T o quantitate hepatic glycogenolysis, liver glycogen concentration was measured with 13C nuclear magnetic resonance spectroscopy in seven type II diabetic and five control subjects during 23 h of fasting. Net hepatic glycogenolysis was calculated by multiplying the rate of glycogen breakdown by the liver volume, determined from magnetic resonance images. Gluconeogenesis was calculated by subtracting the rate of hepatic glycogenolysis from the whole body glucose production rate, measured using [6-3H]glucose. Liver glycogen concentration 4 h after a meal was lower in the diabetics than in the controls; 131 +/- 20 versus 282 +/- 60 mmol/liter liver (P < 0.05). Net hepatic glycogenolysis was decreased in the diabetics, 1.3 +/- 0.2 as compared to 2.8 +/- 0.7 mumol/(kg body wt x min) in the controls (P < 0.05). Whole body glucose production was increased in the diabetics as compared to the controls, 11.1 +/- 0.6 versus 8.9 +/- 0.5 mumol/(kg body wt x min) (P < 0.05). Gluconeogenesis was consequently increased in the diabetics, 9.8 +/- 0.7 as compared to 6.1 +/- 0.5 mumol/(kg body wt x min) in the controls (P < 0.01), and accounted for 88 +/- 2% of total glucose production as compared with 70 +/- 6% in the controls (P < 0.05). In conclusion: increased gluconeogenesis is responsible for the increased whole body glucose production in type II diabetes mellitus after an overnight fast.