Last Thursday I gave a public lecture for the No Man’s Land Historical Society in the Oklahoma Panhandle, titled “Oklahoma’s Jurassic Giants: the Dinosaurs of Black Mesa”. It’s now on YouTube, on the No Man’s Land Museum’s channel.

There’s a point I want to make here, that I also made in the talk: we can’t predict the value of natural history collections. The first sauropod vertebrae that Rich Cifelli and Kent Sanders and I CT scanned back in the spring of 1998 belonged to what would become Sauroposeidon, but most of the ones we scanned after that were Morrison specimens collected by J. Willis Stovall’s crews from the Oklahoma Panhandle between 1934 and 1941. Those scans formed the core of the pneumaticity research that fleshed out the Sauroposeidon papers (Wedel et al. 2000a, b), and was more fully developed in my Master’s thesis and the papers that came out of that (Wedel 2003a, b).

OMNH 1094, a mid-cervical vertebra of Brontosaurus in right lateral view. If you’ve seen one of my talks or my first few papers, you’ve seen this vert. I just realized that I have almost all the photos I need to do a proper multi-view; stand by for a future post on that.

So the foundation of my career was built in large part from collections that had been made 60 years earlier, decades before CT was invented. I’ll also note here that Xenoposeidon — Mike’s fourth paper (Taylor and Naish 2007), but the one which really launched his career as a morphologist — is based on a specimen collected in the 1890s. Natural history collections are not only resources for making comparisons, but also the engines of future discovery, and building and maintaining them is one of the most significant contributions to science that we can make.

I thank a bunch of folks at the end of the talk, but I especially want to thank Brian Engh for the use of his art, and Anne Weil for inviting me to collaborate on the sauropod material from the Homestead Quarry. Looking forward to more adventures!

References

Morphological variation in paramedullary airways; yellow = spinal cord, green = diverticula. The spectrum of variation is discretized into four groups: i, branches of intertransverse diverticula contact spinal cord at intervertebral joints; ii, branches of intertransverse diverticula extend partially into the vertebral canal, but remain discontinuous; iii, paramedullary diverticula form sets of tubes that are continuous through vertebral canals of at least two consecutive vertebrae; iv, continuous paramedullary diverticula anastomose with supravertebral diverticula. Each variant is depicted diagrammatically (A, dorsal view; B, E, H, & K, transverse view) and shown in two CT scans; images in each column correspond to the same morphology. Morphology i: C, cormorant; D, scrub jay. Morphology ii: F, bushtit; G, common murre. Morphology iii: I, red-tailed hawk; J, black-crowned night heron. Morphology iv: L, M, pelican. (Atterholt and Wedel 2022: figure 5)

New paper out:

Atterholt, Jessie, and Wedel, Mathew J. 2022. A computed tomography-based survey of paramedullary diverticula in extant Aves. The Anatomical Record, 1– 22. https://doi.org/10.1002/ar.24923

Quick aside, which will soon be of historical interest only: so far, only the accepted-but-unformatted manuscript is available, with the final, fully-formatted ‘version of record’ due along at some point in the future. We’re not sure when that will be — could be next week, could be months from now — which is why I’m following my standard procedure and yapping about the new paper now. This has paid off in the past, when papers that were only available in accepted ms form were read and cited before the final version was published. UPDATE on April 9: the formatted version of record is out now, as an open-access article with a CC-BY license, and I swapped it for the ‘accepted ms’ version in the links above and at the end of this post.

This paper has had a weirdly drawn-out gestation. Jessie and I hatched the idea of it way back in 2017, when we were teaching in the summer anatomy course together. I learned that Jessie had a big war chest of CTs of dead birds, and I’d been obsessed with supramedullary diverticula in birds and sauropods for some time already (e.g., an SVPCA talk: Wedel et al. 2014). There were detailed published descriptions of the supramedullary diverticula in a handful of species — namely chickens, turkeys, and pigeons — but no broad survey of those diverticula across living birds. Jessie had the CT scans to do that big survey, which we got rolling on right away. She presented our preliminary results at SVPCA in 2018 (Atterholt and Wedel 2018), and by rights the paper should have been along shortly thereafter. More on that in a sec.

One thing that may seem odd: we use the term ‘paramedullary diverticula’ instead of the more familiar and established ‘supramedullary diverticula’. That’s because these diverticula are not always dorsal to the spinal cord; sometimes they’re lateral, sometimes they’re ventral, and sometimes they completely surround the spinal cord, like an inflated cuff. So we decided that the term ‘paramedullary’, or ‘next to the spinal cord’, was more accurate than ‘supramedullary’, or ‘above the spinal cord’, for describing this class of diverticula.

Observed variation in the shape, arrangement, and orientation of paramedullary diverticula relative to the spinal cord; yellow = spinal cord, green = diverticula. A, paired diverticula dorsal to spinal cord in an ostrich. B, paired diverticula lateral to spinal cord in a bushtit. C, paired diverticula ventral to spinal cord in a violet turaco. D, three diverticula dorsal to spinal cord in an ostrich. E, four diverticula dorsal to spinal cord in an eclectus parrot. F, single, c-shaped diverticulum dorsal to spinal cord in an ostrich. G, diverticula completely surrounding spinal cord and pneumatizing vertebra in a violet turaco. H, no paramedullary diverticula present in a Pacific loon. I, diverticula completely surrounding spinal cord in a pelican. (Atterholt and Wedel 2022: figure 6)

I will have more to say about the science in other posts, and you can get the scientific backstory in this post and this one and the abstracts cited above and linked below. The rest of this post is mostly about me, so if you stick around, buckle up for some advanced navel-gazing.

There’s no one reason why this paper didn’t come out sooner. In short, I hit a wall. We went through a curriculum change at work, and suddenly the annual schedule that I’d relied on for a decade was completely upended. I had some unexpected challenges in my personal life. But the biggest problem was that my attitude toward research and writing had changed, for the worse.

