Genetic diversity and population structure
Because the inclusion of close relatives can bias estimates of genetic diversity and structure (Goldberg & Waits, 2010), we removed sequences representing first-order relatives identified by our SNP dataset (methods described below) and performed all subsequent mitogenome analyses with the reduced data, which we refer to as the “unrelated dataset”. We defined haplotypes and calculated haplotype diversity (H d), nucleotide diversity \((\pi\)), and Tajima’s D using DNAsp (Librado & Rozas 2009). We estimated the differentiation between MK and MT and between high and low elevations within each peak through analysis of molecular variance (AMOVA) in Arlequin v3.5 (Excoffier & Lischer, 2010) with a permutation test of 10,000 replicates to assess statistical significance. We visualized relationships among haplotypes by generating a median-joining network with the PopART software (Leigh & Bryant, 2015).