Three discriminant scales for diagnosing endogenous depression were examined in 56 depressed inpatients. These scales were the Newcastle 1 (N1), Newcastle 2 (N2), and Michigan Index (MI). The scales agreed on diagnosis in 17 (30%) patients; respective frequencies of endogenous depression were 23%, 57%, and 78% for the N2, N1, and MI scales. Relationships were examined between treatment response to isocarboxazid, drug dose, and diagnostic type for each scale. Significant dose X diagnosis interactions were noted for N1 and N2, but not for MI. A correlational within-dose analysis of all three scales revealed that actual diagnostic score was related to outcome only for high-dose patients on the Newcastle-1 scale; no other significant correlations emerged.