The Emaciated Tyrannosaur – a Reply to Ford, 1997

Velociraptor mongoliensis jaw “fitting.”

There are few published works on the tight subject of whether theropods have “lips.” One group, of which Ashley Morhardt, Tobin Hieronymus and Larry Witmer are principles, have looked at the distribution of soft-tissues of the skull. One work, by Papp & Witmer (1998), is introductory to much which has yet to be published, speculating on the presence of “cheeks” in ornithischians; additionally, Morhardt’s MSc thesis (while at W. Illinois Uni with Matt Bonnan; now at Ohio Uni) was worked into a talk and abstract (Morhardt, Bonnan & Keillor, 2009) in which fleshy tissues are indicated by the numbers of foramina: reaching. Tyler Keillor’s chapter in the Burpee Tyrannosaur volume (on reconstructing Jane with full, fleshy “lips”) captured much of the present literature on this subject, including the “lips” and “no lips” arguments. Tellingly, the lack of firm data either way has made this debate problematic. But earlier than these works an additional paper of note receives little attention, and this is the most pointed work on the subject to date. I admit to having known of this (Ford, 1997) but only recently have I acquired this paper as it is in a non-digital format.

I like this paper for many reasons because it, unlike many casual treatments of the subject, attempts to tackle the argument in an empirical basis. Several of Ford’s arguments I’ve addressed previously, but here I will briefly re-address them. Ford introduces a new argument which has not been dealt with since, and this allows me to discuss this feature in the full argument. Ford’s arguments can be broken into seven pieces: 1) structure of the extra-oral tissues around the jaw, their function, and their need, wouldn’t have been all that important; 2) space needed for teeth within the lower “lips” would have been immense, thus unlikely, but also the jugal “cheeks” flare out really wide, which means there must be even more soft-tissue there; 3) the “problem of the overbite”, but I addressed that here; 4) teeth extend very far, and consequently contact the inside of the upper jaw or too far below the lower jaw, or punch holes in the palatal roof; and 5) the use of the ectopterygoid as a bony stop, against which the mandible rested when the jaw was closed.

Initially, I’d wanted to write a detailed rebuttal on these points, but it would take way too long. You’d be bored. And I’d be saying the same things over and over. See, the first major problem with Ford (1997) is that it uses some tyrannosaur skulls to make its case. Skulls are fragile things; they have lots of intricate parts, and bend easily. Tyrannosaur skulls, despite being sturdy bastions of bone-crushing, are hollow, both inside the head itself (lots of space for sinuses, air passages, the brain, etc.) and that means they can be compressed. The late Maastrichtian formations from whence Tyrannosaurus rex comes from has a tendency of distorting fossils. You have merely to look at skulls like Sue’s to see:

Skull of FMNH PR2081, “Sue,” Tyrannosaurus rex, in right lateral view. Skull is fully prepared and shows dorsoventral crushing resulting in a concave dorsal profile, but also left-right skewing. Photo by Geoffrey Fairchild, courtesy Wikimedia Commons.

Even more important, a jaw is a complex thing. It has a lot of parts, most of them moving. The jaw is bound to the skull by connections of muscle, ligament, tendon; the jaw joint is a series of soft-tissues and cartilage between the two sides. Really fragile stuff, jaw joints. But they get twisted about; skulls get crushed and bent out of shape. Used to be, every new Triceratops skull that came from Montana or Wyoming was a whole new species (or “genus,” like Stereocephalus). These different taxa were based on the merest of variances, the shapes of the frill and orientation of the horns, and the implied nature of missing parts because … because.

And Tyrannosaurus rex would be no different. We don’t have Tyrannosaurus rex alone in its little world of end-Cretaceous tyrants, no: we have a half score, all probably Tyrannosaurus rex. There’s little Stygivenator molnari and Nanotyrannus lancensis; ancient Manospondylus gigas, and Matt Martynuick’s interesting campaign to have us call rex that; there’s Dynamosaurus imperiosus, named by the same guy who named Tyrannosaurus rex. In the same paper. Relatively complete skulls are so few and well-known that you can read them off like you can read off the 11 (now apparently 12) Archaeopteryx lithographica specimens. AMNH 5027. Sue. Peck’s Rex. B-rex. Wankel-rex. Scotty. Stan. Black Beauty. Reasonable completeness means mounting, fame, and the wonders that come from looking at how distorted the skeleton is. The icon for this is Sue, in which the skeleton and skull were preserved twisted about. The mount at the Field Museum for FMNH PR2081 has a fake head on it, the real one lying at its feet, simply because the thing is so distorted.

