Wednesday, June 27, 2018

Etrigansauria, the unnecessary demon

For this week's post, we have a new paper on ceratosaurs by Delcourt (2018).  The paper discusses anatomy, behavior and such (and provides some good photos of Limusaurus), but here I only want to deal with the phylogenetic taxonomy proposed.  Delcourt uses the topology of Analyses 1 and 2 of Wang et al. (2016) from the Limusaurus ontogeny paper, which seems to have been well done from what I've seen.  Tons of taxa, tons of characters, and some sage commentary on OTUs and anatomy.  It's a shame the full tree is never presented in the paper, and I hope that analysis or a derivative is used for a stand alone phylogeny paper in the future.  Wang et al. interestingly find noasaurids to group with Elaphrosaurus, Limusaurus, Spinostropheus and Deltadromeus in a clade outside ceratosaurids plus abelisaurids, which is also what the Lori analysis found weirdly enough.

Ceratosaur portion of Analysis 1 and 2 of Wang et al. (2016).  Numbers are GC jacknife supports from Analysis A (above branches) and B (below branches).  After Wang et al., 2016.

So I have no issue with the phylogeny, but Delcourt's proposed phylogenetic taxonomy is... bad.  Let's start with his new taxon- Etrigansauria, named after DC Comics character Etrigan who is a demon bound to the human Jason Blood.  Now I love the DC Animated Universe as much as the next person, but what's the phylogenetic definition of Etrigansauria? "The most inclusive clade containing Carnotaurus sastrei and Ceratosaurus nasicornis but not Noasaurus leali."  But wait a second, we already have a Carnotaurus plus Ceratosaurus node-based clade- Neoceratosauria from Novas 1991, most recently refined by Hendrickx et al. (2015) as "The least inclusive clade containing Ceratosaurus nasicornis and Carnotaurus sastrei."  So if we have a standard topology with abelisauroid noasaurids, Etrigansauria self destructs, and if we have a Wang et al. style topology then Etrigansauria is just a junior synonym of Neoceratosauria.  [Edit: I noticed when commenting on the journal's website that the definitions are slightly different in that Etrigansauria is stem-based, so that there could be non-neoceratosaurian etrigansaurs under a Wang et al. style topology.  But the only such taxon in Wang et al.'s trees that were used is the skull-less Berberosaurus, which falls out other places in their other analyses- elaphrosaurid, stem-ceratosaur, etc..  Indeed, I doubt Delcourt had this distinction in mind because his figure 1 actually places Etrigansauria at the neoceratosaur node, not the stem containing Berberosaurus as his definition would have it.]  The word "neoceratosaur" only appears once in Delcourt's paper, as a brief mention- "In some analyses, Berberosaurus is considered as a basal ceratosaurian, a neoceratosaurian, ...".  This seems weird, and I urge everyone not to forget Novas's decades of work on these animals, and use Neoceratosauria instead of Etrigansauria published twenty-seven years later.

I don't know why everyone has such a hard time with Phylocode Article 11.7, which reads in part "when a clade name is converted from a preexisting typified name or is a new or converted name derived from the stem of a typified name, the definition of the clade name must use the type species of that preexisting typified name or of the genus name from which it is derived (or the type specimen of that species) as an internal specifier."  Let's keep that rule in mind as we look at the rest of the definitions...

Delcourt's other big blunder is with Abelisauroidea versus Ceratosauroidea.  Delcourt uses Wilson et al.'s (2003) definition of Abelisauroidea which is (Carnotaurus sastrei + Noasaurus leali).  That's a bad definition since it doesn't include Abelisaurus as a specifier, but we'll ignore that for the moment.  He also uses Wilson et al.'s definition for Abelisauridae- (Carnotaurus sastrei < - Noasaurus leali).  Which is also bad in not including Abelisaurus, but whatever.  Using these bad definitions in Wang et al.'s phylogeny gets us the bad result of making Ceratosaurus an abelisaurid abelisauroid.  Which Delcourt correctly notes it can't be, because Ceratosauridae/oidea has priority over Abelisauridae/oidea.  Delcourt's weird solution is to instead use Wilson et al.'s Abelisauroidea definition for Ceratosauroidea, so that Ceratosauroidea is now Carnotaurus plus Noasaurus. Ack!

