The various and many ways to make a “cheek,” and the various facial tissues for which we have primary (preserved remains) and secondary (inferred) evidence for, in fossil sauropsidans. (These images are CC-BY-ND-NC. Please don’t take them without permission.)
The different structures in sauropsidan faces depend largely on what we should expect to find. For instance, we don’t expect to find extensive oral muscles (around the mouth) because these are novelties restricted to mammals. However, for others, such as beaks and thickened skin, or scales, etc., there can be primary (bony, remains) and secondary features of those remains that allow inferences for those tissues. Such as the relationship of muscles, their extent, etc. Some things we can’t know, such as the lability of the skin, or the size of flaps of skin, like the rictus which occurs at the corner of the mouth. For instances of skin shapes, texturing, and protuberances, like the cock’s comb, you need to look further afield, into impressions and such; you will not find them on the bones themselves.
So what happens when you have only the bones? Lately, despite my intentions to share a humorous and wholly fantastic illustration of a pennaceous ankylosaurid, a debate about inference arose. This got tied into the arguments for lips in theropods. It ends up, there’s a bit of a game of which side of a die you’re on involved in this facial tissues of sauropsidans debate.
I say a “die” here, because it’s a multifaceted issue, they’re all related, but each facet seems distinct — and some people seem more concerned to leave the other facets alone. But this is unwise.
Sauropsidan facial anatomy isn’t a bad thing. It should interest most. It’s the difference between known what Quinkana looked like, but also Captorhinus. The reconstruction of Najash, and Placodus. And of dinosaurs. Ever dinosaurs.
Dinosaurs, unfortunately, hold a larger sway to the arguments involved here, not emrely because of their diversity, but also their ubiquity, in media representation and the psyche. The largest land animals ever, the larger land predators and herbivores; the armored, the dome-headed, the horned; the vicious, ichor-drooling, snarling and roaring, phlegm-spitting monsters of cinema. Dinosaurs bring people to science; they are a gateway of sorts, because people want to know. I certainly wanted to, and because of them, I opened myself to wanting to know HOW I could know, and that brought me to biomechanics and the science of … well, their art.
Science is a deeply nestled part of dinosaurs’ being — figuring out how the largest animals moved, grew, lived, ate, mated, told one another apart, behaved, and even looked. We can examine evidence for colors (as long as those are red, grey, black, brown, white, maybe yellow) and the arrangement and laying of fuzzage, whether some were fuzzy or just bristly. Their metabolism, and laying habits, parental care or the lack of it. All grounded in raw science.
But when it comes to the life appearance of these, some tend to slip aside into the realm of fantasy, and this is fine — to a point. Before I continue, let me clarify something: Fantasy is defined here as “wondrous or strange fancies; imaginative conceptualizing.” It’s not intended to be initially pejorative; but I use it to caution that the term can slip into it: “supposition based on no solid foundation; visionary idea; illusion.” We must be careful to consider the limits of our imaginings. A constraint to our imaginings. For some, this is the argument from inference, and for others its an explicit following of some extant template, down to the color pattern and texture.
This is a fantasy. It’s imaginative, it’s whimsical, and it’s wholly unreal. Because it’s an ankylosaurid, an armored dinosaur, bereft of its namesake suit and dressed instead as for a parade in fine, pennaceous feathers. If I had more room, you’d see the club at the end of the tail sprouting enormous feathers. A fantasy. This was not done to jab at people who argued for fluffy ornithischians, but because whimsy struck me as I was discussing the topic with others, and because some time ago I drew a less refined version, here. Had science gripped me longer, I’d not have finished either piece. I balk at purely fantastic renderings: everything starts with the idea of the skeleton and of muscle, and of guts, and how they fit together. Only after does the drawing take shape.
But I could not resist, it seems. The head (as the feet, I suppose) are the only parts devoid of feathers, and because of this, they are unique in representing the idea of an ankylosaurid head … without the omnipresent ornithischian “cheek.”
