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

Comments on Suzuki et al.'s (2002) rejection of arctometatarsalian alvarezsaurids

I picked this up at SVP (only $40 for 18 pages, what a deal!) and thought
the DML might be interested in why Suzuki et al. reject Sereno's topology
placing alvarezsaurids next to ornithomimosaurs.

First, they have two problems with Sereno's methodologies-
1. They claim he coded much of the data from derived alvarezsaurids
(Shuvuuia), not basal ones (Alvarezsaurus, Patagonykus).  Much of this stems
from data not available for basal forms though (eg. cranial characters), so
it shouldn't be a problem.  The state in derived alvarezsaurids should be
assumed to be the same as basal alvarezsaurids until proven otherwise.
2. They say that since Sereno coded Aves as an OTU, alvarezsaurids could not
fall out within that clade (as in Perle et al., 1994).  This also shouldn't
matter though, as Chiappe now (2002) places alvarezsaurids outside Aves, as
do the AMNH team (2001-2002), Holtz (2000-2002) and Novas and Pol (2002).
Thus, I find these complaints to be unwarrented.

Suzuki et al. contest the validity of fifteen characters used by Sereno
(1999, 2001)-
1. dorsal premaxillary process dorsoventrally flattened.
Present in Shuvuuia and ornithomimids (eg. Struthiomimus), but also in
troodontids (eg. Byronosaurus, Saurornithoides mongoliensis).
Suzuki et al. believe that because the character is found in another clade,
the character is weakened.  "After all", they state, "this condition could
be considered equally synapomorphic of a troodontid-alvarezsaurid clade".
This is not a valid argument unless they advocate such a clade exists, which
they do not.  Similarly, their complaint about most ornithomimosaur skulls
being too crushed to determine the state in, and basal alvarezsaurids
lacking preserved crania is inconsequential.  You should not ask that every
taxon be scorable for a character to be valid, and basal taxa should be
assumed to be like derived taxa , as noted above.  As nearly every recent
study places troodontids within the Deinonychosauria, this character is a
potentially valid arctometatarsalian synapomorphy.
2. dorsal prefrontal exposure greater than lacrimal exposure.
First, Suzuki et al. argue the structure in Shuvuuia could be an ectethmoid.
Then they state that in ornithomimids, the exposure is actually subequal, as
in carnosaurs.   After checking the literature, I have to agree.  Gallimimus
is illustrated with a larger prefrontal (Osmolska et al., 1972), but
Dromiceiomimus is illustrated with a larger lacrimal (Russell, 1972).
Garudimimus seems to have a small prefrontal as well (Barsbold, 1984).  I
agree the available evidence suggests Gallimimus and Shuvuuia convergently
evolved this character.
3. orbital flange on prefrontal.
Suzuki et al. claim ornithomimids lack this character, citing Struthiomimus
specimen AMNH 5339 as an example.  However, Sereno (2001) describes its
presence in Pelecanimimus and illustrates it in another Struthiomimus
specimen (TMP 90.26.1).  Sereno's photo is pretty convincing, though the
reported absence in another specimen by Suzuki et al. is interesting.  I
tentatively consider this a potentially valid arctometatarsalian
4. dentary >80% of mandibular length.
Suzuki et al. estimate the length of Shuvuuia's dentary as <77%, and note
Garudimimus has a value of only 70%.  They also say that many other
tetanurines have dentaries as long as ornithomimids', making the character
invalid.  I agree here, as I examined it for my analysis and ended up
rejecting it due to the same problems.
5. maxillary and dentary teeth implanted in a groove.
Suzuki et al. uselessly bring up the fact that within the Ornithomimosauria,
only Pelecanimimus has this character (ornithomimids are toothless and
Harpymimus is poorly preserved/described).  They mention troodontids and
hesperorniformes show the character as well, but like the first character,
this is inconsequential.  By the way, Eoenantiornis also shows this.  You
might think it could support an avian relationship for alvarezsaurids, but
Archaeopteryx, Sinornis, Yanornis and Ichthyornis all lack it.  I conclude
this character is a potentially valid arctometatarsalian synapomorphy.
6. maxillary and dentary teeth uniform in size along jaw.
Suzuki et al. claim that while Shuvuuia shows this, Pelecanimimus has
maxillary teeth that increase in size posteriorly, which they cite the
original description for.  I could not determine this from that paper, but
it did say that the dentary teeth increase in size anteriorly (which is
Sereno's primitive state).  So I agree this is not a valid character.
7. chevrons four to five times longer than neural spines.
The chevrons of Shuvuuia are actually only three times longer, while
ornithomimids have the primitive condition (~2 times).  Even Sereno's (2001)
reconstruction of Struthiomimus has short chevrons.  Thus, Suzuki et al. are
correct in rejecting this character.
8. flexor depression proximal to distal condyles on proximal manual
Suzuki et al. have no problem with this character, except to note it cannot
be determined in digits II and III in Shuvuuia, so should be restricted to
digit I.
9. paired flexor processes on manual phalanx I-1.
Suzuki et al. doubt the homology of the processes in alvarezsaurids and
ornithomimids, based the difference in morphology (distinct ridges versus.
small projections).  As I understand it however, homology should be assumed
until suggested otherwise by a phylogeny.  Thus, I see no basis to doubt
this as a potential arctometatarsalian synapomorphy.
10. Dorsolateral tubercle on manual phalanx I-1.
Sereno originally listed a dorsomedial tubercle (1999), but later corrected
it to indicate a dorsolateral tubercle (2001).  Suzuki et al. state
Patagonykus lacks this character, but a small tubercle seems to be present
(Novas, 1997).  They use the same argument against homology as above, but
the condition in Patagonykus is very close to that in ornithomimids, making
it even less defendable.  I also view this as a potential arctometatarsalian
11. ventral surface of manual ungual I flattened and broad.
Suzuki et al. note that while mononykines show this morphology, Patagonykus
does not (Novas, 1997).  They also say the state in Alvarezsaurus cannot be
confidently accessed, but Novas (1996) describes and illustrates it with a
keel ventrally.  Thus, it appears mononykines and ornithomimids developed
this in parallel.
12. manual ungual flexor tubercles distally placed.
Suzuki et al. reject this character because the flexor tubercles of
alvarezsaurids are reduced, but this has little bearing on their placement.
I've examined this character before and found flexor tubercle placement to
vary in a continuum that is difficult to justify coding separate states for.
Thus, I reject this character, but on different grounds than Suzuki et al..
13. dorsal ilial edges converge.
Found in Shuvuuia and ornithomimids, but Suzuki et al. note it is absent in
Alvarezsaurus.  This is also found in several other taxa (eg.
tyrannosaurids, SMNK 2349 PAL, Nomingia, oviraptorids) and is most
parsimoniously primitive for coelurosaurs, and lost in Segnosaurus and
eumaniraptorans.  Alvarezsaurids have to change their state regardless of
where they are placed (as they exhibit both states), but the derived state
is primitive for Arctometatarsalia and outgroups, so would not help place
alvarezsaurids in the clade in any case.  Therefore, I agree with Suzuki et
al. and reject this character as an arctometatarsalian synapomorphy.
14. metacarpal-phalangeal joint of digit I with approximately 15 degrees
maximum extension.
Suzuki et al.'s main concern with this character seems to be that avians
also exhibit it, based on the fact no specimen has been preserved with digit
I extended very much.  However, it's also true that no described basal bird
has been preserved with the manus in lateral/medial view.  In addition,
Chiappe (2002) describes Sinornis as having a "well developed dorsal
extensor depression" on metacarpal I.  Deinonychus has a maximum extension
of ~30 degrees (Gishlick, 2001) and Citipati even more (Clark et al., 1999),
while Erliansaurus has an extension of 45 degrees for digit II at least.
Regardless of its presence in birds, ornithomimosaurs (Deinocheirus,
Gallimimus) do have reduced extension (~0 degrees), so the character is
provisionally accepted as a potentially valid arctometatarsalian
15. metacarpal III at least 75% the width of metacarpal II.
Like the previous character, Suzuki et al. point out that early avians also
exhibit similar ratios.  This varies widely among coelurosaurs, with
Nqwebasaurus, oviraptorids and some birds (Archaeopteryx, Gobipteryx,
Neuquenornis, Patagopteryx) exhibiting it as well.  Still, the distribution
suggests that if alvarezsaurids were arctometatarsalians, this would be a
valid synapomorphy (possibly including Nqwebasaurus, as suggested by Sereno

