In meat animals there is some interest in lean growth rate or lean food conversion efficiency as selection criteria. These traits may be estimated as the product of growth rate (or efficiency), killing-out proportion and carcass lean proportion. When used as selection criteria these product traits do not require estimates of genetic parameters or economic values for component traits. Hence, they may be more stable, and of longer-term use than classical economic selection indices. The purpose of this study is to compare expected responses to selection on product traits with expected responses from selection, either on individual component traits, or on an economic selection index. Formulae were derived for predicting the phenotypic and genetic correlations between a product and one component, and for predicting the heritability of the product trait; these depend on the coefficients of variation and heritabilities of components, and on the genetic correlations among them. When the coefficient of variation of one component is much higher than that of the other, (× 3), this component will tend to dominate the product trait. In cattle and sheep, and to a lesser extent in pigs, killing-out proportion and leanness are usually less variable than growth rate or efficiency. Thus, in many cases, there is little loss in response in lean growth (or lean efficiency) from selection solely on growth rate (or efficiency) regardless of leanness. Although product traits do not require derivation of economic values, their component traits do have implied economic values. Often more appropriate weightings will be given to components by using an economic selection index, even when economic values and genetic parameters are not estimated precisely.