New view on ‘Paravians’: Agnolin et al. 2019

Agnolin et al. 2019 produced
a new view of early bird and pre-bird relationships. They write, “We here present a review of the taxonomic composition and main anatomical characteristics of those theropod families closely related with early birds, with the aim of analyzing and discussing the main competing hypotheses pertaining to avian origins. We reject the postulated troodontid affinities of anchiornithines, and the dromaeosaurid affinities of microraptorians and unenlagiids, and instead place these groups as successive sister taxa to Avialae.”

By contrast
in the large reptile tree (LRT, 1401 taxa; subset Fig. 1) some troodontids are basal to anchiornithines, which are basal to avians. Other traditional troodontids are not basal to birds and pre-birds.

Agnolin et al. report, “Regarding character evolution, we found that: (1) the presence of an ossified sternum goes hand in hand with that of ossified uncinate processes; (2) the presence of foldable forelimbs in basal archosaurs indicates widespread distribution of this trait among reptiles, contradicting previous proposals that forelimb folding driven by propatagial and associated tendons was exclusive to the avian lineage; (3) in basal paravians and avialans (e.g., Archaeopteryx, Anchiornis) the wings are relatively large and wide, with relatively short rectricial feathers, a rounded alar contour, and a convex leading margin. These taxa exhibit restricted forelimb folding capability with respect to more derived birds, their hands being preserved at angles of flexion (with respect to the radius/ulna) of no less than 90. In more derived birds, however, the rectrices are notably elongate and the angle between the hand and forearm is much less than 90, indicating not only increased forelimb folding capability but also an increased variety of wingbeat movements during flight. Because of the strong similarities in pectoral girdle configuration between ratites and basal avialans and paravians, it is possible to infer that the main forelimb movements were similar in all these taxa, lacking the complex dorsoventral wing excursion characteristic of living neognathans.”

Unfortunately
Agnolin et al. presented a cladogram that was largely unresolved. According to the LRT that loss of resolution can be attributed to one thing: exclusion of taxa. Key taxa missing from the Agnolin et al. tree include:

  1. Compsognathus (both species)
  2. Ornitholestes
  3. The other ten or so ‘Archaeopteryx’ specimens

With the addition of these key taxa theropods (including pre-birds and birds) become completely resolved in the LRT (subset Fig. 1).

Figure 1. More taxa, updated tree, new clade names.

Figure 1. More taxa, updated tree, new clade names, from an earlier blog post.

References
Agnolin FL et al. (4 co-authors) 2019. Paravian phylogeny and the dinosaur-bird transition: an overview. Frontiers in Earth Science 6:252.
doi: 10.3389/feart.2018.00252

SVP 2018: Hindlimb feathers useful as brood covers in oviraptorids?

Hopp and Orsen 2018
bring a novel and well documented hypothesis to light: “Here we present evidence gleaned from our studies of a number of fossils that possess hind-limb feathers, as well as two examples of nesting Citipati. Two well preserved individuals sitting on nests with large egg clutches (IGM-100/979, IGM-100/1004) clearly demonstrate a lack of complete coverage of the eggs by the animals’ bodies and limbs. We previously showed that pennaceous feathers would have aided the coverage of eggs near the ulna and manus. We also noted a deficiency of egg coverage at the rear quarters laterally adjacent to the pelvis and tail. Here we demonstrate how pennaceous feathers, recently described on the tibiae and tarsi of several non-flying theropods and some primitive birds as well, could have served very effectively to cover eggs in these rear quarter positions.”

FIgure 1. From Zheng et al. 2013 showing the maximum extent of hind leg feathers in Anchiornis.

FIgure 1. From Zheng et al. 2013 showing the maximum extent of hind leg feathers in Anchiornis. Pedopenna nests with Anchiornis.

Excellent hypothesis. But…
Zheng et al. 2013 also studied this problem. They wrote, “parallel pennaceous feathers are preserved along the distal half of the tibiotarsus and nearly the whole length of the metatarsus in each hindlimb [of Sapeornis]. The feathers are nearly perpendicular to the tibiotarsus and metatarsus in orientation and form a planar surface as in some basal deinonychosaurs with large leg feathers.”

Zheng et al. 2013 also report similar leg and/or foot feathers are found in
“Basal deinonychosaurians (= Microraptor), the basal avialan Epidexipteryx, Sapeornis, confuciusornithids, and enantiornithines. In these taxa, the femoral and crural feathers are large, and in most cases they are pennaceous feathers that have curved rachises and extend nearly perpendicular to the limbs to form a planar surface.”

The distribution of foot feathers
in theropods in the large reptile tree (LRT, subset Fig. 2) is shown in blue (cyan). Few included taxa preserve feathers. The question is: do foot feathers appear, then disappear, then reappear? Or do all intervening taxa have foot feathers?

Figure 3. Where feathers on the foot are preserved on the LRT.

