Heritage Cases

The objective of this project is to clarify the late Pleistocene/early Holocene sequence of Heuningneskrans, a rock shelter excavated by P. Beaumont in 1968. This project is part of a broader enterprise of clarifying the Middle and Later Stone Age sequence in Limpopo, which began in 2014 with renewed excavations at Bushman Rock Shelter, located about 3 km away from Heuningneskrans. Our field activities at Heuningneskrans will be limited since they will only consist of conducting Ground Penetrating Radar, opening the previous excavation area, recording and sampling the archaeological profiles. No excavation will be carried out at Heuningneskrans besides a very limited one on the profiles that will require stabilizing before closing the site. The field work at Heuningneskrans follows recent preliminary studies that we conducted on the lithic and faunal collections recovered during Beaumont’s excavation. In order to complete these ongoing studies and publish results, and given the complete absence of field documents, the acquisition of new field data is required (i.e., radiocarbon chronology, site formation processes, stratigraphic reappraisal of the profile, detailed description of the layers, etc). Scientific background to the project While current research is uninterruptedly increasing our understanding of the Middle and Later Stone Age developments in southern Africa, several critical aspects are still being debated. One such aspect relates to the nature of the transition from the Middle Stone Age (MSA) to the Later Stone Age (LSA); a transition, which, unlike in the European context, is not associated with the replacement of one hominin taxon by another one, and which does not fit neatly with significant climatic changes since it precedes the Pleistocene/Holocene boundary by at least 10 thousand years. The discovery of microlithic artefacts likely designed to be hafted, as well as evidence for personal ornaments and other expressions of symbolic behaviour in techno-complexes clearly situated well within the MSA (i.e., the Still Bay and Howiesons’ Poort; see for instance: d’Errico et al., 2005; Henshilwood et al., 2002, 2011; Texier et al., 2013) has contributed to blur the lines between what can be considered to be a “typical” MSA industry and a “typical” LSA one. Furthermore, the question of when, where, and why the LSA phenomenon emerged remains largely unanswered. For a start, dates separated by as much as 20 thousand years have been proposed for the beginning of the LSA at sites located across South Africa. Border Cave, in the Lebombo Moutains on the South African south-eastern border with Swaziland, is considered to have yielded the oldest LSA assemblage known so far, in layers dated to 42-44 ka cal. BP. This assemblage has been attributed to the Early LSA (ELSA) and contains abundant unretouched microliths produced via bipolar knapping associated with bored stones, bone points, engraved bones, ostrich eggshell beads, and wood objects (Villa et al., 2012; d’Errico et al., 2012). This assemblage arguably differs from the previous one, attributed to the post-Howiesons Poort and representing a last expression of the MSA at the site. The concept of an “Early” LSA was initially defined by P. Beaumont, who, together with J. Vogel, considered this industry to take place after the end of the Middle Stone Age and before the “classical” LSA (Beaumont & Vogel, 1972). According to his earlier work, the ELSA would have develop over a long period, going from as far as ~40 ka BP, up until the beginning of the Holocene (~7 ka BP). Reappraisal of the LSA sequence in southern Africa has somewhat modified these views, leading to subdivisions within the ELSA. L. Wadley notes for instance that “the term “early LSA” (ELSA), as originally proposed by Beaumont and Vogel (1972), subsumed bladelet-poor microlithic industries dating back to ca. 40,000 B.P., bladelet-rich microlithic industries, and also non-microlithic industries of the terminal Pleistocene and the early Holocene” (Wadley, 1993: p.264). Today, the ELSA is considered to exist between ~40 to 20 ka BP, and is followed by a microlithic, bladelet-rich industry, the Robberg (~20-12 ka BP), itself preceding a non-microlitic industry, the Oakhurst (~12-7 ka BP). Lombard et al. (2012) propose the following definition of the ELSA: “informal designation; also know as transitional MSA-LSA; [...] characterised by unstandardised, often