Okay, first of all, that doesn't work because Ceratosaurus nasicornis needs to be an internal specifier of Ceratosauroidea (Phylocode Article 11.7!).  Second, we have much earlier and better definitions to use than Wilson et al.'s.  Holtz (1994) defined Abelisauroidea as taxa closer to abelisaurids than Ceratosaurus, which can be easily modified to (Abelisaurus comahuensis < - Ceratosaurus nasicornis).  See, that's a good definition that follows Article 11.7, maps correctly on to Novas' (1991) topology when he created the taxon, and never includes Ceratosaurus to avoid that whole kerfuffle.  In Wang et al.'s topology, noasaurids aren't abelisauroids, simple as that.  What is Ceratosauroidea?  It's only ever been defined before this as an alternative to Neoceratosauria when that clade was thought to be sister to coelophysoids, so those definitions make it a junior synonym of Neotheropoda sensu Bakker.  Given current taxonomy, it doesn't seem like a useful clade to redefine unless what we call Ceratosauridae now expands a LOT.

Ceratosauria in a version of the Lori analysis.  I'd ignore Kayentavenator and bissektensis as flukes based on their very fragmentary holotypes.  Not bad for analyzing ceratosaurs using maniraptoromorph characters...

The rest of Delcourt's definitions aren't any better.

Ceratosauria: Most inclusive clade containing Ceratosaurus but not Neornithes

That's the standard ever since Rowe (1989) so is fine.

Noasauridae: Most inclusive clade containing Noasaurus but not Carnotaurus

That's the original definition from Wilson et al. (2003), and is fine.

Elaphrosaurinae: Most inclusive clade containing Elaphrosaurus but not Noasaurus

This is attributed to Rauhut and Carrano (2016), but that's wrong.  Those authors used the better definition "all noasaurids that are more closely related to Elaphrosaurus than to Noasaurus, Abelisaurus, Ceratosaurus, or Allosaurus" to cover alternative topologies so that we don't get stupid results like abelisaurid Ceratosaurus.

Noasaurinae: Most inclusive clade containing Noasaurus but not Elaphrosaurus

Ditto here. Delcourt falsely attributes that definition to Rauhut and Carrano, but they actually used the much better "all noasaurids that are more closely related to Noasaurus than to Elaphrosaurus, Abelisaurus, Ceratosaurus, or Allosaurus."  As an obvious illustration of why Delcourt's definition is bad, in a standard topology where Noasaurus is an abelisauroid but Elaphrosaurus is not (e.g. Carrano and Sampson, 2008), Ceratosaurus and abelisaurids are all noasaurines.

Ceratosauridae: (new definition) the most inclusive clade containing Ceratosaurus but not Carnotaurus.

We actually already have an equivalent definition for this family- Rauhut (2004) wrote "the name is used here for a clade containing all ceratosaurs that are more closely related to Ceratosaurus than to abelisaurids." But even if one disputed that because Rauhut stated "it is premature to give such a formal definition at present", Hendrickx et al. (2015) defined it as "The most inclusive clade containing Ceratosaurus nasicornis but not Carnotaurus sastrei and Noasaurus leali." 

Abelisauridae: (new definition) the most inclusive clade containing Carnotaurus but not Ceratosaurus.

Article 11.7!  Abelisaurus comahuensis needs to be an internal specifier for Abelisauridae, again why is this so hard?  No argument for using Carnotaurus makes any sense.  Sure it's more complete, and may be more deeply nested, but if Abelisaurus somehow ends up not closer to Carnotaurus than Ceratosaurus, you're not going to be calling the Carnotaurus group Abelisauridae anyway.  The sad part is that we actually do need a new good definition for Abelisauridae.  Novas' (1997) definition is a node including the fragmentary Xenotarsosaurus which has an unstable position in recent analyses.  Rowe et al.'s (1997), Wilson et al.'s (2003) and Sereno's (perpetually in press) definitions all use Carnotaurus.  Sereno's (1998) is a node using Abelisaurus and Carnotaurus that would work (except for Rugops) in Wang et al.'s topology, but exclude all taxa except the two specifiers and Aucasaurus in Filippi et al.'s (2016) topology, for instance.  Plus none of the stem-based definitions exclude Ceratosaurus.  Delcourt's definition is bad because if noasaurids are abelisauroids as in most topologies, noasaurids are abelisaurids.  Here's what a good definition of Abelisauridae looks like- All taxa more closely related to Abelisaurus comahuensis than to Ceratosaurus nasicornis, Noasaurus leali, Elaphrosaurus bambergi or Allosaurus fragilis.  Someone publish that.

