these communities. We examined vegetation structure and species distribution patterns in the forest understory community of two topographically similar watersheds to evaluate the predicted roles of abiotic limitation and competition along a productivity gradient. Canonical correspondence analysis identified a dominant compositional gradient that was strongly correlated with moisture, light, and herbaceous-layer cover. Given these relationships, the first axis of the ordination was taken as a reliable index of productivity. Contrary to expectation, species richness increased monotonically along the gradient rather than reaching a maximum and then declining which suggests that species are not excluded by light competition at the high-productivity end of the gradient. Other vegetation patterns, however, indicated both abiotic limitation in regions of low productivity and competition at high productivity. Patterns in species' distributions also suggested resource limitation at low productivity, but they showed mixed evidence of competition at high productivity. Taken together, these patterns in vegetation structure and species' distributions imply that while competition may occur among herbaceous-layer species, particularly at high productivity, competitive exclusion remains low. This may be because the traits that permit survival under tree canopy shading reduce the effects of light competition with other understory species, or because understory productivity is not high enough to induce intense light competition among herbaceouslayer species.