The Open Source Triceratops Paper

Wow! The past few days have been a real whirlwind. As mentioned in my last post, our long-awaited paper (well, long-awaited by my co-authors and me) on Triceratops horn use and paleopathology was unleashed by PLoS ONE on Tuesday evening. We've received some wonderful feedback through this blog and email, and are very ready to move on to the next research project.

Because this blog focuses partly on open access and open source software, I wanted to write briefly about how these issues factored into our research. This is one of the most "open" projects I've attempted to date, from start to finish. Here are the details:

• The data collection and statistical analysis were completed in OpenOffice.org Calc, with the calculations based on a spreadsheet file I found (and tested against known examples) from somewhere on the internet (but can't remember where, now!). R will also do the appropriate calculations, but I stuck with the spreadsheet because my data files were so small and straight-forward.
  
• The manuscript was written in OpenOffice.org Writer, but my co-authors and I batted it back and forth in Word format (because Darren and Ewan are using the latter program). We had no problems with this strategy, and the format conversions were a snap for the relatively simple documents we were using.
• The bibliography was compiled and formatted using Zotero. Zotero even has a readily available style file for PLoS, so this made my life very easy.
  
• The figures were edited for contrast and brightness in GIMP (no other manipulation was performed on the images) and assembled in Inkscape.
• The journal, PLoS ONE, was selected because of its high profile, high impact, and open access. Thanks to the open access policy, our article is readable by anyone who wishes to see it. I hope that the broader exposure will facilitate debate and further research on the topic--only time will tell. If I have any future articles of potentially broad interest, PLoS ONE will definitely be on my list of candidate journals.
• The popularization of the article was initiated by PLoS ONE, with follow-ups by numerous journalists and bloggers. Thanks to the wonders of the internet, many of the articles are available for free. And, I am happy to say, most of them are pretty well-done.
  

Triceratops Combat?

Exactly how did Triceratops and other horned dinosaurs use their cranial weaponry? Today, my co-authors Ewan Wolff, Darren Tanke, and I published new research in the online, open access journal PLoS ONE, giving our take on the issue. In a study spanning four years and over a dozen museums, we have marshalled what we think is the best evidence to date that Triceratops may have locked horns with their own kind.

Restoration of Triceratops in horn-to-horn combat. Image copyright Lukas Panzarin, courtesy Raymond M. Alf Museum of Paleontology.

The Background
The ceratopsids, or horned dinosaurs, present a bewildering array of cranial headgear. Triceratops was one of the earliest discovered and best known, with its distinctive dual brow horns, single nose horn, and a solid frill of bone over the neck. Subsequent discoveries revealed a clade chock full of anatomical diversity - single horns, triple horns, and frills sporting all sorts of odd accessories. Consequently, a whole host of functional explanations have been dreamed up for these bones.

Defense against predators was an early, popular explanation. With threats such as Tyrannosaurus stalking the Cretaceous landscape, Triceratops and relatives surely would have used their horns for defense if they had to. Yet, this is not the primary function of the horns in modern animals. More commonly, horns, antlers, and other appendages are used for intraspecific combat and display--not against predators. A good example is the massive horns of the bighorn sheep. Careful behavioral studies have shown them to be employed in impressive visual display to other sheep, or for combat over territory, mating rights, dominance, etc. These modern analogs were quickly compared with ceratopsids.

In addition to the determination of function by analogy, some paleontologists noted odd markings on the skulls of some horned dinosaurs. Perhaps a chunk of the frill was missing, with evidence of healing. Maybe there was an abnormal, oddly positioned hole on the skull. Inevitably, these oddities were attributed to "horn gouge wounds." These alleged pathologies were interesting--but anecdotal. Without a broader context, any attempts to infer a cause for the "injuries" usually amounted to constructing a "Just So Story."

Recently, the tide has turned against the idea of horns-as-weapons in Triceratops and its relatives. Research by Kevin Padian, Jack Horner, Mark Goodwin, and others have suggested that the odd cranial ornamentation of many dinosaurs was for more peaceful purposes--specifically, species recognition (picking out members of their own kind amongst a sea of similar species). Although this idea certainly has its merits, I was never quite convinced that the horns of Triceratops were completely benign (and to be fair, I don't know that the species recognition proponents were suggesting this as the only function, either). At the same time, I wasn't exactly convinced that every single ding on a fossil skull should be considered evidence for combat.

What We Did
I've been really interested in the issue of horned combat in Triceratops for some time. Several years ago, I published a brief study, using plastic models, that demonstrated such combat was at least physically possible. I made some predictions about where we should expect injuries on the skull, noted similarities with published anecdotes of alleged injuries, and left the issue open for a follow-up study. The models were fun to play with, but I wanted to know if this combat actually happened in real life! And, I knew that paleopathology--the study of disease and injury in fossils animals--was probably the best approach. Unfortunately, I didn't know that much about the topic.

