Go read this: Marugán-Lobón et al. 2013 on semicircular canal orientation and head posture in saurischian dinosaurs
August 7, 2013
Figure 1: Differences in reference systems in skulls.
(A) In the human skull there is a 30° difference between the Frankfurt plane and that of the LSC, thus yielding substantially different head orientations (from de Beer, 1947). (B) When a stork is in alert its LSC is oriented 19° above the horizon, thus when putting the LSC at 0° (horizontal) head posture differs from its alert posture (from Duijm, 1951).
I know it’s been quiet around here for a while. Mike and I have both been on vacation, and before that, we were both up to our necks in day-job work, and after we get back, we’ll be up to our necks in revising accepted manuscripts. So no time for a long post right now, but I couldn’t let this pass without notice: Jesús Marugán-Lobón, Luis Chiappe, and Andy Farke just published a cool paper on semicircular canal orientation in saurischians and its value–or lack thereof–as a reference system. This is something Mike and Darren and I have addressed before (here and here), but Marugán-Lobón et al. have gone waaaaay further than anyone else I know if in addressing the inherent variability in lateral semicircular canal orientation.
The TL;DR, from the abstract:
The variability of LSC relative to skull landmarks is large (ca. 50°) and likely unpredictable, thus making it an inconsistent reference system for comparing and describing the skulls of saurischian (sauropodomorph and theropod) dinosaurs.
But you shouldn’t stop there! The paper is short, straightforward, and freely available on PeerJ, so go read it. Read the review comments, too–like an increasing number of authors, Marugán-Lobón et al. put the whole paper trail up along with the finished paper. Nice work!
Sauropod Vertebra Picture of the Week 
August 8, 2013 at 8:14 am
But who ever suggested using LSC as a reference system for describing/comparing saurischian skulls (besides apparently two sixty year old ornithology papers)? Marugan-Lobon et al. say these “works were inspiring, and have been widely followed not just in the context of interpretations of how extinct animals held their heads but also as a reference system for descriptive morphology and anatomical comparisons”. Of the papers cited as doing this, Rogers (1998) noted the LSC was parallel to the cerebral surface in Allosaurus, but didn’t use any novel reference system. Sampson and Witmer (2007) only use it for the endocast of Majungasaurus, not the skull, and the current papers says nothing about endocast orientation. Sereno et al. (2007) only says the skull of Nigersaurus was held vertically in life, they still use dorsal/anterior and such to describe and label the anatomy in a standard fashion (e.g. the frontal is horizontal). Evans (2006) seemingly never mentions the semicircular canal and orients lambeosaur skulls normally. Witmer et al. (2008) again only use it for endocast illustration, along with a figure of supposed alert head postures. They certainly never claim e.g. Nanotyrannus’ skull should be interpreted in a descriptive fashion as being tilted compared to Tyrannosaurus’, only that the alert posture was supposedly different and that this has implications for the synonymy of both genera. Finally, I lack Witmer et al. (2008) so cannot comment.
Basically, this paper seems to attack a position no one currently holds. If Sereno had said the frontal of Nigersaurus was mainly exposed anteriorly, or Witmer had said Nanotyrannus differed from Tyrannosaurus in the ventrally angled maxilla, ventrally angled nasal, orbit facing anteroventrally, etc., then there might be an issue worth addressing. But since the worst points are Witmer describing Majungasaurus’ endocast using a coordiate system angled 0.2 degrees from standard and illustrating theropod endocasts with the LSC horizontal, it seems a minor point.
August 9, 2013 at 4:06 pm
The paper doesn’t seem to attack any position. It found a substantive inconsistency in the value of using the LSC as a baseline for head-attitude, and plugged that data against cranial morphology, which itself is interesting. This is good stuff; raw data is NEVER bad. People have to get out of the habit of thinking that papers have to find something wrong to be useful.
August 10, 2013 at 6:50 am
Er… even the abstract blatantly states the authors view others as holding a position, so they tested it, and found the position was false. “In recent years, the orientation of the lateral semicircular canal (LSC) has been regularly used to determine skull orientations for comparative purposes in studies of non-avian dinosaurs.” “In light of our results, the lateral semicircular canal is an inconsistent baseline for comparative studies of craniofacial morphology in dinosaurs.” Maybe ‘attack’ could be seen as being too harsh a word, but the authors surely engaged what they saw as a common idea.
I would disagree that raw data is never bad, as we have a limited amount of time and energy to devote to writing and reading papers. It’s easy to think of utterly pointless papers that while providing true raw data also waste peoples’ time. “Are Archaeopteryx bones made of minerals or cheese?”, etc.. The present case isn’t as bad, as apparently Duijm (1951) actually did suggest “craniofacial anatomy was best compared when skulls were oriented according to the animals’ alert posture”, but that’s such an obscure reference that I don’t see there’s much need to engage with it. It’s not even cited on Google Scholar. That said, it’s not as if I begrudge this paper’s existence. I just don’t see why Matt values it so highly.
That’s not to say there aren’t important related issues that could have been discussed, even besides the obvious one of how closely LSC angle is related to alert posture (which the authors say they don’t tackle, and that’s fine). Since Witmer orients endocasts based on a horizontal LSC, we could test whether this is a more neutral coordinate system than e.g. cerebral angle, or Procrustean analyses of endocasts. Or we could test whether Marugan-Lobon et al.’s Procrustean system that averages skull landmarks is more neutral than using e.g. the ventral maxillary long axis, or the ventral jugal long axis. Are oviraptorids better thought of as having ventrally angled snouts, ventrally angled posterior skulls, or a mix of the two? Maybe these questions are next for the authors.
Finally, I certainly never claimed papers have to find something wrong to be useful. Foth and Rauhut’s (2013) new paper in PLoS ONE is a great example. They wondered if differing reconstructions of specimens were confounding morphometric analyses, tested it, and found that the answer was largely no. Different reconstructions of the same specimen usually fall very close together. So now we know, and that’s great.
January 15, 2019 at 6:04 pm
(I am finding this quite late, but I thought it was worth replying)
I read the paper some time ago and I found it great because there is a clear tendency in dinosaur paleobiology to use the lateral semicircular canal as if we were in the 19th century. Obviously, this paper criticizes such approach, using geometric morphometrics (and avant garde methodology) to unambiguously demonstrate that, 1) the assumption that the inner ear lies parallel to the horizon when animals are in alert is definitely wrong even if DeBeer argued in favor of it almost a century ago. Apparently, this idea has pervaded multidisciplinary biological research for decades because people cross-reference wrong interpretations instead of going to the original source (Duijm’s 1951 paper, which this PeerJ paper did). 2) it further stresses (empirically and unambiguously, again) that the inner ear is an incorrect baseline for anatomical comparison in tetrapods (even if it used dinosaurs as an example), just in case people is tempted to use such a baseline (which seems to be the case, judging from the comments on this paper). 3) the paper elegantly demonstrates the power of geometric morphometrics. Chapeau