Two theories of the effects of loss of sleep were compared against existing facts. The hypothesis that sleep deprivation reduces arousal was judged to be more adequate than the theory that arousal is increased by loss of sleep, the former failing only to account for raised physiological indices of arousal in some experiments. These are attributed to experimental conditions which compel the subject to make compensatory efforts to maintain adequate levels of performance. Under such conditions effort, not arousal, is reflected in the indices.An experiment was conducted in which a constant high level of performance was not demanded of the subjects over a 60 hr. period without sleep. Heart rate and performance fell with loss of sleep.