One of its problems is the skull is skewed over to the side. The maxilla tilts over, and the lower jaw is skewed, the one side parallel to the maxilla and pushed so far up the teeth are pressing into the side of the skull and the upper teeth extend well past the lower jaw. And you know this isn’t natural because the other side isn’t like this. We have some reasonably complete tyrant maxillae, and ones that seem mostly undistorted. Some, like CMNH 7860, show a medial palatal process that juts off close to a right angle from the main maxilla but upwards a little, forming a decent vault; it was probably 20-30 degrees or so. In Sue, though, on one side this process is only slightly angled and is in line with the main maxilla, and the teeth of the jaw pass beyond the base of the palatal process, parallel to it. But on the other, because the skull is skewed, this process is perpendicular. This best illustrated below, the “natural” and the decidedly not normal mode.

Cross-section through the snout of Tyrannosaurus rex at the level of the maxillary antrum. A is generalized from AMNH FR 5027 and inferences from other tyrannosaurine skulls; its cross-section is not known. Degree of distortion in B is not shown, but generalized, from Brochu (2003). Stan approximated from A.

Ford (1997) relies heavily on another skull, BHIGR 3033, or “Stan.” Stan, unlike Sue, was not found necessarily all skewed up. Instead, the skull was relatively disarticulated but largely complete. Stan, however, has problems. The first is relatively noticeable: the mandible is shorter than you’d typically see in a tyrant skull: much shorter than the famous AMNH mount of 5027 (which remains intact and articulated), much shorter than in Sue (skew and all). But importantly, Stan’s jaws are mismatched. The one side is noticeably shorter than the other, so on one side the jaw is even shorter. You’d be forgiven for claiming tyrannosaurids have V-shaped jaws if all you really had to go on was Stan. The shortness of the mandibular rami, each side of the jaw, force them to approach at a narrower angle. The jaws of a “normal” tyrant skull like AMNH FR5027 have a noticeable inward curve to the front end, so much so the first tooth is somewhat medial to the second, which is slightly overlapped by the third.

Stan has several … abnormalities to the rest of the skull; the problems aren’t just in the jaw.

Cast of skull of BHIGR 3033, the Tyrannosaurus rex named Stan. A, skull and mandible in lateral view, showing overbite and position and relative shape of ectopterygoid. Hatching on the ectopterygoid shape shows ECP and JUG symphysis; black arrowhead indicates furthest point the jaw can adduct before the bone stops it. B, Skull in ventral view, showing foramina of premaxilla (gold arrowhead), distinct pits (white arrowheads), and slight channels or :”grooves” along the side of the maxilla (blue arrowheads). C, Both mandibular rami views from the left, showing extreme differences in length. In A and B, white arrows point to margins of the teeth where enamel ends, which can occur at half exposed tooth length; gold arrows indicate resorption pits, where replacement teeth begin to form, but these show no such teeth indicating the teeth are abnormally displaced.  A and B roughly to scale; scale bars in A and C in centimeters. Dots in A correspond to position and color of arrowheads in B, and roughly follow the curve of the palate. Images from WitmerLab.

See above: This includes irregular pitting on the inside of the palatal processes and “grooves” on the medial maxilla. Towards the front of the skull, the pits get deeper, until there are two very obvious holes right in front of the end of the vomers. This morphology (the pits; the foramina are poorly understood in tyrannosaurs due to irregular preparation or preservation, and may occur as incisive foramina or openings for sense organs through the secondary palate) is not normal for tyrants, and it’s not normal for theropods, that we know of. It may be normal for Stan, but that may well be due to the superbly short jaws, which would create malocclusion. In longer, “normal” jaws the first pairs of dentary teeth come up inside the premaxillae, which form an arch and which is supported by a shallow portion of the palate. This keeps the jaw from closing too far, if the first tooth is long enough. But in Stan, it is the jaw that isn’t, and this means the first dentary tooth is next to the first maxillary … on one side. It is next to the second maxillary on the other. The rest of the skull is a little skewed, too.

