How the shark lost its bones video on YouTube

From Martin Brazeau and the Imperial College London,
here’s a new YouTube video (53 minutes) on how and maybe why sharks lost their bony exoskeleton.

The phylogenetic context is wrong. Without testing, Brazeau et al. considered placoderms basal to sharks and bony fish.That’s a traditional mistake. In the large reptile tree (LRT, 1795+ taxa) placoderms are bony fish close to catfish. In the LRT sharks evolved from sturgeons (Fig. 2). Bony fish evolved from hybodontid sharks. The Silurian is when all this happened.

We looked at this subject earlier here (Borrell 2014) and here Brazeau et al. 2020.

Unfortunately, as you’ll see
Brazeau et al. include only fossil taxa to determine which taxa were present in the Silurian.

Figure x. Shark skull evolution.

The jawless,
(by reversal) anapsid-mimic placoderm, Minjinia (Fig. 3) was featured in Brazeau’s paper and video.

Figure 1. Subset of the LRT focusing on the branch of the Osteichthys that includes placoderms and their relatives.
Figure 2. Subset of the LRT focusing on the branch of the Osteichthys that includes placoderms and their relatives.
Figure 1. Minjina in 4 views, mirror-image and colors added.
Figure 1. Minjina in 4 views, mirror-image and colors added.

Ironically
Brazeau illustrates his talk with an image of the exoskeleton and endoskeleton of the sturgeon Acipenser. which entered the LRT here. He reports the endochondral bone was lost in sturgeons. That is a traditional mistake as revealed by the LRT.

Brazeau correctly reports
the origin of bone precedes sharks and is lost in sharks. He just did not realize that placoderms are descendants of sharks, not their ancestors.


References
Brazeau et al. (7 co-authors) 2020. Endochondral bone in an Early Devonian ‘placoderm’ from Mongolia. Nature Ecology & Evolution. https://doi.org/10.1038/s41559-020-01290-2
Hu Y, Lu J and Young GC 2017. New findings in a 400 million-year-old Devonian placoderm shed light on jaw structure and function in basal gnathostomes. Nature Scientific Reports 7: 7813 DOI:10.1038/s41598-017-07674-y

https://cosmosmagazine.com/nature/evolution/new-thoughts-on-how-sharks-evolved/http://www.sci-news.com/paleontology/minjinia-turgenensis-08823.html

Adding taxa updates the origin of placoderms

A year ago
when fish (= basal vertebrates) were first added to the large reptile tree (LRT, now with 1757+ taxa; subset Fig. 1), the extant walking catfish, Clarias, nested with the Silurian placoderm, Entelognathus, rather than any other extant bony fish when there were very few other bony fish to nest with. Since then, adding taxa has separated these two, but they still nest as charter members of the unnamed catfish-placoderm clade. That was a heretical hypothesis then, and it remains so today.

Traditional fish paleontologists
consider placoderms basal to sharks and ratfish + bony fish. Stensioella was considered the most basal placoderm by Carr et al. 2009, who did not list outgroup taxa. These hypotheses are not supported by the LRT (subset Fig. 1) where placoderms arise from coelacanths among the bony fish, far from sharks and ratfish.

The LRT divides placoderms into four clades;

  1. Arthrodira (open ocean predators like Dunkleosteus, Coccosteus, Fig. 3)
  2. Antiarchi (armored jawless bottom dwellers like DicksonosteusBothriolepis, Fig. 2)
  3. Ptyctodontida (chimaera-like taxa like Australoptyctodus, Fig. 2)
  4. Phyllolepida (tiny-eye taxa like Entelognathus, Cowralepis)

Several traditional placoderms nest elsewhere in the LRT.

  1. Rhenanida – nests with catfish in the LRT
  2. Wuttagoonaspis – nests with catfish in the LRT
  3. Stensioellida – nests with Guiyu-like lobefins in the LRT
  4. Brindabellspida – nests with the tetrapodomorph Elpistostege

Several traditional placoderms have not yet been tested in the LRT.