microlithic pieces and includes the bipolar technique; described at some sites, but not always clear whether assemblages represent a real archaeological phase or a mixture of LSA/MSA artefacts” (Lombard et al., 2012: p.136). This definition underlines the lack of agreement regarding this techno-complex, still considered by some to represent mixed assemblages rather than a real phenomenon. Another issue related to the ELSA is the fact that, during that same period (~40-20 ka BP), several sites in southern Africa are still producing MSA assemblages (the “final MSA”) (e.g. Porraz et al., 2016). Whether there is a high degree of regional variability, issues with dating, variations in level of preservation of organic material between sites, or problems of typo-technological definitions of the lithic assemblages is unclear and calls for further investigation. In order to clarify the Stone Age sequence in South Africa, one possible strategy consists in promoting regional scenarios based on long, well stratified and well dated archaeological sequences. In that perspective, our field project at Heuningneskrans (referred to as ‘HNK’ hereafter) fits within our current research and excavation at Bushman Rock Shelter (referred to as ‘BRS’ hereafter), which aims at clarifying the MSA/LSA occupations, their chronology as well as their technological and cultural expressions (Porraz et al., 2015). HNK and BRS are only 3 km from one another and provide two complementary archaeological sequences for the study of the Late Pleistocene/early Holocene occupations in Limpopo. In particular, we expect two direct benefits from the comparison between the two sites. Firstly, BRS presents a clear and long hiatus between the end of the MSA, dated to MIS5, and the oldest LSA layers, dated to about 14.5 ka BP. The reasons for this non-depositional and non-occupational phase at BRS remain unclear; yet, contrastingly, HNK does record occupations older than 14.5 ka BP. Secondly, BRS and HNK both document occupations from the end of the Pleistocene and the beginning of the Holocene, allowing us to build a solid reconstitution of past environments and to provide a clear picture of LSA hunters-gatherers life ways. Permission to study the lithic and faunal material recovered during P. Beaumont’s excavation was granted last year by David Morris, curator of the archaeological collections at the McGregor Museum in Kimberley. Re-analyses of these collections have started and confirm the great potential of the HNK site. However, field documentation is a necessary pre-requisite before any conclusions and comparisons can be proposed: to this end, the site therefore needs to be re-opened. Description of the site The Heuningneskrans, or Heuningsneskrans (Afrikaans for ‘the cliff with honey nests’), Shelter is located on the eponym farm, Farm Heuningsneskrans No.476, owned by M. Marnus van der Wal, in the Lydenburg District, 18 km north of Ohrigstad in Limpopo Province, South Africa. It is a very long dolomitic rock shelter facing North-West, which floor has a maximum depth of 14 m and a maximum width of 210 m (Beaumont, unpublished). The vegetation around the site is assigned to the Mixed Bushveld type. The shelter itself belongs to the Savannah Biome and sits at an altitude of about 950 m above sea level, on the fringe of the Great Escarpment. Together with BRS, which is located less than 3 km away from HNK, it is one of a handful of rock shelters in the Limpopo region to contain well-stratified late Pleistocene/early Holocene archaeological deposits. Initially used as a tobacco shed by his previous owner, M. Sakkie van der Wal, the site caught the attention of James Kitching and the late Peter Beaumont, then based at the Bernard Price Institute in the University of the Witwatersrand, when some archaeological artefacts were recovered in a pit dug by the farmer in an attempt to expand his leaf-drying facilities in 1967. Coordinates of the site: 24°36’ S latitude; 30°39’ E longitude Previous work by Peter Beaumont Following this discovery, a single campaign of excavation took place at the site, from the 14th to the 21st of December 1968, under the direction of P. Beaumont, accompanied by a party of seven of his young colleagues, as well as six of the farm workers. A ~13.65 m long trench was dug, from the drip line of the shelter to the back wall (15 squares of 0.91 m2 each, labelled A1 to A15). The “absence of any clear or consistent