At least one layer of light-brown (5 YR 5/6) ferruginous calcirudite occurs near the bottom of the section. This layer contains _ pebbles, but is predominantly composed of neomorphosed mollusc fragments set in micrite and microspar (Fig. 8). The marly sand- and siltstones are mostly pale greenish yellow (10 Y 7/2), poorly consolidated, and calcareous. The calcareous concretionary layers are greyish yellow (5 Y 7/4) to pale greenish yellow (10 Y 7/2) and are moderately well indurated. All layers contain occasional macrofossils and agates. This section is generally much more calcareous than the Buntfeldschuh section. PALAEONTOLOGY Klinghardt discovered the first fossils at Bogenfels (Haughton 19306). Reuning and Lotz collected from the locality in 1909, and supplied the original fossils studied by BOhm & Weissermel (1913). A Miocene age was initially assigned, based mainly on the molluscs and fish teeth. BOhm (1926) later revised the age to middle to upper Eocene, after obtaining a more extensive collection of fossils made by Kaiser and Beetz during 1914-19 (Haughton 19305). 16 ANNALS OF THE SOUTH AFRICAN MUSEUM However, doubt remained in some quarters as to the age of these beds (see Siesser 1977 for a review). Siesser (1977) eventually confirmed the age as upper Eocene, based on calcareous nannofossils. Macrofossils Bohm & Weissermel (1913), B6hm (1926), and Weissermel (1926) provide an extensive list of the macrofossils found near Bogenfels. Bryozoans, bivalves, ES WF Wil SA WR \ SASS Be Dated Tertiary section (kK) Upper Cretaceous inlier \! Precambrian Quaternary & Tertiary BS) thin sand & gravel Tertiary sand & 1:= 30 000 siltstones Tw N A> Fig. 7. Map showing location of upper Eocene rocks north and north-east of Bogenfels (from Kaiser 1926; Klinger 1977). EOCENE MARINE SEDIMENTS IN THE SPERRGEBIET 17 Fig. 8. Photomicrograph of the upper Eocene mollusc calcirudite at Wanderfeld IV (Fig. 7). Groundmass is micrite and microspar; light grains are quartz; large elongate grains are mollusc fragments neomorphosed te pseudospar. gastropods, a nautiloid, cirripeds, a crab, corals, hydrozoans, fish teeth and Callianassa burrows are all present. Bivalves and gastropods are the most abundant fossils: B6hm (1926) named 9 new species in the former class and 13 in the latter from this locality; fish teeth are also numerous and diverse: Bohm (1926) named 13 new species here. The abundance and diversity of the gastropod Turritella is especially _ striking. This genus litters the ground (Fig. 9) in the area, and prompted Haughton (1930a) to coin the term ‘Turritella-beds’ for the Tertiary outcrops. Microfossils Calcareous nannofossils Moderately to poorly preserved calcareous nannofossils (Figs 10-11) occur in the clayey siltstones. A list of the species identified follows. The relative _ percentage of each species, based on specimen counts in smear slides, is also listed. (These species are all well known and their taxonomy is not controversial ; therefore a systematic palaeontology section is not included.) 70 Braarudosphaera bigelowi (Gran & Braarud) Deflandre . . . . i Braarudosphaera discula Bramlette & Riedel . . . . . ~~. trace 2 LES DIDS SU ee nae rem a eee ey a 1 18 ANNALS OF THE SOUTH AFRICAN MUSEUM O /o Coccolithus eopelagicus (Bramlette & Riedel) Bramlette & Sullivan . 28 Coccolithus formosus (Kamptner) Wise s,s 6 Discoaster saipanensis Bramlette & Riedel lo. i 5 Discoaster tani Bramlette & Riedel . : ; : : : ; é 3) isthmolithus recurvus Weflandre ©: 54 . 3. = ea 3 Reticulofenestra bisecta (Hay, Mohler & Wade) Roth . . . . 37 Reticulofenestra coenura (Reinhardt) Roth 2 ny A ee eee Reticulofenestra umbilica (Levin) Martini & Ritzkowski. . . . 12 Lygrhablithus bijugatus (Dellandre)-Dellandre a 4 The age can be narrowly defined by the range overlap of Discoaster saipanensis and Isthmolithus recurvus. The former ranges from middle to upper Eocene and the latter from upper Eocene to lower Oligocene. They co-occur only in the NP 19 and NP 20 zones of Martini’s (1971) biostratigraphic zonal scheme, and only in the J. recurvus subzone of Bukry’s (1975) zonal scheme. Fig. 9. Various species of Turritella weathered out of the upper Eocene rocks at Wanderfeld IV (Fig. 7). EOCENE MARINE SEDIMENTS IN THE SPERRGEBIET Gr. .., Fas eee Fig. 10. Calcareous nannofossils from the upper Eocene rocks at Wanderfeld IV (Fig. 7). Light photomicrographs. A. Braarudosphaera bigelowi, crossed nicols, scale = 6u. B. Coccolithus eopelagicus, crossed nicols, scale = 9u. C. Coccolithus formosus, crossed nicols, scale = 5u. D. Discoaster saipanensis, plane polarized light, scale = 6p. E. Discoaster tani, plane polarized light, scale = 4u. F. Isthmolithus recurvus, plane polarized light, scale = 5u. G. Reticulofenestra bisecta, crossed nicols, scale = Sp. H. Reticulofenestra cf. R. coenura, crossed nicols, scale = 3u. I. Reticulofenestra umbilica, crossed nicols, scale = 6p. 19 20 ANNALS OF THE SOUTH AFRICAN MUSEUM Vail et al. (1977) show the estimated time boundaries of NP 20-NP 19 as 37,2-39,5 m.y.B.P.; Bukry (1975) estimates the boundaries of the J. recurvus subzone as 38-41 m.y.B.P. Benthic foraminifera Poorly preserved benthic foraminifera occur in small numbers in the clayey siltstones at Wanderfeld IV. This is the first report of in situ upper Eocene benthic foraminifera on the mainland of South or South West Africa. (The ‘upper Eocene’ deposits recorded by Chapman (1930) in the eastern Cape Province are now known to be lower Eocene (Siesser & Miles in press).) The fauna is dominated by Glandulina sp. and Lenticulina spp. Most of the other taxa are represented by only a few specimens, or in some cases, a single specimen. Rather formidable taxonomic problems plague the study of benthic foraminifera under the most favourable conditions (Boltovskoy & Wright 1976; Boltovskoy 1978). These difficulties are accentuated here, since specimens are relatively scarce and have been affected by diagenesis (Fig. 1 1c—f). All specimens have been variously affected by mechanical breakage, solution, recrystallization or test infilling. The variation noted among specimens of some of the better- known species suggests that there may be new species present. However, insufficient numbers of specimens available for study precludes describing them as new species. Because there is little previously published information on southern African Palaeogene benthic foraminifera, and because of the taxonomic problems involved, a section on systematics follows, together with illustrations of the taxa identified at Wanderfeld IV (Figs 11-15). Generic assignments have been made following the classification of Loeblich & Tappan (1964). The preferred modern species name is given, followed by a brief synonymy. Family Textulariidae Textularia sp. Fig. 12A Remarks A single arenaceous specimen from which the initial portion of the test is partly missing and the remainder indistinct; therefore the generic position is somewhat uncertain. The test is slightly elongate, biserial and rhomboidal in section. It is finely agglutinated and fourteen chambers are visible. Apertural details are obliterated. Family Nodosariidae Astacolus sp. Fig. 12B Remarks A solitary specimen with slightly compressed planispiral test, the final chamber of which breaks away to become uniserial. The last chamber is shorter EOCENE MARINE SEDIMENTS IN THE SPERRGEBIET 21 E | | F =—— —— Fig. 11. Calcareous nannofossils and benthic foraminifera from the upper Eocene rocks at Wanderfeld IV (Fig. 7). Scanning electron micrographs. A. Coccolithus formosus, scale = 2. B. Reticulofenestra bisecta, scale = 24. Note dissolution effects on C. formosus and accretion of secondary calcite in the central area of R. bisecta. C-—F. Test infilling and test dissolution in Lenticulina. Differential test dissolution is displayed in C (SAM-K5538) and D (SAM-— K5539). The umbo and sutures are more resistant to dissolution and remain prominent. Dissolution eventually produces an internal mould as shown in E (SAM-K5540) and F (SAM-KS5541). Scale = 100z. aD ANNALS OF THE SOUTH AFRICAN MUSEUM in height, but broader than the previous chambers. Twelve chambers are visible and the sutures are flush to very slightly raised. Lenticulina simplex d’Orbigny, 1839 Fig. 12C—D Robulus simplex (d’Orbigny) Cushman & Laiming, 1931: 98, pl. 10 (fig. 5a—b). Remarks A few well-preserved specimens occur, with thick, prominent umbos and keels. Lenticulina cf. L. pseudo-mamilligerus (Plummer, 1926) Fig. 12E-F Robulus pseudo-mamilligerus (Plummer) Cushman, 1951: 13, pl. 4 (figs 1-5). Remarks The illustrated specimen resembles the species figured by Cushman (1951) in test outline and in suture and chamber disposition. Lenticulina cf. L. oblonga (Coryell & Howe, 1930) Lenticulina cf. L. oblonga (Coryell & Howe) fide Fairchild, Wesendunk & Weaver, 1969: 42, pl. 6 (fig. 4a—b). Remarks A poorly preserved eight-chambered specimen was found. The keel is present on the early chambers only, the remainder unpreserved. Lenticulina subalata Reuss, 1854 Fig. 13A-B Cristellaria subalata Reuss, 1854: 68, pl. 25 (fig. 13). Chapman, 1926: 65, pl. 4 (igs 19a—b, 25a—b, 26a—b). Remarks Two specimens of this species were found. This form has been noted from the Aptian by Reuss and from the upper Eocene of New Zealand by Chapman (1926). Lenticulina spp Remarks Twelve very badly preserved specimens were found. None could be definitely identified to species level. EOCENE MARINE SEDIMENTS IN THE SPERRGEBIET 23 < c gies Fig. 12. Benthic foraminifera from the upper Eocene at Wanderfeld IV (Fig. 7). Scanning electron micrographs. All scales = 100u, unless otherwise noted. A. Textularia sp. SAM-— K5520. B. Astacolus sp. SAM-K5521. C. Lenticulina simplex, side view, SAM-K5522, scale = 300u. D. Lenticulina simplex, peripheral view, SAM-K5522. E. Lenticulina cf. L. pseudo-mamilligerus, side view, SAM-K5523. F. Lenticulina cf. L. pseudo-mamilligerus, peripheral view, SAM-K5523. 24 ANNALS OF THE SOUTH AFRICAN MUSEUM Family Glandulinidae Siphoglobulina ? sp. Fig. 13C-D Remarks A single specimen appears referable to this genus, but the state of preserva- tion does not allow an accurate diagnosis. The test is subfusiform and circular in cross-section. Chambers are strongly overlapping, the sutures depressed and the specimen has a terminal radiate aperture. In side view the specimen appears similar to Marginulina, a distinguishing feature of this genus noted by Parr (1950) in his description of the genoholotype Siphoglobulina siphonifera. The genus ranges from lower Tertiary to Holocene. Family Polymorphinidae Glandulina sp. Fig. 13E Remarks The most abundant form present in the Bogenfels section. Chambers and sutures are indeterminate because of poor preservation. The terminal, radiate aperture, however, is often preserved. | Family Elphidiidae Elphidium cf. E. crispum (Linné, 1758) Fig. 13F Nautilus crispus Linné, 1758: 709. Polystomella crispa (Linné) Brady, 1884: 736, pl. 110 (figs 6-7). Elphidium crispum (Linné) Barker, 1960: 220, pl. 110 (figs 6-7). Remarks | A single, broken specimen was found. Although broken, the features of the central boss, chambers and sutures suggest E. crispum, rather than the closely related form E. macellum. Elphidium sp. A Fig. 14A—-B Remarks A small Elphidium with subcircular outline has twelve chambers and raised. sutures bridged by retral processes. The retral processes (four or five present) of the earlier parts of the whorl are confused, giving a pitted appearance to the test. The aperture consists of a number of pores at the base of the apertural face. This specimen is very similar to E. saginatum Finlay, but differs in not having a depressed central area. EOCENE MARINE SEDIMENTS IN THE SPERRGEBIET 25 E ess F cones, Fig. 13. Benthic foraminifera from the upper Eocene at Wanderfeld IV (Fig. 7). Scanning electron micrographs. A. Lenticulina subalata, side view, SAM-K5524. B. Lenticulina subalata, peripheral view, SAM—K5525. C. Siphoglandulina? sp., side view, SAM-—K5526. D. Siphoglandulina? sp., front view, SAM-K5526. E. Glandulina sp. SAM-K5527. F. Elphidium cf. E. crispum side view, SAM--K5528. All scales = 100z. f ‘ ; } | | 5 26 ANNALS OF THE SOUTH AFRICAN MUSEUM Elphidium sp. B Fig. 14C—D Remarks This well-preserved specimen has a flattened test with near-parallel sides. Numerous chambers with raised sutures and retral processes are present, but detail of the earliest chambers of the whorl is confused because the specimen is broken. The umbilical area is slightly depressed and filled with calcareous nodules. Small papillae cover the chambers and the apertural face. The aperture consists of four pores in a circular position at the base of the apertural face and three pores in the areal position in the lower half of the apertural face. The specimen appears somewhat similar to Discorotalia tenuis, but trochospiral coiling cannot be confirmed as the specimen is broken. It also differs from D. tenuis by having the typical circular apertures of Elphidium. Family Nonionidae Nonion costiferum (Cushman, 1900) Fig. 14E-F Nonion costiferum (Cushman) fide Rau, 1964: 16, pl. 5 (fig. 5). Remarks Three specimens of this species were identified. The species is notable in having raised sutures and a depressed umbilical region. The specimens from Wanderfeld IV differ from those illustrated by Rau (1964) in possessing fewer chambers. The previously reported occurrence of this species in the U.S.A. is near the Oligocene-Miocene boundary (Rau 1964). Nonion sloanii (d’Orbigny, 1839) Fig. 15A Nonionina sloanii d’Orbigny, 1839: 68, pl. 6 (fig. 18). Nonion sloanii (d’Orbigny) Cushman, 1930: 9, pl. 3 (figs 6-8). Remarks This species is represented by a single specimen. It has been recorded in sediments of Eocene to Holocene age. Family Discorbidae Valvulineria aegyptina Le Roy, 1953 Fig. 15B Valvulineria aegyptina Le Roy, 1953: 53, pl. 9 (figs 21-23). Remarks The specimen found differs from the type by having dorsal chambers which overlap strongly, making only eight chambers visible. Apertural features EOCENE MARINE SEDIMENTS IN THE SPERRGEBIET oi | E SE | Fig. 14. Benthic foraminifera from the upper Eocene at Wanderfeld IV (Fig. 7). Scanning electron micrographs. All scales = 100u, unless otherwise stated. A. Elphidium sp. A, side view, SAM-K5529. B. Elphidium sp. A, peripheral view, SAM-—K5529, scale = 30u. C. Elphidium sp. B, side view, SAM-—K5530. D. Elphidium sp. B, peripheral view, SAM-K5530. E. Nonion costiferum, side view, SAM-—-K5531. F. Nonion costiferum, peripheral view, SAM-K5532. ; 28 ANNALS OF THE SOUTH AFRICAN MUSEUM are not distinct as the apertural flap is broken; therefore the full extent of the aperture is not known. Other features are typical of this species. A coarsely punctate wall occurs on both dorsal and ventral sides. A typically imperforate area occurs on the lower ventral part of the final chamber. The final chamber is inflated and extends over, and covers part of, the central umbilical region. The dorsal side is flat, the ventral side convex and umbilicate. This species is recorded from the Eocene in Egypt. Family Rotaliidae Ammonia cf. A. beccarii (Linné, 1767) Nautilus beccarii Linné, 1767: 1162. Rotalia beccarii (Linné) Cushman, 1928: 104, pl. 15 (figs 3-7). Ammonia beccarii (Linné) Cifelli, 1962: 119, pl. 21 (figs 1-6). Remarks This genus occurs in two forms at Bogenfels. A. beccarii has a ventral umbilical area which lacks an umbilical boss, but this is not unusual in this species. Ammonia sp. Remarks Broken specimens occur which were impossible to identify because of poor preservation. The features visible are typical of the genus. The umbilical area is closed but a small protruding plug is present. This is not the typical plug found in A. beccarii; the plug appears as a nodule of material on the umbilical covering. Pararotalia inermis (Terquem) emend. Le Calvez, 1949 Fig. 15C-D Rotalia inermis Terquem, 1882: 68, pl. 6 (fig. la—c). Pararotalia inermis (Terquem) emend. Le Calvez, 1949: 32, pl. 3 (figs 54-56). Remarks Badly broken specimens of this species were found. The tests are eroded into the typical mushroom shape in side view. Perforate chambers are separated by flush, oblique, imperforate sutures, and an imperforate central area occurs on the dorsal side. The umbilical plug has been eroded out by removal of some or all the chamber flaps. This species was originally described by Terquem (1882) from the middle Eocene of the Paris Basin. McMillan (1974) described a closely related form, Pararotalia cf. P. inermis, from Holocene sediments on the Agulhas Bank. EOCENE MARINE SEDIMENTS IN THE SPERRGEBIET 29 ‘ § 2. : t J ES Fig. 15. Benthic foraminifera from the upper Eocene at Wanderfeld IV (Fig. 7). Scanning electron micrographs. A. Nonion sloanii, side view, SAM-—K5532. B. Valvulineria aegyptina, dorsal view, SAM-—-K5533. C. Pararotalia inermis, dorsal view, SAM-—K5534. D. Pararotalia inermis, ventral view, SAM—K5535. E. Cibicides pseudoungerianus, dorsal view, SAM—K5536. . F. Cibicides sp., peripheral view, SAM—-K5537. All scales = 100u. . 30 ANNALS OF THE SOUTH AFRICAN MUSEUM Family Cibicididae Cibicides pseudoungerianus (Cushman, 1922) Fig. 1SE Truncatulina pseudoungeriana Cushman, 1922: 97, pl. 20 (fig. 9). Cibicides pseudoungerianus (Cushman) Cushman, 1931: 123, pl. 22 (figs 3-7). Remarks A species that ranges from the Eocene to Holocene; only a single specimen was found. Cibicides sp. Fig. 15F Remarks This specimen is not well preserved, but has features similar to the form Cibicides sp. illustrated by Van Hinte (1963), viz. a central raised area on the dorsal side. Cibicides spp Remarks A number of poorly preserved forms were found which fit the basic description of this genus. . PALAEOECOLOGY Members of the nannofossil genus Braarudosphaera are virtually absent in open-ocean deposits, and are thus strong indicators of nearshore environments; Zygrhablithus bijugatus is also most common in nearshore waters. These taxa make up a small percentage only of the nannoflora in the samples studied, but their presence in even minor amounts suggests deposition in a shallow-water environment. The benthic foraminifera also suggest a normal marine, shallow- water environment. The nannoflora may in general be considered a temperate-water assemblage. However, Isthmolithus recurvus is a cool-water nannoplankter. Miliolid and peneropolid foraminifera, which are normally common, are conspicuously absent among the benthic foraminifera, and this also suggests that ‘cool— temperate’ may be the most appropriate designation for the water mass overlying this site during the upper Eocene. The benthic foraminifera are not closely age diagnostic, but do corroborate a Palaeogene age for the Wanderfeld IV beds. STRATIGRAPHIC STATUS OF THE TERTIARY DEPOSITS Early workers assumed age equivalency among the scattered Tertiary rocks cropping out between Buntfeldschuh and the Bogenfels—Granitberg area. Haughton (1930qa) first expressed doubt, on palaeontologic grounds, noting that the fish teeth reported at Buntfeldschuh do not prove these beds are con- EOCENE MARINE SEDIMENTS IN THE SPERRGEBIET 3] temporaneous with the Bogenfels exposures. Bohm (1926: 55), refers to abundant fish remains at Buntfeldschuh in the introductory remarks to his paper on the Bogenfels fossils. But he does not designate any teeth specifically from Buntfeld- schuh in the text of his paper, and his summary table on p. 86 lists teeth from Bogenfels only. Based on published reports, the numerous geologists who have subsequently visited Buntfeldschuh have failed to find a single marine body fossil (e.g. Haughton 1930a; Stocken 1962; Ziegler 1969; Martin 1973; Siesser 1977), although C. G. Stocken (pers. comm. 1979) has confirmed the presence of fish teeth in the pebble layers. The Buntfeldschuh section is therefore considered to be essentially unfossiliferous, and in marked contrast to the highly fossiliferous sections near Bogenfels. Nor can the two localities be correlated on lithologic grounds: dissimilarities are as apparent as similarities (cf. Beetz 1926; Klinger 1977; and the descriptions in this paper). In fact, the chief basis for age correlation seems to be that both contain lenses of agates and other pebbles (Beetz 1926; Kaiser 1926; Martin 1973). Against this, Haughton (1930a) points out that the Buntfeldschuh beds contain pebbles and boulders of the Pomona beds (pre-Oligocene, possibly Cretaceous (Stocken 1962)), whereas the exposures near Wanderfeld IV do not (Haughton 1930a). Descriptions by Beetz (1926) also suggest lithologic similarity among the various non-fossiliferous outcrops in the area, and among the fossiliferous outcrops, but less similarity between the two. A significant point is the relative elevations of the fossiliferous and non-fossiliferous beds. The base of the exposures at Buntfeldschuh lic 65 m above those at Wanderfeld IV, and those at Advokat Bake and Eisenkieselklippenbake lie up to 100 m above Wanderfeld IV (Kaiser 1926; Haughton 1930a; Stocken 1962). Haughton (1963) has already suggested that there may be an age difference between the higher and lower Tertiary deposits in this area. He suggested that the higher deposits may be Eocene and the lower may be Miocene, presumably believing, with Du Toit (1954), that the original age (Miocene) assigned by Bohm & Weissermel (1913) to the lower deposits was more likely to be correct. The writers also think the higher, non-fossiliferous deposits are probably older than the lower, fossiliferous deposits. But there is now an unequivocal date for the lower deposits: upper Eocene. Therefore the higher deposits may be upper Paleocene-lower Eocene. This suggestion is based on evidence for a late Paleocene—early Eocene transgression of the Cape Province south coast recently documented by Siesser & Miles (1979). Moreover, Siesser & Dingle (1979) have suggested that this transgression was higher and more extensive than the late Eocene transgression around the coast of southern Africa. The writers frankly admit that assigning the Buntfeldschuh outcrops to the upper Paleocene—lower Eocene is based on circumstantial evidence. But it is no more speculative than to assign them to the middle or upper Eocene, or to the Miocene for that matter, as has been done in the past. In fact, what little evidence there is for the age of these unfossiliferous beds (viz. topographically higher than the known upper Eocene deposits; known major late Paleocene-early Eocene 32 ANNALS OF THE SOUTH AFRICAN MUSEUM transgression in southern Africa) supports the age assignment suggested here. Whatever the age relationship of these two units, their lithologic dis- similarity precludes their inclusion in a single lithostratigraphic unit. The generally higher, and essentially non-fossiliferous beds are best exposed at Buntfeldschuh, and this can be regarded as the type locality for this unit. A ‘Buntfeldschuh Formation’ is thus potentially available, but is not herein proposed. SACS (1977: 21) has set out strict requirements for establishing a formal lithostratigraphic unit, and the writers do not have sufficient data on this unit to describe a stratotype. However, the 1977 South African Stratigraphic Code makes clear provision for the erection of informal lithostratigraphic units in such cases. Thus it is proposed that this unit be informally termed the ‘Buntfeldschuh beds’. This name can be upgraded to the ‘Buntfeldschuh Formation’ if and when the requirements for formal status can be met. Topographically lower and richly fossiliferous outcrops occur near Bogenfels, with the best-known exposures at Wanderfeld IV. This unit is also a potential formation, but again cannot be formally proposed in this paper for the same reason: insufficient data available to satisfy requirements for erection of a formal stratotype. It is proposed that this unit also be given informal status. As with a formal unit, an informal unit should carry a geographic name. ‘Bogenfels’ is unsuitable, as it already is used for the Precambrian dolostone formation in the area. Similarly, “Wanderfeld IV’ has already been pre-empted for the Cretaceous informal lithostratigraphic unit occurring in the same area (Klinger 1977). The writers suggest that this unit be informally called the ‘Langental beds’, referring to the long valley adjacent to which the best exposures occur. This name can later be upgraded to the ‘Langental Formation’. ACKNOWLEDGEMENTS Consolidated Diamond Mines of South West Africa Ltd kindly granted permission and provided facilities for one of us (WGS) to visit and sample the outcrop areas. Dr C. G. Stocken and Mr D. Minney of Consolidated Diamond Mines are thanked for their efforts in making the trip a success. The senior author is grateful to the South African Committee for Stratigraphy who provided funds for his flight to Oranjemund. The South African Geological Survey also provided additional samples, which were originally collected by Dr H. C. Klinger, from the area. The senior author began this project while a member of the Department of Geology, University of Cape Town and completed it after he joined the Palaeontology Department, South African Museum. Both organizations are thanked for their support. We wish to thank the Director of the Geological Survey for permission for one of us (DS) to publish. Scanning electron micrographs were taken in the Electron Microscope Unit, University of Cape Town. Dr C. G. Stocken read the manuscript and offered useful comments. EOCENE MARINE SEDIMENTS IN THE SPERRGEBIET 33 REFERENCES BARKER, R. W. 1960. Taxonomic notes on the species figured by H. B. Brady in his report on the Foraminifera dredged by H.M.S. ‘Challenger’ during the years 1873-1876. Am. Ass. Petrol. geol., Spec. Publ. 9: 1-238. Brapy, H. B. 1884. Report on the Foraminifera dredged by H.M.S. ‘Challenger’ during the years 1873-1876. Rep. Sci. Res. Explor. Voyage H.M.S. Challenger (Zool) 9: 1-814. BEETZ, W. 1926. Die Tertidrablagerungen der Kiistennamib. Jn: KAISER, E. ed. Die Diamanten- wiiste Siidwestafrikas 2: 1-54. Berlin: Dietrich Reimer (Ernst Vohsen). Boum, J. 1926. Uber Tertiaire Versteinerungen von den Bogenfelser Diamantfeldern. Jn: KaAIsErR, E. ed. Die Diamantenwiiste Siidwestafrikas 2: 55-87. Berlin: Dietrich Reimer (Ernst Vohsen). Boum, J. & WEISSERMEL, W. 1913. Uber Tertidre Versteinerungen von den Bogenfelser Diamantfeldern. Beitr. geol. Erforsch. dt. Schutzgeb. 5: 1-58. Bo.tovskoy, E. 1978. Late Cenozoic benthonic foraminifera of the Ninetyeast Ridge (Indian Ocean). Mar. Geol. 26: 139-175. BoLtovskoy, E. & WRIGHT, R. 1976. Recent Foraminifera. The Hague: W. Junk. Bukry, D. 1975. Coccolith and silicoflagellate stratigraphy, northwestern Pacific Ocean, Deep Sea Drilling Project, Leg 32. In: Larson, R. L. et al. Initial Reports of the Deep Sea Drilling Project 32: 677-701. Washington: United States Government Printing Office. CHAPMAN, F. 1926. The Cretaceous and Tertiary foraminifera of New Zealand. Palaeontol. Bull., Wellington 2: 1-119. CHAPMAN, F. 1930. On a foraminiferal limestone of Upper Eocene age from the Alexandria Formation, South Africa. Ann. S. Afr. Mus. 28: 291-296. CIFELLI, R. 1962. The morphology and structure of Ammonia beccarii (Linné) Contr. Cushman Fdn foramin. Res. 13 (4): 119-126. CUSHMAN, J. A. 1922. The Foraminifera of the Mint Spring calcareous marl member of the Mariana Limestone: Prof. Pap. U.S. geol. Surv. 129-E: 123-143. CUSHMAN, J. A. 1928. On Rotalia beccarii (Linné). Contr. Cushman Lab. foramin. Res. 4 (4): . 103-107. CUSHMAN, J. A. 1930. The Foraminifera of the Atlantic Ocean. Bull. U.S. natn. Mus. 104 (7): 1-79. CUSHMAN, J. A. 1931. The Foraminifera of the Atlantic Ocean. Bull. U.S. natn. Mus. 104 (8): 1-179. CUSHMAN, J. A. & LAIMING, B. 1931. Miocene Foraminifera from Los Sauces Creek, Ventura County, California. J. Paleont. 5: 79-120. CUSHMAN, J. A. 1951. Paleocene foraminifera of the Gulf Coastal Regions of the United States and adjacent areas. Prof. Pap. U.S. geol. Surv. 232: 1-75. D’OrsiGny, A. D. 1839. Histoire. Physique, Politique et Naturelle de I’Ile de Cuba. Paris: _ A. Bertrand. Du Tort, A. L. 1954. The Geology of South Africa. 3rd ed. Edinburgh: Oliver & Boyd. FAIRCHILD, W. W., WESENDUNK, P. R. & WEAVER, D. W. 1969. Eocene and Oligocene foramini- fera from the Santa Cruz Mountains, California. Univ. Calif. Publs geol. Sci. 81: 1-141. HaucGuton, S. H. 1930a. Note on the occurrence of Upper Cretaceous marine beds in South- West Africa. Trans. geol. Soc. S. Afr. 33: 61-63. HAUGHTON, S. H. 19305. On the occurrence of Upper Cretaceous marine fossils near Bogenfels, S.W.Africa. Trans. R. Soc. S. Afr. 18: 361-365. HauGuton, S. H. 1963. Stratigraphic history of Africa south of the Sahara. Edinburgh: Oliver & Boyd. Kaiser, E. ed. 1926. Die Diamantenwiiste Siidwestafrikas. 2 vols. Berlin: Dietrich Reimer (Ernst Vohsen). -Kuincer, H. C. 1977. Cretaceous deposits near Bogenfels, South West Africa. Ann. S. Afr. Mus. 73: 81-92. Le CALveEz, Y. 1949. Révision des foraminiféres lutétiens du Bassin de Paris; Rotaliidae et familles affines. Mém. Serv. Carte géol. dét. Fr. 2: 1-32. Leroy, L. W. 1953. Biostratigraphy of the Maqfi Section, Egypt. Mem. geol. Soc. Am. 54: 1-73. LINNE, C. 1758. Systema Naturae. 10th ed. Stockholm. LINNE, C. 1767. Systema Naturae. 12th ed. Leipzig. 34 ANNALS OF THE SOUTH AFRICAN MUSEUM LogsLicH, A. R. & TAPPEN, H. 1964. Foraminifera— Morphology and Biology. Jn: Moore, R. C. ed. Treatise on Invertebrate Paleontology, Part C Protista 2: 58-134. Kansas: Geological Society of America, University of Kansas Press. MartTIn, H. 1973. The Atlantic margin of southern Africa between latitude 17° South and the Cape of Good Hope. Jn: NAIRN, E. M. & STEHLI, F. G. eds. The Ocean basins and margins 1: 277-299. New York: Plenum. MartTINi, E. 1971. Standard Tertiary and Quaternary calcareous nannoplankton zonation. In: Farinaccl, A. ed. Proc. II Planktonic Conf. Roma 1970: 739-785. Roma: Edizioni Technoscienza. ~ McMILLAN, I. 1974. Recent and Relict foraminifera from the Agulhas Bank, South African continental margin. Unpublished M.Sc. thesis, University College Wales, Aberystwyth. Parr, W. J. 1950. Foraminifera. Rep. B.A.N.Z. Antarct. Res: Exped. (B) 5 (6): 236-392. PLummMer, H. J. 1926. Foraminifera of the Midway Formation in Texas. Bull. Univ. Tex. Bur. econ. Geol. 2644: 3-206. Rau, W. W. 1964. Foraminifera from the Northern Olympic Peninsula, Washington. Prof. Pap. U.S. geol. Surv. 374-G: 1-33. Reuss, A. E. 1854. Beitrage zur Charakteristik der Kreideschicten in den Ostalpen, besonders im Gosauthale und am Wolfgangsee. Abh. math-naturw. KI. Akad. Wiss Wien 7 (1): 1-156. SIESSER, W. G. 1973. Diagenetically formed ooids and intraclasts in South African calcretes. Sedimentology 20: 539-551. SIESSER, W. G. 1977. Upper Eocene age of marine sediments at Bogenfels, South West Africa, based on calcareous nannofossils. In: Papers on Biostratigraphic Research. Bull. geol. Surv. Rep. S. Afr. 60: 72-74. SIESSER, W. G. & DINGLE, R. V. 1979. Tertiary sea-level movements around southern Africa. (Abstract.) AAPG-SEPM Annual Convention Houston, Abstracts of Papers: 165. SIESSER, W. G. & MILES, G. A. In press. Calcareous nannofossils and planktic foraminifera in Tertiary limestones: Natal and eastern Cape Province, South Africa. Ann. S. Afr. Mus. SOUTH AFRICAN COMMITTEE FOR STRATIGRAPHY 1977. South African Code of Stratigraphic Terminology and Nomenclature. Spec. Publs geol. Surv. Rep. S. Afr. 20. Pretoria: Govern- ment Printer. STOCKEN, C. G. 1962. The diamond deposits of the Sperrgebiet, South West Africa. Field Excursion Guide, 5th Ann. Congr. geol. Soc. S. Afr. TERQUEM, O. 1882. Les foraminiferes de l’Eocéne des environs de Paris. Mém. Soc. géol. Fr. 3: 1-139. VAIL, P. R., MircHum, R. M. & THompson, S. 1977. Global cycles of relative changes of sea level. In: PAYTON, C. E. ed. Seismic stratigraphy—applications to hydrocarbon exploration. Mem. Am. Ass. Petrol. Geol. 26: 83-97. VAN Hinte, J. E. 1963. Zur stratigraphie und micropalaeontologie der Oberkreide und des Eozans des Krappfeldes (Karntes). Jb. geol. Bundesanst. 8: 1-140. WEISSERMEL, W. 1926. Neues uber Tabulate, Hydrozoen und eine Hexakoralle aus dem Tertiar der Bogenfelser Diamantenfelder. In: KaAIserR, E. ed. Die Diamantenwiiste Siidwestafrikas 2: 88-106. Berlin: Dietrich Reimer (Ernst Vohsen). ZIEGLER, W. H. 1969. Cenozoic geology and morphology of the coast of the western Cape Province and southern Namib desert, South Africa and South West Africa. Unpublished Report ESSO Exploration South Africa Inc. 6. SYSTEMATIC papers must conform to the Jnternational code of zoological nomenclature (particularly Articles 22 and 51). Names of new taxa, combinations, synonyms, etc., when used for the first time, must be followed by the appropriate Latin (not English) abbreviation, e.g. gen. nov., sp. nov., comb. nov., syn. nov., etc. An author’s name when cited must follow the name of the taxon without intervening punctuation and not be abbreviated; if the year is added, a comma must separate author’s name and year. The author’s name (and date, if cited) must be placed in parentheses if a species or subspecies is transferred from its original genus. The name of a subsequent user of a scientific name must be separated from the scientific name by a colon. Synonymy arrangement should be according to chronology of names, i.e. all published scientific names by which the species previously has been designated are listed in chronological order, with all references to that name following in chronological order, e.g.: Family Nuculanidae Nuculana (Lembulus) bicuspidata (Gould, 1845 Figs 14-15A Nucula (Leda) aap maa peu. Nae Sis Leda plicifera A. Adams, Laeda bicuspidata Hanley, Maso: 118, pl. 228 (fig. 73). Sowerby, 1871: pl. 2 (fig. 8a—b). Nucula largillierti Philippi, 1861: 87. Leda bicuspidata: Nicklés, 1950: 163, fig. 301; 1955: 110. Barnard, 1964: 234, figs 8-9. Note punctuation in the above example: comma separates author’s name and year semicolon separates more than one reference by the same author full stop separates references by different authors figures of plates are enclosed in parentheses to distinguish them from text-figures dash, not comma, separates consecutive numbers Synonymy arrangement according to chronology of bibliographic references, whereby the year is placed in front of each entry, and the synonym repeated in full for each entry, is not acceptable. In describing new species, one specimen must be designated as the holotype; other speci- mens mentioned in the original description are to be designated paratypes; additional material not regarded as paratypes should be listed separately. The complete data (registration number, depository, description of specimen, locality, collector, date) of the holotype and paratypes must be recorded, e.g.: Holotype SAM-—A13535 in the South African Museum, Cape Town. Adult female from mid-tide region, King’s Beach Port Elizabeth (33°51’S 25°39’E), collected by A. Smith, 15 January 1973. Note standard form of writing South African Museum registration numbers and date. 7. SPECIAL HOUSE RULES Capital initial letters (a) The Figures, Maps and Tables of the paper when referred to in the text 3 ee. . /. the Figure depicting C. namacolus ...’; *. . .in C.namacolus (Fig. 10)...’ (b) The prefixes of prefixed surnames in all languages, when used in the text, if not preceded by initials or full names e.g. DuToit but A.L.du Toit; Von Huene but F. von Huene (c) Scientific names, but not their vernacular derivatives e.g. Therocephalia, but therocephalian Punctuation should be loose, omitting all not strictly necessary Reference to the author should be expressed in the third person Roman numerals should be converted to arabic, except when forming part of the title of a book or article, such as ‘Revision of the Crustacea. Part VIII. The Amphipoda.’ Specific name must not stand alone, but be preceded by the generic name or its abbreviation to initial capital letter, provided the same generic name is used consecutively. Name of new genus or species is not to be included in the title: it should be included in the abstract, counter to Recommendation 23 of the Code, to meet the requirements of Biological Abstracts WILLIAM G. SIESSER & DAVID SALMON EOCENE MARINE SEDIMENTS IN THE SPERRGEBIET, SOUTH WEST AFRICA OLUME 79 PART 3 JULY 1979 ISSN 0303-2515 07-6 sane ilinas SE cage soe * +” . At cy aS PE Hi 5 4 ae tr } OF THE SOUTH CAPE TOWN INSTRUCTIONS TO AUTHORS 1. MATERIAL should be original and not published elsewhere, in whole or in part. 2. LAYOUT should be as follows: (a) Centred masthead to consist of Title: informative but concise, without abbreviations and not including the names of new genera or species Author’s(s’) name(s) Address(es) of author(s) (institution where work was carried out) Number of illustrations (figures, enumerated maps and tables, in this order) (b) Abstract of not more than 200 words, intelligible to the reader without reference to the text (c) Table of contents giving hierarchy of headings and subheadings (d) Introduction (e) Subject-matter of the paper, divided into sections to correspond with those given in table of contents (f) Summary, if paper is lengthy (g) Acknowledgements (h) References (Gi) Abbreviations, where these are numerous 3. MANUSCRIPT, to be submitted in triplicate, should be typewritten and neat, double spaced with 2,5 cm margins all round. First lines of paragraphs should be indented. Tables and a list of legends for illustrations should be typed separately, their positions indicated in the text. All pages should be numbered consecutively. Major headings of the paper are centred capitals; first subheadings are shouldered small capitals; second subheadings are shouldered italics; third subheadings are indented, shouldered italics. Further subdivisions should be avoided, as also enumeration (never roman numerals) of headings and abbreviations. Footnotes should be avoided unless they are short and essential. Only generic and specific names should be underlined to indicate italics; all other marking up should be left to editor and publisher. 4. ILLUSTRATIONS should be reducible to a size not exceeding 12 « 18 cm (19 cm including legend); the reduction or enlargement required should be indicated; originals larger than 35 x 47 cm should not be submitted; photographs should be rectangular in shape and final size. A metric scale should appear with all illustrations, otherwise magnification or reduction should be given in the legend; if the latter, then the final reduction or enlargement should be taken into consideration. All illustrations, whether line drawings or photographs, should be termed figures (plates are not printed; half-tones will appear in their proper place in the text) and numbered in a single series. Items of composite figures should be designated by capital letters; lettering of figures is not set in type and should be in lower-case letters. The number of the figure should be lightly marked in pencil on the back of each illustration. 5. REFERENCES cited in text and synonymies should all be included in the list at the end of the paper, using the Harvard System (ibid., idem, loc. cit., op. cit. are not acceptable): (a) Author’s name and year of publication given in text, e.g.: ‘Smith (1969) describes .. .’ ‘Smith (1969: 36, fig. 16) describes . ‘As described (Smith 1969a, 19695; ae med ‘As described (Haughton & Broom 192 ys ‘As described (Haughton et al. 1927) . Note: no comma separating name aed year Dagination indicated by colon, not p. names of joint authors connected by ampersand et al: in text for more than two joint authors, but names of all authors given in list of references. (b) Full references at the end of the paper, arranged alphabetically by names, chronologically within each name, with suffixes a, b, etc. to the year for more than one paper by the same author in that year, e.g. Smith (1969a, 19695) and not Smith (1969, 1969a). For books give title in italics, edition, volume number, place of publication, publisher. For journal article give title of article, title of journal in italics (abbreviated according to the World list o, scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses, volume number, part number (only if independently paged) in parentheses, pagination (first and last pages of article). Examples (note capitalization and punctuation) BULLOUGH, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan. FISCHER, P. —H. 1948. Données sur la nae wae et de le vitalité des mollusques. J. Conch., Paris 88: 100-140. FiscHER, P.-H., DuvaAL, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines. Archs Zool. exp. zen. 74: 627-634. Koun, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee region of Ceylon. Ann. Mag. nat. Hist. (13) 2: 309-320. Kon, A. J. 19606. Spawning behaviour, egg masses and larval development in Conus from the Indian Ocean. Bull. Bingham oceanogr. Coll. 17 (4): 1-51. THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. In: SCHULTZE, L. Zoologische und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-Afrika 4: 269-270. Jena: Fischer. Denkschr. med.-naturw. Ges. Jena 16: 269-270. (continued inside back cover) ANNALS OF THE SOUTH AFRICAN MUSEUM ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM Volume 79 Band July 1979 Julie Part 3 Deel GY) - VARIN S B:B.3:3:9 \ ue < Aayiy LS S NOV! NOS A SECOND GENUS IN THE MARINE ISOPOD FAMILY BATHYNATALIIDAE By BRIAN KENSLEY Cape Town Kaapstad The ANNALS OF THE SOUTH AFRICAN MUSEUM are issued in parts at irregular intervals as material becomes available Obtainable from the South African Museum, P.O. Box 61, Cape Town 8000 Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM word uitgegee in dele op ongereelde tye na gelang van die beskikbaarheid van stof Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad 8000 OUT OF PRINT/UIT DRUK 1, 21-3, 5-8), 3(1-2, 4-5, 8, t—p.i.), 51-3, 5, 7-9), 6(1, t.—p.i.), 711-4), 8, 9(1-2, 7), 10(1-3), 11(1-2, 5, 7, t.—p.i.), 15(4—5), 24(2), 27, 31(1-3), 32(5), 33 Copyright enquiries to the South African Museum Kopieregnavrae aan die Suid-Afrikaanse Museum ISBN 0 908407 78 5 Printed in South Africa by In Suid-Afrika gedruk deur The Rustica Press, Pty., Ltd., Die Rustica-pers, Edms., Bpk., Court Road, Wynberg, Cape Courtweg, Wynberg, Kaap A SECOND GENUS IN THE MARINE ISOPOD FAMILY BATHY NATALIIDAE By BRIAN KENSLEY Smithsonian Institution, Washington, D.C. (With 3 figures) IMS. accepted 14 June 1979] ABSTRACT Naudea louwae, the second genus and species in the south-western Indian Ocean family Bathynataliidae is described, and compared with Bathynatalia gilchristi. The single specimen of Naudea louwae was collected at 850 m off southern Natal. CONTENTS PAGE Introduction . : ; : ee OD Systematic discussion . : Ww 35 Acknowledgements . : : aP eel References : : : : bl INTRODUCTION The isopod family Bathynataliidae (Kensley 1978) was described when several specimens of Bathynatalia gilchristi Barnard (previously known from a single specimen) became available. These specimens were collected by the South African Museum’s Department of Marine Biology working from the R.V. Meiring Naude off the east coast of South Africa. A single specimen of an unusual isopod has now been found in sorting a later batch of sediments from the same station which yielded the Bathynatalia. Unfortunately this specimen was seen only after the publication of the new family. Its addition to the earlier paper would have been desirable as it slightly modifies the diagnosis of the family. SYSTEMATIC DISCUSSION Suborder FLABELLIFERA Family Bathynataltidae . Diagnosis Body dorsoventrally flattened. Cephalon anterolaterally expanded, fused with pereonite 1 medially, separated by deep slits laterally. Pereonites 2-7 distinct, articulating coxae present at least on pereonites 2-6. Pleon of five pleonites plus large pleotelson. At least two pleonites with free lateral extensions; Ann. S. Afr. Mus. 79 (3), 1979: 35-41, 3 figs. 35 36 ANNALS OF THE SOUTH AFRICAN MUSEUM pleonites | and 5 lacking free lateral margins. Antennule with four, antenna with five peduncular segments, both flagella multiarticulate. Molar process reduced to spiniform process in both mandibles. Lacinia present on one mandible only. Maxilla 1 curved, armed apically with cluster of spines. Maxilla 2 with inner ramus uni- or bilobed. Maxilliped with 3-segmented palp, broad endite, broadly oval or triangular exopod. Pereopod | robust and subchelate in both sexes; remaining pereopods ambulatory. Pleopod 1 indurate, exopod and endopod lying parallel, operculate over branchial chamber. Pleopods 2-5 biramous, membranous. Uropod subterminal, consisting of single segment, obscurely trilobed apically. Naudea gen. nov. Diagnosis Maxilla 2 with inner ramus bilobed. Pereonite 7 lacking coxa or free lateral margins. Pleonites 2 and 3 with broad lateral extensions; pleonite 4 with narrow elongate lateral extension. Pereopod 7 absent. Uropod of single segment with distal rudimentary ramus. Gender Feminine. T ype-species Naudea louwae. Etymology The generic name is derived from the C.S.I.R. Research Vessel Meiring Naude. Remarks In describing the family Bathynataliidae, the diagnosis for the genus Bathynatalia was considered the same as for the family. Now that a second genus is described, the familial diagnosis has been slightly revised. The main differences between the two genera are summarized below: Bathynatalia Barnard, Naudea gen. nov. 1957 Antennular flagellum 12 articles 5 articles Antennal flagellum 11 articles 8 articles Maxilla 2, inner ramus _unilobed bilobed Dorsal integument sculptured unsculptured Pereonite 7 with free coxa lacking free lateral margins Pereopod 7 present absent Pleonite 4 lacking lateral extension with lateral extension A SECOND GENUS IN THE MARINE ISOPOD FAMILY BATHYNATALIIDAE 20 The close affinity of the two genera within the same family is illustrated by the many similarities, especially in pereopodal, pleopodal, and mouthpart structure, as well as in the unusual uropodal structure. The superficial resemblance of Naudea to the serolids is even more marked than in Bathynatalia. Because of the similarity in structure between the serolids and bathynataliids in the maxilliped, maxillae, mandibular palp, and pereopod 1, it is thought that these two families are more closely related to each other than to any other family within the Flabellifera. The strongly depressed body, as in the serolids, is probably an adaptation to allow detrital feeding in the upper few millimetres, without the body sinking too deeply into the fine-sediment substrate. Naudea louwae sp. nov. Figs 1-3 Description Female Integument moderately indurate, brittle, lacking sculpture. Body broadest at pereonite 4; strongly dorsoventrally depressed. Cephalon lacking eyes; anterior margin hollowed to receive antiguous antennal bases; tiny rostral point present; anterolateral corners quadrate/rounded; two circular convexities dorsally marking insertion of mandibular musculature. Cephalon and pereonite | fused medially, separated laterally by deep sinuous slit. Coxa of pereonite | not demarked, but laterally broadly flattened, margin convex, with two circular convexities dorsally marking insertion of pereopod | musculature. Pereonites 2—4 similar, with shallow transverse dorsal groove; coxae demarked, rectangular. Pereonites 5—6 narrower than preceding pereonites; coxae demarked, roughly rectangular. Pereonite 7 very short, lacking coxa. Pleon consisting of five pleonites plus pleotelson. Pleonite | short, similar to pereonite 7, lacking free -lateral margin. Pleonite 2 medially very short, widening laterally, with broad rectangular lateral extension. Pleonite 3 medially very short, widening into broad lateral extension, somewhat posterodistally produced. Pleonite 4 very short, with narrow lateral extension slightly shorter than that of pleonite 3. Pleonite 5 very short, lacking free lateral margins. Pleotelson roughly rect- angular; distolateral corners rounded, separated from short, broad triangular apex by notch for insertion of uropod; proximal half dorsally convex, with low, rounded ridge running to apex. Antennular peduncle 4-segmented, basal segment slightly longer than subequal segments 2 and 3; fourth segment very short; flagellum of five articles, reaching to base of antennal flagellum. Antennal peduncle 5-segmented, segments | and 2 moderately broad; segment 3 inserted almost at right angle to segment 2; segments 4 and 5 moderately broad; flagellum of eight articles. Mandibles indurate; molar process on each side reduced to spiniform process; palp 3-segmented, first and third segments subequal, one-third length of middle 38 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 1. Naudea louwae. A. Holotype in dorsal view. Scale = 1 mm. B. Ventral view of pleon. C. Left mandible. D. Apex of left mandible. E. Apex of right mandible. F. Maxilla 1. G. Maxilla 2. A SECOND GENUS IN THE MARINE ISOPOD FAMILY BATHYNATALITDAE 39 Fig. 2. Naudea louwae. A. Maxilliped. B. Pereopod 1. C. Pereopod 6. D. Uropod. segment; latter with six short, fringed spines in distal half; terminal segment curved, with elongate terminal spine and four shorter fringed spines in distal half. Left mandibular incisor with four rounded cusps, lacinia transversely broad, with five or six short rounded cusps. Right mandible, incisor of five ‘ rounded cusps; lacinia lacking. Maxilla 1 consisting of single strongly curved indurate ramus armed distally with nine spines. Maxilla 2 with bilobed outer ramus, each lobe bearing two elongate fringed spines; inner ramus tipped with five fringed spines. Maxilliped exopod broadly oval, outer margin fringed with fine setules; endite about as broad as palp, with single strong coupling hook on median margin and four setae on distal margin; palp of three segments, first and third shorter and narrower than broadly oval second segment; terminal segment 40 ANNALS OF THE SOUTH AFRICAN MUSEUM tiny, with three elongate setae. Pereopod | robust, subchelate; dactylus meeting spine at distal end of carpus; dactylar unguis very short; propodus proximally broad, palm with slightly convex hyaline border bearing fourteen sensory spines of varying lengths; carpus triangular, with single strong distal sensory spine; merus half length of carpus. Pereopods 2-6 ambulatory, similar; unguis half length of dactylus; propodus twice length of dactylus, with four sensory spines on posterior margin and two fringed spines distally; carpus two-thirds length of SSS > 350 m) off the eastern South African coast are both abundant and diverse. A total of 1 Ctenostome, 2 Cyclostome and 48 Cheilostome species is described. Of these, 23 species— Carbasea mediocris, Notocoryne cervicornis, Notoplites cassidula, N. candoides, Tricellaria varia, Bugulella australis, Cellaria tectiformis, C. paradoxa, Aspidostoma magna, Inversiscaphos setifer, Escharoides distincta, Flustramorpha angusta, Adeonella majuscula, A. cracens, Tessaradoma bispiramina, T. circella, Sertella bullata, Reteporella dinotorhynchus, R. clancularia, Turbicellepora protensa, Costaticella carotica, Batopora lagaaiji and B. nola—are considered to be new. In addition, 3 new genera — Notocoryne, Leiosalpinx and Inversiscaphos—and 1 new family, the Setosellinidae, are intro- duced. Some species have previously been described from regions as far distant as the Eastern ~ Atlantic and Western Pacific Oceans, and are known from very few records. Of the 22 previously reported species, 14 are new records for South Africa, and 4 for the Indian Ocean. Many of the colony growth forms (morphotypes) of species are known to be adapted to life on the surface of fine sediments, and are either anchored by rhizoids or ‘free-living’. Most colony morphotypes are flexible and delicate, and some are extremely small, less than 3 mm in diameter. CONTENTS PAGE Introduction dal eee 44 List of species. . : alia: 44 Systematic account . ; : 44 Discussion . : : 5 sth hey Summary . : : : 37) 124 Acknowledgements . : aD md |S) References . i : ; pat 125 Abbreviations . g ; ew ZO Appendix 1 eee oe 2 130 43 Ann. S. Afr. Mus. '79(4), 1979: 43-130, 21 figs, 4 tables, 1 appendix. 44 ANNALS OF THE SOUTH AFRICAN MUSEUM INTRODUCTION Bryozoa are generally most abundant in the shallow waters of continental shelves, where firm substrata are available for larval settlement and colony growth (Ryland 1970). Recent work on the faunas of the deeper waters of the Atlantic and the outer shelf and slope of western Europe has, however, revealed a far greater abundance of colonies and diversity of species than previously recorded (D’Hondt 1975a, 1977; Hayward 1977, 1978a; Hayward & Ryland 1978). The substrata available for colonization in some of these regions often consist only of the fine sea-bottom sediments. Species therefore tend to show adaptations of structure which enable them to live in this environment. The bryozoan fauna described here, from seventeen of the Meiring Naude stations, with depths ranging between 376 and | 300 m, resembles those from western Europe in showing a remarkable diversity of taxa and colony growth forms. A high proportion of the species does not appear to have been described before, and many of the previously described species have extensive geographical ranges, often with equally broad bathymetric distributions. The occurrence of numerous colonies of very small ‘rooted’ or ‘free-living’ species in the bottom sediments is particularly interesting, and these collections have provided a great deal of information, not only on the South African Bryozoa, but on the potential nature of deep-water faunas from other regions. LIST OF SPECIES The Meiring Naude collections comprised 51 species of Bryozoa: 48 cheilo- stomes, 2 cyclostomes and 1 ctenostome. In Table 1 the species are listed in systematic order, and their occurrence at each of the seventeen stations indicated. The stations are arranged in order of increasing depth, and the sediment type and total number of species recorded is given for each station. Co-ordinates and depth for the Meiring Naude stations which yielded samples of Bryozoa are listed in Appendix 1. Data for all stations are given by Louw (1977). SYSTEMATIC ACCOUNT ORDER CHEILOSTOMATA Family Cupuladriidae Lagaai, 1952 Cupuladriidae Lagaaij, 1952: 31. Cook, 1965a: 154; 19655: 192. Discoporella @Orbigny, 1852 Discoporella @Orbigny, 1852: 472. Cook, 1965b: 219. Discoporella umbellata (Defrance, 1823) Lunulites umbellata Defrance, 1823: 361, pl. 47, fig. 1a—b. Discoporella umbellata: Cook, 1965a: 177, pl. 1 (fig. 7), pl. 3 (figs 1, 3, 5-6), fig. 4; 19655: 221, pl. 3 (fig. 3), fig. 2h. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 45 Material Stations SM 23, SM 31. Description Colonies lunulitiform, free-living. Zooids with an extensive cryptocyst lamina, perforated by opesiules; opesia small. Basal surface of colonies grooved. Avicularia regularly patterned, each distal to a zooid; mandibles setiform, slung from asymmetrical condyles. Remarks Two small, worn colonies only, which may have been transported from shallower water, were found. Family Setosellinidae fam. nov. Colony encrusting very small substrata, often becoming free peripherally. Ancestrula single, budding two primary zooids and two avicularia directly. Zooids in two series, budded spirally for at least the first four astogenetic generations. Avicularia interzooidal, one placed distally or distolaterally to each zooid; subrostral chambers rounded, inflated. Mandible setiform, slung from asymmetrical condyles. Brooding zooids with dimorphic, wide opercula, embryos brooded in a large ovicell with a central, distal foramen, closed by the operculum. Genera included: Setosellina and Heliodoma. The genera Setosellina Calvet and Heliodoma Calvet share a correlation of distinctive colony form, spiral budding pattern and type of zooidai poly- morphism. Some species of Setosellina appear, however, to be similar to encrusting membraniporine forms which have no distinctive budding pattern, but which are assigned to different families. A parallel group of deep-water coilostegan species (the Setosellidae) which display a similar range of colony form, but with zooids with distinct cryptocyst laminae and opesiules, has, however, long been given family status. A family group is therefore introduced here for the genera Setosellina and Heliodoma. Setosellina was placed in the Hincksinidae by Lagaaij (19635) and Cook (1965a). The type species of this heterogenous group, Hincksina flustroides (Hincks), is now usually assigned to the Flustridae (Ryland & Hayward 1977: 86). Setosellina has also been included in the family Lunulariidae, another diverse and artificial grouping, which included the Cupuladriidae (Prenant & Bobin ' 1966: 297), and the Selenariidae (Cheetham 1966: 24). The free-living members of the Cupuladriidae and Selenariidae differ from the Setosellinidae in possessing a basal, colony-wide coelomic cavity (Hakansson 1973). The genus Vibracellina Canu & Bassler (1917: 14) includes fossil and Recent species which may prove to be attributable to the Setosellinidae. V. laxibasis Canu & Bassler, a Pliocene species from Panama which was discussed by Lagaaij (19635), and V. viator Canu & Bassler (1929: 97), a Recent xX XR ANNALS OF THE SOUTH AFRICAN MUSEUM 46 xX X me O68 sug) 4 “(Wu CO‘ [ > ou AIDA) JA :SNY} poyeorpuT sr soporjred yUOUTIpas Jo ozIs JURUTWIOpeld SUL, ‘(Pet1aFUT) (YY) “(PeAtesqo) Y :sny} po}oU ST SPpIOzIYI Jo eouasoid oY, ‘(WIOJUTTJaSOI0s) § pue (WIOJINUN) JT ‘“(WAOFUNPEIeYOsoUOd) OD “(WIIOFMIETIID) VD ‘(WIOJTITE[N]TE9) 9D “(WIOJINSNY) A ‘(WIJOjI10d9}01) Y ‘(WIOJIUOEpe) Y ‘(WIIOFIIOdTuvIquIOUT) J : SNY} poyedrpul oe (g{T °d 99s) sodAjyoydioul AuojoD *| poyx1ew ore ULC UPIPU] 94} 0} Mou so1oeds puv ‘, payxJeUl ore solsods MON ‘yIdap SUISeSIOUI Jo JopiO Ul pesULIIL o1e SUOTIE}S OY} PUL “IOPIO SI}eUID}SAS Ul posi] ore soloods oY I 2qQU0L fo uoyouvjdxq ID a [Dad oe 1D A IOUS Ss. Ss YS ID US Ad OW HA OW d HA AH AAHA A A d dH dA DEG eS a Aa iin oh eee AVS | SO Sli I aC BE CC Cae Oe ah so sorseds JO ‘ON x es 1YJAOMSPIOY “Ie D1S1AD x " * *) paljuDj1D DaUuospIUpyT x 7 5 + ds papyjasssnpfoan?é x 7 + + + "ds uoposoys04, DUDILA{D DUI JAADYISAUDD SMuUdsofiddd DINULAIDT pjou D10d0j0g » thiovsv] v1odojvg » : 2 : 14pAAnu psodojog : : DIDULADUL DAOdOJaJOUp x ; : S14JSOA1JD] DAOdOAaJOUp x "8 991JOADI DIJAINIDISOD » 4 : : : * "ds psagip1ddn[ é psuajodd DAOdAJAI1GANT, DIADINIUD]I 011210431 2Y x snyoudysojouip 111240d212Y » x TS SANA IE ES xX Xx x 7% 19]]8D AI. DULOPDAVSSAL, x DUIUDALdS1G DULOPDADSSAT, » ey . ° cman nN Wnnvaqi: n9nN1N17777I1001 (a) 0) - Ad 0h 0h 0 oe Oe oo fe) fe) O00 x x XX us 25 2S Xx 0S eX ON x X X x x x X X x x QD e:)) Aenean toOnn 6K XK OK x x SK 3K KOK KK LAAPFANATAMNAANCHYTTN Geos x - a 4 TABLE 1 List of species collected. Depth 500 m > 750 m > 1000m > Station 16 2385/8669 92 103 67 1 53 32 31 78 60 61 No; of Mork 41 107 109 stations type —_—-Rhizoids I ors Discoporellaumbellata . . x x +Setosellinaroulei. . . . x tHeliodomaimplicata. . . X X *Carbasea mediocris yee *Notocoryne cervicornis eo x Notocoryne cylindracea . . x *Notoplites cassidula . . . Sees x *Notoplites candoides . . . x *Tricellaria varia. . . . x Eupaxia quadrata F : . x x *Bugulellaaustralis . . . X X X x Leiosalpinx inornata 2m Columnellamagna . . . x Neliaispiuenite tues. se, ie x +Petalostegus bicornis . . . x x *Cellaria tectiformis . 8 ; x= XTX x *Cellaria paradoxa = eee: x x *Aspidostoma magna . . . x x Figulariaphilomena . . . xX *Inversiscaphos setifer . ‘i é x *Escharoides distincta . . . yaa Flustramorpha marginata . . x *Flustramorpha angusta : eo Gigantopora polymorpha . ; x Adeonella coralliformis 5 ; x x * Adeonella majuscula . ‘ A x * Adeonella cracens - 5 . SO Adeonella sp. P : ‘ _ x x ; RB a Se SS ag a IAT a ee eed Se eg Cleidochasma protrusum . ; x au Smittoidea ?hexagonalis *Tessaradoma bispiramina . . *Tessaradoma circella. . . XxX *Sertellabullata . . . . ™X * Reteporella dinotorhynchus * Reteporella clancularia *Turbicellepora protensa ?Turritigera sp. eT ee *Costaticella carotica . . . *X Anoteropora latirostris x Anoteropora inarmata Batopora murrayi *Batopora lagaaiji *Batopora nola a Lacrimula pyriformis . Conescharellina africana Trochosodonsp.. . . . ?Neoflustrellidra sp... % Fi x Idmidronea atlantica . " = x Crisia aff. holdsworthi % x Noyof:species; =. 4 23-23) DI (3a 22) S11 eal S saa A eae ee ee ee ae VF F F F F VF VF F VF F MC VF MC F Sh Gl Sh S S GiGl F Gl F F Gl F Gl x XK XX x x x x x x x — nner am ° ~ Rrrnnnnn -_~ Le WOASNW NVOI8AV HLAOS AHL 40 STVNNV x X WWNWEENNENNWAN HERP RWEEKWNIFKOAN Or PP >>> mNZMErPOOOOOOOOOE OD PARR x xX XX x xX XK XK xX XK K KK XK x x Slslel talallakaa tal) XXX XX x x xX XX x x xX XX xX XK XK xXx XX x lens Gigs XS Cr St tL bo det NU ee Explanation of Table 1 The species are listed in systematic order, and the stations are arranged in order of increasing depth. New species are marked *, and species ae to ne Indian Ocean are marked t. Colony morphotypes (see p. 118) are indicated thus: M (membraniporiform), A (cdeouitoan) R Geiepodtoray : = ri wen Ce (cellulariiform), Ca (cellariiform), Co (conescharelliniform), L (lunulitiform) and S (etoselliniform). The presence of rhizoids e a fa us lus: R (observed), (R) (inferred). The predominant size of sediment particles is indicated thus: VF (very fine <1,0 mm), F (fine <5,0 mm), GL Caleb =e Senate ae a and the dominant type of particle thus: Sh (shell), S (sand), F (benthic foraminiferans with arenaceous test), Siocon: ype foraminiferans). The number of species at each station, and the number of stations at which each species was found is SHSINUO JGAYN ONIVITW SIANASAW NVOrddaY HINOS AHL LY 48 ANNALS OF THE SOUTH AFRICAN MUSEUM species from the Philippines (41-903 m), are both very similar in colony form and zooidal morphology to Setosellina. Setosellina Calvet, 1906 Setosellina Calvet, 1906: 157; 1907: 395. Cook, 1965a: 182. Ancestrula with one proximolateral primary zooid and avicularium, and one distal primary zooid and aviculartum. The two spirally budded series of zooids originating from these zooids surround the ancestrula in either a clock- wise or anti-clockwise direction. Intercalary series of zooids budded after the fourth astogenetic generation. Zooids growing free from the substratum peripherally, but rarely protruding for more than two generations. Peripheral avicularia often enlarged and directed basally. Ovicell large, terminal, with a central foramen, closed by a wide operculum. The earliest known species is S. gregoryi Cheetham (1966: 25, figs 45) from the Upper Eocene of southern England. The zooids were very small, were budded in anti-clockwise series and had distally placed avicularia. Recent species are associated with fine particled sediments (Lagaaij 19635), from deep water. The localities given by Calvet (1906, 1907) are based on a meridian passing through Paris, not Greenwich (see Ryland 1969: 238). Setosellina roulei Calvet, 1906 Figs 1A, 17B, 18B Setosellina roulei Calvet, 1906: 157; 1907: 395, pl. 26 (figs 5-6). Material Stations SM 1, SM 16, SM 69, SM 86, SM 103, SM 109. Description Colonies very small (largest—with 34 zooids—3,0 mm diameter). Zooids with basal walls calcified only peripherally, cryptocyst narrow, gymnocyst distinct laterally, variable proximally. Avicularian subrostral chambers large, rounded, placed distally to zooids. Remarks Cook (1965a) distinguished S. roulei from the Mediterranean species, S. capriensis (Waters), on the basis of zooidal size as well as bathymetrical and geographical distribution (see also Harmelin 1977: 1062). The zooids of South African colonies are of comparable size up to four generations from the ancestrula, but attain larger dimensions later in astogeny (see Table 2). Colonies of S. capriensis therefore have a similar size range to those of S. roulei, but differ in that in S. capriensis size decreases with astogeny in colonies which grow beyond the substratum. In addition, the lateral cryptocyst < a THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 49 of S. capriensis is extensive, and the apertures of the ancestrula and first three zooid generations are closed by a lamina very early in astogeny, when colonies have only twelve zooids. Colonies of S. roulei have closed zooids only when more than thirty zooids have been budded, and those affected include only the ancestrula and primary zooid pair, which have slight extensions of the crypto- cyst. S. goesi (Silén), from the West Indies and Florida, also has zooids with a well-developed lateral cryptocyst (see Lagaaij 19635: 172, pl. 2 (fig. 1), figs 1-2). The zooids are very small and the avicularia budded distal-laterally. Lagaaij (19635) observed that a significant proportion of colonies had anti-clockwise spirals of growth. S. constricta Harmer (1926: 264), from the East Indies, differs from all other species in the absence of pore-chambers. The zooids are very small, and have thin lateral cryptocysts and no closed zooids. Calvet (1907, pl. 26, fig. 5) figured, but did not describe, two zooids of S. roulei with large ovicells, each with a small uncalcified frontal foramen. Very similar ovicells have recently been described in S. capriensis by Harmelin (1977: 1062, fig. 12). TABLE 2 Comparative measurements (in mm) of species of Setosellina and Heliodoma implicata Lz Lap Lop lop Is Setosellina roulei ancestrula . 0,25-0,28 0,23-0,25 0,22-0,24 0,17-0,18 zooid 1 0,30-0,33 0,27-0,30 0,25-0,28 0,18—0,20 0,30 zooid 2 0,35-0,41 0,32-0,33 0,28-0,30 0,18-0,19 0,38 zooid 3 0,39-0,42 0,34-0,36 0,31-0,34 0,19-0,22 0,40 zooid 4 0,40-0,42 0,35-0,36 0,32-0,34 0,23-0,25 0,53 zooid 6 0,43 0,36 0,33 0,25 0,75 zooid 7 0,50 0,43 0,39 0,29 0,75 S. capriensis ancestrula . 0,25-0,30 0,23-0,25 0,20-0,23 0,19-0,20 — zooid 1 0,30-0,35 0,27-0,29 0,20-0,22 0,15-0,16 ~~ zooid 2 0,35-0,41 0,33-0,40 0,30-0,32 0:15-0:17 — zooid 3 0,45—0,50 0,36—0,39 0,32-0,36 0,13-0,16 — zooid 5 0,50-0,56 0,40-0,43 0,32-0,35 0,17—0,20 — zooid 9 0,25-0,28 0,19-0,21 0,17-0,18 0,11-0,13 — zooid 11 0,22-0,24 0,17-0,20 0,16-0,18 0,10-0,12 — Heliodoma implicata ancestrula . 0,20-0,22 0,18-0,19 0,17-0,18 0,15-0,17 — zooid 1 (lateral) 0,20-0,25 0,19-0,21 0,18-0,19 0,11-0,13 — zooid 1 (distal) 0,23-0,30 0,20-0,25 0,20-0,21 0,14-0,18 0,40 zooid 2 : 0,31-0,35 0,25-0,27 0,21-0,23 0,14-0,15 0,43 zooid 3 0,36—-0,40 0,25-0,28 0,20-0,21 0,13-0,14 0,50 zooid 4 0,42-0,43 0,27-0,30 0,20-0,22 0,10-0,14 0,68 zooid 5 0,43-0,45 0,28-0,31 0,20-0,21 0,08-0,10 0,83 zooid 6 0,40-0,43 0,30-0,31 0523-025 0,09-0,10 1,10 zooid 11 0,41-0,45 0,28—0,30 0,27-0,29 0,09-0,11 1,30 zooid 20 0,42-0,43 0,29-0,31 0,27-0,28 0,09-0,10 1,40 50 ANNALS OF THE SOUTH AFRICAN MUSEUM Nearly all the colonies from South Africa were alive when collected; they have the setiform mandibles intact, and tentacles and viscera are present in all zooids except the ancestrula in specimens with fewer than sixteen zooids. Colonies are far less numerous than those of Heliodoma implicata (see below), and show no evidence of larval preference for any particular type of substratum. The size range of sand grains and foraminiferans colonized is 1,00—3,00 mm. The long setiform avicularian mandibles (which reach a length of 0,75 mm) almost certainly have a stabilizing, if not supporting function. They probably also clean the colony surface of deposits, like those of the Cupuladriidae (see Cook 1963) and Selenariidae (see Cook & Chimonides 1978), although the low sedimentation rate in very deep waters makes this an inferred, secondary function. Distribution The only previous records of S. roulei are from 1 900 m off the Cape Verde Islands, and from 2 330 m off Cap Blanco. Heliodoma Calvet, 1906 Heliodoma Calvet, 1906: 157; 1907: 396. Colonies primarily encrusting minute substrata and subsequently becoming free-living. Ancestrula with one proximolateral avicularium, and one distal and one distolateral primary zooid. Two spirals of zooids originate from these zooids, each budded laterally and clockwise, surrounding the ancestrula and alternating with each other. Intercalary zooid series absent. Substratum becoming covered basally by extensions from the basal walls of free-living zooids. Interzooidal avicularia distal. Peripheral avicularian setae supporting colony. Ovicells large, terminal with a central foramen, closed by a wide operculum. Heliodoma implicata Calvet, 1906 Figs 17A, 18A Heliodoma implicata Calvet, 1906: 157, 1907: 396, pl. 26 (figs 7-9). Harmelin, 1977: 1063, pl. 1 (fig: 4), fig. 11. Material Stations SM 1, SM 16, SM 23, SM 31, SM 41, SM 60, SM 61, SM 69, SM 86, SM 103. Description Zooids with well-developed lateral cryptocysts. Ancestrula and primary zooids with thin cryptocysts, which later become extended to form complete closures. Basal walls thinly but completely calcified. Setae of peripheral avicularia very long. aia Pe es ee ee THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 51 eed Fig. 1. A. Setosellina roulei Calvet. Young colony with setiform avicularian mandibles. B. Inversiscaphos setifer gen. et sp. nov. Young colony with paired setiform avicularian mandibles. Scale = 1,0 mm. 52 ANNALS OF THE SOUTH AFRICAN MUSEUM Remarks The differences between Setosellina and Heliodoma are small but consistent. The early astogeny differs fundamentally (see Fig. 18A—B), and the zooids of Heliodoma are always budded laterally. The first primary bud is narrower than the ancestrula (see Table 2) and is laterally budded at right angles; the other primary bud is distal. The pattern figured by Calvet (1907) differs slightly from this, and he did not show any closed zooids. In the South African specimens closures are present in the ancestrula and primary zooids when colonies have six pairs of zooid generations. In larger, older colonies, closures extend to the zooids of the fourth astogenetic generation. Closed zooids have the aperture completely covered by a thick, curved calcified lamina, which has a raised, central umbo, and a faint opercular scar. In contrast, the closures of S. capriensis (see above) are flat with a small central pore. The peripheral setiform mandibles reach 1,40 mm in length and almost certainly have a supporting, as well as a possible cleaning function. None of the colonies is as large as that figured by Calvet, which had 39 zooids. The largest has 34 zooids and a diameter of 2,30 mm. The larvae of H. implicata appear to display a distinct preference for fine substrata. Of a total of 221 colonies, 195 grew on sand grains, 22 on foraminiferans, 3 on shell fragments and 1 on a fragment of echinoid spine. The size range of substrata selected ranged from 0,70 mm to 2,00 mm in diameter. Most of the sediments analysed had sand grains abundantly available, except at stations SM 1, SM 16 and SM 69. At station SM 16, shell fragments were dominant, yet of 138 colonies found, none encrusted shell, and 127 grew on sand grains. Distribution The original description of H. implicata was based on 6 colonies (4 from 1 900 metres off the Cape Verde Islands, and 2 from 3 700 m off the Canary Islands). Recently, 9 more colonies were reported from 200 m from a sea-mount north of the Canary Islands (Harmelin 1977). Family Flustridae d’Orbigny, 1852 Flustridae d’Orbigny, 1852: 324. Smitt, 1868: 357. Ryland & Hayward, 1977: 76. Carbasea Gray, 1848 Carbasea Gray, 1848: 105, 146. Ryland & Hayward, 1977: 79. Carbasea mediocris sp. nov. Fig. 2A—B Material Holotype: SAM-A26294, station SM 86. 27°55.4’S 32°40,8’E. 550 m. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 53 Ss = k = = % = Fig. 2. A-B. Carbasea mediocris sp. nov. A. Zooids from the middle of the branch, frond edge on the right. B. The proximal portion of the colony. C—D. Leiosalpinx inornata (Goldstein). C. Portion of a colony. D. Zooids viewed in profile, showing joints. E-F. Nellia sp. E. A portion of the colony. F. An ovicelled zooid. Scale = 0,5 mm for E-F; 1 mm for A-D. 54 ANNALS OF THE SOUTH AFRICAN MUSEUM Description Colony forming long and narrow, strap-like fronds, unilaminar, up to 3,5 cm long. Zooids in single, linear series, each series bifurcating infrequently; large, oval or linguiform, distal end rounded, proximal end forked. Very lightly calcified, frontal surface almost entirely membranous, with a scarcely discernible area of gymnocyst over each proximal corner. Operculum subterminal, marked by a thin, shallowly curved sclerite. Adjacent zooids are linked by large multi- porous septula in the vertical walls. These linkages are formed between the distal end of the proximal zooid and the forked corner of the distal zooid. At the frond edge, the outer, proximal portion of each zooid may be traced back, along the outer edge of the zooid preceding it, as a slender tube, apparently originating from a particularly conspicuous septulum in the outer proximal fork of that zooid. Etymology Mediocris (L)— ordinary, referring to the lack of strongly defined features. Remarks One frond had a slender, proximal zooid which was inferred to be the ancestrula. It was tapered proximally, but had no apparent anchoring processes. A single series of five zooids succeeded this assumed ancestrula, the first arose from a septulum in its proximal half, traversing it as a simple tube; disto- laterally, on the opposing side, it budded what appeared to be a simple kenozooid. Two colonies were collected; in both, all the zooids lacked polypides and a majority contained distinctive brown bodies. The species of Carbasea present acute taxonomic problems, arising out of their simplified morphological features. Zooid size and shape are the most useful characters, and in these respects C. mediocris is distinct from other known southern hemisphere species. In other members of the Flustridae the form of the colony and the structure of ovicells and heterozooids appear to be valuable characters. The marginal tubular structures seen in C. mediocris, which may prove to be kenozooids, have not been described in other species of Carbasea. Measurements (means of 15 values) in mm Lz WZ, 1,05 0,56 Family Chaperiidae Jullien, 1888 Chaperiidae Jullien, 1888: 61. Brown, 1952: 94. Notocoryne gen. nov. Colony erect, club-shaped or cylindrical, rising from a single elongate ancestrula, secured by rhizoids. Zooids in longitudinal series around the long THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES B axis of the colony. Occlusor laminae well developed, occupying the distal third of the opesia. Adventitious avicularia present; spines present, both cylindrical and branched. Ovicell hyperstomial, closed by the zooidal operculum. Type species: Notocoryne cervicornis sp. nov. The genus Chaperia, as presently constituted, includes encrusting, erect foliaceous and vinculariform species; this assemblage may prove to be unnaturally broad. In Notocoryne the occlusor laminae are more fully developed than in all species of Chaperia, but the principal difference between the two genera lies in the colony form of Notocoryne, which implies a degree of morpho- logical integration and adaptation not found in the former genus. Etymology Notos (G)—south, koryne (G)—a club, referring to the geographical distribution and shape of the colonies respectively. Notocoryne cervicornis sp. nov. Fig. 3 Material Holotype: SAM—A26303, station SM 86, 27°59,5’S 32°40,8’E, 550 m. Other material: stations SM 1, SM 16, SM 32, SM 41, SM 53, SM 60, SM 86. Description Colony erect, in the form of a slender, faceted club, arising from a single ancestrula but budding rapidly to produce six alternating, longitudinal series of zooids within four astogenetic generations. In the extensive material collected, the largest colonies were 5,5 mm long. Zooids broadly oval or pear-shaped, narrowing distally, 0,5—-0,55 mm long by about 0,5 mm broad; zooidal boundary marked by the edges of a broad, flaring cryptocyst, finely granular, gymnocyst - reduced and indistinct. Opesia oval, occupying about one half the total zooid length. Occlusor laminae well developed, prominent in the distal half of the opesia in cleaned specimens; fused medially, forming stout junctions with the lateral walls and delimiting two elliptical lacunae for the passage of the opercular occlusors. Two short cylindrical spines on the distal border of the opesia, absent in ovicelled zooids; on either side of the orifice a pair of large hollow, intricately- branched and antler-like spines, curving inwards over the frontal membrane. These cervicorne spines are variously developed, and may be broken short in ‘ older material, their position marked by a pair of thickened, socket-like bases. Adventitious avicularia present, typically one pair per zooid, situated on the gymnocyst and directed medioproximally, mandible elliptical. Ovicell spherical, imperforate, finely granular, partially immersed in the succeeding zooid, closed by the operculum of the maternal zooid. On zooids succeeding non-ovicelled individuals a small, simple kenozooid, with an oval opesia, occurs between the paired avicularia. 