Friday, February 19, 2010

Citipati or IGM 100/42?

Another mini-post.  Everyone's familiar with IGM 100/42, that oviraptorid that was the classic example of its family through the 80's and 90's.  It's known from an almost complete skeleton first mentioned by Barsbold in 1981.

Skull of IGM 100/42, courtesy of Auditore.

Everyone's also familiar with Citipati osmolskae, the nearly complete specimen that was named in 2001 and whose skull was described by Clark et al. (2002).  The most famous nesting oviraptorid Big Mama (Clark et al., 1999) and the stil undescribed Big Auntie have been assigned to this species as well, as has the most famous oviraptorid embryo (Norell et al., 2001).

Citipati osmolskae holotype (copyright AMNH)

Now these specimens are similar, and may indeed be congeneric, but this hasn't been demonstrated yet.  While Osmolska et al. (2004) found IGM 100/.42 to be the sister taxon of Citipati osmolskae in their cladistic analysis, Lu (2004) found that it was the sister taxon to Oviraptor or Conchoraptor (depending on taxa included) using Maryanska et al.'s (2002) characters.  The point of this post is that some phylogenetic analyses have been quite sloppy in regard to exactly which oviraptorid they are coding.  The first phylogenetic analysis to include either was the TWG analysis, which started with Norell et al. (2001).  They called it Oviraptorid IGM 100/42.  Note the progression of what this OTU is labeled as the TGW analysis is modified in future works-

Norell et al., 2001 (Ostrom volume) Oviraptorid IGM 100/42
Clark et al., 2002 (Mesozoic Birds) Oviraptorid IGM 100/42
Hwang et al., 2002 (Microraptor) IGM 100/42
Xu et al., 2002 (Incisivosaurus) IGM 100/42
Xu et al., 2002 (Sinovenator) IGM 100/42
Makovicky et al., 2003 (Byronosaurus) Citipati osmolskae
Hwang et al., 2004 (Huaxiagnathus) Citipati osmolskae
Lu, 2004 (thesis) Citipati osmolskae
Makovicky et al.; Makovicky and Norell; Norell and Makovicky, 2004 (Dinosauria 2) Citipati osmolskae
Xu and Norell, 2004 (Mei) Citipati osmolskae
Kirkland et al., 2005 (Falcarius) Citipati
Makovicky et al., 2005 (Buitreraptor) Citipati osmolskae
Mayr et al., 2005 (Archaeopteryx) IGM 100/42
Novas and Pol, 2005 (Neuquenraptor) Citipati osmolskae
Gohlich and Chiappe, 2006 (Juravenator) Citipati
Norell et al., 2006 (Tsaagan) Citipati osmolskae
Hwang, 2007 (enamel thesis) Citipati
Martinelli and Vera, 2007 (Achillesaurus) Citipati osmolskae
Senter, 2007 (coelurosaurs) Citipati osmolskae* (see below)
Turner et al., 2007 (Shanag) Citipati osmolskae
Turner et al., 2007 (Mahakala) Citipati osmolskae
Xu et al., 2007 (Gigantoraptor) Citipati osmolskae
Novas et al., 2008 (Orkoraptor) Citipati osmolskae
Xu et al., 2008 (Anchiornis) Citipati osmolskae
Zhang et al., 2008 (Epidexipteryx) Citipati osmolskae* (see below)
Makovicky et al., 2009 (Beishanlong, Xiongguanlong) Citipati osmolskae
Zanno et al., 2009 (Nothronychus graffami) Citipati osmolskae
Zheng et al., 2009 (Tianyuraptor) Citipati osmolskae* (see below)

