3.3 Influence of host plant use and geographic distance in genetic variation
Analyses of the influence of host plants and geographic distance on genetic variation in the H. pungens species complex showed that host plant use was a better predictor of genetic distance than geographic distance (Table 2). Mantel correlation tests between host plant use and genetic distance matrices and between geographic and genetic distance matrices using the nuclear SNPs dataset were significant (r = 0.5254; p = 0.0017 and r= 0.4;p= 0.0128, respectively). In addition, the partial Mantel test indicated that the correlation between host plant distance and genetic distance matrices remained significant when controlling for geographical distance (r = 0.6133; p = 0.0001). Likewise, the results of regressing the matrix of genetic distances on both host plant and geographic distances showed that both factors significantly affected genetic differentiation among sampling localities for nuclear SNPs (host plant coefficient= 0.28497, p= 0.0001; geographic coefficient = 0.00709, p= 0.0001). Host plant and geographic distance together explained 47.6% of genetic variation. In contrast, correlation and multiple regression tests based on mtDNA distance matrix on geographic and host plant distances yielded non-significant results (Table 2).