What's the Deal with Enantiornithean Tails?

You are drawing a scene set in the Cretaceous and you need some birds in the picture, maybe as a way to show off the size of some much larger, more charismatic dinosaur. The setting is not oceanic enough to use something vaguely Ichthyornis-like, so what do you do? You draw in a few enantiornitheans, of course.

That's a good choice. Enantiornitheans, or "opposite birds", were probably the most diverse and abundant bird-like dinosaurs during the Cretaceous. They would have looked quite similar to modern birds and were probably doing many of the same things that modern birds do. However, recent findings have also given us insights into how enantiornitheans differed from modern birds, not only in the details of their skeletons but also in their external appearance and behavior. This means that scientifically rigorous depictions of enantiornitheans should not show them simply as modern birds transplanted into a different time period. And probably the most noticeable difference between most enantiornitheans and modern birds would have been the feathering on their tail.

(You might have noticed that I've switched to using "enantiornitheans" instead of the more commonly used "enantiornithines". As has been noted by Harris et al., 2006 and Matt Martyniuk, the suffix "-ine" is more typically used for the vernacular names of clades that end in "-inae", and Enantiornithes is not one of those clades. Seeing as Harris et al. suggested using "enantiornitheans", that's what I'm going with for now.)

Restoration of the enantiornithean Dapingfangornis, by Jack Wood (used with permission).

Unlike most other dinosaurs, the bony part of a modern bird's tail is very short, even in birds that have long tail feathers. Despite this, the tail serves important functions in most birds. The last few tail vertebrae are fused together into a structure called the pygostyle, which supports a pair of muscular bulbs in which the flight feathers on the tail are embedded. These bulbs are known as rectricial bulbs, and they are responsible for the ability to spread the tail feathers out like a fan. This allows the tail feathers to act as an additional lift-generating surface, which among other things grants finer control during low-speed flight, such as while landing.

A white-necked jacobin showing off its spread tail fan, photographed by Paul Tavares, under CC BY-NC 4.0.

Like modern birds, enantiornitheans had short tails with a pygostyle. As a result, they have traditionally been depicted with tail fans similar to those of modern birds. In fact, several early papers about enantiornitheans outright suggested that they likely had tail fans based on their possession of a pygostyle. However, by the mid-2000s, a good number of enantiornithean fossils had been described from the Jehol Group in China, famous for its preservation of soft tissue structures, including feathers.

How many of these enantiornitheans have been found with tail fans? None.

Julia Clarke and colleagues pointed this out in a 2006 paper describing the anatomy of Yixianornis, an Early Cretaceous avialan from the Jehol that was more closely related to modern birds than enantiornitheans were. Whereas the holotype of Yixianornis was preserved with a tail fan similar to those of modern birds, all enantiornitheans preserved with tail feathering that had been described up to that point either lacked large tail feathers entirely (leaving only a fuzzy stump of a tail) or had just a single pair of long tail feathers that wouldn't have provided much of a lift-generating surface.

Clarke et al. also highlighted the fact that (as had been noted by previous researchers) the pygostyle of enantiornitheans differed from those of Yixianornis and modern birds. Enantiornitheans had a longer, more rod-like pygostyle, whereas Yixianornis and modern birds have a shorter, more tab-like one. Thus, Clarke et al. suggested that the presence of a pygostyle alone does not indicate tail feathering similar to that of modern birds. Instead, it is specifically the short, tab-like pygostyle that correlates with a tail fan.

Comparison of the last few tail vertebrae among different maniraptor groups, mapped onto their phylogenetic relationships, from Wang and O'Connor (2017). In most long-tailed maniraptors (e.g.: Archaeopteryx, dromaeosaurids), the vertebrae are not fused into a pygostyle, but some oviraptorosaurs and therizinosaurs have evolved fused tail vertebrae independently of short-tailed avialans. Note that members of Ornithuromorpha (a group that includes modern birds) have a short, tab-shaped pygostyle.

