An initial assessment of zooarchaeological assemblage sizes from South Africa
MetadataShow full item record
Original versionRevue de Paléobiologie. 2022, 41 (1), 1-10. 10.5281/zenodo.6022753
Large faunal assemblages remain a concern in zooarchaeology. Without sufficiently large assemblages, interpretations about the past are less secure. The number of potential taxa in an assemblage is related to various factors such as the number of species found in an ecological zone, economic activities (hunting vs. farming) and taphonomy. In South Africa, no attempt has been made to determine what constitutes a sufficiently representative assemblage size, yet animal remains are frequently utilised to infer aspects of human behaviour during the past. We use faunal assemblages analysed using a standardised method, and we plot the number of taxa against the number of identified specimens (NISP). This way, we are able to determine the asymptote point of a faunal assemblage. Our initial study suggest that for subsamples, specifically those containing only large mammal remains, the asymptote is reached nearing 1000 NISP, whereas for those assemblages containing a full range of vertebrates and invertebrates, the asymptote is reached when the NISP nears 2000. The assemblages used are from different ecological zones and vastly different time periods, and each has its own taphonomic history. As a result, these asymptote points should not be considered as adequate samples for every faunal assemblage in South Africa. Instead, we advocate that zooarchaeologists analyse the largest possible assemblage to ensure more reliable interpretations are made from faunal remains. In many instances, only smaller assemblages are available; this is often beyond the control of the zooarchaeologist. Nevertheless, our data suggest that often, assemblages are adequate in size for archaeological sites as a whole; however, when comparisons are attempted between layers, occupational phases or spaces at a site, assemblage sizes are often severely reduced, requiring caution when making inferences about the past.