Although density-edge effects are commonly reported, we have only scratched the surface in understanding the mechanisms underlying how habitat edges mediate species interactions. Here, I use a combination of field surveys and experiments to explore the linkages between the presence of host-plant edges (Scolochloa festucacea), the within-patch distribution of planthopper eggs (Delphacodes scolochloa), and the oviposition behavior of their egg parasitoids (Anagrus spp.). The field surveys revealed that densities of hosts and parasitoids at the edge were > or = 2.5 times lower than densities in the patch interior and that the effect was independent of patch size. As a consequence of its edge-avoidance behavior, host within-patch dispersion was significantly more aggregated in the presence, as compared to the absence, of an edge. Also, as patch size decreased, the proportion of the patch that was core (i.e., beyond the influence of the edge) decreased, and the degree of host aggregation increased. In a subsequent field experiment, I found that the dispersion of hosts affected Anagrus oviposition behavior only when hosts were present on islands with discrete edges. Under these circumstances, the proportion parasitized and per capita parasitized were 2.3 and 3.4 times higher, respectively, when hosts were clumped as compared to when hosts were uniformly or randomly distributed. Based on a laboratory experiment using small S. festucacea patches, I found that Anagrus had 40% shorter step lengths, spent 52% more time in the patch, and parasitized 84% more hosts when hosts were clumped as compared to uniformly distributed. These results were indicative of area-restricted search by Anagrus, which is an effective foraging strategy when hosts are clumped. This is the first study to demonstrate that predator foraging behavior in response to prey dispersion can be mediated by the presence of a patch edge. Also, because edge-averse behavior is commonly reported in the literature, an underappreciated effect of fragmentation on predator-prey interactions and stability could arise from edge-mediated effects on prey within-patch aggregation.