Allozyme genotyping and analyses
Our allozyme dataset comprised the same 54 putative allozyme loci as employed by Unmack et al. (2019) and was generated according to the principles and procedures presented in Richardson, Baverstock & Adams (1986) and Hammer et al. (2007). We used PCoA, coupled with assessments of diagnosability (fixed differences, allowing a 10% cumulative tolerance for shared alleles at a locus as advocated by Adams et al., 2014 for allozyme markers) and admixture (intermediate positioning between parental taxa for PCoA, higher levels of heterozygosity, and lack of fixed differences at otherwise diagnostic loci), to explore the broader taxonomic and phylogeographic patterns evident in this dataset. The rationale and methods for these analyses follow Adams et al. (2014) and Unmack et al. (2022).