A number of studies suggest that breast-feeding has beneficial effects on an individual's cardiovascular risk factors in adulthood, although the mechanisms involved are unknown. One possible explanation is that adults who were breastfed differ in their health behaviours. In a historical cohort, adult health behaviours were examined in relation to type of milk feeding in infancy. From 1931 to 1939, records were kept on all infants born in Hertfordshire, UK. Their type of milk feeding was summarised as breastfed only, breast and bottle-fed, or bottle-fed only. Information about adult health behaviours was collected from 3217 of these men and women when they were aged 59–73 years. Diet was assessed using an administered FFQ; the key dietary pattern was a ‘prudent’ pattern that described compliance with ‘healthy’ eating recommendations. Of the study population, 60 % of the men and women were breastfed, 31 % were breast and bottle-fed, and 9 % were bottle-fed. Type of milk feeding did not differ according to social class at birth, and was not related to social class attained in adult life. There were no differences in smoking status, alcohol intake or reported physical activity according to type of milk feeding, but there were differences in the participants' dietary patterns. In a multivariate model that included sex and infant weight gain, there were independent associations between type of feeding and prudent diet scores in adult life (P= 0·009), such that higher scores were associated with having been breastfed. These data support experimental findings which suggest that early dietary exposures can have lifelong influences on food choice.