It has commonly been assumed that grazing decreases the longevity of plants, however, very few studies address this question. Knowledge of mean age, longevity and other aspects of population dynamics would help improve estimates of productivity and nutrient cycling and aid understanding of range condition and trend. These data would also aid in developing a theory of community organization and control. As a step toward these ends, pantograph records from twelve 1 m2 plots established at the U. S. Sheep Station on the Snake River Plains of southeastern Idaho between 1923 and ' l935 were examined. The plots were mapped almost yearly until 1956. These records were analyzed for establishment and survival of populations of major vascular, perennial plant species of the tall threetip sagebrush-grass vegetation type. Average longevities of the grazed plants were significantly different from ungrazed populations for some species. For the pastures grazed by sheep in the fall the average longevities of the dominant grass and two subdominant grasses were increased whereas the average longevity of tall threetip sagebrush was reduced by grazing. The differences in average longevity due to grazing are thought due to alteration of competitive relationships. The photosynthetically active sagebrush is sensitive to grazing in the fall whereas the grasses are dormant. A slight reduction in the total volume of the sagebrush apparently releases resources to the grasses the following spring. The greatest mortality was observed during the first year of life for all species . Survivorship curves for the first eleven years of life of most species could be linearized by logarithmic transformations on both axes indicating that the chance for survival improved with age, up to the physiological maximum. No linear relationships of increasing size to age were found in these plants. This indicates a considerable plasticity in the growth characteristics of these plants, where in poor years negative growth can occur by the reduction of unsupportable biomass. This all ows plant size to fluctuate with fluctuating environment, and allows for stands of even sized plants with great age differences. Labeling recruitment to these plant populations as "pulse" phenomena becomes a matter of definition. If a log-normal distribution is expected, as indicated by some researchers, outliers are not present and "pulse" regeneration does not occur in this system. However, if recruitment is expected to follow a normal distribution then the years with high numbers of seedlings become "pulses," by definition. From these analyses, it appears that these semi-desert communities are closed (the available resources are totally used) and replacement of individuals is somewhat random following more or less random death.