We Owe Our Entire Existence to a Bunch of Long-Necked Mouth-Breathers

Elsa Panciroli Looks Deep Into the Fossil Gap

At the close of the Devonian period (about 359 million years ago) and the beginning of the Carboniferous, there is a gap in the fossil record. It has a name: Romer’s Gap, after Alfred Sherwood Romer, a paleontologist from the United States fascinated with the evolution of vertebrates. Romer was particularly obsessed with the “fish to tetrapod transition”: the transformative evolutionary journey of one group of lobe-finned fish to become the progenitors of all four-limbed animals. His stunningly detailed books—published between 1930 and 1970—cover the structure of all living vertebrates and are so meticulous and well-illustrated that they remain vital texts to this day.

Romer noticed that there was a point in the tetrapodomorph fossil record where we simply didn’t have any fossil remains to tell their story—a gap in our knowledge. This absence of fossils was named after him by later researchers. From 375 to 360 million years ago, at the end of the Devonian, two mass extinctions decimated life on Earth. In the following 15 million years (the start of the Carboniferous) the fossil record goes strangely quiet. It has been suggested that an unusually low concentration of oxygen in Earth’s atmosphere at that time may have reduced rates of fossilization, but there may also have simply been fewer animals around to become preserved. This is Romer’s Gap. After it, tetrapods are diverse landlubbers able to support their own weight out of water without batting an eyelid.

For a long time we didn’t know how this water-to-land transition had happened, but recently the gap has started to be filled in. Many of the newest finds come from the borders of Scotland, where scientists working as part of the TW:eed project (Tetrapod World: early evolution and diversification) have been digging up new early terrestrial vertebrates. The work was driven by pioneering paleontologist Jenny Clack and colleagues, and among their discoveries was Aytonerpeton microps, “the creeping one from Ayton with the small face” (Ayton being the Scottish parish in which she was found).

With nippy teeth and a chin full of dimples, Aytonerpeton—or “Tiny” as she was christened by the team—has a skull just five centimeters long. In other words she was small, certainly when compared with some of the other tetrapods from the time period. Her rows of sharp little teeth probably snatched up the copious invertebrates that surrounded her. The success of arthropods would have provided a tempting menu for early tetrapods venturing from the waterways.

It is, of course, easy to slip into language that suggests evolution had a goal in mind when the first backboned animals took to the land. Of course, there is no end-goal to evolution’s journey. It is random, the route forged by happenstance. Divergences occur when unexpected mutations or behaviors turn out to be useful at any moment in time. This uncaring randomness frightens some people—as evidenced by the anti-evolutionary arguments given by some religious devotees. But it is the serendipity of it all that makes evolution so utterly glorious.

Our tetrapod ancestors didn’t evolve any of their adaptations for life on land, but the adaptations they already had turned out to be useful platforms from which to launch themselves out there. This phenomenon is called exaptation by scientists, and it is an important feature of how evolution works. A trait that evolves for one purpose is retooled to serve another. It wasn’t remotely inevitable that Tiny would stride on her way through the Carboniferous Scottish undergrowth, let alone that a strange group of animals that puts all its squishy bits on the outside (crazy idea, just ask an arthropod) would ever amount to anything.

And so this foray through the swamp-forests has brought us to our origins—arguably one of many origins or forks in the trunk of the vertebrate tree of life. The creature you met earlier in the wet undergrowth is Westlothiana, named for the region in which the fossil was discovered, West Lothian in the Scottish Borders. We could trace our evolutionary history back further, slithering into the brackish estuaries in search of the common ancestor between lobe-finned sarcopterygian fish and their ray-finned sisters (the actinopterygians, which include the rest of the bony fish on the planet).

Or indeed further still, to the insanity of the Ediacaran where the first multi-cellular organisms included flattened blobs that looked like hernia support cushions. But Westlothiana is as good a place as any to begin our mammal tale. It was not only one of the first animals on land, but features in the skeleton also tell us that it is related to the group we belong to, the amniotes.

