One hundred patients with postpartum mental illness were compared with a group of 100 female patients, matched for age and year of admission, who had been admitted to the same institution for mental illness not associated with childbirth. No significant difference in symptomatology was found. Postpartum mental illnesses are not a psychiatric entity; the term should be understood to refer simply to the usual psychiatric disorders but in the setting of childbirth. Current beliefs regarding etiology were found to need correction. The illness was not especially associated with first pregnancies; almost half of the patients had experienced childbirth before without recognizable mental illness. Among the most frequent precipitating factors were unstable marriages, long-standing maladjustment of the patient herself, immaturity of the patient, unstable family history of the patient, and the creation of new financial and other strains by the birth of the baby. The same factors were found important in determining the prognosis. The obstetrician's role includes early recognition of symptoms, use of nursing precautions to prevent suicidal and homicidal attempts, and use of a prognostic guide described by the authors. Textbooks of obstetrics need revision in the light of new facts about postpartum mental illness.