Predators commonly structure natural communities, but predation effects can vary greatly. For example, increasing predator densities may not reduce prey populations as expected if intraspecific predator interactions suppress foraging efficiency or if prey size refuges exist. In northeastern Florida (USA), outbreaks of the predatory crown conch Melongena corona have contributed to declines in oyster populations and the commercial oyster fishery. However, despite expectations of oyster population collapse, reefs have persisted, albeit with reduced adult oyster size and living reef biomass. To investigate the mechanism(s) underlying this unexpected persistence, we used field observations and experiments to examine the effects of predator density and prey size on predation rates. Multi-year surveys indicated that large oysters did not experience a predation size refuge, and further suggested that predation rates declined with increased predator density. Consistent with field surveys, field experiments demonstrated that conchs selectively consumed larger oysters (potentially explaining the absence of large oysters on natural reefs) and that high conch densities suppressed per capita predation rates, likely due to intraspecific antagonistic interactions. A Type III ratio-dependent model best described the experimental conch functional response, explaining >50% of the variation in per capita prey consumption and including a signal of reduced attack rates at high predator densities. Thus, although large aggregations of predators have the potential to deplete prey populations, our study illustrates intraspecific predator interactions that possibly prevent the local extirpation of an important habitat-forming prey species.