Years of over-fishing combined with increased nutrient pollution have had a catastrophic effect on the ecology of the Chesapeake Bay. The Holocene record of bay mollusks may provide a baseline for ecological restoration, but the effects of taphonomic bias on these assemblages first must be assessed. In this study, a live-dead comparison was carried out on four sites distributed in the main channel of the upper bay. Molluscan death-assemblage data were obtained from replicate box-core samples from which whole specimens and fragments were sorted, identified, and counted. Data on live communities at the same sites, sampled over the past twenty years, were provided by the Chesapeake Bay Program, making it possible to examine the degree to which death assemblages reflect long-term changes in the live community. Traditional live-dead metrics document a strong agreement between live-community and death-assemblage estimates of species composition, richness, and abundance—77% of the species in the live community are found in the death assemblage, and 99% of the individuals of species found in the death assemblage are found in the live community. Correlations between live and dead estimates of species richness are not statistically significant, although they do improve with longer-term sampling of the live community. Rank abundance of taxa in the death assemblage is correlated strongly and significantly with live rank abundance regardless of the duration of live sampling. These results suggest that Holocene molluscan assemblages may provide useful estimates of richness and abundance for Chesapeake Bay restoration.