Part 1. Beta-cell secretion was stimulated maximally by an intravenous glibenclamide-glucose load in four groups of 10 adultonset diabetics divided according to their therapy, which was determined by clinical criteria. Under the standardized conditions of this test, the 20 patients requiring insulin had a mean fasting blood glucose (BG) of 255 mg./100 ml. and showed no stimulation of immunoassayable C-peptide (IMCP) by glibenclamide-glucose. In contrast, the mean fasting BG was 160 mg./100 ml., but the IMCP rose in patients who were able to be managed with sulfonylurea tablets. These differences in IMCP proved to be of value for the clinician in predicting the efficacy of diabetes therapy with either insulin or sulfonylurea tablets. Part 2. Twenty-four-hour profiles of BG and IMCP were per performed in 28 insulin-requiring diabetics (age range: 14 to 74 years) who were 66 to 111 per cent of normal weight and had had diabetes for 0.2 to 31 years. These patients were divided into diabetics with IMCP (n = 14) and without IMCP (n = 14); the mean age and weight were similar in the two groups. The mean IMCP value and its standard deviation correlated inversely with the mean blood glucose (MBG) in all diabetics and healthy subjects. According to clinical criteria, diabetes control was better in the patients with IMCP. Analysis of variance showed that the mean BG concentrations for diabetics with and without IMCP were statistically different (pwith IMCP were found to inject less insulin daily, and six of them were adequately treated with only one daily injection.