sign in sign up
<<
Home , Border Cave, Chasmaporthetes, Makapansgat, Plovers Lake, Sterkfontein, Swartkrans

AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 96235-250 (1995)

Faunal Assemblage Seriation of Southern African and Deposits

JEFFREY K. McKEE, J. FRANCIS THACKERAY, AND LEE R. BERGER Department of Anatomy and Biology, University of the Witwatersrand, (J.K.M, L.R.B.) and Department of Palaeontology, Museum, Pretoria (J.F.T.),South KEY WORDS Faunal seriation, Southern African biochronology

ABSTRACT Fossil assemblages from the Pliocene and Pleistocene of southern Africa were seriated in order to give a better idea of their relative chronology. Time-sensitive were selected for calculation of the Fau- nal Resemblance Index among 17 site units. On the basis of a logistical seriation and subsequent site analysis, the following sequence of sites was deemed most probable: Member 3, Makapansgat Member 4, Dart deposits, Member 4 and Taung HrdliEka deposits, Sterkfontein Member 5 (in part) and Kromdraai B, Kromdraai A and Swart- krans Member 1, Member 2, Swartkrans Member 3, , Cornelia, Elandsfontein Main Site, of Hearths Acheulian levels, Florisbad and Equus Cave and Klasies River Mouth. o 1995 Wiley-Liss, Inc.

Associations of the early hominid and Partridge, 1983; Vogel, 1985) led to the with a wealth of Pliocene and Pleistocene suggestion that the may have fauna from southern Africa provide vital been close to the hypodigm of A. robustus clues to the past environments and evolu- (Tobias, 1973,1975,1978; Wolpoff, 1974) or tionary processes that shaped human evolu- that it represented a derived form of A. afri- tion. Due to the of the cave deposits canus (Tobias, 1991a,b). Recent faunal dat- in which these ancient assemblages are typ- ing of Taung (Delson, 1988; Cooke, 1990; ically found, where absolute geochronologi- McKee, 1993a), placing the site in time cal dating techniques are inhibited, the ab- closer to Sterkfontein and Makapansgat, solute and even relative temporal positions opens the possibility that the Taung homi- of these sites have proved difficult to ascer- nid is a more typical or even primitive repre- tain. Nevertheless, it is of pivotal impor- sentative ofA. africanus. tance in studies of mammalian evolution, Turner (1983, 1990) and Vrba (1985a, and particularly in paleoanthropological re- 1988,1992) have argued that changes in glo- search, to establish a chronology with which bal climate during the Pliocene caused we can assess the evolutionary roles of the mammalian evolutionary changes in Africa. varied fauna found in southern African cave Recently Hill et al. (1992) have suggested deposits. that the origin of corresponded to Understanding the timing of deposition in these putative climatic changes on the basis southern African hominid fossil sites is of of the age of a specimen they feel represents continuing importance, as illustrated by the the . However, in southern Africa it is influence dates have on interpretations of hominid phylogeny and theories of the causes of evolution. For example, the pro- Received August 21,1993; accepted August 17,1994. posed late date for the type specimen ofdus- Address reprint requests to Dr. J.K. McKee, Department of Anatomy and Human Biology, University of the Witwatersrand tralopithecus africanus from Taung, at circa Medical School, 7 York Road, Parktown, 2193 Republic of South 1 mya (Partridge, 1973, 1982, 1985; Vogel Mrica.

0 1995 WILEY-LISS, INC. 236 J. K. MCKEE ET AL. not clear whether the apparent faunal turn- referred to as Sterkfontein Member 5), overs seen in the fossil record predate or Swartkrans, and Kromdraai. The Vaal-Cor- postdate any climatic change, or whether or nelia Faunal Span included the Vaal not some of the faunal turnovers are indeed Younger gravels, Elandsfontein (formerly real (McKee, 1991, in press). Thus an under- known as Hopefield or Saldanha), and Cor- standing of cause and effect in mammalian nelia, followed by the Florisbad-Vlakkraal evolution requires at least an understand- Faunal Span encompassing the important ing of the succession of sites if not the actual Homo-bearing sites of Florisbad and Broken age of deposition. Hill. Opportunities to determine the ages of As progress is made in both excavation sites by geological dating have been limited and taxonomic identification of southern Af- by the nature of cave deposits. Sedimento- rican fossil fauna, sufficient samples are logical correlations of sites (Butzer, 1974), emerging for the refinement of the early nick-point analysis of cave openings (Par- methods that simply recognized faunal tridge, 19731, sedimentological rates (Par- spans. Toward this end, biochronologies tridge, 1985), and isotopic techniques (Vogel within taxonomic families have been estab- and Partridge, 1983; Vogel, 1985) have pro- lished for bovids (Vrba, 1975, 1982,1985b1, vided unsatisfactory results (De Swardt, cercopithecids (Delson, 1984, 1988), carni- 1974; Butzer et al., 1978; Tobias et al., 1993; vores (Hendey, 1974; Turner, 1990), and McKee, 1993a). Palaeomagnetic chronolo- suids (Harris and White, 1979). A combined gies have given mixed results, with some picture of all the known mammalian fauna success at Makapansgat (Brock et al., 1977; may then be useful in the seriation of indi- McFadden et al., 1979; McFadden, 1980) vidual site units within and between the and Kromdraai (Jones et al., 19861, but few faunal spans. clues at other sites due to the dynamics of There are numerous factors that may lead cave sediments. Electron spin resonance to similarities or differences of faunal repre- (ESR) has provided age estimates for vari- sentation in site assemblages that must be ous sites, including (Grun et considered when attempting to establish al., 1990a), Klasies River Mouth (Grun et biochronologies across taxonomic bound- al., 1990b), and Sterkfontein (Schwarcz et aries. Temporal determinants, those most al., 1993), but these are dependent on as- useful for faunal dating, interplay with sumptions regarding rates of uptake of ura- other sampling biases so that only a fraction nium (Griin and Stringer, 1991). Neverthe- of the total faunal community living at any less, there is considerable potential in the one time would be represented in a cave de- application of ESR dating with necessary ca- posit. A summary of the key factors that veats and corroborating evidence from other may lead to similarities or differences techniques. among faunal assemblages in southern Afri- Chronologies based on faunal correlations can cave deposits can be found in Table 1. have thus been the primary means of dating Although faunal dating may have its limi- southern African fossil sites. Many attempts tations due to the complex network of fac- have been made (reviewed by McKee, tors that determine site assemblages, it re- 1993a), but the most relevant work has its mains as the most reliable tool for cornerstone in the faunal spans delineated determining the sequence of fossil sites in by Ewer and Cooke (1964) and Cooke (1967). southern Africa. Our objective here is to es- The oldest of these temporal associations tablish a chronological framework for un- was the Sterkfontein Faunal Span, compris- derstanding mammalian evolution in the ing the -bearingdeposits of Pliocene and Pleistocene of southern Africa Sterkfontein, Makapansgat, and Taung, as through the seriation of site units. In order well as the Older Gravels of the Vaal River to achieve this with a total mammalian per- Basin. The subsequent Swartkrans Faunal spective, issues arising from the complicat- Span encompassed deposits from the Sterk- ing factors discussed above must be resolved fontein extension site (most of which is now with an appropriate method. SOUTHERN AFRICAN FAUNAL ASSEMBLAGE SERIATION 237 TABLE 1. Factors causing similarities or differences in sampled or faunally diverse site is more faunal representation at southern African caue sites likely to yield species identified from else- Environmental factors where than is a site with a paucity of fauna. Faunal community Floral community Simpson’s (1960) Faunal Resemblance In- Demography dex (FRI) provides a partial solution to the Climate sampling problem and is employed here. For Altitude Terrain each pair of sites, the FRI is calculated by Geographic proximity of sites taking the number of species shared (C) di- Availability of water Cave size and depth vided by the total number of species found in Taphonomic factors the smaller sample (N1),or simply FRI = C/ Taphonomic agents Nl x 100. These values are displayed in Ta- Bone preservation Mixing of deposits ble 3B. The advantage of the method is that Interpretive factors a small sample can exhibit an FRI value Selective excavation and preparation comparable to that of a much larger sample, Taxonomic biases Temporal factors thus standardizing the figures in the ma- Origins trix. It should be noted, however, that a site Radiations Extinctions with a large sample still has a greater prob- ability of sharing species with other sites and the use of an FRI does not address this issue. METHODS AND ASSUMPTIONS Complications from nontemporal factors Lists of identified mammalian species and of site similarities can cause spurious corre- genera were compiled from published litera- lations. For example, a long-lived but ecolog- ture for 23 Neogene site units (Fig. l and ically limited species, Hippopotamus am- Table 2). This amounted to 246 species of phibius, appears at the oldest and youngest 159 genera and 44 families. sites, but not at sites that span the time in It was our primary goal to seriate the sites between. Thus the analysis must be limited based on similarities of identified species be- to the species that are most sensitive to time tween site unit assemblages. The initial step and not restricted to highly specialized envi- is to create a matrix exhibiting the number ronments. of identified species shared between each Evaluation of temporally sensitive species pair of site units (Table 3A); the diagonal can be a somewhat subjective task. Any cri- from top left to bottom right shows the num- teria used to label a species as time-sensi- ber of identified species for each site unit. tive must be seen as an underlying assump- Specific identifications with “cf.” or “aff.” tion of the analysis. For this analysis we were included for the analysis, as in Turner established the following objective criteria and Wood (1993). for identifying such species: (1)The species The ideally seriated matrix should exhibit must occur at more than one fossil site; spe- an ordering of figures within each row, in- cies known from only one site that show evo- creasing from left to right toward the diago- lutionary relationships to species from other nal, and decreasing away from the diagonal. sites can be used as corroborating evidence. In other words, sites closer in time should (2)The species must not be present through- share a successively greater number of spe- out the entirety of the sequence (given the cies than more temporally distant sites. The assumption of an oldest site). (3) Species nontemporal factors listed in Table 1, how- that are highly sensitive to local environ- ever, severely complicate the ideal, and ad- mental conditions may cause spurious corre- ditional steps are necessary to place the fo- lations among sites of similar conditions cus of the seriation on the temporal from vastly different times, and thus should components of the data. be excluded. Known habitat preferences of Intrinsic standardization of the data is extant species can be used to infer similar necessary in order to alleviate the problems tendencies of the same or related species in engendered by differential sampling; a well the past, but the potential for bias in exclud- 238 J. K. MCKEE ET AL

