Due to their prevalence and relationship to agoraphobia, panic attacks are receiving considerable attention. The DSM-III requires that they include at least four associated symptoms and occur at least weekly for a diagnosis of panic disorder. This study of panic symptoms and frequency in 68 panic sufferers addresses the validity of these DSM-III requirements in terms of symptom patterns and phobic avoidance. Using chi-square, Fisher, and t-test analyses, this study found no clinical distinction between frequent (occurring at least weekly) and infrequent panic attacks. When comparing limited-symptom attacks (associated with less than four symptoms) with full-blown panic attacks, as expected, symptom frequencies were greater (p <.05) in full-blown attacks, with the exceptions of feelings of unreality and fear of dying, going crazy, or losing control. Phobic avoidance did not differ between limited-symptom and full-blown attacks. This study not only demonstrates the arbitrary nature of the DSM-III requirements for a diagnosis of panic disorder but recognizes the importance of limited-symptom attacks to the development of agoraphobia. © 1990 by William & Wilkins.