TU ‘ daly er oy my ie i. Fo RATT RS LARS ENS BRC AILS Ghd aU? es Wa tethiat tl mls A tte stam aoeaa mma Arent CAG MLO WE eee . A Ne ones " ing gttt ae met ee NATE . Sf ViaNt nals, cinta ate We dabei y wren On tien aca Anko Me Ae BY tena reeeys serne 7 ee a ett VON eons enngetyi oy 'y® vee * neg OnWMN EE Abe Me aN y i Lee OA A Ota ea Doon uate ok aS Ld | Dar nary NT b ere ho wy aMe Hens ‘ et hiydd fl pe Oy FEU 42 y ¥ * x x eh Ne Sahes ia reby de AIDA Bene aia aa ate he ashes oath Beat: Nah Baten’ HEIN Ory BO Ino : Paty rely ! i SV alaaib aM bnbelyigth | Vovitsay Nap ett phn Pao Pavel wm ane Mang cha on” ey a " Hat bits i 4 Raw Me var Vente MALS Ree 65 Ny ebay hs S00! w Cetnay Toe Le aD Ahm Ee hehe nay Ayal AMO ey 4 r Sbathetirtbobeet WUT peated RIN AT ah aM Slee vate NAA Eel oer eh talent Ae- EMled gH hyate Aa eS! 4 % VEY ats helt ‘ 2 Yaa st vig egn Rel fV 918 in Bach ANS hm CRP eee Peel PTT Lae We Ue fre pon ite ately Mt Welly © (he Vik Me Stead 0 Hoy Roe eas Wed SG Ne eet Dn et Cee rn n Syd MAE EDEN ae % r ef neha Path Sed Mien sal wNede aia 2 Bat etna Bay wil Ph Geek aN om Mad ely We Cabin Prater dts ee, ay Dh eee EEN AS vie PO ene ae hs D : PU eae Pe : Dat ED eee eet Rohe atime oh a thomon 4 leanne fo pc Ale Mo ede (inegperd sete Hage tee “ EUS EWN ZN A eB Seah We [og Me saad AR ot te Po oe ee ee Mee Miele tip dt PORE He Meh {La Budo eddy lt Seer et way r y Fi As ANT SRE bite 9 “Ne Myedinthe K eltedia tte sede whey Dies Ne Be oh We et atte Me teble he a SON (Sof og) eres viens me ay eye / Tie ei aa Med pe nah Manat den 3 ‘ eotgnedredgungit kid tec ph tye Un Year Gh aot Pr rn eAw a i MeN Se bat ae 4 Vath eyo NS DONT Ay ete Moskoihth, Og heriond Wedded Parner re pnp NEES sas DPN TD ek ee » Prats WD INA sy: Dig ena S Hm Ceofa meet Hemel ety recep tins NEN TOL NNN Aah dP ASAE PRHONar Tor Ree At UNL Tt Rae SU yet OS Peal eval ys Fd ant ehh 8 ofSeenk ree Wank tl altho! EL GSLs, He eka sro tu adtyines Hae eM Geotyr nates ASEM a ha? Sti ll eae a Me lol LL SS Ng FM ak yt oa ae catedl i Ay WACO HP ED gs teen ann dart sa hye oer Ra te MALE Be Yo OM Pat ta He he thoes ely ted pmsl “ shal tt Hiab elhe Malte We bm Be Fults Heato? wpe fla ek swe agen ng apete Ca pePod OSES . Hues WRU ove DAF Mean hl Mg Marj oit gialn paul ei hig Ry Well Ay CoP eg tee Me te F ype hotest AUlab ie tsb ce ledeDiveathiyae Po tSathn ts 145 00 Ve de as § Beer nara) Krvatyes wet Ladies ty omen woot ebtty dtoether te dingd a prot tasin dt getter dare tie sedge dlaatady dReddete Mra WoW eam ee Tic ea E . Ne Vth ats wed Fath oA ho hy tet bg en pad fem Meme tae e thet . Nee on Cdn OA SENAY. bens eee SI Al . a sit Nag ne tea“ teg Sat B (0 1 AF Sa deena PN Yost Yanb mas “i Sect yt cpireiaet ted y WEL hight ‘ Sete Mp te teres Dae A Ue WY Mowe I he PW Nt Sore Lge bad AM ohgai gt erie Lyd be 4 Macdhythdte Ma tte Arete vege Oi ado ON iy oe nhs SSR ae ae Per ier arr Peer emt Pron ene eae : Darras mothe an gas eh Meee dd eet ee a NaslaytinKty am anh \otta dt theta aay ey my MEME a Fe tw FEI dco sone Sora Ney Pia nie Malte hn de Qn Ni inp Oe Pin OAF ee Vaedie ety gh hal re One Pete Da dee | Ape MAB DLE Rr emt gegen In fee F wets ele tot a tot ¥ feo Denied AN be NG : c " ASIEN Medio NAS Hg ie OE Panne get 5 ry ody fF etines ds Heed ane Lava St Hy! : : Pereret Vee why Vinh em Sell ott 8 ay - Pre Fnlacwe Nee Cette foe 80 BE TB Oe eS * 4 WEN PN Ee Gels ye won oy SPAM ANE erA orotate) te AVL SYNE AM! POEM easy Cs Seg SAN Phe sesiy ow ha Rina Mate Se yetedt em alae tee att Ms Cole Get nde Haves effete Or abe ya ts Suid ede boty ° SRSar We WDE EMC IN Son Pay a a - SFe pty sth okt vty a eg tp Me nth Sea setae ot A hye ma Oe foes Oy nile he Cea ABET a LedsdNIT IG Nat é er ne whet Jun t S ra Wear rs TL POC wired 7 5 APP haat ata Mat nit (ted te ete t deaths Me NAgte eav pun tO, 1 PAREN ASO Hae ber eons MeN) Ip Ff Peayhg aah e talc ws Naty “ nisin Eat em vamee LS ten SP Fa Bete PS Pn AN ween 4 x F Thayer ns a mettle me Wat ot cea gh eth te, Muah eer OYA: Ae tna wai ain ceed Holes ; . Lan yiey bol Nope patbt tbat tapattelia/t vthad efter MOST goa manent Shin tre Danity e Weer mA ti Hie 6 Sao Lyd wey RE IR? ; : : FaOgBCANTBL Gs atl) Moa aioe SP ath : ste PN , HT IAP TT OLE TS aa " + yruwatotves ot miaPhen x 1 PENA SRL oI Mita hud Ls Me ‘ shag fle te Phead ding Eon EN pene, pe are Tine ng temlly 9 em tet . Jiri ae ITN A NS 7 ate Bw z sie a8 Ph yAb Ait poem eb * FaNinp Maia Ew Vial aid ph aon? 4 Sete othe ‘ bs Mearntp i a ¥ Heated tam wt batige Mie ork eta rit 4 : Bah te plnnthntiatey : 5 ATG Ria! at pa , igo a LCS en ig mpd ned ont otrenins tritium, eens . ane VL ime ie a ’ Groth eer Wk we ewtl he es nee? ari ol Oe ir Pate £78 Oe Is Aen Ome Re Renter UF Bice teens : ett ’ s . Denn te ee eae eT Pete ee ° : Lda la’ BEN TE ‘ents pe pos. ink wifi D Fyre tN eg eg ego NN a evi wate es en eine 8/9 4 aaah duteg aaneryrentsagidmiie LNG STA Ne cabgeory iat Liat ion ws onan Goan ke . Beene er tl a a A as bs SRDS tye) Pe . VOLUME 99 PART 7 JULY 1990 ISSN 0303-2515 7 we id EE as & OF THE SOUTH AF RICAN 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 3 cm margins all round. First lines of paragraphs should be indented. Tables and a list of captions for illustrations should be typed separately, their positions indicated in the text. All pages should be num- bered 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 18cm (19 cm including caption); the reduction or enlargement required should be indicated (and preferably uniform); orig- inals 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 caption; 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. If Letraset is used authors are requested to use Helvetica-style letter- ing, if possible. 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, 1969b; Jones 1971)’ ‘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, 1969b) 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 (according to the World list of scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses, volume number, part number in parentheses, pagination (first and last pages of article). Examples (note capitalization and punctuation) BuLLouGu, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan. FiscHER, P. H. 1948. Données sur la résistance et de la vitalité des mollusques. Journal de conchyliologie 88 (3): 100-140. FiscHer, P. H., DuvAL, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines. Archives de zoologie expérimentale et générale 74 (33): 627-634. Koun, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee region of Ceylon. Annals and Magazine of Natural History (13) 2 (17): 309-320. Koun, A. J. 1960b. Spawning behaviour, egg masses and larval development in Conus from the Indian Ocean. Bulletin of the Bingham Oceanographic Collection, Yale University 17 (4): 1-51. THIELE, J. 1910. Mollusca. B. Polyplacophora, Gastropoda marina, Bivalvia. In: SCHULTZE, L. Zoologische und anthro- pologische Ergebnisse einer Forschungsreise im westlichen und zentralen Stid-Afrika ausgeftihrt in den Jahren 1903-1905 4 (15). Denkschriften der medizinisch-naturwissenschaftlichen Gesellschaft zu Jena 16: 269-270. (continued inside back cover) ANNALS OF THE SOUTH AFRICAN MUSEUM ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM Volume 99 Band July 1990 Julie Part 7 Deel x SS ro BENTHIC COMMUNITIES AND SEDIMENTARY FACIES IN THE LOWER WITTEBERG GROUP (DEVONIAN, SOUTH AFRICA) By NORTON HILLER 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 ils AES), 53), SUED, “5, B Beso.f), SUES, 8, 79), AG, tat, HOD), 3, OC 7), WOGES), 1D, 3, 7, tsi.) 142), 15(4=5), 24@; 5) 27, 31023), 526), 33, 36(2), 43(1), 45(1), 67(5), 84(2) Copyright enquiries to the South African Museum Kopieregnavrae aan die Suid-Afrikaanse Museum ISBN 0 86813 113 X 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 C633 BENTHIC COMMUNITIES AND SEDIMENTARY FACIES IN THE LOWER WITTEBERG GROUP (DEVONIAN, SOUTH AFRICA) By NORTON HILLER Department of Geology, Rhodes University, Grahamstown, South Africa (With 10 figures) [Paper presented at the Palaeontological Society of southern Africa Symposium, Cape Town, September 1986] ABSTRACT Rare, Middle—Upper Devonian invertebrate fossils are recorded from a few scattered localities in the lower part of the Witteberg Group. The specimens represent a number of shallow-water communities similar to those recognized in Silurian and Devonian rocks from other parts of the world. ~ At a single locality near the top of the Weltevrede Formation near Grahamstown, four co-existing communities are recognized: a linguloid—orbiculoid community, a Tropidoleptus community, a homalonotid—Plectonotus community, and a community of largely infaunal bivalves. Such an assemblage is interpreted as representing the restrictive conditions of an intertidal flat environment. Analysis of the sedimentary facies shows that the fossils come from a sequence of interbedded shales, siltstones and sandstones arranged in a number of thin upward-fining cycles and displaying flaser, lenticular and wavy bedding. This sequence rests on lithic arenites at the top of an upward-coarsening unit and is overlain by thick cross-bedded quartz arenites of the Witpoort Formation. The top part of the Weltevrede Formation is interpreted as having formed in back-barrier tidal flats during transgressive reworking of a delta top. In the Western Cape, two localities in the Wagen Drift Formation have yielded a number of brachiopods, including Tropidoleptus, Australospirifer, chonetaceans and linguloids, as well as bivalves and possible bryozoa. The greater diversity of this brachiopod assemblage suggests somewhat deeper water than the Weltevrede Formation assemblage, probably subtidal. The sedimentary facies of the Wagen Drift Formation are interpreted as having formed in a delta slope environment. The different interpretations of the faunal assemblages from the two areas accord well with the different interpretations of the containing sedimentary facies. CONTENTS PAGE AGRO MU CHO Dipamr ree entre an ena in hee SAE Gee MOS vie alg 216 STEN TEYON ONY’ 5 ea 3 SGAGNCUL tb ccs CRO ORTON OP On ee ee es V7 Sedimentary tacteSsa 4). 4. 5 Chg Sai, AE PE ee He rar fe ER 218 FcicA UID aerate eer te My Me ihr a es St orice a Gosedlndunididl @eolep Wid herd DD PEW EC SCOOT a5. crak dG 6 od lp Ps aH Nne es ea vo eR ae a Jip) INTE 55 Ble i BR Che otchas Gee ot TE ee ee ee eee 228 (CONES OMNG 6:5 sh Setec a oh de icin ce le ae eee aaa a a rae 228 INGHSCEINGSS cacerousoves of exci GPs cg an ea begat ee ed a a i ee 229 215 Ann. S. Afr. Mus. 99 (7), 1990: 215-230, 10 figs. 216 ANNALS OF THE SOUTH AFRICAN MUSEUM INTRODUCTION The Witteberg Group, topmost division of the tripartite Cape Supergroup, is not noted for its invertebrate fossils, although its plant fossils (Plumstead 1967, 1969) and fish fossils (Jubb 1965; Gardiner 1969) are fairly well documented. However, rare marine invertebrates have been recovered from a number of widely spaced localities in the lower part of the group (Fig. 1). Swart (1950) recorded the presence of the inarticulate brachiopod Lingula and further indeterminate brachiopods and bivalves in the lowest part of the Witteberg Group in the Wuppertal area. D. K. Toerien (pers. comm. 1978) recalled finding brachiopods and trilobites in Witteberg shales at Howison’s Poort, near Grahamstown, some forty years ago. Theron (1962) described moulds of the nautiloid Orthoceras and what are possibly bivalve impressions, along with trace fossils, in rocks of the lower Witteberg Group in the Willowmore district. Loock (1967) reported the discovery of a ‘Chonetes’ (brachiopod) in sandstones in the lower divisions of the Witteberg Group near Robertson. Hiller & Dunlevey (1978) recorded brachiopods, bivalves and possible bryozoa from one of several localities in the basal unit of the Witteberg Group in the vicinity of Touws River. Theron (1970, 1972), who had also collected in the Touws River area, used brachiopods and a trilobite as a basis for assigning a Lower to Middle Devonian or possibly an Upper Devonian to Lower Carboniferous age to the basal Witteberg beds. The purpose of the present study was to re-investigate some of the localities, in particular those near Grahamstown in the Eastern Cape and near Touws River in the Western Cape, to see if it was possible to recognize any of the benthic communities that have been described from Silurian and Devonian rocks from other parts of the world, and to see if the environmental information provided by the fossil record was consistent with proposed environmental interpretations N WLo&> SOUTH AFRICA | Sart fr2 aes ee = TR < {eo} @ > a 4,5 km) are probably limited to two small sills: via the central Walvis Ridge to the Angola Basin, and adjacent to the southern tip of Africa to the Agulhas and Mozambique basins. Our samples come from the continental slope and rise off south-western Africa, which has been a sediment-starved margin since mid-Tertiary times (Dingle et al. 1987). The principle physiographic features of this region are slumps. These produce minor sea-floor irregularities north of 34°S, and major relief, with numerous small canyons, south of 34°S. Only in the southern region is the continental shelf narrow and shallow enough to have allowed the advance of the coastline (and river discharge) to within 30 km of the shelf break during glacial sea-level lows. Consequently, the potential for allochthonous shelf contaminants on the slope is relatively low over most of the study area, although at 28°S (just south of Lide- ritz) the head structures of a large slump lie in a re-entrant of the shelf edge. The regional physical oceanography along the outer continental margin of south-western Africa has been summarized by Shannon (1985—see pp. 122-125 mixed layer Antarctic asa a EN Intermediate salinity minimum Water zone poe North Atlantic depth range Deep studied for Water ° ostracods water depth, km Antarctic Bottom Water km 600 Fig. 2. Schematic water column structure off south-western Africa superimposed on a bathy- metric profile across the continental margin south of Liideritz (28°S). Depths for the various water masses are taken from Shannon (1985). Circled symbols indicate direction of water flow: + is towards reader, — is away from reader. Depth range of the ostracod faunas studied is indicated on the right hand side. DEEP-WATER QUATERNARY OSTRACODA 254 for full citation of previous works), who identified three major water masses by their temperature and salinity characteristics. Superimposing these water masses on a bathymetric profile of the south-eastern Atlantic at 28°S (Fig. 2) shows that the outer shelf and upper continental slope (<1,5 km) lie under the northward flowing Antarctic Intermediate Water (AAIW), whereas all the middle to lower slope, and upper continental rise (>1,5 km) lie under the southward flowing North Atlantic Deep Water (NADW). Southward flowing Antarctic Bottom Water (AABW) overlies the lower continental rise and deep ocean floor (>4,0 km), with the consequence that these deeper regions are below the car- bonate compensation depth (CCD). The salinity minimum zone within the AAIW lies between 0,6 and 1,0 km on the upper slope, where its base coincides with the bathyal thermocline. As we will discuss later, these two physico- chemical features coincide with the boundary that separates the bathyal from the neritic ostracod faunas. TAXONOMY Abbreviations: ACA = anterior cardinal angle; AM = anterior margin; ATE = anterior terminal element; DM = dorsal margin; LV = left valve; ME = median element; MS = muscle scars; PCA = posterior cardinal angle; PM = posterior margin; PTE = posterior terminal element; RV = right valve; TE = terminal elements; VM = ventral margin. Figured specimens are stored at the South African Museum under the cata- logue numbers SAM—POQ-MF-. SEM numbers refer to unique scanning electron microscope numbers in the collection of RVD. Subclass OSTRACODA Latreille, 1806 Order PODOCOPIDA Miller, 1894 Suborder PLATYCOPINA Sars, 1866 Family Cytherellidae Sars, 1866 Genus Cytherella Jones, 1849 The genus Cytherella is widely distributed along the continental margin of south-western Africa (Fig. 3). Four species have been recognized and these occupy well-defined depth ranges: Cytherella dromedaria Brady, 1880, and Cytherella sp. nov. in the Neritic Zone (40-300 m and 115-736 m, respectively), and C. serratula (Brady, 1880) and Cytherella sp. 3027 in the Bathyal and Abyssal zones (Fig. 4) (1 000-2 070m, and 2 916m, respectively). Unlike Krithe and Buntonia, two important genera that range from shallow- to deep- water environments, neritic species of Cytherella do not extend their depth ranges into the Bathyal Zone. In this respect Cytherella is similar to Cytheropteron. Dp ANNALS OF THE SOUTH AFRICAN MUSEUM TABLE 2 List of genera and species discussed in this paper. Local depth Species range Page (m) Cytherella Jones, 1849 Dil C. serratula (Brady, 1880) 1 000—2 070 Cytherella sp. 3027 2 916 Cytheropteron Sars, 1866 258 C. cronini sp. nov. 990—2 070 Cytheropteron sp. 2909 945 Cytheropteron sp. 2914 2 070 Krithe Brady et al., 1874 263 K. capensis sp. nov. 238-1 430 K. spatulanis sp. nov. 392-1 662 K. rex sp. nov. 2 916 K. peypouqueti sp. nov. 2 916-4 736 Krithe sp. 8 BWR 353) Krithe sp. 9 430-900 Krithe sp. 4 1 600-2 916 Krithe sp. 6 1 662-2 916 Krithe sp. 7 1 600-2 916 Krithe sp. 19 1 662-2 916 Krithe sp. 22 2 926 Parakrithe van den Bold, 1958a 286 Parakrithe sp. 10 945-1 353 Buntonia Howe, 1935 286 B. rosenfeldi sp. nov. 186—2 070 Dutoitella Dingle, 1981 293 D. suhmi (Brady, 1880) 2 916 Abyssocythere Benson, 1971 298 A. australis Benson, 1971 2 916 Ambocythere van den Bold, 19585 301 Ambocythere sp. 3057 2 070 Echinocythereis Puri, 1954 302 E. whatleyi sp. nov. 730-2 916 Trachyleberis Brady, 1898 309 Trachylebens sp. 3017 2 916-4 736 Henryhowella Puri, 1957 310 H. melobesioides (Brady, 1869) 100-2 916 Rugocythereis gen. nov. 318 R. horridus (Whatley & Coles, 1987) 730-7, 916 Poseidonamicus Benson, 1972 B22 P. major Benson, 1972 2 070-2 916 Indeterminate taxa 328 Indeterminate sp. 62 1 026 Indeterminate sp. 23 1 060 DEEP-WATER QUATERNARY OSTRACODA 253 a) C. sp. 3027” ABYSSAL BATHYAL water depth, km C. serratula C. dromedaria + C.sp. nov. latitude °S Fig. 3. Latitude and water depth of samples bearing Cytherella species. There are two species in each of the Neritic and Bathyal zones: C. dromedaria Brady and Cytherella sp. nov., and C. serratula Brady and Cytherella sp. 3027, respectively. NERITIC | BATHYAL | ABYSSAL UPPER LOWER 100 | | | | | | C. dromedaria 50 % Cytherella | | | | | | | + |S: sp. nov | | | | | | | | | | | | | C. serratula | | | | | C. sp. S021 water depth, km Fig. 4. Cytherella species as percentage of total ostracod fauna plotted against water depth. Note concentrations of samples on the continental shelf (<300 m), and at the Upper/Lower Bathyal Zone boundary. 254 ANNALS OF THE SOUTH AFRICAN MUSEUM Cytherella serratula (Brady, 1880) Fig. SA-C Cythere (?) serratula Brady, 1880: 77, pl. 43 (figs 7a—d). Puri & Hulings, 1976: 288-289, pl. 24 (figs 15-16). Cytherella serrulata Brady & Norman, 1896: 713-716, pl. 66 (figs 3-6). ?Cytherella sp. 11 Ducasse & Peypouquet, 1979, pl. 1 (figs 3-4). ?Cytherella sp. Guernet, 1985: 281, pl. 1 (figs 2, 4). Cytherella sp. B Cronin, 1983, pl. 6 (fig. E). Cytherella serratula (Brady) Whatley & Coles, 1987: 81, pl. 6 (figs 30-31). Illustrated specimens MF-0422, LV, TBD 3344, 1 430 m. MF-—0423, RV, TBD 3344, 1 430 m. MF-0424, LV, TBD 3344, 1 430 m. Remarks Our material shows slight shape and ornamentation differences to the type specimens described by Brady (1880), and the lectotypes illustrated by Puri & Hulings (1976). These, however, are within the range of intraspecific variation found on an inter-ocean basis (R. C. Whatley citing his own data base, pers. comm. 1988). Positive identifications of C. serratula (Brady) have been made from the following sites and depths: Brady (1880) (Recent)— ‘Challenger’ site 24, Caribbean, 390 fm (713 m); ‘Chal- lenger’ site 85, Canary Islands, 1 125 fm (2 057m); ‘Challenger’ site 335, Tristan da Cunha, 1 425 fm (2 605 m). Brady & Norman (1896) (Recent)— West Africa, 466-1 168 fm (852-2 135 m); Mauretania, 418-675 fm (764-1 234m); Canary Islands, 487 fm (890 m); Morocco, 600 fm (1 097 m). Cronin (1983) (Recent)—south-eastern USA continental slope, 462—1 070 m. Whatley & Coles (1987) (Miocene to Holocene)—DSDP Leg 94 sites, North Atlantic, 2 417-3 022 m. Present study (Recent)—south-western Africa, 1 000-2 070 m. These records give a modern depth range of 462-3 022 m in the Atlantic. Cytherella sp. Guernet (1985) from bathyal Eocene sediments at DSDP site 219 on the 90 East Ridge in the Indian Ocean is very similar to C. serratula (Brady) in shape and ornamentation, and may be the same species, but differs in having a small, pronounced dorso-median depression. Cytherella vulgata Rug- gierl, 1962 (as illustrated by Benson 1978, pl. 2 (fig. 3)), from the Upper Pliocene at DSDP site 371 also has a similar shape to C. serratula, but has its highest point in the posterior part of the valve and a more extensive covering of small spines and granules in the posterior area. Probably the closest species pre- viously recorded, but not placed in C. serratula (Brady), is Cytherella sp. 11 of Ducasse & Peypouquet (1979) from bathyal (c. 2 300 m water depth) Pliocene sediments at DSDP site 403 in the Rockall Basin. These specimens differ slightly DEEP-WATER QUATERNARY OSTRACODA DS) BRK NN < —— AS Fig. 5. A-C. Cytherella serratula (Brady, 1880), TBD 3344, 1 430 m. A. SAM-PQ-MF-0422, LV, SEM 2584. B. SAM—-PQ-MF-0423, RV, SEM 2589. C. SAM—POQ-MF-0424, LV internal view, SEM 2587. D-F. Cytherella sp. 3027, SAM-PQ-MF-0425, RV, TBD 6581, 2 916 m. D. External view, SEM 3027. E. Detail, postero-ventral area, SEM 3028. F. Detail, antero- dorsal area, SEM 3029. Scale bar = 100 microns. 256 ANNALS OF THE SOUTH AFRICAN MUSEUM from our material in having coarser and more extensive spines in the posterior part of the valve. Rosenfeld & Bein (1978) recorded smooth species of the genus from the continental slope off north-west Africa, but illustrated only a coarsely pitted species, whereas Peypouquet & Benson (1980) noted the genus at 439-2 154 m in the Cape Basin and 527-2 754 m in the Angola Basin. In neither case did they illustrate or discuss the species present. Figures 3 and 4 show the geographical and water-depth distributions of Cytherella serratula (Brady) in relation to other species of the genus Cytherella along the continental margin of south-western Africa. It has been found in sedi- ment samples between approximately 19°S and 33°S (Fig. 3). There is a very well-defined partitioning of the Cytherella species with water depth (Fig. 4). In the Neritic Zone, Cytherella sp. nov. extends to 736 m and, together with the more restricted C. dromedaria Brady, 1880, commonly constitutes greater than 30 per cent of the total ostracod population. There seems to be an hiatus in occurrence of the genus over the depth range 750-1 000 m, coinciding with the Salinity Minimum Zone of the Antarctic Intermediate Water Mass. Cytherella serratula appears at the top of the Bathyal Zone, to which it is restricted, and for which its appearance and disappearance constitute one of the defining par- ameters. In the Bathyal Zone, the genus Cytherella (as represented by C. serratula) is numerically less important than it is higher up the continental slope, and only reaches 29 per cent in the region of the Upper/Lower Bathyal Zone boundary (1 430-1610 m). Overall, it forms 3 per cent of the total bathyal ostracod assemblage (fourth most abundant species), but is somewhat more important in the Lower Bathyal Zone, where it averages 9 per cent. Clearly, Cytherella is a sensitive indicator of the different water masses, but is less tolerant of the conditions in the colder, more saline Bathyal and Abyssal zones than it is of those in the Neritic Zone. These results are in broad agree- ment with the data presented by Peypouquet & Benson (1980) from a transect across the continental margin west of Walvis Bay. Plotting their values we obtain Figure 6, which shows a decline in percentage of Cytherella in the ostracod populations from greater than 50 per cent in shallow water (<500 m) to less than 10 per cent in the water-depth range 974-1 546 m. In deeper water (2 094 m and 2 117 m) values rise again to approximately 30 per cent. Cytherella sp. 3027 Fig. 5D-F ?Cytherella sp. gr. ovata (Roemer, 1840) Guernet, 1985: 281, pl. 1 (fig. 1). Illustrated specimen MF-0425, RV, TBD 6851, 2 916 m. DEEP-WATER QUATERNARY OSTRACODA DST oO is) NERITIC BATHYAL % Cytherella spp. water depth, km Fig. 6. Cytherella species as percentage of total ostracod fauna plotted against water depth for a profile off Walvis Bay. Data computed from Peypouquet & Benson (1980). Remarks A single RV, possibly a juvenile. It has a distinctive lateral outline, with DM sloping towards the PM, a short straight VM, and broadly rounded AM. The lateral surface has an overall delicate, bead-like reticulation, which is more pronounced in the anterior and posterior areas. There are small spines along the postero-ventral margin, and the AM has a very narrow compressed border with well-separated small pores with setae. There is a triangular depression on the dorso-median surface that extends as a weak sulcus to the valve centre over the MS area. This is probably a new species, which is confined to the upper part of the Abyssal Zone (2 916 m, TBD 6581). Guernet (1985) illustrated a very similar specimen from the bathyal Eocene sediments at DSDP site 219 on the 90 East Ridge in the Indian Ocean. Cytherella sp. 3027 is the only representative of the genus Cytherella that we encountered in the Abyssal Zone (i.e. >2 070 m, Fig. 4), where it constitutes <0,3 per cent of the total ostracod population. In contrast, Peypouquet & Benson’s (1980) samples from similar depths in the northern Cape Basin off Walvis Bay contained approximately 30 per cent Cytherella, although they did not illustrate or describe the species involved. This may indicate a shallowing of the CCD in the south-eastern Cape Basin, because below 2 154 m, where they reported very strong dissolution, Peypouquet & Benson (1980, fig. 2) no longer recorded the genus. 258 ANNALS OF THE SOUTH AFRICAN MUSEUM Suborder PODOCOPINA Sars, 1866 Superfamily Cytheracea Baird, 1850 Family Cytheruridae Muller, 1894 Genus Cytheropteron Sars, 1866 On the continental margin off south-western Africa, this genus is represented by 12 species, which occur over a depth range of 40-2 070 m and over a latitudi- nal range 19°S to 35°S (Figs 7, 8). Nine of these species are confined to the Neritic Zone (where they occur in two distinct groups—uinner shelf: 40—90 m, and inner shelf/slope: 80-738 m), and three to the Bathyal Zone. We did not find the genus in the Abyssal Zone (i.e. deeper than 2 070 m). In the Neritic Zone, Cytheropteron is locally relatively abundant (up to 45% total ostracod population in raw data; c. 12% on smoothed curves, Fig. 8), whereas in the Bathyal Zone it does not exceed 6 per cent in any one sediment sample. However, the means of its occurrence in both regions are remarkably constant: 5 per cent in the Neritic Zone, and 4 per cent in the Bathyal Zone. Overall, we found Cytheropteron to be numerically more important in the Bathyal Zone: 3,4 per cent of the total ostracod population from samples in which it occurred in the Bathyal Zone, compared to 1,2 per cent from the Neritic Zone. There is, however, a significant difference in the distribution of the various species within the two depth zones. In the Neritic Zone four species BATHYAL-— water depth, km latitude °S @ — C. cronini Fig. 7. Latitude and water depth of samples bearing Cytheropteron species. There are nine species in the Neritic Zone, and three in the Bathyal Zone (see Fig. 8 for depth ranges). DEEP-WATER QUATERNARY OSTRACODA 259 | | nae C.sp. 2914 # M@C.sp. 2909 mC.sp.9 -/———_—__———— C. sp.8 1 C. sp. 7 orn C, SD, 6 bret , SP). © mC. sp. 4 ™C.sp.3 mC.sp.2 FIC. sp.1 | | | | | | | | | | | | | | | | NERITIC BATHYAL fn UPPER Bi as % Cytheropteron spp. (0) 1 2 water depth, km Fig. 8. Cytheropteron species as percentage of total ostracod fauna plotted against water depth. Values are five point running means. Depth ranges of species are shown by bars. N.B. Neritic species, Cytheropteron species 1—9, are not further discussed herein. occur with extensively overlapping depth ranges in the outer shelf/upper slope region, whereas in the Bathyal Zone only one species has an extensive depth range (C. cronini sp. nov., 990-2 070 m), with the other two species occurring only at the upper and lower limits of the Bathyal Zone (Fig. 8). The upslope limit of the Bathyal Zone is marked by a decrease in the per- centage of the genus relative to the overall ostracod population (mean 12% to 3%) and, with the exception of sample TBD 3355 (2 070 m), all our records for the genus are from the Upper Bathyal Zone (Fig. 7). These data indicate that the genus is sensitive to the physico-chemical changes in the vicinity of the base of the AAIW low salinity layer, and between the AAIW and the NADW masses. 260 ANNALS OF THE SOUTH AFRICAN MUSEUM The depth ranges of the three bathyal Cytheropteron species from off south- western Africa can be summarized as: Cytheropteron sp. 2909—945 m (uppermost Bathyal Zone) Cytheropteron cronini sp. nov.—990-2 070 m (Bathyal Zone) Cytheropteron sp. 2914—2 070 m (lowermost Bathyal Zone) Well-documented Quaternary records of the genus from deep-water sites elsewhere are sparse, and confusing. As Whatley & Masson (1979) have observed, there have been few detailed revisions of the nineteenth century works on this genus, with the result that records of earlier-named species are frequently in error. In their survey in the south-eastern Atlantic, Peypouquet & Benson (1980) recorded the following depth ranges for undefined species of the genus: Angola Basin—527 m; Cape Basin—439-974 m, so that effectively they did not record it from the Bathyal Zone as we have defined it. Off north-west Africa, Rosen- feld & Bein (1978) recorded Cytheropteron sp. from 470m, which was the uppermost part of their ‘deep water’ Association B fauna. Cronin (1983) listed ten species of Cytheropteron from various depths off south-eastern United States, seven of which extend into water depths greater than 1 000 m (his Cythe- ropteron spp. S, D & V, E, B, P, R, and Q). Of these, Cytheropteron sp. P is probably conspecific with our new species C. cronini from the Cape Basin. Tressler (1941) illustrated three deep-water species of the genus from the North Atlantic: Cytheropteron alatum Sars, 1866 (1 280-4 700 m); C. hamatum Sars, 1869 (1 280 m); and C. inflatum Brady et al., 1874 (1 955-3 230 m), but comparison of his illustrations with those of Whatley & Masson (1979) indicate that none are conspecific with the original species. Similarly, Benson et al.’s (1983) records of C. alatum Sars, 1866 (2 560-2 743 m), and C. testudo Sars, 1869 (1 380-2 758m), off Newfoundland are probably mis-identifications. Whatley & Masson (1979) noted that dead specimens of C. alatum Sars, 1866, have been found in 830 m in the Rockall Trough (north-eastern Atlantic). Deep-water species from atypical deep-water settings have been made by Breman (1975a, 19756), who described Cytheropteron sp. and C. ‘adriaticur’ from samples between 144-1 216 m in the Adriatic, and Bonaduce et al. (1983), who recorded three species originally described by Bonaduce et al. (1976) from ‘deep water’ sites in the Gulf of Aqaba and the Red Sea: C. alabarda, C. excisum, and C. pulcinella. No water-depth ranges were allocated to individual species. Cytheropteron cronini sp. nov. Fig. 9A-C ?Cytheropteron sp. P Cronin, 1983, pl. 8 (fig. B). Derivation of name This species is named for Dr T. M. Cronin (US Geological Survey), who first recorded the species or a very close relative of it. DEEP-WATER QUATERNARY OSTRACODA 261 Fig. 9. A—C. Cytheropteron cronini sp. nov, SAM—PQ-MF-0426, holotype, LV, TBD 2880, 1 026 m. A. External view, SEM 2911. B. Detail, ala, SEM 2912. C. Internal view, SEM 3093. D. Cytheropteron sp. 2909, SAM-PQ-MF-0427, LV, TBD 3341, 945m, SEM 2909. E-F. Cytheropteron sp. 2914, SAM-PQ-MF-0428, RV, TBD 3355, 2070 m. E. External view, SEM 2914. F. Detail, ala, SEM 2915. Scale bar = 100 microns. 262 ANNALS OF THE SOUTH AFRICAN MUSEUM Holotype MF-0426, LV, TBD 2880, 1 026 m. Diagnosis A dolphin-shaped species with delta-like ala that have a large dimple on their dorsal leading edge. In the dorso-median area there is vertical, slit-shaped reticulation. Description In lateral outline the AM is asymmetrically rounded, and is inclined ven- trally. PM is asymmetric and bluntly caudate, with the apex dorsally directed. DM is strongly arched, VM is asymmetrically convex, sloping ventrally towards the posterior. The central area of the valve is inflated, with a delta-shaped ala that carries a small spine at its apex. There is a large dimple in the anterior proximal area of the ala. Valve surface dorsal to the ala is ornamented with low vertical ribs and slit shaped reticulation. In lateral view there is a small, promi- nent nick at the posterior end of the DM. Dimensions (mm) length height MF-0426 0,42 0,28 Remarks This distinctive species (or a close relative) was recorded as Cytheropte- ron sp. P from 347-1 034 m off Florida by Cronin (1983). Species with similar features are: C. abyssorum Brady, 1880, from 4 753 m (2 600 fm) off Tasmania, which has a less triangular ala and stronger surface reticulation; C. trifossata Whatley & Coles, 1987 (>3 000 m late Quaternary, North Atlantic), which has more prominent ornamentation and a less posteriorly directed ala; and C. para- latissimum Swain, 1963 (Pleistocene, Arctic Alaska), whose ala are less posteri- orly directed and which have ‘subvertical furrows’ on their lateral surface. Cythe- ropteron porterae Whatley & Coles, 1987 (= C. alatum Sars of Benson et al. 1983), differs in possessing a thickened leading edge to the ala, and is more coarsely ornamented than C. cronini sp. nov. Whatley & Coles (1987) recorded C. porterae from the early Pliocene to Quaternary of the North Atlantic. Cytheropteron cronini is the only species of the genus that has wide geo- graphical (21°-33°S) and depth (990-2 070 m) ranges in the Bathyal Zone of the south-eastern Cape Basin (Figs 7, 8). Cytheropteron sp. 2909 Fig. 9D Illustrated specimen MF-0427, LV, TBD 3341, 945 m. DEEP-WATER QUATERNARY OSTRACODA 263 Remarks One poorly preserved specimen with a distinct bevel along the outer edge of its ala, in the centre of which there is a small depression. The lateral surface may have originally been coarsely punctate. Despite its battered appearance, we con- clude that this specimen is not allochthonous because no provenance population has been located farther upslope. It occurs at the top of the Upper Bathyal OIC: In general shape this species is similar to Cytheropteron sp. Q, recorded by Cronin (1983) from 584 m (?382 m) to 1070 m off the south-eastern United States. Cytheropteron sp. 2914 Fig. 9E—F Illustrated specimen MF-0428, RV, TBD 3355, 2 070 m. Remarks A fragile species with a small, sharp ala that bears two small spines, and has a corded leading edge. The sculptured postero-dorsal margin is similar to that of Cytheropteron sp. 8 of Ducasse & Peypouquet (1979, pl. 4 (fig. 7)) from the Quaternary of DSDP site 400A (north-eastern Atlantic). The latter has an area of coarse reticulation dorsal to the ala, which in Cytheropteron sp. 2914 is covered by a feint, longitudinally sinuous ridge. This species is a close relative of C. syntomoalatum Whatley & Coles, 1987 (see pl. 2 (fig. 27)), from the late Pliocene to Quaternary of the North Atlantic, particularly in possessing a mid- dorsal ‘cross’. Differences in ornamentation and structure of the ala apex may indicate that the two are not conspecific. Cytheropteron sp. 2914 was encountered only in the lowermost part of the Lower Bathyal Zone. Family Cytherideidae Sars, 1925 Genus Krithe Brady, Crosskey & Robertson, 1874 The genus Krithe and the closely related genus Parakrithe have convention- ally been regarded as indicators of ‘deep’ water (e.g. Van Morkhoven 1962), and use has recently been made of variations in shell architecture to semi- quantitively predict palaeo-water depths and dissolved oxygen levels (e.g. Peypouquet 1975, 1979; Donze et al. 1982). The latter studies have been based on numerous morphotypes that, whilst they have not attempted to systematically isolate species, have shown that this group is taxonomically complex and diverse (e.g. Peypouquet 1979). Whatley (1983) recorded that, with the exception of 264 ANNALS OF THE SOUTH AFRICAN MUSEUM Cytheropteron (57 species), Krithe (54 species) is the most diverse genus in the Quaternary ostracod faunas of the bathyal and abyssal regions of the south- western Pacific. Similarly, Krithe (together with one species of Parakrithe) and Cytheropte- ron are the two most diverse genera in the deep-water faunas off south-western Africa (12 species each), whereas in the Abyssal Zone the genus Krithe is the most abundant taxon. It occurs across the entire latitudinal range of our study area (17°S to 38°S), and is found over a more extensive water-depth range than any Other genus (238—4 736 m, Fig. 10). With one exception (Krithe sp. 9), all the species are found in either bathyal or abyssal depths, and five also occur in the Neritic Zone (Fig. 13). On a regional scale, variations in the abundance of Krithe within the overall ostracod populations have been used (together with Henryhowella melobesioides (Brady, 1869) and Buntonia rosenfeldi sp. nov.) to help identify the limits of the faunal zones off south-western Africa (see Figs 54, 55). Figure 11 shows that the abundance of Krithe species (as a percentage of the total ostracod fauna) varies greatly over the water-depth range of the genus, and that there is not a simple progressive increase in values oceanward. This can be further emphasized when mean values for individual depth zones are exam- ined: Figure 12 shows that, although the overall trend is Neritic (19%) through ——® = K. capensis —-—D = Parakrithe sp.10 water depth, km latitude °S Fig. 10. Latitude and water depth of samples bearing Krithe species and Parakrithe sp. 10. The distributions of Krithe capensis sp. nov. and Parakrithe sp. 10 are outlined. DEEP-WATER QUATERNARY OSTRACODA 265 | | _ @K | | mJ | ees ee ey depth Moe cee ees ranges SSS SS SS | | | | | | -}-———1E ee ae 7 RAE gee Da Ee I = dominant | fi ea AE uh 4sp.4 peypouqueti i Spatularis species ea | ->——————_icapensis | | NERITIC BATHYAL ABYSSAL UPPER | LOWER m oi fo) if] % Krithe spp. 0) 1 2 3 4 5 water depth, km sp. 7 . sp. 6 . sp. 19 rex . sp. 22 . peypouqueti . Capensis . spatularis sp. 9 . sp. 8 arakrithe sp. 10 . sp. 4 mmodowowp RUAKRAZA Ros & — 2c ®) RAKRARKRA Fig. 11. Krithe species (including Parakrithe sp. 10) as percentage of total ostracod fauna plotted against water depth. Values are three point running means. In the upper part of the diagram ‘depth ranges’ show bars for total range of individual species of Krithe and one species of Parakrithe identified A-L in the list in lower part of figure. Krithe sp. 9, which is a neritic taxon, is included for completeness. ‘Dominant species’ shows depth range bars for particular species dominating the Krithe assemblage, and are derived from variations in abundance shown in Figure 13. 