. ,......

.. ...* ’. .. ..*. ..*: . *...... *

‘Cornelia *...- Border Cave / .. . -** ,.**.* ...... vaa~River 4 + .....‘ ’. Florisbad :’ .:

Republic of

Fig. 1. Map of southern African fossil sites. ing species on this basis must be acknowl- produced unrealistically large FRI values. edged. Thus species that occur at sites These sites included Langebaanweg, the which, on the basis of independent evidence, Taung Dart Deposits, the Vaal Older River appear to have been environmentally dis- Gravels, and Border Cave. Additional sites tinct (e.g., Taung and Makapansgat), were were excluded because it was known a priori not excluded. It should be remembered that that fauna from deposits of vastly different most of the southern African fossil sites are time periods were mixed into single assem- cave deposits that have sampled mammals blages; these sites were Bolt’s Farm and from a mosaic of local environments. Gladysvale (Berger, 1993). It must be ac- Virtually all of the 66 time-sensitive spe- knowledged at the outset, however, that cies identified using the above criteria some mixing of fossils may have occurred at showed some degree of irregularity in their other sites with multiple deposits. appearances at various sites, as they are not Logistical seriation of the sites was based free from the factors listed in Table 1. It thus on the matrix of FRI values for time-sensi- must be assumed that sampling of such spe- tive species. A computer program was writ- cies from site units is random with respect to ten which tested different site sequences in time. The time-sensitive species are listed in order to minimize differences among adja- Table 4 along with their distribution among cent sites showing “backward steps”: figures the sites used in the analysis. The matrices that decrease in value toward the diagonal in Table 5 are based on time-sensitive spe- are “backward steps” because in an ideal cies alone. matrix the FRI values should always in- Once time-sensitive species were consid- crease toward the diagonal. This was accom- ered, a number of sites had to be excluded plished by summing the absolute value of from the analysis due to small sample sizes. differences between adjacent sites in a row The small samples of time-sensitive species showing “backward steps.” The site unit or- SOUTHERN AFRICAN FAUNAL ASSEMBLAGE SERIATION 239

