Nesting success is an important vital rate affecting the reproductive fitness of birds, and predation typically is the single most important factor affecting nesting success. Presumably, birds should nest in locations that maximize nest survival. If specific nest characteristics increase the probability that a nest will hatch, natural (phenotypic) selection could favour use of sites with these features, producing nonrandom patterns of nest site use. Alternatively, birds that are highly selective in nest site choices might be at a disadvantage if predators learn to forage preferentially in these locations and improve their efficiency in depredating nests; in this case, random nesting patterns could be favoured. Finally, it has been hypothesized that predation pressure can influence nest site selection patterns of entire bird communities. If predators develop a search image to hunt for bird nests, then nests that are most similar to each other, irrespective of species, should sustain higher mortality. To evaluate these hypotheses, I quantified nest site selection patterns of multiple species of ground-nesting dabbling ducks in areas where predation pressure was normally high, and compared these patterns to those on areas where predation was relaxed. Predation pressure was experimentally reduced by removing common predators of duck nests and females (mainly red foxes, coyotes, skunks and raccoons) on some study areas and not on others (controls). Predator removal and natural causes produced a 10-fold difference in duck nesting across study sites, allowing for investigation of effects of predation pressure on nest site selection of ducks. Coarse scale habitat selection patterns were similar to results reported in previous studies; blue-winged teal and northern shoveler were found more often in native grassland than in other habitat types, while gadwall and mallard nests occurred more frequently in shrub patches when compared with other habitat patches. A difference in nest site characteristics was observed between hatched and depredated nests for gadwall and northern shoveler but not for blue-winged teal and mallard. However, in all species, the nest site selection patterns were non-random. Thus, the process of nest predation did not shape patterns of nest site choice. Contrary to predictions, inter-specific overlap in nest site features was not related to predation pressure: nests that overlapped most with features of other species did not suffer higher predation, nor did inter-specific overlap in nest characteristics decrease during the nesting season. These findings were inconsistent with the hypothesis that community-level patterns of nest site use are differentiated as a result of predation pressure. Long-term work on nest site use by individually marked females of numerous ground-nesting bird species would be informative, as would experimental studies of other hypotheses about factors affecting nest site choices in birds.

Nesting success is an important vital rate affecting the reproductive fitness of birds, and predation typically is the single most important factor affecting nesting success. Presumably, birds should nest in locations that maximize nest survival. If specific nest characteristics increase the probability that a nest will hatch, natural (phenotypic) selection could favour use of sites with these features, producing nonrandom patterns of nest site use. Alternatively, birds that are highly selective in nest site choices might be at a disadvantage if predators learn to forage preferentially in these locations and improve their efficiency in depredating nests; in this case, random nesting patterns could be favoured. Finally, it has been hypothesized that predation pressure can influence nest site selection patterns of entire bird communities. If predators develop a search image to hunt for bird nests, then nests that are most similar to each other, irrespective of species, should sustain higher mortality. To evaluate these hypotheses, I quantified nest site selection patterns of multiple species of ground-nesting dabbling ducks in areas where predation pressure was normally high, and compared these patterns to those on areas where predation was relaxed. Predation pressure was experimentally reduced by removing common predators of duck nests and females (mainly red foxes, coyotes, skunks and raccoons) on some study areas and not on others (controls). Predator removal and natural causes produced a 10-fold difference in duck nesting across study sites, allowing for investigation of effects of predation pressure on nest site selection of ducks. Coarse scale habitat selection patterns were similar to results reported in previous studies; blue-winged teal and northern shoveler were found more often in native grassland than in other habitat types, while gadwall and mallard nests occurred more frequently in shrub patches when compared with other habitat patches. A difference in nest site characteristics was observed between hatched and depredated nests for gadwall and northern shoveler but not for blue-winged teal and mallard. However, in all species, the nest site selection patterns were non-random. Thus, the process of nest predation did not shape patterns of nest site choice. Contrary to predictions, inter-specific overlap in nest site features was not related to predation pressure: nests that overlapped most with features of other species did not suffer higher predation, nor did inter-specific overlap in nest characteristics decrease during the nesting season. These findings were inconsistent with the hypothesis that community-level patterns of nest site use are differentiated as a result of predation pressure. Long-term work on nest site use by individually marked females of numerous ground-nesting bird species would be informative, as would experimental studies of other hypotheses about factors affecting nest site choices in birds.