sign in sign up
<<
Home , Diepkloof Rock Shelter, Elands Bay Cave, Revil Mason

Article

Bushman (Limpopo, ): a perspective from the edge of the highveld

PORRAZ, Guillaume, et al.

Abstract

In this paper, we introduce a recently initiated research project conducted at Bushman Rock Shelter, on the northeastern edge of the Highveld plateau in Limpopo Province, South Africa. Previous excavations carried out at the site during the 1960s and 1970s exposed a deep and -stratified sequence of c. 7 metres of archaeological deposits associated with Later and Middle occupations (LSA and MSA). Owing to the lack of contextual information, Bushman Rock Shelter remains poorly studied despite recording cultural and palaeoenvironmental data that are key for the understanding of the South African Stone Age. Here, we propose a synthesis of the 1967–1976 excavations led by Hannes Eloff and provide general background information that will serve as a reference for future research. Our synthesis is based on previous publications by Ina Plug, as well as on Eloff ’s field diaries, which were thought to be lost. We complement these observations with data from our own 2014 field season, and pay tribute to the work previously done at the site. Finally, we discuss some aspects of the LSA/MSA contact at the site and [...]

Reference

PORRAZ, Guillaume, et al. Bushman Rock Shelter (Limpopo, South Africa): a perspective from the edge of the highveld. South African Archaeological Bulletin, 2015, vol. 70, no. 202, p. 166-179

Available at: http://archive-ouverte.unige.ch/unige:105881

Disclaimer: layout of this document may differ from the published version.

1 / 1 166 South African Archaeological Bulletin 70 (202): 166–179, 2015

Research Article BUSHMAN ROCK SHELTER (LIMPOPO, SOUTH AFRICA): A PERSPECTIVE FROM THE EDGE OF THE HIGHVELD

GUILLAUME PORRAZ1,2*, AURORE VAL2,3, LAURE DAYET1, PALOMA DE LA PEÑA2,3, KATJA DOUZE2,3, CHRISTOPHER E. MILLER4,5, MAY L. MURUNGI2, CHANTAL TRIBOLO6, VIOLA C. SCHMID7,8 & CHRISTINE SIEVERS9 1CNRS, USR 3336, UMIFRE 25, French Institute of South Africa, Johannesburg, South Africa *Corresponding author. E-mail: [email protected] 2Evolutionary Studies Institute, University of the Witwatersrand, Johannesburg, South Africa 3DST/NRF Centre of Excellence in Palaeosciences 4Institute for Archaeological Sciences, University of Tübingen, 5Senckenberg Centre for Evolution and Paleoenvironment, University of Tübingen, Germany 6CNRS, UMR 5060, IRAMAT-CRP2A, CNRS-Université Bordeaux Montaigne, 7Abteilung für Ältere Urgeschichte und Quartärökologie & Senckenberg Centre for and Paleoenvironment, Universität Tübingen, Germany 8UMR 7041, ArScAn-AnTET, Université Paris Ouest Nanterre La Défense, France 9School of Geography, Archaeology and Environmental Studies, University of the Witwatersrand, Johannesburg, South Africa (Received March 2015. Revised July 2015)

ABSTRACT emergence of our species seems to coincide with the generali- In this paper, we introduce a recently initiated research project sation throughout the African continent of conducted at Bushman Rock Shelter, on the northeastern edge of the (MSA) , based primarily on predetermined blanks Highveld plateau in Limpopo Province, South Africa. Previous and hafted (Goodwin & Van Riet Lowe 1929; Clark 1988, excavations carried out at the site during the 1960s and 1970s exposed McBrearty & Brooks 2000). Following Clark (1988), the MSA a deep and well-stratified sequence of c. 7 metres of archaeological also appears to coincide with the first appearance of regional deposits associated with Later and Middle Stone Age occupations expressions of and culture, suggesting significant (LSA and MSA). Owing to the lack of contextual information, changes in the way these societies were organised. This process Bushman Rock Shelter remains poorly studied despite recording of group differentiation and territoriality would have helped cultural and palaeoenvironmental data that are key for the under- strengthen cultural identities and social networks, represent- standing of the South African Stone Age. Here, we propose a synthesis ing a major evolutionary step in the history of hunter- of the 1967–1976 excavations led by Hannes Eloff and provide general gatherers. In South Africa, the MSA is considered to have background information that will serve as a reference for future started developing between 300 and 250 ka. research. Our synthesis is based on previous publications by Ina Plug, South Africa is a 1 220 000 km² region composed of varied as well as on Eloff’s field diaries, which were thought to be lost. We topography and nine different biomes based upon the complement these observations with data from our own 2014 field dominant forms of plant life, which broadly, but not exclu- season, and pay tribute to the work previously done at the site. Finally, sively, correlate with different climatic conditions. More than we discuss some aspects of the LSA/MSA contact at the site and 440 vegetation types are found within the biomes, distributed comment on the presence of a bifacial lithic component in the upper across altitudes ranging from the shorelines of the Atlantic and MSA layers, which is reminiscent of the later Pietersburg. Indian Oceans to over 3000 m elevation above sea level (Mucina & Rutherford 2006). One of the current research ques- Keywords: Middle Stone Age, Pietersburg technocomplex, tions is the possible influence of these varied parameters on the bifacial technology, Later Stone Age, burial. cultural history of hunter-gatherers. For the past two decades, field-based studies in South INTRODUCTION Africa have intensified dramatically with clear focus on the Still South Africa documents the appearance of early innova- Bay (SB) and (HP). This focus has emphasised tive practices, both technological and symbolic, some of the high degree of innovation of these two technocomplexes, which occur as early as c. 100 ka (e.g. Deacon & Deacon 1999; but has also reinforced two preconceived ideas (Conard et al. McBrearty & Brooks 2000; Wurz 2000; Mitchell 2002; 2012): 1) the pre-SB and post-HP technologies are “conven- Henshilwood et al. 2002, 2004; Texier et al. 2010; Wadley 2013). tional“, and 2) the MSA is “homogeneous” throughout South This archaeological record fuels and challenges our models of Africa. the history of anatomically modern (AMHs): when Recently, questions have arisen about the cultural sequenc- and under which circumstances these societies developed new ing of the MSA, and the definition of its (sub)phases and their behaviours is a matter of intense discussion. succession. Researchers now tend to agree that there is broader The earliest occurrences of AMHs come from East Africa temporal and spatial variability than previously thought and date to between 200 and 150 ka ago (Ingman et al. 2000; (Soriano et al. 2007; Villa et al. 2010; Mackay 2011; de la Peña et al. White et al. 2003; McDougall et al. 2005; Pearson 2008). The 2013; Porraz et al. 2013a,b; Wurz 2013; Henshilwood et al. 2014), South African Archaeological Bulletin 70 (202): 166–179, 2015 167