  1. Petalichthyida (includes Diandongpetalichthys)
  2. Acanththoraci (closely related to rhenanids, nesting with catfish)
  3. Pseudopetalichthyida (similar to rhenanids, nesting with catfish)

After testing
in the LRT (subset Fig. 1) placoderms are still bony fish close to catfish and this clade still arises from coelocanths.

Figure 1. Subset of the LRT focusing on the branch of the Osteichthys that includes placoderms and their relatives.

Figure 1. Subset of the LRT focusing on the branch of the Osteichthys that includes placoderms and their relatives.

The pertinent taxa in the first list
(Fig. 2) start with the small, Early Devonian spiny shark Diplacanthus and end with the rather flat nearly jawless placoderm, Dicksonesteus also from the Early Devonian. That tells us that every taxon between them was part of the Early Devonian fauna. That also tells us the radiation of taxa in figure 2 must have occurred much earlier, sometime in the middle of the mysterious Silurian, which preserves very few gnathostome fish fossils.

Figure 2. Taxa from the LRT nesting prior to the clade Placodermi.

Figure 2. Taxa from the LRT nesting prior to the clade Placodermi. See figure 3 for the arthrodire clade within Placodermi. Robustichthys is basal to catfish and lacks a squamosal.

Phylogenetic miniaturization
occurs at the origin of placoderms with the smallest specimen in figure 2, Romundina, half the size of its predecessor, Eurynotus. In like fashion, the smallest placoderm in figure 3is the unnamed ANU V244 specimen, is also half the size of its predecessor, the aforementioned Eurynotus.

Figure 3. Arthrodires and their ancestor, Euryodus. See figure 2 for Euryodus ancestors. Note the phylogenetic miniaturization at the origin of the arthrodires.

Figure 3. Arthrodires and their ancestor, Euryodus. See figure 2 for Euryodus ancestors. Note the phylogenetic miniaturization at the origin of the arthrodires.

Phylogenetically, the lack of marginal teeth
in placoderms goes back to a late-surviving taxon from the Jurassic, the angelfish-mimic,  Cheirodus (Fig. 1). Note the hidden palatine teeth in Cheirodus that in the arthrodires, Coccosteus and Dunkleosteus become visible and act as marginal teeth/plates. The Silurian ancestors of Cheirodus may not have been so uniquely angelfish-like. That shape is apomorphic due to the separation in time.

Ptyctodonts, like Austroptyctodus,
(Fig. 2) do not nest with other traditional placoderms in the LRT, but nest closer to Cheirodus.  These are the sort of results the LRT recovers only because it tests more taxa.


References
Carr RK, Johanson Z and Ritchie A 2009. The phyllolepid placoderm Cowralepis mclachlani: Insights into the evolution of feeding mechanisms in jawed vertebrates. Journal of Morphology. 270 (7): 775–804.
Hu Y, Lu J and Young GC 2017. New findings in a 400 million-year-old Devonian placoderm shed light on jaw structure and function in basal gnathostomes. Nature Scientific Reports 7: 7813 DOI:10.1038/s41598-017-07674-y
Miles RS and Young GC 1977. 
Placoderm interrelationships reconsidered in the light of new ptyctodontids from Gogo Western Australia. Linn. Soc. Symp. Series 4: 123-198.
Young GC 1980. A new Early Devonian placoderm from New South Wales, Australia, with a discussion of placoderm phylogeny: Palaeontographica 167A pp. 10–76. 2 pl., 27 fig.
Zhu et al. 2012. An antiarch placoderm shows that pelvic girdles arose at the root of jawed vertebrates. Biology Letters Palaeontology 8:453–456.
Zhu M, Yu X-B, Ahlberg PE, Choo B and 8 others 2013. A Silurian placoderm with osteichthyan-like marginal jaw bones. Nature. 502:188–193.
Zhu M et al. 2016. A Silurian maxillate placoderm illuminates jaw evolution. Science 354.6310 (2016): 334-336.

wiki/Entelognathus
wiki/Bothriolepis
wiki/Dicksonosteus
wiki/Romundina
wiki/Qilinyu
wiki/Parayunnanolepis
wiki/Lunaspis
wiki/Coccosteus
wiki/Mcnamaraspis
wiki/Dunkleosteus

Revisions to the catfish + placoderms subset of the LRT

Things were not quite right in the catfish-placoderm clade,
so a critical examination of the traits and scores was due.

