To assess the contributions of leucine and alpha-ketoisocaproate (KIC) derived from the portal vein vs. hepatic artery for hepatic protein synthesis, 14-postabsorptive dogs were infused simultaneously with [1-14C]- and [4,5-3H]leucine or [1-14C]- and [4,5-3H]KIC. On one occasion one tracer was infused via a leg vein and the other via mesenteric infusion catheters, and dogs were restudied with both tracers infused systemically. The ratios of systemically to portally infused tracers in portal and arterial plasma leucine were used as indexes of the radioactivity in the potential precursor pools and in fibrinogen-bound leucine as a paradigm of hepatic protein synthesis. In the dogs given leucine tracers, the relative proportions of systemically to portally infused radioactivity in portal free leucine (0.50 +/- 0.06) were lower (P less than 0.001) than in arterial free leucine (1.22 +/- 0.03) and not different from that bound in fibrinogen (0.43 +/- 0.02). In the dogs infused intraportally with KIC, these values were 0.81 +/- 0.04, 0.97 +/- 0.05, and 0.74 +/- 0.05, respectively. In the control studies these ratios were not significantly different from the expected value of 1.0. The results suggest that, in postabsorptive dogs, fibrinogen is exclusively synthesized from portally delivered leucine with little or no contribution from the hepatic artery, whereas portally delivered KIC contributes little directly to fibrinogen synthesis. These data are consistent with zonation of hepatic amino acid metabolism and/or protein synthesis.