Plant competition experiments have historically used designs that are difficult to interpret due to confounding problems. Recently, designs based on a 'response function' approach have been proposed and tested in various plant mixture settings. For this study, 3 species were used that are important in current revegetation practices in the Intermountain West. 'Nordan' (Agropyron desertorum [Fish. ex Link] Shult.) and 'Hycrest' (A. cristatum [L.] Gaertn. x desertorum) crested wheatgrass are commonly-used revegetation species on rangelands susceptible to cheatgrass (Bromus tectorum L.) invasion, although little quantitative data exist that compare their competitive abilities. We evaluated the competitive ability of Hycrest and Nordan seedlings in 2-species mixtures with cheatgrass in a greenhouse study. Linear and nonlinear models were developed for a range of densities (130- 520 seeds m-2) for each species to predict median above-ground biomass and tiller numbers and to further test the usefulness of this design for evaluating species to rehabilitate rangelands. In both experiments, increasing Hycrest and Nordan densities reduced their own biomass and tiller production while increasing Hycrest densities reduced cheatgrass biomass and tiller production. Nordan did not affect cheatgrass biomass and tiller production. However, increasing cheatgrass densities reduced Hycrest and Nordan biomass and tiller production, and its own biomass and tiller production. The competition index i.e. substitution rate, indicated that Hycrest seedlings were better competitors with cheatgrass than Nordan, although in all mixtures, cheatgrass plants were the superior competitors. Further field research using this design, where environmental inputs are less optimal and diverse, is needed to validate these results and to further evaluate the use of this approach in examining effects of intra- and interspecific competition.