Saturday, March 26, 2016

Peters' theropod craziness

So David Peters, whose analyses I have thoroughly trashed on this blog before, has branched out into theropods.  In my prior foray into his basal dinosaurian phylogeny, I was surprised by how poorly coded (~30% miscoded) taxa were for his tiny and badly formed analysis.  Now that he's trying to interpret MY turf, it's just sad.  Or laughable.  Or laughably sad, take your pick.  Even in the best of circumstances, taking 228 characters designed for amniote phylogeny and coding theropods for them is likely to give you a terrible result.  But add Peters' DGS to the mix, where he takes photos and discovers missing bones in a slab or traces his own outlines, and you have hilarious fiction.  Peters might know pterosaurs, but when he tries tracing theropods, it's obvious he does NOT know theropods.  He mixes bone outlines, identifies random sediment texture or feathers as bones, and based on his results just has no idea how theropods look.  I commented on this on his Archaeornithura skull post last week (see last figure below), but Peters deleted the comment and changed his reconstruction to be wrong in different ways (new ways include that the pubis is the ischium, the ischium has a huge obturator process, the manus has a complete digit, as does the fifth pedal digit(!)).

Reconstructions of Sinosauropteryx prima by me (top), and Peters (bottom), scaled to same femoral length.
As an example, here's Sinosauropteryx as reconstructed by him vs. by me.  Now, Peters is no doubt the more experienced artist, and my reconstruction isn't perfect.  I made it back in 2005, and if I were to redraw it today, I'd give the tibia and fibula less generic shapes (e.g. larger cnemial crest, flatter distal tibia, expanded proximal fibula), orient the humerus to not be in posterior view, and get rid of the post-obturator notch in the ischium, which was later found to be erroneous.  But compare that to Peters' monstrosity.  The squamosal is just a triangular plate, lacking any posterior process so that the quadrate head is impossibly just floating freely posterior to it.  The quadrate has a narrow orbital/pterygoid process unlike anything but derived birds.  There are twenty-five presacral vertebrae unlike (?)all non-maniraptoran avepods, the caudals vary randomly in length by large degrees unlike any theropod, and the chevrons are all short as if it were a derived paravian.  In the pectoral girdle, there's some huge crescent where the sternum would be, but Sinosauropteryx lacks an ossified sternum.  The ilium is just wacky- tiny postacetabular process and concave dorsal margin.  The ischium is a thick blob with nary an obturator process.  Perhaps most sadly, Peters can't even draw the tibia as longer than the femur, which is (?)universal in small theropods, stated even in the crappy original Chinese description and is character 195 in Peters' analysis.

With his reconstructions being more or less fictional animals, it's no surprise the cladograms based on them will be equally unrealistic.  And lo, they are!  For fun, I give his cladogram from March 14th and applied phylogenetic nomenclature to see how it fares.  I listed the assumed position of taxa not yet included, as they are specifiers of various clades.  I also list where family-level taxa would have to be renamed due to ICZN rules.

