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

The plausibility of alternative placements for theropod taxa



David Marjanovic wrote-

>> I should try adding the holotype to a non-maniraptoriform supermatrix I've
>> been working on (including all codings from Smith et al., 2007; Tykoski,
>> 2005; Carrano and Sampson, 2007; Rauhut, 2003; Azuma and Currie, 2000;
>> Allain, 2002; Ezcurra and Novas, 2006; etc.).

> Wow. Is that supermatrix your thesis?

Nah.  As you know, ensuring a preexisting matrix is coded correctly and has 
objective character states takes as much time as creating your own matrix.  So 
I would never use such a supermatrix for more than just seeing where the 
currently published matrices suggest taxa belong.  I made it because I was 
curious about ceratosaur and coelophysoid monophyly.  So far it contains all 
the characters, taxa and codings from-

Smith et al.'s (2007) large theropod matrix in the Cryolophosaurus paper. (347 
characters)
Tykoski's (2005) matrix focusing on coelophysoids and ceratosaurs from his 
thesis. (95 additional characters)
Carrano and Sampson's (2007) ceratosaur matrix. (39 additional characters)
Excurra and Novas' (2006) matrix focusing on coelophysoids from their 
Zupaysaurus paper. (61 additional characters)
Rauhut's (2003) theropod matrix, including Xu et al.'s (2006) codings for 
Guanlong and Dilong, Lamanna's (2004) codings for Megaraptor, and Yates' (2005) 
codings for Dracovenator and several other taxa. (20 additional characters)
Allain's (2002) matrix focusing on spinosauroids and carnosaurs from his 
Dubreuillosaurus paper (7 additional characters)
Azuma and Currie's (2000) matrix from their Fukuiraptor paper (31 additional 
characters), which is basically the same as Currie and Carpenter's (2000) 
Acrocanthosaurus paper.
Langer and Benton's (2006) matrix of basal dinosaur relationships (25 
additional characters).
Yates' (2006) sauropodomorph matrix (134 additional characters so far), in 
order to get good representation of the theropod sister group.

I only checked character accuracy when two matrices disagreed, and generally 
didn't code taxa for characters nobody else had.  Thus while superior to the 
studies listed above, it's by no means a finished product and represents more 
of a compilation of prior work than my own work.  Coding 131 taxa for 763 
characters the proper way is too large a project for me at the moment- that's 
what my coelurosaur analysis is and it's taking years. :)  Obvious future steps 
include adding Holtz et al.'s (2004), Carrano et al.'s (2005), Sereno et al.'s 
new carcharodontosaurid stuff, and the info from the Lophostropheus, 
Tyrannotitan and Mapusaurus papers.  Important taxa to add would be 
Scleromochlus, Lewisuchus, Sacisaurus, Eucoelophysis, Agnostiphys, 
Chindesaurus, Sarcosaurus, Chuandongocoelurus, Berberosaurus, Brontoraptor, 
Spinosaurus, Marshosaurus, Erectopus, Gasosaurus, Lourinhanosaurus, 
Yangchuanosaurus, Mapusaurus, Saurophaganax, Huaxiagnathus, Sinocalliopteryx, 
Tanycolagreus, Eotyrannus,
 Aniksosaurus, Juravenator, Nedcolbertia, Nqwebasaurus, Santanaraptor and 
Scipionyx.  Also, coding Falcarius, Protarchaeopteryx, Incisivosaurus, 
Sinovenator, Mei, Buitreraptor and/or Rahonavis in addition to the two birds 
and two derived dromaeosaurids would quite probably affect the placement of 
Ornitholestes and Bagaraatan.

