Population densities of the grey-sided vole Myodes rufocanus fluctuate greatly within and across years in Japan. Here, to investigate the role of individual dispersal in maintaining population genetic diversity, we examined how genetic diversity varied during fluctuations in density by analyzing eight microsatellite loci in voles sampled three times per year for 5 years, using two fixed trapping grids (approximately 0.5 ha each). At each trapping session, all captured voles at each trapping grid were removed. The STRUCTURE program was used to analyze serially collected samples to examine how population crashes were related to temporal variability, based on local-scale genetic compositions in each population. In total, 461 and 527 voles were captured at each trapping grid during this study. The number of voles captured during each trapping session (i.e., vole density) varied considerably at both grids. Although patterns in fluctuations were not synchronized between grids, the peak densities were similar. At both grids, the mean allele number recorded at each trapping session was strongly, positively, and non-linearly correlated with density. STRUCTURE analyses revealed that the proportions of cluster compositions among individuals at each grid differed markedly before and after the crash phase, implying the long-distance dispersal of voles from remote areas at periods of low density. The present results suggest that, in grey-sided vole populations, genetic diversity varies with density largely at the local scale; in contrast, genetic variation in a metapopulation is well-preserved at the regional scale due to the density-dependent dispersal behaviors of individuals. By influencing the dispersal patterns of individuals, fluctuations in density affect metapopulation structure spatially and temporally, while the levels of genetic diversity are preserved in a metapopulation.

For the genotype data,

1. Liver samples were collected three times per year (late May, early August, and early October) for 5 years (2002–2006) at two fixed sites on Nemuro Peninsula, Japan (grids A and I). During each trapping session, 50 Sherman-type traps were arranged in a 5 × 10 grid pattern at an interval of 10 m (~0.5 ha) for 3 nights (i.e., a total of 150 trap-nights per session).

2. Genomic DNA was extracted using the conventional phenol-chloroform method. Eight microsatellite loci (MSCRB-01, -04, -06, -07, -09, -10, -11, and -13) were amplified with ABI AmpliTaq Gold DNA polymerase and an ABI 2720 Thermal Cycler system.

3. Genotyping was performed with an ABI PRISM 310 Genetic Analyzer using GeneScan Analysis 3.1.2 and Genotyper 2.5.

There is a missing value in a single individual at grid I (ID: 2006L136; Locus: MSCRB11).