We investigated the movement of 13N-labeled ammonia from blood to brain, in vivo, in adult rhesus monkeys. For this purpose, we monitored the behavior of the tracer by using external detection of the 511-keV annihilation-radiation photons of nitrogen-13 following rapid bolus injection into the internal carotid artery. Our data reveal a diffusion limitation for the transport of ammonia from blood to brain. We attribute this to the low permeability of the blood-brain barrier for the ammonium cation. At a measured cerebral blood flow of 51 (ml/min) per hg (hg = hectogram), for example, only 35% of the injected tracer leaves the vasculature and is incorporated into brain tissue. Further, this extracted fraction not only decreases with increasing cerebral blood flow, but is also influenced by the pH of the blood perfusing the brain and by the integrity of the blood-brain barrier. We have interpreted our data on the basis of a new regional model of the cerebral circulation that takes into account both capillary heterogeneity within an external detector spatial-resolution element and the effects of shifts in the degree of ionization of vascular radioammonia due to the existence of pH gradients along the direction of flow in capillary blood. We have thus obtained, apparently for the first time, estimates for the individual permeability coefficient-specific surface-area products for diffusive transport across the blood-brain barrier of the two aqueous solution ammonia species, NH3 and NH4+. These estimates, denoted PoS and P+S, respectively, are regional averages; our values and associated standard deviations are PoS = 0.12 ± 0.02 ml/sec per g and P+S = (5.5 ± 4.9) x 10-4 ml/sec per g. Assuming a regional average brain capillary specific surface area, S, of 100 cm2/g, these data yield the tentative values Po = (1.2 ± 0.4) x 10-3 cm/sec and P+ = (6±5) x 10-6 cm/sec.