Sereno's (2005) new definitions

[Michael Mortimer <mickey_mortimer111@msn.com>]

Taxon Search reveals the new definitions in press by Sereno, which he has 
requested commentary on.  Here are my thoughts, including some suggestions 
for modifications.

First order revisions involving substituting species for genera-
These were welcome definitions, as all definitions should use species-level 
specifiers if they were published subsequent to the draft Phylocode.  But 
they are nothing new, just more specific versions of prior definitions.

Coelophysidae
(Coelophysis bauri + Procompsognathus triassicus)

Procompsognathinae
(Procompsognathus triassicus <- Coelophysis bauri)

Coelophysinae
(Coelophysis bauri <- Procompsognathus triassicus)

Ceratosauria
(Ceratosaurus nasicornis <- Passer domesticus)

Allosauroidea
(Allosaurus fragilis <- Passer domesticus)

Therizinosauroidea
(Beipiaosaurus inexpectus + Therizinosaurus cheloniformis)

Caenagnathidae
(Chirostenotes pergracilis <- Oviraptor philoceratops)

Oviraptoridae
(Oviraptor philoceratops <- Chirostenotes pergracilis)

Ornithurae
(Passer domesticus <- Archaeopteryx lithographica)

Ornithothoraces
(Sinornis santensis + Passer domesticus)

Enantiornithes
(Sinornis santensis <- Passer domesticus)

Euornithes
(Passer domesticus <- Sinornis santensis)

Hesperornithes
(Hesperornis regalis <- Passer domesticus)

Carinatae
(Ichthyornis dispar + Passer domesticus)

First order revisions involving adding new specifiers-

Coelophysoidea
(Coelophysis bauri <- Ceratosaurus nasicornis, Carnotaurus sastrei, Passer 
domesticus)
The presence of Passer as a specifier is useful, but Carnotaurus seems 
superfluous, since there has never been a (Ceratosaurus (Passer 
(Carnotaurus, Coelophysis))) topology suggested.

Noasauridae
(Noasaurus leali <- Coelophysis bauri, Carnotaurus sastrei, Passer 
domesticus)
Neither Coelophysis nor Passer seem particularily useful here, as noasaurids 
have been universally considered abelisaurians since Paul (1988) and 
Bonaparte et al. (1990), and were never placed anywhere else besides the 
polyphyletic pre-cladistic Coelurosauria.

Abelisauridae
(Carnotaurus sastrei <- Coelophysis bauri, Noasaurus leali, Passer 
domesticus)
For the same reason noted above for Noasauridae, the inclusion of 
Coelophysis and Passer as specifiers seems useless.  Abelisaurus comahuensis 
should be the internal specifier (Phylocode 11.8), so I don't accept this 
definition.  If Abelisaurus is a carcharodontosaurid, a possibility 
suggested by Lamanna et al. (2002), we'd have Abelisauridae without 
Abelisaurus under Sereno's definition.

Tetanurae
(Passer domesticus <- Ceratosaurus nasicornis, Carnotaurus sastrei)
The addition of Carnotaurus as an external specifier actually seems 
counter-productive to me.  Tetanurae was designed as a stem away from 
Ceratosaurus, and abelisaurids were not explicitly discussed (having been 
named only a year prior).  In fact, technically, Indosaurus and Indosuchus 
were classified as tetanurines by Gauthier (1986), since he lists them as 
carnosaurs.  If abelisaurids are megalosauroids (as in Paul, 1988), it 
shouldn't stop megalosauroids from being tetanurines.

