RODENT CYCLES IN RELATION TO BIOMASS AND PRODUCTIVITY OF GROUND VEGETATION AND PREDATION IN THE PALEARCTIC

We have synthesized the data on population dynamics and densities of r odents in seven biomes of the Palearctic (mainly western part), and re lated them to the data on standing crop of biomass and net productivit y of ground vegetation (as rough indicators of food availability to ro dents). Analysis of 44 long-term (greater than or equal to 5 years) se ries of rodent trapping showed that there was a continuum from highly cyclic to non-cyclic populations. Rodents inhabiting tundra, taiga, st eppe, and farmlands (wintercrops) in the temperate zone have highest c yclicity indices. Definitely non-cyclic are rodents in the temperate f orests (mixed and deciduous forests, steppe woodland) and desert. Stan ding crop of biomass of ground vegetation (analysis of 63 data points) correlated positively with latitude; it was highest in the northern t undra and decreased towards South. Variation within biomes was most pr onounced in the temperate zone, with forests having ground vegetation biomass as low as that in deserts, whereas farmlands in that zone - as high as that in tundra. In various habitats (natural open, farmland, and forested) located in seven biomes, the mean index of rodent cyclic ity was significantly positively correlated to the mean standing crop of ground vegetation. Net productivity of ground vegetation (30 data p oints) did not show latitudinal trends. It was lowest in desert, tundr a, and all types of forests, and highest in open habitats of the tempe rate zone and steppes. Mean densities of rodents (calculated as averag ed spring and autumn estimates) were lowest in tundra, desert, and all types of forests (8-29 rodents/ha). The highest average densities wer e recorded in the farmlands of temperate zone and steppe (143-490 rode nts/ha). Mean and maximum densities of rodents were strongly positivel y correlated with the mean productivity of ground vegetation. Dichotom y between seasonal (non-cyclic) and multiannual (cyclic) fluctuations in rodent numbers was not found. The magnitude of seasonal changes in rodent densities (from spring to autumn) was a continuous variable rel ated to the propitiousness of climate for plant growth. Irrespectively of the type of population dynamics, seasonal changes in rodent densit ies were small in the coolest and the hottest biomes (tundra, desert) and big in the temperate zone. Results of our long-term study on preda tion on rodents in the temperate deciduous forests did not support the hypothesis on the role of specialist and generalist predators in shap ing rodent dynamics. We found no qualitative difference between predat ory impacts by generalist and specialist predators. All predators exer ted the heaviest impact at low or moderate densities of rodents (inver sely density-dependent predation). Predation may be an important facto r of rodent mortality but it does not shape the pattern of rodent popu lation dynamics. Based on the observed vegetation-rodent correlations, we have proposed an interpretation of the mechanisms of rodent popula tion dynamics in the Palearctic biomes. A prerequisite for rodent cycl es to occur is abundant winter food, which enables rodents to continue an increase phase beyond one growing season (by winter breeding). Hab itats with mean standing crop of ground vegetation of over 4000 kg dry weight/ha in summer are expected to harbour cyclic populations of rod ents.