The effectiveness of cognitive and cognitive-relaxation conditions in reducing general anger were compared. By the 5-week follow-up, no differences were found between the two treatments. Relative to controls, both treatments reported significantly less general anger, anger across different situations, person-specific anger, anger-related physiological reactivity, and trait anxiety. In an analogue provocation, both groups reported less unconstructive coping than controls. Subjects in the cognitive condition reported significantly less tendency to suppress anger than controls; cognitive-relaxation subjects did not differ from cognitive or control subjects on this measure. No between-group differences were found for state anger and pulse rates in the analogues though both were low in the initial assessment, daily anger intensities, or the tendency to express anger outwardly, though the latter approached significance. A 15-month follow-up revealed maintenance of anger and anxiety reduction relative to controls, and no differences between active treatments. Implications for the treatment of general anger are discussed.