PREDATION ON COMPETING RODENT SPECIES - A SIMPLE EXPLANATION OF COMPLEX PATTERNS

1. We extend a predator-prey model previously parameterized for voles and weasels to two prey species, a 'Microrus' (field vole) type which is competitively superior to but more vulnerable to predation than a ' Clethrionomys' (bank vole) type. 2. The model explains four patterns i n the dynamics of multispecies rodent assemblages in Fennoscandia: a p redictable shift in the relative abundances of different prey species during one multiannual population cycle; long-term (supracyclic) varia tion in relative prey abundances; an association between the amplitude of population oscillations and the type of the numerically dominant p rey species; and increasing rodent species number with increasing lati tude, 3. The model results illustrate the complex and often unexpected behaviour of strongly connected multispecies assemblages, of which th e Fennoscandian rodent-predator community is a prime example. 4. Since the mid 1980s, rodent oscillations in many, though not all, parts of northern Fennoscandia have become distinctly less regular (non-cyclic) , a change which is reflected in the entire animal community linked to the keystone species. the arvicoline rodents, We demonstrate that suc h long-term changes in the amplitude and regularity of rodent oscillat ions are not unexpected in multispecies prey-predator assemblages.