The competitive exclusion principle poses the pressing question of how biodiversity is maintained in nature. Many mechanisms have been proposed to explain diversity and to resolve what has become known as the “paradox of the plankton”. We propose a dichotomy among these mechanisms in order to enable empiricists to begin testing their relative importance. Specifically, the mechanisms can be categorized as being internally generated or as depending on forces external to the competitive community. Here we tested whether the internal competitive dynamics of a phytoplankton assemblage or externally generated resource variability (a disturbance) were more effective at maintaining species diversity over time. We also tested whether the species composition of assemblages was important in determining the persistence of species diversity. We employed controlled microcosm experiments in which we either imposed exogenous variability in nutrient availability via serial dilution or allowed the communities to remain completely undisturbed. We found that species diversity was maintained most effectively in undisturbed microcosms in which only internal dynamics regulated coexistence. We also found that the community composition of the assemblage significantly interacted with the disturbance regime in determining species diversity. This confirmed the importance of internal dynamics and community composition in maintaining species diversity.