Control of urea synthesis was studied in rat hepatocytes incubated with physiological mixtures of amino acids in which arginine was replaced by equimolar amounts of ornithine. The following observations were made.1. Intramitochondrial carbamoyl phosphate was always below 0.1 mM. Only when ornithine was absent and when, in addition, the concentration of amino acids was higher than four times their plasma concentration, intramitochondrial carbamoyl phosphate rose up to about 3 mM ; under these conditions ammonia accumulated in the medium.2. The relationship between ornithine-cycle flux and the concentration of the cycle intermediates at varying amino acid concentration indicated that under near-physiological conditions the ornithine-cycle enzymes are far from being saturated with their substrates.3. Moderate concentrations of norvaline had no effect on the rate of urea synthesis unless the cells were severely depleted of ornithine.4. Activation of carbamoyl-phosphate synthetase (ammonia) by addition of N-carbamoylglutamate only slightly stimulated urea production at all amino acid concentrations. However, in the presence of the activator the curve relating ornithine-cycle flux to the steady-state ammonia concentration was shifted to lower concentrations of ammonia.5. The intramitochondrial concentration of carbamoyl phosphate in rat liver in vivo was below 0.1 mM. This value is far below the concentration required for substantial inhibition of carbamoyl-phosphate synthetase.6. It is concluded that in vivo the function of activity changes in carbamoyl-phosphate synthetase, via the welldocumented alterations in the intramitochondrial concentration of N-acetylglutamate, is to buffer the intrahepatic ammonia concentration rather than to affect urea production per se. At constant concentration of ammonia the rate of urea production is entirely controlled by the activity of carbamoyl-phosphate synthetase.Short-term control of urea-cycle flux has in the past been the subject of many studies. In reviewing the literature on this topic [I] we have concluded that: (a) the capacity of the urea cycle, i.e. with saturating intracellular concentrations of ammonia, ornithine and aspartate, is determined by the V of argininosuccinate synthetase; and (b) that in vivo, with nonsaturating substrate supply, flux through the urea cycle must be primarily controlled by the activity of carbamoyl-phosphate synthetase, a conclusion based on qualitative rather than on quantitative considerations. It is also clear from the literature that in vivo modulation of carbamoyl-phosphate synthetase activity occurs by variations in the intramitochondrial concentration of N-acetylglutamate, the essential activator for the enzyme [l -31. However, this latter concept has been questioned [4].