Restoring the skull of Quetzalcoatlus sp.

We know of at least three partial skulls of Quetzalcoatlus sp. (Kellner and Langston 1996), the smaller version of Q. northropi (Lawson 1975), for which no skull is known. This is how Kellner and Langston (1996) reconstructed the skull.

Figure 1. Quetzalcoatlus sp. skull according to Kellner and Langston (1996).

Since this is such a popular pterosaur, many have attempted restorations based on their access to the specimen and their own artistic license. See one here and another here.

Here’s how I do it.
Below are images traced from Kellner and Langston (1996), scaled to the same scale (they were nearly identical in length and proportion) then combined to create the restoration (Fig.1).

Figure 2. Quetzalcoatlus skulls individually and combined to create a restoration. Not preserved possible soft tissue in gray. Back of the skull based on chaoyangopterids and Zhejiangopterus. Not sure what to make of the two bony crest shapes (red and black), whether a result of taphonomy or individual variation. Considering the angle between the cranial and rostral parts of this skull, apparently Q kept its snout down, not out. 

The skull turns out to be a little taller due to a low rostral crest, with a taller antorbital fenestra. Kellner and Langson (1996) assumed that the dentary continued on to more of a point and that may be so considering the example of Azhdarcho. The details and sutures of the skull are sometimes easy to make out but otherwise the texture of the gnarled bone  makes things more difficult.

All the same size? Yes.
And what does that mean? Were all three specimens juveniles of the same age? Or adults of the same age but of a species distinct from Q. northropi? I don’t know.

Figure 3. The dentary tip of the TMM 42161 specimen, twisted and skewed in dorsal view. Not sure what to make of this. If the tip does extend further (in pink), it likely does not extend very much further. The tip currently has the shape of a yardstick with some added ridges and valleys. Color represent distinct areas, not distinct bone. This is the anterior part of paired fused bones, the dentaries. Bottom view skewed segments unskewed. This dentary tip is wider than tall, unlike the tip of Eopteranodon and Eoazzhdarcho. Green areas are hypothetical lines to join and continue apparent missing areas. 

The dentary tip
I can only wonder what is going on at the tip of this dentary. It was skewed left and right several times, but in lateral view it appears undistorted. Distinct from sharp-jawed and tooth-tipped Eopteranodon and Eoazhdarcho (which are not azhdarchids), this rostral tip Fig. 3) is wider than tall, like a yardstick. It doesn’t appear strong enough to battle prey that fights back, but supports a gentle wading after tiny crustacean lifestyle.

References
Kellner AWA and Langston W 1996. Cranial remains of Quetzalcoatlus (Pterosauria, Azhdarchidae) from late Cretaceous sediments of Big Bend National Park, Texas. – Journal of Vertebrate Paleontology 16: 222–231.
Lawson DA 1975. Pterosaur from the latest Cretaceous of West Texas: discovery of the largest flying creature. Science 187: 947-948.

wiki/Quetzalcoatlus

A tree topology change – turtles and pareiasaurs move from diadectids to millerettids

I spent last week adding taxa and running through potential problems with the large reptile tree. Several matrix boxes were rescored. The result shifted turtles + pareiasaurs from Diadectides + Procolophon  to Milleretta RC70 + Odontochelys (a near-turtle now, not a real turtle), which we discussed earlier.

An interesting shift. 
Moving pareiasaurs + Proganochelys back to Diadectes + Procolophon now adds 22 steps. Moving pareiasaurs + turtles to Eunotosaurus (following the results of Lyson et al. 2013) adds 28 steps.

The RC14 specimen of Milleretta is still in the same clade as Acleistorhinus + Eunotosaurus + Austraolthyris + Feeserpeton + Belebey + Bolosaurus. 

Maybe not as crazy as it sounds
It’s the “plain brown sparrows” like Milleretta, that lie at the bases of major clades, not the highly derived taxa, like Procolophon and Diadectes. Those become extinct. The various specimens of Milleretta have long been ignored, but they really are the keys to understanding the reptile family tree.