You can see it changed inexplicably in 2003.  For unknown reasons, it was dropped from the Graciliraptor and Pedopenna versions, which I assume to mean it's absent in Xu's (2002) thesis.  You might say this kind of relabeling is harmless, but I find it misleading.  If someone didn't follow the sequence of permutations of the TWG matrix, they could think the codings were based on specimens described in the literature.  After all, Citipati's holotype is a complete skeleton, though it remains undescribed except for the furcula.  Lu (2004) referred to it as Citipati osmolskae in his revisions of the TWG matrix, but coded IGM 100/42 and Citipati separately for his versions of Maryanska et al.'s matrix.  So one of his Citipati's was not comparable to his other Citipati.  Even worse, Senter (2007) used the IGM 100/42 codings as Citipati osmolskae, then coded IGM 100/42 as a separate taxon!  So really his matrix has two copies of IGM 100/42.  Oddly, there are coding differences and they are not sister taxa.  Zhang et al. (2008) copied this in their Epidexipteryx analysis, as did Zheng et al. (2009) in their Tianyuraptor analysis.  So this problem can worm its way through the literature unnoticed.

In conclusion, if anyone is going to use the TWG matrix (and I know you are), please label your OTU IGM 100/42.  Or else defend the assertion IGM 100/42 belongs to Citipati osmolskae.  Taxonomy via data matrix should be avoided.

Monday, February 15, 2010

Pterosaur Relationships

Pterosaurs on a theropod blog?!  What has the world come to?  Well, a consequence of having a blog based on material for my website and projects is that when I'm too busy to work on them, the post count lightens.  But I had some thoughts based on a recent DML message that's been repeated numerous times in the past.  It concerns the sister taxon to Pterosauria.  In the literature, we find two hypotheses- David Peters thinks they're related to protorosaurs within Lepidosauromorpha; everyone else thinks they're related to dinosauromorphs within Archosauromorpha.  Now it would be easy enough to just say Peters is a fringe researcher and that we should go with the professional consensus view.  But I don't believe that's justified here.  Certainly Peters has methodological problems.  His Photoshop method has been shown to fail multiple times (Bennett, 2005) and I don't believe half the things he sees are real.  Plus he doesn't understand how to perform a phylogenetic analysis correctly in multiple respects.  So do I trust his analyses?  Absolutely not.  But does that mean he's wrong?  No.  The ironic thing, and the point of this post, is that the other side is almost as bad.  Oh sure they (probably) make less anatomical errors and have a better understanding of correlation and ordering for characters, but their matrices are flawed in other ways.  Bennett (1996) had Lepidosauromorpha and Prolacertiformes (now recognized as polyphyletic) as OTUs, and the former was the outgroup.  Sereno's (1990, 1991) "analyses" never include conflicting data so are demonstrations of hypotheses instead of tests.  Benton (1999) only included archosauriforms.  Hone and Bennett's (2008) supermatrix method oddly included both lower level taxa (e.g. Squamates, Champsosaurus) AND the higher level taxa they belong to (e.g. Lepidosauromorpha, Choristodera), but even stranger these taxa did not group with themselves, showing the matrix is highly flawed.  One obvious way is actually the exact thing I complained about earlier on this blog- many taxa are only coded for a small percentage of characters.  What good are 324 characters if Dinosauria is only coded for 90 of them, or Lepidosauromorpha for 78 of them?  Thus no published study is a useful test of pterosaur origins and they might as well be left as Sauria incertae sedis until someone takes the time to do an analysis properly.

Wednesday, February 10, 2010

Avialan alvarezsaurs I: Cranial Characters

In this series of posts examining the evidence for placing alvarezsaurs in various parts of the theropod tree, I decided to break up the bird part due to the huge number of characters suggested.  Various analyses have suggested alvarezsaurids are members of Avialae, here used in the sense of closer to birds than Deinonychus.  In addition, some of these have placed alvarezsaurids closer to modern birds than Archaeopteryx, which means they fall into Ornithurae sensu Gauthier.  This alvarezsaurid-ornithothoracine clade was usually called Metornithes, but that name is valid for the smallest group containing Aves and Mononykus, regardless of where the latter falls.  So if alvarezsaurids are arctometatarsalians, Metornithes is a senior synonym of Maniraptoriformes.  If alvarezsaurids are basal maniraptorans, Metornithes is a subset of Maniraptora.  Thus the term Metornithes will not be used in the discussion below.  One problem with these early phylogenies is that many basal avialans were not yet known, except Archaeopteryx, ornithothoracines and in 1995 and later, confuciusornithids (not actually included in analyses until 2000).  Thus exactly where alvarezsaurids would fall in comparison to taxa like scansoriopterygids, Shenzhouraptor or omnivoropterygids is not obvious.  The following is a list of every cranial character found to support placing alvarezsaurids in Avialae (first 18 characters) and its subgroups (characters 19-26), based on every analysis to support this arrangement.  These were (not including later variations)- Perle et al., 1993; Chiappe et al., 1996; Novas, 1996; Chiappe et al., 1998; Forster et al., 1998; Chatterjee, 1999; Holtz, 2000; Chiappe, 2001; Maryanska et al., 2002.  Most found them to be ornithurines, though the last two found them to be outside the Archaeopteryx+Aves clade.

