There is an epidemic of obesity, with associated type 2 diabetes and metabolic syndrome in developed societies, and in some developing societies such as parts of the Indian sub-continent and China. The epidemic brings an enormous burden of chronic illness including cardiovascular disease and bone and joint disorders, which will have tremendous economic and humanitarian cost. Prevention of disease is difficult if its causes are not known. The speed of development of this problem rules out a genetic origin, although the action of early environmental factors on gene expression must be involved. It clearly results from some environmental change, and this underlines its prominence in populations in transition. In part the changing environment relates to nutrition, or probably more likely the balance between energy expenditure and calorific intake. The focus on the early origins of such disease has drawn attention to the ways in which predictive adaptive responses may play a part. These are induced by cues from the mother about the external environment including her dietary balance, body composition and metabolism. The effects can be mediated even on the pre–implantation embryo and determine the ways in which the late gestation fetus and the placenta interact. Maternal adaptations to pregnancy also play a role. During evolution these predictive responses conferred survival advantage in a poor or uncertain nutritional environment. But when the prediction is inappropriate, e.g. in a modern energy-rich environment, they are associated with greater risk of later disease. The mechanisms involved include epigenetic gene-environment interactions, e.g. changes in DNA methylation. In addition the effects of gene polymorphisms which confer susceptibility to disease in affected individuals become manifest when coupled with prenatal restriction of growth. Animal and human physiological studies are giving insights into these processes – one particular focus being on endothelial function. They also show that both the metabolic and the cardiovascular effects of an impaired intrauterine environment can be passed across two or more generations, questioning the ways in which we have traditionally thought about ‘inherited’ components of disease risk.