Species relative abundance distributions are one of the most well-documented patterns in ecology. Study of these distributions has been limited because model predictions do not vary enough to be distinguishable by goodness-of-fit tests. We present a simple model of relative abundance distributions, the shape of which depend on the species-area relationship, and make testable predictions for how the distributions should change as community processes change. Others have shown that the geometric sequence model can be derived from an empirical energetic trade-off. It is shown here that almost all possible shapes for relative abundance distributions can be fit by altering the strength of this trade-off. The shape of the distribution is parameterized by the exponent of the species-area relationship. Moreover, we suggest how the relative abundance distribution should change shape as different processes alter the intensity of competition, opening the way for an experimental analysis of relative abundance distributions. This model of relative abundance distributions does not depend on any assumption of competitive equilibrium, as do, for example, the geometric sequence and Hubbell's neutral theory. We suggest instead that as a community moves away from equilibrium the shape of its relative abundance distribution will change in a predictable manner. Finally, we present examples of field data which this model fits well.