Mouse population eruptions in New Zealand forests: the role of population density and seedfall

1. Sporadically eruptive house mouse populations provide an opportunity to test ideas about how intrinsic and extrinsic factors interact to limit rode nt populations. New Zealand house mouse populations inhabiting beech forest s and associated mosaics erupt when autumn beech seedfall is high. We analy sed 25 years of data on mouse density and beech seedfall from a beech and p odocarp/hardwood forest mosaic in New Zealand to test the effects of mouse density, beech seedfall and rat density on rates of change in mouse abundan ce, estimated as the rate of change in a density index for sequential seaso nal transitions (r). 2. For autumn to winter and winter to spring transitions, I was related to the magnitude of autumn seedfall and mouse density at the start of each tra nsition, regardless of whether forest types were considered separately or t ogether. For spring to summer and summer to autumn transitions, r. was rela ted to mouse density alone. Density dependence could reflect the influence of intrinsic social factors, intraspecific competition for food, or the eff ects of parasites, diseases or predators. 3. We modified a conceptual model of variation in rodent reproductive effor t to derive a general model of how numerical responses to density-independe nt increases in food availability, modified by density-dependent mechanisms and extrinsic precursors to population increase, drive sporadic eruptions in the density of some rodent populations. For house mice in New Zealand, t he numerical response to seedfall was modified by some unidentified density -dependent mechanism(s) to generate sporadic eruptions. Extrinsic precursor s to population increase appear to be important in other eruptive house mou se populations. 4. House mouse density in beech forest was more volatile than in podocarp/h ardwood forest, maximum rates of increase and decrease being higher in the former. Density tended to be higher in podocarp/hardwood forest when overal l density was low, and higher in beech forest when overall density was high . This suggests that between eruptions, mice persist in podocarp/hardwood f orest at higher densities than in adjoining beech forest.