Re: neotheropod classification

["Mickey Mortimer" <Mickey_Mortimer11@msn.com>]

Ken Kinman wrote-

>       Okay, I am leaving therizinosaurians as a clade (with Beipiaosaurus
as
> sister group to the others).  But still splitting them off as a separate
> clade before the oviraptorosaur clade.
>       Now if I could only decide what the H**L to do with Family
> Troodontidae.  Anyway, below is the revised portion of my neotheropod tree
> (from which I exclude Eoraptor and herrerasaurids).  The coding shows a
> definite "Mortimer" influence, n'est ce pas?

Hooray!  You've seen the light! ;-) I'm glad you finally accept segnosaurs
as monophyletic and close to oviraptorosaurs.  But I must ask why you would
put them below oviraptorosaurs.  What you must ask yourself is "what
characters do segnosaurs lack, but maniraptorans have?".

The only analysis to have this arrangement was Sereno (1999), who also had
tyrannosauroids above ornithomimosaurs and alvarezsaurids sister to
ornithomimosaurs.  This was based on several characters supposedly lacking
in segnosaurs, alvarezsaurids and non-maniraptorans, present in but present
in tyrannosauroids, oviraptorosaurs and paravians.  This formed the basis
for Sereno's Tyrannoraptora (all descendents of the most recent common
ancestor of Tyrannosaurus and birds), though in his phylogeny it was a
junior synonym of Maniraptora.  Most of these have problems, as can be seen
below-
- large hypapophyses on anterior dorsals.  Sereno codes ornithomimosaurs as
lacking these, despite the fact they are comparable to those of
tyrannosaurids and carnosaurs (Makovicky, 1995), which he codes as having
them.  He also codes alvarezsaurids as lacking them, but Mononykus has them.
Alvarezsaurus doesn't, so they should be coded as polymorphic.  Finally, he
codes segnosaurs as lacking them, but the evidence isn't good, being based
on Russell and Dong's (1993) skeletal reconstruction of Alxasaurus.  Because
the remains were crushed and described poorly, Makovicky coded them as "?".
- furcula present.  Because neither the furculae of Segisaurus, Coelophysis
or Beipiaosaurus were published when Sereno wrote the papers, ceratosaurs
are coded 0 and segnosaurs ?.  Similarily, Sereno codes alvarezsaurids as 0,
despite the fact no articulated specimen has been found yet.  Also,
Scipionyx was not included.  This all adds up to PAUP placing this character
at an incorrect position.
- semilunate made of single carpal.  The more basal Coelurus has this, but
was not included in Sereno's study.  So do Acrocanthosaurus and many
specimens of Allosaurus (Madsen, 1976), leaving me to wonder why carnosaurs
are coded as 0.  Beipiaosaurus also has this, as do ornithomimosaurs(?) and
Mononykus.  Another victim of incorrect placement by PAUP due to unincluded
taxa, incorrect codings and post-publication descriptions.
- lacrimal-frontal contact present.  This character's just fine.
- quadrate height 40-50% of mid-orbital skull height (primitive state
70-90%).
Here's the data-
non-tyrannoraptorans: outgroups (60-94), Ornitholestes (93), Segnosauria
(79), Ornithomimosauria (62-80)
tyrannoraptorans: Tyrannosauridae (58-68), Oviraptoridae  (57),
Alvarezsauridae (83), Troodontidae (45), Dromaeosauridae (58-79),
Archaeopteryx (60), Confuciusornithidae (56),  Enantiornithines (42)
not examined: Scipionyx (51), Avimimus (47),
I think you'll agree there is no definable point and the character is
faulty.
- coracoid dorsoventral length five times or more coracoid glenoid diameter.
Basal oviraptorosaurs (Caudipteryx, Microvenator) have shorter coracoids
(3.4 and 3.49 times glenoid length respectively), so Sereno coded
Caudipteryx incorrectly.  I'm confused why caenagnathids are coded as
unknown, despite the fact they have the derived state (Russell, 1988).  I
think that caenagnathoids and the troodontid-eumaniraptoran clade developed
this independently, as did tyrannosaurids, whose glenoids are smaller due to
reduced arms.  In the other taxa, the coracoid elongates.
- distal end of radius anteroposteriorly flattened
This is one of those characters nearly impossible to find in the literature,
but as Gallimimus and Mononykus have distal radii 62% and ~60% as wide as
they are broad, compared to Deinonychus' 63%, I don't think it's valid.  It
may define Coelurosauria or Maniraptoriformes though, as Acrocanthosaurus'
is 97%.
- distal end of ilial pubic peduncle horizontal (primitive state sloping
posteroventrally)
Sereno codes segnosaurs, alvarezsaurids, oviraptorosaurs and arguably
tyrannosaurids incorrectly.  The first two have the derived state (hell,
segnosaurs' are sloping anteroventrally), while the third and probably the
fourth have the primitive state.
- reduced supracetabular crest of ilium
This is an artifact of setting the character states to DELTRAN.  The crest
is absent in tyrannosaurids, segnosaurs and other maniraptorns, as Sereno
correctly codes.  It is present in more basal theropods, alvarezsaurids and
ornithomimosaurs.  Thus, DELTRAN interprets this as developing parallel in
segnosaurs and tyrannoraptorans.  As the basal tyrannosauroid Stokesosaurus
has the supracetabular crest, I think there's a good chance it evolved in
parallel in tyrannosaurids and maniraptorans.  In any case, as segnosaurs
have the character, it's no reason to keep them more basal than
oviraptorosaurs.
- ischial foot absent
Sereno correctly codes alvarezsaurids as lacking an ischial foot, but failed
to notice segnosaurs do too.   The distal expansion is the obturator
process, as can be seen by the unexpanded tip distal to it.  Even some
ornithomimosaurs lack the ischial foot (Rauhut, 2000).
- anterior trochantor reaches level of greater trochantor
Sereno codes both segnosaurs (Xu et al., 1999) and alvarezsaurids (Novas,
1997) incorrectly for this character, as they both have it.
- anterior trochantor splitting from greater trochantor closer to femoral
head.
Sereno codes tyrannosaurids incorrectly, as they have a condition
near-identical to carnosaurs, ornithomimosaurs, Alxasaurus and
Alvarezsaurus.  Patagonykus and mononykines have the derived condition, so
they should be coded as polymorphic.  This character would be diagnostic of
Sereno's Maniraptora (actually an unnamed clade), discussed below.

