Abstract
Few attempts have been made to experimentally address the extent to which temporal or spatial variation in food availability influences avian habitat use. We used an experimental approach to investigate whether bird use differed between treated (arthropods reduced through insecticide application) and control (untreated) forest canopy gaps within a bottomland hardwood forest in the Upper Coastal Plain of South Carolina, USA. Gaps were two‐ to three‐year‐old group selection timber harvest openings of three sizes (0.13, 0.26, and 0.50 ha). Our study was conducted during four bird use periods (spring migration, breeding, post‐breeding, and fall migration) in 2002 and 2003. Arthropods were reduced in treated gaps by 68% in 2002 and 73% in 2003. We used mist‐netting captures and foraging attack rates to assess the influence of arthropod abundance on avian habitat use. Evidence that birds responded to arthropod abundance was limited and inconsistent. In 2002, we generally captured more birds in treated gaps of the smallest size (0.13 ha) and fewer birds in treated gaps of the larger sizes. In 2003, we recorded few differences in the number of captures in treated and control gaps. Foraging attack rates generally were lower in treated than in control gaps, indicating that birds were able to adapt to the reduced food availability and remain in treated gaps. We conclude that arthropod abundance was not a proximate factor controlling whether forest birds used our gaps. The abundance of food resources may not be as important in determining avian habitat selection as previous research has indicated, at least for passerines in temperate subtropical regions.