This investigation was undertaken partly on account of its physiologic interest and partly in the hope that it might yield information which would enable the clinician to interpret blood sugar values taken at any time of the day. The blood sugar was determined at hourly intervals in cases of diabetes mellitus. These patients were ordered diets which were adjusted to the therapeutic needs of the individual. That is, the diets were either "carbohydrate-free," containing no starch except that found in green vegetables, or, save for a very few instances, limited in starch content, so that the glycosuria was held in abeyance or at a low level. The results obtained under these circumstances, while they do not exhaust the subject from the physiologic or pathologic-physiologic point of view, are applicable to the practical interpretation of blood sugars in the treatment of diabetes mellitus. The method of Lewis and Benedict1 was employed