Exercise elicits circadian phase-shifting effects, but additional information is needed. The phase-response curve describing the magnitude and direction of circadian rhythm phase shifts, depending on the time of the zeigeber (time cue) stimulus, is the most fundamental chronobiological tool for alleviating circadian misalignment and related morbidity. Fifty-one older and 48 young adults followed a circadian rhythms measurement protocol for up to 5.5 days, and performed 1 h of moderate treadmill exercise for 3 consecutive days at one of eight times of the day/night. Temporal changes in the phase of 6-sulphatoxymelatonin (aMT6s) were measured from evening onset, cosine acrophase, morning offset and duration of excretion. Significant phase-response curves were established for aMT6 onset and acrophase with large phase delays from 7:00 pm to 10:00 pm and large phase advances at both 7:00 am and from 1:00 pm to 4:00 pm. Delays or advances would be desired, for example, for adjustment to westward or eastward air travel, respectively. Along with known synergism with bright light, the above PRCs with a second phase advance region (afternoon) could support both practical and clinical applications. Although bright light is regarded as the primary circadian zeitgeber, its limitations support exploring alternative zeitgebers. Exercise elicits significant circadian phase-shifting effects, but fundamental information regarding these effects is needed. The primary aim of the present study was to establish phase-response curves (PRCs) documenting the size and direction of phase shifts in relation to the circadian time of exercise. Aerobically fit older (n = 51; 59-75 years) and young adults (n = 48; 18-30 years) followed a 90 min laboratory ultrashort sleep-wake cycle (60 min wake/30 min sleep) for up to 5½ days. At the same clock time on three consecutive days, each participant performed 60 min of moderate treadmill exercise (65-75% of heart rate reserve) at one of eight times of day/night. To describe PRCs, phase shifts were measured for the cosine-fitted acrophase of urinary 6-sulphatoxymelatonin (aMT6s), as well as for the evening rise, morning decline and change in duration of aMT6s excretion. Significant PRCs were found for aMT6s acrophase, onset and duration, with peak phase advances corresponding to clock times of 7:00 am and from 1:00 pm to 4:00 pm, delays from 7:00 pm to 10:00 pm, and minimal shifts around 4:00 pm and 2:00 am. There were no significant age or sex differences. The amplitudes of the aMT6s onset and acrophase PRCs are comparable to expectations for bright light of equal duration. The phase advance to afternoon exercise and the exercise-induced PRC for change in aMT6s duration are novel findings. The results support further research exploring additive phase-shifting effects of bright light and exercise and health benefits.