Spinosauroidea
(Spinosaurus aegyptiacus + Torvosaurus tanneri, - Allosaurus fragilis, 
Passer domesticus)
According to ICZN rules, the superfamily containing Megalosaurus should be 
called Megalosauroidea, as Megalosauridae has priority over Spinosauridae.  
Despite worries about the validity of Megalosaurus, it has been universally 
placed with 'torvosaurids' and 'eustreptospondylids' when the two groups are 
viewed as part of a larger clade exclusive of birds (Allain, 2002; Holtz et 
al., 2004).  Furthermore, a stem-based definition (as given for 
Spinosauroidea by Holtz et al., 2004) is preferrable, as it's possible 
megalosaurids are paraphyletic with respect to spinosaurids (e.g. Sereno et 
al., 1994; Rauhut, 2000).  Spinosauroidea can be a node-based clade within a 
stem-based Megalosauroidea, letting both names co-exist, though nested 
-oidea clades are distinctly un-Linnaean.  Sereno's suggestion to let 
Spinosauria be the stem-based clade if so needed in the future is more 
elegant, as would be using Megalosauria.  Both names already exist, and 
Megalosauria has a century of priority.  I do like Sereno's idea to use 
Allosaurus as an external specifier, as this would ensure carnosaurs aren't 
megalosauroids (though the inverse is possible, as in Rauhut, 2000).  I 
would suggest (Megalosaurus bucklandii <- Ceratosaurus nasicornis, 
Allosaurus fragilis, Passer domesticus) as a definition for Megalosauroidea, 
and just drop Spinosauroidea altogether.

Torvosauridae
(Torvosaurus tanneri <- Spinosaurus aegyptiacus, Allosaurus fragilis, Passer 
domesticus)
According to ICZN rules, the family containing Megalosaurus should be called 
Megalosauridae, as it has priority over Spinosauridae.  Megalosaurus has 
been considered to fall within the definition of Sereno's Torvosauridae by 
almost all workers with the exception of Holtz (who placed Megalosaurus and 
Torvosaurus as successively further from birds in 1994, and successively 
closer to birds in 2000) and Kurzanov (who placed Torvosaurus further from 
spinosaurids and allosaurids than Megalosaurus).  So I feel this definition 
is good, but would advocate a first order revision of Holtz et al.'s (2004) 
definition of Megalosauridae as- (Megalosaurus bucklandii <- Spinosaurus 
aegyptiacus, Allosaurus fragilis, Passer domesticus).  Then Megalosauridae 
would have priority over Torvosauridae if the two refer to the same clade.

Eustreptospondylinae
(Eustreptospondylus oxoniensis <- Torvosaurus tanneri, Spinosaurus 
aegyptiacus, Allosaurus fragilis)
Holtz et al.'s (2004) definition only includes Megalosaurus bucklandii as an 
external specifier, and I agree with Sereno that additional ones are useful 
in case eustreptospondylines are closer to allosaurids than to 
megalosauroids (Paul, 1988; Kurzanov, 1989; Molnar et al., 1990; Holtz, 
2000).  Not sure about Spinosaurus' utility, since I've never seen 
(Torvosaurus,Allosaurus(Spinosaurus, Eustreptospondylus)) suggested.

Spinosauridae
(Spinosaurus aegyptiacus <- Torvosaurus tanneri, Allosaurus fragilis, Passer 
domesticus)
This is the same as Holtz et al.'s (2004) except for the substitution of 
Torvosaurus for Megalosaurus.  Sereno's may have an advantage that 
Torvosaurus is more often included in analyses than Megalosaurus (Sereno, 
1999; Rauhut, 2000; Allain, 2002), but if Megalosaurus is used as the 
internal specifier for megalosaurids (as it must be) there is an unevenness 
involved with using Torvosaurus as the external specifier for their sister 
clade.  Incidentally, I believe a node-based Spinosauridae may be a better 
idea, for this stem-based one could include Chilantaisaurus (Rauhut, 2000), 
coelophysoids (Paul, 1988), tyrannosaurids (Walker, 1964), 
carcharodontosaurids (Bonaparte et al., 1990) and other taxa proposed to be 
more closely related to Spinosaurus than megalosaurids, allosaurids or 
birds.

Neotetanurae
(Sinraptor dongi + Carcharodontosaurus saharicus + Allosaurus fragilis + 
Passer domesticus)
This differs from the original definition (Sereno, 1998) by adding Sinraptor 
and Carcharodontosaurus as internal specifiers.  I suppose it would preserve 
content better if sinraptorids or carcharodontosaurids end up just basal to 
carnosaurs + coelurosaurs (Paul, 1988; Coria and Salgado, 1995; Longrich, 
2001; Paul, 2002).  However, if carcharodontosaurids are ceratosaurs 
(Bonaparte et al., 1990) or sinraptorids are megalosauroids (Kurzanov, 
1989), the original intent of Neotetanurae would be lost.  The latter two 
situations seem less likely than the former, so Sereno's redefinition may be 
more advantageous than not.