As longtime readers know,
every new taxon added to the LRT is a new experience, scored to the best of my nascent ability each time. When the first few taxa were scored, I had little to no experience with any fish. Now, with a substantial taxon list, comparisons can be reexamined that were overlooked or not present before.

Because evolution works gradually,
bones and proportions that appear on one taxon should also appear on closely related taxa. Here tetrapod labels were put on all fish skull bones, so the traditional published fish skull bone labels were not as helpful as they will be once other workers adopt this several times earlier proposed nomenclature standard.

Figure 1. Subset of the LRT focusing on the branch of the Osteichthys that includes placoderms and their relatives.

Figure 1. Subset of the LRT focusing on the branch of the Osteichthys that includes placoderms and their relatives.

The basic tree topology in this clade has not changed. 
A few of the taxa have been rescored (Fig. 1).

Figure 2. Menaspis armatas in situ. Colors added to bones and skin.

Figure 3. Menaspis armatas in situ. Colors added to bones and skin. White area above restores the displaced mandibles relative to one another.

Change #1:
A former odd Permian ‘placoderm’ with barbels, Menaspis (Fig. 2), moved over to the Siluriformes (catfish clade). That only makes sense.

Menaspis armata (Ewald 1848; Late Permian; > 15cm long) was described as the ‘last known arthrodire placoderm’. Here it nests with the catfish, Clarias and Wuttagoonaspis. The former skull spine is the displaced mandible. The former ‘horns’ are barbels. The orbit is somewhere under the barbels. The entire ventral half of the skull is missing here on the counterplate. This is a ventral view of the dorsal skull plates.

Figure 3. Tiny unnamed arthrodire, ANU V244-3 in various views.

Figure 3. Tiny unnamed arthrodire, ANU V244-3 in various views. The upper left image lacks jaws. The jaws are upper right. Palatal view at middle right.

Change #2:
A tiny arthrodire ANU 244  now nests basal to the open water predatory clade of large to giant placoderms.

ANU V244 (Hu, Lu and Young 2017; Early Devonian) is a tiny basal arthrodire. The authors provided several views of the skull, even dividing it in half to show upper and lower elements separately (Fig. 3). The authors followed tradition in the proposal that placoderms were basal to gnathostomes not realizing placoderms have lost the maxilla, like their sisters, the catfish.

Change #3:
Several cheek bones on other placoderms were re-identified following this holistic look at several taxa all at once. Each specimen contributed to the understanding of the clade. Placoderms are highly derived leaving no descendants in the Mesozoic or thereafter. Traditional cladograms nesting placoderms basal to sharks and bony fish are in error, according to the LRT, which tests a wide gamut without prejudice.

On that note: the traditional ptyctodontid ‘placoderms’, Astroptyctodus and Campbellodus (Fig. 4), still nest outside the clade that includes the other placoderms.

Figure 2. Cheirodus and Campbellodus to scale. These two nest together in there LRT.

Figure 4. Cheirodus and Campbellodus to scale. These two nest together in there LRT.

The radiation of catfish and placoderms
must have happened deep in the Silurian with late survivors among the tested taxa.

Placoderms developed internal fertilization
with claspers and live birth of a few large young, convergent with sharks and manta rays.

On the other hand, catfish retained external fertilization
with thousands of eggs produced by a single female through several spawning periods. Typically 10% develop and survive. The first few spawnings produce none or fewer than five eggs.


