Geoarchaeology and zooarchaeology of Border Cave, South Africa: Initial multiproxy considerations of stratigraphy and site formation processes from the Backwell et al. excavations

Abstract

Border Cave (BC) has accumulated over 200,000 years of archaeological deposits that document remarkable evidence of human behaviour during the Middle and Later Stone Age. For nearly fifty years, researchers have relied on the stratigraphic framework established by Peter Beaumont in 1973, in which the deposits are lithostratigraphically categorized into a sequence of alternating ‘Brown Sand’ (BS) and ‘White Ash’ (WA) members. Geoarchaeological work in the 1970s focused on stratigraphic sequencing of the anthropogenic assemblages, and proposed broad correlations between autogenic contributions and environmental conditions. The research presented here was undertaken as part of a new excavation campaign at Border Cave started in 2015 under the direction of Backwell at al. Re-examining the stratigraphic context of the deposits and assessing site formation processes are among the key goals of this project; this will enable finer-scale intra- and inter-member comparative analyses of the artefacts and ecofacts recovered at the site. In this paper, we apply a facies and allostratigraphic approach to assess the stratigraphic sequence exposed through the Backwell et al. excavations. We also provide an initial assessment of the prevailing site formation processes active in the deposition and modification of the sediments. The geoarchaeological data are integrated with new zooarchaeological and taphonomic evidence in order to explore inter- and intra-unit patterns throughout the sequence. Results of this work are: (1) exposed sediments can be broadly correlated to members of the Beaumont sequence; (2) we clearly define member boundaries, reassess member stratigraphic complexity and recognise finer intra-member layering; (3) geoarchaeological and taphonomic studies demonstrate that the sediments have been subjected to greater post-depositional disturbance than was previously recognised and affect all levels of the sequence; (4) overall, faunal density at BC appears to be much lower than that at other Middle Stone Age sites such as Blombos and Sibudu; (5) multiproxy analysis suggests that WA and BS members have distinctive taphonomic histories that cross-cut the identified archaeological industries. As such, caution is warranted when combining BS and WA members for analysis of artefacts and ecofacts.