In all methods for the assay of secretin now in use, the hormone is intravenously injection into the experimental animal and the response of the pancreas is recorded. The animals used have been dogs, cats, rabbits and rats (1). The rat method was elaborated in 1957 by Love (2), and Svatos and Jelinek (3). Although this rat method was shown impractical by Lin and Alphin in consequence of their comparative experiments between rats and dogs (4), it has been used by some groups (5-7). The technique of the rat method has been recently reinvestigated by Heatley and he shows that secretin can be assayed in rats (8). For several years we have worked up the purification and economic production of secretin. In the course of this work the necessity arose for developing a simple method for the assay of secretin content in large numbers of preparations. Mutt and Soderberg found that if certain precautions were taken, anesthetized cats could be used for the assay of secretin for several days, thereby eliminating the necessity of time-consuming daily operations (9). But their valuable method which can be used to follow the activity of secretin in the course of purification is not suitable for determining precisely the quantity of secretin in preparations. Except Dorchester and Haist who used rabbits, none of the experiments have been statistically done for the estimation of secretin potencies (10). We chose rats for the assay of secretin, because we could easily obtain the rats whose strain, age and body weight were nearly similar. The following facts were confirmed. Secretin could be satisfactorily assayed in anesthetized rats by the measurement of the volume of pancreatic juice secreted. The technique of the rat method was very simple, and if the twin crossover design which was used to determine various hormone potencies was applied, the expected standard error of the estimate was 10 to 15 percent. Furthermore secretin could be also assayed if the bicarbonate output in the pancreatic juice was determined, which changed linearly with the log doses of secretin. Because of the lack of the international unit and a stable standard of secretin, there is no precise comparison between the potencies of the various units reported such as rat, dog, cat and Clinical unit yet (1, 5). The variation in the secretin potency between two brands of commercial secretins has been recently found to differ significantly (11). The relative potency between Crick, Harper and Raper unit (12) and Swedish Clinical unit (13), both of which were used in the clinical diagnosis for the pancreatic disorders, was determined by this rat method. The potency of 1 Clinical unit was 2 to 4 times as strong as that of the former.