References
Broom R 1913. On the Structure and Affinities of Bolosaurus. Bulletin of the American Museum of Natural History 32:509-516
Broom R 1938 On the Structure of the Reptilian Tarsus: Proceedings of the Zoological Society of London, v. 133, 108, p. 535-542.
Broom R 1948. A contribution to our knowledge of the vertebrates of the Karroo beds of South Africa: Transactions of the Royal Society of Edinburgh, Endinburgh 61: 577-629.
Case EC 1907. Description of the Skull of Bolosaurus striatus Cope. Bulletin of the American Museum of Natural History 23:653-658
Cope ED 1878. Descriptions of extinct Batrachia and Reptilia from the Permian formations of Texas. Proceedings of the American Philosophical Society 17:505-530
Gow CE 1972. The osteology and relationships of the Millerettidae (Reptilia: Cotylosauria). Journal of Zoology, London 167:219-264.
Watson DMS 1954. On Bolosaurus and the origin and classification of reptiles.Bulletin of the Museum of Comparative Zoology at Harvard College,, v. 111, no. 9444-449.

wiki/Milleretta
wiki/Bolosaurus

Broomia – A sister for Milleropsis – not Milleretta

Figure 1. Broomia. A long-recognized sister to Milleropsis, an early possible biped. Check out those thighs!

Broomia perplexa (Watson 1914, Thommasen and Carroll 1981, Middle Permian) was considered a millerettid and a descendant of romerid - catorhinomorphs, considerably older than other millerettids. The large reptile tree nested Broomia away from the millerettids among the new Lepidosauromorpha, but right next to Milleropsis, among the new Archosauromorpha close to the origin of the Diapsida and Petrolacosaurus. A shame we can’t see the top of the skull! Those long legs, robust femora, strong tarsals and large feet give the impression that Broomia was a strong runner.

Millerettids perplexed Carroll in 1981 and again in 1988 in the days before phylogenetic analysis. Today, due to the large reptile tree, we know the original members are diphyletic (not related), but developed similar traits by convergence.

Broomia has a lateral temporal fenestra derived from basal synapsid ancestors like Aerosaurus and Heleosaurus. mimicking millerettids and caseids. Unlike lepidosaurs, Broomia lacked an ossified sternum or fenestrated pectoral girdles.

Thommasen and Carroll (1981) note the disappearance of the millerettids about the time that lizards first appeared in the Late Permian. Millerettids gave rise to turtles and lepidosaurs via owenettids and early lepidosauriformes. Milleropsids, like Broomia, gave rise to enaliosaurs and protorosaurs, which ultimately produced archosaurs.

References
Thommasen H and Carroll RL 1981. Broomia, the Oldest Known Millerettid Reptile. Paleontology 24(2): 379-390.
Carroll RL 1988. Vertebrate Paleontology and Evolution. W. H. Freeman and Company, New York 1-698.
Watson  DMS 1914. Broomia perplexa gen et sp. nov., a fossil reptile from South Africa. Proceedings of the Zoological Society of Londont 995-1010.

The origin of pareiasaurs (and turtles).

Figure1. Pareiasaur (Anthodon) and its phylogenetic predecessor, Stephanospondylus, a robust millerettid and an ancestor to turtles. Note the close correspondence of dorsal skull elements.

Earlier we looked at the ancestry of turtles (starting with Proganochelys and Stephanospondylus) and near-turtles (starting with Odontochelys and the RC70 specimen of Milleretta). According to the large reptile tree, the closest sister taxon to Stephanospondylus and turtles is the clade Pareiasauria, represented above by Anthodon (Fig. 1). Stephanospondylus thus represents the closest known sister to pareiasaurs followed more distantly by the RC70 specimen of Milleretta. So pareiasaurs AND turtles are really just robust millerettids, which raises millerettid status from “huh? and “obscure” to “waiting in the wings.”

