How long induced moods last is a critical question for mood research but has been only poorly addressed. In particular, physiological parameters have been rarely included to assess the effectiveness of mood induction procedures. Adopting a dimensional model of mood, we investigated the persistence of four different moods (positive higharousal, positive low-arousal, negative high-arousal, negative lowarousal) induced by four film clips ("sport", "nature", "torture", "slum") during a 9-minute computer task. We measured subjective mood state (valence and arousal), respiration, skin conductance level (SCL), heart rate, and corrugator activity in 76 subjects. Viewing of the selected film clips induced the expected effects both subjectively and physiologically. Corrugator activity was higher at the end of the negative clips than the positive clips, and ventilation and SCL were higher for the arousing clips than for the less arousing clips. People who watched the negative clips still reported more negative valence after the computer task and also showed more facial frowning (cf. figure) and lower SCL during the task than people who watched the positive clips. No arousal effects persisted throughout the task. The results suggest that induced changes in the valence dimension of moods are maintained throughout an intervening task and are physiologically best reflected by corrugator activity and SCL, whereas induced changes in the arousal dimension dissipate quickly. The findings of this study enrich, first, our knowledge concerning the relationships between subjective feelings and their physiological substrate. Second, they inform us about the effectiveness of film clips as a mood induction instrument. Third and most important, they suggest that induced changes in valence last longer than induced changes in arousal. High-arousal moods can last for an extended period of time in daily life, but they seem to be short-lived when induced in the lab. An important methodological consequence is that investigating the effect of the arousal dimension of a person's mood induced in the lab may be only possible when the subsequent task is relatively short. Finally, the findings show which physiological measures may be useful in tracking mood states.