Figure 2. Where feathers on the foot are preserved on the LRT.

Back to the brooding question:
Citipati is an oviraptorid and oviraptorids are outside of the occurrences of foot feathers in theropods in the LRT. Note: all specimens with foot feathers are a magnitude smaller than oviraptorids. Hopp and Orsen do not differentiate (in their abstract, I did not see their presentation) between tibial feathers and foot feathers. Citipati nests outside of the current phylogenetic bracket for foot feathers. Tibial feathers have a much wider distribution in fossils. Tibial feathers are more likely to be present in Citipati, but note: tibial and foot feathers are not present in Caudipteryx (Fig. 3) an oviraptorid sister in the LRT .

Figure 3. Caudipteryx preserves forelimb and tail feathers, but no leg or foot feathers. It nests with oviraptorids in the LRT.

Figure 3. Caudipteryx preserves forelimb and tail feathers, but no leg or foot feathers. It nests with oviraptorids in the LRT.

Back to the question of pennaceous hind limb feathers in pre-birds:
Here’s one answer, perhaps convergent with the presence of large uropatagia in flapping, but non-volant fenestrasaurs (like Cosesaurus Fig. 4). And look at the long legs and large uropatagia of the basalmost pterosaur, Bergamodactylus (Fig. 4)! It was just learning how to flap and fly and could use a little aerodynamic help in keeping steady.

When pre-birds, like Anchiornis,
and other convergent theropods, like Microraptor, first experimented with flapping and leaving the ground, they were necessarily new at it, not perfect at coordinated symmetrical flapping. Perhaps pre-birds used a bit of aerodynamic stabilization in the form of hind limb feathers as they phylogenetically became better and better at flapping, then flying. Tibial and foot feathers may have provided that aerodynamic stability, acting like vertical stabilizers in most airplanes. Exceptionally, present-day flying wing-type airplanes no longer require a vertical stabilizer because computers assist the pilot in controlling the aircraft, just as modern birds control flight without vertical stabilizers. That’s because modern birds with unfeathered feet have established neural networks not present or only tentatively present in pre-birds.

Figure 1. Bergamodactylus compared to Cosesaurus. Hypothetical hatchling also shown.

Figure 4. Bergamodactylus compared to Cosesaurus. Hypothetical hatchling also shown. Look at those large uropatagia. Those are for stability in this student pilot, not yet as coordinated as in later, more derived pterosaurs.

References
Hopp TP and Orsen MJ 2018. Evidence that ‘four-winged’ paravian dinosaurs may have used hindlimb feathers for brooding.” SVP abstracts.
Hu D, Hou L, Zhang L and Xu X 2009. A pre-Archaeopteryx troodontid theropod from China with long feathers on the metatarsus. Nature 461(7264):640-3. doi: 10.1038/nature08322.
Longrich N 2006. Structure and function of hindlimb feathers in Archaeopteryx lithographica. Paleobiology 32 (3), 417-431
Xu X and Zhang F 2005. A new maniraptoran dinosaur from China with long feathers on the metatarsus. Naturwissenschaften. 92(4): 173–177.
Zhang F-C and Zhou Z-H 2004. Palaeontology: Leg feathers in an Early Cretaceous bird. Nature 431, 925(2004). doi:10.1038/431925a
Zheng X-T et al. 2013. Hind wings in basal birds and the evolution of leg feathers. Science 339:1309-1312. DOI: 10.1126/science.1228753

“Kinematics of wings from Caudipteryx to modern birds”: Talori et al. 2018

A new paper without peer-review by Talori, Zhao and O’Connor 2018
seeks to “better quantify the parameters that drove the evolution of flight from non-volant winged dinosaurs to modern birds.”

Unfortunately
they employ Caudipteryx, an oviraptorosaur. They correctly state,
Currently it is nearly universally accepted that Aves belongs to the derived clade of theropod dinosaurs, the Maniraptora.” They incorrectly state, “The oviraptorosaur Caudipteryx is a member of this clade and the basal-most  maniraptoran with pennaceous feathers.” In the large reptile tree (LRT, 1269 taxa) oviraptorosaurs nest with therizinosaurus, and more distantly ornithomimosaurs. This clade is separated from bird ancestor troodontids by the Ornitholestes/Microraptor clade.

Figure 1. More taxa, updated tree, new clade names.

Figure 1. Caudipterys is in the peach-colored clade, far from the lineage of birds.

The Talori team
mathematically modeled Caudipteryx with three hypothetical wing sizes, but failed to provide evidence that the Caudipteryx wing was capable of flapping. In all flapping tetrapods the elongation of the coracoid  (or in bats of the clavicle) signals the onset of flapping… and Caudipteryx does not have an elongate coracoid. Rather, it remains a disc.

So, no matter the math, or the accuracy of the mechanical model,
the phylogeny is not valid and the assumption of flapping is inappropriate. It would have been better if they had chosen a troodontid and several Solnhofen birds to test.

