Vole Cycles, Snow Depth and Fox Predation

Abstract

Throughout the years, Oikos has recurrently lent its pages to discussions over Fennoscandian vole cycles in general and the role of predation in particular. One conceptual model that has emerged considers delayed density dependent predation during the decline to low phase by specialist predators as one possible driving force behind the cycles in the north, and an overall high predation by generalist predators as the cause of stability or noncyclicity in the south (Hansson 1979, 1987; see also Erlinge et al. 1983, 1984, 1988, and Erlinge 1987). An intermediate zone is characterized by weakly cyclic or semi-cyclic vole populations (Hansson and Henttonen 1985, 1988, Korpimaki 1986). The delayed predation pressure in the north is suggested to be partly due to low densities and low diversity of alternative prey, which cannot buffer the variations in vole abundance. For example, rabbit (Oryctolagus cuniculus L.), brown hare (Lepus europaeus, Pall.), and wood mice (Apodemus spp.) are alternative prey species that occur only or mainly in the southern parts of Sweden (Siivonen 1976). Consequently, the population density of the red fox (Vulpes vulpes L.) exhibits cyclical fluctuations in northem Sweden. In southern Sweden, the diet of foxes is more diverse (Angelstam et al. 1984), and the population density is more stable (Lindstrom 1982, Angelstam et al. 1985). Another important component of the assumed mechanism is the lack of a reliable snow-cover during winter in the south. This should increase the vulnerability of voles to predation. The average duration of snow-cover varies between 20 and 120 d in southern Sweden and between 150 and 220 d in most of northern Sweden (Angstrom 1974). Grimso Wildlife Research Area (59'40'N, 15'25'E) is situated in the intermediate zone concerning vole fluctuations. There, we studied field vole (Microtus agrestis L.) fluctuations by snap-trapping, and red fox diet by scat and stomach analyses, during 1973-1992. According to the model, we should expect a cyclic-like pattern of fluctuations in which delayed density dependent predation by foxes and other predators plays a role. However, this pattern might be disturbed by an increased overall predation pressure during winters of little snow. Field voles did indeed exhibit a fluctuation pattern resembling regular 3-4-yr cycles, though with variable peak densities (Fig. 1). However, the interrupted increase during 1980-81 and the regular winter decreases and low densities during the last 3 yr of the series (1989-1992) all coincided with winters of little snow (measured on 15 Feb., Fig. 1). In fact, over-winter population change, calculated as spring index/previous autumn index, was significantly positively related to snow depth when adjusted for increase, decrease or no change from previous to present spring (multiple linear regression: n = 14 [5 cases omitted: autumn index = 0; 0 or 1 vole caught following spring]; R2 = 0.58; total F = 7.70, P = 0.004; snow depth F = 4.82, P = 0.025; spring-spring change F=7.47, P=0.01, Fig. 2). All P-values in this text are 1-tailed. The density of adult foxes (indexed during spring) in the Grimso area tracked field vole density, but with a time lag of one year (Lindstr6m 1989). The fluctuations in density were primarily due to variations in reproductive success. Field voles constituted 66% of all small rodents determined to species in fox scats during 1975-1979 (Lindstrom 1982). The only other species that occurred in significant amounts (27% of all small rodents) was the bank vole (Clethrionomys glareolus Shreb.). Both these species exhibited synchronous population fluctuations in the Grimsb area as well as in northern Sweden (Hornfeldt 1991, in press). Field voles were equally important in fox stomachs, whereas bank voles and mice (Apodemus spp. and Mus musculus L.) each constituted 13% of all small rodents. During 1982-1986, the predominance of field voles in the diet of foxes was still higher (ratio field