In fishes the sugar of the blood and the glycogen of the liver have been found to vary considerably even in individuals of the same species, and still more so in those of different species. Practically nothing is known definitely of the causes for these variations, and this we consider an important problem to investigate, especially since light might thereby be thrown on the nature of the metabolism of carbohydrates in cold-blooded animals in which the intermediary stages proceed more slowly than in warm-blooded animals. Our interest was aroused in the behaviour of the blood sugar of fishes for other reasons as well. In certain of the bony fishes (Teleostei) the islets of Langerhans exist as definite glands which have come to be known as the “principal islets.” Being more or less separated from the pancreatic tissue itself, these can readily be excised, thus making it possible, by examination of the blood sugar, to determine whether a diabetic condition can be induced by isletectomy without removal of any of the pancreas proper. It was of interest also to see whether insulin can affect the blood sugar. Before such investigations could be undertaken it was necessary to know exactly the degree to which the blood sugar of different fishes of the same species may vary independently of such an operation. Lang and Macleod (1), in confirmation of earlier work by Diamare (2) and of Bierry and Fandard (3), found that there are usually only traces of sugar in the blood of the Elasmobranchi, such as Squalus (dog-fish), but that considerable amounts may occur in the blood of representative Teleostei, such as Cyprinus (carp). In the latter fish it was also noted that the amounts may vary from 0·058 to 0·300 per cent. Fandard and Ranc (4) have stated that the blood sugar in fishes is peculiarly susceptible to asphyxial conditions, but so far as we have been able to find they have published no details of their observations. The most important recent work is that of E. L. Scott (5), who has observed the blood sugar in Mustelis canis , the fish prior to the observations being kept in traps which were exposed to tide water and, during them, in shallow tanks. The percentage of oxygen was also frequently determined in the water of the tanks. It was found that no blood sugar, or only traces, could be detected in six out of eight individuals, which are described as having been in a subnormal condition. On the other hand, when the fish were asphyxiated by keeping them out of water for varying periods of time, the blood sugar rose rapidly, to attain, in two specimens, a maximum of about 0·240 per cent, after four minutes, followed by a gradual decline, so that a level of 0·032 was reached in one specimen after 15 minutes. The degree of variability in the results is, however, very great, and they do not seem to us to justify the conclusion that the sugar rises within a few minutes and then falls again during the asphyxial period.