Sinraptoridae
(Sinraptor dongi <- Allosaurus fragilis, Carcharodontosaurus saharicus, 
Passer domesticus)
This differs from Holtz et al.'s (2004) definition by including Passer as an 
external specifier, which I view as superfluous, since a (Allosaurus, 
Carcharodontosaurus (Sinraptor, Passer)) topology has never been advocated.  
Megalosaurus might be a better choice for a tertiary external specifier, to 
cover traditional phylogenies prior to 1993.

Carcharodontosauridae
(Carcharodontosaurus saharicus <- Allosaurus fragilis, Sinraptor dongi, 
Passer domesticus)
This also differs from Holtz et al.'s (2004) definition by including Passer 
as an external specifier.  The only times carcharodontosaurids have been 
placed in Coelurosauria is when tyrannosaurids were as well (Bakker et al., 
1988; Paul, 1988), and they have often been placed closer to tyrannosaurids 
than to Allosaurus or Passer (Paul, 1988; Kurzanov, 1989; Molnar et al., 
1990).  So Tyrannosaurus would be a more useful tertiary external specifier 
than Passer.  The remote possibility of a relationship to ceratosaurs 
(Bonaparte et al., 1990) might suggest Carnotaurus should be used as an 
additional external specifier.

Allosauridae
(Allosaurus fragilis <- Sinraptor dongi, Carcharodontosaurus saharicus, 
Passer domesticus)
A third carnosaurian family redefinition that differs from Holtz et al.'s 
(2004) by including Passer as an external specifier.  In this case, I agree 
it's useful for cases like Paul (2002), Longrich (2001) and Coria and 
Salgado (1995).  I wonder if Tyrannosaurus might be a useful external 
specifier as well, as tyrannosaurids and allosaurids have often been posited 
as sister groups (Paul, 1988; Kurzanov, 1989; Molnar et al., 1990).  
However, in all these cases, Carcharodontosaurus was seen as an intermediate 
between Allosaurus and Tyrannosaurus, which would keep tyrannosaurids from 
being allosaurids under Sereno's and Holtz et al.'s definitions.

Coelurosauria
(Passer domesticus <- Allosaurus fragilis, Sinraptor dongi, 
Carcharodontosaurus saharicus)
This differs from the standard definition by including Sinraptor and 
Carcharodontosaurus as external specifiers.  Sinraptor's inclusion is 
superfluous, as mentioned under Sinraptoridae above.  If Carcharodontosaurus 
is a tyrannosauroid (Paul, 1988; Kurzanov, 1989; Molnar et al., 1990), this 
redefinition would exclude tyrannosauroids from Coelurosauria.  This 
wouldn't necessarily be a bad thing, especially as most of the evidence 
indicates tyrannosauroids are basal to anything else called a coelurosaur 
nowdays except perhaps Tugulusaurus, Coelurus, Tanycolagreus, Calamosaurus, 
Proceratosaurus, Bagaraatan and Dryptosaurus.  The latter is especially true 
if carcharodontosaurids have anything to do with tyrannosaurids, as that set 
of character transformations leaves tyrannosaurids developing their 
coelurosaurian characters convergently with maniraptoriformes.  So the only 
topology which suffers by using Carcharodontosaurus as an external specifier 
is Paul's (1988), which would exclude compsognathids, Ornitholestes and 
Coelurus from Coelurosauria (in addition to Proceratosaurus and 
tyrannosauroids).  But since Allosaurus would have to be a coelurosaur for 
any of these latter taxa to be coelurosaurs in Paul's topology, 
Carcharodontosaurus' inclusion as an external specifier doesn't add any 
further harm.  One thing I object to is the use of Passer as an internal 
specifier for Coelurosauria, as birds were not originally classified as 
coelurosaurs in Huene, 1914 or by anyone until the 1970's at least.  Huene 
included what would today be called coelophysids, coelurids, compsognathids, 
Ornitholestes and ornithomimids.  The best internal specifier for 
Coelurosauria in my opinion is Ornithomimus.  It's always been a 
coelurosaur, and has always been placed closer to birds than Allosaurus 
(unlike Compsognathus, Coelurus or Ornitholestes- Paul, 1988; Novas, 1992).  
Thus I would suggest (Ornithomimus velox <- Allosaurus fragilis, 
Carcharodontosaurus saharicus) as a definition for Coelurosauria.