1. Premaxilla long and pointed with long nasal process (Holtz, 2000). A composite of three characters, none of which are in TWG and only one of which is in Choiniere et al. (premaxilla angled <70 degrees). However, it is composited there with the proportion of the premaxilla which is anterior to the external nares so is difficult to evaluate the codings for. In any case, Shuvuuia has a very short premaxilla which is blunt and has a short nasal process (Sereno, 2001), unlike most birds. It was thus miscoded by Holtz. Haplocheirus is similar.

2. Premaxillary teeth conical (Holtz, 2000). In both TWG and Choiniere et al.. Miscoded- Allosaurus, Giganotosaurus, Incisivosaurus, Troodon, Sinornithosaurus, Saurornitholestes, Tsaagan, Velociraptor, Deinonychus, Utahraptor. Not coded- Daspletosaurus, Sinovenator. Premaxillary teeth are unknown in Shuvuuia (Chiappe et al., 2002), so Holtz miscoded it. It turns out that Haplocheirus has the character though, potentially supporting avialan alvarezsaurs. However, the character is not unique to avialans. Basal troodontids and oviraptorosaurs also have it (the condition in therizinosaurs is unknown), so this could be congruent with a basal maniraptoran placement as well.

3. Nasal shorter than frontal (Maryanska et al., 2002). Not in TWG, but in Choiniere et al. as part of a character coding for rostrum length and composited with maxilla length compared to premaxilla length. Again, the compositing means I won't be evaluating its accuracy, but again this was miscoded in Shuvuuia which has longer nasals than frontals (Sereno, 2001). Haplocheirus is similar.

4. Jugal shallow or rod-like under orbit (Maryanska et al., 2002). Not in TWG, but in Choiniere et al. when specified to define the cross section, not necessarily the depth. Miscoded- Garudimimus, Rahonavis. Not coded- Acrocanthosaurus, Carcharodontosaurus, Dromiceiomimus, Byronosaurus, Saurornithoides, Buitreraptor, Velociraptor, Shenzhouraptor, Sapeornis, Confuciusornis, Yixianornis. Haplocheirus lacks this, so does not support avialan alvarezsaurs.

5. Frontals anteriorly triangular (Forster et al., 1998). In both TWG and Choiniere et al.. This is not true in Shuvuuia, where the nasofrontal suture is W-shaped (Sereno, 2001), or Haplocheirus where it is transverse. It was thus miscoded by Forster et al. and its accuracy in Choiniere et al.'s matrix is not examined here. It is present in the derived Ceratonykus though.

6. Squamosal-quadratojugal contact absent (Chiappe et al., 1996). In both TWG and Choiniere et al.. Miscoded- Garudimimus, Deinonychus, Archaeopteryx. Not coded- Daspletosaurus, Harpymimus, Oviraptor, Conchoraptor, "Ingenia", Mei, Saurornitholestes, Dromaeosaurus, Scansoriopteryx, Epidexipteryx, Sapeornis, Confuciusornis. As no avialans were coded, except Archaeopteryx which was miscoded as having the contact, it couldn't test the homology of Shuvuuia's condition. Absent in Haplocheirus, so does not support avialan alvarezsaurs.

7. Quadratojugal joined to quadrate by ligament (Chiappe et al., 1998). Not in TWG or Choiniere et al.. Not reported in Haplocheirus, so potentially supports avialan alvarezsaurs. However, the distribution of this character is largely uncertain since many maniraptoriforms have at least loosely sutured articulations which make determination difficult or ambiguous.

