Well, I’m a moron again. In the new preprint that I just published, I briefly discussed the six species of sauropod for which complete necks are known — Camarasaurus lentus (but it’s a juvenile), Apatosaurus louisae (but the last three and maybe C5 are badly damaged), Mamenchisaurus hochuanensis (but all the vertebrae are broken and distorted), Shunosaurus lii, Mamenchisaurus youngi and Spinophorosaurus nigerensis.

I did have the wit to say, in the Author Comment:

Although I am submitting this article for formal peer-review at the same time as publishing it as a preprint, I also solicit comments from readers. In particular I am very keen to know if I have missed any complete sauropod necks that have been described in the literature. In the final version of the manuscript, I will acknowledge those who have offered helpful comments.

Happily, several people have taken me up on this (see the comments on the preprint), but one suggestion in particular was a real D’oh! moment for me. Oliver Demuth reminded me about Kaatedocus — a sauropod that we SV-POW!sketeers love so much that it has its own category on our site and we’ve held it up as an example of how to illustrate a sauropod specimen. More than that: we have included several illustrations of its vertebrae in one of our own papers.

Aaanyway … the purpose of this post is just to get all the beautiful Kaatedocus multiview images up in one convenient place. They were freely available as supplementary information to the paper, but now seem to have vanished from the publisher’s web-site. I kept copies, and now present them in the conveniently viewable JPEG format (rather the download-only TIFF format of the originals) and with each image labelled with its position in the column.

Please note, these images are the work of Tschopp and Mateus (2012) — they’re not mine!

Atlas and axis (C1-2)

Atlas and axis (C1-2)

























C15 (and the rest of the skeleton) is missing, which makes this a very nearly, but not quite, complete sauropod neck.


  • Tschopp, Emanuel, and Octávio Mateus. 2012. The skull and neck of a new flagellicaudatan sauropod from the Morrison Formation and its implication for the evolution and ontogeny of diplodocid dinosaurs. Journal of Systematic Palaeontology 11(7):853-888. doi:10.1080/14772019.2012.746589
Kaatedocus by Brian Engh

Kaatedocus is heading to the sidebar to help the cause.

We have a new page on the sidebar – here – where we’re collecting as many museum abbreviations as possible, the idea being that you can copy and paste them into your papers to rapidly populate the ‘Museum Abbreviations’ section. I grabbed about 100 from my own previous papers and a handful of others, so currently the list is highly skewed toward museums with (1) sauropods (2) that I’ve had reason to yap about. I’ve probably missed tons of museums that are important for people working on hadrosaurs or stegosaurs or (shudder) mammals. From here on out the list will grow as people suggest additions and edits in the comments on that page. So please get on over there and contribute!

Completely unrelated eyeball-bait art courtesy of Brian Engh, who writes,

I don’t even remember drawing this, I just found it lying around and spruced it up a bit today. It’s supposed to be some kinda diplodocid, maybe Kaatedocus, but I think the main goal of the drawing was to draw one with a sense of weight that felt right given that their center of mass is supposed to be so far back. I like the idea of them getting startled and popping up every now and again… [see also–MJW]

Introduction and Background

2005-09-27 CM 555 c6 480

An epipophysis in a neural arch of a juvenile Apatosaurus, CM 555. From this post.

I have three goals with this post:

  1. To document the range of variation in epipophyses in the cervical vertebrae of sauropods.
  2. To show that the “finger-like processes” overhanging the cervical postzygapophyses in the newly described Qijianglong are not novel or mysterious structures, just very well developed epipophyses.
  3. Finally, to show that similar long, overhanging epipophyses are present in other mamenchisaurids, although as far as I can tell no-one has noted them previously.

Epipophyses are muscle attachment points dorsal to the postzygapophyses, for the insertion of long, multi-segment epaxial (dorsal) neck muscles in birds and other dinosaurs. I know that they turn up occasionally in non-dinosaurian archosaurs, and possibly in other amniotes, but for the purposes of this post I’m only considering their distribution in sauropods. For some quick background info on epipophyses and the muscles that attach to them, see the second half of this post, and see Wedel and Sanders (2002) and Taylor and Wedel (2013a) for further discussion and more pictures.

