1. Imidazoline alpha 2-antagonist drugs such as efaroxan have been shown to increase the insulin secretory response to sulphonylureas from rat pancreatic B-cells. We have investigated whether this reflects binding to an islet imidazoline receptor or whether alpha 2-adrenoceptor antagonism is involved. 2. Administration of (+/-)-efaroxan or glibenclamide to Wistar rats was associated with a transient increase in plasma insulin. When both drugs were administered together, the resultant increase in insulin levels was much greater than that obtained with either drug alone. 3. Use of the resolved enantiomers of efaroxan revealed that the ability of the compound to enhance the insulin secretory response to glibenclamide resided only in the alpha 2-selective-(+)-enantiomer; the imidazoline receptor-selective-(-)-enantiomer was ineffective. 4. In vitro, (+)-efaroxan increased the insulin secretory response to glibenclamide in rat freshly isolated and cultured islets of Langerhans, whereas (-)-efaroxan was inactive. By contrast, (+)-efaroxan did not potentiate glucose-induced insulin secretion but (-)-efaroxan induced a marked increase in insulin secretion from islets incubated in the presence of 6 mM glucose. 5. Incubation of rat islets under conditions designed to minimize the extent of alpha 2-adrenoceptor signalling (by receptor blockade with phenoxybenzamine; receptor down-regulation or treatment with pertussis toxin) abolished the capacity of (+)- and (+/-)-efaroxan to enhance the insulin secretory response to glibenclamide. However, these manoeuvres did not alter the ability of (+/-)-efaroxan to potentiate glucose-induced insulin secretion. 6. The results indicate that the enantiomers of efaroxan exert differential effects on insulin secretion which may result from binding to effector sites having opposite stereoselectivity. Binding of (-)-efaroxan (presumably to imidazoline receptors) results in potentiation of glucose-induced insulin secretion, whereas interaction of (+)-efaroxan with a second site leads to selective enhancement of sulphonylurea-induced insulin release.