Theropoda
|--Tawa
`--+--+--Staurikosaurus
   |  `--+--Segisaurus
   |     `--+--Guaibasaurus
   |        `--+--Marasuchus
   |           `--Procompsognathus
   `--Avepoda
      |--Coelophysoidea
      |  |--Dracoraptor
      |  `--+--Coelophysidae
      |     |  Coelophysis
      |     `--Dilophosauridae
      |        Dilophosaurus
      `--Neotheropoda (= Orionides, Avetheropoda, Neotetanurae)
         |--Ceratosauria (= Spinosauroidea, Carnosauria, Deinocheiridae)
         |  Megalosauria
         |  |--Spinosauridae (= Proceratosauridae)
         |  |  |--Proceratosaurus
         |  |  `--+--Deinocheirus
         |  |     `--+--Xiongguanlong
         |  |        `--+--Sinocalliopteryx
         |  |           `--+--Suchomimus
         |  |              `--Spinosaurus
         |  `--+--+--Dilong
         |     |  `--Guanlong
         |     `--Allosauroidea (= Neoceratosauria) (would need to be renamed Megalosauroidea)
         |        |--Abelisauroidea (= Metriacanthosauridae)
         |        |  |--Yutyrannus
         |        |  `--+--Sinraptor (assuming Metriacanthosaurus)
         |        |     `--Abelisauridae
         |        |        Majungasaurus (assuming Abelisaurus)
         |        `--Ceratosauridae (would need to be renamed Megalosauridae)
         |           Allosauria
         |           |--Allosauridae
         |           |  Allosaurus
         |           `--Megalosauridae (= Megalosauroidea, Carcharodontosauridae)
         |              |--+--Sinosaurus
         |              |  `--Monolophosaurus
         |              `--+--Acrocanthosaurus (assuming Carcharodontosaurus)
         |                 `--+--Eustreptospondylus (assuming Megalosaurus)
         |                    `--Ceratosaurus
         `--Coelurosauria (= Tetanurae)
            Maniraptoriformes (= Aviremigia, Tyrannoraptora, Bullatosauria)
            |--Compsognathidae (= Tyrannosauroidea, Arctometatarsalia, Microraptoria)
            |  |--Ornithomimosauria
            |  |  |--Compsognathus
            |  |  `--Struthiomimus (assuming Ornithomimus)
            |  `--+--+--Ornitholestes
            |     |  `--+--Microraptor
            |     |     `--Sinornithosaurus
            |     `--+--Fukuivenator
            |        `--+--Tianyuraptor
            |           `--+--Zhenyuanlong
            |              `--+--Alioramini
            |                 |  Alioramus
            |                 `--Tyrannosauridae
            |                    |--Gorgosaurus
            |                    `--Tyrannosauridae
            `--Maniraptora
               Pennaraptora (= Chuniaoae)
               |--Caenagnathiformes (= Oviraptoriformes)
               |  |--Therizinosauria
               |  |  |--Falcarius
               |  |  `--+--Jianchangosaurus (assuming Therizinosaurus)
               |  |     `--Rahonavis
               |  `--Oviraptorosauria
               |     |--+--Juravenator
               |     |  `--Sinosauropteryx
               |     `--+--Limusaurus
               |        `--Khaan (assuming Caenagnathus and Oviraptor)
               `--Paraves
                  |--Coeluridae
                  |  |--Tanycolagreus (assuming Coelurus)
                  |  `--Eotyrannus
                  `--Metornithes (= Eumaniraptora)
                     |--Dromaeosauridae (= Alvarezsauroidea, Alvarezsauria, Deinonychosauria)
                     |  |--+--Haplocheirus
                     |  |  `--Shuvuuia (assuming Alvarezsaurus)
                     |  `--+--Velociraptor (assuming Dromaeosaurus and Deinonychus)
                     |     `--Balaur
                     `--Avialae (= Averaptora)
                        |--Sinornithoides (assuming Troodon)
                        `--+--Jinfengopteryginae
                           |  Jinfengopteryx
                           `--+--Unenlagiinae
                              |  |--Anchiornis
                              |  `--+--Aurornis
                              |     `--Buitreraptor (assuming Unenlagia)
                              `--+--Sinovenator
                                 `--+--+--Eosinopteryx
                                    |  `--Xiaotingia
                                    `--+--Thermopolis "Archaeopteryx"
                                       `--Ornithothoraces (= Ornithes, Ornithopectae)
                                          |--Archaeopterygidae (= Saurornithes, Enantiornithes)
                                          |  |--Archaeornis
                                          |  `--+--+--Archaeopteryx
                                          |     |  `--Yixian embryo IVPP V14238
                                          |     `--+--Protopteryx
                                          |        `--+--Cathayornis (assuming Iberomesornis and Enantiornis)
                                          |           `--+--Pengornis
                                          |              `--Sulcavis
                                          `--Ornithurae
                                             |--Omnivoropterygiformes
                                             |  |--"Archaeopteryx" bavarica
                                             |  `--+--Mei
                                             |     `--Scansoriopterygidae
                                             |        |--Scansoriopteryx
                                             |        `--+--Yi
                                             |           `--+--Epidexipteryx
                                             |              `--Omnivoropteryx
                                             `--+--Jeholornis
                                                `--+--Jurapteryx
                                                   `--Pygostylia
                                                      |--Confuciusornithiformes
                                                      |  |--Wellnhoferia
                                                      |  `--Confuciusornis
                                                      `--Avebrevicauda
                                                         |--Chiappeavis
                                                         `--+--Sapeornis
                                                            `--+--Archaeornithura
                                                               `--+--Ichthyornis
                                                                  `--Aves
                                                                     |--Struthio
                                                                     `--Gallus (assuming Passer)


That's pretty funny.  Nomenclature fails in Allosauroidea and Compsognathidae, due to the weird topologies there where taxa not seen as closely related enough to need their neighbors as external specifiers suffer.  Ditto for dromaeosaurids vs. alvarezsaurs.  Also for archaeopterygids vs. enantiornithines, which no BAD analysis has ever recovered. 

