Abstract
Reliable methods are needed to distinguish anthropogenic from non-anthropogenic causes of proboscidean limb bone breakage in fossil assemblages because of theoretical uncertainty about human-proboscidean relationships in the Pleistocene. This paper compares experimentally broken bones of African elephants (Loxodonta africana) and mammoths (Mammuthus spp.) after establishing that limb bone fracture dynamics are the same for those proboscidean taxa. We show that features thought exclusively diagnostic of percussive fracturing of green proboscidean long bones such as notched fracture edges, smooth fracture surfaces, and curvilinear fracture outlines also can be created on non-green bones and on bones affected by non-anthropogenic processes. The information reported here can be applied in analyses or re-analyses of fossil proboscidean bone assemblages and may either support or potentially alter current interpretations of hominin behavior.