Résumé : Both control and hereditarily diabetic (Goto-Kakizaki) rats were administered twice daily for 7 days with an oral solution of carboxymethylcellulose containing, when required, glibenclamide (1.0 microgram g-1 body wt.) or nateglinide (50.0 micrograms g-1 body wt.). The increase in plasma D-glucose concentration and decrease in insulinogenic index caused by the bleeding and handling of the rats prior to sacrifice was more pronounced in the hyperglycaemic and hyperinsulinemic diabetic rats than in the control animals. Eighteen hours after the last oral loading, a sizeable fall in plasma D-glucose concentration and increase in plasma insulin concentration was only observed in the glibenclamide-treated control rats, indicating a more prolonged biological effect of the hypoglycaemic sulphonylurea, as compared to the meglitinide analog. This coincided with the fact that the insulin content of the islets, their secretory response to a high concentration of D-glucose and their basal biosynthetic activity were more severely affected in glibenclamide than nateglinide-treated animals, especially in the control rats. It is proposed, therefore, that the meglitinide analog, considered as a new insulinotropic tool for the treatment of non-insulin-dependent diabetic subjects, may offer the far-from-negligible advantage of minimising the risk of a sustained decrease in both islet insulin content and glycaemia.