8. Parietals and laterosphenoids unfused (Forster et al., 1998). Not in TWG or Choiniere et al.. Forster et al. miscoded basically every taxon, as their Ornithurae has fusion and their non-avialan taxa lack it (e.g. tyrannosaurids, Compsognathus, ornithomimids, oviraptorids). The condition in Shuvuuia is unreported otherwise, though even if I trusted Forster et al.'s coding it would be compatible with any topology where they are outside Ornithurae sensu Chiappe (Hesperornis and more derived birds). The condition in Haplocheirus is unknown.

9. Posterior tympanic recess opens in otic recess (Chiappe et al., 1996). In both TWG and Choiniere et al.. Miscoded- Tyrannosaurus, Sinovenator, Byronosaurus, Saurornithoides, Troodon. Not coded- Acrocanthosaurus, Gorgosaurus, Pelecanimimus, Chirostenotes, "Ingenia". The condition in Haplocheirus is unknown, so potentially supports avialan alvarezsaurs. However, it is also present in Pelecanimimus, Falcarius and troodontids, meaning it could be congruent with other hypotheses as well.

10. Occiput directed ventrally (Holtz, 2000). Not in TWG or Choiniere et al.. This seems true in Haplocheirus, so potentially supports avialan alvarezsaurs.

11. Foramen magnum taller than wide (Holtz, 2000). In both TWG and Choiniere et al.. However, Holtz miscoded Alvarezsauridae since Shuvuuia does not have this character (Chiappe et al., 1998). Haplocheirus also lacks it and its accuracy in Choiniere et al.'s matrix is not examined here.

12. Neck of occipital condyle not constricted (Holtz, 2000). In both TWG and Choiniere et al.. Miscoded- Sinraptor, Dilong, Tyrannosaurus, Avimimus, Troodon, Dromaeosaurus. Not coded- Acrocanthosaurus, Carcharodontosaurus, Giganotosaurus, Daspletosaurus, Struthiomimus, "Ingenia", Microraptor. This is true in Haplocheirus, so potentially supports avialan alvarezsaurs. However, it is also present in most maniraptoriforms except some oviraptorosaurs, so was miscoded by Holtz.

13. Triradiate palatine (Chiappe et al., 1998). In both TWG and Choiniere et al.. Miscoded- Shenzhousaurus, Archaeopteryx. Not coded- Neovenator, Gorgosaurus, Daspletosaurus, Juravenator, Chirostenotes, Sinornithosaurus, Dromaeosaurus, Sapeornis. Absent in Haplocheirus, so does not support avialan alvarezsaurs.

14. Splenial not widely exposed laterally (Holtz, 2000). In both TWG and Choiniere et al.. Miscoded- Compsognathus. Not coded- Daspletosaurus, Juravenator, Pelecanimimus, Caudipteryx, Buitreraptor, Saurornitholestes, Shenzhouraptor. This is true in Haplocheirus, so potentially supports avialan alvarezsaurs. Yet this is also found in almost every theropod except most deinonychosaurs so was only an avialan character in Holz's analysis due to his lack of deinonychosaur monophyly.

15. Coronoid absent (Chiappe et al., 1998). In both TWG and Choiniere et al.. Miscoded- Shuvuuia, Citipati, Scansoriopteryx. Not coded- Acrocanthosaurus, Daspletosaurus, Compsognathus, Segnosaurus, Caudipteryx, Conchoraptor, "Ingenia", Shenzhouraptor, Sapeornis. Since Shuvuuia was miscoded, this character could not function to support avialan alvarezsaurs. The condition in Haplocheirus is unknown, so potentially supports this relationship. However, ornithomimosaurs and therizinosaurs show this as well.

16. Elongate and tapering retroarticular process (Holtz, 2000). In TWG, but not Choiniere et al.. Both Shuvuuia and Haplocheirus have this, but Holtz miscoded Archaeopteryx as having it as well. In actuality, almost all avialans lack it, so it does not support this relationship.

