One of the great things about conferences is that you do get to see new work coming though. As a researcher you are generally familiar with major projects and the research themes of your colleagues and collaborators, but people entering the field for the first time, especially PhD students, can spring a surprise. At the Beijing Flugsaurier, I was delighted to meet W. Scott Persons IV who was there talking about pterosaur tails, but is looking at dinosaur tails primarily for his thesis under Phil Currie. Scott recently had a great paper published about the structure of tyrannosaur tail musculature and was kind enough to pen this for me to explain more. The photos are Scott’s property with the exception of art generously on loan from Scott Hartman. Right onto the tails:
The holiday season is here, and for me that means making a yearly migration from grad school in Alberta back to my home in sunny North Carolina. It will also mean slogging through predictable yuletide conversations with my visiting relatives: “So, you’re still going to school in Canada? It’s cold up there, aye?” Followed by: “Still into dinosaurs, huh? What about them are you studying?” I’ll say that I’m doing a project on the evolution and functional morphology of carnivorous dinosaur tails. After that, there’ll be polite smiles, an awkward silence, and finally: “Well . . . isn’t it interesting how specialized you can be, nowadays?”
I think my extended family members are always expecting and hoping that I’ll say my research is on a sexy topic, like mass extinction, the origin of birds, or at least something related to jaws or claws, and I’m always left feeling a little bitter by their disappointment. Of course, my relatives will be right. My research project is extremely specialized and several aspects of it would not have been possible ten years ago. But, investigating the posteriors of theropods (the group that includes the bipedal carnivores, like Tyrannosaurus and Velociraptor) is nontrivial, and the results have broad implications.
Mammalian Bias
We mammals suffer from inherent tail prejudice. Big modern terrestrial mammals, like elephants and rhinos (which we often use as analogs, when thinking about dinosaurs), have short, scrawny, flyswatter tails, and our own species (with nothing but a vestigial nubbin) is at the furthest extreme. As a result, when considering dinosaur behavior and locomotion, there has been a tendency to not give tails their due.
The definitively unimpressive tail of an African elephant.
Pick up a kids book on Tyrannosaurus (of which there are many), and, if the function of T. rex’s tail gets mentioned at all, it will probably be in the context of what I call the “Seesaw Explanation”. That goes something like this: T. rex was built like a seesaw, with its two legs acting like the central fulcrum (about which the two teeters totter). Sitting on one end of the seesaw was the weight of T. rex’s big head, tiny arms, and torso. On the other half was its big tail. According to the explanation, T. rex needed a big tail to keep it balanced, because, without it, the seesaw would have too much weight on one end, and the ‘tyrant lizard king’ would fall flat on its face!
The classic Seesaw Explanation. I have no idea who the two kids are, but they sure are cute. (Tyrannosaurs image courtesy of Scott Hartman.)
But consider this saurian seesaw from the perspective of Darwinian economics. The tail of T. rex really was big. It accounted for more than half of the animal’s total length, and it was made of muscle and bone, which are metabolically expensive materials for an animal to produce. It isn’t that the Seesaw Explanation is wrong — the horizontal body position of Tyrannosaurus and most other theropods did give them a seesaw-like build, and the tail of T. rex was probably important for balance. But the explanation is incomplete and implies that carnivorous dinosaurs were so biomechanically maladapted that they needed to carry around an enormous counter weight, just to keep from toppling over. In fact, the tails of theropods were much more than dead weight, and (as the popular press has become found of hearing me say) there was more than junk in T. rex’s trunk.
Reptilian Perspective
Unlike modern mammals, most modern lizards and crocodiles have, undoubtedly, a superior appreciation for the benefits of a robust tail. The first step in my theropod tail study was to perform tail dissections on a series of modern reptiles. I began with a caiman (a South American crocodilian), and before even putting it under the knife, I was surprised by its tail morphology. The caiman, like all crocodilians and most other reptiles, has a tail that is really girthy just past the base of the hips. The tail literally buldges out and is wider than the hips themselves. I was expecting a tail shape similar to how the tails of theropod dinosaurs are commonly depicted – that is laterally compressed and tapering in girth smoothly from the hips.
Figure 1. Tyrannosaurus dorsal silhouette with traditionally thin tail (A) vs. modern Alligator dorsal silhouette with beefy tail (B).
