Predators can influence prey abundance and traits by direct consumption, as well as by non-consumptive effects of visual, olfactory, or tactile cues. The strength of these non-consumptive effects (NCEs) can be influenced by a variety of factors, including predator foraging mode, temporal variation in predator cues, and the density of competing prey. Testing the relative importance of these factors for determining NCEs is critical to our understanding of predator-prey interactions in a variety of settings. We addressed this knowledge gap by conducting two mesocosm experiments in a tri-trophic intertidal oyster reef food web. More specifically, we tested how a predatory fish (hardhead catfish, Ariopsis felis) directly influenced their prey (mud crabs, Panopeus spp.) and indirectly affected basal resources (juvenile oysters, Crassostrea virginica), as well as whether these direct and indirect effects changed across a density gradient of competing prey. Per capita crab foraging rates were inversely influenced by crab density, but they were not affected by water-borne predator cues. As a result, direct consumptive effects on prey foraging rates were stronger than non-consumptive effects. In contrast, predator cue and crab density interactively influenced indirect predator effects on oyster mortality in two experiments, with trait-mediated and density-mediated effects of similar magnitude operating to enhance oyster abundance. Consistent differences between a variable predator cue environment and other predator cue treatments (no cue and constant cue) suggests that an understanding of the natural risk environment experienced by prey is critical to testing and interpreting trait-mediated indirect interactions. Further, the prey response to the risk environment may be highly dependent on prey density, particularly in prey populations with strong intra-specific interactions.