FIG. 1. Location of Bushman Rock Shelter, Limpopo, South Africa. demonstrating the necessity to build independent cultural between the high-altitude grasslands to the west, called the sequences throughout South Africa. The challenge is to Highveld, and the low altitude plains to the east, called the correctly assess the diversity of the South African record; Lowveld (or Bushveld). The shelter is located on the edge of the one possibility is to characterise and compare long regional Highveld plateau, at an altitude of about 1500 m above mean sequences. sea level (amsl) within a mountainous area with local altitudes With these perspectives in mind, we recently started a new up to c. 2000 m amsl (Badenhorst & Plug 2012). research project at Bushman Rock Shelter (BRS), in the Much of this part of the Great Escarpment is drained by Limpopo Province. Its geographic position – inland and in eastward-flowing tributaries of the Limpopo River, which the northeast of South Africa – together with the nature of its flows into the Indian Ocean at a distance of c. 350 km from the material record, offer a new dataset to compare to other South site. BRS opens towards the south and overlooks the alluvial African Stone Age sites. Neither HP nor SB technologies have plains of the Molapong River, a small tributary of the perennial been recognised at BRS, and with the exception of the LSA Ohrigstad River, which flows northwards (Fig. 1). deposits, the chronocultural sequence of the site remains The shelter is carved from the dolomites of the Malmani unclear. We consider the BRS project an opportunity to build a Subgroup, Transvaal Supergroup, and is made of dark-grey to new and empirical scenario for cultural changes and AMH black well-bedded shales. The geology of the area is character- adaptations in the interior of southern Africa. However, before ised by the presence of many faults orientated south to north, BRS can be included in large-scale comparisons, many ques- and southwest to northeast. The shelter is about 55 m wide, tions need to be clarified. The site has long been part of the 25 m deep and 13 m high. Its geological formation is the result South African research tradition, but has fallen into obscurity in of tectonic events and the progressive collapse of its bedded the past few decades, mainly because of the lack of field data. lithology. The current drip line of the shelter, which corre- Besides Plug’s papers (Plug 1981a,b, 1982; Badenhorst & Plug sponds to the modern cliff line of the hill, and the absence of 2012) and a series of field reports (Eloff 1969; Louw 1969; Mason major rockfall suggest the morphology of the shelter did not 1969; Vogel 1969), little information has been published on the change radically over the . site and the archaeological material it has yielded. The region falls within the summer rainfall zone, character- In this paper, we aim to present general background infor- ised by an annual average of c. 900 mm of rain. It experiences mation about the site together with a review of the excavations warm summers and mild winters. BRS lies on the ecotone of conducted by Hannes Eloff during the 1960s and 1970s and of the grassland and savanna biomes (Mucina & Rutherford 2006) the associated archaeological record. This paper is based on the and the local vegetation comprises bushveld (Ohrigstad original Eloff field diaries, the papers written by Ina Plug (1981, Mountain Bushveld) and wooded grassland. The degree of 1982; Badenhorst & Plug, 2012), as well as on the field data we grass cover and density of wooded vegetation varies between collected in 2014. In the discussion, we clarify the nature of the steep mountain slopes and flatter alluvial plains, such as the LSA/MSA contact at the site and briefly introduce the along the Molapong River, which in turn contributes to the techno-typological characteristics of the lithic assemblages variety of plant communities in the local vegetation mosaic. from the upper MSA layers, which include bifacial pieces and Amongst the taxa present are the fruits marula, Sclerocarya are reminiscent of the later Pietersburg (Mason 1957, 1962; birrea subsp. caffra, stamvrug, Englerophytum magalismontanum Sampson 1974). and the jacket-plum, Pappea capensis (also noted for the superior oil of its seeds); excellent woods for firewood and tools such as BACKGROUND TO BUSHMAN ROCK SHELTER the wild olive, Olea europea subsp. capensis and several Acacia BRS is located in the Ohrigstad district, in the southeastern and Grewia species; and grasses of differing palatability. The corner of the Limpopo Province, close to the border with local fauna is diverse and comprises a wide range of ungulates Mpumalanga Province (Fig. 1). The site is located on the edge of (bovids such as blue , eland, bushbuck, roan the Great Escarpment in the Drakensberg mountain chain. The antelope, and impala; and suids, namely warthogs and bush- Great Escarpment represents a natural ecological boundary pigs), carnivores (small and medium taxa), primates (chacma 168 South African Archaeological Bulletin 70 (202): 166–179, 2015

FIG. 2. Planimetry of previous excavations at Bushman Rock Shelter (CAD by G. Porraz, from Plug 1978, and from Eloff's personal documents). baboons and vervet monkeys), rodents (including large species hole previously dug by the farmer (Fig. 2). Louw reached a depth such as scrub hares, porcupines, African mole-rats, and greater of about 2.4 m (8 ft) and stopped the excavation in his layer 43. cane rats), birds, reptiles and invertebrates (especially the giant The excavation followed a stratigraphy composed of artificial African land snail, Achatina immaculata) (Branch 1998; Skinner layers of approximately 7 cm (3 inches) thick (Plug 1981a). & Chimimba 2005; Sinclair et al. 2011; Herbert & Kilburn 2004). J.F. Eloff, leading a team from the University of Pretoria, Until recent times, the region was populated mainly by started excavating the site in July 1967, in response to a request Shona-speaking people. Iron Age and historical periods are by B.D. Malan from the Monuments Council. Eloff started by documented at BRS by a few poorly preserved monochrome opening a new square (C8) of c. 1.5 m² (5 ft × 5 ft) adjacent to paintings (red or black). More striking at BRS is the presence of Louw’s excavation and stopped at about the same depth as a few thousand incised on the dolomite walls of the Louw (“about 10 inches below”, Eloff 1969: 60). He then exca- shelter, likely produced during the sharpening of metal blades vated squares A6, B6 and C6, stepping in from the sides for or during Iron Age times (Plug 1981a; Louw 1969). safety reasons. Once the excavation area had reached the same Regarding the Stone Age record, the region remains poorly depth everywhere, he shifted to a metric system (Fig. 2) to explored though Revil Mason predicted “a rich reward from ultimately attain a final depth of 7 to 8 m, ending with his level [the known in the Transvaal Dolomite in the eastern 105. It is unclear whether or not bedrock had been reached escarpment]” (Mason 1969: 59). Apart from BRS, the only other (Plug 1981a). site documented and studied nearby is Heuningneskrans Louw and Eloff used the same numbering system but fol- (HGS), located about 2 km east of BRS, on the east side of the lowed different excavation strategies. Thus, their sequencing Ohrigstad River. Excavations undertaken by Peter Beaumont of the archaeological deposits do not correlate well (Plug 1978). in 1968 (Beaumont 1981) exposed a sequence greater than 6 m To date, the best record to evaluate the scientific potential deep with Iron Age, LSA and MSA deposits. At a regional scale, represented by BRS comes from Eloff’s excavation. With the second exception is of , located about 170 km invaluable help from the staff of the Department of Archaeol- northwest of BRS (Mason et al. 1988; McNabb & Sinclair 2009). ogy at the University of Pretoria, we were able to locate the The archaeological potential of BRS seems to have aroused Eloff’s field diaries, which were thought to be lost (Badenhorst interest for the first time in the 1960s, after J.J. Malan collected & Plug 2012). These notebooks offer the first opportunity to artefacts at the site. At that time, the shelter was used as a glean some contextual information on the old collections. tobacco barn and some deposits were removed by the farmer, presumably to obtain stones to build a wall along the outer ELOFF’S EXCAVATIONS: SUMMARY AND TRIBUTE perimeter of the shelter (Plug 1981a). Some deposits located at The observations presented here rely on six manuscripts the front of the shelter were also used to build the road to Echo that were compiled by Eloff during his fieldwork, almost on a Caves, situated about 2 km away from the site. A museum, the daily basis, starting from 17 July 1967. These notebooks, trans- ‘Museum of Man’, was erected inside the site a few years later. lated from Afrikaans to English (by Gerhard Jordaan, an A.W. Louw from the University of the Witwatersrand was archaeology student at the University of Pretoria), record ten the first to investigate BRS (Louw 1969). From 20 June to 4 July field seasons, running from 1967 to 1976. With the exception 1965, he excavated a trial trench at the back of the shelter. A of 1967, the excavations took place once a year, in November– datum line, still visible today on the northeastern wall of the December, for one to two weeks’ duration. The notes end in shelter, was established to set up the excavation. The trench 1976 with layer 59, but the fieldwork continued until layer 105 was about 4.5 m (15 ft) long and 1.5 m (5 ft) wide (squares A/B7, was reached. Together with the field notes, there are a few C7, D7 and E7; see Fig. 2). An additional square (A/B8) was photographs and sketches, as well as one unpublished opened to rectify the disturbance caused by the presence of a stratigraphic profile of layers 1 to 72 (Fig. 3). From layer 59 to South African Archaeological Bulletin 70 (202): 166–179, 2015 169 iptions on the right are taken from Eloff's field diaries. The descriptions of . Unpublished stratigraphic drawing of the West section from Bushman Rock Shelter (from Eloff's personal documents). Short sedimentological descr layers 1 to 28(a) are those from square C6, the descriptions of layers 28(b) to 58 are those from square K1. FIG. 3 170 South African Archaeological Bulletin 70 (202): 166–179, 2015 layer 105, we have at our disposal only a few sketches, some- sieving, sorting and recording (Ina Plug, pers. comm. 2014; times accompanied by comments. Fig. 4). The sediments were dry sieved using 12 and 3 mm Eloff’s excavations can be subdivided into three distinct screen meshes (Plug 1978), and special attention was paid phases: to organic pockets when they were found: “[layer 25] The • Phase 1 (1967 to 1972). This first phase of the excavations fo- material from this pocket [25(b)(i)] is sifted and packed cused on the upper deposits, namely layers 1 to 28. Various separately. A lot of charcoal and burnt seeds were found. […] squares were excavated at different times, with initial atten- The seeds are brittle and are therefore handled with care and tion paid to square C8 in order to secure correlations with treated.” (1972: book 4). Louw’s excavation. • Excavation record. The installation of metal rods together • Phase 2 (1972 to 1975). This second phase corresponds to with the use of sketches, photographs and daily notes when Eloff switched to a metric system, “to determine distri- highlight the importance Eloff attached to the context of the bution of artefacts” (1972: book 3). During these four years, finds, as well as his search for methodological consis- layers 28(b) to 36(ii) were excavated consecutively in squares tency over the years. Personal interpretations, questions and J-K1 (1972–1973) and squares J-K2 and J3 (1974–1975). observations made a posteriori were also reported: “[layer • Phase 3 (1976 onwards). This last phase of excavation is consis- 17] It seems that what we have here is a transition phase tent with the exploration of the lower deposits in Test Pit 1 between Middle Stone Age and Later Stone Age. An un-sifted (called T/1). “Even though it is named T/1 (test pit) it is still a soil sample (P42) as well as sifted charcoal samples (B30) were controlled excavation based on stratigraphical layers. […] taken. […] It is difficult to see where one layer ends and The reasons for digging the test pit are: (a) to determine the the next one starts. […] during the previous excavations we depth of the unit and (b) to understand the material culture did not realise that the lower level should be divided.” (1969: coming out and (c) to identify the stratigraphy of the unit and book 3). the different layers that are present, also to create a known • Stratigraphic observations. In his notebooks, Eloff provided point from which to work from” (1976: book 5). Excavation of detailed descriptions of the colour and contents of each of the T/1 focused on 2 m², astride squares J1-K1 and J2-K2. layers (cave roof debris, ash, charcoal, lithics, etc.). He also The excavation followed the natural stratigraphy (i.e. paid attention to their thickness and lateral variation and changes in sediment type, colour and content). The layers were noted the presence of post-depositional effects (i.e. water numbered from 1 to 105 and when specific features were recog- circulation, animal activities, topographic depression and nised, some of them were subdivided using letters. The layers, secondary crystal formation). He often discussed strati- their subdivisions and lateral variations, are carefully described graphic correlations. Small test pits were excavated when the in the field books. Stratigraphic descriptions provided by Eloff stratigraphy was unclear. are summarised in Fig. 3. • Archaeological descriptions. Eloff systematically described the In many ways, the excavation methods comply with ‘mod- nature of the archaeological remains (i.e. lithics, ochre, fauna, ern standards’: seeds and charcoal). He estimated relative densities of finds • Excavation techniques. The field notes document the use of and proposed preliminary spatial analysis and associated trowels and shovels during the excavations. Archaeological behavioural interpretations (e.g. occurrence of hearths). The finds were systematically recorded according to their square description of the child burial (cf. infra) constitutes one of the and layer of provenance. The excavation team was divided best examples of the high level of precision in Eloff’s field into four groups, each composed of four to five people who notes. In terms of sampling, Eloff adopted a rigorous proto- were assigned to one of the following tasks: excavating, col, whereby each loose sample (mostly for sedimentological