Figure 2 Arganaceras, as originally reconstructed and modified.

Stephanospondylus doesn’t have much cheek flare because its closer to turtles. For that we look to the basal pareiasaur, Arganaceras. Stephanospondylus also begat Elginia and Sclerosaurus, the “horned toads” of the Middle Triassic.

Figure 3. Sclerosaurus, a sister to pareiasaurs that perhaps gives some insight into the postcrania of Stephanospondylus , Milleretta RC70 and the ancestry of turtles and Odontochelys.

Sues and Reisz (2008) considered Sclerosaurus a procolophonid, but their concept of a procolophonid was much greater than the results of the large reptile tree.

Figure 4. The skulls of (left to right) the RC70 specimen of Milleretta, Odontochelys, Stephanospondylus and Proganochelys. The two on the left belong to a distinct clade and the two on the right belong to a distinct clade along with pareiasaurs. Odontochelys is more closely related to this specimen of Milleretta than to Stephanospondylus and Proganochelys.

References
Broom R 1938 On the Structure of the Reptilian Tarsus: Proceedings of the Zoological Society of London, v. 133, 108, p. 535-542.
Broom R 1948. A contribution to our knowledge of the vertebrates of the Karroo beds of South Africa: Transactions of the Royal Society of Edinburgh, Endinburgh 61: 577-629.
Gow CE 1972. The osteology and relationships of the Millerettidae (Reptilia: Cotylosauria). Journal of Zoology, London 167:219-264.
Hartmann-Weinberg AP 1933. Evolution der Pareiasauriden: Trudy Palaeontological institute Academe Nauk, SSSR, 1933, n. 3, p. 1-66.
Lee MSY 1997. Pareiasaur phylogeny and the origin of turtles. Zoological Journal of the Linnean Society 120: 197-280.
Owen R 1876. Descriptive and Illustrated Catalogue of the Fossil Reptilia of South Africa in the Collection of the British Museum. London, British Museum (Natural History).
Sues H-D and Reisz RR 2008. Anatomy and Phylogenetic Relationships of Sclerosaurus armatus (Amniota: Parareptilia) from the Buntsandstein (Triassic) of Europe. Journal of Vertebrate Paleontology 28(4):1031-1042. doi: 10.1671/0272-4634-28.4.1031 online

wiki/Anthodon
wiki/Deltavjatia
Deltavjatia paleocritti
wiki/Pareiasaur
Sclerosaurus paleocritti
wiki/Milleretta

Dorygnathus – where are the postorbital bones? DGS to the rescue.

My first encounter with the UUPM R156 (Uppsala Museum, Sweden) Dorygnathus was in Wellnhofer (1991), the famous pterosaur encyclopedia. The image was small and produced with halftone dots. Nevertheless I produced a reconstruction from it and I used the fuzzy data in the large pterosaur tree.

The resolution question.
There are a whole raft of pterosaur workers who dismiss such efforts gleaned from photographs, both of poor quality and excellent. Some photographic data comes from publications. Other data comes from photographs I’ve taken on various trips to visit the specimens. Sometimes those photos come in handy long after the trip is over as new insights come in randomly.

Now let’s draw a parallel. There was a time, before the advent of the Hubble telescope and the Voyager and other flyby satellites, when the best images we could get of the planets came form Earth-bound telescopes beneath an ocean of atmosphere. Fuzzy is the best way to describe them. The broiling atmosphere was the problem. Even in photos from the largest telescopes there’s not a a lot of resolution. Then, after 1990, Hubble images provided a magnitude leap in resolution because they were taken far above the atmosphere.

Figure 1. The planet Jupiter as seen from above the atmosphere (Hubble) and below (Hale). Having a poor resolution photo did not impede astronomers from gathering data on Jupiter.

