Factors Associated with Duck Nest Success
in the Prairie Pothole Region of Canada
Adult Mortality
We compared species composition of ducks found dead with species composition of ducks in the breeding population using a chi-squared goodness-of-fit test. Comparisons were done separately for each year, 1983-85. For the 5 common species, we used an r-sample binomial test (Bain and Engelhardt 1987) to examine the null hypothesis that male and female ducks were killed in proportion to their abundance in the breeding population. We assumed that breeding populations contained equal proportions of males and females, although continental populations of these species are thought to contain more males than females (Bellrose 1980). We first simultaneously tested the null hypothesis for all years for each species to determine overall significance (P = 0.05). Significant results for a species were then examined by individual year to isolate the source of significance. For each species, we excluded years with <10 specimens identified to sex, because distribution of the test statistic is unknown for small samples.
Previous Section -- Seasonal Variation in Nest Predation Rate
Return to Contents
Next Section -- Components of Mallard Production