FIG. 4. Bushman Rock Shelter during excavation (courtesy I. Plug, centre left). South African Archaeological Bulletin 70 (202): 166–179, 2015 171

FIG. 5. Two photographs of the southwestern corner taken four decades apart (left: J.F. Eloff; right: G. Porraz). The correspondence between the photographs is striking. The white circle indicates a rounded rock which is likely a grindstone and which is clearly described in Eloff's diaries (1972): “At the D6/7 wall, a round stone is sticking out. This rock is drawn on the profile drawing and numbered 19(ii). The rock is not removed”.

analysis and radiocarbon dating) was described and kept and appear to be thicker and include laterally discontinuous separately. lenses. These two schematic ‘blocks’ are separated by a rocky Several questions still need to be addressed thereby justify- unit, which corresponds to layers 18 and 19 (Fig. 6). Based on ing the reopening of the site for renewed excavations. Such Butzer’s interpretation (1984), this rocky unit that marks the questions relate to the nature of the small components of the contact between the LSA and the MSA coincides with the Last fauna and botanical assemblage (<3 mm) and to the validity of Glacial Maximum (Badenhorst & Plug 2012). the gradual transformations that can be observed between Throughout the sequence, differences in artefact densities some specific layers. The latter question is especially relevant were noticed by Plug as well as Eloff during his excavation: for the MSA deposits, which have been more affected by “[layer 30(ii)] A lot less cultural material was found in this layer diagenetic processes than the LSA ones, and have received compared to 30(i) and 29. Not a lot of charcoal found. […]The field descriptions less detailed than the recent units. layer contains some bone and a few flakes […]. Marula plugs During a three-week field season that took place in April were also found. Within the ash, small thin red brown 2014, the site was cleaned and the deposits were secured with soil pockets were found, probably burnt pockets” (1973: scaffolding. This first season allowed us to refine our under- book 4). standing of the stratigraphy and of the previous work at the One striking element in this description is the nature and site. Based on the significant number of stratigraphic tags still quality of organic preservation at the site. The deposits have present on the profiles (Fig. 5), we were able to establish direct yielded abundant vertebrate and invertebrate remains, correspondence between our field observations and the charcoal and seeds, from the surface to the base of the descriptions provided by Eloff. These preliminary observa- sequence. According to the notes, it seems that most of the tions confirm the high stratigraphic resolution of the LSA and botanical remains were collected from ‘burnt pockets’. They MSA deposits (Fig. 6), as well as the quality of the excavations were sifted separately during the excavations. The preliminary led by Eloff. study conducted by Wadley (1986, 1987) on the macrobotanical remains from BRS provides valuable palaeoenvironmental THE ARCHAEOLOGICAL SEQUENCE OF BRS indications. Hence, the alternating presence and absence of fruits of the marula, a Bushveld tree that does not tolerate frost, GENERAL OBSERVATIONS and of Hypoxis, an underground edible corm, suggest the No sterile layers have been recognised at BRS, but radiocar- occurrence of major ecological shifts between layer 1 and 28. bon dates (Plug 1978) suggest the occurrence of hiatuses in All the faunal remains recovered during Eloff’s excava- sedimentation. The stratigraphic sequence can be schemati- tions were studied by Plug (1978, 1981b; see also Brain 1969). cally divided into two main parts. The upper part, associated Taxonomic descriptions of the LSA fauna are provided in Plug with the LSA layers, corresponds to fine layers composed of (1981a), while a recent paper (Badenhorst & Plug 2012) ash, charcoal and sand. The lower part, consistent with the presents a synthesis on the MSA fauna from the site, based MSA units, is well-stratified and includes both anthropogenic exclusively on data from Plug’s Masters dissertation (Plug and natural inputs, but the layers contain more cave roof debris 1978). The use of 12 and 3 mm sieves allowed for fairly good 172 South African Archaeological Bulletin 70 (202): 166–179, 2015

