Three previous studies at Chernobyl, Ukraine, documented elevated mitochondrial DNA diversity in bank voles (Clethrionomys glareolus) from radioactively contaminated sites. Little evidence was found to link patterns of diversity in contaminated areas to radiation exposure, but the experimental design precluded discriminating among alternative explanations for elevated diversity in exposed groups. Reference sites selected for the studies were relatively distant from contaminated sites and, additionally, were separated from contaminated sites by large river systems; thus, we hypothesized that differences among sites were correlated with geographic isolation rather than with radiation exposure. For the present study, we added three reference sites, which were selected based on minimal radioactive contamination, proximity to contaminated sites, and absence of obvious barriers to dispersal. We hypothesized that neighboring reference sites should exhibit levels and patterns of diversity similar to those of contaminated sites if the previously detected differences were, in fact, caused by geographic isolation. Indeed, levels of diversity in nearby reference sites are comparable to levels in contaminated sites. Additionally, nearby reference sites contain several haplotypes not observed at other study sites. Our results suggest that levels of diversity in contaminated regions are more plausibly explained by ecological and historical factors than by increased mutational pressure resulting from exposure to Chernobyl radiation. © 2007 SETAC.