17. Unserrated teeth (Chiappe et al., 1996). In both TWG and Choiniere et al.. Miscoded- Compsognathus, Avimimus, Sinovenator, Microvenator, Adasaurus. Not coded- Daspletosaurus, Alxasaurus, IGM 100/44, Scansoriopteryx, Shenzhouraptor. Oddly, only some toothless taxa are coded as inapplicable. Absent in Haplocheirus, so does not support avialan alvarezsaurs.

18. Teeth uncompressed and unrecurved (Forster et al., 1998). Not in TGW, but in Choiniere as the same composite of characters, with an additional state added compositing crown elongation as well. Miscoded- Shenzhousaurus, Harpymimus, Anserimimus, Buitreraptor, Sinornithosaurus, Microraptor, Archaeopteryx. Not coded- Acrocanthosaurus, Carcharodontosaurus, Giganotosaurus, Mapusaurus, Beipiaosaurus, IGM 100/44, Sinornithoides, Adasaurus, Utahraptor, Scansoriopteryx, Yixianornis. Absent in Haplocheirus, so does not support avialan alvarezsaurs.

Skull of Haplocheirus (modified from Choiniere et al., 2010) showing features which agree (green) and disagree (red) with placing alvarezsaurs as birds in various analyses.  Numbers match character numbers in this post.

19. Maxillary fenestra absent (Perle et al., 1993). In both TWG and Choiniere et al.. This was based on the partial maxilla of Mononykus, found before complete skulls of Shuvuuia showed a fenestra in a different position. Thus Mononykus may have had one as well, and Choiniere et al.'s codings are not evaluated for accuracy, though a glance shows taxa left uncoded like Allosaurus and Archaeopteryx, which obviously have the fenestra. Haplocheirus has one as well.

20. Slot in ventral process of lacrimal for jugal absent (Holtz, 2000). Not in TWG or Choiniere et al.. This has never been explicitly described in Shuvuuia, though Sereno (1999) agreed with Holtz's coding and Holtz et al. (2004) later switched Alvarezsauridae's coding. Thus the coding for Alvarezsauridae is controversial. In any case, examination shows basically all maniraptoriforms lack it, contra Holtz's original coding, but agreeing with the 2004 coding. Haplocheirus' condition is unknown, but the character is ignored as it has no affect on alvarezsaurs' position given current data.

21. Postorbital-jugal contact absent (Chiappe et al., 1996). Not in TWG or Choiniere et al.. Absent in Haplocheirus, so does not support ornithurine alvarezsaurs.

22. Dorsal jugal process absent (Forster et al., 1998). Not in TWG or Choiniere et al.. Absent in Haplocheirus, so does not support ornithurine alvarezsaurs.

23. Quadrate articulates with braincase (Novas, 1996). Not in TWG or Choiniere et al. except as a corollary of a character coding for a divided quadrate head. Unknown in Haplocheirus so potentially supports ornithurine alvarezsaurs. This character is more widespread than usually given credit though as even most tyrannosaurids have a surface extending distally from the quadrate head that articulates with the braincase. Even ignoring this primitive condition, at least oviraptorosaurs, troodontids and Mahakala also have the condition among maniraptoriforms, though eudromaeosaurs lack it.

24. Separate otic head of quadrate (Forster et al., 1998). In both TWG and Choiniere et al.. Miscoded- Deinonychus, Yixianornis. Not coded- Acrocanthosaurus, Giganotosaurus, Mapusaurus, Gorgosaurus, Daspletosaurus, Compsognathus, Harpymimus, Garudimimus, IGM 100/44, Scansoriopteryx, Sapeornis, Apsaravis. Unknown in Haplocheirus so potentially supports ornithurine alvarezsaurs. However, among birds only Confuciusornis, Enaliornis and some neoavians have well separated heads like Shuvuuia, so convergence is most likely by far.

