In many U.S. federally designated wilderness areas, wildfires are likely to burn of their own accord due to favorable management policies and remote location. Previous research suggested that limitations on fire size can result from the evolution of natural fire regimes, specifically in places where fuels were recently reduced by previous burning. To explore the broader-scale importance of fire management on wilderness landscapes, we selected three study regions representing diverse ecosystems in the western U.S. and modeled the change in fire size distributions across a gradient defined by wilderness/non-wilderness boundaries. For randomly selected locations across the gradient, we derived a scaling parameter (α) using fire size-frequency data for public lands (1984-2007); the parameter reflected the magnitude of change in the right tail of the fire size distribution where the largest fires reside. We then used quantile regression to model changes in α across the wilderness gradient, interpreting the results in terms of constraints on the relative role of large fires in structuring the fire size distribution. In the Southwest study region, the influence of large fires on size distributions decreased across the gradient toward wilderness at some places, suggesting that increased occurrence of natural burning, favored by wilderness management, led to limitations on fire sizes within recent timeframes. In contrast, we were unable to support the expectation that wilderness fire management limits the role of large fires in the Sierra Nevada and Northern Rockies study regions. Rather, the predominance of large fires increased toward wilderness interiors. Among spatial climate and topographic roughness variables included in our study, only winter and fire season precipitation limited fire size in the Northern Rockies, whereas several constraints on large fire occurrence operated in other regions. In southwestern ecosystems, evidence is needed to document stability in fire size distributions through time. In ecosystems of the Sierra Nevada and Northern Rockies, a longer temporal extent of observations may better match scales of disturbance and recovery. Our findings reflect the role of wilderness in addressing a fire deficit which has resulted from strong human influences on forests and fires over the past 150 years.