References
Ewald J 1848. Über Menaspis, eine neue fossile Fischgattung. Berichte Über die zur Bekanntmachung Geeigneten Verhandlungen der Königlich-Preussischen Akademie der Wissenschaften zur Berlin 1848:33-35.
Hu Y, Lu J and Young GC 2017. New findings in a 400 million-year-old Devonian placoderm shed light on jaw structure and function in basal gnathostomes. Nature Scientific Reports 7: 7813 DOI:10.1038/s41598-017-07674-y

https://animaldiversity.org/accounts/Clarias_batrachus/

Kryptoglanis: a catfish mimic

Workers wondered, where did those big teeth come from?
Catfish don’t have such long, interlocking teeth. No wonder Vincent and Thomas 2011 called Kryptoglanis shajii (Figs. 1–3; 6cm) an ‘enigmatic’ subterranean catfish. No wonder Britz et al. erected a new clade of catfish to put it in.

Figure 1. The uncatfish-like teeth of Kryptoglanis in anterior view.

Figure 1. The uncatfish-like teeth of Kryptoglanis in anterior view.

According to Wikipedia
“[Kryptoglanis] has also been seen in dense vegetation in paddy fieldsThe species strongly avoids light and feeds on small invertebrates.”

“The morphology of K. shajii differs from all other known species of catfish and includes such features as the absence of dorsal fin; the presence of four pairs of barbels; an upwardly directed mouth, with a distinctly projecting lower jaw with 4 set[s] of teeth; subcutaneous eyes; anal fin completely confluent with the caudal fin; anal and caudal fins together carry 70–74 fin rays; and no spines in any of the fins.”

Figure 2. Kryptoglanis in 3 views. Note the catfish-like barbels.

Figure 2. Kryptoglanis in 3 views. Note the catfish-like barbels.

Figure 3. Skull of Kryptoglanis, a knife fish, not a catfish.

Figure 3. Skull of Kryptoglanis, a knife fish, not a catfish.

Here’s where we put on our lab coats and act like scientists.
Whenever a taxon is described as ‘differs from all other known species’, that’s the time to expand the taxon list. We’ve seen this so many times before. And every time the LRT has found less traditional and more parsimonious sister taxa simply by adding taxa.

Figure x. Gymnotus carapo in vivo.

Figure 4. Gymnotus carapo in vivo.

Figure 8. Skull of Gymnotus.

Figure 5. Skull of Gymnotus.

According to
the large reptile tree (LRT, 1750+ taxa) those catfish-like barbels on Kryptoglanis developed by convergence on an eel-like knife fish, nesting between the eel, Anguilla and two knifefish, Gymnotus and Electrophorus, the electric eel. This clade of fish DO have large interlocking teeth and a long list of other traits shared with Kryptoglanis. They just don’t have barbels. Seems like prior authors were caught “Pulling a Larry Martin” by focusing on the barbels to the exclusion of all the other traits.

Figure 1. Subset of the LRT focusing on the ray fin only clade of bony fish. Fundulus (yellow) is the new taxon. It attracted Anableps. Various convergent eel-like taxa are shown in baby blue.

Figure 1. Subset of the LRT focusing on the ray fin only clade of bony fish. Fundulus (yellow) is the new taxon. It attracted Anableps. Various convergent eel-like taxa are shown in baby blue.


References
Vincent M and Thomas J 2011. Kryptoglanis shajii, an enigmatic subterranean-spring catfish (Siluriformes, Incertae sedis) from Kerala, India. Ichthyological Research. 58 (2): 161–165. doi:10.1007/s10228-011-0206-6.
Britz R, Kakkassery F and Raghavan R 2014. Osteology of Kryptoglanis shajii, a stygobitic catfish (Teleostei: Siluriformes) from Peninsular India with a diagnosis of the new family Kryptoglanidae. Ichthyological Exploration of Freshwaters. 24 (3): 193–207.

wiki/Kryptoglanis_shajii

Two taxa moved from here to there in the LRT yesterday

A little light housekeeping over the weekend
I falsified two nestings in the large reptile tree (LRT, 1680+ taxa; subset Fig. 1) and made the corrections. Falsifying hypotheses is what scientists do, or should do. It’s almost as rewarding correcting my own errors as anyone else’s… after the head slap, of course.

