A dexamethasone suppression test (DST) and a thyrotropin releasing hormone stimulation test (TRHST) were given to 100 affectively ill inpatients with a mean age of 54.8 years and 16 healthy controls matched for age and sex. Of the affectively ill patients, 54 had primary major depressive disorders. Sensitivity and specificity, respectively, were 41% and 100% for DST; 44% and 88% for blunted TRHST; and 24% and 94% for augmented TRHST. The combined sensitivity for all three responses was 87%. DST nonsuppression discriminated between major and minor depression and between unipolar endogenous and unipolar nonendogenous subtypes. However, it failed to discriminate among primary depression, depression secondary to serious medical illness, or organic brain syndrome with depression. A blunted TRHST response was significant only for unipolar major depressives. Augmented TRHST response was significant only for bipolar depressives, suggesting that the TRHST may discriminate bipolar from unipolar depression.