Revisiting the Dendromaia tiny den-mate

Earlier I attempted a tracing of the 2cm skull of the Dendromaia (Maddin et al. 2020; Figs. 1–3) small den-mate using a low resolution image. That didn’t work out well due to using only one plate and misinterpreting the subtle grays in the photo. Even so, oddly enough, the error-filled scoring nested the small skull close to the same taxa that Maddin et al. nested the specimen.

As you might remember,
the much larger den-mate nested in the large reptile tree (LRT, 1628+ taxa) with Acleistorhinus and other skull-only taxa between the more complete Casea, Eocasea and Eunotosaurus taxa. The large den-mate was probably an herbivore based on phylogenetic bracketing. The tiny den-mate was a likely insectivore.

Today
with higher resolution images of the part and counterpart mated together in Photoshop layers (Fig. 1), the skull of the small den-mate is re-traced and reconstructed in much greater detail. (Still far from perfect.) The resulting plate and counter plate preserve the palate and the mandibles respectively in ventral view. Dorsal sutures are unknown.

Figure 1. The small den mate assigned to the genus Dendromaia traced using DGS methods and reconstructed in figure 2.

Figure 1. The small den-mate assigned to the genus Dendromaia traced using DGS methods and reconstructed in figure 2. Details are difficult to interpret. As before, this is a best guess based on current data.

Details are still difficult to interpret.
As before, this is a best guess based on current data. Now the small Late Carboniferous den-mate nests between two other much larger skull taxa both assigned to the genus Varanosaurus, known from Early Permian skeletons. So the small den-mate must be congeneric with Varanosaurus. The small size of the small den-mate is probably due to its young ontogenetic age.

These taxa are basal synapsids (in the lineage of humans), not protodiapsids.

Figure 3. The small den mate nests between these two specimens assigned to Varanosaurus.

Figure 2. The small den-mate nests between these two specimens assigned to Varanosaurus.

Morphologically flat skulls,
like those in Varanosaurus (Fig. 2) and the small den-mate (Fig. 3), tend to fossilize in dorsal or palatal view (Fig. 1). The shape of the mandible informs the reconstructed width of the skull. The dorsal sutures are best guesses based on phylogenetic bracketing.

Figure 3. Reconstruction of the small den mate based on DGS tracings in figure 1.

Figure 3. Reconstruction of the small den-mate based on DGS tracings in figure 1.

The number of mistakes I’ve made
and corrected over the last eight years is now deep into six figures. These corrections are just the latest set to get corrected. Few other workers are attempting to identify bones to this degree on such tiny specimens. There’s no blueprint for this. Everyone who attempts such tracings are on their own. You might try practicing on some roadkill for starters.


References
Maddin HC, Mann A and Hebert B 2020. Varanopid from the Carboniferous of Nova Scotia reveals evidence of parental care in amniotes. Nature ecology & evolution 4:50–56.

wiki/Varanosaurus
wiki/Dendromaia (not yet posted)

The affinities of ‘Parareptilia’ and ‘Varanopidae’: Ford and Benson 2020

Readers will know the knives are out for this one
by Ford and Benson 2020 since the large reptile tree (LRT, 1625+ taxa) finds the Parareptilia is polyphyletic and the Varanopidae (1940) is a junior synonym for Synapsida (1903). And yes, Ford and Benson’s cladogram (Fig. 1) suffers from (altogether now): taxon exclusion. The Ford and Benson paper, like many before it, keeps perpetuating the myth of the Parareptilia and other traditional clades.

Figure 1. Cladogram by Ford and Benson 2020, with orange overlay showing taxa in the Archosauromorpha in the LRT. Massive taxon exclusion is the problem with the Ford and Benson tree.

Figure 1. Cladogram by Ford and Benson 2020, with orange overlay showing taxa in the Archosauromorpha in the LRT. Massive taxon exclusion is the problem with the Ford and Benson tree.

From the abstract:
“Amniotes include mammals, reptiles and birds, representing 75% of extant vertebrate species on land. They originated around 318 million years ago in the early Late Carboniferous and their early fossil record is central to understanding the expansion of vertebrates in terrestrial ecosystems.

By contrast, in the LRT the last common ancestor of all amniotes (= reptiles) is Silvanerpeton from the Viséan (Early Carbonferous, 335mya, not listed in Fig. 1) with a likely genesis earlier since the Viséan includes several other  amphibian-like reptiles, also not listed. Ford and Benson need to dip much deeper into the basal Tetrapoda to figure out which taxon is the last common ancestor of the Amniota and which taxa precede it. They make the mistake of considering Tseajaia and Limnoscelis pre-amniotes.The LRT nests them both deep within Amniota / Reptilia.

