This study describes the contribution of de novo glucose synthesis by the kidney to blood glucose homeostasis in rats. The net glucose release by the kidney in vivo was measured by an isotope-dilution method, which calculated the extent of dilution of injected [(14)C]glucose by glucose newly synthesized in the kidney. The extent of dilution was determined from the difference between the decrease of the actual blood glucose concentration and that of the radioactivity of [(14)C]glucose, after injecting [(14)C]glucose into functionally hepatectomized rats. The results indicate that the net glucose release by the kidney in vivo in normal fed rats was 0.75+/-0.13 mg/dl per min, and that its contribution to blood glucose was 25.9+/-5.0%. When unilateral nephrectomy was performed, under the same conditions, renal net glucose release was one-half of that in rats with two intact kidneys, which indicates the quantitative accuracy of the isotope-dilution method employed in this study. In rats starved for 24 h, the renal net glucose release increased to 0.99+/-0.08 mg/dl per min. Diabetic rats showed a remarkably higher renal net glucose of 2.28+/-0.33 mg/dl per min, which was 360% of the normal level. Treatment of diabetic rats with insulin, restored the renal net glucose release to the normal level. In acidotic rats, renal net glucose release was as great as 1.03+/-0.15 mg/dl per min, which suggests that the acid-base balance participates in control of renal glucose output. Measurements every 6 h throughout the day showed that glucose was supplied from the kidney at a constant rate without any circadian rhythm. These data suggest that renal gluconeogenesis is of physiological importance in the maintenance of homeostasis of blood glucose.