When I was fresh out of grad school I had this kinda snotty attitude that my research was MINE, and wherever I was teaching was just, like, a paycheck, man, but they don’t own me, or my research. And as my teaching and committee responsibilities ramped up I still felt like research and writing was something I did for myself, and that my mission was to steal however many hours I could away from the “day-job work” to get done the things that I really wanted to do. Like a guerilla insurgency. In retrospect, it was a pretty good attitude for getting stuff done.

But somewhere along the way, I stopped thinking about research as something that belonged to me, something that I did for myself, and started thinking about it as part of my job. (This also maybe is not so flattering in what it reveals about how I think, or at least thought, about my job.) Instead of using my research time as a source of energy and a wellspring of satisfaction and positivity, I starting thinking of it only as a sink. And it happened so insidiously that I didn’t even realize it. My productivity plummeted, and I didn’t understand why. I was restless and depressed, and I didn’t understand that either. At the level of my superficial thoughts I still wanted to get research done, but my subconscious was turned off to it, so I just spun my wheels.

Then the pandemic hit. I’d always been a pretty optimistic, upbeat person, but I found myself just staring off into space franticizing about all the horrible things going on in the world, or staying up too late doom-scrolling the news. I slept too little, and poorly, and by the end of 2020 I felt worn down to a nub.

Osteological evidence of paramedullary diverticula. A, pocked texturing inside the vertebral canal of a pelican (LACM 86262). B, pneumatic foramen on the roof of the vertebral canal of an albatross (Phoebastria nigripes, LACM 115139). C, pneumatic foramina in the floor of the vertebral canal of an ostrich (Struthio camelus, LACM 116205). D, deep pneumatic fossae in the roof of the vertebral canal of an Eastern moa (Emeus sp., LACM unnumbered). (Atterholt and Wedel 2022: figure 7)

Then a series of positive things happened:

  • I received a long, heartfelt email from Jessie (fittingly!), asking after me and laying out a plan for getting the paper done and out. It was the kick I needed to look inside and start picking myself apart, to figure out what the heck was going on. Much of this post is cribbed from my reply to her.
  • I got a little break from lecturing in the spring of 2021, and that gave me the space to get a couple of things finished and submitted — the pneumatic variation paper with Mike in January (Taylor and Wedel 2021), and the Haplocanthosaurus neural canal paper, which was submitted even earlier in January, although it came out much later (Wedel et al. 2021; more on that publication delay in a future post).
  • Finally, I had young, energetic coauthors who were moving projects forward that required modest levels of effort from me, but which paid off with highly visible publications that I’m proud to be an author on, including the saltasaur pneumaticity paper (Aureliano et al. 2021) and the ‘Sauro-Throat’ paper (Woodruff et al. 2022).

It’s impossible to overstate how energizing it was to get new things done and out, and how much it helped to have collaborators who were putting wins on the board even when I was otherwise occupied. One of those collaborators was Jessie, who just kept pushing this thing forward — and, sometimes, pushing me forward — until it was done. So the paper you can read today is a testament not only to her acumen as a morphologist, but also to her tenacity as a scholar, and as a friend.

The part of the paper I’m happiest about is the “Conclusions and Directions for Future Research”, which points the way toward a LOT of further studies that need to be done, both on extant birds and on fossil archosaurs, ranging from bone histology to functional morphology to macroevolution. As we wrote in the concluding sentence of the paper, “We hope that this study serves as a foundation and an enticement for further studies of this most unusual anatomical system, in both extinct and extant archosaurs.”

I can’t wait to see what comes next.

References

For reasons that would be otiose, at this moment, to rehearse, I recently found myself in need of a hemisected turkey cervical. Happily, I own five skeletonised turkey necks, so it was with me the work of a moment to select a candidate. But now what? How to hemisect it? We have  discussed plenty of hemisected things here at SV-POW!, but they tend to have been produced using heavy machinery such as a band saw: something that I singularly lack.

SPOILER: I found a way. Here is a domestic turkey Meleagris gallopavo domesticus, 9th cervical vertebra, hemisected, in right medial view. Read on to discover the extremely high-tech approach that yielded this prize. It’s propped up on some kind of turkey bone to help me get a good medial perspective, I am thinking maybe the pygostyle?

One idea was to use an angle-grinder: not to cut down the midline of the vertebra — it would be much too blunt and powerful for a small, delicate vertebra — but to use as a sanding surface, locking the grinder in place and holding the vertebra up against the spinning plate. That might work well, assuming I could find a way to secure the angle grinder safely, but as it happened my need for a hemisected vertebra came up during a power cut. (Thanks, Storm Eunice!)

So I did it the way the Amish do their vertebral hemisections: by hand, simply by rubbing the vertebra against a sheet of sandpaper:

CT scanning: the Amish method.

This is not as laborious as you might think. I used a single sheet of medium-grade sandpaper, and it took maybe 15–20 minutes. And I just rubbed back and forth while exerting downward pressure. Initially I worked my way only through the prezygapophyseal ramus, which is the part of the turkey cervical that extends the furthest laterally. Once I was satisfied that the plane between eroded prezyg and the intact postzyg was parasagittal, I just kept the vertebra parallel to the sandpaper and kept rubbing. (Sorry I didn’t think to get a photo at this stage.)

One thing that took me by surprise is that there was so very much bone dust. I mean, I am an idiot that this surprised me, since the whole purpose of this exercise was to reduce one half of this vertebra to bone dust. But the lesson to be learned here is to do it on the easily-cleaned bathroom floor — not on the desk next to the computer keyboard and above a carpet. Learn from my mistakes, folks!

Anyway, after some work on the prezyg/postzyg pair, here’s how the vertebra was looking:

You can see straight away that the prezyg ramus, postzyg ramus and parapophyseal ramus are extensively pneumatized, honeycombed with small, irregular air-spaces. In this image it looks like the region of bone between the pre- and postzygs is much more solid, but this is an illusion: what we’re seeing here is a section through the cortical  bone of the neural arch, cut parallel to the surface. Let this be a warning not to over-interpret individual slices of CT-scans!

Once we get a little deeper, we see that the whole wall of the neural arch — and indeed the centrum and the neural spine — is honeycombed, just like the zyg rami:

Now we have another area of what I’m going to call Phantom Apneumaticity: the posterior part of the centrum looks like solid bone, apart from a few pneumatic spaces in the posteroventral extremity. Again, this is an illusion.

