Models of sexual selection usually assume that variation in the expression of sexual ornaments is determined largely by genetic, rather than environmental, factors. However, empirical support for this assumption comes from studies of species with little parental care, in which the influence of environmental factors may be limited, and from studies of just two species with parental care, in both of which heritability estimates vary hugely between years or populations. In the remaining studies of species with parental care, it is not known whether resemblance in sexual ornamentation between relatives was due to shared genes or shared patterns of care. Here we use cross-fostering experiments in house sparrows, Passer domesticus, to examine the relative roles of these effects. We demonstrate thai, although sons resemble their fathers with respect to sexual ornamentation, this resemblance is mainly due to post-hatching environmental effects rather than shared genes. We also show that sons hatching early in the year have the largest ornaments. These results support models that emphasize the importance of environmental sources of variation, such as direct paternal effects, on the expression of sexual ornaments, and agree with the general observation that sexually selected traits tend to be condition dependent. We urge the incorporation of gene-environment interactions into future models of sexual selection.