Tyrannosauroidea
(Tyrannosaurus rex <- Ornithomimus edmontonicus, Troodon formosus, 
Velociraptor mongoliensis)
This is a revision of Holtz's (2004) definition, substituting Ornithomimus 
edmontonicus for O. velox, Velociraptor for Deinonychus, and Troodon for 
Allosaurus.  Ornithomimus velox is the better ornithomimosaur specifier, as 
discussed under Maniraptoriformes.  Being the namesake of Deinonychosauria, 
Deinonychus is a better specifier than Velociraptor, but Dromaeosaurus might 
be better than either due to its priority.  Still, I have no problem with 
Deinonychus.  Replacing Allosaurus with Troodon was a bad choice, since 
numerous topologies have had allosaurids sister to tyrannosaurids but none 
I'm aware of have had troodontids sister to tyrannosaurids.  The only other 
taxa that have been suggested to be sister to tyrannosauroids are 
spinosaurids (Walker, 1964), carcharodontosaurids (Paul, 1988; Kurzanov, 
1989; Molnar et al., 1990) and compsognathids (Olshevsky), but the former 
was explicitly placed in Tyrannosauroidea and I don't see placing the others 
in Tyrannosauroidea as counter-intuitive.

Tyrannosauridae
(Gorgosaurus libratus + Albertosaurus sarcophagus + Tyrannosaurus rex)
This is a revision of Holtz's (2004) definition, after deleting 
Daspletosaurus and Tarbosaurus.  I agree with Sereno that their inclusion is 
useless, as none are ever placed outside the clade defined here.

Tyrannosaurinae
(Tyrannosaurus rex <- Gorgosaurus libratus, Albertosaurus sarcophagus)
This is a revision of Currie et al.'s (2003) definition, adding Gorgosaurus 
as an external specifier.  It does do a better job at maintaining stability 
if albertosaurines are paraphyletic.  And since Tyrannosauridae has multiple 
internal specifiers, this isn't part of a node-stem triplet, so I 
tentatively agree with Sereno.

Ornithomimosauria
(Ornithomimus edmontonicus <- Tyrannosaurus rex, Shuvuuia deserti, 
Therizinosaurus cheloniformis, Oviraptor philoceratops, Troodon formosus, 
Passer domesticus)
This revises Sereno's (1998) definition which only included Passer as an 
external specifier, and Osmolska's (1997) which only included Troodon.  The 
only other suggested definition is Makovicky et al.'s (2004), which is a 
node-based first level redefinition of Padian et al.'s (1999) using 
Pelecanimimus and Ornithomimus edmontonicus.  Sereno has a good point that a 
taxon directly outside Pelecanimimus + Ornithomimus (such as Deinocheirus in 
Makovicky et al. 2004) should be an ornithomimosaur, but wouldn't be using 
the node-based definition.  Furthermore, Pelecanimimus is sometimes closer 
to alvarezsaurids in my analysis, which would force alvarezsaurids inside 
Ornithomimosauria.  So I agree Sereno's definition is superior, and 
unproblematic as far as I can tell besides the use of Ornithomimus 
edmontonicus instead of O. velox.  See the discussion under 
Maniraptoriformes for details.

Alvarezsauridae
(Shuvuuia deserti <- Tyrannosaurus rex, Ornithomimus edmontonicus, 
Therizinosaurus cheloniformis, Oviraptor philoceratops, Troodon formosus, 
Passer domesticus)
This revises Sereno's earlier (1999) definition by adding 
non-ornithomimosaur external specifiers, a very good choice considering 
alvarezsaurids may be basal maniraptorans, oviraptoriformes, paravians, 
avialans or ornithurines.  It appears to cover all the published topologies, 
though the inclusion of Tyrannosaurus seems superfluous, as it's never been 
placed more closely to alvarezsaurids than at least one of the other 
external specifiers.  Once again though, Sereno didn't use an eponymous 
taxon - Alvarezsaurus calvoi in this case.  His rationale is that Shuvuuia 
is more completely known and "clearly related", but if the relationship is 
so clear, why not just use Alvarezsaurus?  In his new paper, Sereno states 
"Well-known (and/or more complete), nested specifiers are critical because 
they are least likely to shift significantly in phylogenetic position", but 
if Alvarezsaurus shifts outside Sereno's defined Alvarezsauridae (such as in 
Lu et al., 2002), we'd have to redefine the family anyway.  I suggest 
(Alvarezsaurus calvoi <- Ornithomimus velox, Therizinosaurus cheloniformis, 
Oviraptor philoceratops, Troodon formosus, Passer domesticus) as a first 
order redefinition of Alvarezsauridae.

