Nuclear and concatenated phylogenies.
Topologies resulting from the Fgb-I7 and IRBP nuclear genes were inconsistent with each other and with the cytb tree, showing low support values for the ML and BI reconstruction methods. Neither of the nuclear trees recovered the Aporodon species groups for the ML optimality criteria, and most clades collapsed into polytomies (Appendix 3 and 4). In the Fgb-I7 tree, individuals from R. mexicanus clade IIIB were grouped within the R. mexicanus clade I, with generally low support values (pP ≤ 0.50, UFB ≤ 50). Specimens of the R. mexicanus clade II did not form a monophyletic group and their relationships with the species with which they were grouped were mostly poorly supported (pP ≤ 0.80, UFB ≤ 85). The IRBP tree showed minor inconsistencies with the cytb tree, and relationships betweenAporodon species were, in general, better supported, compared to the Fgb-I7 tree, at least for the BI analysis (Appendix 4).
Tree topologies obtained with the cytb + Fgb-I7 concatenated data were almost identical to the cytb topology (Fig. 4). The only difference was in the ML analysis, where specimens of R. gracilis from El Salvador were grouped as the sister clade of R. mexicanus clade I, but with low support values (UFB = 43, Appendix 5). Phylogenies obtained with the cytb + IRBP concatenated data were generally congruent with the cytb tree (Fig. 4). Minor differences were found with respect to the tree resulting from the ML analysis, where R. garichensis+ R. sp. Volcan Poas were recovered as a sister clade to R. microdon , although these relationships were not well supported (UFB = 44, Appendix 6).