OMNH emu vert 480

Before we start with the pictures, a fiddly nomenclatural point: this muscle attachment point dorsal to the postzyg has traded under at least six names to date.

  1. The ‘Owenian’ term, used by virtually all non-avian theropod workers, by Sereno et al. (1999) for Jobaria, and probably by loads of other sauropod workers (including myself, lately) is epipophysis.
  2. Beddard (1898) referred to this feature in birds as the hyperapophysis; this term seems to have fallen completely out of use.
  3. Boas (1929), again referring to birds, called it the processus dorsalis. Zweers et al. (1987: page 138 and table 1) followed this terminology, which is how I learned of it when I was an undergrad at OU.
  4. Baumel and Witmer (1993) called this feature in birds the torus dorsalis (note 125 on page 87), which some authors have informalized to dorsal torus (e.g., Harris 2004: page 1243 and fig. 1). Baumel and Witmer (1993: page 87) note that, “the use of ‘Torus’ is preferable since it avoids confusion with the spinous [dorsal] process of the neural arch”.
  5. In my own early papers (e.g., Wedel et al. 2000b) and blog posts I called this feature the dorsal tubercle, which was my own attempt at an informal term matching ‘processus dorsalis’ or ‘torus dorsalis’. That was unfortunate, since there are already several other anatomical features in vertebrates that go by the same name, including the dorsal-facing bump on the dorsal arch of the atlas in many vertebrates, and a bump on the humerus in birds and some other taxa. In more recent papers (e.g., Taylor and Wedel 2013a) I’ve switched over to ‘epipophysis’.
  6. In the last post, Mike coined the term parapostzygapophysis for this feature in Qijianglong. [Note: he now regrets this.]

As usual, if you know of more terms for this feature, or additional history on the ones listed above, please let us know in the comments.

Now, on to the survey.


Leonerasaurus_cervical_vertebrae - Pol et al 2011 fig 5

I haven’t seen very many prominent epipophyses in basal sauropodomorphs. Probably the best are these in the near-sauropod Leonerasaurus, which is very sauropod-like in other ways as well. Modifed from Pol et al. (2011: fig. 5).

This combination of photograph and interpretive drawing neatly shows why it’s often difficult to spot epipophyses in photos: unless you can make out the postzygapophyseal facet, which is often located more anteriorly than you might guess, you can’t tell when the epipophysis projects further posteriorly, as in the last of these vertebrae. In this case you can make it out, but only because the interpretive drawing shows the facet much more clearly than the photo.

Basal sauropods

Tazoudasaurus cervical - Allain and Aquesbi 2008 fig 9i-j

The most basal sauropod in which I have seen clear evidence of epipophyses is Tazoudasaurus. They’re not very apparent in lateral view, but in posterior view the epipophyses are clearly visible as bumps in the spinopostzygapophyeal laminae (SPOLs). Modified from Allain and Aquesbi (2008: fig. 9).

Jobaria epipophyes

In addition to Qijianglong, some other basal eusauropods have prominent epipophyses. Probably the best known is Jobaria; Sereno et al. (1999: fig. 3) figured and labeled the epipophysis in one of the cervical vertebrae. The vertebra image in that figure is tiny (nice work, glam-magz!), so here are some sketches of Jobaria mid-cervicals (from two different individuals) that I made back in the day when I was doing the research for Gary Staab’s Jobaria neck sculpture (see Sanders et al. 2000 for our SVP abstract about that project).

Turiasaurus also has prominent, overhanging epipophyses in at least some of its cervical vertebrae. You can just make one out as a tiny spike a few pixels long in Royo-Torres et al. (2006: fig. 1K). I have seen that cervical firsthand and I can confirm that the epipophyses in Turiasaurus are virtually identical to those in Jobaria.

Other mamenchisaurids

It’s not air-tight, but there is suggestive evidence of projecting epipophyses in some other mamenchisaurids besides Qijianglong.

