The Cambrian explosion

Foundations

The event, as usually told

Five hundred and thirty-eight million years ago, the story goes, the fossil record abruptly fills with animals. Before it, almost nothing. After it, arthropods, molluscs, worms, echinoderms, sponges, and the first chordates, which is the group you belong to. Almost every major body plan alive today, arriving within about twenty million years, and no fundamentally new one since.

That summary is accurate. It is also three different phenomena stacked on top of each other, and separating them is the whole lesson.

Phenomenon one: something real happened

This is not an artefact. Trace fossils are the cleanest evidence, because a burrow cannot be argued about the way a body can.

Before the boundary, burrows are shallow, simple, mostly horizontal scrapes. After it, they are deep, branching, vertical, and made by something that was clearly moving with purpose and going somewhere. Whatever else is true, animals began behaving differently, and behaviour does not fossilise selectively the way anatomy does.

Ecosystems changed too. Seafloors that had been sealed under microbial mats for a billion years were churned open by burrowers. That change, sometimes called the substrate revolution, altered sediment chemistry worldwide and destroyed the environment the previous inhabitants had been built for.

Phenomenon two: skeletons became fossilisable

Here is the part that makes the explosion look sharper than it was.

An animal without hard parts almost never fossilises. Jellyfish, flatworms, and most larvae leave essentially nothing. Shells, spines, teeth and carapaces fossilise readily. And across the Cambrian boundary, many unrelated lineages started building mineralised skeletons at roughly the same time.

So the record does not switch from no animals to animals. It switches from animals that could not fossilise to animals that could. Some of the explosion is an explosion in the availability of evidence.

We know this because of rare deposits where soft tissue survived. The Burgess Shale in British Columbia, about 508 million years old, and Chengjiang in Yunnan, older still, preserve guts, eyes, gills, antennae and limbs. These deposits show complete, sophisticated animals, and they also show many soft-bodied forms that ordinary rock would have lost entirely. Where preservation is good, the diversity is higher. That is a warning about every other interval.

Phenomenon three: the lineages are older than the fossils

The third complication comes from genetics. Count the differences between the genomes of living animal groups, calibrate the rate against fossils of known age, and you can estimate when lineages split.

Molecular clocks consistently put the divergence of major animal groups earlier than their first fossils, often well into the Ediacaran and sometimes earlier. The methods have real uncertainties, and different assumptions shift the answers by tens of millions of years, which is exactly why careful papers on this are titled around the limits of precision rather than the answer.

But the direction of the discrepancy is consistent, and the implication is uncomfortable for the simple story. Animals may have existed for tens of millions of years as small, soft, unfossilisable creatures before anything appeared in the rocks.

So what was the explosion?

Putting the three together: a long fuse and a short flare. Animal lineages probably originated well before the Cambrian and spent a long time small and invisible. Then, over roughly twenty million years, they became large, mineralised, mobile, ecologically diverse, and preservable, more or less at once.

Twenty million years is not sudden by any human standard. It is about eighty thousand human generations. It is sudden only against the three billion years of microbial life that preceded it, and that comparison is what the word “explosion” is really doing.

Why then?

There is no agreed answer, and the honest version of this lesson is that the candidate explanations are all partial.

Oxygen. Large active animals need a lot of it, and oxygen rose in the Neoproterozoic. But the timing is awkward: some analyses indicate levels adequate for early animals were reached well before the radiation, which means oxygen may have been a permission rather than a trigger.

Predation. The first clear evidence of one animal eating another appears here, including shells with holes bored through them. Predation sets off an arms race, and armour, eyes, speed and burrowing are all responses to being hunted. Once the first predator exists, everything else must change or die.

Eyes. Compound eyes appear in this interval and are astonishingly good almost immediately. A world in which animals can see each other is a different world from one in which they cannot.

Development. The genetic toolkit for building bodies, including the Hox genes that specify what goes where along an axis, is shared across animal phyla and predates the radiation. So this was not the invention of new genes so much as new ways of deploying existing ones. Why that flexibility was exploited then rather than earlier is unresolved.

Ecology itself. Once a few niches are filled, filling more becomes possible, because every new organism is an environment for another. Some models treat the radiation as largely self-propelling once it started, which shifts the question from what caused it to what permitted it.

The current synthesis combines these rather than choosing among them, which is less satisfying than a single cause and more likely to be true.

What it means for you

Your body plan, a bilaterally symmetric animal with a through-gut, a nerve cord along the back and a stiffening rod, is Cambrian. The chordate that first shows that arrangement was a few centimetres long, probably a poor swimmer, and not remotely the most impressive animal in its neighbourhood.

Five hundred million years later, nothing has been added at the level of body plan. Vertebrates elaborated the design considerably. Nobody invented a new one.

