Foster and Wedel 2014 fig 3 - dorsals

Fig. 3. MWC 8028, Haplocanthosaurus dorsal vertebrae. A. Lateral view of dorsal centrum with bottom edge of lateral pneumatic fossa preserved. B. Dorsal view of same centrum as in A, showing the median septum between the paired lateral fossae. C. Lateral view of dorsal centrum with smaller segment of the lateral pneumatic fossa margin preserved. D. Dorsal view of same centrum as in C, again showing the median septum and paired lateral fossae. E. Lateral view of dorsal centrum with partial pleurocoel preserved. F. Cross-sectional (posterior) view of same dorsal as in E. G. Dorsal neural spines in lateral (top) and anterior or posterior (center, bottom) views. Scale bars = 10 cm.

Right on the heels of Aquilops last week, my paper with John Foster on the new specimen of Haplocanthosaurus from Snowmass, Colorado, was just published in Volumina Jurassica. I’ll have more to say about it later, but right now I am up against a deadline on a big project and I need to go work on that. I’m only popping up here to note two quick things.

First, if you’re not familiar with Volumina Jurassica – and I wasn’t, before this project – it’s a free-to-access* journal that publishes papers on all aspects of the Jurassic. The current issue is specifically dedicated to the Jurassic formations of the American West. There’s a lot of interesting stuff in there, but of special interest to SV-POW! readers will be the paper by Cary Woodruff and John Foster on the legendary and possibly apocryphal Amphicoelias fragillimus.

* But not truly open access since the journal claims to retain exclusive rights to distribute the papers. That seems like a quaint affectation now that the internet is here, but whatever – at least they let anyone download the PDF for free, which is primarily what I care about.

Foster and Wedel 2014 fig 4 - sacrum

Fig. 4. Sacra of Haplocanthosaurus.  A. MWC 8028, sacrum in right lateral view. B. MWC 8028, close-up of S4 and S5 centra highlighting pneumatic fossae. C. MWC 8028 with divisions between the vertebrae overlaid. D. CM 879, sacrum in right lateral view with divisions between the vertebrae overlaid. E. CM 572 in right lateral view, after Hatcher (1903c: plate 4). B–E are not shown at the same scale, scale bar for A = 20 cm. Note that the neural arches in CM 572 were restored during preparation, and the sacral neural spines as shown here are probably lower than they would have been in life.

Second, the figure resolution in the PDF of the Haplocanthosaurus paper is not stellar, so as is the case with almost all of our papers, the full-color, high-resolution figures are available at the paper’s page on the sidebar.

Gotta run.

For our previous posts on Haplocanthosaurus, go here; for those on Amphicoelias, including Mike’s very popular, “How big was Amphicoelias fragillimus? I mean, really?”, go here.


No time for anything new, so here’s a post built from parts of other, older posts.

The fourth sacral centrum of Haplocanthosaurus CM 879, in left and right lateral view. This is part of the original color version of Wedel (2009: figure 8), from this page. (Yes, I know I need to get around to posting the full-color versions of those figures. It’s on my To Do list.)

Note the big invasive fossa on the right side of the centrum. The left side is waisted (narrower at the middle than the ends) like most vertebrae of most animals, but has no distinct fossa on lateral face of the centrum. What’s up with that? Here’s an explanation from an old post (about another sauropod) that still fits:

Now, this asymmetry is also weird, but it’s expected weirdness. Pneumaticity seems to just be inherently variable, whether we’re talking about human sinuses or the facial air sacs of whales or the vertebrae of chickens. It appears that the form of pneumatic features is entirely determined by local tissue interactions, with little or no genetic control of the specific form. Think of it this way: genes prescribe certain developmental events, and those events bring tissues into contact–such as pneumatic epithelium and bone. The morphology of the bone arises out of that interaction, and each interaction of bone and pneumatic epithelium has the potential to produce something new. In this case, the diverticula on the left side of the vertebral column come from the lungs or air sacs on the left, and those on the right side come from the lungs or airs sacs on the right, so it’s really two sets of diverticula contacting the bone independently. The wonder, then, is not that pneumatic bones are so variable, but that we see any regularities at all.

Matt and I have been looking in more detail at indications of maturity in sauropod skeletons, as we prepare the submission of the paper arising from our response to Woodruff and Fowler (2012) [part 1, part 2, part 3, part 4, part 5, part 6].  Here is an oddity.

Sacra of Haplocanthosaurus. Top, H. utterbacki holotype CM 879 in right lateral view, from Hatcher (1903:fig. 15). Bottom, H. priscus holotype CM 572 in left lateral view (reversed), from Hatcher (1903:pl. IV, part 3). To the same scale.

H. priscus is the type species of Haplocanthosaurus; H. utterbacki is the second species, named by Hatcher in the 1903 monograph that described the original material in detail.  As previously noted, the type species is based on adult material, and the referred specimen on subadult material.  This is shown by their different stages of neurocentral fusion, and corroborated by the size of the specimens as indicated in the composite illustration above.

There is a lot of fusion going on in the sacra of dinosaurs:

  1. sacral neural arches fused to their centra
  2. consecutive sacral centra fused together
  3. consecutive sacral neural spines fused together
  4. sacral lateral processes fused to ilia

As we would expect, the less mature of the two Haplocanthosaurus individuals is less fused in most respects: none of the centra were fused either to each other or their respective neural arches, and the ilium was not fused to any of the lateral processes, whereas in the adult all neural arches are fused to their centra, the five sacral centra are all fused together, and the ilium is fused to the lateral processes.

How strange, then, that the consecutive neural spines are more fused in the juvenile!  Not only are spines 1, 2 and 3 fused along their entire dorsolateral length, as in the adult, but spine 4 is similarly fused.  And more: the neurapophysis of spine 5 is fused to that of 4, even though the spines are not fused more ventrally.

What does this mean?  Hatcher (1903:27-28) took it as indicative of species-level separation.  After briefly noting that the posterior dorsal centra of H. utterbacki are more opisthocoelous than those of H. priscus, and speculating that the adult of the referred species was probably larger than that of the type, he continued:

But the most distinctive character is to be found in the sacrum which, in the present species, has the five neural spines normally coössified.  The first four are cocoössified throughout their entire length, forming a long bony plate.  The union between the fourth and fifth is limited to the extremities while medially [sic, presumably meaning half way up the spines] they are separated by an elongated foramen.  In H. priscus only the spines of the three anterior sacrals are coössified, those of the first and second [sic, presumably intending fourth and fifth] sacrals remaining free.  This difference exists notwithstanding that the type of the present species was scarcely adult, the sacral centra being neither coössified with one another nor with their neural arches. By some this character might be considered as of generic importance although I prefer to consider it as of only specific value since in all other parts of the skeleton preserved, there are no distinguishing characters which could be considered as of generic value.

At present, however, the synonymy of H. utterbacki with the type species, proposed by McIntosh and Williams (1988:22), seems to be universally accepted.  If they truly belong to the same taxon then the only realistic possibility is that we are seeing individual variation in the timing of fusion.  That certainly seems to have been the opinion of McIntosh and Williams (1988:14), writing about the sacrum of their own specimen, the H. delfsi holotype CMNH 10380:

As in CM 572 the short to moderately long spines of sacrals one through three are firmly united throughout, and those of sacrals four and five are firmly united to midheight. In CM 572 spines four and five are free, but this is probably an individual character because in the even younger CM 879 all five spines are united.

All of which means: we need to be really careful when drawing conclusions about taxonomy or ontogeny from individual observations of skeletal fusion.

