Social and local environments of the vole Microtus townsendii

Why do vole populations usually stop increasing before they run out of food, then decline to a phase of low numbers and low body weights? Here I test the hypothesis that indefinite increase is prevented by spacing behaviour and that the decline is due to selection against the phenotypes present during the increase and early peak phases. This hypothesis implies that, by establishing a stock of voles born in a phase of increase, one can produce peak populations out of phase with a decline in a natural cycle. Using such voles, I established increasing and peak populations on trapped-out areas of grassland near Vancouver, British Columbia. At first these populations remained abundant during a decline in a natural population in spring 1976. This decline, however, was not followed by a definitive low phase but by an immediate return to a peak. Also, despite good reproduction and survival before the middle of March 1976, numbers of the experimental animals then declined further than expected. Failure to produce out-of-phase populations seems to have been due to the falsity of two assumptions: (i) that increase-phase voles are innately docile and (ii) that selection necessarily occurs at high densities. Instead of being docile during the increase phase, voles are, I now suggest, so aggressive towards strangers that they maintain discrete family groups. But amicable behaviour within groups, if disrupted by hostile behaviour between them, would explain why increase stops or is interrupted by a temporary decline. By contrast, a 2- to 3-year period of decline or a refractory phase of low numbers and low body weights is more plausibly explained by a change in behaviour due to natural selection. A decline in 1977–1980, associated with poor survival and reduced body weight, may therefore be attributed to a change in the social rather than the local environment.