Tutorial 15: the bones of the sauropod skeleton

September 7, 2011

We should have done this long ago.  Back in the early tutorials, we covered skeletal details such as regions of the vertebral column, basic vertebral anatomy, pneumaticity and laminae, but we never started out with an overview of the sauropod skeleton.

Time to fix that.  This is numbered as Tutorial 15 but you can think of it as Tutorial Zero if you prefer.  Thanks to the wonder of homology, it doubles as a primer for dinosaur skeletons in general.


Here is a complete, labelled sauropod skeleton, modified from Erwin S. Christman’s reconstruction of Camarasurus supremus in Ostrom and Mook 1921:plate LXXXIV:

Click through for the full-sized version (2897 by 1280 pixels), which you are welcome to print out and hang on your wall as a handy reference.  (Christman’s original is out of copyright; I hereby make my modified version available under the CC-BY-NC-SA licence.)

Since that’s a lot to take in all at once, we’ll walk through the regions of the skeleton: the head and neck, the rest of the vertebral column, the forelimb and girdle, and the hindlimb and girdle.  But first, a little bit of …

Skeletal nomenclature

Skeletons consist of bones.  The study of skeletons and of bones is called osteology.  There are several ways of dividing up the skeleton into manageable chunks.  One is to consider cranial vs. postcranial bones.  In this division, cranium just means skull (though see below) and postcranium means “everything except the skull”.  Here at SV-POW!, of course, we consider skulls beneath our notice, so this division seems silly to us.  We have been known to refer to the skull as the prepostcranium on occasion.

A more useful division of the skeleton is into axial and appendicular.  The axial skeleton includes the skull, hyoid apparatus (little bones in the neck that anchor tongue and throat muscles), vertebrae, ribs and chevrons (i.e. everything on the midline), and the appendicular bones are those of the limbs and their girdles, i.e. shoulders and hips.  (I learned only very recently that, although they seem to be part of the forelimb girdle, the sternal plates are actually part of the axial skeleton, being related to the ribs rather than the shoulders.)

Head and neck

Let’s start with the head.  Although “cranium” is sometimes used to mean the whole head, as noted above, it more strictly refers to the rigid upper portion of the skull which attaches to the neck and includes the upper jaw.  The lower jar, which moves independently, is called the mandible. Both of these units are made up of many smaller bones.  There is of course much, much more to say about skull anatomy, but that is another tutorial for another day.  For now, we will just pretend that the skull is made of two lumps of bone and move swiftly onwards.

The back of the skull articulates with the neck, which is part of the spine, or vertebral column.  All vertebrates have a spine; and in all tetrapods it’s divided into neck, trunk (or torso), sacrum and tail.  The spine is composed of vertebrae: those in the neck are called cervical vertebrae, or cervicals for short; those in the trunk are called dorsal vertebrae (in crocs and mammals these are further broken down into the thoracic vertebrae, which bear mobile ribs, and the lumbar vertebrae which do not); those in the sacrum are called sacral vertebrae and those in the tail are called caudal vertebrae.  But you already know that if you read Tutorial 1.

In some kinds of tetrapods, including all dinosaurs, the cervical vertebrae have backward-pointing ribs; these are called the cervical ribs.  Birds have these (in reduced form) and so do crocs and mammals, but they are absent in at least some lizards and turtles. Contrary to popular belief, mammals do have bicipital (two-headed) cervical ribs, they are just very short and fused to the vertebrae. Even most human osteology textbooks refer to them as transverse process. But developmentally and functionally they are ribs; they bound the transverse foramina through which the vertebral arteries pass, and they anchor deep neck muscles. The “cervical ribs” that occasionally crop up as a pathology in humans are large, mobile, thoracic-style ribs, and represent segmentation anomalies during early development.

The cervical vertebrae are numbered backwards from the head. Each cervical can be identified by number, so that the tenth is called “cervical 10″, or C10 for short.  Sauropods have between eleven and nineteen cervicals — a lot more than the feeble seven that nearly all mammals have, but well short of the seventy or so that Elasmosaurus could boast.

