The Evolution of Fur - A Historic Perspective

Fig. 1: A 1988 illustration by Gregory S. Paul, showing two South African therocephalians, Trochosaurus, feeding on their distant tapinocephalian relative Styracocephalus. The small animal in the foreground is the therocephalian Blattoidealestes. All animals are notably drawn with a thin coat of hair.

When did hair first evolve? I do not know. Neither does anyone really at this point. Compared to the evolution of feathers, which is fairly well understood thanks to the discovery of amazingly preserved dinosaur-fossils, the first appearance of mammalian fur is quite a bit more mysterious. Is it something that is unique to mammals and their immediate ancestors or did it have its origins farther back in time? Were famous stem-mammals, such as Dimetrodon, Inostrancevia or Lystrosaurus scaly, leathery or hairy in appearance? It is a contentious question and one that continues to be relevant when we try to reconstruct these animals. While this article will not offer a definitive answer, it is nonetheless interesting to see how the different takes on these questions have evolved over time and how they might change in the future.

The mammal-like “reptiles” (Late 19th century – 1968)

Fig. 2: Dimetrodon as portrayed by Charles R. Knight in 1897. The skin and posture is strongly inspired by modern living reptiles. Guessing by the spiny sail, the pelycosaur in the background may be an early rendition of Edaphosaurus.

Amniotes, the animals which develop their young inside an amniotic sac (which includes us), were classically split up into three major groups: Anapsida, Diapsida and Synapsida, named after the number or form of their post-orbital skull openings or fenestra. Anapsida, sometimes called Parareptilia, have none and it was once thought that this group gave rise to turtles. Diapsida have two openings and the group includes lizards, snakes, crocodilians and dinosaurs (birds included), as well as a whole host of extinct forms. Synapsida, the group that would eventually give rise to mammals, have only one such opening (in humans this was eventually closed off, but you can still feel where it used to be if you touch your sphenoid bone or temple). The name means “fused arches”, as it was originally thought that synapsids descended from the diapsids and evolved their opening by combining the two fenestra. All synapsids that were not mammals were therefore classified as a sub-class of reptiles in classic taxonomy. It was only later realized that they did not descend from diapsids, that they instead just shared a common ancestor among the stem-amniotes and that they evolved their skull openings independently. The name, however, stuck. To quote zoologist John C. McLoughlin: “Like the Indians of North America, our remote ancestors were named by bumbling neophytes” (McLoughlin 1980, p. 29). What also stuck was the perception of early synapsids as reptiles or reptile-mammal missing links. They were most commonly called “mammal-like reptiles”, a term that one can still find in modern books, documentaries and museums. This then had a great effect to how they were represented in art. Even though no skin-impressions were known at the time, famous artist Charles R. Knight firmly believed that these animals had scales, with hair being a development of only mammals proper, and illustrated them as such (Knight 1946, p. 10). Most famous is his illustration of Dimetrodon from 1897, showing the animal with a very crocodile-like skin and a sprawling gait. His 1946 drawing also has some elements of a tuatara. His paintings of other synapsids, such as dicynodonts and gorgonopsians, were in similar vein. Works like these of course had a great influence on public perception, which came to view these animals as examples of early reptiles to the point where Dimetrodon is even today often mistaken as a dinosaur. That many of these species were given names ending in -saurus (Ancient Greek for lizard) such as Varanosaurus, Edaphosaurus or Lystrosaurus of course did not help (in my opinion the suffix saurus is way too overused in nomenclature and often inadequate at describing a species’ appearance).

Hippo-skin, whiskers and cladistics (1968 - 2012)

Fig. 3: “He has no style, he has no grace, this therapsid has a funny face”, to paraphrase the DK Rap. Estemmenosuchus as illustrated by Julius Csotonyi in 2011, shown with a smooth, glandular skin. My scanner unfortunately could not catch the entire picture, showing the rest of the herd. Splashing around in the background is distant relative Eotitanosuchus. 