Tossing those issues aside,
the Talori team did an excellent job of setting their mechanical model (which could be a troodontid) in a wind tunnel, extracting data from three different wing shapes and presenting their findings. Feathers would have been more flexible than their mold manufactured wings, but the effort is laudable.

References
Zhao J-S, Talori YS, O’Connor J-M 2018. Kinematics of wings from Caudipteryx to modern birds. [not peer-reviewed] bioRXiv
https://www.biorxiv.org/content/early/2018/08/16/393686

http://reptileevolution.com/reptile-tree.htm

Serikornis: Pre-bird or flightless bird?

Several authors have wondered over the years
how we might be able to tell (or nest) a flightless post-Archaeopteryx  bird from a flightless pre-Archaeopteryx troodontid. Earlier we nested a very large flightless sapeornithid bird, Jianianhualong, distinct from its original nesting as a troodontid. So it can be done.

Figure 1. Serikornis and Jurapteryx (Archaeopteryx) recurva to scale. These two nest as sisters in the LRT.

Figure 1. Serikornis and Jurapteryx (Archaeopteryx) recurva to scale. These two nest as sisters in the LRT. The larger Serikornis was non-volent.

Lefèvre et al. 2017 bring us
a new Late Jurassic ground-dwelling theropod from China, Serikornis sungei (Figs. 1–3; PMOL-AB00200; 50 cm in length) that they nested with the derived troodontid, Eosinopteryx. They reported, “The plumage of this new specimen brings new information on the structure and function of the feathers in basal paravians and consequently on the early evolution of flight.”

By contrast
in the large reptile tree (LRT, 1050 taxa) Serikornis nests strongly with the Eichstätt specimen of Archaeopteryx, aka Jurapteryx recurva. That Solnhofen bird has large wing and tail feathers. The latest Jurassic, earliest Cretaceous formations from which Serikornis came are chronologically appropriate to this relationship. Apparently taxon exclusion by the Lefèvre team is the cause of the disparate nestings.

Figure 2. Serikornis in situ, with original drawing, skull under DGS and reconstructed.  As you can see, the metatarsus was feathery, not scaly, and the wing feathers were reduced. The teeth were longer, curved and sharper.  DGS did a pretty good job with the skull.

Figure 2. Serikornis in situ, with original drawing, skull under DGS and reconstructed.  As you can see, the metatarsus was feathery, not scaly, and the wing feathers were reduced. The teeth were longer, curved and sharper.  DGS did a pretty good job with the skull.

As earlier authors have noted
the most likely time for an early volant bird to go back to flightlessness is when they are still not very good at flight. And that seems to be the case here. Serikornis probably got too big to fly. And its teeth were larger, opposite the general trend for volant birds. And so its flight feathers, like those of any number of extant and extinct flightless birds, became less able to perform aerial duties.

What about that short coracoid?
It is not long and strap-shaped, a common shape in flapping tetrapods. The coracoids in the Eichstätt specimen are lost in a crack so coracoids could not be scored for that Solnhofen bird. That short coracoid of Serikornis must have been a reversal, an atavism. That happens. It’s only one trait out of 228.

Maybe a sternum was overlooked.
The two putative coracoids (Fig. 3) do not have the same outline. So I wonder if one of them was a sternum? Certainly part of the large furculum is buried.

Figure 3. Serikornis pectoral girdle. Here one of the putative coracoids is rei-dentified as a sternum rotated from its in vivo position.

Figure 3. Serikornis pectoral girdle. Here one of the putative coracoids is rei-dentified as a sternum rotated from its in vivo position. A tiny portion of the bottom coracoid peeks out (in indigo).

At first I scored Serikornis
by copying the row for Eosinopteryx then renaming it. Soon distinct scores started appearing. The list became long. PAUP nested Serikornis apart from Eosinopteryx, among the very early birds and with Jurapteryx recurva, close to the base of the clade that includes all extant birds.

An abbreviated list of birdy traits in Serikornis include:

  1. orbit in the posterior half of the skull
  2. ascending process of premaxilla extends to frontals
  3. tail longer than presacral spine
  4. that long gracile pubis
  5. fibula poorly ossified to absent at mid length
  6. metatarsal 5 lacking phalanges

So the claim to fame for this taxon
should have been yet another one of the earliest flightless birds –- not a transitional troodontid documenting the advent of flight feathers. These flight feathers were on their way out, not on their way in.

References
Lefèvre U, Cau A, Cincotta A, Hu D-Y, Chinsamy A, Escuillié F and Godefroit P 2017. A new Jurassic theropod from China documents a transitional step in the macrostructure of feathers. Sci Nat 104:74. DOI 10.1007/s00114-017-1496-y

to soon yet for a Wikipedia article

Liaoningvenator: Bird-like troodontid? Or troodontid-like bird?