In most tetrapods, the cervicals from C3 backwards are similar in shape, although they tend to get bigger as they approach the torso; but the first two are distinctive, so they have special names.  C1 is called the atlas — easy to remember as it holds up the head, just as the titan Atlas held up the sky (not the Earth as often thought).  It doesn’t really look like a vertebra at all, being ring-shaped and (in sauropods) tiny.  C2 is called the axis.  It looks much more like a normal vertebra, but has an odd articulation at the front, a distinctive blunt spike that the atlas sits on (it also has small prezygapophyses for the neural arch elements of the atlas–these little bits of bone are often lost in fossil skeletons).  It’s smaller than the succeeding vertebrae — unlike the situation in mammals, in which the axis is ususally the largest cervical — and has a big, blade-like neural spine.

Torso and tail

The vertebral column continues back from the base of the neck, as the torso, which consists of dorsal vertebrae.

In the region of the hips, several vertebrae fuse together: this is true to some extent in most or all tetrapods, but in many groups it’s only two or three vertebrae that fuse, whereas in sauropods (and most dinosaurs) it’s four or more.  This set of fused vertebrae is the sacrum, and the vertebrae that make it up are the sacral vertebrae.

Behind the sacrum is the tail, which is composed of caudal vertebrae.  Hanging beneath these — or, specifically, between the intervertebral joints — are transversely flattened bones called chevrons or haemopophyses.  These exist in most reptiles, but have been lost in most mammals. (They do exist in wallabies, but they are a very different shape.) Developmentally the chevrons mirror the neural arch, and form a canal for the caudal aorta in the same way that the neural arch forms a canal for the spinal cord.

Just as the cervical vertebrae have cervical ribs, so the dorsal vertebrae have dorsal ribs.  These are longer and more vertically oriented than the cervical ribs.  The sacral vertebrae, too, have sacral ribs, but you rarely see them because in lateral view they are obscured by the ilium — as is the case here.  You might, then, wonder whether the caudal vertebrae have caudal ribs, but the answer is not clear.  The first few caudals, at least, do have lateral processes, but surprisingly there is no consensus about what they actually are: ribs that are fused to the vertebrae, or paraphophyses/diapophyses that are fused together.  See the overview in Wilson (1999:642).

How can you tell where the neck ends are the torso begins?  The traditional answer is that the first dorsal vertebrae is the first one with a “free” (i.e. unfused) rib, but it’s not always that clear.  Although cervical ribs generally fuse to their vertebrae and dorsal ribs rarely or never do, there are plenty of exceptions — for example, the last few cervical ribs of the Mamenchisaurus hochuanensis holotype appear unfused.  Also, in specimens where the cervicodorsal transition is well preserved, it’s apparent that the switch from short backward-directed cervical ribs to long downward-directed dorsal ribs may be abrupt, between adjacent vertebrae, or a gradual transition spread out over several vertebrae. Since the shoulder girdle bones don’t articulate with the torso, that clue’s also unavailable, so all in all it can be hard to nail down where the transition was.  You just sort of know it when you see it.

The final axial bones are the sternal plates, which belong somewhere in the breast area.  The exact placement and orientation of these bones is not agreed, and they are rarely if ever preserved in place.

Shoulder and forelimb

The bones of the shoulder are the elongate scapula, or shoulder-blade, on the side of the torso; and the coracoid, lower down wrapping round to the front.  Together, these bones make up the shoulder girdle.  Unlike the pelvis, the shoulder is not fused to the bones of the torso, but would have been bound to it by ligament and muscle.  Because of this, the exact position of the scapula and coracoid are not known, and remain the subject of controversy.  The reconstruction above shows a fairly vertical scapula; some others make it more nearly horizontal.

Where the scapula and coracoid meet, they form a hollow on the underside, called the glenoid.  The head of the humerus fits in here; two parallel bones form the lower limb segment: the ulna and radius.  In sauropods, the ulna is a rounded triangle in cross-section, with a hollow on the front face of the triangle which the radius fits into.

At the bottom of the lower limb segment are the carpals, or wrist bones; then the manus, or hand.  The upper bones of the manus are the metacarpals, which in sauropods are held near-vertical in a semi-circular arcade with the hollow directed backwards and slightly inwards.  Below the metacarpals are the phalanges (singular phalanx); each finger may have multiple phalanges, but sauropods tend to have very few.  When the last phalax of a digit is claw-shaped, it’s called an ungual.