Here’s the next place I stopped:

Here, the Phantom Apneumaticity is even more striking: seeing just this as a CT slice could easily mislead someone into thinking that almost the whole of the posteroventral part of the centrum is solid bone. But again, it’s just that we’re very close to the surface of the bone, and seeing a slice parallel to that surface.

This last image also shows an important point of technique: there is a low convex ridge running across the phantom apneumatic area from the top of the cotyle to the base of the centrum. This is where I had changed the angle I was holding the vertebra at, so I accidentally sanded the posteroventral part of the vertebra more than the rest. I found that it was important during this process to keep checking the angles, and to adjust: making sure I wasn’t sanding more from the front than the back, or from the top than the bottom, or leaving a ridge like this.

Also in this last photo you can see that I was just beginning to break through into the neural canal: the anterior part of it is now exposed, between the anterior part of the neural spine and the anterior articular surface. At this stage I sighted along from in front to get a sense of how much further I had to go:

Domestic turkey Meleagris gallopavo domesticus, 9th cervical vertebra, most of right side removed, in anterior view with dorsal to the right. Propped up on the coracoid of a different, larger, turkey.

Quite a way, I guess. Here it is rotated and cropped, so you can more easily recognise it:

Domestic turkey Meleagris gallopavo domesticus, 9th cervical vertebra, most of right side removed, in anterior view. You can see that the neural canal is still mostly intact.

More sanding was required. I sanded some more.

You’ve already seen the final result up at the top of the page, but here is a cleaned-up version of that image, oriented according to Definition 3 of Taylor and Wedel (2019):

Domestic turkey Meleagris gallopavo domesticus, 9th cervical vertebra, hemisected, in right medial view.

And if that isn’t beautiful, what is?

The exciting thing is, anyone can make one of these. Matt’s already explained how to extract and clean up bird vertebrae and given you some ideas of what to do with them. Prepare out some turkey vertebrae and get going with the sandpaper!

I leave you with one more image: the hemisected vertebra in anterior view, oriented with dorsal to the top, and mirrored so it make up a complete vertebra once more. Enjoy!

References

 

I was looking more closely at the turkey skeleton from my recent post, and zeroed in on the last two dorsal (= thoracic) vertebrae. They articulate very well with each other and with the first vertebra of the sacrum, with the centra and zygapophyses both locking in so that there can only have been very little if any movement between them in life. Here they are, in right lateral view:

Last two dorsal (= thoracic) vertebrae of a mature domestic turkey Meleagris gallopavo domesticus, in right lateral view.

Before we move on, it’s worth clicking through to the full-size version of this image and wondering at both the quality of modern phone cameras (a Pixel 3a in this case) and the variety of textures on these little bones. There is smooth, finished bone on the sides of the neural spines; very fine pits and bumps on the zygapophyseal facets where the thin layer of hyaline cartilage attached; rougher texture in the parapophyseal facets where thicker cartilage attached; and very rough texture on the ends of the transverse processes, where there was relatively thick cartilage.

And there is, unsurprisingly in a bird, pneumaticity everywhere. In the more anterior vertebra alone (to the right) the photo shows pneumatic openings (from bottom to top) low on the centrum (below the parapophysis), high on the centrum (below the lateral process),  in the hollow between the lateral process, the posyzyg and the centrum, on the lateral surface of the prezygapophyseal ramus, and on the rear surface of the lateral process. There are others that are obscured in this photo, including on top of the lateral process where it meets the neural spine. Here they are, pointed out for you (with the hidden one shown translucently):

Last two dorsal (= thoracic) vertebrae of a mature domestic turkey Meleagris gallopavo domesticus, in right lateral view. Pneumatic openings on penultimate vertebra highlighted with red lines; obscured opening above lateral process shown as translucent.

OK, that was the B-movie. Now to the main feature. The next photo shows the same two vertebrae, folded away from each other so that we see the anterior face of the posterior vertebra (on the left) and the posterior face of the anterior vertebra (on the right).

Last two dorsal (= thoracic) vertebrae of a mature domestic turkey Meleagris gallopavo domesticus. Left: last dorsal vertebra in anterior view; right: penultimate dorsal vertebra in posterior view.

Again, do click through to see the exquisite detail, especially the complex of pneumatic features on the anterior face of the neural spine of the last dorsal (on the left) and on the posterior face of the left lateral process of the penultimate dorsal (on the right).

And … in the articular facets of the centra?

Seriously, what the heck is going on here? It doesn’t make sense  that there would be pneumatic openings in articular surfaces, because by definition something else (in this case the adjacent vertebra) is abutted hard up against then, so there is no way for a diverticulum to get in. For the same reason, you don’t get vascular foramina in articular surfaces because there is no way to get an artery in there. And there is no hint in these vertebrae of channels along either articular surface that diverticula or arteries could  possibly have laid in.

And yet, there those big openings are. What are they?

I discussed this with Matt, in case it’s Well Known Phenomenon that I’d somehow not heard about but it seems it is not. What we know for sure is that these openings are present, and that they are not mechanical damage inflicted during preparation. So what are they?

What else even is there for them to be? What penetrates bone apart from diverticula and blood vessels? Nerves follow the blood vessels, so it can’t be nerves in the absence of blood vessels.

By the way, there are similar but smaller openings in the posterior face of the last dorsal (the one on the right in the photo), but none anywhere else along the postcervical column: not on the anterior surface of the penultimate dorsal, not on the front or back of the sacrum, and not in any of the other dorsals.

One possibility we considered is that the vertebrae were locked together in life and that a pneumatic space inside the centrum of the last dorsal worked right through into the penultimate one. But that doesn’t work: the openings are not aligned. Also, those in the penultimate dorsal are definitely blind (i.e. they do not connect to deeper internal air-spaces) and those in the last dorsal probably are, too.

We do not know what is going on here.

Help us! Is this kind of thing common in turkeys? Have people seen it in other taxa? Do we know what it is?

