In the β-fibint7 dataset, the number of haplotypes (h),nucleotide diversity index (π, Nei 1987), haplotype diversity (Hd) and the average number of nucleotide differences (k) were used as estimates of the genetic diversity of the total dataset and within the main haplogroupsdetected. Genetic diversity calculations were performed excluding recombinant sequences for theglobal data set and within distinctive haplogroups using DnaSP v4.10.9 (Rozas et al., 2003). We used the software PHIPACK (Bruen et al., 2006) to evaluate the presence of recombination in β-fibint7 sequences using the pairwise homoplasy index. This method has been shown by simulation studies to be less sensitive than other available statistics to falsely infer recombinationwhen levels of recurrent mutation are high (Bruen et al., 2006). We used a permutation test (1000 permutations) to estimate the p-values for the null hypothesis of no recombination.Intraspecific median-joining haplotype networks were constructed for both β-fibint7 and the combined nad4 and control region sequences in Network version 4.2 (Bandelt et al., 1999).