Four connexin (Cx) molecules, namely Cx37, Cx40, Cx43 and Cx45, are expressed in the gap junctions that exist within and between the cellular layers of arteries. 2. Endothelial cells are well coupled by large gap junctions expressing Cx37, Cx40 and, to a lesser extent, Cx43, whose expression may be more subject to regulation by physical factors. 3. Smooth muscle cells are more heterogeneously coupled by gap junctions that are small and rare. The identity of the Cx expressed in the media may vary among different arteries. 4. Myoendothelial gap junctions are small and more common in resistance arteries with fewer layers of smooth muscle cells. Given the small size of these gap junctions and the rapid turnover rate of Cxs, homocellular coupling in the media and heterocellular coupling between the cell layers may be subject to more dynamic control than coupling in the endothelium. 5. Vascular gap junctions have been implicated in a number of vasomotor responses that may regulate vascular tone and blood pressure. These include the mechanism of action of the vasodilator, endothelium-derived hyperpolarizing factor (EDHF), the myogenic constriction to intramural pressure increase, the spontaneous or agonist-induced vasomotion of arteries and arterioles and the spreading vasodilation and constriction observed in microcirculatory networks. 6. Few data are available on Cx expression in the media of resistance arteries during hypertension. Changes in the expression of Cx43 described in the media of the aorta of hypertensive rats vary with the hypertensive model studied and are likely to represent adaptations to structural changes in the vascular wall. 7. In contrast, in the endothelium of the caudal and mesenteric arteries of spontaneously hypertensive rats, expression of Cxs is significantly decreased compared with arteries from normotensive rats and this decrease is reversed by inhibitors of the renin–angiotensin system. 8. During hypertension, the activity of EDHF is decreased in the mesenteric artery, but this occurs much later than the initial increase in blood pressure and the decrease in endothelial Cxs, suggesting that changes in EDHF may not be causally related to hypertension or to the changes in endothelial Cxs. 9. Upregulation of the myogenic response and the incidence of vasomotion has been reported in hypertension. Little is currently known of the effects of hypertension on spreading vasomotor responses. 10. Deletion of specific Cxs in genetically modified mice is complicated by neonatal lethality or coordinate regulation and compensatory changes in the remaining Cxs. Nevertheless, mice in which Cx40 has been deleted are hypertensive and spreading vasodilatory responses are significantly impaired. 11. Determination of a role for specific Cxs in the control of blood pressure must await the development of animals in which Cx expression can be modulated in a more complex temporal and tissue-specific manner.