natural stratification” (Beaumont, unpublished: p.3) led to an excavation that proceeded by spits of 7.5 cm in the upper part and 15 cm in the lower part of the deposits. Depth of deposits varied greatly due to the general morphology of the excavation, but a maximum depth of 6.35 m was reached in square A12. At this point, large slabs of dolomite were encountered and interpreted as the bedrock. A total of 15.21 m3 of sediments were removed and sieved using 3 mm screen meshes. Stratigraphy of the deposits P. Beaumont identified three main stratigraphic strata, from top to bottom, each of them subdivided into two or more layers – it is important to note here that such subdivisions were decided a posteriori and based exclusively on variations in density and type of archaeological content rather than on clear sedimentary changes. Thus: “Analysis revealed that artefacts were not evenly distributed with depth in the deposit. They tend rather to occur in zones of lesser or greater abundance. These variations were used to split up Stratum 1 into eight sublevels (1a-h), Stratum 2 into two sublevels (2a+b), and Stratum 3 into two sublevels (3a+b)” (Beaumont, unpublished: p.4). Below is a short description of each of these three strata as provided by P. Beaumont in his report.  Stratum 3 (depth: 0 to 69/76 cm): this first unit contains grey/brown sand interspersed with small and sporadic ash lenses. It is subdivided into substrata 3a and 3b;  Stratum 2 (depth: 69/76 to 145/152 cm): the second unit is characterized by more abundant ash lenses of various colour (white, grey and black), and smaller quantities of brown sand. It is subdivided into substrata 2a and 2b;  Stratum 1 (depth: 145/152 to 635 cm): this last and oldest unit contains mostly fine brown sand, with minor and sporadic ash lenses. It is subdivided into eight substrata, from 1a to 1h. During the excavation, a change in the calcification intensity and degree of humidity was noted: “much of Stratum 1 was perceptibly calcified. Marked vertical variations in calcification intensity occur within in.” And: “the deposit below about 480 cm was found to be increasingly moist with depth” (Beaumont, unpublished: pp.3-4). Cultural sequence The general impression, when looking at mentions of HNK in the literature of the past 50 years, is one of confusion when it comes to its techno-cultural assignment. Various and sometimes contradictory attributions have been proposed by different authors (see for instance: Beaumont, unpublished; Beaumont & Vogel, 1972; Klein, 1984; Wadley, 1993; Lombard et al., 2012). This is partly due to ongoing redefinition and refining of the South African Middle and Later Stone Age sequence, but this also clearly highlights the crucial need to provide new, up-to-date, description and analysis of the archaeological sequence of the site. Initially, the whole sequence was attributed by P. Beaumont to the Early Later Stone Age (ELSA) (Beaumont, unpublished report; Beaumont & Vogel, 1972), with the exception of the very top layer, stratum 3a, which contains abundant potsherds and glass beads, consistent with an Iron Age/modern occupation. However, in a paper postdating the excavation by about 15 years, R. Klein, presumably following consultations with P. Beaumont, describes a somewhat different sequence from the one initially proposed, insofar as he does not even mention an ELSA occupation anymore. Rather: “Beaumont found Early Iron Age potsherds at the very top of the sequence and a few Middle Stone Age flakes among the dolomite blocks near bedrock. He assigned the artefacts in between to three successive LSA industries. The oldest, in layers 3B to 3H, includes bladelets and associated pyramidal cores similar to those found in the Robberg Industry of the southern Cape. [...] The youngest industry, in levels 1A and 1B, contains large scrapers and other flake artefacts that resemble ones now often subsumed in the ‘Oakhurst Complex’ of Sampson (1974). [...] Between the Robberg-like and Oakhurst Complex industries is a third industry which Beaumont regards as typologically and technically intermediate between them.” (Klein, 1984: p.109). Interestingly, no mention of bladelets and pyramidal cores is made anywhere else by P. Beaumont himself - it is remarkably absent from his unpublished report. Besides, no drawings of such artefacts exist. This is actually true of the whole lithic assemblage from the site, for which no pictures