Small-rodent dynamics and predation

The hypothesis that the regular multiannual population oscillations of bore al and arctic small rodents (voles and lemmings) are driven by predation is as old as the scientific study of rodent cycles itself. Subsequently, for several decades, the predation hypothesis fell into disrepute, possibly bec ause the views about predation and rodent dynamics were too simplistic. Her e we review the work that has been done on the predation hypothesis primari ly in Fennoscandia over the past decade. Models of predator-prey interaction have been constructed for the least wea sel (Mustela nivalis) and the field vole (Microtus agrestis), which are con sidered to be the key specialist predator and the key prey species in the m ultispecies communities in the boreal forest region in Fennoscandia, The ba sic model has been parameterized with independent field data, and it predic ts well the main features of the observed dynamics. An extension of the mod el also including generalist and nomadic avian predators predicts correctly the well-documented and striking geographic gradient in rodent oscillation s in Fennoscandia, with the amplitude and cycle period decreasing from nort h to south. These geographic changes are attributed to the observed latitud inal change in the density of generalist and nomadic predators, which are e xpected to have a stabilizing effect on rodent dynamics. We review the other observational, modeling, and experimental results beari ng on the predation hypothesis and conclude that it accounts well for the b road patterns in rodent oscillations in Fennoscandia. We discuss the applic ation of the predation hypothesis to other regions in the northern hemisphe re. The predation hypothesis does not make predictions about multiannual an d latitudinal changes in body size, behavior, and demography of rodents, wh ich may have some population-dynamic consequences. With the current evidenc e, however, we consider it unlikely that the phenotypic and genotypic compo sition of populations would be instrumental for generating the broad patter ns in rodent oscillations.