Hiwassee Island: The Research Value and Limitations of Legacy Collections

This thesis examines the research value and limitations of WPA-era archaeological collections at the University of Tennessee’s McClung Museum of Natural History and Culture from the Hiwassee Island site (40MG31) in east Tennessee. Excavations on Hiwassee Island were conducted from 1937–1939 and uncovered a multicomponent site with Woodland, Mississippian, and historic Native American occupations. The most common artifact from all time periods was pottery, numbering more than 80,000 sherds and 70 whole vessels (Lewis and Kneberg 1946:80). This ceramic assemblage was used to determine the research significance of the Hiwassee Island legacy collection by comparing it to modern excavation samples from this site and by applying new analytical techniques in an attempt to extract new data from old collections.

Sherds were compared for size, surface decoration, and vessel area between the 1930s legacy collection and a 1997–1999 excavation assemblage to determine data limitations caused by excavation and recovery methods. Unlike modern excavations, WPA-era investigations at Hiwassee Island did not employ screening or water flotation to recover artifacts. Instead, artifacts were hand sorted with a focus on larger or decorated sherds and an emphasis on rims, appendages, and effigies; the most prominent difference was sherd size.

The ability to collect new data with old collections was tested with pilot studies in absorbed residue and portable X-ray fluorescence analyses (pXRF). Absorbed residue analysis was conducted to determine if the avenue is worth pursuing with legacy collections. Results indicated that although interpretation can be difficult, absorbed residue analysis can provide insight into vessel use for legacy collection ceramics. The most interesting result was the presence of pine resin in most of the sherds tested.

A pXRF study was conducted to determine if paste differed between ceramic types. A discriminant function analysis revealed the ceramic types clustered in three distinct groups, suggesting that at least three separate clay sources were utilized. This research demonstrates that new technology does allow for the collection of new types of data from legacy collections that supplements, supports, and aids in the interpretation of old data sets, enhancing the research potential of these collections.