Plasmolysis bays, induced in Escherichia coli by hypertonic treatment, are flanked by zones of adhesion between the plasma membrane and the cell wall. To test the proposition of Cook et al. (W. R. Cook, F. Joseleau-Petit, T. J. MacAlister, and L. I. Rothfield, Proc. Natl. Acad. Sci. USA 84:7144-7148, 1987) that these zones, called periseptal annuli, play a role in determining the division site, we analyzed the positions of these zones by phase-contrast and electron microscopy. In situ treatment of cells grown in agar showed that the youngest cell pole was the most susceptible to plasmolysis, whereas the constriction site was resistant. Lateral bays occurred only at some distance from a polar bay or a resistant constriction site. Orienting cells with their most prominently plasmolyzed polar bay in one direction showed that the lateral bays were always displaced away from the polar bay at about half the distance to the other cell pole. If no poles were plasmolyzed, lateral bays occurred either in the centers of nonconstricting cells or at the 1/4 or 3/4 position of cell length in constricting cells. The asymmetric positions of lateral plasmolysis bays, caused by their abrupt displacement in the presence of polar bays or constriction sites, does not confirm the periseptal annulus model (Cook et al.), which predicts a gradual and symmetric change in the position of lateral bays with increasing cell length. Our analysis indicates that plasmolysis bays have no relation to the development and positioning of the future division site.