Effects of Habitat Characteristics on Gadwall Nest Predation and Nest-Site Selection

We compared characteristics of successful gadwall (Anas strepera) nests and those destroyed by mammalian predators (i.e., striped skunks [Mephitis mephitis]). Lateral cover density, understory cover height, species richness, vegetative penetrability, and patch size were significant determinants of the fate of a nest. Nest success was also influenced by 3 nonvegetative variables: minimum distance to water, dike width, and nest initiation date. Predation rates on nests differed (P 1.0 m), dense biennial and perennial weeds such as thistle (Cirsium arvense), stinging nettle (Urtica dioica), and teasel (Dipsacus spp.), and lacked well-developed understory (ground vegetation). The proportional occurrences of habitats A, B, and C were 33, 42, and 25%, respectively. Beginning in late May, we located nests by flushing females with a slow moving truck and occasional bursts of the truck's horn. This procedure was effective because all nests occurred within 10 m of the dike road. Density of waterfowl nests on the study area was determined by a complete enumeration verified by intensive ground searches in each area. Data recorded for each nest included species, number of eggs, stage of development (Weller 1956), and habitat. Pin flags were placed perpendicular to each nest at the edge of the road. One hundred seventy artificial nests were placed at random distances perpendicular to the dike road at regular intervals along the dike. They consisted of 2 chicken eggs the size of gadwall eggs placed in a shallow, excavated bowl. The same procedures employed for natural nests were used for artificial nests. Because predation rates for natural nests This content downloaded from 157.55.39.186 on Tue, 12 Apr 2016 09:03:55 UTC All use subject to http://about.jstor.org/terms J. Wildl. Manage. 53(1):1989 HABITAT NEST PREDATION * Crabtree et al. 131