Abstract
There has been a firm consent among traditional arthropod systematists that hexapods share a common most recent ancestor with myriapods. Molecular studies, however, question this view by supporting a closer relationship of hexapods with crustaceans. As there is only one evolutionary history, either molecular or morphological data must have been incorrectly interpreted. Morphology can be misleading when structural similarity results from convergent evolution. Molecular tree estimation can also be misleading when the underlying sequence evolution assumptions violate actual sequence evolution parameters. We therefore reexamined the validity of a previously published arthropod ribosomal DNA phylogeny using a new generation of molecular phylogenetic analysis tools. Attention was directed towards potential tree estimation bias arising from the combination of strong rate heterogeneity across sites in ribosomal sequences and elevated substitution rates of insect and crustacean sequences. Consistency analysis with the Monte Carlo simulation based parametric bootstrap method revealed that substitution rate differences in arthropod ribosomal DNA were large enough to cause tree estimation bias for methods that insufficiently correct for multiple substitutions. Maximum likelihood, however, proved immune against long-branch attraction when rate heterogeneity across sites was taken into account. When we estimated trees with maximum likelihood algorithms that take across-sites rate heterogeneity into account, the hexapod/crustacean sistergroup clade remained very strongly supported corroborating confidence in this clade. As recent analyses of protein coding genes and mitochondrial genome rearrangements also converge on uniting hexapods with crustaceans, a revision of the traditional ideas about the origin and evolution of hexapods seems inevitable.
