Mammals in the shadows

Foundations

The story you were told

Mammals were small, timid, primitive creatures that scurried underfoot for a hundred and sixty million years, hiding in burrows, waiting for the dinosaurs to be removed so that the age of mammals could begin.

Three things in that sentence are wrong: primitive, waiting, and the implication that mammals were next in line.

Not new, and not late

Mammals do not follow dinosaurs. They arrive at the same time. Both groups originate in the Late Triassic, around 225 million years ago, and both are survivors of the same catastrophe.

Before the dinosaurs, the dominant land vertebrates were synapsids, the lineage mammals belong to. Some were large, some were formidable predators, and the group ran terrestrial ecosystems for tens of millions of years. Then the end-Permian extinction of lesson 5.7 removed almost all of them. In the world that followed, the archosaurs, which became dinosaurs and crocodiles, took the large-bodied roles, and the surviving synapsids got smaller.

So mammals did not start small out of inadequacy. They started small because the previous extinction had reset the board, and a different group won the large-animal niches.

Mammals then coexisted with dinosaurs for about 160 million years. They have existed without them for 66. Most of mammalian history happened in the shadow.

Not primitive

The stereotype says generalised insectivores. The fossils, especially from China over the past twenty-five years, say something else.

Castorocauda, from about 164 million years ago, had a broad flattened tail with scales, webbed feet, seal-like teeth for catching fish, and preserved fur. It was semi-aquatic in the Middle Jurassic, more than a hundred million years before beavers or otters.

Volaticotherium had a gliding membrane. Fruitafossor had forelimbs built for digging into insect nests, with teeth reduced the way anteaters’ are. Others were arboreal, others were specialised burrowers.

And Repenomamus, from the Early Cretaceous of China, weighed something like twelve to fourteen kilograms, about the size of a badger. One specimen was preserved with the articulated remains of a juvenile Psittacosaurus in its abdominal cavity. It had eaten a dinosaur.

By the Cretaceous, mammals were swimming, gliding, burrowing, digging and, at least occasionally, hunting the neighbours. What they were not doing was getting big. Something kept a ceiling on body size for a hundred and sixty million years, and the most likely something is that the large niches were occupied by animals extremely good at holding them.

What they were building instead

If mammals could not get large, what did selection do with them? It refined them, in ways that turned out to matter enormously later.

Teeth that chew. A reptile’s teeth are mostly interchangeable pegs for grabbing and swallowing. Mammals evolved incisors, canines, premolars and molars, with upper and lower molars that occlude precisely, so food is processed before swallowing. That extracts far more energy per mouthful, which is exactly what an expensive metabolism needs. It is also why palaeontologists can identify a Mesozoic mammal species from a single tooth, which is fortunate, because teeth are usually all that survives.

Hearing. In mammal ancestors the jaw joint was formed by two bones that, over a long sequence of well-documented fossils, shrank, detached and migrated into the middle ear, becoming the malleus and incus. Mammals hear a wider range of frequencies than most reptiles because of it. Every mammal embryo still repeats the transition: the bones begin near the jaw and move.

Warmth. Endothermy costs roughly ten times as much food as being a comparable-sized reptile. It buys activity at night, in cold weather, and sustained rather than burst effort.

Night vision, and its price. Most mammals have poorer colour vision than most birds and reptiles, and much better hearing and smell. The usual explanation is a long nocturnal phase in mammal ancestry, during which colour was useless and the other senses were not. Primates later recovered a third colour channel, which is why you can see a red fruit and a dog cannot.

The day everything changed

Sixty-six million years ago the asteroid of lesson 5.7 struck. Every non-avian dinosaur died. So did the flying pterosaurs, the marine reptiles, and about three quarters of all species.

Mammals were not spared. Most mammal species died too. What survived skewed towards small, generalist burrowers that could eat seeds and insects rather than fresh vegetation, and that could shelter underground through a collapse in photosynthesis. Being unimportant was, for once, a survival strategy.

Then the niches were empty.

