Our ability to link microecology and macroevolution depends on whether we can study speciation and extinction with small-scale experiments. Such experiments will likely be most successful if they use metrics closely related to population growth rates while enabling behavioural choices that either enhance speciation or forestall extinction. I designed field experiments with meadow voles that combined observations of foraging behaviour and dispersal from rapidly deteriorating habitats. Voles emigrated from low to higher-quality habitat as expected, but did not match densities with those expected from supplemental food. Movement and giving-up densities suggest that the voles’ habitat selection was modulated by sex-dependent differences in dispersal conditioned by additional components of fitness including predation risk. Even so, carefully designed behavioural experiments have a demonstrated ability to inform processes of speciation and extinction and thereby provide crucial insights into macroevolution and the conservation of Earth’s biodiversity.