If we had only a few fossils to base these conclusions on, these observations might not mean much. Even the Jehol Group rarely preserves a complete record of plumage in any individual specimen. However, dozens if not hundreds of feather-preserving Mesozoic avialan fossils have now been studied, and for most part, the pattern observed by Clarke et al. has held up under subsequent discoveries. So far, all Mesozoic avialans confirmed to have had a modern-style tail fan are species more closely related to modern birds than enantiornitheans were, with short, stumpy pygostyles.

Among enantiornitheans, on the other hand, the most commonly seen condition is the absence of any large tail feathers, which has been documented in Eoenantiornis, Longipteryx, Cruralispennia, and Avimaia. Enantiornithean genera that have been found with a single pair of long tail feathers include Protopteryx, Dapingfangornis, Bohaiornis, Parapengornis, Eopengornis, Cratoavis, Junornis, and Orienantius. Longirostravis was described as preserving these long tail feathers as well, though Mesozoic avialan specialist Jingmai O'Connor is reportedly doubtful of this interpretation. There are even very small juvenile enantiornithean specimens that apparently show a pair of tail feathers that had begun to grow, as seen in one of the amber-preserved enantiornitheans. (Yeah, we have amber-preserved enantiornitheans.)

Two enantiornithean specimens preserved with feathers, from O'Connor et al. (2012). Protopteryx (A) preserves a single pair of streamer-like tail feathers, whereas Eoenantiornis (B) preserves no large tail feathers at all. (C) shows isolated body feathers of Protopteryx, which are not especially relevant to this post.

These forms of tail plumage are similar to that of confuciusornithiforms, which were more distant relatives of modern birds compared to enantiornitheans. It's perhaps no accident that, like enantiornitheans, confuciusornithiforms had rod-shaped pygostyles. Currently, we don't have enough specimens described of any enantiornithean to determine whether long tail feathers were only present in some individuals of the same species, as has been documented in Confuciusornis and Eoconfuciusornis. At least in the case of confuciusornithiforms, it appears likely that the individuals with the long tail feathers were males.

The long tail feathers in enantiornitheans and confuciusornithiforms have been described as being ribbon-like or streamer-like. They are known by the technical names Rachis-Dominated Feathers (RDFs) or Proximally Ribbon-like Pennaceous Feathers (PRPFs). Recent findings of such feathers preserved in amber have been very informative regarding their detailed structure, which had been difficult to interpret from flattened Jehol-style fossils. It turns out that the central shaft of RDFs is an extremely thin sheet that is less than 50 micrometers thick. The underside of the shaft is open such that it forms an arch, unlike the enclosed tube seen in typical feathers.

Examples of RDFs preserved in amber, from Xing et al. (2018). (D) is a schematic showing the structure of RDFs, as inferred from these specimens.

Several groups of modern birds, from hummingbirds to parrots, have evolved superficially similar, streamer-like feathers, but none of these plumes exhibits the same type of structure seen in RDFs. It appears that RDFs are an entirely extinct form of feather that modern birds have "forgotten" how to make. Like most other types of feathers, RDFs have barbs coming off the central shaft. In some enantiornitheans, such as Eopengornis and Parapengornis, the barbs are present along almost the entire length of the tail feather, but in many others the barbs are restricted towards the feather tip.

So enantiornitheans either lacked flight feathers on the tail or had a single pair of ribbon-like feathers not found in any modern bird. Simple, right? Yet, as happens so often in science, some new discoveries have arisen that complicate the picture. For starters, we now also know of enantiornitheans that had two pairs of RDFs on the tail, as first discovered in Paraprotopteryx.

Furthermore, a few enantiornitheans described since 2006 have been purported to challenge the notion that enantiornitheans lacked tail fans. The first of these was Shanweiniao, described in 2009 by O'Connor and colleagues. Shanweiniao was found with at least four (possibly six) tail feathers closely aligned with one another, as though forming a tail fan. It was even named for this feature, Shanweiniao meaning "fan-tailed bird" in Chinese.

