Realized evolvability: quantifying phenotypic evolution in a Drosophila clade

Understanding the evolutionary potential of morphological evolution is still a major problem in evolutionary biology. In this study, we tried to quantify the amount of variation of different traits among species of a Drosophila clade reared under standard conditions. Nineteen different traits have been measured on nine species of the same clade, the Neotropical saltans group of Drosophila. Measured traits can be distributed into five categories: size‐traits (wing and thorax), shape indices (ratios among the size traits), number of sternopleural bristles on the thorax, number of abdominal bristles on successive sternites, and dorsal pigmentation of abdomen. All species are of medium size with a generally dark pigmentation. A remarkable feature is the presence of numerous bristles on the 6th sternite of the males, while this segment is bare in other Drosophila species. A multivariate analysis revealed that it was possible to discriminate all the investigated species by using the complete set of measured traits. For each trait, phenotypic variability was investigated at the within‐ and between‐species levels. As a rule, the interspecific coefficient of variation (CV) was much greater than the within species CV, and it is proposed to call it realized evolvability. All possible correlations were calculated between traits across species, providing many unexpected results. Size traits were highly correlated among them, but not correlated with shape indices. Abdominal traits (bristles and pigmentation) were correlated, but often in opposite directions, with thoracic shape indices. Tergite pigmentation was negatively correlated with bristle number on sternite. For the moment, most of the correlations cannot be explained by developmental processes or parallel selective pressures. Nonetheless, mapping the evolution of the two characters on a molecular phylogeny of the studied species revealed two opposite phylogenetic trends for abdominal pigmentation and setation, respectively. Our data suggest a need for similar studies in other well‐defined Drosophila clades.