Tests were made on three out of the four bottles. The fourth bottle was not opened, as we desired to learn what change it might undergo in the way of protein precipitation and bacterial decomposition. There is nothing in the Beveridge method of preparation that insures a sterile secretin unless it is passed through a Berkefeld filter. In all our crucial experiments the animals (dogs) were kept under light ether anesthesia, a cannula inserted into the pancreatic duct, the blood pressure recorded from the carotid artery and the various secretin preparations injected intravenously. When inactive secretin preparations were encountered, control tests were always made with active solutions of secretin to eliminate possible individual peculiarities of the animal. Thus when the pancreas of a dog reacts to the injection of preparation _A_, but not to preparation _B_, it is evident that absence of response to _B_ is due to this preparation and not to the animal or to the experimental conditions. [Illustration: Fig. 1.--Records of carotid blood pressure and secretion of pancreatic juice on intravenous injection of Beveridge’s secretin in dogs. _X_, injection of 10 c.c. secretin; _b_, record of flow of pancreatic juice in drops. Tracing _A_, injection of 10 c.c. of one sample secretin (ten days old) furnished by Dr. Beveridge. Tracing _B_, injection of 10 c.c. of second sample of secretin (ten days old) furnished by Dr. Beveridge. Tracing _C_, injection of 10 c.c. of secretin (twenty hours old) made by us according to the Beveridge method. Showing that the secretin preparations sent us by Dr. Beveridge contained no secretin.] Each of the three samples of secretin sent us by Dr. Beveridge was tested in the above manner on five dogs. The first tests were made June 27, 28 and 29, respectively, that is, within nine days of the preparation of these samples of secretin. _None of the samples was active (Fig. 1), even when injected intravenously in quantities up to 50 c.c._: 40-50 c.c. of Beveridge’s secretin mixture may kill a dog by too great lowering of the blood pressure. A good secretin preparation yields a copious secretion of pancreatic juice on intravenous injection of a few cubic centimeters. It is not difficult to prepare a secretin, by the original Bayliss or Starling method or by the Beveridge method, that retains some activity for a longer period than nine days. Hence we cannot account for the absolute inactivity of these preparations except on the assumption that they did not contain any secretin to start with; that is, faulty preparation and absence of physiologic standardization.