TABLE 2. Site units and codes used in Tables 36' sites often changes the sequence among Site unit Code other sites, and these situations must be scrutinized carefully. However, penecon- Langebaanweg 'E' Quarry LANG Makapansgat Member 3 MAK3 temporaneous sites may be expected to seri- Makapansgat Member 4 MAK4 ate in alternating positions. (3) Each site Taung Dart deposits TAUD Taung HrdliEka deposits TAUH must be considered in light of corroborating Sterkfontein Member 4 STS4 evidence from evolutionary sequences of lin- Sterkfontein Member 5 STS5 eages, human artifacts (stone tools, etc.), Older Vaal River Gravels VAAL Bolt's Farm BOLT and absolute dating, whenever such infor- Kromdraai A KROA mation is available. Kromdraai B KROB Gladysvale GLVH Swartkrans Member 1 SWTl RESULTS Swartkrans Member 2 swT2 Swartkrans Member 3 swT3 The seriation of the FRI values presented Plovers Lake PLOV in Table 5B led to the following optimum Cornelia CORN Elandsfontein Main Site EFTM series of 10 sites per run: Cave of Hearths (Acheulian levels) COHA Border Cave (MSA levels) BCVM Florisbad springs FLOS A. MAK4 MAK3 TAW STS4 STS5 Equus Cave (MSA levels) EQW KROB KROA SWTl SWT3 SWT2 Klasies River Mouth (MSA I and I1 levels) KLAS Extant soecies in southern Africa EXTA B. MAK3 MAK4 STS4 TAUH KROB KROA SWTl SWT2 SWT3 PLOV 'Species identified from each site were compiled by one of US (JKM) from the following references: Berger (19931, Brain (19811, Brain and Watson (19921, Brickhill (19761, Brink (19871, Butzer et al. (19741, Churcher (19701, Collings (1972), Collings et al. (19761, Cooke (1978, These two sequences maintained consis- 1988,1990,1991,1993a,b),Cooke and Wells (19471,de Graaf (1960, 1988), Freedman (19571, Greenwood (19551, Helgren (1979), Hendey tency with the further removal of individu- (19811, Hwijer (19581, Klein (1976, 19771, Mein et al. (19911, Klein als sites. The position of the Taung HrdliCka and Cruz-Uribe (19911, McKee (1991,1993a,c),Randall (19751, Pocock (19871, Skinner and Smithers (19901, Szalay and Delson (19791, assemblage alternated with Sterkfontein Turner (19861, Vrba (1981, 1987a,b), Watson (19931. An unpublished Member 4, depending on the inclusion of species list was provided hy J.F. Thackeray for Plovers Lake. Sterkfontein Member 5. As discussed below, this may be because of near contemporane- ity of the two alternating sites. Although der producing the minimum sum was Makapansgat Member 4 was sometimes po- deemed the best seriation. As the number of sitioned before Member 3, and Swartkrans possible seriated orders is a factorial of the Member 3 sometimes seriated before Mem- number of site units, only 10 sites could be ber 2, forcing the stratigraphic orders in se- seriated at any one time and the complete quence did not affect the relative positions of sequence was compiled from overlapping se- the other site units. riations. For the overall ordering of the se- In the presumed middle of the site se- quences, an a priori assumption was made quence, the seriation results were quite con- that sites with fewer extant species were sistent and confirmed the relations sug- earlier in time. gested by sequences A and B above. For Use of the method outlined above mini- example: mizes the effects of nontemporal factors but is far from foolproof. Initial tests of the pro- cess revealed that a number of principles C. TAUH STS5 KROA SWTl SWT3 must be imposed on the logistcal seriation SWT2 PLOV CORN EFTM COHA and interpretation of the matrix. (1) When D. TAUH KROB KROA SWTl SWT2 available, stratigraphic relations among site CORN EFTM COHA EQUU KLAS units must be used. For example, Swart- krans Member 3 must follow Swartkrans Optimal seriations of the more recent Member 2, although this detracts from an sites produced a more variable picture, de- ideal sequence of FRI values. (2) Where mul- pending on the sites used for analysis. The tiple sequences are possible, each must be main variants are illustrated by the follow- treated as tenable. The exclusion of certain ing sequences: TABLE 3. List of shared species at southern African site units: (A) absolute frequencies of shared species and (B)faunal resemblance in& (FRI)' A LANG MAK3 MAK4 TAUD STS4 VAAL TAUH BOLT STS5 KROB KROA GLVH SWTl SWTZ SWT3 PLOV CORN EFTM COHA FLOS EQUU KLAS BCVM EXTA LANG34 4 2 132 12 11 2 2 2 11 2 11 10 0 0 0 1 MAK3451 13 3 24 115 13 911 15 1514 5 6 2 3 6 2 2 7 5 2 14 MAK4 2 13 34 2 14 0 9 8 12 11 8 9 7 3 3 3 11 3 13 0 0 14 TAUD1 3 2 7 30 2 2 0 0 12 11 11 0 0 0 0 0 0 0 0 STS4 3 24 14 3 45 0 18 16 18 16 20 21 20 9 9 6 1 6 6 4 7 4 1 15 VAAL 2 10 0 0 5 0 2 0 11 11 2 2 12 2 0 0 0 0 0 0 TAUH 1 15 9 2 18 0 33 16 10 13 11 13 13 6 5 5 0 2 5 1 3 3 1 13 BOLT 2 13 8 2 16 2 16 31 11 11 15 14 19 9 9 7 4 5 4 3 3 1 0 8 STS5 1 9 12 0 18 0 10 11 31 15 17 11 13 6 8 3 2 6 8 5 6 2 1 17 KROB 1 11 11 0 16 1 13 11 15 33 20 9 16 7 8 6 2 4 5 2 4 3 0 15 KROA 2 15 8 1 20 1 11 15 17 20 48 22 25 13 17 12 6 11 9 7 13 10 3 19 GLVH 2 15 9 2 21 1 13 14 11 9 22 54 20 12 15 12 7 12 12 10 19 11 8 27 SWTl 2 14 7 1 20 1 13 19 13 16 25 20 46 19 23 9 5 13 7 6 15 10 3 20 SWT21 5 3 192 6 9 6 71312192524 8 511 3 613 6 3 14 SWT31 6 3 192 5 9 8 8171523243410 512 5 719 9 5 21 PLOV 2 2 3 161 5 7 3 6 12 12 9 8 10 24 3 7 8 5 9 5 3 13 CORN1 3 10 12 0 4 2 2 6 7 5 5 5 3 18 11 5 8 6 3 1 6 EFTM 1 6 1 0 6 2 2 5 6 4 11 12 13 11 12 7 11 40 10 14 19 17 4 26 COHA 12 3 0 6 0 5 4 8 5 9 12 7 3 5 8 5 10 29 9 13 11 6 21 FLOS 0 2 10 4 0 13 5 2 7 10 6 6 7 5 8 14 9 22 18 8 4 16 EQUU 0 7 3 0 7 0 3 3 6 4 13 19 15 13 19 9 6 19 13 18 46 22 10 43 KLAS 0 5 0 0 4 0 3 12 3 10 11 10 6 9 5 3 17 11 8 22 34 9 32 BCVMO 2 0 0 10 10 10 3 8 3 3 5 3 14 6 4 10 9 15 15 EXTA 1 14 14 0 15 0 13 8 17 15 19 27 20 14 21 13 6 26 21 16 43 32 15 101 B LANG MAK3 MAK4 TAUD STS4 VAAL TAUH BOLT STS5 KROB KROA GLVH SWTl SWT2 SWT3 PLOV CORN EFI'M COHA FLOS EQUU KLAS BCVM EXTA LANG100 12 6 14 940 3 6 3 3 6 6 6 4 3 8 6 3 3 0 0 0 0 3 MAK3 12 100 38 43 53 20 45 42 29 33 31 29 30 20 18 8 17 15 7 9 15 15 13 27 MAK4 6 38 100 29 41 0 27 26 39 33 24 26 21 12 9 13 6 3 10 5 9 0 0 41 TAUD 14 43 29 100 43 0 29 29 0 0 14 29 14 14 14 14 0 0 0 0 0 0 0 0 STS4 9 53 41 43 100 0 55 52 58 48 44 47 44 36 26 25 6 15 21 18 16 12 7 33 VAAL 40 20 0 0 0 100 0 40 0 20 20 20 20 40 40 20 40 40 0 0 0 0 0 0 TAUH 3 45 27 29 55 0 100 52 32 39 33 39 39 24 15 21 0 6 17 5 9 9 7 39 BOLT 6 42 26 29 52 40 52 100 35 35 48 45 61 36 29 29 22 16 14 14 10 3 0 26 STS5 3 29 39 0 58 0 32 35 100 48 55 35 42 24 26 13 11 19 28 23 19 6 7 55 mon 3 33 33 o 48 20 39 35 48 100 61 27 48 28 24 25 11 12 17 9 12 9 o 45 KROA 6 31 24 14 44 20 33 48 55 61 100 46 54 52 50 50 33 28 31 32 28 29 20 40 GLVH 6 29 26 29 47 20 39 45 35 27 46 100 43 48 44 50 39 30 41 45 41 32 53 50 SWTl 6 30 21 14 44 20 39 61 42 48 54 43 100 76 68 38 28 33 24 27 33 29 20 43 SWT2 4 20 12 14 36 40 24 36 24 28 52 48 76 100 96 33 28 44 12 27 52 24 20 56 SWT3 3 18 9 14 26 40 15 29 26 24 50 44 68 96 100 42 28 35 17 32 56 26 33 62 PLOV 8 8 13 14 25 20 21 29 13 25 50 50 38 33 42 100 17 29 33 23 38 21 20 54 CORN 6 17 6 0 6 40 0 22 11 11 33 39 28 28 17 28 100 61 28 44 33 17 7 33 EFTM 3 15 3 0 15 40 6 16 19 12 28 30 33 44 35 29 61 100 34 64 48 50 27 65 COHA 3 7 10 0 21 0 17 14 28 17 31 41 24 12 17 33 28 34 100 41 45 38 40 72 FLOS 0 9 5 0 18 0 5 14 23 9 32 45 27 27 32 23 44 64 41 100 82 36 27 73 EQUU 0 15 9 0 16 0 9 10 19 12 28 41 33 52 56 38 33 48 45 82 100 65 67 93 KLAS 0 15 0 0 12 0 9 3 6 9 29 32 29 24 26 21 17 50 38 36 65 100 60 94 BCVM 0 13 0 0 7 0 7 0 7 0 20 53 20 20 33 20 7 27 40 27 67 60 100 100 EXTA 3 27 41 0 33 0 39 26 55 45 40 50 43 56 62 54 33 65 72 73 93 94 100 100

'Site codesare listed~n Table 2 SOUTHERN AFRICAN FAUNAL ASSEMBLAGE SERIATION 241 TABLE 4. Time-sensitiue mammalian species with site unit occurrences marked by an asterisk' MMSTSKKSSSPCECFEKE AATATRRWWWLOFOL LX KKSUSOOTTTORTHO$AT 344H5BA123VNMASUSA

