Monday, August 31, 2015

How the ankylosaur got its tail club

Ankylosaur tail clubs are odd structures, odder than they are usually given credit for. They represent substantial modifications to two different skeletal systems – the endoskeleton, in the form of the caudal vertebrae, and the dermal skeleton, in the form of the caudal osteoderms. The centra of the caudal vertebrae lengthen but stay robust, and the neural arches undergo huge changes, such that the prezygapophyses, postzygapophyses, and neural spine become a robust, V-shaped structure on the top of the centrum, and which creates a tightly interlocking vertebral series with almost no flexibility. We call this the handle of the tail club. The osteoderms at the tip of the tail smush together and two of them become huge: although the tail club knob is small in some species, there are colossal knobs exceeding 60 cm in width. The ankylosaur tail club represents one of the most extreme modifications to the tail in terrestrial tetrapods.

Look at that thing. That is a weird thing.
(This is UALVP 47273, a really nice club that I studied for my MSc work on tail club biomechanics.)

One of the questions I became interested in during my MSc research on ankylosaur tail club biomechanics was how the tail club evolved in the first place. Most ankylosaurs with tail clubs are known from a relatively narrow slice of time right at the end of the Cretaceous, but when and where did the tail club first evolve? Did the stiffening of the tail occur before the enlargement of the tail osteoderms, or vice versa? Or did both changes happen at about the same time? This was a fun question to address during my PhD research, once I had a fairly well resolved phylogeny of ankylosaurids, and once I had looked at tons of ankylosaurid fossils.

So, how did the ankylosaur get its tail club? Well, based on what we see in the fossil record, it looks like the changes to the vertebrae predate the changes to the osteoderms – in other words, the handle comes first and the knob comes later. There is at least one ankylosaur out there that seems to have a tail club handle but not a knob: Gobisaurus!

Hello Gobisaurus! Many many thanks to my friend and colleague Sydney Mohr for preparing this awesome illustration of Gobisaurus for me.

Gobisaurus, a shamosaurine ankylosaurid, has a really nice complete tail club handle that is indistinguishable from other ankylosaurid tail club handles, but does not have a knob. And it's not just because the knob is broken off – it seems as though the last vertebrae in the tail are preserved, because they look very similar to the terminal vertebrae in a CT scan of a tail club from the University of Alberta collections. It's likely that Gobisaurus had osteoderms along the sides of the tail like we see in most other ankylosaurs, but it doesn't appear that there were osteoderms tightly enveloping the tip of the tail.

An even earlier ankylosaur seems to show some changes towards acquiring a tail club handle, as well. Liaoningosaurus, a basal ankylosaurid known only from a very small juvenile, has distal caudal vertebrae where the prezyapophyses extend about 50% the length of the adjacent vertebra. This is what we see in ankylosaurid tail clubs, but not in more basal taxa like Mymoorapelta where the prezygapophyses are much shorter. Liaoningosaurus is missing the tip of the tail and also lacks osteoderms on most of its body because it's a juvenile, so it's harder to say whether or not it had a tail club knob based just on the fossil alone.

I also did a cool and relatively simple thing with my phylogenetic tree to see if I could better understand the likelihood that some ankylosaurs without preserved tail material had a tail club handle or full tail club with a knob. Unsurprisingly, all shamosaurine and ankylosaurine ankylosaurids probably had a tail club handle. Liaoningosaurus is part of a basal polytomy of ankylosaurids, and it was a bit more equivocal whether or not any of these taxa was likely to have a tail club handle or not, partly because another basal ankylosaurid in this region of the tree, Chuanqilong, does not have modified distal caudal vertebrae.

All ankylosaurine ankylosaurids more derived than Pinacosaurus (so including things like Tsagantegia, Saichania, Euoplocephalus, etc.) almost certainly had a tail club knob, and shamosaurine ankylosaurids probably did not. Crichtonpelta, the most basal ankylosaurine, may or may not have had a tail club – we'll need more data to know for sure. There is amounted skeleton of Crichtonpelta at the Sihetun visitor center in Liaoning, and it is shown with a tail club, but it isn't clear whether or not this is sculpted or original material belonging to this specimen, and a full description of this material is necessary.