But did that stop astronomers from studying Jupiter?  No. You take what you’re given. And when you’re given better data you refine your hypotheses. What you don’t do is denigrate others for gathering data using the best available data, fuzzy though it may be. That is what the opposing camp of traditional pterosaur experts (Naish, Witton, Bennett, Hone, Unwin) do. Those are the experts you’ll recall, who are most responsible for disfiguring pterosaurs. They are still hoping that pterosaurs had deep chord wing membranes, fingers that faced palms forward in flight, babies that did not look like grownups, strong sexual dimorphism, eggs that were buried under rotting vegetation, a cruropatagium controlled by the lateral digits and, perhaps worst of all, they still have no idea what pterosaurs are despite being given the answer some 12 years ago (Peters 2000). They could have discovered what pterosaurs are, just by testing, looking and comparing. But they refuse to. 

Getting back to Dorygnathus R156
The skull of the R156 specimen (Fig. 2) appeared online and it offered better resolution than the Wellnhofer (1991) print. So I applied DGS to it and discovered several previously “missing” bones. None of these have been documented yet, as far as I know. Padian (2009) did not illustrate this specimen in his recent treatise on Dorygnathus, but described it nevertheless. Padian (2009) reported that Wiman (1925) wrote a detailed paper on the specimen.

Figure 2. Click to enlarge. Here is the R156 specimen of Dorygnathus. It looks like a complete skull, but can you find the postorbital. pterygoid and squamosal? Image has been corrected for perspective from the original posted photo. That’s the left femur in the teeth.

Padian (2009) wrote that Wiman (1925) noted, “the bones of the left side of the skull behind the premaxilla are missing, so that one sees the posterior part of the skull from inside the right side.” Padian also considered the squamosal missing but made few comments about the skull other than the teeth and jaw symphysis, the most easily seen elements. He did not comment on the palate or occiput elements, which are more difficult to determine.

Figure 3. Click to enlarge. Digital Graphic Segregation applied to the skull revealing the location of the displaced postorbital and palatal elements. Note much of the left maxilla is missing, revealing the right maxilla in medial view. Look closely to see the replacement tooth coming up laterally on the longest dentary tooth.

DGS Step-by-step
Digital Graphic Segregation helps one understand crushed fossils by removing areas of chaos and segregating bones by color and layer. Coloring the easy bones first ultimately reveals the difficult ones. And that’s the beauty of it. Later, making a reconstruction of the elements lifted and placed digitally, confirms the fit of the rest.

The parietal lateral elements were broken off and slightly displaced. The postorbital lay inside the jugal (Fig. 3). I would be surprised if the palatal elements have ever been identified. They are currently folded up in the parasagittal plane. The major elements of the occiput are probably washed away along with the left side of the skull.

Figure 4. Dorygnathus R156 reconstructed in three views. Elements from the insitu image were lifted intact and reassembled here in the second phase of DGS. A reconstruction confirms the identification of the elements as the puzzle pieces fit back together in patterns that resemble sister specimens. There were probably more dentary teeth, but the present view (Fig. 3), taken from slightly below. does not reveal them.

Phylogenetic analysis
The original reconstruction was refined by the new reconstruction (Fig. 4), but only two or three traits changed scores in the large pterosaur tree. The result of these rescorings nested R156 with Sericipterus, which it nested next to previously.

Figure 5. Doryganthus UUPM R156 revised with new data coming from the online image in high resolution of the skull and cervicals. Science marches on.

The nearly parallel pterygoids are atypical for pterosaurs in general, but become even more parallel in ctenochasmatids. R156 is in the lineage of ctenochasmatids according to the large pterosaur tree, something that should be obvious from its similarly protruding teeth. A while back those teeth were the first clues I had of a possible direct ancestry with ctenochasmatids. Later, by adding taxa, I realized that the tiny pre-ctenochasmatids transitioned larger forms like Angustinaripterus to Ctenochasma.