Of course, there is no end-goal to evolution’s journey. It is random, the route forged by happenstance.

As the Carboniferous period continued, disaster struck those sultry lycopsid swamp-forests. An event dubbed the Carboniferous Rainforest Collapse felled the wetland trees. It’s unclear what caused the collapse, but evidence suggests it was a change in climate, driven by the eruption of volcanoes in what is now northwest Europe. There were still forests, but they were different after that: fragmented, and filled with gymnosperms. Their cones and seeds are familiar to us now as conifer trees, the unusual and endangered gingko, and the palm-like cycads much beloved of Victorian botanical collections. Beneath their branches those ambiguous tetrapodomorphs finally came into focus as full-blown, four-footed land animals. There was no doubt any more, these were the progenitors of the two great lineages of vertebrate life on land: amniotes and anamniotes.

As the climate altered, one group of animals retained many of the prototype evolutionary outlines of the first tetrapodomorphs. They continued to rely on water for reproduction, and as a safe place to keep their eggs moist and provide them with oxygen. This group is called the anamniotes, and their need for moisture meant the drier world of the Late Carboniferous was a challenging place. But they hung on: this group survives today in the form of frogs, salamanders and caecilians.

For the rest of the tetrapods, however, two novel adaptations meant they were able to break their reliance on water, severing ties to the depths completely. The first adaptation, which was long given precedence, also gave them their name: the amniotes. It comes from the fluid-filled membrane called an amnion that encloses their developing embryos. The amnion evolved from the jelly-like outer layers of egg you can see around the frogspawn in your pond. In amphibians, this jelly lets waste and gases pass between the spawn and water, but in amniotes the amnion performs the same functions out of water. With two other membranes, the chorion and allantois, the amnion forms a cushion of membranes enclosing the developing embryo in its own portable pond: the amniotic fluid. All of this is placed inside an egg-shell (and much later, in some animals, occurs inside the mother’s body). It is a feature amphibian and fish eggs lack. Thanks to this eggy innovation, the amniotes could raise their young away from their riparian homelands.

But there is another change in the amniote body that was arguably even more crucial to their transition. As vertebrates left the water, they took advantage of the fact that air has up to 30 times more available oxygen than water. Fish have to flush huge volumes of water over their gills to breathe, but on dry land it is thought that the first tetrapods employed something called buccal pumping. It’s a kind of bellows mechanism, created by lifting and dropping the oral cavity. This also explains why a lot of these first tetrapods have wide flat heads that look a little as if they’ve been stepped on. The shape creates a bigger oral cavity. Basically the first tetrapods were a bunch of mouth-breathers.

Basically the first tetrapods were a bunch of mouth-breathers.

Amniotes, however, stuck their necks out, quite literally. Examining the differences in rib mobility in these animals, paleontologists Christine Janis and Julia Keller noticed in 2001 that the ribs became more mobile in the first amniotes. Many of them had narrower heads and longer necks. They realized these changes were all connected, and reflected a change in breathing mode, from mouth-based buccal pumping to costal breathing, using chest muscles. This would have a profound impact on the history of life on Earth.

Not only did it mean that they were more efficient breathers, but amniotes were no longer using their stout ribs to maintain their posture. They could stand more upright and elongate their necks—something that would have hindered a buccal pumper from getting enough air into the lungs. With the mouth no longer employed in ventilation, amniotes were free to re-use some of their skull and jaw muscles for new types of feeding. For the first time they could eat vegetation, a diet that required skilled biting at the front of the jaw. With a chesty sigh, amniotes strode away from the water’s edge, laid their eggs and took on the world.

At most points in our evolutionary history all we have to go on to understand evolution are the bones. It is clear from the fossil record that amniotes are one of a complex array of animals living alongside one another at this time in Earth’s history, all experimenting with life on land. Gradually, this lineage accumulated changes in the skeleton that helped them survive out of water: stronger vertebrae, larger limbs, the restructuring of the ankle to support their body weight and optimize foot movements for walking. Paleontologists closely examine and chart such changes, observing which ones provide the tell-tale mark of a distinct group. Such features are called synapomorphies, and they distinguish one group from another.

