Getting to the Root of Fur

First thing in the morning, my mind is on autopilot. I’m mostly relying on muscle memory to get me through the routine of feeding the cats, washing up, and getting dressed as my brain boots up. But the other day an unexpected question smacked me out of my usual stupor. On that particular morning I needed to shave, and my cat Margarita jumped up onto the sink to see what strange thing I was doing. When I looked into the sink and saw the scraped-off stubble, with my furry assistant trying to swat the stream of water from the faucet, I was struck by the unique body covering that signifies that the cat and I are part of the same evolutionary family – hair.

But where did hair come from? I’ve written plenty about the integument of dinosaurs and the ever-growing record of enfluffled saurians, but I couldn’t recall when hair became a defining feature of my ancestry. So with a few pointers from paleontologist Adam Huttenlocker, I started to sift through the literature to find out why I have to rake blades across my face every few days to remove itchy hairs instead of grooming feathers or shedding scaly skin.

We know that fur has to be ancient. The evolutionary tree of modern mammals is a very furry one, and the last time the three major lineages of today’s beasts – the monotremes, marsupials, and placentals – shared a common ancestor was in the Triassic, over 220 million years ago. There’s no indication that hair evolved more than once, so fur has to be at least that old.

Eomaia, surrounded by remnants of fur. Image from Kielan-Jaworowska and Hurum, 2006. CC BY 2.0
Eomaia, surrounded by remnants of fur. Image from Kielan-Jaworowska and Hurum, 2006. CC BY 2.0

Some spectacular fossils back up the notion that Mesozoic mammals were furry. A skeleton of Eomaia, a 125 million year old mammal found in China’s Liaoning Province, was preserved with a halo of fossil fur. Even older, the 165 million year old mammal Megaconus and the almost-mammal Castorocauda were fossilized with parts of their pelts, as have other exceptionally-preserved Mesozoic beasts found in China.

But fur didn’t necessarily start with mammals. If we look to feathers, for example, plenty of non-avian dinosaurs had them. Evolution just allowed birds to co-opt the insulating, flashy structures to take to the air. The same was probably true of fur, with sparse protohairs preceding full pelage. To get down to the root of hair, we need to go back beyond the first mammals.

Mammals belong to a much larger group of vertebrates called synapsids. For a long time the non-mammalian parts of this family tree – like the sail-backed Dimetrodon and weasel-like cynodonts – were called “mammal-like reptiles”, but this was off the mark. Better to call them protomammals, as they were more closely related to mammals than any reptile. Fur probably evolved within their protomammal ranks. The question is where.

The problem is that the fossil record isn’t always obliging when it comes to direct evidence. For some of the earliest synapsids, bones make up the majority of their fossil record. But there are a few scattered clues about what they looked like on the outside.

In 1975, paleontologist Robert Reisz described the anatomy of Archaeothyris – a small, lizard-like synapsid that lived about 306 million years ago. Among the preserved material were belly scales. These weren’t the same as the beta-keratin-coated scales of today’s lizards and snakes, but analogous structures. Basically, pseudoscales. And additional finds, including body impressions, suggest that the bellies of early synapsids were dotted with these same structures. What the rest of their bodies looked like, we don’t know.

A cast of the skull of Estemmenosuchus, a synapsid that may have had smooth skin. Photo by Brian Switek.
A cast of the skull of Estemmenosuchus, a synapsid that may have had smooth skin, at the Arizona Museum of Natural History. Photo by Brian Switek.

During their evolution, protomammals eventually lost their pseudoscales. In 1968 the Russian paleontologist P. Chudinov described skin impressions for a wildly-ornamented protomammal named Estemmenosuchus. This synapsid didn’t appear to have any scales, but instead smooth “glandular” skin similar to that of a frog or a hairless mammal.

Unfortunately, though, the original material Chudinov studied hasn’t been reevaluated in some time, and similar specimens from other animals have been lacking. Multiple fossil finds are needed to clear up the ancestral protomammal state and when hair started to become a synapsid feature. Yet, even as they await additional direct evidence, paleontologists can look to other hints to see when protomammals started to get fluffy.

Hair isn’t just for insulation or display. Hairs are also an essential part of the mammalian sense of touch, and having so many little sensors over the body certainly changed the brains of the first synapsids to bear fur. By combining direct evidence with the shape of prehistoric brains, University of Texas paleontologist Tim Rowe and colleagues were able to make the case that hair was already an essential mammal trait by the time the group scurried out onto the evolutionary stage.

Photo by Brian Switek.
A restoration of Morganucodon at the National Museum of Natural History. Photo by Brian Switek.
A restoration of Morganucodon at the National Museum of Natural History.

