Wednesday, March 18, 2020

Catch a Falling Star: Meet Asteriornis!

Today marks the publication of a very exciting study that I have been waiting to talk about for a long time. Led by my supervisor Daniel Field, we describe a new species of fossil bird from the Late Cretaceous of Europe, and it's a doozy.

Doesn't look like much, does it? But read on... (Photo by Daniel Field.)

Among the most highly sought-after discoveries in the avian fossil record are specimens of the oldest modern-type birds. Although we now know of numerous bird-like dinosaur fossils that document the evolutionary acquisition of many notable avian features, the origin of the modern bird group (Neornithes) itself remains cryptic.

Molecular clock analyses, which use molecular sequences from different organisms and estimated mutation rates to infer the age of divergence between different lineages, place the origin of modern-type birds in the Cretaceous. For reasons that I won't go into here (lest I wander too far off topic), timing the origin of modern birds is a subject fraught with controversy. However, all recent molecular clock analyses of modern-type birds agree in placing the divergences between the three main extant avian lineages—paleognaths (ostriches, emus, and their kin), galloanserans (chickens, ducks, and their kin), and neoavians (all other living birds)—by the end of the Cretaceous. Despite this, fossil evidence for these hypothetical modern-type birds in the Cretaceous has been limited.

There have been a number of specimens purported to represent Cretaceous modern-type birds, but most of them are so fragmentary that their supposed affinities are best taken with an unhealthy serving of salt. By far the best represented (and probably the most convincing) example of a Mesozoic modern-type bird is Vegavis iaai, known from two partial skeletons from the Late Cretaceous of Antarctica. However, the skull of Vegavis is largely unknown. Furthermore, how Vegavis relates to living birds is not well understood. Although it was originally described as a close relative of modern ducks and geese, this has been questioned by some recent studies.

By the end of the Cretaceous (66 million years ago), modern-type birds had likely diverged into at least three major lineages.

In 2018, Daniel arranged a loan of a fossil bird specimen with the Maastricht Natural History Museum (Natuurhistorisch Museum Maastricht or NHMM) in the Netherlands. My labmate Juan Benito Moreno traveled to Maastricht to borrow the specimen from curator and coauthor John Jagt, who also provided us with information about the geologic setting in which the fossil had been discovered.

The specimen had been recovered from a quarry in Belgium, hailing from the Valkenburg Member of the Maastricht Formation. This site dates to 66.8-66.7 million years old, very close to the end of the Late Cretaceous (66 million years ago). Another fossil from this site had been previously described briefly as an Ichthyornis-like bird. Seeing as it preserves at least one tooth, this previously-studied specimen was certainly not a modern-type bird.

The new specimen is a very unassuming fossil, composed of four blocks of rock matrix that together take up about as much space as a deck of playing cards. In fact, the only reason we paid any attention to it was because one of the blocks (pictured at the beginning of this post) had an avian leg bone exposed on its surface. Bits and pieces of other bones were visible as well, but nothing that looked particularly exciting at a glance. Little did we know what we would find when we took a look inside.

To visualize the fossil material still hidden inside the rock, Daniel and Juan CT-scanned the specimen at the Cambridge Biotomography Centre. Shortly after they'd done so, Juan sent me a photo of what they saw inside the aforementioned block.

Holy fucking shit.

It was the nearly complete skull of a modern-type bird! Although it was distinct from any other bird skull known, we immediately noticed similarities with those of extant galloanserans. For reasons that I will explain near the end of this post, we eventually decided to give this bird the scientific name Asteriornis maastrichtensis. However, while we were in the process of studying the specimen, we called it by the nickname "Wonderchicken", for its suspected galloanseran affinities.

Due to the delicate nature of the fossil, we chose not to physically separate the bones from the surrounding rock. (For some perspective, the preserved length of that skull is less than 5 cm long, shorter than my thumb.) Instead, we spent much of 2019 digitally isolating each of the bones in the CT scans. In addition to the amazing skull, we were able to identify several wing and leg bones contained within the blocks, even though these bones were broken or incomplete.

