A standardized intravenous regimen has been assessed, in 25 insulin-treated diabetic women, for insulin and dextrose therapy in labour and delivery. Adjustments to insulin infusion rate are determined by trends in blood glucose as well as by absolute concentration, in order to approach normoglycaemia. Blood glucose was 5.0 (SD 1.7) mmol l-1 on arrival in labour (or at 0800 h before planned delivery) and was maintained at 6.0 (SD 1.8) mmol l-1 with insulin 0-5 U h-1 for up to 29 h before delivery, when it was 6.3 (SD 2.1, range 3.0-9.0) mmol l-1 with insulin infusion rate 0-4 U h-1. Neonatal blood glucose (less than 2.0 mmol l-1 in 11 babies) correlated with both maternal HbA1c (rs = -0.47, p less than 0.02) and maternal blood glucose at delivery (rs = -0.58, p less than 0.01). During 12 months observation on the intravenous regimen, 339 measurements of blood glucose were made; 10 were less than 3.0 mmol l-1, 242 were 3.0-8.0 mmol l-1, and 81 were greater than 8.0 mmol l-1 (mean 6.5, range 2.7-13.5 mmol l-1). Insulin infusion rate ranged from 0 to 5 U h-1, with 139 rate adjustments. Only one mild clinical hypoglycaemic episode, responding to increased dextrose infusion, was recorded. This simple flexible regimen proved clinically reliable for both midwifery and medical staff.