Runoff and sediment yield wer monitored from 1983 through 1986 on a range site dominated by creosotebush. The site was rootplowed and seeded, treated with tebuthiuron at 36 kg ai ha ^-1, or left untreated. Runoff from rootplowed and herbicide-treated plots was no different from untreated plots for 1 yr, but sedimetn yield from treated plots was lower than that from the untreated plots. Rootplowing and seeding increased sediment yield in the second year (1984), whereas treating with herbicide decreased sediment yield. Runoff and sediment yield during 1985 and 1986 were greatest from the untreated areas and least from the rootplowed and herbicide-treated areas. Total cover was nearly the same for each treatment. The untreated plots had 13% shrub cover, the others had none. Between shrubs in untreated plots, bare ground occupied large interconnected areas that contributed to the higher runoff and erosion rates.