In an experiment that factorially manipulated plant diversity, CO2, and N, we quantified the effects of the presence of species on assemblage biomass over 10 time points distributed over 5 years. Thirteen of the 16 species planted had statistically significant effects on aboveground and/or belowground biomass. Species differed dramatically in their effects on biomass without any relationship between aboveground and below-ground effects. Temporal complementarity among species in their effects seasonally, successionally, and in response to a dry summer maintained the diversity–biomass relationships over time and may be the cause behind higher diversity plots having less variation in biomass over time. The response of plant biomass to elevated N, but not CO2, was at times entirely dependent on the presence of a single species.