Estimates of nutrient intakes based on food purchasing records modelled on the National Food Survey (NFS) were compared with nutrient intakes calculated from food consumption records based on a semi-weighted method, a combination of weighing and household measurement techniques. Of eighty-two families in Cambridge who completed the study, thirty-two were volunteers and fifty were from a random sample in which the co-operation rate was 73%. The estimated energy and nutrient contents of the Cambridge food purchases were very similar to those reported by the NFS for families of similar composition and income. The energy intakes obtained by the semi-weighed method were compared with results from twenty-five studies of energy intake based on quantitative measurements of food consumption: there was no evidence to suggest that the semi-weighed method consistently under- or overestimated intakes in the Cambridge subjects. Purchases adjusted to allow for waste and consumption of food by visitors contained significantly more energy, protein, carbohydrate, calcium, iron and dietary fibre than measured home food consumption. There was no significant difference in the nutrient content of purchases and consumption per 4.184 MJ (1000 kcal), with the exception of Fe and ascorbic acid. Measured wastage of edible food in thirty-one families averaged 3.2% of purchases. Estimate of wastage in all eighty-two families was 3.8%, and consumption of food by visitors accounted for 3.0% of purchases. The excess of purchases over measured home food energy intake is probably accounted for by a net increase in larder stocks rather than wastage, consumption of food by visitors, or under-recording of intakes.