266 ANNALS OF THE SOUTH AFRICAN MUSEUM Bathyal (24%) to Abyssal (42%), there is a decrease from Neritic into Upper Bathyal (13%), and a decrease from Lower Bathyal (51%) into Abyssal. These fluctuations can be related to the environmental tolerances of particular species, and we recognize two groups: an Upper Krithe Fauna that occurs in the Neritic and Upper Bathyal zones (238-1 500 m); and a Lower Krithe Fauna that occurs in the Lower Bathyal and Abyssal zones (1 500-4 736 m) (Figs 11, 13A—B, 14). Throughout most of the Neritic Zone, the Upper Krithe Fauna is domi- nated by Krithe capensis sp. nov., but this species declines sharply in abundance across the Neritic/Bathyal boundary, below which it is effectively replaced by Krithe sp. 8. Krithe capensis also suffers a temporary decline in abundance between approximately 500 m and 700 m water depth, where there is also a sharp decline in the importance of the genus as a whole across the upper bound- ary of the salinity minimum zone of the AAIW. Within this narrow depth range, as well as over the lower part of the Upper Bathyal Zone (where the genus as a whole is again relatively poorly represented), Krithe spatularis sp. nov. replaces K. capensis as the dominant species. The genus Krithe reaches a low level of abundance within the ostracod populations at the base of the Upper Bathyal Zone (i.e. at the base of the AAIW mass) where, over the depth range 1 430 m to 1 525 m, all four extant LOWER BATHYAL 50 ABYSSAL total ostracods NERITIC mean % 6) % range = 1—64 Samos 20-100 14-66 n = 30 10 5 3 Fig. 12. The mean percentage of Krithe species for the four depth zones. Values are calculated on samples in which Krithe occurs, and not on the overall ostracod fauna. Mean value for the Bathyal Zone is 24%; n = number of samples. % Krithe spp. DEEP-WATER QUATERNARY OSTRACODA 267 | | NERITIC | BATHYAL | ABYSSAL | | UPPER | LOWER 100 UPPER KRITHE FAUNA = K. capensis © = K.sp.8 K. spatularis 50 100 | LOWER KRITHE FAUNA | . peypouqueti | | | | 50 0) water depth, km Fig. 13. Variation in abundance of individual Krithe species with water depth (plotted as a percentage of the Krithe assemblage). A. Upper Krithe Fauna (three-point means). B. Lower Krithe Fauna (three-point means, and raw data in deepest sample). Ranges over which particu- lar species dominate are shown in Figure 11. 268 ANNALS OF THE SOUTH AFRICAN MUSEUM LOWER BATHYAL water depth, km NERITIC latitude °S — O = K. capensis --— += K. spatularis Fig. 14. Latitude and water depth of samples bearing Krithe capensis, K. spatularis, and Krithe sp. 8. The open arrow locates a possible allochthonous occurrence of K. capensis. species of the Upper Krithe Fauna die out (K. capensis, K. spatularis, Krithe sp. 8, and Krithe sp. 9). There is a dramatic increase in the abundance of the genus in the Lower Bathyal Zone with the incoming of four species of the Lower Krithe Fauna (Krithe sp. 4, Krithe sp. 6, Krithe sp. 19, and Krithe sp. 7). Of these, Krithe sp. 4 dominates the Lower Bathyal and upper parts of the Abyssal zones, before it and the three other species are themselves replaced by the true abyssal species at 2 916 m. The most important abyssal taxon is the relatively small, globular Krithe peypouqueti sp. nov. A summary of the species of Krithe that dominate at the various depths off south-western Africa is given in Figure 11. Peypouquet & Benson (1980) recorded the distribution of Krithe and Para- krithe in their transect west of Walvis Bay, but the relative sparsity of data points precludes a detailed comparison with our data (Fig. 15). The overall picture is of a significant increase in abundance across the Neritic/Bathyal boundary, and a decrease in Abyssal depths. The highest values occur within the Bathyal Zone, and in this respect their data agrees with our own. DEEP-WATER QUATERNARY OSTRACODA 269 35 NERITIC BATHYAL ABYSSAL UPPER NO [o) % Krithe spp. 1 2 3 water depth, km. Fig. 15. Krithe species as percentage of total ostracod fauna plotted against water depth for a profile off Walvis Bay. Data computed from Peypouquet & Benson (1980). Krithe capensis sp. nov. Figs 16A-—C, 17A, 18D Krithe spp. Boomer, 1985: 57-58, pl. 4 (fig. 63). Derivation of name From the Cape Basin. Holotype MF-0429, LV, TBD 2879, 530 m. Paratypes MF-0431, LV, TBD 2879, 530 m. MF-0430, RV, TBD 3577, 453 m. MF-0450, LV, TBD 3577, 453 m. Diagnosis In lateral view, the highest point of the valve lies in the posterior third, over a broadly rounded postero-dorsal arch. VM is straight, AM is broadly rounded. 270 ANNALS OF THE SOUTH AFRICAN MUSEUM Description In external view, the valves have a high, broadly rounded arch over the pos- terior dorsal region. AM in the LV is broadly rounded and almost symmetrical, whereas the AM in the RV is slightly upturned and there is a subtle antero- dorsal step. The PM depression is prominent and elliptical, but partly hidden in internal lateral view. The anterior vestibule is moderately large and has two lobes on the dorsal side. The anterior inner margin descends from the dorsal margin in almost a straight line, and forms an acute angle antero-ventrally. MS have a partly subdivided dorsal-most adductor, and a lobed anterior scar. Dimensions (mm) length height MF-0429 0,91 0,50 MF-0431 0,90 0,51 MF-0430 0,95 0,50 MF-0450 0,99 0,50 The mean length/height ratio of the type specimens is 1,86, which distin- guishes Krithe capensis in the local deep-water faunas from all but Krithe sp. 4 (Figs 19, 20). Remarks We have difficulty in assigning K. capensis sp. nov. to the ‘ecotypes’ described by Peypouquet (1979), but on balance it has most in common with his category C. Krithe capensis is quite close to K. nibelaensis Dingle from the Campanian to Eocene of southern Africa (Dingle 1981; Frewin 1987), but their MS patterns differ in the shape of the anterior scar, and the former species has a more arched postero-dorsal valve outline. Rosenfeld & Bein (1978, pl. 1 (fig. 20)) illustrated a species with a similar inner margin outline to K. capensis, which they allocated to K. producta? Brady, 1880. Our species differs from theirs in its more arched postero-dorsal outline and less prominent and incised posterior depression. The lectotype of K. producta Brady, 1880 (Puri & Hulings 1976), has a more rounded PM and broadly arched DM outline than K. capensis. Benson & Peypouquet (1983, pl. 5 (fig. 5)) illustrated a specimen (Krithe sp. C23) from the Lower Miocene of DSDP site 516 that has a similar shaped vestibule to K. capensis, but this species has parallel VM and DM. Off south-western Africa, we have found Krithe capensis within the latitudi- nal range 17°S to 35°S (Fig. 10). Its depth range is 238-1 430 m (a total of 1 192 m), which includes the outer shelf (Neritic Zone) through to the upper- most part of the Lower Bathyal Zone. This suggests that K. capensis is the most euryhaline and eurythermal Krithe species in our study area (34,9-34,6%o to <34,4%o, 10°C to 3,2-3,6°C). Variations in its abundance in relation to other species of the genus (Fig. 13) indicate that K. capensis is the dominant species DEEP-WATER QUATERNARY OSTRACODA OFM Fig. 16. A-C. Krithe capensis sp. nov. A. SAM-PQ-MF-0429, holotype, LV, TBD 2879, 530m, SEM 2709. B. SAM-—PQ—-MF-0430, RV, TBD 3577, 453m, SEM 2714. C. SAM-PQ-MF-0431, LV internal view, TBD 2879, 530m, SEM 2710. D-F. Krithe spatularis sp. nov. D. SAM—PQ-MF-0432, holotype, LV internal view, TBD 2978, 736 m, SEM 2707. E. SAM-—PQ-MF-0433, RV, TBD 2978, 736m, SEM 2705. F. SAM- PQ-MF-0434, LV, TBD 3177, 1 000 m, SEM 2702. Scale bars = 100 microns. 272 ANNALS OF THE SOUTH AFRICAN MUSEUM on the continental shelf and uppermost slope, but over the depth range c. 500-700 m it is replaced by K. spatularis sp. nov. This coincides with the part of the AAIW mass that lies immediately above the Salinity Minimum Zone (see Fig. 2). Within the Salinity Minimum Zone, K. capensis reverts to its dominant position but, from about 950 m to the depth at which it dies out (1 430 m), it declines rapidly in abundance. At the Neritic/Bathyal boundary (950 m) it is replaced by Krithe sp. 8., and at deeper levels by K. spatularis sp. nov. Its lower limit is apparently defined by the AAIW/NADW shear zone boundary (Upper/ Lower Bathyal boundary). All our evidence suggests that in the south-eastern Cape Basin, K. capensis is the most tolerant of low-salinity conditions of all the species of the genus. A more critical assessment of the geographical distribution of Krithe capen- sis (Fig. 14) suggests that the deepest site at which the species has been located may in fact represent an allochthonous occurrence. We suspect this because the three sites astride the Neritic/Bathyal boundary lie along a wide front from which station TBD 3344 (1 430 m) is isolated. If this is the case, then the depth range of this species is 238-1 071 m (a total of 833 m), and K. capensis would more correctly be considered a neritic species that only straggles into the deep- water assemblages in the Upper Bathyal Zone, just below the base of the Salin- ity Minimum Zone of the AAITW. At present we have insufficient data to confirm this suspicion, and take the data at face value. Krithe spatularis sp. nov. Figs 16D-F, 17B, 18E Derivation of name Latin spatula—spoon, allusion to spoon- or spatula-like lateral outline. Holotype MF-—0432, LV, TBD 2978, 736 m. Paratypes MF-0433, RV, TBD 2978, 736 m. MF-0434, LV, TBD 3177, 1 000 m. MF-0451, RV, TBD 3555, 590 m. Diagnosis Species with a compressed, spatula-shaped anterior lateral outline, and a deep, cleft-like posterior depression. Description In external lateral view, AM is broadly and symmetrically rounded. The antero-dorsal and AM areas are compressed, giving a spatula-like appearance. DEEP-WATER QUATERNARY OSTRACODA DAS 3 aes Bs Q R Fig. 17. Outlines of various species of Krithe and Parakrithe. A. Krithe capensis sp. nov., SAM-PQ-MF-0450, LV, TBD 3577, 453m. B. Krithe spatularis sp. nov., SAM-—PQ —MF-0451, RV, TBD 3555, 590 m. C. Krithe sp. 8, SAM—PQ-MF-0452, LV, TBD 3177, 1000 m. D. Krithe sp. 4, SAM—PQ-MF-0453, RV, TBD 3355, 2 070 m. E-F. Krithe sp. 7, TBD 6851, 2 916m. E. SAM-—PQ-0454, LV. F. SAM-PQ-MF-0446, RV. G-H. Krithe rex sp. nov., TBD 6851, 2 916 m. G. SAM—PQ-MF-0456, RV. H. SAM—PQ-MF-0438, LV. I. Parakrithe sp. 10, SAM-—PQ-MF-0449, RV, TBD 3553, 1003m. J-K. Krithe sp. 6, TBD 6851, 2 916 m. J. SAM—PQ-—MF-0457, LV. K. SAM—PQ-MF-—0445, RV. L-M. Krithe peypouqueti sp. nov., TBD 6851, 2916m. L. SAM-—POQ-MF-0441, LV. M. SAM- PQ-MF-0439, RV. N. Krithe sp.9, SAM-PQ-MF-0443, RV, TBD 3524, 475 m. O-P. Krithe sp. 19, TBD 6851, 2916m. O. SAM-PQ-MF-0447. Q-R. Krithe sp. 22, TBD 6851, 2 916 m. OQ. SAM—PQ-MF-0458, LV. R. SAM-PQ-MF-0448, RV. Scale bar = 500 microns. 274 ANNALS OF THE SOUTH AFRICAN MUSEUM D ¥ » ig « « » Fig. 18. Details of anterior marginal areas of Krithe and Parakrithe, placed in order of lower depth limit of species. All internal views. A. Krithesp.9, SAM-—PQ-MF-0443, RV, TBD 3524, 475 m. B. Parakrithe sp. 10, SAM—PQ-—MF-0449, RV, TBD 3553, 1 003 m. C. Krithe sp. 8, SAM—PQ-—MF-0452, LV, TBD 3177, 1 000m. D. Krithe capensis sp. nov., SAM-PQ-MF-0450, LV, TBD 3577, 453m. E. Krithe spatularis sp. nov., SAM-—PQ- MF-0451, RV, TBD 3555, 590 m. F. Krithe sp. 4, SAM-PQ-MF-0453, RV, TBD 3355, 2070m. G. Krithe sp.6, SAM-PQ-MF-0457, LV, TBD 6851, 2916m. H. Krithe rex sp. nov., SAM-—-PQ-MF-—0438, ?female, LV, TBD 6851, 2916m. I. Krithe sp. 19, SAM-PQ-MF-0447, RV, TBD 6851, 2916m. J. Krithe sp. 7, SAM-—POQ-MF-0454, LV, TBD 6851, 2916m. K. Krithe sp. 22, SAM-—PQ-MF-0448, RV, TBD 6851, 2 916m. L. Krithe peypouqueti sp. nov., SAM-PQ-MF-0439, holotype, RV, TBD 6851, 2 916 m. Scale bar = 500 microns. DM and VM parallel, although there is a slight concavity at about mid-length in the VM. PM asymmetric, sloping steeply to a ventral apex. In internal view, the posterior depression is very prominent and lies in a deep, ventrally open cleft. Anterior vestibule is large and posteriorly wide. Outline of inner margin is sym- metrically rounded, with an anterior apex. DEEP-WATER QUATERNARY OSTRACODA Das) Dimensions (mm) length height MF-0432 1,09 0,55 MF-—0433 IBO2 0,50 MF-0434 let OSs MF-0451 Le 0,49 The mean length/height ratio of the type specimens is 2,07, which is the highest value we have recorded amongst the large species of Krithe in our study (Figs 19, 20). Remarks The distinctive shape of the new species has no close analogues from southern Africa, but is similar to the following species from elsewhere: K. hiwanneensis Howe & Lea, 1936, from the Oligocene of Louisiana (which is not compressed anteriorly); K. oertlii Dieci & Russo, 1967, from the Miocene of Italy (which does not have a straight DM); and K. vandenboldi Steineck, 1981, from the Miocene of Jamaica (which is a smaller species with a slightly concave VM outline, and a differently shaped anterior vestibule). Krithe spatularis sp. nov. would seem to fit into the ‘ecotype’ Krithe sp. F of Peypouquet (1979), which he suggested is an indicator of low dissolved O2 in the water column. 0-6 2 (a) 7 / fl /, peypouqueti / ,ao, S27 E ve Qo) Mal ; { \_ E05 eee ao iS . . 2 capensis spatularis £ 0:4 sp 224 Tey) 2 ah > oa a a ‘ol = eae 2S exe Nee en Z (a o> sp.19 0:3 length, mm ©) = capensis N = sp.7 & = spatularis @ = rex Fig. 19. Length versus height scattergram of Krithe species. Data points are specimens illustrated herein. 276 ANNALS OF THE SOUTH AFRICAN MUSEUM O spatularis co Capensis sp. 4 mean length/height O peypouqueti 0-7 0:8 0-9 1:0 11 mean length, mm Fig. 20. Ratios of mean length/mean height plotted against mean lengths for the illustrated specimens of the various species of Krithe. Off south-western Africa, K. spatularis has a latitudinal range 19°S to 35°S, and a depth range 392-1 662 m (a total of 1 272 m) (Figs 13, 14). It is most abundant (relative to other species of the genus) at those levels where K. capen- sis is relatively unsuccessful, i.e. immediately above the Salinity Minimum Zone in the AAIW (lower Neritic Zone), and in the lower part of the Upper Bathyal Zone, although we never recorded it exceeding 3 per cent of the total ostracod population. Despite a greater depth range than K. capensis, K. spatularis prob- ably tolerates a narrower salinity and temperature range (34,8-34,5%o to <34,6%0; 8,5—6°C to 3,2—2,8°C), because of its deeper upper depth limit. Krithe rex sp. nov. Figs 17G—H, 18H, 21A—D ?Krithe sp. C Cronin, 1983, pl. 10 (figs B—C). Derivation of name Latin rex—king, allusion to large, bold species. Holotype MF-—0435, RV, male, TBD 6851, 2 916 m. DEEP-WATER QUATERNARY OSTRACODA Die, Paratypes MF-0436, RV, female, TBD 6851, 2 916 m. MF-0437, RV, male, TBD 6851, 2 916 m. MF-0438, LV, ?female, TBD 6851, 2 916 m. MF-0456, RV, male, TBD 6851, 2 916 m. Diagnosis Large species with humped DM and convex VM. There is a prominent antero-dorsal step in lateral outline. Description A large species, with a distinctive, bold shape and marked sexual dimor- phism. In external lateral view, AM is broadly rounded, PM is truncated and short, sloping to a postero-ventral apex. The DM is strongly arched, particularly in the shorter (presumed) females, in which the convex VM is also more promi- nently developed. The central area of the valves is somewhat inflated. In internal view, the posterior depression is clearly visible but is neither large nor incised. The anterior inner lamella is wide but the vestibule is relatively small and complex in shape, lying between the anterior margin and two large lobe-like re-entrants in the line of concrescence. The inner margin outline has a small neck-like incision that links the vestibule to the interior of the shell. This may represent a weak zone that preferentially suffers abrasion. The MS consist of a forward pointing V-shaped anterior scar, with a small super-adjacent scar, and four elongate adductors, the most dorsal of which is partially subdivided. Dimensions (mm) length height MF-0435 1,15 0,60 MF-0436 1,18 OMS MF-0437 1,12 0,51 MF-0438 1,18 0,65 MF-—0456 Ly 0,60 The mean length/height ratio of the type specimens is 1,96. When plotted against valve length (Fig. 20), Krithe rex is seen to be a more elongate taxon than Krithe sp. 7 and Krithe sp. 8, with which it has an overlapping length/height field on the scattergram (Fig. 19). Remarks This is the largest species of the genus Krithe in the ostracod populations off south-western Africa (the holotype is 1 150 microns in length), and can be assigned to Peypouquet’s (1979) category D2. No species analogous to Krithe rex has been recorded from southern Africa, but the following species from elsewhere are similar in general appearance: Krithe cubensis van den Bold, 1946, from the Oligocene of Cuba (which is less inflated, and possesses relatively large vestibules); Krithe morkhoveni van den 278 ANNALS OF THE SOUTH AFRICAN MUSEUM ad Fig. 21. A-D. Krithe rex sp. nov. TBD 6851, 2 916m. A. SAM—PQ-MF-0435, holotype, male, RV, SEM 2691. B. SAM-—PQ-MF-0436, female, RV, SEM 2694. C. SAM- PQ-MF-0437, male, RV internal view, SEM 2701. D. SAM—POQ-MF-0438, ?female, LV. E-F. Krithe peypouqueti sp. nov. TBD 6851, 2 916m. E. SAM—PQ-MF-0439, holotype, RV. F. SAM—POQ-MF-0440, LV, SEM 2695. Scale bars: B = 200 microns, others = 100 microns. DEEP-WATER QUATERNARY OSTRACODA 279 Bold, 1960, from the Miocene of Trinidad (which is a much smaller species, has a different MS pattern, and possesses relatively large vestibules); K. trinidaden- sis van den Bold, 1958a, from the Oligocene to Miocene of Trinidad (which has a less strongly arched DM and the males have a more acuminate PM outline than K. rex sp. nov.). Brady (1880, pl. 27 (fig. /)) illustrated a species as Krithe producta (probably not conspecific with the lectotype designated by Puri & Hulings 1976), which lacks the LV antero-dorsal step. Cronin (1983) illustrated a species (Krithe sp. C, pl. 10 (figs B—C)) from his deepest sample off Florida (1 070 m), which may be conspecific with Krithe rex sp. nov. and is the closest relative that we have observed in the literature. Guer- net’s Krithe sp. 1 (1985, pl. 1 (fig. 16)) from the Upper Eocene of DSDP site 214 (90 East Ridge, Indian Ocean) also has a similar lateral outline, although its VM is somewhat less convex. Off south-western Africa, we have recovered Krithe rex at only one site (TBD 6851) at a water depth of 2 916 m, where it constitutes 3 per cent of the total ostracod population and 6 per cent of the Krithe population. In our area it is clearly an abyssal species, but if Cronin’s (1983) Krithe sp. C is conspecific, then in the north-western Atlantic at least, the species ranges into the depths equivalent to the Upper Bathyal Zone as we define it in the south-eastern Atlantic. Krithe peypouqueti sp. nov. Figs 17L—M, 18L, 21E-F, 22A Derivation of name The species is named for Dr J.-P. Peypouquet (University of Bordeaux) for his work on the genus Krithe. Holotype MF-0439, RV, TBD 6851, 2 916 m. Paratypes MF-0440, LV, TBD 6851, 2 916 m. MF-0441, LV, TBD 6851, 2 916 m. Diagnosis A globular species with a rounded DM, and a strong antero-dorsal step in the RV lateral outline. Description A squat, globular species in lateral view. AM broadly rounded, PM asym- metrically rounded, truncated in LV. DM strongly arched and rounded, with a prominent antero-dorsal step in the RV. VM broadly convex. Central valve area is inflated. In internal view, the inner margin runs approximately parallel to the AM, and the inner lamella and vestibule are narrow. 280 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 22. A. Krithe peypouqueti sp. nov., SAM—PQ-—MF-0441, LV, TBD 6851, 2 916 m. B-C. Krithe sp. 8, SAM—PQ-MF-0442, TBD 3177, 1000 m. B. RV internal view, SEM 2715. C. External view. D. Krithe sp. 22, SAM-PQ—MF-0448, RV, TBD 6851, 2 916 m. E. Krithe sp. 4, SAM—POQ-MF-—0444, RV, TBD 3355, 2 070 m. F. Krithe sp. 6, SAM-PQ-—MF-0445, RV, TBD 6851, 2 916 m. Scale bars = 100 microns. DEEP-WATER QUATERNARY OSTRACODA 281 Dimensions (mm) length height MF-—0440 0,81 0,60 MF-0439 0,78 0,56 MF-0441 0,79 0,56 The mean length/height ratio of the type specimens is 1,38, which combined with its mean length values (0,79 mm) easily distinguishes Krithe peypouqueti from all other species of the genus that occur off south-western Africa (Figs 19, 20). Remarks The species most similar in lateral outline to Krithe peypouqueti sp. nov. that we have encountered in the literature is a specimen recorded by Guernet (1983, pl. 1 (fig. 15)) under Krithe sp. 1 (his category is clearly poly-specific) from the Upper Eocene of DSDP site 214 on the 90 East Ridge. However, this species has a wide inner lamella. Off south-western Africa, Krithe peypouqueti has a depth range 2 916—4 736 m, and is one of the most abundant ostracod species in the Abyssal Zone. At Site TBD 6851 (2 916 m), it constitutes 13 per cent of the total ostra- cod fauna (second in abundance to Poseidonamicus major) and 28 per cent of the Krithe population (the most abundant species of the genus). At site 6852 (4 736 m, where only three valves were recovered), Krithe peypouqueti forms 66 per cent of the total ostracod fauna. OTHER SPECIES OF KRITHE Neritic/Bathyal species Krithe sp. 8 Figs 17C, 18C, 22B-—C Illustrated material MF-0442, RV, TBD 3177, 1 000 m. MF-0452, LV, TBD 3177, 1 000 m. Remarks This is a relatively large species (see Fig. 19) that has a distinctive postero- dorsal arch, giving it a ‘humped back’ appearance in lateral view. Its overall outline is similar to Krithe rex sp. nov., but it is somewhat less elongate (mean length/height ratio = 1,84) (Fig. 20). The posterior depression is deep, rounded, and set in a wide flat furrow. The inner lamella is relatively wide, with large anterior vestibules, and the outline of the inner margin is asymmetrically 282 ANNALS OF THE SOUTH AFRICAN MUSEUM rounded, with an antero-ventral apex. MS consist of an irregular U-shaped and small elongate anterior set, and four elongate adductors, the dorsalmost of which is almost subdivided. Krithe sp. 8 has a depth range of 530-1 353 m off south-western Africa (i.e. Neritic to Upper Bathyal) and, although it is generally rare (1-6% total ostra- cod fauna), the species appears to opportunistically exploit an environmental niche at the Neritic/Bathyal boundary. Here, straddling the base of the Salinity Minimum Zone, it is the most abundant species in the Krithe population (up to 11% of the total ostracod fauna, and >80% of the Krithe fauna), locally sup- planting the two important taxa in the Upper Krithe Fauna (Krithe capensis and K. spatularis) (see Figs 11, 13). Krithe sp. 9 Figs 17N, 18A, 23C Illustrated material MF-0443, RV, TBD 3524, 475 m. Remarks Although not a bathyal species, we include a record of Krithe sp. 9 for com- pleteness. This is a sub-quadrate, elongate species with a large and complex anterior vestibule. It has a depth range of 430—900 m, so that its lower range overlaps with Krithe capensis, K. spatularis, and Krithe sp. 8. Bathyal/Abyssal species Four species of Krithe appear immediately below the Upper/Lower Bathyal boundary (i.e. at the top of the NADW mass). These constitute the taxa in the upper part of the Lower Krithe Fauna: Krithe sp. 4, Krithe sp. 6, Krithe sp. 7, and Krithe sp. 19 (Figs 11, 13). Krithe sp. 4 Figs 17D, I8SE, 22E Illustrated material MF-0444, RV, TBD 3355, 2 070 m. MF-0453, RV, TBD 3355, 2 070 m. Remarks This is a relatively large quadrate species, with a distinctively truncated PM outline in lateral view. On the length/height scattergrams, this species occupies a field between the species with mean lengths >1,00 mm and K. capensis (Figs 19, 20). The anterior vestibule is moderately large and widens anteriorly from a DEEP-WATER QUATERNARY OSTRACODA 283 narrow posterior ‘neck’. The outline of the inner margin has a rounded apex at about mid-height. None of our adult specimens had well-preserved hinges but there appears to be a strong, elongate, tooth-like structure at the anterior end of the RV ME. The single anterior MS has three anteriorly directed lobes. Krithe sp. 4 has a depth range 1 600-2 916 m off south-western Africa. It is the dominant species of the Lower Krithe Fauna in the Lower Bathyal and upper Abyssal levels (2% of total ostracod fauna), below which it is replaced by K. peypouqueti (Figs 11, 13). This distribution pattern suggests that Krithe sp. 4 is tolerant only of the physico-chemical environments found in the upper part of the NADW mass. Krithe sp. 6 Figs 17J-K, 18G, 22F Illustrated material MF-0445, RV, TBD 6851, 2 916 m. MF-0457, LV, TBD 6851, 2 916 m. Remarks A moderate-sized species with a semi-elliptical outline in lateral view, in which the RV posterior margin has a small re-entrant in the vicinity of the pos- terior depression. With the exception of Krithe peypouqueti, the mean length/height ratio of 1,58 is the lowest that we have recorded for a species of Krithe and gives this species an isolated position on scattergrams (Figs 19, 20). The vestibules of Krithe sp. 6 are very narrow, with the inner margin lying close and parallel to the outer margin. MS appear complex, with three small anterior scars and four adductors. Krithe sp. 6 has a depth range of 1 662-2 916 m. The species is a minor component of the Lower Krithe Fauna, but is relatively more abundant in the Lower Bathyal Zone. Krithe sp. 7 Figs 17E-F, 18J, 23A Illustrated material MF-0446, RV, TBD 6851, 2 916 m. MF-0454, LV, TBD 6851, 2 916 m. Remarks A relatively large species with a mean length/height ratio of 1,89 (Fig. 20). In external lateral view, the species is characterized by a broadly rounded AM that contrasts with a somewhat extended PM, which in the RV has a small 284 ANNALS OF THE SOUTH AFRICAN MUSEUM re-entrant above the posterior depression. The DM is gently convex, which helps to distinguish it from Krithe sp. 4. The anterior inner lamella is moderately wide but the vestibule is small and almost rectangular in shape. Krithe sp. 7 has a depth range 1 600-2 916 m. The species is a minor com- ponent of the Lower Krithe Fauna but is relatively more abundant in the Lower Bathyal Zone, particularly in the uppermost part. Krithe sp. 19 Figs 170-P, 181, 23B Illustrated material MF-0447, RV, TBD 6851, 2 916 m. Remarks A small elongate species that has the highest mean length/height ratio within the Krithe populations that we studied (2,30—Figs 19, 20). In external lateral view, Krithe sp. 19 is characterized by prominent re-entrants in both valves, above the posterior depression, that give the valve a ‘pleated’ appear- ance. The anterior vestibules are small, and largely lie anterior to two lobes in the marginal zone. The anterior MS appears to consist of a cluster of three small scars. Krithe sp. 19 has a depth range 1 662-2 916 m. Although it occurs in small numbers (maximum of 4% total ostracod fauna at TBD 6851), it is the second most abundant taxa of the Lower Krithe Fauna throughout most of the Lower Bathyal Zone (Fig. 13). Krithe sp. 22 Figs 17Q-R, 18K, 22D Illustrated material MF-0448, RV, TBD 6851, 2 916 m. MF-0458, LV, TBD 6851, 2 916 m. Remarks A small species with a gently rounded DM outline in lateral view. Its overall shape is very similar to that of Krithe capensis and the two have similar length/height ratios (Krithe sp. 22 = 1,93; Krithe capensis = 1,86). However, the great difference in size between the two species facilitates identification (Fig. 20). Internally, the two species differ in shape of the vestibule; in Krithe sp. 22 it is small, with a short, straight, inner margin post-adjacent to it. Krithe sp. 22 is a rare abyssal form that we encountered only in sample 6851 (2 916 m), where it constitutes 12 per cent of the total Krithe population. _ AN DEEP-WATER QUATERNARY OSTRACODA 285 ose AS SS WW Se AE Bee, Fig. 23. A. Krithe sp. 7, SAM—PQ-MF-0446, RV, TBD 6851, 2 916m. B. Krithe sp. 19, SAM-—PQ-MF-0447, RV, TBD 6851, 2 916m. C. Krithe sp. 9, SAM—PQ-—MF-—0443, RV, TBD 3524, 475m. D. Parakrithe sp. 10, SAM-—PQ-MF-0449, RV, TBD 3553, 1 003 m. E-F. Buntonia rosenfeldi sp. nov. TBD 3338, 990 m. E. SAM—PQ-MF-—0460, LV, SEM 2799. F. SAM—PQ-MF-0459, holotype, RV, SEM 2792. Scale bars = 100 microns. 286 ANNALS OF THE SOUTH AFRICAN MUSEUM Genus Parakrithe van den Bold, 1958a Parakrithe sp. 10 Figs 171, 18B, 23D Illustrated material MF-0449, RV, TBD 3553, 1 003 m. Remarks This relatively large (0,72 mm) species has a slight bulge on the antero- dorsal margin in lateral view, and a less pointed postero-ventral outline than in most of the examples discussed by Peypouquet (1979). The anterior vestibule is relatively large and widens anteriorly. We place it within Peypouquet’s (1979) ecotype A3. Van den Bold (1966) illustrated three species of the genus from the Miocene of Gabon (P. robusta sp. nov., P. datylomorpha Ruggieri, 1962, and P. vermunti (van den Bold, 1946)), but none are conspecific with our species. Off south-western Africa, Parakrithe sp.10 has a depth range of 900-1 353 m, which limits it to the portion of the AAIW that lies between the base of the Salinity Minimum Zone, and the top of the NADW mass. Family Buntoniidae Apostolescu, 1961 Genus Buntonia Howe, 1935 (in Howe & Chambers, 1935) Ruggieri (1958) split the genus Buntonia and separated off the large, rela- tively thin-shelled and smooth taxa into Quasibuntonia. Van Morkhoven (1963) did not accept this as taxonomically valid, and we follow his example. However, the forms that were represented in Ruggieri’s Quasibuntonia certainly constitute a geographically well-defined group: they inhabit relatively deep water, and are (and were) confined to areas adjacent to Africa, and in the Mediterranean (Benson & Sylvester-Bradley 1971). Representatives of Buntonias.s. have a world-wide distribution at all water depths. The genus Buntonia is an important component of the ostracod populations off south-western Africa, where five species occur over a latitudinal range 17°S to 35°S, in water depths of 95-2 070 m (Fig. 24). All five species are found in the Neritic Zone, where mean values of the genus’s abundance within the ostra- cod population (in samples that contain the genus) are 9,8 per cent (range 0,1-66%). Only two species (B. rosenfeldi sp. nov. and Buntonia sp. 34) extend beyond the continental shelf on to the slope, and only B. rosenfeldi sp. nov. occurs in the Bathyal Zone. The latter species belongs in the group Quasibunto- nia, as understood by Ruggieri (1958) and Benson & Sylvester-Bradley (1971). We did not find the genus in the Abyssal Zone. Despite the reduction in diversity of the genus into progressively deeper water, mean values of abundance in the Bathyal Zone are higher than in shallow-water areas: 30 per cent (range 3-100%). A three-point running mean DEEP-WATER QUATERNARY OSTRACODA 287 water depth, km latitude °S @- B.rosenfeldi ™@— modern specimens Fig. 24. Latitude and water depth of samples bearing Buntonia spp. Solid points and outline indicate the distribution of Buntonia rosenfeldi sp. nov. plot of the percentage of all Buntonia species within the overall ostracod popu- lations from all our samples off south-western Africa (Fig. 25) shows a general increase in abundance of the genus from the inner shelf to the lower part of the Neritic Zone (i.e. 95—900 m). We detect three populations within these shallow- water assemblages but will defer further discussion to a later publication. At the Neritic/Bathyal boundary (c. 950 m), there is fluctuation at the apex of the curve, which peaks at around 25 per cent of the overall ostracod population. Below this depth, in the Upper Bathyal Zone, values fall steadily for about 250 m. Clearly, there is an important physio-chemical barrier at around 950 m that has the following effects: (a) the disappearance of Buntonia sp. 34; (b) a reduction in the relative size of the population of B. rosenfeldi sp. nov. Buntonia rosenfeldi sp. nov. is the only bathyal species of the genus, although valves of Buntonia sp. 34 occur as deep as 1 050 m. Mean values of percentage of the ostracod population show a high in the Upper Bathyal Zone compared to the Lower Bathyal Zone (35%, range 6—-100%; 20%, range 3—40%, respectively), but these figures hide the true distribution, which is closely linked to the boundary zones of the water masses. We will discuss this under ‘Remarks’ on B. rosenfeldi sp. nov. 288 ANNALS OF THE SOUTH AFRICAN MUSEUM Peypouquet & Benson (1980) recorded the genus (as Quasibuntonia spp.) from their transects in the Cape and Angola basins. There are too few data points on their curves to make a detailed comparison with our results (cf. Figs 25, 26), but they suggest a decline from high values of percentage of overall ostracod populations in shallow water (20% in 450 m) to a low of 1 per cent at around 1 000 m. They recorded three sites with relatively high values (20-30%) in depths that we classify as Lower Bathyal/uppermost Abyssal, and a deeper site (2 800 m) at 6 per cent. These data are broadly compatible with our results, but we suspect that they are too scattered to resolve the actual distribution. Benson & Sylvester-Bradley (1971) discussed the distribution of the related, relatively large species, B. sulcifera (Brady, 1887) and _ B. radiatopora (Seguenza, 1880), and concluded that these deep-water buntoniids represent the remnants of a “Tethyan’ fauna that is restricted to a circum-Africa zone. t+— B. sp. 72 UPPER LOWER NERITIC BATHYAL BATHYAL mean % =9:°8 mean % = 35 mean % = 20 % Buntonia spp. ol jo) water depth, km Fig. 25. Buntonia species as percentage of total ostracod fauna plotted against water depth. Values are five point running means. Depth ranges for all Buntonia species on the continental margin are shown by bars. Mean % = mean percentages of Buntonia species in total ostracod fauna for each depth zone. DEEP-WATER QUATERNARY OSTRACODA 289 30 NERITIC BATHYAL ABYSSAL | UPPER LOWER i) (oe) % Buntonia spp. _ fo) { 2 3 water depth, km Fig. 26. Buntonia species as percentage of total ostracod fauna plotted against water depth for a profile off Walvis Bay. Data computed from Peypouquet & Benson (1980). Neither Cronin (1983) nor Benson et al. (1983) recorded taxa from the con- tinental margin of eastern North America that would fall into the category Quasibuntonia. Buntonia rosenfeldi sp. nov. Figs 23E-F, 27A—D Buntonia sulcifera? (Brady, 1887) Rosenfeld & Bein, 1978: 18, pl. 1 (fig. 21). Buntonia sp. 1 Boomer, 1985: 34-35, pl. 2 (figs 27-28). (These are probably all juveniles.) Buntonia sp. 2 Boomer, 1985: 35-36, pl. 2 (figs 33-34). (These are probably all adults.) Derivation of name The species is named for Dr A. Rosenfeld (Geological Survey of Israel) for his work on deep-water ostracods from north-western Africa. Holotype MF-0459, RV, TBD 3338, 990 m. Paratypes MF-0460, LV, TBD 3338, 990 m. MF-0461, LV, TBD 3338, 990 m. MF-0462, RV, TBD 3109, 900 m. 290 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 27. A-D. Buntonia rosenfeldi sp. nov. A-C. TBD 3338, 990 m. A. SAM-—PQ-MF-0459, holotype, LV seta, SEM 2793. B. SAM-—PQ-—MF-0461, LV internal view, SEM 2795. C. SAM-PQ-MF-0461, LV, MS, SEM 2796. D. SAM-—PQ-MF-0462, RV, TBD 3109, 900 m, SEM 2790. E-F. Dutoitella suhmi (Brady 1880), TBD 6851, 2 916 m. E. SAM—PQ-MF-0463, LV, SEM 2936. F. SAM—PQ-MF-—0464, RV, SEM 2940. Scale bars: A = 10 microns; C-F = 100 microns. DEEP-WATER QUATERNARY OSTRACODA 291 Diagnosis Species with well-developed longitudinal ventro-lateral keel and reticulate ornamentation in posterior part of lateral valve surface. Description Typical ovate buntoniid in lateral outline with broadly rounded AM and narrow, bluntly truncated PM. High, arched DM that has an abrupt step to the PM in LV, but is continuous with PM outline in RV. VM is straight or slightly concave. There is a prominent, keeled, ventro-lateral ridge that passes anteriorly into a faint rib running parallel to the AM. AM area is strongly compressed. Posterior lateral surface is reticulate, with indistinct longitudinal ribbing and prominent conjugate pores. Anterior valve surface is faintly reticu- late, with a delicate tracery of muri and inter-mural pitting. Hinge is amphidont, with both ATE in LV open ventrally. There are narrow anterior vestibules. MS consist of four adductors, the ventral and dorsal of which are ovate, and a hooked anterior scar, above which lie two small rounded scars. Dimensions (mm) length height ME-0459 0,66 0,43 MF-—0460 0,69 0,45 MF-0461 0,68 0,46 MF-0462 0,67 0,44 Remarks Three species are closely related to B. rosenfeldi sp. nov.: B. radiatopora (Seguenza, 1880), B. pyriformis (Brady, 1880), and B. sulcifera (Brady, 1887). The closest is B. radiatopora (Seguenza, 1880) from the Neogene of south- ern Italy. Seguenza (1880) recognized two varieties, but stated that there is a continuous transition between them and that he was unable to differentiate separate species. Buntonia radiatopora radiatopora has longitudinal ridges in the posterior half of the valves, whereas the variety B. r. sculpta has ridges over the whole valve surface. The former variety has been illustrated by Benson & Sylvester-Bradley (1971), and both varieties by Colalongo (1965). Buntonia rosenfeldi sp. nov. differs from B. radiatopora (Seguenza) in having reticulate ornamentation in the posterior half of the valve. Buntonia sulcifera (Brady, 1887) is very close to B. radiatopora (Seguenza), because both species possess ornamentation of longitudinal ribs in the posterior half of the valve. They may be conspecific, although the rib pattern in the former may be sharper if the illustration by Benson & Sylvester-Bradley (1971) of a specimen from the Mozambique Channel is the same species as Brady’s. Whatever the case, B. rosenfeldi differs from B. sulcifera by lacking the longi- tudinal median ribs in the posterior half of the valve. 292 ANNALS OF THE SOUTH AFRICAN MUSEUM Buntonia pyriformis (= B. mackenziei nom. nov. Puri & Hulings, 1976) has a similar outline to B. rosenfeldi but is more elongate, has a smooth lateral valve surface (Brady 1880: 78), and an eyespot (Puri & Hulings 1976: 281). Buntonia rosenfeldi sp. nov. occurs over a depth range of 186-2 070 m off south-western Africa, and 1 418-2 859 m off north-western Africa (recorded as B. sulcifera? by Rosenfeld & Bein 1978, fig. 21). Buntonia radiatopora is reported as living at 2 816 m in the eastern Mediterranean and fossil in the Plio—Pleistocene of southern Italy and Sicily (Benson & Sylvester-Bradley 1971). Buntonia sulcifera (Brady, 1887) was recorded by Brady (1887) from 3 655 m off Mauritius, and 2 980 m from the Mozambique Channel by Benson & Sylvester-Bradley (1971), and B. pyriformis (Brady, 1880) was first recorded in 675 fm (1 234 m) off Brazil (Brady 1880—‘Challenger’ site 120), and has been found between 400 m and 830 m off north-western Africa by Rosenfeld & Bein (1978). Buntonia cf. B. pyriformis (Brady) (recorded as cf. B. mackenziei Puri & Hulings) has been reported from late Miocene and late Pliocene strata at DSDP site 608 in the North Atlantic by Whatley & Coles (1987). Figure 28 shows the abundance of Buntonia rosenfeldi sp. nov. across its depth range off south-western Africa. In the Neritic Zone it constitutes a mean | | NERITIC ! BATHYAL : ! UPPER | LOWER mean %=6:°5 mean % = 29 mean % = 20 % Buntonia rosenfeldi 0:5 1-0 1:5 2:0 water depth, km Fig. 28. Buntonia rosenfeldi sp. nov. as percentage of total ostracod fauna plotted against water depth. Values have not been smoothed. DEEP-WATER QUATERNARY OSTRACODA 293 of 6,5 per cent (range 1-25%) the ostracod population (in samples in which it occurs) and, for depths shallower than approximately 900 m, there are few samples with values above 10 per cent, although there is a slow, progressive increase with depth. Within the Bathyal Zone (i.e. 950—2 070 m), the mean abundance of B. rosenfeldi is 27 per cent (range 1-100%), so that it appears to represent a characteristic element of the deep-water fauna. However, it is at water mass boundaries that the species is most successful in establishing itself: at the Neritic/Upper Bathyal boundary values are c. 20 per cent, and at the Upper/ Lower Bathyal boundary values are over 35 per cent. Nevertheless, we did not record it from the Abyssal Zone, and values in the Lower Bathyal Zone steadily decline below about 1 600 m. From these data we conclude that B. rosenfeldi prefers relatively cold, saline water, and that it is particularly successful, in comparison with other ostracod species, at tolerating the unstable conditions that occur at the major water mass boundaries (i.e. at the base of the AAITW low salinity zone, and the AAIW/NADW shear zone). It cannot tolerate the adverse conditions of the AABW. Family Trachyleberididae Sylvester-Bradley, 1948 Subfamily Unicapellinae Dingle, 1981 Genus Dutoitella Dingle, 1981 Recognition of this genus in Quaternary deep-sea sediments in the south- eastern Atlantic necessitates expansion of the concept of this genus to include reticulate species. This in turn allows the accommodation within Dutoitella of several species of previously uncertain affinity: D. eocenica (Benson, 1977) from the South Atlantic, ‘Suhmicythere’ sp. Benson et al., 1983, from the north- western Atlantic, and D. crassinodosa (Guernet, 1985) from the Indian Ocean. Phylogenetic implications of these new data are that, between Maastrichtian and Eocene times, the genus migrated from south-eastern African outer continental- shelf environments into deeper-water habitats on both sides of the South Atlantic and in the Indian Ocean. There was a concomitant development of reticulation. A similar trend can be seen in Aflanticythere, with spinose species appearing by Eocene times. Architectural similarity between the two genera strongly sug- gests a common ancestor (Fig. 29). Dutoitella suhmi (Brady, 1880) Figs 27E-F, 30A-B, 31A, C, F Cythere suhmi Brady, 1880: 106-107, pl. 26 (fig. 3a—h). Puri & Hulings, 1976: 290-291, pl. 17 (figs 7-12), text-fig. 10. ‘Suhmicythere’ suhmi (Brady, 1880): Whatley & Coles, 1987, pl. 6 (figs 18-21). ‘Suhmicythere’ sp. Benson et al., 1983, pl. 1 (fig. 8). 