OBJECTIVE Although clinical trials have documented the importance of identifying individuals with major depression with atypical features, there are fewer epidemiological data. In a prior report, the authors used latent class analysis (LCA) to identify a distinctive atypical depressive subtype; they sought to replicate that finding in the current study. METHODS Using the National Comorbidity Survey data, the authors applied LCA to 14 DSM-III-R major depressive symptoms in the participants' lifetime worst episodes (N=2,836). Validators of class membership included depressive disorder characteristics, syndrome consequences, demography, comorbidity, personality/attitudes, and parental psychiatric history. RESULTS The best-fitting LCA solution had six classes. Four were combinations of atypicality and severity: severe atypical, mild atypical, severe typical, and mild typical. Syndrome severity (severe atypical and typical versus mild atypical and typical classes) was associated with a pronounced pattern of more and longer episodes, worse syndrome consequences, increased psychiatric comorbidity, more deviant personality and attitudes, and parental alcohol/drug use disorder. Syndrome atypicality (severe and mild atypical versus severe and mild typical classes) was associated with decreased syndrome consequences, comorbid conduct disorder and social phobia, higher interpersonal dependency and lower self-esteem, and parental alcohol/drug use disorder. CONCLUSIONS As in prior reports, the atypical subtype of depression can be identified in epidemiological samples and, like typical depression, exists in mild and severe variants. Atypical depressive subtypes were characterized by several distinctive features. However, the correspondence between epidemiologically derived typologies of atypical depression and DSM-IV major depression with atypical features is not yet known.