Filipin, a mixture of polyene antibiotics which form complexes with cholesterol, perturbs membrane lipid organization, and causes hemolysis of erythrocytes, is increasingly used as a cytochemical probe for the distribution of cholesterol in cell membranes. We used light (phase-contrast, dark-field and fluorescence) and electron microscopical techniques (whole-mount shadowing, negative staining, and freeze-fracture) to study the interaction of filipin with unfixed and glutaraldehyde-fixed human red blood cell (RBC) membranes. Lysis time and extent depended upon the cholesterol:filipin (C:F) ratio. Lysis was prevented by osmotic protection with high MW dextran. Filipin treated cells fluoresced, but variation in fluorescence intensity among unfixed as well as among fixed cells was evident both at low and high C:F ratios. Negatively stained preparations of unfixed cells lysed on grids or in suspension revealed ring- or C-shaped filipin-induced lesions (FIL) equipped with a veil-like appendage; single FIL, and FIL fused by their veils into aggregates, were shed from membranes. FIL at the surface proper of shadowed whole-mounts and of freeze-etched preparations of prefixed cells appeared as single, dispersed or aggregated cylinders protruding to variable heights above the membrane's plane; aggregated FIL were shed from cells. The freeze-fracture appearance of FIL differed in membranes fixed before or after filipin treatment. E- and P-faces of post-fixed membranes exhibited cylindrical protrusions and depressions, respectively; in essence, the reverse was found in pre-fixed RBC. Both pre- and post-fixed membranes showed considerable variation in the number of FIL on individual cells whether incubated at high (1:1) or low (1:5) C:F ratios, or for a short (10 min) or a long (80-180 min) time. Aggregation and shedding of FIL was evident in all preparations. Thin layer chromatography of the incubation fluid after sedimentation of cells showed that membrane cholesterol was shed from incubated cells. The presented data question the feasibility of filipin as a probe for the topographical distribution of cholesterol in cell membranes.