Abstract Prairie restoration at the northern edge of the Great Plains can be frustrated by previously established non-native perennial grasses. We compared the emergence of a widely introduced grass, Agropyron cristatum, and a common native grass, Bouteloua gracilis, in a 4-year-old field experiment in which the Agropyron-dominated vegetation had either been left intact or treated annually with herbicide. This was done at two levels of water supply, reflecting conditions expected in wet and dry years, to examine the effects of among-year variability in precipitation. Water addition significantly increased the emergence of both surface-sown and buried (1 cm deep) seeds. Herbicide treatment of neighbors did not increase the emergence of experimentally added seeds. Emergence was much greater for buried (80%) than surface-sown seeds (20%). Significantly more Bouteloua than Agropyron germinated from experimentally buried seeds. Whereas only a single seedling of Bouteloua emerged from the existing seed bank, the mean density of Agropyron seedlings emerging from the seed bank was 930/m2 (range, 0 to 6,455/m2). Surprisingly, the emergence of Agropyron from the seed bank was not decreased by 4 years of herbicide treatment, possibly because herbicide may release Agropyron from intraspecific competition and allow increased seed production to compensate for decreased plant abundance. In summary, we found few differences between Agropyron and Bouteloua in spring and summer emergence at high or low water availability. The persistence of Agropyron stands despite repeated herbicide application may be partly due to increased seed production.