Starting a series on mammals

Having basically run out of reptiles to do,
I thought I’d run through some mammals to see what the large reptile tree would come up with. So far we’ve looked at a few monotremes. And the pangolin. As other taxa are being added the convergence factor seems to be arising (or else [[shudder]] placentals had dual origins). So, I’m going to keep adding taxa and see where the data lead. As before, with theropods, the new taxa will arrive largely one at a time.

Thank you for your curiosity…

The big and small Estemmenosuchus (Dinocephalia, Therapsida, Synapsida, Archosauromorpha)

Figure 1. There are two known Estemmenosuchus species, the smaller famous one, E. mirabilis, and the larger less famous one, E. uralensis, here shown to scale. The ectopterygoid (Tr here) is oddly placed in E. uralensis, posterior to the pterygoid, but of the standard shape, if I’m reading this drawing correctly. the palatine/pterygoid pads are part of the herbivorous chewing apparatus.

Earlier we looked at
the baroque skull of Estemmenosuchus (late Permian, Tchudinov 1960, 1968) here, here and here. Today they are shown together to scale, with one about twice the size of the other. E. mirabilis is the more famous one because it is the more bizarre one, yet it is smaller than E. uralenesis. Both the postfrontals and the jugals expand distally to proceed these skull ‘horns’. A smaller one is produced by the premaxillary ascending process. And other smaller bumps are produced by the postorbital and frontal.

Looking at the palate drawings
we see a row of palatine teeth, a pterygoid with a row of teeth on the transverse process and more on the medial process, and some oddly placed ectopterygoids posterior to the pterygoids.

Figure3. The skull of Estemmenosuchus mirabilis, the hippopotamus of the Late Permian. Note the large procumbent teeth.

The maxilla has
two parallel rows of post canine teeth in E. mirabilis, essentially one row of marginal maxillary teeth in E. uralensis. But the maxillary/palatine row in E. uralensis is a new twist on the same construction. We’ve seen multiple rows of maxillary teeth in the taxa that lead up to rhynchosaur lepidosaurs.

References
Tchudinov PK 1960. Diagnosen der Therapsida des oberen Perm von Ezhovo: Paleontologischeskii Zhural, 1960, n. 4, p. 81-94.
Tchudinov PK 1968. Structure of the integuments of theriomorphs. Doklady Acad. Nauk SSSR. 179:207-210.

Just ran the numbers: Eocasea is not a sister to Casea

Reisz and Fröbisch 2014
considered Eocasea martini (Late Pennsylvanian, Fig. 1) the basalmost caseid, despite its long slender appearance and small size.

Figure 1. Eocasea in situ with anterior skull imagined based on phylogenetic bracketing. This long, low taxa does not nest with large, big-bellied caseasaurs but with more similar sister, including Delorhynchus. These are not wide dorsal ribs, but belong to a specimen with a standard, narrow torso.

And at the time (2 years ago), I wrote in ReptileEvolution.com,
“[Eocasea] had a long narrow torso and short legs. Note the resemblance to millerettids like Australothyris and Oedaleops.” 

Reisz and Fröbisch did not include those two
in their phylogenetic analysis because they were so sure they had a caseasaur. They also did not include Eunotosaurus, Acleistorhinus, Microleter, Delorhynchus, or Feeserpeton for the same reason.

But they should have done so, because
that’s where Eocasea nests in the large reptile tree (Fig. 2) close to, but not with caseasaurs.

Figure 2. Eocasea nests between Feeserpeton + Australothyris and Delorhynchus in this subset of the large reptile tree, not close to Casea.

The problems with the Reisz and Fröbish data set
is that it was too small. The authors presumed Eocasea would be a caseasaur and so included only the following taxa: Reptilia, Diadectes, Limnoscelis, Mycterosaurus, Varanops, Oromycter, Casea, Cotylorhynchus, Angelosaurus, Ennatosaurus and Eocasea. As you can imagine, providing scores to a clade named, “Reptilia” is inappropriate and, frankly, dangerous, whether cherry-picking traits or scoring all zeroes. Diadectes, Limnoscelis, Mycterosaurus, Varanops are not related to each other or to caseasaurs. Reisz and Fröbisch were making guesses without having a large gamut study from which to draw subsets.

Reisz and Fröbisch report,
“Eocasea changes significantly our understanding of the evolutionary history of both caseids and caseasaurs.” Not with the current nesting. “The discovery of Eocasea extends the fossil record of Caseasauria and Caseidae significantly, well into the Pennsylvanian, in line with the fossil record of other early synapsid clades, indicating that the initial stages of synapsid diversification were well under way by this time.” Eocasea is not a caseid and caseids are not synapsids, so it doesn’t extend anything related to Caseidae.

