The electrokinetic behaviour of normal erythrocytes is compared with that of trypsin-, N-bromosuccinimide-, and tosyl-treated erythrocytes. Reduction in the net negative charge with reduction in ionic strength of the suspending medium and also on treatment with N-bromosuccinimide and trypsin is discussed using a porous non-rigid polyanion as a model for the periphery of the cell membrane. It is deduced from the equivalent binding of chloride and thiocyanate ions and the absence of any effect on treatment of red cells with tosyl chloride, that normal, N-bromosuccinimide- and trypsin-treated cells are polyanionic in character. Reduction in erythrocyte charge on treatment with N-bromosuccinimide or trypsin is probably not due to the removal of phosphate groups from the interface, nor to physical adsorption of N-bromosuccinimide or trypsin. The charge reduction is probably produced by bond fission with possibly a net disappearance of carboxyl groups from the electrophoretic plane of shear either by loss from, or reorientation of, the membrane. The loss or reorientation of material associated with these carboxyl groups does not lead to any basic change in the character of the surface of the cell, nor to any obvious structural instability. The biconcave discoid form is maintained and there is no significant hemolysis of the erythrocytes even after contact with a solution of trypsin for 24 hours.