25. Parietal sagittal crest absent (Forster et al., 1998). In both TWG and Choiniere et al.. However, Choiniere et al. have two sagittal crest characters, one coding for height and one for axial length. This unfairly weights the presence of a crest (since both characters have a state for 'crest absent'). To mitigate this, all taxa coded as lacking a crest in one of the characters are changed to unknown in my recoding. Miscoded- Dilong, Ornitholestes, Citipati, Troodon, Tsaagan. Not coded- Allosaurus, Acrocanthosaurus, Carcharodontosaurus, Giganotosaurus, Gorgosaurus, Daspletosaurus, Juravenator, Pelecanimimus, Shenzhousaurus, Harpymimus, Garudimimus (oddly coded as inapplicable), Dromiceiomimus, Struthiomimus (also coded inappicable), Erlikosaurus, Incisivosaurus, Caudipteryx, Avimimus, Oviraptor, Conchoraptor, "Ingenia", Sinovenator, Saurornitholestes, Velociraptor, Deinonychus, Scansoriopteryx, Archaeopteryx, Sapeornis, Confuciusornis. Unknown in Haplocheirus so potentially supports ornithurine alvarezsaurs. Yet this is also found in ornithomimosaurs, Erlikosaurus and basal troodontids among maniraptoriforms.

26. Dentary teeth set in groove (Chiappe et al., 1996). In both TWG and Choiniere et al.. I already covered this character in the ornithomimosaur post. Unknown in Haplocheirus so potentially supports ornithurine alvarezsaurs. However, it is also present in Pelecanimimus and troodontids, and more importantly is only found in a few derived ornithuromorphs among birds.

As before, a summary of the codings for relevent taxa.  They are divided into supposed avialan (1-18) and supposed ornithurine (19-26) characters.

                          111111111 12222222

                 123456789012345678 90123456

Haplocheirus     010000???10101?100 0?00????
Tugrik           ????????????1????? ????????
Shuvuuia         0?01011?1101111111 0?111111
Ceratonykus      ???111???1???1?1?? ??1111??
Mononykus        ????????1???????11 ????1???

Ornithomimosaur  00001001p001011010 0100?01p

Scansoriopteryx  ????01???????11011 ??00?01?
Epidexipteryx    010??1?????????11p ??10??0?
Archaeopteryx    11011111111101101p 00001010
Shenzhouraptor   ???1?????????1101? 0?00???0
Dalianraptor     0????????????1???? ????????
Jixiangornis     0???????????????1? ???0???-
Yandangornis     1-??????????????-- ???????-
Zhongjianornis   1-01????????????-- ???1???-
Sapeornis        010001??????11101? 0?0010?-
Confuciusornith  1-11011????1?1?0-- 0?00111-
Ornithothoraces  110p011011p1p1p010 0-p010p0

Of the 26 suggested cranial characters, only nine hold up as even potentially supporting alvarezsaurids as birds.  Of these, several (2, 12, 14, 23) are present basally in Maniraptora.  Of the others, several are also found in ornithomimosaurs (9, 15, 25), therizinosaurs (9, 15, 25) or troodontids (9, 25).  Of the remaining, character 7 is hard to unambiguously score for most maniraptoriforms as noted above, leading only character 10 (occiput directed ventrally) as strong evidence for the relationship.  Of the nine potentially valid characters, Choiniere et al.'s analysis includes six.  Yet they did not code alvarezsaurids for one of these, making the total functionally five.  The effect of recoding in Choiniere et al.'s matrix will be examined at the end of this series.

Thursday, February 4, 2010

The First Asian Alvarezsaur

... ever found, that is.  The avialan alvarezsaur post is taking a while since there have been 95 proposed characters supporting it, and Choiniere et al. include more of them.  In the mean time, now that I have a blog, I get to do filler posts pictures of the day.  So here's AMNH 6524, the first alvarezsaurid discovered, back in 1923.  Back then it was only identified as a bird-like dinosaur. It was not until 1993 that it was identified as an alvarezsaurid (Norell et al., 1993). Originally identified as Mononykus, it is more likely Shuvuuia or Parvicursor as it is from the Djadokhta Formation.  It's mostly too damaged to see useful details in, but in the upper right you can see the parallel pubis and ischium with broad contact, a parvicursorine synapomorphy.

Norell, Chiappe and Clark, 1993. New limb on the avian family tree. Natural History. 9/93, 38-43.