“More significantly, Eocasea also allows us to re-evaluate the origin and evolution of herbivory within this clade, and terrestrial vertebrates in general.” It’s not an herby and it’s not a clade member. “Thus, we can identify Paleozoic herbivores because their rib cages are typically significantly wider and more capacious than those of their closest insectivorous or carnivorous relatives. Not in Eocasea. “Nevertheless, it is likely that the ability to process this kind of plant matter precedes the skeletal correlates that can be found in the fossil record.” This is an unsupported supposition in light of the new nesting. “It is therefore possible that we are underestimating the extent of herbivory that existed in the Paleozoic, but this does not invalidate our results because the clades of herbivores that we examine here are widely separated by successive clades of non herbivorous vertebrates.” There is no ‘therefore’ when the setup if invalid. 

To their credit, 
Reisz and Fröbisch did not nest Edaphosaurus or Protorothyris as outgroup taxa to the Caseasauria (Eothyris at its base). Perhaps that is so because they did not include these taxa! And I wonder why? But they did nest the unrelated Varanops and Mycterosaurus as caseasaur outgroups. Both nest about twenty nodes away on the other major branch of the Reptilia, the new Archosauromorpha.The Reisz and Fröbisch tree is bogus because their outgroup taxon list was based not on testing, but on tradition.

If you’re looking for
osteological evidence for herbivory in the ancestry of the Caseasauria, you won’t find big bellies and flat teeth, but you will find several herbivores arising from Milleretta (late Permian late survivor of a Carboniferous radiation) in the clade Lepidosauromorpha. These include diadectids and their allies bolosaurids (post-crania unknown) and procolophonids, pareiasaurs and their allies turtles, along with caseasaurs.

Working on their ‘wish list’
Reisz and Fröbisch continue with their hypothesis: “Whereas other caseids also show dental specializations, with leaf-like large serrations being present in the marginal dentition, Eocasea, Oromycter, and the undescribed Bromacker Quarry caseid lack these serrations. Interestingly, both Oromycter, and the Bromacker caseid show skeletal evidence for herbivory, raising the possibility that oral processing in the form of puncturing vegetation may have evolved within Caseidae after the acquisition of herbivory.”  Only the latter two are indeed caseasaurids. Eocasea definitely is not one. You can’t derive homolog conclusions from unrelated taxa.

Reisz and Fröbisch continue
“Late Pennsylvanian and Early Permian diadectids also show convincing evidence of dental and skeletal adaptations for herbivory. These enigmatic [not any more] Paleozoic forms are part of Diadectomorpha, a sister group to crown Amniota  [not any more]. A preliminary phylogeny of diadectids indicates that Ambedus, a small diadectid from the Early Permian, tentatively identified as omnivorous because of its labiolingually expanded cheek teeth (but no evidence of dental wear) is the sister taxon to all other diadectids. Ambedus may not be a diadectid as noted here. However, the oldest known diadectid from the late Pennsylvanian of Oklahoma is already clearly an herbivore and older than the edaphosaur Edaphosaurus novomexicanus. As is the case with the caseid and edaphosaur synapsids, the sister taxon of Diadectidae, the Early Permian Tseajaia from New Mexico, was faunivorous.” That oldest known diadectid is not identified.

Reisz and Fröbisch add to their unsupported hypothesis with,
“Although the holotype of Eocasea certainly represents a juvenile individual [actually, and you can check this, it is the same size as sister taxa, but smaller than basal and other caseasaurs], it is diminutive, with an estimated snout-vent length of 125 mm. In contrast, the smallest known herbivorous caseid with a comparable ontogenetic age, based on level of ossification of the vertebrae and pedal elements, is a basal, undescribed form from Germany and has an estimated snout-vent length of 400 mm.” Not sure which specimen this is…

If Reisz and Fröbisch had just
increased the size of their taxon list, they would/could have correctly nested Eocasea, and avoided making the many subsequent mistakes based on that bad nesting, including the unfortunate and inappropriate naming of the taxon and the bogus headline that got tacked to the article and all the PR that attended it.

We don’t have a name yet
for the enanticaseasaurs or paracaseasaurs (Fig. 2), but we need one!