FIG. 6. Southern profile of Bushman Rock Shelter (drawn by G. Porraz, K. Douze, P.de la Peña, V.C.Schmid and A. Val;CAD by G. Porraz and V.C.Schmid). recovery of faunal remains, including elements such as yielded 66 358 remains (Badenhorst & Plug 2012). The diversity microfaunal bones and teeth, which have not yet benefited of the faunal spectrum, both for the LSA and the MSA (39 taxa from detailed taxonomic identifications. Bone and tooth identified in the LSA; 50 in the MSA assemblage) (Plug 1978, fragments are abundant in all archaeological units. Remains 1981b; Badenhorst & Plug 2012), suggests the existence of a recovered from LSA units are relatively well-preserved while wide range of acquisition techniques, including active hunting, those found in MSA layers present a higher level of fragmenta- snaring, trapping and gathering. The fauna is numerically tion and a poorer degree of cortical preservation. The dominated by giant land snails (Achatina sp.) and tortoises (no anthropogenic origin of the faunal assemblage is indisputable taxonomic identification available), followed by ungulates and for the LSA, as illustrated by the abundance of cut-marked and monitor lizards (Varanus sp.). The consumption of tortoises, burnt bones (Plug 1978, 1981b). Given the depth of the deposits and especially angulate tortoises (Chersina angulata), is for the MSA sequence, one cannot exclude the possibility that common at several other MSA/LSA sites from the Western some bone-accumulating animals such as leopards, jackals and Cape, including , Klipdrift Shelter, Pinnacle porcupines used the shelter during non-anthropogenic occu- Point Cave 13B and (Thompson 2010; pation phases. However, carnivore and rodent modifications Steele & Klein 2013; Henshilwood et al. 2014; Thompson & of the bone remains are extremely scant since only 13 speci- Henshilwood 2014). Interestingly, the significant place occu- mens exhibit gnawing marks, and carnivore contribution to the pied by giant land snails and lizards in the faunal spectrum at faunal assemblage seems limited (Badenhorst & Plug 2012). BRS seems to be a specific to the site. Ungulates are very Besides, burnt bones are also abundant in the MSA units and common in the diet of MSA and LSA hunter-gatherers, across cut marks have been noted. Consequently, Badenhorst and South Africa (see for instance: Klein 1978; Clark & Plug 2008; Plug (2012) argue in favour of a predominantly human origin Thompson 2010; Steele & Klein 2013); species represented at for the faunal remains in the MSA units. BRS include mainly large bovids, equids and suids. Trapping at The LSA faunal assemblage (layers 2 to 18) comprises 41 659 BRS is suggested by the presence, mostly in the LSA assem- faunal remains (Plug 1978), while the MSA units (19 to 88) have blage, of small bovids, namely duikers (Sylvicapra grimmia), South African Archaeological Bulletin 70 (202): 166–179, 2015 173

