A Fairly Slow Explosion

COLD OPEN

The Cambrian begins with a burrow.

There are grander ways to start an age of animals. A predator with too many teeth. The first eye opening on the first morning. Something acquiring enough legs to become a maintenance problem. Geologists went with a burrow.

Specifically, a trace fossil called Treptichnus pedum. It is a burrow, the trace of something probing through soft sediment. The animal that made it, is not preserved with it. We have only the record of its searching.

In nineteen ninety-two, the international committee responsible for such things ratified a point in the rocks at Fortune Head, on the Burin Peninsula in Newfoundland, as the official start of the Cambrian. The point sits two point four metres above the base of one layer in a rock unit called the Chapel Island Formation, which is the kind of precision you get when a committee is involved. It was chosen because that is where Treptichnus pedum first showed up. Below the line, the Ediacaran. Above it, the Cambrian. A clean beginning, marked by a burrow.

In two thousand and one, James Gehling and colleagues went back to Fortune Head and found the same burrow about four metres below the line.

The fossil that marks the start of the Cambrian had turned up before the start of the Cambrian.

The line stayed where it was. Some stratigraphers think it should be moved. So far, it hasn't been. This is less awkward than it sounds. The official boundary is defined by the marked point in the rock, not by the fossil. The fossil helped choose the point, and once the point is chosen it stays put, so that every geologist on Earth is measuring from the same place. The burrow is welcome to wander. The calendar does not follow it.

Still, it is a fitting start. Before we have met a single Cambrian animal, the paperwork is already out of order.

That is roughly the story of this whole episode. The popular version of the Cambrian explosion goes like this: an empty sea, then, suddenly, trilobites. The record turns out to be more crowded, more gradual at the edges, and a good deal more argued over than that. There were large, complex living things before the Cambrian. Some of them were probably animals. Some things were already digging. The explosion itself, measured in the fossil record, takes something like twenty million years, which is quick for a planet and a long time for a bang.

None of which makes the question go away. Something really did change. In a geologically short window, the fossil record fills up with animals built on most of the body plans that leave fossils, and a few that, thankfully, do not. Darwin saw this as a serious problem for his theory and said so in print. People have been arguing about it ever since, with considerable stamina.

In Part One we reached the Ediacaran and promised to come back. This is us coming back. Today: what was already there, what the Cambrian rocks actually contain, how sudden it really was, and what might explain it. Including one argument about the machinery that builds animal bodies, which turns out to be the most interesting argument of all.

We will try to be brief. We have been here before.

ACT ONE — BEFORE THE BANG

The sea was already occupied

Here is the claim. The Cambrian did not begin in an empty ocean. It began in an ocean that had been home to large, complex living things for tens of millions of years.

The evidence starts, again, in Newfoundland, which for a quiet province has a great deal to answer for. At a place called Mistaken Point, there are surfaces of rock covered in the impressions of soft-bodied organisms. In twenty twenty-one, Jack Matthews and colleagues published a careful dating study of the site. The oldest of what they called architecturally complex fossils there are about five hundred and seventy-four million years old.

That is about thirty-five million years before the official start of the Cambrian. The same study notes that the site holds early evidence of animal-type bodies and of animal-style movement. Whatever the Cambrian was, it did not arrive to find the house empty.

Here is the constraint. Large and complex are descriptions. Animal is a classification. Many Ediacaran organisms, including a group called rangeomorphs found at Mistaken Point, do not fit comfortably into any group alive today. An organism can leave a perfectly good fossil and still be very hard to place. Its existence is secure. Its family is a matter for discussion.

And being older is not the same as being an ancestor. As we said in Part One, this is a bush, not a ladder. A fossil from before the Cambrian may be a great-great-grandparent of something later, or a distant cousin, or a branch that went nowhere at all. A timeline tells you the order things lived in. It does not, on its own, tell you who is related to whom.

So the open question is this. Some of these earlier organisms were there. Which of them were animals, and how would you tell, if the shape alone is not enough?

Dickinsonia leaves a chemical clue

One answer is to stop looking at the shape and look at the chemistry.

