Both plasma and dietary measures of vitamin A status were investigated along with previously established risk factors (number of sexual partners, age at first intercourse, smoking, and oral contraceptive use) in a study of 117 in situ cervical cancer patients and 196 matched community controls in Sydney, Australia. Neither total calories nor retinol from foods was related to cancer risk, nor was plasma retinol. When plasma and dietary indexes were considered together, vitamin C, fruit juices, and plasma beta-carotene showed protective effects. Plasma beta-carotene reduced risk from top to bottom quartile by 80%, vitamin C by 60%, and fruit juices by 50%. Thus the evidence suggests that cancer risk is associated with some aspect of diet that is reflected in the effect of plasma beta-carotene. There is no clear effect of any one nutrient but fruit juices appear protective. Thus vitamin C and beta-carotene are likely candidates.