LAUSR

An optimal habitat-selecting organism should use a dispersal strategy that enables occupation of the habitat yielding greatest fitness. The strategy is complicated when habitat quality varies through time. Theory predicts that the long-term distribution of individuals will match mean habitat quality while undermatching current habitat quality. I tested the prediction with experiments on controlled populations of meadow voles occupying two pairs of field enclosures. I released equal numbers, and equal sexes, of voles in each enclosure, and varied resource abundance between enclosures by supplemental feeding. I measured the voles' response with giving-up densities (GUDs) in artificial foraging patches, and with live-trapping at the end of the experiment. The data were consistent with only one of four a priori dispersal models. Giving-up densities declined with resource supply because short-term supply had no effect on population density. GUDs were invariant to the time-course of the experiment because densities were proportional to each enclosure's long-term mean quality. Similar patterns in sex ratios and patterns of habitat occupation by juvenile voles born during the experiment reinforce the interpretation of time-averaged habitat matching. This study adds to the cumulating evidence that strategies of space use converge towards behavioral and evolutionary optima.