Bonus Pneumaticity Observation: In the image at top, you’ll see that the centrum of sacral 4 in CM 879 has a couple of pneumatic fossae. For more than you probably wanted to know about those specific holes in that specific bone, see this post and the linked paper.


  • Hatcher, J.B. 1901. Diplodocus (Marsh): its osteology, taxonomy, and probable habits, with a restoration of the skeleton. Memoirs of the Carnegie Museum 1:1-63.
  • McIntosh, J.S., and Williams, M. E. 1988. A new species of sauropod dinosaur, Haplocanthosaurus delfsi sp. nov., form the Upper Jurassic Morrison Fm. of Colorado. Kirtlandia 43:3-26.
  • Woodruff, D.C, and Fowler, D.W. 2012. Ontogenetic influence on neural spine bifurcation in Diplodocoidea (Dinosauria: Sauropoda): a critical phylogenetic character. Journal of Morphology, online ahead of print.


Last time around, Matt walked through a lot of the detailed cervical morphology of Suuwassea and known diplodocids to show that, contra the suggestion of Woodruff and Fowler (2012), Suuwassea is distinct and can’t be explained away as an ontogenomorph of a previously known genus.

Although Suuwassea is singled out for special treatment in this paper, other genera do not escape unscathed.  From the Conclusions section on page 9:

Just as particularly large diplodocid specimens (e.g., Seismosaurus; Gillette, 1991) have been more recently recognized as large and potentially older individuals of already recognized taxa (Diplodocus; Lucas et al., 2006; Lovelace et al., 2007), taxa defined on small specimens (such as Suuwassea, but also potentially Barosaurus, Haplocanthosaurus, and ‘‘Brontodiplodocus’’), might represent immature forms of Diplodocus or Apatosaurus.

I have to admit I more or less fell out of my chair when I saw the suggestion that poor old Haplocanthosaurus might be Diplodocus or Apatosaurus.  I think this idea comes from a misstatement in the very first sentence of the abstract:

Within Diplodocoidea (Dinosauria: Sauropoda), phylogenetic position of the three subclades Rebbachisauridae, Dicraeosauridae, and Diplodocidae is strongly influenced by a relatively small number of characters.

As a statement of fact, this is simply the opposite of the truth: in all the major phylogenetic analyses, the arrangement of subclades with Diplodocoidea is the most stable part of the tree, supported by more characters than all the other clades.

For example, in the analysis of Upchurch et al. (2004) in The Dinosauria II, fig. 13.18 shows that the nodes with the highest bootstrap percentages are Diplodocinae (96%), Dicraeosauridae (95%) and Diplodocidae (93%).

Or consider the analysis of Wilson (2002).  While it’s getting on a bit, it still scores highly by being the most explicit published sauropod analysis, with comprehensive lists of apomorphies.  Table 12 lists the decay indexes for the 24 nodes in the strict consensus tree.  Apart from the three very basal nodes separating sauropods from their outgroups, the two highest-scoring clades are Diplodocidae and Diplodocinae (DI=7), followed by four clades all with DI=5 of which two are Dicraeosauridae and Flagellicaudata (which Wilson just called “Dicraeosauridae + Diplodocidae” as it had not yet been named).  (It’s well worth reading Wilson’s Appendix 3 to see the synapomorphies supporting these nodes in the MPTs: he lists 14 separating Diplodocimorpha from the node it shares with Haplocanthosaurus, 18 separating Flagellicaudata from the node it shares with Rebbachisauridae, 16 separating Diplodocidae from the node it shares with Dicareosauridae, and seven separating Diplodocinae from the node it shares with Apatosaurus).*

* Why are the lists of apomorphies longer than the decay indexes?  Because they list the apomorphies as they occur in the specific topology of the consensus tree.  Nodes within that tree can be made to collapse without wiping out all the apomorphies by rejuggling other parts of the tree to move character-state transitions around.  So although (for example) 26 characters separate Flagellicaudata from Rebbachisauridae (18 + 8 synapomorphies respectively) you can rejuggle the whole tree to break the monophyly of Flagellicaudata while making the entire tree only five steps longer.

Anyway, for whatever reason, Woodruff and Fowler felt that the stability of the diplodocoid clades was in question, and this presumably influenced their hypothesis that Haplocanthosaurus could be easily moved down into one of the diplodocid genera.

Next time we’ll be considering the implications for the tree.  But today, let’s take a moment to do this the old-fashioned way, by looking at …



Hatcher (1903), ever helpful, included a comparative plate in his monograph which should help us to evaluate the idea that Haplo is a known diplodocid:

Pelves of diplodocids and Haplocanthosaurus. 1. Pelvis of Brontosaurus excelsus (No. 568); 2. Pelvis of Diplodocus carnegii (No. 94); 3. Pelvis of Haplocanthosaurus priscus (No. 572).  All seen from left side.  1, 2, 3, 4, 5 indicate neural spines of respective sacral vertebra.  Presumably to scale.  Direct from Hatcher (1903:plate IV).

Based on this, the pelvis of Haplocanthosaurus differs from those of the diplodocids in having a proportionally lower ilium, in the absence of the laterally facing rugosity on the posterodorsal margin of the ilium, in the very small distal expansion of the pubis and in the almost non-existent distal expansion of the ischium.  These are all characters of the limb-girdle elements, which do not change greatly through ontogeny in sauropods.

But the evidence from the sacral vertebrae is just as significant: the neural spines in the sacral area are less than half as tall as in the diplodocids — and this in an animal whose dorsal neural spines are conspicuously tall.  The spines are also more anteroposteriorly elongate and plate-like.  What’s more, sacral spines 1, 2 and 3 have fused into a single plate in Haplocanthosaurus, while the spine of S1 remains well separated from 2 and 3 in the diplodocids.  So the ontogenetic hypothesis would have to say that the spine of S1 unfuses through ontogeny.  Which is not something I’ve heard of happening in any sauropod, or indeed any animal.

So the pelvis and sacrum seem distinct.  But Woodruff and Fowler’s (2012) notion of ontogenetic synonymy is built on the idea that the differences in the cervical and dorsal vertebrae are ontogenetic.  So let’s take a look at them.

Cervical vertebrae

Posterior, mid and anterior cervical vertebrae, in right lateral view, of (top to bottom), Haplocanthosaurus, Apatosaurus louisae CM 3018 (from Gilmore 1936:plate XXIV, reversed for ease of comparison) and Diplodocus carnegii CM 84 (from Hatcher 1901:plate III), scaled to roughly the same size.  For the diplodocids, we illustrate C13, C9 and C4.  For Haplocanthosaurus, we illustrate C14 of H. priscus (from Hatcher 1903:plate I) and C9 and C4 of H. utterbacki (from plate II).

It should be immediately apparent that the Haplocanthosaurus cervicals have less extensive pneumatic features than those of the diplodocids, but that is one feature which we know does vary ontogenetically.  There are other differences: for example, the cervical ribs in Haplocanthosaurus are level with the bottom centrum rather than hanging below.  Still, if you kind of squint a bit, you could probably persuade yourself that the Haplocanthus vertebrae look like possible juveniles of Diplodocus.

Unless you look at them from behind:

Posterior cervical vertebrae C15 and C14, in posterior view, of (top to bottom), Haplocanthosaurus priscus CM 572 (from Hatcher 1903:plate I), Apatosaurus louisae CM 3018 (from Gilmore 1936:plate XXIV) and Diplodocus carnegii CM 84 (from Hatcher 1901:plate III), scaled to the same centrum-to-neural-spine height.

