Early hominid bone tools from Drimolen, South Africa

Journal of Archaeological Science 35 (2008) 2880–2894

Contents lists available at ScienceDirect

Journal of Archaeological Science journal homepage: http://www.elsevier.com/locate/jas

Early hominid bone tools from Drimolen, South Africa Lucinda Backwell a, *, Francesco d’Errico b, c a

Bernard Price Institute for Palaeontological Research, School of Geosciences, University of the Witwatersrand, Private Bag 3, Wits 2050, South Africa Institut de Pre´histoire et de Ge´ologie du Quaternaire, CNRS UMR 5199 PACEA, Universite´ Bordeaux I, Avenue des Faculte´s, F-33405 Talence, France c Institute for Human Evolution, University of the Witwatersrand, Private Bag 3, Wits 2050, South Africa b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 30 January 2008 Received in revised form 24 May 2008 Accepted 28 May 2008

The earliest use of bone tools is a topic of ongoing debate that concerns the criteria used to identify utilised or minimally modified bone tools, and if verified, the implications for hominid adaptation and cognition. Here we present the first description of 22 possible bone tools from the early hominid site of Drimolen (Gauteng Province, South Africa), dated w1.5–2 Mya. We compare the results of a taphonomic, morphometric and microscopic analysis of these pieces with those obtained from the study of faunal assemblages modified by a variety of non-human agents, and experimentally modified bones. None of the naturally modified assemblages contained pieces bearing the wear pattern observed on specimens from Drimolen and on bones experimentally used in digging activities. Fourteen pieces from Drimolen bear a pattern comparable to one previously described on early hominid bone tools from Sterkfontein and Swartkrans. This suggests that Drimolen bone tools were involved in a similar, if not exactly the same, task. Other common features include favoured bone types, fracture patterns, and the length and position of the worn area. Larger bone tools known from Swartkrans are absent at Drimolen, perhaps due to less availability of large mammal bones. The association of a high number of Paranthropus remains with bone tools at Drimolen, and the exceedingly low number of stone tools at the site supports the hypothesis that Paranthropus robustus used these bone tools. Based on implement-assisted termite foraging strategies amongst chimpanzees, we have inferred similar social and cultural behaviours for early hominids. Gorillas were recently proposed as a model for P. robustus social structure due to the high degree of sexual dimorphism observed. According to female aggregation practices present in both models, one can speculate that if P. robustus was the user of the bone tools, the foraging activity in which they were used may have been conducted mainly by females. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: Bone tools Early hominid behaviour Drimolen South Africa

1. Introduction The oldest use of bone implements, whether intentionally shaped or not, is a controversial issue. Claims that bone tools were in use since the dawn of humanity were first made by Schmidtgen (1929), Breuil (1932, 1938) and Bastin (1932), and culminated in Dart’s (1957) Osteodontokeratic culture of Australopithecus. Since then the earliest use of bone tools has been the subject of much debate, concerning the criteria used to identify intentionally used or minimally modified bone tools, and if verified, the implications of such innovation for hominid adaptation and cognition. Taphonomic studies have revealed that natural processes can produce pseudo bone artefacts. Vascular grooves have been shown to mimic cut marks (Shipman and Rose, 1984) and engravings (d’Errico and Villa, 1997), and some types of wear on teeth have been

* Corresponding author. Tel.: þ27 11 717 6672; fax: þ27 11 717 6694. E-mail addresses: [email protected] (L. Backwell), [email protected] (F. d’Errico). 0305-4403/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jas.2008.05.017

misinterpreted as carved notches (Gautier, 1986). Other mimics include natural breakage and wear of deer antler (Olsen, 1989) and elephant tusk tips (Haynes, 1991; Villa and d’Errico, 2001), gnawing or digestion by carnivores, rodents or herbivores (Pei, 1938; Sutcliffe, 1973, 1977; Binford, 1981; Villa and Bartram, 1996; d’Errico and Villa, 1997), fracture for marrow extraction by hominids and carnivores (Bunn, 1981, 1983; Gifford-Gonzalez, 1989; Backwell and d’Errico, 2004), trampling (Haynes, 1988; Olsen and Shipman, 1988), root etching (Binford, 1981), weathering (Brain, 1967), and the abrasive action of sediment (Brain, 1981; Lyman, 1994). In light of these findings it has become widely accepted that in order to distinguish between pseudo and true tools, it is necessary to adopt an interdisciplinary approach that combines analysis of bones modified by known agents, taphonomic analysis of the fossil assemblages from which the purported bone tools derive, microscopic studies of possible traces of manufacture and use, experimental replication of purported tools, and quantification of wear patterns. Employing a comparative, contextual approach has enabled authors such as Klein (1975), Shipman and Phillips (1976), Brain (1981) and Maguire et al. (1980) to contest Dart’s

L. Backwell, F. d’Errico / Journal of Archaeological Science 35 (2008) 2880–2894

‘Osteodontokeratic’ hypothesis. This has created scepticism as to whether early hominids truly used bone tools, but at the same time has provided a frame of reference for assessing new potential evidence. Robinson (1959) proposed that a metapodial shaft fragment with a smoothed pointed end from Sterkfontein Member 5 West (w1.7–1.4 Ma) was a tool. Nearly 30 years later Brain described 68 similarly modified bones and horn cores from Members 1–3 at Swartkrans, dated w1.8–1 Ma (Brain et al., 1988; Brain, 1989). Comparative scanning electron microscope inspection of replicas of the worn area on archaeological specimens and experimental shaft fragments used to extract tubers from the ground and work skins, suggested to Brain and Shipman (1993, 2004) that the wear patterns on the smoothed tips of the Swartkrans specimens matched those produced experimentally. Reanalysis of these pieces confirmed the anthropogenic origin of the use-wear (Backwell, 2000; Backwell and d’Errico, 2001; d’Errico et al., 2001). The analysis was based on (1) comparison with a large reference collection of fauna modified by 10 non-human agents, (2) morphometric and fracture analysis of bone flakes from Members 1–3 at Swartkrans, (3) experimental use of shaft fragments in a variety of tasks, (4) microscopic analysis of natural, experimental and archaeological wear patterns, and (5) quantification of striation width and orientation on archaeological and experimental specimens. The breakage patterns and size of the bone tools, compared with the remainder of the faunal remains, indicated that early hominids selected weathered, elongated and robust bone fragments for use as tools. The wear pattern on the Swartkrans tools did not match those produced experimentally to extract tubers or work skins. It fitted more closely with one created experimentally when excavating termite mounds in the area (Fig. 1), which suggested this was