Hoisting Stan up as the model of tyrant skullness would be as problematic as doing so for Sue. They are too different from the expected norm. The problem of the overbite is that it isn’t necessarily a normal thing. A tyrant with normal “lips” and an overbite is gonna have more problems than the fact the upper and lower oral tissues don’t meet at the front. And this issue with Stan, by way of Sue and “normal” tyrant skulls, helps deal with some of the arguments of Ford (1997). It means the overbite, accommodation of teeth, are not an issue. But there’s more.

Ford (1997) takes the step of using the extended teeth of Stan as limits on jaw adduction. In Stan, most of the teeth are seemingly lose in their sockets without apparent replacement teeth pushing them out, a sign to Ford the teeth seem to be “naturally” high. But of all the teeth, only half of each exposed tooth is crown, the rest being root – and this tells me the teeth are “slipped.” It would take an x-ray of the jaws to confirm this. When the crown-root junction forms a limit to the “gumline” (no lower than it), the mere length of the teeth of the upper jaw would allow high exposure of this soft tissue, and that does pose some problems. But this is part of Ford’s “accommodation” problem, where there is too much tooth and too deep a bite for there to be soft-tissues, forming a gap around the teeth. Ford also pulls in the broad back of the skull, but really this isn’t much of an issue. Rexes had huge jaw muscles, which would themselves be covered in skin, regardless of how far out from the upper jaw they would bulge. As can also be seen above, the jugals do not flare as far as is implied: the space present merely indicates the m. adductor mandibular medialis, which attaches on the outside lower jaw above the ridge for attached of the m. pterygoideus ventralis,  was very large.

Ultimately, these aren’t problems on the presence of “lips” in tyrants, much less any theropod. They aren’t problems for most theropods, in fact. They might be a problem for tyrants, or just one species of tyrant, but that tells us nothing about the multitude. This is thus a case of generalizing the specific.

As I said, Ford (1997) has brought a novel argument to the table. This argument amounts to the function of the ectopterygoid as a “brace” against which the lower jaw rested as the jaw closed. As this argument goes, the position of the ectopterygoid shows how deep the jaw adducts. The first issue with this is the dual function of the ectopterygoid. In tyrants, the bone is hollow. In fact, it’s the most hollow cranial bone in tyrannosaurids, exemplified in Alioramus altai (Brusatte et al., 2013). The bone is thus somewhat of a balloon, built of a thin – very thin – membrane of bone. This means it’s fragile, and delicate. On top of this, the bone is oriented sideways, with a flange that hangs down but otherwise crossing from the palate/pterygoid to the jugal. If the lower jaw comes up against the bone, and the bone is supposed to stop it (mind you, this being a jaw that can produce over ten thousand Newtons of force (Bates & Falkingham, 2013), which is calculated at the back of the jaw, near where Ford (1997) argues the “stop” occurs) it is likely the bone would be somewhat damaged. The ectopterygoid’s role in the skull is two-fold, being the dual functions of forming a mediolateral brace in the skull, acting as a spring for transverse (not vertical!) forces. In tyrants, the bone is pushed forward of where it normally sits in other theropods, and is close to the contact for maxilla, jugal, and lacrimal. In short, it’s positioned at a crux of skull bones where sutures don’t fuse, allowing forces to be dissipated. The ectopterygoid is a spring.

But the other function is more important. The ectopterygoid is a waypoint for the m. pterygoideus dorsalis muscle. The ectopterygoid in tyrants is very large, and the pterygoid is relatively narrow, with the attachment site for mPTD being anterior and above where the ectopterygoid lies. This forces the muscle to pass over and behind the bone before it turns downward to meet the jaw. On top of this, there’s the muscles of the medial mandible, namely the m. pseudotemporalis complex and m. adductor mandibulae complex, which insert on the mandible behind the dentary/surangular contact.