References
Reisz R and Fröbisch J 2014. The oldest caseid synapsid from the Late Pennsylvanian of Kansas, and the evolution of herbivory in terrestrial vertebrates. PLoS ONE 9(4): e94518. doi:10.1371/journal.pone.0094518

 

Hypsibema missouriensis – a Late Cretaceous Appalachia duckbill dinosaur

Figure 1. Model of Hypsibema missouriensis, a hadrosaurid dinosaur

Hypsibema missouriensis
(Cope 1869; Gilbert and Stewart 1945; Gilbert 1945; Baird and Horner 1979; Darrough et al. 2005; Parris 2006; Campanian, 84-71 mya, Late Cretaceous) is a fairly large hadrosaurid dinosaur discovered in 1942, at what later became known as the Chronister Dinosaur Site near Glen Allen, Missouri. At present this literal pinprick in the map of Missouri is the only site that preserves dinosaur bones.

Figure 2. Where the Hypsibema maxilla chunk (Figure 3) came from modeled on the skull of Saurolophus.

Small pieces of broken bone and associated caudals and toes
were first discovered when digging a cistern. They had been found about 8 feet (2.4 m) deep imbedded in a black plastic clay. The area is in paleokarst located along downdropped fault grabens over Ordovician carbonates.

Gilmore and Stewart 1945 described a series of Chronister caudal centra (now at the Smithsonian) as sauropod-like, reporting, “The more elongate centra of the Chronister specimen, with the possible exception of Hypsibema crassicauda Cope, and the presence of chevron facets only on the posterior end appear sufficient to show that these vertebral centra do not pertain to a member of the Hadrosauridae.”

First named Neosaurus missouriensis,
the caudals were renamed Parrosaurus missouriensis by Gilmore and Stewart 1945 because “Neosaurus” was preoccupied. The specimen was allied to Hypsibema by Baird and Horner 1979.

Figure 3. Back portion of a Hypsibema maxilla showing tooth root grooves and cheek indention close to jugal.

Back in the 1980s
I enjoyed going to the Chronister site with other members of the local fossil club, the Eastern Missouri Society for Paleontoogy. I was lucky enough to find both a maxilla fragment (Fig. 3) and a dromaeosaurid tooth. I remember the horse flies were pesky and  one morning, before the other members got there, I was met by a man with a shot gun who relaxed when I identified myself. A friend found a series of hadrosaur toe bones, each about as big as a man’s hand (sans fingers). The bone was so well preserved you could blow air through the porous surfaces.

References
Baird D and Horner JR 1979. Cretaceous dinosaurs of North Carolina. Brimleyana 2: 1-28.
Cope  ED 1869. Remarks on Eschrichtius polyporus, Hypsibema crassicauda, Hadrosaurus tripos, and Polydectes biturgidus“. Proceedings of the Academy of Natural Sciences of Philadelphia 21:191-192.
Darrough G; Fix M; Parris D and Granstaff B 2005.  Journal of Vertebrate Paleontology 25 (3): 49A–50A.
Gilmore CW and Stewart DR 1945. A New Sauropod Dinosaur from the Upper Cretaceous of Missouri. Journal of Paleontology (Society for Sedimentary Geology 19(1): 23–29.
Gilmore CW 1945. Parrosaurus, N. Name, Replacing Neosaurus Gilmore, 1945. Journal of Paleontology (Society for Sedimentary Geology 19 (5): 540.
Parris D. 2006. New Information on the Cretaceous of Missouri. online

wiki/Hypsibema_missouriensis
bolinger county museum of natural history
More info and links

Hipposaurus: close to the ancestry of man, but off a wee bit

Figure 1. Therapsida includes the pangolin, Manis, which nests here with Notharctus. one of only a few mammals tested so far.

At the very base of the Therapsida
(Fig. 1) we have a split between the plant-eating Anomodontia (dicynodonts, dromasaurs and kin) and the meat-eating Kynodontia (new name for a new clade that encompasses all other therapsids, including cynodonts and mammals). At the base of the Kynodontia is the rarely discussed, but obviously important taxon, Hipposaurus boonstrai (Fig. 2, Haughton 1929, 21 cm skull. SAM 8950). Biarmosuchus is a sister.

Figure 2. Published material on Hipposaurus permits one to create a reconstruction like this. Not far removed from its ophiacodont / haptodine / pelycosaur precursors, Hipposaurus had longer, more gracile limbs and a distinct sabertooth canine, like Haptodus or Cutleria on steroids!

Long-legged, saber-toothed Hipposaurus
was originally thought to be a gorgonopsian, but in a note from Dr. Jim Hopson  (U Chicago) who xeroxed Boonstra 1965 for me, Hipposaurus (“horse lizard”) has been considered a biarmosuchian within the Ictidorhinidae since the 1980s.