Next, we have those characters diagnosing Sereno's Maniraptora, which
consists of oviraptorosaurs and paravians.  As Maniraptora is stem-based, it
would also include tyrannosauroids in Sereno's phylogeny, so this clade is
named inappropriately.  Unfortunately, there is no current name for an
oviraptorosaur-paravian clade if there are more basal maniraptorans.  The
following characters are supposedly absent in tyrannosauroids, segnosaurs,
alvarezsaurids, ornithomimosaurs and more basal theropods, but present in
oviraptorosaurs and paravians-
- lacrimal-jugal articulation slotted
This is nearly impossible to tell from the literature, but Tom Holtz
included it in his analysis (1999) and coded it from specimens.  Thus, we
can see that Sereno coded ornithomimosaurs, dromaeosaurids and troodontids
differently from Holtz.  Sereno has these lacking the state, while Holtz has
them with the derived state.  In any case, the character is very homoplasic
even in Sereno's phylogeny.
- ischium 50% or less of pubic length
Sereno coded oviraptorosaurs incorrectly (Caudipteryx- 58%; Nomingia- 60%;
Ingenia- 83%), although Chirostenotes' is 45%.  Same with troodontids and
dromaeosaurids- Sinornithoides' is 52%, Velociraptor's is 57%,
Achillobator's is 69%.  Obviously not an accurate character.
- cervical centra weakly opisthocoelous to amphicoelous
Sereno correctly has both segnosaurs and ornithomimosaurs as having this,
PAUP just interprets it as a parallel situation.  He codes alvarezsaurids as
lacking it, although they are polymorphic, as Alvarezsaurus has it.
Ornitholestes is coded incorrectly, as it has platycoelous centra.  This
character also doesn't help segnosaurs be below oviraptorosaurs.
- lacrimal ventral process anteriorly directed (primitive state laterally
directed)
I must admit to not understanding this character.  It is ventrally directed
in all theropods.  Perhaps he meant ventral to be the primitive state and
the derived state would be anteroventral.  Or perhaps it's the direction the
long axis of the ventral process's cross section.  I don't know.
- jugal anterior process tapered anteriorly
Don't know what segnosaur skull Paul was looking at, but Erlikosaurus' is
tapered.  Another character coded incorrectly.
- laterosphenoid head small, snovial joint small or absent
A braincase character I'm not yet skilled enough to identify.  Sereno codes
it as unknown in segnosaurs though, so it can't keep them out of
Maniraptora.
- uncinate processes present
Unknown in segnosaurs, so it can't keep them out of Maniraptora either.
- sternal anterior margin slotted for coracoid
Also unknown in segnosaurs.
- acromion height less than 40% of the distance from the middle of the blade
to the acromial edge
Here's the statistics-
Allosaurus (73), Deltadromeus (70), Scipionyx (~55), Compsognathidae (~65),
Tyrannosauridae (74), Segnosauria (75-78), Ornithomimosauria (66-74),
Alvarezsauridae (52),  Caudipteryx (20), Deinonychus (26-65)
This character's fine, only Achillobator has the primitive condition among
known taxa in Sereno's clade.
- humeral deltoid depression broad and triangular.  Not familiar with this
character, or detailed humeral anatomy.  Segnosaurs are conveniently coded
as unknown for it though.
- olecranon process absent
As much as this is true of any coelurosaurian, it's true of Therizinosaurus
and probably Beipiaosaurus and Alxasaurus.
So another character coded incorrectly.
- articular surface for humerus on ulna an equilateral triangle (primitive
condition anteroposteriorly elongate triangle)
Another character I'm unsure of.  Is this the entire ulnar proximal end?  I
doubt it, as Gallimimus' width/depth ratio is 90% compared to Deinonychus'
92%.  That wouldn't be much of a difference.
- articular surface for humerus on ulna divided by low intercondylar crest
Another hard character to find in the lit, but Ostrom says of Deinonychus,
"The proximal articular surface .... is nearly flat, with very finely rugose
texture.  The surface is triangular in shape with a slight concavity marking
the radial margin."  Seems like it was absent in Deinonychus, contra Sereno.
- bowed ulnar shaft
True, though also present in the basal Coelurus and Scipionyx, which Sereno
did not test.  However, it is present in Ornitholestes as well, which Sereno
did test and coded incorrectly.
- distal end of ulna 3 times as broad as tall
Oh, I love these proximal/distal forelimb element characters ;-) .  This
isn't true of Deinonychus, but at least it has a more flattened distal end
than Acrocanthosaurus, Gallimimus or Patagonykus.  Don't know where you'd
get measurements for segnosaurs.  If we assume Sereno examined a segnosaur
to get this ratio, and adjust the character to "more than 1.5 times as broad
as tall", can could assume it keeps segnosaurs out of Maniraptora.
- carpal trochlea of metacarpal I extends distally down medial side of
metacarpal
This looks to be best developed in Gallimimus, then Chirostenotes and
Patagonykus.  Deinonychus and Allosaurus have small medial exposure.  I
can't see how it would define the clade Sereno calls Maniraptora.
- medial condyle of metacarpal I dorsal to lateral condyle
