SUMMARY Fourteen studies of gastric acid secretion in a basal hour and in the two hours after a single intravenous injection of soluble insulin (0.01 to 0.40 units/kg) were performed in a healthy man. The peak acid output after insulin (measured as the two consecutive 15-minute samples giving the highest acid output) was significantly correlated with the lowest concentration of blood glucose, the fall in blood glucose, the rate of fall of blood glucose, and the maximum fall of blood glucose in any 15 minutes. Peak acid outputs after insulin were similar over the range 0.1 to 0.2 units/kg, and greater than at lower or higher doses.These results are contrary to the accepted assumption that insulin-stimulated acid secretion is an 'all-or-none' phenomenon. They support instead the hypothesis that insulin hypoglycaemia provides a quantitative glycopenic stimulus producing quantitative vagal acid response. Extreme hypoglycaemia, below about 15 mg/100 ml of bloodglucose,inhibitsinsulin-stimulated acid secretion.Certain assumptions about the insulin test have been widely held (Roholm, 1930;Hollander, Jemerin, and Weinstein, 1942) in the 40 years since intravenous insulin was found to be an hypoglycaemic stimulus of gastric acid secretion (Simici, Popescu, and Diculesco, 1927). These assumptions are that the acid response is an 'all-or-none' phenomenon which is (1) initiated when the blood sugar falls to a threshold value of 40-50 mg/100 ml; (2) not related to the degree of hypoglycaemia below this threshold; (3) not dependent on the fall in blood sugar; and (4) not related to the rate of fall of blood sugar.