(Note above: I’m probably being way less generous with the extent of the “lips” than I should be, as in squamates with them they’d cover the teeth utterly in lateral view. These are shown this way to compare the shapes of the tissues relative to the jaw.)
See, the problem with ornithischians with “cheeks” is that for a very complicated reason, “cheeks” seem obvious. Too obvious. It’s a peculiar structure, and the history of it is fairly basic, but followed by almost 100 years of iterated, memetic art. I’m still untangling it. Needless to say, at one point, a scientist made an observation that presumed, to his mind, that hadrosaurs and ceratopsians and yeah, even ankylosaurs, had this peculiar, muscular cheek. One piece of evidence was offered for it, the presence of inset tooth rows, considered an analogue to the presence of a cheek[n1], because mammals. So good, so far. Science followed, but before it could, he made another observation, that a cheek was necessary, somehow, for the management of food.
(That note isn’t an aside, scroll to the bottom for clarity.)
This immediately results in a red flag — because this claim is asserted, not demonstrated. Remember the addage: Show, not tell. But we are told this fact, and it is not proven. The earlier observation is seemingly sound, it follows logical deduction: mammals have inset tooth rows, and mammals have cheeks; ornithischians have inset tooth rows; therefore, ornithischians have cheeks. But this is a logical deduction. It’s a logical fallacy, and however that we use deductions to fuel our factual findings, it’s a starting point, not the end, of logic. There’s a missing component to this deduction, and that’s the isolation of the inset tooth row as a class unique to animals with cheeks. It’s also completely and utterly false, but first, the isolation of the class.
To logically isolate insets to cheeks, you have to demonstrate this without much in the way of variation: cheeks should tend to appear in animals with inset tooth rows, and animals with inset tooth rows should have cheeks. Neither is true.
For instance, mammals with cheeks almost all have inset toothrows, but only at the back of the jaws, in the location of the masseter muscle, which also correspond to the position of the ascending process of the mandible and the zygomatic arch. Further towards the tip of the jaws, the tooth rows, including amongst highly herbivorous taxa such as horses, cattle, and deer, the tooth row becomes flush with the lateral margins of the jaw. This is comparable to many mammalian carnivores, in which the external masseter complex is reduced generally in favor of larger emphasis on the temporalis muscle complex. In mammalian carnivores, the extent of the fleshy cheek is often much, much shorter, and the oral opening extends much further back. Yet both have equally flush dental rows and lateral margins of the jaws, with an inset towards the rear of upper and lower jaws. The correspondence of cheek to inset in mammals cannot be made consistent with “cheeks” but with size of the masseter complex.
In squamates, and especially in lacertilians, tooth rows become inset in a variety of carnivorous, omnivorous, and herbivorous taxa; but rather than be restricted to the rear, and despite flaring of the jugal arch laterally, the tooth rows become inset consistently along the tooth row. This is despite the extent of the upper teeth passing lateral to the lower teeth, or even beyond the edge of the dentary, as in some varanids.
In ornithischians, inset of the tooth row is often associated with a lateral “ridge” on the maxilla and dentary, the first associated with the jugal arch and which either tapers off or forms a slight shelf. The shelf occurs in ankylosaurs, but otherwise tends to vanish mid-maxilla. A ridge extends from the coronoid process of the mandible forwards, but becomes a shallow elevation and indistinct rapidly in most ornithischians, resulting in a laterally-bowed surface. Except in ankylosaurs, which lack this for the most part. You can see the difference here (excerpted from an earlier post on ornithischian cheeks):
Much ado is made of ornithischian “cheeks,” but ultimately, the model is flawed due to the incidence of lateral ridges, and inset tooth rows, in animals without (also excerpted from an earlier post):
As you can see, the lateral ridge of the mandible and even maxilla are clearly visible. Uromastyx lizards aren’t notable for having cheeks.