Shen et al. 2017 describe
a new troodontid, Liaoningvenator curriei (DNHM D3012; Dalian Natural History Museum; Figs. 1-2; Early Cretaceous), they nest Liaoningvenator outside of the Aves (birds).

Figure 1. Liaoningvenator has a long neck and short torso. It nests as a secondarily flightless bird in the LRT, rather than as a troodontid.

Figure 1. Liaoningvenator has a long neck and short torso. It nests as a secondarily flightless bird in the LRT, rather than as a troodontid.

From the abstract:
“A new troodontid, Liaoningvenator curriei gen. et sp. nov., is described based on a complete skeleton from the Early Cretaceous Yixian Formation of Beipiao City, Liaoning Province. It bears the following characteristics of Troodontidae: numerous and more closely appressed maxillary and dentary teeth; the teeth markedly constricted between the roots and crowns; the nutrient foramina in groove on the external surface of dentary; distal caudal vertebrae having a sulcus on the dorsal midline rather than a neural spine. Unlike other troodontids, Liaoningvenator exhibits a sub-triangular ischial boot in lateral view and slender ischial obturator process; transition point in caudal vertebrae starts from the seventh caudal vertebra. A phylogenetic analysis recovers Liaoningvenator and Eosinopteryx as sister taxa that belong to the same clade.”

Figure 2. Troodontid-like birds and bird-like troodontids shown together to scale.

Figure 2. Troodontid-like birds and bird-like troodontids shown together to scale. Note the robust hind limbs  in the secondarily flightless birds, Jianianhualong and Liaoningvenator.

By contrast,
the large reptile tree (LRT, 1011 taxa) nests Liaoningvenator with Jianianhualong as a large flightless basal sapeornithid bird—and all birds nest within the Troodontidae. Size-wise Liaoningvenator is midway between the smaller Archaeopteryx recurva (Fig. 2) and the larger Jianianhualong. So this might be a transitional taxon between the two.

Unrelated
Eosinopteryx (Fig. 2) continues to nest outside of Aves (birds). Distinct from Eosinopteryx, Liaoningvenator has a much shorter torso and much longer neck, as in other birds. Like Jianianhualong metarsal 4 is longer than 3 in Liaoningvenator, among many other traits (see below). Shen et al. did not mention Jianianhualong, probably because the two taxa were published within a few weeks of each other. You might remember earlier Xu et al. 2017 also nested Jianianhualong with the non-avian troodontids. Shen et al. included Sapeornis in their phylogenetic analysis. Not sure why they nested apart in the LRT.

A reconstruction of the Liaoningvenator skull
(Fig. 2) has a large openings and gracile bones. What Shen et al. identified as a maxillary foramen is identified here as the base of the naris. The in situ tail curls anteriorly and several caudal vertebrae are visible over the torso.

From the Shen et al. diagnosis:
“A new troodontid dinosaur bears the following unique combination of characters including autapomorphies indicated with an asterisk and new characters indicated with a double asterisk: prominent slender triradiate postorbital*; deltopectoral crest distinctly extended to the half of the humeral shaft*; no posterior process on the dorsodistal end of ischium**; slender obturator process of ischium**; manual phalanx I-1 longer than metacarpal II**, the length ratio of phalanx I-1 to metacarpal II about 1.49**; the width of metatarsus distally distinctly decrease**; transition point in caudal series starts from the seventh caudal vertebra**.

Troodontid or not?
The large flightless basal birds share a long list of traits in common with troodontids and a few that show they are distinct. Here is a list of the differences between bird-like troodontids, like Sinornithoides and Anchiornis, and the troodontid-like sapeornithid birds, like Jianianhualong and Liaoningvenator.

Liaoningvenator bird traits not shared with non-avian troodontids:

  1. Ventral aspect of premaxilla > 1/3 preorbit length
  2. Ascending process of premaxilla extends beyond naris and contacts frontals (nasal separated)
  3. Lacrimal deeper than maxilla
  4. Major axis of naris 30-90º
  5. Posterolateral premaxilla absent (also in Xiaotingia and Eosinopteryx)
  6. Nasals not longer than frontals (also in Xiaotingia and Eosinopteryx)
  7. Antorbital fenestra without fossa
  8. Manual mc2 and 3 do not align with joints on digit 1
  9. Metatarsal 5 not shorter than pedal digit 5

Shifting
Liaoningvenator and Jianianhualong to Sinornithoides adds 14 steps.

Paul 2002
considered the possibility of secondarily flightless (neoflightless) birds, unfortunately without the benefit of a phylogenetic analysis. Paul wrote: “Reversal normally associated with loss of flight is observed in ornithomimids, therizinosaurs and dromaeosaurs.” The LRT found possibly volant bird-like taxa associated with therizinosaurus (Rahonavis), Ornitholestes (microraptorids) and troodontids (birds), but not ornithomimids (related to Compsognathus) and dromaeosaurs (related to Shuvuuia).