So, I teamed up with two experts. Darren Tanke works at the Royal Tyrrell Museum of Paleontology, and is an expert in both paleopathology and ceratopsian dinosaurs. Ewan Wolff recently finished his dissertation on oral pathology in archosaurs at Montana State University, and now is a student at the University of Wisconsin-Madison School of Veterinary Medicine. Ewan and Darren provided expertise in identifying and interpreting the markings we observed on the ceratopsid skulls.

Our logic was simple--if ceratopsid dinosaurs used their horns against each other in combat, evidence of this should be found in bony lesions on the skull. And, assuming that animals with different horn configurations fight differently (a fair assumption from observations of modern horned critters), we should find differing rates and patterns of lesions between different species of horned dinosaurs. If the horns were not used in combat (or specifically, patterns of horn locking that could result in injury to the skull), we should see no difference between various species.

So, we surveyed many, many specimens at over a dozen museums across North America. We looked for any evidence of pathology, and recorded its location on the skull. Two genera were the focus of our research--Triceratops (pictured at the top of the post) and Centrosaurus (pictured below this paragraph). Centrosaurus has only a single nasal horn (and no massive brow horns), so we assumed that it would be most likely to show differences from Triceratops if there were any to be found. After years of data collection, we subjected all of the numbers to statistical analysis. Lots of folks (Darren and I included) had described anecodotal reports of pathology before--but nobody had subjected them to the rigors of statistics.

One artist's concept (courtesy Wikimedia Commons) of Centrosaurus in combat.

What We Found
Briefly, we found that most bones of the skull that we examined did not show statistically significant differences (i.e., P>0.05) in rate of pathology between the two genera. Except, that is, for the squamosal bone of the frill (shown on the figure below). But what does this mean?

Schematic showing rates of pathology (number of pathological elements/total elements examined) for Triceratops (top) and Centrosaurus (bottom). The squamosal rates, highlighted here, were different from each other at P<0.05.

Did Triceratops Fight Each Other?
As you can probably guess by now, we postulate that the differences in rates of pathology between Triceratops and Centrosaurus were caused by behavioral differences. In other words, because Triceratops were locking horns with either other, they got injured. However Centrosaurus used its horns (whether for combat or display, or whatever), they didn't commonly cause injury to the squamosal bone. We do consider some alternative explanations for the lesions (predator attack, microbial origin, etc.), but generally discount these (read the paper for a full explanation).

As a reminder, if the horns were only being used for display (species recognition, or whatever), we wouldn't expect to have seen these differences in pathology rates between the two species. Thus, I am reasonably comfortable saying that the horns weren't "just for looks."

What Does This Mean for Dinosaur Research?
The statistical study presented here is some of the best evidence to date (in our opinion) for intraspecific combat in Triceratops. Centrosaurus may well have done the same, just in a very different manner (flank butting, or something like that). Or, maybe Centrosaurus was a more peaceful taxon--we just don't know yet. Lots more research is needed in this area.

Another big question is why these animals were fighting. Mating rights? Sexual dimorphism hasn't been demonstrated in the taxon, so it doesn't seem like males and females are set up any differently in terms of horn configuration (but further study may change this). Territory? Who knows (and good luck investigating this with fossils!). General bad temper? Your guess is as good as mine. We may never know the why in this case.

Centrosaurus came from an ancestor with big old brow horns--perhaps this ancestor lost the horns in favor of a kinder, gentler approach. The latest centrosaurines (living millions of years after Centrosaurus), such as Pachyrhinosaurus, ditched their horns altogether in favor of big bony pads on the skull. So, did centrosaurines evolve towards less risky forms of combat? Again, this is another hypothesis that should be tested more thoroughly. There's lots more work to do!

On a more methodological note, we hope that more studies of dinosaurian paleopathology will adopt a statistical approach. Anecdotes are interesting, and can be very informative in some cases, but they only tell part of the story. There is strength in numbers!

What Aren't We Saying?
Because cranial function is a contentious issue, I just want to make clear on what we aren't claiming. (I am speaking for myself here--my co-authors are probably largely in agreement)

• All injuries on ceratopsid skulls are attributable to combat. No, no, no. Just because there is a ding on the skull doesn't mean it's a combat wound. Almost certainly, some non-combat injuries are included in our sample--this is why we looked at overall patterns, rather than anecdotal cases. I would be willing to say, though, that the overall pattern of lesions (at least on the squamosal) suggests that most of them were caused by the horns of other ceratopids.
• Ceratopsid horns were only for combat. Again, no. This isn't the case in modern animals, and it likely wasn't the case for ceratopsids. As I (and Darren) have said elsewhere, ceratopsid horns were probably like Swiss Army Knives--multi-purpose tools. Fighting, defending, showing off, cooling off, you name it. I am just saying that now we shouldn't exclude fighting as a function.
• All ceratopsids used their horns identically. No. Different horn morphologies suggest different uses--and I would be willing to believe (as others have suggested) that animals such as Pachyrhinosaurus focused on flank-butting behavior (rather than head-to-head combat). Again, this will need more work.
  

