In 1939, Marble1 set forth criteria for the diagnosis of renal glucosuria—proved glucosuria in the postabsorptive state, relative independence of the urine sugar on blood level of glucose, no progression to diabetes for at least 3 years, normal glucose tolerance test, and constancy of fasting glucosuria. If these criteria are accepted, the subsequent development of diabetes mellitus is a rare event. The combined experience of Marble, Robbers and Rümelin, and Wilder,2-4 numbering several hundred cases of renal glucosuria followed for as long as 64 years, yields no more than 3 patients who later developed diabetes mellitus. In 1950, Callaway5 could find only 6 patients in the literature in whom renal glucosuria was later associated with diabetes. Gräfe and Hering6 reported 2 more patients, and Reubi7 cites one. Ackerman, Fajans, and Conn8 applied Lawrence's9 less rigorous criterion for the diagnosis of renal glucosuria—appearance of