Neither older (Raiman, 1902; Ehrenberg, 1909; Wigert, 1918; Reiter, 1925) nor more recent papers (Diethelm, 1936; McFarland and Goldstein, 1939; Holmgren and Wohlfarhrt, 1944) provide general agreement about impairment of glucose tolerance in mental disease. Plum (1958) considers a high incidence of decrease among epileptics established, but while Freeman and Zaborenke (1949) reported a decrease in schizophrenics Shattock (1950) found an increased tolerance in such patients. Waelsch and Weil-Malherbe (1964) comment that a reduction in the rate of glucose utilization in cases of severe depression stands out as perhaps the most significant and constant finding, but Coppen (1967) claims that quite recent work by him and his co-workers (Herzberg, Coppen and Marks, 1968) provided evidence that there was no impairment in glucose tolerance when depressed patients were given diets high in carbohydrate for three or four days before testing. It would be expected that much conflicting evidence could be due to difficulties inherent in psychiatric nosology. But a review of the surprisingly large number of reports which have left the problem unsolved could not but impress one with the rather casual attitudes of some workers to the chemistry of different methods for “blood sugar” determination.