It is not generally appreciated that the diagnosis of chemical diabetes by oral glucose tolerance testing has many pitfalls and that many patients with a diagnosis of diabetes based solely on abnormal glucose tolerance testing in fact do not have true diabetes mellitus. Once the clinician sees an abnormal glucose tolerance test his major objective is to exclude all the nondiabetic factors which may have influenced the testing procedure, thus giving a false-positive result. Furthermore, the standards used in the interpretation of the test remain uncertain in older patients thus markedly reducing the usefulness of the procedure in this group. In addition, the test is of limited value and therefore probably should not be performed in hospitalized or chronically (or acutely) ill patients. Finally the detection of glucose intolerance (in presence of fasting normoglycemia) is rarely of benefit to the patient in the absence of obesity and may prove a hardship for psychosocial reasons. Thus the clinician should carefully evaluate the medical indications and the potential benefits derived prior to ordering an oral glucose tolerance test. He should be very conservative in making a diagnosis of diabetes mellitus based on this test.