56 ANNALS OF THE SOUTH AFRICAN MUSEUM Ancestrula elongate, up to 1 mm in length, posteriorly cylindrical, smooth and bifid. Frontal membrane occupying about three-quarters of the length of the zooid, underlain by a flaring, finely granular cryptocyst with a few small marginal spines. Opesia oblong. Etymology Cervus (L)—a deer, cornu (L)—a horn, referring to the spines. Remarks The ancestrula buds two zooids from its distobasal surface; from these a triplet of zooids is budded, thereafter new zooids are added in alternating whorls of three, giving a hexagonal section to the colony. Zooids of the first two astogenetic generations may be more slender than later ones, but otherwise differ from them in no significant way, although the small kenozooids are more frequent on the proximal parts of the colony. In the largest colonies the ancestrula is frequently damaged and the earliest generations of zooids have their opesiae occluded by convex laminae of granular calcite, usually each with a small central foramen. The small kenozooids are inferred to give rise to supporting rhizoids, although none were found in the material studied. Notocoryne cylindracea (Busk, 1884) Electra cylindracea Busk, 1884: 78, pl. 33 (fig. 2). Material Stations SM 16, SM 41. Description Colony erect, cylindrical, branching irregularly. Zooids oval, closely spaced, boundaries indistinct. Frontal surface largely occupied by an oval opesia; gymnocyst small and obscured, cryptocyst narrow, a distinct mural rim surrounding the distal half of the zooid. Occlusor laminae closely applied to the terminal wall. Four distal oral spines present; the distalmost pair short, thick and cylindrical, the proximal pair flattened, blade-like, curving over the frontal membrance. A single avicularium occurs on the gymnocyst, the rostrum elongate, lanceolate, directed distally, over the frontal membrane, occasionally laterally or proximally. A broad, cervicorne spine arises from the base of the avicularium and extends over the proximal half of the opesia. A second type of avicularium arises from the distal wall, single or paired, columnar, with a short, acute mandible; rarely this may replace the elongate, proximal type. Ovicell hemispherical, smooth, with a triangular frontal lacuna bounded by raised ridges. The ovicell is usually intimately associated with the distal avicularia. Remarks The principal justification for assigning Electra cylindracea to the new genus Notocoryne lies in the form of its colony, although zooid morphology is THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES Fig. 3. Notocoryne cervicornis gen. et sp. nov. A. Zooids at the distal tip of a colony. B. The characteristic appearance of older zooids, with spines lost. C. The ancestrula of a young colony. D. The same in lateral view. E. A complete colony, the most proximal zooids with occluded opesiae. Scale = 0,5 mm for A-D; 4 mm for E. 58 ANNALS OF THE SOUTH AFRICAN MUSEUM strikingly similar to that of the type species. The colony of N. cylindracea is developed from a single, elongate ancestrula which, in larger specimens, becomes enclosed by basally directed rhizoids. Two fragments only were obtained by the Meiring Naude. Described by Busk (1884) from Prince Edward Island, N. cylindracea has not been reported before from South African waters and these records possibly mark the as limit of its geographical range. Family Scrupocellariidae Levinsen, 1909 Scrupocellariidae Levinsen, 1909: 130. Ryland & Hayward, 1977: 128. Notoplites Harmer, 1923 Notoplites Harmer, 1923: 348. Ryland & Hayward, 1977: 131. Notoplites cassidula sp. nov. Fig. 4A—C Material Holotype: SAM-A26299, station SM 23, 27°44,4’S 32°42,8’E, 400-450 m. Other material: Stations SM 60, SM 86. Description Colony erect, forming diffuse tufts up to 1,5 cm high, branching dichotomously; bifurcations of Type 15 (Harmer 1923). Zooids elongate, slender, biserially arranged; opesiae oval, each occupying half the frontal surface of the zooid, not overlapping but strictly alternating along the length of the branch. Cryptocystal rim narrow, scutum absent; four or five delicate oral spines present, frequently broken short. An adventitious avicularium present on the outer distal angle of each zooid, mandible short, triangular, directed laterally. Frontal avicularia also present, with short, triangular mandibles orientated perpendicularly to the long axis of the branch; when borne by ovicelled zooids, the orientation of the frontal avicularia is reversed, they are closely associated with the ovicells and directed distally. Enlarged frontal avicularia sporadic; rostrum hooked, supporting a short, triangular mandible. Ovicell small, domed, resembling a small helmet (hence, cassidula), but not projecting prominently from the branch; with a transversely elongate frontal fenestra, not closed by zooidal operculum. Basal avicularia absent. Each zooid has a number of conspicuous septula on the basal surface, at the points of origin of long rhizoids. These traverse the length of the colony, closely applied to the basal surface, forming proximally a thick bundle of anchoring rootlets. Etymology Cassidula (L)—a little helmet, referring to the ovicell. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 59 Fig. 4. A-C. Notoplites cassidula sp. nov. A. Part of a branch. B. Ovicelled zooids at a Lifurcation. C. Basal view of a bifurcation, two rhizoids proximally. D-E. Notoplites candoides sp. nov. D. Part of a colony, including a bifurcation, showing parts of three lateral processes. E. A bifurcation in basal view. Scale = 0,5 mm for A-D; 1 mm for E. 60 ANNALS OF THE SOUTH AFRICAN MUSEUM Remarks The genus Notoplites Harmer is defined precisely by its method of bifurca- tion, shown clearly in the present species (Fig. 4C). The scutum may be present or absent, but the lateral avicularia, fenestrate ovicell and oral spines of N. cassidula are typical of the genus. The antarctic and subantarctic species of Notoplites were monographed by Hastings (1943), and four Indo-West-Pacific species were described by Harmer (1926). N. cassidula differs from all the species included by these two authors, and bears no resemblance to any cellularine species described from this area by earlier writers. It is characterized by the lack of scutum and basal avicularium, the shape and size of the ovicell and the form of the two types of frontal avicularia. Notoplites candoides sp. nov. Fig. 4D-E Material Holotype: SAM-A26300, station SM 23, 27°44,4’S 32°42,8’E, 400-450 m. Description Colony erect, branching dichotomously, bifurcations of Type 15 (Harmer 1923); branches linked laterally by stout chitinous rhizoids, giving a loose, reticulate structure, reminiscent of a species of Canda. Zooids elongate, biserially arranged; opesiae oval, each occupying about one half the frontal surface of the zooid; angled to the long axis of the branch, thereby imparting a saw-toothed outline. Cryptocystal rim broad, scutum absent; two or three distal oral spines present, delicate and frequently broken short. A single adventitious avicularium present on the gymnocyst of each zooid, immediately proximal to the opesia, acute to frontal plane; mandible acute triangular, directed distally or proximally. Enlarged avicularia occur sporadically, replacing the frontal type and essentially of the same form. A second avicularium present on the basal surface of each zooid, situated on the outer distal corner and directed disto- laterally. The broad, tubular rhizoids linking the branches arise from enlarged kenozooids, with small circular opesiae; these are developed on the basal, frontal or lateral surfaces of the colony and their position varies from zooid to zooid. Occasionally a kenozooid is developed on the distal wall of the terminal zooid in a row, when the small opesia may be missing. The rhizoids typically extend laterally, probably fusing with, and certainly continuous with, those arising from neighbouring branches. Others trail proximally and perhaps serve to secure the colony to the substratum. Etymology Canda—a bryozoan genus, oides (G)—like, referring to the appearance of the colony. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 61 Remarks A single, damaged, colony of this species was found, comprising four branching fragments up to 1 cm long. Regrettably, both early stages and brooding zooids were absent but the morphology of this species is so unusual that it is certain that it has not been described before. The cross-linking rhizoids are a feature of the genus Canda Lamouroux, but the method of bifurcation and the type and distribution of avicularia place this species unequivocally in Notoplites Harmer. Tricellaria Fleming, 1828 Tricellaria Fleming, 1828: 540. Ryland & Hayward, 1977: 143. Tricellaria varia sp. nov. Fig. 5 Material Holotype: SAM-A26295, station SM 103, 28°31,7’S 32°34,0’E, 680 m. Description Colony erect, branching, jointed; bifurcations of Type 9 (Harmer 1923). Internodes of three to nine zooids, a variable but always uneven number, separated by well developed brown, chitinous joints. Zooids in biserial rows, slender; opesiae oval, occupying just less than half the frontal surface, alternating and not overlapping. Cryptocystal rim narrow, scutum broad, consisting largely of an extensive proximal lobe. Two spines present on the outer distal edge of each zooid. Presence, frequency and size of adventitious avicularia very variable; most frequent are lateral avicularia, situated on the outer distal corner of the zooid. These are of varying size, may be present on a few or all of the zooids of ~acolony, or may be absent altogether. Even rarer are minute frontal avicularia which may occur on the gymnocyst immediately proximal to the opesia. This variability may impart remarkably different appearances to different colonies, but intermediate branches may always be found. Rhizoids issue from large kenozooids sporadically present basally, at the proximal end of the zooid. Ovicell globular, smooth, with a small fenestra close to its opening. Etymology Varia (L)—different, referring to the variability of the avicularian characters. Remarks The material of this species was extensive but fragmentary, comprising portions of several different colonies, the largest fragment being about 7 mm long. The range of internode size and the variation in avicularia caused initial 62 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 5. Tricellaria varia sp. nov. A. An internode, with well-developed lateral avicularia. B. Three zooids, with a single small lateral avicularium. C. Ovicelled zooids. D. Three zooids, one with a minute frontal avicularium. E. Basal view of a bifurcation. Scale = 0,5 mm. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 63 confusion in the determination of this species, but sufficient intermediate specimens were found to demonstrate that only one species was present. T. varia is distinguished from other species of Tricellaria by the variable lateral avicularia, and the almost complete absence of frontal avicularia. It has fewer spines than most southern hemisphere species and, as in many of the Scrupocellariidae, the morphology of the ovicell appears to be characteristic. Eupaxia Hasenbank, 1932 Eupaxia Hasenbank, 1932: 321, 363. Eupaxia quadrata (Busk, 1884) Fig. 6C-F Cellularia guadrata Busk, 1884: 18, pl. 5 (fig. 5). Eupaxia incarnata Hasenbank, 1932: 363, fig. 30A-C. Material Stations SM 16, SM 92, SM 103. Description Colony erect, slender, branching dichotomously, up to 5 cm high; deep carmine in alcohol preserved specimens. Zooids in two alternating, longitudinal series, the frontal planes at an obtuse angle; elongate, rectangular, approximately 1 mm long by 0,35 mm wide, lightly calcified with a completely membranous frontal surface. Distolaterally each zooid bears a small adventitious avicularium, mandible short, triangular, orientated at a right angle to the branch and directed laterally. Basally, the avicularium may be seen to be linked to the outer proximal corner of the zooid by a slender tube, most clearly visible at the growing tips of the colony. Elsewhere, additional tubes arise in the same region of each zooid, traverse the edges of the basal surface of the colony as rhizoids, gathering proxi- mally into a bundle and anchoring the colony. At a bifurcation the basal rhizoids cross from the inner edges of the two branches to the outer edges of the supporting ramus (Fig. 6F). Ovicells were not found, and perhaps do not occur. Ancestrula of similar size to later zooids with, additionally, a slender proximal portion continued as an elongate tube. The ancestrula and earliest zooids are typically ensheathed by rhizoids. The coloration of the colony is very characteristic and seems to be derived from pigment in the polypide and from coarse granules in the coelomic fluid. Distribution E. quadrata was reported by Busk (1884) from Kerguelen and Heard Islands at depths of 51 m (28 fms) and 137 m (75 fms). Hasenbank’s material was collected off the Somali coast, at the far greater depth of 1 668 m, where it possibly approaches the northern limit of its range. 64 ANNALS OF THE SOUTH AFRICAN MUSEUM Family Bicellariellidae Levinsen, 1909 Bicellariellidae Levinsen, 1909: 93. Ryland & Hayward, 1977: 146. Bugulella Verrill, 1879 Bugulella Verrill, 1879: 472. Maturo & Schopf, 1968: 36. Bugulella australis sp. nov. Fig. 6A-B Erymophora sp. Hastings, 1943: 469, fig. 56. Material Holotype: SAM—A26302, station SM 23, 27°44,4’S 32°42,8’E, 400-450 m. Other material: stations SM 16, SM 23, SM 32, SM 86. Description Colony erect, branching; delicate, forming diffuse, tangled clumps, frequently attached to sponges or to other bryozoans (Adeonella spp.). Zooids in single, linear series, branching frequently; elongate, a slender tubiform proximal portion broadening abruptly to an oval distal portion bearing an oval opesia, with a narrow cryptocystal rim. Ten delicate marginal spines present, the two distal pairs short and erect, the rest incurved over the frontal membrane. Avicularia infrequent; pedunculate, attached to the disto-basal wall of the zooid, rostrum prominent, supporting a semi-elliptical mandible. Ovicell globular, thin walled, with an irregular, tessellate surface and a few indistinct pores. Each zooid develops a short peduncle on its distobasal wall from which the next zooid is budded, separated from the peduncle by a joint or constriction. A bifurcation is initiated when a slender, unjointed tube arises close to the origin of the new zooid and grows parallel to it. As the new zooid broadens distally a short tube develops from its proximal lateral wall, fusing with the tube adjacent to it. Above the point of fusion a joint is developed and from this a second zooid buds, diverging from the first at an angle of about 60°. Each zooid bears four lateral septula, two distal and two proximal; it is from the proximal septula that the short lateral tubes are derived; they do not arise from the distal septula, the significance of which is consequently unclear. Numerous specimens were obtained; all were alive when collected, and many bore embryos. Etymology Australis (L)—southern, referring to the geographical distribution. Remarks B. australis was formerly known by a single small specimen described by Hastings (1943) as ‘Erymophora sp. indet’. Her material was from an unknown locality. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 65 Fig. 6. A-B. Bugulella australis sp. nov. A. Part of a branch including a bifurcation. B. Two zooids in lateral view showing an avicularium. C—F. Eupaxia quadrata (Busk). C. Three zooids in frontal view. D. The ancestrula, enveloped by rhizoids. E. Zooids from the tip of a branch in basal view. F. Basal view of a dichotomy. Scale = 0,5 mm for A-C, E; 1 mm for D, F. 66 ANNALS OF THE SOUTH AFRICAN MUSEUM Leiosalpinx gen. nov. Colony erect, branching dichotomously, unjointed. Zooids in single linear series. No avicularia, spines or ovicell. Type species: Alysidium inornata Goldstein, 1882. Alysidium inornata Goldstein (= Catenaria attenuata Busk) was placed by Hastings (1943) in the genus Brettia. As presently constituted Brettia embraces a considerable number of species, from most parts of the world, whose only common feature seems to be an erect, uniserial habit. Ryland & Hayward (1977) pointed out that the type species of Brettia, B. pellucida Dyster, was based on a single, dead fragment collected at Tenby, South Wales. The type specimen, although extant, is unrecognizable and no further material has ever been collected. The definition and systematic status of Brettia is thus open to doubt and there seems to be little excuse for continuing its use, particularly in describing faunas from regions remote from the British Isles. It is appropriate, therefore, to introduce a new genus for Alysidium inornata Goldstein in the hope that a critical re-examination of the species currently assigned to Brettia Dyster may be encouraged. Etymology Leios (G)—smooth, salpinx (G)—a trumpet, referring to the shape of the zooids. Leiosalpinx inornata (Goldstein, 1882) Fig. 2C—D Alysidium inornata Goldstein, 1882: 42, pl. 1 (fig. 1). Catenaria attenuata Busk, 1884: 14, pl. 2 (fig. 1). Brettia inornata: Hastings, 1943: 476. Material Station SM 107. Description Colony erect, diffuse, branching dichotomously; zooids in single linear series, slender, very elongate, horn-shaped, calcification thin and translucent. The zooid is broadest distally, where it bears an oval opesia, comprising no more than one-sixth of the total zooid length. New zooids budded from the disto-terminal walls, linked to their predecessors by an uncalcified node. Dichotomies arise simply when the terminal zooid produces two buds, widely spaced and distolaterally orientated ; there is no connection between the two new zooids (cf. Bugulella). As noted by Hastings (1943), the surfaces of the zooids are often covered by detritus, usually obscuring the borders of the opesia. The operculum was unclear in all specimens studied. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 67 Remarks Several small colonies were found, each tangled among specimens of Bugulella australis. Distribution At present L. inornata is known only from Marion and Heard Islands. Family Farciminariidae Busk, 1852 Farciminariidae Busk, 1852: 32. Harmer, 1926: 401. Columnella Levinsen, 1914 Columnella Levinsen, 1914: 571. Levinsenella Harmer, 1926: 402. Columnella magna (Busk, 1884) Fig. 9A Farciminaria magna Busk, 1884: 49, pl. 5 (fig. 1). Columnella magna: D’Hondt, 1975a: 563. Material Station SM 67. Description Colony erect, branching dichotomously, delicate and very lightly calcified. Zooids in four longitudinal series, each comprising one facet of a quadrate branch section; rectangular, frontal surface entirely membranous, transparent, 1,4 mm long by 0,3 mm broad. Operculum thin, lightly chitinized, no spines, -ovicells not present in material collected. Avicularia small, adventitious, infrequent; occurring, rarely, at the proximal end of the zooid, mandible semi- circular, perpendicular to frontal plane. Remarks A single fragment was collected, 2 cm long and including two bifurcations. Distribution Columnella magna is a deep water species with a wide geographical distribution. Described by Busk from off Uruguay and from south of Heard Island, it has since been reported from the North Atlantic (Silén 1951, D’Hondt 1975a) and the Bay of Biscay (Hayward 1978a). Nellia Busk, 1852 Nellia Busk, 1852: 18. Harmer, 1926: 240. 68 ANNALS OF THE SOUTH AFRICAN MUSEUM Nellia sp. Fig. 2E-F Material Station SM 60. Description Colony erect; jointed? Branches quadrate in section, composed of four longitudinal series of zooids, spiralling around the branch axis. Zooids rectangular, frontal surface comprising a large oval opesia with a narrow, granular cryptocyst, raised into a crenellated mural rim. Gymnocyst reduced, bearing one or two adventitious avicularia, mandibles semicircular, directed proximolaterally. No spines. Ovicell spherical, granular and imperforate; immersed, and closed by zooidal operculum. Remarks Three very small fragments were found, none of which included a joint. Clearly attributable to Nellia, this species is distinct from other known species, but in view of the extreme paucity and fragility of the material available it is impossible to obtain a complete idea of its morphology. Consequently, it is necessary to await the collection of further specimens before a name can be applied. Family Alysidiidae Levinsen, 1909 Alysidiidae Levinsen, 1909: 201. Harmer, 1957: 641. Petalostegus Levinsen, 1909 Petalostegus Levinsen, 1909: 114. Harmer, 1957: 642. Petalostegus bicornis (Busk, 1884) Catenaria bicornis Busk, 1884: 14, pl. 2 (fig. 2). Petalostegus bicornis: Harmer, 1957: 642, pl. 51 (figs 13-18), fig. 118. Material Stations SM 1, SM 107. Description Colony erect, branching, jointed, diffuse; each internode comprising a single zooid. Each zooid spatula shaped, a slender, smooth proximal portion expanding abruptly to form a club-like distal portion containing the polypide. Frontal shield of distal portion formed of several flattened and overarched plates. New zooids budded distally and laterally. A single avicularium present on each of the distolateral corners of the zooid. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 69 Distribution Described from the Society Islands, in the mid-Pacific, P. bicornis has also been reported from two localities in Indonesia; the present records are the first from the western side of the Indian Ocean. Family Cellariidae Hincks, 1880 Cellariidae Hincks, 1880: 103. Ryland & Hayward, 1977: 119. Cellaria Ellis and Solander, 1786 Cellaria Ellis & Solander, 1786: 18. Ryland & Hayward, 1977: 119. The phylogenetic relationships of the southern hemisphere species of Cellaria present certain problems. The first of the two species described below, C. tectiformis sp. nov., has a distinctive node formation, unlike that of the type species, C. sinuosa, or, indeed, of the majority of the northern species. The second species, C. paradoxa, displays a pattern of early astogeny which is perhaps unique among described species. The joints of C. tectiformis are similar to those described in C. tecta by Harmer (1926), and in Cellariaeforma aurorae by Moyano (1969). However, Cellariaeforma Rogick is at present poorly defined and until a comparative study of the southern Cellariidae is accom- plished, the two species described here are most usefully accommodated within Cellaria Hincks. Cellaria tectiformis sp. nov. Fig. 7 Material Holotype: SAM-—A26292, station SM 23, 27°44,4’S 32°42,8’E, 400-450 m. Other material: stations SM 1, SM 16, SM 41, SM 86, SM 103. Description Colony erect, branching, jointed, forming a rigid, flat and open fan; up to 4,5 cm high in the material studied. Internodes square in section, about 0,6 mm wide, broadening in older specimens or where ovicells develop; basal internodes up to 14 mm long, distal internodes typically shorter. Dichotomies smooth, symmetrical, U-shaped: subsequently, fractures develop around the base of each new internode which is then secured to its origin by a tangled knot of rhizoids, arising from the epithecae of the zooids distal to the fracture. ' Basally, the colony is supported by a single internode, which in the largest colonies is entirely obscured by a mass of brown rhizoids. This forms a broad holdfast, securing the colony to the substratum. Zooids in four alternating longitudinal series, each comprising one face of the branch; hexagonal, truncate distally and proximally, laterally extending on to two other faces of the branch. Boundaries marked by distinct raised sutures. Cryptocyst depressed medially, delimited laterally and distally by a low ridge; opesia occupying about one- 70 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 7. Cellaria tectiformis sp. nov. A. Part of an internode, including an avicularium. B. Part of a branch, including an ovicelled zooid. C. An avicularium with an unusual orientation. D. A-new bifurcation, showing the avicularium at the base of one ramus. E. Outline diagram of colony form. F. Old bifurcations, and the base of the colony, showing rhizoids. Scales = 0,5 mm for A—C; 1 mm for D; 10 mm for E; 2,5 mm for F. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES Jb quarter of the total frontal surface, broadly crescentic, twice as wide as long, the proximal border forming a rounded or quadrate lip, beyond which two blunt, proximal denticles are visible. Ovicells endotoichal, position indicated by a broadening of the ramus in fertile areas; opening of ovicell a narrow, transverse slit, arched over by a short, domed canopy arising immediately distal to it. Avicularium vicarious, as large as an autozooid; opesia oval, comprising less than one-third of total length of heterozooid, rostrum acute, raised, supporting a hooked, triangular mandible. Avicularia are found infrequently along the length of the branch, but one occurs constantly at the base of each internode, immediately above the dichotomy. The rostrum is usually orientated transverse to the branch axis, but occasionally it may be directed distally. No complete ancestrulae were found, but in two specimens sufficient detail was visible to indicate the pattern of early astogeny. The first eight zooids budded are biserially arranged ; buds 9 and 10 assume diametrically opposite orientations to those preceding them and with buds 11 and 12 the four-faceted branch is established. Etymology Tecta—a species of Cellaria, forma (L)—shape, referring to the similarity with C. tecta. Remarks The flat, fan-like colony of this species is characteristic of several southern Cellariidae, and is in complete contrast to the dense, bushy form assumed by, particularly, the European species of Ce/laria. In colony form, structure of the joints and the form of the ovicellar orifice this species most resembles the Indo- West-Pacific Cellaria tecta Harmer (hence, tectiformis), but differs from it in the size and proportions of the zooids and avicularia, and in the structure of the _zooidal opesia. Measurements (means of 10 values) in mm Lz Iz Lop lop Lav lav 0,63 0,62 0,13 0,2 0,68 0,35 Cellaria paradoxa sp. nov. Fig. 8 ‘ Material Holotype: SAM-A26291, station SM 1, 27°00,8’S 33°03,1’E, 688 m. Other material: stations SM 86, SM 103. Description Colony erect, attached by chitinous rootlets derived from frontal eipithecae of the lowest zooids; possibly branching, although only single rami were found, AD ANNALS OF THE SOUTH AFRICAN MUSEUM without indication of nodes. Up to 1,8 cm high, with a maximum width of about | mm. Proximally the colony is very slender, 0,2-0,4 mm broad, com- mencing as a slender, fusiform ancestrula, the epitheca of which, proximally, forms a fine rootlet. First generations of buds similar in size to ancestrula, biserially arranged, frontal surfaces at approximately 45° to each other. After about eight astogenetic generations a third series of zooids develops; the branch now assumes a rectangular section, three facets being occupied by zooid frontal surfaces, the fourth comprising the lateral walls of the two outer series. After a further eight or ten astogenetic generations three more series of zooids are developed and the branch section is elongated, giving a slim, rectangular shape. The two narrowest faces are occupied each by a single series of zooids, with obliquely orientated opesiae. The third face of the branch is occupied by three alternating, longitudinal series of essentially similar zooids (‘A’ zooids), and the branch here attains its greatest width. The former basal surface is now occupied by a single series of very broad zooids (‘B’ zooids). This disposition is maintained through the whole of later growth stages. : Zooids at proximal end of colony elongate, pyriform, later approximately hexagonal. In the biserial areas of the colony a raised rim surrounds the depressed cryptocyst of each zooid. In later ‘A’ zooids this is absent, the frontal surface being generally concave between the raised lateral sutures, but they are retained in the single marginal series. In “B’ zooids there is a distinct cryptocystal rim, often completely encircling the opesiae. In all zooids the opesia is semi- circular, with a quadrate proximal lip and paired, blunt denticles. Additional paired denticles present distally, but less distinct. Opesiae larger in ‘B’ zooids than in ‘A’ zooids. Avicularia not observed, although small, triangular keno- zooids were present, each with a circular central opesia. These occurred constantly distolateral to the outer series of “A’ zooids, infrequently among the marginal zooids, and never in association with ‘B’ zooids; they occurred also among the biserial and triserial regions of the colony and appear to be the origin of the anchoring rhozoids. Etymology Paradoxus (G)—strange, referring to the zooidal dimorphism. Remarks Dimorphic zooids do not appear to have been described before for any species of Cellaria. The differences between the ‘A’ and ‘B’ zooids of C. paradoxa are so striking that the two aspects of the branch, viewed side by side, suggest two distinct species. There seems little doubt that the dimorphism is sexual in nature as ovicellar orifices occurred only among the series of “B’ zooids. However, not all of these large zooids seem to develop ovicells and other functional differences may be involved. The astogenetic changes apparent in a large colony are also extraordinary, the proximal, biserial regions closely resembling a species of Euginoma. Such is the degree of difference between THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 13 ee Fig. 8. Cellaria paradoxa sp. nov. A. Two ovicelled ‘B’ zooids. B. The reverse of the branch, ‘A’ zooids. C. The ancestrula. D. The proximal part of a colony, showing the biserial arrangement developing into a triserial form. E. Part of a branch showing the inception of ‘B’ zooids. Scale = 0,5 mm for A—-C; 1 mm for D-E. 74 ANNALS OF THE SOUTH AFRICAN MUSEUM astogenetically old and young fragments that it would not be possible to consider them to constitute a single species, if a complete colony, displaying transitional stages, had not been obtained. Measurements (means of 10 values) in mm Lz Iz Lop lop ‘A’ zooids _ 0,80 0,47 0,14 0,20 ‘*B’ zooids 0,82 0,74 0,15 0,24 Family Aspidostomatidae Jullien, 1888 Aspidostomatidae Jullien, 1888: 77. Harmer, 1926: 322. Aspidostoma Hincks, 1881 Aspidostoma Hincks, 1881: 159. Harmer, 1926: 323. Aspidostoma magna sp. nov. Fig. 9C-D Material Holotype: SAM-A26310, station SM 23, 27°44,4’S 32°42,8’E, 400-450 m. Other material: station SM 41. Description Colony erect, cylindrical, branching; stout, up to 1,5 mm wide, the largest fragment being 6 mm long. Zooids in three alternating, longitudinal series, their frontal surfaces occupying two-thirds of the circumference of the branch; hexagonal to polygonal, 1,2 to 1,4 mm long by about 0,8 mm broad, boundaries marked by raised sutures. Calcification thick, finely granular; frontal surface convex proximally, dipping distally to a deep, oval opesia occupying about one-third of the zooid length, bounded distally by a thickened, raised rim. Distolaterally, on each side, the rim is produced into a prominent, projecting lobe, often broken short. Proximal border of opesia forming a flared, bilobed lip, perpendicular to the frontal plane; a prominent conical knob proximal to it, often extended as a longitudinal rib. Ovicells large, subimmersed, with one or more frontal umbones. Etymology Magnus (L)—large, referring to the size of the zooids. Remarks The material of this species was fragmentary, but sufficient of its morphology was preserved to demonstrate its uniqueness. Few species of Aspidostoma are known, and only one living species, A. cylindricum Harmer (1926: 323), has an erect habit. A. magna is distinguished by its large size, the zooids being almost twice as large as any other species, and by the disposition of the zooid series which, in A. cylindricum, open on all surfaces of the branch. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES (BS) Fig. 9. A. Columnella magna (Busk). Part of the colony, including a bifurcation. B. Figularia philomela (Busk). Two damaged zooids and an avicularium. C—D. Aspidostoma magna sp. nov. C. A group of zooids. D. A fragment showing two ovicelled zooids. Scale = 0,5 mm for B; 1 mm for A, C-D. 716 ANNALS OF THE SOUTH AFRICAN MUSEUM Family Cribrilinidae Hincks, 1880 Cribrilinidae Hincks, 1880: 182. Hayward & Ryland, 1979: 56. Figularia Jullien, 1886 Figularia Jullien, 1886: 608. Hayward & Ryland, 1979: 70. Figularia philomela (Busk, 1884) Fig. 9B Cribrilina philomela Busk, 1884: 132, pl. 17 (fig. 6). Material Stations SM 16, SM 41. Remarks Two minute fragments only were found. One was unilaminar, but detached, and the other was bilaminar and evidently part of an erect, foliaceous colony. The zooids of both fragments were extensively damaged and all chitinous parts were missing. However, the avicularian rostrum is an important character in distinguishing the different species of Figularia, and the avicularium present in this material corresponds to that illustrated by Busk for F. philomela. This latter species, characterized by an erect, foliaceous habit, was collected by the Challenger from Marion Island. Busk distinguished encrusting material of his species, from the same locality, as var: adnata; but, as ‘hemescharan’ colonies are usually secured to the substratum by encrusting, unilaminar sheets of zooids, there seems to be little use in such a distinction. Inversiscaphos gen. nov. - Colony minute, encrusting very small substrata. Zooids budded in alter- nating distolateral radial series. Gymnocyst extensive, costae simple, fused terminally forming a median keel (hence Jnversiscaphos). Zooids communicating by large septulae (or pore-chambers with frontal windows?). Avicularia adventitious and terminal, interzooidal in position, paired. Mandibles long and setiform. Ovicells terminal, closed by the operculum, brooding zooids wide. Type species: Jnversiscaphos setifer sp. nov. Etymology Skaphe (G), scapha (L)—a boat, inversus (L)—upside down, referring to the shape of the zooids. Inversiscaphos setifer sp. nov. Figs 1B, 19 Material Holotype: SAM—A26307, station SM 16, 27°33’S 32°44,6’E, 376-384 m. Other material: station SM 16. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES T7 Description Colony minute, encrusting sand grains. Ancestrula similar to subsequent zooids, primary zooids budded proximally or proximolaterally, subsequent zooids budded in alternating lateral series. Zooids small, with a distinct peripheral gymnocyst. Two pairs of large lateral septulae (which may be part of a lateral pore-chamber complex) with uncalcified windows present and facing frontally early in ontogeny. Windows obscured distolaterally by next zooid series. Frontal shield formed by seven to eight flat, wide, simple costae, fused terminally forming a distinct keel, lacunae absent. Intercostal areas simple slits, slightly enlarged peripherally to form rounded foramina, but with no intercostal fusions. Secondary calcified orifice elongated, narrowing proximally. Avicularia adventitious, budded in pairs from the terminal wall of each zooid, not com- municating with other zooids. Basal wall of subrostral chamber reaching the substratum. Avicularia present on ancestrula, mandibles setiform, slung on asymmetrical condyles. Brooding zooids very wide, ovicell terminal, hyper- stomial, imperforate, closed by the operculum. Etymology Setifer (L)— bristly, referring to the avicularian setae. Remarks The colony structure and zooid form of J. setifer do not appear to have been described before. It is inferred that the species is adapted to life in deep water and on minute substrata, and that the setiform mandibles have a stabilizing function similar to those of Setosellina. Cribrimorph bryozoans with very long, setiform avicularia are rare (see Waters 1888: 22). Jolietina latimarginata (Busk), which encrusts corals, and dead, erect bryozoans from very deep water, also has setiform avicularia, but a different growth pattern from J. setifer, which includes large, interzooidal kenozooids. The frontal shield has numerous intercostal - fusions (see Busk 1884: 131, pl. 22 (fig. 10); Waters 1888: 22, pl. 1 (figs 11-12)). The ancestrula and primary zooids of J. setifer all have the same morphology, and the primary buds are produced proximally from the ancestrula. The avicularian chambers are terminal, and do not appear to develop distal or lateral septulae. All subsequent zooids are budded alternately from the lateral- distal septulae of preceding zooid pairs. Avicularia are thus adventitious in origin, but interzooidal in position. The occurrence of avicularia as part of the ancestrula complex is unusual although it is known in the Cupuladriidae ‘ (Lagaay 1963a). Avicularia with acute but not setiform mandibles are also known to be developed on the ancestrulae of some encrusting colonies, such as Metrarabdotos (Cook 19736). The presence of the setiform mandibles in the ancestrular complex of both the Setosellinidae and Cupuladriidae, and in I. setifer, is inferred to indicate the importance of the establishment of a clearing and stabilizing function early in astogeny, in colonies living in sandy or muddy environments. 78 ANNALS OF THE SOUTH AFRICAN MUSEUM Ovicells are prominent, and appear to be a product of the distal wall of the maternal, brooding zooid. No distal zooids are budded from the ovicelled zooid, which are present in colonies with from nine to eleven zooids. Measurements in mm Lz 0,30-0,39 Iz 0,20-0,24 Lbrz 0,39-0,41 Ibrz 0,30-0,32 Ls 0,48-0,75 Family Exochellidae Bassler, 1935 Exochellidae Bassler, 1935: 33. Hayward & Ryland, 1979: 78. Escharoides Milne-Edwards, 1836 Escharoides Milne-Edwards, 1836: 218, 259. Hayward & Ryland, 1979: 78. Escharoides distincta sp. nov. Fig. 11D Material Holotype: SAM-A26308, station SM 23, 27°44,4’S 32°42,8’E, 400-450 m. Other material: station SM 16. Description Colony encrusting. Zooids oval, convex, up to 1,2 mm long. Frontal wall thick and smooth, imperforate centrally, with a double row of marginal pores. Aperture broader than long, proximal border with a projecting, quadrate lip; a single avicularium on each side of the aperture, mandible oval, acute to plane of aperture and directed distolaterally. Two spines present on the distal border of the aperture, and one on each side; each lateral spine arises close to the avicularium and projects proximally to it. Ovicell globular, smooth, with a distinct frontal ridge. Etymology Distinctus (L)—different, referring to the distinctive character correlations of the species. Remarks Two small, damaged specimens only of this species were found, each comprising no more than six zooids. However, Escharoides is a very distinct and Fig. 10. A-D. Adeonella coralliformis O'Donoghue. A. Zooids from a growing tip. B. Later zooids. C. Diagram of the primary orifice. D. A branch viewed from the edge. E-H. Adeonella majuscula sp. nov. E. Two young zooids. F. Diagram of the primary orifice. G. Two older zooids. H. A branch viewed from the edge. I-L. Adeonella cracens sp. nov. I. Three young zooids. J. Diagram of the primary orifice. K. Three older zooids. L. A branch viewed from the edge. Scale = 0,5 mm for A-C, E-G, I-K; 1 mm for D, H, L. Us THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES Fig. 10 80 ANNALS OF THE SOUTH AFRICAN MUSEUM well-defined genus; few species are at present described and each of these is characterized by the morphology of the apertural rim, the size and slope of lateral avicularia and the number and disposition of the oral spines. On these characters E. distincta may be distinguished from all presently known species. Family Microporellidae Hincks, 1880 Microporellidae Hincks, 1880: 204. Hayward & Ryland, 1979: 220. Flustramorpha Gray, 1872 Elustramorpha Gray, 1872: 168 Busk, 1884: 135. Flustramorpha marginata (Krauss) Flustra marginata Krauss, 1837: 35, fig. 3. Flustramorpha marginata: Busk, 1884: 135, pl. 20 (fig. 8). Material Stations SM 23, SM 41. Description Colony erect, bilaminar, foliaceous. Zooids broad and flat, hexagonal, separated by raised sutures. Primary orifice semicircular. Frontal wall granular, with numerous small pores and a median crescentic ascopore. Avicularium adventitious, proximolateral to the orifice, rostrum oval, supporting a whip-like setiform mandible. Remarks Two small, damaged and very worn fragments were found, possibly an indication that they have been transported, perhaps, from shallower water. Flustramorpha angusta sp. nov. Fig. 11E Material Holotype: SAM-A26309, station SM 16, 27°33’S 32°44,6’E, 376-384 m. Description Colony erect, bilaminar, formed of narrow, strap-like branches. Zooids hexagonal, flat, separated by raised ridges; 0,7 mm long, by 0,4 mm wide. Primary orifice with straight proximal edge and curved distal edge, constituting less than a semicircle. Frontal wall minutely granular, with small scattered pores. Ascopore situated in the distal third of the zooid, immediately proximal to the orifice, crescentic, with a finely denticulate edge. Avicularium lateral, single, the cystid partially immersed in the zooid; rostrum broadly oval, with a median pivotal bar, mandible not seen. Ovicell globular, recumbent on the distally- succeeding zooid, prominent, surface granular and imperforate. The edges of the branch are channelled to receive supporting rhizoids. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 81 Etymology Angustus (L)—narrow, referring to the branches of the colony. Remarks A single specimen was found, 4 mm long and 1,5 mm broad, and conse- quently little can be said about the colony. However, this species is clearly referrable to Flustramorpha and differs from the two species presently known, F. marginata and F. flabellaris, in the slender, straplike colony form, the smaller size of the zooids, shape of the primary orifice, and in the structure of the ovicell, which is broader and bears marginal flutings in the latter two species. Measurements (means of 10 values) in mm Lz Iz 0,73 0,47 Family Gigantoporidae Bassler, 1935 Gigantoporidae Bassler, 1935: 32. Harmer, 1957: 878. Gigantopora Ridley, 1881 Gigantopora Ridley, 1881: 47. Harmer, 1957: 879. Gigantopora polymorpha (Busk, 1884) Gephyrophora polymorpha Busk, 1884: 167, pl. 34 (fig. 2). Adeonella ponticula O’Donoghue, 1924: 54, pl. 4 (fig. 23). Gigantopora polymorpha: Brown, 1952: 208, figs 145-146. Material Station SM 103. Description Colony encrusting, or developing erect, bilaminar, cylindrical, foliaceous ~ or reteporiform growths. Zooids broad, flat and polygonal, separated by raised sutures; frontal wall granular, with scattered, minute pores. Primary orifice orbicular, proximal edge concave. An elongate, triangular avicularium develops on each side of the orifice, directed medially, arching above the primary orifice and fusing to form a slender bridge. Distribution The species is known only from South Africa, although Tertiary fossil . specimens have been reported from New Zealand by Brown (1952). Family Adeonellidae Gregory, 1893 Adeonellidae Gregory, 1893: 241. Cook, 1973a: 246. Adeonella Busk, 1884 Adeonella Busk, 1884: 183. Cook, 1968: 180. 82 ANNALS OF THE SOUTH AFRICAN MUSEUM Adeonella coralliformis O’ Donoghue, 1924 Fig. 10A-D Adeonella coralliformis O’Donoghue, 1924: 55, pl. 4 (fig. 24). Adeonella coralliformis: Cook, 1973a: 254. Material Stations SM 16, SM 41, SM 86. Description Colony erect, branching, rigid. Branches flat, bilaminar, up to 4 mm broad; fragments only recovered, no complete colonies. Zooids in regular, quincuncial series, hexagonal or rhombic, becoming irregular in older parts of colony; convex, separated by shallow grooves rapidly infilled by secondary calcification. Primary orifice orbicular, with a deep, narrow sinus. Peristome low, thickened but not particularly prominent; spiramen situated immediately proximal to the secondary orifice, separated from it by a narrow bridge of calcite, thickening with age but often broken in dead material. Avicularia typically paired, lateral to secondary orifice, arising beside, or immediately distal to, the spiramen; rostrum acute triangular, directed distomedially and just extending on to the distal border of the peristome. Frontal wall finely granular, pierced by numerous small pores. Additional avicularia occur elsewhere on the frontal wall, but are not common. Vicarious avicularia, similar to and little larger than the adventitious types, may be distributed along the edge of the branch, but are not common; the edge is largely composed of normal autozooids overlapping from branch faces. Sexual polymorphs were not found, although they are described by Cook (1973a). Distribution This species is known only from South Africa. Both O’Donoghue (1924) and Cook (1973a) have commented on its resemblance to A. regularis Busk (1884), described from south-east of Cape Town, but Busk’s description and figure are not easy to understand and it is not certain that the two species are the same. Adeonella majuscula sp. nov. Fig. 1OE-H Material Holotype: SAM—A26311, station SM 86, 27°59,5’S 32°40,8’E, 550 m. Other material: station SM 41. Description Colony erect, branching, rigid, attached by an encrusting base; at least 10 cm high and probably with an equivalent spread. Branches flat, narrow, bladelike, bilaminar. Zooids in numerous, alternating longitudinal series, THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 83 E Fig. 11. A-B. Bifaxaria longicaulis Harmer. A. The proximal part of the colony showing rhizoids with their tuber-like swellings. B. The worn, main part of the specimen. C. Bifaxaria submucronata Busk. Part of the specimen. D. Escharoides distincta sp. nov. A group of zooids, including one with an ovicell. E. Flustramorpha angusta sp. nov. A group of zooids. Scale = 0,5 mm for B, E; 1 mm for A, C_D. 84 ANNALS OF THE SOUTH AFRICAN MUSEUM bifurcating frequently and diverging towards the edges of the branch, the outer series thus orientated transverse to the branch axis; oval to hexagonal, convex, separated by deep grooves. Primary orifice approximately semicircular, with a broad, shallow sinus proximally. Peristome prominent, thickened, fronto- terminal, secondary orifice orbicular; spiramen medially situated, occasionally displaced laterally, half-way between proximal border of secondary orifice and proximal edge of zooid. Frontal calcification thick and smooth, thickly punctured by numerous pores, in zooids at growing edge marginal pores are large and distinct, but are later little larger than frontal pores. Adventitious frontal avicularia single or paired, developing from marginal pores lateral, or just distal, to the spiramen; typically inclined distomedially, the elongate triangular mandible passing between spiramen and peristome. Frequently the avicularium is distally, or even proximally, directed; in older zooids they are obliterated and there is a proliferation of smaller frontal avicularia, with random orientation. Large vicarious avicularia present in continuous series along edges of the branch, and occasionally replacing marginal autozooids, rostrum identical to frontal adventitious type. | Etymology Majusculus (L)—somewhat greater, referring to the size of the zooids. Remarks In the oldest parts of the colony peristomial orifices and spiramina became deeply immersed, narrowed and eventually obliterated; the surface of the branch develops a uniform, coarsely porous surface with numerous small adventitious avicularia. Secondary calcification appears to proceed most rapidly in the central parts of the branch, which thus develops a distinct keel. Dimorphic zooids were not found in any of the material studied, but are a feature of the genus Adeonella. The most useful taxonomic characters among the species of Adeonella appear to be the shape of the primary orifice, and the position of the spiramen and the avicularia relative to the peristome, although zooid size and colony form are also important. A. majuscula differs in these respects from all presently known species; it is the largest Adeonella occurring in these collections, and the wide gap between the secondary orifice and the spiramen is perhaps its most distinctive feature. Measurements (means of 20 values) in mm Lz Iz 0,97 0,51 Distribution The genus Adeonella is widespread in tropical and subtropical waters. From a centre of distribution in the Indo-West-Pacific it ranges through the Indian Ocean, to the east and west coasts of Africa, and into the Western THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 85 Atlantic. A single species was described by Busk (1884) from South Africa (A. regularis, above), and several others by O’Donoghue (1924) and O’ Donoghue & De Watteville (1944). Adeonella cracens sp. nov. Fig. 10I-L Material Holotype: SAM-A26312, station SM 86, 27°59,5’S 32°40,8’E, 550 m. Other material: stations SM 23, SM 41. Description Colony erect, branching, rigid, attached by an encrusting base; at least 5 cm high, with an equivalent spread. Branches flat, strap-like, bilaminar, up to 6 mm broad; becoming cylindrical in older parts of the colony, where they are often narrower than later growth. Zooids in alternating linear series, bifurcating frequently and diverging towards the margin of the branch; elongate, rectangular or pyriform, becoming irregular, strongly convex and separated by deep grooves when newly developed. Primary orifice wider than long, proximal border deeply concave. Peristome low and thickened, little raised above the frontal plane of the zooid, secondary orifice approximately semicircular. Spiramen situated immediately proximal to the secondary orifice, in the distal third of the zooid; oblong, distoproximally orientated, becoming deeply immersed, the secondary opening developing a more circular outline. Avicularia single or paired, situated laterally between the spiramen and the peristome, distally directed; mandible acute triangular. Vicarious avicularia present in a continuous series along the edge of the branch, rostra identical to, and little larger than, those of the adventitious type. Frontal wall finely granular, closely punctured with numerous small pores. Dimorphic zooids were not seen. Etymology Cracens (L)—slender, referring to the shape of the branches. Remarks With continuing secondary calcification the peristome and spiramen become deeply immersed, as do the frontal avicularia. In this state the zooids have a very characteristic appearance, reminiscent of a species of Micropora. Additional ’ frontal avicularia are developed on older zooids, apparently randomly distri- buted and with no discernible orientation. Adeonella cracens differs from the other species of Adeonella in the fauna most markedly in the shape of the primary orifice and the position of the spiramen and avicularia, relative to the secondary orifice. The shape of the spiramen, though obscured in the oldest parts of the colony, is also an important character. The zooids are smaller than those of A. majuscula, and the colony as a whole tends to be less extensive, with 86 ANNALS OF THE SOUTH AFRICAN MUSEUM narrower, rather delicate branches, not developing the characteristic midrib of the latter species. Measurements (means of 20 values) in mm Lz Iz 0,97 0,34 Adeonella sp. Fig. 12E Material Stations SM 16, SM 41, SM 103. Distribution Fragments of a small species of Adeonella occurred at three stations. It could not be assigned with confidence to any described species, and appeared to be distinct from each of the three species discovered in the present survey. However, the paucity of material, and the indifferent state of preservation of the few specimens available, prevent an adequate description, and further elucidation of the systematic status of this species must await the collection of more representative samples. The specimens obtained represented Piciene (up to 4,5 mm long and 1,5 mm wide) of branches, cylindrical in section and comprising just four series of zooids, the orifices opening all around the branch axis. In a few instances the branch was broadened to six or eight series. The zooids are rectangular and elongate, smaller than any of the other species recorded here, with coarsely granular and densely punctured frontal walls. Secondary orifice semicircular, with an elliptical spiramen immediately proximal to it; avicularia paired, lateral to spiramen and orientated distally or distomedially. Marginal avicularia were not found. Family Bifaxariidae Busk, 1884 Bifaxariidae Busk, 1884: 79. Harmer, 1957: 859. Bifaxaria Busk, 1884 Bifaxaria Busk, 1884: 79. Harmer, 1957: 860. Bifaxaria submucronata Busk, 1884 Fig. 11C Bifaxaria submucronata Busk, 1884: 80, pl. 13 (fig. 1). Bifaxaria submucronata: Harmer, 1957: 861, pl. 57 (figs 1-3, 19, 22). Material Station SM 60, a single damaged internode. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 87 | C D Fig. 12. A-B. Smittoidea hexagonalis (O'Donoghue). A. Two zooids, showing different types of suboral avicularia. B. An ovicelled zooid. C—D. ?Turritigerasp. C. Zooids at the growing tip. D. Older zooids with occluded orifices. E. Adeonella sp. Two young zooids. Scale = 0,5 mm. 88 ANNALS OF THE SOUTH AFRICAN MUSEUM Description Colony erect, jointed. Internodes biserial; zooids in alternating back-to- back sequence, vase-shaped. Primary orifice terminal, obscured by a prominent, projecting proximal lip. Frontal wall smooth, with a single series of marginal pores, and a variable number of pores in longitudinal sequence frontally. Epitheca thick, clearly visible. A single, small adventitious avicularium situated on each side of the orifice, mandible semicircular, directed distolaterally. Bifaxaria ?longicaulis Harmer Fig. 11A-B Bifaxaria longicaulis Harmer, 1957: 863, pl. 57 (figs 5-6, 14-15, 17, 20). Material Station SM 31, a single damaged colony. Description Colony erect, jointed and branching; anchored by long branching rhizoids, locally swollen and forming inflated, tuber-like structures. Internodes slender, biserial, zooids in alternating, back-to-back sequence. Primary orifice sub- terminal, circular, without a pronounced peristome. Frontal wall smooth, fine- grained, with sporadic frontal and marginal pores. Adventitious avicularia small, lateral to orifice or on proximal frontal wall of adjacent zooid. Remarks The solitary specimen obtained represented the proximal portion of an old and thickened colony. Details of zooidal morphology were unclear, but in broadest terms they corresponded most closely to B. longicaulis Harmer. In particular, the curious swellings of the rootlets are described by Harmer (1957) for this species. However, the condition of the specimen prevents confident identification with the latter species. Family Cleidochasmatidae Cheetham & Sandberg, 1964 Cleidochasmatidae Cheetham & Sandberg, 1964: 1032. Cleidochasma Harmer, 1957 Cleidochasma Harmer, 1957: 1038. Cook, 1964a: 11. Several species of this genus have the ability to colonize very small sub- strata, and some appear to be confined to this type of environment. Lunulitiform colonies of C. mirabile were described by Harmer (1957: 1045, pl. 71 (figs 15, 17-18), fig. 113) from the East Indies, and the minute, encrusting C. rotundorum (Norman) was described by Cook (1964a: 20, pl. 1 (fig. 2), fig. 5B—C) from Madeira. Another similar encrusting species, C. gilchristi Cook, is known from South Africa (see Cook 1966: 212, pl. 1 (fig. 1[A—B), fig. 2A—B). C. gilchristi was reported from 101—275 metres off Durban, but has not been found in the present THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 89 South African collections, its place being taken by colonies of C. protrusum, which is otherwise known from shallow water. Cleidochasma protrusum (Thornely, 1905) Gemellipora protrusa Thornely, 1905: 119, pl. 7. Cleidochasma protrusum: Harmer, 1957: 1040, pl. 71 (figs 1-4), fig. 112. Material Stations SM 23, SM 41. Description Colony encrusting, often on small substrata. Zooids with marginal frontal pores and lateral and distal septulae. Orifices with a deep, triangular sinus and large, paired condyles. Adventitious avicularia suboral, derived from marginal septula, mandible acute, orientated laterally or proximally. Ovicells prominent, tuberculate, hyperstomial, not closed by the operculum. Remarks C. protrusum has a very wide bathymetrical and geographical range and is found on a variety of substrata. The colonies from South Africa are very small and encrust sand grains. They range in diameter from 0,90-2,00 mm and comprise from 6 to 35 zooids. Ovicells are present in colonies with only 8 zooids. The specimens described by Harmer (1957) were much larger (diameter 10 mm) and originate on shell fragments 3-4 mm in diameter. The colonies consist of multilaminar spheres formed by successive overgrowth, and as many as 16 zooid layers are present. Distribution C. protrusum has also been found in shallow-water sediments of Upper Miocene age from the Chake clay beds of Pemba, Zanzibar (British Museum Collections). The Recent distribution extends from the Indo-West-Pacific through the Indian Ocean to Mauritius and east and South Africa. Family Smittinidae Levinsen, 1909 Smittinidae Levinsen, 1909: 335. Hayward & Ryland, 1979: 98. Smittoidea Osburn, 1952 Smittoidea Osburn, 1952: 408. Hayward & Ryland, 1979: 108. Smittoidea ?hexagonalis (O’Donoghue) Fig. 12A-B Smittia hexagonalis O’Donoghue, 1924: 46, pl. 3 (fig. 15). Material Station SM 86. 90 ANNALS OF THE SOUTH AFRICAN MUSEUM Description Colony encrusting. Zooids broad and flat, quadrate or hexagonal, separated by raised sutures; 0,7-0,9 mm long by about 0,6 mm broad. Primary orifice orbicular, with a short, quadrate lyrula occupying about half the proximal border; paired, blunt and downcurved, lateral condyles present. Orifice surrounded laterally by a thin, raised peristome, incomplete proximally where it incorporates a small avicularium; mandible semicircular, directed proximally. Three or four distal oral spines present. Frontal wall smooth, thin and almost flat, with distinct marginal pores. Frontal avicularia occur sporadically, similar to the suboral type but larger. Frequently the suboral avicularium is replaced by an elongate, parallel-sided avicularium which extends from the proximal edge of the orifice laterally, almost to the edge of the zooid, the peristome being deformed in the process. Ovicell hyperstomial, thin, hyaline, with numerous small frontal pores. Remarks A single small colony only was found. In most respects it is closest to O’Donoghue’s species, described from eastern South Africa, but more informa- tion on the South African Smittinid fauna is required before a firm identification may be made. Family Tessaradomidae Jullien, 1903 Tessaradomidae Jullien, 1903: pl. 14. Hayward & Ryland, 1979: 242. Tessaradoma Norman, 1869 Tessaradoma Norman, 1869: 309. Lagaaij & Cook 1973: 494. Hayward & Ryland, 1979: 242. Tessaradoma bispiramina sp. nov. Fig. 13A—D Material Holotype: SAM-A26296, station SM 86, 27°59,5’S 32°40,8’E, 550 m. Other material: stations SM 23, SM 85, SM 103. Description Colony attached by an encrusting base, erect, cylindrical, branching irregularly; branches composed of triple whorls of zooids, tapered distally, thickening steadily by continuous frontal calcification. Typically producing paired lateral branches, at right angles to the main stem; all branches tending to curve distally. Colonies up to 1,5 cm tall, with maximum width of 2 mm. Zooids oval, boundaries marked by distinct sutures. Frontal wall thick, vitreous, with tessellated surface, distinct marginal pores present. Primary orifice orbicular, obscured by a tall, cylindrical peristome—frequently broken short—at the base THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 9] of which two tubular spiramina arise, developing separately, close together, on the outer, proximal side of the peristome. In older zooids the spiramina are covered by secondary calcification and appear to be enclosed within the peristome. Avicularia adventitious, single or paired, distolaterally situated on the frontal wall of the zooid, developing from the marginal pores; mandible semicircular, directed laterally or proximolaterally. Ovicell broader than long, smooth and imperforate, opening into the peristome; completely obscured by development of the peristome and continued secondary calcification. The ovicell is apparent only at the growing tips, or in broken areas of the colony. Etymology Spiraculum (L)—an air-hole, referring to the paired spiramina of this species. Remarks With continued calcification the boundaries of the zooids become indistinct, sutures are increasingly undulated and wander over large areas of the colony. Orifices are deeply immersed, hiding the characteristic doubled spiramen, and there is a proliferation of small adventitious avicularia. The triple zooid whorls and the doubled spiramen distinguish 7. bispiramina from other described species of Tessaradoma. Measurements (means of 10 values) in mm Lz Iz 0,8 0,4 Tessaradoma circella sp. nov. Fig. 13E-H Material Holotype: SAM-A26313, station SM 103, 28°31,7’S 32°34’E, 680 m. Other material: stations SM 16, SM 86, SM 92. Description Colony erect, slender, attached by a narrow, ring-shaped base, encircling hydroid stems; branching irregularly, up to 1 cm long in material collected, with a maximum width of 0,5 mm. Zooids in alternating; back-to-back pairs; elongate, oval, separated by raised sutures with a deep peristome constituting ‘ approximately half the zooid length. Frontal wall gently convex, smooth, with _ distinct marginal perforations; a short, tubular spiramen medially sited at the base of the peristome, in the apparent middle of the zooid. Peristome transversely oval. Avicularia adventitious, small, lateral, developed along the margins of the zooids, proximal to the peristome, apparently developing from the marginal pores; at least two, but up to six or more per zooid, mandible short, semi- circular, acute to the frontal plane of the zooid and directed laterally. Ovicell 92 ANNALS OF THE SOUTH AFRICAN MUSEUM broader than long, smooth surfaced and imperforate, opening into the peristome; rarely clearly visible, usually immersed and hidden by the peristome. The colony becomes progressively smoother, and zooid outlines indistinct, with continued secondary calcification. The primary orifice is deeply immersed and the peristomial opening lies almost completely flush with the colony surface. Marginal pores are similarly deeply immersed but the spiramen remains prominent. Additional avicularia are developed as earlier ones are obliterated. Etymology Circellus (L)—a little ring, referring to the early astogeny of the colony. Remarks Of the 7 specimens with complete bases, 5 were attached to the stems of a hydroid, 1 old colony stump had become detached from its support, and 1 was attached to another colony of the same species. In all cases, the basal portion formed a narrow ring. Although both young and old colonies were found, the astogeny was not clear. The ring in the smallest colony appeared to be formed partly by a single zooid, perhaps the ancestrula, and partly by expansions from the zooids budded from it. Secondary calcification rapidly thickens the ring and early developmental stages are thus obliterated. 7. circella is easily distinguished from other species of Tessaradoma by the very slender zooids, the length of the peristome, relative to total zooid length, and the median position of the spiramen. Its basal attachment is also unique. Measurements (means of 10 values) in mm Lz Iz 0,91 0,41 Family Sertellidae Jullien, 1903 Sertellidae Jullien, 1903: 57. Hayward & Ryland, 1979: 260. Sertella Jullien, 1903 Sertella Jullien, 1903: 57. Hayward & Ryland, 1979: 260. Sertella bullata sp, nov. Fig. 1SA—D Material Holotype: SAM-A26293, station SM 23, 27°44,4’S 32°42,8’E, 400-450 m. Other material: stations SM 16, SM 41. Fig. 13. A-D. Tessaradoma bispiramina sp. nov. A. Zooids at the tip of a colony. B. An ovicelled zooid. C. A group of young zooids, one with a developing spiramen. D. Part of an old thickened branch. E-H. Tessaradoma circella sp. nov. E. Zooids at the tip of a colony. F. Ovicelled zooids. G. The proximal part of an old colony, showing its ring-like base. H. The proximal part of a young colony, attached to a hydroid. Scale = 0,5 mm for A-F, H; 1 mm for G. 93 THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES Fig. 13 94 ANNALS OF THE SOUTH AFRICAN MUSEUM Description Colony erect, reticulate; trabeculae, up to 0,4 mm wide, composed of two alternating series of zooids, occasionally doubled where two trabeculae fuse; fenestrulae large, diamond-shaped, 1,5 mm long by 0,9 mm wide. Zooids oval, elongate, slightly convex, separated by shallow grooves, later partially immersed and less distinct. Frontal wall smooth, fine grained, with inconspicuous marginal pores, later partially covered by small granular papillae. Primary orifice broader than long, transversely oval, distal half minutely denticulate, proximal half gently concave between blunt lateral condyles. Peristome thin, erect, distal edge flared, deep, completely obscuring the orifice; a narrow median fissure extends the whole of its length, communicating with a small round pseudosinus proximally. Two short oral spines distolaterally, visible only in newly developed zooids. Adventitious avicularia present on different areas of each zooid, usually close to the margin, very numerous in older zooids; cystid inflated, rostrum fusiform, pivotal bar approximately median with a foramen on each side; mandible triangular, acute, variously orientated. There is little variation in size, although small individuals, with semicircular mandibles, occur rarely. Ovicell very large, almost as long as the bearing zooid, pear-shaped, broadest distally, slightly flattened frontally with a very narrow fissure extending almost the whole of its length. In the present material, ovicells seem to occur only on zooids immediately proximal to trabecular fusion, the ovicell being supported on the joint. Basal surface with numerous avicularia, and thickly covered with papillae. Etymology Bullatus (L)—inflated, referring to the ovicell. Remarks Species of Sertella are characteristic of the benthos of the outer continental shelf and slope in various parts of the world. Systematic problems are worsened by the often fragmentary nature of the material obtained, and earlier records are often difficult to corroborate. The Sertellidae (= Reteporidae) of the Indo-West- Pacific have been well monographed by Harmer (1934) but in most other regions they require considerable investigation. Sertella bullata may be distinguished from other described species by the relatively large size of the ovicell, the shape of the avicularia and its characteristically papillose surface. Measurements (means of 10 values) in mm LZ Iz 0,49 0,21 Reteporella Busk, 1884 Reteporella Busk, 1884: 126. Harmer, 1934: 572. This genus was created by Busk (1884: 26) for a single species, R. flabellata, collected by the Challenger from Heard Island. The two species described below THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 95 conform to Busk’s, and later Harmer’s (1934) diagnosis of the genus, but are distinctive in being bilaminate, with zooids on both sides of the flat branches. Reteporella dinotorhynchus sp. nov. Fig. 14A—D Material Holotype: SAM-A26305, station SM 23, 27°44,4’S 32°42,8’E, 400-450 m. Other material: station SM 41. Description Colony erect and branching, not reticulate. Branches flat and broad, up to 3 mm wide, largest unbranched fragment measuring 13 mm; bilaminar, zooids opening on both faces. Zooids oval to hexagonal, strongly convex and separated by deep grooves. Frontal calcification smooth and vitreous, with a few scattered, inconspicuous marginal pores. Primary orifice longer than broad, approximately bell-shaped, distal rim finely denticulate, posterior border gently concave below blunt, lateral condyles. Peristome low and thick, with a small mucro medio- proximally adjacent, on the right or left, to a shallow notch. Five or six tall, flattened and jointed (antenniform) spines around the distal and distolateral borders of the orifice. Adventitious avicularia sparsely distributed, occurring on the frontal wall immediately proximal to the peristome; cystid inflated, mandible semi-elliptical, acute to frontal surface and directed proximally. Vicarious avicularia present along the branch edges, often common, identical to adventitious type but larger. Ovicells were found only in the oldest specimens, only the lateral walls remained and nothing of their structure could be deduced. In the more proximal parts of the colony secondary calcification causes a thickening and smoothing of the colony surface, and adventitious avicularia are more frequent. Etymology Dinotos (G)—rounded, rhynchos (G)—a snout, referring to the avicularium. Remarks No complete colony was found, but a number of fragments, of both living and dead material, was obtained from each of the two stations. Measurements (means of 10 values) in mm Lz Iz 0,86 0,52 Reteporella clancularia sp. nov. Fig. 14E-I Material Holotype: SAM-A26304, station SM 23, 27°44,4’S 32°42,8’E, 400-450 m. Other material: station SM 41. 96 ANNALS OF THE SOUTH AFRICAN MUSEUM Description Colony erect, branching, not reticulate. Branches flat, narrow, commonly less than 2 mm broad, largest unbranched fragment 8 mm long; bilaminar, zooids opening on both faces of the branch. Zooids oval, convex, separated by deep grooves, distinct at the growing edges but later immersed. Frontal wall smooth, fine-grained, with a few small and inconspicuous marginal pores. Primary orifice longer than wide, bell-shaped, distal rim finely denticulate, proximal border concave below blunt lateral condyles. Peristome erect, thin, with a rounded notch proximally and frequently a small columnar avicularium adjacent to it; mandible semicircular. Six or seven slender, jointed (antenniform) spines on the distal border of the orifice. Avicularia numerous and varied, becoming more frequent in older parts of the colony. Small frontal avicularia present on most zooids, with semicircular mandible, often two or three in older zooids; enlarged avicularia, with broadly spatulate mandibles, may also occur. These large avicularia also occur consistently along the edges of the branch. Ovicell thin, hyaline, with a broad frontal fissure; prominent when newly developed, later immersed and obscured. Etymology Clancularius (L)—unknown, referring to the specific character correlations. Measurements (means of 10 values) in mm Lz Iz 0,64 0,29 Family Celleporidae Busk, 1852 Celleporidae Busk, 1852: 85. Hayward & Ryland, 1979: 274. Turbicellepora Ryland, 1963 Turbicellepora Ryland, 1963: 34. Hayward & Ryland, 1979: 284. Turbicellepora protensa sp. nov. Fig. 1SE-H Material Holotype: SAM—A26301, station SM 86, 27°59,5’S 32°40,8’E, 550 m. Other material: stations SM 23, SM 41, SM 103. Fig. 14. A—D. Reteporella dinotorhynchus sp. nov. A. Two young zooids with complete spines. B. Diagram of primary orifice. C. Later zooids, with marginal avicularia on right. D. Immersed orifices of old zooids. E-I. Reteporella clancularia sp. nov. E. Young zooids with complete spines. F. The branch edge, showing vicarious avicularia. G. Ovicelled zooids. H. Old zooids with immersed orifices. I. Diagram of primary orifice. Scale = 0,5 mm. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES Fig. 14 oT 98 ANNALS OF THE SOUTH AFRICAN MUSEUM Description Colony attached by a small encrusting base, forming a flat, branching and spreading growth; branches slender, cylindrical and tapering, up to 2,5 mm thick, with a spread of 2,5 cm in the largest specimen. Zooids oval, strongly convex, in regular series at the distal ends of the branches, elsewhere randomly orientated as branches thicken by frontal budding. Primary orifice orbicular, with a V-shaped proximal sinus, encircled by a thin, erect peristome enclosing a small avicularium; mandible semi-elliptical, acute to frontal plane of zooid and directed laterally. Frontal wall smooth, finely granular, with small, widely-spaced marginal pores. Vicarious avicularia spatulate, palate with an oval foramen, pivotal bar very thin, without a columella. Ovicell spherical, smooth, with about ten small, round pores frontally. As in all species of this genus, the appearance of the colony alters with age. Continuous frontal budding produces a multilaminar colony with orifices opening at all levels; the primary orifice becomes deeply immersed in older parts of the colony, but where the rate of frontal budding slows the peristomes are commonly worn or broken and the orifice has a more open appearance. Colonies commence branching at an early stage and the form seems to be particular to the species. Nodular, massive colonies were not found and the open fan-like growth occurred in all the specimens found. Initially attached wholly to the substratum, the branches are independent of it; the actual point of attachment is very small and in the largest specimen was not apparent at all. Etymology | Protensus (L)—extended, referring to the branching pattern of the colony. Remarks The genus Turbicellepora, and indeed a majority of the Celleporidae, presents severe systematic problems which will not be resolved without critical re-examination of described species, and of specimens of species with purportedly broad geographical ranges. Many of the Celleporidae reported from South African waters have been identified with European or north Atlantic species (e.g. O’Donoghue 1924; O’Donoghue & De Watteville 1944). Some of these reports may refer to the above species, which has not been independently described, as far as it is possible to judge, but without accurate descriptions and illustrations synonymy is impossible. Fig. 15. A-D. Sertella bullata sp. nov. A. Young zooids, with numerous avicularia. B. An ovicellate zooid. C. Basal view of a branch. D. Diagram of primary orifice. E-H. Turbicellepora protensa sp. nov. E. A group of zooids. F. Later zooids, and vicarious avicularia. G. Ovicelled zooids. H. Typical form of the colony. Scale = 0,5 mm for A-G; 2 mm for H. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES Figeals 99 100 ANNALS OF THE SOUTH AFRICAN MUSEUM Orifice dimensions (means of 20 values) in mm Lor lor 0,138 0,139 Turritigera Busk, 1884 Turritigera Busk, 1884: 129. ?Turritigera sp. Fig. 12C—D Material Stations SM 16, SM 23. Description Colony erect and branching, composed of four longitudinal series of zooids, orifices opening alternately around the whole periphery. Zooids 0,7 to 1,1 mm long by about 0,3 mm broad, convex, boundaries indistinct. Primary orifice obscured by a tall, funnel-like peristome, curving out perpendicularly from the branch when undamaged. Peristomial orifice longer than broad, approximately quadrate in shape; an adventitious avicularium on its distal border with acute triangular mandible, directed distally. A smaller avicularium occurs on the proximal border of the peristome, with a minute semicircular mandible. Additional avicularia of the second type occur elsewhere on the frontal surface. Frontal wall thick and smooth with discontinuous longitudinal striations, small marginal pores present. Secondary calcification proceeds swiftly, giving a smooth, uniform surface to the colony. Orifices of the proximalmost zooids are obliterated, with only the distal avicularium remaining visible. Remarks - Fragments of this curious species were recovered from two stations, but were insufficient to give a complete account of its morphology; none of the fragments bore ovicells, and the shape of the primary orifice could not be determined. Turritigera is represented by a single species, 7. stellata Busk, recorded from the Patagonian shelf (Busk 1884; Moyano 1974) and the Cape of Good Hope (Busk 1884). The present species differs from T. stellata principally in having zooids disposed around the whole circumference of the branch, in the latter there is a defined basal surface with zooid orifices opening only on one side of the branch. The structure of the peristome and the type and distribution of avicularia suggests an affinity with Turritigera, but conclusive evidence will be supplied only by the collection of better material. Family Vittaticellidae Harmer, 1957 Vittaticellidae Harmer, 1957: 765. Wass & Yoo, 1975: 286. Costaticella Maplestone, 1899 Costaticella Maplestone, 1899: 9. Wass & Yoo, 1975: 288. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 101 Costaticella carotica sp. nov. Fig. 16 Material Holotype: SAM-A26306, station SM 85, 27°59,5’S 32°40,8’E, 550 m. Other material: stations SM 16, SM 23, SM 86. Description Colony erect, ramifying, composed of chains of single zooids, linked proximally and distally by chitinous, tubular nodes. Branching at irregular intervals, where a short daughter zooid is budded distolaterally from a normal zooid, and initiates a new chain. Rarely, the daughter zooid may bud a second series of zooids laterally, in addition to the distal series. Zooids elongate, vase- shaped; primary orifice lepralioid, with a straight or slightly convex proximal border. Frontal wall with an elliptical series of 8-13 small, round or irregular, fenestrae, often with faint sutures extending medially from them. Proximal to the orifice, half of the area enclosed by the fenestrae is apparently formed by the fusion of seven costae; distinct lateral and medial sutures are visible between them. Scapular chambers developed as short, squat avicularia, with triangular mandibles orientated parallel to the long axis of the zooid. Suprascapular chambers short; two infrascapular chambers on each side, elongate; each with two or three uniporous pore plates. Gonozooids were not present. The colony is anchored by bundles of chitinous rootlets arising from the basal surfaces and the infrascapular chambers of the lowest zooids. Etymology Caroticus (L)—stupefying, referring to the monotonous character correlations of the genus. ~ Remarks The structure of the frontal wall, the shape of the orifice and the form of the infrascapular chambers suggest that this species is most appropriately placed in Costaticella Maplestone. It appears to be most similar to C. benecostata (Levinsen), described from southern Australia, but differs from it in its more prominent avicularia. The morphological complexities of this group of Ascophorans makes the evaluation of many early records extremely difficult _ However, C. carotica is certainly distinct from any of the Australian species recently described or redescribed by Wass & Yoo (1975), and is quite different from the few Vittaticellidae known from the east African coast. Measurements (means of 20 values) in mm Lz Iz 0,82 0,33 102 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 16. Scanning Electron Micrographs. Costaticella carotica sp. nov. A. A single zooid. B. Daughter zooid budded from a maternal zooid. C. Enlarged view of the central area of the frontal wall. D. The proximalmost infrascapular chamber, showing uniporous pore plates. Scale = 0,092 mm for A; 0,07 mm for B; 0,0874 mm for C; 0,026 mm for D. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 103 Family Mamilloporidae Canu & Bassler, 1927 Mamilloporidae Canu & Bassler, 1927: 9. Harmer, 1957: 887. Anoteropora Canu & Bassler, 1927 Anoteropora Canu & Bassler, 1927: 10. Harmer, 1957: 888. Colony lunulitiform, but anchored by basal rhizoids. Zooid series developed radially from a fan-shaped group of primary and secondary zooids derived from a single ancestrula. Zooids with greatly extended vertical walls, and small, hexagonal basal walls. Zooids communicating through septula, placed at the basal part of the zooid. Orifices large, usually with paired condyles. Avicularia, if present, interzooidal, regularly patterned. Ovicells large, hyperstomial, closed by the operculum, orifices of brooding zooids variously dimorphic. Anoteropora latirostris Silén, 1947 Anoteropora latirostris Silén, 1947: 58, pl. 5 (figs 25-27), figs 49-50. Cook, 1966: 210. Material Stations SM 16, SM 23. Description Anoteropora with large avicularia placed laterally to the orifice of both autozooids and brooding zooids. Remarks Anoteropora latirostris differs from A. inarmata (see below) in possessing avicularia, and from A. magnicapitata, the other species found in the Indian Ocean, in having avicularia associated with the brooding zooids (see Canu & Bassler 1929: 476; Harmer 1957: 888). A. smitti (Calvet), known from only two colonies from the Cape Verde Islands, is very similar to A. latirostris (Cook 1966: 291, 1968 :-183). - Distribution A. latirostris has been reported from the Red Sea and the Western Indian Ocean to the East Indies, from depths ranging from 30 to 450 m. Anoteropora inarmata Cook Figs 17C, 18C Anoteropora inarmata Cook, 1966: 211, fig. 1. ‘ Material Stations SM 53, SM 60. Description Anoteropora without avicularia. Frontal shield with a reticulate pattern of calcification. Operculum of brooding zooids with a subperipheral sclerite which is straight distally. 