Saturday, August 10, 2019

Ornithomimosaurs in the Lori matrix

Next up is Ornithomimosauria.  The Theropod Database has been updated with new ornithomimosaur info too.  The published cladogram here is-

I included the Angeac taxon as an OTU, which had very few illustrated elements at the time of scoring- the cervical, tibia and distal metatarsus with a few other reported details like the toothless dentary.  It emerged as an ornithomimosaur but since then good views of the known material have been made public and it's pretty obviously a ceratosaur similar to Limusaurus or Elaphrosaurus.  This was always an easy possibility in the matrix, as it can move to Ceratosauria in the published matrix with only two steps.  After rescoring it for the more complete remains, the Angeac  taxon emerges in Ceratosauria sister to Deltadromeus.  Here's its new scorings

????????1? ?????????? ?????????? ?????????? ?????????? 0????????? ?????????? ???2??0??? ?????????- ????--???? ??0?0??10? ???{01}??0??0 000?1?1{01}0? ???????{01}01 ?01?02{01}0?{01} ?????????? ????????11 1000?0?110 1100??{12}?00 ?1?10?1000 1??0??001? ??0??000{01}0 ???001???? ???0?????? {01}?021??0?1 ?0?110???? ?000?0{012}0?? ??{01}0???1?? ?????0?002 ??1?0?0??0 ??100?0?0? ??????11{01}1 ?????????? 1??1??{23}??? ?????????? ????-22?00 1001??0??? ?????????? 0??0011??? ?2000????? ?0?????00? ?0?000??0? ????{12}???0? ????000??? ??0-?1???? 0?1?1{01}???? ??1-?????0 ?0?????0?? ??00010?0? ?00??????? 0???????00 ???011???? ??00?????? ??0???0??0 ??0??????? ?????????? ?????????? ????0??010 ?????????? ???000-?0? ?0?-01?00? ???1?0??{01}? ??0?1????? ?????????? ??000?-?1? 00???0?{01}?? ???000???- -??00-???? ??????0??0 0???????0?

When Angeac moves in the new matrix with added taxa, we get a little change where Harpymimus is more basal and the polytomy is resolved-

Btw, I published a list of phylogenetic definitions in the supplementary material, including the first to define Ornithomimosauria on the type species- (Ornithomimus velox < - Tyrannosaurus rex, Shuvuuia deserti, Therizinosaurus cheloniformis, Oviraptor philoceratops, Troodon formosus, Passer domesticus).

Ours is one of the few published studies using the new Deinocheirus skeletons, and unlike Lee et al. (2014) I recovered it basal to garudimimids and any other ornithomimosaurs except for Hexing.  It only takes 4 steps to move with other toothless ornithomimosaurs, so that's still quite possible.  What seems implausible is it being a garudimimid, as that takes 14 steps.  The supporting evidence for that in Lee et al.'s analysis was never great, and as shown in the Sciurumimus section the scoring in Choiniere's analysis is pretty bad.

This is the first time Hexing has been in an analysis besides the ornithomimosaur-only one in its original description, which did not include Deinocheirus.  It groups with Deinocheirus, which is funny because of the size difference, but there would have to be something to fill up that ghost lineage if Deinocheirus is so basal.

Shenzhousaurus and Harpymimus are next.  Kinnareemimus takes three steps to move to Ornithomimosauria in the old matrix, but only one step in the new matrix.  It emerges in a trichotomy with Shenzhousaurus and Harpymimus. An extra step moves it to Alvarezsauroidea, so that's still pretty uncertain.

Archaeornithomimus asiaticus paratype left femur in posterior view (AMNH 21800) (Courtesy AMNH).

