September 15, 2016:

by Justin Tweet

By now you may have heard about the new paper by Ji Q. et al. on Liaoningosaurus, proposing some surprising things about the paleobiology of this ankylosaur. Andrea Cau already had a go at it over at Theropoda. I'm afraid I won't be as concise or insightful, but possible gut contents are involved, and having fought my own gut-content dinosaur to a truce, I think I can bring a distinct perspective to this paper.

Let's get what is by far the most frustrating part out of the way quickly: Ji Q. et al. (2016) mention that "a number" of specimens are known, but we only really get to hear about two, the holotype and the new fishy individual. I would like to propose a couple of guidelines: if you are going to make paleobiological inferences about a fossil organism, please please please use as much of your sample size as you can, and please please please include it in the paper! In this case, I would have really liked to know if there were more specimens with possible gut contents, and if other specimens can show if the holotype's belly armor is actually armor or skin impressions. (Also, if the ages of your specimens are important for your propositions, please please please do histology!)

The paper has three significant proposals about the paleobiology of Liaoningosaurus. In approximately the order in which they are presented, they are:
1. That it was carnivorous
2. That it was adapted for an aquatic lifestyle
3. That it is known from tiny adult ankylosaurs and not young juveniles

Hypothetically #3 should be the easiest proposition to test, via histology, but the other two are much harder to address if you're looking for firm conclusions. Eventually you end up arguing from first principles, which is only fun as long as you don't look down and see that you're running out over empty space. Some considerations on #1 and #2 follow.

Care and feeding of Liaoningosaurus
The guest of honor is Xinghai Museum of Paleontology (xhpm)-1206, part-counterpart of a nearly complete skeleton, reasonably articulated. It is slightly larger than the holotype, which itself is only 17.3 cm from skull through sacrum (i.e., tail not included). In and around the rib cage are numerous tiny fish skeletons, and there is a tiny tail fragment hanging out approximately where the ankylosaur would have ended along the left abdomen. Ji Q. et al. interpret the fish and tail as things the ankylosaur ate. Before getting there, they consider three general possibilities to explain the presence of the fish:

1) Dead fish were swept into the body—the authors eliminate this on the grounds that fish are not found outside of the body (with the exception of one next to the right femur), and also that the preservation of the fish differs from other fish found in the same horizon without benefit of ankylosaur. These points are reasonable but not sufficient. The preservation of the fish comes up later as well, and I would have liked consideration of the possibility that the chemical micro-environment of the decaying dinosaur could produce the different preservation of the fish. In short, if some dead fish (or live fish that become dead fish) end up in the former guts of a recently deceased and decaying dinosaur, how will they look upon fossilization? Will they look "digested" compared to control fish? I do however concur that these fish were not swept in, because the scenario would require a supply of unprotected but reasonably intact fish corpses in the environment, and these things don't have an extremely long shelf life even without the hazards of transportion (see Hellawell and Orr 2012 for some interesting experiments relative to the Green River Formation).

2) The dinosaur was dead, a bunch of fish came to scavenge or shelter in it, and died while within—the authors eliminate this on the basis of the preservation, the apparent absence of openings in the body, and the presence of a belly shield in the holotype. We talked about the preservation above. In regard to the body being unopened, all we can say is that the ribs aren't broken. We don't have soft tissue impressions, and we don't have any way to be certain that the body had not split along the sides, or that openings formed between the ribs. Plain and simple, it's practically impossible to show that the body wall of something the size of a dinosaur was completely intact even when you have extensive skin impressions. Unfortunately, xhpm-1206 is bare bones. The belly armor plate is something that crops up several times in the paper. The presence of such a plate is not reported for xhpm-1206, and it has been argued recently (Arbour et al. 2014; Arbour and Currie 2015) that the "plate" in the holotype is not an armor plate at all, but more likely skin impressions (see here for a nice photo). If this is correct, it puts something of a crimp here and at several other points in this paper.

3) The dinosaur ate the fish—this is the scenario that the authors find most convincing. They interpret the preservation of the fish as evidence of digestion, and find it unlikely that the reptile tail is evidence of scavenging. The locations of the fish around the margin of the abdomen is interpreted as evidence of having been expelled or otherwise scattered post-mortem.

