After my short post on what to leave out of a conference talk, here are few more positive thoughts on what to include, based on some of the SVPCA talks that really stayed with me.

First, Graeme Lloyd’s talk in the macroevolution symposium did a great job of explaining very complex concepts well (different ways of mapping morphospace onto phylogeny). It was a necessarily difficult talk to follow, and I did get lost a few times. But, crucially, Graeme offered lots of jump-back-on points, so I was never out of the game for more than a minute or so.

I think that concept of jumping-back-on points is important (albeit clumsily named). It’s easy, if someone is describing for example a detailed osteological point about a bone in basal tetrapods that doesn’t even exist in the animals we know and love, to tune out and lose the thread of the rest of the talk. There is an art in making it easy for people in this situation to tune back in. I’m not sure how it’s done: it might be more a matter of style than of content. I’ll think more on this one.

Also: several times as I watched Graeme’s talk, I internally raised an objection (such as low explanation-of-variation values of PC1 and PC2 in the plots he was showing) only for him to immediately go on to note the issue, and then explain how he deals with it. This should not be too difficult to emulate: anticipate possible objections and meet them in advance. This is something to have in mind when rehearsing your talk.

It was a talk that had obviously had a lot of work put into it. Another talk, which I shall not attribute, had very obviously been thrown together in 24 hours, which I think is flatly unacceptable. When you know you’re going to have a hundred professionals gathered in a room to listen to you, do the work to make it worth the audience’s while. Putting a talk together at the last minute is not a ninja move, or a mark of experience. It’s simple unprofessionalism.

Neil Brocklehurt’s talk was based on a taxon that was of very little interest to me: Milosaurus, and the pelycosaur-grade synapsid group to which it belongs. But his presentation was a textbook example of how to efficiently introduce a taxon and make it interesting before launching into details. There is almost certainly video out there somewhere — the SVPCA talks were filmed — and I recommend it highly when it becomes available. For fifteen glorious minutes, I was tricked into thinking that Carboniferous synapsids are fascinating. And it’s left me thinking that, hey, maybe they are interesting.



I deliberately left a lot of things out of the poster I presented at SVPCA: an abstract (who needs repetition?), institutional logos (who cares?), references (no-one’s going to follow them up that couldn’t find what they need in other ways), headings (all the text was in figure captions) and generally as much text as I could omit without compromising clarity.

In the same way, I found myself thinking a lot of the talks at his conference could have done with leaving some conventional things out — especially as talks now take place in 15-minute slots rather than 20 minutes.

Here are some things you don’t need to do:

  • Don’t start by saying the title. We can read it. Instead, while the title slide is up, tell us something about why we should care about your talk.
  • Don’t introduce yourself. It doesn’t matter if you’re in the last year of your Ph.D, or starting a postdoc with such-and-such person. We care about your science, not your biography (at least during your talk).
  • Don’t reiterate your conclusions at the end. We just heard them: if we can’t remember what you told us less than 15 minutes ago, we have bigger problems.
  • Don’t say “thanks for listening”. We’re here to listen to you. It’s why we came to the conference. You’re doing us a favour, not the other way around.
  • Don’t read the acknowledgements out loud. Nothing is more boring to listen to(*). Just leave the acknowledgements up on the screen as you finish, and we can read them if we’re interested.
  • Don’t say “I’ll be happy to take questions”. It’s the moderator’s job to invite questions — and indeed to judge whether there enough time.

Why omit these things? Most importantly, because they waste time, which you want to use to tell us your story. Your work is fascinating and we want to hear all about it. Do all you can to make space for it.

[See also: Tutorial 16: giving good talks (in four parts)]


(*) Except talks about mammal teeth, of course.


Revising a poster

September 14, 2017

Yesterday, Matt and I showed our posters in a two-hour session at SVPCA, as two of about thirty. It was actually a hugely positive experience, and it’s left me wondering whether to prefer posters to talks at future conferences — one of us will write about it separately

My poster (left) and Matt’s (right), in their natural environment. Phil Mannion (Mammalia: PrimatesHomoninae) for scale.

