EDGE EFFECTS ON NEST PREDATION IN THE SHAWNEE-NATIONAL-FOREST, SOUTHERN ILLINOIS

Edge habitats may be considered 'ecological traps' for breeding birds if they attract many birds because of apparently favorable nesting con ditions but have higher nest predation levels than interior habitats, Four alternative, nonexclusive hypotheses have been suggested to expla in why edges might have higher predation levels than interior habitats : (1) predator activity is higher in areas with higher prey density (d ensity-dependent predation), (2) predators are more abundant on edges than in forest interior, (3) the predator community is richer in speci es on edges than in forest interior; and (4) predators forage along tr avel lanes (linear geographical features) such as edges. Here we evalu ated whether forest-farm edges in southern Illinois ave ecological tra ps, and examined the relevance of these four hypotheses at our study s ire with several different experiments during May-July 1992 using arti ficial nests (n = 605) bailed with quail eggs and placed on the ground , in shrubs, or in saplings. Our results showed that, in general, the forest-farm edges of southwestern Illinois did not attract significant ly more individuals or species of nesting songbirds, but they did have higher nest predation levels than forest interior sites, primarily as a result of higher predation levels on sapling nests. We did not find evidence strongly supporting any of the four hypotheses suggested as explanations for higher nest predation levels near edges. Two data set s showed that predation levels on artificial nests were density-indepe ndent. Forest-farm edges had neither more total species of potential n est predators nor more individual predators. However, there were more species of avian predators on edges than in interior sites. Correlatio ns between predator abundance and nest predation levels on individual transects were weak. The travel lane hypothesis was not supported beca use nest predation levels were either not affected by distance from li near geographical features (roads and ravines) or were significantly l ess when close to than when far from presumptive travel lanes. High sp atial heterogeneity in predation levels, numbers of singing birds, and potential nest predators may have obscured general patterns and sugge st a need for larger sample sizes. Edges may be detrimental to some sp ecies of singing birds but nor to others, and for different reasons.