Concentrations of insulin, proinsulin, and C-peptide were measured in portal and peripheral venous blood in six nondiabetic, nonobese subjects. Portal vein samples were obtained by umbilical vein catheterization. Three subjects were studied with intravenous infusion of 25 g glucose, and three with 30 g arginine. Insulin and proinsulin were determined in the insulin immunoassay after separation by gel filtration, and C-peptide was measured by direct immunoassay. With both glucose and arginine stimulation, portal vein levels of all three peptides peaked at 90-120 s after the onset of the stimulus. Relative increases in insulin concentration were greater than those of proinsulin or C-peptide. In peripheral venous blood, maximal levels of the three peptides were observed later (2-5 min), and the increase in insulin relative toproinsulin and C-peptide was not as great. At the time of peak secretion, portal vein insulin and C-peptide approached equimolar concentrations, and proinsulin, as measured against an insulin standard, comprised approximately 2.5% of the total immunoreactive insulin. After stimulation by glucose or arginine, portal insulin, proinsulin and C-peptide levels were not correlated with the concentrations measured in simultaneously drawn peripheral samples. At all sampling times, however, significant correlation was found between insulin and C-peptide in both peripheral and portal blood. The results indicate that under the conditions studied, insulin and C-peptide are secreted in equimolar concentrations in man, and that proinsulin is secreted in the same proportion to insulin as found in the pancreas. Consideration of the relative secretory and metabolic rates of the three beta cell peptides explains their peripheral concentrations. The data further support the use of plasma C-peptide as an indicator of beta cell secretory function.