| Abstract | Given its title, length, and scope, one might expect The Amphibian Tree of Life to be the single most important contribution to amphibian systematics ever. Unfortunately, despite much data and many pages, it is a disaster. The data must be reanalyzed,and the phylogeny and taxonomy should not be used unless the specific results are confirmed by other analyses. So what went wrong? Frost et al. present a new taxonomy for all amphibia based (more or less) on a phylogenetic analysis of 522 species of amphibians from combined parsimony analysis of two mitochondrial gene regions (12S and 16S), five nuclear genes, and an eclectic morphological data set (mostly larval anuran characters). Their study suffers from several fatal flaws. The first is the design. Although the taxon sampling might be reasonable for a study of family-level relationships, it is inadequate for the generic-level changes that are made (e.g., many changes are made without including the type species of genera). The sampling of characters also has two bizarre omissions. The first is the recombination activating gene (RAG-1), unquestionably the most widely used nuclear gene in amphibian phylogenetics. The second is the characters of adult morphology (e.g., osteology, external morphology), which formed the basis for most of the previous amphibian taxonomy. It seems strange indeed to erect a new taxonomy of amphibians by simply ignoring the evidence that was used to construct the previous taxonomies. Problematic data are analyzed with questionable methods. For example, the authors use equally weighted parsimony, which assumes that all characters are evolving at equal rates, an assumption that is demonstrably false. In many cases, the problematic methods seem to have lead to clearly erroneous results (e.g., they find marsupial frogs to be polyphyletic, despite morphological synapomorphies and strong support for their monophyly in molecular studies using modern phylogenetic methods; see J. J. Wiens, J. W. Fetzner, C. L. Parkinson, and T. W. Reeder. 2005. Systematic Biology 54(5):719-748). Rather than pointing out questionable results, they instead use them as their basis for their new classification (e.g., marsupial frogs are divided into three families). The authors paint the picture that any resistance to their taxonomy must be based on "sociology" not science, and that their major innovation is an amphibian taxonomy based on "evidence." However, the evidence they use is questionable at best. For example, they erect a family "Batrachophrynidae" for the leptodactylid genera Batrachophrynus, Caudiverbera, and Telmatobufo. But only the latter two genera are actually included in the phylogeny. Many changes are not only poorly justified, but also unnecessary. To be fair, some of the phylogenetic results and taxonomic changes will almost certainly prove to be correct. But at this point, how does one know which are right and which are not? Given all of these problems, it seems that the safest bet is to simply ignore this study until someone takes the time to do it right. What a waste.
|
