The new monster redescription of Dilophosaurus by Adam Marsh and Tim Rowe came out in the Journal of Paleontology last week. I’m blogging about it now because the OA link just went live yesterday. So you can get this huge, important paper for free, at this link.

There’s a lot of stuff to love here: beautiful, clear photos of every element from every specimen from multiple angles, interesting anatomical and phylogenetic findings, and of particular interest on this blog, some very cool documentation of serial variation in pneumatic features. Here in Figure 62 we see serial changes in the posterior centrodiapophyseal laminae, which in some of the vertebrae are split around an intermediate fossa, or have accessory laminae.

One thing that I’ve thought a lot about, but written not so much about (yet), is pneumatic features on the ventral surfaces of vertebrae and how they change along the column. So I was excited to see Figure 64, which shows how fossae change serially on both the lateral and the ventral surfaces of the presacral centra. As far as I know, no-one has ever done something like this for a sauropod (please correct me in the comments if I’ve forgotten any examples), but it could be done and the results would be interesting, particularly for taxa like Haplocanthosaurus or Dicraeosaurus that have both lateral and ventral fossae and keels in at least some of the vertebrae.

Here’s Figure 66, a beautiful new skull reconstruction and life restoration, both by Brian Engh. There’s a lot of Engh/Dilophosaurus stuff going on right now, including a new video for the St. George Dinosaur Discovery Site museum (short version here, longer version available at the museum, and I think on Brian’s Patreon page), and, uh, another thing that will be revealed in the not-too-distant future.

I hope everyone is well and safe. When I first realized we were going into quarantine back in March, I had big plans for doing various series of posts here, but almost immediately the demand of getting med school anatomy online ate up all my time and creative energy. Just barely getting back on my feet now. I know Mike has been busier than normal, too. So please be patient with us, and we’ll try to remember to feed the blog now and then.


Marsh, Adam D., and Rowe, Timothy B. 2020. A comprehensive anatomical and phylogenetic evaluation of Dilophosaurus wetherilli (Dinosauria, Theropoda) with descriptions of new specimens from the Kayenta Formation of northern Arizona. Journal of Paleontology Volume 94, Supplement S78: 1-103. DOI:

Nature’s CT machine

January 28, 2020

Because I’ve worked a lot on the anatomy and evolution of air-filled bones in sauropod dinosaurs, I’ve spent most of my career looking at images like this:

CT sections through a cervical vertebra of an apatosaurine (Apatosaurus or Brontosaurus), OMNH 1094. Wedel (2003b: fig. 6). Scale bar is 10cm.

…and thinking about images like this:

Physical sections through pneumatic vertebrae of Giraffatitan. Janensch (1950: figs. 71-73).

Turns out, that’s pretty good practice for fossil prospecting in the Salt Wash member of the Morrison Formation, where we frequently find things like this:

That’s a bit hard to read, so let’s pull it out from the background:

This is almost certainly a pneumatic vertebra of a sauropod, sectioned more-or-less randomly by the forces of erosion to expose a complicated honeycomb of internal struts and chambers. The chambers are full of sandstone now, but in life they were full of air. I say “almost certainly” because there is small chance that it could belong to a very large theropod, but it looks more sauropod-y to me (for reasons I may expand upon in the comments if anyone is curious).

I’m not 100% certain what section this is. At first I was tempted to read it as a transversely-sectioned dorsal, something like the Allosaurus dorsal shown in this post (link) but from a small, possibly juvenile sauropod. But the semi-radial, spoke-like arrangement of the internal struts going to the round section at the bottom looks very much like the inside of the condyle of a sauropod cervical or cervico-dorsal–compare to fig. 71 from Janensch (1950), shown above. And of course there is no reason to suspect that the plane of this cut is neatly in any of the cardinal anatomical directions. It is most likely an oblique cut that isn’t purely transverse or sagittal or anything else, but some combination of the above. It’s also not alone–there are bits and bobs of bone to the side and above in the same chunk of sandstone, which might be parts of this vertebra or of neighboring bones. Assuming it is a sauropod, my guess is Diplodocus or Brachiosaurus, because it looks even more complex than the sectioned cervicals and dorsals I’ve seen of Haplocanthosaurus, Camarasaurus, or the apatosaurines.