The synapomorphies of amniotes require a keen eye. In the skull, they include the arrangement of the bones: a bone called the frontal widened and formed part of the orbit of the eye. Inside the mouth, part of the palate, the roof of the mouth, had a flange covered in teeth reaching towards the back of the throat. In the shoulder, the bones become more complicated (including the development of two coracoids, which in mammals are now part of our shoulder blades), likely linked to the change in limb-use with life on land. Where the shoulder bones grew more complex, the ankle and wrist simplified, and multiple bones fused to form the astragalus (part of the ankle).

Other features paint an intriguing picture of these animals. For example they had no specialized ear for hearing out of the water, so theirs was a world of heady vibrations. In some early amniotes the structure of the skull appears flexible, and lacks the kind of muscle attachments we associate with precise biting, suggesting they weren’t able to eat anything that couldn’t be swallowed more or less whole. Some may have continued to feed in water. There are also aspects of their biology that remain hidden from us, such as the keratinization of their skin, or how the amniotic egg evolved. These are soft-tissue features, and unlike bones they seldom leave any trace in the rock record. We can infer their presence through common ancestry, but for most soft-tissue features there are no fossils to prove their presence or absence conclusively.

We now come to the final fork in the tetrapod tale—after this, we hit the mammal highway. Around 300 million years ago, when the lycopsids were falling and the seeds of the first gymnosperms sprouted their claim on the forest floor, our own lineage had already parted ways with our cousins, the reptiles. It is a common misconception that mammals evolved from reptiles. We now know this is not remotely true. But mammals and reptiles do share a common ancestor. This first amniote tetrapod was neither mammal nor reptile—neither of those groups had evolved yet. In the Carboniferous, our last common ancestor with turtles, crocodiles, dinosaurs, birds and lizards said goodbye, and set off into the evolutionary sunset.

The amniote tetrapods were cleaved into two mighty lineages: Synapsida and Sauropsida. These Romulus and Remus groups looked alike back then, you would struggle to tell them apart at a glance. Traditionally though, we have recognized them in the fossil record by one feature in particular: the number of holes in their heads.

The synapsids include us and all our mammal brothers and sisters, as well as an incredible host of extinct creatures that we will meet in chapters to come. The sauropsids (reptiles), on the other hand, are arguably the more successful of the great tetrapod houses—if you discount the so-called “success” of humankind. From equally humble beginnings Sauropsida has given rise to a startling diversity of forms, from turtles to pterosaurs, lizards and tuatara, ichthyosaurs to crocodiles, and of course dinosaurs—who have received more than their fair share of attention. Birds, the living descendants of dinosaurs, are twice as speciose as mammals, setting us all a-twitch with their diversity. But there are plenty of books outlining the reptile evolutionary journey, especially in the Mesozoic. I needn’t tire you with that tale here. Let’s focus on their sister-group, for they are the protagonists of this particular evolutionary tale. They are the synapsids.


There are currently several strong candidates for the earliest synapsid, the lineage of four-limbed animals that includes mammals and their relatives. All of them have been found in the Carboniferous rocks of Nova Scotia. It would seem our most ancient synapsid ancestors were Canadians.

The latest fossil believed to be from our lineage is Asaphestera. The next sounds more like a preventative medication than an animal: Protoclepsydrops. Then there is Archaeothyris (perhaps some kind of throat malady?) and Echinerpeton (Scottish slang, for something naughty no doubt). The names hint at the difficulty their scattered and fragmentary fossils pose for the scientists who study them.

Asaphestera has only recently joined our mammal-line ranks. The name means “less distinct one,” which is appropriate because it was so indistinct, it was originally lumped in with the bones of several unrelated animals. In May 2020 a study came out by a group of scientists from Canada and Germany who had taken a second look at the specimen. They realized the bones of its small wide skull were arranged in a pattern associated with the earliest synapsids. If this is the case, Asaphestera is one of the oldest stem-mammals (as in the stem of the family tree).