294 ANNALS OF THE SOUTH AFRICAN MUSEUM MIO PLI QUA Dutoitella * mimica' $ eocenica2® *# suhmi*® * sp. indet.® * crassinodosa’ * dutoiti' Atlanticythere *~ maestrichtia2 * murareticulata2 * prethallasia? Uncertain affinity ~ carlitae® * = eastern Atlantic Ocean $ = western Atlantic Ocean # = north-western Atlantic Ocean * = Indian Ocean sn--=-=- = known range Saeae = assumed range '= Dingle 1981; * = Benson 1977; ° = Frewin 1987; ¢ = Brady 1880; ° = Benson & Peypouquet 1983; ® = Benson et a/. 1983; ” = Guernet 1985; ® = Peypouquet & Benson 1980 (= ‘Shumicythere’ [sic]); 9 = Miocene record originally designated Atlanticythere? neogenica by Benson (1977). Fig. 29. Geological ranges of modern and fossil species of Dutoitella and Atlanticythere from the Atlantic and Indian oceans. Illustrated specimens MF-—0463, LV, TBD 6851, 2 916 m. MF-0464, RV, TBD 6851, 2 916 m. MF-0465, RV, TBD 6851, 2 916 m. MF-0466, LV, TBD 6851, 2 916 m. Remarks The species, first described by Brady (1880) from the north-western Pacific and Prince Edward Island (Southern Ocean) as Cythere suhmi Brady, 1880, is externally similar to the type species of Dutoitella, having the typical lateral DEEP-WATER QUATERNARY OSTRACODA 295 Fig. 30. A-B. Dutoitella suhmi (Brady, 1880), TBD 6851, 2 916 m. A. SAM—PQ-MF-0465, RV internal view, SEM 2942. B. SAM—-PO-MEF-0466, LV internal view, SEM 2938. C-F. Abyssocythere australis Benson, 1971, TBD 6851, 2916m. C. SAM-—PQ-—MF-0467, LV, SEM 2946. D. SAM—PQ-MF-0468, RV, SEM 2960. E. SAM—PQ-MF-0469, LV internal view, SEM 2948. F. SAM—PO-MF-0470, RV internal view, SEM 2957. Scale bars = 100 microns. 296 ANNALS OF THE SOUTH AFRICAN MUSEUM 2) QD bd @ or £4 Fig. 31. Comparative morphology of various species of Dutoitella. A. D. suhmi (Brady 1880), SAM-PQ-MF-—0464, RV, TBD 6851, 2 916m, SEM 2940. B. D. mimica Dingle, 1981, SAM-K5748, RV, TBD 818, Alphard Formation, Agulhas Bank, Maastrichtian III. C. D. suhmi (Brady 1880), SAM-PQ-MF-0463, LV, TBD 6851, 2 916m, SEM 2936. D. D. neogenica (Benson 1977), USNM 190300, LV, DSDP site 22, Lower Miocene, from Benson (1977, pl. 1 (fig. 8)). E. ‘Suhmicythere’ sp., USNM 247710, 3000 m off New- foundland, Quaternary, from Benson et al. (1983, pl. 1 (fig. 8)); is probably conspecific with D. suhmi (Brady, 1880). F. MS of D. suhmi (Brady, 1880), SAM—PQ-MF-0465, RV, TBD 6851, 2 916m, SEM 2944. G. MS of D. mimica Dingle, 1981, SAM-K5449, LV, TBD 818, Alphard Formation, Agulhas Bank, Maastrichtian III. Scale bars = 100 microns. outline, with contrasting eared LV and stepped RV antero-dorsal outline, of the Cretaceous (Campanian—Maastrichtian) type species (D. dutoiti Dingle, 1981). Its MS pattern is very similar to that of D. mimica Dingle, 1981 (Santonian— Maastrichtian). Dutoitella suhmi (Brady) differs from both Cretaceous species in being reticulate, having a less prominent SCT, and having the lateral ridge or node post-adjacent to the SCT replaced by three short, indistinct ribs. With the exception of the reticulation, the surface architecture, including location of pore conuli and dorsal margin spines of D. suhmi (Brady) and D. mimica, is very similar (Fig. 31). The internal features of these two species are also very close, the major difference being that the LV PTE of D. suhmi lies in a slightly more postero-ventral position. DEEP-WATER QUATERNARY OSTRACODA 297 The illustrations of Puri & Hulings (1976: 290-291, pl. 17 (figs 7-12), text- fig. 10) of lectotypes of D. suhmi (Brady) show a slightly different MS pattern to our specimens of both D. suhmi (Brady) and D. mimica Dingle, 1981. The former show four non-divided adductors, whereas our specimens have a sub- divided second adductor. R. C. Whatley (pers. comm. 1988) informs us, how- ever, that this variation in D. suhmi is within the range that he has observed for the species worldwide, and appears to have no taxonomic significance. Brady (1880) reported Cythere suhmi from water depths of 100-4 600 m, and Whatley & Coles (1987) have found it in deep-water (>3 000 m) Miocene— Quaternary sediments of the North Atlantic. Dutoitella suhmi (Brady, 1880) is known from site TBD 6851 within the Abyssal Zone (2 916 m) of the Cape Basin, where it forms 6 per cent of the ostracod fauna. Benson et al.’s (1983) record of ‘Suhmicythere’ sp. from 3 000 m on the con- tinental slope off Newfoundland also probably refers to this species, and it is possible that Peypouquet & Benson’s (1980) citation of “‘Shumicythere’ [sic] from a range of 3 797-4595 m in the Angola Basin at least refers to species of Dutoitella. Benson (1977) listed ‘Suhmicythere’ sp. from late and mid-Miocene horizons at DSDP site 357 on the Rio Grande Rise, but did not illustrate the material. These may, therefore, also be records of reticulate species of Dutoitella. When he originally erected the two species Atlanticythere? eocenica Benson and A.? neogenica Benson, Benson (1977: 877) queried their generic assign- ment: “This species [A.? eocenica] and A.? neogenica Benson, n. sp. may be considered later as generically distinct [from other species of Atlanticythere].’ The continuous AM and ventro-lateral ridge indicates that Benson’s species belong in Dutoitella. In addition, we consider that these two species are conspe- cific and, in revision, designate D. eocenica (Benson, 1977) to be the valid taxon, because it is described earlier in the publication (i.e. D. neogenica (Benson, 1977) is a junior synonym of D. eocenica (Benson, 1977)). The species D. eocenica (Benson, 1977) and D. suhmi (Brady) have very similar patterns of surface reticulation, with the main points of difference being in the routes of the reticulation muri adjacent to the AM, and the presence of a short ridge post-adjacent to the SCT in D. eocenica compared to three narrow ribs in D. suhmi. In this latter feature, D. eocenica is close to the Cretaceous species of the genus (Fig. 31). We also refer ‘Cythereis’ crassinodosa Guernet, 1985, to Dutoitella and con- sider it to be another close relative of D. suhmi. Guernet (1985) compared his species generically to ‘Suhmicythere’ sp. Benson et al., 1983, and recorded it from bathyal assemblages of early (Ypresian) to late (Priabonian) Eocene age at DSDP site 214 (Chagos Ridge). He also mentioned ‘C.’ aff. C. crassinodosa from the Lower Eocene at DSDP site 245 (Madagascar Basin, south-western Indian Ocean), and ‘C.’ cf. C. crassinodosa from the middle Eocene at DSDP site 214 on the central 90 East Ridge. Dutoitella crassinodosa (Guernet, 1985) 298 ANNALS OF THE SOUTH AFRICAN MUSEUM differs from D. suhmi (Brady) in having a more prominent SCT and prominent nodes on its DM and VM. Phylogenetic implications of our new data are summarized in Figure 29 and suggest that in the early Tertiary, D. mimica migrated from its preferred habi- tats on the outer continental shelf (?100—200 m water depths—see Dingle 1981, 1985) and colonized deep-water sites in the western South Atlantic and in the south-western, north-western and central Indian Ocean. During the course of this, mutation into at least two species occurred: D. eocenica (Eocene—Miocene) and D. crassinodosa (Eocene). Subsequent development produced D. suhmi (Miocene—Quaternary) and possibly other species that have been recorded as ‘Suhmicythere’ (e.g. Benson 1977; Peypouquet & Benson 1980). At this stage it is premature to speculate on possible links between Dutoitella and Atlanticythere but, if the latter did evolve from the former, then the main architectural modifications necessary before Maastrichtian time were the re-organization of the route of the AM and ventro-lateral ridges, a reduction in the size of the SCT and post-adjacent nodes and ridges, and a general increase in valve size. So far, the only record of the genus Aflanticythere in the eastern South Atlantic is Aftlanticythere sp. BO58 from the Eocene of the Agulhas Bank (Frewin 1987). Neither Rosenfeld & Bein (1978) nor Cronin (1983), in their studies of deep-water faunas, referred species to the genera Dutoitella, Atlanticythere, or ‘Suhmicythere’ . Subfamily Trachyleberidinae Sylvester-Bradley, 1948 Genus Abyssocythere Benson, 1971 Abyssocythere australis Benson, 1971 Figs 30C-F, 32A Abyssocythere australis Benson, 1971: 18, pl. 3 (fig. 7), text-fig. 12. Illustrated specimens MF-0467, LV, TBD 6851, 2 916 m. MF-0468, RV, TBD 6851, 2 916 m. MF-0469, LV, TBD 6851, 2 916 m. MF-0470, RV, TBD 6851, 2 916 m. Remarks Externally, the specimens from off south-western Africa differ from those off southern Australia only by possessing a slightly weaker AM rim. In his orig- inal description of A. australis, Benson (1971) did not give details of the internal features of the species, but we can report that the MS pattern (Fig. 32) is similar to that of the type species A. casca Benson, 1971. Also, the hinges of the two DEEP-WATER QUATERNARY OSTRACODA uS) sy SS Fig. 32. MS of Abyssocythere. A. A. australis Benson, 1971, SAM-PQ-MF-0470, RV, TBD 6851, 2 916 m , SEM 2958. Scale bar = 100 microns. B. A. casca Benson, 1971, no local- ity given, c. 3 000 m, off Madagascar (from Benson 1971, fig. 6). No scale given. Species are similar but, in comparison with A. casca, our material exhibits a weaker ATE in the RV, where the anterior tooth is more elongate and in lateral view projects above the valve margin at the ACA. In addition, the socket at the anterior end of the narrow ME groove is small and does not project below the general line of the hinge, as occurs in A. casca. When he originally proposed the genus, Benson (1971) recorded six species of Abyssocythere, all of which had modern representatives. Subsequently, four more have been reported in the literature, with the result that ten separate species are now known, six of which have a fossil record that ranges Eocene to Recent (Table 3). Figure 33 shows the distribution of the Quaternary species of Abyssocythere. Our new data pose an interesting problem in that the geographical range of A. australis can now be extended across the Southern Ocean into the Cape Basin, but that the range of A. casca seems to be limited to the northern Mozambique/Somali basins. The species have overlapping depth ranges (Table 3), so presumably there is a physico-chemical barrier to the migration of A. casca southward through the deep-water passages into the Cape Basin, although the nature of this barrier is not obvious, particularly in view of the fact that another of the typical Abyssal Zone taxa (Poseidonamicus major) does extend from the Mozambique Basin into the South Atlantic. This species is a characteristic element of the Abyssal Zone ostracod assemblage. Our specimens were recovered from a single station (TBD 6851) at 2 916 m, where Abyssocythere australis constituted 6 per cent (23 valves) of the total ostracod fauna, with 17 per cent of the specimens considered modern. At the only previously known site (3 390 m, south of Australia), none of the speci- mens were modern (i.e. they were ‘relict’, presumably Pleistocene) (Benson 1971). No representatives of the genus Abyssocythere were recovered from the deep-water surveys of Rosenfeld & Bein (1978), Peypouquet & Benson (1980), Benson et al. (1983), and Cronin (1983). ‘Ue YON [eU9D—909 dasa ‘neoield one OkS—gTS dS ‘[ouURYD eUIEA —/1¢ dCSC ‘ostY 9pueIH OrY—9I1S ddSa ‘Yue sopjoyokas—/¢z ddSC ‘spuels] sp19A edeg—d7T dasa gusoo1y oddq. = (81S ddSq) onuefy (¢€861) JoNbnodAdg = UOSUDG ~=_ FSH ‘UOSUDG SisuaIIZDIG adayJKOOSSKqy gus00l[q JOMOT (LIS ddSd) MuenVy (¢€861) enbnodAdg 29 uOSUDg ‘ds asaysoosskqy 9ud90)SI9[q—2UD90I[q (91S ddSqd) uUeNY (€g61) JenbnodAag 2 UOsUsg ‘ds asaysoossiqy 9ua00q soddq (LEZ dASC) ueipuy (pL6L) uosusg ‘ds asaysMoosskqy qusd0lld (909 dGSq) nueny (L861) S109 29 Aopeym aus00l{g (971 ddSd) onuepy (1L61) uosueg LL6I ‘UOsudg voNUDYD asaysMOosskqy (LS6I ‘PIOd QUDDOIJN—9U9007 uvoqque) (TL61) uosusg udp UkA) SISUapopIULY asayiossiqy Issoj (q) onueny AS roded siy} 06 £-916 7 elyensny S (1L61) uosueg IL6I ‘UOSUdg syvsnD adayOosskqy 0S8 €-S09 I uvIpUu] AAS (1L6L) uosusg IL6I ‘UOsUdg vISDd asaysMOOsSskqy OZI p-€ZL Z onueny (1L6L) uosusg IL6] ‘UOSUudg voNUD]ID asayMOossiqy Z6~ €-680 Z aloe d (1L61) uosusg IL6I ‘uOsusg vadnuund asayMossiqy (OLS6L ‘PIO S00 I onueny MA ([L61) uosusgq Uap UkA) SisuapHpluLy ‘Jye asayOosskqy 99/ €-00F € aloe d (1L61) uosusg IL6L ‘Uosusg voIUOdDv! asayiMo0sshqy Areusrajend (e) ANNALS OF THE SOUTH AFRICAN MUSEUM (uw!) a8uei yidoq o3V AN [BOOT JOUDIIJOY so1oods 300 ‘soroods asaysOossdgp JO uonnqiuisiq ‘¢€ AIAVL DEEP-WATER QUATERNARY OSTRACODA 301 Ph inal Fig. 33. Distribution of modern species of Abyssocythere. Re-plotted from data in Benson (1971—dots), with sample site TBD 6851 off south-western Africa shown as an open circle. Genus Ambocythere van den Bold, 19586 A listing of Recent species of this genus that have previously been described from relatively deep water, shows that the genus Ambocythere has a depth range from the inner continental shelf to the Abyssal Zone: Bathyal/Abyssal A. ramosa—lIceland, c. 1000 m (Van den Bold 1965); Newfoundland, 2 938-3 210 m (Benson et al. 1983). Ambocythere cf. A. ramosa—north Atlantic, 2 445m (DSDP site 610, Whatley & Coles 1987). Ambocythere sp. 3057—south-eastern Atlantic, 2 070 m (present study). Ambocythere caudata—Iceland, c. 1 000 m (Van den Bold 1965). Neritic Ambocythere keiji— Venezuela, continental shelf (Van den Bold 1958b) Ambocythere stolonifera—False Bay, South Africa, 30-40 m (Brady 1880) Ambocythere spp. A, B, and C—south-eastern USA, 261-1 034 m (Cronin 1983). 302 ANNALS OF THE SOUTH AFRICAN MUSEUM Including fossil records, the genus has a Neogene range that includes the Caribbean, North and South Atlantic, Indonesia, Japan, Australia, and South Africa (Van den Bold 1965; Benson 1983). Ambocythere sp. 3057 Figs 34A, 35A Illustrated specimen MF-0471, LV, TBD 3355, 2 070 m. Remarks This is a new species, but we have insufficient material (2 juveniles) to warrant a formal description. It was originally noted (but not described or illus- trated) by Boomer (1985: 66) as Ambocythere cf. A. stolonifera, but its closest relative is A. subreticulata van den Bold, 19585, from the Oligocene—Miocene of the Caribbean. Ambocythere sp. 3057 and A. subreticulata differ in the pattern and number of lateral ribs: A. subreticulata has numerous small riblets in the ventro-lateral area. Ambocythere stolonifera (Brady, 1880) from False Bay has a similar lateral outline and main rib arrangement to Ambocythere sp. 3057, but is more elongate and has several small ancillary riblets between the main dorsal and median ribs. The two other known deep-water species differ in the following points: A. ramosa is elongate, with numerous fine longitudinal ribs and a VM-AM-DM rim that is complete; A. caudata is also more elongate and has a marked re-entrant on the postero-ventral margin. Ambocythere sp. 3057 is a rare species that has been recorded only from the Lower Bathyal/Abyssal Zone boundary off south-western Africa (TBD 3355, 2 070 m), where it constitutes 7 per cent of the total ostracod fauna. This value compares to a mean of 8,5 per cent of the total ostracod fauna at the two abyssal sites off Newfoundland for Ambocythere ramosa (Benson et al. 1983). The wide sector of continental margin off south-western Africa that separ- ates the sample site with Ambocythere sp. 3057 from the False Bay population of Ambocythere stolonifera, recorded by Brady (1880), is devoid of any record of the genus, so it is reasonble to assume that the deep- and shallow-water popu- lations of Ambocythere in this region have a long history of genetic isolation. The only known possible record of the genus in the Tertiary of southern Africa is from the Agulhas Bank (TBD 819; Middle Eocene—Middle Oligocene), where Frewin’s (1987) Indeterminate Genus 10 sp. B298 bears some resemblance to A. stolonifera. Genus Echinocythereis Puri, 1954 This genus is distinguished from Henryhowella primarily on the possession of a split anterior MS. On this basis, the taxonomic position of several species DEEP-WATER QUATERNARY OSTRACODA 303 B == SS D = Se —~_——""’ a arc no cae RUNES DOTS) Fig. 34. Outlines of various species of Ambocythere. A. Ambocythere sp. 3057, SAM-PQ-MF-0471, LV, TBD 3355, 2070 m. B. A. subreticulata van den Bold, 1958b, Miocene, Cuba (from Van den Bold 19586, fig. 15). C. A. stolonifera (Brady, 1880), ‘Challenger’ station 140, 15-20 fm, False Bay (from Brady 1880, pl. 21 (fig. 3a)). D. A. ramosa van den Bold, 1965, USNM 342104, 1 400 m off Newfoundland (from Benson et al. 1983, pl. 2 (fig. 1)). E. A. caudata van den Bold, 1965, HVH-7897, holotype, c. 1 000 m off Iceland (from Van den Bold 1965, pl. 1 (fig. 12)). Scale bar = 100 microns. similar to the material available to us has been clarified by the re-illustration of Brady’s (1880) ‘Challenger’ ostracods by Puri & Hulings (1976) (Table 4). Echinocythereis whatleyi sp. nov. Figs 35B-F, 36E-G, I-J ‘Xandarosina’ sp. Boomer, 1985: 64, fig. 7. Echinocythereis echinata (non Sars, 1866) Benson et al., 1983, pl. 2 (fig. 8). Derivation of name This species is named for Professor R. C. Whatley (University College of Wales, Aberystwyth) for his work on deep-water ostracod faunas. 304 ANNALS OF THE SOUTH AFRICAN MUSEUM TABLE 4 Previously-described species allocated to Echinocythereis and Henryhowella on the basis of the central muscle-scar pattern, following the revision of Puri & Hulings (1976). A _ Split anterior muscle scar Genus Echinocythereis Cythere irpex Brady, 1880 Cythereis echinata Sars, 1866 B Single, hooked anterior muscle scar Genus Henryhowella Cythere circumdentata Brady, 1880 Cythere melobesioides Brady, 1869 Cythere ericea Brady, 1880 Cythere dasyderma Brady, 1880 Holotype MF-0472, RV, TBD 3821, 1 525 m. Paratypes MF-0473, LV, TBD 6851, 2 916 m. MF-0474, LV, TBD 3109, 900 m. MF-0475, RV, TBD 3177, 1 000 m. Diagnosis Plump, blind species with broadly rounded AM and PM, and delicate surface ornamentation of small spines and lace-like reticulation. Description The species has a relatively delicate shell with broadly rounded AM and narrower rounded PM, more acuminate in RV. DM is short and straight, VM slightly concave. Antero-dorsal areas are somewhat compressed. Central valve area is plump, with the ‘belly’ of the lateral area just reaching to the VM. Highest point of the valve is over the ACA. No eyespots or ocular sinus. Valve surface is covered in fine spines arranged concentrically, being larger and more numerous postero-ventrally. Surface covered in a very fine, lace-like recticula- tion. AM has two, closely parallel rows of fine spines. In internal view, marginal areas are narrow, with no vestibules. Hinge is weak, modified amphidont, with dorsally open terminal elements in LV. All elements appear smooth. MS consist of four elongate adductors, the second scar being the largest, with two small rounded anterior scars. DEEP-WATER QUATERNARY OSTRACODA 305 Fig. 35. A. Ambocythere sp. 3057, SAM—PQ-MF-0471, LV, TBD 3355, 2 070 m, SEM 3058. B-F. Echinocythereis whatleyi sp. nov. B. SAM—PQ-—MF-0472, holotype, RV, TBD 3821, 1525 m. C. SAM—-PQ-MF-0473, LV internal view, TBD 6851, 2 916m. D. SAM-—PQ- MF-0474, LV, TBD 3109, 900 m. E. SAM—PQ-MF-0475, RV dorsal view, TBD 3177, 1 000 m. F. SAM—-PQ-MF-0472, holotype, RV internal view, TBD 3821, 1 525 m. Scale bars = 100 microns. 306 ANNALS OF THE SOUTH AFRICAN MUSEUM oF 3% Fig. 36. Outlines of Echinocythereis. A, H. E. echinata (Sars, 1866). A. LV. H. Dorsal view. From Sars (1928, pl. 90). Localities unknown. B, D. E. echinata (Sars, 1866). B. DSDP site 611D-12 c.c. Pliocene. D. DSDP site 611D-1 c.c. Quaternary, NE Atlantic. From Whatley & Coles (1987, pl. 5 (figs 7-8)). C. E. echinata (Sars, 1866), specimen 13290, 2859, Quaternary, 2 859m off NW Africa. From Rosenfeld & Bein (1978, pl.1 (fig. 17)). E. E. whatleyi sp. nov. USNM 342109, station 77034-15, Quaternary, 3 000m off New- foundland. From Benson et al. (1983, pl.2 (fig.8)). F. E. whatleyi sp. nov. Holotype, SAM-PQ-MF-0472, TBD 3821, 1525m. G. E. whatleyi sp. nov., SAM-—PQ-MF-0474, TBD 3109, 900m. I. E. whatleyi sp.nov., SAM-PQ-MF-0475, TBD 3177, 1000 m. J. E. whatleyi sp. nov., SAM—PQ-—MF-0473, MS, LV, TBD 6851, 2916m. Scale bars: B-I = 100 microns; J = 30 microns; A, H = scales unknown. DEEP-WATER QUATERNARY OSTRACODA 307 Dimensions (mm) length height MF-0472 0,56 0,39 MF-0473 0,51 0,38 MF-—0474 0,46 O33 Remarks Echinocythereis whatleyi sp. nov. is closely related to Echinocythereis echinata (Sars, 1866) and E. irpex (Brady, 1880). Sars (1928) considered E. irpex to be synonymous with E. echinata, and certainly these two species are more similar in lateral outline to each other than either is to E. whatleyi. Echinocythereis whatleyi differs from both these species in having a plumper, less quadrate outline in lateral view. In particular, it has a shorter DM, with a less pronounced angle in the postero-dorsal region, and a more pronounced, rounded VM ‘belly’. In addition, the degree of compression of the anterior marginal area in E. whatleyi is moderately strong in comparison to E. echinata (see dorsal views), and there is a difference in the shape of the postero-ventral and PM outlines, which are rounded and swept dorsally in E. whatleyi and ‘obtusely blunted’ in E. echinata (Sars 1928). Both E. echinata and EF. irpex have a spinose ventro-lateral keel that is strongest posteriorly. This feature is lacking in E. whatleyi sp. nov. Other species that have some similarities to Echinocythereis whatleyi are: E. jacksonensis (Howe & Pyeatt, 1935, in Howe & Chambers, 1935) from the Middle Eocene to Oligocene of south-eastern USA and the Caribbean (which is more elongate, and is sighted—see Howe & Howe 1973); E. madremaestrae van den Bold, 1988, from the Upper Miocene to Pliocene of the Caribbean (which has a very similar outline and ornamentation, but possesses a prominent eye tubercle); and a specimen referred to E. irpex (Brady) by Sylvester-Bradley & Benson (1971, fig. 24), which has two strong, short, antero-ventral ridges, and a denser overall covering of spines than our species. Van den Bold (1966) described a new species of Echinocythereis from the Miocene of Gabon, but this taxon (E. ecphyma) has a hooked anterior MS, and consequently belongs to another genus. Echinocythereis echinata (Sars) has been widely reported from the North Atlantic and the Mediterranean by numerous authors, in depths ranging from 60-600 m off Norway (Sars 1928: 195) to 4 700m in the Central Atlantic (Tressler 1941). Whatley & Coles (1987) recorded it from Late Miocene to Quaternary sediments in the central North Atlantic. Rosenfeld & Bein (1978) recorded this species between 574 m and 2 859 m off north-western Africa. Brady (1880) recovered Echinocythereis irpex from three sites in the central and South Atlantic, at depths between 900 m and 2 850 m. We have recorded E. whatleyi over a latitudinal range of 19°S—36°S (Fig. 37) and a depth range of 730-2 916 m (Lower Neritic to Abyssal zones— Fig. 38). Most of our records lie within the Upper Bathyal Zone but the species 308 ANNALS OF THE SOUTH AFRICAN MUSEUM km water depth, NERITIC 20 ABYSSAL LOWER BATHYAL 25 latitude’S 30 35 Fig. 37. Latitude and water depth of samples bearing Echinocythereis whatleyi sp. nov. UPPER NERITIC | BATHYAL | | | | | | | | | | | | Echinocythereis whatleyi low at base of salinity minimum zone \ %o PS maximum at AAIW/NADW boundary LOWER BATHYAL 2 water depth, km ABYSSAL Fig. 38. Echinocythereis whatleyi sp. nov. as percentage of total ostracod fauna plotted against water depth. Values are three point running means. DEEP-WATER QUATERNARY OSTRACODA 309 is most abundant at the boundary of the AAIW and NADW masses (c. 1 500 m). There is an abundance low at the top of the Upper Bathyal Zone (i.e. immediately below the Salinity Minimum Zone of the AAIW mass), and the species decreases in abundance through the Lower Bathyal into the Abyssal zones. Benson et al. (1983) recorded the species (as E. echinata) between 2 800 m and 3 000 m off Newfoundland. Off Walvis Bay, Peypouquet & Benson (1980) noted Echinocythereis between 974 m and 2 864 m, although they did not differentiate species. There are too few data points to make a detailed comparison with Figure 38, but their highest value does lie in the vicinity of 1500 _m. In contrast to our profile, however, the Walvis Bay data indicate a sharp rise in abundance in the vicinity of 3 km. Genus Trachyleberis Brady, 1898 Trachyleberis sp. 3017 Fig. 42A-B ?‘Thalassocythere’ sp. B Cronin, 1983, pl. 4 (figs B, E, G). Illustrated specimens MF-0476, RV, TBD 6851, 2 916 m. MF-0477, LV, TBD 6851, 2 916 m. Remarks The type species of Trachyleberis Brady, 1898, is Cythere scabrocuneata Brady, 1880, and the holotype is lost. Puri & Hulings (1976, pl. 26 (figs 6, 8)) illustrated a topotype from the Inland Sea of Japan. Our species has less robust spines, no AM and PM rims, and lacks the antero-dorsal ocular ridge, and is not conspecific. In general aspect, however, it conforms with the genus. Benson (e.g. Benson 1977, in Benson & Peypouquet 1983) seems to use the nomen nudum ‘Thalassocythere’ for deep-water Trachyleberis-like species and, judging from Benson (1977), when he lists (but does not illustrate) ‘Thalasso- cythere’ acanthoderma from the Miocene at DSDP sites 356 and 357, it is implied that Cythere acanthoderma Brady, 1880, is the comparative species for this group. An illustration of ‘Thalassocythere’ acanthoderma (Brady, 1880) from 3 000 m off Newfoundland (Benson et al. 1983, pl. 2 (fig. 9)), bears a close resemblance to Brady’s (1880, pl. 18 (figs S5a—e)) original illustrations, but differs from the lectotype of Cythere acanthoderma figured by Puri & Hulings (1976), which is probably a juvenile, in possessing a pointed PM outline; the lectotype has a rounded outline. Coles & Whatley (1989) have formalized the taxonomy of this group by erecting the genus Legitimocythere, with the type species Cythere acanthoderma Brady, 1880. Our species differs from L. acantho- derma (Brady, 1880) (as in Brady’s original illustrations) in lacking an AM rim, 310 ANNALS OF THE SOUTH AFRICAN MUSEUM in having a concave VM outline in lateral view, and having fewer and less massive and ‘ragged’ spines. It may be conspecific with ?‘Thalassocythere’ sp. B (Cronin, 1983), although our species has a less dense pattern of spines. R. C. Whatley (pers. comm. 1988) has suggested that our specimens are probably juveniles of Legitimocythere acanthoderma (Brady, 1880). Rosenfeld & Bein (1978) did not record any trachyleberid-like species from off north-western Africa. In the south-eastern Atlantic, Trachyleberis sp. 3017 is confined to the Abyssal Zone, where we have recorded it from two sites at 2 916 m and 4 736 m water depth. Here it consitutes 1 per cent and 33 per cent of the total ostracod fauna, respectively. Table 5 summarizes the geographical and depth ranges of the documented deep-sea species of Trachyleberis. TABLE 5 Distribution of deep-sea species of the genus Trachyleberis (including the nomen nudum ‘Thalassocythere”™). Species Reference Age Location Baty Trachyleberis Brady (1880) Quaternary Atlantic, Indian, 1 600-5 500 acanthoderma Pacific Benson et al. (1983) Quaternary Newfoundland 2 600-3 210 Ducasse & Pey- Miocene— NE Atlantic (DSDP ? pouquet (1979) Quaternary 403, 405) Whatley & Coles Miocene- N Atlantic(DSDP 2 445-3 884 (1987) Quaternary 606-611) Trachyleberis sp.3017 this paper Quaternary SE Atlantic 2 916-4 736 Trachyleberis sp. B Cronin (1983) Quaternary SE USA 341-739 * ‘Thalassocythere’ = Legitimocythere Coles & Whatley, 1989. Genus Henryhowella Puri, 1957 Henryhowella has been widely reported in the literature from modern and Cenozoic deep-water sediments. The type species for Henryhowella is Cythere evax Ulrich & Bassler, 1904, and Puri’s original diagnosis (1957 new name = Howella Puri, 1956), together with the type description of Cythereis garretti Howe & Mcguirt, 1935 (in Howe & graduate students 1935) (which is the type species for Echinocythereis Puri, 1954), distinguishes the genus Henryhowella from its close relative Echinocythereis on two criteria: 1. the possession of ‘three well-developed longitudinal rows of spines in the posterior half of the carapace’; 2. having a single, hooked anterior MS (Echinocythereis has two small rounded anterior MS). DEEP-WATER QUATERNARY OSTRACODA ot Conventional wisdom is to ascribe certain spinose taxa with a single hooked anterior MS to Henryhowella, irrespective of whether they possess a triplicate posterior ornamentation (e.g. Uffenorde 1981; Whatley & Coles 1987; Steineck et al. 1988). Those that do possess the latter feature are invariably placed in Henryhowella asperrima (Reuss, 1850) (e.g. Van den Bold 1960; Cronin 1983; Whatley & Coles 1987). This has the effect of modifying the generic concept and blurring its definition to the point of causing confusion with genera such as Rocaleberis Bertels, 1969. It also reduces the only substantive distinguishing feature between Echinocythereis and Henryhowella to their different MS patterns. As a consequence, Cythere evax Ulrich & Bassler, 1904, should be considered a synonym of Cypridina asperrima Reuss, 1850 (e.g. Van den Bold 1957b, 1960), which becomes the type species of the genus Henryhowella. A problem here is that in his type description, Reuss (1850: 74), specifically stated that Cypridina asperrima has two posterior ridges and a median longitudinal furrow in the posterior part of the valve. Keij (1957) and Van den Bold (1960) have examined topotypes for Reuss’s original material and both assumed that Reuss misidentified the valve architecture, although there has been no modern re-illustration of his topotypic material, nor the erection of a lectotype, if indeed the holotype is lost. Henryhowella melobesioides (Brady, 1869) Figs 42C-F, 43A-F, 44A-D, 47A Cythere melobesioides Brady, 1869: 162, pl. 12 (figs 10-11); 1880: 108, pl. 18 (figs le-g). Puri & Hulings, 1976, pl. 25 (figs 1-2). Cythere nodulifera Brady, 1869: 163, pl. 19 (figs 24-25). Henryhowella sp. Keeler, 1981: 162-163, pl. 9 (fig. 14). Henryhowella sp. Boomer, 1985: 25-27, pl. 1 (figs 6-8, 18). non Henryhowella sp. Boomer, 1985: 25-27, pl. 3 (figs 38-39). non Cythere melobesioides Brady, 1869. Brady, 1880, pl. 18 (figs 1a—d). Illustrated specimens MF-0478, RV, TBD 311, 184 m. MF-0479, LV, TBD 311, 184 m. MF-0480, LV, TBD 3561, 655 m. MF-0481, RV, TBD 3561, 655 m. MF-0482, LV, TBD 3704, 941 m MF-0483, RV, TBD 3704, 941 m. MF-0484, RV, TBD 3383, 990 m. MF-0485, RV, TBD 3383, 990 m. MF-0486, LV, TBD 3383, 990 m. MF-0487, LV, TBD 6851, 2 916 m. MF-0488, LV, TBD 6851, 2 916 m. MF-0489, LV, TBD 3383, 990 m. 312 ANNALS OF THE SOUTH AFRICAN MUSEUM Remarks Accepting the current interpretation of non-plicate morphology, this species can be accommodated in the genus Henryhowella, and there are four species to which our material can be compared: Henryhowella asperrima (Reuss, 1850), Henryhowella dasyderma (Brady, 1880), Henryhowella digitalis Levinson, 1974 (in LeRoy & Levinson 1974), and Henryhowella melobesioides (Brady, 1869). As discussed above, tri-plicate species (i.e. the ‘typical’ Henryhowella mor- phology) are generally automatically assumed to belong within H. asperrima (Reuss). This has led to the grouping of taxa whose synonomy appears doubtful but for which resolution requires redefinition of the types. To gauge the problem compare the following: Keij (1957, pl. 12 (figs 1-2)); Van den Bold (1960, pl. 4 (fig. 10)); Colalongo (1965, pl. 11 (figs 3-8)); LeRoy & Levinson (1974, pl. 12 (fig. 1)); Rosenfeld & Bein (1978, pl. 1 (fig. 23)); Cronin (1983, pl. 4 (fig. F)); Whatley & Coles (1987, pl. 5 (figs 9-11)). None of our material fits comfortably into the taxon currently interpreted as Henryhowella asperrima, even allowing for the range of morphological variation accepted by other authors. All the specimens that we have recovered have very weak or no lineation of spines in the posterior part of the valves; there are never three well-defined rows or ridges. The closest approximation to a tri-plicate morphology is shown in Figure 43C, but such are isolated examples in populations where there is grada- tion to the completely non-plicate state. As observed by Steinick et al. (1988) for their deep-sea Pacific populations, we feel unable to place specimens into a poss- ibly polytaxonomic category. Tertiary species referred to H. asperrima have been recorded from the vicinity of south-western Africa by Van den Bold (1966—Lower Miocene of Gabon) and Frewin (1987—Middle—Upper Eocene of the Agulhas Bank). Dingle (1976) recorded a form very close to Frewin’s species as Henryhowella sp. This ranges from Lower Eocene to Upper Oligocene in the J(c)—1 borehole on the continental shelf off Natal (south-eastern Africa), and is probably the same species as a worn specimen referred to Indet. sp. 2314 from Upper Ceno- manian strata in the same borehole (Dingle 1985). The latter record was possibly a downhole contaminant from overlying Tertiary strata. These local records are all probably of the same species (whether or not they can ultimately be referred to Reuss’s species is an open question), which seems to have a range Lower Eocene to Miocene. Henryhowella digitalis Levinson, 1974 (in LeRoy & Levinson 1974), was erected as a variety of H. asperrima to accommodate specimens with a reticu- late/spinose ornamentation. LeRoy & Levinson (1974) speculated that this was a ‘deep-water variant’ of H. asperrima. Some of our specimens do show a degree of reticulation, but it is not a consistently developed feature. Henryhowella dasyderma (Brady, 1880) is a non-plicate species. Brady (1880) included several morphotypes in his original description, one of which (1880, pl. 17 (fig. 4e-f)) has a sharp ventro-lateral ridge. The lectotype desig- nated by Puri & Hulings (1976, pl. 11 (figs 10—11)) from ‘Challenger’ site 296 DEEP-WATER QUATERNARY OSTRACODA 313 (south-eastern Pacific) is very similar to our material, but we have not assigned our specimens to it, because Brady himself (1880) believed that the populations that he recovered from off the south-western Cape were closer to Cythere melobesioides Brady, 1869. However, he did remark (1880: 105) that Cythere dasyderma was a deep-water taxon that occurs world-wide. It has been subsequently recorded by Rosenfeld & Bein (1978) from 1 029-2 480 m off north-western Africa, and Whatley & Coles (1987) from early Pliocene to Quaternary of the North Atlantic. Henryhowella melobesioides (Brady, 1869) was recorded by Brady (1880) from ‘Challenger’ site 142 off the Cape Peninsula in 150 fm (274m) water depth. The type species was from Mauritius and Brady (1880) also identified it from Australia, but these (pl. 18 (fig. la—d)) appear not to be conspecific. Our material is conspecific with the specimen illustrated by Puri & Hulings (1976, pl. 25 (figs 1-2)), but there is some variation in ornamentation within the popu- lation, particularly in the prominence of a weak ridge that lies almost parallel to the postero-ventral margin, a feature noted in the illustration in Brady (1880, pl. 18 (fig. 1g)). The lateral surface spines are arranged concentrically in the anterior part of the valve, and randomly or with a weak elongation that tends to converge into a chevron with its apex towards the PM. No consistent morpholo- gical variations are observed between specimens from different depths, although individuals from shallower water tend to have the more nodose spines described by Brady (1880). MS patterns are identical when observed over a range 655- 2 916 m (no good views were obtained from shallow-water examples). It is poss- ible that we have more than one species in our populations but, until the uncer- tainty of the taxonomic position of forms currently referred to H. asperrima (Reuss) and their relationship to H. dasyderma (Brady) have been clarified, there seems little point in attempting to subdivide the various morphotypes of our H. melobesioides (Brady) population. Off south-western Africa, Henryhowella melobesioides (Brady, 1869) has been found over a latitudinal range 19°S~36°S (Fig. 39), and a depth range of 100-2 916 m (Fig. 40). In the Neritic Zone, it is confined to water greater than 600 m north of 29°S, where we suspect it cannot survive in either low oxygen- ated water (associated with the upwelling cells north of 25°S) or influxes of less saline and/or suspensate-rich water from the Orange River on the continental shelf between 25°S and 29°S. The upwelling cells on the shelf south of 29°S do not sustain sufficiently high productivity in the water column to result in low oxygen layers. Figure 40 shows that on the continental shelf (i.e. less than 300 m), FH. melobesioides is a minor element in the ostracod population (generally <10%). Its abundance steadily increases with water depth, so that overall within the Neritic Zone it constitutes a mean of 31 per cent (range 0,1-100%). In the Upper Bathyal Zone it is the dominant ostracod taxon, generally forming more than 60 per cent of the total population (mean of 73%, range 19-100%) but, between about 800m and 1 000m, which includes the Neritic/Bathyal 314 ANNALS OF THE SOUTH AFRICAN MUSEUM " oO —-— = modern specimens ABYSSAL ere Ae ih es eat gente olin) realm eenay ere ny oy a (an ee 2 E x = LOWER BATHYAL fe oO ret =a oa. ae iw oO pe eee a ee See ce) 0 ee a) a) oO bs oes 5 / oO 2 of of ay o UPPER BATHYAL = | ww ees oO = oO --* " 1 o wane, pass | a Eg areae NA — Se oO a aa o-—#X\ — — = DO - ao ie “oO \ Z oO oO Ff B \ oO ea Oo \\ | O \ NERITIC C8 Ja: eo ce RS: Go. L OW Ou APE) WN _9 ao So penn ay V3 : eae 20 25 30 35 latitude °S Fig. 39. Latitude and water depth of samples bearing Henryhowella melobesioides (Brady, 1869). Note the absence of this species from the low O2, and low salinity/high suspensate conti- melobesioides Henr yhowella %o Fig. 40. Henryhowella melobesioides (Brady, 1869) .as percentage of total ostracod fauna 100 = 50 nental shelves off Walvis Bay/Liideritz, and Orange River, respectively. | | NERITIC ! BATHYAL | ABYSSAL | UPPER | LOWER | decrease in | i | abundance ay | fo 7 | at fh | Av We | Vi | = h i decrease in | co / |) oh ane | abundance at yy es AAIW/ NADW ip boundary i" | 1 | | | | 1 : re t = = | | Aa | ir ui | it | | | | Lind A Wise | | @ | al om | | | | he | | | Ua rh | | | n= 1 Gee low abundance | | ae mn < on shelf | | | O 1 2, water depth, km plotted against water depth. Values are three point running means. DEEP-WATER QUATERNARY OSTRACODA 315 boundary, there are large fluctuations in its abundance. Similarly, on approaching the Upper/Lower Bathyal Zone boundary (i.e. the shear zone between the AAIW and NADW masses), there is a large fall in abundance, that continues through the Lower Bathyal Zone (mean 39%, range 28-71%) into the Abyssal Zone (mean 20%, range 6-34%). Clearly, this species has a wide toler- ance of temperature and salinity, but it is relatively less successful in unstable, or mixed zones. In this respect it shows a distribution pattern that is the reverse of Buntonia rosenfeldi sp. nov. Modern specimens were identified on two criteria: shell transparency and good preservation of spines. They were encountered in two distinct areas. North of 26°S the entire distribution zone contains modern valves (i.e. in both the Lower Neritic and Upper Bathyal Zones), whereas farther south the modern population is confined between 29°S and 35°S and almost solely to the Neritic Zone (Fig. 39). Peypouquet & Benson (1980) recorded Henryhowella from their traverse off Walvis Bay. There are too few data points to make a meaningful comparison with our results, but a plot of the abundance of Henryhowella (Fig. 41) indicates similar trends to Figure 40. We do not Know if the species they recorded was H. melobesioides but suspect so, because the latitudinal ranges of the data sets overlap. NERITIC BATHYAL ABYSSAL UPPER LOWER or (2) Henryhowella spp. %. | | | | | | | | | | | | | | | water depth, km Fig. 41. Henryhowella species as percentage of total ostracod fauna plotted against water depth for a profile off Walvis Bay. Data computed from Peypouquet & Benson (1980). 