Because both forefeet and hindfeet have phalanges and unguals, we distinguish by saying manual phalanges and manual unguals for the bones of the forelimb, and pedal phalanges and pedal unguals for those of the hindlimb.

Hip and hindlimb

The pelvis, or hip girdle, is made up of three bones on each side: the ilium, on top, is roughly semi-circular; the pubis, at the front, and the ischium, at the back, are more elongate.  Where these three bones meet, they form a circular hole called the acetabulum, or hip socket.  Unlike the shoulder girdle, the pelvis is fused to the torso: specifically, the ilium is fused to the sacrum via the transverse processes of the sacral vertebrae and their sacral ribs.  The pubes and ischia do not fuse.

The femur, or thigh bone, has a head that projects into the acetabulum.  At the knee, it meets two parallel lower-limb bones, the tibia and fibula.  The former is the main weight-bearing bone and is nearest the midline.  The fibula sits to the side of it.  Unlike mammals, most reptiles including non-avian dinosaurs have no kneecap, or patella; but birds do. Sesamoids or “floating” bones like the patella seem to be evolved and lost more readily than the normally-connected bones of the skeleton.

Below these two bones are the tarsals, or ankle bones.  In sauropods there are one or two of these: a large, disc-shaped astragalus beneath the tibia, and sometimes a smaller globular calcaneum below the fibula.  (For some reason, the carpals don’t seem to have names.)  Beneath these is the pes, or hindfoot.  The upper bones of the pes are the elongate metatarsals.  Beyond these are the short pedal phalanges and unguals.

What did we miss?

The bones listed account for nearly all the skeleton.  There are, however, a few extra bones that are rarely recovered or not always present.  Clavicles, or collar bones, have been reported in the limb girdles of some sauropods.  Gastralia, or belly ribs, were probably present in all sauropods, but are fragile and very rarely preserved.  Finally, some sauropods had osteoderms — small, isolated bones embedded in the skin and serving as armour.  None of these are illustrated in Christman’s Camarasaurus.

Comparative osteology

Because the basic tetrapod body-plan is so conservative — many bones change size and shape, but it’s comparatively rare for bones to evolve away or for new ones to evolve — you can look at skeletons of all sorts of animals in a museum and recognise nearly all the bones I’ve listed here.  Birds, the closest living relatives of sauropods, have everything I’ve listed here, though their sternal plates have merged into a single big sternum and their forelimbs are obviously highly modified.  Crocs have everything.  Lizards have everything except cervical ribs.  Even mammals are surprisingly similar, though all the pelvis bones fuse together and the coracoid is lost (the coracoid process of the scapula in humans and other mammals is a different, non-homologous bit of bone).

In particular, you have nearly all the bones in a sauropod skeleton, though of course many of the bones are very different in shape, or fused together, and your tail is contemptible.  You might like to try re-reading this tutorial, finding all the relevant bones in your own body.  You have a few extras as well: most obviously, your kneecaps, but also extra bones in the wrist and ankle.

SEE ALSO: the same thing done for Tyrannosaurus.


Osborn, Henry Fairfield, and Charles C. Mook.  1921.  Camarasaurus, Amphicoelias and other sauropods of Cope.  Memoirs of the American Museum of Natural History, n.s. 3:247-387, and plates LX-LXXXV.

Wilson, Jeffrey A.  1999.  A nomenclature for vertebral laminae in sauropods and other saurischian dinosaurs.  Journal of Vertebrate Paleontology 19(4): 639-653.  [Wilson used to have a freely available PDF on his site, but he seems to have removed it, and substituted a link to a paywalled PDF.]

About these ads

35 Responses to “Tutorial 15: the bones of the sauropod skeleton”

  1. Allen Hazen Says:

    Nice post. Should somehow be indexed so lots of (students and other) people can refer to it.

    A couple of comments from a mammal:
    —Whales have, if not haemal arches/ chevrons, something similar in the same position under the joints between caudal vertebrae.