Sometimes we can do a little better. This is one of my favorite finds from the Salt Wash. This boulder, now in two parts, fell down out of a big overhanging sandstone cliff. When the boulder hit, it broke into two halves, and the downhill half rolled over 180 degrees, bringing both cut faces into view in this photo. And there in the boulder is what looks like two vertebrae, but is in fact the neatly separated halves of a single vertebra. I know I refer to erosion and breakage as “Nature’s CT machine”, but this time that’s really on the nose. Let’s take a closer look:

Here’s what I see:

It’s a proportionally long vertebra with a round ball at one end and a hemispherical socket at the other end: a cervical vertebra of a sauropod. Part of the cervical rib is preserved on the upper side, which I suspect is the left side. The parapophysis on the opposite side is angled a bit out of the rock, toward the camera. Parapophyses of sauropod cervicals tend to be angled downward, and if we’re looking at the bottom of this vertebra, then the rib on the upper side is the left. The right cervical rib was cut off when the boulder broke. All we have on this side are the wide parapophysis and the slender strut of the diapophysis aiming out of the rock toward the missing rib, which must still be embedded in the other half of the boulder–and in fact you can see a bit of it peeking out in the counterpart in the wide shot, above.

Can we get a taxonomic ID? I think so, based on the following clues:

  • The cervical ribs are set waaay out to either side of the centrum, by about one centrum diameter. Such wide-set cervical ribs occur in Camarasaurus and the apatosaurines, Apatosaurus and Brontosaurus, but not typically in Diplodocus, Brachiosaurus, or other Morrison sauropods.
  • The cervical rib we can see the most of is pretty slender, like those of Camarasaurus, in contrast to the massive, blocky cervical ribs of the apatosaurines (for example).
  • We can see at least bits of both the left and right cervical ribs in the two slabs–along with a section right through the centrum. So the cervical ribs were set wide from the centrum but not displaced deeply below it, as in Camarasaurus, and again in contrast to the apatosaurines, in which the cervical ribs are typically displaced far below (ventral to) the centrum (see this).
  • This one is a little more loosey-goosey, but the exposed internal structure looks “about right” for Camarasaurus. There is a mix of large and small chambers, but not many small ones, and nothing approaching the coarse, regular honeycomb we’d expect in Apatosaurus, Brontosaurus, or Diplodocus, let alone the fine irregular honeycomb we’d expect in Barosaurus or Brachiosaurus (although I will show you a vert like that in an upcoming post). On the other hand, the internal structure is too complex for Haplocanthosaurus (compare to the top image here).
  • As long as Camarasaurus is on the table, I’ll note that the overall proportions are good for a mid-cervical of Cam as well. That’s not worth much, since vertebral proportions vary along the column and almost every Morrison sauropod has cervicals with this general proportion somewhere in the neck, but it doesn’t hurt.

So the balance of the evidence points toward Camarasaurus. In one character or another, every other known Morrison sauropod is disqualified.

When it’s too dark to hunt for sauropods, you can look at other things.

Now, Camarasaurus is not only the most common sauropod in the Morrison, it’s also the most common dinosaur of any kind in the formation. So this isn’t a mind-blowing discovery. Still, it’s nice to be able to get down to a genus-level ID based on a single vertebra fortuitously sectioned by Mother Nature. In upcoming posts, I’ll show some of the more exciting critters that we’ve been able to ID out of the Salt Wash, ‘we’ here including Brian Engh, John Foster, ReBecca Hunt-Foster, Jessie Atterholt, and Thuat Tran. Brian will also be showing many of these same fossils in the next installment of Jurassic Reimagined. Catch Part 1 here (link), and stay tuned to Brian’s paleoart channel (here) for more in the very near future.


Way back in 2009–over a decade ago, now!–I blogged about the above photo, which I stole from this post by ReBecca Hunt-Foster. It’s a cut and polished chunk of a pneumatic sauropod vertebra in the collections at Dinosaur Journey in Fruita, Colorado.

This is the other side of that same cut; you’ll see that it looks like a mirror image of the cut at the top, but not quite a perfect mirror image, because some material was lost to the kerf of the saw and to subsequent polishing, and the bony septa changed a bit just in those few millimeters.