Artiodactyla Alcelaphus buselaphus * *** Antidoreas australis * Antidorcas bondi ** * *** Antidorcas marsupialis *** ** Antidorcas recki * ****** ** Connochaetes gnou ** ** Damaliscus d0rea.s ** Damaliscus niro ** Hippotragus cwkei * * Hippotragus leucophaeus * ** Makapania broomi * * Megalotmgus priscus ****** Oreotragus major ***** Pelea capreolus ** ** *** Pelornuis antiquus **** * Raphicerus campestris *** * * Redunca arundinum ** *** I* Redunca darti * Redunca fulvorufula *** Simatherium kohllarseni ** Symerus caffer *** Taumtragus oryx *********I Tragelaphus strepsicems ** ** ** I** Kolpochoerus paiceae ** Metridiochoerus andrewsi *** Notochoerus scotti ** Phacochoerus aethiopicus ***** Phacochoerus modestus ***I* * Potamochneroides shawi * Lycaan pictus * ** Vulpes pulcher ** barlowi * * Homotherium crenatidens * * Megantereon cultridens **** * Panthera pardas ** ****** *** Chasmaporthetes silberbergi * * Chasmaporthetes nitidula * *** Hyaena brunnea * **** *** Pachycmuta bellax * * * Aonyx capensis *** **** Melliuora capensis * *** Atilax paludinosus **** Herpestes ichneumon * I* *** Suricata suricatta ** * * * Primates Australopithecus africanus *** Australopithecus robustus * *** Homo sapiens ****** Cercopithecoides willinmi **** * Papio izodi *I Papio angusticeps ** Papw mbinsoni ****** Papw cynocephalus * *** Parapapw bmomi *** Parapapio whitei *** Parapapw jonesi *** *** Theropithecus oswaldi *** * Proboscidea Elephas recki * Perissodactyla Equus burchelli * * * *** Equus capensis * ********** Hipparwn libycum * **** * Hyracoidea Pmcavia transvaalensis * * ***** Rodentia Mystromys antiquus * ** * * Pmodontomys cwkei **** *** Hystrir makapanensis * Elephantulus antiquus *****I* Mvosora robinsoni * * **

