CHANGES IN THE WINTER DISTRIBUTION OF THE ROUGH-LEGGED HAWK 
We chose Red-tailed Hawk abundance to test the potential effect of in- 
terspecific competition on the Rough-legged Hawk. The Red-tailed Hawk 
is a documented interspecific competitor (Schnell 1968, Bildstein 1987) 
and is, by far, the most abundant raptor across nearly all the winter range 
of the Rough-legged Hawk (Root 1988). We determined Red-tailed Hawk 
abundance by using data from the same CBC circles used for the Rough- 
legged Hawk's abundance. 
Statistical Analysis 
We screened the data to check assumptions related to normality and het- 
erogeneity by using probability plots in Systat Version 9 (SPSS, Chicago). 
We also screened the continuous co variates (average December temperature, 
December days with snow cover >5 cm, and numbers of Red-tailed Hawks 
per party hour) for multicollinearity by using PROC REG in SAS 9.1 (SAS 
Institute 2005) and removed any variables with tolerance values <0.40 
(Cody and Smith 2006). 
We used a paired t test to determine the statistical significance of changes in 
Rough-legged Hawk abundance between the two periods (first 14 years ver- 
sus second 14 years) for each of the ten regions. We applied the Bonferroni 
method to yield a threshold of significance of <0.005 for these analyses. 
We used logistic regression to analyze the effects of single predictor vari- 
ables on Rough-legged Hawk numbers. We used the change in Rough-legged 
Hawks per party hour from the first 14 years to the second 14 years as the 
binary response variable. That is, if the average number of Rough-legged 
Hawks per party hour over the second 14 years was greater than the average 
over the first 14 years, that CBC circle was assigned a value of one. If the 
average over the second 14 years was less than over the first 14 years, that 
CBC circle was assigned a value of zero. We performed logistic regression 
versus the following four predictor variables: 
• The change in December average temperature between the two pe- 
riods, 
• The change in the average number of December days with snow cover 
>5 cm, 
• County population increase (percent) from 1980 to 2000, and 
• The number of Red-tailed Hawks per party hour over the entire 28-year 
period (count years 79-106). 
Because climate or population data were not available for some CBC circles, 
not every circle was analyzed for each variable. Also, in evaluating the effects 
of snow cover, we used only CBCs that averaged at least five December days 
with snow cover >5 cm during one of the two periods. As a result, we used 
259 CBC circles for analysis of the change in December temperature, 199 
for analysis of December snow cover, 245 for analysis of change in human 
population, and 293 for analysis of Red-tailed Hawks per party hour. 
RESULTS 
Figure 1 shows all 293 CBCs used. Numbers of the Rough-legged Hawk 
on nearly all (101/112, 91%) of the CBCs on the coasts and in the three 
southern regions decreased. In all but seven of the 43 CBCs in the North 
213 
