Holloway and McNally (1987) found that normals with high scores on the Anxiety Sensitivity Index (ASI), an instrument developed to assess beliefs regarding the adverse consequences of anxiety, reported more anxiety and more frequent and intense somatic sensations following hyperventilation than did normals with low scores on the ASI. They concluded that this result provides support for the construct validity of the ASI and thus for the construct of anxiety sensitivity. Nevertheless, we argue that (a) the developers of the ASI have conflated beliefs regarding the adverse consequences of anxiety with fear of these consequences, (b) the accumulated evidence for the construct validity of the ASI is weak, and (c) Holloway and McNally's design and analyses do not permit them to exclude the more parsimonious explanation that trait anxiety accounts for their findings. Implications for research on anxiety sensitivity are discussed.