In tallgrass prairie reconstruction, the way desired seeds are arranged on the landscape may affect species establishment, species persistence, and the establishment and persistence of undesired (nonseeded) species from the local propagule pool. To test effects of species seeding pattern on how grasslands develop spatially, we seeded 20—4 × 4–m bare soil plots with 16 tallgrass prairie species. Treatment plots were divided into 16—1 × 1–m subplots, 64—0.5 × 0.5–m subplots, 256—0.25 × 0.25–m subplots, or 1,024—0.125 × 0.125–m subplots. Each species was hand broadcast into separate subplots (1 m2 total area/species) within each plot. An additional treatment included uniformly mixing and broadcasting all seeds across a plot. We recorded species cover at the 0.125 × 0.125–m scale within each plot at the beginning of the second and third growing seasons. While species persistence was greatest within plots seeded with larger subplots, plots with smaller subplots were more spatially diverse and less occupied by nonseeded species over time than larger subplot and mixed plots. As is common in reconstruction efforts, establishment was variable among species and seeding with monospecific subplots enhanced colonization of desired rhizomatous species (e.g., Heliopsis helianthoides, Monarda fistulosa, Elymus virginicus) into unoccupied locations at the expense of species from the local propagule pool. Results suggest that seeding species in smaller, monospecific patches could result in grasslands with a more balanced native species composition than those established with a seed mixture approach.