In 1994 two 0.65 ha deer exclosures were built in a second growth forest in northeastern Pennsylvania, one under an oak-maple canopy and the other under eastern hemlock. We hypothesized that populations of white-footed mice (Peromyscus leucopus) outside the exclosures were depressed because of herbaceous cover removal by white-tailed deer (Odocoileus virginianus) and/or by acorn competition with deer and would be higher inside the exclosures. For 10 y, the small mammal communities were censured with one 10 × 4 Sherman live-trap grid in each exclosure and one control grid of identical dimensions immediately outside each exclosure. Paired-T comparisons between the oak-maple exclosure and control grids, and the hemlock exclosure and control grids indicated that more total males were captured in the oak-maple exclosure grid than in the oak-maple control grid. Although repeated-measures ANOVA indicated that habitat (tree canopy type) produced more differences in mice numbers than did the presence of the exclosures, it also demonstrated that numbers of large females fluctuated less in the oak-maple exclosure than in the corresponding control. A repeated measures ANOVA of Jolly-Seber population estimate indicated that more mice were trapped in the combined oak-maple and hemlock exclosure grids than in the combined control grids.