ASSESSING SCLEROCHRONOLOGICAL ESTIMATES OF SEASON-OF-CAPTURE FROM ARCHEOLOGICAL SHELL MIDDENS: MODELS AND OBSERVATIONS

Archeological middens—ancient rubbish accumulations—are valuable sources of information about prehistoric humans and their interactions with the environment. Bivalve mollusk shells, which are abundant faunal constituents of shell middens, can be used to establish season-of-capture (SC). In turn, SC can be used to infer the timing of annual migrations and resource utilization. Because bivalves grow by accretion, their shells record environmental conditions in the form of geochemical variation. Analysis of δ¹⁸O variation from the oldest ontogenetic portion of the shell can be used to determine when the clams were harvested. These archives, however, are complicated by biologically controlled variations in growth rate and periodic growth cessations. These phenomena may combine to obfuscate the true SC.

To evaluate the magnitude of this bias, we modeled the growth of bivalves using a variety temperature models, and intra- and interannual growth models. Next, we sampled each modeled profile using one of two approaches. First, we sampled the entire profile with using contiguous samples. This process was repeated with progressively large distances between samples. Second, we sampled profiles stochastically within predefined window widths. This procedure was subsequently repeated with progressively wider windows. These two sampling procedures resulted in numerous hypothetical isotope profiles, each of which was time-averaged to a different degree. To determine SC, we used the last three δ¹⁸O samples from each profile. SC was determined using the following criteria: spring = negative slope; summer = V-shaped, autumn = positive slope; winter = inverted V-shape. Our results indicate that as age increases and sampling resolution decreases, the predicted SC is more likely to be incorrectly assigned, suggesting that sampling young ontogenetic specimens with high temporal resolution is likely to provide the best estimates of SC. Sampling of 25 Mercenaria mercenaria specimens, all harvested on the same day, supports this finding; accurate season of capture estimates were obtained from the youngest/smallest specimens. Our results are significant because ancient humans may collect larger specimens potentially biasing SC estimates from shell middens.