The site where osmotically active substances act to depress food intake was determined in sheep. After 5.5 h of food deprivation, solutions of sodium chloride or polyethylene glycol-200 (PEG-200) were added to either the reticulo-rumen or the abomasum. The sheep were then immediately offered pelleted lucerne (Medicago sativa). Water was withheld during the first 60 min of feeding but was available from 60 to 90 min. There was a linear inhibition in food intake in the first 10 min after loading 2.37, 6.25, 12.5, 25.0 or 50.0 g NaCl into the rumen according to a 5 x 5 Latin square design (P = 0.0001). The intake reduction was 3.49 g food/g NaCl. An osmotic load of PEG-200 equivalent to 50 g NaCl also significantly inhibited food intake in the first 10 min of the meal compared with a control treatment. The inhibition of food intake after loading 55 g NaCl into the rumen was not affected by injecting lidocaine hydrochloride into the reticulum immediately before NaCl loading. NaCl injected into the abomasum did not significantly affect food intake in the first 10 min of feeding even though the tonicity of abomasal digesta was increased to unphysiological levels. There was no consistent relationship between food intake and the change in the tonicity of jugular plasma following solute loading and drinking. The sensing site of hypertonicity was localized to the wall of the reticulo-rumen where neuronal receptors appear to be capable of detecting osmotic pressure within the physiological range to depress food intake. These receptors should be identified and characterized because of their possible significance in limiting food intake by ruminants.