Ceratosauria phylogeny from Filippi et al. (2016), after Filippi et al. (2016).

Carnotaurinae: Most inclusive clade containing Carnotaurus but not Abelisaurus

That's fine and classic, taken from Sereno (1998) who named the clade.  Honestly, I think this and Abelisaurinae have proven to be pretty useless due to the varying position of Abelisaurus and should probably be ignored by future authors.  In Filippi et al.'s trees, Carnotaurus is the only carnotaurine, but in Wang et al.'s trees all abelisaurids except Abelisaurus and Rugops are.  Instead of abelisaurines vs. carnotaurines, the more useful split seems to be between majungasaurs and brachyrostrans (in the Lori trees too, incidentally).  It would be great to have a name for the Majungasaurus plus Carnotaurus node too, so we could easily refer to "basal" abelisaurids like Rugops, Abelisaurus and Ilokelesia in Wang et al.'s trees, Kryptops and Rugops in Filippi et al.'s trees, or Rugops, Genusaurus and Eoabelisaurus in the Lori tree.

Majungasaurini: Most inclusive clade containing Majungasaurus but not Carnotaurus

Here, Delcourt correctly notes that in Wang et al.'s topology, Majungasaurinae as originally defined falls out inside Carnotaurinae.  So it's yet another case where the Phylocode clashes with the ICZN.  Which is a fair observation, but I think the better solution would be to propose Majungasauria for that clade, which could go in either position and work fine. 

Brachyrostra: Most inclusive clade containing Carnotaurus but not Majungasaurus

The original definition, so that's fine.

Furileusauria: Most inclusive clade containing Carnotaurus but not Skorpiovenator.

This is credited to Filippi et al. (2016), but those authors actually had Ilokelesia and Majungasaurus as external specifiers too.  Since those two are also outside Furileusauria in Wang et al.'s trees, I don't see why the change was made.  I'm also not sure how useful the clade Furileusauria is yet.  It obviously works in Filippi et al.'s phylogeny, but has uncertain content in Wang et al.'s since Carnotaurus and Skorpiovenator are part of a polytomy.  I should note here that it's not at all certain whether Wang et al.'s or Filippi et al.'s trees are better supported.  Filippi et al. includes 6-8 more ceratosaur taxa and 416 characters with just a few outgroups, while Wang et al. have 744 characters but also densely sample and test coelophysoids and have quite a lot of tetanurines, so many of those characters are probably not parsimony-informative for ceratosaurs.

So that's Delcourt's (2018) phylogenetic taxonomy.  I honestly don't see how Neoceratosauria was never brought up in peer review.  It's also ironic that Delcourt goes through hoops to try to adjust Wilson et al.'s bad definitions because they only work in Wilson et al.'s topology, only to propose new definitions that only work in Wang et al.'s topologies.  I don't know about the other portions of the paper, but the phylogenetic taxonomy section is an utter failure of peer review in my opinion.  If people start using Etrigansauria and this butchered Ceratosauroidea... *cringe*

Look, if you were tasked with making good definitions for Ceratosauria, it's easy-
Ceratosauria- (Ceratosaurus nasicornis < - Passer domesticus)
Noasauridae- (Noasaurus leali < - Abelisaurus comahuensis)
Elaphrosaurinae- (Elaphrosaurus bambergi < - Noasaurus leali, Abelisaurus comahuensis, Ceratosaurus nasicornis, Allosaurus fragilis)
Noasaurinae- (Noasaurus leali < - Elaphrosaurus bambergi, Abelisaurus comahuensis, Ceratosaurus nasicornis, Allosaurus fragilis)
Neoceratosauria- (Ceratosaurus nasicornis + Abelisaurus comahuensis)
Ceratosauridae- (Ceratosaurus nasicornis < - Abelisaurus comehuensis)
Abelisauroidea- (Abelisaurus comehuensis < - Ceratosaurus nasicornis)
Abelisauridae - (Abelisaurus comahuensis < - Ceratosaurus nasicornis, Noasaurus leali, Elaphrosaurus bambergi, Allosaurus fragilis)
NEW CLADE- (Majungasaurus crenatissimus + Carnotaurus sastrei)
Majungasauria- (Majungasaurus crenatissimus < - Carnotaurus sastrei)
Brachyrostra- (Carnotaurus sastrei < - Majungasaurus crenatissimus)

Follows Article 11.7, works in everyone's topologies, ta da. 

