Selection was carried out for 14 months in populations ofDrosophila melanogaster containing at the outset equal numbers ofe ande+ genes. High temperature and to a less extent low humidity as well as selection for propagation of early flies tended to favour the light wild type, whereas the opposite conditions favoured the dark (ebony) flies. However, only a low temperature brought about an actual increase ofebony. In none of the eight lines was one allele completely eliminated, the only difference being the different level of equilibrium and the speed with which it was approached. This indicates that in all the conditions tested in the laboratory the heterozygotes are superior to both homozygotes, the wild type being fitter than ebony at high temperatures and less fit at lower temperatures.Ebony flies were slower to develop under all conditions. More flies emerged from a bottle at high temperature than at low temperature, but they were lighter in weight. The latter difference, however, did not appear to be inherited.