A correction to a paper on Roman <i>k</i>-domination in graphs