Current results are-

|--Lagerpeton
|--Marasuchus
`--+--Ornithischia (5 taxa)
   `--+--Silesaurus
      |--Sauropodomorpha (4 taxa)
      |--Guaibasaurus
      `--+--Eoraptor
         `--+--+--Staurikosaurus
            |  `--Herrerasaurus
            `--+--+--+--Dracovenator
               |  |  |--Dilophosaurus wetherilli
               |  |  |--"Dilophosaurus" sinensis
               |  |  `--Cryolophosaurus
               |  `--+*-Gojirasaurus
               |     |--Lophostropheus
               |     |--Liliensternus
               |     |--Zupaysaurus
               |     `--+*-Shake'n'Bake taxon
               |        |--+--Segisaurus
               |        |  `--"Syntarsus" kayentakatae
               |        `--+--Coelophysis
               |           `--+--Megapnosaurus
               |              `--Procompsognathus
               `--+--+--+--Elaphrosaurus
                  |  |  `--Spinostropheus
                  |  `--+--Ceratosaurus
                  |     `--+--+--Noasaurus
                  |        |  |--Genusaurus
                  |        |  |--Masiakasaurus
                  |        |  `--Laevisuchus
                  |        `--+--Ekrixinatosaurus
                  |           |--Rugops
                  |           `--+--Abelisaurus
                  |              `--+--+--Rajasaurus
                  |                 |  |--Majungasaurus
                  |                 |  `--Indosuchus
                  |                 `--+--Ilokelesia
                  |                    `--+--Aucasaurus
                  |                       `--Carnotaurus
                  |--Streptospondylus
                  `--+--+--Poekilopleuron
                     |  `--+--+--+--Eustreptospondylus
                     |     |  `--+--Piveteausaurus
                     |     |     |--Dubreuillosaurus
                     |     |     `--Afrovenator
                     |     `--+--Torvosaurus
                     |        `--+--Chilantaisaurus
                     |           |--Irritator
                     |           |--Baryonyx
                     |           `--Suchomimus
                     `--+--Piatnitzkysaurus
                        |--"Szechuanoraptor"
                        |--Condorraptor
                        |--Xuanhanosaurus
                        `--+--Monolophosaurus
                           `--+--+--Tyrannotitan
                              |  |--Megaraptor
                              |  |--Carcharodontosaurus
                              |  `--Giganotosaurus
                              `--+--+--Neovenator
                                 |  `--+--Acrocanthosaurus
                                 |     `--+--Allosaurus
                                 |        `--Sinraptor
                                 `--+--Fukuiraptor
                                    |--Tugulusaurus
                                    |--"Alashansaurus"
                                    `--+--+--Deltadromeus
                                       |  `--+--Stokesosaurus
                                       |     `--Tyrannosaurus
                                       `--+--+--Dilong
                                          |  |--Mirischia
                                          |  `--Aristosuchus
                                          `--+--+--Guanlong
                                             |  `--Coelurus
                                             `--+--Proceratosaurus
                                                `--+--+--Sinosauropteryx
                                                   |  `--+--Compsognathus
                                                   |     `--Siamotyrannus
                                                   `--+--Ornithomimosauria (2 
taxa)
                                                      `--+--Ornitholestes
                                                         `--+--Bagaraatan
                                                            `--Paraves (4 taxa)

I forced various constraint trees to estimate the rough liklihood that 
alternative hypotheses are correct.  I tried to test as many possibilities as I 
could recall were ever suggested in the last thirty years, with some older ones 
in there for good measure.  Of course, with near certain coding errors, not 
every taxon coded for every character possible, and missing characters and 
taxa, these numbers shouldn't be taken as the final word.  But a topology that 
needs 5 more steps is far more likely to be true than one that needs 20 more 
steps.  So, starting with the most likely alternatives and working to the the 
least likely ones, with my subjective divisions of how liklihood correlates to 
extra steps in the trees...



Extremely possible-

The current tree, with the following non-consensus aspects-

Silesaurus as a basal saurischian.  Never suggested before, but all the data 
from Langer and Benton (2006) which kept it out of Dinosauria was included.

Eoraptor and herrerasaurids as successively closer outgroups to Avepoda, as 
suggested by Sereno et al. (1993).  This despite the fact Langer and Benton's 
data were included.

Dilophosaurids as coelophysoids, as in Paul (1988) and most cladistic analyses. 
 With Rauhut (2003), Yates (2005) and Smith et al. (2007) all contributing to 
the data, it certainly seems to support coelophysoid dilophosaurids.

"Syntarsus" kayentakatae being closer to Segisaurus than to Megapnosaurus, as 
suggested by Tykoski (2005).

