We have studied 15 children referred to St. Göran's Children's Hospital because of suspected ketotic hypoglycemia. The patients were investigated according to a program designed to test several hypotheses--old and new--postulated to explain the etiology of ketotic hypoglycemia. We have used the classical ketogenic provocation with a low calorie, high fat diet and measured the blood levels of several substrates and hormones as well as the urinary excretion of certain metabolites and hormones. Out of the 15 children, 6 will fill the criteria of ketotic hypoglycemia at the time of study. The most remarkable finding in these 6 children in contrast to the other children studied was that they did not decrease their peripheral glucose utilization (measured as Kg) during starvation. These 6 children seemed to be more "advanced" in their adaptation to ketogenic diet in all other parameters studied. The children with ketotic hypoglycemia did not differ from the other children in plasma level of cortisol or urinary excretion of nitrogen, urea, 3-methylstidine and catecholamines. We favour the concept that the children with ketotic hypoglycemia represent the tail of the gaussian curve in the normal age-dependent development of the adaptation to starvation.