As I cut into the caiman, I discovered what was responsible for this tail base budge: a single muscle, called the M. caudofemoralis. (The basal bulge of many crocodilians is often further exaggerated by a layer of fat, but my caiman only had a thin fatty layer). The reptilian M. caudofemoralis is unlike any tail muscle we mammals have. [Although, some mammals, including cats, do have a tail muscle that goes by the same name, the identically-named muscles are altogether unrelated (muscle terminology is confusing that way).] The reptilian M. caudofemoralis intimately ties the tail to locomotion, and it is better to think of it as an out-of-place limb muscle, rather than an axial or spinal-column muscle. The M. caudofemoralis usually extends less than halfway down the tail. Along its run, the muscle is attached to the tail vertebrae, and it is anchored, via a tendon, to the femur (the upper leg bone). When the M. caudofemoralis contracts, it swings the hind limb backwards.
The M. caudofemoralis of the caiman revealed.
The insertion of the M. caudofemoralis onto the fourth trochanter of the femur, and the auxiliary tendon that inserts at the knee joint.
If you think for a moment about how you run, you’ll realize (if it hadn’t already occurred to you) that it’s the force of the backwards strokes of your legs that propel you forward. The same is true of most reptiles, so the M. caudofemoralis makes a direct contribution to the locomotive power stroke. And not a small contribution ether! The M. caudofemoralis is usually the largest and the single most important femoral retractor. It really is fair to say that most reptiles are primarily tail propelled.
For those of you thinking phylogenetically, many modern birds also have an M. caudofemoralis, but, along with the rest of the tail, the muscle has become greatly reduced (probably relating to the evolution of forelimb powered flight and the need to minimize weight).
On to Dinosaurs
This raises the question: “Did dinosaurs have an M. caudofemoralis?”, followed immediately by: “How could you tell?” Fortunately, the M. caudofemoralis leaves some telltale signs. Most obvious is the fourth trochanter on the femur. The fourth trochanter is a usually prominent crest of bone that served as the femoral attachment site of the M. caudofemoralis. Being mammals, you and I don’t have femoral fourth trochanters, but most modern reptiles do. As it turns out, so did most dinosaurs.
The femur of an Alligator (A) and a Tyrannosaurus (B). Brackets demark the fourth trochanter of each. (Images not to scale.)
Just noting that most dinosaurs had an M. caudofemoralis is nothing new. The big fourth trochanters are so obvious that the great anatomist Louis Dollo pointed out the presence of a M. caudofemoralis in Iguanodon way back in 1833. More recently, biomechanic Steve Gatesy published several papers discussing the muscle in theropods. What wasn’t known is just how big the muscle got. Unfortunately, the size and shape of the fourth trochanter is not an indicator of the size and shape of muscle that attached to it. Did T. rex have a tiny M. caudofemoralis, like modern birds? Or did it have a beefy budging M. caudofemoralis like crocodiles?
To come up with a method for estimating tail muscle mass, I took measurements of the skeletons and muscles of my dissection specimens. Next, I began work on a computer, using the digital modeling software Rhinoceros®. I created 3-D models of the reptile tail skeletons, and, using the muscle attachment sites as a guide, I digitally modeled each tail muscle over top of the skeleton. The computer was able to give me a volume estimation for each muscle, and, because muscle has a known and fairly consistent density, I could use the muscle volume to calculate the muscle mass. Then, I could check the accuracy of these estimated masses with the true masses measured during the dissections. Fortunately, once you’re familiar with the muscle arrangements and know what you’re looking for, tail muscles have a consistent size relative to specific portions of the skeleton. The modeling techniques developed by the project could estimate (with +/- 6% accuracy) the mass of the M. caudofemoralis and other tail muscles based only on the skeleton for a variety of tail morphologies, from crocodilians to chameleons.
The next steps were straightforward enough: visit museums, measure some dino tails, digitally model the tail skeletons, and estimate the muscle masses, but the end results took me by surprise!
Surpassed Expectations
Just from having done the preliminary modeling and the dissections, it was readily apparent to me that the tail of Tyrannosaurus was proportionately no less beefy than a modern crocodile’s, but, when I calculated the final modeling results, T. rex’s tail turned out to be significantly beefier. In a crocodile, the M. caudofemoralis accounts for roughly 2-3 % of the total body mass (obviously that number varies with the individual, age, and species). In T. rex the M. caudofemoralis was estimated to comprise somewhere in the neighborhood of 9-10 % of the total body mass (although Tyrannosaurus total body mass estimations are a tricky subject). Looking just at the tail muscles, the M. caudofemoralis of a croc makes up roughly 40 % of the total tail musculature. In a T. rex, it made up nearly 60 %. I also estimated exceptionally large M. caudofemoralis masses for other theropods.
