Abstract. Developmental integration is the covariation among morphological structures due to connections between the developmental processes that built them. Here we use the methods of geometric morphometrics to study integration in the wing of Drosophila melanogaster. In particular, we focus on the hypothesis that the anterior and posterior wing compartments are separate developmental units that vary independently. We measured both variation among genetically diverse individuals and random differences between body sides of single individuals (fluctuating asymmetry, FA). For both of these sources of variation, the patterns of variation identified by principal component analyses all involved landmarks in both the anterior and posterior compartments simultaneously. Analyses focusing exclusively on the covariation between the anterior and posterior compartments, by the partial least‐squares method, revealed pervasive integration of the two compartments, for both individual variation and FA. These analyses clearly indicate that the anterior and posterior compartments are not separate units of variation, but that the covariation between compartments is sufficient to account for nearly all the variation throughout the entire wing. We conclude that variation among individuals as well as the developmental perturbations responsible for FA generate shape variation primarily through developmental processes that are integrated across both compartments. In contrast, much less of the shape variation in our sample can be attributed to the localized processes that establish the identity of particular wing veins.

The body plan of Drosophila, and presumably that of other insects, develops under the control of anteriorp osterior and dorsal^ventral axes, but no evidence for a left^right axis has yet been found. We used geometric morphometrics to study the wings in three species of £ies: Drosophila melanogaster, Musca domestica and Glossina palpalis gambiensis. In all three species, we found that both size and shape showed subtle, but statistically signi¢cant directional asymmetry. For size, these asymmetries were somewhat inconsistent within and between species, but for shape, highly signi¢cant directional asymmetry was found in all samples examined. These systematic left^right di¡erences imply the existence of a left^right axis that conveys distinct positional identities to the wing imaginal discs on either body side. Hence, the wing discs of Drosophila may be a new model to study the developmental genetics of left^right asymmetry. The asymmetries of shape were similar among species, suggesting that directional asymmetry has been evolutionarily conserved since the three lineages diverged. We discuss the implications of this evolutionary conservatism in conjunction with results from earlier studies that showed a lack of genetic variation for directional asymmetry in Drosophila.