FIG. 7. Unpublished photography of the child burial “Skankwan” from Bushman Rock Shelter (from Eloff's personal documents). The descriptions on the right are those taken from Eloff's field diary. together with small mammals such as leporids and other large bearing grinding facets. New examination reveals the use of rodents (e.g. Thryonomys swinderianus, the giant cane rat); a various types of iron-bearing rocks – massive hematite, lateritic similar hypothesis was proposed for the HP at crusts, shale, mudstone, siltstone, etc. For the production of (Clark & Plug 2008; Wadley 2010). lithic tools, several rock types were selected but four are espe- Based on the composition of the faunal spectrum, there cially abundant: quartz, quartzite, hornfels and dolerite. Con- does not seem to be any significant difference between the LSA sidering that various degrees of metamorphism have affected and the MSA in the taxa exploited and therefore in terms of hornfels, it is sometimes difficult to distinguish this rock type – subsistence strategies. Detailed taphonomic and zoo- at a macroscopic scale – from fine dolerite. Information from archaeological studies of the fauna from the LSA and MSA the geological map together with foot surveys conducted units will enable us to further test this preliminary hypothesis during the 2014 field season support the hypothesis that all the and provide new data regarding butchery practices at the site, raw materials were locally available. sensu Binford (1981). A good synthesis of the bone tools and beads is available in One of the most spectacular discoveries at BRS is a child Plug (1981a, 1982). Most of them come from LSA layers, but a burial in the upper part of the LSA sequence (Fig. 7). An excel- few were also recovered from MSA units. The occurrence of lent field description is provided by Eloff, who was extremely bone tools in MSA deposits was already noted by Louw (1969), meticulous in recording the context of the burial: “The origins who identified two pointed bone implements in his layers 34 of the grave were probably discovered for the first time in Layer and 41. In his field diaries, Eloff also reported possible bone 6 when a soft pocket was identified. Because of the softness of tools. MSA beads were found in Louw’s layer 35 (layer 19/20 of the soil it is possible that the grave originated even higher up Eloff) as well as in Eloff’s layers 21–22 and 28 (Plug 1982). Our (in Layer 5(a) or in Layer 5). It seems highly unlikely that it orig- recent overview of the archaeological collections revealed the inated any higher for instance in Layer 4 […]” (1971, book 3). presence of six MSA beads recovered from layers 27(i) and Layer 5 was dated to 9 500 BP.The unearthing of the child burial 28(a). Five specimens are made of ostrich eggshell and one of should not be confused with the discovery in 1969 of an infant Achatina sp. shell. Neither field notes nor direct dating are avail- mandible aged between six to eight months (Protsch & De able for these. Beaumont and Bednarick (2013) note the pres- Villiers 1974). This mandible, collected by a tour guide while it ence of engraved dolomite fragments, sometimes densely was apparently protruding from the excavation profile, was marked, collected from layer 36. said to have been found somewhere between layers 14 and 18, On several occasions while he was excavating, Eloff but doubts have been expressed by researchers concerning its inferred the existence of features such as hearths. He also provenance (Protsch & De Villiers 1974). Problems regarding proposed some preliminary interpretations concerning activi- the exact origin of the mandible in the site are aggravated by ties and spatial organisation: “[layer 4] In Square A there is an the discrepancy between the two sets of radiocarbon dates abundance of grass, burnt seeds and some bone and shell available (those from UCLA and those from Gröningen, pieces. It would seem that Square C consists of similar material, Protsch & De Villiers 1974). Another two human remains in but with less grass and fewer seeds than found in Square A. the form of cuboid fragments were collected from MSA units Therefore it seems that for centuries fire was made in Square B, (layers 28(a) and 29(1)) (Badenhorst & Plug 2012). and food was eaten around the fire (the seeds that were found MSA and LSA mineral artefacts recovered at BRS include are evidence for this) and the grass was used to sleep on.” (1968: ochre pieces and stone tools. The MSA ochre component has book 1). The type of sediments composing the upper LSA been previously described by Watts (2002) who identified a layers suggests that this part of the sequence is consistent with majority of red shale and hematite and the presence of pieces a dumping area (“an ash ” according to Eloff), while the 174 South African Archaeological Bulletin 70 (202): 166–179, 2015 lower layers seem to be associated with knapping and/or evidence of the Younger (layers 9–13) and Older (layers 14–18) butchering events. The specific location of the excavated area, Dryas Events (Fig. 8). The LSA deposits of BRS cover the termi- at the back of the shelter, should be noted because it has bear- nal Pleistocene to early succession and are therefore ing on our understanding of the activities recorded at the site. contemporaneous with some other well-known sequences such as those from Cave (Parkington 1988), THE CHRONOCULTURAL SEQUENCE Klipdrift Shelter (Ryano 2014), Boomplaas (Fairhall et al. 1976), To date, detailed cultural analyses have been undertaken (Fairhall et al. 1976), Rose Cottage Cave on material recovered only from layers 1 to 18 (Plug 1981a). A (Wadley & Vogel 1991) and Heuningneskrans (Beaumont preliminary analysis of the MSA lithic sequence was published 1981). by T. Volman in his synthesis on the southern African MSA Plug subdivided the LSA into two main phases, which she (Volman 1981, 1984). More recently, D. Underhill (2012) pub- considered to represent various developmental stages within lished a comparative study of the lithics from layers 95 to 105 the same cultural context (Plug 1981a). The upper (layers 2–5) at BRS (together with the lithics from layers 26–28 of Louw’s and lower (6–14) phases differ in their relative number of excavation) and of those from Bed IV at Cave of Hearths. quartz pieces and bone tools. Thus, quartz pieces are found in Layers 1 and 2 at BRS yielded some fragments greater numbers in the upper phase than in the lower one, and showing Lydenburg characteristics (Plug 1982). These frag- are more frequently retouched. Quartz backed pieces seem to ments present deeply incised diagonal, horizontal and vertical be restricted to the upper phase. Bone tools are more abundant lines (Louw 1969). The recovered from these layers is in the lower phase than in the upper one. According to Plug Iron Age. However, some intrusive stone artefacts from earlier (1982: 57), “some of the bone tools bear great resemblance to excavations and material of recent origin (including articles of stone artefacts such as scrapers, burins, drills and scaled pieces” European manufacture, according to Louw 1969) also occur in and might have compensated for the small number of formal these layers. stone tools (Plug 1981b). Polished points, linkshafts and orna- At present, dating of the BRS sequence is exclusively based ments are also present in the sample. Different researchers on radiocarbon and is reliable only for the LSA deposits (Vogel have discussed and labelled the cultural definition of the 1969; Plug 1981a; Protsch & De Villiers 1974). Based on the dates LSA phases differently. Both Eloff (1969) and Mason (1969) obtained from Eloff’s excavations, the age of the LSA occupa- interpreted the assemblages as part of the Transvaal Smith- tions can be bracketed between 9 500 and 12 500 BP (Plug field; Sampson (1974) places the LSA of BRS as part of the Oak- 1981a). The LSA chronology is supported by isotopic analysis hurst. Plug (1981a: 20–21) pointed out noticeable typological conducted on shell fragments of the giant land snails Achatina variations and concluded: “BRS has very few of the typological sp. collected in abundance at the site (Abell & Plug 2000). Smithfield A (Oakhurst complex), B or C tools in the assem- Changes in oxygen isotope ratios have been interpreted as blage […] The industry belongs to the /early Holocene and may be referred to for the present as the Bushman rock phase of the Transvaal Later Stone Age”. More recently, the LSA at BRS has been attributed to the Oakhurst technocomplex by Lombard et al. (2012). The identification of a microlithic bladelet industry in layers 14–18 led Mitchell to propose the presence of a Robberg Industry at the base of the LSA sequence (Mitchell 1988). We support this hypothesis, but also underline the concomitant occurrence of some components that are reminiscent of a MSA technology (cf. infra). There are two radiocarbon dates available for the MSA deposits (Louw’s excavation), both beyond the range of the technique (Vogel1969). The results indicate an age greater than 53 000 BP for layer 38 (i.e. Eloff’s layer 21, in Plug 1978) and an age greater than 47 000 BP for layer 41. Besides these dates, the sole chronological framework we have at our disposal relies on botanical remains (Wadley 1986, 1987) and sedimentological interpretations (Butzer & Vogel 1979). In her study of the distribution of marula fruit remains throughout the archaeo- logical sequence, Wadley(1986, 1987) found that they are abun- dant during the early Holocene (layers 1–8) and the upper part of the MSA (layers 25–28). Their presence reflects warm clima- tic conditions that, for layers 25–28, could correspond either to a period within MIS3 or to the last interglacial episode. A similar basis of interpretation is proposed by Badenhorst and Plug (2012) who hypothesise that MSA layers with marula correspond to MIS3 while the following layers (31–37) could correspond to MIS4. Butzer and Vogel’s (1979) interpretations suggest that the leached surface of layer 31 might be consistent with MIS5e, while the lower layers might be consistent with MIS6. FIG. 8. Changes in 18O of land snail shells (presumably Achatina immacu- The MSA at BRS has been discussed by lata) from the upper deposits of Bushman Rock Shelter, together with 14C several authors and compared with industries from Cave of dating (after Abell & Plug 2000). Hearths. Mason (1969), using Louw’s collection (layers 28–43), South African Archaeological Bulletin 70 (202): 166–179, 2015 175 assigned the upper MSA industry from BRS to a developed indicating the existence of a hiatus of at least 30 000 years MSA technology. He noted that the lithics collected by Eloff in between the top of layer 18 and layer 21. the layers directly underlying those excavated by Louw shared At this stage of the analysis, we favour the hypothesis of an a style similar to those from Cave of Hearths Bed IV. This is artificial and fortuitous technological association between the stated by Eloff himself (1969: 60), who described an industry MSA-like and Robberg-like components. The rocky nature of below 1.5 m deep that “seems to represent a south-eastern the layers and the presence of large rocks in square C6 (Fig. 6) expression of the Pietersburg industrial complex. […] it is possi- could have led to such an association. However, as stated by ble to discern a distinct typological resemblance between the Plug (1982), although they could explain why some small Cave of Bed 4 Earlier Pietersburg artefacts and the remains could have migrated down the sequence, they can material from the lowest part of BRS.” hardly account for larger elements going up. Two main The main chrono-cultural subdivision for BRS is the one hypotheses are currently considered: 1) the LSA inhabitants of proposed by T. Volman (1984). Referring to his own typology, the shelter dug pits into the MSA layers; or 2) MSA implements he assigned layers 15–18 to the ‘MSA 3’, levels 19–30 to the were scattered on the floor of the shelter when the people who ‘MSA 2’ and levels 31–105 to the ‘MSA 1’. According to this produced the Robberg Industry occupied the site, implying classification, BRS would then represent one of the few that the rocky unit is associated with a phase of erosion or South African sites with an early MSA lithic technology, non-deposition. together with and Peers Cave. Singer and Wymer (1982: 204) also attempted a comparison, suggesting TECHNOLOGICAL OVERVIEW OF THE MSA FROM that BRS “might equate with KRM I or II”. More recently, LAYERS 21–24 Badenhorst and Plug (2012) suggested BRS possibly records Layer 21 at BRS is the first layer below the rocky horizon MSA1 occupations, but also MSA 2 and MSA3, with deposits that follows clearly across the profile (Fig. 6). The lithics recov- that are contemporaneous with the pre-SB, the SB, the HP and ered from this upper part of the MSA sequence were classified the post-HP. as developed MSA technology by Mason (1969) and MSA 2 by Volman (1984), although neither technological nor typological THE LSA/MSA CONTACT studies have yet been published. Badenhorst and Plug (2012) “This period is going to be difficult. We are proposing to suggested these layers may be contemporaneous with what is excavate the area where the Middle Stone Age moves over to informally referred as post-HP,but no direct data support this the Later Stone Age, in other words the transition period. (…) hypothesis. Here, we provide a short description of Eloff’s This part stretches from the layer above the rock level to under- lithic assemblage collected from layers 21 to 24 in squares A6, B6 neath the level with the large rocks (but still contains some and C6, B8 and C8. As there are some uncertainties regarding small rocks). These are Layers 17–23. We are going to excavate the integrity of the collection, we do not provide any detailed these layers with great care and make an even smaller division and quantified analysis but, rather, highlight a few technologi- of each layer than what was done in the past.” (Eloff 1972: cal and typological aspects. book 3). Altogether, the lithic assemblage from these layers com- The exact nature of the sedimentary contact between the prises 4369 artefacts greater than 20 mm in maximum dimen- LSA and the MSA units, as well as the formation processes that sion. The richest horizon is layer 21 with a total of 1848 lithic have contributed to shape it are still poorly understood. This artefacts while the poorest corresponds to layer 24 with a total LSA/MSA contact is very coarse-grained on the eastern profile, of 740 pieces. Two main raw materials comprise the lithic where Eloff excavated, but appears to be much better stratified assemblage, namely a hornfels category (including fine on the southern profile, and notably in the western corner dolerite) and quartz. Both rock types are represented in similar where we observe a decreasing presence of rocks from east to proportions within the layers. The other rock types, quartzite west (Fig. 6). and chert especially, represent about 2% of the lithic assem- Eloff suggested the possible existence of a transitional blage. phase between the MSA and the LSA: “there are some indica- Reduction sequences were oriented towards the produc- tions of a transitional phase between the MSA and LSA hori- tion of flakes and blades of various morphologies, though zons, but no definite conclusions can be made at this stage” rarely triangular. Our first classification of the cores indicates (Eloff 1969: 60). Plug initially considered layers 14 to 18 to be the presence of a discoid, a Levallois (centripetal and unidirec- MSA (Plug 1978, 1981a), as did Volman (1984), but she later tional) and a prismatic reduction sequence. Blanks often questioned the homogeneity of these layers: “Levels 15 to 18 at present faceted platforms and some percussion marks illus- BRS represent a period about which some confusion exists.” trate the use of a soft hammer. Quartz and hornfels appear to (Plug 1982: 61). She questioned the integrity of these layers, but have been exploited in a similar way. rejected the hypothesis of mixing: “This is, unlikely however,in One of the main differences between hornfels and quartz that the deposit appears to be undisturbed and level 18, a thick concerns the frequency of retouched pieces. Between 10 and level of tightly packed rock spalls and ash, is probably impene- 20% of hornfels blanks have been retouched whereas this pro- trable and would have prevented contact between the levels portion is always less than 5% for the quartz. Different mecha- above and below it.” (Plug 1982: 62; but see Badenhorst & Plug nisms may explain this difference with properties of the rock 2012). types being a likely possibility (see Wadley & Kempson 2011). From our preliminary techno-typological observations, it The main typological components of the formal appears that all the material including and below layer 19 can sample (n = 363) are convergent and lateral scrapers. Of partic- be classified as MSA. Layers 15 to 18 include MSA-like compo- ular interest is the way they have been shaped. About half of nents (i.e. blades and toolkits) and Robberg-like (micro) the formal tools have been transformed on their ventral face bladelet implements. The existing series of radiocarbon dates (Fig. 9): 11% have been retouched only on their ventral face, for layers 15 to 18 (n = 3) are coherent and indicate an age of 17% have been bifacially retouched and 14% bifacially shaped. between 12 000 and 13 000 BP (Plug, 1978). Radiocarbon dating We present the difference between the bifacially retouched from the lower layers gives an age greater than 40 000 BP, pieces and the bifacially shaped pieces as a preliminary insight 176 South African Archaeological Bulletin 70 (202): 166–179, 2015