Naturally I was grateful when Cary invited me to be part of the team working on Dolly, the diplodocid with lesions in its neck vertebrae (Woodruff et al. 2022; see previous posts on Dolly here and here). I was also intellectually excited, not only to see air-filled bones with obvious pathologies, but also for what those pathologies could tell us about Dolly and other sauropods. That’s the part of our new paper I want to unpack in this post.

We have a lot of evidence that air-filled bones in birds are a good model for air-filled bones in extinct dinosaurs. And we have several lines of evidence (not just air-filled bones; see Schachner et al. 2009, 2011, 2020) that the respiratory systems of many dinosaurs functioned broadly like those of living birds. But we have less direct evidence than we’d like, so every additional bit of information is welcome.

Diving into Diverticula

In birds, the air-filled bones in the neck and body are connected to the respiratory system by air-filled tubes. These tubes sometimes get called air sacs, in the sense that they are sacs filled with air, but we also refer to them pneumatic diverticula, to distinguish them from the respiratory air sacs in the torsos of birds that ventilate the lungs. Imagine blowing up some rubber gloves and sticking them inside a bird* and you’ll have a pretty good mental model of the system — the inflated ‘palm’ area of each glove is like one of the respiratory air sacs, the air-filled glove fingers are the diverticula, and the rubber material of the glove is the pneumatic epithelium that lines the air sacs and the diverticula alike. 

* Please don’t actually try that.

The cartoon above presents an unrealistically simplified picture of the respiratory system in dinosaurs and birds. For one thing, I omitted the windpipe or trachea — that blue tube going up the neck represents the diverticula that run alongside, and often inside, the neck vertebrae, parallel to the trachea but separated from it by whole sheets of muscle. Also, the cartoon only shows diverticula of the air sacs, but diverticula can also originate from the lungs themselves (see O’Connor 2006: 1211 and Schachner et al. 2020: 16-19). Here’s the actual respiratory system of a pigeon, with the trachea and lungs in pink and the air sacs and their diverticula in blue (Muller 1908 fig. 11):

I think it’s pretty natural to look at that illustration and wonder where the heck the guts go, since it certainly looks like the air sacs are occupying the entire volume of the torso. The answer is that the air sacs enclose the viscera “as do the shells of a nut”, in the memorable formulation of Wetherbee (1951: p. 243), describing the air sacs of the English sparrow.


Fig. 4. Reconstruction of the distribution of pneumatic diverticula in diplodocids and dicraeosaurids. A. Schematic drawing of midcervical vertebra of Diplodocus in left lateral aspect (A1), in dorsal aspect with single neural spine (A2) and in dorsal aspect with bifurcate neural spine (A3). The partitioning of pneumatic diverticula at the lateral surface of the vertebral corpus is hypothetical, based on the strongly divided pneumatic fossae. Schwarz et al. (2007: fig. 4A).

Furthermore, the pneumatic diverticula around the vertebrae in birds are complex, and we are fairly certain that they were also complex in sauropods, because they left so many distinct traces. The most detailed reconstructions of the cervical diverticula in sauropods that I know of are those of Daniela Schwarz and colleagues (2007), as shown above. For what it’s worth, I think those reconstructions are not just reasonable but perhaps conservative; I think there’s a good chance that the diverticular network around the vertebrae was even more complex and extensive. 

The rubber-glove model also lets us see that the diverticula are cul-de-sacs. We know that diverticula can anastomose, or merge, to form networks, and there is a possibility that if diverticula from different air sacs anastomosed, different pressures in those air sacs might allow some air to circulate through the diverticular network. Maybe — the anterior and posterior air sacs fill and empty at the same time, so there might not be a pressure differential to exploit. If air circulates in the diverticula at all in birds, it probably happens in the dorsal vertebrae, where diverticula from different parts of the respiratory system have the best opportunity to anastomose. But the far ends of the diverticular network are always dead ends, and we assume that air diffuses in and out of those terminal diverticula fairly slowly. We’re stuck with assumptions because no-one’s ever checked, experimentally, to determine the rate of diffusion or circulation of air in the diverticula. But it’s hard to imagine much circulation in the terminal diverticula, with no air reservoir or pump at the far end.

Reconstruction of soft−tissues in the neck of Diplodocus. A. Transverse cross−sections through cervical vertebra with bifurcate neural spine in the diapophysis region (A1) and in caudal third of vertebra (A2). B. Transverse cross−sections through cervical vertebra with single neural spine in diapophysis region (B1) and in caudal third of vertebra (B2), dashed outlines representing possible craniocervical extensor muscle analogous to m. biventer cervicis of extant birds or m. transversospinalis capitis of extant crocodylians. Schwarz et al. (2007: fig. 7A-B).

Here’s another great illustration from Schwarz et al. (2007), showing cross-sections of the neck of Diplodocus with hypothetical soft tissues restored. Bone is black, muscle is pink, and the pneumatic diverticula are blue. As this diagram makes clear, the air spaces in the bones are themselves extensions of the diverticula (that much is true regardless of how extensive we make the reconstructed diverticula outside the vertebrae). Instead of smooshing an inflated rubber glove into a duck, imagine smooshing one into a vertebra of a duck — or a Diplodocus — so that all of the empty spaces are occupied by some blobby bit of inflated-glove finger. All of air spaces in the bone would be lined by rubber-glove material, which in this metaphor is the same pneumatic epithelium that lines both the respiratory air sacs and their pneumatic diverticula, outside the bones or inside them.

I get to see this firsthand in the gross anatomy lab in our unit on head and neck anatomy. As we open up the skulls of the cadavers, the air-filled epithelial balloons that fill the sinuses sometimes pull away from, or completely out of, their bony recesses. (I’ll bet I could demonstrate the same thing with the sinuses of a pig or sheep head — I should give that a shot and post the resulting photos or videos here.) The point is, the pneumatic epithelium is in intimate contact with the bone, lining every pneumatic fossa, foramen, and internal chamber; this will be really important later on.

