Squamata: Acrodonta

Priscagamids, Amphibolurines, Draconines, & Chameleons

Taxa on This Page

1. Acrodonta
2. Agamidae
3. Amphibolurinae
4. Chamaeleonidae
5. Draconinae
6. Priscagamidae X

In another context, we looked at the molecular phylogenetics work of J. Robert Macey and his coworkers. Cuculiformes. Dr. Macey recently (1999) completed his Ph.D. at Washington University in St. Louis, Missouri where he studied with Profs. Allan Larson and Jonathan Losos. Since we are presently on the subject of lizard phylogenies, this seemed an auspicious time to look in a bit more detail at what Macey has been doing.

In the prior discussion, we discussed Macey's work tracing phylogeny through changes in gene order in mitochondrial DNA. Macey et al (2000). Here, we examine the companion paper, in which Macey and a large group of coworkers developed trees based more conventional DNA sequencing work. Macey et al 2000a). This paper turns out the way molecular phylogenetics ought to turn out, but seldom does. The results give us a very clear picture -- not only of the phylogeny, but of the biogeography of these lizards. In truth, one suspects Macey's methods stack the deck a little, but only a little, because geography is explicitly considered in the development of the tree structure. However the correspondence of geography to genetics seems to have been statistically tested with some rigor. Surely this is no worse than the usual cladistic methods which combine genetic and morphometric information; and, in fact, the use of geographical information is just another sort of grist for the cladistic mill.

The resulting cladogram (simplified to the generic level) is as follows.

What Macey's work suggests is that each of the various taxa of agamid lizards crossed the Tethys on different continental plates, like so many groups of immigrants arriving on crowded steamers. Each of the seven major acrodont groups can be identified with a particular fragment, or, at worst, series of fragments of the Gondwanan plate which drifted across the Tethys Sea to contact Laurasia now Eurasia) over the last 120 My or so.

Unfortunately, when we compare the cladogram to Figure 1, it is immediately clear that there are problems. For example, the Chamaeleonidae and Agaminae are, respectively, the first and last branches from the acrodont stem. Yet they appear together on the same continental block. One very basal species of the amphibolurine (Physignathus cocinius), which presumably developed in Australia, is found in Southeast Asia. These are not insurmountable concerns, but they do suggest that the picture may not be quite as straightforward as it first appears. It would be very interesting, for example, to look at the evolution of the chameleons in a bit more detail.

Despite minor rough spots like this, the paper is a rich and elegant piece of work, and only the broad conclusions are summarized here. However, it does leave us with a real difficulty. Here is, undoubtedly, a piece of mtDNA phylogeny that worked. Why didn't it work for birds? See Mindell et al. (1997). Macey and coworkers seem to have taken a much more cautious approach to aligning the mitochondrial tRNA gene sequences which make up a large portion of the mtDNA segment used for the study. Ultimately, Macey discarded much of the tDNA sequence as "phylogenetically uninformative," primarily because it was too variable to be useful. This may, according to the views of the reader, be considered either a good use of judgment or hopelessly biased data selection.

For what it may be worth, I am inclined to the former opinion for reasons I have explained in the other essay. Proteins and tRNAs do not function as linear information. They function because of their tertiary and quaternary structure. This means that the DNA sequences coding for these structures are inter-related in ways we absolutely cannot predict from primary sequence data alone. The best we can do is to use judgment to find portions of these genes that seem to behave as if they were independent variables over the time spans of interest. The process is, even at its best, no more objective than cladistic analysis from morphological characters. That does not mean that it is worthless. It simply means that, like a lot of science, absolute and mechanical objectivity is an asymptote we must always approach without quite reaching. The greatest danger is not failing to reach it, but in believing that we already have.

Descriptions