Lithified marine sediments are not equitably distributed through time, raising the possibility that lithification masks biological signals when data from unlithified and lithified sediments are compared or combined. Using mollusk-dominated assemblages from the early Cenozoic of the Gulf Coastal Plain, we find that lithification conceals small taxa, decreases taxonomic resolution, and exacerbates the undersampling of rare taxa. Lithified assemblages appear less diverse and have less even abundance distributions than coeval unlithified samples. These limitations cannot be overcome by standardization procedures, nor are they likely to be circumvented by collecting larger samples. The effects of this bias, however, can be mitigated by restricting analyses to a single lithification state or to specific size classes. Since lithification selectively obscures small taxa, the magnitude of this bias will be most severe when organisms are particularly small, such as in the aftermath of mass extinctions. In the study area, lithification artificially protracts the recovery period following the Cretaceous-Paleogene mass extinction by ~7 m.y.