Testing the thermal melanism hypothesis: a macrophysiological approach