] have pieced together a number of disparate observations to provide a fascinating and novel hypothesis on the aetiology of Type 2 (non-insulin-dependent) diabetes mettitus. Central to their hypothesis is the notion that beta-cell function is diminished as a result of under-nutrition at a critical point in the development of the islets of Langerhans during fetal and early infant life.We are unable to reconcile this hypothesis to observations of increased insulin levels (in relation to Caucasian populations) in normal, glucose intolerant and diabetic individuals in developing populations, such as Pacific islanders According to Hales and Barker [1], these populations (or the individuals within them) are at increased risk of developing glucose intolerance with modernization, presumably because early under-nutrition resulting in beta-cell damage has left the pancreas unable to cope with the increased insulin requirements of the diet, obesity and reduced activity levels of a modern lifestyle.However, the very populations with the greatest susceptibility to Type 2 diabetes, are those with the highest beta-cell function, based on their total immunoreactive insulin levels. Hales and colleagues [1,6] claim that much of this "immunoreactive" insulin is biologically less active proinsulin and des 31,32-split proinsulin. This appeared to be the case in newly diagnosed Type 2 diabetic patients, particularly 30 min after an oral glucose load [6], but is not the case in mild non-insulin-dependent diabetes [7]. If a high proportion of insulin in susceptible populations could be shown to be composed of physiologically inactive proinsulin and its split products, then the Hales and Barker hypothesis of relative insulin deficiency might still hold. To date, there is no confirmation that population differences in total insulin levels can be explained by differences in concentrations of true insulin compared to proinsufins.Within populations, the Hales and Barker hypothesis also runs into difficulties. In several populations it has been shown that increased (not decreased) insulin levels (fasting, 2-h, etc.), predict deterioration in glucose tolerance from normal to impaired or to Type 2 diabetes [8][9][10][11]. While admittedly these studies need to be repeated measuring true insulin, they suggest that the beta cells of individuals with deteriorating glucose tolerance tend to have increased (but possibly still "abnormal") rather than decreased function. If it is only first phase secretion [12] that needs to be diminished to provide the susceptibility (and this is not outside the bounds of possibility), it is not clear that the "thrifty phenotype" hypothesis relating infant nutrition to pancreatic development adequately explains such a specific lesion.Furthermore, we should point out that Hales and Barker have misinterpreted our data showing a recent decline in the incidence of glucose intolerance in the high prevalence Nauruan population [13]. We certainly did not attribute the recently observed changes in prevalence and incidence t...