What is Citipati gracilis?

In paleobiology, you get an interesting mix of personalities when it comes to taxonomic treatment. Various workers, often depending on the school they attended or the personal choices and views that they have on taxonomy or evolution, develop “quirks” of theirs in which they seem to “lump” or “split” a little too much. I prefer a somewhat more moderate, conservative approach to taxonomy (by not naming things) but that’s largely just a personal choice in where I intend to go.

On the one hand, you get people like O. C. Marsh, who was willing to name every little piece he could as a new fish, lizard, dinosaur or pterosaur, often in an attempt to preempt his rival E. D. Cope. Not that Cope wasn’t prolific, I just have a soft spot in using Marsh as the exemplar for this particular habit. The reasons for this perspective can be manifold, including historical observation on what seems diagnostic, or just an attempt to name something quicker than another guy, especially if, as in Marsh’ case, some of your guys are working the same quarry, are rushing their material back to you in New England, and you’ve just finished trying to destroy your rival’s camp’s store of fossils. Oh well. They didn’t call them the Bone Wars for nothing.

Heroes of the Bone Wars (that's Othniel Charles Marsh on the left, and Edward Drinker Cope on the right). From Wikipedia commons.

Marsh (and Cope) were what might be called splitters, or folks who rendered taxonomy into a game of darts by erecting many names for what could be considered fewer species (or “genera,” if you swing Linnaeus’ way). They may egotistically attempt to sink (or “lump”) the other guys’ taxa, but they’re still splitters.

On the other hand, you get people who attempt to use various reasons to merge taxa, often with an over-riding prevarication, but sometimes with a real sense of what’s an unnecessary taxon, such as for the purpose of recognizing an ontogenetic sequence for which various stages have names attached to them. I recently just posted on one of these, where John Scannella and Jack Horner [1] resolved a sequence to show that Torosaurus is a late ontogenetic variation of Triceratops. But there are some trickier cases. There are people who “lump” with very little justification, or with an assumption of anatomical variation (such as Greg Paul’s comparison of theropod variation to Canis, Homo, Anas, Gazella, Panthera, Ursus and monitor variation in Varanus [2:201-203]), while withholding or refraining from including more detailed metrics of variation, such as genetics. Granted, at the time Predatory Dinosaurs of the World [2] was written, genetic testing and cladistic analysis were in their infancy. There is no excuse now.

Before I continue, a short piece on cladistics

Cladistic analysis is made up of two parts: A character-taxon data matrix, and a mathematical analysis in which algorithms are applied to the dataset in such a way as the closest relationships between every single pair of taxa in the matrix can be found. This produces results that can range in the hundreds of thousands of trees (also known as cladograms, which are simply graphical outputs of pairwise closest relationships). One can look at each individual tree, or one can perform various secondary analyses, one of which is called a concensus (there are a couple of different ways you can do this, but I’ll skip those). In such an analysis, ALL of the trees are crunched together, and find what each of them has in common; it will also produce a set of numbers that relate to how strongly each of these positions are supported by the number of trees that find them, so that the more there are, and the more characters that support each relationship, the better supported the output data (or, again, “tree”). In many ways, the output data is only as good as the input data (the taxa included, and the characters chosen; if you include only some data, but exclude some specific point that could be included, your analysis is flawed because of this, but fortunately someone else can come along and fix this by adding them in. The inclusion of as much characters and as much taxa as possible (in what is a total-data or all-in analysis) is the extreme and preferred goal of an analysis that is chosen, a priori, piecemeal.

Some have chosen to reconstruct phylogenies without such a tool, using inferrence of importance of a few select features, ignoring their context in relation to other data, or specifically excluding them with assumptions that they are unimportant [now or ever]. Phylogenetics, the practice of organizing taxa based on their relationships, has a tendency to ignore ranks (not that it perceives they do not exist, but because they are not necessary to explain the relationships of animals). “Genera” can in this scheme be arranged in such a way that it may seem to many to be nested within one another. This occurs only because different people would label each of the nodes that connect species differently; a single phylogeny can be labeled in enough ways to provide ample room for taxonomists to not run out of things to name for decades, if not centuries.

What this is all about

Greg Paul has recently published a book, The Princeton Field Guide to Dinosaurs [3], which is a 320 page volume deomstrating Paul’s mastery of the paleontological artist’s skill. Unfortunately, as from his previous works Predatory Dinosaurs of the World [2] and Dinosaurs of the Air [4], a concerted effort to ignore cladistics and erect specific and “generic” arguments based on little more than “these animals seem similar.” This resulted in Paul arguing that many apparently or assumed closely-related taxa are in fact the same thing — to a level. This is generally derived from the asusmption of the reality of the “genus” as a special entity apart from many other clades. It is generally not an issue in many cases, but one of the problems with the use of the “genus,” and the referral of taxa up to that point, is that it ignores the looser nesting of taxa regardless of “rank.”