What happened next, in detail

For a long time the recovery was a blur. A fossil sequence from the Corral Bluffs area in Colorado, published in 2019, resolved the first million years in unusual detail, because the site preserves mammal skulls, plant fossils and dateable ash together.

The pattern is stepwise rather than smooth. Immediately after the boundary, mammals are small, at most a few hundred grams. Within about a hundred thousand years the maximum roughly doubles. By three hundred thousand years it doubles again. By seven hundred thousand years there are mammals of around fifty kilograms, and the steps track the appearance of new plant groups, including legumes, which offer protein-rich food.

Within ten million years there were mammals over a hundred kilograms. Within twenty, there were whales.

Nothing about mammals improved at the boundary. The teeth, the ears, the metabolism, the parental care were all in place, some for over a hundred million years. What changed was who else was in the room. That is ecological release, and it is one of the clearest demonstrations in the fossil record that success is contingent rather than earned.

And they did not inherit the Earth

Two corrections to the triumphant version.

First, the dinosaurs did not all die. Birds are dinosaurs, they survived, and there are roughly eleven thousand bird species against about sixty-five hundred mammal species. By that measure the dinosaurs are still ahead.

Second, mammals did not immediately dominate. For several million years afterwards, some of the largest land predators were snakes and crocodilians. Titanoboa, from Colombia around fifty-eight million years ago, was a snake estimated at twelve metres and over a tonne, in a world whose mammals were still modest.

The age of mammals took time, and it was not inevitable at any point in it.

Deeper

Where the boundary of mammal lies

“Mammal” is defined by a suite of characters, chiefly the dentary-squamosal jaw joint, and the fossil record contains a graded series of forms acquiring them in different orders. This is why the literature uses “mammaliaform” for the wider group. The gradation is itself the evidence: an unambiguous boundary would be more surprising than the mosaic that is actually found.

The molecular and morphological disagreement

Molecular clock analyses of placental mammals generally recover deep divergences among the major orders well before the boundary, with a long fuse of lineage splitting followed by post-extinction morphological radiation. Large morphological analyses, including the reconstruction of a hypothetical placental ancestor from over four thousand characters, instead place both the divergences and the radiation after the boundary. The disagreement has narrowed but not closed; current syntheses generally accept some pre-boundary lineage divergence with the bulk of ecological and morphological diversification afterwards.

Multituberculates as a counterexample

Analysis of dental complexity in multituberculates shows their diversification and dietary expansion beginning tens of millions of years before the extinction, correlating with the radiation of flowering plants rather than with the removal of dinosaurs. This matters because it demonstrates that Mesozoic mammals were responding to ecological opportunities that had nothing to do with dinosaurs, and it argues against a simple suppression model in which nothing could happen until the competition left.

Reading the Corral Bluffs record

The site’s value lies in the co-occurrence of vertebrate fossils, palynology and radiometrically dateable horizons within a continuous continental section, which is rare. Concretions preserve skulls rather than isolated teeth, allowing body mass estimates and taxonomic identification. The correlation of mammal body-size steps with the arrival of particular plant lineages supports a bottom-up explanation in which recovering plant communities set the pace of mammalian recovery.

The energetic argument for the size ceiling

Endothermy sets a high minimum food requirement, and juveniles of large dinosaurs occupied the small and medium body-size classes that mammals might otherwise have expanded into. This has been proposed as a mechanism for the Mesozoic size ceiling: rather than direct predation, mammals faced competition across the size range from young individuals of species whose adults were enormous. The hypothesis is plausible and difficult to test directly.

Frontier

When did placentals diversify

The remaining gap between molecular and morphological timescales is the main open question. Improved fossil calibration, better models of rate variation across the tree, and new Palaeocene fossils are all narrowing it, but no consensus timescale yet satisfies both lines of evidence.

How warm-blooded, and when

Reconstructing the origin of mammalian endothermy is difficult because the trait does not fossilise. Approaches include bone histology, nasal turbinate structure and, more recently, inner-ear canal dimensions as a proxy for body temperature. Estimates for when full endothermy appeared span tens of millions of years, and the methods do not yet agree.

