Selection for a threshold character in Drosophila I. An analysis of the phenotypic variance on the underlying scale

Abstract

The short-term response to artificial selection in a wild-type population for increased scutellar bristle number, has been interpreted in terms of two alternative threshold models. Both models have been shown to give a satisfactory transformation of the scale of measurement, in that the underlying variable shows an effectively linear response in terms of the accumulated selection differential, and the pheno-typic variance on the underlying scale remains virtually unchanged over a period of generations sufficient to increase the incidence of extra-bristles to almost 100%.A genetic analysis based on the bivariate model, which to a large extent takes account of the location of the supernumerary bristles, has shown the phenotypic variance on the underlying scale to have the following composition: the additive genetic variance is 27·2 ± 0·9% of the total, and the developmental error or ‘chance’ variance accounts for a further 63·8 ± 2·6%, so that the remaining 9·0 ± 2·7% of the variance must be due to environmental differences among individuals and to non-additive genetic effects.The relative magnitudes of these parameters are remarkably similar to those observed for two quasi-continuous variables in the same population, viz. abdominal and sternopleural hair number. In view of the ease with which scutellar bristles may be scored, the character can therefore serve as an extremely useful model character for quantitative genetic experiments.