And this is the reverse face of the section shown above. As you can see, it is a LOT more complex. What’s going on here? This unpolished face must be getting close to either the condyle or the cotyle, where the simple I-beam or anchor-shaped configuration of the centrum breaks up into lots of smaller chambers (as described in this even older post). It’s crazy how fast that can happen–this shard of excellence is only about 4 or 5 cm thick, and in that short space it has gone from anchor to honeycomb. I think that’s amazing, and beautiful.

It’s probably Apatosaurus–way too complex to be Camarasaurus or Haplocanthosaurus, not complex enough to be Barosaurus, no reason to suspect Brachiosaurus, and although there is other stuff in the DJ collections, the vast majority of the sauropod material is Apatosaurus. So that’s my null hypothesis for the ID.

Oh, back in 2009 I was pretty sure these chunks were from a dorsal, because of the round ventral profile of the centrum. I’m no longer so certain, now that I know that the anchor-shaped sections are so close to the end of the centrum, because almost all vertebrae get round near the ends. That said, the anchor-shaped sections are anchor-shaped because the pneumatic foramina are open, and having foramina that open, that close to the end of the vertebra still makes me think it is more likely a dorsal than anything else. I’m just less certain than I used to be–and that has been the common theme in my personal development over the last decade.

Spotted this beauty in the collections at Dinosaur Journey this past summer. With the front end of the centrum blown off, taphonomy once again proves to be the poor man’s CT machine, giving us a great look at the pneumatic spaces inside the vertebra.

EDIT, Oct. 13, 2019 — WHOOPS! That ain’t a cervical. Based on the plates in Madsen (1976), it’s a dead ringer for the second dorsal vertebra.

Allosaurus fragilis cervicodorsal transition - Madsen 1976 plates 14-16

Vertebrae C7 through D3 of Allosaurus fragilis in anterior view, from plates 14-16 in Madsen (1976). Abbreviations: dp, diapophysis; li, interspinous ligament scar; nc, neural canal; ns, neural spine; pp, parapophysis; pr, prezygapophysis.


Madsen, Jr., J.H. 1976. Allosaurus fragilis: a revised osteology. Utah Geological and Mining Survey Bulletin 109: 1-163.

We’re just back from an excellent SVPCA on the Isle of Wight. We’ll write more about it, but this time I just want to draw attention to a neat find. During a bit of down time, Matt and Vicki were wandering around West Cowes (the town where the scientific sessions were held), when they stumbled across a place called That Shop. Intrigued by all the Lego figures in the window, they went in, and Matt found a small section of fossils. Including … an Iguanodon pelvis, supposedly certified as such by the Dinosaur Isle museum.


Here it is: I imagine that whoever classified it read this elongate concave surface as part of the acetabulum. Matt’s hypothesis is that they mistook it for a sacral vertebra and that became “pelvis” via over-simplification.

It’s about 18 cm in a straight line across the widest part, or 20 cm around the curve.

Here is an actual documentary record of Matt’s moment of discovery:

Yep, you got it! It’s a sauropod vertebra! (Matt would never have spent good money on a stinkin’ appendicular element of a stinkin’ ornithopod.)

Specifically, it’s the bottom half of the front part of the centrum of a dorsal vertebra:

Eucamerotus” dorsal vertebra NHMUK PV R88 in right lateral and anterior views. Non-faded portions show the location of the Wedel Specimen. Modified from Hulke (1880: plate IV).

In these photos, we’re looking down into it more or less directly dorsal view, with anterior to the left. Click through the photos, and — once you know what you’re looking at — you can clearly see the pneumatic spaces: nice patches of finished bone lining the camellae, with trabecular bone in between.

Clearly there’s nowhere near enough of this to say what it is with any certainty. But our best guess is that it seems compatible with a titanosauriform identity, quite possibly in same space as the various Wealden sauropod dorsals that have been assigned to Ornithopsis or Eucamerotus.


