Plant competitive balance often changes with perturbations in soil fertility and atmospheric carbon dioxide (CO2). I reviewed the literature on responses of mixture-grown species to elevated CO2, evaluating the importance of differences in plant competitive ability for species responses. I also examined responses of California annual serpentine grassland species to experimental changes in soil fertility and CO2. Elevated CO2 and increased soil fertility enhanced ability of the introduced European grass, Bromus hordeaceus, to compete with the two native herbs, Plantago erecta and Lasthenia californica. This suggests that increases in these resources could increase Bromus' relative abundance in this native species-dominated grassland. High root allocation was associated with superior competitive ability at low soil fertility, and high shoot allocation was advantageous at high soil fertility. Variability in root and shoot allocation in response to soil fertilization appeared to be important to the native herbs' ability to compete successfully at both low and high soil fertility. Strong competitive interactions between species can lead to species exclusions. One mechanism allowing local coexistence of plant species in the face of strong competitive interactions is variation in competitive ability with soil characteristics. I measured moisture, depth, and nitrogen availability in soil (microsites) beneath patches of serpentine grassland species to determine whether these factors varied among patches dominated by different species. I also measured ability of these species to compete with one another in these different microsites. Patches of the native herb, Calycadenia multiglandulosum, occurred in the deepest microsites, patches of Plantago erecta occurred in microsites of intermediate depth, and patches of Lasthenia californica occurred in the shallowest and wettest microsites. Presence of Plantago erecta reduced soil nitrogen availability, even when Plantago was grown in microsites of initially higher nitrogen availability. Thus, both differential species responses to and differential species effects on soil characteristics were responsible for associations between species and microsites. All species performed significantly better against at least one of the other study species in the microsites where they occurred most abundantly, suggesting that variation in competitive ability with variation in soil characteristics can be an important mechanism of coexistence.