Although only a few studies have explicitly evaluated live-dead agreement of species and community responses to environmental and spatial gradients, paleoecological analyses implicitly assume that death assemblages capture these gradients accurately. We use nine data sets from modern, relatively undisturbed coastal study areas to evaluate how the response of living molluscan assemblages to environmental gradients (water depth and seafloor type; “environmental component” of a gradient) and geographic separation (“spatial component”) is captured by their death assemblages. We find that:

1. Living assemblages vary in composition either in response to environmental gradients alone (consistent with a species-sorting model) or in response to a combination of environmental and spatial gradients (mass-effect model). None of the living assemblages support the neutral model (or the patch-dynamic model), in which variation in species abundance is related to the spatial configuration of stations alone. These findings also support assumptions that mollusk species consistently differ in responses to environmental gradients, and suggest that in the absence of postmortem bias, environmental gradients might be accurately captured by variation in species composition among death assemblages. Death assemblages do in fact respond uniquely to environmental gradients, and show a stronger response when abundances are square-root transformed to downplay the impact of numerically abundant species and increase the effect of rare species.

2. Species' niche positions (position of maximum abundance) along bathymetric and sedimentary gradients in death assemblages show significantly positive rank correlations to species positions in living assemblages in seven of nine data sets (both square-root-transformed and presence-absence data).

3. The proportion of compositional variation explained by environmental gradients in death assemblages is similar to that of counterpart living assemblages. Death assemblages thus show the same ability to capture environmental gradients as do living assemblages. In some instances compositional dissimilarities in death assemblages show higher rank correlation with spatial distances than with environmental gradients, but spatial structure in community composition is mainly driven by spatially structured environmental gradients.

4. Death assemblages correctly identify the dominance of niche metacommunity models in mollusk communities, as revealed by counterpart living assemblages. This analysis of the environmental resolution of death assemblages thus supports fine-scale niche and paleoenvironmental analyses using molluscan fossil records. In spite of taphonomic processes and time-averaging effects that modify community composition, death assemblages largely capture the response of living communities to environmental gradients, partly because of redundancy in community structure that is inherently associated with multispecies assemblages. The molluscan data sets show some degree of redundancy as evidenced by the presence of at least two mutually exclusive subsets of species that replicate the community structure, and simple simulations show that between-sample relationships can be preserved and remain significant even when a large proportion of species is randomly removed from data sets.