Paul wrote:
“The less sharply flexed, broad coracoids of flightless birds recapitulate the dino-avepod condition. The loss of any sternal keel and shortening of the arms area also normal reversals for flightless birds. The semilunate carpal block and arm folding mechanism…are sometimes lost in flightless birds.”

References
Paul G 2002. Dinosaurs of the Air. Johns Hopkins Press
Shen C-Z, Zhao B, Gao C-L, Lü J-C and Kundrat 2017. A New Troodontid Dinosaur (Liaoningvenator curriei gen. et sp. nov.) from the Early Cretaceous Yixian Formation in Western Liaoning Province. Acta Geoscientica Sinica 38(3):359-371.
Xu X, Currie P, Pittman M, Xing L, Meng QW-J, Lü J-C, Hu D and Yu C-Y 2017. Mosaic evolution in an asymmetrically feathered troodontid dinosaur with transitional features. Nature Communications DOI: 10.1038/ncomms14972.

Jianianhualong: not a bird-like troodontid — it’s a troodontid-like bird

Revised May 11, 2017 with new nesting for Chiappeavis. 

Once again, taxon exclusion issues arise
Colleagues, we have to let the taxa nest themselves. Don’t restrict your inclusion sets to the short list of taxa you prefer! In the Jianianhualong paper a long list of candidate taxa were excluded, including its actual proximal sisters (Fig. 6). And they missed the other big headline that should have attended this new taxon. See below.

Figure 1. Jianianhualong tengi in situ. This is the largest among the early birds, a fact overlooked by the Xu et al. 2017. Think of Jianianhualong as a giant Archaeopteryx!

Figure 1. Jianianhualong tengi in situ. This is the largest of the early birds, a fact overlooked by the Xu et al. 2017. Think of Jianianhualong as a giant Archaeopteryx!

Xu et al. 2017
bring us a new genus of theropod dinosaur with feathers, Jianianhualong tengi (DLXH 1218; Yixian Formation, Early Cretaceous; Fig. 1). They nested their new find in an unresolved clade including the non-bird troodontid, Sinornithoides (Fig. 5). Notably they did not resolve Solnhofen birds (Archaeopteryx’ specimens), troodontids and dromaeosaurids. That should have been a red flag that more effort was needed to weed out bad scores in their matrix. Maybe a reconstruction would have helped? (Fig. 3).

FIgure 2. GIF animation of the skull of Jianianhualong showing original tracing in line art and colorized bones (DGS) used to create a reconstruction (Fig. 3).

FIgure 2. GIF animation of the skull of Jianianhualong showing original tracing in line art and colorized bones (DGS) used to create a reconstruction (Fig. 3).

Here
in the large reptile tree (LRT, 1004 taxa) Jianianhualong tengi nests strongly with sapeornithid birds, despite its long bony tail, short forelimbs and large size, all atavistic traits retained in this one of the first flightless birds and certainly one of the first large flightless birds. This aspect was overlooked by Xu et al. 2017 as they mistakenly considered this a feathered non-bird troodontid. It is a bird. A big flightless bird.

Figure 3. Reconstruction of the skull of Jianianhualong based on DGS tracings in figure 2.

Figure 3. Reconstruction of the skull of Jianianhualong based on DGS tracings in figure 2.

If you think the long tail of Jianianhualong is an issue…
Archaeopteryx recurva (the Eichstaett specimen) nests with Jianianhualong and it has a long bony tail, too.

Figure 4. Colorizing the bones of Jianianhualong helps separate them from other elements on the matrix.

Figure 4. Colorizing the bones of Jianianhualong helps separate them from other elements on the matrix better than simply dropping a two letter abbreviation somewhere on the bone.

If you think the large size of Jianianhualong is an issue…
think of it like an early ostrich, flightless with no sternum, a giant Archaeopteryx in the Early Cretaceous, running from more primitive dinosaur-eating theropods.

Figure 5. The Xu et al. cladogram that nests Jianianhualong with troodontids. Note the loss of resolution at important nodes. Compare to the LRT in figure 6.

Figure 5. The Xu et al. cladogram that nests Jianianhualong with troodontids. Note the loss of resolution at important nodes. Compare to the LRT in figure 6. The LRT is fully resolved with more taxa.

Unfortunately
Xu et al. did not test taxa that actually nest closer to Jianianhualong, using an antiquated matrix with only two Solnhofen birds. Xu et al. report, “The discovery of Jianianhualong provides direct evidence for the presence of pennaceous feathers in an unquestionable troodontid theropod.” Since all birds are troodontids in the LRT this statement is true. However, Xu et al. were not thinking that birds arose from troodontids (Fig. 5), so this became a surprising discovery for them. As in so many other cases discussed herein, character traits come as no surprise when the taxon in question is correctly nested.

Fgure 6. Subset of the LRT focusing on birds and their immediate ancestors. Note the nesting of Jianianhualong with Sapeornis.

