Genetic diversity and variation

For the plastid rbc spacer, number of haplotypes (N _h), haplotype (h) and nucleotide (π) diversity were estimated for each locality using ARLEQUIN 3.5 [39]. This program was also applied to calculate pairwise values of genetic differentiation (F _ST). All results for significance of covariance components were tested using 10⁵ permutations. A median-joining network among plastid haplotypes was constructed using NETWORK 4.5 [40].

For microsatellite loci, allelic richness (N _e), private allelic richness (A _p), inbreeding coefficients (F _IS), mean expected (H _e) and observed (H _o) heterozygosity were calculated in GENETIX 4.05 [41]. Hardy-Weinberg equilibrium (HWE) and linkage disequilibrium (LD) tests were assessed using GENEPOP 4.3 [42]. The significance level for multiple simultaneous comparisons was adjusted using the sequential Bonferroni technique [43]. Pairwise F _ST and Jost’s D (D _est, [44]) were performed using the program SMOGD [45]. The overall D _est for each population comparison was calculated as the arithmetic mean across loci. The hierarchical analysis of molecular variance (AMOVA) was conducted to detect the proportion of genetic differentiation among regions, and within and among populations [46].