Transcript and shot list

Target runtime about 8 minutes. Built as three corrections to a familiar story: not primitive, not waiting, and not first in line. Shot types: ANIM is rendered from code with Manim, AI is a generated shot from Higgsfield with the shared style reference, REAL is agency imagery with credit, CAM is the narrator.

[00:00] S01 · AI · A small furred animal watching from under a fern as a large dinosaur foot passes; night, low light.

The story goes like this. Mammals were small, timid, primitive things that scurried underfoot for a hundred and sixty million years, hiding, waiting for the dinosaurs to be removed so the age of mammals could begin. Three things in that sentence are wrong. Primitive. Waiting. And the idea that mammals were next in line.

[00:35] S02 · ANIM · A timeline: synapsids dominant before 252 Ma; the end-Permian removing most of them; archosaurs taking the large niches; mammals and dinosaurs both appearing at 225 Ma.

Start with the timing, because mammals do not follow dinosaurs. They arrive at the same time. Both groups originate in the Late Triassic, about two hundred and twenty-five million years ago, and both are survivors of the same catastrophe. Before the dinosaurs, the dominant land animals were synapsids, which is the lineage mammals belong to. Some were large. Some were formidable predators. Then the end-Permian extinction from the last lesson removed almost all of them, and in the world that followed, the archosaurs took the big-animal roles and the surviving synapsids got smaller. Mammals did not start small out of inadequacy. They started small because the board had been reset and somebody else won.

[01:25] S03 · ANIM · A bar showing 160 million years alongside dinosaurs and 66 million without; the first bar dominating.

And then they coexisted with dinosaurs for about a hundred and sixty million years. They have existed without them for sixty-six. Most of mammal history happened in the shadow.

[01:45] S04 · ANIM · Three fossils in turn: Castorocauda with flattened scaly tail, webbed feet and preserved fur; Volaticotherium with a gliding membrane; Fruitafossor with digging forelimbs and reduced teeth.

Now the second correction, and the fossils that did it mostly came out of China in the past twenty-five years. Castorocauda, a hundred and sixty-four million years old, had a broad flattened scaly tail, webbed feet, teeth for catching fish, and preserved fur. Semi-aquatic in the Middle Jurassic, more than a hundred million years before beavers. Volaticotherium had a gliding membrane. Fruitafossor had forelimbs built for tearing into insect nests, with teeth reduced the way an anteater’s are. Others were tree-dwellers. Others were specialised burrowers.

[02:30] S05 · REAL · The Repenomamus specimen with juvenile Psittacosaurus remains preserved in the abdominal cavity (museum image; confirm credit line).

And then there is this one. Repenomamus, Early Cretaceous, about twelve to fourteen kilograms, roughly badger-sized. This specimen was preserved with the articulated remains of a young Psittacosaurus in its abdomen. A mammal that ate a dinosaur.

[02:55] S06 · ANIM · A size chart: most Mesozoic mammals shrew to rat sized, with a hard ceiling around a metre, held flat across 160 million years.

So by the Cretaceous, mammals were swimming, gliding, burrowing, digging and occasionally hunting the neighbours. What they were not doing was getting big. Something held a ceiling on body size for a hundred and sixty million years, and the likeliest something is that the large niches were occupied by animals that were extremely good at keeping them. Including, probably, the juveniles of enormous species, which filled the small and medium sizes too.

[03:30] S07 · ANIM · A reptile jaw of interchangeable pegs beside a mammal jaw with incisors, canines, premolars and molars occluding precisely; food being processed rather than swallowed whole.

So what did selection do with them instead? It refined them, in ways that mattered enormously later. Teeth first. A reptile’s teeth are mostly interchangeable pegs for grabbing and swallowing. Mammals evolved four kinds in one mouth, with uppers and lowers that meet precisely, so food gets processed before it is swallowed. That extracts far more energy per mouthful, which is exactly what an expensive metabolism needs. It is also why a palaeontologist can name a Mesozoic mammal from a single tooth, which is lucky, because teeth are usually all that is left.