While I didn't recover Deinocheirus by Garudimimus, I did get Beishanlong there as in Lee et al..  Also interesting is that Archaeornithomimus groups there.  That genus is known from a whole ton more specimens than suggested by Smith and Galton (1990), which I was able to examine at the AMNH.  There are whole boxes of tibiae, and metatarsals, and phalanges, that could really use a new osteology.  Where's Archaeornithomimus bissektensis you ask?  Its holotype femur went way too many places to include, lacking a unique combination of scores.  Sues and Averianov (2015) did assign a lot of Bissekty material to the taxon, but I didn't include any of the Bissekty isolated composite taxa in case they're chimaerical.  I still scored them though, so if anyone wants to experiment, the Bissekty Ornithomimosauria is-

???????0?? ?????????? ?????????? ?????????? ?????{01}0000 ?0???????? ?????????? ?????????? ?????????? ?????????? ??011010(01)0 ?011????0? 0?0??0???? ?????????? 10{01}0?2{01}00? ?01??0??{01}1 0??1?0???? ?0?0?0?00? ?????????? ?????01000 1100210000 1100000020 ?????????? ????0????? ???2?0?1?? 0???0????? ?00??0??0? ??00?????0 ???10??00? 1?1??????? ????0???0? ????????11 ?????????? ??010????? ??????0??? ?????2???? {01}???0?0?01 ??????10?? ???0?11??? ?210?????? 1??????00? 0????0?0?1 ????????0? ??1?000??? ?????0???? ??1??{01}001? ??1??00??0 10??????10 0(01)1000?0?0 ?????0???? 0???????0? ???1?????? ??0?0??0?? ??00?????? ?????????? ?????????? ?????????? ???00??010 ??0?-????? ??????-?00 ?00-010??1 -11100??00 ???????--? ???00?00?? 0?00?????? ?00?000100 0{01}0000???- -???{01}--??? ??-?0????? ?10?100??0

Archaeornithomimus? bissektensis holotype right femur (CCMGE 726/12457) (Courtesy of Averianov).

 The "Grusimimus" specimen GIN 960910KD ends up in Garudimimidae too, despite generally being seen as a potential juvenile Harpymimus.  Joining the two takes 4 steps, so is still quite possible.  I wonder if this really is an endemic Asian clade.  Interestingly, Arkansaurus (which we were able to score based on the new description) is a basal garudimimid here, but it can move either a node outside the garudimimid plus ornithomimid clade, or to Ornithomimidae or to Tyrannosauroidea (where it ends up similar to where Suskityrannus was placed by its authors) with one step.  The Lori matrix isn't that great with pedal characters yet.

We also provide a definition for Ornithomimidae using O. velox and not dependent on the controversial Deinocheirus- (Ornithomimus velox < - Garudimimus brevipes). At the base of Ornithomimidae, we have Nedcolbertia.  While an ornithomimosaurian Nedcolbertia's been recently proposed by Brownstein (2017) and Hunt and Quinn (2018), this is the first published analysis to find the result.  That being said, only 3 steps move it outside Maniraptoriformes, where it falls out by Zuolong and megaraptorans.  Guess there was some real signal keeping it outside Tyrannoraptora in Dal Sasso and Maganuvo's (2011) TWiG analysis.  Brusatte's redescription should shed light on Nedcolbertia's anatomy, as should the description of BYU 19114 from the Cedar Mountain Formation, said to be similar by Hunt and Quinn. 

Interestingly, Ornithomimus velox does not group with Dromiceiomimus (which includes O. edmontonicus here), but is instead down by Sinornithomimus.  The only character really grouping traditional Ornithomimus species together is metacarpal I being longer than metacarpal II, which was not a character in the Lori matrix but it still takes 3 steps to combine them.  Note Claessens and Loewen (2015) in their excellent redescription just assume Ornithomimus sensu lato is monophyletic based on the metacarpal length, and I don't think anyone's actually used O. velox as its own OTU before this.  It would be funny if Sinornithomimus ended up actually being the Chinese sister taxon to OrnithomimusAepyornithomimus is sister to this pair in the published matrix, but in a trichotomy with Ornithomimus, Sinornithomimus and more derived taxa in the new version (not shown).  It wasn't resolved in its original description's analysis using the Choiniere matrix, merely more derived than Archaeornithomimus.

Next is a novel clade of Tototlmimus and "Gallimimus" "mongoliensis", the first time the latter has been included in an analysis.  Only 3 steps are needed to move it sister to Gallimimus bullatus, and Tototlmimus follows.  Tototlmimus was poorly resolved in its initial analysis based on Kobayashi characters.  Another new clade follows- Rativates plus Kaiparowitz supposed O. velox specimen MNA Pl.1762A plus "Ornithomimus" sedens, the latter based only on the holotype.  Again, Rativates has only been analyzed once before in its initial description, using Choiniere's heavily misscored matrix, while sedens and the Kaiparowitz specimen have never been analyzed before.

