Phylogenetic analyses
Phylogenetic analyses were conducted with the COI fragments amplified with the Jerry and Pat primers—GenBank accession numbers: MW041883-MW0441891. The sequences obtained with the primers dgLCO1490 and dgHCO2198 correspond to the COI fragment traditionally used in DNA barcoding efforts (Waugh, 2007) and are available as references for further comparisons—GenBank accession numbers: MW0441892-MW0441898. Our newly generated DNA sequences were aligned with previously published sequences (Garcia et al., 2019; Mapondera et al., 2012) and with other sequences available in GenBank for the G. scutellatus species complex using the Geneious Alignment tool in Geneious Prime 2020.0.3 (https://www.geneious.com/). This alignment tool is versatile because it can detect, and automatically transform, the direction of the sequences. The final matrix consisted of 152 sequences, including two sequences ofOxyops sp. as outgroups, and it had a length of 420 bp. We used IQ-Tree 1.6.12 (Nguyen et al., 2015) to determine in the same analysis the best-fitting evolutionary model, the best maximum likelihood tree, and three measures of nodal support. With the option ModelFinder (Kalyaanamoorthy et al., 2017), we selected the best-fitting model of codon substitution (MGK+F3X4+G4), because codon substitution models are more realistic than other types of substitution models (i.e., nucleotide, amino acid) in protein-coding sequences (Gil et al., 2013; Seo & Kishino, 2009). Nodal support was estimated with the SH-like approximate likelihood ratio test (SH-aLRT), using 1000 bootstrap replicates (Guindon et al., 2010), the aBayes test, which is a Bayesian-like transformation of aLRT (Anisimova et al., 2011), and the ultrafast bootstrap approximation, using 1000 replicates (Hoang et al., 2018; Minh et al., 2013). For the six Eucalyptus snout beetle species with most samples (n ≥ 10), we calculated haplotype diversity (H) and nucleotide diversity (π) with the functions hap.div and nuc.div from the package pegas in R v3.6.3 (Paradis, 2010; R Core Team, 2022). H is the probability that two randomly drawn DNA sequences from the sample would be different; thus, the values of H range from 0 to 1, with 0 indicating that all sequences are identical and 1 corresponding to a highly diverse sample because all sequences are different. π is the average number of differences per nucleotide site, among pairwise comparisons of DNA sequences of the sample; thus, larger values indicate greater nucleotide diversity (Nei, 1987).