Abstract
"Bottom-up" influences, that is, masting, plus population density and climate, commonly influence woodland rodent demography. However, "top-down" influences (predation) also intervene. Here, we assess the impacts of masting, climate, and density on rodent populations placed in the context of what is known about "top-down" influences. To explain between-year variations in bank vole Myodes glareolus and wood mouse Apodemus sylvaticus population demography, we applied a state-space model to 33 years of catch-mark-release live-trapping, winter temperature, and precise mast-collection data. Experimental mast additions aided interpretation. Rodent numbers in European ash Fraxinus excelsior woodland were estimated (May/June, November/December). December-March mean minimum daily temperature represented winter severity. Total marked adult mice/voles (and juveniles in May/June) provided density indices validated against a model-generated population estimate; this allowed estimation of the structure of a time-series model and the demographic impacts of the climatic/biological variables. During two winters of insignificant fruit-fall, 6.79 g/m(2) sterilized ash seed (as fruit) was distributed over an equivalent woodland similarly live-trapped. September-March fruit-fall strongly increased bank vole spring reproductive rate and winter and summer population growth rates; colder winters weakly reduced winter population growth. September-March fruit-fall and warmer winters marginally increased wood mouse spring reproductive rate and September-December fruit-fall weakly elevated summer population growth. Density dependence significantly reduced both species' population growth. Fruit-fall impacts on demography still appeared after a year. Experimental ash fruit addition confirmed its positive influence on bank vole winter population growth with probable moderation by colder temperatures. The models show the strong impact of masting as a "bottom-up" influence on rodent demography, emphasizing independent masting and weather influences; delayed effects of masting; and the importance of density dependence and its interaction with masting. We conclude that these rodents show strong "bottom-up" and density-dependent influences on demography moderated by winter temperature. "Top-down" influences appear weak and need further investigation.