Since I started taking photographs of sauropod vertebrae back in 2004, I’ve got much, much better at it, and for the last few months I’ve been meaning to write an article about what I’ve learned along the way.  A few weeks ago, fellow SV-POW!er Ranger Matt Wedel posted an article on his 10 Minute Astronomy blog on how to photograph the moon through binoculars, and that served as a prod to get back into blogging gear in the post-Christmas season.

Before I launch in, let me be really clear that I am not a proper photographer — not at all.  I don’t even know what an F-stop is or what Single Lens Reflex means.  Probably I should invest some time into learning some of this, since specimen photographs are so important in the world of sauropod vertebrae.  (After all, the specimens are more than a little cumbersome to loan, so photos often have to stand as proxies for the actual specimens.)  Nevertheless, what I’ve learned in the last five or six years has got me to the point where I am producing much, much better specimen photographs than when I started, and I hope at least some of you can benefit from what I’ve learned.

The very best (and still very bad) of the first batch of Archbishop photographs I took, back in July 2004. Note that it's not square on, doesn't fit in the frame, that it's over-exposed and (as you'll see if you click through to the full-sized version) both blurry and infested with artifacts. Compare with the recent photo at the end of this article. Copyright the NHM since it's their material.



First up, get a decent camera.  However skilled you are, you can’t take better photos than the hardware allows.  Although I am to blame for the composition above and for some of blurriness, the over-exposure, poor definition and artifacts are the fault of the camera.  I was using a truly horrible camera back then — some super-cheap list-of-features-on-a-discount-website piece of kit.

The good news is that a “decent” camera doesn’t need to break the bank: for our purposes you don’t need to spend a fortune on professional-photographer standard equipment.  I am looking on ebay right now, and it seems you can get my model of camera for £100 in the UK or $150 in the US (second-hand of course) which is a level of investment we really should be prepared to put into one of the most important aspects of descriptive work.

What constitutes a decent camera?  Mostly, optics.  These days, every camera has more than enough megapixels for most purposes, so you can just forget about that statistic altogether.  It’s about the quality of the lens and the size of the CCD — those are the factors that determine how much information the camera can capture, and if it puts out more bits than that, then all it’s doing is wasting disk-space and bandwidth.

Can I justify the claim that all modern cameras have enough megapixels?  I think so.  Suppose you’re preparing a full-page plate for the Journal of Vertebrate Paleontology.  In practice, plates are nearly always composites of several photos, but suppose you want a single shot filling the whole plate.  The printable area of a JVP page is 182 x 233 mm, which is 7.2 x 9.2 inches.  At 300 dpi, that’s 2161 x 2752 pixels, which is 5947072, or a slice under 6 megapixels.  So 6 Mp is enough for a full-page plate.  (For what it’s worth, my camera does 2272 x 1704 = 3.8 megapixels, and I have never found myself feeling a need for more resolution.)

For the same reason, you definitely want optical zoom rather than digital zoom, which really amounts to just blowing up the image.


Another big win: get a spare battery, so that one can be recharging while you’re using the other.  If you don’t do that, your camera is out of commission half the time.

And get a big enough memory card.  What’s “big enough”?  For me, that means enough space to hold a whole day’s images so I can do a single dump onto the laptop in the evening, rather than having to keep stopping to transfer.  I can take maybe a maximum of 300 photos a day.  With 1 Mb images, that means I need a 300 Mb card, which is chickenfeed.  You literally can’t buy cards that small any more, so this is not really a factor these days and I might just as well not have mentioned it.  (The reason I did mention it is that my camera originally came with a 16 Mb card or something similarly stupid, which meant ten minutes or so of photography before downloading.)

Horrible photograph of a Brachiosaurus altithorax dorsal (holotype specimen FMNH P25105, natch), showing how NOT to compose a picture.


In the photo above, I did everything wrong.  The vertebra is cropped partly out of the frame, it’s viewed from an uninformative angle, it has a scalebar obscuring part of the bone, and the background is a mess.  Here are five simple rules to avoid badgering it up like I did here:

Get the specimen in frame

I know it sounds obvious, but I can’t tell you how many times I’ve reviewed my photos, picked one that is good in other respects, and realised that I’ve trimmed a bit off the end of a diapophysis or something.

Shoot from cardinal directions

Also  really important.  I am not (of course) saying that you should never get photos from any directions but the cardinals, but if you come home from photographing a vertebra and you don’t have shots from in front, behind, above and left and right lateral, you’d better have a good reason why not.  Only by getting all of these can you make informative composites like the ones of the Archbishop that I’ve been posting lately.

Don’t put anything in front of the specimen

Again this sounds terribly obvious, but I’ve got it wrong many, many times.  The most common culprits are scalebars (as in the picture above) and the tops of the sandbags that a specimen is resting on, obscuring the bottom of the centrum.  I know some people find it useful to have photos with scalebars in them: that’s fine; just don’t forget to also take some without the scalebars.

Use a plain background when possible.

Of course you don’t always have this luxury, but some collections have big white sheets of pleasantly rigid styrofoam that you have prop up behind your specimens to good effect — see the last photo in this post for an example.  Yes, you’re probably going to photoshop the background out later anyway, but it is much, much quicker and easier to remove a near-white more-or-less solid background than a busy one — especially if the background is similar in colour to the specimen, as for example when a brown bone has wood behind it.

But the good news is that all these problems can be ameliorated if you follow the last and most important rule in this section which is:

Take many shots and keep only the good ones

I remember reading once, long ago, that the single biggest factor in the difference of quality between a professional photographer’s work and an amateur’s is that the pro takes ten times as many shots and throws 90% of them away.  In these days of digital cameras with huge memory cards, we can all make like professionals now.  When Matt and I were at the Field Museum in Chicago, we took 168 photos of those Brachiosaurus dorsals alone.  Of those, maybe a dozen or so are really worth keeping.  But at least I have those dozen.

In general, I take every photograph twice.  As I’ve got better at taking the photos, I am increasingly finding that both come out well and it’s a toss-up which to keep, but maybe one time in ten or twenty, one of them just doesn’t come out right — something is wrong with the focus, or the camera shakes, or something — and that’s when I’m glad I have the spare.

Another terrible photo, this time with the flash washing out all the detail of the neural spine of Giraffatitan brancai lectotype HMN SII, 8th cervical, in left lateral view.



