The passage of substances across the endothelial barrier may occur—theoretically at least [1]—by two main pathways: a) Through the cells. Particles of colloidal gold [2] or molecules of ferritin [3] injected into the plasma are taken up by the pinocytic vesicles of the endothelial cells, and thereafter appear in the extravascular spaces. This and other forms of trans-cellular passage will not be dealt within this paper, b) Between the cells. Interest in the intercellular junction as a potential pathway is stimulated by physiologic studies, which indicate that traus-endothelial passage of ions and small molecules is a passive process. Dyes injected intra-arterially escape at discrete points along venules and capillaries [4], in a manner which could not be adequately explained by pinocytic transfer [1]. It has been stated that the endothelial cells are connected by “tight junctions” or zonulae occludentes [5, 6]. It has not, however, been shown that these junctions form a continuous band around each cell, and that they do indeed represent a tight seal. The zone of membrane “fusion” could represent, rather than a seal, a very fine filter permeable to water and to very small molecules [7, 8]. By electron microscopy, the passage of tracer particles along junctions which appeared to be still “closed” has been observed [9] : these studies were done in vitro, on rat hearts perfused with a saline solution containing particles of saccharated iron oxide (S.I.O.).