Analysis of a Paleo-Indian occupation floor at the Duck River site ELMB-10, Manitoba

Abstract

This thesis deals with a proposed analytic procedure designed to aid in the analysis of marginal recovery sites, which present obstacles to analysis such as poor preservation or small sample size. ElMb-10, an Agate Basin site in Manitoba containing only lithic artifacts, was selected as the test case for the analysis. The analytic procedure consisted fo cluster analysis, a technique for grouping data on the basis of similarity, and nearest-neighbour analysis, which generate a measure of spatial dispersion. the cluster analysis was used to investigate the nature of artifact production at the site and to generate sets of functional tools. The tool sets were combined into activity-related kits and submitted to the spatial analysis. The presence of activity areas is inferred from the results of the nearest-neighbour analysis. The results of the flake cluster analysis indicate that no flake types were sought by the artisan(s), and a direct percussion knapping technique is inferred. The knapping technique is discussed in relation to technological parameters and the availability of raw materials. The second cluster analysis isolates several tool classes. When these are combined as tool kits and submitted to nearest-neighbour analysis the results indicate the spatial aggregation of three kits: butchering, hideworking and woodworking. Maps of the distribution of the kits delineate three activity areas. A rock configuration is also submitted to spatial analysis and the results indicate that it was a man-made structure--possibly a windbreak. Interpretations of site structure are discussed in relation to environmental conditions and the analytic procedures are reviewed and evaluated.