STUDIES ON NATURAL POPULATIONS OF DROSOPHILA. II. HERITABILITY AND RESPONSE TO SELECTION FOR WING LENGTH IN DROSOPHILA MELANOGASTER AND D. SIMULANS AT DIFFERENT TEMPERATURES

ELECTION under different environmental conditions varying in nutrition temperature poses a problem to be solved on sound genetical grounds. The problem is whether the best response to selection can be achieved when selection is carried out under optimal conditions or when it is carried out in some other environment that is especially favourable for the expression of genes affecting the desired trait. Most selection experiments, on Drosophila, have been carried out under optimal conditions; for recent references see LATTER and ROBERTSON (1962). The existence of genotype-environment interaction may mean that the “best” genotype in one environment is not the best in another environment. Evidence is accumulating that such interaction is common (PARSONS 1959; ROBERTSON 1960a; for more general references see HULL and GOWE 1962). The situation is important in breeding policy and raises two problems: (a) To what extent is the response to selection which is achieved in one environment also expressed in another environment? (b) Does the attenuation in the heritability resulting from genotype-environment interaction interfere seriously with the response to selection? Actually, the problem is much broader than expressed here. Natural populations and also domesticated species inhabit many environments and, therefore, several environments have to be taken into account when selecting (see DICKERSON 1962 for a good discussion of the problem). The present experiments were carried out to examine the response to selection for wing length in Drosophila melanogaster and zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA D. simulans at different temperature conditions. In particular, the effects of temperature on the response to selection as well as on the heritability of wing length were studied by transferring the selected lines from one temperature to another. Observations were taken on other traits to determine their correlated responses, if any.