The bone 2 cartilage 2 bone transition from sturgeons 2 sharks 2 bony fish

Short summary for those in hurry:
There is support in Pehrson 1940 for the origin of facial (dermal) bones on a cartilaginous template (contra Hall 2005) in a proximal shark descendant.

  1. Sturgeons (shark ancestors in the LRT) have facial bones sheathed to a cartilage template.
  2. Sharks lose all trace of bone, but keep the cartilage.
  3. Bony fish (shark descendants in the LRT) reacquire facial bones on a cartilage template

Backstory
Several recent reader comments disputed and/or cast doubt on the identity of shark skull bones (Fig. 2) and the shark-to-bony fish transition recovered by the large reptile tree (LRT, 1771+ taxa, see Fig. 1 diagram). Objections were  based on developmental grounds. One reader (CB) wrote: Most of the bones you’re trying to identify on shark chondrocrania are dermal bones. That means they don’t pre-form in cartilage. Which means animals without a bony skull cannot have them.”

That is the traditional view found in current textbooks.

First:
my guess is this comment resulted after reading any of several authors all citing Hall 2005, who wrote, “The vertebrate dermal skeleton includes the plate-like bones of the skull, and, in reptiles and fishes, also includes various scales, scutes, denticles and fin rays. Dermal bone forms via a process known as intramembranous ossification, with mesenchymal condensations differentiating directly into bone without a cartilaginous template.”

Second:
As everyone knows, no part of shark skulls is bone. It’s all cartilage. Nevertheless and despite obliteration and/or fusion of most skull sutures, shark ‘nasal’ templates still cover the snout and nares. Shark ‘frontal’ templates are still located between the eyes. I have retained tetrapod skull nomenclature for shark skull template elements in order to include shark taxa in the LRT.

Third:
A valid phylogenetic context, like the LRT (diagram in Figs 1, 4), is vital in matters like this. Taxon exclusion leading to an improper cladogram is the root cause of most prior misunderstandings, as readers well know.

Wagner and Aspenberg 2011 wrote:
“Bone is specific to vertebrates, and originated as mineralization around the basal membrane of the throat or skin, giving rise to tooth-like structures and protective shields in animals with a soft cartilage-like endoskeleton.”

That’s not correct. In sharks dentine and enamel from the skin and teeth are not bone. Instead, bone first appears in sturgeons and kin. Then it disappears in sharks only to reappear in bony fish + tetrapods, according to the LRT. Traditionally and mistakenly sturgeons were considered relatives of derived bony fish, which is part of the problem.

In sturgeons and paddlefish, Bemis et al. 1997 report, 
“the bones more or less closely ensheath the underlying endochondral rostrum”. Sharks lack this sheath of bone on the rostrum. Instead, remaining more flexible cartilage supports the skull and skeleton.

Figure 2. Acipenser brevirostrum, 1m typical length. Records up to 1.47m.

Figure 2. Acipenser brevirostrum, 1m typical length. Records up to 1.47m.

Keys to understanding this issue include:

  1. Elements of the dermocranium in shark outgroup taxa: sturgeons (Fig. 1) and paddlefish = bone sheath over cartilage.
  2. Elements of the dermocranium in sharks (Fig. 2) = prismatic cartilage
  3. Elements of the dermocranium in proximal shark descendants: the bowfin, Amia (Figs. 2, 3) = bone patches develop around sensory cells over a cartilage template, according to Pehrson 1940.
Figure 2. Fish evolution from Hybodus to Amia documenting the shark to bony fish transition.

Figure 2. Fish evolution from Hybodus to Amia documenting the shark to bony fish transition.

Pehrson 1940 examined
a series of embryonic stages of the extant bowfin, Amia calva (Fig. 3), one of the most primitive bony fish in the LRT. Pehrson 1940 reports: “Three different stages of the formation of the premaxillary are shown. The anterior, dental part of the bone is clearly distinguishable from the posterior and dorsal part, situated above the cartilage.”

