Models that describe the dispersion patterns of predators between a series of patches that vary in prey density frequently assume that predators, in the absence of interference, will aggregate in patches with the highest prey density, at any point in time. This assumption has important implications for patterns of prey mortality, and the extent to which prey mortality is density dependent. In natural predator-prey systems, it is likely that environmental factors interact with spatial variation in prey density to influence the aggregative response of predators. We show data consistent with this idea on a population of overwintering oystercatchers foraging on cockles. There was no evidence that birds aggregated in patches with the highest biomass density of cockles. The biomass density of cockles was highest in muddy patches at the start of winter, and birds aggregated in patches that switched from being muddy at the start of winter to being sandy at some point during the winter. We argue that sediment type influences foraging costs experienced by the birds, so birds avoid feeding in muddy patches unless the fine sediment is removed from a patch, as happens during winter storms. When this happens a high biomass density of cockles suddenly becomes available and the birds aggregate in such patches. The rate of biomass loss was greatest in patches used intensively by birds for feeding, suggesting that the birds’ aggregative response influences cockle mortality. We discuss the implications of our results for ideal free models.