104 ANNALS OF THE SOUTH AFRICAN MUSEUM Remarks A. inarmata was originally described from a single colony from Zanzibar. The additional colonies from South Africa show the numerous basal rhizoids, and the pattern of early astogeny. The colonies from station SM 60 are all regenerated from fragments. . The zooids are very deep, and the basal part of each one forms a small compartment which communicates distally and laterally with other zooids. Basally, the compartment walls make a pattern of hexagonal partitions, similar in appearance to those of Cupularia guineensis (Cook 1965a: 170). The compart- ments are not kenozooidal, however, and the structure of Anoteropora is fundamentally different from that of the Cupuladriidae (Hakansson 1973). The rhizoids are numerous (Lr 0,30 mm, Ir 0,04 mm) and originate from a small septulum at the distal end of each basal zooid hexagon. The investing cuticle appears to be distended above the calcification of the frontal shields and ovicells. Distribution A. inarmata occurs from Zanzibar to eastern South Africa, from a narrow bathymetrical range, 720-810 m. Family Orbituliporidae Canu & Bassler, 1923 Orbituliporidae Canu & Bassler, 1923: 186. Cook & Lagaaij, 1976: 349. The budding patterns of the colony forms occurring in both the Orbituliporidae and Conescharellinidae (see below) were described by Cook & Lagaaij (1976). Colonies are constructed of an ancestrular complex including a kenozooidal rooting element. In the genera Batopora and Lacrimula, Conescharellina and Trochosodon, all budding is frontal, and there is no zone of astogenetic repetition. Colonies are inferred to be orientated with the ancestrular, adapical region anchored in the sediment by one or more rhizoids, and with the proliferal, antapical region, where new zooid buds are formed, facing upward. The conventional representation of these colonies with the adapical region upward is therefore the reverse of the living orientation (see Figs 17, 20). Cook & Lagaaij (1976) also discussed the combination and correlation of character states which were variously shared among species nominally assigned to Batopora, Lacrimula and Conescharellina. The type species of these genera Fig. 17. A. Heliodoma implicata Calvet. Colony from frontal side, inferred to be uppermost in life, showing setiform avicularian mandibles. x21. B. Setosellina roulei Calvet. Colony encrusting sand grain, showing setiform avicularian mandibles. x 16. C. Anoteropora inarmata Cook. Basal side of colony, showing rhizoids. x7. D. Batopora nola sp. nov. Lateral view of holotype colony, showing adapical rhizoid. x24. E. B. lagaaiji sp. nov. Lateral view of holotype colony, showing adapical rhizoid. x24. 105 THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 17 Fig 106 ANNALS OF THE SOUTH AFRICAN MUSEUM differ in many characters, but an increasing series of species has been found which has reduced the overall differences among generic groupings belonging to the two families. Batopora Reuss, 1867 Batopora Reuss, 1867: 233. Cook & Lagaaij, 1976: 349. There are a few correlated character states which are used here to distinguish Recent species, at least, of Batopora from those of Lacrimula. The existence of living forms of Batopora was unknown until relatively recently, and the increasing numbers of fossil and living species of both Batopora and Lacrimula are the direct result of examination of sea-bottom sediments. Until many more samples have been analysed, it is advisable to delay revision of genera, and both Batopora and Lacrimula are retained here. Species of Batopora have a single, relatively undifferentiated adapical rooting kenozooid which may become immersed as a ‘pit’, in contrast to the kenozooidal complex of Lacrimula. The peristomial ovicells are probably not closed by the operculum, also in contrast to those of Lacrimula. The primary orifice does not appear to have any condyles. Batopora murrayi Cook, 1966 Figs 18F, 20A Batopora murrayi Cook, 1966: 216, pl. 1 (fig. 3A—B). Cook & Lagaaij, 1976: 329, pl. 1 (fig. 2). Material | Stations SM 16, SM 23, SM 32, SM 53, SM 60, SM 78, SM 86. Description Colonies small (1,60—3,40 mm in diameter), and rather flat. Ancestrular area with a central or eccentric rooting kenozooid, slightly raised at first, becoming immersed in a ‘pit’. Whorls of four to five zooids budded alternately and becoming irregularly direct later in astogeny. Secondarily budded, small kenozooids and avicularia developing late in astogeny. Mandibles rounded, slung on paired condyles. Zooidal primary orifices straight antapically, peristome elongated and tubular in proliferal region, secondary orifices rounded. Ovicells very large, producing swollen peristomes in the proliferal zooids. Remarks Two colonies from station SM 16, with 30 and 12 zooids respectively, each have a delicate rhizoid (Lr 1,00 mm, Ir 0,20 mm) originating from the adapical Fig. 18. Scanning Electron Micrographs. A. Heliodoma implicata Calvet. x46. B. Setosellina roulei Calvet. X37. C. Anoteropora inarmata Cook. Frontal side of colony, note ovicells. x8. D. Batopora lagaaiji sp. nov. x29. E. Lacrimula pyriformis Cook. Zanzibar, BMNH 1965.8.24.12. x30. F. B. murrayi Cook. x12. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES Fig. 18 107 ANNALS OF THE SOUTH AFRICAN MUSEUM 108 a a us 06 X ‘QJe]) PI00z SUIPOOIG JO MOIA [eIDNIeT ‘C “06x “(01}U99) PIOOZ SUIPOOIG JO MIA [eWIXOId-[eJUOIT ‘DO 06x ‘“so[Apuod UBLILINSIAV [BOINEWIUUASe 9JON[ “UI9}}3ed SuIppnq SUIMOYS sprooz JUSNbesqns suWIOS pu (JYSII) B[NISOOUYW “Gq “ZI[X ‘PIIe[NOIAR [eISIp ‘poued pue vjnjdos [eiojv] “pjorys [ejUOIJ SUIMOYS PIOOZ ‘yY ‘AOU ‘dS dafijas soydvdsissaauyT “SYdeIBSOIN IAL UOII[_ Suruuvsg 6] ‘SI Di Yy MO: Yi WHOLE WHY? Ze Ly THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 109 kenozooid. Many of the colonies were alive when collected, and the dark coloured viscera are visible through the peristome walls. Uncalcified buds, appearing as ‘bubbles’ of cuticle are present in the proliferal region. Distribution B. murrayi was first described from Zanzibar, from a depth of 805 m, and subsequently reported from Fiji, from 384 m. Batopora lagaaiji sp. nov. Figs 17E, 18D, 20C Material Holotype: SAM-—A26297, station SM 53, 26°51,5’S 33°12,5’E, 720 m. Other material: stations SM 16, SM 31, SM 60, SM 69. Description Colonies very small (1,80—2,20 mm in diameter, 1,30-1,50 mm high). Adapical rooting kenzooid prominent, surrounded by five proximally directed primary zooids. Subsequent whorls are regularly composed of five alternating zooids. Small, antapical axial kenozooids present. Avicularia absent. Ovicells very large, peristomes of brooding zooids inflated but not curved. Etymology The species is named for the late Dr Robert Lagaaij, who initiated recent work on these minute bryozoan colony forms. Remarks B. lagaaiji differs from B. murrayi in the budding pattern and the absence of avicularia, and from B. nola (see below), by the number of zooids in each whorl. The holotype colony from station SM 53 has a rhizoid (Lr 0,40 m, Ir 0,10 m) emanating from the adapical kenozooid. The differences between B. lagaaiji and B. nola are small but consistent, and stem from the astogenetic pattern. Colonies of both species live under similar conditions and even occur in the same sample (SM 16). The rooted mode of Itfe, and rigid, highly integrated early astogenetic budding pattern of conescharel- liniform colonies, both make it unlikely that differences are attributable to microenvironmental (intracolony) influences. The number of zooids in each whorl is therefore regarded as genetically determined and as a specific difference between the two taxa (see below). Measurements in mm Lk 0,25-0,30 Lz,! 0,70 Iz! 0,60 lk 0,32-0,38 Iz? 0,50 Iz? 0,40 Ik aperture 0,10-0,20 lov 0,36—-0,40 1 secondary orifice 0,14 110 ANNALS OF THE SOUTH AFRICAN MUSEUM Batopora nola sp. nov. Figs 17D, 20B, 21A Material Holotype: SAM-A26298, station SM 16, 27°33’S 32°44,6’E, 376-384 m. | Other material: stations SM 16, SM 23, SM 41, SM 86. Description Colonies elongated, very small (0,90-1,10 mm in diameter, 0,80—1,00 mm high), shaped like a small handbell. Adapical rooting kenozooid large and prominent, surrounded by four proximally directed primary zooids. Sub- sequent whorls of four zooids alternating, orifices directed antapically and laterally. Small, antapical axial kenozooids present. Avicularia absent. Ovicells large, peristomes of brooding zooid curved adapically. Etymology Nola (L)—a little bell, referring to the shape of the colony. Remarks The peristomes of the autozooids are elongated and directed antapically, distinguishing colonies from those of the superficially similar species, Lacrimula pyriformis (see below). Brooding zooids are present in colonies of only twelve zooids; the peristomes are strongly curved adapically. One of the colonies (holotype) from station SM 16 has a rhizoid (Lr 0,80 mm, Ir 0,06 mm) emanating from the adapical kenozooid. Measurements in mm Lk 0,30-0,33 Lz! 0,50-0,53 Iz! 0,40-0,42 Ik 0,35-0,40 Lz? 0,30-0,35 — Iz? 0,30-0,33 Ik aperture 0,10 lov 0,25-0,30 1 secondary orifice 0,06—-0,07 Lacrimula Cook, 1966 Lacrimula Cook, 1966: 217. Cook & Lagaaij, 1976: 355. The genus was introduced for L. burrowsi Cook which was reported from depths of 101-207 m from Zanzibar and South Africa (Cook 1966: 218, pl. 2 (figs 2-4) fig. 4A). L. burrowsi has large colonies (1,0-2,0 mm in diameter, 2,8—3,2 mm high), and differs from L. pyriformis in the characters of the adapical region and the ovicells. Lacrimula pyriformis Cook, 1966 Figs 18E, 20D Lacrimula pyriformis Cook, 1966: 219, pl. 2 (fig. 1), fig. 4B. Cook & Lagaaij, 1976: 342, pl. 5 (fig. 3), pl. 6 (fig. 5). THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 111 Fig. 20. Sketches of conescharelliniform colonies, orientation in life the reverse of that shown. A. Batopora murrayi Cook. Note rhizoid arising from adapical ‘pit’. B. B. nola sp. nov. C. B. lagaaiji sp. nov. D. Lacrimula pyriformis Cook. E. Conescharellina africana Cook. F. Trochosodon sp. Scale = 1 mm. h2 ANNALS OF THE SOUTH AFRICAN MUSEUM Material Stations SM 1, SM 16, SM 23, SM 86. Description Colonies very small (0,8-1,5 mm in diameter, 1,0—2,2 mm high). Adapical kenozooidal complex with a ring of secondarily budded avicularia. Whorls of four zooids alternating, peristomes not elongated. Ovicells hyperstomial, closed by the operculum, frontal wall with a row of subpheripheral pores. Remarks Colonies of L. pyriformis are superficially similar to those of B. nola. They differ in the nature of the adapical region and ovicell, in the area of exposed frontal shield, and in the overall size of zooids. Colonies of comparable size will include 16 zooids in L. pyriformis, but only 12 zooids in B. nola. This difference is a function of the angle of zooidal axes to colony axis and is regarded as genetically determined. Zooid peristomes are not elongated and those of the brooding zooids are not curved adapically. One colony, from station SM 16, has a rhizoid originating from the adapical kenozooid (Lr 0,80 mm Ir 0,06 mm). Distribution L. pyriformis was originally reported from Zanzibar, from a depth of 310 m. Family Conescharellinidae Levinsen, 1909 Conescharellinidae Levinsen, 1909: 308. Harmer, 1957: 722. Conescharellina d’Orbigny, 1852 Conescharellina d’Orbigny, 1852: 446. Harmer, 1957: 726. Colonies have an adapical ancestrular region with secondarily budded kenozooids, heterozooids and extrazooidal tissue, including special ‘rootlet pores’. Zooids are budded frontally in alternating or direct rows, and surrounded by regularly patterned, interzooidal avicularia. Primary zooid orifices with a distinct sinus, secondary orifices with a raised lateral peristome. Ovicells hyperstomial, prominent, not closed by the operculum. Conescharellina africana Cook, 1966 Figs 20E, 21B Conescharellina africana Cook, 1966: 214, pl. 1 (fig. 2A-B), fig. 3. Fig. 21. Scanning Electron Micrographs. A. Batopora nola sp. noy. Adapical view of colony, note large adapical kenozooid. x90. B. Conescharellina africana Cook. Adapical view of colony. Note adapical kenozooidal area surrounded by avicularia, and large hyperstomial ovicell. x90, THE SOUTH AFRICAN MUSEUM’S Fig. 21 MEIRING NAUDE CRUISES 13 114 ANNALS OF THE SOUTH AFRICAN MUSEUM Material Stations SM 1, SM 16, SM 23, SM 41, SM 86. Description Colony very small (1,0-1,4 mm in diameter, 1,20-1,6 mm high) with alternating whorls of six zooids. Adapical area of kenozooids and extrazooidal tissue, surrounded by a circlet of eight to nine avicularia. Primary orifices with a narrow antapically directed sinus, peristome developed laterally and antapically. Frontal shield calcification reticulate, with large frontal septula. Avicularia paired, lateral, mandibles short and rounded, orientated laterally and slung on a complete bar. Occasionally an antapical avicularium, with antapically orientated mandible present. Ovicells globular, very prominent, developed by zooids of the third and fourth whorls late in astogeny, not present on proliferal region zooids. Ovicell frontal wall with large, irregular pores. Remarks The ovicells are not present in very young colonies, but are unusual in that they apparently develop later, on astogenetically early zooids only. C. africana also differs from other species in having an adapical region of extrazooidal tissue which is inferred to be the origin of rhizoids rather than distinct ‘rootlet pores’. Distribution C. africana was originally described from South Africa, at a depth of 101 m. Trochosodon Canu & Bassler, 1927 Trochosodon Canu & Bassler, 1927: 11. Harmer, 1957: 744. The species assigned to this genus differ very little from those placed in Conescharellina, and require revision. Trochosodon has a generally more exten- sive, extrazooidal adapical region than Conescharellina; the primary orifices are not sinuate and the zooidal peristomes are elongated and tubular. Trochosodon sp. Fig. 20F Material Station SM 60. Description Colony very small (1,40 mm in diameter, 1,0 mm high); with twenty-four zooids. Adapical region domed, with ‘rootlet pores’. Zooids in alternating series. Avicularia absent. Ovicells prominent, present on zooids of the proliferal region. THE SOUTH AFRICAN MUSEUM’S MEIJRING NAUDE CRUISES Lis Remarks One colony only of this distinctive species has been found. It greatly resembles unnamed specimens from Cape York, Queensland figured by Cook & Lagaaij (1976: 329, pl. 1 fig. 2). Examination of several other samples in the British Museum collections has revealed the presence of a complex of fossil and Recent forms from the east African area, and until these are analysed this species is left unnamed. ORDER CTENOSTOMATA The Ctenostomata are represented in these collections by one species only, a contrast to the deep-water faunas of the North-eastern Atlantic (D’Hondt 197565; Hayward & Ryland 1978; Hayward 1978a, 19786) which have recently been shown to include substantial numbers of new or little-known species. Family Flustrellidridae Bassler, 1953 Flustrellidridae Bassler, 1953: 33. Neoflustrellidra d’Hondt, 19756 Neoflustrellidra @ Hondt, 19756: 320. ?Neoflustrellidra sp. Material Station SM 86. Description Colony erect, branching profusely by dichotomy. Zooids elongated, arranged in alternating biserial series, back-to-back. Orifices not noticeably strengthened. Kenozooids absent. Remarks The material is fragmentary and shrunken. After treatment with trisodium phosphate solution, the zooids can be seen to be very similar in character and budding pattern to those of N. schopfi D’Hondt (1975b: 320, fig. 3), collected from the North Atlantic in 4 779 m. The orifices, however, lack any prominent, strengthened lower lip, and resemble those of Bockiella angusta Silén (see Hayward 1978a: 219), another erect, deep water ctenostome reported from more than 4000 m from the North-eastern Atlantic, and 135-700 m from the western Pacific. Species of Bockiella usually have regularly disposed interzooidal kenozooids (Cook 1964). Until further material becomes available this species is left unnamed. 116 ANNALS OF THE SOUTH AFRICAN MUSEUM ORDER CYCLOSTOMATA Some Cyclostomata from east Africa, and including some species from deep water, have recently been described by Brood (1976). Specimens are rare in these collections, and most fragments are too worn for identification. Family Tubuliporidae Johnston, 1838 Tubuliporidae Johnston, 1838: 247. Idmidronea Canu & Bassler, 1920 Idmidronea Canu & Bassler, 1920: 784. Harmelin 1976: 181. Idmidronea atlantica (Forbes, in Johnston, 1847) Idmonea atlantica Forbes, in Johnston, 1847: 278. Idmidronea atlantica: Harmelin, 1976: 182. Material Station SM 86. Description Colony erect, branching. Zooids arranged in alternating series of three to four facing frontally. Zooid peristomes connate, curved frontally. Basal surface of colony flat. Gonozooids frontal, slightly inflated, ooeciostome associated with the nearest, most median autozooid aperture. Remarks The complex of forms usually included in /. atlantica has a wide geographical and bathymetrical distribution, and some records have been revised by Harmelin (1976). The specimens from South Africa consist of 3 fragments, 1 of which has 3 gonozooids. Family Crisiidae Johnston, 1847 Crisiidae Johnston, 1847: 282. Crisia Lamouroux, 1812 Crisia Lamouroux, 1812: 183. Ryland, 1967: 272. Crisia aff. holdsworthi Busk, 1854 Crisia holdsworthi Busk, 1886: 6, pl. 3 (fig. 2). Material Station SM 60. Remarks This species is represented by several specimens in a good state of preserva- tion. Zooids are very attenuated, and the orifices alternating and widely spaced. No gonozooids are present, and further, fertile material 1s needed before a certain identity with Busk’s species can be established. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES EY DISCUSSION General review of collection The majority of species present in the Meiring Naude collections were cheilostomes, of which the number of anascan and cribrimorph species (20) was less than that of the ascophorans (28). The relatively low number of cyclo- stomes (2) is not unusual, but in view of the increasing number of ctenostome forms now known from deep water, it is surprising that only one species was found in this survey (see D’Hondt 1975a, 19755; Hayward & Ryland 1978). Relative abundance in terms of numbers of colonies is not easy to estimate, except in the case of those species with very small, and therefore obviously discrete, colonies. Setosellina and Batopora are good examples. In these instances the numbers of anascan colonies (333) outweigh those of the Ascophora (173). The presence of such minute colonies was established only after careful examination of sediment samples. These species are very different in appearance from most other forms of Bryozoa, and for this reason may be overlooked when benthic samples are sorted. Further, because of their small size they are probably not recovered by collecting techniques other than those specifically designed to sample benthic substrata (see also Cook 1979). Harmelin (1977) has reported a diverse fauna of nineteen species, including several rare, free-living or rooted forms, from the Canary Islands region. The sample was collected accidentally during a plankton haul at 200 m over a sandy substratum, and the bryozoans were found only when the sediment was examined. The shallow-water Bryozoa of South Africa have been reviewed by O’ Donoghue (1957) and by Day et al. (1970), but require taxonomic revision and redescription. Generally, the deeper waters off South Africa have not been investigated before. Murray (1910) gave details of several expeditions to the Indian Ocean which were made in the last century. Of these, only the Challenger expedition (1872-6) and the Valdivia (Deutschen Tiefsee Expedition, 1898-9) have produced published descriptions of the Bryozoa. None of the Challenger stations from South Africa were deeper than 270 m (see Busk 1884: viii—xi), and none of the Valdivia stations exceeded 318 m (Hasenbank 1932). Both expeditions, however, made dredge hauls in very deep water from the South Atlantic and Indian Oceans. Some Bryozoa from deep sediments collected by the John -Murray off Zanzibar, and a collection made off Durban by J. D. F. Gilchrist in 1903-4, have been described by Cook (1966). Recently, D’Hondt & Redier (1977) have ’ described some Bryozoa from the Kerguelen Islands, from depths of 29 to 270 m. The Cyclostomata of the Zanzibar area have been described by Brood (1976). Colony form and environment The correlation of bryozoan colony forms, or morphotypes, with environ- mental parameters such as substratum, turbulence and rate of sedimentation 118 ANNALS OF THE SOUTH AFRICAN MUSEUM have been discussed by Lagaaij & Gautier (1965), Cook (1968), Labracherie (1973), Harmelin (1976) and Brood (1976). The significance of colony morphotypes in ecological analyses of faunas depends on the degree to which mode of life (whether known or inferred) is reflected in colony structure. Some forms are distinctive and have a high correlation with certain environments (e.g. the lunulitiform and conescharellini- form morphotypes). Others are capable of inhabiting several types of environ- ment (e.g. the cellariiform morphotype), and have a lower correlation. In addition, some species may display one morphotype early in astogeny and another in later stages of growth. The colony morphotypes particularly asso- ciated with sea-bottoms of mud, sand or foraminiferal ooze, where substrata available for larval settlement are restricted, have been reviewed by Cook (1979). In deep waters, turbulence and sedimentation rate are low, and substrata may be limited to the sediment particles alone. Briefly, species may be adapted to direct colonization of the sediments (primary fauna), or to general and/or specific utilization of other substrata, such as hydroids, which are themselves often primary colonizers (secondary fauna). The morphotypes present in these collections fall into the following groups (see also Table 1): Membraniporiform (4 species): encrusting, unilaminar, often on stones. Except for Cleidochasma protrusum, which in these collections has minute colonies encrusting sand grains, none of the species showing this morphotype are particularly adapted to either deep water or fine particled substrata. Adeoniform and reteporiform (14 species): erect, rigid, branched, often bilaminar and arising from a small encrusting base. Colonies of some of these species originate from stones, but many have grown from flexible substrata such as hydroids or other, jointed Bryozoa. Colonies of Tessaradoma, and the ‘reteporid’ species generally, often occur on this type of substratum and are typical secondary fauna forms (Cook, 1968: 246). Flustriform (3 species): erect, bilaminar, flexible, often anchored by rhizoids. The presence of rooting systems enables these species to colonize a range of habitats, but in deep water they may be part of either primary or secondary faunas. Cellulariiform and Cellariiform (17 species): erect, flexible, jointed, branched, usually delicate and attached or anchored by rhizoids. Most of the species typical of abyssal faunas have these colony morphotypes. In shallow waters they are able to colonize unstable substrata, and it may be inferred that in deep water they are also a major part of the primary fauna. Conescharelliniform (7 species): small, conical or globular, anchored by one or more rhizoids. This morphotype is typical of the primary fauna of fine sediments, and is generally, although not exclusively, associated with deep water. Notocoryne THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 119 cervicornis is included among the conescharelliniform species, because of the small size of its colonies. Lunulitiform (3 species): cup-shaped, unilaminar, either free living and supported by setiform avicularian mandibles, or anchored by basal rhizoids. These colonies have the classical ‘sand-fauna’ morphotype and are exclusively part of the primary fauna, although those with setiform avicularia are not generally found in deep waters (cf. setoselliniform). The forms which have basal rhizoids do not seem to have ‘cleaning’ setiform avicularia, a fact possibly related to the flexibility of their attachment and to the lower sedimentation rates and turbulence of deep waters. Setoselliniform (3 species): discoid, very small, free living, supported or stabilized by setiform avicularian mandibles. This morphotype is exclusively associated with primary faunas, particularly of deep waters, and colonies are often of the same order of magnitude as the surrounding sediments. The classification of colony morphotype is, of necessity, occasionally arbitrary. For example, Notocoryne cylindracea is here considered as a cellariiform species, as is the fragment of Nellia sp., although both differ markedly from Cellaria and related genera. Turbicellepora protensa has a very unusual and interesting growth form (p. 98), but has been placed in the reteporiform group, with which it appears to have functional similarities. Of the 30 species with morphotypes particularly associated with fine-particled sea- bottoms, 25 are known or inferred to have rhizoid systems, and 3 have setiform avicularian mandibles capable of stabilizing their very small colonies. The sediments which have been examined (from stations SM 1, SM 16, SM 23, SM 31, SM 32, SM 41, SM 53, SM 60, SM 61, SM 67, SM 69, SM 78, SM 103, SM 109) vary considerably, both in the nature of the constituents and in the range of particle size (Tables 1, 3). There is no apparent correlation between sediment type and geographical or bathymetric distribution, or between sediment type and abundance and diversity of bryozoans. For example, the sediments at stations SM 31 and SM 61 consisted principally of very fine foraminiferan ooze, with a particle size of less than 1,0 mm. At stations SM 32 and SM 60, there was a preponderance of larger benthic foraminifera, with sand accreted tests, more than 5,0 mm in length or diameter. At station SM 41 there was hardly any sand, and the foraminifera were larger than 5,0 mm, whereas at SM 103 nearly all the sediment consisted of sand grains. Relatively large stones (> 20 mm diameter) were present in the samples from several stations (SM 23, . SM 67, SM 103, SM 109) from a broad depth range. Similarly, large fragments (> 20 mm length) of coelenterate skeleton were abundant at stations SM 1, SM 23 and SM 41. Significant quantities of brachiopod shell and echinoderm test were present in sediments from stations SM 1, SM 41 and SM 60. Generally, the abundance and diversity of bryozoans was highest at the two shallowest stations (SM 16, 376-384 m, and SM 23, 400-450 m), but almost as many species were present at station SM 41, from 880 m (Table 1). ANNALS OF THE SOUTH AFRICAN MUSEUM 120 a OSS 1G I OE= Shere 6 29 Siete 7S te C Ley 1 Se SC STeIOL I I I ID Ad OOfl 601 WS g 8€ (4 vl Gl g ¢ Hq OW 088 Iv WS I v) 9 ID dA 0¢8 19 WS Y) €C I I I 9 € IT H OW 018-008 09 WS I I I al d OSL 8L WS to Si I I € ID HA OL T€ WS ¢ (4 I I cl Hd OSL-O€L ce WS Vv ¢ I I I C ID dA OCL es WS ¢ Sc e I IL Eee ID HAA 889 LT WS C 6V ty 9 Ss d 089 €OI WS € te I I I YS fa 099 69 WS IL le C I I I ") Se my a. Oss 98 WS 8 8V CC I Sle ae, I Cc I (4 ID dH OSv-00r €c WS Ol 69¢ Ol CCunO esc (|b ofl I Le coke th (OIL US ale = VStaee 9T WS Ours Oly See ye eS ee Sere Oe Clas ea Ree eg eS soiseds soruojoo S 5 Be a > iS = : Sto Eh = = S = odiy/ozis w‘yideq uonrs JOON JO'ON §& aoe. 8 S = < S 5 & < ‘ Soa JUSUTPSS SI S 8 Ss ‘ = = a at iS S Ses a =. a 8 8 g W 0 92 OE OF FF OF ODO T T TT S §-— §_ sdk oydI0/W "[ JIQUL Ul Se Poyisse[d ‘UOI}e}S YORS JOJ PoJeOIPUI SI 9ZIS puw 9dA} JUSUIPag ‘saloeds Jo Joquinu ou} pue ‘so1oads [ev JO SOTUOJOO Jo JoquINU [e}0} OY} UII JOYIESO} ‘UOI}e}S YORd 3 SoIOodS YOvO JOJ pojSI] SI puNoJ saruojos Jo Joquinu oy], “YIdop ZUISBOIOUI JO JOpiO Ul posueIIeE ore suOT}E}S odAyoYdIOW 0} SUIpIO.v SdnoJs Ul posuvIIE oe SaIseds OY], “SOTUOTOS 9]NUTUT JO AjISIOATP pue souRpUNnqY ¢ ATaV THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 12] Zoogeographical and faunistic considerations The bryozoan faunas of the outer continental shelf and slope, in most of the world’s seas, are little known. Early systematic works rarely discriminate between the slope faunas and those of the shallower shelf waters, or between the slope and the abyssal depths beyond 2 000 m. However, it is becoming apparent that the composition of slope and deep shelf bryozoan faunas, as with other animal groups (Briggs 1974: 360), may differ markedly from those of both shallower and deeper waters. Recent interest in the bryozoan faunas of the continental slope between 200 and 2 000 m has been largely focused on the North-eastern Atlantic and the Meiring Naude collections presented a valuable opportunity for comparison. Such conclusions as have been reached in recent works are still tentative and subject to modification as further data accumulate. The most significant result of the present survey, namely the astonishing number of undescribed species, is both an aid and a hindrance to current discussion. The works of D’Hondt (1974, 1975a, 19756, 1977) and others have shown that even in the well-studied North- eastern Atlantic waters a significant proportion of new species may be expected from samples collected along the continental slope. The complete lack of exploration of the South African slope is demonstrated here by the discovery of a quite unknown fauna. However, the preponderance of undescribed species makes it difficult to assess their importance to current ideas, as no comment may be made on their geographical and bathymetric ranges, beyond that revealed by the survey. A further problem is the relative paucity of information on the bryozoan benthos of South African coastal waters, compared, for example, to those of Europe and North America. It is essential to have a sound background in the vertical and geographical distribution of coastal and shelf species before the significance of deep water records may be considered. For example, from present understanding of the vertical distribution of colony types, and of certain taxonomic groupings, it seems unlikely that the species of Adeonella recorded here reflect their actual bathymetric distributions within the series of samples available. It is probable that all four species are here at the lower limit of their vertical ranges. In considering the composition of slope faunas it is necessary to be able to isolate different components, demonstrating which may be shelf species, declining steadily with depth; which are perhaps truly abyssal species, at the upper limits of their range; and which species represent the indigenous slope fauna. For the most part this approach is not possible here, although some information may be gleaned from the known distributions of the minority of previously described species. It is perhaps permissible, also, for some of the new species, to draw analogies with related species in better known regions. In Table 4 the 22 previously described species are listed, together with their known geographical and bathymetric ranges, in some cases taken from specimens in the British Museum (Natural History). The bathymetric range of specimens from the Meiring Naude stations which were collected alive is also given for 122 ANNALS OF THE SOUTH AFRICAN MUSEUM TABLE 4 Geographical and bathymetric ranges for 22 species of Bryozoa. Geographical Distribution = s s 2 Bros : 5 Ss Depth range in m < S 8 . 3 - Meiring oes = FA Ss 3 S eS Naude Previously = , - S = Z 8 living known 5 2 B° So See ss) colonies A ff oo 89 eae — 8-130 x Xx D. umbellata (50-207) (x) 0) S.roulei . 376-1 300 1 900-2 330 K H. implicata 376-880 200-3 700 x N. cylindracea ' — 144-270 x L. inornata 1 000 135-270 Xx E.. quadrata 376-680 50-1 668 x x C. magna 680-700 1 900-4 850 x x x Xx P. bicornis 688-1 200 469-3 500 x F. philomela — 90-135 SK F.marginata . ; — 90-270 x x G. polymorpha 680 71-265 x A. coralliformis — 270-750 x B. submucronata . 800-810 567-1 158 x x B. longicaulis . . 740 1 158-2 796 Xx C. protrusum . . 400-450 10-18 4 < S. hexagonalis 50 48 x A. latirostris . 376-450 30-310 x x x A. inarmata 720-810 732 4 B. murrayi 376-810 805 x L. pyriformis . 376-688 310 x C. africana . 376-880 101 x Iatlanica +. 5 550 6-300 x x comparison. Most of the small, setoselliniform, lunulitiform and conescharel- liniform colonies are here reported from depths well within their known vertical range (Heliodoma, Anoteropora inarmata and Batopora) or from deeper waters (A. latirostris, Lacrimula and Conescharellina). Setosellina occurs from a very wide range, slightly shallower than that reported before. Meiring Naude results further strengthen the contention that these small animals are particularly adapted to life on the soft, unconsolidated sediments of the slope, and that the latter four genera may be restricted to a fairly narrow range of depths. Discoporella umbellata is essentially a shallower-shelf species, although the South African population, which differs in some characters from the typical form, is known from deeper water than specimens from the Mediterranean and North-eastern Atlantic (Table 4). D. umbellata is associated with coastal sandy deposits, but the present material was dead and had been transported far THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 123 beyond its normal vertical limits. Several species occurred at depths greater than their known range. In the case of Leiosalpinx inornata, Gigantopora polymorpha, Smittina hexagonalis, and the shallower record of Cleidochasma protrusum living colonies were collected, but several others were represented only by fragments (Notocoryne cylindracea, Figularia philomela, Flustramorpha marginata), suggesting that the records reflect transported debris. Both Columnella magna and Bifaxaria longicaulis normally occur in deeper waters than was sampled by the present survey. Of the newly described species, few show any immediate evidence of restricted vertical distributions: most occurred at more than one station, and many were distributed over a considerable depth range. However, Escharoides distincta was represented by fragments only at the two shallowest stations, suggesting that it might prove to be a shelf species at the lower limit of its distribution. Sertella bullata was likewise recorded only at the two shallowest stations, and the two species of Reteporellina were recorded also at two stations, at 400 m and 880 m. Much of the material of the latter two species was frag- mentary; both of these genera are more usually associated with the outer shelf (although slope species are known) and the deeper records, again, probably represent transported material. The same comments may be applied to the two species of Cellaria. Conversely, Turbicellopora protensa was collected as living material from the same two stations, and this genus is most often found in shallow coastal waters. The genus Notoplites is widespread around the world, and in the North-eastern Atlantic is represented by several deep shelf and slope species (Hayward & Ryland 1978). The virtual absence of Ctenostomata has been remarked upon before. It might be added here that, in the North-eastern Atlantic, significant numbers of deep benthic ctenostomes become evident only below 1 000 m. Table 1 shows the number of species recorded for each of the stations, and exemplifies the usual decline in species diversity with depth. The unexpectedly high number of species (18) recorded from station 41 includes many, discussed above, which were apparently dead or recorded only as fragments. Only six species (H. implicata, N. cervicornis, C. magna, T. protensa, B. nola and C. africana) were represented by specimens which may be inferred to have been alive when collected. It is interesting to note that the total fauna at each station would have been far lower without the small, free-living or rooted, species (Anoteropora, Batopora etc.). Species diversity in Bryozoa is, generally, largely a function of the availability of hard substrata and, consequently, niche diversity (Eggleston 1972). Yet, a very high proportion of this fauna was essentially independent of hard substrata, except for the smallest particles. On the basis of present knowledge it is likely that the Bryozoan fauna of the continental slope of South Africa represents an admixture of shelf species, with varying lower vertical limits, and a distinct component of slope species, whose vertical ranges may or may not be completely encompassed by the range of depths sampled here. But, with the exception of Batopora, Anoteropora, 124 ANNALS OF THE SOUTH AFRICAN MUSEUM Lacrimula and Conescharellina, it is not possible to separate the two clearly until further information becomes available. The various limitations discussed above must be considered also when discussing the geographical distribution of the species. The known geographical ranges for the described species are given in Table 4. The first three species listed, and Idmidronea atlantica, are reported for the first time remote from their presently known centres of distribution, and represent interesting new records. Four species, G. polymorpha, S. hexagonalis, A. coralliformis and C. africana, are still known only from South African waters, and several others are pre- viously known only from the South Indian Ocean (from Kerguelen, for example), or from east Africa. Generally, the Table supports the opinion that the South African marine fauna represents the western fringe of the Indo-West-Pacific zoogeographical realm. Briggs (1974) summarizes earlier works which demon- strate that the east South African fauna constitutes a distinct zoogeographical province, largely influenced by the southward-flowing, warm Agulhas current, and with a high degree of endemism. To a certain extent these collections support this suggestion. Bryozoan faunas to the south are generally well known, with considerable published data on the islands of the southern, cold-temperate Indian Ocean, and the Southern Ocean generally; yet only six of the species in this collection have been reported from these regions. To the north, Waters (1909, 1910, 1913, 1914) reported on the Bryozoa of the Sudan, east Africa and Zanzibar, but again none of the species described by him were found in this survey. It has been emphasized above that these regions are still under-studied, yet it might have been expected that a few of the species described by Waters would have been found, especially considering the common factor of the Agulhas current. Further research will no doubt result in broader geographical distributions being established for many of the species described in this paper, but it is probable that a significant proportion will prove to be endemic to this region. SUMMARY A total of 51 species of Bryozoa has been found in a collection from 17 Meiring Naude stations, ranging in depth from 376 to 1 300 m. Nearly half the species described, 23 in total, are considered to be new. In view of the large number of hitherto undescribed forms found recently in the deeper shelf and slope benthos of the North-eastern Atlantic, this high proportion of new forms is not unexpected, and it is interesting that it is similar to the proportion reported for other groups from the Meiring Naude collection (see Griffiths 1977; Millard 1977; Kensley 1978). The greatest diversity of bryozoans was found from 376 to 550 m, although several species are inferred to have been transported from shallower waters. Some deep-water forms (e.g. Setosellina roulei, Columnella magna and Bifaxaria longicaulis) have previously been reported from much greater depths, but the lower limits of bathymetric range of others (e.g. Leiosalpinx inornata, Cleido- THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 125 chasma protrusum and Conescharellina africana) have been considerably extended by these collections. Specimens which were alive when collected have provided valuable information on both early astogenetic stages and later astogenetic changes of colonies (e.g. Tricellaria varia, Cellaria paradoxa and Tessaradoma circella). Many colonies, particularly those capable of direct colonization of fine sediments, also showed evidence of rhizoid systems for anchorage. The demonstration of rhizoids in Anoteropora, Batopora and Lacrimula allows stronger inferences to be made about the environmental parameters of similarly constructed colonies over a wide range of time and space (see also Cook & Lagaaij 1976). The discovery of a setoselliniform cribrimorph species (nversiscaphos setifer) illustrates the remarkable similarity of adaptations of unrelated genera which are a response to the ‘sand fauna’ environment. The adaptations of growth form of Turbicellepora protensa are also very interesting, although as yet less understood. The abundance of minute colonies is a direct result of the detailed examination of bottom sediments, and these collections have provided large numbers of some ‘rare’ species (e.g. Heliodoma implicata) which have hitherto been known from a few colonies only. Relatively large quantities of delicate, erect branching species (e.g. Bugulella australis and Eupaxia quadrata) were also present, the last named being remarkable for its brilliant red pigmentation, which persists in preserved specimens. It is possible that the deeper shelf fauna of south-eastern Africa may prove to include a significant proportion of endemic forms. Generally, however, the area appears to constitute the extreme westerly limit of the Indo-West-Pacific faunal realm (see also Clark 1977, but compare Millard 1978). As with the Hydroida (see Millard 1977), however, a few of the Bryozoa have previously been reported only from the North-eastern Atlantic (e.g. Setosellina and Heliodoma), and this indicates both how much information has been gained from study of these collections, and how much further work remains to be done on the deeper shelf and slope faunas of the world’s seas. ACKNOWLEDGEMENTS We should like to thank Dr N. A. H. Millard and Dr P. A. Hulley (South African Museum) for the opportunity to work on the Meiring Naude collections, and Mr M. R. Fordy (University College of Swansea) and Mr P. J. Chimonides (British Museum, Natural History) for the Scanning Electron Microscopy. REFERENCES BASSLER, R. S. 1935. Fossilium catalogus 1: Animalia; Part 67, Bryozoa: 1-229. ’s-Gravenhage: W. Junk. BASSLER, R. S. 1953. Bryozoa. In: Treatise on invertebrate paleontology. Moore, R. C. ed. Part G: 1-253. New York: Geological Society of America. BriacGs, J. C. 1974. Marine Zoogeography. New York: McGraw-Hill. Broop, K. 1976. Cyclostomatous Bryozoa from the coastal waters of East Africa. Zoologica Scr. 5: 277-300. 126 ANNALS OF THE SOUTH AFRICAN MUSEUM Brown, D. A. 1952. The Tertiary Cheilostomatous Polyzoa of New Zealand. London: British Museum (Nat. Hist.). Busk, G. 1852. Catalogue of marine Polyzoa in the collection of the British Museum 1. Cheilostomata (part). London: Trustees of the British Museum. Busk, G. 1884. Report on the Polyzoa. Part 1. The Cheilostomata. Rep. scient. Results Voy. Challenger Zool. 10 (30): i—xxiii, 1-216. Busk, G. 1886. Report on the Polyzoa. Part 2. The Cyclostomata, Ctenostomata and Pedicellinea. Rep. scient. Results Voy. Challenger Zool. 17 (50): i—viti, 1-47. CaLveET, L. 1906. Note préliminaire sur les Bryozoaires recueillis par les expéditions du ‘Travailleur’ (1881-1882) et du ‘Talisman’ (1883). Bull. Mus. natn. Hist. nat., Paris 12 (3): 154-166. CALVET, L. 1907. Bryozoaires. Expéd. Sci. ‘Travailleur’ et ‘Talisman’ 1880-1883 8: 355-495. CANU, F. & BASSLER, R. S. 1917. A synopsis of American Early Tertiary Cheilostome Bryozoa. Bull. U.S. natn. Mus. 96: 1-87. CANU, F. & BASSLER, R. S. 1920. North American early Tertiary Bryozoa. Bull. U.S. natn. Mus. 106: 1-879 (2 vols). CANU, F. & BASSLER, R. S. 1923. North American Later Tertiary and Quartenary Bryozoa. Bull. U.S. natn. Mus. 125: 1-302. CANU, F. & BASSLER, R. S. 1927. Classification of the Cheilostomatous Bryozoa. Proc. U.S. natn. Mus. 69 Art. 14: 1-42. CANu, F. & BASSLER, R. S. 1929. Bryozoa of the Philippine Region. Bull. U.S. natn. Mus. 9 (100): ii—xi, 1-685. CHEETHAM, A. H. 1966. Cheilostomatous Polyzoa from the Upper Bracklesham Beds (Eocene) of Sussex. Bull. Br. Mus. nat. Hist. (Geol.) 13 (1): 1-115. CHEETHAM, A. H. & SANDBERG, P. A. 1964. Quaternary Bryozoa from Louisiana Mudlumps. J. Paleont. 38 (6): 1013-1046. CLARK, A. M. 1977. The South African Museum’s Meiring Naude Cruises. Part 4. Echinoderms. Ann. S. Afr. Mus. 73: 133-147. Cook, P. L. 1963. Observations on live lunulitiform zoaria of Polyzoa. Cah. Biol. mar. 4: 407-413. Cook, P. L. 1964a. Polyzoa from west Africa. Notes on the genera Hippoporina Neviani, Hippoporella Canu, Cleidochasma Harmer and Hippoporidra Canu & Bassler. Bull. Br. Mus. nat. Hist. (Zool.) 12 (1): 1-35. Cook, P. L. 1964b. Notes on the Flustrellidridae (Polyzoa, Ctenostomata). Ann. Mag. nat. Hist. (13) 7: 279-300. Cook, P. L. 1965a. Notes on the Cupuladriidae (Polyzoa, Anasca). Bull. Br. Mus. nat. Hist. (Zool.) 13 (5): 151-187. Cook, P. L. 1965b. Polyzoa from west Africa. The Cupuladriidae (Cheilostomata, Anasca). Bull. Br. Mus. nat. Hist. (Zool.) 13 (6): 189-227. Cook, P. L. 1966. Some ‘sand fauna’ Polyzoa (Bryozoa) from Eastern Africa and the Northern Indian Ocean. Cah. Biol. mar. 7: 207-223. Cook, P. L. 1968. Bryozoa (Polyzoa) from the coasts of tropical West Africa. Atlantide Rep. 10: 115-262. Cook, P. L. 1973a. Preliminary notes on the ontogeny of the frontal body wall in the Adeonidae and Adeonellidae (Bryozoa, Cheilostomata). Bull. Br. Mus. nat. Hist. (Zool.) 25 (6): 243-263. Cook, P. L. 1973b. Settlement and early colony development in some Cheilostomata. In: Larwoop, G. P. ed. Living and Fossil Bryozoa. London & New York: Academic Press. Cook, P. L. 1979. The potential of minute bryozoan colonies in analysis of deep sea sediments. In: Gray, J., Boucot, A. J. & Berry, W. B. N. eds. Communities of the past. Stroudsburg: Dowden, Hutchinson & Ross. (In press.) Cook, P. L. & CHIMONIDES, P. J. 1978. Observations on living colonies of Selenaria (Bryozoa, Cheilostomata). Part 1. Cah. Biol. mar. 19 (2): 147-158. Cook, P. L. & LAGAAW, R. 1976. Some Tertiary and Recent conescharelliniform Bryozoa. Bull. Br. Mus. nat. Hist. (Zool.) 29 (6): 317-376. Day, J. H., Fiecp, J. G. & PENRITH, M. J. 1970. The benthic fauna and fishes of False Bay, South Africa. Trans. R. Soc. S. Afr. 39: 1-108. DEFRANCE, J. L. M. 1823. Dictionnaire des Sciences Naturelles 27. Paris. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES i 7) EGGLESTON, D. 1972. Factors influencing the distribution of sub-littoral ectoprocts off the south of the Isle of Man (Irish Sea). J. nat. Hist. 6: 247-260. ELuis, J. & SOLANDER, D. 1786. The natural history of many curious and uncommon zoophytes. London: Benjamin White and Son. FLEMING, J. 1828. A history of British animals. Edinburgh: Bell & Bradfute. GOLDSTEIN, J. R. Y. 1882. Some new species of Bryozoa from the Marion Islands, with notes on Bicellaria grandis. Trans. Proc. R. Soc. Vict. 18: 39-46. Gray, J. E. 1848. List of the specimens of British animals in the collection of the British Museum. Part 1. Centroniae, & radiated animals. London: Trustees of the British Museum. Gray, J. E. 1872. On Flustra marginata of Krauss and an allied species, forming a new Genus (Flustramorpha). ... Ann. Mag. nat. Hist. (4) 10: 167-169. Greoory, J. W. 1893. On the British Palaeogene Bryozoa. Trans. zool. Soc. Lond. 13: 219-279. GRIFFITHS, C. 1977. The South African Museum’s Meiring Naude Cruises. Part 6. Amphipoda. Ann. S. Afr. Mus. 74: 105-123. HAKANSSON, E. 1973. Mode of growth of the Cupuladriidae (Bryozoa, Cheilostomata). In: LARwoop, G. P. ed. Living and Fossil Bryozoa. London & New York: Academic Press. HARMELIN, J.-G. 1976. Le sous-ordre des Tubuliporina (Bryozoaires Cyclostomes) en Mediteranée, écologie et systématique. Mém. Inst. océanogr. Monaco 10: 1-326. HARMELIN, J.-G. 1977. Bryozoaires du banc de la Conception (nord de Canaries). Campagnes Cineca I du ‘Jean-Charcot’. Bull. Mus. natn. Hist. nat., Paris (3) Zool. 341: 1057-1074. Harmer, S. F. 1923. On Cellularine and other Polyzoa. J. Linn. Soc. 35: 293-361. Harmer, S. F. 1926. The Polyzoa of the ‘Siboga’ Expedition. Part 2. Cheilostomata, Anasca. Siboga Exped. 28b: 181-501. Harmer, S. F. 1934. The Polyzoa of the ‘Siboga’ Expedition. Part 3. Cheilostomata, Ascophora 1, Family Reteporidae. Siboga Exped. 28c: 502-640. HARMER, S. F. 1957. The Polyzoa of the ‘Siboga’ Expedition. Part 4. Cheilostomata Ascophora. . . . Siboga Exped. 28d: 641-1147. HASENBANK, W. 1932. Bryozoen der Deutschen Tiefsee Expedition. Dt. Tief. Exped. 21 (2): 319-380. Hastincs, A. B. 1943. Polyzoa (Bryozoa) 1. Scrupocellariide, Epistomiidae, Faciminariidae, Bicellariellidae, Aetidae, Scrupariidae. ‘Discovery’ Rep. 22: 301-510. HAYWARD, P. J. 1977. The morphology of Euginoma vermiformis Jullien (Bryozoa Cheilosto- mata). J. nat. Hist. 12: 97-106. HAYWARD, P. J. 1978a. Bryozoa from the West European continental shelf. J. zool. Lond. 184: 207-224. HAyYwarbD, P. J. 19785. Two new species of Ctenostomata (Bryozoa) from the Norwegian Sea. Sarsia 63 (3): 159-162. HAYWARD, P. J. & RYLAND, J. S. 1978. Bryozoa from the Bay of Biscay and Western Approaches. J. mar. biol. Ass. U.K. 58: 143-159. HAYWARD, P. J. & RYLAND, J. S. 1979. British Ascophoran Bryozoans. Synopses of the British Fauna n.s. 14. London: Academic Press for the Linnean Society. Hincks, T. 1880. A history of the British marine Polyzoa. 2 vols. London: Van Voorst. Hincks, T. 1881. Contributions towards a general History of the Marine Polyzoa, IV. Foreign Membraniporina. Ann. Mag. nat. Hist. (5) 7: 147-161. Honpt, J.-L. D’. 1974. Bryozoaires récoltés par la ‘Thalassa’ dans le golfe de Gascogne (campagnes de 1968 a 1972). Cah. Biol. mar. 15: 27-50. HonptT, J.-L. D’. 1975a. Bryozoaires Cténostomes et Cheilostomes . . . de la campagnes océanographique Biacores du ‘Jean-Charcot’. Bull. Mus. natn. Hist. nat., Paris (3) Zool. 209: 553-600. Honpt, J.-L. D’. 19755. Bryozoaires cténostomes bathyaux et abyssaux de |’Atlantique du : Nord. Docums Lab. Géol. Fac. Sci. Lyon HS. 3 (2): 311-333. HonptT, J.-L. D’. 1977. Bryozoaires récoltés en 1972 et 1973 par les campagnes ‘Polymede IT’ en Mediterranée occidentale et ‘Thalassa’ 1973 dans le Golfe de Gascogne (Cheilostomes et Cyclostomes). Cah. Biol. mar. 18: 59-70. Honpt, J.-L. D’ & RepieR, L. 1977. Bryozoaires récoltés lors des campagnes d’été 1972 et 1974 aux iles Kerguelen .... CNFRA 42: 215-236. JOHNSTON, G. 1838. A history of British zoophytes. Edinburgh, London & Dublin. JOHNSTON, G. 1847. A history of British zoophytes. 2nd. ed 2 vols. London: John van Voorst. 128 ANNALS OF THE SOUTH AFRICAN MUSEUM JULLIEN, J. 1886. Les Costulidées, nouvelle famille de Bryozoaires. Bull. Soc. zool. Fr. 11: 601-620. JULLIEN, J. 1888. Bryozoaires. Mission scient. Cap. Horn 1882-1883 6 (1): 1-92. JULLIEN, J. 1903. Bryozoaires provenant des campagnes de I’Hirondelle (1886-1888). Résult. Camp. scient. Prince Albert I 23: 1-188. KENSLEY, B. 1978. The South African Museum’s Meiring Naude Cruises. Part 7. Isopoda. Ann. S. Afr. Mus. 74: 125-157. Krauss, C. F. F. 1837. Beitrag zur Kenntniss der Corallineen und Zoophyten der Siidsee nebst Abbildungen der neuen Arten. Stuttgart. LABRACHERIE, M. 1973. Functional morphology and habitat of Bryozoa in the Eocene of the North Aquitaine Basin. In; Larwoop, G. P. ed. Living and Fossil Bryozoa. London & New York: Academic Press. LAGAAIJ, R. 1952. The Pliocene Bryozoa of the Low Countries. Meded. geol. Sticht. (C) 5 (5): 1-233. LAGAAL, R. 1963a. Cupuladria canariensis (Busk)— portrait of a Bryozoan. Palaeontology 6 (1): 172-217. LAGAAI, R. 1963b. New additions to the bryozoan fauna of the Gulf of Mexico. Pubs. Inst. mar. Sci. Univ. Tex. 9: 162-236. LAGAAIJ, R. & Cook, P. L. 1973. Some Tertiary to Recent Bryozoa. In: HALLAM, A. Atlas of Palaeobiogeography. Amsterdam, London & New York.: Elsevier. LAGAAI, R. & GAUTIER, Y. V. 1965. Bryozoan assemblages from marine sediments of the Rhone delta, France. Micropaleontology 11 (1): 39-S8. LAMOUROUX, J. V. F. 1812. Extrait d’un mémoire sur la classification des Polypiers corallingénes non enti¢rement pierreux. Nouy. Bull. Scient. Soc. Phil. 3: 181-188. LEVINSEN, G. M. R. 1909. Morphological and Systematic Studies on the Cheilostomatous Bryozoa. Copenhagen. LEVINSEN, G. M. R. 1914. Bryozoa, Endoprocta, Pterobranchia og Enteropneusta [of Greenland]. Meddr Gronland 23: 547-634. Louw, E. 1977. The South African Museum’s Meiring Naude Cruises. Part 1. Station data, 1975, 1976. Ann. S. Afr. Mus. 72: 147-159. MAPLESTONE, C. M. 1899. Further descriptions of the Tertiary Polyzoa of Victoria, II. Proc. R. Soc. Vict. n.s. 12 (1): 1-13. Maturo, F. & Scuopr, T. J. M. 1968. Ectoproct and Entoproct type material. Postilla 120: 2-95. MILLARD, N. A. H. 1977. The South African Museum’s Meiring Naude Cruises. Part 3. Hydroida. Ann. S. Afr. Mus. 73: 105-131. MILLARD, N. A. H. 1978. The geographical distribution of southern African hydroids. Ann. S. Afr. Mus. 74: 159-200. MILNE-EDWarbDs, H. 1836. Histoire des Polypes. In: LAMARCK, J. B. P. A. ed. Histoire naturelle des Animaux sans Vertebres. 2nd ed. 2. Paris. MoyAano, H. 1969. Bryozoa colectados por la Expedicion Antarctica Chilena 1964-65. III. Familia Cellariidae Hincks, 1880. Boln. Soc. Biol. Concepcion 41: 41-77. Moyano, H. 1974. Briozoos marinhos chilenos II, Briozoos de Chile austral I. Gayana Zool. 30: 3-41. Murray, J. 1910. On the depth and marine deposits of the Indian Ocean... . Reps Percy Sladen Exped. Indian Ocean 2. Trans. Linn. Soc. Lond. Zool. 13 (24): 355-396. Norman, A. M. 1869. Last report on dredging among the Shetland Islands. Polyzoa. Rep. Br. Ass. Advmt. Sci. 1868: 303-312. O’Donocuug, C. H. 1924. The Bryozoa (Polyzoa) collected by the S.A. ‘Pickle’. Rep. Fish. mar. biol. Sury. Un. S. Afr. 3 Spec. Rep. 10: 1-63. O’DonoGuuE, C. H. 1957. Some South African Bryozoa. Trans. R. Soc. S. Afr. 35: 71-95. O’DonocuuE, C. H. & DE WATTEVILLE, D. E. 1944. Additional notes on Bryozoa from South Africa. Ann. Natal Mus. 10 (3): 407-432. ORBIGNY, A. D’. 1852. Paléontologie Francaises, Terrains Crétacés, V, Bryozoaires: 185-492. Paris: Masson. OsBurn, R. C. 1952. Bryozoa of the Pacific Coast of America, 2. Cheilostomata Ascophora. Allan Hancock Pacif. Exped. 14: 271-611. PRENANT, M. & BoBIN, G. 1966. Bryozoaires. Part 2. Chilostomes Anasca. Faune Fr. 68: 1-647. THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES 129 Reuss, A. E. 1867. Die Bryozoen, Anthozoen und Spongarien des braunen Jura von Balin bei Krakau. Denkschr. Akad. Wiss. Wien math.—naturw. Cl. 27 (1): 1-26. RIDLey, S. O. 1881. Account of the Zoological Collections ... of H.M.S. ‘Alert’... . Part V. Polyzoa. Proc. zool. Soc. Lond. 1881: 44-61. RYLAND, J. S. 1963. Systematic and biological studies on Polyzoa (Bryozoa) from western Norway. Sarsia 14: 1-59. RYLAND, J. S. 1967. Crisiidae from western Norway. Sarsia 29: 269-282. RYLAND, J. S. 1969. A nomenclatural index to ‘A History of the British Marine Polyzoa’ by T. Hincks (1880). Bull. Br. Mus. nat. Hist. (Zool.) 17 (6): 207-260. RYLAND, J. S. 1970. Bryozoans. London: Hutchinson University Library. RYLAND, J. S. & HAywarp, P. J. 1977. British Anascan Bryozoans. Synopses of the British Fauna n.s. 10. London: Academic Press for the Linnean Society. SILEN, L. 1947. Conescharellinidae (Bryozoa Gymnolaemata) collected by Prof. Dr. Sixten Bock’s Expedition to Japan and the Bonin Islands, 1914. Ark. Zool. 39A (9): 1-61. SILEN, L. 1951. Bryozoa: Rep. Swed. deep Sea Exped. 2. Zool. (5): 63-69. SmiiT, F. A. 1868. Kritisk forteckning 6fver Skandinaviens Hafs-Bryozoer, III. Ofvers. K. Vetensk Akad. Forh. 24: 279-429. THORNELY, L. R. 1905. Report on the Polyzoa. In: W. A. Herdman ed. Rep. Pearl Oyster Fisheries, Gulf of Manaar (4) Suppl. Rep. 26: 107-130. VERRILL, A. E. 1879. Brief contributions to zoology from the museum of Yale College. No. 42. Notice of recent additions to the marine fauna of the eastern coast of North America, No. 5, Polyzoa. Amer. J. Sci. (3) 17: 472-474. Wass, R. E. & Yoo, J. J. 1975. Distribution and taxonomy of some Recent catenicelliform Bryozoa from Australia. Docums Lab. Géol. Fac. Sci. Lyon HS. 3 (2): 281-297. Waters, A. W. 1888. Supplementary report on the Polyzoa.... Rep. scient. Results Voy. Challenger Zool. 79: 1-41. Waters, A. W. 1909. Reports on the Marine Biology of the Sudanese Red Sea.. ., XII. The Bryozoa, Part I, Cheilostomata. J. Linn. Soc. (Zool.) 31: 123-181. Waters, A. W. 1910. Reports on the Marine Biology of the Sudanese Red Sea .. ., XII. The Bryozoa, Part II, Cyclostomata, Ctenostomata and Endoprocta. J. Linn. Soc. (Zool.) 31: 231-256. Waters, A. W. 1913. The Marine Fauna of British East Africa and Zanzibar. . ., Bryozoa— Cheilostomata. Proc. zool. Soc. Lond. 1913: 458-537. Waters, A. W. 1914. The Marine Fauna of British East Africa and Zanzibar . . ., Bryozoa— Cyclostomata, Ctenostomata and Endoprocta. Proc. zool. Soc. Lond. 1914: 831-858. ABBREVIATIONS Lz length of zooid Iz width of zooid Lap length of aperture Lop length of opesia lop width of opesia Ls length of seta Lbrz length of brooding zooid Ibrz width of brooding zooid Lov length of ovicell lov width of ovicell Lav length of avicularium lav width of avicularium Lvic.av length of vicarious avicularium : Lr length of rhizoid Ir width of rhizoid Lk length of kenozooid Ik width of kenozooid 130 ANNALS OF THE SOUTH AFRICAN MUSEUM APPENDIX 1 Meiring Naude stations which produced Bryozoa Station Co-ordinates Depth, m Date og of | SM 1 DT OS: Sessile 688 23.519 SM 16 2IBSi 32°44,6’ 376-384 D5 SM 23 27°44,4’ 32°42,8' 400-450 2635.75 SM 31 28°4,5’ 32°42,8’ 740 27.5015 SM 32 2336.2 32°43;5° 730-750 21-5415 SM 41 28°41,7’ 32°34,5’ 880 29S SM 53 26-5 1,14 Son 12s5, 720 18.5.76 SM 60 27 936) SY) Sy 800-810 19°5:76 SM 61 2h 10;3% BP Sol 820 195576 SM 67 27°14,8’ 32°54,6’ 680-700 20.5.76 SM 69 QT 1222: 32°56’ 660 20.5.76 SM 78 27 Slko* 32750! 750 DiEDS16 SM 85 Des9-5) 32°40,8’ 550 22-5516 SM 86 25955" 32°40,8’ 550 22.5:16 SM 92 28°14,5’ 32°40,6’ - 650-720 23:5.76 SM 103 2337: 32°34’ 680 24.5.76 SM 107 23318 32°38,4’ 1 000-1 200 20-16 SM 109 28°41’ 3236;8- 1 300 2525416 6. SYSTEMATIC papers must conform to the Jnternational code of zoological nomenclature (particularly Articles 22 and 51). Names of new taxa, combinations, synonyms, etc., when used for the first time, must be followed by the appropriate Latin (not English) abbreviation, e.g. gen. nov., sp. nov., comb. nov., syn. nov., etc. An author’s name when cited must follow the name of the taxon without intervening punctuation and not be abbreviated; if the year is added, a comma must separate author’s name and year. The author’s name (and date, if cited) must be placed in parentheses if a species or subspecies is transferred from its original genus. The name of a subsequent user of a scientific namé must be separated from the scientific name by a colon. Synonymy arrangement should be according to chronology of names, i.e. all published scientific names by which the species previously has been designated are listed in chronological order, with all references to that name following in chronological order, e.g.: Family Nuculanidae Nuculana (Lembulus) bicuspidata (Gould, 1845) Figs 14-15SA Nucula (Leda) bicuspidata Gould, 1845: 37. Leda plicifera A. Adams, 1856: 50. Laeda bicuspidata Hanley, 1859: 118, pl. 228 (fig. 73). Sowerby, 1871: pl. 2 (fig. 8a—b). Nuceula largillierti Philippi, 1861: 87. Leda bicuspidata: Nicklés, 1950: 163, fig. 301; 1955S: 110. Barnard, 1964: 234, figs 8-9. Note punctuation in the above example: comma separates author’s name and year “semicolon separates more than one reference by the same author full stop separates references by different authors figures of plates are enclosed in parentheses to distinguish them from text-figures dash, not comma, separates consecutive numbers Synonymy arrangement according to chronology of bibliographic references, whereby the year is placed in front of each entry, and the synonym repeated in full for each entry, is not acceptable. In describing new species, one specimen must be designated as the holotype; other speci- mens mentioned in the original description are to be designated paratypes; additional material not regarded as paratypes should be listed separately. The complete data (registration number, depository, description of specimen, locality, collector, date) of the holotype and paratypes must be recorded, e.g.: Holotype SAM-—A13535 in the South African Museum, Cape Town. Adult female from mid- tide region, King’s Beach Port Elizabeth (33°51’S 25°39’E), collected by A. Smith, 15 January 1973. Note standard form of writing South African Museum registration numbers and date. 7. SPECIAL HOUSE RULES Capital initial letters ~(a) The Figures, Maps and Tables of the paper when referred to in the text e.g. *... the Figure depicting C. namacolus ...’; *. . . in C. namacolus (Fig. 10)...’ ' (b) The prefixes of prefixed surnames in all languages, when used in the text, if not preceded by initials or full names e.g. Du Toit but A.L.du Toit; Von Huene but F. von Huene (c) Scientific names, but not their vernacular derivatives e.g. Therocephalia, but therocephalian Punctuation should be loose, omitting all not strictly necessary Reference to the author should be expressed in the third person Roman numerals should be converted to arabic, except when forming part of the title of a book or article, such as ‘Revision of the Crustacea. Part VIII. The Amphipoda.’ Specific name must not stand alone, but be preceded by the generic name or its abbreviation to initial capital letter, provided the same generic name is used consecutively. Name of new genus or species is not to be included in the title: it should be included in the abstract, counter to Recommendation 23 of the Code, to meet the requirements of Biological Abstracts. P. J. HAYWARD & P. L. COOK THE SOUTH AFRICAN MUSEUM’S MEIRING NAUDE CRUISES PART 9 BRYOZOA _ ANNALS OF THE SOUTH AFRICAN MUSEUM CAPE ‘TOWN ~ INSTRUCTIONS TO AUTHORS 1. MATERIAL should be original and not published elsewhere, in whole or in part. 2. LAYOUT should be as follows: (a) Centred masthead to consist of Title: informative but concise, without abbreviations and not including the names of new genera or species Author’s(s’) name(s) Address(es) of author(s) (institution where work was carried out) Number of illustrations (figures, enumerated maps and tables, in this order) (b) Abstract of not more than 200 words, intelligible to the reader without reference to the text (c) Table of contents giving hierarchy of headings and subheadings (d) Introduction (e) Subject-matter of the paper, divided into sections to correspond with those given in table of contents (f) Summary, if paper is lengthy (g) Acknowledgements (h) References (i) Abbreviations, where these are numerous 3. MANUSCRIPT, to be submitted in triplicate, should be typewritten and neat, double spaced with 2,5 cm margins all round. First lines of paragraphs should be indented. Tables and a list of legends for illustrations should be typed separately, their positions indicated in the text. All pages should be numbered consecutively. Major headings of the paper are centred capitals; first subheadings are shouldered small capitals; second subheadings are shouldered italics; third subheadings are indented, shouldered italics. Further subdivisions should be avoided, as also enumeration (never roman numerals) of headings and abbreviations. Footnotes should be avoided unless they are short and essential. Only generic and specific names should be underlined to indicate italics; all other marking up should be left to editor and publisher. 4. ILLUSTRATIONS should be reducible to a size not exceeding 12 x 18 cm (19 cm including legend); the reduction or enlargement required should be indicated; originals larger than 35 x 47 cm should not be submitted; photographs should be rectangular in shape and final size. A metric scale should appear with all illustrations, otherwise magnification or reduction should be given in the legend; if the latter, then the final reduction or enlargement should be taken into consideration. All illustrations, whether line drawings or photographs, should be termed figures (plates are not printed; half-tones will appear in their proper place in the text) and numbered in a single series. Items of composite figures should be designated by capital letters; lettering of figures is not set in type and should be in lower-case letters. The number of the figure should be lightly marked in pencil on the back of a illustration. 5. REFERENCES cited in text and synonymies should all be included in the list at the end of the paper, using the Harvard System (ibid., idem, loc. cit., op. cit. are not acceptable): (a) Author’s name and year of publication given in text, e.g.: ‘Smith (1969) describes...’ ‘Smith (1969: 36, fig. 16) describes . ‘As described (Smith 1969a, 19695; at sp ‘As described (Haughton & Broom oe ‘As described (Haughton et al. 1927) . Note: no comma separating name and: year Pagination indicated by colon, not p. names of joint authors connected by ampersand et al. in text for more than two joint authors, but names of all authors given in list of references. (b) Full references at the end of the paper, arranged alphabetically by names, chronologically within each name, with suffixes a, b, etc. to the year for more than one paper by the same author in that year, e.g. Smith (1969a, 19695) and not Smith (1969, 1969a). For books give title in italics, edition, volume number, place of publication, publisher. For journal article give title of article, title of journal in italics (abbreviated according to the World list o, scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses, volume number, part number (only if independently paged) in parentheses, pagination (first and last pages of article). Examples (note capitalization and punctuation) BULLOUGH, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan. FISCHER, p. —H. 1948. Données sur la résistance et de le vitalité des mollusques. J. Conch., Paris 88: 100-140. FiscHerR, P.-H., DuvAL, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines. Archs Zool. exp. gen. 74: 627-634. Koun, A. J. 1960a. Ecological notes on Carus (Mollusca: Gastropoda) in the Trincomalee region of Ceylon. Ann. Mag. nat. Hist. (13) 2: 309-320. Konn, A. J. 19605. Spawning behaviour, egg masses and larval development in Conus from the Indian Ocean. Bull. Bingham oceanogr. Coll. 17 (4): 1-51. THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. In: SCHULTZE, L. Zoologische und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-Afrika 4: 269-270. Jena: Fischer. Denkschr. med.-naturw. Ges. Jena 16: 269-270. (continued inside back cover) ANNALS OF THE SOUTH AFRICAN MUSEUM ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM Volume 79 Band September 1979 September Part 5 . Deel A REDESCRIPTION OF THE KELP CURLER _ AMPITHOE HUMERALIS (CRUSTACEA, AMPHIPODA) FROM SOUTH AFRICA AND ITS RELATIONSHIP TO MACROPISTHOPOUS By CHARLES L. GRIFFITHS Cape Town Kaapstad The ANNALS OF THE SOUTH AFRICAN MUSEUM are issued in parts at irregular intervals as material becomes available Obtainable from the South African Museum, P.O. Box 61, Cape Town Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM word uitgegee in dele op ongereelde tye na gelang van die beskikbaarheid van stof Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad OUT OF PRINT/UIT DRUK 1, 2(1-3, 5-8), 3(1-2, 4-5, 8, t.—p.i.), 5(1-3, 5, 7-9), 6(1, t.-p.i.), 7(1-4), 8, 9(1-2, 7), 10(1-3), 11(1-2, 5, 7, t.-p.i.), 15(4-5), 24(2), 27, 31(1-3), 32(5), 33 Copyright enquiries to the South African Museum Kopieregnavrae aan die Suid-Afrikaanse Museum ISBN 0 908407 81 5 Printed in South Africa by In Suid-Afrika gedruk deur The Rustica Press, Pty., Ltd., Die Rustica-pers, Edms., Bpk., Court Road, Wynberg, Cape Courtweg, Wynberg, Kaap A REDESCRIPTION OF THE KELP CURLER AMPITHOE HUMERALIS (CRUSTACEA, AMPHIPODA) FROM SOUTH AFRICA AND ITS RELATIONSHIP TO MACROPISTHOPOUS By CHARLES L. GRIFFITHS Zoology Department, University of Cape Town (With 3 figures) [MS. accepted 24 July 1979] ABSTRACT Ampithoe humeralis Stimpson, to date known only from west North America, is redescribed and figured from material collected from the west coast of the Cape Peninsula, South Africa. Individuals of this species appear to co-operate in the construction of nests made from the folded fronds of living kelp plants and to consume the walls of these nests while progressively extending them back along the fronds. The relationship between Ampithoe and Macropisthopous K. H. Barnard 1916, is discussed and Macropisthopous reduced to a junior synonym of Ampithoe. CONTENTS PAGE Introduction q : ‘ eto Description of material . me US Notes on ecology. : ew 3 Relationships. : ; ha Si Acknowledgements . : ee ley, References . ; 3 ; sft Leh A INTRODUCTION Amphipods of the family Ampithoidae are usually large-bodied, con- spicuous species normally associated with algae in the intertidal or shallow sublittoral zones. Most members of the group are tubicolous, spinning soft parchment-like tubes amongst the holdfasts and fronds of various algae (J. L. Barnard 1965). Exceptions include Pseudoamphithoides incurvaria (Just, 1977) (formerly Amphyllodomus—Just’s 1977 genus being synonymous with that of Ortiz 1976), which cuts oval sections from algal fronds and glues these together to form a ‘mobile-home’, and the kelp curler, Ampithoe humeralis, which constructs nests from the curled blades of the giant kelp Macrocystis (Jones 1971) or lives in Macrocystis holdfasts (J. L. Barnard 1969a). Ampithoe humeralis has to date been recorded only from the west coast of North America, where it may on occasion reach sufficient densities significantly to damage the Macrocystis beds (North 1971). The range of A. humeralis is 131 Ann. S. Afr. Mus. 79 (5), 1979: 131-138, 3 figs. 132 ANNALS OF THE SOUTH AFRICAN MUSEUM extended herein to include the west coast of South Africa, where it is associated with a different species of kelp—Ecklonia maxima—and appears to have somewhat modified its habits accordingly. A redescription from this material has highlighted the similarity between A. humeralis and the monospecific genus Macropisthopous K. H. Barnard and necessitated a re-examination of the taxonomic status of that genus. DESCRIPTION OF MATERIAL Ampithoe humeralis Stimpson, 1864 Figs 1-3 Amphithoe humeralis Stimpson, 1864: 156. Calman, 1898: 271, pl. 33 (fig. 4). Ampithoe humeralis: Stebbing, 1906: 636. J. L. Barnard, 1954: 29; 1965: 7, figs 2-3; 1969a: 83. Description (of male, 15 mm) Head length of first two pereon segments, rostrum absent, ocular lobes rounded, eyes oval, red; antenna | length of pereon, articles 1 and 2 subequal, sparsely setose, article 3 a quarter size of 2, accessory flagellum absent, flagellum 35-articulate; antenna 2 approximately 60 per cent length of 1, sparsely setose, gland cone small, acute, articles 4 and 5 of peduncle equal in size, flagellum 16-articulate. Upper lip apically rounded, distally densely setulose; article 3 of mandibular palp shorter than 2, bearing eight terminal pectinate setae, primary cutting edge of mandible of eight teeth, lacinia mobilis with seven teeth, spine row of thirteen spines, molar triturative; lower lip with outer lobes deeply notched, outer portion considerably the larger; inner plate of maxilla 1 with one apical plumose seta, outer plate with ten strong terminal spines ranging from thick and smooth laterally to slender and strongly comb-like medially, palp exceeding outer plate, biarticulate, article 2 with six apical spines and one pectinate seta; plates of maxilla 2 equal in length, inner more slender than outer and setose medially as well as apically; maxilliped bearing stout 4-articulate palp, outer plate extending to centre of article 3 of palp, thirteen short stout serrate spines along medial margin, seven pectinate setae on lateral margin, inner plate with numerous medial and apical setae but no spines. Pereon dorsally smooth, coxae 1-4 progressively longer, oval, ventrally rounded, a few setae at posterodistal corner; coxa 5 slightly deeper than 4, with broad posterior lobe, 6 bilobed, 7 semicircular; gnathopods 1 and 2 of similar structure, 2 slightly the larger and heavier, article 2 not lobed, 5 slightly lobed in gnathopod 1, more strongly in 2, strongly setose posteriorly, article 6 hardly wider than 5, slightly chelate, margin of palm minutely crenulate, defined by a small spine largely concealed by dense pectinate setae along palm and posterior margin, dactyl almost twice length of palm, bearing closely appressed serrations; article 2 of pereiopods 1 and 2 greatly expanded and filled with glandular A REDESCRIPTION OF AMPITHOE HUMERALIS 133 material, article 4 broadly lobed anteriorly; pereiopod 3 shorter than | and 2, article 2 subcircular, posterior margin of article 6 bearing seven short stout spines, dactyl short, strongly curved; pereiopod 4 longer than 3, article 2 broadly oval, none of distal segments greatly expanded, article 4 somewhat longer than 5 or 6, dactyl moderately curved; pereiopod 5 slightly longer than 4 but of similar structure, none of segments greatly enlarged or expanded. Pleon and urosome dorsally smooth; pleonal epimera rounded with slight lateral ridges; peduncle of uropod 1 with stout interramal spine about 25 per cent length of rami, laterodorsal margin of peduncle with four short spines, mediodorsal margin with seven small spines, outer ramus slightly the shorter, with two dorsal and two apical spines, inner ramus narrower, with one dorsal and four apical spines; uropod 2 extending as far as 1, peduncle with one lateral and one apical spine on each dorsal margin, outer ramus with four short dorsal and three apical spines, inner ramus slightly the longer, with four dorsal and four apical spines; uropod 3 extending slightly beyond 1 and 2, peduncle stout with a few dorsal and apical setae, outer ramus somewhat less than half length of peduncle, dorsally setulose, bearing two large strongly recurved apical hooks, inner ramus slightly shorter, broadly oval, distally truncated and setose with one small spine; telson semicircular, two small plumose setae on each margin and a small spine and plumose seta at each distal corner. Variation Females are very similar to males except for presence of brood-plates. The largest individual recorded, a female of 19,5 mm, had flagellum of antenna 1 40-articulate, of antenna 2 20-articulate, while the smallest juvenile, of 6 mm, had flagellum of antenna | 18-articulate, of antenna 2 8-articulate. Material 136 individuals from Oudekraal (33°58’S 17°21’E). Collected from the fronds of Ecklonia maxima at 5 m depth, 29 May 1978. Representative material has been deposited in the collections of the South African Museum (SAM-— A13660) and the University of Cape Town (CP 838A). Distribution Pacific North America (Puget Sound to Guadalupe Island), west coast of South Africa. Intertidal to approximately 80 m, usually associated with kelp species, rarely with other algae. NOTES ON ECOLOGY The nests of South African Ampithoe humeralis are formed by folding a secondary blade of the frond of the large kelp Ecklonia maxima longitudinally 134 ANNALS OF THE SOUTH AFRICAN MUSEUM a ill PHO Fig. 1. Ampithoe humeralis Stimpson, 1864. Male, 15 mm. Lateral aspect and sketch of nest formed from frond of Ecklonia maxima. and sealing together the adjoining surfaces some 10-20 mm above the fold. A tubular chamber is thus formed along the middle of the blade, the marginal portions of which extend freely. The two intact dwellings examined were constructed from fronds of 70 and 90 mm total width and the enclosed areas measured 55 x 10 and 80 x 20 mm respectively. The larger chamber contained a tightly-packed colony of 121 individuals of all sizes (the smaller chamber was preserved together with loose individuals and its population could not be accurately assessed). The interior of both chambers was strongly eroded, especially around the distal margins where the walls were paper-thin. The long, tattered streamers of frond adjacent to both chambers (Fig. 1) clearly indicate that the nests are initially formed at the tip of the blades and progressively extended backwards as the occupants feed upon their walls. The method of chamber formation is unknown but the rigidity of the blade and turbulence encountered at the collection site would indicate that a considerable communal effort must be required. The sealant used to form the chamber is invisible along most of its length, where the walls are closely appressed, but emerges as a triangular area of transparent mucous-like material, containing distinct trans- verse fibres, where it stretches between the diverging walls of the blade proximally. A REDESCRIPTION OF AMPITHOE HUMERALIS 135 Fig. 2. Ampithoe humeralis Stimpson, 1864. Male, 15 mm. A. Mandible. B. Lower lip. C. Maxilla 1, with tips of spines enlarged. D. Maxilla 2. E. Maxilliped. F—G. Gnathopods 1, 2. H-I. Articles 6 and 7 of pereiopods 3, 5. 