I have found that it is generally best to avoid using the camera’s flash unit: more often than not it just washes out all the detail, as in the Giraffatitan cervical above.  You’d never guess it from this photo, but the lateral faces of that spine are delicately and elaborately sculpted.  Having said that, using flash does sometimes seem to improve a photo — I’ve not been able to put together a mental model of when it does and doesn’t, so I will often take a photo (or pair) without flash and an otherwise identical one with, and see which works better.

On the other hand, my camera’s built-in flash is pretty lame.  Expensive flash units might do much better.

Other lights

I have had varying success in posing external light-sources to illuminate vertebrae.  The lights at the Oklahoma Museum of Natural History are excellent, for example, and allowed me to get stellar picture quality in some of my photos of the Hotel Mesa sauropod material.  [Note to self: we should show some of that material here some time.]  On the opposite extreme, the old angle-poise lamps in the sub-basement of the Natural History Museum, when they worked at all, and could be posed without falling over, seemed to do little more than cast a sickly yellowish pall over the specimen.  But things are better down there since pterosaurophile curator and part-time cephalopod Lorna Steel managed to persuade the department to spring for a few daylight lamps.  They fall apart distressingly easily, but do cast good diffuse light if you can persuade them to go into, and stay in, the position you want.

As with flash, it seems that the only thing to do is try photos with and without external lights, and with the lights in various different positions, and see what comes out best.

Giraffatitan brancai paralectotype HMN SI, cervical vertebra 6 in right anterolateral view. Not a bad photo -- click through to the full-sized version to appreciate the awesome.


If you’re not using flash or external lights, you have a problem, because most sauropod bones are kept in dimly lit basements with no natural light and low ambient light levels that make photography difficult.  If you use your camera in automatic mode (and I admit that I do), it will compensate by lengthening the exposure time, which means that camera-shake becomes a much bigger deal.  With flash, or in good daylight, the shutter will typically open for 1/250 or 1/125 of second; but in low light, your exposure can easily be as much as 1/4 second, and it’s pretty much impossible to keep a camera truly still for that long.

So what can you do?  Well, there are several levels of compensation.

Simply being aware of remaining still

When I have to hold the camera in my hands and I know it’s going to be a long exposure I find myself going into a sort of zen state — I become aware of my heartbeat and try to time the shutter release so that the camera doesn’t get moved by my pulse.  It’s error-prone, but at least being aware of it can help.

Brace against a door-frame or similar

Better, if you can do it, is to brace the camera against an immovable object such as a door frame or a specimen cabinet.  The photograph above was taken using what Matt and I came to call “The Wedel Method”: the camera was held in place on the shelf across the aisle from the specimen, but with the barrel rotated 180 degrees so that the LCD screen faced back into the aisle.  I stood between the camera and vertebra, slightly off to one side and facing away from the vertebra so I could use the screen.  In that position, I zoomed and panned to the the composition I wanted, then let the shelf keep the camera rock-steady as I released the shutter.  This only works with a camera such as a CoolPix 4500 that has a rotating barrel, but that is a useful feature for other reasons, too, and I recommend that you get a camera that has it if possible.  (For example, when you need to get a photo from directly above a specimen, you can often frame it by looking at the rotated screen, even if the specimen is in a cabinet can’t can’t be moved.)


Of course, much better than ad-hoc bracing like door-frames is a proper tripod, and I feel mortified that it took me about five years of specimen photography before I invested in a half-decent one.  I got a Hama Star 61 from Amazon, where you can currently get them at the absurdly low price of £7, and I am really happy with it: it it hits the sweet-spot between being too heavy to lug around comfortable and too light to stabilise the camera properly.  Listen: whatever you’re doing, stop it RIGHT NOW and go buy a tripod instead.  Not a little table-top one, a proper floor-standing one.  You’ll thank me.

Shutter delay

The other thing that can make a huge difference in avoiding camera shake is to arrange that the shutter is released a few seconds after you press the button — so that you eliminate the movement associated with the press itself.  On my camera, for some reason, you can only do this in macro mode (used for close-ups, also known as “flower mode”), but since the camera is happy to focus on large far-off objects in this mode, that’s not a problem.

The combination of tripod mounting and shutter delay means that you can get good exposure in almost any light.

The Archbishop in all its glory, with everything working right. The much-loved dorsals 8 and 9 in right lateral view. Click through to see the detail. Compare with the horrible photo of the same bones at the top of this article. Copyright the NHM since it's their material.


Get a camera with decent optics, and a tripod.  Compose your photos so that the element is fully in frame and unobscured, in orthogonal aspect, with a solid black or white background if possible.  Turn off the flash; use external lighting if it’s available and helpful.  Use shutter delay, and take several photos, keeping only the good ones. That’s what I’ve learned in six years of photographing sauropods, and I am a bit disappointed to find that it can be summarised in 58 words.

… And finally …

I was asked to pass this message on a while back, and I’m glad to finally do so:

From: Carol Brown<>

Hi Michael,

We just posted an article, “100 Best (Free) Science Documentaries Online” ( I thought I’d drop a quick line and let you know in case you thought it was something you’re audience would be interested in reading. Thanks


When I visited Dinosaur National Monument in October with Brian Engh and Yara Haridy, we spent a decent amount of time checking out DNM 28, a skull and associated bits of Camarasaurus. In particular, I got some shots of the axis (the second cervical vertebra behind the head), and it got me thinking about pneumaticity in this unusual element. Why I failed to get a full set of orthogonal shots is quite beyond my capacity, but we can roll with what I have. Before we go on, you might want to revisit Tutorial 36 to brush up on the general parts of the atlas-axis complex.

Here’s the axis in left lateral view (so, anterior to the left).

And a labeled version of the same. A few things to note:

  • One oddity of sauropod axes (and of axes of most critters) is that not only are the articular facets of the prezygapophyses not set forward of the neural arch, they’re set backward, well behind the forward point of the arch.
  • The dens epistrophei or odontoid process is sticking out immediately below the neural canal. This is the tongue of bone that articulated with the atlas (first cervical vertebra) in life.
  • Check out the prominent epipophysis above the postzygapophysis, which anchored the long dorsal neck muscles. (For more on epipophyses, see these posts, and especially this one.)
  • The diapophysis and parapophysis articulated with a cervical rib, which is not shown here. In fact, I don’t remember seeing it in the drawer that this vert came from. The atlantal and axial cervical ribs are small, apparently fused late in life if they fused at all, and are easily lost through taphonomic processes.
  • At least three pneumatic features are visible in this lateral view: the lateral fossa on the centrum, which is referred to as the “pleurocoel” in a lot of older literature; a ventral fossa that lies between the parapophysis and the midline ventral keel; and a fossa on the neural arch, behind the postzygodiapophyseal lamina. In the nomenclature of Wilson et al. (2011), this is the postzygocentrodiapophyseal fossa.

