To characterize the abnormal B-cell response to glucose in type II diabetes, five diet-treated diabetic and six weight-matched non-diabetic subjects were studied using the hyperglycemic clamp technique on three separate days at glycemic levels of 7.5, 10 and 15 mmol/L for 150 minutes with assessment of plasma insulin and C-peptide responses. To reduce possible secondary effects of hyperglycemia, diabetic subjects on a weight-maintaining diet were chosen who had only a slight elevation of the fasting plasma glucose, mean 6.0 mmol/L. They had a normal time-course of both first- and second-phase responses, but both were impaired at each glucose clamp concentration. The first-phase and second-phase C-peptide responses of the diabetic subjects were similarly reduced to mean 49% and 59% of normal, respectively, and the first- and second-phase insulin responses were also reduced to mean 39% and 44% of normal, respectively. The ratio of second- to first-phase plasma C-peptide responses were similar in the diabetic and normal subjects, median 1.6 and 1.5, respectively, as were the same ratios for the insulin responses, 1.4 and 1.1, respectively. The previously described selective reduction of the first-phase response in type II diabetes may be partly a function of the bolus intravenous glucose tests used, in which impaired glucose tolerance in the diabetics gave a greater glycemic stimulus to the second phase than in normal subjects, and partly secondary to long-term hyperglycemia. The diabetic subjects were re-studied after treatment with a sulphonylurea, gliclazide, with a normal fasting plasma glucose, mean 5.1 mmol/L. Sulphonylurea treatment increased both the first and second phases to near normal when assessed by C-peptide but not when assessed by insulin levels. Insulin sensitivity, assessed by the glucose disposal relative to the insulin concentrations during the clamp, was not altered by sulphonylurea treatment.