“We present a phylogenetic hypothesis that challenges the widely accepted consensus about early amniote evolution, based on parsimony analysis and Bayesian inference of a new morphological dataset.”

That would be great, so long as they include pertinent taxa, which they do not.

“We find a reduced membership of the mammalian stem lineage, which excludes varanopids.”

That’s odd because when you add pertinent taxa, the LRT finds an increased membership in the diapsid/mammal stem lineage, the new Archosauromorpha.

“This implies that evolutionary turnover of the mammalian stem lineage during the Early–Middle Permian transition (273 million years ago) was more abrupt than has previously been recognized.”

No one can make valid implications from the Ford and Benson cladogram. It is largely incomplete.

“We also find that Parareptilia are nested within Diapsida.”

This is only possible due to massive taxon exclusion. Ford and Benson omit many taxa that would change the topology of their tree. The Parareptilia include a diverse and polyphyletic assembly of taxa according to the LRT. Ford and Benson are not aware that Lepidosauria are no longer members of the archosauromorph Diapsida.

“This suggests that temporal fenestration, a key structural innovation with important functional implications, evolved fewer times than generally thought, but showed highly variable morphology among early reptiles after its initial origin.”

Just the opposite. In the LRT fenestration evolved MORE times than generally thought.

“Our phylogeny also addresses controversies over the affinities of mesosaurids, the earliest known aquatic amniotes, which we recover as early diverging parareptiles.”

That can only happen with massive taxon exclusion. We’ve known for several years that mesosaurs nest as derived pachypleurosaurs close to thalattosaurs and ichthyosaurs in the LRT. Those pertinent taxa are omitted in Ford and Benson’s paper.

From the introduction:
“The current paradigm of early amniote evolution was established in the late twentieth century. It includes a deep crown group dichotomy between Synapsida (total group mammals) and Reptilia (total group reptiles, including birds), followed by an early divergence of Parareptilia from all other reptiles (Eureptilia).”

Add taxa and the first dichotomy separates the new Archosauromorpha from the new Lepidosauromorpha. This has been online since July 2011 and represents the current paradigm. Ford and Benson are digging into old myths and traditions.

“Furthermore, both molecular and morphological studies have recovered turtles, which lack fenestrae, as diapsids.”

Since molecular studies do not replicate trait studies in deep time molecular studies must be wrong (probably due to epigenetics) and do not employ fossil taxa. So forget genomics in paleontology. Genomics delivers false positives.

“Our analysis includes 66 early fossil members of the amniote crown group, and four crownward members of the amniote stem group, giving a total of 70 operational taxonomic units.” 

By contrast the LRT includes 1625+ taxa not biased by prior studies, including dozens of basal vertebrates and basal tetrapods.

“The goal of our study is to examine the deep divergences of the amniote crown group.” 

If so, then Ford and Benson need to add dozens to hundreds of more taxa to their incomplete study. A suggested list is found here.

“We excluded recumbirostrans from our analysis. Recumbirostrans have generally been assigned to non-amniote microsaurs, but were recently recovered as early crown group amniotes.”

By contrast the LRT includes seven taxa listed by Wikipedia/Recumbirostra. We learned earlier that previous workers have deleted taxa that otherwise deliver unwanted results. Not sure what is happening in the Ford and Benson paper after their omission of this clade. Those seven recumbirostran taxa nest outside the Reptilia /Amniota in the LRT.

From the Results:
“All our analyses recover parareptiles and neodiapsids as a monophyletic group within Diapsida.”

These are false positive results due to taxon exclusion as shown here.

From the Discussion:
“The sister relationship between parareptiles and neodiapsids, and their relationship to Varanopidae, implies a single origin of temporal fenestration before the common ancestor of these clades.” 

These are false positive results due to taxon exclusion as shown here. We’ve known the clade Diapsida is polyphyletic since July 2011 with a last common ancestor in Early Carboniferous amphibian-like reptiles.

Happy holidays, dear readers. 


References
Ford DP and Benson RBJ 2020. The phylogeny of early amniotes and the affinities of Parareptilia and Varanopidae. Nature ecology & evolution 4:57–65. SuppData

Modesto SP 2020. Rooting about reptile relationships. Nature Ecology & Evolution 4:10–11.

 

SVP 2018: The evolution of varanopids

Reisz 2018 reports,
“Varanopidae is a clade of small to medium sized carnivorous synapsids whose fossil
record spans the Late Pennsylvanian to late Permian, one of the longest known temporal
ranges of any Paleozoic eupelycosaur clade. It has also been recently suggested that this clade may not be part of Synapsida, but may instead nest within Diapsida.”

