After we reviewed the literature to identify the clinical and phenomenologic correlates of neurotic depression, we constructed a 6-item operational definition to distinguish neurotic unipolar major depressive disorder from non-neurotic major depression. The neurotic depressives were characterized by a low rate of abnormal dexamethasone suppression test (DST) results and a strong family history of alcoholism. Neurotic depressives improved less than non-neurotic depressives during the index hospitalization, and were more frequently rehospitalized during a 6-month prospective follow-up. Neurotic subtyping was significantly negatively associated with DSM-III melancholia. Neurotic classification remained significantly associated with the above validating variables after melancholic status was held constant, whereas melancholic subtyping did not predict DST results, familial alcoholism rates, or outcome when neurotic status was controlled.