(Unfortunately, these are the only Haplocanthosaurus cervical vertebrae that Hatcher had illustrated in posterior view, so we can’t compare more anterior ones.)

From this perspective, we can immediately significant differences:

  • First, that unsplit spine.  Yes, we know that Woodruff and Fowler (2012) have argued that it could be ontogenetic, but these are vertebrae from the most deeply bifurcated region of a diplodocid neck, in a decent sized animal, and there is nothing that so much as hints at bifurcation.
  • That whacking great ligament scar running right down the back (and also the front, not pictured) of the neural spine.  There is nothing like this in any diplodocid — neither on the metapophyses nor running though the trough.  And remember, scars like these tend to become more prominent through ontogeny.
  • The neural arch (i.e. the region between the postzygapophyses and the centrum) is taller in Haplocanthosaurusmuch taller in the case of C15.
  • The plates running out to the diapophyses are less dorsoventrally expanded in Haplocanthosaurus.
  • The centrum is smaller as a proportion of total height — especially, much smaller than in Diplodocus.
  • The parapophyses extend directly laterally rather than ventrolaterally (hence the position of the cervical ribs level with the bottom of the centrum).

So it doesn’t look good for the juvenile-diplodocid hypothesis.  But let’s take a look at the …

Dorsal vertebrae

Posterior, mid and anterior dorsal vertebrae, in right lateral view, of (top to bottom), Haplocanthosaurus, Apatosaurus louisae CM 3018 (from Gilmore 1936:plate XXV, reversed for ease of comparison) and Diplodocus carnegii CM 84 (from Hatcher 1901:plate VII), scaled to roughly the same size.  For the diplodocids, we illustrate D9, D5 and D2.  For Haplocanthosaurus, which has four more dorsals, we illustrate D13 and D7 of H. priscus (from Hatcher 1903:plate I) and D2 of H. utterbacki (from plate II).

Here we see that Haplocanthosaurus has dorsolaterally inclined diapophyses (which we’ll see more clearly in a minute), a prominent spinodiapohyseal lamina in posterior dorsals, and no infraparapophyseal lamination.  Also, the dorsal vertebrae have reached their full height by the middle of the series (in fact the last nine dorsals are startlingly similar in proportions), whereas in diplodocids, total height continues to increase posteriorly.

Now let’s see those vertebrae in posterior view:

Posterior, mid and anterior dorsal vertebrae, in posterior view, of (top to bottom), Haplocanthosaurus priscus CM 572 (From Hatcher 1903:plate I), Apatosaurus louisae CM 3018 (from Gilmore 1936:plate XXV) and Diplodocus carnegii CM 84 (from Hatcher 1901:plate VII), scaled to the same height of the mid dorsal.  For the diplodocids, we illustrate D9, D5 and D1.  For Haplocanthosaurus, which has four more dorsals, we illustrate D13, D6 and D1.

Here is where it all falls apart.  The Haplocanthosaurus dorsals differ from those of the diplodocids in almost every respect:

  • Of course we have the non-bifid spine in again, in the anterior dorsal, but let’s not keep flogging that dead horse.
  • In the mid and posterior dorsals, the neurapophysis is rounded in posterior view rather than square.
  • In the posterior dorsal, the neural spine has laterally directed triangular processes near the top.
  • All three Haplocanthosaurus neural spines have broad, rugose ligament scars, whereas those of the diplodocids have narrow postspinal laminae.
  • The neural spines (measured from the diapophyses upwards) are much shorter than in the diplodocids; but
  • The neural arches (measured from the centrum up to the diapophyses) are much taller.
  • The diapophyses have distinct club-like rugosities at their tips.
  • the diapophyses of the mid and posterior dorsals are inclined strongly upwards
  • The hyposphenes of mid and posterior dorsals have very long centropostzygapophyseal laminae curving up in a gentle arch.
  • The centra are smaller than those of Apatosaurus, and much smaller than those of Diplodocus.

(By the way, it’s interesting how very different the D5s of Apatosaurus and Diplodocus are.  Since both are from uncontroversially adult specimens, bifurcation was evidently very different between these genera.)

So based on the vertebrae alone, the case of Haplocanthosaurus as an immature form of Diplodocus or Apatosaurus is blown right out of the water.  And this is without even looking at the appendicular material — for example, the scapula and coracoid illustrated by Hatcher (1903:figs 17-19), which are so completely different from those of diplodocids.

But there’s more.  Tune in next time for the rest.

The rest of the series

Links to all of the posts in this series:

and the post that started it all:


  • Gillette, D.D. 1991. Seismosaurus halli, gen. et sp. nov., a new sauropod dinosaur from the Morrison Formation (Upper Jurassic/Lower Cretaceous) of New Mexico, USA. Journal of Vertebrate Paleontology 11(4):417-433.
  • Gilmore, C.W. 1936. Osteology of Apatosaurus with special reference to specimens in the Carnegie Museum. Memoirs of the Carnegie Museum 11:175-300.
  • Hatcher, J.B. 1901. Diplodocus (Marsh): its osteology, taxonomy, and probable habits, with a restoration of the skeleton. Memoirs of the Carnegie Museum 1:1-63.
  • Hatcher, J.B. 1903. Osteology of Haplocanthosaurus with description of a new species, and remarks on the probable habits of the Sauropoda and the age and origin of the Atlantosaurus beds; additional remarks on Diplodocus. Memoirs of the Carnegie Museum 2:1-75.
  • Lovelace, D.M., Hartman, S.A., Wahl, W.R. 2008. Morphology of a specimen of Supersaurus (Dinosauria, Sauropoda) from the Morrison Formation of Wyoming, and a re-evaluation of diplodocid phylogeny. Arquivos do Museu Nacional, Rio de Janeiro 65(4):527-544.
  • Lucas, S.G., Spielmann, J.A., Rinehart, L.F., Heckert, A.B., Herne, M.C., Hunt, A.P., Foster, J.R., Sullivan, R.M. 2006, Taxonomic status of Seismosaurus hallorum, a Late Jurassic sauropod dinosaur from New Mexico. New Mexico Museum of Natural History and Science Bulletin 36:149-162.
  • Upchurch, P. Barrett, P.M., Dodson, P. 2004. Sauropoda. pp. 259-322 in D.B. Weishampel, P. Dodson and H. Osmólska (eds.), The Dinosauria, 2nd edition. University of California Press, Berkeley and Los Angeles. 861 pp.
  • Wilson, J.A. 2002. Sauropod dinosaur phylogeny: critique and cladistic analysis. Zoological Journal of the Linnean Society 136:217-276.
  • Woodruff, D.C, and Fowler, D.W. 2012. Ontogenetic influence on neural spine bifurcation in Diplodocoidea (Dinosauria: Sauropoda): a critical phylogenetic character. Journal of Morphology, online ahead of print.

Special bonus illustrations

I composited the cervical and dorsal series above into the following compound illustrations.  As always, click through for full resolution.

Lateral view:

Posterior, mid and anterior dorsal vertebrae and cervical vertebrae, in right lateral view, of (top to bottom), Haplocanthosaurus, Apatosaurus louisae CM 3018 (from Gilmore 1936:plates XXIV and XXV, reversed for ease of comparison) and Diplodocus carnegii CM 84 (from Hatcher 1901:plates III and VII), scaled to roughly the same size. For the diplodocids, we illustrate D9, D5, D2, C13, C9 and C4. For Haplocanthosaurus, which has four more dorsals, we illustrate D13, D7 and C14 of H. priscus (from Hatcher 1903:plate I) and D2, C9 and C4 of H. utterbacki (from plate II).

