Glucose utilization studies were performed in 17 normal subjects, in 14 patients with mild diabetes, and in 9 patients with severe diabetes by two techniques,i.e.,singleinjection and continuous infusion of glucose. Data were analyzed by a new method which yields a valid specific rate constant of glucose utilization (K) not dependent upon the equation describing blood glucose disappearance. It was shown that previous methods of data analysis resulted in aggravated errors when applied to diabetics. Older normal subjects had significantly higher fasting blood glucose (Cf) and reference (Ceq) levels than younger normal subjects. In contrast to the younger group, the older group showed no significant difference between Ceq and Cf. K and glucose space (V) were similar in both groups. In all diabetics the single-injection method yielded a Ceq higher than Cf. This probably resulted from an increased hepatic glucose output despite a rising concentration of blood glucose. With the continuous-infusion method there was no evidence of this paradoxical response in the patients with mild diabetes. However, in those with severe diabetes, hepatic glucose output was so great that K and V were markedly distorted. K values obtained with the single-injection method in the patients with mild diabetes were not significantly different from those in the controls. This was interpreted as an artifact of the method of glucose administration. With the continuous-infusion method in the patients with mild diabetes, both K and V were significantly less than in the controls. It is concluded that: 1) the single-injection method yields an erroneous K value in all diabetics, whereas the continuous-infusion method yields a valid and clinically useful K value in patients with mild diabetes; 2) the relationship between Ceq and Cf (single-injection method) may also be used clinically; 3) in patients with more severe diabetes, gly-cosuria and/or increased hepatic glucose output often invalidate both methods; and 4) in the diabetic, glucose utilization and glucose space are significantly reduced and the normal hepatic control of blood glucose is disturbed.