Placing the mountain goat: a total evidence approach to testing alternative hypotheses

The interpretation of a group's evolutionary history can be altered based on the phylogenetic placement of problematic taxa. Mountain goats (Oreamnos americanus) epitomize a 'rogue taxon' as many placements within the Caprini tribe have been suggested. Using a total evidence approach, we reconstructed the Caprini phylogeny using parsimony, likelihood, and Bayesian methods. Bayesian and likelihood methods placed mountain goats as an independent lineage sister to all Caprini except muskox and goral. Maximum parsimony placed mountain goats in a derived Caprini clade. Closer examination revealed that parsimony analysis failed to integrate over phylogenetic uncertainty. We then tested our mountain goat placement against nine published alternatives using non-parametric tests, and the parametric SOWH test. Non-parametric tests returned ambiguous results, but the SOWH test rejected all alternative hypotheses. Our study represents the first explicit testing of all hypotheses for the placement of mountain goats and supports a relatively basal position for the taxon.