Isotope Shift in Mg I

The lines of the Mg I spectrum, lying in the visible and in the photographic infrared regions, have been examined and some of them found to show structure. It is pointed out that the structure should be attributed to isotope shift, which, in the case of the $3s3p{}_{}{}^1{}_{}{}^{}P-3s3d{}_{}{}^1{}_{}{}^{}D$ series, owes its origin to the fact that the $\mathrm{smd}{}_{}{}^1{}_{}{}^{}D$ terms, perturbed by the $p^2{}_{}{}^1{}_{}{}^{}D$ term, have passed on to this term some of their isotope separation, which, in the unperturbed case, would have been the same as that of $3s3p{}_{}{}^1{}_{}{}^{}P$ and thus unobservable. This perturbation is discussed in connection with the observed shifts, and it is found possible to correlate the observed shifts with the calculated displacements of the $3smd{}_{}{}^1{}_{}{}^{}D$ energy levels produced by the $p^2{}_{}{}^1{}_{}{}^{}D$ perturbation, with satisfactory agreement. Since the observed patterns show two components, they cannot be attributed to mass effect, which would give three nearly equally spaced components. Furthermore, the shift is in the wrong direction to be accounted for by the usual assumption that heavier isotopes have larger radii, and, consequently, looser binding of electrons than the lighter isotopes.