After decades of illustrating early synapsids as scaly reptiles, an interesting discovery was made in 1968. Skin-impressions of Estemmenosuchus, a dinocephalian from the Middle Permian first described in 1960, were found. These showed that the animal was in fact not scaled, at least in the parts found, but instead possessed a smooth, glandular skin, very similar to what you would expect from a hairless mammal. This made sense from an evolutionary viewpoint, as this means it likely possessed a form of sweat-gland, out of which milk-glands would one day evolve (or perhaps already did?). The problem was the question of how representative Estemmenosuchus was for the rest of the early synapsids. As a dinocephalian therapsid it was fairly midway on the road from the first synapsid to the first mammal. It was more closely related to us than to Dimetrodon, however too basal to give much clues about later therapsids like dicynodonts or gorgonopsians. Further compounding was the fact that Estemmenosuchus is theorized to have had a specialized semi-aquatic lifestyle similar to a hippopotamus (to the point where I fear that Estemmenosuchus-Hippo comparisons will one day become one of those cliché paleoart-memes). Its skin-type might have been similarly derived and not representative of the ancestral state. That said, if it descended from a scaled ancestor it would make little sense for it to become scaleless, as many aquatic and marine reptiles keep their scales. However, it is possible that, just like a hippo, it may have descended from a hairy ancestor and secondarily became smooth-skinned. Or all its ancestors simply were already smooth-skinned and there was nothing special about its skin-type. Artists largely went with the latter hypothesis and in many great artworks since the 60s and up until very recently non-mammalian synapsids were often reconstructed with smooth, hair- and scaleless skin like that of hippos, elephants, rhinos or naked mole rats. The art of Zdeněk Burian is a good example of this.

Fig. 4: The Triassic dicynodont Lystrosaurus, painted by Zdeněk Burian and apparently imagined to have had a semi-aquatic lifestyle and smooth, hippo-like skin. Fun Fact: At one point in time this genus was actually the most abundant tetrapod on Earth, due to the ecologic imbalance left by the mass-extinction at the end of the Permian.

Some of course went a bit further than that and it was not uncommon to see illustrations of cynodonts with hair. From my own childhood books I particularly remember Cynognathus often being shown as this quasi-mammal with fur and ears, but also sprawling legs and lizard-tail. Also common was Thrinaxodon with whiskers (often only whiskers), the most prominent example being in the first episode of Walking with Dinosaurs. The speculation about whiskers in cynodonts goes all the way back to the 1950s when it was speculated that foramina found in their snouts were attachment-sites for vibrissae. However very similar structures were also found in the skulls of modern reptiles, such as the tuatara, which obviously do not have whiskers. We can therefore not definitely say if they had whiskers or not based on these foramina.

Fig. 5: The Cynognathus of my childhood (ca. 2005). I remember no book back then about prehistoric animals that did not have its own version of the supposedly furry cynodont. Today however it feels like it is only rarely featured in books. I wonder why that is. By the way, the German text states that the animal had fur as if it were a fact, even though this would be viewed as speculative even today.

Of course, speculating about hair in non-mammalian cynodonts was not particularly spectacular, as they were the immediate relatives of mammals and showed many signs of already being warm-blooded. During the 70s and 80s some went a bit further however. In his 1979 book Archosauria, John C. McLoughlin for example illustrated an anteosaur-like dinocephalian (the exact species is not given) not only with fur but also with little ear-pinnae, something that would be seen as highly speculative even today, given how “primitive” dinocephalians still were. In the semi-sequel Synapsida from 1980 he however seems to have doubled down on this, illustrating most synapsids as smooth-skinned and speculating that fur was unique to cynodonts (McLoughlin 1980, p. 84). While 1986’s The Dinosaur Heresies was, as you guessed it, primarily about dinosaurs, author Robert T. Bakker dedicates a chapter to the evolution of warm-bloodedness and speculates about endothermy in the Middle Permian synapsids of the Karoo:

“In Kazanian times [today this would be called the Guadalupian epoch], the Karoo experienced cool winters because it was closer to the South Pole than it is now. To remain active all year round, the small animals might well have needed some kind of insulation. Orthodox paleontology insists that hair is a uniquely mammalian invention, the adaptive badge of our own Class Mammalia. Several dissenters have however suggested that perhaps hair evolved long before the first true mammal appeared in the Late Triassic. Perhaps hair did evolve at the very beginning of warm-bloodedness, back in the Kazanian, fully forty million years before the earliest true mammals. Maybe our picture of the Kazanians should include shaggy protomammals stalking their prey through a winter snowstorm, hot breath steaming from their nostrils.” (Bakker 1986, p. 414)

The illustrations in the book however still show them as smooth-skinned, as was the norm of the time. His vision however became prophetic for later paleoart.

