In the course of several studies of hypothalamic-pituitary-adrenal axis activity in depression, 43 patients presented on separate admissions with definite depression and, on both admissions, received dexamethasone suppression tests (DSTs). DST results were discordant across admissions in 21% of cases; among patients who were nonsuppressors on either admission, results were discordant in 40.9%. The correlation between postdexamethasone values obtained on two admissions was highly significant, however. Distinctions between bipolar and unipolar depression and between primary and secondary depression by rates of nonsuppression were inconsistently significant across admissions but were clearer when results from both admissions were pooled and patients with an abnormal DST on either admission were considered nonsuppressors. While abnormal escape from dexamethasone suppression occurs in a significant proportion of depressed patients, this phenomenon may only partially overlap the depressive syndrome in time. Negative DST results in patients with primary depression must be interpreted in this light.