This is true of maniraptorans (eg. Deinonychus, Chirostenotes), but also of
carnosaurs (Eg. Allosaurus, Acrocanthosaurus) and even Torvosaurus.
Gallimimus and alvarezsaurids lack it though.  With polarity reversed, it
could be a good ornithomimosaur-alvarezsaurid character, but it doesn't
concern the current post.  Sereno codes it as absent in caenagnathids for
some reason, though Chirostenotes? rarus has it (Osmolska, 1987).  And of
course he misscodes torvosaurids and allosaurids.
- postacetabular process half-concentric (primitive condition trapezoidal)
At last, an easily observable character.  I wonder how he coded segnosaurs
as being primitive if only Beipiaosaurus exhibits the primitive condition,
and he didn't use it for other characters (eg. furcula).  I can see
Chirostenotes and Oviraptor being derived, but not Caudipteryx, Nomingia or
Ingenia.  Confusing distribution, but ignoring some oviraptorosaurs and
accepting the development in therizinosauroids as parallel, I suppose it
works.
- ischial obturator notch greater than acetabular diameter
Alvarezsaurids lack an obturator process where known, so should really be
uncodable.  Segnosaurs have very distally placed obturator processes, so are
coded incorrectly by Sereno.  Ornitholestes also has it, so is coded
incorrectly as well.  This character is therefore invalid when defining
Sereno's Maniraptora.
- fourth trochantor reduced or absent
Another victim of DELTRAN, this is only here because all coelurosaurs he
analyzed have it except tyrannosauroids and segnosaurs, so it's most
parsimoniously seen to be developed in Deltadromeus, the
alvarezsaurid-ornithomimosaur clade and the oviraptorosaur-paravian clade in
parallel.  Developing it once and having it lost by tyrannosaurids and
segnosaurs would be equally parsimonious (three steps) and would be found by
ACCTRAN.
- anterior muscle scar on distal femur absent
This is hard to see in most taxa due to illustration quality, but
alvarezsaurids (Patagonykus, Mononykus) lack the scar, so are coded
incorrectly.  Even Tyrannosaurus seems to lack a scar (Osborn, 1916).  The
condition in segnosaurs is not described or illustrated.  I would not trust
Sereno's coding for this character.
- tibial malleoli absent
Simply untrue.  Perhaps reduced malleoli, but I can't even find evidence of
this.  Nearly all theropods I examined (including overaptorosaurs,
troodontids and dromaeosaurs) have malleoli, often stronger medially.  The
single exception is a juvenile Daspletosaurus that lacks the medial
malleolus (Lehman and Carpenter, 1990).
- astragalar surface for distal tibia flat
Absent in Allosaurus and Gorgosaurus, but I cannot find information for
other taxa.
- anterior vascular fossa of astragalus absent
I'm assuming Sereno is talking about the fossa on the base of the ascending
process.  Segnosaurus is like other maniraptorans in lacking the fossa, but
the rest of his codings are accurate.  So this character is also not
indicative of segnosaurs being outside Maniraptora.
- vaned feathers
Sereno has both segnosaurs and ornithomimosaurs coded as lacking feathers,
but only a small area of throat pouch is known from any ornitomimosaur
(Pelecanimimus- Briggs et al., 1997) and the details regarding how much the
filaments of Beipiaosaurus resemble feathers is obscure.  Sinornithosaurus
and cf. Sinornithosaurus have both filaments and feathers, so it's not
improbable Beipiaosaurus was similar in this regard.
- ischial shafts flattened in cross section
The shafts of most maniraptorans, including segnosaurs, oviraptorosaurs,
Patagonykus, troodontids and eumaniraptorans, are flattened to some degree
due to the enlarged, distally placed obturator process.  Therizinosauroids,
Patagonykus and some advanced eumaniraptorans have even flatter ischia.
Thus, Sereno codes segnosaurs and alvarezsaurids incorrectly.
- ischial symphysis confined to ventral edges (primitive condition
completely sutured)
Sereno codes alvarezsaurids as lacking this, despite the fact they have no
ischial symphysis!  However, it is impossible to make further comments based
on the literature, as the details of ischial fusion are not often discussed
or figured..
- acromion reduced to an everted edge
This is basically the "acromion height less than 40% of the distance from
the middle of the blade to the acromial edge" character from above, so I
count it as a repeat.
- coracoid glenoid convex
This is a valid character, as Gorgosaurus, Gallimimus, Therizinosaurus and
Segnosaurus have concave or flat glenoids, while Caudipteryx, Microvenator,
Chirostenotes and Deinonychus have convex glenoids.  He codes alvarezsaurids
inaccurately though as Patagonykus appears to have a convex surface and
Mononykus is said to (Perle, et al., 1994).
- narrow anterior trochantor (primitive condition wing-like)
This is possibly a valid character, but Sereno is incorrect in giving
alvarezsaurids the primitive condition.  Beipiaosaurus has a wing-like
anterior trochantor, but Alxasaurus and Segnosaurus don't.  Again, the
confusion of whether Sereno knew about Beipiaosaurus at the time he wrote
this.  In any case, segnosaurs should be coded as polymorphic, as we don't
know which state is primitive for them.

