OBJECTIVE The comorbidity between panic disorder and major depression (MDD) in individuals has been amply documented. However, data from family studies to determine whether panic disorder and MDD aggregate separately or together in families have been inconclusive, in part because of the absence of studies with the full range of proband groups. This report presents results from a family study with the necessary mutually exclusive groups: panic disorder without MDD, panic disorder with MDD, MDD without panic disorder, and normal controls. METHODS Diagnostic information was obtained from 193 probands and 1047 of their adult relatives with the Schedule for Affective Disorders and Schizophrenia--Lifetime Version for Anxiety Disorders by direct interview, and/or from multiple informants, without knowledge of proband diagnoses. Best-estimate diagnoses were based on all available information by clinicians independently of data collection and without knowledge of probands' and other relatives' status. RESULTS Findings indicated the specific and independent transmission of panic disorder and MDD, the separation of panic disorder from MDD, and the nonfamilial nature of late-onset MDD. The pattern of results was unaffected by the use of different diagnostic criteria, number of informants, interview status of relatives, presence of substance abuse or agoraphobia or the sequence of MDD and panic disorder in probands, or whether probands were selected from treatment clinics or community sample. CONCLUSIONS We conclude that panic disorder and MDD are separate disorders with substantial co-occurrence in individuals, and that panic comorbid with MDD is not a single, distinct disorder. Finally, we illustrate an approach to examining comorbidity in family data through analysis of mutually exclusive, parallel diagnoses in probands and relatives.