Tuesday, November 4, 2014

Deinocheirus > Spinosaurus

There’s a new giant, sail-backed, piscivorous theropod in town.

Just a month after Ibrahim et al. (2014) revealed their new interpretation of Spinosaurus aegypticus, Lee et al. have done one better: they've published a description of Deinocheirus mirificus, one of the most mysterious dinosaurs ever found. For those not keeping score at home, Deinocheirus was discovered in 1965 during a Polish-Mongolian expedition into the southern Gobi Desert. Unfortunately, all that was found was a pair of enormous arms—indelibly seared into the minds of every man, woman, and child who opened a dinosaur book over the next forty years. It was pretty easy to see that Deinocheirus was an ornithomimid, but it must have been absolutely enormous. It wasn't until 2006 and 2009 that more skeletal material was discovered—both in the field and in a private collection. For more on this, Ed Yong and Brian Switek have fascinating write-ups about the discovery and its history.


It's a duck-billed, hump-backed, horse-headed, slow-moving ominovorous ostrich dinosaur.

I’ll talk about the actual animal another time. Frankly, this picture (by Michael Skrepnick) may be all you need. It’s a giant, duck-billed, horse-headed, hump-backed, air-filled, slow-moving ornithomimid that probably ate anything it could get its beak around. And between the two specimens, there’s basically an entire skeleton in there. That’s exciting.

But here, I’d like to talk about how this paper absolutely trumps Ibrahim et al.’s Spinosaurus paper (2014). Yeah, I know, it’s low-hanging fruit, but somebody has to point these things out. Also, you'll recall that I'm an English major (with a rhetoric focus) so it's within my wheelhouse.

The first thing I noticed: Lee et al. manage to cram more data into their paper than Ibrahim et al., even though the former is significantly shorter than the latter.

Deinocheirus (Lee et al.): Four total pages in the main description, only two of which are text, eighteen pages of text in the supplementary file, and eight pages of "extended data" figures and tables. These may as well be counted as parts of the supplementary file, so let's say four pages of description (only two of which are text) and twenty-six pages of supplement. Total: thirty pages.

Spinosaurus (Ibrahim et al.): Three pages in the main description, only two of which are text, a whopping forty-eight-page supplement, and three pages of "extended data" figures and tables. These may as well be counted as parts of the supplementary file, so let's say three pages of description (only two of which are text) and fifty-one pages of supplement. Total: fifty-four pages.

Notice that it's been more than a week, and any sort of social media/blogosphere backlash against Deinocheirus is nowhere to be found. This despite the fact that Lee et al. are describing a similarly-legendary dinosaur previously known from intriguing but frustrating remains and rebooting it entirely. The two even share a love of piscivory. Lee et al. have produced a bizarre-looking, sail-backed, piscivorous animal that is very different from its relatives. In a way, Spinosaurus and Deinocheirus couldn't be more similar.

But nobody is batting an eye about the ostrich dinosaur. Meanwhile, the paleo community is STILL discussing Spinosaurus, although discussion has slowed considerably now that Ibrahim et al. have told us all stop talking while they write the monograph.

Here's the difference: Lee et al. are very reasonable--conservative, even--in their conclusions, do not make questionable assumptions, and wrap up every loose end. They don’t say anything like “it must have been a quadruped” and then never discuss how a quadrupedal theropod would work. The Internet took that ambiguity and ran with it. By contrast, Lee et al.’s most daring assertion is that Deinocheirus was a slow-moving omnivore. And they back that statement up—it has a long femur, broad pedal unguals, a therizinosaur-like ilium, sacrum that tilts the dorsal spine upward, and relatively short metatarsals. Stomach contents include over a thousand gastroliths and fish bones.

Miraculously, all of these details are written up in the main body of the description. You don’t need to go to the supplementary file to find the details. The supplementary file is really just an extension of the description proper, with a more thorough bone-by-bone analysis and a discussion about how the skeletons came to be found.


