The study reexamined the relationship between anxiety sensitivity (AS) and drinking behavior in a sample of university women (N = 197) taking into account recent theoretical considerations regarding the possible hierarchical nature of the AS construct. To assess the incremental validity of this construct, the amount of variance in drinking behavior accounted for by measures of AS and trait anxiety was investigated with several multiple regression analyses. Contrary to expectations, measures of both constructs failed to account for significant variance in drinking behavior. McNally's [1996: Rapee RM, editor. Current Controversies in anxiety disorders. p 214-244] hypothesis regarding a trait anxiety by AS interaction received weak support as the combination of these two variables accounted for a small (i.e., 1.6%) portion of variance in one drinking variable (viz., monthly intoxication). Correlational analyses revealed only two small and marginally significant associations between the lower-order components of AS (i.e., fears of physical, cognitive, and publicly observable symptoms of anxiety) and the drinking measures. Potential reasons for the failure to replicate or extend Stewart et al.'s [1995: J Anxiety Disord 9:283-392] findings of significant positive associations between AS and both weekly alcohol consumption and excessive drinking are discussed.