In animals, circadian pacemakers respond to seasonal changes in day length by making corresponding adjustments in the durations of diurnal and nocturnal periods of circadian rhythms; these adjustments mediate effects of photoperiod on breeding and other seasonally recurring phenomena. Little is known about photoperiod responses of human circadian pacemakers. To investigate this question, we recorded and compared circadian rhythm profiles of 15 individuals after chronic exposures to short (8 h) and long (14 h) nights. As occurs in animals, durations of nocturnal periods of active melatonin secretion (11.9 +/- 1.6 vs. 10.3 +/- 1.3 h, df = 14, t = 4.583, P < 0.0005, paired t test), high prolactin secretion (12.9 +/- 2.1 vs. 9.9 +/- 2.2 h, df = 11, t = 2.917, P < 0.01), and sleep (10.6 +/- 0.8 vs. 7.6 +/- 0.4 h, df = 14, t = 17.122, P < 0.0005) were longer after exposure to long nights than after short ones. Durations of nocturnal periods of low rectal temperature (11.6 +/- 2.3 vs. 9.5 +/- 1.6 h, df = 12, t = 3.912, P < 0.001) and rising cortisol secretion (10.8 +/- 1.6 vs. 9.3 +/- 1.9 h, df = 14, t = 3.130, P < 0.005) were also longer. Some of these differences persisted during 24-h periods of enforced wakefulness in constant dim light, indicating that prior exposure to the two regimes induced abiding changes in the timing of internal processes, such as circadian pacemaker oscillations, that control the durations of nocturnal and diurnal periods of the rhythms.