“Postzygocentrodiapophyseal fossa” is a mouthful, but I think it’s the only way to go. To be unambiguous, anatomical terminology needs to references specific landmarks, and the schemes proposed by Wilson (1999) for vertebral laminae and Wilson et al. (2011) for vertebral fossae are the bee’s knees in my book.

Nomenclatural issues aside, how do we know that these fossae were all pneumatic? Well, they’re invasive, there’s no other soft-tissue system that makes invasive fossae like that in archosaur vertebrae (although crocs sometimes have shallow fossae that are filled with cartilage or fat), and the same fossae sometimes have unambiguous pneumatic foramina in other vertebrae or in other sauropods.

Most of the features labeled above are also visible on the right side of the vertebra, although the ventral fossa is a little less well-defined in this view, and I can’t make out the prezyg facet. Admittedly, some of the uncertainty here is because of my dumb shadow falling across the vertebra. Specimen photography fail!

The paired ventral fossae are more prominent in this ventral view, on either side of the midline ventral keel (anterior is to the top).

And here’s a labeled version of the same ventral view.

Finally, here’s the posterior view. It’s apparent now that the neural spine is a proportionally huge slab of bone, like a broad, tilted shield between the postzygapophyses (which are also quite large for the size of this vertebra). The back side of the neural spine is deeply excavated by a complex fossa with several subfossae (kudos again to Jeff Wilson [1999] for that eminently useful term).

Here’s the same shot with some features of interest labeled. If I’ve read Wilson et al. (2011) correctly, the whole space on the back side of the neural spine and above the postzygs could be considered the spinopostzygapophyseal fossa, but here I’ve left the interspinous ligament scar (ILS) unshaded, on the expectation that the pneumatic diverticula that created that fossa were separated on the midline by the interspinous ligament. I might have drawn the ILS too conservatively, conceivably the whole space between the large deeply-shadowed subfossae was occupied by the interspinous ligament.

I’m particularly interested in those three paired subfossae, which for convenience I’m simply calling A, B, and C. Subfossa A may just be the leftover space between the spinopostzgyapophyseal laminae laterally and the interspinous ligament medially. I think subfossa B is invading the ramus of bone that goes to the epipophysis and postzygapophysis, but I didn’t think to check and see how far it goes (that might require CT anyway).

Subfossa C is the most intriguing. Together, those paired fossae form a couple of shallow pits, just on either side of the midline, and aimed straight forward. They can’t be centropostzygapophyseal fossae, which used to be called peduncular fossae, because they’re not in the peduncles on either side of the neural canal, they’re up above the lamina that connects the two postzygapophyses. Could they be ligament attachments? Maybe, but I’m skeptical for at least four reasons:

  1. Although interspinous ligament attachments often manifest as pits in the cervical vertebrae of birds, in sauropods they usually form rugosities or even spikes of bone that stick out, not inward. Furthermore, these pits are smooth, not rough like the interspinous ligament scars of birds.
  2. The interspinous ligament in tetrapods is typically a single, midline structure, and these pits are paired.
  3. Similar pits in front of the neural spine are present in some sauropod caudals, and they appear to be pneumatic (see Wedel 2009: p. 11 and figure 9).
  4. Pits at the base of the neural spine seem to be fairly uncommon in sauropod vertebrae. If they were attachment scars from the universally-present interspinous ligaments, we should expect them to be more prominent and more widespread.

But if these paired pits are not ligament scars, what are they? Why are they present, and why are they so distinct? Sometimes (often?) subfossae and accessory laminae look like the outcome of pneumatic diverticulum and bone reacting to each other (I almost wrote ‘playing together’), in what looks like a haphazard process of adaptation to local loading. But the symmetry of these pits argues against them being incidental or random. They don’t seem to be going anywhere, so maaaybe they are the first hoofbeats of the embossed laminae and “unfossae” that we see in the vertebrae of more derived sauropods (for which see this post), but again, their symmetry in size and placement isn’t really consistent with the “developmental program gone wild” appearance of “unfossae”. I really don’t know what to make of them, but if you have ideas, arguments, or observations to bring to bear, the comment field is open.

In summary, sauropod axes are more interesting than I thought, even in a derpasaurus like Cam. I’ll have to pay more attention to them going forward.



As noted in the last post, Matt and I are off to spend a week at the Carnegie Museum from 11th-15th March. We expect to see many, many fascinating specimens there: far more than we’ll be able to do proper work on in the five days we have. So our main goal is to exhaustively document the most important specimens that we see, so we can work on them later after we’ve got home. I think of this as the “harvesting” phase of research, with the grinding and baking to follow.

I was going to write a checklist for myself, to ensure that I cover all the bases and we don’t find ourselves in six months’ time looking at our records and saying “I can’t believe we forgot to do X for this specimen” — because, believe me, we have spent far too much of our lives doing this already. But then I realised I should share it with the world, in case it’s helpful to others, too.

So here’s what to do when dealing with, for example, an apatosaurine cervical like this one. Let me know in the comments if I forgot anything!

BYU 20178, cervical vertebra from an apatosaurine sauropod. ventral view, anterior to the left. Note that the scalebar is held at approximately half the height of the vertebra; and that the catalogue card is in view and legible, giving a record who collected the specimen, when, and where.

Sketch the specimen, even if (like me) you are a terrible artist. The process of sketching forces you to really look at it — at each part of it in turn — and often results in you noticing something you would otherwise have missed. It would be worth doing this even if you immediately threw the sketch away: but don’t do that, because you’re going to want to …

Measure the specimen, using a tape measure, digital calipers or both as appropriate. You want to get at least all the measurements that you’ll include in a formal description — total length, total height, width across zygapophyses, etc. — but it’s often useful to also get other, more obscure measurements, just to make sure you’ve got your head around the shape. For example, in the vertebra above, you might measure the diagonal distances from the anteriormost projection of each cervical rib to to opposite side’s posterolateralmost part of the centrum. You record measurements in a table in your notebook, but some measurements are hard to describe: so just write them straight onto your sketch. To keep things straight, it can be useful to do the sketch in one colour and the measurements in another; or the sketch in pencil and the measurements in pen.

