|| Checking for direct PDF access through Ovid
In small rodent populations with wide-amplitude fluctuations and low-density bottlenecks, the individuals that survive through the bottlenecks may gain major fitness advantages as they will be the founders of the following population expansion. Most hypotheses assume that there exists a physical or behavioural refuge from increased predation risk, and that the survivors are most likely individuals adapted to use such refuges. A recent hypothesis suggests that survival probability is habitat-dependent so that some otherwise sub-optimal habitats provide a spatial refuge from predation risk by the main predator(s). We used spatially replicated long-term (1981–2004) trapping and tracking data of voles (field vole Microtus agrestis and sibling vole M. rossiaemeridionalis) and their main predators (weasel Mustela nivalis and stoat M. erminea) to test predictions based on this hypothesis. We did not find support for the hypothesis. We did not find marked phase-dependent differences in the habitat-level distribution of Microtus voles. Habitat types with low Microtus vole abundance had, on average, comparable predator activity than the main Microtus vole habitats, indicating that there were no habitat-level refuges from predators. There appeared to be no permanent site-level refuges: the spatial distribution of voles varied from one bottleneck to another. This suggests that survival through bottlenecks is at least partly determined by chance events. We propose that in this kind of systems, where relatively short-lived prey are hunted by nomadic or widely ranging predators, short-term anti-predator responses may increase survival prospects as efficiently as more costly anti-predator adaptations, and there is no apparent need to maintain special adaptations to bottleneck situations that occur at infrequent intervals.