It is axiomatic that early discovery of disease is desirable. Diabetic-detection drives have been launched in an effort to discover early cases of this disease so that treatment can be instituted before dangerous complications occur. Since there is a definite familial tendency in this disease, it has been our custom to examine all siblings of diabetics under treatment at the clinic. Occasionally, we have discovered an unsuspected case, but, even more frequently, we have found such relatives normal by present testing methods, only to discover some years later that they were diabetics. On several occasions such discovery took place only when a serious complication of the disease appeared. Consequently, we have performed glucose tolerance tests on all close relatives over the age of 50 of diabetics in a further effort to discover early diabetes. While these efforts were slightly more salutary, there were still many patients with normal tests who