Our knowledge of the epithelial permeability of different sections of the colon as well as of the surface and crypt epithelium is patchy and contradictory. Therefore, movement of radiolabelled urea, mannitol and Cr-EDTA between the lumen and the plasma of rats was studied, and expressed as clearance. In experiments studying movement from the lumen to the plasma, only the clearance of urea was significant. In experiments on the movement from plasma to the lumen, all three permeability probes exhibited significant clearance in the proximal colon, while in the distal colon the clearance of Cr-EDTA was not significant and the other clearance values were lower than in the proximal colon. Thus, the two methods are proposed to mainly reflect the permeability of two different parts of the epithelium, i.e. the surface and the crypt epithelium. Furthermore, it is proposed that the rat surface epithelium only allows passage of hydrophilic substances smaller than monosaccharides [radius below 0.35 nm (3.5 A] while the crypt epithelium, particularly in the proximal colon, is a heteroporous membrane of higher permeability containing pores corresponding to radii of >3.5-4.0 nm (35-40 A) and 0.4-0.5 nm (4-5 A). Moreover, the results indicate that in vivo luminal fluid solution has no access to the crypt epithelium, a conclusion strengthened by the observation that Evans-blue-labelled albumin and FITC-dextran 4000 do not seem to reach the crypt lumina.