RODENT POPULATION-DYNAMICS IN A PRIMEVAL DECIDUOUS FOREST (BIALOWIEZA-NATIONAL-PARK) IN RELATION TO WEATHER, SEED CROP, AND PREDATION

Population dynamics of forest rodents, bank voles Clethrionomys glareo lus (Schreber, 1780) and yellow-necked mice Apodemus flavicollis (Melc hior, 1834) were studied in 1959 - 1991 in the pristine mature forest dominated by hornbeam Carpinus betulus and oak Quercus robur in Bialow ieza National Park (eastern Poland). The whole 33-year series, and par ticularly the detailed data from 1971 - 1991, were related to weather data and seed crop of hornbeam, oak, and maple Acer platanoides. Roden t numbers were very low in spring (April), then grew through summer du e to reproduction. The annually highest numbers of both species were r ecorded in autumn. Winter mortality was on average 77% of autumn numbe rs of voles and 86% of mice. Multiannual variations in numbers were la rge; the combined numbers of voles and mice showed a regular pattern o f 2 years of outbreak-crash (triggered by heavy mast production) and 4 - 7 years of moderate, though variable, densities. Heavy seed crops ( synchronous in oak, hornbeam and maple) occurred at 6-9-year intervals (in 1958, 1967, 1976, 1982, and 1989), and were triggered by warm Jun e - July temperature in the preceding year (bud formation year). Outbr eaks of rodents were always preceded by winter breeding in mice and so metimes also in voles. Rodents reached highest numbers in autumn of th e year following the mast peaks. Then, they declined rapidly over wint er to extremely low numbers in the following spring, summer and autumn . Such crashes were recorded after 4 out of 5 outbreaks. In moderate y ears, summer and autumn numbers of rodents correlated with food-relate d factors (seed crop, temperature affecting vegetation biomass), where as spring numbers were shaped by density dependent winter mortality. S ummer (July) numbers of mice in moderate years were a direct function of spring numbers of overwintered adults (R2 = 91%). Autumn (September ) numbers of mice were determined by tree seed crop of the previous ye ar (R2 = 32%). Summer numbers of voles depended on temperature in June - July (R2 = 29.5%), which most probably acted through an increased p roduction of herbaceous vegetation biomass. No ambient factors were fo und to explain variation in autumn numbers of voles. In both species, the intensity of reproduction in autumn was inversely related to autum n numbers. In bank voles, high temperature in July - September was con ducive to prolonged breeding. Winter mortality of rodents was density dependent (R2 = 99% in mice and 92% in voles). Mast increased overwint er survival of mice but not voles. Snow cover increased survival of ba nk voles. Density-dependent predation was the main agent of rodent mor tality in moderate years, whereas its role in outbreak-crash years has not yet been studied. The described pattern of population dynamics of forest rodents is regarded as typical for the deciduous forests with Quercus in lowland, temperate Europe. Review of literature showed that the heaviest seed crops of oaks (and beech) and the subsequent outbre aks of forest rodents occurred synchronously from Kazan region (Russia ) in the north-east to Oxford (England) in the west. Destruction of na tive forests made the described pattern of rodent population dynamics a relic that remained in scattered patches of the formerly widespread oak-hornbeam and oak-beech forests of Europe.