References- Rowe, 1989. A new species of the theropod dinosaur Syntarsus from the Early Jurassic Kayenta Formation of Arizona. Journal of Vertebrate Paleontology. 9(2), 125-136.

Novas, 1991. Phylogenetic relationships of ceratosaurian theropod dinosaurs. Ameghiniana. 28, 401.

Holtz, 1994. The phylogenetic position of the Tyrannosauridae: Implications for theropod systematics. Journal of Paleontology. 68(5), 1100-1117.

Novas, 1997. Abelisauridae. In Currie and Padian (eds.). Encyclopedia of Dinosaurs. Elsevier Inc. 1-2.

Rowe, Tykoski and Hutchinson, 1997. Ceratosauria. In Currie and Padian (eds.). Encyclopedia of Dinosaurs. Elsevier Inc. 106-110.

Sereno, 1998. A rationale for phylogenetic definitions, with application to the higher-level taxonomy of Dinosauria. Neues Jahrbuch für Geologie und Paläontologie Abhandlungen. 210(1), 41-83.

Wilson, Sereno, Srivastava, Bhatt, Khosla and Sahni, 2003. A new abelisaurid (Dinosauria, Theropoda) from the Lameta Formation (Cretaceous, Maastrichtian) of India. Contributions from the Museum of Paleontology. 31(1), 1-42.

Rauhut, 2004. Provenance and anatomy of Genyodectes serus, a large-toothed ceratosaur (Dinosauria: Theropoda) from Patagonia. Journal of Vertebrate Paleontology. 24(4), 894-902. 

Carrano and Sampson, 2008. The phylogeny of Ceratosauria (Dinosauria: Theropoda). Journal of Systematic Palaeontology. 6, 183-236. 

Hendrickx, Hartman and Mateus, 2015. An overview of non-avian theropod discoveries and classification. PalArch's Journal of Vertebrate Palaeontology. 12(1), 1-73.

Filippi, Mendez, Juarez Valieri and Garrido, 2016. A new brachyrostran with hypertrophied axial structures reveals an unexpected radiation of latest Cretaceous abelisaurids. Cretaceous Research. 61, 209-219. 

Rauhut and Carrano, 2016. The theropod dinosaur Elaphrosaurus bambergi Janensch, 1920, from the Late Jurassic of Tendaguru, Tanzania. Zoological Journal of the Linnean Society. 178(3), 546-610.

Wang, Stiegler, Amiot, Wang, Du, Clark and Xu, 2016. Extreme ontogenetic changes in a ceratosaurian theropod. Current Biology. 27(1), 144-148.

Delcourt, 2018. Ceratosaur palaeobiology: New insights on evolution and ecology of the southern rulers. Scientific Reports. 8:9730.


  1. Left a link to this post as an annotation on the paper too, nice work! :)

  2. I can't really speak to the taxonomic portion (I learned a lot reading this). The soft tissue stuff seems fine, following Hieronymus et al. (for centrosaurines) and Carr et al. (for Daspletosaurus). Not personally a big fan of the Oliveira illustrations; in particular, Pycnonemosaurus looks like a cartoon. I haven't gotten to the behavior section yet. I do like that Delcourt is actually discussing paleobiology and functional morphology, because it seems like publications regarding those basic tenants of paleontology are becoming increasingly rare.

    I dislike the clade name Etrigansauria for a different reason: I assume he chose this name because Etrigan has horns, but Ceratosaurus and Carnotaurus are the only horned etrigansaurians (unless I'm forgetting someone). There are more widespread features that distinguish the group, like gnarly-looking heads and bulldog faces. I suppose Lockjawsauria doesn't exactly roll off the tongue but it would be more fitting...

  3. I'd adjust the definition of Ceratosauria to include Allosaurus fragilis as an external specifier.