Merodon aureus group

Linear morphometrics

ANOVA did not find significant differences between the sexes (ANOVA: F 1,75 = 2.01; P> 0.05). However, highly significant differences between species were found (ANOVA: F 1,75 = 38.09; P <0.01). Merodon ortus Šašić Zorić, Ačanski & Vujić, 2020 had a larger angle at the intersection of the R4+5 vein and the middle line, indicating that M. ortus has a less sinuate R4+5 vein than M. aureus Fabricius, 1805 and M. calidus Vujić, Ačanski & Šašić, 2020 (Fig. 3).

Geometric morphometrics analysis of R4+5 vein shape

Due to sexual dimorphism, analyses were performed separately on males and females (ANOVA: F 16,102 = 4.596; P <0.01).

Regarding male specimens, DA showed that all species pairs differed highly significantly in R4+5 vein shape (P <0.01). Importantly, 89.16% of the male specimens were correctly classified into a priori defined groups. CVA produced two highly significant axes (CV1: Wilks’ Lambda = 0.173; χ 2 = 116.623; P <0.01; CV2: Wilks’ Lambda = 0.474; χ 2 = 49.603; P <0.01). CV1, with 61% of the total shape variation, separated M. aureus from M. calidus, whereas CV2, with 39% of the total shape variation, separated M. ortus from M. aureus and M. calidus (Fig. 4A). The superimposed outline drawings showed that M. aureus and M. calidus had the most similar R4+5 vein shape (Fig. 4D). Merodon aureus had the most sinuate R4+5 vein, whereas M. ortus had the least sinuate R4+5 vein (Fig. 4D).

DA revealed that female specimens of M. aureus and M. calidus highly significantly differed in R4+5 vein shape (p <0.01). All female specimens were correctly classified into a priori defined groups based on the R4+5 vein shape. Additionally, CVA produced one highly significant axis (CV1: Wilks’ Lambda = 0.101; χ 2 = 71.195; P <0.01). As in males, M. aureus had more sinuate R4+5 vein than M. calidus (Fig. 4D).

Combined geometric morphometrics analysis of R4+5 vein shape and wing shape

All male specimens were correctly classified into a priori defined groups. Additionally, all species pairs differed highly significantly using both semilandmarks and landmarks (DA: P <0.01). CVA produced two highly significant exes, which clearly separated all species in the space defined by these two axes (CV1: Wilks’ Lambda = 0.0085; χ 2 = 264.718; P <0.01; CV2: Wilks’ Lambda = 0.148; χ 2 = 106.119; P <0.01) (Fig. 4C). Regarding females, M. aureus and M. calidus highly significantly differed in the combined semilandmark and landmark analysis (DA: p <0.01), and all female specimens were correctly classified into a priori defined groups. CVA produced one highly significant axis (CV1: Wilks’ Lambda = 0.011; χ 2 = 102.98; P <0.01).