In the large reptile tree (LRT, 1306 taxa, subset Fig. 1) Vaughnictis is the last common ancestor of Diapsida and Synapsida. Varanodon nests within the Synapsida. A series of former varanopids nest as pre-diapsids.

Figure 6. Subset of the large reptile tree showing the nesting of Vaughnictis at the base of the Synapsida and Prodiapsida.

Figure 1. Subset of the large reptile tree showing the nesting of Vaughnictis at the base of the Synapsida and Prodiapsida. Higher synapsids arise from Ophiacodon. Diapsids arise from the Broomia clade. If Reisz isn’t getting this topology, he may have to add taxa. 

Reisz 2018 concludes,
“A revised and expanded data matrix and phylogenetic analysis that integrates Permo-Carboniferous synapsids and reptiles does recover a monophyletic Varanopidae within Synapsida, with Varanodon and its varanodontine sister taxa, Watongia, Varanops, Tambacarnifex, as apex, gracile predators of the early Permian, contemporaries of the larger, more massively built sphenacodontid synapsids.”

Unfortunately taxon exclusion
prevents Dr. Reisz from seeing the big picture (subset Fig. 1), published online in 2015 here and expanded since then.

References
Reisz RR 2018. Varanodon and the evolution of varanopid synapsids. SVP abstracts.

You heard it here first: Orovenator had diapsid AND varanopid traits—for good reason!

This is a YouTube video of a
talk given by postgraduate David Ford recorded at The 65th Symposium on Vertebrate Palaeontology and Comparative Anatomy, University of Birmingham. His incredibly detailed  observations found diapsid traits AND varanopid traits, which was cause for consternation. Click to view.

Ford used µCT data
to recover in Ororvenator what the large reptile tree (LRT, 1181 taxa) was able to recover from published drawings. Ford nested Orovenator and Synapsida within Diapsida. Although heretical, that’s not the correct solution when you add more pertinent taxa.

By contrast, in the LRT
basal synapsids split at their genesis between Synapsida and Prodiapsida following Vaughnictis, another late-surviving taxon. Ford was unaware of that split at the time. In the LRT, late-surviving early Permian Orovenator was derived from basal synapsids (varanopids) AND ancestral to basal diapsids like Petrolacosaurus in the Late Carboniferous.

We looked at Orovenator relationships earlier
here in 2014 and here in 2017. Key to testing any taxonomic relationships is appropriate taxon inclusion. Let’s hope Ford has expanded his taxon inclusion set appropriately when the paper comes out. He’s got a good handle on the details, but the big picture evidently was not in his ken due to the exclusion of pertinent taxa.

Figure 2. The Prodiapsida now include the holotypes of Ascendonanus and Anningia.

Figure 2. The Prodiapsida now include the holotypes of Ascendonanus and Anningia.

Oscar Reig: a paleoprophet separates archosaurs from lepidosaurs in 1967

But… for the wrong reasons.

Reig 1967 prophetically wrote:
“Archosaurs and lepidosaurs apparently have different origins; the former come from the pelycosaurs, and the latter come from the captorhinomorph cotylosaurs through the Millerettiformes.”

Considered heretical at the time,
Reig’s pronouncement echoes in the large reptile tree (LRT, 1151 taxa).

Here’s the full abstract:
“The characteristics of the first archosaurs, the proterosuchian thecodonts, show that neither of the supposed common ancestors of archosaurs and lepidosaurs could actually be an ancestor of archosaurs. Instead, the evidence seems to indicate that the archosaurian ancestors are probably in the ophiacodont-varanopsid group of the pelycosaurian synapsids. In particular, the Varanopsidae are strongly indicative of proterosuchian relationships, as they have evolved some characters which are elsewhere found only in archosaurs. Archosaurs and lepidosaurs apparently have different origins; the former come from the pelycosaurs, and the latter come from the captorhinomorph cotylosaurs through the Millerettiformes.”

The only thing he got wrong
(as everyone else got wrong until seven years ago) was not splitting the Varanopsidae into the Synapsida and the Prodiapsida, as demonstrated in the LRT. He also thought proterosuchids arose directly from varanopsids like Varanodon (Fig. 1), which converge with proterosuchids in size and skull shapes. There’s even an antorbital fenestra, or elongated naris and a drooping premaxilla in Varanodon. No wonder Reig got excited.

Figure 1. Varanodon the synapsid compared to its analog, Proterosuchus, the archosauriform.

Figure 1. Varanodon the synapsid compared to its analog, Proterosuchus, the archosauriform.

Archosauriforms do arise from
former
varanopsids, like Heleosaurus and Mycterosaurus, but not directly. They have to pass through the diapsid grade, then the basal terrestrial younginiform grade before evolving into proterosuchids.