The ontogenetic origin of bone in Amia (Fig. 3) first appears in embryos as tiny islands on the skull surface over a cartilage or pre-cartilage template. This proximal descendant of hybodontid sharks (Fig. 2) documents many skull homologies.

Figure x. Embryo development in the bowfin, Amia. The facial bones develop as buds surrounding dermal sensory organs 'floating' on top of a cartilage base.

Figure 3. Embryo development in the bowfin, Amia. The facial bones develop as buds surrounding dermal sensory organs ‘floating’ on top of a cartilage (chondral) and prechondral base.

It is noteworthy
that the appearance of bone surrounding sensory cells all over the skull in bony fish followed the reduction of the long, sensory-cell-filled rostrum in bony fish. Taking the other evolutionary route, other shark descendants (e.g. hammerheads, skates, rays, goblin sharks, elephant-nosed chimaera, sawfish), further elongated the rostrum for increased acuity in finding bottom-dwelling prey.

Pehrson also described
the appearance of ossification where prior cartilage dissolved, convergent with the process of fossilization. Thereafter some embryos began to develop ossified skull bones without a cartilaginous template, in accord with Hall 2005, who did not cite Pehrson 1940.

Surprisingly,
Pehrson was keen on naming fish bones in accord with those of pre-tetrapods. He reports, “There seems to be no doubt that the intertemporal and supratemporal parts of the developing composite bone correspond to the similarly named bones in Osteolepidae and Rhizodontidae.” Not sure if Pehrson was the first to do this, but it should be standard.

Supporting evidence that sturgeons are shark ancestors:
According to Wikipedia, notable characteristics of Acipenseriformes include:

  1. Cartilaginous endoskeleton – as in sharks and fish more primitive than sharks
  2. Lack of vertebral centrum – as in fish more primitive than sharks
  3. Spiral valve intestine – as in sharks, bichirs, gars and lungfish, the last two by reversals.
  4. Conus arteriosus = infundibulum, a conical pouch found in the heart from which the pulmonary trunk artery arises (not sure how this relates, but there it is).

Bemis et al. report,
“Acipenseriforms are central to historical ideas about the classification and evolution of fishes.”

Indeed. The LRT comes to the same conclusion.

“Acipenseriforms also are noteworthy because of their unusual mixture of characters, which caused early debate about their classification. Two aspects of living Acipenseriformes were especially problematic for early ichthyologists: (1) reduced ossification of the endoskeleton combined with presence of an extensive dermal skeleton; and (2) the presence of a hyostylic jaw suspension and protrusible palatoquadrate recalling the jaws of sharks.”

These aspects are not problematic of sturgeons and paddlefish are basal to sharks.

The palatoquadrate is neither a palatine nor a quadrate. It is largely homologous to the lacrimal with fusion of the tiny quadrate and tall, curved, preopercular in most taxa, fusion of the premaxilla and maxilla (tooth-bearing elements) on taxa with teeth. The former and future jugal is also typically fused.

Figure 5. Sturgeon mouth animated from images in Bemis et al. 1997. This similar to ostracoderms, basal to sharks.

Figure 5. Sturgeon mouth animated from images in Bemis et al. 1997. This similar to ostracoderms, basal to sharks.

“The current conventional view (developed and refined by many authors… holds that Acipenseriformes evolved from a ‘paleonisciform’ ancestor via paedomorphic reduction of the skeleton and specialization of the feeding system, but there is much more to the history of ideas about the systematics of this group.”

That is incorrect according to the LRT, which tests a wider gamut of fish and nests traditional acipenseriformes basal to unarmored sharks and derived from armored osteostracoderms (Fig. 4). There was no paedomorphic reduction of the skeleton at the origin of sturgeons. The sturgeon feeding system is not ‘specialized’. It is primitive.