Digital model of the tail of Tyrannosaurs. The M. caudofemoralis is shown in red.
Such abnormal mass estimations merit explanation. Naturally, the first thing I checked was that I hadn’t made some measurement or modeling error, but everything was in order. Upon close inspection, there are a number of theropod tail adaptations that facilitate an enlarged M. caudofemoralis, but the primary one, and the one most important in the case of T. rex, has to do with the position of the caudal ribs. In modern reptiles, the anterior tail vertebrae support small ribs that stick out transversally. The M. caudofemoralis is positioned directly below these ribs. The same arrangement was true of theropod dinosaur tail vertebrae. Except, in most theropods, the ribs were attached much higher on the vertebrae (anterior caudal ribs of most theropods are attached to the neural arch, rather than to the centrum). This left more room below the ribs to be filled by muscle, and the only muscle in this position was the M. caudofemoralis. Elevated caudal ribs are present on even primitive theropods, like Coelophysis and Herrerasaurus, so it seems theropods raised their ribs to allow the M. caudofemoralis to expand early in their evolutionary history.
Tail vertebra of the theropod Allosaurus (A) compared to the tail vertebra of a modern lizard (B). Arrows point to the caudal ribs.
Thin-tailed illustrations of most theropods couldn’t be more wrong. T. rex and company had powerful rear ends!
Hightailing Theropods
So, we have taken a little kid off one end of the seesaw and replaced him/her with a big kid. So, what? Well, remember, the new big kid is not some pudgy couch-potato, the kid is a junior Swarchenegger – he’s all muscle. Recognizing the true mass of T. rex’s tail muscles is important for considering questions relating to the animal’s athleticism. Again, the M. caudofemoralis is the biggest contributor to the locomotive power stroke . . . and it just got a lot bigger.
Ironically, I think the reason theropods have been drawn with thin laterally-compressed tails is because they look more aerodynamic with thin tails and appear superficially faster and more athletic. In reality, a lean-tailed theropod would be intrinsically less athletic than one with a beefy tail. A bigger M. caudofemoralis could support greater maximum running speeds and greater endurance. This is true not just in T. rex, but in the majority of theropod dinosaurs.
Just how fast was T. rex? Well, that’s a hard question, and it will take more than the tail to tell, but I think speed estimations that only put T. rex on par with similarly sized modern elephants are missing a big piece of the puzzle. I’ll say to you what, in a few short weeks, I’ll be saying over-and-over to my relatives: When you think about T. rex and its tail, don’t think of a seesaw. Instead, think of a souped-up Volkswagen, because it’s what was in the trunk that provided all the locomotive power.
A Volkswagen and a Tyrannosaurus with a tail of appropriate beefiness (courtesy of Scott Hartman).
W. Scott Persons, Philip J. Currie. The Tail of Tyrannosaurus: Reassessing the Size and Locomotive Importance of the M. caudofemoralis in Non-Avian Theropods. The Anatomical Record: Advances in Integrative Anatomy and Evolutionary Biology, 2010; DOI: 10.1002/ar.21290
Dave Hone's Archosaur Musings 
Thanks for a great article Scott, really informative and well explained.
Just in time for your new Tarbosaurus as well Matt.
Very interesting. What about Ostrichs and Emus etc. How do they run? Have they developed alternative musculature to compensate for the loss of their tails?
Yup, now with more ‘junk’! Thanks Scott!
To answer David Stern’s question (and it’s a really interesting question). The M. caudofemoralis is often absent entirely in ratites. Ratites and other flightless birds (that are capable of running) have developed the “Groucho running” style. This requires less femoral retraction and places greater emphasis on knee flexors. So, the hamstring muscles have become more important.
— Scott
Thanks from me too… great walk-through!
The photos are really helpful too – take it from someone whose been trying to model this muscle from image slices.
Dave: can I make a request?
Inerview Scott Hartman! He seems to be a rare case (such as Greg Paul and Mark Witton) of being a scientist and artist in one person.
Really well-written article, thanks very much. A new Scott Hartman illustration too!
This is a great response to those estimates downplaying a high(or even moderate)-speed T. rex based on the excessive leg muscle mass that would be required. If a lot of the necessary muscles were in the tail, less of a problem.
Looking at T. rex as a teeter-totter, if the tail is heavier doesn’t that mean the head,etc. (the parts anterior to the legs) need to be bigger than previously thought?
