How Can Two Trees Look So Different?

A recent paper by Tsuji, Müller and Reisz (2012) reexamined the nycteroleterid, Emeroleter, a long-legged sister to Nyctiphruretus and Macroleter. Their traditional tree (Figure 1) nested Diadectidae and Limnoscelidae outside the Amniota, among many other odd nestings that, unfortunately, recover “sisters” that too often don’t resemble one another. Such odd nestings may be due to misinterpretations, to too many missing taxa between the oddballs or to too few characters employed (136 vs. 228). A contributing factor has to be the lack of an understanding of the diphyletic nature of the complete reptile family tree. The lack of strong support between the clades in the Tsuji, Müller and Reisz tree(Fig. 1) is telling. There are far too few taxa here for the gamut of included taxa.

Sister Problems
In the Tsuji, Müller and Reisz (2012) tree Diadectes nests in its traditional position, outside the Amniota, a whopping 16 nodes away from its sister Procolophon in the large tree. In the Tsuji, Müller and Reisz (2012) tree mesosaurids nest as sisters to Captorhinidae and Millerettidae + Eunotosaurus, none of which resemble mesosaurids.

I’d like to see the Tsuji, Müller and Reisz reconstruction of the the Eudibamus skull, which nests in their tree as a sister to Belebey (known only from a skull). No skull reconstruction or closeup of Eudibamus has ever been published other than here. The slender and long-toed post-crania of Eudibamus is much closer to that of other basal diapsids, like Petrolacosaurus, rather than to short-toed and thick-ribbed millerettids and Eunotosaurus, the only sisters of Belebey with known post-crania.

Lanthanosuchus should have nested with Macroleter. To their credit, Diadectes is indeed a “close enough”sister to Limnoscelis. Younginiformes are indeed sisters to Araeoscelida. Millerettidae are indeed sisters to Eunotosaurus. The pareiasaurs are related to one another as are the nycteroleterids.

Tsuji, Muller and Reisz (2012) tree of basal reptiles.

Figure 1. Tsuji, Muller and Reisz (2012) tree of basal reptiles. Only those branches with support over 66 are highlighted with yellow circles. Note how few branches are highlighted.

Results of the Large Heretical Tree, Pruned to Match
By greatly expanding the taxon list more confidence can be placed in the nesting of individual taxa because there are so many more opportunities available for them to nest. Figure 2 presents a greatly reduced version of the large tree, pruned to more closely match the taxon list of Tsuji, Müller and Reisz (2012). I would have thought the two trees would be closer to each other in topology, but they are not. I’m not surprised by the lack of resolution here (Fig. 2) at the bases of the three recovered main clades (in color). There are many other taxa excluded between the clades, as recovered in the large tree. This is what happens to small studies with too large a gamut and too few taxa.

The large tree pruned to match the taxon list of Tsuji, Muller and Reisz (2012).

Figure 2. The large tree pruned to match the taxon list of Tsuji, Muller and Reisz (2012). While the pruned tree reflects the results recovered from the large tree, it does not match the Tsuji, Muller and Reisz tree. One must be very wrong. Here it matters little how much support is present to support the branches because so many taxa have been excluded here from the large tree with complete resolution.

How Do Scientists Test One Tree Against Another?
How can one judge the validity of one tree versus another? The Tsuji, Müller and Reisz (2012) tree was created by professional paleontologists. These PhDs had direct access to the specimens. I did not have any of these advantages. However…

In counterpoint, everyone knows that adding taxa and character traits to a cladistic analysis is like adding greater diameter to a telescope lens or mirror. More taxa and characters = greater resolution. The large study is a magnitude larger than the Tsuji, Müller and Reisz (2012) tree. Larger size provides greater confidence in tree topology because more possibilities are covered rather than excluded.

As blogged several times earlier, sister taxa should look like one another. Take mesosaurs. Sister taxa should have long jaws, a retracted naris, an elongated torso and tail, short paddle-like limbs and a long pedal digit 5. We find such traits in Wumengosaurus, Askeptosaurus, Utatsusaurus and other enaliosaurs. We don’t find those traits in sisters recovered by Tsuji, Müller and Reisz (2012), captorhinids and millerettids.

I’m stymied to see so many tree topology differences in the two similar taxon list trees. I wish the character list of Tsuji, Müller and Reisz (2012) followed the traditional order of: skull, mandible, axial and appendicular characters, but they are all mixed up. Given this disorganization, it will be difficult to wade through the data point by point. When I have more to say on this, I will do so.

As always, I encourage readers to see specimens, make observations and come to your own conclusions. Test. Test. And test again.

Evidence and support in the form of nexus, pdf and jpeg files will be sent to all who request additional data.

References
Tsuji, Müller and Reisz 2012. Anatomy of Emeroleter levis and the Phylogeny of the Nycteroleter Parareptiles. Journal of Vertebrate Paleontology 32 (1): 45-67. doi:10.1080/02724634.2012.626004.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.