Mamenchisaurus epipophyses - lateral view

If you’re really hardcore, you may remember that back in 2005, Mike got to go up on a lift at the Field Museum of Natural History to get acquainted with a cast skeleton of Mamenchisaurus hochuanensis that was mounted there temporarily. During that adventure he took some photos that seem to show projecting epipophyses in at least two of the mid-cervicals. At least, if they’re not epipophyses, I don’t know what they might be.

Mamenchisaurus epipophyses - medial view

Here they are again in medial view. My only reservation is that these vertebrae were distorted to begin with, and some features of the cast are very difficult to interpret. So, probably epipophyses, but it would be nice to check the original material at some point.

Mamenchisaurus youngi epipophyses

Something similar may be present in some posterior cervical vertebrae of Mamenchisaurus youngi. Here’s Figure 17 from Ouyang and Ye (2002). The “poz” label does not not seem to be pointing to the articular facet of the postzygapophysis, which looks to be a little more anterior and ventral, below the margin of the PODL. If that’s the case, then C15 has long, overhanging epipophyses like those of Jobaria. C16 has a more conservative bump, which is to be expected – the epipophyses typically disappear through the cervico-dorsal transition.

Omeisaurus epipophysis

Finally, here’s a cervical vertebra of Omeisaurus junghsiensis from Young (1939: fig. 2). I don’t want to hang very much on just a few pixels, but my best guess at the extent of the postzygapophyseal articular facet is shown in the interpretation above. If that’s correct, then this specimen of Omeisaurus had really long epipophyses, rivaling those of Qijianglong. Unfortunately that’s impossible to check, because this specimen has been lost (pers. comm. from Dave Hone, cited in Taylor and Wedel 2013).


Haplocanthosaurus epipophyses - Hatcher 1903

Haplocanthosaurus nicely shows that the epipophyses can be large in terms of potential muscle attachment area without projecting beyond the posterior margins of the postzygapophyses. Here is C14 of H. priscus, CM 572, in posterior and lateral views, modified from Hatcher (1903: plate 1).

diplodocid epipophyses

Epipophyses that actually overhang the postzygapophyses are not common in Diplodocidae but they do occasionally occur. Here are prominent, spike-like epipophyses in Diplodocus (upper left, from Hatcher 1901: plate 3), Barosaurus (upper right), Kaatedocus (lower left, Tschopp and Mateus 2012: fig. 10), and Leinkupal (lower right, Gallina et al. 2014: fig. 1).

NIgersaurus cervical - Sereno et al 2007 fig 3

Of course, the champion epiphysis-bearer among diplodocoids is the weird little rebbachisaurid Nigersaurus. Here’s a Nigersaurus mid-cervical, from Sereno et al. (2007: fig. 3). Note that the projecting portions of the epipophysis is roughly as long as the articular surface of the postzygapophysis.


Australodocus epipophysis

The epipophysis in this cervical of Australodocus just barely projects beyond the posterior margin of the postzygapophysis.

Giraffatitan c8 epipophyses

In Giraffatitan, epipophyses are absent or small in anterior cervicals but they are prominent in C6-C8. Here’s a posterolateral view of C8, showing very large epipophyses that are elevated several centimeters above the postzygapophyses. You can also see clearly in this view that the spinopostzygapophyseal lamina (SPOL) and postzygodiapophyseal lamina (PODL) converge at the epipophysis, not the postzygapophysis itself.

Sauroposeidon epipophyses

The holotype of Sauroposeidon, OMNH 53062, is similar to Giraffatitan in that the two anterior cervical vertebrae (possibly C5 and C6) have no visible epipophyses, but epipophyses are prominent in the two more posterior vertebrae (possibly C7 and C8). Click to enlarge – I traced the articular facet of the postzygapophysis in ?C8 to more clearly separate it from the epipophysis. For a high resolution photograph of that same vertebra that clearly shows the postzyg facet and the epipophysis dorsal to it, see this post.