Ornithomimosaur coracoids in right lateral and proximal views. Top left- Beishanlong (after Makovicky et al., 2010). Bottom left- Anserimimus (after Barsbold, 1988). Top right- Sinornithomimus (after Kobayashi, 2004). Bottom right- Gallimimus (after Kobayashi, 2004). Peach dot indicates coracoid tubercle, green line indicates lateral edge of infraglenoid buttress.

Here's probably a good place to say that ornithomimosaur phylogeny has suffered a similar fate to coelurosaur phylogeny lately because everyone reuses Kobayashi's characters and scores just like how everyone reuses TWiG characters and scores.  So everyone gets an Anserimimus plus Gallimimus clade, then an American clade of Struthiomimus and 'Ornithomimus'.  But the former clade is only based on two coracoid characters, both of which are flawed.   Above on the right we have Kobayashi's (2004) figure 88 from his thesis illustrating the characters.  Top right is Sinornithomimus and bottom right is Gallimimus.  The first character is "laterally offset infraglenoid buttress of the coracoid", represented by how much the green curve protrudes downward here.  A bit more in Gallimimus, but compare to the then unknown Beishanlong coracoid in the upper left.  It has a hugely protruding process but isn't a part of the Anserimimus plus Gallimimus clade.  It's not a commonly shown perspective, but is also found in Nqwebasaurus, Allosaurus, etc..   The second character is "biceps tubercle positioned more anterior to base of posterior process", which is the peach dot in the figure.  Here note that Beishanlong also has this anteriorly positioned, but more problematically Anserimimus in the lower left does not.  Maybe the drawing's wrong, but the rest of that figure seems accurate (e.g. the manus) and detailed unlike some of the more schematic ones in Barsbold's works (e.g. Adasaurus' pelvis).  Note that the more recent Xu et al. (2011) ornithomimosaur analysis that finds the Kobayashi arrangment misscores Beishanlong for both characters.  Instead, the Lori analysis (and my previous unpublished TWiG analysis incorporating Kobayashi's and other ornithimomid-relevent characters) recovers a Struthiomimus plus Gallimimus clade and an Anserimimus plus Dromiceiomimus clade.  Forcing my pairing of these four in Xu et al.'s matrix only adds 2 steps, but getting the standard Kobayashi arrangement of them in my matrix takes 8 steps, so here I think I really might be on to something.

Tyrannoraptoran femora in lateral and/or anterior views. Top left- Xiongguanlong (after Li et al., 2010). Top right- Alioramus (after Brusatte et al., 2012). Center- Timimus (after Benson et al., 2012). Bottom left- Archaeornithomimus (courtesy AMNH). Bottom right- Garudimimus in lateral view (after Kobayashi and Barsbold, 2005) and Gallimimus in anterior view (after Osmolska et al., 1972).  Peach indicates accessory trochanter in anterior view, green outline indicates accessory trochanter in lateral view.

Also in the Struthiomimus plus Gallimimus group, we get Timimus as the sister taxon to the latter genus.  But of course Timimus was reassigned to Tyrannosauroidea by Benson et al. (2012).  So what gives?  Well, if you look at Benson et al.'s reasoning for rejecting an ornithomimosaurian identification, they say "The morphology of the accessory trochanter and the relatively anteroposteriorly narrow lesser trochanter of NMV P186303 are similar to those of derived tyrannosauroids such as Xiongguanlong and tyrannosaurids. They are unlike the anteroposteriorly broad, ‘aliform’ lesser trochanter and prominent, triangular accessory trochanter of allosauroids, ornithomimosaurs ..."  "NMV P186303 lacks several features present in all ornithomimosaurs, such as the ‘aliform’ lesser trochanter and prominent accessory trochanter. In contrast, the lesser trochanter of some tyrannosauroids is anteroposteriorly narrower, and the accessory trochanter forms a transversely thickened region, similar to the condition in T. hermani (e.g., Tyrannosaurus)."  "T. hermani also possesses a proximomedially inclined (‘elevated’) femoral head, a synapomorphy of derived tyrannosauroids (e.g., Tyrannosaurus; Xiongguanlong), that is absent in ornithomimosaurs." 

