Despite extensive experimental studies on total starvation, many of the findings relating to protein, fat (plus ketone body), and carbohydrate metabolism remain confusing, although they become more consistent when considered in relation to the degree of initial obesity. During prolonged starvation, protein loss and percent energy derived from protein oxidation are 2- to 3-fold less in the obese than in the lean; percent urine N excreted as urea is 2-fold less in the obese; and the contribution of protein to net glucose production is only about half in the obese compared to lean subjects. During short-term starvation (first few days) the following differences are reported: hyperketonaemia is typically 2-fold greater in lean subjects, but associated with a 2-fold lower uptake of ketone bodies by forearm muscle; glucose tolerance becomes impaired more in lean subjects; and both protein turnover and leucine oxidation increase in the lean, but may show no significant change in the obese. It is no longer acceptable to describe the metabolic response to starvation as a single typical response. The differences between lean and obese subjects have important physiological implications, some of which are of obvious relevance to survival.