To investigate experimentally the relationship between dispersal and population regulation in small mammals, voles were removed continuously from two plots in southern Indiana for 2 years. Three control populations of two Microtus species were monitored over the same period, and animals dispersing onto the experimental areas were compared with resident control animals for the following characteristics: (1) age, weight, and sex; (2) genotype for two polymorphic plasma proteins, leucine aminopeptidase (LAP) and transferrin (Tf); and (3) exploratory, aggressive, and general activity behavior of males. Dispersal was most common during the fall and winter, and in the phase of population increase 59% of male and 69% of female Microtus pennsylvanicus loss from two control populations could be accounted for by dispersal. In contrast, little of the high loss during the population decline could be associated with dispersal. In the late peak and decline periods male M. pennsylvanicus of the Tf—E and LAP—S phenotypes were more common among dispersing animals. In Microtus ochrogaster males the very rare Tf—F type was found only in dispersing ales. Young females were more common in dispersing populations of both species, and sexual maturation and dispersal were related in both vole species. However, dispersal by adults was only common. Dispersing males tended to be less active than males of resident populations. During a period of peak population density, dispersing male M. pennsylvanicus were more aggressive. As population density increased, "intolerant" animals moved into less densely populated areas. The genetic differences between dispersing and resident populations tend to support Howard's hypothesis that a genetic polymorphism influences the tendency to disperse. The association between increasing populations and abundant dispersal agrees with Lidicker's theory of population regulation by emigration and Chitty's hypothesis that a behavioral polymorphism regulates population density.