Natural regulation of native ungulates was initiated in 1968 in Yellowstone National Park (YNP) based on the premise that ungulates would reach an equilibrium with their plant resources. The natural-regulation management model stated: density dependence will regulate ungulates (i.e., a dynamic equilibrium will result between ungulates and their food supply, within some bounds of vegetation and soil effects); and no retrogression of soil and vegetation will occur from elk (Cervus elaphus) grazing during this process. The historical record indicated that elk were abundant in the system and elk were primarily food limited before settlement by European man (i.e., wolves [Canis lupus] and Native Americans were only an adjunct to the density dependent population regulation of ungulates). Density dependence was demonstrated in elk, but not in bison (Bison bison). No widespread evidence of overgrazing was observed through 1993 in study sites within vegetation communities that comprised about 97% of the winter range. No evidence of increased exotics, increased sediment yield, warming or drying of the soil, changes in soil nutrients, or differences in aboveground standing-crop biomass of plants was found between grazed and ungrazed plots. Ungulate herbivory apparently stimulated aboveground production of grasses, enhanced nitrogen and macronutrients in grasses, increased nutrient cycling, and enhanced measures of fitness in 6 common plants. However, exposed soil surface (bare ground and pebbles combined) was 11-18% greater on grazed than ungrazed plots, apparently due to a 71% decline in dead and standing litter on grazed plots. Percent live-plant basal cover, however, did not differ on grazed versus ungrazed plots, and there was no difference in soil microclimate or sediment yield. Differences in the abundance of 12% of the herbaceous species were found in grazed versus ungrazed sites (16 of 128 species); 10 were declines and 6 were increases. Willow (Salix spp.) and aspen (Populus tremuloides) declines predated the new management policy, but their slow declines also continued after 1968. Three uncommon plant species (aspen, willow, and Wyoming big sagebrush [Artemisia tridentata tridentata]) and 1 herbivore (moose [Alces alces]) declined under natural-regulation management. Two uncommon species of woody browse (aspen, Wyoming big sagebrush) were overutilized by ungulates (consumption of >2/3 current annual growth occurred). We conclude the natural-regulation model for YNP was flawed in its assumptions of a single, steady state for the park, based on conditions presumed to exist in 1870 prior to establishment of the national park. The period selected as a standard (1870) was an unusual period characterized by frequent large fires and floods, common wolves, few elk, and a cooler, wetter climate. We also conclude there is a high level of uncertainty surrounding what elk densities were in pre-Columbian times (element 3 of the management model) and what effects wolves and Native Americans had in regulating the elk population.