So there you have it.  If you've read this far, you know that there is
little evidence for placing segnosaurs below oviraptorosaurs.  The valid
characters lacking in segnosaurs are-
1- lacrimal does contacts frontal
2- acromion height less than 40% of the distance from the middle of the
blade to the acromial edge (reversed in Achillobator)
3- coracoid dorsoventral length more than five times coracoid glenoid
diameter (allowing for 2 reversals)
4- coracoid glenoid convex
5- bowed ulnar shaft (assuming parallel development in a few basal
coelurosaurs)
6- postacetabular process half-concentric (assuming parallel development in
therizinosauroids and reversals in a few oviraptorosaurs)
7- anterior trochantor splitting from greater trochantor closer to femoral
head
Three are found in tyrannosaurids (1, 3, 7) and one in alvarezsaurids (4).
The following may be valid, but I could not verify them, besides the last,
which is polymorphic for segnosaurs.
8?- anterior dorsal hypapophyses enlarged
9?- distal end of ulna more than 1.5 times as broad as tall
10?- astragalar surface for distal tibia flat
11?- ischial symphysis confined to ventral edges
12?- narrow anterior trochantor
Two are lacking in tyrannosaurids (8, 12), the others unknown.  One is
absent in alvarezsaurids (9), two present (11, 12) and one equivocal (8).
So, of Sereno's 42 characters, seven are demonstably valid and five more
potentially so.  Even the number of demonstrable segnosaur-oviraptorosaur
synapomorphies is greater, not to mention the maniraptoran characters they
have.  The following are all shared by segnosaurs and oviraptorosaurs.
- posterior section of premaxilla toothless (also in ornithomimids and some
birds)
- promaxillary fenestra absent (reversed in some oviraptorids)
- nasal subequal in length to frontal (also in some basal coelurosaurs and
derived eumaniraptorans)
- vomers extend posteriorly past maxillae (also in birds)
- ectopterygoid lateral to palatine
- reduced basipterygoid processes (also in ornithothoracines)
- dentary symphysis decurved (also in Ornitholestes and derived
ornithomimids)
- dentary symphysis deflected medially (also in ornithomimosaurs, Avimimus
and derived troodontids)
- two pairs of cervical pleurocoels (also in Achillobator and Variraptor)
- ventral sulcus on cervical centra flanked by ventrolaterally directed
ridges
- cervical prezygopophyses separated by a U-shaped space
- caudal vertebrae decrease in length posteriorly
?- ulnar facet of humerus expanded, merges with entepicondyle
- pubic foot expanded anteriorly (also in Achillobator)
Thus, I think it is wise to keep segnosaurs as an oviraptorosaur sister
group.  Of course, if anyone finds any more maniraptoran synapomorphies
lacking in segnosaurs, they're welcome to share them.