References
Bemis WE, Findeis EK and Grande L 1997. An overview of Acipenseriformes. Environmental Biology of Fishes 48: 25–71, 1997.
Gillis JA 2019. ‘Secondary’ cartilage and the vertebrate dermal skeleton in Reference Module in Life Sciences.
Hall BK 2005. Bones and Cartilage. Academic Press, London. ISBN: 978-0-12-319060-4
Maisey JG 1983. Cranial anatomy of Hybodus basanus Egerton from the Lower Cretaceous of England. American Museum Novitates 2758:1–64.
Maisey JG 1987. Cranial Anatomy of the Lower Jurassic Shark Hybodus reticulatus
(Chondrichthyes: Elasmobranchii), with Comments on Hybodontid Systematics. American Museum Novitates 2878: 1–39.
Pehrson GT 1940. The development of dermal bones in the skull of Amia calva. Acta Zoologica 21:1–50.
Wagner DO and Aspenberg P 2011. Where did bone come from? An overview of its evolution. Acta Orthopaedica. 82(4):393–398.
The Skull, Volume 1. Eds. Hanken J and Hall BK University of Chicago Press Books, 1993.

https://en.wikipedia.org/wiki/Acipenseriformes
https://www.zoology.ubc.ca/~millen/vertebrate/Bio204_Labs/Lab_3__Skull.html
G Torsten Pehrson bio

Unwin and Martill 2019 find pterosaurs ‘naked’ and ‘ugly’

Unwin and Martill 2019 report:
“With key roles in flight, thermoregulation and protection of the body, the integument was of fundamental importance to pterosaurs. Determination of the basic anatomy of this structure could provide a range of new insights into the palaeobiology of these enigmatic volant reptiles. Presently, however, there are several conflicting hypotheses regarding the construction of the integument, all founded on limited numbers of specimens, and not one of which is fully consistent with the available fossil evidence.

As mentioned yesterday, pterosaurs are not enigmatic. Unwin and Martill have chosen to avoid the scaly lepidosaurian ancestors of pterosaurs cited by Peters (2000, 2007). The integument found on pterosaurs has similar precursor integument on sister fenestrasaurs like Sharovipteryx (Fig. 1) and Longisquama, adding two taxa to their short list of pterosaurs preserving scaly integument and pycnofibers exclusive of the extradermal membranes (wings and uropatagia).

Figure 1. Note the neck skin (integument) of Sharovipteryx, a pterosaur sister.

Figure 1. Note the neck skin (integument) of Sharovipteryx, a pterosaur sister.

Unwin and Martill continue:
“We have developed a new 
model based on investigations of more than 100 specimens all of which show some form of exceptional preservation. This data set spans the entire temporal and systematic ranges of pterosaurs and a wide variety of preservational modes.”

So… “a limited number of specimens” (see above) just turned into “more than 100 specimens.” Did they just want to see if anyone was paying attention?

“The model has three principal components:
(1) A thin epidermal layer. The external surface of the integument was glabrous [= free from hair or down, smooth] with a smooth, slightly granular, or polygonal texture.

Attenuate ‘bristles’ fringed the jaws in two anurognathids and small tracts of filaments may have adorned the posterior cranium in some pterosaurs.

Perhaps these jaw and skull filaments should have been separately numbered because they are different than glabrous tissue.

(2) A layer of reticular and filamentous collagen and of variable thickness and complexity, formed much of the dermis.

Helically wound bundles of collagen fibres (aktinofibrils), were present throughout all flight patagia. Variation of aktinofibrils in terms of their dimensions, packing, orientation and stiffness permitted localized variation in the mechanical properties and behaviour of the flight patagia whichvaried from relatively stiff distally to more extensible and flexible proximally.

‘Feather-like’ structures reported in Jeholopterus appear to be partially unraveled or decayed aktinofibrils.

Again, these are all distinct tissues worthy of their own numbers.