Fgure 6. Subset of the LRT focusing on birds and their immediate ancestors. Note the nesting of Jianianhualong with Sapeornis.

I just add taxa
and the software/cladogram does the rest. No initial bias. Reconstructions help. So does colorizing the bone. In this case, at least, working from the photo with DGS was more instructive and better able to demonstrate observations to others than traditional firsthand access labeled with small two-letter abbreviations.

Xu et al. 2014
made a headline out of the asymmetric feathers found with Jianianhualong. In the present context, Jianianhualong is derived from volant ancestors (Figs. 1, 6). So, asymmetry is not exceptional, but expected. Xu et al. reported, “Most significantly, the taxon has the earliest known asymmetrical troodontid feathers, suggesting that feather asymmetry was ancestral to Paraves.” The entire statement is false under the present hypothesis of interrelationships.

The unfortunate return of ‘modular evolution.”
Xu et al cite references to the concept of ‘modular (mosaic) evolution‘ which is based on invalid phylogeny. Please avoid ‘modular evolution’. That’s not how evolution works in the real world.

References
Xu X, Currie P, Pittman M, Xing L, Meng QW-J, Lü J-C, Hu D and Yu C-Y 2017. Mosaic evolution in an asymmetrically feathered troodontid dinosaur with transitional features. Nature Communications DOI: 10.1038/ncomms14972.

Paravian phylogeny revisited – SVP abstracts 2016

Pei et al. 2016
reveal the origin of birds in a new phylogenetic analysis. Some aspects confirm earlier recoveries in the large reptile tree (LRT) made about a year ago. Not sure about other aspects given the brevity of the abstract and lack of cladogram imagery.

From the Pei et al. 2016 abstract
“Paraves are theropod dinosaurs comprising of living and fossil birds and their closest fossil relatives, the dromaeosaurid and troodontid dinosaurs. Traditionally, birds have been recovered as the sister group to Deinonychosauria, the clade made up of the two
subclades Dromaeosauridae and Troodontidae. However, spectacular Late Jurassic paravian fossils discovered from northeastern China – including Anchiornis and Xiaotingiapreserve anatomy that seemingly challenges the status quo. (1) To resolve this debate we performed an up-to-date phylogenetic analysis for paravians using the latest Theropod Working Group (TWiG) coelurosaur data matrix which we supplemented with new data from recently described Mesozoic paravians from Asia and North America (e.g., Zhenyuanlong and Acheroraptor). This includes data from the unnamed dromaeosaurid IVPP V22530 and Luanchuanraptor, which are included in a phylogenetic analysis for the first time. We also incorporate new data from iconic paravians such as Archaeopteryx and Velociraptor based on firsthand study. (2) The analysis adopted the maximum parsimony criterion and was performed in the phylogenetic software TNT. Our preliminary results support the monophyly of each of the traditionally recognized paravian clades. (3) The Late Jurassic paravians from northeastern China (e.g., Anchiornis and Xiaotingia) are recovered as avialans rather than deinonychosaurians, at a position more basal than Archaeopteryx and other derived avialans (4). The traditional sister group status of Troodontidae and Dromaeosauridae is reaffirmed (5) and is supported by a laterally exposed splenial and a characteristic raptorial pedal digit II. Recently reported Early Cretaceous dromaeosaurids from northern and northeastern China, including Zhenyuanlong, Changyuraptor and IVPP V22530, are closely related to other microraptorines as expected. (6) Luanchuanraptor, a dromaeosaurid from the Late Cretaceous of central China is recovered as a more advanced eudromaeosaurian. By tracing character evolution on the current tree topology we report on the latest insights into the adaptive radiation amongst early paravians, including the origin of flight and changes in body size and diet. (7)

Notes

  1. In the LRT Xiaotinigia and Anchiornis have nested as derived troodontids, basal to birds since their insertion into the LRT more than 3 years ago. So that’s confirmation that troodontids are basal to Archaeopteryx and other birds with Xiaotinigia and Anchiornis as proximal outgroup taxa.
  2. But did they include five or more Archaeopteryx specimens, as in the LRT? They don’t say so…
  3. In the LRT there is a clade that includes Velociraptor, but the Troodontidae does not produce a clade that does not include birds. Rather birds are derived troodontids in a monophyletic clade.
  4. If avialans are usually defined as all theropod dinosaurs more closely related to modern birds (Aves) than to deinonychosaurs, all troodontids are avialans in the LRT. Since Troodontidae was named by Gilmore in 1924, the term Avialae (Gauthier 1986) is a junior synonym.
  5. Troodontidae and Dromaeosauridae are also sisters in the LRT.
  6. This confirms the topology recovered in the LRT from about a year ago. Microraptorines, like Microraptor and basal tyrannosauroids like Zhenyuanlong are not related to troodontids or birds, but to tyrannosaurs and compsognathids.
  7. I’d like to see their tree whenever it is published to compare the two.