316 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 42. A-B. Trachyleberis sp. 3017, TBD 6851, 2 916m. A. SAM—PO-MF-0476, RV, SEM 3017. B. SAM-PQ-MF-0477, LV, SEM 3024. C-F. Henryhowella melobesioides (Brady, 1869). C. SAM—PQ-MF-0478, RV, TBD 311, 184m, SEM 2625. D. SAM—-PQ- MF—0479, LV, TBD 311, 184 m, SEM 2626. E. SAM—PQ-MF-0480, LV, TBD 3561, 655 m, SEM 2591. F. SAM-PQ-MF-0481, RV internal view, TBD 3561, 655 m, SEM 2594. Scale bars = 100 microns. DEEP-WATER QUATERNARY OSTRACODA Bik Fig. 43. A-F. Henryhowella melobesioides (Brady, 1869). A. SAM—-PQ-MF-0482, LV, TBD 3704, 941 m, SEM 2630. B. SAM—PQ-MEF-0483, RV internal view, TBD 3704, 941 m, SEM 2632. C. SAM-PQ-MF-0484, RV, TBD 3383, 990m, SEM 2552. D. SAM-— PQ-MF-—0485, RV internal view, TBD 3383, 990 m, SEM 2561. E. SAM-POQ-MEF-0486, LV, TBD 3383, 990m, SEM 2551. F. SAM-—POQ-MF-0487, LV, TBD 6851, 2 916 m, SEM 2636. Scale bars = 100 microns. 318 ANNALS OF THE SOUTH AFRICAN MUSEUM —- Ho om ra) O VE Ja rs eee ae fe) C D o Fig. 44. MS of Henryhowella melobesioides (Brady, 1869). A. SAM—PQ-MF-0481, RV, TBD 3851, 655 m, SEM 2621. B. SAM—PQ-MF-0483, RV, TBD 3704, 941 m, SEM 2634. C. SAM-PQ-MF-0489, LV, TBD 3383, 990 m, SEM 2556. D. SAM—-PQ-MF-0488, LV, TBD 6851, 2 916 m, SEM 2639. Open dots are normal pores. Scale bars are 100 microns. Subfamily Pennyellinae Neale, 1975 Neale (1975) established this taxonomic category to accommodate certain blind, reticulate trachyleberids, and identified three genera that belonged here: Pennyella Neale, 1974, Santonian (Western Australia—Neale 1974) to Maastrichtian (north-western Pacific—Swain 1973); Agulhasina Dingle, 1971, Maastrichtian (southern Africa); Agrenocythere Benson, 1972, Eocene (Atlan- tic) to Recent (pandemic). We now add a fourth: Rugocythereis gen. nov., Miocene to Quaternary (pandemic—Brady 1880; Whatley & Coles 1987). Rugocythereis gen. nov. Derivation of name Latin rugosa—rough, reference to coarse texture of surface ornamentation. Type species ‘Oxycythereis’ horridus Whatley & Coles, 1987. Diagnosis Blind, elongate-sub-quadrate trachyleberid with coarsely reticulate and/or stout spinose ornamentation. Strongly sexually dimorphic with males signifi- cantly more elongate than females. Females typically possess an antero-dorsal marginal frill. Valves are inflated postero-ventrally and compressed in the AM DEEP-WATER QUATERNARY OSTRACODA 319 areas, with strong AM and PM rims. Typically there is a prominent cleft between the PM rim and the area of postero-ventral inflation. Hinge is modified amphidont, with weak ATE in RV. Remarks Rugocythereis gen. nov. is a pandemic genus that includes several taxa pre- viously placed in the nomen nudum ‘Oxycythereis’ Benson, 1974. It is closest to Pennyella Neale, 1974, but the two genera differ on the following features: 1. Rugocythereis lacks the prominent vertical postero-dorsal ridge. 2. Rugocythereis lacks the prominent ventro-lateral ridge that terminates in a posterior elevation. 3. Pennyella has a more triangular-shaped PM outline, with a distinctive postero-dorsal concavity. 4. Females of Rugocythereis have a distinctive antero-dorsal marginal frill. 5. The hinges are somewhat different: Pennyella has a peg-like ATE in the RV, with a post-adjacent rounded socket, whereas in Rugocythereis the RV ATE is low and elongate, has a narrow grove on its dorsal side, and has no post- adjacent rounded socket. The two genera have very similar MS patterns, which include a ventrally deflected second adductor and two very close, almost fused, third and fourth adductors. Rugocythereis gen. nov. is a deep-water genus, for which four species have been formally described: R. horridus (Whatley & Coles, 1987), R. dorsoserrata (Brady, 1880, emend. Puri & Hulings, 1976), and two species originally placed in Pennyella (P. fortedimorphica Coles & Whatley, 1989, and P. praedorsoser- rata Coles and Whatley, 1989). In addition, Whatley & Coles (1987) recorded two possible species in open nomenclature (‘Oxycythereis’ sp. 1 and ‘“Oxycy- thereis’ sp. 2). Several other modern and fossil species of Rugocythereis may be included in the nomen nudum ‘Oxycythereis’ (see Kempf 1986a, 1986b), but their status is uncertain. This is because the references have lacked either illustrations or descriptions or both, which introduces confusion over the comparability of R. dorsoserrata (Brady, 1880) and the validity of additional specific taxa (e.g. Benson 1977; Peypouquet & Benson 1980; Benson et al. 1983). Table 6 lists known and possible taxa that belong in Rugocythereis gen. nov. Given the limitations of the uncertain taxa, the genus may have an age range of Eocene to Recent and a water depth range for modern species of 730-3 526 m. Guernet (1985, pl. 4 (fig. 9)) described a species under Wichmanella? cf. W.? reticulata, which has a similar valve outline and ornamentation to R. horridus. This taxon was recorded in Lower Eocene strata at DSDP site 245 in the south-western Indian Ocean. No species referable to Rugocythereis (including the nomen nudum taxa ‘Oxycythereis’) was reported from Quaternary deep-water sites off the south- eastern USA (Cronin 1983) or north-western Africa (Rosenfeld & Bein 1978). 320 ANNALS OF THE SOUTH AFRICAN MUSEUM TABLE 6 Geographical and water depth distribution of species of Rugocythereis gen. nov. Quaternary Species Reference Age Location depth (m) Rugocythereis dorso- Brady (1880) Quaternary S Atlantic 2 850 serrata (Brady, 1880) Rugocythereis horridus Whatley & Coles Miocene- N Atlantic 2 445-3 526 (1987) Quaternary Pliocene— SW Pacific Quaternary Miocene SW Indian Ducasse & Pey- Quaternary NE Atlantic 3 000 pouquet (1979) (DSDP 405) this paper Quaternary SE Atlantic 750—7 NG Rugocythereis? spp. Benson (1974) Eocene— W Indian (DSDP v Pliocene D338) Benson et al. Quaternary NW Atlantic 3 210 (1983) Benson (1977) Eocene— S Atlantic (DSDP u Pleistocene 350, 357) Peypouquet & Quaternary SE Atlantic 2 094—2 644 Benson (1980) Rugocythereis sp. 1 Whatley & Coles | Miocene N Atlantic (1987) Rugocythereis sp. 2 Whatley & Coles _—_ Pliocene N Atlantic (1987) Rugocythereis horridus (Whatley & Coles, 1987) Figs 47B-F, 48 ‘Oxycythereis’ horridus Whatley & Coles, 1987: 76-78, pl. 5 (figs 18-22). Henryhowella sp. Boomer, 1985: 25-27, pl. 3 (figs 38-39). non Henryhowella sp. Boomer, 1985: 25-27, pl. 1 (figs 6-8). Illustrated specimens MF-0490, RV, TBD 2677, 1 662 m. MF-0491, LV, TBD 6851, 2 916 m. MF-—0492, RV, TBD 6851, 2 916 m. MF-—0493, RV, TBD 6851, 2 916 m. Remarks Rugocythereis horridus (Whatley & Coles) was originally described from Miocene to Quaternary strata at DSDP sites from Leg 94 in the north central Atlantic, and was also recorded from Pliocene to Quaternary in the south- DEEP-WATER QUATERNARY OSTRACODA By) western Pacific, and Miocene from the Indian Ocean (Whatley & Coles 1987). We record it in the south-eastern Atlantic between 29°S and 36°S, in water depths of 730-2 916 m (Fig. 45). The species was found in only 16 per cent of the bathyal and abyssal sites, and Figure 46 shows that R. horridus ranges from lower neritic to abyssal depths and that it is most abundant in the Upper Bathyal Zone. In the lower part of the Neritic Zone, abundances are low (mean: 4% total ostracod fauna, range nil), but in the Bathyal Zone the species is relatively abundant (mean: 24%, range 4-100%). There appears to be a decrease in abun- dance with increasing depth through the Bathyal Zone, and the single site in the Abyssal Zone (2 916 m) has the lowest abundance of our data set (2%). The species recorded by Ducasse & Peypouquet (1979) as ‘Oxycythereis’ dorsoserata [sic] from DSDP site 405 in the Rockall Basin (north-eastern Atlan- tic) probably belongs in R. horridus. Rugocythereis horridus is closely related to R. dorsoserrata (Brady, 1880), which has been re-illustrated by Puri & Hulings (1976), who selected a lectotype from Brady’s original ‘Challenger’ material. Positive identifications of Rugo- cythereis dorsoserrata (Brady, 1880) have been: Brady’s original record from 2 605 m (1 425 fm) north of Tristan da Cunha in the South Atlantic; Whatley & Coles’s (1987) record from Miocene—Quaternary strata from DSDP Leg 94 sites in the North Atlantic; and two specimens identified as Cythere suhmi Brady, 1880, by Ducasse & Peypouquet (1979, pl. 2 (figs 5—6)) from the late Pliocene @ = modern specimens ABYSSAL LOWER BATHYAL water depth, km BATHYAL NERITIC 30 35 latitude °S Fig. 45. Latitude and water depth of samples bearing Rugocythereis horridus (Whatley & Coles, 1987). 322 ANNALS OF THE SOUTH AFRICAN MUSEUM 100 % NERITIC ABYSSAL % Rugocythereis rogersi 1 2 3 water depth, km Fig. 46. Rugocythereis horridus (Whatley & Coles, 1987) as percentage of total ostracod fauna plotted against water depth. Values have not been smoothed. of DSDP site 403 on Hatton Bank (north-eastern Atlantic), which are probably conspecific. These records suggest an age range Miocene to Quaternary for R. dorsoserrata. Benson et al. (1983) recorded ‘Oxycythereis’ dorsoserrata (Brady, 1880) (which they equated with ‘O.’ dorsoserrata of Ducasse & Peypouqueti 1979) from a sample at 3 210 m off Newfoundland, but did not illustrate the specimen. Peypouquet & Benson (1980) recorded ‘Oxycythereis’ from their traverses in the Cape and Angola basins. These may refer to species of Rugocythereis, although there were no accompanying illustrations. The data are too sparse to plot, but abundances at the three sites were: Walvis transect—2 094 m, 0,2 per cent; 2 117m, 1 per cent; Angola Basin—2 644 m, 0,7 per cent. These are lower values than we have determined for similar depths, but indicate that the taxon is rare. Subfamily Bradleyinae Benson, 1972 Genus Poseidonamicus Benson, 1972 Whatley et al. (1983) have expressed reservations about the suprageneric placement of Poseidonamicus, but we will follow Benson’s original classification for the present. DEEP-WATER QUATERNARY OSTRACODA 323 A . Ny ¥ : < & : ow ae : * Fig. 47. A. Henryhowella melobesioides (Brady, 1869), SAM—PQ-MF-0488, LV internal view, TBD 6851, 2 916m, SEM 2638. B-F. Rugocythereis horridus (Whatley & Coles, 1987). B. SAM-PQ-MF-0490, RV, TBD 2677, 1662 m, SEM 3030. C-F. TBD 6851, 2 916 m. C. SAM-PQ-MF-0491, LV, SEM 3008. D. SAM-—PQ-MF-0492, RV, SEM 3009. E. SAM- PQ-—MF-0493, RV internal view, SEM 3012. F. SAM—PQ-MF-0493, RV ATE and PTE, SEM 3013 and 3014. Scale bars = 100 microns. 324 ANNALS OF THE SOUTH AFRICAN MUSEUM O f- So 90 Fig. 48. MS of Rugocythereis horridus (Whatley & Coles,:1987). SAM—PQ-—MF-0493, RV, TBD 6851, 2 916 m, SEM 3015. Open dots are normal pores. Scale bar = 100 microns. 30 NERITIC BATHYAL ABYSSAL be) je) P. major Poseidonamicus spp. — oO %o water depth, km Fig. 49. Latitude and water depth of samples bearing Poseidonamicus species. The populations of P. major Benson, 1972 (deep-water species), and P. panopsus Whatley & Dingle, 1989 (shallow-water species) are outlined. They are separated by 1500 m of the water column, which contains several effective physico-chemical barriers. DEEP-WATER QUATERNARY OSTRACODA 325 Whatley (1985) discussed the colonization of bathyal and abyssal environ- ments by species of Poseidonamicus, and he concluded that since early Palaeogene time the genus has been confined to water depths greater than 1 000 m. However, recent work on the ostracod fauna offshore south-western Africa has located two species of the genus, one of which is blind and lives at abyssal depths (P. major Benson, 1972), and one of which is sighted and lives on the outer continental shelf/upper slope (P. panopsus Whatley & Dingle, 1989). Figure 49 shows the distribution and abundances of these two species in the south-eastern Cape Basin. It is significant that the habitats of the two species, which are both extant, are separated by 1500 m of water that includes the Salinity Minimum Zone of the AAIW and the boundary between the AAIW and NADW water masses. These must be very effective barriers in maintaining the identity of the two species. Poseidonamicus major Benson, 1972 Figs 50A—F, 51A-B Poseidonamicus major Benson, 1972: 52-53, pl. 8 (fig. 5), pl. 10 (figs 1-6), text-figs 20, 22. Illustrated specimens MF-0494, RV, TBD 6851, 2 916 m. MF-0495, LV, TBD 6851, 2 916 m. MF-0496, LV, TBD 6851, 2 916 m. MF-0497, RV, TBD 6851, 2 916 m. Remarks Our material is identical to the holotype (USNM 174335), except for its MS pattern, which in the Cape Basin specimens show an indented ventral scar in the anterior pair, and two ‘dog’s bone’-shaped central scars in the adductors. Poseidonamicus major differs from the shelf/upper slope species of the genus that has recently been discovered living in neritic depths (120-545 m) off south-western Africa (P. panopsus Whatley & Dingle, 1989) in several respects. The new species has a prominent eye tubercle, is generally more quadrate in lateral outline, and lacks the strong, curved anterior dorso-lateral ridge that is a characteristic feature of P. major. Their MS are very similar, although in P. panopsus the lower anterior scar is more V-shaped than in our specimens of P. major. Frewin (1987) recorded an Eocene species from the Agulhas Bank as ?Poseidonamicus sp. A126. This is a sighted form that has some architectural features similar to the type species of Poseidonamicus, but whose overall morphology suggests that it belongs to an undescribed taxon. The holotype of Poseidonamicus major is from the Mozambique Channel (south-west of Europa Island) at a depth of 2 995 m, where Benson (1972) recorded a bottom water temperature of 1,6°C. Off south-western Africa, we 326 ANNALS OF THE SOUTH AFRICAN MUSEUM SS SS Ss ee Se Fig. 50. A-F. Poseidonamicus major Benson, 1972, TBD 6851, 2916m. A. SAM-—PQ- MF-0494, RV, SEM 2929. B. SAM—POQ-MF-0495, LV, SEM 2926. C. SAM-—PO-MF-0496, LV internal view, SEM 2931. D. SAM-—PC-—MF-0497, RV internal view, SEM 2933. E. SAM-—PQ-MF-0496, LV MS, SEM 2932. F. SAM—PQ-MF-0497, RV MS, SEM 2935. Scale bars = 100 microns. DEEP-WATER QUATERNARY OSTRACODA B77) A B Fig. 51. MS of Poseidonamicus major Benson, 1972. A. SAM—PQ-MF-0497, RV, TBD 6851, 2 916 m, SEM 2935. B. USNM 174354, RV, IIOE 363B, 2 995 m, Mozambique Channel (from Benson 1972, fig. 22B). Scale bars = 100 microns. BAK Fig. 52. A-B, D. Indet. sp. 62, TBD 3870, 1 026m. A. SAM-PQ-MF-0498, LV. B, D. SAM-PO-MF-0499, RV. B. Internal view. D. MS. C. Indet. sp. 23, SAM-—PQ-—MF-0500, TBD 3174, 1 050 m. Scale bars: A—C = 100 microns; D = 30 microns. 328 ANNALS OF THE SOUTH AFRICAN MUSEUM found the species at two sites within the Abyssal Zone: TBD 3355 (2 070 m) and TBD 6851 (2 916 m) (Fig. 49). In the former, which lies at the top of the Abyssal Zone, P. major accounts for 14 per cent of the total ostracod fauna and is the third most abundant taxon, but all the specimens were relict. At TBD 6851 (2 916 m), this species is the second most abundant taxon (29%), 6 per cent of which were modern specimens (1,7% of total ostracods). In his original discussion of Poseidonamicus, Benson (1972) erected four new species, all of which are extant: P. major, P. minor, P. nudus, and P. pintoi. He did not quote depth ranges for individual species, but reference to his table 2 shows the ranges that he recorded for the modern genus in various oceans, as follows: Atlantic 1 227—5 020 m; Indian 1 190-3 995 m; and Pacific 2 089-3 292 m. Whatley (1985) has discussed the evolutionary history of the genus and concluded that it originated in the south-western Pacific, probably in the Palaeocene, from a species such as Hermanites sagitta Bate, 1972, which had a continental-shelf habitat. During a phase of rapid mutation in the early Tertiary, the earliest forms of Poseidonamicus migrated into bathyal and later (in the Miocene) into abyssal habitats. Whatley (1985) did not have sufficient data to postulate the evolution of P. major, which is morphologically distinct from the other Tertiary forms. He recorded it in the Miocene of the south-western Pacific, but Benson (1972: 53) stated that it occurs in the Oligocene of DSDP site 117 (north-eastern Atlantic). Benson & Peypouquet (1983) recorded several species of the genus from Neogene horizons at DSDP sites in the western South Atlantic (sites 516, 517, and 518), including two new species (P. miocenicus and P. riograndensis), but did not mention P. major specifically. Whatley & Coles (1987) recorded Poseidonamicus sp. cf. P. major and Poseidonamicus sp. cf. P. pintoi as ranging late Miocene to Quaternary at DSDP Leg 94 sites in the North Atlantic. As far as we can gauge from the literature, P. major has the following age ranges: south-western Pacific—Miocene to Recent (bathyal—abyssal); south- western Atlantic—Miocene to ?Recent; north-eastern Atlantic— Oligocene; south-western Indian—Recent (abyssal); and, south-eastern Atlantic— Quater- nary (abyssal); and North Atlantic, as Poseidonamicus cf. P. major—late Miocene to Quaternary. Indeterminate taxa Indet. sp. 62 Fig. 52A-B, D Two living valves were recovered from sample TBD 3870 at a water depth of 1 026 m. This ovate species has a short, straight DM and prominent anterior and posterior cardinal angles. It has a merodont hinge and a MS pattern consist- ing of a small rounded anterior scar and three elongate adductors above a smaller oval fourth adductor. It may be related to Krithe (R. C. Whatley, pers. comm. 1988). DEEP-WATER QUATERNARY OSTRACODA 329 Indet sp. 23 Fig. 52€ Two fragments of the anterior ends of RV were recovered from sample TBD 3174 at a water depth of 1 050 m. The species has a broadly rounded AM with a narrow, spinose rim. There is a prominent spinose ridge in the antero- ventral region, but otherwise the anterior part of the valve surface is smooth. It may belong in Bathycythere (R. C. Whatley, pers. comm. 1988). DISCUSSION A total of 1 023 autochthonous specimens were recovered from 45 sediment samples in water depths greater than 900 m off south-western Africa (only one sample of the original 46 in our data set was barren of ostracods). These represent 31 species, of which four are neritic taxa whose ranges extend 100 m or less into depths that we equate with the Bathyal Zone. The 27 genuinely deep-water species are assigned to 16 genera and two indeterminate categories. DEPTH RANGES AND FAUNAL ZONES Figure 53 shows the depth ranges of all the species we have isolated, and the inset plots the turnover rate (appearances/disappearances) against depth (summed for 100 m intervals). These data indicate that important changes in composition of the ostracod faunas occur within three depth zones: 900-1 100 m; 1 300-1 700 m; and 2 000-2 100 m. (The fourth peak on the right-hand-side of the inset is an artefact caused by the small number of samples available in water deeper than 2 100m.) Figures 54-56 and Table 7 show further details of these faunal changes and allow us to isolate with more precision the depths at which they occur. Between 900 m and 1 050 m, ten species appear or disappear, with a further seven species passing through the zone unaffected (Fig. 54). Four of the species that die out are taxa that also occur in relatively shallow water, and have no sig- nificant presence in the deep-water faunas; we consider these to be stragglers from the neritic assemblages, and they have not been treated in the taxonomic section of this paper (Buntonia sp. 34, Xestoleberis sp. nov., Bythocypris sp. 42, and ?Bradleya sp. 56). Considering species that occur in more than one sample (i.e. those for which we can determine a depth range), 57 per cent of the turnover events take place between 950 m and 1 000 m, 29 per cent between 900 m and 950 m, and 14 per cent between 1 000 m and 1 050 m. The appearance and disappearance, respectively, of Parakrithe sp. 10 and ?Bradleya sp. 56 at 945 m suggests that the significant faunal change occurs closer to 950 m than to 1 000 m. It is unlikely, in fact, to occur at a specific depth for individual or all species, but for the purposes of discussion and diagrammatic presentation, we nominally take 950 m as the depth at which the boundary between the neritic and bathyal ostra- cod assemblages lies. The reality of a major faunal break at approximately this 330 ANNALS OF THE SOUTH AFRICAN MUSEUM Trachyleberis sp. 3017 ABYSSAL | | Krithe peypouqueti | | Krithe rex | | s Cytherella sp. 3027 : eit | e Krithe sp. 22 | BATHYAL | mt Dutoitella suhmi | a AN SEES / MEE custeuls NERITIC : UPPER : : a 20 7 Poseidonamicus major BATHYAL g Ambocythere sp. 3057 | | . Turnover Z Cytheropteron sp. 2914 | | il g g Krithe sp. 6 | | | o g Krithe sp. 19 | | | : y Krithe sp. 7 | | | j WEDS) | | Ss Tous 155 2-05 3-0 Tel G9. 28 | W | | water depth, km Indet. sp. 62 C0 Cytherella serratula | RC es em Cytheropteron cronini | es ital a a Parakrithe sp. 10 ey Cytheropteron sp. 2909 a Echinocythereis whatleyi es a a ey Rugocythereis horridus pS OO ADs | Hee Si pee a Krithe sp. 8 2s ay, ! ?Bradleya sp. 56 * -———_ l l Bythocypris sp. 42 * i l l 8 = patie fea Krithe spaiularis ar ia l Krithe capensis — l l Buntonia rosenteldi a Buntonia sp. 34 * ss | l Henryhowella melobesioides Xestoleberis sp. nov. x | | (0) 1 water depth, .km 4 Fig. 53. Depth range chart for Quaternary deep-water ostracods off south-western Africa. Inset shows turnover of appearances and extinctions (events) summed for 100 m intervals with bars centred on the median 50 m depth. ” on =) o ao) a AS © 3° = > — oO = - o =) oO re 5 = ® o = o a = > 2 — co c = 5 @) a © a] > ‘= - 2 © t o o o 2) Be o rs) ° a a wo € c o 5 o o ” wt o ry a Oo) o x . oe i ee tes Ge) oS Gg 2B & BB ® ae 2 ° o nN a o o = ie o o © © o =- o o © N © Ce = = © ro) er Qa a = xo = 5 io eS SP 8B S FR g 2 S&S & 8 g ® eS ss o 5 5 = € 2 8 So] = 5 o © ) ° 9° en ar a A et ONS re re eye 5 = Ww Q © c rr = o c fe = = = 5 ic > > o c c > o > ® 3 c c c ° 6) re) ) a = = iva) x a S a5 a x x x a w m 5) NERITIC 95 oes BATHYAL 1000 1050 appear disappear continue Fig. 54. Depth ranges for species at the Neritic/Bathyal Zone boundary. depth is also suggested by changes in the abundances of three diagnostic ostracod groups that dominate the deep-water populations (Henryhowella melobesioides, Krithe spp., and Buntonia rosenfeldi—referred to as the HKB assemblage) DEEP-WATER QUATERNARY OSTRACODA Baill eis ho Krithe sp. 22 Dutoitella suhmi Abyssocythere australis Krithe peypouqueti Trachyleberis sp. 3017 Krithe sp.4 Krithe sp.7 Krithe sp.6 Krithe sp.19 Echinocythereis whatleyi Rugocyther Ambocythere sp.3057 Cytherella serratula Cytheropteron cronini Krithe rex Henryhowella melobesioides Cytherella sp. 3027 Cytheropteron sp. 2914 Buntonia rosenfeldi Poseidonamicus major STATION No. 3354 me __ _ BATHYAL ABYSSAL 6851 6852 appear disappear Fig. 55. Depth ranges for species at the Bathyal/Abyssal Zone boundary. (Fig. 56): the commencement of a sharp decline into deeper water in the abun- dance of H. melobesioides; the temporary reversal of the abundances of Krithe spp. and B. rosenfeldi relative to each other; and a sharp decline into deeper water in the combined abundance of all three taxa. The faunal change between 2 000 m and 2 100 m, in fact, relates to one sample (TBD 3355) at 2 070 m, but the size in turnover in the fauna that is defined by this point (8 events), coupled with the apparently coincidental extinc- tion of Buntonia rosenfeldi and the appearance of Poseidonamicus major, indicate that a major shift in the character of the ostracod population occurs at approximately this level. Thirteen species probably appear or disappear in the 1 200 m interval straddling this depth (Fig. 55), so we nominally take the location of this sample to mark the depth of the boundary between the bathyal and abyssal ostracod assemblages. The faunal changes that are represented on the inset in Figure 53 by a cluster of turnover events between 1 300 m and 1 700 m are accompanied by a dramatic alteration in the character of the HKB assemblage (Fig. 56). Across this zone, the previously dominant taxon (H. melobesioides) rapidly declines in abundance from more than 85 per cent to c. 30 per cent of the total ostracod population, whereas Krithe spp. progressively, and B. rosenfeldi temporarily, become more important. In addition, the contribution to the total ostracod population of the HKB assemblage declines across this zone from more than 95 per cent to c. 70 per cent. We nominally take the point at which H. melobesioides ceases to be dominant to mark the boundary between these upper and lower bathyal assemblages, viz. 1 500 m. The vertical changes in the physical properties of the water column off south-western Africa have briefly been mentioned in the introduction, and are 332 ANNALS OF THE SOUTH AFRICAN MUSEUM shown in Figure 2. On Figure 56 they have been related to the ostracod distri- butions, and it is suggested that the important faunal changes described above can be directly correlated with the boundaries between and within the major water masses. Shannon (1966, 1985, figs 11, 12) has plotted the vertical limits of the low salinity zone within the AAIW and, although these fluctuate slightly with latitude, the base lies at approximately 1 000 m. This is close to the depth at which we have located an important change in the composition of the ostra- cod fauna (950 m), and consequently we suspect that the level of the neritic/ bathyal faunal boundary is controlled by the depth of the base of the low salinity zone within the AAIW. The AAIW is underlain by southward-flowing, low-temperature, high-salinity NADW, and the boundary between these two major water masses lies at approximately 1500 m off south-western Africa (Shannon 1985). This level coincides with the depth at which we have identified important changes in the HKB assemblage, and consequently we suspect that the level of the the upper/ lower bathyal faunal boundary is controlled by the depth of the AATIW/NADW shear zone. Within the NADW mass, Shannon (1985) has located the core at between 2km and 3 km water depth, and we suggest that the bathyal/abyssal faunal boundary coincides with the top of this zone. In the Cape Basin, Antarctic Bottom Water (AABW) lies below the NADW in water depths greater than about 4 km (Shannon 1985). It is colder and less saline than the overlying water mass, and its higher dissolved CO: values and relatively vigorous circulation are corrosive to carbonate valves (e.g. Tucholke & Embley 1984). The top of the AABW forms the carbonate lysocline, across which the calcareous skeletons of dead organisms are progress- ively more rapidly dissolved with increasing depth. This is an ‘aggressive’ physico-chemical environment but, because we have only one sample from below this level (TBD 6852: 4 736 m), we have insufficient data to establish if it coincides with a further change in the ostracod fauna. Certainly, the dead fauna was sparse and poorly preserved but we have no data on the living fauna. Table 8 summarizes the physical properties of the three major water masses mentioned above. Fig. 56. Variations in the abundances of the HKB ostracod assemblage (as % total ostracod fauna) from neritic to abyssal depths. Data are five point running means. Triangles (upper curve) = total assemblage; squares = Krithe spp.; diamonds = Henryhowella melobesioides; crosses = Buntonia rosenfeldi. Lower part of the diagram correlates aspects of the abundance curves with various water mass properties. Zone over which H. melobesioides is more than 50 per cent of total ostracod fauna is stippled, and the depths at which it dominates the HKB assemblage are shown as black bars. SMZ = salinity minimum zone; AAIW = Antarctic Inter- mediate Water mass; NADW = North Atlantic Deep Water mass. water -masses ® Henryhowella DEEP-WATER QUATERNARY OSTRACODA 333 UPPER LOWER NERITIC BATHYAL BATHYAL ABYSSAL 100 % 50 Henryhowella ae melobesioides Buntonia 7 rosenfeldi water depth, km NERITIC U. BATHYAL L. BATHYAL ABYSSAL ixed ayes AAIW NADW melobesioides | A | steady increase Krithe spp. { ; I | | | eer { | steady decline I ' Combined In20 461 B76 o1 8 7 | 6 5 | | | 9 ' Henryhowella ; : f | | Krithe spp. | I | ! Buntonia em i a | mixed assemblage NERITIC U. BATHYAL L. BATHYAL ABYSSAL @ data from Shannon 1985 p.124-125 = sharp increase = salinity minimum zone sharp decline ll SMZ CORE | | = H. melobesioides dominant = core of NADW = peak = low AAIW = Antarctic Intermediate water mass NADW = North Atlantic Deep water mass x10% 334 ANNALS OF THE SOUTH AFRICAN MUSEUM TABLE 7 Summary of changes at faunal boundaries. Species Depth (m) Upper limit of Upper Bathyal Zone (950 m) Appearance of: Parakrithe sp. 10 945 Cytheropteron sp. 2909 945 Cytheropteron cronini 990 Cytherella serratula 1 000 Indeterminate sp. 23 1 050 Indeterminate sp. 62 1 026 Disappearance of: Buntonia sp. 34 1 050 Xestoleberis sp. nov. 1 000 Bythocypnis sp. 42 1 000 ?Bradleya sp. 56A 945 Upper limit of Lower Bathyal Zone (1 500 m) Appearance of: Krithe sp. 7 1 600 Krithe sp. 4 1 600 Krithe sp. 6 1 662 Krithe sp. 19 1 662 Disappearance of: Krithe capensis 1 430 Krithe spatulanis 1 662 Upper limit of Abyssal Zone (2 070 m) Appearance of: Poseidonamicus major 2 070 Dutoitella suhmi 2916 Trachyleberis sp. 3017 2 916 Abyssocythere australis 2916 Krithe sp. 22 2916 Krithe peypouqueti 2 916 Krithe rex 2916 Cytherella sp. 3027 2 916 Disappearance of: Cytherella serratula 2 070 Cytheropteron cronini 2 070 Cytheropteron sp. 2914 2 070 Buntonia rosenfeldi 2 070 Ambocythere sp. 3057 2 070 TABLE 8 South-eastern Atlantic deep-water masses. (Data from Shannon 1985.) Water mass Depth Temperature Salinity Direction* (km) (°C) (P70) Antarctic Intermediate Water (AAIW) — 0,2—1,5 11-6 34,7-34,3 NW Salinity minimum zone 0,6—-1,0 SAS) North Atlantic Deep Water (NADW) LS=40 Ot 34,93—34,87 SE Antarctic Bottom Water (AABW) >4,0) 11-5) <34,77 SE * direction of flow off south-western Africa. DEEP-WATER QUATERNARY OSTRACODA 335 BATHYAL FAUNAS A total of 20 ostracod species has been identified from the Bathyal Zone off south-western Africa, three of which are neritic species at the limits of their depth ranges (Table 9A, Fig. 54). The fauna of this zone (950-2 070m) can be considered transitional between the taxonomically diverse (at least 120 species), but geographically endemic, neritic faunas of the continental shelf and upper slope, and the taxono- mically relatively restricted (17 species), but cosmopolitan, faunas of abyssal depths. Only eight species are restricted to the Bathyal Zone (Fig. 53) and, of these, five (Ambocythere sp. 3057, Cytheropteron sp. 2914, Cytheropteron sp. 2909, and Indet. spp. 23 and 62) occur at the limits, leaving three species only that range between the base of the AAIW salinity minimum zone and the AAIW/NADW shear zone: Cytherella serratula, Cytheropteron cronini, and Parakrithe sp. 10. In addition, these eight species constitute only 7 per cent of the total bathyal ostracod population. Consequently, the bulk of the fauna is composed of species that range into the over- and underlying zones. Neverthe- less, the composition of the fauna of the Bathyal Zone is sufficiently different to distinguish it as an identifiable population that has similarity indices of only 30 per cent and c. 10 per cent with the abyssal and neritic faunas, respectively. Throughout the Bathyal Zone the HKB assemblage constitutes a minimum of 65 per cent of the ostracod fauna, and over most of the depth range this figure exceeds 70 per cent (Fig. 56). The individual categories are dominated by Henryhowella melobesioides (61%), with Krithe spp. (13%) and Buntonia rosenfeldi (9%) in relatively minor roles (Fig. 57A), but these mean values mask fluctuations across the zone that indicate an important subdivision into upper and lower populations. This change occurs at approximately 1 500 m, above which the total HKB assemblage typically constitutes more than 80 per cent of the fauna and is dominated by H. melobesioides (>70%), and below which the HKB assemblage constitutes less than 80 per cent of the fauna and its three con- stituent taxa are more equally mixed. As we will discuss later, these faunal changes can be correlated with alterations in environmental parameters that define the Upper and Lower Bathyal zones. The most diverse genus within the Bathyal Zone is Krithe, with seven species, but no individual species ranges throughout the zone and there is a marked faunal discontinuity within the genus at the Upper/Lower Bathyal Zone boundary. UPPER BATHYAL ZONE Sixteen ostracod species occur within the Upper Bathyal Zone but three of these are stragglers from the Neritic Zone (Table 9A, Fig. 54). Of the 13 bathyal species, nine (representing 82% of the total fauna) are forms that also occur in shallower depths. The ostracod faunas of this zone are numerically dominated by the three taxa of the HKB assemblage (Figs 56, 57B). Although together they typically 336 ANNALS OF THE SOUTH AFRICAN MUSEUM TABLE 9 Percentage composition of fauna at bathyal and abyssal depths. U Lower overall A. Bathyal Zone (950-2 070 m) (950-1 500) (1500-2 070) Yo %o (4) ON NWP ON Henryhowella melobesioides #* 31 6 Buntonia rosenfeldi # Krithe sp. 8 # Echinocythereis whatleyi #* Cytherella serratula Rugocythereis horridus #* Krithe capensis # Parakrithe sp. 10 Krithe spatularis # Cytheropteron cronini Cytheropteron sp. 2909 Indeterminate sp. 23 Indeterminate sp. 62 Indeterminate fragments Krithe sp. 4 Krithe sp. 6 Krithe sp. 7 Krithe sp. 19 Ambocythere sp. 3057 Cytheropteron sp. 2914 Poseidonamicus major * Total for Krithe spp. Neritic forms (at lower limit of their depth range) | pe OP Ree — oF EP Ge Ge ey —_ Sy pa YJ FP OORrROOCRrR ORR OR RP RPP NWARWOH lo i) — Buntonia sp. 34 8 Xestoleberis sp. 35 1 = Bythocypnis sp. 42 1 =~) B. Abyssal Zone (051 aay % % Krithe Poseidonamicus major ef 29 Krithe rex Krithe peypouqueti Henryhowella melobesioides sd Dutoitella suhmi Abyssocythere australis Knithe sp. 4 Krithe sp. 7 Krithe sp. 6 Krithe sp. 19 Krithe sp. 22 Indeterminate Krithe Total Krithe spp. 4 Rugocythereis horridus - Echinocythereis whatleyi - Trachyleberis sp. 3017 Bradleya sp. 56 Cytherella sp. 3027 Indeterminate fragment me wy Ge Gy —_ SOR RB SB NAYNWNDBUMNAADADAD WW v — ja 17 species, 354 valves, diversity of 5 per cent * = common to bathyal and abyssal faunas; # = common to bathyal and neritic faunas Note: Percentages are based on total number of valves per zone, and not merely the number of valves in those samples in which a particular species occurs. DEEP-WATER QUATERNARY OSTRACODA 337 A 70 Bathyal fauna (950 — 2070m) % neritic species | ) 0 WAAL aes Iles = DA 123 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 B 70 bathyal fauna (950 —1500m) % neritic 0 A Nialictee a J sae 12 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 C 70 Lower bathyal fauna (1500—2070m) % GQ ; l (A Fig. 57. Species histograms (percentage of total ostracod fauna) for whole Bathyal (A), Upper Bathyal (B), and Lower Bathyal (C) zones. Species identification: 1— Henryhowella melobe- sioides; 2—Buntonia rosenfeldi,; 3— Krithe sp.8; 4—Echinocythereis whatleyi; 5—Cytherella serratula, 6—Rugocythereis horridus; 7—Krithe capensis; 8—Parakrithe sp. 10; 9—Krithe spatularis,; 10—Cytheropteron cronini; 11—Cytheropteron sp. 2909; 12—Indeterminate species; 13—Krithe sp. 4; 14—Krithe sp. 6; 15—Krithe sp. 7; 16— Krithe sp. 19; 17—Ambo- cythere sp. 3057; 18—Cytheropteron sp. 2914; 19—Poseidonamicus major; 20—total Krithe spp.; 21— Buntonia sp. 34; 22— Xestoleberis sp. nov.; 23—Bythocypris sp. 42. 