Unwin and Martill have no idea that Jeholopterus was a vampire bat analog (Peters 2008) covered like no other pterosaur with fluffy, silent, owl-like extradermal integument. Neither Unwin nor Martill seem to make reconstructions, so neither has any idea what Jeholopterus looked like, unless they looked here (Fig. 2).

Finally, Unwin and Martill are mixing in flight membranes here. Perhaps THAT is where they get so many examples because otherwise dermal material is exceedingly rare. Integument generally means ‘covering’, so their inclusion of wing membranes is a little misleading, especially considering the ‘naked and hairless’ portion of their abstract headline.

Figure 2. Reconstruction of Jeholopterus. This owl-like bloodslurper was covered with super soft pycnofibers to make it a silent flyer.

Figure 2. Reconstruction of Jeholopterus. This owl-like bloodslurper was covered with super soft pycnofibers to make it a silent flyer.

Collagen fibre bundles were also present in footwebs, and in the integument of the neck and body. These structures have often been mis-identified as ‘hair’ (pycnofibres).

Again, this variety of tissues should have been numbered separately because they are different than tissue forming much of the dermis.

(3) A deep dermal layer with muscles fibres, blood vessels and nerves.

This variety of demal tissues were already described for the flight membranes, but it could also apply to normal tetrapod skin, like our own.

The pterosaur integument was profoundly different from that of birds and bats, further emphasizing the sharp disparity between these volant tetrapods.”

Why didn’t Unwin and Martill compare pterosaur integument to lepidosaur integument, specifically that of Sphenodon and Iguana (Fig. 3)? These are the two closest living relatives of pterosaurs in the large reptile tree. According to the LRT, Unwin and Martill are looking in the wrong places.

The spines of Iguana.

Figure 3. The dorsal and gular spines of Iguana are homologous with those in Sphenodon.

Not sure where Unwin and Martill
are getting data for pterosaur skin exclusive of the extradermal membranes. They don’t say. The dark wing Rhamphorhychus (Fig. 4) has the most incredible preservation of extradermal membranes, but the skull, neck and torso were prepared down to the bone.

Figure 1. The darkwing specimen of Rhamphorhynchus. Top: in situ. Middle: Soft tissues highlighted. Bottom: Neck and forelimb restored.

Figure 4. The darkwing specimen of Rhamphorhynchus. Top: in situ. Middle: Soft tissues highlighted. Bottom: Neck and forelimb restored.

So, why do Unwin and Martill think the Mesozoic got ugly?
Their abstract does not seem to answer their click-bait headline, which describes naked, hairless and featherless pterosaurs without giving one example of same based on evidence. On the contrary, employing phylogenetic bracketing, between Sharovipteryx (Fig. 1), Scaphognathus and Sordes (the hairy devil, Fig. 5), basal pterosaurs were not naked. Their fibers were not the same as hair or feathers, but unique to fenestrasaurs.

The hind limbs and soft tissues of Sordes.

Figure 5. The hind limbs and soft tissues of Sordes. Above, color-coded areas. Below the insitu fossil.

Finally…
Why were pterosaurs considered naked by Unwin and Martill when hairy Sordes (Fig. 5) was studied by Unwin, known to Martill, and not mentioned in the abstract? Very strange, indeed coming from these two.


References
Peters D 2000b. A Redescription of Four Prolacertiform Genera and Implications for Pterosaur Phylogenesis. Rivista Italiana di Paleontologia e Stratigrafia 106 (3): 293–336.
Peters D 2002. A New Model for the Evolution of the Pterosaur Wing – with a twist. – Historical Biology 15: 277–301.
Peters D 2007. The origin and radiation of the Pterosauria. In D. Hone ed. Flugsaurier. The Wellnhofer pterosaur meeting, 2007, Munich, Germany. p. 27.
Unwin D and Martill D 2019. When the Mesozoic got ugly – naked, hairless, (and featherless) pterosaurs. SVPCA abstracts.

