Density-dependent life histories in female bank voles from fluctuating populations.

Using long-term data sets from two lowland floodplain forest bank vole populations in central Europe, we tested two predictions that with increasing densities: (i) proportions of nonparous females in winter populations would increase; and (ii) age of both nonparous and parous females would increase. These two predictions follow from the assumption that changes in age structure are driven by density-dependent shifts in age at first reproduction. Both populations were sampled by snap trapping between 1956 and 1976. For each year, we examined samples of animals collected between November and April for proportions of females with and without breeding experience and analysed their variation in age relative to the population density. The presence or absence of placental scars was used to discriminate between the parous and nonparous animals. Age was determined by measuring the length of the first mandibular molar (M1 ) roots. Both populations exhibited multi-annual fluctuations in numbers closely resembling those in northern Fennoscandia. The proportion of nonparous females in our total sample was 0.73, suggesting that it is uncommon for parous female bank voles to breed in two successive years and that their life histories are largely designed for breeding in one season only as a major reproductive strategy. Using a logistic-binomial regression model, we found that the probability of females being nonparous or parous at capture varied significantly with time, space and population density. The final model producing the best fit to data predicted that the proportion of nonparous females would be slightly larger over winter and substantially larger after high-density breeding seasons, which is consistent with the tested prediction. With increasing densities during the breeding seasons, both the nonparous and parous females became older at the onset of winter. Again, the field evidence was consistent with the predicted pattern. The age of parous females in the autumn at high densities was older, mainly because of an increased proportion of females entering their second winter. We propose a new intrinsic mechanism by which a delayed density dependence in age-specific fitness components can arise in microtines. This mechanism, based on an increased bimodality in age at first reproduction towards northern latitudes, has a potential to explain the south-north gradient in cyclicity observed in microtine populations in Fennoscandia.