2881

the most common function for the Swartkrans bone tools. Evidence of possible intentional shaping through grinding is also identified by d’Errico and Backwell (2003) on the tips of six horn cores and an ulna in the Swartkrans collection. These pieces bear flat facets close to the tip that are covered by parallel spindle-shaped striations oblique to the tool main axis, comparable to bone experimentally ground on the surfaces of termite mounds. In another paper Backwell and d’Errico (2003) described 16 new bone tools from Swartkrans and used the enlarged collection to search for patterns of variation between Members. No significant differences were observed between Members 1–3 in the type and size of the bone fragments used as tools, as well as in the length and type of the wear pattern, demonstrating that no major changes occurred through time in the bone type favoured, and the motion in which the tools were used. In addition, these authors showed that the bone tools do not share the same spatial distribution as the stone tools, suggesting the two types of artefacts did not enter the cave at the same time and may have been used in different activities, at different times, and by different members of a hominid group, or different hominid taxa. Members 1–3 at Swartkrans contain the remains of the robust australopithecine, Paranthropus robustus. Members 1 and 2 have yielded the remains of Homo erectus (Brain et al., 1988). The absence of this taxon in Member 3, from where most of the bone tools derive, suggests, but does not prove, that these implements were used by P. robustus (Backwell and d’Errico, 2003). The discovery of bone tools similar to those found at Swartkrans and Sterkfontein was announced for the nearby early hominid site of Drimolen by Keyser in 2000 (Keyser, 2000a; d’Errico et al., 2001). The aim of this paper is to provide a description of the Drimolen specimens and assess their artefactual nature. We also intend to

Fig. 1. Transparent resin replicas of wear patterns on Swartkrans, Drimolen and experimental bone tool tips photographed in transmitted light. (a) Bone tool from Swartkrans Member 2 (SKX 1142). (b) Newly identified bone tool (DN 414) from the early hominid site of Drimolen showing a rounded tip covered with longitudinal striations and flake scars smoothed by use-wear. (c) Experimental bone tool used to dig in a termite mound. (d) Bone tool used in Brain’s experiment to dig up Scilla marginata bulbs. Note the similarity in the orientation and width of the striations in (a, b and c). Scales ¼ 1 mm.

2882

L. Backwell, F. d’Errico / Journal of Archaeological Science 35 (2008) 2880–2894

Table 1 Provenance and relevant features of Drimolen bone tools Catalogue No.

Excavation date

Square

Depth (m)

Mammal size classa

Bone typeb

Wear positionc

Recent breakagec

Wear length (mm)

Figures

DN DN DN DN DN DN DN DN DN DN DN DN DN DN

7/14/1995 7/14/1995 10/8/1995 11/14/1996 10/8/1996 10/8/1996 11/19/1996 5/19/1997 11/10/1998 11/26/1998 2/15/1999 4/24/1999 7/9/1999 –

N196 E204 N196 E204 N198 E204 N195 E207 N200 E210 N200 E210 N199 E207 N195 E207 N200 E212 N199 E214 N196 E214 N197 E213 N198 E213 –

1.9 to 2.0 1.9 to 2.0 2.3 to 2.4 2.8 to 3.9 3.6 to 3.7 3.6 to 3.7 3.8 to 3.9 2.2 to 2.3 2.7 to 2.8 2.2 to 2.3 2.5 to 2.6 2.6 to 2.7 2.4 to 2.5 –

II–III II–III III II–III III–IV III II–III III II–III II–III II I II–III II

Com Com Com Sp Com Com Com ComSp Com Com Com Com Com Com

a–d a, d a–d a, b a–d a–d a–d a–d a–d a–d b–d b, c a–d a–d

e c, e

27.0 39.8 25.0 27.2 36.0 21.7 33.7 21.9 47.6 28.9 24.0 19.4 39.0 30.4

Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig.

411 412 408 406/407 413 414 409 964 963 1053 1049 1050 1040 405

e e e e e e

3b 3c 3g 3a 3d 3h 4a 4c 4b 4e 4f 3f 4d 3e

Square coordinates: N, North; E, East. a Following Brain (1974), unidentified antelope species are divided into four size classes (I–IV) based on the live weight of the animal. b Bone type: Com, compact; Sp, spongy; ComSp, compact and spongy. c Wear and breakage position: a, left side; b, periosteal surface; c, right side; d, medullary surface; e, bottom end.

explore their significance in the context of what is known about early hominid cultural traditions in eastern and southern Africa. The recent reappraisal of the Olduvai bones interpreted as tools by Leakey (1971) and Shipman (1984, 1989) confirms that at this site

shaft fragments of very large mammals were intentionally knapped, and that some elements such as the astragalus and patella were used as hammers (Backwell and d’Errico, 2004). The absence of knapped bone flakes in South Africa, and of digging implements

Fig. 2. Selection of heavily altered long bone shaft fragments from the site of Kalkbank. The pieces exhibit gross similarity in size and shape to the Swartkrans and Drimolen early hominid bone tools. See text. Scale ¼ 10 mm.

L. Backwell, F. d’Errico / Journal of Archaeological Science 35 (2008) 2880–2894

in East Africa, lends support to the hypothesis that two distinct bone tool traditions existed in Africa between 1 and 2 Ma (Backwell and d’Errico, 2004), either as variations of a single species behaviour, or due to manufacture by different hominid taxa. Considering that only a single piece is reported from Sterkfontein, this hypothesis is based on the findings from only two sites, Swartkrans and Olduvai. The existence of an additional South

2883

African Plio-Pleistocene site yielding the same type of bone tools as those from Swartkrans would demonstrate that the Swartkrans bone tools are not an isolated case, and that the digging activity in which they were used was relatively widespread and well established in the region. Comparison of bone tools between sites can reveal variation in bone type choice, applied motion and tool function. Drimolen has yielded an abundance of Paranthropus

Fig. 3. Bone tools from Drimolen. When applicable, the periosteal aspect is on the left and the medullary aspect is on the right. Lines indicate breaks. Scale ¼ 10 mm.

