Three models of hypertension were induced in Wistar rats: (1) aortic ligature between renal arteries, (2) uninephrectomy and Na-rich diet, (3) uninephrectomy, 0.9 per cent NaCl as drinking fluid and subcutaneous administration of desoxicorticosterone acetate (DOCA). We studied in aortic endothelium during the early (7 to 10 days) and late (40 days) phases of these models: (1) volume, surface, and ratios of tight and gap junctions to lateral membrane surface by means of morphometric analysis using thin-sectioning electron microscopy, and (2) organization of tight and gap junctions by means of freeze fracture electron microscopy. Compared with normotensive rats: (1) volume of the endothelial cell layer was increased in all hypertensive situations with the exception of the late group after aortic ligature; the increase was most marked early after aortic ligature and early and late after DOCA; (2) endothelial surface to volume ratio was markedly decreased early after aortic ligature and early and late after DOCA; the ratio of luminal, abluminal, and lateral to total endothelial surface remained unchanged in all hypertensive situations; (3) in all hypertensive situations the ratio of tight junctions to lateral membrane surface increased, whereas the ratio of gap junctions to lateral membrane surface showed no significant change; (4) in all hypertensive situations lateral endothelial plasma membranes contained, in addition to zonular and macular type tight junctions, many small aggregates of intramembrane particles in the P face and particle-free grooves in the E face; these structures were particularly numerous early after aortic ligature and early and late after DOCA; (5) in all hypertensive situations (most early and late after DOCA and least after sodium-rich diet) gap junctions were irregularly shaped in contrast to those of normotensive animals. Our results suggest that : (1) different types of hypertension produced different degrees of hypertrophy of the endothelial cell layer ; (2) hypertension increases the ratio of tight junctions to lateral endothelial membrane surface; the presence of small aggregates of intramembrane particles (probably representing assembly and/or disassembly of tight junctions) in the lateral endothelial plasma membranes is responsible for this phenomenon; (3) hypertension does not change the ratio of gap junctions to lateral membrane surface but results in abnormal gap junction morphology.