338 ANNALS OF THE SOUTH AFRICAN MUSEUM constitute more than 80 per cent of the population, there is considerable variation across the zone from less than 70 per cent immediately below the upper boundary, to more than 90 per cent at 1,2 km and 1,45 km. These fluctu- ations, to a large extent, reflect the changes in abundance of Henryhowella melobesioides, which is the dominant taxon in the Upper Bathyal Zone. This species becomes progressively more abundant with increasing depth down the continental slope and reaches a peak (c. 75%) within the salinity minimum zone at the base of the Neritic Zone. It suffers a sharp, but temporary, decline at the top of the Upper Bathyal Zone but, with increasing depth, has a second abundance peak between 1 200 m and 1 300 m (>80%). Below this depth its importance rapidly declines, particularly across the Upper/Lower Bathyal boundary. Both Krithe spp. (6%) and Buntonia rosenfeldi (8%) are relatively minor components of the HKB assemblage in the Upper Bathyal Zone, in terms of overall abundance (Fig. 57B), but at particular depths are temporarily more abundant, where they respond to certain environmental factors. In the case of B. rosenfeldi, this occurs across the Upper/Lower Bathyal boundary (and in the salinity minimum zone at the base of the Neritic Zone), whereas Krithe species become more abundant immediately below the base of the salinity minimum zone in the AATIW (Fig. 56). The three species of Krithe that occur within the Upper Bathyal Zone (Upper Krithe Fauna) are all inherited from the overlying Neritic Zone (Figs 11, 13, 53). Their variations in relative abundance can be correlated with changes in environmental parameters (see next section), and allow a zonation of the Upper Bathyal Zone. Krithe capensis is the dominant species within the Neritic Zone but rapidly declines in abundance across the Neritic/Bathyal boundary, below which it is replaced by Krithe sp. 8. The latter in turn is replaced by K. spatularis in depths greater than approximately 1 000 m. Of the three Upper Bathyal Zone species of Krithe, only K. spatularis extends across the AAIW/NADW shear zone into the Lower Bathyal Zone, but it dies out near the top, where it is replaced by the four species of the Lower Krithe Fauna. Within the minor elements of the bathyal ostracod fauna, several species appear near the top of the Bathyal Zone, although only three have an extended range into deeper water, and of these only Parakrithe sp. 10 is confined to the Upper Bathyal Zone. Cytherella serratula is the sole representative of the genus within the Bathyal Zone, and is only relatively abundant either side of the Upper/Lower boundary, where it appears to favour the unstable physico- chemical environments associated with the shear zone. Across the continental margin off south-western Africa, the genus Cytheropteron is as diverse as Krithe (12 spp.), but the majority of its species are confined to the shelf and upper slope (see Fig. 8). Deep-water representatives are confined to three species, but only Cytheropteron cronini has been found over a wide depth range. Finally, Echinocythereis whatleyi and Rugocythereis horridus are both characteristic members of the Upper Bathyal Zone fauna and, although the latter has its DEEP-WATER QUATERNARY OSTRACODA 339 highest abundance within this zone, neither is numerically important, except within individual samples. LOWER BATHYAL ZONE Fourteen species have been recovered from the Lower Bathyal Zone, 50 per cent of which extend into the Upper Bathyal Zone, 57 per cent extend into the Abyssal Zone, and 36 per cent extend into the Neritic Zone (Tables 9, 10). The character of the Lower Bathyal Zone fauna contrasts strongly with that of the Upper Bathyal Zone fauna by having a mixed Henryhowella melobesioides/ Krithe spp./Buntonia rosenfeldi assemblage, as opposed to one dominated by H. melobesioides (Figs 56, 57C). In terms of total specimens recovered from the Lower Bathyal Zone, H. melobesioides is the most important species (31%) but, except for isolated samples, it is subordinate in abundance to one or other of the two other taxa. In addition, with increasing water depth H. melobesioides becomes progressively less important, whereas Krithe species steadily become more abundant and diverse (Figs 11, 13), with the appearance of four species in the top part of the zone. Including K. spatularis, which extends downwards from the Upper Bathyal Zone, 36 per cent of the species recorded from the Lower Bathyal Zone belong to the genus Krithe. The most abundant of these is Krithe sp. 4. Buntonia rosenfeldi occupies a transitional position in the switch from H. melobesioides to Krithe spp. dominated faunas at the Upper/Lower Bathyal Zone boundary, by temporary increases in abundance at water depths where H. melobesioides declines but where Krithe spp. do not expand to fill the environmental niches. This opportunistic expansion is also displayed by Echinocythereis whatleyi and to a lesser extent by Cytherella serratula, both of which show marked increases in abundance at approximately the same levels as B. rosenfeldi. The depth at which we nominally take the Lower Bathyal/Abyssal Zone boundary (2 070 m) coincides with our first record of the species Poseidonamicus major, which is the link to the typical abyssal assemblage in sample TBD 6581. This is also the deepest occurrence of several of the bathyal species— Buntonia rosenfeldi, Cytherella serratula, and Cytheropteron cronini—and the sole record of two rare taxa, Ambocythere sp. 3057 and Cytheropteron sp. 2914. ABYSSAL FAUNAS It is in the Abyssal Zone that the true deep-water, cosmopolitan ostracod faunas are encountered. The character of the population is markedly different from that found in the Bathyal Zone and the fauna is dominated by various species of the genus Krithe (8 species), which combined comprise 47 per cent of the ostracod population in our largest sample (TBD 6851, 2 916 m, 354 valves). The most abundant single species are Poseidonamicus major (29%) and Krithe peypouqueti (13%), with the overall importance of the HKB assemblage within the total ostracod population showing a steady decline from the upper part of 340 ANNALS OF THE SOUTH AFRICAN MUSEUM TABLE 10 Comparison of Upper Bathyal, Lower Bathyal and Abyssal faunas. A. SIMILARITY Upper Bathyal* Lower Bathyal Abyssal 20 species, 7 common 24 species, 8 common similarity = 35% similarity = 33% Bathyal Abyssal 20 species 17 species 27 species, 8 common similarity = 30% ates =e No. species common to both units Similarity Total no. different species in both units nD common species x 100 * — values do not include the three neritic taxa B. PERCENTAGE SPECIES COMMON Abyssal Lower Upper ee $ Bathyal Bathyal Namie Abyssal Lower 57 2 50 36 Bathyal Upper = Bathyal 23 54 69 Neritic® * = No. of neritic species taken as 120 (unpublished data). Examples: What percentage of Upper Bathyal species occur in the Lower Bathyal Zone? ...54 What percentage of Lower Bathyal species occur in the Upper Bathyal Zone? ...50 the Abyssal Zone (65% at 2 070 m) into deeper water (50% at 2 916 m), in which depths Buntonia rosenfeldi does not occur. Forty-seven per cent of the abyssal species are endemic to the Abyssal Zone, which contrasts strongly with 31 per cent and 14 per cent endemism for DEEP-WATER QUATERNARY OSTRACODA 341 the Upper and Lower Bathyal Zones, respectively. In addition, several of the genera do not presently occur in the shallower zones off south-western Africa: Trachyleberis (? = Legitimocythere), Abyssocythere, and Dutoitella. All of the Krithe species that comprise the Lower Krithe Fauna, and which appear beneath the AAIW/NADW boundary in the Lower Bathyal Zone, persist into the Abyssal Zone, but they are subordinate in abundance to the abyssal endemic K. peypouqueti. Rugocythereis horridus and Echinocythereis whatleyi also occur in the Abyssal Zone but form only minor elements of the overall population (2% and 1%, respectively). Our deepest sample (TBD 6852, 4 736 m) lies beneath the carbonate com- pensation depth, and contains a sparse fauna of Krithe peypouqueti and Trachyleberis sp. 3017 that shows signs of dissolution. This is the only record we have of the ostracod fauna within the AABW mass, which is such an important feature of the circulation in the Cape Basin and has vigorously scoured or main- tained an omission surface in the region since at least late Miocene time (Embley & Tucholke 1984; Dingle et al. 1987). Preservation of the shallower assemblages within the overlying NADW is generally good. SEDIMENTARY AND OCEANOGRAPHIC ENVIRONMENTS The studies of Rosenfeld & Bein (1978), Cronin (1983), and Benson et al. (1983) on the ostracod faunas of the continental margins off north-western Africa and eastern North America, suggest that variations in dissolved oxygen, salinity, and temperature are the main factors in determining the water-depth ranges of individual deep-sea species. We will assess the influence of these parameters on the faunas off south-western Africa, and then briefly compare our results with their work. OSTRACOD ABUNDANCE AND WATER DEPTH The only physical parameter that is directly linked to water depth is hydro- static pressure. All other factors that are likely to limit the vertical distribution of ostracods (e.g. temperature, salinity, dissolved O2 and CO2, nutrients, light, substrate, etc.) vary with depth only through a general vertical zonation imposed by the structure of the water column. Nevertheless, within any area, the latter phenomenon will maintain a relatively steady-state depth-related population structure that allows a vertical zonation based on either assemblages or individual species. Projecting such depth-zonations to other areas must, by definition, be attempted with caution. Murray (1973: 168) discussed the same problem in relation to benthic foraminifera and concluded ‘. . . that depth zones recognized in one area on the basis of certain indicator species are only applicable to that area and to adjacent areas where the environmental conditions are essentially the same’. However, because globally there are fundamental similarities in the vertical structuring of the deep water masses, a similar, relative depth-related succession of ostracod taxa can be anticipated world-wide. _ 342 ANNALS OF THE SOUTH AFRICAN MUSEUM Figure 58A shows the variation in abundance of ostracods (measured as number of valves per 100g of the original, dry sediment sample) on the continental margin off south-western Africa. The highest abundances (>500 valves/100 g) all lie on the continental shelf (<220 m water depth, above the thermocline), with the majority of sites on the slope containing less than 100 valves/100 g (mean = 29). There is a cluster of higher values (>100) either side of 400 m water depth, and a further ‘peak’ (up to 230 valves/100 g; mean = 98) at the base of the AAIW salinity minimum zone (Neritic/Bathyal Zone boundary). Otherwise, maximum values for the Upper Bathyal Zone lie between 40 and 50 valves/100 g (mean = 24) (Fig. 58B). Between 1 450 m and 1500 m, immediately above the AAIW/NADW shear zone (i.e. the Upper/ Lower Bathyal boundary), values drop to less than 10 valves/100 g, and recover only to 15-20 valves/100 g between this level and our deepest data point at 1 780 m. These data suggest that there is no relationship between variation in ostra- cod abundance and water depth, but there is strong correlation between variations in abundance and structure of the water masses. OSTRACOD ABUNDANCE AND MUD CONTENT OF SEDIMENTS The mud (silt and clay: <63 micron) content of 218 sediment samples from the continental margin off south-western Africa is plotted against water depth in Figure 59. There is a general increase in the mud content of sediments with increasing depth from the continental shelf to 1 km, both in terms of number of samples with more than 50 per cent mud content, and in mean percentage mud. However, although the former parameter maintains a value of more than 90 per cent in depths greater than 1 km, the mean percentage mud content of the sedi- ments reaches a high of 84 per cent at 1 km, but decreases slightly farther downslope (to 80% at 1,7 km). Mean percentage mud values for the various zones are: Neritic—49 per cent; Upper Bathyal—82 per cent, Lower Bathyal— 82 per cent. A comparison of Figures 58A—B (details of the mean percentage mud and ostracod abundance for the Bathyal Zone), 59 and 60 (scattergrams of percent- age mud plotted against ostracod abundance as valves/100g), reveals the following trends: 1. Ostracod abundances on the continental margin off south-western Africa are greatest (>500 valves/100 g) on the mid to outer continental shelf (150-220 m water depth). Although the mean mud content of the sediments in this depth range is less than 50 per cent, all the samples with abundances greater than 300 valves/100 g occur in samples with a mud content of more than 50 per cent. This indicates that the mud content of the shelf samples is very variable but that ostracods are most abundant in muddy samples, with maximum populations (up to 1 900 valves/100 g) occurring in sediments with mud contents between 75 and 90 per cent (Fig. 60A). Until we have studied these faunas in detail, we cannot comment on the specific diversity of such populations. DEEP-WATER QUATERNARY OSTRACODA 343 = (50%; triangles = mean % mud: values are plotted at median depth point (e.g. data for the 0-200 m interval are plotted at 100 m level). Mud values are from Birch (1975), Rogers (1977) and Bremner (1981). 2. On the upper continental slope (water depths between 223 m and 900 m), the mean ostracod abundance is 29 valves/100 g, whereas the mean mud content steadily increases to more than 70 per cent, so that in this region the sympathetic relationship between high mud values and high population numbers that we detect on the shelf does not hold. 3. At the base of the salinity minimum zone (SMZ)/bathyal thermocline of the AAIW mass (i.e. at the Neritic/Upper Bathyal Zone boundary), there is a narrow zone (c. 950-1050 m) of high ostracod abundance (mean = 98 valves/100 g). The mean percentage mud value for the 800-1 000 m interval is 74 per cent, but where it rises to 85 per cent in the underlying 200 m wide sector, the ostracod abundances decline significantly (mean = 24 valves/100 g— Fig. 58B). A further decline in abundance occurs in the lowermost Upper Bathyal and Lower Bathyal zones (mean = 8), whereas in the same interval the mean mud values decline slightly (84-77%, and in detail hover around a plateau of about 80%). In general, therefore, although the mud content of the Bathyal Zone sediments is high (>80%—Fig. 60), the ostracod abundances are low overall (<40 valves/100 g), and even the larger populations are relatively small in comparison with those on the continental shelf. DEEP-WATER QUATERNARY OSTRACODA 345 all samples 0 —1800m valves/ 100g Oo ro) ° deep—water samples 200 900 — 1800m valves/100g % mud Fig. 60. Number of ostracod valves/100 g plotted against mud content of sample. The total ostracod fauna has been used. A. Samples from a water depth range of 0-1 800 m. The rectan- gular block in the lower right corner outlines the data field used in Fig. 60B. B. Samples from a water depth range of 900-1 800 m. Taken together, the data on total ostracod population abundances in relation to water depth and mud content of sediments suggest that, in water depths greater than approximately 200 m, variations in overall abundance are not directly related to the mud content of the sea-floor sediments but are con- trolled by the structure of the water column. In as much as changes in the water 346 ANNALS OF THE SOUTH AFRICAN MUSEUM column will also influence the mud content of the sediments, locally there is a sympathetic relationship in alterations in mud content and overall ostracod abundance. On the continental shelf (above the thermocline and within the influence of surface currents), there is a direct correlation between mud content and total population abundance, but this is a topic we will address in a later contribution on the shallow-water faunas. _ The above observations are based on data that monitor variations in the total ostracod population, but these may mask correlations for individual species. In Figure 61A—C, we have plotted the abundance (as percentage total ostracod population) against mud content of sediments for the three taxa in the HKB assemblage and, in Figure 61D, have summarized the percentage of records for each species that occurs in successive 20 per cent mud categories. From these data, we conclude that all three taxa have a preference for sediments with a mud content of more than 50 per cent but that the occurrences of Henry- howella melobesioides and Krithe spp. are more mud-specific (57% and 55% of records occur in sediments with >60% mud, respectively) than Buntonia rosen- feldi (48%). However, the latter has a lower tolerance of low mud contents. We did not record Buntonia rosenfeldi in sediments with less than 20 per cent mud, whereas 7 per cent and 3 per cent respectively of occurrences of H. melo- besioides and Krithe spp. were within this category. Further, B. rosenfeldi appears to tolerate mud contents of more than 85 per cent less well than either of the other two species. Its abundance (as percentage of total ostracod popu- lation) in this category is never more than 25 per cent, whereas in very muddy sediments both H. melobesioides and Krithe spp. are frequently the dominant taxa. This is especially the case with H. melobesioides. From this brief survey we can conclude that, although the abundance of the overall ostracod population is less influenced by the mud content of the bottom sediments than by factors associated with the ambient water mass, individual species are so influenced. OCEANOGRAPHIC FACTORS Figure 62 summarizes the vertical changes in dissolved oxygen, salinity, and temperature in transects across the continental margin off south-western Africa. Using these profiles in conjunction with the vertical distribution of selected ostracod taxa (Fig. 63) and variations in the composition of the overall fauna (Figs 53, 56), we can assess the effectiveness of vertical changes in the physico- chemical properties of the water column as barriers to ostracod distribution, and hence the maintenance of the composition of the faunas of the various depth zones. On the continental shelf, temperature and salinity values are relatively high (>12°C and >34,9%o, respectively) and sediment textures variable. In addition, the pattern of dissolved oxygen values is complicated, with a single minimum zone in the south (Stander 1964), and a double minimum in the north (Chapman & Shannon 1985), with the result that the ostracods of this shallow zone have DEEP-WATER QUATERNARY OSTRACODA 347 B 100 B. rosenfeldi % 20 100 Krithe spp. H. melobesioides B. rosenfeldi 0 100 0) 20 40 60 80 100 % mud % mud Fig. 61. Abundance of individual species (as % total ostracod fauna) of the HKB assemblage plotted against mud content of samples. A. Buntonia rosenfeldi. B. Krithe spp. C. Henry- howella melobesioides. D. Summary of variation of abundance versus percentage mud for the three species. Values are shown as percentage of records within successive 20 per cent intervals of mud, and have been equalized to avoid sampling bias. a complex distribution pattern. We will be describing these faunas in a later publication and defer further discussion of this zone until then. The top of the AAIW mass lies at about 200 m (Shannon 1966, 1985) and is associated with a relatively thick thermocline that is steepest between approxi- mately 200 m and 400 m. At the top of the AAIW, water temperatures vary 348 ANNALS OF THE SOUTH AFRICAN MUSEUM Vite layer | AAIW NADW A ABW shear] zone A Sy 9, OP i} Dissolved oxygen water depth, km 34:7 ae . } Mi Salinity 34°3 vA (0) ry 4 Hy water depth, km c i Or i) 1 Hy 5 HY Temperature 1 0) 15 4 water depth, km U. L. NERITIC BATHYAL ABYSSAL Fig. 62. Variations in physico-chemical parameters on the sea floor along transects off south- western Africa. Water-column structure and faunal zones are correlated along the top and bottom margins of the diagram, respectively. A. Dissolved oxygen (mf/€) at 20°S, 34°S, and a composite profile between 28°S and 33,5°S (the deflection in the curve at 2,5 km is an artefact caused by juxtaposition of data from two transects). B. Salinity (%o) at 24°S and 34°S. C. Tem- perature (°C) at 24°S and 34°S. Constructed with data from Fuglister (1960), Stander (1964), Bubnov (1966), Shannon & Van Rijswijk (1969), Welsh & Visser (1970), Gorshkov (1978), Shannon (1985), and Chapman & Shannon (1985). Data points taken at 100 m intervals. Zones have been shaded to aid correlation. Fig. 63 (see facing page). Variations in abundance (% total ostracod population) of various ostracod species correlated with water column barriers, and faunal zones. A—Henryhowella melobesioides (five-point means); B—Krithe species (five-point means); C—Rugocythereis horridus (raw data); D—Cytheropteron species (five-point means); E—Cytherella species (five-point means); —Buntonia rosenfeldi (five-point means); G—Echinocythereis whatleyi (three-point means). DEEP-WATER QUATERNARY OSTRACODA anne layer AAIW NA DW Ue zone 100 A Bi oa Poe z Vo eerie Eee () H. melobesioides a =) core Krithe spp. R. horridus 30 o =a (o} Cytheropteron spp. Cytherella spp. percentages of total ostracod fauna B. rosenfeldi 1 2 water depth, km U. L. NERITIC BATHYAL ABYSSAL 349 350 ANNALS OF THE SOUTH AFRICAN MUSEUM from north to south between 11°C and 9°C, and these fall to c. 3,4°C at the base. The gradient of the temperature changes is very low below 1 000 m but increases steadily into shallower water, with the main temperature ‘break’ at about 3,5°C. This bathyal thermocline approximately coincides with the base of the salinity minimum zone (SMZ) at the core of the AAIW mass (Shannon 1966, 1985), which itself is a relatively low salinity body of water sandwiched between the higher salinity continental shelf waters and the NADW mass (Fig. 62). The lower limit of the SMZ lies at about 1 000 m along the whole of the continental margin off south-western Africa and coincides with a dissolved oxygen minimum zone in which values vary from 4,25 ml/I at 30°S to 3,8 ml/l at 34°S (Fig. 62A). The SMZ is thicker in the south (32°S), where the upper limit occurs at about 500 m, in comparison with 600 m at 24°S. Minimum salinity values in the SMZ vary from 34,36%o in the south to 34,48%c in the north. The physico-chemical hiatus that controls the depth of the Neritic/Bathyal Zone boundary appears, therefore, to be a combination of the bathyal thermocline (below which the water temperature is <3,5°C), a steep increase in salinity below the SMZ (with the ‘critical’ level around 34,50%c), and an oxygen minimum zone. The boundary between the AAIW and NADW masses lies at about 1 500 m off south-western Africa (Shannon 1985). There is no apparent change in either the temperature or salinity gradients across the contact but, because these bodies are flowing in opposite directions, there will be a relatively intense zone of shearing across it (marked by turbulence and a sharp velocity gradient). The slightly lower mud values detected between 1 400 m and 1 500 m (Fig. 58B) may be related to turbulence in the lowermost part of the AAIW. Variability in temperature, salinity and current strengths, rather than significant breaks in their gradients, seem to be the main factors controlling the location of the Upper/Lower Bathyal zone faunal boundary. The core of the NADW lies between 2 000 m and 3 000 m (Shannon 1985) and its upper boundary is marked by a sharp change in the salinity values, where the gradient decreases rapidly and below which salinity values peak at about 34,9%o in the north and 34,86%o in the south. Below the depths at which these peaks occur (2 000 m and 2 300 m, respectively), the salinity values decrease very slowly. There are no accompanying temperature changes at the top of the NADW core, so that the Bathyal/Abyssal Zone faunal boundary that is main- tained by this feature must be related to the relatively large upslope decrease in salinity and any water turbulence that is caused by the velocity gradient across the top of the NADW core zone. The NADW/AABW boundary lies at about 4 000m and is marked by small, but significant, increases in both temperature and salinity gradients. Because we have no ostracod-rich samples below 4 000 m, we cannot comment on whether or not this water-mass contact causes a significant faunal hiatus. Certainly, because the top of the AABW mass marks the upper limit of the DEEP-WATER QUATERNARY OSTRACODA 351 carbonate lysocline, the chemical environment below this depth can be expected to be severe. To assess the effectiveness of these boundaries as barriers to the habitats of various ostracod species, we have plotted the abundances of the numerically most important species and genera against depth (Fig. 63), and can extract the following salient points. 1. Base of salinity minimum zone/bathyal thermocline of AAITW EFFECTIVE: Krithe spp. (including Parakrithe)—there is a decrease in abundance of the overall population into and out of the SMZ, with a mid- zone high. Two species turn over at lower boundary. Cytherella spp.—very effective barrier; high abundances and number of species in Neritic Zone do not persist across SMZ; bathyal species are isolated by barrier. Echinocythereis whatleyi—does not range above SMZ. Rugocythereis horridus—does not range above SMZ. Cytheropteron—ranges of individual species restrained by SMZ, but not effective in altering overall population abundances, which reach a high in the SMZ. INEFFECTIVE: Henryhowella melobesioides—the drop in abundance at base of SMZ may be related to a decrease in mud content. Buntonia rosenfeldi—favours the conditions therein and reaches a minor peak of abundance just above base of SMZ. 2. Boundary of AAIW and NADW masses (shear zone) EFFECTIVE: Henryhowella melobesioides—very effective; dramatic decrease of abundance into NADW. Krithe species—very effective; dramatic increase in abundance into NADW;; high species turnover. Rugocythereis horridus—confines high abundances to the Upper Bathyal Zone. INEFFECTIVE: Buntonia rosenfeldi—reaches peak abundance in the shear zone. Cytherella species—C. serratula reaches maximum abundance in the shear zone. Echinocythereis whatleyi—reaches maximum abundance in the shear zone. Cytheropteron species—no significant changes across this barrier. 3. Top of NADW core EFFECTIVE: Krithe species—small abundance change across barrier, but apparently no turnover of species. Buntonia rosenfeldi—is severely restricted below boundary. Cytherella species—is severely restricted below boundary. Echinocythereis whatleyi— abundances low beneath barrier. 352 ANNALS OF THE SOUTH AFRICAN MUSEUM Rugocythereis horridus— abundances low beneath barrier. Cytheropteron species—C. cronini does not range below barrier. Poseidonamicus major—does not range above barrier. INEFFECTIVE: Henryhowella melobesioides—no_ significant change in abundance across barrier. Table 11 lists the barriers and the responses to them by individual species. We can summarize the physico-chemical features of each barrier and how it maintains the character of the various faunal zones as follows: 1. Neritic/Upper Bathyal Zone boundary (AAIW—base of SMZ/bathyal thermocline). This barrier consists of: (a) change in temperature gradient (‘critical’ temperature c. 3,5°C); (b) steep salinity gradient (i.e. a zone of rapid change); and (c) oxygen low zone; and has the following effects: (a) prevents the neritic populations of the diverse genera Cytheropteron, Buntonia, and Cytherella from moving downslope and (b) prevents Rugocythereis horridus and Echinocythereis whatleyi from moving upslope. 2. Upper/Lower Bathyal Zone boundary (AAIW/NADW contact and associated shear zone). This barrier consists of turbulence, resulting in variable tempera- TABLE 11 Summary of barrier effectiveness. SMZ AAIW/ NADW Approximate Temp. (°C) 4,0-3,4 3; ‘critical’ Salinity (°/,) 34,5 34, parameters Oxygen O, low Henryhowella melobesioides Krithe spp. as = as Buntonia rosenfeldi Cytherella spp. +3 #2 Echinocythereis whatleyi oD #2 Rugocythereis horridus aD *1 me *3 * Cytheropteron spp. Poseidonamicus major * — effective barrier moving from shallow to deeper depths, cannot tolerate: ‘ —lower temperature and higher salinity —low temperature and high salinity > — lower temperature and low oxygen moving from deep to shallower depths, cannot tolerate: a — higher temperature and lower salinity > — higher temperature, higher salinity, and low oxygen © — isolated peak in SMZ Ranges in ‘critical’ parameters relate to north to south variations DEEP-WATER QUATERNARY OSTRACODA 353 tures, salinities and current strengths, and has the following effects: (a) separ- ates the Henryhowella melobesioides-dominated Upper Bathyal fauna from the mixed H. melobesioides/Krithe spp./Buntonia rosenfeldi Lower Bathyal fauna; (b) limits the depth ranges of several Krithe species to produce the Upper and Lower Krithe faunas; and (c) allows three species to opportunistically increase their abundances— Buntonia rosenfeldi, Cytherella serratula, and Echino- cythereis whatleyi. 3. Lower Bathyal/Abyssal Zone boundary (top of NADW core). This barrier consists of: (a) change in salinity gradient (‘critical’ salinity c. 34,89%o in the north and 34,86%oc in the south); and (b) ?turbulence, and has the following effects: (a) maintains peak Krithe spp. population abundances downslope; (b) prevents Buntonia rosenfeldi and Cytheropteron cronini from extending farther downslope; (c) prevents Poseidonamicus major from extending farther upslope. This regulates the upper level of a true abyssal taxon. Finally, it is clear that, although we have been able to identify the efficiency of these physico-chemical changes in the water column as regulators of overall faunal character, the taxa involved react to these changes in different ways. To some they are solid barriers and to others sieves, and it is not possible at this stage to identify a predictable pattern. A particularly good example of this apparently random response occurs at the AAIW/NADW shear zone. Here, within what we can only tentatively suggest is a narrow, turbulent layer of mixed water with variable temperatures, salinities, and current strengths, there are relatively large increases in abun- dances of Buntonia rosenfeldi, Cytherella serratula, and Echinocythereis what- leyi. In this region, therefore, the combination of narrow, but unstable, temper- ature and salinity ranges involved (3,18-3,25°C and 34,75-34,84%o in the north, to 2,91—2,79°C and 34,68—34,73%o in the south) afford a highly favourable ecological niche for at least three species to exploit opportunistically. At the same time, this zone separates AAIW—in which H. melobesioides is abundant and Krithe species relatively sparse, from NADW-—in which the roles of these - two taxa are almost reversed. However, neither of these species finds the con- ditions within the shear zone at all favourable. A further aspect of this complex response can be seen in the distribution of H. melobesioides, where the portion of the water mass in which it is most abundant (AAIW) has temperature and salinity characteristics of 2,9-3,3°C (mean 3,1°C) and 34,65-34,68%o (mean 34,67%o), whereas the underlying water mass, which it finds so unfavourable, has charactersitics of 2,78-3,1°C (mean 2,94°C) and 34,77-34,87%o (mean 34,82%c). The differences involved are 0,16°C and 0,15%c, respectively, yet in the Neritic Zone (across the upper part of the SMZ) changes in temperature and salinity five times these values result in abundance changes only half as large. 354 ANNALS OF THE SOUTH AFRICAN MUSEUM COMPARISON WITH RESULTS FROM OTHER AREAS We will briefly review four recent studies on deep-water ostracod faunas in which attempts were made to correlate distribution with various environmental parameters, and then summarize the results in the light of our findings. Cronin (1983) worked off the south-eastern USA but, because his survey did not assess environmental factors below 1 100 m, his data relate primarily to shallower depths than we have investigated. Nevertheless, there are similarities in faunal content that make comparison relevant. Cronin found that the two prime controls on depth ranges are the 15—8°C thermocline at 150 m, and the oxygen minimum zone (3 ml/l) between 200-800 m. These barriers define two main ostracod faunas: one within the O2 minimum zone, and one below it. The latter includes some taxa similar to our bathyal faunas. Benson et al. (1983) worked off the continental margin of north-eastern North America. They identified three ostracod faunas (biofacies), which they correlated with particular water masses. These faunas lie at 400-1 400 m (Labrador Sea Water mass; Upper Slope Biofacies); 1 500—2 399 m (North East Atlantic Deep Water mass; Transitional Biofacies); and >2 400 m (Denmark Strait Overflow Water mass; Lower Slope and Rise Biofacies). Benson et al. (1983) concluded that the limiting parameters are probably temperature, and possibly salinity. No correlation was observed between sediment type or the organic carbon content of sediments. Rosenfeld & Bein (1978) worked off north-western Africa. Their study did not extend to abyssal depths but, as with Cronin’s (1983) work, some similarity with our faunas makes comparison relevant. On the north-western African margin, there is a major faunal break on the upper slope that separates a ‘shallow-water’ (100-483 m) fauna from a ‘deep-water’ (470-2 859 m) fauna. Rosenfeld & Bein (1978) did not discuss the nature of the environmental par- ameters that maintain this boundary but assumed that it was temperature controlled (i.e. it marks the top of ‘psychrosphere’). A northward elevation of this level was attributed to local upwelling. No linkage was detected between ostracod distribution and sediment texture. Peypouquet & Benson (1980) compared the ostracod faunas from two traverses in the south-eastern Atlantic (off Walvis Bay, and off northern Angola). Their objective was to assess the role of the Walvis Ridge in limiting the distribution of various taxa. They identified three major depth-related faunas in the Angola Basin and two in the Cape Basin (no samples were collected below 3 000 m in the latter), and these were correlated with major water masses: epibathyal stage (400-1 500 m) = AAIW; mesobathyal stage (1 500-3 000 m) = NADW,; and infrabathyal stage (3 700-4 700 m) = AABW. They detected differences in the epibathyal and mesobathyal faunas on either side of the Walvis Ridge, which they related to nutrient rich (particulary P and Si), Antarctic-derived water affecting higher bathymetric levels on the southern side of the gidge. It is the contrast in dissolved O2 values, rather than minimal DEEP-WATER QUATERNARY OSTRACODA B55 variations in temperature and salinity, that is thought to maintain the faunal differences between the two basins. Temperature, salinity, and dissolved oxygen are all cited as parameters that control the depth ranges of ostracod species, as well as regulating population abundances and diversity. However, the four previous surveys that we have reviewed, as well as our own results, illustrate that their roles vary from species to species, and also vary for a given species within the water column. Peypouquet & Benson (1980) also suggested that variations in nutrient levels and dissolved oxygen can be responsible for major lateral taxonomic differences on a regional scale. As barriers to vertical distribution, local high gradients in any of the parameters we have mentioned can be expected to be limiting, and this is particu- larly the case in relatively shallow water. Consequently, the shelf/slope thermocline (e.g. Cronin’s 1983 data), and shelf oxygen minima (e.g. eastern boundary upwelling systems, such as that off south-western Africa) can be expected to constitute major faunal barriers. In deeper water, lower gradients may not appear to have the same potential but nevertheless are equally effective. It follows, therefore, that the deep-water taxa are less tolerant of, say, temperature and salinity changes, or critical combinations thereof. The latter pre- sumably must be the case for species, such as Henryhowella melobesioides, to cite a local example, that have a wide depth range. In fact, the concept of a ‘depth range’ is misleading because it merely reflects a range over which certain physico- chemical parameters are tolerable. As all the studies we have cited confirm, these tolerance ranges are governed by the characteristics of the ambient water masses. In the deep sea, therefore, the vertical faunal zonation merely reflects the local structure of the water column. INTRA-OCEANIC RELATIONSHIPS Table 12 shows the vertical distribution of key taxa from the deep-water studies in the Atlantic Ocean. Although certain taxa are universally present and often locally abundant, both in the lower neritic, bathyal and abyssal zones (in particular various species of Henryhowella, Cytherella, and Krithe), there are some regional contrasts, which can be related to differences in water-column structure. UPPER BATHYAL FAUNAS Here we include ostracod populations between depths of 900-1 500 m. At the lower level, faunal breaks have been recognized off north-eastern America and south-western Africa (Peypouquet & Benson 1980 recorded a faunal bound- ary at this depth off Angola, although they had no data between 650-2 000 m), and this depth coincides with the upper limit of the NADW mass, which in the south-eastern Atlantic is overlain by AAIW, and in the North Atlantic by the Labrador Sea Water mass. ANNALS OF THE SOUTH AFRICAN MUSEUM 356 alaysMosskqy SladayIA08ny sladaysMoulyor snoupuopiasog slasaysQooulyoq C-dYJUy 009 Z ojyjanoinq PEG MOWIUE) ap Te eo [-ojjasayi) suaqajAyood [ 00r Z L-ayiuly é-vjjasaysay suaqgajAYy od piuojung 00Z Z 9-IYJLy] 9-SIDIIO ee ee ee ee ae oudonopaoooceAn SuUaqgajAYIVA J, 0007 ----------- C-aYyJLy [-ojjasayia) slasayjKo08ny elep ou Z-IYJUy OO8T jedmeg ¢-SIOMIO piuojung ) L-DyjasayiKy IOMOT stasayMoulyoy Dyjamoydiuay] piuojung 009 I wenenennn ene meee nnn nnn SlasayMOOUlyOT won--=----=---- won-------- piuojung C-OYIUY 00r I s1adaypMOO8NY Dyjamoyduuayy ciorerecceecaes StadaysAOUlYyIA €7-SIOMIO. «SCOOT TSCeAtIeg 6-SISYIO €-dYJUy DJasayya) toddq vyjamoyduuazy [-ojjasayi) Byep ou 6-SI9YIO 6-IYIUYY 000 I wannnnn nanan ne [-oyjasayyay vyjamoyciuayy waonnn=-=- SladayAOUIYOA [-ayJuy sudio010W 008 SETYINOIN I cere Lt: L€-SIaMIO 009 vyjamoyuiuayy P-SIOUIQ 4 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 3 cm margins all round. First lines of paragraphs should be indented. Tables and a list of captions for illustrations should be typed separately, their positions indicated in the text. All pages should be num- bered 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 18cm (19 cm including caption); the reduction or enlargement required should be indicated (and preferably uniform); orig- inals 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 caption; 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. If Letraset is used authors are requested to use Helvetica-style letter- ing, if possible. 