A self-administered food use questionnaire which included 276 food items and mixed dishes and a portion size picture booklet with 122 photographs was developed for a large lung cancer intervention trial among approximately 27,000 Finnish men aged 50-69 years. The reproducibility and validity of this questionnaire were studied from March to October 1984. In the reproducibility study, 121 men aged 55-69 years completed the questionnaire three times, at three-month intervals. The intraclass correlations varied from 0.56 for vitamin A to 0.88 for alcohol, with most falling between 0.60 and 0.70. In the validity study, 190 men of similar age kept food consumption records for 12 two-day periods, distributed evenly over a period of six months, and filled in the questionnaire both before and after this period. Correlations between nutrient intake values from the food records and the food use questionnaires ranged from 0.40 for selenium to 0.80 for alcohol. Among subjects who belonged to the lowest quintile on the basis of the food record measurement, an average of 51 per cent fell into the same quintile and 76 per cent fell into the lowest two quintiles when they were categorized on the basis of the food use questionnaire. Findings were similar for the upper tail of the distribution. These data indicate that the self-administered food use questionnaire is useful for measuring individual or group intakes for a variety of nutrients.