The main problem with the use of the ectopterygoid as a “bony stop” is the muscles around the jaw stop it from ever getting close to the ectopterygoid. In short, the ectopterygoid is not there to block the mandible. That’s not to say it doesn’t stop the mandible from closing, just that its presence can stop the mandible from closing too far. The ectopterygoid is positioned on the ventral margin of the jugal immediately posterior to the maxilla/jugal contact, and nearly touches the maxilla. It arches somewhat upward in the transverse plane, but this provides little room for the jaw to move. When the jaw closes and comes close to it, the upper margin of the lower jaw barely passes the teeth of the lower margin. You get something that looks more like a monitor lizard than you get “classic” jaw positions:

UTA 13015, skull of Varanus acanthurus, the ridge-tailed monitor. Image from Digimorph.

Deeper adduction can occur when the ectopterygoid is dispaced, as it represents the only major limiter here; but adduction can be halted when the teeth are too long, and only malocclusion with the palate stops the jaw from adducting further, as seems the case in Stan.

Once you look beneath the surface of the arguments, it is difficult to determine the problems Ford (1997) argues are present. Mostly, that these features necessarily reject the premise of large, lizard-style sessile extra-oral tissues (i.e., “lips”). The irregular shape of upper and lower jaw seems to be one of the bigger issues, the position (or existence) of the ectopterygoid firmly a non-issue. Ford relies on tyrannosaurs to make his case, especially specimens (like Stan) that involve several problems, including slipped teeth and a strong overbite.

If we can reasonably argue that in a large or small nonavian theropod dinosaur without the “problems” of tyrannosaurs may have had distinct, extra-oral tissues that covered the teeth (no deep pockets, no wide “cheeks,” no tooth/palate occlusion, no irregular pits, and certainly no ectopterygoid occlusion), then why wouldn’t tyrannosaurs, necessarily. More to the point, if tyrannosaurs didn’t and this was affirmed (maybe, as has been illustrated, it was all croc-like), why wouldn’t this be peculiar to tyrannosaurs, but not relevant to other theropods.

Here’s an interesting point, one I don’t think many have considered. Applying the Extant Phylogenetic Bracket, we get a strong, first degree inference that all dinosaurs had no facial scales whatsoever. Living crocs lack them; birds lack them. Thus, extinct nonavian dinosaurs must have lacked them, but these are commonly illustrated nonetheless. We have some evidence that some dinosaurs had facial scales, or a complex, scute-like facial integument. Ankylosaurs, certainly. But why isn’t that an exception to the rule? Because we make assumptions that living crocs are an exception themselves, one convergent on living birds that facial skin lost its covering of scales; that older, extinct croc relatives kept facial scales, but at some point lost them. There’s some good data that points to reasonable hypotheses why both birds and crocs lost floppy facial integument the way they have: for both, the other tissues involved in the face are more important. Birds, it’s beaks; for crocs, it’s their habitat, as all living crocs are more or less sedentary sit-and-wait aquatic predators.

I’m willing to entertain the idea of exceptions, but they need to be based on better data. Not generalizations or misinterpreting the functions of some bones. But ultimately, we have to start with a null hypothesis, and that hypothesis arises from the most parsimonious explanation amongst sauropsidans, just as bone homologies must. Amphibians, the outgroup to amniotes, have fleshy extra-oral tissues, as do “reptiles” such as tuataras (despite having fang-like teeth or giant rostral hooks); other lizards, certainly, and if monitors are lippy, mosasaurs might also be lippy and we make no issue with that; snakes, definitely, though there are some snakes that can poke their teeth outside their lips (stiletto snakes, Atractaspididae, which are also among the most toothless of snakes, and are burrowers and feeders upon burrowers) while keeping the mouth closed.

The non-lippy sauropsidans (turtles, crocs, and birds) all give rise their their specialisations differently, so using them to enforce a baseline is unparsimonious. This argument also supports that the lineage from amphibians to mammals should indicate that all intervening animals should be implied to have “lips” – or, actual lips – including “pelycosaurs” like Dimetrodon — which has been done, thanks to this lovely illustration by the Royal Ontario Museum’s Danielle Dufault:

A lippy little bugger.