But it was eye-opening to road-test what I’d thought was a pretty good poster with real visitors. I quickly realised that the “Biconcavoposeidon” poster was missing two important things.

First, there’s an insert that shows schematic views of amphiplatyan, opisthocoelous, procoelous and amphicoelous vertebra. But it should show two or three of each kind of vertebra, in articulation, so that the insert shows not only the isolated shapes, but how they fit together (or how, in the case of biconcave centra, they do not fit together.)

The second problem: another insert shows — again in schematic form — how vertebrae articulate in amphibians and mammals. But I really needed a third part of that figure showing how the articulation works in birds. I kept needing to point to such an illustration, and I didn’t have one.

But all is not lost: I am not bound to an imperfect poster for all time. The poster is published as part of a PeerJ Preprint, and I can revise that preprint as often as I like. Which means I can revise the poster.

And that is what I plan to do. I’ll make both of the changes described above, and update the published version. The question then becomes: in my publications list, where the poster is explicitly tied to the presentation at SVPCA 2017, should I continue to point to the version that I actually used at the meeting? I am inclined to think so.

But wait: there’s more.

It turned out at the conference that Matt was right when he advised me that I should have made the anaglyph bigger, and placed it at the top of the poster, at eye-level. Instead, I had a sequence of visitors who had to painfully kneel down and peer myopically at the small, low-down version that I used. (Or, more often, they didn’t bother looking at the anaglyph at all — which is a shame, because it’s really informative.)

(Seriously, folks: buy yourself some dirt-cheap red/cyan glasses, and start making use of these incredibly useful, and very simple-to-make, visuals. Everyone who looked at the anaglyph, without exception, was startled at how much more informative it was than the regular image.)

But I am not going to re-organise the poster along those lines in the forthcoming revision. Why not? Because I don’t expect to present the new version at a conference, where eye-level is an issue. Instead, I expect it to stand as a research artefact in its own right, to be viewed on screens or printouts — and for those purposes, the present composition is better. That’s true especially because most people downloading the poster won’t have the red/cyan glasses necessary to view what would be centrepiece of the putative revision; but when presenting the poster at a conference, I can provide the glasses (and I did).

So I think I have now landed on the notion that a poster as a research artefact is a fundamentally different thing from a poster for presentation at a conference. I didn’t see that coming.

If you don’t get to give a talk at a meeting, you get bumped down to a poster. That’s what’s happened to Matt, Darren and me at this year’s SVPCA, which is coming up next week. My poster is about a weird specimen that Matt and I have been informally calling “Biconcavoposeidon” (which I remind you is not a formal taxonomic name).

Here it is, for those of you who won’t be at the meeting (or who just want a preview):

But wait — there’s more. The poster is now also formally published (Taylor and Wedel 2017) as part of the PeerJ preprint containing the conference abstract. It has a DOI and everything. I’m happy enough about it that I’m now citing it in my CV.

Do scientific posters usually get published? Well, no. But why not? I can’t offhand think of a single example of a published poster, though there must be some out there. They are, after all, legitimate research artifacts, and typically contain more information than published abstracts. So I’m happy to violate that norm.

Folks: it’s 2017. Publish your posters.


  • Taylor, Michael P., and Mathew J. Wedel. 2017. A unique Morrison-Formation sauropod specimen with biconcave dorsal vertebrae. p. 78 in: Abstract Volume: The 65th Symposium on Vertebrate Palaeontology and Comparative Anatomy & The 26th Symposium on Palaeontological Preparation and Conservation. University of Birmingham: 12th–15th September 2017. 79 pp. PeerJ preprint 3144v2. doi:10.7287/peerj.preprints.3144v2/supp-1


Or, how a single lateral fossa becomes two foramina: through a finely graded series of intermediate forms. Darwin would approve. The ‘oblique lamina’ that separates the paired lateral foramina in C6 starts is absent in C2, but C3 through C5 show how it grows outward from the median septum. How do I know it grows outward, instead of being left behind during the pneumatization of the more posterior cervicals? Because with very few exceptions, all neosauropod cervicals start out with a single lateral fossa on each side, as illustrated in this post. But many of them end up with two or more foramina. Diplodocus is a nice example of this (from Hatcher 1901: plate 3):

I should clarify that the vertebrae above show that character transformation in this individual, at this point in its ontogeny. The vertebrae of CM 555 are about two-thirds the size of those of CM 3018, the holotype of A. louisae. In CM 3018, even C4 and C5 have completely divided lateral fossae, corresponding to the condition in C6 of CM 555.

