Friday, February 8, 2019

Finches Before There Were Finches: Eofringillirostrum and the Diversity of Stem-Passerines

Many types of modern birds eat seeds from time to time. It's a concept so familiar to us that the idea of "bird food" is likely to conjure up imagery of seeds, and indeed seeds probably comprise the majority of food that we offer to both pet and wild birds. It has even been suggested that seed-eating helped the ancestors of modern birds survive the end-Cretaceous mass extinction. However, living on a diet composed primarily of seeds is something that only a relatively small number of bird groups do.

Many of these seed-eating birds are passerines. Though most modern birds can perch, passerines are often called "perching birds" because their feet are particularly specialized for this task. As a whole, passerines account for about 60% of modern bird diversity, but most seed-eating specialists belong specifically to a group of passerines called Passeroidea. Seed specialist passeroids include finches, sparrows, buntings, cardinals, weaverbirds, estrildids (such as the colorful Gouldian finch of Australia), some tanagers (including Darwin's "finches", which are not really finches), and more. All of these birds have heavy-duty, cone-shaped beaks that they use for cracking open seeds. It's perhaps not surprising that bites from seed-eating passeroids are among those most dreaded by bird banders.

Based on the latest estimates, seed specialist passeroids evolved fairly recently during the Miocene, roughly 15 million years ago. Thanks to a new discovery, however, we now know that other birds led similar lifestyles to finches and sparrows long before these groups had even appeared. In a new study, Daniel Ksepka and colleagues named two new species of Eocene birds that exhibit adaptations for seed eating similar to those of seed-eating passeroids.

One of these new species, Eofringillirostrum boudreauxi, came from the early Eocene Green River Formation in North America, making it about 52 million years old. It was a small bird, about the size of a red-breasted nuthatch, and is known from an excellent specimen, a nearly complete skeleton preserved with feathers. Its most notable feature, however, is its stout, cone-shaped bill, which bears a strong resemblance to that of finches.

The holotype of Eofringillirostrum boudreauxi, from Ksepka et al. (in press).

The other new species was also assigned to the genus Eofringillirostrum, and was named Eofringillirostrum parvulum. This species came from the other side of the globe, the Messel Shale in Germany (which dates to about 47 million years ago). It was even smaller than E. boudreauxi, though its head was proportionately larger. The type specimen of E. parvulum is not quite as well preserved as that of E. boudreauxi, but the finch-like skull is evident.

The Green River and Messel are two of the richest fossil sites when it comes to preserving Eocene bird fossils, and Eofringillirostrum is not the only Eocene bird genus that has been found at both localities. Some other birds that are known to have had similar distributions include the stem-roller Primobucco and the rail relative Messelornis.

The holotype of Eofringillirostrum parvulum, from Ksepka et al. (in press).

Though their similarity to finches is striking, the skull of both Eofringillirostrum species is notably different from those of finches in having a prominent projection at the back of the lower jaw. This is a feature typically found in birds that can open their jaws widely. The describers of Eofringillirostrum speculate that this ability allowed it to swallow large seeds and deposit them in its crop (a pouch for temporary food storage at the base of the throat in birds), or helped it gulp down fruits as an alternative food source.

The skull of Eofringillirostrum (B), compared to that of a speckled mousebird (A), which has a similar projection behind the lower jaw, and an American goldfinch (C), which has a similar cone-shaped bill, from Ksepka et al. (in press).

To find out how Eofringillirostrum was related to modern birds, the describers included it in a phylogenetic dataset along with many other species of telluravians, a diverse group of mainly tree-dwelling birds including passerines, parrots, birds of prey, woodpeckers, and more. When this dataset was analyzed, Eofringillirostrum turned out to be a stem-passerine. In other words, passerines as a whole are its closest living relatives, but it was not a member of the group exclusive to extant passerine lineages. It certainly was not particularly closely related to finches or any of the other seed-eating passerines today.

