Comparative methods for distinguishing flakes from geofacts: a case study from the Wenas Creek Mammoth site

Archaeologists have long struggled with distinguishing lithic materials modified by humans (artifacts) from natural objects (e.g., geofacts or zoofacts). This problem is especially difficult for finds of small numbers of flake-like lithic specimens, and particularly for very old finds. We attempt to address the artifact versus geofact problem at a paleontology site by employing three systematic and objective tests on the two recovered possible artifacts. First, they are compared with debitage attributes typically expected of artifacts and geofacts based on published experimental and actualistic data. Second, they are compared in terms of nine of these attributes with a toolstone sample from the site excavation matrix. Third, the two possible artifacts are scored for these nine attributes and graphed against the toolstone matrix sample and two samples of flintknapped debitage assemblages. In all three comparisons, the two specimens are more like artifacts than geofacts. While this does not prove the specimens are artifacts, it at least shows they cannot be easily dismissed as the sort of geofacts typically expected in the site matrix. We argue that this distinction is an important first step in the evaluation of possible lithic artifacts.