Research to date on effects of fire exclusion in ponderosa pine (Pinus ponderosa) forests has been limited by narrow geographical focus, by confounding effects due to prior logging at research sites, and by uncertainty from using reconstructions of past conditions to infer changes. For the work presented here, reference stands in unlogged ponderosa pine/Douglas-fir (Pseudotsuga menziesii) forests at seven sites across a broad region of Idaho and Montana were selected for study. Within each site, understory and overstory community structure and composition were surveyed in 2–4 stands exposed to different numbers of fires during the 20th century. Student's t-tests were used to test differences between the most fire excluded ("unburned") and the most frequently burned ("burned") stands across the entire study. Multivariate analyses were used to characterize successional changes resulting from lack of fire. Understory community composition varied from site to site and was not related to fire exposure. However, species richness was higher in unburned stands. Overall tree densities and densities of shade-tolerant tree species increased with the absence of fire. There was a marginally significant association between xeric understories and ponderosa pine dominated overstories and mesic understories and Douglas-fir or grand fir dominated overstories. Our multivariate analysis suggests that while general trajectories of succession with absence of fire may be predictable, the structure and composition of ponderosa pine/Douglas-fir communities across complex landscapes may be difficult to relate to specific exposure to fire or time-since-fire. This study highlights the importance of natural variability and heterogeneity in ponderosa pine/Douglas-fir forests of the inland Northwest, and supports other recent research calling for cautious approaches to restoration in these forests.