Dim light melatonin onset (DLMO) is the most reliable measure of human central circadian timing. Its modulation by light exposure and chronotype has been scarcely approached. We evaluated the impact of light changes on the interaction between melatonin, sleep, and chronotype in university students (n = 12) between the Antarctic summer (10 days) and the autumn equinox in Montevideo, Uruguay (10 days). Circadian preferences were tested by validated questionnaires. A Morningness–Eveningness Questionnaire average value (47 ± 8.01) was used to separate late and early participants. Daylight exposure (measured by actimetry) was significantly higher in Antarctica versus Montevideo in both sensitive time windows (the morning phase-advancing and the evening phase-delaying). Melatonin was measured in hourly saliva samples (18–24 h) collected in dim light conditions (<30 lx) during the last night of each study period. Early and late participants were exposed to similar amounts of light in both sites and time windows, but only early participants were significantly more exposed during the late evening in Antarctica. Late participants advanced their DLMO with no changes in sleep onset time in Antarctica, while early participants delayed their DLMO and sleep onset time. This different susceptibility to respond to light may be explained by a subtle difference in evening light exposure between chronotypes.