Three Dimensional Aggregate Flake Scar Analysis and Hominin Behavior at Tabun Cave, Israel

Abstract

In the field of archaeological lithic analysis, dorsal flake scar directionality has been used for decades to infer methods of core reduction or flake production. The established method of analyzing these data has been with an approach that places patterns of scar directions into discreet, two dimensional categories, which can then be simply quantified and compared. The present work grew out of a desire to study this aspect of lithic variability as a continuous variable with enhanced statistical rigor, and to do so in three dimensions (3D). My study involves creating three dimensional virtual flake models, created using a NextEngine Desktop 3D laser scanner and 3D Systems Geomagic software suite for model assembly, orientation and data extraction. The use of 3D models allows for the collection of flake scar directions using 3D vectors. As individual flakes have sample sizes of scars that are too small for statistical analysis, the vectors from the individual flakes are combined into pooled assemblages, from which a host of variables are calculated and compared via bootstrapping. Sample analyses using these new variables are compared to results derived from the traditional two dimensional, categorical method. In this study I used a combination of experimentally produced assemblages representing bifacial, blade, discoidal and levallois reduction methods, and archaeologically derived samples from Tabun Cave, Israel. The experimental samples provide a baseline of comparison for illustrating potential areas of inter- and intra-assemblage variability that are investigated in the archaeological samples. Archaeological samples from Tabun Cave, Israel, representing Units I, IX, respectively Tabun C-type and Tabun D-type, and some of their respective Beds, are subjected to the same analysis, as a test case for the applicability of the method to an archaeological assemblage which is noted both for its layers characterized by certain flake scar patterning that change during the site’s sequence and for its status as a touchstone assemblage for the Levantine Paleolithic. Results from the experimental assemblages suggest different methods of reduction result in statistically distinguishable scar patterning in some cases. Results from the archaeological assemblage analysis suggest some assemblages are indistinguishable, such as the two Beds from Unit I, while other comparisons, such as between Units I and IX as well as between some of the Beds within Unit IX, suggest the utility of the method to differentiate archaeological samples by their 3D scar pattern variables. In particular, the ability of the method to distinguish between individual Beds within Unit IX calls into question the legitimacy of treating Tabun “types” as monolithic categories.