Figure 7. Bird cladogram with the latest additions. Here the referred specimen of Yanornis nests with enantiornithes while Archaeovolans nests within the Scansoriopterygidae, not with Yanornis.

Figure 1. Bird cladogram from several months ago. Here Avialae is a junior synonym for Troodontidae.

References
Pei R, Pittman M, Norell M and Xu X 2016. A review of par avian phylogeny with new data. Abstract from the 2016 meeting of the Society of Vertebrate Paleontology.

2015 SVPCA abstract supports troodontid-bird clade

Nice to get confirmation
for a subset of the large reptile tree in a SVPCA poster (Brougham 2015).

From the Brougham results:
“The modified matrix strongly supports a Troodontidae + Avialae clade rather than a monophyletic Deinonychosauria, a topology remarkably convergent on that seen in modified Godefroit phylogeny, in which Aurornis, Eosinopteryx and the Tiaojishan paravians form a sister clade to Anchiornis and more derived avialans, the two of which in turn form a sister clade to Troodontidae.”

Figure 1. Basal theropod subset of the large reptile tree showing troodontids basal to birds and separate from dromaeosaurs.

Figure 1. Basal theropod subset of the large reptile tree showing troodontids (light red) basal to birds (red) and separate from dromaeosaurs (white).

References
Brougham T 2015. Multi-matrix analysis of new Chinese feathered dinosaurs supports troodontid-bird clade. researchgate.net/publication/280728942

Sinovenator: the troodontid that was almost a bird

Sinovenator changii (Xu et al. 2002, Fig. 1) was originally correctly identified as a troodontid. That’s the good news.

Figure 1. The skull of Sinovenator, a short-faced troodontid basal to birds.

Figure 1. The skull of Sinovenator, a short-faced troodontid basal to birds.

Unfortunately
Turner, Makovicky and Norell (2012) nested Sinovenator with Mei (a scansoriopterygid bird in the large reptile tree of 654 taxa) and Xixiasaurus (a long-rostrum troodontid), both far from Avialae (birds). By contrast, the large reptile tree (Fig. 2) nests Sinovenator in the troodontid grade (not a distinct clade) close to two other short-faced troodontids, Xiaotingia and Eosinopteryx, which are proximal to Archaeopteryx and birds (Fig. 2).

Figure 2. Cladogram of the Paraves. Note the position of Sinovenator basal to birds.

Figure 2. Cladogram of the Paraves. Note the position of Sinovenator basal to birds.

Turner et al. 2012 reported,
“The discovery of Sinovenator and the examination of character distributions along the maniraptoran lineage indicate that principal structural modifications toward avians were acquired in the early stages of maniraptoran evolution.”

By contrast,
in Figure 2, Sinovenator represents not an early stage of maniraptoran evolution, but a later stage, which, in turn, is an early stage in avian evolution. That was a taxonomic flip-flop. See below for details.*

Turner et al. 2012 concluded: “The morphological gap between the par avian clades has blurred to the point that basal dromaeosaurids, troodontids, and avialans are nearly indistinguishable from one another, and in life these animals would appear extremely similar. However, important morphological divisions exist that allow us to understand the basic relationship between these three clades. 

“Taxon sampling within Paraves is the most exhaustive to date, but the phylogenetic hypotheses discussed herein will certainly not be the last word on par avian or coelurosaurian relationships. Indeed, changes and additions to similar data matrices are already yielding interesting results within the various clades of Coelurosauria. Moreover, the potential for new discovery that will modify these results is a given.”

Therein we may be seeing part of the problem with the Turner et al. 2012.
The large reptile tree was built taxon by taxon over several years and theropods were only seriously considered in the last year. I started with a few theropod taxa that were essentially complete and later added taxa that were also complete. In this way the correct tree was built from the beginning with incrementally increasing resolution. Afterwards less complete taxa were added or avoided. As everyone knows, incomplete sisters lead to loss of resolution, which is a party killer.

If Turner et al. had started with a dozen or two well-known taxa,
then added taxa incrementally, perhaps there would have been more lumping and splitting from the get go*.

Moreover
every new taxon was scored without bias. I did not know where the lesser known taxa should nest. As theropod experts, Turner et al. sort of knew how their tree would be built.

Moreover part 2
Turner et al 2012 did not test individual Archaeopteryx specimens as unique taxa. Thus they were not able to identify Mei as a scansoriopterygid bird. They also did not test the basal paravaian, Tanycolagreus, but they did test the basal paravian (not a tyrannosauroid), Eotyrannus. They did not test Sinocalliopteryx, Suchomimus, Guanlong or Xiongguanlong, which are not paravians, but would have attracted Dilong and Proceratosaurus, causing their deletion from the inclusion set. Turner et al. erroneously nested Juravenator and Sinosauropteryx with Compsognathus. They nest elsewhere in the large reptile tree (Fig. 2).