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, 1969b; Jones 1971)’ ‘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, 1969b) 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 (according to the World list of scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses, volume number, part number 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 la vitalité des mollusques. Journal de conchyliologie 88 (3): 100-140. FiscHeR, P. H., Duvat, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines. Archives de zoologie expérimentale et générale 74 (33): 627-634. Koun, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee region of Ceylon. Annals and Magazine of Natural History (13) 2 (17): 309-320. Koun, A. J. 1960b. Spawning behaviour, egg masses and larval development in Conus from the Indian Ocean. Bulletin of the Bingham Oceanographic Collection, Yale University 17 (4): 1-51. THIELE, J. 1910. Mollusca. B. Polyplacophora, Gastropoda marina, Bivalvia. In: SCHULTZE, L. Zoologische und anthro- pologische Ergebnisse einer Forschungsreise im westlichen und zentralen Stid-Afrika ausgeftihrt in den Jahren 1903-1905 4 (15). Denkschriften der medizinisch-naturwissenschaftlichen Gesellschaft zu Jena 16: 269-270. (continued inside back cover) ANNALS OF THE SOUTH AFRICAN MUSEUM ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM Volume 99 Band December 1990 Desember Part 10 Deel Kas SZ WARIO D S B:9.8:3.9) & oo So = A TOOTH-BEARING MAXILLA REFERABLE TO LYCORHINUS ANGUSTIDENS HAUGHTON, 1924 (DINOSAURIA, ORNITHISCHIA) By Cc. E. GOW 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, HORA 59), HOLD, 4S, 8, toi,), S03, 5, TO), AGL, tapi), TOD), 8, OGD, 7), 1OGES), 12, S, 7, tp.) MIGED), 15S), 240), D7, Bi-B), SS), 38, 36(2), 43(1), 45(1), 84(2) Copyright enquiries to the South African Museum Kopieregnavrae aan die Suid-Afrikaanse Museum ISBN 0 86813 117 2 Printed in South Africa by In Suid-Afrika gedruk deur the, Rustica Press, Bty. lids. Die Rustica-pers, Edms., Bpk., Court Road, Wynberg, Cape Courtweg, Wynberg, Kaap C973 A TOOTH-BEARING MAXILLA REFERABLE TO LYCORHINUS ANGUSTIDENS HAUGHTON, 1924 (DINOSAURIA, ORNITHISCHIA) By C. E. Gow Bernard Price Institute for Palaeontological Research, University of the Witwatersrand, Johannesburg, South Africa (With 7 figures) [MS accepted 5 December 1989] ABSTRACT The taxonomic status of the genera Lycorhinus and Heterodontosaurus, and the various specimens referred to them, is in an unsatisfactory state, due to incomplete preparation and description of otherwise good material, a tendency to diagnose specimens rather than species, and a lack of understanding of the anatomy and functioning of the teeth. A new tooth-bearing maxilla is described in detail, and both it and the type and only specimen of Lanasaurus scalpri- dens Gow, 1975, are referred to Lycorhinus angustidens Haughton, 1924, which is also restudied and reinterpreted here. Although detailed descriptions of the dentition of Hetero- dontosaurus tucki Crompton & Charig, 1962, have not yet appeared, L. angustidens and H. tucki are readily distinguishable on postcanine tooth morphology, angle of wear facets, and pattern of occlusion. Authors have ranked these differently specialized contemporary species as primitive and advanced; this practice may be questioned. CONTENTS PAGE Mr tnO GC WCEO Re ieee ae eta nee hae etna coe eh ras eon ee te 367 PMS se NWalM xd epee rn ee ON i ac PENA in AR wae te irc Bh ah 368 Lycorhinus angustidens ............... Bk ane ae 371 IBIS CUS SHO Marry ere rs Sree a as ee Ma Nt ary te a a 379 PNCKNOWICAEEMEMUSs arr sees eee ere ames sie ueAnon ies Gla Maen e oa eu § 379 IRCiCREINCCS tears Wr ee ee rn eel ee ee ae 5S 380 INTRODUCTION In 1984 James Kitching and the author collected a tooth-bearing left maxilla of an Early Jurassic ornithischian dinosaur. Any new material of these rare and incompletely known animals is to be welcomed. As the specimen was studied it became apparent that it is a larger specimen of the species Lanasaurus scalpri- dens Gow, 1975, and that both are referable to Lycorhinus angustidens Haughton, 1924. Impressions were accordingly made of the type of L. angusti- dens, thus enabling a detailed study of the three specimens. The study shows that Lycorhinus angustidens differs from Heterodontosaurus tucki Crompton & Charig, 1962. Most described material resides comfortably in one or the other of 367 Ann. S. Afr. Mus. 99 (10), 1990: 367-380, 7 figs. 368 ANNALS OF THE SOUTH AFRICAN MUSEUM these species, with the exception of Lycorhinus consors Thulborn, 1974. The much discussed specimen UCL A100 (Thulborn 1970) was probably correctly identified as L. angustidens; it cannot be grouped with L. consors as Abricto- saurus Hopson, 1975. The intention of this paper is to refrain as far as possible from discussing the work of previous authors but rather to concentrate on presenting new facts and inferences. The following abbreviations are used to indicate the repositories of the material studied: BP — Bernard Price Institute SAM — South African Museum UCL — University College, London. THE NEW MAXILLA Locality The farm Bamboeskloof, Lady Grey: 30°45’S 27°12'E, map reference Floukraal 3027CC. This locality is less than 15 km from two Heterodontosaurus tucki localities in the Herschel district (Crompton & Charig 1962; Santa Luca et al. 1976) and approximately 130 km from the type locality (Mount Fletcher) of Lycorhinus angustidens Haughton, 1924. Lanasaurus scalpridens Gow, 1975, was found about 250 km to the north (Golden Gate Highlands National Park). (See outcrop and locality map in Kitching & Raath 1984, fig. 1.) Material The specimen (BP/1/5253) had been exposed to the elements for some time prior to collection, with the result that the more delicate dorsal and anterior pro- jections of the maxilla are missing. The cutting edges of the teeth are also damaged and two crowns are missing. Preparation Only a little mechanical preparation was necessary. This was followed by treatment with thioglycolic acid, but this was discontinued as some damage to the specimen became evident; this was in any case only cosmetic preparation. Useful X-ray plates were made from the specimen. Description (Figs 1, 3-5) In occlusal view (Fig. 5) three important features are seen: the pit for recep- tion of the lower canine, the deep cheek region, and the pronounced curvature of the dental arcade. The dentition is fully developed and well worn, indicating that this was a mature individual. There are 14 functional teeth and a rudimen- tary 15th. The teeth invite several descriptive analogies; they are broadest linguolabially and closely packed like a row of kernels on a maize cob. (The TOOTH-BEARING MAXILLA OF LYCORHINUS ANGUSTIDENS 369 teeth of Lanasaurus scalpridens are broadest mesiodistally, but this difference can be attributed to age—compare Fig. 6.) Lingually and labially the crowns stand out from the roots (as evident in the photographs and indicated by dotted lines in Fig. 1). This swelling of the crowns is reciprocated by swellings of the roots at the mesial and distal ‘gum lines’ (for example see tooth 5). The final analogy is that the teeth have a symmetrical cold-chisel shape with a constant included angle of about 75° between wear facets and labial crown surfaces. (The narrower teeth of L. scalpridens have an included angle of 45°. Thus, although the angle is age dependent, it is worth stressing for comparison with Heterodon- tosaurus that the new maxilla belonged to a mature individual.) This included angle is a useful means of comparison as it is not affected by damage to the cutting edge of the crown and it eliminates subjective reference to vertical and horizontal axes. More anterior teeth have taller crowns, whereas mesiodistal crown width increases towards the back of the tooth row. The crowns have mesial and distal ridges on their lingual and labial surfaces separated by grooves from the main body of the crown; the distal ridges and grooves are more pronounced; on the lingual surface grooves persist for some time as wear proceeds and are thus important for assessing ages of teeth. In the following text teeth are referred to by numerals for convenience. Part of a wear facet is preserved on 2 but this tooth and 3 have lost much of the crown tips; 5 is lightly worn with part of the posterior groove still present; 6 is younger than its neighbours, being very little worn; 7 appears to have complex wear, but the two small basal facets were probably induced by trapped food rather than direct tooth on tooth contact (this is an old tooth); 8 is also well worn. Here one begins to see the pattern that persists from this point pos- teriorly, whereby adjacent wear facets on successive teeth were formed by a lower tooth in staggered occlusion. This pattern becomes very clear when the teeth are viewed normal to the wear facets (Fig. 1B—lines on the left in the figure separate inferred lower teeth). Steps between adjacent teeth in this occlu- sal view immediately show up young teeth and correlate with those teeth (in Fig. 1C) that retain traces of a posterior groove (notably 6 and 9; 12 is more worn). Tooth 9 is lightly worn, 10 is heavily worn with some blurring between the two main facets, possibly the result of polishing by food but also possibly the remnant of an earlier facet (as argued for Lycorhinus angustidens—see below). Tooth 11 has a well-developed pair of wear facets. Tooth 12, though a moder- ately young tooth (presence of posterior groove), is complicated, as its anterior facet is actually paired—the result of being opposed by two successive lower teeth. This tooth also has a large food polish facet. Teeth 13 and 14 are well worn (retention of the posterior groove on the latter possibly due to delayed eruption of a suitable antagonist). The rudimentary tooth 15 indicates that this is a fully elaborated, mature dentition. Teeth 6, 9 and 12 form a series of increasing age and are clearly younger than the two teeth that follow each. Replacement thus proceeds from back to 370 ANNALS OF THE SOUTH AFRICAN MUSEUM A B C e— SS 8 ——— —=> SS (f L i Clap) Fig. 1. Lycorhinus angustidens, BP/1/5253, left maxillary dentition. A. Labial view. B. Viewed normal to the wear facets. C. Lingual view. Note: In A and B hatching denotes broken areas. In C wear facets are hatched. Lines to the left of B indicated where lower teeth met each other. TOOTH-BEARING MAXILLA OF LYCORHINUS ANGUSTIDENS 3 front in the row. This is the same as the pattern described by Gow (1975) for Lanasaurus scalpridens (Fig. 6) and similar to that by Hopson (1975) for Lyco- rhinus angusticeps but with a slightly modified interpretation (see below). This has interesting implications. In the described dentition, teeth within triplets are arranged in order of increasing age from front to back, but after two more replacements a stage would be reached when this order would be reversed, and it was just such a stage that pertained in the maxilla that opposed the type dentary of Lycorhinus, as demonstrated below. X-rays of the specimen reveal root canals filled with dense haematite; these show that only a very thin layer of maxillary bone roofs the deep tooth sockets: the canal fillings terminate at the alveolar border. In the tooth sequence 6 to 13, root-canal fillings are present for all except tooth 7—this is probably a quirk of preservation as 7 should be the last in the series 13—10—7 to be replaced. X-rays of L. scalpridens reveal a full complement of roots. Although these X-rays show no signs of replacement activ- ity, this is not a firm indication that replacement had ceased. X-raying is a non- destructive technique that should be routinely applied and improved. Lycorhinus angustidens lanes ZS, 1 The specimen, SAM-—3606, has been well described by Hopson (1975, 1980) but was re-examined for this study owing to the possibility (now considered confirmed) that the new maxilla belongs to the same species. In order to take impressions, the specimen was thoroughly wet and the excess water removed with compressed air; a fabric-reinforced latex impression was then made, the first layer being of a very watery consistency. Three impressions were taken and all are equally good. The impressions were coated with a fine film of sublimating ammonium chloride, and it is these that were studied and photographed. The present interpretation differs slightly but significantly from that of Hopson (1980). The first point, which has not been stressed previously, is the marked curvature of the postcanine tooth row. When the canine is oriented with its cutting edges in a sagittal plane, the postcanine row curves back strongly labiad. (The new maxilla matches this curvature. The best way to see this is to orientate the photographs of the occlusal view with the first three teeth in the sagittal plane.) 3 The canine bears serrations on both edges (four per millimetre) as illus- trated by Hopson (1980, fig. 1). However, most of the distal edge of the tooth is missing. Postcanine 1 bears a small mesial cusplet and above it the margin of the crown is damaged (i.e. there may have been other cusplets). The posterior half of the labial surface of the crown is damaged and this looks like wear, as the damaged area has a sharp but smooth enamel edge (the worn area is covered with matrix grains firmly adhering to the dentine surface). Postcanine 2 has a worn occlusal edge to the crown; it also has a mesiolabial wear facet almost certainly formed when the erupted tooth made contact with — 372 ANNALS OF THE SOUTH AFRICAN MUSEUM the opposing upper tooth that had earlier been responsible for making the facet on 1M—the first indication of a staggered pattern of occlusion of upper and lower cheek teeth, which is argued in detail later. The condition of the labio- distal surface of 2 is not clear due to adherent matrix, but it does seem confluent with the anterior facet on 3. Postcanine 3 has a small distal cusplet high on the crown. There is some conchoidal fracture of the dentine at the tip of the crown but this does not mask two distinct wear facets dipping slightly away from each other. Tooth 4 is very similar though more worn and better preserved. Hopson’s (1980) interpretation of tooth 5, 1.e. one major wear facet and a small polished area, is accepted. Hopson interpreted tooth 6 in the same way but the larger lower facet is in fact in perfect contiguity with the anterior facet on tooth 7, thus demonstrating the presence of an upper tooth in overlapping occlusion with 6 and 7. Postcanine 7 is an old tooth; its posterior wear facet bears a wide, deep, smoothly rounded groove. This groove must have been formed by a step between adjacent edges of occluding upper teeth at different stages of wear. A small facet is present on the mesial edge of 8. By tilting the specimen it is poss- ible to see that this facet lies on the same arc as the distal facet on 7—these facets are thus attributable to the same upper tooth. The author is not convinced that there is sufficient evidence for the same situation pertaining between 8 and 9, but agrees with Hopson (1980) that it seems likely. The large wear facets on 8 and 9 have deliberately been left unhatched in Figure 2 because these teeth | 1cm | Fig. 2. Lycorhinus angustidens, SAM-—3606. Impression of left dentary teeth. Hatching indi- cates wear facets but has been deliberately omitted from teeth 8, 9 and 10. Teeth 8 and 9 bear striations, 9 and 10 have heels shown by shading, and 10 bears a raised ridge with the same orientation as the striations on 8 and 9. Wear facet on canine is on the lingual surface of the tooth. TOOTH-BEARING MAXILLA OF LYCORHINUS ANGUSTIDENS 373 bear patches of striations indicating direction of bite. These striations are helpful to understanding the bite, which is seen to have a posteriad component. It is important to note that the striations have the same orientation as the ridge on tooth 10 discussed below. That such striations are rare suggests a degree of imprecision in the bite such that occluding surfaces are continuously roughly pol- ished. Teeth 9 and 10 have heels worn into the base of their facets. The present interpretation of 10 and 11 differs from Hopson’s but is made with the benefit of the hindsight afforded by the new maxilla. Tooth 10 has a raised ridge between facets, such as would result if the edges of occluding uppers did not quite meet. The preserved portion of the 11th tooth was clearly part of a perfectly normal full-sized tooth; it is faceted and is raised labiad of 10, and it was thus opposed by the successor to the tooth responsible for the posterior facet on 10. It is suggested that the differences in wear facet orientation that Hopson (1980) recorded (supposedly increasingly horizontal with age) are illusory, as witness the continuity of facets on 7 with those of its neighbours. Indeed tooth 4 seem- ingly has the most nearly horizontal wear facets, but is less worn than tooth 5, which apparently has more oblique facets. This specimen represents a mature animal of a species characterized by very oblique wear facets. Three more teeth could have been present in the living dentary (see Fig. 3). This dentition contains ample evidence of a staggered occlusal arrangement between upper and lower teeth. In Figure 2 vertical lines above teeth indicate where upper teeth would meet each other. Some of the most interesting and instructive lower teeth are those that at first sight apparently do not conform to this staggered pattern. The best place to begin is with tooth 6: here it is seen that a second wear facet has started to encroach on a previously existing single facet—the new facet would continue to enlarge and migrate forward as indi- cated by the arrow. One can postulate that exactly the same thing would happen in time to tooth 5. Teeth 8 and 9 differ in that they have very well-developed 1cm | Fig. 3. Lycorhinus angustidens. Composite drawing of BP/1/5253 and SAM—3606. The speci- mens fit rather well and give an indication of the degree of incompleteness of the dentary tooth row. For reasons explained in the text, wear facets cannot be directly compared. 374 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 4. Lycorhinus angustidens, BP/1/5253. Above. Labial view. Below. Lingual view. Scale bar = 1 cm. TOOTH-BEARING MAXILLA OF LYCORHINUS ANGUSTIDENS 375 Fig. 5. Lycorhinus angustidens, BP/1/5253. Above. Occlusal view. Below. View normal to wear facets. Scale bar = 1 cm. 376 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 6. Lycorhinus angustidens (Lanasaurus scalpridens), BP/1/4244. Above. Labial view. Below. Lingual view. Scale bar = 1 cm. TOOTH-BEARING MAXILLA OF LYCORHINUS ANGUSTIDENS 37 Fig. 7. Lycorhinus angustidens, SAM-3606. Original above. Positive impression below. Scale bar = 1 cm. 378 ANNALS OF THE SOUTH AFRICAN MUSEUM single facets, and incipient facets on their anterior edges. These latter facets would migrate posteriad in time until these teeth reached the condition seen in tooth 10, which has two distinct but very well-worn facets. After this the tooth would be shed. The difference in direction of facet migration has little to do with position in the tooth row, but seems rather to be related to the extent of wear of the teeth. From the above, the following sequence can be inferred. (a) A single wear facet forms fairly symmetrically over the labial surface of the crown of a dentary tooth (as previously noted by Hopson such a facet is concave—the opposing upper teeth would be well worn and would present a convex surface). This is not apparent from Figure 1B and perceptions of facet curves change as the specimen is rotated about its longitudinal axis. At its best development each of a pair of facets on adjacent teeth is concave, hence a con- vexity is formed where they meet. This is seen in the occlusal stereophotograph (Fig. 5) between teeth 7 and 8, and 10 and 11. (b) A second facet forms posteriorly when a new upper tooth comes into occlu- sion. This facet migrates forward and eventually dominates the crown as the tooth anterior to it is shed (this facet extends further down the crown—quite obviously this must be so). (c) As a new tooth comes into occlusion in the anterior position, a third facet forms, this time on the anterior edge of the tooth, and this migrates backwards to result in the condition seen on tooth 10. At this stage the tooth would be replaced. This interpretation highlights, and is itself supported by, the pattern of trip- lets in the Lycorhinus jaw. Arranged from youngest to oldest, these are 5, 6 and 7, and 8, 9 and 10. This interpretation differs from that of Hopson (1975, 1980), who proposed the following triplets: 4, 5 and 6, and 7, 8 and 9. Turning to the anterior teeth, it appears that 4 is more worn than 3, but both have two facets, whereas 2 has a single (first wear stage) facet; thus these teeth conform to the pattern of triplets proposed here. We now have the interesting situation where both maxilla and dentary bear triplets of teeth that consistently range in age from front to back. For the lower jaw one can demonstrate that each triplet would require to be opposed by a battery of teeth in which the reverse situation pertained. To do this we can look at the hypothetical maxillary (M) triplet 6, 7 and 8 that occluded with dentary (D) teeth 6, 7, 8 and 9 at the time the bearer of the Lycorhinus type died. Tooth M6 was well ground in, M7 had only recently made contact with D8, whereas M8 may just have made contact with D9. We can also look at D9, 10 and 11: the oldest tooth in the next maxillary series should be M9—that fits; M10 should be mature but not as old as M9 and again this is borne out by the wear facet on D11. All this makes eminently good sense, as, if occluding teeth were to erupt together, the amount of attrition would presumably be greater and the teeth would wear faster. The pattern of wear on the teeth in the new maxilla has been frozen at a different stage in the cycle, which makes it look different and more difficult to TOOTH-BEARING MAXILLA OF LYCORHINUS ANGUSTIDENS 379 interpret. The teeth appear more regularly bifaceted. That the same sequence as demonstrated for the dentary still applies, is shown by the following teeth: 9 has first and second wear stage facets, 12 is unusual in that first, second and third stage facets are present, whereas 7 has a partially obliterated second wear stage facet and a well-developed third wear stage facet. (Polish facets on 7 and 12 are ignored.) DISCUSSION As a result of the foregoing description and analysis, tooth morphology, function and replacement are well understood, and the variation inherent in the system must lead to the conclusion that there is a high probability that Lana- saurus scalpridens and the new maxilla BP/1/5253 belong to the species Lycorhinus angustidens. The much discussed but poorly known specimen UCL A100 (Thulborn 1970) may also belong to this species. The very similar species Heterodontosaurus tucki is known from a complete skeleton (Santa Luca et al. 1976), a complete skull (Crompton & Charig 1962), and possibly an undescribed jaw fragment (Charig & Crompton 1974). Lycorhinus consors Thulborn, 1974, is an enigma, apparently possessing the teeth and wear pattern of Lycorhinus and the occlusal pattern of Heterodonto- saurus. Heterodontosaurus has more strongly ridged teeth with more transverse wear facets, and occlusion is nearly if not entirely opposite. Hopson (1980) demonstrated the same pattern of replacing triplets for Heterodontosaurus, as 1s now well documented for Lycorhinus. Hopson (1980) remarked that the Lanasaurus maxilla was larger than the maxilla of Heterodontosaurus. The new maxilla is still larger, and fits the Lyco- rhinus dentary rather well; thus Lycorhinus could well have been a larger animal than Heterodontosaurus. On the basis of the foregoing discussion, the maxilla BP/1/5253 is referred to Lycorhinus angustidens Haughton, 1924, and a revised synonymy for this species is presented below. Lycorhinus angustidens Haughton, 1924 Lycorhinus angustidens Haughton, 1924: 343-344, fig. 8. Thulborn, 1970: 236-241, figs 1-5. Abrictosaurus consors (Thulborn, 1974) Hopson, 1975: 304 (part. —-UCL A100 only). Lanasaurus scalpridens Gow, 1975: 336-339, text-figs 1-2, pl. 1. ACKNOWLEDGEMENTS I wish to thank the Director of the South African Museum, Dr M. A. Cluver, for the loan of the Lycorhinus type, Dr J. van den Heever who packed it so beautifully and Mrs Ann Lawton who kindly added the package to her family holiday luggage. 380 ANNALS OF THE SOUTH AFRICAN MUSEUM REFERENCES Cuaric, A. J. & Crompton, A. W. 1974. The alleged synonomy of Lycorhinus and Hetero- dontosaurus. Annals of the South African Museum 64: 167-189. Crompton, A. W. & Cuaric, A. J. 1962. A new ornithischian from the Upper Triassic of South Africa. Nature, London 196 (4859): 1074-1077. Gow, C. E. 1975. A new heterodontosaurid from the Redbeds of South Africa showing clear evidence of tooth replacement. Zoological Journal of the Linnean Society of London 57 (4): 335-339. Haucuton, S. H. 1924. The fauna and stratigraphy of the Stormberg Series. Annals of the South African Museum 12 (8): 323-497. Hopson, J. A. 1975. On the generic separation of the ornithischian dinosaurs Lycorhinus and Heterodontosaurus from the Stormberg Series (Upper Triassic) of South Africa. South African Journal of Science 71 (10): 302-305. Hopson, J. A. 1980. Tooth function and replacement in early Mesozoic ornithischian dino- saurs: implications for aestivation. Lethaia 13 (1): 93-105. KITCHING, J. W. & RaatH, M. A. 1984. Fossils from the Elliot and Clarens formations (Karoo sequence) of the Northeastern Cape, Orange Free State and Lesotho, and a suggested biozonation based on tetrapods. Palaeontologia africana 25: 111-125. SANTA Luca, A. P., Crompton, A. W. & Cuaric, A. J. 1976. A complete skeleton of the Late Triassic ornithischian Heterodontosaurus tucki. Nature, London 264 (5584): 324-328. THULBORN, R. A. 1970. The systematic position of the Triassic ornithischian dinosaur Lyco- rhinus angustidens. Zoological Journal of the Linnean Society of London 49 (3): 235-245. THULBORN, R. A. 1974. A new heterodontosaurid dinosaur (Reptilia: Ornithischia) from the Upper Triassic Redbeds of Lesotho. Zoological Journal of the Linnean Society of London 55 (2): 151-175. 6. SYSTEMATIC papers must conform to the International code of zoological nomenclature (particu- larly 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 trans- ferred 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 specimens 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, descrip- tion 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 Eliza- beth (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 preferably 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. The generic name should not be abbreviated at the beginning of a sentence or paragraph. 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. C. E. GOW A TOOTH-BEARING MAXILLA REFERABLE TO LYCORHINUS ANGUSTIDENS HAUGHTON, 1924 (DINOSAURIA, ORNITHISCHIA) 7x VOLUME 99 PART 11 FEBRUARY 1991 ISSN 0303-2515 K ‘ i b. “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 3 cm margins all round. First lines of paragraphs should be indented. Tables and a list of captions for illustrations should be typed separately, their positions indicated in the text. All pages should be num- bered 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 12x 18cm (19 cm including caption); the reduction or enlargement required should be indicated (and preferably uniform); orig- inals 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 caption; 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. If Letraset is used authors are requested to use Helvetica-style letter- ing, if possible. 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, 1969b; Jones 1971)’ ‘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, 1969b) 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 (according to the World list of scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses, volume number, part number 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 la vitalité des mollusques. Journal de conchyliologie 88 (3): 100-140. FiscHer, P. H., Duvat, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines. Archives de zoologie expérimentale et générale 74 (33): 627-634. Koun, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee region of Ceylon. Annals and Magazine of Natural History (13) 2 (17): 309-320. Koun, A. J. 1960b. Spawning behaviour, egg masses and larval development in Conus from the Indian Ocean. Bulletin of the Bingham Oceanographic Collection, Yale University 17 (4): 1-51. THIELE, J. 1910. Mollusca. B. Polyplacophora, Gastropoda marina, Bivalvia. In: ScHULTZE, L. Zoologische und anthro- pologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-Afrika ausgeftihrt in den Jahren 1903-1905 4 (15). Denkschriften der medizinisch-naturwissenschaftlichen Gesellschaft zu Jena 16: 269-270. (continued inside back cover) ANNALS OF THE SOUTH AFRICAN MUSEUM ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM Volume 99 Band February 1991 Februarie Part 11 Deel NSA | S BBdBe = ww! { ae Q. & Svouig novi NS A REVISED DESCRIPTION OF THE SKULL OF MOSCHORHINUS (THERAPSIDA, THEROCEPHALIA) By J. F. DURAND 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 ih, ACES, GB), CD, 4-5, 8, taal), SAB, 5, 7D), Ga, tat), TO), 8, O02, 7), IOMES), GD, 3, 7, Ga.) MCD), ICS), WO, 5), V7, SB), 226), 2, 36(2), 43(1), 45(1), 67(5), 84(2) Copyright enquiries to the South African Museum Kopieregnavrae aan die Suid-Afrikaanse Museum ISBN 0 86813 116 4 Printed in South Africa by In Suid-Afrika gedruk deur The ‘Rustica Press; Ptys,) etd:; Die Rustica-pers, Edms., Bpk., Court Road, Wynberg, Cape Courtweg, Wynberg, Kaap C970 A REVISED DESCRIPTION OF THE SKULL OF MOSCHORHINUS (THERAPSIDA, THEROCEPHALIA) By J. F. DURAND Bernard Price Institute for Palaeontology, University of the Witwatersrand* (With 17 figures) [Paper presented at the Palaeontological Society of southern Africa Symposium, Cape Town, September 1986 | ABSTRACT Certain aspects of the external morphology of the Moschorhinus skull have been misinter- preted or overlooked in previous studies. In this paper the external morphology of the posterior half of the Moschorhinus skull is discussed in detail. The bony elements forming the braincase and the morphology of the foramina, fossae and grooves visible on the exterior surface of the braincase are described, and their possible functions are discussed. Certain misconceptions con- cerning the morphology of the prootic, opisthotic, quadrate, pterygoid, squamosal and epipterygoid are corrected. In the light of these findings the taxonomic position of Moscho- rhinus relative to other therocephalians is discussed. CONTENTS PAGE MMO CICLO Ieee ere aer ent ees a erent e vA ah ONE 381 WiatenralvandsniGthiod Savy ete sesh erceerect cs eee SNe nee irRlemsl lance aes 382 DCS CHILO Meer ar crate et wiftiaceate cette any Siauccger a uate Alera 383 DISCUSSION Seer ee eee ey ieee ee eo en Mme ere he ee ek Eo 406 PXCKMOWIEASEMEN(S arene Gein oe mor caheern ae var etna nar aie renee Bak 410 IRRENETRSINCES Oe ie 1 neh Nears ae ane AE MONAT Sn a TU peer a ge ee 411 JNO OFREREUBIOIINS: aaron cscs veneer aCe ean OURO sae ORR MIR) atten ei een 413 INTRODUCTION The interrelationships of the Therocephalia and their relationship with the Cynodontia are not adequately known. Certain authors, such as Broom (1938), Brink (1951), Hopson & Crompton (1969), and Kemp (1982), are of the opinion that the cynodonts had a therocephalian ancestor, whereas others such as Romer (1969) and Kermack & Kermack (1984) have argued that the cynodonts arose independently of the Therocephalia, from a more primitive ancestor. Most of the former authors accepted scaloposaur ancestry for the cynodonts, whereas Kemp (1972, 1982) argued that the cynodont ancestor was closely related to the whaitsiids. To unravel therapsid phylogeny it is essential to know more about thero- cephalian morphology. The present study attempts to broaden our knowledge of * Present address: Geological Survey, Private Bag X112, Pretoria. 381 Ann. S. Afr. Mus. 99 (11), 1991; 381-413, 17 figs. 382 ANNALS OF THE SOUTH AFRICAN MUSEUM Moschorhinus kitchingi, which is an interesting therocephalian with a mixture of primitive and advanced characteristics. Moschorhinus seems to be more advanced than the pristerognathids and scaloposaurids, even though it has certain features in common with the gor- gonopsians. Moschorhinus shares more characteristics with the whaitsiids than with any other therocephalian group. Moschorhinus is also more primitive than the whaitsiids but does not seem to be ancestral to them. The taxonomic pos- ition of Moschorhinus will be discussed in detail later. The skull of Moschorhinus has been described by Broom (1920), Boonstra (1934), Brink (1959), and Mendrez (1974a). The elements forming the anterior half of the skull are well known from these descriptions. However, due to the poorly preserved braincase in most Moschorhinus specimens, or insufficient preparation thereof, certain misconceptions arose concerning the relations of the elements constituting the posterior half of the skull. Two Moschorhinus skulls were selected for this study. Although these skulls are somewhat damaged and distorted, the posterior parts of the skulls are in such a condition that, with careful preparation it was possible to discover a wealth of information that adds to our knowledge of the Moschorhinus skull. In this paper the elements constituting the posterior half of the skull and the inter- relationships between these elements are described, and the possible courses of certain blood-vessels and nerves are discussed. MATERIAL AND METHODS Two previously undescribed specimens were selected for this study: Moschorhinus kitchingi (Broom) BP/1/2788 and BP/1/4636. BP/1/2788 was found by J. W. Kitching in Stoffelton, Afdeel Native Trust (now part of KwaZulu), near Bulwer, Natal, in the Daptocephalus zone (Kitching 1977) (Dicynodon lacerticeps—Whaitsia Assemblage-zone—S.A.C.S. 1980). BP/1/4636 was found by J. W. Kitching on the farm Fairydale in the Bethulie district, Orange Free State in the Lystrosaurus zone (Kitching 1977) (Lystrosaurus—Thrinaxodon Assemblage-zone—S.A.C.S. 1980). Moschorhinus kitchingi, BP/1/2788 (Figs 1-4) Most of the matrix surrounding this skull had been removed with a hammer and chisel prior to this study. The matrix within the temporal cavities had not been removed. This distorted specimen is 21,5 cm long. The major parts of the skull roof and occiput are missing and the jugal and postorbital arches are damaged. Teeth are present in the damaged and distorted mandible. Aspects of this specimen’s teeth and mandible were used in the reconstruction of the lateral view of the skull (Fig. 14). Moschorhinus kitchingi, BP/1/4636 (Figs 5-11) The whole skull was prepared for this study by means of an air-hammer and engraving tool. This distorted skull is 25 cm long. Parts of its jugal and post- REVISED DESCRIPTION OF MOSCHORHINUS 383 orbital arches are missing and the occiput is damaged. Although this skull is severely damaged, it yielded vital information. The descriptions and reconstruc- tion of the posterior part of the skull and the dental formula are those of this specimen (Figs 12-17). DESCRIPTION The posterior half of the Moschorhinus skull consists of the following endochondral elements: the epipterygoid, prootic, opisthotic, quadrate, supra- occipital, exoccipital and basioccipital, and the following dermal elements: the squamosal, quadratojugal, jugal, interparietal, parietal, postorbital, tabular and pterygoid. The parasphenoid and basisphenoid are of endochondral and dermal origin. The lateral wall of the braincase can be seen within the jugal arch (Fig. 16). The large parietal forms the sharp-crested roof of the braincase and the dorsal border of the temporal fossa. The posterior wall of the temporal fossa is largely formed by the squamosal. The medial wall of the temporal fossa is formed by the epipterygoid, the prootic, the ventrolateral part of the pterygoid, the lateral part of the opisthotic, and anterior parts of the supraoccipital and interparietal. Several features relating to blood-vessels and nerves can be seen within the temporal fossa. The posterior surfaces of the parietal, interparietal and supraoccipital form the medial surface of the occiput, dorsal to the foramen magnum (Fig. 15). These elements are flanked by the tabulars, which cover part of the squamosal posteriorly. The lateral part of the occiput is formed by the posterior parts of the squamosal and the opisthotic. The foramen magnum is flanked by the exoccipi- tals. The ventromedial border of the occiput is marked by the basioccipital. Two of the most salient features of the occiput are the large post-temporal fenestra and the paroccipital fossa. EPIPTERYGOID (Figs 12-14, 16) In lateral view (Figs 14, 16), the flattened, blade-like epipterygoid can be seen. It contacts the parietal, supraoccipital and prootic dorsally and the ptery- goid, prootic and squamosal ventrally. The upper part of the processus ascendens (dorsal lamina—Mendrez 1972, 1974a, 1974b) expands anteriorly to form an anterodorsal process and posteriorly to form a posterodorsal process. The basal part of the epipterygoid expands anteriorly to form an anteroventral process and posteriorly to form a posteroventral process (Mendrez 1972, 1974a, 1974b). A small posterior apophysis is present on the posterior edge of the pro- cessus ascendens, which probably made contact with the lateral part of the base of the anterodorsal process of the prootic (unfortunately damaged in all Mos- chorhinus specimens examined). The ventromedial part of the posterodorsal pro- cess of the epipterygoid contacts the anterolateral part of the anterodorsal process of the prootic just above the contact of the posterior apophysis with the prootic, 384 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 1. Moschorhinus kitchingi. BP/1/2788. Dorsal view. Fig. 2. Moschorhinus kitchingi. BP/1/2788. Ventral view. REVISED DESCRIPTION OF MOSCHORHINUS 385 Fig. 3. Moschorhinus kitchingi. BP/1/2788. Lateral view. Fig. 4. Moschorhinus kitchingi. BP/1/2788. Oblique anterior view. 386 ANNALS OF THE SOUTH AFRICAN MUSEUM thus forming a circular foramen—the posterior foramen of the epipterygoid (‘foramen veineux’—Mendrez 1974a) (see Fig. 16). There is a shallow funnel- like indentation surrounding the foramen on the lateral surface of the epiptery- goid. A low ridge runs diagonally across the lateral surface of the epipterygoid from the tip of the posterodorsal process, passes anterior to the foramen, and terminates in the middle of the ventral part of the epipterygoid as a small tuber- osity. This posterior foramen of the epipterygoid should not be confused with the dorsal venous foramen (see discussion). The dorsal border of the epiptery- goid fits in snugly under the parietal. The central part of the dorsal border of the epipterygoid is overlapped laterally by the ventrolateral descending flange of the parietal. The tip of the posterodorsal process curves slightly downwards, away from the parietal, exposing the interparietal upon which the dorsal part of the posterodorsal process lies. The anteroventral process of the epipterygoid is quite small in relation to the posteroventral process. It originates anteroventrally from the base of the processus ascendens. The anteroventral process terminates anteriorly to the dorsolateral ridge of the pterygoid and is in confluence with the posterior corner of this ridge. The term anteroventral process of the epipterygoid is preferred to the ‘pterygoid process of the epipterygoid’ (Crompton 1955) since the whole of the ventral border of the epipterygoid contacts the pterygoid. The foot of the epipterygoid covers the dorsal surface of the anterolateral third of the quadrate ramus of the pterygoid (see Figs 12, 16). The posteroventral process of the epipterygoid originates at the base of the processus ascendens from where it flares out posterolaterally as an elongated, horizontal fan, overlying the middle third of the quadrate ramus of the ptery- goid. These two processes are confluent laterally and posteriorly but not medially and anteriorly. The posteroventral process is slightly raised medially along its whole length, producing a medially facing groove that originates under the vertically inclined anterior part of the posteroventral process. The groove tapers off as it approaches the posterior border of the posteroventral process. The term posteroventral process of the epipterygoid is preferred to the ‘quadrate process of the epipterygoid’ (Crompton 1955), since it is doubtful whether the epipterygoid actually did contact the quadrate in Moschorhinus. The posterior part of the posteroventral process of the epipterygoid is fan- shaped. The lateral half of the posterior border stretches across the dorsal surface of the quadrate ramus of the pterygoid. The posterior border contacts the anterior border of the anteroventral process of the squamosal medially, the contact being visible in dorsal and ventral views. This region of the epipterygoid forms part of the anterolateral corner of the pterygo-paroccipital foramen (see Jers 1172, 113),. ING): In lateral view (Fig. 16), the ventral suture of the epipterygoid runs in the middle of the lateral side of the structure formed with the quadrate ramus of the pterygoid. The suture dips anteriorly and then curves upwards delimiting the border of the anteroventral process. Posteriorly the suture runs diagonally REVISED DESCRIPTION OF MOSCHORHINUS 387 Fig. 5. Moschorhinus kitchingi. BP/1/4636. Dorsal view. Fig. 6. Moschorhinus kitchingi. BP/1/4636. Ventral view. 388 ANNALS OF THE SOUTH AFRICAN MUSEUM upwards in a straight line delimiting the ventral border of the posteroventral process laterally on the dorsal side of the quadrate ramus of the pterygoid. The processus ascendens of the epipterygoid juts upwards and slightly inwards. The middle part of the processus ascendens is relatively constricted in comparison with the dorsal and ventral parts, giving the epipterygoid an hour- glass shape in lateral view. The ventral part of the epipterygoid is directed out- wards posteriorly and inwards anteriorly. This closely reflects the orientation of the quadrate ramus of the pterygoid, whereas the dorsal part of the epipterygoid is more parasagittally inclined. A very distinct and large cavum epiptericum is present, bordered medially by the anteroventral process of the prootic and laterally by the epipterygoid. Certain nerves and veins traverse the cavum epiptericum (see discussion). | Fig. 7. Moschorhinus kitchingi. BP/1/4636. Left lateral view. Fig. 8. Moschorhinus kitchingi. BP/1/4636. Right lateral view. REVISED DESCRIPTION OF MOSCHORHINUS 389 Fig. 9. Moschorhinus kitchingi. BP/1/4636. Occipital view. PROOTIC (Figs 12-17) The prootic and opisthotic are two separate elements. The sutures dividing these bones will be discussed later. The prootic is a complex bone with five major processes. The terminology used by Mendrez (1972) to describe these processes will be followed here. The prootic meets the basisphenoid ventrally, the squamosal and opisthotic postero- laterally, the epipterygoid anterolaterally, and the supraoccipital dorsally. In lateral view (Fig. 16), two distinct ridges can be seen running diagonally across the prootic, more or less parallel to each other. The ridge running from the central process to the anterodorsal process is here called the central ridge of the prootic (the ‘delicate rising crest’ of Mendrez 1972: 205). The ridge running from the lip of the fenestra ovalis to the anteroventral process of the prootic, is here called the ventral ridge of the prootic (the ‘sharp crest—Mendrez 1972: 203; the ‘strong crest’ —Mendrez 1972: 205; 1974b: 76). In lateral view it can be seen that the prootic has two distinct anterior pro- cesses directed diagonally anterodorsally, viz: the anterodorsal process above and anteroventral process below. These two processes are separated by the incisura prootica. Olson (1944), Crompton (1955), Mendrez (1972) and others used these terms to describe the anterior part of the prootic. The anterior part of the anterodorsal process makes contact with the medial surface of the epipterygoid, whereas the anteroventral process passes medially to the epipterygoid, forming a large vacuity between it and the epipterygoid— 390 ANNALS OF THE SOUTH AFRICAN MUSEUM Fig. 10. Moschorhinus kitchingi. BP/1/4636. Detail of left temporal region. Fig. 11. Moschorhinus kitchingi. BP/1/4636. Dorsolateral view of left temporal region. REVISED DESCRIPTION OF MOSCHORHINUS 391 the cavum epiptericum, which will be discussed later. The anterodorsal process runs more or less parallel to the sagittal plane, whereas the anteroventral process points inwards. The anterodorsal process of the prootic (processus anterior superior— Siebenrock 1893; posterior prootic process—Boonstra 1934) is a flattened, broad, projection that originates more or less in the middle of the prootic. Its dorsal border is continuous with the dorsal border of the rest of the prootic, and its ventral border is a continuation of the central ridge of the prootic. The antero- dorsal process is rather broad posteriorly but tapers anteriorly, the thinnest part being its anterior edge, which contacts the posteromedial edge of the epi- pterygoid laterally. This region is damaged in the specimen described by Mendrez (1974a). The anteroventral edge of the anterodorsal process forms the posterior border of the posterior foramen of the epipterygoid. There is a slight lateral protrusion on the posterior part of the ventral border of the anterodorsal process, causing a ventrolaterally directed protuberance in the central ridge of the prootic. This is probably where the prootic made contact with the posterior apophysis of the epipterygoid, because of its inclination towards, and proximity to, the apophysis. Bordering the anterodorsal process of the prootic dorsally and meeting the anterodorsal process of the epipterygoid is the flat, finger-like anterolateral process of the supraoccipital. The suture between the supraoccipital and prootic is not continuous. Posterior to the ventral edge of the posterodorsal process of the epipterygoid, a small triangular gap is formed between the supraoccipital and prootic. This is the dorsal venous foramen commonly found in many therapsids (see discussion). The plane of the anterodorsal process is diagonally inclined in cross-section. The ventral border flares out laterally, whereas the dorsal border is medially inclined, reflecting the orientation of the epipterygoid. The anteroventral process of the prootic (processus anterior inferior— Siebenrock 1893; anterior prootic process—Boonstra 1934) originates below the contact between the posterior apophysis of the epipterygoid and the antero- dorsal process. This process is the ossified pila antotica (pleurosphenoid) (De Beer 1937; Olson 1944; Save-Sdderbergh 1947; Crompton 1955). The dorsal border of the anteroventral process curves upwards in a crescent shape. The anteroventral process is vertically inclined in cross-section and curves inwards anteriorly. The anteroventral process is traversed anteriorly by a horizontal groove. Above this shallow indentation, a low ridge runs from the posterior border of the incisura prootica anteriorly. This low ridge flares out anteriorly, forming two small, anteriorly jutting projections. The incisura prootica is wide and deep. It is bordered ventrally by the concave dorsal border of the anteroventral process and dorsally by the straight ventral border of the anterodorsal process. The foramen for the facial nerve (VII) is situated between the central and ventral ridges of the prootic. This foramen is nearer the former ridge and on the 392 ANNALS OF THE SOUTH AFRICAN MUSEUM dpSq ipSq O 3cm loe Fig. 12. Moschorhinus kitchingi. Dorsal view. same level as the ventral border of the anteroventral process of the prootic anterior to it, and the central process of the prootic posterior to it. There is a small ventral notch between the ventral border of the antero- ventral process and the braincase floor. Its posterior border is formed by the anterior border of the basal region of the prootic. The ventral border is formed by the basisphenoid. This is the same as the notch described by Crompton (1955) for the Scalaposauridae, though here it is more open anteriorly. The prootic has three prominent lateral processes of approximately the same length. The central process of the prootic (lateral process of the prootic— REVISED DESCRIPTION OF MOSCHORHINUS 393 mpSq ptpf Fig. 13. Moschorhinus kitchingi. Ventral view. Kemp 1972) can be seen in lateral (Fig. 16), dorsal, ventral and occipital views. It juts out laterally and slightly posteroventrally towards the squamosal. Its anterolateral corner contacts the posteromedial corner of the posteroventral process of the epipterygoid. The medial part of the anteroventral process of the squamosal (prootic process of the squamosal—Mendrez 1974b) contacts this central process in a complex manner: the distal part of the central process forms two flanges, one anterodorsally, the other posteroventrally, between which the thin medial blade of the anteroventral process of the squamosal is wedged. 