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

Re: Revising Hou et al, 96 (long)



Nicholas Gardner wrote-

> --Aves
>   |--Sauriurae
>   |  |--Archaeopteryx
>   |  `--+--Confuciusornis
>   |     `--Enantiornithes
>   `--Ornithurae
>      |--Liaoningornis
>      `--+--Chaoyangia
>         `--+--Hesperornis
>            `--+--Ichthyornis
>               `--Modern birds

Ah, the ridiculousness.

> It looks from the caption that they ran 31 characters to get this
phylogeny.
>   Some of them seem shockingly like plesiomorphies :-).  At least one
would
> be hard to tell what taxa had it or not.  And some are plain-out
worthless.

Let's be fair and use the actual characters as they appear in Hou et al.,
shall we?  The following are Hou et al.'s characters in order, but with some
excluded (marked with *), so that the numbers do not correspond.  The names
above the characters represent the taxon that group of characters is
supposed to diagnose in Hou et al.'s phylogeny.

Aves-
1. feathers
I leave this unknown in dromaeosaurids, though taxa some consider members of
this family (Sinornithosaurus, Microraptor) have them, as does the new
possible Yixian dromaeosaurid.  They are assumed present in those avians
with elongate arms and well developed flight adaptations.
2. postorbital absent
Though not known in any of the enantiornithine taxa Nick lists, the Spanish
nestling has one.
*. hypopubic cup
This was not included because it is merely the ABSRDers excuse for
Archaeopteryx's pubic foot.  Archaeopteryx does not have a hypopubic cup,
though Velociraptor ironically has a similar structure.
3. opisthopubic pelvis
4. reversed hallux

Euornithes-
5. laterally flexible furcula
Although the flexibility of fossil furculae is difficult to determine, Hou
et al. are undoubtedly referring to the thin non-boomerang-shaped furculae
of post-confuciusornithid birds for this character they use to diagnose
euornithines.  However, many coelurosaurs also had thin furculae
(tyrannosaurids, segnosaurs, Protarchaeopteryx, Caudipteryx, troodontids,
Velociraptor), though a few had boomerang-like ones (oviraptorids,
Bambiraptor, Sinornithosaurus).  My tentative phylogeny suggests that the
latter are primitive for eumaniraptorans, though this is far from certain.
6. concave facet for scapula on coracoid
7. procoracoid process
8. sternal keel extending anteriorly
This is not seen in the enantiornithines that Nick included, but is known in
Neuquenornis for example.
9. longitudinally elongate sternum
Again, not seen in Nick's taxa, but is present in some enantiornithines like
Eoalulavis.
*. coracoid sulci on sternum
This character is excluded because it is not documented in most taxa and was
already present in Velociraptor.
10. ossified uncinate processes
11. fully fused metatarsus
See the excluded character after 17 for details.
12. distally fused metacarpus
*. tarsal cap on the metatarsus
I admit to not understanding how this is supposed to be derived, the distal
tarsals always cap the metatarsus.

Chaoyangia + Ornithurae-
*. small pedal claws
This character is rejected because it is not quantified and highly
correlated with non-perching habits.
13. distal tarsals fused to metatarsus

Ornithurae-
14. no pubic symphysis

Carinatae-
15. hypotarsus

Neornithes-
16. supratendinal bridge on tibiotarsus

Sauriurae-
*. broad grooved furcular arms
This is just the opposite of character 5 and is a probable plesiomorphy.
*. outer condyle of femur with dorsal crest
Chiappe (2001) notes this supposed sauriurine character is absent in
Archaeopteryx and unknown in Confuciusornis, but is found in a few
enantiornithines (Concornis, Neuquenornis) and Vorona, though not Sinornis.
As the taxa with this character are not present in Nick's list, and it would
only make enantiornithines polymorphic and everything else plesiomorphic in
Hou et al,'s original analysis, it is excluded.
17. proximodorsal ischial process
*. proximal to distal metatarsal fusion
Excluded because the presence of only proximal fusion is simply the
plesiomorphic state before character 11, so it is modified to be "1" in that
character, while "fully fused metatarsus" is state "2" of character 11.

Confuciusornis + Enantiornithes-
18. shortened tail with pygostyle
Technically, there are taxa like Caudipteryx with long tails and pygostyles,
and we're not certain if things like Rahonavis and Yandangornis had such
small pygostyles, but for the purposes of this analysis, the character is
left as is.
19. dorsal pleurocoels
Hou et al. made this and the previous into a single character, which is
ludicrous as they are not correlated at all.