Now we come to photography. You want a lot of different kinds of photo, so lets consider them separately.

Take photographs of the specimen with its specimen label, ideally from several different aspects. This will make it easy to remember later which specimen is which. In a typical museum visit — especially a reconnaisance visit like our upcoming Carnegie trip — you’re going to see a lot of different specimens, and when you revisit your photos in six months it’ll be hard to keep them all straight. Make it easy on yourself. Also: the specimen label often contains other  useful information such as the quarry where the specimen was found. Capture that. Get a good close-up photo of the label alone, to ensure all the text is captured cleanly.

Take photographs from the cardinal directions. To illustrate a specimen nicely in a descriptive paper, you will at minimum want photos from anterior, posterior, dorsal, ventral and left and right lateral aspects (or as many of these are possible to obtain: you can’t always turn big specimens). Since these are the photos you’re likely to use in a publication, take extra care with these. Set up a plain-coloured background when possible so it’s easier to crop out later. Set up the best lighting you can. Take each photo several times so you can keep the best one. Use a tripod if you have one. (For much more on this, see Tutorial 8 on how to photograph big bones.)

Take photographs with a scalebar. This will give you a way to sanity-check your measurements later. Think carefully about scalebar placement. If you put it on top of the specimen so it obscures part of the fossil, be sure that’s not your only photo from that aspect: you won’t want to be left without good images of the whole bone. A scalebar placed on top of the specimen will appear larger than the same scalebar placed on the floor or the bench next to the specimen, thanks to perspective, which means your measurements are more trustworthy than photos of the scalebar. If you can easily arrange for it to be raised to half the total height of the specimen, you’ll get a more honest reading.

Photograph individual features of the bone with some kind of note. The reason I say “with some kind of note” is that I have hundreds of close-up photos of bits of sauropod vertebra which I evidently took in the hope of highlighting some specific bit of morphology, but I have no idea what morphology. Get a scrap of paper and scribble something like “big nutrient foramen”, draw an arrow on it, and place the scrap on the bone so that the arrow points at the feature. Take a photo; then remove the paper and take another photo. The first one is your note to yourself; the second is the raw material for an illustration that you might prepare later, highlighting the relevant feature in a more elegant way.

Do a video walkaround with narration. For some reason, we didn’t start doing this until very recently, but it’s a great way to get a rough-and-ready reminder of important aspects of the specimen. You can just do this with a phone, moving it around the specimen, pointing to interesting bits and saying things about them. Here’s an example of Matt pointing out some features of the preserved cervical vertebrae of Suuwassea, and here he is again pointing out how pelican vertebrae are made of nothing.

Take a shedload of undifferentiated photos from every possible angle. Your goal here is that you’ll be able to use photogrammetry later to make a 3D model of the fossil. I admit to my shame that I’ve still never successfully done this — but thanks to the kindness of my good friend Heinrich Mallison, who is an expect in this area, I do have some fine models, including the Xenoposeidon model that was published as a supplementary file to my 2018 paper. Even if you don’t have access to someone as helpful as Heinrich, it’s worth getting these comprehensive photo-sets because photogrammetry software is likely to get progressively easier to use. Hopefully in a couple more years there will be nothing to it but loading a bunch of photos and pressing a button.

Up till here, we’ve been concentrating on gathering information about the specimen in a form that we’ll be able to return to later and use in comparisons and illustrations. But we can do more than that now we’re here with the physical specimen:

Look at the bone texture. Figure out how much of it is real, and how much is reconstructed — a particular problem with older specimens. Keep an eye out for rugosities for muscle and ligament attachments, smooth areas and pockets for pneumatic diverticula (or fat pads in boring mammal verts), and any odd growths that might be ossified soft tissues or pathological reactive bone growth. These kinds of things are often much easier to see in the actual specimens than in even the very best photographs.

Check for areas where the specimen is under-prepared. It’s very common for a neural canal to remain filled with matrix — and easy to spot, so in a sense not a problem. But how often is a pneumatic feature obscured because it’s still full of matrix? This is probably part of the reason that caudal pneumaticity so often goes unobserved, and it will very often obscure foramina within the neural canal. Similarly, I don’t know whether the huge club on the end of the right cervical rib of NHMUK PV R173b (formerly BMNH R173b) is pathological bone or a mineral concretion, because all I have to go from is my lame photos. I should have figured that out while I was with the actual specimen.

Discuss the specimen with a friend. I just can’t overstate how important this is. When Matt and I visit a collection together, we discover much, much more than twice as much as either of us would alone. Isaac Asimov is said to have observed “The most exciting phrase to hear in science, the one that heralds new discoveries, is not “Eureka!” (I found it!) but “That’s funny …””. Whether or not he ever actually said it (it’s not in any of his written works) it’s certainly true that the key moment in investigating a specimen is frequently when one person says “Hey, take a look at this”. Two minds can spark off each other in a way that a single mind can’t.

Last of all, it’s worth giving a little bit of thought to the possibility that you’ll one day be doing publicity for this specimen. So:

Get someone to take photos of you with the specimen. You’ll need them for press releases and media packs. I’ve only once in my life been in physical proximity with the Brontomerus specimen: during the three-day 2007 visit when I did much of the descriptive work for the paper. Idiotically, although I was there with three colleagues (Matt, Randy Irmis and Sarah Werning), I didn’t get anyone to take a photo of me with the material. So when we needed a photo for the publicity:

The Brontomerus mcintoshi holotype specimen OMNH 27761-27800, 61248 and 66429-66432 with the authors of the paper that described it. Back row (L to R): Mike Taylor, Matt Wedel, Rich Cifelli.

There was no good way to get it. I certainly wasn’t going to fly back out to the USA just to get a photo. So we got our Emmy award-winning special-effects-wizard friend Jarrod Davis to photoshop me into a photo that the museum had been able to take of Matt and Rich. (You can see the evidence here and here if you want to see how it was done. And, yes, before he could even start composing me in, Jarrod had to rescue a ludicrously under-exposed base image.)

Much better to avoid such nonsense. Get good photos of you with the specimens, like the one at the top of the Sauropocalypse post, and then if you ever need ’em you’ve got ’em.


