In three counties in western New York, a case-control study of incident, pathologically confirmed, single, primary cancers of the rectum was conducted from 1978 to 1986. Cases were matched with neighborhood controls on age and sex; 277 case-control pairs of males and 145 case-control pairs of females were interviewed regarding usual quantity and frequency of consumption of foods. Risk of rectal cancer increased with increasing intake of kilocalories, fat, carbohydrate, and iron. Risk decreased with increasing intake of carotenoids, vitamin C, and dietary fiber from vegetables. Fiber from grains, calcium, retinol, and vitamin E were not associated with risk. Associations of intake with risk were generally stronger for males than for females except for vitamin C. The association for carotenoids, vitamin C, and vegetable fiber persisted after stratification on intake of either kilocalories or fat.