Figure x. Newly revised fish subset of the LRT

Figure 1. Newly revised fish subset of the LRT

The Middle Triassic ‘holostean,’
Robustichthys (Fig. 2) now nests with the extant armored catfish, Hoplosternum (Fig. 3). The bones are nearly identical. The (cyan) cheek plate originally identified as a maxilla on Robustichthys is re-identified as a jugal in pieces. The (dark red) large pterygoid is visible beneath the taphonomically displaced cheek bones. The only teeth visible are palatine and other palatal teeth. As readers know, catfish, like their placoderm sisters, have no maxilla.

Figure z. Skull of Robustichthys and reconstruction of same. Note resemblance to Hoplosternum (Fig. y).

Figure 2. Skull of Robustichthys and reconstruction of same. Note resemblance to Hoplosternum (Fig. y).

Figure y. Hoplosternum skull with bones identified as homologs to those in Robustichthys.

Figure 3. Hoplosternum skull with bones identified as homologs to those in Robustichthys.

Robustichthys luopingensis (Xu et al. 2014; Xu 2019; Middle Triassic) was described as the largest holstean fish of the Middle Triassic. Here the clade Holostei is polyphyletic. The jugal is further split up. The mandible has a tall coronoid process.

Hoplosternum littorale (Hancock 1828; 24cm length) is the extant kwi kwi, a South American catfish armored from nose to tail. It breathes air. The incurrent and excurrent nares are enormous.

Figure 2. The AMNH specimen of Stensioella in dorsal view along with a diagram in ventral view.

Figure 4. The AMNH specimen of Stensioella in dorsal view along with a diagram in ventral view.

The other change 
involved a crushed Early Devonian, Stensiolla (Fig. 4, mistakenly identified as a possible placoderm). It now nests with Late Silurian Guiyu (Fig. 5), a flat coelacanth relative in the LRT (Fig. 1) originally considered the oldest articulated osteichthyian (= bony fish). The big fore paddles in this lobefin were the giveaway. The rest became a matching game.

Figure 2. Guiyu in situ, DGS colors added here and used to create the flatter, wider reconstruction with paddles preserved.

Figure 5. Guiyu in situ, DGS colors added here and used to create the flatter, wider reconstruction with paddles preserved.

Guiyu oneiros (Zhu et al. 2009; late Silurian; V17914) is the basalmost crossopterygian in the LRT, yet the lobes are barely developed and large, placoderm-like cleithra precede them. The revised restoration differs somewhat from the original in having a smaller orbit, a tall jugal, a convex maxilla, a concave dentary and no teeth in the dentary. A flatter oblate cross-section is envisioned for the torso.

Stensioella heintzi (Broili 1933; Emsian, late Early Devonian; 27 cm est.) is widely considered the most basal placoderm, but here nests with Guiyu, with which it shares most traits including enlarged lobe fins and a flattened head with fused roofing bones.  Stensioellawould have had a lifestyle like Squatina, the extant angel shark.

Both overall and in the smallest detail
sister taxa should document microevolutionary (= gradual, small) changes. That’s what phylogenetic software does for us, but the scoring must be accurate. Apologies for the earlier inaccuracies whether the errors were my own or copied from those with firsthand access to the fossils.


