Abstract Crown rust of oats ( Puccinia coronata f. sp. avenae ) was investigated as a biocontrol agent for wild oats ( Avena fatua ) on San Clemente Island, California. Successful restoration of the native habitats of this island will involve the reduction of wild oats and revegetation with native grasses such as stipa ( Nassella pulchra ). Determination of the outcome of interference between wild oats and stipa is important in the prediction of the success of the biocontrol agent. An addition series design was used to investigate these interactions, with densities of each species ranging from 0 to 2000 seeds per m 2 . Eight replicates were established, four of which were randomly chosen and infected with the pathogen. After 75 days, plant height, dry weight, and seed production were measured. The data were fit to a hyperbola surface model by use of a nonlinear regression procedure. Results indicate that wild oat is the superior competitor at the seedling stage; however, early rust infection greatly reduces fitness of wild oats, causing an increased fitness for stipa. Simulations with a plant community model constructed from the results of the greenhouse experiment and information in the literature indicated that an equilibrium may be established between wild oat and stipa if high initial seeding rates of stipa are used in revegetation.