Lepidosaurs do arise from
captorhinomorphs and millerettids in the LRT, but again, not directly. First they have to pass through the nycteroleterid, owenettid, and basal lepidosauriform grades before evolving into lepidosaurs.

The LRT recovered
two clades of diapsids one closer to lepidosaurs and another closer to archosaurs.

References
Reig OA 1967. Archosaurian reptiles: a new hypothesis on their origins.
Science 157(3788):565-8.

Varanopids: the need for a larger inclusion set

Varanopids (Fig. 1) are those vaguely monitor-like basal synapsids, that are the plain brown sparrows of the clade. Yet from them arise the spectacular pelycosaurs and the less spectacular, but no less important, basal diapsids, according to the large reptile tree.

Figure 1. Varanodon, Varanops and Varanosaurus, three varanopids to scale along with the non-varanopids, Archaeothyris, Apsisaurus, Ophiacodon, Secodontosaurus and Haptodus.

Figure 1. On the left, varanopids according to the large reptile tree: Elliotsmithia, Aerosaurus, Varanops and Varanodon. On the right, non-varanopids: Archaeothyris, Apsisaurus, Varanosaurus, Ophiacodon, Secodontosaurus and Haptodus. Secodontosaurus appears to be close to Varanosaurus. As much as Secodontosaurus looks like Varanosaurus, they are not related. But evidently the genetic material was present as this represents a reversal to earlier morphologies following Haptodus as a direct ancestor.

The problem among paleontologists has been and continues to be: they never add basal diapsids to their inclusion sets. So they don’t know this varanopid-diapsid relationship exists. In the large reptile tree, some varanopids nest within the clade synapsida. Others do not. They are basal to diapsids, so they are called proto-diapsids.

Here’s the breakdown among tested taxa. On the synapsid branch:

  1. Elliotsmithia, Aerosaurus, Varanodon and Varanops form a basal synapsid clade.
  2. Then Archaeothyris.
  3. Then Apsisaurus, Varanosaurus and Ophiacodon, followed by the sailbacks and therapsids.

On the other protodiapsid branch we find:

  1. Mycterosaurus, Heleosaurus, Aracheovenator and Mesenosaurus
  2. Milleropsis, Millerosaurus and Broomia.

Wiki divides Varanopids traditionally into:

  1. Varanopidae: Apsisaurus, Archaeovenator, Varanosaurus and others.
  2. Subfamily-Mycterosaurinae: Elliotsmithia, Heleosaurus, Mesenosaurus and Mycterosaurus
  3. Subfamily-Varanopinae: Aerosaurus, Varanodon, Varanops and others.

Tomorrow and the next day we’ll look at proto-synapsids and proto-diapsids to scale.

New Elliotsmithia semi-skull

This doesn’t add a lot to what we knew of Elliotsmithia (Broom 1937), but it does provide a canine tooth, and an upturned mandible tip. The question is: is this also Elliotsmithia? Or something else?

Figure 1. The new specimen referred to Elliotsmithia. It's the same size and shares many traits. In the original Elliotsmithia (Fig. 2) the posterior teeth appear larger. The part and counterpart are overlaid here. I'm guessing that the anterior lower dentary is buried here. Otherwise there would be no room for the next erupting tooth to develop.

Figure 1. The new specimen referred to Elliotsmithia. It’s the same size and shares many traits. In the original Elliotsmithia (Fig. 2) the posterior teeth appear larger. The part and counterpart are overlaid here. I’m guessing that the anterior lower dentary is buried here. Otherwise there would be no room for the next erupting tooth to develop.

The type Elliotsmithia had giant serrated teeth and a forward leaning squamosal. Even so, this is a very close match in most respects.

Figure 2. The type of Elliotsmithia. Those are really big teeth beneath the orbit. The pineal is more centrally placed. And the jugal appears to lean in more.

Figure 2. The type of Elliotsmithia. Those are really big teeth beneath the orbit. The pineal is more centrally placed. And the jugal appears to lean in more.

Elliotsmithia nests at the base of the Synapsida in the large reptile tree. As such it represents the most primitive appearance of the lateral temporal fenestra, but not the earliest. Living among therapsids in the Middle Permian, it was a relic taxon, like the living Sphenodon.

References
Broom R 1937. A further contribution to our knowledge of the fossil reptiles of the Karroo. Proceedings of the Zoological Society, Series B 1937:299-318.
Modesto S, Sidor CA, Rubidge BS and Welman J. 2001. A second varanopseid skull from the Upper Permian of South Africa: implications for Late Permian ‘pelycosaur’ evolution. Lethaia 34: 249-259.