Fig. 6: A fuzzy dinocephalian, perhaps meant to be Anteosaurus, as illustrated by John C. McLoughlin in 1979’s Archosauria. His later depictions of synapsids in Synapsida were surprisingly more conservative, perhaps because he was criticized for a lot of his speculative takes in the previous book (although many of his predictions, like feathered theropods and rearing diplodocoids, later turned out to be true).

The late twentieth century also saw the advent of cladistic classification, which made the classification of early synapsids as reptiles obsolete. They could now be viewed as their own lineage in the amniote family-tree, distinct from reptiles (which themselves became regrouped into Sauropsida, consisting of diapsids and parareptiles). Despite its still widespread usage in popular media, the term “mammal-like reptiles” became outdated and it has become more common for non-mammalian synapsids, such as Dimetrodon, to be simply called stem-mammals by experts. Colloquially the term protomammal is also used. The notion of glandular-skinned stem-mammals was further complimented by the discovery of skin-impressions by sphenacodontids such as Haptodus in 2009, which showed they possessed dermal scutes, likely similar to those found in crocodilians or turtles. These also found their way into illustrations of pelycosaur-grade synapsids (synapsids that are not therapsids).

Nobody expects the Soft-Tissue Inquisition (2012 – 2018)

Fig. 7: What? Never seen a furry Moschops? You prude. This image was posted on Tumblr by someone with the username Alphynix in 2016, right during the Soft Tissue Revolution which we are still experiencing. It follows the common trend of taking a fairly well-known creature and illustrating it in novel and speculative ways to break up tropes and traditions.

In the wake of the Dinosaur Renaissance, paleontology entered a new golden age of discoveries and artwork. However not everything was fine and dandy, especially in the reconstruction-department. Beginning in the 2010s it was noticed that paleoart ran the danger of becoming too reliant on tropes and unspoken rules that were not necessarily based on facts. Especially shrink-wrapping, the act of neglecting the influence of soft-tissues on reconstructions, became a major problem. In response to all of this All Yesterdays, written and illustrated by Darren Naish, John Conway and Memo Kosemen, was released in 2012 to point out and subvert various unfounded tropes and clichés in paleoart and to encourage more well-informed speculation on things that are not preserved in the fossil-record. While the book was mainly focused on dinosaur-reconstructions it started a broader trend in paleoart as a whole called the Soft-Tissue Revolution. Reconstructions of stem-mammals were not untouched by this and many new artworks started to appear which showed these animals in new, speculative and unique ways. Perhaps in emulation of the trend of reconstructing increasingly more dinosaurs with feathers, it became more common to show therapsids outside of Cynodontia with fur. Gorgonpsians were a favorite candidate, as they already resembled dogs and wolves in overall appearance, but also dicynodonts and even dinocephalians became furrier. The squirrel-like arboreal dromasaurs also made more sense in a coat of hair. Some went even further than that and illustrated pelycosaur-grade synapsids with distinctively mammalian traits. A good example comes from All Your Yesterdays, a fan-sourced sequel to All Yesterdays. Here an illustration by Mike Keesey shows a Dimetrodon-mother with tufts of hair and a fatty sail, sweating out proto-milk from her sides which is then licked up by her young, a behaviour very much inspired modern monotremes such as the platypus. Around the internet more radical examples can be found, such as Dimetrodon with full fur-coats, whiskers and even mammalian muzzles. How likely all these speculations are remains debatable, but recent finds partially supported some of it. In 2016 Piotr Bajdek et al. described Late Permian coprolites (fossilized dung) from Russia, produced by either therapsids or archosaurs. In one of them they found traces of what was most likely hair, probably coming from the predator’s last meal. This presents the best evidence of fur having evolved long before true mammals appeared, but to whom this hair belonged is frustratingly unknown, as neither the prey nor the predator could be identified. The likeliest candidate would be a cynodont, however dicynodont bones were also abundantly present in the coprolites. A year before that, a study done on Ophicacodon, a very basal synapsid from the Early Permian that essentially looked like a Dimetrodon without the iconic sail, also showed distinctive signs of endothermy in this species. Endothermy, the ability to control one’s body-temperature through internal metabolic processes (not to be confused with homeothermy, the ability to keep a stable body-temperature), is a major step towards true warm-bloodedness and eventually necessitated the evolution of insulation, such as fur. Years before this study, in 2009, evidence for another distinctively mammalian trait, parental care, was found in the varanopid Heleosaurus, a synapsid even more basal than either Ophiacodon or Dimetrodon.

