RE: Ah ha! That's where therizinosaurs came from

[Jaime Headden <qi_leong@hotmail.com>]

Greg Paul writes:

<The potential link between jeholornids and therizinosaurs is much weaker than 
that between omnivoropterygids and oviraptorosaurs. The reason that jeholornids 
are interesting is because they show that long tailed herbivorous fliers were 
out and about, and were potential ancestors for neoflightless long tailed 
herbivores like basal therizinosaurs.>

  See, it's that matter-of-fact statement there about therizinosaurs being 
neoflightless that makes me think Paul isn't just throwing this one out to 
provoke debate. This is despite me refuting the argument about tailedness, Paul 
continues to use it! So a second time (and I think this is actually the third, 
since the issue was raised about a year or so ago):

  Caudal vertebrae in the four specimens currently described (Ji et al., 2002; 
Zhou and Zhang, 2002; Zhou and Zhang, 2003) vary between 25 and 27, with the 
complete complete sequence (IVPP V13353) counting up to 27. That's at what 
appears to be adult size, coming close to the sizes of two additional specimens 
with one very small IVPP V13350 juvenile, and with a sacrum having six 
vertebrae. Despite this, the caudal vertebrae form a tail sequence that itself 
is 7.4x the length of the ilium, compares to *Archaeopteryx lithographica*'s 
6.3x (but that's using the Solnhofen, which has an incomplete and likely 
truncated caudal series, regardless of whether the length is diagnostic as 
Elzanowski has argued on the matter), and which has fewer (~24) vertebrae. 
*Shenzhouraptor* (which Paul ignores as the effectively correct name) has a 
relatively long tail based solely on vertebral elongation, while the issue is 
ambiguous and likely very different in therizinosaurs.

  Therizinosaurs suffer from lack-of-tail-itis, which is congenital. 
*Alxasaurus elesitaiensis* preserves upwards of 14 caudal vertebrae, with 
hypotheses (Zanno, 2010) implying a low end of 30 and likely more. 
*Beipiaosaurus inexpectatus* has a minimum of 31 caudals (Xu et al., 2003, with 
an additional caudal from the original slab added in), comprised of two 
articulated sequences of 14 and 11, and five floating vertebrae near the ilium; 
the two gaps bear a sizeable gap between them corresponding (based on the size 
of the vertebrae alone) to a measurably large sequence of about 10 or so 
vertebrae, of which five are present scattered on the slab. I would thus infer 
a low end of 36 [Zanno, 2010 infers 35, but states the count as 30], despite 
their length, with a possible vertebra adding to the sequence at the tip (which 
may be incomplete); the methodology of Grillo & Azevedo (2011) may actually 
produce an accurate number. 

  There is a high level of ambiguity in the sequence of caudal vertebrae 
present in the Crystal Geyser Quarry (n=67) that produced the holotype of 
*Falcarius utahensis*, and the actual count is not inferred, largely due [I'm 
sure] to age variation in the quarry. Taking into account the aspect of distal 
caudals in the Quarry, which become elongated, it should be apparent that this 
is true in *Gallimimus bullatus*, *Tyrannosaurus rex*, *Allosaurus fragilis* 
and any number of other tetanurans for which this morphology is diagnostic. It 
is, in short, plesiomorphic. What Paul is trying to argue is that tail _length_ 
matters on its face, rather than take into account the quanta of caudal count 
or even morphology.

<Jeholornids make poor ancestors for therizinosaurs in part because they lacked 
tooth rows (which are unlikey to have reevolved although it cannot be ruled 
out).>

  *Shenzhouraptor sinensis* lacks tooth rows? When was _this_ published? The 
mandible does actually preserve teeth in IVPP V13274, and their absence is 
arguable in the other specimens (especially the holotype: Ji et al., 2002, who 
concede the poor preservation up front). And yes, they make extremely poor 
ancestors, which is why we're all scratching our heads when Paul pronounces 
therizinosauroids are neoflightless.

<Nor could the omnivoropterygids we know and love be actual ancestors for 
oviraptorosaurs because the latter were already extant.>

  I've heard this argument before. "How could birds come from dinosaurs if 
there are still dinosaurs?" Or "How could humans evolve from apes when there 
are still apes?" Certainly not both arguments from YOU, sir, but they are 
essentially the same, and essentially equally poorly conceived and illogical 
when considering the data involved (with which Greg Paul is aware of, having 
cited it in _DotA_). The response is always the same, as you should be well 
aware of, and the subject is in the TalkOrigins archive 
(http://www.talkorigins.org/faqs/archaeopteryx/challenge.html), but there's a 
handy-dandy reference at 
http://www.askabiologist.org.uk/answers/viewtopic.php?id=2491

  I know you're not joking on a basic premise of the problem of the 
superposition of the geological colum in regards to recovery of _known_ 
fossils, because you continue in all seriousness:

<At best there was a common ancestor whose morphology would be more 
transitional than the fossils we got. If those creatures existed hopefully they 
will show up in the earlier sediments present in NE China or elsewhere. One can 
hope -- and it basically worked for deinonychosaurs so its not a long shot.>

  Tongue-in-cheek this is not. You transition through several phylogenetic 
inferences in this section, and I baffles me how this argument, made way back 
in _PDW_ could continue to be made?