and no drawings seem to have been produced. The only information currently available is contained in tables presenting percentages of various categories of lithic artefacts (i.e. irregular flakes, blade flakes, scaled pieces and several types of scrapers) in P. Beaumont’s unpublished report. This renders the correct techno-cultural attribution of the lithics from HNK virtually impossible without a detailed re-analysis of the archaeological collections. Finally, in a much more recent synthesis of the Stone Age sequence in South Africa and Lesotho, Lombard et al. (2012) place HNK in the list of sites associated with the ELSA – but do not include it in their list of sites containing Robberg and Oakhurst industries. Despite clear issues regarding the cultural attribution of the HNK archaeological deposits and the lack of detailed description of the lithic assemblage, the site is well dated. Two series of radiocarbon dates on charred bones and charcoals were obtained from most layers in the 1970’s, by two different laboratories: the National Physical Research Laboratory of the C.S.I.R. in Pretoria (Vogel & Marais, 1971) and the Mt Soledad Radiocarbon Laboratory of La Jolla, in California (Linick, 1977) (Table 1). These two sets of ages seem to be in accordance with one another and reliable. The basal age of >33 ka for the oldest layer (Table 1) is based on estimates of sedimentation rates rather than direct dating: “at the Heuningsneskrans Shelter 6,4 ma of “Early L.S.A.” overlies bedrock. The lower three datings (69a-c) indicate that the upper and middle levels of Stratum 1 accumulated uniformly at a rate of 1 m in about 5 100 years. Extrapolation on this basis suggests that the base of the deposit here must date to c. 32500 B.C.” (Beaumont & Vogel, 1972). Regardless of the validity or not of this maximum age, the sequence covers the end of the Pleistocene, from at least 24 ka BP, all the way well into the Holocene (ca. 7 ka BP). Table 1. Radiocarbon dates available for the HNK sequence. Stratum Material dated Cultural attribution Beaumont* Klein (1984) Age estimate Source 3b Charred bone ELSA Oakhurst 7200 ± 70 BP Vogel & Marais, 1971 3b - 8880 ± 100 BP Linick, 1977 2a Charcoal ELSA Intermediate industry 9780 ± 85 BP Vogel & Marais, 1971 2a - 9230 ± 100 BP Linick, 1977 1a Charred bone ELSA Robberg-like 10 430 ± 150 BP Vogel & Marais, 1971 1b - 12.260 ± 110 BP Linick, 1977 1b Charcoal 13 100 ± 110 BP Vogel & Marais, 1971 1c - 19.320 ± 240 BP Linick, 1977 1d - 19.980 ± 260 BP Linick, 1977 1e - 20.510 ± 270 BP Linick, 1977 1e Charred bone 24 630 ± 300 BP Vogel & Marais, 1971 1f - 23.400 ± 500 BP Linick, 1977 Base N/A ca.33 ka Beaumont & Vogel, 1972 *Initial attribution by P. Beaumont in his unpublished report, and in Beaumont & Vogel (1972) Rationale for reopening the site As explained above, clarification of the HNK archaeological sequence is required. The 1968’s excavation conducted by P. Beaumont lacked formal field documentation in terms of field notes, photographs and drawings. This was corroborated by a recent verbal conversation (2015) between one of us (G.P.) and P. Beaumont, who confirmed that no such information was recorded during his excavation. Besides, very few publications have dealt with the HNK sequence. The set of existing radiocarbon dates positions HNK as a key site for the second part of the Late Pleistocene. Our preliminary study of the lithic collections clearly documents the presence of ELSA, Robberg and Oakhurst industries at HNK, thus making it one of the very few South African sites with ‘continuous’ deposits from the ELSA to the Oakhurst. Regarding the importance of such cultural and palaeoenvironmental information (it is worth noting that Mitchell considers HNK as one of the oldest Robberg sites in South Africa: Mitchell, 1988, 1995), the acquisition of new field data is crucial. In order to provide a good stratigraphic context for the collections of P. Beaumont (collections currently under study), we propose to re-open the site, re-expose, clean, record and study the archaeological profiles left by him. HNK and BRS share important similarities and some of the specialists from our team involved in the study of BRS already started to analyse the old collections from HNK. This is notably the case for L. Feyfant (PhD candidate, co-registered at the University of the Witwatersrand in Johannesburg and the Université Jean Jaurès in Toulouse, France; see Annex 1 for more details on this PhD project), who is analysing the LSA faunal assemblages from both sites for her PhD work. Similarly, lithic assemblages from the LSA layers of both localities are studied following the same methodological approach (studies directed by G. Porraz). We propose to address and clarify the following specific points in the field:  What are the site formation processes at play? In particular, what is the degree of anthropogenic input (the occurrence of ash lenses was mentioned by P. Beaumont)? How has the nearby Ohrigstad River impacted the sedimentation at the site, in terms of erosion and deposition?  