Most importantly, the presence of inset tooth rows occurs in a variety of squamates, including but not limited to, mosasauroids, varanoids, tuataras, and various lacertilians (such as above). The “inset”, rather, is a result of the lateral margin of the mandible or maxilla bowing outward, and forming a distinct facet between the alveolar face, which was likely covered in oral epithelium, and the ab-alveolar face, which was likely covered in non-epithelial skin, and the texture gradient likely differs markedly. In some dinosaurs, this transition is often marked by a distinct change in texture, so in ankylosaurids it is covered in rugose texturing or even dermal ossifications; whereas in some theropods the texturing is pitted and spiculate, as in tyrannosauroids and abelisaurids. So it is in crocodilians, but at a higher degree, as the skin likes close to the bone surface and is highly cornified, and in birds is covered by a rhamphotheca. Yet even in these taxa, the jugal flares laterally, and a “cheek” is present.
It is of further note that a mechanical, rather than integumental, relationship of the expansion of the lateral surface of tooth-bearing bones is responsible for “ridges” like those shown above, but modeling of the difference hasn’t been done; yet, shouldn’t be ruled out.
Some have questioned that because of the absence of a fleshy membrane to the sides of a jaw, an ornithischian might just lose its food. Starve to death, because it could keep food in its mouth. Which is patently ridiculous, as no animal requires such a feature to contain its food. Such membranes are used in some animals to help maintain high pressure within their mouths during suction. For example, flamingos and their AMAZING mouths possess an extensive rictus, a flap of highly vascularized skin that lies at the corner of the mouth and extends in most animals not much further anterior than the eye, but in flamingos reaches nearly to the naris. (That post is important, because it arose as a consequence of yet another online discussion in which this whole topic was raised, and raises both the warning flag that people need to get their definitions straight and stop corralling animals into explicit morphology categories. Animals are diverse and various in their shapes, and saying that X model precludes or excludes such a morphology of the skin from one group or another is a problem of overstepping the inference of what we have available.)
It seems questionable that any animal would starve were it a messy easier, with food falling out every which way but for a tissue on the sides of the jaws. As if sloppy spillage by dogs and crocs has brought them to extinction. Others raise instead the complexity of ornithischian “chewing,” the oral processing being a complex task itself as indicated by jaw mechanics, analysis of wear on teeth, and on the component tissues of those teeth. The argument goes, that the process of ornithischian dietary structure is so complex, that they would lose food merely by biting. This is because of the outward bevel of mandibular teeth would cause the outer portions of the bolus to fallow to the side during processing. A lip, it is then argued, cannot contain this bolus. No experimental data is offered to support this point; it is merely asserted. And this is the problem.
Certainly, it would be preferable to model a skull or two of an ornithischian with a model for “lips” and one for “cheeks” to see whether “lips” would be effective to contain food. But it’s not necessary, when one has living taxa to consider. It’s not necessary, even if one considers it true, so long as the the bulk of the bolus remains within the mouth during processing. The partial loss of food wouldn’t be enough to spur a membrane. This is shown below.
Part of the problem is the assumption of spectacular exception in ornithischian biting. They may be far more complex than has been assumed, but it is a far, far different thing to claim that they are so exceptional amongst sauropsidans. As if the diversity of lacertilian biting, with translational, propalinal, lingual-palatal occlusion, and palatal biting components, is somehow inferior. Different, yes; inferior, no. This ignores the wide diversity even amongst archosauromorphans for the various types of teeth without any implication of containment “cheeks,” especially in taxa like rhynchosaurs and Trilophosaurus. Transversely-arranged cusps imply translational dentary movements, whereas longitudinal furrows imply propalinal movement, and the complexity involved therein, especially in rhynchosaurs with their prominent coronoid processes in some taxa and flaring jugals, suggests that ornithischian dietary complexity is merely ornithischian dietary complexity. Much of which is restricted to tooth-tooth action (thegosis) and affecting wear. Mandibular rotational moments, variable positions, all minutely accumulate to demonstrate that ornithischians typically likely did NOT have simple orthal biting, nor did they precisely act as shears. Instead, a strike-slip model, of a vertical shear of the upper dentition into a basin formed on the lower teeth, then a lateral “slip,” acted along with a slight propalinal movement of a few tooth positions, allowing the efficient destruction of plant matter with minimal mandibular motion. My model above assumes potentially more radical motion than was necessary, by several degrees. So is the “food will fall out without cheeks” model viable because ornithischian ditary actions are “complex”? Again, asserted without evidence.
