The field work for this project was carried out from 26 June to 2 August, 1995, and 8 to 18 July, 1996. Ten areas of the park were inventoried for birds. Basic weather data were collected daily so that the weather during these periods could be compared to long-term averages. For every field day, a checklist of bird species with approximate numbers was recorded. This allowed us to participate in the CWS Northwest Territories Bird Checklist Program. Breeding records were recorded only if there was a sighting of an active nest with eggs or offspring, or there was a sighting of an adult bird with flightless young (Scotter et al., 1987). During the fieldwork for this project these criteria for breeding records were strictly adhered to.

        Point counts were carried out on three biologically productive study areas within the park: Eames River, Muskox River and the central portion of the Thomsen River (see Henry and Mico, 1997, for details). A standard point count methodology for open environments (100 m radius, 5 minute observation times) was employed during this study (Ralph et al., 1995). Point counts were carried out mostly during the morning activity peak (0500 hr to 1300 hr). A total of 20 to 30 point counts were recorded for each of the eight habitat types (Ferguson, 1991) on each of these study areas.

        An arbitrary cut-off point was established in order for the most abundant species to be included in the analyses of relative abundance and habitat use patterns. Thirteen bird species met or exceeded the cut-off point which required that a species had to be recorded on five or more point counts on at least two out of the three study areas. The relative abundance data were not normally distributed; thus Friedman's Nonparametric Two Way ANOVA test was used to analyze the data (alpha = 0.05). If the null hypothesis was rejected, a multiple comparison procedure (modified Newman-Keuls test; Daniel, 1990) was utilized to determine which species had significantly different abundance values (setting the experiment-wide error at alpha = 0.10; Daniel, 1990).

        For each of the three study areas, the mean number of individual birds for all species was calculated for each habitat type. In addition, the habitat association patterns for each of the thirteen bird species were also analyzed. In each case, a Friedman's Two Way Nonparametric ANOVA test was used to analyze the data. If the null hypothesis is rejected, the modified Newman-Keuls test was used to make multiple comparisons. As above, for these multiple comparisons, alpha was set at 0.10 (Daniel, 1990).

        In order to study if hummock tundra provides shelter for small birds during windstorms, wind speed was measured with a hand-held anemometer on moderately windy days (15-20 km/h) at several different heights above ground level: 150 cm, 100 cm, 50 cm, 20 cm, 10 cm, and ground level. It was also measured at the bottom of micro depressions located between the hummocks. Average wind speed measurements were normally distributed and were tested for significant differences using a single factor ANOVA analysis (Sokal and Rohlf, 1969).

        To test whether coulees were selectively used by large birds (and large mammals) as compared to areas of surrounding tundra, paired 100 m transects were located with one transect randomly placed in the bottom of a sheltered coulee and the other transect randomly placed on nearby flat, open tundra. On each transect, all scat piles that occurred on the transect or within 1 m of either side of it were recorded. Five pairs of transects were carried out in each of three different sheltered coulees. Paired sample t-tests were used to analyze for significant differences (Sokal and Rohlf, 1969).