Gobisaurus and Liaoningosaurus both lived much earlier than the more familiar tail-clubbed ankylosaurs: Gobisaurus is no younger than 92 million years old, and Liaoningosaurus is about 122 million years old. The earliest ankylosaurid with a tail club in the fossil record is Pinacosaurus (from the Campanian), although there is a caveat to this: Talarurus, which is a bit older than Pinacosaurus, should have a full tail club based on its position in the phylogenetic tree, and while a tail club handle is known for this taxon, we haven't found a tail club knob for Talarurus. Talarurus is in kind of a weird spot phylogenetically, since it's from Mongolia but comes out as closely related to North American ankylosaurines, so I think it's worth keeping an eye on this taxon in the future – perhaps Talarurus is another taxon with only a handle and not a knob, which would fit a bit better with its chronologic position if not its phylogenetic position.

Regardless, the changes to the vertebrae of ankylosaurs, starting with Liaoningosaurus at least 122 million years ago and continuing on towards Gobisaurus about 92 million years ago, seem to have occurred long before ankylosaurs evolved a huge osteodermal knob at the end of the tail. Was a stiff tail as good a weapon as a full tail club with a knob? What drove the evolution of the knob so long after the evolution of a stiff handle? And why did ankylosaurs even evolve a tail club at all? Now that I've had fun investigating how ankylosaurs might have used their tails, and how the tail club evolved, the next question feels like it should be 'why' stay tuned for more tail club fun over the next year or so as I make an attempt at that question!

Read it for yourself! Arbour VM, Currie PJ. In press. Ankylosaurid dinosaur tail clubs evolved through stepwise acquisition of key features. Journal of Anatomy.

Sunday, August 23, 2015

Know Your Ankylosaurs: Everybody's in this Together Edition

So with all of those posts about ankylosaur taxonomy over the last few weeks, what have we learned about the evolution of this group? Over the course of my PhD research, I was able to identify a bunch of new characters that seemed useful for understanding ankylosaur phylogenetic relationships, including characters related to the cranial ornamentation, pelvis, and osteoderms. Although ornamentation and osteoderms can be tricky, they can still yield useful information if you're careful about how you construct the characters.

Here's a sampling of some of the new characters from the supplementary file that goes along with the paper. Long live rainbow ankylosaur skulls.

With all the new information, here's what the results of the analyses gave us (click to embiggen):

From this, we can take away some interesting points:

1. Gondwanan ankylosaurs are probably not ankylosaurids, but they also don't form a single evolutionary group. Whatever "Minmi" is, it's a very basal kind of ankylosaur, possibly outside the split between Ankylosauridae and Nodosauridae. It's a little bit harder to say what's going on with "Antarctopelta" (previously considered an ankylosaurid), and the Argentinian ankylosaur: both came out as relatively derived nodosaurids, but my dataset isn't designed to test the interrelationships of nodosaurids. I wouldn't be surprised if future analyses incorporating more nodosaurids and more nodosaurid-based characters found that these two species were closely related. It would also be interesting to know which lineage of nodosaurids (probably a lineage from North America) dispersed into South America in the Late Cretaceous in order to give us these two ankylosaurs.

2. There are nodosaurids in the early-mid Cretaceous of Asia, but not necessarily the ones that have been proposed previously. Zhongyuansaurus, for example, was first described as a nodosaurid but is instead a junior synonym of the shamosaurine ankylosaurid Gobisaurus. However, a couple of taxa, like Taohelong, Sauroplites, and Dongyangopelta, are recovered as basal nodosaurids. At present, there doesn't seem to be much overlap between Asian nodosaurids and ankylosaurids, which is interesting! Why didn't nodosaurids hang on in Asia once ankylosaurids evolved, when the two groups seem to have coexisted pretty happily in North America later on?

3. The ankylosaurids from the Late Cretaceous of North America represent a dispersal of Asian ankylosaurines sometime during the early-mid Late Cretaceous. The earliest ankylosaurine is probably Crichtonpelta, from China, and North American ankylosaurines are a deeply nested clade within Ankylosaurinae. We propose the new name Ankylosaurini for the North American ankylosaurines (plus Talarurus, for now).

