The substrates for hepatic ureagenesis are equimolar amounts of ammonium and aspartate. The study design mimics conditions in which the liver receives more NH(+)(4) than aspartate precursors (very low-protein diet). Fasted dogs, fitted acutely with transhepatic catheters, were infused with a tracer amount of (15)NH(4)Cl. From arteriovenous differences, the major NH(+)(4) precursor for hepatic ureagenesis was via deamidation of glutamine in the portal drainage system (rather than in the liver), because there was a 1:1 stoichiometry between glutamine disappearance and NH(+)(4) appearance, and the amide (but not the amine) nitrogen of glutamine supplied the (15)N added to the portal venous NH(+)(4) pool. The liver extracted all this NH(+)(4) from glutamine deamidation plus an additional amount in a single pass, suggesting that there was an activator of hepatic ureagenesis. The other major source of nitrogen extracted by the liver was [(14)N]alanine. Because alanine was not produced in the portal venous system, we speculate that it was derived ultimately from proteins in peripheral tissues.