Objective: To investigate the association of breast feeding with height and body mass index in childhood and adulthood. Design: Historical cohort study, based on long term follow up of the Carnegie (Boyd-Orr) survey of diet and health in pre-war Britain (1937-1939). Setting: Sixteen urban and rural districts in Britain. Subjects: A total of 4999 children from 1352 families were surveyed in 1937-1939. Information on infant feeding and childhood anthropometry was available for 2995 subjects. Main outcome measures: Mean differences in childhood and adult anthropometry between breast and bottle fed subjects. Results: Breast feeding was associated with the survey district, greater household income, and food expenditure, but not with number of children in the household, birth order, or social class. In childhood, breast fed subjects were significantly taller than bottle fed subjects after controlling for socioeconomic variables. The mean height difference among boys was 0.20 standard deviation (SD) (95% confidence interval (CI) 0.07 to 0.32), and among girls it was 0.14 SD (95% CI 0.02 to 0.27). Leg length, but not trunk length, was the component of height associated with breast feeding. In males, breast feeding was associated with greater adult height (difference: 0.34 SD, 95% CI 0.13 to 0.55); of the two components of height, leg length (0.26 SD, 95% CI 0.02 to 0.50) was more strongly related to breast feeding than trunk length (0.16 SD, 95% CI - 0.04 to 0.35). Height and leg length differences were in the same direction but smaller among adult females. There was no association between breast feeding and body mass index in childhood or adulthood. Conclusions: Compared with bottle fed infants, infants breast fed in the 1920s and 1930s were taller in childhood and adulthood. As stature is associated with health and life expectancy, the possible long term impact of infant feeding on adult mortality patterns merits further investigation.