These models incorporated, in their earliest days of the hypothesis, a muscle complex that originated on the maxilla and inserted on the dentary, from lateral ridge to lateral ridge. This muscle is entirely fictitious. No such muscle exists in sauropsidans, as no muscle attaches on the lateral surface of the maxilla to insert on the dentary.
Not that long ago, Paul Sereno assumed this was true, however, in his reconstruction of the possible “cheek” in psittacosaurids:
The problem with this is that the muscle that forms the “pseudomasster” in this position in mammals is the m. infraorbital zygomaticomandibularis, and originates on the palatal and ethmoid on the internal surface of the skull at or near the infraorbital foramen. In rodents, this is especially large; and in rabbits, it is enormous, with a broad lateral fossa into which this muscle passes. And that fossa has been confused with the lateral fossa found on the psittacosaurid skull, and while correctly in glirans (rodents and rabbits) associated with a muscle and the tissues associated with innervating and vascularizing the oral tissues, in sauropsids it is quite a stretch to assume this is true.
In parrots, an actual m. pseudomasseter muscle exists, but this muscle is derived natally from a branch of the adductor mandibulae as it passes the jugal. Over time, as in some colies (no, not Lassie; but Coliiformes), this forms fibers that attaches to the jugal; and in parrots, this became well-developed and moved to cover much of the jugal ventral margin and dentary lateral surface. Parrots however have done many wonderful things with their cheeks, and deserving of their novelties, include neomorph bones of the cheek, and a novel ethmoid/palatal muscle comparable to the mammalian m. infraorbital zygomaticomandibularis.
Let’s work with some models for how the facial tissues might be arranged. The possibilities various for sauropsid “cheeks.”
Above is the basic bauplan, the one followed by the typical tetrapod prior to the invention of prongs, processes, and muscular features in theropsids and sauropsids. It’s what amphibians have, and what lizards and snakes have, and what turtles have, and what crocs and birds have. The muscular complex involves the pterygoideus and mandibular adductor suites and the pseudotemporal complex. These msucles work merely to close the mouth and variously pull it forward or backward during this process, or assist in bracing the mouth while opening. The small band of skin immediately behind the teeth and in front the jaw muscles is the rictus.
Mammals are a bit unique in that they developed a secondary set of muscles, some of it adapted from previous muscles, due to the loss of the posterior jaw bones and suspensorium to form the bony ear. Almost all those bones around the red in mammals are now tiny, absent, and the muscles associated with them. Instead, the dentary itself bears the insertion of muscles that descends from the jugal, palate, and ethmoid bones, called the masseter complex. These extend the relative position of the muscular “cheek” forward. The rictus is larger in comparison, on this mapping of these structures to a sauropsid skull.
Alternatively, squamates and ornithischian dinosaurs modified the “cheek” by developing a coronoid process which raised the attachment of several muscles on the jaw and thus the angle by which they approach the skull. This is also apparent in some turtles. Here, the “cheek” contains a bony component, and the rictus is pushed forward relatively. Oviraptorid theropods may also be said to compare to this model, due to the arching of the dentary and thus the position of the muscles during jaw closure. More on them later.