136 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 3. Ampithoe humeralis Stimpson, 1864. Male, 15 mm. A. Urosome, lateral aspect. B. Uropod 3, medial view. C. Telson. Macropisthopous stebbingi K. H. Barnard, 1916. Male, 5,5 mm. D-E. Articles 5-7 of gnathopods 1, 2. F. Pereiopod 5. G. Uropod 3, medial view. A REDESCRIPTION OF AMPITHOE HUMERALIS 137 RELATIONSHIPS The material described here agrees in almost every detail with that illus- trated by J. L. Barnard (1965) and there can be little doubt as to its identification. Ampithoe humeralis is unique in Ampithoe for having the male gnathopod 2 almost as small as | (J. L. Barnard 1965) and there is thus a temptation when running it through a generic key (J. L. Barnard 19695) to allocate it to Macro- pisthopous. This genus, a monospecific one erected by K. H. Barnard (1916) for M. stebbingi from South Africa, is distinguished from Ampithoe primarily by the feebly chelate gnathopods (the enlarged pereiopod 5 cannot be regarded as generically significant—e.g. U. platypoda in Urothoe, various Orchestia species). A re-examination of M. stebbingi (Fig. 3) shows that the gnathopods of this species are very similar to those of A. humeralis (Fig. 2), the original sketches of K. H. Barnard (1916) being somewhat misleading in this regard. On the basis of the above evidence it appears that a generic distinction between M. stebbingi and A. humeralis cannot be justified. Moreover, although the weakly chelate gnathopods of these two species do distinguish them from those of other Ampithoe species, this condition appears to be merely the extreme of a wide range of gnathopod types occurring in the genus. For this reason a transfer of A. humeralis to Macropisthopous would not be justified and hence Macropisthopous must be incorporated into Ampithoe. A. stebbingi (K. H. Barnard, 1916) would then be distinguished from other Ampithoe species by the enlarged oar-like pereiopod 5 (Fig. 3F) and from all except A. humeralis by the condition of the gnathopods. ACKNOWLEDGEMENTS My thanks to Mr G. S. Dieckmann, who collected the material of Ampithoe humeralis and brought it to my attention. Mrs S. Hardman kindly typed the manuscript. REFERENCES BARNARD, J. L. 1954. Marine Amphipoda of Oregon. Ore. St. Monogr. Stud. Zool. 8: 1-103. BARNARD, J. L. 1965. Marine Amphipoda of the family Ampithoidae from southern California. Proc. U.S. natn. Mus. 118: 1-46. BARNARD, J. L. 1969a. Gammaridean Amphipoda of the rocky intertidal of California: Monterey Bay to La Jolla. Bull. U.S. natn. Mus. 258: 1-230. BARNARD, J. L. 1969b. The families and genera of marine gammaridean Amphipoda. Bull. U.S. natn. Mus. 271: 1-535. BARNARD, K. H. 1916. Contributions to the crustacean fauna of South Africa. 5. The Amphi- poda. Ann. S. Afr. Mus. 15: 105-302. CALMAN, W. T. 1898. On a collection of Crustacea from Puget Sound. Ann. New York Acad. Sci. 11: 259-292. JONES, L. G. 1971. Studies on selected small herbivorous invertebrates inhabiting Macrocystis canopies and holdfasts in southern California kelp beds. In: NorTH, W. J. ed. The biology of giant kelp beds (Macrocystis) in California. Nova Hedwigia (Suppl.) 32: 1-600. Just, J. 1977. Amphyllodomus incurvaria gen. et sp. n. (Crustacea, Amphipoda), a remarkable leaf-cutting amphithoid from the marine shallows of Barbados. Zool. Scr. 6: 229-232. 138 ANNALS OF THE SOUTH AFRICAN MUSEUM NortTH, W. J. ed. 1971. The biology of giant kelp beds (Macrocystis) in California. Nova Hedwigia (Suppl.) 32: 1-600. OrtTIz, M. 1976. Un nuevo genero y una nueva especie de anfipodi de aguas Cubanas (Amphipoda, Gammaridea, Ampithoidae). Ciencies La Habana (8) 27: 3-12. STEBBING, T. R. R. 1906. Amphipoda 1. Gammaridea. Das Tierreich 21: 1-806. STIMPSON, W. 1864. Descriptions of new species of marine invertebra from Puget Sound, collected by the naturalists of the North-west Boundary Commission, A. H. Campbell, Esq., Commissioner. Proc. Acad. nat. Sci. Philad. 16: 153-165. 6. SYSTEMATIC papers must conform to the International code of zoological nomenclature (particularly Articles 22 and 51). Names of new taxa, combinations, synonyms, etc., when used for the first time, must be followed by the appropriate Latin (not English) abbreviation, e.g. gen. nov., sp. nov., comb. nov., syn. nov., etc. ‘An author’s name when cited must follow the name of the taxon without intervening punctuation and not be abbreviated; if the year is added, a comma must separate author’s name and year. The author’s name ‘(and date, if cited) must be placed in parentheses if a species or subspecies is transferred from its original genus. The name of a subsequent user of a scientific name must be separated from the scientific name by a colon. Synonymy arrangement should be according to chronology of names, i.e. all published scientific names by which the species previously has been designated are listed in chronological order, with all references to that name following in chronological order, e.g.: Family Nuculanidae Nuculana (Lembulus) bicuspidata (Gould, 1845) Figs 14-15A Nucula (Leda) bicuspidata Gould, 1845: 37. Leda plicifera A. Adams, 1856: 50. Laeda bicuspidata Hanley, 1859: 118, pl. 228 (fig. 73). Sowerby, 1871: pl. 2 (fig. 8a—b). Nucula largillierti Philippi, 1861: 87. Leda bicuspidata: Nicklés, 1950: 163, fig. 301; 1955: 110. Barnard, 1964: 234, figs 8-9. Note punctuation in the above example: comma separates author’s name and year semicolon separates more than one reference by the same author full stop separates references by different authors figures of plates are enclosed in parentheses to distinguish them from text-figures dash, not comma, separates consecutive numbers Synonymy arrangement according to chronology of bibliographic references, whereby the year is placed in front of each entry, and the synonym repeated in full for each entry, is ~ not acceptable. In describing new species, one specimen must be designated as the holotype; other speci- mens mentioned in the original description are to be designated paratypes; additional material not regarded as paratypes should be listed separately. The complete data (registration number, depository, description of specimen, locality, collector, date) of the holotype and paratypes must be recorded, e.g.: Holotype SAM-—A13535 in the South African Museum, Cape Town. Adult female from mid-tide region, King’s Beach Port Elizabeth (33°51’S 25°39’E), collected by A. Smith, 15 January 1973. Note standard form of writing South African Museum registration numbers and date. T. SPECIAL HOUSE RULES Capital initial letters (a) The Figures, Maps and Tables of the paper when referred to in the text e.g. *... the Figure depicting C. namacolus ...’; *. .. in C. namacolus (Fig. 10). . (b) The prefixes of prefixed surnames in all languages, when used in the text, if not preceded by initials or full names e.g. Du Toit but A.L.du Toit; Von Huene but F. von Huene (c) Scientific names, but not their vernacular derivatives e.g. Therocephalia, but therocephalian Punctuation should be loose, omitting all not strictly necessary Reference to the author should be expressed in the third person Roman numerals should be converted to arabic, except when forming part of the title of a book or article, such as ‘Revision of the Crustacea. Part VIII. The Amphipoda.’ Specific name must not stand alone, but be preceded by the generic name or its abbreviation to initial capital letter, provided the same generic name is used consecutively. Name of new genus or species is not to be included in the title: it should be included in the abstract, counter to Recommendation 23 of the Code, to meet the requirements of Biological Abstracts. CHARLES L. GRIFFITHS A REDESCRIPTION OF THE KELP CURLER AMPITHOE HUMERALIS (CRUSTACEA, AMPHIPODA) FROM SOUTH AFRICA AND ITS RELATIONSHIP TO MACROPISTHOPOUS { - VOLUME 79 PART 6 DECEMBER 1979 OF THE SOUTH AFRICAN J ; MUSEUM |f CAPE TOWN > INSTRUCTIONS TO AUTHORS 1. MATERIAL should be original and not published elsewhere, in whole or in part. 2. LAYOUT should be as follows: (a) Centred masthead to consist of Title: informative but concise, without abbreviations and not including the names of new genera or species Author’s(s’) name(s) Address(es) of author(s) (institution where work was carried out) Number of illustrations (figures, enumerated maps and tables, in this order) (b) Abstract of not more than 200 words, intelligible to the reader without reference to the text (c) Table of contents giving hierarchy of headings and subheadings (d) Introduction (e) Subject-matter of the paper, divided into sections to correspond with those given in table of contents (f) Summary, if paper is lengthy (g) Acknowledgements (h) References (i) Abbreviations, where these are numerous 3. MANUSCRIPT, to be submitted in triplicate, should be typewritten and neat, double spaced with 2,5 cm margins all round. First lines of paragraphs should be indented. Tables and a list of legends for illustrations should be typed separately, their positions indicated in the text. All pages should be numbered consecutively. Major headings of the paper are centred capitals; first subheadings are shouldered small capitals; second subheadings are shouldered italics; third subheadings are indented, shouldered italics. Further subdivisions should be avoided, as also enumeration (never roman numerals) of headings and abbreviations. Footnotes should be avoided unless they are short and essential. Only generic and specific names should be underlined to indicate italics; all other marking up should be left to editor and publisher. 4. ILLUSTRATIONS should be reducible to a size not exceeding 12 « 18 cm (19 cm including legend); the reduction or enlargement required should be indicated; originals larger than 35 x 47 cm should not be submitted; photographs should be rectangular in shape and final size. A metric scale should appear with all illustrations, otherwise magnification or reduction should be given in the legend; if the latter, then the final reduction or enlargement should be taken into consideration. All illustrations, whether line drawings or photographs, should be termed figures (plates are not printed; half-tones will appear in their proper place in the text) and numbered in a single series. Items of composite figures should be designated by capital letters; lettering of figures is not set in type and should be in lower-case letters. The number of the figure should be lightly marked in pencil on the back of each illustration. 5. REFERENCES cited in text and synonymies should all be included in the list at the end of the paper, using the Harvard System (ibid., idem, loc. cit., op. cit. are not acceptable): (a) Author’s name and year of publication given in text, e.g.: ‘Smith (1969) describes...’ ‘Smith (1969: 36, fig. 16) describes . “As described (Smith 1969a, 19695; ee a ‘As described (Haughton & Broom eee ‘As described (Haughton et al. 1927) . Note: no comma separating name etd year Pagination indicated by colon, not p. names of joint authors connected by ampersand et al. in text for more than two joint authors, but names of all authors given in list of references. (b) Full references at the end of the paper, arranged alphabetically by names, chronologically within each name, with suffixes a, b, etc. to the year for more than one paper by the same author in that year, e.g. Smith (1969a, 19695) and not Smith (1969, 1969a). For books give title in italics, edition, volume number, place of publication, publisher. For journal article give title of article, title of journal in italics (abbreviated according to the World list o, Scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses, volume number, part number (only if independently paged) in parentheses, pagination (first and last pages of article). Examples (note capitalization and punctuation) BULLOUGH, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan. FISCHER, P. —H. 1948. Données sur la résistance et de le vitalité des mollusques. J. Conch., Paris 88: 100-140. FiscHER, P.-H., DuvaL, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines. Archs Zool. exp. gen. 74: 627-634. Konn, A. J. 1960a. Ecological notes on Gans (Mollusca: Gastropoda) in the Trincomalee region of Ceylon. Ann. Mag. nat. Hist. (13) 2: 309-320. Konn, A. J. 19605. Spawning behaviour, egg masses and larval development in Conus from the Indian Ocean. Bull. Bingham oceanogr. Coll. 17 (4): 1-51. THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. In: SCHULTZE, L. Zoologische und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-Afrika 4: 269-270. Jena: Fischer. Denkschr. med.-naturw. Ges. Jena 16: 269-270. (continued inside back cover) ANNALS OF THE SOUTH AFRICAN MUSEUM ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM Volume 79 Band December 1979 Desember Part 6 Deel CALCAREOUS NANNOFOSSILS AND PLANKTIC FORAMINIFERS IN TERTIARY LIMESTONES, NATAL AND EASTERN CAPE, SOUTH AFRICA By WILLIAM G. SIESSER & GREGORY A. MILES Cape Town Kaapstad The ANNALS OF THE SOUTH AFRICAN MUSEUM are issued in parts at irregular intervals as material becomes available Obtainable from the South African Museum, P.O. Box 61, Cape Town Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM word uitgegee in dele op ongereelde tye na gelang van die beskikbaarheid van stof Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad OUT OF PRINT/UIT DRUK 1, 2(1-3, 5-8), 3(1-2, 4-5, 8, t.-p.i.), 5(1-3, 5, 7-9), 6(1, t.-p.i.), 711-4), 8, 9(1-2, 7), 10(1-3), 11(1-2, 5, 7, t.-p.i.), 15(4—-5), 24(2), 27, 31(1-3), 32(5), 33 Copyright enquiries to the South African Museum Kopieregnavrae aan die Suid-Afrikaanse Museum ISBN 0 908407 83 1 Printed in South Africa by In Suid-Afrika gedruk deur The Rustica Press, Pty., Ltd., Die Rustica-pers, Edms., Bpk., Court Road, Wynberg, Cape Courtweg, Wynberg, Kaap CALCAREOUS NANNOFOSSILS AND PLANKTIC FORAMINIFERS IN TERTIARY LIMESTONES, NATAL AND EASTERN CAPE, SOUTH AFRICA By WILLIAM G. SIESSER* South African Museum, Cape Town & GREGORY A. MILES Exxon Company, U.S.A. Houston (With 12 figures) [MS. accepted 28 August 1979] ABSTRACT Sixteen taxa of Palaeogene calcareous nannofossils and seven of planktic foraminifers have been identified in limestones from the eastern Cape Province. Four taxa of Neogene calcareous nannofossils and nine of planktic foraminifers have been found in limestones from Zululand. The assemblages indicate that the limestones at Birbury are lower Eocene, and suggest a probable Eocene age for the limestones in the upper quarry at Needs Camp. Nannofossils and planktic foraminifers confirm that the limestones north of the Umfolozi River in Zululand are upper Miocene—lower Pliocene. The Birbury and Zululand dates reflect the mid Paleocene—-early Eocene and late Miocene— early Pliocene transgressions round South Africa. The Tertiary limestones at Needs Camp could have been deposited during either the early or the late Eocene transgression. CONTENTS PAGE Introduction : ‘ P ; : , 2140 Birbury : : ; : : : a 14 Calcareous nannofossils . : . 144 Planktic foraminifers : : . 144 Age . ‘ E : ‘ : ~~ 144 Needs Camp : ; 3 : : ss AS Calcareous nannofossils . P ie SAS Age. : : : ‘ : 4. 450 Zululand . : ; : ‘ : S50 Calcareous nannofossils . : ; 152 Planktic foraminifers ’ : eee Wel ie) APS. f : ; : ‘ sie JE Tertiary sea-level movements . . . 154 Acknowledgements . : t : 8 References . 3 5 ‘ : ; ; ST. * Present address: Department of Geology, Vanderbilt University, Nashville, Tennessee, 31235. Ann. S. Afr. Mus. 79 (6), 1979: 139-158, 12 figs. 139 140 ANNALS OF THE SOUTH AFRICAN MUSEUM INTRODUCTION Tertiary limestones are exposed intermittently along the coast of the Cape Province from at least Saldanha Bay in the west to East London in the east. A second belt of limestones crops out intermittently from near the Umfolozi River in Natal Province (‘Zululand’) northward into Mozambique (Fig. 1). ; These limestones are not well dated, despite the efforts of numerous investigators. Age assignments of certain outcrops in the eastern Cape and / aes Maputo / /SWAZI- 1 Q otakeyiew 72S eMTOTI PAN MKUZEe SY MORRISVALE® ee u LOT 178 ©—_- 9° 7 UMKWELANE HILL 5@ HAS ZDRAIN THE JUNGLE _ Richards NATAL Bay CAPE NEEDS CAMPe ast London BIRBURYe Port Elizabeth 26 Fig. 1. Location map showing the outcrop localities of the marine Tertiary limestones mentioned in the text. CALCAREOUS NANNOFOSSILS AND PLANKTIC FORAMINIFERS 141 Natal are especially contentious. Most earlier workers relied on molluscs, echinoderms, benthic foraminifers or sharks’ teeth as the bases for their age assignments. The first three groups of animals are strongly facies controlled and are not particularly good for narrow age determinations. Nor are sharks’ teeth very suitable, since their stratigraphic usefulness suffers from a lack of precisely known age ranges for different species. Planktic foraminifers and calcareous nannofossils are the two groups most widely used for high-resolution Tertiary biostratigraphy today. They have several advantageous characteristics: (i) they are very abundant in marine sediments, (ii) their evolution has been rapid, which allows establishment of many narrow and discrete zones on the basis of morphotypes, and (111) as floating organisms they are widely distributed and thus are useful for interregional correlations. Planktic foraminifers from some of the outcrops described in this study have been investigated in recent years, and several of the former age assignments have already been modified (especially in Zululand). Calcareous nannofossils have not previously been reported from any of these rocks. The purposes of this paper are (i) to record the nannofossils and planktic foraminifers present in these Tertiary limestones, (ii) to assess the ages of the rocks, based on the nannofossils and foraminifers, and (iii) to comment on Tertiary sea-level movements round South Africa in the light of the ages now assigned to the limestones. BIRBURY Only cursory remarks on the lithology of the Birbury section (Fig. 1) have previously been published. Figure 2 shows a stratigraphic section measured by one of us (WGS) at Birbury. The section is exposed in an abandoned quarry near the Birbury homestead. It consists of 1 to 1,5 m of calcareous conglomerate and coarse calcarenite, overlain by about 3 m of fine to medium calcarenite which becomes increasingly nodular and ‘chalky’ up section. The limestones contain scattered sharks’ teeth and mollusc, echinoid and bryozoan fragments. The general lithology of this outcrop is not similar to the Tertiary Alexandria Formation extensively exposed to the west (see measured sections in Siesser 1972), nor to the Needs Camp exposures to the east. Chapman (1930) examined ten thin sections of the Birbury limestones which were sent to him by Haughton in 1925. Chapman noted the presence of *. . . beautifully preserved and abundant tests of Discocyclina pratti and D. varians . . ... Chapman (1930) assigned the rocks to the upper Eocene on the basis of the more abundant of the two species, D. pratti, which he stated *. . . is a well-recognized Bartonian species’. He also mentioned that the less numerous D. varians seemed to indicate a slightly lower horizon (upper Lutetian, i.e. middle Eocene). More recent investigations have extended the age range of D. pratti, and it is now known to occur from the Paleocene through the Eocene (D. Salmon, pers. comm. 1978). Nevertheless, an Eocene age seemed evident and was widely accepted until 142 ANNALS OF THE SOUTH AFRICAN MUSEUM 4m Mostly pinkish grey, poorly consolidated, chalky fine calcarenite. = Irregularly shaped, well indurated calcarenite nodules common to abundant. Sparse fossils. 3m = Well consolidated fine calcarenite. Abundant nodules as above. Poorly consolidated fine calcarenite. Sparse nodules as above. _ Pale greenish yellow, poorly consolidated fine calcarenite. Scattered coarse quartz grains. Macrofossil fragments abundant. — Pale greenish yellow, poorly consolidated fine calcarenite. Macrofossil fragments abundant: fish teeth, echinoids, 2m bivalves, bryozoans. - Pale greenish yellow, quartzose medium calcarenite. Pale greenish yellow, well consolidated calcareous conglomerate. — Granules and pebbles of quartz, quartzite and siltstone abundant. Sparse fossils. im — Yellowish grey, well consolidated glauconitic coarse calcarenite Granules of quartzite and siltstone common. Sparse fossils. Yellowish grey, poorly consolidated medium. calcarenite. Sparse granules. Poorly exposed. Fig. 2. Stratigraphic section of lower Eocene limestones at Birbury, eastern Cape Province, 33°28'18’S 26°55’30’E. Base of the section is about 205 m above sea-level. Colours are from the Geological Society of America Rock-Color Chart. Bourdon & Magnier (1969) reported their investigation of the foraminifers at Birbury. Largely on the basis of test preservation, they concluded that two different assemblages are present at Birbury: (1) an abundant, but poorly preserved Eocene fauna represented by at least twenty taxa, and (ii) a sparse, but well-preserved Miocene fauna represented by five species of benthic fora- minifers. They, therefore, assigned the rocks to the Miocene, suggesting that the CALCAREOUS NANNOFOSSILS AND PLANKTIC FORAMINIFERS 143 diverse and abundant Eocene foraminifers were reworked (but correctly noted that such a mixture was ‘quite remarkable’). Their Miocene age assignment seems to have been uncritically accepted by most interested workers in South Africa, even though Bourdon & Magnier’s five ‘Miocene’ species were not conclusively identified (all were listed as ‘cf’). Moreover, even if the identifications of these five species had been definite, it seems somewhat presumptuous to assign a Miocene age on the basis of such long-ranging forms as Cibicides lobatulus (Eocene to Holocene) and Lagena gibbera (? Eocene to Holocene). Finally, Bourdon & Magnier (1969: 123) point out that their colleagues at the Paris Museum examined sharks’ teeth from the same samples and assigned them to the lower Eocene. Fig. 3. Quarry face at Birbury. Pebble conglomerate at bottom of photograph. Hammer lies near top of fossiliferous calcarenite. Nodular, ‘chalky’ zone at top of photograph. 144 ANNALS OF THE SOUTH AFRICAN MUSEUM Calcareous nannofossils (Fig. 4) Very rare, poorly preserved nannofossils occur throughout the section at Birbury. However, only a poorly consolidated fossiliferous layer about 2 m above the base contains nannofossils in sufficient numbers and with adequate preservation to justify study. Species present include: Braarudosphaera bigelowi (Gran & Braarud) Chiasmolithus solitus (Bramlette & Sullivan) Chiasmolithus sp. Coccolithus eopelagicus (Bramlette & Riedel) Coccolithus formosus (Kamptner) Cruciplacolithus sp. Cycloccolithus gammation Bramlette & Sullivan Discoaster sp. ? Lophodolithus nascens Bramlette & Sullivan Pontosphaera spp. Reticulofenestra coenura (Reinhardt) Transversopontis pulcher (Deflandre) Zygodiscus plectopons Bramlette & Sullivan Zygrhablithus bijugatus (Deflandre) Planktic foraminifers (Figs 4-6) A sparse population of planktic foraminifers is found throughout the section, but the best assemblage occurs in a fossiliferous layer about 2,5 m above the base. The following species were identified in the series of samples collected at Birbury: Acarinina esnaensis (LeRoy) A. nitida (Martin) A. primitiva (Finlay) A. pseudotopilensis Subbotina A. soldadoensis soldadoensis (Bronnimann) Morozovella subbotinae (Morozova) Subbotina eocaena (Giimbel) Age The nannofossil assemblage is clearly Palaeogene. Cruciplacolithus sp., Cyclococcolithus gammation, Lophodolithus nascens and Zygodiscus plectopons all range from Paleocene to Eocene, and Coccolithus formosus, Reticulofenestra coenura and Zyerhablithus bijugatus range from Eocene to Oligocene. Trans- versopontis pulcher is probably restricted to the Eocene, and Chiasmolithus solitus occurs only in the lower and middle Eocene. Braarudosphaera bigelowi and Coccolithus eopelagicus range throughout the Tertiary. The planktic foraminifers also indicate a Palaeogene age. Acarinina esnaensis, A. nitida, A. pseudotopilensis and A. soldadoensis soldadoensis all range CALCAREOUS NANNOFOSSILS AND PLANKTIC FORAMINIFERS F ' G Fig. 4. Calcareous nannofossils and planktic foraminifers from Birbury (lower Eocene). A. Chiasmolithus sp., plane polarized light, scale = 6u. B. Coccolithus formosus (right) and Pontosphaera sp. (left), crossed nicols, scale = 8u. C. Transversopontis pulcher, crossed nicols, scale = 5u. D. Acarinina esnaensis, spiral view, SAM-K5542. E-G. A. nitida. E. Spiral view, SAM-K5543. F. Umbilical view, SAM—-K5544. G. Side view, SAM-K5544. Scales D-G = 200z. 145 146 ANNALS OF THE SOUTH AFRICAN MUSEUM E pee. F A SSL Fig. 5. Planktic foraminifers from Birbury (lower Eocene). A-—C. Acarinina primitiva. A. Spiral view, SAM-K5545. B. Umbilical view, SAM-K5546. C. Side view, SAM-K5547. D. A. pseudotopilensis, spiral view, SAM—K5548. E-F. A. soldadoensis soldadoensis. E. Spiral view, SAM-K5549. F. Umbilical view, SAM—K5550. Scales = 200p. CALCAREOUS NANNOFOSSILS AND PLANKTIC FORAMINIFERS 147 E F bee ea ae | Fig. 6. Planktic foraminifers from Birbury (lower Eocene). A-B. Morozovella subbotinae. A. Spiral view, SAM-—K5551. B. Umbilical view, SAM—K5551. C—E. Subbotina eocaena. C. Spiral view, SAM—K5552. D-E. Umbilical views, SAM—K5553, SAM-K5554. F. Acarinina soldadoensis soldadoensis, umbilical view, SAM—K5555. Scales = 200p. —————EEE_eai 148 ANNALS OF THE SOUTH AFRICAN MUSEUM from upper Paleocene to lower Eocene, A. primitiva ranges from upper Paleocene to middle Eocene, Subbotina eocaena from lower Eocene to lower Oligocene, and Morozovella subbotinae is restricted to the lower Eocene. It is difficult to believe that these diverse Palaeogene assemblages of calcareous nannofossils and planktic foraminifers are entirely the result of reworking, especially since not one strictly Neogene nannofossil or foraminifer was found in the section. The section is, therefore, assigned to the lower Eocene, based on the overlapping ranges of various species of planktic foraminifers and calcareous nannofossils. The presence of M. subbotinae corroborates the age and allows assignment to Berggren & Van Couvering’s (1974) planktic fora- miniferal Zones P6 to P8 (54,5 to 50 m.y.B.P.; years from Hardenbol & Berggren 1978). NEEDS CAMP Marine limestones crop out in two quarries near Needs Camp (Fig. 1). The lower quarry contains limestones, which, on the basis of calcareous nannofossils, are probably upper Campanian to lower Maestrichtian (Siesser, unpublished data). The upper quarry is about 17 m above the lower quarry, and the contact between the limestone units in the two quarries is not exposed. The exposed section in the upper quarry is about 4 m thick; it is a hard, recrystallized, coarse calcarenite which becomes more flaggy towards the top (Fig. 7). The calcarenite is mostly yellowish-grey 5Y 7/3 (GSA Rock Color Chart) on fresh surfaces but weathers to a darker grey. A thin section of the calcarenite shows it is a skeletal grainstone, rich in the remains of molluscs, cirripeds, bryozoans and coralline algae, cemented by microspar (Siesser 1971). Large Perna valves are conspicuous in these rocks. Additional notes on the lithology of the upper quarry are found in Lock (1973). _ The limestones in the upper quarry were originally correlated with the Alexandria Formation (Bullen-Newton 1913; Du Toit 1954). An Eocene age was proposed for part of the Alexandria Formation, based on the identification of upper Eocene foraminifers (Chapman 1930) and sharks’ teeth (Haughton 1925) at Birbury. Haughton (1969), therefore, suggested an Eocene age for the upper quarry in his textbook on South African geology. King (1972) considered both quarries to be Upper Cretaceous, based on fossil and geomorphic evidence. Lock (1973) reviewed the literature in detail with regard to the dating of the upper quarry. He concluded, again, that these rocks are Tertiary, but did not assign them to a specific series. Lock (pers. comm. 1977) believes the upper quarry could be as yo..ng as Miocene, based on the presence of Carcharodon angustidens. (This shark is, however, now known to range from middle Eocene to Pliocene (Jubb & Gardiner 1975).) Calcareous nannofossils (Fig. 8) Most of the rocks in the upper quarry are too recrystallized to yield identifiable nannofossils; only a few heavily overgrown species (Coccolithus CALCAREOUS NANNOFOSSILS AND PLANKTIC FORAMINIFERS 149 Fig. 7. Calcarenite in the upper quarry at Needs Camp, eastern Cape Province. Probably Eocene. eopelagicus, Cyclococcolithus gammation and Toweius sp.) were found in a soft layer about 1,75 above the base of the exposure. A slightly better assemblage was found in a calcareous siltstone from a pit dug below the exposed base of the outcrop. This sample was collected and made available by B. E. Lock. The sample contains a mixture of poorly to moderately preserved Upper Cretaceous and Tertiary species. Extensive reworking of the nannofossil-rich underlying beds is obvious, since Upper Cretaceous specimens outnumber Tertiary specimens by a ratio of about 10 to 1 in the siltstone. The following Tertiary species are present: Braarudosphaera bigelowi (Gran & Braarud) Chiasmolithus cf. C. solitus (Bramlette & Sullivan) Coccolithus eopelagicus (Bramlette & Riedel) 150 ANNALS OF THE SOUTH AFRICAN MUSEUM A ——s B | ene | Fig. 8. Calcareous nannofossils from Needs Camp upper quarry (probably Eocene). A. Cycloccolithus gammation, crossed nicols, scale = 3p. B. Chiasmolithus solitus, plane polarized light, scale = Su. C. Micula staurophora (Gardet) Stradner (reworked from Upper Cretaceous), crossed nicols, scale = 3p. C. cf. C. formosus (Kamptner) Cyclococcolithus gammation Bramlette & Sullivan ? Cyclicargolithus floridanus (Roth & Hay) Reticulofenestra coenura (Reinhardt) ? Zygrhablithus bijugatus (Deflandre) No identifiable planktic foraminifers were found in the samples. Age An unequivocal age cannot be assigned because of the obvious reworking and the questionable identification of several species. Nevertheless, the overall assemblage is clearly Palaeogene (Eocene—Oligocene) (see Birbury p. 144 for the age ranges of these species). The age is probably Eocene. Cyclococcolithus gammation is considered by most (but not all) nannofossil workers to be limited to the Eocene; Toweius sp. is not considered to range above the lower Eocene (Bukry 1973; Gartner 1977); and Chiasmolithus solitus is restricted to the lower and middle Eocene (Gartner 1977). ZULULAND Scattered outcrops of Tertiary limestones occur in north-eastern Natal (‘Zululand’), from the Umfolozi River northward to the Mozambique border and beyond. The best-known exposures are at Uloa, Sapolwana, Warners Drain, Lot U-178, Umkwelane Hill and The Jungle (Fig. 1). Detailed studies of these outcrops have been made by King (1953 et seq.) and Frankel (1960 et seq.). Lesser-known exposures occur at or near Mtoti Pan, Lakeview, Morrisvale and Mkuze. The last two are previously unreported Tertiary localities. Their co-ordinates are: Morrisvale—27°40'36”S 32°22'03”E; Mkuze—27°33’50"S 32°15'25”E. B. du Cann (pers. comm. 1976) believes the Zululand limestones may reach their thickest development (about 15 m) at Lakeview; Truswell (1977), however, has reported up to 60 m of the limestones at an unspecified CALCAREOUS NANNOFOSSILS AND PLANKTIC FORAMINIFERS 151 locality in Zululand. In general, these rocks consist of a basal nodular limestone which rests on a planed Cretaceous surface. The nodular limestone contains phosphatized pebbles of Cretaceous siltstone and is intensely iron-stained at some localities. Reworked Eocene foraminifers and Cretaceous ammonites have also been found in this unit (Frankel 1968; King 1970). The basal unit is not everywhere present, but does occur at Uloa, where it is overlain by an unbedded calcirudite (“Pecten bed’) (Fig. 9). The Pecten bed is a richly fossiliferous unit which has yielded over 100 species of macrofossils and 45 species of microfossils (King 1970). The fossil assemblage is dominated by the abundant Aeqguipecten uloa King. The Pecten bed is the oldest unit exposed at some localities, but is always overlain by bedded, upward-fining calcarenites (Frankel (1966) gives a detailed description of the lithology of the Uloa section). Fig. 9. Pecten calcirudite at Uloa (upper Miocene to lower Pliocene). Hammer lies above nannofossil-bearing zone. 152 ANNALS OF THE SOUTH AFRICAN MUSEUM The stratigraphic position of these outcrops has been the subject of con- siderable discussion (see Stapleton (1977) for a review). Briefly, King (1953 et seq.) believes the basal and Pecten beds are lower Miocene, based on macro- and microfossil evidence. He states that an unconformity separates the lower Miocene unit from the overlying calcarenites, which he dates as Pliocene on the basis of mollusc fragments. Frankel (1968) prefers a middle or upper Miocene to Pliocene position for these limestones, based on sharks’ teeth and foraminifers. He recognizes intra-formational channelling of the Pecten bed, but not a major hiatus. Much of the controversy concerning the age of these outcrops resulted from the use of fossils which are not closely age-diagnostic. Two recent papers described assemblages of planktic foraminifers which may have settled the question. Maud & Orr (1975) reported a stratigraphic sequence in boreholes at Richards Bay that is the lithologic equivalent of the Uloa—Sapolwana sequence. The Richards Bay strata consist of a mollusc-fragment coquina (containing ferruginous Cretaceous siltstone pebbles at its base) overlain by quartzose calcarenite. The lowermost portion of the calcarenite consists of well-bedded coarse sand, whereas higher up the calcarenite is cross-bedded. This sequence unconformably overlies Cretaceous and Paleocene rocks and has a maximum thickness of 5 to 6 m in the Richards Bay area. A good assemblage of upper Miocene benthic and planktic foraminifers is found in the calcarenite (Maud & Orr 1975). The foraminiferal assemblage in the coquina is apparently not as age-diagnostic, although Maud & Orr (1975) feel the similarities between the calcarenite and coquina assemblages are sufficient to place the ato in the upper Miocene also. Stapleton (1977) found an assemblage of thirteen species of planktic foraminifers in a sample collected at the base of the Uloa Pecten bed. He concluded that its age is latest Miocene, or at the Miocene—Pliocene boundary. He reassessed published information on the planktic foraminifers in the over- lying calcarenites and concluded that they could be little younger than the Pecten bed, and are probably of nearly the same age. Calcareous nannofossils (Fig. 10) Samples examined from Sapolwana, Warners Drain, Umkwelane Hill, The Jungle, Mkuze and Mtoti Pan were barren of nannofossils. Samples from Lakeview, Morrisvale and Lot U-178 contain rare to sparse, poorly preserved specimens of Reticulofenestra pseudoumbilica (Gartner). A soft pocket in the Pecten bed just south of the railway line at Uloa yielded a sparse, but moderately well-preserved assemblage including Coccolithus pelagicus (Wallich), Discoaster surculus Martini & Bramlette, Discoaster sp. and Reticulofenestra pseudoumbilica. Planktic foraminifers (Figs 10-12). The only samples containing identifiable planktic foraminifers are those from Lakeview, Lot U-178 and Uloa. Globigerinella aequilateralis (Brady) and CALCAREOUS NANNOFOSSILS AND PLANKTIC FORAMINIFERS 153 F eae eee | G [ake vel Fig. 10. A—C. Calcareous nannofossils from Morrisvale (A) and Uloa (Pecten bed) (B & C), upper Miocene-lower Pliocene. A. Reticulofenestra pseudoumbilica, crossed nicols, scale = 3u. B. Discoaster surculus, plane polarized light, scale = 8p. C. Reticulofenestra pseudoumbilica, crossed nicols, scale = 3p. D-G. Planktic foraminifers from Uloa (Pecten bed). Scales = 200u. D-E. Globigerina cf. G. praedigitata. D. Spiral view, SAM-K5556. E. Side view, SAM-—K5556. F-G. Globigerinella aequilateralis. F. Spiral view, SAM-K5557. G. Side view, SAM-KS5558. 154 ANNALS OF THE SOUTH AFRICAN MUSEUM Globigerinoides obliquus extremus Bolli & Bermudez were found at Lakeview and G. trilobus (Reuss) at Lot U-178. The Pecten bed at Uloa yielded a better assemblage, including: | Globigerina cf. G. praedigitata Parker (single specimen) Globigerinella aequilateralis (Brady) Globigerinoides conglobatus (Brady) G. obliquus extremus Bolli & Bermudez G. trilobus (Reuss) Globoquadrina altispira altispira (Cushman & Jarvis) G. altispira globosa Bolli Neogloboquadrina humerosa (Takayanagi & Saito) Orbulina universa d’Orbigny This assemblage is indicative of warm-temperate water. Age The calcareous nannofossils and planktic foraminifers in these rocks strongly support the stratigraphic position assigned to them by Maud & Orr (1975) and Stapleton (1977), i.e. upper Miocene to lower Pliocene. The calcareous nannofossil Reticulofenestra pseudoumbilica ranges from middle Miocene to lower Pliocene; Discoaster surculus ranges from upper Miocene to upper Pliocene. Overlapping age ranges of these two species indicate that the Pecten bed at Uloa is upper Miocene to lower Pliocene. Furthermore, the Pecten bed can be placed in Martini’s (1971) calcareous nannofossil Zones NN 11-NN 15 (9,5 to 3,0 m.y.B.P.; years from Vail et al. (1977)). The planktic foraminiferal assemblage also indicates an upper Miocene to Pliocene assignment. A Pliocene age may be suggested by the presence of Globigerinella aequilateralis and Neogloboquadrina humerosa, since these species are substantially more abundant in the Pliocene than in the Miocene, and thus are more likely to be encountered in sparse Pliocene, rather than in sparse Miocene, samples. However, this is very tenuous ‘evidence’, and only the conservative range of upper Miocene-lower Pliocene is assigned at this time. TERTIARY SEA-LEVEL MOVEMENTS The firmly dated limestones at Birbury (lower Eocene) and in Zululand (upper Miocene-lower Pliocene) and the tentatively dated limestones at Needs Camp (Eocene) provide evidence as to the timing of sea-level movements round the south and east coasts of South Africa. Siesser & Dingle (1979) presented a generalized scenario of Tertiary sea- level movements round southern Africa, based on evidence from continental shelf and on-shore deposits. They suggested that a transgression began in mid Paleocene times and continued into the early Eocene. A middle Eocene regression was followed by another transgression in the late Eocene (see also Siesser 1977). CALCAREOUS NANNOFOSSILS AND PLANKTIC FORAMINIFERS 155 E | oaceaa tant F ‘tesla dae Fig. 11. Planktic foraminifers from Uloa (Pecten bed) (A, C-F) and Lakeview (B), upper Miocene-lower Pliocene. A. Globigerinoides conglobatus, spiral view, SAM-—K5559. B. G. obliquus extremus, umbilical view, SAM—K5560. C-D. G. trilobus. C. Spiral view, SAM-K5561. D. Umbilical view, SAM-K5562. E-F. Globoquadrina altispira globosa. E. Spiral view, SAM-K5563. F. Umbilical view, SAM-K5563. Scales = 200p. 156 ANNALS OF THE SOUTH AFRICAN MUSEUM The early Eocene transgression is recorded on-shore by the outcrops at Birbury. The Needs Camp limestones can be dated only as probably Eocene; thus it cannot yet be suggested whether they were deposited during the early or the late Eocene transgression. Other evidence presented by Siesser & Dingle (1979) indicates that, after a long Oligocene-early Miocene regression, the seas began to move shoreward again in middle Miocene time. This transgression probably reached its greatest extent in the late Miocene-early Pliocene, as represented by the sequence of marine limestones in Zululand. C Ee reeeeeeer cs) Fig. 12. Planktic foraminifers from Uloa (Pecten bed), upper Miocene-lower Pliocene. A-B. Neogloboquadrina humerosa. A. Spiral view, SAM-K5564. B. Umbilical view, SAM-— K5565. C. Orbulina universa, SAM-K5566. Scales = 200p. CALCAREOUS NANNOFOSSILS AND PLANKTIC FORAMINIFERS 57 ACKNOWLEDGEMENTS This work began while the senior author was a member of the Marine Geoscience Group, University of Cape Town, and was completed after he joined the South African Museum. Both organizations are thanked for their support. The senior author also wishes to thank the Council for Scientific and Industrial Research and the University of Cape Town for research grants which made the work possible. Dr H. C. Klinger directed one of us (WGS) to previously unreported Tertiary outcrops in Zululand (Morrisvale and Mkuze). His assistance is gratefully acknowledged. Scanning electron microscopy of the foraminifers was funded by Exxon Company, U.S.A. REFERENCES BERGGREN, W. A. & VAN COUVERING, J. A. 1974. Biostratigraphy, geochronology and paleoclimatology of the last 15 million years in marine and continental sequences. Paleogeogr., Paleoclimat., Paleoecol. 16: 1-216. Bourbon, M. & MAGNIER, PH. 1969. Notes on the Tertiary fossils at Birbury, Cape Province. Trans. geol. Soc. S. Afr. 72: 123-125. Bukry, D. 1973. Coccolith stratigraphy, Eastern Equatorial Pacific, Leg 16 Deep Sea Drilling Project. In: VAN ANDEL, T. H., HEATH, G. R., ef al., Initial Reports of the Deep Sea Drilling Project 16: 653-711. Washington: United States Government Printing Office. BULLEN-NEWTON, R. 1913. On some Kainozoic shells from South Africa. Rec. Albany Mus. 2EBiS=352. CHAPMAN, F. 1930. On a foraminiferal limestone of Upper Eocene age from the Alexandria Formation, South Africa. Ann. S. Afr. Mus. 28: 291-296. Du Torr, A. L. 1954. The geology of South Africa. 3rd ed. Edinburgh: Oliver & Boyd. FRANKEL, J. J. 1960. The geology along the Umfolozi River, south of Mtubatuba, Zululand. Trans. geol. Soc. S. Afr. 63: 231-252. FRANKEL, J. J. 1966. The basal rocks of the Tertiary at Uloa, Zululand, South Africa. Geol. Mag. 103: 214-230. FRANKEL, J. J. 1968. Tertiary sediments in the lower Umfolosi River Valley, Zululand. Trans. geol. Soc. S. Afr. 71: 1135-1146. _ GARTNER, S. 1977. Nannofossils and biostratigraphy : an overview. Earth Sci. Rev. 13: 227-250. HARDENBOL, J. & BERGGREN, W. A. 1978. A new Paleogene numerical time scale. Jn: CoHEE, G. V., GLAESSNER, M. F. & HEDBERG, H. D. eds. Contributions to the Geological time scale. Studies in geology 6: 213-234. Tulsa: American Association of Petroleum Geologists. HaucGuton, S. H. 1925. The Tertiary deposits of the south-eastern districts of the Cape Province. Trans. geol. Soc. S. Afr. 28: 27-32. HAuGurton, S. H. 1969. Geological history of southern Africa. Cape Town. Geological Society of South Africa. JusBB, R. A. & GARDINER, B. G. 1975. A preliminary catalogue of identifiable fossil fish material from southern Africa. Ann. S. Afr. Mus. 67: 381-440. Kina, L. C. 1953. A Miocene fauna from Zululand. Trans. geol. Soc. S. Afr. 56: 59-91. Kina, L. C. 1970. Uloa revisited: a review of the essential data from this classic locality and a statement of Tertiary history of the Zululand coastal area. Trans. geol. Soc. S. Afr. 73: 151-157. KING, L. C. 1972. Geomorphic significance of the Late Cretaceous limestones at Needs Camp, near East London. Trans. geol. Soc. S. Afr. 75: 1-3. Lock, B. E. 1973. Tertiary limestones at Needs Camp, near East London. Trans. geol. Soc. S. Afr. 76: 1-5. 158 ANNALS OF THE SOUTH AFRICAN MUSEUM MarTINI, E. 1971. Standard Tertiary and Quaternary calcareous nannoplankton zonation. In: Farinaccl, A. ed. Proc. II Planktonic Conf. Roma 1970: 739-785. Roma: Edizioni Technoscienza 739-785. Maubp, R. R. & Orr, W. N. 1975. Aspects of Post-Karroo geology in the Richards Bay area. Trans. geol. Soc. S. Afr. 78: 101-109. SIESSER, W. G. 1971. Petrology of some South African coastal and offshore carbonate rocks and sediments. SANCOR Mar. Geol. Prog. Bull. 3. Cape Town: Dept. Geology, University of Cape Town. SIESSER, W. G. 1972. Petrology of the South African Coastal limestones. Trans. geol. Soc. S. Afr. 75: 177-185. SIESSER, W. G. 1977. Upper Eocene age of marine sediments at Bogenfels, South West Africa, based on calcareous nannofossils. In: Papers on biostratigraphic research Bull. geol. Surv. Rep. S. Afr. 60: 72-74. SIESSER, W. G. & DINGLE, R. V. 1979. Tertiary sea-level movements around southern Africa. (Abstract.) AAPG-SEPM annual Convention Houston, Abstracts of Papers: 165. STAPLETON, R. P. 1977. Planktonic foraminifera and the age of the Uloa Pecten bed. In: Papers on biostratigraphic research. Bull. geol. Surv. Rep. S. Afr. 60: 11-17. TRUSWELL, J. F. 1977. The geological evolution of South Africa. Purnell: Cape Town. VaIL, P. R., MircHuM, R. M. & THompson, S. 1977. Seismic stratigraphy and global changes of sea level. Part 4: global cycles of relative changes of sea level. In: PAYTON, C. E. ed. Seismic stratigraphy—applications to hydrocarbon exploration. Mem. Am. Assn. Petrol. Geol. 26: 83-97. 6. SYSTEMATIC papers must conform to the /nternational code of zoological nomenclature (particularly Articles 22 and 51). | Names of new taxa, combinations, synonyms, etc., when used for the first time, must be followed by the appropriate Latin (not English) abbreviation, e.g. gen. nov., sp. noy., comb. nov., syn. nov., etc. An author’s name when cited must follow the name of the taxon without intervening punctuation and not be abbreviated; if the year is added, a comma must separate author’s name and year. The author’s name (and date, if cited) must be placed in parentheses if a species or subspecies is transferred from its original genus. The name of a subsequent user of a scientific name must be separated from the scientific name by a colon. Synonymy arrangement should be according to chronology of names, i.e. all published scientific names by which the species previously has been designated are listed in chronological order, with all references to that name following in chronological order, e.g.: Family Nuculanidae Nuculana (Lembulus) bicuspidata (Gould, 1845) Figs 14-1S5SA Nucula (Leda) bicuspidata Gould, 1845: 37. Leda plicifera A. Adams, 1856: 50. Laeda bicuspidata Hanley, 1859: 118, pl. 228 (fig. 73). Sowerby, 1871: pl. 2 (fig. 8a—b). Nucula largillierti Philippi, 1861: 87. Leda bicuspidata: Nicklés, 1950: 163, fig. 301; 1955: 110. Barnard, 1964: 234, figs 8-9. Note punctuation in the above example: comma separates author’s name and year semicolon separates more than one reference by the same author full stop separates references by different authors figures of plates are enclosed in parentheses to distinguish them from text-figures dash, not comma, separates consecutive numbers Synonymy arrangement according to chronology of bibliographic references, whereby the year is placed in front of each entry, and the synonym repeated in full for each entry, is ‘not acceptable. In describing new Species, One specimen must be designated as the holotype; other speci- mens mentioned in the original description are to be designated paratypes; additional material not regarded as paratypes should be listed separately. The complete data (registration number, depository, description of specimen, locality, collector, date) of the holotype and paratypes must be recorded, e.g.: Holotype SAM-—A13535 in the South African Museum, Cape Town. Adult female from mid-tide region, King’s Beach Port Elizabeth (33°51’S 25°39’E), collected by A. Smith, 15 January 1973. Note standard form of writing South African Museum registration numbers and date. 7. SPECIAL HOUSE RULES Capital initial letters (a) The Figures, Maps and Tables of the paper when referred to in the text e.g. ©... the Figure depicting C. namacolus ...’; *. . . in C. namacolus (Fig. 10)...’ (b) The prefixes of prefixed surnames in all languages, when used in the text, if not preceded by initials or full names e.g. Du Toit but A.L.du Toit; Von Huene but F. von Huene (c) Scientific names, but not their vernacular derivatives e.g. Therocephalia, but therocephalian Punctuation should be loose, omitting all not strictly necessary Reference to the author should be expressed in the third person Roman numerals should be converted to arabic, except when forming part of the title of a book or article, such as “Revision of the Crustacea. Part VIII. The Amphipoda.’ Specific name must not stand alone, but be preceded by the generic name or its abbreviation to initial capital letter, provided the same generic name is used consecutively. Name of new genus or species is not to be included in the title: it should be included in the abstract, counter to Recommendation 23 of the Code, to meet the requirements of Biological Abstracts. WILLIAM G. SIESSER & GREGORY A. MILES CALCAREOUS NANNOFOSSILS AND PLANKTIC FORAMINIFERS IN TERTIARY LIMESTONES, NATAL AND EASTERN CAPE, SOUTH AFRICA S, , Oc VOLUME 79 PART 7 FEBRUARY 1980 OF THE SOUTH AFRICAN MUSEUM CAPE TOWN INSTRUCTIONS TO AUTHORS 1. MATERIAL should be original and not published elsewhere, in whole or in part. 2. LAYOUT should be as follows: (a) Centred masthead to consist of Title: informative but concise, without abbreviations and not including the names of new genera or species Author’s(s’) name(s) Address(es) of author(s) (institution where work was carried out) Number of illustrations (figures, enumerated maps and tables, in this order) (b) Abstract of not more than 200 words, intelligible to the reader without reference to the text (c) Table of contents giving hierarchy of headings and subheadings (d) Introduction (e) Subject-matter of the paper, divided into sections to correspond with those given in table of contents (f) Summary, if paper is lengthy (g) Acknowledgements (h) References (i) Abbreviations, where these are numerous 3. MANUSCRIPT, to be submitted in triplicate, should be typewritten and neat, double spaced with 2,5 cm margins all round. First lines of paragraphs should be indented. Tables and a list of legends for illustrations should be typed separately, their positions indicated in the text. All pages should be numbered consecutively. Major headings of the paper are centred capitals; first subheadings are shouldered small capitals; second subheadings are shouldered italics; third subheadings are indented, shouldered italics. Further subdivisions should be avoided, as also enumeration (never roman numerals) of headings and abbreviations. Footnotes should be avoided unless they are short and essential. Only generic and specific names should be underlined to indicate italics; all other marking up should be left to editor and publisher. 4. ILLUSTRATIONS should be reducible to a size not exceeding 12 « 18 cm (19 cm including legend); the reduction or enlargement required should be indicated; originals larger than 35 x 47 cm should not be submitted; photographs should be rectangular in shape and final size. A metric scale should appear with all illustrations, otherwise magnification or reduction should be given in the legend; if the latter, then the final reduction or enlargement should be taken into consideration. All illustrations, whether line drawings or photographs, should be termed figures (plates are not printed; half-tones will appear in their proper place in the text) and numbered in a single series. Items of composite figures should be designated by capital letters; lettering of figures is not set in type and should be in lower-case letters. The number of the figure should be lightly marked in pencil on the back of each illustration. 5. REFERENCES cited in text and synonymies should all be included in the list at the end of the paper, using the Harvard System (ibid., idem, loc. cit., op. cit. are not acceptable): (a) Author’s name and year of publication given in text, e.g.: ‘Smith (1969) describes... ‘Smith (1969: 36, fig. 16) describes . “As described (Smith 1969a, 19695; fae aoe “As described (Haughton & Broom Sn ‘As described (Haughton et al. 1927) . Note: no comma separating name and year Dagination indicated by colon, not p. names of joint authors connected by ampersand et al. in text for more than two joint authors, but names of all authors given in list of references. (b) Full references at the end of the paper, arranged alphabetically by names, chronologically within each name, with suffixes a, b, etc. to the year for more than one paper by the same author in that year, e.g. Smith (1969a, 19695) and not Smith (1969, 1969a). For books give title in italics, edition, volume number, place of publication, publisher. For journal article give title of article, title of journal in italics (abbreviated according to the World list o, scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses, volume number, part number (only if independently paged) in parentheses, pagination (first and last pages of article). Examples (note capitalization and punctuation) BULLOUGH, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan. FISCHER, P.-H. 1948. Données sur la résistance et de le vitalité des mollusques. J. Conch., Paris 88: 100-140. FIscHER, P.-H., DuvAL, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines. Archs Zool. exp. gen. 74: 627-634. Koun, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee region of Ceylon. Ann. Mag. nat. Hist. (13) 2: 309-320. Konn, A. J. 19606. Spawning behaviour, egg masses and larval development in Conus from the Indian Ocean. Bull. Bingham oceanogr. Coll. 17 (4): 1-51. THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. In: SCHULTZE, L. Zoologische und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-Afrika 4: 269-270. Jena: Fischer. Denkschr. med.-naturw. Ges. Jena 16: 269-270. (continued inside back cover) ANNALS OF THE SOUTH AFRICAN MUSEUM ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM Volume 79 Band February 1980 Februarie Part”. 7. Deel THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCK! (REPTILIA, ORNITHISCHIA) FROM THE STORMBERG OF SOUTH AFRICA By A. P. SANTA LUCA Cape Town Kaapstad The ANNALS OF THE SOUTH AFRICAN MUSEUM are issued in parts at irregular intervals as material becomes available Obtainable from the South African Museum, P.O. Box 61, Cape Town 8000 Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM word uitgegee in dele op ongereelde tye na gelang van die beskikbaarheid van stof Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad 8000 OUT OF PRINT/UIT DRUK 1, 2(1=3, 588), 30122)4-5..8 tp.) 5032579) 6(1, t—p.i.); 744), 8, 9(1—2, 7), 10(1-3), 1101-2;5; 7, tpi), -15G-5), 2402): 27, 31-3), 326), 33 Copyright enquiries to the South African Museum Kopieregnavrae aan die Suid-Afrikaanse Museum ISBN 0 908407 86 6 Printed in South Africa by In Suid-Afrika gedruk deur The Rustica Press, Pty., Ltd., Die Rustica-pers, Edms., Bpk., Court Road, Wynberg, Cape Courtweg, Wynberg, Kaap THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI (REPTILIA, ORNITHISCHIA) FROM THE STORMBERG OF SOUTH AFRICA By A. P. SANTA LUCA The University of Texas Health Science Center, Dallas (With 23 figures, 1 table and 1 appendix) LMS. accepted 30 August 1979] ABSTRACT Heterodontosaurus tucki (SAM—K1332), from the Upper Red Beds of the Stormberg Series, comprises the only known complete postcranial skeleton of an early ornithischian dinosaur. It is characterized by: length just over 1 m; a short presacral, especially dorsal, region; six fused sacrals; ossified tendons only in the dorsal region; humerus with large deltopectoral crest and entepicondyle; ulna with an olecranon process; nine carpal elements; three functional, parallel, manual digits; elongated tibiofibula; functional tibiotarsus and tarsometatarsus; small, robust prepubis; and ischium without an obturator process. H. tucki was bipedal but probably capable of slow quadrupedal progression. The hand was a grasping organ and the forelimb possessed powerful flexor musculature. The hind limb was abducted and protracted, but definitely not parasagittal nor vertical. The tail was not rigid. H. tucki differed in many important characters from fabrosaurids, indicating a long period of ornithischian evolution still unknown. Resemblances are found to Jurassic-Cretaceous ornithopods and non-ornithopods, particularly small Ceratopsia. Ornithopods are redefined as only those ornithischians having an obturator process. H. tucki is classified as a non- ornithopod of unknown subordinal status. It argues for the existence of a non-ornithopodous radiation possibly ancestral to some later non-ornithopods. CONTENTS PAGE Introduction Steg ts : : ; : ‘ ‘ . Psk6o Historical survey ; : ; 2 : : ~ 160 Material . ‘ 3 j . : ’ ‘ oe) GS Description of Heterodontosaurus tucki . : . =. 163 Vertebral column : ; g : é : «163 Pectoral girdle and forelimb . : y : vee? Pelvic girdle and hind limb . ‘ : 3 . »h84 Discussion . : : : ‘ ; : é ; 53 HAOT Diagnosis . ; ; : : ‘ : - . 1 OT Morphological interpretation . : : : LO Heterodontosaurus and Fabrosaurus ' : «ft PASS The importance of H. tucki . ; : ‘ 2 & 200 The significance of the obturator process. cu 20 H. tucki and Ornithischian classification : S202 Summary . ‘ : : : : : gat 3 5 20S Acknowledgements . 3 é ; . : : . » 204 References . : : ; : ; : : : fe m2O5, Abbreviations . ; ; E : : : : Ei eeO7 Appendix 1 ; ‘ : 5 ; : : : ere 209 159 Ann. S. Afr. Mus. 79 (7), 1980: 159-211, 23 figs, 1 table, 1 appendix. 160 ANNALS OF THE SOUTH AFRICAN MUSEUM INTRODUCTION All known Late Triassic ornithischians are classified either as fabrosaurids or heterodontosaurids. The family Fabrosauridae has recently been surveyed, in part, by Galton (1978). In this family he included Fabrosaurus australis Ginsberg, 1964, Echinodon Owen, 1861, Nanosaurus Marsh, 1877, and Lesotho- saurus diagnosticus Galton, 1978 gen. et sp. nov. This last is the fabrosaurid material described by Thulborn (1970a, 1972) as Fabrosaurus australis; it is the only specimen of the above with sufficient postcranial material for comparison with Heterodontosaurus tucki. The family Heterodontosauridae includes Heterodontosaurus tucki Crompton & Charig, 1962, Lycorhinus angustidens Haughton, 1924, Abricto- saurus consors (Thulborn, 1974) Lanasaurus scalpridens Gow, 1975, Gerano- saurus atavus Broom, 1911, and Pisanosaurus mertii Casamiquela, 1967. P. mertii has been classified as a hypsilophodontid by Galton (1972) but as a heterodontosaurid by Bonaparte (1976). Abrictosaurus consors (Hopson 1975) comprises the material which Thulborn (1974) described as Lycorhinus consors (specimen B54, Department of Zoology, University College, London). But apart from SAM-K1332, the specimen described here, little heterodontosaurid post- cranial material exists. Previous discussions of Heterodontosaurus have dealt primarily with its cranial and dental anatomy (Crompton & Charig 1962; Charig & Crompton 1974). Little attention was given to the available postcranial skeleton (but see Santa Luca, Crompton & Charig 1976), though it is the most complete of any known Triassic ornithischian. This is the first study to describe heterodonto- saurid morphology in detail; the purpose is to present a thorough analysis of the postcranial anatomy and of its implications for ornithischian phylogeny. HISTORICAL SURVEY Though no detailed study of Heterodontosaurus has previously appeared, it has been discussed frequently in the press. Most of the controversy has centred around the familial status and generic synonymies of Heterodontosaurus. Crompton & Charig’s (1962) announcement of the first Heterodontosaurus skull described the dentition and diagnosed the specimen as ornithischian. After short comparisons with iguanodonts and hadrosaurs, they provisionally assigned Heterodontosaurus to the suborder Ornithopoda but not to any family within that suborder. Romer (1966: 370) made it the monotypic genus of the family Heterodonto- sauridae. However, Thulborn (1970a: 430) assigned it to the Hypsilophodontidae in his study of a fabrosaurid skull, but he gave no reasons for so doing. Subse- quently, the systematic position of Heterodontosaurus became greatly confused as Thulborn (1970b, 1974) described another South African ornithischian as congeneric with Lycorhinus angustidens Broom, 1911, and then argued that the 161 THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI sjseo oie ydeisojoyd siy} Ul s[qipueu pue [[Nys oy, ‘[eUIsIIO dy} JO MOIA [BIO}e] WoT Ul CEE TO WVS (s11eYyD 2 UoJdWIOID) 149n] SnAnvsojuopodsajayy “| “Biz 162 ANNALS OF THE SOUTH AFRICAN MUSEUM name Heterodontosaurus was but a junior synonym of Lycorhinus and therefore not valid. In answer to this, Charig & Crompton (1974) and Hopson (1975) adequately demonstrated that generic distinctions in the dentition did exist between Heterodontosaurus and Lycorhinus. Furthermore, Galton (1972), Charig & Crompton (1974), and Hopson (1975) have all shown that Heterodontosaurus is sufficiently distinct from the Hypsilophodontidae to warrant separate familial status. Thulborn (1974) later accepted a familial distinction, but continues to refer to this genus as ‘Lycorhinus’ (Thulborn 1978). Because of its dentition and some of its cranial characters, Heterodonto- saurus has always been considered a rather specialized ornithischian. From this Thulborn (1970a, 1971a, 1972, 1974) has inferred that heterodontosaurids were a short-lived evolutionary divergence from the basal ornithischian stock. In a previous publication Santa Luca et al. (1976), only assumed that H. tucki itself could not be ancestral to later ornithischians. This hypothesis will be thoroughly examined at the end of this study since some important similarities in the post- cranial skeleton of H. tucki and later ornithischians do exist. The result of previous work has been to clarify the familial and generic status of Heterodontosaurus. However, the question of the subordinal status of Heterodontosaurus has never been examined. It has simply been standard practice to classify all bipedal ornithischians as Ornithopoda. This is unsatis- factory for some bipedal ornithischians (e.g. pachycephalosaurs) and so the question of subordinal status will be taken up in the discussion. Fig. 2. H. tucki. SAM-—K1332. Skull, right lateral view. Xx 1. THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 163 MATERIAL The specimen described here (South African Museum K 1332) is on loan to the Museum of Comparative Zoology, Harvard University. It was discovered in December 1966 in the Upper Red Beds of the Stormberg series, about 1 770 m above sea-level on the northern slopes of Krommespruit Mountain near Voisana in the District of Herschel, Republic of South Africa. The specimen consists of a virtually complete and articulated skeleton of an adult ornithischian dinosaur (Figs 1-2). The precise extent of preservation of each skeletal element has been noted at the beginning of the descriptive sections. A complete list of measurements is provided in the appendix. Comparisons have been made with a cross-section of published ornithischian material. This includes ankylosaurs (Coombs 1978a, 19786), Camptosaurus (Gilmore 1909), ceratopsians (Hatcher, Marsh & Lull 1907; Lull 1933), Fabro- saurus (Thulborn 1972; Galton 1978), hadrosaurs (Lull & Wright 1942), Hypsilophodon (Galton 1974), Iguanodon (Hooley 1925), Microceratops (Maryarska & Osmdlska 1975), pachycephalosaurs (Maryarska & Osmdolska 1974), Protiguanodon and Psittacosaurus (Osborn 1923, 1924), Protoceratops (Brown & Schlaikjer 1940), stegosaurs (Gilmore 1914), and Thescelosaurus (Gilmore 1915). DESCRIPTION OF HETERODONTOSAURUS TUCKI VERTEBRAL COLUMN The vertebral column of Heterodontosaurus is virtually complete and in articulated condition. However, it has been left in a bed of matrix and only the left lateral surfaces of the vertebrae are generally visible (Fig. 3); a few of the vertebral bodies, in the anterior dorsal region, can be seen from the right side (Fig. 4). In the presacral column, the most notable absence is the atlas which cannot be reconstructed from the few fragments which remain; however, the axis is exceptionally well preserved. The transverse processes of the posterior cervicals are cracked so the precise angle of the processes relative to the neural arch is uncertain. The anterior part of the centrum of C9 is missing, thus the shape of this centrum and its effect on curvature in the neck are indeterminate. The posterior cervical and the dorsal ribs overlay these vertebral bodies and obscure their structure. The most severe deformation is in the middle and posterior dorsals. Here the neural arches are collapsed downwards over the centra; the transverse processes, instead of being horizontal, are now flush _ against the centra and point downward. The ilia have been squeezed together, displacing the sacral ribs. The caudal vertebrae are the best preserved but are also embedded in matrix (Figs 6, 8) so only the left lateral surface is visible. Cervical vertebrae (Figs 1, 3, 5A) The cervical vertebrae of H. tucki can be divided into two groups based on serial changes in centrum and in neural arch shape. C2-5 have longer centra, 164 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 3. H. tucki. Stereophotograph of main matrix block, left side. Scale = 5 cm. longer neurocentral junctions (10-11 mm) and more widely separated pre- and postzygapophyses. Each centrum has a moderate ventral keel, concave in lateral outline, which is not strongly differentiated from the centrum itself. The axis has a rather more elongate centrum with a less marked lateral concavity and ventral keel than the other anterior cervicals. The odontoid process is 7,5 mm long. C6—-9 have shorter centra, narrower neurocentral junctions (decreasing from 8 mm in C6 to 5 mm in C9), more closely apposed pre- and post- zygapophyses and much more strongly developed diapophyseal processes. The keeling and concave ventral outline are more pronounced in C6-9. Furthermore, a ridge outlines the anterior and posterior intercentral margins. The anterior ridge continues up the side of the centrum and joins the parapophyseal promi- nence. The posterior ridge continues along the ventral margin of the centrum. These ridges outline a much deeper concavity below the parapophyses in C6—9 than in C2-5. They also increase the transverse width of the ventral keel which is narrow and sharp in C2-5, but flat and several millimetres wide in C6-9. While some of these distinctions typify the cervical vertebrae of other orni- thischians (e.g. Hypsilophodon), the division into two groups is more pronounced in A. tucki. Fig. 4. H. tucki. Stereophotograph of main matrix block, right side. Scale = 5 cm. THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 165 Only two features of the cervical centra do not show the dichotomous variation noted above. First, the height of the centra is approximately constant throughout the series. Second, the absolute position of the parapophysis is constant in all the cervical vertebrae. The position of the diapophysis varies but not in the fashion noted above for the cervical centra: in C3—5 (the axis has no rib facet) it lies just above and behind the parapophysis, but the distance between the two facets progressively increases so that the diapophysis lies at the level of the zygapophysis on C7 and above that level on C8 and C9. Three different kinds of neural spines are found in successive groups of cervical vertebrae: C2, 3, 4; C5, 6; C7, 8, 9. The neural arch of the axis has a very prominent spine, the long axis of which is almost horizontal, parallel to the long axis of the centrum. The arch is lateromedially compressed except at the posterior ventral margin. Here the arch develops two lateral processes or flanges; these extend from the distal tip of the arch anteriorly and inferiorly to the postzygapophyses. The neural spines of C3 and 4 are successively smaller versions of this form. C5 and 6 have small, narrow spines which project anteriorly and dorsally between the postzygapophyses of the preceding vertebra. The spine of C4 is inclined about 45° to the horizontal, C5 about 30°. The spines of C7-9 are narrow vertical processes; in C7 and 8 the tip of the spine is broken and the height uncertain; in C9 it rises about 7 mm above the level of the zygapophyses. _ The orientation and position of the zygapophyses also vary in the cervical region. The prezygapophyses of the axis for articulation with the atlantal neural arch are flat and face laterally. The transverse axis of the joint is horizontal at C2/3, but it becomes successively more angulated until at C6/7 it is about 70° above the horizontal. The zygapophyseal joints of C7/8 and C8/9 are covered by matrix, but the transverse axis of C9/D1 is less erect than the axis of C6/7. The transverse axis is horizontal again at the D2/3 articulation (see D3/D4 in Fig. 5B). Since the distance between the pre- and postzygapophyses of any single vertebra varies with the length of the centrum, the zygapophyses are closer together in the posterior than in the anterior group of cervical vertebra. In the cervical series, the anterior and posterior surfaces of the centra are not perpendicular to the long axis of the centrum. Thus, the centra of C3 and 4 have a parallelogram-shaped profile; those of C5 and 6 are approximately rectangular, those of C7 and 8 trapezoidal (that of C9 is indeterminate). The differently shaped centra, when articulated, automatically impart a curvature to - the neck. Though facets for the cervical ribs begin with C3, only the ribs of C4, 5, 8 and 9 are preserved or worked out of matrix (Figs 1, 3), but this is sufficient to infer the structure of the entire series. The ribs of C3-—5 are alike, being very short with capitular and tubercular processes of about equal length; both arms are equally divergent from the axis of the rib shaft and so form a Y. The rib of C9 (Fig. 5A) is incomplete distally, the preserved portion is 70 mm long; the 166 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 5. H. tucki. A. Cervical vertebrae, left lateral view, atlas missing. B. Dorsal vertebrae, left lateral view. Scales = 5 cm. tubercular process (6 mm long) is only half the length of the capitular process (12 mm). The tubercular process is not divergent from but lies on the long axis of the rib shaft; the head is flattened anteroposteriorly and bony excrescences indicate a strong attachment to the diapophysis. Comparable ridges are found on the diapophyses and these are especially well developed-on C8 and 9. The capitular process of C9 diverges at about 60° from the long axis of the shaft; its head is rounded and marked by a bony ridge along only the anterior margin of the articular facet. Dorsal vertebrae (Figs 1, 3, 5B) Several features of the tenth vertebra indicate that it is the first dorsal; counting thus, there are then 12 dorsal vertebrae. Most importantly, compared to C6-9, the tenth centrum is elongated and has a longer neurocentral junction; the intercentral margins are not raised into strong ridges and the ventral keeling is considerably smaller. However, the first three dorsals are, in some features, intermediate between the posterior cervicals and the remaining dorsals. The presence of a small ventral keel is an intermediate condition in that the succeeding dorsals have none. The level of the zygapophyses above the centrum decreases gradually in the first three dorsals from their position high above the centrum in the posterior cervicals (D1, Fig. 5B) to just above the centrum in most of the other dorsals (D3/4, Fig. 5B). The angle of inclination of the zygapophyseal facets gradually THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 167 decreases from D1 so that it is horizontal between D3 and 4. The diapophyses and transverse processes also become lower in the first three dorsals so that on D3 they are immediately above the parapophysis and connected to it by an oblique crest. Finally, the neural spines of the first three dorsals change from a very narrow-based process to a long-based process, about as long as the centrum itself. D4-10 may be considered typical dorsal vertebrae. The centra are rect- angular in lateral outline (the outlines were determined by radiograph for Figure 5B); consequently, the shape of the centra does not impart a curvature to the dorsal region. The centra lack the strongly marked vertical ridges on the intercentral margins seen in the posterior cervicals. The transverse processes lie approximately at the same level as the zygapophyses; furthermore, the two rib facets also lie on this level, the parapophyses (having risen completely off the centrum) on the ventromedial surface and the diapophysis on the lateral surface of the transverse process. The transverse processes are horizontal and angled posteriorly. In D6—10 the transverse process becomes bifid: that is, an incisure develops between that part carrying the parapophysis anteriorly and the part carrying the diapophysis posteriorly. The two rib facets are closer together in the posteriormost dorsals, still divided in D11 but completely merged into a single facet in D12. While orientation of the zygapophyseal facets varies with position in the dorsal series, exact orientation is unknown because the neural arches were broken just above the centra and displaced ventrally in most of the middle dorsals. The transverse processes were also broken and folded downward, so as to lie in contact with the lateral surface of the centra. Only the processes of D3, 4, 5, 11 and 12 retain their original horizontal orientation; thus, only in these vertebrae can the orientation of the zygapophyses be determined. The facets between D3/4 and D4/5 are horizontal and those between D10/11 and D11/12 are inclined at about 45°. In the other vertebrae, if the transverse _ processes were restored to their horizontal position, then the inclination of the facets would also be about 45°, similar to that of D11 and 12. The height of the zygapophyses on the neural arch decreases in the first three dorsals; but from D4~-10 the height is indeterminate since the neural arches were displaced ventrally as described above. In D11 the prezygapophysis is 4-5 mm above the centrum, the postzygapophysis about 10 mm above this level. The zygapophyses of D12 are about 7 mm above the centrum. The size and shape of the neural spine vary throughout the dorsal region. . The precise height in D4~9 is uncertain, but the spines are clearly antero- posteriorly elongated at both the base and vertex. The height of the spines above the zygapophyses is about 10 mm in D3 and 14 mm in D12. The difference in height between D3 and D12 is actually greater because the zygapophyses them- selves articulate at a higher level in the more posterior dorsal vertebrae. The ribs of the transitional dorsals differ from those of the posterior cervicals only in having shorter tubercular and capitular processes. Total rib 168 ANNALS OF THE SOUTH AFRICAN MUSEUM length is indeterminate: the preserved part of dorsal rib 3? is 90 mm long. In the remaining dorsal ribs except the last, the tubercular process is so reduced that the tuberculum lies on the dorsal surface of the rib a short distance behind the capitulum. A line connecting capitulum and tuberculum makes a 45° angle with the proximal part of the shaft. The last dorsal rib is very short and has a single head which articulates with a reduced transverse process. Sacral vertebrae (Figs 3, 7A) The sacrum consists of six vertebrae, the centra of which are completely fused. Only a few details about the structure of the sacrum can be obtained since the sacrum is only partially exposed (and only on the left side). Sacrals 1 and 2 articulate with the anterior iliac process. The transverse processes of SI and 2 resemble those of D12: they are short, horizontal, dorso- ventrally flattened and arise from the middle of the centrum. The first two sacrals do not have typical sacral ribs which cover a large portion of the lateral central surface; rather, the ribs connect the transverse processes with the ilia. The prezygapophyses of S1 are exactly like those of D12, the transverse and anteroposterior axes angled about 45° to the horizontal. The prezygapophyses of S2 are completely obscured by matrix. The neural spines are as high as those of the posterior dorsals but narrower; D12 resembles S1 and 2 in this last respect more than it does D7-11. S3 articulates with the ilia immediately dorsal and anterior to the pubic peduncle. The structure of the sacral rib is indeterminate since the centrum has a vertical fracture along which it has sheared; thus, only the point of articulation is clear, but not the shape of the bones forming it. Dorsally, only the spinous portion of the neural arch is visible: it is anteroposteriorly longer than that of Sl and 2. In dorsal view only the basal part of the neural spine of S4 is visible, the dorsal tip being eroded away. The remaining portion resembles that of S3. S5 has an anteroposteriorly elongated transverse process which seems to be continuous with a ventrolateral projection from the centrum; this may indicate a true sacral rib. S6 has a narrow transverse process which angles caudally and meets the posterior iliac process. The ventral surface of each centrum is concave anteroposteriorly; it is somewhat flattened transversally in S1-3 but has a slight ridge in S5 and 6. Ossified tendons are found on the sides of the neural spines beginning abruptly at D4 (Figs 1, 3). They continue throughout the dorsal series and are found on the sides of the sacral neural spines. They disappear at S5 or 6 and are not found at all posterior to the sacrum. Caudal vertebrae (Figs 6, 7B—C, 8-9) The caudal series is not complete, but a total of 28 vertebrae remain. The first 12 are preserved in two separate but contiguous blocks of matrix. A further block contains another group of 16 articulated caudals. The number of vertebrae THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 169 Fig. 6. H. tucki. Stereophotograph of caudal vertebrae 3-12, left lateral view. Note distal portions of ischium and postpubic rod in lower left corner of matrix block. Scale = 5 cm. that would bridge the gap between the two groups can be estimated by com- paring the mid-central heights of the two vertebrae at the ends of the gap. (This measure seems to decrease uniformly from anterior to posterior caudals, while centrum length increases and decreases several times within the caudal series.) The difference in height is 1,5 mm; this corresponds to 6 vertebrae in the first group and to 9 in the second. Presumably, then, at least 6 but not more than 9 caudals intervened between the two segments preserved. The caudal cenira do not have a consistent pattern of variation in shape or length. In the first seven caudals the anterior central surface is inclined postero- dorsally while the posterior surface is perpendicular to the long axis of the centrum. The remaining centra in the first group are rectangular. Centra 4 to 7 of the last block have a parallelogram outline, the dorsal margin anterior to the ventral; the other centra are rectangular. The length of the centra increases from about 15 mm for sacral 1 to about 18 mm for sacral 11. In the second group, length is approximately constant at 16 mm. Several features are common to the first nine caudal centra. The ventral - margin is markedly concave; this is accentuated by the inferiorly projecting articulation with the chevrons. Beyond caudal 9 the area of heamal arch articulation decreases and the inferior border becomes more gently concave. On the lateral surface of the centra a fossa lies below the