Spinostropheus clading with Elaphrosaurus to the exclusion of other 
ceratosaurs, as in Lapparent (1960) and Carrano and Sampson (2007).

Ceratosaurus closer to abelisauroids than Elaphrosaurus is, as in Holtz (2000).

Ilokelesia being a carnotaurine, as in Carrano and Sampson (2007).

Streptospondylus being a basal tetanurine outside the 
Spinosauroidea+Avetheropoda clade, which hasn't been suggested before.

Poekilopleuron being a basal spinosauroid, which has not been suggested before.

Torvosaurus being closer to spinosaurids than to eustreptospondylids, as in 
Rauhut (2003).

The generally basalmost tetaurines Piatnitzkysaurus, Condorraptor, 
Xuanhanosaurus and "Szechuanoraptor" being closer to avetheropods than 
spinosauroids are.  This has been suggested previously for Piatnitzkysaurus 
(Novas, 1992) and "Szechuanraptor" (Chure, 2000).

Monolophosaurus being outside Avetheropoda, as in Smith et al. (2007).

Carcharodontosaurids being outside Avetheropoda, as in Coria and Salgado (1995).

Sinraptor being closer to Allosaurus than to Neovenator or Acrocanthosaurus, 
which I don't think has been suggested before.

Neovenator being closer to allosaurids than carcharodontosaurids, as in Hutt et 
al. (1996).

Acrocanthosaurus being closer to allosaurids than carcharodontosaurids, as in 
Stovall and Langston (1950).

Fukuiraptor being a coelurosaur, as in Longrich (2001).

Deltadromeus being a tyrannosauroid, which hasn't been suggested yet.  Though 
the possibly synonymous Bahariasaurus has (Paul, 1988; Chure, 2000).

Dilong being closer to birds than tyrannosaurids, as in Turner et al. (2007).  
This leaves no reason to postulate secondarily featherless tyrannosaurids.

Mirischia and Aristosuchus being closer to Dilong than Compsognathus (as in 
Niaish, online 2006; though he had Dilong as a tyrannosauroid).

Guanlong being closer to birds than tyrannosaurids (and Dilong!), which hasn't 
been suggested before and is frankly the opposite of what I expected.

Siamotyrannus being a compsognathid, which hasn't been suggested before.

Bagaraatan being a maniraptoran, as in Rauhut (2003).

Forcing Elaphrosaurus to be closer to abelisaurids than Ceratosaurus is (as in 
Holtz, 1994) adds only one step.

Forcing Streptospondylus to be sister to Eustreptospondylus (as in Allain, 
2001) adds only one step.  Streptospondylus moves into Spinosauroidea, whose 
topology remains the same.

Forcing Piatnitzkysaurus to be outside the spinosauroid-avetheropod clade (as 
in Rauhut, 2003 and Smith et al., 2007) adds only one step.  Condorraptor is 
its sister taxon in this case, as in Smith et al. (2007), and Xuanhanosaurus 
and "Szechuanoraptor" move along with it (as in Rauhut, 2003).

Forcing Neovenator to be a carcharodontosaurid (as in Rauhut, 2003 and others) 
adds one step.  The (Acrocanthosaurus(Sinraptor+Allosaurus)) clade remains, 
while Neovenator is the sister to Megaraptor in basal Carcharodontosauridae 
outside Avetheropoda.

Forcing Siamotyrannus to be a carnosaur (Pharris, 1997) adds one step.  The 
topology among paraphyletic carnosaurs is unchanged and it is a sinraptorid 
(which is where Pharris predicted it would be).

Forcing Mirischia to be a compsognathid (as in Martill et al., 2000) adds one 
step.

Forcing Coelophysis and Megapnosaurus rhodesiensis to be sister taxa (as in 
Tykoski, 2005 and others) adds 2 steps.

Forcing Eustreptospondylus to be closer to spinosaurids than Torvosaurus (as in 
Smith et al., 2007) adds 2 steps.  The other eustreptospondylids move with 
Eustreptospondylus.

Forcing Poekilopleuron to be a torvosaurid (as in Galton and Jensen, 1979) adds 
2 steps.  The newly formed Torvosauridae are still sister to Spinosauridae.

