The masses of the radioactive nuclei 22 Mg and 22 Na have been measured with the Canadian Penning trap on-line mass spectrometer to a precision of 3 ϫ 10 −8 and 1 ϫ 10 −8 , respectively. A Q EC value of 4124.39͑73͒ keV for the superallowed  decay of 22 Mg is obtained from the difference of these two masses. With this precise Q value, the Ft value for this decay is determined with improved precision and is found to be consistent with the existing precision data set of superallowed Fermi emitters. This provides an important test of the isospin symmetry-breaking corrections. If the mass of 22 Mg determined here is used in the calibration of a recent 21 Na͑p , ␥͒ 22 Mg measurement, part of the discrepancy observed in that measurement is removed. Superallowed 0 + → 0 + decays play a key role in a number of tests of the electroweak theory. Angular momentum conservation forces these decays to be of a pure vector (Fermi decay) character. The fact that they occur between isobaric analog states, members of the same T = 1 isospin multiplet, means that the nuclear transition matrix element is almost independent of nuclear structure: with perfect isospin symmetry, it would simply be given by a Clebsh-Gordan coefficient. In reality, it exhibits some effects of isospin symmetry breaking but only at the percent level, the same level at which the radiative corrections contribute to the transition. The ft value for such a decay, obtained from measurements of the lifetime, Q value, and branching ratio, can be corrected for these small effects to yield an Ft value, which has the simple form [1]with f being the statistical rate function, t the partial half-life, G V the weak vector coupling constant, and K is a numerical constant. The small correction terms include ␦ C , the isospin symmetry-breaking correction, ␦ R Ј and ␦ NS , the transitiondependent parts of the radiative correction, and ⌬ R V the transition-independent part. Only ␦ C and ␦ NS depend on nuclear structure. The conserved vector current (CVC) hypothesis states that the vector current interaction is not modified by the presence of the strongly interacting nuclear system so that exactly the same value of G V should be obtained from each superallowed decay. The existing set of nine precisely determined transitions confirms this CVC expectation at the 3 ϫ 10 −4 level [1]. The resultant average value of G V , together with the Fermi coupling constant from the pure-leptonic decay of the muon, yields the most precise available value for V ud , the up-down quark mixing element of the CabibboKobayashi-Maskawa (CKM) matrix. This matrix is a rotation matrix connecting the weak and mass eigenstates of the quarks and, as such, must be unitary within the standard model. Currently, the most demanding test of CKM unitarity [1,2] comes from the top-row sum,
V ud2 + V us 2 + V ub 2 = 0.9968 ± 0.0014, ͑2͒which actually violates unitarity at the 2.3 level. The main contributors to that unitarity test are V ud and V us , which contribute almost equally to the total uncertainty. (Alth...