An 11-year competition experiment among combinations of six prairie perennial plant species showed that resource competition theory generally predicted the long-term outcome of competition. We grew each species in replicated monocultures to determine its requirements for soil nitrate (R*) and light (I*). In six pairwise combinations, the species with the lower R* and I* excluded its competitor, as predicted by theory. In the remaining two pairwise combinations, one species had a lower R*, and the second had a lower I*; these species pairs coexisted, although it is unclear whether resource competition alone was responsible for their coexistence. Smaller differences in R* or I* between competing species led to slower rates of competitive exclusion, and the influence of R* differences on the rate of competitive exclusion was more pronounced on low-nitrogen soils, while the influence of I* differences was more pronounced on high-nitrogen (low-light) soils. These results were not explained by differences in initial species abundances or neutrality. However, only a few of our paired species coexisted under our experimentally imposed conditions (homogeneous soils, high seeding densities, minimal disturbance, regular water, and low herbivory levels), suggesting that other coexistence mechanisms help generate the diversity observed in natural communities.