Fig. 8: A speculative depiction of Dimetrodon, titled Texan Mama, illustrated by Mike Keesey and first appearing in 2013’s All Your Yesterdays. The animal has a fatty hump surrounding its sail, tufts of hair and feeds its children with proto-milk-like secretions from its skin. Believe it or not, but I have seen speculative drawings that were way more bizarre than this. 

Both in science and art, the stem-mammals were beginning to take on a distinctive new look. No longer did they look and act like primitive reptiles, but instead came to resemble their mammalian descendants a lot more. Many of them also stopped being ugly or weird-looking and the desire to get a furry Diictodon as a pet became irresistible. Then everything changed when the Varanopid-Nation attacked.

Revenge of the Scales (2018 - 2019)

In the year 2018 a new species, Ascendonanus nestleri, was discovered. It was a member of the Varanopidae, one of the most basal clades of the Synapsida (though read on). It was significant in that it was preserved with an outline of the body and complete impressions of the skin, the first pelycosaur-grade synapsid to do so. It showed that it was covered in scales very similar to those of a lizard. This was rather surprising, as the lack of reptilian scales was up to that point used as a defining trait of synapsids. It also meant that the evolution of fur likely did not occur at the pelycosaur-grade but a lot later and that the smooth skin of Estemmenosuchus may have just been a transitional phase, not the ancestral state. This put a lot of the smooth-skinned and furry reconstructions of pelycosaurs into question. Artists were however quick to adapt. Shortly after the discovery Mark Witton produced a new rendition of Dimetrodon, clad in scales and obviously inspired by Charles Knight’s 1946 version, which is set to appear in the as of yet unreleased book Life through the Ages II. It is supposed to come out this year, however something awkward has happened in the meantime.

Fig. 9: A scaly Dimetrodon illustrated recently by Mark Witton, very much inspired by Knight’s 1946 rendition of the animal. It is however not entirely retro, as the sail-tips are now exposed (something which is currently believed to have been the case) and the animal is surrounded by juveniles, presumably its children.

The Big Sike (2019 – point of writing)

Almost like an early Christmas present, on December 23 2019, palaeontologists David P. Ford and Roger B. J. Benson presented a new cladistic analysis of early amniotes. Not only did they conclude that parareptiles were in fact a group of derived diapsids, but that the same was true for varanopids. This is not that surprising, considering that similar results had been produced by earlier studies for individual members of both, but this was the first extensive study that used both groups in their entirety. While at this point it remains to be seen if this becomes consensus, if true, it has many implications for the early evolution of amniotes, but also for the recent reconstructions of stem-mammals. Ascendonanus and its relatives would not be synapsids anymore, its scales being nothing out of the ordinary since it would be more closely related to lizards than to mammals. It therefore could not be used anymore as a reference to reconstruct early synapsids. Where does this leave us? On one hand it seems safe to go back to pre-2018 speculations, meaning smooth, glandular skin and perhaps fur, until skin-imprints of definitive pelycosaurs are discovered. On the other we might also have to abandon some older concepts. Heleosaurus, used as evidence for parental care, was also a varanopid and could therefore not be used as reference anymore as well. 

Fig. 10: I sometimes make half-assed memes in my free time. It is a peaceful life.

