To study the metabolic effects of pulsatile glucagon administration, six male volunteers were submitted to a 260-min glucose-controlled glucose intravenous infusion using the Biostator. The endogenous secretion of the pancreatic hormones was inhibited by somatostatin (100 micrograms X h-1), basal insulin secretion was replaced by a continuous insulin infusion (0.2 mU X kg-1 X min-1), and glucagon was infused intravenously in two conditions at random: either continuously (125 ng X min-1) or intermittently (812.5 ng X min-1, with a switching on/off length of 2/11 min). Blood glucose levels and glucose infusion rate were monitored continuously by the Biostator, and classical methodology using a D-[3-3H]glucose infusion allowed us to study glucose turnover. While basal plasma glucagon levels were similar in both conditions (122 +/- 31 vs. 115 +/- 18 pg X ml-1), they plateaued at 189 +/- 38 pg X ml-1 during continuous infusion and varied between 95 and 501 pg X ml-1 during pulsatile infusion. When compared with continuous administration, pulsatile glucagon infusion initially induced a similar increase in endogenous (hepatic) glucose production and blood glucose, did not prevent the so-called "evanescent" effect of glucagon on blood glucose, and after 3 h tended to reduce rather than increase hepatic glucose production. In conclusion, in vivo pulsatile hyperglucagonemia in normal man fails to increase hepatic glucose production.