The test case is Dickinsonia, a flat, ribbed Ediacaran fossil whose shape, for a long time, did not settle what it was.

In twenty eighteen, Ilya Bobrovskiy and colleagues reported ancient steroid molecules preserved with Dickinsonia fossils. The organisms and their relatives had produced cholesteroids, a type of steroid that the authors describe as a hallmark of animals. Their conclusion was that Dickinsonia was one of the earliest confirmed animals.

That is a real result, and it is a nice one, because for once the fossil's shape is not the only witness. The tag here is confident, as a description of what the study found, and the interpretation is that chemical traces point to an animal.

The constraints are about scope. A chemical signature tells you what kind of organism made it. It does not tell you how the organism grew, what it ate, or which later animals descend from it. And it applies to Dickinsonia and its close relatives, not to every strange frond on every Ediacaran sea floor. It is tempting, when one member of a confusing neighbourhood gets properly identified, to assume the rest of the street has been sorted out too. The street has not been sorted out.

So: at least one large, Ediacaran organism has a strong claim to being an animal. The open question is how many others do, and what they were doing there.

Kimberella changes departments

A second Ediacaran fossil shows how much an interpretation can move while the fossil stays exactly where it is.

Kimberella was first reconstructed as a jellyfish. Later, it was reconstructed as a box jellyfish, which is a more specific kind of jellyfish and, as rebrands go, a modest one. Then, in nineteen ninety-seven, Mikhail Fedonkin and Benjamin Waggoner described more than thirty-five new specimens from the White Sea coast of Russia. The new material, they wrote, did not support a jellyfish at all. Instead it showed a bilaterally symmetrical animal, meaning it had a left side and a right side, that lived on the sea floor under a single shell that wasn't mineralized. In many respects, they said, it resembled a mollusc.

So, over the years, Kimberella went from the jellyfish department to something closer to a slug. It was not consulted.

The careful phrase is mollusc-like. The reinterpretation itself is confident as published. Calling it a mollusc, full stop, would go further than the authors did, and that is the disputed part. But even the cautious version matters. An animal with a front, a back, and a left and right side, living on the sea floor before the Cambrian, is not what the empty-sea story predicts.

The open question is how securely we can place it. Was Kimberella an early relative of molluscs, or something that happened to arrive at a similar arrangement by a different road?

Bodies and burrows

Body fossils tell you what was there. Trace fossils tell you what was going on.

A trace fossil is a record of activity: a track, a trail, a burrow. It preserves behaviour without preserving the animal that did the behaving. That is a limitation, but it is also an advantage, because soft animals that almost never fossilize as bodies can still leave tracks in the mud, and mud is patient.

We have already met one. Treptichnus pedum at Fortune Head records an animal probing through sediment, and Gehling's discovery tells us that this behaviour was going on a few metres below the official line. Nothing obliges an animal to start digging only after the period has been named.

In twenty fourteen, Gabriela Mángano and Luis Buatois looked at the trace-fossil record across the whole Ediacaran-to-Cambrian transition. They argued that two things people tend to roll together actually happened on different schedules. A wide range of body plans and behaviours was in place by the Fortunian, the first stage of the Cambrian. But the kind of ecology typical of all the seas that came afterwards arrived only in the following stage. Bodies first, they argued. The neighbourhood got busy later.

They also treated the trace record as strong support for a real Cambrian explosion, while allowing for what they called a short time of phylogenetic fuse beforehand, meaning a brief run-up in which the family lines were already splitting. So this is not an argument that nothing happened. It is an argument that several different things happened, close together, and that the word explosion has been asked to cover all of them at once.

That is confident as a description of their reading. The open question is how those schedules connect. Did new bodies make new ecology possible, or did new ecology make new bodies worthwhile? It is possible the answer is yes.

Before does not cancel after

So here is where the first act leaves us.

Mistaken Point gives us complex life about five hundred and seventy-four million years ago. Dickinsonia gives us chemistry pointing to an animal. Kimberella gives us a mollusc-like animal with a left and a right. The trace fossils give us animals digging, and a change in the way they dug.

