Last time on The Theropod Database we looked at Xu et al.'s (2026) new supposed basal avialan Yuezhuornis. But not really, we just looked at their terrible, TERRIBLE skeletal reconstruction, decided ChatGPT doesn't seem capable of making it bad in the ways it was bad (so it was probably human failure), and checked how peer review covered it. I didn't really argue about anatomy except to point out that the authors' own tracings of the bones didn't match their skeletal reconstruction, so at least one had to be wrong. As I said at the time in response to the illustrious Anonymous #X (surely not connected to the authors) "The only things I reinterpreted are what fenestra is in the maxilla (and neither fenestra in the reconstruction matches the one they traced for the in situ fossil) and the postorbital rotation (which again wouldn't match their reconstruction no matter how you orient it)." But now I HAVE looked at the figures and read the text, so what do we find? My paleontological muse Anonymous #X said "you need a feature-by-feature comparison: which reconstructed traits are avialan-like, which character states the authors code differently, and why the fossil supports your reading rather than theirs. A visual impression of a 2D drawing is not that comparison." So Anonymous #X, that's what we're getting today! (minus the scoring part, which will be the third and final post on this topic) I'm going to divide this into two sections- explicit statements by the authors that I think are incorrect, and explicit statements by the authors I think can't be demonstrated. Then a review of how this affects the diagnosis and a conclusion.
Now the unfortunate thing about the Nanyuan Formation is that the preservation sucks. Elements flatten into each other so determining edges is often impossible, and bones are almost always visible via cross sections. Luckily, Xu et al. provided very high resolution figures for most of the specimen, so I was able to interpret a lot! Kudos to them!
 |
| Yuezhuornis skull (top to bottom, scaled to 'jugal' size) Xu et al.'s (2026) reconstruction from Extended Data Figure 1; Xu et al.'s tracing from Figure 1b; specimen in situ from Xu et al.'s Extended Data Figure 3a; my tracing over the in situ specimen; my reconstruction from that tracing. Unlike my last post, DO trust my reconstruction here, as it is my best attempt. Black > blue > red outlines represent my interpretations of a natural bone edge to a broken bone edge. |
 |
| Yuezhuornis skeleton (top to bottom, scaled basically the same) Xu et al.'s (2026) tracing from Figure 1b; specimen in situ from Xu et al.'s Figure 1a with my tracings over it; my tracing layer with identified elements shaded gray. Black > blue > red outlines represent my interpretations of a natural bone edge to a broken bone edge. *NOTE I didn't see the pedal digits on the upper right until after compositing this image, so those are missing from the bottom tracing. |
1. Explicit statements by the authors that I think are incorrect.
Autapomorphic jugal morphology
Well the funniest thing is, you'll recall I said the jugal was the one cranial element accurately represented in their skeletal reconstruction? Supposedly diagnostic, Xu et al. include in their diagnosis "slender jugal that has a reduced quadratojugal process and a blunt-ended postorbital process" as an autapomorphy. They get a lot of runway from this element, saying (numbered refs excluded, as they are for the rest of this post) "The rod-like jugal tapers anteriorly, and the suborbital bar is markedly more slender than in anchiornithines and non-avialan theropods (Extended Data Fig. 3c–h). The postorbital process of the jugal ends bluntly (Extended Data Fig. 3c), contrasting with the tapering form as in deinonychosaurs. As in anchiornithines and deinonychosaurs, the postorbital process is directed posterodorsally, rather than being oriented dorsally as in other early avialans. Similar to some Cretaceous avialans, the quadratojugal process of the jugal is reduced, contrasting with the prominent form seen in Archaeopteryx, anchiornithines and most non-avialan paravians."
