Wednesday, December 3, 2014

Norman's nomenclature's notoriously negative

Norman (2014) just published a revision of the iguanodont Hypselospinus, and while I have no opinion on the issues of synonymization, his treatment of nomenclature is deeply flawed.

First, Norman uses the term "nomen dubium in two different ways.  The correct way is illustrated by his discussion of Delapparentia, where he declares it a nomen dubium because it can't be distinguished from other iguanodonts.  This is also how he uses it for Siamodon and Penelopognathus.  However, he uses it incorrectly when he claims junior synonyms are nomina dubia.  A nomen dubium cannot be distinguished from two or more taxa, so that we can't know what it was.  But if you can't distinguish it from only one other taxon, then we know what it was and it's a junior synonym instead.  So contra Norman, he views Huxleysaurus, Darwinosaurus, Dollodon, Proplanicoxa, Sellacoxa, Mantellodon, Vectisaurus, Sphenospondylus and Kukufeldia as junior synonyms, NOT nomina dubia.  It's particularly difficult to figure out what Norman's concept of 'nomen dubium' is since there are cases where the same species is given different status depending on which genus it's paired with.  Iguanodon seeleyi and Wadhurstia fittoni are listed only as synonyms, yet Dollodon seeleyi and Huxleysaurus fittoni are listed as synonyms AND nomina dubia?!  What?  That makes no sense.

Secondly, Norman doesn't seem to 'get' phylogenetic nomenclature.  He brings up numerous ornithopod examples of how "topological change can occur in trees resulting from different systematic analyses", which is true.  He finds Tenontosaurus closer to Hypsilophodon than to Iguanodon, whereas most other modern analyses find the opposite, etc..  But then he says "Topological change will generate nomenclatural inconsistency that compromises the technique of clade anchoring. Consistency (a universal aspiration amongst taxonomists) underpins the advocacy of phylogenetic definitions but can only be assured if (and when) phylogenetic trees maintain consistent relational topologies."  This doesn't compromise the technique, it's the point of the technique!  If we all have a concept of Iguanodontia as (Iguanodon < Hypsilophodon) and Hypsilophodontidae as (Hypsilophodon < Iguanodon)*, then we can just say Norman recovered Tenontosaurus as a hypsilophodontid while Butler found it in Iguanodontia and know what that means.  The clades aren't supposed to have stable membership, they're supposed to have stable positional relationships to certain taxa.  So when Norman states Iguanodontia as defined above "is misleading in the sense that it clusters OTUs as anatomically dissimilar (dentally, cranially, and postcranially) as Tenontosaurus (and other hypsilophodontians, in the usage employed here - Figs 50, 52) with Dryosaurus and Camptosaurus that have definitively Iguanodon-like teeth", he's missing the point.  It doesn't matter how dissimilar the taxa in a clade are, the fact they form that clade with Iguanodon as opposed to Hypsilophodon is all that matters.

* All definitions I use here are simplified in not listing the type species for each genus for clarity's sake.

Norman's (2014) nomenclature on left vs. traditional/suggested nomenclature on right.  Modified after Norman (2014).

Finally we have Norman's taxa and definitions themselves.

Clypeodonta is defined as "Hypsilophodon foxii, Edmontosaurus regalis, their most recent common ancestor, and all of its descendants."  So the only non-clypeodont ornithopods are orodromines, Asian thescelosaurines and sometimes American thescelosaurines (in Boyd-based and Scheetz-based analyses but not Butler-based ones).

Norman defines Hypsilophodontia as "Hypsilophodon foxii, Tenontosaurus tilletti, their most recent common ancestor, and all of its descendants", but this is a senior synonym of Clypeodonta in most topologies since Tenontosaurus is an iguanodont in them.  It's also problematic because Hypsilophodontidae is already defined as (Hypsilophodon < Parasaurolophus) thanks to Sereno (1998).  Hypsilophodontidae thus covers the same taxa as Hypsilophodontia in Norman's cladogram, and in the absence of any proposed non-hypsilophodontian hypsilophodontids it seems unnecessary.  Note also that this puts an -ia clade inside of an -idae clade, which is unintuitive thanks to our Linnaean training.  Ironically, Cooper (1985) first proposed Hypsilophodontia as a clade equivalent to Clypeodonta.