However, these feathers were not very well preserved, so their structure and arrangement are difficult to confirm. In a 2016 paper describing another enantiornithean, Chiappeavis (which will factor into this story shortly), O'Connor et al. reevaluated the evidence for a tail fan in Shanweiniao. They noted that there were narrow gaps between its tail feathers, unlike what was preserved in avialans with more definite tail fans. Therefore, they concluded, it was more likely that the tail feathering of Shanweiniao consisted of at least two pairs of RDFs, similar to Paraprotopteryx. Shanweiniao may have had more tail feathers than the average enantiornithean, but there is little reason to think that it had a modern-style tail fan.

The tail feathers of Shanweiniao, from O'Connor et al. (2009).

The next new find that seemed to question the dearth of enantiornithean tail fans was Feitianius, described in 2016. Feitianius was found with not just two or even four large tail feathers, but many. It had a single pair of RDFs, but it was also preserved with 5-7 shorter feathers that appear to have been barbed for only about half their length, as well as at least 5 more tail feathers that were shorter still. In addition, an organic mass preserved around the pygostyle was interpreted as the remnants of the soft tissues that include rectricial bulbs.

In 2017, Wang Wei and O'Connor put forth a more detailed study on the relationship between pygostyle form, tail musculature, and tail feathers. They found that it was unlikely for any enantiornithean pygostyle to have supported well-developed rectricial bulbs. However, they did find features of the pygostyle in enantiornitheans (and confuciusornithiforms) that correlate with strong muscles for raising the tail, and suggested that the soft tissues surrounding the pygostyle in Feitianius may instead include these muscles instead of rectricial bulbs. Thus, Feitianius was likely well adapted to raising its elaborate set of tail feathers (perhaps in display), but not for spreading them apart as a true tail fan.

The tail of Feitianius with preserved soft tissues, from O'Connor et al. (2016a).

Only a day after the paper naming Feitianius was posted online, along came the aforementioned description of Chiappeavis. The preserved tail feathers of Chiappeavis really do resemble a tail fan. There are no visible gaps between the feathers as in Shanweiniao, and though their detailed structure is poorly preserved, they appear to have been standard flight feathers instead of RDFs.

When O'Connor et al. presented a more detailed study* on the anatomy of Chiappeavis in 2017, they also estimated the aerodynamic qualities of its supposed tail fan. They found that its lift-generating capabilities were very limited compared to those of tail fans in modern birds and their close relatives. Wang and O'Connor (2017), which was published a few months later, also found no features that made Chiappeavis any more likely to have had rectricial bulbs than other enantiornitheans. So even though the tail feathers of Chiappeavis were arranged in a similar manner to a tail fan, they probably didn't function like the spreading tail fan of modern birds.

*Interestingly, O'Connor et al. did not exclude the possibility that enantiornitheans had rectricial bulbs in this study, claiming that some modern birds also lack the pygostyle features that typically support rectricial bulbs. They cite the Wang and O'Connor study (in-prep at the time) for this, but the final version of Wang and O'Connor (2017) does not repeat the claim, instead concluding that rectricial bulbs were most likely absent in enantiornitheans.

The tail of Chiappeavis with preserved feathers, from O'Connor et al. (2016b).

In the end, Feitianius and Chiappeavis demonstrate that enantiornithean tail feather arrangements came in more varieties than the two that Clarke et al. (2006) originally identified. Given the great diversity of enantiornitheans, it's not unlikely that more will be found in the future. However, after so much back and forth, it appears that Clarke et al.'s central contention that a true tail fan is found only in avialans more closely related to modern birds remains valid.

So if enantiornitheans weren't using their tail feathers like modern birds do, what were their tail feathers for? In many cases, the most obvious possibility is some form of visual signalling. That would certainly be consistent with the evidence for well-developed tail-raising muscles, as previously noted with respect to Feitianius. In addition, modern birds that have superficially similar, streamer-like feathers tend to use them in display.

