Bats are not squirrels: Revisiting the cost of cooling in hibernating mammals.

Many species use stored energy to hibernate through periods of resource limitation. Hibernation, a physiological state characterized by depressed metabolism and body temperature, is critical to winter survival and reproduction, and therefore has been extensively quantified and modeled. Hibernation consists of alternating phases of extended periods of torpor (low body temperature, low metabolic rate), and energetically costly periodic arousals to normal body temperature. Arousals consist of multiple phases: warming, euthermia, and cooling. Warming and euthermic costs are regularly included in energetic models, but although cooling to torpid body temperature is an important phase of the torpor-arousal cycle, it is often overlooked in energetic models. When included, cooling cost is assumed to be 67% of warming cost, an assumption originally derived from a single study that measured cooling cost in ground squirrels. Since this study, the same proportional value has been assumed across a variety of hibernating species. However, no additional values have been derived. We derived a model of cooling cost from first principles and validated the model with empirical energetic measurements. We compared the assumed 67% proportional cooling cost with our model-predicted cooling cost for 53 hibernating mammals. Our results indicate that using 67% of warming cost only adequately represents cooling cost in ground squirrel-sized mammals. In smaller species, this value overestimates cooling cost and in larger species, the value underestimates cooling cost. Our model allows for the generalization of energetic costs for multiple species using species-specific physiological and morphometric parameters, and for predictions over variable environmental conditions.