No pygostyle for Baminornis

This week's big news was Baminornis (Chen et al., 2025), a bird with a supposed large pygostyle from a well dated Tithonian locality (Nanyuan Formation- in the same layer Fujianvenator is from, which emerges sister to Serikornis in the Lori matrix btw). This supposed pygostyle is at least 74% of ilial length, so is not like the comparatively undeveloped fused structures in Deinocheirus, Beipiaosaurus and some oviraptorosaurs. Instead, it's out of place in the Jurassic and out of place phylogenetically given the authors recover Baminornis as just crownward of Archaeopteryx in both the TWiG and O'Connor matrices, stemward of Jeholornis in both. Chen et al. call this "The observed ‘chaos’ of bird tail evolution" and note the Late Jurassic Solnhofen and Yanliao Biotas lack short-tailed birds, thus concluding "It would be intriguing to explore why no short-tailed avialans have been found from these two localities."

The obvious answer is- that structure in Baminornis isn't a pygostyle. It's pretty obviously a synsacrum. The first thing to note is that despite the clean, rounded end drawn in their Extended Data Fig. 1b (see below), the posterior end is clearly broken in the photo Extended Data Fig. 1a above it, so it may have included six or more vertebrae in life. This isn't an argument for either identification, as Archaeopteryx can have five (most specimens) or six (Thermopolis specimen), while Jeholornis has six. In any case, the arguments for identification as a synsacrum are-

1. Size. As noted above, even with its broken posterior end, the structure in Baminornis is 74% of ilial length despite only including five vertebrae. As Chen et al. say "As in confuciusornithids, the pygostyle is longer than metacarpal II, but the opposite is true in most other early avialans such as Sapeornis and Cratonavis" yet confuciusornithids fuse eight-ten vertebrae to make their longer pygostyle and enentiornithines usually fuse more as well (e.g. Iberomesornis, IVPP V15664) (Rashid et al., 2018). Excluding the very different elongated distal caudals of basal paravians, bird caudals are short, so fusing so few together wouldn't result in such a large structure. Using Wang and O'Connor's (2017) Table 1, the pygostyle/femoral length ratio of Sapeornis with 4 fused caudals is 19-29%, Jehol euornithines' with 3-5 fused caudals is 19-35%, confuciusornithids' is 49-71%, and non-pengornithid Jehol enantiornithines' is 39-80%. Baminornis' ratio is >44%.

2. Curvature. Chen et al. state the bone "curves dorsally—reminiscent of, but to a lesser extent than, the derived condition in ornithuromorphs in which the pygostyle is characteristically plough-shaped. By contrast, the bone is straight in other early avialans such as confuciusornithids and enantiornithines." It actually looks more curved than their figured 'ornithuromorphs' (= euornithines) Bellulornis and 'Abitusavis' (= Yanornis), but in any case I reinterpret the concave side to be ventral, which matches many paravians from Saurornitholestes to Ornithodesmus to Zhyraornis. It's usually not obvious in Lagerstatten birds where the synsacrum is often preserved in dorsoventral view or below the ilium in lateral view. It's of course possible that a basal avialan pygostyle could be curved either dorsally or ventrally, but it definitely matches the listed paravian synsacra more than any pygostyle I can think of.

Top row- upside down synsacra in right lateral view of (left) Zhyraornis kashkarovi holotype TsNIGRI 42/11915 (after Nessov, 1992); (center) Ornithodesmus cluniculus holotype NHMUK R187 (after Howse and Milner, 1993); (right) Saurornitholestes langstoni (?) NHMUK R4463 (after Howse and Milner, 1993). Bottom rows- (a, b) photo and drawing of Baminornis zhenghensis holotype IVPP V33259, note the posterior end in the photo is broken unlike the drawing; tails of- (c) Confuciusornis sp. IVPP V16066; (d) Jinguofortis perplexus holotype IVPP V24194; (e) Jeholornis prima IVPP V13353; (f) Parabohaiornis martini paratype IVPP V18690; (g) Pterygornis dapingfangensis holotype IVPP V20729; (h) Yanornis martini (Abitusavis lii holotype) IVPP V14606); (i) Bellulornis rectusunguis holotype IVPP V17970 (after Chen et al., 2025).

3. Supposed proximoventral processes. The authors say "The pygostyle has a pair of proximally distributed ventrolateral processes which terminate posterior to the proximal articular facet", but they are asymmetrical with the supposed left one being a small prong while the supposed right one is a more posteriorly placed and ventrally projected, longer blade. Taphonomy maybe, but you know what these perfectly match if the structure is flipped dorsoventrally? A prezygapophysis and a neural spine lamina, respectively. Chen et al. note Sapeornis lacks these processes and obviously Jeholornis doesn't have them, so it would be less parsimonious if Baminornis converged with pygostylians in developing them (or they were lost in Sapeornis) anyway.

4. Anteroposteriorly concave vertebral edges. At least the first, third and fourth vertebrae have concave edges along the concave side. This matches the ventral edges of each centrum in most archosaurian vertebrae, while the dorsal edge of pygostyles is either smooth or convex at each vertebra to reflect the neural spines (see Chen et al.'s euornithine examples, or Fukuipteryx).

5. Taphonomic position. Under my interpretation, you only have to rotate the synsacrum 90 degrees instead of also flipping it upside down, plus the only well-preserved free caudal is next to its posterior end. These aren't in themselves good arguments, since the right pectoral girdle, ischium etc. obviously suffered more displacement than just rotation, but they do match this hypothesis better. Similarly, the synsacrum would be missing otherwise, and as a centrally placed element we would expect it to stay close to the skeleton more than the distal tail.

6. Parsimony, as mentioned above. A pygostyle in an early-diverging avialan like this introduces homoplasy, especially such an elongate one with proximoventral processes like confuciusornithids and most enantiornithines, unlike the shorter and simpler form in Sapeornis and basal euornithines. Whereas a synsacrum is exactly expected.

Once you see the synsacrum in Baminornis, it's pretty hard to un-see. The details all make sense, like the simple and smooth centra and the more complicated topology where the sacral ribs are. Indeed, I'd say if it were a pygostyle my orientation would still make more sense with the centra, prezygapophysis and neural spine lamina, but the vertebrae would still be too big, and why wouldn't it be a synsacrum at that point since it looks just like one and a pygostyle in this taxon is unexpected. Another case of supposed giant Nature falling for sensationalism *cough Oculudentavis cough*.

References- Nessov, 1992. Mesozoic and Paleogene birds of the USSR and their paleoenvironments. In Campbell (ed.). Papers in Avian Paleontology Honoring Pierce Brodkorb. Natural History Museum of Los Angeles County Science Series. 36, 465-478.

Howse and Milner, 1993. Ornithodesmus - a maniraptoran theropod dinosaur from the Lower Cretaceous of the Isle of Wight, England. Palaeontology. 36, 425-437.

Wang and O'Connor, 2017. Morphological coevolution of the pygostyle and tail feathers in Early Cretaceous birds. Vertebrata PalAsiatica. 55(4), 289-314.

Rashid, Surya, Chiappe, Carroll, Garrett, Varghese, Bailleul, O’Connor, Chapman and Horner, 2018. Avian tail ontogeny, pygostyle formation, and interpretation of juvenile Mesozoic specimens. Scientific Reports. 8:9014.

Chen, Wang, Dong, Zhou, Xu, Deng, Xu, Zhang, Wang, Du, Lin, Lin and Zhou, 2025. Earliest short-tailed bird from the Late Jurassic of China. Nature. 638(8050), 441-448.