Taken together, they make the empty sea a poor place to start. What they do not do is explain the Cambrian. Finding predecessors changes the question. It does not answer it. We are no longer asking how an empty ocean acquired animals all at once. We are asking how an ocean that already had some animals became the very crowded ocean of the Cambrian rocks.

Which means we should probably look at the Cambrian rocks.

ACT TWO — THE RECORD

A boundary is not a birthday

Here is the claim. The start of the Cambrian is a line geologists agree to measure from. It is not the day animals were born.

The current international chart puts that line at five hundred and thirty-eight point eight million years ago, give or take about six hundred thousand years. That number is confident, as the chart's best estimate. But as we saw in the cold open, the number does not define the boundary. The marked point at Fortune Head does. The number is what that point is thought to be worth in years, and if the dating gets better, the number can change while the point stays put.

This matters because a precise-looking number invites a precise-looking story: before five hundred and thirty-eight point eight, one world, and after it, another. The rocks do not support that kind of story. The burrow below the line already tells us that animal activity did not wait for the official opening. Nor should four metres of rock be turned into a number of years. Metres and years are different measurements, and the conversion depends on how fast the mud piled up, which we are not going to guess at.

So the open question is not when the Cambrian began. That has been decided, by committee. The question is which biological changes happened before the line, which after, and how many of them we have quietly filed under one name.

Small shells, considerable consequences

The first animal fossils of the Cambrian are not the spectacular ones. They are small shelly fossils.

This is one of the great names in palaeontology, because it describes exactly what they are and promises nothing more. After a word like explosion, it is a relief. They are shells, tubes, and skeletal bits, from the very end of the Precambrian and the start of the Cambrian, and they are mostly small, as advertised.

In nineteen seventy-five, two palaeontologists, Matthews and Missarzhevsky, reviewed the work on these fossils, much of it from sections in Siberia. Two of their observations matter here. Some shelly fossils, they wrote, are found high in the Precambrian. And trilobites, the animals most people picture when they hear the word Cambrian, first appear later, at the base of a stage called the Atdabanian. So the shells came first, and the trilobites arrived after the shells had been at it for some time.

A twenty twelve review by Artem Kouchinsky and colleagues put more detail on the sequence. Sponges and some relatives of jellyfish already had skeletons before the Cambrian began. Several groups started building hard parts mostly in the Fortunian, the first Cambrian stage. Others, including the group that trilobites belong to, followed later, and starfish relatives later still. Hard parts did not arrive all at once, like a delivery. They arrived group by group, like a slow and rather disorganized delivery.

That is confident. And it matters for the argument coming in Act Four, because the shells are part of what makes the explosion look either sudden or not, depending on how much attention they get. We will come back to that.

The open question is how much had already happened before the famous animals show up, and how much of the apparent suddenness comes from skipping the small stuff.

Chengjiang

The famous animals show up, in their best-known form, in two places. The first is in southern China.

In July nineteen eighty-four, at a hill called Maotianshan near Chengjiang in Yunnan Province, the palaeontologist Hou Xianguang found a deposit of fossils with their soft parts preserved. Not just shells, but guts, legs, eyes, and the outlines of animals that had no hard parts at all. This is the Chengjiang biota.

Its age needs a little care. A twenty eighteen dating study by Chuan Yang and colleagues gives a maximum age for the deposit of about five hundred and eighteen million years. That means the fossils are no older than that. For our purposes, about five hundred and eighteen million years is the number, with the understanding that it is a ceiling.

UNESCO, which made the site a World Heritage property in twenty twelve, describes at least sixteen phyla and about one hundred and ninety-six species. A phylum is one of the big divisions of the animal kingdom, roughly one basic body plan. The earliest known specimens of several phyla are found here, including relatives of jellyfish, comb jellies, a group of burrowing worms called priapulids, and the vertebrates, the group that eventually includes us. We will come back to that, briefly, in Part Three.

Here is the important constraint. Most of these are early relatives of groups alive today, not modern animals in old rocks. The accurate summary is that most animal phyla that leave fossils have shown up by the middle of the Cambrian, many of them in unfamiliar early forms. That is confident. The popular version, in which every animal alive today appears at once, fully assembled, is myth-adjacent. The early members of a group often look quite unlike its living members, which is part of what makes these fossils hard to read.

