Experimental studies suggest that melatonin has a protective effect against breast cancer. Exposure to light suppresses melatonin secretion, but to a lesser degree in totally blind persons. Breast cancer was investigated in a cohort of 15 412 Norwegian visually impaired women. The risk among totally blind women was 0.64 (95% CI = 0.21-1.49, 5 cases only), and for those who became blind before age of 65, the SIR was 0.51 (95% CI = 0.11-1.49). Our findings give support to the 'melatonin hypothesis'.