Native plant communities invaded by cheatgrass (Bromus tectorum L.) are at risk of unnatural high intensity fires and conversion to cheatgrass monocultures. Management strategies that reduce cheatgrass abundance may potentially allow native species to expand and minimize further cheatgrass invasion. We tested whether the selective herbicide imazapic is effective in reducing cheatgrass and “releasing” native species in a semiarid grassland and shrub steppe in north-central Oregon. The experiment consisted of a completely randomized design with two treatments (sprayed with 70 g ai · ha⁻¹ of imazapic and unsprayed) and three replicates of each treatment applied to either 2.5 or 4 ha plots. We repeated this experiment in three different sites dominated by the following native species: 1) bluebunch wheatgrass (Pseudoroegneria spicata [Pursh] A. Löve ssp. spicata) and needle and thread (Hesperostipa comata [Trin. & Rupr.] Barkworth), 2) needle and thread and Sandberg bluegrass (Poa secunda J. Presl), and 3) big sagebrush (Artemisia tridentata Nutt.). Nested frequency of all plant species in 1-m² quadrats was collected for 1 yr pretreatment and 4 yr posttreatment. In all three sites, cheatgrass frequencies were significantly lower in sprayed plots than unsprayed plots for 3–4 yr posttreatment (P < 0.1). Other annual plant species were also impacted by imazapic, but the effects were highly variable by species and site. Only two native perennial species, hoary tansyaster (Machaeranthera canescens [Pursh] Gray) and big sagebrush, increased in sprayed plots, and increases occurred only at two sites. These results suggest that a short-term reduction in cheatgrass alone is not an effective strategy for increasing the abundance of most native perennial plant species.