Sleep and circadian rhythms are the primary determinants of arousal state, and torpor is the most extreme state change that occurs in mammals. The view that torpor is an evolutionary extension of sleep is supported by electrophysiological studies. However, comparisons of factors that influence the expression of sleep and torpor uncover significant differences. Deep sleep immediately following torpor suggests that torpor is functionally a period of sleep deprivation. Recent studies that employ post-torpor sleep deprivation, however, show that the post-torpor intense sleep is not homeostatically regulated, but might be a reflection of synaptic loss and replacement. The circadian system regulates sleep expression in euthermic mammals in such a way that would appear to preclude multiday bouts of torpor. Indeed, the circadian system is robust in animals that show shallow torpor, but its activity in hibernators is at least damped if not absent. There is good evidence from some species, however, that the circadian system plays important roles in the timing of bouts of torpor.