In the course of some experiments on the quantitative excretion of sugars after parenteral injection, we had occasion to use Bang's1 method for the determination of glucose and experienced, as have others who have used this method, the difficulty of determining sharply the end point of the reaction when dealing with urines that were even moderately colored. This difficulty is obviated by clearing the urine before titration and, according to Bang and Bohmannsson,2 this clearing is readily accomplished without the loss of any appreciable amount of glucose by shaking the urine with blood charcoal after having rendered the urine 5 per cent. acid by means of hydrochloric acid. The loss as indicated by Bang and Bohmannsson under these conditions is so small that it can be neglected. Our first control experiment, however, showed quite the contrary, and further determinations indicated losses that amounted to 61 per cent.