Wednesday, February 24, 2016

Fukuivenator thoughts

So we have another new theropod, Fukuivenator paradoxus (Azuma et al., 2016).  With Fukuiraptor, Fukuititan and Fukuisaurus also in existence, we just need a Fukuipelta, Fukuimimus and Fukuiornis to round out the Kitadani Formation.  This was first announced in an anonymous 2009 msn article (now defunct) and Shibata and Azuma's (2010) SVP abstract.  While those sources called it a dromaeosaurid, Azuma et al. interestingly say "it in fact represents a bizarre, basally branching maniraptoran theropod with a large number of autapomorphies."  Intriguing.  Let's check it out...

A partial skull is present (though the jugal, palatine, ?ectopterygoid and pterygoid are unillustrated) as well as a dentary fragment.  The weirdest thing is the gigantic premaxilla, which is so odd that I would question its identity and propose it's actually the right maxilla, except that it lacks the pneumatic fossa and foramina present on the left maxilla.  Contra the authors, the posterolateral process of the premaxilla is broken so (barring the presence of some really well preserved sutures on the maxilla) it cannot be determined if it extends posterior to the external naris like in dromaeosaurids.  Similarly, the supposedly large promaxillary fenestra (cited in the diagnosis as well) is actually small.  The text claims the lacrimal is T-shaped, but the figure shows the posterodorsal process is unpreserved, and the character is coded unknown in their matrix.  Although the authors claim the frontals have dromaeosaurid-like anterolateral notches and sigmoid supratemporal fossa edges, neither is visible in the figured left element.  Amusingly, Azuma et al. say one premaxillary tooth "exhibits a "ridge," on which denticles are absent."  Ah, you mean a carina? ;)  The other premaxillary tooth is indeed interesting in being D-shaped.  Surely the fenestra labeled "IX?" in figure 4  is the otic fenestra, and the basipterygoid processes are confusingly labeled as laterosphenoids twice in that figure.  The supposed medial eustachian foramina are the paired foramina of the basisphenoid recess.  Oddly, neither of the eustachian tube characters are coded, but the basisphenoid recess characters are coded correctly.  It's an interesting braincase, with no obvious dorsal or posterior tympanic recesses, a reduced basisphenoid recess and laterally diverging basipterygoid processes, most of which are troodontid-y.

Reconstructed skull of Fukuivenator paradoxus holotype, premaxilla and dentary flipped, and frontal shown in dorsal view (modified from Azuma et al., 2016).

An almost complete vertebral series is present.  The description states "ten cervical vertebrae are preserved, missing at least the atlas but the materials list only says "eight cervical vertebrae" are present.  Hmm.  The text states "the middle cervical vertebrae have a highly modified hyposphene-hypantrum with the hypantrum extending ventrally below the dorsal border of the neural canal (Fig. 6b)", but the figure shows nothing extending ventrally into the neural canal and I thought hyposphene-hypantrum articulations were absent from most theropod anterior dorsals, let alone cervicals.  "Dorsal centra are longer anteroposteriorly than tall dorsoventrally, unlike the dorsal centra of typical predatory theropods."  :|  Tell that to coelophysoids, compsognathid-grade taxa, troodontids, microraptorians...  "Likely pleurocoels are present in all dorsal vertebrae in the form of longitudinal fossae on the lateral surfaces of centra."  Haven't we progressed since the days of Osmolska et al. and Welles calling central fossae pleurocoels?  This is another character coded as dromaeosaurid-like, but the posterior dorsals (at least) actually lack pleurocoels as is standard for theropods.  Azuma et al. also state "the parapophyses of the dorsal vertebrae including the posterior ones are stalk-like as in derived alvarezsauroids and dromaeosaurids, though they are not as prominent as in the latter groups", but they're actually short.  My matrix uses a ratio of parapophyseal length (measured from the ventromedial corner to the apex) and centrum width, so that even Alxasaurus and some Sinraptor and Allosaurus vertebrae count as apomorphic (contra the unchanging TWG matrix), but Fukuivenator still falls short.  The authors state "the [sacral] zygapophyses are fused to each other to form a platform lateral to the neural spines, a feature also known in dromaeosaurids", but the figured sacrum lacks neural arches in the first four vertebrae.  Azuma et al. say the sacral ribs+diapophyses are "bifurcated distally to contact the ilium, a feature previously unreported in any other theropod", but this is true in e.g. the middle four sacrals of Gallimimus.  They also say "the most unusual feature is that the prezygapophyses of the middle caudal vertebrae are distally bifid (Fig. 6i), which has not been reported in any dinosaurs", but this is a standard dromaeosaurid character reported in e.g. Deinonychus (Ostrom, 1969) and Velociraptor (Norell and Makovicky, 1999).

Pes of Fukuiraptor paradoxus holotype, as shown in figure 7h on left, and with my reidentification of the phalanges on right (modified from Azuma et al., 2016).  Note the longer phalanges on digit II as in other theropods, and how digit II isn't particularly developed into a sickle-claw.
Contra the text and coding, the coracoid is proximodistally shallow, unlike pennaraptorans.  Surely, figure 7's caption is incorrect and the humeri shown are a right in anterior view and left in lateral view, not the left "in lateral (left) and posterior (right) view."  Also, the femur in figure 7f is in medial and posterior views, not lateral and posterior.  In figure 7h, the pedal phalanges are almost certainly placed incorrectly, with II-1 and II-2 switched with IV-1 and IV-2, which explains why digit II looks ridiculously short (see figure above).  "IV-1" has the standard shape of II-1 with the medial side highly concave (so it actually belongs on the other foot), plus I don't think any terrestrial Mesozoic theropod has IV-1 longer than II-1.  The measurement table is partly inconsistent as it has IV-2 subequal to IV-3 and IV-4 in length, unlike the figure.  Contra the text and coding, I don't think the second pedal digit looks particularly deinonychosaurian- Tanycolagreus has the same dorsally prominent distal articular surface on II-1, and the ungual in Ornitholestes is comparatively larger.  Alas, the text continues to use the II-III-IV manual digit identification, thanks to Xu, which at least Choiniere has disavowed (in his recent Ceratosaurus forelimb paper with Carrano).  It was an intriguing idea Xu, but it's long past time to let it die.