This is not the same as...

This. One of them is a misleading composite, and one is a responsible composite.

The figures are also much better. The front page includes photographs of the two specimens in a neutral, non-ambiguous pose so that readers can see how much of each specimen exists, then a skeletal reconstruction in a neutral, non-ambiguous pose as a composite. This is exactly the opposite of how Spinosaurus was shown: a CG model in a swimming pose that misrepresented the length of the femur* and did not adequately explain how much of a composite the new skeletal was. And while there are some photographs in the supplementary file, there weren't nearly enough to get an idea of exactly how much Spinosaurus aegypticus material was known.

I sense that if we saw the new Spinosaurus specimen by itself, there wouldn’t be much to see. No such guesswork is required for Deinocheirus: what you see is what actually exists.

As I said, the supplementary file includes an interesting narrative about how the Deinocheirus specimens were discovered, tracked down, and brought together. The Spinosaurus supplementary file includes a lengthy series of specimen referrals and a poorly-justified case for designating a neotype. The Deinocheirus paper includes excellent photographs of the skull and other bones, along with interpretive line drawings. The Spinosaurus supplementary file includes a drawing of Spinosaurus’ skull with known skull bones colored blue. There are no photographs of those bones. I know from reviewing past literature that these blue bones are a composite, but Ibrahim et al.'s paper does not state that fact.

This is not the same as...

This. One of them is a misleading composite. The other is a photograph and line drawing.

To me, it boils down to this:

Lee et al. say “Look at this thing we found.”

Ibrahim et al. say “Take our word for it, and wait an indeterminate amount of time for the monograph.”

Only one of those statements is good science.

Here’s what the Lee et al. paper proves: you can have good figures, good description, good conclusions, and unambiguous assertions in a brief paper. The Deinocheirus paper does NOT read like an extended abstract or poster presentation. My previous assertion that you simply can’t have a good, detailed paper in Nature or Science is wrong. You clearly can. While I think we all would have preferred to read about Deinocheirus in an open-access journal with more room for multiple, large figures, I'm impressed that Lee et al. were able to work within the restrictions of Nature and produce a quality paper. Despite being twenty-four pages shorter (in all), there is simply no question that Lee et al. have produced a far more comprehensive, reasoned, thoughtful, and descriptive scientific article than Ibrahim et al.

There are lessons to take from this comparison, and I hope writers do take them.

I hope you enjoyed this post, and congratulations to Lee et al. on this stunning discovery. I may have a new favorite giant dinosaur.

*Recall that, in their post at Scott Hartman's website, the authors note the following:

“Also, in the 3D model, the femur is not hanging straight down—it is angled outward/laterally slightly and so foreshortened slightly in lateral view, making it look shorter.”
As I said before, there's no use in putting ambiguity in your description. It leads to Scott Hartman and others trying to figure out where things went wrong.

7 comments:

  1. I'm all for blogs providing reasoned critiques and responses to papers, but some comments here are misleading and quite insulting (i.e. suggesting the Spinosaurus paper imagery is useless, that the femur was incorrectly scaled - we thought it was, but it wasn't).

    Comparing papers as if there's some sort of palaeontological league table isn't right. The construction of scientific papers is not, as implied here, a simple, straightforward affair. They're written by people from different backgrounds and at different professional levels. They have to consider professional politics, the preferred interpretations and biases present in diverse authorship teams, as well as research plans which may not be immediately obvious to readers. Throw referees and editors into that mix and the situation becomes more complex. I don't doubt that there are better and worse ways to approach scientific papers, and that some papers are more useful than others. But all authors deserve a basic level of respect for putting new data out there, if nothing else for having worked and trained hard to get to the position where they can write and publish a paper. That respect is lacking from this article.

