In 1993-94 a nutritional survey of 15-year-old adolescents was carried out in Uppsala, a university city in eastern Sweden, and in Trollhättan, an industrial town in the western part of the country. The study was the beginning of a longitudinal nutritional survey of 193 boys and 218 girls randomly selected from the official population register. The dietary methods used were a food frequency questionnaire (FFQ) and a combined estimated and weighed 7-day dietary record. According to the FFQ both boys and girls consumed cereals, butter or margarine and something from the group meat/fish/eggs every day. Milk, milk products, sweets and snacks were consumed more than once per day. Vegetables and fruit/roots were consumed less often. There was relatively good correlation between the FFQ and the 7-day record results. The mean daily energy intakes of the 15-year-old boys and girls were 10.2 and 8.3 MJ, respectively, in Uppsala, and 9.8 and 7.4 MJ in Trollhättan. The mean daily intakes of sucrose were 74 and 58 g in boys and 64 and 52 g in girls. The average daily dietary fibre intake was 1.8 g/MJ. The daily intake of energy obtained from breakfast was 18% on weekdays and 22% on weekends; 18 and 11% from prepared lunch; 24 and 30% from dinner, 21 and 14% from light meals; and 16 and 28% from snacks, in Uppsala and Trollhättan respectively. The lowest energy intake from a prepared lunch meal was noted during weekends. On both weekdays and weekends, considerable energy was obtained from light meals and snacks. However, nearly 40% of the girls and 28% of the boys had an energy intake from fat of < 30 energy%. The mean intakes of vitamin D and selenium and, in the case of girls, iron and zinc, were below the official Swedish Nutrition Recommendations. The daily median intake of iron was 18.7 mg in boys and 14.1 mg in girls. The iron intakes varied between 6 and 35 mg per day. Low serum ferritin concentrations, defined as s-ferritin < 12 micrograms/L, were found in 7 boys (3.7% and in 29 girls (13.9%). Significant negative correlations were found between smoking and frequency of consumption of vegetables, roots, fruits and meat. A negative correlation was also found between smoking and the intake of energy and a number of nutrients. Socioeconomic factors seem to be less important for the food habits of teenagers than for those noted during childhood. Only frequent consumption of vegetables was positively correlated to the mothers' educational level.