Iguanodontia is defined as "Edmontosaurus regalis and all taxa more closely related to E. regalis than to the taxa subtended to the clade (Hypsilophodontia) that includes Hypsilophodon foxii and T. tilletti."  This is a horrible definition.  An obvious issue is that Edmontosaurus is used instead of Iguanodon, which violates Article 11.7 of the Phylocode.  The more pressing issue is that the definition is self-destructive if the consensus is correct in placing Tenontosaurus closer to Edmontosaurus than to Hypsilophodon.  If that consensus is correct, Edmontosaurus is itself subtended by the clade that includes Hypsilophodon and Tenontosaurus, so it refers to impossible taxa that are more closely related to themselves than to themselves.  Funny that after railing against definitions that only work in certain phylogenies, Norman creates a definition for a commonly used name that only works in his own heterodox phylogeny.

Ankylopollexia is redefined as "Edmontosaurus regalis and all taxa more closely related to E. regalis than to Dryosaurus altus" instead of the original (Camptosaurus + Parasaurolophus).  Why bother to redefine this when both definitions cover the same known taxa?  Also how about using a taxon with a actual pollex in the definition?  It's as if someone used Harpymimus as the internal specifier for Arctometatarsalia.

Styracosterna is redefined as "Batyrosaurus rozhdestvenskyi, E. regalis, their common ancestor, and all of its descendants" instead of the original (Parasaurolophus < Camptosaurus).  Again, both definitions cover the same taxa in Norman's cladogram, so there was no reason to make a new one.  Norman says application of the old definition "is compromised by
the increased complexity of camptosaur-grade (Camptosauridae sensu Sereno) iguanodont interrelationships (McDonald, 2011: fig. 1), as exemplified by the positions of Uteodon and Cumnoria."  But what's the problem if Uteodon and Cumnoria are styracosternans?  That's an interesting fact, and it's not like Sereno (1986) mentioned either genus when naming Styracosterna.  More importantly, why choose Batyrosaurus as your internal specifier?!  That was named 26 years after Styracosterna and found to be a hadrosauroid close to Probactrosaurus and Eolambia in its original description.  If that's right, Norman's Styracosterna would be a much less inclusive clade than historically recognized.  So again Norman creates a definition that only works in his heterodox phylogeny.  I don't know which topology is right, though Norman's does create a huge ghost lineage for the Santonian Batryosaurus if it's that basal.

Norman unofficially calls a clade of taxa 'iguanodontoids', but this clade already has a valid and defined name- Iguanodontidae.  The latter family was defined as (Iguanodon < Parasaurolophus) by Sereno (1998), so if Norman were effectively using the power of phylogenetic nomenclature, he could just say he found Proa, Jinzhousaurus, Bolong, Barilium and Mantellisaurus to be iguanodontids and we'd know what he means.

Hadrosauriformes is redefined as "Altirhinus kurzanovi, E. regalis, their common ancestor, and all of its descendants", instead of the original (Iguanodon + Parasaurolophus).   This changes the clade to be less inclusive for no reason, excluding iguanodontids.  Note both definitions violate Article 11.7, which requires Hadrosaurus to be used as an internal specifier.  Like the styracosternan situation, Norman's redefinition uses an internal specifier which was not even named when the clade was named (Altirhinus named in 1998 vs. Hadrosauriformes in 1997). 

Hadrosauromorpha is a new name for (Edmontosaurus < Probactrosaurus), which should again use Hadrosaurus as per Article 11.7.  Besides that, I think this taxon is a good one subsecting a long stem and using a classic genus.

