DSM IV is a simple, reliable diagnostic system with many advantages. However, DSM diagnostic criteria may not provide sufficient characterization of clinically significant symptoms. We have undertaken a project to assess an array (spectrum) of clinical features associated with different DSM Disorders. The purpose of this paper is to report on reliability of assessment instruments for Panic-Agoraphobic Spectrum (PAS), to document convergent validity of PAS symptom groupings, and to confirm the relationship between PAS and DSM IV Panic Disorder (PD). We studied 22 normal controls and 95 outpatients who met criteria for Panic Disorder with and without lifetime Major Depression, and Major Depression or Obsessive Compulsive Disorder without lifetime Panic Disorder. Assessment instruments had excellent reliability and there was good concordance between interview and self-report formats. PAS scores were highest in subjects with PD, followed by outpatients without PD, and were lowest in normal controls. PAS scores varied among PD patients, and a subgroup of patients without PD scored high on PAS. We conclude that PAS can be reliably assessed, and that it describes a valid, coherent constellation of features associated with DSM IV Panic Disorder, but providing additional important clinical information.