Paul, for example, argues that virtually all oviraptorids are a single “genus:” Citipati. This is personally interesting to me because the various oviraptorids are actually well-defined from one another (and because I study them to some degree, so there is an applied level of assault on my assumptiosn). Paul argues [3:153-154] that several of these forms are crestless juveniles of larger, crested forms, and lumps them all into a large, crested form. This is all done without analysis, merely assumption of ontogeny. Secondarily, Paul lumps “Ingenia” yanshini in with Conchoraptor gracilis (as if they were the same form) and Khaan mckennai as a juvenile of Citipati osmolskae.

Mickey Mortimer had a short piece [5] to say about this book, speaking more generally of the taxonomic acts which Paul performed (of which the oviraptorid lumping is a broad-swath effect of):

Citipati barsboldi, gracilis, huangi and mongoliensis– Instead of Nemegtomaia, Conchoraptor, Heyuannia and Rinchenia.  Synonymize derived oviraptorid genera if you like, but Conchoraptor has priority over Citipati.  Paul seems to think Conchoraptor gracilis and “Ingenia” yanshini are synonyms, so yanshini should be the name he uses for that species.  Where’s Khaan you ask?  A juvenile of Citipati osmolskae according to Paul.  I bet Jaime will have some harsh words for these synonymies.

You betcha.

In total, Paul creates the following behemoth:

Citipati Clark, Norell & Barsbold, 2001
= “Ingenia” Barsbold, 1981
= Conchoraptor Barsbold, 1986
= Rinchenia Barsbold, 1997
= Khaan Clark, Norell & Barsbold, 2001
= Heyuannia Lü, 2002
= Nemegtomaia Lü, Tomida, Azuma, Dong & Li, 2005
Citipati osmolskae Clark, Norell & Barsbold, 2001
= Khaan mckennai Clark, Norell & Barsbold, 2001
Citipati gracilis Barsbold, 1986
= Conchoraptor Barsbold, 1986
= “Ingenia” yanshini Barsbold, 1981
Citipati mongoliensis Barsbold, 1986
= Rinchenia mongoliensis Barsbold, 1986
Citipati huangi Lü, 2002
= Heyuannia huangi Lü, 2002
Citipati barsboldi Lü, Tomida, Azuma, Dong & Li, 2004
= Nemegtomaia barsboldi Lü, Tomida, Azuma, Dong & Li, 2004

So there are a few things to say about this:

First, as Mickey [5] notes, Conchoraptor has priority (by 26 years) over Citipati, so this is Paul ignoring priority. Does he think that only the perceived adult form is valid for taxonomy? If so, why include the species, but not the genus, as valid enough to use? Furthermore, if “Ingenia” yanshini is a synonym of Conchoraptor gracilis, the correct species would be yanshini, not gracilis.

Second, all of these taxa differ from one another in various groupings postcranially, especially in the manus. In oviraptorids, there are actually THREE types of manus organization:

1) where the manus is elongated, greater than the length of the radius by 150%, and md2 and md3 are virtually the same length, with md1 shorter (this is seen in Oviraptor philoceratops and Citipati osmolskae);
2) where md1 is longer relative to md2 or md3, but all digits are equally slender, and the manus is closer to 100% the length of the radius (as seen in Khaan mckennai and Conchoraptor gracilis);
3) where md1 is robust compared to slender md2 and md3, and also the manus is also only 120% of the radius, while digits md2 and md3 are actually shorter than md1 (as seen in “Ingenia” yanshini and Heyuannia huangi).

The manus is unknown in Nemegtomaia barsboldi.

The pelvis is known in most of these taxa, and they also differ from one another in substantial ways, but I’ll get into that some other time.

Thirdly, several of these taxa (Heyuannia, Khaan, “Ingenia,” Conchoraptor) are the same size as one another, and given their overlapping shared material, it is unlikely that they represent the same form (as in “Ingenia” yanshini and Conchoraptor gracilis, with drastically different forelimb, pelvic, and pes morphology) but different age classes in order to account for morphological differentiation. Heyuannia huangi is known from multiple described specimens, and they are all generally consistent, supporting lack of individual variation, so it seems odd to potentially support variation in Paul’s “Citipati gracilis” to support inclusion of “Ingenia” yanshini.

Fourth, looking at the cranial morphology, presence or absence of a crest is irrelevant to most oviraptorid discrimination: Citipati osmolskae can be differentiated on the basis of the slope of the occiput and related orientation of various bones in the skull, whiel Khaan mckennai shares with various other taxa (but not with Citipati osmolskae) the shape of the external naris, the slope of the premaxilla’s anterior margin, and the shape of the various bones of the mandible.I am not going into explicit detail, because much of this is either published (see the original work diagnosing Citipati and Khaan, Clark, Norell & Barsbold, [6]) or isn’t (my personal observations).