Forcing Chilantaisaurus to be a tyrannosauroid (as in Paul, 1988) adds 2 steps. 
 Interestingly, this is rather close to where "Chilantaisaurus" maortuensis 
ends up.

Forcing Allosauroidea (allosaurids, carcharodontosaurids and sinraptorids) (as 
in Sereno et al., 1996) to be monophyletic adds 2 steps.  The topology within 
it stays the same, so that carcharodontosaurids are most basal and Sinraptor is 
nested within allosaurids.  Monolophosaurus is still basal to Avetheropoda, 
while Fukuiraptor and Siamotyrannus are coelurosaurs.

Forcing Monolophosaurus to be a spinosauroid (as suggested by Headden on the 
DML, 2002) adds 2 steps.  It is the most basal spinosauroid in these trees.

Forcing Fukuiraptor to be sister to Siamotyrannus (as in Holtz et al., 2004) 
adds 2 steps.  The two are basal coelurosaurs outside Tyrannoraptora.

Forcing Procompsognathinae sensu Sereno (Procompsognathus + Segisaurus) (as in 
Sereno and Wild, 1992) adds 3 steps.

Forcing Dracovenator to be closer to Ceratosauria + Tetanurae than to 
Coelophysoidea (as in Yates, 2005) adds 3 more steps.  Oddly, it ends up as a 
spinosaurid, while Cryolophosaurus is sister to Ceratosauria+Tetanurae and 
"Dilophosaurus" sinensis is a basal tetanurine.  Dilophosaurus wetherilli stays 
in Coelophysoidea.

Forcing Deltadromeus to be a ceratosaur (as suggested by Sereno et al., 2002) 
adds 3 more steps.  It falls in the Elaphrosaurus + Spinostropheus clade.

Forcing Carnotaurus and Majungasaurus to clade to the exclusion of Rajasaurus 
(as in Wilson et al., 2003) adds 3 steps.

Forcing Sinraptoridae to be basal to Avetheropoda (as in Paul, 1988) adds 3 
steps.  In this tree, carcharodontosaurids are coelurosaurs.

Forcing Fukuiraptor to be a carnosaur (as in Azuma and Currie, 2000 and Holtz 
et al., 2004) adds 3 more steps.  It is placed sister to Allosaurus, while 
carcharodontosaurids and Monolophosaurus are still outside Avetheropoda.

Forcing Aristosuchus to be a coelurid (as in Lydekker, 1889) adds 3 steps.  In 
this tree, Guanlong is also a coelurid.

Forcing Coelurus to be a compsognathid (as in Lull, 1911; technically, 
Compsognathus would be a coelurid...) adds 3 steps.

Forcing Eusaurischia to exclude herrerasaurids and Eoraptor (as in Langer and 
Benton, 2006) adds only 4 more steps.  These trees have the same topology as 
Langer and Benton's paper, with Silesaurus outside Dinosauria, Eoraptor closer 
to eusaurischians than herrerasaurids are, and Guaibasaurus as a basal theropod.

Forcing Procompsognathus to be the most basal avepod (as in Paul, 1988) adds 4 
steps.

Forcing Cryolophosaurus to be a basal tetanurine (as in Smith et al., 2005) 
adds 4 more steps.  In these trees, "Dilophosaurus" sinensis is its sister 
taxon.

Forcing Piatnitzkysaurus to be a spinosauroid (as in Holtz et al., 2004) adds 4 
steps.  It falls out (often with Condorraptor) as the basalmost spinosauroid, 
even more basal than Poekilopleuron.

Forcing Neovenator to be the sister taxon to Allosaurus (as in Hutt et al., 
1996) adds 4 steps.  Acrocanthosaurus, Sinraptor and Megaraptor form successive 
outgroups, while carcharodontosaurids are coelurosaurs.

Forcing Coelurus to be a maniraptoran (as in Gauthier, 1986) adds 4 steps.  It 
is directly basal to Ornitholestes in this tree.

Forcing Proceratosaurus to be sister to Ornitholestes (as in Paul, 1988) adds 4 
steps.  This pairing ends up in basal Maniraptora.