    Furthermore, I disagree that the two projects are as comparable as you make out. The Deinocheirus project has significantly more material, and clear stratigraphic and geographic provenance. They have much less to prove, in that sense. The Spinosaurus project, on the other hand, is attempting to make sense of a much poorer dataset, spread across space and time. In that respect, it's a more ambitious project. Was it wrong to attempt to do this all in a relatively short Science paper, and did that affect presentation of the data? You could argue that case, but given that a single Science paper can make a career, I fully understand why the authors took the opportunity to publish where they did. I suggest considering the bigger picture - not to mention the fact that you're talking about real people here, not anonymous hypotheses and facts - when writing articles such as this one.

    ReplyDelete
    Replies
    1. Mark, thank you for the reasoned comment. You're right--the figures in the Spinosaurus paper aren't "useless" and I'll change that in the body of the text. However, I believe they are misleading (feel free to disagree). In both the skull and multi-colored skeletal restoration, the figures are making big assumptions about what material actually exists and what it might look like. I covered that in my original Spinosaurus post, and I won't go into it again here. But the point still stands: Lee et al. provide several photographs of bones with line drawings, while Ibrahim et al. simply do not, even in cases where they HAVE the bones.

      I will repeat one thing that I said in that post: Ibrahim et al. are making claims that they don't back up in the text. Maybe they can't in such a short amount of space, or they haven't worked out all the details yet. Don't make claims you can't support. Period. I don't necessarily blame them for publishing in Science--as you write, it is a career-boosting move. That particular issue (impact factor) is a problem for academia generally, and must be addressed at some point. However, that doesn't mean the authors should be saying things like "it must have been quadrupedal" without discussing how that would work, or justifying why so many isolated bones were referred to Spinosaurus (hasn't Andrea Cau already suggested the humerus is from a sauropod?).

      And remember, the authors commented on Scott Hartman's blog and said that:

      “Also, in the 3D model, the femur is not hanging straight down—it is angled outward/laterally slightly and so foreshortened slightly in lateral view, making it look shorter.”

      That's not cool, guys. It create ambiguity where there's no need for it.

      You talk about how much more material Lee et al. had to work with. And I completely agree. Of course they did. And they made perfectly reasonable conclusions based on that material. Ibrahim et al. had far LESS material to work with, yet they made more ambitious assertions that were perhaps less testable. And, look, Ibrahim et al. had fifty-something pages to work with. They could have gone into a LOT more detail in the supplementary file, but they didn't. It's not even a formatting issue. My point is they had MORE room to strengthen their conclusions than Lee et al. did.

      "Was it wrong to attempt to do all this in a relatively short Science paper, and did that affect the presentation of the data?" At the risk of sounding improper or disrespectful, I would answer that in the affirmative.

      And I AM talking about real people here, and the decisions they make, and the things they could do better. I realize this isn't Zanadu--we can't all have nice things and there are funding concerns and politics to consider--but I'm not going to hold back and play nice when I see thing that I interpret as irresponsible or poorly drafted.

      Feel free to call me out, as you've done, where you disagree. I do apologize for characterizing the Ibrahim et al. figures as "useless," and like I said, I will change that.

      I also don't think we necessarily need "referees," but it's been my experience that every writer needs an editor.

      Delete
  2. I can't edit my own comment?!

    Something I forgot to add before: In case it's not clear, my thesis here is that the Spinosaurus paper is just plain poorly-written by comparison to the Deinocheirus paper, for a variety of reasons. I mean no disrespect; how is this any different from pointing out deficiencies in any other paper--which I've read many of on the blogosphere over time?

    ReplyDelete
  3. "Lee et al. provide several photographs of bones with line drawings, while Ibrahim et al. simply do not, even in cases where they HAVE the bones."

    OK, but it's not like this is uncommon. That doesn't make it right, but team Spino don't deserve to be dragged across the coals for it.

    "Ibrahim et al. are making claims that they don't back up in the text. Maybe they can't in such a short amount of space, or they haven't worked out all the details yet. Don't make claims you can't support. "

    Again, I don't necessarily disagree - even the authors have suggested some details could have been clearer - but this has all been said elsewhere in more reasoned critiques. As with any criticism, these things can only be said so many times before they stop being constructive.

