Delayed Density Dependence as a Determinant of Vole Cycles

Abstract

Largely synchronous population fluctuations of Clethrionomys glareolus, C. rufocanus, and Microtus agrestis were monitored by snap—trapping in spring and autumn in 1971—1988 in a strongly seasonal environment near Umea, northern Sweden. All species were cyclic in the sense that they showed fairly regular (3—4 yr) fluctuations, but amplitudes (Nmax/Nmin) varied, averaging °200—fold in each species. This conclusion was supported by autocorrelation and spectral analysis, and by fitting time series data to a model for phase—forgetting cycles. By contrast, data did not conform to a model for phase—remembering cycles (with fixed period and amplitude). The transition between cycles, i.e., from the low to increase phase, was characterized by a distinct shift in rate of change in numbers from low to high or markedly higher values both in summer and winter. Generally, rate of change in summer declined continuously from the increase phase through each cycle. Moreover, there was a similar decrease of rate of change in winter, although rate of change (mainly in C. rufocanus and M. agrestis) first frequently increased early in the cycle. Rate of changes was delayed density dependent in all species both in summer and winter, as revealed by high negative correlations with density in previous autumn and spring for summer and winter changes, respectively. These new finding of delayed density dependence (DDD) support the suggestion that vole cycles are generated by a time—lag mechanism. Possible mechanisms of the DDD are discussed. Regression analyses of rate of change in the different voles suggest that, besides the strong dependence on previous density, rate of change in numbers was also affected by current seed supply (in C. glareolus) and/or weather variables (temperature and precipitation sums) that may have affected the quantity or quality of food.