The significance of the alteration in circadian rhythm in endogenous depression is discussed. 23 patients with endogenous depression, 11 patients with neurotic depression and 23 healthy subjects were deprived of sleep for one night. Those with endogenous depression responded favourably though the duration of the improvement was variable. In those with neurotic depression the response was less reliable and when it occurred the improvement was less marked. In one patient the condition was made worse. Patients with endogenous depression showed a rise in systolic bloodpressure in the ophthalmic artery following sleep deprivation. This did not occur in those with neurotic depression nor in healthy subjects. Sleep deprivation is to be recommended for the treatment of endogenous depression both in the initial stage and in later stages when the condition has proved resistant to thymoleptic agents.