First of all, Gallimimus has an elevated femoral head too (Osmolska et al., 1972: Plate XLVII).  Not as much as most tyrannosauroids', but neither does Timimus.  Of course tyrannosauroids have always been recognized as having aliform anterior trochanters as well, so this is a matter of degree.  Their figure 19B does look anteroposteriorly narrower (~49% of total femoral width; ~49% in Xiongguanlong, ~52% in Alioramus), unlike Archaeornithomimus (~69%).  But Garudimimus' ratio is ~50%.  Also note figure 19C, also labeled as lateral view, looks broader (~64%) and no doubt had a slightly more anterior angle to the photo.  Even ignoring Garudimimus, something that depends so heavily on exact angle of perspective, especially considering taphonomy and how theropod femoral heads phylogenetically vary in their anterior angle compared to the distal end (basal forms are famously more anteromedially directed), is not great evidence in my opinion.  What about that accessory trochanter?  I agree Timimus' is more tyrannosauroid in side view, but ironically because they're larger than at least Archaeornithomimus and Gallimimus (green highlight), contra Benson et al.'s statement.  And again some ornithomimosaurs like Garudimimus have large accoessory trochanters too.  Regarding transverse width, I can't see a difference between e.g. Alioramus and Gallimimus above (peach highlight).  I certainly wouldn't say Alioramus' is thicker.  So is Timimus a tyrannosauroid or an ornithomimosaur?  I don't think the evidence is great either way, and certainly no published analysis scores for these difficult to quantify degrees of trochanter size.  Honestly, the biostratigraphy makes me think it will ultimately be some coelurosaur convergent with both.  Maybe something like Aniksosaurus, also Gondwanan Early Cretaceous with a tall and narrow anterior trochanter.

Finally, Qiupalong joins the Anserimimus plus Dromiceiomimus clade. In its description, it grouped in the American clade, but that's the same Xu et al. (2011) analysis noted above that misscores Beishanlong as lacking the supposed Anserimimus plus Gallimimus characters.

So that's the Ornithomimosauria.  I think the Lori analysis does a good job here doing one of the things it's meant to- include a ton of taxa that have either never been analyzed or were only added singly and separately to existing analyses.  Another point I like to emphasize is the hidden instability of our consensus.  You might be thinking 'well your analysis seems very poorly supported if all of these tested changes only take 3 to 4 steps each'.  Yet you can rearrange the entire tree of Xu et al.'s (2011) ornithomimosaur analysis to my topology and it just needs 5 more steps in total.  And Brusatte et al.'s (2014) tree doesn't even find resolution between Harpymimus, Beishanlong, Garudimimus, Archaeornithomimus, Sinornithomimus and the derived clade.  Overall I'd say this is the best ornithomimosaur analysis published, in taxon number, character number and robusticity of results.

Next, alvarezsauroids...

References- Osmólska, Roniewicz and Barsbold, 1972. A new dinosaur, Gallimimus bullatus n. gen., n. sp. (Ornithomimidae) from the Upper Cretaceous of Mongolia. Palaeontologica Polonica. 27, 103-143.

Barsbold, 1988. A new Late Cretaceous ornithomimid from the Mongolia People’s Republic. Journal of Paleontology. 1988(1), 122-125.

Smith and Galton, 1990. Osteology of Archaeornithomimus asiaticus (Upper Cretaceous, Iren Dabasu Formation, People's Republic of China). Journal of Vertebrate Paleontology. 10(2), 255-265.

Kobayashi, 2004. Asian ornithomimosaurs. PhD thesis. Southern Methodist University. 340 pp.

Kobayashi and Barsbold, 2005. Reexamination of a primitive ornithomimosaur, Garudimimus brevipes Barsbold, 1981 (Dinosauria: Theropoda), from the Late Cretaceous of Mongolia. Canadian Journal of Earth Sciences. 42(9), 1501-1521.

