Probability Models of Bone Surface Modification and Application to Fossil Evidence from Ledi-Geraru (2.82 Ma) and Dikika (3.39 Ma), Afar Ethiopia

Two of the defining behaviors associated with the hominin lineage are an increased reliance on tool use and the routine incorporation of animal tissue in the diet. These adaptations have been linked to numerous downstream consequences including key physiological adaptations as well as social and cognitive effects associated with modern humans. Thus, a critical issue in human evolution is how to determine when hominins began incorporating significant amounts of meat into their diets. Bone surface modifications (BSM) have long been recognized as a powerful inferential tool in identifying the differential involvement of actors responsible for altering assemblages of bone recovered from both archaeological and paleontological contexts and remain a primary source of direct evidence for butchery activities. Thus, determining the spatiotemporal context of increased carnivory in the hominin lineage relies on the accurate identification of fossil BSM.

Multidecade-long debates over the agents responsible for individual BSM indicate systemic flaws in historical approaches to identification. These debates are in part due to the extreme morphological overlap between BSM produced by certain agents of modification. The primary goal of this dissertation project therefore, is to construct probability models of BSM capable of identifying individual marks with an associated probability of assignment. Using a multivariate Bayesian approach to analyze experimentally-generated BSM data, this dissertation uses two different models, one incorporating both two and three-dimensional (3D) metric and attribute data associated with individual BSM and a second model comparing 3D geometric morphometric (GM) shape data associated with BSM.

The 2D/3D attribute model of BSM is used evaluate an assemblage of fossil BSM recovered from the Ledi-Geraru research area, Ethiopia (2.82 Ma) in spatiotemporal association with early Homo. The results of the analysis reveal compelling evidence for early butchery activities, suggesting hominins may have been using both modified and unmodified stone implements to process carcasses.

The second model, based upon 3D GM data, was used to evaluate the earliest purported evidence for stone-mediated butchery at Dikika, Ethiopia (3.39 Ma). The Dikika marks have been argued to be the result of crocodile feeding, trampling, and butchery by three different research groups. The 3D GM model evaluates the likelihood of each of these actors in the production of the controversial Dikika marks.