Furthermore, Eofringillirostrum was found to be a member of a specific group of stem-passerines, the psittacopedids. This group includes several other Eocene birds, including Psittacopes and Pumiliornis from the Messel and Morsoravis from the Fur Formation in Denmark. Psittacopedids have not always been recognized as stem-passerines, partly because they had a fourth (outermost) toe that was at least partially reversed. This feature (known as zygodactyly) is not found in modern passerines, in which only the first or innermost toe points backwards (as is typical of most modern birds). However, genetic data have consistently shown that the closest living relatives of passerines are parrots, which do have zygodactyl feet. In light of this, it is not so surprising that passerines appear to have evolved from zygodactyl ancestors.

There are other noteworthy aspects of the phylogeny recovered by this study. One of the oldest known true passerines, Wieslochia from the early Oligocene of Germany, was found to be a suboscine, one of the two main passerine lineages. This makes sense given that the other main passerine lineage, the oscines or songbirds, is thought to have been confined to Australia during the early Oligocene. In addition, the halcyornithids, a group of Eocene birds once thought to be most closely related to parrots specifically, were found to be stem-members of Psittacopasserae, the group uniting both parrots and passerines.

The results of the phylogenetic analysis run by Ksepka et al. (in press), from their study. Note that "Afroaves" should be labeled Australaves.

In fact, the phylogeny of psittacopasserans found by this study is strikingly consistent with the results of Mayr (2015), despite the latter having used a much smaller dataset. However, the analysis from the description of Eofringillirostrum still lacks a few more early telluravians that might be interesting to include (such as the possible stem-falcon Masillaraptor and the parrot-like, apparently raptorial Messelastur). I am curious to see this dataset expanded further in the future.

As far as we know, Eofringillirostrum was unique among psittacopedids for its seed-eating adaptations. Other psittacopedids had quite different skulls. Morsoravis had a generalized, thrush-like beak, suggesting a generalist diet of invertebrates and fruit. Psittacopes had a short, slightly downcurved beak, which is found in birds that mainly feed on insects but also eat seeds. Pumiliornis had a long beak and has been found with pollen as gut contents, indicating that it likely fed on nectar. It's often easy to imagine stem-groups as little more than intermediates "on their way" to becoming modern species, but Eofringillirostrum and other psittacopedids show that stem-passerines had their own independent burst of diversification, taking on ecological niches that true passerines wouldn't occupy until millions of years later.

The skulls of stem-passerines (left) compared to those of extant passerines (right) that exhibit similar adaptations, from Ksepka et al. (in press). Morsoravis (A-B) is compared to a hermit thrush (I-J), Eofringillirostrum (C-D) is compared to an American goldfinch (K-L), Pumiliornis (E-F) is compared to a black-throated sunbird (M-N), and Psittacopes (G-H) is compared to a bearded reedling (O-P).

Between Eofringillirostrum, fellow stem-passerine Zygodactylus ochlurus, the stem-hoopoe Laurillardia smoleni, the recently extinct penguin Eudyptes warhami, and the early waterfowl Conflicto, neornithine birds have so far had a strong showing among the new paleontological discoveries of this year. I can only hope that the rest of the year is just as good!

Reference: Ksepka, D.T., L. Grande, and G. Mayr. In press. Oldest finch-beaked birds reveal parallel ecological radiations in the earliest evolution of passerines. Current Biology in press. doi: 10.1016/j.cub.2018.12.040

Monday, February 4, 2019

Conflicto and the Evolution of Waterfowl

Estimating when a specific group of organisms appeared in Earth history is never a simple task. Fossils provide the most direct evidence of when specific organisms were around, but the fossil record is far from complete. As a result, we can't assume that the oldest fossils known from a given clade were the oldest members of that clade to have existed. Fossil taxa can only provide a minimum constraint, telling us that a clade must be at least of a certain age.

If the group we're interested in is still extant (or lived recently enough for genetic material to be recovered), then molecular clocks can help. Molecular clock analyses compare the differences between the molecular sequences of different organisms and use estimated mutation rates to approximate the amount of time that has passed since their lineages diverged. However, when dealing with extremely long timescales (such as tens of millions of years), we usually cannot assume that rates of genetic mutation have remained constant for all that time. As such, most divergence time studies make use of fossils to provide minimum constraints for when specific lineages must have diverged, setting calibrations for their molecular clock. For the results of these studies to be considered reliable though, the fossils used in molecular clock studies need to be well supported as members of the respective lineages they calibrate. After all, using a fossil species to calibrate the age of a certain group does little good if the species is not actually a member of that group.

