The zonal dome model has previously been shown to fit the shape of the poles of the Gram-positive Streptococcus faecium quite accurately, but measurements of the angle that the pole of Bacillus subtilis makes with the cylinder portion of the cell show that the poles of this organism do not fit this model. Furthermore, the diameters of curvatures in different parts of the completed and nascent pole are contrary to the predictions of the zonal dome model and several other proposed models. The results suggest that the poles of Gram-positive rods are generated by a mechanism designated the ‘split-and-stretch’ model. The model assumes that no additional murein is inserted as the septal wall is split and externalized.