As for the topology, there's that sister clade to avepods, whose oldest family name would be Procompsognathidae, though none of the members have phylogenetic definitions attached to them.  The division in neotheropods is sort of like Rauhut (2003), who had a huge Carnosauria.  The most basal megalosaurians are all actually basal tyrannosauroids.  Ornithomimids plus tyrannosaurids is a classic clade from Huene to Holtz, and Compsognathus being there reminds me of Olshevsky (1995) having it as a tyrannosaur.  Funny how Peters recovered Enigmosauria, though no doubt for completely different reasons than the dinosaur community, given the inclusion of compsognathids, Limusaurus and Rahonavis.  Troodontids and unenlagiines being avialans has been more popular recently (e.g. Agnolin and Novas, 2013), and the Archaeopteryx plus Enantiornithes pairing hearkens back to the BANDits' Sauriurae.  Similarly, Confuciusornis being closer to Aves than enants reminds me of Kurochkin's (2006) ideas.  And those are the only parallels I can make between Peters' non-consensus phylogeny and science.  My head's full of non-traditional phylogenetic proposals, but Deinocheirus as a spinosaur, or microraptorans as basal tyrannosaurs while Velociraptor is by alvarezsaurs?  That's just nuts.

So David, first I'd say you should start using the names Archaeornis, Jurapteryx and Wellnhoferia if you find these taxa away from the Archaeopteryx holotype.  But also, I think it would be amusing if you added the following- Megalosaurus, Hexing, Nothronychus/Erlikosaurus, Caudipteryx, Achillobator, Dromaeosaurus and Patagopteryx.


Peters' reconstruction of Archaeornithura meemannae from 3-26-16, preserved in case it's deleted like the first was (though Peters to his partial credit does say "Updated March 16, 2016 with new images. The beak, if present, is ephemeral, questionable. Only two scores changed.").

References- Olshevsky, 1995. The origin and evolution of the tyrannosaurids. Kyoryugaku Saizensen. 9, 92-119 (part 1); 10, 75-99 (part 2).
 

Rauhut, 2003. The interrelationships and evolution of basal theropod dinosaurs. Special Papers in Palaeontology. 69, 1-213.  

Kurochkin, 2006. Parallel evolution of theropod dinosaurs and birds. Entomological Review. 86(suppl. 1), S45-S58.

Agnolin and Novas, 2013. Avian ancestors: A review of the phylogenetic relationships of the theropods Unenlagiidae, Microraptoria, Anchiornis and Scansoriopterygidae. Springer. 96 pp.

8 comments:

  1. Replies
    1. Because I find it entertaining to analyze phylogenetic hypotheses, plus more anti-Peters media might help dissuade people from trusting him.

      Delete
  2. I don't know if I should laugh or cringe at that AWFUL cladogram. Seriousy, it's just...bleh.

    ReplyDelete
  3. I think your attempt at applying phylogenetic nomenclature to Peters' cladogram is doomed to failure, as it is based on the assumption that Peters' phylogeny is in any way going to follow a traditional topology. I mean, Rahonavis is related to basal therizinosaurs and Deinocheirus is a spinosaurid so I have no expectation that, for example, Therizinosaurus, will end up somewhere predictable.

    ReplyDelete
    Replies
    1. Ah, but the goal of phylogenetic nomenclature is that clade names will tell us where taxa fall in a phylogeny, and hypothetically _should_ be applicable in any realistic topology. That's why Sereno definitions are terrible, because they only function in his topology. So by utterly failing in several areas, it shows just how far from reality Peters is.

      Delete
  4. I think the Spinosauridae is probably the best part. Not just Deinocherius, but tyrannosauroids and oviraptorosaurs!

    ReplyDelete
  5. As I discussed in my blog, Peters' topologies are statistically mid-way between pure randomly-built trees and parsimony. They are evidently wrong and evidently biased (someway).

    ReplyDelete
  6. The lack of chevrons is really something!

    taking 228 characters designed for amniote phylogeny and coding theropods for them is likely to give you a terrible result

    Case in point: I once put myself (plus some human anatomy from Wikipedia) into a matrix of close to 300 characters designed for tetrapod phylogeny. One or two other synapsids were in the matrix. Instead, I came out next to the albanerpetid lissamphibians. I'm too far out for that character sample to handle. :-)

    ReplyDelete