NONREPRESENTATIVE SAMPLING DURING WATERFOWL BANDING - EMPHASIS ON BODY CONDITION

We trapped mallards (Anas platyrhynchos) at 3 sites on the eastern sho re of Lake St. Clair, Ontario, during August and September of 1989 and 1990 to investigate the potential for nonrepresentative sampling in t ypical autumn (i.e., pre-hunting season) duck-banding operations. Inte rsite comparisons showed that the local population was spatially segre gated with respect to sex, age, molt chronology, and body condition, i ndicating a potential danger when the spatial distribution of the band ing effort is limited. Sex composition of trapped samples varied accor ding to size of the daily catch; smaller catches contained proportiona lly more females. However, size of the catch had no apparent influence on age ratios. Analyses conducted separately by sex and age provided little evidence for a relationship between size of the catch and condi tion of captured individuals. These results are generally inconsistent with the hypothesis that food-stressed individuals are among the firs t to respond to baited areas and enter traps. Thus, condition-related trap selectivity is unlikely to vary with size of the trapped sample, as some authors have suggested. Additional research is needed to deter mine whether bait-trapped samples of ducks are generally skewed toward individuals in poor condition.