Learning from the games animals play: using behavior to assess spatial structure and stochasticity in natural populations

Population densities are heterogeneous across a variety of spatial scales. The variation in density reflects a similar variety of processes ranging fr om density-dependent habitat selection at small scales to independently reg ulated populations at much larger ones. I measured each scale with experime nts capitalizing on the behavior of individual deer mice foraging in badlan d habitats in Alberta, Canada. First, I used patterns in rodent density alo ng transects crossing badland and prairie habitats to measure the scale of habitat selection. Consistent with theoretical predictions, differences in the intercepts of isodars (graphs of density in adjacent habitats assuming ideal habitat selection) comparing prairie and badland densities revealed a maximum scale of habitat selection on the order of only 140 m. Second, I u sed foraging experiments to estimate density-dependent declines in fitness measured by the surrogate of giving-up-density of mice foraging in artifici al foraging patches. Habitat selection should tend to equalize giving-up-de nsities among replicated, but spatially segregated, grids containing differ ent numbers of foragers. Contrary to predictions from habitat selection the ory, giving-up-densities declined with increased forager density in the maj ority of grids. Giving-up densities in nine of 12 grids increased linearly as population density was reduced in 1997. Giving-up densities in eight of 10 grids increased linearly with resource supplements in 1998. The results of both experiments are consistent with independent resource harvest by var ying numbers of foraging mice. The identity of "outlier" grids, that showed little response to either manipulation, varied between years. The combined results document spatially-structured populations and allow us to estimate the frequency of stochastic dynamics that may have a profound influence on evolution and conservation strategies in heterogeneous landscapes.