Coelurosaur Ghosts!

Ok, so this isn't really news, but I thought it would be an interesting topic for the Goss to review: Coelurosaur ghost lineages. Rumors abound of coelurosaurs creeping their way back out of the Late Jurassic and Cretaceous where they belong, and into the murky shadows of the Early Jurassic or even Triassic. Nobody has ever gotten a glimpse of such specters, but little bits of evidence that can't be explained away so easily keep poking little, irritating holes in our picture of how the lineage that lead to birds came about. Spooky!

[Top image: Nothronychus by NTamura, lisenced. Modified by M.Martyniuk]

You all know coelurosaurs: those little (ok, not so little) carnivorous (but not all of them, or maybe even most) dinosaurs (in all likelihood). The first coelurosaurs were discovered in the Late Jurassic Solnhoffen and Morrison formations, familiar little critter like Compsognathus, Coelurus, Ornitholestes, and Archaeopteryx (thanks to phylogenetic nomenclature, birds are classified as coelurosaurs too), all found during the 19th Century. In the 20th Century, coelurosaur discoveries exploded. Turns out, coelurosaurs really flourished in the Cretaceous, with the dromaeosaurids ("raptors"), troodontids, oviraptorosaurs, ornithomimosaurs, therizinosaurs, alvarezsaurs, and tyrannosaurs pretty much monopolizing the theropod game. Despite the diversity of fossils and number of specimens, however, the evolution of these dinosaurs is a bit murky. We have a generally good idea of how they were all related on a large scale (Tyrannosaurs and coelurids are the most primitive, followed by ornithomimids, alvarezsaurids, therizinosaurids, oviraptorosaurs, deinonychosaurs and birds). The problem is the timing.

The most advanced coelurosaur lineage is Avialae, the birds, and their close relatives the deinonychosaurs. These groups reached their apeces in the Cretaceous, but we know they were around in the Late Jurassic thanks to Archaeopteryx. So how could avialan birds, which are more advanced than the deinonychosaurs, have been around before the deinonychosaurs?

The answer: ghost lineages. These are segments of the dinosaur family tree that we know must exist thanks to phylogenetic studies of relationships, but for which no fossils have been found so far. Often, fossils are found later to prove the relationship studies right. For example, there's a Late Jurassic troodont from about the same time as Archaeopteryx currently in press (nicknamed "Lori"). Lori proves the ghost lineage was right: the more advanced Archaeopteryx was not around before the more primitive troodonts, in fact they lived at the same time, and troodonts (and by extension, all the other Cretaceous coelurosaur groups) probably go back even further than that.

But how much further? The earliest definite coelurosaurs on record come from the Dauhugou Beds of China, which depending on who's doing a study this week, date to either about the same time as Archie or a few million years earlier. Coelurosaurs from here include Scansoriopteryx and Pedopenna. Pedopenna, interestingly enough, is a basal paravian, essentially what you'd expect the ancestor of troodontids and Archie to look like. Not too surprising. Based on this, and the presence of dromaeosaur-like teeth from the Middle Jurassic, we'd expect that the common ancestor of all coelurosaurs lived around the Middle Jurassic, about 170 million years ago.

So ghost lineages exist, and we still have a very poor record, if any, of the other coelurosaur groups over this time span. But in general, the ghosts seem to be appeased. Or are they?

Enter Eshanosaurus. When described in 2001, the scientists were baffled. This specimen, a long lower jaw and teeth, looked for all the world like a therizinosaurid. Not a therizinosaur in general mind, but a member of the advanced clade Therizinosauridae, more advanced than primitive, Early Cretaceous therizinosaurs like Falcarius. This would not be very surprising if it hadn't been for one thing: the bone hailed from the Early Jurassic, 60 million years before the previous, earliest known therizinosaur. Uh oh.

The Eshanosaurus jaw was found beneath a bonebed of the prosauropod Lufengosaurus. Mystery solved, you may say. Prosauropods and therizinosaurs have famously been confused in the past. Their skulls and dentition tend to be similar. Jim Kirkland thought so, and in a 2001 review said one aspect of the teeth, a mid-line ridge, screamed prosauropod. But he hadn't seen the specimen himself, and was going from a bit of unpublished goss (in scientific circles, when you use goss in an official, published paper it's called "personal correspondence", or pers. cor.). The authors of Eshanosaurus themselves, Zhao and Xu, had thought of the prosauropod thing, and tried every method to test that hypothesis, but could only rule it out.

[Image: Skull of Lufengosaurus, photo by FarleyKatz, licensed.]

Now for more pers. cor. The debate raged on for years, and still nobody has published a real rebuttal to the idea that Eshanosaurus is a ridiculously early therizinosaur. In fact, according to comments posted on Darren Naish's blog, Xu is sticking by his identification, and so are Richard Butler and Paul Barrett (the later a bona-fide prosauropod expert). Barrett has seen the specimen personally, and says it's no prosauropod. It's a therizinosaur.

Could it be, that therizinosaur existed throughout the entire Jurassic and even into the Late Triassic, and left no fossils (that we've found anyway)? That seems to be the case. Talk about a ghost lineage.

Incidentally, that same comments section discusses Larry Witmer's (2001, 2002) assessment of the infamous Triassic "bird" Protoavis. Likely a chimera made of a jumble of bones from all kinds of animals, Witmer seems adamant that, at the very least, the braincase of Protoavis is too coelurosaurian to ignore. And Witmer's opinion is nothing to be sneezed at, he's only the guru of dinosaur skull anatomy. Well, if therizinosaurs lived in the Early Jurassic, a Late Triassic coelurosaur doesn't seem so far fetched after all.

In conclusion, it looks like there's no good reason to think Eshanosaurus is not a coelurosaur, ghost lineages or no, and lots of good reasons to think it probably is unless other evidence proves otherwise. But what does all this mean for dinosaur evolution as a whole? If coelurosaurs, or even maniraptorans were stomping around the Late Triassic, shortly after dinosaurs first appeared, it would seem dinosaurs radiated much, much more quickly than anybody had thought, unless they have their own ghost lineage back into the mid-Triassic. Either that, or Steven Czerkas (2002) was right about Scansoriopteryx and it's mysteriously closed acetabulum (the hole in the hip for which an open state has always been a key characteristic of all dinosaurs). Maybe coeulurosaurs brached off of the dinosaurian tree much earlier than our cladograms would like us to think. Only more fossils will tell.

- Czerkas, S.A., and Yuan, C. (2002). "An arboreal maniraptoran from northeast China." Pp. 63-95 in Czerkas, S.J. (Ed.), Feathered Dinosaurs and the Origin of Flight. The Dinosaur Museum Journal 1. The Dinosaur Museum, Blanding, U.S.A.
- Witmer, L.M. (2001). "The role of Protoavis in the debate on avian origins"; pp. 537-548 in Gauthier, J., and Gall, L.F. (eds.), New Perspectives on the Origin and Evolution of Birds. Yale University Press, New Haven.
- Witmer, L.M. (2002). "The debate on avian ancestry: phylogeny, function, and fossils"; pp. 1-29 in Chiappe, L.M., and Witmer, L.M. (eds.), Mesozoic Birds: Above the Heads of Dinosaurs. University of California Press, Berkeley.
- Xu, X., Zhao, X. and Clark, J.M. (2001). "A new therizinosaur from the Lower Jurassic Lower Lufeng Formation of Yunnan, China." Journal of Vertebrate Paleontology, 21(3): 477–483.