The origin of feathers and hair (part 1: skin and scales)

Three clades
developed extra dermal hair-like structures: mammals, dinosaurs (reaching an acme in birds) and pterosaurs. Traditional thinking holds that reptile scales evolved early, along with the origin of the amniotic membrane. Both of these were viewed as adaptations to a non-aquatic, (i.e. ‘dry’) environment. Unfortunately there’s very little evidence for scales in the earliest reptiles (see below). They appear to have lived in a moist coal forest leaf litter environment throughout the Carboniferous.

Basal amniotes
Dhouailly 2009 reports: “The common ancestor of amniotes may have presented both a glandular and a ‘granulated integument’, i.e. an epidermis adorned with a variety of alpha-keratinized bumps, and thus may have presented similarities with the integument of common day terrestrial amphibians. Whereas the glandular quality of the integument was retained and diversified in the mammalian lineage, it was almost completely lost in the sauropsid lineage (non-mammalian amniotes). When the amniote ancestors started to live exclusively on land in the late Carboniferous, they derived from a group of basal
amphibiotic tetrapods, and it is plausible that they evolved a skin barrier similar to that of modern toads to prevent desiccation”

Two dermal proteins
are key to discussions on reptile skin: alpa-keratins and beta-keratins.

Alpha-keratins
Dhouailly 2009 reports: “In all living vertebrates, at least from trout to human, specific types of alpha-keratins characterize the epidermis and corneal epithelium showing a strong homology in the different lineages.In all amniotes, the last supra-basal layers of the epidermis are cornified, meaning they are formed of dead cells filled entirely with alpha-keratin filaments coated with specific amorphous proteins and lipids, providing a barrier to water loss.”

Beta-keratins
Dhouailly 2009 reports: “Only the sauropsids (birds and reptiles) possess an additional capacity for beta-keratin synthesis, an entirely different type of intermediate filament, which appears to result from a phylogenetic innovation that occurred after that of the alpha-keratins.”

Perhaps a correction here: In the large reptile tree there is no clade “Sauropsida.” Rather synapsids are more derived than the basal reptiles that ultimately evolved into the other amniotes. So, if beta-keratins had a single origin, mammals and perhaps their closest ancestors lost the capacity to produce beta-keratins. Phylogenetic bracketing indicates this could have happened at any node between Protorothyris and Megazostrodon.

Scales
When did scales arise? And are lizard scales homologous with those of turtles, crocs, birds, pangolins and opossum tails? Unfortunately fossils of skin and scales are rare.

Carroll and Baird 1972
traced long interwoven ventral ‘scales’ for the basal lepidsosauromorph, Cephalerpeton,  similar to those found in basal amniotes like Gephyrostregus watsoni. Carroll and Baird also report, “The skin impressions along the forelimb [of Cephalerpepton] have a slightly pebbly texture—rougher than the limb bones but smoother than the broken surface of the matrix. There is no evidence of discrete scales. An indication of epidermal scales would be expected in this type of preservation, if they were present in the animal.epidermal scales can only be recorded as impressions and this type of preservation is rare and apparently not reported in other Paleozoic reptiles.”

As you’ll recall, reptiles divide at the start into two lineages, the Lepidosauromorpha and the Archosauromorpha.

Lepidosauromorphs 
The most primitive appearance of dermal tissue in the lepidosaurorph line occurs with the scutes of pareiasaurs, Sclerosaurus and basal turtles. These include a bony base. At some point turtles developed scales that covered the face and limbs, but when is not known. My guess is as far back as Stephanospondylus because it was a large and tasty herbivore.

Otherwise nothing on scales appears until  Xianglong, a gliding basal lepidosaurifom (not a squamate). Li et al. 2007 report, “The entire body including the skull is covered with small granular scales, which show little size variation.” Perhaps noteworthy, this is the node at which some short-legged, ground-dwelling flattened owenettids evolved to became large-limbed and arboreal, exposed to the dry air above the damp leaf litter.