transverse process; a ridge marks the middle of the ventral margin. The size of the fossa decreases from the first to ninth caudal and disappears at caudal 10. Transverse processes are found on all twelve caudals of the first group and on the first twelve caudals of the second. The processes maintain the same ’ relative position throughout, projecting from the middle of the centrum just below the neurocentral junction. The processes of the first nine caudals are horizontal and angled posteriorly; all the rest are successively smaller projections but perpendicular to the centrum. The neural arches of the first ten caudals have spinous processes, the remaining caudals do not. In the first ten, the spine both decreases in height and inclines more posteriorly. The first seven spines are about 15 mm high while the 170 ANNALS OF THE SOUTH AFRICAN MUSEUM { —- + - 1 2 3 4 Fig. 7. H. tucki. A. Sacral vertebrae, left lateral view; outline of S 2-5 taken from radiograph. B. First two caudal vertebrae, from main matrix block, left lateral view. C. Caudal vertebrae 3-12, from matrix block illustrated in Figure 6, left lateral view. Scale = 5 cm. remainder diminish rapidly in height. The inclination of the spines to the horizontal plane decreases from about 90° in the first sacral to about 45° in the tenth; in this last, the spine is almost parallel to the postzygapophyseal process. The zygapophyseal processes are set at about 40° to the horizontal plane in the first group of caudals; this angle decreases to about 30° in the second group. The transverse axis of the articular facets increases from about 45° relative to the horizontal plane at the sacrocaudal junction to almost 90° after the first six or seven caudals. All the caudal vertebrae preserved had chevrons. The first six chevrons are expanded proximally, having relatively large articular contacts with the centra; Fig. 8. H. tucki. Stereophotograph of second block of caudal vertebrae. Scale = 5 cm. THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 171 distally, these chevrons narrow to a small rod. From the seventh chevron of the first group to the sixth of the second, the distal end of the chevron is antero- posteriorly expanded into a knob. The length of the chevrons decreases pro- gressively from about the ninth. The first chevron preserved on the last block is 20-21 mm long; the fifth behind that is about 18 mm long; and the last complete chevron is still 16 mm long (on the third from last vertebra). If the reduction in chevron length were a linear function, then 15-20 vertebrae would have com- pleted the caudal series if the smallest chevron were 8-10 mm long. Comparisons The total number of presacral and sacral vertebrae in H. tucki (9+12+6) _ cannot be matched in any of the well-known ornithopods such as Hypsilophodon (9+15+6), Camptosaurus (9+16+4/5), or Jguanodon (11+17-+6). Only ceratopsians (including Protoceratops), Psittacosaurus and Protiguanodon have a similar sacral and pre-sacral count. As a whole, the vertebral structure of H. tucki resembles that of Hypsilo- Phodon most closely, though similarities to the non-ornithopods are frequent. The axis has a structure similar to that of the ornithopods Hypsilophodon and Camptosaurus, but also to that of Stegosaurus. In the remaining anterior cervicals, H. tucki and Hypsilophodon are quite similar. However, the deeply concave and short posterior cervical centra are not matched in Hypsilophodon but rather in Protoceratops. The morphology of the dorsal vertebrae agrees with that of typical ornitho- pods such as Hypsilophodon, Camptosaurus and Thescelosaurus and with what is known of Fabrosaurus. The dorsals, however, are unlike those of hadrosaurs, Fig. 9. H. tucki. Caudal vertebrae on second matrix block illustrated in Figure 8. Scale = 5cm. 172 ANNALS OF THE SOUTH AFRICAN MUSEUM iguanodonts and some non-ornithopods (Stegosaurus, Centrosaurus) which have high and relatively narrow neural spines. The sacral vertebrae of H. tucki are virtually unknown since they are buried in matrix between the ilia. The caudal vertebrae resemble those of Hypsilophodon, Thescelosaurus and Fabrosaurus. In Camptosaurus, the anterior caudals are considerably shorter but the posterior caudals much more like those of H. tucki. The caudals of H. tucki do not have the anteroposteriorly compressed centra nor the high neural spines of iguanodonts and hadrosaurs. PECTORAL GIRDLE AND FORELIMB Scapula (Figs 1, 3-4, 10D) Both scapulae are preserved, the left presenting the external aspect, the right the inferior-external aspect. The glenoid cavity is clearly visible on the right (Figs 1, 3) as the humerus is disarticulated from the scapula, but the humeral head lies in the glenoid on the left side (Fig. 4). Both scapulae lie approximately parallel to the vertebral column, the anterior extremity somewhat more ventral than the posterior. In the reconstruction the scapula has been reoriented parallel to the vertebral column in a position comparable to that seen in well-preserved hadrosaurs (Lull & Wright 1942). An anteroventral- posterodorsal orientation of the scapular long axis would also be quite accept- Fig. 10. H. tucki. A. Medial view of olecranon process, right ulna. B. Proximal articular surface, right radius and ulna. C. Right distal tarsals and proximal portion of metatarsals with reduced fifth digit, ventral view. D. Left scapula, lateral view. Scales: A-B = 2,5 cm; C-D = 5 cm. THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 173 able. In either case, the glenoid long axis should be primarily ventral in position, not posterior. The scapula is the longest element of the shoulder girdle-forelimb complex. The caudal portion is thin and broadened into a blade; the bone at the margin here has a roughened and unfinished appearance which probably indicates a cartilaginous suprascapular extension. Along the blade-like caudal portion of the scapula, the ventral margin is strongly concave while this part of the dorsal margin is straight. Cranially, the scapula tapers considerably: it becomes sub- circular in cross-section about 20 mm proximal to the scapulocoracoid suture; it broadens out again above the glenoid but remains thick in cross-section. The scapula has only a gentle curvature to conform to the thoracic wall: the arc 1s 90 mm, the chord 85 mm; most of the curvature occurs just proximal to the glenoid fossa. Along the dorsal margin above the glenoid fossa, a prominent acromial process rises about 9 mm above the scapula. The coracoid edge of the process is damaged so there is no evidence of a clavicular facet. A well-developed glenoid tubercle appears at the posterior lip of the glenoid fossa, separated from the lip by a small fissure. It probably marks the attachment of the scapular head of triceps. The glenoid itself is anteroposteriorly concave (10 mm wide) and transversely flat (5 mm thick); the cavity is 3-4 mm deep. The scapula and coracoid are firmly fused, each contributing about half of the articular area. Coracoid (Figs 1, 3, 10D) The left coracoid is complete except for a small area opposite the acromion and for the distal end of the plate; only the glenoid portion of the right coracoid is preserved. The coracoid plate has flat proximal and distal halves, bent about a perpendicular to the long axis of the scapula. The proximal half is in line with the glenoid portion of the scapula, the distal half is bent medially relative to the proximal. An ovoid tubercle lies at the ventral margin approximately 10 mm distal to the glenoid fossa, at the junction of the proximal and distal coracoid halves; it is perhaps associated with the coracobrachialis or costocoracoideus muscle. Sternum (Figs 1, 3) What seems to be a thin sternal plate lies in matrix within the angle formed by the left humerus and scapula. Its approximate dimensions are 35 mm by 18 mm. However, the exact shape and orientation of the plate are unknown since its complete outline is not discernible: it may be either quadrangular like the sternal plates of Hypsilophodon (Galton 1974), or quadrangular with a handle-like process like the plates of Iguanodon atherfieldensis (Hooley 1925). Humerus (Figs 1, 3-4, 11A) Both humeri are complete but still in matrix, the left gives a posterior view and the right an anterior view so no details are lost. The shaft is only slightly 174 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 11. H. tucki. A. Right humerus, dorsal view. B. Right manus, medial view. Seales: A> — Si cmis B — 25cm twisted about its long axis, so the transverse axis of the proximal and of the distal articular surfaces are virtually parallel. The proximal part of the shaft with the deltopectoral crest is retroflexed relative to the distal part; the angle formed is about 30°. Proximally, the humerus has a moderately well-developed articular head, bounded laterally by the superior margin of the deltopectoral crest and medially by a tuberosity lying just below the articular eminence. The head lies in the middle of the superior surface; its diameter is greater than the thickness of the proximal end of the shaft and a small fossa occurs just below the head anteriorly while a buttress of bone lies below the head posteriorly. The deltopectoral crest occupies about 40 per cent of the lateral margin of the humerus; it ends abruptly and joins the shaft at nearly a right angle. The crest is directed anteriorly as well as laterally, so the anterior surface of the crest and shaft is concave, the posterior surface convex. The edge of the crest is thin except superiorly where it forms part of the tuberosity (for the insertion of the deltoid) and inferiorly where it is rugose and thickened for the attachment of the pectoralis muscle. Below the deltopectoral crest the shaft is subcircular in cross-section; it becomes more rectangular distally as the supracondylar area is approached. The medial, ulnar condyle is gently rounded both transversely and antero- posteriorly; it extends somewhat lower than the radial condyle. The radial condyle is ridge-like transversely, not rounded; a transverse section through the condyles would thus show a U-shaped ulnar condyle meeting a V-shaped radial condyle. While the anteroposterior axis of the ulnar condyle is parasagittal, the THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 175 ridge of the radial condyle is offset in an anterolateral-posteromedial direction. The radial condyle is limited above and anteriorly by a small transverse ridge, creating a shallow fossa between condyle and ridge. A small supracondylar ridge widens the humerus radially. The medial surface of the humerus just above the ulnar condyle has a strong entepicondyle sharply demarcated from the surface of the shaft. This indicates a correspondingly strong development of forearm flexor musculature. Neither condyle extends on to the dorsal surface of the shaft (Figs 1, 3); the humerus is here completely flat with no olecranon fossa though the ulna has a well-developed olecranon process. Consequently, it would be impossible for the forearm to have been fully extended on the humerus. Radius (Figs 1, 3, 10B, 12-13) Both radii are complete but still partially contained in matrix. The right radius is composed of several realigned fragments and is somewhat distorted, whereas the left is complete and undistorted. On the left radius the distal articulation is turned about 20°-30° medially relative to the proximal; a similar torsion is found in the ulna. In section, the shaft of the radius is generally subcircular; at the extremities it becomes quadrangular. The superior articular surface (Fig. 10B) is semicircular in outline: a straight medial edge in contact with the ulna and a convex lateral margin. The fossa for the radial condyle of the humerus is an elongated shallow sulcus which matches the ridge-like nature of the condyle. Such an arrangement would stabilize the radiohumeral articulation and limit rotation of the radius. The posterior margin of the superior articular surface is raised and extended Fig. 12. H. tucki. Stereophotograph of right radius, ulna, carpus and manus, dorsal view. Scale = 5 cm. 176 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 13. H. tucki. Right radius, ulna, carpus and manus, dorsal view. Scale = 5 cm. THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 177 posteriorly (Fig. 12); the articular surface is thus lengthened and inclined anteroinferiorly. Distally, a small rounded tubercle lies on the dorsal surface of the shaft, 5 mm above the distal end (Fig. 13); this may be associated with an insertion of the extensor carpi radialis muscle. Distally, the shaft is expanded toward the ulnar side; here it makes an oblique contact with the ulnare. The inferior articular surface is planar, in contact with the radiale. Uina (Figs 1, 10A, 10B, 12-13). Both ulnae are preserved, in articulation proximally and distally, but not freed from matrix. The left ulna presents a dorsolateral view, the right a dorso- medial view. Proximally, the ulna has a pronounced olecranon process which rises about 10 mm above the coronoid process; the coronoid process itself projects anteriorly as an almost horizontal shelf. The olecranon part of the articular surface is wider than the coronoid part. The posterior surface of the olecranon bears a rugose, uplifted area of bone for the attachment of the triceps tendon. The radial side of the proximal ulna is convex; a small tubercle lies on this side of the olecranon process, above the coronoid (Fig. 12). The medial surface of the proximal ulna is marked by the jutting ridge for the triceps attachment and by the overhanging ridge of the articular surface; these ridges give the proximal ulna a concave appearance. The ulnar shaft narrows below the coronoid process but widens gradually towards the distal end. A long, low ridge is found on the dorsal part of the shaft beginning below the area of triceps attachment (ur, Fig. 13); it continues to the distal third of the shaft and then terminates. The distal articular surface is transversely convex, fitting into the concave proximal surface of the ulnare. Both ulna and ulnare are in contact with the pisiform laterally. Carpus (Figs 1, 12-15) Both left and right carpi are complete and in virtually undistorted articula- tion with the forelimb and metacarpus of each side. Consequently, the relative position of the carpal elements is certain; this is important since the arrangement of the proximal row of carpals differs from that known in all other ornithischian groups. The carpus is composed of nine ossified elements, arranged in a proximal and distal row with one element sandwiched centrally between these rows. The proximal row contains the radiale, ulnare and pisiform; the distal row contains - one carpal for each of the five metacarpals; finally, one carpal lies beneath the medial part of the ulnare, above distal carpal 2 and the medial part of distal carpal 3. This arrangement contrasts sharply with the carpal construction of known ornithischians. Properly speaking, the bone intervening between the two rows of carpals does not correspond to an os intermedium; rather, it corresponds in position to an os centrale. Comparisons with other ornithischians are hindered because so few com- 178 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 14. H. tucki. Stereophotograph of right carpal region, dorsal view. Scale = 5 cm. plete carpi exist; however, the complete carpus of Camptosaurus dispar (Gilmore 1909) does not contain an os centrale, only an os intermedium. An os centrale does exist in the carpus of the alligator (but along with an os intermedium) so it would not be unusual to find its retention in an early ornithischian archosaur. A number of possibilities exist with respect to the fate of the os intermedium in Heterodontosaurus, but it is impossible to choose among them: (i) the os inter- medium has been lost in the Heterodontosaurus lineage; (i1) the os intermedium has fused with the ulnare, as in Stegosaurus (Gilmore 1914); (iii) the os inter- medium has fused with the radiale; (iv) the os intermedium has become dis- placed inferiorly by the growth of the radiale and/or ulnare, to occupy the position of an os centrale. The phylogenetic history of the carpus in Heterodonto- saurus thus remains unknown for the present. The radiale is a flat, plate-like bone, articulating with the entire distal surface of the radius; it is transversely elongated to match the distal, expanded shaft of the radius. The ulnare, proximodistally thicker than the radiale, contacts the radiale on the inferior half of its medial margin while the superior half of the medial margin is in contact with the distal end of the radial shaft. As noted above, the superior surface of the ulnare is transversely concave to accept the rounded end of the ulnar shaft; distally, the ulnare has a flat articular surface. The lateral surface of the ulnare has two facets, the inferolateral for distal carpal 5, the superolateral for the pisiform. The pisiform is a cuboidal element in contact with both ulnare and ulna. The intervening bony element in the carpus is biplanar, in contact proximally only with the ulnare, but with distal carpals 2, 3 and 4 below. Distal carpals 1 and 2 are both thin, biplanar elements, sandwiched between the carpals above and the metacarpals below. Distal carpal 3 is proximodistally thicker than 1 and 2, rectangular medially, but the lateral margin is diagonal. The diagonal margin permits distal carpal 4 to intervene between the ulnare Wie THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI mct _— -. a. -* mand) e 4 2 ‘ , Ab Fig. 15. H. tucki. Detail of right carpus and manus. Scale = 2,5 cm. 180 ANNALS OF THE SOUTH AFRICAN MUSEUM above and distal carpal 3 below. The inferolateral surface of distal carpal 4 is offset in such a way that digit 4 is highly abducted with respect to digits 1-3. Distal carpal 5 is cuboidal, articulating with the inferior facet of the ulnare’s lateral surface. The inferior facet of the ulnare faces laterally as well as inferiorly; consequently, the fifth digit is abducted and lies parallel to digit 4. The orienta- tion of the carpal bones results unquestionably in a manus with the 3 medial digits parallel to each other and digits 4 and 5 abducted with respect to the first three but parallel to each other. Manus (Figs 1, 11B, 12-13, 15) Both left and right manus are virtually complete. The left lacks only the distal portions of the metacarpal and phalanges of digit 2. The right lacks the distal portion of phalanx 1, digit 1, and of phalanx 3, digit 3; none of the right fifth digit is visible (though the proximal part may still be in matrix, the distal portion is certainly missing). Most of each manus is still embedded in matrix, so only the dorsal surface of the digits is usually visible. The phalangeal formula is 2-3-4—3-2. The base of metacarpal | is transversely expanded and dorsoventrally flattened, making it a thin rectangle in section (metacarpal | is seen in oblique view in Fig. 11B). The metacarpal narrows below the base so the medial and lateral margins are both concave. A tubercle, about 3 mm long, lies at the upper- most medial margin of the metacarpal base, in the position of attachment of the supinator manus muscle of modern reptiles (Fig. 15). On the laterai half of the metacarpal base the articular facet extends a short distance on to the dorsal surface of the metacarpal, allowing extension of the metacarpal on the distal carpal. The metacarpal bears another tubercle at the uppermost lateral margin of the dorsal surface, in contact with a similar protuberance at the base of metacarpal 2. The heads of metacarpals 2-4 each bear such a tubercle at both the uppermost medial and lateral dorsal margin. These tubercles correspond to the attachments of the humerodorsalis muscle of reptiles and amphibians. The condition of Heterodontosaurus indicates at least seven slips of attachment, not a reduced number as seen in Varanus, Sphenodon and Alligator (Haines 1939). Distally, each of the first three metacarpals bears a deep pit on the dorsal surface just proximal to the articular head. This receives a well-developed dorsal process of the phalangeal base. The distal articular surface of the first three metacarpals continues on to the dorsal surface, just distal to the pit. This prolongation of the articular surface permits hyperextension of the proximal — phalanges on the metacarpal heads of digits 1-3. This is important during quadrupedal progression (plantigrade) when the forelimb is weight-bearing, and supports the hypothesis that Heterodontosaurus was capable of quadrupedal locomotion. The distal articular surface of metacarpal 1 is complex and asymmetric (Figs 11B, 15). The outline of the medial condyle is subelliptical, its long axis proximodistal, that is, oriented on the long axis of the metacarpal. The lateral THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 181 condyle is also subelliptical, but its long axis is perpendicular to the medial condyle, that is, directed dorsoventrally. Thus, the lateral condyle rises above the dorsal surface of the metacarpal shaft while the medial condyle is flush with this surface. In addition, the pits for the collateral ligaments have relatively different positions due to the difference in condyle orientation. The pit for the medial collateral ligament lies midway between the dorsal and ventral margins; the pit for the lateral collateral ligament lies just at or above the dorsal surface of the metacarpal. Dorsally, the intercondylar groove lies medial to the midline of the joint, the lateral condyle being wider than the medial. The metacarpal is also asymmetric in length: the lateral edge is longer than the medial edge. This produces an oblique transverse axis of rotation which has two effects: first, the phalanges of digit 1 would be medially offset (though the metacarpal is not) in extension; second, the phalanges would be laterally offset during rotation around the oblique axis of the joint. The bases of metacarpals 2-5 differ from that of metacarpal 1 in being almost square in section, not dorsoventrally flattened. These metacarpals have an almost square dorsal outline resulting from the planar articulations with the distal carpals and with the adjacent metacarpals. The carpometacarpal articu- lation of metacarpal 2 carries on to the dorsal surface for a short distance; this would permit some extension of the metacarpal on the carpus. In metacarpal 3, the articulation with the distal carpal does not extend on to the dorsal meta- carpal surface; the boundary between the carpal and dorsal surfaces is a sharp ridge, not a smooth, curved transition as in metacarpals 1 and 2. The metacarpophalangeal joint of digit 2 is symmetrical, the two condyles being of equal size. Flexion-extension here would result only in movement along the long axis of the metacarpal. The metacarpophalangeal joint of digit 3 is smaller but also seems to be symmetrical (the view of the medial surface is limited since digit 3 lies very close to digit 2). Digits 4 and 5 are extremely reduced and apparently non-functional. The - total length of digit 4 hardly exceeds the length of metacarpal 3. However, the base of metacarpal 4 is almost as large as the bases of metacarpals 2 and 3. This may be correlated with the still large attachment for the humerodorsalis on metacarpal 4. The head of metacarpal 4 is quite small; it is covered with a black concretion which obscures the details of the metacarpophalangeal joint. The base of metacarpal 5 is convex dorsally without any tubercles, indicating the humerodorsalis had lost all attachment to the fifth digit. Atrophy of digits 4 and 5 occurs in both known Late Triassic ornithischians, - Heterodontosaurus and Fabrosaurus, and in almost all other Jurassic-Cretaceous ornithischians as well. Consequently, the reduction of the lateral digits occurred very early in the phylogenetic history of the group or was characteristic of the ancestral group which gave rise to ornithischians. Gilmore’s interpretation, for example, that digit 5 of Camptosaurus ‘was becoming atrophied’ (1909: 256) is thus not accurate. The digit had already become reduced in previous forms and was merely the expression of an ancestral trait. 182 ANNALS OF THE SOUTH AFRICAN MUSEUM Apart from the unguals, the phalanges of the first 3 digits can be placed in 2 categories. The simpler has a symmetric articular facet both proximally and distally; these are the penultimate phalanges of digits 1-3. The base of each has a well-developed median dorsal and median ventral process. Distally, the outline of the articular facet for the ungual phalanx is not uniformly rounded but is somewhat flattened dorsally. The articular surface extends farther on to the ventral than the dorsal aspect of the phalanx. The pits for the collateral liga- ments are found near the dorsal surface of the phalanx on both lateral and medial sides. In the other category of nonungual phalanges are those which intervene between the metacarpal and the penultimate phalanx: phalanx 1 of digit 2, and phalanges 1 and 2 of digit 3. These phalanges have a symmetric base proximally but an asymmetric head distally (Fig. 12). The asymmetry involves, firstly, a torsion of the head relative to the base which turns the ventral surface of the digit somewhat medially. Secondly, the length of the phalanx along the medial margin is a little less than along the lateral margin: this directs the longitudinal axis of the succeeding digit medially. Thirdly, the condyles themselves are asymmetric: the medial condyle is larger than the lateral, and the trochlear surface of the medial condyle is not as steep as that of the lateral condyle. The outline of the medial condyle is a half ellipse,-its long axis pointed ventrally about 45° to the long axis of the phalanx; the outline of the lateral condyle is also a half ellipse, but its long axis parallels that of the phalanx (see the head of phalanx 1, digit 2, Fig. 11B). The depressions for the collateral ligaments are found near the ventral margin medially but near the dorsal border laterally. The transverse axis of the joint is thus dorsolateral-ventromedial, with the dorso- lateral edge being slightly distal to the ventromedial. Rotation about this axis would produce a medial-to-lateral movement during flexion like that of digit 1. The first three ungual phalanges are large, lateromedially compressed claws. The proximal ventral surface bears a protruding flexor tubercle for the attach- ment of the long flexor tendons (Fig. 11B). A comparable development of flexor tubercles is found nowhere else in the Ornithischia. The lateral and the medial surface of the unguals has an irregular depression midway between the dorsal and ventral margins about 3 mm distal to the articular surface. This probably marks the distal attachment of the collateral ligaments. The phalanges of digits 4 and 5 are quite diminutive; those of digit 4 have recognizable articulations permitting flexion-extension, but are very simplified. The articular surfaces are dorsoventrally rounded; the phalangeal heads have a single uniform articular surface without clearly defined medial and lateral condyles. The terminal phalanx of digit 4 is dorsoventrally flattened, triangular in outline. The proximal phalanx of digit 5 is much too small for details to have been retained during preparation. It has a concave proximal and convex distal articular surface. A small fragment of bone distal to the phalanx may be part or all of the next and presumably terminal phalanx: it is only about 1,5 mm long. THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 183 Comparisons ‘The scapular morphology of H. tucki approaches that of conservative ornithopods (Hypsilophodon, Thescelosaurus, and Camptosaurus) more than that of other ornithopods (Anatosaurus or Iguanodon). It differs from that of ceratopsians, stegosaurs and ankylosaurs which have a more nearly uniform width of the scapular blade. However, the non-ornithopods are no more different from H. tucki than Anatosaurus and Iguanodon. The humerus of H. tucki is relatively more robust than that of conservative ornithopods and Fabrosaurus. Most of the differences are probably related to the forelimb capabilities of H. tucki, both quadrupedal and prehensile. The relatively larger deltopectoral crest and the entepicondyle are more reminiscent of the large ornithischian quadrupeds such as Stegosaurus and Triceratops. However, hadrosaurs have a deltopectoral crest larger than that of H. tucki. In one feature, the humerus of H. tucki seems unlike that of any other ornithischian: the lack of a posterior intercondylar groove or depression between the radial and ulnar condyles. This would have severely limited elbow extension; consequently, H. tucki may have assumed a semi-sprawling posture with the forelimbs. In H. tucki, the extremities of the radius are expanded relative to the shaft, somewhat more than they are in the radii of Hypsilophodon, Camptosaurus, Thescelosaurus, and Anatosaurus, but very similar to those of the radius of Iguanodon. In fact, as a whole the forearm of Jguanodon is more similar to that of H. tucki than are the forearms of other ornithischians. The radii of the large quadrupedal forms are variable: those of ceratopsians are but little expanded at the extremities but that of Stegosaurus is much more so. The ulna of H. tucki has a relatively well-developed olecranon process, a feature usually found in the heavy quadrupedal ornithischians ; however, the ulna of the small Microceratops also has an olecranon process. Only Jguanodon among the ornithopods seems to have a comparably developed process. Because the carpus is incomplete in so many ornithischians, the precise orientation of the metacarpals remains uncertain in these cases. Digits 4 and 5 are clearly deviated to the ulnar side in H. tucki; however, in Fabrosaurus and Hypsilophodon the manus was reconstructed with only digit 5 abducted. In the latter two genera, however, digit 4 may also have been abducted since the carpals, which would have determined digit orientation, are missing. The only previously published illustration of the carpus and manus of H. tucki (Bakker & Galton 1974, fig. 1H) is completely inaccurate. On the basis of that inaccurate reconstruction, the authors argued that the hand of H. tucki was identical to that of Triassic saurischians and was most likely inherited from Triassic saurischians. In fact, the properly reconstructed hand of H. tucki is more reminiscent of that of thecodontians than of early saurischians. The phalangeal formula in Heterodontosaurus is 2-3-4-3-2. This agrees with that of Hypsilophodon (according to Gilmore 1915), Thescelosaurus and the ceratopsians; it differs from that of Camptosaurus (2-3-3-3-2), Trachodon (0O-3-3-3-3) and Jguanodon (1-3-3-3-4). Trachodon and Iguanodon are 184 ANNALS OF THE SOUTH AFRICAN MUSEUM anomalous in the number of phalanges in the fifth digit; it seems unlikely that such a phalangeal formula could have been derived from that of H. tucki since it would require an increase in the number of phalanges in a non-functional digit and a change of this digit to a functional role in the animal’s behaviour. By inference, then, Trachodon and Iguanodon could not be derived from any genus having only two phalanges in digit 5; thus, they represent a deviation in ornithischian phylogeny about the origin of which we have no information. PELVIC GIRDLE AND HIND LIMB Pelvis (Figs 1, 3-4, 16-18A) Both pelves are preserved but the left is somewhat damaged. The left ilium has been displaced ventrally while the right has been displaced dorsally. The ilia are also lateromedially compressed; this has compacted the sacral ribs and transverse processes and considerably narrowed the interacetabular width of the pelvis. The left anterior iliac process is broken and shifted ventrally; this creates a greater convexity than the ilium actually had. The left prepubic process has been completely crushed, its original form destroyed; fortunately, the right prepubic process is well preserved. A fracture through the left ilio-ischial suture distorts this region, but the right side is undamaged here. The fracture separating the main matrix block from the first caudal vertebrae block (Fig. 1) also divides the postpubes and ischia in midlength but no bone has been lost and the shafts are complete. Tlium The ilium of H. tucki has a shallow, elongated anterior and posterior process. The anterior process extends about 45 mm anterior to the pubic peduncle; this constitutes 44 per cent of total iliac length. Anteriorly the process veers away from the vertebral column and ends opposite the last dorsal vertebra. A small ventral flange or convexity gives a bulbous appearance to the last 10 mm of the process. The first three sacral vertebrae join the slender anterior process; furthermore, the rib of the last dorsal vertebra was probably fused to the tip of the process. This rib is short (19 mm) and could not have extended beyond the ilium. The posterior iliac process is shorter than the anterior (28 mm from the ischial peduncle) but equally shallow; it ends in a small, rounded expansion. The brevis shelf is horizontal but shallow, extending 4-6 mm medially from the ventral margin of the process to join the last sacral vertebra. The acetabulum of H. tucki is 22 mm high and 20 mm long at its base. The pubic peduncle is relatively long compared to other ornithischians (18 mm from the notch between the anterior process and peduncle to the ilio-pubic suture). The peduncle is almost vertical, inclined only about 20° anterior to a perpen- dicular from the long axis of the ilium. The anterior and posteroventral acetabular margins are raised into a sharp ridge while the dorsal and ventral acetabular margins are rounded and flush with the surface of the ilium and puboischium. THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 185 Fig. 16. H. tucki. Stereophotograph of right pelvis, lateral view. Scale = 5 cm. Fig. 17. H. tucki. Detail of right pelvic region. Scale = 5 cm. 186 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 18. H. tucki. A. Right pelvis, lateral view. B. Left femur, lateral view. Scale = 5 cm. The single most important feature of the acetabulum is the expanded articular surface at the posterodorsal corner. Here the ischiadic peduncle of the ilium flares out, creating an horizontal articular shelf and buttress against which the femoral head would have rested. The ventral surface of this shelf is a continuation of the acetabulum and would have borne articular cartilage during life. The ilio-ischial suture lies below this buttress and is separate from it. Structurally, this articular buttress closely resembles the antitrochanter of birds (completely different from the ‘antitrochanter’ of dinosaurs). Very strong ligaments connect the iliac antitrochanter and the femoral greater trochanter in birds. This system resists the collapse of the body on the femur when the body is supported by only one leg. Unquestionably, the structural similarity to the avian acetabulum argues that some similar mechanical system in the pelvis and hind limb of H. tucki prevented excessive pelvic tilt when weight was borne by a single hind limb. The dorsal margin of the ilium is slightly convex; its arc is 101 mm while the chord (that is, maximum iliac length) is 96,7 mm. A ridge for muscle attachment runs from the tip of the anterior to the tip of the posterior process: beginning on the ventral flange of the anterior process, it passes obliquely upwards to the dorsal margin of the ilium, 20 mm behind the tip of the anterior process; the very last 15 mm of the ridge descends from the dorsal rim to the middle of the lateral surface of the posterior iliac process. In Romer’s reconstruc- tion of the hind limb musculature of Thescelosaurus (1927a) this ridge marks the attachment of the ilio-tibialis and ilio-fibularis muscles. The ilio-tibialis 1 THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 187 would be above the ridge on the anterior iliac process; the ilio-tibialis 2 would be attached above the ridge along the dorsal margin; the ilio-fibularis would lie below the ridge on the posterior process; the ilio-caudalis would lie above the ilio-fibularis and the ridge. A small tubercle situated ventrally on the tip of the posterior process may indicate the attachment of flexor tibialis externus (ilioflexorius). The coccygeofemoralis brevis is, of course, attached to the small, horizontal brevis shelf. Below the ilio-tibialis ridge, the lateral iliac surface presents no muscular ridges. The cortical bone has been fractured in many places; on the right, over- lapping fractures simulate a dividing ridge between muscle masses, but it is purely artificial. This means that no demarcation can be found between the supposed attachments of ilio-femoralis externus and ilio-trochantericus 1. Romer (1927a: 264) could not find any limiting ridges either and so based his reconstruction of Thescelosaurus on the position of the antitrochanter of hadrosaurs and, probably, on the position of these two muscles in birds. Recently, Walker (1977) argued that the iliotrochanterici should be considered part of the ilio-femoralis externus and should not be reconstructed as a separate muscle. Thus, he proposed a single deep dorsal muscle mass, the ilio-femoralis externus, originating below the ilio-tibialis ridge and inserting on the lesser trochanter. The confusion develops because the embryonic origin of the ilio- trochanterici is uncertain (Romer 1927b, 1942). Walker’s argument is weak because he drew his analogy with the development of the thigh musculature of Lacerta (described by Romer 1942); however, the avian condition is certainly a better model for ornithischian musculature than the lacertilian. Ischium The ischium of H. tucki has a long, columnar iliac peduncle and a flat, deep pubic peduncle. The iliac peduncle is 15 mm high and 5 mm in diameter; the pubic peduncle is 9 mm deep. The ischial rods are straight and do not seem to be fused together. The rod bears a robust, laterally projecting ridge (if, Fig. 18A) beginning about 35 mm behind the acetabular border and continuing to the end of the rod; a similar feature is present on the ischium of Protiguanodon. In mid- length, the ridge becomes drawn out laterally into a shelf which may have provided attachment for the flexor tibialis internus and probably the ischio- trochantericus. Romer noted (1927a: 248) that the flexor tibialis internus probably arose from the dorsal margin of the ischial rod half-way along its length; since this position corresponds with that of the shelf in H. tucki, the . flexor tibialis internus may have attached here. A short, roughened line on the proximal part of the ischial shaft and pubic peduncle may mark the adductor musculature; this line lies below and anterior to the above ridge. The area below the major ridge on the ischial rod was probably occupied by the obturator internus (pubo-ischio-femoralis externus). The ischial rod of H. tucki does not have an obturator process. However, the only real criterion for classification within the Ornithopoda seems to have been a bipedal form of locomotion; it is a 188 ANNALS OF THE SOUTH AFRICAN MUSEUM functional category in which several independent phylogenetic lineages may be included. Pubis The pubis of H. tucki is the first which shows the configuration of the early ornithischian prepubis. The prepubic process is short and deep, 11 mm long from the anterior edge of the pubic peduncle of the ilium, and about 8 mm deep. The postpubis is thin and fragile, as long as the ischial rod, lying parallel to and about 5-8 mm below it. The postpubis seems devoid of muscle markings except on the internal surface opposite the shelf-like process on the ischium. Here a small, longitudinal, roughened area may indicate the obturator internus muscle. The obturator foramen ventral to the acetabulum is closed posteriorly by the pubis itself, not by the ischium. A tubercle lies above the obturator foramen on the ventral margin of the acetabulum; no corresponding tubercle exists in the alligator but in Struthio (Gadow, in Gregory & Camp 1918, pl. 46) a tubercle in this position marks the attachment of the mm accessorii. While the postpubis has few signs of muscle attachments, the prepubis has several prominent tubercles and ridges which indicate its importance for muscle attachment (Figs 16-18A). A small, smooth ridge parallels the dorsal margin of the prepubis and terminates at the anterosuperior edge in a small but distinct tubercle. Another tubercle below the former lies on the anterior edge of the prepubis surrounded dorsally and ventrally by small damaged and pitted areas. These two tubercles have rounded, finished edges showing that they were not part of a single continuous ridge on the anterior margin of the prepubis. The anteroventral corner bears another, smaller, tubercle; its original extent is indeterminate because this corner is also slightly pitted and damaged. A short ridge runs along the ventral margin of the prepubis but ends below the obturator foramen and does not continue on to the postpubis. The question of which muscles attached to the prepubis has never been resolved. Romer (1927a) thought only the abdominal musculature would have attached here. Galton (1969) disagreed and, in addition to the abdominal musculature, placed the ambiens, pubo-tibialis and part of the pubo-ischio- femoralis internus (ilio-femoralis internus of birds) on the prepubis. The muscle markings of H. tucki cannot themselves solve this problem, but they show that Romer was wrong, at least in the case of H. tucki, in placing only the abdominal musculature here. Romer rejected the idea of other muscles attaching to the prepubis because the most primitive ornithischians known, the hypsilopho- dontids, had a long, thin prepubis; this would have placed some muscle origins too far forward to have provided a firm support for muscular contractions. These are not problems in H. tucki because the prepubis is short and sturdy. The ambiens and a head of pubo-ischio-femoralis internus could each have origi- nated from one of the tubercles. However, the pubo-tibialis is an unlikely muscle to attach to the prepubis; its absence in both birds and crocodiles certainly means it had a low probability of appearing in ornithischians. More likely, a THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 189 further head of pubo-ischio-femoralis internus or an embryonic derivative of the same muscle mass, the ilio-trochanterici, also originated from the prepubis. The avian model for prepubic musculature may be sounder than realized before. The prepubic process of the earliest known ornithischians was clearly short and stout, not long and thin. This is precisely the shape of the pectineal process in the embryo chick. Only during later embryonic development does the pectineal process become relatively and absolutely small. This occurs as the pectineal process stops growing and becomes incorporated into the other pelvic cartilages during their expansion (Johnson 1883). But originally, the pectineal process of the chick has the same relative size, shape and position as the prepubis of H. tucki (though the pectineal process is an iliac, not a pubic, derivative in birds). Because of this similarity the musculature of the avian pre-acetabular area could be used as a model for the ornithischian prepubic musculature. Femur (Figs 1, 3-4, 18B, 19) Both femora are complete but each has been damaged. The left is fractured somewhat proximal to the condyles; the proximal portion is on the main matrix block while the distal end is on a separate block with the tibia—fibula. The fourth trochanter is on a third block, that containing the first group of caudal vertebrae. The right femur is in articulation with tibia—fibula and pelvis, but is only partially exposed from the matrix. The right fourth trochanter is still buried in matrix except for its lateral edge. The proximal tips of the right greater and lesser trochanters have not been preserved. The femur of H. tucki differs from that of ornithopods in that the lesser trochanter is not separated by a cleft from the femoral shaft. Instead, the lesser trochanter is a protuberant crest at the anterolateral femoral margin (Figs 16—18B). It begins just below the level of the femoral head and is about 20 mm long. The anteromedial surface is smooth and continuous with preaxial surface. The lateral surface, however, is extremely rough and irregular. Romer (1927a: - 256) remarked that the lesser trochanter was independent of the greater in the more primitive ornithischians, since only some stegosaurs and the ankylosaurs did not have independent lesser trochanters. But this is not a question simply of primitiveness. As with the obturator process, both conditions of the lesser trochanter appear in the Upper Triassic of South Africa: H. tucki with the lesser trochanter joined to the femoral shaft, Fabrosaurus with a cleft between lesser trochanter and shaft. The two configurations may thus be independent of each other, representing two different ornithischian lineages. Neither does the greater trochanter of H. tucki correspond to Romer’s (1927a: 254) description of the primitive form. Ordinarily, a depression separates the femoral head from the greater trochanter. But in H. tucki both head and greater trochanter are on the same level; in fact, the greater trochanter is distinguished from the lateral femoral surface only by a low, horizontal ridge and an uplifted area above this for tendinous attachment. The 4th trochanter is a pendant, rod-like process, 14 mm long; it makes an 190 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 19. H. tucki. Stereophotograph of matrix block containing left distal femur and tibiotarsus. Note transversely oblique proximal tibial surface. Scale = 5 cm. approximate 45° angle with the shaft. Two parallel ridges run along the lateral surface of the trochanter, creating a shallow sulcus between them; the sulcus presumably marks the attachment of coccygeo(caudi)-femoralis brevis. The medial surface of the femur is not visible at the trochanter so the insertion of caudi-femoralis longus cannot be checked. Distally, the lateral femoral condyle is the smaller, about 18 mm long, the inner condyle much larger, 24 mm long (Figs 1, 19). Anteriorly and posteriorly, the condyles are not separated by an intercondylar groove. While Fabrosaurus does not have an anterior groove it does have a posterior intercondylar groove. Presumably, the posterior groove is absent in H. tucki because the outer condyle is so poorly developed posteriorly. The transverse axis of the distal femoral articular surface is oblique, that is, the transverse axis is inclined, the lateral edge lower than the medial. However, the superior articular surface of the tibia—fibula is horizontal. Consequently, the femur must be abducted relative to the pelvis to keep the articulation with the tibia—fibula horizontal. A perfectly parasagittal orientation of the femur, often depicted in reconstructions of bipedal dinosaurs, was impossible in H. tucki. Furthermore, the femoral head must have rested against the articular surface of the antitrochanter-like buttress of the ilium to support the body weight. Thus, the femur, in addition to being abducted, would be protracted, so the long axis of the femur passed through the antitrochanter-like articular area. The femur could not have been held vertically for several reasons: firstly, the interace- THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 191 tabular width of H. tucki was not great, consequently, vertically oriented femora would have impinged on the abdomen; secondly, the inferior femoral articular surface is oblique relative to the long axis of the shaft, so a horizontal, stable articular surface could be obtained only with the femur abducted; thirdly, even if a wedge-shaped articular cartilage intervened between tibia and femur to produce a vertical limb, loading would produce a sliding of the femur relative to the tibia and thus an unstable knee joint. Thus, the only possible position of the femur in H. tucki is abducted and protracted. Tibia and fibula (Figs 1, 3-4, 19-20A) Both tibiae are complete, but the left is fractured so that the fibula and calcaneum are displaced ventrally. The cortical bone is eroded in many places, particularly on the distal surface of the right tibia. Structurally, the tibia—fibula of H. tucki is a tibiotarsus: the astragalus and calcaneum are completely fused with the tibia and fibula and the fibula is fused distally with the tibia. No sutures remain distally between the tibia—fibula and the proximal tarsals. Fusion occurred in stegosaurs and certaopsians but not in the known ornithopods. Proximally, the inner tibial condyle is lateromedially compressed, only 7 mm wide, but anteroposteriorly expanded, 31 mm long. The great length is partially due to the cnemial crest on the medial side of the tibial head. The outer condyle is small, 12 mm long. It sits on a strong lateral buttress (hidden by the displaced fibula in Figs 1, 19-20; see Fig. 4) which extends 40 mm down the tibial shaft. This creates a deep sulcus between it and the cnemial crest medially. No intercondylar groove is present anteriorly, but a small cleft is found posteriorly. Thulborn (1972: 46) noted a smaller, accessory condyle anterior to the outer in Fabrosaurus but H. tucki does not have a similar accessory condyle. The head of the fibula is longer than the outer tibial condyle against which it rests. The fibular shaft narrows progressively and is fused with the tibia; it terminates in a blunt end, about 3 mm in diameter, immediately above the fused calcaneum. The distal tibiotarsal surface permitted only flexion-extension of the tarso- metatarsus. The joint is a pulley, anteroposteriorly rounded, with lateral and medial ridges which prevented long axis rotation. The depressions for the collateral ligaments are well developed (laterally; the medial astragaler surface is damaged). The joint surface rises 12 mm on to the anterior tibiotarsal surface and about 10 mm on to the posterior. The tibiotarsus of H. tuckiis remarkably convergent with the tibiotarsus of birds and quite unlike that of other ornithischians. Muscle markings on the tibia—fibula are obscured by fracturing and the flaking of the outer cortical bone, but some surface features remain (Fig. 19). The cnemial crest received the common extensor tendon. A small vertical rugose area on the posterolateral surface of the fibular head may mark the attachment ANNALS OF THE SOUTH AFRICAN MUSEUM 192 Lea | ee, re, area Be ( ae \ \ y | Fig. 20. H. tucki. A. Left tibiotarsus, lateral view. B. Right pes and distal tarsals, dorsal view. Scale = 5 cm. THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 193 of the ilio-fibularis. Three other ridges can be made out clearly: a vertical ridge on the posterior surface of the fibula, 16 mm below the fibular head and about 12 mm long; a smaller ridge on the lateral fibular surface just below and anterior to the former; another vertical ridge continuous with the second on the edge of the lateral buttress described above. The first may be associated with a head of the digital flexors and the others with the peroneal muscles. Pes (Figs 1, 10C, 20B, 21-22) The right pes is complete and well preserved save for a transverse fracture proximally and some displacement near the metatarsal bases. The left pes is virtually incomplete except for the phalanges and part of the distal tarsals. The following description, therefore, depends only on the right pes. The distal tarsals of H. tucki differ considerably from those preserved in other ornithischians. In the latter, they are usually flat, disc-shaped bones not ankylosed with the metatarsals. Fabrosaurus is similar to other ornithischians in this respect and in no way resembles H. tucki. The three distal tarsals of H. tucki are fused with each other and with the four metatarsals; the latter are also fused with each other: So just as the tibia—fibula is a structural tibiotarsus, the foot is a structural tarsometatarsus. The fifth digit is only a small splint of bone on the proximoventral surface of digit 4. Though fused, the individual tarsals are still distinguishable: distal tarsal 1 caps metatarsals 1 and 2, distal tarsal 2 caps metatarsal 3, and distal tarsal 3 caps metatarsal 4. A ridge rises along the medial and posterior margins of distal tarsal 1; the anterior margin is rounded and the articular surface permitted e = Fig. 21. H. tucki. Stereophotograph of right pes and distal tarsals, dorsal view. Scale = 5 cm. 194 ANNALS OF THE SOUTH AFRICAN MUSEUM extension of the tarsometatarsus on the tibiotarsus. Posteriorly, the medial edge of the tarsus bears a vertical process or flange (Fig. 10C). The tendon of gastrocnemius may have passed over the tarsus lateral to this tubercle before expanding into the plantar aponeurosis. Distal tarsal 2 also has a rounded anterior margin but a ridge does not appear along the posterior margin; the articular surface is slightly concave. Distal tarsal 3 has a strong ridge round its free margin. The articular surface is a shallow, elliptical depression, antero- posteriorly oriented; this accepts the articular ridge of the calcaneum. The anterior margin of distal tarsal 3 is strongly lipped and overhangs slightly the body of the tarsal; the articular surface thus does not extend on to the dorsal tarsal surface. A foramen pierces the third distal tarsal in the middle of its dorsal (anterior) surface. The rest of the pes generally resembles a small ornithopod such as Hypsilophodon. The head of metatarsal 1 faces medially so the first digit lies in an abducted (i.e. relative to the axis of the pes) position. The first digit was much too short to have been weight-bearing as the tip of the ungual reached only the middle of phalanx 1, digit 2. The distal articular surfaces of the three weight-bearing metatarsals produce a bird-like stance of the digits. On metatarsals 2 and 4, the surfaces are transversely oblique so both digits were abducted relative to digit 3; on meta- tarsal 3, the articular surface is horizontal so digit 3 was aligned along the pedal axis. The distal ends of metatarsals 3 and 4 bear a small dorsal pit just above the articular surface; neither metatarsal 1 nor 2 has such a pit. Presumably, digits 3 and 4, being much longer than digit 2, required greater extension to shorten their effective length and allow digit 2 to reach the ground. The dorsal pits on metatarsal 3 and 4 are indicative of this greater extension. The phalanges of H. tucki are characterized by deep interphalangeal pulley Fig. 22. H. tucki. Stereophotograph: detail of right distal tarsals and proximal portion of metatarsals, dorsal view. Scale = 5 cm. THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 195 Fig. 23. Reconstruction of Heterodontosaurus tucki. 196 ANNALS OF THE SOUTH AFRICAN MUSEUM joints. Proximally, a deep intercondylar groove separates the steep-walled condylar articular surfaces; a cross-section of the joint surface here would show a narrow W. A deep pit lies on the dorsum of the phalanx proximal to the articular surface. The condyles are quite extensive; in side view, they describe an arc of somewhat more than 180°. To fit the deep intercondylar groove, the base of the succeeding phalanx has a steep, V-shaped articular surface. The base also has a prominent dorsal keel which reaches the extensor pit of the opposing phalanx. The collateral ligaments were apparently quite robust: a deep pit for the ligament on the lower half of the condyle lies opposite a tubercle for its attachment near the ventral margin of the apposed phalanx. The terminal joint between ungual and penultimate phalanx differs from the interphalangeal joints only in that the pit for the collateral ligament lies near the dorsal margin of the penultimate phalanx. The ungual phalanges are trans- versely compressed claws but are not greatly recurved. They have a groove along their length both laterally and medially. Comparisons The ilium of H. tucki resembles that of Hypsilophodon most closely amongst ornithopods (also that of Protiguanodon and Psittacosaurus should they be classed as ornithopods). The ilia of Camptosaurus and Thescelosaurus have a much deeper brevis shelf. Hadrosaur ilia have the dinosaurian antitrochanter which the ilium of H. tucki does not have. The ilium of Jguanodon differs in having a slightly reflected dorsal supra-acetabular margin and deeper post-acetabular blade. Amongst the non-ornithopods, the ilium of Protoceratops is remarkably similar to that of H. tucki. | The prepubis of H. tucki is similar to that of Fabrosaurus and Scelidosaurus. The pubis of ornithopods generally has a much longer prepubic process, some- times associated with a greatly reduced postpubic process (hadrosaurs). A short prepubic process (but with a short postpubic process unlike that of H. tucki) can be found in the pelvis of Psittacosaurus, Protiguanodon and Protoceratops. The ischium of H. tucki has few parallels within the ornithopods; it is similar only to Psittacosaurus, Protiguanodon and the pachycephalosaurs, all of which are doubtful ornithopods. Conversely, all non-ornithopodous ornithi- schians resemble H. tucki in the lack of an obturator process. The femur of H. tucki, with the lesser trochanter not demarcated from the femoral shaft, cannot be matched in any ornithopod. The femora of some ankylosaurs, some stegosaurs (S. ungulatus (Gilmore 1914)) and Triceratops are similar (the trochanters are separated by a small notch in Protoceratops). In this respect, H. tucki is more similar to theropod dinosaurs. In general morphology, the tibia of H. tucki resembles that of Hypsilophodon and Fabrosaurus. However, the fusion of tibia—fibula and astragalus—calcaneum into a functional tibiotarsus does not seem to have an equivalent within the ornithischians. The pes of H. tucki is less distinctive than the tibiotarsus and similar to that f "SSS a ar a in THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 197 of many light ornithopods: Fabrosaurus, Hypsilophodon, and Dryosaurus (Galton 1977). The pes also resembles that of Thescelosaurus and Camptosaurus in the closely applied proximal metatarsals; but these heavy bipedal forms have proximodistally compressed phalanges unlike those of H. tucki. Outside the ornithopods, the pes most closely resembles that of Microceratops and to a lesser extent that of Protoceratops. ‘DISCUSSION Diagnosis Order ORNITHISCHIA Suborder incertae subordinis Family Heterodontosauridae Romer, 1966; Kuhn, 1966 The following diagnosis of Heterodontosaurus tucki Crompton & Charig, 1962, is based on the postcranial skeleton of SAM-—K1332: short presacral column of 21 vertebrae (9 cervical + 12 dorsal), sacrum with 6 fused vertebrae; at least 34 caudal vertebrae; ossified tendons present in dorsal and sacral region, but absent from caudal. Scapula elongate relative to trunk; humerus with large deltopectoral crest and large entepicondyle; humerus lacks posterior intercondylar fossa; ulna with olecranon process; nine carpal bones, one of which occupies a position analogous to the os centrale; digits 1-3 parallel, digits 4 and 5 reduced and abducted. Ilium with articular boss analogous to avian antitrochanter; prepubis short but deep, postpubis as long as ischium; ischium without obturator process; greater and lesser femoral trochanters not separated by cleft; transverse axis of distal femoral surface obliquely oriented; fibula reduced and fused with tibia; astragalus—calcaneum fused with each other and to the tibia—fibula; three distal tarsals present, all fused to each other and metatarsal heads; metatarsals 1-4 fused. ~ Morphological interpretation The functional importance of several morphological features may not be entirely clear from the foregoing descriptions. In particular, the orientation of the vertebral column and the posture of the forelimb and hind limb should be discussed. In the vertebral column, a strong flexion is induced in the cervical region by the shape of the centra (Fig. 5A). The posterior cervicals are especially . important in this: their trapezoidal outline combined with their shortness _ creates an abrupt flexion at the transition from dorsal to cervical region. Secondly, ossified tendons are present only in the dorsal region of the vertebral column; the back was therefore a rigid structure, the ossified tendons presumably acting to resist flexion and support the trunk during bipedal progression. Thirdly, the tail of H. tucki was not strengthened by ossified tendons. This is probably not a vagary of preservation since the caudals are by far the best 198 ANNALS OF THE SOUTH AFRICAN MUSEUM preserved vertebrae. Consequently, the tail was not a rigid structure as inferred for some other ornithischians such as Hypsilophodon and hadrosaurs, but was flexible and mobile. The hind limb of H. tucki cannot be articulated at a right angle to the long axis of the illum. The presence of the avian-like antitrochanter required the femoral long axis to lie at 45° or less to the iliac long axis. In normal resting position, the femur would thus be protracted. There is a further similarity to birds since the femur must also be abducted relative to the midsagittal plane. This clearly follows from the oblique orientation of the inferior femoral surface: only in abduction of the femur would the femorotibial joint of H. tucki be horizontal and thus stable. This is completely analogous to the structure of birds in which the inferior femoral articular surface is also oblique. It might be argued that since known dinosaur trackways are narrow, the hind limb could not have been abducted in any dinosaur. However, it must be remembered that in birds, in spite of femoral abduction, rotation about the knee joint of the supporting limb brings the body weight over this limb and close to the centre of gravity, creating a narrow trackway. A similar system is to be expected in H. tucki because of the very close structural similarity to the avian hind limb and joint surfaces. Many features of the forelimb can be interpreted as quadrupedal adaptations; however, the evidence is not unequivocal. As already noted, the ulna of H. tucki has a relatively large olecranon process. Such a feature is usually considered a quadrupedal adaptation since it increases the lever arm of | the ulna; indeed, the forelimb itself is relatively large compared to that of Fabrosaurus and Hypsilophodon. The large entepicondyle of the humerus indicates powerful forearm flexor and rotational musculature, a further sign of quadrupedal capabilities. In addition, presence of a large entepicondyle has been interpreted by Bakker (1971) as a key feature in the sprawling gait of primitive tetrapods. Finally, the large flexor tubercle of the unguals may be associated with a powerful propulsive stroke during push-off. Thus the forelimb of H. tucki quite clearly had the structure requisite for quadrupedal locomotion. However, each trait may also be interpreted as a feature of a powerful, grasping manus. In the case of a manipulative hand, the skeletal features associated with forelimb flexion would also be emphasized. Only the olecranon process does not seem to fit this interpretation; yet, the coelurosaur Syntarsus (Raath 1969) has a grasping hand and also an elongate olecranon. Furthermore, it is quite possible that the forelimb and hand of H. tucki performed a dual function, in locomotion and feeding, as in living sciurids. Unfortunately, the orientation of the forelimb in H. tucki cannot be precisely determined. The humeral head is smooth and rather amorphous and the glenoid is somewhat saddle-shaped. Haines (1952) has pointed out that in living reptiles the ligaments surrounding the shoulder joint, not the joint surfaces themselves, are primarily responsible for determining the range of humeral movements. The long axis of the scapula would have been more or less THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 199 parallel to the vertebral column, as found in articulated skeletons of hadrosaurs ; consequently, the long axis of the glenoid would have been horizontal. The lack of a posterior intercondylar groove on the humerus is most unusual since that would have severely restricted forelimb extension. Furthermore, the extent of forelimb rotation which would have been possible is indeterminate: the radial condyle of the humerus is not rounded but ellipsoid or ridge-like; this would seem to restrict rotation, yet a similar condylar structure in Sphenodon permits about 45° pronation/supination (Haines 1946). The orientation of the humeral condyles is similar to those of Sphenodon and Varanus; thus, even with 45° rotation the humerus of H. tucki would have to have been abducted somewhat (approaching a semi-sprawling position) for the palmar surface of the manus to contact the ground. Thus, a fully erect gait (as described by Bakker 1971) is questionable for the forelimb of H. tucki. Heterodontosaurus and Fabrosaurus The extent of the difference between these two genera must be appreciated for an understanding of ornithischian phylogeny (see Table 1). Heterodonto- saurus and Fabrosaurus represent a schism in ornithischian structure which cannot be contained within a single family. A whole series of anatomical characters separates these genera, many of which reflect the differences in ornithopods and non-ornithopods of the Jurassic—Cretaceous. _ These distinctions are important because Galton (1978: 154) has contended that fabrosaurids were either directly or indirectly ancestral to heterodonto- TABLE 1 Differences in the postcranial skeleton of H. tucki and Fabrosaurus (Thulborn 1972) Feature forelimb and hand humerus humerus ulna ilium ilium ischium trochanters fourth trochanter ~femoral condyles - tibia—fibula proximal tarsals distal tarsals proximal metatarsals ossified tendons H. tucki relatively and absolutely larger posterior intercondylar groove absent entepicondyle large olecranon process present posterior process shallow presence of ‘avian antitro- chanter’ obturator process absent greater and lesser continuous rod-shaped no intercondylar grooves fused fused to each other and tibia— fibula fused to each other and meta- tarsal heads square in X-section absent in caudal region Fabrosaurus relatively and absolutely smaller posterior intercondylar groove present entepicondyle absent no olecranon process posterior process deep dorsal acetabular margin roofed obturator process present greater and lesser divided by cleft triangular, blade-like posterior intercondylar groove only separate separate separate lateromedially compressed present in caudal region 200 ANNALS OF THE SOUTH AFRICAN MUSEUM saurids. However, he offered no anatomical comparisons to substantiate how such a derivation could have occurred. When the distinctions listed in Table 1 are taken into account, it would be almost impossible to derive H. tucki from a fabrosaurid: the smaller forelimb skeleton, the acetabular morphology, the obturator process, the configuration of the femoral trochanters make Fabro- saurus a most unlikely ancestor for H. tucki. However, a more primitive heterodontosaurid, without the specializations of H. tucki (e.g. without the caniniform teeth, without the jugal boss, without the functional tibiotarsus and tarsometatarsus) could be ancestral to the fabrosaurids or hypsilophodonts. It also follows from the above distinctions that H. tuckiis not a hypsilopho- dontid as Thulborn (1970a, 1970, 1971a, 19716, 1972) has previously contended. The absence of the obturator process is alone sufficient to distinguish unequi- vocally the two forms. The taxonomic significance for ornithischians of differences in pelvic structure cannot be ignored. Yet, Thulborn and Galton have done precisely this, the former in trying to make H. tucki a hypsilopho- dontid, the latter in trying to derive H. tucki from a fabrosaurid. The importance of H. tucki The distinctions between Fabrosaurus and H. tucki and the improbability that the latter evolved from a fabrosaurid have important implications for ornithischian evolution. Firstly, since H. tucki is a specialized ornithischian, particularly in comparison with Fabrosaurus, its specialized nature implies a derivation from a more conservative and stratigraphically older ornithischian. H. tucki thus implies the existence of an heterodontosaurid-like radiation of which it is a product: that is, a radiation of non-fabrosaurid, non-hypsilopho- dontid ornithischians. Thus, the distinctions between the ornithopods and the non-ornithopods of the Cretaceous appear in incipient form in the Triassic. | Secondly, the existence of heterodontosaurids discredits the notion of a ‘hypsilophodont plexus’ (Thulborn 19715); that is, that hypsilophodonts were ancestral to all other ornithischians, including the major groups of Jurassic— Cretaceous non-ornithopods. Galton recently advanced a similar hypothesis (1978), that the fabrosaurids were ancestral to all other ornithischian dinosaurs. The existence of heterodontosaurids discredits both hypotheses, primarily because the heterodontosaurids themselves cannot be derived from either hypsilophodonts or fabrosaurids. In addition, because heterodontosaurids lack an obturator process but have quadrupedal as well as bipedal capabilities, they are better structural precursors for the later non-ornithopod groups than are fabrosaurids/hypsilophodonts. Thus, though H. tucki itself could not be ancestral to a later non-ornithopod such as Microceratops, a heterodontosaurid or a derivative of the heterodontosaurid radiation (without the derived specializa- tions of H. tucki) is a much more likely ancestor than a fabrosaurid or hypsilo- phodont. Consequently, the notion of a hypsilophodont plexus or fabrosaurid basal stock should be restricted to the phylogeny of those ornithischians with an obturator process. THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 201 The significance of the obturator process The existence of a fundamental evolutionary dichotomy within the ornithischia, based on the presence or absence of the ischial obturator process would be controverted, at least in part, if it could be shown that some non- ornithopod had an ornithopod (sensu stricto) ancestry. The case in point is the Ceratopsia. It has been speculated that Psittacosaurus and Protiguanodon represent the ancestral group (Romer 1966) of the ceratopsians or at least a related group (Maryanska & Osmdlska 1975). However, by the definitions proposed in this paper, Psittacosaurus and Protiguanodon are clearly not ornithopods and their assumed ancestry to ceratopsians proves nothing about a supposed ornithopod ancestry of ceratopsians. Maryariska & Osmdlska (1975) even place the psittacosaurids within the suborder Ceratopsia which makes the question of ‘ornithopod’ ancestry moot. This, however, merely throws the question back to the origin of psittaco- saurs. Thulborn (19715) stated that psittacosaurs resemble hypsilophodonts in postcranial anatomy but gave no specifics. Unfortunately, Thulborn included both heterodontosaurids and hypsilophodonts (sensu stricto) in his category Hypsilophodontidae. In postcranial structure, H. tucki clearly resembles the psittacosaurs as much as the hypsilophodonts (sensu stricto) do. For instance, and most importantly, the psittacosaurs and H. tucki lack the obturator process which all hypsilophodonts have; the precaudal vertebral count (27) is reduced relative to Hypsilophodon (30) but is the same as H. tucki; the ossified tendons of psittacosaurs extend only from the anterior dorsal to the anterior caudal region, not through the entire caudal region as in Hypsilophodon; and the tibiotarsus is closely joined, though not united, in psittacosaurs, but completely free in hypsilophodonts and fused in H. tucki. At this simplistic level of analysis, psittacosaurs are no more similar to hypsilophodonts than to H. tucki. In a further attempt to derive ceratopsians from hypsilophodonts, Thulborn derived the protoceratopsids from hypsilophodonts on the basis of certain “primitive characters, the only one of which he mentioned was the presence of premaxillary teeth. However, such similarities mean nothing since they are symplesiomorphies. Other characters Thulborn used to join ceratopsians with hypsilophodonts also fall into this category, i.e. a nasal-maxilla contact. In fact, no sound evidence exists to support a hypsilophodont (sensu stricto) ancestry of ceratopsians in preference to a non-hypsilophodont ancestry. The existence of a fundamental evolutionary dichotomy within the Ornithischia would also be controverted if it could be shown that the absence of -an obturator process were due to secondary loss. However, the hypothesis of _ secondary loss of the obturator process in non-ornithopod Jurassic—Cretaceous ornithischians is but an assumption; the fact is, no evidence exists that these forms have ever possessed such a process. But the hypothesis will be examined anyway; it will be shown that secondary loss is less plausible than a hypothesis of original absence for several reasons. Of the five major divisions within the Ornithischia, only the ornithopods 202 ANNALS OF THE SOUTH AFRICAN MUSEUM possessed an obturator process; according to the loss hypothesis, therefore, stegosaurs, ceratopsians, ankylosaurs and pachycephalosaurs first possessed and then lost this trait. For a structure which must have had some adaptive significance to develop in the first place, this is a poor record of adaptive value. It could be assumed that the process was lost because it was not necessary for quadrupedal forms but was for bipedal forms. However, if the obturator process were important for bipedal progression, why did it first appear and then disappear in the lineages represented by pachycephalosaurs, psittacosaurs, H. tucki and Microceratops which were all bipedally adapted ornithischians? The most plausible, logical and evolutionarily sound answer is that the obturator process never existed in these forms. According to the loss hypothesis, the only reason the obturator process would have disappeared was that an animal had taken up a quadrupedal mode of locomotion; the bipedal ancestors of these animals should, therefore, have possessed an obturator process. However, in the only test case available, the Ceratopsia, all the related bipedal forms (psittaco- saurs and Microceratops) do not have an obturator process. This is certainly contradictory; it is more plausible to assume that the obturator process never existed in these forms than to assume it was lost secondarily. The obturator process is clearly related to bipedalism within the Ornithischia since no known quadrupedal form possesses this process. While some bipeds did not have an obturator process, the most successful bipedal ornithischians (in terms of diversity) did have this process. In addition, the only large bipedal ornithischians (e.g. Jzwanodon, Camptosaurus and hadrosaurs) all possessed the process. Bipedal ornithischians without the obturator process, H. tucki, Microceratops, psittacosaurs, and the pachycephalosaurs of Mongolia, are all relatively small dinosaurs. If the obturator process were really functionally important in an efficient bipedal gait in larger animals, then, as descendants of the small bipedal forms increased in size, they assumed a quadrupedal gait. This is presumably what occurred during the evolution of ceratopsians; thus, the lack of an obturator process was not due to secondary loss. H. tucki and ornithischian classification While the ornithischian status of H. tucki cannot be questioned, its sub- ordinal classification is problematical. As defined and used, the Ornithopoda are bipedal ornithischians (Romer 1956: 627-628). The inadequacy of placing all bipedal ornithischians within the Ornithopoda has become apparent recently as more varied ornithischian types are discovered and described. For example, both the pachycephalosaurs and Microceratops differ greatly from typical ornithopods such as Hypsilophodon yet they are certainly bipedal. Should these forms be included in the Ornithopoda, the meaning of this category in terms of representing ornithischian evolution would be almost nil. H. tucki presents the same classificatory difficuity as pachycephalosaurs, Microceratops, Protiguanodon and Psittacosaurus. Since the ancestors of the quadrupedal non-ornithopods were very likely bipeds, as Microceratops THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 203 indicates for the ceratopsids, then bipedalism is not limited to any one phylo- genetic lineage in the Ornithischia but is distributed throughout the various phyletic lines. Consequently, to classify all bipedally adapted ornithischians as ornithopods only confuses the phylogeny of the Ornithischia by creating a paraphyletic group. Since H. tucki is phylogenetically divergent from orni- thopods such as hypsilophodonts and iguanodonts, then it makes little sense to classify H. tucki as an ornithopod. Furthermore, since H. tucki could not have been derived from a fabrosaurid, it is even less desirable to classify it as an ornithopod. However, the real problem here is not H. tucki, but rather the definition of the Ornithopoda and this should be dealt with first. The Ornithopoda could be defined as only those ornithischians which possess an obturator process on the ischium. This basic dichotomy in the ornithischians should finally be recognized, particularly since pelvic structure has tremendous taxonomic value for both groups of dinosaurs. The Ornithopoda would thus be defined on the basis of a probable derived character and approximate a natural group much more than under the other definition. The obturator process is usually considered a primitive structure for the ornithischians, while its absence has usually been attributed to loss. However, the converse seems more likely: the obturator process is a new, derived structure, not present in other archosaurs; conse- quently, the presence of the process must be explained, not its absence. The Ornithopoda thus defined could be placed in two infraorders, one comprising a lineage represented by hypsilophodonts and related forms such as fabrosaurids, the other comprising the lineage represented by iguanodonts and hadrosaurs. The non-ornithopods such as stegosaurs and ankylosaurs could be included in a single suborder as separate infraorders, but this would define the suborder on the basis of a symplesiomorphy (lack of an obturator process) and would not be equivalent to the suborder Ornithopoda. It would be preferable to leave these groups as separate suborders since these can be defined on the basis of - derived characters and are thus equivalent to the suborder Ornithopoda. As aconsequence of this redefinition, H. tuckiis certainly not an ornithopod, yet it does not clearly fit into any other suborder. This should not cause surprise since the cranial and postcranial material for the heterodontosaurids and Triassic ornithischians is still somewhat limited. The family has been known only for a short time; until further material is found which defines the extent of the heterodontosaurid radiation, the family should simply be ‘incertae subordinis’. The situation is like that of pachycephalosaurs which were too - poorly known to classify for a long time. SUMMARY H. tucki was a very small Late Triassic ornithischian dinosaur about 1 m long. The postcranial skeleton combines an elongate hind limb adapted to bipedal locomotion and a moderate but not reduced forelimb adapted to 204 ANNALS OF THE SOUTH AFRICAN MUSEUM quadrupedal locomotion and/or grasping movements of the hand. H. tucki was undoubtedly a facultative biped; a quadrupedal gait was probably used during slow locomotion, perhaps while foraging. The elbow structure suggests that a semi-sprawling attitude of the forelimb was possible. The pelvic structure indicates that the femur was both abducted and protracted, creating a stance similar to, but not the same as, that of birds. The hind limb proportions of H. tucki show an elongated tibia and meta- tarsus relative to the femur. This is usually interpreted as a cursorial adaptation (e.g. Galton 1974), though no sound evidential basis exists for this inference. Certainly, to infer cursorial habits in a bipedal reptile from the hind limb proportions of living quadrupedal mammals is questionable. Different taxa have different base levels from which cursorial limb proportions develop; thus, comparisons across groups may mean very little. H. tucki may or may not have been cursorial: this inference could only be substantiated by comparisons with the hind limb proportions of other heterodontosaurids; these, however, are presently unknown. 