Verney's hypothesis of cerebral osmoreceptors controlling the renal excretion of water via vasopressin was reinvestigated in conscious trained dogs provided with bilateral skin loops containing the common carotid arteries. In multiple experiments in two dogs, bilateral intracarotid injections (0.25 ml.(kg b. wt.)‐1per artery in 10 s) of a hyperosmotic solution of sodium chloride (0.257 mol/l) during transient water diuresis failed to produce an antidiuretic response, although it is estimated that the injections elevated the osmolality of the carotid blood by 12–15%. In another 5 dogs, bilateral intracarotid infusions of hyperosmotic saline (45 μmol. (kg b. wt. min)‐1per artery for 10 min) during sustained water diuresis resulted in a 3% increase in jugular venous osmolality and an antidiuretic response without detectable changes in heart rate or mean arterial pressure. Equal intravenous hyperosmotic or intracarotid isosmotic infusions were not associated with antidiuretic responses. Analysis of the concomitant concentrations of vasopressin in plasma fell short of supporting the hypothesis that the antidiuretic response to intracarotid hyperosmotic infusions was exclusively or mainly due to liberation of vasopressin. although the renal response could be mimicked by exogenous vasopressin. It is concluded that the present results—although discordant with several of Verney's results and assumptions—nevertheless support the concept of a cerebral solute receptor influencing the rate of renal water