Using geometric morphometrics to investigate wing dimorphism in males and females of Hymenoptera– a case study based on the genusTachysphexKohl (Hymenoptera: Sphecidae: Larrinae).

The shape of an organism and variation among related groups lies at the core of understanding evolution and interactions between phenotype, genotype and environmental spaces. Traditional morphometrics, used in the study of shape, are currently complimented by an additional statistical method, geometric morphometrics. This technique is based on the quantification of shape usingx,ycoordinates or landmarks. Important features of this technique are obtaining an exact geometric description of the differences or similarities of a biological structure, as well as the ease in visually determining exactly which landmarks are responsible for differences/similarities. The current study investigates the amount of variation between the forewings of males and females of 24 species in the genusTachyspex(Hymenoptera). Wing morphology and its diversity have been used frequently in taxonomic as well as systematic studies of insects and most groups can be identified based solely on the morphology of one wing. This research also aims to investigate the extent of the morphological similarities/differences between males and females of the same species, in order to make recommendations regarding inclusion of both males and females in a systematic or geometric morphometrics study. Mantel statistics revealed a good correlation between the male and female data set; however, Hotelling'st-test and canonical variates analyses (using the Integrated Morphometric Package) indicated that there were indeed significant differences between the two data sets. Although visual analysis of the relative warp analysis plots of the two sexes indicated significant correlations between sexes of the same species, differences in the position of the clusters (or species) on the plots suggests that sexual dimorphism in wings are present. Results from the thin-plate spline analysis of the mean shape of the male and female data set revealed which landmarks are responsible for the differences. It is concluded that, in an analysis of members of a genus, only one of the sexes be used in the analysis, as small-scale differences may influence the results. [ABSTRACT FROM AUTHOR]