3 The classification of H. tucki within the Ornithischia is complicated by the inadequate definition of the Ornithopoda. Once it is accepted that not all bipeds must be classed as ornithopods, then a better definition can be given based on the presence or absence of the obturator process. H. tucki need not be placed in the Ornithopoda; rather, it is taken as a representative of an early non- ornithopod radiation which is presently too poorly known to warrant subordinal or infraordinal distinction. However, if H. tucki is representative of other heterodontosaurids, then they are more likely structural and phyletic precursors — to at least some non-ornithopods than are the fabrosaurids. H. tucki seems to represent a basic cleavage in the Ornithischia. The only two well-known Triassic ornithischians fall on either side of this division which mirrors the differences in pelvic structure of Jurassic-Cretaceous ornithopods and non-ornithopods. Though H. tucki probably represents an early non- ornithopod radiation, it does not itself seem to be ancestral to any known later ornithischian. Similarities with primitive ceratopsians are suggestive but difficult to interpret. Nor does H. tucki help in the search for ornithischian origins, because its structure is already typically ornithischian. Its features do not point to any special group in the thecodonts. This implies a considerable but indeterminate independent phylogenetic history for the Ornithischia. ACKNOWLEDGEMENTS I wish to thank Dr T. H. Barry, Director of the South African Museum, Cape Town, for permission to study the material described here. I am very grateful to Dr A. W. Crompton, Director of the Museum of Comparative Zoology, Harvard University, for the use of the material, permission to photo- graph it and for many valuable discussions and comments during the prepara- tion of this manuscript. Dr A. J. Charig, British Museum (Natural History), THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 205 also offered valuable criticisms. The postcranial material described here was prepared by Mr Arnie Lewis (now of the U.S. National Museum, Washington, D.C.), and by Mr Chuck Schaff and Mr Bill Amaral of the Museum of Com- parative Zoology. The skull was prepared at the South African Museum. The drawings were prepared by Ms M. L. Estey at the Museum of Comparative Zoology, Harvard University; the stereophotographs were taken by Mr Alan Coleman, also of the Museum of Comparative Zoology. The photograph of the skull was taken by Mr Neville Eden of the South African Museum. The remaining photographs were prepared by the author. This research was con- ducted while the author was in the Department of Anthropology, Peabody Museum, Harvard University, and later in the Department of Anatomy, Harvard Medical School. The Department of Celi Biology, The University of Texas Health Science Center at Dallas, provided funds for the completion of the manuscript. REFERENCES BAKKER, R. T. 1971. Dinosaur physiology and the origin of mammals. Evolution, Lancaster, Pa. 25: 636-658. BAKKER, R. T. & GALTON, P. M. 1974. Dinosaur monophyly and a new class of vertebrates. Nature, Lond. 248: 168-172. BONAPARTE, J. F. 1976. Pisanosaurus mertii Casamiquela and the origin of the Ornithischia. J. Paleont. 50: 808-820. Broom, R. 1911. On the dinosaurs of the Stormberg, South Africa. Ann. S. Afr. Mus. 7: 291-308. Brown, B. & SCHLAIKJER, E. M. 1940. The structure and relationships of Protoceratops. Ann. N.Y. Acad. Sci. 40: 133-266. CASAMIQUELA, R. M. 1967. Un nuevo dinosaurio ornitisquio tridsica (Pisanosaurus mertii; Ornithopoda) de la Formacion Ischigualasto, Argentina. Ameghiniana 5: 47-64. CHARIG, A. J. & CrompTON, A. W. 1974. The alleged synonymy of Lycorhinus and Hetero- dontosaurus. Ann. S. Afr. Mus. 64: 167-189. Coomss, W. P., Jun. 1978a. The families of the ornithischian dinosaur order Ankylosauria. Palaeontology 21: 143-170. Coomss, W. P., Jun. 1978b. Forelimb muscles of the Ankylosauria (Reptilia, Ornithischia). J. Paleont. 52: 642-657. CroMpPTON, A. W. & CHARIG, A. J. 1962. A new ornithischian from the Upper Triassic of South Africa. Nature, Lond. 196: 1074-1077. GALTON, P. M. 1969. The pelvic musculature of the dinosaur Hypsilophodon (Reptilia: Ornithischia). Postilla 131: 1-64. GALTON, P. M. 1972. Classification and evolution of ornithopod dinosaurs. Nature, Lond. 239: 464-466. GALTON, P. M. 1974. The ornithischian dinosaur Hypsilophodon from the Wealden of the Isle of Wight. Bull. Br. Mus. Nat. Hist. (Geol.) 25(1): 1-152. GALTON, P. M. 1977. The ornithopod dinosaur Dryosaurus and a Laurasia-Gondwanaland connection in the Upper Jurassic. Nature, Lond. 268: 230-232. GALTON, P. M. 1978. Fabrosauridae, the basal family of ornithischian dinosaurs (Reptilia: Ornithopoda). Paldont. Z. 52: 138-159. GILmMoreE, C. W. 1909. Osteology of the Jurassic reptile Camptosaurus, with a revision of the species of the genus, and descriptions of two new species. Proc. U.S. natn. Mus. 36: 197-332. GiLmorE, C. W. 1914. Osteology of the armored dinosauria in the United States National Museum, with special reference to the genus Stegosaurus. Bull. U.S. natn. Mus. 89: 1-143. 206 ANNALS OF THE SOUTH AFRICAN MUSEUM GiILMorE, C. W. 1915. Osteology of Thescelosaurus, an orthopodous dinosaur from the Lance formation of Wyoming. Proc. U.S. natn. Mus. 49: 591-616. GINSBERG, L. 1964. Découverte d’un scélidosaurien (dinosaure ornithischien) dans le Trias supérieur du Basutoland. C. r. hebd. Séanc. Acad. Sci., Paris 258: 2366-2368. Gow, C. E. 1975. A new heterodontosaurid of the Red Beds of South Africa showing clear evidence of tooth replacement. Zool. J. Linn. Soc. 57: 335-339. GREGORY, W. K. & Camp, C. L. 1918. Studies in comparative myology and osteology. No. III. Bull. Am. Mus. nat. Hist. 38: 447-563. HAINES, R. W. 1939. A revision of the extensor muscles of the forearm in tetrapods. J. Anat. 73: 211-233. Haines, R. W. 1946. A revision of the movements of the forearm in tetrapods. J. Anat. 80: 1-11. HAINES, R. W. 1952. The shoulder joint of lizards and the primitive reptilian shoulder mechanism. J. Anat. 86: 412-422. HATCHER, J. B., MARSH, O. C. & LULL, R. S. 1907. The Ceratopsia. Monogr. U.S. geol. Surv. 49: 1-300. HAuGHTON, S. H. 1924. The fauna and stratigraphy of the Stormberg Series. Ann. S. Afr. Mus. 12: 323-497. Hoo ey, R. W. 1925. On the skeleton of Iguanodon atherfieldensis sp. nov., from the Wealden shales of Atherfield (Isle of Wight). QO. J] geol. Soc. Lond. 81: 1-60 Hopson, J. A. 1975. On the generic separation of the ornithischian dinosaurs Lycorhinus and Heterodontosaurus from the Stormberg Series (Upper Triassic) of South Africa. S. Afr. J. Sci. 71: 302-305. JOHNSON, A. 1883. On the development of the pelvic girdle and skeleton of the hindlimb in the chick. Q. JI Microsc. Sci. 23: 399-411. Kuan, O. 1966. Die Reptilien, System und Stammesgeschicte. Krailling bei Mitinchen: Oeben. LuLL, R. S. 1933. A revision of the Ceratopsia or horned dinosaurs. Mem. Peabody Mus. Yale 3: 1-135. LULL, R. S. & WRIGHT, N. E. 1942. Hadrosaurian dinosaurs of North America. Spec. Pap. geol. Soc. Am. 40: 1-242. MARSH , 0. C. 1877. Notice of some new vertebrate fossils. Amer. J. Sci. 14: 249-256. MARYANSKA, T. & OSMOLSKA, H. 1974. Pachycephalosauria, a new suborder of ornithischian dinosaurs. Palaeont. pol. 30: 45-102. MaryANnskA, T. & OsmorsKA, H. 1975. Protoceratopsidae (Dinosauria) of ee Palaeont. pol. 33: 133-181. OsBORN, H. F. 1923. Two Lower Cretaceous dinosaurs of Mongolia. Am. Mus. Novit. 95: 1-10. OsBorn, H. F. 1924. Pssitacosaurus and Protiguanodon: two Lower Cretaceous iguanodonts from Mongolia. Am. Mus. Novit. 127: 1-16. OwEN, R. 1861. On the fossil Reptilia of the Wealden and Purbeck formations: Part 5. Palaeontogr. Soc. (Monogr.): 31-39. RaatH, M. A. 1969. A new coelurosaurian dinosaur from the Forest Sandstone of Rhodesia. Arnoldia 4 (28): 1-25. Romer, A. S. 1927a. The pelvic musculature of ornithischian dinosaurs. Acta Zool., Stockh. 8 225-275. Romer, A. S. 1927b. The development of the thigh musculature of the chick. J. Morph. 43: 347-385. Romer, A. S. 1942. The development of tetrapod limb musculature—the thigh of Lacerta. J. Morph. 71: 251-297. Romer, A. S. 1956. Ostenlogy of the Reptiles. Chicago: University of Chicago Press. Romer, A. S. 1966. Vertebrate Paleontology. 3rd ed. Chicago, London: University of Chicago Press. SANTA Luca, A. P., CROMPTON, A. W. & CHARIG, A. J. 1976. A complete skeleton of the Late Triassic ornithischian Heterodontosaurus tucki. Nature, Lond. 264: 324-328. THULBORN, R. A. 1970a. The skull of Fabrosaurus australis, a Triassic ornithischian dinosaur. Palaeontology 13: 414-432. THULBORN, R. A. 1970b. The systematic position of the Triassic ornithischian dinosaur Lycorhinus angustidens. Zool. J. Linn. Soc. 49: 235-245. THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 207 THULBORN, R. A. 197la. Toothwear and jaw action in the Triassic ornithischian dinosaur Fabrosaurus. J. Zool., Lond. 164: 165-179. THULBORN, R. A. 19715. Origins and evolution of ornithischian dinosaurs. Nature, Lond. 234: 75-78. THULBORN, R. A. 1972. The post-cranial skeleton of the Triassic ornithischian dinosaur Fabrosaurus australis. Palaeontology 15: 29-60. THULBORN, R. A. 1974. A new heterodontosaurid dinosaur (Reptilia: Ornithischia) from the Upper Triassic Red Beds of Lesotho. Zool. J. Linn. Soc. 55: 151-175. THULBORN, R. A. 1978. Aestivation among ornithopod dinosaurs of the African Trias. Lethaia 11: 185-198. WALKER, A. D. 1977. Evolution of the pelvis in birds and dinosaurs. Jn: ANDREWS, S. M.., Migs, R. S., & WALKER, A. D. Problems in Vertebrate Evolution: 319-358. London: Academic Press. ABBREVIATIONS ap acromial process ar anterior ridge of cervical centra at avian-like antitrochanter C coracoid Cc calcaneum ca capitulum ch chevron cn cnemial crest cp coronoid process cr calcaneal ridge ct coracoid tubercle d diaphophysis de deltopectoral crest dt _ distal tarsal ee entepicondyle i. vertical flange, ventral surface of metatarsals ft flexor tubercle gt glenoid tubercle hh humeral head if ischial flange: possible attachment of flexor tibialis internus and ischio-trochantericus IL _ilium ir _ iliac ridge: attachment of ilio-tibialis and ilio-fibularis IS = ischium it infra-acetabular tubercle: possible attachment of accessorius met metacarpal tubercle ms neural spine odp odontoid process of axis op olecranon process ot _ ossified tendons p pubis pa parapophysis pd pit on dorsal surface of metacarpal or metatarsal pi _ pisiform pr posterior ridge of cervical centra poz postzygapophysis prz prezygapophysis for atlantal neural arch pt tubercles on prepubic process R radius if ridge above radial condyle of humerus re __radiale tt _— radial tubercle 208 ANNALS OF THE SOUTH AFRICAN MUSEUM scapula sulcus between cnemial crest and lateral buttress of tibial shaft tuberculum fourth trochanter greater trochanter lesser trochanter tuberosity medial to humeral head ulna ulnare ulnar ridge ventral keel ventral ridge delimiting fossa beneath transverse process THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 209 APPENDIX 1 Measurements of the postcranial skeleton of H. tucki (SAM-—K1332), in mm. (NA = not available; the lengths of some vertebral centra were determined from radiographs and these are noted in the table; parentheses indicate approximate measurements.) SCAPULA Max. length Max. proximal width Max. distal width . Min. blade breadth CORACOID Max. length Max. width Min. width HUMERUS Max. length : : Proximal transverse width Max. distal width (transverse) Least shaft diameter : Length: deltopectoral crest . RADIUS Max. length Max. proximal width Max. distal width . Least shaft diameter ULNA Max. length : Max. proximal width Max. distal width . Least shaft diameter Length: base of sigmoid notch-distal end . ILIUM Max. length : ; Length: anterior end to middle of acetabulum Length: posterior end to middle of acetabulum Min. height above acetabular rim ISCHIUM Max. length Max. height Least shaft diameter PUBIS Max. length A Height at anterior end of prepubic process Length: to anterior wall of obturator foramen. Length: from anterior wall of obturator foramen to distal end FEMUR Max. length a Proximal transverse width Distal transverse width Least shaft diameter 2 Max. proximal A-P width . Max. distal A—P width . LEFT 67,7 15,2 including coronoid process) 210 ANNALS OF THE SOUTH AFRICAN MUSEUM FEMUR (cont.) LEFT RIGHT Distal attachment of 4th trochanter to proximal end of femur ~ -. : A f : : . 46,1 NA Length of 4th trochanter : 5 ; 2 : i OBS 14,0 (broken) TIBIA Max. length . . . (144) 145,0 Max. proximal transverse width (without fibula) . NA 11,7 Max. distal transverse width : : ; : hs 9:8 2D Max. proximal A—P width . : ; : ‘ . 30,6 NA Max. distal A—P width . : ; : : ; a Sst 13,9 Least shaft diameter . : ; ’ : ; Br RED. 8,3 MANUS— RIGHT DIGIT I MC Ph 1 UNGUAL R L R L R L Max. length .. mire een ea UNleeD 17,6 16,6 16,5 16,5 18,2 Proximal transverse width Se, Acoma es Lee Ors 9,1 6,6 NA 4,6 NA Distal transverse width Sith ee ee | 16,0 NA 4,8 NA NA NA Length along outer curve . ; : 23 23 DIGIT II MC Ph 1 -— Ph UNGUAL R L R L R Ec R L Max. length . ees. NA 15,6 isa 16,7 NA (18) NA Proximal transverse Width. 3° es 2) °6.0 a3 6,1 NA 5,0 NA 4,0 NA Distal transverse , width . ae OO NA Syl NA 4,2 NA NA NA Length along outer CULVE” © eee (21) NA DIGIT III MC Ph 1 Ph 2 Ph 3 UNGUAL R L R L R L R EL R & Max. length .. . 21,4 22.4 14,1. 13,0 12,1 12,6 “NA GSS Nee ig0 Proximal transverse width . : . 58 5,7. 62 NA 47 NA. “4,05 INAS =o Distal transverse width . 64 NA °5,0 NA ° 4,2’ -_NA -NA© NASSNARINA Length along outer curve NA_ (20) DIGIT IV MC Ph 1 Ph 2 UNGUAL R L R IG R L R L Max. length . . 14:5 (15) 6,8 6,6 4,6 4,6 PIG | NA Proximal transverse width . Pets AGS 5,8 NA 3,3 NA 2,6 1,8 NA Distal transverse WiGthi i.e = "5 sae3h7, 3,8 NA 2,6 NA NA NA NA PES—RIGHT ONLY DIGIT I MT Ph 1 UNGUAL Max. length 5 eis 3 Sl wad ee ONL 173 17,6 Proximal transverse width i 5 . . NA NA NA Distalitransverse width- ~~ * - <2,“ =6:1 4,7 NA Length along outer curve ee ees 20 DIGIT II MT Ph 1 Ph2 UNGUAL Max. length see Nie BOS o 2 + peepee 351) | 19,4 155 20,9 Proximal transverse width ; : ‘ . NA NA NA NA Distal transverse width °° 5) 575° 2) (8) NA NA NA Length along outer curve eee 24 THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI 21] DIGIT III Max. length “ee ae Proximal transverse width Distal transverse width Length along outer curve DIGIT IV Max. length . , ; Proximal transverse width Distal transverse width Length along outer curve Vertebra No. Presacrals Axis MT Ph Ph 2 Ph 3 UNGUAL 67,9 21,8 15,6 14,4 (18) 5,9 (9) 7,8 6,7 (5,5) 8,9 7,4 6,5 57 NA (20) MT Ph 1 Ph 2 Ph 3 Ph 4 UNGUAL 61,4 16,8 i Psee2 10,6 oN 16,0 (6) NA 6,2 5,5 5,0 NA (6) 339 39 5,0 NA NA VERTEBRAL COLUMN Max. length Vertebra No. Max. length of Centrum Sacrals of Centrum 16,2 1 14,2 14,2 D, (13) 15,9 3 NA (15,5) 4 NA 13,4 5 NA 13,0 6 14,2 Set Caudals* NA Al 14,2 (13) A2 NA Be? A3 14,8 13:5 A4 (15) (13) A5 15.5 14,5 A6 15,8 15,2 A7 16,0 15,0 A8 17,0 15 AQ 17,8 14,8 A10 17,8 NA All 18,2 (15) Al2 fragmentary 14,6 * Sequentially numbered on each block of matrix, A and B. Vertebra No. Max. length Caudal of Centrum Bl 16,0 B2 16,0 B3 16,2 B4 16,5 B5 16,4 B6 16,7 B7 16,3 B8 16,3 B9 16,3 B10 tI Bil 16,4 B12 16,0 B13 NA B14 tS55 B15 15,6 B16 fragmentary 6. SYSTEMATIC papers must conform to the International code of zoological nomenclature (particularly Articles 22 and 51). Names of new taxa, combinations, synonyms, etc., when used for the first time, must be followed by the appropriate Latin (not English) abbreviation, e.g. gen. nOv., sp. nov., comb. nov., SyN. Nov., etc. An author’ Ss name when cited must follow the name of the taxon without intervening punctuation and not be abbreviated; if the year is added, a comma must separate author’s name and year. The author’s name ‘(and date, if cited) must be placed in parentheses if a species or subspecies is transferred from its original genus. The name of a subsequent user of a scientific name must be separated from the scientific name by a colon. Synonymy arrangement should be according to chronology of names, i.e. all published scientific names by which the species previously has been designated are listed in chronological order, with all references to that name following in chronological order, e.g.: Family Nuculanidae Nuculana (Lembulus) bicuspidata (Gould, 1845) Figs 14-15A Nucula (Leda) bicuspidata Gould, 1845: 37. Leda plicifera A. Adams, 1856: 50. Laeda bicuspidata Hanley, 1859: 118, pl. 228 (fig. 73). Sowerby, 1871: pl. 2 (fig. 8a—b). Nucula largillierti Philippi, 1861: 87. Leda bicuspidata: Nicklés, 1950: 163, fig. 301; 1955: 110. Barnard, 1964: 234, figs 8-9. a punctuation in the above example: comma separates author’s name and year semicolon separates more than one reference by the same author full stop separates references by different authors figures of plates are enclosed in parentheses to distinguish them from text-figures dash, not comma, separates consecutive numbers Synonymy arrangement according to chronology of bibliographic references, whereby the year is placed in front of each entry, and the synonym repeated in full for each entry, is not acceptable. In describing new species, one specimen must be designated as the holotype; other speci- mens mentioned in the original description are to be designated paratypes; additional material not regarded as paratypes shou!d be listed separately. The complete data (registration number, depository, description of specimen, locality, collector, date) of the holotype and paratypes must be recorded, e.g.: Holotype SAM-—A13535 in the South African Museum, Cape Town. Adult female from mid-tide region, King’s Beach Port Elizabeth (33°51’S 25°39’E), collected by A. Smith, 15 January 1973. Note standard form of writing South African Museum registration numbers and date. 7. SPECIAL HOUSE RULES Capital initial letters (a) The Figures, Maps and Tables of the paper when referred to in the text e.g. *... the Figure depicting C. namacolus ...’; *. .. in C. namacolus (Fig. 10)...’ (b) The prefixes of prefixed surnames in all languages, when used in the text, if not preceded by initials or full names e.g. Du Toit but A.L.du Toit; Von Huene but F. von Huene (c) Scientific names, but not their vernacular derivatives e.g. Therocephalia, but therocephalian Punctuation should be loose, omitting all not strictly necessary Reference te the author should be expressed in the third person Roman numerals should be converted to arabic, except when forming part of the title of a book or article, such as - “Revision of the Crustacea. Part VIII. The Amphipoda.’ Specific name must not stand alone, but be preceded by the generic name or its abbreviation to initial capital letter, provided the same generic name is used consecutively. Name of new genus or species is not to be included in the title: it should be included in the abstract, counter to Recommendation 23 of the Code, to meet the requirements of Biological Abstracts. A. P. SANTA LUCA THE POSTCRANIAL SKELETON OF HETERODONTOSAURUS TUCKI (REPTILIA, ORNITHISCHIA) FROM THE STORMBERG OF SOUTH AFRICA OF THE SOUTH AFRICAN MUSEUM CAPE ‘TOWN INSTRUCTIONS TO AUTHORS 1. MATERIAL should be original and not published elsewhere, in whole or in part. 2. LAYOUT should be as follows: (a) Centred masthead to consist of Title: informative but concise, without abbreviations and not including the names of new genera or species Author’s(s’) name(s) Address(es) of author(s) (institution where work was carried out) Number of illustrations (figures, enumerated maps and tables, in this order) (b) Abstract of not more than 200 words, intelligible to the reader without reference to the text (c) Table of contents giving hierarchy of headings and subheadings (d) Introduction (e) Subject-matter of the paper, divided into sections to correspond with those given in table of contents (f) Summary, if paper is lengthy (g) Acknowiedgements (h) References (i) Abbreviations, where these are numerous 3. MANUSCRIPT, to be submitted in triplicate, should be typewritten and neat, double spaced with 2,5 cm margins all round. First lines of paragraphs should be indented. Tables and a list of legends for illustrations should be typed separately, their positions indicated in the text. All pages should be numbered consecutively. Major headings of the paper are centred capitals; first subheadings are shouldered small capitals; second subheadings are shouldered italics; third subheadings are indented, shouldered italics. Further subdivisions should be avoided, as also enumeration (never roman numerals) of headings and abbreviations. Footnotes should be avoided unless they are short and essential. Only generic and specific names should be underlined to indicate italics; all other marking up should be left to editor and publisher. 4. ILLUSTRATIONS should be reducible to a size not exceeding 12 « 18 cm (19 cm including legend); the reduction or enlargement required should be indicated; originals larger than 35 X 47 cm should not be submitted; photographs should be rectangular in shape and final size. A metric scale should appear with all illustrations, otherwise magnification or reduction should be given in the legend; if the latter, then the final reduction or enlargement should be taken into consideration. All illustrations, whether line drawings or photographs, should be termed figures (plates are not printed; half-tones will appear in their proper place in the text) and numbered in a single series. Items of composite figures should be designated by capital letters; lettering of figures is not set in type and should be in lower-case letters. The number of the figure should be lightly marked in pencil on the back of each illustration. 5. REFERENCES cited in text and synonymies should all be included in the list at the end of the paper, using the Harvard System (ibid., idem, loc. cit., op. cit. are not acceptable): (a) Author’s name and yea of publication given in text, e.g.: ‘Smith (1969) describes . ‘Smith (1969: 36, fig. 16) aerate “As described (Smith 1969a, 19695; ae ti, ee ‘As described (Haughton & Broom 1927). ‘As described (Haughton et al. 1927)...’ Note: no comma separating name and year Pagination indicated by colon, not p. names of joint authors connected by ampersand - et al. in text for more than two joint authors, but names of all authors given in list of references. (b) Full references at the end of the paper, arranged alphabetically by names, chronologically within each name, with suffixes a, b, etc. to the year for more than one paper by the same author in that year, e.g. Smith (1969a, 19695) and not Smith (1969, 1969a). For books give title in italics, edition, volume number, place of publication, publisher. For journal article give title of article, title of journal in italics (abbreviated according to the World list o, scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses, volume number, part number (only if independently paged) in parentheses, pagination (first and last pages of article). Examples (note capitalization and punctuation) BULLOUGH, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan. FISCHER, P.-H. 1948. Données sur la résistance et de le vitalité des mollusques. J. Conch., Paris 88: 100-140. FISCHER, P.-H., DuvAL, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines. Archs Zool. exp. gen. 74: 627-634. Koun, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee region of Ceylon. Ann. Mag. nat. Hist. (13) 2: 309-320. Konn, A. J. 19606. Spawning behaviour, egg oe and larval development in Conus from the Indian Ocean. Bull. Bingham oceanogr. Coll. 17 (4): 1-51 THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. In: SCHULTZE, L. Zoologische und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-Afrika 4: 269-270. Jena: Fischer. Denkschr. med.-naturw. Ges. Jena 16: 269-270. (continued inside back cover) ANNALS OF THE SOUTH AFRICAN MUSEUM ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM Volume 79 ~~ Band April 1980 April Part 8 Deel : SSS CoO ES ID SS Q S Sour wow FOSSIL BOVIDAE (MAMMALIA) FROM LANGEBAANWEG SOUTH AFRICA By LOW. GENTRY. Cape Town Kaapstad The ANNALS OF THE SOUTH AFRICAN MUSEUM are issued in parts at irregular intervals as material becomes available Obtainable from the South African Museum, P.O. Box 61, Cape Town Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM word uitgegee in dele op ongereelde tye na gelang van die beskikbaarheid van stof Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad OUT OF PRINT/UIT DRUK 1, 2(1=3,-5-8); 31-2, 4-5, 8; t—pi)s 5(1=3, 5: 79 6(1, t--p.i.), 7(1—-4), 8, 9(1-2, 7), 10(1-3), 11(1-2, 5, 7, t—p.i.), 15(4-5), 24(2), 27, 31(1-3), 32(5), 33 Copyright enquiries to the South African Museum Kopieregnavrae aan die Suid-Afrikaanse Museum ISBN 0 908407 89 0 Printed in South Africa by In Suid-Afrika gedruk deur The Rustica Press, Pty., Ltd., Die Rustica-pers, Edms., Bpk., Court Road, Wynberg, Cape Courtweg, Wynberg, Kaap FOSSIL BOVIDAE (MAMMALIA) FROM LANGEBAANWEG, . SOUTH AFRICA By A. W. GENTRY British Museum (Natural History), London (With 63 figures and 6 tables) [MS. accepted 4 September 1979] ABSTRACT Fossil Bovidae are described from the Quartzose Sand Member and Pelletal Phosphorite Member of the Varswater Formation in ‘E’ Quarry, Langebaanweg. The new genus Damalacra and the new species Simatherium demissum, Kobus subdolus, Damalacra neanica, D. acalla and Raphicerus paralius are named. The bovids best fit a very late Miocene age of about 6 million years B.P. according to faunal correlations with other sites for some of which radiometric dates are available. Evidence from other mammalian groups present in ‘E’ Quarry may suggest alternative ages. _ The Bovidae of Baard’s Quarry, Langebaanweg, are also discussed. The lower assemblage is younger than the ‘E’ Quarry faunas, perhaps even of Pleistocene age; the upper assemblage is of Middle Pleistocene age or later. CONTENTS PAGE Introduction , ; : : : i et Bovidae from ‘E’ Quarry 4 : , ; SAS Tribe Tragelaphini . ‘ ‘ 3 on 2h Tribe Boselaphini F ; : : Vt be Tribe Bovini : : : ‘ : Pee, Tribe Reduncini . : ; ; : oe a DaT Tribe Alcelaphini : : ; : poe! Tribe Neotragini : : : 5 299 Tribe Antilopini . : : : ‘ «2 SO Tribe Ovibovini . ‘ : : : ero Enigmatic horn-cores : Sst: Sad, PLUS eL Bovidae from Baard’s Quarry . F : eal Tribe Boselaphini : ; 3 : ni et Tribe Reduncini . : ‘ : : eo Tribe Hippotragini . é : : bee S24 Tribe Alcelaphini ; : 3 ; . ~324 Tribe Neotragini ; 5 : : PACs Pe Tribe Antilopini . ‘ 326 Age of the Bovidae from Baard’s s Quarry sh S27 Discussion . , : z S328 Acknowledgements . : : d i ses? 333 References . 5 é 5 : 3 : . 334 213 Ann. S. Afr. Mus. 79 (8), 1980: 213-337, 63 figs, 6 tables. 214 ANNALS OF THE SOUTH AFRICAN MUSEUM INTRODUCTION The late Tertiary fossil vertebrate site of Langebaanweg is situated 105 km north-north-west of Cape Town and about 15 km inland from Saldanha Bay at 32°58’S 18°09’E. Faunal and geological studies of the site have been made by Hendey (1970, 1973, 1974, 1976) and Tankard (1974). Commercial mining for phosphates led to the opening of a number of quarries in which Tertiary gravels, sands and clays were exposed beneath the covering of Pleistocene and Recent sands. Most of the abundant remains of freshwater, marine, terrestrial and flying vertebrates have come from deposits constituting the Varswater Forma- tion in the New Varswater Mine or ‘E’ Quarry. A smaller number came from unnamed deposits in Baard’s Quarry, but this has now been backfilled and the relationships of its deposits to those in ‘E’ Quarry are uncertain. All unqualified references to Langebaanweg fossils are to those from ‘E’ Quarry. ‘E’ Quarry is the type locality for the Varswater Formation, and its stratigraphy is as follows: Largely or entirely Pleistocene/ Surface bed Holocene Latest Miocene/early Pliocene Varswater Formation: Pelletal Phosphorite Member Quartzose Sand Member Gravel Member Miocene Saldanha Formation The Varswater Formation is only about 10 m thick in ‘E’ Quarry, but © elsewhere it can reach 39-43 m. Its deposition was initiated bya marine trans- gression probably in the late Miocene during which the Gravel Member was deposited in a rocky and sandy marine beach environment. This member consists of boulders, cobbles and pebbles of phosphate rock in sands and has yielded marine invertebrates and vertebrates. Terrestrial vertebrates are rarer and their bones have usually been heavily rolled, evidently the result of wave action. There followed a stillstand in the transgression, during which the Quartzose Sand Member was accumulated in a variety of depositional environments in and near an estuary. Lithologically it is the most complex of the Varswater members in ‘E’ Quarry and the demarcation of its upper and lower limits has been difficult. Three main facies are recognized within the Quartzose Sand Member. The first, of which there are extensive exposures, is composed of largely non-phosphatic quartz sands, believed to represent a floodplain environ- ment. This unit contains vertebrate fossils which accumulated both subaerially and subaqueously. The second is an horizon of carbonaceous sand and clay (the ‘peat bed’), which probably represents a marsh environment. The third facies is a muddy silt rich in invertebrate fossils but without significant verte- brates, which apparently represents a tidal mudflats accumulation. Exposures of the marsh and tidal mudflats facies are limited in extent. FOSSIL BOVIDAE FROM LANGEBAANWEG 215 The final phase of the transgression was the period of accumulation of the Pelletal Phosphorite Member—medium grade phosphatic sands whose base truncates the Quartzose Sand Member. The lowermost part of it in ‘E’ Quarry was laid down in shallow marine water, and the remains of both terrestrial and aquatic vertebrates are numerous in restricted areas in, and immediately adjacent to, a river channel. It is believed that the course of the river shifted northward as the transgression progressed and two distinct channels have been exposed. These deposits are informally termed bed 3aS and bed 3aN (Hendey 1976: 226), the former being the older. The duration of the intervals between deposition of the Quartzose Sand Member and bed 3aS and between beds 3aS and 3aN are not known. The mammal fauna from ‘E’ Quarry has been discussed by Hendey (1976: 231-243). The larger mammals of the Quartzose Sand Member are those on which the relative dating originally depended, and Hendey (1973: 13; 1974: 61, 62) inferred an age of about 4,5 m.y. by correlation with dated east African faunas. There is now more uncertainty about the age, and the Varswater Formation may have accumulated over an appreciable period. It is, however, generally agreed that none of the sediments is likely to be younger than 3,5 m.y. and none older than about 7 m.y. (Hendey 19785: 267-269). The varied her- bivores suggest a more luxuriant vegetation than exists in the area at the present time, and there is evidence of fire damage to some of the bones in the Quartzose Sand and Pelletal Phosphorite Members. Other African localities African localities other than Langebaanweg which are mentioned in this paper are: Afar, Ethiopia, comprising mainly the Hadar Formation which appears by radiometric methods to date from 3,1 to less than 2,6 m.y. (Taieb et al. 1978; Aronson et al. 1977). The Amado Formation is of unknown age. Beglia Formation, Tunisia, aged about 12-13 m.y. (Robinson & Black 1969). Elandsfontein, near Langebaanweg, a rich site of Middle Pleistocene age with some later fossils (Klein 1978; Hendey 1974: 26). Fort Ternan, Kenya, dated to 14 m.y. (Gentry 1970a; Bishop, Miller & Fitch 1969). Kaiso Formation, Uganda, thought to span about 5,0-2,5 m.y. and to have an earlier and a later faunal level (Cooke & Coryndon 1970). Karmosit Beds, Kenya, probably a little older than 3,4 m.y. (Bishop et al. 1971): Laetoli, formerly called Laetolil, Tanzania. The Laetolil Beds date from 3,59 to 3,77 m.y., and the later Ndolanya Beds are older than 2,4 m.y. (M. D. Leakey et al. 1976; M. D. Leakey & Hay 1979). Lothagam, Kenya. The Logatham 1 fauna may be about 5,5 m.y. (Behrens- meyer 1976; Smart 1976). 216 ANNALS OF THE SOUTH AFRICAN MUSEUM Lukeino Formation, Kenya, 6,0-6,7 m.y. (Pickford 1975, 1978b; Thomas 1979b). , Makapansgat Limeworks, Transvaal, South Africa, where the bovids appear to come between Langebaanweg and Olduvai middle and upper Bed II. The fauna derives from more than one stratigraphic level (Wells & Cooke 1956; Gentry & Gentry 1978: 66; Vrba 1977). Mpesida Beds, Kenya, about 7 m.y. (Bishop et al. 1971; Thomas 19795). Mursi Formation, Omo, Ethiopia, where a basalt overlying the fossili- ferous levels has been dated to 4,05 m.y. (Butzer & Thurber 1969). Ngorora Formation, Kenya, spanning 12-9 m.y. (Bishop & Pickford 1975; Pickford 1978a.) Olduvai Gorge, Tanzania, where Beds I to IV span 2,1-0,6 m.y. (M. D. Leakey 1971; Gentry & Gentry 1978). Peninj, Tanzania, which correlates faunally and radiometrically with Olduvai middle and upper Bed II (Gentry & Gentry 1978: 292, 62-63). Sahabi, Libya, of latest Miocene age, perhaps slightly younger than Wadi Natrun (Maglio 1973: 68, 70; Boaz et al. 1979). Shungura Formation, Omo, Ethiopia, with an approximate time span from 3,2-0,8 m.y. (Coppens et al. 1976; Brown et al. 1978). Member B has an age of about 2,8 m.y. and member G an age slightly younger than 2 m.y. Sterkfontein Type Site or Main Quarry, Transvaal, South Africa, at which most of the mammalian fauna comes from member 4 with a probable age of about 3,0-2,5 m.y. (Vrba 1976, fig. 20; Partridge 1978). Swartklip, southern Cape Province, South Africa,- of Upper Pleistocene age (Hendey & Hendey 1968; Klein 1975). Wadi Natrun, Egypt, with a poorly known vertebrate fauna, perhaps about 6 m.y. (Andrews 1902; Maglio 1973: 70). The most frequently mentioned locality outside Africa is the Siwaliks Group in India and Pakistan, which has a sequence of deposits ranging from Miocene to Pleistocene. Pilbeam et al. (1977) give a condensed history and much new information on these deposits. At present it appears likely that the bulk of known Lower and Middle Siwaliks faunas fall into two groups. An earlier fauna comes from the upper two-thirds of the Chinji Formation at its type locality and from Ramnagar. It best resembles Astaracian faunas of Europe and west Asia and its age is judged to be about 12-13 m.y. It contains the lopho- dont pig Listriodon but no Hipparion. A later fauna comes from the middle part of the Nagri Formation at Nagri, the upper part of the same Formation in the Dhok Mila—Gandakas area and Haritalyangar, and continues through the succeeding Dhok Pathan Formation in its type and adjacent areas. (The Dhok Pathan Formation is known only from its upper levels in the type area.) This later fauna agrees best with Vallesian and Turolian faunas elsewhere and has a likely age range from about 10 to 7,5 m.y. ‘Dhok Pathan’ fossils collected FOSSIL BOVIDAE FROM LANGEBAANWEG 217 in the Hasnot area are somewhat younger, perhaps about 7,0 m.y. Hipparion enters the sequence at the poorly fossiliferous base of the Nagri Formation. Classification The classification of bovids used here is modified from that of Simpson (1945), with some improvements from Ansell (1971) and some new features: Family Bovidae Subfamily Bovinae Tribe Tragelaphini Tribe Boselaphini Tribe Bovini Subfamily Cephalophinae Tribe Cephalophini . Subfamily Hippotraginae Tribe Reduncini Tribe Hippotragini Subfamily Alcelaphinae Tribe Alcelaphini Subfamily Antilopinae Tribe Neotragini Tribe Antilopini Subfamily Caprinae Tribe ‘Rupicaprini’ . Eland, bongo, kudus, mountain nyala, sita- tunga, nyala, bushbuck. Mainly browsers in bush and forest. Now represented by only the nilgai and four-horned antelope in India, but formerly occurred in Africa Cattle and buffaloes, the largest bovids Duikers, mainly small forest antelopes which are rarely fossilized Waterbuck, lechwes, kob and reedbucks. Grazing antelopes always found in the vicinity of water Roan, sable, oryxes and addax Wildebeests, hartebeests, bastard harte- beests, Hunter’s antelope or Tana River hartebeest. Grazing, cursorial antelopes of open country. Includes Aepyceros, the impala, usually placed in the Antilopini Royal antelope, Bates’ dwarf antelope, suni, dik-diks, steenboks, grysbok, klipspringer, oribi, beira. Small antelopes not found in such dense cover as duikers Gazelles, springbok, blackbuck, gerenuk and dibatag. Also includes tribe Saigini (containing Saiga and Pantholops of Asia). Small to medium sized, cursorial antelopes often adapted to conditions of water shortage Goral, serow, Rocky Mountain goat. Rupi- capra itself, the chamois, might be better placed in the Caprini. Not found in Africa 218 ANNALS OF THE SOUTH AFRICAN MUSEUM Tribe Ovibovini ; Musk ox and takin. More abundant earlier in bovid history than they are today ‘Fribe Caprint =yoex: Sheep, goats, tahrs The first three subfamilies have been thought to comprise a group called the Boodontia and the second three the Aegodontia (Gentry 19785: 564; Simpson 1945: 270; Pilgrim 1939: 10), but these groups are not used in the formal classification adopted here. Abbreviations The present report covers Bovidae which had been incorporated into the collections of the South African Museum by early 1977. All fossils from Lange- baanweg are in the South African Museum. Catalogue numbers of these specimens begin with the letters SAM-—PQ-L, in which SAM refers to the South African Museum, PQ is a departmental prefix, and L stands for Lange- baanweg. Other abbreviations in the text are: QSM Quartzose Sand Member of the Varswater Formation PPM Pelletal Phosphorite Member of the Varswater Formation 3aS bed 3aS of the Pelletal Phosphorite Member 3aN bed 3aN of the Pelletal Phosphorite Member BM(NH) British Museum (Natural History), London BPI Bernard Price Institute for Palaeontological Research, Johannesburg KNM Kenya National Museum, Nairobi m.y. millions of years Specimens from the Omo and Afar in Ethiopia have yet to be lodged permanently in an institution, and only locality and field catalogue numbers are given for them. Measurements Measurements are given in millimetres. Tooth measurements were taken on specimens in the earlier or later parts of middle wear. They were not taken on specimens in early or late wear unless this is stated. Length measurements of limb bones were taken as follows: Femur—from the lateral end of the articular head to the lowest level of the distal medial condyle Tibia—from the lowest point of the top medial facet to the projecting tip of bone behind the medial malleolus Metatarsal—from the highest point behind the medial part of the ecto- cuneiform facet to the medial side of the most projecting part of the distal medial condyle Humerus—from the top of the lateral tuberosity to the lowest point of the medial side distally Radius—from the centre of the medial edge of the proximal medial facet FOSSIL BOVIDAE FROM LANGEBAANWEG 219 to the lowest point of the ridge on the scaphoid facet medially Metacarpal—from the edge of the proximal articular facet above the insertion for the extensor carpi radialis to the median side of the most projecting part of the distal medial condyle Terms Five terms used frequently in the text need explanation: The basal index of a horn-core is a pair of measurements given in the form 46,9 x 37,2 in which the first is the anteroposterior diameter at the base and the second the mediolateral diameter at 90° to the first Horn-cores are said to be obliquely inserted when their inclinations are low in side view. It is the opposite condition from upright insertions Horn-cores with any degree of curvature frequently have torsion. This may be clockwise or anti-clockwise and is described as it exists in the right horn-core A basal pillar, when it occurs, is found in the centre of the medial side of upper molars or the lateral side of lower molars, completely or partly separate from the rest of the occlusal surface. In the Cope—Osborn nomenclature it is the entostyle of an upper and ectostylid of a lower molar A goat fold is a transverse flange at the front of the lower molars. BOVIDAE FROM ‘E’ QUARRY Tribe Tragelaphini Tragelaphus spp indet. Figs 1-4. Material A number of tragelaphine horn-cores from ‘E’ Quarry at Langebaanweg belong to a species about the size of the living nyala, Tragelaphus angasi. They are: L5252—basal half of a right horn-core L5868—base of a left horn-core L5922—base of a right, index 42,1 x 41,4 L5924— base of a right, index 39,1 x 38,8 L6568— basal half of a right, index 40,1 x 42,3 (Fig. 1) L6574— basal half of a left L40056— base of a right, index 39,2 x 41,6 L4620, L5255, L6079, L6081, L6083, L6084, L6379, L6571, L6576, L41039 —parts of left horn-cores L5253, L5716B, L5920, L6435, L6569, L6570, L6575, L6583, L6584, L13983 —parts of right horn-cores In addition, there is an occipital surface, L5085, of appropriate size to be conspecific with the horn-cores. The height of the occipital, measured from the dorsal edge of the foramen magnum, is 41,3. 220 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 1. Tragelaphus sp. L6568, anterior view of right horn-core. Scale = 50 mm. L40056 and L41039 of the above listed material are definitely from bed 3aS of the PPM. The rest is probably also from 3aS but a few may be from the QSM. Parts of left and right horn-cores, L13164 and L22556 additional to the above list, do come from the QSM. Parts of two left and right horn-cores, L33779 and 133833, come from bed 3aN of the PPM. Also probably from 3aN is L40759 (Fig. 2), the greater part of a left horn-core with a rather damaged base and an index of 42,2 x 46,3. Horizon Tragelaphus is much better represented in bed 3aS than in either 3aN or the QSM. Nearly all the horn-cores were picked up by mine workers, many in 1966-7, and were not recovered from controlled excavations. Description On horn-core L6574 the posterolateral keel is weaker than the anterior one and can hardly be seen at all. This may have been an unusual individual in life or the fossil may have suffered water rolling after death. In the other horn- cores the posterolateral keel is always strong. The anterior keel is rarely as strong as the posterolateral one, but it can be seen to be so at the very bases of L5252, L5868 and L6568. A part of a left horn-core, L33779, is about 140 mm long and has a strong groove alongside, or instead of, a keel. In comparison with the living, similarly sized species Tragelaphus angasi and T. spekei, the horn-cores are less compressed anteroposteriorly, inserted more uprightly, and look as if they were inserted less far behind the orbits. The insertions must FOSSIL BOVIDAE FROM LANGEBAANWEG 221 Fig. 2. Tragelaphus sp. 40759, lateral view of left horn-core. Scale = 50 mm. have been wider apart than at the present day, and this could be linked with the anteroposterior level of the insertions above the back of the orbits rather than just behind them. L6568 and L44056 certainly show a greater width of frontals on the medial side of the horn bases than can be seen in living Tragelaphus. The degree of basal divergence can be seen in anterodorsal view on the horn-cores L5922, L6568 and L40056, and it was greater than in 7. angasi and T. spekei or any other tragelaphine alive today. The degree of spiralling in the Langebaan- weg horn-cores is about as strong as in J. angasi and spekei. There are no sinuses within the horn pedicels. Most of the characters whereby this species differs from equivalent sized living species can be paralleled in the smaller horn-cores of an extinct bushbuck, L144—1 and 2 from member C of the Shun- gura Formation. (However the width apart of the insertions and the basal divergence are not known in the Omo fossils.) Such characters appear to be primitive in the small and medium sized lineages of Tragelaphus. Two of the three horn-cores from bed 3aN deserve particular mention. _ L40759, probably from 3aN, is larger than the main mass of 3aS horn-cores (Fig. 3) and has a preserved length of about 280 mm. It is strongly spiralled, in fact, almost as much as in the living greater kudu, Tragelaphus strepsiceros. Its anterior keel is stronger basally than the posterolateral one, unlike other ‘E’ Quarry Tragelaphus, but by the tip of the preserved part of the horn-core the posterolateral keel has become strong and sharp. Through being more spiralled, the line of the posterolateral keel near the base is more strongly concave than in the other horn-cores. Finds from the Shungura and Mursi Formations (Gentry 1976: 276-7, 288; a fuller account is awaiting publication) suggest that T. strepsiceros evolved from ancestors of smaller size and with horn-cores showing a stronger posterolateral keel and a weaker anterior keel, no medio- lateral compression, weaker spiralling and less upright insertions. One could anticipate difficulties in distinguishing members of this lineage from relatives or ancestors of T. spekei at sites coeval with, or earlier than, the Mursi Forma- tion. Possible kudu horn-cores from the Mursi Formation, particularly the base 222 ANNALS OF THE SOUTH AFRICAN MUSEUM Mediolateral diameter re) 60 50 40 50 Anteroposterior diameter 20 30 40 30mm Fig. 3. Basal dimensions of Tragelaphus horn-cores. X = E Quarry Langebaanweg, O = extant T. spekei, M = Mursi Formation YS 68.2078, L = Makapansgat Limeworks BPI M 490, S=T. ?pricei from Shungura Formation, U = Lukeino Formation (from Thomas 19795). The highest X is L40759. The lower diagonal line is that along which medio- lateral diameter is 100 per cent of anteroposterior diameter; the upper line is 125 per cent. A cross-section of the right horn-core L6568 is shown, taken 20 mm above its base, with lateral side to the left and anterior side to the base. Scale = 10 mm. of a right horn-core YS 1968-2078 with a basal index 38,7 x 44,3 (Fig. 3), are about the same size as L40759, have a posterolateral keel stronger than the anterior one, weaker spiralling than L40759, and about the same inclination and degree of compression. Thus, if L40759 were a different species from other Langebaanweg horn-cores and on the kudu lineage, it would be at about the level of member E or F of the Shungura Formation by its keels and even later FOSSIL BOVIDAE FROM LANGEBAANWEG 223 by its strong spiralling. If it is a different species from the other ‘E’ Quarry horn-cores, it is probably a precocious southern Cape development unrelated to the 7. strepsiceros lineage. L33833, the fragment of a large right horn-core from bed 3aN, has an index of 48,5 x 49,5 at its lowest level which is not known to be the original base. This is larger than other tragelaphine horn-cores likely to be from 3aS. It appears to have little spiralization, but its large size makes the comparison difficult. The posterolateral keel is the most prominent as in most Langebaan- weg horn-cores, and the anterior keel is also well developed. Either or both L40759 and L33833 could represent additional species of tragelaphines in the Langebaanweg fauna. The occipital surface L5085 shows its tragelaphine affinities in its rather flat surface, a median vertical ridge without flanking hollows, traces of a small, narrow mastoid and a horizontal top edge centrally. One of the interesting features of the Langebaanweg assemblage is the problem of finding teeth which might be conspecific with the tragelaphine horn-cores. This question will be taken up later and all that need be stated here is that only a few teeth such as the right M, L4628 (Fig. 4) are candidates for being tragelaphine. This tooth probably comes from the PPM, is in middle wear, and has an occlusal length of 29,2. It resembles M,s of Mesembriportax acrae except in being rather small. A moderately developed metastylid is present, there is a tiny basal pillar, and the medial wall of the back lobe of the tooth is set obliquely. The wear is abnormal in that the occlusal surface slopes steeply down towards the buccal edge. Comparisons The base of a tragelaphine right horn-core, BPI M490 with basal index 37,6 < 38,3, and possibly another fragment, M491, from Makapansgat Lime- works are about the size of the Langebaanweg species, and agree with it in - having the anterior keel no better developed than the posterolateral one and in being less compressed anteroposteriorly than in living 7. angasi and T. spekei. Dentitions of appropriate size to go with the horn-cores have already been assigned to 7. cf. angasi by Wells & Cooke (1956: 10). A right horn-core of a Tragelaphus, BM(NH) M 26402, from the early assemblage of the Kaiso Formation (Cooke & Coryndon 1970: 200; Gentry & Gentry 1978: 305) is about the same size as the Langebaanweg species but more anteroposteriorly compressed, having a basal index of 41,0 x 50,0. The early Kaiso assemblage has been thought to have an age of about 5 m.y. (Cooke & Coryndon 1970: 220, fig. 17; Maglio 1973: 70), but Gentry & Cenlty (1978: 64) thought that the bovids Sold as easily fit a later age. Thomas (19798, fig. 2) has recorded a fine frontlet and two partial horn- cores of Tragelaphus cf. spekei from Lukeino which are very like the Langebaan- weg horn-cores, but slightly less compressed anteroposteriorly. He gives basal indices of 38 x 37 and 44 x 41. Tragelaphine teeth are also present at Lukeino 224 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 4. Tragelaphus sp. L4628, occlusal and lateral views of right M3. Scale = 10 mm. and Mpesida (Thomas 1979), pl. 2 (figs 4, 8-9, 11, 14)). Gentry (1978a: 297) pointed to the appearance of tragelaphine-like teeth as early as Ngorora. Tribe Boselaphini Genus Mesembriportax Type species Mesembriportax acrae Gentry, 1974. Mesembriportax acrae Gentry, 1974 Figs 5-6 Remarks Since Gentry (1974) described Mesembriportax acrae from Bed 2 (= QSM) of ‘E’ Quarry, Langebaanweg, a number of new specimens have come to light from both the QSM and PPM. The two chief ones (Figs 5-6) are: L25870—cranium with both horn-cores attached, other fragmentary skull FOSSIL BOVIDAE FROM LANGEBAANWEG 225 Fig. 5. Mesembriportax acrae. L40071, dorsal view of cranium; L25870, lateral view of cranium Scales = 50 mm. bones, the maxillary tooth rows, both mandibles, and parts of cervical vertebrae. It comes from the QSM L40071—cranium with right horn-core attached, much of the left horn- core, and fragmentary skull bones. This comes from bed 3aS. Other remains include the partial frontlets and skull fragments L20918 and L22005, both from the QSM. The latter has some upper teeth as well as quite a lot of other dental remains. In both L25870 and L40071 the frontals are preserved further forward than in the holotype L13101. Thus it can easily be seen that their horn-core insertions 226 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 6. Mesembriportax acrae. 25870, occipital and ventral views of cranium. Scales = 25 mm. FOSSIL BOVIDAE FROM LANGEBAANWEG p94 | PPM: 3aN x x9 x x X PPM: 3aS 0 x x QSM ‘ KX SR Gs. 0 + 0 cy sn yee an BSP wen ga 40mm Fig. 7. Occlusal length of M; in Mesembriportax acrae. X = right side, O = left side, + = both sides. Dotted readings are in later wear. are slightly closer than on the holotype (Gentry 1974, fig. 2). Moreover, the divergence of the horn-cores is much less, the torsion is less strong, and the dorsal parts of the orbital rims project more strongly. These characters bring L25870 and L40071 closer to Miotragocerus, especially M. amalthea, a fairly large and advanced species from Pikermi, Greece, and other sites of Turolian age. Other characters show the variation among the Langebaanweg specimens but do not correlate with the geology of the site. Compared with L13101, L25870, likewise from the QSM, shows a stronger approach to a posteromedial keel, more oblique horn-core insertions, a more sharply localized raising of the frontals between the horn-core bases, almost certainly a less sloping braincase roof, and a braincase which does not widen posteriorly. However, L40071, which is more like the QSM holotype for these characters, itself comes from the PPM. Besides L40071, there are other, less well preserved, remains of Mesembriportax acrae from the PPM. No valid differences between them and the remains from the QSM were found, but it appears from a comparison of Ms; occlusal lengths that there could have been a size increase (Fig. 7). This is also shown by a statistical comparison of both M, and Ms; occlusal lengths: Number Standard Standard measured Mean Range deviation error MemrOsM - . 12 2371 208-259 1055) 0,42 PPM :3aS 3 24,5 23,9-25,3 0,7 0,42 PPM :3aN 3 24,7 23,8-25,7 1,0 0,55 M; QSM 9 32,3 oh p292= 35:3 2,0 0,66 PPM :3aS 2 32,0 30,2-33,8 — —- PPM :3aN 6 SO VCR 233 0:92 The values of T for M,s and M,s in the QSM and bed 3aN were 1,14 and 228 ANNALS OF THE SOUTH AFRICAN MUSEUM 2,65, the latter alone being significant at the 5 per cent level. The measurements were confined to teeth in middle wear, and each sample was from either the left or right side alone. Gentry (1974: 180) compared the dental characters of Mesembriportax acrae and Miotragocerus and stated that the former had straighter medial walls of its lower molars. In addition, the P, of the Mesembriportax acrae holotype had a more massive metaconid with less differentiation into a neck and strong anterior and posterior flanges, the paraconid had the form of a low protuberance from the parastylid rather than a flange, and the hypoconid projected more strongly than in many Miotragocerus. In the P, of L20508 only the last differ- ence appeared valid. This statement can be modified now that more is known of variation during wear in Mesembriportax acrae. Of a dozen mandibles of M. acrae with P,, six were held to be in early or early middle wear and the remainder in late middle or late wear. Five P,s out of the six in earlier wear had flanges anteriorly and/or posteriorly on the metaconid, while four of the six in later wear had scarcely any traces of flanges. Again, four of the first group had the paraconid of P, in the form of a flange and five of the second group had it merely as a low protuberance from the parastylid. Seven out of the whole twelve had a fairly projecting hypoconid on P, and four had a fusion between paraconid and metaconid (the latter character not mentioned by Gentry 1974), but these characters were uncorrelated with wear. A similar loss of definition of the various flanges can be seen during wear in Miotragocerus amalthea from Pikermi, but it does seem that the paraconid flange is more marked in this species. In the author’s revised opinion this becomes one of two good differences in the P,s of the two species. The other is that paraconid—metaconid fusion is far more infrequent in M. amalthea; it occurred in only one out of twenty-six examples from Pikermi. The straighter medial walls of the lower molars in Mesembriportax acrae appear to be present in only about a third of the speci- mens—seven out of twenty-four. Revised diagnosis The diagnosis of Mesembriportax acrae (Gentry 1974: 148) should there- fore be modified and reduced to the following. A moderate to large boselaphine with short to fairly long horn-cores. Horn-cores with varying insertion angle, basal divergence and width apart of insertion positions. Horn-cores compressed mediolaterally and with a posterolateral keel and a strong slightly helical anterior keel in their lower part, the anterior keel being stepped at its top and the succeeding distal part of the horn-core being of small circular cross-section. In so far as any torsion exists, its direction is anti-clockwise in the right horn- core. Divergence of the horn-cores lessens distally, and there is little or no backward curvature. Frontals extensively hollowed internally, and their top surface raised much above the level of the top of the orbits; braincase in line with or slightly angled on the face axis; top of braincase not curved downward posteriorly above the occipital surface; strong temporal ridges on braincase FOSSIL BOVIDAE FROM LANGEBAANWEG 229 roof not approaching closely posteriorly and with a rugose surface between them; orbits sometimes without a projecting dorsal rim; small supraorbital pits; nasals long and narrow with large central flanges anteriorly but no lateral flanges; large preorbital fossa; infraorbital foramen low and situated above the posterior margin of P?; premaxillae narrow anteriorly but with strong ascending rami of approximately even width throughout and with a wide contact on the nasals; palate very wide; median indentation at the back of the palate well behind the level of the lateral indentations; occipital surface with a squared outline; large mastoid exposure of periotic; anterior tuberosities of basioccipital fairly wide apart and not very large. Brachyodont or only moderately hypsodont cheek teeth, with not very rugose enamel; small basal pillars on upper and lower molars; medial lobes of upper molars not joined to one another or to the lateral side of the tooth until late in life; mesostyles quite strong on upper molars; central cavities of upper molars not very complicated in outline; medial walls of lower molars sometimes rather flat; lower molars sometimes with a small goat fold (a transverse flange at the front of the tooth); M, often with a large central cavity in the rear lobe during early wear and often with a rear flange; long premolar rows with large anterior premolars; paraconid and metaconid more often unfused than fused on P,; paraconid of P, has the shape of a low protuberance or small flange on the back of the parastylid; hypoconid sometimes projecting on P,; I,s not greatly enlarged. Measurements Measurements on the two most complete new crania of M. acrae are: L25870 L40071 Length of horn-core along anterior keel . . 310 420 Anteroposterior diameter at base of horn-core . 66,3 12,2 Lateromedial diameter at base of horn-core : 47,2 51,2 Minimum width across lateral sides of horn PE iecisong orate ~ osmium” | reuee’ be code TAO) — Occipital height from dorsal edge of foramen magnum . : ; : i : : 49,5 47, ooo Skull width across mastoids behind external auditory meatus cs cute ats gi eracanteg in cle wean een ee eee 124,6 Width across anterior tuberosities of basioccipital 28-2 28,0 Width across posterior tuberosities of basioccipital 46,0 40,6 Weelusat lensth:Me=Me 67,9 — Mechs lense Me ee. us cee ee 24,3 — Meeinsanienetae Ps =o 8. re eee al eB oe meee 14,7 —- 230 ANNALS OF THE SOUTH AFRICAN MUSEUM Measurements on new more fragmentary skull remains are: n en a eee SS a3 85 82 82 Sz $2 HO WO YO 12 Length of horn-core along anterior keel . : — — — — — 330 Anteroposterior diameter at base of horn-core. 60,1 — — 69,5 — 79,8 Lateromedial diameter at base of horn-core . 43,6 — — 41,6 — 38,5 Width across anterior tuberosities of basi- occipital . : et QO 27-6 *- BAsG — 28,0 — Width across _ posterior tuberosities of same 37,7 36,1 39,6 o 42,9 — Some measurements on the more complete dentitions in middle wear are shown in Table 1. All are likely to be from different individuals. TABLE | Measurements of boselaphine dentitions. L20985 125870 L28327 L46059 L20542 L21744 QSM QSM QSM 3aN QSM QSM Occlusal length M,—-M, : ; 4 69,0 70,8 76,8 81,6 — — Occlusal length M, ; 3 : : 21,6 23,0 25,0 Dip! — — Occlusal length M, ; ; \ d DD 30,6 BO 33,7 — — Occlusal length P,-P, . 3 : : 48,7 Sy OSes SS)57/ 53,225, aeoihes Saal Occlusal length P, . : : : Shae Se 16,2 15:5 7. PeROre = 12.9 Occlusal length P, . , : ; : 17,4 Leg) — 20,3 729509 or 2D, | Ramus depth below P, . ; : ; 34,8 36,9 — 31,9 — — Ramus depth below M, : ‘ 3 39,7 39,3 37,8 — — — Ramus depth below M; - : 5 44,0 41,1 39,2 —_— — — L25033 L25870 L32401 QSM QSM 3aN Occlusal length M1—-M? ‘ : : 70,1 67,9 68,4 Occlusal length M? : ; : é D5), 3) 24,3 230), Occlusal length P?—-P* . : : , — — S515 /i Occlusal length P* . ‘ ; : ; 15,3 14,7 ISS) * — deciduous dentition Occlusal lengths of other teeth in middle wear are as follows: L14253 left P4 15,7 QSM; L20982 right P* 15,4 QSM; L22253 right P4 15,2 QSM; L41720 left P* 17,9 PPM L21828 left dP? 20,9, dP* 20,2 QSM £11350: left. P,- 21593 1.22369 left PP, 20,6 QSM; 132792 emehim yy 13:0 PPM; L40976A right P, 19,4 PPM; L41689 left P, 15,4 PPM; L41720 right P, 21,4 PPM. L24651 left dP, 25,3 QSM; L30948 left dP, 29,9 PPM; L31076 right dP, 27,5 PPM FOSSIL BOVIDAE FROM LANGEBAANWEG 231 Comparisons Because of the resemblances of the new material to Miotragocerus it is neces- sary to revise the account of differences between Mesembriportax acrae and other boselaphines from that given in Gentry (1974: 179-181). The most important comparison lies with Miotragocerus amalthea, a species which has considerable variation of its horn-cores (Pilgrim & Hopwood 1928: 46-49). The Langebaanweg species is now seen to have differences in its larger size, less mediolateral compression of the horn-cores, even less backward curvature of its horn-cores, a longer terminal portion of its horn-cores distal to the top of the anterior keel, horn-cores usually more divergent, frontals raised to a higher level between the horn-core bases relative to the dorsal part of the orbital rims and hence with a more extensive system of internal sinuses, perhaps a better rugose surface on the braincase roof behind the horn-cores, and a squarer outline of the occipital surface. For most of the differences the South African species can be plausibly supposed to be more advanced. Another Miotragocerus species at Samos is larger than M. amalthea, and includes material named M. curvicornis and M. recticornis (Andree 1926). This species is unlike both M. amalthea and the Langebaanweg species in that its horn-cores have little or no demarcation of a distal portion with a rounded cross-section. The remaining principal species of Miotragocerus were listed by Gentry (1974: 175). Among these M. gradiens is a small, primitive species and M. pannoniae and M. leskewitschi are later but still small species. M. spectabilis is a larger and later species from China similar to M. amalthea. M. valenciennesi is a small to moderate sized species coexisting with M. amalthea at Pikermi, and M. browni is a moderate sized Siwaliks species, perhaps descended from the earlier M. gradiens. The only fossils of Miotragocerus from Africa are the record from Lothagam (Smart 1976: 365) and a frontlet with much of its horn-cores from Sahabi, now -in Rome. The latter was taken by Thomas (1979a: 268, pl. 1 (figs 5a—5b)) as holotype of a new species M. cyrenaicus. It differs from the Langebaanweg species by being probably somewhat smaller and its horn-cores more strongly compressed mediolaterally, slightly curved backward, and with no sharp diminution of anteroposterior diameter distally. This last character causes it to resemble the Samos examples of M. curvicornis and M. recticornis rather than M. amalthea. Its horn-cores are strongly divergent and with little sign of a posteromedial keel as in the Mesembriportax acrae holotype, but their insertion - angle is lower and they have less torsion. It also differs from the holotype by _ having projecting rims to its orbits dorsally. It is not likely to belong to the caprine Pachytragus which occurs in the earlier pre-Hipparion and Hipparion levels of the Beglia Formation (Robinson & Black 1969; Robinson 1972) because of the wider insertion of its horn-cores, and the anterior keel extending low on the pedicel to below the level of the lateral and medial sides. The low inclination and wide divergence of the horn-cores are also unlike Pachytragus. 232 ANNALS OF THE SOUTH AFRICAN MUSEUM Mesembriportax acrae differs from Miotragocerus browni by its greater size, wider skull and the horn-cores being shorter, not curved backward and with the anterior keel terminating well below the horn-core tip. The characters distinguishing Mesembriportax acrae from Protragocerus as a whole are fewer than given by Gentry (1974: 179). They now comprise greater size, sinuses in the frontals, braincase roof not curved downward posteriorly, presence of a rugose surface between the temporal ridges, and a larger basi- occipital with stronger anterior tuberosities. These differences continue to be more impressive than those separating M. acrae from Miotragocerus. It seems that in Europe Miotragocerus does not usually occur as early as Protragocerus. It is known back to the Vallesian and has one late Astaracian record (M. monacensis Stromer, 1928, of ‘Sarmatian’ age, the type species of Miotragocerus). Thus, according to the biozones set up by Mein (1975), it occurs from zones 12 or 13 back to 9 or perhaps 8. Protragocerus, as recorded from such sites as Belomechetskaya, Despotovac, Atzgersdorf, Hollabrunn, Sommerein, and La Grive St Alban (Thenius 1956, 1959), is of Astaracian (= late Vindobonian and ‘Sarmatian’) age, equal to Mein’s zones 6 to 8. In the Siwaliks Miotragocerus gradiens occurs together with Protragocerus in the Chinji fauna, thought to be of Astaracian-equivalent age, but Miotragocerus also survives until the Dhok Pathan Formation (Pilbeam et.al. 1977). Thus, Miotragocerus is generally younger than 12 m.y. whereas Protragocerus is from 12 to 14 m.y. Relationships of these early boselaphines are poorly understood, but it seems that Protragocerus could well be a stem genus from which Miotragocerus and other genera took their origins—cf. Gentry (1974, fig. 27) in which Miotrago- cerus is shown having an ancestry independent of Protragocerus. If Mesembriportax acrae were to be described as a new species in this paper, one might well decide to put it into Miotragocerus as, indeed, Thomas (1979a: 273) has suggested. However, for the present Mesembriportax will be retained. M. acrae is not very like the north African Miotragocerus cyrenaicus and it is still possible that it is an independent line of descent from an African Protragocerus. Tribe Bovini Simatherium Dietrich, 1941 Simatherium Dietrich, 1941: 221. Simatherium Dietrich, 1942: 119. Type species Simatherium kohllarseni Dietrich, 1942: pl. 20 (figs 161, 163, 165). Generic diagnosis Extinct moderate to large sized African Bovini with short to moderately long horn-cores, rather massive for the size of the skull. Horn-cores slightly compressed mediolaterally or without compression, sometimes with an anterior keel, of irregular or rounded rather than neatly triangular cross-section, inserted FOSSIL BOVIDAE FROM LANGEBAANWEG 233 just behind the orbits, inserted widely apart, with moderate to strong divergence, gently curved backward in side view, and without torsion. Horn-cores some- times with deep longitudinal grooves. Frontals and horn pedicels with quite extensive, irregularly shaped internal sinuses, braincase short, braincase roof sloping a little downward posteriorly, temporal ridges present, a rugose raised area at the back of the braincase roof where temporal ridges converge toward the top of the occipital surface, occipital broad and low, with horizontal top edge and with some development of hollows dorsally on either side of the median vertical ridge. Nuchal crests strong. Moderate to large mastoid. Basioccipital wide posteriorly and triangular, with a short central longitudinal valley between the posterior tuberosities, with a central longitudinal ridge in the area just behind the anterior tuberosities, with small localized anterior tuberosities and poor or no longitudinal ridges behind them. Remarks The single species hitherto known of Simatherium is represented by a poorly preserved cranium from the Vogel River, Laetoli, kept in the Palaeonto- logical Museum of Humboldt University, East Berlin, no. Vo 670. Its precise stratigraphical provenance is unknown but Dietrich (1942; 1950: 49) assigned it to the oldest of the faunas from this area. Some isolated bovine teeth (Dietrich 1950, pl. 1 (fig. 5), pl. 3 (figs 32, 36)) are probably of S. kohllarseni. By invitation of M. D. Leakey, the writer has been able to see a bovine cranium recently recovered from the Laetolil Beds (as defined by M. D. Leakey et al. 1976), which appears to be a second specimen of S. kohllarseni. Simatherium is a possible ancestor for Pelorovis, the extinct long horned ‘buffaloes’ of the African Pleistocene (Gentry & Gentry 1978: 311). It is more primitive than Pelorovis, and, therefore, has some similarity to Ugandax, a genus probably ancestral to the extant Syncerus. The choice for generic identity of the Langebaanweg bovine lay between Ugandax and Simatherium and finally the latter was chosen, although with mainly primitive forms the balance of evidence is not overwhelming. Simatherium demissum sp. nov. Figs 8-13 Holotype L45001—nearly complete left and right cores with midfrontals suture, left mandible with P,-Ms;, left M’*, right P?, parts of occipital surface, basi- occipital, all found associated (Figs 8-11). Referred material The main specimens assigned to this species are as follows. From QSM: L20905—left M3, occlusal length c. 37,4, early middle wear L21297—right dP?, dP*, M'-M®, left dP?+dP*; unworn right P?+ P? and left P?—P4 234 ANNALS OF THE SOUTH AFRICAN MUSEUM L25861—right and left mandibles with P,-Ms, upper dentitions, early middle wear. Skull fragments and postcranial parts (Fig. 10) L28328—left mandible with M,+ Ms, early middle and distorted wear Probably from QSM: L23400—crushed cranium with horn-cores. Isolated teeth of upper denti- tions in late middle wear. Fragments of skull, vertebrae and BED st: cranial bones From bed 3aS: L40094—right mandible fragment with Ms, occlusal length 39,3, late middle wear L41709—left mandible with P,-M;, middle wear L41736—left mandible with P,-M3, early middle wear Probably from 3aS: L11981—right M?, occlusal length 35,1, late middle wear L12116—left mandible fragment with damaged Ms, occlusal length c. 43,0, early middle wear Probably from bed 3aS, but a few possibly from QSM: L4615—fragment of right horn-core, index = 82,9 x c. 77,0. Piece of horn-core about 160 mm long L6586—fragment of left horn-core L1843—left maxilla with dP?-dP4, right sendble with dP,-M, L2051—right Ms, occlusal length 39,5, late middle wear L4774—right M2, occlusal length 37,2, late middle wear L6599—left mandible fragment with P,;+P,, occlusal length P, 2533, 1m late « middle wear L7255—left P*, occlusal length 18,7, early middle wear From bed 3aN: L30174—complete but weathered right horn-core, left horn-core incomplete at base, basioccipital, back of braincase L30175—frontlet with complete horn-cores, skull fragments, basioccipital and part of sides of braincase, right and left maxillae each with P*-M? in late middle wear (Fig. 9) L30880—fragment of left horn-core L30888, L30889—bases of horn-cores L32609—left and right mandibles with P,-M;, middle wear L33380—right mandible with dP,—M,, occlusal lengths dP,-dP, 65,2, dP. 307 L33841—right mandible with P,-Ms, late middle wear L45029—right mandible with P,, occlusal length 16,2, ramus depth below Po 3n9 L46058—right mandible with M,—Msg, early wear L46073—left mandible with P,-Ms, early middle wear Locality The holotype is from bed 3aN of the PPM. The provenance of other speci- FOSSIL BOVIDAE FROM LANGEBAANWEG 235 Fig. 8. Simatherium demissum. 145001, holotype. Dorsal view of left horn-core and frontal; dorsal and occipital views of cranium. Scale = 25 mm. 236 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 9. Simatherium demissum. 30175, anterodorsal view of frontlet with cross-section of right horn-core at level shown. Lateral side of cross-section to the left and anterior side to the base. Scale = 50 mm for frontlet and 25 mm for cross-section. L45001, holotype. Lateral view of right horn-core. Scale = 50 mm. mens has already been given, and it can be seen that Simatherium demissum is represented in the QSM and both beds of the PPM. Diagnosis Horn-cores with a strong anterior keel, without much tendency to a postero- lateral keel, with some rounding of the cross-section on either side of the anterior keel so that its shape, even near the base, is not neatly triangular, horn-cores inserted at a low inclination in side view, strongly divergent basally in anterodorsal view and curving strongly so that at the tips they are parallel or even slightly convergent. Slight backward curvature of the horn-cores is not confined to their basal sector. Sometimes with deep longitudinal grooves on posterior surfaces of horn-cores. Localized rugose areas on the front of the pedicels below and in line with the anterior keels. FOSSIL BOVIDAE FROM LANGEBAANWEG 237 Fig. 10. Simatherium demissum. 25861, occlusal view of right upper cheek tooth-row. L45001, holotype, occlusal view of left lower cheek tooth-row. Scale = 25 mm. Fig. 11. Simatherium demissum. 45001, holotype. Lateral view of left mandible. Scale = 25 mimi: Internal sinuses of frontals and horn pedicels reaching as much as 15 mm above the top of the pedicels, small supraorbital pits, frontals not raised between horn-core bases and no higher than dorsal part of orbital rims, dorsal part of orbital rims projecting quite strongly. The parietofrontals suture has a shallow V-shaped outline pointing forward. Temporal ridges well marked. Median vertical occipital ridge with some development of flanking hollows at - the top. Mastoid moderate sized. Basioccipital long, narrowing fairly abruptly _ just in front of the posterior tuberosities. Cheek teeth large and only moderately hypsodont with rugose enamel, basal pillars present and of moderate size but simple outline, diminishing in size from front to back of the upper and lower tooth-rows, central cavities with fairly simple outline, styles moderately strong on upper molars, ribs quite large but not very localized on lateral walls between styles, transverse goat folds 238 ANNALS OF THE SOUTH AFRICAN MUSEUM practically absent at front of lower molars but mesostylid well developed, out- bowings on front and back of parts of medial walls of lower molars quite well localized as ribs, back lobe of M, often with a posterior flange and with only a small or no central cavity, P, with hypoconid not projecting very far, metaconid slanted diagonally backward but sometimes with an incipient forwardly directed flange, paraconid quite distinct from parastylid. Etymology The name is from the Latin demissus, drooping, and refers to the low inclination of the horn-cores. Remarks Most of the fossils of Simatherium demissum are fairly broken up, but the parts which have survived are well preserved, and the total assemblage allows one to acquire a good idea of much of the cranial morphology. The two horn- core pieces L4615 and L6586 were previously misidentified as from a kudu of similar size to the Olduvai Bed II Tragelaphus strepsiceros grandis (Gentry in Hendey 1970: 114). Two characters in this species are primitive among bovines: the rather low-crowned cheek teeth and the closeness of the horn-core insertions to the back of the orbits, indicating that the insertions have not started their evolutionary migration backward on the skull. Other characters which can reasonably be taken as primitive are the backward curvature of the horn-cores in profile, temporal ridges approaching relatively closely posteriorly on the braincase roof, relatively high rather than low and wide occipital surface, and rather smooth anterior tuberosities of the basioccipital. Isolated teeth of Simatherium demissum are not always easily told from the teeth of Mesembriportax acrae especially those of larger size from bed 3aN. One may hope that bovine teeth will show all or many of the following characters: larger size, more hypsodonty, larger basal pillars, more rugose enamel, ribs stronger in relation to the mesostyles on the upper molars, less flattened medial walls on the lower molars, stronger mesostylids (contrasting with less of a tendency to goat folds on the lower molars), central cavity absent or small and more restricted to the anterior part on the third (rear) lobe of M; (Fig. 12). A number of limb bones are likely by their size and morphology to belong to the Bovini. Two associated sets are, firstly, distal left humerus L12764, distal left radius L12762 and complete left metacarpal L12763 from the QSM, and, secondly, distal left humerus and complete left metacarpal from the QSM or bed 3aS, both numbered L41704 (Fig. 13). The two metacarpals have lengths and least transverse thicknesses of 268 x 32,0 and 253 x 34,8 respectively, and are less short and thick than in Pleistocene and Recent Bovini (Fig. 14). A right calcaneum and left astragalus, L40773 from bed 3aS, are also associated with one another. Other bovine limb bones are: L9992—proximal right radius probably from PPM L21306—proximal right radius from QSM (Fig. 13) FOSSIL BOVIDAE FROM LANGEBAANWEG 239 Fig. 12. Occlusal and lateral views of left M3s. From the left: Mesembriportax acrae L46592; Bovini, presumably Simatherium demissum 50612, L50663. Scales in millimetres. L6094, L9740—distal right radii probably from PPM L20445—complete right metacarpal, probably from QSM, with length and least transverse thickness of 225 x 35,0. L12279—left proximal metacarpal probably from PPM The distal humeri have slanted condyles, a fairly deep hollowing for the lateral humeroradial ligament, a wide distal end of the lateral surface behind the ridge demarcating the hollow for the humeroradial ligament, and a coronoid fossa which is moderately deep. Both proximal radii show a moderately large lateral tubercle which is set low, a rim on the medial side of the medial facet, and a lateral facet which sticks well forward anteriorly. L9992 has an angled edge to its medial facet, whereas L21306 is rounded, and L9992 is wider front to back across its medial facet than L21306. The anterior flanges on the distal radii are wide apart and not strongly developed, the posteromedial facet for the scaphoid at maximum flexion is poorly hollowed, the anterior facets for lunate and scaphoid and the posterior one for the lunate are, however, better marked, that part of the articular facet for the cuneiform which lies on the radius is quite _ wide, and the distal end as a whole is swollen in side view. The articular surfaces of the metacarpal proximally do not fill the whole available area at the top of the bone, and the edge of the magnumtrapezoid facet does not have a single, clearly angled anteromedial corner. Distally there are poor hollows on the anterior surface above the condyles. The outer edges of the condyles are nearly parallel to one another. 