[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

FW: McDonald on Rubeosaurus [was Re: Last Dinosaur of 2011]

I wrote this message from my laptop where, I realized, the input field is not 
set to plain text. This has been fixed. 2012 glitch or something.

  I thank Andy Farke and Andrew McDonald for taking the time to respond in 
detail on the matter at hand. I want to emphasize my amateurishness on the 
matter, and note that if I gloss something over, I expect to be firmly 
upbraided as a result. Please, do not hesitate to affirm my amateurishness. 

  I'd like to note that I do not actually care what label to apply to the taxon 
as a personal matter. My reasoning is simple: the label is not relevant. 
*"Styracosaurus" ovatus*, *Rubeosaurus ovatus*, etc., represent the same 
essential information, a point I've tried to make in regards to *Brachiosaurus 
brancai* vs. *Giraffatitan brancai* here on the list on in discussions on my 
blog or over at SV-POW!. One is inclined, however, to treat "genera" 
differently from their species, a behavior which works for taxa like 
*Centrosaurus* with multiple included species, but less so for monotaxic 
"genera." The fault here, I think, lies in treating "genera" as special; that 
there is special meaning behind supporting *Rubeosaurus*. This leads to use of 
*Rubeosaurus* when discussing the taxon, rather than *ovatus*, and has little 
to do with splitting or lumping of taxa.  It endorses the fiction that genera 
are more valuable than species, by emphasizing them over the species once 
created: The discussion is not in the value or effect of *ovatus* on the 
analysis, but that of *Rubeosaurus*, as though they are two different entities. 
In the classic Linnaean sense, this is true, but as I like to remind people, 
the Linnaean System is a horribly subjective and unscientific system, and 
endorsement of its ideals, especially that of different "categories" of 
nonspecies names merit different levels of attention, is one of its worse 
descendants in the Sciences.

  Cladistic analysis is somewhat at fault for this, although the practice of 
upholding species into different immediate least-inclusive clades, or "genera," 
did not originate with the use of cladistic analysis. But they are still 
related because it has become common practice to produce an analysis in support 
of this behavior, as though the analysis supports the opinion of whether there 
were 5 or 10 "genera" involved or not. *Rubeosaurus ovatus* is limited, its 
material rendered to a very small, specific, and otherwise considered highly 
diagnostic region of the body. So much weight, in fact, has been left on the 
back of the skull in centrosaurine systematics that it comprises the lions' 
share of all diagnstic information provided for included taxa; after this, the 
remainder of the skull, then the rather systematic exclusion of the postcrania. 
It resulted in Scott Sampson's 1995 analysis coining *Achelousaurus horneri* 
and *Einiosaurus procurvicornis* excluding description of the postcrania from 
the paper he wrote (it is still a good analysis, though). This limitation on 
material results in multiple problems:

1. In prevents us from assessing variation in the diagnostic range of 
characters among individuals, with respect to plasticity of appearance; for 
example, the precise appearance of how "popcorny" the frill gets in 
*Centrosaurus brinkmani* tends to be individual.
2. It prevents us from assessing ontogeny in the diagnostic range of characters 
among individuals, especially with regards to rare specimens; it was the 
originating problem that lead to doubt over the authenticity of *Torosaurus 
latus* as a unique form from *Triceratops* spp.
3. It can warp cladistic analysis, which typically favors more complete 
material than not, except when the complex of included characters are focused 
towards a specific region, in which case the analysis is not approximating a 
phylogeny (a misconception) but rather a likelihood of distribution of 
characters and their organization through projected lineages.

