In normal fasting dogs, rabbits, rats and mice the maximal hypoglycaemia produced by intravenous administration of CS476 is comparable on a weight basis to that produced by Glibenclamide. Randomized Latin square experiments in dogs show that 0.03 mg/kg orally of CS476 and of Glibenclamide cause the same maximal decrease of blood glucose and that CS476 has the shorter duration of action. CS476 has no hypoglycaemic effect in totally pancreatectomized dogs nor in streptozotocin diabetic dogs and rats. The insulin releasing activity is studied in dogs after intravenous and oral administration of equipotent doses of CS476, Tolbutamide and Glibenclamide.