Figure 3. The skull of Hpposaurus was larger than that of its sisters and predecessors among the basal Therapsida, including Stenocybus and Cutleria. The fangs were longer too.

There are some odd details
in the manus and pes of this mid-sized carnivore that indicate this is a derived late survivor of an earlier radiation.

1. Hipposaurus has a large pisiform (post axial carpal, Fig. 1)
2. The first centrale is quadrant shaped
3. The second centrale is shaped like a squat chevron
4. The radiale is twice as long as wide
5. The fourth and fifth carpals are fused
6. A small circular sternum present (none in sister taxa)
7. The posterior calcaneum has a hook like tuber
8. Two wedge-shaped centralia extend the width of the tarsus
9. The first distal tarsal is the size of a metatarsal and shifts the proximal metatarsal distally, almost to the mid length of metatarsal 2.
10. Two mid phalanges are fused on pedal digit 4

Speaking of oddities at clade bases…
as we’ve seen before, clade bases are, by definition, when novelties arise. In the case of Hipposaurus, these novel carpal and tarsal oddities went nowhere. A sister taxon without such novelties, Biarmosuchus, produced all the descendants we all know and love. Hipposaurus became a mere footnote and a short Wikipedia page.

References
Boonstra LD 1952. Die Gorgonospier-geslag Hipposaurus en die familie Ictidorhinidae: Tydskr. Wet. Kuns., v. 12, p. 142-149.
Boonstra LD 1965. The girdles and limbs of the Gorgonopsia of the Taphinocephalus Zone. Annals of the South African Museum 48:237-249.
Haughton SH 1929. On some new therapsid genera: Annals of the South African Museum, v. 28, n. 1, p. 55-78.

Vaughnictis (Brocklehurst et al. 2016): a new last common ancestor of birds and bats

Recently Brocklehurst et al. 2016
renamed ‘Mycterosaurus’ smithae (Lewis and Vaughn 1965; early Permian, MCZ 2985). The new name is Vaughnictis smithae. The specimen was originally considered a varanopid close to the holotype of Mycterosaurus (Fig. 1). Now the Brocklehurst team nest the MCZ 2985 specimen between Eothyris and Oedaleops (Fig. 1) at the base of the Caseasauria, which they consider a clade at the base of the Synapsida.

Unfortunately, 
the large reptile tree nests Vaughnictis at the bases of two major clades: between Protorothyris and the Synapsida (represented here (Fig. 1) by Elliotsmithia) + the Prodiapsida (represented here by Mycterosaurus). The Caseasauria, as noted five years ago here, does not nest with the Synapsida when the taxon list is expanded, and the Prodiapsida (former varanopids) split from the Synapsida at their base when the taxon list is expanded.

Figure 1. Vaughnictis nests between Protorothyris and the Synapsida (Elliotsmithia) + Prodiapsida (Mycterosaurus) in the large reptile tree – not the Caseasauria (Eothyris + Oedaleops). Despite the many similarities, the narrow skull with parallel sides, upturned mandible tip and longer rostrum are a few traits that split Vaughnictis from caseasaurs and lump it with prosynapsids.

Lewis and Vaughn got it right.
Vaughnictis is a sister to Mycterosaurus in the LRT. It is not a caseasaur.

Brocklehurst et al. got it right
in that Vaughnictis is distinct enough from Mycterosaurus to warrant its own genus.

Despite their phylogenetic distance
only 22 additional steps are needed when Vaughnictis is force shifted over to Eothyris. This is largely due to convergence. Both clades developed lateral temporal fenestrae in similar patterns, had large eyes and a short rostrum at this stage.

Figure 2. Vaughnictis skull in situ with color tracings. See figure 3 for reconstruction. The jaws shifted posteriorly during taphonomy The parietal and its opening are difficult to read.

The phylogenetic importance of Vaughnictis
was overlooked by Brocklehurst et al. It is the most primitive known specimen in the lineage of synapsids and diapsids to have a lateral temporal fenestra. That’s why it is the last common ancestor of bats and birds, an honor formerly earned by Protorothyris, but now superseded by a taxon with lateral temporal fenestrae.

Figure 3. Color tracings of bones moved to their in vivo positions and traced. Note the anterior shifting of the jaws to their in vivo positions based on posterior dentary and jugal positions common to most if not all candidate sister taxa. 