Figure 6. The Vienna specimen of Dorygnathus portrayed in Wiman 1925, which, ironically, is a study of the R156 Uppsala specimen, not the Vienna specimen shown here. The R156 specimen was not illustrated by either Wiman or Padian.

Wiman (1925) likewise did not figure the R156 skull
Like Padian (2009), Wiman (1925) did not illustrated the the R156 skull. Instead Wiman employed the Vienna specimen skull, after Arthaber. Compared to figure 3, several of the suture differ here, perhaps attributable to a more primitive knowledge of the pterosaur skull back in 1925. No known pterosaur has such a large quadratojugal, nor such an oddly shaped jugal.

References
Andres B, Clark JM and Xing X 2010. A new rhamphorhynchid pterosaur from the Upper Jurassic of Xinjiang, China, and the phylogenetic relationships of basal pterosaurs, Journal of Vertebrate Paleontology 30: (1) 163-187.
Padian K 2009. The Early Jurassic Pterosaur Dorygnathus banthenis (Theodori, 1830) and The Early Jurassic Pterosaur Campylognathoides Strand, 1928, Special Papers in Paleontology 80, Blackwell ISBN 9781405192248
Wellnhofer P 1991. The Illustrated Encyclopedia of Pterosaurs, London (Salamander Books Ltd)192 pp.
Wiman C 1925. Über Dorygnathus und andere Flugsaurier. Bulletin of the Geological Institute of Uppsala, 19 (for 1923), 23–54.

wiki/Dorygnathus
wiki/Sericipterus

Turtles and Eunotosaurus, rib comparisons – Lyson et al. 2013

Lyson et al. (2013) recently homologized the odd ribs and vertebrae of Eunotosaurus with those of turtles (Fig. 2) and nested Eunotosaurus with turtles as the sister taxa of pareiasaurs (Fig. 1).

Figure 1. From Lyson et al. (2013) in which turtles nest with Eunotosaurus and pareiasaurs. Light orange taxa are suprageneric. Green boxes highlight nodes that are different from the large reptile tree.

Lyson et al. (2013) reported, “The goal here is not an exhaustive description of Eunotosaurus but rather one focused on shell related features and novel morphologies not apparent in previous descriptions. A more comprehensive treatment will be provided in a later publication.”

To their point, the expanded dorsal ribs of Eunotosaurus and Odontochelys are remarkably similar and T-shaped in cross-section. Unfortunately, the large reptile tree, which relies on no suprageneric taxa and includes a much larger number of pertinent taxa, found Eunotosaurus nested with Acleistorhinus (in Fig. 1 separated by 3 nodes in Lyson et al) and then Milleretta (in Fig. 1 separated by an additional node in Lyson et al).

Lyson et al. (2013) based their analysis on deBraga and Rieppel (1997), which relies heavily on suprageneric taxa, which always brings problems (like nesting gliding kuehneosaurs as sisters to swimming sauropterygians in figure 1). To their credit, Lyson et al. (2013) nested pareiasaurs and turtles as Milleretta descendants. Lyson et al. (2013) listed diadectids as outgroups outside the Reptilia along with Seymouriadae. As reported before, diadectids are basal reptiles, not amphibians.

Figure 2. Eunotosaurus compared to Odontochelys. While remarkably alike in many respects, especially with regard to the dorsal ribs, the large reptile tree nested these two widely separated. The long toes, long tail and slender limbs of Eunotosaurus were inherited from Milleretta ancestors and they both share a lateral temporal fenestra. Taxa must always be considered in toto, not just on the basis of their ribs.

We looked at Lyson et al. (2010) earlier here, in which they nested Eunotosaurus with turtles. Unfortunately in the large reptile tree moving Eunotosaurus to the base of turtles  adds 144 steps.

Figure 3. Skull of Eunotosaurus compared to turtles, Milleretta and Stephanospondylus. The odd bedfellow here in Eunotosaurus, which retains the lateral temporal fenestra of its Milleretta (RC14) ancestors.