Once we had digitally extracted the bones, we could look at them more thoroughly and make detailed comparisons with the anatomy of other birds. Here our coauthor Dan Ksepka brought his extensive experience in studying bird skeletons to bear, and we were able to confirm the presence of several features in Asteriornis that are also seen in most galloanserans. These include long, curved retroarticular processes (projections at the back of the lower jaws), narrow, upward-pointing medial processes (projections on the inside of the lower jaws), and a shallow groove along the top of the skull. In addition, the quadrate (a bone at the back of the upper jaw) of Asteriornis closely resembles that of the Paleogene waterfowl Presbyornis.

The skull of Asteriornis in multiple views. Juan spent a copious amount of time preparing the anatomical figures in our paper. This image is Extended Data Fig. 2 from Field et al. (2020).

To test our phylogenetic hypothesis and to give us a better handle on exactly what type of galloanseran Asteriornis might have been, we added Asteriornis to a modified version of the phylogenetic dataset used by Tambussi et al. (2019) in their description of the Paleocene waterfowl Conflicto. (As an aside, the last sentence in my blog post about Conflicto was intended to be an oblique reference to Asteriornis.) This dataset includes a range of extant and extinct galloanserans, and originated from Worthy et al. (2016).

The phylogenetics section of our study was my primary contribution to the description of Asteriornis. I implemented many of the changes to the dataset that we'd agreed to make, set up the phylogenetic analyses, and wrote up our phylogenetic methods and results. (Readers interested in those details are encouraged to check out our 136-page supplementary material. It contains more than just phylogenetics, but my input does take up a pretty big chunk!)

Ultimately, one of our analyses found Asteriornis to have most likely been a stem-galloanseran (more closely related to galloanserans than to any other living group, but not a member of either the landfowl or the waterfowl lineages). However, some of our other analyses found it best supported as an early stem-landfowl (more closely related to chickens than to ducks, but outside of the group uniting all modern landfowl), and we found that an early stem-waterfowl position was potentially plausible as well. This is not terribly surprising; species that retain many ancestral traits of a given group are often very difficult to confidently place on one specific branch or another. Yet regardless of exactly where Asteriornis goes in galloanseran phylogeny, no other fossil has been identified as a better representative for what the ancestral galloanseran probably looked like.

Landfowl and waterfowl have not always been widely accepted to be close relatives. Some previous authors have even provided long lists of anatomical differences between these two groups while expressing skepticism of such a relationship. Although the close ties between landfowl and waterfowl are now strongly supported by more recent studies, the distinct anatomy between the two groups has made it difficult to infer what their ancestral form was like. Asteriornis exhibits a mixture of traits from both lineages. Like landfowl, the bones in its snout are only weakly joined to each other and there is no obvious hinge at the base of its upper beak. On the other hand, its rounded, unhooked bill tip is more commonly found in waterfowl, as are the forward-pointing bony projection behind each eye socket and the hourglass-shaped depression on the top of its skull.

The skull of Asteriornis (with particular focus on the quadrate bone) compared to those of other galloanserans, including an Australian brushturkey (Alectura lathami), an... actual turkey (Meleagris gallopavo), the Eocene waterfowl Presbyornis, the Paleocene waterfowl Conflicto, a southern screamer (Chauna torquata), and a mallard (Anas platyrhynchos). This image is Extended Data Fig. 4 from Field et al. (2020).

We recovered a few additional phylogenetic results that I found especially interesting. First of all, none of our analyses found a particularly close relationship between Vegavis and waterfowl, supporting recent studies that have raised questions about its affinities. In fact, one of our analyses even found it outside of modern-type birds, though support for this result was very weak (and we are pretty skeptical of it). Secondly, in some (but not all) of our analyses, we recovered the extinct, long-legged presbyornithids as stem-waterfowl instead of their more traditional position as extinct members of the modern waterfowl group, similar to the findings of Tambussi et al. (2019). Having seen the similarities between Asteriornis and Presbyornis, I personally feel quite amenable towards the possibility that presbyornithids were stem-waterfowl.

We were also interested in adding two species of pelagornithids to our phylogenetic analyses. These large, extinct seabirds with tooth-like projections in their beaks have been suggested to have been aberrant galloanserans or close relatives thereof, but had not been included in previous versions of the dataset we used. However, our analyses did not find any consistent position for them. Resolving their evolutionary relationships will probably require a detailed study for another time.

