[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]
Fun with Partitions and Constraints
Just for fun, I decided to see the effect of only using subsets of
characters in a large theropod analysis. I used Holtz's (2000, 2001)
because it's the least biased large analysis available and I have a NEXUS
file of it. There were some very interesting results.
First, his analysis when all characters are used-
strict consensus of 189 mpt's-
|--allzero
|--Herrerasauridae
`--+--Sauropodomorpha
`--+--+--Coelophysidae
| |--Dilophosaurus
| |--Elaphrosaurus
| `--+--Ceratosaurus
| `--+--Abelisaurus
| `--Carnotaurus
`--+--Torvosaurus
`--+--Eustreptospondylus
`--+--Piatnitzkysaurus
`--+--Afrovenator
`--+--Megalosaurus
`--+--+--"Megalosaurus" hesperis
| `--Spinosauridae
`--+--Monolophosaurus
`--+--+--+--Sinraptor
| | `--Yangchuanosaurus
| `--+--+--Neovenator
| | `--Allosaurus
| `--+--Acrocanthosaurus
| `--+--Carcharodontosaurus
| `--Giganotosaurus
`--+--Gasosaurus
|--Proceratosaurus
`--+--+--Bagaraatan
| `--Ornitholestes
`--+--Scipionyx
`--+--+--Compsognathidae
| `--+--Tyrannosauridae
| `--+--Troodontidae
| `--+--Pelecanimimus
| `--Ornithomimidae
`--+--Dryptosaurus
`--+--Coelurus
`--+--Deltadromeus
|--Protarchaeopteryx
|--Caudipteryx
`--+--+--Segnosauria
|
`--+--Caenagnathidae
| |--Microvenator
| `--Oviraptoridae
`--+--Dromaeosauridae
`--+--Archaeopteryx
|--+--Unenlagia
| `--Rahonavis
`--+--Alvarezsauridae
`--+--Confuciusornis
`--Ornithothoraces
Note a few differences from the 2000 phylogeny-
- Coelophysoidea not supported, nor a neoceratosaurian Elaphrosaurus.
- Completely different ordering of basal tetanurines.
- "Megalosaurus" hesperis stabilized as sister to Spinosauridae.
- Monolophosaurus not an avetheropod
- Dryptosaurus and Coelurus as maniraptorans.
- Compsognathids as arctometatarsalians, and troodontids as well in all
mpt's.
- Deltadromeus stabilized as sister to enigmosaurs+paravians.
- Unenlagia stabilized as sister to Rahonavis.
These are presumably caused by the inclusion of the three taxa from his 2001
paper- Protarchaeopteryx, Caudipteryx and Confuciusornis.
Now, using only the 135 cranial characters results in 449 mpt's. Some taxa
had to be excluded since they don't have enough cranial material to be
placed in the trees.
The 90% majority rule consensus was-
|--allzero
|--Herrerasauridae
`--+--Sauropodomorpha
`--+--Coelophysidae
`--+--+--Ceratosaurus
| `--+--Abelisaurus
| `--Carnotaurus
`--+--Dilophosaurus
`--+--+--Afrovenator
| `--+--Torvosaurus
| |--Eustreptospondylus
| `--Piatnitzkysaurus
`--+--+--Monolophosaurus
| `--+--+--Sinraptor
| | `--Yangchuanosaurus
| `--+--Neovenator
| `--+--Allosaurus
| `--+--Acrocanthosaurus
| `--+--Giganotosaurus
| `--Carcharodontosaurus
`--+--"Megalosaurus" hesperis
|--Spinosauridae
`--+--Proceratosaurus
|--Ornitholestes
`--+--Bagaraatan
|--Tyrannosauridae
|--Scipionyx
|--Compsognathidae
|--Dromaeosauridae
`--+--+--Archaeopteryx
| `--+--Alvarezsauridae
| |--Confuciusornis
| `--Ornithothoraces
`--+--+--Troodontidae
| `--+--Pelecanimimus
| `--Ornithomimidae
`--+--Segnosauria
`--+--Caudipteryx
|--Caenagnathidae
|--Microvenator
`--Oviraptoridae
Note several important differences-
Ceratosaurs are paraphyletic.
Megalosaurs are monophyletic, and spinosaurids are coelurosaurs.