Posterior view:

Posterior, mid and anterior dorsal vertebrae and posterior cervical vertebrae C15 and C14, in posterior view, of (top to bottom), Haplocanthosaurus priscus CM 572 (From Hatcher 1903:plate I), Apatosaurus louisae CM 3018 (from Gilmore 1936:plates XXIV and XXV) and Diplodocus carnegii CM 84 (from Hatcher 1901:plates III and VII), scaled to the same height of the mid dorsal. For the diplodocids, we illustrate D9, D5 and D1. For Haplocanthosaurus, which has four more dorsals, we illustrate D13, D6 and D1.

Broadly speaking, pneumatic sauropod vertebrae come in two flavors. In more primitive, camerate vertebrae, modeled here by Haplocanthosaurus, the centrum is a round-ended I-beam and the neural arch is composed of intersecting flat plates of bone called laminae (lam above; fos = fossa, nc = neural canal, ncs = neurocentral suture; Ye Olde Tyme vert pic from Hatcher 1903).

In more derived, camellate vertebrae, the centrum and neural arch are both honeycombed with many small air spaces. This inflated-looking morphology is very similar to that seen in birds, like the turkey we recently discussed. The fossae and foramina on the outside tend to be smaller and more numerous than in camerate vertebrae, as shown here in a titanosauriform axis from India (Figure 3 from Wilson and Mohabey 2006). The green arrows show that the fossae visible on the external surface are excavations or depressions into the honeycombed internal structure of the bone.

External fossae on bones can house many different soft tissues, including muscles, pads of fat or cartilage, and pneumatic diverticula (O’Connor 2006). Pneumatic fossae are often strongly lipped and internally subdivided and may contain pneumatic foramina, which makes them easier to diagnose (but they may also be simple, smooth, and “blind”, which makes them harder to diagnose as pneumatic). But in all of these cases we are usually talking about the same thing: a visible excavation into a corpus of bony tissue, which may have marrow spaces inside if it is apneumatic, or air spaces inside if it is pneumatic (the corpus of bone, not the dent). That’s probably how most of us think about fossae, and it would hardly need to be explained…except that sometimes, something much weirder happens.

Consider this cervical of Brachiosaurus (this is BYU 12866, from Dry Mesa, Colorado). Brachiosaurus and Giraffatitan have an in-between form of vertebral architecture that my colleagues and I have called semicamellate (Wedel et al. 2000); the centrum does have large simple chambers (camerae), but smaller, thin-walled camellae are also variably present, especially along the midline of the vertebra and in the ends of the centrum. As in Haplocanthosaurus, the neural arch is composed of intersecting plates of bone; unlike Haplocanthosaurus, these laminae are not flat or smooth but are instead highly sculpted with lots of small fossae. Janensch (1950) called these “Aussenkaverne”, or accessory outside cavities, because and they are smaller and more variable than the large fossae and foramina that invade the centrum.

And that’s the weird thing. As the red arrows in the above image show, the “Aussenkaverne” are not excavations or depressions into anything, except the space on the other side of the lamina (which in life would have been occupied by another diverticulum). The neural arches of Brachiosaurus and Giraffatitan are not excavated by fossae, they’re embossed, like corporate business cards and fancy napkins.

What’s up with that!? We tend to think of pneumaticity as reducing the mass of the affected elements, but the shortest distance between two vertebral landmarks is a smooth lamina. These embossed laminae actually require slightly more bony material than smooth ones would.

As you can see above, the outer edges of the laminae are thick but the bone everywhere else is very thin. Maybe, like the median septa in pneumatic sauropod vertebrae, the thin bone everywhere except the edges of the laminae was just not loaded very much or very often, and was therefore free to get pushed around by the diverticula on either side, in the sense of being continually and quasi-randomly remodeled into shapes that don’t strike us as being very mechanically efficient. But also like the median septa, the thin parts of the laminae are only rarely perforated (but it does happen), for possible (read: arm-wavy) reasons discussed in the recent FEA post. And maybe the amount of extra bone involved in making embossed laminae versus smooth ones was negligible even by the very light standards of sauropod vertebrae.

Another question: since these thin sheets of bone were sandwiched in between two sets of diverticula, why are the “unfossae” always embossed into them, in the medial or inferior direction? Why don’t any of them pop out laterally or dorsally, looking like domes or bubbles instead of holes, like Mount Fist-of-God from Larry Niven’s Ringworld? Did the developmental program get accustomed to making fossae that went down and into a corpus of bone, and just kept on with business as usual even when there was no corpus of bone to excavate into? I’m only half joking.

I don’t have good answers for any of these questions. I scanned this vert a decade ago and I only noticed how weird the “unfossae” were a few months ago. I’m putting all this here because “Hey, look at this weird thing that I can only wave my arms about” is not a great basis for a peer-reviewed paper, and because I’d like your thoughts on what might be going on.

In Other News

The Discovery Channel’s Clash of the Dinosaurs premiered last night. I would have given you a heads up, except that I didn’t get one myself. I only discovered it was on because of a Facebook posting (thanks, folks!).

COTD is intended to be the replacement, a decade on, for Walking With Dinosaurs. I’m happy to report that one of the featured critters is Sauroposeidon. I grabbed a couple of frames from the clips posted here.

I haven’t seen the series yet, because I don’t have cable. But I’m hoping to catch it at a friend’s place next Sunday night, Dec. 13, when the entire series will be shown again. With any luck, I’ll have more news next week.

Finally, I got to do an interview at Paw-Talk, a forum for all things animal. I’m very happy with how it turned out, so thanks to Ava for making it happen. While you’re over there, have a look around, there’s plenty of good stuff. Brian Switek, whom you hopefully know from this and this, is a contributor; check out his latest here.


Bifid Brachiosaurs, Batman!

September 6, 2009

These are the days of miracle and wonder, especially for all you right-minded people out there who are lovers of fine brachiosaurs.  I heard yesterday evening about a new paper in Proceedings of the Royal Society B: You and Li’s (2009, duh) description of a new brachiosaur, the first one known from the Cretaceous of Asia: Qiaowanlong kangxii. Best of all, it’s based primarily on vertebral material:

You and Li (2009:fig. 2)  Cervical vertebrae of Qiaowanlong kangxii holotype FRDC GJ 07-14.

You and Li (2009:fig. 2) Cervical vertebrae of Qiaowanlong kangxii holotype FRDC GJ 07-14. (a) Photograph and (b) interpretative line drawing of C4-C7 in left lateral view; (c) a distal portion of a cervical rib; C9 in (d) cranial, (e) left lateral, (f) caudal, (g) right lateral, (h) dorsal and (i) ventral views. di, diapophysis; f1-f5, fossa 1-fossa 5; pa, parapophysis; poz, postzygapophysis; prz, prezygapophysis; sp, neural spine. Scale bars, 10 cm.

Brachiosaur aficionados will be gazing slack-jawed at parts d, f and h of this figure (the anterior, posterior and dorsal views of C9), which clearly show that the neural spines of the new taxon are bifid (i.e. have two peaks side by side and a trough between them) — just like the cervical neural spines of flagellicaudatans (diplodocids and dicraeosaurs) and camarasaurs.  And mamenchisaurs.  And some titanosaurs.  And Erketu.  Finding this feature yet again — apparently independently evolved in brachiosaurs — makes it about the most plastic character in the matrix.  Very exciting.