'Site edea are listed in Table 2. TABLE 5. List of shared time-sensitiue species at southern African site units: (A) absolute frequencies of shared species and (B) faunal resemblance in& (FRI)l A MAK3 MAK4 STS4 TAUH STS5 KROB KROA SWTl SWT2 SWT3 PLOV CORN EFTM COHA FLOS EQUU KLAS EXTA MAK3 21 10 17 6 4 5 8 7 1 1 0 1 0 0 0 0 0 0 MAK4 10 11 8 6 2 4 4 4 1 1 1 0 0 0 0 0 0 0 STS4 17 8 26 7 9 9 11 12 4 5 2 1 2 1 1 3 2 2 TAUH 6 6 7 9 2 5 4 5 2 2 2 0 0 0 0 1 1 1 STS5 4 2 9 2 11 3 7 5 3 5 1 2 2 2 2 2 0 1 KROB 5 4 9 5 3 13 10 10 5 6 3 2 1 1 1 2 1 1 KROA 8 4 11 4 7 10 24 16 9 13 8 6 6 6 4 8 6 7 SWTl 7 4 12 5 5 10 16 27 14 17 6 5 8 4 3 9 7 9 sWT2 1 1 4 2 3 5 9 14 17 17 6 5 8 2 4 8 4 7 swT3 1 1 5 2 5 6 13 17 17 21 8 5 9 4 5 11 6 10 PLOV 0 1 2 2 1 3 8 6 6 8 11 3 6 5 4 7 4 4 CORN 1 0 1 0 2 2 6 5 5 5 3 11 9 5 6 4 3 3 EFTM 0 0 2 0 2 1 6 8 8 9 6 9 21 8 9 12 10 12 COHA 0 0 1 0 2 1 6 4 2 4 5 5 8 13 8 11 9 9 FLOS 0 0 1 0 2 1 4 3 4 5 4 6 9 8 15 13 6 10 EQUU 0 0 3 1 2 2 8 9 8 11 7 4 12 11 13 25 15 22 KLAS 0 0 2 1 0 1 6 7 4 6 4 3 10 9 6 15 18 17 EXTA 0 0 2 1 1 1 7 9 7 10 4 3 12 9 10 22 17 24 B MAK3 MAK4 STS4 TAUH STS5 KROB KROA SWTl SWT2 SWT3 PLOV CORN EFTM COHA FLOS EQUU KLAS EXTA MAK3 100 91 81 67 36 38 38 33 6 5 0 9 0 0 0 0 0 0 MAK4 91 100 73 67 18 36 36 36 9 9 9 0 0 0 0 0 0 0 STS4 81 73 100 78 82 69 46 46 24 24 18 9 10 8 7 12 11 8 TAUH 67 67 78 100 22 56 44 56 22 22 22 0 0 0 0 11 11 11 STS5 36 18 82 22 100 27 64 45 27 45 9 18 18 18 18 18 0 9 KROB 38 36 69 56 27 100 77 77 38 46 27 18 8 8 8 15 8 8 KROA 38 36 46 44 64 77 100 67 53 62 73 55 29 46 27 33 33 29 SWTl 33 36 46 56 45 77 67 100 82 81 55 45 38 31 20 36 39 38 sWT2 6 9 24 22 27 38 53 82 100 100 55 45 47 15 27 47 24 41 swT3 5 9 24 22 45 46 62 81 100 100 73 45 43 31 33 52 33 48 PLQV 0 9 18 22 9 27 73 55 55 73 100 27 55 45 36 64 36 36 CORN 9 0 9 0 18 18 55 45 45 45 27 100 82 45 55 36 27 27 EFTM 0 0 10 0 18 8 29 38 47 43 55 82 100 62 60 57 56 57 CQHA 0 0 8 0 18 8 46 31 15 31 45 45 62 100 62 85 69 69 FLOS 0 0 7 0 18 8 27 20 27 33 36 55 60 62 100 87 40 67 EQUU 0 0 12 11 18 15 33 36 47 52 64 36 57 85 87 100 83 92 KLAS 0 0 11 11 0 8 33 39 24 33 36 27 56 69 40 83 100 94 EXTA 0 n 8 11 9 8 29 38 41 48 36 27 57 69 67 92 94 100 ‘Site codes are listed in Table 2 SOUTHERN AFRICAN FAUNAL ASSEMBLAGE SERIATION 243 E. STS5 KROA SWTl SWT2 PLOV cus africanus as well as the ancient artio- CORN EFTM FLOS EQWKLAS dactyls, Simatherium cf. kohllarseni and F. SWT2 SWTl KROA PLOV CORN Notochoerus scotti, and Eisenhart (1974) EFTM FLOS COHA EQWKLAS had noted that the cercopithecids are virtu- G. SWTl SWT2 SWT3 PLOV KROA ally identical. Makapansgat Member 4 does CORN EFTM EQUU COHA FLOS pick up more extant fauna, but this is largely due to the microfauna from “rodent Because of the high FRI value for Plovers corner” which may have only a tenuous asso- Lake and Kromdraai A, two of the sequences ciation with the large mammals in other place the latter site after Swartkrans. But parts of the cave. Although a stratigraphic note in sequence F that the first three sites separation between the two “members” ex- are in reverse order from that expected; they ists in some parts of the cave, the difference seriated correctly among themselves in or- may not have been of a geologically signifi- der to allow for the similarities between cant time period and Members 3 and 4 may Kromdraai A and Plovers Lake. be better viewed as different parts of the None of the Middle (MSA) sites same cave exhibiting slightly different ta- (Florisbad, Equus Cave, and Klasies) takes phonomic processes (Maguire et al., 1985). consistent positions in the seriation. A close look at the species from the Taung sites shows why: there are few distinctive Most of the Taung fauna came from the faunal trends following the time of Elands- HrdliEka deposits, which have their greatest fontein deposition, thus site correlations are affiliations with Sterkfontein Member 4. somewhat more spurious. Nevertheless, The faunal seriation of the time-sensitive there is consistency in the sequence of Plo- species supports the recently espoused view vers Lake, Cornelia, Elandsfontein, and that the HrdliEka deposits are roughtly con- Cave of Hearths, and one of the MSA sites temporaneous with Sterkfontein Member 4 always comes out as the most recent site. (McKee, 1993a). As the association is as This is illustrated by the following sequence strong with Swartkrans Member 1 as it is covering the entire time span: with Makapansgat Member 3, a slightly later age may be tenable. H. MAK3 TAUH STS4 KROB SWTl The Taung hominid skull, however, prob- PLOV CORN EFTM COHA FLOS ably came from a cave system associated with the Dart deposits, and has been posited Site analysis to be considerably older than the age repre- Makapansgat sented by the Hrdlicka faunal assemblage (McKee, 1993b). There are few time-sensi- Makapansgat Member 3 has long been tive species yet known from the Dart depos- thought to be the oldest of the southern Afri- its, so a faunal age is difficult to ascertain by can fossil sites, and this is borne out by the the seriation method used here. All of the faunal seriation. Over two-thirds of its spe- seven species likely to be from the Dart de- cies are extinct, and it shares no time-sensi- posits are extinct, and only two time-sensi- tive species with the late Pleistocene Homo- tive species (A. africanus and Parapapio bearing sites. The assemblage includes broomi) occur at other sites, so an age pre- particularly ancient species from East Af- dating Makapansgat is conceivable. On the rica such as the chalicotheriid, Ancylothe- other hand, the occurrence of four unique, rium hennigi, a primitive buffalo, Simathe- extinct species, as witnessed in the Dart de- rium kohllarseni, and two primitive suids, posits, is not unexpected at a time following Potamochoeroides shawi (Bender, 1992) and deposition at Makapansgat (McKee, 1994). Notochoerus scotti (Cooke, 1993). The Dart assemblage, including the type Member 4 at Makapansgat has its great- specimen of Australopithecus africanus, est faunal associations with Member 3, most likely falls between the times of Maka- hence their alternating positions in the seri- pansgat (Members 3 and 4) and Sterkfon- ations. Both deposits contain Australopithe- tein Member 4 (McKee, 1993a). 244 J. K. MCKEE ET AL. Sterkfontein likelihood of this interpretation. The time- sensitive species, however, fit conformably Sterkfontein Member 4 has been ex- within the seriated position of Member 5 tremely well sampled through 27 years of (Table 4). continuous excavation. As it has yielded con- siderably more of the late Pliocene fauna Kromdraai known from other Transvaal sites than does Kromdraai A and B are most similar to Makapansgat, it most likely postdates each other, but Kromdraai B, with Australo- Makapansgat. However, Sterkfontein Mem- pithecus () robustus, is most ber 4 shares many ancient species with likely older. Both units are the same FRI of Makapansgat, and thus shows the last ap- time-sensitive species with Makapansgat pearance of seven time-sensitive species Members 3 and 4, and Kromdraai B has pro- (Table 4), including A. africanus. portionally stronger ties to Sterkfontein Member 5 at Sterkfontein marks the first Member 4. Kromdraai A, on the other hand, appearance in southern Africa of the genus is the first site to show a large increase in Homo as well as of stone tools. Its strongest the number of extant time-sensitive species faunal affinities are with Sterkfontein as well as those species characteristic of the Member 4, followed by Kromdraai A. Being Cornelia Faunal Span. This is largely due to that it has lost some of the older species, the increased proportion of bovid species notably A. africanus and Notochoerus scotti, known from the site. it is best placed following the Taung A case could be made for Kromdraai A HrdliEka deposits. It does not have as strong following Swartkrans Member 1, as sug- an affinity to Swartkrans as does Krom- gested by the seriation. This is largely due to draai B, so the seriation places Sterkfontein the similarity between Kromdraai A and Member 5 before Kromdraai B. Plovers Lake, but there are other consider- Sterkfontein Member 5 thus falls between ations. Dinofelis piveteaui of Kromdraai A is two sites with which it has very little in com- seen to be an evolutionary descendent of D. mon. There are a number of possible expla- barlowi, known from Swartkrans Member 1 nations for this. It may have had a distinctly and earlier sites (Cooke, 1991). Swartkrans different environment from that of Taung, Member 1 also has stronger faunal ties to as it does today, and thus would have more Sterkfontein Member 4 and the Taung in common with Sterkfontein Member 4. HrdliEka deposits. On the other hand, However, it seems to be nearly contempora- Kromdraai A has a strong resemblance to neous with Kromdraai B, yet shares fewer Kromdraai B and Sterkfontein Member 5, than half of all its species, and only 27% of and has not yielded as many extant species its time-sensitive species with this nearby as Swartkrans Members 2 and 3, so we sug- site in the same valley. This may be attribut- gest that on the current evidence it comes able to taphonomic considerations in which just before or very near in time to Swart- early Homo is a major contributor to the krans Member 1. Sterkfontein Member 5 deposit (Vrba, 1975, 1976, 1980; Brain, 1981, 1985; McKee, Swartkrans 1991), but not to Kromdraai which has been viewed as a part death trap and part feeding The Swartkrans Formation harbors both lair of opportunistic predators (Vrba, 1981; A. robustus and Homo together for