2884

L. Backwell, F. d’Errico / Journal of Archaeological Science 35 (2008) 2880–2894

remains and only two possible stone tools, an association that may help elucidate the early hominid user of the bone tools. 1.1. Archaeological context Drimolen is a former cave within the dolomites of the Monte Christo Formation of the Malmani Subgroup of the Chunniespoort Group (Palmer, 1991). The site, located 7 km North of Sterkfontein, was discovered by Keyser in 1992 (Keyser, 2000b; Keyser et al., 2000). Most of the ceiling of the cave has been removed by weathering and erosion. The large mammal fossil-bearing deposits consist of in situ cave sediment and collapsed blocks of dolomite, chert, large boulders of Blocky Breccia and cave siltstone, together termed Collapsed sinkhole Fill. Miners in search of limestone reportedly dumped some of the decalcified waste in the centre of

the site. Although large numbers of fossils, including hominids, have been recovered from the dump, the majority of hominid remains and bone tools have been found in the decalcified Collapsed sinkhole Fill. However, one specimen (DN 406, Fig. 3a) has calcified breccia adhering to it, suggesting that all of the bone tools derived from in situ calcified cave sediment (Keyser, personal communication). The age of the cave fill is estimated to be 1.5 to 2 Ma, based mostly on faunal correlation with Member 1 at Swartkrans (Brain, 1993). Excavation of this site has yielded 79 hominid specimens of which only two are firmly attributed to Homo sp. Interestingly, nearly half of the P. robustus specimens, which constitute the remainder of the hominid collection, are assigned to juveniles (Keyser et al., 2000). Other large mammal taxa identified (Keyser et al., 2000; Table 1) include cercopithecines, carnivores and

Fig. 4. Bone tools from Drimolen. The delineated area in (d) marks a weathered surface altered by porcupine gnawing. Lines indicate breaks. Scale ¼ 10 mm.

L. Backwell, F. d’Errico / Journal of Archaeological Science 35 (2008) 2880–2894

bovids. Noteworthy are the absence of suids and equids, and the abundance of small- to medium-sized bovids. Twenty-three bone fragments were earmarked in the field as possible bone tools. Of these, 22 were made available for analysis and form the subject of this paper.

2885

1.2. Comparative collections Human and non-human modification of bone is the topic of a large and constantly growing literature (see for example Brain, 1981; Olsen and Shipman, 1988; White, 1992; Fisher, 1995; Hannus

Fig. 5. Pseudo bone tools from Drimolen. (a) DN 1038 showing a smoothly rounded end without any evidence of striations. (b) Close up of the same tip. (c) DN 765 and (d) DN 72 showing smooth, rounded tips; (e, f) DN 1039 and (g, h) DN 1042 showing smooth, rounded ends without multiple fine parallel striations. Scale ¼ 5 mm in (a, e, and g) and 10 mm in (b–d, f, and h).

2886

L. Backwell, F. d’Errico / Journal of Archaeological Science 35 (2008) 2880–2894

et al., 1997; Plummer and Stanford, 2000; Selvaggio and Wilder, 2001; Njau and Blumenschine, 2006) which has defined the criteria to be used in the identification of agents responsible for bone alteration. Experimental replication has also refined the criteria used to identify shaping techniques and bone tool function (LeMoine, 1997; Scheinsohn, 1997; Choyke and Bartosiewicz, 2001). We have previously examined 35 reference collections of bone from both modern and fossil contexts, modified by known non-human agents (hyaena, dog, leopard, cheetah, porcupine, river gravel, spring water, plain flooding, trampling) in search for mimics of modifications recorded on the Swartkrans specimens interpreted

as bone tools (Backwell and d’Errico, 2001, 2004). All bone from these collections was analysed at macroscopic and microscopic scale using optical and scanning electron microscopy. Our research on the recognition of shaping techniques includes the reproduction and microscopic analysis of marks produced with different techniques and motions (d’Errico et al., 1984; d’Errico and Backwell, 2003; Backwell and d’Errico, 2004). We have also created and microscopically analysed a reference collection comprising bones used in a variety of tasks (Backwell and d’Errico, 2004), experimentally manipulated and transported (d’Errico, 1993a,b). In the framework of this study we have increased our comparative

Fig. 6. (a) Newly identified Drimolen bone tool (DN 413) showing a smooth, rounded tip covered with use-wear. Scale ¼ 5 mm. (b–d) Close up of wear pattern on the same object showing multiple, fine, individual striations oriented longitudinally. Scales ¼ 1 mm.

L. Backwell, F. d’Errico / Journal of Archaeological Science 35 (2008) 2880–2894

collection to include the bone assemblage from the site of Kalkbank, excavated by Mason in 1954 (Mason, 1958). Located 64 km north-west of Polokwane (formerly Pietersburg), Limpopo Province, South Africa, Kalkbank is an open air site comprised of a number of layers of calcrete, indicative of former pan surfaces. Several layers of hardpan calcrete interspersed with layers of silty sand and sandy clay overlie a 0.3 m fossil horizon, which sits atop

2887

clayey sand containing artefacts (Netterberg, 1974). It has yielded a bone assemblage composed of 3619 fragments, described by Dart and Kitching (1958) as containing Osteodontokeratic tools, and lithics attributed by Mason (1958) to the Middle Stone Age. The interest of this assemblage lies in the large number of elongated shaft fragments with pointed or rounded tips, and an overall gross morphology that is reminiscent of the Swartkrans bone tools

Fig. 7. Selection of newly identified bone tools from Drimolen showing wear pattern consisting of parallel or sub-parallel individual striations oriented longitudinally. (a, b) DN 963; (c, d) DN 409; (e, f) DN 412. Scale ¼ 1 mm in (a, b, and d–f) and 5 mm in (c).

2888

L. Backwell, F. d’Errico / Journal of Archaeological Science 35 (2008) 2880–2894

Fig. 8. Newly identified bone tool specimen DN 414 from Drimolen showing (a) wear pattern on the periosteal aspect of the specimen, and (b) resin replica of the same area observed close up in transmitted light. (c, d) Medullary aspect of the same specimen showing fine individual sub-parallel striations oriented longitudinally. Scales ¼ 1 mm.

