Some background notes on inequality indices are presented. Equality is seen not as a distribution in which each cell receives an equal share of a resource but as one in which each share is proportional to a criterion or weighting variable for the cell. Instances are noted where the coefficient of variation has been calculated without full consideration of the weighting variable. This can result in misleading comparisons between distributions. A specially constructed data set illustrates that an unweighted coefficient of variation fails to satisfy the Pigou-Dalton principle of transfers and is therefore invalid as an inequality index.