the first Brain, 1981). Indeed, Kromdraai B has re- time. Each of the three members share the markably few bovids, and Sterkfontein most species, including time-sensitive spe- Member 5 has a paucity of cercopithecids. cies, among themselves, and it has been sug- A final possibility for the interpretation of gested that they do not span a long period of Sterkfontein Member 5, as suggested by time (Brain, 1988, 1993; Delson, 1988). Vrba (19851, is that temporally distinct de- However, Member 1 is clearly distinct from posits are mixed within the “member.” Ar- the others. It is the last site unit in the seri- chaeological evidence from recent excava- ation to have strong faunal ties to Maka- tions (Kuman, 1994) has increased the pansgat, Taung, and Sterkfontein, and thus SOUTHERN AFRICAN FAUNAL ASSEMBLAGE SERIATION 245 sees the last appearance of eight time-sensi- nelia has its greatest affinities with the tive species (Table 4). Recently Grine et al. Elandsfontein Main Site. Remarkably, only (1993) have suggested a further link be- 33% of all its species are extant. It has tween Swartkrans Member 1 and Sterkfon- yielded many of the species characteristi- tein Member 5 with their identification of cally associated with early Homo in south- similarities in the hominid material, which ern Africa, but predates the Homo sites they tentatively assign to where extant species appear in greater pro- sensu lato; Clarke (1985), however, refers portions. the SK 847 cranium from Member 1 to . It should be noted as well that Elandsfontein Swartkrans Member 1 also picks up five The Elandsfontein Main Site ushers in ar- first appearances in the fossil record (or chaic Homo sapiens with the “Saldanha” more if Kromdraai A follows it in time), thus skull. The fossils include the last evidence of clearly separating it from the early Austral- some key Plio-Pleistocene boundary species opithecus sites. including Antidorcas recki, Theropithecus Swartkrans Members 2 and 3 then lose oswaldi, and Metridiochoerus andrewsi, as most of the fauna from sites of the Sterkfon- well as the last appearance of the genus tein Faunal Span. By the time of Member 3, Megantereon, thus attesting to its great age. nearly half of the time-sensitive species, and However, it is the first site to have extant 62% of all the species, are those that are species represented in over half of the time- extant today. Member 3 also provides the sensitive fauna. last glimpse ofA. robustus. There is a most remarkable similarity be- Cave of Hearths tween Members 2 and 3 at Swartkrans, with an FRI for time-sensitive species of 100 and The Acheulian levels near the bottom of an overall FRI of 96. This may be because of the Cave of Hearths deposits in the Maka- a very close time frame covered by the depo- pansgat valley, yielded a mandible and a ra- sition of the two members, but it is also pos- dius of a human that have been referred by sible that there has been some inadvertent Tobias (1971) to Homo sapiens rhodesiensis. mixing of specimens from the two members The relative position in time of this Early due to the decalcification of Member 3. Stone Age (ESA) site is difficult to ascertain. Cornelia, Elandsfontein, and Cave of Plovers Lake Hearths have ”Upper Acheulian” (Butzer et al., 1974; Mason, 1988), but the latter site Plovers Lake, a cave site in the Blaaubank certainly postdates Cornelia and Elandsfon- River Valley with Sterkfontein, Swartkrans, tein on the basis of the fauna. The seriation and Kromdraai, is the most recent addition results suggest, as did Cooke (1988), that to the known faunal assemblages of south- the faunal assemblage is typical of the Mid- ern Africa. The fauna from Plovers Lake re- dle Stone Age (MSA) sites. It has very close sembles most closely Kromdraai A and affinities to the MSA site of Equus Cave and Swartkrans Member 3, and fits most consis- is consequently seriated in variable posi- tently at or near the time of the latter. Plo- tions following Elandsfontein. As the site of vers Lake presents the last appearance of Equus Cave has apparent fully modern Papio robinsoni (or Papio hamadryas robin- Homo sapiens (Grine and Klein, 19851, it fits soni) and Procavia transvaalensis. that Cave of Hearths, with ESA artifacts and archaic Homo sapiens, should be older. Cornelia Florisbad Cornelia is the last faunal site before the appearance of archaic Homo sapiens in the Florisbad comes out in the seriation in fossil record of southern Africa. It is espe- variable positions following Elandsfontein, cially significant for the last appearance of with a strong similarity to Equus Cave. As Hipparion libycum, the last time-sensitive with the other Late Pleistocene MSA sites, species from Makapansgat Member 3. Cor- the faunal seriation alone can tell us little 246 J. K. MCKEE ET AL. about its relative age. However, the MSA existed could not be picked up by the seria- deposits of Florisbad has five extinct species tion technique. (Table 4), and sometimes seriates before the ESA deposits of Cave of Hearths, suggesting SUMMARY AND CONCLUSIONS that it may be older than some of the other The faunal seriation of southern Afi-ican MSA sites. Kuman and Clark (1986) esti- cave site assemblages has proven to be a mate the age of the Florisbad cranium and useful tool. Although it confirms the general associated fauna to be in the range of validity of the faunal spans devised in the 100,000 to 200,000 years old. 1960s by Ewer and Cooke, greater resolu- tion within those spans is now possible. Cave Moreover, as with fossil species, the naming Equus of faunal spans can obscure the evolutionary Faunal remains from Equus Cave show continuity of many mammals through time, that a somewhat more lush local environ- so a sequence of individual sites is prefera- ment once existed in what is now the desert ble. margin at Taung. The site has yielded only On the basis of the total analysis, both three extinct species: Antidorcas bondi, with the logistical seriation of time-sensitive which lasts until at least 38,000 B.P. at Bor- mammals and subjective scrutiny of the der Cave (Klein, 1977); Equus capensis, the data, the following sequence of sites is most large equid which survives to the very end of likely: Makapansgat Member 3, Makapans- the Pleistocene; and Megalotragus priscus, a gat Member 4, Taung Dart deposits, Sterk- giant alcelaphine last seen at 16,000 BP fontein Member 4 and Taung Hrdliirka de- (Klein, 1980). The faunal seriation thus tells posits, Sterkfontein Member 5 (in part) and us little about the place in time of Equus Kromdraai B, Kromdraai A and Swartkrans Cave; the presence of MSA artifacts, how- Member 1, Swartkrans Member 2, Swart- ever, places it prior to about 40,000 BP krans Member 3, Plovers Lake, Cornelia, (Klein et al., 1991). Elandsfontein Main Site, Cave of Hearths Acheulian levels, Florisbad and Equus Cave and Klasies River Mouth. Klasies River Mouth An important caveat emerges from analy- Although Klasies is an important MSA sis of the FRI values across the sites. For site with some evidence for Homo sapiens in most of the sites at which there is more than a Late Pleistocene context, the faunal analy- one deposit or “member,” the FRI values for sis of the hominid-associated MSA I and each unit are highest for other units of the MSA I1 levels gives few clues as to its rela- same site. This is true of Makapansgat, tive age. The site bears its strongest resem- Sterkfontein, Kromdraai, and Swartkrans, blance to Equus Cave, but has only one ex- and thus seems to be more than coinciden- tinct species, Pelorovis antiquus, which tal. At each site there are a number of possi- went extinct between 12,000 and 10,000 BP ble explanations including (1) temporal con- (Klein, 1980). Attempts have been made to tinuity across adjacent members, (2) similar seriate the total faunal assemblage from the external and internal environments yielding cave complex (Binford, 19861, but these taphonomic consistency, andor (3)inadvert- studies have been shown to be problematic ent mixing of the fossils and breccias from (Thackeray, 1986, 1987). Despite such un- separate deposits. The need to discern certainty, there is little question that the among these options highlights the impor- lower-most deposits are on the order of tance of continuing field work and analysis. 100,000 years, perhaps between 80,000 and The site sequence proposed here may 115,000 years old (Deacon and Geleijnse, change for some of the closely related sites, 1988). Thus Florisbad, Klasies River Mouth, as the seriation for those sites is sensitive to and perhaps Equus Cave, fall within a rela- the inclusion or exclusion of particular sites tively short time range during which what- and species. The overall picture, however, is ever mammalian faunal trends may have consistent for most sites and certainly the SOUTHERN AFRICAN FAUNAL ASSEMBLAGE SERJATION 247 proposed sequence may be used as a work- Brock A, McFadden PL, and Partridge TC (1977) Prelim- ing framework for the interpretation of evo- inary palaeomagnetic results from Makapansgat and lutionary patterns among the spe- Swartkrans. Nature 266:249-250. Butzer KW (1974) Paleoecology of South African Aus- cies and as corroborating evidence for tralopithecines: Taung revisited. Curr. Anthropol. 