New radical dinosaur cladogram: Baron, Norman and Barrett 2017

Baron, Norman and Barrett 2017
have just allied Ornithischia with Theropoda to the exclusion of Sauropodomorpha. That radical hypothesis was not recovered by the large reptile tree (LRT, 980 taxa) nor any other study in the long history of dinosaurs. Despite the large size of their study, it was not large enough. And so taxon exclusion bites another group of well-meaning paleontologists who used traditional small inclusion sets.

From the Baron et al. abstract:
“For 130 years, dinosaurs have been divided into two distinct clades—Ornithischia and Saurischia. Here we present a hypothesis for the phylogenetic relationships of the major dinosaurian groups that challenges the current consensus concerning early dinosaur evolution and highlights problematic aspects of current cladistic definitions. Our study has found a sister-group relationship between Ornithischia and Theropoda (united in the new clade Ornithoscelida), with Sauropodomorpha and Herrerasauridae (as the redefined Saurischia) forming its monophyletic outgroup. This new tree topology requires redefinition and rediagnosis of Dinosauria and the subsidiary dinosaurian clades. In addition, it forces re-evaluations of early dinosaur cladogenesis and character evolution, suggests that hypercarnivory was acquired independently in herrerasaurids and theropods, and offers an explanation for many of the anatomical features previously regarded as notable convergences between theropods and early ornithischians.”

As a reminder, the fully resolved cladogram
at finds Herrerasaurus as a basal dinosaur arising from the Pseudhesperosuchus clade. Tawa (Fig. 1) and Buriolestes lead the way toward Theropoda. Barberenasuchus and Eodromaeus are basal to Phytodinosauria, which includes Sauropodomorpha + Ornithischia. So the Nature piece is totally different due to taxon exclusion and improper taxon inclusion.

Earlier heretical dinosaur origins were presented here with images and complete resolution with high Bootstrap scores at every or virtually every node.

Problems with the Baron et al. report

  1. Lack of resolution: Over dozens of nodes, only 5 bootstrap scores were over 50 (the minimum score that PAUP shows as fully resolved).
  2. Lack of correct proximal outgroup taxa (taxon exclusion) and they chose several wrong outgroup taxa (see below) because they had no large gamut analysis that established the correct outgroup taxon out of a larger gamut of choices
  3. Lack of several basal dinosaur taxa. (again, taxon exclusion, see below)
  4. Improper taxon inclusion: poposaurs, pterosaurs and lagerpetons are not related to dinos or their closest kin
  5. Lacking reconstructions for all pertinent basal/transitinal taxa so we can see their data at a glance, see if a gradual accumulation of traits can be observed and not have to slog through all the scores
Figure 1. Unrelated archosaurs. Silesaurus is a poposaur. Eoraptor is a phytodinosaur (note the big belly). And Tawa is a lean theropod.

Figure 1. Unrelated archosaurs mentioned in this blog. Silesaurus is a poposaur. Eoraptor is a phytodinosaur (note the big belly). And Tawa is a lean theropod.

LRT differences with the Baron et. al results

  1. Carnivorous Staurikosaurus, Herrerasaurus, Chindesaurus and Sanjuansaurus nest at the base of the herbivorous Sauropodomorpha.
  2. Herbivorous Eoraptor nests at the base of the Theropod with Tawa.
  3. Poorly known Saltopus sometimes nests as the last common ancestor of Dinosauria.
  4. Six taxa nest basal to dinosaurs in SupFig1 including the poposaur Silesaurus and kin. Silesaurus has ornithischian and theropod traits and so appears to make an ideal outgroup taxon,  but nests with neither clade when more taxa are included. This is the key problem with the study: pertinent taxon exclusion. 
  5. The lack of Gracilisuchus and other bipedal basal crocs that nest basal to dinos in the LRT certainly skewed results.

In an effort to understand Baron et al. I duplicated their outgroup taxon list
but retained all the LRT dinosaurs to see what would happen. The SupFigs are available free online at

  1. SupFig 1: When Euparkeria is the outgroup and Postosuchus is included: 3 trees result and (theropods Herrerasaurus + Tawa + Buriolestes) + (poposaurs Sacisaurus + Silesaurus) nest as the base of the Phytodinosauria, while bipedal croc Saltopus nests at the base of the Theropoda.
  2. SupFig 2: When the lepidosaur pterosaur Dimorphodon is the outgroup and Euparkeria + Postosuchus are excluded: 12 trees and basal scansoriopterygid birds (come to think of it, they DO look like Dimorphodon!) nest as basal dinosaurs, then the bird cladogram gets reversed such that basal becomes derived, but Phytodinosauria is retained.
  3. SupFig. 3: when Silesaurus is the outgroup: 12 trees and Phytodinosauria is retained in the LRT
  4. SupFig. 4: when no characters were treated as ordered. Neither does the  LRT order any characters, so this test was moot.