I wish that Brocklehurst et al. had 

1. created a multi-view reconstruction
2. showed candidate sisters side by side compared to Vaughnictis
3. not excluded pertinent taxa (diapsids for the former varanopids and millerettids for the caseasaurs)
4. used colors to identify bones, rather than lines, which helps when bones overlap. They did color the teeth (Fig. 4), but not all the teeth.

Figure 4. Teeth scanned by Brocklehurst fit to dorsal view of skull. Premaxillary and maxillary teeth were not published. Note the scale bar for the teeth appears to be off by a factor of 2. Also note the premaxillary teeth appear to be jammed back from their in vivo position. 

The interesting thing about their cladogram
is that the Brocklehurst team nested Captorhinus, Limnoscelis and Tseajaia as outgroup taxa — which is correct for Caseasauria, but not for Synapsida. Protorothyris is also listed as the proximal outgroup to the Caseasauria (incorrect) + Synapsida (correct). It is clear they rely on tradition, rather than testing for their inclusion set.

In Vaughnictis, as opposed to Eothyris, note the 

1. relatively narrow skull
2. the rising mandible tip
3. the lack of maxilla/orbit contact
4. the shorter temporal length
5. the lower rostrum

These traits ally Vaughnictis with Elliotsmithia to the exclusion of basal caseasaurs.

Brocklehurst et al. note:

1. Vaughnictis lacks these mycterosaurine and varanopid traits (but it is not a member of either of these clades in the LRT) :
   a. slender femur
   b. linguo-labially compressed and strongly recurved teeth – I disagree, the teeth are indeed recurved
   c. lateral boss on the postorbital – I don’t see this on candidate taxa
2. Vaughnictis has these caseasaur traits:
   a. coronoid teeth – plesiomorphic for synapsids, but they have been lost in derived caseids, ophiacodontids, varanopids and sphenacodontians. Most workers do not include these in their tracings of pertinent taxa, so are rarely noted.
   b. large supratemporal – actually they are long, as in synapsids, not large (and wide) as in caseasaurs
   c. large pineal foramen – unable to confirm, but Elliotsmithia and Mycterosaurus also have a large pineal foramen.

Teeth
Broklehurst et al. published synchrotron scans of the palatal, dentary and coronoid teeth (perhaps the scale bar should be 1 cm, not 5mm, Fig. 4), but did not publish scans of the maxillary teeth. All of the palatal teeth form shagreen fields, not single rows. That’s different than all candidate sister taxa, whether caseasaurid or protorothyrid. What they label as “vomerine teeth” may be premaxillary teeth based on the posterior displacement of the jaws. The dentary teeth are recurved and robust. Coronoid and parasphenoid teeth are present.

Figure 5. Subset of the large reptile tree showing the nesting of Vaughnictis at the base of the Synapsida and Prodiapsida. Also note that the Synapsida is NOT the first clade to branch off from the base of the Amniota. Far from it.

This Brocklehurst team was led by
the venerable Robert Reisz, who has made dozens of great discoveries, but has resisted testing candidates suggested by the large reptile tree. And that sort of paleoxenophobia is unfortunate. Outsiders can make valuable contributions.

Finally, kudos and credit to the Brocklehurst team,
for finding the one best specimen closest to the advent of the Synapsida + Prodiapsida. It should be in every textbook from here on out.

References
Brocklehurst N, Reisz RR, Fernandez V and Fröbisch 2016. A Re-Description of ‘Mycterosaurus’ smithae, an Early Permian Eothyridid, and Its Impact on the Phylogeny of Pelycosaurian-Grade Synapsids. PLoS ONE 11(6):e0156810. doi:10.1371/journal.pone.0156810
Lewis GE, Vaughn PP 1965. Early Permian vertebrates from the Cutler Formation of the Placerville Area, Colorado. Geological Survey Professional Paper 500C: 1–50.
Reisz RR, Dilkes DW and Berman DS 1998. Anatomy and relationships of Elliotsmithia longiceps Broom, a small synapsid (Eupelycosauria: Varanopseidae) from the late Permian of South Africa. Journal of Vertebrate Paleontology 18(3):602-611.

Pangolins nest as basal precocial placentals in the LRT

Updated May 3, 2022.
Click here to see the latest nesting of pangolins as basal precocial placentals when the LRT has grown to 2083 taxa.

Figure 2. Two subsets of the LRT focusing on bats and pangolins.

Note added July 31, 2016:
The addition of more taxa preserves the close relationship of pangolins to primates, 

Note added June 20, 2018
The reevaluation of Zhangheotherium as a basal pangolin nests this clade between dermopterans and bats. Click here for more details.

