Whole representative groups or randomly selected samples of Indians (of Asian origin) (1,520), Malays (1,252) and African Bantu (1,029) living in the Cape Town area were interviewed by social workers and screened by urine and blood sugar estimations after either a heavy meal or 50 gm. of glucose. Subjects who “screened positive” were subjected to full glucose tolerance tests, on which the final diagnosis of “discovered diabetes” was based. The screening methods used were of approximately comparable sensitivity,and the two sexes were almost equally represented in each age group and race. The difference between postprandial and post-glucose blood sugar was not consistent, and rose with age. The meanscreening blood sugar levels of adults rose with age, but in young adults it was no higher than in children. There was no difference between the sexes, except apparently in the Malays. Africans had the lowest and Malays the highest mean blood sugar levels at all ages. (Known diabetics were excluded from these observations.) The known diabetes rate among Indians was 4 per centover age fifteen, or 10.2 per cent when age-corrected. This is one of the highest prevalence rates known. Known diabetes (0.9 per cent) among the Africans was very similarto that usually reported among white populations. The total (known and discovered) diabetes prevalence among Indians was 10.4 per cent over age fifteen, rising to 60 per cent in women over fifty-five; among Africans it was3.6 per cent over age fifteen, with more diabetes in middleaged than in elderly people. Among the Malays the prevalenceof diabetes was intermediate (6.6 per cent over age fifteen). Comparisons with other reported prevalence rates are extremely difficult because of great discrepancies due to differences in methods and interpretation. Thus whenever glucose loads and blood sugar estimations are used for screening, the prevalence is comparatively high; it is also high when a single blood glucose level is used for diagnosis. Juvenile diabetes was rare; we found no known diabetic under age twenty and no African diabetic under thirty-five. Two seventeen-year-old Malay girls and one fifteen-year-old Indian were discovered to have asymptomatic diabetes. Glycosuria was rare under age fifteen; it is argued thatthe renal threshold may fall with age. Glycosuria prevalence rose with age and was more common in men than women. The presence of glycosuria even after glucose was inconsistent. Glycosuria after glucose was more frequent than after a meal; among the Indians, postprandial glycosuria screening would have missed two thirds of the discovered diabetics, whereas all but one had glycosuria after glucose. Ninety-one per cent of the total newly discovered diabetics had glycosuria at one or another time of testing