The importance of disturbances in time experience in endogenous depression (melancholia) has long been under discussion. In the present study alterations of time experience were examined by means of time estimation experiments. 25 endogenous depressive patients (according to ICD-9) underestimated prospectively a 30-second interval by 6 s, whereas 12 healthy controls overestimated this interval by more than 10 s. The results are based on eight successive measurements over a period of 2 days. In the endogenous depressives, a feeling of being unwell was accompanied by a more pronounced time estimation error than a relatively good state of well-being. The influence of various patient data on the results of time estimation is also discussed in detail. The findings show clearly that time estimation provides a satisfactory delineation between depressive patients and healthy controls.