What Kind of Data Are in the Back Dirt? An Experiment on the Influence of Screen Size on Optimal Data Recovery

Abstract

Zooarchaeological assemblages are non-renewable resources that provide a record of human culture and the paleoenvironment. Optimal sampling of zooarchaeological remains forms a foundation from which questions about human ecology are asked and answered. It has been known for over three decades that fine-gauge screen sieving (1/l6 inch = 1.58 mm) has an influence on the recovery of biological remains. However, archaeologists are not always fully aware of what types of faunal data may be lost into the back dirt, or the impact of those losses on interpreting the zooarchaeological record when fine-gauge screens are not used. As zooarchaeology has evolved to include studies of the paleoenvironment, sampling concerns have become even more important. This paper presents the analysis of the results of an experiment on the recovery of 10 zooarchaeological assemblages from coastal Alabama; Georgia; Florida; St. John, U.S. Virgin Islands; and Puerto Rico. When fine-gauge screen sampling protocol is used, large and small taxa have an equal chance of being recovered, and there is an exponential increase in the numbers of fish taxa and individuals recovered. Measurements of the lateral width of fish vertebrae in the 10 assemblages show that over 80% of the specimens are smaller than 6 mm and potentially would have been lost through coarse-gauge (1/4 inch = 6.35 mm) screen. The inclusion of remains recovered through finer meshes of screen significantly changes biomass calculations. Furthermore, descriptive statistics for sample diversity, equitability, trophic level, and similarity indices are affected. Fine-gauge screen recovers an assemblage of faunal remains that more closely resembles the midden population than do coarse-screened samples; therefore the fine-screened samples more accurately reflect the relative percentage of taxa and the size classes of the represented animals. The statistical data show that faunal assemblages recovered with different screen gauges are not directly comparable.