The modified dietary history, which was used for assessment of food consumption in the Euronut SENECA study, was validated against a 3-day weighed record in a subsample of 82 elderly subjects from 11 of the 19 participating centres. The modified dietary history provided consistently higher intakes of energy and nutrients than the weighed record, with a median difference of 14% (energy) of the record mean. Unattenuated correlation coefficients for nutrient intakes varied from 0.18 (vitamin A) to 0.79 (water) with a median coefficient of 0.58 (energy). On the whole there was a fair agreement between the dietary history and the 3-day weighed record when nutrients were expressed in weight units and a good agreement when standardized for the apparently lower energy intake by the record method.