In rodents, increased activity due to running-wheel access is associated with a change in observed circadian period. In humans, exposure to exercise has failed to demonstrate similar effects on period. Methodological issues with prior studies such as light exposure during exercise, length of study, and method of measuring period confounded those evaluations of the effect of exercise on period in humans. In the present experiment, the authors examined the effect of exercise on period in 8 subjects using a 44-day within-subjects inpatient study. They used a 20-h forced desynchrony protocol, in which subjects were exposed to exercise across circadian phases under dim light conditions. Exercise consisted of three 45-min sessions per wake period on an ergometer. Target exercise intensity was ~65% of maximal heart rate. Intrinsic circadian period was measured using both core body temperature and hourly plasma melatonin samples. Consistent with previous reports, the authors find no effect of exercise on endogenous circadian period as measured by either core body temperature or melatonin. Exercise distributed across biological day and night does not appear to affect circadian period.