That is, it’s exciting if this really is a brachiosaurid.  Now as it happens, Matt was one of the reviewers for this paper (and by the way did an amazingly professional job of not telling me about it until it came out, the git).  He’s told me in email that he’s satisfied that Qiaowanlong really is a brachiosaur, and I hesitate to question that identification given that (A) unlike the authors I’ve never seen the material, and (B) unlike Matt, I’ve spent most of my brachiosaur-presacral quality time with dorsals rather than cervicals.  But, with that caveat, I’m not sure that a compelling case has yet been made for a brachiosaurian identity.

The authors cite three characters in support of a brachiosaurid identity:

  • The most persuasive is the deeply excavated cervical neural spines.
  • Next is a transition in neural spine height: this is quite abrupt in “Brachiosaurusbrancai between cervicals 6 and 7, and also in Sauroposeidon — presumably also between C6 and C7, but that can’t be known for sure, since it’s only the assumption that this is the case that led to the identification of the four preserved Sauroposeidon cervicals as C5-C8 in the first place.  In Qiaowanlong, this transition is “much less pronounced”, with spines increasing in height by only 25% rather then 100% in the other taxa — and occurs between C8 and C9.  All in all, not really very similar to the condition in “B.” brancai.
  • The final character supporting the brachiosaurid identity of Qiaowanlong is the absence of an anterior centrodiapophyseal lamina.  As the authors point out, though, this lamina does exist in “B.” brancai and is absent only in Sauroposeidon; so if this is evidence of anything, it’s a synapomorphy of a clade uniting Qiaowanlong and Sauroposeidon to the absence of other brachiosaurs — something that seems very unlikely given the proportions of the vertebrae.

Putting it all together, there seems to be only one convincing brachiosaur character cited; and that stands against several non-brachiosaur characters, most obviously the bifurcation of the neural spine and the low Elongation Index (centrum length divided by cotyle height) but also by a few other characters that are not discussed in the paper.  For example, Matt has previously noted that in brachiosaur cervicals, the diapophyses are more anteriorly positioned than the parapophyses whereas in diplodocids the opposite is the case: as shown in fig 2(b) above, C6 at least of Qiaowanlong resembles diplodocids in this respect.

To try to get more of a handle on this, I put together a comparative figure of the 8th and 9th cervicals of various sauropods:

8th/9th cervicals vertebrae of various sauropods, scaled to the same centrum length.  "Brachiosaurus" brancai, Sauroposeidon; Qiaowanlong, Diplodocus; Haplocanthosaurus, Camarasaurus

8th/9th cervicals vertebrae of various sauropods, scaled to the same centrum length. From top to bottom and left to right: "Brachiosaurus" brancai, Sauroposeidon; Qiaowanlong, Diplodocus; Haplocanthosaurus, Camarasaurus. Six sauropod vertebrae for the price of one!

Based on overall proportions, I don’t find it intuitively obvious that the Qiaowanlong (middle row, left) more closely resembles the brachiosaurs (top row) than it does the other three.

What does all this mean?  Probably nothing: most likely there are further reasons for the brachiosaurid identification of the new taxon, and lack of space prevented their explanation and illustration.  We can hope that the authors, having placed an initial brief description in Proc. B, will follow it up with a more comprehensive description and analysis in a journal that does not impose such tight length restrictions.  But for now at least, my feeling is that the case for a bifid brachiosaur has yet to be made.

Moving on … Qiaowanlong is also represented by some nice appendicular material: the entire right side of the pelvis (ilium, ischium and pubis).  The ilium certainly looks brachiosaury, so that is another bit of support for the systematic hypothesis, but the proportions of the pelvic bones are very odd:

Right pelvis of "Brachiosaurus" brancai (left), based on composite of Janensch's (1961) figures, and Qiaowanlong (from You and Li 2009: fig. 3a).  Scaled to same ilium length.

Right pelvis of "Brachiosaurus" brancai (left), based on composite of Janensch's (1961) figures, and Qiaowanlong (from You and Li 2009: fig. 3a). Scaled to same ilium length.

You and Li (2009) describe their pelvis as having a “much reduced ischium”, but as is apparent by comparison with the pelvis of “Brachiosaurusbrancai, the ischium is in reasonable proportion to the ilium, and the oddity is more that the pubis is enormous.  So much so that it makes me feel a little ill looking at it, and it makes me wonder how certain it is that all three of these bones are from the same individual — sadly, the paper doesn’t discuss the association of the material.

[Not to flog a dead horse, but this kind of omission shows once more the perils of publishing new taxa in general-interest journals such as Proc. B that impose draconian length limits.  This paper just creeps onto page 7, and I simply don’t believe that it’s possible to do anything like justice to the description of a new taxon in that little space, especially when there is also geography, geology, phylogeny and discussion to be got through.  I don’t want to go all This Is How To Do It, but I can’t help remembering that Darren and I took 18 pages, nearly three times as long, to describe the single partial vertebra that is Xenoposeidon (Taylor and Naish 2007), and it’s not as though that paper wastes a lot of words.  To give You and Li credit, they did squeeze in photos of a representative vertebra from all six cardinal directions, which is great; but only at the cost of the photos being too tiny to be much use.  Please, folks: send your new taxon descriptions to a proper descriptive journal, not to a tabloid!  </hobbyhorse>]

Back on the Dinosaur Mailing List, B tH asked how big Qiaowanlong was.  According to the BBC, the authors say that “the dinosaur would have been a relatively small sauropod about 12m long, 3m high, and weighing perhaps 10 tonnes”.  Can we confirm that?  Well, the excellently comprehensive table of measurements in the paper gives centrum lengths, not counting the condyle, totalling 267 cm for the seven vertebrae C5-C11.  Janensch (1950a:44) gave measurements for the corresponding vertebrae of “Brachiosaurusbrancai HMN SII totalling 577 cm, which is more than twice as long.  If Qiaowanlong was 267/577 = 0.46 times as long as HMN SII, which Janensch (1950b:102) gave as 22.46 m, then it would have been 10.4 m long; it’s not obvious how the authors got the larger figure of 12 m unless they had reason to think the neck was proportionally shorter than in HMN SII.  If Qiaowanlong was isometrically similar to HMN SII, then it was 0.46^3 = 0.99 0.099 times as heavy.  Using my own in-press mass of 23337 kg for HMN SII, this would make Qiaowanlong only 2312 kg in mass — pretty pathetic for a sauropod.

That’s it for now.  I’d be the first to admit that there’s an awful lot of speculation in this post based on relatively little published information.  Hopefully You Hai-Lu will drop by and comment — I’ll be letting him know that I’ve posted this.


  • Janensch, Werner.  1950.  Die Wirbelsaule von Brachiosaurus brancai. Palaeontographica (Suppl. 7) 3: 27-93.
    Janensch, Werner.  1950.  Die Skelettrekonstruktion von Brachiosaurus brancai.  Palaeontographica (Suppl. 7) 3: 97-103.
    Janensch, Werner.  1961.  Die Gliedmaszen und Gliedmaszengurtel der Sauropoden der Tendaguru-Schichten.  Palaeontographica, suppl. 7 (1), teil 3, lief. 4: 177-235.
    Taylor, Michael P. and Darren Naish.  2007.  An unusual new neosauropod dinosaur from the Lower Cretaceous Hastings Beds Group of East Sussex, England.  Palaeontology 50 (6): 1547-1564.  doi: 10.1111/j.1475-4983.2007.00728.x
    You, Hai-Lu, and Li, Da-Qing.  2009.  The first well-preserved Early Cretaceous brachiosaurid dinosaur in Asia.  Proceedings of the Royal Society B: Biological Sciences.  doi: 10.1098/rspb.2009.1278.
  • Janensch, Werner.  1950.  Die Wirbelsaule von Brachiosaurus brancai. Palaeontographica (Suppl. 7) 3: 27-93.
  • Janensch, Werner.  1950.  Die Skelettrekonstruktion von Brachiosaurus brancai.  Palaeontographica (Suppl. 7) 3: 97-103.
  • Janensch, Werner.  1961.  Die Gliedmaszen und Gliedmaszengurtel der Sauropoden der Tendaguru-Schichten.  Palaeontographica, suppl. 7 (1), teil 3, lief. 4: 177-235.
  • Taylor, Michael P. and Darren Naish.  2007.  An unusual new neosauropod dinosaur from the Lower Cretaceous Hastings Beds Group of East Sussex, England.  Palaeontology 50 (6): 1547-1564.  doi: 10.1111/j.1475-4983.2007.00728.x
  • You, Hai-Lu, and Li, Da-Qing.  2009.  The first well-preserved Early Cretaceous brachiosaurid dinosaur in Asia.  Proceedings of the Royal Society B: Biological Sciences.  doi: 10.1098/rspb.2009.1278.

