other small mammal species were captured infrequently. In the case of 

 the Wyoming pocket mouse, this probably is an accurate reflection of the species' 

 low abundance. However, for the two vole species and the least chipmunk^the 

 paucity of captures reflects strong habitat selection for western bulrush and 

 western snowberry-prairie rose associations for the former, and silver buffalo- 

 berry/western snowberry-prairie rose communities for the latter. A single cap- 

 ture of the thirteen-lined ground squirrel does not accurately reflect its true 

 abundance or strong habitat selection. It was seen in a variety of grassland 

 associations, often in the immediate vicinity of baited and set snap-traps. It 

 showed little affinity for the traps and apparently managed to free itself if 

 caught. 



Habitat Effects 



Table 37 shows the abundance of rodents captured in 24 snap-trap sites which 

 correspond to 15 separate habitat categories. The maximum number of rodent species 

 recorded for any snap-trapped site was three (one site) and the least zero (one 

 site). Sites yielding two species numbered seven and those with one species num- 

 bered fifteen. When control and experimental lines within similar habitat cat- 

 egories are combined, the maximum number of species per habitat category is four, 

 recorded in the riparian western bulrush habitat category. This type is confined 

 to an area less than 8 m (26 feet) wide along either bank of a permanent stream; 

 consequently, occurence of some species in this type is probably limited to trans- 

 ients and the actual number of resident species is probably less than four. 



Monthly changes in numbers of individuals of deer mice and other rodent 

 species captured in live-trap grids are shown in figure 51. 



Both sagebrush live-trap grids yielded only deer mice and western harvest 

 mice. Deer mice and Wyoming pocket mice were captured on the grassland control 

 grid, and the grassland experimental grid yielded all species found on other grids 

 with the addition of the thirteen-lined ground squirrel. 



In all snap-trapped and live-trapped habitat categories, deer mice occurred 

 with 100% frequency. The western harvest mouse was found in both live-trapped 

 habitats and in three of those snap-trapped. The Wyoming pocket mouse was con- 

 fined to the grassland live-trap grids, but was found in one grassland and one 

 sagebrush snap-trap transect. The sagebrush transect which produced Wyoming poc- 

 ket mouse captures (silver sagebrush/green needlegrass-western wheatgrass) was 

 more open than the two grids (94.3% vs 114.7% live plant cover), and its under- 

 story was dominated by green needlegrass rather than western wheatgrass as in 

 the grids. Percentage of absolute litter cover was slightly higher in the tran- 

 sect than the grid (34.8% vs 28.8%). The thirteen-lined ground squirrel was 

 captured only on the grassland experimental grid and one snap transect (blue 

 grama/thread- leaved sedge--needle-and-thread, #16) . Captures of the prairie vole 

 occurred in two snap-trapped habitats, riparian bulrush and western snowberry- 

 prairie rose, but this species was not found on the grids although the latter 

 vegetative association occurred on both grids. The meadow vole was captured only 

 once on a snap transect in riparian bulrush within 100 m (321 feet) of dense 

 grass cover. The least chipmunk was also not captured on grids but occurred 

 once in a silver buffaloberry/western snowberry-prairie rose snap transect in a 

 low coulee (table 36) . 



175 



