We investigated the membrane of En(a-) human erythrocytes as part of a study of the structure and biochemical function of the surface glycoproteins of the mammalian cell. 2. En(a-) erythrocytes were selected because they have more extensive changes at the cell surface than any other known erythrocyte variant. 3. Our results show that in En(a-) erythrocytes: (a) the major membrane sialoglycoprotein is lacking; (b) the other major membrane-penetrating glycoprotein (band 3) has an altered electrophoretic mobility. 4. The apparent clinical normality of En(a-) cells suggests that the change in band 3 may compensate for the loss of the membrane sialoglycoproteins. It is clear that a viable erythrocyte can exist despite the absence of one of its major surface components.