During a Diabetes Detection Drive, 559 persons were tested for hyperglycemia (Auto Analyzer and Dextrostix methods) or glycosuria (Clinistix and Dreypak methods) ninety minutes after a glucose load. Of 164 screened positive, 154 returned for standard glucose tolerance testing. Using the criteria of Fajans and Conn, forty-two new cases of diabetes were diagnosed; by age-adjusted criteria there were twenty new cases. Approximately one-third of these would have been missed by reliance on screening by glycosuria alone. Comparison of screening methods reveals the Auto Analyzer measurement of blood sugar to be more sensitive and therefore better for the intensive study of limited populations; however, the Dextrostix method is cheaper and quicker and may have some advantages in the usual mass detection efforts.