Serum levels of acetoacetate, 3-hydroxybutyrate and the 3-hydroxybutyrate/acetoacetate ratio were determined in Type 1 (insulin-dependent) and Type 2 (non-insulin-dependent) diabetic patients by a new sensitive method. Efforts were made to differentiate Type 1 and Type 2 diabetes by serum levels of ketone bodies and to determine whether their measurement is a useful way of monitoring diabetic control. In Type 2 diabetes, serum levels of total ketone bodies did not exceed 2.0 mmol/l even if the patients were untreated or poorly controlled. In Type 1 diabetic subjects, treated with once or twice daily injections of insulin, morning serum levels of acetoacetate, 3-hydroxybutyrate and total ketone bodies were significantly elevated by four-, ten- and sevenfold, respectively. In Type 2 diabetic subjects treated with diet or sulphonylureas, serum levels of 3-hydroxybutyrate were highest before breakfast, next highest before dinner and decreased after each meal. The changes were roughly inversely proportional to serum insulin levels. In addition, insulin treatment normalized fasting serum levels of ketone bodies better than diet or sulphonylurea treatment. Acetoacetate was also significantly increased in both types of diabetes to a lesser extent, but no apparent diurnal rhythm was observed. Determination of serum levels of ketone bodies is useful for the diagnosis of Type 1 diabetes (those with total ketone bodies greater than 2 mmol/l) and for detecting insufficient insulin therapy.