130 William C. McComb and Robert L. Rumsey 



to captures of T. striatus, but they also found high plant species richness 

 a characteristic important to this species. 



Peromyscus leucopus. — The habitat-wide model selected crown 

 cover (-0.28) and log diameter (+0.17) as important to capture of P. 

 leucopus (/K0.05). Captures were most strongly correlated (P<0.01) 

 with crown cover (-0.31), log diameter (+0.23), basal area (-0.22), dis- 

 tance to log (-0.21), log length (+0.20), and distance to a tree (+0.20) 

 (Table 4). Understory structure and log characteristics were correlated 

 (ZK0.05) with capture on four and three of the treatments, and three 

 and two of the aspects, respectively. The importance of each microsite 

 characteristic to capture of P. leucopus varied among treatments and 

 probably varies geographically. Dueser and Shugart (1978) reported a 

 deciduous overstory, wide tree dispersion, high species richness, and 

 large stumps important to P. leucopus in Tennessee. Geier and Best 

 (1980) found logs, brushpiles, and low plant species richness important 

 to P. leucopus abundance in Iowa. The most likely place to capture P. 

 leucopus in our plots was on the edge of a clearcut with an understory 

 of > 8 species and > 50 stems per 4 m 2 , and within 2.8 m of a log over 

 20 cm in diameter, on a site covered more than 4% by rocks (R 2 =47.3). 



Ochrotomys nuttalli. — This species selected sites on clearcuts with 

 slopes > 21%, < 146 m from water, with < 25% leaf cover, > 50% 

 understory cover, > 1.1 m from a log and > 5.4 m from a rock 

 (R 2 =62.1). Dueser and Shugart (1978) and Linzey (1968) identified dense 

 understory cover as important to O. nuttalli in Tennessee, but they did 

 not report leaf cover (SPRC = -0.86), logs (SPRC = +0.29), or rocks 

 (SPRC = +0.52) as important to this species. 



Microtus pinetorum. — A habitat-wide model selected understory 

 diversity (-0.98), understory density (+0.90), understory richness (+0.76), 

 basal area (+0.81), and tree density (-0.69) as important to capture M. 

 pinetorum. Kirkland (1978) also reported microtines responding to 

 changes in vegetative structure at West Virginia sites. Microtus pineto- 

 rum was most likely captured on the M91 plot on northfacing slopes 

 with a basal area > 24 m 2 /ha, near a snag < 28 cm dbh, on a 4 m 2 site 

 with < 1 tree, and with logs < 5.5 m long (R 2 =32.6). 



DISCUSSION 



Differential response of small mammals to treatments and aspects 

 is a function of differences in the relative abundance of individuals, and 

 of habitat components among sites. Results of studies of habitat prefer- 

 ences and relative abundances of small mammals may differ from ours 

 if conducted in a contiguous rather than a discontiguous forest. Four 

 species responded to treatments by change in relative abundance and 

 two species responded to aspects by the same change. Biotic microsite 

 characteristics (vegetative cover, logs, stumps, leaf cover, etc.) were 



