Phylogenetic, Blanchard, 1845

Phylogenetic View in CoL and population genetic analyses

The concatenated dataset used for the phylogenetic reconstruction contained 657 bp of COI, 583 bp of 28S ribosomal RNA, and 435 bp of Wingless, for a total of 1675 bp of nucleotide sequence. The Bayesian inference and ML inference trees showed the same topology, with strongly supported major nodes ( Fig. 2 View Figure 2 ). We considered both ML bootstrap values (UFB) and Bayesian posterior probability from BEAST of 95%–100% as strong, 80%–95% as moderate, and <80% as low support. Trictenotomidae were recovered as monophyletic with strong support [100/100 (bootstrap support 100%, posterior probability 100%)]. The two genera were recovered as monophyletic with moderate to high supporting values ( Autocrates , 85/100; Trictenotoma , 88/100). All the species were recovered as mono-clade with strong support values (100/100). Most of the outgroup taxa have a messily clustered topology (Supporting Information, Fig. S1 View Figure 1 ), largely different from the most up-to-date and credible phylogenies ( Zhang et al. 2018, McKenna et al. 2019) and with most nodes poorly supported.

The analysis of haplotype networks in A. maqueti revealed a relatively high level of diversity in their COI haplotypes, with eight unique haplotypes identified among the 12 specimens studied ( Fig. 2 View Figure 2 ). The subpopulations of A. maqueti from China and Korea were found to be genetically different, with a range of 8–16 genetic mutational steps between them. Additionally , the Korean specimens were closely related to each other, with only one or two mutational steps separating them, whereas the Chinese subpopulations exhibited greater genetic distance, which is consistent with their disjunct distribution and the large geographical distance between collection sites.

Molecular species delimitation

Results of species delimitation from the four approaches are summarized in Figure 2 View Figure 2 . The number of molecularly delimited species of all four approaches was the same. All four approaches revealed nine molecular clades of Trictenotomidae , and they fitted perfectly with current morphospecies. Two most recently described species, A. ivanovi and A. lini , were confirmed as valid using molecular techniques. Autocrates maqueti was demonstrated to be autochthonous in the Korean Peninsula.

Molecular dating and phylogeography

The age estimates for the family Trictenotomidae suggest that it originated in the late Palaeogene ~60.3 Mya (with 95% HPD confidence intervals of 45.02–77.45 Mya), as estimated by the secondary calibration point of Boridae + Trictenotomidae . The divergence time of the genus Autocrates was estimated to be in the mid-Eocene, ~49.7 Mya (with 95% HPD of 36.3–64.6 Mya), and Trictenotoma in the late Oligocene, ~34.1 Mya (with 95% HPD of 22.7–46.2 Mya). The most recent common ancestor of the Korean and Chinese A. maqueti subpopulations is inferred to have existed ~1.79 Mya (with 95% HPD of 0.75–2.96 Mya) during the Pleistocene ( Fig. 3 View Figure 3 ).

The DIVALIKE model was determined as the best-fitting model with the lowest AIC score (Supporting Information, Table S5). Our biogeographical results suggest that Trictenotomidae had originated from a wide geographical range including India, South Asia, and mainland East Asia. The genus Autocrates was estimated to have originated in Indochina through vicariance, followed by a few dispersals, range expansion, and vicariance events that might have led to speciation. Autocrates maqueti was derived from the most recent common ancestor of itself and A. oberthueri from mainland China and went through range expansion to the Korean Peninsula. The divergence between the Korean and Chinese A. maqueti subpopulations can be explained by a vicariance event that occurred during the Pleistocene, resulting in their distinct, disjunct distributions.