    On the femur bowing: it doesn't seem to affect the scaling, so it can't be that misleading. We all know how the referred partial-skeleton was scaled now, and I've not seen anyone saying that they cannot replicate it from the published material. Is it time to stop mentioning this as a problem?

    "Ibrahim et al. had fifty-something pages to work with. They could have gone into a LOT more detail in the supplementary file, but they didn't. It's not even a formatting issue. My point is they had MORE room to strengthen their conclusions than Lee et al. did."

    Both authors would have infinite space - there's no limit on supplementary material. But again, the Spino paper is not alone in perhaps not fleshing out its methods and conclusions as much as it could: why continue to pick on it and the authors?

    "I'm not going to hold back and play nice when I see thing that I interpret as irresponsible or poorly drafted. "

    But is it really necessary to stirring? I understand that the Spinosaurus paper left a lot of readers wanting more, but the continued bashing of the paper - which now includes being personally critical of the authors, it seems - is getting a bit much. There is some good stuff in the paper: Ibrhaim et al. presented a lot of new data which can now be evaluated and tested, and, as consequence of their work, we are moving towards an improved understanding of Spinosaurus and its world. That's a good thing, and we need to remember that.

    "how is this any different from pointing out deficiencies in any other paper--which I've read many of on the blogosphere over time?"

    Because critiques of science - which I think we all agree are fair game - are spilling over to attacks on the professionalism and integrity of Ibrahim et al., which isn't on. The constant recycling of issues with the Spinosaurus paper, or attempting to show how other works are 'superior', is not only gratuitous, but becoming increasingly rude to the real people behind the paper.

    ReplyDelete
    Replies
    1. I agree that the issues with the Spino paper have been picked clean by other blogs and social media. It wasn't necessarily my intention to revisit them--even though I basically ended up doing that. Look, I was happy to never bring up the Ibrahim et al. paper again, but when I read Lee et al., I was immediately taken aback by the contrast between the two papers.

      In my previous post, I basically said "the authors could have done better." In this post, I'm saying "here's an example of how this can be done better."

      I'm not sure if calling the Ibrahim et al. paper "poorly written" is a personal attack (or increasingly rude) as much as it's a criticism that could be leveled at ANY piece of poorly-written literature. Do better next time, that's all. But you're right in that I haven't done a great job of pointing out that there IS good information in the Spino paper. There certainly is--especially the bone density stuff, pedal morphology, and leg proportions. It's a dachshund dinosaur!

      I also wonder if some of the paper's deficiencies could have (should have?) been caught in peer review.

      But I probably won't discuss it further. Thanks for the interesting comments; they've given me a lot to consider. Also, Blogger's comment text window is ridiculously small--irritatingly so.

      Delete
    2. Agree to disagree but Zach has a point. Ibrahim et al. was somewhat half-assed. Especially since the paper read like an SVP abstract.

      Like Zach said in his first point. Ibrahim et al. have no one to blame but themselves for ambiguities in the paper. Especially since the model gave poor views of the femur for reference.

      The point of a skeletal model is to provide clear, un-misleading data NOT to go "Oh look guys our Spinosaurus model is swimming".

      Furthermore as Cau pointed out the authors did not make their "knuckle-walking" animation available for third-party testing. Why is that?

      The authors also never provided a reason for their tallest vertebral spine placement or their line of reasoning for placing the ambiguous caudal of the holotype in the most anterior position possible.

      All in all, the Spinosaurus paper had a lazy,rushed feel to it in some places.

      Delete
    3. Furthermore Lee et al. DID provide clear views of their skeletals without any ambiguities.

      Ibrahim et al. made the poor choice of using a 3D model with a perspective effect that makes measuring of the femur difficult. This whole scaling controversy would have never happened if Ibrahim et al. didn't make so many poor decisions.

      Delete