The normal rates of glucose excretion in man and dog were measured, and the effects of water diuresis, osmotic diuresis, and plasma glucose concentration on glucose excretion were studied by the clearance method. The possibilities of more complete glucose reabsorption and glucose secretion were examined with the stop-flow technique in dogs. In unfasted human subjects, glucose excretion rate was higher in the afternoon and evening and correlated with urine flow rate. In fasted human subjects, water loading did not alter glucose excretion; whereas, water loading of fasted trained dogs increased glucose excretion and clearance. With mannitol osmotic diuresis induced in dogs undergoing maximal water diuresis, glucose clearance was approximately proportional to the urine volume flow rate. In both human subjects and dogs, glucose clearance was not significantly altered by moderate variation in plasma glucose levels. Stop-flow experiments on anesthetized dogs demonstrated that under these conditions glucose is added to the urinary fluid stopped in renal pelvis or collecting duct; however, this phenomenon is probably not a significant determinant of glucose excretion during normal free-flow. Glucose reabsorption by the proximal tubule was not as complete during stop-flow as during free-flow and indicates that failure of glucose to be completely reabsorbed is not caused by insufficient time for reabsorption. The findings are consistent with the concept that glucose excretion is normally determined by the volume flow rate from the proximal tubule and by the glucose concentration gradient finally effected across the epithelium.