(Fig. 2). These pieces were interpreted by Dart and Kitching as evidence that long bones were longitudinally broken to be used as tools. However, these authors mention that 38.7% of the postcranial bone fragments show porcupine gnawing. This, together with the predominance of weathered fractures (Brain, 1981) points to porcupines as the main modifying agents and accumulators of the Kalkbank faunal assemblage. In our analysis of this material we have focused on shaft fragments morphologically and dimensionally similar to those identified as bone tools from Swartkrans and Drimolen, and analysed them with the same procedures described below for the Drimolen site’s purported bone tools. 2. Method Drimolen specimens selected as potential tools were analysed with an optical microscope in search of the localised wear pattern recorded on the tips of Swartkrans, Sterkfontein and experimental bone tools. On the specimens retained as tools we recorded a range of variables related to mammal size class, type of bone used,

Table 2 Orientation of striations on each aspect of the Drimolen bone tools Catalogue Periosteala Right Medullary Left No. Long. Obl. Trans. Long. Obl. Trans. Long. Obl. Trans. Long. Obl. Trans. DN DN DN DN DN DN DN DN DN DN DN DN DN DN

411 412 408 406 413 414 409 964 963 1053 1049 1050 1040 405

1 1 1 1 1 1 1 1 1 1 1 1 1

1 1 1 1 1 1 1 1 1 1 1

1 1

1

1

1

1

1 1 1 1 1 1 1 1 1 1

1 1

1

1

1 1 1 1 1 1 1 1

1 1 1 1

1

1

1

1

1

1 1 1 1

1

1

1 1 1 1 1 1 1 1 1 1

1 1

1 1 1 1

1 1 1 1

1 1

1

1

1

Right and Left represent the sides of a bone flake when viewing the periosteal surface. Our sequence of analysis begins on the periosteal aspect, and is followed by the right side, medullar and left, as the flake is rotated counter-clockwise. Striation orientation: Long, longitudinal; Obl, oblique; Trans, transverse. a In the few instances for which the anatomical orientation is unknown, the broadest convex aspect is taken as the periosteal surface.

L. Backwell, F. d’Errico / Journal of Archaeological Science 35 (2008) 2880–2894

fracture patterns, shape and size of the bone flake, including compact bone thickness and the length and position of the worn area. The orientation of the striations (longitudinal, oblique, transverse) on the worn end was recorded for each aspect of the piece (periosteal, right, medulla, left). The same variables were recorded on the 84 specimens from Swartkrans, and the single piece from Sterkfontein. The tips of 11 specimens from Drimolen and 14 from Kalkbank, selected for their overall similarity to Swartkrans bone tools, were

2889

moulded with Coltene President light body dental elastomer, and cast using Araldite M resin (Plastomax, South Africa). All original specimens from Drimolen and transparent replicas of their tips, as well as 32 of the 590 shaft fragments studied from Kalkbank, were examined microscopically, and selected pieces photographed with a motorised Leica Z6 APOA microscope equipped with a DFC420 digital camera. Microscope and camera are connected to a desktop computer installed with Leica Application Suite (LAS) and Multifocus module software. LAS integrates the

Fig. 9. Resin replicas observed in transmitted light of the ends of pseudo bone tools from the site of Kalkbank (a–d). Microscopic examination reveals (a) porcupine gnaw marks in a good state of preservation, (b) eroded traces of porcupine or carnivore gnawing, (c) first stage of weathering, and (d) root acid etching. Scales ¼ 1 mm.

2890

L. Backwell, F. d’Errico / Journal of Archaeological Science 35 (2008) 2880–2894

Fig. 10. (a) Tip of original Drimolen bone tool (DN 1040). (b) Transparent resin replica of the same area viewed under transmitted light, revealing an unusually high proportion of transverse and oblique striations. Scales ¼ 1 mm.

Leica automated microscope, digital camera and software into a common environment. The Multifocus module is designed to acquire extended depth of field images from the microscope. Exposure, gain, shading and all other camera parameters are individually set to optimise the quality of image acquisition. Once digital images have been collected at different Z-positions, adapted algorithms combine them into a single, sharp, composite image with a significantly extended depth of field.

Fig. 11. Frequency of mammal size classes present in the (a) Drimolen and (b) Swartkrans bone tool collections. Both sites show a high representation of medium- to large-sized mammals. Size class III–IV mammals are comparatively under-represented at Drimolen.

3. Results 3.1. Assessing the artefactual nature of the Drimolen collection Microscopic inspection of the Drimolen specimens reveals that only 14 (Figs. 3 and 4, Table 1) of the 22 pieces analysed bear the features characteristic of the Swartkrans bone tools (Fig. 1a), namely a single smoothed end covered with individual striations confined to an average of ca. 30 mm from the tip. On the Swartkrans specimens most of these striations are oriented parallel or subparallel to the main axis of the object, and decrease in number away from the tip. They generally vary in width from 5 to 40 mm, and can

Fig. 12. Frequency distribution of compact bone thickness recorded in tools from (a) Drimolen and (b) Swartkrans. The Swartkrans distribution includes pieces from very large mammals, absent at Drimolen.

L. Backwell, F. d’Errico / Journal of Archaeological Science 35 (2008) 2880–2894

2891

Table 3 Ancient breakage types recorded on bone tools from Drimolen, Swartkrans and Sterkfontein Site

Drimolen Swartkrans Sterkfontein Total

Left edge

Right edge

Spiral (fresh)

Longitudinal (weathered)

Total

Spiral (fresh)

Longitudinal (weathered)

