The experimental data were taken from
(1) S. Takano, S. Yamamoto, and S. Saito, 1991, J. Chem. Phys. 94, 3355.
Even though fine and hyperfine structure splitting are of the same order of magnitude at low values of N, the conventional coupling scheme has been employed.
Very small "uncertainties" have been tripled if their values were very small and the residuals much larger. One line was omitted. Blended lines usually fit to nearly the quoted "uncertainties" if they were treated as intensity-weighted averages. The "uncertainties" of these blends were adjusted slightly.
The predictions should be reliable up to at least 600 GHz, possibly even throughout.
The dipole moment was taken from an ab initio calculation published by
(3) M. Guichemerre and G. Chambaud, 2000, J. Phys. Chem. A 104, 2105.