Rat hindquarter preparations were perfused with a plasma substitute for 20 min at high venous pressure (+10 cm H 2O; HVP rats) or low venous pressure (−4 cm H 2O; LVP rats). Perfusion-fixed specimens for electron microscopy were obtained from the rectus femoris, biceps femoris, tibialis anterior, and gastrocnemius (medial head) muscles. Endothelial thickness and frequency of plasmalemmal vesicles (luminal, abluminal, and free vesicles) were assessed on randomly selected capillaries. Mean endothelial thickness was: HVP, 0.26 μm; LVP, 0.28 μm ( P < 0.02). In both groups the abluminal vesicles outnumbered the luminal ones ( P < 0.01), by 35% (LVP) and by 8% (HVP). This difference in abluminal vesicle excess was significant ( P < 0.01). There were no statistically demonstrable differences between the groups in luminal or free vesicle counts per unit luminal surface area or in total vesicle count per unit cytoplasmic volume. The depletion of abluminal vesicles in HVP rats conformed quantitatively with a concept of vesicle membrane incorporation into the plasmalemma proper at pressure-induced capillary dilatation. Provided that the vesicular turnover rate does not change with pressure, the net effect of the reduction in size of the abluminal vesicle population could be an augmented vesicle-bound transendothelial transport of macromolecules from the capillaries to the interstitium. The study indicates that vesicular transport should be regarded as a balance, influenced by hemodynamics, between two separate transport processes in the respective directions.