Monolophosaurus is a carnosaur, and Neovenator is the basalmost allosaurid.
Compsognathids, tyrannosaurids and dromaeosaurids are not maniraptoriformes.
Enigmosaurs are arctometatarsalians.
Using only the 75 axial characters results in 1951 mpt's. Again, there were
some necessarily excluded taxa. The 90% majority rule consensus was-
|--allzero
|--Sauropodomorpha
`--+--Herrerasauridae
`--+--+--+--Ceratosaurus
| | `--Carnotaurus
| `--+--Dilophosaurus
| `--+--Elaphrosaurus
| `--Coelophysidae
|--Spinosauridae
|--Torvosaurus
|--Eustrepospondylus
`--+--Afrovenator
`--+--+--Monolophosaurus
| |--Giganotosaurus
| `--+--+--Neovenator
| | `--Allosaurus
| `--+--+--Acrocanthosaurus
| | `--Carcharodontosaurus
| `--+--Sinraptor
| `--Yangchuanosaurus
`--+--Piatnitzkysaurus
|--Compsognathidae
`--+--+--Tyrannosauridae
| `--Ornitholestes
`--+--Scipionyx
|--Coelurus
`--+--+--Pelecanimimus
| `--Ornithomimidae
`--+--+--Segnosauria
| `--+--Microvenator
| `--+--Caenagnathidae
| `--Oviraptoridae
|--Protarchaeopteryx
|--Caudipteryx
`--+--+--Troodontidae
| `--Dromaeosauridae
`--+--Alvarezsauridae
`--+--Archaeopteryx
`--+--Rahonavis
`--+--Unenlagia
`--+--Confuciusornis
`--Ornithothoraces
Differences here include-
Theropod herrerasaurids.
Coelophysoidea, with Elaphrosaurus.
Sereno et al.'s derived megalosaurs are not necessarily tetanurines,
Monolophosaurus is a carnosaur, and Piatnitzkysaurus is a coelurosaur.
Carcharodontosaurines are sinraptorids, and Giganotosaurus is a basal
carnosaur.
Coelurus and tyrannosaurids are non-maniraptoriformes.
Troodontids are deinonychosaurs, and Caudipteryx and Protarchaeopteryx are
part of the enigmosaur-paravian crown.
Alvarezsaurids are not avians, and rahonavis and Unenlagia form successively
closer outgroups yo pygostylians.
Using only the 74 forelimb characters really messes the tree up. Here's a
90% majority rule consensus of the first 50000 mpt's or so.
|--allzero
|--+--Herrerasauridae
| `--Carnotaurus
|--Sauropodomorpha
|--+--Coelophysidae
| `--Dilophosaurus
|--Elaphrosaurus
|--Ceratosaurus
`--+--Megalosaurus
|--Eustreptospondylus
|--+--Torvosaurus
| `--+--Dryptosaurus
| `--Piatnitzkysaurus
|--+--Sinraptor
| `--+--Acrocanthosaurus
| `--+--Neovenator
| `--Allosaurus
|--+--+--Spinosauridae
| | `--Alvarezsauridae
| `--+--Segnosauria
| `--Caenagnathidae
|--+--Compsognathidae
| `--+--Tyrannosauridae
| `--+--Pelecanimimus
| `--Ornithomimidae
|--Scipionyx
`--+--Afrovenator
`--+--Coelurus
|--Ornitholestes
|--Protarchaeopteryx
|--Caudipteryx
|--+--Microvenator
| `--Oviraptoridae
|--Troodontidae
`--+--Dromaeosauridae
`--+--Archaeopteryx
`--+--Unenlagia
|--Rahonavis
|--Confuciusornis
`--Ornithothoraces
There are some very strange clades here- herrerasaurids and Carnotaurus;
Dryptosaurus with megalosaurs; spinosaurids and alvarezsaurids, both sister
to segnosaurs and caenagnathids (all outside Maniraptora).
Of the 'backbone clades', only Tetanurae, Maniraptora, Eumaniraptora and
Aves were supported.
Torvosaurus is sister to Piatnitzkysaurus, and Afrovenator is a
maniraptoran.
Ornitholestes and troodontids are maniraptorans.
