The relation between portal vein insulin concentrations and suppression of hepatic glucose production, as well as peripheral venous insulin level and increase of peripheral glucose utilization, was compared in conscious, depancreatized, diabetic dogs after infusion of insulin at 0.25 and 0.5 mU/kg/min into either the portal system or the peripheral circulation. Glucose appearance and clearance was measured using [3-³H]-glucose. Before infusion of insulin, portal vein insulin concentrations were undetectable. The intraportal infusion of insulin at 0.25 mU/kg/min increased portal vein insulin to 16 ± 1 μU/ml, significantly higher than the arterial concentration (9 ± 1 μU/ml). Infusion of the same amount of insulin into a peripheral vein raised peripheral insulin levels to 14 ± 1 μU/ml and portal vein concentrations to 12 ± 1 μU/ml. When 0.5 mU/kg/ min of insulin was infused into the portal system, the portal vein insulin level was 28 ± 2 μU/ml and significantly greater than the arterial concentration (16 ± 1 μU/ml). After the same amount of insulin was infused into a peripheral vein, the arterial insulin level was higher than that of the portal vein (25 ±1 μU/ml versus 20 ± 1 μU/ml, respectively). The ensuing hypoglycemie was greater after the 0.5 mU/kg/min infusion compared with the 0.25 mU/kg/min infusion. At each dose there was no significant difference between the peripheral venous or the portal route. Despite different portal vein insulin levels, suppression of glucose appearance (as measured by the area under the extended basal value) during insulin infusion was similar in all four groups (29 ± 6% at portal vein insulin concentration of 12 ± 1 μU/ml after 0.25 mU/kg/min insulin into the peripheral vein, 36 ± 4% at 16 ± 1 μU/ml after 0.25 mU/kg/min insulin into the portal system, 37 ± 6% at 20 ± 1 μU/ml after 0.5 mU/kg/min insulin into the pe-ripheral vein, and 31 ± 7% at 28 ± 2 μU/ml after 0.5 mU/kg/min insulin into the portal system, respectively). Glucose clearance did not increase after infusion of insulin at 0.25 mU/kg/min into either the portal or the peripheral system despite an increase of peripheral venous insulin levels (9 ± 1 μU/ml and 14 ± 1 μU/ ml, respectively). Glucose clearance (measured as the area above the curve) did increase significantly when 0.5 mU/kg/min was infused. The increase after peripheral infusion (197 ± 27 ml/kg) was significantly greater than after portal infusion (141 ± 22 ml/kg) of insulin. Under these two circumstances, the peripheral insulin concentrations were 25 ± 1 μU/ml and 16 ± 1 μU/ml, respectively. These results demonstrate that the liver in diabetic dogs is more sensitive than peripheral tissue to small changes in the plasma insulin concentration. While the peripheral administration of small amounts of insulin is equally effective as portal insulin infusion in suppressing hepatic production of glucose, glucose clearance is more directly related to peripheral insulin concentrations.