Questions: Grazing may speed or slow secondary succession, and the direction may depend on seral stage and relative tolerance of native perennial grasses compared with annual invasive species. How does grazing affect succession where undisturbed communities have a long evolutionary history of grazing by native herbivores and are tolerant to livestock grazing? Location: Shortgrass steppe, North American Great Plains, Colorado (40°49′N, 104°46′W), USA. Methods: Ungrazed and grazed early-seral (4–6 yr) and mid-seral (18–20 yr) seeded fields (Conservation Reserve Program) and traditionally grazed native steppe were compared for effects on plant composition in relation to changes expected from regional succession models. Results: Recovery towards undisturbed native shortgrass steppe for early- and mid-seral communities, respectively, was 19% and 36% for total vegetation cover, 5% and 21% for planted native species, 10% and 88% for non-planted native perennial grasses, only 0.2% and 13% for short grasses, and overall dissimilarity in community species compositions was 97% and 68%. In general, grazing effects were neutral or most often not significant in all years and/or were small in overall community magnitude. The early-seral community displayed more changes indicative of a slowing of succession with grazing (total vegetative and grass basal cover) rather than reducing invasive species (species targeted by timing of grazing), although drought had limited the establishment of grazing-tolerant short grasses. The mid-seral community showed more changes consistent with advancing successional recovery with grazing (overall community dissimilarity, forbs, planted native perennial grasses, tall grasses and target species). However, non-planted native perennial grasses responded negatively to grazing in the mid-seral community and positively in native shortgrass steppe where outside seed would originate. Conclusions: Grazing effects on particular functional groups and species were not the same across seral stages, were mixed in terms of speeding or slowing succession, and were generally not large at the community level. Evolutionary history of grazing may serve as a general guide but decisions on whether to graze successional grasslands may best be made after assessing whether tolerant perennial short grass species are significant components. Monitoring may then be necessary to determine species responses in particular community matrixes and effects on subsequent immigration of non-seeded native perennial species.