Oddly enough, I’ve never seen prominent epipophyses in a titanosaur. In Malawisaurus, Trigonosaurus, Futalognkosaurus, Rapetosaurus, Alamosaurus, and Saltasaurus, the SPOLs (such as they are – inflated-looking titanosaur cervicals do not have the same crisply-defined laminae seen in most other sauropods) merge into the postzygapophyseal rami and there are no bumps sticking up above or out beyond the articular facets of the postzygs. I don’t know what to make of that, except to note that several of the animals just mentioned have mediolaterally wide, almost balloon-shaped cervical neural spines. In our 2013 PeerJ paper, Mike and I argued that the combination of tall neural spines and tall epipophyses in the cervical vertebrae of sauropods made them functionally intermediate between crocs (huge neural spines, no epipophyses) and birds (small or nearly nonexistent neural spines, big epipophyses). Perhaps most titanosaurs reverted to a more croc-like arrangement with most of the long epaxial neck muscles inserting on the neural spine instead of the postzygapophyseal ramus. I’ve never seen that possibility discussed anywhere, nor the apparent absence of epipophyses in most titanosaurs. As usual, if you know otherwise, please let me know in the comments!


Cervical vertebrae of Malawisaurus from Gomani (2005: fig. 9): not an epipophysis in sight. But check out the spike-like neural spines – these are so wide from side to side that from the front they look like party balloons.

And as long as we’re discussing the phylogenetic distribution of epipophyses, it is interesting that long, overhanging epipophyses are so broadly but sporadically distributed. They turn up in some non-neosauropods (Jobaria, Turiasaurus, Omeisaurus) and some diplodocoids (Nigersaurus, the occasional vertebra in Diplodocus and Leinkupal), but not in all members of either assemblage, and they seem to be absent in Macronaria (although many non-titanosaurs have shorter epipophyses that don’t overhang the postzygs). I strongly suspect that a lot of this is actually individual variation that we’re not perceiving as such because our sample sizes of almost all sauropods are tiny, usually just one individual. Epipophyses are definitely muscle attachment sites in birds and no better hypothesis has been advanced to explain their presence in other archosaurs. Muscle attachment scars are notoriously variable in terms of their relative development and expression among individuals, and it would be odd if epipophyses were somehow exempt from that inherent variability.

It also seems more than likely that ontogeny plays a role: progressive ossification of tendons attached at the epipophyses would have the effect of elongating the preserved projection. And since for some aspects of sauropod vertebral morphology, serial position recapitulates ontogeny (Wedel and Taylor 2013b), it shouldn’t be surprising that we see differences in the prominence of the epipophyses along the neck.

Back to Qijianglong

By now it should be clear that the “finger-like processes” in Qijianglong are indeed epipophyses, and although they are quite long, they aren’t fundamentally different from what we see in many other sauropods. I haven’t gone to the trouble, but one could line up all of the vertebrae figured above in terms of epipophysis size or length, and Qijianglong would sit comfortably at one end with Omeisaurus and Mamenchisaurus, just beyond Nigersaurus and Jobaria.

FIGURE 11. Anterior cervical series of Qijianglong guokr (QJGPM 1001) in left lateral views unless otherwise noted. A, axis; B, cervical vertebra 3; C, cervical vertebra 4; D, cervical vertebrae 5 and 6; E, cervical vertebra 7 and anterior half of cervical vertebra 8 (horizontally inverted; showing right side); F, posterior half of cervical vertebra 8 and cervical vertebra 9; G, cervical vertebra 10; H, cervical vertebra 11; I, close-up of the prezygapophy- sis-postzygapophysis contact between cervical vertebrae 3 and 4 in dorsolateral view, showing finger-like process lateral to postzygapophysis; J, close- up of the postzygapophysis of cervical vertebra 5 in dorsal view, showing finger-like process lateral to postzygapophysis. Arrow with number indicates a character diagnostic to this taxon (number refers to the list of characters in the Diagnosis). All scale bars equal 5 cm. Abbreviations: acdl, anterior centrodiapophyseal lamina; cdf, centrodiapophyseal fossa; plc, pleurocoel; pocdl, postcentrodiapophyseal lamina; poz, postzygapophysis; pozcdf, post- zygapophyseal centrodiapophyseal fossa; pozdl, postzygodiapophyseal lamina; ppoz, finger-like process lateral to postzygapophysis; ppozc, groove for contact with finger-like process; przdl, prezygodiapophyseal lamina; sdf, spinodiapophyseal fossa.

