Previously recorded and published photoelectron spectroscopic data for mercury in the gas phase has been reanalyzed. The life-time broadenings have been determined for a large number of core levels. It is then seen that a recent detailed derivation of core-level line-widths based on X-ray emission spectroscopy give life-time widths that are generally too large. The 4d(3/2)4d(5/2)nd Coster-Kronig (CM) transition is also discussed. We find that the additional broadening Of the 4d(3/2) level for mercury metal is indeed due to a CM decay, in contrast to recent claims. In atomic mercury, however, the CM process in energetically forbidden. In spite of this we find that the 4d(3/2) level is broadened also in this case. We propose that this is due to a mixing between the 4d(3/2) hole state and discrete 4d(5/2)nd states.