One of the most influential ideas in schizophrenia research is that schizophrenia may be a syndrome with significant pathophysiological heterogeneity, rather than a single disease. Among patients with schizophrenia, presence or absence of the deficit syndrome has been suggested as a method for defining relatively homogeneous groups. The criteria for the deficit syndrome require the presence of negative symptoms that are judged primary to the illness, rather than to factors, such as depressive mood, that may resemble the negative symptoms of schizophrenia. To test one aspect of the validity of the primary/secondary judgment, we tested the relationship of depressive symptoms to the deficit/nondeficit categorization. Using independent clinician ratings, the depressive mood of deficit and nondeficit patients was compared at the time the categorization was made, and at an average follow-up of 2 1/2 years. Using patients' self ratings, deficit and nondeficit patients were compared at an average follow-up of 1 1/2 years. Deficit patients had significantly less severe depressive symptoms by clinicians' ratings both cross-sectionally and at follow-up, and less severe self-rated symptoms at follow-up. These differences were not due to confounding by age, race, sex, socioeconomic status, or chronicity. These results support the validity of the deficit/nondeficit categorization.