    —Camarasaurus seems to have 11 dorsals, and other Sauropods have similarly short dorsal series. This seems very odd! Mammals typically have something like 19 dorsals and lumbars, and a couple of clades that include large herbivores (Perissodatyls, Proboscideans) have extras. Where in that short body did Sauropods accommodate the innards needed to digest plant food? Did they do something odd to allow full digestion with a short gut: maybe refection of droppings (like rabbits)?

  2. Matt Wedel Says:

    Whales have, if not haemal arches/ chevrons, something similar in the same position under the joints between caudal vertebrae.

    Yep, those are haemal arches.

    Where in that short body did Sauropods accommodate the innards needed to digest plant food?

    Wait, you’re complaining that sauropod bodies are too small? :-) Surely it is the relative proportions of the body parts that matter, and not the sheer number of vertebrae. Sauropod vertebrae were big compared to the size of the body (corn on the cob, not shish kebabs). And they had long ribs and deep body profiles compared to mammals (tacos not corndogs).

    But don’t take my word for it: Franz et al. (2009) actually did the work and found that the guts needed to run a sauropod would fit very comfortably inside the ribcage. Happily that paper is free to the world here.

  3. Whales, and in fact various mammals, often have their haemapophyses separated distally. This makes them seem different. This is because, as in some fish, and many, many, amphibians and some reptiles, the haemapophyses are paired elements on either side of the haemal canal that fused in their distal ends, forming a “haemal arch.” This structure (termed the “haemal spine” in some sources) is often bent backwards in lateral view, giving it the name “chevron,” and the haemapophyses become a term more useful for the portion of the arch that articulates with the adjacent vertebrae in distinct facets. Sometimes an ossification forms across the arch, enclosing the haemal canal. The best is when sundry and all fusion occurs and the arches fuse with their respective vertebrae (primarily, I think it is the more anterior of the two articulating centra). Of course, this etymological gibberish is pointless in face of the larger expanse of the tutorial, which is fine as is. It skips unnecessary qualification on the use and comparison of various terms (“skull” vs. “cranium,” vs. “whatever”) because it’s largely meaningless unless you’re going to devote a lot of time to the discussion.

    Nice work, Mike.

  4. Nathan Myers Says:

    With your permission, I’m inclined to think of this as tutorial 0, instead of 15.

  5. Mark Wildman Says:

    Excellent post Mike. This would make great SV-POW television! If only……

  6. Mike Taylor Says:

    Thanks, Allen and Jaime, for comments on chevrons. I tweaked the article to mention whales.

  7. Great tutorial!

    Note that C2 (axis) is often called the epistropheus in older and medical texts.

  8. Allen, one point about vertebral count is that mammals have very mobile trunks because they use bending and extending a lot for rapid locomotion. Sauropod bodies were quite stiff, thus there was less need to have a large number of dorsals. Fewer but longer verts did the same job. And this freed anterior dorsals for incorporation into the neck (termed cervicalization).

  9. Really excellent post: I’m going to link to it from my osteology lecture notes.

    As for the difference in vertebral count, check out this paper:
    Johannes Müller, Torsten M. Scheyer, Jason J. Head, Paul M. Barrett, Ingmar Werneburg, Per G. P. Ericson, Diego Pol, and Marcelo R. Sánchez-Villagra
    Homeotic effects, somitogenesis and the evolution of vertebral numbers in recent and fossil amniotes
    PNAS 2010 107: 2118-2123.

    Differences occur both in the total number of presacrals generated in development, and in the regionalization assigned to those vertebrae.

  10. Nice post guys. I will definitely tweet this.

  11. Matt Cobley Says:

    So… question for you guys.

    What is the plural of manus?

  12. Kurt Kohler Says:

    The plural of manus is…
    wait for it….

  13. I can this Kurt — I did a whole dissertation on sauropod feet and limbs! It is manus, having a U with two dots over it. It is pronounced manOOs.

    Matt Bonnan

  14. Mike Taylor Says:

    You must never use “mani” as the plural of manus, otherwise Heinrich will hunt you down and kill you like a dog.

  15. Matt Cobley Says:

    I was going with manūs, but I guess I can just swap it out for an umlaut. Thanks guys

  16. [...] time, we looked at the bones of the sauropod skeleton, and I mentioned that “thanks to the wonder of homology, it doubles as a primer for dinosaur [...]