Incidentally, one question I get a lot is whether the diverticula, inside or outside the bones, contributed to gas exchange in sauropods. The answer is, probably not. We know from dissections and histological work on birds that the respiratory air sacs, their diverticula, and the diverticular spaces inside the skeleton are all relatively avascular, meaning that the tissues get enough blood to stay alive, but aren’t specialized for gas exchange. Furthermore, physiological experiments on living birds have shown that about 95% of the gas exchange happens in the lungs, and almost all of the remaining 5% happens in the paired abdominal air sacs (Magnussen et al. 1976), probably because they are so large and so intimately in contact with the guts (Wetherbee’s nutshell metaphor), which are well-supplied with blood. We also know from bone histology that the air-filled bones of extinct dinosaurs are essentially identical to those of modern birds (Lambertz et al. 2018), so there’s no evidence that they functioned any differently.

A simplified diagram of the sauropod respiratory system. What I’ve labeled “air tubes” here are the pneumatic diverticula. Air holes in the vertebrae are also known as pneumatic foramina. The shapes of the lungs and air sacs are speculative, but the minimum extent of the pneumatic diverticula is not–although it could be an underestimate (e.g., diverticula might have gone even further down the tail, and just not left any diagnostic traces on those vertebrae).

A final piece before we get back to Dolly: we know from lots of anecdotal observations, and some actual experiments, that air-filled bones have to stay connected to the outside to form in the first place, and to stay healthy afterward. This is true of both human sinuses and postcranial pneumatic bones in birds, so it’s reasonable to assume that it’s a general feature of all air-filled bones (see Witmer 1997 for lots of relevant citations and discussion). This is a pretty handy thing to know, because if we find an air-filled vertebra way out in the tail, we know pneumatic diverticula of the respiratory system got at least that far. ‘At least’ because pneumatic diverticula can make diagnostic traces on bones, but they don’t always do so. That means that the diverticular network can easily be more extensive than its skeletal traces, but not less so — see Wedel and Taylor (2013) for more on that.

To sum up, we suspect the following things about pneumatic diverticula around the vertebrae of sauropods, including Dolly:

  1. The diverticula were complex, based on the traces they left on the bones, and similarly complex diverticula in birds.
  2. The diverticula were patent, or open, maintaining an open connection to the outside by way of the respiratory air sacs, lungs, and trachea, because that’s how air-filled bones work in living birds and mammals.
  3. Despite being complex and ultimately open to the outside in one direction, the diverticula were cul-de-sacs, with little or no active circulation of air — especially in the neck.

With that in mind, what does the distribution of infected bone in Dolly tell us about sauropods?

Infections and inferences

Here’s another illustration of the respiratory system of the pigeon, this time a dorsal or top-down view, from Muller (1908: fig. 12):

Okay, that was a bit of a bait-and-switch: I promised you Dolly and gave you another pigeon. But that’s only to help you understand this similar cartoon I drew, which represents Dolly’s respiratory system and neck vertebrae, also seen from the top down:

Like the earlier cartoon, this is pretty simplified. For instance, I got lazy and didn’t draw all of the neck vertebrae. In life, Dolly probably had 15 or 16 neck vertebrae, like other diplodocids, and we know that the three with lesions are C5-C7 because they were found articulated. Here I drew just enough vertebrae to make my points, and I left off the head and all the other extraneous bits. Also, I’ve drawn the diverticula that run up the neck originating from cervical air sacs, as in pigeons (Muller 1908), but there is evidence that in ostriches those diverticula may originate from the lungs themselves (Schachner et al. 2020). Whether the diverticula come from the lungs or the air sacs is probably not an answerable question for sauropods, and for my purposes here, it doesn’t matter, only that the diverticula are connected back to the core respiratory system.

Three things struck us about the distribution of the infected bone in Dolly’s neck:

  1. The lesions are all in vertebrae that are a long way up the neck, far from the lungs and respiratory air sacs in the torso.
  2. The lesions are clustered in serially-adjacent vertebrae, instead of being scattered up and down the neck randomly.
  3. The lesions are present bilaterally, on both left and right sides of the affected vertebrae.

Well, as opposed to what? We can imagine a scenario in which the lesions were scattered randomly, not just up and down the neck, but also on left and right sides, like so:

If the infection had been carried in the blood, we might expect such a random pattern. In that case, it would be an extreme coincidence if a blood-borne infection, which could go anywhere in the body, only manifested in the air spaces on the sides of three consecutive vertebrae. The clustering of the Dolly’s lesions, in the air spaces on both sides in three adjacent vertebrae, far up the neck, points to a different cause.

Recall that diverticula are lined by epithelium, and that in air-filled bones, the epithelium is right up against the bone tissue. The infection in Dolly almost certainly started out as an infection in the diverticulum, which was so severe that it spread to the underlying bone. In exactly the same way that the air spaces in the bones are the skeletal footprints of the diverticula, the lesions in Dolly’s vertebrae are the skeletal footprints of infected epithelium lining the diverticula, like so:

The infection may have gotten so severe, far up Dolly’s neck, precisely because there was little airflow so far from the lungs and air sacs. Airborne bacteria or fungal spores could have floated into the diverticula by diffusion, come to rest against the epithelium in warm, dark, humid conditions, and gone wild. It’s also possible that a huge swath of Dolly’s respiratory system was infected, but the infection only got severe enough to spread to the underlying bone in cervical vertebrae 5-7, in which case the actual infection might have looked something like this:

Just like a diverticulum can contact a bone without producing a distinct trace, the pneumatic epithelium could be infected without producing a bony lesion. Thought experiment: how many times have you had a sinus infection, and how many people do you know who have had sinus infections? And how many of those sinus infections were severe enough, and lasted long enough, to produce bony lesions like we see in Dolly? Probably very few — such things do happen in humans, and the medical literature has plenty of cases (and if this has happened to your or a loved one, you have my full sympathy) — but on a population level, the fraction of respiratory infections that produce bony lesions is miniscule. Similarly, it’s very likely that much more of Dolly’s respiratory system was infected than we can tell from the skeleton.

A cervical vertebra of an ostrich with some of the pneumatic diverticula traced on.

