The site of action of phlorhizin in phlorhizin diabetes is now recognized as being in the kidneys, the disturbance caused by the drug being apparently due to inability of the tubule to reabsorb glucose. This, according to Lundsgaard, arises from the fact that the poisoned tubule is unable to carry out phosphorylation of glucose. Moreover, several glucosides, including phlorhizin, are now known to inhibit phosphorylation in vitro, and glucose resorption in the isolated perfused frog kidney is also inhibited. It occurred to us that the glycosuric effect of phlorhizin might be a property common to all these glucosides. The action of arbutin, amygdalin, and salicin on healthy, fasting dogs was studied according to the procedure and methods of urine and blood analyses customary with phlorhizin. Arbutin is, like phlorhizin, a phenol glucoside. When injected subcutaneously in sterile distilled water twice daily in fasting dogs, it caused a diabetes similar to that produced by phlorhizin with lowered blood sugar and acetone bodies in the urine, the ratio of dextrose to nitrogen in the urine being 1.45, 1.56, 1.50, and 1.92 in 4 different dogs. The urinary reducing substance was dextrose, for after yeast fermentation the urine no longer reduced Benedict's reagent. In contrast to the diuresis of phlorhizin diabetes, the urine of the animals receiving arbutin, likewise of those receiving the amygdalin and salicin, was very scanty unless water was administered by stomach tube, but the voluntary intake of water also was very small. Three fasting dogs injected subcutaneously with salicin, an alcohol glucoside, in sterile distilled water exhibited no glycosuria nor ketonuria; nor did 3 dogs similarly injected with amygdalin, an aldehyde glucoside.