The permeability to protein of the vessels of the renal medulla in the rat was studied by intravenous injection of Evans' Blue or fluorescent-labelled serum, followed by perfusion of the whole vascular system with saline. The vessels of the inner medulla and the vasa recta of the vascular bundles of the outer medulla were found to be rapidly and freely permeable to protein. Electron microscopy, after the injection of ferritin, Thorotrast or colloidal gold, showed the vessels probably concerned in the protein leak to be ascending vasa recta and the capillaries of the medulla. The findings support suggestions that extravascular albumen is important in the process of concentration of the urine.THERE is considerable evidence for the existence of a large pool of exchangeable albumen in the kidney, and the results of a number of studies suggest that much of this is extravascular. This has been shown by the many workers who have used radioactive albumen and labelled red cells to investigate the relative distribution of cells and plasma in the kidney Collings and Swann, 1958;Lassen et al., 1958;Goldberg and Lilienfield, 1963;Ochwadt, 1963;Chinard et al., 1964]. More recently, Giirtner et al. [1968] have investigated the problem using 1311 albumen and polyvinylpyrrolidon to follow the fate of large molecules. They found a rapid loss of protein from post-glomerular vessels; 50 per cent of extravascular protein exchanged in 12 minutes -a very rapid rate of exchange compared to that of other organs. Pinter [1967], using isotope-labelled chylomicrons as an indicator of intravascular plasma space, found a large quantity of extravascular albumen in the kidney, equivalent to 31 per cent of the total renal albumen.A more direct method was that of Slotkoff and Lilienfield [1967] who injected 131I albumen and 51Cr red cells into dogs and then perfused the vascular system with dextran to clear the vessels of blood. They found a significantly higher proportion of albumen than of red cells remaining in the kidney, presumably indicating an extravascular albumen pool.Attempts to visualize the extravascular albumen directly have met with conflicting results. Pomerantz et al. [1965], using fluorescin-labelled albumen in rats, found that fluorescence microscopy showed extravascular protein, particularly in the medulla, as little as 40 seconds after the injection. Carone et al. [1963, 1967], on the other hand, found albumen only within the vessels. Wilde and Vorburger [1967], who used Evans' Blue in hamsters, described and illustrated extravascular dye in freeze-dried kidneys embedded in Epon. A number of other authors who have used Evans' Blue to study the renal circulation have noticed staining of the kidneys as an incidental 60