The elemental analysis of clays has become a standard method for sourcing archaeological
ceramics. Problems encountered include the wide natural variability in clays and the contribution
of inclusions to the elemental signature. With shell-tempered ceramics, this latter problem has
been dealt with by removing the shell with acids prior to analysis or by mathematically correcting
for the effects of shell. Ironically, both procedures discard data that pertain directly to the topic of
interest: the elemental signature of the shell itself, which theoretically can be used to
discriminate sources of temper material and, by extension, of the pottery itself.
Freshwater mussel shells have chemical signatures related to the waterway of origin. If
elemental variation between waterways is greater than that between species or shell parts, then
analyzing shell temper particles should allow for pottery to be sourced to the drainage where it
LA-ICP-MS was used to analyze shell-temper particles in pottery from the Lyon's Bluff site
(22OK520), a Mississippian to Protohistoric period mound site in north-east Mississippi.
Several plain sherds were used to represent local wares; possible imports analyzed included
sherds of Moundville Engraved, Nodena Red-on-White, negative painted sherds, and a
fragment of a "cat monster" figurine. Each of these might be considered non-local based on
Principal Components plot based on the elemental analysis of archaeological
mussel shells using a PE-Optima 4300 DV inductively coupled-plasma emission
spectrometer. Elements used in analysis are Ca, K, Mg, Mn, Na, and Sr. Note
the separation between drainages.
Clear grouping is displayed when different elements are compared on bivariate plots.
The plainwares constitute a presumably local group, while most of the "exotic"
ceramics fall together in one or more groups.
Interestingly, the negative painted sherds- arguably the most likely imports - consistently
fall in with the plain sherds. This may be due to geological similarities between the Black Prairie,
with a chalk bedrock, and the Nashville Basin, with a limestone bedrock. Alternatively, the
negative-painted sherds may in fact be of local manufacture. LA-ICP-MS provides a rapid method
for further exploring these possibilities, as large numbers of samples can be analyzed rapidly and a
wide array of elements examined for purposes of discrimination. Large amounts of archaeological
shell are available to provide background data. This pilot study demonstrates the potential for using
shell as a sourcing agent in ceramic studies.