Neornithine birds are one clade that has been at the center of controversies about the timing of their origin and diversification. However, one conclusion that all divergence time studies on neornithines agree on is that by the end of the Cretaceous, they had diverged into their three major lineages: paleognaths (ostriches, emus, etc.), galloanserans (land- and waterfowl), and neoavians (all other modern birds). This post will focus specifically on the origins of waterfowl.

Modern waterfowl can in turn be split into three main lineages: screamers (an unusual South American group), the magpie goose (a single extant species from Australia), and anatids (the most diverse group, including the ducks and geese we are most familiar with). Compared to other neornithines, the fossil record appears to have been kinder to waterfowl when it comes to preserving traces of their early evolutionary history. Whereas the only potential fossils of Cretaceous paleognaths, landfowl, and neoavians consist of fragmentary specimens of ambiguous affinities, several decently complete skeletons have been posited as strong evidence of waterfowl antiquity.

The phylogenetic relationships among living waterfowl.

The most famous of these ancient purported waterfowl is probably Vegavis, which lived in Antarctica at the very end of the Cretaceous. It is known from two partial skeletons (including one that preserves a syrinx, the vocal organ of modern birds). The original description of Vegavis found it to be more closely related to anatids than to the magpie goose or screamers, which would imply that waterfowl had already diverged into their three modern lineages by the end of the Cretaceous. However, this result has not been replicated by many recent analyses, with some researchers arguing that even galloanseran affinities for Vegavis are not strongly based. A few other Southern Hemisphere birds from around the Cretaceous-Paleogene (K-Pg) boundary have been suggested to be closely related to Vegavis. These include Polarornis from the Late Cretaceous of Antarctica, Neogaeornis from the Late Cretaceous of Chile, and Australornis from the Paleocene of New Zealand. This supposed close relationship has also been questioned though, and in any case these birds are known from far less complete material than Vegavis, limiting their potential in elucidating waterfowl evolution.

Anatalavis is another fossil bird that might provide evidence for an early radiation of modern waterfowl. The type species, A. rex, comes from the North American Hornerstown Formation, which appears to straddle the K-Pg boundary. A. rex is only known from incomplete arm bones, but a second species (A. oxfordi) from the early Eocene of the United Kingdom is known from a partial skeleton. The broad, flattened bill of A. oxfordi is certainly quite duck-like, and its original description suggested that it was a close relative of the magpie goose.

Then there are the presbyornithids, a group of extinct, long-legged waterfowl. Presbyornithids are best known from Paleogene fossils, but a few possible Cretaceous records have been reported, including Teviornis from the Late Cretaceous of Mongolia (known from a partial forelimb). The phylogenetic position of presbyornithids is disputed, but they are often found to be crown-waterfowl (i.e.: nested among the extant waterfowl lineages), usually as close relatives to anatids.

If all of these aforementioned ancient birds were crown-waterfowl as has been suggested, that would indicate that all three extant lineages of waterfowl originated in the Cretaceous and made it through the K-Pg extinction, along with a few extinct waterfowl groups. But were there really modern-type ducks paddling around at the same time that Tyrannosaurus rex was alive? The oldest fossils of unambiguously anatid-like waterfowl are much younger, hailing from the late Eocene, and the oldest unambiguous magpie goose fossil comes from the Oligocene. This by itself does not falsify an early origin of crown-waterfowl; after all, we are familiar with the concept that the fossil record contains many gaps. However, without a better understanding of how the ancient waterfowl relate to extant ones, it is difficult to determine when the lack of fossils reflects true absence and when we're simply looking at a missing record.

