In this paper we present the results of an investigation designed to measure diurnal variation of mood in a group of unipolar depressives using a sensitive self-assessment inventory and a behavioral measure of the depth of depression. Both measures correlate well with the Hamilton Rating Scale which was designed to quantify depression during a psychiatric interview. Although all patients included in our sample reported that they experienced diurnal variation in mood, this fluctuation was not manifest on behavioral measures and only one patient showed a significant difference on the depressive inventory. Our data indicate that self-report and behavior may not be affected by diurnal variation in eight of these nine depressed patients. One possible explanation is that self-report in depression is subject to perceptual distortion leading to retrospective falsification.