Mononykinae
(Mononykus olecranus + Shuvuuia deserti)
Sereno claims this is the first definition suggested for the clade, but 
Chiappe et al. (1998) defined it earlier as "the common ancestor of 
Mononykus, Shuvuuia, and Parvicursor, plus all their descendants."  As no 
one has examined mononykine interrelationships, I think it's best to include 
all three well known genera in the definition.  According to ICZN rules, the 
clade should be called Parvicursorinae though, as it has priority over 
Mononykinae (1996 vs. 1998).

Therizinosauria
(Therizinosaurus cheloniformis <- Tyrannosaurus rex, Ornithomimus 
edmontonicus, Shuvuuia deserti, Oviraptor philoceratops, Troodon formosus)
This is a redefinition of Russell's (1997) attempt, deleting all internal 
specifiers except Therizinosaurus and adding Tyrannosaurus and Shuvuuia as 
external specifiers.  The first change is good (why bother forcing 
Alxasaurus, Enigmosaurus, Erlikosaurus, Segnosaurus and Nanshiungosaurus to 
be in the clade?), but the second doesn't appear very pointful to me.  I'm 
unaware of any suggested topology- (Ornithomimus, Oviraptor, Troodon 
(Tyrannosaurus, Shuvuuia, Therizinosaurus)).  If anything, I might suggest 
using Plateosaurus and Stegosaurus as additional specifiers instead, just to 
ensure views like Paul's (1984, 1988), Sereno's (1992) and Olshevsky's are 
covered.

Oviraptorosauria
(Oviraptor philoceratops <- Tyrannosaurus rex, Ornithomimus edmontonicus, 
Therizinosaurus cheloniformis, Troodon formosus, Passer domesticus)
This adds a lot of non-avian specifiers to Maryanska et al.'s (2002) 
definition, which was sorely needed.  The only issue I have (besides O. 
edmontonicus vs. O. velox) is the inclusion of Tyrannosaurus as an external 
specifier, as I'm unaware of any suggested topology positioning it closer to 
Oviraptor than the other taxa.  Maybe Dromaeosaurus albertensis would be 
better, to cover the possibility figured by Barsbold et al. (1990).

Troodontidae
(Troodon formosus <- Ornithomimus edmontonicus, Velociraptor mongoliensis, 
Passer domesticus)
This differs from Sereno's (1998) definition which only had Velociraptor as 
an external specifier, and Senter et al.'s (2004), which had Mononykus, 
Therizinosaurus and Oviraptor as additional external specifiers.  Keeping 
Oviraptor seems useful, in the case Osmolska and Barsbold (1990), Russell 
and Dong (1993) or Norell et al. (2001) are correct.  I'm not aware of any 
topology placing Shuvuuia or Therizinosaurus closer to Troodon than the 
other taxa, but they both seem relatively plausible as troodontid sister 
taxa.  Another external specifier that might be useful considering recent 
discoveries is Archaeopteryx.

Deinonychosauria
(Troodon formosus + Velociraptor mongoliensis, - Ornithomimus edmontonicus, 
Passer domesticus)
There have been two basic suggested definitions for Deinonychosauria, one 
stem-based (Deinonychus <- Passer) by Padian (1997) and the other node based 
(Troodon + dromaeosaurids) by Sereno (1997).  This is a modification of the 
latter, but explicitly excludes birds and ornithomimosaurs.  I prefer 
Padian's definition because it is based on the eponymous genus, and Colbert 
and Russell (1969) did not originally specify the inclusion of troodontids.  
They only include dromaeosaurids in the taxon, and only mention 
Dromaeosaurus, Deinonychus and Velociraptor as members of that family.

