Trampling experiments were conducted in five high-elevation plant communities in the Wind River Mountains. In one experiment (simulating infrequent acute disturbance), plots were trampled once and recovery was followed for 3 yr. In another experiment (simulating chronic disturbance), plots were trampled for three successive years and recovery was followed for one year. Yearly trampling intensities were 0 to 500 passes (up to 800 and 1000 in two communities). Structural responses (reductions in cover and height) were more pronounced than compositional responses. Low levels of trampling caused substantial reductions in cover and height, but rates of change decreased as trampling intensity increased. The ability of different plant communities to tolerate trampling disturbance varied by at least an order of magnitude, suggesting that impacts can be reduced by directing use to more tolerant plant communities. Moreover, tolerance can be gauged on the basis of readily observable plant characteristics. Plant communities with groundcover dominated by chamaephytes (primarily low, woody shrubs) or by erect, caulescent forbs were less resistant than those with groundcovers dominated by turf-forming or caespitose graminoids, or by forbs with matted, caespitose or rosette growth forms. Plant communities with woody, chamaephytic groundcover were less resilient than other plant communities.