Ancestral range reconstruction of Galliformes: the effects of topology and taxon sampling

Abstract

AbstractAimWe examined divergence times and biogeography of the avian order Galliformes (which, despite a nearly world‐wide distribution, includes many weak fliers), to test whether current distributions reflect vicariance or long‐distance dispersal. We also tested the impact of taxon sampling and tree topology on our estimates of historical biogeography.LocationWorld‐wide.MethodsWe generated timetrees for all major galliform lineages using several fossil calibrations and a combination of mitochondrial and nuclear data. We compared divergence times and reconstructed ancestral ranges using this timetree and the galliforms from a recent synthetic tree for all birds. Thus, we explored the sensitivity of our conclusions to differences in topology and taxon sampling.ResultsThe results suggest late Cretaceous origins for Megapodiidae and possibly for Cracidae, the earliest diverging families. The other families diversified in the Eocene after the break‐up of Gondwana, contrary to previous suggestions. Both topology and taxon sampling affected the ancestral area estimates, although many nodes were consistent among the approaches.Main conclusionsDivergences of Numididae, Odontophoridae and Phasianidae occurred in Africa, with subsequent dispersal to other continents. Reconstructed ranges for ancestors of the earliest diverging families, Megapodiidae and Cracidae, are less conclusive, but may have involved a South American origin, then dispersal to other continents. Thus, long‐distance dispersal is likely to have been important in galliforms, possibly reflecting more vagile ancestors and repeated range expansions and contractions. There appears to be a trade‐off between the use of trees with rich taxon sampling but limited data, where key taxa can be misplaced, and more robust trees with missing taxa. When biogeographically important taxa are included, complete taxon sampling does not appear necessary for accurate ancestral area reconstruction. Robust trees that include taxa targeted based on their distribution may be the best way to improve our understanding of historical biogeography.