The presence of infected bone on both left and right sides of C5-C7 in Dolly is also telling. If the diverticula on the left and right sides of the neck were separate, the symmetrical pattern of infection would be another extreme coincidence. But in birds there are opportunities for diverticula from the left and right sides of the neck to meet and anastomose, especially the supravertebral diverticula on the neural arch (shown above), and the supramedullary diverticula inside the neural canal. Based on pneumatic traces on the vertebrae we infer that the same diverticula were present in sauropods, as shown up above in the Diplodocus figures from Schwarz et al. (2007), and those left-and-right communications probably allowed the infection to develop more or less symmetrically. Or to put it another way, the symmetrical infections are additional evidence that the diverticula on the left and right sides of the neck were connected across the midline, and birds show that there are several ways that could have happened.

CT scans of cervical 7 of MOR 7029. Photograph and scan model of the vertebra ((A,B) respectively). The colored lines in (B) correspond to the scan slices (and scan interpretative drawings below). White arrows point to the external feature, while black arrows denote the hyperintense bone and irregular voids. (C) Comparison of the abnormal tissue composition of MOR 7029 (left), compared to that of a ‘normal’ diplodocine (right). Text and white arrows indicate the various features different shared/differentiated between the two. For the interpretative drawings, white = ‘normal’ bone, grey = hyperintense bone, black = irregular voids. Woodruff et al. (2022: fig. 2).

Another possibility is that a good chunk of the internal structure Dolly’s vertebrae was infected, and the lesions that we see on the surface are just the groady tips of big, disgusting icebergs of infected bone. In fact, that’s pretty much what the CT scans show. So possibly the infection started on one side of each vertebra and basically burrowed through to reach the other side. That would probably take weeks or months, whereas the infection could have spread across the midline through diverticula in hours or days, so I think the latter scenario is still the most plausible explanation for the presence of the lesions on both sides of the affected vertebrae.

In summary, I don’t think Dolly tells us anything surprising that we didn’t suspect before. Rather, the pattern of infection in Dolly makes perfect sense if the diverticula of sauropods were essentially bird-like, and that pattern is difficult to explain any other way.

Finding skeletal traces of a respiratory infection in Dolly was still a crazy lucky break, and that’s something I’ll discuss more in the next post in this series.

References

Turkey skeleton audit

February 16, 2022

Back in at least 2008 — maybe earlier — I kept all the bones from our good-sized Christmas turkey. Of course, it’s missing the head, neck and feet, but otherwise it’s pretty much all there. (I may also have the neck, but if so then it was supplied as a separate item, and prepared separately.) Here is the box of postcervical bones:

Postcervical skeleton of a mature domestic turkey Meleagris gallopavo domesticus, complete except for feet.

As I was transferring them to a better box today, it occurred to me to lay them  out and see how much sense I could make of them. Here’s what I did with the bones I was happy about:

Postcervical skeleton of a mature domestic turkey Meleagris gallopavo domesticus, complete except for feet. Some small bones omitted (see below). Laid out roughly as in life, in dorsal view.

I should  have put something in the photo to act as a scale-bar, because it’s not apparent from this photo that a turkey is a pretty big thing. From top to bottom of the skeleton as I laid it out here is about 90 cm.

Here are my (in some cases tentative) identifications of the bones:

Postcervical skeleton of a mature domestic turkey Meleagris gallopavo domesticus, complete except for feet. Some small bones omitted (see below). Laid out roughly as in life, in dorsal view. Bones are tentatively labelled. DO NOT USE FOR TUTORIAL PURPOSES.

Are there any obvious mistakes in there? And have I got any of the bones left-right reversed?

Now, here are all the other bones that I was less confident about the positions of:

Postcervical skeleton of a mature domestic turkey Meleagris gallopavo domesticus, complete except for feet. Small bones only.

On the left of course we have the dorsal ribs, but I’ve not been able to arrange them all into near pairs, nor figure out what order I should impose on the pairs that I do have. I’m not even sure how many pairs I should have. On the right are other paired bones whose identity I can’t figure out. I am guessing that the longer ones are probably sternal ribs and that the irregularly shaped ones might be parts of the wrist, but I would welcome corrections and clarifications. Finally, the middle column contains the bones whose idea I have little or no idea about, and which don’t appear to be paired. Any ideas?


By the way, I found this image useful in figuring out the identities of the appendicular bones:

Skeletal reconstruction of a domestic turkey, published in 1808. Source unknown (leave a comment if you can identify it).

And this one useful for the bones of the wing and especially the hand:

Human and bird arm skeletons compared. Source unknown.

More from this skeleton another time!

Skull audit: Wedel responds

February 9, 2022

Left to right: alligator, beaver, black bear, armadillo, cat, ostrich. I know, the archosaurs aren’t mammals, and the alligator isn’t even a skull. But if you can’t have a lounge lizard crash your mammal skull party, what are you even doing with your life? Not pictured: about four rabbit skulls I forgot I had boxed up, plus a couple of turtles (yeah, yeah) sitting on a friend’s desk, in their locked office.

It warmed my crooked little heart to see Mike Taylor, noted sauropodologist and disdainer-of-mammal-heads, return mammal skulls to the blog’s front page yesterday. Naturally I had to support my friend and colleague in this difficult time, when he may be experiencing confusing feelings regarding nasal turbinates, multi-cusped teeth, and the dentary-squamosal jaw joint.

My skull collection is split across home and office, but I had to go in to campus this afternoon for a video recording thing, so I got most of the office set, shown above, on that jaunt.

After the workday ended, I had just enough time before the light faded to assemble and photograph the home collection:

Back row: peccary, pig, deer, sheep, dog. Middle row: opossum, rabbit. Front row: opossum, marten (both hemisected). Not pictured: emergency backup sheep, moar rabbits

I’ve blogged about the bear, the pig, and the hemisected skulls, but I think that’s it. I should do more skull blogging, most of these have a story:

  • I prepped the armadillo, cat, rabbit, and sheep skulls myself (besides the bear and pig). The first two I found in the woods, the mostly-decomposed rabbit was a gift from my father-in-law, and the sheep head I obtained from the market down the street ($10, and I ate the meat).
  • The alligator head and deer skull were gifts, from Vicki and from my brother Ryan, respectively.
  • The rest I purchased here and there over the years, usually when they were on deep discount. The peccary is a memento of a trip to Big Bend back in 2007 (I bought it at a taxidermy shop a long way outside the national park), and the dog came from the seconds bin at the Museum of Osteology — I plan to saw off the top of the braincase to see the cranial nerve exits, just as in the preparation by Peter Dodson shown in this post.

