Studies that have found a positive influence of moderate, non-exhaustive exercise on life expectancy contradict the rate-of-living theory, which predicts that high energy expenditure in exercising animals should shorten life. We investigated effects of exercise on energy metabolism and life span in male mice from lines that had been selectively bred for high voluntary wheel-running activity and from the nonselected control lines. Mice were divided into the following three groups (n = 100 per group): active high-runner mice (housed with wheels; HR +), sedentary high-runner mice (no wheels provided; HR -), and active control mice (C +). Sixty animals from each group were left undisturbed throughout their lives to create survival curves. In the remaining 40 animals in each group, energy metabolism and body composition was measured at 2, 10, 18, or 26 mo of age. Wheel-running activity was increased by similar to 50% throughout life in HR + mice compared with C + mice, and mass-specific daily energy expenditure was increased by similar to 30% in HR + mice compared with both C + mice and HR - mice. Median life span was similar in HR + mice and HR - mice (740 and 733 d, respectively), and it was significantly shorter in these mice than it was in C + mice (828 d). Thus, increasing the amount of voluntary aerobic exercise (as a result of selective breeding or housing with wheels) did not result in extended life span in mice, and we found no evidence for a direct link between energy expenditure and life span.