Diel changes in circulating melatonin were measured in juvenile Atlantic salmon, Salmo salar, maintained under natural and out-of-phase seasonal photocycles. Under natural daylengths of autumn, winter, spring, and summer circulating melatonin levels were inversely related to light intensity, with levels low during the day and high at night. The duration of the nocturnal increase in circulating melatonin was related to the duration of darkness, i.e., longer in winter than in summer. Under simulated seasonal photocycles circulating melatonin concentrations measured in August, October, and December were also elevated for the duration of darkness, irrespective of whether the photoperiods were synchronized or 6 months out-of-phase with the natural light and temperature cycles. Circulating melatonin also provided an accurate representation of the prevailing photoperiod in fish initially maintained on simulated natural photocycles, either synchronized or 6 months out-of-phase with the natural light cycle, and then held for 3 months on daylengths approximating the summer and winter solstices. Well-defined melatonin rhythms were always present, irrespective of time of year, photoperiod, and temperature. The amplitude of the nocturnal increase in circulating melatonin was similar in groups of fish maintained under simulated seasonal photoperiods 6 months out-of-phase with each other, but otherwise identical conditions, indicating that daylength per se did not influence the amplitude of the melatonin rhythm. The amplitude of the melatonin rhythm was slightly higher during the summer months, suggesting that temperature may modify circulating melatonin levels. These results demonstrate that circulating melatonin profiles always reflect the prevailing daylength and hence have the potential to provide the Atlantic salmon with accurate information on daily and calendar time, which could be utilised to time daily and seasonal events.