Phylogenetic relationships and divergence time estimates
In the phylogenetic tree, we observed that the haplotypes were clearly divided into two clades, i.e., clade 1 and clade 2, and the S morph was derived from the ancestral L morph (Figure 2). Clade 1 included only Korean L morph haplotypes, while clade 2 included haplotypes from Korea, Japan, and the USA, as well as both L and S morph haplotypes. The Korean and Japanese S morph haplotypes were monophyletic, implying that the wing transformation event took place only once in the history of the species. The USA haplotype (Hap 36) diverged earlier and was not monophyletic with the S morph haplotypes. Based on molecular dating, the two clades diverged approximately 2.7078 MYA (95% HPD: 0.0053–9.278 MYA), and the divergence of the S morph from the L morph occurred around 0.2 MYA (95% HPD: 0.0003–0.7164 MYA) (Figure 2).
Wing morphology and Wolbachia infection
Unlike all of the L morph individuals, which proved to be infected withWolbachia , Wolbachia infection was polymorphic in the S morph individuals. None of the Korean S morph populations (Hap 7) harbored Wolbachia , but the Japanese S morph populations collected from the mid-northern part of Japan, i.e., Ishikawa and Toyama (Hap 22), were completely infected whereas populations from Tokyo (Hap 25, 26) and Gifu (Hap 24) were free of Wolbachia (Figure 2). The wing development pattern correlated strongly with Wolbachiainfection status in this ant species (n = 142, Pearson χ2 = 100.339, df = 1, P < 0.001). These results also suggest that Wolbachia is not involved in clonal reproduction in the ant species because clonal reproduction occurs in both wing morphs (Ohkawara et al., 2006).