Posttraumatic stress disorder (PTSD) and alcohol use disorders are frequently comorbid conditions (Stewart, 1996). Alcohol use may serve a “negatively-reinforcing” function among traumatized individuals with PTSD (Stewart, 1996; Stewart, Conrod, Pihl, & Dongier, 1999a; Stewart, Pihl, Conrod, & Dongier, 1998). As such, the heavy drinking behavior of those with PTSD should be relatively situation-specific (i.e., more frequent in “negative” discriminative contexts than in other types of contexts). To test this “situational-specificity” hypothesis, a lifetime measure of trauma exposure (Everstine & Everstine, 1993), the PTSD Symptom Self-Report Scale (Foa, Riggs, Dancu, & Rothbaum, 1993), and the 42-item Inventory of Drinking Situations (Annis, Graham, & Davis, 1987) were administered to a community-recruited sample of 294 adult women substance abusers. PTSD symptoms were significantly positively correlated with frequency of heavy drinking in negative situations, but unrelated to frequency of heavy drinking in positive and temptation situations. At the level of specific drinking situations, PTSD symptoms were significantly positively correlated with frequency of heavy drinking in the negative situations of Unpleasant Emotions, Physical Discomfort, and Conflict with Others. PTSD symptoms were unrelated to frequency of heavy drinking in the positive situations of Pleasant Times with Others and Social Pressure to Drink, or in the temptation situations of Testing Personal Control and Urges and Temptations. Additionally, PTSD symptoms were significantly negatively correlated with frequency of heavy drinking in positive situations involving Pleasant Emotions. Anxiety sensitivity (fear of anxiety-related sensations; Peterson & Reiss, 1992), but not Neuroticism (tendency to experience negative affect; Costa & McCrae, 1992), mediated the observed associations between PTSD symptoms and situation-specific heavy drinking in negative contexts in general, and Conflict with Others and Physical Discomfort situations in particular. Implications for designing potentially more effective interventions for women with comorbid PTSD-alcohol use disorders are discussed.