Summary
We examined trap bias relative to body condition, age, and sex using data collected from Brown-headed Cowbirds (Molothrus ater) decoy trapped during the 1987 and 1988 breeding seasons. A condition bias was evident, as decoy-trapped birds were in poor condition relative to those caught by mist net in an adjacent roost. This occurred in both sexes and independently of study year and male age. Paradoxically, however, males in better condition when first trapped were more likely to be recaptured in the same season, possibly reflecting more widespread movement, and hence fewer encounters with traps, by birds in poor condition. Hatching-year birds were more likely than adults to be recaptured in the same season. Trapped samples were strongly male biased, which probably reflects a difference in nutritional requirements, and thus foraging strategies, between males and females during the breeding season. This study indicates that biases among birds trapped by methods that rely on a feeding response are likely to be widespread and difficult to predict, suggesting that interpretation of data from such samples always requires caution.
Methodology
We conducted the study in 1987 and 1988 at the Queen's University Biological Station in eastern Ontario. We used three decoy traps (woodland, intermediate, and farmland), located within 6 km of each other. The traps were similar to those used by Weatherhead and Greenwood (1981), the design of which was based on the "New York starling trap" (U.S. Department of the Interior, unpublished report, 1962). All traps were in open areas with no trees within 5 m. However, the habitat surrounding each trap differed substantially among locations. Using aerial photographs, we estimated that the area in a 200-m radius was 90% forest around the woodland trap, 55% forest around the intermediate trap, and 30% forest around the farmland trap. The forest was predominantly mixed deciduous, and the remaining habitat was a mix of old field, hayfield, and mowed lawn. The significance of the habitat differences is that most cowbird hosts in the study area nest in forest and forest edge habitat (Teather 1983; P. J. Weatherhead, personal observation). Thus, the woodland trap was in cowbird breeding habitat, the farmland trap was in predominantly feeding habitat, and the intermediate trap was in a mix of the two.
We operated traps from 14 April to 3 1 July in 1987 and from 4 April to 2 August in 1988. Comprehensive analyses of cowbird migration patterns are lacking. However, Weatherhead et al. (1980) used bandrecovery data to show that cowbirds encountered in the St. Lawrence Valley (north of our study area) during April generally remained in that area through the summer. Thus, most cowbirds encountered in our study area likely had completed spring migration. Traps were emptied daily between 16:30 and 17:30. Two male cowbirds, one female cowbird, and usually one male Red-winged Blackbird (for another study) were left as decoys. Water and cracked corn were always available in the traps. Upon initial capture, each cowbird was banded with a U.S. Fish and Wildlife Service aluminum leg band. Birds already banded when first encountered in our study (n = 25) were excluded from analyses to avoid any effects of possible prior exposure to traps. Adult (after hatching year) males were classified as SY (second calendar year) or ASY (after second year), based on plumage (Selander and Giller 1960). In the case of adult females, age-classes could not be distinguished. HY (hatching year) birds were easily distinguished from adults but were not sexed. For a random subsample of the daily catch, body mass (k0.5 g) and wing length (+ 1 mm) were also recorded.
We used body mass, corrected for variation in structural size (i.e., wing length), as an index of condition. Body mass is generally correlated with lipid reserves in birds (see Blem 1990 for a review), and this correlation is generally improved by controlling variation in body mass associated with structural size differences (Bailey 1979; Johnson et al. 1985; Ringelman and Szymczak 1985; see also Alisauskas and Ankney 1987).