Imagery has had an important role in cognitive behavior therapy. One largely untested assumption has been that vivid imagery is necessary for adequate emotional arousal and successful treatment outcome. This study explored the assumed connection between imagery vividness and emotional arousal. It was hypothesized that imagery vividness was not related to degree of emotional arousal and that another individual difference dimension, typical use/non-use of imagery, was a more salient factor. In the first session, sixty five volunteer subjects answered several questionnaires, including the revised Betts Questionnaire Upon Mental Imagery as a measure of imagery vividness, and the Individual Differences Questionnaire as a measure of typical use of imagery. In the second session subjects were asked to visualize one neutral and two negatively-toned scenes. Relaxation instructions were given before each imaging phase. Heart rate and skin resistance were recorded during relaxation and imaging. Subjects also reported on their affective state during each phase by filling out the Affect Rating Scale. Data were analyzed using analysis of variance and multiple regression procedures. Consistent with prior research, it was found that negative imagery produced greater arousal than did neutral imagery. The effect was found for self-reported emotional arousal only; physiological measures showed non-significant trends in the expected direction. The two individual difference dimensions of imagery (vividness and typical use/non-use) were found to be unrelated to degree of arousal as measured by self-report and physiological response. Implications for Paivio's conceptualization of imagery and directions for future research with therapy outcome were discussed.