Tyrannosaurs aren’t a problem because if theropods would have their tissues, then tyrannosaurs would likely have to have some form of them. And not the half-assed way Greg Paul envisioned them, the tissues would need to accommodate teeth, the flare of the jugal, “gaps” in the jaws, etc., and there is no reason given that demonstrates otherwise. I’d even imagine that, instead of accommodating an overbite, the tissues would be irregular and deformed despite them: the overbite would be apparent in the flesh, too, and the first few upper teeth visible, noticeably.

I started this project on tissues by using Velociraptor and gave it a joker’s grin. Velociraptor doesn’t suffer the problems Tyrannosaurus has that I’ve noted in this post. Even though the ectopterygoid is further back, so is the attachment of the adductor muscles, and the shallower skull also means the muscles pass the ectopterygoid at a shallower angle; the cheeks don’t flare out; the mandible and upper jaw curve much the same way; the jaws largely fit neatly into the maxillary arcade. The null hypothesis is that sessile tissues surrounding each jaw and meeting and covering the teeth, barring direct indicators otherwise: crocs have closely-adhering keratinized skin and teeth that pass one another, preventing extensive soft tissues, whilst birds have beaks. This doesn’t mean nonavian theropods aren’t doing something yet even differently, but without direct indicators, we have to fall back to the parsimonious position, and that’s the null. Ford (1997) offers good data that reverts the null to a peculiar condition, especially in the light of arguments for indicators otherwise (Morhardt et al., 2009; Keillor, 2013). But the argument doesn’t hold up to the addition of more, better data. Whatever tyrannosaurs may be doing they’re doing the same thing many other theropods are doing, and that means extra-oral tissues in the form of “lips.”

Finally, I thank Tracy Ford very much for supplying a copy of the paper in print form, without charge and thus at his personal expense. Without this generosity, this post would not have happened. Tracy knew I’d look at this work in a very critical light, and his willingness to let me go at it is appreciated.

Bates, K. T. & Falkingham, P. L. 2012. Estimating maximum bite performance in Tyrannosaurus rex using multi-body dynamics. Biology Letters 8, 4: 660-664.
Brochu, C. A. 2003. Osteology of Tyrannosaurus rex: Insights from a nearly complete skeleton and high-resolution computed tomographic analysis of the skull. Society of Vertebrate Paleontology, Memoir 7 — Journal of Vertebrate Paleontology: 1-138.
Brusatte, S. L., Carr, T. D. & Norell, M. A. 2012. The osteology of Alioramus, a gracile and long-snouted tyrannosaurid (Dinosauria: Theropoda) from the Late Cretaceous of Mongolia. Bulletin of the American Museum of Natural History 366: 1-197. [PDF]
Ford, T. L. 1997. Did theropods have lizard lips? Southwest Paleontological Symposium – Proceedings 1997: 65-78.
Holliday, C. M. 2009. New insights into dinosaur jaw muscle anatomy. The Anatomical Record 292: 1246-1265.
Hurum, J. H. & Sabath, K. 2003. Giant theropod dinosaurs from Asia and North America: Skulls of Tarbosaurus bataar and Tyrannosaurus rex compared. Acta Palaeontologica Polonica 48 (1): 161-190.
Keillor, T. 2013. Jane, in the flesh: The state of life-reconstruction in paleoart. pp.:157-176 in Parrish, J. M., Molnar, R. E., Currie, P. J. & Koppelhus, E. B. (eds.) Tyrannosaurid Paleobiology. (Indiana University Press, Bloomington.)
Morhardt, A., Bonnan, M. & Keillor, T. 2009. Dinosaur smiles: Correlating premaxilla, maxilla, and dentary foramina counts with extra-oral structures in amniotes and its implications for dinosaurs. Society of Vertebrate Paleontology, Abstracts – Journal of Vertebrate Paleontology 29 (supp. to 3): 152A.
Papp, M. J. & Witmer, L. M. 1998. Cheeks, beaks, or freaks: a critical appraisal of buccal soft-tissue anatomy in ornithischian dinosaurs. Society of Vertebrate Paleontology, Abstracts – Journal of Vertebrate Paleontology 18 (supp. to 3): 69A.
Witmer, L. M. & Ridgely, R. C. 2010. The Cleveland tyrannosaur skull (Nanotyrannus or Tyrannosaurus): New findings based on CT scanning, with special reference to the braincase. Kirtlandia 57: 61-81.