From the same rocks comes Protoclepsydrops, which means “first Clepsydrops.” The slightly younger fossil of Clepsydrops is yet another Carboniferous Canadian whose vertebrae looked like an hourglass, or Greek klepsydra. Although it is clear from its skeleton that Clepsydrops is a synapsid, opinion remains divided over Protoclepsydrops. Its fossils comprise only a few vertebrae and upper arm bones, but they seem to resemble other early synapsids in shape. So this cough-medicine animal may indeed be among the very first of our line.

Archaeothyris means “ancient window” in Greek, and it’s a poetic choice. This fossil provides not only a glimpse of the ancient amniote past, but outlines a defining feature of our synapsid family. We, along with our shared relatives all the way back to the Carboniferous, have a single hole in each side of our skulls called a temporal fenestra. In anatomy, a fenestra (meaning “window” in Latin) refers to any hole in a bone, but this particular hole is pivotally different. It lies just behind the eye, and synapsids have just one of them on each side. Synapsida means “one arch,” because this single fenestra creates a single arch in the skull bones.

Archaeothyris means “ancient window” in Greek, and it’s a poetic choice. This fossil provides not only a glimpse of the ancient amniote past, but outlines a defining feature of our synapsid family.

Our temporal fenestra is rimmed by the temporal, squamosal and postorbital bones of the skull. You can feel your own fenestra by placing your fingers in the hollow behind your eye, above your cheekbone. Now clench and unclench your jaw, and you should feel the muscles running through your temporal fenestra. This opening provides attachment areas for the muscles that open and close the mouth, so it could be that different arrangements of holes in early tetrapods are linked to different ways of biting and feeding.

Recent research has shown that the pattern of how many holes there are in the skulls of reptile lineages is more complicated than traditionally thought, with fenestrae being acquired and lost again in multiple groups. Most reptiles are diapsids with two holes, but turtles are anapsids and have none. Early groups may have gained and lost one or more holes through their history. In Synapsida however, the single fenestra is almost unequivocal. The main split between synapsids and sauropsids is clear-cut and our single skull-hole on each side is a consistent defining feature. Archaeothyris’s “window” is a window into the skull and into our own past, a clear indicator that it belongs at the base of our tree.

Echinerpeton, on the other hand, rather unhelpfully means “spiny lizard.” Scientific names, once given, cannot be changed even if their meaning is subsequently proven incorrect or misleading. This poses a problem for modern paleontologists doing their best to make clear the crowbar separation between synapsids and their reptile cousins. Although Echinerpeton belongs to the mammal line, it is unfortunately named a lizard in perpetuity.

As well as scientific names, colloquial ones can be infuriatingly diehard. For synapsids, the once commonly used term for these ancestors of mammals is “mammal-like reptiles.” I cringe as I write it. Some of the best books on the origin of mammals use this, keeping the term alive in new generations of students and the public. It is a relic from an extinct terminology, but it’s easy to find yourself falling back on it as a familiar touchstone, which only keeps the misnomer going. In truth, we need the term “mammal-like reptiles” about as much as a hole in the head.

You might think I’m over-reacting, but I’m not. “Mammal-like reptiles” reflects a fundamental wrongness about where we and all our milky brothers and sisters come from. It’s the taxonomic equivalent of insulting your mother. The truth of mammal origins is so much more fantastic, and knowing it transforms how we see ourselves and the animals we share our planet with.

In the quest to understand how evolution works, mammals that we are, we have tended to search for mammalness in the fossil record, to trace our lineage back to its genesis. We are creatures that perceive time as a linear experience with a start, middle and end, and we structure our stories about the world accordingly. Our language is excellently utilized for delighting us with these tales, but it is also riddled with ambiguities that are less than ideal for scientific accuracy. This is why scientific language seems opaque to those who don’t habitually use it. A very specific terminology is needed to ensure every researcher is discussing exactly the same thing, and prevent misunderstanding. This can look like pedantry, but it is not. Attention to wording has a purpose, a clarity.