Drawing from brain endocasts created from CT scans of the 200-195 million year old mammals Morganucodon and Hadrocodium, Rowe and colleagues were able to track how mammal brains evolved. Among other alterations, the researchers saw that the early mammals had an expanded neocortex – part of the brain dominated by the sense of touch. Paired with the fact that other Mesozoic mammals and their almost-mammal relatives, such as Castorocauda, definitely had thick fur, the enlargement of the neocortex in Morganucodon and Hadrocodium hints that fur had originated sometime earlier so that it could play such an important sensory role by the time the first mammals evolved.

Whether or not prehistoric bone bears other hallmarks of hair requires further research to discern. Associating external integument to internal scaffold is no simple task. But the fossil record is full of surprises and no doubt contains additional evidence to explain how we and our mammalian kin came to be, including the long backstory behind why I have a fluffy companion who is fascinated by the ritualized removal of the hair evolution gave me.


Blackburn, D., Hayssen, V., Murphy, C. 1989. The origins of lactation and the evolution of milk: a review with new hypotheses. Mammal Review. 19, 1: 1-26.

Carroll, R. 1969. Problems of the origin of reptiles. Biological Reviews. 44, 3: 393-431.

Kielan-Jaworowska, Z., Hurum, J. 2006. Limb posture in early mammals: sprawling or parasagittal. Acta Palaeontologica Polonica. 51, 3: 393-406

Ji, Q., Luo, Z-X., Yuan, C-X.,Wible, J., Zhang, J-P., Georgi, J. 2002. The earliest known eutherian mammal. Nature 416: 816–822. doi:10.1038/416816a

Niedźwiedzki, G., Bojanowski, M. 2012. A supposed eupelycosaur body impression from the Early Permian of the Intra-Sudetic Basin, Poland. Ichnos. 19, 3: 150-155

Reisz, R. 1975. Pennsylvanian pelycosaurs from Linton, Ohio and Nýřany, Czechoslovakia. Journal of Paleontology. 49, 3: 522-527.

Rowe, T., Macrini, T., Luo, Z. 2011. Fossil evidence on the origin of the mammalian brain. Science. 332: 955-957.

Zhou, C., Wu, S., Martin, T., Luo, Z. 2013. A Jurassic mammaliaform and the earliest mammalian evolutionary adaptations. Nature. 500: 163-167. doi: 10.1038/nature12429

6 thoughts on “Getting to the Root of Fur

  1. This is a subject I’ve often wondered about. I imagine that protofeathers, which are relatively large structures attached to reasonably large animals, have a much easier time fossilizing than protohairs, which probably evolved around the same time mammaliform synapsids went through a period of miniaturization. But as you photo of Eomaia demonstrates, hair doesn’t fossilize the same way feathers do. No individual traces there–just a blob of black carbon around the body.

    I’ve always liked the idea of basal synapsids having skin like naked mole-rats–something not reptilian but also not entirely modern, either. I imagine that before hair started growing, changes in the skin had to happen first.

  2. A total speculation that I have had with John Merck: the traditional interpretation has been that whiskers (vibrissae) are hair modified for sensory data. But what if vibrissae evolved first? Given that burrowing goes to nearly the base of Synapsida (so far as we can tell), a set of such detectors would be quite useful. Then, at some later phase, the capacity to grow long integumentary structures gets co-opted for insulation.

  3. There has been at least some speculation that various pre-mammaliamorph cynodonts (maybe Thrinaxodon?) had vibrissae. (Evidence was, i.i.r.c., pores in the bones of the snout that might have been nutrient foramina.)

    One of my favorite paleontological “methods” stories is how we got confirmation that (maybe early Cenozoic, rather than Mesozoic) Multituberculates were furry: coprolites were found containing Multit. bones and also impressions — detailed enough to reveal the “scale” structure of individual hairs — of hairs! (1980 or 1990s, I think a short communication or “letter” in “Nature.”)

  4. Thank you for this; hair before mammals is something I’ve long been interested in. I recently asked Darren Naish what the chances of furry Dimetrodon were, and he was skeptical, though I can’t remember his reasons, through lack of a clear idea of the names of geological periods in my head.

  5. This has almost certainly been addressed in the literature – but are the genes that regulate feathers in birds and scutes different than the genes that regulate hair? Is there a chance these structures originated earlier and were perhaps present in the LCA for diapsids and synapsids?

  6. Here’s the paper Allen mentioned:

    Meng, Jin, and André R. Wyss, 1997. Multituberculate and other mammal hair recovered from Paleogene excreta. Nature 385, pp. 712-714.

    Neat stuff. By focusing on coprolites and regurgitolites that contained skeletal remains from only one species, the authors determined they had hair traces that could be assigned to four Paleocene mammals, including the multituberculate Lamdopsalis.

    As near as I can find, the idea of whiskered therapsids based on “snout pores” was first proposed by D.M.S. Watson in 1931 for the therocephalian Ericiolacerta.

Leave a Reply

Your email address will not be published. Required fields are marked *