Li, Norell, Gao, Smith and Makovicky, 2010. A longirostrine tyrannosauroid from the Early Cretaceous of China. Proceedings of the Royal Society B. 277(1679), 183-190.

Makovicky, Li, Gao, Lewin, Erickson and Norell, 2010. A giant ornithomimosaur from the Early Cretaceous of China. Proceedings of the Royal Society B. 277, 191-198.

Dal Sasso and Maganuco, 2011. Scipionyx samniticus (Theropoda: Compsognathidae) from the Lower Cretaceous of Italy: Osteology, ontogenetic assessment, phylogeny, soft tissue anatomy, taphonomy, and palaeobiology. Memorie della Società Italiana di Scienze Naturali e del Museo Civico di Storia Naturale di Milano. 281 pp.

Xu, Kobayashi, Lu, Lee, Liu, Tanaka, Zhang, Jia and Zhang, 2011. A new ornithomimid dinosaur with North American affinities from the Late Cretaceous Qiupa Formation in Henan Province of China. Cretaceous Research. 32(2), 213-222.

Benson, Rich, Vickers-Rich and Hall, 2012. Theropod fauna from Southern Australia indicates high polar diversity and climate-driven dinosaur provinciality. PLoS ONE. 7(5), e37122.

Brusatte, Carr and Norell, 2012. The osteology of Alioramus, a gracile and long-snouted tyrannosaurid (Dinosauria: Theropoda) from the Late Cretaceous of Mongolia. Bulletin of the American Museum of Natural History. 366, 197 pp.

Brusatte, Lloyd, Wang and Norell, 2014. Gradual assembly of avian body plan culminated in rapid rates of evolution across the dinosaur-bird transition. Current Biology. 24(20), 2386-2392. DOI: 10.1016/j.cub.2014.08.034

Lee, Barsbold, Currie, Kobayashi, Lee, Godefroit, Escuillie and Tsogtbaatar, 2014. Resolving the long-standing enigmas of a giant ornithomimosaur Deinocheirus mirificus. Nature. 515, 257-260.

Claessens and Loewen, 2015. A redescription of Ornithomimus velox Marsh, 1890 (Dinosauria, Theropoda). Journal of Vertebrate Paleontology. e1034593 DOI: 10.1080/02724634.2015.1034593

Sues and Averianov, 2015. Ornithomimidae (Dinosauria: Theropoda) from the Bissekty Formation (Upper Cretaceous: Turonian) of Uzbekistan. Cretaceous Research. 57, 90-110.

Brownstein, 2017. Redescription of Arundel Clay ornithomimosaur material and a reinterpretation of Nedcolbertia justinhofmanni as an "ostrich dinosaur": Biogeographic implications. PeerJ 5:e3110. DOI: 10.7717/peerj.3110

Hunt and Quinn, 2018. A new ornithomimosaur from the Lower Cretaceous Trinity Group of Arkansas, Journal of Vertebrate Paleontology. e1421209. DOI: 10.1080/02724634.2017.1421209

Hartman, Mortimer, Wahl, Lomax, Lippincott and Lovelace, 2019. A new paravian dinosaur from the Late Jurassic of North America supports a late acquisition of avian flight. PeerJ. 7:e7247. DOI: 10.7717/peerj.7247


  1. Are we sure Timimus isn’t a ceratosaur?

    1. Nope, femora can converge so much. Just look at Gualicho.

  2. Nice work Mickey.

    Having the Angeac taxon as a ceratosaur perhaps makes more sense, given the highly reduced forelimbs reported for this taxon. (Should we call it _Valdoraptor_, in your opinion?)

    What's happened to _Pelecanimimus_?

    "Grusimimus", should it ever be named after a crane (_Grus_), should be spelled "Gruimimus". It would be nice if the genus could be named properly (seems to be a difficult thing to do lately.)