As Mike and I discussed in our 2013 neural spine bifurcation paper, isolated sauropod cervicals require cautious interpretation because the morphology of the vertebrae changes so much along the series. The simple morphology of anterior cervicals reflects both earlier ontogenetic stages and more primitive character states. As Mike says, in sauropod necks, serial position recapitulates both ontogeny and phylogeny. So if you have a complete series, you can do something pretty cool: see the intermediate stages by which simple structures become complex.

If you’re thinking this might have something to do with my impending SVPCA poster, you’re right. Here’s the abstract.

For more on serially increasing complexity in sauropodomorph cervicals, see this post.


August 15, 2017

Here is a fascinating sequence of five consecutive posterior dorsal vertebra — AMNH FARB 291 from the”Big Bone Room” at the AMNH:

AMNH FARB 291, five consecutive posterior dorsal vertebrae of a probably brachiosaurid sauropod, in right lateral view. The vertebrae are embedded in a plaster block, which has been desaturated in this image.

Matt and I first saw this specimen back in February 2009, when we were mostly there to look at Apatosarusminimus (and then again in 2012). As soon as our eyes lit on it, we couldn’t help but be captivated by its bizarre biconcave centra. We immediately started flippantly referring to it as “Biconcavoposeidon” — the ugliest name we could come up with — and in our subsequent discussions the name has stuck (often abbreviated to “BCP”).

  • Taxonomic note: for avoidance of doubt, “Biconcavoposeidon” is not and will never be a formal taxonomic name, only an informal specimen nickname. If at some future point we conclude that this specimen represents a new taxon, and name it, we will definitely not use the name “Biconcavoposeidon”. If you ever use the name, please do not set it in italics.

As you can see in this front view, the specimen is sheared: the upper part of the vertebrae have been displaced to their left (which is the right as we see it in this image):

AMNH FARB 291, most anterior of five consecutive posterior dorsal vertebrae of a probably brachiosaurid sauropod, in anterior view.

Apart from the shearing, though, and the truncation of the neural spines shortly above the transverse processes, the specimen is in pretty good nick. Crucially, it’s not been “restored” in plaster to conceal what is and is not real bone — unlike many specimens of that era. It came out of the Bone Cabin quarry in 1898, back when scientific information was routinely discarded in order to obtain a more beautiful-looking specimen.

This is the specimen that I’ll be presenting at SVPCA this year — though only as a poster, unfortunately: there’s no talk for me, Matt or Darren this year. We’ve posted our abstract (including the illustration above) to the nascent PeerJ collection for SVPCA 2017, and we’re looking forward to seeing more of the materials from that conference — abstracts, then manuscripts, then papers — appearing in the collection.

So far as we know, there’s no other sauropod specimen with biconcave posterior dorsal vertebrae. (And, no, Amphicoelias is not an exception, despite its name.) But have we missed any?

I got an email this morning from Jim Kirkland, announcing:

All of the lectures (with permission to be filmed) will be available on the NHMU YouTube channel. I just wrapped the edit of the 6th video which should be available later today. However, 5 of the lectures are now edited and already available for viewing. They can be found here.

And by the time I read that message, the sixth talk had appeared!

Each talk is 20-25 minutes long, so there’s a good two and a quarter hours of solid but accessible science here, freely available to anyone who wants to watch them. Here, to get you started, is long-time friend of SV-POW!, Randy Irmis, on Discovering Dinosaur Origins in Utah:

It’s great that the DinoFest people are doing this. In 2017, it should really be the default — and yet I can’t think of a single vertebrate palaeo conference that has done this before. (Did I miss some? Links, please!)

I know it’s one more thing for conference organisers to have to think about (or, more optimistically, one more thing for them to delegate). but I hope we’ll be seeing a lot more of it!