394 ANNALS OF THE SOUTH AFRICAN MUSEUM Soc Fig. 14. Moschorhinus kitchingi. Lateral view. These two elements form an anterodorsally curving bar that forms the anterior border of the pterygo-paroccipital foramen. The dorsal crest-like border of this bar is a continuation of the central ridge of the prootic. The anteroventral border of this bar forms a sharp concave crest running from the dorsal lip of the fenestra ovalis medially to the posteromedial corner of the posteroventral process of the epipterygoid laterally. The distal part of the céntral process is spindle-shaped in cross-section. The base of the central process, however, is triangular in cross-section, because of a short, sharp crest that forms the posteromedial corner of the process. This crest originates on the posteromedial part of the central process, curves posteromedially and terminates on the anterior surface of the posteroventral process of the prootic. Behind the central process of the prootic, a more posteriorly inclined, flattened process, the posteroventral process of the prootic, originates. This process can be seen in occipital view. It contacts the opisthotic ventrolaterally. This unified structure forms the posterior wall of the pterygo-paroccipital foramen, the ventral border of the post-temporal fenestra and the anterior wall of the paroccipital fossa. The posteroventral process forms only the dorsomedial quarter of the posterior wall of the pterygo-paroccipital foramen and the medial half of the ventral border of the post-temporal fenestra. A third process, the posterodorsal process of the prootic, contacts the intermediate process of the squamosal dorsally. It can be seen in lateral view Pro REVISED DESCRIPTION OF MOSCHORHINUS 35 fm pdpPro Ic PvpPro ipOp ioc pvfpp qpOp tom Tipe Bs O 38cm mpep Fig. 15. Moschorhinus kitchingi. Occipital view. (Fig. 16). The posterodorsal process forms most of the anterodorsal rim of the post-temporal fenestra. This process tapers off from a relatively broad base medially to a point jutting laterally, terminating in the lateral part of the roof of the post-temporal fenestra. This process protrudes from under the intermediate process of the squamosal anteromedially but, as it progresses laterally, it is gradually covered by the intermediate process of the squamosal anteriorly. Its posteromedial half is covered by the posterodorsal process of the opisthotic. In lateral (Figs 16, 17) and ventral (Fig. 13) views, it can be seen that the basal region of the prootic is in sutural contact with the basisphenoid. The lateral suture runs diagonally from the floor of the braincase anterodorsally to the fenestra ovalis posteroventrally. The anterodorsal part of the lip surrounding the fenestra ovalis is formed by the basal region of the prootic (see Fig. 17). As Mendrez (1972, 1974a) has remarked, the opisthotic and prootic are two quite separate bones. A clearly distinguishable suture divides them. The posterodorsal process of the opisthotic does not make contact with the lateral border of the posterodorsal process of the prootic, nor is it visible in anterior view as Mendrez (1974a) stated to be the case. The posterodorsal process of the opisthotic covers the posteromedial half of the posterodorsal process of the prootic posteriorly. In occipital view, one can observe within the post-temporal fenestra a part of the suture dividing the posterodorsal processes of the prootic and the opisthotic. This suture runs vertically for a short distance, skirting the lateral border of the posterodorsal process of the opisthotic, and then curves medially along its ventral border to where the posteromedial borders of the posteroventral and posterodorsal processes of the prootic originate. Laterally to 396 ANNALS OF THE SOUTH AFRICAN MUSEUM this, the suture divides the medial lip formed by the posteroventral process of the prootic and the posteroventral flange of the paroccipital process of the opisthotic. Ventrally to this medial lip, the suture that divides the posteroventral process of the prootic from the anterodorsal flange of the opisthotic, runs laterally. It can be traced posteriorly in the paroccipital fossa (Mendrez 1972: 203) and anteriorly between the anterodorsal flange of the paroccipital process of the opisthotic and the posteroventral process of the prootic. The suture curves first ventrally then medially along the border of the posteroventral process of the prootic. The suture then skirts the base of the central process of the prootic anteroventrally and the posterior corner of the dorsal lip of the fenestra ovalis, before it enters the roof of the fenestra ovalis medially (see lee, WY). The concave, dorsal border of the prootic can be seen in lateral view, curving posterodorsally to meet the medial border of the intermediate process of the squamosal. The central part of the dorsal border of the prootic (dorsal limit of the ‘lame dorsale’ of the prootic—Mendrez 1974a) contacts the supra- occipital. The largest part of the ventral border of the lateral supraoccipital fossa is formed by the dorsal border of the prootic. (This fossa is discussed below. ) OPISTHOTIC (Figs 12-17) The opisthotic contacts the exoccipital and basioccipital posteromedially, the prootic anteromedially, the tabular, supraoccipital and squamosal dorsally, the squamosal, pterygoid and quadrate laterally, and the stapes ventrally. The opisthotic consists of a robust transverse bar—the paroccipital process, a T-shaped ventromedial tuberosity—the internal process, and a small, flattened dorsomedial projection—the posterodorsal process. This is the terminology Mendrez (1972, 1974a, 1974b) used to describe the opisthotic. The paroccipital process of the opisthotic is V-shaped in parasagittal section. This V-shaped process is formed by two flanges joined anteroventrally. The posteroventral flange is more massive than the anterodorsal flange. The V-shaped cavity formed by these two flanges is the paroccipital fossa of the opisthotic. This fossa can be seen in dorsal and occipital views (Mendrez 1972, 1974a, 1974b). The paroccipital process is laterally subdivided into two processes that can be distinguished in ventral view, viz: the mastoid process of the opisthotic posteriorly, and the quadrate process of the opisthotic anteriorly (see Fig. 13). The mastoid process is marked by a ventral ridge originating approximately in the middle of the paroccipital process and terminating near the bulbous lateral end of the mastoid process. This ventrolaterally curving ventral ridge adds to the robustness of the mastoid process. The mastoid process is thickest near its distal end where the ventral ridge terminates. The posterior margin of the mastoid process marks the posteroventral border of the rim of the paroccipital fossa. The posterolateral part of the mastoid process contacts the squamosal laterally, whereas its anterolateral part is free. A shallow indentation REVISED DESCRIPTION OF MOSCHORHINUS 397 doppf Isfo ae |p dvf dpSq pr eopc bspPt kPs tso virPt alpSoc adpPro papEpt crPro pdpPro pvpPro cpPro pasEpt avpPro_ ipro O 3cm Fig. 16. Moschorhinus kitchingi. A. Detail of left temporal region. B. Detail of left temporal region. 398 ANNALS OF THE SOUTH AFRICAN MUSEUM separates the mastoid and quadrate processes of the opisthotic and forms the so- called roof of the middle ear (‘toit de ’oreille moyenne’— Mendrez 1974a). The quadrate process of the opisthotic, which juts outwards anterolaterally, is thinner, broader and longer than the mastoid process. It becomes broader laterally, similarly to the mastoid process, to form a bulbous lateral end. The quadrate process projects further laterally than the mastoid process. The pos- terior third of the dorsolateral surface of the quadrate process of the opisth- otic contacts the quadrate process of the squamosal and the anterior two-thirds contact the posterior part of the quadrate ramus of the pterygoid. This can be seen in lateral (Fig. 16) and dorsal (Fig. 12) views. The lateral surface of the quadrate process of the opisthotic loosely articulates with the quadrate and its anteroventral surface loosely contacts the stapes. The anterodorsal flange of the paroccipital process (anterior wall of the paroccipital process—Mendrez 1972) contacts the posteroventral process of the prootic medially. This combined structure forms the posterior wall of the pterygo-paroccipital foramen (seen in dorsal view), the ventral border of the post-temporal fenestra and the anterodorsal border of the paroccipital fossa (both seen in occipital view). The anterior ridge of the paroccipital process, marking the anterior border of the opisthotic, originates at the anterolateral edge of the quadrate process of the opisthotic and terminates near the dorsal lip of the fenestra ovalis. The base of the central process of the prootic and the dorsal lip of the fenestra ovalis (also formed by the prootic) contact the anteromedial part of the opisthotic (see Fig. 17). The suture dividing the prootic and opisthotic has already been described. The opisthotic forms the posterior third of the fenestra ovalis. The suture dividing the opisthotic and the basioccipital can been seen in ventral and occipi- tal view. It emerges from the posteroventral corner of the fenestra ovalis and then turns medially across the ventral surface of the lip of the fenestra ovalis (Fig. 17). The suture then curves posteromedially behind the tuberculum spheno-occipitale, runs around the internal process of the opisthotic and then curves laterally after passing medially to the jugular foramen. On reaching the posteroventral lip of the jugular foramen, the suture extends into the jugular foramen in an anterodorsal direction. The internal process of the opisthotic can be seen in ventral and occipital views. It is formed by a ridge originating on the ventromedial part of the par- occipital process, curving and expanding ventromedially and terminating as a ventromedial tuberosity between the jugular foramen and fenestra ovalis. From this tuberosity a thin anterolateral and thicker, blunter posteromedial extension project. The posterior extension forms the ventral lip of the jugular foramen and the anterior extension forms the posteroventral corner of the lip of the fenestra ovalis (Fig. 17). There is a small groove dividing the anterior extension from the medial part of the opisthotic that forms the posterior border of the fenestra ovalis. Similarly the posterior extension is separated ventrally from the part of the opisthotic that forms the anterior part of the roof of the jugular foramen by REVISED DESCRIPTION OF MOSCHORHINUS 399 Bs eopc Fig. 17. Moschorhinus kitchingi. Detail of left fenestra ovalis. a shallow groove (see Fig. 17). This groove runs from inside the jugular foramen anteromedially, more or less parallel to the ventrolateral suture of the exoccipi- tal, and terminates on the posteromedial surface of the paroccipital process. The part of the opisthotic that is overlain by the posterior extension of the internal process projects into the jugular foramen and, with the exoccipital posteriorly, forms the roof of the jugular foramen. A short groove is present on the pos- terior face of the ridge and tuberosity, ventral to the groove at the jugular foramen (see Figs 13, 15). In occipital view, the suture between the anterolateral part of the exoccipi- tal and the posteromedial surface of the opisthotic can be seen. This suture runs from inside the jugular foramen, around the exoccipital ventrolaterally, and then dorsomedially. Also visible in occipital view is the suture dividing the medial corner of the posterodorsal process of the opisthotic and the ventro- lateral corner of the supraoccipital. It originates half-way along the dorsal border of the exoccipital. The posterodorsal process of the opisthotic (visible in occipital view) orig- inates medially to the paroccipital fossa and the lip (formed by the postero- ventral process of the prootic and the posteroventral flange of the par- occipital process) covering the dorsomedial part of the paroccipital fossa. The posterodorsal process of the opisthotic is a flat projection that curves dorso- laterally. It forms the posteromedial surface of the post-temporal fenestra. Its lateral border stretches as far medially as the lip covering the dorsomedial part of the paroccipital fossa (a medial indentation separates these two structures). The posterodorsal process of the opisthotic makes sutural contact along its dorsomedial border with the supraoccipital. This suture continues in a dorso- lateral direction as it follows the dorsal border of the posterodorsal process. The suture dividing the supraoccipital and tabular originates half-way along this border. Part of the dorsal border of the posterodorsal process contacts the 400 ANNALS OF THE SOUTH AFRICAN MUSEUM ventral border of the tabular. The lateral border of the posterodorsal process contacts the posteroventral part of the intermediate process of the squamosal. The short ventral border of the posterodorsal process of the opisthotic contacts the posterior part of the posterodorsal process of the prootic. The posterodorsal process of the opisthotic is not visible in anterior view as Mendrez (1974a) has observed in Moschorhinus kitchingi, SAM-—K118. It rather resembles the con- dition found by Mendrez (19746) in Promoschorhynchus platyrhinus, RC 116, where the posterodorsal process is completely covered anteriorly by the inter- mediate process of the squamosal and the anterodorsal process of the prootic. SQUAMOSAL (Figs 12-16) The squamosal is a large, complex bone with several processes; the termin- ology used by Mendrez (1972) will be used to describe these processes. The squamosal makes contact with the parietal, interparietal, supraoccipital, prootic, opisthotic and tabular medially, the pterygoid and epipterygoid anteriorly, and with the jugal and quadrate laterally. The major part of the posterior wall of the temporal fossa is formed by the three large medial processes of the squamosal. The laterally sloping dorsal border of the squamosal forms the posterodorsal border of the temporal fossa and anterior part of the lambdoid crest. The three medial processes occur one above the other (see Fig. 16). The anteroventral process (third squamosal process—Crompton 1955) is separated from the intermediate process (second squamosal process of Crompton) by the post-temporal fenestra. The inter- mediate and dorsal processes (first squamosal process of Crompton) are separated by the posterior fold forming part of the lateral supraoccipital fossa. The medial part of the broad dorsal process of the squamosal covers the posterolateral part of the parietal (see Fig. 12). The lateral surface of the dorsal process is confluent with that of the parietal, as are their dorsal and ventral borders. The dorsal process of the squamosal is fused to the tabular along most of its posterior surface. Its ventral border is marked by the fold forming the posterior part of the lateral supraoccipital fossa (see Fig. 16). Part of this fold is occupied by the anterior extension of the interparietal, which contacts the ventral border of the dorsal process posterodorsally. In lateral view it can be seen that the dorsal border of the intermediate process of the squamosal forms the ventral border of the above-mentioned fold and contacts the posteroventral area of the part of the interparietal that is laterally exposed. The anterior part of the intermediate process makes sutural contact with the anterolateral process of the supraoccipital dorsally and with the prootic ventrally. The medial part of the ventral border of the intermediate process of the squamosal forms a V-shaped notch in which the posterodorsal process of the prootic is wedged. This fused structure forms the dorsomedial part of the roof of the post-temporal fenestra, whereas the dorsolateral part of the roof is formed by the lateral part of the ventral border of the intermediate process. REVISED DESCRIPTION OF MOSCHORHINUS 401 In occipital view it can be seen that the medial part of the intermediate process contacts the ventrolateral part of the tabular, and the dorsolateral part of the posterodorsal process of the opisthotic ventromedially. In lateral view the ventral border of the intermediate process can be seen. It curves ventrolaterally in a crescent-shape, delimiting the dorsolateral rim of the post-temporal fenestra, until it becomes the dorsal border of the anteroventral process of the squamosal. The dorsal border of the anteroventral process forms the lateral part of the pterygo-paroccipital foramen. The medial blade of the anteroventral process of the squamosal (the prootic process of the squamosal— Mendrez 1974b) is wedged into a V-shaped notch formed by the lateral part of the central process of the prootic. The posteromedial border of the postero- ventral process of the epipterygoid contacts the anteroventral process of the Ssquamosal anteromedially. The whole anterolateral border of the anteroventral process of the squamosal, except for its distal end, contacts the posteromedial edge of the quadrate ramus of the pterygoid. This suture runs posterolaterally from the contact between the posterior part of the posteroventral process of the epipterygoid and the squamosal, to the suture connecting the squamosal and opisthotic. The laterally directed flange of the squamosal overlies most of the opisthotic, the only contact formed being where their distal borders meet (see Fig. 15). The ventrolateral part of the anteroventral process of the squamosal (the quadrate process of the squamosal—Mendrez 1974b) contacts the dorsal part of the posterolateral tip of the quadrate process of the opisthotic posterior to the squamosal-pterygoid contact (see Fig. 16). This suture is visible anteriorly to the quadrate notch in the squamosal. (This notch is described below.) The ventrolateral part of the squamosal seen in occipital view is the mastoid process of the squamosal. The suture connecting the dorsolateral border of the mastoid process of the opisthotic to the ventromedial border of the mastoid process of the squamosal is visible in occipital and ventral views. Except for these two above-mentioned distal sutures and the suture with the posterodorsal process of the opisthotic, the rest of the squamosal overlies, but does not contact, the opisthotic. It thus forms the lateral parts of the roof of the pterygo-paroccipital foramen and the roof of the paroccipital fossa. A deep, dorsally directed quadrate notch of the squamosal is present anteriorly, in the distal part of the squamosal; it houses the dorsal part of the quadrate (see Fig. 16). This notch is surrounded by the jugal laterally, the pos- terolateral part of the anteroventral process of the squamosal (which covers the quadrate process of the opisthotic) anteroventrally and the thick lip formed by the jugal process of the squamosal dorsally. A shallow indentation is present anteroventrally to this notch in the bones covering the quadrate process of the opisthotic, namely, the quadrate process of the squamosal and the posterior end of the quadrate ramus of the pterygoid. This indentation, the quadrate recess of the squamosal, is probably synonymous with that described by Kemp (1969) and with the squamosal recess (‘recessus squamosal’) of Mendrez (1974a). The 402 ANNALS OF THE SOUTH AFRICAN MUSEUM posterior part of the quadrate fits into this recess. The quadrates are lost in most of the Moschorhinus specimens, because they were loosely articulating bones with no sutural connections. In occipital view the following features can also be seen: the posteroventral border of the intermediate process of the squamosal curves ventrolaterally, forming the posterolateral border of the post-temporal fenestra and part of the posterolateral border of the paroccipital fossa. A ridge runs in a dorso-ventral plane on the mastoid process of the squamosal laterally to its contact with the opisthotic. A posteroventral facing indentation on the mastoid process of the squamosal borders this ridge laterally. Ventrally to this indentation and ridge, and medially to the posterior part of the quadrate, a notch is situated posteriorly between the two lateral processes of the opisthotic. The dorsolateral part of the squamosal (the jugal process) curves anteriorly to join the posterior end of the jugal. Unfortunately, in those specimens studied, the jugal arch is either lost or damaged to such a degree that a detailed descrip- tion is impossible. SUPRAOCCIPITAL AND INTERPARIETAL (Figs 14, 15) Both the supraoccipital and the interparietal are visible in occipital and lateral view. In occipital view, the broad supraoccipital contacts the interparietal dorsally, the tabulars dorsolaterally, the posterodorsal processes of the opis- thotics ventrolaterally and the exoccipitals ventrally. The exoccipitals cover the ventromedial part of the supraoccipital, except for a narrow gap between the exoccipitals where the supraoccipital forms the dorsomedial part of the roof of the foramen magnum. The suture between the tabular and supraoccipital runs diagonally in a dorsomedial direction from its origin at the junction of the ven- tral borders of the tabular and supraoccipital, to the dorsal border of the supraoccipital. The suture between the ventral border of the interparietal and the dorsal border of the supraoccipital is horizontal and short. A large, deep occipital indentation is present in the region of the interparietal. Two smaller, ventral indentations, forming part of the larger indentation, are present on the dorsal part of the supraoccipital. These flank a short ridge originating in the middle of the dorsal part of the supraoccipital and continuing dorsally on the interparietal. In occipital view, the interparietal is a small, laterally ovate bone bordered ventrally by the supraoccipital, laterally by the tabulars and dorsally by the parietal. The aforementioned occipital indentation causes the interparietal, parietal, medial part of the tabular, and the dorsal part of the supraoccipital to be set deeper than the rest of the surrounding elements. This indentation was for the attachment of certain neck muscles. In lateral view, ventral to the posteromedial angle of the temporal foramen (Mendrez 1974a: 80), is the lateral supraoccipital fossa, a large oval indentation bordered dorsally by the parietal, whose ventral border forms a concave over- REVISED DESCRIPTION OF MOSCHORHINUS 403 hang. The ventral border of this fossa is formed by the concave dorsal edge of the anterodorsal process of the prootic, and the anterior border by the antero- ventrally curving posterior edge of the posterodorsal process of the epi- pterygoid. The fossa tapers off posteriorly into a short, horizontal fold. The anteromedial part of the intermediate process of the squamosal forms the pos- terior border of the fossa and the ventral part of its posterior fold. The dorsal border of this fold is formed by the ventral border of the parietal and the dorsal process of the squamosal. The dorsal venous foramen is situated in the anterior region of this fossa. The anterior extensions of the supraoccipital and interparietal are visible within the lateral supraoccipital fossa. The anterolateral process of the supra- occipital lies at a more medial level than any of the surrounding elements, forming the medial wall of the lateral supraoccipital fossa. The anterolateral process of the supraoccipital contacts the parietal dorsally, the posterodorsal process of the epipterygoid anteriorly, the anterodorsal process of the prootic ventrally, the intermediate process of the squamosal posteroventrally, and the anterior extension of the interparietal posterodorsally. The dorsal venous foramen is visible ventral to the epipterygoid-supraoccipital contact and anterior to the prootic-supraoccipital contact. The anterior extension of the interparietal fills the posterodorsal corner of the lateral supraoccipital fossa. The interparietal is triangularly shaped and con- tacts the anterolateral process of the supraoccipital anteroventrally, the dorsal border of the intermediate process of the squamosal posteroventrally, the ventral border of the parietal anterodorsally, and the ventral border of the dorsal process of the squamosal posterodorsally. PARIETAL (Figs 12, 14-16) The parietal forms the posterodorsal part of the skull. It contacts the squam- osal and tabular posteroventrally, the prootic and epipterygoid ventrolaterally, the postorbital anterolaterally, and the frontal anteriorly. In occipital view the parietal is situated between the dorsomedial borders of the tabulars and the dorsal border of the interparietal. Its dorsal border is in confluence with those of the tabular and squamosal. These borders form the ventrolaterally curving, dorsal border of the occiput. The parietal has a pronounced sagittal crest. The anterodorsal rim of the temporal fossa is formed by an acute curving ridge on the posterodorsal and dorsolateral surfaces of the postorbital (see Fig. 12). These postorbital ridges bow posteriorly and are continued on the dorsomedial surface to produce the sagittal crest. The sagittal crest is widest at its origin, anterior to the parietal foramen. Posteriorly it becomes narrower and splits into two posterolaterally flaring lambdoid crests, which form the dorsomedial border of the occiput pos- teriorly and part of the posterodorsal rim of the temporal fossa laterally. The parietal, in dorsal view, has an hour-glass shape. It is broad and robust anteriorly, constricted in the middle above the epipterygoid, and forms two 404 ANNALS OF THE SOUTH AFRICAN MUSEUM posterolaterally flaring flanges that form the anterior parts of the lambdoid crests. The broad anterior part of the parietal contacts the frontal and postorbitals (see Figs 12, 14, 16). In dorsal view, it can be seen that the suture between the posterodorsal border of the frontal and the anterodorsal border of the parietal has a zig-zag arrangement. The anteroventrally sloping area between the anterior border of the parietal and the parietal foramen, i.e. the broad origin of the sagittal crest, is corrugated. In BP/1/4636, four small but distinct parasagittal ridges are present in this region. The two medial ridges join up with the sagittal ridge of the frontal anterior to them. This sagittal ridge runs on the dorsal surface of the skull, from the middle of the nasals, over the frontals, and joins the medial ridges of the parietal, which terminate on the slope anterior to the parietal foramen. The posteromedial flange of the postorbital and the anterolateral part of the parietal are separated by a suture that can be seen in dorsal and lateral view. The parietal foramen for the pineal organ is situated in the anterior part of the sagittal crest, on the same level as the posterior border of the posteromedial flange of the postorbital (see Fig. 12). The external opening of the parietal foramen is a narrow spindle-shaped slit, similar to the condition in the Moschorhinus specimens described by Brink (1959) and Mendrez (1974a). In lateral view, the vertically curving suture between the dorsal process of the squamosal and the posterior flange of the parietal can be seen. This suture closely reflects the occipital suture between the parietal and interparietal and the tabular. The ventral border of the parietal forms the roof of the braincase. The anterior third of the ventral border of the parietal does not make sutural contact with any bony elements, since this part of the braincase was unossified. The middle part of the parietal is triangular in cross-section. The lateral edges of the ventral border of this triangle contact the dorsal border of the epipterygoid and the dorsal part of the anterolateral process of the supraoccipital on each side. The short ventrolateral descending flange of the parietal overlaps the anterior two-thirds of the dorsal border of the epipterygoid. The posterior part of the ventral area of the parietal contacts the dorsal border of the interparietal. The ventrolateral edge of the parietal, posterior to its suture with the epipterygoid, forms the dorsal border of the lateral supraoccipital fossa. EXOCCIPITAL (Figs 12, 13, 15, 17) In occipital view, it can be seen that the exoccipital contacts the posteroventral part of the supraoccipital dorsally, the posteromedial part of the opisthotic laterally, and the posterior part of the basioccipital ventromedially. The concave medial side of the exoccipital forms the lateral wall of the foramen magnum. The exoccipital is divided externally into a flat anterodorsal part and a posteroventral boss. The ventrolateral half of the anterodorsal part overlaps the posteromedial part of the opisthotic and the dorsolateral half contacts the REVISED DESCRIPTION OF MOSCHORHINUS 405 posteroventral part of the supraoccipital. The anterodorsal part has two pronounced, acute rims; one medially, the other ventrally. The medial rim forms the dorsolateral lip of the foramen magnum. The ventral rim forms a ridge demarcating the posterodorsal lip of the jugular foramen. The posteroventral bosses of the exoccipitals form, together with the posterior part of the basioccipital, the occipital condyle. The exoccipitals form the dorsolateral parts of the occipital condyle and the basioccipital the ventral third. The occipital condyle in BP/1/4636 has a central indentation on its articular surface, not described before in Moschorhinus (Figs 12, 15). This indentation involves the posterodorsal part of the basioccipital third of the condyle and the posteromedial parts of the exoccipitals. The indentation and the associated lateral bosses hint at a double condyle condition, and are similar to those described by Watson (1913). The suture dividing the exoccipital and the basioccipital runs parasagittally from dorsomedially inside the foramen magnum (see Fig. 12), over the dorsal rim of the occipital condyle, and diagonally to a point ventrolaterally on the convex ventral rim of the occipital condyle. From here the suture runs anteriorly for a short distance on the ventral surface of the occipital condyle and then curves dorsolaterally, over the ventromedial lip of the jugular foramen, from whence it plunges into the jugular foramen in an anterodorsal direction. The dorsomedial side of the posteroventral boss of the exoccipital forms the posterior part of the ventrolateral wall of the foramen magnum (and the concave dorsomedial border of the occipital condyle), and its posterodorsal rim demarcates the posterior border of the foramen magnum. The ventrolateral side of the posteroventral boss forms the ventrolateral border of the occipital condyle, and the anterior border of the ventrolateral side forms the posterior lip of the jugular foramen (Fig. 13). Near the mouth of the jugular foramen, the lateral wall of the posteroventral boss is penetrated by two small foramina for the hypoglossal nerve (XII) (see Fig. 17). BASIOCCIPITAL (Figs 12, 13, 15-17) The basioccipital forms the posterior part of the basicranium and the ventral part of the occiput. In ventral view, the basioccipital contacts the basisphenoid anteriorly, the opisthotic laterally, and the exoccipitals posterolaterally. The anteroventral part of the basioccipital and the posteroventral part of the basisphenoid form the two spheno-occipital tubercles. The suture between the anterior border of the basioccipital and posterior border of the basisphenoid can be seen in ventral view. It dips diagonally in an anteroventral direction from the ventral lip of the fenestra ovalis, curves medially, and surrounds the posterior part of the spheno-occipital tubercle. The suture in the indentation between the two tubercles is set further posteriorly than those parts that bisect the tubercles. Of the four elements that form the lip surrounding the fenestra ovalis, viz: the basioccipital, opisthotic, basisphenoid and prootic, the basioccipital 406 ANNALS OF THE SOUTH AFRICAN MUSEUM contributes least to the formation thereof after the basisphenoid (see Fig. 17). Posterolateral to each spheno-occipital tubercle occurs the small laterally directed process of the basioccipitals that forms part of the ventral lip of the fen- estra ovalis. This small process is wedged between the posterolateral corner of the basisphenoid anteriorly and the anterior border of the internal process of the opisthotic posteriorly. The suture between the internal process of the opisthotic and the basioccipi- tal has a roughly diagonal arrangement. Seen in ventral view, it emerges from the fenestra ovalis and runs medially across its ventral surface for a short dis- tance. The suture turns posteromedially and skirts the anterolateral extension of the internal process of the opisthotic, then curves slightly posterolaterally around the ventromedial tuberosity and the posteromedial extension of the internal process of the opisthotic. Between the posterior border of the internal process of the opisthotic and the posteroventral edge of the occipital condyle, a short parasagittal flange sep- arates the ventrolateral border of the basioccipital and the ventromedial border of the occipital boss of the exoccipital. This suture can be seen in ventral view. The posterior part of the basioccipital has a rugose ventral surface. This rectangular part of the basioccipital, situated posteriorly to the level of the jugal foramina, forms the convex base of the occipital condyle. In dorsal view, the posterior part of the basioccipital that participates in the formation of the occipital condyle is visible as a thin strip flanked by the postero- ventral bosses of the exoccipitals. The basioccipital is wedge-shaped in occipital view; the broad concave base of the wedge is formed by the ventral surface of the basioccipital. The suture between the basioccipital and exoccipital has been described above. DISCUSSION In his paper on the Scaloposauridae, Crompton (1955) describéd a fused periotic and mentioned that Olson (1944) had not found any dividing suture in the periotic of those therocephalians he had studied either. The specimen known as “Therocephalian A’ (Olson 1944) was discovered in the Tapinocephalus zone. Its locality (Boonstra 1969; Kitching 1977) and size indicate that it is most prob- ably a pristerognathid. Olson (19385) described this specimen as having a periotic, but Boonstra (1954) and Van den Heever (pers. comm.) found a prootic and opisthotic in the Pristerognathidae. It has been shown by Van den Heever & Hopson (1982) that ‘Therocephalian B’ (Olson 1944) is actually a gor- gonopsian. Olson (1944) described a periotic in this specimen, as well as in the other gorgonopsians he studied. Authors such as Sigogneau (1970, 1974) described a prootic and opisthotic in the Gorgonopsia. However, it is possible that, in certain adult Gorgonopsia and Pristerognathidae, the suture between the prootic and opisthotic is difficult or impossible to detect in the region sur- REVISED DESCRIPTION OF MOSCHORHINUS 407 rounding the fenestra ovalis, but the co-ossification of these elements is not complete enough to consider these groups as having a periotic. The Scalopo- sauridae also have an opisthotic and prootic as described by Mendrez (1972), and not a periotic as Crompton (1955) described. Neither the Whaitsiidae (pers. obs.) nor the Moschorhinidae have a periotic, a clearly distinguishable prootic and opisthotic being present. It seems quite clear that the Therocephalia have a prootic and opisthotic, and not a periotic. The pterygoid process of the quadrate, as it is described by Mendrez (1974a), is actually not part of the quadrate but is the posterior end of the quad- rate ramus of the pterygoid. The quadrates have been lost in the Moschorhinus kitchingi specimen described by Mendrez (1974a) and the specimens described here. Judging from the shape of the squamosal recess, the quadrate was a rela- tively large, broad bone, approximately the same shape and size as that described by Mendrez (1974b) in Promoschorhynchus platyrhinus. In contrast to the condition understood in Promoschorhynchus, the quadrate in Moschorhinus seems to have had a small dorsal process that fitted into the quadrate notch of the squamosal (described below). This notch and slot arrangement allowed the qua- drate to articulate with the squamosal in a hinge-like manner. The quadrate lay upon the quadrate process of the squamosal and the lateral tip of the quadrate process of the opisthotic that is not covered by the squamosal or pterygoid. The shape of the quadrate notch of the squamosal indicates that the quadrate had a posterolaterally directed process that articulated medially with