Objective: To evaluate reproducibility and validity of a self-administered food frequency questionnaire (FFQ) to determine the dietary supplement use. Design: Questionnaire data were compared with dietary records (DR) in four 7 day periods for validity. FFQ1 and FFQ2 administered at an interval of approximately one year were compared for reproducibility. Setting and subjects: A total of 361 samples were selected for a validation study of a questionnaire from the subgroup of the Japan Public Health Center-based prospective Study on cancer and cardiovascular disease (JPHC Study) cohort II. Categories of dietary supplements in the FFQ were`multivitamins',`b-carotene',`vitamin C', vitamin E', and`other supplements'. For each category, inquiries were made as to the brand name, frequency and duration of use. A dietary supplement user was de®ned as a subject who used a dietary supplement one or more times a week for a year or longer. Results: Prevalence of overall supplement use was 10.9%. Percentage agreement between FFQ1 and FFQ2 was 91.7%. Sensitivity and speci®city of supplement use by FFQ were 80.6% and 89.2%, respectively. Sensitivity was high for`vitamin C'. When the self-reported categories were corrected, sensitivity for`multivitamins' and vitamin E' improved. Conclusions: The FFQ on dietary supplements used for the JPHC 5 y follow-up survey was valid to determine overall use and well-de®ned dietary supplement use such as vitamin C. Categories must be corrected when determining multivitamin and`vitamin E'. Sponsorship: Grant-in-Aid from the Ministry of Health and Welfare of Japan. Descriptors: dietary supplements; questionnaires; validity of results; reproducibility of results; epidemiologic measurements