  • Hulke, J. W.  1880.  Supplementary Note on the Vertebræ of Ornithopsis, Seeley, = Eucamerotous, Hulke. Quarterly Journal of the Geological Society 36:31–35.  doi:10.1144/GSL.JGS.1880.036.01-04.06

As Mike noted in the last post, many (all?) of the talks from SVPCA 2018 are up on YouTube. Apparently this has been the case for a long time, maybe most of the past year, and I just didn’t know. But I’m glad I do now, because I can encourage you to take 14 minutes and watch Jessie Atterholt’s talk on air spaces inside the neural canal in birds and other archosaurs:

This will not only be interesting in itself — assuming you are interested in pneumaticity, animals, or just how weird the natural world can be at times — but it will be good homework for the Atterholt and Wedel talk at this year’s SVPCA. That talk, also to be delivered by Jessie, will be on a different weird thing about archosaur neural canals, and one that neither of us have yapped about yet on social media.

Here’s the full rundown of talks by SV-POW!sketeers and affiliates at this year’s SVPCA:

Thursday, September 12

  • 11:00-11:20 – Vicki Wedel, “Validating the use of Dental Cementum Increment Analysis to determine season-at-death in humans and other mammals”
  • 11:20-11:40 – Matt Wedel, “How to make new discoveries in (human) anatomy”

Friday, September 13

  • 10:10-10:30 – Mike Taylor and Matt Wedel, “The past, present and future of Jensen’s Big Three sauropods”
  • 15:00-15:20 – Jessie Atterholt and Matt Wedel, “Neural canal ridges: a novel osteological correlate of post-cranial neurology in dinosaurs”

Presumably most or all of these will become PeerJ Preprints in time, just like Mike’s and my presentations from SVPCA 2017 (link, link) and Jessie’s presentation last year (link). I haven’t heard anything yet about livestreaming or recording of the talks this year — fingers firmly crossed.

Anyway, we look forward to seeing at least some of you at SVPCA or at other points on our trip to England, and to having more stuff to talk about here in the near future. Stay tuned!

Ray Wilhite posted this gorgeous image on a Facebook thread, and we’re re-posting it here with his permission.

It’s taken from a poster that Ray co-authored (Roberts et al. 2016). We’re looking here at a coronal cross-section of a hen (age not specified), with anterior to the left. Latex has been injected into the air sacs and lungs, highlighting them in shocking pink.

FInding your way around: the big yellow blobs near the middle are vitelline follicles. Just to their left, the two rounded red triangles that look like networks are the lungs. All the rest of the pink is diverticula and air-sacs: the interclavicle air-sac to the left, the caudal thoracic air-sac right behind the left (lower) lung, and abdominal air-sacs running backwards from the tips of the lungs. The big white oval is a calcified egg.

More from this poster in a subsequent post!


  • Roberts, John, Ray Wilhite, Gregory Almond, Wallace D Berry, Tami Kelly, Terry Slaten, Laurie McCall and Drury R. Reavill. 2016. Gross and histologic diagnosis of retrograde yolk inhalation in poultry. The American Association of Avian Pathologists, San Antonio, Texas. doi:10.13140/RG.2.2.28204.26246


If you followed along with the last post in this series, you now have some bird vertebrae to play with. Here are some things to do with them.

1. Learn the parts of the vertebrae, and compare them with those of other animals

Why are we so excited about bird vertebrae around here? Mostly because birds are reasonably long-necked living dinosaurs, and although their vertebrae differ from those of sauropods in relative proportions, all of the same bits are present in roughly the same places. If you know the parts of a bird vertebra and what each one does, you have a solid foundation for inferring the functions of sauropod vertebrae. Here’s a diagram I made for my SVP poster with Kent Sanders way back in 1999. I used an ostrich vertebra here but you should be able to find the same features in a cervical vertebra of just about any bird.

These are both middle cervical vertebrae in right lateral view. A middle cervical vertebra of a big ostrich will be between 3 and 4 inches long (7.5-10 cm), and one from a big brachiosaur like Giraffatitan will be about ten times longer.

I should do a whole post on neck muscles, but for now see this post and this paper.

2. Put the vertebrae in order, and rearticulate them

It is often useful to know where you are in the neck, and the only way to figure that out is to determine the serial position of the vertebrae. Here’s an articulated cervical series of a turkey in left lateral view, from Harvey et al. (1968: pl. 65):

Harvey’s “dorsal spine” is the neural spine or spinous process, and his “ventral spine” is the carotid process. The “alar process” is a sort of bridge of bone connecting the pre- and postzygapophyses; you can see a complete version in C3 in the photo below, and a partial version in C4.