Forcing Megapnosaurus rhodesiensis and "Syntarsus" kayentakatae to be sister 
taxa (as in Rowe, 1989) adds 5 more steps.  In these trees, Procompsognathus is 
closest to Megapnosaurus sensu lato, with Coelophysis one step further out.  
Segisaurus is now in the Zupaysaurus-Liliensternus mess.

Forcing Deltadromeus to be sister to ornithomimosaurs (as in Rauhut, 2003) adds 
5 more steps.

Forcing Spinostropheus to be sister to abelisaurs (but keeping Elaphrosaurus as 
the basalmost ceratosaur; as in Sereno et al., 2004) adds 5 more steps.

Forcing Megaraptor to be a spinosauroid (as suggested by Calvo et al., 2004) 
adds 5 steps.  It ends up sister to Torvosaurus+Spinosauridae.

Forcing Siamotyrannus to be a tyrannosauroid (as in Buffetaut et al., 1996) 
adds 5 steps.  Guanlong and Dilong are still closer to birds than to 
tyrannosaurids.

Forcing Bagaraatan to be a tyrannosauroid (as in my analyses from 2003 onward) 
adds 5 steps.

Forcing Torvosaurus to be closer to eustreptospondylids than to spinosaurids 
(as in Sereno et al., 1994) adds 6 steps.  In these trees, the four 'basalmost 
tetanurines' (Piatnitzkysaurus, Xuanhanosaurus, Condorraptor and 
"Szechuanoraptor") all move outside the spinosauroid-avetheropod clade.

Forcing Afrovenator to be outside the Eustreptospondylus + Torvosaurus + 
Spinosauridae clade (as in Sereno et al., 1994) adds 6 steps.  Dubreuillosaurus 
and Piveteausaurus stay with Afrovenator, and Torvosaurus is still closer to 
spinosaurids than Eustreptospondylus.

Forcing Monolophosaurus to be a carnosaur (as suggested by Zhao and Currie, 
1993) adds 6 steps.  It is the basalmost carnosaur, and the topology of 
Allosauroidea is the same as when its monophyly is forced, with Fukuiraptor and 
Siamotyrannus still coelurosaurs.

Forcing Ornitholestes to be a coelurid (as in Matthew and Brown, 1922) adds 6 
steps.

Forcing Ornitholestes to be outside Maniraptoriformes (as in Paul, 1988) adds 6 
steps.

Forcing Proceratosaurus to be a tyrannosauroid (as suggested by me on the DML) 
adds 6 steps.  

Forcing Dilophosaurus to be sister to Ceratosauria + Tetanurae (as in Rauhut, 
2003) adds 7 more steps.  In these trees, Dracovenator is sister to 
Dilophosaurus, while "Dilophosaurus" sinensis and Cryolophosaurus form a clade 
one node closer to ceratosaurs+tetanurines.

Forcing Deltadromeus to be a noasaurid (as in Wilson et al., 2003) adds 7 more 
steps.

Forcing Xuanhanosaurus to be a torvosaurid adds 7 steps.

Forcing Acrocanthosaurus to be a carcharodontosaurid (as in Sereno et al., 
1996) adds 7 steps.  In these trees, Neovenator is a more basal 
carcharodontosaurid, while Sinraptor and Allosaurus form a clade sister to 
Carcharodontosauridae.  Monolophosaurus and Megaraptor are outside 
Avetheropoda, Fukuiraptor and Siamotyrannus are coelurosaurs, and Tyrannotitan 
is basal to Carcharodontosaurus+Giganotosaurus.

Forcing Guanlong (as in Xu et al., 2006), Dilong (as in Xu et al., 2004) and/or 
Mirischia (as in Naish, online 2006) to be tyrannosauroids adds 7 steps.  They 
form a clade with Aristosuchus to the exclusion of Tyrannosauridae, with 
"Alashansaurus" as the basalmost tyrannosauroid.

Forcing Compsognathus to be outside Tyrannoraptora (as in Holtz, 1994) adds 7 
steps.

Forcing Proceratosaurus to be basal to Coelurus and tyrannosauroids (as in 
Holtz, 2000) adds 7 steps.