Total

3 3 0 7

10 37 1 48

13 40 1 55

2 3 0 5

10 43 1 54

12 46 1 59

reach in rare cases 150 mm. A few striations oriented perpendicular to the main axis, generally posterior to the longitudinal parallel ones, are also recorded on some specimens, and range between 100 and 400 mm in width. The eight Drimolen specimens interpreted as pseudo tools (Fig. 5) show a rounded and in some cases smoothed/polished end broadly similar to those observed on true tools. At microscopic scale, however, their tips lack the superimposition of individual oriented striations diagnostic of use as digging implements in a fine sedimentary environment (Fig. 1a–c). While these pieces may well be bone tools that have lost their striations through surface deterioration, or were implements used for a purpose other than digging, they could equally have acquired their rounded ends from another process altogether, and so are considered here as pseudo tools and eliminated from further analysis. In contrast, the 14 pieces interpreted as tools from Drimolen (Figs. 6–8) record exactly the same wear pattern observed at Swartkrans. All record multiple fine parallel or sub-parallel striations oriented longitudinal with the main axis, and some show a few broad transverse striations generally posterior to the longitudinal parallel ones (Table 2). Two pieces (Figs. 3h and 4e) show flake scars originating from the tip, the edges of which are partially smoothed by wear (Figs. 1b and 8a). Also recorded on Swartkrans (Backwell and d’Errico, 2001; Fig. 1a) and experimental tools, these flake scars probably result from impact of the tip during use. A single specimen (Fig. 4d) shows, in addition to the fine longitudinal striations, an unusually high proportion of transverse and oblique striations (Fig. 10). Many pieces from Kalkbank have smoothed rounded tips. However, the smoothing generally extends over the entire piece (Fig. 2). Microscopic analysis was unable to find a single specimen from this site bearing the wear pattern observed on the Swartkrans and Drimolen bone tools. The surface of a high proportion of these fragments is damaged by porcupine gnawing, which is well preserved in some specimens (Fig. 9a), and eroded/ smoothed in others. The tips of several specimens bear shallow pits and grooves, some of which may be eroded traces of porcupine and carnivore gnawing (Fig. 9b). Others show weathering of the outer bone surface (Fig. 9c), dissolution and possible root marks (Fig. 9d).

Fig. 13. Frequency distribution of the length of the worn area on the (a) Drimolen and (b) Swartkrans bone tools. At both sites most pieces have a wear length ranging between 20 and 40 mm.

3.2. Comparing Drimolen, Sterkfontein and Swartkrans bone tools Besides the wear pattern, the Drimolen bone tools exhibit a number of other features similar to those recorded on the Sterkfontein and Swartkrans bone tools. Some differences are also apparent. At the latter sites the tools derive predominantly from mammal size classes II–III and III–IV, while at Drimolen class III–IV is comparatively under-represented (Fig. 11). The use of more robust bone fragments at Swartkrans is also reflected in the relatively high number of fragments with a compact bone thickness in excess of 10 mm (Fig. 12). As at Swartkrans, the large majority of Drimolen tools derive from long bone shaft fragments. The four exceptions – two mandible fragments (Fig. 4d and f), a rib (Fig. 4c), and a horn core (Fig. 3a) – are not in contradiction with the trend observed at Swartkrans, where one mandible, seven ribs, and 14 horn cores out of 84 pieces were used as tools. The choice for weathered, elongated, straight bone flakes (as opposed to small spirally fractured fresh fragments) already noted for Swartkrans, is also evident at Drimolen, where almost all the shaft fragments bear longitudinal breakage (Table 3). As at the former site, almost all Drimolen tools are represented only by their tips (Table 1), which frustrates analysis of intra- and inter-site length variation. However, analysis of the width and thickness of complete tools at 5, 10, 15 and 20 mm from the tip (Table 4) reveals a remarkable dimensional similarity between the three sites. The slightly more slender trend observed at Drimolen is in accordance with the low incidence of size class III–IV mammal shafts used as tools.

Table 4 Size of bone tools at 5, 10, 15 and 20 mm from the tip Site

Size at 5 mm from the tip No. of tools

Size at 10 mm from the tip

Size at 15 mm from the tip

Size at 20 mm from the tip

Mean (mm)

s.d.

No. of tools

Mean (mm)

s.d.

No. of tools

Mean (mm)

s.d.

No. of tools

Mean (mm)

s.d.

Bone tool width Drimolen 13 Swartkrans 65 Sterkfontein 1

9.1 8.9 8.0

2.9 2.8

13 60 1

11.0 11.7 11.5

3.0 3.4

13 56 1

12.5 13.9 13.0

3.0 4.1

11 49 1

14.0 15.5 14.5

2.9 4.6

Bone tool thickness Drimolen 13 Swartkrans 76 Sterkfontein 1

6.5 6.4 4.0

1.9 2.1

13 76 1

7.4 8.3 7.0

1.9 2.9

13 73 1

7.8 9.2 9.0

2.0 3.2

12 65 1

8.7 10.5 12.0

2.5 3.5

2892

L. Backwell, F. d’Errico / Journal of Archaeological Science 35 (2008) 2880–2894

Table 5 Number of bone tools recording longitudinal, oblique and transverse striations on the four aspects Site

Drimolen Swartkrans Sterkfontein

No. of tools

14 81 1

Periosteala

Right

Medullary

Left

Long.

Obl.

Trans.

Long.

Obl.

Trans.

Long.

Obl.

Trans.

Long.

Obl.

Trans.

12 (86) 70 (86) 1

11 (79) 38 (47) 0

3 (21) 16 (20) 0

12 (86) 56 (69) 1

5 (36) 23 (28) 0

4 (29) 6 (7) 0

12 (86) 64 (79) 1

5 (36) 22 (27) 0

3 (21) 8 (10) 0

12 (86) 55 (68) 1

7 (50) 20 (26) 0

6 (43) 4 (5) 0

Right and left represent the sides of a bone flake when viewing the periosteal surface. Our sequence of analysis begins on the periosteal aspect, and is followed by the right side, medullar and left, as the flake is rotated counter-clockwise. Striation orientation: Long, longitudinal; Obl, oblique; Trans, transverse. Numbers in parentheses are the percentage (%) of bone tools that record the different striation orientations on the four aspects. a In the few instances for which the anatomical orientation is unknown, the broadest convex aspect is taken as the periosteal surface.