The open question is how much of this diversity was new at the time, and how much was older diversity that had never had the luck to be buried in the right kind of mud.

Walcott and the Burgess Shale

The second place is in western Canada, and it was found first.

On the thirtieth of August, nineteen oh nine, Charles Doolittle Walcott found the Burgess Shale in British Columbia. The date comes from his field notebook, by way of the Royal Ontario Museum's history of the site.

The Burgess Shale is younger than Chengjiang. The museum gives it an age of about five hundred and five to five hundred and ten million years and notes that it cannot be dated directly. The often-quoted single figure is a midpoint, not a measurement. So it is a later snapshot of the Cambrian, by roughly ten million years.

It is two snapshots, then, not a film. Chengjiang and the Burgess Shale give us unusually detailed views of Cambrian life at two moments. Neither of them is the beginning of the Cambrian. Between and around them, most of what lived left no soft-bodied record at all.

A fossil becoming known is also not the same as an animal becoming new. Walcott found the Burgess Shale in nineteen oh nine. The animals had been there for more than five hundred million years and were in no hurry. And finding a fossil is only the start of understanding it. Some Burgess Shale animals took most of the following century to be put together the right way up. One in particular.

Hallucigenia, the right way up

In nineteen seventy-seven, Simon Conway Morris described a small Burgess Shale animal and named it Hallucigenia, a name that tells you how the description went. As reconstructed, it walked on long, stiff spines, carried a row of soft tentacles along its back, and had a round blob at one end that was taken for a head.

In nineteen ninety-one, Lars Ramsköld and Hou Xianguang, the same Hou who found Chengjiang, published a new reading based on Chengjiang fossils of related animals. Hallucigenia, they argued, was upside down. The spines were armour on its back. The tentacles were its legs. They placed it with the armoured lobopods, a group related to today's velvet worms. In nineteen ninety-two, Ramsköld reported that the legs came in pairs, the second row having been hidden in the rock. The round blob that had been the head was reinterpreted as a stain left behind as the body decayed, which meant the animal also had to be turned end to end.

So Hallucigenia had been upside down, and its head was not a head. In twenty fifteen, Martin Smith and Jean-Bernard Caron finally found the actual head, at the other end, with a pair of simple eyes and a ring of plates and teeth around the mouth.

The sequence is confident: spines down with a blob for a head in nineteen seventy-seven, turned over in nineteen ninety-one, turned end to end in nineteen ninety-two, and its real head found in twenty fifteen. In thirty-eight years, it changed orientation once and ends once.

The lesson is not that scientists are bad at fossils. It is that fossils which look baffling are sometimes baffling because we have very few of them, crushed flat, and seen from one angle. More specimens changed the picture. The open question is how much of the strangeness of Cambrian animals belongs to the animals, and how much belongs to our view of them.

Eyes enter the record

Two more animals, both with eyes, and both worth the trouble.

Opabinia, from the Burgess Shale, was studied in detail by Harry Whittington in nineteen seventy-five. It had five eyes. We have checked, and we are not going to improve on that.

Anomalocaris was reconstructed by Whittington and Derek Briggs in nineteen eighty-five, in a paper titled, without any visible excitement, as the largest Cambrian animal. It was a predator. For the details of its eyes, though, the best evidence comes from a third site, the Emu Bay Shale in Australia.

In twenty eleven, John Paterson and colleagues described fossil eyes of Anomalocaris from there, about five hundred and fifteen million years old. A single eye had at least sixteen thousand lenses, packed in a honeycomb pattern. The eye, they wrote, rivalled the most acute compound eyes in modern arthropods, the group that includes insects, crabs, and spiders.

That is confident. Sophisticated vision existed by about five hundred and fifteen million years ago. What it does not tell us is when vision began, or whether eyes caused the explosion. Those are separate claims, and we will meet the second one in Act Four.