Finally, the 102 hindlimb characters create over 50000 mpt's, with the
following 90% majority rule consensus-
|--allzero
|--Herrerasauridae
`--+--Sauropodomorpha
`--+--+--Dilophosaurus
| `--+--Elaphrosaurus
| `--+--Coelophysidae
| `--+--Ceratosaurus
| `--Carnotaurus
`--+--Piatnitzkysaurus
|--Megalosaurus
|--Torvosaurus
|--Spinosauridae
|--Monolophosaurus
`--+--Eustreptospondylus
|--Afrovenator
`--+--Sinraptor
|--Yangchuanosaurus
`--+--Gasosaurus
|--Carcharodontosaurus
|--+--Giganotosaurus
| `--Neovenator
|--Acrocanthosaurus
|--Allosaurus
|--Coelurus
`--+--Deltadromeus
|--Bagaraatan
|--Scipionyx
`--+--Ornitholestes
`--+--+--Troodontidae
| `--+--Tyrannosauridae
| `--Ornithomimidae
`--+--Dryptosaurus
|--Compsognathidae
|--Protarchaeopteryx
`--+--Caudipteryx
`--+--+--Segnosauria
| |--Caenagnathidae
| |--Microvenator
| `--Oviraptoridae
`--+--Dromaeosauridae
`--+--+--Unenlagia
| `--Rahonavis
`--+--Archaeopteryx
`--+--Alvarezsauridae
`--+--Confuciusornis
`--Ornithothoraces
Differences from the total evidence tree include-
Coelophysids sister to neoceratosaurs, with Dilophosaurus as the basalmost
ceratosaur.
Eustreptospondylus and Afrovenator as the closest megalosaurs to
Avetheropoda.
Sinraptors outside of Avetheropoda, and Giganotosaurus sister to Neovenator.
Coelurus basal to other coelurosaurs, Deltadromeus outside
Maniraptoriformes, Ornitholestes more derived than Scipionyx and Bagaraatan.
Compsognathids in Maniraptora, Caudipteryx more derived than
Protarchaeopteryx.
Finally, I tried constraining his analysis to come up with a topology
congruent with my preferred one, on my website (
http://students.washington.edu/eoraptor/Phylogeny%20of%20Taxa.html ). Note
this is not the same as my analysis' (constantly changing) topology. This
resulted in 72 mpt's that were 106 steps longer (7.6%). I have many
polytomies in my tree, and Holtz's matrix resolved some. Piatnitzkysaurus
was a tetanurine outside Tyrannoraptora and Monolophosaurus + Allosauroidea.
"Megalosaurus" hesperis ended up as a megalosauroid. Neovenator ended up
closer to derived carcharodontosaurines than Acrocanthosaurus. Allosaurus
was closer to carcharodontosaurines than to sinraptorids. Yangchuanosaurus
and Sinraptor formed a clade exclusive of Gasosaurus. The 'basal
coelurosaurs' formed a Hennigian comb, successively closer to
maniraptoriformes- Proceratosaurus, Ornitholestes, Coelurus, Scipionyx,
Compsognathidae. Finally, Protarchaeopteryx was a basal paravian.
And since I was already doing all this, I constrained Holtz's trees to match
Rauhut's (2003) topology ( http://dml.cmnh.org/2001Jul/msg00110.html )as
well, since it's the only extensive theropod phylogenetic analysis published
since. 35 mpt's 72 steps longer (2.5%) than Holtz's were recovered.
Polytomous areas and unclassifed taxa in Rauhut's topology that were
resolved using Holtz's data included the following. Megalosaurus was more
derived than Piatnitzkysaurus, but outside Avetheropoda. "Megalosaurus"
hesperis was the sister taxon of Torvosaurus + Spinosauridae, with
Eustreptospondylus sister to these three. Monolophosaurus was the sister of
Afrovenator + Allosauroidea. Afrovenator was outside Allosauroidea.
Acrocanthosaurus was outside the Carcharodontosaurus + Giganotosaurus clade.
Gasosaurus and a Dryptosaurus + Scipionyx clade were between Proceratosaurus
and tyrannoraptorans. Bagaraatan was the sister clade of enigmosaurs +
paravians. Protarchaeopteryx was a basal paravian again. Aves had the
topology (Rahon(Archa(Alvare(Confuc,Ornitho)))).
Mickey Mortimer
Undergraduate, Earth and Space Sciences
University of Washington
The Theropod Database - http://students.washington.edu/eoraptor/Home.html