The length of the worn area on the Drimolen specimens falls well within the range of that observed at Swartkrans (Fig. 13) with the majority of the wear ranging between 20 and 40 mm. The lack of pieces with a very short or long worn area at Drimolen is most likely a function of the small sample size. At all three sites the occurrence and location of differently oriented striations shows an abundance of longitudinal relative to oblique, and rarity of transverse striations (Table 5). Microscopic inspection of the worn tips has highlighted another feature that parallels what we have observed at Swartkrans. The only horn core from Drimolen records an area close to the tip covered with parallel striations perpendicular to the main axis (Fig. 14). These closely match those described on six horn cores from Swartkrans (d’Errico and Backwell, 2003), interpreted as possible evidence of grinding, for the purpose of tapering or re-sharpening the functional area of the tool. Finally, among the pieces interpreted as bone tools, two have punctures probably produced by small carnivores (Fig. 4a and f) and one piece (Fig. 4d) shows at the end opposite to the worn tip, damage consisting of shallow grooves reminiscent of porcupine gnawing. Carnivore damage is recorded on one of the bone tools from Swartkrans (d’Errico and Backwell, 2003: p. 1565, Fig. 3), where it is quite clearly covered with use-wear. 4. Discussion The wear pattern found on the tips of bone fragments from the sites of Swartkrans, Sterkfontein and Drimolen has so far not been found in any comparative collections that we have studied. Kalkbank is amongst the collections in which the proportion of elongated bone fragments with rounded tips is high. Still, a close scrutiny reveals no match with the bones interpreted as early

hominid tools. Considering the close similarity in location, extent and microscopic features between the wear on the Swartkrans, Sterkfontein and Drimolen specimens and that produced experimentally in digging activities, an anthropogenic origin for the observed modifications is retained as the best-fit hypothesis. In order to strengthen this interpretation, detailed inspection of faunal remains and morphometric analysis of shaft fragments from Drimolen will be conducted as soon as the material becomes available for study. Functional interpretation of the tools from these three sites may also be refined. Quantification of the wear pattern has shown termite foraging to be the closest match. However, this interpretation must be verified by enlarging the variety of tasks in which similar bone tools may be used experimentally, and by refining the methods used to quantify the resulting surface modification. This is the goal of our ongoing research project (Backwell and d’Errico, 2005). For the purposes of this paper we observe a striking similarity between bone tools from southern African hominid sites, which suggests that they were involved in a similar, if not exactly the same, task. The only possible difference between the bone tools from Drimolen and those from Swartkrans – rarity of very robust pieces at the former site – may be due either to the small size of the sample, or to different availability of weathered shaft fragments, or to a preference for slightly smaller fragments at Drimolen. The second hypothesis is partially supported by the rarity of very large mammals reported for this site (Keyser et al., 2000). The third hypothesis is almost contradicted by the morphometry of the tool tips, showing, in spite of the use of thinner blanks at Drimolen, a strong similarity between sites. These hypotheses may be tested in future, when results of taphonomic analysis of the faunal assemblage from this site become available. More detailed contextual information may also elucidate why, in three instances, bone tools have been found in very close proximity,

Fig. 14. Newly identified bone tool (DN 406) from Drimolen, representing the only horn core in the collection. It shows a smooth rounded end (left), which at microscopic scale records possible evidence of grinding (right). Scales ¼ 1 mm.

L. Backwell, F. d’Errico / Journal of Archaeological Science 35 (2008) 2880–2894

and why others cluster in a 4 m2 area, within a 50 cm depth of each other (Table 1). Spatial analysis of hominid and other mammal remains, bone and possible stone tools, is also required to help address the question of who used the bone tools. Lockwood et al. (2007) recently proposed that the high degree of sexual dimorphism observed in P. robustus from Swartkrans, Drimolen and Kromdraai, all situated in the Cradle of Humankind, South Africa, may reflect social structures comparable to that of gorillas, and that stable groups of females formed in response to the distribution of food sources. Female aggregation is also known amongst chimpanzees, where in some populations tools are used to forage for termites (Whiten et al., 1999). Based on implement-assisted termite foraging strategies amongst chimpanzees, we have inferred similar cultural behaviours for early hominids (d’Errico et al., 2001). In light of the female aggregation models mentioned, we can speculate that if P. robustus was the user of the bone tools, then the foraging activity in which they were used may have been conducted mainly by females. 5. Conclusion Although the presence of bone tools at Drimolen has been mentioned before in the literature, a detailed assessment of this material was needed. Our study has reduced from 23 to 14 the number of pieces bearing modifications similar to those recorded on bone tools from other southern Africa early hominid sites, and confirmed – by comparing the Drimolen specimens to an expanded comparative collection – that the wear pattern recorded on the tips most likely results from use in digging, which we posit was associated with termite foraging. This finding demonstrates that the use of bone tools at early hominid sites was apparently widespread, and considering available information on fauna, geochronology and site formation processes at Swartkrans and Drimolen (Vrba, 1982; Brain, 1993; Watson, 1993; Turner, 1997; Keyser et al., 2000; Curnoe et al., 2001; de Ruiter, 2003), lasted unchanged for at least 300,000 years, and possibly as long as one million years. The association of a high number of Paranthropus remains with bone tools and the virtual absence of stone tools at Drimolen supports the hypothesis, first proposed by Brain and Shipman (1993) and followed by us (2001, 2003), that P. robustus may have used these bone tools. Acknowledgements We thank Andre Keyser and Colin Menter for inviting us to describe the material, and for polishing the final version of the manuscript. We are grateful to Dominique Armand for assistance with bone identification, Cynthia Kemp for editing a draft of this manuscript, and two anonymous reviewers for their helpful comments. This research was funded by the National Research Foundation (NRF) in South Africa, University of the Witwatersrand Research Council (URC), Palaeontological Scientific Trust (PAST), Cultural Service of the French Embassy in South Africa, French Ministry of Research and Education, and Origin of Man, Language and Languages (OMLL) Programme of the European Science Foundation. References Backwell, L.R., 2000. A Critical Assessment of Southern African Early Hominid Bone Tools. MSc. dissertation, University of the Witwatersrand, South Africa. Backwell, L.R., d’Errico, F., 2001. Evidence of termite foraging by Swartkrans early hominids. Proceedings of the National Academy of Sciences of the United States of America 98 (4), 1358–1363. Backwell, L.R., d’Errico, F., 2003. Additional evidence on the early hominid bone tools from Swartkrans with reference to spatial distribution of lithic and organic artefacts. South African Journal of Science 99, 259–267. Backwell, L.R., d’Errico, F., 2004. A reassessment of the Olduvai Gorge ‘bone tools’. Palaeontologia Africana 40, 95–158.