Here, have some frowny-faced rainbow ankylosaurs. Ankylosaurs are very serious dinosaurs.

4. Where do ankylosaurids first evolve? Unfortunately, that question isn't easy to answer right now: down at the base of Ankylosauridae, there's a mix of taxa from North America and Asia. The position of Gastonia as an ankylosaurid tips the scales slightly in favour of a North American origin for the clade, but some analyses recover this taxon as a nodosaurid, so I think we should be a little cautious about this result. One step up the tree, we've got a polytomy of Aletopelta and Cedarpelta (both from North America) and Liaoningosaurus and Chuanqilong (both from China). Does Ankylosauridae originate in North America with something like Cedarpelta, with a subsequent migration and diversification into Asia? Or does this group originate in Asia with something like Liaoningosaurus and Chuanqilong, and Cedarpelta represents an immigration into North America?

5. And finally, what's going on with ankylosaurids in the mid-Cretaceous of North America? Why don't we find any ankylosaurids between Cedarpelta and the later ankylosaurins? Did 'endemic' North American ankylosaurids go extinct during that time? And why does Aletopelta have such a weird basal phylogenetic position despite being from the Campanian? I don't really have answers for some of these questions, although if you come to the Society of Vertebrate Paleontology meeting in Dallas this October I'm going to try addressing some of them. For now, Aletopelta remains the biggest ankylosaurid enigma to me – it really shares very few things in common with the other Campanian ankylosaurids and I doubt it is an ankylosaurin from the Asian immigration into North America – could it represent a distinctive lineage of North American ankylosaurids stemming from things like Gastonia or Cedarpelta, for which we just don't have other representatives at the moment? Or, is it a nodosaurid masquerading as an ankylosaurid because I haven't sampled the right taxa or characters?

Darn you Aletopelta, why must you vex me so?

As usual, I wind up with more questions than answers every time I try to figure something out.

That wraps up the summaries for this paper, but stay tuned for some more cool research coming out in the next few weeks, and some summer fieldwork recaps!

Arbour VM, Currie PJ. In press. Systematics, phylogeny and palaeobiogeography of the ankylosaurid dinosaurs. Journal of Systematic Palaeontology.

Wednesday, August 19, 2015

Know Your Ankylosaurs: Mongolian Odds and Ends Edition

I'm back in civilization, so let's get back to ankylosaurs! Ready Set Go!

Gobisaurus, Zhongyuansaurus, and Shamosaurus

Shamosaurus is a really interesting ankylosaurid from the Zuunbayan Formation of Mongolia. Unlike later ankylosaurids, it still has a relatively long snout like you see in basal ankylosaurs and nodosaurids, and it lacks the distinctive tile-like skull ornamentation of ankylosaurs like Euoplocephalus or Saichania, instead just having a granular, pebbly texture on the skull surface. Gobisaurus, from the Ulansuhai Formation of China, is nearly identical in appearance, and only a few features distinguish these two taxa, namely the length of the tooth row relative to skull length and the orientation of the pterygoids. (Indeed, I think you could make an argument for subsuming Gobisaurus into Shamosaurus as Shamosaurus domoculus, but I'm generally reluctant to start making new combinations given that generic separation is pretty arbitrary anyway.)

Shamosaurus and its too-cool-for-school cervical half rings, on display in Moscow.

Gobisaurus and Shamosaurus are sister taxa; the name Shamosaurinae was proposed at one point and there's no reason to discard it at present even though it only contains two taxa. Shamosaurinae is the sister taxon to Ankylosaurinae. I also identified one new character that links Gobisaurus and Shamosaurus together which isn't present in other ankylosaurids: both taxa have a distinctive groove on each premaxilla, the purpose of which is unknown but there you go. There have been some suggestions that Cedarpelta (from North America) is also a shamosaurine ankylosaurid, and while I find the overall morphology of Cedarpelta to be pretty compelling for placing it in a clade with Gobisaurus and Shamosaurus, I didn't recover it with those taxa in my analysis (it came out more basally-positioned). However, I wouldn't be surprised if Cedarpelta winds up in Shamosaurinae at some point in the future as we find more specimens of both it and Gobisaurus and Shamosaurus.

