Background/Question/Methods High-diversity communities are often characterized by large numbers of rare species coexisting with few dominant species. Although theory suggests that local interactions among dominant and rare species influence biodiversity, empirical tests of the underlying mechanisms are limited. We explored how two dominant guilds of large-stature plants (bunchgrasses and shrubs) influence recruitment limitation and local species diversity in a species-rich longleaf pine savanna (often >30 spp/m²) in Louisiana. We experimentally tested the hypotheses that competition from dominant guilds reduces diversity by limiting recruitment of seedlings and lowering survival of established plants. Our 2-year experiment consisted of two factorial treatments (competition reduction, seed addition) applied to 0.25-m² plots occupied by either a dominant bunchgrass (Schizachyrium tenerum) or shrub (Ilex glabra). When present at this scale, each species usually accounts for >90% of total aboveground biomass. In the competition-reduction treatment, we removed bunchgrasses or shrubs via herbicide or repeated clipping, respectively. In the seed-addition treatment, we added seeds of 35 focal small-stature, relatively rare species (approximating realistic dispersal rates) to test for competitive effects on germination and establishment. Control plots retained dominant species and received natural seed rain. For bunchgrasses, we also partitioned effects of aboveground competition from complete plant removal via a foliage reduction treatment. Results/Conclusions After 1 year, seed additions significantly increased total species richness in both bunchgrass and shrub plots. Competition reductions, in contrast, increased richness only in shrub plots. There was no significant effect of bunchgrass removal or foliage reduction on richness after 1 year. Species richness of the focal species paralleled patterns of total richness, with the exception that shrub removal had no significant effect on focal richness, indicating a positive effect of shrub removal on non-focal species. Overall, there were no significant interactive effects of treatments on species richness. In contrast, interactions influenced focal seedling density, especially in shrub plots, where seedling densities in the seed addition treatment were significantly higher in the competition reduction relative to control treatments. As the experiment continues, we hypothesize that competition from these dominant species will reduce local diversity by limiting recruitment of established seedlings. These patterns would suggest two important mechanisms of community assembly. First, seedling recruitment of sub-dominant species would occur primarily in low-competition habitats. Second, initial population growth in these new habitats would be enhanced through short-distance dispersal. Collectively, these mechanisms could help maintain coexistence of rare and dominant species in this and potentially other high-diversity plant communities.