Species diversity is expected to increase with environmental heterogeneity, and this pattern has been confirmed by numerous observational studies. For plants, this spatial heterogeneity is likely driven by differences in soil substrate and by changes in soil conditions that occur as different species colonize and alter the soils and the invasibility of the community. Using a field experiment, we tested how the spatial pattern of seed arrival (aggregated versus uniform seed mixtures), soil heterogeneity, and patch size influence species richness in grassland plant communities in south-central Kansas. We established 96, 4x4.8 m plots each divided into 120 large-scale (0.4x0.4 m) or 480 small-scale (0.2x0.2 m) patches. We then manipulated the soils within a plot to create either homogenous or heterogeneous soil conditions. We also sowed 40 species into a plot with either one species per soil patch, to create aggregated species distributions within plots, or were sown uniformly across plots. After four growing seasons, these communities were influenced by interactive effects of seed arrival, patch size, and soil heterogeneity. Species richness along with species composition and abundance differed when the seed arrival of each species was spatially aggregated rather than in uniform mixtures particularly when the patch scale was large. However, in spatially uniform seed arrival plots, richness was higher when soils were homogeneous compared with heterogeneous soils. These differences were primarily driven by the arrival of unsown species from the surrounding plant communities or the seed bank, which were more weedy than those utilized in the seeding treatments. These results suggest that the spatial structure of seed arrival and size of the seed patches may have stronger effects on plant communities than soil heterogeneity.