Azuma et al. add Fukuivenator to Turner et al.'s (2012) TWG analysis, which is unfortunately flawed by having numerous taxa left uncoded for entire sections of the matrix, and most of the codings retained from the original TWG matrix of Norell et al. (2001).  So don't put much credit into e.g. the joining of oviraptorosaurs and therizinosaurs, because the latter are only represented by the 2001 codings for Alxasaurus, Segnosaurus and Erlikosaurus.  No Beipiaosaurus, let alone Falcarius or Jianchangosaurus.  In any case, contra the text, they do not recover Fukuivenator as a maniraptoran.  In their trees, it's actually in a polytomy with compsognathids, Ornitholestes, ornithomimosaurs and maniraptorans.  They also didn't find "Anchiornis and Xiaotingia outside the Troodontidae", instead the base of Deinonychosauria is unresolved with those genera part of the polytomy.  Frustratingly, the authors never try excluding taxa a posteriori, so we don't know if e.g. the underlying structure of Turner's tree is still there and Fukuivenator can fall out in multiple places, or if it creates an actual polytomy between some taxa there.  Running the matrix myself, I find the latter is true- Fukuivenator's character combination creates a polytomy there even if it's pruned from the tree a posteriori.  Also, the Anchiornis+Xiaotingia situation exists because in a minority of trees these are sister to microraptorians+eudromaeosaurs instead of being basal troodontids.  Interestingly, the authors do check how parsimonious it is to place Fukuivenator in alternative positions and find it only takes three more steps to make it a paravian, deinonychosaur or dromaeosaurid. 

Adding Fukuivenator to the Lori matrix recovers it in a trichotomy with Ornithomimosauria and Maniraptora.  Makes sense considering its general basal coelurosaur grade anatomy.  There are a few dromaeosaurid-like characters- the fossa around the maxillary fenestra, the short ventral postorbital process, the squamosal shelf over its ventral process, the twisted paroccipital processes, the reduced crista prootica, the bifurcated caudal prezygapophyses.  But the vast majority are quite unlike paravians, and as seen above, many of the supposedly dromaeosaurid-like characters aren't present.  Andrea Cau recovered Fukuivenator as sister to Pennaraptora in his Megamatrix, which isn't far from my result as only alvarezsauroids and therizinosaurs are between the two Fukuivenator positions, basal members of which (e.g. Haplocheirus, Falcarius) are like Ornitholestes-grade coelurosaurs in much of their anatomy. 

References- Ostrom, 1969. Osteology of Deinonychus antirrhopus, an unusual theropod from the Lower Cretaceous of Montana. Bulletin of the Peabody Museum of Natural History. 30, 165 pp.

Norell and Makovicky, 1999. Important features of the dromaeosaurid skeleton II: Information from newly collected specimens of Velociraptor mongoliensis. American Museum Novitates. 3282, 45 pp.

Anonymous, 2009. [3rd new species? Small-sized meat diet dinosaur to restoration Fukui] 3/18/2009 

Shibata and Azuma, 2010. New dinosaurs from the Lower Cretaceous Kitadani Formation of the Tetori Group, Fukui, Central Japan. Journal of Vertebrate Paleontology. Program and Abstracts 2010, 163A-164A.

Turner, Makovicky and Norell, 2012. A review of dromaeosaurid systematics and paravian phylogeny. Bulletin of the American Museum of Natural History. 371, 1-206.

Azuma, Xu, Shibata, Kawabe, Miyata and Imai, 2016. A bizarre theropod from the Early Cretaceous of Japan highlighting mosaic evolution among coelurosaurians. Scientific Reports. 6, 20478.


  1. Fukuiceratops, Fukuityrannus, Fukuiavis, Fukuidon, Fukuinykus... yes.

  2. A new osteology of Fukuivenator has just been published and they corrected the arrangement of the pedal phalanges. Unfortunately they did not credit you.

    1. Note also I stated "The weirdest thing is the gigantic premaxilla, which is so odd that I would question its identity and propose it's actually the right maxilla", which turned out to be exactly what happened.

      Turns out they misidentified a lot more as well- "right lacrimal, originally identified as right postorbital; atlantal neurapophyses, originally identified as axial ribs; partial neural arch of fifth dorsal vertebra, originally identified as left squamosal; right manual phalanx III-1, originally identified as right pedal phalanx I-1; right pubis, originally identified as left pubis; left metatarsals II and V, originally identified as right metatarsals IV and V, respectively". Then there are the mysterious "Withdrawn elements: left pterygoid; posterior caudal vertebra; left ischium."

      Resolving other things I mentioned, the lacrimals aren't T-shaped and the frontals lack anterolateral notches, although they do have sinuous supratemporal fossa margins. IX was the otic fenestra and the eustachian tubes were basisphenoid foramina. Nine cervicals are preserved, and do not have hyposphenes as the ventrally projecting median process is separated from the postzygapophyses. None of the dorsals have pleurocoels and their parapophyses are short, but the newly figured sacral neural arches do show a platform. Among new issues, what Hattori et al. (2022) call a subotic recess in figure 12C is just what they (correctly) call an anterior tympanic recess in figure 12E.