It is well known that intensive insulin treatment of non-insulin-dependent diabetics (NIDDM) suppresses endogenous insulin secretion and thereafter improves it. To determine whether 'peripheral normo-insulinemia' or 'normoglycemia' established by the treatment is responsible for this suppression, the following five experiments were conducted on 15 well-controlled non-obese NIDDM patients. Experiment 1: a 100 g oral glucose load (OGL) was performed and blood glucose was monitored by an artificial endocrine pancreas (AP). Experiment 2: a 100 g OGL was done and blood glucose was normalized by AP-controlled insulin infusion. Experiments 3 and 4: a 100 g OGL was conducted while 'hyperglycemia' seen in experiment 1 was mimicked by AP-controlled glucose infusion with pre-programmed insulin infusion at the same rates as those in experiment 2 ('normoinsulinemia') or at rates 1.5 times higher than those in experiment 2 ('relative hyperinsulinemia'), respectively. Experiment 5: a 40 g OGL was conducted while AP-controlled insulin and glucose infusions were administered to make the plasma insulin level lower than in experiment 2 ('hypoinsulinemia') and to mimic the normoglycemic profile observed in experiment 2, respectively. In experiments 3 and 4, neither 'normoinsulinemia' nor 'relative hyperinsulinemia' suppressed the increase in plasma C-peptide after a 100 g OGL. In experiment 5, where the plasma insulin level showed a significantly (P less than 0.05) lower level than in experiment 2 and glycemia was normalized, C-peptide did not show a significant rise after OGL. These results indicate that 'normoglycemia' rather than 'normoinsulinemia' attained during exogenous insulin therapy, is responsible for suppressing endogenous insulin secretion against orally administered glucose.