Saurischian laminae and fossae v2 - Adam Marsh 2015

[Hi folks, Matt here. I’m just popping in to introduce this guest post by Adam Marsh (UT Austin page, LinkedIn, ResearchGate). Adam is a PhD student at UT Austin’s Jackson School of Geosciences, currently working for a semester as a Visiting Student Researcher at my old stomping ground, Berkeley’s UCMP.  Adam’s been working at Petrified Forest National Park in the summers and most of his research is on the Navajo Nation in Arizona. His major interest is in how we perceive extinctions in the fossil record. Specifically, he’s exploring the geochronology of the Glen Canyon Group to look at the biotic response to the end-Triassic mass extinction. He’s also working on an overhaul of the early saurischian dinosaurs of western North America – hence this post. It’s timely because I was just talking in the last post about putting together infographics to spread your ideas; here Adam’s very nice diagram serves as a quick guide and pointer to several papers by Jeff Wilson and colleagues. Many thanks to Sarah Werning for suggesting that Adam and I get acquainted over vertebrae. Update the next day: both the diagram above and the PDF linked below have been updated to fix a couple of typos. Also, there are now black and white versions – see below.]

– – – – – – – – – –

If you’re like me, you don’t count sheep when you fall asleep, you count laminae. These struts of bone and their affiliated fossae connect and span between major structural features on vertebral neural arches such as prezygapophyses, postzygapophyses, parapophyses, diapophyses, hyposphenes, hypantra, and the neural spine. Presumably, laminae bracket and fossae house outgrowths of pneumatic diverticula from the respiratory system, which has been covered extensively on this blog in sauropodomorph dinosaurs.

Talking about these complicated structures is cumbersome; they’ve been called buttresses, ridges, struts, etc. throughout descriptive skeletal literature. But what we call things is important, especially when we introduce laminae and other vertebral structures to the rigors of phylogenetic systematics, where homologous apomorphies reign supreme. In order to avoid arguing about whether one structure is called the potato or the tomato, Jeff Wilson and others introduced a strategy of naming vertebral laminae (Wilson, 1999) and the fossae (Wilson et al., 2011) that they surround using the same vertebral landmarks that most tetrapod anatomists agree upon (see the parade of –apophyses above). The process is very simple. Vertebral laminae are named for the two structures that they connect; the prezygodiapophyseal lamina (prdl) connects the prezygapophysis and the diapophysis, so each neural arch will have two prdls. Vertebral fossae are named for the two major laminae that constrain them; the prezygocentrodiapophyseal fossa (prcdf) opens anterolaterally and is delineated dorsally by the prezygodiapophyseal lamina and ventrally by the anterior centrodiapophyseal lamina. Again, each neural arch will have two prcdfs. Those of you who are black belt vertebral anatomists, to borrow a favorite phrase from my advisor, might be interested in serial variation and how these structures change up and down the vertebral column. Until I get my act together and publish some cool new saurischian data, I will refer you to Wilson (2012). [We’ve also touched on serial variation in laminae in this post and this one. – MJW]

Saurischian laminae and fossae v2 bw - Adam Marsh 2015

Same thing in black and white, with labels


You might have noticed that the names are a mouthful and take up their fair share of typed characters. In my research of early saurischian dinosaurs, I’ve run across quite a few of these laminae everywhere from herrerasaurids to sauropodomorphs to coelophysoids to Dilophosaurus. Even though I’ve drawn, photographed, and written about various laminae and fossae, I still need to remind myself of what goes where and what it’s called. Believe me, vertebral lamina nomenclature does not lend itself well to Dem Bones covers. As a result, I’ve put together a reference figure that might be useful for those of you who are dealing with this or even teaching it to students. At the very least, you can put it on the ceiling above your bed so that it’s the first thing you see when you open your eyes in the morning.

Four main vertebral laminae are present plesiomorphically in archosaurs: the anterior and posterior centrodiapophyseal laminae, the prezygodiapophyseal lamina, and the postzygodiapophyseal lamina. This means that the prezygocentrodiapophyseal, postzygocentrodiapophyseal, and centrodiapophyseal fossae are present, and sometimes the top of the transverse process is concave between the neural spine and the zygapophyses to form the spinodiapophyseal fossa. I know that a certain sister group of Sauropodomorpha can get disparaged around these parts, but the truth is that theropods build long necks, too, and sometimes in very different ways than sauropodomorphs. When you are writing about the various vertebral buttresses and chonoses, don’t get frustrated with the names, because Wilson and his colleagues have actually made it much easier for us to talk to one another about presumably homologous structures without needing an additional degree in civil engineering.

– – – – – – – – – –

Here’s the figure again in PDF form: Marsh, Adam 2015 saurischian laminae and fossae diagram v2

And in black and white for those who prefer it that way: Marsh, Adam 2015 saurischian laminae and fossae diagram v2 bw


Last October, Mike posted a tutorial on how to choose a paper title, then followed it up by evaluating the titles of his own papers. He invited me to do the same for my papers. I waited a few days to allow myself to forget Mike’s comments on our joint papers – not too hard during my fall anatomy teaching – and then wrote down my thoughts.

And then did nothing with them for three and a half months.

The other day I rediscovered that draft and thought, hey, I don’t remember anything I wrote back then, I should redo the experiment and see if my evaluations will be consistent. And this time without looking at Mike’s post at all, so the risk of contamination would be even lower.

BUT FIRST I thought I should write down what I admire in paper titles, so I could see whether my titles actually lived up to my ideals. So now we can compare:

  • what I say I like in paper titles;
  • what I actually titled my papers;
  • what I had to say about my titles last October;
  • what I have to say about them now;
  • and, for some of my papers, what Mike had to say about them.

What I Admire In Paper Titles

Brevity. I first became consciously aware of the value of concise titles when I read Knut Schmidt-Nielsen’s autobiography, The Camel’s Nose, in 2004 or 2005. (Short-short review: most of the book is a narrative about scientific questions and it’s great, the self-congratulatory chapters near the end are much less interesting. Totally worth reading, especially since used copies can be had for next to nothing.) Schmidt-Nielsen said he always preferred short, simple titles. Short titles are usually punchy and hard to misunderstand. And I like titles that people can remember, and a short title is easier to recall than a long one.

Impact. In short, maximum information transfer using the minimum number of words. This is a separate point from sheer brevity; a paper can have a short title that doesn’t actually tell you very much. But brevity helps, because it’s difficult to compose a long title that really hits hard. Whatever impact a title might have, it will be diluted by every extraneous word.