Fascinating article, Scott. Your study pretty much resolves something that has bugged me for a while – life depictions of moderate to large theropods often look front-heavy to me, like they have to keep moving forward to stop from toppling over.
I know that the tail being longer would counteract this to some extent, and that the poses used are often dynamic (eg lunging forward), but it never seemed enough. (I had even done some back-of-the-envelope calculations on various representations of T. rex and it’s c of g would often be forward of it’s feet when standing still). A beefier tail addresses my concerns in this regard, and it’s implications regarding speed are also very interesting.
BTW, Hartman’s green T. rex looks much more balanced to my untrained eye.
Amen! Especially when the front-heavy drawings were in books, NEXT to the “teeter-totter” description!
“T. rex needed a large, powerful tail to balance it’s massive head”, next to a drawing with a waif of a tail that wouldn’t even balance its massive ribcage.
Super-interesting article. Thanks!
Great post and compliments my current thoughts regarding how the basic theropod body plan is continually underestimated as an extremely successful and adaptive lifestyle and not simply one that prohibited slow speeds and certainly did not induce theropods to continually fall flat on their faces either!
Great article Scott. I wouldn’t downplay your contribution to paleontology around your family. Frankly, bird origins are over publicized, and (IMO) a little boring. Your study, on the other hand, is likely to cause a major change in the way dinosaurs are portrayed in future documentaries. No more wispy tailed “Walking with Dinosaurs” style sauropods and theropods. That alone should be enough to impress your family.
Excellent and interesting article.
This is great! I’m trying to design fictional creatures, and I want a realistic base for them. Your easy-to-understand explanation will keep me from perpetrating the thin-tail misunderstanding in my writings. One of my main “races” is very dinosaur-ish, and I don’t want to come off as another idiot writer who didn’t bother with real science or research. Thank you for saving the world from sloppy character design!
Does this also apply to non-theropods?
Does it mean that most theropod tails would lash back and forth as they ran? That seems slightly wasteful of energy, though I suppose it could help with balance.
So what about oviraptorids and their reduced tails – do we assume they went the same way as emus? And do we know anything about the other maniraptors (I’m particularly thinking of dromaeosaurs and their famously odd tail design)?
Oh, boy. These are some serious questions.
“Does this apply to non-theropods?”
Like the majority of modern reptiles (or at least those with legs), the majority of dinosaurs probably did have big caudofemoral muscles (although caudofemoral muscle mass does appear to have been reduced in Neoceratopsia and Ankylosauria). However, in most theropods these muscles were exceptionally large.
“Does it mean that most theropod tails would lash back and forth as they ran? That seems slightly wasteful of energy, though I suppose it could help with balance.”
I would guess the opposite: theropod tails probably did sway as the animals ran, but, far from being energetically wasteful, they may have been massive stores of elastic energy.
“So what about oviraptorids and their reduced tails – do we assume they went the same way as emus? And do we know anything about the other maniraptors (I’m particularly thinking of dromaeosaurs and their famously odd tail design)?”
Funny you should ask! I currently have two papers in review — one on oviraptorosaur tails and one on dromaeosaur tails. Because these papers have not yet been published, I hesitate to give your question the full response it deserves. Here is what I will let slip:
Most oviraptorosaur tails are not as reduced as is commonly thought. Oviratorosaur tails are short (anteroposteriorly), but the portion of the tail that was lost is the region posterior to the tapering of the M. caudofemoralis. The tails of oviraptorosaurs were extremely stocky — that is, all their tail muscles were unusually beefy – which has unique functional implications . . .
As for dromaeosaurs, their tails are odd indeed. All, but the most anterior dromaeosaur tail vertebrae, have what I call “caudal rods” – thin elongations of the chevrons and prezygapophyses. Sometimes, these caudal rods are seven times as long as the vertebral centra. Traditionally, it has been assumed that the caudal rods made dromaeosaur tails inflexible but useful as balancing poles. I disagree. Interestingly, it may be possible to phylogenetically bracket the evolution of dromaeosaur caudal rods between the primitive troodontid Anchiornis and the dromaeosaur Microraptor (both commonly accepted as winged gliding/flying animals), and caudal rods did convergently evolve in another group of archosaurs . . . the rhamphorhynchoid pterosaurs.
Sorry that I can’t say more.
— Scott
This is very cool, how long did it take you to get a good grasp on Rhino3D? And would you recommend this method for future studies trying to estimate muscle volume/mass in extinct animals?
