Chert studies, traditionally, were undertaken either through the use of visual sourcing or through Instrumental Neutron Activation Analysis (INAA). Visual sourcing has limitations, as it requires years of training to identify differences in materials. Furthermore, the appearance of source materials varies greatly across the source, making individual identifications difficult. INAA analysis, by contrast, has a high degree of accuracy, but is an expensive analysis, and often results in equivocal results. As pXRF technology becomes more common, the opportunity for museums and research institutions to use equipment to analyze existing collections could result in a much greater understanding of the artifacts already in accessioned collections.
The research funded by the NCPTT grant has begun the development of a broad chert database, to provide a mechanism for comparing artifacts to known chert sources. The development and publication on the web of this database will make possible comparisons from across the country. As more XRF instruments become available at research institutions, the database will become increasingly useful.
Analysis of the different source materials was pretty convincing. Principal component analysis of the elemental variation successfully differentiated the source materials, showing groupings of each source that overlapped, but distinct from one another.
The analysis was, however, imperfect. Because the database is in its infancy, the connections between artifacts and the source materials were flawed. Broader analysis of source materials would help ‘broaden the net’, but additional resolution in the analysis will also be necessary for the attribution of artifact to source. Material overlap between sources from vastly different locations means that there are fewer opportunities for quickly tying artifacts to source.
The collections analyzed were not able to be attributed to source. The technology is still not available to make the attribution on a consistent basis, and chert attribution will likely always be difficult. But the analysis did succeed in identifying differences among source materials. And that is a necessary first step in building the database.