What these animals do for us now is make the record concrete. The Cambrian explosion is not just a phrase. There are bodies, armour, eyes, revised drawings, and dates with error bars. Something really did fill the seas. The question for the next act is how fast.

ACT THREE — HOW SUDDEN

Darwin's difficulty

Before the clocks, the man who first made this a problem.

In the first edition of On the Origin of Species, in eighteen fifty-nine, Darwin raised it himself. "There is another and allied difficulty," he wrote, "which is much graver. I allude to the manner in which numbers of species of the same group, suddenly appear in the lowest known fossiliferous rocks." A page later: "To the question why we do not find records of these vast primordial periods, I can give no satisfactory answer." And a page after that: "The case at present must remain inexplicable; and may be truly urged as a valid argument against the views here entertained."

That is an unusually honest thing to put in your own book. It is also only half of what he said. In the very next sentence, Darwin offers his answer: that the earlier record is missing, not the earlier animals. So this is not a surrender. It is a difficulty, conceded in public, followed straight away by a hypothesis about it. Both halves belong in the story. Quoting the first half and stopping is a popular hobby, and we are not taking it up.

One small note on wording. Darwin did not call these rocks Cambrian. He called them the lowest Silurian, which is what the leading geologists of eighteen fifty-nine called them. The names have since been rearranged, after a quarrel between two geologists, Adam Sedgwick and Roderick Murchison, whose systems overlapped and who, by one account, ended up no longer speaking to each other. In eighteen seventy-nine Charles Lapworth sorted it out by putting a new system, the Ordovician, in the disputed middle. The rocks with that first big burst of animal fossils are now called Cambrian. The question has kept its name throughout.

Darwin's question was, roughly: where are the ancestors? We have now met some of them, or at least some of their neighbours. The next question is how long the change took.

Twenty million years, with edges

Here is the claim. In the fossil record, the Cambrian explosion takes about twenty million years.

The figure comes from Rachel Wood and colleagues, in twenty nineteen, who give the fossil-record window as roughly five hundred and forty to five hundred and twenty million years ago. Chuan Yang and colleagues, dating Chengjiang, describe the explosion as a drawn-out process lasting more than twenty-two million years, and that duration is their interpretation of the dates.

Twenty million years is a strange thing to call an explosion. It is about a third of the time since the dinosaurs, other than birds, disappeared. On the other hand, the history of life on Earth runs to something like four billion years, and against that, twenty million is a short chapter with an unusually large cast. Both descriptions are fair. It depends on which clock you are holding, and this act is mostly about which clock you are holding.

Here is the constraint. A window of twenty million years is not the same as everything starting at the beginning of it. Inside the window there are several different events: animals first appearing as fossils, shells being acquired group by group, new kinds of burrowing, new kinds of ecology. They need not start together, and they need not end together. The word explosion is a useful heading for a period of very concentrated change. It is not a measurement of any one of them.

And twenty million years tells us nothing, on its own, about whether any proposed mechanism was up to the job. A duration is a limit. It is not an explanation.

So the open question is how that visible window relates to the history of the animals in it. Did the animals begin when they first become visible, or is the fossil record showing us a later chapter of a longer story?

Old roots, quick appearance

One way to answer that is to ask the animals themselves, or rather their genes.

A molecular clock works by comparing the genes of living animals. The more two groups differ, the longer ago their lines are thought to have split, assuming changes pile up at a rate that can be calibrated using fossils of known age. Molecular clocks consistently place the splits between the major animal groups earlier than the first fossils of those groups. That is confident as a broad result. How much earlier is disputed.

In twenty eleven, Douglas Erwin and colleagues set out the picture in a paper called The Cambrian Conundrum. The major animal groups, they argued, split from one another many tens of millions of years before their first appearance as fossils. The basic genetic toolkit for building animal bodies was in place earlier still, in a cold interval called the Cryogenian, between about eight hundred and fifty and six hundred and thirty-five million years ago. And yet, in the fossil record, the groups appear apparently within a few million years in the early Cambrian.

So: deep roots, and a quick appearance. Erwin and colleagues called the later part ecological success. The family lines were old. What changed in the Cambrian was that they became abundant, large, varied, and visible.