Cervical vertebrae of Qijianglong (Xing et al. 2015: fig. 11)

The strangest thing about the epipophyses in Qijianglong is that they seem to be bent or broken downward in two of the vertebrae (B and H in the figure above). I assume that’s just taphonomic distortion – the cervical shown in H wouldn’t even be able to articulate with the vertebra behind it if the epipophysis really drooped down like that. The epipophyses in Qijianglong seem to mostly manifest as thin spikes of bone (or maybe plates, as shown in B and I), so it’s not surprising that they would get distorted – most of the vertebrae shown above have cervical ribs that are incomplete or missing as well.

One more noodle-y thought about big epipophyses. I wrote in the last section that I’ve never seen them in titanosaurs, possibly because titanosaurs have big neural spines for their epaxial muscles to attach to. Maybe long, overhanging epipophyses are so common in mamenchisaurids because their neural spines are so small and low. Although we tend to think of them as a basal group somewhat removed from the “big show” in sauropod evolution – the neosauropods – mamenchisaurids did a lot of weird stuff. At least in terms of their neck muscles, they may have been the most birdlike of all sauropods. Food for thought.


Wedel and Taylor 2013 bifurcation Figure 7 - small Diplodocus cervical

Figure 7. BYU 12613, a posterior cervical of Diplodocus or Kaatedocus in dorsal (top), left lateral (left), and posterior (right) views. It compares most favourably with C14 of D. carnegii CM 84/94 (Hatcher, 1901: plate 3) despite being only 42% as large, with a centrum length of 270 mm compared to 642 mm for C14 of D. carnegii.

The original version of the PDF of our new paper (Wedel and Taylor 2013) had a couple of obvious errors: Kaatedocus was misspelled in the caption to Figure 7 (as Kaatedocu), and the submission date was given as June 24, 2012, not the correct date of June 14. Both of these errors were introduced during the editorial handling, so I politely asked  if they could be fixed, and thanks to the kind offices of the folks at PalArch, now they have been. However, to avoid confusion (or perhaps propagate it, depending on your feelings), the corrected PDF has a different filename. The original version will continue to be available at:


and the corrected version (with an extra ‘1’ on the end of the filename) is at:


Two things:

  • I will go around changing the links here and elsewhere (FigShare, etc.) to the new version, but I probably won’t have time today, as I have an all-afternoon community outreach at the local public library to help organize.
  • I realize that some people, including possibly my coauthor, will hate this because now we have created some uncertainty about which is the version of record. So we’re not going to ask for any more changes, no matter how egregious the errors we find (and we are certain to find a few more, that’s just the nature of the beast); as far as I’m concerned, the second corrected version is the final version. Also, the changes made are tiny and don’t affect the science at all, so it’s not like we’ve moved any important goalposts here.

If you have strong feelings about this either way, feel free to sound off in the comments.

A couple of days ago, a paper by Tschopp and Mateus (2012) described and named a new diplodocine from the Morrison Formation, Kaatedocus siberi, based on a beautifully preserved specimen consisting of a complete skull and the first fourteen cervical vertebrae.

Unfortunately, the authors chose to publish their work in the Journal of Systematic Palaeontology, a paywalled journal, which means that most of you reading this will be unable to read the actual paper — at least, not unless you care enough to pay £27 for the privilege.

So you’ll just have to take my word for it when I tell you that it’s a fine, detailed piece of work, weighing in at 36 pages. It features lavish illustrations of the skull, but we won’t trouble you with those. The vertebrae are illustrated rather less comprehensively, though still better than in most papers:


Tschopp and Matteus (2012: figure 9). A, Photograph and B, drawings of the mid-cervical vertebrae of the holotype of Kaatedocus siberi (SMA 0004). Photograph in lateral view and to scale, CV 8 shown in the drawings is indicated by an asterisk. Drawings of CV 8 (B) in dorsal (1), lateral (2), ventral (3), posterior (4) and anterior (5) views. Scale bars = 4 cm.

It should be immediately apparent from these lateral views that the vertebra are rather Diplodocus-like. But the hot news is that there is a great raft of free supplementary information, including full five-orthogonal-view photos of all fourteen vertebrae!