Well I'm embarrassed to not notice this until I traced the skeleton, but there's a reason it's weird- it's a first dorsal rib, NOT a jugal. This element is remarkably conservative in at least avetheropods, the capitulum ('dorsal process') is expanded distally and usually angled toward the tuberculum ('posterior process') which is short and bluntly rounded. The distal ('anterior') tip tapers to a point (since it doesn't articulate with a sternal rib). Here are first dorsal ribs from some classic monographed avetheropods compared to Yuezhuornis-
 |
| First dorsal rib of (top to bottom) Yuezhuornis (from Xu et al., 2026 Extended Data Figure 3a), Deinonychus (from Ostrom, 1969: Figure 51A), Tyrannosaurus (from Brochu, 2003: Figure 62M), and Allosaurus (from Madsen, 1976: Plate 39A). |
 |
Most complete dorsal ribs of Yuezhuornis (some flipped to be in the same view), extracted from Xu et al.'s (2026) Figure 1a. Supposed jugal on far left is actually the first dorsal rib. |
Xu et al. say "The premaxillary teeth are smaller than the anterior maxillary teeth—a feature present in some dromaeosaurids and troodontids but absent in early avialans. As in basal troodontids and Anchiornis, but not in dromaeosaurids and early avialans such as Archaeopteryx, the anterior maxillary teeth are packed more closely and larger than the posterior teeth." As the authors correctly show in their tracing, the anterior maxillary teeth are not preserved, though fragmentary bits of the anterior maxilla itself do seem to exist. I estimate about five teeth are missing anterior to the first preserved tooth, so that about fifteen maxillary teeth were originally present. So what's actually true is the middle maxillary teeth are larger than the posterior maxillary teeth, as is typical in theropods including dromaeosaurids. As for spacing, there are two gaps between the ~11th and ~13th and the ~13th and ~15th tooth in the left maxilla, but this could easily be due to shed teeth instead of absent alveoli. The rest of the posterior teeth are similar in spacing to Sinovenator, contra their statement "Unlike troodontids, but as in early avialans such as Archaeopteryx, the middle and posterior maxillary teeth are spaced sparsely." The reconstruction is fictional as usual with seven teeth and a tooth row shorter than they trace for their eight-toothed right maxilla. Having fewer teeth spaced more sparsely as in their reconstruction makes it appear more bird-like.
Anterior lacrimal process much shorter than posterior process
The authors then say "Like Jeholornis and some enantiornithines, the triradiate lacrimal has an anterior process that is much shorter than the posterior process (Fig. 2c), but the opposite is true in Archaeopteryx and non-avialan paravians." The actual specimen has an anterior process 110% the length of the posterior process. Their tracing has a ratio of ~71%, and their reconstruction gives the lacrimal a strongly obtuse angle between the ventral and posterior processes that does not exist. That angle would imply a more triangular snout, which would be more bird-like.
Postorbital ventral and posterior processes rod-like
Xu et al. state "Like early avialans, the T-shaped postorbital has slender, rod-like processes (Fig. 2a,c), contrasting with the stout, strap-like condition in non-avialan theropods." However, the posterior and ventral processes are triangular, contra their tracing for the latter. Their reconstruction makes up a new tiny postorbital, ironically with robust ventral and posterior processes like the real one. Nonetheless, a tiny postorbital is more bird-like than the real thing.
Dentary lateral groove absent
The authors claim "A deep lateral dentary groove, which is present characteristically in anchiornithines and deinonychosaurs, is absent (Fig. 2), as in Archaeopteryx and more crownward taxa." This groove is very obvious in both dentaries, and these aren't Meckelian grooves, as teeth and position show the orientation of the dentaries and Meckelian grooves are located more ventrally in paravians (e.g. Troodon- Currie, 1987: Fig. 1c; Archaeopteryx- Elzanowski and Wellnhofer, 1996: Fig. 12).
Yuezhuornis dentaries in lateral view (top- right, flipped; bottom- left) showing longitudinal grooves (extracted from Xu et al.'s 2026 Extended Data Figure 3a).
V-shaped furcula with interclavicular angle of 110°
Xu et al. state "The V-shaped furcula has an interclavicular angle of 110° (Fig. 2d), comparable with that of Anchiornis and non-avialan paravians, but this angle is acute in other avialans." This is based on the preserved median part of the furcula, but neither end seems complete and I'd estimate given its curvature the real angle would be about 95°. More importantly (since angle can vary this much even within a genus), the furcula is clearly U-shaped even in their own tracing, not V-shaped like ornithothoracines. Note Sapeornis has an obtuse interclavicular angle (e.g. 96°, 104°), so Yuezhuornis' angle is known in some basal avialans.