This is further interesting as recently Paul “split” several species referred to Iguanodon [8] for reasons that had themselves (as summarized in that paper) been argued as ontogenetic or regional variation. Evolutionarily, oviraptorids appear grouped into a range of grades: a “citipatiine” grade, a “conchoraptorine” grade, and a “ingeniine” grade, and they can be differentiated from one another through a variety of cranial and postcranial comparisons.

Paul is attempting, without justification, to determine what is essentially Oviraptoridae sans Oviraptor and making it all a single “genus,” without attempting to determine the internal composition, or ignoring it. This is then a matter of changing labels around, and applying a Linnaean philosophy to the labels. And it is utterly meaningless. There is also a reasoned debate over whether such an act actually increases or decreases our ability to converse with one another about these taxa, and it tends to favor greater communication of the individual species rather than some ambiguous (or absent) concept of a “genus” being applied to it.

[1] Scannella, J. B. & Horner, J. R. 2010. Torosaurus Marsh, 1891, is Triceratops Marsh, 1889 (Ceratopsidae: Chasmosaurinae): Synonymy and ontogeny. Journal of Vertebrate Paleontology 30(4):1157-1168.
[2] Paul, G. S. 1988. Predatory Dinosaurs of the World. (Simon & Shuster, New York City.)
[3] Paul, G. S. 2010. The Princeton Field Guide to Dinosaurs. (Princeton University Press, Princeton.)
[4] Paul, G. S. 2002. Dinosaurs of the Air: The Evolution and Loss of Flight in Dinosaurs and Birds. (Johns Hopkins University Press, Baltimore.)
[5] Mortimer, M. 2010. GSP’s new taxon combinations from his dinosaur field guide.
[6] Clark, J. M., Norell, M. A. & Barsbold R. 2001. Two new oviraptorids (Theropoda:Oviraptorosauria), upper Cretaceous Djadokhta Formation, Ukhaa Tolgod, Mongolia. Journal of Vertebrate Paleontology 21(2):209-213.
[7] Paul, G. S. 2008. A revised taxonomy of the iguanodont dinosaur genera and species. Cretaceous Research 29(2):192-216.

This entry was posted in Paleontology, Science Reporting, Taxonomy and tagged , . Bookmark the permalink.

9 Responses to What is Citipati gracilis?

  1. Heinrich Mallison says:

    Why am I not surprised?

    Paul appears to work with a load fo preconceived notions. Facts only get in the way of those…..

    on how screwed up his take on prosauropod locomotion is.

  2. You do not disappoint. :) I’ll note that I didn’t consider it problematic to sink Rinchenia into Citipati priority-wise, as I view the 1997 usage as being a nomen nudum, with it only being officially named in 2004.

    And as a minor nitpick, if you’re going with the species citation for “Rinchenia mongoliensis Barsbold, 1986”, then “Nemegtomaia barsboldi Lü, Tomida, Azuma, Dong & Li, 2005” should be 2004. Since barsboldi was named in 2004, albeit as Nemegtia instead of Nemegtomaia at the time.

  3. Mike Taylor says:

    “There is also a reasoned debate over whether such an act actually increases or decreases our ability to converse with one another about these taxa.”

    Is there really any debate over that? It seems perfectly obvious to me that GSP’s actions can only possibly have the effect of muddying future communications. The best case would be tacit agreement between theropod workers to just ignore all these reassignments.

    • In this case, it is about the effect of either combining similar species together under a common header, which can be a “family,” or a “genus,” or splitting them off with “genera” containing similar or formerly “related” species. That there IS a debate among well-reasoning people (e.g., Horner or yourself) suggests there is merit to both sides.

      • Mike Taylor says:

        I can — of course, given my recent publication record — see how there can be debate over a particular case. Hence Horner’s eleven-page argument in favour of sinking Torosaurus into Triceratops, and my own twenty-page argument that Giraffatitan is not congeneric with Brachiosaurus.

        What I can’t see is anyone who thinks the kind of scattergun, minimally justified lumping that Greg’s done here can be a good thing for taxonomic clarity or stability. Even if he happens to be right about some or all of these synonymisations, the fact that he’s not supported them with argument means that they won’t be adopted by anyone who’s actually working on the animals in question, but will get endlessly perpetuated on Internet sites and books written by science popularisers. Recipe for confusion.

        • I agree. My comment was to derive on the very fact that Paul failed to justify his argument in more substantial a way than he does. The statement I make about the debate excludes Paul unless, as in some of his works (on Giraffatitan or Dollodon), he can support it with some form of substantive list of features; and you agree. This is the case in the specific of a failure.

          • As an addendum to this reply, I’d like to note that Pauls’ previous volumes (Predatory Dinosaurs of the World and Dinosaurs of the Air) show more detail in the realm of argumentation when explaining the “why” when lumping taxa — as well as his previous work on splitting, when done — although most of these have lacked any sort of phylogenetic context, which is what is continued here as more classic of Paul’s style than the list of details in which two forms may be similar to differ.

  4. Pingback: Are Scansoriopterygids Oviraptorosaurs? | The Bite Stuff

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s