Earlier we looked at Baron et al. 2017, who heretically allied ornithischians with theropods to create a new clade ‘Ornithoscelida.’ After a week of study and several tests, I came away with the realization that the Baron et al. tree topology did not depend on taxon exclusion, based on comparable results from subsets of the large reptile tree (LRT). Reducing the list of LRT taxa to match the Baron et al. taxon list did not change the dinosaur tree topology. Rather scoring inaccuracies were the problem, and that goes back to the original matrix presented by Nesbitt 2011, which Baron et al. continued using. You might remember, that’s the study that nests pterosaurs with parasuchians with no intervening taxa, among many other odd nestings. We’ll detail those inaccuracies later this week.
One of the reasons
Baron et al. allied theropods with ornithisichians is because they nested two phytodinosaurs, Eodromaeus (Fig. 1) and Eoraptor, as basalmost theropods. So when they thought they were comparing the ‘theropod’ Eoraptor with the ornithischian, Heterodontosaurus (their figure 2), they were actually comparing two phytodinosaurs. So, of course they shared a long list of characters. To really test their assertion, Baron et al. should have tested Heterodontosaurus against a real basal theropod, like Tawa (Fig. 2). (Tomorrow we’ll compare Tawa with Eodromaeus.)
Along the way,
I was able to better understand what was going on at the base of the Dinosauria and at the basal Theropod/Phytodinosaur split. For instance, as noted earlier, Herrerasaurus (Fig. 3) has a smaller number of cervicals than other dinos do, matching the number found in ancestral basal crocodylomorpha, like Lewisuchus, not silesaurids, like Silesaurus.
How do basal dinosaurs differ from the basal croc ancestors?
Here’s the latest list:
- skull shorter than cervical series (not Herrerasaurus):
- skull width less than 1.2x height;
- posterolateral premaxilla narrower than naris (not Herrerasaurus);
- quadrate curls posterodorsally (not Herrerasaurus);
- dentary tip straight or descends (does not rise);
- deep canine maxillary teeth absent (not Herrerasaurus);
- nine or more cervicals (not Herrerasaurus);
- cervical ribs slender, parallel centra (Herrerasaurus unknown);
- 3–4 sacral ribs (not Herrerasaurus, which has 2);
- mid-caudal centra 3x longer than tall where known (not Herrerasaurus);
- clavicles poorly ossified or absent (reappear in later taxa);
- interclavicle poorly ossified or absent;
- manual unguals long with penultimate phalanges longer than proximal phalanges;
- tibia not shorter than femur (not Herrerasaurus);
- fourth trochanter sharp (also in the proximal dinosaur outgroup taxon PVL 4597);
- tibia not < 2x ilium length (not Herrerasaurus);
- advanced and simple metatarsal joint without a calcaneal tuber (also in Lewisuchus, convergent in Silesaurus);
- compact metatarsus;
- longest metatarsal: #3;
- metatarsal #1 < 75% of metatarsal #3 length;
- metatarsals 2–4 ≥ tibia (not Herrerasaurus);
- proximal metatarsals 1 and/or 5 reduced in diameter.
Summary and significance
Arising from the morphology of basal bipedal crocodylomorphs, many of these basal dinosaur traits document a longer, leaner bipedal morphology with an increasingly robust fulcrum (pelvis and sacrum), along with longer distal hind limb elements with more reliance on the middle digits for faster locomotion. By contrast the proximal outgroup clade, Junggarsuchus and kin, documents a return to quadrupedalism terminating with Trialestes and its long forelimbs. The increasingly elongate radiale and ulnare of the Junggarsuchus clade turn out to be convergent, rather than homologous, with those of other bipedal crocodylomorphs, like Terrestrisuchus and Hesperosuchus, a clade that also transitioned toward a quadrupedal configuration, as seen in extant crocodylomorphs. We’ll take a closer look at this clade and this transition later this week.
Baron MG, Norman DB, Barrett PM 2017. A new hypothesis of dinosaur relationships and early dinosaur evolution. Nature 543:501–506.