Unenlagiinae
(Unenlagia comahuensis <- Microraptor zhaoianus, Velociraptor mongoliensis, 
Dromaeosaurus albertensis, Passer domesticus)
This replaces Makovicky et al.'s (2005) definition, which only included 
Velociraptor as an external specifier.  As I detailed in 
http://dml.cmnh.org/2005Oct/msg00372.html , I don't find support for a clade 
of Unenlagia, Rahonavis and Buitreraptor exclusive of dromaeosaurids, 
troodontids and birds very convincing.  I also don't find the position of 
these taxa within Eumaniraptora to be very stable, so I'd recommend adding 
Troodon and Archaeopteryx as external specifiers.

First order redefinitions involving different specifiers-

Maniraptoriformes
(Ornithomimus edmontonicus + Passer domesticus)
Maryanska et al. (2002) used Ornithomimus velox, the type species, as 
dictated by Phylocode.  To illustrate why this is a good idea, consider the 
fact that Makovicky et al. (2004) synonymized O. edmontonicus with 
Dromiceiomimus.  They listed the species as O. edmontonicus, but 
brevitertius has priority, so the species should be Ornithomimus 
brevitertius.  Ornithomimus velox, on the other hand, remains valid.  
Makovicky et al. also considered the possibility O. brevitertius (as O. 
edmontonicus) may be a junior synonym of O. velox, and deCourten and Russell 
(1985) suggested it (again as O. edmontonicus) may warrant generic 
separation from O. velox if the specimen they describe is properly referred 
to the latter species.  Then Sereno's redefinitions of taxa eponymous with 
Ornithomimus would not be based on Ornithomimus.  Sereno claims O. 
edmontonicus is the taxon represented by most analyses, not O. velox, but 
only the TWG matrix (from Ji et al., 2003 onward) and Kobayashi's work 
(Kobayashi and Lu, 2003; Kobayashi, 2004; Kobayashi and Barsbold, 2005; 
Kobayashi and Barsbold, 2005) have used Ornithomimus as an OTU, and the 
latter uses both species as references.  So this is not a valid rationale.

Maniraptora
(Passer domesticus <- Ornithomimus edmontonicus)
This differs from Maryanska et al.'s (2002) definition in the same way 
Sereno's (2005) Maniraptoriformes does, so the above comments apply here as 
well.

Caenagnathoidea
(Chirostenotes pergracilis + Oviraptor philoceratops)
This is the same as Maryanska et al.'s (2002) definition, except it replaces 
Caenagnathus with Chirostenotes.  This is a poor decision, as the taxa are 
not definitely synonymous.  Chirostenotes pergracilis and Elmisaurus elegans 
co-occur in the same formation, and the only reason Caenagnathus is 
synonymized with pergracilis instead of elegans is size.  Until taxonomic 
problems are solved for caenagnathids, it's best to associate the family 
with it's eponymous species, even ignoring Phylocode rules.

Dromaeosauridae
(Dromaeosaurus albertensis <- Ornithomimus edmontonicus, Troodon formosus, 
Passer domesticus)
This modifies Sereno's (1998) definition by using the eponymous genus (yay!) 
as an internal specifier, and adding Ornithomimus and Passer as external 
specifiers.  This contrasts with the two published node-based definitions- 
(Dromaeosaurus albertensis + Velociraptor mongoliensis) by Padian et al. 
(1999), and (Microraptor zhaoianus + Sinornithosaurus millenii + 
Velociraptor mongoliensis) by Norell et al. (2004).  Norell et al.'s is 
problematic for not including Dromaeosaurus, and for explicitly including 
Microraptor.  The latter would make pygostylians dromaeosaurids in Mayr et 
al.'s (2005) topology, and troodontids and birds dromaeosaurids in my 
corrected version of the TWG matrix.  Padian et al.'s is equivalent to the 
use of Dromaeosauridae prior to 1999, but microraptorians weren't known and 
'unenlagiines' were considered avialans.  Once microraptorians were 
discovered, they were assigned to Dromaeosauridae by most authors except 
Senter et al. (2004).  So using a definition which won't probably exclude 
them does have some precedence, but Sereno's should have more external 
specifiers if it is adopted, namely Archaeopteryx (Paul, 1988; 2002), 
Tyrannosaurus (Matthew and Brown, 1922; Russell and Dong, 1993), 
Ornitholestes (Makovicky, 1995) and Oviraptor (Barsbold et al., 1990).  I 
actually think this stem-based definition is better than Padian et al.'s 
node-based one for the additional reason that exactly which taxa fall into 
the (Dromaeosaurus + Velociraptor) crown is uncertain.  With the recognition 
that velociraptorine sensu lato characters are symplesiomorphic for 
dromaeosaurs (e.g. found in Bambiraptor and microraptorians), taxa like 
Deinonychus or Saurornitholestes could easily fall just outside the defined 
node, not to mention more fragmentary taxa such as Variraptor or Pyroraptor. 
 So I advocate the following redefinition of Sereno's- (Dromaeosaurus 
albertensis <- Tyrannosaurus rex, Ornithomimus velox, Oviraptor 
philoceratops, Troodon formosus, Archaeopteryx lithographica, Passer 
domesticus).