Speaking of that photo, here’s my best attempt at rearticulating the vertebrae from the smoked turkey neck I showed in the previous post, with all of the vertebrae in left dorsolateral view.

These things don’t come with labels and it can take a bit of trial and error to get them all correctly in line. C2 is easy, with its odd articular surface for the atlas and narrow centrum with a ventral keel. Past that, C3 and C4 are usually pretty blocky, the mid-cervicals are long and lean, and then the posterior cervicals really bulk out. Because this neck section had been cut before I got it, some of the vertebrae look a little weird. Somehow I’m missing the front half of C6. The back half of C14 is also gone, presumably still stuck to the bird it went with, and C7 and C12 are both sectioned (this will come in handy later). I’m not 100% certain that I have C9 and C10 in the right order. One handy rule: although the length and neural spine height change in different ways along the column, the vertebrae almost always get wider monotonically from front to back.

And here’s the duck cervical series, in right lateral view. You can see that although the specific form of each vertebra is different from the equivalent vert in a turkey, the same general rules apply regarding change along the column.

Pro tip: I said above that these things don’t come with labels, but you can fix that. Once you have the vertebrae in a satisfactory order, paint a little dot of white-out or gesso on each one, and use a fine-point Sharpie or art pen to write the serial position (bone is porous and the white foundation will keep the ink from possibly making a mess). You may also want to put the vertebrae on a string or a wire to keep them in the correct order, but even so, it’s useful to have the serial position written on each vertebra in case you need to unstring them later.

3. Look at the air spaces

One nice thing about birds is that all of the species that are readily commercially available have pneumatic traces on and in their vertebrae, which are broadly comparable to the pneumatic vertebrae of sauropods.

The dorsal vertebrae of birds are even more obviously similar to those of sauropods than are the cervicals. These dorsal vertebrae of a duck (in left lateral view) show a nice variety of pneumatic features: lateral fossae on the centrum (what in sauropods used to be called “pleurocoels”), both with and without foramina, and complexes of fossae and foramina on the neural arches. Several of the vertebrae have small foramina on the centra that I assume are neurovascular. One of the challenges in working with the skeletal material of small birds is that it becomes very difficult to distinguish small pneumatic foramina and spaces from vascular traces. Although these duck vertebrae have small foramina inside some of the lateral fossae, the centra are mostly filled with trabecular, marrow-filled bone. In this, they are pretty similar to the dorsal vertebrae of Haplocanthosaurus, which have fossae on the neural arches and the upper parts of the centra, but for which the ventral half of each centrum is a brick of non-pneumatic bone. For more on distinguishing pneumatic and vascular traces in vertebrae, see O’Connor (2006) and Wedel (2007).

This turkey cervical, in left posterolateral view, shows some pneumatic features to nice advantage. The lateral pneumatic foramina in bird cervicals are often tucked up inside the cervical rib loops where they can be hard to see and even harder to photograph, but this one is out in the open. Also, the cervicals of this particular turkey have a lot of foramina inside the neural canal. In life these foramina are associated with the supramedullary diverticula, a set of air-filled tubes that occupy part of the neural canal in many birds — see Atterholt and Wedel (2018) for more on this unusual anatomical system. The development of foramina inside the neural canal seems to be pretty variable among individuals. In ostriches I’ve seen individuals in which almost every cervical has foramina inside the canal, and many others with no foramina. For turkeys it’s even more lopsided in my experience; this is the first turkey in which I’ve found really clear pneumatic foramina inside the neural canals. This illustrates one of the most important aspects of pneumaticity: pneumatic foramina and cavities in bones show that air-filled diverticula were present, but the absence of those holes and spaces does not mean that diverticula were absent. Mike and I coined the term “cryptic diverticula” for those that leave no diagnostic traces on the skeleton — for more on that, see the discussion section in Wedel and Taylor (2013b).

