Physiological exploration has disclosed profound disturbances in the regulation of sleep in depression. The finding that depression can be relieved or intensified by manipulation of sleep has inspired the investigation of the possible pathogenetic significance of these dysregulations. Three hypotheses play a leading role in this context: the "phase-advance", the "S-deficiency" and the "acetylcholine-monoamine imbalance" hypothesis. They explain the therapeutic effects of a variety of sleep wake manipulations as consequences of the normalization of depressogenic sleep regulation disturbances. The issue discussed in this paper is whether these claims are valid. To what extent can the hypotheses explain the existing sleep physiological disturbance in depression as well as the results of therapeutic interventions in sleep? The empirical data provide firm evidence for the importance of particularly the timing of sleep in relation to the course of depression, both in respect to the spontaneous and the experimentally induced fluctuations of the clinical state. Without considerable extensions none of the hypotheses provide an adequate explanation of these facts.