Monday, February 1, 2010

A plea to theropod workers- Code the Taxa in Your Analyses

In the middle of writing the next installment of my alvarezsaur relationships series, I keep coming across a particular problem that's important enough to deserve its own post.  I've often complained about workers leaving out conflicting characters (e.g. Sereno, 1999) or relevent taxa (e.g. Maryanska et al., 2002), but recently I've noticed a more insidious problem.  Somebody will include a character and plenty of taxa, but then only code the character for a few of those taxa.  Oh sure, I don't expect everyone to track down every obscure coding available, so I'd forgive you if you didn't code Avimimus as having a fused sternum or Adasaurus as having an unfused one.  And there are certainly instances where a coding is controversial enough that I could agree being conservative and coding an uncertainty is a valid choice.  But not coding for the presence of a pubic symphysis in Archaeopteryx, or for fibular-tarsal contact in Allosaurus?  That's just lazy.  Now sure you might say that coding Allosaurus for fibular-tarsal contact won't add anything to the analysis, since carnosaurs and taxa closely related to carnosaurs all have the contact, and hey you coded Sinraptor for that character anyway.  But that's assuming you have the phylogeny correct before you even run the analysis, which begs the question of why you're even going through the motions of testing relationships in the first place.  In this particular case, I agree it's relatively harmless, but what about that same analysis (Choiniere et al.'s new Haplocheirus paper) where no birds with closed popliteal fossae were coded for that character?  Instead, all pygostylians were incorrectly left unknown.  This makes it so that derived alvarezsaurids are no longer even attracted to pygostylians in the matrix based on that character, which means that despite including the relevent taxa and character, you're not really testing that hypothesis.  It just breeds complacency and a false impression of consensus, leaving us unlikely to ever find unexpected relationships from our data.  Recent papers which are guilty of this problem are the Smith et al. (2007) Cryolophosaurus analysis, the Xu et al. (2009) Limusaurus analysis and the Choiniere et al. (2010) Haplocheirus analysis.  I know coding for so many taxa can be tedious, but these papers all gave the impression they performed impressively large analyses when they each actually took shortcuts and presented cladograms which aren't representative of the data they purport to include.  As scientists we have a duty to present accurate data to the public.  Virtually no one will bother looking through your matrix to make sure you did things correctly, and that includes peer reviewers and editors unfortunately.  How else to explain how Choiniere et al. could code Rahonavis for the presence of a mandibular character, when anyone who studies theropods knows it doesn't preserve a skull.  They just copied Makovicky et al.'s (2005) codings without thought, assuming they were correct.  But the public sees your cladogram and (justifiably in my opinion) assumes you did the work to create it.

I'll repeat a point that needs to be made depressingly often- the function of cladistic analysis (and science in general) is to question, not to confirm.  It does no good to only feed PAUP the data which supports your idea.  You might as well not even waste the time, and just list the characters instead.  The entire point of cladistic analysis is that it can somewhat objectively weigh competing ideas and tell us which is most parsimonious.  So if you're going to run an analysis, take some time to track down the characters which disagree with your hypothesis and code them for every available taxon.  Hell, send your data to me and I'll code the taxa for you in exchange for being slapped on as a coauthor.  I have a database of theropod codings at my disposal anyway.  That way you save time and your analysis will actually be a useful contribution to science.  Otherwise the general populace may be impressed by your cladogram, but your peers will realize they can ignore it because it didn't include the relevent data.

Smith, Makovicky, Hammer and Currie, 2007. Osteology of Cryolophosaurus ellioti (Dinosauria: Theropoda) from the Early Jurassic of Antarctica and implications for early theropod evolution. Zoological Journal of the Linnean Society. 151, 377-421.

Xu, Clark, Mo, Choiniere, Forster, Erickson, Hone, Sullivan, Eberth, Nesbitt, Zhao, Hernandez, Jia, Han and Guo, 2009. A Jurassic ceratosaur from China helps clarify avian digital homologies. Nature. 459, 940-944.

Choiniere, Xu, Clark, Forster, Guo and Han, 2010. A basal alvarezsauroid theropod from the Early Late Jurassic of Xinjiang,China. Science. 327, 571-574.