15: independent geochronological and biochro- 367-426. nological techniques. Ongoing research at Butzer KW, Clark JD, and Cooke HBS (1974) The geol- Sterkfontein, Taung, and Plovers Lake, as ogy, and fossil mammals of the Cornelia well as recently initiated excavations at Gla- Beds, O.F.S. O.F.S. Nat. Mus. (Bloemfontein) Mem. 9. dysvale, Makapansgat, and Kromdraai, un- Butzer KW, Stuckenrath R, Bruzewica AJ, and Helgren doubtedly will yield evidence that can refine DM (1978) Late Cenozoic paleoclimates of the Gaap the picture of Neogene mammalian evolu- Escarpment, Kalahari Margin, South Africa. Quat. tion and the place of hominids in it. Res. 10:310-339. Churcher CS (1970) The fossil Equidae from the ACKNOWLEDGMENTS . Ann. Transvaal Mus. 26:14&168. Clarke RJ (1985) Austrulopithecus and early Homo in Professor James W. Etching is gratefully southern Africa. In E Delson (ed.): Ancestors: The acknowledged for his help with the Maka- Hard Evidence. New York Alan R. Liss, pp. 171-177. pansgat faunal list. Funds for the develop- Collings GE (1972) A new species of Machaerodont from ment of the faunal data base and associated Makapansgat. Palaeont. Afr. 14:87-92. Collings GE, Cruikshank ARI, Maguire JM, and Randall software were provided by a grant to J.K. RM (1976) Recent faunal studies at Makapansgat McKee from the Foundation for Research Limeworks, Transvaal, South Africa. Ann. S. Afr. Development (South Africa). Mus. 71:153-165. Cooke HBS (1949) Fossil mammals of the Vaal River LITERATURE CITED deposits. Geol. Surv. (S. Afr.) Mem. 35:l-117. Bender PA (1992) A reconsideration of the fossil suid, Cooke HBS (1967) The Pleistocene sequence in South Potunochoeroides shawi, from the Makapansgat Africa and problems of correlation. In WW Bishop and Limeworks, Potgeitersrus, Northern Transvaal. JD Clark (eds.): Background to Evolution in Africa. Navors. nas. Mus. (Bloemfontein) 8:1-67. Chicago: University of Chicago Press, pp. 175-184. Berger LR (1993) A preliminary estimate of the age of Cooke HBS (1978) Suid evolution and correlation ofAfri- the Gladysvale site. Palaeont. Afr. can hominid localities: An alternative . Sci- 30:51-55. ence 201:460-463. Binford LR (1986) Reply to comments by R. Singer and Cooke HBS (1988) The larger mammals from the Cave of J. Wymer. Curr. Anthropol. 27:5742. Hearths. In RJ Mason (ed.): Cave of Hearths, Maka- Brain CK (1981) The Hunters or the Hunted? Chicago: pansgat, Transvaal. Archaeological Research Unit University of Chicago Press. (University of the Witwatersrand), Occasional Paper NO.21, pp. 507-534. Brain CK (1985) Cultural and taphonomic comparisons of Hominids from Swartkrans and Sterkfontein. In E Cooke HBS (1990) Taung fossils in the University of Delson (ed.): Ancestors: The Hard Evidence. New California Collections. In GH Sperber (ed.): From York: Alan R. Liss, pp. 72-75. to Angels. New York: Wiley-Liss, pp. 119-134. Brain CK (1988) New information from the Swartkrans Cooke HBS (1991) Dinofelis burlowi (Mammalia, Car- cave of relevance to “robust” . In FE nivora, Felidae) cranial material from Bolt’s Farm, Grine (ed.): Evolutionary History of the “Robust” Aus- collected by the University of California African Expe- tralopithecines. New York: Aldine de Gruyter, pp. dition. Palaeont. Afr. 28:9-21. 311-316. Cooke HBS (1993a) Undescribed suid remains from Brain CK (1993) Structure and stratigraphy of the Bolt’s Farm and other Transvaal cave deposits. Palae- Swartkrans cave in the light of new excavations. In ont. Afr. 30:7-23. CK Brain (ed.): Swartkrans: A Cave’s Chronical of Cooke HBS (1993b) Fossil Proboscidean remains from Early Man. Pretoria: Transvaal Museum Monograph Bolt’s Farm and other Transvaal cave deposits. Palae- No. 8, pp. 23-33. ont. Afr. 30:25-34. Brain CK, and Watson V (1992) A guide to the Swart- Cooke HBS, and Wells LH (1947) Fossil mammals from krans early hominid cave site. Ann. Trans. Mus. 35: the Makapan valley, Potgeitersrus, I11 Giraffidae. S. 343-365. Afr. J. Sci. XL111:232-235. Brickhill JAJ (1976) Some small fossil felids from Maka- De GraafT G (1960) A preliminary investigation of the pansgat Limeworks. Unpublished B.Sc. Hons. Thesis, mammalian microfauna in Pleistocene deposits of University of the Witwatersrand. caves in the Transvaal system. Palaeont. Afr. 7.59- Brink JS (1987) The archaeozoology of Florisbad, Or- 118. ange Free State. Mem. nas. Mus. (Bloemfontein) 24: De Graaff G (1988) The smaller mammals of the Cave of 1-151. Hearths from basal underlying the 248 J. K. MCKEE ET AL. deposits (circa 200,000 BPI. In RJ Mason (ed.):Cave of Hill A, Ward S, Deino A, Curtis G, and Drake R (1992) Hearths, Makapansgat, Transvaal. Archaeological Earliest Homo. Nature 355:719-722. Research Unit (University of the Witwatersrand), Oc- Hooijer DA (1958) Fossil Rhinoceroses from the lime- casional Papers No. 21, pp. 535-548. works cave, Makapansgat. Palaeont. Afr. 6:l-13. De Swardt AMJ (1974) Geomorphological dating of cave Jones DL, Brock A, and McFadden PL (1986) Palaeo- openings in South Africa. Nature 250:683. magnetic results from the Kromdraai and Sterkfon- Deacon HJ, and Geleijnse VB (1988) The stratigraphy tein hominid sites. S. Afr. J. Sci. 82:160-163. and sedimentology of the main site sequence, Klasies Klein RG (1976) The mammalian fauna of the Klasies River Mouth, South Africa. S. Afr. Arch. Bull. 43:s River Mouth site, southern Cape Province, South Af- 14. rica. S. Afr. Arch. Bull. 31:75-98. Delson E (1984) Cercopithecid biochronology of the Afri- Klein RG (1977) The mammalian fauna from the Middle can Plio-Pleistocene: Correlation among eastern and and Later Stone Age (Later Pleistocene) levels of Bor- southern hominid-bearing localities. Cour. Forsch. der Cave, Natal Province, South Africa. S. Afr. Ar- Inst. Senckenberg 69:199-218. chaeol. Bull. 32:14-27. Delson E (1988) Chronology of South African australop- Klein RG (1980) Environmental and ecological implica- ith site units. In FE Grine (ed.): Evolutionary History tions of large mammals from Upper Pleistocene and of the “Robust”Australopithecines. New York: Aldine Holocene sites in Southern Africa. Ann. S. Afr. Mus. de Gruyter, pp. 317-324. 81:223-283. Eisenhart WL (1974) Fossil Cercopithecoids of Maka- Klein RG, and Cruz-Uribe K (1991) The bovids from pansgat and Sterkfontein. Unpublished B.A. Thesis, Elandsfontein, South Africa, and their implications Harvard University. for the age, palaeoenvironment, and origins of the Ewer RF, and Cooke HBS (1964) The Pleistocene mam- site. Afr. Archaeol. Rev. 9.21-79. mals of Southern Africa. In DHS Davis (ed.): Ecologi- Klein RG, Cruz-Uribe K, and Beaumont PB (1991)Envi- cal Studies in Southern Africa. The Hague: Dr. W. ronmental, ecological, and paleoanthropological im- Junk Publishers, pp. 35-48. plications of the Late Pleistocene mammalian fauna Freedman L (1957) The fossil Cercopithecoideaof South from Equus Cave, Northern Cape Province, South Af- Africa. Ann. Transvaal Mus. 23:122-262. rica. Quat. Res. 36:94-119. Greenwood M (1955) Fossil Hystricoidea from the Kuman K (in press) The archaeology of Sterkfontein: Makapan Valley, Transvaal. Palaeont. Afr. 3t77-85. Past and present. J. Hum. Evol. Grine FE, Demes B, Jungers WL, and Cole TM, 111 Kuman K, and Clarke RJ (1986) Florisbad-new inves- (1993) Taxonomic affinity of the early Homo cranium tigations at a hominid site in South from Swartkrans, South Africa. Am. J. Phys. Anthro- Africa. Geoarchaeology 1:103-125. pol. 92:411426. Maguire JM, Schrenk F, and Stanistreet IG (1985) The Grine FE, and Klein RG (1985) Pleistocene and Ho- lithostratigraphy of the Makapansgat Limeworks locene human remains from Equus Cave, South Af- australopithecine site: Some matters arising. Ann. rica. Anthropology 8.55-98. Geol. Surv. S. Afr. 19:37-51. Grun R (1989) Electron spin resonance (ESR) dating. Mason RJ (1988) Cave of Hearths-Makapansgat, Quat. Int. 1:65-109. Transvaal. Archaeological Research Unit (University Grun R, Beaumont PB, and Stringer CB (1990a) ESR of the Witwatersrand), Occasional Paper No. 21. dating evidence for early modern at Border McFadden PL (1980) An overview of palaeomagnetic Cave in South Africa. Nature 344:537-539. chronology with special reference to the South African Grun R, Shackleton NJ, and Deacon H (1990b) Electron- hominid sites. Paleont. Afr. 23:3540. spin-resonance dating of tooth enamel from Klasies McFadden PL, Brock A, and Partridge TC (1979) Palae- River Mouth Cave. Curr. Anthropol. 31 :427432. omagnetism and the age of the Makapansgat hominid Grun R, and Stringer CB (1991) Electron spin resonance site. Earth Plant. Sci. Let. 44:373-382. dating and the evolution of modern humans. Archae- McKee JK (1991) Palaeo-ecology of the Sterkfontein ometry 33:153-199. hominids: A review and synthesis. Palaeont. Afr. 28: Harris JM, and White TD (1979) Evolution of the Plio- 41-51. Pleistocene African Suidae. Trans. Am. Phil. Soc. 69: McKee JK (1993a) The faunal age of the Taung hominid 1-128. deposit. J. Hum. Evol. 25:363-376. Helgren DM (1979) River of Diamonds: An Alluvial His- McKee JK (199313) The formation and geomorphology of tory of the Lower Vaal Basin, South Africa. Chicago: caves in calcareous tufas and implications for the University of Chicago Research Paper no. 185. study of the Taung fossil deposits. Trans. Roy. SOC.S. Hendey QB (1974) Faunal dating of the Late Cenozoic of Afr. 48:307-322. southern Africa, with special reference to the car- McKee JK (1993~)Taxonomic and evolutionary affini- nivora. Quat. Res. 4:149-161. ties of Papio izodi fossils from Taung and Sterkfon- Hendey QB (1981) Palaeoecology of the Late Tertiary tein. Palaeont. Afr. 30:4349. fossil occurrences in ’Equarry, Langebaanweg, South McKee JK (in press) Faunal turnover patterns in the Africa, and a reinterpretation of their geological con- Pliocene and Pleistocene of southern Africa. S. Afr. J. text. Ann. s.Afr. Mus. 84:l-104. Sci. SOUTHERN AFRICAN FAUNAL ASSEMBLAGE SERIATION 249