Here is the constraint. Moving the roots back in time does not explain why the fossils arrive in a cluster. It moves the question. Instead of asking how the animals appeared so quickly, we ask what they were doing for all those tens of millions of years, and why they then became conspicuous together. Extra time can matter to an explanation. It does not explain what happened during the time.

The clocks have limits

The clocks have their own troubles.

In twenty fifteen, Mario dos Reis and colleagues took a hard look at how precisely the clocks can date the origin of animals. Earlier studies had put some of the main splits more than a hundred million years before the Cambrian. Using four different ways of calibrating against fossils, dos Reis and colleagues concluded that the group including all living animals began during the Cryogenian. But they also concluded that the uncertain nature of very old fossils, and the fact that genes do not change at perfectly steady rates, put a limit on how precise any such date can be.

That is confident, as their conclusion. It is also a good example of uncertainty being reported honestly, which deserves the courtesy of being heard correctly. It does not mean the clocks tell us nothing. The direction of the result, that the roots are older than the first fossils, is fairly consistent. What is limited is the precision. A result can point clearly in one direction and still come with a wide range.

The fossils have the same problem from the other end. A first appearance is the first appearance we have found. It is evidence that a group existed by then, not evidence that it began then.

So the honest summary is: older roots, quick appearance, and a gap between them whose size nobody can yet state exactly. The open question is what could narrow that gap, and how much of it will stay uncertain because the earliest animals were small, soft, and inconsiderate about fossilization.

A fast handover, disputed

Now for a number that goes the other way.

In twenty nineteen, Ulf Linnemann and colleagues published high-precision dates from rocks in Namibia. They concluded that the time between the last Ediacaran organisms and the first complex burrowing animals there was no more than about four hundred and ten thousand years. For geology, that is very fast. For a person waiting for it, less so.

That is their result, and it is a careful one. As a claim about the whole world, it is disputed. It comes from one composite section of rock in one region, and extending it to a global handover needs more than one place.

The same year, Rachel Wood and colleagues argued the other way. They proposed that the early history of animals should be seen as a series of successive radiations that started in the late Ediacaran and continued on into the early Palaeozoic, and that the Cambrian explosion was one phase among several. That is also disputed. Their paper's title says it challenges the Cambrian explosion, and titles that say challenge tend to get replies.

Both can be partly right. A quick handover in one place can sit inside a longer, messier story somewhere else. What the evidence does not allow is picking the most dramatic number and letting it stand for everything.

What sudden means

By now it should be clear that sudden needs an object. Sudden appearance of fossils is one question. When the lineages split is another. When ecology changed is a third. How fast one set of organisms replaced another in one place is a fourth.

Wood gives us the twenty-million-year window, and a longer series of radiations. Erwin gives us old roots and a quick appearance. Dos Reis gives us limits on what the clocks can say. Linnemann gives us a very fast handover in Namibia. Mángano and Buatois give us bodies first and ecology later.

That is less tidy than the word explosion. It is a better set of questions. There was a real burst of change, concentrated in a short span of geological time, built on roots that go back further than the fossils. What still needs explaining is why it happened when it did, and how it produced so many different kinds of body.

Which brings us to causes.

ACT FOUR — CAUSES, AND THE ARGUMENT

Oxygen and ecology

Here is the claim. There is no single agreed trigger for the Cambrian explosion. There are several proposals, and the leading ones each explain only part of it.

A useful way into the proposals comes from Erik Sperling and colleagues, in twenty thirteen. They pointed out that the two most popular kinds of explanation each do half the job. Ecological explanations, such as animals starting to eat each other and evolving defences in response, an arms race, do a good job of explaining the pattern of what evolved, but not the timing. Rising oxygen does the reverse. It can explain why something happened when it did, but not why it produced the particular variety of animals it did.

Their proposal was to tie the two together. They linked low oxygen to relatively few carnivores, and proposed that oxygen acted as a physiological control on carnivory. When there was enough of it, predators became possible, and the arms race could begin.

That is disputed as the cause. It is one proposal. It is confident as a description of the problem: timing needs one kind of explanation, and pattern needs another.

