A comment is made on a previous paper by the same author. It is further shown that in family selection, the optimum number of families chosen, giving the maximum total change due to selection and inbreeding depression, is independent of the size of the testing programme over a wide range.