Parting Thoughts
This has been a really, really fun study, and one that I hope will spur more research on the possibilities for skull function in horned dinosaurs. We haven't heard the last on this issue, rest assured. There are lots more specimens, new species to describe and discover, and many other people working to better understand these magnificant animals. I would also like to publicly thank my co-authors--I have learned so much about paleopathology and ceratopsids from you! Finally, a big thank you to our artist, Lukas Panzarin, who skillfully brought to life a long-vanished Triceratops battle.

Farke AA, Wolff EDS, Tanke DH (2009). Evidence of Combat in Triceratops. PLoS ONE, 4 (1):e4252. DOI: 10.1371/journal.pone.0004252
Read and download it for free at: http://dx.plos.org/10.1371/journal.pone.0004252

Coming soon. . .other goodies related to this paper, including the open source software used in the research, the open access publishing experience, and much more!

It's Almost Here. . .

All shall be revealed later this week.

Open Access Publishing and the Paleontologist

Open access publishing provides free and unrestricted access to scientific works. This contrasts with the traditional model of publishing, in which an interested reader either needed to purchase a personal copy of the publication or access to a library with a subscription. The rise of the internet has made open access a reality--and a thorny issue for the paleontologist looking to get published.

In this post, I'll talk about some of the benefits and drawbacks of open access publishing from the perspective of a paleontologist. I'll also discuss how to choose an open access journal.

Kinds of Open Access
Open access comes in many different flavors. From least to most restrictive, they are:

• Open Access Immediately. Your paper is available, for free to the reader, from the instant of publication. Palaeontologia Electronica and PLoS One are widely known examples.
• Open Access After Awhile. The paper is available to paid subscribers for a set length of time (perhaps six months or a year), and then becomes freely available. Journal of Experimental Biology, as one example, uses this method.
  
• Hybrid Access. The article is in a normally closed-access journal, but the author pays a fee to allow open access for all readers. Journal of Systematic Palaeontology allows this option.
  
• Author's Choice Access. The author pays a fee to a closed-access journal for the privilege of posting a PDF of the said article on a personal website. Journal of Vertebrate Paleontology is a well-known example of this mode.
  
• Author's Email Access. The author pays to get a PDF file that they can then distribute (legally) via email only. JVP also allows this mode.

Benefits of Open Access
The most striking benefit of open access is availability--the article is free to download for anyone who has an internet connection. The paper is picked up by Google and other search engines pretty much automatically, and people who are looking for that sort of research will usually find it.

From the author's standpoint, open access publishing is also usually faster than the standard model. Everything is done electronically, including the final publication. There is no lag time to mail the journals out, and some open access journals have a "rolling" publication schedule (meaning that once your article is approved, it's out there). Because many of the journals aren't bound by page constraints, they can place as many articles in a single issue as are available. There's no waiting 14 months for space to be available in the journal (a frighteningly common problem in many journals - if you've published papers, you've probably experienced this at least once).

Finally, the electronic-only mode of many open access journals allows some creativity in presentation. Need to have a movie as Figure 1 in your paper (rather than just a supplement)? Go right ahead (in some cases)!

The Costs of Open Access
Like many "regular" publications, some (but not all) open access journals have page charges. Depending on the journal, these can run up to $3,000 for a single article. A partial list of typical fee schedules is available here (thanks to Dr. Randy Irmis for passing along this link).

What is a poor paleontologist to do? In some cases, universities (that is, if you're based at a university) will partially subsidize the cost. In other cases, the journals (such as PLoS One) may partially or completely waive the open access fee for those with financial need. Finally, you can always find an open access journal with no fees (and they are numerous!). The most ethical of open access journals will not consider ability to pay in making decisions on publication.

A more critical concern, from the standpoint of relatively unestablished authors, is that some in the scientific community don't consider open access journals "real" or as good as their closed access counterparts. Yes, there are some open access journals out there with questionable editorial practices (they are in the minority, fortunately) --but there are also printed journals with loose standards (and these are also in the minority). And, there are many quite excellent open access journals, which are widely cited and recognized as such! So, the "quality" argument doesn't hold a lot of water in my view. Unfortunately, this reality doesn't necessarily hold up to the perception of the search, award, or tenure committee (rightly or wrongly).