At present, the large reptile tree (LRT) includes very few mammals
so keep that in mind. The LRT (now at 704 taxa) is also not fine tuned to mammal traits, like molar shapes, so keep that in mind.

So here it is:
the large reptile tree nests the pangolin, Manis, with the basal lemur, Notharctus. And Notharctus is derived from basal carnivores like Vulpavus.

Pangolins have been difficult to nest.
Recent DNA tests (Murphy et al. 2001, Beck et al. 2006) nested pangolins with carnivores, but could be no more specific than that because fossil taxa cannot be tested for DNA.

Here’s the early morphological evidence
linking Manis to Notharctus using traits that are NOT in the LRT.

1. Flexible vertebral column – pangolins use to roll up, lemurs use to wind up then jump from tree to tree and land without a jolt
2. Circumorbital ring in some species of pangolin
3. Long, clawed fingers (toes), short opposable thumb (big toe)
4. Procumbent dentary teeth at tip (some species)
5. Arboreal habitat
6. Prehensile tail
7. One usually, but up to three infants born at a time.
8. Infants ride mother’s back and tail

Figure 1. Notharctus, an Eocene adapid (lemur) and likely sister to Manis.

Figure 2. Manis, the Chinese Tree Pangolin along with other views of other pangolins

Manis/Notharctus synapomorphies from the LRT:

1. Dorsal nasal shape: widest at mid length (here posterior to mid length, but identical in Manis and Notharctus).
2. Pmx/Mx notch: > 45º
3. Posterolateral Pmx not narrower than nares
4. Mx ventrally convex
5. Fr/Pa suture straight and > Fr/Na suture width (with Homo, too)
6. Posterior parietal angle in dorsal view > 40º to transverse plane
7. Suborbital fenestra (with Homo, too)
8. Ectopterygoid, cheek process larger (with Homo, too)
9. Ectopterygoid continues aligned along pterygoid lateral edge
10. Premaxillary teeth tiny to absent
11. Cervical centra taller than long (with Homo, too)
12. Cervicals cerntra decrease toward skull
13. Femuir < half glenoid – acetabulum length
14. Pedal 3.1 > p2.1
15. Longest pedal digits: 3 and 4
16. Metatarsals 2 and 3 align with mt1
17. Metatarsals 3 and 4 align with mt5

There are several traits
in the LRT that pangolins share with people to the exclusion of lemurs, all by convergence, so not worth going into.

Some atavisms (genetic reversals) in Manis
that most other mammals don’t have include the following:

1. Chevrons
2. Scales
3. Low to absent coronoid process
4. Elongate caudal transverse processes

The important thing here
is that given the opportunity to nest with the basal carnivores, Vulpavus, Nandinia and Chriacus, Manis nested instead with Notharctus.

Keratin scales
What opossums and rats have on their tails, pangolins have all over their bodies.

The order of the loss of facial bones
provides clues to the chronology of evolutionary events in pangolins. The loss of the lateral temporal bar (posterior jugal + squamosal) occurred in all pangolins, but the loss of the jugal is apparent in ground forms, so this was a trees down order, with burrowing following tree climbing. The clavicle is also lost in pangolins.

Diet: ants and termites.
So this is what happens when a lemur changes diet and becomes solitary, and depends on sense of smell, rather than sight. Elongate tongue is convergent with that of chameleons, woodpeckers, anteaters and nectar bats. Some pangolins burrow. Loss of the lower temporal bar and loss of most of the jugal in some species goes along with loss of the coronoid process in this anteater. Manis doesn’t need chomping muscles. Nor does it need speed and leaping ability. Given an ant diet and solitary social life, perhaps that makes it easier to visualize how Manis could be derived from a more active, social lemur-like ancestor.

So…here’s the evolutionary scenario:

1. Vicious and crafty arboreal carnivore: Vulpavus
2. Frisky and social arboreal omnivore: Notharctus
3. Slow and antisocial arboreal (grading to burrowing) anteater: Manis

New data:

Figure 1. Basal placentals at two scales, all arising from a Middle Jurassic sister to Monodelphis, based on the Earliest Cretaceous appearance of Zhangheotherium, in the lineage of pangolins.

Figure 1. Subset of the LRT focusing on basal placentals, including bats.

