Ecologists would like to explain general patterns observed across multi-species communities, such as species-area and abundance-frequency relationships, in terms of the fundamental processes of birth, death and migration underlying the dynamics of all constituent species. The unified neutral theory of biodiversity and related theories based on these fundamental population processes have successfully recreated general species-abundance patterns without accounting for either the variation among species and individuals or resource-releasing processes such as predation and disturbance, long emphasized in ecological theory. If ecological communities can be described adequately without estimating variation in species and their interactions, our understanding of ecological community organization and the predicted consequences of reduced biodiversity and environmental change would shift markedly. Here, I introduce a strong method to test the neutral theory that combines field parameterization of the underlying population dynamics with a field experiment, and apply it to a rocky intertidal community. Although the observed abundance-frequency distribution of the system follows that predicted by the neutral theory, the neutral theory predicts poorly the field experimental results, indicating an essential role for variation in species interactions.