Possible but not well supported-

Forcing Liliensternus to be more closely related to Dilophosaurus than to 
Coelophysis (as in Paul, 1988) adds 8 steps.  Dracovenator (and sometimes 
Lophostropheus) joins this 'halticosaur' clade, while Zupaysaurus and 
Gojirasaurus stay closer top Coelophysis.  Cryolophosaurus and "Dilophosaurus" 
sinensis are then closer to Ceratosauria+Tetanurae.

Forcing "Szechuanoraptor" to be an allosaurid (as in Molnar et al., 1990) adds 
8 steps.

Forcing Afrovenator to be closer to Avetheropoda than Piatnitzkysaurus, 
Torvosaurus, Eustreptospondylus and Spinosauridae (as in Holtz, 2000) adds 8 
steps.  Dubreuillosaurus and Piveteausaurus stay in Spinosauroidea.

Forcing spinosauroids to be carnosaurs (as in Rauhut, 2003 and older sources 
with more extensive Carnosauria's) adds 8 steps.  Fukuiraptor also falls out in 
this clade, though Siamotyrannus remains a coelurosaur.

Forcing Zupaysaurus to be closer to ceratosaurs+tetanurines than to 
coelophysoids (as in Arcucci and Coria, 1997) adds 9 steps.  As in that study, 
dilophosaurids end up even closer to birds.

Forcing Piatnitzkysaurus to be a basal carnosaur (as in Harris, 1998) adds 9 
steps.

Forcing Coelurus to be a tyrannosauroid (as in Senter, 2007) adds 9 steps.  The 
topology is similar to when Guanlong is forced to be a tyrannosauroid, with 
Coelurus outside the Dilong+Guanlong+Aristosuchus+Mirischia clade.

Forcing Coelurus to be basal to tyrannoraptorans (as in Makovicky, 1995) adds 9 
steps.  

Forcing Bagaraatan to be outside Tyrannoraptora (as in Holtz, 2000) adds 9 
steps.

Forcing Proceratosaurus to be a ceratosaurid (as in Huene, 1926) adds 9 steps.



Implausible-

Forcing "Szechuanoraptor" to be a sinraptorid (as in Paul, 1988) adds 10 steps.

Forcing Piatnitzkysaurus to be a eustreptospondylid (as in Paul, 1988) adds 10 
steps.  It ends up sister to other eustreptospondylids.

Forcing Sinraptor to be basal to allosaurids and carcharodontosaurids within 
Allosauroidea (as in Harris, 1998) adds 10 steps. Oddly, this forces 
spinosauroids to be carnosaurs, in addition to Monolophosaurus.  While 
Neovenator and Acrocanthosaurus are carcharodontosaurids, Fukuiraptor is an 
allosaurid.

Forcing Dilong to be a compsognathid (as suggested by Olshevsky I believe, 
online) adds 10 steps.  

Forcing Eoraptor to be a basal theropod, but herrerasaurids to be outside 
Eusaurischia (as in Ezcurra, 2006) adds 11 steps.

Forcing Indosaurus to be a tyrannosaurid (as in Chatterjee, 1978) adds 11 steps.

Forcing Coelurus to be basal to avetheropods (as in Paul, 1988) adds 11 steps.  
Oddly, instead of being just outside Avetheropoda, it moves to the base of the 
Tetanurae with Tugulusaurus.

Forcing Aristosuchus to be a compsognathid (as in Naish, 2002) adds 12 steps.  

Forcing Tyrannosaurus to be a maniraptoran (as in Sereno, 1997) adds 12 steps.

Forcing coelophysoids to be ceratosaurs (as in Gauthier, 1986) adds 13 more 
steps.  In these trees, Cryolophosaurus and "Dilophosaurus" sinensis are basal 
tetanurines.

Forcing Elaphrosaurus to be a coelophysoid (as in Paul, 1988) adds 13 more 
steps.  It falls out with Spinostropheus between Liliensternus+Lophostropheus 
and Zupaysaurus+derived coelophysids.  Amusingly right where Paul put it.

Forcing spinosaurids to be more basal than Piatnitzkysaurus, 
Eustreptospondylus, Afrovenator and Torvosaurus (as in Holtz, 2000) adds 13 
steps.  Spinosauroidea becomes paraphyletic to Avetheropoda, with Torvosaurus, 
Eustreptospondylidae, Poekilopleuron and Piatnitzkysaurus successively closer 
to it.