Zhongyuansaurus was originally described as a nodosaurid ankylosaur partly because of its long snout, but it's indistinguishable from Gobisaurus (except for being smashed and flattened). The holotype is also a subadult (or at least not fully skeletally mature), since some of the cranial sutures are still visible towards the back of the skull. There are some interesting things going on with the postcrania of Zhongyuansaurus, but that's a story for a few weeks from now so STAY TUNED NO SPOILERS IF YOU'VE READ MY THESIS.


Of all of the more obscure ankylosaurs I looked at during my PhD, Tsagantegia might be my favourite for being the most surprising in person compared to what I had read about it. Tumanova included a line drawing of the specimen in her original description, which has been oft reproduced, but interestingly it doesn't really do justice to the original specimen (despite being a pretty nice drawing). The line drawing shows a long-snouted ankylosaur with amorphous cranial ornamentation, not dissimilar to Shamosaurus, but with a wider premaxillary beak more typical of later ankylosaurs. In person, however, the skull has distinct cranial caputegulae like we see in Euoplocephalus and Ankylosaurus! It's a pretty cool ankylosaur and I think it's probably really important to understanding the dispersal of ankylosaurs from Asia into North America and the diversification of ankylosaurids in the Campanian-Maastrichtian of Asia, but it's really hard to pin down the age of the Bayan Shiree Formation, and we don't have any postcrania for this taxon. I'm sure I'll be revisiting this guy in the future.

Heck yeah Tsagantegia!

Here it is again but in a more different view!


Way back when I originally started this blog in 2010, I had travelled to Korea to spend some time working in the Hwaseong paleo lab preparing Talarurus material and generally studying the ankylosaur material they had collected from the Gobi. Talarurus, like Tsagantegia, is also from the Bayan Shiree Formation but is clearly distinct. The holotype skull has very subtle cranial ornamentation that takes the form of small cones, rather than flat hexagonal tiles like Euoplocephalus, or bulbous pyramids like Saichania. Weirdly, this configuration is also present in the North American taxon Nodocephalosaurus – either this ornamentation style has convergently evolved, or, as I recovered in my analysis, these two taxa are closely related despite being fairly widely separated geographically and temporally. This is another ankylosaur that I'm sure we'll talk about again.
Talarurus butt in Moscow. The skeleton on display is a composite of several individuals from the same locality, and the skull is totally sculpted and a bit out of date.

Here's the holotype skull, with its weird, weird ornamentation.


I've talked about Saichania fairly extensively here last year, but there were a few new things added in this most recent paper: Tianzhenosaurus and Shanxia (both from China) are, most likely, junior synonyms of Saichania, making this the most geographically widespread of the Asian ankylosaurids. Tianzhenosaurus has a nearly identical cranial ornamentation pattern when compared to Saichania, and I couldn't identify any differences that were outside of the usual ornamentation pattern variation we see in something like Euoplocephalus. Shanxia is known from the same formation but from a less well preserved skull, but the morphology of the squamosal horn is consistent with that of both Tianzhenosaurus and Saichania and therefore it probably represents the same taxon.

Next up: what's the big picture here, anyway?

Friday, August 7, 2015

Know Your Ankylosaurs: North American Odds and Ends Edition

I've covered many of the North American ankylosaurs in my previous papers and blog posts. In 2013, I argued that what we thought was Euoplocephalus was more likely 4 taxaAnodontosaurus, Dyoplosaurus, Scolosaurus, and Euoplocephalus proper. Then in 2014 we described a newankylosaurid, Ziapelta, from New Mexico. There are a few other taxa that had previously been proposed to be ankylosaurids, so let's take a look at them here.