The results of one of our phylogenetic analyses. This one recovered Asteriornis as a stem-galloanseran. This image is modified from Extended Data Fig. 9 from Field et al. (2020), and the restorations of Asteriornis were painted by Phillip Krzeminski.

Given that Asteriornis lived near the end of the Cretaceous, it provides us with a potential test of previous ideas about Cretaceous modern-type birds. Recent studies have predicted that the modern-type birds at the end of the Cretaceous were likely small-bodied, ground-dwelling animals, and that these factors may have been critical to their survival across the Cretaceous-Paleogene (K-Pg) mass extinction. To determine the body size of Asteriornis, Daniel used equations for estimating avian body mass from the width of their limb bones, and calculated its body mass at less than 400 g. That's much smaller than the average domestic duck or chicken, falling within the size range of a teal or partridge: relatively small by galloanseran standards.

As for the preferred habitat of Asteriornis, the preserved limb bones in our specimen are mostly incomplete, but enough is present to show that the hindlimbs of Asteriornis were relatively long and slender, well within the expected proportions for a ground-dwelling bird. This, along with the nearshore marine rocks that the specimen was preserved in, may even suggest that it mainly lived and foraged on the beach. Interestingly, the idea that all modern birds evolved from shore-dwelling ancestors has been put forth by other paleontologists in the past. Our current understanding of avian evolution does not support the notion that the group of living birds popularly known as "shorebirds" (Charadriiformes) gave rise to other modern birds, but Asteriornis suggests that a shorebird-like ancestral ecology may well be plausible for at least some modern bird groups.

Finding a single species of modern-type bird from near the end of the Cretaceous that fits these predictions certainly does not demonstrate that these features were widespread in end-Cretaceous modern-type birds, nor that they contributed to avian survival across the K-Pg. However, it is at least consistent with these hypotheses. Here's hoping more well-preserved modern-type bird specimens from Late Cretaceous rocks will be found that can further inform us about this subject!

Wonderful restoration of Asteriornis in what was likely its natural habitat, by Phillip Krzeminski. Two Ichthyornis-like birds squabble over a mosasaur carcass in the background.

One final takeaway from our study that I would like to highlight pertains to where Asteriornis was discovered. Some researchers have proposed that modern-type birds most likely originated in the Southern Hemisphere, largely based on the distribution of bird groups today and seemingly corroborated by the discovery of Vegavis and similar birds from the Cretaceous of Antarctica. Being from Europe, however, Asteriornis shows that modern-type birds were present in the Northern Hemisphere during the Late Cretaceous as well. This does not necessarily indicate that modern-type birds instead arose in the north, but it does warn us that the available data does not unambiguously point to a southern origin.

I promised that I would explain why we named Asteriornis the way we did. The species name, maastrichtensis, naturally reflects the Maastricht Formation in which the fossil had been found. Selecting a genus name, however, was a source of internal debate within our team for some time, until Dan Ksepka, in a stroke of genius, came up with "Asteriornis". This name translates to "Asteria's bird", after the Titaness Asteria from Greek mythology. Asteria relates to our fossil bird in three different ways: she was the goddess of falling stars (referencing the K-Pg asteroid impact that would occur shortly after the time that Asteriornis was alive), in mythology she transformed into a quail (referencing the galloanseran affinities of Asteriornis), and she threw herself into the sea to escape an amorous Zeus (referencing the preservation of Asteriornis in marine rocks).

Fossils from the Maastricht Formation have been collected and studied for over 200 years. That a remarkable find like Asteriornis had gone undiscovered for so long is testament to the fact that even well-excavated fossil sites can continue to surprise us. As for Asteriornis itself, there is still much left to learn about our Wonderchicken and we are excited to continue studying it.

Also... yeah, I suppose I've helped describe a new Mesozoic dinosaur in Nature now.

Reference: Field, D.J., J. Benito, A. Chen, J.W.M. Jagt, and D.T. Ksepka. 2020. Late Cretaceous neornithine from Europe illuminates the origins of crown birds. Nature 579: 397-401. doi: 10.1038/s41586-020-2096-0