the lateral end of the quadrate process of the opisthotic and the quadrate process of the squamosal, and posteriorly with the posterior wall of the quadrate notch that is formed by the squamosal. This process is probably synonymous with the ‘squamosal process of the quadrate’ described by Mendrez (19745) in Promoschorhynchus platyrhinus. The posteroventral part of the quadrate probably contacted the squamosal in the same manner as in Promos- chorhynchus. If this was the case, the stapes would be longer than Mendrez (1974a) indicated in Moschorhinus kitchingi, and would extend laterally past the lateral end of the opisthotic. The lateral side of the quadrate would have contacted the medial part of the quadratojugal. Because of the absence of the stapes and quadratojugal, it is not possible to describe the relations between these elements and the quadrate. In her paper on Moschorhinus, Mendrez (1974a) referred to the opening dorsal to the posterior apophysis of the epipterygoid and the anterodorsal process of the prootic as the ‘foramen veineux’. It would seem that the dorsal part of the anterodorsal process of the prootic is damaged in her specimen. The posterior foramen of the epipterygoid and dorsal venous foramen are separated by the anterodorsal process of the prootic. This process is expanded dorsally to contact the supraoccipital and the posterodorsal process of the epipterygoid. (The term ‘dorsal venous foramen’ is preferred to venous notch—Boonstra 1934; Cox 1959; dorsal notch—Mendrez 1972; and venous foramen— Mendrez 408 ANNALS OF THE SOUTH AFRICAN MUSEUM 1974b, because it is a foramen completely surrounded by bone, and distinguishes between the two above-mentioned venous foramina.) The root of the trigeminal nerve (V) exits through the incisura prootica into the cavum epiptericum, which housed the trigeminal ganglion. From this ganglion the three trigeminal rami branched. The ramus opthalmicus passed mesial to the processus ascendens of the epipterygoid into the orbit. The ramus maxillaris and ramus mandibularis passed posterior to the processus ascendens into the temporal cavity (see Presley & Steel 1976). Certain authors (Brink 1957; Mendrez 1972, 1974a, 1974b) argue that the posterior apophysis of the epipterygoid divided these two rami. Others (Watson 1920; Kemp 1972) argued that both rami emerged through the foramen dorsal to the posterior apophysis (the posterior foramen of the epipterygoid). Crompton (1955) proposed a third alternative, namely that both rami emerged ventral to the posterior apophysis. The greatest part of the cavum epiptericum (and therefore also the trigeminal ganglion) lies below the level of the posterior foramen of the epipterygoid. Since the ramus maxillaris must have been directed ventrally, as is the ramus mandibularis, it is unlikely that it would first be deflected dorsally from the ganglion to pass through the posterior foramen of the epipterygoid and then ventrally towards the maxilla. It is more feasible that the ramus maxillaris passed, together with the ramus mandibularis, ventral to the posterior apophysis. The root of the abducens nerve (VI) usually exits through a foramen in the base of the anteroventral process of the prootic (ossified pila antotica) (see Haughton 1918; Goodrich 1958; Starck 1979). In Moschorhinus (Mendrez 1974a; pers. obs.) and Promoschorhynchus (Mendrez 1974b) this foramen is absent. Olson (1938a) mentioned that a foramen for the abducens nerve may be absent in certain gorgonopsids and would, in this case, pass anterior to the prootic. This seems to have been the case in Moschorhinus as well. The root of the facial nerve (VII) exits through its foramen between the central and ventral ridges of the prootic. No impression for the geniculate ganglion (gasserian ganglion—Mendrez 1972) could be found on the lateral surface of the prootic. There is no separate glossopharyngeal foramen in Moschorhinus. The glossopharyngeal (IX) exited through the jugular canal together with the vago- accessory (X and XI). This is a common feature in the therapsids (see Watson 1911; Haughton 1918; Kemp 1979). The hypoglossal (XII) enters into the jugular foramen through two foramina in its dorsomedial wall (see Fig. 17). The primary head vein of Moschorhinus probably ran mesially to the epi- pterygoid (see Goodrich 1958; Presley & Steel 1976), then laterally to the otic capsule where it received the vena cerebralis media. This united vessel passed ventral to the paroccipital process of the opisthotic (see Presley & Steel 1976). It is postulated that the vena cerebralis media most probably exited through the incisura prootica via the posterior foramen of the epipterygoid. The posterior foramen of the epipterygoid coincides with the dorsal part of the incisura REVISED DESCRIPTION OF MOSCHORHINUS 409 prootica, as the anteroventral border of the anterodorsal process of the prootic forms the foramen’s posterior border (see Fig. 16). If the vena cerebralis media passed through this foramen, it would be in line with the dorsal opening of the pterygo-paroccipital foramen through which it would pass ventrally to join the primary head vein. Many authors (Watson 1920; Parrington 1946; Cox 1959; Fourie 1974) described a groove running from the pterygo-paroccipital foramen to the incisura prootica (foramen for the trigeminal nerve—Parrington 1946). Most authors since Watson (1920) have claimed that the vena capitis lateralis ran in this groove, but it seems unlikely that this large vein could pass through the small posterior foramen of the epipterygoid in Moschorhinus. Moreover, the vena cerebralis media, which is a very important vein in extant reptiles (usually ignored by these writers), must have left some trace on the lateral wall of the prootic. The vena capitis dorsalis, which was situated in the sinus canal in cynodonts (Watson 1911), seems to have been expanded anteriorly to form a broad sinus that was situated in the lateral supraoccipital fossa in Moschorhinus. This sinus would have been confluent anteriorly with a vein that passed through the dorsal venous foramen. The vena capitis dorsalis would have been connected to the vein that passed through the pterygo-paroccipital foramen (probably the vena cerebralis media) and the vein that passed through the post-temporal fenestra (see Watson 1920; Parrington 1946; Cox 1959; Fourie 1974). The internal carotid artery ran anteriorly, ventral to the paroccipital process of the opisthotic, and entered the external opening of the parabasal canal. The stapedial artery branched off from the internal carotid in the proximity of the stapes. It probably ran in an anterodorsal direction in the depression below the central ridge of the prootic towards the cavum epipterycum, where it ramified into three branches, each of which accompanied a trigeminal nerve ramus (see O’Donoghue 1920). Moschorhinus has many primitive characteristics, such as large suborbi- tal vacuities similar to those of the pristerognathids (see Boonstra 1969), a gorgonopsid-like dentition (see Parrington 1955), and a robust skull compared to other Therocephalia (see Crompton 1955; Romer 1956; Brink 1959). Its epipterygoid is not as expanded, and therefore not involved to the same degree in the formation of the lateral wall of the braincase as are those of Promoscho- rhynchus (Mendrez 19746) or whaitsiids (Kemp 1972). No ossified ethmoid or orbitosphenoid elements could be found, as in gorgonopsids (see Olson 1944; Kemp 1969) or in whaitsiids (see Kemp 1972). Moschorhinus has large post- temporal fenestrae compared to cynodonts (see Watson 1920; Romer 1969; Kemp 1979), but this may be a characteristic peculiar to the Therocephalia (see Kemp 1972) because they are also larger than those of primitive therapsids (see Romer 1956). Derived and advanced characteristics in Moschorhinus include the par- occipital fossa of the opisthotic, which seems to be shared with all the other therocephalians except the pristerognathids (see Hopson & Barghusen 1986). 410 ANNALS OF THE SOUTH AFRICAN MUSEUM The epipterygoid and its relations with the surrounding bony elements are more advanced in Moschorhinus than in primitive Therocephalia (see Boonstra 1934; Crompton 1955; Mendrez 1972). Moschorhinus has a much larger epipterygoid than the gorgonopsids, scaloposaurids or certain pristerognathids. Its epiptery- goid makes sutural contact with the parietal and supraoccipital dorsally and the prootic posterodorsally, thus forming a substantial part of the lateral wall of the braincase. The posterior apophysis of the epipterygoid is present in certain Therocephalia (see Brink 1957; Kemp 1972; Mendrez 1974b), but in Moschorhi- nus it most probably made contact with the prootic, a condition unique amongst the Therocephalia. The posterior foramen of the epipterygoid is shared with the whaitsiids (see Kemp 1972). The venous notch of the primitive therapsids (see Boonstra 1934; Olson 1937; Mendrez 1972) is closed anteriorly by the epiptery- goid in Moschorhinus forming a venous foramen, as happens in whaitsiids (see Kemp 1972). Looking at all the above-mentioned characteristics it would seem that Mos- chorhinus was more advanced than the pristerognathids or scaloposaurids. Mos- chorhinus was more primitive than, but not ancestral to, the whaitsiids. Moschorhinus has a more primitive palatine region (see Mendrez 1974a), large suborbital vacuities, postcanines, and a robust skull. Theriognathus, on the other hand, does not possess suborbital vacuities nor postcanines and has a more deli- cate skull. Moschorhinus did not have an ossified orbitosphenoid or interorbital septum as in whaitsiids. Its epipterygoid is also much smaller and participates less in the formation of the lateral wall of the braincase (see Kemp 1972). The posterior foramen of the epipterygoid may have a different function in Moscho- rhinus than in whaitsiids, because in Moschorhinus it is formed differently and the posterior apophysis in the whaitsiid described by Kemp (1972) obstructs the passage between the posterior epipterygoid foramen and the pterygo-parocci- pital foramen. In whaitsiids the posteroventral process of the epipterygoid (quadrate ramus of the pterygoid—Kemp 1972) is also apparently much closer to the prootic than in Moschorhinus. Moschorhinus has too many derived characteristics to be a cynodont ancestor. It has paroccipital fossae, large post-temporal fenestrae and suborbital vacuities. It has few postcanine teeth, a small dorsal parietal foramen and, as in all Therocephalia, no stapedial foramen (see Mendrez 1974a). Furthermore, the posterior epipterygoid foramen is unique to the moschorhinids and the whaitsiids, and not homologous to the cynodont trigeminal foramen. ACKNOWLEDGEMENTS I wish to thank Dr C. E. Gow and Prof. J. W. Kitching for their help and advice, and my wife Juanita for typing the manuscript. REVISED DESCRIPTION OF MOSCHORHINUS 411 REFERENCES Boonstra, L. D. 1934. A contribution to the morphology of the mammal-like reptiles of the suborder Therocephalia. Annals of the South African Museum 31 (2): 215-267. Boonstra, L. D. 1954. The pristerognathid therocephalians from the Tapinocephalus-zone in the South African Museum. Annals of the South African Museum 42 (2): 65-107. Boonstra, L. D. 1969. The fauna of the Tapinocephalus zone (Beaufort beds of the Karoo). Annals of the South African Museum 56 (1): 1-73. Brink, A. S. 1951. Studies of Karroo reptiles. 1. Some small cynodonts. South African Journal of Science 47 (12): 338-342. Brink, A. S. 1957. On Aneugomphius ictidoceps Broom and Robinson. Palaeontologia afri- cana 4: 97-115. Brink, A. S. 1959. Notes on some whaitsiids and moschorhinids. Palaeontologia africana 6: 23-49. Broom, R. 1920. On some new therocephalian reptiles from the Karoo Beds of South Africa. Proceedings of the Zoological Society of London 2 (18): 343-355. Broom, R. 1935. On some new genera and species of Karoo fossil reptiles. Annals of the Transvaal Museum 18 (1): 55-72. Broom, R. 1938. The origin of the cynodonts. Annals of the Transvaal Museum 19 (2): 279-288. Cox, C. B. 1959. On the anatomy of a new dicynodont genus with evidence of the position of the tympanum. Proceedings of the Zoological Society of London 132 (3): 321-367. Crompton, A. W. 1955. A revision of the Scaloposauridae with special reference to kinetism in this family. Navorsinge van die Nasionale Museum 1 (7): 149-183. DE BEER, G. 1937. The development of the vertebrate skull. Oxford: Oxford University Press. Fouriz, S. 1974. The cranial morphology of Thrinaxodon liorhinus Seeley. Annals of the South African Museum 65 (10) 337-400. GoopricH, E. S. 1958. Studies on the structure and development of vertebrates. New York: Dover. Haucuton, S. H. 1918. Some new carnivorous Therapsida, with notes upon the brain-case in certain species. Annals of the South African Museum 12 (6): 175-216. Hopson, J. A. & BarGHusen, H. R. 1986. An analysis of therapsid relationships. Jn: Horton, N. ET AL. eds. The ecology and biology of mammal-like reptiles. Washington, D.C.: Smithsonian Institution Press. Hopson, J. A. & Crompton, A. W. 1969. Origin of mammals. Evolutionary Biology 3 (2): 15-72. Kemp, T. S. 1969. On the functional morphology of the gorgonopsid skull. Philosophical Transactions of the Royal Society (B) 256 (801): 1-8. Kemp, T. S. 1972. Whaitsiid Therocephalia and the origin of cynodonts. Philosophical Trans- actions of the Royal Society (B) 264 (857): 1-54. Kemp, T. S. 1979. The primitive cynodont Procynosuchus: functional anatomy of the skull and relationships. Philosophical Transactions of the Royal Society (B) 285 (1005): 73-122. Kemp, T.S. 1982. Mammal-like reptiles and the origin of mammals. London: Academic Press. KERMACK, D. M. & KerMack, K. A. 1984. The evolution of mammalian characters. USA: Kapitaan Szabo. KITCHING, J. W. 1977. The distribution of the Karoo vertebrate fauna. Memoirs. Bernard Price Institute for Palaeontological Research 1: 1-131. MENpDREZ, C. H. 1972. On the skull of Regisaurus jacobi, a new genus and species of Bauria- morpha Watson and Romer, 1956 (= Scalopsauria Boonstra, 1953), from the Lystrosaurus-zone of South Africa. In: JoysEy, K. A. & Kemp, T. S. eds. Studies in verte- brate evolution: 191-212. Edinburgh: Oliver & Boyd. MeENprEz, C. H. 1974a. Etude du crane d’un jeune specimen de Moschorhinus kitchingi Broom, 1920 (?Tigrisuchus simus Owen, 1876), Therocephalia, Pristerosauria, Moscho- rhinidae d’ Afrique australe. (Remarques les Moschorhinidae et les Whaitsiidae.) Annals of the South African Museum 64: 71-115. 412 ANNALS OF THE SOUTH AFRICAN MUSEUM MENDREZ, C. H. 1974b. A new specimen of Promoschorhynchus platyrhinus Brink, 1954 (Moschorhinidae) from the Daptocephalus-zone (Upper Permian) of South Africa. Palaeontologia africana 17: 69-85. O’DonoGHuE, C. H. 1920. The blood vascular system of the tuatara, Sphenodon punctatus. Philosophical Transactions of the Royal Society (B) 210 (377): 175-252. Otson, E. C. 1937. The cranial morphology of a new gorgonopsian. Journal of Geology 45 (5): 511-524. OLson, E. C. 1938a. The occipital, otic, basicranial and pterygoid regions of the Gorgonopsia. Journal of Morphology 62 (2): 141-175. OLson, E. C. 1938b. Notes on the brain case of a therocephalian. Journal of Morphology 63 (1): 75-86. Otson, E. C. 1944. The origin of mammals based on the cranial morphology of the therapsid suborders. Special Papers of the Geological Society of America 55: 1-136. PARRINGTON, F. R. 1946. On the cranial anatomy of cynodonts. Proceedings of the Zoological Society of London 116 (2): 181-197. PARRINGTON, F. R. 1955. On the cranial anatomy of some gorgonopsids and the synapsid inner ear. Proceedings of the Zoological Society of London 125 (1): 1-40. PRESLEY, R. & STEEL, F. L. D. 1976. On the homology of the alisphenoid. Journal of Anatomy 121 (3): 441-459. Romer, A. S. 1956. Osteology of the reptiles. Chicago: University of Chicago Press. Romer, A. S. 1969. The Chanares (Argentina) Triassic reptile fauna. V. A new chiniquodon- tid cynodont, Problesodon lewisi—cynodont ancestry. Breviora 333: 1-24. SAVE-SODERBERGH, G. 1947. Notes on the brain-case in Sphenodon and certain Lacertilia. Zoologiska Bidrag fran Uppsala 25: 489-516. SIEBENROCK, F. 1893. Zur osteologie des Hatteria-Kopfes. Sitzungsberichte der Osterreich- ischen Akademie der Wissenschaften (Matematisch-naturwissenschaftliche Klasse. Abteilung 1) 102 (1): 250-268. SIGOGNEAU, D. 1970. Contribution 4 la connaissance des ictidorhinidés (Gorgonopsia). Palaeontologia africana 13: 25-38. SIGOGNEAU, D. 1974. The inner ear of Gorgonops (Reptilia, Therapsids, Gorgonopsia). Annals of the South African Museum 64: 53-69. SOUTH AFRICAN COMMITTEE FOR STRATIGRAPHY (S.A.C.S.) 1980. Stratigraphy of South Africa. Part 1. (Comp. L. E. Kent). Handbook. Geological Survey. Republic of South Africa 8: 1-690. STARCK, D. 1979. Cranio-cerebral relations in recent reptiles. Jn: GANS, C., NORTHCUTT, R. G. & Uinski, P. eds. Biology of the Reptilia 9 (1): 1-38. London and New York: Academic Press. VAN DEN HEEVER, J. A. & Hopson, J. A. 1982. The systematic position of ‘Therocephalian B’ (Reptilia: Therapsida). South African Journal of Science 78 (10): 424—425. Watson, D. M.S. 1911. The skull of Diademodon, with notes on those of some other cyno- donts. Annals and Magazine of Natural History (8) 8 (45): 293-330. Watson, D. M. S. 1913. Further notes on the skull, brains and organs of special sense of Diademodon. Annals and Magazine of Natural History (8) 12 (68): 217-228. Watson, D. M. S. 1920. On the Cynodontia. Annals and Magazine of Natural History (9) 6 (36): 506-524. REVISED DESCRIPTION OF MOSCHORHINUS 413 ABBREVIATIONS adfpp —anterodorsal flange of the paroccipital process adpEpt —anterodorsal process of the epipterygoid adpPro —anterodorsal process of the prootic aelIp —anterior extension of the interparietal alpSoc —anterolateral process of the supraoccipital Ang —angular Art —articular avpEpt —anteroventral process of the epipterygoid avpPro —anteroventral process of the prootic avpSq —anteroventral process of the squamosal Bo —basioccipital Bs —basisphenoid bspPt —basisphenoid process of the pterygoid ce —cavum epiptericum cpPro —central process of the prootic crPro —central ridge of the prootic D —dentary dirPt —dorsolateral ridge of the pterygoid doppf —dorsal opening of the pterygo- paroccipital foramen dpSq —dorsal process of the squamosal dvf —dorsal venous foramen Ec —ectopterygoid Eo —exoccipital eopc —external opening of the parabasal canal Ept —epipterygoid F —frontal fEpt —foot of the epipterygoid fm —foramen magnum fo —fenestra ovalis 10C —indentation in occipital condyle Ip —interparietal ipOp —ainternal process of the opisthotic ipro | —incisura prootica ipSq —intermediate process of the squamosal J —jugal jf —jugular foramen jpSq —Jjugal process of the squamosal kPs —keel of the parasphenoid IG —lacrimal Ic —lambdoid crest Isfo —lateral supraoccipital fossa M —maxilla mpOp —mastoid process of the opisthotic mpSq —mastoid process of the squamosal —nasal Op — opisthotic J? —parietal Pal —palatine papEpt —posterior apophysis of the epipterygoid parf —parietal foramen pasEpt —processus ascendens of the epipterygoid pdpEpt—posterodorsal process of the epipterygoid pdpOp —posterodorsal process of the opisthotic pdpPro—posterodorsal process of the prootic pfEpt —posterior foramen of the epipterygoid pfo —paroccipital fossa Pm —premaxilla Po —postorbital ppOp —paroccipital process of the opisthotic pr —parasphenoid rostrum Prf — prefrontal Pro —prootic Pt —pterygoid ptf —post-temporal fenestra ptpf —pterygo-paroccipital foramen pvipp —posteroventral flange of the paroccipital process pvpEpt —posteroventral process of the epipterygoid pvpPro —posteroventral process of the prootic qnSq —dquadrate notch of the squamosal qpOp —dquadrate process of the opisthotic qr —quadrate recess qrPt —dquadrate ramus of the pterygoid Sm —septomaxilla Soc | —supraoccipital Sq —squamosal SV —suborbital vacuity al —tabular tpPt |—transverse process of the pterygoid tso —tuberculum spheno-occipitale Vv —vomer vidfP —ventrolateral descending flange of the parietal virPt —ventrolateral ridge of the pterygoid vn —ventral notch vrPro —ventral ridge of the prootic VII | —foramen for the facial nerve (VII) XII © —foramina for the hypoglossal nerve (XII) at the mouth of the jugular foramen BP/1 —Bernard Price Institute for Palaeontology catalogue number RC —Rubidge collection SAM —South African Museum ——. / re’ y a, ~~ oe fi om ; a ' ; TORT. % ' oF os a a ; hte! ‘4 : , ; , . ..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 preferably 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. The generic name should not be abbreviated at the beginning of a sentence or paragraph. 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. J. F. DURAND A REVISED DESCRIPTION OF THE SKULL OF MOSCHORHINUS (THERAPSIDA, THEROCEPHALIA). TY. - VOLUME 99 PART 12 MARCH 1991 | ISSN 0303-2515 em THSOM ages GW JUN 9 © 1991 OF THE SOUTH AFRICAN | 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 3 cm margins all round. First lines of paragraphs should be indented. Tables and a list of captions for illustrations should be typed separately, their positions indicated in the text. All pages should be num- bered 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 xX 18cm (19 cm including caption); the reduction or enlargement required should be indicated (and preferably uniform); orig- inals 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 caption; 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. If Letraset is used authors are requested to use Helvetica-style letter- ing, if possible. 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, 1969b; Jones 1971)’ ‘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, 1969b) 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 (according to the World list of scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses, volume number, part number 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 la vitalité des mollusques. Journal de conchyliologie 88 (3): 100-140. FISCHER, P. H., DuvaL, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines. Archives de zoologie expérimentale et générale 74 (33): 627-634. Koun, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee region of Ceylon. Annals and Magazine of Natural History (13) 2 (17): 309-320. Koun, A. J. 1960b. Spawning behaviour, egg masses and larval development in Conus from the Indian Ocean. Bulletin of the Bingham Oceanographic Collection, Yale University 17 (4): 1-51. THIELE, J. 1910. Mollusca. B. Polyplacophora, Gastropoda marina, Bivalvia. In: SCHULTZE, L. Zoologische und anthro- pologische Ergebnisse einer Forschungsreise im westlichen und zentralen Stid-Afrika ausgefiihrt in den Jahren 1903-1905 4 (15). Denkschriften der medizinisch-naturwissenschaftlichen Gesellschaft zu Jena 16: 269-270. (continued inside back cover) ANNALS OF THE SOUTH AFRICAN MUSEUM ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM Volume 99 Band March 1991 Maart Part 12 Deel GEOGRAPHY AND CLIMATOLOGY OF THE LATE CARBONIFEROUS TO JURASSIC KAROO BASIN IN SOUTH-WESTERN GONDWANA By JOHAN N. J. VISSER 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 il, DOES, 53), 302, 25, 3B tan), SS, 8, 7-9), G(s tp), 71=4)5 8, OGD: a0 E38) is 11D seat prin 14(1-2), 115425), 240; 5), 27031123) 326) 33s 36(2), 43(1), 45(1), 67(5), 84(2) Copyright enquiries to the South African Museum Kopieregnavrae aan die Suid-Afrikaanse Museum ISBN 0 86813 120 2 Printed in South Africa by In Suid-Afrika gedruk deur The Rustica’ Press; Bty> td? Die Rustica-pers, Edms., Bpk., Court Road, Wynberg, Cape Courtweg, Wynberg, Kaap D118 GEOGRAPHY AND CLIMATOLOGY OF THE LATE CARBONIFEROUS TO JURASSIC KAROO BASIN IN SOUTH-WESTERN GONDWANA By JOHAN N. J. VISSER Department of Geology, University of the Orange Free State, Bloemfontein (With 9 figures) [Paper presented at the Palaeontological Society of southern Africa Symposium, Cape Town, September 1986 | ABSTRACT The major late Palaeozoic to early Mesozoic basins of south-western Gondwana were located on a platform partly surrounding a highland interior. A palaeo-Pacific Ocean formed the margin of the platform. The Karoo Basin had an oblong shape with a long axis of more than 2 000 km but, following the late Palaeozoic glaciation, it changed in size to small enclosed fluvial basins during the early Mesozoic. A migrating tectonic region, attributed to subduction of the palaeo-Pacific plate, caused the shrinking of the basin. The climate in the Karoo Basin during the late Carboniferous varied from polar to cold. The early Permian cold stage showed distinct climatic fluctuations, resulting in glacials and interglacials, and a short-term warm period during which part of the Prince Albert Formation was deposited. Temperatures rose sharply during the late Permian and early Triassic. An abnormal cold, wet, middle Triassic resulted in the deposition of the Molteno Formation. During the late Triassic and early Jurassic, warm to warm desert conditions prevailed. Tectonic- ally and geographically induced regional climatic patterns in the Karoo Basin were super- imposed on global climatic trends. The basin was also consistently positioned at high to inter- mediate latitudes, suggesting that the principle of uniformitarianism cannot be applied without constraint. CONTENTS PAGE RMtROGUCH OME eet ee ye ee Sere ano Ree ae eek nee as 415 alacOPCOsraphyerese eto een ae Meranda a Mit Bees fe 416 alacochimatolopyna arn hinge sare iad Re oe via ea a aN 425 CONCUSSIONS eRe eras ct RNs oh aa s Benes 429 ENE THOWMSUTSTUA TIONS 5 Se Ubi Gate Aa oo 50 Bom Enc oe 430 INCISIONS 5 008 6 weme eas CHEE Ch atm oOn Oe One ha 430 INTRODUCTION Knowledge on the configuration of the late Carboniferous to Jurassic Karoo Basin in space and time is absolutely necessary in understanding the distribution of lithofacies, climatic zones, and biological provinces extending beyond the domain of the southern African continent. At the present stage of our know- ledge of the Palaeozoic, it is impossible to separate climatology from geography (Spjeldnaes 1981). Furthermore, sedimentation in the Karoo Basin was primar- ily controlled by the palaeoclimate and tectonism. 415 Ann. S. Afr. Mus. 99 (12), 1991: 415-431, 9 figs. 416 ANNALS OF THE SOUTH AFRICAN MUSEUM The objectives of this study are to define the position of the Karoo Basin in south-western Gondwana; to illustrate basin evolution from the late Carbonifer- ous to the early Jurassic; to draw a macro-scale climatic curve for Karoo sedimentation; and to focus attention on anomalous climatic—geographic relationships. For the palaeogeographic and palaeoclimatological analysis, raw data were taken from Tankard et al. (1982), Smith (1984), Visser (1983, 1984, 1987) and Anderson & Anderson (1985), as well as from field notes made by the author over several years of study on Karoo rocks. The palaeogeographic reconstruc- tions are based on palaeotopographic maps, sediment dispersal patterns, thickness of stratigraphic units, depositional environments and lithofacies. In the interpretation of the palaeoclimate, use was made of palaeolatitudinal maps (Irving 1977; Smith et al. 1981; Hallam 1985) and the lithology of the Karoo rocks, as well as their fossil content. Although maps based on palaeomagnetic evidence suffer from considerable uncertainties in areas where reliable determi- nations are scarce, the apparent polar wander curve for the late Carboniferous to Jurassic shows a fairly consistent trend from various sources of literature. A modified Gondwana reconstruction, based on one by Norton & Sclater (1979), was used for south-western Gondwana. In this reconstruction the Falk- land Islands were considered as part of a rotated microplate and were repositioned alongside the Transkei coast of southern Africa to achieve a better fit for the palaeomagnetics, geology and palaeontology between the two regions (Mitchell et al. 1986; Visser 1987). In the palaeoclimatic reconstruction a pro- visional stratigraphic time scale for the Karoo Sequence, based on fossil evi- dence, depositional rates and isotopic age determinations, was used. However, more refined age data are needed, particularly for the Ecca Group above the Whitehill Formation, which forms a highly significant basin-wide climatic marker at the end of the early Permian. The age of the Dwyka Formation was partly taken from Loock & Visser (1985). The subdivision of the Palaeozoic and Meso- zoic follows that of Harland et al. (1982), except for the Triassic where an informal subdivision of ‘early’, ‘middle’ and ‘late’, based on climatic trends, is used. PALAEOGEOGRAPHY Karoo Basin in south-western Gondwana To understand the tectonic evolution of the Karoo Basin, it is essential to briefly refer to the regional geography of south-western Gondwana during the late Palaeozoic. This part of Gondwana consisted of southern South America, southern Africa, Falkland Islands, East Antarctica, and other microplates in West Antarctica (cf. Storey et al. 1988), as well as inundated microplates in the southern Atlantic Ocean (Fig. 1). South-western Gondwana consisted of an elevated continental interior (e.g. Transvaal and Windhoek highlands) with trough-faulted, intracratonic basins 417 GEOGRAPHY AND CLIMATOLOGY OF THE KAROO BASIN ‘(6L61) 1218[9S 2 UOVION Joye poyipow UOTIONIJSUODOT BUBMPUOL) “OIOZOVEIed 9] OY} sulinp BUBPMPUOL) UIO}JSOM-YINOS UI SUISEG 194JO pUk OOIkY OY], CT ‘SIq Ww» OOS a aoe Nf Ss. AABDWIMOG| 7 Sey OF 7 A1B ld Ssh » \ —! . \ ~e, A» e SS —- Miu SNISV4 See Se ee Boone. eS 2 ae oem e ammo oo ° a N74 WOO Nel Barrie MOOS! SW, e © 0%e%,0°? aya ae O ¥y "y oe €>~- bare 7 ~ / S ee HLYOMS 114 es / : oe 9 ee Nees al ies A N ; ~ eoee PES MN ay MIS ES ANT es Dd Saar ds ROOT Cais ENS Seg ! i ESS NS h o— ! | ) 1 b ulBuew uiseg “Saf , i < ! ey i & 2 000 km) east-west depository extending into Antarctica (Fig. 3). It was bounded on the north by highlands, extending from East Antarctica across southern Africa to South America, and on the south by an alpine-type mountain range (proto-Precordillera). The northern mountainous plateau probably attained elevations of 2 000 to 3 000 m, whereas the alpine-type mountains were about 1 500 m above sea-level (Martin 1981; Visser 1987). A palaeo-escarpment, up to 400 m high, separated the northern highlands from the basin in the south. The subsidiary intracratonic Kalahari Basin was fault controlled. Both basin and highlands were completely ice covered for most of the time, but the western sea arm probably only had a sea-ice cover during winter. During two interglacials this sea transgressed eastwards, almost up to the Falkland EAST ANTARCTICA Magnetic South + Pole © ee CP FALKLAND Shallow {¢ “AI SLANDS sea Major source area |ce-f low direction 500 km Fig. 3. The Permo-Carboniferous Dwyka basin during maximum glaciation. Palaeolatitudes (at +290 Ma) after Smith ef al. (1981). 420 ANNALS OF THE SOUTH AFRICAN MUSEUM Islands, along the basin axis that was isostatically depressed by the weight of the ice as well as accumulated glacial debris. At the beginning, ice flow into the basin was mainly from the north and south. Where the ice flowed over bedrock, its thickness at the basin centre was probably in the order of 4 000 m but later, when the ice advanced over diamicton, its thickness decreased to about 1 000 m. Ice thickness over the northern highlands at maximum glaciation varied from about 2 600 to 3 000 m (Visser 1987). The early Permian post-glacial basin shows rapidly changing geography (Fig. 4). After ice retreat only small local ice caps remained on the highest mountains. The inundation of large parts of south-western Gondwana can be attributed to a combination of sea-level rise (Stavrakis (1986) suggested a rise of 100 to 150 m on deglaciation) and isostatic depression. Although post-glacial rebound took place, the depth of isostatic subsidence was such that this uplift did not elevate the sediment—water interface above sea-level. The Irati Sea, which covered the Parana Basin, was in continuity with the Whitehill sea in the Karoo Basin (Oelofsen 1981). An eastern limit for the Whitehill sea is defined by the absence of black shales on East Falkland (pre- _ SOUTH sais: Shallow sea Fluvial and deltaic deposition Region of isostatic rebound Fig. 4. Post-glaciation (+260 Ma) palaeogeography of the Karoo Basin and surrounding areas. Shoreline of the Irati and Whitehill seas modified after Oelofsen (1981). Palaeolatitudes after Smith et al. (1981). GEOGRAPHY AND CLIMATOLOGY OF THE KAROO BASIN 421 rotation and drift), but the original extent of the sea to the south-west is unknown. The shallow seas had a typical fjord coastline in the north, whereas in the south uplift along the proto-Precordillera caused northward regression of shorelines, especially in South America. Water conditions ranged from fresh through brackish to normal marine, depending on the rate of meltwater inflow from the mountains. Isostatic rebound along sections of the northern highlands caused extensive erosion of the glacial deposits and basement rocks. Debris was deposited as deltaic and fluvial beds in fjord heads and shallow embayments (Fig. 4) that were favourable for coal formation (Falcon 1986). Epicontinental Ecca Basin The large marine to non-marine basin formed a transition from an open shelf to an enclosed basin with major source areas in the south, west and north Source area Palaeocurrent direction Fig. 5. The later Permian Ecca basin. Palaeolatitudes (at +255 Ma) after Smith et al. (1981). NA = Namibia, BO = Botswana, ZI = Zimbabwe, MO = Mogambique, FI = Falkland Islands, RSA = South Africa. 422 ANNALS OF THE SOUTH AFRICAN MUSEUM (Fig. 5). Subsidence of part of the northern highlands led to large-scale inun- dation of the region. Water depths in the south were up to 700 m (Kingsley 1981), although it was much shallower in the north. The bottom sediments con- sisted mostly of black mud. The presence of abundant pyrite in the black shale is also suggestive of highly reducing benthic conditions, probably well above the water—sediment interface. Such a soft muddy bottom with toxic conditions would have been unfavourable to a benthic fauna and, if such conditions were associated with a low pH, the destruction of all organisms settling on the bottom after death would have occurred. This could account for the scarcity of body fossils in these rocks. A new development in the basin evolution was the appearance of a source area in the west that may be attributed to uplift caused by hot spot migration (cf. Anderson 1982) preceding the break-up of Gondwana. The prominent southern mountains, located a few hundreds of kilometres from the present outcrop area of the Ecca Group, consisted of low-grade metamorphic rocks Source area Palaeocurrent direction \ 2 4 i Flood-‘plain basin ; -- / ~ Basin margin Fig. 6. Early Triassic intracratonic Beaufort basins. Palaeolatitudes (at +245 Ma) after Smith et al. (1981). NA = Namibia, BO = Botswana, ZI = Zimbabwe, MO = Mocambique, RSA = South Africa, FI = Falkland Islands. GEOGRAPHY AND CLIMATOLOGY OF THE KAROO BASIN 423 associated with synorogenic intrusive and extrusive magmatic activity (Elliot & Watts 1974; Kingsley 1981). Intracratonic Beaufort basins During the Triassic, south-western Gondwana underwent a dramatic change in basin evolution with the formation of enclosed intracratonic basins. In addition to the major Karoo Basin in the south, a much smaller one developed in northern Zimbabwe (Fig. 6). The true extent of the basins will never be known, except where prominent highlands defined their margins. These fluvial basins had a largely centripetal drainage but the size of the rivers depositing the widespread flood-plain muds and silts suggests there must have been a basin outflow, probably towards the north-west and west (Botswana and northern Namibia?), where the most distal fluvial facies were deposited. However, these sediments are not fully preserved, as they were probably removed by uplift and erosion during the pre-Stormberg hiatus. Fluvial 7 Pv ial basin / Source area Palaeocurrent direction Basin margin aes u SOugherm we Range? Fig. 7. Middle to late Triassic fluvial Stormberg basins. Palaeolatitudes (at +225 Ma) after Smith et al. (1981). NA = Namibia, BO = Botswana, ZI = Zimbabwe, MO = Mocambique, RSA = South Africa, FI = Falkland Islands. 424 ANNALS OF THE SOUTH AFRICAN MUSEUM Plate tectonics in the far south greatly influenced the palaeogeography. Crustal deformation and uplift migrated northwards so that during Beaufort deposition prominent mountain ranges were then located close to the present southern African coastline. Crustal uplift in the west also resulted in basinward migration of source areas. Fluvial Stormberg* basins An erosional period of up to 10 Ma separated the fluvial Stormberg sedi- mentation from the underlying Beaufort beds in the south. During this period large areas were probably stripped of their Beaufort sediment cover. The Storm- berg beds were deposited in small isolated fault-controlled basins (Fig. 7) with mountainous sources located mostly in the south-east, east and north. Typical Major wind direction Basin margin Fig. 8. The early Jurassic aeolian Stormberg Basin. Palaeolatitudes (at +205 Ma) after Smith et al. (1981). NA = Namibia, BO = Botswana, ZI = Zimbabwe, MO = Mocambique, RSA = South Africa, FI = Falkland Islands. * The term ‘Stormberg’ is informally used for the combined Molteno, Elliot and Clarens formations. GEOGRAPHY AND CLIMATOLOGY OF THE KAROO BASIN 425 thick clastic wedges accumulated at depocentres that showed a progressive shift towards the east and north (Visser 1984). The noticeable absence of deposition in the west may be attributed to those regions preceding break-up of Gondwana. The mountainous sources in the south-east had an elevation of up to 4000 m (Turner 1975) and were initially drained by predominantly braided streams during deposition of the Molteno Formation and by meandering streams at a later stage. The distal fluvial facies of these stream systems probably accumulated in Okavango-type swamps in the west (Visser 1984). Aeolian Stormberg Basin During the early Jurassic, fluvial sedimentation was largely replaced by aeolian deposition over a large part of south-western Gondwana, as a result of a climatic change (Fig. 8). The lateral extent of the aeolian basin margin is highly speculative, as deposition was independent of the palaeoslope. Westerly winds reworked and transported the unlithified fluvial sediments (including the distal facies) of the Stormberg basins (Visser 1984). The highlands in the south-east, east and north formed a barrier to the transportation of sand and airborne silt. This resulted in thick accumulations of sediment on the windward side of the uplands. PALAEOCLIMATOLOGY Climatology is the net result of a number of integrated parameters (admis- sion of solar energy, atmospheric composition, configuration of landforms and ocean basins, sea-level, pole position, and oceanic and atmospheric circulation), the relative importance of which we do not yet fully understand. In the dis- cussion of late Palaeozoic and early Mesozoic climates, reference will be made to some of these parameters, although those of extraterrestrial origin are not dealt with. Furthermore, to simplify discussion, only polar, cold, temperate, warm and warm desert climatic zones are referred to. Carboniferous Globally the early Carboniferous (320-360 Ma) had a warm climate with small latitudinal gradients (Frakes 1979; Boucot & Gray 1982). No Karoo rocks of this age were preserved and the upper part of the Witteberg Group thus constitutes the only source of evidence for reconstructing the pre-Karoo climatology, which appears to have been cool to cold (pers. comm. J. C. Loock; Fig. 9). This conclusion is substantiated by a palaeolatitude of >60°S for the region (Smith et al. 1981). Although global temperatures were apparently high during the early Carboniferous, the proximity of the basin to the south pole and the location of the pole largely over land, resulted in the anomalous climate of this region. As suggested by Spjeldnaes (1981), it is suspected that the cold climate was confined to a narrow zone around the pole during globally warmer periods. ANNALS OF THE SOUTH AFRICAN MUSEUM 426 STRATI- Lu Ke