Eight dogs were prepared with unilateral carotid loops, and trained to stand quietly in a modified Pavlov stand. They were deprived of water for 24 h, which significantly depleted both the cellular and extracellular fluid (ECF) compartments. When they were allowed access to water for 5 min, they promptly drank amounts sufficient to make up their fluid deficits. Infusion of water into the unilateral carotid loop to reduce the cerebral blood osmolality to normal did not consistently reduce drinking because of inadequate mixing in the Circle of Willis. The eight dogs were then prepared with bilateral loops and deprived of water. Infusion of water at 0.6 ml-kg-1-,min-1 reduced the jugular plasma osmolality to control levels but did not affect systemic osmolality and reduced the drinking by 72%. Intracarotid infusion of water at differing rates reduced drinking in a dose related fashion. Intravenous infusion of 0.15 M NaCl in an amount designed to expand the ECF volume to normal reduced drinking by 27%. Expansion of the ECF volume plus bilateral intracarotid infusion of water to remove the central osmotic stimulus completely inhibited drinking. It is concluded that drinking following water deprivation in dogs is controlled by both central osmotic and by extracellular fluid volume factors, and that the osmotic factor is the more important. The drinking stopped well before these fluid deficits, as judged by body fluid analysis, had been restored. Some other mechanisms, therefore, must be involved in satiety.