References
Murphy WJ., et al. 2001-12-14. Resolution of the Early Placental Mammal Radiation Using Bayesian Phylogenetics. Science 294 (5550): 2348–2351. doi:10.1126/science.1067179. PMID 11743200.
Beck R, Bininda-Emonds ORP, Cardillo,M; Liu, F-G and  Purvis A 2006. A higher-level MRP supertree of placental mammals. BMC Evolutionary Biology 6 (1): 93. doi:10.1186/1471-2148-6-93. PMC 1654192. PMID 17101039.

wiki/Pangolin

What it takes to be a mammal, according to the LRT

Earlier we nested
the monotremes, Akidolestes and Ornithorhynchus,  as basalmost mammals (contra traditional nestings) in the large reptile tree (LRT). It is notable that both of these taxa display atavistic (reversed) traits, including splayed hind limbs and the loss of teeth. Akidolestes also appears to converge with therians in molar occlusion patterns, while Ornithorhynchus converges with expansion of the braincase and fusion of the cranial elements.

It is possible
that known monotremes developed advanced molar occlusal patterns by convergence with derived mammals. This may be one reason for the difference in the nesting of monotremes in other studies and the LRT.

Basic and traditional traits
associated with being a mammal include:

1. Mammary glands on the female
2. Single tooth replacement
3. Dentary/squamosal jaw joint

Rowe (1988) defined Mammalia phylogenetically 
as the crown group of mammals, the clade consisting of the most recent common ancestor of living monotremes and therians and all descendants of that ancestor. According to Wikipedia, that excludes all pre-Middle Jurassic forms.

Kemp (2005) defined mammals by their key traits:
Synapsids that possess a dentary–squamosal jaw articulation and occlusion between upper and lower molars with a transverse component to the movement. This becomes the last common ancestor of Sinoconodon and living mammals. The earliest known synapsid satisfying Kemp’s definitions is Tikitherium, (Late Triassic, Datta 2005).

Here
in the large reptile tree, not all known taxa are listed. Here monotremes, like Ornithorhynchus and Akidolestes, appear after Chiniquodon. Sinocodon nests within the Mammalia. The following are traits separate monotremes from Chiniquodon in the large reptile tree, which employs traits not specific to synapsids and mammals. Other mammals may not have these traits. Traits specific only to monotremes are not listed.

1. Parietals fused, frontals fused
2. Quadrate becomes an ear ossicle
3. Supraoccipital fused to tabulars and opithotics (data missing for Chiniquodon)
4. Premaxillary teeth, more than four (lost in Ornithorhynchus, of course).
5. Last maxillary tooth below posterior orbit
6. Sacral vertebrae reduced to two (not Akidolestes)
7. Interclavicle evolves from long ‘I’ shape to short ‘T’ shape
8. Sternum present
9. Ulna > 3x radius + ulna width
10. Longest metacarpals: 3 and 4 (except Ornithorhynchus where 2-4 are subequal)
11. Longest manual digit is not 4
12. Ilium posterior process is not longer
13. Pubis orientation is not strictly medial
14. Prepubis bone present
15. Tarsus < 0.6x pedal digit 4 length
16. Longest pedal digits 2-4 (rather than 4, not known from Chiniquodon)
17. Pedal 4.1 proportions: length/width not < 3:1

Most of these, 
it goes without saying, are provisional subject to additional taxa and more precise data.

References
Datta PM 2005. Earliest mammal with transversely expanded upper molar from the Late Triassic (Carnian) Tiki Formation, South Rewa Gondwana Basin, India. Journal of Vertebrate Paleontology 25 (1): 200–207. doi:10.1671/0272-4634(2005)025[0200:EMWTEU]2.0.CO;2.
Kemp TS 2005. The Origin and Evolution of Mammals. United Kingdom: Oxford University Press. p. 3. ISBN 0-19-850760-7. OCLC 232311794.
Rowe T 1988. “Definition, diagnosis, and origin of Mammalia”(PDF). Journal of Vertebrate Paleontology 8 (3): 241–264. doi:10.1080/02724634.1988.10011708.

Former reptile: Gymnarthrus. Former reptile, former amphibian: Diadectes. Both from Case 1910.

Case 1910
described several skulls from what he presumed were Permian deposits in Archer County, Texas. Yes, they are Early Permian and home to many a Dimetrodon.

Among the several skulls
was Gymnarthrus willoughbyi (Fig. 1), known from a tiny 1.6cm skull. Case reported: “It was thought at first that both the basisphenoid and the parasphenoid process constituted the the parasphenoid bone and that the animal was an amphibian, but this is impossible… the animal approaches the intermediate form between the amphibians and reptiles.” Today we know Gymnarthrus to be one of the lizard mimics, the lepospondyl microsaurs. Case also wrote, “The nearest approach to this form is the small amphibian skull described by Broili as Cardiocephalous sternbergii, but this is described as having the skull complete, no parietal foramen, teeth regularly diminishing in size anteriorly but with cutting edges and lyra present.” I don’t know what lyra are in this context.