Trapezoidal coracoid
The authors write "The coracoid is trapezoidal (Fig. 2d)—an ancestral condition for Paraves", but their own accurate tracings in both Figures 1b and 2d show it's not the case. A trapezoid is of course defined by one side being longer than its opposite, which is the distal (sternal articulating) edge in paravians compared to the proximal (scapula articulating) edge. The classic example is Deinonychus, contrasting with the 'semicircular' coracoids of most non-maniraptorans. This morphology for Yuezhuornis is unusual for paravians and makes it a LOT less bird-like. Their reconstruction gives it a tall Archaeopteryx-like coracoid.
Radius thick compared to ulna
Classic Figure 2 from Ostrom 1974 in part, with Yuezhuornis coracoid added at right from Xu et al.'s (2026) Figure 2d. Note it is NOT trapezoidal.
Xu et al. claim "The robust radius measure around 90% of the width of the ulna, comparable with that of basal theropods and Anchiornis; whereas the bone is relatively slender in other paravians", but this is because the right elements are in anterior view. The left elements in lateral view show a ratio of ~70%, similar to e.g. Archaeopteryx, Rahonavis or Bambiraptor. Their reconstruction gives it a thin radius (~46% of ulnar width), so I guess the artist didn't bother with the authors' interpretation. Having a thin radius is bird-like, by the way.
Preacetabular process markedly longer than postacetabular process
The authors say "The preacetabular process is markedly longer than the postacetabular process—a derived condition for avialans—whereas the two processes are subequal in Anchiornis and most non-avialan theropods." But even using their own Extended Data Figure 5 meant to compare ilia, their basically accurate tracing shows Yuezhuornis falls out in the non-avialan category. It's also funny that they traced Adasaurus from the very inaccurate drawing in Barsbold (1983) instead of Turner et al.'s (2012) photo that's been available for fourteen years. Needless to say, a shorter preacetabular process is less bird-like. Oh, and the reconstruction's ilium is fictional with a longer preacetabular process (~116%), so is again more birdy.
Ilia of paravians modified from Extended Data Figure 5 of Xu et al. (2026). Blue lines represent pre- and postacetabular lengths, with the number above being the pre/post - acetabular ratio. At bottom is Adasaurus' actual ilium (from Turner et al., 2012).
Metatarsal IV as long as metatarsal III
Xu et al. claim "Like enantiornithine longipterygids, metatarsal IV extends as far distally as metatarsal III, and metatarsal II is much shorter (Extended Data Fig. 6c)." This simply isn't true, as seen in their Extended Data Figure 6d. Metatarsal III is longest as in almost all Mesozoic theropods.
 |
| Left metatarsus of Yuezhuornis in posterior view showing metatarsal III extending further distally than IV (red lines) and potential for a ginglymoid metatarsal II (blue lines) (modified from Xu et al., 2026: Extended Data Figure 6d). |
2. Explicit statements by the authors I think can't be demonstrated.
Teeth unserrated mesially
The authors say "The maxillary teeth are serrated distally but not mesially", but their Extended Data Figure 3b shows these teeth are poorly preserved with serrations only observable on parts of distal carinae of three teeth. Mesial serrations are often confined to apical areas in deinonychosaurs and potentially smaller than distal ones, so their absence could easily be preservational. Lacking mesial serrations would make the taxon more bird-like, of course.
 |
| Maxillary teeth of Yuezhuornis in lateral view (extracted from Extended Data Figure 3b of Xu et al., 2026). |
Lacrimal ventral process short
Xu et al. say "The jugal process of the lacrimal is slightly longer than the posterior process, contrasting with the large difference seen in other early avialans and some troodontids", but the ventral (jugal) process is clearly overlapped distally by the maxilla (then potentially the matrix). It would be almost unheard of to have a lacrimal not reach the maxilla/jugal in a non-ornithothoracine, and even their extra birdy reconstruction shows a ventral process 60% longer than the posterior process. It doesn't connect the lacrimal to the maxilla/jugal though, completely without evidence and making it look birdier than evidence suggests.
