Applied zooarchaeology provides baselines which can be used in modern conservation biology to better understand how faunal communities have changed over time. This goal can only be accomplished, however, by first accounting for the multiple biases present within the archaeological record, and how they may affect sample representativeness. Taxonomic analysis was conducted on freshwater mussel shell from the late prehistoric (ca. A.D. 700 - 1200) Kinlock site, Sunflower County, Mississippi. Species-area curves and biodiversity indices demonstrate that random sampling of surface clusters of shell, up to about 4,000 valves, provides an adequate picture of the overall surface assemblage. Comparison of surface and subsurface contexts shows a highly significant difference in species numbers and proportions, indicating a need for multi-context sampling when dealing with archaeological shell deposits.