Dispersal of deer mice, Peromyscus maniculatus, was measured as immigration to and emigration from two control areas, and as immigration to a removal area. The number of mice dispersing was linearly related to the densities on the control areas, while the proportion of the population dispersing (rate of dispersal) was correlated primarily with the rate of increase of control populations. High rates of dispersal were also associated with a breakdown of the established social structure in the spring and fall. Dispersing animals were compared to residents with respect to sex ratio, weight, age, and breeding condition. The types of animals dispersing varied seasonally: light-weight, non-breeding males dispersed in the spring and summer; juveniles and breeding males dispersed at the end of the breeding season; and light-weight mice of both sexes dispersed over the winter. It is proposed that the animals that dominated the dispersal samples each season were moving in response to social pressure from residents, or local limitations of some resource, and thus, that dispersal was adaptive for the individuals concerned. Some tests of the hypotheses concerning resource limitation are suggested.