The H+-ATPase (F1F0) of Escherichia coli was purified from cells labeled with either [35S]sulfate or [U-14C-D] glucose, and the molar ratio of subunits in the complex determined. The molar ratio was calculated from the radioactivity incorporated into each subunit, using either the subunit sulfur content or subunit molecular weight. These labeling experiments confirm an alpha 3 beta 3 gamma 1 delta 1 epsilon 1 ratio of subunits in F1, and indicate a chi 1 psi 2 omega 10 ratio of subunits in F0. The chi, psi, and omega designations used here refer to the subunits of F0 in order of decreasing molecular weight. Staining with Coomassie brilliant blue gave a reliable indication of the molar ratio of subunits in F1, but very erroneous values for each of the subunits of F0. We attempted to estimate the ratio of subunits in the native membrane, since the stoichiometry determined for the purified complex could be an anomaly of purification. These estimates were made after labeling cells with [35S]sulfate during amplification of the ATPase genes carried on a lambda transducing phage. The subunit ratios in the native membrane were reasonably close to those obtained with purified F1F0. We conclude that the stoichiometry determined reflects the composition of F1F0 in the native membrane. The most surprising conclusion from this study is that there are 10 +/- 1 omega ("proteolipid") subunits in each F1F0 complex. This is considerably more than had been assumed previously.