A new fossil described by Claudia Tambussi and colleagues might shed light on this question. The fossil comes from the early Paleocene of Antarctica, about 64.5 million years ago, putting it shortly after (by the standards of geologic time) the K-Pg mass extinction of 66 million years ago. Assigned to a new genus and species, Conflicto antarcticus, the specimen is extremely well preserved for a bird fossil. The bones are preserved in three dimensions (instead of being flattened) and represent much of the skeleton, including the skull and most of the major limb bones other than the feet.

The skull of Conflicto, from Tambussi et al. (in press).

Conflicto was about the same size as an extant magpie goose (and thus larger than typical ducks). Its overall anatomy, most prominently its flattened bill, makes it clear that it was a waterfowl, but which modern waterfowl was it most closely related to? To find out, its describers entered it into the phylogenetic dataset used by Worthy et al. (2017), probably the most comprehensive morphological dataset focused on galloanserans so far. When they did so, they found that Conflicto was equally closely related to all extant waterfowl; in other words, it fell outside of the group exclusive to the extant waterfowl lineages. Conflicto was a stem-waterfowl, not a crown-waterfowl.

Possibly even more interesting, however, was what happened to other early waterfowl in this analysis. Anatalavis (also included in this dataset for the first time, as far as I'm aware) was recovered not as a close relative of the magpie goose, but as another stem-waterfowl. In fact, it was found to be the closest known relative to Conflicto, though the authors of the study point out that statistical support for this result is weak. The presbyornithids (formerly recovered as crown-waterfowl by Worthy et al.) turned out to be stem-waterfowl as well. Vegavis was found to be yet another stem-waterfowl, but most curiously it was found to be a close relative of gastornithiforms, which were giant flightless galloanserans so far only known from the Cenozoic Era. It should be noted that the relationship between Vegavis and gastornithiforms is not entirely new, as it was also found by some of the analyses run by Worthy et al. In addition, it does not have strong statistical support.

In any case, the analysis including Conflicto excludes all of these ancient birds from the radiation of modern waterfowl (and that's not to mention the aforementioned skepticism of Vegavis being a galloanseran at all). Thus, they probably should not be used as calibrations for the age of modern waterfowl in future molecular clock studies. If this phylogeny is correct, we have no fossil evidence of modern-type waterfowl from before the K-Pg boundary. Modern waterfowl could well have originated later during the Paleogene, which would be consistent with their known fossil record.

The results of the phylogenetic analysis run by Tambussi et al. (in press), from their study.

In addition to the timing of waterfowl evolution, Conflicto also has implications for the evolution of notable waterfowl features. Among modern waterfowl, the screamers stand out in lacking a flattened beak (along with other strange characteristics), instead having one superficially similar to that of landfowl such as chickens. Given that the magpie goose and anatids are more closely related to each other than to screamers, it might be reasonable to assume that the characteristic flattened bill of these waterfowl originated relatively recently, after their lineage had diverged from that of screamers. However, it has long been noted that screamers have vestigial versions of the filter-feeding plates found inside the mouths of other waterfowl, raising the possibility that the ancestral waterfowl was a flat-beaked filter feeder and that these features were later lost by the screamer lineage.

Conflicto supports this second possibility. With the finding that the flat-billed Conflicto, Anatalavis, and presbyornithids might have been stem- rather than crown-waterfowl, it appears likely that all modern waterfowl (including screamers) descended from a filter-feeding bird with a duck-like bill. Furthermore, Conflicto and presbyornithids both had long hindlimbs, suggesting that this might have been another feature present on the line leading to modern waterfowl.

Skeletal of Conflicto, from Tambussi et al. (in press). Preserved bones are shown in white.

The describers of Conflicto named it for their prediction that its phylogenetic position and evolutionary implications are likely to become the subject of heated debate. They are probably correct. However, for the time being I personally find the results of this study to be very appealing in terms of its congruence with the known fossil record and previous observations regarding extant waterfowl. Let's see what the future brings.

Reference: Tambussi, C.P., F.J. Degrange, R.S. De Mendoza, E. Sferco, and S. Santillana. In press. A stem anseriform from the early Palaeocene of Antarctica provides new key evidence in the early evolution of waterfowl. Zoological Journal of the Linnean Society in press. doi: 10.1093/zoolinnean/zly085