That being said, a few other possibilities have been put forth. The abundance of isolated RDFs found in amber has been suggested as evidence that these feathers detached easily, which may have allowed them to misdirect attacks from predators (much like the detachable tails of many lizards). A variety of modern birds certainly have tail and rump feathers that shed readily if grabbed by a predator. However, as with the feathers of these modern birds, I suspect that distracting predators was not the primary function of RDFs.

The RDFs of Parapengornis in particular have been proposed to have functioned as props, similar to the tail feathers of some modern climbing birds such as woodpeckers. This idea was purportedly backed up by pygostyle anatomy, but Wang and O'Connor (2017) later determined that the pygostyle of Parapengornis wasn't especially similar to that of woodpeckers after all.

Restoration of Feitianius showing it using its tail feathers for display, by Scott Reid (used with permission).

How would the absence of a lift-generating tail fan impact the flight of enantiornitheans? That a tail fan is not a requirement for avian flight is evident from observations of modern birds. Individual birds that have lost their tail feathers can usually still adapt to flying without them. There are even flying birds today that naturally lack tail fans, namely grebes (though their large, lobed feet perform some of the functions that a tail fan normally would).

Even so, I wonder if enantiornitheans were generally clumsier fliers compared to similarly-sized modern birds, or perhaps needed to expend more energy to perform equivalent aerial maneuvers. Martyniuk has speculated that enantiornitheans may have preferred to land on large surfaces such as tree trunks before climbing to desired perches, instead of trying to precisely target small twigs and branches from the get-go. Given that many enantiornitheans appear to have been small, forest-dwelling animals, they might not have needed much endurance or agility if most of their flight involved flitting from tree to tree in this manner. Some recent research has focused on inferring the likely flight behavior of Mesozoic avialans from their wing shape and body mass. It would be interesting for future studies to compare the effects of different types of tail feathering on flight ability.

Restoration of the enantiornithean Alethoalaornis in flight, by Scott Reid (used with permission).

One final subject I'll remark on here is the evolutionary history of enantiornithean tail feathers. The structure of RDFs suggests that they evolved from more typical shafted feathers. Most early avialans and other long-tailed maniraptors (including troodonts, dromaeosaurids, and oviraptorosaurs) certainly had large shafted feathers along the length of their tail. Such an array of tail feathers has often been called a "tail fan", but there is no evidence that long-tailed maniraptors had structures similar to rectricial bulbs that could fan out their tail feathers. As a result, I prefer to use Stephen Gatesy's term "tail frond" to distinguish this type of tail from the mobile tail fan of modern birds.

On the lineage leading to enantiornitheans (as well as modern birds), the bony part of the tail shortened. What would an immobile tail frond look like when retained on a shortened tail? Well, probably something similar to the immobile fan of Chiappeavis. (That this represents the ancestral state for short-tailed avialans is potentially supported by the fact that a similar fan is also known in Sapeornis, which was more distantly related to modern birds than enantiornitheans were.) A plausible evolutionary scenario may be that most enantiornithean lineages then lost the majority of their tail feathers, modifying the remainder into RDFs. In 2014, Wang Xiaoli and colleagues further suggested that the fully barbed RDFs seen in Eopengornis might represent a precursor to RDFs that had a mostly naked shaft.

Comparison of tail plumage in different Mesozoic avialans, from Wang et al. (2014). (A) shows the "tail fan" of Sapeornis, (B) shows the RDFs of Dapingfangornis (which only have barbs near their tip), (C) shows the true tail fan of Hongshanornis, and (E) shows the fully barbed RDFs of Eopengornis. ("D" featured the forked tail of Schizooura, which I cropped out as it is not directly relevant to this post.)