Flip-Flopped Troodontidae
In phylogenetic analysis it is easy to flip flop taxonomic lineages. As an example, in the large reptile tree the proximal basal taxa to the Solnhofen birds (Archaeopteryx, etc.) are the troodontids, beginning with the primitive dromaeosaurid-like Sinornithoides and ending with the derived  bird-like Anchiornis, Sinovenator, Xiaotingia and Eosinopteryx. The Turner et al. study flipped this order by nesting the bird-like Xiaotingia and Anchiornis as basal troodontids while nesting the dromaeosaurid-like Sauronithoides as the most derived troodontid taxon.

Unfortunately
Turner et al. did not include Aurornis, which was published a year later. So, that’s not their fault. The long-snouted troodontid Aurornis would have attracted even longer snouted Buitreraptor away from the dromaeosaurids (Fig. 2). (See above for the Turner et al. prediction of this sort of thing happening after their publication.)

Among troodontids
(and I’m just now learning this) there were basal forms with medium length skulls like Sinoornithoides, and (so far) two derived clades. One tended toward ever longer skulls, like Buitreraptor. The other tended toward shorter, more bird-like skulls, like Sinovenator. The Turner et al. study flip-flopped and otherwise confused this.

In Conclusion
Taxon inclusion does solve problems. But I can see where too many incomplete taxa can cause problems. Flip-flopping taxonomic lineages is a problem one has to avoid. It is also important to reconstruct all taxa, which is something Turner et al. did not do, or at least did not publish.

*The Turner et al study reminds me of the mechanical pterosaur ornithopter video in which the experts added fur, teeth and eyeballs to their flying model before affirming that the wings worked! They didn’t work, unfortunately for everyone in the video and watching the video.

References
Turner AH, Makovicky PJ Norell MA 2012. A Review of Dromaeosaurid Systematics and Paravian Phylogeny. Bulletin of the American Museum of Natural History 371:1–206. doi:10.1206/748.1. Online here.
Xu X, Norell MA, Xiao-lin W, Makovicky PJ, Xiao-chun W 2002. A basal troodontid from the Early Cretaceous of China. Nature 415: 780–784. doi:10.1038/415780aPMID 11845206.

wiki/Sinovenator

Eosinopteryx – part 3 – to scale

Updated November 5, 2015 with a new interpretation of the pectorals,a new nesting and images of Eosinopteryx and Xiaotingia

Sometimes it just helps
to see taxa to scale with possible sisters (Eosinopteryx in this case, Fig. 1). Smaller than Aurornis, Eosinopteryx also had a shorter snout, shorter torso, shorter tail, more robust clavicle and shorter pubis.

Figure 1. Eosinopteryx and kin, including Xiaotingia, Aurornis and Archaeopteryx (Thermopolis).

Figure 1. Eosinopteryx and kin, including Xiaotingia, Aurornis and Archaeopteryx (Thermopolis).

Godefroit et al. (2013) reported, “The straight and closely aligned ulna-radius of Eosinopteryx also means that pronation/supination of the manus with respect to the upper arm would have been limited; combined with the absence of a bony sternum and weakly developed proximal humerus, these attributes suggest that Eosinopteryx had little or no ability to oscillate the arms to produce a wing beat.”

Funny that they didn’t even mention the short coracoid.
The locked down elongate coracoid is a hallmark of flapping tetrapods (pterosaurs and birds) and an elongate clavicle does the same thing in bats.

The ulna is not bowed
in Aurornis or Eosinopteryx. It is bowed in Xiaotinigia and Archaeopteryx and more greatly bowed in subsequent flapping taxa, including oviraptorids by convergence. The coracoid is strut-like and locked down to the sternum in Xiaotingia and Archaeopteryx, perhaps by convergence because overall Archaeopteryx has proportions more similar to Aurornis.

Figure 2. Xiaotingia is an outgroup taxon to basal birds. The left coracoid is broken and reconstructed here. The coracoid should be as deep as the furcula. The coracoid is longer here than in Eosinopteryx implying a greater ability to flap.

Figure 2. Xiaotingia is an outgroup taxon to basal birds. The left coracoid is broken and reconstructed here. The coracoid should be as deep as the furcula. The coracoid is longer here than in Eosinopteryx implying a greater ability to flap.

The bowed antebrachium
produces a parallelogram in living birds that serves to automatically extend and fold the manus bearing the outer flight feathers with flexion/extension of the elbow. Prior to this, muscle power would have to extend and bend the wrist, independent of the flexion/extension of the elbow.

References
Godefroit P, Demuynck H, Dyke G, Hu D, Escuillié FO and Claeys P. 2013. Reduced plumage and flight ability of a new Jurassic paravian theropod from China. Nature Communications 4: 1394. doi:10.1038/ncomms2389

wiki/Eosinopteryx