A note on the analysis of covariance

In recent papers which refer to the use of covariance there has still appeared an occasional lack of understanding of this method of statistical analysis. There have, for example, been references in the literature to methods of analysis which the writer has used in certain problems as if they were in competition with the use of covariance as described by Fisher (1, 2). It would therefore seem to be of value to indicate briefly the different situations in which covariance may prove of use, in order that the distinction between some of these situations may be clearly realised.