Melatonin may protect against breast cancer. Light and other factors influence melatonin, but the evidence is limited. The authors conducted a study to determine factors related to melatonin. Women volunteers recruited in Toronto, Canada, from 2002 to 2004 collected urine for three nights (winter and summer), took periodic light measurements, and recorded exposures in a diary. The relation of each variable to log-transformed creatinine-corrected 6-sulfatoxymelatonin in overnight urine was determined by use of generalized estimating equation linear regression. The final model was based on 1,054 measurement days from 213 participating women. None of the light variables was related to the log of 6-sulfatoxymelatonin. A significant interaction between season and day length was included in the final model. The most significant factor was duration of exercise (beta = 0.072; p = 0.004, two-tailed), which increased the amount of melatonin produced. Exercise duration later in the day was more significant (beta = 0.108; p = 0.0009, two-tailed). There was no difference between moderate or strenuous exercise. The failure to find a relation between light brightness and melatonin may be due to the difficulty of measuring this, as well as the importance of the light spectrum, which could not be measured. It is possible that the protective effect of exercise with respect to breast cancer may operate in part through an effect on melatonin.