A shift in the topology of the large reptile tree

As loyal readers know…

  1. I have challenged others to find taxa that are improperly nested.
  2. So far this year no one has stepped up to the challenge. Well, they had their chance…
  3. Yesterday I discovered a mistake in the large reptile tree and made the correction (Fig. 1). The diadectomorph clade (including procolophonids) have moved closer to the bolosaurids and pareiasaurs.
Figure 1. A shift in the tree topology moves diadectids (and procolophonids) closer to pareiasaurs and away from limnoscelids. But Orobates stayed back.

Figure 1. A shift in the tree topology moves diadectids (and procolophonids) closer to pareiasaurs and away from limnoscelids. But Orobates stayed back. At the bottom of the chart, that line leads to macroleterids and nycteroleterids, then on to the lepidosauriformes and lepidosaurs.

The problem was
so much of the data for the taxa around these nodes are represented by drawings, sometimes with errors. The most difficult taxa and perhaps the least interesting of all reptiles (judging by the number of papers written about them) are also at this node: Saurorictus, Milleretta (both specimens) and the bolosaurids (sans Eudibamus, which nests with Petrolacosaurus). While a single tree was found then, and is also found now, the bootstrap scores were not strong, now or then. The present bootstrap scores could use a boost owing to the skull-only and otherwise unfortunately incomplete data in several taxa often represented only by drawings. Many taxa were rescored.

The other problem was
diadectomorphs nest pretty well with Orobates, millerettids, caseasaurs and limnoscelids. Now they nest just a little better with bolosaurs and pareiasaurs. At one point it was either way.

What sparked this change?
When I added Bashkyroleter, added data to Procolophon and Belebey, and created a new skull restoration of Sauropareion. Issues arose and I took another look at several dozen taxa.

On the plus side,
Stephanospondylus (still known from very poor data) has traditionally been considered a diadectomorph. Now it nests as a sister to that clade. Some diadectids had some widely expanded ribs, as did Stephanospondylus. Both were experimenting with a morphology that would be perfected in their now closer relatives, the turtles.

The limnoscelids and their kin, including Orobates, are all long-bodied taxa now. The diadectids plus pareiasaurs plus turtles are all shorter bodied taxa with progressively wider bodies and shorter tails.

The true procolophonids (Procolophon and kin) are now closer to the nycteroleterids and owenettids, which were traditionally associated in a large single clade. They’re still not directly related.

As noted earlier, one of the earmarks of good Science is correcting errors. I encourage the finding of errors in reptileevolutiion.com. Behind the scenes, as you already know, I make corrections and additions all the time. As mentioned earlier:

Carl Sagan (in the Demon Haunted World) wrote:
“Science has built-in error-correcting mechanisms—because science recognizes that scientists, like everybody else, are fallible, that we make mistakes, that we’re driven by the same prejudices as everybody else. There are no forbidden questions. Arguments from authority are worthless. Claims must be demonstrated. Ad hominem arguments—arguments about the personality of somebody who disagrees with you—are irrelevant; they can be sleazeballs and be right, and you can be a pillar of the community and be wrong.”

In other words,
watch out for those who hold dearly to their paradigms, whether religious or scientific. It’s okay to test those paradigms.

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 )

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.