Perhaps more misunderstood, those wing spars are actually ossified dermal extensions, as in a sister taxon, Coelurosauravus, not extended ribs, as we carefully considered earlier here and here.

Figure 1. Xianglong, a basal lepidosauriform with dermal extensions, not ribs, with which it used to glide.

Figure 1. Xianglong, a basal lepidosauriform with dermal extensions, not ribs, with which it used to glide.

Then there’s Sphenodon, an extant basal lepidosaur with a variety of large and small scales, some were overlapping and others were not. Basalmost sphenodontids, like Pleurosaurus, were trending toward an aquatic niche. Sphenodon is a burrowing and foraging reptile. the best clue to basal sphenodontid squamation comes from a tritosaur sister, Tijubina (see below).

Of course, all living lizards (squamates) have scales.
and they shed their skin in whole or in patches during ontogeny.

The tritosaur lepidosaurs are a special case,
The basal tritosaur, Tijubina, preserves rhomboid scales on the neck, large rhomboid scales on the trunk and annulated ones on the ventral side of the entire caudal region. Not far removed from Sphenodon with regard to squamation.

By contrast, a more derived tritosaur, Huehuecuetzpalli preserves tiny disassociated calcified granular scales over its dorsal neural arches. A more derived tritosaur, Macrocnemus (Renesto and Avanzini 2002), had a scale covering in the sacral and proximal caudal region.

Now things get more than interesting…
The scales of Cosesaurus (Ellenberger and DeVillalta 1974, Fig. 2) were about the size of the matrix particles in its mold, so they have not been described. However, Cosesaurus had extra dermal tissues in the form of a gular sac, a dorsal frill, fibers streaming from the posterior arm, uropatagia trailing the hind limbs and long hairs emanating from the tail. A larger sister, Kyrgyzsaurus had scales and similar extra dermal ornaments. Sharovipteryx shares these traits and accentuates the uropatagia. Longisquama shares these also but accentuates the dorsal plumes. The latter two taxa also have pycnofibers (hairs) at least surrounding the cervical series. Their sisters, the pterosaurs, accentuate the trailing arm fibers, which become fiber-embedded foldable wings. Pterosaur ‘hair’ reaches its acme in Jeholopterus, which may have used its ‘hair ball’ as a barrier to insects likewise attracted bloody patches of dinosaur skin. In certain basal pterosaurs the tail hairs coalesce to become tail vanes.

Figure 1. Click to enlarge. The origin and evolution of Longisquama's "feathers" - actually just an elaboration of the same dorsal frill found in Sphenodon, Iguana and Basiliscus. Here the origin can be found in the basal tritosaur squamate, Huehuecuetzpalli and becomes more elaborate in Cosesaurus and Longisquama.

Figure 2. Click to enlarge. The origin and evolution of Longisquama’s “feathers” – actually just an elaboration of the same dorsal frill found in Sphenodon, Iguana and Basiliscus. Here the origin can be found in the basal tritosaur squamate, Huehuecuetzpalli and becomes more elaborate in Cosesaurus and Longisquama.

It is clear in fenestrasaurs that extra dermal membranes were secondary sexual traits, decorations that enhanced their chances for mating. Hairs ultimately became barriers or acted as insulation. Arm fibers ultimately became wings.

Archosauromorphs
Like the basal lepidosauromorph, Cephalerpeton, basal archosauromorphs like Eldeceeon  had ossified belly scales in V-shaped patterns, but not coalesced to form gastralia.

According to Carroll and Baird (1972) in Brouffia, “many ventral scales are present in the blocks. They are quite broad, rather than being narrowly wheat-shaped, as has been considered typical in early reptiles. A faint impression of dorsal scales is evident also, but these are too insubstantial to illustrate.”

