The purpose of this study was to test the theory put forth by Zvolensky et al. [Clin. Psychol. Sci. Pract. 10 (2003) 29] that smoking is specifically associated with panic disorder (PD) and not more generally associated with other anxiety disorders. Smoking behaviors were examined across three anxiety disorders: panic disorder with or without agoraphobia, social phobia (SP), and obsessive-compulsive disorder (OCD). A greater proportion of the PD group (40.4%) reported smoking compared to the SP (20%) and OCD (22.4%) groups. Those in the PD group were also more likely than those in the other groups to report being a heavy smoker (greater than 10 cigarettes daily). No significant interaction between diagnosis and smoking status was found for any of the symptom measures. However, participants who smoked had significantly higher scores than nonsmokers on a number of scales, including measures of depression, general anxiety, and distress. Differences in anxiety sensitivity between smokers and nonsmokers approached significance. These findings provide support for Zvolensky et al.'s [Clin. Psychol. Sci. Pract. 10 (2003) 29] theoretical conceptualization and suggest a specific link between smoking and panic disorder. Further investigation is warranted to determine the causal direction of this association.