Wednesday, May 9, 2018

Did Confuciusornis Really Eat Fish? The Mysteries of Mesozoic Bird Diets

Without a doubt, the Jehol Biota dating to the Early Cretaceous of northeastern China has been one of the greatest contributors to the exponential increase in our understanding of Mesozoic bird biology and evolution in recent years. Hundreds of complete dinosaur skeletons have been recovered from the Jehol, often preserved with soft tissue structures such as feathers and scales. These fossils have shed much light on the anatomy and life appearance of Mesozoic birds and other feathered dinosaurs. Naturally, such complete specimens sometimes also preserve gut contents, providing us with information about the feeding habits of these dinosaurs.

A recent paper by Jingmai O'Connor, a leading expert on Jehol fossil birds, reviews our current knowledge of Mesozoic bird diets. I was reminded of Darren Naish's 2014 review paper on evidence of avian behavior (including feeding) preserved in the fossil record... which isn't cited by this new paper. Hmm.

Nonetheless, the detailed overview and evaluation of previous claims makes O'Connor's paper an important publication on the topic of Mesozoic bird biology. As expected, Jehol specimens have contributed the most direct evidence on this subject: fish have been preserved as gut contents or pellets in the euornithines Piscivoravis and Yanornis, whereas seeds have been preserved as gut contents in the long-tailed avialan Jeholornis, the basal short-tailed avialan Sapeornis, and the euornithine Eogranivora. A few Mesozoic birds outside of the Jehol have also been reported with gut contents: the enantiornithine Eoalulavis from Spain was preserved with crustacean exoskeletons and the hesperornithine Fumicollis (not mentioned in O'Connor's paper) from the United States was preserved with fish remains.

Seeds preserved in the gut of Jeholornis, from O'Connor (in press). Insets show the skull and teeth of Jeholornis.

In addition, a large number of Jehol birds have been preserved with gastroliths (gizzard stones). These include specimens of Jeholornis, Sapeornis, Archaeorhynchus, Bellulornis, Changzuiornis, Dingavis, Eogranivora, Gansus, Iteravis*, and Hongshanornis. Gastroliths are also known in some non-avian theropods (such as Caudipteryx), but it appears that not all Mesozoic theropods used gastroliths given that some have never been preserved with them despite being known from numerous specimens. It's likely that a muscular gizzard was widespread in theropods, yet only some lineages independently evolved the habit of swallowing stones as digestive aids.

*This paper is yet another recent publication that accepts the synonymy of Iteravis and "Gansus" zheni (even though they're still listed separately in the table listing preserved gut contents in Mesozoic birds). Looks like this taxonomic revision can now be comfortably considered the general consensus.

Cretaceous euornithines that preserve gastroliths, from O'Connor (in press). (A) is Dingavis, (B) is Archaeorhynchus, (C) is Eogranivora, (D) is Hongshanornis, and (E) is Gansus.

The positioning of gut contents preserved in the specimens can sometimes provide evidence for the anatomy of their digestive system. Ingested food in both Sapeornis and Yanornis have been found clustered near the base of the neck, suggesting that these birds had a crop, an expansion at the base of the esophagus that is used to temporarily store food.

Despite the wealth of information they contain, however, interpreting Jehol fossils is not always straightforward. They are generally preserved as crushed slabs, which can distort anatomical information in spite of the completeness of the specimens. When it comes to inferring diet of fossil taxa, chance associations between the specimens and external objects can also mislead. O'Connor identifies several instances of purported gut contents in Jehol birds that are not as convincing as the previous examples.

Fish remains associated with the euornithine Jianchangornis** and the enantiornithine Piscivorenantiornis, for example, are not located directly within their body cavity and cannot be definitively considered evidence for diet, contrary to previous interpretations. In fact, the mass of fish bones preserved near Piscivorenantiornis (originally interpreted as a pellet) more strongly resembles the fossilized droppings of carnivorous fish or aquatic reptiles that have previously been found in the Jehol. This renders the name of this bird (meaning "fish-eating opposite bird") particularly awkward.

**An additional tidbit provided in the paper is that the skull of the only known specimen of Jianchangornis is a composite constructed from several different animals, not all of which may even be birds! Given that features of its teeth were claimed to be one of its distinguishing features in its original description, a restudy of the specimen may be in order.

Another enantiornithine, Bohaiornis, has been suggested to have been a carnivorous, raptorial bird, based primarily on the fact that one specimen was preserved with a few small gastroliths. Raptorial birds today are known to consume small numbers of stones to aid in cleansing their digestive system, and it was thought that this represented evidence of Bohaiornis having done the same. (In contrast, the gastroliths of herbivorous birds tend to be more numerous and are used to help grind up food.) However, based on personal examination of the specimen, O'Connor concludes that the supposed gastroliths are more likely to be precipitated minerals.

Enantiornithines purportedly preserved with gut contents, from O'Connor (in press). (A) is Bohaiornis, (B) shows its supposed gastroliths, and (C) is Piscivorenantiornis.

