Investigators in the field of water metabolism have generally taken it for granted that all the water in the body is fixed by the colloids and electrolytes in a firm combination and that changes in this colloid-water complex must be postulated to account for any variations in the tissue fluid concentrations. Fischer,1 indeed, postulated a "free" and "combined" water in the tissues, but the "free" water, according to his conception, seemed to be merely the superfluous fluid derived from metabolic processes or from ingestion which was promptly excreted. During the course of studies of the so-called dehydration fevers in dogs it was noted that the diuresis produced by injections of equivalent amounts of glucose varied within wide limits and in many instances failed entirely. These injections were made with the Woodyatt pump. Merck's "white granular glucose" was used. Balcar2 injected 36 per cent. glucose solution at the rate of 15