References
Broili F 1933. Weitere Fischreste aus den Hunsrickschiefern. Situngsbirechte der bayerischen Akademie der Wissenschaften, Mathematisch-Naturewissenschaftliche Klasse 2: 269–313.
Hancock J 1828. Notes on some species of fishes and reptiles, from Demerara, presented to the Zoological Society by John Hancock, Esq., corr. memb. Zool. Soc. In a letter addressed to the secretary of the Society. Zoological Journal, London v. 4: 240-247.
Xu G-H, Zhao L-J and Coates MI 2014. The oldest ionoscopiform from China sheds new light on the early evolution of halecomorph fishes. Biology Letters 10(5):20140204
DOI 10.1098/rsbl.2014.0204.
Zhu M, Zhao W, Jia Lu J, Qiao T and Qu Q 2009. The oldest articulated osteichthyan reveals mosaic gnathostome characters. Nature 458:469-474

wiki/Hoplosternum
wiki/Robustichthys
wiki/Guiyu
wiki/Stensioella

Origin and evolution of gnathostome dentitions

Updated January 2, 2021
with a new reconstruction of Gemuendina (Fig. x), which now appears to nest basal to Manta, close to Jagorina, but will not be entered into the LRT due to the large amount of skin and scale covering bone.

Figure x. Gemuendina in situ. So much skin and ornament cover the bone, this taxon has been withdrawn from  the LRT.

Figure x. Gemuendina in situ. So much skin and ornament cover the bone, this taxon has been withdrawn from the LRT.

Johanson and Smith 2005
looked at the questions of teeth and pharyngeal denticles in placoderms.

Unfortunately
the large reptile tree (LRT, 1597+ taxa; subset Fig. 1) does not confirm the first sentence of the authors’ abstract: “The fossil group Placodermi is the most phylogenetically basal of the clade of jawed vertebrates but lacks a marginal dentition comparable to that of the dentate Chondrichthyes, Acanthodii and Osteichthyes (crown group Gnathostomata).”

The LRT nests placoderms along with catfish
between Hybodus and spiny sharks, deep into the Gnathostomata. Catfish are not mentioned in the Johanson and Smith text. They do mention, “the rounded or pointed denticles described for the Arthrodira may only be present in a limited number of taxa (Gemuendina (Fig. 2) Traquair, 1903).” Regrettably the authors did not know that some members attributed to this generic wastebasket of Gemuendina are catfish (Fig. 1), a clade closely related to traditional placoderms. So, taxon exclusion, once again, becomes a major issue.

Figure 7. the KGM 1983 306 specimen referred to Gemuendina. This one is closer to the extant channel catfish, Ictalurus.

Figure 1. The KGM 1983 306 specimen referred to Gemuendina. This one is closer to the extant channel catfish, Ictalurus. Note the tiny teeth in the mandible, but not at the edge or rim of the mandible.

Johanson and Smith also err
when they state, “The Arthrodira is a derived taxon within the Placodermi, hence origin of teeth in placoderms occurs late in the phylogeny and teeth are convergently derived, relative to those of other jawed vertebrates.” The LRT notes that Coccosteus is a basal placoderm, one that is closer to the outgroup taxon, Gregorius than are other less predatory taxa. This exemplifies a problem with this, and many other papers in that without a proper and validated cladogram, it is nearly impossible to determine whether the absence of teeth, or any other trait, represents a vestigial loss or a vestigial genesis situation.

Johanson and Smith report, 
“Tooth sets and tooth whorls in crown-group gnathostomes are suggested to derive from the pharyngeal denticle whorls, at least in sharks, with the patterning mechanisms co-opted to the oral cavity. A comparable co-option is suggested for the Placodermi.”

Figure 1. Whale shark (Rhincodon) tooth pads, not that much different from catfish tooth pads (Fig. 2).

Figure 2. Whale shark (Rhincodon) tooth pads, not that much different from catfish tooth pads (Fig. 2).

The authors do not mention
the tooth carpets of Rhincodon (Fig. 2) and Manta. The LRT indicates that these taxa represent the origin of teeth within the jaws, not on the margins, which remain toothless, but on the palate, reusembling shark skin.

The authors likewise do not mention
the angel shark Squatina. The LRT indicates this taxon represents the origin of teeth along the margins of the jaws.