And finally … two announcements!

Traumador the Tyrannosaur has asked us to point out that over on ART Evolved (the palaeo-art blog), the next big art gallery is to be sauropod themed.  Details are on the site, so get over there and submit your sauropod art!

And Matt and I will shortly be teaming up with Andy Farke, the open-source paleontologist, on a new project where we plan to actually do some of this Shiny Digital Future that we keep on talking about.  Andy will be announcing the details on Tuesday 8th September.  Mark the date well!  For now, I shall say no more …

The respiratory system of a pigeon injected with pink latex. Click for the unlabeled gory version.

The respiratory system of a pigeon injected with pink latex. Click for the unlabeled gory version.

In case you’ve missed it, William Miller has been asking some great questions over in the comment thread for “Brachiosaurus: both bigger and smaller than you think“. Here’s his most recent, which is so good that the answer required a post of its own:

…in birds, the air sacs are obviously useful for flight, and they might have been useful for weight lightening in sauropods: but the common ancestor would have been flightless and too small to need the lightening. So what drove their evolution in the first place, I wonder?

To which I say: oh, Alice, the rabbit hole is a lot deeper than that.

Introduction to the Three Mysteries

First, in birds the diverticula that enter the bones are a comparatively small subset of all diverticula. Visceral, intermuscular, and subcutaneous diverticula run between the guts, between muscles, and under the skin, respectively. These are usually more numerous and more extensive than the diverticula that enter the bones, and with rare exceptions, like the subcutaneous “bubble wrap” in pelicans, we have no idea what they do. If, indeed, they do anything. All a character needs to do to be hereditarily propagated is not compromise the survival and reproduction of its bearer. Diverticula could be mostly functionless products of developmental processes that are usually invisible to selection but sometimes produce useful exaptations, like lightening the skeleton, insulating the body, etc. Sort of the evolutionary equivalent of the fire extinguisher in your kitchen: most of the time it does absolutely nothing, but once in a while it is really, really useful.

Second, postcranial skeletal pneumaticity (PSP) starts in the cervical vertebrae in basal theropods and sauropodomorphs, and possibly also in pterosaurs (Butler et al. 2009). The vertebrae adjacent to the lungs and air sacs are not the first to be pneumatized. Rather, the pneumatic diverticula must have gotten out of body cavity and traveled a ways before they started impacting the skeleton. Assuming that one thing had to come before another and it didn’t happen in one saltatory leap, diverticula must have evolved before they started pneumatizing the skeleton.

Third, in the earliest evolutionary stages of pneumatization in saurischians, the amount of bone removed is completely negligible. In Wedel (2007) I calculated that in Pantydraco (Thecodontosaurus caducus at the time) and Coelophysis the pneumatic spaces in the bones accounted for 0.0017% and 0.17%, respectively, of the body volumes. The fossae in the Pantydraco vertebrae are not absolutely diagnostic for PSP, but they’re in the right place and hard to explain otherwise. The holotype individual is a juvenile, and it is possible that PSP might have been more extensive in an adult, but it could increase one hundred-fold and still only be 1/500 of the animal’s volume, as in Coelophysis. Although I haven’t run the numbers, a similar result probably hold for the basalmost sauropods with definitive PSP.

Pneumatic fossae in the cervical vertebrae of Coelophysis

Pneumatic fossae in the cervical vertebrae of Coelophysis

To sum up:

  1. Most diverticula in birds are not involved with pneumatizing the skeleton, so PSP can’t be the reason for their existence.
  2. In basal saurischians, the diverticula that pneumatized the skeleton must have evolved before they could start pneumatizing the skeleton, so PSP can’t be the reason for their existence, either.
  3. In the early stages of the evolution of PSP in saurischians, the amount of mass saved was negligible and could not plausibly have influenced natural selection, so PSP didn’t initially evolve to lighten the skeleton.

Lighten Up, Fatso

There is a complication on that last point, which requires a little digression on fat.

In birds, pneumatic diverticula don’t just replace bone tissue, they also take up space that would be occupied by fat in mammals, for example in the spaces between muscles and around plexuses of nerve and blood vessels. Any of you who have had the misfortune to dissect the brachial plexus of a mammal know whereof I speak–you spend most of your time carefully picking fat out from around the nerves and blood vessels. This isn’t gross subcutaneous fat that means an animal or person is obese, this is adipose tissue doing its other job of being a lightweight packing material. Mammal bodies put fat in those spaces because they need to occupied by something light and squishy and fat is the cheapest thing your body can build.

That may seem backwards; we think of fat as an energy store and therefore energetically expensive. But it’s cheaper to build than muscule or cartilage or skin, and lighter than any other tissue or fluid in the body. It has been observed that even when mammals are starving, they do not use the fat in the yellow marrow that fills the marrow cavities of long bones. This is utterly unsurprising if you think about how bodies work. Nature really does abhor a vacuum, at least biologically (cosmically, it seems to be the biggest thing ever). If a starving body used the fat in the marrow cavity, it would have to replace it with something else, and all of the alternatives are heavier and more expensive to build. If the fat was not replaced, a partial vacuum would develop which would cause serous fluid to weep into the space, and that would also be heavier and more expensive, and a great site for infection to boot (ask someone who has an edema).

Birds cheat the system by replacing the lightest of tissues with something even lighter: air, held in diverticula that are basically super-thin layers of epithelium. Possibly diverticula had been running around replacing fat for a long time before they first entered the skeleton, in which case the earliest stages of pneumatization would have been a continuation of pre-existing function of replacing superfluous connective tissue (fat and bone are both forms of connective tissue, along with cartilage, ligaments, tendons, mesenteries, fascia, and blood; blood is connective tissue the way snakes are tetrapods).

Although that will be difficult or impossible to test, it actually makes quite a bit of sense. Getting fat out of the way ought to be easy; the lipids can be mobilized into the bloodstream and the flattened cells could either be pushed out of the way or resorbed. Getting bone out of the way requires increasing parathyroid hormone, mobilizing blood-born multinucleated osteoclasts, and convincing them to digest bone where it needs to be digested, which I assume is a more complicated process from a regulatory standpoint (physiologists or cell biologists, please correct me if I’m wrong!). So it seems plausible that diverticula might have acquired the ability to replace fat early on, and the ability to replace bone much later, and that by the time they got started on the skeleton diverticula could have been lightening the body for a long time by removing little bits of superfluous fat.

