Serum levels of insulin (IRI), total glucagons (GLI), cortisol, glucose and free fatty acids were measured in 14 normal subjects and in 6 diabetic patients. All of them were hospitalized, but not con- , fined to bed and were given 3 daily meals of identical carbohydrate, fat and protein content. Blood samples were collected through an indwelling venous catheter at hourly intervals for 72 h. In addition, three 15-min samples were collected after each meal. The study was repeated about 1 month later, during the last 3 days of a period of chlorpropamide treatment. The results confirm the existence of a diurnal cycle in serum cortisol in normal and diabetic subjects. Diurnal variations in the postprandial glucose response of normal subjects and in the postprandial IRI response of diabetic patients were observed, but they may have been drug-induced. Chlorpropamide treatment had no significant effects in normal subjects, but was followed by a decrease in the concentration of serum GLI and by an increase in the fasting levels of IRI and in the IRI response to meals in diabetic patients. These changes may explain the restraining effect that the drug had on postprandial hyperglycemia. The results of our experiments suggest that the therapeutic effect of chlorpropamide in adult onset diabetic patients may be due to a combination of an increased release of insulin and a decreased release of total gucagon in response to the ingestion of food. The difference in the response of normal and diabetic subjects to prolonged chlorpropamide treatment remains unexplained, but may account for some of the conflicting reports on the effects of sulfonylurea derivatives on insular function.