The study of thirst in rats and rabbits by a controlled technique showed that either of two kinds of stimuli was necessary to induce drinking: depletion of body water or excess in the body of a solute in a concn. hypertonic to blood plasma. Diverse doses of NaCl (or urea) were added by three routes. Drinking followed upon their absorption into the circulation. Diverse rates of chloride absorption from the three sites of admn. largely accounted for the differences of drinking. Water was drunk in sufficient amts. to dilute the salt to about 0.4/meq./l., and to excrete it in that concn., but never in sufficient amts. to render the salt isotonic with control body fluids. Drinking could be partially inhibited by severe restraint, filling of the stomach, of water by other routes, injn. of large doses of Pitressin, cocainization of the mouth, or placing high concns. of solutes in the drinking water. The only method found of enhancing drinking in rats was to place dilute solutes (especially NaCl) in drinking water. Physiological adaptations were identified in wt. losses during deprivation of water, and in drinking after NaCl admn.