Velociraptorinae
(Velociraptor mongoliensis <- Microraptor zhaoianus, Dromaeosaurus 
albertensis, Unenlagia comahuensis, Passer domesticus)
Since Sereno uses a stem-based definition for Dromaeosauridae, the node stem 
triplet advocated by Padian et al. (1999) no longer functions.  I don't have 
a problem with this definition.  It's slightly superior to Padian et al.'s 
by adding Microraptor, Unenlagia and Passer, but I think the chance of a 
(Dromaeosaurus (Velociraptor, Passer)) topology is very low, and wouldn't 
have a problem with microraptorians being velociraptorines.  But since 
Unenlagiinae is active, I suppose it's best to keep subfamilies out of 
subfamilies.

Dromaeosaurinae
(Dromaeosaurus albertensis <- Microraptor zhaoianus, Velociraptor 
mongoliensis, Unenlagia comahuensis, Passer domesticus)
The comments for Velociraptorinae apply here as well, though I consider 
Passer even less necessary in this definition.

Second order redefinitions-

Ornithomimidae
(Ornithomimus edmontonicus <- Pelecanimimus polyodon, Harpymimus 
okladnikovi, Shenzhousaurus orientalis, Garudimimus brevipes)
About time someone defined Ornithomimidae to be equivalent to its usage over 
the last two decades.  Sereno had the only prior attempts, defining it as 
equivalent to Arctometatarsalia in 1998 and Ornithomimosauria in 1999.  I'm 
very happy with this defintion besides the use of Ornithomimus edmontonicus 
instead of O. velox.

Therizinosauridae
(Therizinosaurus cheloniformis + Nothronychus mckinleyi + Neimongosaurus 
yangi)
Previously, the family has been given a stem-based definition equivalent to 
Maniraptora (Sereno, 1998) or Therizinosauria (Sereno, 1999), or a different 
node-based one- (Therizinosaurus + Segnosaurus + Erlikosaurus + 
Nanshiungosaurus) in Zhang et al. 2001; (Therizinosaurus + Erlikosaurus) in 
Clark et al. 2004.  Unfortunately, the exact interrelationships between 
therizinosauroids are still far from clear.  The only analysis has been 
Clark et al. (2004), which found a large polytomy of taxa more derived than 
Beipiaosaurus.  When Erlianosaurus, Falcarius and new information about 
Beipiaosaurus (Xu et al., 2003) and Nothronychus (Gillette et al., 2005; 
Kirkland et al., 2005) is added, more structure exists- (Falcarius 
(Beipiaosaurus (Alxasaurus (Nanshiungosaurus (Neimongosaurus, Nothronychus 
(Erlianosaurus, Erlikosaurus, Segnosaurus, Therizinosaurus)))))).  
Enigmosaurus can go anywhere more derived than Beipiaosaurus.  According to 
this topology, the three suggested Therizinosauridae definitions would each 
encompass a different clade.  Since the point of Therizinosauridae seems to 
be to separate more derived taxa from Alxasaurus, I would actually recommend 
(Therizinosaurus cheloniformis <- Alxasaurus elesitaiensis).

