We have examined the effects of two types of insulin therapy (continuous i.v. infusion (group A: 1.25 u. h-1) and direct i.v. bolus administration of 10 u. every 2 h (group B: 10 u./2 h)) on the metabolic and endocrine responses to surgery in 60 adult diabetic patients undergoing general anaesthesia for elective procedures. Blood glucose concentrations were measured every 15 min from just before induction of anaesthesia until 2 h after surgery. Plasma ketone bodies, lactate, pyruvate, insulin, C-peptide and counter-regulatory hormone concentrations were measured also. Blood glucose concentrations were comparable in both groups, except at 60 min, because of fluctuating blood glucose concentrations in group B. In each group, one patient became hyperglycaemic. One case of hypoglycaemia occurred in group B. There were no statistically significant differences for the other data except for C-peptide. We conclude that, during the operative period, the administration of a direct i.v. bolus of insulin 10 u. every 2 h is a simple and effective method to control blood glucose concentrations; the method can be used when an insulin infusion pump is not available.