FIG. 9. MSA lithic tools from layers 21–24 of Bushman Rock Shelter (all hornfels). From top to base, left to right: no. 630, convergent tool with thinned base; no. 1241: convergent tool bifacially retouched at its tip; nos. 1428, 1441 and 84: convergent tools bifacially retouched; nos. 1238, 1237 and 70: bifacial pieces; no. 1411: Kostienki tool; nos. 39 and 48: shaping-like flakes laterally retouched (drawings by M. Grenet). and not an argument in favour of any specific functions or de- assemblage. There is no specific selection in terms of blank signs. The difference is based on the degree of invasion of the morphologies and dimensions; a few of the blanks might even removals: bifacially shaped pieces, unlike bifacially retouched originate from a bifacial reduction sequence (Fig. 9). Finally, it pieces, have at least one surface entirely shaped. In the layers is worth mentioning the presence of so-called “Kostienki” considered here (21–24), the bifacial reduction strategy repre- pieces (see Klaric et al. 2015), which are blanks that have been sents a minor component, as reflected by the low proportion of retouched, truncated and exploited on their dorsal surface. shaping flakes, which compose between 2 and 5% of the lithic Technologically and typologically, the MSA from layers 21 South African Archaeological Bulletin 70 (202): 166–179, 2015 177 to 24 is reminiscent of the Pietersburg industry and especially Echo Caves, who facilitate greatly the practical aspects of the its later phase that is characterised by an “increase in the use of fieldwork. Many thanks to Ceri Ashley, Ndukuyakhe Ndlovu, bifacial technique” (Mason 1957: 135; 1962; Sampson 1974). and Alexander Antonites from the University of Pretoria for Bifacial pieces in Pietersburg contexts have been notably their assistance during the location of the archaeological mate- described in the upper beds (6–9) of Cave of Hearths (Sampson rial excavated by Eloff, as well as to Cameron Penn-Clarke from 1974) as well as in the pre-HP layers of which was the School of Geosciences at the University of the Witwaters- dated using ESR between 82 ka and 200 ka (Beaumont et al. rand for his help. We would like to thank the Evolutionary 1978; Grün & Beaumont 2001; Wadley 2007). The technology of Studies Institute at the University of the Witwatersrand, for the Pietersburg and the definition of its three main phases need housing the collections and for allowing us to use some equip- to be clarified, but some lithic attributes such as prepared cores, ment during our fieldwork. This research is supported by the blades and flakes, faceted platforms (for the latter phase, see Ministère des Affaires Étrangères in France and the Fyssen Mason 1957) and elongated unifacial points (Clark 1982; Foundation; the support of the DST/NRF Centre of Excellence McBrearty & Brooks 2002) – sometimes ventrally retouched – in Palaeosciences (CoE in Palaeosciences) towards this research typify and individualise the Pietersburg within the current is hereby acknowledged. Two reviewers gave valuable com- understanding of technology within the southern African ments and helped improve the manuscript. Opinions ex- MSA. pressed and conclusions arrived at, are those of the authors and are not necessarily to be attributed to the CoE in Palaeo- PERSPECTIVES sciences. Any information about the excavation history at BRS The aim of this paper was to briefly introduce our new field and its archaeological collections are most welcome and can be project at BRS and to clarify previous work conducted at the sent to the corresponding author. site. Wepay tribute to the professionalism of Eloff, the quality of his fieldwork and his legacy (Meyer 2007). Eloff’s excavation REFERENCES methods conform to modern standards and guarantee, four Abell, P.I.& Plug, I. 2000. The Pleistocene/Holocene transition in South Africa: evidence for the younger Dryas event. Global and Planetary decades later, the scientific value of the old archaeological Change 26: 173–179. collections. Conversely, we are concerned about the subse- Badenhorst, S. & Plug, I. 2012. The faunal remains from the Middle quent curation of the collections. Stone Age levels of Bushman Rock Shelter in South Africa. South Afri- The diversity and richness of the archaeological record, the can Archaeological Bulletin 67: 16–31. quality of the organic preservation at the site together with the Beaumont, P.B.1981. The Heuningneskrans shelter. In: Voigt,E.A. (ed.) resolution of the sedimentary events provide an excellent Guide to Archaeological Sites in the Northern and Eastern Transvaal: 133–145. Southern African Association of Archaeologists, Transvaal opportunity to decipher cultural changes and human adapta- Museum, Pretoria tions over a long period. Chronology, palaeoenvironmental Beaumont, P.D. & Bednarik, R.G. 2013. Tracing the emergence of change, subsistence and technology are the four main that paleoart in sub-Saharan Africa. Research 30(1): 1–22. structure the project. The nature of the Terminal Pleistocene to Binford, L.W. 1981. Bones: Ancient Men and Modern Myths. New York: Holocene succession, the LSA/MSA contact and the characteri- Academic Press. Brain, C.K. 1969. Faunal remains from the Bushman Rock Shelter, East- sation of the MSA layer 28 are the short-term goals that will ern Transvaal. South African Archaeological Bulletin 24: 52–55. direct our fieldwork. Branch, W.R. 1998. Field Guide to Snakes and Other Reptiles of Southern Long-term goals are to describe the 5 to 6 m deep deposits Africa. : Struik Publishers. with MSA occupations, to document the technological trans- Butzer, K.W. 1984. Archaeology and Quaternary environment in the formations and to establish comparisons with other regional interior of southern Africa. In: Klein, R.G. (ed.) Southern African Pre- records. The presence of bifacial pieces in the upper MSA layers history and Paleoenvironments: 1–64. Rotterdam: A.A. Balkema. Butzer, K.W. & Vogel, J.C. 1979. Archaeo-sedimentological sequences of BRS, recorded from layers 20 to 26, fuels a discussion about from the sub-montane interior of South Africa: Rose Cottage Cave, the bifacial technological phenomenon in South Africa, its Heuningneskrans, and Bushman Rock Shelter. Unpublished paper extension, definition and significance. The increasing number presented at a Southern African Association of Archaeologists work- of lithic studies recording bifacial pieces (e.g. Wadley 2007, shop on: ‘Towards a better understanding of the Upper Pleistocene 2012; Lombard et al. 2010; MacKay et al. 2010; Vogelsang et al. in sub-Saharan Africa’. Stellenbosch, 27–29 June 1979. Clark, J.D. 1988. The Middle Stone Age of East Africa and the 2010; De la Peña et al. 2013; Porraz et al. 2013a; De la Peña & beginnings of regional identity. Journal of World 2(3): Wadley 2014) show that bifacial technology is not restricted to a 235–305. single cultural unit nor a specific temporal phase (Conard & Clark, J.L. & Plug, I. 2008. Animal exploitation strategies during the Porraz 2015) and the later Pietersburg of BRS provide another South African Middle Stone Age: Howieson’s Poort and post- example of this. There is a need for more detailed descriptions Howieson’s Poort fauna from Sibudu Cave. Journal of Human Evolu- of these bifacial technologies as well as of their technological tion 54: 886–898. Conard, N.J., Porraz, G. & Wadley, L. 2012. What is in a name? Charac- associations, a need to provide new insights on the relation- terising the ‘post-Howieson’s Poort’ at Sibudu. South African Archaeo- ships that unify and distinguish the different variants of the logical Bulletin 67: 180–199. MSA bifacial technologies. Conard, N.J. & Porraz G. 2015. Revising models for the cultural strati- The site at Bushman Rock Shelter opens a window facing graphic sequence of the Middle Stone Age. South African Archaeologi- north, looking at issues of contacts, influences and relation- cal Bulletin 70: 124–130. Deacon, H.J. & Deacon, J. 1999. Human Beginnings in South Africa: ships between south and east Africa, and between the coast Uncovering the Secrets of the Stone Age. New York:Rowman Altamira. and the Highveld. It also opens a new window that provides de la Peña, P., Wadley, L. & Lombard, M. 2013. Quartz bifacial points in for a retrospective into the legacy of previous researchers, and the Howieson’s Poort of Sibudu. South African Archaeological Bulletin perspectives on the origins of the MSA and some of its regional 68: 119–136. phases that have disappeared from the literature. de la Peña, P. & Wadley, L. 2014. Quartz knapping strategies in the Howieson’s Poort at Sibudu (KwaZulu-Natal, South Africa). PLOS ONE 9(7): e101534. ACKNOWLEDGEMENTS Eloff, J.F. 1969. Bushman Rock Shelter, Eastern Transvaal: excavations, We would like to acknowledge Ina Plug and Lyn Wadley 1967–8. South African Archaeological Bulletin 24: 60. for their invaluable support. We are very grateful to the staff at Fairhall, A.W.,Young,A.W.& Erickson, J.L. 1976. University of Washing- 178 South African Archaeological Bulletin 70 (202): 166–179, 2015