Confuciusornithidae
(Confuciusornis sanctus <- Passer domesticus)
The only previous definition for this family is that of Chiappe et al. 
(1999), who used a node-base- (Confuciusornis sanctus + Changchengornis 
hengdaoziensis).  I agree with Sereno that as more members of the 
confuciusornithid stem are discovered, it would be ideal to be able to refer 
them to the family.  "Proornis" is a relevent example.  Another solution 
would be to use Confuciusornithiformes (Hou et al., 1995) for the stem.  I 
don't mind either possibility, but would suggest a stem-based definition 
include Enantiornis leali as an additional external specifier, to satisfy 
the Sauriurae crowd (such as the result of running Zhou and Zhang's [2005] 
complete matrix).  As a common stand-in for Pygostylia, Confuciusornis has 
had some odd relationships in a couple recent analyses (sister to 
Oviraptorosauria in Maryanska et al., 2002; sister to Microraptor in Mayr et 
al., 2005).  But without additional pygostylians tested in those analyses, 
it's difficult to justify using those sister taxa as additional specifiers 
for Confuciusornithidae.

New definitions-

Archaeopterygidae
(Archaeopteryx lithographica <- Passer domesticus)
This is the first published definition of Archaeopterygidae.  I would 
recommend a few additional external specifiers- Dromaeosaurus albertensis 
(Paul, 1988; 2002), Troodon formosus and Enantiornis leali (Martin, 
Feduccia, Hou, et al.).

New taxa-

Ornithomimiformes
(Ornithomimus edmontonicus <- Passer domesticus)
This rather blatantly replaces Arctometatarsalia, though the latter has 
never been more explicitly defined than (Ornithomimus <- Neornithes).  I 
actually think it's a better name for the clade in question, especially as 
the arctometatarsus evolved so many times and most authors would only 
include ornithomimosaurs in the taxon anyway (exceptions are Perez-Moreno et 
al., 1993-1994; possibly Russell and Dong, 1993; Holtz, 1994-2000; Sereno, 
1998-2005).  Also, Arctometatarsalia was originally defined as an 
apomorphy-based clade, and "arctometatarsalian" is a non-taxonomic word as 
well.  The problem is priority, as Arctometatarsalia has about a decade of 
it both nomenclaturally and definitionally.  What does Holtz think about 
Sereno's proposition?

Oviraptoriformes
(Oviraptor philoceratops <- Passer domesticus)
At last, a defined replacement for Enigmosauria (so we can stop having those 
tedious debates on the DML; but Enigmosauria is still listed on TaxonSearch- 
I told you people don't like pretending things don't exist).  I like this 
clade and definition, which even work in Maryanska et al. (2002) and Lu et 
al. (2002) where oviraptorosaurs are ornithurines.

Microraptorinae
(Microraptor zhaoianus <- Velociraptor mongoliensis, Dromaeosaurus 
albertensis, Unenlagia comahuensis, Passer domesticus)
Senter et al. (2004) erected Microraptoria, with a non-family level suffix 
to ensure ICZN rules are followed whether or not the clade was within 
Dromaeosauridae (depending on both the tree's topology and the latter's 
definition). They listed Microraptoridae and Microraptorinae as alternatives 
considered prior to deciding upon Microraptoria, and Microraptorini as a 
mispelling of Microraptoria. Makovicky et al. (2005) cited Senter et al. for 
Microraptorinae, but using the subfamily name was a mistake by Makovicky et 
al. (Headden, pers. comm. to Makovicky, 2005). Sereno (2005) erected 
Microraptorinae, attributing it to Senter et al. (though I don't believe the 
ICZN allows 'potential' names to be attributed in that way), giving it a 
more restrictive definition that supposedly overcomes Senter et al.'s 
perceived difficulties with family-level names. However, in topologies such 
as Mayr et al.'s (2005), Microraptorinae does not fall under any named 
family, which was just the problem Senter et al. anticipated when they 
rejected the name. On the other hand, Microraptoria includes Confuciusornis 
and potentially all pygostylians in Mayr et al.'s topology, unlike 
Microraptorinae. So Sereno's definition is more stable due to its extra 
specifiers. At present, however, the consensus is that Microraptorinae and 
Microraptoria are synonymous, and the latter has priority.

Mickey Mortimer