Density dependence in fluctuating grey-sided vole populations

1. The presence of direct and delayed (1 year) density dependence in fluctuating grey-sided vole populations (Clethrionomys rufocanus) in Hokkaido, Japan, were tested using 90 time series spanning 31 years (1962-92). 2. The autocorrelation test of Bulmer, the randomization test of Pollard, Lakhani and Rothery, and the bootstrap test of Dennis and Taper using two different models (Ricker and Gompertz) detected direct density dependence in most of the time series (90-0 100%) at the 5% level (one-tailed test). 3. Plotting population growth rates as functions of abundance suggested that the Gompertz model is more appropriate than the Ricker model for the studied populations. The tests for direct density dependence using the Gompertz model also rejected the null hypothesis of density-independent growth more frequently than the tests using the Ricker model. 4. We extended the randomization and the bootstrap methods to investigate delayed density dependence by using partial correlation and partial regression, respectively. The extended tests detected delayed density dependence in 8-15 time series (8.9%- 16.7%) at the 5% level (two-tailed test). 5. The high incidence of direct density dependence and the low, but significant, incidence of delayed density dependence in these vole populations are discussed. Delayed maturation at high densities by social interference and a rich generalist fauna are thought to represent plausible mechanisms generating direct density dependence. As a candidate mechanism for delayed density dependence, specialist predator and climate conditions (affecting the exposure to predators) are discussed.