Planting native grasses can provide a source of seed for prairie restorations, but requires knowledge of how the plants that establish will perform. This study sought to determine variation in fitness of population sources of Andropogon gerardii, a dominant grassland species, when grown in a mesic common garden. Using multiple population sources, we tested the hypothesis that plants from populations within the local region would exhibit a ‘home-site advantage’ as measured by higher fitness compared to plants from populations collected from drier regions of the tallgrass prairie ecosystem (up to 986 km west of the common garden). Plants collected from four pristine, never restored population sources from each of three regions (central Kansas, eastern Kansas, and southern Illinois) were raised in the greenhouse from seeds and planted in a common garden in Illinois. To estimate fitness, we used commonly measured traits related to seed production, including flowering tiller number and number of flowering raceme branches, seed number, viability, and percentage germination. There was no evidence of a ‘home-site advantage’ for populations originating from southern Illinois. Rather, there was high within-region variability in fecundity. Plants from southern Illinois had the largest number of raceme branches per plant. Plants from eastern Kansas had the highest number of vegetative tillers per plant. Plants from central Kansas produced the most germinable seeds. Under the current climate, plants from any one of the three regions may be suitable to propagate seeds for restoration, but other traits may vary among populations to affect height, cover, and productivity.