The technique of Mitchison and Swann (1954) was modified for determining the resistance to deformation, or "stiffness," of the red cell membrane and the pressure gradient across the cell wall. It requires a measure of the pressure needed to suck a portion of the cell into a micropipette. Stiffness of hypertonically crenated cells was less than that of biconcave discs or hypotonically swollen cells. Crenated cells showed zero pressure gradient and a stiffness, probably due to pure bending, equivalent to 0.007 +/- 0.001 (SE) dynes/cm. Normal and swollen cells showed a pressure gradient of 2.3 +/- 0.8 (SE) mm H(2)O and a stiffness, due to bending and tension in the membrane, equivalent to 0.019 +/- 0.002 (SE) dynes/cm. No difference in stiffness was found between the rim and the biconcavity of the cell or between biconcave discs and hypotonically swollen cells. Micromanipulation showed that the membrane can withstand large bending strains but limited tangential strains (stretching). These results have significant implications in any theory explaining the cell shape. For example, the data give no indication that the physical properties of the membrane are different at the rim from those of the biconcavities, and the existence of a positive pressure in the normal cell is established.