DescriptionEvidence indicates that exercise improves mood, but not enough is known about the level of exertion required for optimum mood benefit. The present study examined the nature of the relation between exertion level and mood improvement in the theoretical context of the dual-mode hypothesis and opponent-process theory by testing mood changes in highly active and sedentary college-age participants in both assigned and self-selected conditions. As expected, exercise produced in-task arousal, and post-task mood improvement. As predicted by the dual-mode hypothesis and opponent-process theory, at low levels of exertion, in-task and post-task mood improvement was observed, and at high levels of exertion, in-task mood worsened, but post-task mood improved. Participants chose speeds close to 5% below lactate threshold. Theoretical and practical implications of these findings are discussed.