  Let me emphasize this point about influence of portions of material on the 
analysis: The cladistic analysis in McDonald & Horner (2011) consists of 11 
taxa and 27 characters; all of the included OTUs are species-level taxa, though 
most approximate "genus" level taxa, while the characters are all cranial. In 
the latter, the features of the frill (squamosal and parietal) make up over 
half of the characters, 12-27 [19 out of 27, or 70% of the analysis], while the 
parietal alone makes up characters 14-27 [or 51% of the analysis]. This 
emphasizes, I think unfairly, the use of parietal ornamentation as diagnostic 
of not just species, but of segregating species from one another as the 
analysis did. The analysis is small and its components brief. In contrast, 
Farke et al. (2011) used 18 taxa and 97 characters, included some postcrania, 
but noted "[b]ecause cranial characters are the most useful for elucidating 
phylogenetic relationships among these taxa, the matrix focused on this 
subset." [pg.967]; of these, 1-76 are cranial [78%], and of those characters, 
28 were related to the frill [37%, 29% of the whole], and the parietal 
comprised 40, 46-52, 57-68, including the epiossifications [71% of the frill, 
26% of the cranial, 20% of the whole]. These metrics, while better, are still 
weighted towards recovery of particular regions of the skeleton, providing 
greater weight and emphasis on using these regions in further analyses. I have 
little to quantify on the issue of whether the postcrania or other portions of 
the skeleton of centrosaurine ceratopsids are useful for systematics, save that 
the prevailing opinion is that while preparation and examination are useful, 
the postcrania has almost never been strictly quantified in this regards. I 
think it's a mistake. That said, Farke et al. do a better job in supporting 
segregation of *ovatus* from *Styracosaurus* proper than did McDonald & Horner, 
but this merely means that they tested the inclusiveness of the analysis with 
new data; that said, resolution was only gai!
ned in th
f characters, altering over 10 of them related to the parietal; *Rubeosaurus 
ovatus* often falls into a polytomy including *Styracosaurus albertensis* when 
these modifications are not made, which is all well and good in testing the 
analysis; moreover, the practice of enforcing topologies to examine the 
influence of creating slightly less parsimonious outcomes, to examine what 
variations in the character distribution occur, is not always typical of these 
analyses, and I'd like to see a more severe complex analysis on centrosaurs, 
and one which can break taxa down to specimens as OTU's, as this will test the 
robustness of including problematic specimens in the matrix. 

  Cladistic analyses change over time, and the inclusion or exclusion of new 
characters and taxa can and do warp earlier results, enough so that support can 
dwindle for an earlier held proposition, or that the "key" piece in the 
argument of erecting new taxonomy disappears (or conversely, becomes 
strengthened), and little attention is actually paid to this process (the 
cladistic analysis itself), merely its result and the main factors that seem 
relevant (consistency and retention indeces, subordinate Bremer analysis and 
bootstrapping to test the "robusticity," "inclusiveness," or "completeness" of 
the included material, etc.). I am a critic of one thing here, and only one 
thing, and that is the use of limited material to support new taxonomy, and 
have nothing to say on the legitimacy of *Rubeosaurus ovatus* as a distinct 
taxon, just as I have nothing to say on the strength of any other taxon in 
which cladistic analysis is used to support it. My desire is that we use more, 
and better analyses to recover better resolution of problematic specimens and 
apparent species, and find a robust answer, before we produce unique 
nomenclature above or beyond that which already exists. 

  Let me finish this overly long post with the statement that I do not disagree 
with McDonald or Farke on the essential natures of their arguments, that their 
taxa are unique and warrant taxonomy. My interest is in one of removing the 
subjectivity, whenever possible from the situation; in the one case, the 
specimen already had taxonomy produced for it, while in the second it seemed 
useful to have to coin new taxonomy. And honestly, while I would prefer that 
one try to dump new taxonomy into older containers whenever possible, I have no 
problem with forming a unique binomen. However, I'd err on the side of not 
making new taxonomy if I can avoid it, and would prefer this result regardless 
of the analysis. If I produce an analysis which supports novel taxonomy (and I 
have), I would refrain from this practice, simply because it is just one point 
of data, and I would rather it be stronger, more robust, than otherwise. This 
was my essential regard toward the analysis provided in Gates, Horner, Hanna 
and Riley earlier last year (2011) on *Acristavus gagslarsoni* where, using a 
cladistic analysis, the authors eschewed one nearly twice its size to test the 
material at hand and which had already produced an alternative hypothesis, one 
which would surely have influenced whether the authors named the taxon or not 
(http://qilong.wordpress.com/2011/08/09/o-crest-less-one/). To me, this 
enforces the problem of using limited data sets in formulating taxonomy, which 
I think is the case.