Pelycosaurs lacked scales. They were naked. So were basal therapsids as far as the fossil record goes. Estemmenosuchus (Chudinov 1970), an herbivorous therapsid, preserves no scales, hair or hair follicles. However, the preserved skin was well supplied with glands.

Since all living basal mammals (Fig. 3) are richly endowed with fur, that trait probably extends to the first tiny egg-laying mammals, denizens of the leaf litter. In tiny animals, so in contact with the substrate, the leaf litter and water, hair appears to have developed not only to insulate its little warm-blooded body, but also to act as a barrier to all dermal contact with the environment. Insects, like fleas, had to lose their wings to burrow past the hair to get to the skin.

Figure 2. This is Amphitherium a basal mammal.

Figure 3. This is Amphitherium a basal mammal.

In basal diapsids, the sisters of basal synapsids no dermal material has been found.

Plesiosaurs and ichthyosaurs were naked, so pachylpleurosaurs, thalattosaurs and mesosaurs were likely naked as well. A rare exceptions, the thalattosaur Vancleavea, was covered with large bony scales. Hupehsuchids had short bony plates over the neural spine tips.

For protorosaurs or proterosuchids no dermal scales have been reported.  Dorsal armor developed as large plates in the likely piscivore Doswellia, and to a less degree in Champsosaurus, Diandongosuchus, and then again to a greater degree in parasuchids, proterochampsids and chanaresuchids.

For euarchosauriformes a line of dorsal scutes also appeared on the dorsal midline of Euparkeria and many descendant taxa (except finbacks and poposaurs). The herbivorous Aeotosaurs,  Revueltosaurus and Simosuchus independently expanded their armor in similar ways. So did the carnivorous extant alligators and crocodiles and their ancestors. However basal bipedal and near-bipedal croc taxa, like Gracilisuchus do not preserve scales, other than their dorsal scutes. These may have enhanced the strength of the backbone. Otherwise, basal bipedal crocs were likely not heavily scaled, and neither were the oceanic swimmers, like Metriorhynchus.

That brings us to dinosaurs.
Kaplan (2013) reports “the overwhelming majority had scales or armor.” We’ll cover dinosaur scales and dinosaur feathers in more detail in part 3: feathers.

References
Kaplan M 2013. Feathers were the exemption rather than the rule for dinosaurs. Nature News. doi:10.1038/nature.2013.14379
Renesto S and Avanzini M 2002. Skin remains in a juvenile Macrocnemus bassanii Nopsca (Reptilia, Prolacertiformes) from the Middle Triassic of Northern Italy. Jahrbuch Geologie und Paläontologie, Abhandlung 224(1):31-48.
Carroll RL and Baird D 1972. Carboniferous Stem-Reptiles of the Family Romeriidae. Bulletin of the Museum of Comparative Zoology 143(5):321-363. online pdf
Bennett AF and Ruben JA 1986. The metabolic thermoregulatory status of therapsids. In The Ecology and Biology of Mammal-like reptiles (Hottom, Roth and Roth eds) 207-218. Smithsonian Institution Press, Washington DC
Chudinov PK 1970. Skin covering of therapsids [in Russian] In: Data on the evolution of terrestrial vertebrates (Flerov ed.) pp.45-50 Moscow: Nauka.
Dhouailly D 2009. A new scenario for the evolutionary origin of hair, feather, and avian scales. Journal of Anatomy 214:587-606.
Lecuona A and Desojo, JB 2011. Hind limb osteology of Gracilisuchus stipanicicorum(Archosauria: Pseudosuchia). Earth and Environmental Science Transactions of the Royal Society of Edinburgh 102 (2): 105–128.
Persons WC4 and Currie PF 2015. Bristles before down: A new perspective on the functional origin of feathers.Evolution (advance online publication)
DOI: 10.1111/evo.12634
http://onlinelibrary.wiley.com/doi/10.1111/evo.12634/abstract

* too bad I did not know this when I painted Estemmenosuchus with scales for the cover of a book.