Comparing turtle skulls to turtle ancestor candidates (Fig. 3) graphically demonstrates the differences that put Eunotosaurus as the odd man out.

We don’t know what the ribs of Stephanospondylus or Milleretta (RC70) look like. Thankfully there is no such thing as modular evolution. Rather phylogenetic bracketing hints that these two likely had broad ribs too, but then… pareiasaurs do not have such broad ribs. And, Odontochelys does not have transverse processes, but Proganochelys does. So…we’ll have to wait for that data.

References
DeBraga M and Rieppel O 1997. Reptile phylogeny and the affinities of turtles. Zoological Journal of the Linnean Society 120, 281–354.
Lyson T, Bever GS, Scheyer TM, Hsiang AY and Gauthier JA 2013. Evolutionary Origin of the Turtle Shell.

Where are the internal nares in plesiosaurs?

Figure 1. Simosaurus and Anningasaura. Somewhere between these two the internal and external nares of these protoplesiosaurs became much smaller, almost useless vestiges. With such a tiny nostril in Anningsaura, apparently breathing continued through the mouth alone. The placement of the internal nares did not shift much. Is that a secondary choana (internal naris) between the pterygoids? Probably not because it’s not much larger than the real choana, so no advantage. There is no medial extension of the maxilla here as in other reptiles with a secondary palate (see below).

The recent paper on pliosaur palates by Schumacher et al. (2013) considered the placement of the internal nares in plesiosaurs (Figs. 1, 3). They noted Buchy et al. (2006) questioned the various placements and offered evidence for a functional secondary palate in plesiosaurs, shifting the internal nares to the back of the palate, posterior to the pterygoids, similar to the situation in crocodilians (Fig. 2). However in crocs, as you can see, the maxillae and palatines contact medially producing a standard sort of secondary palate.

Figure 2. The palate of Alligator. Note the posterior placement of the internal nares with paired maxillae, palatines and pterygoids forming the secondary palate.

The traditional view
holds that a pair of very small openings in the anterior half of the plesiosaur palate (Figs. 1,3), anterior to the palatines, represent the internal nares through which respiration took place.

Another traditional view
Williston (1903) accepted such a position for Dolichorhynchops, but not for Brachauchenius. Instead, despite the retention of internal nares in the traditional place, Williston placed the internal nares between the pterygoids separated by the parasphenoid (see Plesiosaurus in figure 3) as in crocodilians.

More recently, Schumacher et al. (2013) reported, “We have independently concluded that the posterior interpterygoid vacuity…should be called the internal nares.”

The Schumacher Hypothesis
Such a posterior position of the internal naris would be due to the development of a secondary palate in plesiosaurs, according to Schumacher et al. (2013). Benefits: Shifting the internal nares posteriorly, as in crocodiles, separates nasal respiration from oral functions (like biting, chewing, reorienting and swallowing). Problems: Small nares restrict air passage and ventilation. Since the posterior openings are typically slightly larger than the anterior ones Schumacher et al. (2013) suggest that the shift was made by soft tissue within the skull and the tiny anterior openings would have been covered with mechanosensory or chemosensory tissue, thereby completely blocking respiration there. According to Schumacher et al. (2013) respiration would then have taken place at the back of the palate where the slightly larger interpterygoid opening is.

Figure 3. Click to enlarge. Enaliosaur palates beginning with Claudiosaurus (upper left). The internal nares shrink in Anningasaurus, Pisotosaurus and Plesiosaurus compared to Simosaurus and Pachypleurosaurus. The palate is open posteriorly only in Plesiosaurus, but that does not appear to be a channel for respiration.

Figure 4. Sea turtle with secondary palate. Here the palatines join medially to shelve the respiratory tract and shifting the internal nares to mid palate. From Brown and Madara 2000. See Proganochelys for a turtle without a secondary palate.

