Plasma and dietary levels of retinol and beta-carotene were evaluated in a consecutive series of 47 females with histologically proven primary lung cancer and 159 nonneoplastic hospital controls. The dietary questionnaire included 69 different items: special care was given to foods rich in vitamin A and seasonal foods (e.g., vegetables and fruits), whereas serum analysis was focused on retinol and beta-carotene. Age-adjusted mean values for cases and controls were, respectively, 458.3 vs. 551.3 mg for plasma retinol, 276.1 vs. 390.1 mg for plasma carotene; 598.1 vs 820.6 mg for daily retinol, and 628.0 vs. 882.5 mg for dietary carotene. The odds ratios for low vs. high tertile, adjusted for age, smoking, retinol or carotene, cholesterol, and triglycerides by multivariate analysis were, respectively, 1.13 for plasma retinol, 5.04 for plasma carotene, 3.27 for dietary retinol, and 2.93 for dietary carotene. For all the examined items, there was a trend of increased risk for the second and third tertile, and statistical significance was reached for plasma beta-carotene (p less than 0.05). The hypothesis that a higher risk of lung cancer is related to a low vitamin A consumption is supported by these data.