Dr. Kevin Padian said, 
“‘original and provocative reassessment of dinosaur origins and relationships”. And because Baron and his colleagues used well-accepted methods, he notes, the results can’t simply be dismissed as a different opinion or as mere speculation. “This will send people back to the drawing board,” he added in an interview.”

“There have been a lot of studies on the phylogenetic relationships, the family tree of the dinosaurs, but they’ve mostly been on individual dinosaurian groups. They haven’t really examined the entire dinosaur tree in such depth. And so this analysis had the advantage of using a different and larger set of critters than most previous trees. They’ve analyzed the characters used by others before and then also adding their own characteristics and getting their selves quite different configurations, radically different in fact.

The LRT has had, for several years, an even larger set of taxa, so large that any bias in selecting an outgroup taxon list has been minimized. Unfortunately, Baron et al. were biased and used traditional outgroup taxa that skewed their results.

Dr. Hans-DieterSues reported,
“For one thing, palaeontologists’ analyses of relations among species are keenly sensitive to which species are considered, as well as which and how many anatomical features are included, he says.”

Many more outgroup taxa would have minimized the inherent bias clearly present in Baron et al. When Silesaurus is your outgroup, herbivores will nest with carnivores. When you start your study with a goal in mind (read and listen to Baron’s comments) that’s never good. When you exclude taxa that have been shown to be pertinent to your study, that’s never good.

That’s what is here for (on the worldwide web). Free. Testable. And with a demonstrable gradual accumulation of traits along with minimal bias due to its large gamut.

I was surprised to see Nature print this
because they have not published relationship hypotheses in favor of  new specimens of note. Co-author Dr. David Norman has published for several decades and has a great reputation.

Baron MG, Norman DB, Barrett PM 2017. A new hypothesis of dinosaur relationships and early dinosaur evolution. Nature 543:501–506.


6 thoughts on “New radical dinosaur cladogram: Baron, Norman and Barrett 2017

  1. I haven’t yet had an opportunity to read the paper (later today!), but I’ve read Darren Naish’s blog post about it. I quote:

    Let me emphasise that: this study comes as the result of the careful and thorough collection and analysis of a ton of new data.

    For example, Ornithoscelida has twenty autapomorphies, many of them being states of characters that have never been included in a phylogenetic analysis of dinosaurs before.

    So there’s a huge advantage of this paper over the LRT.

    BTW, there’s no way the silesaurids are poposaurs. Just look at them; even the head is completely different. Calling Eoraptor herbivorous is an exaggeration (omnivorous fits much better), and the oddly short trunk you reconstruct in combination with the oddly long pubis you reconstruct give it the shape of an ornithomimosaur, oviraptorosaur or alvarezsaurid… are those “phytodinosaurs” too?

    Lacking reconstructions for all pertinent basal/transitinal taxa so we can see their data at a glance, see if a gradual accumulation of traits can be observed and not have to slog through all the scores

    That’s a really unreasonable amount of work you’re demanding.

    • Besides, why do you want to eyeball the “gradual accumulation of traits” in the first place? Why not simply load the matrix and the tree into Mesquite (or MacClade, I guess) and trace the characters on the tree?

      And why do you keep making the assumption that the fossil record is dense enough and uniform enough for all branches to have the same morphological length!?!

    • Ornithoscelida has twenty autapomorphies

      Oops, sorry. I’ve now read the paper: it has 21 unambiguous and 5 ambiguous ones. I see you’ve written a post about them, so I’ll stop here.

      Concerning the outgroup sample, on the one hand I agree with you: it’s very small indeed. In the supp. inf. the authors justify their taxon inclusion, and it reads like they think the default is to include no taxa at all, and every exception to this rule requires a special reason… Their argument that the outgroup taxa should be as complete as possible so that all characters can be polarized is simply nonsense; that’s not how trees are rooted.

      On the other hand, however, they’ve done the experiment of replacing all outgroups with Dimorphodon. They got practically the same ingroup topology. Evidently, the outgroups don’t matter much in this particular dataset.

      • I found the same results despite outgroups, except for the Dimorphodon experiment where scansoriopterygid birds are basal theropods. : ) Having examined the Nesbit 2011 matrix they base their work on they made lots of mistakes with regard to pterosaurs, first of all my including them despite their lack of synapomorphies with related taxa. A classic ‘by default’ nesting.

    • As I mentioned earlier, outgroups should choose themselves. They were cherry-picked in the Baron et al. study. There is a way silesaurids are poposaurs. Just let them nest in a large gamut analysis. Poposaurs come in many shapes. Of the 21 synapomorphies I found 14 exceptions, they found others.

Leave a Reply

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

You are commenting using your 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.