Potential for predator learning of artificial arboreal nest locations

We examined the potential for predators to learn the location of artificial arboreal (1.5 m above ground) nests in a managed forested landscape of central Pennsylvania from June-July 1995. We tested the hypothesis that predators do not learn the location of artificial arboreal nests placed repeatedly at the same sites (fixed nests) versus those placed at random sites in three habitats created by clearcutting (forested patches, forested corridors, contiguous forest). Sixty-nine (23%) of 299 total nests in five combined trials were disturbed by predators; 11 (16%) of these disturbances were attributed to corvids. Predation rates were greater on nests placed at random (28%) compared to fixed sites (18%, P < 0.05), indicating predators did not learn or return to the location of arboreal nests during our study. Predation rates varied significantly (P < 0.001) among habitats, with 49% of the nests disturbed in the forested-patch habitat versus only 7% and 13% in forested-corridor and contiguous-forest habitats, respectively. We propose that predation was higher in forested patches than in the other two habitats because the former had greater amounts of edge.