With a normal muscle complex, but a large rictus, we arrive at what may be deemed an adequate model for ornithischian “cheeks,” minus the muscles that some assume must be present. This is “adequate,” in that it involves known tissue relationships and assumes the model is true without inventing novelties, like muscles in rabbits that are then extrapolated to psttacosaurs. An extended rictus is the least problematic model for such a structure, if it can be assumed to be true, and I illustrated it far above or ornithischians as a possible means of retaining a bolus.
A common reconstruction of ornithischian “cheeks” is that the skin is smooth and continuous with the lateral surface of the rest of the head, and so covered in a consistent pavement of scales or scutellae. This structure gives the appearance of a smooth “cheek” that doesn’t imply a muscle underlying it, as shown above. Instead, the rictus is covered in squamous integument, which implies the tissue is stiffened. Here, a highly collagenous rictus forms a semi-rigid “cheek” without incorporating the more restricted use of a muscle. Problematically, this model is one of the least viable I might note, because no sauropsid bears extraoral integumental structures (feathers, bristles, scales, “fur”).Or one can instead involve larger, but more sparsely arranged bands of stiffening tissue in the rictus, allowing expansion and also durability for the processing of food. Because consider: the rictus in birds is a very thin, and not a durable structure prone to the abuse of being on the external surface, heavily innervated. In crocodilians, it’s covered in a cornified dermis, but only in the area where the rictus is exposed because of the gap between upper and lower jaws beneath the jugal.
These models suggest various ways to en-cheek a sauropsid, should one need to, but only two of them occur amongst living taxa. And those models are applicable across all sauropsidans, as none of them so far appear to demonstrate a unique reason to consider otherwise. Here’s Uromastyx, as shown above but again for emphasis.
Look closely at the upper jaw, and note the dental arcade. This lies not only well within the margin formed by the fleshy “lips,” with an appreciable gap to the sides, but the “lips” extend further than the teeth do. The same is true of the upper jaw. There are no muscles involved. When the jaws close, the upper and lower “lips” meet before the teeth do. This creates a closed oral margin prior to the teeth passing one another. And as an herbivore, uromastyx lizards are prone to the same issues as ornithischians. They bear similar dental wear, albeit with less dental complexity, and have as diverse jaw behavior, without the mandibular rotational capability. Do uromastyx lizards need “cheeks” to eat? Doesn’t seem so.
I’ve spent a considerable amount of time on the topic, covering the question of “cheeks” in dinosaurs various, not merely in ornithischians, but sauropods as well. Theropods involve the necessity of oviraptorid models. There’s some features in other sauropsids mentioned above, such as the mammal-like jaws of placodonts, that raise the specter of various tissue complexities in sauropsids. As Casey Holliday argues, it’s a problem to assume one tissue type fits all, but it is further a problem to assume the fantasy of inventiveness when one is faced with exceptionalism. There isn’t much room for inventing tissues out of whole cloth to solve problems that aren’t there. We have much more room to consider the absence of extraoral integument in theropods than we do the question of whether ornithischians could have “cheeks.”
[n1] The term “cheek” used here is problematic. On the one hand, it corresponds to the muscular component similar to a mammalian jaw; and on the other, it refers to the region of the face or skull below, around, or just behind the eye associated with jaw closing muscles, or a bony expansion of the suborbital rim. It has been used in tetrapod and fish description to refer to the region of the skull or hear around the jaw joint, or just the jugal bone. Lately, the term has taken on the use of any tissue around this region, but in the shape of the muscular, mammalian masseter muscle complex, and often with corresponding anatomy implied. The terms for components of the tissues in this region are and should be explicitly separated, but they correspond in general to the “cheek region,” alternatively the suspensorial or suborbital region, the “jowls,” etc. We’re a very mammal-centric species. “Cheek” is used here in quotation marks in order to separate the regions or structures in non-mammals from the “true cheeks” (or, cheeks) of mammals. While this argument may be deemed semantic — pejoratively so — the semantics of the issue involving homology and comparison make the separation of terms important. Without clear specification, this topic will become confusing very quickly.