Soaring with Sapeornis

Among the diverse avifauna of the Cretaceous, there were none (that we know of) quite like Sapeornis from the Early Cretaceous of China. Though a few other avialans have been described as its close relatives, including "Omnivoropteryx", "Didactylornis", and "Shenshiornis", all of them almost certainly represent additional specimens of Sapeornis itself. Most phylogenetic analyses recover Sapeornis as having been more distantly related to modern birds than Confuciusornis (which is known from the same geologic formations) was, but Sapeornis possessed a strange mosaic of features, some of which were convergently similar to those of modern birds.

Like modern birds, Sapeornis had a reduced third finger that lacked a claw, unlike Confuciusornis, which retained claws on all three of its fingers. Sapeornis was also similar to modern birds in having a backwards-facing first toe that would have allowed it to easily grip a perch, whereas Confuciusornis and other avialans more distantly related to modern birds had a first toe that pointed inward at most. However, unlike Confuciusornis and modern birds, but similar to earlier paravians such as Anchiornis and Archaeopteryx, Sapeornis lacked a bony sternum (breastbone), and may not even have had a cartilaginous one. The sternum anchors many of the chest muscles used for flapping in modern birds, so this suggests that Sapeornis was not a very strong flapper.

A well-preserved specimen of Sapeornis showing the long wing feathers, from Serrano and Chiappe (2017).

Sapeornis was very large for a Mesozoic avialan, likely weighing around 1 kg, comparable in size to a common raven or a red-tailed hawk. Its forelimbs were very long and had long wing feathers attached to them, giving it an estimated wingspan of more than a meter wide, as well as a wing planform very unlike the relatively short, broad wings of most other Mesozoic avialans. So Sapeornis could likely perch skillfully with its reversed first toe, and it likely flew from perch to perch, as indicated by the reduced grasping ability in its hands and its extremely long wings. Yet its anatomy also suggests that couldn't flap its wings very well.

The size of Sapeornis compared to a human, by Matt Martyniuk, licensed.

Perhaps this apparent paradox could be explained if Sapeornis only did quick bursts of flapping flight to get from tree to tree, but such a lifestyle is more common among birds with short, broad wings, more similar to those of other Mesozoic avialans. Another possibility is that Sapeornis was specialized for soaring, given that soaring flight doesn't require a whole lot of flapping. Extant soaring birds tend to have very long wings, similar to Sapeornis, and some previous studies have certainly concluded on anatomical grounds that Sapeornis could likely soar. Until recently, however, detailed investigation into the aerodynamics of Sapeornis had not been carried out to further test these conclusions.

In a newly-published study, Francisco J. Serrano and Luis M. Chiappe used multiple lines of evidence to infer and model the flight style of Sapeornis. First, they found that the size of the deltopectoral crest (a ridge on the humerus, or upper arm bone, that helps anchor the chest muscles) is negatively correlated with flapping frequency in modern birds. In other words, birds that flap less frequently (such as soaring birds) have a larger deltopectoral crest. Sapeornis had a sizeable deltopectoral crest relative to its body size. Score one for soaring.

The forelimb of Sapeornis showing its large deltopectoral crest (DPC), from Serrano and Chiappe (2017).

Next, they used computational modeling to calculate the amount of power necessary for Sapeornis to support itself using sustained flapping as well as the amount of power it could actually generate. These results showed that it would be extremely inefficient for Sapeornis to use sustained flapping flight. This is again similar to extant soaring birds such as hawks and ravens. These birds are capable of sustained flapping, but it is energetically costly for them to do so, and as such they prefer to soar when they can. Score another for soaring.

In addition, Serrano and Chiappe estimated the wing aspect ratio (the length of the wing relative to the width of its chord) of Sapeornis using both multivariate equations as well as direct measurements of reconstructed wing area. In both cases, the wing aspect ratio of Sapeornis plotted among soaring birds, and specifically among thermal soarers such as vultures that primarily soar on columns of warm air rising off the land. This makes sense. There are other soaring birds that are dynamic soarers, specialized for soaring at high speeds over the ocean by exploiting differences in wind velocity at different altitudes. Most of these species are seabirds with narrow, pointed wings. Given that fossils of Sapeornis were preserved in inland lakes near forested environment, it would have been a big surprise if it was found to have been a dynamic soarer!

The reconstructed wing planform of Sapeornis, from Serrano and Chiappe (2017).

All of these features are consistent with Sapeornis having been a thermal soarer, but could it, in fact, soar? Thermal soarers need to be able to make tight turns so that they remain near the center of a thermal where lift is the strongest. Using their computational models, Serrano and Chiappe were also able to calculate the turning radius and minimum sinking speed of Sapeornis. They found that the smaller specimens of Sapeornis could have had turning radii comparable to those of turkey vultures, some of the most masterful extant thermal soarers. Larger specimens had turning radii closer to those of white storks, which are also capable soarers.

Meanwhile, the large wings of Sapeornis would have been more than enough to slow its sinking speed so that it could be offset by the typical speed of rising air generated by thermals, and Sapeornis is also known to have had several large feathers on its tail, which would have given it even more lift. (In contrast, Confuciusornis had a single pair of long tail feathers at most.) Some specimens of Sapeornis preserve large feathers on the feet, which weren't incorporated into this analysis. I wonder what aerodynamic function, if any, they might have had.

Diagrams showing how Sapeornis likely soared and a graph comparing its estimated turning radii and sinking speeds to those of extant soaring birds, from Serrano and Chiappe (2017).

Thermal soaring also meshes well with the likely ecology of Sapeornis. The shape of its teeth indicate that it was most likely primarily herbivorous, and seeds have been found as its gut contents, suggesting perhaps that it ate fruit. Herbivory being a relatively low-energy diet, Sapeornis would benefited from the low-energy flight style of thermal soaring. Serrano and Chiappe suggest that good ecological analogues for Sapeornis are the screamers, a group of bizarre South American waterfowl that grow daggers on their wings, given that screamers are among the few herbivorous extant birds that both soar and perch in trees. (Though I personally suspect that Sapeornis was less terrestrial than screamers, considering the relative lengths of its forelimbs and hindlimbs as well as its large, curved foot claws.)

On the whole, this study does a convincing job of showing that both the anatomy and biomechanics of Sapeornis support thermal soaring as its main flight style, providing a plausible explanation for its unusual combination of aerially-adapted characteristics with the lack of flapping adaptations. It's easy to envision Mesozoic avialans as mere intermediates, gaining progressively more impressive flying abilities as they approached modern birds. However, the specialized soaring adaptations of Sapeornis, so far unique among Early Cretaceous avialans, set a good example in showing that many of them forged their own innovations that were not directly inherited by the line ancestral to extant birds.

I look forward to seeing this kind of multi-pronged approach applied to studying the locomotion of other extinct flying dinosaurs. Perhaps we can get an enantiornithine next...? ;)

Reference: Serrano, F.J. and L.M. Chiappe. 2017. Aerodynamic modelling of a Cretaceous bird reveals thermal soaring capabilities during early avian evolution. Journal of the Royal Society Interface 14: 20170182. doi: 10.1098/rsif.2017.0182