This does not contradict my statement above that by and large we don’t know what diverticula do. Some diverticula run where there is no fat to replace. And healthy birds do carry some fat, like any healthy tetrapod. One would think that this energy-reserve fat would need to be protected from diverticula that would otherwise resorb it, but I don’t know how or if that happens, and I don’t know if anyone else does either. The amount of research on diverticula is basically nil.

I also think that fat-resorbing diverticula don’t solve the third mystery, they just pushes it back a level. The amount of mass saved by replacing the “packing” fat with air is probably negligible in most animals, and it certainly would have been so in the earliest stages of replacement, so the third mystery still holds if it  is restated as:

3b. In the early stages of the evolution of diverticular replacement of connective tissue in saurischians, the amount of mass saved was negligible and could not plausibly have influenced natural selection, so PSP didn’t initially evolve to lighten the body.

A dorsal vertebra of Haplocanthosaurus in anterodorsal oblique view. It's pneumatic, sure, but not THAT pneumatic.

A dorsal vertebra of Haplocanthosaurus in anterodorsal oblique view. It's pneumatic, sure, but not THAT pneumatic.

The Problem is the Solution

So, we seem to be stuck. We don’t know why diverticula evolved in the first place, and we don’t know what most diverticula do, and even the diverticula that lighten the body could not have initially evolved to do so.

One upshot of all this is that we need more research on possible  physiological functions of diverticula in birds. Oy! Ornithologists and avian physiologists! We’ve thrown you a bone, now throw us some data. Please?

Another upshot is that the erratic evolutionary pattern of PSP in Triassic and early Jurassic ornithodirans is maybe not entirely unexpected. Pterosaurs and theropods seem to have had PSP right out of the gate, but at least in theropods it was not enough to have done any good. Basal sauropodomorphs had little or no PSP, and not enough to have done any good below about the level of Eusauropoda. No non-dinosaurian dinosauromorphs have been found with PSP, but then we only have a handful of them and they’re all pretty dinky, so it’s possible it just hasn’t been recognized yet. Silesaurids, at least, had very pronounced, very thin laminae, which in derived saurischians are almost always associated with PSP. And ornithischians never had PSP at all, as far as we know.

My opinion is that an air sac system is probably primitive for Ornithodira, and that most of these lineages had pneumatic diverticula, but the speed with which they “discovered” extensive, skeleton-lightening PSP–ranging from “almost immediately” in pterosaurs to “after a while” in theropods to “after a long while” in sauropodomorphs to  “never” in ornithischians–varied because it was such an evolutionarily haphazard process. Basically, PSP had to evolve as a developmental accident, and in some lineages it got far enough to become visible to selection, and in others it did not, or took a long time to do so. That’s a pretty picture that makes a lot of sense to me. If I ever figure out a way to test it, I’ll let you know.

What's going on here? Why are ornithischians so lame?

What's going on here? Why are ornithischians so lame?

The Solution is the Problem

The absence of PSP in Ornithischia is still a right sod. Pterosaurs, theropods, and sauropodomorphs all evolved some level of PSP in the Late Triassic, even if it wasn’t enough to significantly lighten their skeletons at first. Why not ornithischians? If air sacs are primitive for Ornithodira, then ornithischians had the gear for 160 million years and never exploited it, when the other three major lineages of ornithodirans discovered PSP pretty fast out of the gate. And if air sacs are not primitive for Ornithodira, three out of four ornithodiran lineages still discovered PSP on their own, so why not Ornithischia? It’s a big mystery, any way you slice it.

What do you think?


I had a new paper come out today. Unofficial supplementary info here, PDF here. I would have had all this ready to go sooner, but the paper came out sooner than I expected. In fact, I didn’t even know that it had been published until Andy Farke (aka the Open Source Paleontologist) wrote me for a PDF. Turnabout’s fair play, I suppose, because last year I congratulated Stuart Sumida on his Gerobatrachus paper before he knew it was out. I guess letting the authors find out through the grapevine that their stuff has been published is part of the “value added” that commercial journals provide. ;-)

Anyway, I’m happy the paper is out, finally. It’s the third chapter of my dissertation, but with teaching and traveling to Spain and such I didn’t get it submitted until last January. I had to forcibly bite my tongue during the Aerosteon saga last fall, when such a big deal was made about the absence of pneumatic hiatuses in non-avian dinosaurs. This despite the facts that there are several good reasons to expect pneumatic hiatuses to be rare, and that pneumatic hiatuses are not the Rosetta Stone or magic bullet for air sacs in saurischian dinosaurs. They’re more like the cinderblock that broke the camel’s back, given all the other evidence for air sacs.

In fact, the structure of the new paper is built around the idea that there are several tiers of evidence for bird-like air sacs in saurischians. Those tiers are:

  1. The presence of postcranial pneumaticity at all. Some of the first authors to get interested in the implications of pneumaticity for dinosaurs argued that pneumaticity probably implies an air sac system, and left it at that. Later workers have tended to denigrate this argument as overly simplistic–just because some of the postcranial skeleton is pneumatic does not mean that the animal’s air sac system was necessarily like that of birds–but it’s not actually a bad argument. We can imagine lots of ways to get air into the postcranial skeleton, but for tetrapods the only system that we have any evidence for is diverticula of a lung/air sac system like we see in birds.
  2. The distribution of pneumaticity in the skeletons of most saurischians and pterosaurs is  diagnostic for specific air sacs, namely the cervical, clavicular, and abdominal air sacs that we see in birds. This is what Pat O’Connor and Leon Claessens established so firmly with their work on mapping parts of the respiratory system to skeletal domains in birds.
  3. The evolutionary patterns of pneumatization in sauropods and theropods parallel the development of pneumatization during ontogeny in birds. Or, more economically, ontogeny recapitulates phylogeny in this system. This is more evidence that the observed patterns of pneumaticity in the skeletons of birds and non-avian saurischians are produced by the same underlying process of diverticula developing from different air sacs in a highly conserved order–even if we don’t know why things evolved, and continue to develop, in the order that they do. And it’s better evidence, because it accounts for more observations (points 1 and 2 can be established from single specimens) and ties postcranial pneumaticity in all saurischians, living and extinct, into a more coherent picture.
  4. Pneumatic hiatuses are more evidence that the postcranial skeleton is pneumatized by diverticula from more than one part of the respiratory system. Not the only evidence–we already suspect this quite strongly based on points 2 and 3–but more evidence. It is possible that the diverticula of extinct animals behaved differently than those of all extant birds, and diverticula from a single source could conceivably pneumatize the whole vertebral column. Possible. Conceivably. How likely? Dunno–our n on this is either 1 (if you count all living birds as a batch) or several hundred (if you count each of the species that Pat O’Connor has dissected and injected). Pneumatic hiatuses offer another level of evidence, because they can potentially show that the patterns of pneumaticity in fossil taxa are inconsistent with pneumatization from a single point. That’s how they work in chickens, and that’s how they may work in non-avian dinosaurs, as long as diverticula don’t leapfrog over  some bones without leaving any traces, or at least don’t do that very often.

For the record, I don’t think that pneumatic hiatuses are stronger evidence than point 3; if I was ranking the tiers based on importance I would put 3 at the top. Pneumatic hiatuses ended up being last in the paper because 1-3 were basically review material, and it made sense to group them together before the big  bolus of description.

[UPDATE the next day: also, I just realized that those 4 are not the same as I used in the paper! In the paper I left out 1, advanced 2 and 3, and added a different number 3, which is pneumatization of the pelvic girdle and hindlimb. I tend to forget about that one because the evidence in sauropods is underwhelming so far. And arguably this is just another aspect of 2 (above), or if you like you can think of 5 tiers. They say consistency is the hobgoblin of small minds!]

