The relationships between thiamin intake, erythrocyte transketolase (EC 2.2.1.1) activity coefficient (ETK-AC) and total erythrocyte thiamin were investigated in a group of adolescents (13 to 14 years old; nineteen boys, thirty-five girls). Thiamin intakes were calculated from 7 d weighed records, using food composition tables, and compared with those obtained by direct analysis of duplicate diets. Average 7 d calculated thiamin intakes were significantly lower than analysed intakes for both sexes. On an individual basis, calculated intakes ranged from 30 to 143 % of corresponding analysed values. Analysed and calculated intakes were significantly correlated when expressed as mg/d; however, when expressed in terms of energy intake, the correlation was significant for males only. Thiamin intake appeared largely adequate when compared with current UK dietary recommendations (Department of Health, 1991), but the limitations of such comparisons are considered. The major food groups contributing to thiamin intake were examined and showed breakfast cereals to contribute more than 25% of dietary thiamin. A proportion of the subjects had ETK-AC values in ranges usually associated with marginal or severe thiamin deficiency. There was, however, no statistically significant relationship between erythrocyte thiamin and basal or stimulated transketolase activity, or between thiamin intake and either of the methods used to assess status. The need to re-evaluate indices of thiamin status is discussed.