    1. Thanks Tim. I wouldn't call it Valdoraptor yet. As I say on the Database- "If Allain et al. were correct that the genus was almost identical to and shared a synapomorphy with their specimens, it would make their unnamed taxon Valdoraptor, it wouldn't make Valdoraptor indeterminate. However, they fail to mention that the Angeac taxon lacks the dorsolateral ridge on metatarsal II Naish and Martill used to diagnose Valdoraptor, or the suboval interosseous space between metatarsals II and III that Naish used to diagnose the genus in 2011. They also misunderstand Naish's short metatarsal II length as only meaning it was shorter than III (which is near universal in theropods), whereas he was referring to the extreme shortness (81%) that is shorter than other ornithomimosaurs. The length of metatarsal II in the Angeac taxon is unpublished still, as only partial metatarsals have been figured. In any case, the ridge and interosseous space validate Valdoraptor and distinguish it from the Angeac taxon, though they do appear to be related." Also "Carrano et al. (2012) refer it to Avetheropoda based on the trapezoidal cross section of metatarsal III", and I'm not sure if any of these tetanurine-like ceratosaurs have that character.

      Pelecanimimus emerged as an alvarezsauroid, so expect that next time.

      The funny thing about "Grusimimus" is that the description's already written in Kobayashi's (2004) thesis. "All he'd have to do" is port it directly to a published paper like he did for the Harpymimus, Garudimimus and Sinornithomimus sections of his thesis. But after my experience with Lori's peer review, maybe this isn't such an obviously easy road after all.

  3. Timimus is already classified as a tyrannosaur by Benson et al. (2012) and Delcourt & Grillo (2018).

    1. I addressed Benson et al.'s arguments above.

      Delcourt and Grillo included it in three analyses. In Porfiri et al.'s it emerged sister to Tyrannoraptora, not in Tyrannosauroidea.

      In Choiniere et al.'s it is tyrannosauroid based on the dorsally angled femoral head (addressed above), cylindrical anterior trochanter (not true in Timimus or tyrannosauroids) and anterior trochanter "that is as proximal or more proximal than the greater trochanter", which is not true in basal tyrannosauroids like Xiongguanlong and is probably why they recovered it within Tyrannosaurus.

      Finally, in Carr et al.'s it is a tyrannosauroid based on anterior "trochanter and the greater trochanter extending to approximately the same level proximally" and "Proximal margin of the femur is concave in posterior view" which are also present in e.g. Archaeornithomimus but their Ornithomimosauria OTU is scored 0 for both. The final supporting character is "shallow femoral extensor groove on the anterior surface of the distal end that is expressed as a broad concave anterior margin in distal view but present as an extensive depression on the anterior surface of the femur" which is wrong and should have been scored 0 (compare to Juratyrant, the only other taxon scored 1).

      So I don't think Delcourt and Grillo's analyses support this either.

  4. The Shishugounykus cladistic analysis still recovers Nqwebasaurus and Pelecanimimus in their traditional ornithomimosaur placement rather than as alvarezsaurs. Bear in mind that Ornithomimosauria hasn't yet been reported from the Jurassic, and since Compsognathidae is recovered as part of Maniraptora in the Xu et al. (2018) and Qin et al. (2019) analyses, it is possible that some compsognathids are closer to alvarezsaurs than to Compsognathus.

    1. The Shishugounykus analysis is Choiniere's, which recall from the Sciurumimus reanalysis is heavily misscored. So for instance, Pelecanimimus is misscored for both otic recess placement and metacarpal II proximal expansion, noted as supporting alvarezsauroid placement in the Lori paper. Btw, I added Shishugounykus yesterday and it groups with Haplocheirus as a compsognathid.

      I'm skeptical compsognathids are maniraptorans, since it takes 8 more steps to move them there in the Lori matrix. Also, while a compsognathid plus alvarezsauroid clade seemed intuitively neat with so many taxa being recovered in Choiniere's matrix as alvarezsauroids and in mine as compsognathids, it still takes 12 more steps to enforce that even after Shishugounykus.

  5. what are your thoughts of "Coelosaurus" antiquus?

    1. Nothing past , since there aren't enough tibial characters in the Lori matrix to place it.

  6. Nice to see Rativates come out separate from the other named Albertan ornithomimids, it was in a polytomy with them in our original flawed analysis. I agree that the recovered position of Timimus is biogeographically and stratigraphically implausible. Does the rest of the ornithomimid topology change if Timimus is simply ignored?