The importance of pneumatic hiatuses remains to be seen; there might not be enough of them to tell us very much, or we might find that leapfrogging diverticula exist and are common (we’d then need a way to sort hiatuses caused by multiple sources of diverticula from those caused by leapfrogging diverticula). But they’re important to me, for a couple of reasons.

First, they’re probably one of the two or three best ideas that I’ve had in my life. When I realized that pneumatic hiatuses could potentially indicate pneumatization from multiple sources it really was like a light going on in my head. I walked around seeing stars all week.

I got the idea from this figure, from King (1957):

king_57_figs1-2-480On the left King has drawn the vertebral columns of several chickens, and shaded in the pneumatic regions. Blocks of pneumatic verts separated by apneumatic gaps represent pneumatization from different sets of diverticula. I remember very vividly sitting in the Padian lab reading this paper and thinking, “if we found one of those in a dinosaur it would be the money.” Then I suddenly sat up straight, then stood up, then paced around the room a few times to burn off the discovery energy. I had a very profound need to tell someone. I don’t remember who I told, but it was probably Mike.

The other reason that pneumatic hiatuses are important to me: they are now one of those cool little cases that show that paleontology can be a predictive science. If you want to test a hypothesis in the experimental sciences you manipulate the conditions and see what happens. Historical sciences don’t usually give you that option. But you can play What If? As in, “If hypothesis A is true then we ought to see such-and-such evidence.” In 2003, I predicted that if sauropods had abdominal air sacs we ought to see pneumatic hiatuses once in a while. Finding the evidence that validated the prediction was almost as much of a rush as having the idea in the first place.

haplo-skeleton-with-me-for-scale-480The owner of Sauropod Pneumatic Hiatus #1 is Haplocanthosaurus CM 879, which is a cool animal but fairly pathetic as sauropods go. In my dissertation/job talks I would show the above picture and joke that I could probably beat up that animal on a good day. I found out about the pneumatic hiatus by accident, when I was poring over Hatcher (1903). In one of the figures near the end of the paper, Hatcher shows the centra of the fourth and fifth sacral vertebrae. I noticed that sacral 4 had a pneumatic chamber of some sort but sacral 5 did not. Then a few minutes later I had gotten to the plates at the back of the paper, and saw what looked like a pneumatic chamber on the first caudal. Somewhere in the dank, beer-flooded grottoes of my skull, the neuron fired.

haplo-verts-v3-480This is the figure I put together, using images from Hatcher (1903), for a Jurassic Foundation grant to go see the material in the Carnegie Museum in 2005. It worked; they came through with $1500 for that trip and a week at BYU the same fall (to my immense shame, although the Jurassic Foundation is credited for funding on the first page, I see that I forgot to thank them in the acknowledgments. Belatedly: thanks, you guys rock, I suck). The pneumatic cavities are labeled as foramina because that’s what they look like in the drawing, and not having seen them I didn’t know any better. In fact they are fossae, but they are deep, invasive fossae and their morphology is not consistent with anything other than pneumatic invasion. (Pneumatic invasion!? Flee for your lives!!) See the paper for all the excruciating details. Note that the sacrals have unfused neurocentral sutures, so the animal was not fully mature when it died (there is probably a whole post ahead just on the neurocentral weirdness in this animal).

So that’s the story, for now at least. There are more pneumatic hiatuses coming, but those papers are still in the pipe so I can say no more for now. I’m sure when they come out some alert blogger will notice and e-mail me for a PDF, and then you’ll get the news here.

The moral of the story is that you can make real progress by reading lots of old, obscure stuff. Support–and abuse–your local academic library!


Just got back from SVP, the annual meeting of the Society of Vertebrate Paleontology. SVP is always a blast; beyond the inherent coolness of four solid days of intense paleo-everything, I get a chance to catch up with my “SVP friends”–the pool of friends, some of them close, that I only see for a few days each year.

This year there was an extra bonus, which was making friends with a lot of paleobloggers that I’d previously only known as e-quaintances. There was even a paleoblogger lunch, which drew a big enough crowd that people at opposite ends of the table couldn’t effectively converse. That’s okay, though. I have a ‘critical mass’ theory about shared meals and parties, which is that they are usually lame unless there are enough people to sustain more than one conversation at the same time. By that standard, the paleoblogger lunch was a flying success. Unfortunately there weren’t any sauropods at the lunch, so it’s not really germane to our purpose here.

However, I did engineer a photo with a subset of paleobloggers that would be fair game, and you can see it above. This was at the welcome reception at the Cleveland Museum of Natural History. From left to right we’ve got Haplocanthosaurus (sauropod vertebra picture of the week: check), a disembodied floating T. rex head (it’s an oracle), a mosasaur, Megaloceros, Mammuthus (one of the least stinky of the stinkin’ mammals), and Allosaurus. Oh, and down in front are Julia Heathcote (Ethical Palaeontologist), moi, Amanda Northrop (Self-Designed Student), and Neil Kelley (microecos). Neil is the only real paleontologist of the bunch; you can tell because he’s holding a beer.


Mike drew the straw for our first birthday post, but I didn’t want to let it pass unremarked. This blog started as a joke on e-mail. “We ought to make a blog of nothing but sauropod vertebra pictures.” “Like APOD?” “Yeah, but weekly so we don’t drive ourselves crazy.” Next thing I knew, Mike had set up the blog and registered Darren and me as contributors. Then I posted a couple of times and BAM! we’re a year into it.

At first I thought we’d be struggling to get a year’s worth of material. I even made up a list of stuff I wanted to blog about, just in case I ever ran out of inspirado. I won’t tell you what’s on the list, but I will tell you that there are 27 things on the list and so far I’ve only gotten to six of them. So the outlook is good.

I have no idea how long we’ll be doing this. It never occurred to me after about the first month. If there is a Life Lesson (TM) in paleontology, it’s that nothing lasts forever. No sense getting an ulcer about it. Just enjoy what you have, and have a little fun.

We’ll do the same.

Hottt news

May 15, 2008

Mike Taylor has a loooong interview up at Laelaps. It’s sauropawesome.

The picture above has nothing to do with that, we just like to put sauropod vertebrae in every post. Here are some CT sections of a Haplocanthosaurus cervical (abbreviations: fos – fossa, lam laminae, nc neural canal, ncs neurocentral suture). I like them because they look nothing like what I expected. Not that the internal structure or laminae are unusual for sauropods, just that sauropod vertebrae themselves are unusual and sometimes the best way to be confronted with that is to see them from new vantages. I recycled this from Wedel (2007:fig. 13), and it basically just a rearrangement of Wedel (2005:fig. 7.3). The nice drawing of the Haplo cervical is from Hatcher (1903:pl. 2).

Well, now I’ve blabbed on for a paragraph and managed to cite myself twice in a post that is ostensibly about Mike. Seriously, go read the interview, it’s great.


Hatcher, J.B. 1903. Osteology of Haplocanthosaurus, with a description of a new species, and remarks on the probable habits of the Sauropoda, and the age and origin of Atlantosaurus beds. Mem Carn Mus 2:1-72.

Wedel, M.J. 2005. Postcranial skeletal pneumaticity in sauropods and its implications for mass estimates; pp. 201-228 in Wilson, J.A., and Curry-Rogers, K. (eds.), The Sauropods: Evolution and Paleobiology. University of California Press, Berkeley.

Wedel, M.J. 2007. Aligerando a los gigantes (Lightening the giants). ¡Fundamental! 12:1-84. [in Spanish, with English translation]