I have more heads awaiting skull-ization in various freezers, too. Couple more pig heads at work, and at the house a strategic reserve sheep head, plus skunk, squirrel, and rat. Plus a partially-mummified but mostly defleshed armadillo whose saga deserves a detailed recounting:

NB: the stray bits toward the bottom of the image are from a cat. Mr. Armadillo’s limb bones and vertebrae are still in the armadillo kit.

In the first comment on Mike’s post yesterday, I expressed envy that he had the better skull collection. After pulling together all my critters, I think I just have a worse memory. In my defense, it’s been almost two years since I was in the office regularly, and about half the skulls in the home collection are recent-ish acquistions (~last three years), so a lot of stuff had either fallen out of memory or not gotten properly established yet. But Mike has definitely prepped more — and more exotic — skeletons, and it was his enthusiastic collecting and blogging of dead animal bits that inspired me to start my recent-ish spate of skull preparations. More to come on that front as time and opportunity allow, probably starting with this:

 

Windpipe and lungs in pink, air sacs in teal. Steps 1 and 3 happen at the same time — one breath of air is moving through the lungs and into the air sacs in back (1) at the same time as an earlier breath of air is moving out of the lungs and into the air sacs up front (3). Steps 2 and 4 happen at the same time as well — the air sacs in back are blowing air through the lungs (2) while the air sacs in front are blowing air out the windpipe (4). Each breath of air is inside the bird for two inhalations and exhalations.

Our lungs are made up of millions of tiny bags. Breath in, fill the bags with fresh air, breathe out, empty the bags of spent air. But bird lungs are very different. They’re made up of millions of tiny tubes, like bundles of drinking straws, and those tubes are connected to big, empty air sacs, like balloons that spread throughout the body. When birds breathe in, some of the air goes through the lungs, and some skips the lungs and goes into the air sacs. Then when the bird breathes out, the air in the air sacs gets pushed through the tubes in the lungs. So birds get oxygen-rich air blown through their lungs both when they inhale and when they exhale. The lungs and air sacs of birds also send mini air sacs into the skeleton, and these create air-filled spaces inside the bones, akin to our sinuses. These air spaces in the skeleton are the footprint of the respiratory system. A lot of extinct dinosaurs have the same pattern of air spaces in their skeletons, so we think they breathed like birds.

— Jessie Atterholt and Matt Wedel

Starlings are amazing

October 29, 2021

Back in May, Amy Schwartz posted a photo of a starling that shethat had ringed that morning:

Impressed by the subtlety of the coloration, I wondered what would happen if I increased the colour saturation. I did this very simply: in the free image editor GIMP, I selected the parts of the photo that were starling (omitting the human hand and the background), and using the Hue-Saturation tool I wound the saturation up to 100%. Then I did the same thing again. Here is the result, with no other editing at all:

What an extraordinary riot of colour, in a bird that we mostly think of as “basically black with dots.”

So I thought I’d try the same trick on another starling photo, this one from the All About Birds page on the European Starling. Here is the original:

And here is the result of saturating the colours — this time through three cycles.

So my question is this: can other starlings see all this colour? In their own closed starling-centric world, are they fabulously colourful? Is this something close to what is perceptually apparent to animals whose eyes are attuned to different wavelengths from ours?that

Anatomical features of the neural canal in birds and other dinosaurs. A. MWC 9698, a mid caudal vertebra of Apatosaurus in posterodorsal view. Arrows highlight probable vascular foramina in the ventral floor of the neural canal. B. LACM 97479, a dorsal vertebra of Rhea americana in left anterolateral view. Arrows highlight pneumatic foramina inside the neural canal. C. A hemisected partial synsacrum of a chicken, Gallus domesticus, obtained from a grocery store. Anterior is to the right. The bracket shows the extent of the dorsal recess for the glycogen body, which only spans four vertebrae. Arrows highlight the transverse grooves in the roof of the neural canal for the lumbosacral organ. D. Sagittal (left) and transverse (right) CT slices through the sacrum of a juvenile ostrich, Struthio camelus. The bracket shows the extent of the lumbosacral expansion of the spinal cord. Indentations in the roof of the neural canal house the lumbosacral organ. In contrast to the chicken, the ostrich has a small glycogen body that does not leave a distinct osteological trace. Yellow arrows show the longitudinal troughs in the ventral floor of the neural canal that house the ventral eminences of the spinal cord. Wedel et al. (2021: fig. 4).

This is the second in a series of posts on our new paper about the expanded neural canals in the tail vertebrae of the Snowmass Haplocanthosaurus. I’m not going to talk much about Haplo in this post, though. Instead, I’m going to talk about chickens, and about how you can see a lot of interesting spinal anatomy in a living dinosaur for about two bucks.

You know by now that Academia Letters publishes peer reviews, which is one of the things that drew me to this fairly new journal. More on that in a later post, but in the meantime, the peer reviews for the Haplo paper are on the right sidebar here. I confess, I had a total forehead-slap moment when I read the opening lines of Niels Bonde’s review: 

This paper is interesting, and should be published and discussed by others with interest in dinosaur-bird relations. However, as these publications are also meant for the general public, I would recommend that 2 – 3 illustrations were added of the features mentioned for birds under nos. 3 – 6, because the general public (and many paleontologists) have no ideas about these structures, and what they look like.

The original submission only had figures 1 and 2. And this request is totally fair! If you are going to discuss six alternative hypotheses for some mysterious anatomical structure, it’s just responsible reporting to illustrate those things. That goes double if, as Niels Bonde noted, the anatomy in question is unfamiliar to a lot of people, even many paleontologists. Huxley’s quote after first reading Darwin’s Origin of Species flashed through my head: “How extremely stupid not to have thought of that.”

Slide 21 of my 2014 SVPCA talk on supramedullary diverticula in birds and other dinosaurs, illustrating pneumatic foramina in the roof, walls, and floor of the neural canal.