Distal caudals have long postzygapophyses and prezygapophyses that don't extend beyond centra
Xu et al. say of mid/distal caudals, "The prezygapophyses terminate level with the anterior articular facet, whereas postzygapophyses project beyond the posterior margin of the centra." There are potentially seven elongate caudals with preserved zygapophyses, one of which is poorly preserved in white (like the jugal and palatine) and not not shown in a magnified figure, so won't be considered here. I've labeled the others A-F below. Of these, I think D is more likely to be a pedal phalanx based on the strongly concave and convex ends and association with the pes, though it could be a very broken caudal. Because the tail is disarticulated and the bones are poorly preserved, the anteroposterior orientation of each vertebra can't be determined. Paravians can show the condition interpreted by Xu et al. (e..g. 10th caudal of Rahonavis) or the opposite one (e.g. 11th caudal of Sinornithoides). They can even have both conditions in different parts of the tail, with e. g. Archaeopteryx (Eichstatt specimen) having longer postzygapophyses in caudals 4-16, then longer prezygapophyses in caudals 17-21, or Buitreraptor (holotype) having longer postzygapophyses in caudals 7-9 but longer prezygapophyses in some more distal caudals. Only caudal F of Yuezhuornis even preserves both zygapophyses, with one much longer than the other. So really the most we can say is that it doesn't have serraraptorian-like hyperelongate zygapophyses. Their reconstruction gives every elongate caudal short postzygapophyses and even shorter prezygapophyses, even though caudals A and B clearly have at least one zygapophysis longer than a third of the centrum.
 |
| Elongate caudals of Yuezhuornis to scale, rotated to be horizontal with ventral side down (extracted from Xu et al.'s 2026 Extended Data Figure 6a). Blue marks the ends of the centra, while red marks the ends of the zygapophyses. |
Coracoid with pitted external texture
The authors claim "The anterior surface of the coracoid is sculped by numerous pits—a feature observed among paravians only in Anchiornis." The external surface of the coracoid (and most other bones) is unpreserved, with only a broken internal section exposed. Compare its texture to other adjacent elements in that figure.
 |
| Coracoid (co at right) of Yuezhuornis in situ showing similar rough pitted texture caused by the exposure of bone cross sections (extracted from Figure 2d of Xu et al., 2026). |
Humerus with small deltopectoral crest
This one's simple. Xu et al. write "The deltopectoral crest is small", but the right humerus is preserved in posterior view with the deltopectoral crest projecting unseen away from us into the sediment. The left humerus seems to be in lateral view (contra the authors' tracing), but even if what I think is a proximal fragment is correctly identified, who knows if the anterior edge is complete.
Interpubic fenestra absent
The authors claim "A slit-like opening on the apron, which is present in basal dromaeosaurids and troodontids (for example, Microraptor and Sinovenator), is absent (Fig. 2g)." However there certainly seems to be an area of matrix where the fenestra would be expected, although the apron proximal to this is poorly preserved. The left pubis does seem to extend medially over it, but whether it contacted the right pubis cannot be said. Contra the authors, Sinovenator lacks an interpubic fenestra (Xu, 2002- Fig. 44-45), as do some deinonychosaurs like Anchiornis (PKUP V1068), the Archaeopteryx type, Jianianhualong, Buitreraptor and Unenlagia. Regardless of its actual distribution, Xu et al. write "The avialan affinity of Yuezhuornis was supported by ... extensive pubic apron without a slit ..."
 |
| Pubes of Yuezhuornis in anterior or posterior view (extracted from Figure 2g of Xu et al., 2026). Red line points to possible interpubic fenestra. |
Fibula doesn't reach tarsus
Xu et al. say "The fibula terminates far proximally to the distal end of the tibia (Fig. 2i and Extended Data Fig. 6)—a derived feature absent in most non-avialan paravians, anchiornithines and Archaeopteryx, but seen widely in Cretaceous avialans." This character actually gets rather messy in taxa like Troodon and Sapeornis, but in Yuezhuornis the authors devote an entire figure to demonstrate this (Extended Data Figure 6c) where "the blue arrowhead denotes the distal end of the fibula." But if you look at the more complete Extended Data Figure 6b, extending distal to that is a hemicylindrical concavity just like the more proximal fibula. Distal to the slab break, the hindlimb is shifted medially so that what's labeled 'ti' certainly seems like a separate fibula continuing. After the next break things are too fragmented to tell. I do question the decision to include Extended Data Figure 6c if it's only going to be 22% larger than 6b. If your intent is to demonstrate that's the end of the fibula, make it twice the magnification at least! In any case the left fibula also extends past Xu et al.'s supposed distal tip while being thicker, supporting the idea that the apparent thinness of the right fibula at mid-shaft is caused by the cross section being closer to the far edge of the cylindrical shaft. Again, things become too fragmentary towards the distal end, so we can't know if Yuezhuornis' fibula contacted the tarsus.