So what is your view on the function of dromaeosaur caudual rods?
It seems unlikely that the theropod tail will move from side to side as they run, because theropods and all other dinosaurs have limbs tucked underneath their bodies, like mammals, not splayed out to the sides. When we watch a horse run, we don’t see its tail swinging from side to side, for example, so why would a theropod swing its tail from side to side? The hindlimbs of a theropod can rotate forward and backward, unlike the limbs of a lizard. therefore there is minimal lateral motion when a theropod runs. We do know that Archaeopteryx and Caudipteryx lack the M. caudofemoralis. That makes sense if Archaeopteryx is a flyer and Caudipteryx is a secondarily flightless bird.
Early birds like Archaeopteryx and Microraptor had long tails and hindlimb wings to help generate additional lift. A long tail with feathers can generate more lift if it is light weight than if it is heavy and muscular. Since a flyer does not need a muscular tail to balance it as it runs, losing the M. caudofemoralis is not maladaptive but adaptive in a primarily arboreal and volant animal. Even though Caudipteryx is most likely a flightless animal, it still retains the ancestral condition of the loss of the M. caudofemoralis. As Louis Dollo points out, evolution is irreversible. Caudipteryx cannot simply re-evolve the lost M. caudofemoralis. OTOH, if Caudipteryx was really a dinosaur that did not have a flying ancestor, then it makes no sense for it to lose such an adaptive feature as the M. caudofemoralis.
The comparison between the relative positions of the caudal ribs of modern quadrupedal reptiles and bipedal theropods ignores their different modes of locomotion and therefore it can be potentially misleading. Modern reptiles move their bodies in a sinusoidal curve and their limbs are splayed out to the side, but theropods have limbs tucked underneath their bodies, just like mammals, and their gait is probably more mammal-like even if most mammals are quadrupeds, not bipeds. Crocodilians also swim by swinging their tails from side to side. In fact, the power the crocodilian’s tail can generate is so enormous that they can be used as a defensive weapon.
Therefore, the M. caudofemoralis of crocodilians may well have evolved its massiveness for swimming, rather than running. Besides, mammals can run well enough without a M. caudofemoralis, so the current claim that T. rex locomotion was powered by the M. caudofemoralis would need further investigation, because comparing quadrupedal, semi-aquatic crocodilians with bipedal, terrestrial theropods is like comparing the proverbial apples and oranges. Further, when crocodilians swim, they fold their hindlimbs backwards and flush with their bodies to streamline their bodies. The hindlimbs do not move even when the tail is moved side to side. It shows that when the M. caudofemoralis is contracting, the legs do not have to move. Therefore there is evidence to contradict the claim that the M. caudofemoralis is the muscle used by theropods or crocodilians to walk. Since most theropods probably did not swim, that could not have been the reason for the massive M. caudofemoralis in T. rex. Did T. rex use its tail as a weapon? Or did T. rex really did use its massive tail to balance its front end? After all. T. rex went to extraordinary lengths to reduce weight at the front end by shrinking its forelimbs. Therefore do not trivialize the seesaw theory so casually. It may well be the real reason T. rex has a massive tail.
Im no paleontologist but one thing to note about Tyrannosaurus is that the posterior iliac crest flares outwards distally (unlike many carnosaurs,) this is also an indication for exposure of increased muscle mass. Im not too sure about the correct T.Rex gait, but it appears that with so much medially centered mass (the center of gravity seems to be slightly anterior of that of the pelvis) and that its gait would be slightly bent kneed with the legs at pointing out laterally at around 15-25 degrees. I suppose a squatting stance was the preferred stance of a robust T.rex. Having a beefy M. caudofemoralis would be beneficial especially concerning that it was attached to the femur, which I have hypothetically stated as being flared out. Such a beefy muscle could have acted like a stabilizer.
Tyrannosaurus looks as if it had a center of gravity centered around its legs, a particularly solid one most likely. Its iliac surface for muscle attachments and deep pelvis are substantial in comparison to lets say a theropod of its size like Giganotosaurus. Giga seems to be a better analogy for the teeter-totter explanation of large therapod mechanics especially concerning the elongated tail and laterally compressed rib cage. One thing I noticed was how compact T.rex (relative to other large therapods) was, especially how spherical its rib cage is as well as its S-shaped neck. The anatomy looks as if its gravity conscious.
The Article on the Tyrannosaurus tail mass matter is great and the details you provide here is very informative.