The increase in body weight in the USA over the past several decades is now commonly referred to as the ‘obesity epidemic’. An empirical analysis of the literature suggests that the increased weight can be accounted for by an increase in food intake. The solution to the obesity epidemic, therefore, must centre on a reduction in food consumption, a position well accepted by the American population who think that they, as individuals, are responsible for their adiposity by holding the belief that the decision as to what and how much to eat is determined by their own free will. The evidence demonstrates, however, that this is not true. Variables such as portion size, variety of foods offered, fat content of the diet, the number of people eating, the location where eating occurs and even watching food advertisements act as ‘food primes’ causing individuals to increase their energy intake. Despite the plethora of diets, weight-loss clubs, drugs and mechanical devices available to facilitate weight loss, once treatment is terminated and people return to the ‘free’ environment, their weight returns to pre-treatment levels. Only when individuals are protected from environmental variables by gastric surgery or limited to consume only portion-controlled meals can they successfully maintain a reduced weight. Combining the technique of daily weight monitoring with accepting that our eating behaviour is not determined totally by our free choice, we may be able to curb the obesity epidemic.

Use of the energy balance equation for understanding the causation of obesity is discussed. Its basis on the thermodynamic laws is expressed in mathematical models for body-weight changes. Only a very small net energy surplus per time unit constitutes the energy deposition during weight gain, making measurements of its components difficult. The physical laws provide exact quantitative relationships between energy intake, energy expenditure and deposition of energy, but cannot disentangle the initiating and driving forces of the energy imbalance, which may also be an active storage of fat in adipose tissue. These and various other limitations of the energy balance model warrant cautiousness in using the model in studies of obesity causation. Weight gain may be self-promoting and mathematical feedback models allowing estimation of such effects show that they are realistic. Predisposition and susceptibility should be distinguished, and susceptibility as a modifiable predisposition, the genetic and environmental contribution to predisposition and its usefulness as targets for prevention and treatment are discussed. Current progress in unravelling genetic predisposition, the complex genetically-determined mechanisms, the slower progress in unravelling the environmental influences, the different nature of genetic and environmental influences, the possible pathways of environmental influences and the environmental influences as mediators of genetic effects are addressed. The evidence behind the prevailing concept of the ‘obesogenic’ environment is critically analysed. Finally, particular opportunities for the identification of the causes of the obesity epidemic by detailed analysis of an observed irregular development of the epidemic over long time periods are presented, and evidence for predisposition as a result of postnatal environmental influences is inferred from these studies.

The epidemic of ‘obesity’ in India is not appreciated because BMI underestimates the adiposity of Indians. Specific adiposity measurements are necessary for recognition of the adiposity of ‘thin’ Indians. The origin of this adiposity is only beginning to be understood. In addition to a possible genetic predisposition, intrauterine ‘programming’ might be responsible, although in the ‘thrifty phenotype’ hypothesis the adiposity of the ‘thin’ fetus has not been appreciated. Dutch men who faced ‘winter hunger’ during the first trimester of their in utero life have become more obese as adults. Low birth weight predicts central obesity in some studies, including studies in urban children. It has also been shown that small and thin Indian newborns (weight 2·7?kg and ponderal index 2·4?kg\m3) have poor muscle and visceral mass but higher adiposity for a given weight compared with white Caucasian babies. This body composition is influenced by maternal adiposity before pregnancy and by aspects of maternal nutritional intake and circulating nutrient concentrations during pregnancy. There are no strong paternal determinants of adiposity at birth. Adiposity may be an integral part of the orchestrated adjustments made to support ‘brain preservation’ during intrauterine growth, because brain tissue is predominantly fat. Increased nutrition in the face of a genetic predisposition or multigenerational undernutrition increases maternal insulin resistance in late pregnancy and promotes fetal adiposity even in absence of marked hyperglycaemia. Further research is necessary to define the role of specific nutrients and metabolites in the intrauterine processes promoting adiposity before maternal interventions to curtail the epidemic of obesity and diabetes are planned.