Experiments were carried out on two groups of diabetic patients treated (1) by diet alone (group A) and (2) by diet and oral agents (group B), the latter being discontinued 3 days before each test. All patients were tested twice. To measure rates of glucose metabolism, 3H-3-glucose was infused before and during a standard 3-h 50'g oral glucose tolerance test. 14C-1-glucose was added to the glucose solution and the percentage of ingested glucose that appeared in peripheral blood was calculated. Bran improved glucose tolerance only in group B patients by delaying the peripheral appearance of ingested glucose but had no effect on glucose absorption in group A. In contrast, patients in group A showed a marked reduction in their insulin response when bran was mixed with the glucose. Despite this reduction, glucose tolerance, the metabolic clearance rate (MCR) of glucose, arid all other rates measured were unaffected by bran. Thus we conclude that in patients adequately controlled by diet alone, the effect of insulin was potentiated when bran was ingested, but the mechanism involved remains obscure.