The intraportal injection of 3–12 g of glucose through a chronic cannula produced a 40–75 min anorexia in 22 hr fasted dogs that would normally eat 500 g of dry food without interruption. The same amount of glucose injected into a jugular cannula did not elicit any anorexia. The intraportal injection of an equal volume of isosmotic NaCl did not produce any noticeable change in the feeding behavior of the same dogs. Therefore, the satiating effect was obtained only when the glucose went directly into the liver and not when distributed throughout the organism.