Hope for the Future
How does a researcher support open access and circumvent this last concern? As a compromise, I might recommend using open access journals as one part of a publishing portfolio. After all, there are some really good closed-access publications out there (e.g., Journal of Vertebrate Paleontology, etc.) that do well at getting research into the hands (and minds) of the right audience. Some open access journals (Acta Palaeontologica Polonica, PLoS One, etc.) also do a good job at this. Choose the right journal for your paper--and an open access journal if that will get the research to the right people. Out of all the papers I have published, the one that gets cited the most was in Palaeontologia Electronica. I submitted it there on a whim, and have been pleasantly surprised by just how widely read it has been. Finally, just write good papers! A good paper is recognized as one no matter what, regardless of the journal.

Time will be the most important factor in changing attitudes. As open access publishing becomes more common, it will be seen as less freaky by the powers-that-be. And, this is a very, very good thing.

It's Coming. . .

Look for a major research announcement in one week. . .

Welcome, Longhorns!

I hear from a reliable source that your Digital Methods in Paleontology course has this blog listed as a recommended website for some supplemental readings. So, to all of you UT Austin students, welcome! I hope that the content here is at least somewhat useful. . .don't hesitate to post if you have any questions or comments. I (and my readers) are particularly eager to hear if you run across any other good software tools that aren't listed here, or if you have your own feedback on some of the software I've reviewed previously.

Your first reading assignment (if I'm reading the syllabus correctly) is found here. . .it gives some good background info on the blog. If you're completely bored, I would strongly recommend this post as a logical follow-up. Despite the name of this blog, I am not an open source zealot, and the referenced post gives some of the pros and cons of using open source software. Don't know what open source software is, exactly? This page is as good as any for a succinct introduction to the concept.

Good luck in the coming semester, and enjoy the class!

Workshop for the Digital Paleontologist

If you're a grad student interested in using mathematical techniques in your paleontological research, this workshop is for you! John Alroy and colleagues are presenting the fifth annual Paleobiology Database Summer Course in Analytical Paleobiology, hosted at the campus for the University of California's National Center for Ecological Analysis and Synthesis in Santa Barbara.

The rest of this post is taken directly from the announcement to the VRTPALEO list by John Alroy. I never had the chance to participate in this workshop (field season and all, and then I was too old of a grad student), but wish I had!

About the course
Since 2005 the Paleobiology Database has conducted a five-week intensive course in analytical paleobiology at the University of California's National Center for Ecological Analysis and Synthesis in Santa Barbara. In 2009 the course is scheduled to run from 30 June to 4 August, following NAPC. It will be supported primarily by the Paleontological Society with additional contributions from NESCent, the Palaeontological Association, and the Society of Vertebrate Paleontology.

Topics will include community paleoecology, quantitative biochronology, diversity curves, speciation and extinction, phylogenetics, phenotypic evolution, and morphometrics. Both simulation modelling and data analysis methods will be employed. The course will combine lectures and labs. Students will be given hands-on instruction in programming using R and trained in other analytical software. In addition to the course coordinator, each week a new instructor will be present. The instructors are expected to be John Alroy, Gene Hunt, Tom Olszewski, Pete Wagner, and Mark Webster.

There is no fee for registration, and students will be housed for free in apartments on the UCSB campus. Students are urged to apply for travel funds from their home institutions. If such funds are not available, travel expenses may be reimbursed for up to $400 if coming from the United States, $600 if coming from Western Europe, or $800 if coming from other countries. Students are responsible for meal expenses. There are no other charges of any kind, and no other major expenses are likely.

How to apply
Participating students should be in the early stages of their own research in any area related to paleontology. They should have a background in basic statistics, and preferably also programming. The ability to understand rapidly spoken English is essential. The course is open to undergraduates and advanced graduate students, but first or second year graduate students are particularly encouraged to apply. We also strongly encourage applications from women, minorities, and international students. Applications from professionals who have completed their studies will be considered, but strong preference will be given to students.

Applications should be submitted in PDF format to John Alroy (alroy@nceas.ucsb.edu). The review process will begin on 15 February 2009, and applications received by midnight Pacific time on that day will receive priority. Applications should consist of a one page statement. Do not include separate documents such as a curriculum vitae. No form needs to be filled out.

The statement should include a brief description of current research plans, a list of degrees earned stating the year of graduation in each case, a brief list of relevant classes taken, and an account of the student's previous use of statistics and programming. Students who do not employ English as a primary language should describe their experiences learning and speaking it. Applicants are encouraged to explain why the topics addressed by the course are of special interest to them, and which of these subjects are taught at their home institutions.

Applications must be accompanied by a recommendation letter, also in PDF format, written by the student's academic advisor and e-mailed separately. Obtaining a recommendation from anyone who is not an advisor must be explained. It is important that the recommendation give details about the applicant's personal character and abilities, not just credentials and descriptions of research projects. Recommendation letters also must be received by the end of the due date.