Converging evidence suggests that anxiety sensitivity (i.e., threatening beliefs regarding autonomic arousal) is a risk factor for anxiety pathology. Specification of premorbid risk factors requires exclusion of individuals with a history of spontaneous panic to ensure that anxiety sensitivity is not merely a consequence or concomitant of the experience of panic. However, the psychometrics and dimensional nature of anxiety sensitivity in such a sample is undetermined. The present study evaluated the factor structure of the Anxiety Sensitivity Index (ASI), a measure of anxiety sensitivity, in a community sample (N = 233) with no history of psychiatric illness or spontaneous panic. Exploratory factor analyses (EFA) suggested a two- or three-factor solution (1, Fear of Mental Catastrophe; II, Fear of Cardiopulmonary Sensations; III, Fear of Vasovagal Sensations). Confirmatory factor analyses (CFA) comparing alternative models indicated that a hierarchical two-factor solution (I, Fear of Mental Catastrophe; II, Fear of Cardiopulmonary Sensations) best accounted for the data. This model generalized well to a nonclinical college sample (N = 809).