Duration judgements for intervals of different lengths and content were studied in depressive in-patients (n = 47) and a control sample of surgical in-patients (n = 16). As suggested by research on non-clinical subjects, tasks during the intervals influenced the depressed patients' duration judgements. Severely depressed endogenous depressives (n = 17) over-estimated time when left completely unoccupied or when attending to tasks requiring concentration. Endogenous depressives (n = 17) remitted with regard to subjective depression but, exhibiting signs of psychomotor retardation, selectively over-estimated time when required to concentrate under time pressure. Neurotic/reactive depressives (n = 13) with an intermediate level of subjective depression and almost normal psychomotor functioning did not over-estimate any of these intervals. Time estimations of patients and controls did not differ for intervals in the range of seconds and minutes requiring attention to time only, and for a longer part of the experimental session. Alteration of time estimation and results of a time experience inventory corresponded for endogenous depressives but not for neurotic/reactive depressives. Results are discussed in terms of the influence of affective state and subjective concentration effort on the over-estimations observed.