The relative contribution of socioeconomic, behavioural and biological factors operating in fetal and infant life, childhood and adulthood to risk for cardiovascular disease, respiratory diseases and non-insulin-dependent diabetes in middle age has become an important research issue. All 1142 babies born in Newcastle upon Tyne in May and June 1947 were recruited into a prospective cohort study of child health (the 'Thousand Families' study) and followed in great detail to the age of 15 y, with a brief further follow up at age 22 y. Children from poorer families were at greatest risk of severe respiratory tract infection in infancy. Children from professional and managerial families were on average taller and heavier throughout childhood than those from semi- and unskilled manual social classes. Repeated infections in early childhood greatly increased the risk of developing chronic respiratory disease by age 15 y. This paper outlines a new investigation designed to trace surviving members of this cohort and to chart the relationships between their socioeconomic circumstances, lifestyles, experiences and health from birth through to the present day. Existing data on socioeconomic circumstances and infections in infancy and childhood, infant nutrition, birthweight and physical development to age 22 y will be linked to information gained from a new study. This comprises a postal questionnaire survey of study members' adult health, socioeconomic circumstances and lifestyle, and a hospital based clinical examination including heart and lung function, glucose tolerance, blood lipids and anthropometric measurements at age 49-51 y. Out of a target sample of 979 people for whom sufficient data are available on the first year of life, 866 (88%) have been traced and 649 are still resident in the North of England. Those study members who have been traced are highly representative of the original cohort. The Thousand Families cohort provides a unique opportunity for detailed epidemiological study because of the wealth of data available on infant and childhood socioeconomic and family circumstances, all of which was collected prospectively. In addition, there has been comparatively little loss to follow-up since 1948.