Of course
All that presupposes that plesiosaurs actually breathed through those tiny nostrils and internal nares. Maybe they didn’t. Maybe, once the nares became sufficiently tiny (vestigial), plesiosaurs began to breathe through their mouth, like the sea turtle in figure 5. This makes all the more sense in hyper-long-necked elasmosaurs, as the mouth offers no respiratory restrictions while elevating the skull above the water surface (even slightly above as in fig. 5). This also makes sense in giant-jawed pliosaurs where breathing might have taken place with the jaws just barely open and air respiring between the slightly gaped giant teeth, perhaps while “spy-hopping,” or during less obvious maneuvers.

Figure 5. Sea turtle breathing at the surface. Both the nares and the mouth are open. (Photo by Joe Raedle/Getty Images)

Phylogeny Clues
If we look at a series of plesiosaurs and their ancestors (Fig. 6) we see that Claudiosaurus, with its tiny skull, had the relatively largest internal nares. The nothosaurs (Pachypleurosaurus, Lariosaurus, Simosaurus (Fig. 1) had a smaller, but still substantial internal naris. In contrast, Anningasaura and Pistosaurus had vestigial internal nares, obviously unusable for respiration. Other marine taxa (Fig. 3) also had tiny internal nares.

Something changed.
Between Simosaurus and Anningasaura both the internal and external nares shrank to vestiges (Fig. 1). I think it’s likely that Anningasaura and its plesiosaur and pliosaur descendants were breathing through their mouth.

Figure 6. Secondary palate development in a series of synapsid cynodonts leading to mammals. In  the eutheriodont and Procynosuchus the secondary palate is incomplete. In Thrinaxodon and all subsequent cynodonts, including all mammals, the secondary palate is complete when the maxillary palatal processes meet each other along with the palatine palatal processes. Note the gradual posterior shift of the internal naris from Thrinaxodon to Morganucodon. Image from Hopson 1991.

In a real secondary palate,
as in synapsids (Fig. 6), the development of a secondary palate (maxillae and palatines meet medially) can be traced through a series of fossils. The same holds true for crocodilians (Fig. 2, see Scleromochlus and Terrestrisuchus for taxa without a palate), sea turtles (see Proganochelys for a taxon without a palate) and pterosaurs (Fig. 7, see Cosesaurus for the primitive condition). Champsosaurus doesn’t have a secondary palate, just a longer snout and the internal nares stayed put.

Figure 7. Click to enlarge. Evolution of the pterosaur palate from Eudimorphodon to Pterodaustro. The secondary palate formed by maxillary palatal processes meeting at the midline shift the internal nares posteriorly. A similar expansion of the maxillae and migration of the internal nares is not documented in sauropterygians.

Sauropterygians do not document a similar gradual shift of the internal nares. Rather the nares simply shrinks reflecting a lack of usage while shifting to respiration through the mouth. There is no medial extension of the maxilla or palatine. Various valves would have been present to open and shut the esophagus (to the stomach) and epiglottis (to the lungs). Otherwise, no bony changes document the development of a secondary palate as in other reptiles (contra Schumacher et al. 2013). And there’s no real benefit to that marginally larger interpterygoid opening. The presupposition that the nares were used for breathing is at the heart of the problem. The problem goes away when that supposition goes away.

References
Buchy M-C, Frey E and Salisbury  2006. Internal cranial anatomy of Plesiosauria (Reptilia, Sauropterygia): evidence for a functional secondary palate. Lethaia 39:290-303.
Hopson JA 1991. Systematics of the nonmammalian Synapsida and implications for patterns of evolution in synapsids, in H-P Schultze & L Trueb [eds], Origins of the Higher Groups of Tetrapods: Controversy and Consensus. Comstock, pp. 635-693.  Schumacher BA, Carpenter K and Everhart MJ 2013. A new Cretaceous Pliosaurid (Reptilia, Plesiosauria) from the Carlile Shale (middle Turonian) of Russell County, Kansas. Journal of Vertebrate Paleontology 33(3):613-628.