There are also the characters not discussed by Suzuki et al.-
16. maxillary teeth end far anterior to dentary teeth.
Potentially valid if defined specifically (maxillary teeth do not extend
posteriorly past antorbital fenestra).
17. dentary teeth smaller than maxillary teeth.
The primitive condition for theropods.
18. metacarpals with 60-70% of shafts in contact.
19. metacarpal I at least 60% the length of metacarpal II.
20. subnarial foramen absent.
21. preantorbital portion of maxilla more than 50% length of antorbital
Not valid as far as I can determine.
22. invaginated ventral margin of antorbital fossa.
Does not seem to be true in Shuvuuia, and is controversial in ornithomimids.
23. nasal anteroventral process absent.
Also in Erlikosaurus, some troodontids and a few birds, but seems valid.
24. reduced internal mandibular fenestra.
Also true in Erlikosaurus, oviraptorosaurs and basal birds, so it's possibly
plesiomorphic for maniraptoriformes and reversed in deinonychosaurs.
25. posterolateral surangular ridge absent.
Seems to be present in Dromiceiomimus at least (Russell, 1972).

As you can see, more work needs to be done, but at least ten of Sereno's
characters could potentially unite alvarezsaurids with ornithomimosaurs.
The idea should be seriously considered, and more people should include
these characters in their matrices.

Mickey Mortimer