There is an increasing interest in the origins of adult disease in early life. The elucidation of such explanations of current morbidity and mortality must depend upon the follow-up of previously established cohorts. This paper describes the design of and background to such a follow-up of one of the richest data sources for this type of research: Sir John Boyd Orr's survey of diet and health in pre-war Britain. 1,352 families from 16 centres in England and Scotland were surveyed; 3,762 children aged up to 19 years from these families were examined. Socio-economic information and detailed one week dietary diary records are available for all families. Detailed medical examinations (including anthropometry) were undertaken on children in 14 of the centres. Most of the information is cross-sectional although 1,322 children were examined on two or more occasions one year apart to assess the effects of dietary supplementation. Dietary records were retrieved for 1,343 (99.3%) of the families. Medical examination records were found for 3,560 (94.6%) of the children who were examined in the survey and attempts have been made to trace 4,973 children who were either examined or whose family participated in the dietary survey. The data demonstrate relationships between family food expenditure and height in childhood and housing conditions. Eighty-five per cent (4211/4973) of the children have been traced and flagged on the National Health Service Central Register, Southport. The characteristics of those traced do not differ significantly from those we have been unable to trace although untraced females were slightly heavier. To date 696 (16.5%) of the cohort have died. The cohort will be used to investigate the relationship between diet, nutritional status (height, weight, cristal height), health and social circumstances in childhood, and mortality and morbidity in adulthood.