Here is just one of them, C6, in glorious high resolution (click through for the full awesome):


Now, folks, that is how to illustrate a sauropod in 2012! The goal of a good descriptive paper is to be the closest thing possible to a proxy for the specimen itself, and you just can’t do that if you don’t illustrate every element from multiple directions. By getting this so spectacularly right, Tschopp and Matteus have made their paper the best illustrated sauropod description for 91 years. (Yes, I am talking about Osborn and Mook 1921.)

It’s just a shame that all the awe-inspiring illustrations are tucked away in supplementary information rather than in the paper itself. Had the paper been published in a PLOS journal, for example, all the goodness could have been in one place, and it would all have been open access.

Is Kaatedocus valid?

There’s a bit of a fashion these days for drive-by synonymisation of dinosaurs, and sure enough no sooner had Brian Switek written about Kaatedocus for his new National Geographic blog than comments started cropping up arguing (or in some cases just stating) that Kaatedocus is merely Barosaurus.

It’s not. I spent a lot of time with true Barosaurus cervicals at Yale this summer, and those of Kaatedocus are nothing like them. Here is Tschopp and Mateus’s supplementary figure of C14:


And here is a posterior vertebra — possibly also C14 — of the Barosaurus holotype YPM 429, in dorsal and right lateral views:



Even allowing for a certain amount of post-mortem distortion and “creative” restoration, it should be immediately apparent that (A) Barosaurus is much weirder than most people realise, and (B) Kaatedocus ain’t it.

There may be more of a case to be made that Kaatedocus is Diplodocus — but that’s the point: it there’s a case, then it needs to be actually made, which means a point-by-point response to the diagnostic characters proposed by the authors in their careful, detailed study based on months of work with the actual specimens.

There seems to be an idea abroad at the moment that it’s somehow more conservative or sober or scientific to assume everything is a ontogenomorph of everything else — possibly catalysed by the Horner lab’s ongoing “Toroceratops” initiative and subsequent cavalier treatment of Morrison sauropods — maybe even by the Amphidocobrontowaassea paper. Folks, there is no intrinsic merit in assuming less diversity. Historically, the Victorian sauropod palaeontologists of England did at least as much taxonomic damage by assumptions of synonymy (everything’s Cetiosaurus or Ornithopsiswhatever that is) as they did by raising new taxa. The thing to do is find the hypothesis best supported by evidence, not presupposing that either splitting or lumping is a priori the more virtuous course.

Sermon ends.

Morrison sauropod diversity

As we’ve pointed out a few times in our published work, sauropod diversity in the Kimmeridgian-Tithonian in general, and in the Morrison Formation in particular, was off-the-scale crazy. There’s good evidence for at least a dozen sauropod genera in the Morrison, and more than fifteen species. Kaatedocus extends this record yet further, giving us a picture of an amazing ecosystem positively abundant with numerous species of giant animals bigger than anything alive on land today.

Sometimes you’ll hear people use this observation as a working-backwards piece of evidence that Morrison sauropods are oversplit. Nuh-uh. We have to assess taxonomy on its own grounds, then see what it tells us about ecosystem. As Dave Hone’s new paper affirms (among many others), Mesozoic ecosystem were not like modern ones. We have to resist the insidious temptation to assume that what we would have seen in the Late Jurassic is somehow analogous to what we see today on the Serengeti.

Hutton’s (or Lyell’s) idea that “the present is the key to the past” may be helpful in geology. But despite its roots as a branch of the discipline, the palaeontology we do today is not geology. When we’re thinking about ancient ecosystems, we’re talking about palaeobiology, and in that field the idea that the present is the key to the past is at best unhelpful, at worst positively misleading.

Sermon ends.

But isn’t the Kaatedocus holotype privately owned?

You’ve had two sermons already, I’m sure we can all agree that’s plenty for one blog post. I will return to this subject in a subsequent post.

Sermon doesn’t even get started.


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.

Tschopp, Emanuel, and Octávio Mateus. 2012. The skull and neck of a new flagellicaudatan sauropod from the Morrison Formation and its implication for the evolution and ontogeny of diplodocid dinosaurs. Journal of Systematic Palaeontology. doi:10.1080/14772019.2012.746589