Flavonoids are effective antioxidants and, in theory, may provide protection against cancer, although direct human evidence of this is scarce. The relation between the intake of antioxidant flavonoids and subsequent risk of cancer was studied among 9,959 Finnish men and women aged 15-99 years and initially cancer free. Food consumption was estimated by the dietary history method, covering the total habitual diet during the previous year. During a follow-up in 1967-1991, 997 cancer cases and 151 lung cancer cases were diagnosed. An inverse association was observed between the intake of flavonoids and incidence of all sites of cancer combined. The sex- and age-adjusted relative risk of all sites of cancer combined between the highest and lowest quartiles of flavonoid intake was 0.80 (95% confidence interval 0.67-0.96). This association was mainly a result of lung cancer, which presented a corresponding relative risk of 0.54 (95% confidence interval 0.34-0.87). The association between flavonoid intake and lung cancer incidence was not due to the intake of antioxidant vitamins or other potential confounding factors, as adjustment for factors such as smoking and intakes of energy, vitamin E, vitamin C, and beta-carotene did not materially alter the results. The association was strongest in persons under 50 years of age and in nonsmokers with relative risks of 0.33 (95% confidence interval 0.15-0.77) and 0.13 (95% confidence interval 0.03-0.58), respectively. Of the major dietary flavonoid sources, the consumption of apples showed an inverse association with lung cancer incidence, with a relative risk of 0.42 (95% confidence interval 0.23-0.76) after adjustment for the intake of other fruits and vegetables. The results are in line with the hypothesis that flavonoid intake in some circumstances may be involved in the cancer process, resulting in lowered risks.