Finally, Euhadrosauria is defined as "Parasaurolophus, Saurolophus, Edmontosaurus, their most common ancestor, and all of its descendants."  It hasn't been defined before to my (limited) knowledge, and this does correspond to the clade in Weishampel et al.'s (1993) phylogeny where they named it.  The problem is that in Prieto-Marquez-based phylogenies (which are 98% of all recent hadrosaur phylogenies), Hadrosaurus itself falls just outside this clade.  A Euhadrosauria excluding Hadrosaurus is counterintuitive to say the least.  Weishampel et al. created Euhadrosauria as "the traditional grouping of hadrosaurines and lambeosaurines", as opposed to the more basal Telmatosaurus.  Since Telmatosaurus is still considered more basal than Hadrosaurus, the phylogenetic definition (Hadrosaurus + Lambeosaurus) seems obvious.  This way it will always include the same taxa whether Hadrosaurus is in its traditional position by Kritosaurus or just basal to the main hadrosaur split.

It seems pretty obvious the ideal definitions using eponymous taxa and original concepts are-
Clypeodonta- (Hypsilophodon + Hadrosaurus)
Hypsilophodontidae- (Hypsilophodon < Hadrosaurus)
Iguanodontia- (Iguanodon < Hypsilophodon)
Dryomorpha- (Dryosaurus + Hadrosaurus)
Dryosauridae- (Dryosaurus < Hadrosaurus)
Ankylopollexia- (Camptosaurus + Hadrosaurus)
Camptosauridae- (Camptosaurus < Hadrosaurus)
Styracosterna- (Hadrosaurus < Camptosaurus)
Hadrosauriformes- (Iguanodon + Hadrosaurus)
Iguanodontidae- (Iguanodon < Hadrosaurus)
Hadrosauroidea- (Hadrosaurus < Iguanodon)
Hadrosauromorpha- (Hadrosaurus < Probactrosaurus)
Euhadrosauria- (Hadrosaurus + Lambeosaurus)

I don't know why no one's suggested those yet.  All eponymous, match original definitions except for replacing genera with standardized ones, plenty of node-stem triplets...  Phylogenetic nomenclature isn't rocket science.

Reference-  Norman, 2014. On the history, osteology, and systematic position of the Wealden (Hastings group) dinosaur Hypselospinus fittoni (Iguanodontia: Styracosterna). Zoological Journal of the Linnean Society. DOI: 10.1111/zoj.12193


  1. Thank you Mickey! I'd like to add that the node ={Hypsilophodon + Iguanodon} was widely called Euornithopoda by most authors already.
    Additionally, I'm not sure why Norman would choose to upend a perfectly stable clade definition by using Batyrosaurus as an internal specifier. It's not based on much at all (skeletal restoration:, and like you said, Styracosterna already has a perfectly stable definition.
    Lastly, I wish Sereno and Norman would stop using Parasaurolophus and/or Edmontosaurus as internal specifiers for clades that are modifications of the names Iguanodon and Hadrosaurus. I mean seriously....

  2. I thought Euornithopoda had that definition, but then checked Sereno (1986) and saw he created it nonsensically as a clade within Ornithopoda (excluding heterodontosaurids). When Sereno (1997, 1998) fixed this, he changed it to a stem-based clade (Parasaurolophus < Heterodontosaurus) which becomes a junior synonym of Ornithopoda itself in recent phylogenies where heterodontosaurids are outside Cerapoda. So I could understand Norman wanting to avoid this messy history and creating a new name for (Hypsilophodon + hadrosaurid), since besides Cooper's unintuitive Hypsilophodontia, none had been proposed before AFAIK.

    As for Batyrosaurus, at least it has a hatchet-shaped sternal plate. ;)

    Sereno's rationale for using more complete and deeply nested taxa as internal specifiers instead of eponymous taxa misses the very obvious objection that if somehow e.g. Hadrosaurus is found to fall outside a Hadrosauriformes defined as (Iguanodon + Parasaurolophus), then we couldn't call that taxon Hadrosauriformes anyway.

  3. Sorry about being off the topic, but do you know about the details of this tyrannosaur specimen?

  4. Hi Pete,

    The one reason that Sereno and Norman declined to use Hadrosaurus as an internal specifier is because the holotype of Hadrosaurus largely lacks a skull, with the exception of maxillary fragments and teeth, which offered little phylogenetic signal about whether Hadrosaurus is closely related to uncrested duckbills.