240 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 13. Limb bones of Simatherium demissum, all shown as of the right side. A. Anterior view of distal humerus L41704. B. Proximal articular surface of metacarpal L41704. C. Proximal articular surface of radius L21306. D. Anterior view of same radius. Anterior sides of B and C towards the base of the illustration. a = ridge behind hollow for lateral humero-radial ligament, b = coronoid fossa, c = lateral tubercle, d = medial rim of medial facet, e = lateral facet, f = medial facet, g = magnum- trapezoid facet, h = unciform facet. Measurements Measurements on the three best preserved skulls are: L45001 L30174 #130175 Total length of horn-core .. . 299 292 348 Anteroposterior diameter at base of ent i core ; A ; : : 74,8 59,6 70,0 Lateromedial diameter at base ofhorn-core 58,2 50,4 64,5 Minimum width across lateral sides of horn PIR GUCET Sie Oe ie ae oe 187 — 192 Width across lateral edges of aipeor itl foramina ; : : ; 90,4 — —- Minimum width across temporal lines on skull roof ; : : ine 63D —_ 66,5 Skull width across mastoids behind external auditory meati fe ed wa 170 — Occipital height from dorsal edge of foramen magnum ew . 61,4 — — Width across anterior tuberosities of basi- OCCHDIEA. 0 sg mer ng eee ey A: 28,8 23,0 Width across posterior tuberosities of basi- occipital . ; : : : 60,9 c. 64,0 48,5 Tooth measurements of these and other specimens in middle wear are given in Table 2 FOSSIL BOVIDAE FROM LANGEBAANWEG 24] oe Least peeine 60 4 thickness go e ae ee ow | ae : @ & 50 > " aa f S e S s 4 S 100 200 300mm Fig. 14. Proportions of bovine metacarpals. X = E Quarry Langebaanweg, e = Pelorovis ?antiquus from Elandsfontein, S = extant Syncerus caffer, O = S. acoelotus BK 1952.218 from upper Bed II Olduvai Gorge, at present in Nairobi, + = Pelorovis oldowayensis asso- ciated skeleton from upper Bed II Olduvai Gorge. ANNALS OF THE SOUTH AFRICAN MUSEUM 242 v6 6 LE ZTE 0°76 JOMO} 12] 9‘TS Orr 2 €‘LE L‘€Z CsI Z'€9 47 ZTE 600T JOMO]T 1J2] 9°19 v'Lb v'LE 9°61 L'8€ ace L'%6 JOMO] JP] ELO9VT LOOSVI SELIVI G6OLIVI Iv8ttel 609cET SLIOCT 0°Or OLE (GNIC6 CSE 6LZ ‘16 JOMO] IPT (KG O16 2 IOMO] yysII yer = t ys = | S07 = rd WS “LEE = GCC scl 0°8¢ €°8E 9°87 8°76 JOMOT 12] ‘6 7 It 9°0€ 9°6€ 9°67 7 87 9°78 a Joddn JoMOj yysiI 49] uoIjUEp snonpidep = aN 39] YISUI] [esN}D90 [OOSHT y°S9 L‘9v 0‘Or 8°77 g*LI 16S €‘Ep TEE L‘Sol JOMO] SII 7 AIAVL Eale 6 £7 Eve 6 VE TL6 9 Joddn 14311 Joddn yYysI4 8ZE8TI T98S71 198S71 OOVECT crsi'l ‘SUOI}IJUSP SUIAOG JO S}UOWOINSPO|] JOMO] JST Joddn IJ9] evs SI MOjog WIdep snweYy TA MOOG YIdop snueyY ®q MOloq YIdop snuIeY bd YISUdN| [esn]s0O * Zdq Yul [esn[s909 bd-Zd YIsUI] [esn[I9O * €W YIsuUs] [esn[s9O ZW Ui8uU9] [esn[s909 EW-TW Yisu9] [esnfs99 FOSSIL BOVIDAE FROM LANGEBAANWEG 243 Comparisons Simatherium demissum is obviously similar to other African and Eurasian bovines of Miocene and Pliocene age, and comparisons must be made with the following forms: Leptobos syrticus Petrocchi, 1956, from Sahabi. It is a primitive species but its generic attribution seems reasonable. Parabos Arambourg & Piveteau, 1929, containing one or more species from the Ruscinian of France. Among these are the boselaphine-like P. cordieri Christol (Gervais), 1852, from Montpellier and P. boodon (Gervais), 1853, from Perpignan, generally considered to be a later site (Guérin 1975). In Europe Parabos predates Leptobos, but the primitive L. syrticus in North Africa is likely to be still older. Proamphibos Pilgrim, 1939, containing one or more species from the Siwaliks. P. Jachrymans Pilgrim (1939: 271) is the type species from the Tatrot Formation and P. kashmiricus Pilgrim (1939: 278) is apparently a more advanced Hemibos-like species, either contemporaneous or later. Simatherium demissum will not be compared with P. kashmiricus. Ugandax gautieri Cooke & Coryndon, 1970, from deposits of unknown age in the Kazinga Channel of the Kaiso Formation. An unnamed species of Ugandax from the Hadar Formation. Simatherium kohllarseni from Laetoli. Simatherium demissum differs from Leptobos syrticus by being slightly pealion its horn-core cross-section less clearly triangular (it has more expansion on either side of the anterior keel), horn-cores less divergent, horn-cores curved backward, supraorbital pits wider apart, braincase roof more strongly angled (in the Sahabi bovine it is almost horizontal), less strong temporal ridges and less of a temporal fossa, occipital surface probably less low and wide and certainly with a less long top surface, mastoids smaller and without deep pit-like excavations dorsomedially, median vertical ridge on occipital, and basioccipital more triangularly shaped with a more definite valley between the posterior tuberosities, narrowing more abruptly in front of the posterior tuberosities, and with smaller anterior tuberosities. The two forms are similar only in having a strong anterior keel, dorsal orbital rims which project well and horn-core insertions above the back of the orbits. The last resemblance is certainly a character which is primitive in bovines. Most of the differences can. be recog- nized as being primitive or advanced in one or other of the species, and one is aware of two bovine lineages showing different characters advancing at different ‘rates. More advanced Leptobos, like L. falconeri and the European Villa- franchian species, have an almost horizontal braincase roof and very strong temporal ridges, and it is clear that the Langebaanweg bovine cannot be assigned to Leptobos. A cranium of Parabos cordieri and a cranium and skull of P. boodon (Depéret 1890, pl. 7 (fig: 4); Piveteau 1961: 1052 fig. 145) were examined in the Institut de Paléontologie, Paris, and some horn-cores of P. cordieri in Basle. 244 ANNALS OF THE SOUTH AFRICAN MUSEUM Simatherium demissum differs from the skull of P. boodon by its shorter horn- cores, anterior keel stronger and posterolateral keel weaker, horn-core cross- section not so clearly triangular in shape, horn-core divergence more strongly diminished distally (i.e. a more curved course in anterior view), horn-cores probably inserted closer behind the orbits, horn-cores curve slightly backward, frontals not raised between the horn-core bases, better marked temporal ridges, more development of a rugose raised area posteriorly on the braincase roof, occipital surface with a horizontal top edge, teeth probably larger, styles larger on upper molars, slightly more localized ribs on anterior parts of lateral walls of upper molars. It seems likely that the facial part of the P. boodon skull has been displaced downward and that this has reduced the apparent inclination of the horn-cores in profile. Originally they would have been more upright than in S. demissum. The large cranium of P. boodon from Perpignan has horn-cores with no posterolateral keel at all, more strongly diminished distal divergence and a slight degree of backward curvature, flatter frontals, horn-cores set more closely to the orbits, better temporal ridges, braincase roof less slanted, and is thus less different from Simatherium demissum than is the complete skull. It is possible that its frontals’ morphology is more representative of the species than that of the complete skull, and its horn-cores are more uprightly inserted than in the Langebaanweg bovine. The absence of a posterolateral keel is puzzling. The basioccipital is preserved on this specimen. It is unlike that of S. demissum by being wider anteriorly, not narrowing abruptly just in front of the posterior tuberosities, not having a narrow and deep conta groove between them, and not having a central longitudinal ridge. The cranium of Parabos cordieri is well removed from resemblance to S. demissum by its smaller size, and horn-cores which are less divergent basally and less robust. It also has a more clearly triangular cross-section and sometimes a better marked posterolateral keel. It agrees with the South African species in having little compression of the horn-cores, horn-cores set widely apart and close above the orbits, temporal crests present, a rugose area at the back of the braincase roof, mastoids not very large, and basioccipital narrowed immediately in front of the posterior tuberosities. The Basle examples have a low inclination of the horn-cores. It is interesting that the teeth of Parabos appear to have about the same level of brachyodonty and occlusal complexity as S. demissum. One can see Parabos cordieri to P. boodon as a bovine lineage evolving in Europe before the appearance there of Leptobos which displaced it. From the facts just given, one can sum up the differences of Parabos as a whole from S. demissum. Parabos has variable development of the posterolateral keel and generally an anterior keel, whereas S. demissum had already lost its supposed ancestral posterolateral keel, but has a strong anterior keel. Again, S. demissum has evolved horn-cores with strong basal divergence and distal recurvature. Their low inclination is matched only by the Basle example of P. cordieri. The FOSSIL BOVIDAE FROM LANGEBAANWEG 245 upper molars on the Parabos skull had evolved rather small styles but retained less localized anterolateral ribs than S. demissum. Such differences seem sufficient to show that Parabos must have been a different lineage, and this is compatible with the geographical circumstances. Simatherium demissum differs from Proamphibos lachrymans as from Lepto- bos syrticus and Parabos by its shorter and more robust horn-cores, their less regularly triangular cross-section, greater basal divergence, and a more curved course in anterior view. These can all be considered advanced characters. Other differences are horn-cores inserted less widely apart in S. demissum, supra- orbital pits less wide apart, more projection of the orbital rims, horn-core insertions probably less far behind the orbits, occipital surface lower and wider, presence of a central longitudinal ridge on the basioccipital, and smaller anterior tuberosities. S. demissum agrees with P. /achrymans in the degree of mediolateral compression and a strong anterior keel on the horn-cores, in the inclination of the horn-cores in side view, and even in the slight swelling of the cross-section laterally to the anterior keel. It also agrees in the not very large mastoids and many cranial characters. Simatherium demissum is more advanced than the Ugandax gautieri holo- type by its larger size, horn-core insertions wider apart, horn-cores more divergent basally and with a more curved course, probably a shorter braincase and a more triangular basioccipital. It is more primitive in retention of a strong anterior keel, horn-core insertions closer to the orbits, a rugose area on the braincase at the top of the occipital, and smaller anterior tuberosities of the basioccipital. Other differences lie in its transverse constriction of the basi- occipital immediately in front of the posterior tuberosities, the valley between the posterior tuberosities of the basioccipital, the central longitudinal ridge less pronounced on the basioccipital and the smaller mastoids. It is similar in its short to moderate horn-core length, degree of mediolateral compression, lack of a very orderly triangular cross-section, inclination of the insertions, and _ most of the other cranial characters. Thomas (19798, pl. 2 (figs 1, 3, 5)) has identified bovine teeth at Lukeino as Ugandax cf. gautieri. They are very similar to those from ‘E’ Quarry but a little smaller. The Afar Ugandax is generally a more advanced species than U. gautieri. It may have occurred later, although the time level of U. gautieri is unknown. S. demissum has many differences from the Afar Ugandax sp. It is more primitive in its horn-cores with a stronger anterior keel, usually more backwardly curved, the horn insertions closer behind the orbits, the braincase less low and wide, _ the smaller anterior tuberosities of the basioccipital, and the occlusal surface of the cheek teeth less complicated both in the outline of the central cavities and in the absence of constrictions across the lateral lobes of the lower molars. It is more advanced in the greater divergence of its horn-cores, their insertions being wider apart, the occipital with perhaps a straighter top edge, its longer molars (Fig. 15), the relatively shorter premolar rows (Fig. 16) and in the metaconids of P; and P, perhaps less diagonally slanted backwards. Other 246 ANNALS OF THE SOUTH AFRICAN MUSEUM Elandsfontein KXXXXX XK X X Olduvai middle a upper Bed I X KX MX X x Shungura Fm. member 6G WOK X OOK XX OK KORE Oe Shungura Fm. member C XXX XOX ROOHOKX Shungura Fm. member B ja Sette Hadar Formation x OOK XX Langebaanweg ROOK XX a a a ia a) a a i nna a) (arama nS Si SSCSCS*C~S~S 30 40 50 60mm Fig. 15. Occlusal length of bovine M;s. Dotted Langebaanweg readings are QSM, others are PPM. Elandsfontein sample is Pelorovis ?antiquus. The four largest readings for Olduvai and for Shungura member G are likely to be P. oldowayensis. Other Olduvai and Omo readings are Syncerus, and Hadar Formation ones are Ugandax sp. Not until Olduvai does Syncerus become comparable in length with the Langebaanweg bovine. differences from the Afar species are the strong projection of the orbital rims, more of a median vertical ridge on the occipital, smaller mastoids, and a longer -basioccipital with a central longitudinal ridge and a valley between its posterior tuberosities. A comparison of skull measurements between the Afar and ‘E’ Quarry bovines is shown in Figure 17. Simatherium demissum differs from the Berlin S. kohllarseni by its smaller size, anterior keel, horn-cores inserted less extremely widely apart, lower inclination of the horn insertions, less divergent horn-cores, horn-cores with a slight backward curvature which is not confined to the basal sector, a less irregular surface of the horn-core, no consistent shallow longitudinal groove running along the horn-core, frontals not transversely arched between the horn-core bases, clearer temporal ridges and no temporal fossa below over- hanging horn-core insertions. It differs from the more recently recovered example by an anterior keel, lower inclination of horn-cores, less divergent horn-cores, horn-cores with slight backward curvature, a less irregular surface of the horn-core and no deep longitudinal grooves anteroventrally, more of a median vertical occipital ridge, and a smaller mastoid. It is similar to both these Tanzanian fossils in its short robust horn-cores, without compression, loss of a triangular cross-section, horn-cores inserted FOSSIL BOVIDAE FROM LANGEBAANWEG 247 9q 4 Length 70 50 Length M,-Mz 60 80 100 120 mm Fig. 16. Occlusal lengths of lower premolar and molar rows in Bovini. X = bed 3aN Lange- baanweg, X = QSM Langebaanweg, e = Pelorovis ?antiquus from Elandsfontein, n = P. antiquus from Naivasha, Kenya, + = P. oldowayensis from upper Bed II Olduvai Gorge, S = extant Syncerus caffer, S = S. caffer from latest Pleistocene Kibish Formation, Omo (see Gentry & Gentry 1978: 308-322 for this and other bovine fossils), O = S. acoelotus from upper Bed II Olduvai Gorge, O = S. ?acoelotus from Shungura Formation member G, m = Syncerus sp. from Melkbos, a = Ugandax sp. from Hadar Formation. Lower diagonal line = 50 per cent, upper one = 66,7 per cent as in Figure 3. widely apart, the slope of the braincase, temporal ridges, a raised area on the brain roof at the top of the occipital, occipital proportions similar, a flat-topped - occipital with side edges formed by the nuchal crests, basioccipital with a central longitudinal ridge in the area just behind the level of the anterior tubero- sities, and a very deep, short groove between the posterior tuberosities. A Simatherium such as the Langebaanweg species could be ancestral to the Laetoli one. Nearly all its differences from the Laetoli examples can be thought of as more primitive, and this contrasts with the comparisons previously made with other bovines. It is interesting that at Langebaanweg the horn-core inser- tions are more inclined. Either this character is primitive for the genus or it - could be an indication of regional specialization at an early time level. Tribe Reduncini Kobus A. Smith, 1840 Type species Kobus ellipsiprymnus (Ogilby, 1833). 248 ANNALS OF THE SOUTH AFRICAN MUSEUM Anteroposterior diameter at horn core base 1 I Mediolateral diameter at horn core base ); Width across supraorbital pits Width across horn bases Minimum width across temporal ridges ie nas Skull width across mastoids : : se Fa cg 3 \ \ e . ee 60 80 100 120 140 Occipital height Width across anterior tuberosities of basioccipital Js ae \, Width across posterior tuberosities of basioccipital Fig. 17. Percentage diagram of skull measurements in Bovini. The standard line at 100 per cent is the holotype of Proamphibos lachrymans (M 26576, cast) and readings on other lines are expressed as percentages of their values on the standard line. The second continuous line is the holotype of Simatherium demissum, and the dashed line is Ugandax sp. AL 194-1 from the Hadar Formation. The African forms are wider across the temporal ridges, have lower occipitals, and have basioccipitals with relatively narrow anterior tuberosities. Generic diagnosis Larger sized reduncines; horn-cores usually long, their bases sometimes curving backward instead of being concave anteriorly, usually with a flattened lateral surface but no tendency towards a flattened posteromedial surface; frontals sometimes with a small system of internal sinuses. Remarks A majority of the Kobus-like horn-cores and possibly the back of a reduncine cranium from ‘E’ Quarry, Langebaanweg, belong to a short-horned species which is given a new name below. Two problems surround these remains. A major one is that dentitions apparently associated with these horn-cores are very unlike other known reduncines. If the association could be proved beyond question, i.e. by the recovery of a skull with both horn-cores and teeth, then a new generic name would be needed for this species. Meanwhile it is put into Kobus. A lesser problem is that the two most completely preserved crania with horn-cores show a small suite of character differences from the other horn-cores and cranium. They may represent another species and are described under the heading Kobus sp. 2. Kobus subdolus sp. nov. Figs 18—20. Holotype L30878—right horn-core with part of frontal, dorsal part of orbital rim and supraorbital pit (Fig. 18). FOSSIL BOVIDAE FROM LANGEBAANWEG 249 Referred material The main specimens assigned to this species are as follows: From bed 3aS: L41387—frontlet with nearly complete horn-cores. Index 48,7 x 41,0, minimum width across lateral sides of horn pedicels 92,9, width across lateral edges of supraorbital foramina 45,7 L40248—left horn-core, index 43,8 x 37,5; same individual as right horn- core L40371 L40049—right horn-core, index 41,5 x 37,0 L40051—left horn-core, index 39,2 x 33,6 Probably from bed 3aS, but a few possibly from QSM: L1847, L2609, L6076, L10672—left horn-cores, the second and fourth with indexes 50,2 x 41,2 and 43,5 x 35,6 L2611, L2612—right horn-cores, indexes 46,5 x 35,9 and 41,4 x 36,6 L2604, a reduncine cranium preserved only from a level behind the horn bases, may be conspecific with the horn-cores (Fig. 19). Its measure- ments are: Maximum braincase width . : ; : : 80,2 Minimum width across temporal lines on skull roof : oe 28,1 Skull width across mastoids behind external auditory meatus . 9253 Occipital height from dorsal edge of foramen magnum .__. 44.0 Width across anterior tuberosities of basioccipital . .. 31,0 Width across posterior tuberosities of basioccipital. . . 3559 From bed 3aN: L40726—left horn-core, index 40,2 x 33,0; same individual as right horn-core L40728 L40870—left horn-core, same individual as right horn-core L40872, index c. 50,5 x 43,0 L46062—right and left horn-cores, index 50,3 x 46,0, length c. 190 L30879—index 46,7 x 40,3, L41754 45,7 x 39,8, L46044 41,1 x 36,3, L46045 43,0 x 40,3, L46048, L46060 47,7 x 41,2 and length 197, L46069 50,9 x 43,0 and length 159—left horn-cores L30029—index 47,7 x 36,1, L30878 44,4 x 37,6, L33383 47,7 x 40,0, 33746 43,3-<-35,0, LA0873. 46,0 x 32,7, . LAI/55 -42,6 x 36,5, L46049 39,7 x 35,6, L46070 44,4 x 36,6 and length 172, L46071 54,4 x 46,3 and length 178—right horn-cores L32163—partial occipital surface and basioccipital — Locality The holotype comes from bed 3aN of the PPM. The provenance of other specimens has already been given, and it can be seen that the species is not definitely known from the QSM. 250 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 18. Kobus subdolus. 30878, holotype. Right horn-core in lateral and anterior views. Scale = 25 mm. Fig. 19. Kobus subdolus. 2604, cranium. From the left: occipital view, lateral view, ventral view. Scale = 25 mm. FOSSIL BOVIDAE FROM LANGEBAANWEG 251 Diagnosis An antelope about the size of Kobus kob or K. leche. Horn-cores short, little compressed mediolaterally, with a flattened lateral surface, their widest mediolateral diameter lying rather anteriorly and the cross-section narrowing behind this level to a posterolateral edge which does not quite assume the form of a keel. Deep grooves run longitudinally just medial to this posterior edge. A minority of horn-cores show transverse ridges which are V-shaped and close together. Horn-cores inserted above the orbits, at a low inclination and close together, not very divergent and with divergence perhaps lessening distally. They show little curvature but the tips curve forward slightly. Anterior edge of horn pedicels sometimes set uprightly in profile and thus at an angle to the anterior edge of the horn-core proper. Postcornual fossa small and deep, frontals little raised between horn-core bases, supraorbital pits large. Etymology The specific name taken from the Latin subdolus, is somewhat deceitful, and refers to the species not having teeth which are clearly reduncine. Remarks The most striking features of Kobus subdolus horn-cores is that they are so short. Transverse ridges may be seen on 5 out of 17 left horn-cores and 3 out of 14 right ones. The tendency towards an upright anterior edge of the horn pedicel can be seen in 5 among 16 and 6 among 15 right horn-cores. It is possible that those from bed 3aN are larger than from 3aS (Fig. 20), but the difference is not statistically significant, even if the two 3aN horn-cores of ‘Kobus sp. 2’ are, indeed, a separate species. The reduncine cranium 12604 is presumably conspecific with the horn- cores. There is no central indentation in the parietofrontals suture as it passes across the top of the skull, the temporal lines approach fairly closely posteriorly, the braincase widens posteriorly, its roof is inclined in profile and not curved downward posteriorly, the nuchal crests are not very strongly developed, the - occipital surface is fairly high and narrow and faces backward, its edge is not evenly rounded, the median vertical ridge of the occipital appears to have been present and had shallow depressions flanking it dorsally, the fairly small mastoids lie mostly within the bounds of the occipital surface and do not have a pronounced dorsal or ventral rim, the anterior tuberosities of the basioccipital are very large, longitudinal ridges pass backward from the anterior tuberosities and converge posteriorly (giving the appearance of a transverse constriction _ across the centre of the basioccipital), the posterior tuberosities take the form of posterolaterally directed transverse ridges, and the auditory bulla is large and inflated. The large size of the auditory bullae is noteworthy. Another fragmentary occipital surface with part of the basioccipital, L32163, appears to be conspecific. One or two limb bones from ‘E’ Quarry are identified as possibly reduncine. L30392 from bed 3aN includes a distal right humerus with much of its shaft 252 ANNALS OF THE SOUTH AFRICAN MUSEUM and an associated proximal right radius and reduncine dental remains. Two distal humeri likely to be conspecific are L31937B and L32629 from bed 3aN and two further proximal radii are L31937C from bed 3aN and L24932 from the QSM (Fig. 21). The humeri show the following characters: l. Dep 3. 504 Mediolateral 40 50 The distal part of the lateral surface is wide behind the ridge bounding the hollow for the lateral humeroradial ligament. | The ridge on the distal part of the lateral side for the origin of the extensor carpi radialis is marked, especially in L31937B. The coronoid fossa is deep in L30392 and L32629 but not in L31937B. In L31937B and L32629 the condyles are slanted but this is not true for L30392. . The medial groove on the condyles is deep in L30392 and L32629 but shallow in L31937B. The hollow for the lateral humeroradial ligament is deep in L30392 and L32629 but shallow in L31937B. The medial condyle passes high into the coronoid fossa in L30392 and L32629 but this is less clear in L31937B. diameter Anteroposterior diameter 20 30 40 50 60 mm Fig. 20. Basal diameters of reduncine horn-cores. O = bed 3aS Langebaanweg, X = bed 3aN, X = Kobus sp. 2, + = Kobus sp. from Sahabi, r = K. ?porrecticornis from Baard’s Quarry, S = K. sigmoidalis from members C-F inclusive of the Shungura Formation. Upper diagonal line = 100%, lower one = 66,7 % as in Figure 3. A cross-section of the right horn- core L2612 is shown, taken 20,7 mm above its base, with lateral side to the left and anterior side to the base. Scale = 10 mm. FOSSIL BOVIDAE FROM LANGEBAANWEG 253 Fig. 21. Limb bones of Reduncini, Alcelaphini and Neotragini, all shown as of the right side. Anterior sides of proximal and distal views are towards the base of the illustration. A. Anterior view of distal reduncine humerus L24932. B. Proximal articular surface of associated radius L24932. C. Proximal articular surface of alcelaphine radius L41482. D. Proximal articular surface of associated metacarpal L41482. EE. Distal articular surface of neotragine tibia L41684. F. Proximal articular surface of associated metatarsal L41684. G. Anterior view of distal neotragine humerus L40787. H. Proximal articular surface of neotragine radius L40088 with part of its back edge taken from L40021. a = medial malleolus, b = fibula facets, c = main facet for naviculocuboid, d = main facet for ectocuneiform. For the first four characters two or all three of the humeri are unlike alcelaphines, but for the last three characters L30392 and L32629 resemble them. None of the characters is definitely unlike later Reduncini, and although the second and third could be held to resemble Tragelaphini, the last three do not agree very well with that tribe. Most characters of the proximal radii are unlike Alcelaphini: the lateral facet is broader and longer (not in L31937C), not pointed anteriorly, and its _ posterior edge is not stepped forward from the level of the back edge of the medial facet. The lateral tubercle is small and low. The indent in the back edge of the medial facet is less deep than in Alcelaphini, and the anteromedial part of that facet is more extensive in L30392 (but not in L31937C or L24932) than in Alcelaphini. It is interesting that there is no medial rim on the medial facet —unlike later Reduncini or Tragelaphini. Comparisons The Langebaanweg fossils agree well with a frontlet and three horn-core pieces from Sahabi, now in Rome. The Sahabi horn-cores are about the same size, not very compressed mediolaterally, with some flattening of the lateral surface, hardly any backward curvature basally but with some upward curva- ture towards the tips, and their widest mediolateral diameter lying rather anteriorly. The supraorbital pits are close together and may have been rather large, the dorsal part of the orbital rim is strongly projecting, and the parieto- 254 ANNALS OF THE SOUTH AFRICAN MUSEUM frontals suture is transversely straight across the top of the skull. The Sahabi horn-cores although short are less short than the Langebaanweg ones and they diverge more. There are no transverse ridges on the horn-cores and no angling of the front edge of the pedicel on the line of the horn-core proper, but the condition of both these characters is variable in the Langebaanweg fossils. The cast of a right horn-core, BM(NH) M 8200, from Wadi Natrun is a little smaller than the Sahabi and Langebaanweg ones but similarly rather short, little compressed and with its widest transverse diameter lying at rather an anterior level. Measurements on the Sahabi frontlet are: horn-core index 47,4 x 42,0, horn-core length 218, minimum width across lateral sides of horn pedicels c. 102, and width across supraorbital pits probably near 50. The basal index for another Sahabi horn-core is 49,0 x 41,5, and for the Wadi Natrun cast 38,8 x 30,6. The length of the Wadi Natrun cast is c. 150. Kobus subdolus is unlike the early reduncine K. porrecticornis of the middle Siwaliks by its larger size, shorter and less divergent horn-cores which are without backward curvature and are inserted at a lower inclination, and by its occasional possession of transverse ridges. K. porrecticornis is discussed again in the account of Baard’s Quarry fossils. Compared with Redunca darti of Makapansgat Limeworks, the Langebaan- weg species shows closer insertions of the horn-cores and probably closer supraorbital pits as well, more oblique insertions, the tendency to have a front edge of the pedicel at an angle to the front edge of the horn-core, and horn-cores with a completely different cross-sectional shape. This last feature arises from the horn-cores having a flattened lateral surface, narrowing posteriorly in cross-section so that there is an approach to having a posterior keel, having no hollowing or flattening posteriorly or posteromedially at the base, and being slightly compressed mediolaterally. The Langebaanweg species also differs in having no central indentation of the parietofrontals suture, a less pronounced ventral rim of the mastoid, and possibly better marked hollows flanking the median vertical ridge of the occipital surface. It must be a different lineage from R. darti. Kobus subdolus differs from undescribed reduncine crania and horn-cores from the Hadar Formation, Afar, by its short horn-cores with flattened lateral surfaces and a slight degree of mediolateral compression, less narrow mastoids, larger bullae, larger anterior tuberosities of the basioccipital, less narrow posterior tuberosities of the basioccipital, and no anterior indentation of the parietofrontals suture. It is clear that the Afar species is a different lineage not very closely related to the Langebaanweg form. Compared with Siwaliks reduncines from the Pinjor Formation which may be assigned to Sivacobus palaeindicus (BM(NH) 17437, 39559, M 487, and M 2402 which Gentry & Gentry (1978: 337) discussed), Kobus subdolus shows horn-cores with a flattened lateral surface, more mediolateral compres- sion, their widest mediolateral diameter lying more anteriorly, and more of an FOSSIL BOVIDAE FROM LANGEBAANWEG 255 approach to a posterolateral keel. The cranium L2604 also has weaker temporal ridges, no central indentation in the parietofrontals suture, not such pro- nounced posterior widening of the braincase, and not such a strong median vertical ridge on the occipital. Despite a number of differences such as shortness and less mediolateral compression (Fig. 20), the ‘E’ Quarry horn-cores recall the species Kobus sigmoidalis and K. ancystrocera from the Shungura Formation, Omo, by their cross-section with a flattened lateral surface and an approach to a posterolateral keel, the very low insertion angle in side view, and the manner in which the front of the pedicel is angled on the axis of the horn-core proper. The two Omo species seem to belong to a ‘modern’ group of reduncines related to the living waterbuck and Central African lechwe and to living and fossil kob. On the Langebaanweg cranium L2604 the occipital is higher and narrower than in the Shungura species, which one could imagine to be a primitive character befitting a more ancient species. However, the evidence of the redun- cine teeth at Langebaanweg must be considered before making a final assessment of likely phylogenies. The question of reduncine teeth at Langebaanweg In 1975 the most northerly exposures of bed 3aN were excavated and among the bovid horn-cores there was the following representation of individuals: Tragelaphus two, Mesembriportax one, Simatherium three, Kobus six, Damalacra one. No teeth were found which were unquestionably reduncine in appearance, despite the predominance of Kobus horn-cores, but the com- monest group of teeth, coming from at least seven individuals, appeared to be most like Tragelaphini. They agreed with teeth found in earlier years in ‘E’ Quarry and assigned to Tragelaphini, and showed the following characters: 1. Dentition rather low crowned . like Tragelaphini, also a primi- tive character 2. Hardly any development of ribs like Tragelaphini between the styles on the lateral walls of the upper molars 3. No basal pillars on upper molars like Tragelaphini and only tiny ones on lowers 4. Anteromedial lobe of upper a primitive character molars is not connected in early wear to the junction of the posteromedial with the antero- lateral lobes 5. Premolar row moderately long . intermediate between Reduncini and Tragelaphini (Fig. 22) 6. P® and especially P? large in com- like Tragelaphini parison with P* 7. Central cavities of P? and P* sited like Tragelaphini, also a primi- rather anteriorly tive character =," * 256 40 30 20 ANNALS OF THE SOUTH AFRICAN MUSEUM 40 50 Length M,-Mz 60 70mm Fig. 22. Occlusal lengths of lower premolar and molar rows in Reduncini and Tragelaphini. X = bed 3aN Langebaanweg, a = extant Tragelaphus angasi, S = extant T. spekei, e = extant Kobus ellipsiprymnus, O = extant K. kob, r = extant Redunca redunca. Lower diagonal line = 50%, upper one = 66,7 % as in Figure 3. 10. ee 13% 14. Small goat folds present on lower molars, presumably not a primi- tive character Medial walls of lower molars with slight outbowings only Lateral lobes pointed but not narrowed in their lateral part Medial wall of back lobe of M; set more laterally than medial wall of anterior lobes Hypoconid of P, often not pro- jecting laterally and never with a deep valley in front of it Paraconid-metaconid on _ P, usually not fused but metaconid growing forward A backwardly directed flange is usually present on P, metaconid a tendency towards Reduncini like Tragelaphini—flatter than in 7. spekei and buxtoni* like Tragelaphini like Reduncini like Tragelaphini, also a primi- tive character like Tragelaphini and Redunca arundinum frequent in Tragelaphini * Extant Tragelaphus spekei and buxtoni have less flattened walls than T. imberbis, strepsi- ceros and eurycerus and Taurotragus oryx. Tragelaphus angasi and scriptus are probably intermediate. Se _ ee -S ers. 251 FOSSIL BOVIDAE FROM LANGEBAANWEG Ee — rs ———— 8°8 TOI 8°6 rOI "+ gd BUI] TeSNIDIO 601 ee == — " * gd YSU] [esN[I9O 0OE = — = j rd-zd YISUI] [esN{I90 (SHi\h LI gc 6°91 " * ZW Yi8u9] jesn[s90 €‘Ly 7 6r ply? S9p°2 ° WW YisuU9] [esn[s90O Nee Nee Nee WSO OS8cel LEEIEIL 6SpOel coOlell 0°97 €“LZ CLC OL =; ed ee €“LZ ae = — -- -- * £~w Moye ydep snuey 74 oar L°v7Z YANG = = =e BONSG aa — -- = — * Ty MOpOQG Ydep snueYy EC a = 0°61 ae as ‘ST SVT == = — — = - 8g Mojaq ydep snuey vil ee Lc a) Sb Ge ra EZ ay Oe] wal VC = 7 + Fg yysusy] [esnjooO 76 ie 78 €°8 ¢°8 cae = L3 = — — —- — "+ &y yySua] pesnjooO y‘TE ae 60. 01602 O0CE = meee “Pile < = Se Re aa aa "dd Y}3u9] [esnjs9O SEZ SEZ 6°€7 G6 SAG ee C356 eee Ober tee aCle 6° as 610C 7° SAT YASUI] [esN[I9O GS S°LT OO} -°_ LV ‘Ol 781 = 0° 6°96 E“ys S‘6r Ors aE ra ae 9°€S L’8v 0°sS pw Nee See Nee Nee Ne®e Nee Nee Nee Nee WSO See L909VT IVEIVI €OCEET OS8ZET COLZET SI6GIET 990IET COEOET OO00ET 809771 SO9STT LE69ET CLSZT "IBOM [PPI 932] Ul EOZEET PUL 99DTET “OOODET ydeoxe ‘1eeM o[ppIw A[ies Ul o1e ASU] “(S1JO] POANSLOWTUN YPIA\ PO}BIDOSSe SJYBII JO IO) Opis Io] sy} JO oIe [JY ‘SUONMUOP suTOUNpeI Jo s}UDWOINsKO/ 8°91 6 v1 eal = 6 SI : "SAL YIsUI] [esn[s9O = FWA yysus] [esnjs9O € aTav LL 258 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 23. Kobus subdolus. L32850, occlusal view of right upper tooth row. Scale = 10 mm. Fig. 24. Kobus subdolus. L15605, occlusal view of left lower dentition. Scale = 10 mm. Fig. 25. -Kobus subdolus. L32850, L46067, lateral views of right mandibles. Scale = 10 mm. FOSSIL BOVIDAE FROM LANGEBAANWEG 259 Measurements for these teeth are given in Table 3 and they are illustrated in Figures 23-25. Readings for occlusal length and metastylid crown height in M,s in early middle wear are: P7055); riecht 21,7 x 12,5 ESLIT) right «ce. 23,2:% 14,0 E20906 left 23,4 x 14,1 31916 left 25,6 x 17,4 Other readings of occlusal length are: L1843C left P, 13,3 probably PPM, L10292 left dP, 18,4 3aS or QSM. Some characters not mentioned in the list do not show a clear trend towards one tribe rather than the other. For example, the anterior part of the lateral wall of P, sometimes bends round to become almost transversely oriented as in Reduncini, and sometimes remains mainly anteroposteriorly directed as in Tragelaphini. Again, some of the mandibles, e.g. L32850, are rather shallow anteriorly while others, e.g. L30392, are deeper, and it is not possible to say which condition is more like Tragelaphini and which is more like Reduncini. For assessment of the tribal affinity of these teeth, much depends on the ‘direction’ of their evolution. In some cases, for example the first character in the list, one can feel confident that the brachyodonty of the cheek teeth is not only more like Tragelaphini than Reduncini, but also that it is more primitive than the high-crowned teeth of later Reduncini. In other cases it is harder to guess whether or not a condition is primitive without a prior knowledge of phylogeny. The only indication of changes within the span of ‘E’ Quarry deposits is that the measured mandible and maxilla from the QSM have relatively small molar teeth, but one cannot know that larger samples would validate this difference. The tooth assemblage cannot be assigned satisfactorily to more than one species and if this species should be Kobus subdolus, one tooth only (p. 223) is left to go with Tragelaphus. It is possible that these teeth show an early stage in the evolution of redun- cine characters. This implies that Reduncini had ancestors with tragelaphine-like teeth, and that at Langebaanweg recognizably reduncine teeth had not yet evolved. Hitherto it has seemed likely that Reduncini would have evolved from boselaphine ancestors (Gentry 19785: 567) although Pilgrim (1939: 21) and Simpson (1945: 272) were not definite about their boodont affinities. In the Siwaliks succession one finds small boselaphines (Protragocerus gluten, Mio- tragocerus gradiens) in the Chinji Formation, and then in the Nagri and Dhok Pathan Formations boselaphines of varying but generally larger size—Pachy- portax latidens, Selenoportax vexillarius, Miotragocerus punjabicus and Trago- _ portax salmontanus. Towards the top of the Dhok Pathan Formation, or shortly afterwards, these overlap or give way to the early bovine Proamphibos, Redun- cini, and the supposed hippotragine teeth assigned to Sivatragus. Undoubted boselaphines are rare in the Tatrot Formation (Pilgrim 1939: 6). In this suc- cession it is easy to visualize the derivation of Tatrot reduncine teeth from boselaphine ancestors without the need for any diversion towards a tragelaphine- like morphology. Such a derivation clashes with that favoured by the Langebaan- “ser =" nent inetinetiieeemi 260 ANNALS OF THE SOUTH AFRICAN MUSEUM weg teeth in which characters 2, 3, 6, 9, 13 and perhaps 10 and 14 are unlike either reduncines or early boselaphines. A certain number of fossil reduncine teeth were examined for signs of descent from Langebaanweg-like ancestors, but the results were not conclusive. Among pre-Pinjor fossils from the Siwaliks were BM(NH) M 34568 a cast of a palate B 815 with left P?-M® and right P®-M? (Pilgrim 1939: 107, fig. 11a); M 15372, part of a left mandible with M,—M, and a fragment of M, (Pilgrim 1939, pl. 3 (fig. 13)); M 15385, various lower teeth; M 34569, a cast of part of a right mandible B 816 with P,—P, (Pilgrim 1939: 119, fig. 11b). All come from the Hasnot-Tatrot area and the first three were believed to come from the Tatrot Formation. The last may be from deposits equivalent in age or slightly younger than the type Dhok Pathan Formation but this is not certain. They could belong to Kobus porrecticornis, a species known from horn-cores and first known to occur in the upper Dhok Pathan Formation (Thomas 19795; Pilbeam et al. 1977: 687). (A left P,, M 15371 (Pilgrim 1939, fig. 11c), was earlier identified as reduncine but appears to be alcelaphine.) Other available reduncine fossils came from the middle of the Hadar Formation and from various parts of the Shungura Formation. The oldest reduncine teeth in Africa are from Lukeino and Mpesida, and these, too, could belong to K. porrecti- cornis or K. aff. porrecticornis (Thomas 19795, pl. 1 (fig. 6), pl. 2 (figs 6-7)). Compared with living reduncines, these fossils showed that earlier members of the tribe had lower crowned teeth, smaller basal pillars, less strong ribs on the lateral walls of the upper molars, smaller goat folds and less narrowed lateral lobes on the lower molars, longer premolar rows and the entoconulid more distinct from the entoconid on the back of P,. There is also interesting evidence for a closer approach of paraconid and metaconid on P,. The lower part of the paraconid on the nearly unworn P, and P, of B 816 has a flange growing backwards. Afar P,s have paraconid and metaconid growing very close to one another (e.g. AL 153-3, AL 156-1A and AL 167-5) and a Shungura example, L2-46B, shows a flange on the metaconid of P,. Among later redun- cines only Redunca arundinum and ontogenetically older Kobus ellipsiprymnus have an approach to fusion of paraconid and metaconid. In a sample of 31 Redunca arundinum P,s only 15 show close approach or fusion of paraconid and metaconid. They effect this more by forward growth of the metaconid than by backward growth of the paraconid, but in the Afar examples it looks as if backward growth of the paraconid is the more important. None of these characters is taken so far as to demand descent from a form with teeth similar to those at Langebaanweg. Even the approach or fusion of paraconid and metaconid on P, can evolve in boselaphines and its appearance in some reduncines need not imply affinity with the Langebaanweg teeth. Reduncine phylogeny Returning to the question of possible phylogenies, one sees that Langebaan- weg, Sahabi, and Wadi Natrun have one or more reduncine species with horn- FOSSIL BOVIDAE FROM LANGEBAANWEG 261 cores which would be good structural ancestors for modern Kobus and the extinct Shungura species K. sigmoidalis and K. ancystrocera. However, tooth morphology suggests that K. porrecticornis or K. aff. porrecticornis may be a better ancestor. Its teeth are known back to Lukeino and Mpesida and perhaps to the upper part of the Dhok Pathan Formation, so Langebaanweg could be earlier than these sites, or its reduncine retained primitive teeth late in geological time, or its reduncine teeth have been incorrectly identified. The most interesting aspect of the Langebaanweg reduncine fossils is that they are more like the aegodont Damalacra than the boodonts Tragelaphus sp., Mesembriportax acrae or Simatherium demissum. This is shown in their narrower skull proportions, absence of keels other than the trace of a posterolateral one, close insertions of the horn-cores, poor basal divergence of the horn-cores, deep postcornual fossa, inclined cranial roof, absence of strong temporal ridges, absence of a rugose area at the back of the braincase roof, and the good-sized anterior tuberosities of the basioccipital. Only in the teeth can even a few boodont characters be seen: basal pillars on the lower molars, retarded joining of the lobes of the upper molars in ontogeny, and the long premolar rows. However, poor development of ribs relative to styles on the upper molars and rather flat medial walls of the lower molars could be seen as aegodont characters. Some of the dentitions, e.g. L32850, show signs of transverse wear ridges across the dentine of the upper molars, and this too parallels caprines. One has to ask whether the Reduncini are properly to be considered as a boodont group, and what the evolutionary relationship is between those of the Siwaliks and of Langebaanweg, but for the present there seem to be no answers. Kobus sp. 2 Figs 26-27 Material Two reduncine crania with horn-cores from bed 3aN could belong to a smaller, more kob-like species of Kobus. They are L30391 and L31287, which -are both frontlets with almost complete horn-cores and separate, almost complete crania (Figs 26-27). Their measurements are: £30391. 131287 Anteroposterior diameter at base of horn-core. 43,8 37,6 Lateromedial diameter at base of horn-core . 3 39,7 30,1 Minimum width across lateral sides of horn pedicels Sor. 88,6 Width across lateral edges of supraorbital foramina. c. 45,0 52:1 Occipital height from dorsal edge of foramen magnum 37,6 37,8 Skuli width across mastoids behind external auditory meatus. : : E ) : : . 93,3 94,0 Width across anterior tuberosities of basioccipital . 22,9 28,6 Width across posterior tuberosities of basioccipital . 27,0 35:0 Description The horn-cores of L30391 differ from those of Kobus subdolus by being “ser” ~~ —— “se “s76a0r°" 264 ANNALS OF THE SOUTH AFRICAN MUSEUM the Nagri or Dhok Pathan stages of the Siwaliks succession. The Langebaanweg fossils are larger and relatively wider across the occipital surface, but share with the Siwaliks fossil a slightly inclined and anteroposteriorly convex cranial roof and temporal lines which approach closely posteriorly and are not very promi- nent. The anterior tuberosities of the basioccipital take the form of distinctive longitudinal crests in both the Siwaliks fossil and L30391. The Siwaliks fossil has a moderately sized and inflated auditory bulla, but the state of this character is unknown in Kobus sp. 2. It is possible that B798 belongs to Kobus porrecticornis, already mentioned in the discussion of reduncine teeth and to be mentioned again in the account of Baard’s Quarry fossils. If this were so, its non-bosela- phine characters (inclined braincase roof, poorly developed temporal lines) would align it with the ‘E’ Quarry reduncine fossils in throwing doubt on the close relationship or descent of reduncines from boselaphines. Tribe Alcelaphini There are two similar sized species of Alcelaphini at Langebaanweg. There is no doubt of their tribal affinity on account of a suite of characters including frontals set at a high level between the horn-core bases in comparison with the dorsal parts of the orbital rims, extensive internal hollowing of the frontals and a single large smooth-walled sinus extending into the horn-core pedicel, the shallow and narrow postcornual fossa, small supraorbital pits, occipital surface facing laterally as well as backward and having a median vertical ridge and flanking hollows dorsally, basioccipital with a central longitudinal groove having its sides formed by ridges behind the anterior tuberosities, hypsodont cheek teeth, basal pillars on the teeth small or absent, and short premolar rows often accompanied by reduction and disappearance of P,s. — A nearly complete skull belonging to one of the species allows assessment of its facial characters and they, too, are clearly alcelaphine as appears from its diagnosis. For both species there are associations between horn-cores and teeth and in some cases postcranial bones as well, but it was found impossible to allocate teeth or postcranial bones at species level. There is no doubt of the marked primitiveness of the teeth, and this is one of the most interesting charac- teristics of these alcelaphines. Damalacra gen. nov. Type species Damalacra neanica sp. nov. Generic diagnosis Moderate sized alcelaphines, a little smaller than the living Alcelaphus buselaphus or Damaliscus lunatus. Skull rather narrow as in those species and not wide as in Connochaetes. Horn-cores moderately long and without keels or transverse ridges. Horn-cores inserted fairly uprightly and close together. Shallow, elongated postcornual fossa. Frontals set at a high level between the horn bases in comparison with the dorsal parts of the orbital rims, orbital FOSSIL BOVIDAE FROM LANGEBAANWEG 265 rims project quite strongly, little or no central indentation of the parietofrontals suture, temporal lines on cranial roof do not approach closely posteriorly, braincase sides parallel, small supraorbital pits set close together. Occipital surface is wide and low and has a median vertical ridge, the mastoid has a large exposed area and is entirely contained within the occipital surface, the basi- occipital is only slightly narrowed anteriorly if at all, it has anterior tuberosities of moderate size and a central longitudinal groove, the basisphenoid rises fairly sharply in front of the basioccipital, and the auditory bullae are moderate to large sized. : Hypsodont cheek teeth with not very rugose enamel, small basal pillars on M,s and dP,s and occasionally on upper molars, central cavities of upper molars not very complicated in outline, upper molars with rather strong styles but poor development of ribs between them, medial lobes of upper molars less well rounded than in Pleistocene and Recent alcelaphines, medial walls of lower molars with less pronounced outbowings and with more prominent metastylids than in later alcelaphines, lower molars without goat folds, central cavities of lower molars with hardly any transverse constrictions centrally, P,s reduced and often absent in life. P,s with posterior part of tooth (behind level of metaconid) less reduced than in later alcelaphines, generally with transverse orientation of the valley between entoconid and entostylid, and with paraconid and metaconid growing towards one another but not usually fusing. The tibia has only a shallow posterior indentation in its distal articular facet, the metatarsal has a strong anterior longitudinal groove, otherwise the limb bones agree with those of later small or medium sized alcelaphines. Etymology The generic name comes from the Greek damalis, young cow or heifer, and acra, a cape. Damalacra neanica sp. nov. Figs 28-30, 34, 37 _ Holotype L7257—a complete skull with horn-cores and upper dentitions comprising P?—M®? on the right and a broken M!—-M? on the left (Figs 28-29, 37). Referred material From bed 3aS: L2573—cranium with horn-cores broken at their bases L2680—cranium with the lower part of the left horn-core and base of the right L12694A —cranium with most of the right horn-core and the lower part of the left; right maxilla with P*-M? (Fig. 30) L40072—cranium with left horn-core L40083—frontlet with basal half of horn-cores L40154—cranium with basal half of right horn-core “nner? ’ a ed a. ~wy, 44 266 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 28. Damalacra neanica. L7257, holotype. Dorsal and ventral views. Scale = 50 mm. FOSSIL BOVIDAE FROM LANGEBAANWEG 267 gtr =» * “2” ee Fig. 29. Damalacra neanica. L7257, holotype. Lateral view. Scale = 50 mm. ~~ “a0e>* —- icien eateries Fig. 30. Damalacra neanica. L12694A, cranium with horn-cores in lateral view. Scale = 50 mm. . aN ee 268 ANNALS OF THE SOUTH AFRICAN MUSEUM L40166—left horn-core with midfrontal suture (Fig. 34) L40275—female frontlet with both horn-cores, possibly of this species L40537—left and right horn-cores, back of skull, skull fragments L41329—left horn-core and base of the right one from one individual L41710—left horn-core with mid-frontal suture L2608, L2614, L4617, L4618, L6590, L6595, L15720, L40054, 40055, L40168, L40173, L40514, L40752, L40754, L40756, L41019, L2619 (immature)—left horn-cores L2607, L2613 (Fig. 37), L2668, L2670, L15015, L15809, L40110A, L40750 —right horn-cores L6582A —horn-core of indeterminate side L40126A—left horn-core, a female, possibly of this species From bed 3aN: L41414, L45085—left horn-cores Of uncertain origin (picked up by mine workers): 140761—cranium preserved from behind the level of the horn-cores L41330—back of cranium 5 Dental and postcranial remains wili be considered under the next species. Horizon The holotype and nearly all the other material of Damalacra neanica comes from bed 3aS of the PPM. No remains are known from the QSM. Etymology The specific name comes from the Greek neanicos, youthful, and refers to the place of this species early in the history of Alcelaphini. Diagnosis A species of Damalacra in which the horn-cores are without compression or slightly compressed anteroposteriorly, have no flattened lateral surface, taper fairly sharply from base to tip, show much increased divergence distally, have either slight backward or slight forward curvature in profile, and are inserted behind or above the back of the orbits. They have no very apparent torsion but it would have been anti-clockwise on the right side. The boundary between the pedicel top and the base of the horn-core is higher on the medial than on the lateral side of the horn-cores. It is probable that female horn-cores are smaller than those of males as in extant alcelaphines. Frontals set at a notably high level between the horn bases in comparison with the dorsal part of the orbital rims, braincase roof strongly angled on face, straight in profile, and without a parietal boss. Nasals transversely domed, without lateral flanges anteriorly, wider relative to their length than in living Alcelaphus and Damaliscus, preorbital fossae moderately large and deep and with an upper rim, ethmoidal fissure absent, zygomatic arch somewhat deepened in its anterior parts, infraorbital foramen set high above the back of P®, pre- maxillae contacting the nasals and narrowing only slightly as they rise, median indentation at back of palate passing further anteriorly than the lateral ones, FOSSIL BOVIDAE FROM LANGEBAANWEG 269 cheek tooth-row less anteriorly positioned than in living Alcelaphus and Dama- liscus lunatus, palatine foramina wide apart. The large mastoid exposure is especially expanded in its medioventral part, and the moderate to large auditory bullae are little inflated. Remarks The holotype skull is well preserved and lacks only the anterior parts of the premaxillae. The posteroventral surfaces of the skull around the basioccipital have been worn smooth by water rolling (Hendey 1970, fig. 3). The paired right and left mandibles, L7257C and D, were once thought to have come from the same individual as the skull, but their wear states are too early. However, it is possible, although not definite, that another right mandible, L7257B with teeth in a later state of wear, is from the same individual as the holotype. The holotype and L12694A are specimens in which there is an association between horn-cores and teeth. The cranium L40761 has the largest mastoid exposure seen in any Dama- lacra of either species, and it has also preserved its auditory bulla. 1 cranium, 3 frontlets and 33 horn-cores in the collection of Damalacra were taken as females. Of these, only the left horn-core L40126A and the frontlet L40275 are thought to be possibly of D. neanica. The insertion of their horn-cores is such that the longest cross-sectional diameter does not lie more or less parallel to the longitudinal midline of the skull. The remaining pieces will be considered under the next species. The combination of a suite of specialized characters with a number of primitive characters is notable in this species. Among its primitive characters are the transversely straight parietofrontals suture, supraorbital pits set closely together, the large preorbital fossa, the nasals not very long or narrow, the tooth-row not positioned very anteriorly, and most of the tooth characters mentioned in the diagnosis. Against this one sees the specializations of horn- cores compressed anteroposteriorly if at all, without much backward curvature - but with some distal divergence, inserted behind the orbits, the frontals well raised between the horn-core bases, and the shortness and steep inclination of the braincase roof. The shortness and steepness of the braincase roof must be a consequence of the short posterior migration of the horn-core bases, and one can appreciate that all the specialized characters are a simple unified set of changes. The cranium L12694A is interesting as having the most primitive aspect of _ any specimen of this species. Its horn-cores are less anteroposteriorly com- pressed, their distal divergence is less, they still show appreciable backward curvature, and the braincase roof is longer. A cranium, L40761, preserved forward to just behind the horn-cores, has a very large mastoid with the greatest expansion lying in the medioventral part of the bone. This matches D. neanica crania in which the mastoids are more expanded medioventrally than in D. acalla. Of six specimens of D. neanica, wy’ ~ tee rani a ee waster" 270 ANNALS OF THE SOUTH AFRICAN MUSEUM only one, L40072, has mastoids apparently both small and not expanded medio- ventrally. Hence L40761 may be assigned to D. neanica, and it has the largest mastoid known in that species. The top of the braincase is less smooth in profile in L40761 than in other D. neanica, but straight enough to fit this species. The anterior tuberosities of its basioccipital are fairly narrow and this, too, may fit D. neanica better than D. acalla. L40761 has an auditory bulla and becomes the only specimen of D. neanica with this structure preserved. It is moderately large but not very inflated, and thereby differs from that of D. acalla, as seen in L40474, which is only slightly larger but much more inflated. Measurements Measurements on L7257 other than those shown in Table 4 are as follows. Length of nasals 133, breadth of nasals 32,8, length of frontals 128, minimum width of palate between medial borders of M?s 55,2. Both L7257 and L12694A had upper tooth rows with the following respective measurements: occlusal lengths M1—M? 61,5, 55,4; occlusal lengths M? 23,2, 21,0; occlusal lengths P# dO eri: Measurements on the horn-cores and frontlets of D. neanica from bed 3aS, including specimens shown in Table 4, are: Number Standard Standard measured Mean’ Range _ deviation error Anteroposterior dia- meter at base of horn-core, left side 17 40,8 35,6-45,2 Bl 0,75 Mediolateral diameter at base of horn- core, left side gue igh 41,5 36,2-51,3 4,6 0,99 Horn-core length. 4 218 199-232 14,3 viel Minimum width across lateral sides of horn pedicels ; 6 99,9 95,8-107,3 4,7 1,92 Width across lateral edges of supra- orbital foramina . 4 49,3 44,4-55,3 4,5 Dy The female horn-core L40126A has an index 33,3 x 28,8. Horn-core indices for L41414 and L45085 from bed 3aN are 43,3 x 433, and 39,8 x 44,5 respectively. Comparisons The only extant genus of alcelaphine which might be thought to resemble Damalacra neanica is Beatragus. The single living species, B. hunteri, had become restricted to a very small area of northern Kenya by the time of its discovery by Europeans in the last century (Ansell 1971). B. antiquus L. S. B. Leakey, 1965, is a larger extinct species known from Beds I and II at Olduvai Gorge and from uppermost member G of the Shungura Formation, Omo. AIA PD. 271 FOSSIL BOVIDAE FROM LANGEBAANWEG v0S €°78 0°67 rar o[eugy SLCOVI OFEIVI I9LOVI LESOVT STOP €800V1 CLOOVI P69TIT 08971 ee Ft 1 i OLE L‘°9€ VTE I‘SZ €“CE OSE TOIL = 9°ZOL = IpL ze €“8p See 6‘LV ae OSE (lS 1‘6€ 9IZ = vry 0°66 69 9°CE 708 VLp E“Ss 6 £01 90 ULE tT 1 = EL me Qo TEE CVE €*€€ = OLZ 9°97 = — OEE €°O€ OCE * — OLOL O‘OOT — S601 9°78 76L DVL p78 0°E9 pes ae C*€S os = = b6r ¢‘L6 8°C6 sts €*LOL ZOE L‘9€ Ss Orr L‘@p I‘8¢ as (AS 7 TET =a ae C7 CST -LSCLT ‘DIIUDAU DAIDIDUUDG JO SJUSWIOINSeOU |[NYAS p I1aV tA IEP a. Boo [eqd1990 ede jo pepeoroa TOPOS SSOIOW YIPIM [e1Id1950 “IS¥q JO SOI}ISOIOqn} JONINJUe SSOIDe YIPIAA ee 1a DerEZO go do} wody WYSIOY [eWdI09G Tyeour AroyIpne ARE PUIYOq SPIO}SeL SSO1OR YIPIM ][NyS hee OSvouIeIg WINWIxeLy s[ejuory JO yorq 0} do} jeydi090 wiosy do} [Nys Jo yIsuST sjid TeyqreRians ge SPE [¥19} 8] SSOIOV YIPIAA ; s[oorpod UIOY JO SOOBJANS [8.1972] SSO1OV YIPIA WINUITTUTPT 9109-U1OY JO 9Seq 7B JOJOWIVIP [eID L[OIPo|] EMO BOI | jo a 7@ IOJOWVIP IOLN}sodo19jUy : : : YISUds] 9109-UWIOF{ 272 ANNALS OF THE SOUTH AFRICAN MUSEUM The anteroposterior compression of so many of the horn-cores of Damalacra neanica causes them to resemble the horn-cores of B. antiquus and, less closely, the larger horned males of B. Aunteri. The torsion of D. neanica horn-cores, in so far as it exists at all, is anti-clockwise on the right side, and this, too, is a resemblance to Beatragus. However, the Langebaanweg alcelaphine is. too primitive to be satisfactorily related to Beatragus. It is difficult to visualize two horn-core characters being held relatively steady over several million years while all other skull characters underwent evolutionary advance. Further com- parisons with other alcelaphines will be made in the discussion of Damalacra acalla. Damalacra acalla sp. nov. Figs 31-35, 37-38, 40 Hippotragini sp. Gentry in Hendey, 1970: 115. Holotype L40001—a cranium with much of the right and the lower part of the left horn-core (Fig. 31). Referred material | The major specimens assigned to this species are as follows: From bed 3aS: L1799—frontlet with skull fragments L1836—frontlet with some isolated, partly fragmented left and right upper teeth, skull fragments, vertebrae (Hendey 1970, pl. 4, fig. C). (Figs 32, 37) L12427—frontlet with horn-core bases, occipital surface, braincase roof, basioccipital, two atlas vertebrae and a crushed palate with P?-broken M? L20187—frontlet with both horn-cores, back of skull with basioccipital L40096—cranium with horn-core bases (Fig. 35) L40225—female cranium with both horn-cores (Fig. 38) L40319—frontlet with complete horn-cores, right and left maxillae with _ P8_M8, metatarsal, atlas and axis L40474—cranium with parts of both horn-cores L41832—frontlet with horn-cores, vertebrae L2615, L15928, L40288, L40493—frontlets L2616 (Figs 34, 37), L2622, L2624, L2625, L2632, L2649, L3587, L16280, L40025, L40120A, L40155A, L40163, L40165, L40167, L40171, L40172, L40178, L40278, L40723, L41327, L41368—left horn-cores L2622, L2623, L2628, L2629, L2633, L2638, L2665, L3489, L6592, L40048, L40050, L40074, L40123A, L40161, L40177, L40187, L40248, L41092, L41202, L41706—right horn-cores L10563—female left horn-core FOSSIL BOVIDAE FROM LANGEBAANWEG 273 Fig. 31. Damalacra acalla. L40001, holotype. Cranium in dorsal and lateral views. Scale = 25 mm. From QSM or bed 3aS: L12856—damaged cranium with base of right horn-core L22278—left horn-core, basioccipital, atlas, vertebral fragments, left maxilla with P?-M?® and right mandible with P,-M; (no P, in life). (Fig. 39) L24809—right horn-core, part of the left horn-core, basioccipital, skull fragments, three left upper molars, two right upper molars From bed 3aN: L46075—cranium with horn-cores L33832, L41412, L46040 (Fig. 33)—frontlets 146042, L46061, L46072—left horn-cores L30215, L45012, L46039, L46043, L46047, L46061, L46068—right horn- i cores _ From beds 3aS or 3aN: L40751, L40757—left horn-cores L40758, L40793—right horn-cores. L46072 and L40793 may belong to Damalacra neanica. Associations between horn-cores and teeth are provided by L1836, L12427, 122278, L24809, and L40319. a’ // ved rTy “h/t . “M4 274 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 32. Damalacra acalla. 1836, frontlet in anterior view. Scale = 50 mm. QE PIE LER a oe ET GIT ey OT Hite ET: ae — nee ae 4 ieee. Winco. 1 TN te Np AOS a AN a IA * 9 ee Oo! FOSSIL BOVIDAE FROM LANGEBAANWEG 275 Fig. 33. Damalacra acalla. 46040, frontlet in anterior view. Scale = 25 mm. Fig. 34. From the left: Damalacra neanica 40166, left horn-core in anterior and lateral views. D. acalla L2616. same views. Scale = 25 mm. 276 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 35. Damalacra acalla. 40096, cranium in occipital and ventral views. Scale = 25 mm. Horizon The holotype comes from bed 3aS of the PPM. Most of the other material comes from 3aS, but this species is better represented in bed 3aN than is D. neanica and three fossils are possibly from the QSM. A left mandible, L12883, is from the QSM, but it cannot be ascertained to which species of Damalacra it belongs. Diagnosis An alcelaphine in which the cranial size and proportions are similar to Damalacra neanica. Horn-cores differ from those of D. neanica by being com- pressed in the mediolateral plane if at all, often with a localized, usually medial, swelling at their bases, sometimes tending to have a flattened lateral surface ‘along part of their length, with more definite backward curvature and less marked distal divergence, inserted closer behind the orbits, and with a more nearly horizontal boundary between the top of the pedicel and the horn-core proper. Other differences from D. neanica are that the braincase roof is less steeply inclined, more curved in profile, and with an insignificant parietal hump which is about as well developed as in living Damaliscus; the mastoid exposure large but less expanded especially medioventrally, and the auditory bullae slightly larger and much more inflated. Female horn-cores are smaller than those of the males as in extant alcela- phines. Tooth characters are taken not to differ from D. neanica. Etymology The specific name is from the Greek acalles, without charms, and refers to the primitive state of nearly all the characters in this species. Remarks No fossil has been preserved with a face, so no facial characters appear in the diagnosis. One cannot assume that the face would differ very much from FOSSIL BOVIDAE FROM LANGEBAANWEG 9 | that of Damalacra neanica. ’ There can be little doubt that the two alcelaphine species are closely related. It is interesting that D. acalla has many characters which appear to be primitive just as in D. neanica, but hardly any which are specialized. Primitive characters are the transversely straight parietofrontals suture, the supraorbital pits not set wide apart, most of the characters of the horn-cores, the fairly long brain- case and its not very inclined roof, and the characters of the cheek teeth. Fig- ure 36 shows the difference in horn-core compression between D. acalla and D. neanica, and also that D. neanica has different cranial proportions in the form of a shorter braincase and a lower and wider occipital surface. The extant Damaliscus dorcas, shown on the same figure, has cranial proportions similar to Damalacra acalla, but rather small horn-cores and a relatively narrow width across the anterior tuberosities of the basioccipital. Figure 37 shows cross- sections of Damalacra horn-cores. Anteroposterior diameter at horn core base Mediolateral diameter at horn core base aN : SS me ~ Width across horn core bases ae Braincase length ; : ! Ce Skull width across mastoids ees ve Occipital height : : a 7 Width across anterior tuberosities of basioccipital “a Width across posterior tuberosities of basioccipital 80 90 100 110 Fig. 36. Percentage diagram of skull measurements in Alcelaphini. The standard line at 100 per cent is for seven Damalacra acalla from bed 3aS. The other continuous line is for six D. neanica including L12694A, and the dashed line is for four male Damaliscus dorcas. Not all measurements were available on all the fossils. Among the female remains of Damalacra only a horn-core and a frontlet, L40126A and L40275, are thought to be of D. neanica. Another two are thought to be definitely of D. acalla; these are the cranium L40225 and the left horn- core L10563. The remaining horn-cores appear to be attenuated versions of D. acalla male horn-cores, but one cannot be certain that the females of D. neanica might not have more primitive horn-cores than the males. 278 ANNALS OF THE SOUTH AFRICAN MUSEUM OOO Fig. 37. Cross-sections of Damalacra horn-cores at a distance above the pedicel top equal to half the anteroposterior basal diameter. All are shown as if they were of the right side. Lateral sides are towards the left and anterior sides towards the base of the illustration. From the left: D. neanica L7257 (holotype) and L2613, D. acalla L1836 and L2616. Scale = 10 mm. Fig. 38. Damalacra acalla. 140225, female cranium in lateral and anterior views. Scale = 25 mm. FOSSIL BOVIDAE FROM LANGEBAANWEG 279 L46075 is an alcelaphine cranium with horn-cores from bed 3aN, apparently belonging to D. acalla, but differing from those in bed 3aS by its horn-cores being slightly less compressed mediolaterally, the braincase roof shorter, and the anterior tuberosities of the basioccipital wide apart. Its most instructive comparison is with L12694A which has the most primitive aspect of any D. neanica cranium. L46075 is larger, its horn-cores differ little in their degree of compression, they diverge distally but less than in L12694A, they have a slight backward curvature but less than L12694A, their insertion is less upright in side view than in L12694A, the pedicel is not higher on the medial than the lateral side of the horn-core as in L12694A (and other D. neanica), the braincase roof has a slightly curved profile, whereas L12694A is straight, and the mastoid is not expanded medioventrally and is certainly less so than in L12694A. The anterior tuberosities of the basioccipital are much wider apart than in L12694A or any other examples of D. neanica. It is interesting that the relative lack of backward curvature of the horn-cores makes them more like D. neanica than are those of L12694A. The total sample of horn-cores of D. acalla does not show that antero- posterior compression has increased in bed 3aN in comparison with 3aS, but a size increase has taken place (Fig. 39). The differences between the 3aS and 3aN samples of horn-cores for both anteroposterior and mediolateral diameters (measurements given below) are significant at the 5 per cent level. The values of T were 2,39 and 2,88 for 27 degrees of freedom, and confirm the size increase. Samples for the measurement of minimum width across the lateral sides of the horn-core pedicels were smaller but similarly showed a size increase from bed 3aS to 3aN. The value of T was 3,13 for 10 degrees of freedom which is significant at the 5 per cent level. Measurements also suggest that the horn-cores decreased in length from 3aS to 3aN but in this case the difference was not statistically significant. In general the D. acalla in bed 3aN looks more robustly built across the frontals and horn-core bases than in bed 3aS, and the frontals - are more raised between the horn-core bases. As the horn-cores and their supporting pedicels become larger, so the projection of the dorsal part of the orbital rims becomes less obvious. On the paired horn-cores L46061, the localized basal swelling is confined to the middle part of the medial side and the top of the sinus in the pedicel comes level only with the junction of the pedicel top and horn-core proper. Their basal index is 55,3 x 50,3. Numbers of such horn-cores are coming to light in the material being processed from bed 3aN (Q. B. Hendey, pers. comm. 29 August 1978). The characters of the teeth of Damalacra have been mentioned in the generic diagnosis and the teeth are illustrated in Figures 40-42. They are about the size of the teeth of Damaliscus lunatus, whereas the size of horn-cores, crania and postcranial bones of Damalacra is in closer agreement with Dama- liscus dorcas, a species smaller than D. Junatus. Even the mandibular rami of Damalacra are only about as large as in Damaliscus dorcas. It was not found possible to differentiate between the two species of Damalacra on their teeth. 280 ANNALS OF THE SOUTH AFRICAN MUSEUM Mediolateral x 504 diameter x A Anteroposterior diameter 30 Ai 50 60mm Fig. 39. Basal diameters of Damalacra horn-cores. X = D. neanica lefts from bed 3aS, X = the same from bed 3aN, O = D. acalla rights, the solid ones from bed 3aN and the rest from bed 3aS, + = D. neanica 22278 and L24809 from QSM, C = Damalacra female left horn-cores, the underlined one from bed 3aN and the rest from 3aS, Upper diagonal. line = 100%, lower one = 75% as in Figure 3. _ Fig. 40. Damalacra acalla 22278, occlusal view of left upper dentition. Damalacra sp. L40534, occlusal view of right lower dentition. Scale = 10 mm. ee FOSSIL BOVIDAE FROM LANGEBAANWEG 281 Fig. 41. Damalacra sp. L41482, lateral view of left mandible. Scale = 50 mm. Fig. 42. Damalacra sp. L11612, occlusal view of immature right lower dentition with dP;—M,, alveoli for dP, and a fragment of Mbp. Scale = 10 mm. 20 30 40 mm Fig. 43. Dimensions of unworn and almost unworn M;s of Alcelaphini. X = PPM Lange- baanweg, the lowest one being bed 3aS and the other two bed 3aN, O = Cape Province sites of Middle Pleistocene and later age, the underlined ones being not completely unworn, S = Sterkfontein Type Site, n = Wadi Natrun. Diagonal lines = 200%, 150% and 100% as in Figure 3. 282 ANNALS OF THE SOUTH AFRICAN MUSEUM 40 4 Length 30. 