Enantiornithes-
20. reduced fingers
This supposed enantiornithine synapomorphy desperately needs quantification,
so I'll take it to mean that the phalangeal count must be at least as
reduced as enantiornithines (2-3-1).
21. elongated coracoid
Needs quantification again, but I'll take it to be at least as elongate as
enantiornithines.  There is a second part to Hou's version of this
character, the convex articulation on the coracoid for the scapula, but this
is just the opposite of character 6.
22. dorsal coracoid fossa
23. elongate hypocleidium
*. sternal keel much posterior to anterior margin
Excluded because it is just the plesiomorphic state of character 8.
24. metacarpal III extends past metacarpal II

Dromaeo ?0101?000100000000p00000
Archaeo 10010?000010000010000000
Confuci 10110?000110100011100000
Enantio 1011100pp01p100011p11111
Liaonin 1??1???11?1?1??0????1?1?
Chaoyan 1?1??????1???0??1???????
Hespero 111??100112?11011111100?
Ichthyo 111???11??21111?11111000
Neornit 111111111121111111011000

> So anyway, I decided to correct, update, and run their characters on
PHYLIP.

Running their taxa and corrected characters in PAUP with dromaeosaurids as
an outgroup results in 9 most parsimonious trees, because Chaoyangia can
have any non-ornithurine position.  It is not as complete as presented in
Hou et al., much of their figure being based on the possibly synonymous
Songlingornis.  Removing Chaoyangia results in one most parsimonious tree of
38 steps (CI .8421).  It has the standard avian phylogeny, though
Liaoningornis is closer to enantiornithines as I argued earlier.
|--Dromaeosauridae
`--+--Archaeopteryx
   `--+--Confuciusornis
      `--+--+--Liaoningornis
         |  `--Enantiornithes
         `--+--Hesperornis
            `--+--Ichthyornis
               `--Neornithes

>   And also incorporate some extra recent taxa in.  The basic point is,
could
> someone affirm that the codings are correct?

Sure.  You coded too many characters as being known, when we don't have
enough data to say.  For instance, character 1 (postorbital absent) is
unknown in Yandangornis, "Archaeoraptor", Longipteryx, Yandangornis,
Iberomesornis, Gobipteryx, "Cathayornis" (=Sinornis), Boluochia,
Patagopteryx, Yanornis, Liaoningornis and Ichthyornis.  But you have these
coded as 0 or 1.  Also, you have Protopteryx coded as lacking one, when it
has one.  Looking at the small amount of ?'s, I'm guessing the rest of your
codings are similar.  Here are the (hopefully) correct codings for the other
taxa you added, using the characters I listed above-

Rahonav 1?01?1????0?00??101?????
Yandang ???0???01?1?1????0?0????
Sapeorn 1???0?0???1000?011000?00
Protopt 10??1010001000???1101111
Longipt 1?111??001101???11?01?11
Jibeini 1?111???0?1000???1?01??0
Boluoch ??1???????1?1??011??????
Iberome 1?111????01?10??110?111?
Gobipte 1??1100???101000?1??1111
Sinornis 1?1110000?10100011111111
Patagop ??10??00??2111101?0?10??
"Archa" 1???1??11??0??????1?1?00
Yanorni 111?10111?2110???1111100
Yixiano 1?111?11?12010?0?1?11100
Apsarav 1?1??10???2?1111?10?11??

> Running the current matrix generates 7 trees.  The consensus tree is as
> follows-
>
> --+--Sapeornis
>   |--Yandangornis
>   |--Confuciusornis
>   |--Rahonavis
>   |--Deinonychus
>   |--Archaeopteryx
>   `--+--Jibeinia
>      |--+--"Archaeoraptor"
>      |  `--+--Yixianornis
>      |     `--+--Yanornis
>      |        `--+--Ichthornis
>      |           |--Hesperornis
>      |           |--Chaoyangia
>      |           |--Liaoningornis
>      |           |--Neornithes
>      |           |--Patagopteryx
>      |           `--Apsaravis
>      `--+--Longipteryx
>         `--+--Protopteryx
>            `--+--Cathayornis
>               `--+--Iberomesornis
>                  |--Boluochia
>                  `--Gobipteryx

There are far too many taxa for 24 characters to be able to separate, so
over 100000 trees resulted, though a clade containing Patagopteryx,
Apsaravis, Hesperornis, Ichthyornis and Neornithes was found.  Deleting the
poorly known Yandangornis, Boluochia, "Archaeoraptor" and Chaoyangia
resulting in "only" 962 trees.  The consensus manages to have a basal
polytomy of taxa (Dromaeosauridae, Archaeopteryx, Rahonavis, Sapeornis,
Confuciusornis), and a polytomy of "enants" (Protopteryx, Longipteryx,
Jibeinia, Iberomesornis, Gobipteryx, Sinornis, Liaoningornis) and the
Euornithes a bit further up.  The Euornithes has Yixianornis and Yanornis
successively closer to an ornithurine group (Patagopteryx, Apsaravis,
Hesperornis, Ichthyornis, Neornithes).  Clearly this is consistant with the
standard avian phylogeny, but such a small amount of characters will not do.

> It does seem strange that when you correct the characters the phylogeny no
> longer supports the Sauriurae-Ornithurae dichotomy. :-)

Agreed.

Mickey Mortimer