In 35 hospitalized depressed patients afternoon plasma cortisol (15:00-15:20 h) was measured in 3 ways: spontaneous, 2 h after administration of 45 mg oxazepam, and 16 h after administration of 2 mg dexamethasone. Dexamethasone and oxazepam caused a marked suppression of cortisol secretion, the former being the most pronounced. A significant correlation between spontaneous and suppressed cortisol levels was found and a correlation between the suppressed cortisol levels was also demonstrated. Both spontaneous and suppressed cortisol values correlated with measures of type (Newcastle scales) and severity (Hamilton rating scale) of depression, and the highest correlations were obtained with cortisol values after dexamethasone suppression. Analysis of the items of the Hamilton rating scale for depression showed a high correlation with retardation (items 7 + 8) and somatic complaints (items 12 + 13). A high correlation with the items 8 + 9 + 10 + 12, as found in an earlier study, was also found in this study. On the basis of these findings it is concluded that hypersecretion of cortisol measured spontaneously or after suppression is not a true diagnostic measure (within a group of depressed patients), but reflects differences in the intensity of certain psychiatric symptoms constituting the depressive syndrome.