ton Dates IV. In: Flint, R.F., Ogden, J.G., Rouse, I. & Stuiver, M. (eds) and Swaziland. Strelitzia 19. Pretoria: South African National Radiocarbon 18(2): 221–239. Biodiversity Institute: 1–808. Goodwin, A.J.H. & Van Riet Lowe, C. 1929. The Stone Age Cultures of Parkington, J.E. 1988. The Pleistocene/Holocene transition in the West- South Africa. Trustees of the South African Museum. ern Cape, South Africa, observations from Verlorenvlei. BAR Interna- Henshilwood, C. S., D’Errico, F., Yates, R., Jacobs, Z., Tribolo, C., Duller, tional Series 405: 197–206. G.A.T., Mercier, N., Sealy, J.C., Valladas, H., Wattas, I.. & Wintle A.G. Pearson, O.M. 2008. Statistical and biological definitions of “anatomi- 2002. Emergence of modern human behaviour: Middle Stone Age cally modern” humans: suggestions for a unified approach to from South Africa. Science 295(5558): 1278–1280. modern morphology. Evolutionary Anthropology 17: 38–48. Henshilwood, C., D’Errico, F., Vanhaeren,M., Van Niekerk, K. & Jacobs, Plug, I. 1978. Die Latere Steentydperk van die Boesmanrotsskuiling in Z. 2004. Middle Stone Age shell beads from South Africa. Science Oos-Transvaal. Unpublished Masters dissertation. Pretoria: Univer- 304(5669): 404–404. sity of Pretoria. Henshilwood, C.S., Van Niekerk, K.L., Wurz, S., Delagnes, A., Plug, I. 1981a. Bushman Rock Shelter. In: Voigt, E.A. (ed) Guide to Armitage, S.J., Rifkin, R.F., Douze, K., Keene, P., Haaland, M.M., Archaeological Sites in the Northern and Eastern Transvaal: 111–131. Reynard, J., Discamps, E. & Mienies, S. 2014. Klipdrift Shelter, south- Transvaal Museum, Pretoria. ern Cape, South Africa: preliminary report on the Howieson’s Poort Plug, I. 1981b. Some research results on the Late Pleistocene and Early layers. Journal of Archaeological Science 45: 284–303. Holocene deposits of Bushman Rock Shelter, Eastern Transvaal. Herbert, D. & Kilburn, D. 2004. Field Guide to the Land Snails and Slugs of South African Archaeological Bulletin 36: 14–21. Eastern South Africa. Pietermaritzburg: Natal Museum. Plug, I. 1982. Bone tools and shell, bone and ostrich eggshell beads from Högberg, A. & Larsson, L. 2011. Lithic technology and behavioural Bushman Rock Shelter (BRS), Eastern Transvaal. South African Archae- modernity: new results from the Still Bay site, Hollow Rock Shelter, ological Bulletin 37: 57–62. Province, South Africa. Journal of Human Evolution Porraz, G., Texier, P-J., Archer, W.,Piboule, M., Rigaud, J.P.& Tribolo, C. 61(2): 133–155. 2013a. Technological successions in the Middle Stone Age sequence Ingman, M., Kaessmann, H., Pääbo, S. & Gyllensten, U. 2000. Mitochon- of Diepkloof Rock Shelter, Western Cape, South Africa. Journal of drial genome variation and the origin of modern humans. Nature 408: Archaeological Science 40(9): 3376–3400. 708–713. Porraz, G., Parkington, J., Rigaud, J-P., Miller, C.E., Poggenpoel, C., Klaric, L., Lev, S., Giria, Y. & Polanska, M. 2015. “Couteaux de Kostienki Tribolo, C., Cartwright, C., Charrié-Duhaut, A., Dayet, L., De Araujo, et lames aménagées par technique de Kostienki” Retour sur un M., Klein, R.G., Steele, T., Schmidt, P., Verna, C. & Texier, P-J. 2013b. malentendu historique. Bulletin de la Société Préhistorique Française The MSA sequence of Diepkloof and the history of southern African 112(3): 421–474. Late Pleistocene populations. Journal of Archaeological Science 40(9): Klein, R.G. 1978. Stone Age predation on large African bovids. Journal of 3542–3552. Archaeological Science 5: 195–217. Protsch, R. & De Villiers, H. 1974. Bushman Rock Shelter, Ohrigstad, Lombard, M., Wadley, L., Jacobs, Z., Mohapi, M. & Roberts, R.G. 2010. Eastern Transvaal, South Africa. Journal of Human Evolution 3: Still Bay and serrated points from Umhlatuzana Rock Shelter, 387–396. Kwazulu-Natal, South Africa. Journal of Archaeological Science 37: Ryano, K.P. 2014. The Later Stone Age in the Southern Cape, South 1773–1784. Africa, during the Terminal Pleistocene/Early Holocene with a focus Lombard, M., Wadley, L., Deacon, J., Wurz, S., Parsons, I., Mohapi, M., on Klipdrift Cave. Unpublished PhD thesis. Johannesburg: Univer- Swart, J. & Mitchell, P. 2012. South African and Lesotho Stone Age sity of the Witwatersrand. sequence updated (1). South African Archaeological Bulletin 67: 123–144. Sampson, C.G. 1974. The Stone Age Archaeology of Southern Africa. New Louw, A.W.1969. Bushman Rock Shelter, Ohrigstad, eastern Transvaal: York: Academic Press. a preliminary investigation, 1965. South African Archaeological Bulletin Sinclair, I., Hockey, P., Tarboton, W. & Ryan, P. 2011. Birds of Southern 24: 39–51. Africa, 4th edn. Cape Town: Struik Nature. Mackay, A. 2011. Nature and significance of the Howiesons Poort to Singer, R. & Wymer, J.J. 1982. The Middle Stone Age at Klasies River Mouth post-Howieson’s Poort transition at Klein Kliphuis rockshelter,South in South Africa. Chicago: University of Chicago Press. Africa. Journal of Archaeological Science 38(7): 1430–1440. Skinner,J.D. & Chimimba, C.T.2005. The Mammals of the Southern African Mackay, A., Orton, J., Schwortz, S. & Steele, T.E. 2010. Soutfontein Sub-Region. Cambridge: Cambridge University Press. (SFT)-001: preliminary report on an open-air site rich in bifacial Soriano, S., Villa, P.& Wadley, L. 2007. technology and tool forms points, southern Namaqualand, South Africa. South African Archaeo- in the Middle Stone Age of South Africa: the Howieson's Poort and logical Bulletin 65: 84–95. post-Howieson's Poort at Rose Cottage Cave. Journal of Archaeological Mason, R.J. 1957. The Transvaal Middle Stone Age and statistical analy- Science 34(5): 681–703. sis. South African Archaeological Bulletin 12: 119–137. Steele, T.E.& Klein, R.G. 2013. The Middle and Later Stone Age faunal Mason, R.J. 1962. Prehistory of the Transvaal. Johannesburg: Witwaters- remains from Diepkloof Rock Shelter, Western Cape, South Africa. rand University Press. Journal of Archaeological Science 40: 3453–3462. Mason, R.J. 1969. Tentative interpretations of new radiocarbon dates Texier P-J., Porraz G., Parkington J., Rigaud J-P., Poggenpoel C., Miller, for stone artefact assemblages from Rose Cottage Cave, O.F.S. and C.E., Tribolo, C., Cartwright C., Coudenneau A., Klein R., Steele T.& Bushman Rock Shelter, Tvl. South African Archaeological Bulletin 24: Verna, C. 2010. A Howiesons Poort tradition of ostrich egg- 57–59. shell containers dated back to 60,000 years ago at Diepkloof. Proceed- Mason, R.J., Brain, C.K., Brown, A.J.V., Cooke, H.B.S., Deacon, H.J., ings of the National Academy of Sciences of the of America Ewer,R.F.,De Graaff, B., Hofmeyr,I., Malan, B.D., Tobias, P.V., du Toit, 107(14): 6180–6185. A.P., Turton, K. & , L.H. 1988. Cave of Hearths, , Thompson, J.C. 2010. Taphonomic analysis of the Middle Stone Age Transvaal. Johannesburg: University of the Witwatersrand. Archaeo- faunal assemblage from Cave 13B, Western Cape, logical Research Unit, Occasional Papers 21. South Africa. Journal of Human Evolution 59: 321–339. McBrearty, S. & Brooks, A.S. 2000. The revolution that wasn’t: a new Thompson, J. & Henshilwood, C.S. 2014. Tortoise taphonomy and interpretation of the origin of modern human behaviour. Journal of tortoise butchery patterns at Blombos Cave, South Africa. Journal of Human Evolution 39(5): 453–563. Archaeological Science 41: 214–229. McNabb, J. & Sinclair, A. 2009. The Cave of Hearths: Makapan Middle Pleis- Underhill, D. 2012. The Earliest Middle Stone Age of northern South tocene Research Project: Field Research by Anthony Sinclair and Patrick Africa: the Cave of Hearths and Bushman Rock Shelter. In: Ruebens, Quinney, 1996–2001. Oxford: Archaeopress. K., Romanowska, I. & Bynoe, R. (eds) Unravelling the Palaeolithic. BAR McDougall, I., Brown, F.H. & Fleagle, J.G. 2005. Stratigraphic placement International Series 2400: 13–26. and age of modern humans from Kibish, Ethiopia. Nature 433: Villa, P., Soressi, M., Henshilwood, C. S. & Mourre, V.2009. The Still Bay 733–36. points of Blombos Cave (South Africa). Journal of Archaeological Science Meyer, A. 2007. Professor Hannes Eloff 1918–2007. South African Archae- 36(2): 441–460. ological Bulletin 62: 176–177. Villa, P., Soriano, S., Teyssandier, N. & Wurz, S. 2010. The Howieson's Mitchell, P.1988. The Late Pleistocene early microlithic assemblages of Poort and MSA III at Klasies River main site, cave 1A. Journal of Archae- southern Africa. World Archaeology 20(1): 27–39. ological Science 37(3): 630–655. Mitchell, P. 2002. The Archaeology of Southern Africa. Cambridge: Cam- Vogel, J.C. 1969. Radiocarbon dating of Bushman Rock Shelter, bridge University Press. Ohrigstad district. South African Archaeological Bulletin 24: 56. Mucina, L. & Rutherford, M.C. 2006. The vegetation of South Africa, Vogelsang,R., Richter,J., Jacobs, Z., Eichhorn, B., Linseele, V.& Roberts, South African Archaeological Bulletin 70 (202): 166–179, 2015 179

