Bats seldom soar because it is behaviour generally associated with the use of thermals, which are normally of insufficient strength at night to support the behaviour. Daylight flying bats, however, such as the Samoan flying fox Pteropus samoensis may be able to exploit thermals for soaring. This may give the bats one of two advantages. It may reduce the energy costs of transport because gliding flight is much cheaper than active flapping flight. However, because less endogenous heat is generated by soaring, a second advantage may be that it reduces the thermal stress placed on these bats. Thermal stress is a factor that we have shown previously probably constrains the daylight flying behaviour of this species. Observations of the patterns of soaring behaviour at two sites on American Samoa in March and October 1995 supported the predictions of the energy saving but not the hyperthermia avoidance hypothesis. Soaring was a common behaviour under all conditions and was used extensively when conditions did not pose a threat of hyperthermia. In March, the bats also adopted flight patterns over time that exposed them to areas of the valleys where insolation was greatest, presumably increasing their risk of hyperthermia but bringing energy saving benefits. Modelling the expected heat flows during soaring and flapping flight using an established model revealed that soaring reduced the risk of hyperthermia, when flying in the shade of clouds, because of the energy savings resulting from reduced endogenous heat production. However, when soaring in sunlight, these savings are more than offset by the increased exogenous heat uptake, because a greater proportion of the wing surface is exposed when soaring. Despite its low endogenous energy cost, soaring in sunlight is not thermally advantageous, and the behaviour of the bats reflected this fact.