Full sentences as titles. This is taking the information-transfer aspect of the last admirable quality to its logical extreme, although often at the expense of brevity. I was heavily influenced here by two things that happened while I was at Berkeley. First, I taught for a year in an NSF GK-12 program, where graduate students went out into local elementary, middle, and high schools and taught biology enrichment classes. One thing that was drilled into us during that experience is that we were teaching concepts, which ideally would be expressed as complete sentences. Also about that same time I read James Valentine’s book On the Origin of Phyla. The table of contents of that book is several pages long, because every chapter title, heading, and subheading is a complete sentence. This has a lovely effect: once you’ve read the table of contents of the book or any of its parts, you’ve gotten the TL;DR version of the argument. Sort of like a distributed abstract. I’d like to do that more.

How Did I Do?

Time to see if my actions match my words. Full bibliographic details and PDFs are available on my publications page. I stuck with Mike’s red-blue-green color scheme for the verdicts. My October 2014 and February 2015 thoughts are labeled. For joint papers with Mike, I’ve copied his assessment in as well. Any comments in brackets are my editorializing now, comparing what I said in October to what I said a few days ago before I’d looked back at my old comments or Mike’s.

* * * * * * * * * * * *

Sauroposeidon proteles, a new sauropod from the Early Cretaceous of Oklahoma. (11 words)

Oct 2014: Like it. Short, to the point, includes the taxon name.
Feb 2015: Good, gets the job done with a minimum of fuss

Osteology, paleobiology, and relationships of the sauropod dinosaur Sauroposeidon. (9 words)

Oct 2014: This title was inspired by the papers from the early 20th century
Feb 2015: It gets the job done, I suppose. I can’t help but wonder if there might have been a more elegant solution. Part of my unease is that this title is an example of the same attitude that produced the next monstrosity.

Osteological correlates of cervical musculature in Aves and Sauropoda (Dinosauria: Saurischia), with comments on the cervical ribs of Apatosaurus. (19 words)

Oct 2014: Ugh. It gets the job done, I suppose, but it’s waaaay long and just kind of ugly.
Feb 2015: Ugh. Waaay too wordy. I had a (fortunately brief) fascination with long titles, and especially the phrase, “with comments on”. Now I would cut it down to “Bony correlates of neck muscles in birds and sauropod dinosaurs” (10 words)

Vertebral pneumaticity, air sacs, and the physiology of sauropod dinosaurs. (10 words)

Oct 2014: Like it. Would be better made into a sentence, like, “Vertebral pneumaticity is evidence for air sacs in sauropod dinosaurs.”
Feb 2015: Fairly clean. Does what it says on the tin. I’m having a hard time seeing how it could be turned into a sentence and still convey so much of what the paper is about in so few words.

[Heh. As we will see again later on, I was evidently smarter last fall than I am now.]

The evolution of vertebral pneumaticity in sauropod dinosaurs. (8 words)


Oct 2014: Like it. It couldn’t really be any shorter without losing crucial information. Happy to have a decent title on my second-most-cited paper!
Feb 2015: Short, clean, probably my best title ever.

First occurrence of Brachiosaurus (Dinosauria: Sauropoda) from the Upper Jurassic Morrison Formation of Oklahoma. (14 words)

Oct 2014: Yep. once you’ve read the title, you barely need to read the paper. Even better would have been, “A metacarpal of Brachiosaurus from the Upper Jurassic Morrison Formation of Oklahoma.” (12 words)
Feb 2015: Does what it says, but like my other PaleoBios pub, it’s a long title for a short paper. Now I would title it, “First record of the sauropod dinosaur Brachiosaurus from Oklahoma” (9 words)

[my October title was better!]

Postcranial skeletal pneumaticity in sauropods and its implications for mass estimates. (11 words)

Oct 2014: It’s not elegant but it gets the job done. I wanted that paper to be one-stop shopping for sauropod PSP, but of course the real payoff there is the ASP/mass-estimate stuff, so I’m happy to have punched that up in the title.
Feb 2015: Good enough. I like it. It’s a little long–I could reasonably have just titled this, “Postcranial skeletal pneumaticity in sauropods”, but I wanted to draw attention to the implications for mass estimates.

Sauroposeidon: Oklahoma’s native giant (4 words)

Feb 2015: Nice and short. Not terribly informative, but since this was a narrative about the discovery and description of Sauroposeidon aimed mostly at an Oklahoma audience, it’s not obvious how it could be improved.
[Note sure how missed this one last October, but I did.]

Origin of postcranial skeletal pneumaticity in dinosaurs. (7 words)

Oct 2014: About all I would change now would be to add the word “early” at the beginning of the title.
Feb 2015: Great. Could not be shortened further without losing information.

What pneumaticity tells us about ‘prosauropods’, and vice versa. (9 words)


Oct 2014: Love this title. I used it for the abstract of the SVP talk that the paper was derived from, too.
Feb 2015: Kind of a gimmick title, but it’s accurate–the SVP abstract this paper was based on was built around a bullet list. And it’s still nice and short.

Evidence for bird-like air sacs in saurischian dinosaurs. (9 words)


Oct 2014: Along with Wedel (2003b) and Wedel (2006), this has a short (7-9 words apiece) title that tells you what’s in the paper, simply and directly. For once, I’m glad I didn’t turn it into a sentence. I think a declarative statement like “Saurischian dinosaurs had air sacs like those of birds” would have been less informative and come off as advertising. I wanted this paper to do what the title said: run down the evidence for air sacs in saurischians.
Feb 2015: I like it and wouldn’t change it. The “evidence for” part is key – I didn’t want to write a paper primarily about the air sacs themselves. Instead I wanted to lay out the evidence explaining why we think sauropods had air sacs.

Head and neck posture in sauropod dinosaurs inferred from extant animals. (8 words)

Oct 2014: It’s not horrible but it would be better as a declarative statement like, “Sauropod dinosaurs held their necks and heads elevated like most other tetrapods.” (12 words)
Feb 2015: Good. Reads almost telegraphically brief as it is. Does what it says on the tin.


[October Matt wins again!]

