In a previous study, we showed that exposure of rats to dexamethasone (Dex) selectively in late pregnancy produces permanent induction of hepatic phosphoenolpyruvate carboxykinase (PEPCK) expression and hyperglycaemia in the adult offspring. The mechanisms by which glucocorticoids cause this programming are unclear but may involve direct actions on the fetus/neonate, or glucocorticoids may act indirectly by affecting maternal postnatal nursing behaviour. Using a cross-fostering paradigm, the present data demonstrate that switching the offspring at birth from Dex-treated dams to control dams does not prevent induction of PEPCK or hyperglycaemia. Similarly, offspring born to control dams but reared by Dex-treated dams from birth maintain normal glycaemic control. During the neonatal period, injection of saline per se was sufficient to cause exaggeration in adult offspring responses to an oral glucose load, with no additional effect from Dex. However, postnatal treatment with either saline or Dex did not alter hepatic PEPCK activity. Prenatal Dex permanently raised basal plasma corticosterone levels, but under stress conditions there were no differences in circulating corticosterone levels. Likewise, Dex-exposed rats had similar plasma catecholamine concentrations to control animals. These findings show that glucocorticoids programme hyperglycaemia through mechanisms that operate on the fetus or directly on the neonate, rather than via effects that alter maternal postnatal behaviour during the suckling period. The hyperglycaemic response does not appear to result from abnormal sympathoadrenal activity or hypothalamic-pituitary-adrenal response during stress.