Deeper

Dating the boundary

The base of the Cambrian is defined by the first appearance of the trace fossil Treptichnus pedum at a reference section in Newfoundland, and is currently calibrated at 538.8 million years by uranium-lead dating of interbedded volcanic ash. Precision on Cambrian ages has improved substantially with chemical-abrasion methods, which is what allows the radiation to be constrained to a few tens of millions of years rather than a vague interval.

What the small shelly fauna show

Before the main radiation, the record contains an interval of millimetre-scale mineralised fragments: tubes, cones, spines and sclerites, many of unknown affinity because they are pieces rather than animals. Their appearance indicates that biomineralisation spread across lineages before large skeletal animals appear, and their diversity rises through the earliest Cambrian. They are the clearest evidence that skeletonisation was a phased process rather than a single event.

The Ediacaran problem

The Ediacaran biota comprises large, mostly sessile, soft-bodied organisms, some with fractal branching architectures unlike any living group. Interpretations range from stem-group animals to a separate multicellular experiment with no descendants. Dickinsonia was shown in 2018 to carry cholesteroids, which are characteristic of animals, supporting an animal affinity for at least that genus. Most Ediacaran forms disappear at or near the boundary, and whether that reflects extinction, ecological displacement by burrowers, or a change in preservation conditions is unsettled.

Why molecular clocks and fossils disagree

Molecular estimates depend on the assumed model of rate variation across the tree, the calibration fossils chosen, and the priors on those calibrations. Analyses that explore these choices systematically find the credible intervals for animal origins are wide, spanning tens to over a hundred million years, and that additional sequence data does not narrow them much because the limiting uncertainty is in the calibrations rather than the sequences. The disagreement with the fossil record is therefore real but its magnitude is poorly constrained.

The oxygen question, quantified

Modelling of oxygen requirements for early animals suggests small, thin-bodied forms could tolerate concentrations of a few percent of present atmospheric levels. Iron-speciation datasets analysed statistically have been read as indicating that late Proterozoic oxygenation was more limited than previously supposed, and that the deep ocean remained largely anoxic well into the Palaeozoic. If low oxygen persisted through the radiation, then oxygen cannot be the proximate trigger, though it may still set a ceiling on body size and activity.

Frontier

Cause, permission or contingency

Whether the radiation had a trigger at all is genuinely open. An alternative framing treats it as the expected consequence of crossing an ecological threshold, in which the first predators and the first burrowers generate the selection pressures that produce everything else, with no external cause required beyond whatever permitted the first steps.

Filling the Ediacaran gap

The most valuable evidence would be more exceptional deposits from the interval between 600 and 540 million years ago, where molecular clocks predict animals should exist and the record is nearly silent. Phosphatised embryos from the Doushantuo Formation in China are the closest available, and their interpretation is contested.

Transcript and shot list

Target runtime about 8 minutes. Built as a reframe: take the famous event apart into three separate phenomena, then reassemble it honestly. 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 · ANIM · A stratigraphic column: barren rock below, then a sudden crowd of animal fossils above the boundary; the word EXPLOSION stamped across it, then the word slowly questioned.

Five hundred and thirty-eight million years ago the fossil record fills up with animals. Below the line, almost nothing. Above it, arthropods, molluscs, worms, echinoderms, and the first chordates, which is your group. Almost every major body plan alive today, arriving within about twenty million years. And nothing fundamentally new since. That summary is accurate. It is also three completely different things stacked on top of each other, and taking them apart is the whole story.

[00:45] S02 · ANIM · Trace fossils before the boundary: shallow horizontal scrapes. After: deep branching vertical burrows.

Start with the part that is definitely real. Trace fossils, because a burrow cannot be argued with the way a body can. Before the boundary, burrows are shallow, simple, horizontal scrapes. After it, they are deep, branching, vertical, made by something moving with purpose and going somewhere. Whatever else is true, animals started behaving differently.

[01:20] S03 · ANIM · A seafloor sealed by a microbial mat for a billion years; burrowers churn it open; sediment chemistry changes.

And ecosystems changed with them. Seafloors that had been sealed under microbial mats for a billion years were churned open. That single change altered sediment chemistry worldwide, and it destroyed the world the previous inhabitants were built for.

[01:50] S04 · ANIM · A gallery of Cambrian body plans, each labelled; a chordate highlighted with a nerve cord, a through-gut and a stiffening rod; a counter of new phyla since: zero.

Here is what actually turns up. Arthropods. Molluscs. Echinoderms. Annelids. And chordates: bilaterally symmetric, with a through-gut, a nerve cord along the back, and a stiffening rod. That is your body plan. It shows up here, in something a few centimetres long that was probably a poor swimmer and not remotely the most impressive animal in the neighbourhood. Five hundred million years later, nothing new has been added at this level. Vertebrates elaborated the design enormously. Nobody invented another one.

