Delimiting species can be challenging, but is a key step for the critical examination of evolutionary history and for prioritizing conservation efforts. Because systematic relationships are often determined iteratively using tests based on taxonomy, such methods can fail to detect cryptic variation and result in biased conclusions. Conversely, discovery-based approaches provide a powerful way to define operational taxonomic units and test species boundaries. We compare both approaches (taxonomy-based delimitation - TBD and discovery-based delimitation - DBD) within North American jumping mice (Zapodinae) using broad sampling, multilocus analyses, and ecological tests. This group diversified through the dynamic glacial-interglacial periods of the Quaternary and phylogeographic tests reveal 28 lineages that correspond poorly with current taxonomy (4 species, 32 nominal subspecies). However, neither the 4-species or 28-lineage hypotheses are optimal for species-level classification. Rather, information theoretic approaches (Bayes Factors) indicate a 15-species hypothesis is best for characterizing genetic variation in this group, with subsequent iterative pairwise ecological tests failing to confirm four species pairs. Taken together, evolutionary and ecological tests capture divergence among 11 putative species that, if upheld by additional tests, will lead to taxonomic revision and reevaluation of conservation plans.
Keywords: Bayes factors; Cryptic diversity; Napaeozapus; Species distribution model; Species tree; Zapus.
Copyright © 2017 Elsevier Inc. All rights reserved.