[04:15] S08 · ANIM · A sequence of fossil jaws: two bones forming the jaw joint shrinking, detaching and migrating into the middle ear; then a mammal embryo repeating the same migration.

Then hearing. In mammal ancestors, the jaw joint was made by two bones. Across a long series of well-documented fossils, those bones shrank, detached and migrated into the middle ear, where they became two of the three bones you hear with. Every mammal embryo still does it: they start near the jaw and move. That is why mammals hear a wider range of frequencies than most reptiles.

[04:50] S09 · ANIM · An energy comparison: an endotherm eating roughly ten times a comparable ectotherm; then night-time and cold-weather activity unlocked; then a colour-vision comparison, birds and reptiles with four channels, most mammals with two, primates recovering a third.

And warmth. Endothermy costs something like ten times as much food as being a comparable reptile. What it buys is activity at night, in the cold, and sustained effort instead of bursts. Which brings up a price you are still paying. Most mammals have worse colour vision than most birds and reptiles, and far better hearing and smell. The usual explanation is a long stretch of night-time living, when colour was useless and the other senses were not. Primates later got a third colour channel back, which is why you can see a red fruit on a green tree and your dog cannot.

[05:35] S10 · ANIM · The impact; non-avian dinosaurs, pterosaurs and marine reptiles removed; then the mammal survivors: small, burrowing, seed and insect eaters.

Then, sixty-six million years ago, the asteroid. Every non-avian dinosaur died. So did the pterosaurs, the marine reptiles, and about three quarters of all species. Mammals were not spared, and most mammal species died too. What got through skewed small, burrowing, and able to live on seeds and insects rather than fresh vegetation, in a world where photosynthesis had collapsed. Being unimportant turned out to be a survival strategy.

[06:15] S11 · REAL · The Corral Bluffs section in Colorado (research image; confirm credit line). S12 · ANIM · Stepwise body-size increase over the first million years: a few hundred grams, doubling by 100 kyr, again by 300 kyr, around 50 kg by 700 kyr, correlated with new plant groups appearing.

For a long time what happened next was a blur. Then a site in Colorado, published in 2019, resolved the first million years, because it preserves mammal skulls, plant fossils and dateable ash together. And the pattern is stepwise, not smooth. Straight after the boundary, mammals are at most a few hundred grams. Within about a hundred thousand years the maximum doubles. By three hundred thousand it doubles again. By seven hundred thousand there are mammals around fifty kilograms. And each step tracks the arrival of new plant groups, including legumes, which are protein-rich. Within ten million years, mammals over a hundred kilograms. Within twenty, whales.

[07:05] S13 · ANIM · A checklist of mammalian traits, each with a date long before the boundary: teeth, ears, endothermy, parental care; then a single change at the boundary: the competition removed.

And here is the point. Nothing about mammals improved at the boundary. The teeth, the ears, the metabolism, the parental care were all in place, some for over a hundred million years. What changed was who else was in the room. That is called ecological release, and it is one of the clearest demonstrations in the entire fossil record that success is contingent rather than earned.

[07:35] S14 · ANIM · A species count: about 11,000 birds against about 6,500 mammals. Then Titanoboa at twelve metres beside a modest Palaeocene mammal.

Two final corrections to the triumphant version. The dinosaurs did not all die. Birds are dinosaurs, they survived, and there are around eleven thousand bird species against about six and a half thousand mammal species. By that measure they are still ahead. And mammals did not take over immediately either. For several million years afterwards, some of the biggest land predators were snakes and crocodilians. Titanoboa, from Colombia, was a snake of about twelve metres and over a tonne, in a world where mammals were still modest.

[08:10] S15 · AI into ANIM · A small modern mammal at dusk; pull back to the calendar, the last hours of 31 December approaching.

So the age of mammals took time, and at no point was it inevitable. Which is worth remembering, because everything in the next chapter, including you, descends from an animal that spent a hundred and sixty million years being no threat to anybody. Next chapter: what happened when one of those animals started learning from the others.

Shot list

Two AI shots, eleven code-rendered shots, two research images. Estimated Higgsfield use: 3 to 6 generations including retries.