When it comes to evolution, it is really hard to use the correct words rather than the best narrative ones. We talk about animals that “turn into” or “become” other animals through the process of evolution—I’ll probably do it multiple times in this book. These are Just So stories. This happens because we view evolution from back to front, comparing everything to what lives today. This approach has its uses—paleontology is founded on comparative anatomy, and in later chapters we’ll see how such comparisons in the fields of biomechanics and ecology can tell us how extinct animals might have lived. But it also has the consequence that we see inevitability in what is actually just randomness. It means we miss important distinctions like the one we must make now, before we go any further, between the outward appearance of reptiles and their shared common ancestors with mammals.

Reconstructing Asaphestera, Protoclepsydrops, Archaeothyris or Echinerpeton, they would have looked rather like lizards. They were all small, not bigger than your forearm. Their legs stuck out to the side and their long pointy tails would have swayed and dragged as they waddled along in the characteristic “reptilian” wiggle (reptiles still move more like fish, with side-to-side motions of the body). There would have been no fur or feathers on these first synapsids, just a tough textured surface to the skin, locking all their precious moisture in. Their long lipless mouths were filled with rows of simple pointed teeth to crunch down on insects or fish.

A suite of subtle skeletal features define the earliest synapsids. Aside from the main synapomorphy of having one large hole in the skull on each side of the head, the bone that forms part of the back of the eye region became broader and tilted, and the septomaxilla bone in the nose enlarged. Outwardly however, they looked like reptiles. They probably acted like reptiles, and so did the first sauropsids that lived alongside them. But technically none of them are really reptiles.

Despite appearances, the early amniote tetrapods—the synapsids and sauropsids—were not “reptile-like.” Instead, we ought to say that reptiles today are “early amniote-tetrapod-like.” Most modern reptile groups (particularly lizards) have retained a lot of characteristics that are reminiscent of their forebears; they’ve not superficially changed that much, they’re just a bit retro. Meanwhile the mammals have changed radically and obviously (as have birds). This change is clear even to the casual observer. No one ever mistook a cow for a chameleon, or an eagle for an iguana. But one might easily mistake a cow’s early ancestor in the Carboniferous for an eagle’s early ancestor, because they all looked pretty similar. This is where a detailed understanding of anatomy is crucial. Anatomy is the only way to tell the difference between them and piece together how they emerged into distinct groups.

The repercussion of this similarity was that the first paleontologists and anatomists traced mammals back in time, and perceived their fossils as looking increasingly “reptilian.” They therefore assumed mammals evolved from a branch of reptiles, and so the term “mammal-like reptiles” was born. With more fossils and increasingly detailed analyses, we now know mammals did not evolve from reptiles at all. The reptile-ness of the early synapsids was just a hangover from the early amniote tetrapod body plan, which they share with reptiles.


Beasts Before Us

Excerpted from Beasts Before Us: The Untold Story of Mammal Origins and Evolution. Copyright (c) 2021 by Elsa Panciroli. Used with permission of the publisher, Bloomsbury Sigma, a division of Bloomsbury. All rights reserved.

Elsa Panciroli
Elsa Panciroli
Elsa Panciroli is a palaeontologist who studies the evolution and ecology of extinct animals. She is a researcher based at the University of Oxford and associate researcher at the National Museums of Scotland. Elsa is a keen science speaker and communicator. She has contributed to the Guardian and Biological Sciences Review, as well as to radio and podcast programmes such as Crowdscience, The John Beatty Show and Our Lives.

More Story
Wakanda Forever: How the Black Panther Salute Links the Black Community We catch it at about the 45-second mark of the second Black Panther trailer. Chadwick Boseman, aka T’Challa, aka the Black...