However, there's a wrinkle to this tale. Phylogenetic analyses that include Chiappeavis find it to be more closely related to Eopengornis and Parapengornis (which had fully barbed RDFs) than to other enantiornitheans. If these relationships are correct, they may imply that the RDFs of Eopengornis and Parapengornis evolved independently from those of other enantiornitheans, instead of representing an ancestral state for RDFs. An alternative possibility is that Chiappeavis evolved its "tail fan" from an ancestor that had RDFs and is uninformative about the ancestral state of tail plumage in enantiornitheans. Or perhaps our current understanding of enantiornithean phylogeny is wrong. Given that enantiornitheans have some of the most poorly resolved interrelationships among all Mesozoic dinosaurs, that wouldn't surprise me. At present, I don't think we have enough information to rule out any of these possibilities.

Long story short, enantiornithean tail plumage came in a wide variety of forms, all of which were different from that of most modern birds. We still have some ways to go when it comes to understanding their function and evolution, and it's likely that they will continue to surprise us in years to come.

References

• Clarke, J.A., Z. Zhou, and F. Zhang. 2006. Insight into the evolution of avian flight from a new clade of Early Cretaceous ornithurines from China and the morphology of Yixianornis grabaui. Journal of Anatomy 208: 287-308. doi: 10.1111/j.1469-7580.2006.00534.x
  
• Gatesy, S.M. and K.P. Dial. 1996. From frond to fan: Archaeopteryx and the evolution of short-tailed birds. Evolution 50: 2037-2048. doi: 10.2307/2410761
  
• O'Connor, J.K., X. Wang, L.M. Chiappe, C. Gao, Q. Meng, X. Cheng, and J. Liu. 2009. Phylogenetic support for a specialized clade of Cretaceous enantiornithine birds with information from a new species. Journal of Vertebrate Paleontology 29: 188-204. doi: 10.1080/02724634.2009.10010371
  
• O'Connor, J.K., L.M. Chiappe, C. Chuong, D.J. Bottjer, and H. You. 2012. Homology and potential cellular and molecular mechanisms for the development of unique feather morphologies in early birds. Geosciences 2: 157-177. doi: 10.3390/geosciences2030157
  
• O'Connor, J.K., D.-Q. Li, M.C. Lamanna, M. Wang, J.D. Harris, J. Atterholt, and H.-L. You. 2016a. A new Early Cretaceous enantiornithine (Aves, Ornithothoraces) from northwestern China with elaborate tail ornamentation. Journal of Vertebrate Paleontology 36: e1054035. doi: 10.1080/02724634.2015.1054035
  
• O'Connor, J.K., X. Wang, X. Zheng, H. Hu, X. Zhang, and Z. Zhou. 2016b. An enantiornithine with a fan-shaped tail, and the evolution of the rectricial complex in early birds. Current Biology 26: 114-119. doi: 10.1016/j.cub.2015.11.036
  
• O'Connor, J.K., X.-T. Zheng, H. Hu, X.-L. Wang, and Z.-H. Zhou. 2017. The morphology of Chiappeavis magnapremaxillo (Pengornithidae: Enantiornithes) and a comparison of aerodynamic function in Early Cretaceous avian tail fans. Vertebrata PalAsiatica 55: 41-58.
  
• Wang, W. and J.K. O'Connor. 2017. Morphological coevolution of the pygostyle and tail feathers in Early Cretaceous birds. Vertebrata PalAsiatica 55: 289-314. doi: 10.19615/j.cnki.1000-3118.170118
  
• Wang, X., J.K. O'Connor, X. Zheng, M. Wang, H. Hu, and Z. Zhou. 2014. Insights into the evolution of rachis dominated tail feathers from a new basal enantiornithine (Aves: Ornithothoraces). Biological Journal of the Linnean Society 113: 805-819. doi: 10.1111/bij.12313
  
• Xing, L., P. Cockx, R.C. McKellar, and J. O'Connor. 2018. Ornamental feathers in Cretaceous Burmese amber: resolving the enigma of rachis-dominated feather structure. Journal of Palaeogeography 7: 13. doi: 10.1186/s42501-018-0014-2