  17. Allen Hazen Says:

    Matt Cobley–
    No! Don’t do it! Keep the overbar (“macron”) over the u! (Latin doesn’t use umlauts; the overbar — not used in classical times — is meant to signify that the vowel in the plural is drawn out longer than that in the singular, not that it has a different vowel-quality like the umlauted u in German.)

    Matt W, Heinrich M, Thomas H: thanks for comments and references! … Now that you mention it, the individual vertebrae in Sauropods ARE sort of big. ;-)

  18. Allen Hazen Says:

    Sorry, I was only thinking about comments on vertebral count when I thanked the rest. Thank you for the further information on Cetacean haemals.

    Especially for the remark that when haemals fuse with centra, it is usually with the anterior: I spent some time staring at whale skeletons a few months back, and wondered if there was a principled way of “allocating” haemals to centra!

  19. Mike Taylor Says:

    On chevrons fusing with their caudals, see this earlier SV-POW! article.

  20. Mark Robinson Says:

    Another great article, Mike. I think there’s a typo in the 4th para of “Head and neck” – surely the eleventh cervical is denoted as C11 not C10?

    Also, in the very last para, is it definitely the case that the human coracoid process is homologous with the coracoid in dinos?

  21. Allen Hazen Says:

    I’ve had a skim of the Mueller et al. paper Tom Holtz referred to: interesting survey of different amniote taxa. Interestingly, most (*) Sauropods are like mammals in having close to the ancestral Amniote count of pre-sacral vertebrae, but whereas mammals have kept close to the original division between cervicals and “dorsals” (= thoracics + lumbars in mammals), the Sauropods (on Mueller et al.’s hypothesis about the developmental processes involved) have, as it were, moved their shoulders back: converted what in ancestral Amniotes were forward dorsals into rearward cervicals!

    (*) Mamenchisaurus being an exception: it really has some extra vertebrae. Mamenchisaurus is weird-looking: when I first saw a mounted specimen, what with the large number of cervicals and the variable length of individual cervicals, I wondered if it was somehow cobbled together out of bones of more than one individual! I take it that it is known from, if not articulated, securely enough associated bones that professionals are confident of the reconstruction?

  22. Matt Wedel Says:

    Also, in the very last para, is it definitely the case that the human coracoid process is homologous with the coracoid in dinos?

    Augh, good catch. No, it’s a different thing. I’m experiencing the anguish because I put in that dumb bit when I was proofing the post for Mike. Fixed it now–thanks.

    Mamenchisaurus is weird-looking: when I first saw a mounted specimen, what with the large number of cervicals and the variable length of individual cervicals, I wondered if it was somehow cobbled together out of bones of more than one individual! I take it that it is known from, if not articulated, securely enough associated bones that professionals are confident of the reconstruction?

    I’m virtually certain that the skeleton you saw was a cast of the holotype of M. hochuanensis–there are mounts in Los Angeles and Copenhagen and probably quite a few other places–and it is based on a mostly complete, articulated skeleton that lacked the forelimbs and distal tail and not much else. So the neck to trunk proportions are accurate (Omeisaurus is equally wacky). I can’t believe we’ve never blogged about the image of that skeleton as it was found in the field. It’s just a drawing, from Young and Zhao (1972), but there is a lot of good stuff in there (there is also a photo of the skeleton in the field, but it was taken from ground level and suffers from some foreshortening–the drawing is cooler). One more for the To Do list!

  23. Mike Taylor Says:

    Yes, the Mamenchisaurus hochuanensis holotype is one of surprisingly few sauropod specimens found with the complete, articulated necks. The neck:trunk proportions in the many casts are definitely correct. By happy coincidence, several other Mamenchisaurus specimens have also been found with the necks complete and articulated, something that is vanishingly rare with sauropods in general. I don’t know whether there was something taphonomically weird about that part of China at that time, or — and this idea is occurring to me literally only as I type this — maybe there was something about the construction of mamenchisaur necks, such as extra-strong ligaments, that helped to keep the necks intact after death?

