Micronutrient deficiencies occur most commonly in poor countries and, therefore, are most likely to be associated with cancers common in these countries. Epidemiological studies are hampered by inaccurate measurement of micronutrient intake and by the correlations between intakes of many nutrients. The strongest evidence for a protective effect of micronutrients is for oesophageal cancer. The identity of the micronutrients is not certain, but may include retinol, riboflavin, ascorbic acid and Zn; alcohol, smoking and dietary nitrosamines increase the risk for oesophageal cancer. For stomach cancer there is good evidence that fruit and vegetables are protective. The protective effect of these foods might be largely due to ascorbic acid, but other nutrients and non-nutrients may also be important; the risk for stomach cancer is increased by salt, some types of preserved foods, and by infection of the stomach with the bacterium Helicobacter pylori. The risk for lung cancer appears to be reduced by a high intake of fruit and vegetables, but it is not clear which agents are responsible and the major cause of lung cancer is cigarette smoking. Diet is probably the major determinant of the risk for colo-rectal cancer; there is evidence that fruit and vegetables and fibre reduce risk and that meat and animal fat increase risk, but there is no convincing evidence that these relationships are mediated by micronutrients. The risk for cervical cancer is inversely related to fruit and vegetable consumption and, therefore, to consumption of carotenoids and ascorbic acid, but the major cause of this cancer is human papillomavirus and it is not yet clear whether the dietary associations indicate a true protective effect or whether they are due to confounding by other variables. The evidence that micronutrients are important in the aetiology of either breast cancer or prostate cancer is weak, but the possible roles of 1,25-dihydroxycholecalciferol and alpha-tocopherol in prostate cancer require further study.