Outside of China, the enantiornithine Enantiophoenix from Lebanon has been interpreted as a sap eater based on globs of amber associated with the holotype, but the specimen is highly disarticulated and as such the amber cannot be confirmed as gut contents. With these reevaluations, the only direct evidence of diet in enantiornithines is found in the aforementioned Eoalulavis. Enantiornithines likely exhibited a variety of specialized feeding strategies given the diversity of their tooth and skull shapes, but it appears that they mainly fed on foods that would not preserve easily as fossils.

One Jehol bird in which the absence of preserved gut contents has been especially perplexing is Confuciusornis. Known from hundreds of specimens, more specimens of Confuciusornis have been collected than likely any other Mesozoic dinosaur. Despite this, no specimens that preserve definite gut contents have been identified. One individual was thought to preserve fish remains in a crop. However, this association is by no means unequivocal, especially given that it has only been reported in one out of hundreds of Confuciusornis specimens. O'Connor points out that, in contrast, about half of all known Yanornis specimens preserve fish in their gut. Regardless of whether the fish associated with that one specimen represents ingested material, evidence for fish being the main diet of Confuciusornis is lacking.

One of the many, many, many known specimens of Confuciusornis, photographed by Tommy from Arad, under CC BY 2.0.

Even without direct evidence from gut contents, one might think it wouldn't be so hard to get at least a general idea of what Confuciusornis ate. Inferring diet in many Mesozoic birds can be tricky because they had teeth, which makes it difficult to compare them to modern birds (all of which are toothless). Confuciusornis, however, had a beak instead of teeth, convergent with modern birds. Despite this, no living bird has a beak that closely resembles that of Confuciusornis, setting us back to square one.

Not that paleontologists haven't tried drawing conclusions from the beak morphology of Confuciusornis. A common interpretation is that Confuciusornis ate seeds, which on the surface seems reasonable. Its beak was quite robust and tooth loss is commonly associated with herbivory in other theropods. However, this runs into the same problem as the suggestion that it fed mainly on fish: definite seed-eating birds found in the Jehol regularly preserve gut contents, yet the same cannot be said of Confuciusornis.

Another recent study may shed light on this problem. Andrzej Elżanowski and colleagues have redescribed the skull anatomy of Confuciusornis based on detailed examination of multiple specimens and comparison with modern birds. As mentioned previously, the anatomical features of Jehol fossils aren't always easy to interpret, and the reconstruction presented in this new paper differs from those of previous authors in some notable ways. It's probably no coincidence that one of the authors on the paper is Gerald Mayr, known for his research on similarly crushed bird fossils of the Eocene Messel Shale.

Reconstructed skull of Confuciusornis, from Elżanowski et al. (2018).

One of the most surprising discoveries made by the researchers was that, despite not being particularly closely related to modern birds among Mesozoic avialans, Confuciusornis resembled certain living birds in more than just its lack of teeth. In their new interpretation, Confuciusornis had a bony rim behind its eye socket, and this rim was connected to the back of the skull by a bony bridge. Among modern birds, the prominent rim behind the eye is also found in the tawny frogmouth, the courol, puffbirds, some rollers, and some kingfishers. In addition, the tawny frogmouth also has a bony connection between the rim and the back of the skull.

All of these modern birds feed in a similar manner: watching from a perch before flying forth to grab their prey (often insects) in their jaws. This method of hunting is called sally-striking. The bony expansions of the skull behind their eyes likely provide room to attach larger jaw-closing muscles, increasing the grip strength of their beak.

Some modern sally-striking birds, including a tawny frogmouth (top left), white-whiskered puffbird (top right), Indian roller (lower left), and brown-headed paradise kingfisher (lower right), composited from photographs by JJ Harrison, Len Blumin, Shantanu Kuveskar, and "markaharper1", under CC BY-SA 3.0.

Is a sally-striking lifestyle consistent with the rest of Confuciusornis anatomy? Probably! Like modern sally-strikers, its beak was relatively large and deep. A close relative of Confuciusornis, Changchengornis, even had a slightly hooked bill like many modern sally-strikers do. The foot anatomy of Confuciusornis suggests that it was capable of perching, and its wing shape would have allowed it to maneuver well in its forest environment while making its prey-catching flights. Confuciusornis was also about the same size as some of the larger sally-striking birds today (such as the tawny frogmouth). And if Confuciusornis mainly fed on invertebrates whose remains might not have survived for long in the digestive system, that would explain why its gut contents are so rarely preserved.

Sally-strikers may snatch their quarry from the ground, from vegetation, or in mid-air. Though Confuciusornis could probably perch, some of its limb proportions have been likened to birds that spend part of their time on the ground. And while it may have had relatively maneuverable flight, it still lacked many of the flying specializations that modern birds possess, so it may not have been quite as acrobatic as some living sally-strikers. Perhaps it primarily apprehended prey on the ground, which might also explain how two individuals ended up as gut contents of the terrestrial theropod Sinocalliopteryx!

On the whole, the suggestion that Confuciusornis was a sally-striker appears plausible in light of what was previously inferred about its biology. Given this model, I do wonder if we'll ever find direct evidence of Confuciusornis feeding on vertebrates. Many modern sally-strikers also readily prey on smaller vertebrates in addition to insects, and Confuciusornis certainly was large enough to have done so. Even if vertebrates did not account for a large proportion of its diet, it is perhaps not a huge stretch to predict that such a discovery will one day be found.