The behavior and ecology of Uinta ground squirrels (Spermophilus armatus) at the Utah State University Forestry Field Station northeast of Logan, Utah, were studies 1964—71 to determine the role of behavior in population regulation. In 1968 the population was reduced experimentally to about one—half the previous density. The study area consisted of a lawn area where resident squirrels more than replaced themselves, and an area of mixed shrubs and grasses where they did not. Surplus squirrels from the lawn raised the population density off the lawn. Also, habitat suitabilities varied with population densities. Hence, densities in the two areas were not proportional to habitat preferences. Before the reduction the population density fluctuated widely but the mean change for 5 yr was approximately zero. A potentially high rate of increase was curtailed by disappearance of juveniles and yearlings; juveniles tended to disperse from the natal burrow. The principal changes following the reduction were increases in percentage of yearling squirrels breeding and of juveniles remaining on the study area. Litter size and proportion of adult females breeding did not change significantly, but losses during hibernation, due in part to badger predation, decreased. Dispersal of squirrels from the study area played a key role in population regulation. Dispersers probably suffered higher mortality than sedentary squirrels, but dispersal was not necessarily maladaptive for the individual. Dispersers may have been unable to produce young at high densities; thus emigration offered the only opportunity for reproduction.