This paper describes the clinical methodology and currently collected pedigrees from an ongoing genetic study of panic disorder. The main objectives are to (1) document the clinical aspects of the study for current [Knowles et al., 1998: Am. J. Med. Genet. (Neuropsychiatr. Genet.) 81:138-147] and future reports of genetic analysis; (2) assist other investigators working on the genetic aspects of panic disorder who wish to compare results; and (3) illustrate the numerous judgment calls required in such studies that may lead to methodological variability and could account for differences in findings between studies. We also describe initial strategies to identify more genetically homogeneous panic disorder subtypes. Families were recruited through letters to magazines, word of mouth referral, and screening in anxiety disorder clinics and were asked to participate if at initial diagnostic screening they appeared to have at least three members (in two generations) affected with panic disorder. Diagnostic evaluations included a lifetime clinician-administered semistructured psychiatric interview, family history assessment, and pertinent medical records. Diagnoses for linkage analysis are derived from a best-estimate procedure that includes independent review of all materials by two senior investigators. All clinical evaluations were done blind to genotypes. Fifty-five pedigrees including 679 individuals have been collected to date. DNA is available from 500 family members of whom almost half (48%) are definitely or probably affected with panic disorder. Most (93%) of these subjects were directly interviewed, and the clinical data include not only lifetime psychiatric and medical diagnoses and but also detailed narrative histories describing sequence and context of symptoms. Family sizes range from 4-36 individuals (mean = 12.3) and the number of affected individuals per family from 2-12.