 |
| Left tibia and fibula of Yuezhuornis showing fibula continuing distally (red lines) and displacement of slab sections (blue lines). Extracted from Xu et al.'s (2026) Extended Data Figures 6b and 6c. |
The authors claim "The metatarsals are not ginglymoid (no sign of intertrochlear incisures)—a unique condition unknown among early paravians, in which at least one metatarsal is ginglymoid." But while mtIV is convex and mtIII only weakly concave distally, mtII is incomplete distally. What could easily be a convex and prominent lateral condyle is preserved if that is the actual distal end to the bone, so that a medial condyle clould easily make the element ginglymoid (see above figure under metatarsal length). Contra the authors, non-ginglymoid metatarsals are common in early paravians including Serikornis, Anchiornis, Eosinopteryx and Archaeopteryx.
Metatarsal V short and straight
Xu et al. say "The vestigial metatarsal V is straight", but I didn't notice one. Nor is one indicated in their tracing.
3. Their diagnosis
Xu et al. list several characters in their diagnosis.
- "closely packed premaxillary teeth that are smaller than anterior maxillary teeth", yet as noted above the anterior maxillary teeth are unpreserved, with the anteriormost preserved being around the sixth. The closely packed premaxillary teeth are similar to e.g. Anchiornis (Wang et al., 2025).
- "maxillary teeth that are serrated distally and not constricted below the crowns*", is true but would be expected for a serraraptorian.
- "slender jugal that has a reduced quadratojugal process and a blunt-ended postorbital process*", is based on the first dorsal rib as noted above, with the actual jugal less slender, with a longer quadratojugal process and tapered postorbital process.
- "trapezoidal coracoid with an anterior surface sculped [sic] by numerous pits." False and unknown respectively, as shown above.
- "furcula with an interclavicular angle of 110°", which is debated above. It doesn't really matter though because plenty of taxa have angles similar to both (e.g. Sinornithosaurus 108°; Anchiornis 110°; Microraptor 82-108°).
- "humerus shorter than femur" is true of all non-avialan paravians except Archaeopteryx, Fujianvenator, some Anchiornis (e.g. LPM B00169) and probably Rahonavis.
- "penultimate manual phalanges that have enlarged ventral heels proximally*", is only preserved in III-3. I haven't examined this extensively, but Anchiornis can show the same condition (e.g. BMNHC PH822 in Figure 30 of Pei et al., 2017). And the authors mention Buitreraptor shows it.
 |
| Manual phalanx III-3 of Anchiornis (left; after Figure 30 of Pei et al., 2017) and Yuezhuornis (right; after Extended Data Figure 4b of Xu et al., 2026). Red line points to supposedly autapomorphic enlarged proximoventral heel. |
- "pubic apron mediolaterally broad and imperforated." It's narrower (width of pubes <23% of length) than e.g. Anchiornis (34% in PKUP V1068), Archaeopteryx (~47% in the London specimen), Microraptor (~37% in CAGS 20-8-001) and Sinovenator (37%). Maybe it's ironically diagnostically narrow? See above for the ambiguous interpubic fenestra.
- "fibula that ends far proximal to the distal end of tibia", which is not necessarily true as demonstrated above.
- "metatarsals II–IV not ginglymoid." As shown above, the condition of metatarsal II is at best uncertain.
- "metatarsal II more robust than other metatarsals". The metatarsus is in posterior view, so the actual width of metatarsal III is not certain compared to its posterior exposure, but both Graciliraptor and Overoraptor also have distal ends of metatarsal II >130% wider than metatarsal III.
- "pedal phalanx II-1 (or II-2) that has collateral ligament fossa placed dorsally and large proximoventral heel." This is very obviously pedal phalanx II-2, and the same morphology is present in most deinonychosaurs and Balaur.
- "forelimb one-third shorter than hind limb." Does this include phalanges, and regardless it's of course correlated with the humerofemoral ratio used above. Without wasting a lot of time calculating measurements, Yuezhuornis is clearly not anywhere unique among paravians in having a supposed ~66% forelimb/hindlimb ratio. Plus it doesn't even have a complete metacarpus to compare, so at best we're judging humerus+radius to femur+tibia.