  But on this specific further point:

  Except we are only now finding sister taxa to the nodes where troodontids and 
dromaeosaurs split, and it turns out it's a lot more complex than you even 
thought of, and still plastic (you know, prone to constant shifting). For 
instance, the uncertainty results in us treating *Microraptoria* (NOT 
*Microraptor zoui* alone) as a clade of basal dromaeosaurs, but so too 
*Unenlagiidae/Unenlagiinae* [I'd prefer a rank-neutral variant name]. By 
constrast, troodontids show less defined variation, but it was there, too. 
These groups are more complex, and they keep moving around, sometimes shifting 
taxa to the other side and sometimes forcing them apar with respect to 
*Avialae* and sometimes taxa move into the *Avialae*. To say that it "worked 
for deinonychosaurs" is to assume we actually HAVE the correct picture, or a 
semblance of it. How long did we think segnosaurs weren't theropods before 
cladistic analysis, or the descriptive analysis of the skull of *Erlikosaurus 
andrewsi* (Perle et al.)?
 
<Someone said something about it not being possible to tell if a flightless 
tetrapod with flight related features is neoflightless or not.>

  That was me! Yay! Maybe now if what I'm about to say on this subject, like my 
repeating about therizinosaur tails, were to be spoken by someone that Paul 
likes, he'll not denegrate that individual and name him/her as he named Jason 
earlier in his email.

<We know that all current flightless birds are neoflightless.>

  Bullhockey. We have basal taxa to the entirety of avian-ness (*Neornithes*) 
which show flight retained based on our assumptions on the anatomy involved. 
*Ichthyornis dispar*? Once inferred to be basal to terns, flighted due to 
environment and the high disparity of leg/wing length, and extreme development 
of the forelimb and pectoral girdle. *Hesperornis regalis*? Trickier, once 
inferred to be a loon, or grebe, or with both; but also because it is 
phylogenetics that infers flight at the base (phylogenetic bracketing_; basal 
taxa do not appear to be flighted, but we assume it so because of sister-taxa. 
And so on. When, however, we arrive at the base of *Avialae,* we find taxa 
lacking pygostyles in general and with long caudal counts, small sacral counts 
and very, very short ilia, small pectoral girdles, no keels or no ossification, 
no alulae, unreversed halluces with few adaptatations for perching or with 
often terrestrial indicators in phalangeal/ungual morphology/proportions, 
retained third manual digits and fingers often with large claws, 
feather-strength indicators (although that's been questioned), broad and now 
narrow furcula, two completely different forms of arrangement for a triosseal 
canal indicating it is convergent, and a questionable gradient on expansion and 
orientation of the scapular glenoid.

  What is the problem here? Paul infers that the process for losing flight 
shows the _same_ features as for gaining flight, based on the phylogenies 
produced by others, which he then uses his "method" to contradict by arguing 
that the entire system is convergent across multiple lineages. Thus, when he 
maps the above features to his anti-phylogeny, he can arrive at loss-of-flight 
characteristics. This becomes the primary hypothesis. But how to test it ... I 
know, look at birds! Wait, isn't that circular reasoning? That's why we ask 
that such an argument find a test to the loss of flight. You have to find a way 
to discriminate a developing flight condition from a developing loss of flight 
condition, and when Paul writes:

<The loss of flight in early fliers may have been as easy as pie,>

  He makes an argument that _requires_ some method to assess, and offers none. 
What we have is the inference of a single gain in flight, without any actual 
data to show otherwise. Things seem to get busy, but animals get less-leggy, 
less-feathery, larger and ever so larger from a basal small seeming scansorial 
gradient. Even *Ornithomimosauria* seems to arise from this, along with 
*Tyrannosauroidea*, as they all stem from a lineage that looked an aweful lot 
like *Sinosauropteryx prima*.

<Someone said I am inconsistent in rejecting cladograms that contradict the 
neoflightless hypothesis while accepting those that do. But this is illogical. 
Say the cladograms for years contradict the neoflightless hypothesis and I pay 
them little mind because it is my conclusion that they are defecitve due to 
lack of sufficient fossil data and because the nonneoflightless hypothesis>

  This was also me, but I think Tim Williams also said it.

