Northern Prairie Wildlife Research Center
Thirty plots were selected for Breeding Bird Censuses during the initial year of the study. Site selection involved choosing three draws in each of 11 areas (Fig. 2). This selection included draws set at different aspects and with varying slope and grazing intensities. Although attempts were made to select draws exhibiting a wide array of tree species, the geographic location of the study site appeared to preclude much variation in species composition.
Avian Census Methods
Breeding bird populations were censused June 2-June 30, 1982, by using a modified Emlen line transect (Emlen 1971, 1977). Each of the 30 draws finally selected for censusing was visited three times during the count period. Count times were usually from local sunrise to 1100. During each count, data were recorded on the location of each male bird of a species encountered (except that only female Brown-headed Cowbirds were counted).
Each census plot was 50 x 400 m, covering an area of two ha. In all plots but one, this size was maintained. The centerline and edges of each plot were marked with colored plastic surveyor's tape.
In addition to census plots, periodic investigations of habitats adjacent to the wooded draws were also made. These data were used to supplement observations made in each area, and to provide further information on the numbers and kinds of birds using these habitats.
The method used for this investigation was a modification of the James and Shugart (1970) circular plot method, as described by Noon ( 1981). Vegetation within each draw was sampled in five 0.04-ha circular plots, which provided an analysis of vegetation on 0.2 ha of each bird census plot. These plots were at 0, 50, 150, 200, and 350 m, along each bird census transect.