Technical Abstract: Manipulation of plant species diversity may provide a way to improve the ecosystem functioning of managed systems by increasing productivity and suppressing weedy species. As yet, the functional role of species richness is not well-enough understood to enable practical application. We investigated the effects of differing species richness on cover and plant community composition in a grazed temperate grassland. Nine plant species from three different functional groups (grass, legume, forb) were planted in the fall of 2001, in mixtures of 2, 3, 6, and 9 species. Two replicates of each treatment were planted in 1 ha plots, and rotationally grazed by dairy cattle during 2002 and 2003. Species composition was assessed during April and October of each year. The summer was particularly dry in 2002, while 2003 was unusually wet. Higher richness conferred an establishment advantage; richness was positively correlated with total cover in fall 2001 and spring 2002. Although cover of unplanted (weedy) species declined with increasing richness on these dates, there was no difference in the number of unplanted species. After the first year there was little difference in cover or richness between the treatments. The forb Cichorium intybus, part of the 3, 6 and 9 species mixtures, established rapidly and contributed greatly to the early success of these treatments. Several other grasses and legumes increased in abundance over the first year, then declined. By the end of the study, the same species (Dactylis glomerata in all treatments and Trifolium repens in 2, 3 and 9-species mixes) were dominant regardless of initial seeding. Species identity was more important to cover and invasion resistance than was species richness. Management recommendations based on particular combinations of species may be more effective at improving ecosystem function than those based solely on species richness.