Winter is becoming warmer and shorter across the northern hemisphere, and reductions in snow depth can decrease tree seedling survival by exposing seedlings to harmful microclimates. Similarly, herbivory by small mammals can also limit the survival and distribution of woody plants, but it is unclear whether winter climate change will alter small-mammal herbivory. Although small-scale experiments show that snow removal can either increase or decrease both soil temperatures and herbivory, we currently lack snow-removal experiments replicated across large spatial scales that are needed to understand the effect of reduced snow. To examine how winter herbivory and snow conditions influence seedling dynamics, we transplanted Acer saccharum and Tsuga canadensis seedlings across a 180 km latitudinal gradient in northern Wisconsin, where snow depth varied seven-fold among sites. Seedlings were transplanted into one of two herbivory treatments (small-mammal exclosure, small-mammal access) and one of two late-winter snow removal treatments (snow removed, snow unmanipulated). Snow removal increased soil freeze-thaw frequency and cumulative growing degree-days (GDD), but the magnitude of these effects depended on forest canopy composition. Acer saccharum survival decreased where snow was removed, but only at sites without conifers. Excluding small mammals increased A. saccharum survival at sites where the smallmammal herbivore Myodes gapperi was present. Excluding small mammals also increased T. canadensis survival in plots with < 5 cm snow. Because variation in canopy composition and M. gapperi presence were important predictors of seedling survival across the snow-depth gradient, these results reveal complexity in the ability to accurately predict patterns of winter seedling survival over large spatial scales. Global change scenarios that project future patterns of seedling recruitment may benefit from explicitly considering interactions between snow conditions and small-mammal winter herbivory.