A new sauropod dinosaur from the Lower Cretaceous Cedar Mountain Formation, Utah, USA. (13 words)

Oct 2014: Two things about this one. First, I wish we’d been able to include the taxon name in the title, as we were allowed to do back in the day for Sauroposeidon. Second, I know some people whinge about us using the CMF in the title and in the paper instead of the Burro Canyon Fm, which is what the CMF is technically called east of the Colorado River. But srsly, how many people search for Burro Canyon Fm versus CMF? All of the relevant faunal comparisons are to be made with the CMF, so I don’t feel the least bit bad about this.
Feb 2015: Fine. About as short as it could be and still be informative.


The long necks of sauropods did not evolve primarily through sexual selection. Journal of Zoology. (12 words)

Oct 2014: Perfect. The abstract and the paper expand on the title, but if all you read is the title, you know what we found. That’s a worthy goal.
Feb 2015: My first sentence title. Every word does work, so even though this is one of my longer titles, I like it. The length relative to my other titles is not a knock against this one; rather, it emphasizes how well I did at keeping my early titles short and to the point (with a couple of regrettable exceptions as noted above).


The early evolution of postcranial skeletal pneumaticity in sauropodomorph dinosaurs. (10 words)

Oct 2014: Not bad. I wonder if something like, “Widespread vertebral fossae show that pulmonary pneumaticity evolved early in sauropodomorphs” might be better. It’s hard, though, to put so many long, polysyllabic words in a title that doesn’t sound like a train wreck. At a minimum, this paper does what it says on the tin.
Feb 2015: Short and to the point. Another one that couldn’t be any shorter without losing valuable information.

A monument of inefficiency: the presumed course of the recurrent laryngeal nerve in sauropod dinosaurs. (15 words)

Objectively: BAD to OK
Subjectively: GOOD to FREAKIN’ AWESOME
Oct 2014: I readily admit that I could have fashioned a more informative title, but I dearly love this one. It’s derived from a TV commercial for cheeseburgers (true story), and it warms my heart every time I read it.
Feb 2015: This is definitely a gimmick title that is longer than it has to be (it would be a concise 11 words without the unnecessary intro clause) BUT I love it and I’d do it exactly the same if I could do it again. So there!

Why sauropods had long necks; and why giraffes have short necks. (11 words)

Oct 2014: This is one of those ‘draw the reader in’ titles. I like it.
Feb 2015: We both liked the even shorter, “Why giraffes have short necks” but we really felt that a paper about sauropod necks needed sauropod necks in the title. I feel about this one like I feel about my 2007 prosauropod paper: it’s a gimmick title, but it’s short, so no harm done.


Neural spine bifurcation in sauropod dinosaurs of the Morrison Formation: ontogenetic and phylogenetic implications. (14 words)

Oct 2014: Blah. It’s okay, not great. Maybe better as, “No evidence for increasing neural spine bifurcation through ontogeny in diplodocid sauropods of the Morrison Formation”, or something along those lines.
Feb 2015: This one is long but I think the length is necessary. It’s also kinda boring, but it was addressing a fairly dry point. I think any attempt to shorten it or sexy it up would come off as gratuitous.

Mike: WEAK

The effect of intervertebral cartilage on neutral posture and range of motion in the necks of sauropod dinosaurs. (18 words)

Oct 2014: Probably better along the lines of, “Intervertebral spacing suggests a high neutral posture and broad range of motion in the necks of sauropod dinosaurs” or something like that.
Feb 2015: My second-longest title ever! Looking at it now, I think we could have titled it, “Effects of intervertebral cartilage on neck posture and range of motion in sauropod dinosaurs” and gotten it down to 14 words, but the word ‘neutral’ is doing real work in the original so maybe that’s a bust.

Mike: UGH, rubbish.

[October Matt is up by three points at least]

Caudal pneumaticity and pneumatic hiatuses in the sauropod dinosaurs Giraffatitan and Apatosaurus. (12 words)

Oct 2014: Along the same lines as the previous: “Caudal pneumaticity and pneumatic hiatuses show that pulmonary diverticula in the tails of sauropod dinosaurs were pervasive and complex” or something.
Feb 2015: Good. Long only by comparison with some of my earlier titles. Does what it says.


The neck of Barosaurus was not only longer but also wider than those of Diplodocus and other diplodocines. (18 words)

Feb 2015: My second sentence-as-title, and another entry in the run of mostly long titles from 2012 onward. I like how precise it is, despite the length.

Mike: GOOD

A ceratopsian dinosaur from the Lower Cretaceous of Western North America, and the biogeography of Neoceratopsia. (16 words)

Feb 2015: I had no say in this one (by choice, I’m sure Andy et al. would have listened if I had had any suggestions about the title, but I didn’t). If I could rewrite it, I’d probably make it even longer by adding in the word ‘new’ between A and ceratopsian

Haplocanthosaurus (Saurischia: Sauropoda) from the lower Morrison Formation (Upper Jurassic) near Snowmass, Colorado. (13 words)

Feb 2015: Feels a lot longer than its 13 words, mostly because so many of the words are polysyllabic. Normally I like pulling the words in parentheses out, but in this case I can’t see that doing that would actually improve the title. Sometimes descriptive papers need plain titles. It’s okay.

* * * * * * * * * * * *


First, Mike graded harder than I did. In fact, I only rated one of my titles as BAD, which seems a bit feeble. I think we were using different criteria. If a title was boring but serviceable, I gave it an OK, whereas Mike tended to flag any suboptimal title as RUBBISH. But I didn’t remember that about his post, and I deliberately avoided looking at it until I’d made my evaluations.

Second, except for the two PaleoBios papers, all of the titles from the first half of my career (2000-2007) are 12 words or fewer, including a substantial bundle from before I’d read either The Camel’s Nose or Strunk & White. I’m sure that being a Cifelli student and then a Padian student had something to do with that; Rich and Kevin made me into the word choice and grammar pedant that I am today (my rhetorical excrescences on this site are my fault, not theirs).

Third, much to my surprise and consternation, my titles have gotten longer over time, not shorter. Partly that’s because my little corner of the science ecosystem is getting increasingly subdivided, so it’s hard for me to write a paper now with a title as broad as, “The evolution of vertebral pneumaticity in sauropod dinosaurs.” (Possibly a prod to keep seeking out new, more open horizons?) And I suppose there is some tension between brevity, informativeness, and precision. For example, saying in the title of a descriptive paper than a specimen is “from the Upper Jurassic Morrison Formation of [Location], [State or Country]” adds 11 words, but the title really does need those words. That could be a segue into a whole other discussion about descriptive versus analytical work, but that will be a topic for another time.

