A New Statistical Tool for Comparing Animal Bone Assemblages

Abstract Recent statistical developments have led to the creation of a distinctive approach to the comparison of quantified assemblages of pottery sherds (the pie-slice approach). This paper investigates, in both theory and practice, whether the same approach can be applied to assemblages of animal bone fragments. The conclusion is that, provided an appropriate means of quantification is used, the transfer of the approach is valid, and that useful results are generated. Assemblages from sites in Roman, Saxon, mediaeval and post-mediaeval London are compared in terms of both species and bone elements represented, using correspondence analysis to display the results. Chronologial trends, and functional differences between sites of the same period, are observed, and problems caused by differential retrieval or recording are high-lighted by this approach.