[02:35] S05 · ANIM · A jellyfish, a flatworm and a larva decaying to nothing; a shell, a spine and a carapace fossilising cleanly; then many unrelated lineages starting to mineralise at once.

Now the part that makes the explosion look sharper than it was. An animal without hard parts almost never fossilises. Jellyfish, flatworms, larvae: essentially nothing survives. Shells, spines, teeth and carapaces fossilise easily. And across this boundary, many unrelated lineages started building mineralised skeletons at roughly the same time. So the record does not switch from no animals to animals. It switches from animals that could not fossilise to animals that could. Part of this explosion is an explosion in the availability of evidence.

[03:20] S06 · REAL · A Burgess Shale specimen showing preserved soft tissue: gut trace, eyes, limbs (Royal Ontario Museum or Smithsonian; confirm credit line). S07 · ANIM · A map marking Burgess Shale, British Columbia and Chengjiang, Yunnan.

We know that because of rare places where soft tissue survived. The Burgess Shale in British Columbia, about five hundred and eight million years old. Chengjiang in Yunnan, older still. These preserve guts, eyes, gills, antennae, legs. Complete, sophisticated animals. And crucially, lots of soft-bodied forms that ordinary rock would have lost completely. Where preservation is good, diversity is higher. That is a warning about every other interval in the record.

[04:05] S08 · ANIM · Genetic differences counted between living animal groups; a clock calibrated against dated fossils; divergence estimates landing well before the Cambrian; wide uncertainty bars.

And then genetics complicates it further. Count the differences between the genomes of living animal groups, calibrate the rate against fossils of known age, and you can estimate when the lineages split. Molecular clocks consistently put those splits earlier than the first fossils. Often well into the Ediacaran. The uncertainty bars are wide, and honest papers on this are titled around the limits of precision rather than the answer. But the direction is consistent. Animals may have existed for tens of millions of years as small, soft, unfossilisable things before they show up in the rocks.

[04:50] S09 · ANIM · A long fuse burning slowly through the Ediacaran, then a short bright flare across 20 million years; a scale bar: 80,000 human generations.

So put the three together and you get a long fuse and a short flare. Lineages that originated well before the Cambrian, spent a long time small and invisible, and then over roughly twenty million years became large, mineralised, mobile, diverse and preservable, more or less at once. And twenty million years is not sudden by any human measure. It is about eighty thousand human generations. It is sudden only against the three billion years of microbes that came before it. That comparison is what the word explosion is really doing.

[05:35] S10 · ANIM · Five candidate causes as cards, each with a tick and a caveat: oxygen, predation, eyes, developmental toolkit, ecology itself.

Why then? There is no agreed answer, and the honest version is that every candidate is partial. Oxygen: large active animals need a lot of it, and it rose beforehand. But some analyses say adequate levels were reached well before the radiation, which makes oxygen a permission rather than a trigger. Predation: the first clear evidence of one animal eating another shows up right here, including shells with holes bored through them. Once the first predator exists, everything else must change or die. Eyes: compound eyes appear in this interval and are astonishingly good almost immediately. A world where animals can see each other is a different world. Development: the genetic toolkit for building bodies is shared across phyla and predates the radiation, so this was not new genes, it was new ways of using old ones. And ecology itself: once a few niches are filled, every new animal is an environment for another one, so some models have the whole thing largely self-propelling once it starts.

[06:45] S11 · ANIM · The five cards combining rather than competing; a single-cause arrow crossed out.

The current synthesis combines these instead of choosing between them. Which is less satisfying than a single cause, and considerably more likely to be true.

[07:10] S12 · ANIM · The Ediacaran gap highlighted on a timeline between 600 and 540 Ma, marked as the place the answer would come from.

And if you want to know where the answer will come from, it is here. The interval between six hundred and five hundred and forty million years ago, where the clocks say animals should exist and the rocks are nearly silent. More deposits with exceptional preservation from that window would settle more arguments than anything else in this field.

[07:35] S13 · AI into ANIM · Anomalocaris moving through a Cambrian sea; pull back to the calendar.

So it was not an explosion out of nothing, and it was not gradual either. It was a threshold, crossed once, after which the seas contained things that hunted, saw, burrowed, and swam. And every animal since, including the one watching this, has been a variation on what was worked out in those twenty million years. Next lesson: the five times it was very nearly all undone.

Shot list

One AI shot, eleven code-rendered shots, one museum specimen image. Estimated Higgsfield use: 2 to 4 generations including retries.