Detection of Density‐Dependent Effects in Annual Duck Censuses

Annual censuses of duck populations from two localities differing in environmental variability were used to test the effectiveness of six tests of density dependence. The magnitude of Type I error for each test was estimated from simulated density—independent data having the same mean and variance as the observed censuses. The actual census data were used to evaluate the ability to detect density dependence. Only autoregression had the desired level of type I error, but this test was ineffective when populations grew or declined. The standard major axis was effective in the latter cases but was ineffective when populations did not grow or decline. The only effective test applicable to all cases was a regression test with a rejection region based oin simulated data. Results of these latter analyses agree with earlier descriptions of interspecific variation in life—history tactics among duck species and the importance of competition as a factor affecting guild structure in prairie—nesting ducks. Diving ducks, which occupy the most temporally stable wetlands, show a greater degree of density dependence than do dabbling ducks, and competition influences the guild structure of diving ducks more than that of dabbling ducks.