Abstract
Results obtained by Young for the expectation of genetic gain in an arbitrary linear function of several traits under selection by independent culling levels, under tandem selection, and under index selection have been obtained in slightly more general form and their dependence on basic genetic and phenotypic parameters exhibited. A warning is given about the effects of selection in modifying the distribution of traits; when the distribution has become appreciably non-normal, any calculation of genetic gains from formulae based on normality will tend to overestimation.