Community structure and relative abundance of 27 species of small mammals and forest birds were compared among three types of residual forest stands and unharvested control forest (CO). Treatments were young and old mosaics (a checkerboard pattern of residual and logged forest units of 85-100 ha each) and megablocks (residual stands of 250300 ha isolated within a logged area of 2500-3000 ha). Relative abundances were also used to establish habitat use models (HUMs). We found no statistical difference in species relative abundances between treatments and COs, although small sample sizes limited statistical power. HUMs explained a large amount of variation in habitat use for 15 mammal and bird species (mean 57.4 3.5%, ranging between 22.3% and 75.7%). Variance partitioning emphasized the importance of stand structure characteristics as the principal predictors of abundance for censused species. Our results suggested that mosaics and megablocks are both suitable configurations to maintain studied species because no species exhibited lower relative abundances in such residual forest stands than COs. We suggest that residual forest stands planning should shift from a strictly landscape perspective toward a more holistic approach that considers residual forest structure as well as landscape characteristics.