An experimental model is described which can be used to study substrate and hormone responses to normal meals administered in very near normal circumstances. After 500, 300 and 125 calorie meals, the relative proportion of fat or protein content did not influence the plasma glucose except for minor differences between the high protein-high fat meals. The insulin response to such meals was correlated positively with the increment in glucose but reduction of protein content below 8 g caused a signficant reduction in the increment in plasma insulin per unit increase in plasma glucose. Alterations in protein content above 8 g made no difference. Fat content of the meal did not significantly alter the insulin response. No evidence was obtained for a major component of insulin release attributable to either bulk or preabsorption phenomena such as sight or smell. It is concluded that a significant accentuation of the insulin response to meals is dependent on a minimum amount of protein and that this is probably mediated by one of the gastro-intestinal hormones. Glucagon release is dependent on protein and carbohydrate content of the meal and is independent of the fat content. There may also be an early stimulation of glucagon release, regardless of content, which may also be hormonally mediated.