Sampling to redundancy in an applied zooarchaeology: A case study from a freshwater shell ring in the Mississippi Delta, southeastern USA

Abstract

Different archeological research goals necessitate different target and measured variables. In applied zooarchaeology, the target can be 1) a full list of taxa present at a site, to allow for biogeographical mapping; 2) species proportions representing those present in site deposits, which in some situations can be argued to represent past faunal communities, one possible target state for habitat reconstruction; or 3) both. For any of these goals, understanding the effects of sampling and differential preservation is key. We investigate the effects of repeated random sampling of a surface assemblage of freshwater mussel shells acquired from a prehistoric site in Sunflower County, Mississippi, southeastern USA. Evaluation of Species Area Curves shows that redundancy in taxonomic presence is reached at an average of ca. 8200 shells, but that meaningful values for diversity measures can be obtained with smaller samples. Comparison with sub-plow zone materials shows that denser, more robust species are over-represented, and thin, rod-shaped species under-represented, in the surface assemblage, while interpretation of the excavated materials is hampered by insufficient spatial sampling. Regardless of such issues, both data sets have value depending on the target variables of interest.