Estimating Densities of Breeding Canvasbacks and Redheads

Methods of estimating breeding densities of canvasbacks (Aythya valisineria) and redheads (A. americana) were evaluated by comparing variances of estimates arising from ground counts on 12 259ha sections. Variances were similar among indexes used to indicate breeding pairs. Mainly by reason of past use and simplicity, the pair-plus-lone-male index is recommended for canvasbacks, and the totalfemale index for redheads. For isolated blocks, precision of density estimates improved with increasing pair densities, up to about 2.3 indicated pairs/km2. When sections were selected at random for extrapolation purposes, little was gained by making >3 counts. Nest counts are considered impractical for monitoring breeding pair densities because of the large, time-consuming samples needed; the difficulty in identifying renests; and, in redheads, the fact that some hens lay parasitically. J. WILDL. MANAGE. 44(4):814-821 For many years, waterfowl biologists have been estimating densities of breeding ducks from ground counts. Although refinements in technique have been made, some basic questions concerning reliability of results remain unanswered. Dzubin (1969) reviewed the art of counting ducks on small wetlands and concluded (page 225), "An urgent need exists for more testing of estimate variances and evaluating the magnitude of sampling errors and biases in pair population estimates." Sauder et al. (1971) determined the effort needed to estimate densities of dabbling ducks at various levels of precision from roadside counts in South Dakota. Our study examined variances associated with counts of canvasbacks, redheads, and their nests when the counts are used in traditional ways to derive an index of breeding-pair density. We thank our field assistants and the farmers upon whose land we worked. We appreciate the helpful manuscript reviews by D. J. Nieman, J. H. Patterson, G. E. J. Smith, and an anonymous refer-