The LRT indicates
placoderms lose teeth and sometimes develop sharp, turtle-like gnathal plates, some of which retain vestigial tooth-like bumps. Their sister clade, the Siluriformes (catfish) lose the maxilla and retain tooth carpets only in the mandible (Fig. 1). This begins with the basalmost catfish, traditionally considered a basal placoderm, Entelognathus.


References
Johanson Z and Smith MM 2005. Origin and evolution of gnathostome dentitions: a question of teeth and pharyngeal denticles in placoderms. Biology Review 80:1–43.

Lunaspis: placoderm transitional to catfish

Always considered an odd sort of placoderm,
(Fig. 2) Lunaspis heroldi (Fig. 1) nests in the large reptile tree (LRT, 1597+ taxa; subset Fig. 3) basal to catfish, derived from another traditional placoderm without maxillae, Entelognathus.

Figure 1. Early Devonian Lunaspis lacks a maxilla, like sisters Entelognathus and catfish.

Figure 1. Early Devonian Lunaspis lacks a maxilla, like sisters Entelognathus and catfish. Photo from Long 1995.

Lunaspis heroldi (Broili 1929; latest Early Devonian, 405mya) is traditionally considered a petalichthyid placoderm, but here nests between Entelognathus (Zhu et al. 2013) and the armored catfish, Hoplosternum (Figs. A, B). The tail, missing from this fossil, but known from other specimens, was long and whip-like, as in extant rays. None of the specimens preserve mouth parts. As in catfish and Entelognathus, the maxilla is absent. Note the serrated leading edge of the spike-like pecotral fins,  also found in catfish.

FIgure 2. Traditional placoderm cladogram with LRT notes added

FIgure 2. Traditional placoderm cladogram with LRT notes added.

The LRT continues
to nest catfish with placoderms, taxa that were not tested together prior to this ongoing online experiment. Throughout the LRT taxa nest conventionally (turtles with turtles, birds with birds), but occasionally new topologies are recovered.

Figure 4. From Zhu et al. 2016, overlay added based on LRT topology.

Figure 3. From Zhu et al. 2016, overlay added based on LRT topology. Entelognathus is basal to the catfish clade. Bothriolepis and the antiarchs are osteostraci mimics.

Osteostraci and Galespida|
(Figs. 2, 3) are widely considered basal to placoderms, which, in turn, are widely considered basal to sharks (Fig. 3). Unfortunately Osteostraci and Galespida skulls generally do not have bone sutures and if any sutures are present, they are not homologous to those in gnathostomes (jawed vertebrates). Missing from the above taxon lists are thelodonts, like Thelodus, which is shaped like a tiny whale shark (Rhincodon) and angel shark (Squatina). These taxa preserve gill bars alongside jaw bones and document the appearance of jaws prior to the appearance of marginal teeth. The LRT (subset Fig. 4) shows that figures 2 and 3 are upside-down, nesting basal taxa as derived taxa and vice versa, and they are missing the taxa that would upend them.

Figure 1. Hoplosternum in vivo. You can see the armor/bone beneath its shiny skin.

Figure B. Hoplosternum in vivo. You can see the armor/bone beneath its shiny skin.

Figure A. The skull of Hoplosternum in lateral view with bones identified by their tetrapod colors.

Figure A. The skull of Hoplosternum in lateral view with bones identified by their tetrapod colors.

Figure 3. Subset of the LRT focusing on catfish and placoderms.

Figure 4. Subset of the LRT focusing on catfish and placoderms.


References
Broili F 1929. S. B. Bayer. Akad. Wiss., 1.
Long JA 1995. The Rise of Fishes. 500 million years of evolution. The Johns Hopkins University Press, Baltimore and London.
Zhu M, Yu X-B, Ahlberg PE, Choo B and 8 others 2013. A Silurian placoderm with osteichthyan-like marginal jaw bones. Nature. 502:188–193.
Zhu M et al. 2016. A Silurian maxillate placoderm illuminates jaw evolution. Science 354.6310 (2016): 334-336.

wiki/Lunaspis
wiki/Entelognathus