2893

Backwell, L.R., d’Errico, F., 2005. The origin of bone tool technology and the identification of early hominid cultural traditions. In: d’Errico, F., Backwell, L.R. (Eds.), From Tools to Symbols. From Early Hominids to Modern Humans. Wits University Press, Johannesburg, pp. 238–275. Bastin, A.H., 1932. Nouvelle contribution a` l’e´tude de l’utilisation de l’os au Pale´olithique infe´rieur. 1. L’e´pieu acheule´en de Saint-Mard (Aisne). 2. L’hume´rus de´bite´ de Saratov (U.R.S.S.). Bulletin de la Socie´te´ Pre´historique Française 29, 458–465. Binford, L.R., 1981. Bones, Ancient Men and Modern Myths. Academic Press, New York. Brain, C.K., 1967. Bone weathering and the problem of bone pseudo-tools. South African Journal of Science 63, 97–99. Brain, C.K., 1974. Some suggested procedures in the analysis of bone accumulations from southern African Quaternary sites. Annals of the Transvaal Museum 29 (1), 1–8. Brain, C.K., 1981. The Hunters or The Hunted? An Introduction to African Cave Taphonomy. University of Chicago Press. Brain, C.K., 1989. The evidence of bone modification by early hominids in southern Africa. In: Bonnichsen, R., Sorg, M.H. (Eds.), Bone Modification. University of Maine Centre for the Study of the First Americans. Thompson-Shore Inc., Orono, United States of America, pp. 291–297. Brain, C.K., 1993. Swartkrans. A Cave’s Chronicle of Early Man. In: Transvaal Museum Monograph, No. 8 Pretoria. Brain, C.K., Shipman, P., 1993. The Swartkrans bone tools. In: Brain, C.K. (Ed.), Swartkrans: a Cave’s Chronicle of Early Man. Palaeontology Scientific Trust, pp. 195–215. Brain, C.K., Shipman, P., 2004. The Swartkrans bone tools. In: Brain, C.K. (Ed.), Swartkrans, a Cave’s Chronicle of Early ManTransvaal Museum Monograph, second ed., No. 8, pp. 195–215. Brain, C.K., Churcher, C.S., Clark, J.D., Grine, F.E., Shipman, P., Susman, R.L., Turner, A., Watson, V., 1988. New evidence of early hominids, their culture and environment from the Swartkrans cave, South Africa. South African Journal of Science 84, 828–835. Breuil, H., 1932. Pointe d’e´pieu en os du Mouste´rien de la caverne du Castillo (Santander). L’Anthropologie 42, 679–680. Breuil, H., 1938. Bone and antler industry of the Chou-Koutien Sinanthropus site. Paleontologia Sinica n.s. 6 Pe´kin. Bunn, H.T., 1981. Archaeological evidence for meat-eating by Plio-Pleistocene hominids from Koobi Fora and Olduvai Gorge. Nature 291, 574–577. Bunn, H.T., 1983. Comparative analysis of modern bone assemblages from a San hunter-gatherer camp in the Kalahari desert, Botswana, and from a spotted hyena den near Nairobi, Kenya. In: Clutton-Brock, J., Grigson, C. (Eds.), Animals and Archaeology: 1. Hunters and their Prey. British Archaeological Reports International Series 163, pp. 143–148. Choyke, A.M., Bartosiewicz, L., 2001. Crafting Bone: Skeletal Technologies Through Time and Space. British Archaeological Reports International Series 937, Oxford. Curnoe, D., Gru¨n, R., Taylor, L., Thackeray, F., 2001. Direct ESR dating of a Pliocene hominin from Swartkrans. Journal of Human Evolution 40, 379–391. Dart, R.A., 1957. The Osteodontokeratic Culture of Australopithecus prometheus. Transvaal Museum Memoirs, No. 10, Pretoria. Dart, R., Kitching, J., 1958. Bone tools at the Kalkbank Middle Stone Age site and the Makapansgat australopithecine locality, central Transvaal. Part 2. The Osteodontokeratic contribution. Archaeological Bulletin 13 (51), 94–116. d’Errico, F., 1993a. Identification des traces de manipulation, suspension, polissage sur des objets d’art mobilier en os, bois de cervide´s, ivoire. In: Traces et fonction: les Gestes Retrouve´s. ERAUL, Lie`ge, pp. 177–188. 50. d’Errico, F., 1993b. Criteria for identifying utilised bone: the case of the Cantabrian ‘Tensors’. Current Anthropology 34 (3), 298–311. d’Errico, F., Villa, P., 1997. Holes and grooves: the contribution of microscopy and taphonomy to the problem of art origins. Journal of Human Evolution 33, 1–31. d’Errico, F., Backwell, L.R., 2003. Possible evidence of bone tool shaping from the early hominid site of Swartkrans, South Africa. Journal of Archaeological Science 30, 1559–1576. d’Errico, F., Giacobini, G., Puech, P.-F., 1984. Varnish replicas. A new method for studying worked bone surfaces, OSSA. International Journal of Skeletal Research 9/10, 29–51. d’Errico, F., Backwell, L.R., Berger, L.R., 2001. Bone tool use in termite foraging by early hominids and its impact on understanding early hominid behaviour. South African Journal of Science 97, 71–75. Fisher, J.W., 1995. Bone surface modifications in zooarchaeology. Journal of Archaeological Method and Theory 2 (1), 7–68. Gautier, A., 1986. Une histoire de dents: les soi-disant incisives travaille´es du Pale´olithique moyen de Sclayn. He´linium 26, 177–181. Gifford-Gonzalez, D., 1989. Ethnographic analogues for interpreting modified bones: some cases from East Africa. In: Bonnichsen, R., Sorg, M.H. (Eds.), Bone Modification. University of Maine Centre for the Study of the First Americans. Thompson-Shore Inc., Orono, United States of America, pp. 179–246. Hannus, L.A., Rossum, L., Winham, R.P., 1997. Proceedings of the 1993 Bone Modification Conference, Hot Springs, South Dakota. In: Occasional Publication, No. 1. Archeology Laboratory, Augustana College, Sioux Falls, USA. Haynes, G., 1988. Longitudinal studies of African elephant death and bone deposits. Journal of Archaeological Science 15, 131–157. Haynes, G., 1991. Mammoths, Mastodonts, and Elephants: Biology, Behaviour and the Fossil Record. Cambridge University Press. Keyser, A., May 2000a. New finds in South Africa, National Geographic, pp. 77–83.