Charles Marshall, whom we will meet again shortly, reviewed the explanations in two thousand and six. His summary was that the variety of Cambrian animals is best understood as an interplay between the developmental system shared by animals with a left and a right side, which can be combined in many ways, and the growing number of ecological needs as animals began to interact. In other words, a flexible system for building bodies, and a world suddenly full of reasons to build different ones.

For anyone wanting a whole book, Douglas Erwin and James Valentine published The Cambrian Explosion: The Construction of Animal Biodiversity in twenty thirteen. We mention it as the reading list, not as a source for any particular conclusion, since we have not read it for this episode.

The open question is whether environment and ecology can account for both the timing and the variety, or whether something about the developmental system has to be part of it.

Holes in the tubes

If predation drove the explosion, it helps to know when predation started. There is a candidate from before the Cambrian.

Cloudina was a small tube-shaped animal, described as the earliest known animal to build a mineralized skeleton. In nineteen ninety-two, Stefan Bengtson and Yue Zhao examined more than five hundred Cloudina tubes and found that about two point seven percent of them had holes bored through them. The holes were size-selective, which is what you would expect if something was choosing its prey.

The authors were careful. If the holes were made by predators, they wrote, this would be the oldest case of predation known. The if belongs in the sentence. A hole is an observation. A predator is an interpretation. That is confident as a published finding, and conditional as the oldest predation.

And notice what it does to the argument. If predators were already drilling holes before the Cambrian, then predation existing cannot be the whole explanation for what happened in it. What would matter is how much predation grew, and what it did to everything else. Being present and being decisive are different things.

Parker's eyes

Vision is another candidate.

In two thousand and three, Andrew Parker published a book called In the Blink of an Eye. His idea, in brief, is that vision kick-started the explosion; later editions say so in the subtitle, How Vision Kick-Started the Big Bang of Evolution. In Parker's account, vision set off the explosion.

That is Parker's hypothesis, paraphrased, and it is disputed. It is one proposal among several.

The fossil evidence is narrower. As we heard, Anomalocaris had very sharp compound eyes by about five hundred and fifteen million years ago. That establishes that good vision existed by then. It does not establish when vision began, or that it started everything. An eye being there and an eye being the cause are different claims, which by now is starting to sound like the theme of the episode.

The open question is how much weight vision can carry. A main driver, one contributor among several, or a result of changes already under way?

The machinery

Which brings us to the developmental machinery itself, and to the most interesting argument in the whole subject.

Here is the claim. An animal's body is built during development by networks of genes switching one another on and off in a particular order, called gene regulatory networks. That is confident. Changes in how those networks are wired are how body plans change. That too is the mainstream view.

In two thousand and six, Eric Davidson and Douglas Erwin made this the centre of a paper on how animal body plans evolved. They wrote that body plan evolution must depend on changes in the architecture of these networks. They also identified a core part of each network, which they called a kernel, and described these kernels as the parts most resistant to change. They suggested this may be why basic body plans have stayed so stable: the kernels, they proposed, have been kept since before the Cambrian.

That is worth pausing on, because it is a mainstream result, and both sides of the coming argument accept that the core of today's networks resists change. They disagree about whether it was always so, and what that means for the Cambrian.

So the open question is how the networks themselves came to be, and how much they could be changed when they were young.

Meyer and Marshall

What follows is an argument about how to read the record, not a disagreement about what is in it. Both sides accept the fossils and the dates.

Stephen Meyer, a philosopher of science and a leading proponent of intelligent design, makes the strongest version of the case in his twenty thirteen book Darwin's Doubt. His argument, in summary, goes like this. New body plans need new instructions: new genes, new proteins, and new networks that switch genes on in the right order as an embryo develops. The networks are, in his view, the hardest part, because changing their core tends to break the animal rather than redesign it. Known unguided processes, he argues, cannot supply that kind of information in the time available. And in everything we have direct experience of, information of that kind comes from a mind. So he argues that design is the best explanation for the Cambrian.

