The effect of a 5-day continuous intravenous infusion of somatostatin (4.6 ng min-1 kg-1) was studied, using anoestrous ewes given 791 g dry matter per day of a 60:40 lucerne hay:oat grain pelleted diet from a continuously moving belt. 51Cr-EDTA, 103Ru-phenanthroline and lignin were used as markers to determine digesta mean retention times (MRT) by a continuous infusion-total sampling procedure. The somatostatin infusion increased the concentration of somatostatin in venous plasma within the physiological range from 10 to 76 ng/l, decreased plasma concentrations of prolactin and thyroxine, but had no effect upon plasma concentrations of insulin and glucagon. It had no effect upon digesta-free weight of the rumen and omasum but consistently decreased the weight of all post-ruminal segments of the gastrointestinal (GI) tract. The infusion increased the accumulation of digesta in the abomasum and caecum. Total MRT of all three markers in the entire GI tract was unaffected by somatostatin infusion, but the proportion of total MRT spent in the abomasum + small intestine + caecum increased and the proportion spent in the large intestine and rumen decreased. Somatostatin infusion decreased apparent endogenous abomasal secretion, increased water flow from the rumen and into the abomasum and decreased voluntary water consumption. It is proposed that the prime site of somatostatin action was in the abomasal to caecal region, where somatostatin-secreting D cells are found in greatest concentration, that effects observed in the large intestine and rumen may represent secondary compensatory mechanisms and that the effects observed were due to direct action of somatostatin and were not mediated by other GI hormones.