The study aimed to quantify a seasonal change in circadian rhythms and its relationship to the social/sleep regimen in humans living in Novosibirsk (55°N), using the naturalistic situation that daylight saving time transitions have been abolished in Russia. Sixty-three volunteers entered the study, and 46 completed it. One group got up at ~6 a.m. and another at ~7 a.m. during their regular 5-workdays schedule. They collected 19 saliva samples at home over 24 h (including 2 samples during the night) on July 3–4, and December 18–19, 2015. Salivary melatonin was measured using radioimmunoassay; the times of evening onset and morning offset were objectively determined using the hockey-stick algorithm and served as circadian phase markers. Nearly all melatonin profiles were normal in summer (high nighttime and low daytime levels), whereas in winter, significantly more – 8 profiles – were abnormal (additional daytime peak, out-of-phase daytime secretion, or absence of secretion), of which 3 (plus 1 for other reasons) could not be included in the further analysis. The duration of melatonin secretion (somewhat less than 12 h) and amount of melatonin secreted did not differ between seasons. In winter compared to summer the melatonin rhythm, on average, significantly phase delayed by half-an-hour, with a tendency for greater inter-individual phase variability. The phase delay was attributable to those subjects who got up at ~7 a.m. (and who were longer sleepers). The melatonin rhythm reflected well the sleep timing difference between the two groups in summer, whereas in winter this coherence was lost. In summary, timing of the circadian system is strictly synchronised in summer by the long light: short dark photoperiod (with sleep as a constituent of the 7 h 10 min dark phase of the cycle), whereas in winter, with the long dark nights (17 h 12 min), an inter-individual phase desynchrony and even abnormal melatonin patterns emerge, despite a constant sleep/social regimen, suggesting that the winter season is unfavourable for circadian status.