The 1 mg and 2 mg dexamethasone suppression tests (DST) were evaluated in two groups of endogenously depressed patients (n = 39 and n = 30, respectively) who also had a 1300-1600 hr basal cortisol assessment. Non-suppressors (on both DSTs) had significantly higher basal plasma cortisol levels and thus were significantly associated with relative cortisol hypersecretion. However, there was only a partial overlap between DST response and basal plasma cortisol, with a large variation of cortisol levels among non-suppressors. The 2 mg DST appears to be more specific for cortisol hypersecretion than the 1 mg DST. If cortisol hypersecretion is to be identified, neither the 1 mg or 2 mg DST is an adequate assessment nor a substitute for a basal cortisol assessment.