That is an inference, and he presents it as one. Part of his case is a claim about what mechanisms can and cannot do, and that part is testable. The step from there to a designer is interpretation.

The palaeontologist Charles Marshall reviewed the book in the journal Science in twenty thirteen. His central reply is that new body plans came mostly from rewiring gene networks that animals already had, not from large amounts of new genes. He also argued that today's networks are a poor guide to how flexible the early ones were. Modern networks have had about half a billion years of evolution piled on top of them. The networks around when the phyla first appeared need not have been so hard to change. That is aimed squarely at the key step in Meyer's argument, the step from damage done by changes in living animals to limits on what was possible back then.

Marshall also wrote that Meyer "completely omits mention" of the early Cambrian small shelly fossils, which makes the explosion look more sudden than it was. Meyer replies that the book does mention them.

Meyer then published a four-part response to the review, disputing both the idea that early networks were more flexible and the idea that rewiring existing genes is enough. Which is roughly how book reviews go.

What would settle it

Both sides say parts of this are testable, and each says what it thinks would count.

For the design reading, the relevant evidence would be experiments showing that functional new genes, or new wiring in developmental networks, are too rare to be reached by known mechanisms in the time available. For the mainstream reading, it would be demonstrations, in living animals, that rewiring existing networks can produce large, workable changes in body form, along with earlier and simpler versions of the toolkit genes turning up before the Cambrian.

They disagree about what the tests have shown so far.

A few constraints apply to both. Showing that a change damages a modern animal does not show that the same change would have damaged an ancient one. That is Marshall's point. And showing that rewiring can change a body does not recover the actual route to any particular Cambrian animal. That is a fair point in the other direction. A possible pathway and a reconstructed history are different achievements.

Gene regulatory networks are the mainstream scientific framework here, and the mainstream view is that rewiring them is how body plans changed. Meyer offers a design interpretation, together with claims about the limits of mechanisms that he thinks the evidence supports. Each gets its best version. The historical reconstruction is not finished either way.

Two readings of the Burgess Shale

One more argument, about a different question.

In nineteen eighty-nine, Stephen Jay Gould wrote Wonderful Life, a book about the Burgess Shale. He argued that if the history of life could be run again from the Cambrian, the result would almost certainly be different, and probably nothing like us. History, in this reading, is mostly accident.

Simon Conway Morris, the same Conway Morris who described Hallucigenia, looked at the same fossils and argued the opposite, in The Crucible of Creation in nineteen ninety-eight, and more fully in Life's Solution in two thousand and three. Evolution, he argued, keeps arriving at the same solutions, eyes, brains, intelligence, again and again in unrelated groups, so something like us was likely to turn up.

Convergence itself, the same solution evolving more than once, is well established. What it implies about whether we were likely is the disputed part. Both are interpretations of the same record, and neither has been settled by it.

OUTRO

So that is the fairly slow explosion.

The sea was not empty. There were complex organisms at Mistaken Point about five hundred and seventy-four million years ago, Dickinsonia, whose chemistry points to an animal, and a mollusc-like sea-floor dweller called Kimberella. Something was digging below the official start of the Cambrian, at the very spot chosen to mark it. Then came small shells, then trilobites, then the extraordinary soft-bodied animals of Chengjiang and the Burgess Shale, some of which had to be turned over before they made sense.

In the fossils, the explosion takes about twenty million years. The family lines go back further. How much further, the clocks cannot say precisely. In Namibia, one handover may have taken only about four hundred thousand years. Whether that was true everywhere is still argued.

As for why, there are good proposals and no single trigger: oxygen, predators, eyes, and a set of developmental networks that can be rewired to build very different bodies. Whether rewiring is enough is where Stephen Meyer and Charles Marshall part company.

That leaves the question this series keeps asking: what had to happen next, and how sure are we? Here the next question is already inside the answer. The mainstream account says new animals came from rewiring an old toolkit. Which raises two things. Is rewiring an old toolkit enough, and where did a toolkit that rewires so well come from?

Next time, one group of Cambrian animals, with a stiff rod along its back, starts on the long road towards a backbone, a jaw, and eventually a face. It will be brief. It will not be brief.