At the time I read that review, I already had images illustrating five of the six hypotheses. A juvenile ostrich synsacrum that Jessie Atterholt and I had CT scanned gave us three of them all by itself: the lumbosacral expansion of the spinal cord to run the hindlimbs, as in all limbed tetrapods and in some fish with sensitive fins; the transverse channels in the dorsal wall of the neural canal to accommodate the lumbosacral balance organ; and the paired troughs in the floor of the neural canal that house the ventral eminences of the spinal cord (Figure 4D in the image at the top of this post). I had good photos of pneumatic foramina in the walls and floor of the neural canal in a dorsal vertebra of a rhea from my 2014 SVPCA talk (Figure 4B), and some photos of small foramina, presumably for blood vessels rather than air spaces, in the floor of the neural canal in a caudal vertebra of Apatosaurus (Figure 4A).

What I did not have is a photo illustrating the fairly abrupt, dome-shaped space in the sacral neural canal that houses the glycogen body of birds. I mean, I had published images, but I didn’t want to wrestle with trying to get image reproduction rights, or with redrawing the images. Instead, I went to the grocery store to buy some chicken.

I don’t know how universally true this is, but IME in the US when you buy a quartered chicken, the vertebrae are usually nicely hemisected by the band saw that separated the left and right halves of the animals. So you can see the neural canal in both the dorsal and sacral parts of the vertebral column. Here are the hemisected dorsal vertebrae in the breast quarter from a sectioned rotisserie chicken:

That’s just how it came to lie on my plate, but it’s not in anatomical position. Let’s flip it over to sit upright:

And label it:

I could and probably should do a whole post just unpacking this image, but I have other fish to fry today, so I’ll just note a couple of things in passing. The big interspinous ligament is the same one you can see in transverse section in the ostrich dissection photos in this post and this one. Also, the intervertebral joints heading toward the neck, on the left of the image, have much thicker intervertebral cartilage than the more posterior dorsals. That’s because the posterior ones were destined to fuse into a notarium. You can see a diagram and a photograph of a chicken notarium in figures 4 and 5, respectively, here. And finally, the big takeaway here is that the neural canal is normal, just a cylindrical tube to hold the spinal cord.

The thigh quarter usually has the pelvis and the hemisectioned synsacrum attached. Here’s a lateral view of the left half of the pelvis and synsacrum:

And the same thing labeled:

And now flipped around so we can see it in medial view:

And now that image labeled:

And, hey, there are three of our alternative hypotheses on display: the long (many vertebral segments) lumbosacral expansion of the spinal cord, which is reflected in a gradually expanded neural canal in the synsacrum; the shorter, higher dome-shaped recess for the glycogen body; and finally the transverse spaces for the lumbosacral balance organ.

As a refresher, there’s nothing terribly special about the lumbosacral expansion of the spinal cord — you have one, labeled as the ‘lumbar enlargement’ in the above diagram. Where the spinal cord has adjacent limbs to run, it has more neurons, so it gets fatter, so the neural canal gets fatter to accommodate it. The cord itself doesn’t look very expanded in the chicken photo above, but that chicken has been roasted rotisserie-style, and a lot of lipids probably cooked out of the cord during that process. What’s more important is that the neural canal is subtly but unmistakably expanded, over the span of many vertebrae.

The lumbosacral spinal cord of a 3-week-old chick in dorsal view. The big egg-shaped mass in the middle is the glycogen body. Watterson (1949: plate 1).

That’s in contrast to the recess for the glycogen body, which is colored in blue in the chicken photo. Glycogen bodies, like the egg-shaped one in the young chicken in the image immediately above, tend not to go on for many vertebral segments. Instead they balloon up and subside over the space of just 4 or 5 vertebrae, so they leave a different skeletal trace than other soft tissues.

Finally, there are the transverse spaces for the lumbosacral balance organ, which I discussed in this post. Those are the things that look like caterpillar legs sticking up from the sacral endocasts in the above figure from Necker (2006). In life, the spaces are occupied by loops of meningeal membranes, through which cerebrospinal fluid can slosh around, which in turn puts pressure on mechanoreceptive cells at the edge of the spinal cord and gives birds a balance organ in addition to the ones in their heads. In the photo of the cooked chicken, the delicate meninges have mostly fallen apart, leaving behind the empty spaces that they once occupied.

I really liked that chicken synsacrum, and I wanted to use it as part of Figure 4 of the new paper, but it needed a little cleaning, so I simmered it for a couple of hours on low heat (as one does). And it promptly fell apart. At least in the US, most of the chickens that make it to table are quite young and skeletally immature. That particular bird’s synsacrum wasn’t syn-anything, it was just a train of unfused vertebrae that fell apart at the earliest opportunity. I had anticipated that might be an issue, so I’d gotten a lot of chicken, including a whole rotisserie chicken and four thigh quarters from the deli counter at the local supermarket. Happily this fried chicken thigh quarter had a pretty good neural canal:

And it cleaned up nicely:

And with a little cropping, color-tuning, and labeling, it was ready for prime time:

I didn’t label them in the published version, for want of space and a desire not to muddy the waters any further, but the jet-black blobs I have colored in the lower part of that image are the exit holes that let the spinal nerves out of the neural canal so they could go serve the hindlimbs, pelvic viscera, and tail. We have them, too.

At my local grocery store, a fried chicken thigh costs about $1.65 if you get it standalone, or you can buy in bulk and save. You get to eat the chicken, and everything else I’ve done here required only water, heat, soap, and a little time. The point is that if I can do this, you can do this, and if you do, you’ll get to see some really cool anatomy. I almost added, “which most people haven’t seen”, but given how much chicken we eat as a society these days, probably most people’s eyes have fallen on the medial surface of a cooked chicken thigh quarter at one time or another. Better to say, “which most people haven’t noticed”. But now you can. Go have fun. 

Way back in January of 2019, I finished up “Things to Make and Do, Part 25b” with this line: “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!” Here we are, 2.3 years later, and I’ve finally made good. So if there’s a promised post you’ve been waiting for, stick around, we may get to it yet.

References