I'm glad the authors excluded Sigilmassasaurus from their Carcharodontosaurus OTU, though I would have been interested to see where the genus went when coded separately. Carrano et al. recognize the referred Chuandongocoelurus specimen is Elaphrosaurus-like, which I first determined back in 2001. Like me, they keep "yandonensis" and zigongensis separate. Interestingly, both are close to Yangchuanosaurus, and indeed the authors use the combination Yangchuanosaurus zigongensis. I finally learn why the "Dilophosaurus" sinensis skull I saw at the RTMP doesn't look like the illustration in Hu's paper- there are two sinensis specimens! Young Dilophosaurus specimen TMM 43646 differs from the other specimens, so may be a new species. We learn that new specimens confirm the snout elements are correctly referred to Marshosaurus. Turns out the type of Allosaurus? sibiricus is metatarsal II, not IV. Erectopus is stated to be a non-carcharodontosaurid carnosaur, though this isn't tested. Oddly though stated to be possibly metriacanthosaurid, it's placed above allosaurids in their phylogram. Gasosaurus lacks a proximally extensive anterior trochanter and is being restudied (cited as Hone pers. comm.). Carrano et al. retain "Poekilopleuron" schmidti's identification as a theropod humerus, but I disagree. Unquillosaurus' pubis is correctly flipped so that the supposed medial side is lateral, something which Headden and I found in 2002. They think it's possibly carcharodontosaurid, which is interesting because back then I said it resembled Acrocanthosaurus most in anterior view. The pelvis of Kryptops is carcharodontosaurid, probably belonging to Eocarcharia.
|Phylogeny from Carrano et al. (2012) taken from their figure 7A. Numbers to the left of nodes are unambiguous character support and numbers to the right are branch support.|
The phylogeny is interesting in several ways. Since Carrano and Benson have done good cladistic work in the past, I'm betting all codable characters are coded and with few miscodings, unlike several other recent tetanurine phylogenies (Holtz et al. 2004 Smith et al. 2007 *cough cough*). I will note all characters are unordered though, which should be fixed. One of the coolest things is supplementary table 1, which shows what prior analyses have used each character and how they phrased it.
Cryolophosaurus and "Dilophosaurus" sinensis are not only not coelophysoids, they're basal tetanurines. This has happened sometimes in my saurischian supermatrix as well, but I always figured it was due to not all characters being coded for all taxa.
Eustreptospondylus is a basal megalosaurid, and Streptospondylus may be related, but can go anywhere in Megalosauria outside Baryonychinae+Spinosaurinae and Megalosaurinae+Afrovenatorinae. The latter clade is another new name, since it turns out Dubreuillosaurus, Magnosaurus and Afrovenator aren't closely related to Streptospondylus or Eustreptospondylus (countering Rauhut's synonymization of the latter with Magnosaurus) after all. Other afrovenatorines are Leshansaurus (which Cau correctly placed in Megalosauridae), Poekilopleuron and Piveteausaurus. Duriavenator and Torvosaurus are megalosaurines.
Metriacanthosauridae turns out to be huge, with Yangchuanosaurus most basal, then a Metriacanthosaurinae structured as follows- (Shidaisaurus (Metriacanthosaurus, hepingensis (Sinraptor, Siamotyrannus). So Sinraptor is not monophyletic. Xuanhanosaurus is also a metriacanthosaurid, but not a member of the Metriacanthosaurus+Sinraptor clade. Saurophaganax's allosaurid position is finally tested and confirmed. Concavenator is a basal carcharodontosaurid by Eocarcharia. Lourinhanosaurus oddly emerges as a basal coelurosaur.
We get a "new" theropod taxon- "Saurocephalus" monasterii Muenster, 1836. As they say, it's "based on a recurved, serrated tooth from the Oxfordian Korallenkalk of the Lindner Berge, Hanover, Germany. Windolf (1997) transferred the taxon to Megalosaurus but Muenster’s tooth cannot be identified past the level of Theropoda indet."
There are a few things I disagree with or that could use work though. One is that additional basal and derived taxa are necessary to confirm some relationships. Placing Cryolophosaurus and sinensis in Tetanurae might be caused by excluding Sarcosaurus, Dracovenator and/or Zupaysaurus, for instance. And three coelurosaurs (Proceratosaurus, Compsognathus and Ornitholestes) are way too few to tell if megaraptorans are really carcharodontosaurids. I'd want Zuolong, Coelurus and/or Tanycolagreus, Guanlong, Dilong, Stokesosaurus, Eotyrannus and Huaxiagnathus included too. That might also reposition Lourinhanosaurus, as it is a basal carnosaur or sister to Avetheropoda with two more steps.
Also, alternative topologies are barely tested, except that making megalosauroids carnosaurs takes seven more steps (so not that unlikely, really) and several genera are stated to move to other places with one more step (Monolophosaurus a basal megalosauroid instead of non-orionidan; Poekilopleuron elsewhere in Megalosauroidea or in Allosauria; Cryolophosaurus outside Neotheropoda; etc.). I'd like to know how many more steps are needed for alternatives- non-orionidan piatnitzkysaurids or Xuanhanosaurus, coelophysoid Cryolophosaurus and/or sinensis, non-neotheropod sinensis, carnosaurian Monolophosaurus, 'afrovenatorine' Eustreptospondylus and/or Streptospondylus, Eustreptospondylus sister to Magnosaurus, afrovenatorine Duriavenator, monophyletic Sinraptor, coelurosaurian Siamotyrannus or megaraptorans, allosaurids sister to metriacanthosaurids+carcharodontosaurids, allosaurid Acrocanthosaurus and/or Neovenator, allosaurians sister to coelurosaurs, etc..
Megalosauria and Allosauria are never explicitly defined. Metriacanthosaurinae is listed as a new name, but Paul (1988) named it. It's odd they missed that, since the family is correctly attributed to him. I don't like their use of Averostra (for the cerato+tet node, when promaxillary fenestrae evolved before that), Neotheropoda (for the coelophysoid+ node instead of the cerato+tet node) or Allosauroidea (instead of Carnosauria). Though Carrano et al. agree Suchomimus is probably Cristatusaurus, they incorrectly call the latter indeterminate. Megalosaurus? "dapukaensis" and "tibetensis" are wrongly cited as Zhao 1986 instead of 1985. I would have liked matrix entries for the more complete taxa not included, like Erectopus, Kaijiangosaurus, Kelmayisaurus and Cruxicheiros. Finally, while the completeness of their taxon review is commendable, their analysis of tooth-based taxa is weak, since most are just bypassed as Theropoda indet. instead of examining dental features that vary between theropod clades.
These caveats aside, Carrano et al. (2012) is good enough that everyone should cut out the Basal Tetanurae chapter in their copy of The Dinosauria 2nd Ed. and put this there instead.
References- Muenster, 1836. Ueber die Korallenkalk das Linder Berges bei Hanover vorkommenden Ueberreste von Fischen, mit Beschreibung und Abbildung einiger neuen Arten. in Muenster and Wissman (eds.). Beitrage zur Petrefacten-Kunde. 7, 36-50.
Windolf, 1997. Theropoden-Zahne aus dem Oberen Jura Niedersachsens. in Sachs, Rauhut and Weigert (eds.). 1. Treffen der deutschsprachigen Palaoherpetologen, Terra Nostra, Extended Abstracts. Dusseldorf, Germany. 33-34.
Carrano, Benson and Sampson, 2012. The phylogeny of Tetanurae (Dinosauria: Theropoda). Journal of Systematic Palaeontology. 10(2), 211-300.