R.G. 2010. New excavations of Middle Stone Age deposits at Apollo Wadley, L. 2012. Two moments in time during Middle Stone Age occu- 11 Rockshelter, : stratigraphy, archaeology, chronology and pations of Sibudu, South Africa. Southern African Humanities 24: past environments. Journal of 8(2): 185–218. 79–97. Volman,T.P.1981. The Middle Stone Age in the Southern Cape. Unpub- Wadley, L. & Vogel,J.C. 1991. New dates from Rose Cottage Cave. South lished PhD thesis. Chicago: University of Chicago. African Journal of Science 87: 605–608. Volman, T.P. 1984. Early Prehistory of southern Africa. In: Klein, R.G. Wadley, L. & Kempson, H. 2011. A review of rock studies for archaeolo- (ed.) Southern African Prehistroy and Paleoenvironements: 1–64. Rotter- gists, and an analysis of dolerite and hornfels from the Sibudu area, dam: A.A. Balkema. KwaZulu-Natal. Southern African Humanities 23(1): 87–107. Wadley, L. 1986. A social and ecological interpretation of the Later Watts, I. 2002. Ochre in the Middle Stone Age of southern Africa: ritual- Stone Age in the southern Transvaal. Unpublished PhD thesis. ised display or hide preservative? South African Archaeological Bulletin Johannesburg: University of the Witwatersrand. 57: 1–14. Wadley, L. 1987. Later Stone Age Hunters and Gatherers of the Southern White, T.D., Asfaw, B., DeGusta, D., Gilbert, H., Richards, G.D., Suwa, Transvaal: Social and Ecological Interpretation. BAR International Series G. & Howell, F.C.2003. Pleistocene Homo sapiens from middle Awash, 380. Ethiopia. Nature 423(6941): 742–747. Wadley, L. 2007. Announcing a Still Bay industry at Sibudu Cave, South Wurz, S. 2000. The Middle Stone Age at Klasies River, South Africa. Un- Africa. Journal of Human Evolution 52(6): 681–689. published PhD thesis. Stellenbosch: Stellenbosch University. Wadley, L. 2010. Were snares and traps used in the Middle Stone Age Wurz, S. 2013. Technological trends in the Middle Stone Age of South and does it matter? A review and a case study from Sibudu, South Africa between MIS 7 and MIS 3. Current Anthropology 54(S8): Africa. Journal of Human Evolution 58: 179–192. S305–S319.