Eliciting information about a patient's family history is a component of a standard diagnostic evaluation. Since depression runs in families, a question arises as to whether family history information should be considered when diagnosing depression. The text of DSM-IV indicates that a family history of a mood disorder should sometimes be considered when trying to distinguish between major depressive disorder and other conditions such as catatonic schizophrenia. The questions posed herein are how well family history of depression performs as a diagnostic criterion, and how its performance compares with the DSM-IV symptom criteria. One thousand eight hundred psychiatric outpatients were evaluated with a semistructured diagnostic interview as part of a research assessment infrastructure that has been embedded in the Rhode Island Hospital Department of Psychiatry outpatient practice. Family history diagnoses were based on information provided by the patient. The interview followed the guide provided in the Family History Research Diagnostic Criteria. We constructed a continuum of family history morbidity based on the number of first-degree family members with a history of depression and whether the family member was treated for their depression. Thus, we determined the presence or absence of the family history diagnostic criterion in different ways. Family history information was collected on 9763 first-degree relatives of 1776 patients. The sensitivity of the family history criterion was lower than each of the symptoms. Based on the broadest definition of the family history variable, the specificity was also lower than all other symptoms. Based on the narrowest definition (two or more family members who were treated for depression), the specificity was higher than all of the symptom criteria though sensitivity dropped to 15%. Overall, as a diagnostic criterion, a family history of depression did not perform as well as the DSM-IV symptom criteria. Consistent with the familial nature of depression, the family history variable performed better as a diagnostic criterion when considering diagnosis from a lifetime, rather than a current, perspective. This has implications for the future consideration of genetic markers as diagnostic criteria.