Figure 1. Gymarthrus willougbyi, drawn by Case 1910 on the left and von Huene 1913 on the right. These are apparently freehand sketches and, judging by the perspective implied by the large orbit on the right, sketched from two distances.

Carroll and Gaskill 1978
allied Gymnarthrus with Cardiocephalus, another microsaur.

Figure 2. Diadectes phaseolinus in situ, as originally illustrated and as reillustrated above according to phylogenetic bracketing. Case reported the tail was as long as the presacral portion of the column, but did not illustrate it that way for this specimen. No intercentra were present.

Case also identified Diadectes as a reptile
(order Cotylosauria), but later authors (and currently Wikipedia, taken from a PhD thesis by R Kissel 2010) considered it a reptile-like amphibian. The large reptile tree nests Diadectes as derived from Milleretta and all the “Diadectomorpha” listed in Wikipedia are reptiles. Limnoscelis, Orobates and Tseajaia do not nest with Diadectes in the large reptile tree, but bolosaurids and procolophonids do. So we’ve got some housecleaning to do at that node.

The interesting thing about this Diadectes specimen,
according to Case 1910, is the set of expanded dorsal ribs beneath the scapulae. He writes, “The ribs of the third, fourth and fifth vertebrae show a well defined articular end with a distinct neck. The bodies of these ribs are expanded into thin triangular plates, with the front edge straight and the posterior edge drawn out into a point which overlaps the succeeding rib; this forms a strong protection for the anterior thoracic region. The sixth, seventh and eighth [ribs] are overlain by thin, narrow, plates which continue backward the protection of the thoracic region to a point opposite the posterior end of the scapula.” Some, but not all Diadectes specimens have such expanded ribs.

Case presumed
that gastralia (his ‘abdominal ribs’  were present. They are not. Case notes “the animal was distinctly narrow chested, with the bones of the the girdle strongly interlocked. Diadectes had practically no neck.”

Based on the mounted skeleton, Case reiterated
“the suggestions previously made by the author that these animals are the nearest discovered forms to the ancestors of turtles.” That old hypothesis has not been confirmed by the large reptile tree, as noted earlier.

References
Carroll RL and Gaskill P 1978. The Order Microsauria. Memoirs of the American Philosophical Society 126:1-211 [J. Mueller/T. Liebrecht/T. Liebrecht]
Case EC 1910. New or little known reptiles and amphibians from the Permian (?) of Texas. Bulletin of the American Museum of Natural History 28 (17):163-181.
Huene FRF von and Gregory WK 1913. The skull elements of the Permian Tetrapoda in the American Museum of Natural History, New York. Bulletin of the AMNH ; v. 32, article 18.: 315-386.

A little eutherian/metatherian convergence

Convergence
between marsupials (metatherians) and placentals (eutherians) is rather commonplace. Today we’ll take a quick look at two mice-like taxa with arboreal affinities and omnivorous diets, Dromiciops (metatherian) and Ptilocercus (eutherian) (Fig. 1).

Figure 1. Dromiciops (marsupial, left) is similar to Ptilocercus (placental, right), until you look at the teeth and other details. Here, the marsupial has a larger braincase. The processes angularis is the hook-like shape at the bottom rear of the mandible.

Despite their similarities
these two nest separately in the large reptile tree. The NHC website notes the following differences between marsupial and palcental skulls:

1. large face/small braincase – not true in figure 1.
2. rear part of jaw turned inward (inflected jaw angle) in marsupials. Judging by the shadows this appears to be true.
3. marsupials have more teeth including 3 premolars + 4 molars – true in figure 1.
4. placental mammals typically have two sets of teeth, marsupials replace only some of their teeth

According to Wikipedia
distinct from placentals, marsupials have:

1. foramen lacrimale in the front of the orbit- in both taxa in figure 1
2. the cheekbone is enlarged and extends further to the rear – not true in figure 1
3. the angular extension (processus angularis) of the lower jaw is bent toward the center. Judging by the shadows this appears to be true.
4. hard palate has more openings.
5. different number of incisors in the upper and lower jaws (except wombats)
6. more premaxillary incisors than placentals – true in figure 1.
7. more molars than premolars – true in figure 1.

Hope this helps.
This is all fresh data for me. And BTW, none of these traits divide metatherians from eutherians in the large reptile tree. There’s a whole other list — but, it works only with the few taxa I’ve worked with so far. What would happen with additional taxa is currently beyond the realm of this study.