  There is no "nonneoflightless hypothesis," it is the null. This is what all 
other arguments apart from _DotA_ are: "Ground up," "Trees down," "WAIR" and 
"the Hopp-Orsen hypothesis" are hypotheses for developing flight; your argument 
is to pretend flight came first, and this simply differs from the initial 
hypothesis that flight evolved in birds once. Your argument differs from 
Huxley, Marsh, Ostrom, Gauthier, and even Feduccia and the other BANDits, etc., 
in which progressive birdiness developed on a lineage resulting in 
more-bird-like taxa, rather than several less-bird-like taxa. You reverse the 
clinal trope from what is assumed to be the null, no one does the opposite.

<does not appear as logical. Say that as new fossils come online some 
cladograms start supporting the neoflightless hypothesis. Am I supposed to 
automatically say that since cladistics is offering support for the 
neoflightless hypothesis I must now reject the latter? I do not think 
cladistics is useless. My concern is that cladistics has become overly dominant 
to the point it prevents thinking outside the box of the cladistic results at 
any given time.>

  Cladistic output improves with more data. It is a separate inference that 
newer fossils will improve the analysis, as they may result in branches that 
have nothing to do with major lineages and thus seldom shake the tree. A new 
oviraptorid described tomorrow will not see us with a novel oviraptorosaurian 
position or array of the basal forms; but finding new basal taxa does cause the 
tree to adjust, and may (as in *Xiaotingia zhengi*) pull otherwise "stable" 
taxa with it. Despite this, such trees are looked on with caution, while the 
stable trees which hold despite other inclusions of such data are given higher 
levels of confidence due to integrity of data measures and mean output of 
similar trees with different data sets. This is akin to five different groups 
using different sets of data to arrive at the same conclusion; when a sixth 
group uses another set of data and arrives at a divergent conclusion, we trust 
that analysis a lot less.

<Here's the thing. Are those who oppose my opposition to over reliance of 
cladistics really telling me I should be a cladist?>

  There is no opposition to reliance on the cladistic methodology here. Paul 
rejects the methodology, and does not make a case for occassional use of the 
processes involved. He eschews it, and the results, in favor of his own 
methods, inferring that his are more useful or competitive, despite not showing 
a way to distinguish the processes. When cladistic methodology recovers a 
topology that conforms with some of his previous-published arguments, it is 
declared he is right (again) and Paul crows eventual victory. These methods are 
not competitive; Paul lacks a methodology entire, merely a phyletic lumping 
process which is subjective to the extreme, while the cladistic methodology 
uses parsimony as its discriminating mechanism. Regardless of whether we 
_should_ be using parsimony and mathematic processes to recover phylogenetic 
topologies and affirm or deny hypotheses of descent, cladistic methodology can 
be tested using known processes and math, while Paul's cannot. Further, 
confirmation biases in data selection can be fixed in various points of the 
data matrix used, including taxon selection, specific coding of sequences, and 
amount of sequence to code. Are such biases apparent when Paul's "scientific" 
argument avers "therizinosaurs are neoflightless"? I say yes.

  This makes the cladistic methodology _scientific_, and Paul's "method" 
decidedly not so.

  No one is asking Paul to become a cladist; we are saying he should have some 
faculative understanding of the system he disregards before calling it or its 
results into question. It is a tool, like any else.

Clark, J. M., Perle A., & Norell, M. A. 1994. The skull of *Erlicosaurus [sic] 
andrewsi*, a Late Cretaceous "segnosaur" (Theropoda: Therizinosauridae) from 
Mongolia. _American Museum Novitates_ 3115:1-39.
Grillo, O. N. & Azevedo, S. A. K. 2011. Recovering missing data: Estimating 
position and size of caudal vertebrae in *Staurikosaurus pricei* Colbert, 1970. 
_Anais da Academia Brasileira de Ciências_ 83(1):61-71. 
[http://www.scielo.br/pdf/aabc/v83n1/aop0311.pdf]
Ji Q., Ji S.-a., You H.-l., Zhang J.-p., Yuan C.-q., Ji X.-x., Li J.-l. & Li 
Y.-x. 2002. [Discovery of an avialae bird - Shenzhouraptor sinensis gen. et sp. 
nov - from China.]. _Geological Bulletin of China_ 21(7):363-369 [in Chinese]
Xu X., Cheng T.-n., Wang X.-l. & Chang C.-s. 2003. Pygostyle-like structure 
from *Beipiaosaurus* (Theropoda, Therizinosauroidea) from the Lower Cretaceous 
Yixian Formation of Liaoning, China. _Acta Geologica Sinica_ 77(3):294-298.
Zhou Z.-h. & Zhang F.-h. 2002. A long-tailed, seed-eating bird from the Early 
Cretaceous of China. _Nature_ 418:405–409.

Cheers,

  Jaime A. Headden
  The Bite Stuff (site v2)
  http://qilong.wordpress.com/

"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 
Backs)