The development of mechanistic, predictive ecological theory will entail the explicit inclusion of organismal tradeoffs, of environmental constraints, and of the basic mechanisms of interspecific interaction. This approach was used to address the causes of species dominance and successional dynamics in sandplain vegetation in Minnesota. The major constraints on plants were soil N and disturbance, with N competition being a major force. Nutrients other than N, herbivory and light were of minor importance. As predicted by theory, the superior N competitors were the species that, when growing in long-term monocultures in the field, lowered soil extractable N the most. These species had high root biomass and low tissue N levels. Seven alternative hypotheses of succession, each named after its underlying tradeoff, were tested. The colonization-nutrient competition hypothesis provided the best explanation for the initial dominance (years 0-40) of herbs, whereas the nutrient versus light competition hypothesis best explained the long-term dominance by woody plants. Hypotheses involving transient dynamics caused by differences in maximal growth rates were rejected. -from Author