How relevant are the three stratigraphic units defined – a posteriori – by P Beaumont?  How is Beaumont’s trench positioned within the general setting of the site?  What is the degree of preservation for archaeobotanical remains (e.g. pollens and phytoliths) at the site? The potentially good degree of pollen preservation in the basal layers is suggested by the relative high humidity content (P. Beaumont, unpublished report and see above). High levels of humidity within archaeological deposits tend to be conducive for pollen preservation.  Where are hiatuses at the site and where are the ‘contacts’ (e.g. ELSA/Robberg)? How do depositional processes relate – or not – with cultural changes? Available dates for the various strata of the site suggest the existence of a chronological gap of about 7 000 years, between the Last Glacial Maximum (strata 1c-1f; ~24.5 to 19.5 ka BP) and the Pleistocene/Holocene boundary (strata 1b to 3b, ~13 to 7 ka). The exact nature of this apparent hiatus (i.e. stop in sedimentation, phase of erosion, other) will be investigated by the geoarchaeological team. Objectives of the field project The objectives of this new field project taking place at HNK are straightforward and highly feasible. They are presented in details below.  Ground Penetrating Radar exploration Prior starting the actual field work, we will investigate the extent and depth of remaining archaeological deposits within the shelter using a Ground Penetrating Radar (GPR). As pointed out by P. Beaumont, the rock shelter is 210 meters long and he estimated that “the entire site contains about 10 000 cu. m. of deposit” (Beaumont, unpublished). GPR exploration offers the double advantage of being completely non-destructive and of providing preliminary results rapidly – i.e. directly on site. Such geophysic analysis will also help circumscribing the exact limits of the excavated trench.  Exposing and cleaning the stratigraphy Based on recent visits to the site and with the help of the plan drawn by P. Beaumont, the trench appears to still be visible. The excavated zone seems to have been backfilled and to have undergone little change since the original excavation. The strategy will consist in exposing the profiles, sieving disturbed sediments (or a sample of them, depending on the richness of the archaeological content), and filling military bags that will be used during excavations. Of importance will be to match our field investigation with P. Beaumont’s data, as well as to record any landmarks (e.g. nails, sampling areas, etc) related to previous field activities. A first photogrammetric model will be produced.  Stabilizing the profiles We expect to find the archaeological profile left by P. Beaumont in various states of preservation. We therefore might have to stabilize part of the profile and/or ensure that it is straightened before any descriptions and sampling can take place. When limited excavation will be needed, classical excavation procedures will be adopted, following the ones used at BRS (for more details on excavation and post-excavation protocol, see Annex 2). Tasks related to excavation, including sieving using 3 and 1 mm meshes, cleaning and sorting will take place directly inside the shelter, as the field work progresses. Position of archaeological remains equal to and larger than 2 cm will systematically be plotted using a Total Station (Annex 2).  