Forcing Sinraptor to be closer to carcharodontosaurids than to Allosaurus (as 
in Coria and Currie, 2002) adds 13 steps.  I  these trees, Neovenator and 
Acrocanthosaurus are carcharodontosaurids, while Allosaurus is the basal 
carnosaur.

Forcing Compsognathus to be a tyrannosauroid (as in Olshevsky, 1991) adds 13 
steps.  This also makes Sinosauropteryx, Guanlong, Coelurus, Dilong, Mirischia 
and Aristosuchus tyrannosauroids.

Forcing Ilokelesia to be outside Noasauridae+Abelisauridae (as in Coria and 
Salgado, 2000) adds 14 steps.

Forcing Afrovenator to be sister to Allosauroidea (as in Rauhut, 2003) adds 14 
steps.  Dubreuillosaurus and Piveteausaurus stay in Spinosauroidea.

Forcing Ornitholestes outside Tyrannoraptora (as in Holtz, 1994 and others) 
adds 14 steps.

Forcing Procompsognathus to be outside Dinosauria (as in Allen, 2004) adds 15 
steps.

Forcing Deltadromeus to be closer to birds than Ornitholestes (as in Sereno et 
al., 1996) adds 15 more steps, mainly from moving Ornitholestes down the tree.

Forcing Piatnitzkysaurus to be an allosaurid (as in Molnar et al., 1981) adds 
15 steps.



Near certainly untrue-

Forcing Ceratosaurus to be closer to tetanurines than abelisaurids are (as in 
Carrano and Sampson, 1999) adds 18 steps.  Elaphrosaurs and noasaurids stay 
with abelisaurids.

Forcing Compsognathus outside Avetheropoda (as in Paul, 1988) adds 18 steps.

Forcing Piatnitzkysaurus to be an abelisaur (as in Currie and Zhao, 1993) adds 
21 steps.

Forcing Monolophosaurus to be sister to Guanlong (as suggested by Carr, 2006) 
adds 22 steps.  The pair are placed as coelurosaurs just outside Tyrannoraptora.

Forcing Cryolophosaurus to be a carnosaur (as in Sereno et al., 1996) adds 23 
more steps.  It changes the tree a lot, with most basal tetanurines moved into 
Carnosauria.

Forcing Monolophosaurus to be a tyrannosauroid (as suggested by me on the DML) 
adds 24 steps.  In these trees, Guanlong and Dilong are still closer to birds 
than to tyrannosaurids, though Siamotyrannus is a tyrannosauroid.

Forcing Guanlong to be a carnosaur (as in Carr, 2006) adds 24 steps.  
Siamotyrannus, Fukuiraptor, Dilong and Aristosuchus are also carnosaurs in 
these trees.

Forcing Tyrannosaurus to clade with ornithomimosaurs (as in Huene, 1923) adds 
24 steps.

Forcing Torvosaurus to be a ceratosaur (as in Britt, 1991) adds 26 steps.  It 
emerges outside the elaphrosaur-ceratosaurid-abelisauroid clade, but other 
spinosauroids stay in Tetanurae.

Forcing Guanlong to be a carnosaur sister taxon to Monolophosaurus (as in Carr, 
2006) adds 27 steps.

Forcing Acrocanthosaurus to be a spinosaurid (as in Walker, 1964)adds 30 steps. 
 Spinosaurids are moved into Carnosauria.

Forcing Elaphrosaurus to be an ornithomimosaur (as in Galton, 1982) adds 33 
steps.

Forcing abelisaurids to be sister to carcharodontosaurids (as in Novas, 1997) 
adds 35 steps.  The clade ends up in Ceratosauria, and Acrocanthosaurus and 
Neovenator remain in Carnosauria.

Forcing Zupaysaurus to be a tetanurine (as in Arcucci and Coria, 2003) adds 37 
steps.

Forcing Megaraptor to be a dromaeosaurid (as suggested in Rauhut, 2003) adds 38 
steps.  

Forcing Abelisaurus to be a carcharodontosaurid (as suggested by Lamanna et 
al., 2002) adds 39 steps.

