Human manual distal phalanges from the Middle Stone Age deposits of Klasies River Main Site, Western Cape Province, South Africa

Abstract

Two new distal manual phalanges from the Middle Stone Age deposits of Klasies River Main Site are described. One (SAM-AP 6387) likely derives from ray II or ray III, whereas the other (SAM-AP 6388) is from the thumb. Both derive from a late adolescent or fully adult individual. They were recovered by H. Deacon from the same stratigraphic unit (submember W or possibly submember R) of the Shell and Sand Member of Cave 1, which places them between 100 and 90 ka. Both are comparatively small elements, and the possibility that they came from the same hand cannot be discounted at this time. These bones add to the meager and all too fragmentary postcranial human fossil sample from the Late Pleistocene of South Africa. These two specimens provide some additional evidence pertaining to the morphological attributes of the distal phalanges of the Middle Stone Age inhabitants of South Africa. Together with the distal pollical phalanx from Die Kelders (SAM-AP 6402), they are relatively small in comparison with homologs from recent human samples as well as Late Pleistocene specimens from Eurasia. Given their small sizes, the distal pollical phalanges from Klasies and Die Kelders are not dissimilar to Holocene Khoesan homologs. As expected, the Klasies elements differ noticeably from Neandertal homologs, especially in the narrowness of their shafts and distal tuberosities.

Introduction

Klasies River Main Site (KRMS; 34°6′29.69″S, 24°23′25.95″E) is a large cave system comprising a series of interrelated recesses referred to as Caves 1, 1A, 1B, and 2 that have eroded into a Table Mountain Sandstone cliff bordering the Indian Ocean along the Tsitsikamma coast of the Eastern Cape Province of South Africa (Fig. 1). Among the dozen or so southern African Middle Stone Age (MSA) sites with human fossils that date to Marine Isotope Stage (MIS) 6–3 (Grine, 2016), KRMS has featured prominently in discussions of the emergence of modern human behavior and morphology (Grine et al., 2017a). This is owing to its rich artifactual and faunal records (e.g., Klein, 1976, 1989; Singer and Wymer, 1982; Thackeray and Kelly, 1988; Thackeray, 1989; Turner, 1989; Deacon, 1992, 2008; Klein and Cruz-Uribe, 1996; Milo, 1998; Bartram and Marean, 1999; Klein et al., 1999; Wurz, 1999, 2002, 2008, 2012; van Pletzen, 2000; Wurz et al., 2003; von den Driesch, 2004; Deacon and Wurz, 2005; McCall, 2006, 2007; Dusseldorp, 2010; d’Errico et al., 2012) together with a comparatively abundant assemblage of human remains (e.g., Singer and Wymer, 1982; White, 1987; Rightmire and Deacon, 1991, 2001; Bräuer et al., 1992; Smith, 1992; Frayer et al., 1993; Churchill et al., 1996; Lam et al., 1996; Bräuer and Singer, 1996a, b; Pearson and Grine, 1997; Grine et al., 1998; Groves, 1998; Pearson et al., 1998; Groves and Thorne, 2000; Rightmire et al., 2006; Royer et al., 2009; Grine et al., 2017a).

The site was the focus of 14 months of intensive excavations by J. Wymer in 1967–1968 (Singer and Wymer, 1982). It was subsequently excavated by H. Deacon on a seasonal basis from 1984 to 1995 (Deacon, 1989, 1995, 2001; Deacon and Wurz, 2005), and S. Wurz has been working the site since 2013. Most of the work has focused on Cave 1 and Cave 1A deposits. Wymer recovered human remains from Caves 1, 1A, and 1B, with the majority coming from Cave 1; Deacon obtained human fossils from Caves 1 and 1A.

The hominin assemblage from KRMS comprises nearly 50 fossils spread across three stratigraphic members that span some 40 kyr (Grine et al., 2017a). However, there is a very strong bias toward cranial remains, even to the exclusion of isolated teeth. Thus, only nine postcranial elements have been documented so far (Table 1). This meagerness is in keeping with other hominin-bearing MSA sites in southern Africa, where postcranial elements are also sparsely manifest (Grine et al., 2017b). As a result of this paucity, every discovery is worthy of documentation. We here provide descriptions and comparative analyses of two distal manual phalanges from the MSA deposits of KRMS.

Singer and Wymer (1982) enumerated some seven or eight MSA artifact-bearing horizons in Cave 1 (numbered from top to bottom: 13 through 39), with the bottom of the sequence (layer 40) being clean beach gravel and shingle. They recognized 39 MSA layers in Cave 1A (numbered 1–39 from top to bottom), 15 MSA layers in Cave 1B (numbered 1–15 from top to bottom), and five MSA layers in Cave 2.