  Thank you for reading,

Farke, A. A., Ryan, M. J., Barrett, P. M., Tanke, D. H., Braman, D. R., Loewen, 
M. A. & Graham, M. R. 2011. A new centrosaurine from the Late Cretaceous of 
Alberta, Canada, and the evolution of parietal ornamentation in horned 
dinosaurs. _Acta Palaeontologica Polonica_ 56(4):691-702.
Gates, T. A., Horner, J. R., Hanna, R. R. & Nelson, C. R. 2011. New unadorned 
hadrosaurine hadrosaurid (Dinosauria, Ornithopoda) from the Campanian of North 
America. _Journal of Vertebrate Paleontology_ 31(4):798-811.
McDonald, A. T. & Horner, J. R. 2010. New material of *“Styracosaurus” ovatus* 
from the Two Medicine Formation of Montana. p142-168 in Ryan, Chinnery-Allgeier 
& Eberth (eds.) _New Perspectives on Horned Dinosaurs: the Royal Tyrrell Museum 
Ceratopsian Symposium_. (Indiana University Press, Bloomington.)


  Jaime A. Headden
  The Bite Stuff (site v2)

"Innocent, unbiased observation is a myth." --- P.B. Medawar (1969)

"Ever since man first left his cave and met a stranger with a
different language and a new way of looking at things, the human race
has had a dream: to kill him, so we don't have to learn his language or
his new way of looking at things." --- Zapp Brannigan (Beast With a Billion 

> Date: Mon, 2 Jan 2012 16:13:19 -0800
> Subject: McDonald on Rubeosaurus [was Re: Last Dinosaur of 2011]
> From: afarke@gmail.com
> To: dinosaur@usc.edu
> CC: qi_leong@hotmail.com; michael.osullivan@port.ac.uk
> Andrew McDonald (first author on the paper naming Rubeosaurus) asked
> me to forward the following:
> *********************
> First, a general thought. Neither splitting nor lumping is a valid
> taxonomic philosophy. Taxonomic decisions should be made on a
> case-by-case basis using the evidence at hand. Yes, I split
> *Rubeosaurus ovatus* from *Styracosaurus* and *Uteodon aphanoecetes*
> from *Camptosaurus*, but I've also sunk *Dollodon* and probably
> *Proplanicoxa* into *Mantellisaurus*, and I agree with Norman (2011)
> that *Sellacoxa* is a junior synonym of *Barilium*. Creation of a new
> genus for an existing species or the sinking of a genus into another
> is no reason to be driven up the wall.
> Second, the specifics. Jaime is quite right that centrosaurine
> phylogeny is not very robust at the moment. Referral of USNM 14765 to
> *Rubeosaurus* added a great deal to the cranial codings of that taxon
> (McDonald 2011), but I would not be shocked if the next paper
> describing a new centrosaurine reshuffles things, or, for that matter,
> if it doesn't reshuffle things. Who knows? My reason for separating
> and continuing to separate *ovatus* from *Styracosaurus albertensis*
> is simply that none of the recent phylogenetic analyses (McDonald and
> Horner 2010; McDonald 2011; Farke et al. 2011; Fiorillo and Tykoski in
> press) recovered a sister taxon relationship between *ovatus* and
> *albertensis*; instead, *Rubeosaurus* is a member of the
> 'pachyrhinosaur clade'. In my estimation, keeping *ovatus* in
> *Styracosaurus* implies preference for a phylogenetic hypothesis for
> which there is no support. With its own genus, *ovatus* can appear in
> any position in future analyses without necessitating further
> taxonomic revision. *Rubeosaurus ovatus* lacks P1 processes, whereas
> all *Styracosaurus albertensis* parietals that preserve that region
> exhibit P1 processes of varying prominence. Furthermore, not a single
> *Styracosaurus albertensis* parietal, i.e., none of the isolated
> skulls or bonebed specimens, bears a medially-inclined P3 spike as in
> the holotype of *Rubeosaurus ovatus*. Do you think that some arbitrary
> measure of difference should overrule phylogeny? What are "generic
> level" versus "specific level" differences? If you disagree with the
> splitting of *Rubeosaurus ovatus*, then I think you must present
> evidence that supports a sister taxon relationship between *ovatus*
> and *albertensis*.
> Finally, I want to thank Andy Farke for posting this on my behalf;
> this does not represent an endorsement or refutation of any kind on
> Andy's part. Also, it's never wise to become too complacent in one's
> conclusions, so I thank Jaime and Michael for the interesting
> challenge.
> Best regards,
> Andrew McDonald