    BYT., there is, sadly, no mamenchisaur in Copenhagen any more. The cast that I saw at the Copenhagen Geological Museum a few years back was on loan from the Homogea Museum in Trzic, Slovenia, and returned there not long afterwards. It’s the same one that I subsequently saw again in the Basement Of Doom.

  24. Mike Taylor Says:

    Thanks, Mark. I am not not sure where the bizarre “eleventh cervical is called C10″ typo came from from, but it’s fixed now.

  25. Even if Mike says I’ll kill you if you don’t get the plural of manus right – I won’t! I just once threatened a group of very lazy students, saying I’d flunk them if they made the plural “Manie”, which is German for “mania”.


  26. Allen Hazen Says:

    Thanks, Matt and Mike, for reassurances about Mamenchisaurus!

    What I saw may have had some original stuff in it: there was a visiting “Dinosaurs from China” exhibition in Melbourne (Australia) in the 1980s; part of the deal (which had been negotiated, on the Australian side, by Pat Vickers-Rich) was that the Melbourne museum got to make casts of the skeltons: so there may now be a Mamenchisaurus cast in Melbourne.

    Specific name for M. is (well, was in the exhibition catalogue) “constructus,” apparently because the first specimen was found on a construction site. Good to know it wasn’t “constructus” in the sense of “put together”!

  27. Matt Wedel Says:

    Specific name for M. is (well, was in the exhibition catalogue) “constructus,” apparently because the first specimen was found on a construction site. Good to know it wasn’t “constructus” in the sense of “put together”!

    Wow, I did not know that the M. constructus skeleton had ever been cast or exhibited outside of China. The skeleton Mike and I have been blabbing about is M. hochuanensis, which is arguably what most people think of when they think of Mamenchisaurus. The taxonomy of all the Omeisaurus and Mamenchisaurus species is overdue for revision; it’s very hard to tell from an ocean’s remove who actually belongs with whom. M. constructus is the genotype, so it’s not going anywhere, but it wouldn’t surprise me if someone made a case for splitting or lumping of several species (I think Mamenchisaurus is up to seven or so species). I’m not advocating that, it just wouldn’t surprise me.

  28. Allen Hazen Says:

    About the trivial name…
    I could be misremembering: the exhibit(*) was back in the 1980s, and I don’t think I have the catalogue any more: it’s just possible that I remember “constructus” from another source.

    (*) called, I think, “Chinese Dinosaurs.” I remember mounts of Mamenchisaurus and Tsintaosaurus, and an immense isolated cervical from some sauropod or other which fascinated me just as a sculptural form. There was another visiting fossil exxhibition in Melbourne a bit later — late 1980s– called “Dinosaurs from Russia” that included lower tetrapods from European Russia and dinosaurs from Mogolia as well as some neat Synapsid stuff: early Therapsids from European Russia and Multituberculates from Mongolia. I think the Melbourne Museum now has casts of specimens from both exhibits.

  29. Allen Hazen Says:

    O.k., my mistake: the Melbourne Mamenchisaurus, according to the museum’s WWWebsite


    is a hochuanensis.

  30. MartinB Says:

    Very nice – I remember I once did something similar but stupidly used non-free skeleton pictures, so I can’t post it anywhere.

    One small suggestion: Perhaps you could add a “coordinate system” with directions dorsal/ventral etc. – that might also help people a lot in reading the literature.

  31. [...] you’re not a regular and some of these terms are unfamiliar, check out these handy guides [1, 2] to the vertebrate skeleton). That leaves marrow in everything else, although the only bones [...]

  32. [...] printing and laminating a giant Camarasaurus skeleton and cutting it into pieces (print your own here). Sunday morning, the Community Garden exhibit became our dig site, and the kids loved burying the [...]

  33. […] memories might recall that, nearly two years ago, we published annotated skeletal reconstructions of Camarasaurus and of Tyrannosaurus, with all the bones labelled. At the time, I said that I’d like to do an […]

  34. […] skeleton of a bird. We’ve never done either of those — but we should, to go with our Camarasaurus, Tyrannosaurus and Triceratops. Skeletal homology for the […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


Get every new post delivered to your Inbox.

Join 346 other followers

%d bloggers like this: