Previous studies have indicated that the relationship between community biomass and species density can be represented by a multivariate model in which abiotic variables influence species density both through effects on biomass and through effects on the species pool. In this paper, we use data from grazed and ungrazed coastal meadows in Finland to evaluate and extend this general conceptual model of the factors controlling species density. Structural equation analysis was used to evaluate a model for all meadows and then to perform a multigroup analysis to determine how grazed and ungrazed meadows differ. By itself, biomass could explain only 12% of the variation in species density while the multivariate model was able to explain 47% using five types of predictor variables: site, soil, flooding, grazing, and biomass. Analyses found that flooding explained the greatest amount of variability in species density, primarily through negative effects on the species pool. Grazing was also found to have a strong effect on species density and results suggest that its negative influence may be largely through reductions in the species pool in grazed meadows. The most important difference found between grazed and ungrazed meadows was that species density had a strong negative relationship to biomass in the ungrazed meadows but no significant relationship in the grazed ones. Thus, it appears that the influence of competition on species density was much greater in ungrazed meadows compared to grazed ones.