Partridge TC (1973) Geomorphological dating of cave carnivores in Africa. In JC Vogel (ed.): Late Cainozoic: opening at Makapansgat, Sterkfontein, Swartkrans Palaeoclimates of the Southern Hemisphere. Rotter- and Taung. Nature 246:75-79. dam: A. A. Balkema, pp. 49%506. Partridge TC (1982) The chronological positions of the Turner A (1986) Miscellaneous carnivore remains from fossil hominids of Southern Africa. Proc. Cong. Int. Plio-Pleistocene deposits in the Sterkfontein Valley Pal. Humaine, Nice 2:617-675. (Mammalia: Carnivora). Ann. Transvaal Mus. 34: Partridge TC (1985) Spring flow and tufa accretion at 203-226. Taung. In PV Tobias (ed.): Hominid Evolution: Past, Turner A (1987) New fossil carnivore remains from the Present and Future. New York: Alan R. Liss, pp. 171- Sterkfontein hominid site (Mammalia: Carnivora). 187. Ann. Transvaal Mus. 34:319-347. Pocock TN (1987) Plio-Pleistocene fossil mammalian mi- Turner A (1990) The evolution of the guild of larger crofauna of Southern Africa-A preliminary report in- terrestrial carnivores during the Plio-Pleistocene in cluding description of two new fossil Muroid genera. Africa. Geobios 23:349-368. Palaeont. Afr. 26:69-91. Turner A, and Wood B (1993) Taxonomic and geographic Randall RM (1981) Fossil Hyaenidae from the Maka- diveristy in robust australopithecines and other Afri- pansgat Limeworks deposit, South Africa. Palaeont. can Plio-Pleistocene larger mammals. J. Hum. Evol. Afr. 24:75-85. 24:147-168. Schwarcz HP, Griin R, and Tobias PV (1993) ESR dating Vogel JC (1985) Further attempts at dating the Taung studies of the australopithecine site of Sterkfontein, tufas. In PV Tobias (ed.): Hominid Evolution: Past, South Africa. J. Hum. Evol. 26:175-181. Present and Future. New York: Alan R. Liss, pp. 189- Simpson GG (1960) Notes on the measurement of faunal 194. resemblance. Am. J. Sci. 258-A:30&311. Vogel JC, Partridge TC (1983) Preliminary radiometric ages for the Taung tufas. In JC Vogel (ed.): Late Cain- Skinner JD, and Smithers RHN (1990) The Mammals of ozoic: Palaeoclimates of the Southern Hemisphere. the Southern African Subregion. Pretoria: University Rotterdam: A.A. Balkema, pp. 507-514. of Pretoria. Vrba ES (1975) Some evidence of chronology and palae- Szalay FS, and Delson E (1979) Evolutionary History of oecology of Sterkfontein, Swartkrans and Kromdraai the Primates. New York Academic Press. from the fossil Bovidae. Nature 254:301-304. Thackeray JF (1986) Further comment on fauna from Vrba ES (1976) The fossil Bovidae of Sterkfontein, Klasies River Mouth, South Africa. Curr. Anthropol. Swartkrans and Kromdraai. Transvaal Mus. Mem. 27.511. 21, pp. 1-166. Thackeray JF (1987) Statistical analyses of faunal data Vrba ES (1980) The significance of Bovid remains as from Klasies River Mouth. CUR. Anthropol. 29:149- indicators of environment and predation patterns. In 151. AK Behrensmeyer and AP Hill (eds.): Fossils in the Tobias PV (1971) Human skeletal remains from the Making. Chicago: University of Chicago Press, pp. Cave of Hearths, Makapansgat, Northern Transvaal. 247-271. Am. J. Phys. Anthropol. 34:335-367. Vrba ES (1981) The Kromdraai Australopithecine site Tobias PV (1973) Implications of the new age estimates revisited in 1980: Recent investigations and results. of the early South African hominids. Nature 246:7% Ann. Transvaal Mus. 33:1740. 83. Vrba ES (1982) Biostratigraphy & chronology, based Tobias PV (1975) New African evidence on the dating particularly on Bovidae, of southern hominid-associ- and phylogeny of the Plio-Pleistocene . In ated assemblages: Makapansgat, Sterkfontein, RP Suggate and MM Cresswell (eds.): Quaternary Taung, Kromdraai, Swartkrans; also Elandsfontein Studies. Wellington: Royal Society of , (Saldanha), Broken Hill (now Kabwe) and Cave of pp. 289-296. Hearths. In H deLumley and MA deLumley (eds.): Tobias PV (1978) The South African australopithecines Proc. Congres international de paleontologie hu- in time and hominid phylogeny, with special reference maine, Vol2, pp. 707-752. Nice: Union Internationale to the dating and affinities of the Taung skull. In CJ des Sciences Prehistoriques et Protohistorique. Jolly (ed): Early Hominids of Africa. London: Duck- Vrba ES (1985a) Environment and evolution: Alterna- worth, pp. 45-84. tive causes of the temporal distribution of evolution- Tobias PV (1991a) Nouvelles hypotheses sur ary events. s. Afr. J. Sci. 81:229-236. l'apparition de l'hominide dans un environment Afri- Vrba ES (1985b) Early hominids in Southern Africa: can instable. L'Anthropologie 95:379-390. Updated observations on chronological and ecological Tobias PV (1991b)Olduvai Gorge, Volume 4: The Skulls, background. In PV Tobias (ed.): Hominid Evolution: and Teeth of Homo habilis. Cambridge: Past, Present and Future. New York: Alan R. Liss, pp. Cambridge University Press. 195-200. Tobias PV, Vogel JC, Oschadleus HD, Partridge TC, and Vrba ES (1987a) New species and a new genus of Hippo- McKee JK (1993) New isotopic and sedimentological tragini (Bovidae) from Makapansgat Limeworks. measurements on the Thabaseek deposits and their Palaeont. Afr. 26:47-58. bearing on the dating of the Taung hominid. Quat. Vrba ES (1987b)A revision of the Bovini (Bovidae)and a Res. 40:360-367. preliminary revised checklist of Bovidae from Maka- Turner A (1983) Biogeography of -Recent larger pansgat. Palaeont. Afr. 26:33-46. 250 J. K. MCKEE ET AL. Vrba ES (1988)Late Pliocene climatic events and homi- accumulations, in terms of skeletal parts and nid evolution. In FE Grine (ed.): Evolutionary History represented. In CK Brain (ed.): Swartkrans: A Cave’s of the “Robust”Australopithecines. New York: Aldine Chronicle of Early Man. Pretoria: Transvaal Museum de Gruyter, pp. 405-426. Monograph No. 8, pp. 35-73. Vrba ES (1992) Mammals as a key to evolutionary the- Wolpoff MH (1974) CA* comments on Butzer. Curr. An- ory. J. Mamm. 73t1-28. thropol. 15t412413. Watson V (1993) Composition of the Swartkrans bone