Ultimately, this has been a fun exercise and it’s made me more aware of how I title my papers. This is useful because I have some manuscripts in the works that deal with really detailed anatomy, and I need to figure out how to give them titles that are precise and informative but still punchy. It’s not easy.

Parting thought: after I posted the slides from my photography and illustration talk, Mike and I talked about posting some of our figures and dissecting them to see how they could be improved (it’s axiomatic that almost all figures could be improved in one way or another). We should really get started on that.

Illustration talk slide 51

Here’s a working version of that link.

Illustration talk slide 52

Illustration talk slide 53

Working link.

Illustration talk slide 54

Illustration talk slide 55

Illustration talk slide 56

Illustration talk slide 57

Working links:

The rest of this series.


  • Powell, Jaime E.  2003.  Revision of South American Titanosaurid dinosaurs: palaeobiological, palaeobiogeographical and phylogenetic aspects.  Records of the Queen Victoria Museum 111: 1-94.

Illustration talk slide 32

Illustration talk slide 33

Illustration talk slide 34

The links in the first slide:

Mike’s post on desaturating the background in specimen photos is here, and previous posts in this series are here.

Illustration talk slide 19

Illustration talk slide 20

Illustration talk slide 21

Illustration talk slide 22

This whole section, including the title, is mostly swiped from Mike’s Tutorial 17.

Other posts in this series are here.

Papers referenced in these slides:

Illustration talk slide 9

Illustration talk slide 10

Illustration talk slide 11

Illustration talk slide 12

That link in the first slide will take you here, and the rest of the series is here.

OMNH baby Apatosaurus

I was at the Oklahoma Museum of Natural History in March to look at their Apatosaurus material, so I got to see the newly-mounted baby apatosaur in the “Clash of the Titans” exhibit (more photos of that exhibit in this post). How much of this is real (i.e., cast from real bones, rather than sculpted)? Most of the vertebral centra, a few of the neural arches, some of the limb girdle bones, and most of the long bones of the limbs. All of the missing elements–skull, neural arches, ribs, appendicular bits–were sculpted by the OMNH head preparator, Kyle Davies. Kyle is one of those frighteningly talented people who, if they don’t have what they need, will just freaking build it from scratch. Over the years he has helped me out a LOT with the OMNH sauropod material–including building a clamshell storage jacket for the referred scapula of Brontomerus so we could photograph it from the lateral side–so it’s about time I gave him some props.

Atlas-axis model with Kyle

Case in point: this sweet atlas-axis complex that Kyle sculpted for the juvenile Apatosaurus mount.

Atlas-axis model by Kyle Davies

Most fish, amphibians, and other non-amniote tetrapods only have a single specialized vertebra for attaching to the skull. But amniotes have two: a ring- or doughnut-shaped first cervical vertebra (the atlas) that articulates with the occipital condyle(s) of the skull, and a second cervical vertebra (the axis) that articulates with the atlas and sometimes with the skull as well. Mammals have paired occipital condyles on the backs or bottoms of our skulls, so our skulls rock up and down on the atlas (nodding “yes” motion), and our skull+atlas rotates around a peg of bone on the axis called the odontoid process or dens epistrophei (shaking head “no” motion). As shown in the photos and diagrams below, the dens of the axis is actually part of the atlas that fuses to the second vertebra instead of the first. Also, reptiles, including dinosaurs and birds, tend to have a single ball-shaped occipital condyle that fits into the round socket formed by the atlas, so their “yes” and “no” motions are less segregated by location.

Anyway, the whole shebang is often referred to as the atlas-axis complex, and that’s the reconstructed setup for a baby Apatosaurus in the photo above.  In addition to making a dull-colored one for the mount, Kyle made this festive version for the vert paleo teaching collection. Why so polychromatic?

Atlas-axis model key

Because in fact he built two: the fully assembled one two photos above, and a completely disassembled one, some of which is shown in this photo (I had to move the bigger bits out of the tray so they wouldn’t block the key card at the back). I originally composed this post as a tutorial. But frankly, since Kyle did all of the heavy lifting of (a) making the thing in the first place, (2) making a color-coded key to it, and (d) giving me permission to post these photos, it would be redundant to walk through every element. So think of this as a self-study rather than a tutorial.

UPDATE in December, 2019: oh heck with it, I’m very belatedly renaming this a tutorial, so I can tag on a follow-up post as Tutorial 36b and curate this where it belongs, on our Tutorials page. The URL will stay the same, like a digital fossil.

Atlas-axis model by Kyle Davies - labeled

Oh, all right, here’s a labeled version. Note that normally in an adult animal the single piece of bone called the atlas would consist of the paired atlas neural arches (na1) and single atlas intercentrum (ic1), and would probably have a pair of fused cervical ribs (r1). Everything else would be fused together to form the axis, including the atlas pleurocentrum (c1), which forms the odontoid process or dens epistrophei (etymologically the “tooth” of the axis).

Romer 1956 fig 119 atlas-axis complex

Here’s the complete Romer (1956) figure from the key card, with a mammalian atlas-axis complex  for comparison. Incidentally, the entire book this is drawn from, Osteology of the Reptiles, is freely available online.

Apatosaurus axis-atlas complex Gilmore 1936 figs 5 and 6

And here’s the complete Gilmore (1936) figure. Sorry for the craptastic scan–amazingly, this one is NOT freely available online as far as I can tell, and Mike and I have been trying to get good scans of the plates for years. Getting back on topic, single-headed atlantal cervical ribs have been found in several sauropods, especially Camarasaurus where several examples are known, so they were probably a regular feature, even though they aren’t always preserved.

Also, as noted in this post, it is odd that in this specimen of Apatosaurus the cervical ribs had not fused to the first two vertebrae, even though they normally do, and despite the fact that the vertebrae had fused to each other, even though they normally don’t. Further demonstration, if any were needed, that sauropod skeletal fusions were wacky.

Varanops atlas-axis complex Campione and Reisz 2011 fig 2C3

For comparison to the above images, here is the atlas-axis complex in the synapsid Varanops, from Campione and Reisz (2011: fig. 2C).

Those proatlas thingies are present in some sauropods, but that’s about all I know about them, so I’ll say no more for now.

There is a good overview of the atlas-axis complex with lots of photos of vertebrae of extant animals on this page.

Previous SV-POW! posts dealing with atlantes and axes (that’s right) include: