This study addressed the problem of monitoring serum glucose levels in persons suffering with diabetes mellitus.The purpose was to investigate the relationship between the renal threshold for glucose in diabetics and age, sex, type and duration of diabetes. The effects of insulin dose, race, and serum creatinine were investigated. Difference between double-void and triple-void urine glucose concentration was studied.The sample consisted of 40 adult diabetics who met selected criteria. The independent variables were determined from subjects' hospital records. The renal threshold for glucose was estimated by comparison of serum glucose concentrations and urine glucose grades. During a one-hour period blood was drawn and a double-void urine specimen and a triple-void urine specimen were collected. Blood and urine were analyzed for glucose with glucose oxidase as the reagent. Beginning with the morning after hospital admission, blood and urine were collected and analyzed three times each day for four days.During this study the term “estimated threshold interval” was used and defined. The terms “negative boundary point,” “positive boundary point,” and “inverted boundary points,” which are closely associated with the estimated threshold interval concept, were also used.Analysis of the data revealed that during 21 percent of the test periods the glucose concentration of double-void and triple-void urines differed. Decreases accounted for 73 percent and increases accounted for 27 percent of the changes. By employing the negative and positive boundaries of the estimated threshold interval associated with double-void and triple-void urine, four separate analyses of variance were made. The renal threshold for glucose had wide variability among individuals. The means of the negative boundary associated with double-void urine was 213; the positive boundary of double-void urine was 216; the negative boundary of triple-void urine was 215, and the mean of the positive boundary of triple-void urine was 243. The greatest probability found that any of the independent variables influenced the renal threshold for glucose was the effects of age (p = .08) and insulin dose (p = .07)on the negative boundary point associated with double-void urine. It was evident from theR2= 0.328 that the variability of the renal threshold for glucose is influenced by variables other than those studied.Concepts defined in this study may be useful to researchers who want to study the renal threshold for glucose of patients in the clinical area.