>   8  Guaibasauridae
>   9  Coelophysidae
> 10  Ceratosauridae
> _a_  Abelisauridae (incl. nodosaurs)
> 11  Spinosauridae (spinosaurs)
>   B  Torvosauridae ("megalosaurs")
> 12  Allosauridae ("carnosaurs")
> 13  Coeluridae (basal coelurosaurs)
>          1  Proceratosaurus
>          ?  Deltadromaeus (carnosaur??)
>          2  Dryptosaurus
>          3  Nedcolbertia
>          4  Ornitholestes
>          ?  Santanaraptor
>          5  {{Tyrannosauridae}}
>          6  Compsognathus
>          B  Sinosauropteryx
>          7  Scipionyx
>          8  Coelurus
>          9  Nqwebasaurus
>         10  {{Ornithomimidae to AVES}}
> _a_  Tyrannosauridae (sensu lato)
> _1_  Ornithomimidae
> 2A   Plesion _Beipiaosaurus_
>   B   Therizinosauridae
>   3   Plesion _Protarcheopteryx_
> 4A   Caudipteridae (probably paraphyletic)
>   B   Caenagnathidae
>   C   Oviraptoridae
>   5   Plesion _Bagaraatan_
>   6   Avimimidae
>   ?   Troodontidae (misfits??)
>   7   Alvarezsauridae
>   8   Dromaeosauridae (deinonychosaurs)
>   9   Plesion _Unenlagia_
> 10   Plesion _Rahonavis_
> 11   {{AVES/AVEA}}

I really like the coelurosaur section, but I don't think you have enough
non-coelurosaur families to properly represent their diversity.  Really
should keep sinraptorids and carcharodontosaurids separate families, and
you'd need a lot more plesia (Dilophosaurus, Monolophosaurus,
Piatnitzkysaurus, Magnosaurus, etc.).  I'd put troodontids above
alvarezsaurids, that's where they are in both mine and Tom's most recent
analyses.  And of course there's the other eumaniraptoran plesia
(Microraptor, Sinornithosaurus, Bambiraptor, etc.).

Mickey Mortimer