2894

L. Backwell, F. d’Errico / Journal of Archaeological Science 35 (2008) 2880–2894

Keyser, A., 2000b. The Drimolen skull: the most complete australopithecine cranium and mandible to date. South African Journal of Science 96, 189–193. Keyser, A., Menter, C.G., Moggi-Cecchi, J., Pickering, T.R., Berger, L.R., 2000. Drimolen: a new hominid-bearing site in Gauteng, South Africa. South African Journal of Science 96, 193–197. Klein, R.G., 1975. Palaeoanthropological implications of the non-archaeological bone assemblage from Swartklip I, south-western Cape Province, South Africa. Quaternary Research 5, 275–288. Leakey, M.D., 1971. Olduvai Gorge. Vol. 3. Excavations in Beds I and II. Cambridge University Press. LeMoine, G., 1997. Use Wear Analysis on Bone and Antler Tools of the Mackenzie Inuit. British Archaeological Reports International Series 679, Oxford. Lockwood, C., Menter, C., Keyser, A., Moggi-Cecchi, J., 2007. Extended male growth in a fossil hominin species. Science 318, 1443–1446. Lyman, R.L., 1994. Vertebrate Taphonomy. Cambridge University Press. Maguire, J.M., Pemberton, D., Collett, M.H., 1980. The Makapansgat Limeworks Grey breccia: hominids, hyaenas, hystricids or hillwash? Palaeontologia Africana 23, 75–98. Mason, R., 1958. Bone tools at the Kalkbank Middle Stone Age site and the Makapansgat australopithecine locality, central Transvaal. Part 1. The Kalkbank site. Archaeological Bulletin 13 (51), 85–93. Netterberg, F., 1974. Observations on calcretes at some archaeological and palaeontological sites in southern Africa. The South African Archaeological Society Goodwin Series 2, 20–24. Njau, J.K., Blumenschine, R., 2006. A diagnosis of crocodile feeding traces on larger mammal bone, with fossil examples from the Plio-Pleistocene Olduvai Basin, Tanzania. Journal of Human Evolution 50, 142–162. Olsen, S.L., 1989. On distinguishing natural from cultural damage on archaeological antler. Journal of Archaeological Science 16, 125–135. Olsen, S.L., Shipman, P., 1988. Surface modification of bone: Trampling versus butchery. Journal of Archaeological Science 15, 535–553. Palmer, A.N., 1991. Origin and morphology of limestone caves. Geological Society of America Bulletin 103, 1–21. Pei, W.C., 1938. Le roˆle des animaux et des causes naturelles dans la cassure des os. Palaeontologia Sinica 7, 1–16. Plummer, T.W., Stanford, C.B., 2000. Analysis of a bone assemblage made by chimpanzees at Gombe National Park, Tanzania. Journal of Human Evolution 39, 345–365. de Ruiter, D.J., 2003. A revised faunal list for the Australopithecine-bearing Members of Swartkrans. Annals of the Transvaal Museum 39, 29–41. Robinson, J.T., 1959. A bone implement from Sterkfontein. Nature 184, 583–585.

Scheinsohn, V.G., 1997. Use-wear patterns on bark removers. In: Hannus, L.A., Rossum, L., Winham, R.P. (Eds.), Proceedings of the 1993 Bone Modification Conference, Hot Springs, South Dakota. Occasional Publication No. 1. Archeology Laboratory, Augustana College, Sioux Falls, USA, pp. 265–276. Schmidtgen, O., 1929. Knochenartefakte aus den Mosbacher Sanden. Jahrbu¨cher des Nassauischen Vereins fu¨r Naturkunde, Wiesbaden. 80, 1–6. Selvaggio, M.M., Wilder, J., 2001. Identifying the involvement of multiple carnivore taxa with archaeological bone assemblages. Journal of Archaeological Science 28, 465–470. Shipman, P., 1984. The earliest tools: re-assessing the evidence from Olduvai Gorge. Anthroquest 29, 9–10. Shipman, P., 1989. Altered bones from Olduvai Gorge, Tanzania: techniques, problems, and implications of their recognition. In: Bonnichsen, R., Sorg, M.H. (Eds.), Bone Modification. University of Maine Centre for the Study of the First Americans. Thompson-Shore Inc., Orono, United States of America, pp. 317–334. Shipman, P., Phillips, J., 1976. On scavenging by hominids and other carnivores. Current Anthropology 17 (1), 170–172. Shipman, P., Rose, J., 1984. Cutmark mimics on modern and fossil bovid bones. Current Anthropology 25 (1), 116–117. Sutcliffe, A.J., 1973. Similarity of bones and antlers gnawed by deer to human artefacts. Nature 246, 428–430. Sutcliffe, A.J., 1977. Further notes on bones and antler chewed by deer and other ungulates. Deer 4 (2), 73–82. Turner, A., 1997. Further remains of Carnivora (Mammalia) from the Sterkfontein hominid site. Palaeontologia Africana 34, 115–126. Villa, P., Bartram, L., 1996. Flaked bone from a hyena den. Pale´o 8, 143–159. Villa, P., d’Errico, F., 2001. Bone and ivory points in the Lower and Middle Paleolithic of Europe. Journal of Human Evolution 41, 69–112. Vrba, E.S., 1982. Biostratigraphy and chronology, based particularly on Bovidae, of southern hominid-associated assemblages: Makapansgat, Sterkfontein, Taung, Kromdraai, Swartkrans; also Elandsfontein (Saldanha), Broken Hill (now Kabwe) and Cave of Hearths. In: De Lumley, M.A. (Ed.), Homo erectus et la Place de l’Homme de Tautavel parmi les Hominide´s Fossiles. Premie`re Congre`s International de Pale´ontologie Humaine, Vol. 2. UNESCO Colloque International du Centre National de la Re´cherche Scientifique, France, pp. 707–752. Watson, V., 1993. In: Brain, C.K. (Ed.), Swartkrans. A Cave’s Chronicle of Early Man. Transvaal Museum Monograph, No. 8, pp. 35–73. Pretoria. White, T.D., 1992. Prehistoric Cannibalism at Mancos 5MTUMR-2346. Princeton University Press, New Jersey. United States of America. Whiten, A., Goodall, J., McGrew, W.C., Nishida, T., Reynolds, V., Sugiyama, Y., Tutin, C. E.G., Wrangham, R.W., Boesch, C., 1999. Cultures in chimpanzees. Nature 399, 682–685.