Finally, it’s worth taking a look at the air spaces inside the vertebrae. Here’s a view into C12 of the turkey cervical series shown above. The saw cut that sectioned this neck happened to go through the front end of this vertebra, and with a little clean-up the honeycomb of internal spaces is beautifully displayed. If you are working with an intact vertebra, the easiest way to see this for yourself is to get some sandpaper and sand off the front end of the vertebra. It only takes a few minutes and you’ll be less likely to damage the vertebrae or your fingers than if you cut the vertebra with a saw. Similar complexes of small pneumatic cavities are present in the vertebrae of some derived diplodocoids, like Barosaurus (see the lateral view in the middle of this figure), and in most titanosauriforms (for example).

I have one more thing for you to look for in your bird vertebrae, and that will be the subject of the next installment in this series. Stay tuned!


What it says on the tin. This is a specimen from the UCMP comparative collection.

I was just going to post this photo with zero commentary, but I can’t help myself. Note that on the two vertebrae in the middle, the crista transverso-obliqua (what in non-avian dinos would be the spinopostzygapophyseal lamina or SPOL) rises higher than either the neural spine apex or the epipophyses. That’s crazy. And it demonstrates something we also see in sauropods, which is that laminae are not merely the plates of bone left behind after pneumatization has scooped all of the unnecessary material out of a normal vertebra–sometimes they are additive structures, too.

If all of that sounded like gibberish, I can sympathize. I spent my first few months as a sauropodologist just learning the lingo (another thing I should blog about sometimes). Here’s a labeled version:

As long as I’m yapping, note the light shining through the honeycombed internal structure of these highly pneumatic vertebrae. For more on the ridiculous pneumaticity of pelican bones, see this post and this one. For more on the homology of bird and sauropod vertebrae, see Wedel and Sanders (2002), and for more on laminae as additive versus reductive structures, see the discussion on pages 210-212 of Wedel (2007).



In a comment on the last post, Mike wrote, “perhaps the pneumaticity was intially a size-related feature that merely failed to get unevolved when rebbachisaurs became smaller”.

Caudal pneumaticity in saltasaurines. Cerda et al. (2012: fig. 1).

Or maybe pneumaticity got even more extreme as rebbachisaurids got smaller, which apparently happened with saltasaurines  (see Cerda et al. 2012 and this post).

I think there is probably no scale at which pneumaticity isn’t useful. Like, we see a saltasaurine the size of a big horse and think, “Why does it need to be so pneumatic?”, as if it isn’t still one or two orders of magnitude more massive than an ostrich or an eagle, both of which are hyperpneumatic even though only one of them flies. Even parakeets and hummingbirds have postcranial pneumaticity.

Micro CT of a female Anna’s hummingbird. The black tube in the middle of the neck is the supramedullary airway. Little black dots in the tiny cervical centra are air spaces.

We’re coming around to the idea that the proper way to state the dinosaur size question is, “Why are mammals so lousy at being big on land?” Similarly, the proper way to state the pneumaticity question is probably not “Why is small sauropod X so pneumatic?”, but rather “Why aren’t some of the bigger sauropods even more pneumatic?”

Another thought: we tend to think of saltsaurines as being crazy pneumatic because they pneumatized their limb girdles and caudal chevrons (see Zurriaguz et al. 2017). Those pneumatic foramina are pretty subtle – maybe their apparent absence in other sauropod clades is just because we haven’t looked hard enough. Lots of things have turned out to be pneumatic that weren’t at first glance – see Yates et al. (2012) on basal sauropodomorphs and Wedel and Taylor (2013b) on sauropod tails, for example.

Back of the skull of a bighorn sheep, showing the air spaces inside one of the broken horncores.

Or, even more excitingly, if the absence is genuine, maybe that tells us something about sauropod biomechanics after all. Maybe if you’re an apatosaurine or a giant brachiosaurid, you actually can’t afford to pneumatize your coracoid, for example. One of my blind spots is a naive faith that any element can be pneumatized without penalty, which I believe mostly on the strength of the pneumatic horncores of bison and bighorn sheep. But AFAIK sauropod girdle elements don’t have big marrow cavities for pneumaticity to expand into. Pneumatization of sauropod limb girdles might have come at a real biomechanical cost, and therefore might have only been available to fairly small animals. (And yeah, Sander et al. 2014 found a pneumatic cavity in an Alamosaurus pubis, but it’s not a very big cavity.)

As I flagged in the title, this is noodling, not a finding, certainly not certainty. Just an airhead thinking about air. The comment thread is open, come join me.