Thermal benefits of clustering during hibernation:: a field test of competing hypotheses on Myotis sodalis

1. Clustering lessens heat loss in euthermic animals by decreasing exposed surface area. However, clustering is also used during heterothermic periods such as hibernation, when heat loss should be minimal because body temperature (T(b)) is near ambient temperature (T(a)). 2. We tested two hypotheses that attempt to explain the thermal benefit of clustering by cave-dwelling bats during hibernation: (i) clustering buffers individuals against fluctuations in T(a) ('thermal inertia hypothesis'), and (ii) clustering minimizes heat loss during energetically expensive arousals and subsequent periods of euthermy ('arousal hypothesis'). 3. We created path models parameterized with data obtained from thermal and digital images of naturally hibernating Myotis sodalis clusters and temperature-sensitive dataloggers placed near each cluster. 4. Akaike weights (omega(i)) indicate the arousal hypothesis has nearly complete support (omega(i) = 0.9999), while the thermal inertia hypothesis has little support (omega(i) = 0.0001). As predicted by the arousal hypothesis, mean T(a) is the strongest predictor of cluster size. Further, after accounting for mean T(a), cluster size is negatively related to variation in T(a). 5. Our results indicate that most of the thermal benefit of clustering during hibernation is from decreased heat loss during brief euthermic periods. Further, our results predict that clustering may lessen mass loss during hibernation in stable environments, may be minimized during hibernation when there are large fluctuations in the thermal environment, and that the propensity to cluster may be negatively related to body condition.