several other methods of quantifying habitat selection, 

 assumed that all vegetation cover types were equally available 

 to all deer on the study area. That assumption was not met. 

 For example, the Douglas fir types were distributed primarily 

 on the western 60% of the area, where all other types also 

 occurred. Deer on the eastern portion did not have the option 

 of using fir, but had to choose among pine-juniper and other 

 apparently less selected or preferred types. Because of this, 

 the fir types probably were even more preferred and selected 

 for than indicated by our analysis. 



Distributional biases also were apparent in the 

 relatively low selection indicated for other types, especially 

 the river riparian and sagebrush-grassland types. The former 

 was usually available only to deer ranging within about 3 km 

 of the northern boundary of the area and could be entered only 

 from one direction. Although our analysis indicated the river 

 riparian type was used less than expected based on 

 availability, general observations and movements of marked 

 deer suggested it received relatively heavy use at least 

 during autumn, by deer whose home ranges included some river 

 bottomland. 



Our analysis also indicated that the sagebrush-grassland 

 type was not preferred during any season, whereas Mackie 

 (1970) showed heavy and apparently preferential use of the 

 type, especially during winter and spring. This type covered 

 more of the study area than any other vegetation type. It 

 occurred throughout, interspersed among forested and other 

 types and as large vegetationally homogeneous blocks on major 

 ridgetops and along the southern and southwestern fringes . 

 Lacking diversity, such large blocks were not attractive to 

 and received little use by deer. Further analysis showed that 

 more deer than expected were observed in blocks containing 1% 

 to 49% aerial coverage of sagebrush-grassland, and fewer than 

 expected (P < 0.0001) occurred only in blocks containing more 

 than 75% coverage (Fig. 9.2). This indicated that, although 

 large blocks of the type were avoided, small patches 

 interspersed with forested types were important to deer. This 

 was especially true during spring green-up when general 

 observations and data from earlier studies (Mackie 1970) 

 indicated that a majority of the deer observed feeding were in 

 sagebrush-grasslands . 



In reporting earlier studies on habitat use by mule deer 

 on the study area, Mackie (1970) noted that observability bias 

 probably resulted in underestimates of deer use and the 

 relative importance of the timbered pine-juniper and Douglas 

 fir-juniper vegetation types as compared with open types such 

 as sagebrush-grassland. We compared relative use of major 

 vegetation types as determined by observations along vehicle 

 routes during 1960-1964, by direct recording of vegetation 



249 