20 Length M—M, 40 50 60 70 mm Fig. 44. Occlusal lengths of lower premolar and molar rows in Alcelaphini and in bovidS from Fort Ternan. X = Damalacra rights, O = lefts, + = both sides, underlined readings are bed 3aN, the rest are bed 3aS, solid circle is L12883 from QSM. a = extant AlcelaphuS buselaphus, c = extant Damaliscus dorcas, u= extant D. lunatus,. s = ?Pseudotragus potwaricus from Fort Ternan, t = Oioceros tanyceras from Fort Ternan. Upper diagonal line = 50%, lower one = 40% as in Figure 3. The readings for all species below the 40 per cent line are those without P, in life. The most interesting feature of the teeth is the primitive state of so many of their characters. Figure 43 shows how their hypsodonty is less than in extant and other fossil alcelaphines. In Langebaanweg times unworn M,s were only about two-thirds as high crowned as in the Middle Pleistocene and later. The difference much exceeds that between Sterkfontein Type Site and the most recent sample. It is interesting against this background that trends to the reduction and loss of P, were already present. Out of 36 specimens in which its presence or absence could be ascertained, 21 had it and 15 were without. Of the 21 with it only 4 were in later middle or late wear, while of the 15 without it 10 were in later middle or late wear. The trend to reduction and loss of P, may exist about as much as in Parmularius altidens of Bed I, Olduvai Gorge, in which three out of eight specimens lacked it, and is certainly in advance of modern Damaliscus lunatus and Alcelaphus in which P, is practically always present. However, Parmularius has reduced its whole premolar row more than in Damalacra (Fig. 44) and hence P, is smaller relative to M, (Fig. 45). FOSSIL BOVIDAE FROM LANGEBAANWEG 283 Alcelaphus buselaphus aetna ae sie 2 6 Damaliscus dorcas re ; : Damaliscus agelaius [J J Parmularius altidens values Parmularius sp (Laetoli) Sax X Damalacra aes een eee : 18 Oioceros tanyceras Ss SE ES ] 50 60 70 Fig. 45. Occlusal length of P, expressed as a percentage of that of M, in alcelaphines. For each species the range is shown by a horizontal line, mean value by a short vertical line, standard deviation by squared brackets, and the number of specimens by the figure adjacent to the mean. The Langebaanweg alcelaphine teeth are the earliest known which are definitely of this tribe, and they make an interesting comparison with those of the much earlier Caprini, Oioceros tanyceras and Pseudotragus? potwaricus which are candidates for alcelaphine ancestry (and might eventually have to be transferred to that tribe to preserve monophyly). These forms are from Fort Ternan and Ngorora (Gentry 1970a, 1978a). The Langebaanweg teeth are larger, more hypsodont than the Ojoceros if not also than the Pseudotragus ?, the basal pillars are much less evident, the paraconid and metaconid of P, are either growing toward one another or fused, the entoconid and entoconulid remain separate on P, later in wear, the medial wall of the back lobe of Msg is perhaps more clearly offset laterally in earlier wear, and the central cavities of the lower molars show less sign of being transversely constricted centrally. The first four of these characters are more advanced at Langebaanweg, the next two may be linked with larger size and the last appears to be more primitive. Some characters are unchanged: the degree of rounding of the walls of the medial and lateral lobes on the lower molars, the shape of the medial lobes of the upper molars, and the level of development of styles on the upper molars. Some examples of both Fort Ternan species are already without P,s in life (Gentry 1970a: 265, 285) but this trend has been taken further at Langebaanweg. More- over, there is less tendency to reduction of P, at Fort Ternan, as can be seen in Figure 45 where the relatively small size of P, is a consequence of the relatively unreduced P,. Overall the Fort Ternan premolar rows are as long as at Lange- baanweg (Fig. 44). Some of the morphological differences between Langebaan- 284 ANNALS OF THE SOUTH AFRICAN MUSEUM weg alcelaphine teeth, those of other alcelaphines and the Fort Ternan caprines are shown in Figure 46. Alcelaphine postcranial bones from ‘E’ Quarry are of appropriate size to be conspecific with the cranial and dental remains of the Damalacra species. The following notes are based on an associated skeleton, L41482 (Figs 21, 47-48), from bed 3aS with complete examples of all the long limb bones, as well as on other examples. The lengths and least transverse thicknesses of the long limb bones of L41482 are as follows: femur 224 x 22,2, tibia 269 x 22,7, metatarsal 223 x 16,9, humerus 195 x 20,4, radius 226 x c. 23,6, metacarpal 201 x 17,9. Chief among the other limb bones are the following: From bed 3aS: L12456—much of right humerus, proximal and distal right radius L12463—distal right tibia, associated with above L15000—complete right metacarpal with length and least transverse thickness of 226 x 19,1 L15031—right scapula L15075, L15155—complete left and right uation with lengths and least transverse thicknesses of 197 x 22,0 and 176 x 19,3 L15271—most of right humerus L15276—distal metatarsal L15963—distal right humerus L40279—complete left metacarpal with length and least transverse thick- ness of 199 x 17,3 L40319—complete metatarsal with length and least transverse thickness of 224 x 17,9, atlas vertebra L41216—distal left humerus, distal left radius, proximal left metacarpal, left and right proximal metatarsals Probably from bed 3aS: L1848, L3037—distal left and distal right femora L2189, L3042—complete left and right metatarsals with engi and least transverse thicknesses of 220 x 16,8 and 228 x 17,4 L2167, L12280—distal right and distal left tibiae From bed 3aN: L30769—distal left humerus, distal right radius; probably alcelaphine L31388—distal right humerus L31684—distal right tibia and proximal right metatarsal L32707—distal left tibia and complete left metatarsal with length and least transverse thickness of c. 237 X c. 18,6 Generally the Langebaanweg bones show the typical characters of alcela- phines, most of which are cursorial adaptations (Gentry 1970a: 277-282), but often less sharply defined than in extant species. The metapodials may be thicker than in the similarly sized Damaliscus dorcas and are about as thick as in later, larger alcelaphines (Fig. 49). The metacarpals are not as long relative FOSSIL BOVIDAE FROM LANGEBAANWEG 285 ey GY Si Fort Ternan Langebaanweg Omo Extant Fig. 46. Occlusal views of M? (top row), M, (middle row) and P, (bottom row) of Caprini from Fort Ternan, Damalacra from E Quarry Langebaanweg, Alcelaphini from the Shungura Formation Omo, and extant Alcelaphus buselaphus. All teeth are of the right side and the anterior direction lies to the right. Fig. 47. Damalacra sp. L41482, long limb bones of hind leg. From the left: right femur in lateral and anterior view, left tibia in medial and anterior view, right metatarsal in lateral and anterior views. Scale = 25 mm. 286 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 48. Damalacra sp. 41482, long limb bones of foreleg. From the left: left humerus in lateral and anterior view, right radius in medial and anterior view, left metacarpal in posterior and anterior view. Scale = 25 mm. Least = thickness 20 200 250 mm Fig. 49. Proportions of alcelaphine metacarpals. XK = bed 3aS, E Quarry Langebaanweg. a = extant Alcelaphus buselaphus, c = extant Damaliscus dorcas, u = extant D. lunatus, o = Parmularius altidens, rights underlined. Upper diagonal line = 10%, lower one = 7% as in Figure 3. FOSSIL BOVIDAE FROM LANGEBAANWEG 287 to the metatarsals as in extant Damaliscus lunatus or Alcelaphus buselaphus. The only characters in which they are definitely different or less advanced have been mentioned in the generic diagnosis. The femur shows an anteroposteriorly long lateral part of the articular head, a deep hollow between the articular head and the great trochanter, well-marked insertion positions for muscles and ligaments on the distal lateral condyle, and a medial condyle which projects well anteriorly. However, it is possible that the patellar fossa distally is less extremely wide and the lateral roughened fossa less deep than in living alcela- phines. The tibia shows a strong central swelling on the top articular surface and a depression medial to it just in front of the level of the paired flanges, themselves well marked. There is an upcurved edge of the lateral facet at the proximal end, and the medialmost muscle scar at the top of the posterior surface is long, strong, high on the shaft and in a relatively medial position. Distally the medial malleolus is not clearly shorter than in living alcelaphines, and the front fibula facet is well outlined and distinct from the rear one. The presence of a patellar groove at the front of the proximal articular surface is like Conno- chaetes rather than Alcelaphus and Damaliscus. The back edge of the distal arti- cular facets is less indented centrally than in most extant alcelaphines in L2167, L41482 and L32707, but L31684 does not appear to be different. The metatarsal has the posterior part of its top articular surface transversely narrower than _ the central parts, a small main facet for the naviculocuboid, no deep hollow between the anterior and posterior naviculocuboid facets, parallel outer edges of the distal condyles, deep hollows above the distal condyles anteriorly, and strong paired flanges distally on the anterior surface. The ridges on the condyles pass high posteriorly. However, the posteromedial part of the main naviculo- cuboid facet may have been more strongly raised than in living alcelaphines, the hollow deeper around the foramen at the top of the posterior surface (a character as in Connochaetes), and the distal condyles frequently have the appearance of being less high and narrow overall. The single complete metatarsal from bed 3aN is larger than examples from 3aS. In the scapula the tuber scapulae is situated near to the lateral side in ventral view, the area for the origin of the teres minor muscle is well hollowed, and there is a slight flattening of the posterolateral edge of the glenoid facet. In the humerus the bicipital groove is wide and set back from the front edge of the lateral tuberosity, and a sharp ridge down the front of the tuberosity is variably developed. The distal end is completely alcelaphine with upright con- dyles of the articular surface, a strong medial groove, an indentation in the top of the medial condyle and a V-shaped prolongation of the lateral surface. However, the last character is not well marked in L15075, L31388A or L41482; it is better in L12456 and L41216. Two other possible differences from living alcelaphines are that the top of the medial tuberosity forms a less upstanding point and that the lateral tuberosity rises higher above the top of the infra- spinatus scar in lateral view. In the radius the proximal lateral tubercle is large and set high, there is no rim on the medial side of the proximal medial facet, ——— ae 288 ANNALS OF THE SOUTH AFRICAN MUSEUM the back of the lateral facet is set forwards, and distally the anterior flanges are strong and set close together. However, L41216 has rather a poor medioposterior hollowing for the scaphoid and L12456A is rather swollen distally in side view for an alcelaphine. The metacarpal of L41482 shows an angled anteromedial corner to its magnumtrapezoid facet as is usual in Alcelaphini but the unciform | facet is perhaps larger. Another metacarpal, L15000, shows the converse con- ditions for these characters. In the two atlas vertebrae, L40319 and L41482 (Fig. 50), the side edges are not concave over a very great length, and the front edge of the dorsal surface is not very indented. An alcelaphine character which both do show is a forwardly directed spike centrally at the front of the ventral surface. SSS Fig. 50. Damalacra sp. L41482. From the left: atlas vertebra in dorsal and ventral views, axis vertebra in left lateral view. Anterior sides towards the left. Scale = 25 mm. Measurements Measurements on three frontlets of D. acalla are: L1799 L1836 L15928 Skull width across posterior side of orbits — 133,8 — Length of horn-core along its front edge — 265 — Anteroposterior diameter at base of horn-core . ; : : : : 45,2 42,9 44,4 Mediolateral diameter at base of horn- core. f ; . i ‘ ce 38,8 38,2 37,0 Minimum width across lateral sides of horn pedicels. 2 . : ; 90,3 90,3 94,3 Width across lateral edges of supra- orbital foramina : f : : — 46,3 51,3 — ews tittten an _ — FNS w am Fin) 0 YP ia MéMdhbhs es <7. fist —_——— ca aaaae soe a“: om Swe Oe 289 oO 5 oe COE CSE 6°0€ (6165 VCE 8°ZE SOF : ; ‘ ; " — [eyIdos0Iseq JO SaIjIsOIogn} IOLI91sod sso10e UIPIM Zi — 9°ZE 767 9°97 WG 9°TZ =< ClEZ 7 5 Ft JeyEdIOo0Iseq JO SarIsoJaqn} JOFI9}Ue sso1Oe YIPIAA 2 = eer €°9€ viv a €“9€ 9°9€ 7 OV "7 Fs wnuseuT UoUTeIOy Jo do} wouy WyStey eydI009 Hu ie a oa ae ae 0669 — OOOW = — — o COO eee 2 " * YYPIM osevouresq wunUIxeEy < +09 8°rS 8°TS ae = — — 6°€9 "51 syesuoay Jo eed 0} doy Eidine Hoy do} [[nys Jo yI3ueT a = FE: = ae == — 8 eee "syd [eyIqroeidns Jo sospa [e19}e] SSOIDe YIPIAA a 9°8L (0) a as £°96 a 8°96 aa 0°66 es ee A Sleoiped UJOY JO SOdBJINS [e19}e] SSOIOV YIPIM WINUTTUTY Ee £6 8S a 0°6€ O*EE v°6E lig ‘bb ee egos * * 9109-UJOY JO eseq 7 JO}OUIRIP [eIOIL[OIPSW, 4 S*vE 60S er Ly O'Er L’vv L‘Or [LASS ey ee re SS SOTOae “urOY JO aseq ye JoJoUILIP JO}1e}sodos9}Uy 2 ae 917 ce ar CST =a ae a : i ; z : : i : YJsus[ 9109-UIOH{ g[euray fe) A Ssccovt SLoorT plow 9600VT 608771 L8IOZI LZ@PZIT 1000r1 S| B ‘DI[VIV DAIDIDUIDG JO SyUoWOINSeOUL |[NyS e) Ik ¢ ATEVL 290 ANNALS OF THE SOUTH AFRICAN MUSEUM The widths across the anterior and posterior tuberosities of the basi- occipital of L12856 are 26,9 and 30,9. Additional measurements on L12427, not shown in Table 5, are occlusal lengths of M? 24,8, of P?-P* 39,4, of P2 10,9, of P* 12,2. The female horn-core L10563 has an index of 32,9 x 27,4. Measurements on horn-cores and frontlets of D. acalla, including specimens listed in Table 5, are: Number Standard Standard measured Mean Range deviation error Anteroposterior dia- meter at base of horn-core, bed 34S, rishtsides . 21 47,5 40,7-53,7 aS 0,72 Mediolateral diameter of above . eae | 39,3 = 31,7-44,3 2,9 0,63 Anteroposterior dia- meter at base of horn-core, bed 3aN, right side . 8 51,4 44,7-58,2 5,3 1,89 Mediolateral diameter of above . 8 43,4 36,7-49,9 4,6 1,63 Horn-core length, bed 3aS_. : 4 240 216-265 216 10,8 Minimum width across lateral sides of horn pedicels, bed 3aS 9 96,4 90,3-101,7 4,1 1,35 Width across lateral edges of supra- orbital foramina, bedasaSh sie] (ae 5 49,3 44,1-54,2 4,1 1,81 Readings on the female horn-cores of Damalacra from bed 3aS, apart from the four assigned to species, are: Number Standard Standard measured Mean Range deviation error Anteroposterior dia- meter at base of horn-core . sek XG) 34,3 28,9-38,9 2,6 0,51 Mediolateral diameter at base of horn- core bu MENS ES ce) 29,0 22,4-32,6 22 0,44 Individual readings for specimens from bed 3aN were: Horn-core lengths PS PAKS Minimum width across lateral sides of horn pedicels 98,9, 107,9 111,9 Width across lateral edges of supraorbital foramina 55,7 FOSSIL BOVIDAE FROM LANGEBAANWEG 291 Readings for occlusal lengths of alcelaphine lower teeth in middle wear are: ‘=| = 6 5 2 Ue ee ee 25 o aie S Be Zee es fe Bee oe eS M,-Ms;, bed 3aS, Ponte wets il 669s 59029743 ~ 3:48 O75) 0510 M,-Ms, bed 3aN, Fee eh oS 65,9 60,6-68,9 3,1 140° 24570 M,, bed 3aS, right 29 18: SHO 225 0 Gas OA. 5196 M,, bed 3aN, left 5 D2 Wier OLD8 Sire TS eI ISOH 2155.88 M,, bed 3aS, left. 49 28,8 25,7-32,3 1,7 0,24 5,90 M,, bed 3aN, left 8 DBA 96:8230:4e 00,4 048 4893 P,-P, (with P,), bed 3aS, right 11 28,00 2153-3333" 139 1S aa (SS P,—P, (without P,), bed 3aS, right 6 24,2 22,6-29,9 2,9 |e) cya PALE) P,-P, (with P,), bed 3aN, left 2 31,6 31,2-32,.0 — — — P,, bed 3aS, right. 8 6:8 VSB CNE! eG 20230 RS P,, bed 3aN, left. 2 Tih COONS snus ue P,, bed3aS, right. 40 13.9}. M2655. Ou! OF oom S04 P,, bed 3aN, left. 5 14.00 128215,00 0,9 2426408 » 6,43 The coefficients of variation show that measurements of the occlusal lengths of P,P, and P, are particularly variable. Either the reduction in size of P, or the difficulty of identifying measuring points at the front and back of the premolar row may have an effect here, but it is also possible that the large coefficient of variation shows that one of the two species of Damalacra has a shorter premolar row than the other one. There is no indication that tooth size increased from bed 3aS to 3aN, in contrast to the indication from horn-cores. _ Comparisons It is informative to compare both species of Damalacra with a number of alcelaphine specimens from the Hadar Formation, Afar, which may represent an early member of the group in which Parmularius, Damaliscus and Alcelaphus evolved. The conclusion reached will be that Damalacra represents a still more primitive stage of alcelaphine evolution and is of uncertain relationship to the 292 ANNALS OF THE SOUTH AFRICAN MUSEUM Afar remains. The main Afar remains are: AL 208-7 largely complete skull from SH-3 surface AL 353-3 cranium with left horn-core from SH-2 surface AL 16125 cranium with base of left horn-core from DD-3 surface AL 120-2 pair of horn-cores and parts of | from DD-3 surface associated skeleton AL 310-18 — cranium with base of right horn-core The Afar alcelaphine shows the following characters: size about equal to extant Damaliscus lunatus and Alcelaphus; horn-cores of short to moderate length, without much compression, without a flattened lateral surface, without Keels or transverse ridges, base often squared off posteromedially, thickness of cross-section often diminishing rapidly above the base and producing a tapered appearance, inserted close together above the back of the orbits, little divergent basally but more so distally, boundaries between pedicel tops and horn-core bases higher on the medial than the lateral sides; postcornual fossa long and shallow, parietofrontals suture with little central indentation, frontals raised between horn-core bases, orbital rims not projecting as a separate structure from the descending lateral surface of the horn pedicels, braincase roof short and inclined, near absence of a Parmularius-like parietal boss, braincase widening posteriorly, a deep face, long and narrow nasals, a large and fairly deep preorbital fossa, occipital surface with a strong median vertical ridge, and ventral border of the mastoid angled instead of straight. It is possible that horn-core shape evolves during the Hadar Formation from the condition of rather low insertions, little curvature or distal divergence and slight medio- lateral compression to more upright insertions, a forwardly curved course with clockwise torsion on the right side, stronger distal divergence, and slight antero- ‘posterior compression. It is also possible that the braincase shortens and that the slight indication of a parietal boss disappears altogether, but these are not very certain. There also seem to be evolutionary changes at the back of the skull in the Afar alcelaphines. In AL 208-7 and AL 353-3 the junction between the base of the nuchal crest and the back of the zygomatic arch is not posteriorly placed, each side of the occipital surface faces partly laterally as well as back- wards, there is a strong median vertical ridge on the occipital, and the mastoid is wholly contained within the occipital surface. These characters are all linked and the converse conditions are found in AL 161-5 and AL 310-18 in which the junction of nuchal crest and zygomatic arch is more posterior, the occipital faces more wholly backwards and has less of a median vertical ridge, and the lateral parts of the mastoids have the appearance of being deflected to face laterally just in front of the occipital edge. AL 161-5 and AL 310-18 are Closer to extant Alcelaphus and many Damaliscus and one imagines that they are more highly evolved. These Pliocene alcelaphines differ strikingly from extant Alcelaphus by the absence of extreme braincase shortening and face lengthening, more extensive FOSSIL BOVIDAE FROM LANGEBAANWEG 293 preorbital fossae, and the absence of horn-core specializations as well as by more minor characters. They also show differences from Damaliscus: transverse ridges usually absent on the horn-cores, the horn-core compression evolving to become slightly anteroposterior rather than slightly mediolateral, horn- cores not curved backward, horn-cores more divergent distally (not a difference from D. lunatus lunatus), the diminishing parietal boss, the braincase roof not curved in profile, the external auditory meatus set lower in relation to the occipital surface behind it than in male D. Junatus, a narrower basioccipital, and the suture at the back of the parietal protruding further forwards in its central parts (not a difference from D. dorcas). They differ from both Alcelaphus and Damaliscus by the angled ventral border of the mastoid. The teeth of the Afar alcelaphines differ from extant alcelaphines by their more primitive characters. They are probably less hypsodont; on the upper molars the central cavities are less complicated, the ribs are less prominent in relation to the styles, the medial lobes are perhaps less rounded; on the lower molars the central cavities are less curved and the medial walls perhaps less outbowed; on P, the rear part (the region of the hypoconid, entoconid and entostylid) is less reduced, and the valley between entoconid and entostylid is perhaps oriented more nearly transversely. Insufficient of them are known for the reduction or otherwise of P, to be assessed. Both species of Damalacra differ from the Hadar Formation species by slightly smaller size, horn-cores probably longer and with bases not usually squared off posteromedially, dorsal orbital rims more strongly projecting, braincase sides parallel and supraorbital pits closer together. Both species would probably also differ in shorter and wider nasals, a less deep face and a more posterior setting of the tooth-row—characters which are so far known only from a unique face of D. neanica. Nearly all the characters mentioned so far are likely to be primitive. D. neanica differs additionally from the Afar species in the more posterior insertions and weak anti-clockwise torsion of the horn-cores, and the more inclined braincase roof without a hint of a parietal boss. These characters remove D. neanica from likely ancestry to the Afar species. D. acalla differs additionally by sometimes showing a flattened lateral surface of its horn-cores, sometimes having a localized swelling at the base of the horn-cores, with more definite backward curvature and pedicels of about equal height on their lateral and medial sides. The last two characters are probably primitive in the Lange- baanweg form and the first two, even if advanced, do not appear very imposing. The Langebaanweg alcelaphine teeth as a whole differ from the Hadar Formation ones by being still more primitive. They are definitely less hypsodont, small basal pillars exist on M,s and dP,s and occasionally on upper molars, the central cavities of the upper molars are even less complicated, the ribs still weaker in relation to the styles, and the medial lobes are pointed rather than rounded. The medial walls of the lower molars are straighter, the central cavities not very curved and with almost no transverse constriction centrally (except 294 ANNALS OF THE SOUTH AFRICAN MUSEUM perhaps in later wear), and the lateral lobes are more pointed. The P, has poorer fusion between metaconid and paraconid, a larger rear part, and the valley between entoconid and entostylid is oriented transversely. Some characters of horn-cores and braincase were mentioned earlier as possibly undergoing change during the span of the Hadar Formation. Dama- lacra acalla is more like the earlier than the later form except in its more upright horn-core insertions. D. neanica is more like the earlier in its occipital characters but more like the later in its horn-core characters and complete absence of a parietal boss. A partial cranium from the Laetolil or Ndolanya Beds, 1959.233 (Gentry & Gentry 1978, pl. 22 (fig. 1)) at present in Nairobi, could be conspecific or a close relative of the Afar species, and Damalops palaeindicus (Falconer), 1859, from the Pinjor Formation of the Siwaliks and Tadzhikistan (Dmitrieva 1977) could also be a close relative. Details of D. palaeindicus are given in Lydekker (1886, pl. 4 (figs 3, 3a, 5)) and Pilgrim (1939: 67-70), and it was discussed by Gentry & Gentry (1978: 406, 412) in comparison with Olduvai alcelaphines. D. palae- indicus differs from the Afar species in that its horn-cores curve backward, they show no rapid tapering above the base and the sides of the braincase appear to be parallel instead of showing posterior widening. The horn-cores may be longer, the braincase shorter, and the tooth-row positioned more anteriorly, but this is not certain. So far as can be seen, the Laetoli specimen agrees more closely with the Afar species than with D. palaeindicus. It is apparent that both species of Damalacra will differ from Damalops about as much as from the Afar alcelaphine. . Apart from the Afar alcelaphine and its possible close relatives, a smaller alcelaphine is represented at Laetoli by a cranium with horn-cores, 1959.277 at present in Nairobi, discussed by Gentry & Gentry (1978: 382, pl. 21, pl. 22 (fig. 2)) who thought it was an early Parmularius. Since 1974 some conspecific horn-cores have been recovered from the Laetolil Beds by M. D. Leakey, so it can be taken as a member of the fauna dating from before 3,5 m.y. It differs from the Hadar Formation species by smaller size, possibly longer horn-cores, backward curvature and little distal divergence of the horn-cores, less definite posteromedial squaring off at the horn-core bases, occipital perhaps facing even more strongly laterally on each side, closer supraorbital pits with a more concave area of the frontals in between them, and dorsal orbital rims projecting more strongly as a separate structure from the lateral sides of the horn pedicels, all of which could be conceived as primitive. The slight mediolateral compres- sion and absence of rapid tapering of the horn-cores is more like earlier than later Afar specimens and could also be primitive. Such primitive characters can be attributed either to the greater geological age of the Laetoli species or to its smaller size. Other differences are quite a sharp backwards bend of the horn-core just over half-way from base to tip, a posterolateral basal swelling on the horn-core, higher pedicels, parallel sides of the braincase and a more prominent parietal boss. Some of these may be advanced and others primitive. ‘- ’ t ‘ ‘ il } ‘ FOSSIL BOVIDAE FROM LANGEBAANWEG 295 The parietal boss and the sharp bend, basal swelling and high pedicels of the horn-cores could all foreshadow Parmularius and suggest ancestry to it. Damalacra neanica horn-cores differ from the Laetoli species by the slight anteroposterior compression, their more posterior insertions, no backward curvature and more distal divergence which appear to constitute their own set of advanced characters. They also lack the sharp alteration in course, a localized basal swelling and high pedicels—the supposed advanced characters of the Laetoli species. The cranial roof is also quite different in D. neanica by being shorter, more inclined, straight and without a parietal boss. Damalacra acalla horn-cores differ by sometimes having a flattened lateral surface, which may be advanced but is unlikely to be a constant or evolutionarily irreversible character. They also lack some advanced characters of the Laetoli species: the sharp alteration in course, high pedicels, and a strong difference between the heights of the pedicel on its medial and lateral sides. Their basal swelling is not always present nor is it localized posterolaterally. The parietal boss is less obvious. It should also be reiterated that alcelaphine teeth from Langebaanweg are definitely more primitive than those from the Laetolil Beds. Once again D. acalla is better fitted for ancestry than D. neanica. The condition of the alcelaphine teeth at Langebaanweg indicates that Damalacra is more primitive than the alcelaphines hitherto considered from the Hadar Formation, Siwaliks or the Laetolil Beds. The clear implication is that Damalacra at Langebaanweg existed in an earlier time span. D. neanica had acquired some specializations of its own which make it unlikely to be ancestral to these forms. It was probably the end of an evolutionary line and it would be unwise to use its advanced characters to look for relationships with later alcelaphines (cf. Gentry in Hendey 1970: 116 in which D. neanica was con- fused with advanced Parmularius). The combination of some specializations of the horn-cores with an otherwise primitive skull is analogous to the living Connochaetes gnou which, despite its advanced horn-cores, has a short face and teeth which are occlusally simpler than other comparably-sized extant alcela- phines. One imagines that such combinations may have arisen with some frequency in alcelaphine evolution. There is some reason to believe that alcelaphine horn-core characters can change relatively rapidly. A possible interpretation of events in the Hadar Formation is the evolution of more upright insertions, more forward curvature, more distal divergence and a small degree of anteroposterior compression, all of which are a repeat of characters which had already appeared in Damalacra - neanica. Even at Langebaanweg itself D. acalla horn-cores of bed 3aN seem to be losing their backward curvature and acquiring more divergence distally, almost in imitation of the D. neanica which had been so abundant in bed 3aS. It is interesting that a Damalacra acalla horn-core such as L30215 in bed 3aN not only has a degree of distal divergence and little backward curvature, but also a base which has become squared off posteromedially, all of which foreshadow the Afar alcelaphine. The tendency of horn-cores to become 296 ANNALS OF THE SOUTH AFRICAN MUSEUM shorter from 3aS to 3aN could also be a means for them acquiring a more tapered appearance like the Afar alcelaphine. It is wise not to read too much into these supposed tendencies. More material from Langebaanweg is still being accessioned at the South African Museum and will provide a basis for a more thorough examination. One can have little confidence that all alcelaphine horn-core characters are stable or that, once acquired, they need be irreversible. However, it is unquestionable that Damalacra acalla, particularly as known from bed 3aS, is better fitted than D. neanica to be an ancestor of later alcela- phines by reason of its lack of specialized characters. Its only known locality at the southern end of Africa is an unlikely venue for evolutionary enterprise, but D. acalla or a closely related species further north in Africa is potentially an ancestor for the Afar alcelaphine and even for other species such as the Laetoli species represented by the cranium 1959.277. As yet there is little evidence of Damalacra-like alcelaphines further north in Africa. But there is some. A horn-core cast from Wadi Natrun, BM(NH) M8199, is the base of a much damaged horn-core, probably of the right side, and appears to be the one mentioned by Studer (1898: 76) and Andrews (1902: 438). It is labelled Hippotragus ?cordieri, but could belong to a Damalacra. A left upper molar from Garet el Muluk at Wadi Natrun figured by Stromer (1907: 120, pl. 20 (fig. 1)) and identified by him as perhaps tragelaphine, appears to represent an alcelaphine at an evolutionary level comparable with Dama- lacra. Its basal length is given as 21 mm. A cast of a left M; also from Wadi Natrun, BM(NH) M 12361, which Andrews (1902: 439, p. 21 (fig. 9)) thought was from a large gazelle-like form, could also represent an alcelaphine at the Damalacra \evel. Its occlusal length is 28,3 and its height (early wear) is c. 32 mm. It begins to look as if the evolutionary history of medium sized alcelaphines has been of successive replacements of one dominant group by another. One could conjecture that the living Alcelaphus and Damaliscus have replaced the various Parmularius species and Damaliscus niro of the Pleistocene. (They could also have come close to replacing Beatragus which seems to be a lineage on the verge of extinction.) Now it looks as if Parmularius and Damaliscus niro themselves had replaced earlier Pliocene alcelaphines such as the Afar species and Damalops, while Damalacra is giving us the first intimation of a still older stratum of alcelaphines. The actual phylogenetic path from one dominant group to its successor is still a conjectural matter. Parmularius can be con- vincingly derived from Laetoli 1959.277, but Damaliscus and Alcelaphus may come either from this species or from the stock containing the Afar species and Damalops palaeindicus. There is a possibility that the Afar species and Damalops palaeindicus are a Pliocene dispersal not ancestral to any later forms. It has been noted above that they could be descended from Damalacra acalla or some more northern Damalacra species, and it is also possible that 1959.277 has a similar ancestry. FOSSIL BOVIDAE FROM LANGEBAANWEG 297 Fig. 51. Lateral views of skulls and horn-cores of Damaliscus lunatus (left) and Alcelaphus buselaphus (right), shown at their normal inclinations when not feeding. _ Functional skull morphology and evolution in alcelaphines The two functions which mainly lead to variation of skull morphology among bovids are feeding (ingestion and mastication) and horn support. Other functions, such as breathing or input of sensory information, must not be impaired by changes in feeding habits or horn support, but are not themselves the cause of larger scale morphological changes. In living Damaliscus and Alcelaphus (Fig. 51) the most notable feature of the skull is the long face. The most likely explanation for this is that animals grazing at ground level need to have eyes as high as possible to avoid being surprised by predators. In line with this requirement they also show long diastemata, tooth-row forward of orbital level (especially in Alcelaphus), and the brain cavity becoming realigned diagonally instead of horizontally. When not feeding, Alcelaphus, and to a lesser extent Damaliscus, hold their heads more nearly vertical than non-alcelaphine antelopes. (When asleep while standing (Plessis 1972, fig. 11) their heads can swing even further to an almost inverted position.) This is mechanically more 298 ANNALS OF THE SOUTH AFRICAN MUSEUM convenient with long skulls in which the entire weight has to be supported on the rest of the body at the occipital condyles, and it also causes less obstruction to the field of vision. It is desirable in bovids for the horn-core insertions to be as high as possible on the skull, presumably to ensure maximum visual effect. Consequently, an antelope with a long face held vertically could well have insertions in an extreme position behind (now = above) the orbits. This has happened in Alcelaphus buselaphus where the insertions are close together on a united pedicel, and in A. lichtensteini where they are wide apart but also high. The changed position of the insertions is also linked with horn-core curvature being forward rather than backward. This is a means of ensuring that the distribution of weight in relation to the condyles continues to be balanced. If a Damaliscus evolved posterior/high insertions like Alcelaphus, then it would have to evolve either a more vertical carriage of its head or forward curvature of its horn-cores to avoid a weight imbalance on the occipital condyles. It is difficult to assess how the different structure of the horns in Alcelaphus and Damaliscus affects dominance-testing encounters between conspecific males. In high intensity exchanges Damaliscus kneels on its carpal joints, may even hold its forehead to the ground and the horns of the opponents interlock (David 1973, fig. 11f; Lynch 1974: 37; Monfort-Braham 1975, fig. 6). Alcelaphus also locks horns but they do not appear to get to the stage of pressing their foreheads to the ground. The tips of all Alcelaphus horn-sheaths are turned backward, unlike Damaliscus, so in this position Alcelaphus would more readily injure one another. Walther (1972: 403) notes that hartebeests do injure themselves more often than most other horned ungulates, so perhaps this is what actually happens. It is apparent from the foregoing comments that the major differences between Damaliscus and Alcelaphus skulls are a single suite of functionally and mechanically linked characters. This must apply also to the differences of Damalacra neanica from D. acalla. There must have been ecological opportunity for two species to coexist, just as at the present day, and one of them evolved similar but less extreme differentiating characters of posterior horn-core inser- tions, shortened and inclined braincase roof, horn-cores curving less backward and even forward, and horn-core compression being more anteroposterior than mediolateral. It may be that pairs of sympatric alcelaphines have repeatedly evolved similar differentiating characters. In middle and upper Bed II at Olduvai Gorge Parmularius angusticornis had a very short braincase and horn-cores which lack backward curvature. Damaliscus niro in the same deposits had backwardly curved horn-cores and may also have had a longer braincase. If the above reasoning is correct then early alcelaphines at Langebaanweg had already become grazers at ground level. This conclusion is compatible with the presence of horn-cores in female Damalacra. It seems from extant bovids that. horned females are more characteristic of larger than smaller species and that in Africa horned females are more frequent in species living in open habitats. FOSSIL BOVIDAE FROM LANGEBAANWEG 299 Comparison of Langebaanweg alcelaphines with early caprines _ The early caprine Pachytragus is known from the Turolian fauna of Samos, Greece, where it has two species, P. crassicornis and P. laticeps, discussed by Gentry (1971). Pachytragus shares with Damalacra the basic characters of aegodont antelopes such as the rather narrow skull, inclined braincase roof, hypsodonty of the cheek teeth, reduction of basal pillars on the molars and shortening of the premolar row. P. crassicornis and D. neanica are easily dif- ferentiated by a number of individual specializations, but P. Jaticeps and D. acalla are morphologically more primitive, and it is necessary to point out how they differ enough to avoid being placed in one genus. D. acalla shows: 1. A tendency to basal thickening of its horn-cores usually on the medial surface | 2. Its horn-cores are less compressed mediolaterally 3. Their distance apart, measured across the lateral sides of the pedicels, is greater; this character is probably linked with the last 4. The internal hollowing of the horn pedicels has been carried much further 5. The frontals are more raised between the horn-core bases (linked with the last character) 6. The midfrontals suture is less complex and less raised into a ridge 7. The supraorbital pits are smaller 8. The supraorbital pits are situated more widely apart 9. The back of the braincase is wider 10. The occiput is lower 11. The mastoid is probably larger 12. The lower molars have less flattening of their medial walls 13. The P, is more strongly reduced Characters 1, 4, 5 and 13 look like specializations in D. acalla and, on the hypothesis that Langebaanweg is the younger site, could have been acquired _ during descent from an earlier Pachytragus. Characters 2, 3, 6, and 12 seem to be specializations in P. Jaticeps and to indicate a different direction of evolu- tion from Damalacra. In the case of 7 to 11 either species could show the more advanced condition, nevertheless the characters add to the ‘morphological distance’ between the two species. The total morphological difference between the two species is too great for them to be regarded as congeneric. Their common ancestry, if one assumes the monophylety of aegodont antelopes, must lie _ further back in time. Tribe Neotragini Genus Raphicerus H. Smith, 1827 Type species Raphicerus campestris (Thunberg). Generic diagnosis Moderate sized to large neotragines. Horn-cores short to moderately long 300 ANNALS OF THE SOUTH AFRICAN MUSEUM with little mediolateral compression, inserted widely apart above the back of the orbits, parallel to one another, and having a slightly concave front edge in profile. Postcornual fossa present. Supraorbital pits wide apart, back of braincase roof not very strongly turned down, temporal lines wide posteriorly on cranial roof, preorbital fossa moderate sized to large, premaxilla wide and rising to contact nasals, auditory bulla inflated, median indentation at back of palate level with or forward of lateral ones, palatal ridges on maxilla anterior to the tooth row approach one another closely. Upper molars with quite small styles, central cavities of lower molars disappear early in wear, medial walls of lower molars fairly flat and metastylids not strong, M,s with moderate to large back lobes. Metaconid of P, passes transversely then backwards, front of lateral wall of P, bends round into a transverse plane, P, not greatly reduced. Remarks The type species occurs in most of southern Africa and Rhodesia, and also in Tanzania and Kenya. There are two other living species, R. melanotis (Thun- berg) which is largely confined to the Cape Biotic Zone of South Africa, and R. sharpei O. Thomas found in Mozambique, Malawi, Rhodesia, Zambia and parts of surrounding countries. Klein (1976) has shown that R. melanotis and R. campestris were already separate species in the early Upper Pleistocene of the southern Cape Province. R. campestris lives in more open country than the other two species and grazes more frequently (Klein 1976: 171-172 and references). Raphicerus paralius sp. nov. Figs 52-53, 55-57 ?Madoqua sp. Gentry in Hendey 1970: 116. Holotype L12238—right horn-core, index 15,5 x 15,2, and associated right maxilla with P?—M® in early middle wear (Fig. 52). Referred material From QSM: L21143, L21146 (Fig. 53)—left horn-cores, 18,1 < 14,7 and length c. 53,0, 753, Xone L22504, L41606—right horn-cores, 15,7 x 16,0, 17,6 x 17,2 L41643—left and right horn-cores, the left with an index 17,4 x 15,6, also another left horn-core with index 19,7 x 19,2 L12513—left and right maxillae with P?-M® and P?—P* respectively in early middle wear L22635—left upper molar L41607—right maxilla with damaged P?-M? L41645—right mandible with M, and Mg in early middle wear L41666—left mandible with dP, in early middle wear FOSSIL BOVIDAE FROM LANGEBAANWEG 301 Fig. 52. Raphicerus paralius. L12238, holotype. Ventrolateral view of right horn-core. Ventral view of most of right palate and right cheek tooth-row. Scale = 10 mm. Fig. 53. Raphicerus paralius. L21146, L6565, anterior views of left horn-cores. Scale = 10 mm. ANNALS OF THE SOUTH AFRICAN MUSEUM Probably from QSM: L41703—right horn-core, 17,8 x 13,4 and length 67,6 From bed 3aS: L40413—right horn-core, 17,3 x 16,7 L40132A—left maxilla with dP?—M? L40270—left mandible with dP,-M, L40443—left maxilla with P?-M®? and associated right upper teeth in late middle wear L41245—left mandible with M,—Msg, part of right mandible in early wear (Figs 55, 57) L41320—right mandible with P,—M, in early middle wear (Fig. 56) L41526—two left upper molars, left lower molar, other teeth L40088—right radius with proximal surface and most of shaft (Fig. 21) L40021—complete left radius with length and least transverse thickness of 145 x 14,9 mm Probably from bed 3aS: L10788, L10789, L10931—right upper molars L10790—left upper molar L11008—right M, L11197, L11198—left lower molar, right dP® Probably from bed 3aS, but a few possibly from QSM: L5412, L6565 (Fig. 53), L6566, L6573—left horn-cores, 20,1 x 16,2, 20,4 <8. 7 1505 13,7... 16:3 > 16 L6567—left horn-core L2931—right horn-core, 18,4 x 18,2 L3132—right upper molar L5312—right mandible with damaged M,—M, L6600—left mandible with M, and M, From QSM or bed 3aS: L11157—right horn-core, 17,1 < 15,7 L11978—left mandible with M, L41686—fragmentary right and left dP?s L9939—distal left humerus L41684—distal right tibia, partial left naviculocuboid, proximal left meta- tarsal, much of distal right radius, fragmentary distal metacarpal, terminal phalanx (Fig. 21). All associated with a pair of Raphicerus horn-cores but the distal tibia, naviculocuboid and proximal meta- tarsal are from a larger animal than the distal radius and metacarpal. Also one side of a juvenile distal metapodial which must be a third individual From bed 3aN: L45170—right mandible with remains of dP,-dP, in early middle wear above P,-P,, M, and M, FOSSIL BOVIDAE FROM LANGEBAANWEG 303 From QSM or bed 3aN: - L40787—distal left humerus (Fig. 21) Horizon The holotype comes from the QSM. The species is almost confined to the QSM and bed 3aS of the PPM. Diagnosis A Raphicerus considerably larger than the three living species. Horn-cores short and thickened basally, generally with a posterolateral keel, a tendency towards a longitudinal concavity in front of it on the posterior half of the lateral surface, sometimes a medial or anteromedial keel and other irregular ridges all of which combine to give an irregularly shaped cross-section. Insertions of horn-cores at a lower angle than in living Raphicerus. Postcornual fossa well marked. Supraorbital pits not obscured by overgrowth of the frontals. Preorbital fossa large and deep. Infraorbital foramen above P? or the front part of P®. Basal pillars present on M, and traces of them on other lower molars. Premolar row long. P? and P?® larger than in living Raphicerus. Short diastema. Etymology The specific name comes from the Greek paralios, by the sea, and refers to the type locality for this species being in a coastal region. Remarks These fossils are larger than any living neotragine. There is no evidence from the horn-cores (Fig. 54) of a size increase having taken place from the QSM to the PPM. The horn-cores are nearly as short as in R. sharpei. Their keels and irregular cross-section are more pronounced than in any other Raphi- cerus, but this may be an allometric feature of large horn-cores in the Neo- tragini, as indicated more faintly in larger examples of Neotragus, Dorcatragus and Oreotragus. The large and deep or moderately deep postcornual fossae are most like R. campestris among living species. There are no sinuses in the frontals. The triangular supraorbital pit is large but shallow round the two foramina in L12238A, but in L6565 and L41643 the pits are smaller and have more resemblance to later Raphicerus. The absence of overgrowth by the frontals is like R. sharpei and unlike R. campestris or melanotis. The lowness of the horn-core insertions is only apparent from a few specimens, e.g. L21146, L41606, and the left side of L41643, in which sufficient of the frontals posteromedially to the horn-core has been preserved to show a somewhat Madoqua-like aspect of the insertion. Such low insertion angles are more unlike R. campestris than R. melanotis or sharpéi. The maxillae L12238 and L12513B show that there was probably a large and deep preorbital fossa which evidently passed low on the face and far anteriorly. This large fossa would be a resemblance to R. melanotis. The infra- orbital foramen is low over the back of P? or front of P’, unlike most R. sharpei but resembling R. campestris and melanotis. The palatal ridges in front of the 304 ANNALS OF THE SOUTH AFRICAN MUSEUM 207 Mediolateral diameter Anteroposterior diameter ne I5 20 mm Fig. 54. Basal diameters of neotragine horn-cores. O = Raphicerus paralius from QSM, X = the same from bed 3aS; underlined readings are lefts, others rights. a = Raphicerus from lower assemblage in Baard’s Quarry, m = Makapansgat Limeworks, s = R. melanotis from Swartklip, c = extant R. campestris, dots = Elandsfontein Raphicerus sp. of the right side. Upper diagonal line = 100%, lower one = 66,7 % as in Figure 3. tooth-row on the maxilla L12238 converged and touched at the midline. Judged by its sockets, the missing P? was a large tooth on L12238, as it is on L12513B where it is still present. P? was also large. The large size of these anterior pre- molars is reminiscent of Madoqua or Raphicerus melanotis and sharpéi. It is not easy to distinguish the teeth of Raphicerus paralius from those of the gazelle at Langebaanweg. This problem will be discussed on page 313. The teeth accepted as neotragine show the following characters. They are large FOSSIL BOVIDAE FROM LANGEBAANWEG 305 19: CE oe: Ci en eeinaene ean NTE EO . a Fig. 55. Raphicerus paralius, L41245, left lower dentition in occlusal view. Gazella sp., L40603, right lower dentition in occlusal view. Scale = 10 mm. se ath sm myn. ath ie pT Meee My om Oe a lanai ee Sl te Fig. 56. Raphicerus paralius. L41320, right lower dentition in occlusal view. Scale = 10 mm. MG a TE AS A nego Fig. 57. Raphicerus paralius, L41245, left mandible in lateral view. Gazella sp., L40603, right mandible in lateral view. Scale = 10 mm. 306 ANNALS OF THE SOUTH AFRICAN MUSEUM 307 Length p¢_p4 ; Length M'-m> 20 25 30 35mm Fig. 58. Occlusal lengths of upper premolar and molar rows in Neotragini. X = Raphicerus paralius L12513 from QSM Langebaanweg, C = extant R. campestris, m = extant R. melanotis, S = extant R. sharpei. Upper diagonal line = 100%, lower one = 75% as in Figure 3. compared with living neotragines (Fig. 58). The upper molars have quite small styles. On the lower molars the central cavities disappear early in wear, the lateral lobes are only drawn out a little in a transverse direction, the medial walls are fairly flat, and metastylids are not strong. There are small basal pillars on M, and sometimes persistent traces of the basal pillars on other molars as in L11978. The rear lobe on Mg is small to moderate sized and in three out of four specimens it shows a flange posteriorly. The premolar row is rather long as deduced from a number of incomplete specimens. On P, the hypoconid projects slightly, the metaconid tends to be oriented transversely in its lateral part, the paraconid is joined to the parastylid and shows no approach towards the metaconid behind it, and the front of the lateral wall of the tooth tends to bend round into a transverse plane. The metaconid tends to be in a trans- verse plane (at least in early wear) in P, as well. The P, is large and shows little sign of having been reduced in size. The diastema is short and curved upwards and the lower edge of the mandible is curved. Nearly all these characters are FOSSIL BOVIDAE FROM LANGEBAANWEG 307 very similar to those of living Raphicerus. The tendency for the metaconid to be oriented transversely on P; as well as P, is one of the few resemblances to R. campestris. The curved lower edge of the mandible and the long premolar row are both resemblances to R. melanotis and sharpei. Otherwise the Lange- baanweg species differs from Raphicerus only in its large size, the presence of basal pillars on M, and possibly on more posterior teeth as well, and in its shorter diastema. It differs from other living neotragines as follows. Compared with Madoqua it is much larger, basal pillars are present at least on M,, M3; always has a third (rear) lobe, the hypoconid on P, projects less than it sometimes can in Madoqua, and P, is less reduced. It differs from Neotragus by much larger size, poorer styles on upper molars, basal pillars present at least on M,, normally a flange on the back of M;, longer premolar row, the metaconid of P, less clearly diagonal, and a short diastema. It differs from Oreotragus by being a little larger, basal pillars present at least on M,, normally a flange on the back of M.,, and the metaconid of P, less clearly diagonal. It differs from Dorcatragus by larger size, longer premolar row, absence of a slight tendency for metaconid and paraconid of P, to approach, and a short diastema. The distal tibia, L41684, has a long medial malleolus in side view; however its main facets are not greatly indented in the middle of their posterior edge so it is unlikely to be of the gazelle. The front facet for the fibula is small but not minute in comparison with the back one, more or less in front of the back one instead of anteromedial to it, and without a deep indentation between it and the back facet. The back of the rear fibula facet is not much anterior to the rear edge of the bone as a whole. The tibiae of three out of four representatives of extant Raphicerus campestris were more indented at the back of the articular facets, two of them were more deeply indented between the fibula facets, and one had the front fibula facet sited anteromedially rather than anteriorly to the posterior one. A single example of Dorcatragus megalotis, the living beira of Somalia which is about the size of R. campestris, has a very small front fibula facet but is otherwise like R. campestris. Three examples of Oreotragus oreotragus, another living neotragine of about the size of R. campestris, tend to have more massive medial malleoli and more attenuated central ones than in either R. campestris or the fossil Raphicerus. O. oreotragus has rather distinctive limb bones, presumably because its preferred habitat is rocky slopes and outcrops. The metatarsal associated with the tibia L41684 has no longitudinal groove on its anterior surface but some development of a posterior longitudinal groove. - The foramen at the top of the posterior surface is set very deeply. The main facet for articulation with the naviculocuboid curves rather strongly upwards at the back and in side view the top articular surface appears far from flat. There is no groove between the front and back naviculocuboid facets. there are clear anterior and medial sides of the main ectocuneiform facet, and there is a swollen rugose area at the top of the medial side of the shaft. The bone may not have been very long when complete. The metatarsals of three R. campestris show 308 ANNALS OF THE SOUTH AFRICAN MUSEUM more differences from the Langebaanweg fossil than did the tibiae. There is perhaps less development of the posterior longitudinal groove, the foramen at the top of the posterior surface is set less deeply, the main ectocuneiform facet does not show clear medial and anterior edges, there is no swollen rugose area at the top of the medial side of the shaft, and the top of the articular surface as a whole is anteroposteriorly longer. The single Dorcatragus megalotis is like the living Raphicerus except that the rear of the main naviculocuboid facet is less upcurved and the articular surface consequently appears flatter in medial profile. The three Oreotragus oreotragus metatarsals are notably short and have more of a groove behind the main naviculocuboid facet. However, two of them have distinct anterior and medial edges to the main ectocuneiform facet as in the Langebaanweg fossil, and in all three of them the articular surface is even more compressed anteroposteriorly. On the distal radius L41684 the back of the medial facet for articulating with the scaphoid is not very deep and hence is unlike a gazelle. The cuneiform articulation has quite a large area on the radius. The distal end as a whole looks a little swollen in side view, and the flanges on the anterior surface are wide apart and the surface between them not very hollowed. The distal end of the complete radius L40021 is larger than that of L41684, but its characters are similar. The articular surface for the lunate is not deeply excavated. The proximal radius L40088 has hardly any development of a medial rim to its medial facet, the back edge of the lateral facet is not set anteriorly, and the tubercle at the top of its lateral surface is small and situated below the level of the articular facets. The radii of four available examples of R. campestris and the single D. megalotis have the distal flanges on the anterior surface closer together than in the Langebaanweg fossils but are otherwise similar. The three ‘O. oreotragus have even less of a rear medial facet for the scaphoid and a very small or non-existent radial articulation for the cuneiform. However, the anterior flanges are wide and little pronounced as at Langebaanweg. The distal humeri L9939 and L40787 have slanted not upright condyles, a deep coronoid fossa, only a shallow hollow for the o11gin of the lateral humero- radial ligament, the posterior ridge for this hollow situated well forward from the rear edge of the bone, and a moderately developed ridge on the lateral surface marking the origin of the extensor carpi radialis. The humeri of R. campestris and D. megalotis have a deeper hollow for the lateral humeroradial ligament. Less certain differences are the possibility of more upright condyles, a higher medial condyle, and the ridge demarcating the posterior limit of the origin of the lateral humeroradial ligament being less anteriorly placed. O. oreotragus has a distinctive appearance in anterior view by being transversely wide across the medial condyle and by having a low lateral condyle. However, it does have a shallow hollow for the humeroradial ligament and the posterior ridge to this hollow is situated anteriorly. 309 TOC SI SLLOT'T 405 "WW Mojeq yidop snuey “HTT "AW 9680IT {9'PI ‘2 IW SOLS sore “ds D1J2ZDH JO Syysus] [esnjs00 J9yIO OEL "AP OLISVT ‘ECE "AP 999TH ‘0°91 “WW SOOTIT sore smposvd snaaoiydoy JO syydug] [esnfd90 JOyIO o “smijpddd Snaaaiydoy OJ SULUN[OD se14} 4SI OY} ydooxo suOTUEp JAMO] 9B [[V Bee Ie ee Ge ee ES ee og == 7° SW 4ojeq yidep snuey < LS = .9n => VL ef 61. Z81 = hie, 6L9. SPE = " "TAL Mojeq yidep snuey ee Gy Saec Ul = =" 406 LS == 9'6 ras. aL 6 Sy SC 78 9°L 7% * pe YBUA] [eSN[INO O = = ie oe ey ee. = ee lb i PY aster TEsNIa9@ Ze = a eee Ce. = = a oe Se = =z -5.0'6¢ = " * * pd-Zd YSU] [esn[s99 - ay ae 5 a | = =o BC bh =i. OPI =P Fle Sl = = = "7 * WW y8ue] [esnjs99 2 OcCl-wicel , “Let - Uri a — <0 2 LOL -CCh Sil Orel ae Ie « Orel SCC 7 "ZW yisusy Tesnjs9O e rae e CY regen. Ger. = = = Gs Le eee ao. CVE EOE 2 OSCE " "EWTN Wi8u9] [esnjs99 = te ea ew ey ea S Es Hel el Is fee See ee eee Se 2 ee Bee Ge Q No) Oo (ore) Vo) o a Q NS R S S g 1S G2 e8 > fo) Ww 1oS) \O Nn [=e) NH [e.2) Oo (oS) (oe) 2 =| 2 ‘ds pjjazvy Snyosod snsaaiydoy fy ‘SuUOI}USp ouIdo][Ue pue sUISeIJOOU JO s}UOWOINSLO;| 9 AIEV 310 ANNALS OF THE SOUTH AFRICAN MUSEUM Measurements Measurements of horn-cores have already been given. Those of teeth are given in Table 6. Comparisons The Langebaanweg Raphicerus is larger than the living species of that genus. As has been seen, it differs most from R. campestris among living species, and rather less from R. melanotis or sharpei. It would not be very convincing to link the Langebaanweg with Mcderh Raphicerus except by way of congeneric Elandsfontein fossils, comprising one or two crania, some frontlets, many horn-cores and many dentitions. Most of the Elandsfontein horn-cores differ from living species only by their larger size (Fig. 54) and less upright insertions. They often have an approach to a posterolateral keel and the right horn-core on the frontlet SAM—PQ-E14153 has a mid-lateral keel with a longitudunal hollow between it and the postero- lateral keel which suggests a link with the Langebaanweg species. Another Elandsfontein horn-core, SAM-—PQ-E927, has an anterolateral keel. A right horn-core from Makapansgat Limeworks, BPI M478, is like those at Langebaanweg and may be identified as Raphicerus ?paralius. Its basal index is 16,0 x 17,5 and length is 80 mm. It was published as Cephalophus pricei by Wells & Cooke (1956: 12, fig. 6), but the tooth-rows assigned to this species by the same authors are from a bushbuck-sized tragelaphine and one of them is the holotype. A pair of large neotragine horn-cores from member G of the Shungura Formation, Omo 280 71-1168 and 1169, have an index of 18,7 x 15,9 and are like R. paralius in their shortness, poor degree of compression and irregularly shaped cross-section arising from strong longitudinal ridging and grooving. ‘However, their insertion is at a less low inclination, they show slight backward curvature and the postcornual fossa is weak, so their identification as Raphicerus is not very secure. Moreover, neotragine tooth-rows from low in member G which may or may not be conspecific with the horn-cores, e.g. Omo 75i 70-1106 and L 5044, have smaller teeth than at Langebaanweg and the former specimen is complete enough to show that the premolar row is much shorter. Tribe Antilopini Gazella sp. Figs 55, 57, 59 Material A number of horn-cores and dentitions belong to a gazelle. The horn-cores and their basal indices are as follows: From QSM: Left Right L13208 14,3 x 11,4 (female) L13984 part of horn-core L20510 13,5 x 11,8 (female) FOSSIL BOVIDAE FROM LANGEBAANWEG 311 From bed 3aS: Left Right 140324 26,4 x 20,5 L40097 = 32,1 x 24,3 L40390 =31,8 x 24,0 L40179 30,2 x 24,3 P41325 30,8 x’ 22,] LA0277 © 2836 < 20,5; L41528 distal part only length c. 130 L40389 14,7 x 13,0 (female) Probably from bed 3aS, but a few possibly from QSM: Left Right Weest -25,3' x 18:5 E261 ae 295222 L3491 B07 < 2032 (Fig. 59) L2620 26,6 x 19,9 Mai? ~ 27;7. x 22,6 MSU25 PZ S222 Ha07s ~ 28,3) :23;1 L3196 Rose! 243°< 19,1 £3206)" 28) wae a + Si settee a a =o 334 ANNALS OF THE SOUTH AFRICAN MUSEUM the preparation of this paper. I also thank my wife for much help. Financial support for visits to Cape Town was given by the Wenner Gren Foundation for Anthropological Research, New York. Dr H. Thomas kindly allowed me to quote from an unpublished paper. REFERENCES ANDREE, J. 1926. Neue Cavicornier aus dem Pliocan von Samos. Palaeontographica 67: 135-175. ANDREWS, C. W. 1902. Note on a Pliocene vertebrate fauna from the Wadi-Natrun, Egypt. Geol. Mag. (4) 9: 433-439. ANSELL, W. F. H. 1971. Artiodactyla. Jn: MEESTER, J. & SETZER, H. W., eds. The mammals of Africa: an identification manual. 15: 1-93. Washington D.C.: Smithsonian Institution Press. ARAMBOURG, C. 1947. Contribution a l’étude géologique et paléontologique du bassin du lac Rodolphe et de la basse vallée de ’'Omo. Mission scient. Omo 1 Géol. Anthrop. (3): 232-562. ARAMBOURG, C. & PIVETEAU, J. 1929. Note préliminaire sur un Ruminant du Pliocéne inférieur du Roussillon. C.r. somm. Séanc. Soc. géol. Fr. 10: 144-146. ARONSON, J. L., SCHMITT, T. J., WALTER, R. C., TAIEB, M., TIERCELIN, J-J., JOHANSON, D. C., NAESER, C. W. & Nairn, A. E. M. 1977. New geochronologic and palaeomagnetic data for the hominid-bearing Hadar formation of Ethiopia. Nature, Lond. 267: 323-327. BEHRENSMEYER, A. K. 1976. Lothagam Hill, Kanapoi, and Ekora: a general summary of stratigraphy and faunas. Jn: CoppENs, Y., HOWELL, F. C., ISAAC, G. Ll. & LEAKEY, R. E. F., eds. Earliest man and environments in the Lake Rudolf basin: 163-170. Chicago: University of Chicago Press. BERGGREN, W. A. & VAN COUVERING, J. A. 1974. The late Neogene. Palaeogeogr. Palaeoclimat. Palaeoecol. 16: 1-216. BisHop, W. W., CHAPMAN, G. R., HILL, A. & MILLER, J. A. 1971. Succession of Cainozoic vertebrate assemblages from the northern Kenya Rift Valley. Nature, Lond. 233: 389-394. BisHop, W. W., MILLER, J. A. & Fitcu, F. J. 1969. New potassium-argon age determinations relevant to the Miocene fossil mammal sequence in East Africa. Am. J. Sci. 267: 669-699. BisHop, W. W. & PICKFORD, M. H. L. 1975. Geology, fauna and palaeoenvironments of the Ngorora Formation, Kenya Rift Valley. Nature, Lond. 254: 185-192. Boaz, N. T., GAziry, A. W. & EL-ARNAUTI, A. 1979. New fossil finds from the Libyan Upper Neogene site of Sahabi. Nature Lond. 280: 137-140. Broom, R. 1937. Notices of a few more new fossil mammals from the caves of the Transvaal. Ann. Mag. nat. Hist. (10) 20: 509-514. Brown, F. H., HowELL, F. C. & Eck, G. G. 1978. Observations on problems of correlation of late Cenozoic hominid-bearing formations in the North Lake Turkana Basin. Jn: BisHop, W. W., ed. Geological background to fossil man: 473-498. Geol. Soc. Spec. Publ. no. 6. Edinburgh: Scottish Academic Press. BUTZER, K. W. & THURBER, D. L. 1969. Some late Cenozoic sedimentary formations of the Lower Omo Basin. Nature, Lond. 222: 1138-1143. CooKkE, H. B. S. & CorYNDON, S. 1970. Pleistocene mammals from the Kaiso Formation and other related deposits in Uganda. Jn: LEAKEY, L. S. B. & SAVAGE, R. J. G., eds. Fossil Vertebr. Afr. 2: 107-224. London: Academic Press. CopPENS, Y., HOWELL, F. C., IsAAc, G. LI. & LEAKEY, R. E. F., eds. 1976. Earliest man and environments in the Lake Rudolf basin. Chicago: University of Chicago Press. Dav, J. H. M. 1973. The behaviour of the bontebok, Damaliscus dorcas dorcas (Pallas, 1766), with special reference to territorial behaviour. Z. Tierpsychol. 33: 38-107. DELson, E. 1975. Paleoecology and zoogeography of the Old World monkeys. Jn: TUTTLE, R. H., ed. Primate functional morphology and evolution: 37-46. The Hague: Mouton. DEPERET, C. 1890. Les animaux pliocénes du Roussillon. Mém. Soc. géol. Fr. Paléont. 3: 1-194. DieTRIcH, W. O. 1941. Die sdugetierpalaontologischen Ergebnisse der Kohl-Larsen’schen Expedition 1937-1939 im n6rdlichen Deutsch-Ostafrika. Zentbl. Miner. Geol. Puldont. (B) 8: 217-223. FOSSIL BOVIDAE FROM LANGEBAANWEG 335 DietricH, W. O. 1942. Altestquartare Saugetiere aus der siidlichen Serengeti, Deutsch- Ostafrika. Palaeontographica 94A: 43-133. Dietricu, W. O. 1950. Fossile Antilopen und Rinder Aquatorialafrikas. Palaeontographica 99A: 1-62. DItTRIEVA, E. L. 1977. Tajikistan’s and India’s fossil Alcelaphinae. J. Palaeont. Soc. India 20: 97-101. ERDBRINK, D. P. B., PRieM, H. N. A., HEBEDA, E. H., Cup, C., DANKERS, P. & CLOETINGH, S. A. P. L. 1976. The bone bearing beds near Maragheh in N.W. Iran. Proc. Kon. Ned. Akad. Wet. (B) 79, 2: 85-113. FALCONneER, H. 1859. Descriptive catalogue of the fossil remains of Vertebrata... in the museum of the Asiatic Society of Bengal. Calcutta. GENTRY, A. W. 1970a. The Bovidae (Mammalia) of the Fort Ternan fossil fauna. In: LEAKEY, L. S. B. & SAVAGE, R. J. G., eds. Fossil Vertebrates of Africa 2: 243-324. London: Aca- demic Press. GENTRY, A. W. 19705. Revised classification for Makapania broomi Wells & Cooke (Bovidae, Mammalia). Palaeont. afr. 13: 63-67. GENTRY, A. W. 1971. The earliest goats and other antelopes from the Samos Hipparion fauna. Bull. Br. Mus. nat. Hist. (Geol.) 20: 229-296. GENTRY, A. W. 1974. A new genus and species of Pliocene boselaphine (Bovidae, Mammalia) from South Africa. Ann. S. Afr. Mus 65: 145-188. GENTRY, A. W. 1976. Bovidae of the Omo Group deposits. Jn: CopPENS, Y., HOWELL, F. C., Isaac, G. LI]. & LEAKEY, R. E. F., eds. Earliest man and environments in the Lake Rudolf basin: 275-292. Chicago: University of Chicago Press. GENTRY, A. W. 1978a. The fossil Bovidae of the Baringo area, Kenya. Jn: BisHop, W. W.., ed. Geological background to fossil man: 293-308. Geol. Soc. Spec. Publ. no. 6. Edinburgh: Scottish Academic Press. GENTRY, A. W. 19785. Bovidae. In: CooKE, H. B. S. & MAGLIO, V. J., eds. Evolution of African mammals: 540-572. Cambridge, Mass.: Harvard University Press. ‘GENTRY, A. W. & GENTRY, A. 1978. Fossil Bovidae (Mammalia) of Olduvai Gorge, Tanzania. Parts I-II. Bull. Br. Mus. nat. Hist. (Geol.) 29: 289-446; 30: 1-83. GERVAIS, P. 1852. Zoologie et paléontologie Frangaises. Paris. GERVAIS, P. 1853. Description des ossements fossiles de mammiféres rapportés d’Espagne par M M. de Verneuil, Collomb et de Loriére. Bull. Soc. géol. Fr. (2) 10: 147-167. GUERIN, C. 1975. Les rhinocéros (Mammalia, Perissodactyla) des gisements Pliocénes Frangais: intérét biostratigraphique et paléoécologique. Colloques int. Cent. natn. Res. scient. 218 Problémes actuels de paléontologie (évolution des vertébrés): 739-747. HENDEY, Q. B. 1970. A review of the geology and palaeontology of the Plio/Pleistocene deposits at Langebaanweg, Cape Province. Ann. S. Afr. Mus. 56: 75-117. HENDEY, Q. B. 1973. Fossil occurrences at Langebaanweg, Cape Province. Nature, Lond. 244: 13-14. HENDEY, Q. B. 1974. The late Cenozoic Carnivora of the south-western Cape Province. Ann. S. Afr. Mus. 63: 1-369. HENDEY, Q. B. 1976. The Pliocene fossil occurrences in ‘E’ Quarry, Langebaanweg, South Africa. Ann. S. Afr. Mus. 69: 215-247. HENDEY, Q. B. 1978a. The age of the fossils from Baard’s Quarry, Langebaanweg, South Africa. Ann. S. Afr. Mus. 75: 1-24. ; HENDEY, Q. B. 19785. Late Tertiary Hyaenidae from Langebaanweg, South Africa, and their relevance to the phylogeny of the family. Ann. S. Afr. Mus. 76: 265-297. HENDEY, Q. B. & HENDEY, H. 1968. New Quaternary fossil sites near Swartklip, Cape Province. Ann. S. Afr. Mus. 52: 43-73. KLEIN, R. G. 1975. Palaeoanthropological implications of the nonarchaeological bone assem- blage from Swartklip I, south-western Cape Province, South Africa. Quartern. Res. 5: 275-288. KLEIN, R. G. 1976. The fossil history of Raphicerus H. Smith 1827 (Bovidae, Mammalia) in the Cape biotic zone. Ann. S. Afr. Mus. 71: 169-191. KEIN, R. G. 1978. The fauna and overall interpretation of the “Cutting 10’ Acheulean site at Elandsfontein (Hopefield), south-western Cape Province, South Africa. Quatern Res. 10: 69-83. Ye | med AH wi / ‘a7 £ 4 Sa i >,>SL_e eee > FF a7. «| + Ss 2 wa Ta. * | = Mani ”* I seeeesaa! —-P) ~~ SS 336 ANNALS OF THE SOUTH AFRICAN MUSEUM LEAKEY, L. S. B. 1965. Olduvai Gorge 1951-61.1 Fauna and Background. Cambridge: University Press. LEAKEY, M. D. 1971. Olduvai Gorge 3 Excavations in Beds I and II, 1960-1963. Cambridge: University Press. Leakey, M. D., Hay, R. L., Curtis, G. H., DRAKE, R. E., JackEs, M. K. & Wuite, T. D. 1976. Fossil hominids from the Laetolil Beds. Nature, one 262: 460-466. LEAKEY, M. D. & Hay, R. L. 1979. Pliocene footprints in the Laetolil Beds at Laetoli, northern Tanzania. Nature Lond. 278: 317-323. LYDEKKER, R. 1878. Crania of ruminants from the Indian Tertiaries, and supplement. Mem. geol. Surv. India Palaeont. indica (10) 1: 88-181. LYDEKKER, R. 1886. Siwalik Mammalia, Supplement I. Mem. geol. Surv. India Palaeont. indica (10) 4: 1-21. Lyncu, C. D. 1974. A behavioural study of blesbok, Damaliscus dorcas phillipsi, with special reference to territoriality. Mem. natn. Mus. Bloemfontein 8: 1-83. Maa1uio, V. J. 1973. Origin and evolution of the Elephantidae. Trans. Am. phil. Soc. (n.s.) 63, 3: 1-149. MEIN, P. 1975. Proposition de biozonation du Néogéne Méditerranéen a partir des mammi- féres. Trab. Neogeno-Cuaternario 4: 112-113. MOoNFORT-BRAHAM, N. 1975. Variations dans la structure sociale du topi, Damaliscus korrigum Ogilby, au Parc National de l’Akagera, Rwanda. Z. Tierpsychol. 39: 332-364. PARTRIDGE, T. C. 1978. Reappraisal of Sterkfontein hominid site. Nature, Lond. 275: 282-287. Petroccui, C. 1956. I Leptobos di Sahabi. Boll. Soc. geol. Ital. 75, 1: 206-238. PICKFORD, M. 1975. Late Miocene sediments and fossils from the northern Kenya Rift Valley. Nature, Lond. 256: 279-284. PICKFORD, M. 1978a. Geology, palaeoenvironments and vertebrate faunas of the mid-Miocene Ngorora Formation, Kenya. Jn: BisHop, W. W., ed. Geological background to fossil man: 237-262. Geol. Soc. Spec. Publ. no. 6. Edinburgh: Scottish Academic Press. PICKFORD, M. 19785. Stratigraphy and mammalian palaeontology of the late-Miocene Lukeino Formation, Kenya. Jn: BisHop, W. W., ed. Geological background to fossil man: 263-278. Geol. Soc. Spec. Publ. no. 6. Edinburgh: Scottish Academic Press. PILBEAM, D., BARRY, J., MEYER, G. E., SHAH, S. M. I., PickForD, M. H. L., BisHop, W. W., THOMAS, H. & JAcoss, L. L. 1977. Geology and palaeontology of Neogene strata of Pakistan. Nature, Lond. 270: 684-689. PILGRIM, G. E. 1939. The fossil Bovidae of India. Mem. geol. Surv. India Palaeont. indica (n.s.) 26: 1-356. PILGRIM, G. E. & Hopwoop, A. T. 1928. Catalogue of the Pontian Bovidae of Europe in the Department of Geology. London: British Museum (Natural History). PIVETEAU, J., ed. 1961. Traité de paléontologie 6 (1) Mammiferes, origine reptilienne, évolution. Paris: Masson. PLEssis, S. S. Du 1972. Ecology of blesbok with special reference to productivity. Wildl. Monogr. 30: 1-70. ROBINSON, P. 1972. Pachytragus solignaci, a new species of caprine bovid from the late Miocene Beglia Formation of Tunisia. Notes Serv. géol. Tunisie 37: 73-94. ROBINSON, P. & BLACK, C. C. 1969. Note préliminaire sur les vertébrés fossiles du Vindo- bonien. Notes Serv. géol. Tunisie 31: 67-70. ScHAvuB, S. 1923. Neue und wenig bekannte Cavicornier von Senéze. Eclog. geol. Helv. 18: 281-295. ScHAuB, S. 1943. Die oberpliocaene Sdugetierfauna von Senéze (Haute-Loire) und ihre verbreitungsgeschichtliche Stellung. Eclog. geol. Helv. 36: 270-289. SIMPSON, G. G. 1945. The principles of classification and a classification of mammals. Bull. Am. Mus. nat. Hist. 85: 1-350. SMART, C. L. 1976. The Lothagam 1 fauna: its phylogenetic, ecological and biogeographic significance. In: COpPPENS, Y., HOWELL, F. C., Isaac, G. LI. & LEAKEY, R. E. F., eds. Earliest man and environments in the Lake Rudolf basin: 361-369, Chicago: University of Chicago Press. ‘é STROMER, E. 1907. Fossile Wirbeltier-Reste aus dem Uadi Faregh und Uadi Natrun in Agypten. Abh. sencken. naturforsch. Ges 19: 99-132. STROMER, E. 1928. Wirbeltiere im obermiocanen Flinz Munchens. Abh. bayer. Akad. Wiss. 32: 1-71. FOSSIL BOVIDAE FROM LANGEBAANWEG 337 STUDER, T. 1898. Ueber fossile Knochen vom Wadi-Natrun Unteregypten. Mitt. naturf. Ges. . Bern 1451-1462: 72-77. TalEB, M., JOHANSON, D. C., CopPpeNns, Y. & TIERCELIN, J-J. 1978. Expédition internationale de l’Afar, Ethiopie (4° et 5° campagnes 1975-1977): chronostratigraphie des gisements a Hominidés pliocénes d’Hadar et corrélations avec les sites préhistoriques du Kada Gona. C. r. hebd. Séanc. Acad. Sci., Paris (D) 287: 459-461. TANKARD, A. J. 1974. Varswater Formation of the Langebaanweg-Saldanha area, Cape Province. Trans. geol. Soc. S. Afr. 77: 265-283. TEILHARD DE CHARDIN, P. & TRASSAERT, M. 1938. Cavicornia of south-eastern Shansi. Palaeont. sin. (n.s.C) 6: 1-98. THENIUS, E. 1956. Zur Entwicklung des Knochenzapfens von Protragocerus Depéret (Bovidae, Mammalia) aus dem Miozdn. Geologie 5: 308-318. THENIUS, E. 1959. Handbuch der stratigraphischen Geologie III Tertidr, 2 Wirbeltierfaunen. Stuttgart: Ferdinand Enke. Tuomas, H. 1979a. Miotragocerus cyrenaicus sp. nov. (Bovidae, Artiodactyla, Mammalia) du Miocéne supérieur de Sahabi (Libye) et ses rapports avec les autres Miotragocerus. Geobios Lyon 12: 267-281. Tuomas, H. 1979b. Les bovidés du Miocéne supérieur des couches de Mpesida et de la formation de Lukeino (District de Baringo, Kenya). In: LEAKEY, R. E. F., ed. Proc. pan-afr. Congr. Prehist. 1977 Nairobi. VAN COUVERING, J. A. & MILLER, J. A. 1971. Late Miocene marine and non-marine time scale in Europe. Nature, Lond. 230: 559-563. VILLALTA, J. F. & CRUSAFONT-PaIRO, M. 1955. Un nuevo Ovicaprino en la fauna villafran- quiense de Villaroya (Logrono). Actes 4th Congr. int. Ass. quatern. Res.: 426-432. VrRBA, E. S. 1976. The fossil Bovidae of Sterkfontein, Swartkrans and Kromdraai. Transv. Mus. Mem. 21: 1-166. Vrs, E. S. 1977. New species of Parmularius Hopwood and Damaliscus Sclater & Thomas (Alcelaphini, Bovidae, Mammalia) from Makapansgat, and comments on faunal chrono- logical correlation. Palaeont. afr. 20: 137-151. WALTHER, F. 1972. Hartebeests, roan and sable antelopes and waterbucks. Jn: GRZIMEK, B., ed. Grzimek’s Animal Life Encyclopaedia 13, Mammals IV: 399-430. New York: Van Nostrand. WELLS, L. H. 1969. Generic position of ‘Phenacotragus’ vanhoepeni. S. Afr. J. Sci. 65: 162-163. WELLS, L. H. & Cooke, H. B. S. 1956. Fossil Bovidae from the limeworks quarry, Makapans- gat, Potgietersrus. Palaeont. afr. 4: 1-55. es Sf Mitts oz \) Pn” ow 7s 6. SYSTEMATIC papers must conform to the /nternational code of zoological nomenclature (particularly Articles 22 and 51). Names of new taxa, combinations, synonyms, etc., when used for the first time, must be followed. by the appropriate Latin (not English) abbreviation, e.g. gen. nov., sp. nov., comb. nov., Syn. nov., etc. ‘An author’s name when cited must follow the name of the taxon without intervening punctuation and not be abbreviated; if the year is added, a comma must separate author’s name and year. The author’s name (and date, if cited) must be placed in parentheses if a species or subspecies is transferred from its original genus. The name of a subsequent user of a scientific name must be separated from the scientific name by a colon. Synonymy arrangement should be according to chronology of names, i.e. all published scientific names by which the species previously has been designated are listed in chronological order, with all references to that name following in chronological order, e.g.: Family Nuculanidae Nuculana (Lembulus) bicuspidata (Gould, 1845) Figs 14-15A Nucula (Leda) bicuspidata Gould, 1845: 37. Leda plicifera A. Adams, 1856: 50. Laeda bicuspidata Hanley, 1859: 118, pl. 228 (fig. 73). Sowerby, 1871: pl. 2 (fig. 8a—b). Nucula largillierti Philippi, 1861: 87. Leda bicuspidata: Nicklés, 1950: 163, fig. 301; 1955: 110. Barnard, 1964: 234, figs 8-9. Note punctuation in the above example: comma separates author’s name and year “semicolon separates more than one reference by the same author full stop separates references by different authors figures of plates are enclosed in parentheses to distinguish them from text-figures dash, not comma, separates consecutive numbers Synonymy arrangement according to chronology of bibliographic references, whereby the year is placed in front of each entry, and the synonym repeated in full for each entry, is not acceptable. In describing new Species, one specimen must be designated as the holotype; other speci- mens mentioned in the original description are to be designated paratypes; additional material not regarded as paratypes should be listed separately. The complete data (registration number, depository, description of specimen, locality, collector, date) of the holotype and paratypes must be recorded, e.g.: Holotype SAM-—A13535 in the South African Museum, Cape Town. Adult female from mid-tide region, King’s Beach Port Elizabeth (33°51’S 25°39’E), collected by A. ‘Smith, 15 January 1973. Note standard form of writing South African Museum registration numbers and date. 7. SPECIAL HOUSE RULES ‘Capital initial letters (a) The Figures, Maps and Tables of the paper when referred to in the text - 6 e.g. *... the Figure depicting C. namacolus ...’; *. . . in C. namacolus (Fig. 10)...’ ’ (b) The prefixes of prefixed surnames in all languages, when used in the text, if not preceded by initials or full names e.g. Du Toit but. A.L.du Toit; Von Huene but F. von Huene (c) Scientific names, but not their vernacular derivatives e.g. Therocephalia, but therocephalian Punctuation should be loose, omitting all not strictly necessary Reference to the author should be expressed in the third person Roman’ numerals should be converted to arabic, except when forming part of the title of a - book or article, such as “Revision of the Crustacea. Part VIII. The Amphipoda.’ - Specific name must not stand alone, but be preceded by the generic name or its abbreviation to initial capital letter, provided the same generic name is used consecutively. Name of new genus or species is not to be included in the title: it should be included in the abstract, counter to Recommendation 23 of the Code, to meet the requirements of Biological Abstracts. ie CC) ea rr DN ee te A \U wee —1 A. W. GENTRY FOSSIL BOVIDAE (MAMMALIA) FROM LANGEBAANWEG, SOUTH AFRICA _y oh a al ae Leh De en) Ee, A. SN ey AP |) I : > be > = ie — a i= 2 ra STITUTION NOILNLILSNI NVINOSHLINS S3SIYVYEIT LIBRARIES SMITHSONIAN | io o- z 2] = we “” z | WS = < = ~x = < : A hh = = z =e z 2 Vy = z = z 3 See ga = > = a = ’ oy eee ; Ww) Pad 77) patina 7) iYVYUSIT_ LIBRARI ES SMITHSONIAN INSTITUTION NOILOLILSNI_NVINOSHLIWS S23 buVas é = .o a \\e a oe. a oc a a = : S = c . a = =i is = oo = o : (@) os ore” oe) —_ (@) oop z oF = _! z Saat STITUTION NOILALILSNI NVINOSHLINS. SS'1¥YVYdIT LIBRARIES SMITHSONIAN Zz — ~ = - = as re) | Oo ae ° =e a \ 2 0 5 3 > | > SS a > = > = a \\ ro a - ps b = W's = ae = z ao m= Z e z ao LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLIWS | Z ” eo nw z 7) < = < = pes = z = 2 + es on 8 3 : 3 8 : = z i 2 = 2. = aN. > = a >" = > 77) = 77) sp 2 ” z NOILNLILSNI_ NVINOSHLIWS S3IYVYGIT LIBRARIES SMITHSONIAN on ei 7 z of = isl ” ne 7 us no ae = a =a, a , a eS = a = o = ' ca rs) ee re) ra o ituvud!it LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3I1YVY E 3 = 5 ae w Baie = iE ce . r= A = ,Y i a ED | A d czy > E: Y > i | > - : 2 6 2 - 2 = y NOILNLILSNI S3J!IYVYUGIT LIBRARIES SMITHSONIAN | UT w Ld = awe w es Oe = = ‘= NSS = ae = CAS = 5 => WSS 5 4 NZ . a x. D9 MK ZF mm) 5 ’ a O % Sm oOo ke) SS] +r is SS FVAST 4 © ene or > Nast nm Nanos = nee m ores " wv SWASni f ae Ww — w —_ BRARIES. SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3luyygi7] LIBRARI P = = 7h —g Fe 2 ? = z= \ =| ie =| = 5% Ny = i Oo x : 2 SN 3 e z = \S 2, = Pak oe ab seer Sse a 2 ILSNI NVINOSHLINS S3!IYVYGIT LIBRARIES SMITHSONIAN INSTITUTION *e = < z= Oo ae ake F = Ww) > w — w ~_ G : é : a “ 4 poe = < | = “2 oe 2 = a sad oo ial zZ : Z o — — w . pa ILNLILSNI NVINOSHLINS S3IYVYEIT LIBRARIES SMITHSONIAN NOILALIL : = g ee Z = ‘'Z = = Ws § Fa a! ee 9 a BN & : bY j re = = igi E 2G [= 2 : z ie = 2 BRARIES SMITHSONIAN _INSTITUTION NOILNLILSNI_NVINOSHLINS S3INVUSIT_LIBRARI uw 2 ws = ES 2 ons te wt OO es wn =e ” ae. ¢Y Z ea a = a za qt *°Yy =I < + = = <> a me > he *. - = = = ae = i -— os = 2 m 2" m g BRARIES SMITHSONIAN INSTITUTION NOILNLILSNI_NVINOSHLINS S31uvygIT_LIBRARI . g = Fe wn = n y =! = ae a Ly, = =| = 2 My. O FORA. F216 : LO » > 2 rr a = z * oe _NVINOSHLIWS saiuvygi LIBRARIES SMITHSONIAN INSTITUTION =e aes - = 77 2. aa ix 2 a a : a ul - us = aw S w pa a = . << = <