The Winter Food Limitation Hypothesis (WFLH) states that winter food abundance is a dominant source of population limitation of migratory birds. Evidence is accumulating that long-distance migratory birds wintering in tropical climates have high overwinter survival probabilities and that winter food limitation mainly affects their fitness nonlethally by limiting energetic reserves necessary for successful reproduction. In contrast, the relative roles of direct mortality vs. indirect effects caused by food limitation have not been investigated thoroughly on short-distance migratory birds wintering in temperate zones, where they face thermal challenges. We performed the first test of the WFLH for a temperate migratory bird in the wild on the Swamp Sparrow (Melospiza georgiana), with a replicated plot-wide food supplementation experiment. In contrast to tropical, but consistent with other temperate-wintering migrants, Swamp Sparrows on unmanipulated plots showed relatively low apparent survival across the winter. Following food addition, birds (1) immigrated to experimental plots, which subsequently supported approximately 50% higher abundances, (2) experienced increases of within-season apparent survival of 8-10%, depending on age/sex class, and (3) had higher-scaled mass index values, all supporting winter food limitation. The last two findings are interrelated because birds with higher scaled mass had higher survival probabilities, further supporting direct effects of winter food limitation. Food limitation of fat reserves might also have indirect effects on reproductive success by limiting migration timing and survival during migration. Increases in scaled mass were higher in females, suggesting that they are disproportionately affected by food limitation, possibly through competition. Based on Robust Design Modeling, we found no support for emigration prior to food addition, indicating that our estimates of mortality are unbiased.