Sleep and mood are clearly interrelated in major depression, as shown by the antidepressive effects of various experiments, such as total sleep deprivation, partial sleep deprivation, REM sleep deprivation, and temporal shifts of the sleep period. The prevailing hypotheses explaining these effects concern the antidepressant potency of the suppression of either REM sleep or non-REM sleep. This issue is discussed in the light of present knowledge of the kinetics of non-REM sleep intensity, REM sleep production, and their interaction. Recent findings have led us to suggest that the suppression of non-REM sleep intensity is the common pathway in the set of experimental data on the antidepressant effects of sleep manipulations.