4. What else?
There's a ton more I could say. The premaxilla is clearly tall like eudromaeosaurs; the dorsally located maxillary fenestra is dromaeosaurid-like; the upcurved dentary is dromaeosaurid-like; there seem to be two preserved stapes; with one dorsal and no sacrals, I doubt we can exclude a long caudal series like eudromaeosaurs; the one well preserved proximal caudal is in lateral view not dorsal/ventral view like the authors' tracing; the supposed metacarpal III is more likely a bowed phalanx I-1; the long bones on the far right are clearly not just the left metatarsals III and IV, as there are four of them, so maybe the ischia are involved there; the right astragalus is potentially preserved in lateral view by the distal right tibia, with the calcaneum a third down the metatarsus and distal tarsals by the proximal metatarsus; in the far upper right are pedal unguals III and IV along with part of at least one of their digits and claw sheaths not figured by the authors, so digit II was definitely hypertrophied. Etc., etc....
After making my last post, I felt a little bad for saying "when reading the next report of the birdiest bird-like thing from the Nanyuan Formation, DON'T TRUST IT!" But now after really looking in detail at Yuezhuornis - OOF. I really didn't think they could get things this wrong. I want to go back and really examine the Baminornis, Zhengheornis and Fujianvenator figures to see what declarations I took for granted. Maybe a skull was actually an ilium?!
In any case, I hope my mentor Anonymous #X is partially satisfied with this detailed exposé. But we have one step left when beating this dead horse even deader- "a feature-by-feature comparison: ... which character states the authors code differently." So NEXT TIME, I'll score Yuezhuornis for my own analysis to see where it goes AND ALSO check Xu et al.'s scorings in the TWiG analysis to see how accurate they are. I haven't checked this team's scorings for any of their Nanyuan taxa, so your guess is as good as mine...
References- Ostrom, 1969. Osteology of Deinonychus antirrhopus, an unusual theropod from the Lower Cretaceous of Montana. Bulletin of the Peabody Museum of Natural History. 30, 165 pp.
Ostrom, 1974. The pectoral girdle and forelimb function of Deinonychus (Reptilia: Saurischia): A correction. Postilla. 165, 11 pp.
Madsen, 1976. Allosaurus fragilis: A revised osteology. Utah Geological and Mineral Survey Bulletin. 109, 1-163.
Barsbold, 1983. Carnivorous dinosaurs from the Cretaceous of Mongolia. Transactions of the Joint Soviet-Mongolian Palaeontological Expedition. 19, 117 pp.
Currie, 1987. Bird-like characteristics of the jaws and teeth of troodontid theropods (Dinosauria, Saurischia). Journal of Vertebrate Paleontology. 7(1), 72-81.
Elzanowski and Wellnhofer, 1996. Cranial morphology of Archaeopteryx: Evidence from the seventh skeleton. Journal of Vertebrate Paleontology. 16(1), 81-94.
Xu, 2002. Deinonychosaurian fossils from the Jehol Group of western Liaoning and the coelurosaurian evolution. PhD Thesis. Chinese Academy of Sciences. 325 pp.
Brochu, 2003. Osteology of Tyrannosaurus rex: Insights from a nearly complete skeleton and high-resolution computed tomographic analysis of the skull. Society of Vertebrate Paleontology Memior. 7, 138 pp.
Turner, Makovicky and Norell, 2012. A review of dromaeosaurid systematics and paravian phylogeny. Bulletin of the American Museum of Natural History. 371, 1-206.
Pei, Li, Meng, Norell and Gao, 2017. New specimens of Anchiornis huxleyi (Theropoda: Paraves) from the Late Jurassic of northeastern China. Bulletin of the American Museum of Natural History. 411, 66 pp.
Wang, Wang, Zheng and Zhou, 2025 (2024 online). Cranial anatomy of Anchiornis huxleyi (Theropoda: Paraves) sheds new light on bird skull evolution. Vertebrata PalAsiatica. 63(1), 20-42
Xu, Wang, Lin, Dong, Deng, Xu, Du, Lin, Chen and Zhou, 2026. Jurassic flightless bird reveals early ecological diversity of avialans. Nature. 658(8136), 690-697.
