Since the first report of a protective effect of dietary vitamin A on lung cancer risk was issued in 1975, a succession of retrospective (case-control) and prospective (cohort) epidemiological studies have examined the association. Food frequency questionnaires have been used to assess dietary intake of food groups and to generate indices of specific dietary components, in particular retinol and carotene. Serum levels of retinol, carotenoids, and other micronutrients have been determined as well, and related to risk of lung cancer. The dietary studies have been notably consistent, finding an approximate 50% reduction in risk associated with high compared to low consumption of carotene-containing fruits and vegetables. Unresolved issues include gender differences, histological specificity and the interaction with cigarette smoking as well as the role of other possible protective factors found in these foods--indoles, vitamin C, trace minerals. Most studies in which serum beta-carotene was assayed in prospectively collected blood samples have found lower levels of beta-carotene in people who subsequently developed lung cancer. Unlike carotene, blood retinol levels do not reflect dietary intake under normal conditions and, as might be expected, have failed to show a consistent relation with risk of lung cancer. Although epidemiological studies have not strongly supported the role of preformed retinol as a protective agent, animal studies have provided convincing evidence that retinol and synthetic retinoids are protective against epithelial tumours including those of the lung. A number of on-going clinical trials are testing the efficacy of retinol, beta-carotene, synthetic retinoids and alpha-tocopherol for preventing cancer of the lung, especially in high risk groups such as heavy smokers, miners and asbestos workers. It is hoped that these trials will help explain the nature of the observed protective effect of fruit and vegetables.