Aletopelta, Stegopelta and Glyptodontopelta
Aletopelta is one of the more tantalizingly enigmatic ankylosaurs from North America. It's from a weird place – California – which may have been much further south 75 million years ago compared to its current position. It was also found in marine sediments, and the decaying carcass had formed a little reef, with oysters encrusting the ribs. The only known specimen of Aletopelta is relatively complete, all things considered, with the osteoderms in situ over part of the pelvis, the legs partially articulated, and with various odds and ends like osteoderms and vertebrae. Unfortunately, the ends of the bones are often chewed apart, and some of the material is a bit hard to interpret.

Here's the articulated pelvis and hindlimbs, and some other armour pieces, on display at the San Diego Museum of Natural History.

Regardless, Aletopelta is a very interesting ankylosaur. It has an unusual osteoderm morphology over the pelvis, with small hexagonal osteoderms closely appressed to each other. Ankylosaur pelvic armour seems to come in two major flavours: fused rosettes, like we saw in Dongyangopelta and Taohelong (and perhaps most famously in Polacanthus), and interlocking hexagons, like in Stegopelta, Glyptodontopelta, and Aletopelta. Tracy Ford suggested that ankylosaurs with these hexagonal pelvic shields might represent a clade (dubbed Stegopeltinae) of ankylosaurids. Glyptodontopelta has since typically been interpreted as a nodosaurid, as has Stegopelta, but the most recent interpretation of Aletopelta was that it was an ankylosaurid. In the revised phylogeny in my new paper, we found Stegopelta and Glyptodontopelta as nodosaurids, but Aletopelta as a very basal ankylosaurid. However, although Ford and Kirkland reconstructed Aletopelta with the typical ankylosaurid tail club, I don't think that it possessed one: the preserved distal caudal vertebrae don't show any of the lengthening or other modifications that are characteristic of ankylosaurid handle vertebrae.

An updated restoration of the known elements in Aletopelta - the main differences between this and Ford and Kirkland's reconstruction are the absence of a tail club, and uncertainty over what the head should look like.


Cedarpelta is an important taxon for understanding the biogeography and evolution of ankylosaurids, and I wish we had more specimens! I don't have many new comments to add about this taxon, since Ken Carpenter published a great description of the disarticulated skull back in 2001. Cedarpelta has been interpreted as a shamosaurine ankylosaur, as a relative of taxa like Gobisaurus and Shamosaurus (which I'll talk about in the next post) from Asia, and thus may point towards a mid Cretaceous faunal interchange between these two continents. In our revised phylogenetic analysis, we didn't find Cedarpelta as the sister taxon to either Gobisaurus or Shamosaurus, but it does come out as a basal ankylosaurid in their general neighbourhood, and I honestly wouldn't be surprised if future analyses or new taxa show support for it as a shamosaurine ankylosaur after all.


Nodocephalosaurus! What a fun ankylosaur. It's really quite unlike the other ankylosaurids from North America, which typically have flat, hexagonal cranial ornamentation. Instead, Nodocephalosaurus has bulbous, conical cranial ornamentation. Bulbous cranial ornamentation is typical of Campanian-Maastrichtian Mongolian ankylosaurs like Saichania and Tarchia, but in those taxa the ornamentation is pyramidal rather than conical. The front end of the snout in Nodocephalosaurus is also unusual, because there's no obvious narial opening and instead the ornamentation has a stepped appearance. Hopefully better specimens with more complete snouts will resolve this weird morphology. I've also reinterpreted the position of the quadratojugal horn compared to Sullivan's original figures – the horn should be rotated forward so that the bottom margin of the orbit is complete.

Nodocephalosaurus holotype skull in dorsal and left lateral views.


I don't have much to say about Tatankacephalus because I didn't look at the original material myself, but the previous phylogenetic analysis by Thompson et al. recovered it as a nodosaurid rather than an ankylosaurid as originally suggested by Parsons and Parsons, and we found the same result. Overall, Tatankacephalus is VERY similar to Sauropelta, so this is perhaps not surprising.

Up next: More odds and ends, but after I return from Utah!

Arbour VM, Currie PJ. In press. Systematics, phylogeny and palaeobiogeography of the ankylosaurid dinosaurs. Journal of Systematic Palaeontology.