Sampling Once the archaeological sequence will have been clarified, the second step of the field work will entail various types of sampling. The first type will include micromorphological blocks that will serve to produce thin sections of specific parts of the deposits, and sediment samples. Both will be subjected to various sets of analyses in order to provide information regarding to the exact composition of the different stratigraphic units (i.e. identification of mineral and organic content) and the nature of the contact between units. This will help, ultimately, to understand the general history of the deposits in terms of accumulation processes. A second type of sampling will be conducted in order to collect potential microscopic palaeoenvironmental proxies in the form of pollens and phytoliths. Sediment samples will be collected to this end, followed by analytical work in the lab. Finally, we will collect new organic samples directly from the profile, in other words from a secure context. This is relevant for the establishment of the chrono-cultural sequence. These organic samples will be further used for radiocarbon dating. Closure and conservation plan of the site Previous excavation conducted at the site in 1968 was backfilled and, while the shelter has been used by the previous and current farmer as a meeting place, no further work (i.e. no extraction of sediments) was performed since. Besides, the rock shelter is located on a private farm and is therefore reasonably protected from undesirable visits. At the end of our field project, we will backfill the site with sand bags, in order to keep it protected in case of future investigations. The various objectives aforementioned should all be reached within a single four-week field campaign (planned for 2018). If needed, consultation with the appropriate SAHRA authorities will be made before any further field work takes place at the site. A trustworthy relationship has already been established with the current owner of the farm, M. Marnus van der Wal, who has granted us access to the site for the duration of the field project (cf. his letter of consent attached with this application). Together with him, we are planning to do some heritage site promotion, directly in the rock shelter, most likely in the form information panels providing a description of the site, its significance within South African prehistory, previous work done by P. Beaumont and aims/results of our project. Finally, progress reports on the field project will be sent to SAHRA in due time. Research team and scientific production As previously stated, the project at HNK closely relates to the ongoing research and excavation at BRS. As such, the research team will be composed of the same specialists (Table 2). Table 2. Composition of the scientific team working on the HNK project. Discipline Specialist Affilation Lithic technology Dr. Guillaume Porraz CNRS, Université Paris Ouest Nanterre, France & Evolutionary Studies Institute (ESI), University of the Witwatersrand, Johannesburg Dr. Paloma de la Peña ESI, University of the Witwatersrand, Johannesburg Lithic use-wear Dr. Marina Igreja Universidade do Porto, Porto, Portugal Residue analysis Dries Cnuts TraceoLab / Prehistory, University of Liège, Belgium Dr. Veerle Rots TraceoLab / Prehistory, University of Liège, Belgium Adhesive analysis Dr. Armelle Charrié-Duhaut CNRS, Université de Strasbourg, France Pigments Dr. Laure Dayet Université Toulouse Jean Jaurès, France Zooarchaeology Léa Feyfant* ESI, University of the Witwatersrand, Johannesburg & Université Toulouse Jean Jaurès, France Personal ornaments Dr. Laure Dayet Université Toulouse Jean Jaurès, France Archaeobotany Dr. Christine Sievers (seeds) School of Geography, Archaeology and Environmental Sciences (GAES), University of the Witwatersrand, Johannesburg May Murungi (phytoliths) ESI, University of the Witwatersrand, Johannesburg Elysandre Puech (pollens) ESI, University of the Witwatersrand, Johannesburg & EPHE, Université de Bordeaux, France Geoarchaeology Prof. Chris Miller Institute for Archaeological Sciences, University of Tübingen, Germany Dr. Susan Mentzer Institute for Archaeological Sciences, University of Tübingen, Germany Geomorphology Dr. Laurent Bruxelles INRAP, Toulouse, France; IFAS, Johannesburg & GAES, University of the Witwatersrand, Johannesburg Ancient DNA Riaan Rifkin Department of Genetics, University of Pretoria Jean-Baptiste Ramond Department of Genetics, University of Pretoria * See Annex 1 Besides the Master and PhD projects linked to HNK and the publication of results of various studies on the site in peer-reviewed, international journals, it is worth mentioning that the field project at HNK precedes the organization of a two day-conference that we are planning to organize on the late Pleistocene and early Holocene record in southern Africa.