Forcing Monolophosaurus to be an ornitholestiid (as in Paul, 2002) adds 40 
steps.  Ornitholestes and Proceratosaurus are moved outside Avetheropoda with 
Monolophosaurus.

Forcing Ornitholestes to be an allosaurid (as in Paul, 1988) adds 41 steps.

Forcing Sinosauropteryx outside Coelurosauria (as in Longrich, DML 2000) adds 
42 steps.

Forcing the less extensive Carnosauria of Molnar et al., 1990 
(Piatnitzkysaurus, Allosaurus, carcharodontosaurids and tyrannosaurids) adds 44 
steps.  Spinosauroids, Monolophosaurus and Sinraptor form successively closer 
outgroups to Carnosauria.  The Allosauridae of Molnar is paraphyletic to 
tyrannosaurids, with Piatnitzkysaurus most basal and carcharodontosaurids most 
closely related.

Russell and Dong's (1993) crazy topology of 
(Coelophys(Ornithomimidae(Baryonyx(Sinraptoridae(Allosauridae(Tyrannosauridae,Dromaeosauridae))))))
 adds 45 steps, mainly from moving ornithomimosaurs so basally.

Bakker et al.'s (1988) topology of 
(Cerato(Allo(Dromaeo(Acrocantho(Ornithimimidae,Tyrannosauridae))))) adds 53 
steps.

Forcing abelisaurids to be tetanurines closer to birds than Torvosaurus (as in 
Forster, 1999) adds 55 steps.  Noasaurids and Deltadreomeus clade with 
abelisaurids here.

Forcing the Allosauria of Paul (Ornitholestes, Proceratosaurus, Allosaurus, 
carcharodontosaurids and tyrannosaurids) adds 56 more steps.  Though 
constrained to be outside Paul's Allosauria, Piatnitzkysaurus, spinosauroids, 
Monolophosaurus and sinraptorids end up forming successively closer outgroups 
to it.  Paul's ornitholestiines end up closer to tyrannosaurids than to 
allosaurids, and include Dilong and Guanlong as well.

Forcing abelisauroids to be torvosaurids within Tetanurae (as in Paul, 1988) 
adds 57 steps.  Spinosaurids and Monolophosaurus join the torvosaur part of 
this clade, while elaphrosaurs hang back with Ceratosaurus.

Forcing spinosaurids to be coelophysoids closest to dilophosaurs (as in Paul, 
1988) adds 60 steps.  

Forcing a traditional Carnosauria of sinraptorids, Torvosaurus, spinosaurids, 
abelisaurids, eustreptospondylids, carcharodontosaurids, allosaurids and 
tyrannosaurids (as in Kurzanov, 1989) adds 82 steps.  Unlike Kurzanov's 
topology, abelisaurids emerge as most basal, Spinosauroidea is intact, and 
sinraptorids are sister to allosaurids.

Forcing paravians outside of Dinosauria (as Feduccia, Martin, etc. suggest) 
adds 87 steps.  Even though I specified Allosaurus as a dinosaur, all 
coelurosaurs are more parsimoniously "birds" when this happens.

Forcing spinosaurids and birds to be sister taxa (as in Elzanowski and 
Wellnhofer, 1992) adds 93 steps.  The clade ends up in a very basal coelurosaur 
position (ignoring for the moment that birds aren't with the rest of what's 
normally Paraves).

Forcing Huene's original Carnosauria-Coelurosauria phylogeny of 1923, where 
coelophysids, ornithomimids, tyrannosaurids and Ceratosaurus are coelurosaurs, 
while Allosaurus and spinosauroids are carnosaurs adds 97 steps.  

Forcing Torvosaurus and Poekilopleuron to be more closely related to 
Plateosaurus than to coelophysoids and coelurosaurs (as in Galton and Jensen, 
1979) adds 101 steps.  The rest of Spinosauroidea, as well as Ceratosauria and 
Carnosauria end up joining the two 'megalosaurs' to form a large-bodied 
theropod clade sister to Sauropodomorpha.  Within Saurischia, coelophysoids are 
more closely related to this clade than coelurosaurs.

Mickey Mortimer