This entire story just goes to show how rapid and fluctuous of a science paleontology and paleontology-illustrations have become. While it means we should never grow too attached to a single piece or style of reconstruction, it is a good thing as the growing data and discussion only helps us better understand deep time and our own origins. The fact that paleoart ages is a feature, not a flaw, to paraphrase Mark Witton. Even when an artwork ages, it is still useful to us, as it literally illustrates what science and the culture around it was like at a given point in time. On a deeper level we have also seen how interested we apparently are in the evolution of fur and our earliest relatives. It is always strange to realize that we share more in common with bizarre creatures such as Dimetrodon or Moschops than one might expect, so it is tempting to familiarize them with hair and parental care, all traits that we do not only recognize in other mammals, but also in ourselves. While we humans like to think of ourselves as naked apes, hair is still very much part of our bodies and our lives (we in fact have more hair-follicles than chimpanzees, our hairs are just a lot thinner and shorter). Our hair is not a vestige from prehistory, it still serves a purpose, the most obvious in ours being that it helps us shape our appearance and personality. And hey, perhaps, just perhaps, back in the Middle Permian the first hairy synapsid got quite a lot of looks from the ladies and that is why his lineage still lives on. 

If you liked this and other articles, please consider supporting me on Patreon. I am thankful for any amount, even if it is just 1$, as it will help me at dedicating more time to this blog and related projects. Patrons also gain early access to the draft-versions of these posts.

Literary Sources:

• Bakker, Robert Thomas: The Dinosaur Heresies. New Theories Unlocking The Mystery of the Dinosaurs and Their Extinction, New York 1986.
• Benton, Michael James: Vertebrate Paleontology, 1990 (3. Edition from 2005, German translation).
• Burian, Zdeněk/Spinar, Zdeněk: Life Before Man, Prague 1972 (Revised Edition from 1995).
• Conway, John/Kosemen, C.M./Naish, Darren: All Yesterdays. Unique and Speculative Views of Dinosaurs and Other Prehistoric Animals, London 2012.
• Conway, John/Kosemen, C.M./Naish, Darren: All Your Yesterdays. Extraordinary Visions of Dinosaurs and Prehistoric Animals from a New Generation of Paleoartists, London 2013 (Second Edition from 2017).
• Chambers, Paul/Haines, Tim: The Complete Guide to Prehistoric Life, London 2005.
• Csotonyi, Julius/White, Steve: The Paleoart of Julius Csotonyi. Dinosaurs, Sabre-Tooths and Beyond, London 2014.
• Knight, Charles Robert: Life through the Ages, New York 1946 (Commemorative Edition).
• McLoughlin, John: Archosauria. A New Look at the Old Dinosaur, New York 1979.
• McLoughlin, John: Synapsida. A New Look into the Origin of Mammals, New York 1980.
• McRae, Anne/Nardi, Marco: Tessloffs Enzyklopädie der Dinosaurier, Nürnberg 2005. 
• Paul, Gregory Scott: Predatory Dinosaurs of the World. A Complete Illustrated Guide, New York 1988.

Papers:

• Botha-Brink J, Modesto SP. A mixed-age classed 'pelycosaur' aggregation from South Africa: earliest evidence of parental care in amniotes?. Proc Biol Sci. 2007;274(1627):2829–2834. doi:10.1098/rspb.2007.0803
• Ford, D.P., Benson, R.B.J. The phylogeny of early amniotes and the affinities of Parareptilia and Varanopidae. Nat Ecol Evol 4, 57–65 (2020). https://doi.org/10.1038/s41559-019-1047-3

Online Sources:

• Ancestry of mammalian 'warm-bloodedness' revealed by Science Daily
• Getting to the Root of Fur by National Geographic
• The Stem-Mammals--a Brief Primer by Tetrapod Zoology (Ver. 3).

Image Sources

• Fig. 1: Paul 1988, p. 52.
• Fig. 2: Wikimedia
• Fig. 3: Csotonyi 2014, p. 6-7.
• Fig. 4: Burian 1972, p. 106
• Fig. 5: McRae 2005, p. 25
• Fig. 6: McLoughlin 1979, p. 102-103
• Fig. 7: Tumblr-user Alphynix
• Fig. 8: Conway 2019, p. 99.
• Fig. 9: Mark Witton's Twitter.