Decreased melatonin production, due to acute suppression of pineal melatonin secretion by light exposure during night work, has been suggested to underlie higher cancer risks associated with prolonged experience of night work. However, the association between light exposure and melatonin production has never been measured in the field. In this study, 24-h melatonin production and ambulatory light exposure were assessed during both night-shift and day/evening-shift periods in 13 full-time rotating shiftworkers. Melatonin production was estimated with the excretion of urinary 6-sulfatoxymelatonin (aMT6s), and light exposure was measured with an ambulatory photometer. There was no difference in total 24-h aMT6s excretion between the two work periods. The night-shift period was characterized by a desynchrony between melatonin and sleep-wake rhythms, as shown by higher melatonin production during work and lower melatonin production during sleep when working night shifts than when working day/evening shifts. Light exposure during night work showed no correlation with aMT6s excreted during the night of work (p > .5), or with the difference in 24-h aMT6s excretion between the two work periods (p > .1). However, light exposure during night work was negatively correlated with total 24-h aMT6s excretion over the entire night-shift period (p < .01). In conclusion, there was no evidence of direct melatonin suppression during night work in this population. However, higher levels of light exposure during night work may have decreased total melatonin production, possibly by initiating re-entrainment and causing internal desynchrony. This interpretation is consistent with the proposition that circadian disruption, of which decreased melatonin production is only one of the adverse consequences, could be the mediator between night shiftwork and cancer risks.