*2.3. DNA Extraction, Amplification and Sequencing*

Genomic DNA from the muscle tissue was extracted using the Wizard® Genomic DNA Purification kit (Promega, Madison, WI, USA) according to the manufacturer's protocol. Partial sequences of the nuclear gene histone H3 (H3) and the mitochondrial gene cytochrome c oxidase subunit 1 (*cox*1) were PCR amplified as follows: H3 gene fragments (347 bp) were amplified using the primers H3af 5 -ATGGCTCGTACCAAGCAGACVGC-3 and H3ar 5 -ATATCCTTRGGCATRATRGTGAC-3 of Lai et al. [25]. *Cox*1 gene fragments (600 bp) were amplified using the primers LCOI490 (5 -GGTCAACAAATCATAAAGAYATYGG-3 ) and HCOI21908 (5 -TAAACTTCAGGGTGACCAAAR AAYCA-3 ) from Folmer et al. [26]. PCR mix comprised 7.23 μL nuclease-free of H2O, 0.75 μL of MgCl2, 0.66 μL of dNTPs, 2.5 μL of 10× buffer, 0.13 μL of each primer, 0.10 μL of Taq polymerase (Promega) and 1.2 μL of DNA0. After an initial denaturation step at 94 ◦C for 5 min, 40 cycles, comprising 94 ◦C for 1 min, 62 ◦C for 1 min (H3) or 48.4 ◦C for 1 min (*cox*1) and 72 ◦C for 1 min, were carried out, with a final extension step of 5 min at 72 ◦C.

PCR products were visualized on 2%-TAE (Tris-acetate–EDTA) agarose gels. The final products were purified with the Wizard SV gel and PCR clean-up system (both from Promega) and sent to Macrogen, Inc. (Seoul, Korea). The forward and reverse sequences obtained were manually edited using Geneious Prime 2019.2.3 software in order to obtain a consensus sequence for each gene. To confirm the species identity, the consensus sequences were queried against the National Center for Biotechnology Information database (i.e., GenBank) via BLAST. Each gene sequence was submitted to NCBI with the following accession numbers: (H3) MT720697, MT720698, MT720699, MT720700, MT720701, MT720702, MT720703; (cox1) MN852247, MN852248, MN852249, MN852250, MN852251, MN852252, MN852253, MN852254.

Some additional sequences from members of the family Trapeziidae were obtained from GenBank and aligned with the sequences generated herein using Clustal W within Mega-X [27]. It is important to emphasize that, to date, there are no *T. corallina* and *T. formosa* H3 and *cox*1 sequences available and that for *T. bidentata* no published H3 sequences exist. Molecular data sets (H3 and *cox*1) were analyzed with maximum likelihood (ML) and Bayesian information (BI) methods.-ML analyses were performed using Mega-X software with 1000 bootstraps. TN93+G+I (Tamura–Nei model+gamma distributed rate of substitution+estimated proportion of invariant sites) was the best-fitting nucleotide substitution model for *cox*1, with T92+G (Tamura 3-parameter+gamma distributed rate of substitution) the ML best-fit for H3. Bayesian analyses were conducted with MrBayes v.3.1.2 [28] and appropriate DNA substitution models were determined separately for the H3 and *cox*1 datasets. The Bayesian analysis was conducted by computing 10,000 Markov chain Monte Carlo generations with 7 H3 and 8 *cox*1 sequences. Using Mega-X software and *cox1* marker, pairwise genetic distances were calculated among the *T. bidentata* and *T. formosa* sequences from the present work, as well as *T. bidentata* sequences from GenBank. Finally, the species *Kempina mikado* (Crustacea: Stomatopoda) was used as the outgroup.