Deacon and Geleijnse (1988) developed a stratigraphic framework in which levels with broadly similar lithological characteristics were combined into members. Deacon and Geleijnse (1988) combined the various MSA artifact-bearing layers identified by Singer and Wymer (1982) into five named members. From bottom to top, these were designated the Light Brown Sand (LBS), Shell and Sand (SAS), Rock Fall (RF), Upper, and White Sand (WS). No single repository in the Main Site complex preserves all five. The LBS, SAS, and WS members are represented in Cave 1, the LBS, SAS, RF, and Upper are represented in Cave 1A, the LBS and SAS are represented in Cave 1B, and the RF and Upper are represented in Cave 2. The stratigraphic units (i.e., members and submembers) identified by Deacon and Geleijnse (1988) and their correlations with the layers designated by Singer and Wymer (1982) are shown in Figure 2. These relationships have been described in detail by Grine et al. (2017a).

Of relevance to the two phalanges described herein, the deposits that comprise the SAS member in Cave 1 were viewed by Singer and Wymer (1982) as representing a series of separate occupational layers (layer 15 and most of layer 17) interspersed with layers of rubble (part of layers 17, 16, and 14) that had rolled down from the large pile of occupational debris that had built up near Cave 1A. In agreement, Deacon and Geleijnse (1988) described SAS-L (layer 17) as an occupation level and SAS-R (layer 14) as talus slope deposits. However, Deacon and Geleijnse (1988) considered SAS-U (layer 16) as an occupational layer and SAS-W (layer 15) as deriving from a talus slope that blocked off the mouth of Cave 1, rendering it uninhabitable through SAS-W and SAS-R times. Subsequently, however, Rightmire and Deacon (1991, 2001) suggested that all the material in the SAS member could have derived from the same depositional event, although the reason for this conclusion was not given.

Singer and Wymer (1982) acknowledged that the provenience of material recovered from this particular area of Cave 1 is questionable, and Rightmire and Deacon (1991) argued that the context of the finds from SAS W (layer 15) is geological rather than an in situ archaeological deposit. They also claimed that SAS R (layer 14) in Cave 1 occupies a broad channel–like feature (i.e., a drainage sump) against the back slope of SAS W such that SAS W represents slope wash from Cave 1A.

Vogel (2001) obtained U-Th estimates of 85.2 ± 2.1 ka, 94.6 ± 3.2 ka, and 100.8 ± 7.5 ka for three stalagmite samples from SAS-R (layer 14). Although Vogel (2001) also obtained a U-Th age of only 36.6 ka for shell from SAS W (layer 15) in Cave 1, he noted that shells produce unreliable U-series dates owing to their tendency to absorb uranium from percolating water long after deposition. This results in an age that is substantially younger than the shell itself. Based on δ¹⁸O values for shell from layer 15 (SAS W) of Cave 1, Shackleton (1982) and Deacon et al. (1988) have argued for its accumulation during MIS 5c, that is, somewhere between ca. 105–96 ka according to the SPECMAP dates (Imbrie et al., 1984) and the chronologies of Wright (2000: fig. 4) and Lisiecki and Raymo (2005). Thus, whether the finger bones derive from SAS W or SAS R, they most likely date to ca. 100–90 ka.

We here document two manual distal phalanges from the MSA deposits at KRMS that have not previously been afforded either description or comparative analysis. Both elements were recovered by Deacon during the course of his excavations of the Witness Baulk in Cave 1. This ‘ridge’ of sediment was left intact by Wymer as a witness section; it has been the focus of subsequent work by Deacon in the 1990s and is currently being excavated by Wurz (Wurz et al., 2018).

The first bone, SAM-AP 6387, was mentioned in passing by Rightmire and Deacon (2001) and discussed briefly by Grine et al. (2017a) in the context of the hominin fossil record from the site. Rightmire and Deacon (2001: 538) recorded that it derived from the talus deposit of submember W (i.e., layer 15) of the SAS member (Fig. 2), and they interpreted this to mean that it had derived from “a horizon higher in the sequence.” This might suggest layer 14 (SAS R). Grine et al. (2017a) opined that the size of the bone suggests that it may have come from the index finger (i.e., ray II), and while they noted that the side from which it derived cannot be determined, they treated it as a right for illustrative purposes only.

The second specimen, SAM-AP 6388, was recovered in 1991 by Deacon, also during his excavation of the Witness Baulk in Cave 1. Deacon's notes indicate that he regarded it as also having come from SAS W. It is possible that it derives from the same stratigraphic context as SAM-AP 6387 (but see the following section). This finger bone was not included by Grine et al. (2017a) in their documentation of the human fossils from the MSA of KRMS because it was unpublished at the time.