I am skeptical of the evidence. I don't think it's impossible that these fish are gut contents, but I consider the taphonomic evidence to be equivocal. It's just really hard to definitively prove or disprove gut contents. With that said, I would not have extended the case from this individual to the species as a whole, not with the equivocal evidence and certainly not in the face of additional possibilities for diet. Just like the difference between weather and climate, one dinosaur with fish in its guts does not show that all members of the species were strict carnivores, or even that the individual was a strict carnivore. Even in cases where we have good reason to infer diet of individual dinosaurs, we have questions about how typical that diet was for the species. For example, Chin (2007) described wood in probable Maiasaura coprolites, which she postulated was a recurring, possibly seasonal diet, as opposed to usual. Carpenter (1987) has suggested that the twiggy, needle-rich gut contents of one of the "Trachodon" mummies actually represented material eaten by a starving animal. In this case, if the dinosaur did indeed eat the fish, we have at least the following possibilities to consider:
Liaoningosaurus as a species was carnivorous
Liaoningosaurus as a species was opportunistic, with this individual eating a bunch of live or dead fish because they were at hand
Liaoningosaurus as a species had a diet that changed with age, perhaps eating more protein as a juvenile
This particular animal ate the fish mistakenly (which brings in the possibility that a meal of fish, perhaps rotten fish, killed it, but this lies entirely in speculation)

The submersible ankylosaur?
A proposed diet of fish leads into the next proposition, that Liaoningosaurus was an aquatic carnivore. The major points of evidence for aquatic habits, apart from a diet of fish, are the belly plate (interpreted as similar to a turtle's plastron), and the poor fusion of the bones, which is usually interpreted as evidence of the animal being juvenile but here is compared to various aquatic and semi-aquatic reptiles. The fusion goes into the proposal about size and development, so I'll hold off, but in regard to the proposal of Liaoningosaurus as an aquatic carnivore, I am skeptical. In the first place, Liaoningosaurus does not have any obvious adaptations for swimming. The hands and feet are small and compact, unlike paddles. The tail is skinny and relatively short. The body is broad, not laterally compressed. To me, if you put this body plan in the water, you'd get at best a wader. In terms of offensive capabilities, Liaoningosaurus lacks the neck adaptations of snapping turtles or the head and neck specializations of choristoderes, which have either long necks with small heads or long narrow snouts with numerous spiky teeth. Pointed unguals are indicated on the hands and feet, but they are not laterally compressed, and it's hard to see how the animal could make use of those on the hind feet for predation, especially against small, fast-moving targets like fish. Pointed unguals are also present in Chuanqilong and Dyoplosaurus (Arbour and Currie 2015).

And after all of that, a closing scenario that involves eating fish
In the end, I suspect that xhpm-1206's fish were scavengers, and the reptile tail was something that drifted in (or was already there when the dinosaur carcass came to rest). However, the reported presence of only small, juvenile-appearing Liaoningosaurus in the Yixian Formation is intriguing. For fun, here's a scenario where Liaoningosaurus includes fish in its diet without being aquatic or represented by tiny adults: perhaps Liaoningosaurus had age-segregated habits, with the younger individuals inhabiting environments near the water's edge and eating a wide variety of foods (which could include fish), while the adults lived farther from the water and had more typically ankylosaurian proclivities. How could this be tested? Well, we couldn't test everything from fossils, but we could test the age issue with histology, where the large reported sample size should tell us a thing or two. If indeed we are looking at an entirely juvenile population in the lake fossils, we'd be partway there, but we'd still need an adult (or at least a mostly grown individual) before we could think about addressing dietary segregation and specialization. I suppose we could try stable isotopes on the teeth to look at trophic level between juveniles and adults.

References
Arbour, V. M., M. E. Burns, P. R. Bell, and P. J. Currie. 2014. Epidermal and dermal integumentary structures of ankylosaurian dinosaurs. Journal of Morphology 275:39–50.

Arbour, V. M., and P. J. Currie. 2015. Systematics, phylogeny and palaeobiogeography of the ankylosaurid dinosaurs. Journal of Systematic Palaeontology 14(5):385–444.

Carpenter, K. 1987. Paleoecological significance of droughts during the Late Cretaceous of the Western Interior. Pages 42–47 in P. J. Currie and E. H. Koster, editors. Fourth symposium on Mesozoic terrestrial ecosystems, Drumheller, August 10–14, 1987. Royal Tyrrell Museum of Palaeontology, Drumheller, Alberta. Occasional Paper of the Tyrrell Museum of Palaeontology 3.

Chin, K. 2007. The paleobiological implications of herbivorous dinosaur coprolites from the Upper Cretaceous Two Medicine Formation of Montana: why eat wood? Palaios 22:554–566.

Hellawell, J., and P. J. Orr. 2012. Deciphering taphonomic processes in the Eocene Green River Formation of Wyoming. Palaeobiodiversity and Palaeoenvironments 92(3):353–365.

Ji Q., Wu X., Cheng Y., Ten F., Wang X., and Ji Y. 2016. Fish-hunting ankylosaurs (Dinosauria, Ornithischia) from the Cretaceous of China. Journal of Geology 40(2):183–190.